PROBING LIGAND EFFECTS FOR TITANIUM-CATALYZED HYDROAMINATION REACTION AND ISOLATION OF RARE EARTH METAL COMPLEXES By Rashmi Jena A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Chemistry – Doctor of Philosophy 2024 ABSTRACT The exploration and utilization of different ligands are crucial for design and development of superior catalysts and isolation of rare metal complexes. Ligands play a crucial role in stabilizing metal complexes with their diverse electronic and steric properties. Therefore, it is essential to understand metal-ligand interactions and electronic structures. Chapters 2 and 3 embark on optimizing ligand’s electronic and steric effects for accelerating titanium-catalyzed hydroamination reactions. By combining theoretical and experimental approaches, we could predict optimal ligand structures and properties for the development of new catalysts. Chapter 4 delves into mechanistic pathways for pyridine synthesis from isoxazole and enamine. Chapters 5, 6, and 7 emphasize understanding the metal-ligand interactions for isolating highly reactive and stable rare-earth metal complexes. Further, the unexpected reactivity of these complexes has led to the discovery of some intriguing and unprecedented metal complexes, enriching the field of organometallic chemistry. To My Family and Friends iii ACKNOWLEDGEMENTS This is probably the most challenging part of my thesis, and I want to acknowledge everyone who has helped me get to this career stage. I was fortunate to have an amazing PI and great friends who have made my journey exciting. First of all, I want to thank my supervisor, Prof. Aaron L. Odom, for his unwavering support, invaluable guidance, and relentless encouragement throughout the entirety of my research journey. Aaron, I want to thank you for responding to my very first email regarding asking for a vacancy in your group. Under your guidance, I have not only grown as a scientist but also as a person. This thesis is a testament to your profound impact on my academic and personal development. I am deeply fortunate to have had the privilege of learning and growing under your guidance. I cannot count how many times I have knocked on your door, and you have always responded, thank you for that. I am profoundly grateful to you for shaping me as a scientist and allowing me to pursue my interests with gratitude. I also want to express my gratitude to my committee members, Prof. Selvan Demir, Prof. Xuefei Huang, and Prof. Thomas W. Hamann. Their invaluable insights and guidance during my committee meetings and comprehensive exam have significantly enriched my research. I am particularly grateful to Prof. Demir for our fruitful collaboration over the past five years. I want to thank Prof. McCracken for his assistance with EPR spectroscopy and for sharing stories from his grad school and the department itself. I have enjoyed working with you. Along those same lines, I would like to thank Dr. Richard J. Staples for always being there for me when it comes to solving crystal disorders or my conflicting thoughts. I am very grateful to you for sharing your knowledge and always cheering me up. Besides that, you have been a true well- wisher for me, always advising me when I go to you with a problem. I would also like to thank Dr. iv Daniel Holmes for all his help with my kinetics, SST, and isocyanide experiments. You have always been there to help me in the sub-basement. I don't think I would have pursued graduate school without the influence of mentors I had during my master's at the Indian Institute of Technology, Delhi. I extend my deepest gratitude to Prof. Selvarajan Nagendran for his guidance and encouragement, which ignited my passion for organometallic chemistry. Thanks to Dr. Darmendra Singh and Dr. Pritam Mahawar for their teachings and insightful mentorship, which have shaped my academic journey and inspired me to pursue my PhD in this field. I sincerely thank previous Odom group members, Dr. Zhilin Hou and Dr. Seokjoo Lee, for their support and discussions on the office board. They both were great mentors to me and created a very welcoming atmosphere in the group. Further, I want to thank Dr. Tanner J. McDaniel, Dr. Brennen S. Billow, and Dr. Kelly E. Aldrich, who have never overlapped with me in the Odom group but still replied to my emails and calls whenever I had questions about their expertise. I am very grateful to our current Odom group, especially Sazia and Gunner, for helping me during my final year of grad school. Thank you for chatting over tea and helping me with my thesis. I will always be grateful for that. To my friends a few rooms away, Ernesto, Florian, TJ, and Saroshan. Ernesto, I enjoyed talking to you about everything. You are such a great person/scientist to be around. I am grateful for all your help and for listening to my blabbering sometimes. TJ and Florian, I am very thankful for what I have learned from both of you. And Saroshan, thank you for your patience and help with the blender; I wouldn’t have a journal cover if you didn’t help me. Thank you all for your patience and support throughout my grad school career. Shahrukh, Morgan, Sneha, Jishnu, Abhishek, Asmita, Ananya, Hannah, and Mo thank you for v being such a great friend and believing in me. In addition, I would like to thank Akhil for helping me achieve my dream of pursuing a Ph.D. I wouldn’t be here without your support, and I am very grateful for all the countless ways you have supported me over the last several years. I need to thank my roommates in grad school. Kriti, Swati, Priya, and Raghvi you guys have been my home for the last 5 years. I enjoyed having intense conversations with you about work, geopolitics, and secularism. To my best friend back home, Roshni, I am blessed to have a friend like you. She has been my constant support throughout my career, always cheering me up and motivating me to believe in myself. Thank you so much for always being there for me whenever I need you. Lastly, but most importantly, my family. Nothing that I have achieved would have been possible without them. They have made a huge sacrifice to help me achieve my dreams. I am lucky to have my parent's support in all the decisions I have made so far. There were definitely more people who contributed to my journey. If I forgot your name, I am sorry about that. At the end, I would like to say thank you to everyone who has helped me in any way. vi TABLE OF CONTENTS LIST OF ABBREVIATIONS AND SYMBOLS ....................................................................... ix CHAPTER 1: IMPORTANCE OF LIGANDS IN DESIGNING CATALYSTS ANDSTABILIZING RARE-EARTH METAL COMPLEXES................................................ 1 1.1 Ligands Dictating Metal’s Reactivity ........................................................................................1 1.2 Ligands Development for Titanium-Catalyzed Hydroamination Reactions .............................1 1.3 Role of Ligands in Isolation of Rare Group 3 and Lanthanides Complexes .............................6 1.4 Choice of Ligand for Single-Molecule Magnet Behavior .........................................................9 1.5 Conclusions ..............................................................................................................................12 REFERENCES ..............................................................................................................................13 CHAPTER 2: MODELING UNSYMMETRICAL LIGANDS FOR TITANIUM- CATALYZED HYDROAMINATION REACTIONS*........................................................... 17 2.1 Introduction ..............................................................................................................................17 2.2 Ligand Donor Parameter (LDP) and Percent Buried Volume (%Vbur) ....................................20 2.3 Synthesis of Unsymmetrical Ligands and Titanium Catalysts ................................................27 2.4 Ligand 1- 5 Isomerization ...................................................................................................33 2.5 Collection of Kinetics Data for 10 New Titanium Catalysts ...................................................37 2.6 Model Consideration for Unsymmetrical Ligands ..................................................................40 2.7 Conclusions ..............................................................................................................................47 2.8 Experimental Details ................................................................................................................50 REFERENCES ............................................................................................................................107 CHAPTER 3: AN INDOLE-EFFECT FOR INCREASING THE RATE OF HYDROAMINATION REACTION USING HIGH-VALENT METAL CATALYSIS* ............................................................................................................................ 110 3.1 Introduction ............................................................................................................................110 3.2 Synthesis of 3-Unsubstituted Indolyl-Based Unsymmetrical Ligands and Titanium Precatalyst ....................................................................................................................................114 3.3 Kinetics Data, Model Progression, and Discussions .............................................................117 3.4 Efficiency 3-Unsubstitutedindolyl based Titanium Precatalyst .............................................136 3.5 Synthesis of Chromium Complexes for LDP and %Vbur Measurements ..............................141 3.6 Conclusions ............................................................................................................................146 3.7 Experimental Details ..............................................................................................................148 REFERENCES ............................................................................................................................252 CHAPTER 4: MECHANISTIC STUDIES OF PYRIDINE SYNTHESIS FROM ISOXAZOLE* ........................................................................................................................... 254 4.1 Introduction ............................................................................................................................254 4.2 Insights into the Reaction Mechanism through DFT Analysis ..............................................255 4.3 Effects of Solvent on the Mechanistic Pathway ....................................................................259 4.4 Conclusions ............................................................................................................................266 vii REFERENCES ............................................................................................................................267 CHAPTER 5: SYNTHESIS, STRUCTURE, AND PROPERTIES OF A RARE YTTRIUM ISOCYANIDE COMPLEX FROM UNPRECEDENTED DIVALENT YTTRIUM COMPLEX* .......................................................................................................... 268 5.1 Introduction ............................................................................................................................268 5.2 Synthesis of Potassium Terphenyl Amide Ligand System ....................................................271 5.3 Synthesis and Properties of Y(NHAr*)2Cl ............................................................................272 5.4 Generation and Characterization of a Novel Neutral Y(II) Complex ....................................274 5.5 Reactivity of Y(NHAr*)2 (2) and Discovery of Yttrium Isocyanide Y(NHAr*)–NC (3) .....283 5.6 Conclusions ............................................................................................................................290 5.7 Experimental details...............................................................................................................292 REFERENCES ............................................................................................................................336 CHAPTER 6: SYNTHESIS, STRUCTURE, AND REACTIVITY OF DYSPROSIUM COMPLEXES* ......................................................................................................................... 341 6.1 Introduction ............................................................................................................................341 6.2 Synthesis and Characterization of Dy(NHAr*)2Cl Complex ................................................343 6.3 Synthesis and Characterization of a Neutral Dy(II) Complex ...............................................346 6.4. Synthesis of Characterization of Dysprosium(III) Cation ....................................................352 6.5 Reactivity of Dy(NHAr*)2 (5) and Discovery of Dysprosium Isocyanide Dy(NHAr*)–NC (7) .....................................................................................................................356 6.6 Reactivity of Dy(NHAr*)2 (5) with Elemental Phosphorous (P4) .........................................358 6.7 Conclusions ............................................................................................................................366 6.8 Experimental Details ..............................................................................................................368 REFERENCES ............................................................................................................................409 CHAPTER 7: CHARGE-CONTROLLED REACTIVITY OF YTTRIUM COMPLEXES ........................................................................................................................... 