INVESTIGATING TANTAL UM NITRIDE AND HEMAT ITE - CATALYST INTERFACE FOR PHOTOELECTROCHEM ICAL WATER OXIDATION By Hamed Hajibabaei A DISSERTATION Submitted to Michigan State University i n partial fulfilment of the requirements for the degree of Chemistry - Doctor of Philosophy 201 8 ABSTRACT INVESTIGATING TANTALUM NITRIDE AND HEMATITE - CATALYST INTERFACE FOR PHOTOELECTROCHEMICAL WATER OXIDATI ON By Hamed Hajibabaei This dissertation is focused on the synthesis and characterization of semiconductor materials for photoelectrochemical (PEC) water oxidation and it is comprised of two sections. In the first section, synthesis and characterization of Ta 3 N 5 thin films for photoelectrochemical water oxidation were investigated . Despite promising properties , the harsh synthesis conditions , involving high - temperature ammonolysis limits the PEC water oxidation efficiency o n Ta 3 N 5 . From synthetic point of view, this method is highly energy intensive, produces a sizable quantity of chemical waste s , inefficient on chemical utilization, and prov ides highly reducing conditions, preventing to integrate it in the tandem cell . As t he first study, the electrodeposition of tantalum oxide from aqueous solution followed by ammonolysis to synthesize Ta 3 N 5 films was investigated . This is a promising approach as it require s fairly simple instrumentation, maximizes the chemical utilization , and allows to realize Ta 3 N 5 on any conductive substrate. In order to eliminate the ammonolysis step, the atomic layer deposition (ALD) was utilized to directly deposit tantalum nitride on transparent conductive oxides (TCO) which otherwise require highly reactive (reducing) amm onolysis conditions. It was discovered that the low - temperature ALD (175 - 280 °C) only results in amorphous TaO x N y films which still need to be crystalized/ nitridized in ammonia but practically at more moderate conditions. This furt her allowed to integrate Ta 3 N 5 with a Ta - doped TiO 2 a newly developed TCO that is stable in reducing conditions - and to realize the first example of a Ta 3 N 5 electrode on TCO. Lastly, a high temperature and fully automated ALD system was designed and built to directly deposit crystalline Ta 3 N 5 on FTO. In the second section of this dissertation , hematite as another promising candidate for PEC water oxidation was investigated . The goal of this project was to understand how catalyst inte rface s with underlying semiconductor and how it affects the performance of catalytically modified electrodes. To this end , the Ni 1 - x Fe x O y catalysts with various composition were utilized to coat hematite electrodes. A combination of structural and electroc hemical techniques such as steady - state and transient photocurrent measurements, electrochemical impedance spectroscopy (EIS), intensity modulated photocurrent spectroscopy (IMPS), and dual - working electrode (DWE) measurements (in collaboration with Prof . Boettcher) were opted to elucidate the role of catalyst. The finding s from these studies are of great importance as they provide a clear pict ure of the device under operando conditions. It was found that the catalyst layer acts as a hole storage layer wher e the photogenerated holes from underlying semiconductor are stored in the catalyst layer, causing the potential of catalyst to drop until a sustainable water oxidation is achieved . In addition, it was discovered that the effect of catalyst (improving or s uppressing) on the PEC performance of the catalyst - coated electrode s strongly relates to the electronic conductivity of the catalyst and the morphology of the underlying hematite photoelectrode . For example, in case of highly conductive catalyst and porous hematite substrate (in presence of pinholes), the potential of the catalyst layer is pinned to the conductive substrate which subsequently limits the built - in po tential in the catalyst layer. iv Dedicate to my wife (Maryam), my mom and dad (Mehri & Masoud), my brother (Vahid), and my teachers for helping and guiding me here. v ACKNOWLEDGEMENTS I would like to thank many important p eople in my life. They are truly my friends and helped me every step of the way. I would like to thank my dad, Masoud, who was the one that introduced me to Chemistry and encouraged me to peruse my education in this area. He was the one pushing me forward and when I failed he was the one encouraging me to try harder next time. I can think of many things in my life that I might not have if it was not for him. But, I was not fortunate enough to have him at my side a bit longer. He passed away in March 2004. H e may rest in peace and he is always in my thoughts. I like to thank my mom, Mehri, who has been my inspiration. She is the one who taught me everything is possible if you try hard enough. She taught me the value of people and taught me to be a better vers ion of myself. I appreciate her for doing a great job of giving me and my brother the joy of a happy family. I still remember, like it happened yesterday, the September of 2004 that she traveled 2000 miles to enroll me for my undergraduate. Still, after al l these years she is one of my big fans and gives me positive energies. I would like to thank my big brother, Vahid. He has been my first friend and has been one of the most influential people in my life. He has been like a father to me and supported me all these years and I consider myself lucky to have him. I like to thank my wife, Maryam, who is the sweetest thing that ever happened to me. We came to the United States together and she has been my only family all these years. Indeed, her unconditional l ove, devotion, and supports have eased the journey of grad school and I will always vi and thank them for their support. I would like to thank my undergraduat e professor, Dr. Eskandari. He was the one who realized my interest in chemistry and encouraged me to pursue my interest and to seek higher education. I like to thank my Ph.D. advisor, Dr. Hamann. Besides being a great advisor he is also been a true friend and he has been my teacher in chemistry and life. Over the past 5 years that I have been working with him, I have learned much quality from him but the two imporrtant that always stays with me are: how to think scientifically and of course always think be fore talk. I would also like to thank the Hamann group members that I have got to know over the last five years and I really enjoyed working with them and be friends with them: Omid, Josh, Mandal, Dan, Yuan, Yuling, Youjue, Abe, Arriana, Faezeh, Austin, Pa risa, Tim, Geletue. I like to thank my Ph.D. committee members Dr. Beaulac, Dr. Jackson, Dr. Blanchard, and Dr. Smith. I also would like to thanks many great and friendly people in the chemistry department at Mic higan state university: Dr. Borhan, Dr. Staples, Morvey, Heidi, Alireza, Hadi, Behnaz, and Olivia. vii TABL E OF C ONTENT S LIST OF TABLES ................................ ................................ ................................ ....................... ix LIST OF FIGURES ................................ ................................ ................................ ...................... x Introduction ................................ ................................ ................................ ............. 1 1.1. Motivation and Approach ................................ ................................ ................................ 2 1.2. Solar Water Splitting ................................ ................................ ................................ ....... 7 1.3. Photoanode Material ................................ ................................ ................................ ....... 9 1.3.1. Tantalum Nitride ................................ ................................ ................................ .. 11 1.3.2. Hematite ................................ ................................ ................................ ............... 26 1.4. Synthesis of Semiconductor Thin Films ................................ ................................ ....... 28 1.4.1. Atomic Layer Deposition ................................ ................................ ..................... 29 1.4.2. Electrodeposition ................................ ................................ ................................ .. 31 1.5. Dissertation Overview and Objectives ................................ ................................ .......... 33 REFERENCES ................................ ................................ ................................ ....................... 35 Selective Electrodeposition of Tantalum(V) Oxide Electrodes ........................ 43 2.1. Abstract ................................ ................................ ................................ ......................... 44 2.2. Introduction ................................ ................................ ................................ ................... 44 2.3. Experimental Section ................................ ................................ ................................ .... 47 2.3.1. Electrodeposition of tantalum oxide ................................ ................................ .... 47 2.3.2. Film Characterization ................................ ................................ ........................... 48 2.3.3. Photoelectrodep osition of CoPi ................................ ................................ ............ 49 2.3.4. Photoelectrochemical Measurements ................................ ................................ ... 49 2.4. Results and Discussion ................................ ................................ ................................ .. 50 2.5. Conclusion ................................ ................................ ................................ ..................... 59 APPENDIX ................................ ................................ ................................ ............................. 61 REFEREN CES ................................ ................................ ................................ ....................... 74 Tantalum Nitride Films Integrated With Transparent Conductive Oxide Substrates via Atomic Layer Deposition for Ph otoelectrochemical Water Splitting ........... 80 3.1. Abstract ................................ ................................ ................................ ......................... 81 3.2. Introduction ................................ ................................ ................................ ................... 81 3.3. Experimental ................................ ................................ ................................ ................. 84 3.3.1. Film preparation ................................ ................................ ................................ ... 84 3.3.2. Film Characterization ................................ ................................ ........................... 85 3.3.3. Photoelectrochemical Measurements ................................ ................................ ... 86 3.4. Results and Discussion ................................ ................................ ................................ .. 87 3.4.1. ALD of TTO ................................ ................................ ................................ ......... 87 3.4.2. ALD of Ta 3 N 5 ................................ ................................ ................................ ....... 91 3.5. Conclusions ................................ ................................ ................................ ................... 98 APPENDIX ................................ ................................ ................................ ........................... 100 REFERENCES ................................ ................................ ................................ ..................... 123 viii D irect Deposition of Ta 3 N 5 via High - Temperature ALD ................................ 128 4.1. Abstract ................................ ................................ ................................ ....................... 129 4.2. Introduction ................................ ................................ ................................ ................. 129 4.3. Experimental ................................ ................................ ................................ ............... 134 4.3.1. Instrumentation ................................ ................................ ................................ ... 134 4.3.2. Deposition of Ta 3 N 5 Film ................................ ................................ ................... 144 4.3.3. Film Characte rization ................................ ................................ ......................... 145 4.3.4. Photoelectrochemical Measurements ................................ ................................ . 146 4.4. Results and Discussion ................................ ................................ ................................ 148 4.5. Conclusions ................................ ................................ ................................ ................. 163 APPENDIX ................................ ................................ ................................ ........................... 165 REFERENCES ................................ ................................ ................................ ..................... 182 Interface Control of Photoelectrochemical Water Oxidation Performance with Ni 1 x Fe x O y Modified H ematite Photoanodes ................................ ................................ .......... 187 5.1. Abstract ................................ ................................ ................................ ....................... 188 5.2. Introduction ................................ ................................ ................................ ................. 188 5.3. Experimental ................................ ................................ ................................ ............... 191 5.3.1. Electrode preparation ................................ ................................ ......................... 191 5.3.2. Deposition of Catalyst ................................ ................................ ........................ 192 5.3.3. Film Characterization ................................ ................................ ......................... 193 5.3.4. (Photo)electroc hemical Measurements ................................ .............................. 194 5.4. Results and Discussion ................................ ................................ ................................ 195 5.5. Conclusion ................................ ................................ ................................ ................... 209 APPENDIX ................................ ................................ ................................ ........................... 213 REFERENCES ................................ ................................ ................................ ..................... 235 The Role of Catalyst Composition on the PEC Water Oxidation Performance of Hematite with Different Crystal Orientations ................................ ................................ ... 241 6.1. Abstract ................................ ................................ ................................ ....................... 242 6.2. Introduction ................................ ................................ ................................ ................. 242 6.3. Experimental ................................ ................................ ................................ ............... 245 6.3.1. Electrode Preparation ................................ ................................ ......................... 245 6.3.2. Photoelectrochemical Measurements ................................ ................................ . 245 6.4. Results and Discussion ................................ ................................ ................................ 247 APPENDIX ................................ ................................ ................................ ........................... 254 REFERENCES ................................ ................................ ................................ ..................... 264 Conclusion and Future Directions ................................ ................................ .... 266 7.1. Summary and Conclusions ................................ ................................ .......................... 267 7.2. Future Directions ................................ ................................ ................................ ......... 270 REFERENCES ................................ ................................ ................................ ..................... 275 ix LIST OF TABLES Table 4 - 1. The summary of the experimental conditions and techniques utilized for deposition of thin film of tantalum nitride. ................................ ................................ ................................ ....... 133 x LIST OF FIGUR E S Figure 1 1. World CO 2 emissions from fossil fuel combustion and global atmospheric concentration adapted from the U.S. Energy Information Administration. 3 ................................ .. 3 Figure 1 2. The U.S. energy consumption in 2016 by energy sources. Graph is reproduc ed from the U.S. Energy Information Administration. 11 ................................ ................................ .............. 5 Figure 1 3. The U.S. energy consumption in 2017 by sector. Graph is reproduced from the U.S. Energy Information Administration. 11 ................................ ................................ ............................ 6 Figure 1 4. The schematic representation of PEC water oxidation in a tandem cell comprised of: 1) a p - type semiconductor as the photocathode, 2) a transparent conductive oxide - this is a degenerately doped wide - bandgap (> 3.2 eV) semiconductor - which is optically transparent and electronically conductive to produce an ohmic contact with the top layer, 3) n - type semiconductor as the photoanode. ................................ ................................ ................................ ........................... 9 Figure 1 5. a) The crystal structure of Ta 3 N 5 along the c - axis, b) the unit cell of tantalum nitride representing different types of nitrogen group and coordina tion environment around tantalum. 12 Figure 1 6. a) The calculated density of states (DOS) of Ta 3 N 5 at the experimental geometry, b) comparison between calculated and experimental spectra of the absorption coefficient of Ta 3 N 5 . These plots are reprinted with permission from ref. 20. Copyright (2014) by the American Physical Society. ................................ ................................ ................................ ................................ .......... 13 Figure 1 7. The schematic representation of band edge positions of a series of common photoanodes suitable for PEC water oxidation. The numbers on the gra ph represent the optical band gap. The vertical two - pointed arrows represent the theoretical photovoltage gained from Ta 3 N 5 and Fe 2 O 3 under water oxidation conditions. The water reduction and oxidation potentials at 0.0 and 1.23 V vs. RHE are shown with ho rizontal dashed lines. The data used for preparation of this scheme were adapted from reference 31 and 24. ................................ ............................... 16 Figure 1 8. The en ergy - level diagram of a n - type semiconductor under illumination representing charge carrier generation and different charge transport mechanisms as a function of distance from the electrolyte. ................................ ................................ ................................ ............................... 18 Figure 1 9. Schematic representation of the synthesis procedure of Ta 3 N 5 photoelectrode. ...... 21 Figure 1 10. The morphology and PEC performance of Ta 3 N 5 . a) cross section of Ba - doped Ta 3 N 5 on Ta substrate, b) the J - V curves of undoped and Ba - doped Ta 3 N 5 coated with CoPi catalyst under 1 sun illumination at pH = 13. Ada pted by permission from Springer Nature: Nature Communication (ref. 41), copyright 2013. ................................ ................................ ................... 23 Figure 1 11 . a) The schematic representation of Ta 3 N 5 electrode comprised of six layers from bottom to top as: (1) Ta (conductive substrate), (2) Ta 3 N 5 , (3) TiO x , (4) FeOOH, (5) Ni(O H) 2 , (6) a combination of molecular catalyst shown on the right hand side; b) J - V curves of the integrated xi electrode with various combination of molecular catalyst in dark and under 1 sun illumination at pH = 13.6 in 1 M NaOH. These figures were adapted fr om ref. 44 with permission from The Royal Society of Chemistry. ................................ ................................ ................................ .................... 24 Figure 1 12. Comparison between the stability of Ta 3 N 5 and Ta 3 N 5 coated GaN at pH = 13 under 1 sun illumination at 1.2 V vs. RHE. The figure is adapted with permission from ref. 48. Copyright 2017 John Wiley & Sons. ................................ ................................ ................................ ............. 26 Figure 1 13. The schematic representation of atomic layer deposition of a binary reaction. ..... 30 Figure 1 14. A typical schematic representation of electrodeposition cell in three - electrode configuration. ................................ ................................ ................................ ................................ 32 Figure 2 1. a) The CVs of solutions containing 1 M KCl (red, pH = 7.0), 1 M KCl + 0.5 M KNO 3 (black, pH = 7.0), 1 M KCl + 50 mM Ta - IPA (green, pH = 0.6), and 1 M KCl + 0.5 M KNO 3 + 50 mM Ta - IPA (blue, pH = 0.6); inset is the magnified CVs of first two solutions. b) The CV of 1 M KCl with pH of 1 adjusted by HCl (dotted orange) and 1 M KCl + 50 mM Ta - IPA (green), c) The CV of 1 M KCl + 0.5 M KNO 3 at pH 1 adjusted by HCl (dotted maroon ) and 1 M KCl + 0.5 M KNO 3 + 50 mM Ta - IPA (blue). ................................ ................................ ................................ .... 51 Figure 2 2. a) The survey XPS spectra of electrodeposited tantalum oxide fil ms (black) and precipitated powder by addition of 1 M KOH (red), b) the detailed spectra of Ta 4f, C 1s, and O 1s peaks. ................................ ................................ ................................ ................................ ........ 54 Figur e 2 3. a) SEM micro - images of as - deposited TaO x H y , the inset is the magnified image, b) the XRD patterns of electrodeposited film (top, cyan) and powder control sample precipitated by addition of 1 M KOH (bottom, dark red) after annealing in air at 850 °C for 1 hour. The vertical dotted lines represent the Bragg positions of Ta 2 O 5 with PDF # 01 - 071 - 0639 and the stars represents the diffraction lines of TiO 2 . ................................ ................................ ........................ 56 Figure 2 4. a) Top view SEM image of the film after niridization, the inset is the magnified image. b) XRD pattern of the Ta 3 N 5 on Ti foil (top, orange), Ti substrate (bottom, dark red). The vertical dotted lines represent the Bragg positions of Ta 3 N 5 (orange diamond) with PDF # 01 - 089 - 5200. ................................ ................................ ................................ ................................ ....................... 57 Figure 2 5. Plots of J - E curves for PEC water oxidation with Ta 3 N 5 prepared from electrodeposited tantalum oxy(hydrate) film followed by ammonolysis (orange) and the control electrode (dark red) coated with CoPi. Inset is a photograph of control and electrodeposited films on a Ta substrate. ................................ ................................ ................................ .......................... 58 Figure 3 1 . The tandem cell configuration for overall water splitting composed of n - type, and p - type semiconductor connected through a transparent and conductive layer. ................................ 82 Figure 3 2. a) Resistivity (brown) and atomic % of O (blue) and N (red) of Ta - doped TiO 2 as a function of Ta%. b) Optical transmittance of Ta - doped TiO 2 films, un - doped TiO 2 (orange), 5.0% Ta (pink), 2.5% Ta (Cyan), 1.67% Ta (red) and 1.25% Ta (black). c) Photograph of the TTO films after annealing in ammonia at 750 °C for 30 min. ................................ ................................ ........ 90 xii Figure 3 3. Growth rate of TaO x N y (green) and TaO x (pink) as a function of deposition temperature. Also shown are the ratios of atomic concentrations of O/N (purple) as a function of the deposition temperature found from EDS analysis. ................................ ................................ . 92 Figure 3 4. a) XPS signals of O 1s, N 1s , and Ta 4f, b) calculated atomic percentages of Ta - N (red) and Ta - O (purple) as a function of deposition temperature. ................................ ................ 93 Figure 3 5. Plots of absorbance of thin films of Ta 3 N 5 on quartz with different thicknesses: 50 nm (blue), 70 nm (orange), 99 nm (red), and 122 nm (green). Inset is the absorbance at 350 nm vs. the thickness of the films. ................................ ................................ ................................ ................... 95 Figure 3 6. PEC performance of CoPi modified Ta 3 N 5 (~ 75 nm) on TTO with 1.6% Ta concentration under 1 sun illumination. The inset is the photograph of the working el ectrode. .. 97 Figure 4 1. The building blocks of an atomic layer deposition system. ................................ ... 134 Figure 4 2. The schematic representation of the Inlet of the ALD reactor. .............................. 137 Figure 4 3. The schematic representation of different generation of ALD reactors: a) 1 st , b) 2 nd , and c) 3 rd generation. The body of all generations is constructed from stainless steel 316. ....... 139 Figure 4 4. The EDS spectra showing the composition of the films deposited on FTO substrate at 500 °C with different generations of the ALD re actor; a) 1 st , b) 2 nd , and c) 3 rd generation. The inset in the figure (a) shows the photograph of the as - deposited film along with the film annealed in air at 750 °C for 1 minutes; also shown is the comparison of the Raman spectra of the as - deposited film - Fe 2 O 3 ) film. 21 The inset in the Figure (c) shows the photograph of the film deposited from 3 rd generation reactor on the custom - built substrate heater. .............................. 142 Figure 4 5. The EDS spectra of the bare FTO (black) and t he deposited films as a function of number of cycles; 500 (red), 300 (green), and 150 cycles (blue). The highlighted graph on the right - hand side represents the peak located at 1.7 kV in a reverse order showing a progressive growth of Ta peak as the number of cycles increases. The inset represent the configuration of the film and the origin of signals. ................................ ................................ ................................ ..... 149 Figure 4 6. The XRD patter ns of the FTO (SnO 2 ) substrates before (black) and after (orange) annealing in ammonia and tantalum nitride films with various thicknesses on FTO substrate. Films with 150 cycles, 300, and 500 cycles are shown in blue, green, and red colors, respectively. The vertical black dashed lines represents the Bragg positions for the standard crystalline Ta 3 N 5 powder with random crystal orientations (PDF # 01 - 079 - 1533). The numbers represent the Miller indices of the corresponding diffraction peaks of Ta 3 N 5 . The yel low diamonds represent the Bragg positions of the standard SnO 2 (PFD # 00 - 046 - 1088). ................................ ............................... 150 Figure 4 7. The SEM and AFM images showi ng the cross - section, morphology, and topology of the bare FTO and Ta 3 N 5 films with different thickness. The scale bars for the SEM images are 200 nm with ~ 200,000X magnification. The 3D AFM images are shown in Figure A 4 - 9. ............. 153 Figure 4 8. a) Plots of the absorptance of Ta 3 N 5 films on FTO substrate with various thicknesses. The corresponding transmittances and reflectances are show n in Figure A 4 - 11. The data shown xiii here are corrected for the pristine FTO substrate using the previously reported procedure. 22 b) The thicknes s of deposited films as a function of number of cycles grown at 550 °C on FTO substrate. The thickness of the films were determined via two independent methods of cross - section SEM analysis (black square) and the optical absorbance at 450 nm (red circle) i n combination with beer - lambert law using the known absorption coefficient reported previously for compact film of Ta 3 N 5 . 23 ................................ ................................ ................................ ................................ ....... 155 Fig ure 4 9. a) The chopped light J - E curves of Ta 3 N 5 on FTO substrate as a function of film thickness in contact with an aqueous solution containing 0.1 M K 4 [Fe(CN) 6 ] as the hole scavenger with the pH of 6.8. The potential was scanned at 20 mV s - 1 and the el ectrodes were illuminated through the electrolyte with 1 sun intensity. b) The photograph of the electrode with various thicknesses of Ta 3 N 5 on FTO substrate. As shown all the edges of the rectangular shape electrode were coated with Ag epoxy and the elect rode was clamped to a costume made cell with an O - ring with diameter of 0.19 cm - 2 (the mark of the O - ring is partially visible on the 500 cycles electrode). ................................ ................................ ................................ ................................ ..................... 157 Figure 4 10. Wavelength dependence of the a) IPCE% and b) APCE% values at 1.0 V vs. RHE for the Ta 3 N 5 films with different thicknesses as a function of illumination direction in contact an aqueous solution containing 0.1 M K 4 [Fe(CN) 6 ] as the hole scavenger with the pH of 6.8. The solid and open shapes represent the values under front (through the solution) and back (through) illumination, respectively. Ta 3 N 5 films with 230, 103, and 70 nm thicknesses are shown in red diamond, green circle, and blue triangle, respectively. ................................ ............................... 160 Figure 4 11. The schematic representation of the profile of charge generation as a function of illumination direction; a) back illumination (through the FTO), b) front illumination (through the solutions). ................................ ................................ ................................ ................................ .... 161 Figure 5 1. a) IR - corrected CVs of Ni 1 - x Fe x O y catalysts on FTO substrates; FeO y (cyan), Ni 0.25 Fe 0.75 Oy (orange), Ni 0.5 Fe 0.5 O y (pink), Ni 0.75 Fe 0.25 O y (green), and NiO y (blue). b) plots of current density at 300 mV overpotential (blue diamonds) and overpotential at 10 mA cm - 2 (red triangle) as a function of the % Ni in contact with 1 M KOH; the compositions, activities, and the error bars are the average and standard de viations acquired from 3 independently prepared samples. ................................ ................................ ................................ ................................ ....... 197 Figure 5 2. J - E curves measured at 20 mV s - 1 under 1 sun illuminatio n for bare and modified a) ALD and b) ED hematite electrodes in contact with 1 M KOH solution; bare ALD (red), bare ED (blue), Ni75 coated hematite (green), Ni25 coated hematite (orange). The dark current are not included for the sake of clarity and are sh own in Figure A 5 - 8. ................................ .................. 200 Figure 5 3. a) MS plots for bare and catalyst modified ED electrodes, b) J - E curve and C ss values for ba re ED electrode under 1 sun illumination in contact with 1 M KOH; bare electrode (blue diamond), Ni75 modified ED electrode (green triangle), and Ni25 coated ED electrode (orange pentagon). The dotted lines are the linear fitting used to extract the dopant density. ................ 202 Figure 5 4. a) C Ni25 and b) C Ni75 obtained from fitting EIS for catalyst coated FTO (1cycle, shown with star symbols ) and ED electrodes with varying thickness: 1 cycle (pentagon), 2 cycles (circle), and 3 cycles (triangle). ................................ ................................ ................................ ................ 205 xiv Figure 5 5. R trap / R ct for bare and R cat /R ct for catalyst modified ED electrodes; bare electrode (blue diamond), Ni75 modified ED electrode (green triangle), and Ni25 coated ED electrode (orange pentagon). ................................ ................................ ................................ ................................ .... 207 Figure 5 6. a) variation of low frequency (LF), and b) high frequency (HF) limits of the IMPS response for bare and catalyst coated ED electrodes under monochromatic illumination (470 nm) in con tact with 1 M KOH solution; bare electrode (blue diamond), Ni75 modified ED electrode (green triangle), and Ni25 coated ED electrode (orange pentagon). ................................ .......... 209 Figure 6 1. The surface configuration of hematite as a function of crystal orientations, a) (100) and b) (001) orientations. The iron centers and the lattice oxygens are shown in orange (large circle) and red (small circle), respe ctively. The singly and doubly coordinated oxygen sites at the surface are shown in yellow and cyan, respectively. Throughout this chapter the (100) and (001) orientations are labeled as M and C, respectively. ................................ ................................ ...... 244 Figure 6 2. Comparison of PEC performance of catalytically modified hematite electrodes with various crystal orientations and different composition of catalysts. a) J - E cu rves of bare, Ni25 - (orange) and Ni75 - (green) coated hematite electrodes with different orientations, b) hole collection efficiency for bare, Ni25 - (orange) and Ni75 - (green) coated hematite electrodes. The (100) (labeled as M), and (001) (labeled as C) orien tations are shown in broken line with empty shapes and solid lines with solid shapes, respectively. Note: for clarity, the dark currents are omitted and are shown in Figure A 6 - 2. The ratio of the low frequency (LF) to high frequency (HF) of IMPS responses were used to calculate the hc %. 6,7 ................................ ................................ ............................. 248 Figure 6 3. The calculated a) R trap , R cat and b) R ct for the bare and catalyst coated hematite electrodes with different crystal orientation of hematite and varying composition of ca talyst as a function of applied potential. The C - and M - orientation are shown in solid and empty shapes, respectively. The Ni75 and Ni25 catalyst coated electrodes are shown in green and orange, respectively. ................................ ................................ ................................ ................................ 251 Figure 7 1. a) Schematic representation and photographs of the reaction products of TaCl 5 with ammonia and subsequent exposure to air and annealing in Ar, b) the XRD patter n of the powder after annealing in argon; the vertical blue dashed lines and red dashed - doted lines represent Bragg position of Ta 2 O 5 (PDF# 00 - 019 - 1299 of and Ta 3 N 5 (PDF# 00 - 019 - 1291), respectively, c) cyclic voltammograms of the electrodeposition bath con taining various components; blank is solvent only ( 1 - butyl - 3 - methylimidazolium tetrafluoroborate) (black), 30 mM TaCl 5 (red), and 60 mM NH 4 PF 6 (blue). The Ionic liquid of 1 - butyl - 3 - methylimidazolium tetrafluoroborate was synthesized using previously report ed procedure. 12 The water impurity was removed by heating at 100 °C under reduced pressure for 48 hours using Schlenk line and subsequently it was stored in the glove box with the concentration of water and oxygen at ~ 1.2 and 3 ppm, respectively. The scan rate is 100 mV s - 1 and the Pt disc electrode was used as the working electrode. ............... 272 xv KEY TO ABBREVIATIONS IPCC Intergovernmental Panel on Climate Change EIA Energy Information Administration IEA International Energy Agency PV Photovoltaic STH Solar to hydrogen E f Fermi level E FB Flat band energy L Diffusion length D Diffusion coefficient C harge carrier life time µ Charge carrier mobility T Temperature k B Boltzman constant Light harvesting efficiency Charge Separation Efficency Hole collection efficiency q Elementary charge J ph Photocurrent density Photon flux DOS Density of State Absorption coefficient xvi % A Percent of absorptance TRMC time resolved microwave conductance TAS Transient absorption spectroscopy IPCE I ncident photon - to - current - efficiency APCE Absorbed photon - to - current efficiency P mono I ntensity of monochromatic light SC Semiconductor WOC Water Oxidation Catalyst RHE Reversible hydrogen electrode MS Mott Schottky 0 P ermittivity of free space D ielectric constant of the semiconductor A Surface area of the electrode C bulk Capacitance of the space charged region V app Applied potential V FB Flat band potential V bi Built in potential N d Dopant density w Thickness of space charged region TCO Transparent conductive oxide HCE Hole collection efficiency FTO Florin doped tin oxide ITO Tin doped indium oxide xvii AZO Alumina doped zinc oxide TTO Tantalum doped t itanium d ioxide Ta 3 N 5 Tantalum n itride TiO 2 Titanium d ioxide Ga 2 O 3 Gallium o xide Ta Tantalum GaN Gallium n itride Fe 2 O 3 Iron o xide Ni1 - xFexO y Nickel i ron o xide FeOOH Iron o xy - h ydroxide SnO 2 Tin Oxide CoPi Cobalt p hosphide CH 4 Methane CO 2 Carbon d ioxide H 2 O Water H 2 O 2 Hydrogen p eroxide HCl Hydrogen c hloride KCl Potassium Chloride KNO 3 Potassium nitrate TaCl 5 Tantalum pentachloride TaBr 5 Tantalum pentabromide TaF 5 T antalum pentafluoride PDMAT Pentakis(dimethylamino)tantalum(V) xviii Ta(NEt 2 ) 5 Pentakis(diethylamino) tantalum(V) MMH Monomethyl h ydrazine (Ga 2 (NMe 2 ) 6 T ris - (dimethylamido)gallium(III) NH 3 Ammonia NH 4 PF 6 Ammonium h exafluorophosphate Pt Platinum Ti Titanium K 4 [Fe(CN) 6 ] Potassium hexacyanidoferrate(II) IPA I sopropanol DI water Deionized water IMPS Intensity modulated photocurrent spectroscopy DWE Dual - working electrode LF Low frequency HF High frequency EIS Electrochemical impedance spectroscopy CV Cyclic voltammetry SE Spectroscopic ellipsometry XRD X - ray Diffraction TEM T ransmission electron microscopy XPS X - ray photoelectron spectroscopy UPS Ultraviolet photoelectron spectroscopy SEM Scanning electron microscop y EDS Energy - dispersive X - ray spectroscopy xix AFM Atomic force microscopy NMR Nuclear magnetic resonance spectroscopy FT - IR Fourier - transform infrared spectroscopy ALD Atomic layer deposition CVD Chemical vapor deposition ED Electrodeposition DC Direct current AC Alternative current DAC Digital to analog convertor ADC Analog to digital convertor SSR Solid state relay 1 Introduction 2 1.1. Motivation and Approach The climate of our planet has changed many times throughout its history . Most of these changes are attributed to small variation s of a ffect ing the amount of solar energy striking the . 1 The chemical composition of the atmosphere surrounding our planet is another important factor that direc tly influences its climate. Many chemicals such as carbon dioxide (CO 2 ) , water (H 2 O), and methane (CH 4 ) contributes to the greenhouse effect by absorbing the infrared radiation (heat) of the solar spectrum and trapping heat in the atmosphere. By far, carbo n dioxide variations have contributed the most to the climate changes due to both natural and artificial (anthropogenic) causes. 2 CO 2 concentration for centuries have been fluctuat ing but because of the human activities in recent history ( since the beginning of the industrial revolution in the mid - 18 th century ) the atmospheric concentration of carbon dioxide has begun to r ise significantly above its historical average . While the concentration of CO 2 in the atmosphere is regulated by natural processes such as photosynthesis that can absorb some of the anthropogen ic CO 2 , the mid - 20th century CO 2 levels has steeply increased ( Figure 1 1 ) with a rate that is far beyond the capacity of such natural processes. As a result, the atmospheric concentration of CO 2 has reached an unprecedented record of ~ 400 ppm in 2017 (~ 40% increase since the 1800s ), see Figure 1 1 . 3 The deleterious effect of increased level s of atmospheric CO 2 has already influenced the global climate and biological life on earth. For example, global temperature s have increased about 1.1 °C since the late 19 th century, 4,5 the average oceanic temperatures (top 700 meters) has increased by as much as 0.17 °C since 1969, 6 and the acidity ocean ic waters at the surface has increased by ~ 30% since the industrial revolutions . 7,8 As a result of global warming , ice she ets are shrinking, 3 snow - covered lands are dimin ishing, sea level s are rising, and the C limate change is now affecting everyone around the world, with the consequences that keep getting worse every year. Figure 1 1 . World CO 2 emissions from fossil fuel combustion and glob al atmospheric concentration ada pted from the U.S. Energy Information Administration . 3 What is/are the source(s) of anthropogenic CO 2 ? To answer this question, we take the energy consumption of the United States, i.e. the largest economy in the world, as an example. Figure 1 2 shows the total energy consumption and resources of the U.S. in 2016 . As shown, the U .S. energy consumption in 2017 was 97.4 quadrillion Btu (quadrillion = 10 15 ) or 3.2 TW, which roughly accounts for 18.6% of the global energy consumption. More than 80% of this energy is supplied from nonrenewable and carbon - based energ y re sources such as petroleum , natural gas, and coal which are ultimately released as anthropogenic CO 2 into the atmosphere. In order to emphasize 4 the magnitude of the released CO 2 , we now consider the U.S. energy consumption by sector ( Figure 1 3 ). As can be seen, ~ 30% of the energy is used for transportation. The U.S. Energy Information Administration (EIA) estimates that combination of gasoline and diesel fuels accoun ts for 82% of the U.S. transportation which emitted ~ 1.5 billion metric tons of CO 2 in 2016. 9 This number alone accounts for ~ 4.2% of the global CO 2 emission in 2016 ( Figure 1 1 ). An Intergovernmen tal Panel on Climate Change (IPCC) analysis published in 2013 2 suggests that in order to limit the global warming caused by anthropogenic CO 2 emissions to less than 2 °C (relative to pre - industrial level, ~ 1860s) with an occurrence probability of 33, 50, and 66% it requires to keep the cumulative CO 2 emissions from all anthropogenic sources below 1.57, 1.21, and 1.00 trillion tons, respectively. By 2011, approximately 515 billion tons of CO 2 have been emitted, leaving the carbon budget below 500 billion tons to limit the global warming within 2 °C in most probable ca se. 2,10 These numbers and reports are overwhelming, further emphasizing the necessity for comprehensive and immediate actions to decarboni ze the global energy diet by improving the efficiencies of existing technologies and inventing alternative and renewable technologies that minimize our reliance on fossil fuels. Biomass, hydropower, geothermal, wind, and solar energy are the five major ren ewable energy sources. Among them, solar energy is the largest source of energy, and the only one exceeding the global energy consumption. In order to utilize this great source of energy, three main pathways are generally adopted: (1) solar to thermal ene rgy, (2) solar to electrical energy via photovoltaic (PV), and (3) solar water splitting which represents the conversion of solar energy into potential energy of the chemical bonds of hydrogen. The PV technology has gained a lot of attention in the past fe w years and is considered as one of the primary routes to convert solar energy in to 5 electricity. Thanks to the technological advances made in the manufacturing of crystalline silicon and improvements in the design of efficient module, the global installed capacity of PV at the end Figure 1 2 . The U.S. energy consumption in 2016 by energy sources. Graph is reproduced from the U.S. Energy Information Administration. 11 of 2016 amounted to 303 GW which has grown by a factor of > 6 since 2010. 11 In addition to the solar energy, other renewable energy sources including wind, and hydropower have made a large contribution to the production of s ustainable and renewable electricity. Based on the most recent report by International Energy Agency ( IEA ) , the global renewable electricity capacity in 2016 12 was 805 GW and is estimated to reach a record number of 1.15 TW by 2022. 6 The common issue with the direct conv ersion of solar energy to electricity, however, is the intermittent, fluctuating output power which also varies with season . Therefore, the energy storage in the form of potential energy of electrons and chemical bonds becomes progressively important, open ing up a new series o f challenges and opportunities. Solar energy can be directly stored in the Figure 1 3 . The U.S. energy consumption in 201 7 by sector. Graph is reproduced from the U.S. Energy Information Administration. 11 chemical bonds of hydrogen by splitting water into hydrogen and oxygen. Hydrogen has a large energy density (142 MJ/kg at 700 bar) 13 , it is not intermittent (it is available any time of the day and night) , and produces a constant energy output. In addition, renewable hydrogen can be coupled with many industrial productions of valuable chemicals - e.g. synthesis of ammonia, hydrochloric acid, reduction of metallic ores allowing to decarbonize these proces ses and moving toward 7 sustainability. The main focus of this thesis is on solar water splitting via semiconductor light absorbers and how the electrocatalyst interface with the underlying semiconductor. 1.2. Solar Water Splitting The water - splitting reaction is an uphill reaction with a change in standard Gibbs free energy of 238 kJ mol - 1 or 1.23 e V (equation 1 - 1 ). The water splitting reaction can be simplified as two half - reactions (water reduction and oxidation reactions) described by equation 1 - 2 and 1 - 3 : 1 - 1 1 - 2 1 - 3 In solar water splitting applications, the energy of this endothermic reaction is provided by the energy of photons from sunlight. Semiconductors films are often used as the light absorber and photocatalyst and are therefore the heart of many solar water s plitting schemes. The solar water splitting reaction, then, can be driven by either (1) a single semiconducting light absorber to power the reaction according to equation 1 - 1 , or (2) photoelectrochemical (PEC) cell that is comprised of two physically separated semiconductors to divide the overall water splitting reaction into two half - reactions according to equation 1 - 2 and 1 - 3 . It is important to remember that the standard Gibbs free energy of the reaction is a state f unction, therefore regardless of the pathway, the overall energy balance of the reaction is ideally constant. Depending on the pathway, however, the overall solar conversion efficiency can be drastically different. In the first scheme, the semiconductor mu st 8 the thermodynamic energy plus activation energy (discussed in next section) - and band edge potentials that straddle both water oxidation and reduction reactions. These requirements greatly limit the selection of materials with the right band gap and band edge positions. In addition, in this scheme, the oxygen and hydrogen gases are generated in the same locale which further limits overall solar conversion efficiency by requiring extra energy to separate and purify hydrogen and by promoting the reverse recombination reaction of hydrogen and oxygen, i.e. hydrogen oxidation and oxygen reduction. Moreover, the recombination reaction of hydrogen and oxygen is extremely exothermic, leading to potentially hazardous conditions. Alte rnatively, in the second scheme, the water oxidation and reduction reactions are physically separated. This readily relaxes the constrain t s of a single light absorber. The most efficient PEC cell to drive solar water splitting to date has a tandem cell co nfiguration. 14, 15 The cell was comprised of a p - type semiconductor wirelessly connected to a n - type semiconductor via a transparent conductive layer ( Figure 1 4 ). In this scheme, the n - type semiconductor (photoanode: drives the water oxidation reaction via energetic photogenerated holes) has a wider band gap than the p - type semiconductor (photocathode: drives the water reduction reaction via energetic photogenerated electrons). The light is therefore shone through the photoanode where the sub - bandgap photons (not absorbed by photoanode) are transmitted to the photocathode. The transparent conductive layer is, therefore, the crucial component o f the tandem cell. This type of the cell is additionally advantageous as it allows researchers to independently study and optimize each half - cell. The multi - electron transfer nature of the water oxidation reaction (equation 1 - 3 ), makes this half - reaction sluggish, and generally limit the overall efficiency of water splitting reactions. Consequently, the majority of efforts in this area have focused on the water oxidat ion 9 Figure 1 4 . The schematic representation of PEC water oxidation in a tandem cell comprised of: 1) a p - type semiconductor as the photocathode, 2) a transparent conductive oxide - this is a degenerately doped wide - bandgap (> 3.2 eV) semiconductor - which is optically transparent and electronically conductive to produce an ohmic contact with the top layer, 3) n - type semiconductor as the photoanode. reaction. This dissertation is therefore focused on inves tigating photoanode material for PEC water oxidation. 1.3. Photoanode Material The photoelectrochemical reaction mediated by a semiconductor is a chain process initiated by light absorption and charge carrier generation (i.e. photogenerated electron and holes) followed by charge separation, charge diffusion, and ultimately charge tra nsfer at the surface of the electrode. 10 Therefore, the photocurrent response of a photoanode for PEC water oxidation reaction is controlled by three efficiencies summarized in equation 1 - 4 : 1 - 4 In this equation, q is the elementary charge, is the photon flux, , , and represent the light harvesting efficiency (LH), charge separation efficiency (CS), and hole collection efficiency (HC), respectively. The latter is basically determined by the catalytic activity of the surface for charge transfer. The first two efficienci es depend on the bulk characteristics of the semiconductor (absorption coefficient, dielectric constant, charge carrier lifetime , charge carrier diffusion length, and charge carrier mobility), and are defined by its composition, structure , and the syntheti c conditions used. Throughout this thesis we utilized atomic layer deposition (ALD), electrodeposition or electroplating (ED), and spin coating to prepare thin films of semiconductor on various substrates. In practice, a ny given photoelectrode must meet se veral criteria to be suitable for efficient PEC water oxidation: (1) its valence band (VB) edge position must be positive compared to the electrochemical potential of water oxidation ( 1.23 V vs. RHE) this is the thermodynamic requirement wh ich is required to ensure that the reaction between photogenerated holes in the VB and water is thermodynamically feasible. (2) The ban d gap of semiconductor should be narrow to allow for maximum overlap with the solar spectrum. (3) In addition to the favo rable bulk properties such as absorption coefficient, charge carrier lifetime , charge carrier diffusion length and mobility, it must have a catalytically active surface toward water oxidation to maximize the charge separation on surface and charge transfer to solution. (4) It also needs to be stable under PEC water oxidation conditions, i.e. under illumination in neutral to basic aqueous 11 solutions. Due to their high stability in neutral and basic solutions - e.g. TiO 2 , Fe 2 O 3 , WO 3 , BiVO 4 , and CuWO 4 - metal oxide semiconductors are promising candidates for PEC water oxidation reaction. On the other hand, non - oxide semiconductors such as metal nitrides, oxynitrides, and phosphides - such as Ta 3 N 5 , TaON, GaN, and GaP suffer from surface corrosion and instability under PEC water oxidation conditions. For example, the photogenerated holes in the VB of Ta 3 N 5 are sufficiently energetic to oxidize the surface nitride groups to dinitrogen forming the wide bandgap tantalum oxide on the surface (equation 1 - 5 ). The instability of tantalum nitride under PEC water oxidation reaction is thus one of its drawbacks that halts efficient water oxidation and is discussed in sectio n 1.3.1.6 . 1 - 5 As for any technology, the large - scale production the photoanode requires the use of earth - abundant , cheap, and non - toxic elements. Based on these criteria, tantalum nitride (Ta 3 N 5 ) and hematite (Fe 2 O 3 ) stand out as promising candidates for PEC water oxidation. In the following sections the pr omises and challenges of these materials is discussed. 1.3.1. Tantalum Nitride 1.3.1.1 Crystal Structure Tantalum (V) nitride (Ta 3 N 5 ) crystalizes in the orthorhombic structure with the space group of Cmcm as shown in Figure 1 5 . 16 As shown, the structure is comprised of the corner - shared irregular octahedral N groups with Ta atom at the center. Two types of N sites with four (type 1: shown as N1 and N3) and three (type 2: shown as N2 in Figure 1 5 .b ) neighboring Ta can be 12 realized. Tantalum in octahedral coordination is therefore coordinated by two N (type1) and four N (type2) with the Ta - N bond length ranging from 1.96 to 2.24 Å. Base d on this crystal structure, a total number of 24 phonon modes with the symmetry of A g and B g have been theoretically (DFPT/PBE) predicted for Ta 3 N 5 . 17 However some of these peaks hav e not been observed experimentally. 18,19 Figure 1 5 . a) The crystal structure of Ta 3 N 5 along the c - axis, b) the unit cell of tantalum nitride representing different types of nitrogen group and coordination environment around tantalum. 1.3.1.2 Optical Properties of Ta 3 N 5 Theoretical methods have been utilized to assess the electronic structure of Ta 3 N 5 . 20 22 For example, Galli and coworkers 20 studied the optoelectronic properties of tantalum nitride via ab initio calculations and ellipsometry measurements. The calculated density of states (DOS) for Ta 3 N 5 at the experimental geometry is shown in Figure 1 6 .a. As depicted, the top of the val e nce 13 Figure 1 6 . a) The calculated density of states (DOS) of Ta 3 N 5 at the experimental geometry, b) comparison between calculated and experimental spectra of the absorption coefficient of Ta 3 N 5 . These plots are reprinted with permission from r ef. 20 . Copyright (2014) by the American Physical Society. band (VB) and the bottom of the conduction band (CB) are primarily comprised of N 2p and Ta 5d orbitals, respectively. 22,23 UV - vis measurements on both powd er and thin - films of Ta 3 N 5 showed an absorption edge of ~ 600 nm corresponding to an optical band gap of ~ 2.1 eV ( Figure 1 6 . b). 18,20,24 In addition, the absorption spectrum exhib its two prominent features with absorption edges occurring at ~ 600 and ~ 550 nm ( Figure 1 6 . b). To assess the origin of these transitions (direct or indirect), Galli and coworkers 20 used the Bethe - Salpeter (BSE) equation to calculate the absorption spectra of T a 3 N 5 ; the importance of this equation is that it does not include the phonon - assisted electronic transition. The experimental and calculated spectra of Ta 3 N 5 are compared in Figure 1 6 . b. As depicted, the 14 calculated spectrum is in good agreement with the experimentally measured spectrum. Therefore, these electronic transitions are direct electronic transition originating from the electron transition fr om N 2p to Ta 5d orbitals ( Figure 1 6 .a ) . In addition to the band gap electronic transitions, a sub - band gap absorption centered around 720 nm has repeatedly been observed. 25 29 This absorption feature is assigned to two distinctive sites: (1) reduced Ta 5+ sites, i.e. Ta 4+ or Ta 3+ , 25,29 and (2) N - vacancies. 26,28 The main difference between these two sites is that the former is basically a trap state where the charges are localized on the Ta - sites while in the second scheme the charge is delocalized over the conduction band. Moreover, the locatio n of these sites, located on the surface or in the bulk, remains an open question. If this is a property of bulk, a high carrier density and low resistivity is expected. These are the physical properties that can be experimentally measured ( see section 1.3.1.4 ). On the other hand, if these sits are dominantly located on the surface, they influence the dynamic of charge at the surface (surface charge t ransfer and surface electron - hole recombination) and/or energetic at the surface (fermi level pinning) which influence the photocurrent onset potential. Therefore, the origin and location of the sub - band gap absorption features and their effect on the PEC performance of Ta 3 N 5 are important questions yet to be addressed. 1.3.1.3 The Band Edge positions in Ta 3 N 5 The Mott - Schottky plot and ultraviolet photoelectron spectroscopy (UPS) are among the most versatile techniques to determine the flat band energy ( E FB ) or the F ermi level ( E f ) of a semiconductor. When coupled with optical measurements, the band edge positions of the semiconductor can be evaluated. In photoelectrochemistry the band edge positions are the important characteristics of a semiconductor that essentially determine the feasibility of the 15 electrochemical reaction at the surface and the magnitude of the photovoltage gained from the photoelectrochemical device. For a n - type semiconductor the photogenerated holes (the minority charge carriers locate d in the valence band) drive the oxidation reaction at the surface of the photoanode. The valence band position of a semiconductor with respect to the potential of a hole acceptor in solution is therefore the thermodynamic quantity that determines the feas ibility of the oxidation reaction. For example, for PEC water oxidation the valence band position of the semiconductor must be more positive than the water oxidation potential. The band edge positions of a series of common photoanodes as well as the oxidat ion and reduction potential of water are compared in Figure 1 7 . As can be seen, the valence band edges of these semiconductors are located at positi ve potential with respect to the water oxidation potential with decreasing overpotentials from Ta 2 O 5 (~ 2 V) > Fe 2 O 3 (~ 1.2 V) > TaON (~ 1 V) > Ta 3 N 5 (~ 500 mV). The magnitude of the difference between the water oxidation potential and the valence band edg e position is another important factor to be considered. This value defines the overpotential (kinetic energy) available to initiate the electrochemical reaction. The minimum overpotential for water oxidation for most of the metal oxides ranges between 30 0 to 500 mV. 30 Therefore, for efficient photoelectrodes the difference between the valence band edge position and water oxidation potential should ideally lie within the same range. The valence band edge position of Ta 2 O 5 and Fe 2 O 3 for example, are ~ 2.0 and ~ 1.2 V mor e positive than the water oxidation potential, respectively. The photogenerated holes in these semiconductors are more energetic than the overpotential required to oxidize water. As a result, the excess energy is lost as heat which further limits the effic iency of the photoelectrode. On the other hand, the valence band edge position of Ta 3 N 5 is located ~ 1.6 - 1.7 V vs. RHE which lies within the ideal potential for water oxidation reaction, making it a promising candidate for efficient PEC water oxidation. 24 16 Figure 1 7 . The schematic representation of band edge positions of a series of co mmon photoanodes suitable for PEC water oxidation. The numbers on the graph represent the optical band gap. The vertical two - pointed arrows represent the theoretical photovoltage gained from Ta 3 N 5 and Fe 2 O 3 under water oxidation conditions. The water reduc tion and oxidation potentials at 0.0 and 1.23 V vs. RHE are shown with horizontal dashed lines. The data used for preparation of this scheme were adapted from reference 31 and 24 . The photovoltage gained from a photoanode is the difference between the electrochemical potential of semiconductor, i.e. E FB , and the potential of donor (reductant). For a given solution this potential is constant, and the magnitude of the photovoltage is thus linearly related to E FB . The E FB for a n - type semiconductor is located more positive than the conduction band edge position. This depends on the density of states of the conduction band and the dopant density of the 17 semiconductor. As depicted in Figure 1 7 , the conduction band edge position of Ta - based semiconductors are relatively similar and are located negative of the water reduction potential while the conduction band edge position of hematite is about 500 mV more positive than the water reduction potentials. Under identical conditions, therefore it is theoretically expected that Ta - based semiconductors produce a larger photovoltage than hemati te electrodes. Accordingly, a photovoltage greater than 1.23 V can theoretically be gained from Ta 3 N 5 . 1.3.1.4 Charge Transport Properties of Ta 3 N 5 In photoelectrochemical devices the majority of the light absorption occurs deep in the bulk of the semiconductor (d epending on the absorption coefficient), while the charge transfer occurs at the surface. Therefore, the photogenerated charges must be transported to the surface. The schematic representation of charge transport in an n - type Ta 3 N 5 as a function of distanc e from the electrolyte is shown in Figure 1 8 . As shown, the photogenerated charges are transported via (1) diffusion and (2) drift. 32 Diffusion refers to the net movement of charge carriers as a result of the gradient in charge carrier concentration in the absence of electrical fields and thus it is the main mechanism of charge tra nsport in the neutral region (bulk) of the semiconductor. The diffusion length and charge carrier life time are two of the fundamental characteristics of a semiconductor that controls its charge separation efficiency. The diffusion length ( L ) is the mean d istance that a carrier moves between generation and recombination and is mathematically defined as: 1 - 6 18 D (cm 2 s - 1 ) and (s) are the diffusion coefficient and the charge carrier life time. The diffusion coefficient relates to mobility of the charge carrier according to Einstein relationship as: 1 - 7 In this equation, µ (cm 2 V - 1 s - 1 ) is the charge carrier mobility, (J K - 1 ) Boltzmann constant, T (K) temperature, and q (C) is the elementary charge. 33 Figure 1 8 . The energy - level diagram of a n - type semiconductor under illumination representing charge carrier generation and different charge tran sport mechanisms as a function of distance from the electrolyte. 19 Consequently to determine the diffusion length, the carrier mobility and lifetime need to be determined. The four - probe (Van der Pauw method) measurement coupled with Hall effect measurements are widely used methods to determine the charge carrier density and mobility. 34 It should be bore in mind that these methods are informative when the characteristics of majority charge carriers are concerned. In order to determine the lifetime of the photogenerated charge carrier, various methods such as transient absorption spectros copy (TAS) and time resolved microwave conductance (TRMC) have been used. 34 36 The latter measurement (TRMC) is appealing as it can be used to simultaneously determine the charge carrier life time and mobility. Alternatively, the space - charge mode l is another approach that have been widely used to extract the diffusion length of minority charge carrier through fitting. 37,38 Recently, Van de Krol and coworkers 35 utilized TRMC to extract the mobility and lifetime of the photog enerated charge carriers in Ta 2 O 5 , TaON, and Ta 3 N 5 . They showed that Ta 3 N 5 exhibits a longlife time of 1.6 ms with a moderate mobility of 0.08 cm 2 V s - 1 corresponding to an exceptionally large diffusion length of ~ 18 µm. In a separate study, a combination of TAS and Hall Effect measurements were utilized to assess the lifetime and mobility of the compact thin films of Ta 3 N 5 with various thicknesses. 39 A mobility of ~ 4 cm 2 V s - 1 and an average life time of 5 picosecond were measured. As mentioned earlier, for Ta 3 N 5 as an n - type semiconductor, the Hall Effect measurements results in the mobility of the majority carriers, i.e. electrons, therefore the diffusion length of holes cannot be calculated. Clearly, these two independe nt studies suggest drastically different charge carrier life time and mobility. This further indicates that the bulk properties of Ta 3 N 5 are poorly understood and its charge carrier lifetime and diffusion length remain elusive. In addition, it is not clear that the transport of which charge carrier, i.e. electrons or holes, is limiting. 20 1.3.1.5 Synthesis of Ta 3 N 5 Ta 3 N 5 (films and powder) is commonly synthesized via ammonolysis crystallization and n itridization by annealing in ammonia - of Ta 2 O 5 (tantalum oxide) at high temperatures (> 850 °C) for a prolonged durations (from 8h to 120h). 40 42 Interestingly, it has been shown t hat regardless of ammonolysis conditions (temperature or duration), all the samples contained oxygen impurity - even after annealing at 900 °C for 120h. 40 Expectedly, they showed that as the annealing temperature and duration (but to less extent) increases more nitrogen is incorporated into the structure and the oxygen contents decreases. However, the rate at which the ratio of N:O increases was less than the theoretical ratio of N/O (0.66, for Ta 3 N 5 to Ta 2 O 5 that corresponds to two N 3 - for every three O 2 - sites) which was found to decrease as the ammonolysis temperature increases. The discrepancy between the experimental and theoretical ratio, therefore, was ascribed to the formation of anion - vacancy where its density increases with ammonolysis temperature. Formation of these vacancies can explain the observed sub - band gap absorption at ~ 720 nm (see section 1.3.1.2 ). In a separate study, Terao 43 utilized electron diffraction to explore the structural and compositional evolution of Ta 3 N 5 film by heating in vacuum (10 - 5 Torr) at the temperature range of 1100 - 1800 °C. It was discovered that tantalum nitride has a dynamic structure and compos ition which varies with temperature (equation 1 - 8 ). With increasing temperatures, nitrogen is successively pulled out of the structure which it is accompanied by concomitant reduction of Ta - resulting in different tantalum nitride. 1 - 8 The common synthesis procedure of Ta 3 N 5 electrode start with the oxidation of Ta - substrate via electrochemical anodization or simply heating in air to oxidize the top Ta layer to a mesoporous/ 21 nanostructure or planner tantalum oxide film, respectively . 26,27,41,44,45 Subsequently, the oxidized sample is nitridized in a flow of ammonia at elevated temperatures (850 - 1000 °C) for a prolonged period of time (2 - 15 hours). T he Ta - substrate serves as the source of Ta, the substrate, and the conductive layer to collect the majority charge carriers necessary to fabricate electrode. Figure 1 9 . Schematic representation of the synthesis procedure of Ta 3 N 5 photoelectrode. Although this method is simple and resulted in the best PEC water oxidation performance of T a 3 N 5 , 41,44 vide infra , from synthetic point of view ho wever, multiple draw backs makes this method unsuitable to realize efficient photoelectrodes as this method: (1) is highly energy intensive, (2) produces a sizable quantity of chemical waste s , (3) is inefficient on chemical utilization, (4) provides highly reducing conditions which limits Ta 3 N 5 to be only compatible to Ta - substrate or Nobel metals, e.g. Pt, 26,45 (5) precludes its application in the tandem cell (see Figure 1 4 ) as Ta or other metal substrates are not transparent to the sub - bandgap photons, (6) results in the formation of electronic resistive phases at the Ta 3 N 5 |Ta junction which further limits the electron collection efficiency. In addition, Ta as a substrate becomes very brittle after heating in ammonia which makes the post - annealing electrode processing very challenging. For example, Jaramillo and coworkers 27 studied the structure and phase transformation of tantalum oxide to tantalum nitride films on Ta foil and fused silica as a function of ammonolysis temperature (850 - 1000 °C). They showed that the synthesized Ta 3 N 5 films on fused silica are pure, but the films prepared on the Ta substrate contained impurity phases where the formation of Ta 2 N and Ta 5 N 6 phases are favored as the ammonolysis temperature increase. A similar behavior was further 22 observed by Van de Krol and coworkers. 26,35,45 They explored the effect of ammonolysis duration on the optical and PEC performance of Ta 3 N 5 photoanode prepared on Pt conductive substrate via a three steps method of: (1) sputtering of Ta on Pt (or fused silica), (2) oxidation of Ta in ai r, and (3) ammonolysis. Interestingly, they did not detect no N - poor phases. As confirmed by the depth profile GIXS measurements 27 , these observations indicate that the N - poor resistive phases are s pecifically formed at Ta 3 N 5 |Ta junction. In a separate study, Domen and coworkers 41 showed that ammonolysis of anodized Ta - substrate results in the formation of N - poor phases, e.g. Ta 5 N 6 , which suppress the elec tron collection efficiency at the back contact. Interestingly, doping with barium excludes the formation of Ta 5 N 6 which it substantially improves the PEC water oxidation performance of the electrode (discussed in the next section). These examples highlights the complexity of the synthesis of Ta 3 N 5 electrodes that requires a precise control over temperature, atmosphere, and duration of ammonolysis. These findings are extremely important as the composition, density, and distribution of vacancies (dop ant) ultimately define the dynamic of charge carrier within and on the surface of semiconductor. To rationalize the fundamental characteristics of Ta 3 N 5 , therefore, a synthesis method where the density of these vacancies can be controlled is highly desirab le. One of the promising approach to circumvent this issue is to directly synthesize Ta 3 N 5 from oxygen free precursors by eliminating the ammonolysis step via vacuum deposition techniques at moderate deposition temperatures. 1.3.1.6 PEC Performance of Ta 3 N 5 One of the best example of efficient water oxidation on Ta 3 N 5 was introduced by Domen et al . 41 The two step procedure of mask anodization followed by ammonolysis at 1000 °C for 2h was utilized to fabricate a high aspe ct ratio films of pristine (undoped) and Ba - doped Ta 3 N 5 on Ta 23 substrate. The morphology along with the PEC water oxidation performance of these electrodes ( J - E curves) are shown in Figure 1 10 . As shown, the mask anodization resulted in the well - defined vertically oriented nanorods of Ta 3 N 5 with a diameter and length of ~ 60 and 600 nm, respectively. The light J - V measurement of un doped Ta 3 N 5 produced a current density of ~ 5 mA cm - 2 at the thermodynamic water oxidation potential. Upon doping with barium, remarkably, a substantial improvement in both photocurrent density and photocurrent onset potential were observed. In line with c onductivity measurements, this improvement was further ascribed to selective suppression of the formation of resistive interlayer at Ta 3 N 5 |Ta junction (see section 1.3.1.5 ). Figure 1 10 . The morphology and PEC performance of Ta 3 N 5 . a) cross section of Ba - doped Ta 3 N 5 on Ta substrate, b) the J - V curves of undoped and Ba - doped Ta 3 N 5 coated with CoPi catalyst under 1 sun illumination at pH = 13. Adapted by permission from Springer Nature: Nature Communication (r e f. 41 ), copyright 2013. The state - of - the - art Ta 3 N 5 photoanode was introduced by Li and coworkers 44 with the photocurrent density approaching its theoretical limit (~ 12.5 mA cm - 2 ). They synthesized the porous Ta 3 N 5 film 24 via ammonolysis of NaTaO 3 film formed by anodization and hydrothermal process on Ta substrate. To promote the charge collection efficiency at the surface of electrode several layers including a thin layer of TiO x as an electron blocking layer to suppress the electron hole recombin ation, a hole storage layer comprised of FeOOH and Ni(OH) 2 to further facilitate the charge separation efficiency at the surface, and finally a layer of molecular catalyst to promote the kinetic of water oxidation reaction at the surface of electrode were successively deposited on the surface (shown in Figure 1 11 .a). Accordingly, the integrated electrode with a combination of molecular water oxidation catalyst at the surface produced 12.5 mA cm - 2 at water oxidation potential with a current onset potential at ~ 0.65 V vs. RHE (shown in Figure 1 11 .b ). Figure 1 11 . a) The schematic representation of Ta 3 N 5 electrode comprised of six layers from bottom to top as: (1) Ta (conductive substrate), (2) Ta 3 N 5 , (3) TiO x , (4) FeOOH, (5) Ni(OH) 2 , (6) a combination of molecular catalyst shown on the right hand side; b) J - V curves of the integrated electrode with various combination of molecular catalyst in dark and under 1 sun illumination at pH = 13.6 in 1 M NaOH. These figures were adapted from r ef. 44 with permission from The Royal Society of Chemistry. While these examples highlight the potentials of Ta 3 N 5 as a promising candidate for PEC water oxidation, the instability under PEC water oxidation conditions is one of the immense challenges 25 limiting the efficiency its efficiency. It is generally accepted that surface oxidation - formation of tantalum oxide or TaO x N y (equation 1 - 5 ) - induced by photocorrosion is the dominant cause of instability. The surface oxide layer lowers the PEC performance by blocking hole transfer to the solution (tantalum oxide is a highly resistive layer with a more positive valence band position with respect to Ta 3 N 5 see Figure 1 7 ), introducing surface trap states that promote surface recombination, and/or by pinning the fermi l evel that lowers the built - in potential. For example, Wang and coworkers 42 utilized a combination of electrochemical, TEM, and ambient pressure XPS spectroscopy measurements to investigate the effect of the surface oxidation on the PEC performance of Ta 3 N 5 . On the basis of their findings, they proposed that the formation of surface oxide layer is self - limiting, and it causes a severe fermi level pinn ing that drastically diminishes the photoactivity of Ta 3 N 5 . Furthermore, to assess the origin of instability they evaluated the PEC in presence of various hole scavengers, e.g. H 2 O 2 , Fe II (CN)/Fe III (CN) 6 . It was shown that Ta 3 N 5 exhibits a relatively stable performance in presence of fast hole scavengers. This observation indirectly suggests that Ta 3 N 5 has a relatively low catalytic activity toward water oxidation reaction and accumulation of holes on the surface is the main cause of surface oxidation. Thus, the catalytic modification of surface is an effective strategy to suppress the photo - induced surface oxidation. In addition, the surface encapsulation or protection is another viable strategy that has widely been used to protect the surface of photoelectr odes against oxidation. 41,42,44,46 48 For example, recently Domen and coworkers 48 explored the PEC performance of Ta 3 N 5 modified with 50 nm of GaN (surface protection layer) and a layer of CoPi catalyst. The schematic configuration of the electrode and its stability under PEC water oxidations are summarized in Figure 1 12 . As depicted, without the protection layer where the surface is only modified with CoPi catalyst, the performance of electrode is diminished within ~ 1 h of continuous illumination. 26 Figure 1 12 . Comparison between the stability of Ta 3 N 5 and Ta 3 N 5 coated GaN at pH = 13 under 1 sun illumination at 1.2 V vs. RHE. The figure is adapted with permission from r ef. 48 . Copyright 2017 John Wiley & Sons. Remarkably, when the surfaces are protected with GaN layer and CoPi catalyst, it exhibited a stable performance for 10 h continuous illuminations which to date it is the most stable Ta 3 N 5 photoelectrode under PEC water oxidation conditions. 1.3.2. Hematite Hematite ( - Fe 2 O 3 ) stands out as another promising photoanode material for PEC water oxidation. Hematite exhibits outstanding stability in a wide pH range from pH 7 to 14, which is favorable fo r oxygen evolution reactions. In addition , Fe is one of the most abundant element s continental crust after O, Si and Al. 49 It also possesses a reasonably small bandgap of ~ 2.1 eV which has a large overlap with the solar spectrum . With the absorption edge extended to 590 nm it can produce a theoretical maximum photocurrent density of 12.5 mA cm - 2 which corresponds to 15% solar - to - hydrogen (STH) efficiency. In addition, the va lence band edge position of - Fe 2 O 3 is sufficiently positive, making it suitable for water oxidation ( Figure 1 7 ). The first example of 27 photolysis of water with hematite was introduced in 1976. 50 Since then, a progressively large number of studies were conducted to identify and improve the bottleneck processes limiting the efficiency of water oxidation on hematite. Despite the remarkable improvements of t he PEC performance of hematite made over the past decade, the solar energy conversion efficiency of hematite is still far from its theoretical value . H ematite has a relatively long absorption depth but a short minority carrier collection length. 51 56 In order to realize a high ly efficient hematite photoanode, therefore, both the LHE and charge separation efficiencies need to be simultaneously maximized . Nanostructuring and doping with foreign elements are th e two well - studied strategies to optimize LHE and charge separation efficiency. 57 62 Although these strategies are highly effective to improve the performance of hematite, 63 65 they are found to have less effect on HCE at the surface of the electrode which is mainly controlled by the surface properties of the semiconductor and it depends on the surface terminal atoms and surface crystal orientations. 66 69 Hematite surface has a low catalytic activity toward water oxidation , thus hole transfer to the acceptor species in solution ( injection ) compete s with the parallel electron - hole recombination reaction on the surface . To suppress the surface recombination, a large overpotential with respect to the flat band potential is thus necessary to achieve a sustainable water oxidation. 41 To realize a high - performing hematite photoanode , it is important to minimize th is overpotential by promoting the kinetic s of water oxidation and /or suppressing the unfavorable surface ele ctron - hole recombination . S urface modification with various catalytic and non - catalytic materials have been shown to be effective strateg ies to impro ve the performance of hematite. 31,70 74 Al though the true mechanisms that these surface modification s reduce the overpotent ial are not fully understood , the common observation among these studies indicates a strong correlation between the concentration of long - lived photog enerated holes at the surface , i.e. reduced elec tron - hole recombination, and 28 the water oxidation overpotential. 71,75 77 T he role of cat alyst in PEC systems can be rationalized by two completely distinctive mechanisms: (1) the catalyst behave as an spectator where it improves the PEC performance of the photoelectrode by suppressing the electron - hole recombination at the surface, and (2) th e catalyst is as a hole storage layer and behave as an adoptive junction where it improve the performance of electrode by promoting (catalyzing) the water oxidation reaction. Additionally, there are some examples in literature where the surface modificatio n of photoanode with catalyst, surprisingly, have shown to suppress the performance of the electrode. 77,78 T herefore, it is highly beneficial to uncover the role of catalyst and the mechanism at which it affect the performance of photoelectrode. 1.4. Synthesis of Semiconductor Thin Films The art of material synthesis holds fundamental importance for both basic science and practical applications. Unlike molecular systems, t he physical characteristics of materials are fundamentally linked to the synthetic conditions used to prepare them. In addition, it is important to note that most materials that are commonly used in advanced technologies and even renewable energy applicati ons are directly or indirectly related to industrial processes that deeply rely on fossil fuels. Therefore, it is highly important to design carbon - free and versatile reaction procedures that provide a good tunability and control over the key material char acteristics - i.e. crystallinity, crystal orientations, size (dimension), morphology, and the surface terminal groups where the ordered bonding is interrupted - of the final products. Importantly, a deep knowledge of the material synthesis provides a path for studying the structure - function relationship of materials. This further allows to design exotic materials with unusual properties which will be the foundation of new technologies for catalysis , energy conversion and energy storage. Furthermore, many cu tting - edge 29 technologies and autonomous materials are comprised of various components, each with a specific functionality. Thus, integrating these materials while preserving their functionality to operate as a whole is fundamentally challenging and requires a thorough control over the synthesis of each component as well as the techniques to interface them together. For example, the wireless solar water splitting system (see section 1.2 ) is comprised of a transparent conductive layer, two semiconductor light absorbers, electrocatalyst, and a separation membrane layer. To realize a high - performance device, it is thus important to minimize any energy loss by elaborate and constructive integration of these layers. The optimization of individual layer and integration of them, of course, covers a wide range of science and research efforts. In this dissertation we utilized atomic layer deposition (ALD) and ele ctrodeposition (ED) to synthesize thin films of semiconductors on various substrates. In addition, spin coating was utilized to modify the electrode surface with a thin layer of electrocatalyst. In the following sections the ALD and ED methods are discusse d in more details. 1.4.1. Atomic Layer Deposition Atomic layer deposition is a promising and robust method for the synthesis of thin films both on industrial and academic scales. ALD is characterized by its precise control over the film thickness and the atomic level control over the composition of the film. In addition, in comparison to the conventional methods, ALD p rovides substantially moderate synthesis condition s and as the ALD process is performed at reduced pressures in inert atmospheres, it is suitable f or the sy nthesis of non - oxide and air - sensitive mater ials which otherwise require harsh and reactive synthesis conditions. From a synthesis point of view, t his is of great interest as it eliminates the necessity for reactive conditions to integrate multipl e layers each with specific functionality. 79 30 The self - limiting ALD deposition reaction occurs at the interface of the gas and solid phase and it is comprised of 4 elementary steps which include sequential introduction of precursor and co - reactant into the deposition chamber in a timely fashion ( Figure 1 13 ) . Initially, the precursor is introduced in the deposition chamber (step 1) where it reacts with the active surface sites by forming chemical bonds (chemisorption) and the excess precursors are physisorbed. Next, the chamber is purged with a chemically inert carrier gas to remove the loosely bonded precursors (step 2). Ideally, this step provides an atomically clean surface ready for the next step. Subsequently, the co - reactant is introduced into the chamber to react with the fist layer (step 3) followed by a purge step to remove the gaseous products and excess co - reactants (step 4). Figure 1 13 . The schematic representation of atomic layer deposition of a binary reaction. 31 Together, these four steps complete one ALD cycle that corresponds to a film growth of a few angstrom thickness (the thickness depends on the material and deposition conditions/ precursors). To grow a given the whole sequence is therefore repeated many times. The time scale of the ALD process depends on material being deposited, thickness of the film, and also the aspect ratio (porosity) of the substrate. For example, the growth r ate of hematite is ~ 0.5 Å /cycle and it takes ~ 83 s to complete 1 ALD cycle. To grow 30 nm hematite film on a planner and non - porous substrate, it takes ~ 18 hours. 57,63,80 82 Slow grow rate and long deposition durations are certainly two in herent limitations of the ALD, limiting its applicability to the films with nm thicknesses. 1.4.2. Electr o deposition Electrodeposition, or electroplating, is a method that can be used to prepare both thin films and nano (crystalline) materials. By controlling th e composition, ionic strength, pH, and temperature of the deposition bath, the physical characteristic of the final product, such as crystallinity, morphology and composition, can be easily and effectively controlled. In addition, electrodeposition is amen able to large - scale production and requires fairly simple instrumentation and is, at least relative to ALD, considerably cheaper overall. Due to these attributes, electrodeposition is one of the promising techniques for the synthesis of thin - film semicondu ctors. Unlike the ALD method, ED is limited to conductive substrates. In a typical ED experiment, electrons are considered as one of the reactants that can directly, or indirectly, react with precursors in the solution. Therefore, by controlling the ele ctron energy and concentration at the surface of working electrode, one can control the nature and rate of the deposition reaction occurring at the surface of the working electrode. A schematic picture of an undivided electrodeposition cell in a three - elec trode configuration is shown in Figure 1 14 . As shown the potential of the working 32 electrode is controlled/ or monitored relative to a reference elec trode while the current is collected/ controlled at the counter electrode. The thickness of the film can simply be controlled by varying the deposition time. The electrodeposition reaction can proceed through two distinctive pathways, namely: (1) the direc t reaction of electrons with solubilized precursors, followed by their deposition on the surface of the electrode; (2) the reaction of electrons with sacrificial reagents in solution, which subsequently reacts with precursor in the solution, to then lead t o the formation of the film at the surface of the electrode. Figure 1 14 . A typical schematic representation of electrodeposition cell in three - electrode configuration. 33 1.5. Dissertation Overview and Objectives This dissertation is focused on the synthesis and characterization of semiconductor materials for photoelectrochemical water oxidation , and consists in two sections. In the first section, the synthesis of tantalum nitride thin films for PEC water oxidation were investigated. To circumvent the issues associated with the conventional two - step synthesis procedure of Ta 3 N 5 (oxidation followed by ammonolysis), the ED and ALD as bottom - up procedures are investigated. The aim of these projects is to realize/integrate Ta 3 N 5 photoelectrodes on a transparent conductive substrates (TCO), one of the crucial requirements for tandem cell designs. In addition, this platform provides a viable system to study the fundamental properties of t antalum nitride, using the illumination direction as another parameter to explore the fundamental characteristics of Ta 3 N 5 electrodes. In the first project, a two - step electrodeposition of tantalum oxide from aqueous solutions followed by ammonolysis was utilized to fabricate nanostructured thin films of Ta 3 N 5 on various substrates. To avoid the ammonolysis step, ALD was utilized to directly deposit thin films of tantalum nitride. It was noted that the films deposited at the temperature range of 1 - 280 °C are amorphous TaO x N y which are nitridized to crystalline Ta 3 N 5 via annealing in ammonia at 750 °C for 30 min. Subsequently, TaO x N y was integrated with Ta - doped TiO 2 (a newly developed stable TCO under reducing conditions of ammonolysis) via ALD to introdu ce the first demonstration of Ta 3 N 5 grown on a TCO. Inspired by this study, we noted that direct deposition of crystalline Ta 3 N 5 requires - temperature ALD system was designed. In this project, the direct deposition of crystalline Ta 3 N 5 on FTO (a commercially available TCO which is not stable under ammonolysis conditions) was studied, 34 allowing us to study the fundamental properties of Ta 3 N 5 independently of the conductivity of the s ubstrate. In the second section of this dissertation, the PEC performance of catalytically - modified hematite electrodes was studied. As discussed earlier, the PEC water oxidation performance of hematite is limited by its low catalytic activity toward water oxidation. The aim of the second part of this dissertation is to understand how catalyst s interface with underlying semiconductor and how it affect s the performance of cat alytically - modified electrodes. For this purpose, hematite electrodes were prepared by ALD and ED, and Ni 1 - x Fe x O y with various compositions were utilized as the water oxidation catalyst. The catalyst - coated hematite electrodes were subsequently studied with various electrochemical and spectroscopic methods. In addition, in collaboration w ith the Rothschild group at Technion, a combinations of two Ni 1 - x Fe x O y catalysts, i.e. Ni - rich and Fe - rich catalysts, prepared by UV light - assisted deposition and hematite electrodes with two distinctive orientations, i.e. (100) and (001) were utilized as model electrodes with systematically different composition / or structure of the interface at the semiconductor|catalyst junction. The preliminary data indicate that the PEC water oxidation performance of the catalyst - coated electrodes depends on both the composition of the catalyst and the crystal orientation of the hematite substrate. Overall, this dissertation highlights the origins of some of the challenges associated with efficient PEC water oxidation on Ta 3 N 5 and hematite electrodes. The knowledge gained from this dissertation are the base of some of the exciting pr ojects carried out in our lab. 35 R EFERENCES 36 REFERENCES (1) NASA - 2. Climate Change: How Do We Know? http://climate.nasa.gov/evidence/ (accessed Jan 10, 2018). (2) Stocker; T.F., D. Qin, G. - K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. B. and P. M. M. (eds. . Summary for Policymakers. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change ; P.M. Midgley, Ed.; Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 1535 pp, doi:10.1017/CBO9781107415324., 2015; Vol. 1542. (3) EIA. Gree - Energy Explained, Your Guide To Understanding Energy - Energy Information Administration https://www.eia.gov/energyexplained/index.cfm?page=environment_how_ghg_affect_cli mate (accessed Jan 10, 2018). (4) GISTEMP Team. GISS Surface Temperature Analysis (GISTEMP) www.data.giss.nasa.gov/gistemp/ (accessed Jan 10, 2018). (5) Sánchez, S. Global Climate Change Indicators https://www.ncdc.noaa.gov/indicators/ (accessed Jan 10, 2018). (6) Levitus, S.; Antonov, J. I.; Boyer, T. P.; Locarnini, R. A.; Garcia, H. E.; Mishonov, A. V. Global Ocean Heat Content 1955 - 2008 in Light of Recently Revealed Instrumentation Problems. Geophys. Res. Lett. 2009 , 36 (7), L07608. (7) NOAA PMEL. What is Ocean Acidification? https://www.pmel.noaa .gov/co2/story/What+is+Ocean+Acidification%3F (accessed Jan 10, 2018). (8) Knoll, A. H.; Fischer, W. W. Ocean Acidification https://www.pmel.noaa.gov/co2/story/Ocean+Acidification (accessed Jan 10, 2018). (9) U.S. Energy Information Administration (EIA). How much carbon dioxide is produced from burning gasoline and diesel fuel? - FAQ - U.S. Energy Information Administration (EIA) https://www.eia.gov/tools/faqs/faq.php?id=307&t=10 (accessed Jan 10, 2018). (10) Raupach, M. R.; Davis, S. J.; Peters, G. P.; Andrew, R. M.; Canadell, J. G.; Ciais, P.; Friedlingstein, P.; Jotzo, F.; Van Vuuren, D. P.; Le Quéré, C. Sharing a Quota on Cumulative Carbon Emissions. Nat. Clim. Chang. 2014 , 4 (10), 873 879. (11) Jäger - Valdau, A. PV Status Report 2017 ; 2017. (12) Lew is, N. S.; Nocera, D. G. Renewables 2017 http://www.iea.org/bookshop/761 - Market_Report_Series:_Renewables_2017 (accessed Aug 2, 2018). 37 (13) Zamfirescu, C.; Dincer, I. Using Ammonia as a Sustainable Fuel. J. Power Sources 2008 , 185 (1), 459 465. (14) Hu, S.; Xiang, C.; Haussener, S.; Berger, A. D.; Lewis, N. S. An Analysis of the Optimal Band Gaps of Light Absorbers in Integrated Tandem Photoelectrochemical Water - Splitting Systems. Energy Environ. Sci. 2013 , 6 (10), 2984 2993. (15) Verlage, E.; Hu, S.; Li u, R.; Jones, R. J. R.; Sun, K.; Xiang, C.; Lewis, N. S.; Atwater, H. A. A Monolithically Integrated, Intrinsically Safe, 10% Efficient, Solar - Driven Water - Splitting System Based on Active, Stable Earth - Abundant Electrocatalysts in Conjunction with Tandem III - V Light Absorbers Protected by Amorphous TiO 2 Films. Energy Environ. Sci. 2015 , 8 (11), 3166 3172. (16) 3 N 5 at 16 K by Time - of - Flight Neutron Diffraction. Acta Crystallogr. Sect. C Cryst. Struct. Commun. 1991 , 47 (11), 2291 2294. (17) Nurlaela, E.; Harb, M.; Del Gobbo, S.; Vashishta, M.; Takanabe, K. Combined Experimental and Theoretical Assessments of the Lattice Dynamics and Optoelectronics of TaON and Ta 3 N 5 . J. Solid State Chem. 2015 , 229 , 219 227. (18) Hajibabaei, H.; Zandi, O.; Hamann, T. W. Tantalum Nitride Films Integrated with Transparent Conductive Oxide Substrates: Via Atomic Layer Deposition for Photoelectrochemical Water Splitting. Chem. Sci. 2016 , 7 (11), 6760 6767. (19 ) Hajibabaei, H.; Hamann, T. W. Selective Electrodeposition of Tantalum(V) Oxide Electrodes. Langmuir 2017 , 33 (41), 10800 10806. (20) Morbec, J. M.; Narkeviciute, I.; Jaramillo, T. F.; Galli, G. Optoelectronic Properties of Ta 3 N 5 : A Joint Theoretical an d Experimental Study. Phys. Rev. B 2014 , 90 (15), 155204. (21) Wang, J.; Fang, T.; Zhang, L.; Feng, J.; Li, Z.; Zou, Z. Effects of Oxygen Doping on Optical Band Gap and Band Edge Positions of Ta 3 N 5 Photocatalyst: A GGA + U Calculation. J. Catal. 2014 , 309 , 291 299. (22) Fang, C. M.; Orhan, E.; De Wijs, G. A.; Hintzen, H. T.; De Groot, R. A.; Marchand, R.; Saillard, J. Y.; De With, G. The Electronic Structure of Tantalum (Oxy)Nitrides TaON and Ta 3 N 5 . J. Mater. Chem. 2001 , 11 (4), 1248 1252. (23) Wu, Y.; Ceder, G. First Principles Study on Ta 3 N 5 :Ti 3 O 3 N 2 Solid Solution As a Water - Splitting Photocatalyst. J. Phys. Chem. C 2013 , 117 (47), 24710 24715. (24) Chun, W. J.; Ishikawa, A.; Fujisawa, H.; Takata, T.; Kondo, J. N.; Hara, M.; Kawai, M.; Matsumoto, Y.; Domen, K. Conduction and Valence Band Positions of Ta 2 O 5 , TaON, and Ta 3 N 5 by UPS and Electrochemical Methods. J. Phys. Chem. B 2003 , 107 (8), 1798 1803. (25) Maeda, K.; Terashima, H.; Kase, K.; Higashi, M.; Tabata, M.; Domen, K. Surface Modification of Ta ON with Monoclinic ZrO 2 to Produce a Composite Photocatalyst with 38 Enhanced Hydrogen Evolution Activity under Visible Light. Bull. Chem. Soc. Jpn. 2008 , 81 (8), 927 937. (26) Dabirian, A.; Van De Krol, R. Resonant Optical Absorption and Defect Control in T a 3 N 5 Photoanodes. Appl. Phys. Lett. 2013 , 102 (3), 033905. (27) Pinaud, B. A.; Vailionis, A.; Jaramillo, T. F. Controlling the Structural and Optical Properties of Ta 3 N 5 Films through Nitridation Temperature and the Nature of the Ta Metal. Chem. Mater. 2014 , 26 (4), 1576 1582. (28) Nurlaela, E.; Ould - Chikh, S.; Harb, M.; Del Gobbo, S.; Aouine, M.; Puzenat, E.; Sautet, P.; Domen, K.; Basset, J. M.; Takanabe, K. Critical Role of the Semiconductor - Electrolyte Interface in Photocatalytic Performance for Wat er - Splitting Reactions Using Ta 3 N 5 Particles. Chem. Mater. 