415 7.1 Motivation and Background ..................................................................................................415 7.2 Approach to Modification in Metal Complexes ....................................................................415 7.3 Conclusions ............................................................................................................................422 7.4 Experimental Details ..............................................................................................................423 REFERENCES ............................................................................................................................435 viii LDP %Vbur SST DFT NBO G16 NRT NPA MO Equiv. GC-MS NMR DCM THF BArF24 EA H2dpm SQuID HOMO LUMO Calcd LAH LIST OF ABBREVIATIONS AND SYMBOLS Ligand Donor Parameter Percent Buried Volume Spin Saturation Transfer Density Functional Theory Natural Bond Orbital Gaussian 16 Natural Resonance Theory Natural Population Analysis Molecular Orbital Equivalents Gas Chromatography-Mass Spectrometry Nuclear Magnetic Resonance (Spectroscopy) Dichloromethane Tetrahydrofuran Tetrakis(3,5-(trifluoromethyl)phenyl)borane Elemental Analysis 5,5-dimethyldipyrrolylmethane Superconducting Quantum Interference Device Highest Occupied Molecular Orbital Lowest Unoccupied Molecular Orbital Calculated Lithium Aluminum Hydride ix θ χ ν ηn Å δ χMT μeff °C h μB s s-1 Tolman’s Cone Angle Tolman’s Electronic Parameter Stretching Frequency Hapticity of n Angstrom Chemical shift in Nuclear Magnetic Resonance Spectroscopy Magnetic Susceptibility Effective Magnetic Moment Degree Celsius Hour(s) Bohr Magnetons Seconds Unit of wavenumber/first-order rate constant x CHAPTER 1: IMPORTANCE OF LIGANDS IN DESIGNING CATALYSTS AND STABILIZING RARE-EARTH METAL COMPLEXES 1.1 Ligands Dictating Metal’s Reactivity Ligands play a crucial role in determining the reaction chemistry of metal complexes. By coordinating with the metal center, they exhibit unique properties and reactivities that are not found in the isolated metal species. This unique potential of ligands is exemplified by the fact that platinum, which is not a chemotherapeutic agent on its own, becomes one when incorporated into specific compounds known as cisplatin, carboplatin, and oxaliplatin, revolutionizing cancer treatment.1–3 Further, the efficiency and selectivity of a catalytic process can be significantly influenced by the choice of ligand around the metal center. For example, TiCl4 is a poor catalyst for intermolecular hydroamination reactions of alkyne and amine.4 However, Ti(NMe2)4 is a much more effective catalyst for the same reaction.5 This demonstrates how the choice of ligands around the metal center significantly dictates the metal's reactivity toward specific reactions. Ligands' ability to tune the properties of metal centers makes them indispensable in fundamental research and practical applications. Therefore, understanding metal-ligand interactions is pertinent for innovation and advancement in science and technology. 1.2 Ligands Development for Titanium-Catalyzed Hydroamination Reactions The hydroamination of alkene and alkynes has been extensively studied to make C‒N bonds (Figure 1.1) efficiently. The imines and amines obtained after this transformation can be used to make a variety of industrially and pharmaceutically relevant heterocycles.6 The direct addition of amine into alkene or alkyne has a high activation barrier owing to the electrostatic repulsion 1 between electron-rich alkenes or alkynes and the lone pair of amines.7 Further, the addition reaction of amine and alkene is symmetry forbidden and thus unfavorable (Figure 1.1).8 The late transition metal complexes,9–12 group 4 metal complexes,13,14 and actinides complexes15 have been utilized as catalysts to overcome these limitations and make hydroamination of alkene and alkyne more feasible at lower temperatures. Figure 1.1. (top) Hydroamination reaction of alkynes and alkenes. (bottom) Addition of amine into alkene is symmetry forbidden.8 Although titanium is the second most abundant transition metal, titanium catalysts for hydroamination were not explored until the late 1990s. In 1990, Rothwell and coworkers reported the first example of a titanium-catalyzed hydroamination reaction. They showed the bis(phenylamide)titanium complex could catalyze a reaction of aniline and 3-hexyne to produce 2 N-phenylimine of 3-hexanone.16 1n 1999, Doye reported regioselective hydroamination of various alkynes and amines using a dimethyltitanocene catalyst.17 At the time, our group was also exploring the titanium-catalyzed hydroamination reactions and discovered that a commercially available Ti(NMe2)4 serves as an effective catalyst for hydroamination of alkynes.5 Not only is Ti(NMe2)4 more stable and easily accessible, but it also demonstrates superior reactivity toward hydroamination reactions in many cases. For instance, Ti(NMe2)4 gave a shorter reaction time for terminal alkynes and showed some interesting functional group tolerance. While Doye’s catalyst gave the selectivity for the anti-Markovnikov product (B, Figure 1.1, R3 = H), our catalyst Ti(NMe2)4 gave Markovnikov product (A, Figure 1.1, R3 = H). At the time, very little was known about the effects of altering the ancillary ligand on the metal’s reactivity and reaction’s regioselectivity. Our group has been interested in finding new ancillary ligands and studying their impact on hydroamination and iminoamination reactions. The rate of hydroamination can be increased by enhancing the Lewis acidity of the metal. This was achieved by replacing dimethylamide with a pyrrolyl ligand, as pyrrolyl ligands are relatively poor donors compared to dimethylamide. In 2001, our group reported di(pyrrolyl-α-methyl)methylamine (H2dpma) ancillary ligand for titanium catalysis.13 Unlike Ti(NMe2)4, where all ligands are protolytically labile, H2dpma stays intact throughout the catalysis, reflected in their different reactivity. For hydroamination of 1-hexyne and aniline, Ti(dpma)(NMe2)2 shows >50:1 selectivity for Markovnikov’s product as compared to Ti(NMe2)4, which gives 3:1 Markovnikov’s to anti- Markovnikov’s product. Further, Ti(dpma)(NMe2)2 exhibited great reactivity towards alkylamines and demonstrated tolerance for various groups. Inspired by Bergman’s zirconocene-based catalysis,18 the precatalyst is proposed to interact with an amine first, forming the active catalyst titanium(IV) imido complex. This enters the 3 catalytic cycle and undergoes [2+2]-cycloaddition with the alkyne in a reversible reaction. The slow step in the catalytic cycle is believed to be the protonolysis of the Ti‒C bond. In the final step, proton transfer releases the enamine product and regenerates the active catalyst. Figure 1.2. Proposed mechanism for titanium-catalyzed hydroamination reaction. Furthermore, in 2003, our group published the use of dipyrrolylmethane derivatives of titanium catalysts Ti(dpm)(NMe2)2.19 This was believed to show better catalytic activity because of increased Lewis acidity of the metal and less steric constraints around the metal center compared to Ti(dpma)(NMe2)2. This has indeed been confirmed; the hydroamination of 1-hexyne with aniline now takes only 5 minutes at room temperature. Generally, Ti(dpm)(NMe2)2 is a more active catalyst and requires a shorter reaction time. Moreover, this catalyst also showed monohydroamination of 1,4- and 1,5-diynes which then undergo cyclization to yield respective pyrroles.20 This methodology to make pyrrole is carried out directly from alkynes and thus atom economical as opposed to Paal‒Knorr pyrrole synthesis which uses 1,4-dicarbonyl compounds.21 4 Catalyst Cp2TiMe2 Ti(NMe2)2 Ti(dpma)(NMe2)2 Ti(dpm)(NMe2)2 Structure Loading 3 mol%a 10 mol% 10 mol% 5 mol% Temp, Time 100 °C, 72 ha 75 °C, 2 h 75 °C, 6 h 25 °C, 5 mins M:anti-M 3:1 aData for 1-dodecyne (1-hexyne not reported). No data >50:1 M selective Figure 1.3. Comparison of various titanium catalysts for hydroamination of 1-hexyne. So far, we have seen how changing ligands on a metal can significantly affect the rate and selectivity (Figure 1.3). Chemical intuition is invaluable, but quantitative parameters for ancillary ligand effects are a more reliable and efficient tool for designing catalysts. Chadwick Tolman developed two experimentally-derived parameters (electronic and steric) for late transition metals.22 Inspired by this, the Odom group reported similar electronic and steric parameters for quantifying ancillary ligand effects on high-valent transition metals in 2012.23 The electronic parameter Ligand Donor Parameter (LDP) describes the ability of a ligand to donate electron density to a high-valent metal center. In NCr(NiPr2)2X system, the enthalpic barrier to rotation of the Cr‒N bond of the diisopropylamido ligand gave an LDP value where X is the ligand under consideration. Sterics of the ligands were measured using the concept of percent buried volume (%Vbur).24 5 Figure 1.4. NCr(NiPr2)2X system used to measure LDP values (X = ligand under consideration). In 2017, the Odom group showed the first application of LDP by employing a correlation between electronic and steric parameters to predict the rate of the alkyne hydroamination reactions. This study included symmetric bidentate ligands based on pyrrolyl, indolyl, and phenoxide. However, it didn’t result in a faster catalyst than previously known Ti(dpm)(NMe2)2.25 Chapters 2 and 3 provide detailed descriptions of how unsymmetrical ligands led to the discovery of faster catalysts, along with valuable mechanistic insights. Density Functional Theory (DFT) is a crucial tool due to its predictive power. Throughout this thesis, we have used DFT calculations to predict molecular geometry, electronic properties, and stability of molecules to enhance the reliability of experimental findings. In Chapter 4, we used DFT calculations to understand the reaction pathway for pyridine synthesis using isoxazole and enamine.26 Here, we proposed a plausible reaction mechanism based on experimental results and DFT calculations. 1.3 Role of Ligands in Isolation of Rare Group 3 and Lanthanides Complexes Lanthanides (Ln) have widespread applications in the fields of catalysis, material science, and medical diagnostics.27 A recent report from market.us has shown that the global rare-earth market is expected to expand at a compound annual growth rate of 10.4% between 2023-2032.28,29 Metal oxides such as cerium, lanthanum, and neodymium oxides are widely used in magnets, metal alloys, catalysts, and other applications (Figure 1.5). Understanding and exploring the reaction 6 chemistry of lanthanides can lead to significant progress, driving innovations and improving quality of life. Figure 1.5. (top) Uses of rare earth elements based on data collected in 2016.27,28 (bottom) Global rare earth metals market predictions.29 7 Understanding a metal’s accessible oxidation states is critical because it significantly impacts reactivity, chemical bonding, and potential applications. The most common oxidation state for lanthanides and group 3 metals is +3, irrespective of the number of f-electrons. Firstly, it was believed that only samarium, europium, and ytterbium could be accessible in the +2 oxidation state. However, cyclopentadienyl (Cp) ligands have enabled access to the +2 oxidation state for the rest of the lanthanides except radioactive promethium (Figure 1.6).30–32 Figure 1.6. Typical structure of Ln(II) and Ac(II) complexes stabilized by Cp ligands. The steric saturation provided by three Cp ligands is suitable for the stabilization of the larger ionic radius of the lanthanides. The use of Cp ligands has been pivotal not only for lanthanides but also for actinides (Ac), such as uranium and thorium.33,34 Even though the group 3 and lanthanide complexes in +2 oxidation states have been isolated, they are not very stable at room temperature. For instance, the most stable divalent yttrium complex lasts only 48 hours at room temperature, making studying their reactivity very challenging.35 Consequently, very little is known about their chemical reactivity. In 2018, our group isolated a rare example of a divalent uranium complex using a terphenyl- based amide ligand system. The large size of the ligand provides steric saturation around the metal center and electronic stabilization through the aryl rings. The neutral metal complex obtained is 8 relatively stable at room temperature. Chapters 5 and 6 discuss the isolation of divalent yttrium and dysprosium complexes with their interesting reactivity towards small molecules. Figure 1.7. The bulky triaryl amide ligand used by our group to isolate rare metal complexes. 1.4 Choice of Ligand for Single-Molecule Magnet Behavior Modern data storage devices such as hard drives use ferro- or ferrimagnetic nanoparticles to store information in binary form (a combination of 1s and 0s). These particles are typically around tens of nanometers in size and behave like tiny magnets below a specific temperature called the Curie temperature (Tc). Below Tc, each particle aligns with the applied external magnetic field and creates a strong magnetic moment. The magnetic moments remain aligned even after removing an external magnetic field at a specific temperature (blocking temperature TB), resulting in a memory effect. The thermal energy below TB is insufficient to overcome the energy barrier for spin reversal (Ueff). In these particles, the direction of magnetization (aligned for “1” and opposite for “0”) is used to store data. High-density data storage devices can be achieved by reducing the size of these nanoparticles. Now the question arises: how small can they be while retaining their magnetic properties? Recently, Colacio and coworkers in their comprehensive review article discuss the relationship between nanoparticle size and the width of the hysteresis loop. Figure 1.8 shows if the nanoparticles are smaller than about 10 nm, their hysteresis loop closes, leading to a loss in magnetization.36 9 Figure 1.8. Correlation between particle size and coercive field.36 Single-molecule magnets (SMMs) are molecules that exhibit a bistable ground state, which allows them to retain magnetic moments in the absence of an external magnetic field. Instead of nanoparticles, a single molecule can exhibit a similar memory effect, thus at least a magnitude smaller in size. This property makes them exciting candidates for high-density data storage devices, quantum computing, and molecular spintronics. Lanthanides are exceptionally well suited for synthesizing SMMs because of their high magnetic anisotropy arising from their considerable unquenched orbital angular momentum. The 4f electrons are effectively shielded by outer 5s and 5p electron shells. This shielding reduces the influence of the external crystal field, allowing spin- orbit coupling to dominate. Dysprosium exhibits one of the largest magnetic moments in the periodic table. Dy(III) is a Kramer’s ion with a bistable ground state irrespective of the ligand field. This is one of the prerequisites for molecular magnetism. Secondly, Dy(III)-based SMMs have larger Ueff due to high magnetic anisotropy. Although the ligand’s crystal field doesn’t impact the electronic structure of the lanthanide ion, it does play a crucial role in modulating the magnetic anisotropy of the lanthanide ion. For instance, the figure below shows that the magnetic anisotropy of the oblate 10 Dy(III) ion can be enhanced by using the axial ligand field below and above the xy plane. However, the magnetic anisotropy decreases drastically if ligands are placed in the equatorial direction (Figure 1.9).37 Figure 1.9. Effect of ligand environment on anisotropy for the oblate metal ion.36 The strategic design of ligands is essential for advancing the field of molecular magnetism. Layfield and coworkers in their seminal work reported the best SMM known to date.38 Their [(CpCp*)Dy]+ cation as a salt where Cp = pentaisopropylcyclopentadienyl, Cp* = pentamethylcyclopentadienyl shows magnetic hysteresis up to liquid N2 temperature. Here the Cp‒ Dy‒Cp angle is 162.507(1)°, slightly deviating from perfectly linear geometry and providing a strong axial field to the metal center. After discovering this record-breaking SMM in 2018, continuous efforts to develop a better SMM are ongoing. However, no successful attempts have been made to break the record TB = 80 K for a mononuclear dysprosium complex. 