2014 , 26 (16), 4812 4825. (29) Yuliati, L.; Yang, J. H.; Wang, X.; Maeda, K.; Takata, T.; Antonietti, M.; Domen, K. Highly Active Tantalum(v) Nitride Nanoparticles Prepared from a Mesoporous Carbo n Nitride Template for Photocatalytic Hydrogen Evolution under Visible Light Irradiation. J. Mater. Chem. 2010 , 20 (21), 4295 4298. (30) McCrory, C. C. L.; Jung, S.; Ferrer, I. M.; Chatman, S. M.; Peters, J. C.; Jaramillo, T. F. Benchmarking Hydrogen Evol ving Reaction and Oxygen Evolving Reaction Electrocatalysts for Solar Water Splitting Devices. J. Am. Chem. Soc. 2015 , 137 (13), 4347 4357. (31) Jang, J. - W.; Du, C.; Ye, Y.; Lin, Y.; Yao, X.; Thorne, J.; Liu, E.; McMahon, G.; Zhu, J.; Javey, A.; et al. En abling Unassisted Solar Water Splitting by Iron Oxide and Silicon. Nat. Commun. 2015 , 6 , 7447. (32) Gärtner, W. W. Depletion - Layer Photoeffects in Semiconductors. Phys. Rev. 1959 , 116 (1), 84 87. (33) Sze, S. M.; Ng, K. K. Physics of Semiconductor Device s ; John Wiley & Sons, Inc.: Hoboken, NJ, USA, 2006. (34) Wahid, H.; Ahmad, S.; Nor, M. A. M.; Rashid, M. A. Prestasi Kecekapan Pengurusan Kewangan Dan Agihan Zakat: Perbandingan Antara Majlis Agama Islam Negeri Di Malaysia. J. Ekon. Malaysia 2017 , 51 (2), 39 54. (35) De Respinis, M.; Fravventura, M.; Abdi, F. F.; Schreuders, H.; Savenije, T. J.; Smith, W. A.; Dam, B.; Van De Krol, R. Oxynitrogenography: Controlled Synthesis of Single - Phase Tantalum Oxynitride Photoabsorbers. Chem. Mater. 2015 , 27 (20), 70 91 7099. (36) Ziani, A.; Nurlaela, E.; Dhawale, D. S.; Silva, D. A.; Alarousu, E.; Mohammed, O. F.; Takanabe, K. Carrier Dynamics of a Visible - Light - Responsive Ta 3 N 5 Photoanode for Water Oxidation. Phys. Chem. Chem. Phys. 2015 , 17 (4), 2670 2677. (37) Butler, M. A. Photoelectrolysis and Physical Properties of the Semiconducting Electrode 39 WO 2 . J. Appl. Phys. 1977 , 48 (5), 1914 1920. (38) Pala, R. A.; Leenheer, A. J.; Lichterman, M.; Atwater, H. A.; Lewis, N. S. Measurement of Minority - Carrier Diffusion Lengths Using Wedge - Shaped Semiconductor Photoelectrodes. Energy Environ. Sci. 2014 , 7 (10), 3424 3430. (39) Ziani, A.; Nurlaela, E.; Dhawale, D. S.; Silva, D. A.; Alarousu, E.; Mohammed, O. F.; Takanabe, K. Carrier Dynamics of a Visible - Light - Responsive Ta 3 N 5 Photoanode for Water Oxidation. Phys. Chem. Chem. Phys. 2015 , 17 (4), 2670 2677. (40) Henderson, S. J.; Hector, A. L. Structural and Compositional Variations in Ta 3 N 5 Produced by High - Temperature Ammonolysis of Tantalum Oxide. J. Solid State Chem. 2 006 , 179 (11), 3518 3524. (41) Li, Y.; Zhang, L.; Torres - Pardo, A.; González - Calbet, J. M.; Ma, Y.; Oleynikov, P.; Terasaki, O.; Asahina, S.; Shima, M.; Cha, D.; et al. Cobalt Phosphate - Modified Barium - Doped Tantalum Nitride Nanorod Photoanode with 1.5% S olar Energy Conversion Efficiency. Nat. Commun. 2013 , 4 , 2566. (42) He, Y.; Thorne, J. E.; Wu, C. H.; Ma, P.; Du, C.; Dong, Q.; Guo, J.; Wang, D. What Limits the Performance of Ta 3 N 5 for Solar Water Splitting? Chem 2016 , 1 (4), 640 655. (43) Terao, N. Structure of Tantalum Nitrides. Jpn. J. Appl. Phys. 1971 , 10 (2), 248 259. (44) Liu, G.; Ye, S.; Yan, P.; Xiong, F.; Fu, P.; Wang, Z.; Chen, Z.; Shi, J.; Li, C. Enabling an Integrated Tantalum Nitride Photoanode to Approach the Theoretical Photo current Limit for Solar Water Splitting. Energy Environ. Sci. 2016 , 9 (4), 1327 1334. (45) Dabirian, A.; Van De Krol, R. High - Temperature Ammonolysis of Thin Film Ta 2 O 5 Photoanodes: Evolution of Structural, Optical, and Photoelectrochemical Properties. Ch em. Mater. 2015 , 27 (3), 708 715. (46) Hu, S.; Shaner, M. R.; Beardslee, J. A.; Lichterman, M.; Brunschwig, B. S.; Lewis, N. S. Amorphous TiO 2 Coatings Stabilize Si, GaAs, and GaP Photoanodes for Efficient Water Oxidation. Science (80 - . ). 2014 , 344 (6187 ), 1005 1009. (47) Hu, S.; Lewis, N. S.; Ager, J. W.; Yang, J.; McKone, J. R.; Strandwitz, N. C. Thin - Film Materials for the Protection of Semiconducting Photoelectrodes in Solar - Fuel Generators. J. Phys. Chem. C 2015 , 119 (43), 24201 24228. (48) Zhong, M.; Hisatomi, T.; Sasaki, Y.; Suzuki, S.; Teshima, K.; Nakabayashi, M.; Shibata, N.; Nishiyama, H.; Katayama, M.; Yamada, T.; et al. Highly Active GaN - Stabilized Ta 3 N 5 Thin - Film Photoanode for Solar Water Oxidation. Angew. Chemie Int. Ed. 2017 , 56 (17), 47 39 4743. (49) The U.S. Geological Survey, https://pubs.usgs.gov/fs/2002/fs087 - 02/, (accessed September 2017). 40 (50) Hardee, K. L. Semiconductor Electrodes. J. Electrochem. Soc. 1976 , 123 (7), 1024. (51) Horowitz, G. Capacitance - Voltage Measurements and Flat - Band Potential Determination on Zr - - Fe 2 O 3 Single - Crystal Electrodes. J. Electroanal. Chem. 1983 , 159 (2), 421 436. (52) - Fe 2 O 3 Electrodes. J. Electrochem. Soc. 1978 , 125 (5), 709. (53) Dare - Edwards, M. P.; Goodenough, J. B.; Hamnett, A.; Trevellick, P. R. Electrochemistry and Photoelectrochemistry of Iron(III) Oxide. J. Chem. Soc. Faraday Trans. 1 Phys. Chem. Condens. Phases 1983 , 79 (9), 2027 2041. (54) Sanchez, C.; Sieber, K. D.; Somorjai, G. A. Th e Photoelectrochemistry of Niobium Doped - Fe 2 O 3 . J. Electroanal. Chem. 1988 , 252 (2), 269 290. (55) Itoh, K.; Bockris, J. O. M. Stacked Thin - Film Photoelectrode Using Iron Oxide. J. Appl. Phys. 1984 , 56 (3), 874 876. (56) Itoh, K. Thin Film Photoelectro chemistry: Iron Oxide. J. Electrochem. Soc. 1984 , 131 (6), 1266 1271. (57) Riha, S. C.; Klahr, B. M.; Tyo, E. C.; Seifert, S.; Vajda, S.; Pellin, M. J.; Hamann, T. W.; Martinson, A. B. F. Atomic Layer Deposition of a Submonolayer Catalyst for the Enhanced Photoelectrochemical Performance of Water Oxidation with Hematite. ACS Nano 2013 , 7 (3), 2396 2405. (58) Sivula, K.; Zboril, R.; Le Formal, F.; Robert, R.; Weidenkaff, A.; Tucek, J.; Frydrych, J.; Grätzel, M. Photoelectrochemical Water Splitting with Mesoporous Hematite Prepared by a Solution - Based Colloidal Approach. J. Am. Chem. Soc. 2010 , 132 (21), 7436 74 44. (59) Kay, A.; Cesar, I.; Grätzel, M. New Benchmark for Water Photooxidation by - Fe 2 O 3 Films. J. Am. Chem. Soc. 2006 , 128 (49), 15714 15721. (60) Lin, Y.; Zhou, S.; Sheehan, S. W.; Wang, D. Nanonet - Based Hematite Heteronanostructures f or Efficient Solar Water Splitting. J. Am. Chem. Soc. 2011 , 133 (8), 2398 2401. (61) Lin, Y.; Yuan, G.; Sheehan, S.; Zhou, S.; Wang, D. Hematite - Based Solar Water Splitting: Challenges and Opportunities. Energy Environ. Sci. 2011 , 4 (12), 4862. (62) Upul Wijayantha, K. G.; Saremi - Yarahmadi, S.; Peter, L. M. Kinetics of Oxygen Evolution - Fe 2 O 3 Photoanodes: A Study by Photoelectrochemical Impedance Spectroscopy. Phys. Chem. Chem. Phys. 2011 , 13 (12), 5264. (63) Klahr, B. M.; Martinson, A. B. F.; Hamann , T. W. Photoelectrochemical Investigation of Ultrathin Film Iron Oxide Solar Cells Prepared by Atomic Layer Deposition. Langmuir 2011 , 27 (1), 461 468. 41 (64) Klahr, B. M.; Hamann, T. W. Current and Voltage Limiting Processes in Thin Film Hematite Electrod es. J. Phys. Chem. C 2011 , 115 (16), 8393 8399. (65) Klahr, B. M.; Hamann, T. W. Voltage Dependent Photocurrent of Thin Film Hematite Electrodes. Appl. Phys. Lett. 2011 , 99 (6), 063508. (66) Grave, D. A.; Klotz, D.; Kay, A.; Dotan, H.; Gupta, B.; Visoly - Fisher, I.; Rothschild, A. Effect of Orientation on Bulk and Surface Properties of Sn - - Fe 2 O 3 ) Heteroepitaxial Thin Film Photoanodes. J. Phys. Chem. C 2016 , 120 (51), 28961 28970. (67) Peter, L. M.; Upul Wijayantha, K. G. Photoelect rochemical Water Splitting at Semiconductor Electrodes: Fundamental Problems and New Perspectives. ChemPhysChem 2014 , 15 (10), 1983 1995. (68) Zandi, O.; Schon, A. R.; Hajibabaei, H.; Hamann, T. W. Enhanced Charge Separation and Collection in High - Perform ance Electrodeposited Hematite Films. Chem. Mater. 2016 , 28 (3), 765 771. (69) Hajibabaei, H.; Gao, Y.; Hamann, T. W. Unravelling the Charge Transfer Mechanism in Water Splitting Hematite Photoanodes. In Advances in Photoelectrochemical Water Splitting: Theory, Experiment and Systems Analysis ; S David Tilley, Stephan Lany, R. van de K., Ed.; 2018; pp 100 127. (70) Thorne, J. E.; Jang, J. W.; Liu, E. Y.; Wang, D. Understanding the Origin of Photoelectrode Performance Enhancement by Probing Surface Kinetic s. Chem. Sci. 2016 , 7 (5), 3347 3354. (71) Barroso, M.; Mesa, C. A.; Pendlebury, S. R.; Cowan, A. J.; Hisatomi, T.; Sivula, K.; Gratzel, - Fe 2 O 3 Photoanodes for Solar Wat er Splitting. Proc. Natl. Acad. Sci. 2012 , 109 (39), 15640 15645. (72) Hisatomi, T.; Le Formal, F.; Cornuz, M.; Brillet, J.; Tétreault, N.; Sivula, K.; Grätzel, M. Cathodic Shift in Onset Potential of Solar Oxygen Evolution on Hematite by 13 - Group Oxide O verlayers. Energy Environ. Sci. 2011 , 4 (7), 2512 2515. (73) Klahr, B.; Gimenez, S.; Fabregat - Santiago, F.; Bisquert, J.; Hamann, T. W. Photoelectrochemical and Impedance Spectroscopic Investigation of Water Oxidation with - - Coated Hematite Electrodes. J. Am. Chem. Soc. 2012 , 134 (40), 16693 16700. (74) Qiu, J.; Hajibabaei, H.; Nellist, M. R.; Laskowski, F. A. L.; Hamann, T. W.; Boettcher, S. W. Direct in Situ Measurement of Charge Transfer Processes During Photoelectrochemical Water Oxidation on Catalyzed Hematite. ACS Cent. Sci. 2017 , 3 (9), 1015 1025. (75) Pendlebury, S. R.; Cowan, A. J.; Barroso, M.; Sivula, K.; Ye, J.; Grätzel, M.; Klug, D. R.; Tang, J.; Durrant, J. R. Correlating Long - Lived Photogenerated Hole Populations with Photocurrent Densities in Hematite Water Oxidation Photoanodes. Energy Environ . Sci. 2012 , 5 (4), 6304 6312. 42 (76) Pendlebury, S. R.; Barroso, M.; Cowan, A. J.; Sivula, K.; Tang, J.; Grätzel, M.; Klug, D.; - Fe 2 O 3 Electrodes for Water Oxidation Probed by Transient Ab sorption Spectroscopy. Chem. Commun. 2011 , 47 (2), 716 718. (77) Hajibabaei, H.; Schon, A. R.; Hamann, T. W. Interface Control of Photoelectrochemical Water Oxidation Performance with Ni 1 - x Fe x O y Modified Hematite Photoanodes. Chem. Mater. 2017 , 29 (16), a cs.chemmater.7b01149. (78) Wang, G.; Ling, Y.; Lu, X.; Zhai, T.; Qian, F.; Tong, Y.; Li, Y. A Mechanistic Study into the Catalytic Effect of Ni(OH) 2 on Hematite for Photoelectrochemical Water Oxidation. Nanoscale 2013 , 5 (10), 4129. (79) George, S. M. At omic Layer Deposition: An Overview. Chem. Rev. 2010 , 110 (1), 111 131. (80) Zandi, O.; Hamann, T. W. Enhanced Water Splitting Efficiency through Selective Surface State Removal. J. Phys. Chem. Lett. 2014 , 5 (9), 1522 1526. (81) Klahr, B.; Gimenez, S.; Zandi, O.; Fabregat - Santiago, F.; Hamann, T. Competitive Photoelectrochemical Methanol and Water Oxidation with Hematite Electrodes. ACS Appl. Mater. Interfaces 2015 , 7 (14), 7653 7660. (82) Zandi, O.; Beardslee, J. A.; Hamann, T. Substrate Dependent Water Splitting with Ultrathin - Fe 2 O 3 Electrodes. J. Phys. Chem. C 2014 , 118 (30), 16494 16503. 43 Selective Electrodeposition of Tantalum(V) Oxide Electrodes Adapted With permission from: Selective Electrodeposition of Tantalum(V) Oxide Electrodes, Hamed Hajibabaei, and Thomas W. Hamann, Langmuir , 2017 , 29 (16), 6674 6683. Copyright (20 17). American Chemical Society. 44 2.1. Abstract Thin films of TaO x H y were cathodically electrodeposited from an aqueous solution containing Ta - IPA precursor and KNO 3 as a sacrificial agent. It was shown that the deposition resulted from a precipitation reaction triggered by the local change of pH at the surface of working electrode. Combined structural and compositional analysis revealed that during the electrodeposition , the oxidation stat e of tantalum remained constant as Ta(V). The as - deposited films are mesoporous amorphous tantalum oxide hydrate films, which can be co nverted to either pure Ta 2 O 5 or Ta 3 N 5 by high - temperature annealing in either air (or Ar) or ammonia, respectively. T he Ta 3 N 5 electrodes exhibited promising PEC activity for water oxidation. These results open the door for the reduced temperature synthesis of Ta 3 N 5 electrodes on TCO substrates which would allow for efficient overall solar water splitting. 2.2. Introduction Tantalum oxide, Ta 2 O 5 , is a wide band gap semiconductor with a high refractive index and dielectric constant, 1 that has been used in a wide range of applications including capacitors, 2 electrochromic devices, 3 5 and catalysis. 6,7 Tantalum oxide is also the predo minant starting material for synthesis of tantalum nitride (Ta 3 N 5 ), which is a nearly ideal n - type semiconductor for solar water splitting. 1,8 13 For the formation of Ta 3 N 5 electrodes, the oxide is first derived from the oxidation or anodization of a Ta substrate, followed by ammonolysis. Despite the simplicity and great performance that has come from this method, there are multiple negative consequences. First, the initial oxidation of tantalum metal is very energy intensive and generally requires hazardous chemicals such as hydrogen fluoride. 14 17 Second, the Ta substrate is not transpa rent to sub - band gap light, Ta 3 N 5 cannot be employed as a photoanode in a tandem configuration. 45 Furthermore, the oxidation of Ta can result in the formation of sub - oxide 18,19 phases which later are converted to nitrogen - poor phases, TaN x (x < 1.67), at the interface between Ta 3 N 5 and Ta substrate. 15,20 These phases are normally electronicall y resistive which can suppress the overall performance of the electrodes. The above drawbacks of Ta oxidation or anodization can be circumvented via the electrodeposition of tantalum(V) oxide(hydrate) films. Electrodeposition is one of the most powerful t echniques to obtain uniform films on conductive substrates, which should also allow use of transparent conductive oxide (TCO) substrates instead of the opaque Ta substrates. Generally, electrodeposition can be divided in two main groups: electroreduction /d eposition and precipitation/deposition. 21 The electroreduction /deposition is widely used for plating a large group of metals on conductive substrates where the metal precu rsor in the solution is reduced and subsequently deposited on the surface of the electrode. Most Ta - precursors, such as TaCl 5 , TaF 5 , and K 2 TaF 7 are highly moisture sensitive and spontaneously react with water. Therefore, the electrodeposition of tantalum h as been limited to the nonaqueous media, with previous studies utilizing either molten salts or ionic liquids with a well - controlled atmosphere. Since the electro - reduction of a tantalum precursor is the dominant reaction occurring at the cathode, this syn thetic route results in the formation of thin films of metallic Ta. 22 25 Recently, Domen and co - workers 6,7 studied the electrodeposition of TaO x nanoparticles from a non - aqueous solution to make catalysts for the oxygen reduction reaction. Their resu lts showed that the TaO x deposited in this way were highly dispersed nanoparticles comprised of sub - oxides of tantalum. In a separate study by Moo and coworkers, 26 it was shown that the cathodic deposition of tantalum in nonaqueous media resulted in films with the Ta reduced to lower oxidation states, e.g. Ta 4+ and Ta 3+ . These results suggest that the electrochemically induced deposition from non - aqueous solution results in the 46 formation of TaO x with a range of compositions, including Ta 2 O 5 , TaO 2 , and TaO, through the electro - reduction of a tantalum precursor. To the best of our knowledge, there has been no report of the electrodeposition of pure Ta 2 O 5 . In order to exploit t he power of electrodeposition tantalum oxide, where the oxidation state of Ta 5+ is preserved, the cathodic reaction at the surface of working electrode must be controlled to suppress the reduction of Ta 5+ . One method to realize the electrodeposition of Ta 2 O 5 is through the two - step precipitation/deposition mechanism. In this method, a co - reactant is electrochemically generated at the surface of the electrode followed by its reaction with a metal - precursor in solution where the products precipitate at the el ectrode surface. It was further shown that this approach can be used for preparation of both thin films and nanocrystalline powders. 27 33 In 1996, Izaki et al . 27,28 and Lincot et al . 31,32 studied the electrodeposition of ZnO films from an aqueous solution via electroreduction of nitrate and dissolved oxygen, respectively. Choi and coworkers 34,35 further studied the electrodeposition of zinc oxide films by reduction of nitrate ions in a slightly acidic zinc nitrate solu tion. It was shown that the electroreduction of these nitrate produced hydroxide ions at the surface of working electrode, thereby increasing the local pH which initiates the precipitation of ZnO films. 28,34 This technique has now been widely utilized to deposit metal oxides such as Fe 2 O 3 , 36,37 ZnO, 34 KNbO 3 , 38 and NiO. 39 In this work, we employ the precipitation/deposition approach to directly deposit tantalum oxy(hydrate) as thin films from aqueous solution, where the oxidation state of Ta 5+ is preserved. The resulting film is highly hydrated, and upon annealing in air o r ammonia atmosphere it can be converted to phase pure films of Ta 2 O 5 or Ta 3 N 5 , respectively . 47 2.3. Experimental Section 2.3.1. Electrodeposition of tantalum oxide All electrodeposition reactions were carried out in an undivided cell using Eco Chemie Autolab potentios tat coupled with No va electrochemical software. H omemade saturated Ag/AgCl and high surface area platinum mesh electrodes were used as the reference and counter electrode, respectively. The thin films of TaO x H y were electrodeposited on different substrates , including Ti ( Sigma - Aldrich , 99.7% trace metal), Pt ( Sigma - Aldrich , 99.99%), and Ta (Alfa Aesar, 99.95% metal basis). The substrates were cleaned by sequential sonication in soap deionized ( DI ) water, and isopropyl alcohol ( IPA ) for 15 minute each. Electrodeposition was performed at ambient temperature in aqueous solutions. The tantalum precursor solution, labeled as Ta - IPA, was separately prepared by dissolving the calculated amount (895 mg) of TaCl 5 (Alfa Aesar 99. 8% metal basis), i n 5 mL 99.9% isopropyl alcohol ( Sigma - Aldrich, used as received without further purifications or drying ) to make a 0.5 M solution. This solution was further aged at room temperature for 24 hours and any particulate formed during aging the solution, presumably tantalum oxide due to the reaction of TaCl 5 with the trace amount of water in IPA or air, was subsequently filtered with 0.45 µm syringe filter ( Lab Depot, Inc.). For electrodeposition, we tested Ta - IPA solution aged for various durat ions, as - prepared and aged for 1h to 1 day. The best results in terms of reproducibility of the deposited film were obtained when the Ta - IPA solution was aged at least for 24 hours. An aliquot was transferred to the plating bath to prepare a 50 mM solution of the Ta precursor containing 1 M KCl and 0.5 M KNO 3 . The resulting solution was sonicated for 1 hour at room temperature until a milky solution was obtained. The cathodic deposition w as carried out at - 1 V vs. Ag/AgCl for a fixed time. The as - 48 deposited films were soaked in DI water for at least 2 hours. The samples were frozen by immersing in liquid nitrogen for 15 minutes and immediately transferred to a freeze dryer (lyophilizer labconco freezone 2.5 plus) and dried for 12 hours. The surface coverage of the film was electrochemically determined by cyclic voltammetry (CV) using an aqueous solution containing 1 mM K 3 Fe(CN) 6 and 0.5 M KCl as the supporting electrolyte. The potential was cathodically swept from 0.6 to 0 V vs. Ag/AgCl with a scan rate of 100 mV s - 1 . The as - deposited films were crystallized by annealing in Ar (99.99%, in tube furnace) and air (in muffle furnace) for 1 h at 850 °C. For nitridization to Ta 3 N 5 , the deposited films were annealed in a flow of dry ammonia (99.99%) with a flow rate of 500 mL/min, at 850 °C for 1 h with the heating rate of 10 °C/min and then le f t to naturally cool down to room temperature. Control samples of TaO x H y were prepared via precipitation by adjusting the pH of the addition of 1 M KOH. The white precipitate was then filtered and washed 3 times with excess amount of DI water and dried under vacuum for 6 hour. The white powder was then annealed in air in a preheated furnace at 850 °C for 1 hour. 2.3.2. Film Characterization The morphologies of the films were examined by scanning electron microscopy (Carl Zeiss Auriga, Dual Column FIB - SEM ). Energy dispersive spectroscopy (EDS) was used to determine the film composition . X - ray photoelectron spectroscopy (XPS) was performed wit h a Perkin Elmer angle of 45°. Survey scans of 0 - 1100 eV binding energy and detailed scans for C 1s, O 1s, and Ta4f regions were measured for all samples. The binding energies were corrected in reference to 49 C 1s peak (284.8 eV). The Raman spectra were collected using inVia Raman Microprobe (Renishaw) equipped with a 45W Cobalt DPSS laser (532 nm line) laser and a 100x microscope to focus the laser on the film surface. The spectrometer was calibrated (wavelength position) with internal s ilicon standard. For comparison, Raman spectra were normalized to the total area. The X - ray diffraction (XRD) patterns were obtained on a Bruker D8 Advanced diffractometer using 2.3.3. Photoelectrodeposition of CoPi Prior to PEC measurements the surface of electrodes were modified with CoPi using a photo - electrodeposition method reporte d previously. 36,40 Briefly, the electrodes were immersed in a solution containing 0.5 mM Co(NO 3 ) 2 6H 2 O in 0.1 M phosphate buffer (pH 7). A bias of 1 .0 V vs. RHE was applied under 1 sun illumination. The thicknes s of the CoPi film was controlled by the deposition time, with a 300 s deposition used for these electrodes. 2.3.4. Photoelectrochemical Measurements The PEC measurements were performed in an aqueous solution of 1 M KOH ( pH = 13.6). H omemade saturated Ag/AgCl ele ctrode and platinum mesh were used as a reference and counter electrode, respectively. The reference electrode was frequently calibrated to a commercial saturated calomel electrode (Koslow Scientific). The potential vs. Ag/AgCl were converted to reversible hydrogen electrode (RHE) by the equation . The chopped light J - E curves were measured at a scan rate of 20 mV s - 1 . The light source was a 450 W Xe arc lamp (Horiba Jobin Yvon). The incident light was chopped using a computer - controlled Thor Labs solenoid shutter. An AM 1.5 solar filter was used to simulate sunlight at 100 mWcm - 2 (1 sun). 50 2.4. Results and Discussion TaCl 5 was dissolved in IPA under ambient conditions, then mixed with water, which resulted in a transparent a nd highly acidic solution (pH < 1). No precipitation was observed upon standing for more than 10 hours ( Figure A 2 - 1 ). Upon raising the pH (> 1.7) or temperature (> 50 °C), however, a white solid precipitated. These observations indicate that the hydrolysis of tantalum is very slow in acidic solutions at room temperature but can be accelerated by increasing the pH and/or the temperature of the solution. In order to characterize the composition of the initial Ta - IPA solution prior to hydrolysis, any excess IPA solvent was evaporated under vacuum for 48 h which resulted a gel - like sample that was characterized via FT - IR and 1 HNMR spectroscopy. The detailed analysis of 1 HNMR and IR spectra are discussed following Figure A 2 - 2 . Briefly, both 1 HNMR and IR collectively indicate that TaCl 5 reacts with IPA to form tantalum isopropoxide. 41 HCl is formed as a by - product, which accounts for the low pH upon mixing with water. Cyclic voltammograms (CVs) of solutions containing 1 M KCl and different combinations of additional species in solution are shown in Figure 2 1 .a. In all cases, the potential was scanned from 0 .0 to - 1.5 V vs. Ag/AgCl using Ti foil as a working elec trode. The CV of the 1 M KCl solution at neutral pH shows negligible current, attributed to the low activity of Ti for the hydrogen evolution reaction. 42 Upon the addition of KNO 3 , a small increase in ca thodic current was observed at potentials negative of - 1.1 V vs. Ag/AgCl. This increase in cathodic current is attributed to the reduction of NO 3 - which causes a local pH increase at the interfacial region of electrode and electrolyte. 43 45 A positive potential shift along with a substantial increase in the cathodic current was observed for the KCl solution containing the Ta - IPA precursor. The increase in cathodic current with the acidic Ta - IPA precursor can be attributed to either reduction of protons or reduction of Ta 5+ to lower oxidation states. In order to identify the reaction responsible for this 51 shift, the CVs of solutions with pH = 1 (adjusted by addition of HCl) with and without the Ta - IPA precursor are compared in Figure 2 1 .b. The CVs of both solutions are similar, thus the cathodic wave is attributed to the reducti on of protons, in agreement with a previous report of proton reduction with Ti substrates in acidic solutions. 46 Figure 2 1 . a) The CVs of solutions containing 1 M KCl (red, pH = 7.0), 1 M KCl + 0.5 M KNO 3 (black, pH = 7.0), 1 M KCl + 50 mM Ta - IPA (green, pH = 0.6), and 1 M KCl + 0.5 M KNO 3 + 50 mM Ta - IPA (blue, pH = 0.6); inset is the magnified CVs of first two solutions. b) The CV of 1 M KCl with pH of 1 adjusted by HCl (dotted orange) and 1 M KCl + 50 mM Ta - IPA (green), c) The CV of 1 M KCl + 0.5 M KNO 3 at pH 1 adjusted by HCl (dotted maroon) and 1 M KCl + 0.5 M KNO 3 + 50 mM Ta - IPA (blue). When KNO 3 was added to the Ta - IPA solution, an even greater increase in ca thodic current was observed at more positive potentials, as depicted in Figure 2 1 .a . The CV of the KNO 3 + Ta - IPA solution was compared to the CV of a 3 , shown in Figure 2 1 .c . As shown both solutions produced nearly identical CVs, suggesting that the cat hodic 52 wave of the electrodeposition bath is due to the reduction of NO 3 - in acidic solutions, not direct reduction of Ta 5+ or protons. Note the different potentials for NO 3 - reduction (black vs. blue lines in Figure 2 1 .a ) are due to the pH dependence of this reaction; CVs of KNO 3 in acidic and neutral solution are compared in Figure A 2 - 3 to support this. 47 The CVs of the same solutions were also measur ed utilizing Pt as the working electrode ( Figure A 2 - 4 ). The CVs of the 1 M KCl solution with and without KNO 3 are very similar, which is due to the h igh activity of Pt toward proton reduction competing with NO 3 - reduction. 48 Unlike the Ti electrode, th e CVs of the solutions containing Ta - IPA precursor with and without KNO 3 are also almost identical ( Figure A 2 - 4 .a ) with the cathodic wave substantial ly shifted to more positive potentials. This behavior is consistent with the CVs of 1 M KCl solution with different concentrations of HCl ( Figure A 2 - 4 .b ) which further indicates that on Pt electrode the reduction of protons is the dominant cathodic reaction. The CVs of the solution containing both Ta - IPA and KNO 3 , however, showed an anodic peak at - 1.3 V vs. Ag/AgCl after the potential was swept to - 1.5 V vs. Ag/AgCl . This new feature was neither observed on Ti electrodes nor on Pt electrodes in contact with solutions containing only Ta - IPA. Interestingly, when the potential was swept to - 1 V, the oxidation peak w as no longer observed ( Figure A 2 - 5 ). Therefore, we hypothesize that at potentials more negative than ~ - 1 .0 V, the Ta 5+ in the electrodeposited films was reduced to Ta 4+ or Ta 2 + , 6,26 and upon scanning anodically it was re - oxid ized to Ta 5+ . To test this hypothesis, two films were deposited by holding the potential at - 1.0 and - 1.5 V vs. Ag/AgCl and subsequently characterized via XPS. The peak positions of Ta 4f 7/2 and Ta 4f 5/2 strongly depend on the oxidation state of Ta and the immediate surrounding atoms, e.g. ~ 26.6 and 28.5 eV for Ta (5+) 11 The detailed XPS spectra of C 1s, O 1s, and Ta 4f are shown in Figure A 2 - 6 . The binding energies of all the peaks were referenced to C 1s at 284.8 eV. It was noted that regardless of the deposition 53 potential, the binding energy of oxygen peaks are not affected. However, at more negative deposition potentials the Ta 4f peak was shifted to the lower binding energies and became broader. This observation readily implies that at more negative deposition potentials, the Ta in the electrodeposited film was reduced which results in the formation of Ta suboxides with lower ox idation states of Ta. This is further consistent with the previous study by Stickney and coworkers where they showed that the surface oxide layer on the tantalum foil can be reduced in aqueous solutions at potentials below - 1.5 V. 49 Thus, electrodeposition was performed at - 1 .0 V vs. Ag/AgCl. We note that at this potential, the electrodeposition reaction triggered by a local pH increase can be induced by reduction of NO 3 - or by local depletion of H + by direct reduc tion at the surface. In order to test the extent of these possible competing reactions, depositions were performed from the plating bath with and without NO 3 - . The EDS spectra of the electrodeposited films prepared from solutions with and without NO 3 - are shown in Figure A 2 - 7 . a . The observed peaks are assigned to Ti (substrate), Ta and O. The sampling depth of the EDS is in the range of micrometers, the refore, in case of thin films the substrate signals are always observed. The Ta signal, located at ~ 1.7 eV, can be seen for both electrodeposited films. This observation readily indicates that the electrodeposition proceeds by the local increase of pH at the surface of the electrode induced by either proton depletion or reduction of NO 3 - . However, the intensity of Ta signal is noticeably lower for the film grown without NO 3 - ions. We note that the atomic percentage found from EDS is a function of film thickness. Thus, to quantify the difference in the film thickness, we calculated the atomic percentage of Ta with respect to Ti (substrate). As depicted in Figure A 2 - 7 . b, the percentage of Ta for the film deposited without NO 3 - is considerably lower than the film deposited 54 in presence of NO 3 - , further implying that the deposi tion rate without NO 3 - is much slower. Therefore, NO 3 - was used in all following electrodeposition reactions. Figure 2 2 . a) The s urvey XPS spectra of electrodeposited tantalum oxide films (black) and precipitated powder by addition of 1 M KOH (red), b) the detailed spectra of Ta 4f, C 1s, and O 1s peaks. The composition of the electrodeposited film, as well as the tantalum oxide pow der precipitated via addition of KOH to the plating bath, were analyzed by XPS. All the peaks from survey scan ( Figure 2 2 .a) are assigned to Ta 4f, O 1s, and C 1s. This observation readily indicates that the electrodeposition reaction taking place at the surface of the electrode is the same reaction as the precipitation reaction induced by raising the pH via addition of KOH. The detailed XPS spectra of Ta 4f, O 1s, and C 1s are summarized in Figure 2 2 .b. It is expected that the oxidization state of Ta remains unchanged in precipitation reaction induced by ex - situ addition of KOH as confirmed by the Ta 4f peak at 26 - 28 eV, consistent with Ta 5+ . 11 The XPS spectra from the electrodeposited film show identical shape and position of the Ta 4f peak as the powder sample. This indicates that during the cathodic electrodeposition at - 1 .0 V vs. Ag/AgCl, tantalum was not involved in the 55 electrochemical reaction and thus its oxidation state was preserved. The atomic ratio of O/Ta of the as - deposited film and the powder were found to be ~ 2:7, which is greater than the theoretical ratio of 2:5 for Ta 2 O 5 . The excess oxygen in as - prepared samples are presumably bonded hydroxide or water groups. Therefore, the as - deposited films were marked as T aO x H y . The surface coverage of the substrate with TaO x H y electrodeposited films were studied by cyclic voltammetry of aqueous solutions containing K 3 [Fe(CN) 6 ]. Since the associated redox peaks of [Fe(CN) 6 ] 3 - /4 - are not detectable on Ti electrodes, a Pt ele ctrode was utilized. The reversible redox peak of Fe III /Fe II on a bare Pt electrode ( Figure A 2 - 8 ) is located at ~ 0.3 V vs. Ag/AgCl with the peak separation of ~ 60 mV. 50 After electrodeposition for 120 s, however, the current of the reversible red ox peaks of Fe III /Fe II was reduced . The ratio of the cathodic and anodic peak area s or of the current at the Pt electrode after electrodeposition relative to the bare Pt electrode were found to be ~ 0.15. Thus, ~ 85% of the Pt substrate is assumed to be co vered by the insulating TaO x H y films, where [Fe(CN) 6 ] 3 - /4 - cannot contact the platinum surface. The as - deposited films are gel - like and are highly hydrated ( Figure A 2 - 9 . a). Therefore, the drying process can greatly influence the final morphology of the film. Freeze drying was used to dry the films since this minimizes the effect of capillary forces induced by water evaporation on the morphology of the f ilm. The SEM micrograph of the dried as - deposited film is shown in Figure 2 3 .a . The film has a nanostructured morphology and almost full coverage of the Ti substrate, consistent with the CVs of [Fe(CN) 6 ] 3 - /4 - described above. Upon heating at 850 °C the morphology of the film was preserved ( Figure A 2 - 9 .b ). 56 Figure 2 3 . a) SEM micro - images of as - deposited TaO x H y , the inset is the magnified image , b) t he XRD patterns of electrodeposited film (top, cyan) and powder control sample precipitated by addition of 1 M KOH (bottom, dark red) after annealing in air at 850 °C for 1 hour. The vertical dotted lines represent the Bragg position s of Ta 2 O 5 with PDF # 01 - 071 - 0639 and the stars represents the diffraction lines of TiO 2 . The Raman spectrum of the dr ied as - deposited films and the KOH precipitated powder are compared in Figure A 2 - 10 . The Raman spectra of both samples were nominally identical, with two broad peaks centered at ~ 250 and 650 cm - 1 , which are consistent with amorphous tantalum oxide. 51,52 In line with the XPS results, the similarity of the Ra man spectra also indicates that the products of the electrodeposition and precipitation reaction induced by ex - situ addition of KOH are similar, therefore, it can be concluded that the electrodeposition is triggered by the local pH - increase and proceeds th rough precipitation/deposition mechanism. In order to crysta l lize the as - prepared samples, they we re annealed under a flow of Ar 99.9 % (an inert atmosphere) or air at 850 °C. The inert atmosphere for annealing was utilized to prevent oxidation of any lowe r valent tantalum that formed during electrodeposition. The XRD patterns of powder and electrodeposited 57 Figure 2 4 . a) Top view SEM image of the film after niridization , the inset is the magnified image. b) XRD pattern of the Ta 3 N 5 on Ti foil (top, orange), Ti substrate (bottom, dark red). The vertical dotted lines represent the Bragg position s of Ta 3 N 5 (orange diamond) with PDF # 01 - 089 - 5200. thin - film tantalum oxide on Ti substrates after annealing in ai r are shown in Figure 2 3 .b. The diffraction patterns for powder and films are unambiguously assigned to the crystalline phase of Ta 2 O 5 . 51 The remaining pea ks can be assigned to the TiO 2 which formed upon oxidation of the Ti substrate. As shown in Figure A 2 - 11 , the diffraction pattern of the film annealed in Ar is also assigned to a single phase of Ta 2 O 5 , and no tantalum suboxide phase is observed. In agreement with XRD patterns, the Raman spectra of the film and powder after crystallization ( Figure A 2 - 12 ) are consistent with the formation of crystalline Ta 2 O 5 . 51 The TaO x H y samples were also annealed under a flow of ammonia at 850 °C for 1 hour to convert the dried as - deposited films into Ta 3 N 5 . The SEM micrograph of Ta 3 N 5 is shown in Figure 2 4 .a, which indicates the substrate is covered with the porous film of Ta 3 N 5 with a comparable morphology to the as - deposited or Ta 2 O 5 films. The XRD patterns of the nitridized film and Ti subst rate are compared in Figure 2 4 . b. It is worth mentioning that the Ti substrate was treated in 58 Figure 2 5 . Plots of J - E curves for PEC water oxidation with Ta 3 N 5 prepared from electrodeposited tantalum oxy(hydrate) film followed by ammonolysis (orange) and the control electrode (dark red) coated with CoPi. Inset is a photograph of control and electrodeposited films on a Ta substrate. the same ammonolysis condition s. The diffraction peaks are in an excellent agreement with the pure phase of Ta 3 N 5 which indicates that the as - deposited film was completely transformed into Ta 3 N 5 and the remaining peaks are due to the Ti substrate. In addition, no diffraction peak assoc iated to the nitrogen - poor tantalum nitrides were observed, which further indicates that during the cathodic electrodeposition the oxidation state of Ta 5+ was preserved. Consistent with XRD, the Raman spectra of the film after nitridization ( Figure A 2 - 13 ) confirms the formation of a pure phase Ta 3 N 5 . In order to test the PEC water oxidation activity of Ta 3 N 5 , the current density ( J ) vs. applied potential ( E ) measurements were performed on the electrodeposited thin films on Ta substrates 59 under chopped illumination. As a control experiment, a fresh Ta substrate was subjected to the same electrodeposition conditions without the Ta - IPA precu rsor followed by ammonolysis at 850 °C for 1 hour. The XRD diffraction of these electrodes are shown in Figure A 2 - 14 . Although, both electrodes have t he signature of Ta 3 N 5 , the peaks intensity for electrodeposited film is higher than the control sample which further suggest a higher degree of crystallinity and higher thickness for electrodeposited film. In addition, besides the associated peaks of Ta 3 N 5 , the control electrode shows some diffraction peaks that can be assigned to the nitrogen - poor phases of tantalum nitride. Interestingly, the sub - nitride layers were not observed for the electrodeposited electrode. The J - E curves of these electrodes are sh own in Figure 2 5 . Control samples produced little - to - no photocurrent at low applied potentials. On the other hand, the electrodeposited film resulted in a substantial increase of photocurrent at lower potentials with an onset photocurrent potential of ~ 0.6 V vs. RHE and a photocurrent of 0.4 mA cm - 2 at 1.23 V vs. RHE. 2.5. Conclusion The work presents the first example of the selective electrodeposition of tantalum(V) oxide thin films. This was achieved by inducing local pH increases at the electrode surface by either proton or NO 3 - reduction, with the latter being a significantly more efficient route. The resultant ele ctrodeposited films were nanostructured and hydrated, and can be converted to pure Ta 2 O 5 or Ta 3 N 5 by annealing in Air or NH 3 , respectively. The Ta 3 N 5 electrodes are active toward solar water oxidation. This methodology thus opens up the possibility of prod ucing Ta 3 N 5 electrodes on a variety of substrates. This is advantageous since the substrate can be nanostructured and/or transparent to improve light harvesting, charge transport and integration in a tandem configuration, which is not feasible when relying on a Ta metal substrate. For example, electrodeposition of 60 TaO x H y on nanostructured Ta - doped TiO 2 substrates, followed by ammonolysis, may allow for realization of highly efficient unassisted water splitting with a suitable small bandgap semiconductor bot tom junction. 11 61 A PPENDIX 62 Figure A 2 - 1 . Photographs of the electrodeposition solution containing Ta - IPA. The top pictures were recorded just after the solutions were prepared, the bottom pictures were taken after 12 hours sitting in ambient condition. 63 Figure A 2 - 2 . a) FT - IR spectra of IPA (black) and Ta - IPA (red), b) 1 H NMR of Ta - IPA. The FT - IR spectra of IPA and tantalum isopropoxide are compared in Figure A 2 - 2 .a characteristic vibration is located at 3330 cm - 1 . As it can be seen, after drying the Ta - IPA solution s are located in the range of 3000 - 1000 cm - 1 , therefore, as expected the 64 spectrum of both sample in this region are the same. The several peaks between 2975 cm 1 and 2850 cm 1 are due to C H stretching vibration of alkoxy groups. The peaks between at 1467 an d 1300 cm 1 correspond to the C H bending vibrations and vibration of methylene. Bands at 1 these bands for Ta - IPA are shifted to the lower energies which can be at tributed to the formation - IPA, the peak at 919 cm - 1 is assigned to in - plane rocking vibration of methyl groups. The broad envelopes of bands below 700 cm 1 stretching modes. 53,54 Figure A 2 - 2 .b shows the 1 HNMR spectrum of Ta - IPA in CDCl 3 with respect to TMS. The solvent peaks are observed at 7.26 ppm. Consistent with the IR spectrum, yet a fter drying there are some residual IPA. The associated peaks of the IPA were observed at 1.2 (d) ppm and 4.0 (m) ppm. These peaks are assigned to H atom of the term inal methyl and H in - position . The two broad peaks centered at 1.5 (m) and 5.0 (m) ppm ar e assigned to the isopropoxide groups bonded to Ta. 54 65 Figure A 2 - 3 . a) The CVs of 1 M KCl aqueous solution without (blue) and with KNO 3 (red) in neutral (dotted line) and acidic (solid line) solutions at pH 1.5. Figure A 2 - 4 . a) The CVs of solutions containing 1 M KCl ( red), 1 M KCl + 0.5 M KNO 3 (black), 1 M KCl + 50 mM Ta - IPA (green), and 1 M KCl + 0.5 M KNO 3 + 50 mM Ta - IPA (blue). b) CVs of 1 M KCl solution with varying concentration of HCl. Pt was used as working electrode with the scan rate of 100 mV s - 1 . 66 Figure A 2 - 5 . The CVs of electrodeposition bath containing 1 M KCl (red), 1 M KCl + 0.5 M KNO 3 (black), 1 M KCl + 50 mM Ta - IPA (green), and 1 M KCl + 0.5 M KNO 3 + 50 mM Ta - IPA (blue). Pt was used as working electrode with scan rate of 100 mV s - 1 and potential was swept from 0 .0 to - 1.0 V vs . Ag/AgCl. 67 Figure A 2 - 6 . The detailed XPS spectra Ta 4f, C 1s, and O 1 s for tantalum oxide deposited films different potential; - 1 .0 (black) and - 1.5 (green) V vs. Ag/AgCl. 68 Figure A 2 - 7 . a) The EDS spectra of the electrodeposited film on Ti substrate with (black) and wit hout (green) KNO 3 , b) the atomic percentage of Ta found for electrodeposited films with and without KNO 3 . Note: the percentage was calculated by only considering the Ta and Ti signals. 69 Figure A 2 - 8 . Cyclic voltammetry of an aqueous solution containin g K 3 [Fe(CN) 6 ] (1 mM) and KCl (0.5 M ) at scan rate of 100 m V s - 1 on Pt (gray) and TaO x H y /Pt (green) electrodes. TaO x H y was electrodeposited at - 1 .0 V vs . Ag/AgCl for 120 second. 70 Figure A 2 - 9 . a ) The photograph of bare Ti substrate (1), as deposited film at - 1 .0 V vs. Ag/AgCl for 1h (2), and the film after freeze - drying (3), b) SEM micro - images of electrodeposited film after annealing in Ar at 850 °C for 1 hour. Figure A 2 - 10 . The Raman spectr a of the as - deposited film (black) and KOH induced precipitation of TaO x H y (red) powder. 71 Figure A 2 - 11 . The XRD pattern of the powder (dark red) and electrodeposited thin film of tantalum oxide on Ti substrate after annealing in Ar (dark blue) at 850 °C for 1 hour. The XRD of Ti substrate (gray) after annealing in Ar is also shown as the reference; the vertical lin es represent crystalline Ta 2 O 5 with the PDF # 01 - 071 - 0639. The open triangle represent the crystalline phase of TiO 2 with PDF # 00 - 034 - 0180. The black star represents crystalline phases of suboxide phase of titanium with a range of compositions such as: T iO 0.48 (01 - 089 - 3074), TiO 0.325 (01 - 073 - 1581), Ti 6 O (01 - 072 - 1471), Ti 3 O 5 (01 - 072 - 0519), Ti 2.5 O 3 (03 - 065 - 6711), Ti 0.72 O 2 (01 - 086 - 1157). 72 Figure A 2 - 12 . The Raman spectra of the powder (dark red) and electrod eposited thin film of tantalum oxide on Ti substrate after annealing in air (cyan) and Ar (dark blue) at 850 °C for 1 hour. Figure A 2 - 13 . The Raman spectrum of the Ta 3 N 5 on Ti foil (purple) and powder Ta 3 N 5 (orange) Note: the films were nitridized by a mmonolysis at 850 °C for 1 hour. 73 Figure A 2 - 14 . The XRD diffraction pattern of Ta 3 N 5 on Ta substrate (orange) and Ta substrate (dark red). The vertic al dotted lines represent the B ra g g position s of Ta 3 N 5 with PDF # 01 - 089 - 5200. Note: the films were prepared by ammonolysis at 850 °C for 1 hour. 74 R EFERENCES 75 REFERENCES (1) Chun, W. J.; Ishikawa, A.; Fuj isawa, H.; Takata, T.; Kondo, J. N.; Hara, M.; Kawai, M.; Matsumoto, Y.; Domen, K. Conduction and Valence Band Positions of Ta 2 O 5 , TaON, and Ta 3 N 5 by UPS and Electrochemical Methods. J. Phys. Chem. B 2003 , 107 (8), 1798 1803. (2) Kamiyama, S.; Lesaicherre, P.; Suzuki, H.; Sakai, A.; Nishiyama, I.; Ishitani, A. Ultrathin Tantalum Oxide Capacitor Dielectric Layers Fabricated Using Rapid Thermal Nitridation Prior to Low Pressure Chemical Vapor Deposition. J. Electrochem. Soc. 1993 , 14 0 (6), 1617. (3) Frenning, G.; Nilsson, M.; Westlinder, J.; Niklasson, G. A.; Mattsson, M. Dielectric and Li Transport Properties of Electron Conducting and Non - Conducting Sputtered Amorphous Ta 2 O 5 Films. Electrochim. Acta 2001 , 46 (13 14), 2041 2046. (4) Frenning, G.; Engelmark, F.; Niklasson, G. a.; Strømme, M. Li Conduction in Sputtered Amorphous Ta 2 O 5 . J. Electrochem. Soc. 2001 , 148 (5), A418. (5) Wang, Y.; Kim, J.; Gao, Z.; Zandi, O.; Heo, S.; Banerjee, P.; Milliron, D. J. Disentangling Photochromis m and Electrochromism by Blocking Hole Transfer at the Electrolyte Interface. Chem. Mater. 2016 , 28 (20), 7198 7202. (6) Seo, J.; Cha, D.; Takanabe, K.; Kubota, J.; Domen, K. Highly - Dispersed Ta - Oxide Catalysts Prepared by Electrodeposition in a Non - Aqueo us Plating Bath for Polymer Electrolyte Fuel Cell Cathodes. Chem. Commun. 2012 , 48 (72), 9074 9076. (7) Seo, J.; Zhao, L.; Cha, D.; Takanabe, K.; Katayama, M.; Kubota, J.; Domen, K. Highly Dispersed TaO x Nanoparticles Prepared by Electrodeposition as Oxyg en Reduction Electrocatalysts for Polymer Electrolyte Fuel Cells. J. Phys. Chem. C 2013 , 117 (22), 11635 11646. (8) Liu, G.; Ye, S.; Yan, P.; Xiong, F.; Fu, P.; Wang, Z.; Chen, Z.; Shi, J.; Li, C. Enabling an Integrated Tantalum Nitride Photoanode to Approach the Theoretical Photocurrent Limit for Solar Water Splitting. Energy Environ. Sci. 2016 , 9 (4), 1327 1334. (9) Ho, C. T.; Low, K. Bin; Klie, R. F.; Maeda, K.; Domen, K.; Meyer, R. J.; Snee, P. T. Synthesis and Characterization of Semiconductor Tantalum Nitride Nanoparticles. J. Phys. Chem. C 2011 , 115 (3), 647 652. (10) Chen, Z.; Jaramillo, T. F.; Deutsch, T. G.; Kleiman - Shwarsctein, A.; Forman, A. J.; Gaillard, N.; Garland, R.; Takanabe, K.; Heske, C.; Sunkara, M.; et al. Accelerating Materials Development for Photoelectrochemical Hydrogen Production: Standards for Methods, Definitions, and Reporting Protocols. J. Mater. Res. 2010 , 25 (1), 3 16. (11) Hajibabaei, H.; Zandi, O.; Hamann, T. W. Tantalum Nitride Films Integrated with Transparent Conductive Oxide Substrates: Via Atomic Layer Deposition for Photoelectrochemical Water Splitting. Chem. Sci. 2016 , 7 (11), 6 760 6767. 76 (12) Ishikawa, A.; Takata, T.; Kondo, J. N.; Hara, M.; Domen, K. Electrochemical Behavior of Thin Ta 3 N 5 Semiconductor Film. J. Phys. Chem. B 2004 , 108 (30), 11049 11053. (13) De Respinis, M.; Fravventura, M.; Abdi, F. F.; Schreuders, H.; Savenije, T. J.; Smith, W. A.; Dam, B.; Van De Krol, R. Oxynitrogenography: Controlled Synthesis of Single - Phase Tantalum Oxynitride Photoabsorbers. Chem. Mater. 2015 , 27 (20), 7091 7099. (14) He, Y.; Thorne, J. E.; Wu, C. H.; Ma, P.; Du, C.; Dong, Q.; Guo, J.; Wang, D. What Limits the Performance of Ta 3 N 5 for Solar Water Splitting? Chem 2016 , 1 (4), 640 655. (15) Li, Y.; Zhang, L.; Torres - Pardo, A.; González - Calbet, J. M.; Ma, Y.; Oleynikov, P.; Terasaki, O.; Asahina, S.; Shima, M.; Cha, D.; et al. Cobalt Phosphate - Modified Barium - Doped Tantalum Nitride Nanorod Photoanode with 1.5% Solar Energy Conversion Efficienc y. Nat. Commun. 2013 , 4 , 2566. (16) El - Sayed, H. A.; Birss, V. I. Controlled Growth and Monitoring of Tantalum Oxide Nanostructures. Nanoscale 2010 , 2 (5), 793 798. (17) Singh, S.; Greiner, M. T.; Kruse, P. Robust Inorganic Membranes from Detachable Ultr athin Tantalum Oxide Films. Nano Lett. 2007 , 7 (9), 2676 2683. (18) Chandrasekharan, R.; Park, I.; Masel, R. I.; Shannon, M. A. Thermal Oxidation of Tantalum Films at Various Oxidation States from 300 to 700 °C. J. Appl. Phys. 2005 , 98 (11), 114908. (19) Lecuyer, S.; Quemerais, A.; Jezequel, G. Composition of Natural Oxide Films on Surf. Interface Anal. 1992 , 18 (4), 257 261. (20) Pinaud, B. A.; Vailionis, A.; Jaramillo, T. F. Contro lling the Structural and Optical Properties of Ta 3 N 5 Films through Nitridation Temperature and the Nature of the Ta Metal. Chem. Mater. 2014 , 26 (4), 1576 1582. (21) Kang, D.; Kim, T. W.; Kubota, S. R.; Cardiel, A. C.; Cha, H. G.; Choi, K. S. Electrochemi cal Synthesis of Photoelectrodes and Catalysts for Use in Solar Water Splitting. Chem. Rev. 2015 , 115 (23), 12839 12887. (22) Arnould, C.; Delhalle, J.; Mekhalif, Z. Multifunctional Hybrid Coating on Titanium towards Hydroxyapatite Growth: Electrodepositi on of Tantalum and Its Molecular Functionalization with Organophosphonic Acids Films. Electrochim. Acta 2008 , 53 (18), 5632 5638. (23) Zein El Abedin, S.; Welz - Biermann, U.; Endres, F. A Study on the Electrodeposition of Tantalum on NiTi Alloy in an Ionic Liquid and Corrosion Behaviour of the Coated Alloy. Electrochem. commun. 2005 , 7 (9), 941 946. (24) Borisenko, N.; Ispas, A.; Zschippa ng, E.; Liu, Q.; Zein El Abedin, S.; Bund, A.; Endres, F. In Situ STM and EQCM Studies of Tantalum Electrodeposition from TaF5in the Air - and Water - Stable Ionic Liquid 1 - Butyl - 1 - Methylpyrrolidinium Bis(Trifluoromethylsulfonyl)Amide. Electrochim. Acta 2009 , 54 (5), 1519 1528. 77 (25) Maho, A.; Delhalle, J.; Mekhalif, Z. Study of the Formation Process and the Characteristics of Tantalum Layers Electrodeposited on Nitinol Plates in the 1 - Butyl - 1 - Methylpyrrolidinium Bis(Trifluoromethylsulfonyl)Imide Ionic Liquid. Electrochim. Acta 2013 , 89 , 346 358. (26) Sheng Moo, J. G.; Awaludin, Z.; Okajima, T.; Ohsaka, T. An XPS Depth - Profile Study on Electrochemically Deposited TaOx. J. Solid State Electrochem. 2013 , 17 (12), 3115 3123. (27) Izaki, M.; Omi, T. Electrolyte O ptimization for Cathodic Growth of Zinc Oxide Films. J. Electrochem. Soc. 1996 , 143 (3), L53 -- L55. (28) Izaki, M.; Omi, T. Transparent Zinc Oxide Films Prepared by Electrochemical Reaction. Appl. Phys. Lett. 1996 , 68 (17), 2439 2440. (29) Cipris, D. Elec trochemical Synthesis of N - Alkylformamides. J. Electrochem. Soc. 1980 , 127 (5), 1045. (30) Switzer, J. A. Electrodeposition of Superlattices and Multilayers. In Electrochemistry of Nanomaterials ; Wiley - VCH Verlag GmbH: Weinheim, Germany, 2001; pp 67 101. (31) Peulon, S.; Lincot, D. Cathodic Electrodeposition from Aqueous Solution of Dense or Open - Structured Zinc Oxide Films. Adv. Mater. 1996 , 8 (2), 166 170. (32) Peulon, S. Mechanistic Study of Cathodic Electrodeposition of Zinc Oxide and Zinc Hydroxychl oride Films from Oxygenated Aqueous Zinc Chloride Solutions. J. Electrochem. Soc. 1998 , 145 (3), 864. (33) Mitchell, P. J.; Wilcox, G. D. An Electrochemical Route to Pre - Shaped Ceramic Bodies. Nature 1992 , 357 (6377), 395 397. (34) Choi, K. S.; Lichtenegger, H. C.; Stucky, G. D.; McFarland, E. W. Electrochemical Synthesis of Nanostructured ZnO Films Utilizing Self - Assembly of Surfactant Molecules at Solid - Liquid Interfaces. J. Am. Chem. Soc. 2002 , 124 (42), 12402 12403. (35) Govinda raju, G. V.; Wheeler, G. P.; Lee, D.; Choi, K. S. Methods for Electrochemical Synthesis and Photoelectrochemical Characterization for Photoelectrodes. Chem. Mater. 2017 , 29 (1), 355 370. (36) Zandi, O.; Schon, A. R.; Hajibabaei, H.; Hamann, T. W. Enhanced Charge Separation and Collection in High - Performance Electrodeposited Hematite Films. Chem. Mater. 2016 , 28 (3), 765 771. (37) Therese, G. H. A.; Kamath, P. V. Electrochemical Synthesis of Metal Oxides and Hydroxides. Chem. Mater. 2000 , 12 (5), 1195 1204 . (38) Papa, C. M.; Cesnik, A. J.; Evans, T. C.; Choi, K. S. Electrochemical Synthesis of Binary and Ternary Niobium - Containing Oxide Electrodes Using the p - Benzoquinone/Hydroquinone Redox Couple. Langmuir 2015 , 31 (34), 9502 9510. 78 (39) Louie, M. W.; Bell, A. T. An Investigation of Thin - Film Ni - Fe Oxide Catalysts for the Electrochemical Evolution of Oxygen. J. Am. Chem. Soc. 2013 , 135 (33), 12329 12337. (40) Klahr, B.; Gimenez, S.; Fabregat - Santiago, F.; Bisquert, J.; Hamann, T. W. Photo electrochemical and Impedance Spectroscopic Investigation of Water Oxidation with - - Coated Hematite Electrodes. J. Am. Chem. Soc. 2012 , 134 (40), 16693 16700. (41) Bradley, D. C.; Charkravarti, B. N.; Chatterjee, A. K.; Wardlaw, W.; Whitley, A. 19. Niobium and Tantalum Mixed Alkoxides. J. Chem. Soc. 1958 , No. 0, 99 101. (42) Callejas, J. F.; McEnaney, J. M.; Read, C. G.; Crompton, J. C.; Biacchi, A. J.; Popczun, E. J.; Gordon, T. R.; Lewis, N. S.; Schaak, R. E. Electrocatalytic and Photocatalytic Hy drogen Production from Acidic and Neutral - PH Aqueous Solutions Using Iron Phosphide Nanoparticles. ACS Nano 2014 , 8 (11), 11101 11107. (43) Cruickshank, A. C.; Tay, S. E. R.; Illy, B. N.; Da Campo, R.; Schumann, S.; Jones, T. S.; Heutz, S.; McLachlan, M. A.; McComb, D. W.; Riley, D. J.; et al. Electrodeposition of ZnO Nanostructures on Molecular Thin Films. Chem. Mater. 2011 , 23 (17), 3863 387 0. (44) Yoshida, T.; Tochimoto, M.; Schlettwein, D.; Wohrle, D.; Sugiura, T.; Minoura, H. Self - Assembly of Zinc Oxide Thin Films Modified with Tetrasulfonated Metallophthalocyanines by One - Step Electrodeposition. Chem. Mater. 1999 , 11 (10), 2657 2667. (45 ) Li, M.; Feng, C.; Zhang, Z.; Sugiura, N. Efficient Electrochemical Reduction of Nitrate to Nitrogen Using Ti/IrO 2 - Pt Anode and Different Cathodes. Electrochim. Acta 2009 , 54 (20), 4600 4606. (46) Popczun, E. J.; McKone, J. R.; Read, C. G.; Biacchi, A. J.; Wiltrout, A. M.; Lewis, N. S.; Schaak, R. E. Nanostructured Nickel Phosphide as an Electrocatalyst for the Hydrogen Evolution Reaction. J. Am. Chem. Soc. 2013 , 135 (25), 9267 9270. (47) Tsuda, T.; Hussey, C. L.; Milhano, C.; Pletcher, D. Modern Aspect s of Electrochemistry, No. 45 ; White, R. E., Ed.; Modern Aspects of Electrochemistry; Springer New York: New York, NY, 2009; Vol. 45. (48) Taguchi, S.; Feliu, J. M. Electrochemical Reduction of Nitrate on Pt(S)[n(1 1 1) × (1 1 1)] Electrodes in Perchloric Acid Solution. Electrochim. Acta 2007 , 52 (19), 6023 6033. (49) Tsang, C.; Kim, Y. G.; Gebregziabiher, D.; Stickney, J. Ta Surface Chemistry in Aqueous Solutions and the Possible Formation of TaTe 2 and TaS 3 . J. Electrochem. Soc. 2013 , 160 (12), D3278 D32 84. (50) Klahr, B. M.; Hamann, T. W. Current and Voltage Limiting Processes in Thin Film Hematite Electrodes. J. Phys. Chem. C 2011 , 115 (16), 8393 8399. (51) Joseph, C.; Bourson, P.; Fontana, M. D. Amorphous to Crystalline Transformation in Ta 2 O 5 Studied by Raman Spectroscopy. J. Raman Spectrosc. 2012 , 43 (8), 1146 1150. 79 (52) Chen, Y.; Fierro, J. L. G.; Tanaka, T.; Wachs, I. E. Supported Tantalum Oxide Catalysts: Synthesis, Physical Characterization, and Methanol Oxidation Chemical Probe Reactio n. J. Phys. Chem. B 2003 , 107 (22), 5243 5250. (53) Turova, N. Y.; Korolev, A. V.; Tchebukov, D. E.; Belokon, A. I.; Yanovsky, A. I.; Struchkov Maltese, Y. T. Tantalum(V) Alkoxides: Electrochemical Synthesis, Mass - Spectral Investigation and Oxoalkoxocompl exes. Polyhedron 1996 , 15 (21), 3869 3880. (54) Yang, S.; Chen, Y.; Yang, H.; Liu, Y.; Tang, M.; Qiu, G. Preparation of High - Purity Tantalum Ethoxide by Vacuum Distillation. Trans. Nonferrous Met. Soc. China 2008 , 18 (1), 196 201. 80 Tantalum Nit ride Films Integrated With Transparent Conductive Oxide Substrates via Atomic Layer Deposition for Photoelectrochemical Water Splitting Adapted With permission from: Tantalum nitride films integrated with transparent conductive oxide substrates via atomic layer deposition for photoelectrochemical water splitting, Hamed Hajibabaei , Omid Zandi and Thomas W. Hamann, Chem. Sci. , 2016 , 7, 6760 - 6767. Copyright 2016 Th e Royal Society of Chemistry. 81 3.1. Abstract Tantalum nitride, Ta 3 N 5 , is one of the most promising materials for solar energy driven water oxidation. One significant challenge of this material is the high temperature and long duration of ammonolysis previously required to synthesize it, which has so far prevented the use of transparent conduct ive oxide (TCO) substrates to be used which would allow sub - bandgap light to be transmitted to a photocathode. Here, we overcome this challenge by utilizing atomic layer deposition (ALD) to directly deposit tantalum oxynitride thin films, which can be full y converted to Ta 3 N 5 via ammonolysis at 750 °C for 30 minutes. This synthesis employs far more moderate conditions than previous reports of efficient Ta 3 N 5 photoanodes. Further, we report the first ALD of Ta - doped TiO 2 which we show is a viable TCO material that is stable under the relatively mild ammonolysis conditions employed. As a result, we report the first example of a Ta 3 N 5 electrode deposited on a TCO substrate, and the photoelectrochemical behavior. These results open the door to achieve eff icient overall water splitting using a Ta 3 N 5 photoanode. 3.2. Introduction Solar - driven photoelectrochemical (PEC) water splitting is a promising rout e to directly store solar energy in the chemical bonds of hydrogen . Due to the limita tion of available material s capable of overall PEC water splitting , a tandem cell is likely required to efficiently convert solar energy into hydrogen . 1,2 One promising tandem cell configuration is comprised of an n - type semiconductor as a photoanode to drive the oxygen evolution reaction that is electrically connected to a p - type photocathode to drive the hydr ogen evolution reaction ( Figure 3 1 ). This type of PEC cell is advantageous as it allows researchers to independently investigate and optimize each half - cell . Alt hough many semiconductor metal oxides have been proposed as a photoanode for solar water 82 oxidation , the majority of them have band ga p that lie s in the UV region which covers a negligible portion of the solar spectrum. 3 6 Metal oxide materials with narrower optical band ga p and absorption edge s that extend to the visible region, e.g. Fe 2 O 3 , 7,8 WO 3 , 9,10 and BiVO 4 , 11,12 have therefore attracted a lot of attention. The state - of - the - art electrodes using these materials have produced promising water oxidation photocurrent densities, with the best examples producing approximately 5 mA cm - 2 at 1.23 V vs. RHE. For example, Wang and co - workers recently employed a solution processed hematite photoanode in combination with an amorphous Si electrode to achieve overall water splitting at an efficiency of 0.91%. No metal oxide photoanode, however, has produced a photocurrent density that would enable achievin g ~ 10% water splitting efficiency. Figure 3 1 . The tandem cell configuration for overall water splitting composed of n - type, and p - type semiconductor connected through a transparent and conductive layer. 83 Domen has recently introduced a new class of nitride semiconductors , specifically tantalum nitride (Ta 3 N 5 ), as promising alternative candidate s to oxides for PEC water oxidation. 13 19 In one impressive example, they demonstrated 1.5 % efficient solar water splitting with Ba - doped Ta 3 N 5 nanorods nitridized at 1000 °C for 2 h . 20 Tantalum nitride is intrinsically a n - type semiconductor with an optical band gap of 2.1 eV that theoretically corresponds to a maximum photocurrent density of 12.5 mA cm - 2 . 21,22 If it is coupled with an appropriate photocathode in a PEC tandem cell, it could perform unassisted water splitting at a solar - to - hydrogen efficiency of ~ 15 % . 23 Strikingly, Li and coworkers recently reported a Ta 3 N 5 photoanode on Ta foil prepared by ammonolysis at 950 °C for 6 h that produced a photocurrent density of ~ 12.1 mA cm - 2 at 1.23 V vs . RHE with a photocu rrent onset potential of ~ 0.7 V vs. RHE . 24 Implementing a Ta 3 N 5 photoanode in a tandem configuration to achieve efficient overall water splitting is hindered by the lack of a synthetic procedure to prepare Ta 3 N 5 electrodes under conditions compatible with a transparent conductive oxide (TCO) substrate. Most of the studies on tantalum nitride (Ta 3 N 5 ) share a similar synthetic route, beginning wit h the oxidation of Ta(0) to Ta( V), followed by ammonolysis at elevated temperatures (> 8 0 0 °C) for long periods of time ( > 6 h ), as noted in the best literature examples provided above. Despite of the simplicity and great performance that comes with this method, there are multiple negative consequences. The initial o xidation of tantalum metal is very energy intensive. Importantly, the subsequent ammonolysis prevents the use of a TCO substrate as noted by others. 18 In our la b, we found that when conventional TCOs such as FTO (F - doped SnO 2 ), ITO (Sn - doped In 2 O 3 ), and AZO (Al - doped ZnO) are exposed to ammonia at 600 °C or higher, they are reduced to metallic phases and become flaky with weak adhesion to the substrate. Consequen tly, their important properties of conductivity and transparency are lost. Therefore , Ta 3 N 5 is commonly prepared on Ta foil which excludes the 84 applicability of the Ta 3 N 5 as a photoanode in a tandem configuration ( Figure 3 1 ), since the substrate is not transparent to sub - bandgap light . Further more , high - temperature ammonolysis makes it difficult to control the morphology, interfaces and the inherent properties of this semiconductor. In order to overcome these issues, we synthesized Ta - doped TiO 2 (TTO) films via atomic layer deposition (ALD) which we found to be a stable TCO in reducing atmospheres. In addition, to circumvent the high - temperature ammon olysis, ALD was also used to directly deposit thin films of Ta 3 N 5 on the TTO substrates. While initial as - deposited films are primarily amorphous TaO x N y , these films can be nitridized to Ta 3 N 5 at far more moderate nitridation conditions, i.e. 750 °C for 30 minutes, compared to previous reports where hours (>6 h) of nitridation at temperatures higher than 800 °C were necessary. The photoelectrochemical properties of the Ta 3 N 5 films deposited on TTO were investigated and the PEC water oxidation performance wa s analyzed. The excellent material control reported here allowed for a detailed material structure - function relationship to be determined and a path to improved performance elucidated. 3.3. Experimental 3.3.1. Film preparation Thin films of Ta - doped TiO 2 TCO film s wer e prepared on quartz substrates (Advalue Technology) by alternating the deposition of TiO 2 and TaO x with four diff erent ratios of TaO x :TiO x ALD cycles, 1:200, 1:150, 1:100, and 1:50, to modify the dopant concentration. TiO 2 was deposited using a modified literature procedure; 25 brief ly, titanium isopropoxide (99.9 %, Aldrich) was heated to 80 °C and pulsed for 2 s. After purging for 10 s, water was pulsed for 15 ms followed by purging for another 10 s. The growth rate of TiO 2 at 250 °C was found to be 0.2 Å /cycle. The deposition of 85 the TaO x sub - cycles is described below. The as - deposited TTO films were subsequently annealed under an ammonia atmosphere at 750 °C for 3 0 minutes with a heating rate of 35 °C/ min and cooled down to the room temperature by opening up the top cover of the tube furnace. TaO x N y and TaO x films were deposited on quartz , silicon (University Wafer, with ~ 16 Å native SiO 2 ) or the TTO coated quart z substrates described above using ALD (Savannah 200, Cambridge Nanotech Inc). All substrates were sequentially sonicated for 15 minutes in soap ( ) , DI water and isopropyl alcohol , then blown dr y under a nitrogen flow and loaded into the ALD chamber. High purity nitrogen was used as a carrier gas, which was further dried and deoxygenated by in - line molecular sieves 3 Å (Sigma Aldrich) and an O 2 scrubber (Restek), respectively. Throughout the deposition, the N 2 flow rate wa s adjusted at 5 SCCM, providing a constant pressure of ~ 350 m T orr. Pentakis(dimethylamine) t antalum(V), Ta(N(CH 3 ) 2 ) 5 (PDMAT), ( 99.9%, Aldrich) was used as the tantalum precursor. Monomethylhydrazine , CH 3 NHNH 2 were used as the co - reactant s . The tantalum precursor, PDMAT, was kept at 90 °C and consecutively pulsed 5 times for 2 s duration with 10 s purging in between pulses . The MMH and DI water c o - reactant s were kept at ambient temperature. Nitridation or oxidation was performed by a 15 ms pulse of MMH or water followed by purging for 15 s to complete one ALD cycle. Films were annealed in an ammonia atmosphere at 750 °C for 30 min to complete the nitridatio n and crystalize the films . 3.3.2. Film Characterization Films thickness were determined via spectroscopic ellipsometry (SE) using Horiba Jobin Yvon, Smart - SE instrument. X - ray photoelectron spectroscopy (XPS) was performed with a Perkin 86 Elmer Phi 5600 ESCA syste takeoff angle of 45°. Survey scans of 0 - 1100 eV binding energy and detailed scans for C 1s, O 1s, N 1s and Ta 4f, Ti 2p regions were measured for all samples. The binding energies were corr ected in reference to C 1s peak (284.8 eV) and shirley background subtraction was performed for fitting for each sample. Absorbance spectra were collected on PerkinElmer Lambda35 UV - vis spectrometer equipped with Labsphere integrating sphere. Raman spectra were recorded using LabRam Armis, Horiba Jobin Yvon instrument equipped with 532 nm laser and a ×50 microscope to focus the laser on the film surface. X - ray diffraction (XRD) patterns were obtained on a Bruker D8 Advanced diffractometer using Cu radiation - probe electrical measurements were performed using computer controlled Pro4 - 440N system equipped with Keithley 2400, and Pro4 software. The film thickness was also measured by cross - section SEM (Carl Zeiss Auriga, Dual Column FIBSEM) and was taken at a tilt angle of 90°. 3.3.3. Photoelectrochemical Measurements All electrodes were coated with the Co - Pi co - catalyst via photoelectrodeposition prior to carrying out further PEC measurements. The Co - Pi co - catalyst was deposited fro m a solution with 0.5 mM Co(NO 3 ) 2 in a 0.1 M potassium phosphate buffer at pH 7 at a constant potential of 1.06 vs. RHE for 180 s under AM 1.5 G simulated sunlight. A Ag/AgCl and hi gh surface area platinum mesh electrodes were used as the reference and cou nter electrode s, respectively. Photoelectrochemical measurements were made with Eco Chemie Autolab potentiostat coupled with Nova electrochemical software. The light source was a 450 W Xe arc lamp (Horiba Jobin Yvon). An AM 1.5 solar filter was used to simulate sunlight at 100 mWcm - 2 (1 sun). All the photoelectrochemical measurements were performed by shining light on the electrodes through 87 electrolyte. Current - voltage curves were measured using a scan rate of 10 mV s - 1 . The incident light was chopped using a computer controlled Thor Lab s solenoid shutter. Electrodes were masked with a 60 µm Surlyn film ( solaronix ) with a 0.28 cm 2 hole which was adhered to the electrode by heating to 120 °C. The protected electrodes were clamped to a custom made glass electrochemical cell with a quartz wi ndow. A homemade saturated Ag/AgCl electrode was used as the reference electrode and was frequently calibrated to a commercial saturated calomel electrode (Koslow Scientific). Potential vs. Ag/AgCl were converted to reversible hydrogen electrode (RHE) by t he equation . An aqueous solution of 0.5 M K 2 HPO 4 was used as the electrolyte. The pH of the electrolyte was adjusted to 13 by adding KOH. A high surface area platinum mesh was used as the counter electrode. 3.4. Results an d Discussion 3.4.1. ALD of TTO ALD was used to deposit Ta - doped TiO 2 (TTO) on quartz substrates. D ifferent Ta concentration s were introduced by varying the relative number of ALD cycles of TiO 2 and TaO x . Samples with TaO x :TiO 2 ALD sub - cycle ratios of 1:50, 1:100, 1:150 and 1:200 were prepared to produce a series of decreasing Ta dopant concentrations in TiO 2 . In addition, pure TiO 2 films were prepared as control substrates. The total number of cycles were controlled to keep the final film thickness constant at 100 nm . Energy dispersive spectroscopy ( EDS ) was used to determine the resultant concentration of Ta in TiO 2 (M: 1.809 keV), TTO films were also deposit - 26 T he EDS spectra of these films with different concentrati on of Ta are shown in Figure A 3 - 1 . The Ta concentration in the 1:200 film was below 88 the detec tion limit of the instrument, so it was not included in this plot, however, all the observed signals for the other three films are readily assigned to Ta, Ti, O, and Al (substrate). As shown in Figure A 3 - 1 , t he atomic percent of Ta was found to increase linearly with the ALD sub - cycle ratio of TaO x :TiO 2 . T he atomic percentage s of Ta in the films, then , were calculated from a linear fit of these data and used to assign the following percent of Ta contained in the films : 5.0 ( ± 0.32 ) , 2.5 ( ± 0.16 ) , 1.67 ( ± 0.11 ) and 1.25 ( ± 0.08 ) . We note that the actual concentration of Ta contained in the TTO films does not correspond simply to the pulse ratios of Ta and Ti precursors. The difference can largely be accounted for by the different growth rates; ~ 0.25 Å/cycle for TiO 2 compared to ~ 0.79 Å/cycles for TaOx, vid e infra . The resistivity of as - deposited Ta - doped TiO 2 films on quartz were o n the order · cm. In addition, consistent with a previous study, we observed that when the Ta - doped TiO 2 films were annealed in air or oxygen, they became more insulating. 27 Prior examples of Ta - doped TiO 2 were prepared at low oxygen pressure, e.g. 1 0 - 5 Torr, or the films were annealed in vacuum. 27 29 Since our ultimate goal is to realize TCO films coated with Ta 3 N 5 , which may have to be annealed under ammonia, vide infra , all TCO films were annealed under a reducing ammonia atmosphere at 750 °C for 30 minutes . XPS measurements were performed on samples deposited on quartz both before and after annealing in ammonia. T he surface concentration of Ta for as - deposited films are higher compared to the results from EDS measurements (see Figure A 3 - 2 ) . Since XPS is a surface sensitive technique, this higher apparent concentration of Ta may be attributed to the fact that the deposition of TaO x was the last ALD cycle of these films . After annealing in ammonia, however, the atomic ratio of Ta/Ti determined by XPS was within error of the ratio determined by EDS on the as - deposited samples. Thus, annealing allows Ta to diffuse and be homogeneously distribute d 89 throughout the film. We therefore take the surface compositional analysis done by XPS after annealing as a good approximation of bulk composition. Details of t he XPS analysis of as - deposited and annealed TTO films with different concentration s of Ta are discussed followi ng Figure A 3 - 3 in the supporting information . The atomic percentages of oxygen and nitrogen as a function of Ta concentration after annealing in ammon ia are shown in Figure 3 2 . a . After annealing in ammonia the atomic percentage of O decreased and a new N signal emerged which indicates oxygen i s substituted by nitrogen in the films . Thus, the annealing step results in TiO 2 co - doped with Ta and N. Interestingly, at high concentration of Ta, i.e. ~ 5 % , a nother N signal is detectable which can be assigned to a Ta - N bond. Further, the Ta signal from t he same film shows two types of Ta present in the films. Therefore, we attribute this to the formation of TaN x as a separate phase at high Ta concentration s . This observation is supported by XRD results of the films an d the resistivity of the films discuss ed below. The XRD diffraction patterns of all annealed samples were unambiguously assigned to anatase TiO 2 . Detailed analysis of the XRD patterns of the N - and Ta - co - doped TiO 2 films with different Ta concentrations are discussed in the supporting inform ation following Figure A 3 - 4 . Depending on dopant concentration, however, the peak positions of anatase are shifted to the lower angles which indicate s an increase in cell volume as expected from doping Ta into TiO 2 . 30 The resistivity of the TTO films as a function of concentration of Ta are shown in Figure 3 2 . a. The resistivity decreases sharply with introduction of Ta, reaching a minimum for the film with 1.6% Ta . This is ~ 3 times smaller than the optimum Ta concentration reported in the literature. 28,31 The main difference between the Ta - doped TiO 2 synthesized in this study to those reported in literature is t he annealing atmosphere. As noted above, use of ammonia as the reducing atmosphere results in TiO 2 films co - doped with Ta and N. As depicted in Figure 3 2 .a , the resistivity of the 90 films has a strong correlation to the atomic concentration of oxygen and n itrogen. The film without Ta exhibits a surprisingly low resistivity which results from the formation of nitrogen - doped TiO 2 or segregation of metallic TiN phases. The lowest resistivity occurs for the film with 1.6% Ta, which has the highest concentration of nitrogen and the lowest concentration of oxygen, i.e. highest concentration of oxygen vacancies. Based on the formal cha rge of oxygen and nitrogen, it can be inferred that substitution of oxygen with nitrogen induces an increase in the concentration of oxygen vacancies. On the other hand, substitution of Ti 4+ with Ta 5+ may reduce the number of oxygen vacancies. Therefore , c o - doping of N and Ta into TiO 2 may have an opposing influence on carrier concentration and conductivity, which explains the difference between the optimal doping concentration found here compared to prior reports . 27 Figure 3 2 . a) Resistivity (brown) and atomic % of O (blue) and N (red) of Ta - doped TiO 2 as a function of Ta%. b) Optical transmittance of Ta - doped TiO 2 films, u n - doped TiO 2 (orange) , 5.0% Ta (pink), 2.5 % Ta (Cyan), 1 .67 % Ta (red) and 1.25 % Ta (black). c) Photograph of the TTO films after annealing in ammonia at 750 °C for 30 min . 91 The optical transmittance of un - doped and Ta - doped TiO 2 thin films after annealing in ammonia are shown in Figure 3 2 . b. Note that t hese transmitta nce s were not corrected for reflectance , which accounts for ~ 25 % loss of incident pho tons ( Figure A 3 - 5 ). The transmittance of the TiO 2 without Ta was below 50% in the visible region, which is in line with numerous reports of N - doped TiO 2 . 32,33 Substitution of oxygen with nitrogen introduces new states in the band gap which results in absorption edge tailing to the visibl e region. Upon Ta - doping, however, the average transmittances in the visible region are increased with a maximum transmittance of ~ 70 % for 1.6 % Ta - doped in TiO 2 . 3.4.2. ALD of Ta 3 N 5 TaO x films were deposited from 50 500 ALD cycles at 250 °C . The resultant film thickness increases linearly with the number of ALD cycles ( Figure A 3 - 6 ). T he growth rate found to be 0.79 Å/cycle , which is in g ood agreement with the previous report of the ALD deposition of TaO x ( 0.85 Å/cycle ) . 34 The ALD of tantalum nitride using PDMAT and MMH has previously been st udied, where an ALD deposition temperature window between 200 - 300 °C was found with a growth rate of ~ 0.3 Å/ cycle. 35 It has been previously reported that PDMA suffers from thermal decomposition at temperatures above 300 °C, therefore, to avoid the decomposition of the precursor and ensure an ALD process, 280 °C was used as the maximum deposition temperature. 34,35 Interestingly, while we confirmed an ALD temperature window over 175 to 280 °C, we found a growth rate approximately three times larger with a small temperature dependence. Plots of thickness vs. number o f ALD cycles are provided in Figure A 3 - 7 . From the slope of these plots, growth rates were found to reproducibly vary from 0.86 Å/cycle at 175 °C to 1 .04 Å/cycle at 280 °C, as shown in Figure 3 3 . 35,36 These growth rates were confirm ed by cr oss - section SEM 92 Figure 3 3 . Growth rate of TaO x N y (green) and TaO x ( pink ) as a function of deposition temperature . Also shown are t he ratios of atomic concentrations of O / N ( purple ) as a function of the deposition temperature found from EDS analysis . measurements of films grown with 1000 ALD cycles at 175 and 280 °C , shown in Figure A 3 - 8 . The images indicate a ~ 28 nm difference in film thicknesses which is consistent with the different growth rates displayed in Figure 3 3 . In addition to the growth rate, the temperature affected the composition of the deposited film s. The bulk composition of the as - deposited films was analyzed by EDS ( Figure A 3 - 9 ). Silicon with ~ 16 Å SiO 2 was used as the substrate. The Ta and Si signals overlap which prevents accurate determinations of these individual elements . The atomic percentage s of nitrogen and oxygen were calculated based on the signal of these two elements and shown in Figure 3 3 . Oxygen was detected in all films. We note that only a mini mal amount of oxygen can be attributed to the ~ 16 Å SiO 2 substrate since the film thicknesses are ~ 100 nm . Thus, despite the lack of oxygen in either 93 ALD precursors, and use of high purity nitrogen as a carrier gas, all deposited films are actually TaO x N y . Thus, there must be some source of oxygen which we were not able to fully eliminate despite significant efforts to control the ALD atmosphere . Further, as the deposition temperature increases from 200 to 2 5 0 °C t he re is a change in the relative percenta ge of oxygen and nitrogen; the relative amount of O compared to N decreases from ~ 65 to ~ 25%. Figure 3 4 . a) XPS signals of O 1s, N 1 s , and Ta 4f , b) calculated atomic percentages of Ta - N (red) and Ta - O (purple) as a function of deposition temperature. The surface composition of the as - deposited films was also analyzed by XPS. Fitted spectra are shown in Figure 3 4 . a. A s the deposition temperature increases , the N1s signal grows and it can only be fitted to a single Ta - N peak . The oxygen signal was fi tted to 3 peaks. Two p eak s with binding energ ies > 53 1 eV were assigned to carbon species , i.e. C - OH and C =O groups . The peak at 529 - 530.5 eV was assigned to the Ta - O group which was correlate d to the Ta 4f peak. The 94 peak positions of Ta 4f 7/2 and Ta 4f 5/2 strongly depend on the immediate surrounding atoms , e.g. ~ 26.6 and 28.5 eV for Ta - O and 25.0 eV and 26.9 eV for Ta - N, respectively . 37 Therefore to avoid complexity arising from carbon species, the surface atomic percentages of Ta - O and Ta - N were estimated from the Ta 4f peaks ( Figure 3 4 . b). As it can be seen, at lower temperatures the film is mostly composed of Ta - O groups . On the other hand, at higher deposition temperatur e s , the Ta - N bec o me s the dominant composition . This result is in line with EDS analysis discussed earlier. Th ese combined result s are also in agreement with the previous study by Ritala et al . who studied the deposition of thin films of Ta 3 N 5 at temperatur es from 200 to 500 °C via ALD using TaCl 5 and NH 3 as the reactants. 38 Their results showed that the composition of the films were strongly correlat ed to the deposition temperature and the concentration of oxygen was decreased from 25 to ~ 5 % as the deposition temperature increased from 200 to 500 °C. The composition and growth rates of the films deposited at 280 and 250 °C are similar; since we found 280 °C to be the edge of the ALD temperature window, all the subsequent depositions of TaO x N y were performed at 250 °C unless otherwise mentioned. A l ack of diffraction peak s in XRD and phonon modes in Raman spectrum of as - deposited films indicate that th ese films are in fact amorphous TaO x N y ( Figure A 3 - 10 ). Therefore, to improve the crystallinity and to modify the composition, th ey were annealed ammon ia . T here are three pa rameters which control the results of annealing; temperature, time and flow rate of ammonia. It was found that the optimum conditions (details discussed surrounding Figure A 3 - 11 to Figure A 3 - 13 ) to form pure crystalline Ta 3 N 5 films from as - deposited films is amm onolysis at 750 °C for 30 min with ammonia flow rate mL/ min. It is worth noting that both ALD deposited thin films of TaO x and TaO x N y were nitridized to Ta 3 N 5 ( Figure A 3 - 14 ) at far more moderate conditions compared to previous reports . 20,39,40 95 Figure 3 5 . Plots of absorbance of thin films of Ta 3 N 5 on quartz with different thicknesses : 50 nm (blue), 70 nm (orange), 99 nm (red), and 122 nm (green). Inset is the absorbance at 350 nm vs. the thickness of the films. Four TaO x N y films of different thicknesses were deposited on quartz followed by ammonolysis at 750 °C for 2 hours. Based on the XRD patterns of the films ( Figure A 3 - 15 ) , they can all be unambiguously matched to Ta 3 N 5 . The thicknesses of Ta 3 N 5 films were evaluated via both cross - section SEM and SE ( Figure A 3 - 16 ). As shown in Figure A 3 - 16 .d, the growth rate found by both methods are in good agreement. However, t he growth rate of pure Ta 3 N 5 , i.e. ALD deposition followed by ammonolysis, was ~ 0. 77 Å/cycle while the growth rate of as - deposited films is ~ 1.0 Å/cycle. This discrepancy in the growth rates is due to the fact that the as - deposited films are amorphous TaO x N y , whereas ammonolysis transforms the films to crystalline Ta 3 N 5 which has 22% smaller molar volume per Ta atom than Ta 2 O 5. 41,42 96 The absorbance of Ta 3 N 5 as a function of the thickness is plotted in Figure 3 5 ( absorptance, transmittance and reflectance are shown in Figure A 3 - 17 ). T he absorbance was corrected for the substrate using a previously developed model. 43 T he absorbance scale s linearly with the film thickness confirming a linear growth of tantalum nitride by ALD/ammonolysis. The absorption coefficient, - 1 ), was calculated from absorbance using the average film thicknesses from SEM and SE ( Figure A 3 - 18 .a ). Ta 3 N 5 has two optical transitions, located at ~ 2.10 eV and ~ 2.50 eV. A recent study on optoelectronic properties of Ta 3 N 5 suggest that both electronic transitions of Ta 3 N 5 are direct. 44 The corresponding Tauc plots for direct tr ansitions is shown in Figure A 3 - 18 b. To study the PEC performance, ~ 75 nm (1000 cycles) of TaO x N y was deposited on 100 nm TTO films with different Ta concentrations, followed by ammonolysis at 750 °C for 30 minutes. Attempts to increase the ammonolysis time to 2 hours resulted in transformation of the anatase - TiO 2 to rutile - TiO 2 in the TTO, based on the XRD patterns of the films, which resulted in an electrode with negligible photocurrent ( Figure A 3 - 19 ). Our initial results produce a photocurrent density of ~ 0.77 mA cm - 2 at 1.23 V vs . RHE with an onset photocurrent potential of ~ 0.8 V vs. RHE. T he PEC performance of these electrodes are strongly correlated to the conductivity of TTO substrates . The photocurrent response of electrodes at 1.23 V vs. RHE as a function of Ta concentration are shown i n Figure A 3 - 21 (the XRD pattern of the films are shown in Figure A 3 - 20 ) . Interestingly, the observed photocurren ts are in good agreement with the conductivity of TTO shown in Figure 3 2 . a - i.e. the least resistive TTO substrate, produces the highest photocurre nt - suggesting that the performance of these electrodes are controlled by the conductivity of the TTO substrate. This performance falls short of the recent report by Li and coworkers, 24 who reported a photo current density of ~ 12.1 mA cm - 2 at 1.23 V vs . RHE with a 97 photocurrent onset potential of ~ 0.7 V vs . RHE for the electrode prepared on Ta foil which was nitridized under ammonia at 950 °C for 6 hrs. Van de Krol et al also recently studied the formation o f Ta 3 N 5 as a function ammonolysis conditions on Pt foil. 14 The maximum ph otocurrent density of ~ 1.1 mA cm - 2 at 1.23 V vs . RHE with an onset photocurrent potential of ~ 0.9 V vs . RHE was found for the Ta 3 N 5 film prepared at 800 °C 10 hr s after IrO 2 treatment. To the best of our knowledge, however, this is the first report of PE C water oxidation of Ta 3 N 5 on any TCO ( Figure 3 6 ). Figure 3 6 . PEC performance of CoPi modified Ta 3 N 5 (~ 75 nm) on TTO with 1.6 % Ta concentration under 1 sun illumination. The inset is the photograph of the working electrode. Finally, since we have not yet eliminated the ammonolysis step in the synthesis, we compared the behavior of the TaO x N y depos ited films to TaO x . 40 nm of TaO x was deposited on the best TTO ( 1.6% Ta ) followed by ammonolysis at 750 °C for 30 minutes. The transmittance/reflectance 98 spectrum of the corresponding films are compared to that of TaO x N y - derived film in Figure A 3 - 22 . The TaO x - derived film is colorless with a take - off transmittance at ~ 450 nm. On the other hand, TaO x N y - derived film is orange with a take - off transmittance at ~ 590 nm which correspond s to the known band gap of Ta 3 N 5 , i.e. 2.1 eV, discussed above . The PEC performance of these electrodes are compared in Figure A 3 - 23 . The TaO x derived fil m shows negligible photocurrent superimposed on a large dark current. Therefore, it can be concluded that the TaO x - derived films require harsher nitridization conditions (higher temperature and longer duration ) where the TTO is not chemically stable. 3.5. Concl usion s The realization of photoactive Ta 3 N 5 films on a TCO electrode was demonstrated for the first time. This required two breakthroughs. First, we established the ALD of TTO and found that it is structurally and chemically stable (unlike conventional TCO materials including FTO, ITO and AZO) unde r the reducing atmosphere employed (ammonolysis at 750 °C for 30 min), and can therefore be used as a conductive transparent layer for tantalum nitride electrodes. The TTO films were not able to withstand the harsher nitridation conditions which must be em ployed for Ta 3 N 5 films produced from TaO x , however, which is the synthetic route of all prior examples of Ta 3 N 5 photoelectrodes. Therefore, the second necessary breakthrough consisted of the direct deposition of TaO x N y films via ALD which can be crystalliz ed and completely converted to Ta 3 N 5 under sufficiently milder ammonolysis conditions to maintain the TCO properties of TTO. The resultant Ta 3 N 5 films on TTO produced promising solar water oxidation performance, especially considering the films are quite t hin and not yet optimized. We found that the performance of the photoelectrodes correlated to the conductivity of the TCO. Thus, it would beneficial to utilize state - 99 of - the - art TCOs such as FTO. Since these are not stable under even the mildest ammonolysis procedures utilized here, it would clearly be advantageous to directly deposit crystalline Ta 3 N 5 films under sufficiently mild conditions on a TCO, which do not require a subsequent annealing (thus ammonolysis) step. While we are actively working on such an ideal synthetic method, the results reported herein represent a significant step towards realizing a high - efficiency solar water oxidation electrode which can be employed in a tandem configuration. 100 A PPENDIX 101 Figure A 3 - 1 . EDS spectrum of Ta - doped TiO 2 as a function of ratio of cycles of TaO x to TiO 2 , a) 1: 150, b) 1 : 100, and c) 1 : 50. d) Atomic percentage of Ta as a function of ratio of cycles of TaO x :TiO 2 . 102 Figure A 3 - 2 . Comparison of atomic percentage of Ta as a function of ratio of cycles of TaO x /TiO 2 found by EDS and XPS. 103 Figure A 3 - 3 . XPS spectra of Ta - doped TiO 2 as a function of Ta % doping, a) as deposited, b) after annealing in ammonia at 750 °C for 30 min. 104 The as - deposited films are only comprised of Ti, O and, Ta. After annealing in ammonia, nitrogen signal is also observed . The main features of XPS spectra for the Ta - doped TiO 2 would be better captured by analyzing each individual elements before and after annealing in ammonia. Ti: The Ti main peak before and after annealing in ammonia were located at the same position. Upon annealing in ammonia, however, two new peaks associated to Ti - O - N (blue) and Ti - N (green peak) groups were appeared which confirms that beside s Ta, N is also doped into TiO 2 . 45 47 O: As expected, oxygen peak ( metal - oxygen bond) at 29 - 30 eV was not affected by annealing in ammonia, however, the atomic percentage of oxygen was changed by both annealing in ammonia and the atom ic percentage of Ta. The other o xygen peaks at h igher binding energies were assigned to C=O and C - OH groups. N: As it can be seen, the as - deposited films did not have any nitrogen, however, after annealing in ammonia the nitrogen signal was emerged. At high concentration of Ta, there is a small shoulde r (orange peak), which further indicates the formation of a different type metal - N bonding. Ta: Before annealing in ammonia, there was only one type of Ta present which can be associated to Ta - O groups. Upon annealing in ammonia, however, Ta signals became broadened which is an indication of doping of Ta into the structure with a range of atomic interaction with its neighbors. The Ta signal for the film with high concentration of Ta, i.e. 5 % Ta, could not be fitted to one peak, and it was fitted to two peaks. Therefore, there are two types of Ta groups . This observation is in line with the extra N peak for the film with high concentration of Ta. Then it was hypothesized that at high Ta % doping, TaN x would segregate off from TiO 2 and form a separate phase. The atomic percentages of Ta, Ti, O and N were calculated using the fo l lowing equation: 105 A 3 - 1 The S x is the normalized peak area associated to each element. For normalization, the raw peak area was divided to the sensitivity factor. The sensitivity factor of Ti, Ta, N, and O are 52.013, 75.608, 12.231, and 18.643, respectively. 106 Figure A 3 - 4 . a) XRD and b) calculated cell volume of TiO 2 as a function of Ta % doping. Un - doped TiO 2 (orange). 5.0 % Ta ± 0.32 ( pink), 2.25 % Ta ± 0.0.16 ( cyan), 1.67 % Ta ± 0.11 (red) , and 1.25 % Ta ± 0.08 (black). The diffraction patterns were unambiguously matched to anatase TiO 2 . Since the ionic radii of Ta 5+ is slightly larger than Ti 4+ (0.64 vs. 0.60), 48 therefore, upon doping TiO 2 with Ta the unit cell would slightly expan d. The films with low and high Ta % doping as well as un - doped samples hav e almost the same peak positions. This observation suggests that these two films have the lowest Ta incorporation. The slight peak shift of lightly doped film, 1.25 % Ta doping, can be ascribed to the low concentration of Ta to make a detectable alteration in anatase structure. On the other hand, the heavily doped film, 5 % Ta doping, has a high concentration of Ta which makes it energetically more favorable to form a separate phase of tantalum (oxy)nitride. This is consistent with the extra N and Ta p eaks ob served in XPS ( Figure A 3 - 3 ). On the other hand, for the films with Ta % doping in between, however, the diffraction peaks are shifted to the lower ang e l s . GSAS 49,50 was used to 107 calculate the cell parameters. The calculated volumes vs . Ta % doping level are shown in Figure A 3 - 4 . Figure A 3 - 5 . a) reflectance and corrected transmittance of TTO as a function of Ta concentration. Films were annealed in ammonia at 750 ° C for 30 min. 200/1 (pink), 150/1 ( c yan), 100/1 (red), 50/1 (black), and un - doped TiO 2 (orange). 108 Figure A 3 - 6 . Thickness of TaO x vs . the number of cycles deposited at 250 °C. Figure A 3 - 7 . Thickness of TaO x N y vs . number of cycles deposited at a) 175, b) 200, c) 250, and d) 280 °C. 109 Figure A 3 - 8 . Cross - section SEM of 1000 cycles of TaO x N y deposited at a) 280 and b) 175 °C. 110 Figure A 3 - 9 . EDS spectrum of as - deposited TaO x N y as a function of deposition temperature. Silicon wafer with 16 Å native SiO 2 was used as the substrate. Ta main peak is buried under large Si signal. 111 Figure A 3 - 10 . a) XRD and b) Raman scattering of 100 nm TaO x N y deposited on FTO at 250 °C. The red curve represent the reference spectrum of crystalline Ta 3 N 5 . 112 Figure A 3 - 11 . The XRD of TaO x N y after annealing in ammonia at different temperature for 10 hours. The vertical das hed lines represent the B ra g g position s of Ta 3 N 5 with PDF number of 01 - 089 - 5200. To eliminate the effect of flow rate, high throughput ammonia (~ 500 SCCM) was used and the annealing temperature and duration were optimized. Based on the XRD, the minimum temperature required to form pure phase Ta 3 N 5 was found to be 750 °C. 113 Figure A 3 - 12 . a) The XRD pattern of TaO x N y deposited film after annealing at 750 °C for different durations. The v ertical dashed lines represent B ra g g position s of Ta 3 N 5 with PDF number of 01 - 089 - 5200. Optimization of annealing dur ation was done by soaking temperature at 750 °C for different durations from 10 hours to 30 minutes. Evidently, the formation of pure Ta 3 N 5 is completed after 30 minutes ammonolysis. However, increasing the annealing duration, results in a more well - define d and sharper peaks (more crystalline films). All the following films were annealed for 2 hou rs unless otherwise mentioned. 4 114 Figure A 3 - 13 . XRD of TaO x N y deposited film after ammonolysis at 750 °C for 2 hours as a function of ammonia flow rate. The vertical dotted lines represent the B ra g g position s of Ta 3 N 5 with PDF number of 01 - 089 - 5200. The effect of the ammonia flow rate on crystallinity and phase puri ty of the films were studied by annealing at 750 °C for 2 hours with different flow rate of ammonia. From Figure A 3 - 13 , films annealed at low flow rat e of ammonia, e.g. 50 and 100 mL min - 1 , had poor crystallinity and were comprised of impure phases (possibly TaON). On the other hand, for the flow rate more than 200 mL min - 1 , the observed peaks were sharper (more crystalline) and Ta 3 N 5 was the only detec table phase. This observation is consistent with previous study, where thin films of TaO x and Ta 3 N 5 were prepared on Ta foil. It was found that at lower flow rates of ammonia the formation of TaON is more favorable where at higher flow rates only Ta 3 N 5 films are observed. 37 Here we found that the optimum ammonolysis conditions to form pure and crystalline Ta 3 N 5 , is annealing at 750 °C for 2 hours min - 1 . 115 Figure A 3 - 14 . XRD of 50 nm ALD deposited TaO x N y (green) and TaO x (gray) films after ammonolysis at 750 °C for 30 minutes as a function. The verti cal dotted lines represent the B ra g g position s of Ta 3 N 5 with PDF number of 01 - 089 - 5200. The XRD of the ALD deposited TaO x and TaO x N y are compared in Figure A 3 - 14 . As it can be seen the ALD deposited TaO x were completely nitridized to Ta 3 N 5 after ammonolysis at the same conditions of 750 °C for 30 minutes. 116 Figure A 3 - 15 . XRD of Ta 3 N 5 with different thickness after ammonolysis. The film thicknesses are: 50 nm (blue), 70 nm (orange), 99 nm (maroon), and 122 nm (green). The vertical dashed lines represent Bra g g position s for Ta 3 N 5 with PDF number of 01 - 089 - 5200. Figure A 3 - 16 . a) Cross - section SEM image of Ta 3 N 5 different number of cycles with the scale bar of 100 nm, b) g rowth rate of pure Ta 3 N 5 found by SE and cross - section SEM. 117 Figure A 3 - 17 . a) absorptance, b) % transmittance, and c) % reflectance of thin films of Ta 3 N 5 as a function of thickness, 50 nm (blue), 70 nm (orange), 99 nm (red), and 122 nm (green). 118 Figure A 3 - 18 . a) absorption coefficient of Ta 3 N 5 as a function of wavelength and b) Tauc plots for direct transitions. Figure A 3 - 19 . XRD of the films as a function of ammonolysis duration at 750 °C, 2 hours (red), and 30 minutes. The dashed line are the B ra g g position s of Ta 3 N 5 (green), anatase TiO 2 (blue), rutile TiO 2 (pink). 119 Figure A 3 - 20 . a) XRD of 40 nm of Ta 3 N 5 on TTO with different Ta concentration after annealing in ammonia at 750 °C for 30 min. The vertical dashed lines represent the Bragg position s of Ta 3 N 5 with PDF number of 01 - 086 - 1155. 120 Figure A 3 - 21 . The photocurrent of 70 nm Ta 3 N 5 on TTO at 1.23 V vs. RHE as a function of Ta percentage in TTO. 121 Figure A 3 - 22 . Transmittance and reflectanc e spectrum of electrodes w ith 5 % Ta - doped TiO 2 ~ 40 nm either TaO x N y ALD - deposited (Orange) or TaO x ALD - deposited (gray) after ammonolysis at 750 °C for 30 minutes. The solid black line represents the transmittance of quartz. 