11 Recently, a different strategy has been employed by using two monoanionic ligands with minimal equatorial coordination. Mills and coworkers have shown the use of a bis(amide) ligand system to isolate bent dysprosium metallocene.39 However, this molecule showed poorer SMM behavior than predicted values due to magnetic relaxation through quantum tunneling. This study focuses on the importance of the rigidity of coordinated ligands in attaining high-blocking temperatures. The flexible ligand environment could lead to magnetic relaxation even in a strong crystal field.39,40 Dysprosium systems with amide ligands and arenes are pretty rare.41–43 The best yet discovered features a see-saw geometry using a bisanilide terphenyl ligand with a TB = 5.8 K.43 Chapter 6 shows the synthesis, structure, and magnetic properties of pseudo-low-coordinate bisamide dysprosium complexes. The ligand system with amide and arene has not yet been explored in the SMM realm. We attempted to convert our amide ligand into an imide to correlate their magnetic properties. It may or may not give us a better SMM, but it will allow us to understand the metal-ligand interactions and their effect on magnetism. Despite multiple attempts, we were unsuccessful. However, during these attempts, we isolated some interesting metal complexes. DFT calculations allowed us to uncover the charge-controlled reactivity of these complexes. Chapter 7 discusses the synthetic strategies and DFT calculations to understand these reactive metal complexes. 1.5 Conclusions This thesis focuses on understanding the influence of ligands on high-valent metal catalysis. With the design principles of the ligands, a systematic study is presented here to enhance the efficiency of titanium catalysts for the hydroamination of alkynes. Furthermore, the choice of proper ligand environment is crucial for the isolation of divalent rare-earth metal complexes. Here we discuss the syntheses, properties, and reactivities of these metal complexes. 12 REFERENCES (1) Corte-Rodríguez, M.; Espina, M.; Sierra, L. M.; Blanco, E.; Ames, T.; Montes-Bayón, M.; Sanz-Medel, A. Quantitative Evaluation of Cellular Uptake, DNA Incorporation and Adduct Formation in Cisplatin Sensitive and Resistant Cell Lines: Comparison of Different Pt- Containing Drugs. Biochemical Pharmacology 2015, 98, 69–77. (2) Calderon, L. E.; Keeling, J. K.; Rollins, J.; Black, C. A.; Collins, K.; Arnold, N.; Vance, D. E.; Ndinguri, M. W. Pt-Mal-LHRH, a Newly Synthesized Compound Attenuating Breast Cancer Tumor Growth and Metastasis by Targeting Overexpression of the LHRH Receptor. Bioconjugate Chem. 2017, 28, 461–470. (3) Zayed, A.; Jones, G. D. D.; Reid, H. J.; Shoeib, T.; Taylor, S. E.; Thomas, A. L.; Wood, J. P.; Sharp, B. L. Speciation of Oxaliplatin Adducts with DNA Nucleotides. Metallomics 2011, 3, 991. (4) Ackermann, L. TiCl 4 -Catalyzed Intermolecular Hydroamination Reactions. Organometallics 2003, 22, 4367–4368. (5) Shi, Y.; Ciszewski, J. T.; Odom, A. L. Ti(NMe 2 ) 4 as a Precatalyst for Hydroamination of Alkynes with Primary Amines. Organometallics 2001, 20, 3967–3969. (6) Müller, T. E.; Hultzsch, K. C.; Yus, M.; Foubelo, F.; Tada, M. Hydroamination: Direct Addition of Amines to Alkenes and Alkynes. Chem. Rev. 2008, 108, 3795–3892. (7) Pohlki, F.; Doye, S. The Catalytic Hydroamination of Alkynes. Chem. Soc. Rev. 2003, 32, 104– 114. (8) Steinborn, D.; Taube, R. Zur Komplexkatalyse Der Aminomethylierung Und Aminierung von Olefinen. Zeitschrift fuer Chemie 1986, 26, 349–359. (9) Åkermark, B.; Bäckvall, J. E.; Hegedus, L. S.; Zetterberg, K.; Siirala-Hansén, K.; Sjöberg, K. to Isolated Double Bonds. Journal of Palladium-Promoted Addition of Amines Organometallic Chemistry 1974, 72, 127–138. (10) Qian, H.; Widenhoefer, R. A. Platinum-Catalyzed Intermolecular Hydroamination of Vinyl Arenes with Carboxamides. Org. Lett. 2005, 7, 2635–2638. (11) Burling, S.; Field, L. D.; Messerle, B. A. Hydroamination of Alkynes Catalyzed by a Cationic Rhodium(I) Complex. Organometallics 2000, 19, 87–90. (12) Field, L. D.; Messerle, B. A.; Vuong, K. Q.; Turner, P. Intramolecular Hydroamination with Rhodium(I) and Iridium(I) Complexes Containing a Phosphine−N-Heterocyclic Carbene Ligand. Organometallics 2005, 24, 4241–4250. (13) Cao, C.; Ciszewski, J. T.; Odom, A. L. Hydroamination of Alkynes Catalyzed by a Titanium Pyrrolyl Complex. Organometallics 2002, 21, 5148–5148. 13 (14) Hou, Z.; Jena, R.; McDaniel, T. J.; Billow, B. S.; Lee, S.; Barr, H. I.; Odom, A. L. Modeling Complex Ligands for High Oxidation State Catalysis: Titanium Hydroamination with Unsymmetrical Ligands. ACS Catal. 2024, 5531–5538. (15) Li, Y.; Marks, T. J. Diverse Mechanistic Pathways and Selectivities in Organo-f-Element- Catalyzed Hydroamination. Intermolecular Organolanthanide-Catalyzed Alkyne and Alkene Hydroamination. Organometallics 1996, 15, 3770–3772. (16) Hill, J. E.; Profilet, R. D.; Fanwick, P. E.; Rothwell, I. P. Synthesis, Structure, and Reactivity of Aryloxo(Imido)Titanium Complexes. Angew. Chem. Int. Ed. Engl. 1990, 29, 664–665. (17) Haak, E.; Bytschkov, I.; Doye, S. Intermolecular Hydroamination of Alkynes Catalyzed by Dimethyltitanocene. Angew. Chem. Int. Ed. 1999, 38, 3389–3391. (18) Walsh, P. J.; Baranger, A. M.; Bergman, R. G. Stoichiometric and Catalytic Hydroamination of Alkynes and Allene by Zirconium Bisamides Cp2Zr(NHR)2. J. Am. Chem. Soc. 1992, 114, 1708–1719. (19) Shi, Y.; Hall, C.; Ciszewski, J. T.; Cao, C.; Odom, A. L. Titanium Dipyrrolylmethane Derivatives: Rapid Intermolecular Alkyne Hydroamination. Chem. Commun. 2003, No. 5, 586–587. (20) Ramanathan, B.; Keith, A. J.; Armstrong, D.; Odom, A. L. Pyrrole Syntheses Based on Titanium-Catalyzed Hydroamination of Diynes. Org. Lett. 2004, 6, 2957–2960. (21) Amarnath, V.; Anthony, D. C.; Amarnath, K.; Valentine, W. M.; Wetterau, L. A.; Graham, D. G. Intermediates in the Paal-Knorr Synthesis of Pyrroles. J. Org. Chem. 1991, 56, 6924–6931. (22) Tolman, C. A. Steric Effects of Phosphorus Ligands in Organometallic Chemistry and Homogeneous Catalysis. Chem. Rev. 1977, 77, 313–348. (23) DiFranco, S. A.; Maciulis, N. A.; Staples, R. J.; Batrice, R. J.; Odom, A. L. Evaluation of Donor and Steric Properties of Anionic Ligands on High Valent Transition Metals. Inorg. Chem. 2012, 51, 1187–1200. (24) Falivene, L.; Credendino, R.; Poater, A.; Petta, A.; Serra, L.; Oliva, R.; Scarano, V.; Cavallo, L. SambVca 2. A Web Tool for Analyzing Catalytic Pockets with Topographic Steric Maps. Organometallics 2016, 35, 2286–2293. (25) Billow, B. S.; McDaniel, T. J.; Odom, A. L. Quantifying Ligand Effects in High-Oxidation- State Metal Catalysis. Nature Chem 2017, 9, 837–842. (26) Lee, S.; Jena, R.; Odom, A. L. Substituted Pyridines from Isoxazoles: Scope and Mechanism. Org. Biomol. Chem. 2022, 20, 6630–6636. (27) Goodenough, K. M.; Wall, F.; Merriman, D. The Rare Earth Elements: Demand, Global Resources, and Challenges for Resourcing Future Generations. Nat Resour Res 2018, 27, 201– 216. 14 (28) Roskill. Rare Earths: Global Idustry, Markets and Outlook Report, 2016. (29) Global Rare Earth Metals Market By Type (Cerium Oxide, Lanthanum Oxide, Neodymium Oxide, Samarium Oxide, and Other Types) By Application (Magnet, Metals Alloys, Polishing, Catalysts, Glass & Ceramics, and Other Applications), By Region and Companies - Industry Segment Outlook, Market Assessment, Competition Scenario, Trends, and Forecast 2023- 2032. Market.US. 2023. (30) MacDonald, M. R.; Bates, J. E.; Fieser, M. E.; Ziller, J. W.; Furche, F.; Evans, W. J. Expanding Rare-Earth Oxidation State Chemistry to Molecular Complexes of Holmium(II) and Erbium(II). J. Am. Chem. Soc. 2012, 134, 8420–8423. (31) MacDonald, M. R.; Bates, J. E.; Ziller, J. W.; Furche, F.; Evans, W. J. Completing the Series of +2 Ions for the Lanthanide Elements: Synthesis of Molecular Complexes of Pr 2+ , Gd 2+ , Tb 2+ , and Lu 2+. J. Am. Chem. Soc. 2013, 135, 9857–9868. (32) Evans, W. J. Tutorial on the Role of Cyclopentadienyl Ligands in the Discovery of Molecular Complexes of the Rare-Earth and Actinide Metals in New Oxidation States. Organometallics 2016, 35, 3088–3100. (33) MacDonald, M. R.; Fieser, M. E.; Bates, J. E.; Ziller, J. W.; Furche, F.; Evans, W. J. Identification of the +2 Oxidation State for Uranium in a Crystalline Molecular Complex, [K(2.2.2-Cryptand)][(C 5 H 4 SiMe 3 ) 3 U]. J. Am. Chem. Soc. 2013, 135, 13310–13313. (34) Langeslay, R. R.; Fieser, M. E.; Ziller, J. W.; Furche, F.; Evans, W. J. Synthesis, Structure, and Reactivity of Crystalline Molecular Complexes of the {[C 5 H 3 (SiMe 3 ) 2 ] 3 Th} 1− Anion Containing Thorium in the Formal +2 Oxidation State. Chem. Sci. 2015, 6, 517–521. (35) Moehring, S. A.; Miehlich, M.; Hoerger, C. J.; Meyer, K.; Ziller, J. W.; Evans, W. J. A Room- Temperature Stable Y(II) Aryloxide: Using Steric Saturation to Kinetically Stabilize Y(II) Complexes. Inorg. Chem. 2020, 59, 3207–3214. (36) Zabala-Lekuona, A.; Seco, J. M.; Colacio, E. Single-Molecule Magnets: From Mn12-Ac to Dysprosium Metallocenes, a Travel in Time. Coordination Chemistry Reviews 2021, 441, 213984. (37) Rinehart, J. D.; Long, J. R. Exploiting Single-Ion Anisotropy in the Design of f-Element Single-Molecule Magnets. Chem. Sci. 2011, 2, 2078. (38) Guo, F.-S.; Day, B. M.; Chen, Y.-C.; Tong, M.-L.; Mansikkamäki, A.; Layfield, R. A. Magnetic Hysteresis up to 80 Kelvin in a Dysprosium Metallocene Single-Molecule Magnet. Science 2018, 362, 1400–1403. (39) Emerson-King, J.; Gransbury, G. K.; Whitehead, G. F. S.; Vitorica-Yrezabal, I. J.; Rouzières, M.; Clérac, R.; Chilton, N. F.; Mills, D. P. Isolation of a Bent Dysprosium Bis(Amide) Single- Molecule Magnet. J. Am. Chem. Soc. 2024, jacs.3c12427. 15 (40) Goodwin, C. A. P.; Ortu, F.; Reta, D.; Chilton, N. F.; Mills, D. P. Molecular Magnetic Hysteresis at 60 Kelvin in Dysprosocenium. Nature 2017, 548, 439–442. (41) Harriman, K. L. M.; Brosmer, J. L.; Ungur, L.; Diaconescu, P. L.; Murugesu, M. Pursuit of Record Breaking Energy Barriers: A Study of Magnetic Axiality in Diamide Ligated Dy III Single-Molecule Magnets. J. Am. Chem. Soc. 2017, 139, 1420–1423. (42) Yang, K.; Sun, R.; Zhao, J.; Deng, C.; Wang, B.; Gao, S.; Huang, W. A Combined Synthetic, Magnetic, and Theoretical Study on Enhancing Ligand-Field Axiality for Dy(III) Single- Molecule Magnets Supported by Ferrocene Diamide Ligands. Inorg. Chem. 2023, 62, 9892– 9903. (43) Harriman, K. L. M.; Murillo, J.; Suturina, E. A.; Fortier, S.; Murugesu, M. Relaxation Dynamics in See-Saw Shaped Dy( III ) Single-Molecule Magnets. Inorg. Chem. Front. 2020, 7, 4805–4812. 16 CHAPTER 2: MODELING UNSYMMETRICAL LIGANDS FOR TITANIUM-CATALYZED HYDROAMINATION REACTIONS* *Reprinted in part with permission from the Journal of the American Chemical Society. Hou, Z.#; Jena, R.#; McDaniel, T. J.; Billow, B. S.; Lee, S.; Barr, H. I.; Odom, A. L. Modeling Complex Ligands For High Oxidation State Catalysis: Titanium Hydroamination with Unsymmetrical Ligands. ACS Catal. 2024, 14, 8, 5531-5538. #These authors contributed equally. 2.1 Introduction Transition metal catalysts have played a vital role in the past few decades in improving quality of life by acting as a tool for accessing intensive chemical transformations. At the beginning of the 20th century, Bosch produced ammonia from nitrogen and hydrogen by using iron catalysts. This catalytic reaction is sometimes called “the reaction that feeds the world since it’s product, ammonia is an inorganic fertilizer” that currently supports the food supply for half of the world’s population.1 Developing tools for efficient catalyst design is crucial for optimizing chemical processes, enhancing industrial productivity, and reducing environmental impacts. Low valent transition metal catalysts are commonly used in olefin metathesis,2 cross-coupling reactions3, and alkene hydrogenation reactions4. These reactions are widely used in industry to produce detergents, pharmaceutical drugs, oils, specialty chemicals, etc. Upon closer examination of the catalysts for the above-mentioned processes, it becomes evident that each catalyst shares a common ligand, which is phosphine. In low-valent transition metal catalysts, phosphines can function as both σ- donor and π- acceptor and are therefore good ligands. Ancillary ligands play a critical role in the activity of the catalysts, therefore, understanding their effect is important for catalyst design. 17 In 1977, Chadwick Tolman developed two experimentally derived parameters to study steric and electronic properties of ancillary ligands toward low-valent late transition metals.5 Tolman’s electronic parameter is measured by using carbonyl stretching frequencies (ν) in a Ni(CO)3(PR3) complex. If phosphine is a good donor to the metal center, CO will have a lower frequency due to the increased backbonding into CO π*-orbital. The steric parameter referred to as Tolman’s Cone Angle (θ) for phosphine ligands is the apex angle of a cylindrical cone keeping P about 2.28 Å from the metal center. The larger the angle (θ), the greater the steric contribution. Figure 2.1. Tolman electronic parameter and cone angle model. By using parameters for electronic and steric contributions, Tolman was able to fit them into an Eq. 2.1 where Z is the property under investigation. 𝑍 = 𝑎 + 𝑏(𝜒) + 𝑐(𝜃) Eq. 2.1 This mathematical model was a valuable tool for designing late transition metal-based catalysts. However, methods for probing catalysts for high valent early transition metals were not well explored. The chemical properties of late-transition metals are often not transferable to early- transition metals due to the lack of d-electrons in high-valent metals. At the time there was a lack of reliable methods to understand the ancillary ligand effect on high-valent metal catalysis. High-valent metal catalysts such as Ti(IV) are pertinent for the synthesis of some commodity chemicals. For instance, Ziegler-Natta polymerization for the production of polyethylene and 18 polypropylene.6 Sharpless Epoxidation is another Ti(IV) catalyzed reaction to generate enantioselective epoxides which are important in the synthesis of natural products.7,8 Our group has extensively studied the one-pot synthesis of heterocycles using titanium-catalyzed iminoamination reactions.9–14 Using these multicomponent coupling reactions, the Odom group later synthesized different heterocycles with biological activity, including a class of NRF2 inhibitors similar to MSU38335,15 and proteasome inhibitors (Scheme 2.1, A). Other high-valent catalysts such as Ta(V) (Scheme 2.1, B) are known to catalyze hydroaminoalkylation of alkenes.16 For this project we chose a widely studied reaction, alkyne hydroamination (Scheme 2.1, C). Here our goal is to find the effect of unsymmetrical ancillary ligands on catalyst activity. Unlike late transition metals where phosphine ligands are dominant within the realm of ligands, early transition metal catalysts use a variety of ancillary ligands such as alkoxides, halides, cyclopentadienyls, amides, etc. In this case, predicting the properties of the catalysts proves challenging due to the distinct mode of ligand interaction with the metal center. For instance, Iodide primarily serve as a sigma donor ligand, but amides exhibit significant π-donation to the metal’s low-lying vacant orbitals. Traditionally, pKa values have been employed to predict the donor ability of the ligand, but they do not encompass the π-effects from the ligand. Consequently, comparing these two ligands directly on a unified scale presents considerable difficulty. At the time, there was no parameterization system known for the high valent metal systems. 19 Scheme 2.1. Recent examples of high oxidation state transition metal reactions. 2.2 Ligand Donor Parameter (LDP) and Percent Buried Volume (%Vbur) To address this, in 2012 Odom group developed a method to experimentally measure the electron-donating ability of monoanionic ligands towards high-valent metal systems such as titanium(IV), vanadium(V), and chromium(VI).17 This donor parameter was based on a chromium(VI) nitride complex and called the Ligand Donor Parameter (LDP) by our group. The LDP value entails both σ- and π-donations from the ligands to the metal center similar to Tolman’s electronic parameter (χ). The LDP value represents the enthalpic barrier to rotation of the Cr−N bond of the diisopropylamide ligand in the NCr(NiPr)2X complex (Figure 2.2, left), where X is the ligand under investigation. In this chromium system, there is an electronic competition between the monoanionic X ligand and diisopropylamide ligand for the chromium’s acceptor orbitals in the 20 xy plane. Depending upon the nature of X ligands i.e., more donating, or less donating, the rate of rotation of the Cr–N bond changes. If X is a stronger donor than diisopropylamide, there will be less donation from diisopropylamide into metal’s vacant empty d-orbitals. Thus, the rotation around Cr–N will be faster, whereas, if X is a weak donor in comparison to diisopropylamide, the lone pairs from N of diisopropylamide would be donating into the metal’s empty orbital. This will give rise to a partial double bond character to the Cr–N bond and a higher barrier to rotation (slow rotation). A higher LDP value suggests X ligand is a weaker donor than diisopropylamide and a lower value of LDP suggests that it is a stronger donor than diisopropylamide to the high valent metal center. The resulting donor parameters for X have been correlated to Hammett parameters for aryloxide ligands, Angular Overlap Model parameters, and 13C NMR data for W(VI) species.17 Figure 2.2 (right) shows some common ancillary ligands with their LDP values. 21 Figure 2.2. (left) NCr(NiPr)2X system for LDP measurement. (right) Collection of LDP values for common ligands.18 The rate of Cr–N bond rotation was measured by a Spin Saturation Transfer (SST) 1H NMR experiment. The rotation rate around Cr–N was used to calculate the Gibbs free energy of rotation. Using the Eyring equation (Eq. 2.2), the enthalpy of activation was calculated (LDP value). Keeping the entropy of activation constant at –9 cal•mol-1k-1, gives an LDP value that is temperature-independent. 𝑙𝑛 ( 𝑘𝑜𝑏𝑠 𝑇 ) = 𝑙𝑛 𝜅𝑘𝐵 ℎ 1 𝑅 ‡ (𝛥𝑆 − 𝛥𝐻 ‡ 𝑇⁄ ) Eq. 2.2 Where 𝜅 = transmission coefficient (assumed 1) Inspired by Cavallo and coworkers, percent buried volume (%Vbur) was used to study the steric effects of the ligands. Percent buried volume is defined as the volume of a sphere occupied by a 22 ligand with a 3.5 Å radius from the metal. This usually encompasses the first coordination sphere around the metal center.19 In 2017, our group reported the first application of LDP values by modeling the donor ability of ancillary ligands for high-valent metal catalysts.18 A correlation between the electronic and steric properties of the ligand was used to build a model to predict the reaction rate. Similar to Tolman’s model (Eq. 2.1), our model had both electronic (LDP) and steric (%Vbur) parameters (Eq. 2.3). Once a model is built by using a few ligands, it can be used to predict the rate of the reaction for a catalyst based on the ancillary ligand. 𝐾𝑜𝑏𝑠 = 𝑎 + 𝑏(𝐿𝐷𝑃) + 𝑐(%𝑉𝑏𝑢𝑟 ) Eq. 2.3 To summarize previous findings, the LDP values and %Vbur were used together for different symmetrical bidentate ligands to measure the rate of hydroamination reaction. The LDP and %Vbur were calculated from half of the symmetrical chelate. For instance, to evaluate the ligand effect from the dipyrromethene (dpm) ligand in Ti(dpm)(NMe2)2, complex, one pyrrole ligand on chromium nitrido complex was used to measure the LDP value. Further, the purple color shows the volume occupied by a pyrrole ligand around a 3.5 Å radius from the chromium center and thus represents %Vbur (Figure 2.3). 23 Figure 2.3. (top) Chromium system to calculate the LDP value of dipyrrolylmethane ligand. (bottom) Space-filling model to measure %Vbur of the pyrrole ligand.18 The preferred kinetic model involves the extensively researched hydroamination reaction of an alkyne, in this case 1-phenylpropyne, employing diverse Ti(NMe2)2X catalysts (Figure 2.4). Here, X represents a spectrum of symmetric, bidentate ligands derived from pyrrole, aryloxide, and indole. All the pyrrolyl-based precatalysts are shown in 1,1 binding mode for easier visualization, in a solid state this might not be true (vide infra). According to this model, Ti(dpm)(NMe2)2 1a is the fastest catalyst so far. The model indicated that smaller, more electron- deficient ligands yield higher reaction rates. 24 Figure 2.4. A list of symmetrical ligand-based titanium precatalysts used in our previous study.18 The discussed parameterization was only applicable to C2v symmetric ligands which limits the scope of the environment around the metal. What if we use unsymmetrical chelate ligands? Expanding our LDP systems to unsymmetrical ligands would give us more tunability and freedom with ligand design and perhaps lead toward a faster catalytic system than 1a. A pictorial representation of unsymmetrical ligand-based catalyst 5a is shown in Figure. 2.5. Figure 2.5. An example of a symmetrical system where both sides of the chelate are the same 1a, an example of an unsymmetrical system where both sides of the chelate are different 5a. Schrock and coworkers have extensively studied the effect of ancillary ligands for d0 molybdenum-catalyzed olefin metathesis reactions. During their quest to find out a synthetic route 25 to synthesize bisalkoxides molybdenum alkylidene complex Mo(NAr)(CHR’)(OR”)2 from a starting precursor Mo(NAr)(CHR’)(X)2, where X is a pyrrolide ligand, they isolated a monoalkoxide pyrrolide (MAP) complex Mo(NAr)(CHR’)(X)(OR”) instead dialkoxide, as shown in Scheme 2.2.20 Scheme 2.2. Synthesis of molybdenum monoalkoxide pyrrolide. These MAP complexes were studied extensively by the Schrock and Hoveyda group, where the chiral nature of the metal center led to the discovery of a new class of chiral olefin metathesis catalysts. Traditionally, the bisalkoxide metal complexes performed better for olefin metathesis than sluggish bispyrrolide metal complexes. Interestingly, they found that the reactivity of MAP ligands towards olefin is not intermediate between those of bispyrrolides and bisalkoxides, but it was even much greater than the bisalkoxides. Using MAP complexes, they were able to carry out ring-closing metathesis of triene in the synthesis of the natural product quebrachamine, which was quite challenging before this. From a theoretical perspective Eisenstein and Coperet21 found that in the Mo(NAr)(CHR’)(X)(Y) complex, ligands X and Y are playing different roles where one of them is acting as a weak donor ligand (i.e., alkoxide) and the other one is acting as a strong donating ligand (i.e., alkyls) – leading to better reactivity of MAP complexes compared to their symmetrical analogs. Having this knowledge made us wonder if we could also enhance the activity of our titanium catalysts by swapping our symmetrical ligands with unsymmetrical ligands. If so, we would have to build a new model to consider each side of the ancillary ligands independent of 26 each other. 2.3 Synthesis of Unsymmetrical Ligands and Titanium Catalysts This project was completed in close collaboration with Dr. Zhilin Hou, Dr. Tanner J. McDaniel, Dr. Brennan S. Billow, Dr. Seokjoo Lee, and Hannah Barr. Zhilin and I carried out the synthesis of ligands, titanium catalysts, and kinetics experiments along with the characterization of all complexes made. Tanner started the project and made a couple of unsymmetrical ligands and their respective catalysts and collected preliminary data to support this project. Seokjoo carried out ligand synthesis and prepared starting material to push the project forward. Hannah discovered the conditions for the mono-borylation of dipyrromethene ligand. For designing unsymmetrical ligand systems for catalysts, we possessed foundational knowledge rather than complete unfamiliarity with the task at hand. We commenced with a definite starting point and formulated the ligand design based on our previous study where symmetrical pyrrole, indole, and phenol-based bidentate ligands were used. There appear to be two viable approaches to accomplish this; the first is desymmetrizing the symmetrical ligand by putting a substituent on one side of the ligand. Secondly, joining two monodentate ligands using a condensation reaction in the presence of a Lewis acid to make an unsymmetrical bidentate ligand system. A pictorial version of the two strategies is shown in Figure 2.6. 27 Figure 2.6. Two routes to synthesize unsymmetrical ligands. (top) Desymmetrizing the symmetrical ligand, (bottom) and condensation of two monodentate ligands. The first route to make an unsymmetrical ligand includes two steps. 1) iridium-catalyzed mono borylation of dipyrromethane (H2dpm) ligand22,23 and 2) Suzuki-Miyaura coupling with a variety of aryl halides.24 The serendipitous discovery of the regioselective monoborylation at the α- position of H2dpm ligand using pinacolborane streamlined our synthetic route. Remarkably, there was a minimal amount of diborylated product observed which was separated by column chromatography. The reaction of titanium tetrakis(dimethylamide) with an unsymmetrical ligand gave a yellow/orange solid which was purified by recrystallization to afford a pure titanium precatalyst (Scheme 2.3). 28 Scheme 2.3. Desymmetrizing a symmetrical ligand using monoborylation followed by Suzuki- Miyaura coupling and synthesis of titanium precatalyst. The second approach to making an unsymmetrical ligand involved condensation of a monodentate ligand containing the leaving group with another monodentate ligand in the presence of Lewis acids such as InCl3 or BF3OEt2.25 This method was utilized to couple indoles with pyrroles or phenols (Scheme 2.4). 29 Scheme 2.4. Condensation of different sides of the unsymmetrical ligand and synthesis of respective titanium precatalyst. Utilizing the aforementioned method, we successfully synthesized 10 new titanium precatalysts (shown in Figure 2.7). All the catalysts were isolated and completely characterized for this study. For clarity, these catalysts were divided into three categories. First H2dpm-based unsymmetrical catalysts (5), second the unsymmetrical complexes with 3-methyl indolyl and pyrrolyl (6), and third phenoxide and indolyl/pyrrolyl-based complexes (7). For clarity, ligands are drawn in 2 30 dimensions and as 1-coordinated to the metal center which might not be the case in solid-state structures of the complex (see later). The linker between two sides of the ligand was either −CH2− or −CMe2− and based on our previous study we measured the rate constant based on electronic and steric parameters of the attached ligand, not the linkers.18 Figure 2.7. Structure of unsymmetrical ligand catalysts used for this study. Out of 10 new unsymmetrical catalysts, we were able to get crystal structures for 7 of them by using single-crystal X-ray diffraction (Figure 2.7, 5a, 5b, 5c, 5d, 6a, 7a, and 7c). In the case of dpm-based unsymmetrical catalysts (5), the 2-aryl pyrrolyl side of the unsymmetrical ligand was always 5-coordinated to the metal center irrespective of the electronic (electron donating/electron withdrawing) and steric properties of the ligand. In other words, the side of the ligand with a higher %Vbur and lower/higher LDP was always found to be 5-coordinated to the titanium center. In the case of 6a, the expected electron-rich pyrrole ring was 5-coordinated to the metal center to 31 stabilize the metal in a high oxidation state. When the unsymmetrical ligand had phenoxide as one side, irrespective of the other side (indolyl/pyrrolyl) 1−1 binding mode was observed. In the case of phenoxides (7a-c), a seven-membered metallacycle is formed which could prevent pyrrole on the other side from going 5 due to steric constraints around the metal center. Furthermore, phenoxides are better donors to the metal center; therefore, an additional electronic stabilization from the pyrrole ring (5) is not needed. Figure 2.8. Structures of unsymmetrical titanium catalysts with their respective ORTEP drawing. 32 2.4 Ligand 1- 5 Isomerization Previously our group has done a study to calculate the enthalpic barrier for dipyrrolylmethane (dpm) isomerization.26 I did a similar study here to understand the energetic preference for substituted pyrrole binding in an 5 fashion. Theoretically, there are three possible structures A, B, and C of the catalyst under consideration (Figure 2.9). I used Density Functional Theory to calculate the energies with B3PW91/def2TZVP. The transition states corresponding from A to B and B to C are represented as TS1‡ and TS2‡, respectively; these were found for catalyst 5d. The ground state structures A, B, and C were located for all the complexes with unsymmetrical bidentate ligands. Based on DFT calculations, isomer A is more stable in most cases and consistent with crystal structures, i.e., the substituted pyrrolyl side being η5 is energetically preferred. In all cases with indoles, this ligand is strongly preferred to be η1. The calculated energy profile for all catalysts is shown in Figure 2.9 to compare the relative stability of each isomer concerning A (set to zero). Unlike other catalysts, 5b shows isomer B has lower energy than isomer A and that could be attributed to the 2-Me groups on both sides, but such η1,η1-dpm derivatives have been observed in other cases, usually with the metal picking up an additional donor ligand.18 Figure 2.9. (top) Possible structures for catalysts. (bottom) Calculated ground state energies for isomers A, B, and C for all catalysts. 