122 Figure A 3 - 23 . PEC performance of TaN x (orange) or TaO x (gray) on 5% Ta - doped TiO 2 (100nm)/quartz after annealing in ammonia at 750 °C for 30 min level under 1 sun illumination and 0.5 M K 2 HPO 4 phosphate solution with pH = 13. Note: pr ior to PEC water oxidation measurements, for all the electrodes, CoPi was photo - electrodeposited at 1.06 V vs. RHE for 60s. 123 REFER E NCES 124 REFERENCES (1) Brillet, J.; Comuz, M.; Formal, F. Le; Yum, J. H.; Grätzel, M.; Sivula, K. Examining Architectures of Photoanode - Photovoltaic Tandem Cells for Solar Water Splitting. J. Mater. Res. 2010 , 25 (1), 17 24. (2) Walter, M. G.; Warren, E. L.; McKone, J. R.; Boet tcher, S. W.; Mi, Q.; Santori, E. A.; Lewis, N. S. Solar Water Splitting Cells. Chem. Rev. 2010 , 110 (11), 6446 6473. (3) Fujishima, A.; Honda, K. Electrochemical Photolysis of Water at a Semiconductor Electrode. Nature 1972 , 238 (5358), 37 38. (4) Wrigh ton, M. S.; Ginley, D. S.; Wolczanski, P. T.; Ellis, A. B.; Morse, D. L.; Linz, A. Photoassisted Electrolysis of Water by Irradiation of a Titanium Dioxide Electrode. Proc. Natl. Acad. Sci. 1975 , 72 (4), 1518 1522. (5) Bolts, J. M.; Wrighton, M. S. Correlation of Photocurrent - Voltage Curves with Flat - Band Potential for Stable Photoelectrodes for the Photoelectrolysis of Water. J. Phys. Chem. 1976 , 80 (24), 2641 2645. (6) McKone, J. R.; Lewis, N. S.; Gray, H. B. Will Sol ar - Driven Water - Splitting Devices See the Light of Day? Chem. Mater. 2014 , 26 (1), 407 414. (7) Jang, J. - W.; Du, C.; Ye, Y.; Lin, Y.; Yao, X.; Thorne, J.; Liu, E.; McMahon, G.; Zhu, J.; Javey, A.; et al. Enabling Unassisted Solar Water Splitting by Iron O xide and Silicon. Nat. Commun. 2015 , 6 , 7447. (8) Zandi, O.; Schon, A. R.; Hajibabaei, H.; Hamann, T. W. Enhanced Charge Separation and Collection in High - Performance Electrodeposited Hematite Films. Chem. Mater. 2016 , 28 (3), 765 771. (9) Liu, X.; Wang, F.; Wang, Q. Nanostructure - Based WO 3 Photoanodes for Photoelectrochemical Water Splitting. Phys. Chem. Chem. Phys. 2012 , 14 (22), 7894. (10) Seabold, J. A.; Choi, K. S. Effect of a Cobalt - Based Oxygen Evolution Catalyst on the Stability and the Selectivi ty of Photo - Oxidation Reactions of a WO 3 Photoanode. Chem. Mater. 2011 , 23 (5), 1105 1112. (11) Park, Y.; McDonald, K. J.; Choi, K. S. Progress in Bismuth Vanadate Photoanodes for Use in Solar Water Oxidation. Chem. Soc. Rev. 2013 , 42 (6), 2321 2337. (12) Seabold, J. A.; Choi, K. S. Efficient and Stable Photo - Oxidation of Water by a Bismuth Vanadate Photoanode Coupled with an Iron Oxyhydroxide Oxygen Evolution Catalyst. J. Am. Chem. Soc. 2012 , 134 (4), 2186 2192. (13) Liu, G.; Shi, J.; Zhang, F.; Chen, Z .; Han, J.; Ding, C.; Chen, S.; Wang, Z.; Han, H.; Li, C. A Tantalum Nitride Photoanode Modified with a Hole - Storage Layer for Highly Stable 125 Solar Water Splitting. Angew. Chemie - Int. Ed. 2014 , 53 (28), 7295 7299. (14) Dabirian, A.; Van De Krol, R. High - Temperature Ammonolysis of Thin Film Ta 2 O 5 Photoanodes: Evolution of Structural, Optical, and Photoelectrochemical Properties. Chem. Mater. 2015 , 27 (3), 708 715. (15) Zhang, P.; Zhang, J.; Gong, J. Tantalum - Based Semiconductors for Solar Water Splitting. Chem. Soc. Rev. 2014 , 43 (13), 4395 4422. (16) Zhen, C.; Wang, L.; Liu, G.; Lu, G. Q. (Max); Cheng, H. - M. Template - Free Synthesis of Ta3N5 Nanorod Arrays for Efficient Photoelectrochemical Water Splitting. Chem. Commun. 2013 , 49 (29), 3019. (17) Li, M.; Luo, W.; Cao, D.; Zhao, X.; Li, Z.; Yu, T.; Zou, Z. A Co - Catalyst - Loaded Ta 3 N 5 photoanode with a High Solar Photocurrent for Water Splitting upon Facile Removal of the Surface Layer. Angew. Chemie - Int. Ed. 2013 , 52 (42), 11016 11020. (18) Higashi, M.; D omen, K.; Abe, R. Fabrication of Efficient TaON and Ta 3 N 5 Photoanodes for Water Splitting under Visible Light Irradiation. Energy Environ. Sci. 2011 , 4 (10), 4138. (19) Kitano, M.; Hara, M. Heterogeneous Photocatalytic Cleavage of Water. J. Mater. Chem. 2 010 , 20 (4), 627 641. (20) Li, Y.; Zhang, L.; Torres - Pardo, A.; González - Calbet, J. M.; Ma, Y.; Oleynikov, P.; Terasaki, O.; Asahina, S.; Shima, M.; Cha, D.; et al. Cobalt Phosphate - Modified Barium - Doped Tantalum Nitride Nanorod Photoanode with 1.5% Solar Energy Conversion Efficiency. Nat. Commun. 2013 , 4 , 2566. (21) Ho, C. T.; Low, K. Bin; Klie, R. F.; Maeda, K.; Domen, K.; Meyer, R. J.; Snee, P. T. Synthesis and Characterization of Semiconductor Tantalum Nitride Nanoparticles. J. Phys. Chem. C 2011 , 115 (3), 647 652. (22) Chen, Z.; Jaramillo, T. F.; Deutsch, T. G.; Kleiman - Shwarsctein, A.; Forman, A. J.; Gaillard, N.; Garland, R.; Takanabe, K.; Heske, C.; Sunkara, M.; et al. Accelerating Materials Development for Photoelectrochemical Hydrogen Production : Standards for Methods, Definitions, and Reporting Protocols. J. Mater. Res. 2010 , 25 (1), 3 16. (23) Hu, S.; Xiang, C.; Haussener, S.; Berger, A. D.; Lewis, N. S. An Analysis of the Optimal Band Gaps of Light Absorbers in Integrated Tandem Photoelectroc hemical Water - Splitting Systems. Energy Environ. Sci. 2013 , 6 (10), 2984 2993. (24) Liu, G.; Ye, S.; Yan, P.; Xiong, F.; Fu, P.; Wang, Z.; Chen, Z.; Shi, J.; Li, C. Enabling an Integrated Tantalum Nitride Photoanode to Approach the Theoretical Photocurrent Limit for Solar Water Splitting. Energy Environ. Sci. 2016 , 9 (4), 1327 1334. (25) Hamann, T. W.; Martinson, A. B. F.; Elam, J. W.; Pellin, M. J.; Hupp, J. T. Atomic Layer Deposition of TiO 2 on Aerogel Templates: New Photoanodes for Dye - Sensitized Solar Cells. J. Phys. Chem. C 2008 , 112 (27), 10303 10307. 126 (26) Friel, J. J. X - Ray And Im age Analysis in Electron Microscopy , 2nd ed.; Princeton Gamma - Tech: Princeton, 2005. (27) Mazzolini, P.; Gondoni, P.; Russo, V.; Chrastina, D.; Casari, C. S.; Bassi, A. L. Tuning of Electrical and Optical Properties of Highly Conducting and Transparent Ta - Doped TiO 2 Polycrystalline Films. J. Phys. Chem. C 2015 , 119 (13), 6988 6997. (28) Hitosugi, T.; Furubayashi, Y.; Ueda, A.; Itabashi, K.; Inaba, K.; Hirose, Y.; Kinoda, G.; Yamamoto, Y.; Shimada, T.; Hasegawa, T. Ta - Doped Anatase TiO 2 Epitaxial Film as T ransparent Conducting Oxide. Jpn. J. Appl. Phys. 2005 , 44 (No. 34), L1063 L1065. (29) Osorio - Guillén, J.; Lany, S.; Zunger, A. Atomic Control of Conductivity versus Ferromagnetism in Wide - Gap Oxides via Selective Doping: V, Nb, Ta in Anatase TiO 2 . Phys. R ev. Lett. 2008 , 100 (3), 036601. (30) Grins, J. Ti x Ta (O,N) y Phases Formed by Ammonolysis of Ti - Ta Gels: Preparation of an Anatase - Type Solid Solution Phase Ti x Ta O 1+x N J. Eur. Ceram. Soc. 1997 , 17 (15 16), 1819 1824. (31) Barman, A. R.; Motapothula, M.; Annadi, A.; Gopinadhan, K.; Zhao, Y. L.; Yong, Z.; Santoso, I.; Ariando; Breese, M.; Rusydi, A.; et al. Multifunctional Ti 1 - X Ta x O 2 : Ta Doping or Alloying? Appl. Phys. Lett. 2011 , 98 (7), 072111. (32) Hoang, S.; Guo, S.; Mul lins, C. B. Coincorporation of N and Ta into TiO 2 Nanowires for Visible Light Driven Photoelectrochemical Water Oxidation. J. Phys. Chem. C 2012 , 116 (44), 23283 23290. (33) Hoang, S.; Guo, S.; Hahn, N. T.; Bard, A. J.; Mullins, C. B. Visible Light Drive n Photoelectrochemical Water Ox idation on Nitrogen - Modified TiO 2 Nanowires. Nano Lett. 2012 , 12 (1), 26 32. (34) Maeng, W. J.; Park, S. - J.; Kim, H. Atomic Layer Deposition of Ta - Based Thin Films: Reactions of Alkylamide Precursor with Various Reactants. J . Vac. Sci. Technol. B Microelectron. Nanom. Struct. 2006 , 24 (5), 2276. (35) Fang, Z.; Aspinall, H. C.; Odedra, R.; Potter, R. J. Atomic Layer Deposition of TaN and Ta 3 N 5 using Pentakis(Dimethylamino)Tantalum and Either Ammonia or Monomethylhydrazine. J. Cryst. Growth 2011 , 331 (1), 33 39. (36) Langereis, E.; Knoops, H. C. M.; MacKus, A. J. M.; Roozeboom, F.; Van De Sanden, M. C. M.; Kessels, W. M. M. Synthesis and in Situ Characterization of Low - Resistivity Ta Nx Films by Remote Plasma Atomic Layer Depo sition. J. Appl. Phys. 2007 , 102 (8), 083517. (37) Hara, M.; Chiba, E.; Ishikawa, A.; Takata, T.; Kondo, J. N.; Domen, K. Ta 3 N 5 and TaON Thin Films on Ta Foil: Surface Composition and Stability. J. Phys. Chem. B 2003 , 107 (48), 13441 13445. (38) Ritala, M.; Kalsi, P.; Riihelä, D.; Kukli, K.; Leskelä, M.; Jokinen, J. Controlled Growth of 127 TaN, Ta 3 N 5 , and TaO x N y Thin Films by Atomic Layer Deposition. Chem. Mater. 1999 , 11 (7), 1712 1718. (39) Hou, J.; Wang, Z.; Yang, C.; Cheng, H.; Jiao, S.; Zhu, H. Cobal t - Bilayer Catalyst Decorated Ta3N5nanorod Arrays as Integrated Electrodes for Photoelectrochemical Water Oxidation. Energy Environ. Sci. 2013 , 6 (11), 3322 3330. (40) Li, Y.; Takata, T.; Cha, D.; Takanabe, K.; Minegishi, T.; Kubota, J.; Domen, K. Vertical ly Aligned Ta 3 N 5 Nanorod Arrays for Solar - Driven Photoelectrochemical Water Splitting. Adv. Mater. 2013 , 25 (1), 125 131. (41) Rugge, A.; Park, J. S.; Gordon, R. G.; Tolbert, S. H. Tantalum(V) Nitride Inverse Opals as Photonic Structures for Visible Wavelengths. J. Phys. Chem. B 2005 , 109 (9), 3764 3771. (42) Leslie, J. D.; Knorr, K. Ellipsometric Study of the Plasma Oxidation of Tantalum. J. Electrochem. Soc. 1974 , 121 (2), 263 267. (43) Klahr, B. M.; Martinson, A. B. F.; Hamann, T. W. Photoelectrochemical Investigation of Ultrathin Film Iron Oxide Solar Cells Prepared by Atomic Layer Deposition. Langmuir 2011 , 27 (1), 461 468. (44) Morbec, J. M.; Narkeviciute, I.; Jaramillo, T. F.; Galli, G. Optoelectronic Properties of Ta 3 N 5 : A Joint Theoretical and Experimental Study. Phys. Rev. B 2014 , 90 (15), 155204. (45) Yoon, J. - W.; Sasaki, T.; Koshizaki, N.; Noguchi, Y.; Miyayama, M. C omparison of Photoelectrode Properties Between TiO 2 Thin Films Doped with Tantalum and Dispersed with Nanosize Gold. Electrochem. Solid - State Lett. 2004 , 7 (7), A172. (46) Chen, T. T.; Liu, H. P.; Wei, Y. J.; Chang, I. C.; Yang, M. H.; Lin, Y. S.; Chan, K . L.; Chiu, H. T.; Lee, C. Y. Porous Titanium Oxynitride Sheets as Electrochemical Electrodes for Energy Storage. Nanoscale 2014 , 6 (10), 5106 5109. (47) Yue - Ping, H.; Yan, H. Physical and Chemical Properties of TiO x N y Prepared by Low - Temperature Oxidation of Ultrathin Metal Nitride Directly Deposited on SiO 2 . Chinise Phys. Lett. 2009 , 26 (3), 3. (48) Shannon, R. D. Revised Effective Ionic Radii and Systematic Studies of Interatomic Distances in Halides and Chalcogenides. Acta Crystallogr. Sect. A 1976 , 32 (5), 751 767. (49) Toby, B. H. EXPGUI, a graphical user interface for GSAS http://www.ncnr.nist.gov/xtal/software/EXPGUI_reprint.pdf (accessed Aug 3, 2015). (50) Larson, A. C.; Von Dreele, R. B. General Structure Analysis System http://www.ccp1 4.ac.uk/ccp/ccp14/ftp - mirror/gsas/public/gsas/manual/GSASManual.pdf (accessed Aug 3, 2015). 128 Direct Deposition of Ta 3 N 5 via High - Temperature ALD 129 4.1. Abstract In this project the direct deposition of crystalline Ta 3 N 5 on FTO substrate via a custom - built ALD system is studied. A combination of TaCl 5 (Ta - precursor) and with ammonia (N - source) were sequentially pulsed into the reactor with the substrate heated to 550 °C to deposit compact and thin films of Ta 3 N 5 with controllable thicknesses on FTO substrates. Importantly, it is shown that the FTO is chemically and structurally stable under the reducing conditions of ammonia at 550 °C. This is a significant progress as it allows to study the fundamental characteristics of Ta 3 N 5 independently o f the conductivity of the substrate. These electrodes produced a photocurrent onset potential of ~ 0.3 V vs. RHE and a maximum photocurrent of ~ 2.4 mA cm - 2 . In addition, the PEC characterizations as a function of film thickness and illumination direction (through the electrolyte and substrate) reveal that the PEC performance of Ta 3 N 5 is controlled by the limited hole diffusion length (<100 nm). 4.2. Introduction Tantalum nitride (Ta 3 N 5 ) is one of the promising semiconducting materials for photoelectrochemical ( PEC) water oxidations and has attracted a lot of attention in the past few year. State of the art Ta 3 N 5 photoanodes for PEC water oxidation were introduced by Li et al. 1 with a photocurrent density approaching the theoretical limit of ~ 12.5 mA cm - 2 . The synthesis of Ta 3 N 5 electrodes commonly start with the oxidation of Ta - substrate via electrochemical anodization or simply by heating in air to oxidize the top Ta layer to nanostructured or planar tantalum oxide films, respectively . 1 5 Subsequently, the oxidized sample is nitridized in a flow of ammonia at elevated temperatures (850 - 1000 °C) for a prolonged period of time (2 - 15 hou rs). The Ta - substrate serves as the source of Ta, the conductive layer to collect the majority charge 130 carriers necessary to fabricate electrode, and the substrate. Although this method is simple and resulted in the best performing T a 3 N 5 photoanodes , 1,2 from a synthetic point of view however, multiple drawbacks make this method unsuitable to realize efficient photoelectrodes; specifically, this method: (1) is highly energy intensive, (2) produces a sizable quantity of chem ical waste, (3) provides highly reducing conditions which limit Ta 3 N 5 to be only compatible with Ta - substrate or Nobel metals like Pt, 4,5 (4) precludes its application in the tandem cell configuration as Ta or other metal substrates are not transparent to the sub - bandgap photons, (5) res ults in the formation of electronically resistive phases at the Ta 3 N 5 |Ta junction which further limits the electron collection efficiency. In addition, Ta as a substrate becomes very brittle after heating in ammonia which makes the post - annealing electrode processing very challenging. For example, Jaramillo et al. 3 studied the structure and phase transformation of tantalum oxide to tantalum nitride films on various substrates including Ta foil and fused silica. They showed that the films on fused silica were single phase Ta 3 N 5 , but that the films prepared on Ta subs trates were comprised of impurity phases and upon rising the ammonolysis temperature, the formation of Ta 2 N and Ta 5 N 6 phases were favored. 3 These observations, along with the depth profiling GIXS analy sis, confirmed that N - poor phases are exclusively formed at the Ta 3 N 5 |Ta junction. 3 6 Domen et al. 2 showed that the ammonolysis of anodized Ta - substrate results in the formation of N - poor phases, e.g. Ta 5 N 6 , which limits the electron collection efficiency at the back contact. Interestingly, doping with barium suppressed the formation of interfacial Ta 5 N 6 resistive layer, which substantially improved the PEC water oxidation performance of the electrode. Importantly, these examples highlights the complexity of the synthesis of pure Ta 3 N 5 which requires a precise control over ammonolysis temperature, atmosphere, and reaction duration. 7,8 For example, Henderson and Hector 7 studied the structural and compositional evolution of Ta 3 N 5 from amorphous tantalum oxide as a function 131 of ammonolysis temperature (from 680 to 900 °C) and duration (from 8h to 120h). Their r esults are quite remarkable, showing that regardless of ammonolysis conditions (temperature or duration), all samples contained oxygen impurities - even after annealing at 900 °C for 120 h. In addition, Terao 8 showed that at elevate d temperatures (>1000 °C) Ta 3 N 5 has a dynamic structure and composition where nitrogen is successively pulled out of the structure upon increasing temperature in vacuum this is accompanied by the concomitant reduction of Ta (V) which subsequently results in the formation of various phases of tantalum nitride. Therefore, it is highly beneficial to directly synthesize crystalline Ta 3 N 5 films at lower temperatures from an oxygen - free precursors and atmosphere by eliminating the a mmonolysis step. This approach is additionally beneficial as it allows to integrate Ta 3 N 5 with different materials, e.g. Ta 3 N 5 on various TCO, and to study its structure - function relationship. The vacuum deposition techniques such as atomic layer depositio n (ALD) or chemical vapor deposition (CVD) provide an oxygen - free atmosphere suitable to synthesize none - oxide materials with a controllable thickness and composition. A series of previous studies for the deposition of tantalum nitride using ALD or CVD are listed i n Table 4 - 1 . Briefly, these studies indicates that (1) tantalum halides, e.g. TaCl 5 and TaBr 5 , coupled with ammonia are commonly used as the tantalum and nitrogen sources to directly deposit crystalline Ta 3 N 5 thin films. (2) The near atmospheric pressure CVD synthesis of Ta 3 N 5 from these sources low - pressure CVD or ALD deposition can be done at considerably lower temperatures in the range of 400 500 °C. Inspired by these reports, we studied the direct deposition of crystalline Ta 3 N 5 on FTO substrates via a custom - built and fully automated ALD system ca pable of operating at elevated temperatures with a maximum temperature of 640 °C. Furthermore, we showed that this technique provides a suitable method to integrate Ta 3 N 5 on a rather chemically u nstable substrate like FTO in reducing 132 atmospheres of ammonia . Subsequently, the photoelectrochemical performance of Ta 3 N 5 |FTO electrodes with various thicknesses as a function of illumination direction were explored, showing that the PEC performance of Ta 3 N 5 is controlled by the limited diffusion length of holes (1 0s of nm). 133 Table 4 - 1 . The summary of the experimental conditions and techniques utilized for deposition of thin film of tantalum nitride. Ta - Source N - Source Method Temp. (°C) Substrate Characterization Crystalline or Amorphous Growth Rate Ref. TaCl 5 N 2 (plasma) PECVD 715 Si(100) TEM / SAED Microcrystalline NA [ 9 ] TaCl 5 NH 3 CVD (400 Torr) 900 - 1300 Fused s ilica E lectron D iffraction, Conductivity Crystalline (Ta 3 N 5 ) 30 nm/min [ 10 ] PDMAT NH 3 CVD 200 Silicon, Glass XPS, XRD, TEM, RBS Amorphous 200 nm/min [ 11 ] PDMAT NH 3 CVD 200 - 400 Various RBS: ( N / Ta ) = 1.7 Amorphous NA [ 12 ] TaBr 5 NH 3 CVD 0.96 Torr 500 Si, SiO 2 RBS, AFM, XRD, AES, XPS Crystalline ( Ta 3 N 5 ) 3 - 7.5 nm/min [ 13 ] Ta(NEt 2 ) 5 NA LPCVD >500 Si(100) XRD NA NA [ 14 ] TaBr 5 NH 3 ALD 500 Soda lime NA NA NA [ 15 ] TaF 5 H 2 /NH 3 ALD 200 NA A morphous NA 0.05 nm /cycle [ 16 ] TaCl 5 NH 3 ALD 400 - 500 Soda lime glass, ITO XRD, C onductivity Crystalline ( Ta 3 N 5 ) 0.025 nm/cycle [ 17 ] TaCl 5 NH 3 ALD 500 Soda g lass XRD Crystalline (Ta 3 N 5 ) ~ 0.02 nm/cycle [ 18 ] PDMAT MMH ALD 175 - 280 SiO 2 , Al, FTO XRD, Raman, XPS, EDS Amorphous 0.07 - 0.1 nm/cycle [ 19 ] PDMAT MMH/NH 3 ALD 200 - 375 Si(100) RBS, Conductivity NA 0.02 - 0.1 nm/cycle [ 20 ] 1) NA: Not Assigned; 2) CVD: Chemical vapor deposition; PECVD: Plasma enhanced chemical vapor deposition; LPCVD: Low pressure chemical vapor Deposition; ALD: Atomic Layer Deposition; PDMAT: Pentakis(dimethylamino)tantalum(V); TEM: Transmission electron microscopy; SAED: Selected area electron diffraction; RBS: Rutherford backscattering spectrometry; MMH : Monomethyl Hydrazine. 3) We note that in these studies a generic term of tantalum nitride as oppose to the exact composition of material is used . In addition, only a few of them f ully characterized the material and some of just gust the composition/crystalline phase based on the color of the films. 134 4.3. Experimental 4.3.1. Instrumentation 4.3.1.1 At omic Layer Deposition Figure 4 1 shows the building blocks of a typical ALD instrument. As shown, it is comprised of three main blocks of: software, electronic box, and the ALD reactor. In following, these components and their functionality will be briefly discussed. Figure 4 1 . The building blocks of an atomic layer deposition system. 4.3.1.2 Software The ALD is comprised of several different components such as heaters, valves, gas flow controller , and vacuum stage which all require a continu ous monitoring and controlling. In addition, the ALD procedure is a well - timed process that is comprised of sequential pulse of precursors into the reactor in a timely fashion (see Figure 1 13 ) . In some cases it may require a complex sequence of pulses of precursors. Moreover, depending on the deposition rate and the desire d thickness, the ALD deposition can be a long process. These requirements, thus, emphasize the importance of a sophisticated program that not only provides a remote access to each component of the instrument but also it produces reproducible pulses and steps in each ALD cycle. For this purpose, Labview progra m ming software was used. Th e screenshot of the program in action is shown in Figure A 4 - 1 . This program allows to control the sequence and the duration of the precursor pulses (with tens of 135 ms resolution) and flow rate of the carrier gas (with a resolution of ± 5 SCCM 1 ). This program provides a remote access and controls over 5 ALD valves, 1 va cuum valve, 11 heaters, 1 mass flow controller, and 1 pressure sensor. 4.3.1.3 Electronic Box The electronic box is the heart of the ALD instrument. As shown in Figure 4 1 , the electronic box bridges the physical part of the ALD reactor to the computer and the program that controls it. Basically, it reads the physical status of each components of the ALD in ter ms of voltage (analog) and convert them into digital signals which can then be processed by the software. Subsequently, it converts the digital commands from the computer into the analog output signals (voltage) to set the instrument at the specified condi tion. The electronic box is comprised of various components ( Figure A 4 - 2 ) , each with specific functionality. The wiring diagram of these components as well as the function of each component are summarized following Figure A 4 - 3 . 4.3.1.4 Carrier Gas Purification One of the advantage of ALD is that it can be utilized to synthesize non - oxide and air - sensitive materials. In addition, most of the precursors used in ALD are organometallic complexes with extreme sensitivity to water and/or oxygen. It is thus important t o remove any trace of oxygen impurity from the nitrogen as the carrier gas. Here, we used three scrubbers in series to remove any oxygen in forms of water and molecular oxygen. The first scrubber was a glass tube (1/4" × 8") which was filled with quartz wo ol and P 2 O 5 (which vigorously react with water) and the second 1 Standard cubic centimeter per minutes 136 scrubber was made out Pyrex (1.5" × 6") was filled with activated molecular sieve (3 Å pellets 1.6 mm, Sigma Aldrich). These two scrubbers were used to remove water from. The last scrubber (1.5 " × 6") was made out Pyrex and was filled with metallic copper powder mixed with quartz wool ( Figure A 4 - 4 ). This scrubber was heated to 300 400 °C with a band heater (from outside). At elevated temperature, the metallic copper reacts with the residual oxygen and forms copper oxide which consequently removes oxygen from the flowing gas. As shown in Figure A 4 - 4 , over time the color of metallic copper was turned to black which indicates that the residual oxygen in the nitrogen line is successfully removed by reacting with copper . 4.3.1.5 Inlet of the Reactor: ALD Valves and Configuration The schematic representation of the inlet of the ALD reactor is shown in Figure 4 2 . All the connection are VCR - type fitting which offers a leak - tight metal to metal seal. To seal the VCR assembly, the 316L s tainle ss steel VCR face seal fitting ( Silver Plated Gasket , SS - 4 - VCR - 2 ) is used. The tubes, fittings, connectors, body of the valves, and the body of the mass flow controller (MFC) are all constructed from stainless steel 316. The tubes and fitting are ¼" in diameter except the inlet and outlet of the reactor that are ½" and 1", respectively. As depicted in Figure 4 2 , two MFCs (Omega, FMA5510A) are used for nitrogen (carrier gas) and ammonia (co - reactant). The MFC for nitrogen has a flow range of 10 - 200 SCCM. The inlet of the MFC for ammonia is directly c onnected to the ammonia cylinder with a pressure set to 40 PSI 1 . A ¼" stainless tube was fashioned in the shape of a spring and it is installed between the MFC and ALD valve (this part 1 Pounds per square inch 137 act as a reservoir for ammonia to provide a constant dose of ammonia th roughout the ALD process). Experimentally, it was observed that even for a short pulse (0.1 s) of ammonia and long Figure 4 2 . The schematic representation of the Inlet of the ALD reactor. - purge time (70 s), ammonia diffuses into the carrier gas and find its way into the Ta - precursor cylinder and react with tantalum inside the tantalum cylinder. In order to circumvent this issue, the ammonia line was isolated from the carrier gas and tantal um precursor line via a 2 - way valve (labeled as the N 2 valve in Figure 4 2 ). The tantalum precursor was kept in a 50 mL stainless steel 316 cylinder a nd the cylinder was heated with a flexible heating jacket equipped with a J - type 138 thermocouple. The tube line from the Ta - precursor to the inlet of the reactor (except for the ammonia line) is wrapped with a flexible band heater and it is thermally insulate d from the surroundings. The temperature of this line is set to 170 °C which is monitored by a K - type thermocouple installed on the N 2 valve. For pulsing the reactant, 2 - way valves (Swagelok, stainless steel bellows sealed v alve , SS - 4BK - V51 - 1C) equipped wi th solenoid valves (24 V DC, MAC Valve, INC., 34C - ABA - GDFC - 1KT) were utilized. 4.3.1.6 Outlet of the Reactor: Vacuum Line and Valves The Alcatel 2005I (purchased from IdealVac) charged with perfluoropolyether (PFPE, Fomblin 25/6) is used as the vacuum pump . This p ump provides a base pressure of ~ 2 × 10 - 3 T orr . To monitor the pressure of the reactor a vacuum gauge ( Edwards Linear Convection Vacuum Gauge, APGX - H - NW16) is installed at the outlet of the reactor. A poppet valve (Genesis Stainless Steel, Model # 051207 - 20) equipped with a solenoid valve (120 V AC , MAC Valve, INC., 34B - AAA - GAAA - 1KC) is used to pump/vent the ALD reactor. Since ammonia is highly reactive and corrosive which cause shortening the lifetime of the vacuum pump, a Metal Wool Foreline Trap made of stainless steel (Kurt J. Lesker, TAR 4 CS 100 QF) was installed just before the inlet of the vacuum pump. This trap was charged with copper wool (react with ammonia) to remove ammonia from the effluent ga s stream. 4.3.1.7 The ALD Reactor The ALD reactor is where the chemical reaction takes place. In terms of chemistry, this is the most important part of the ALD system which defines the purity and the quality of the final product. In addition to the reactor configu ration, the material used in its construction is also an important aspects of the reactor. The latter becomes highly important when the ALD deposition conditions 139 are reactive. In this study, we designed, build, and examined three different ALD reactors. Th e schematic cross - section pictures of these reactors are shown in Figure 4 3 and each on e of them are briefly introduced below. Figure 4 3 . The schematic representation of different generation of ALD reactors: a) 1 st , b) 2 nd , and c) 3 rd generation. The body of all generations is constructed from stainless steel 316. One of the advantage of the commercially available ALD reactor is that loading and unloading of the sample is fairly easy and quick. It is an important aspect of ALD as it minimizes the time required to transfer the substrate which further minimized the po ssibility of surface contamination. On the other hand, one of the limitation of these types of reactors is the necessity of the specially designed and robust O - rings for high - temperature applications. These O - rings have a limited 140 temperature range and are not stable at temperatures beyond 300 °C. To get around this problem for high - temperature ALD reactor, the heated area must be separated from the loading section. The 1 st Generation ALD reactor is shown in Figure 4 3 . a . This generation is a vertically aligned cylindrical reactor with a branch pointing away at the midway. The inlet is located at the top and the outlet of the reactor connected to vacuum p ump is located at the bottom. The inlet of the reactor is equipped with a custom - designed shower head to e nsure a uniform flow of carrier gas and precursors over the substrate. This is a hot - wall reactor and is constructed from stainless steel 316. The mai n body of the reactor is heated with flexible band heaters and its temperature is probed by a K - type thermocouple installed on the outer wall. To maintain the temperature, the heated area is thermally isolated with fiberglass insulation sheets. The rectang ular branch is designed to separate the heated area from the loading area (this is the place where the O - ring is installed). This area is not thermally insulated and is cooled in air. We note that when the temperature of the reactor is s e t at 650 °C, the m easured temperature in vicinity of the O - ring reaches ~ 200 °C (this measurement was done after ~ 5 hours of temperature stabilization ). The pressure of the reactor remained constant during continuous heating for 5 hours which further indicates that the O - ring is stable under this extreme operating condition . For this reactor we used three different sample holders constructed from stainless steel 316, Al (with a maximum temperature of 550 °C), and quartz. The first two holders were fabricated with a screen (~ 1 cm - 2 holes) welded to the bottom of a ring shape holder where the FTO substrates (1.0 cm × 1.0 cm) were placed on the screen. In this configuration the flow of gas is perpendicular to the substrate. For the quartz sample holder, two parallel racks wit h a distance of 2.5 cm were welded to the bottom of a quartz ring and the FTO substrates (cut into small slides with the dimensions of 1.0 - 5.0 cm × 1.2 cm (length × height)) were mounted on the racks (with 0.3 cm space between the substrates). For this s ample holder, the 141 FTO substrates are oriented parallel to the gas flow (see Figure 4 3 .a ) . Figure 4 4 . a shows the EDS spectrum of the film deposited with this reactor. As shown in addition to Ta, FTO (SnO 2 ), glass (SiO 2 ) signals, and a well - resolved peak attributed to iron is clearly detectable. The photograph of the film shows that the as - deposited film has an orange/green color which upon annealing in air at 750 °C it tur ns into a bright orange and the Raman spectrum of the as - deposited - Fe 2 O 3 ) . 4 - 1 These observations are better understood by considering the chemical reaction between tantalum chloride and ammonia. As shown in eq. 4 - 1 , the byproduct of this reacti on is HCl. As the result, at elevated deposition temperature iron is etched from the reactor wall and is deposited on the substrate. In order to circumvent this issue the 2 nd generation reactor was designed. 142 Figure 4 4 . The EDS spectra showing the composition of the films deposited on FTO substrate at 500 °C with different generations of the ALD reactor; a) 1 st , b) 2 nd , and c) 3 rd generation. The inset 143 in the figure (a) shows the photograph of the as - deposited film along with the film annealed in air at 750 °C for 1 minutes; also shown is the comparison of the Raman spectra of the as - deposited - Fe 2 O 3 ) film. 21 The inset in the Figure (c) shows the photograph of the film deposited from 3 rd generation reactor on the custom - built substrate heater. The second generation ALD reactor is a modifi ed version of the 1 st generation where its interior composition is changed to quartz which has a neutral composition ( Figure 4 3 . b). The main body of the reactor is constructed from stainless steel 316, but to change the interior composition a quartz tube is installed inside. Unlike the 1 st generation, this reactor is a horizontal tube reactor. Similar to the 1 st generation, this reactor is heated with high - temperature flexible band heaters from outside and the heated area is separated from the area where the O - ring is installed (at the outlet of the reactor). The outlet (vacuum) of the reactor is designed so it can be easily and quickly attached/removed and thus it additionally serves as loading section. The sample holder is constructed from pyrex (with the softening point of ~ 820 °C) and it is bend to maximize the effective contact of the flowing gas with the FTO substrate. After several trials and inv estigating various deposition conditions, no sign of color change was detected. However, it was reproducibly observed that a white powder (more like dust with no adhesion to the substrate which can be easily washed off with water) was deposited on the back side of the FTO substrate. We speculate that this powder is ammonium chloride which is the side product of the ALD. The EDS spectrum of the substrate after deposition with this reactor is shown in Figure 4 4 . b. This spectrum is similar to that of the bare FTO substrate which further indicates that nothing is deposited. The 3 rd generation of ALD reactor has a similar configuration as the first gen eration and it is constructed from stainless steel 316 ( Figure 4 3 .c ). Unlike the first two generations, however, this 144 generation is a hot - substrate a nd cold - wall reactor where only substrate is heated and the reactor wall is left uninsulated letting the heat to skip the system. As a result, the reactor wall is cooler than the substrate. Accordingly, the sample holder was modified to accommodate a high - temperature substrate heater (Chromalox, Chrome Steel Sheath) to specifically heat the substrate (see inset of Figure 4 4 . c). In this configuration, thereby, the gas flow is perpendicular to the substrate. The EDS spectrum along with the photograph of the film deposited on FTO substrate from this reactor is shown Figure 4 4 c . As depicted, all the signals are assigned to the FTO substrate and tantalum nitride deposited film . Importantly, no signature of iron or any other components of the stainless steel is not detected. This reactor configuration wa s then used to study the deposition of crystalline Ta 3 N 5 . 4.3.2. Deposition of Ta 3 N 5 Film The deposition of Ta 3 N 5 was carried out in a custom built ALD system described in previous section . The films were deposited on FTO (F - doped tin oxide) coated glass substrate s (Tech 15, Hartford Glass Co) with the softening point of ~ 600 °C. The FTO - glass substrates were cut into hexagonal pieces with the diameter of ~ 5.2 cm. Prior to deposition, substrates were cleaned twice by sequential sonication in soap ( Fisher ), water, and isopropyl alcohol for 15 minutes each (with the total time of 90 minutes). Prior to loading into the instrument, the FTO substrate was drie d in a gentle flow of nitrogen. The Ta 3 N 5 films were deposited via sequent ial pulses of TaCl 5 (Alfa Aesar, 99.9%) as the source of tantalum and anhydrous ammonia (Airgas) as the co - reactant. The purified N 2 (99.9%) with the flow rate of 20 SCCM was used as the carrier gas. The TaCl 5 cylinder was kept at 120 °C and it was pulsed for 2 s which was followed by 15 s waiting/purging under the flow of nitrogen. Prior to pulsing ammonia the nitrogen line is 145 closed. This step is required to avoid contamination of carrier gas or even TaCl 5 precursor. After 5 s wait time, ammonia was pulse d for 0.1 s. This step is followed by a 10 s waiting time while the nitrogen line is closed and an additional 15 s purging time while the nitrogen line is in open position. The schematic representation of this recipe is shown in Figure A 4 - 5 . A range of substrate temperature from 450 to 550 °C were investigated. The temperature of the substrate heater was externally calibrated with a secondary thermocoup le (digital meter model 6802II equipped with a K - type thermocouple) and the calibration points can be found in Table A 4 - 1 . The measured temperature by both thermocouple are within ± 10 °C and throughout this chapter we referred to the set point temperature. To study the stability of FTO under the deposition cond itions of tantalum nitride we annealed the FTO|glass substrate at 550 °C under a continuous flow of ammonia with a flow rate of ~ 100 SCCM for 5 hours in a tube furnace equipped with quartz tube (2.5" ×18"). We note that this annealing condition is far mor e extreme than the deposition conditions (as ammonia is contentiously flowing and is not diluted with N 2 ) thus it represent the upper limits for the stability FTO. 4.3.3. Film Characterization The morphology and thickness of the deposited films were examined usin g scanning electron microscopy (SEM, Carl Zeiss Auriga, Dual Column FIB - SEM). For the cross - section analysis the films were coated with 5 - 10 nm tungsten conductive layer via sputtering (Denton Vacuum Desk II sputter coater). Image J was utilized to analyze the thickness of the films from cross - section SEM. The composition of the deposited films were evaluated via energy dispersive spectroscopy ( C ARL ZEISS EVO LS 25 ) equipped with Ametek - EDAX Apollo X detector and TEAD EDS software. For simplicity of compari son, the spectrum for each sample was normalized to its total 146 area. The Raman spectrum were collected via inVia Raman Microscope (Renishaw) equipped with 45 W cobalt DPSS laser (532 nm line) and a 100× microscope. Prior to Raman measurements, the instrumen t was quickly calibrated against internal silicon standard. For simplicity of comparison the Raman spectra were further normalized to the total area. The crystallinity and phase purity of the films were evaluated via X - ray diffraction (XRD) using Bruker D8 advanced diffractometer using Cu radiation at 1. 5118 Å. It is important to note that during the XRD measurement the sample was spanned to eliminate the artifact arising from the 2D structure of film by contributing all the plane and crystallites. The opti cal properties of the Ta 3 N 5 |FTO films were determined utilizing PerkinElmer Lambda 35 UV - vis spectrometer equipped with Labsphere integrating sphere. The % absorptance (% A) was corrected for the substrate using a model previously developed in our lab. 22 The Beer - Lambert law in combination with the absorption coefficient of Ta 3 N 5 at 450 nm (188,000 cm - 1 ) 23 was utilized to calculate the film thickness. Atomic force microscopy ( AFM, Cyphe r Atomic Force Microscope /Scanning Probe Microscope (AFM/SPM) ) in the tapping mode equipped with silicon AFM probe (Budget Sensor) with 300 kHz resonant frequency and 40 N/m force constant was utilized to measure the film roughness and topology. 4.3.4. Photoelectrochemical Measurements T he electrochemical measurements were carried out with a A u tolab potentiostat (PGSTAT128N ) equipped with Nova electrochemical software. The photoelectrochemical performance of the electrodes were carried out in contact with an aqueous solution containing 0. 1 M K 4 [Fe(CN) 6 ] as the hole scavenger with the pH of 6.8. It is worth mentioning that different concentrations of the K 4 [Fe(CN) 6 ] (50 mM to ~ 400 mM) were tested and it was noted that at concentrations beyond 147 100 mM, the PEC response is constant. Throughou t this chapter the current density ( J ) vs. potential ( E ) J - E curves were measured at the scan rate of 20 mV/s. Additionally, the photo - response of the electrodes were measured as a function of illumination direction, the back illumination refering to illum ination through the substrate and the front illumination describing the illumination through the solution. A 4500 W Xe lamp (Horiba Jobin Yvon) and AM 1.5 solar filter was used to simulate sunlight at the power of 100 mW cm - 2 (1 sun). The light intensity a t the position of the electrode was calibrated with a certified reference cell (Oriel Reference Solar Cell & Meter). A homemade Ag/AgCl which was calibrated against commercial calomel electrode (Koslow Scientific) and a platinum mesh electrodes were utiliz ed as the reference and counter electrode, respectively. The electrochemical impedance measurements were carried out at 10 mV amplitude perturbation in the 0.01 to 10,000 Hz frequency range. Under dark conditions, only one semicircle was observed in the Nyquist plot. The data were fit to the Randle circuit model using Zview software. The incident photon - to - current - efficiency (IPCE) measurements were carried out in the three - electrode setup under monochromatic light in the 400 to 700 nm range with the 10 n m intervals. The white light was monochromatized using a Horiba Jobin Yvon MicroHR monochromotor. The entrance and the exit of the monochrom a tor whereas set at 0.75 mm which correspond to 8 nm linewidth. The light intensity at the electrode position under monochromatic light was measured with Nova II Ophir. The IPCE and a bsorbed photon - to - current efficiency (APCE) values were calculated according to equations 4 - 2 and 4 - 3 , respectively. 4 - 2 148 4 - 3 4.4. Results and Discussion In this study a combination of TaCl 5 (Ta - source) and NH 3 (N - source) were utilized to deposit crystalline films of Ta 3 N 5 on FTO using a home - built ALD. A range of deposition tem peratures from 450 to 550 °C were investigated. On the basis of the XRD measurements, it was found that 550 °C is the minimum temperature required to deposit crystalline Ta 3 N 5 . As discussed in previous chapter, at temperature beyond 550 °C the FTO substrat e is not stable under ammonia/ vacuum conditions (a highly reducing atmosphere) as it becomes flaky and loses its inherent property as a TCO. We note that this temperature is higher by 50 - 150 °C in comparison to the previous studies on ALD deposition of Ta 3 N 5 (summarized i n Table 4 - 1 ) . 17,18 The higher tempera ture required to deposit crystalline films found here can be ascribed to the difference in the configuration of the ALD system, and substrate used. 149 Figure 4 5 . The EDS spectra of the bare FTO (black) and t he deposited films as a function of number of cycles; 500 (red), 300 (green), and 150 cycles (blue). The highlighted graph on the right - hand side represents the peak located at 1.7 kV in a reverse order showing a progressive growth of Ta peak as the number of cycles increases. The inset represent the configuration of the film and the origin of signals. Energy dispersive spectroscopy (EDS) was utilized to assess the composition of the deposited films. The EDS spectra for the bare FTO and the as - deposited fil ms as a function of the number of cycles are compared in Figure 4 5 . Due to the large sampling depth of this techniques, in addition to the deposited film (top layer) the FTO substrate (SnO 2 ) is also detected. The observed peaks are then assigned to Ta, Sn, O, and N (the lists of characteristic X - ray for these elements are summarized in Table A 4 - 2 ). We also note that for all thicknesses no signature of chloride as an impurity was observed which further suggests that the reaction between ammonia and tantalum chloride runs to completion. The broad peaks located at 0.69 and 3.4 keV are attributed to tin (Sn) from the FTO substrate. For a similar beam energy (15 kV), this peak has the highest intensity for the bare FTO substrate but its intensity progressively declines as the number of depo sition cycles 150 increases. Concurrently, the intensity of Ta peaks (at ~ 1.7 and 7 > keV) grows with the number of cycles used. This is an important observation as it readily indicates that the deposited material forms a layer on top of the FTO substrate and that its thickness grows with the number of cycles. While these spectra confirms the presence of nitrogen (peaks at 0.4 and 0.825 keV, respectively), due to the low sensitivity of this technique to the light elements the ratio of Ta to N cannot be determi ned. Figure 4 6 . The XRD patterns of the FTO (SnO 2 ) substrates before (black) and after (orange) annealing in ammonia and tantalum nitride films with various thicknesses on FTO substrate. Films with 150 cycles, 300, and 500 cycles are shown in blue, green, and red colors, respectively. The vertical black dashed lines represents the Bragg positions for the standard crystalline Ta 3 N 5 powder with random crystal orientations (PDF # 01 - 079 - 1533). The numbers r epresent the M iller indices of the corresponding diffraction peaks of Ta 3 N 5 . The yellow diamonds represent the Bragg positions of the standard SnO 2 (PFD # 00 - 046 - 1088). 151 The XRD diffraction patterns of the bare FTO substrates before and after annealing in ammonia with a flow rate of 100 SCCM at 550 °C for 2 h are compared to the deposited films with various number of cycles in Figure 4 6 . The diffractio n patterns for the bare FTO substrates before and after ammonolysis (control samples) are identical and are assigned to SnO 2 . The measured diffraction patterns for the deposited films (for all thicknesses) are unambiguously assigned to Ta 3 N 5 and SnO 2 . In a ddition, the peak positions of SnO 2 for the deposited films are identical to that of the control samples. These observations collectively indicate that FTO is chemically and structurally stable under the deposition conditions of the crystalline Ta 3 N 5 . Inte restingly, we note that under identical XRD measurement conditions, the ratio of the diffraction intensities of SnO 2 to Ta 3 N 5 progressively decrease with increasing Ta 3 N 5 thicknesses. In agreement with the EDS data ( Figure 4 5 ) , this observation indicates that the deposited films form a layer over the FTO substrates. A closer inspection of the XRD pattern as a function of film thickness reveals a strong correlation between the pattern of diffraction (relative peak intensities) and the number of deposition cycles. In comparison to the diffraction pattern of the randomly oriented reference Ta 3 N 5 (vertical dashed lines), it can be realized that the diffrac tion peak located at 24.6° ( M iller indices of (110)) - which is one of the intense diffractions for the reference - is missing for all of the films. For this to occur, the (110) plane must be perpendicular to the substrate, implying that the deposited film s exhibit a preferential orientation. A detailed analysis of the crystal orientation of the films are discussed following Figure A 4 - 6 . In addition, a s it can be seen, the relative intensity of (130) plane to (113) plane strongly depends on the number of deposition cycles; it approaches the relative ratio observed for the rand omly oriented reference pattern as the number of cycles increases. It can ther efore be hypothesized that the different lattice structure of Ta 3 N 5 compared to SnO 2 , generated a lattice mismatch, such that the initial growth mode of Ta 3 N 5 is different and 152 proceeds through the formation of a buffer layer on the FTO substrate. Following the formation of the buffer layer, as the number of cycle increases the effect of the lattice mismatch is minimized and Ta 3 N 5 grows more randomly. It is therefore expected that the growth rate and morphology of the crystalline Ta 3 N 5 by this method is likely strongly substrate - dependent, which is beyond the scope of this study. The Raman spectrum of films of different thicknesses are compared in Figure A 4 - 7 . In line with the XRD measurements, the observed phonon modes are consistent with crystalline Ta 3 N 5 . The SEM images of the bare FTO substrates before and after annealing in ammonia are shown in Figure A 4 - 8 . As depicted, while after annealing in ammonia some surface aggregations are observed the film thicknesses are constant ~ 240 nm. In line with the XRD patterns of the FTO, this observatio n further indicates that SnO 2 phase is stable under reducing atmosphere of ammonia. The top view and cross - section SEM as well as the AFM images of Ta 3 N 5 films with various thicknesses are compared to the bare FTO substrate in Figure 4 7 . The cross - section SEM images clearly show that the films are comprised of two distinct layers of Ta 3 N 5 (top layer) and FTO (bottom layer). The Ta 3 N 5 films are compact, and their thicknesses increases with the number of cycles, whereas the thickness of the FTO layer remains constant. Due to the lack of a visually sharp interface between these two layers, the thickness of the Ta 3 N 5 film was calculated by subtracting the original thickness of th e FTO layer (240 nm) from the total thickness. 