33 Figure 2.9 (cont’d) I did further analysis of our fastest catalyst Ti(pyr3,5-CF3Ph-C(CH3)2-pyr)(NMe2)2 5d both computationally and experimentally to understand the interconversion between each isomer. The gas phase structure calculated was consistent with experimentally-obtained crystal structures. I found isomer A has the lowest energy as expected with an η5-substituted pyrrolyl. The energy barrier TS1‡ to go from isomer A (5d) to isomer B (5d) is 3.64 kcal/mol. This energy gap can be easily overcome at room temperature, which is consistent with NMR spectroscopy (vide infra). The transition state energy TS2‡ is higher than TS1‡ suggesting the conversion of isomer B to A is energetically more favorable than the conversion of isomer B to C (Figure 2.10). The η1,η1- isomer (B) and the η5,η1-isomer, where the unsubstituted pyrrolyl is coordinated through the π- system (C) have the same energy (2.8 kcal/mol) according to DFT. 34 Figure 2.10. Calculated energy profile for catalyst Ti(pyr3,5-CF3Ph-C(CH3)2-pyr)(NMe2)2 5d showing interconversion between each isomer (A, B, and C) and associated transition states (TS1‡ and TS2‡). Variable temperature NMR spectroscopy was used to further analyze the 1-5 ligand isomerism of Ti(pyr3,5-CF3Ph-C(CH3)2-pyr)(NMe2)2 5d complex. At 182 K, we believe the ground state structure should be similar to the crystal structure, where substituted pyrrole is coordinated in 5-fashion to the metal center. In this case, the two methyl groups present on the ligand backbone will be in different chemical environments and show two singlets in the NMR spectrum (boxed region in Figure 2.11). As we increase the temperature to 208 K, the coalescence temperature, the two peaks start to broaden and merge into a broad peak with a flat top. Warming the sample to 232 K gives a sharp signal for the methyl groups suggesting fast exchange on the NMR timescale. 35 Figure 2.11. Variable temperature NMR study for the catalyst Ti(pyr3,5-CF3Ph-C(CH3)2- pyr)(NMe2)2 5d. Line shape analysis27 was used to calculate the rate constant associated with different temperatures. At 182 K the rate constant k1 is given by Eq. 2.4; where w = w – wo, w is the width at the half-height of the broadened peak (189 K), and wo is the width of the half-height at 182 K. At the coalescence temperature, 204 K, the rate constant k2 is given in Eq. 2.5; where 𝜗a and 𝜗b are chemical shifts of the different methyl protons in Hz. At 232 K, the rate constant k3 is calculated by using Eq. 2.6; where wf is the half height of the peak at the fast exchange limit. 𝑘1 = 𝜋(Δ𝑤) Eq. 2.4 𝑘2 = 𝜋(𝜗𝑎−𝜗𝑏) √2 Eq. 2.5 36 𝑘3 = 𝜋(𝜗𝑎−𝜗𝑏)2 2(𝑤−𝑤𝑓) Eq. 2.6 Using the Eyring Equation (Eq. 2.7), the slope and intercept of ln(kobs/T) vs 1/T plot was used to calculate enthalpy and entropy associated with 1-5 ligand isomerism (see experimental). 𝑙𝑛 ( 𝑘𝑜𝑏𝑠 𝑇 ) = − Δ𝐻‡ 𝑅𝑇 + 𝑙𝑛 𝑘𝐵 ℎ + Δ𝑆‡ 𝑅 Eq. 2.7 The enthalpic barrier, H‡, for 1-5 ligand isomerism of catalyst 5d = 9.5 ± 0.2 kcal/mol, and S‡ = 1.2 ± 1.1 cal/mol. The Gibbs free energy, G‡, at room temperature is 9.1 ± 0.6 kcal/mol. The lower barrier between A and B suggests that these will be quickly exchanged at the accessible temperatures on the NMR timescale. As a result, we assume we are measuring the barrier (TS2‡) between C and the fast-exchanging B/A system. 2.5 Collection of Kinetics Data for 10 New Titanium Catalysts In the previous paper from the Odom group with symmetrical ligands, the hydroamination of 1-phenylpropyne with aniline under pseudo first-order conditions was used to examine the effect of the ancillary ligand on the rate of the reaction.18 The same conditions were used here for an unsymmetrical ligand system where the hydroamination reaction of 1-phenylpropyne was run with an excess of aniline and the reaction rate was measured by monitoring dissipation in alkyne concentration over time, which gave a good fit for the first-order plots. The fit of kinetic data was done using Espenson’s first-order equation for the instrument response.28 For all the catalysts under investigation, the major product generated was imine as in Figure 2.12 and the other regioisomer was present in a trace amount which could be observed by GC-MS (Figure 2.26). The reaction conditions and a representative plot for catalyst 5e are shown in Figure 2.12. 37 Figure 2.12. (top) Condition for kinetics experiment, (bottom) 1H NMR representative plot with fit (5e) showing the disappearance of 1-phenylpropyne starting material in the reaction. In the previous study, using symmetrical ligands the fastest catalyst was the most common catalyst known at the time 1a without any side reactions. On the other hand, employing the use of unsymmetrical ligands led to the discovery of catalysts faster than 1a. This proves our hypothesis that tuning each side of the ligand independently of each other can give us a better catalyst. It is to be noted here that not all the unsymmetrical catalysts gave us a faster reaction rate than their symmetrical analogs. All catalyst kinetic runs were run in triplicates, and an average value is reported in Table 2.1. 38 Table 2.1. Average rate constant values of 10 new precatalysts. Catalyst Number Observed Structure Rate Constant (10-4 s-1) 3.1 1.9 3.95 5.5 3.46 3.64 5a 5b 5c 5d 5e 6a 39 Table 2.1 (cont’d) 6b 7a 7b 7c 4.4 0.67 0.67 0.46 2.6 Model Consideration for Unsymmetrical Ligands An approach similar to that used for symmetrical ligands was used to model the rate of hydroamination reactions of alkynes using unsymmetrical catalysts. For symmetrical ligands half of the ligand was used to calculate %Vbur and LDP, likewise, in the case of unsymmetrical ligands, both sides of the chelating ligand will be used to measure these values. As mentioned earlier, the building blocks for unsymmetrical ligands were inspired by symmetrical ligands therefore we can leverage LDP and %Vbur values from our previous study and build a model.18 As our unsymmetrical ligand system comprises two distinct components, we established parameters (LDP and %Vbur) for each to construct the model effectively. For instance, indolyl-pyrrolyl-based ligands had the steric and electronic parameters taken from chromium-indolyl and chromium-pyrrolyl 40 complexes. A five-parameter model was obtained for this new system where values from a-e were found by regression. 𝑘𝑜𝑏𝑠(× 104) = 𝑎 + 𝑏(𝐿𝐷𝑃)1 + 𝑐(𝐿𝐷𝑃)2 + 𝑑(%𝑉𝑏𝑢𝑟)1 + 𝑒(%𝑉𝑏𝑢𝑟)2 𝐄𝐪 𝟐. 𝟖 The subscript 1 and 2 refer to different sides of the bidentate ligand. Upon juxtaposing our previous model with the new one (Figure 2.13), it becomes evident that employing this five- parameter, four-variable system could provide the metrics for a better catalyst, thereby facilitating the production of superior catalysts. Figure 2.13. Direct comparison of the previous model used for symmetrical ligands with the new model for unsymmetrical ligands. This model is quite useful and can indicate whether the two sides of the unsymmetrical ligand (for example, the pyrrolyl and indolyl of the bidentate ligand in 6a) are in separate or similar environments, even though it doesn’t describe the nature of the critical transition state. To 41 understand, we can assume that the protonolysis of the Ti−C bond is a crucial step in the catalytic cycle, similar to related systems.29 For this discussion, we will suppose that a trigonal bipyramidal structure like B’ is involved. After that, a mechanism like the one in Scheme 2.5 may be drawn. The electronics and sterics of the ancillary ligands are probably influencing the rate of protonolysis or the equilibrium required to reach B’. The model provides recommendations for the structure but does not provide the precise structure of the complex involved. For instance, at the transition state the model parameters of the two sides of the bidentate ligand (X and X’) differ significantly, indicating the sites they occupy around the metal are also different. This is more compatible with a trigonal bipyramidal or square pyramidal structure where X and X’ are in axial and equatorial positions, as indicated in Scheme 2.5 for B’, rather than the pseudo-tetrahedral complexes like A’ or C’. Scheme 2.5. A possible mechanism for the titanium hydroamination where B’ is represented as a key intermediate. 42 Furthermore, the model can indicate whether the sterics or the electronics of the two sides of the ligand decide the ligand’s residence during the key step of the catalysis. Put another way, if we designate one side of the ligand, say X in B’ (Scheme 2.5) as “1” and the other side X’ as “2”, the modeling will indicate the placement of X or X’ which could be axial or equatorial based on the larger %Vbur or LDP, which are our steric and electronic descriptors, respectively. The model produced was shown to be much better when we assigned side 1 to the greater LDP value rather than using sterics (%Vbur) (more details in the experimental). The 10 new unsymmetrical catalysts under investigation (Figure 2.7) are modeled with previously reported 10 symmetrical catalysts (Figure 2.4). Precatalysts with symmetrical ligands are shown as blue triangles, and precatalysts with unsymmetrical ligands are shown as orange and yellow diamonds (Figure 2.14). The linear fit black line should ideally have a slope of 1.00 but in this case, it is 0.75  0.10 with R2 = 0.75. The extremely poor fit obtained here suggested that something was amiss. We observed phenoxide-based ligand systems (Figure 2.6, 7a−7c) exhibit sluggish behavior and demonstrate slower catalysis (Figure 2.14, yellow diamonds) than other unsymmetrical ligands (Figure 2.14, orange diamonds). After closely analyzing the reaction mixtures of the 7a−7c hydroamination reaction, we determined that these catalysts are undergoing decomposition in the freezer at −35 °C after a few days, highlighting the inherent instability of the phenoxide-based unsymmetrical ligand system. 43 Figure 2.14. The plot of experimental rate constant (kobs) vs calculated rate constant using the model (kcalc) for 20 precatalysts both symmetrical and unsymmetrical ligand systems. It would be prudent to remove catalysts 7a−7c due to their instability and conduct the modeling process again to ascertain if a better fit can be achieved with now 17 precatalysts (7 unsymmetrical and 10 symmetrical ligand catalysts). Simply by excluding the phenoxides-based ligand systems, the model was significantly improved with a slope of 0.97  0.04 with R2 = 0.97 (Figure 2.15). 44 Figure 2.15. The plot of experimental rate constant (kobs) vs calculated rate constant using the model (kcalc) for 17 precatalysts (excluding 7a-7c) in both symmetrical and unsymmetrical ligand systems. Table 2.2. Parameters for Eq. 2.8 for titanium hydroamination and statistics for the regression. Parameter (descriptor) Natural Parameters a (intercept) b (LDP)1 c (LDP)2 d (%Vbur)1 e (%Vbur)2 Scaled Parameters (–1 to +1) –6.56 ± 0.18 –7.6 ± 3.4 2.8 ± 0.46 3.0 ± 0.42 –0.98 ± 0.57 –0.90 ± 0.53 –0.16 ± 0.096 –0.63 ± 0.26 –0.48 ± 0.091 –1.5 ± 0.25 45 Table 2.2 shows the model parameters from the regression analysis using natural parameters and scaled parameters with all the precatalysts with symmetrical ligands (Figure 2.4) and unsymmetrical ligands 5a-e and 6a-b (Figure 2.7). If one wishes to determine the rate of the catalysts not included in this study, substituting the LDP and %Vbur values for the ligand in question with a-e values (natural parameters) from Table 2.2 into Eq. 2.8 will give the weighted rate constant, provided the mechanism is the same as that of all the other ligands in the data set. Using scaled parameters one can also determine the stereoelectronic characteristics of either side of the ligand and how that influences the rate of the reaction. For instance, the electronics of side 1 (higher LDP) make a larger contribution than side 2 (lower LDP). Based on signs and magnitudes of the parameters, it appears that side 1 should have a ligand that is relatively small but doesn’t donate strongly (higher LDP value) in order to maximize the rate of the reaction. On the other hand, the donor ability of the ligand on side 2 seems to be somewhat less critical, although a relatively stronger donor may still be advantageous. However, it is crucial that the ligand on side 2 of the unsymmetrical ligand is small. This combination suggests that both electronic and steric properties of the ligand play a significant role in determining the reaction rate. As we can see here, the unsymmetrical ligands, in general, are faster catalysts than the symmetrical counterparts. This is probably because unsymmetrical ligands have more flexible electronic and steric structures; i.e., they may be able to arrange their different sides in a way that is more energetically favorable than symmetrical ligands. Going back to the catalytic cycle shown in Scheme 2.6, the unsymmetrical ligand can orient itself in different ways in the TBP structure of B’ and impact the rate of protonlysis.30 The employment of an unsymmetrical ligand system on titanium catalysts enhanced their performance for hydroamination reactions. Moreover, we were able to find a better catalyst 5d 46 than the previously reported best symmetrical precatalyst 1a. The structure of the 5d meets the requirement of the model and therefore acts as a faster catalyst. Let’s see what makes 5d a faster catalyst than other symmetrical and unsymmetrical precatalysts. As shown above, the “side 1” of the ligand is the side with a larger LDP value (meaning less donating ligand). Based on scaled parameters, the magnitude of the coefficient of LDP1 (b = +3.0 ± 0.42) is the highest among all coefficients suggesting that it is the most important factor in deciding the rate of the reaction. Further, the positive sign here indicates that increasing the LDP value will increase the rate rapidly. The size of side 1 is relatively less important because of the smaller magnitude of the coefficient d = –0.63 ± 0.26. The negative sign in this context signifies that as the steric hindrance of side 1 decreases, the rate will increase. Consequently, a minor impedance in the rate should be there due to the larger sterics of side 1. The “side 2” has a smaller LDP value and is therefore consistent with the model (c = –0.90 ± 0.53). Unlike the electronics of side 2, sterics play a significant role with e = –1.5 ± 0.25 suggesting that side 2 should be smaller to get a better rate. As we can see in 5d side 2 is pyrrolide which is both a strong donor and sterically small. 2.7 Conclusions Previously our group reported 10 different precatalysts (Figure 2.4) using symmetrical bidentate ancillary ligands. The mathematical model provided valuable insight into the relationship between 47 steric and electronic properties of the ancillary ligand and the alkyne hydroamination rate constant. It was observed that the rate increases as the ancillary ligand becomes more electron-deficient (high LDP) and smaller (low %Vbur). This kind of quantitative analysis can be extremely useful for predicting and optimizing the reaction conditions for high-valent metal catalysis. However, they were unable to discover a catalyst that was faster than the 1a at the time. In this study, 10 new precatalysts using unsymmetrical bidentate ancillary ligands (Figure 2.7) were prepared to explore the effect of these ligands on titanium-catalyzed alkyne hydroamination reaction. The two sides of the ligands (X and X’, Scheme 2.5) differ both electronically and sterically, and it was found that in most cases this asymmetry can lead to accelerated reaction rates as compared to their symmetrical counterparts. Phenoxide-based ligand systems showed decomposition at ambient temperatures and therefore were excluded from the modeling. The cumulative model of the catalysts with symmetrical and unsymmetrical ligands (17 precatalysts) suggests that the position of the ligand in the key transition state is best determined by the electronics where “side 1” of the ligand needs to be a poorly donating ligand (positive coefficient on descriptor LDP1) and small (negative coefficient on descriptor %Vbur1) to get good reaction rates. Whereas a more donating ligand is required on the other side of the unsymmetrical ligand (the negative coefficient on descriptor LDP2), and most importantly, it should be smaller (the larger negative coefficient on descriptor %Vbur2). We were able to make a faster catalyst 5d using an unsymmetrical ligand than the previously reported 1a. An unsymmetrical ligand system provides us with the greater advantage of tuning each side independently of the other to design an efficient catalyst for hydroamination. 48 Moving forward, we are looking at other catalytic systems using symmetrical and unsymmetrical ligand systems. It should be noted that the model given here is valid only when the mechanism is the same for the catalysis; what if the mechanism is changed? In that case, this model cannot be implied. In the next chapter, I will discuss in detail some other factors that can affect the rate of the reaction and how you can spot small details that would otherwise be overlooked if not for this modeling study. 49 2.8 Experimental Details General Considerations All manipulations were carried out under an inert dinitrogen atmosphere in an MBraun glovebox or using standard Schlenk techniques. Toluene was sparged with purified dinitrogen and passed over an activated alumina column prior to use. n-Hexane was dried over sodium benzophenone radical, refluxed, and distilled under dinitrogen prior to use. All deuterated NMR solvents were purchased from Cambridge Isotope Laboratories. C6D6 was dried over CaH2 and distilled under dinitrogen. CDCl3 was dried over P2O5 and distilled under dinitrogen. Ti(NMe2)4 was purchased from Gelest and used as received. Tetrakis(triphenylphosphine)palladium(0) was purchased from Strem and used as received. 1- phenyl-1-propyne was purchased from Combi-blocks and distilled from barium oxide prior to use. Aniline was purchased from Sigma-Aldrich and distilled from KOH and passed through dry alumina prior to use. Spectra were taken on Varian instruments located in the Max T. Rogers Instrumentation Facility at Michigan State University. These include an Agilent DDR2 500 spectrometer equipped with a 5 mm pulsed-field-gradient (PFG) OneProbe and operating at 499.955 MHz (1H) and 125.77 MHz (13C), a Varian Inova 600 spectrometer equipped with a 5 mm PFG switchable broadband probe operating at 599.89 MHz (1H) and 564.30 MHz (19F), a UNITY plus 500 spectrometer equipped with a 5 mm Pulsed-Field-Gradient (PFG) switchable broadband probe and operating at 499.955 MHz (1H) and 125.77 (13C). Single crystal data was collected on XtaLAB Synergy, Dualflex, Hypix diffractometer using CuKα radiation. Data collection was done at 100 K under a continuous flow of liquid nitrogen. In Olex2 program, crystal structures were solved with ShelXT solution 50 using intrinsic phasing and refined with the SheXT refinement package using least squares minimization. All hydrogens are refined anisotropically. All crystals were stable at room temperature for mounting. Synthesis of Ligands Most of the ligands synthesis was carried out by Zhilin Hou and Seokjoo Lee. Synthesis of (3-methyl-1H-indol-2-yl)methanol In a 100 mL flask, LAH (461 mg, 5.0 equiv) was added to 10 mL of dry THF under a constant flow of N2. A solution of methyl 3-methylindole-2-carboxylate (460 mg, 1.0 equiv) in dry THF (15 mL) was added slowly. The reaction was stirred at 40 °C for 6 h. The reaction was slowly quenched with 1 M HCl and extracted with ethyl acetate. The organic layer was washed with brine, separated, and dried over sodium sulfate. The evaporation of the solvent afforded crude product as a white solid (390 mg, 99%). 1H NMR (500 MHz, cdcl3) δ 8.16 (s, 1H), 7.54 (d, J = 8.8 Hz, 1H), 7.33 (d, J = 9.0 Hz, 1H), 7.22 – 7.16 (m, 1H), 7.11 (t, J = 6.9 Hz, 1H), 4.84 (s, 2H), 2.30 (s, 3H). 13C NMR (126 MHz, CDCl3) δ 135.79, 133.00, 128.97, 122.38, 119.39, 119.00, 110.91, 108.60, 57.71, 56.71. Synthesis of 2-((1H-pyrrol-2-yl)methyl)-3-methylindole 51 In a 50 mL Schlenk tube, (3-methyl-1H-indol-2-yl)methanol (395 mg, 1.0 equiv) was dissolved in an excess of pyrrole (3 mL). Slowly, BF3·Et2O (0.15 mL, 0.5 equiv) was added dropwise to the solution under a constant flow of N2. The reaction rapidly turned purple. The mixture was allowed to stir at room temperature for 5 min. The heat was generated after the addition. The reaction was quenched by adding 20 mL of saturated sodium bicarbonate solution. The product was extracted by adding 20 mL of EtOAc. The organic layer was separated and dried over sodium sulfate. The excess pyrrole was removed under vacuum for 5-8 h. The crude product was purified by column chromatography (silica gel, hexanes: EtOAc, 20:1) to afford the desired products. Removal of the solvent afforded the product as a colorless solid (200 mg, 40%). 1H NMR (500 MHz, CDCl3) δ 7.82 (s, 1H), 7.66 (s, 1H), 7.53 (d, J = 7.3 Hz, 1H), 7.21 (d, J = 9.0 Hz, 1H), 7.17 – 7.02 (m, 2H), 6.73 – 6.54 (m, 1H), 6.18 (q, J = 2.8 Hz, 1H), 6.09 (s, 1H), 4.11 (s, 2H), 2.31 (s, 3H). 13C NMR (126 MHz, CDCl3) δ 135.53, 131.58, 129.35, 128.36, 121.71, 119.35, 118.47, 117.56, 110.58, 108.76, 108.03, 106.84, 24.90, 8.61. Synthesis of 2,4-di-tert-butyl-6-((5-fluoro-3-methyl-1H-indol-2-yl)methyl)phenol A 35 mL pressure tube, equipped with a stir bar, was loaded with 5-fluoro-3-methylindole (750 mg, 5 mmol), InCl3 (202 mg, 0.9 mmol), and lastly 2,4-di-tert-butyl-6-(hydroxymethyl)phenol (1.08 g, 4.5 mmol). The reaction was then heated at 65 °C for 16 h. The crude product was purified by column chromatography using hexanes:ethyl acetate (95: 5), concentrated, and kept for recrystallization at −20 ° to give white solid (830 mg, 50% yield). 1H NMR (500 MHz, C6D6) δ 7.47 (d, J = 2.6 Hz, 1H), 7.01 (d, J = 2.7 Hz, 1H), 6.93 – 6.82 (m, 2H), 6.46 (dd, J = 8.8, 4.4 Hz, 52 1H), 4.94 (s, 1H), 3.62 (s, 2H), 2.00 (s, 3H), 1.48 (s, 9H), 1.34 (s, 9H). 13C NMR (126 MHz, C6D6) δ 159.46, 157.59, 151.78, 143.01, 136.74, 132.82 (d, J = 58.2 Hz), 129.94 (d, J = 9.5 Hz), 125.60, 124.42, 123.42, 111.54 (d, J = 9.1 Hz), 110.25 (d, J = 26.2 Hz), 108.74 (d, J = 4.8 Hz), 104.01 (d, J = 22.9 Hz), 35.13, 34.47, 31.89, 30.08, 28.91, 8.32. 19F NMR (470 MHz, C6D6) δ -124.73 (td, J = 9.6, 4.4 Hz). Synthesis of Titanium Catalysts Synthesis of Ti(dpm2-Me)(NMe2)2 (5a) A 35 mL pressure tube equipped with a stir bar was loaded with Ti(NMe2)4 (0.575 g, 2.57 mmol, 1 equiv) and toluene (3 mL). A 20 mL scintillation vial was loaded with H2dpm2Me (0.483 g, 2.57 mmol, 1 equiv) and toluene (3 mL). Both solutions were cooled in a liquid nitrogen cold well for 15 min. The cold solution of H2dpm2Me was added dropwise to the vigorously stirring solution of Ti(NMe2)4. The reaction was allowed to warm and then stir at room temperature for 1 h. The pressure tube was then sealed and heated at 65 °C. The reaction progress was monitored by 1H NMR and was completed after 24 h. Volatiles were removed in vacuo to give a light red solid. This solid was rinsed with cold pentane to yield the product as an orange powder (0.67 g, 82% yield). X-ray quality crystals can be grown by dissolving the complex in the minimum amount of pentane and cooling to –30 °C. 1H NMR (C6D6, 500 MHz): δ = 7.05 – 6.97 (m, 1H), 6.59 – 6.45 (m, 1H), 6.35 (d, J = 2.4 Hz, 1H), 6.20 (d, J = 3.2 Hz, 1H), 6.01 (d, J = 2.6 Hz, 1H), 2.96 (s, 12H), 2.03 (s, 3H), 1.82 (s, 6H) 13C{1H} NMR (C6D6, 125 MHz): δ = 161.93, 161.70, 140.25, 123.61, 114.76, 53 113.67, 107.65, 101.56, 46.91, 39.54, 29.68, 15.01. Elemental Analysis: Calcd for C17H29N4Ti. C, 59.63; H, 8.13; N, 17.38. Found: C, 59.55; H, 8.48; N, 17.89. M.pt.: 127-128 °C. Synthesis of Ti(dpm3-Me)(NMe2)2 (5b) A 35 mL pressure tube equipped with a stir bar was loaded with Ti(NMe2)4 (0.228 g, 1.02 mmol, 1 equiv) and toluene (3 mL). A 20 mL scintillation vial was loaded with H2dpm2,2’,3-TriMe (0.220 g, 1.02 mmol, 1 equiv) and toluene (3 mL). Both solutions were then cooled in a liquid nitrogen cold well for 15 min. The cold solution of H2dpm2,2’,3-TriMe was added dropwise to the vigorously stirred solution of Ti(NMe2)4. The reaction was allowed to warm and stir at room temperature for 1 h. The pressure tube was then sealed and heated at 65 °C. The reaction progress was monitored by 1H NMR and was complete after 36 h. Volatiles were removed in vacuo to give a light red solid. This solid was rinsed with cold pentane to yield the product as a red-orange powder (0.252 g, 71% yield). X-ray quality crystals can be grown by dissolving the complex in the minimum amount of pentane and cooling to –30 °C. 1H NMR (C6D6, 500 MHz): δ = 6.13 (d, J = 2.8 Hz, 1H, pyrr-H) 5.99-6.00 (m, 2H, pyrr-H), 2.85 (s, 12H, N(CH3)2), 2.29 (s, 3H, CH3), 2.14 (s, 3H, CH3), 1.83 (s, 6H, C(CH3)2), 1.80 (s, 3H, CH3). 13C{1H} NMR (C6D6, 126 MHz): δ = 162.12, 160.84, 135.35, 134.62, 121.78, 112.46, 108.29, 105.75, 46.76, 39.61, 30.35, 16.25, 13.68, 11.64. M.pt.: 108- 109 °C. 54 Synthesis of Ti(dpm2-Ph)(NMe2)2 (5c) In 20 mL scintillation vial with a stir bar was loaded with Ti(NMe2)4 (0.068 g, 0.3 mmol, 1 equiv) and diethyl ether (3 mL). A 20 mL scintillation vial was loaded with H2dpm2Ph (0.074 g, 0.3 mmol, 1 equiv) and diethyl ether (3 mL). Both solutions were then cooled in a liquid nitrogen cold well for 15 min. The cold solution of H2dpm2Ph was added dropwise to the vigorously stirred solution of Ti(NMe2)4. The reaction mixture was stirred at room temperature for 12 h. The volatiles were removed in vacuo to give a sticky red-orange solid. This solid was washed with cold pentane three times, which afforded analytically pure orange solid (50 mg, 44%). X-ray quality crystals were grown from a saturated solution in pentane. 1H NMR (C6D6, 500 MHz): δ = 1H NMR (500 MHz, C6D6) δ 7.71 (d, J = 8.1 Hz, 2H), 7.14 – 7.07 (m, 2H), 7.03 (d, J = 8.9 Hz, 2H), 6.60 (d, J = 2.6 Hz, 1H), 6.53 (d, J = 2.7 Hz, 1H), 6.48 (d, J = 2.7 Hz, 1H), 6.29 – 6.16 (m, 1H), 2.77 (s, 12H), 1.87 (s, 6H). 13C NMR (C6D6, 126 MHz,) δ 163.24, 161.74, 140.91, 133.81, 128.35, 128.31, 125.57, 124.14, 115.03, 112.68, 107.91, 101.78, 47.09, 40.08, 30.04. M.pt: 154-156 °C. Synthesis of Ti(pyr3,5-CF3Ph-C(CH3)2-pyr)(NMe2)2 (5d) In 20 mL scintillation vial with a stir bar was loaded with Ti(NMe2)4 (0.058 g, 0.26 mmol, 1 equiv) and diethyl ether (3 mL). A 20 mL scintillation vial was loaded with pyr3,5-CF3Ph-C(CH3)2- 55 pyr (0.100 g, 0.26 mmol, 1 equiv) and diethyl ether (3 mL). Both solutions were then cooled in a liquid nitrogen cold well for 15 min. The cold solution of pyr3,5-CF3Ph-C(CH3)2-pyr was added dropwise to the vigorously stirred solution of Ti(NMe2)4. The reaction mixture was stirred at room temperature for 12 h. The volatiles were removed in vacuo to give a viscous red-orange solid. Recrystallization from ether/n-hexane afforded product as orange crystals (60 mg, 45%). 1H NMR (C6D6, 500 MHz): δ 8.09 (s, 2H), 7.64 (s, 1H), 6.88 (dd, J = 2.3, 1.3 Hz, 1H), 6.37 (s, 2H), 6.31 (dd, J = 3.1, 2.3 Hz, 1H), 6.17 (dd, J = 3.0, 1.3 Hz, 1H), 2.58 (s, 12H), 1.78 (s, 6H). 13C{1H} NMR (C6D6, 126 MHz): δ 164.43, 161.08, 137.53, 136.91, 131.78, 131.51, 125.70, 125.25, 120.28, 113.33, 112.74, 109.61, 104.61, 46.81, 39.97, 29.89. 19F NMR (C6D6, 470 MHz, 25 °C) δ –62.58. M.pt.: 143-144 °C. Synthesis of Ti(pyr3,5-diMePh-C(CH3)2-pyr)(NMe2)2 (5e) In 20 mL scintillation vial with a micro stir bar was loaded with Ti(NMe2)4 (0.292 g, 1.30 mmol, 1 equiv) and diethyl ether (3 mL). A 20 mL scintillation vial was loaded with pyr3,5-diMePh-C(CH3)2- pyr (0.362 g, 1.30 mmol, 1 equiv) and diethyl ether (3 mL). Both solutions were then cooled in a liquid nitrogen cold well for 15 min. The cold solution of pyr3,5-diMePh-C(CH3)2-pyr was added dropwise to the vigorously stirred solution of Ti(NMe2)4. The reaction mixture was stirred at room temperature for 12 h. The volatiles were removed in vacuo to give a sticky red-orange solid. Recrystallization from ether/n-hexane afforded the product as orange crystals (300 mg, 56%). 1H NMR (C6D6, 500 MHz): δ 7.51 (s, 2H), 7.06 (dd, J = 2.6, 1.3 Hz, 1H), 6.71 (s, 1H), 6.67 (d, J = 56 2.7 Hz, 1H), 6.58-6.55 (m, 1H), 6.51 (d, J = 2.7 Hz, 1H), 6.25 (dd, J = 3.1, 1.2 Hz, 1H), 2.80 (s, 12H), 2.13 (s, 6H), 1.89 (s, 6H). 13C{1H} NMR (C6D6, 126 MHz, 25 °C): δ 163.14, 161.72, 141.57, 137.63, 133.67, 129.98, 124.22, 123.62, 115.14, 112.43, 107.88, 101.79, 47.11, 40.12, 30.10, 21.30. M.pt.: 157-159 °C. Elemental Analysis: Calcd for C24H35N4Ti. C: 66.99; H: 7.82; N: 13.59. Found: C: 66.90; H: 7.92; N: 13.30. Synthesis of complex Ti(pyr-CH2-ind3-Me)(NMe2)2 (6a) In 20 mL scintillation vial with a stir bar was loaded with Ti(NMe2)4 (0.320 g, 1.16 mmol, 1 equiv) and diethyl ether (3 mL). A 20 mL scintillation vial was loaded with Hpyr-CH2-Hind3Me (0.300 g, 1.16 mmol, 1equiv) and diethyl ether (3 mL). Both solutions were then cooled in a liquid nitrogen cold well for 15 min. The cold solution of Hpyr-CH2-Hind3Me was added dropwise to the vigorously stirred solution of Ti(NMe2)4. The reaction mixture was stirred at room temperature for 12 h. The volatiles were removed in vacuo to give a sticky red-orange solid. Recrystallization from ether/n-hexane afforded product as orange crystals (250 mg, 50%). 1H NMR and 13C{1H} NMR show peaks for free dimethylamine. 1H NMR (C6D6, 500 MHz): 77.67 (d, J = 7.7 Hz, 1H), 7.62 (d, J = 7.9 Hz, 1H), 7.40 (td, J = 7.5, 1.4 Hz, 1H), 7.35 (td, J = 7.3, 1.2 Hz, 1H), 6.95 (t, J = 1.4 Hz, 1H), 6.13 (dd, J = 2.7, 1.3 Hz, 1H), 5.91 (dd, J = 2.7, 1.5 Hz, 1H), 4.02 (s, 2H), 3.12 (s, 3H), 2.98 (s, 12H). 13C{1H} NMR (C6D6, 126 MHz, 25 °C): δ 155.79, 153.42, 143.92, 130.36, 129.33, 120.52, 119.60, 117.14, 116.76, 116.29, 115.31, 105.99, 47.39, 29.16, 9.24. 57 Synthesis of Ti(pyr-CH2-ind-3-Me-5-OMe)(NMe2)2 (6b) In 20 mL scintillation vial with a stir bar was loaded with Ti(NMe2)4 (0.390 g, 1.66 mmol, 1 equiv) and diethyl ether (3 mL). A 20 mL scintillation vial was loaded with Hpyr-CH2-Hind3-Me-5- OMe (0.400 g, 1.66 mmol, 1 equiv) and diethyl ether (3 mL). Both solutions were then cooled in a liquid nitrogen cold well for 15 min. The cold solution of Hpyr-CH2-Hind3-Me-5-OMe was added dropwise to the vigorously stirred solution of Ti(NMe2)4. The reaction mixture was stirred at room temperature for 12 h. The volatiles were removed in vacuo to give a sticky red-orange solid. Recrystallization from ether/n-hexane afforded product as orange crystals (217 mg, 35%). 1H NMR (C6D6, 500 MHz) δ 7.50 (d, J = 8.2 Hz, 1H), 7.18 (d, J = 8.2 Hz, 2H), 6.96 (s, 1H), 6.13 (s, 1H), 5.91 (s, 1H), 4.02 (s, 2H), 3.67 (s, 3H), 2.99 (s, 13H), 2.22 (s, 3H). 