153 Figure 4 7 . The SEM and AFM images showing the cross - section , morphology, and topology of the bare FTO and Ta 3 N 5 films with di f ferent thickness. The scal e bars for the SEM images are 200 nm with ~ 200,000X magnification. The 3D AFM images are shown in Figure A 4 - 9 . The thickness of the films with 150, 300 , and 500 cycles were found to be 70, 103, and 230 nm, respectively. Based on the top view SEM and AFM topography images, the morphology of the films are strongly correlated to the number of cycles (film thickness). The film with 70 nm thickness has a simi lar spherical morphology as the FTO substrate but with the bigger features and a porous morphology. Interestingly, the morphology of the film with 103 nm thickness begins to deviates from the FTO substrate exhibiting new oval - shape features where the film with 230 nm thickness only exhibit a randomly distributed oval - shape features perpendicular to surface. Remarkably, the morphology transformation as a function of the film thickness described here follows the same trend as the evolution of the XRD pattern which similarly can be ascribed to the difference in the crystal structure of Ta 3 N 5 and FTO as described above. AFM was used to analyze the roughness of the films with various thicknesses. It was found that while the mean roughness 154 of the films with differ ent thicknesses are similar but it is increased by ~ 3 nm with respect to the bare FTO substrate ( Table A 4 - 3 ). The optical properties of bare FTO subs trates before and after annealing in ammonia are shown in Figure A 4 - 10 . As it can be seen, while the reflectance of the FTO substrate is not affected by the annealing in ammonia, however, its transmittance is reduced by ~ 10%. This reduction can be ascribed to the surface agglomeration observed in the SEM image ( Figure A 4 - 8 ). The absorptance spectra for various thicknesses of Ta 3 N 5 on FTO substrate are shown in Figure 4 8 . a. As shown, the absorptance edge for these films occurs at 575 - 600 nm which is in well agreement with the known bandgap of Ta 3 N 5 . 19,24,25 Clearly, the absorptance beyond the bandgap edge does not reach zero, it is specifically more pronounced for the thicker films. The non - zero sub - bandgap absorptance have previously been attributed to the formation of th e reduced Ta - sites. 3 In addition, a well - resolved sub - bandgap absorption peak centered at ~ 700 nm is uniquely observed for the film with 500 cycles. The origin of this absorption feature have been ass igned to two completely distinctive sites: (1) the reduced Ta 5+ sites, i.e. Ta 4+ or Ta 3+ , 26,27 and (2) the N - vacancy as a deep donor accompanied by free electron in the conduction band to compensate the positive charge of the anion vacancy. 5,28 The main difference between these two sites is that the first one is basically a trap state where the charges are localized on the Ta - sites while in the second one the charge is delocalized over the conduction band. 155 Figure 4 8 . a) Plots of the a bsorptance of Ta 3 N 5 films on FTO subs trate with various thicknesses. The corresponding transmittances and reflectances are shown in Figure A 4 - 11 . The data shown here are corrected for the pristine FTO substrate using the previously reported procedure. 22 b) The thicknes s of deposited films as a function of number of cycles grown at 550 °C on FTO substrate. The thickness of the films were determined via two independent methods of cross - section SEM analysis (black square) and the optical absorbance at 450 nm (red circle) i n combination with beer - lambert law using the known absorption coefficient reported previously for compact film of Ta 3 N 5 . 23 Figure 4 8 . a also shows that the absorptance of the film does not scale with the number of cycles and the film with 300 cycles has a slightly higher absorptance than the film with 150 cycles. The plot of thickness vs. number of cycles a cquired by the cross - section SEM and absorbances are shown in Figure 4 8 . b . We note that the film thicknesses found by these methods are different by as much as ~ 50 nm for the thickest film which can be ascribed to the uncertainty in the absorption coefficient an d/or the calculated absorbance or change in the density of the films with various thicknesses. The trend in the film thickness as a function of number of cycles between two methods 156 are remarkably similar, however. As shown, within the window of the number of cycles studied here, the growth of tantalum nitride on FTO substrate is not linear. This phenomena can be better understood by considering the growth of the crystalline tantalum nitride on films FTO. As show n in the SEM images the film with 150 cycle is porous this is further manifested in the high dark current shown in Figure 4 9 whereupon growing more materials, they fill the empty spaces. Con sistent with the XRD and SEM images, the trend of the film growth indicates that the deposition of the tantalum nitride initiates by seeding and formation of a buffer layer for the first 150 cycles. Based on these thicknesses, the apparent growth rate of t he Ta 3 N 5 films on FTO is ~ 0.46 nm per cycle. We note that this growth rate is substantially higher than the previously reported values for the ALD deposition of crystalline tantalum nitride by a factor of ~ 18 (see Table 4 - 1 ). For example, Ritala et al. 17 utilized TaCl 5 and NH 3 to deposit Ta 3 N 5 on soda lime glass and ITO glass substrate s . Under ALD growth conditions, i.e. self - terminating surface reaction, they reported an average growth rate of 0.025 nm per cycles at the temperature range of 400 to 500 °C. It is also worth notin g that a growth rate of 0.46 nm per cycle is larger than the average bond length of Ta - N (0.21 nm) in Ta 3 N 5 by a factor of ~ 2.2. 29 These observations collectively indicate that the deposition of Ta 3 N 5 here is not under self - limiting ALD deposition conditions. However the higher grow th rate of the film with a controlled thickness found here is advantageous as it minimize the processing time to prepare tantalum nitride films . 157 Figure 4 9 . a) The chopped light J - E curves of Ta 3 N 5 on FTO substrate as a function of film thickness in co n tact with an aqueous solution containing 0.1 M K 4 [Fe(CN) 6 ] as the hole scavenger with the pH of 6.8. The potential was scanned at 20 mV s - 1 and the electrodes were illuminated through the electrolyte with 1 sun intensity. b) The photograph of the electrode with various thicknesses of Ta 3 N 5 on FTO substrate. As shown all the edges of the rectangular shape electrode were coated with Ag epoxy and the electrode was clamped to a costume made cell with an O - ring with diameter of 0.19 cm - 2 (the mark of the O - ring is partially visible on the 500 cycles electrode). The deposited tantalum nitride films on FTO substrate are strongly adhered to the substrate, thus, sandpaper was used to expose the FTO conducti ve su bstrate . The J - E curves in presence of the 1 - elec t ron fast hole scavenger along with the photograph of the electrodes with various thicknesses are shown in Figure 4 9 . Previous example of Ta 3 N 5 on FTO substrate was reported by Higashi et al. 30 . They used electrophoretic deposition to prepare mesoporous films from nanoparticles Ta 3 N 5 . 158 The best performance was achieved by treating the as - prepared films with TaCl 5 followed by ammonolysis at 500 °C. The other example of Ta 3 N 5 photoanode on a tra nsparent conductive substrate was reported by our group via ALD stack deposition of tantalum oxynitrides on Ta - doped TiO 2 followed by ammonolysis (described in previous chapter). 19 To the best of our knowledge this is the first example of directly deposited crystalline Ta 3 N 5 photo anode on the FTO substrate. The J - E curves reported here produce a photocurrent onset potential of ~ 0.3 V vs. RHE. This potential is comparable to the onset potential reported by Jaramillo et al. for the tandem core - shell Si - Ta 3 N 5 photoanode and also the high performing nanostructured Ta 3 N 5 photoanodes introduced by Wang and coworkers. 31 33 The early photocurrent onset potential readily implies that (1) Ta 3 N 5 make s an ohmic contact with the FTO substrate and (2) FTO i s sufficiently conductive that the energy lost for electron collection and transport is minimized. This observation is further supported by the conductivity measurement of the FTO substrate before and after annealing in ammonia. As shown in the slope of th e J - E curves for both electrodes are nearly identical confirming that FTO is stable and remains conductive under the reducing conditions of ammonia at 550 °C. Another feature of the chopped light J - E curves shown in Figure 4 9 is the saturation of photocurrent density for the film with 103 nm thickness. The current saturation readily indicates that the transport of the holes (minority charge carrier) is limiting while electrons are efficiently collected at 230 nm thickness. The la t ter can be ascribed to the high dopant density found for Ta 3 N 5 (discussed below). The chopped light J - E curves for Ta 3 N 5 films with various thicknesses as the function of illumination direction (front vs. back) are compared in Figure A 4 - 13 . We note that while the photo current onset potential is independent to the illumination direction, the photocurrent density, however, strongly correlates to the illumination direction. The 70 nm film produces a relatively constant photocurrent regardless of the illumination direction. However, as the film 159 thickness grows the front and back illumination start to diverge with the front illumination constantly producing higher photocurrent and plateauing beyond 103 nm film thickness with a maximum photocurrent density of ~ 2.4 mA cm - 2 at 1.23 V vs. RHE. We note that almost all the previous examples of Ta 3 N 5 photoanodes that exhibited high photocurrent densities were nanostructured electrodes with high aspect ratios ( n anorod , nanotube, or porous cubic morphologies). 1,2,32,34 Jaramillo and coworkers have also shown that the photocurrent density of Ta 3 N 5 predominantly depends on its effective surf ace area. 35 Here we note that for the planar Ta 3 N 5 films, the maximum current density reaches ~ 20% of the theoretical limit (12.5 mA cm - 2 ). These observations collectively indicate that PEC performance of tantalum nitride is dominantly controlled by its finite diffusion of holes (at best between 70 and 103 nm film thicknesses). Figure 4 10 .a shows the wavelength dependence of incident photon to current efficiency (IPCE%) for various thicknesses of Ta 3 N 5 as a function of illumination direction. As depicted, the onset of IPCEs for all the thicknesses of these electrodes is located around 575 nm which is in good agreement with the known band gap value for Ta 3 N 5 19,24 , indicating that the observed photoresponse is due to tantalum nitride. The sharp decline in the IPCE% values which is uniquely observed in the fron t illumination (through electrolyte) at the wavelengths below 450 nm can be ascribed to the absorption of the ferri/ferrocyanide solution. 36 In agreement with Figure A 4 - 13 , the front IPCE% is larger than the back IPCE%, which further indicates that the transport of holes is kinetically limiting. Consistent with the J - E curves, the front IPCE% responses saturates for the films with 103 nm thickness. 160 Figure 4 10 . Wavelength dependence of the a) IPCE% and b) APCE% values at 1.0 V vs. RHE for the Ta 3 N 5 films with different thicknesses as a function of illumination direction in co n tact an aqueous solutio n containing 0.1 M K 4 [Fe(CN) 6 ] as the hole scavenger with the pH of 6.8. The solid and open shapes represent the values under front (through the solution) and back (through) illumination, respectively. Ta 3 N 5 films with 230, 103, and 70 nm thicknesses are s hown in red diamond, green circle, and blue triangle, respectively. The back IPCE% response initially increases as the film thickness grows but it steeply declines for the 230 nm thick film. In addition, the back IPCE% values for the films with 70 and 103 nm thicknesses are plateau ed with average values of approximately 10 and 15%, respectively. Conversely, the back IPCE% for the 230 nm Ta 3 N 5 electrode produced a peak at 525 nm and declined to zero for the lower wavelengths. These observations can be better understood by considering the absorption penetration depth of tantalum nitride and the profile of photo generation of charge carriers as a function of illumination direction ( Figure 4 11 ). The penetration depth of light for Ta 3 N 5 for the photons with the wavelength of 525 and 450 nm are 120 and 52 nm , respectively. 19,23 As a result, for the back illumination photons with larger penetration depths (red 161 photo ns) produce charge carrier within the diffusion length of hole (in vicinity of the surface) while the short wavelength lights (blue photons) with short absorption length predominately generate holes close to the FTO substrate. As the result, for the film w ith 230 nm thickness the back IPEC% produces a peak rather than a constant value. The APCE% values of these films as a function of illumination direction for various thicknesses are shown in Figure 4 10 .b. As shown, the maximum APCE% is observed for the film with 103 nm thickness. For Ta 3 N 5 with 230 nm thickness, even though the absorptance is quantitative ( Figure 4 8 . a ) but the APCE% decreases. In agreement with IPCE%, this observation indicates that for this films the charge carries are produced outside of the holes diffusion length and thus do not contributed to the photocurrent. The APCE% for 70 nm is lower than the 103 nm film which can be ascribed to its lower IPCE%. Figure 4 11 . The schematic representation of the profile of charge generation as a function of illumination direction; a) back illumination (through the FTO), b) front illumination (through the solutions). 162 We used dark EIS to assess the electrical properties of Ta 3 N 5 on FTO substrate. The EIS responses were fitted to the Rundle mode l circuit to extract the capacitance of the space charged region ( C bulk ). The Mott - Schottky (MS) plot of Ta 3 N 5 with 230 nm thickness in dark is shown in Figure A 4 - 14 . The dopant density (N d ) was calculated by linear fitting of the slop to the MS equation 4 - 4 . 37 We note that in the literature there is no agreement on the dielectric constant of Ta 3 N 5 and different values ranging from 7 to 110 have been reported. 2,23 ,24,38,39 The calculated dopant density based on these dielectric constants are summarized in Table A 4 - 4 . As shown, the calculated dopant densitie s span 1 order of magnitude in range with an average value of 8.0 × 10 20 cm - 3 in agreement with previous reports. 2,23 The flat band potential (V FB ) of Ta 3 N 5 was calculated from the intercept of the linear fit according to equation 4 - 4 . In this equation, A, k B 0 , and V app absolute temperature, elementary charge, dielectric constant of the semiconductor, permittivity of free space, and applied potential, respectively. From the intercept of the linear fit the flat band potential was calculated as 0.21 V vs. RHE. With the solution potential of 0.68 V (measured in two - electrode system), therefore the built - in voltage (V bi 4 - 5 The depletion width (w) was calculated according to equation 4 - 5 and the results for various dielectric constants are summarized in Table A 4 - 4 . The average depletion widths is only 4 nm. 4 - 4 163 The narrow depletion width can be ascribed to the high dopant density. Due to the narrow depletion width and the large absorption depth of Ta 3 N 5 , a significantly small number of charge carrie r s are generated in the depletion region which cannot account for the observed photocurrent densities. Since the depletion width is narrow in comparison to the thinnest film here (70 nm) thus th e PEC performance of Ta 3 N 5 is controlled by the diffusion of photogenerated holes. 4.5. Conclusio ns In this chapter, the experimental procedures to build a fully automated ALD instrument is described. In this study, a combination of TaCl 5 and ammonia were used to deposit crystalline Ta 3 N 5 . It was shown that the quality and the composition of the film strongly depends on the configuration and the composition of the reactor. Specifically, it was discovered that for the deposition of crystall ine Ta 3 N 5 , the cold - wall reactor equipped with substrate heater is advantageous over the hot - wall reactor as it minimizes the deleterious side reaction induced by the highly reactive nature of the deposition conditions of crystalline Ta 3 N 5 temperature > 500 °C and the presence of highly reactive reactants (NH 3 ) and by product (HCl). Subsequently, crystalline Ta 3 N 5 with various thicknesses were deposited on FTO substrate at 550 °C. We note that the crystallinity, morphology and topography, growth rate, and optical density of the deposited films on FTO strongly depend on the number of ALD cycles. These correlations were attributed to the crystal mismatch between SnO 2 , i.e. the FTO substrate, and Ta 3 N 5 . Therefore, for the first ~ 150 cycles a buffer layer wit h relatively similar morphology to the substrate is formed on the surface of FTO where after this point Ta 3 N 5 is randomly deposited. Furthermore, it was shown that FTO is chemically and mechanically stable under reducing conditions of ammonia at 550 °C and remains transparent and conductive which further allowed to realize the 1 st example of directly 164 deposited Ta 3 N 5 photoanode on FTO. The Mott Schottky analysis from EIS measurements revealed that Ta 3 N 5 is heavily doped (N d = 8.0 × 10 20 cm - 3 ) which results i n a rather narrow depletion with (< 10 nm). This readily indicates that the charge transport in Ta 3 N 5 is under diffusion control. Furthermore, the PEC performance of these electrodes in presence of the fast hole scavenger as a function of illumination dir ection with various thicknesses have shown that the PEC performance of Ta 3 N 5 is controlled by the limited diffusion length of holes with the diffusion length of 70 - 100 nm. This observation contradicts with the previous study by Van de Krol et al. 6 where they predicted that the diffusion length of charge carriers for Ta 3 N 5 is in the order of ~ 18 ,000 nm . However, we note that this observation is consistent with the f act that the best examples of Ta 3 N 5 in literature have a nanostructured morphology with feature size of ~ 100 nm. 1,2,32,33 In conclusion, here we showed that ALD provides a viable and robust technique to integrate crystalline Ta 3 N 5 with a transparent conductive substrate like FTO. As a result, it allowed us to study the physical properties of pristine Ta 3 N 5 independent of the conductivity of the substrate. Furthermore, the ALD deposited Ta 3 N 5 electrodes can further be in situ protec ted against photocorrosion with an over layer of GaN or TiO 2 prior to exposure to the ambien t atmosphere. 40 42 165 A PPENDIX 166 Figure A 4 - 1 . The screenshot of the program in use. 167 Figure A 4 - 2 . The photograph of the electronic box. 168 Figure A 4 - 3 . The wiring diagram of the electronic box to the LabJack T7 - Pro. 169 The individual components of the electron ic box are briefly described below: 1) Power Supply: two types of power supplies are used here: (i) a 120V a lternating current (AC) power supply is used for the heaters and the vacuum valve, and (ii) a 24 V direct current (DC) power supply is used to power the ALD valves. 2) Solid State Relays (SSR): solid state relays are basically electronic switching devices without a moving part. They turn on/off an external circuit (loads) by applying a small external voltage across the control terminal. Sine here we u sed two different power supplies, i.e. AC and DC, two different types of SSR, DCDC - SSR and DCAC - SSR were accordingly utilized to control the AC and DC loads. It is important to note that the control signal is a 5 V DC voltage supplied by the LabJack ( vide inf ra ). 3) Digital to Analog Convert e r (DAC) and Analog to Digital Convert e r (ADC): The heart of the electronic box is the DAC/ADC. The real world signals, i.e. temperature, position of the solenoid valve, are analog signals in terms of voltage. The ADC co nverts these signals into digital signals that can be processed by the computer. Conversely, the DAC converts the command from computer into analog signal to be executed on the system. For this electronic box, the LabJack T7 - Pro with 14 analog inputs and 2 4 - bit low - speed ADC with resolution of 1 µV was used. Further information about the specifics of configuration of the convert e r can be find from the Labjack website [ 43 ]. 170 Figure A 4 - 4 . The photograph of the meta l lic copper powder filled scrubber equipped with a band heater and thermocouple . The phot o graph was tak e n ~ 4 months after continuous flow of N 2 with an average flow of 20 SCCM. The black area at the inlet represent the formation of the copper oxide and the red copper color at the outlet represent the unreacted copper powder. 171 Figure A 4 - 5 . The schematic representation of the procedure developed to deposit Ta 3 N 5 via a custom - built ALD system. 172 Figure A 4 - 6 . The crystal structure of Ta 3 N 5 along and its 3 major planes with the M iller ind ic es of: (110) shown in blue, (130) shown in orange, and (113) shown in yellow. The uni t cells shown in (a) and (b) are the same but they are represented along different axes: a) along the c axis and b) alon g the b axis that is rotated by 21° counterclockwise. Since the diffractions from all the planes except (110) are observed, the visual inspection of the relative orientations of the major atomic planes of Ta 3 N 5 indicates that Ta 3 N 5 is preferentially deposi ted along the b - axis rotated ~ 21° counterclockwise where the c - axis is parallel to the surface. 29 173 Figure A 4 - 7 . Raman spectra of Ta 3 N 5 films with different thicknesses on FTO substrate. Films grown with 150 cycles, 300, and 500 cycles are shown in blue, green, and red color, respectively. These phonon modes are in agreement with the previously reported Raman spectrum of Ta 3 N 5 . 44 174 Figure A 4 - 8 . The cross - section and top view SEM images of the as received and bare FTO substrate a - b) before and c - d) after annealing in ammonia with a flow rate of ~ 100 SCCM at 550 °C for 2 hours. The scal e bars for the SEM images are 1 00 nm with ~ 200,000 X magnification. 175 Figure A 4 - 9 . The AFM images of the bare and Ta 3 N 5 with different thicknesses on FTO. 176 Figure A 4 - 10 . The transmittance (solid lines) and reflectance (dashed lines) of the bare FTO substrate before (black) and after (orange) annealing in ammonia with a flow rate of ~ 100 SCCM at 550 °C for 2h. As it can be seen, while the reflectance of the FTO substrate i s not affected by the annealing in ammonia, however, the transmittance of the FTO substrate is reduced by ~ 10% after annealing in ammon ia. 177 Figure A 4 - 11 . The optical properties of Ta 3 N 5 on FTO substrate as a function of film thickness; a) transmittance and b) reflectance. Figure A 4 - 12 . The electrical properties of FTO before and after annealing in ammonia with a flow rate of 200 SCCM for 2 hours at 550 °C. The J - E curves were measured from the electrodes with the same geometry and dimensions (details are shown in the inset). 178 Figure A 4 - 13 . The front vs. back illumination J - E curves of Ta 3 N 5 on FTO substrate with various thickne s ses in co n tact with an aqueous solution containing 0.1 M K 4 [Fe(CN) 6 ] as the hole scavenger with the pH of 6.8 ; a) 150 cycles (70 nm), b) 300 cycles (103 nm), c) 500 cycles (230 nm). The back ill u mination (through the FTO substrate) is shown in black and the front illuminations are shown in color. 179 Figure A 4 - 14 . a) the C bulk and b) the calculated M ott S chottky plot of Ta 3 N 5 with 230 nm thickness on FTO substrate in dark. The electrode was in co n tact with an aqueous solution containing 0.1 M K 4 [Fe(CN) 6 ] as the hole scavenger with the pH of 6.8 . The inset represents the equivalent circuit used to fit the EIS data. 180 Set Temperature (°C) Measured Temperature (°C) 450 444 505 500 550 540 Table A 4 - 1 . The calibration data for the substrate temperature. Element Peak position (k eV) Ta 1.709 ( M ), 8.145 ( L ) Sn 3.443 ( L ), 0.691 ( M ) Si 1.739 ( K ) Cl 2.621 ( K ) O 0.523 ( K ) N 0.392 ( K ) Table A 4 - 2 . The characteristic X - ray of Ta, Sn, Cl, Si, O, and N. 45 Sample Mean Roughness / nm Bare FTO 8.053 150 cycles 10.50 300 cycles 10.64 500 cycles 11.53 Table A 4 - 3 . The measured mean rou g hness of the FTO substrate and deposited films with different thickness. 181 a N d × 10 20 (cm - 3 ) b W (nm) b Ref. 110 1.68 6 [ 2 ] 65 2.84 3.5 [ 24 ] 43 4.29 2.5 [ 38 ] 17 10.9 1 [ 23 ] 9 20.5 0.5 [ 39 ] (a) T hese values were taken from the literature . (b) These values are the calculated parameters using the literature dielectric constants. Table A 4 - 4 . The physical properties of Ta 3 N 5 . 182 REFERENCES 183 REFERENCES (1) Liu, G.; Ye, S.; Yan, P.; Xiong, F.; Fu, P.; Wang, Z.; Chen, Z.; Shi, J.; Li, C. Enabling an Integrated Tantalum Nitride Photoanode to Approach the Theoretical Photocurrent Limit for Solar Water Splitting. Energy Environ. Sci. 2016 , 9 (4), 1327 1334. (2) Li, Y.; Zhang, L.; Torres - Pardo, A.; González - Calbet, J. M.; Ma, Y.; Oleynikov, P.; Terasaki, O.; Asahina, S.; Shima, M.; Cha, D.; et al. Cobalt Phosphate - Modified Barium - Doped Tantalum Nitride Nanorod Phot oanode with 1.5% Solar Energy Conversion Efficiency. Nat. Commun. 2013 , 4 , 2566. (3) Pinaud, B. A.; Vailionis, A.; Jaramillo, T. F. Controlling the Structural and Optical Properties of Ta 3 N 5 Films through Nitridation Temperature and the Nature of the Ta M etal. Chem. Mater. 2014 , 26 (4), 1576 1582. (4) Dabirian, A.; Van De Krol, R. High - Temperature Ammonolysis of Thin Film Ta 2 O 5 Photoanodes: Evolution of Structural, Optical, and Photoelectrochemical Properties. Chem. Mater. 2015 , 27 (3), 708 715. (5) Dabi rian, A.; Van De Krol, R. Resonant Optical Absorption and Defect Control in Ta 3 N 5 Photoanodes. Appl. Phys. Lett. 2013 , 102 (3), 033905. (6) De Respinis, M.; Fravventura, M.; Abdi, F. F.; Schreuders, H.; Savenije, T. J.; Smith, W. A.; Dam, B.; Van De Krol, R. Oxynitrogenography: Controlled Synthesis of Single - Phase Tantalum Oxynitride Photoabsorbers. Chem. Mater. 2015 , 27 (20), 7091 7099. (7) Henderson, S. J.; Hector, A. L. Structural and Compositional Variations in Ta 3 N 5 Produced by High - Temperature Ammon olysis of Tantalum Oxide. J. Solid State Chem. 2006 , 179 (11), 3518 3524. (8) Terao, N. Structure of Tantalum Nitrides. Jpn. J. Appl. Phys. 1971 , 10 (2), 248 259. (9) Ganin, A. Y.; Kienle, L.; Vajenine, G. V. Plasma - Enhanced CVD Synthesis and Structural Characterization of Ta 2 N 3 . Eur. J. Inorg. Chem. 2004 , 2004 (16), 3233 3239. (10) Hieber, K. Structural and Electrical Properties of Ta and Ta Nitrides Deposited by Chemical Vapour Deposition. Thin Solid Films 1974 , 24 (1), 157 164. (11) Fix, R.; Gordon, R. G.; Hoffman, D. M. Chemical Vapor Deposition of Vanadium, Niobium, and Tantalum Nitride Thin Films. Chem. Mater. 1993 , 5 (5), 614 619. (12) Hoffman, D. M. Chemical Vapour Deposition of Nitride Thin Films. Polyhedron 1994 , 13 (8), 1169 1179. (13) Chen, X.; Peterson, G. G.; Goldberg, C.; Nuesca, G.; Frisch, H. L.; Kaloyeros, A. E.; Arkles, B.; Sullivan, J. Low - Temperature Chemical Vapor Deposition of Tantalum Nitride from Tantalum Pentabromide for Integrated Circuitry Copper Metallization Applications. J . 184 Mater. Res. 1999 , 14 (5), 2043 2052. (14) Chiu, H. T.; Chang, W. P. Deposition of Tantalum Nitride Thin Films from Ethylimidotantalum Complex. J. Mater. Sci. Lett. 1992 , 11 (2), 96 98. (15) Deposition of Ta(Al)N(C) Thin Films Using Trimethylaluminum as a Reducing Agent. J. Electrochem. Soc. 2001 , 148 (10), G566. (16) Kwon, J. - D.; Yun, J.; Kang, S. - W. Comparison of Tantalum Nitride Films for Different NH 3 / H 2 /Ar Reactant States in Two - Step Atomic Layer Deposition. Jpn. J. Appl. Phys. 2009 , 48 (2), 025504. (17) Ritala, M.; Kalsi, P.; Riihelä, D.; Kukli, K.; Leskelä, M.; Jokinen, J. Controlled Growth of TaN, Ta 3 N 5 , and TaO x N y Thin Films by Atomic Layer Deposit ion. Chem. Mater. 1999 , 11 (7), 1712 1718. (18) Hiltunen, L.; Leskelä, M.; Mäkelä, M.; Niinistö, L.; Nykänen, E.; Soininen, P. Nitrides of Titanium, Niobium, Tantalum and Molybdenum Grown as Thin Films by the Atomic Layer Epitaxy Method. Thin Solid Films 1988 , 166 (C), 149 154. (19) Hajibabaei, H.; Zandi, O.; Hamann, T. W. Tantalum Nitride Films Integrated with Transparent Conductive Oxide Substrates: Via Atomic Layer Deposition for Photoelectrochemical Water Splitting. Chem. Sci. 2016 , 7 (11), 6760 6767. (20) Fang, Z.; Aspinall, H. C.; Odedra, R.; Potter, R. J. Atomic Layer Deposition of TaN and Ta3N5using Pentakis(Dimethylamino)Tantalum and Either Ammonia or Monomethylhydrazine. J. Cryst. Growth 2011 , 331 (1), 33 39. (21) Hajibabaei, H.; Schon, A. R.; Hamann, T. W. Interface Control of Photoelectrochemical Water Oxidation Performance with Ni 1 - x Fe x O y Modified Hematite Photoanodes. Chem. Mater. 2017 , 29 (16), acs.chemmater.7b01149. (22) Klahr, B. M.; Martinson, A. B. F.; Hamann, T. W. Photoelectrochemical Investigation of Ultrathin Film Iron Oxide Solar Cells Prepared by Atomic Layer Deposition. Langmuir 2011 , 27 (1), 461 468. (23) Ziani, A.; Nurlaela, E.; Dhawale, D. S.; Silva, D. A.; A larousu, E.; Mohammed, O. F.; Takanabe, K. Carrier Dynamics of a Visible - Light - Responsive Ta 3 N 5 Photoanode for Water Oxidation. Phys. Chem. Chem. Phys. 2015 , 17 (4), 2670 2677. (24) Morbec, J. M.; Narkeviciute, I.; Jaramillo, T. F.; Galli, G. Optoelectron ic Properties of Ta 3 N 5 : A Joint Theoretical and Experimental Study. Phys. Rev. B 2014 , 90 (15), 155204. (25) Chun, W. J.; Ishikawa, A.; Fujisawa, H.; Takata, T.; Kondo, J. N.; Hara, M.; Kawai, M.; Matsumoto, Y.; Domen, K. Conduction and Valence Band Posit ions of Ta 2 O 5 , TaON, and Ta 3 N 5 by UPS and Electrochemical Methods. J. Phys. Chem. B 2003 , 107 (8), 1798 1803. 185 (26) Maeda, K.; Terashima, H.; Kase, K.; Higashi, M.; Tabata, M.; Domen, K. Surface Modification of TaON with Monoclinic ZrO 2 to Produce a Compos ite Photocatalyst with Enhanced Hydrogen Evolution Activity under Visible Light. Bull. Chem. Soc. Jpn. 2008 , 81 (8), 927 937. (27) Yuliati, L.; Yang, J. H.; Wang, X.; Maeda, K.; Takata, T.; Antonietti, M.; Domen, K. Highly Active Tantalum(v) Nitride Nanop articles Prepared from a Mesoporous Carbon Nitride Template for Photocatalytic Hydrogen Evolution under Visible Light Irradiation. J. Mater. Chem. 2010 , 20 (21), 4295 4298. (28) Nurlaela, E.; Ould - Chikh, S.; Harb, M.; Del Gobbo, S.; Aouine, M.; Puzenat, E.; Sautet, P.; Domen, K.; Basset, J. M.; Takanabe, K. Critical Role of the Semiconductor - Electrolyte Interface in Photocatalytic Performance for Water - Splitting Reactions Using Ta 3 N 5 Particles. Chem. Mater. 2014 , 26 (16), 4812 4825. (29) Keeffe, M.; Rauch, P.; DiSalvo, F. J. Structure of Ta 3 N 5 at 16 K by Time - of - Flight Neutron Diffraction. Acta Crystallogr. Sect. C Cryst. Struct. Commun. 1991 , 47 (11), 2291 2294. (30) Higashi, M.; Domen, K.; Abe, R. Fabrication of Efficient TaON and Ta 3 N 5 Photoanodes for Water Splitting under Visible Light Irradiation. Energy Environ. Sci. 2011 , 4 (10), 4138. (31) Narkeviciute, I.; Chakthranont, P.; Mackus, A. J. M.; Hahn, C.; Pinaud, B. A.; Bent, S. F.; Jaramillo, T. F. Tandem Core Shell Si Ta 3 N 5 Photoan odes for Photoelectrochemical Water Splitting. Nano Lett. 2016 , 16 (12), 7565 7572. (32) He, Y.; Thorne, J. E.; Wu, C. H.; Ma, P.; Du, C.; Dong, Q.; Guo, J.; Wang, D. What Limits the Performance of Ta 3 N 5 for Solar Water Splitting? Chem 2016 , 1 (4), 640 65 5. (33) Cheng, Q.; Fan, W.; He, Y.; Ma, P.; Vanka, S.; Fan, S.; Mi, Z.; Wang, D. Photorechargeable High Voltage Redox Battery Enabled by Ta3N5and GaN/Si Dual - Photoelectrode. Adv. Mater. 2017 , 29 (26), 1700312. (34) Cheng, Q.; Fan, W.; He, Y.; Ma, P.; Vanka, S.; Fan, S.; Mi, Z.; Wang, D. Photorechargeable High Voltage Redox Battery Enabled by Ta 3 N 5 and GaN/Si Dual - Photoelectrode. Adv. Mater. 2017 , 29 (26), 1700312. (35) Pinaud, B. A.; Vesborg, P. C. K.; Jaramillo, T. F. Effect of Film Morphology and Th ickness on Charge Transport in Ta 3 N 5 /Ta Photoanodes for Solar Water Splitting. J. Phys. Chem. C 2012 , 116 (30), 15918 15924. (36) Cohen, S. R.; Plane, R. A. The Association of Ferrocyanide Ions with Various Cations. J. Phys. Chem. 1957 , 61 (8), 1096 1100. (37) Gelderman, K.; Lee, L.; Donne, S. W. Flat - Band Potential of a Semiconductor: Using the Mott Schottky Equation. J. Chem. Educ. 2007 , 84 (4), 685. (38) Nurlaela, E.; Harb, M.; Del Gobbo, S.; Vashishta, M.; Takanabe, K. Combined 186 Experimental and Theor etical Assessments of the Lattice Dynamics and Optoelectronics of TaON and Ta 3 N 5 . J. Solid State Chem. 2015 , 229 , 219 227. (39) Khan, S.; Zapata, M. J. M.; Pereira, M. B.; Gonçalves, R. V.; Strizik, L.; Dupont, J.; Santos, M. J. L.; Teixeira, S. R. Struct ural, Optical and Photoelectrochemical Characterizations of Monoclinic Ta 3 N 5 Thin Films. Phys. Chem. Chem. Phys. 2015 , 17 (37), 23952 23962. (40) Hu, S.; Shaner, M. R.; Beardslee, J. A.; Lichterman, M.; Brunschwig, B. S.; Lewis, N. S. Amorphous TiO 2 Coatings Stabilize Si, GaAs, and GaP Photoanodes for Efficient Water Oxidation. Science (80 - . ). 2014 , 344 (6187), 1005 1009. (41) Singh, A.; Fekete, M.; Gengenbach, T.; Simonov, A. N.; Hocking, R. K.; Chang, S. L. Y.; Rothmann, M.; Powar, S.; Fu, D.; Hu , Z.; et al. Catalytic Activity and Impedance Behavior of Screen - Printed Nickel Oxide as Efficient Water Oxidation Catalysts. ChemSusChem 2015 , 8 (24), 4266 4274. (42) Zhong, M.; Hisatomi, T.; Sasaki, Y.; Suzuki, S.; Teshima, K.; Nakabayashi, M.; Shibata, N.; Nishiyama, H.; Katayama, M.; Yamada, T.; et al. Highly Active GaN - Stabilized Ta 3 N 5 Thin - Film Photoanode for Solar Water Oxidation. Angew. Chemie Int. Ed. 2017 , 56 (17), 4739 4743. (43) T7 Series | LabJack https://labjack.com/products/t7 (accessed Mar 31, 2018). (44) Hajibabaei, H.; Hamann, T. W. Selective Electrodeposition of Tantalum(V) Oxide Electrodes. Langmuir 2017 , 33 (41), 10800 10806. (45) X - ray absorption edges, characteristic X - ray lines... 4.2.1 http://www.kayelaby.npl.co.uk/atomic_and_nuc lear_physics/4_2/4_2_1.html (accessed Apr 2, 2018). 187 Interface Control of Photoelectrochemical Water Oxidation Performance with Ni 1 x Fe x O y Modified Hematite Photoanodes Adapted w ith permission from: Interface Control of Photoelectrochemical Water Oxidation Performance with Ni 1 x Fe x O y Modified Hematite Photoanodes, Hamed Hajibabaei, Abraham R. Schon, and Thomas W. Hamann, Chem. Mater. , 2017 , 29 (16), 6674 6683. Copyright (2017). American Chemical Society . 1 88 5.1. Abstract In this work, Ni 1 - x Fe x O y - coated hematite electrodes are investigated as a model system of different semiconductor / catalyst interfaces. We found that the PEC performance of the electrodes strongly depends on both the way the hematite electrode is prepared and the composition of the catalyst. Two extreme behaviors are observed for electrodeposited hematite electrodes coated wit h slightly different compositions of catalyst. In case of Fe - rich catalyst (Ni 0.25 Fe 0.75 O y ) the performance is substantially enhanced compared to the bare electrode, however the Ni - rich (Ni 0.75 Fe 0.25 O y ) catalyst inhibits the PEC performance. A combination of photoelectrochemical, IMPS, and EIS measurements collectively reveal the critical role that the interface states of the semiconductor and catalyst plays in controlling the key interfacial charge transfer and recombination reactions. The photogenerated holes are efficiently collected and stored into the catalyst layer for the Ni 0.25 Fe 0.75 O y coated hematite electrodes. An unusually large improvement in performance is attributed to this hole collection circumventing recombination at the hematite surface. F or the Ni 0.75 Fe 0.25 O y coated hematite electrodes, however, there is a presence of interface trap states that act as recombination centers and pin the catalyst potential. These combined results provide important new understanding of the role of the interfac es at semiconductor/electrocatalyst junctions. 5.2. Introduction Hematite continues to attract a lot of interest as an electrode material for photoelectrochemical (PEC) water oxidation. Hematite has an optical band gap of ~ 2.1 eV, is composed of very earth - abu ndant elements, is easily prepared and is chemically stable in basic electrolytes. 1 3 Due to these attributes, there has been a lot of effort aimed at improving the performance of this material. 189 Previous studies suggest that one of the key steps preventing efficient PEC water oxidation at hematite electrode is the high overpotential necessary to initiate water oxidation. 3,4 The large overpotential is g enerally attributed to the slow kinetics of water oxidation at the surface of the hematite. Experiments by us and others, showed that after deposition of water oxidation catalysts (WOCs), e.g. CoPi or Ni(OH) 2 , the photoelectrode characteristics (e.g. photo current onset potential, photocurrent, and fill - factor) substantially improve. 2,5 11 The cause for these improvements, however, are not yet fully understood. The enhancement of PEC performance of electrocatalyst coated photoanodes has been attributed to reducing surface state recombination, 12,13 increasing band bending, 14,15 or facilitating charge separation. 10,16 Recently, Boettcher and coworkers developed a dual working electrode experiment to simultaneously measure the current and potential of the semiconductor (SC) and catalyst. 17,18 Their results suggest that the SC/WOC interface is strongly affected by the physical structure of the catalyst, specifically whether it is dense or io n permeable. To the best of our knowledge, all catalysts reported in conjunction with hematite are ion - permeable , including the most often reported Co - Pi. 9 For such ion - permeable WOCs, hole transfer from the semiconductor oxidizes the WOC, and the accumulated holes a re charge compensated by ions in the electrolyte. Thus, the catalyst potential drops until it can sustain water oxidation at a rate matching the flux of holes reaching the SC/WOC interface. This type of junction is termed an adaptive junction. The adaptive junction model assumes an ideal SC/WOC interface and the measurements reported by Boettcher utilized an ideal single crystal TiO 2 semiconductor which did not appear to significantly suffer from surface state recombination. 17 We have shown, however, that surface states on hematite largely control the water oxidation behavior and cannot necessarily be neglected in considering the addition of WOC s to the hematite surface. 19,20 For example, it was shown that 190 surface modification of hematite that eliminates the surface defects resulted in a substant ial improvement of PEC performance of hemati te electrodes modified with WOC s. 21 In addition, it has been reported that many other photoanode materials such as BiVO 4 suffer from surface state recombination as well. 22,23 The presence o f a high density of surface states can result in substantial surface state recombination and F ermi level pinning . 9,24 Their influence on the interface that forms upon addition of a WOC is unknown, however. Further more , the addition of materials to the hematite (and other metal oxide) surfaces can result in new/more interface states that can inhibit the performance. It is also not clear why some hematite/WOC juncti ons produce exceptional results compared to most other junctions which result in modest improvements. 21,25,26 Therefore, in this study we aim to expand the understanding of SC/WOC interfaces, including non - ideal junctions, which can be further generalized for a larger groups of photoelectrode/WOCs combinations. In this work we employ hematite thin films prepared by atomic layer deposition (ALD) and electrodeposition (ED) methods as photoanodes. Recently we have shown that the preparation method of hematite has a great impact on the performance of the photoanode. 2 The ED hematite electrodes outperformed the ALD - made films which was attributed to enhanced hole tra nsport and collection efficiencies . Therefore, it is expected that the de nsity of defects at the surface may also be altered as a result of different preparation method s . The hematite films were coated with Ni 1 - x Fe x O y WOCs. Consistent with prior reports, we found that tuning the composition of the Ni 1 - x Fe x O y catalysts results in the formation of one or two separate phases, including FeOOH and Fe - doped NiOOH. 27 30 Thus, a combination of hematite electrodes prepared by different methods and Ni 1 - x Fe x O y WOCs provides a good system for studying the role of the interface formed at the SC/WOC junction. The Ni 1 - x Fe x O y - coated hematite electrodes were studied using electrochemical 191 impedance spectroscopy (EIS) and intensity modulated photocurrent spectroscopy (IMPS) as a function of catalyst composition to determine the factors controlling the performance of the modified electrodes. These combined results provide new insight s on the SC/WOC junction and allow for the development of more efficient solar water splitting systems. 5.3. Experimental 5.3.1. Electrode preparation Thin films of hema tite electrodes were prepared on F :SnO 2 (FTO) - coated aluminoborosilicate via electrodeposition ( ED) and atomic layer deposition (ALD) methods reported previously. 2,31 Prior to deposition, the FTO glass was cleaned by ~ 15 minutes sequential sonication i n soap ( ) , water, and isopropyl alcohol and dried in a gentle stream of nitrogen. The ED of planar thin films were performed in 0.1 M FeCl 2 ·4H 2 4.2) at 60 °C by applying 1.2 V vs. Ag/AgCl reference electrode under gentle stirring for 30 minutes. Subsequently, the amorphous electrodeposited FeOOH films were converted to crystalline hematite by annealing at 800 °C in a pre - heated furnace for 10 minutes followed by quenching to the room temperature by taking it out of the furnace . Analogous thicknesses of hematite were deposited via ALD on a ~ 2 nm Ga 2 O 3 underlayer. The Ga 2 O 3 was deposited using tris - ( dimethylamido )gallium(III) (Ga 2 (NMe 2 ) 6 ) (Strem Chemicals Inc.) as the Ga precursor and H 2 O as the oxidant using a modif ied version of a previously reported procedure. 32 The Ga cylinder was heated to 150 °C and pulsed for 0.2 s under exposure mode for 8 s, followed by 12 s purge . Then, a 0.015 s pulse of H 2 O was introduced under the same exposure - purge time to oxidize the Ga precursor. A growth rate of ~ 1.1 Å Ga 2 O 3 /cycle was measured by spectroscopic ellipsometry (Horiba Jobin Yvon, Smart - SE) on control Si wafers. The Ga 2 O 3 coated 192 FTO substrates were subsequently coated with the ~ 30 nm of Fe 2 O 3 by alternating pulses of ferrocene as iron precursor and a combination of water and ozone as the oxidant. The ferrocene cylinder, heated to 70 °C, was pulsed for 20 s which was followed by an oxidation cycle that included 10 subcycles of a 0.015 s H 2 O pulse followed by a 2 s ozone pulse, where each subcycle was separated by a 5 s purge . In agreement with our previous report, the growth rate of Fe 2 O 3 is 0.55 Å /cycles. 4 After deposition, the ALD Fe 2 O 3 films were annealed at 500 °C for 2 hours with heating/cooling rate of 17 °C/min. Finally, the electrodes were annealed in a preheated furnace at 800 °C for 4 minutes. 5.3.2. Deposition of Catalyst Thin films of Ni 1 - x Fe x O y were dep osited by spin coating on 1 cm 2 FTO and hematite substrates using a modified version of previously reported procedures. 21,33 Prior to the deposition of the catalyst on FTO, the substrates were sequentially cleaned by soa p ( Detergent ) , DI water, and IPA for 15 min. In order to coat the hematite electrodes with catalyst, the freshly prepared hematite electrodes were only rinsed with DI water followed by drying in a stream of N 2 . Precursor solutions were prepared from iron (III) 2 - ethylhexanoate (50% w/w in mineral spirits, Strem Chemicals), and nickel (II) 2 - ethylhexanoate (78% w/w i n 2 - ethylhexanoic acid, Strem Chemicals) by dissolving the appropriate amount of metal precursor in hexanes ( Table A 5 - 1 ) to give a total concentration of 15% w/w metal complex. These solutions were further diluted with hexane to prepare a solution with a total metal concentration of 50 mM. Approximately 0.25 mL of the solution was added to the substrate, followed by spinning at 5000 rpm for 60 s. Subsequently, the films were irradiated with UV light (254 nm, 4 W) for 1 hour followed by annealing in a preheated furnace in air at 100 °C for 1 hour. We used Raman 193 Spectroscopy to e nsure that the decomposition of organic precursor was completed ( Figure A 5 - 1 ). In order to change the thickness of the catalyst, the deposition procedure was performed 1, 2 or 3 times sequentially to produce three thicknesses of catalysts on the ED hematite electrodes. Note that each deposition cycle involved 1 hr UV light treatment fol lowed by 1 hr annealing at 100 °C. Atomic Force Microscopy (AFM) and Spectroscopic Ellipsometry (ES) were utilized to determine the thickness of the catalyst deposited on silicon substrates using comparable procedures. A s used for ES measurements since the oxides are transparent which allows for more accurate determination of thickness. 34 5.3.3. Film Characterization X - ray photoelectron spectroscopy (XPS) was utilized to determine the composition of the thin films of Ni 1 - x Fe x O y catalyst on FTO. The XPS data were collected utilizing Perkin Elmer Phi 5600 ESCA system equippe angle of 45°. Survey scans of 0 - 1000 eV binding energy and detailed scans for C 1s, O 1s, Ni 2p and Fe 2p, regions were measured for all samples. The binding energies were corrected in reference to C 1s peak (284.8 eV) and Shirley background subtraction was performed for fitting for each sample. For fitting the Ni and Fe 2p regions, the p 1/2 and p 3/2 peaks were bonded together with the relative area of 1 to 2. Raman spectra were col lected via Raman Microprobe (Renishaw) equipped with a 45W Cobalt DPSS laser (532 nm line) laser and a 100 × magnefication objective to focus the laser on the film surface. The surface morphology of the prepared films were examined by scanning electron mic roscopy, SEM (Carl Zeiss Auriga, Dual Column FIB - SEM ). 194 5.3.4. (Photo)electrochemical Measurements Electrochemical measurements were made with an Eco Chemie Autolab potentiostat coupled with Nova electrochemical software . The electrochemical characterization of Ni 1 - x Fe x O y on FTO was carried out by cycling the potential linearly between 1.0 to 2.0 V vs. RHE at a scan rate of 100 mV s - 1 . The electrodes were examined in contact with 1 M KOH solution (pH = 13.6) as was determined with Fisher Scientific Accument pH me ter. Each electrode was activated by cycling the potential between 1 - 2 V vs. RHE (40 CVs). All the presented CVs are measured after activation. The potential was corrected for IR drop, using R values determined by impedance measurements. A homemade satur ated Ag/AgCl and a platinum mesh electrodes were used as the reference and counter electrodes, respectively. The reference electrode was frequently calibrated against commercial saturated calomel electrode (Koslow Scientific). All the potentials were conve rted to reversible hydrogen electrode, RHE, by equation . The electrochemical impedance spectroscopic and photoelectrochemical measurements were made with an Eco Chemie Autolab potentiostat coupled with Nova electroc hemical software. Impedance data were gathered using a 10 mV amplitude perturbation of between 10,000 and 0.01 Hz. Data were fit using Zview software (Scribner Associates). For each thickness of catalyst with different compositions two electrodes were prep ared and the fitted parameters were averaged and the standard deviations were reported on the graphs. The current density vs. applied potential ( J - E ) response of bare and modified hematite electrodes were examined by cycling the potential between - 0.5 and 0.7 V vs. Ag/AgCl at the scan rate of 20 mV s - 1 in contact with 1 M KOH. The light source was a 450 W Xe arc lamp (Horiba Jobin Yvon). An AM 1.5 solar filter was used to 195 simulate sunlight at 100 mW cm 2 (1 sun). Unless otherwise stated, all photoelectroche mical tests were carried out by shining light on the electrodes through the electrolyte. The performance of each electrode was initially tested by measuring the J - E in light and dark. Subsequently, each electrode was conditioned by cycling between 0.5 - 1.7 V vs. RHE under illumination, followed by another J - E measurement. Each electrode was then subjected to EIS ( ~ 2 h under illumination) and IMPS ( ~ 1 h under illumination) measurements, with additional J - E measurements after EIS and IMPS measurements to determine any change in behavior. Negligible changes were observed between initial (after conditioning / before EIS) and final J - E curves for a given electrode. The J - E data shown are from the final measurem ents. IMPS measurements were made with an Eco Chemie Autolab potentiostat equipped with metrohm LED driver accessory which was controlled by Nova electrochemical software. The measurements were performed in 1 M KOH utilizing a three - electrode configuration . A 470 nm L ED li ght was utilized as the light source for all the experiment. The power of the LED was chosen in a way that it provides the same number of photocurrent as was measured under 1 sun illumination. The modulation intensity was set at 10% of the power of the LED to ensure a linear respond and the frequency was swept from 15 kHz to 0.1 Hz with 10 frequency per decade increments. The IMPS data were collected over the potential range of 0.5 to 1.5 V vs. RHE with 0.1 V interval. 