13C{1H} NMR (C6D6, 126 MHz): δ 156.96, 155.14, 153.43, 139.26, 130.61, 128.42, 116.66, 116.19, 115.93, 109.94, 105.93, 99.56, 55.52, 47.42, 29.31, 9.36. Synthesis of Ti(phen-CH2-ind3-Me-5-F)(NMe2)2 (7a) In 20 mL scintillation vial with a stir bar was loaded with Ti(NMe2)4 (0.244g, 1.66 mmol, 1 equiv) and diethyl ether (5 mL). A 20 mL scintillation vial was loaded with Hphen-CH2-Hind3-Me- 5-F (0.400 g, 1.66 mmol, 1 equiv) and diethyl ether (5 mL). Both solutions were then cooled in a 58 liquid nitrogen cold well for 15 min. The cold solution of Hphen-CH2-Hind3-Me-5-F was added dropwise to the vigorously stirred solution of Ti(NMe2)4. The reaction mixture was stirred at room temperature for 12 h. The volatiles were removed in vacuo to give a sticky red-orange solid. Recrystallization from ether/n-hexane afforded the product as orange crystals (310 mg, 62%). 1H NMR (500 MHz, C6D6) δ 7.43 (dd, J = 36.6, 2.4 Hz, 2H), 7.32 – 7.27 (m, 1H), 7.09 (dd, J = 7.6, 1.8 Hz, 2H), 3.00 (s, 12H), 2.21 (s, 3H), 1.53 (s, 9H), 1.38 (s, 9H). 13C NMR (126 MHz, C6D6) δ 160.64, 160.15, 158.29, 144.03, 140.38, 140.05, 135.85, 132.20 (d, J = 9.5 Hz), 130.14, 125.87, 122.21, 115.18 (d, J = 9.5 Hz), 108.88 (d, J = 25.4 Hz), 107.54, 103.26 (d, J = 22.7 Hz), 44.49, 35.40, 34.57, 32.47, 31.89, 30.55, 8.61. 19F NMR (470 MHz, C6D6) δ -124.88 (q, J = 8.0 Hz). General Procedure for Kinetics The procedure from the previous study was followed.7 All manipulations were conducted in an N2 glovebox. All measurements are done using volumetric syringes for accurate volumes. A 2 mL volumetric flask was loaded the titanium precatalyst (10 mol%, 0.1 mmol) and ferrocene (56 mg, 0.3 mmol). Toluene-d8 (~0.75 mL) was added to the volumetric flask to completely dissolve the solids. Next, aniline (911 μL, 10 mmol) and 1-phenylpropyne (125 μL, 1.0 mmol) were added to the volumetric flask. Lastly, the solution was diluted up to 2 mL with toluene-d8. The solution was mixed via a pipette, i.e., the solution was drawn up into the pipette and dispensed back into the volumetric flask, five times to ensure the homogeneity of the solution. Then, 0.75 mL solution was loaded into a threaded J. Young NMR tube that and removed from the dry box. This solution was heated at 75 °C in the NMR spectrometer (Varian Inova 600). The relative 1-phenylpropyne versus ferrocene concentration was monitored as a function of time. The fits are exponential decay of the starting material. The expression used to fit the data was 𝑌𝑡 = 𝑌∞ + (𝑌0 − 𝑌∞)𝑒−𝑘𝑜𝑏𝑠𝑡,8 where Y = concentration at time t (Yt), infinity (Y∞), or at the start of the reaction (Y0). Each kinetic experiment 59 was done in triplicates and the average value was used to represent rate constant for the catalyst under investigation. 60 Representative Plots from Kinetics Figure 2.16. Plot of [1-phenylpropyne] vs time with Ti(dpm2-Me)(NMe2)2 (5a). Figure 2.17. Plot of [1-phenylpropyne] vs time with Ti(dpm3-Me)(NMe2)2 (5b). 61 Figure 2.18. Plot of [1-phenylpropyne] vs time with Ti(dpm2-Ph)(NMe2)2 (5c). Figure 2.19. Plot of [1-phenylpropyne] vs time with Ti(pyr3,5-CF3Ph-C(CH3)2-pyr)(NMe2)2 (5d). 62 Figure 2.20. Plot of [1-phenylpropyne] vs time with Ti(pyr3,5-diMePh-C(CH3)2-pyr)(NMe2)2 (5e). Figure 2.21. Plot of [1-phenylpropyne] vs time with Ti(pyr-CH2-ind3-Me)(NMe2)2 (6a). 63 Figure 2.22. Plot of [1-phenylpropyne] vs time with Ti(pyr-CH2-ind3-Me-5-OMe)(NMe2)2 (6b). Figure 2.23. Plot of [1-phenylpropyne] vs time with Ti(phen-CH2-ind3-Me-5-F)(NMe2)2 (7a). 64 Figure 2.24. Plot of [1-phenylpropyne] vs time with Ti(phen-CH2-ind3-Me)(NMe2)2 (7b). Figure 2.25. Plot of [1-phenylpropyne] vs time with Ti(phen-CH2-py)(NMe2)2 (7c). 65 Gas Chromatography To show that there are no side reactions for the fastest catalyst (5d) examined here, we ran the reaction and examined products by GC. The GC sample was prepared by diluting the crude reaction mixture after passing through silica plug to remove metal. The hydroamination of 1- phenylpropyne and aniline could give 2 regioisomers as shown in Figure 2.26. The chromatogram here shows two peaks at retention time 9.43 mins and 9.37 mins, corresponding to two imine isomers C (major) and D (minor). Some of the imine product was hydrolyzed to ketone during silica plug filtration, leading to ketone E (retention time = 5.23 mins). Peaks at 4.05, 4.70, 10.98 mins corresponds to aniline (A), 1-phenylpropyne (B), and H2dpm2-(3,5-diCF3Ph) (F). No side reactions were observed. Figure 2.26. (top) Possible products from 1-phenylpropyne and aniline hydroamination. (bottom) Gas Chromatograph from the crude reaction mixture. 66 Figure 2.26 (cont’d) Ligand 1-5 Isomerization A 0.03 M solution of 5d was prepared in the J-young NMR tube. 1H NMR was collected between –92 °C to 25 °C (Figure 2.11). Temperature calibration was done using a methanol sample in the NMR probe. Using Eyring Equation (Eq. 2.7), the slope and intercept of ln(kobs/T) vs 1/T plot was used to calculate enthalpy and entropy associated with 1-5 ligand isomerism. 𝑙𝑛 ( 𝑘𝑜𝑏𝑠 𝑇 ) = − Δ𝐻‡ 𝑅𝑇 + 𝑙𝑛 𝑘𝐵 ℎ + Δ𝑆‡ 𝑅 Eq. 2.7 Temperature (K) 181.9 204.7 246.0 -1 Rate (s ) -3 1/T (10 ) ln(k/T) 20.2 499 25,600 5.498 4.886 4.310 -2.199 0.890 4.704 67 The enthalpic barrier, H‡, for 1-5 ligand isomerism of catalyst 5d = 9.5 ± 0.2 kcal/mol, and S‡ = 1.2 ± 1.1 cal/mol. The Gibbs free energy, G‡, at room temperature is 9.1 ± 0.6 kcal/mol. The lower barrier between A and B suggests that these will be quickly exchanged at the accessible temperatures on the NMR timescale. As a result, we assume we are measuring the barrier (TS2‡) between C and the fast-exchanging B/A system. 68 Modeling of the Kinetic Data The data used in the modeling is in Table S1 below. The LDP and %Vbur values are from our previous studies. The rate constants for complexes 1-4 were previously published.7 The parameters for new compounds (5-6) were all known from the previous study. For the unsymmetrical 5-6 complexes, “side 1” is simply the side of the ligand with the larger LDP value. Using the larger %Vbur as “side 1” does not give as good a model, statistically speaking. Table 2.3. Data for the modeling of natural variables. Complex LDP1 LDP2 %Vbur1 %Vbur2 kobs x 104 (s–1) 1a 3a 1b 1c 1d 1e 3b 4 2a 2b 5a 5b 5c 5d 5e 6a 6b u(i) 13.64 12.49 13.46 14.03 13.91 14.32 12.66 11.98 11.87 11.82 13.64 13.46 14.03 14.32 13.91 13.64 13.64 13.07 Delta u(i) 1.25 13.64 12.49 13.46 14.03 13.91 14.32 12.66 11.98 11.87 11.82 13.46 13.09 13.64 13.64 13.64 12.49 12.22 13.07 1.25 20.4 22.6 23.7 27.1 26.7 27.9 22.6 21.6 21.5 21.5 20.4 23.7 27.1 27.9 26.7 20.4 20.4 20.4 22.6 23.7 27.1 26.7 27.9 22.6 21.6 21.5 21.5 23.7 23.1 20.4 20.4 20.4 22.6 23.3 4.16 0.66 1.35 0.52 0.55 0.58 1.08 0.43 0.24 0.05 3.10 1.90 3.95 5.50 3.46 3.64 4.40 kcalc 4.11 0.60 1.66 0.51 0.55 0.53 0.91 0.32 0.18 0.09 2.69 2.32 4.14 4.82 3.86 4.18 4.10 Abs diff 0.05 0.06 0.31 0.01 0.00 0.05 0.17 0.11 0.06 0.04 0.41 0.42 0.19 0.68 0.40 0.54 0.30 24.15 3.75 24.15 3.75 Ave. diff = 0.22 69 The parameters of the multivariate analysis were found using Microsoft Excel (16.78.3). The output is shown below. Regression statistics for Table 2.3. Regression Statistics Multiple R 0.98494985 R Square 0.97012621 Adjusted R Square 0.96016828 Standard Error 0.35649998 Observations 17 ANOVA Regression Residual Total df 4 12 16 SS MS F Significance F 49.5265641 12.381641 97.4224799 4.8481E-09 1.52510686 0.12709224 51.051671 P-value Lower 95% Upper 95% Coeff. Standard Error a b c d e -7.58 2.80 -0.98 -0.159 -0.484 1.57 0.211 0.264 0.0443 0.0421 t Stat -4.84 13.3 0.000408 1.525E-08 -3.734 0.00287 -3.59 0.00374 -11.48 7.94E-08 -10.99 2.344 -1.561 -0.255 -0.575 -4.165 3.262 -0.410 -0.062 -0.392 The fits were done using both the “natural” variables, the direct LDP and %Vbur values, and scaled variables. The natural variables give a model that can be used to calculate the rate constant of a new catalyst if the LDP and %Vbur values are known or can be accurately estimated. The 70 scaled variables allow direct comparison between the different coefficients and comparison of electronic and steric factors. Modeling with the Ligand Site Determined by Sterics We also did the modeling with “side 1” determined by the %Vbur value. The tabulated data then become: Table 2.4. Data for the modeling of natural variables with ligand sites determined by sterics. Complex LDP1 LDP2 %Vbur1 %Vbur2 1a 3a 1b 1c 1d 1e 3b 4 2a 2b 5a 5b 5c 5d 5e 6a 6b 13.64 13.64 12.49 12.49 13.46 13.46 14.03 14.03 13.91 13.91 14.32 14.32 12.66 12.66 11.98 11.98 11.87 11.87 11.82 11.82 13.46 13.64 13.46 13.09 14.03 13.64 14.32 13.64 13.91 13.64 12.49 13.64 12.22 13.64 20.4 22.6 23.7 27.1 26.7 27.9 22.6 21.6 21.5 21.5 23.7 23.7 27.1 27.9 26.7 22.6 23.3 20.4 22.6 23.7 27.1 26.7 27.9 22.6 21.6 21.5 21.5 20.4 23.1 20.4 20.4 20.4 20.4 20.4 71 kobs x 104 (s–1) 4.16 0.66 1.35 0.52 0.55 0.58 1.08 0.43 0.24 0.05 3.10 1.90 3.95 5.50 3.46 3.64 4.40 k(calc) Abs diff 3.82 0.67 1.61 0.56 0.60 0.57 0.94 0.42 0.29 0.21 3.99 1.45 4.30 4.39 4.25 3.77 3.77 average 0.34 0.00 0.26 0.04 0.05 0.01 0.14 0.02 0.05 0.16 0.89 0.45 0.35 1.11 0.79 0.13 0.63 0.32 Table 2.4 (cont’d) Regression Statistics Multiple R R Square Adjusted R Square 0.9639 0.9290 0.9054 Standard Error 0.5495 Observations 17 ANOVA Regressio n Residual Total df 4 12 16 SS MS F Significanc e F 47.4284 3.6233 51.0517 11.8571 0.3019 39.2699 0.0000 Coefficient s Standar d Error t Stat P-value a b c d e -6.5566 0.1578 1.4718 0.0612 -0.6422 2.5051 0.3354 0.3100 0.0963 0.0603 -2.6173 0.4705 4.7485 0.6357 -10.6450 0.0225 0.6464 0.0005 0.5369 0.0000 Lower 95% - 12.0147 -0.5730 0.7965 -0.1485 -0.7737 Upper 95% -1.0985 0.8887 2.1471 0.2709 -0.5108 From these sterically-driven values, we can give a different set of parameter with their 95% confidence intervals. As shown, only two of the parameters (excluding the intercept) is above the confidence interval in this case. 72 Table 2.5. Coefficients for steric model. Coefficients Steric Model with 95% confidence intervals a b c d e –6.55 ± 5.46 0.16 ± 0.73 1.47 ± 0.68 0.061 ± 0.21 –0.64 ± 0.13 The statistics for this sterically-driven case are significantly worse. For example, the R2 here is 0.93, compared to R2 = 0.97 for the case above where the electronics are determining the ligand position. Modeling of the Scaled Values The scaling was done using the equation below, where xi = scaled variable, ui = natural variable, ui 0 = midpoint of the range of the natural variables, and ∆ui = the difference between the midpoint and the high value (half the full range). The equations for the calculation of ui 0 and Dui are shown below. The values in Table S1 were scaled (–1 to +1) in this way and are shown in Table 2.6. In order to determine if the electronic parameter (LDP) or the steric parameter (%Vbur) determined the position of the ligand in the key transition state, we also modelled the data where 73 xi=ui-ui0DuiDui=uihigh-ui0ui0=uihigh+uilow2 the larger %Vbur determined ligand one. The statistics for this model are somewhat worse than if the electronics are used to determine the ligand label, suggesting the electronics of the ligand determine the orientation in the key transition state. Table 2.6. Data for the modeling of scaled variables. Complex LDP1 LDP2 %Vbur1 %Vbur2 kobs x 104 (s–1) 1a 3a 1b 1c 1d 1e 3b 4 2a 2b 5a 5b 5c 5d 5e 6a 6b 0.456 -0.464 0.312 0.768 0.672 1 -0.328 -0.872 -0.96 -1 0.456 0.312 0.768 1 0.672 0.456 0.456 0.456 -0.464 0.312 0.768 0.672 1 -0.328 -0.872 -0.96 -1 0.312 0.016 0.456 0.456 0.456 -0.464 -0.68 -1 -1 -0.4133333 -0.4133333 -0.12 -0.12 0.78666667 0.78666667 0.68 1 0.68 1 -0.4133333 -0.4133333 -0.68 -0.68 -0.7066667 -0.7066667 -0.7066667 -0.7066667 -1 -0.12 0.78666667 1 0.68 -1 -1 -0.12 -0.28 -1 -1 -1 -0.4133333 -0.2266667 4.16 0.66 1.35 0.52 0.55 0.58 1.08 0.43 0.24 0.05 3.10 1.90 3.95 5.50 3.46 3.64 4.40 kcalc 4.11 0.60 1.66 0.51 0.55 0.53 0.91 0.32 0.18 0.09 2.69 2.32 4.14 4.82 3.86 4.18 4.10 74 Table 2.6 (cont’d) Regression Statistics Multiple R R Square Adjusted R Square Standard Error Observations 0.96 0.93 0.91 0.55 17 ANOVA Regression Residual Total df 4 12 16 SS 47.43 3.62 51.05 MS 11.86 0.30 F 39.27 Significance F 8.40E-07 Coeff. -6.56 0.16 1.47 0.06 -0.64 a b c d e Standard Error 2.51 0.34 0.31 0.10 0.06 t Stat P-value -2.62 0.47 4.75 0.64 -10.65 0.02 0.65 0.00 0.54 1.82E-07 Lower 95% - 12.01 -0.57 0.80 -0.15 -0.77 Upper 95% -1.10 0.89 2.15 0.27 -0.51 75 NMR Spectra for Ligands and Titanium Catalysts Figure 2.27. 1H NMR of (3-methyl-1H-indol-2-yl)methanol. 76 Figure 2.28. 13C NMR of (3-methyl-1H-indol-2-yl)methanol. 77 Figure 2.29. 1H NMR of 2-((1H-pyrrol-2-yl)methyl)-3-methylindole. 78 Figure 2.30. 1H NMR of 2-((1H-pyrrol-2-yl)methyl)-3-methylindole. 79 Figure 2.31. 1H NMR of 2,4-di-tert-butyl-6-((5-fluoro-3-methyl-1H-indol-2-yl)methyl)phenol. 80 Figure 2.32. 13C NMR of 2,4-di-tert-butyl-6-((5-fluoro-3-methyl-1H-indol-2-yl)methyl)phenol. 81 Figure 2.33. 19F NMR of 2,4-di-tert-butyl-6-((5-fluoro-3-methyl-1H-indol-2-yl)methyl)phenol. 82 Figure 2.34. 1H NMR of Ti(dpm2-Me)(NMe2)2 (5a). 83 Figure 2.35. 13C NMR of Ti(dpm2-Me)(NMe2)2 (5a). 84 Figure 2.36. 1H NMR of Ti(dpm2,2’,3-TriMe)(NMe2)2 (5b). 85 Figure 2.37. 13C NMR of Ti(dpm2,2’,3-TriMe)(NMe2)2 (5b). 86 Figure 2.38. 1H NMR of Ti(dpm2-Ph)(NMe2)2 (5c). 87 Figure 2.39. 13C NMR of Ti(dpm2-Ph)(NMe2)2 (5c). 88 Figure 2.40. 1H NMR of Ti(pyr3,5-CF3Ph-C(CH3)2-pyr)(NMe2)2 (5d). 89 -1.0-0.50.00.51.01.52.02.53.03.54.04.55.05.56.06.57.07.58.08.59.09.510.010.511.011.512.012.5f1 (ppm)-500050010001500200025003000350040004500500055006000650070006.4412.311.000.972.101.010.931.971.782.586.166.176.176.176.306.316.316.326.366.376.376.886.886.886.887.16(cid:9)C6D67.648.09 Figure 2.41. 13C NMR of Ti(pyr3,5-CF3Ph-C(CH3)2-pyr)(NMe2)2 (5d). 90 -100102030405060708090100110120130140150160170180190200210220230f1 (ppm)-5005010015020025030035040045050055060065070075080085090029.8939.9746.81104.61109.61112.74113.33120.28125.25125.70128.06(cid:9)C6D6131.51131.78136.91137.53161.08164.43 Figure 2.42. 19F NMR of Ti(pyr3,5-CF3Ph-C(CH3)2-pyr)(NMe2)2 (5d). 91 Figure 2.43. 1H NMR of Ti(pyr3,5-diMePh-C(CH3)2-pyr)(NMe2)2 (5e). 92 -1.0-0.50.00.51.01.52.02.53.03.54.04.55.05.56.06.57.07.58.08.59.09.510.010.511.011.512.0f1 (ppm)-100010020030040050060070080090010001100120013001400150016001700180019006.586.8413.661.001.120.951.180.971.162.141.892.132.806.256.256.266.266.516.526.566.566.576.676.686.717.067.067.067.077.16(cid:9)C6D67.51 Figure 2.44. 13C NMR of Ti(pyr3,5-diMePh-C(CH3)2-pyr)(NMe2)2 (5e). 93 Figure 2.45. 1H NMR of Ti(pyr-CH2-ind3-Me)(NMe2)2 (6a). 94 Figure 2.46. 13C NMR of Ti(pyr-CH2-ind3-Me)(NMe2)2 (6a). 95 -100102030405060708090100110120130140150160170180190200210220230f1 (ppm)-200204060801001201401601802002202402602803003209.2429.1644.0647.39105.99115.31116.29116.76117.14119.60120.52128.06(cid:9)C6D6129.33130.36143.92153.42155.79 Figure 2.47. 1H NMR of Ti(pyr-CH2-ind3-Me-5-OMe)(NMe2)2 (6b). 96 -101234567891011121314f1 (ppm)0500100015002000250030003.1713.113.262.181.031.000.932.291.082.222.993.674.025.916.136.967.16(cid:9)C6D67.177.187.497.51 Figure 2.48. 13C NMR of Ti(pyr-CH2-ind3-Me-5-OMe)(NMe2)2 (6b). 