5.4. Results and Discussio n XPS measurements were performed on five catalysts prepared with variable composition of the Ni 1 - x Fe x O y solution (x = 0, 0.25, 0.5, 0.75, and 1) on FTO substrates. The compositions of the films determined by XPS, summarized in Table A 5 - 2 , are in good agreement with the composition of the solutions used for spin coating. The detailed XPS spectra as a function of the composition of 196 the films are shown in Figure A 5 - 2 , followed by interpretation of the spectral features. Depending on the ratio of Ni to Fe, Ni 1 - x Fe x O y may consist of one or more separate phas es. F or iron concentration up to 25% only one phase, presumably Fe - doped Ni(OH) 2 , is discernable. At higher iron content, however, two phases appear to be present, which are assigned to FeOOH and Fe - doped Ni(OH) 2 , i n accord with recent studies on the electrocatalytic activity of Ni 1 - x Fe x O y on different substrates. 35 This observation suggests that by tuning the composition of the catalyst, the interface of the electrocatalyst and underlying substrate may be modified by contact with the different catalyst phases. The cyclic vol tammograms (CV) of the Ni 1 - x Fe x O y catalysts on FTO in contact with 1 M KOH are shown in Figure 5 1 .a . The CV of Ni(OH) 2 exhibits two characteristic fe atures: a redox wave which is attributed to the transition between Ni(OH) 2 and NiOOH, and an anodic wave at more positive potentials corresponding to water oxidation. 29,36,37 The CV of FeOOH, on the other hand, only shows an anodic wave associated with water oxidation. 29 Similar to the Ni(OH) 2 catalyst, the CV of the Ni 1 - x Fe x O y catalysts shows two features: one redox wave for the transition from Ni(OH) 2 to NiOOH and an anodic wave due to water oxidation. Consistent with previous studies, the position and the area of the redox peaks are strongly correlated to the composition of the film. 27,37 As the Fe content increases the redox peak shifts to more positive potentials; at iron concentr ations beyond 25% the redox peak was merged into the water oxidation wave and is no longer observable. 29,36 35 In order to compare the electrocatalytic activity of the Ni 1 - x Fe x O y catalysts, the current density at 300 mV overpotentia l and the overpotential required to produce 10 mA cm - 2 were determined. A plot of these two parameters as a function of Ni (content determined by the XPS) are shown in Figure 5 1 . b. The maximum current density at 300 mV overpotential and the minimum overpotential required to achieve 10 mA cm - 2 , i.e. the highest activity for oxygen 197 evolution reaction (OER), were observed for the film with 75 - 50 % nick el content. We note that the composition of the most active catalyst is comparable to previous reports where thin films of Ni - Fe catalysts were prepared by different methods on different substrates. 29,35,38 The com position of the catalysts after electrochemical conditioning were studied via XPS ( Table A 5 - 3 ). With the exception of the NiO y catalyst, we did not observe a change in the composition of the catalysts after electrochemical conditioning. Iron was electrochemically intercalated into the NiO y films, however, in agreement with prior studies. 29,37 Raman spectroscopy was used to investigate a possible phase change of the catalysts during conditioning, however the catalysts prepared by this method are thin and amorphous and show no distinguishable peaks i n agreements with other reports. 33,39,40 Figure 5 1 . a) IR - corrected CVs of Ni 1 - x Fe x O y catalysts on FTO substrates; FeO y (cyan), Ni 0.25 Fe 0.75 Oy (orange), Ni 0.5 Fe 0.5 O y (pink), Ni 0.75 Fe 0.25 O y (green), and NiO y (blue). b) plots of current density at 300 mV overpotential (blue diamonds) and overpotential at 10 mA cm - 2 (red triangle) as a function of the % Ni in contact with 1 M KOH; the compositions, activities, and the 198 error bars are the average and standard deviations acquired from 3 independently prepared samples. The influence of the electrocatalyst composition on the performance when integrated with hematite electrodes was further investigated. The hematite electrodes were prepared via ED and ALD . As shown by Raman and SEM the morphology and crystallinity of the films remained unaffected after modification with catalyst ( Figure A 5 - 3 and Figure A 5 - 4 ). These observations indicate that the catalyst coating with this met hod is solely a surface modification with no alteration of the bulk characteristic of hematite. The current density ( J ) vs. applied potential ( E ) responses in the dark and under illumination of bare ED and ALD hematite electrodes, as well as ED and ALD ele ctrodes modified with the five compositions of the Ni 1 - x Fe x O y catalysts are shown in Figure A 5 - 5 . Though no obvious change in the PEC water oxidation performance of the FeO y coated electrodes were observed, the performance of both ED and ALD electrodes modified with NiO y were substantially diminished. In addition, it was noted that for the ALD electrodes all three catalysts, i.e. Ni25, Ni50, and Ni75, r esulted in 200 mV cathodic shift of the photocurrent onset potentials and thus improved performance. For the ED electrodes, on the other hand, only Ni25 improved the performance while Ni50 and Ni75, similar to the NiO y , diminished the photocurrent response of the ED hematite electrode. To determine the underlying cause the different category behavior, two representative catalyst compositions we selected for more in - depth investigations described below: Ni 0.25 Fe 0.75 O y (Ni25) as the Fe - rich catalyst, and Ni 0.75 F 0.25 O y (Ni75) as the Ni - rich catalyst. To ensure the reproducibility of the behavior of the modified electrodes described below, J - E measurements for several additional batches of electrodes prepared independently by different people on differen t 199 days were performed. The J - E curves for catalyst coated hematite electrodes are shown in Figure A 5 - 6 and Figure A 5 - 7 . The PEC behavior of the bare ALD electrodes is very consistent, however there is some variability of the behavior of the bare ED electrodes. This is not too surprising given the di fferent preparation methods. In both cases, however, the trend of performance and the effect of the different catalyst compositions is reproducible. The dark J - E curves Ni25 and Ni75 catalysts, are compared in Fig ure A 5 - 8 . The dark current behavior follows the same trends in activity as these catalysts on FTO, with increasing activity in the order Ni75 > Ni25 > bare. This result suggests that for the thin film of catalysts the electrocatalytic activates are nominally independent of the hematite substrate. Representative J - E curves under illumination for the bare and modified electrodes are shown in Figure 5 2 . Consistent with our prior report, the bare ED electrode somewhat outperformed the bare ALD electrode. 2 The effect of the different catalysts on the performance of the ED and ALD hematite electrodes is dramatically different, however. For the ALD hematite electrodes ( Figure 5 2 . a), both catalyst compositions improved the performance. This result agrees with our prior studies that showed surface modification of ALD hematite electrodes by coating with Co - Pi or NiO y enhanced the overall performance of the electrode by increasing charge separation from hematite to the catalyst at the surface, thereby suppressing surface recombination rates. 9,41 This behavior is generally consistent with the formation of an adaptive junction at SC/WOC interfa ce. 9,17 In the case of Ni75 coated ALD electrod e a capacitive peak prior to the water oxidation wave is clearly observable. This peak is generally observed for Ni(OH) 2 coated semiconductor electrodes, i.e. Fe 2 O 3 or TiO 2 , under PEC water oxidation conditions and it is attributed to the Ni 3+ /Ni 2+ redox reaction. 6,17 For the ED hematite electrodes shown in Figure 5 2 . b, on the other hand, the J - E responses are strongly dependent on the composition of the catalyst. For the Ni25 coated ED electrodes, a ~ 300 200 mV negative shift of the photocurrent onset potential is produced. In addit ion, the photocurrent density and fill factors are significantly improved compared to the bare electrode. Conversely, for the Ni75 coated ED electrodes, the performance is suppressed. This behavior contrasts the electrocatalytic activity, where Ni75 outper forms Ni25. In addition to the diminished performance, the J - E curve with Ni75 exhibits two well - separated capacitive features; a redox peak located at ~ 0.7 V vs. RHE, similar to the ALD electrodes, and a second peak at ~ 1.35 V vs. RHE. The first capacit ive wave at ~ 0.7 V vs. RHE is assigned to charging of the catalyst film, which is generally seen for Ni - rich Ni 1 - x Fe x O y catalysts and is attributed to oxidation of Ni 2+ to Ni 3+ . 6 Figure A 5 - 9 shows an increase in the magnitude of this wave in response to increasing incident light intensities from 0.1 sun to 3.5 suns, which further supports this assignment. Figure 5 2 . J - E curves measured at 20 mV s - 1 under 1 sun illumination for bare and modified a) AL D and b) ED hematite electrodes in contact with 1 M KOH solution; bare ALD (red), bare ED (blue), Ni75 coated hematite (green), Ni25 coated hematite (orange). The dark current are not included for the sake of cl arity and are shown in Fig ure A 5 - 8 . 201 Because the dark J - E curve of Ni75 on the ED electrodes ( Fig ure A 5 - 8 ) and the light J - E curve of the bare ED electrodes both show very good activity, whereas the combination suppresses the performance, we hypothesize that deleterious interface states develop between the ED hematite and Ni75 catalyst films. Interestingly, since the ALD hem atite electrode modified with the Ni75 catalyst exhibits good performance, the proposed development of Ni75/hematite interface states is not general, but appears to be a function of the particular hematite surface that is modified with the Ni75 catalyst. A n additional control system was fabricated using a NiO y catalyst, without the intentional incorporation of Fe, deposited on an ED electrode. The J - E curve of this system showed a redox wave with similar magnitude and potential as the ED/Ni75, displayed in Figure A 5 - 10 . The light J - E behavior is suppressed to a similar extent as the ED/Ni75 system. From these results, we deduce that the deleterious inter face is due to Ni bonding with the ED hematite surface. We further deduce that the peak observed at ~ 1.35 V vs. RHE in Figure 5 2 . b for the ED/Ni75 system is due to the ED/NiO y interface states. We therefore focus our attention below on the ED systems. EIS measurements were performed for bare ED as well as electrodes modified with Ni25 and Ni75 catalysts. Examples of Nyquist plots for ED electrodes me asured under illumination at 1, 1.2 and 1.4 V vs. RHE can be seen in Figure A 5 - 11 . At this potential range two semicircles are clearly visible. The e quivalent circuits used to fit the EIS data have been established previo usly and are shown in Figure A 5 - 12 . 9 Plots of C bulk , as well as Mott - Schottky (MS) plots p repared from C bulk , are shown in Figure A 5 - 13 and Figure 5 3 . a, respectively. The C bulk behavior was found to be invariant with re spect to catalyst under illumination. Dopant densities of 2.40 (± 0.14) × 10 20 cm - 3 , 2.75 (± 0.10) × 10 20 cm - 3 , and 2.76 (± 0.12) × 10 20 cm - 3 were determined for ED/Ni75, ED/Ni25, and bare ED electrodes from fitting the capacitance data to the MS equation 42 using a dielectric constant of 32 for hematite and 202 the geometric surface area of the electrodes. 43 This large dopant density is in good agreement with our previous report of electrodeposited hematite electrodes. 2 The small deviation in dopant density cannot account for the observed behavior for the catalyst coated electrodes. We note that thi s high dopant density precludes accurate determination of the flat band potential, 44,45 however the similar behavior of all electrodes indicates that the catalyst does not shift the band positions or affect band bending by pinning the hematite electron Fermi level. This conclusion is further supported by IMPS results discussed below. Figure 5 3 . a) MS plots for bare and catalyst modified ED electrodes, b) J - E curve and C ss values for bare ED electrode under 1 sun illumination in contact with 1 M KOH; bare electrode (blue diamond), Ni75 modified ED electrode (green triangle), and Ni25 coated ED electrode (orange pentagon). The dotted lines are the linear fitting used to extr act the dopant density. The low - frequency capacitance is assigned to surface states ( C ss ) for the bare and the catalyst layers ( C cat ) for the modified electrodes. The C ss , displayed in Figure 5 3 .b , has two peaks. One peak is coincident with the water oxidation onset potential at ~ 0.8 which we assign to the water oxidation 203 intermediate species building up on the electrode surface, and a second peak located at more positive potential, ~ 1.3 V vs. RHE, which we tentatively assign to a surface defect state. 19,46 As shown in Figure 5 3 . b, there is an apparent inflection of the J - E curve at this potential which suggests that recombination at this state mitigates the water oxidation efficie ncy until sufficie nt potential is applied to deplete the trap of electrons. We note that the presence of two surface states following high - temperature annealing procedure is a key difference between the ED and ALD electrodes. 19,20 We therefore hypothesize that this second state is responsible for the dramatically different behavior of catalys ts on the ED vs. ALD hematite electrodes. In order to confirm our assignment and understand the effect of the catalysts, the thickness was varied by repeating the deposition procedure one or two additional times to make the series denoted 1, 2, and 3 cycle s of catalyst. Atomic Force Microscopy (AFM) and Spectroscopic Ellipsometry (ES) were utilized to determine the thickness of the catalyst deposited on silicon substrates using comparable procedures. The thickness of the catalyst on hematite electrodes were estimated by integration of the cathodic peaks of the dark J - E ( Table A 5 - 4 ). For Ni75 catalyst an approximately linear growth as a function of the number of cycles was measured. Consistent with the experimental values the calculated thickness of the catalyst on hema tite was also found to linearly increase with the number of cycles. In addition, the estimated thicknesses via charge integration of modified hematite electrodes were fairly close to the experimentally determined thicknesses of catalyst on silicon substrat es by AFM/ES. Any discrepancy in results can be attributed to either the substrate dependence of the growth of the Ni75 catalyst or the possibility that not all the Ni sites are redox active. For Ni25 catalyst, however, the growth mode of the catalyst was found to be non - linear and the measured thicknesses were generally lo wer than Ni75 ( Table A 5 - 5 ). The light and dark J - E curves of the catalyst coated ED electrodes with varying thickness of catalyst 204 are shown in Figure A 5 - 14 . We note at a given potential the photocurrent was diminished by increasing the thickness of the catalyst. This can be understood by considering the illumination direction; the PEC measurements were all performed by illumination from the electrolyte side, thus a portion of the light is lost by competitive light absorption with th e catalyst layer. This observation is important since it supports the increasing thickness of catalyst. The parameters derived from fitting the impedance spectra as a function of the catalyst thickne ss are shown in Figure A 5 - 15 to Figure A 5 - 18 . The C cat as a function of potential for ED hematite electrodes coated with different thicknesses of Ni25 is compared in Figure 5 4 . a. Also shown is the C cat v s. potential of Ni25 measured on FTO. Boettcher and coworkers previously studied the electrocatalytic activity of Ni 1 - x Fe x O y as a function of the catalyst loading. 47 It was shown that at the potential range where the catalyst is fully oxidized the C cat was increased with the catalyst loading. As it can be seen from Figure 5 4 . a, the C cat increases with increasing the thickness of catalyst which further indicates that at this condition Ni25 is fully oxidized. The peak in C cat is shifted negatively by 0.8 V co mpared to Ni25 on FTO substrate , indicating 0.8 V of photovoltage is produced by the hematite. Interestingly, this is the exact same photovoltage determined by Wang and coworkers for hem atite coated with a NiFeO x catalyst via open circuit potential measurements. 48 Once the potential of the catalyst drops to more positive potentials via oxidation by photogenerated holes in the hematite, represented by the ca pacitance, the photocurrent density is controlled by the flux of holes reaching the interface. This behavior is generally consistent with the adaptive junction model, and is further supported by IMPS results described below. We note that the magnitude of t he C cat for 1 cycle Ni25 on FTO is about 4 times larger than 1 cycle Ni25 on ED electrode, suggesting that the catalyst on FTO is thicker. As the thickness of the catalyst increase, however, the C cat (ED) increases and becomes 205 almost equal to the C cat (FTO) , which indicates that the growth of the catalyst is substrate dependent, and that the catalyst is fully oxidized. Figure 5 4 . b shows C cat for different nominal thickness of Ni75 on hematite as well as Ni75 on FTO. We note that the peak of C cat for Ni75 on FTO is about three times larger than Ni25 on FTO, which is a ttributed to the expected three times larger charge density, which is on the Ni atoms, for a given thickness. The plots of C cat vs. potential for Ni75 exhibits three features: (1) similar to the Ni25, there is an onset of C cat at approximately 0.6 V vs. RH E, however instead of reaching a peak, the C cat plateaus over an additional 0.6 V applied potential, (2) where it increases to values consistent with Ni75 on FTO, and (3) the C cat was found to be independent to the thickness of the catalyst at the potentia l range of 0.7 - 1.4 V vs. RHE. Figure 5 4 . a) C Ni25 and b) C Ni75 obtained from fitting EIS for catalyst coated FTO (1cycle, shown with star symbols) and ED electrodes with varying thickness: 1 cycle (pentagon), 2 cycles (circle), and 3 cycles (triangle). 206 This behavior is consistent with the potential of the catalyst be ing pinned. Since the bulk capacitance, and thus band bending, band positions and hole flux from hematite, is constant for all ED electrodes measured, the catalyst potential must be controlled by recombination to interface (or catalyst) states. Once these states are emptied, for example by controlling the electron Fermi level (applied potential), the photogenerated holes should further oxidize the catalyst, resulting in an abrupt increase in C cat and current. Above we assigned the peak observed at ~ 1.35 V vs. RHE in the J - E curves for the ED/Ni75 system to the interface states, which is coincident with the sudden increase in C cat . This is also approximately the same potential as the second surface state observed for the bare ED electrode displayed in Figure 5 2 . b. Thus, we propose that the interface state results from bonding interactions between the ED surface state and Ni hydroxide species. Further und erstanding of the effect of catalyst can be gleaned by comparing the trap (recombination) resistance to surface states for the bare, R trap , or catalyst layer for modified electrodes, R cat , and the charge transfer resistance at the respective electrochemical interface, R ct , ( Figure A 5 - 17 and Figure A 5 - 18 ). Bisquert and coworkers 49 showed that the R trap / R ct for bare electrodes, which corresponds to R cat / R ct for catalyst coated electrodes, is proportional to the ratio of charge transfer and recombination rate constants. Figure 5 5 shows the ratio of the R trap / R ct and R cat / R ct for bare and catalyst modified ED electrodes, respectively. There is a strong correlation between the ratios of resistances to the J - E curves displayed in Figure 5 2 . b . For the Ni25 coated electrode, improvement in photocurrent and onset potential can be ascribed to an acceleration of charge transfer at the hematite surface compared to the bare electrode. In other words, oxidation of the Ni25 catalyst outcompetes interface recombination to a greater extent than water oxidation competes with surface state recombination at a bare hematite electrode. This result is consistent with our prior report on CoO x coated hematite electrodes. 50 For Ni75 coated electrodes, on the 207 other hand, the R cat /R ct ratio is low and essentially constant, which indicates that fast recombination at the SC/WOC interface inhibits the photocurrent - voltage behavior. Figure 5 5 . R trap / R ct for bare and R cat / R ct for catalyst modified ED electrodes; bare electrode (blue diamond), Ni75 modified ED electrode (green triangle), and Ni25 coated ED electrode (orange pentagon). IMPS measurements were also performed to interrogate the relevant charge dynamics at the surfa ce of the electrode which give rise to the large difference in effects of the Ni25 and Ni75 catalysts. A set of normalized Nyquist plots for bare and modified ED electrodes at 1.3 V vs. RHE are shown in Figure S19. Ideally the low - frequency intercept (LF) with the real axis is equal to the steady - state photocurrent, which is equal to the rate at which the holes are transferred to the catalysts and from catalyst to solution. At sufficiently high frequencies (HF) the surface recombination is p roposed to be frozen out and the HF intercept with real axis thus provides a measure of hole flux toward the SC/electrolyte (or SC/WOC) junction. 51,52 Plots of the LF 208 photocurrent for bare and catalyst coated ED electrodes are compared in Figure 5 6 . a. The LF plots mimic the behavior of the J - E plots, consistent with the assignment of the LF as the steady - state photocurrent. The HF photocurrent is shown in Figure 5 6 .b . For the Ni 25 modified electrodes, the HF response is nominally identical to that of the bare electrode at all potentials. This indicates that after surface modification with Ni25, the flux of holes to the surface are constant. This is consistent with the Mott - Schott ky plots which show that the energetics at the surface (band edges and band bending) that control the hole flux reaching the interface remains essentially constant. For the Ni75 coated .0 V vs. RHE. At higher potentials, however, the HF response plateaus. This means that either the surface hole flux is constant over a relatively large range of applied potentials, from 1 .0 to 1.5 V vs. RHE, or the assignment of the HF is wrong. The validity of the assumption that surface recombination is frozen out at high frequencies obviously depends on the rate of recombination. Just like the bare and Ni25 modified electrodes, the Mott - Schottky plots suggest an increasing hole flux due to increasing band bending over this potential range. The R cat / R ct ratio derived from EIS measurements for the Ni75 electrodes described above suggest fast interfacial recombination, which likewise can explain the HF re sults. Thus, we assign the plateau region of the HF of the IMPS results to an essentially constant, fast 209 Figure 5 6 . a) variation of low f requency (LF), and b) high frequency (HF) limits of the IMPS response for bare and catalyst coated ED electrodes under monochromatic illumination (470 nm) in contact with 1 M KOH solution; bare electrode (blue diamond), Ni75 modified ED electrode (green tr iangle), and Ni25 coated ED electrode (orange pentagon). 5.5. Conclusion The electrocatalytic activity of Ni 1 - x Fe x O y catalysts on FTO substrates and on hematite electrodes in the dark is primarily controlled by the composition of the catalyst, where the nickel - rich Ni75 catalyst is the most active for electrochemical water oxidation. Under illumination, however, the PEC performance of modified hematite photoanodes (prepared by ALD and ED) were found to be dependent both on the substrate as well as the compositio n of the catalyst. Several unique and interesting specific differences were clearly observed. The PEC performance of the modified ALD electrodes improved upon modification with all catalyst compositions, indicated by the ~ 200 mV cathodic shift in photocur rent onset potential. The modified ED electrodes behaved very differently from the modified ALD electrodes for a given catalyst composition. Regardless of the hematite electrode preparation, however, the iron - rich Ni25 catalyst produces superior 210 performanc e compared to Ni75, despite the inferior electrocatalytic activity. These results suggest the overall PEC performance is strongly influenced by the interface that can develop upon contacting semiconductors with electrocatalysts. The most remarkable systems were the modified ED hematite electrodes, where the Ni25 catalyst produced a very large improvement in PEC water oxidation whereas the Ni75 catalyst produced diminished performance. The large improvement of the Ni25 modified ED electrode can be understood by the presence of a surface state defect which influences the J - E behavior of the bare electrode. The a ddition of the Ni25 catalyst apparently either passivates this state or just opens a more favorable path of hole transfer to the catalyst which circumv ents it. This is supported by the catalyst capacitance which reaches a maximum coincident with the water oxidation onset potential. The capacitance increases with increasing catalyst thickness and illumination intensity, showing it originates from oxidatio n of the catalyst throughout the entire thickness from photogenerated holes in the hematite. The ratios of R cat / R ct from EIS measurements and HF/LF from IMPS further suggest the improvement in performance derives from an increase in charge separation and r eduction in recombination at the hematite surface. These results are generally consistent with modified ALD electrodes, 6,9 and other reported hematite electrode catalyst combinations 48,53 which c an be generally interpreted via the adaptive junction model. 18 The most striking system is Ni75 modified ED hematite electrodes. Independently, they are the best WOC and photoanode examined, however the combination results in the worst PEC behavior of all systems investigated. The decreased performance coincides with the appearance of an additional capacitive wave. This capacitance is initially ascribed to an ED h ematite - NiO y interface state that acts as a recombination center. This is supported by the ratio of R cat / R ct from EIS measurements which indicates fast recombination, and the unusual leveling off of the HF intercept 211 of IMPS measurements. The capacitance of the catalyst also remains relatively low and constant over a large potential region from the nominal Ni 3+/2+ potential to the interface state potential indicating the catalyst potential is pinned by the interface state. Thus, even a soft interface tha t is expected from the deposition of an amorphous metal oxide WOC on hematite under mild conditions can produce states which control the behavior of the junction. Finally, it is interesting to note that there are prior reports that the PEC water oxidation is influenced by surface or interface states, which can be perhaps better understood following the insight provided here. 19,20 Wang and coworkers 21 , showed that the surface modification of hematite through a re - growth procedure followed by NiFeO x catalyst coating resulted in a substantial improvement of the photoanode performance. This result is analogous to what we observe with the ED/Ni25 junction. Li and co - workers studied Ni(OH) 2 coated Fe 2 O 3 nanowires. 54 They observed that coating the hematite electrode with Ni(OH) 2 , resulted in an initial cathodic shift of the photocurrent density; however, the photocurrent decayed by 90% to values lower than the bare hematite within 30 s. This behavior was attributed to the fast oxidation of Ni 2+ to Ni 3+ , followed by a rate - limiting step of further oxidation to a Ni 4+ species, thus, resulting in the NiOOH film storing charge but not producing a sustained enhancement of water oxidation. This result can be understood by the NiOOH catalyst being pinned, thus not achieving a potential sufficient to sustain water oxidation, consistent with what we observe with the ED/Ni75 (and NiO y ) system. In another system, Choi and coworkers 55 showed that the PEC characteristic of BiVO 4 is substantially improved when the interface between the semiconductor and NiOOH, was modified with FeOOH. This result is consistent with our findings that the Ni25, which consists of separate FeOOH and Fe - doped NiOOH phases, passivates the ED hematite in terface. This examples emphasize the importance of the SC/WOC interface, where the contact of a Ni - rich phase with otherwise 212 promising photoanode materials may produce a large density of deleterious states which act as recombination centers, whereas contac t with Fe - rich phases may produce a . Clearly more research is needed to understand such phenomena in more detail to develop more general mode ls of this important interface. In follow - up n), the PEC characteristics of the bare and catalyst coated ED - hematite electrode were studied. A combination of electrochemical and spectroscopic characterization including cross - section TEM and dual working measurement were utilized to further assess the relationship between the PEC performance and the composition of the catalyst. 56 The two key observations of this study are: (1) Ni75 is electrically more conductive than Ni25 by 5 order of magnitudes and (2) the electrodeposited hematite electrodes are porous with presence of pin - holes. On the bases of these observations , it is concluded that the performance of the Ni75 coated ED - hematite is suppressed by the fast recombination of the stored holes in catalyst with the free electrons from the conductive substrate (FTO). In another word the potential of the catalyst is pinn ed to the conductive substrate. 213 A P PENDIX 214 Composition Fe (mg) Ni (mg) Hexane (mg) NiO y 0 20 90 Ni 0.75 Fe 0.25 O y 103 141 920 Ni 0.5 0 Fe 0.5 0 O y 44 20 212 Ni 0.25 Fe 0.75 O y 411 63 1440 FeO y 50 0 142 Table A 5 - 1 . The weight of metal precursor used to prepare the catalyst solutions. Composition Theoretical Found % Ni % Fe % Ni % Fe NiO y 100 0 100 0 Ni 0.75 Fe 0.25 O y 75 25 73 ± 6 27 ± 6 Ni 0.5 0 Fe 0.5 0 O y 50 50 55 ± 4 45 ± 4 Ni 0.25 Fe 0.75 O y 25 75 28 ± 5 72 ± 5 FeO y 0 100 0 100 Table A 5 - 2 . Composition of as - prepared Ni 1 - x Fe x O y catalysts found by XPS, the compositions and error bars are the average and standard deviations acquired from 3 independently prepared samples. 215 Composition Before After Theoretical Found % Ni % Fe % Ni % Fe % Ni % Fe NiO y 100 0 100 0 80 20 Ni 0.75 Fe 0.25 O y 75 25 71 29 70 30 Ni 0.5 0 Fe 0.5 0 O y 50 50 52 48 49 51 Ni 0.25 Fe 0.75 O y 25 75 28 72 29 71 FeO y 0 100 0 100 0 100 Table A 5 - 3 . Comparison between catalyst composition before and after electrochemical conditioning on the same electrode. Note: s amples were conditioned by cycling the potential at 1.0 2.0 V vs . RHE for 40 cycles as well as electrochemical characterization by impedan ce at potential range of 1.3 - 1.8 V vs . RHE with 100 mV interval. Afterward the electrodes were rinsed with extensive amount of DI water. 216 Figure A 5 - 1 . Raman spectra of thin films of Ni 1 - x Fe x O y with diff erent composition before and after irradiation with UV - light for 1h. The loss of the bands after UV - irradiation indicates that ligands are decomposed . 217 Figure A 5 - 2 . The detailed XPS spectr a of the Ni 1 - x Fe x O y catalyst on FTO. Note: The C 1s peaks are presented to show that the shift caused by charging the sample during the measurement are correctly accounted for. For FeO y film, two peaks in O1s region (530 and 531.7 eV) and one group of peaks in Fe2p r egion (711 and 724 eV) were observed. This is in agreement with the formation of FeOOH phase. 35 For NiO y , on the other hand, only one O 1s peak (531.7 eV) and one group of Ni 2p peaks (850 and 860 eV) were observed which is consistent with formation of Ni(OH) 2 phase. 35 Interestingly, upon 218 doping Ni(OH) 2 with iron, at the maximum concentration of 25%, t h e peak position of th e Fe 2p and Ni 2p were shifted to the lower binding energies and only one O1s peak ( ~ 531.7 eV) was observed. However, at concentrations beyond 25%, a new O 1s peak emerged which is consistent with the formation of FeOOH. Consistent with the oxygen peak, i n Fe 2p region, up to 25% of the iron, there is only one group of Fe 2p peak. At higher concentration of iron, however, a new group of Fe 2p peaks emerged which were located at the same position as the Fe 2p peaks for FeOOH. In addition, in the Ni 2p region a new group of Ni 2p peaks emerged which was independent to the composition of the films. Therefore, the presence of one O 1s peak and one group of Fe 2p peak for the iron content up to 25% indicates that iron is in fact doped into the nickel phase which formed a single phase of Ni 1 - x Fe x O y . On the other hand, at the higher concentration of iron two distinctive phases of FeOOH and Ni 1 - x Fe x O y were formed. This finding is consistent with previous reports on the electrocatalytic activity of Ni 1 - x Fe x O y on different substrates. 35,37 219 Figure A 5 - 3 . The SEM images of bare and catalyst coated ALD (a - c), ED (d - e) electrodes. The first column represents the bare electrodes, the 2 nd and 3 rd columns represent the electro des coated with Ni 0.25 Fe 0.75 O y and Ni 0.75 Fe 0.25 O y , respectively. The scale bar is 500 nm with a magnification of 50K. 220 Figure A 5 - 4 . The Raman spectrum of a) bare hematite, b) bare and catalyst modified ALD, c) bare and catalyst modified ED hematite electrodes. The Raman spectrum of the bare and catalyzed hematite electrodes are shown in Figure A 5 - 4 . All - Fe 2 O 3 . Clearly, the Raman spectr a of hematite films, i.e. ALD and ED, are almost identical which further indicates that regardless of the preparation method the degree of crystallinity and crystal orient ations for both type of hematite films are similar. In addition, after coating hematite electrodes with catalysts, the Raman spectrum of the bare and catalyzed hematite films, in terms of peak position, shape, and patterns, are identical. 221 Figure A 5 - 5 . J - E curves measured in light (a and c) dark ( b and d) for bare and catalyst coated ALD (top), and ED (below) hematite electrodes in contact with 1 M KOH solution. The plots correspond to bare ALD (red ), bare ED (blue), NiO y coated hematite (violet), Ni75 coated hematite (green), Ni50 coated hematite (pink), Ni25 coated hematite (orange), and FeO y coated hematite (cyan). 222 Figure A 5 - 6 . J - E curves measured in light (left plots, solid lines) and dark (right plots, dotted lines) for three different batch of ALD hematite electrodes coated with catalyst in contact with 1 M KOH solution; bare ALD (red), Ni75 coated hematite (green), Ni25 coated hematite (orange). Note: these electrodes are different from the one shown in the main text. 223 Figure A 5 - 7 . J - E curves measured in light (left plots, solid lines) and dark (right plots, dotted lines) for three different batch of ED hematite electrodes coated with catalyst in contact with 1 M KOH solution; bare ED (blue), Ni75 coated hematite (green), Ni25 coated hematite (orange). Note: these electrodes are different from the one shown in the main text . 224 Fig ure A 5 - 8 . J - E curve measured in dark for bare and catalyst coated a) ALD, and b) ED hematite electrodes in contact with 1 M KOH solution. 225 Figure A 5 - 9 . J - E curves measured at 20 mV s - 1 for ED coated electrodes with Ni75 in contact with 1 M KOH solution as a function of illumination intensity. 226 Figure A 5 - 10 . J - E curves measured at 20 mV s - 1 under 1 sun illumination for NiO y and Ni75 coated ED electrodes in contact with 1 M KOH solution. 227 Figure A 5 - 11 . Nyquist plots for bare and catalyst coated ED electrodes measured under illumination at a) 1.0, b) 1.2, and c) 1.4 V vs. RHE. 228 Figure A 5 - 12 . a) Equivalent circuit used for interpretation of bare and c atalyst coated ED electrodes, b) Randle circuit , when two and one semicircle was observed, respectively. 9 229 Figure A 5 - 13 . C bulk values from EIS measurements for bare and catalyst coated ED electrodes under 1sun illumination in contact with 1 M KOH. 230 Number of cycle Integrated charge from CV (mC cm - 2 ) Calculated thickness from integrated charge (nm) a Measured thickness with AFM (n m) Measured thickness with SE (nm) 1 0.8 15.4 17.8 15 2 1.58 30.5 37.4 29 3 2.2 42.4 48.9 42 Table A 5 - 4 . The measured and calculated thickness of the Ni75 on silicon and hematite. a For calculation of the thickness the measured composition found by XPS was used and assumed that only Ni atoms are involved in the redox reaction. The cell volume of 0.2108 nm 3 (ICSD 159700) for Ni - Fe layered double hydroxide and the geometrical surface a rea were used to calculate the thickness of the film. 57 59 For these calculation it was assumed that there is only one Ni per unit cell. Obviously if the number of Ni ion per unit cell is more, the calculated thickness would be smaller. However it does not violate t he conclusion that the thickness grows with the number of cycles. Number of cycle Measured thickness with AFM (nm) Measured thickness with SE (nm) 1 13.0 10.0 2 14.0 15.5 3 20.5 23.0 Table A 5 - 5 . The measured thicknesses of the Ni25 on silicon. 231 Figure A 5 - 14 . J - E curves measured for bare (blue) and Ni25 (a - b) and Ni75 (c - d) coated ED electrodes; a) Ni25 (light), b) Ni25 (dark), c) Ni75 (light), and d) Ni75 (dark) in contact with 1 M KOH solution. 232 Figure A 5 - 15 . a) C bulk , and b) Mott - Schottky values from EIS data for bare and Ni25 coated ED electrodes with 1cycle (pentagon), 2 cycles (circle), and 3 cycles (triangle pointing down). Also included are the value measured for the bare electrode (blue diamond). Figur e A 5 - 16 . a) C bulk , and b) Mott - Schottky values from EIS data for bare and Ni75 coated ED electrodes with 1cycle (pentagon), 2 cycles (circle), and 3 cycles (triangle pointing down). Also included are the v alue measured for the bare electrode (blue diamond). 233 Figure A 5 - 17 . a) R cat,Ni25 and b) R ct,Ni25 values from impedance response of ED hematite electrodes with 1cycle (pentagon), 2 cycles (circle), and 3 cycles (triangle pointing down). Bare ED hematite fitting R trap,bare and R ct,bare (blue diamond) are shown for comparison. Figure A 5 - 18 . a) R cat,Ni75 and b) R ct,Ni75 values fit from impedance response of ED hematite electrodes with 1cycle (pentagon), 2 cycles (circle), and 3 cycles (triangle pointing down). Bare ED hematite fitting R trap ,bare and R ct,bare (blue diamond) are shown for comparison. 234 Figure A 5 - 19 . Complex plots of normalized IMPS data for bare and catalyst coated ED electrodes at 1.3 V vs. RHE; bare electrode (blue triangle), Ni75 modified ED electrode (green diamond), and Ni25 coated ED electrode (orange circle) 235 REFERENCES 236 REFERENCES (1) Hamann, T. W. Splitting Water with Rust: Hematite Photoelectrochemistry. Dalt. Trans. 2012 , 41 (26), 7830 7834. (2) Zandi, O.; Schon, A. R.; Hajibabaei, H.; Hamann, T. W. Enhanced Charge Separation and Collection in High - Performance Electrodeposited Hemat ite Films. Chem. Mater. 2016 , 28 (3), 765 771. (3) Klahr, B. M.; Hamann, T. W. Current and Voltage Limiting Processes in Thin Film Hematite Electrodes. J. Phys. Chem. C 2011 , 115 (16), 8393 8399. (4) Klahr, B. M.; Martinson, A. B. F.; Hamann, T. W. Photoelectrochemical Investigation of Ultrathin Film Iron Oxide Solar Cells Prepared by Atomic Layer Deposition. Langmuir 2011 , 27 (1), 461 468. (5) Deng, J.; Lv, X.; Zhang, H.; Zhao, B.; Sun, X.; Zhong, J. Loading the FeNiOOH Cocatalyst on Pt - Modified Hematite Nanostructures for Efficient Solar Water Oxidation. Phys. Chem. Chem. Phys. 2016 , 18 (15), 10453 10458. (6) Young, K. M. H.; Hamann, T. W. Enhanced Photocatalytic Water Oxidation Efficiency with Ni(O H) 2 - Fe 2 O 3 via ALD. Chem. Commun. 2014 , 50 (63), 8727 8730. (7) Upul Wijayantha, K. G.; Saremi - Yarahmadi, S.; Peter, L. M. Kinetics of Oxygen Evolution - Fe 2 O 3 Photoanodes: A Study by Photoelectrochemical Impedance Spectroscopy. Phys. Chem. Chem. Phys. 2011 , 13 (12), 5264. (8) Gamelin, D. R. Water Splitting: Catalyst or Spectator? Nat. Chem. 2012 , 4 (12), 965 967. (9) Klahr, B.; Gimenez, S.; Fabregat - Santiago, F.; Bisquert, J.; Hamann, T. W. Photoelectrochemical and Impedance S pectroscopic Investigation of Water Oxidation with - - Coated Hematite Electrodes. J. Am. Chem. Soc. 2012 , 134 (40), 16693 16700. (10) McDonald, K. J.; Choi, K. S. Photodeposition of Co - Based Oxygen Evolution Catalysts on - Fe 2 O 3 Photoanodes. Chem. Mater. 2011 , 23 (7), 1686 1693. (11) Du, C.; Yang, X.; Mayer, M. T.; Hoyt, H.; Xie, J.; McMahon, G.; Bischoping, G.; Wang, D. Hematite - Based Water Splitting with Low Turn - on Voltages. Angew. Chemie - Int. Ed. 2013 , 52 (48), 12692 12695. (12) Zhong, D. K. ; Choi, S.; Gamelin, D. R. Near - Complete Suppression of Surface - - Modified W:BiVO 4 . J. Am. Chem. Soc. 2011 , 133 (45), 18370 18377. (13) Liu, R.; Zheng, Z.; Spurgeon, J.; Yang, X. Enhanced Photoele ctrochemical Water - Splitting Performance of Semiconductors by Surface Passivation Layers. Energy Environ. Sci. 2014 , 237 7 (8), 2504 2517. (14) Barroso, M.; Cowan, A. J.; Pendlebury, S. R.; Grätzel, M.; Klug, D. R.; Durrant, J. R. The Role of Cobalt Phosphate - Fe 2 O 3 toward Water Oxidation. J. Am. Chem. Soc. 2011 , 133 (38), 14868 14871. (15) Barroso, M.; Mesa, C. A.; Pendlebury, S. R.; Cowan, A. J.; Hisatomi, T.; Sivula, K.; Gratzel, M.; Klug, D. R.; Durrant, J. R. - Fe 2 O 3 Photoanodes for Solar Water Splitting. Proc. Natl. Acad. Sci. 2012 , 109 (39), 15640 15645. (16) Zhong, D. K.; Sun, J.; Inumaru, H.; Gamelin, D. R. Solar Water Oxidation by Composite - F e 2 O 3 Photoanodes. J. Am. Chem. Soc. 2009 , 131 (17), 6086 6087. (17) Lin, F.; Boettcher, S. W. Adaptive Semiconductor/Electrocatalyst Junctions in Water - Splitting Photoanodes. Nat. Mater. 2014 , 13 (1), 81 86. (18) Nellist, M. R.; Laskowski, F. A. L.; Lin, F.; Mills, T. J.; Boettcher, S. W. Semiconductor - Electrocatalyst Interfaces: Theory, Experiment, and Applications in Photoelectrochemical Water Splitting. Acc. Chem. Res. 2016 , 49 (4), 733 740. (19) Klahr, B.; Gim enez, S.; Fabregat - Santiago, F.; Hamann, T.; Bisquert, J. Water Oxidation at Hematite Photoelectrodes: The Role of Surface States. J. Am. Chem. Soc. 2012 , 134 (9), 4294 4302. (20) Zandi, O.; Hamann, T. W. Enhanced Water Splitting Efficiency through Select ive Surface State Removal. J. Phys. Chem. Lett. 2014 , 5 (9), 1522 1526. (21) Jang, J. - W.; Du, C.; Ye, Y.; Lin, Y.; Yao, X.; Thorne, J.; Liu, E.; McMahon, G.; Zhu, J.; Javey, A.; et al. Enabling Unassisted Solar Water Splitting by Iron Oxide and Silicon. N at. Commun. 2015 , 6 , 7447. (22) 4 Photoanodes for Improved Solar Water Splitting. J. Mater. Chem. A 2016 , 4 (8), 2919 2926. (23) Liu, G.; Ye, S.; Yan, P.; Xiong, F.; Fu, P.; Wang, Z.; Chen, Z.; Shi, J.; Li, C. Enabling an Integrated Tantalum Nitride Photoanode to Approach the Theoretical Photocurrent Limit for Solar Water Splitting. Energy Environ. Sci. 2016 , 9 (4), 1327 1334. (24) Le Formal, F.; Pendlebury, S. R.; Cornuz, M.; Tilley, S. D.; Grätzel, M.; Durrant, J. R. Back Electron - Hole Recombination in Hematite Photoanodes for Water Splitting. J. Am. Chem. Soc. 2014 , 136 (6), 2564 2574. (25) Kim, J. Y.; Magesh, G.; Youn, D. H.; Jang, J. W.; Kubota, J.; Domen, K.; Lee, J. S. Single - Crystalline, Wormlike Hematite Photoanodes for Efficient Solar Water Splitting. Sci. Rep. 2013 , 3 , 2681. (26) Morales - Guio, C. G.; Mayer, M. T.; Yella, A.; Tilley, S. D.; Grätzel, M.; Hu, X. An 238 Optically Transparent Iron Nickel Oxide Catalyst for Solar Water Splitting. J. Am. Chem. Soc. 2015 , 137 (31), 9927 9936. (27) Louie, M. W.; Bell, A. T. An Investigation of Thin - Film Ni - Fe Oxide Catalysts for the Electrochemic al Evolution of Oxygen. J. Am. Chem. Soc. 2013 , 135 (33), 12329 12337. (28) 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.; et al. Identification of Highly Active Fe Si tes in (Ni,Fe)OOH for Electrocatalytic Water Splitting. J. Am. Chem. Soc. 2015 , 137 (3), 1305 1313. (29) Trotochaud, L.; Young, S. L.; Ranney, J. K.; Boettcher, S. W. Nickel - Iron Oxyhydroxide Oxygen - Evolution Electrocatalysts: The Role of Intentional and Incidental Iron Incorporation. J. Am. Chem. Soc. 2014 , 136 (18), 6744 6753. (30) 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 T rends and Design Principles. Chem. Mater. 2015 , 27 (22), 7549 7558. (31) Zandi, O.; Beardslee, J. A.; Hamann, T. Substrate Dependent Water Splitting with Ultrathin - Fe 2 O 3 Electrodes. J. Phys. Chem. C 2014 , 118 (30), 16494 16503. (32) Dezelah; Niinistö, J.; Arstila, K.; Niinistö, L.; Winter, C. H. Atomic Layer Deposition of Ga 2 O 3 Films from a Dialkylamido - Based Precursor. Chem. Mater. 2006 , 18 (2), 471 475. (33) Smith, R. D. L.; Prevot, 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 (80 - . ). 2013 , 340 (6128), 60 63 . (34) Hilfiker, J. N.; Singh, N.; Tiwald, T.; Convey, D.; Smith, S. M.; Baker, J. H.; Tompkins, H. G. Survey of Methods to Characterize Thin Absorbing Films with Spectroscopic Ellipsometry. Thin Solid Films 2008 , 516 (22), 7979 7989. (35) Klaus, S.; Lou ie, M. W.; Trotochaud, L.; Bell, A. T. Role of Catalyst Preparation on the Electrocatalytic Activity of Ni 1 - x Fe x OOH for the Oxygen Evolution Reaction. J. Phys. Chem. C 2015 , 119 (32), 18303 18316. (36) Swierk, J. R.; Klaus, S.; Trotochaud, L.; Bell, A. T. ; Tilley, T. D. Electrochemical Study of the Energetics of the Oxygen Evolution Reaction at Nickel Iron (Oxy)Hydroxide Catalysts. J. Phys. Chem. C 2015 , 119 (33), 19022 19029. (37) Klaus, S.; Cai, Y.; Louie, M. W.; Trotochaud, L.; Bell, A. T. Effects of F e Electrolyte Impurities on Ni(OH) 2 /NiOOH Structure and Oxygen Evolution Activity. J. Phys. Chem. C 2015 , 119 (13), 7243 7254. (38) Smith, R. D. L.; Prévot, M. S.; Fagan, R. D.; Trudel, S.; Berlinguette, C. P. Water Oxidation Catalysis: Electrocatalytic R esponse to Metal Stoichiometry in Amorphous Metal Oxide Films Containing Iron, Cobalt, and Nickel. J. Am. Chem. Soc. 2013 , 135 (31), 11580 11586. 239 (39) 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 (5), 1025 1035. (40) Andronic, L. S.; Hill, R. H. The Mechanism of the Photochemical Metal Organic Deposition of Lead Oxide Films Thin Films of Lead (II) 2 - Ethylhexanoate. J. Photochem. Photobiol. A Chem. 2002 , 152 (1 3), 259 265. (41) Young, K. M. H.; Hamann, T. W. Enhanced Photocatalytic Water Oxidation Efficiency with Ni(OH) 2 - Fe 2 O 3 via ALD. Chem. Commun. 2014 , 50 (63), 8727 8730. (42) Gelderman, K.; Lee, L.; Donne, S. W. Flat - Band Potential of a Semiconductor: Using the Mott Schottky Equation. J. Chem. Educ. 2007 , 84 (4), 685. (43) Glasscock, J. A.; Barnes, P. R. F.; Plumb, I. C.; Bendavid, A.; Martin, P . J. Structural, Optical and Electrical Properties of Undoped Polycrystalline Hematite Thin Films Produced Using Filtered Arc Deposition. Thin Solid Films 2008 , 516 (8), 1716 1724. (44) Gomes, W. P.; Cardon, F. Electron Energy Levels in Semiconductor Elec trochemistry. Prog. Surf. Sci. 1982 , 12 (2), 155 215. (45) Van Den Meerakker, J. E. A. M.; Meulenkamp, E. A.; Scholten, M. (Photo)Electrochemical Characterization of Tin - Doped Indium Oxide. J. Appl. Phys. 1993 , 74 (5), 3282 3288. (46) Klahr, B.; Hamann, T. Water Oxidation on Hematite Photoelectrodes: Insight into the Nature of Surface States through in Situ Spectroelectrochemistry. J. Phys. Chem. C 2014 , 118 (19), 10393 10399. (47) Batchellor, A. S.; Boettcher, S. W. Pulse - Electrodeposited Ni - Fe (Oxy)Hyd roxide Oxygen Evolution Electrocatalysts with High Geometric and Intrinsic Activities at Large Mass Loadings. ACS Catal. 2015 , 5 (11), 6680 6689. (48) Thorne, J. E.; Li, S.; Du, C.; Qin, G.; Wang, D. Energetics at the Surface of Photoelectrodes and Its Influence on the Photoelectrochemical Properties. J. Phys. Chem. Lett. 2015 , 6 (20), 4083 4088. (49) Bertoluzzi, L.; Bisquert, J. Equivalent Circuit of E lectrons and Holes in Thin Semiconductor Films for Photoelectrochemical Water Splitting Applications. J. Phys. Chem. Lett. 2012 , 3 (17), 2517 2522. (50) Riha, S. C.; Klahr, B. M.; Tyo, E. C.; Seifert, S.; Vajda, S.; Pellin, M. J.; Hamann, T. W.; Martinson , A. B. F. Atomic Layer Deposition of a Submonolayer Catalyst for the Enhanced Photoelectrochemical Performance of Water Oxidation with Hematite. ACS Nano 2013 , 7 (3), 2396 2405. (51) Ponomarev, E. A.; Peter, L. M. A Generalized Theory of Intensity Modulated Photocurrent Spectroscopy (IMPS). J. Electroanal. Chem. 1995 , 396 (1 2), 219 226. 240 (52) Peter, L. M.; Wijayantha, K. G. U.; Tahir, A. A. Kinetics of Light - Driven Oxygen Evolution at - Fe 2 O 3 Electrodes. Faraday Discuss. 2012 , 155 (0), 309 322. (53) Thorne, J. E.; Jang, J. W.; Liu, E. Y.; Wang, D. Understanding the Origin of Photoelectrode Performance Enhancement by Probing Surface Kinetics. Chem. Sci. 2016 , 7 (5), 3347 3354. (54) Wa ng, G.; Ling, Y.; Lu, X.; Zhai, T.; Qian, F.; Tong, Y.; Li, Y. A Mechanistic Study into the Catalytic Effect of Ni(OH) 2 on Hematite for Photoelectrochemical Water Oxidation. Nanoscale 2013 , 5 (10), 4129. (55) Kim, T. W.; Choi, K. - S. Nanoporous BiVO 4 Photo anodes with Dual - Layer Oxygen Evolution Catalysts for Solar Water Splitting. Science (80 - . ). 2014 , 343 (6174), 990 994. (56) Qiu, J.; Hajibabaei, H.; Nellist, M. R.; Laskowski, F. A. L.; Oener, S. Z.; Hamann, T. W.; Boettcher, S. W. Catalyst Deposition on Photoanodes: The Roles of Intrinsic Catalytic Activity, Catalyst Electrical Conductivity, and Semiconductor Morphology. ACS E nergy Lett. 2018 , 3 (4), 961 969. (57) Nakahira, A.; Kubo, T.; Murase, H. Synthesis of LDH - Type Clay Substituted with Fe and Ni Ion for Arsenic Removal and Its Application to Magnetic Separation. IEEE Trans. Magn. 2007 , 43 (6), 2442 2444. (58) Iwasaki, T .; Yoshii, H.; Nakamura, H.; Watano, S. Simple and Rapid Synthesis of Ni - Fe Layered Double Hydroxide by a New Mechanochemical Method. Appl. Clay Sci. 2012 , 58 , 120 124. (59) Li, Y.; Li, H.; Yang, M.; He, X.; Ni, P.; Kang, L.; Liu, Z. - H. Topochemical Synth esis of Ni 2+ Fe 3+ Layered Double Hydroxides with Large Size. Appl. Clay Sci. 2011 , 52 (1 2), 51 55. 241 The Role of Catalyst Composition on the PEC Water Oxidation Performance of Hematite with Different Crystal Orientations 242 6.1. Abstract In this chapter, the PEC water oxidation performance of catalytically modified hematite electrodes with different crystal orientations, i.e. (100) and (001), coated with two compositions of Ni 1 - x Fe x O y catalysts, i.e. Ni75 and Ni25 catalysts, was investigated. A series of photoelectrochemical methods including EIS and IMPS were utilized to assess the dynamics of charge transfer on the catalytically modified electrodes. It was shown that under identical PEC water oxidation conditions, (100) - oriented hematite electrode outperforms the bare electrode with (001) orientation. Interestingly, for the catalyst - coated hematite electrodes, the Ni25 coated - electrodes produced almost identical PEC characteristics fo r both orientations, whereas the PEC performance of Ni75 - coated electrodes exhibited strongly dependent on the crystal - orientation. For example, Ni75 coated (100) and (001) hematite electrodes produced a hole - collection efficiency of ~ 60% and 10%, respect ively, at 1.1 V vs. RHE, which manifested itself in higher photocurrent densities at this potential. Based on the IMPS and EIS measurements, it was shown that the improved charge - separation efficiency at the surface of the electrode is responsible for the enhancement in PEC efficiency . 6.2. Introduction A deep knowledge of the dynamics of photogenerated charges at the semiconductor and catalyst junction is highly beneficial as it allows to engineer the photoelectrode surface to enhance its charge collection effi ciency. It has been shown that the catalytic modification substantially improves the PEC performance of the semiconductor and for some cases the improvement in is only achieved when prior to deposition of catalyst the semiconductor surface is modified with a non - catalytic layer. 1 3 These observations clearly underline the role of the interface at the catalyst and semiconductor junction. As mentioned in chapter 5, in presence of pin - holes and conductive 243 catalyst t he PEC performance of catalytically modified hematite electrodes is dominated by shunting to the conductive substrate. Therefore, pin - hole - free electrodes with various crystal orientations are ultimately the ideal model electrodes to investigate the role o f semiconductor|catalyst junction. Recently, Rothschild et al. 4 studied the effect of the crystal orientation on the bulk and the surface properties of heteroepitaxially - grown Sn - doped hematite by PLD/MBE . They observed that under identical conditions for PEC water oxidation, the photocurrent onset potential s for (110) - and (100) - oriented hematite w ere located ~ 170 mV more negative than the (0 01) - oriented hematite electrode . In the presen ce of a fast hole scavenger such as H 2 O 2 , however, the photocurrent onset potential and the photocurrent density at any given potential were found to be independent of the crystal orientation. Subsequent analysis via EIS measurement s revealed relatively similar bulk properties (dopant density and flat - band potentials) for these series of hematite electrodes. On the basis of these observations , it was concluded that the crystal orientation primarily influences the kinetic s of charge transfer at the surface of the electrode . 4 The heteroepitaxially - grown hematite electrodes by this method are promising candidates for fundamental investigation of the effec t of surface orientation on the catalytically - modified electrodes. In this chapter, the combinations of two compositions of Ni 1 - x Fe x O y catalysts, i.e. Ni - rich and Fe - rich catalysts, prepared by UV light - assisted deposition and hematite electrodes with two distinctive orientations, i.e. (100) and (001), supplied by the Rothschild Technion, were utilized as model electrodes with systematically different interfacial composition/and structure at the semiconductor|catalyst junction. The crystal struct ure of hematite at the surface for these two orientations are shown in Figure 6 1 . As it can be seen, the (100) orientation is comprised of two types of terminal hydroxyl groups at the surface whereas the (001) orientation contains only one type of bridging oxygen groups. In addition, the number of iron sites 244 Figure 6 1 . The surface configuration of hematite as a function of crystal orientations, a) (100) and b) (001) orientations. The iron centers and the lattice oxygens are shown in orange (large circle) and red (small circle), respec tively. The singly and doubly coordinated oxygen sites at the surface are shown in yellow and cyan, respectively. Throughout this chapter the (100) and (001) orientations are labeled as M and C, respectively. per unit area is higher for the (001) orient ations (18 sites vs. 12 sites). As the result, the nature and density of states at the surface of hematite electrodes with different crystal orientations are different. The combination of different surfaces with catalyst therefore creates new states which can modulate the charge transfer from semiconductor to the catalyst. Herein we aim to understand how the PEC water oxidation performance of the catalytically modified hematite electrodes varies as a function of composition and structure of the interface. 245 6.3. E xperimental 6.3.1. Electrode Preparation The hematite electrodes with various orientations were heteroepitaxially grown by PLD. These electrodes were supplied by the Rothschild about the preparation method and their properties can be found in literature. 4 Thin films of Ni 0.75 Fe 0.25 O y (Ni75) and Ni 0.25 Fe 0.75 O y (Ni25) catalysts were deposited by spin - coating on the hematite substrates using a modified version of previously reported procedures. 2,5 Prior to deposit ion of catalysts, the hematite electrode was sequentially cleaned by spin - coating with DI water and hexane for 30 s at 5000 rpm. Precursor solutions were prepared from iron (III) 2 - ethylhexanoate (50% w/w in mineral spirits, Strem Chemicals), and nickel (I I) 2 - ethylhexanoate (78% w/w in 2 - ethylhexanoic acid, Strem Chemicals) by dissolving the appropriate amount of metal precursor in hexanes to give a total concentration of metals of 0.05 M with the ratio of 15% w/w metal complex to hexane ( see Table A 5 - 1 ) . Subsequently, 250 µL of the catalyst solution was spun coated at 5000 rpm for 60 s. The as - prepared films were irradiated with UV light (254 nm, 5.8W ) for 1 h followed by annealing in a preheated furnace in air at 100 °C for 1 h. The ohmic electrical contact was made by gently scratching Ga - In eutectic (99.99% Sigma Aldrich) to the back contact and then attaching a copper wire with silver epoxy on to i t. The electrode was then placed on a preheated hot plate at 100 °C to dry out and harden it. 6.3.2. Photoelectrochemical Measurements The photoelectrochemical and electrochemical impedance spectroscopic measurements were made with an Eco Chemie Autolab potentios tat coupled with Nova electrochemical software. 246 Unless otherwise stated, all photoelectrochemical tests were carried out by shining light on the electrodes through the electrolyte. Prior to PEC characterization each electrode was aged (conditioned) by 40 c yclic voltammetry in the potential range of 0.5 to 1.8 V vs. RHE with the scan rate of the 100 mV s - 1 in 1 M KOH. The current density vs . applied potential ( J - E ) for the bare and modified hematite electrodes were examined by cycling the potential between 0. 5 and 1.5 V vs. RHE at the scan rate of 20 mV s - 1 in contact with 1 M KOH with pH = 13.6 as determined with Fisher Scientific Accument pH meter . A 1 M KOH solution containing 0. 5 M H 2 O 2 was used as a hole scaven ger. Subsequently the ratio of J photo (H 2 O) to J photo (H 2 O 2 ) was used to calculate the hole collection efficiency. The light source was a 450 W Xe arc lamp (Horiba Jobin Yvon) and the AM 1.5 solar filter was used to simulate sunlight at 100 mW cm 2 (1 sun). A homemade saturated Ag/AgCl and a platinum mesh electrode were used as the reference and counter electrode, respectively. The reference electrode was frequently calibrated against commercial saturated calomel electrode (Koslow Scientific). All the potenti als were converted to reversible hydrogen electrode (RHE) using . Electrochemical i mpedance spectroscopic data was measured at the potential range of 0. 8 and 1.5 V vs. RHE with 100 mV intervals using a 10 mV amplitude perturbation between 10,000 and 0.01 Hz. Data were fit to the appropriate equivalent circuit using Zview software (Scribner Associates). The IMPS measurements were performed using Eco Chemie Autolab potentiostat equipped with metroh m LED driver accessory . A 470 nm LED light was utilized as the light source. The power of the LED was chosen to produce the same photocurrent as it was measured under 1 sun illumination. The modulation intensity was set at 10% of the power of the LED to en sure a linear respond and the frequency was swept from 15 kHz to 0.1 Hz with 10 frequency per decade increments. The IMPS data were collected over the potential range of 0. 8 to 1.5 V vs. RHE with 100 mV intervals. 247 The data shown here were collected for a t otal number of 8 electrodes (4 sets × 2 orientations, see Table A 6 - 1 ). Due to the limited number of the electrodes, the reproducibility of the results were only examined for Ni25 coated electrodes (set2 and set 4). The J - E curves, high frequency (HF), and low frequency (LF) IMPS responses for these electrodes are compared to the corresponding bare electrodes (Set1) in Figure A 6 - 1 . For Ni25 coated C - hematite electrodes, while the HF responses are highly reproducible bu t the photocurrent onset potential are different by 100 mV (evident from both J - E and LF). For the Ni25 coated M - hematite electrodes, the photocurrent onset potentials are relatively similar (evident from both J - E and LF) but the photocurrent density at th e higher potentials and the HF responses are different by ~ 25%. While, the reproducibility of the performances are highly desirable and more experiments are required to statistically evaluate any interface - performance relationship, for the following discu ssion the results for the best performing electrodes are utilized. 6.4. Results and Discussion The dark J - E curves of the bare and catalyst coated electrodes exhibit a highly rectifying behavior and produce a negligible dark (leakage) current only at high posit ive potentials (shown in Figure A 6 - 2 ). This is an important observation as it ensures that the behavior of the catalytically - modified electrodes are n ot influenced by shunting. The light J - E curves of the catalyst - coated hematite electrodes with various compositions of catalyst are compared to the corresponding bare electrodes i n Figure 6 2 .a. Under PEC water oxidation conditions the bare M orientation outperforms the bare C orientation hematite electrode. In presence of a fast hole scavenger, however, they produce an indistinguishable J - E curves ( Figure A 6 - 3 ) . These results are consistent with the original report, 4 the superior performance of M orientation being due to its higher catalytic activity. As 248 shown in Figure 6 2 . a , the surface modification with Ni25 and Ni75 results in a cathodic shift in the photocurrent onset potentials. In comparison to the bare electrode, these J - E curves exhibit a wide hysteresis between the forward (anodic sweep) and reverse sca n (cathodic sweep) - especially for Ni75 coated electrodes. To estimate the photocurrent onset potentials, thus, we used c hronoamperometry to determine the photocurrent at various applied potentials under steady - state conditions ( Figure A 6 - 4 ). Interestingly, regardless of the crystal orientation of the hematite electrodes, the photocurrent onset potential for both Ni75 - and Ni25 - coated hematite electrodes are located at relativel y similar potentials. Figure 6 2 . Comparison of PEC performance of catalytically modified hematite electrodes with various crystal orientations and different composition of catalysts. a) J - E curves of ba re, Ni25 - (orange) and Ni75 - (green) coated hematite electrodes with different orientations, b) hole collection efficiency for bare, Ni25 - (orange) and Ni75 - (green) coated hematite electrodes. The (100) (labeled as M), and (001) (labeled as C) orientations ar e shown in broken line with empty shapes and solid lines with solid shapes, respectively. Note: for clarity, the dark currents are omitted and are shown 249 in Figure A 6 - 2 . The ratio of the low frequency (LF) to high frequency (HF) of IMPS responses were used to calculate the hc % . 6,7 The J - E curves of the Ni75 - coated electrodes exhibit two additional features: (1) Ni75 improves the turn - on potential, but the current densities at higher potentials are lower than that of the bare or Ni25 - coated electrodes. This behavior can be ascribed to the absorption of light by the catalyst (electrode was illuminated through the solution) as confirmed by the spectroelectrochemistry analysis of the Ni75 - and Ni25 - coated FTO electrodes (results are discussed following Figure A 6 - 5 ). (2) A unique set of cathodic peaks appeared during the reverse scan, which are not observed for the bare electrodes. This suggests that the photogenerated holes are stored into the catalyst layer. Ideally, the electrochemical properties of the catalyst should be independent of the substrate, thus the redox characteristics, i.e. peak position and area, of the Ni75 catalyst with the same thickness should remain the same. As can be seen the position in Figure 6 2 . a , shape and the area of the peaks depend on the substrate and on the crystal orientation of the hematite substrate. A sharp reduction peak can be seen in Figure 6 2 . a for Ni75 - coated C - hematite electrode, while the Ni75 - coated M - hematite produces a rather broad peak. This can be better understood by considering the surface configuration of hematite grown in different crystal orientations ( Figure 6 1 ). The surface of C - hematite ((001) orientation) is uniquely comprised of only doubly coordinated oxygen groups while the surface of M - hematite ((100) orientations) possesses both singly and doubly coordinated oxygen groups with the ratio of 2:1. 8 Therefore, it can be postulated that the combination of C - hematite with Ni75 forms a narrow density of interface states, as a result th e reduction peak is sharper. 250 The hole collection efficiency for the bare and catalyst coated hematite electrode with varying crystal orientation and different catalyst composition as a function of potential are shown in Figure 6 2 . b. Consistent with the conclusion made in the original work, the better performance of the bare M - hematite in comparison to the C - hematite is due to its superior hole collecti on efficiency. Figure 6 2 .b show s that for the catalyst coated electrodes the hole collection efficiencies (HCE) at the potentials beyond ~ 1.2 V vs. RHE is unity. At 1.1 V vs. RHE however, coated M - and C - hematite produced ~ 60% and 10% hole collection efficiency, respectively. The superior hole collection efficie ncy is further manifested in the slightly higher current density at 1.1 V vs. RHE for Ni75 coated M - hematite (evident from the steady state J - E curves shown in Figure A 6 - 4 ). The reproducibility of these results needs to be evaluated by conducting more experiments. The light EIS measurement was utilized to further study the effect of crystal orientation on the PEC performance of the bare, N i75 - , and Ni2 5 - coated hematite electrodes. The EIS responses were fitted to the equivalent circuit shown in Figure A 5 - 12 . For the potentials with only one semicir cle/peak in the Nyquist/Bode plot the Randles circuit and for the potentials with two semicircle/peaks the previously reported equivalent circuit for CoPi coated hematite electrodes 9 were used to extract the capacitances and charge transfer resistances. Since catalyst coating is solely a surface modification, it is expected that the dopant density and the flat band potential (bulk property) of semiconductor to remain constant. Plots of C bulk and Mott Schottky (MS) for the bare and catalyst coated electrodes with different orientations and catalyst compositions are compared in Figure A 6 - 6 . A nearly identical slope and intercept can be calculated for all electrodes. B y fitting the linear slope to the MS equation 10 and using a dielectric constant of 32 for hematite 11 and th e geometric surface area of the electrodes , a d opant density of 8 - 11 × 10 19 cm - 3 was calculated. 251 Figure 6 3 . The calculated a) R trap , R cat and b) R ct for the bare and catalyst coated hematite electrodes with different crystal orientation of hematite and varying composition of catalyst as a function of applied potential. The C - and M - orientation are shown in solid and empty shapes, respectively. The Ni7 5 and Ni25 catalyst coated electrodes are shown in green and orange, respectively. Due to the highly doped nature of these electrodes, the MS equation cannot be used to determine the true flat band potential of these electrodes. The high frequency (HF) res ponse from the IMPS measurements - corresponding to the flux of the holes toward the surface of the electrode 6,7 - was used to assess how ban d bending changes upon catalyst modification. The value of HF as a function of applied potential for the bare and catalyst coated hematite electrodes with various orientations are shown i n Figure A 6 - 7 . As it can be seen, across different electrodes the HF values are relatively constant and are independent to the composition of the catalyst or crystal orientation of the hematite electrodes (see Figure A 6 - 3 ) which further indicates that the catalyst coating is solely a surface modification. The calculated charge transfer resistance from hematite electrode to the surface state ( R trap ) and catalyst ( R cat ) are shown in Figure 6 3 .a . As shown, R trap and R cat values are relatively constant for all the electrodes with different orientations and catalyst composition. This further implies that the 252 charge transfer from semiconductor to the catalyst (or to the surface state) is independent from the crystal orientation of hematite electrode and the compos ition of the catalyst. The value of the R ct shown in Figure 6 3 .b, however is reduced by 1 - 2 orders of magnitude after catalytic modification. Intere stingly, the lowest R c t value was observed for the Ni75 coated M - hematite. This agrees with the slightly better current density and hole collection efficiency observed in Figure 6 3 . In addition, at 1.1 V vs. RHE (the potential close to the photocurrent onset potential) the calculated R c t value for Ni75 coated C - hematite is about 2 - 5 times higher than that of other catal yst coated electrodes. This observation is consistent with slightly lower current density for the C(Ni75) electrode. The calculated surface state capacitance ( C ss ) and capacitance of the catalyst ( C cat ) for the bare and catalyst coated hematite with differ ent compositions are shown in Figure A 6 - 8 . In line with previous reports, 4,12 the C ss exhibits a peak that coincides with the photocurrent onset potential ( Figure A 6 - 4 ). Consistent with the original work 4 the surface state capacitance of C orientation was found to be higher than for the M orientations. As can be seen, the C cat across all the catalytically modified electrodes are larger than the C ss . This observation indicates that the photogenerated holes are stored in the catalyst. Depending on the catalyst composition and orientation of hematite, C cat vs. applied potential exhibit multiple features: (1) For Ni25 - coated electrodes, while the peak position is independent to the crystal orientation of the hematite substrate the value of C Ni25 increases by a factor of ~ 4 from C - orientation to M - orientation. (2) For the Ni75 - coated electrodes both the value o f C Ni 7 5 and its position depends on the crystal orientation of the hematite substrate. For the Ni75 coated C - hematite, the capacitance of the catalyst peaks around 200 mV after the photocurrent onset potential, while for the Ni75 - coated M - hematite, the ana log peak occurs near the photocurrent onset potential. In addition, the magnitude 253 of C cat for C(Ni75) is ~ 2 times higher than M(Ni75). (3) Due to the higher density of acceptor states in Ni75, i.e. Ni 2+ , it is expected that C Ni75 to be larger than C Ni25 (by a factor ~ 3), however, this behavior was not observed for these electrodes. The discrepancy in latter can be ascribed to the variation in the catalyst thickness. As discussed in chapter 5, the growth rate of catalyst depends on its composition and va ries on different substrates. At this stage, it is premature to draw conclusion on the effect of crystal orientation of hematite and catalyst composition on these catalytically modified electrodes as more experiments are required to confirm the reproducibi lity of these observations. In summary, the PEC performance of a series of the catalyst coated hematite electrodes with different crystal orientations were studied. We note that while the photocurrent onset potential of the bare electrode strongly depend o n the crystal orientation of hematite, the catalyst coated electrodes (with Ni75 and Ni25) produced a nearly identical photocurrent onset potentials (1.0 - 1. 1 V vs. RHE). The EIS and IMPS measurements have collectively shown that the improved hole collect ion efficiency /reduced charge transfer resistance at the surface of the electrode is the origin of the enhancement in PEC. 254 A PPENDIX 255 Sample C (001) M (100) Set1 (Bare) Set2 (coated with Ni25) Set3 (coated with Ni75) Set4 (coated with Ni25) Table A 6 - 1 . The list of the samples studied in this study. 256 Figure A 6 - 1 . The reproducibility of the PEC performances of the Ni25 coated hematite electrodes with various orientations. The J - E curve of a) C - (001) and b) M - (100) oriented hematite electrodes; the high frequency (HF, filled shapes) and low frequency (LF, empty sha pes) responses of different batches of Ni25 coated hematite electrodes with c) (001) and d) (100) orientations. The bare electrodes are shown with solid lines. The Dark J - E curves of corresponding electrodes are shown in Figure A 6 - 2 . 257 Figure A 6 - 2 . The dark current J - E curves of bare and catalytically modified hematite electrodes with various orientations and different catalyst composition. The bottom plot is the zoomed plot indicating residual current at high potentials (>1.6 V vs. RHE). 258 Figure A 6 - 3 . A) J - E curves of the bare hematite electrodes with various orientations in light (solid lines) and dark (dashed line) in contact with 1 M KOH solution containing 0.5 M H 2 O 2 as the hole scavenger. b) Calculated hole coll ection efficiency using: J photo (KOH)/ J photo (H 2 O 2 ); J photo (KOH) or J photo (H 2 O 2 ) was calculated using: J photo (V) = J light (V) J dark (V). 259 Figure A 6 - 4 . The steady - state light J - E curves of the bare and catalytically modified hematite electrode with varying crystal orientation and composition of catalyst. The current was measured in the range of 0.5 1.8 V vs. RHE with 100 mV intervals. At each applied potential the photocurrent w as recorded until a stable current was achieved (~ 120s) and reported as the steady state current density. 260 Figure A 6 - 5 . The transmittance of a) bare, b) Ni25, and c) Ni75 coated FTO electrode as a function of the applied potential in contact with 1 M KOH. For FTO and Ni25| FTO electrodes the potential was scanned from 1.4 to 2.0 V vs . RHE with 100 mV intervals. For Ni75| F TO electrode, the potential was scanned in the same range but with 25 mV intervals. The background for each electrode was measured at open circuit potential at the beginning of the experiment. Prior to transmittance measurement at each potential the curren t was let to stabilize for 60 - 120 s. As it can be seen, Ni25 electrode produce a very similar % transmittances profile as the bare FTO electrode where the transmittance gradually decrease upon an increase in applied potential with a maximum reduction of ~ 20% at 2.0 V vs . RHE (Ni25 catalyst is completely oxidized at this potential, see Figure 5 4 a ). On the other hand, the transmittance of Ni75 sharply decreases by 50% upon increasing potential by 25 mV from 1.425 to 1.500 V vs . RHE (this switching potential is similar to the onset of oxidation potential of Ni75, shown in Figure 5 4 a ). 261 Figure A 6 - 6 . The fitted parameters of EIS me a surements using equi v alent cir c uits shown i n Figure A 5 - 15 for the bare and catalyst coated hematite electrodes with various orientations and different composition of the catalyst. a) C bulk and b) Mott S chottky plots calculated from the C bulk . The C and M or ienta tions are shown with the solid and empty shapes, respectively. 262 Figure A 6 - 7 . The HF response of the bare and catalyst coated hematite electrodes with different orientations in cont act with 1 M KOH. The C - and M - orientation is shown in solid and empty shapes, respectively. The Ni75 and Ni25 catalyst coated electrodes are shown in green and orange, respectively. 263 Figure A 6 - 8 . The surface state capacitance ( C ss ) of the bare hematite electrodes and the capacitance of catalyst ( C cat ) for the catalyst coated electrodes with different crystal orientation and composition of catalyst. The M - and C - orientations are shown with open and clos ed shapes, respectively. The Ni75 and Ni25 are shown in the green and orange colors, respectively. 264 R EFERENCES 265 REFERENCES (1) Liu, G.; Ye, S.; Yan, P.; Xiong, F.; Fu, P.; Wang, Z.; Chen, Z.; Shi, J.; Li, C. Enabling an Integrated Tantalum Nitride Photoanode to Approach the Theoretical Photocurrent Limit for Solar Water Splitting. Energy Environ. Sci. 2016 , 9 (4), 1327 1334. (2) Jang, J. - W.; Du, C.; Ye, Y.; Lin, Y.; Yao, X.; Thorne, J.; Liu, E.; McMahon, G.; Zhu, J.; Javey, A.; et al. Enabling Unassisted Solar Water Splitting by Iron Oxide and Silicon. Nat. Commun. 2015 , 6 , 7447. (3) Kim, T. W.; Choi, K. - S. Nanoporous BiVO 4 Photo anodes with Dual - Layer Oxygen Evolution Catalysts for Solar Water Splitting. Science (80 - . ). 2014 , 343 (6174), 990 994. (4) Grave, D. A.; Klotz, D.; Kay, A.; Dotan, H.; Gupta, B.; Visoly - Fisher, I.; Rothschild, A. Effect of Orientation on Bulk and Surfac e Properties of Sn - - Fe 2 O 3 ) Heteroepitaxial Thin Film Photoanodes. J. Phys. Chem. C 2016 , 120 (51), 28961 28970. (5) Smith, R. D. L.; Prevot, M. S.; Fagan, R. D.; Zhang, Z.; Sedach, P. A.; Siu, M. K. J.; Trudel, S.; Berlinguette, C. P. Pho tochemical Route for Accessing Amorphous Metal Oxide Materials for Water Oxidation Catalysis. Science (80 - . ). 2013 , 340 (6128), 60 63. (6) Ponomarev, E. A.; Peter, L. M. A Generalized Theory of Intensity Modulated Photocurrent Spectroscopy (IMPS). J. Ele ctroanal. Chem. 1995 , 396 (1 2), 219 226. (7) Peter, L. M.; Wijayantha, K. G. U.; Tahir, A. A. Kinetics of Light - Driven Oxygen Evolution - Fe 2 O 3 Electrodes. Faraday Discuss. 2012 , 155 (0), 309 322. (8) Barrón, V.; Torrent, J. Surface Hydroxyl Configuration of Various Crystal Faces of Hematite and Goethite. J. Colloid Interface Sci. 1996 , 177 (2), 407 410. (9) Klahr, B.; Gimenez, S.; Fabregat - Santiago, F.; Bisquert, J.; Hamann, T. W. Photoelectrochemical and Impedance Spectroscopic Investigation of Water Oxidation with - - Coated Hematite Electrodes. J. Am. Chem. Soc. 2012 , 134 (40), 16693 16700. (10) Gelderman, K.; Lee, L.; Donne, S. W. Flat - Band Potential of a Semiconductor: Using the Mott Schottky Equation. J. Chem. Educ. 2007 , 84 (4), 685. (11) Glasscock, J. A.; Barnes, P. R. F.; Plumb, I. C.; Bendavid, A.; Martin, P. J. Structural, Optical and Electrical Properties of Undoped Polycrystalline Hematite Thin Films Produced Using Filtered Arc Deposi tion. Thin Solid Films 2008 , 516 (8), 1716 1724. (12) Klahr, B.; Gimenez, S.; Fabregat - Santiago, F.; Hamann, T.; Bisquert, J. Water Oxidation at Hematite Photoelectrodes: The Role of Surface States. J. Am. Chem. Soc. 2012 , 134 (9), 4294 4302. 266 Conclusion and Future Directions 267 7.1. Summary and Conclusion s As outlined in the first chapter, due to its favorable properties Ta 3 N 5 stands out as a promising candidate for the PEC water oxidation reaction. Experimentally it has been shown that its perform ance can reach its theoretical limits. The best examples of PEC water oxidation with Ta 3 N 5 have shown promising performance only on tantalum substrate as the electron collector back contact. 1,2 Ta - substrate is not transp arent to the sub - band gap photons. In addition, the oxidation of Ta - substrate followed by ammonolysis (the common synthesis procedure) provides a highly reactive environment which preclude to realize Ta 3 N 5 on transparent conductive substrates. To circumven t these issues we utilized electrodeposition and atomic layer deposition as the bottom - up approaches to deposit tantalum oxides and nitride. These methods are promising as they provide a viable approach to fabricate tantalum oxide and nitrides on different substrates. In the first project, the electrodeposition of tantalum(V) oxide from aqueous solution was developed. Electrodeposition of tantalum oxide is challenging because Ta 5+ is highly susceptible to reduction which often results in the deposition of t he metallic Ta or TaO x sub - oxide s . Moreover, most of the Ta precursors vigorously react with water which precipitate as tantalum oxide. To circumvent these issues a highly acidic solution of tantalum precursor was used where the hydrolysis rate of tantalum oxide can be controlled via regulating pH at the electrode surface. KNO 3 was further utilized as the sacrificial electroactive reagent which upon electroreduction increases the solution pH near the electrode surface triggering precipitation of tantalum ox ide. The as - deposited films were amorphous nanostructured tantalum oxide which further were nitridized to Ta 3 N 5 via annealing in ammonia . The resultant Ta 3 N 5 film showed a promising PEC performance with an early turn on potential of 0.5 V vs. RHE. 268 Although electrodeposition is promising but it is beneficial to eliminate ammonolysis step. In the second project, tantalum nitride was integrate with Ta - doped TiO 2 (TTO), i.e. transparent conductive substrate (TCO), via ALD. It was shown that the low - temp erature ALD results in an amorphous film of tantalum oxynitrides where the composition of the films strongly depend on the deposition temperature (more nitrogen is incorporated as the deposition temperature increases). These films were crystalize into Ta 3 N 5 by a quick annealing in ammonia at 750 °C for 30 minutes. Subsequently, Ta - doped TiO 2 was developed as a stable TCO under reducing conditions of ammonolysis via ALD. Finally, these two layers were integrated to synthesize the first example of tantalum ni tride on a transparent conductive substrate. Furthermore, it was shown that the PEC water oxidation performance of tantalum nitride is strongly correlated to the conductivity of the TTO substrate. Thus, it is highly advantageous to directly deposit crystal line phase of Ta 3 N 5 on the state - of - the - art TCO such as FTO. In the subsequent project, a high temperature (T max = 650 °C) and fully automated ALD system was designed and built. TaCl 5 coupled with ammonia were utilized to directly deposit crystalline tanta lum nitride. It was found that 550 °C is the minimum temperature required to deposit crystalline Ta 3 N 5 on FTO substrate. Due to the crystal mismatch between FTO substrate, and Ta 3 N 5 the crystallinity, morphology and topography, growth rate, and optical den sity of the films were correlated on the number of ALD cycles. Therefore, in the first ~ 150 cycles a buffer layer with relatively similar morphology to the substrate is initially formed where after, Ta 3 N 5 is randomly deposited. Furthermore, it was shown t hat FTO remained stable under reducing conditions of ammonia at 550 °C and remains transparent and conductive which further allowed to realize the 1 st example of directly deposited Ta 3 N 5 photoanode on FTO. The PEC measurements as a function of illumination direction coupled with EIS measurements revealed that the PEC performance of Ta 3 N 5 is controlled by the diffusion of hole. 269 In the next section of this dissertation, the effect of the ca talyst on the PEC performance of hematite was studied. The Ni 1 - x Fe x O y coated planner hematite electrodes prepared by ALD and ED were utilized as the model electrodes to investigate how the catalyst interface with the underlying semiconductor. Depending on the catalyst composition and the hematite electrode, two distinctive behaviors were observed. In case of iron - rich catalyst (Ni25), the PEC water oxidation performance of both ALD and ED hematite electrodes were significantly improved. The electrochemical analysis via EIS and IMPS measurement have collectively shown that the Ni25 behave as a charge storage layer which improves PEC performance of the electrode by promoting the charge separation efficiency and reducing the recombination at the surface of elec trode . On the other hand, the PEC water oxidation performance of the Ni - rich (Ni75) coated hematite electrodes strongly depends on the preparation method of hematite electrodes. In case of pin - hole free hematite electrodes prepared by ALD, Ni75 improves th e performance of the hematite electrode similar to that realized for Ni25 coated hematite electrode. This was further confirmed via dual working electrode and potential - sensing AFM measurements in collaborative studies with y of Oregon . 3 5 However, Ni75 substantially diminished the PEC water oxidation performance of ED - hemat ite. A combination of electrochemical and spectroscopic methods, e.g. TEM and DWE, have revealed that the electrodeposited hematite electrodes are porous. 6 Subsequently, upon deposition of relatively thick catalyst (relative to the pore size) the catalyst is shunted to the conductive substrate. As the result, under illumin ation the catalyst is charged (oxidized) by the photogenerated holes from hematite and immediately discharged (recombination) by the electrons from the conductive substrate. This recombination process is induced by shunting which prevents the catalyst pote ntial to sufficiently be dropped to achieve sustainable water oxidation. 270 7.2. Future Directions Electrodeposition is a promising approach to directly deposit tantalum nitride (Ta 3 N 5 ) on a conductive substrate . In chapter 2, the electrodeposition of tantalum oxi de from aqueous solution was described. It was shown that the electro - generation of hydroxide ions at the surface of working electrode triggers the hydrolysis of tantalum precursor. Alternatively for electrodeposition of tantalum nitride, ammonia is electr ochemically generated via electrochemical reduction of ammonium (NH 4 + ), 7 where it immediately reacts with the tantalum precursor in solution and precipitate at the surface of working electrode. For successful electrodeposition of crystalline tantalum nitride without ammonolysis step, several criteria must be met: (1) the electro generation of ammonia must be more feasible than the reduction of Ta(V), (2) the reaction between electro - generated ammonia and Ta - precursor in solution must be spontaneous, (3) the product of the reacti on must be insoluble, and (4) in case of amorphous products it can should be nitridized to crystalline Ta 3 N 5 by annealing in an inert atmosphere (N 2 or Ar). Our preliminary observations consistent with previous studies 8 11 indicate that TaCl 5 (Ta - precursor) in solution or in the solid phase spontaneously reacts with ammonia gas in the room temperature which is accompanied by a color change from opaque white to a bright yellow precipitates according to equation 7 - 1 (as shown in Figure 7 1 .a ). This reaction is very clean and only ammonium chloride is formed as the byproduct which can be easily removed by sublimation under reduced pressure at > 170 °C. 8 The yellow intermediate is insoluble in most of the commonly used solvents (such as DCM, DMSO, DMF, and acetonitrile but reacts with alcohols such as methanol and ethanol) and irreversibly reacts with atmospheric moisture and produce a white powder (presumably tantalum oxide). Elemental analysis by EDS indicated that the yellow powder 271 is comprised of O, N, Ta, and Cl. The presence of oxygen atom (~ 10 atomic percent) precludes the determination of molecular formula of the yellow intermediate. 7 - 1 Upon annealing the yellow intermediate in an inert atmosphere of Ar at 800 °C, interestingly, it was converted to a bright orange/red color that is similar to the bandgap of Ta 3 N 5 , ( Figure 7 1 . a). The XRD pattern of the orange/red - powder is s hown in Figure 7 1 . b. The observed pattern is unambiguously assigned to crystalline Ta 3 N 5 and Ta 2 O 5 . The presence of Ta 2 O 5 can be attributed to the residual oxygen in Ar and/or to the formation of oxides during handling and storage of the yellow intermediate sample. Nonetheless, these experiments indicates that the ammonolysis step can effectively be eliminated and the crystalline phase of Ta 3 N 5 can b e synthesized by annealing the yellow intermediate in an inert atmosphere. This is an important observation as it confirms that this method provide a non - reactive atmosphere that allows to integrate Ta 3 N 5 with various substrates including FTO, and ITO that are otherwise u nstable under ammonolysis conditions. The cyclic voltammogram of the electrodeposition bath containing various components of electrodeposition bath prepared in an ionic liquid are shown in Figure 7 1 . c. As it can be seen, Ta(V) is highly susceptible to electroreduction and its reduction potential is ~ 200 mV more positive than the reduction potential of ammonium. One strategy to circumv ent this issue is to use alternative Ta - precursors or complexes with more negative reduction potentials. Alternatively, electrochemical generation of ammonia from other sources other than electroreduction of ammonium is another strategy that can be followe d. 272 Figure 7 1 . a) Schematic representation and photographs of the reaction products of TaCl 5 with ammonia and subsequent exposure to air and annealing in Ar, b) the XRD pattern of the powder after annealing in argon; the vertical blue dashed lines and red dashed - doted lines represent Bragg position of Ta 2 O 5 (PDF# 00 - 019 - 1299 of and Ta 3 N 5 (PDF# 00 - 019 - 1291 ), respectively, c) cyclic voltammograms of the electrodeposition bath containing various components; blank is solvent only ( 1 - butyl - 3 - methylimidazolium tetrafluoroborate ) (black), 30 mM TaCl 5 (red), and 60 mM NH 4 PF 6 (blue). The Ionic liquid of 1 - butyl - 3 - methylimidazolium tetrafluoroborate was synthesized using previously reported procedure. 1 2 The water impurity was removed by heating at 100 °C under reduced pressure for 48 hours using S chlenk line and subsequently it was stored in the glove box with the concentration of water and oxygen at ~ 1.2 and 3 ppm, respectively. The scan rate is 10 0 mV s - 1 and the Pt disc electrode was used as the working electrode. As discussed in chapter 1, one of the issue halting efficient water oxidation with Ta 3 N 5 is the surface oxidation as a result of photocorrosion under PEC water oxidation condition. In ch apter 4 w e showed that ALD provides a viable and robust technique to integrate crystalline Ta 3 N 5 with a transparent conductive substrate like FTO. Furthermore, prior to exposure to the ambien t atmosphere the ALD deposited Ta 3 N 5 electrodes can be in situ pr otected against photocorrosion with an over layer of GaN or TiO 2 . 273 As discussed in chapter 4, the PEC performance of tantalum nitride is controlled by the diffusion of photogenerated holes (the minority charge carrier). Galli and coworkers 13 have recently shown that the main mechanism for electron and hole transports in tantalum nitride is bandlike which is strongly affected by the effective mass of charge carriers. Moreover, theoretical studies have shown that the charge transport in Ta 3 N 5 s uffers from the large electron and hole effective masses. 14 DFT calculations were subsequently utilized to study the effect of substitutional impurities and strain on the charge transport characteristics of tantalum nitride. 13 It was predicted t hat substitutional impurities - such as Nb and V replacing Ta atoms and P replacing N atoms significantly reduces the effective mass of both electron (by 19%) and holes (by 39%). Therefore, substitutional doping and introducing strain are two plausible st rategies to improve the diffusion length of holes and electrons in Ta 3 N 5 and thus its PEC performance. As discussed in chapter 1, the band edge positions of tantalum nitride straddle both the water oxidation and water reduction reactions. However, most of the research in this area are focused on the water oxidation reaction with few reports only on the ph o tocatalytic hydrogen production. For example, Domen and coworkers 15 have shown that in presence of methanol as a hole scavenger Pt - loaded Ta 3 N 5 nanoparticles evolve hydrogen under visible light irradiations. In a separate study Seo et al. 16 have studied the PEC water oxidation and photocatalytic H 2 evolution (in presence of methanol) performances of Mg - and Zr - doped Ta 3 N 5 under visible light irradiation. They have shown that upon co - doping with Mg and Zr (25 at.%), the H 2 evolution rate has increased by a factor of 15 in comparison to the pure sample. The improvement in photocatalytic activity was attributed to the negative shift of band edge positions which provides a higher driving force for hydrogen ev olution reaction. While these examples indicate that Ta 3 N 5 can photocatalytically generate hydrogen, the applicability of Ta 3 N 5 as a photocathode for PEC hydrogen evolution is 274 yet to be demonstrated, however. The challenge in this area is to find a dopant to synthesis a p - type Ta 3 N 5 films. The two steps procedure involving spin coating of precursor solution on Ta - substrate followed by ammonolysis (developed in our lab) is a promising and fast method to screen various dopant with variable concertation to exp lore their effect(s) on the semiconducting, photophysic s , and photochemistry characteristics of Ta 3 N 5 . 275 R EFERENCES 276 REFERENCES (1) Liu, G.; Ye, S.; Yan, P.; Xiong, F.; Fu, P.; Wang, Z.; Chen, Z.; Shi, J.; Li, C. Enabling an Integrated Tantalum Nitride Photoanode to Approach the Theoretical Photocurrent Limit for Solar Water Splitting. Energy Environ. Sci. 2016 , 9 (4), 1327 1334. (2) Li, Y.; Zhang, L.; Torres - Pardo, A.; González - Calbet, J. M.; Ma, Y.; Oley nikov, P.; Terasaki, O.; Asahina, S.; Shima, M.; Cha, D.; et al. Cobalt Phosphate - Modified Barium - Doped Tantalum Nitride Nanorod Photoanode with 1.5% Solar Energy Conversion Efficiency. Nat. Commun. 2013 , 4 , 2566. (3) Hajibabaei, H.; Schon, A. R.; Hamann, T. W. Interface Control of Photoelectrochemical Water Oxidation Performance with Ni 1 - x Fe x O y Modified Hematite Photoanodes. Chem. Mater. 2017 , 29 (16), acs.chemmater.7b01149. (4) Qiu, J.; Hajibabaei, H.; Nellist, M. R.; Laskowski, F. A. L.; Hamann, T. W.; Boettcher, S. W. Direct in Situ Measurement of Charge Transfer Processes During Photoelectrochemical Water Oxidation on Catalyzed Hematite. ACS Cent. Sci. 2017 , 3 (9), 1015 1025. (5) Nellist, M. R.; Laskowski, F. A. L.; Qiu, J.; Hajibabaei, H.; Sivula, K.; Hamann, T. W.; Boettcher, S. W. Potential - Sensing Electrochemical Atomic Force Microscopy for in Operando Analysis of Water - Splitting Catalysts and Interfaces. Nat. Energy 2018 , 3 (1), 46 52. (6) Qiu, J.; Hajibabaei, H.; Nellist, M. R.; Laskowski, F. A. L.; Oener, S. Z.; Hamann, T. W.; Boettcher, S. W. Catalyst Deposition on Photoanodes: The Roles of Intrinsic Catalytic Activity, Catalyst Electrical Conductivity, and Semiconduct or Morphology. ACS Energy Lett. 2018 , 3 (4), 961 969. (7) Little, D. J.; Smith, III, M. R.; Hamann, T. W. Electrolysis of Liquid Ammonia for Hydrogen Generation. Energy Environ. Sci. 2015 , 8 (9), 2775 2781. (8) Moureu, H.; Hamblet, C. H. The Ammonolysis of Tantalum Pentachloride. J. Am. Chem. Soc. 1937 , 59 (1), 33 40. (9) Sasaki, Y.; Tokuyasu, Z.; Ono, Y.; Iwasaki, M.; Ito, S. Synthetic Conditions and Color Characteristics of Tantalum Oxynitride Prepared via Liquid - NH 3 Process. Adv. Mater. Sci. Eng. 2009 , 2009 (2), 1 4. (10) Sasaki, Y.; Tokuyasu, Z.; Ono, Y.; Iwasaki, M.; Ito, S. Synthetic Conditions and Color Characteristics of Tantalum Oxynitride Prepared via Liquid - NH 3 Process. Adv. Mater. Sci. Eng. 2009 , 2009 , 2 6. (11) Choi, D.; Kumta, P. N. Synthe sis, Structure, and Electrochemical Characterization of Nanocrystalline Tantalum and Tungsten Nitrides. J. Am. Ceram. Soc. 2007 , 90 (10), 3113 3120. 277 (12) Burrell, A. K.; Del Sesto, R. E.; Baker, S. N.; McCleskey, T. M.; Baker, G. A. The Large Scale Synthesis of Pure Imidazolium and Pyrrolidinium Ionic Liquids. Green Chem. 2007 , 9 (5), 449 454. (13) Morbec, J. M.; Galli, G. Charge Transport Properties of Bu lk Ta 3 N 5 from First Principles Juliana. Phys. Rev. B 2016 , 93 (3), 035201. (14) Morbec, J. M.; Narkeviciute, I.; Jaramillo, T. F.; Galli, G. Optoelectronic Properties of Ta 3 N 5 : A Joint Theoretical and Experimental Study. Phys. Rev. B 2014 , 90 (15), 155204 . (15) Hitoki, G.; Ishikawa, A.; Takata, T.; Kondo, J. N.; Hara, M.; Domen, K. Ta 3 N 5 as a Novel Visible Light - Chem. Lett. 2002 , 31 (7), 736 736. (16) Seo, J.; Takata, T.; Nakabayashi, M.; Hisatomi, T.; Shibata, N.; Mine gishi, T.; Domen, K. Mg - Zr Cosubstituted Ta 3 N 5 Photoanode for Lower - Onset - Potential Solar - Driven Photoelectrochemical Water Splitting. J. Am. Chem. Soc. 2015 , 137 (40), 12780 12783.