97 -100102030405060708090100110120130140150160170180190200210220230f1 (ppm)-500501001502002503003504004505005506006507007508008509.3629.3147.4255.5299.56105.93109.94115.93116.19116.66128.06(cid:9)C6D6128.42130.61139.26153.43155.14156.96 Figure 2.49. 1H NMR of Ti(phen-CH2-ind3-Me-5-F)(NMe2)2 (7a). 98 Figure 2.50. 13C NMR of Ti(phen-CH2-ind3-Me-5-F)(NMe2)2 (7a). 99 Figure 2.51. 19F NMR of Ti(phen-CH2-ind3-Me-5-F)(NMe2)2 (7a) . 100 Single Crystal X-ray Diffraction Thermal ellipsoids of titanium precatalysts are drawn with a 50% probability level. The violet, blue, green, grey, and white spheres represent titanium, nitrogen, fluorine, carbon, and hydrogen atoms, respectively. Figure 2.52. Structure of Ti(dpm2-Me)(NMe2)2 (5a) recrystallized from toluene/n-hexane. 101 Table 2.7. Crystallographic data and structural refinement of Ti(dpm2-Me)(NMe2)2 (5a). Empirical formula Formula weight Temperature/K Crystal system Space group a/Å b/Å c/Å α/° β/° γ/° Volume/Å3 Z ρcalcg/cm3 μ/mm-1 F(000) Crystal size/mm3 Radiation 2Θ range for data collection/° Index ranges Reflections collected Independent reflections Data/restraints/parameters Goodness-of-fit on F2 Final R indexes [I>=2σ (I)] Final R indexes [all data] Largest diff. peak/hole / e Å-3 C16H26N4Ti 322.31 100.00(10) monoclinic P21/n 8.09880(10) 12.67000(10) 16.3955(2) 90 98.0760(10) 90 1665.69(3) 4 1.285 4.334 688.0 0.188 × 0.143 × 0.105 Cu Kα (λ = 1.54184) 8.854 to 160.594 -10 ≤ h ≤ 10, -15 ≤ k ≤ 15, -19 ≤ l ≤ 20 24926 3606 [Rint = 0.0352, Rsigma = 0.0219] 3606/0/197 1.101 R1 = 0.0293, wR2 = 0.0791 R1 = 0.0310, wR2 = 0.0804 0.38/-0.35 102 Figure 2.53. Structure of Ti(pyr-CH2-ind3-Me)(NMe2)2 (6a) recrystallized from ether/n-hexane. 103 Table 2.8. Crystallographic data and structural refinement of Ti(pyr-CH2-ind3-Me)(NMe2)2 (6a). Empirical formula Formula weight Temperature/K Crystal system Space group a/Å b/Å c/Å α/° β/° γ/° Volume/Å3 Z ρcalcg/cm3 μ/mm-1 F(000) Crystal size/mm3 Radiation C18H24N4Ti 344.31 100.00(10) monoclinic P21/n 8.63010(10) 19.5622(3) 10.6847(2) 90 108.138(2) 90 1714.20(5) 4 1.334 4.253 728.0 0.184 × 0.101 × 0.091 Cu Kα (λ = 1.54184) 2Θ range for data collection/° 9.042 to 159.724 Index ranges Reflections collected Independent reflections -5 ≤ h ≤ 10, -24 ≤ k ≤ 24, -13 ≤ l ≤ 13 15019 3650 [Rint = 0.0382, Rsigma = 0.0321] Data/restraints/parameters Goodness-of-fit on F2 Final R indexes [I>=2σ (I)] Final R indexes [all data] Largest diff. peak/hole / e Å-3 3650/0/213 1.088 R1 = 0.0323, wR2 = 0.0835 R1 = 0.0343, wR2 = 0.0847 0.31/-0.32 104 Figure 2.54. Structure of Ti(phen-CH2-ind3-Me-5-F)(NMe2)2 (7a)recrystallized from ether/n-hexane. 105 Table 2.9. Crystallographic data and structural refinement of Ti(phen-CH2-ind3-Me-5-F)(NMe2)2 (7a). Empirical formula Formula weight Temperature/K Crystal system Space group a/Å b/Å c/Å α/° β/° γ/° Volume/Å3 Z ρcalcg/cm3 μ/mm-1 F(000) Crystal size/mm3 Radiation 2Θ range for data collection/° Index ranges Reflections collected Independent reflections Data/restraints/parameters Goodness-of-fit on F2 Final R indexes [I>=2σ (I)] Final R indexes [all data] Largest diff. peak/hole / e Å-3 Flack parameter C28H40FN3OTi 501.53 100(2) monoclinic Pc 12.1882(2) 11.0797(2) 20.5023(3) 90 94.452(2) 90 2760.31(8) 4 1.207 2.864 1072.0 0.059 × 0.053 × 0.045 CuKα (λ = 1.54184) 7.274 to 159.56 -15 ≤ h ≤ 15, -13 ≤ k ≤ 14, -25 ≤ l ≤ 26 27942 9697 [Rint = 0.0516, Rsigma = 0.0570] 9697/2/635 1.047 R1 = 0.0434, wR2 = 0.1059 R1 = 0.0507, wR2 = 0.1104 0.32/-0.42 -0.004(6) 106 REFERENCES (1) Smil, V. Enriching the Earth: Fritz Haber, Carl Bosch, and the Transformation of World Food Production; MIT Press: Cambridge, Mass, 2001. (2) Ogba, O. M.; Warner, N. C.; O’Leary, D. J.; Grubbs, R. H. Recent Advances in Ruthenium- Based Olefin Metathesis. Chem. Soc. Rev. 2018, 47, 4510–4544. (3) Miyaura, Norio.; Suzuki, Akira. Palladium-Catalyzed Cross-Coupling Reactions of Organoboron Compounds. Chem. Rev. 1995, 95, 2457–2483. (4) Jourdant, A.; González-Zamora, E.; Zhu, J. Wilkinson’s Catalyst Catalyzed Selective Hydrogenation of Olefin in the Presence of an Aromatic Nitro Function: A Remarkable Solvent Effect. J. Org. Chem. 2002, 67, 3163–3164. (5) Tolman, C. A. Steric Effects of Phosphorus Ligands in Organometallic Chemistry and Homogeneous Catalysis. Chem. Rev. 1977, 77, 313–348. (6) Alt, H. G.; Köppl, A. Effect of the Nature of Metallocene Complexes of Group IV Metals on Their Performance in Catalytic Ethylene and Propylene Polymerization. Chem. Rev. 2000, 100, 1205–1222. (7) Katsuki, T.; Sharpless, K. B. The First Practical Method for Asymmetric Epoxidation. J. Am. Chem. Soc. 1980, 102, 5974–5976. (8) Sharpless, K. B. Searching for New Reactivity (Nobel Lecture) Copyright© The Nobel Foundation 2002. We Thank the Nobel Foundation, Stockholm, for Permission to Print This Lecture. Adapted with the Permission of the Editors from “Coelacanths and Catalysis”: K. B. Sharpless, Tetrahedron 1994, 50, 4235. Angew. Chem. Int. Ed. 2002, 41, 2024. (9) Odom, A. L.; McDaniel, T. J. Titanium-Catalyzed Multicomponent Couplings: Efficient One- Pot Syntheses of Nitrogen Heterocycles. Acc. Chem. Res. 2015, 48, 2822–2833. (10) Majumder, S.; Gipson, K. R.; Staples, R. J.; Odom, A. L. Pyrazole Synthesis Using a Titanium‐Catalyzed Multicomponent Coupling Reaction and Synthesis of Withasomnine. Adv Synth Catal 2009, 351, 2013–2023. (11) Barnea, E.; Majumder, S.; Staples, R. J.; Odom, A. L. One-Step Route to 2,3-Diaminopyrroles Using a Titanium-Catalyzed Four-Component Coupling. Organometallics 2009, 28, 3876– 3881. (12) Dissanayake, A. A.; Odom, A. L. Single-Step Synthesis of Pyrazoles Using Titanium Catalysis. Chem. Commun. 2012, 48, 440–442. (13) Majumder, S.; Odom, A. L. Titanium Catalyzed One-Pot Multicomponent Coupling Reactions for Direct Access to Substituted Pyrimidines. Tetrahedron 2010, 66, 3152–3158. 107 (14) Majumder, S.; Gipson, K. R.; Odom, A. L. A Multicomponent Coupling Sequence for Direct Access to Substituted Quinolines. Org. Lett. 2009, 11, 4720–4723. (15) Zhang, D.; Hou, Z.; Aldrich, K. E.; Lockwood, L.; Odom, A. L.; Liby, K. T. A Novel Nrf2 Pathway Inhibitor Sensitizes Keap1-Mutant Lung Cancer Cells to Chemotherapy. Molecular Cancer Therapeutics 2021, 20, 1692–1701. (16) Garcia, P.; Lau, Y. Y.; Perry, M. R.; Schafer, L. L. Phosphoramidate Tantalum Complexes for Room‐Temperature C H Functionalization: Hydroaminoalkylation Catalysis. Angew Chem Int Ed 2013, 52, 9144–9148. (17) DiFranco, S. A.; Maciulis, N. A.; Staples, R. J.; Batrice, R. J.; Odom, A. L. Evaluation of Donor and Steric Properties of Anionic Ligands on High Valent Transition Metals. Inorg. Chem. 2012, 51, 1187–1200. (18) Billow, B. S.; McDaniel, T. J.; Odom, A. L. Quantifying Ligand Effects in High-Oxidation- State Metal Catalysis. Nature Chem 2017, 9, 837–842. (19) Falivene, L.; Credendino, R.; Poater, A.; Petta, A.; Serra, L.; Oliva, R.; Scarano, V.; Cavallo, L. SambVca 2. A Web Tool for Analyzing Catalytic Pockets with Topographic Steric Maps. Organometallics 2016, 35, 2286–2293. (20) Schrock, R. R. Recent Advances in High Oxidation State Mo and W Imido Alkylidene Chemistry. Chem. Rev. 2009, 109, 3211–3226. (21) Poater, A.; Solans-Monfort, X.; Clot, E.; Copéret, C.; Eisenstein, O. Understanding d 0 -Olefin Metathesis Catalysts: Which Metal, Which Ligands? J. Am. Chem. Soc. 2007, 129, 8207–8216. (22) Ishiyama, T.; Takagi, J.; Yonekawa, Y.; Hartwig, J. F.; Miyaura, N. Iridium‐Catalyzed Direct Borylation of Five‐Membered Heteroarenes by Bis(Pinacolato)Diboron: Regioselective, Stoichiometric, and Room Temperature Reactions. Adv Synth Catal 2003, 345, 1103–1106. (23) Cho, J.-Y.; Tse, M. K.; Holmes, D.; Maleczka, R. E.; Smith, M. R. Remarkably Selective Iridium Catalysts for the Elaboration of Aromatic C-H Bonds. Science 2002, 295, 305–308. (24) Kanwal, S.; Ann, N.-; Fatima, S.; Emwas, A.-H.; Alazmi, M.; Gao, X.; Ibrar, M.; Zaib Saleem, R. S.; Chotana, G. A. Facile Synthesis of NH-Free 5-(Hetero)Aryl-Pyrrole-2-Carboxylates by Catalytic C–H Borylation and Suzuki Coupling. Molecules 2020, 25, 2106. (25) Xu, J.; Rawal, V. H. Total Synthesis of (−)-Ambiguine P. J. Am. Chem. Soc. 2019, 141, 4820– 4823. (26) Swartz Ii, D. L.; Odom, A. L. Effects of 5,5-Substitution on Dipyrrolylmethane Ligand Isomerization. Dalton Trans. 2008, No. 32, 4254. (27) Kegley, S. E.; Pinhas, A. R. Problems and Solutions in Organometallic Chemistry; University Science Books: Mill Valley, Calif, 1986. 108 (28) Espenson, J. H. Chemical Kinetics and Reaction Mechanisms, 2. ed.; MacGraw-Hill series in advanced chemistry; McGraw Hill: New York, NY, 1995. (29) Straub, B. F.; Bergman, R. G. The Mechanism of Hydroamination of Allenes, Alkynes, and Alkenes Catalyzed by Cyclopentadienyltitanium-Imido Complexes: A Density Functional Study. Angew. Chem. Int. Ed. 2001, 40, 4632–4635. (30) McDaniel, T. J. Applications and Optimization of Titanium Catalysis; 2017. 109 CHAPTER 3: AN INDOLE-EFFECT FOR INCREASING THE RATE OF HYDROAMINATION REACTION USING HIGH- VALENT METAL CATALYSIS* *Under revisions Jena, R.; Hou, Z.; Lee, S.; Odom, A. L. An Indole-Effect for Increasing Rates of C–N Bond Formation in a Titanium Catalysis. Under revisions. 3.1 Introduction In the previous chapter, we discussed the effect of desymmetrizing the bidentate ligand for the titanium-catalyzed hydroamination reaction. To summarize, we found moving from a symmetrical bidentate ligand system1 to an unsymmetrical ligand system enabled us to find a catalyst faster than the fastest symmetrical catalyst known.2 As shown in Eq. 2.8, the five-parameter, four- variable model is beneficial for us to design better catalysts. The level of control with individual electronic (LDP) and steric variables (%Vbur) for either side of the unsymmetrical ligand allowed us to develop relatively faster catalysts. The unified model (for symmetrical and unsymmetrical systems) presented in the previous chapter (Table 2.2) can be used to predict the rate of the reaction if we have LDP and %Vbur values for the ligand system under consideration, given the mechanism followed stays the same. 110 Figure 3.1. Graphical representation of kobs vs kcalc for the comparison of symmetrical ligands with unsymmetrical ligands. Moreover, we found that unsymmetrical ligand-based catalysts are generally faster than their symmetrical counterparts (Figure 3.1). The unsymmetrical catalysts had a ligand system based on pyrrolyl, 3-methylindolyl, and phenoxide-based ligand systems (Figure 2.6). We assumed that they all followed the same mechanism for hydroamination as shown in Scheme 2.6 and thus could be represented by the model shown in Eq 2.8. In this chapter, we will discuss the catalysts containing 3-unsubstituted indolyl (3UI) as a part of the bidentate ligand which according to our previous model follows a different mechanism for alkyne hydroamination. 111 Figure 3.2. An example of 3-unsubstituted indolyl-based titanium precatalyst and 3-methyl indolyl-based titanium precatalyst. To understand how we uncovered these mechanistic changes, let’s delve into the differences between 3UI-Ti precatalysts and the catalysts used in the previous chapter (Figure 2.6). Our initial step was to employ the previous model and calculate the rate constant for 3UI-Ti precatalysts (Figure 3.3). Using equation 2.8 (shown below) with the natural parameters the rate constants of 3UI-Ti precatalysts were calculated. In Figure 3.3 the blue dots and the line of best fit are for symmetrical and unsymmetrical ligands presented in the previous chapter and the red diamonds are for the ligands containing 3UI as a part of the bidentate ligand. Clearly the calculated rate constants for 3UI-Ti precatalysts do not fit the previous model, leading us to consider the possibility of a mechanism change and need for a new model. 𝑘𝑜𝑏𝑠(× 104) = −7.6 + 2.8(𝐿𝐷𝑃)1 − 0.98(𝐿𝐷𝑃)2 − 0.16(%𝑉𝑏𝑢𝑟)1 − 0.48(%𝑉𝑏𝑢𝑟)2 Eq. 2.8 112 Figure 3.3. Plot for the experimentally observed (kobs) rate constant vs the model calculated rate constant (kcalc) for symmetrical and unsymmetrical ligands (blue dots), and 3UI-based unsymmetrical ligands (red diamonds). Further analysis of Figure 3.3 reveals that this alternative mechanism allowed us to find a catalyst many times faster than the predicted rate constant value. For instance, the red diamond on the right-hand side of the plot shows approximately 4 times higher observed rate constant (15.13 × 10-4s-1) than the predicted value from the model (4.15 × 10-4s-1). This project aimed to make a library of 3UI-Ti precatalysts and build a model for them to understand the additional factors involved in dramatically increasing the rate constant for the hydroamination reactions. The complex model for 3UI-Ti suggests that the 3-unsubstituted indolyl group can act as a proton shuttle to increase the rate (vide-infra). 113 3.2 Synthesis of 3-Unsubstituted Indolyl-Based Unsymmetrical Ligands and Titanium Precatalyst This project was completed in collaboration with Dr. Zhilin Hou and Dr. Seokjoo Lee. I carried out mechanistic studies, prepared titanium catalysts, examined their kinetics, and characterized many of the compounds in this chapter. Zhilin prepared many of the titanium precatalysts, carried out some of the kinetics, and did the deuterium labeling experiment. Seokjoo Lee prepared some of the ligands for the titanium precatalysts. For the discussion, I will be referencing the data collected by Zhilin and Seokjoo as well. We were able to synthesize a variety of compounds by introducing substituents on either side of the ligand, indolyl and pyrrolyl. The general approach followed for the synthesis of these ligands includes condensation of 2-methanol-indoles with pyrrole using a Lewis acid, which can be done by either of the following routes. Firstly, 2-methanol-indoles were synthesized from indole-2- esters if not commercially available. Then condensation of 2-methanol-indoles with pyrrole using a Lewis acid gave indolyl-pyrrolyl based unsymmetrical ligands. These bidentate ligands can be subjected to iridium-catalyzed borylation3 to put the substituent on the second position of the pyrrolyl side of the ligand followed by Suzuki coupling4 to introduce various -R groups on the pyrrole side of the ligand. 114 Scheme 3.1. Synthesis of unsymmetrica1 ligands using condensation reaction followed by functionalization of pyrrole. Secondly, one can synthesize 2-aryl pyrroles using well-established Sadighi couplings.5 The condensation of 2-methanol-indoles and 2-aryl pyrroles in the presence of Lewis acid gives 3UI- based unsymmetrical ligands. Once ligands were synthesized and purified, they were treated with commercially available Ti(NMe2)4 to give 3UI-Ti precatalysts for hydroamination reactions. 115 Scheme 3.2. Condensation of 2-arylpyrrole with 2-methanol-indoles and synthesis of titanium precatalyst. For this study, we synthesized 11 new precatalysts bearing a 3UI-based moiety on one half of the bidentate ligand (Figure 3.4). The numbering for these precatalysts continues from Chapter 2. Catalysts numbered 8a-f contain 3UI and substituted pyrrolyl on the other side of the ligand. Catalysts 9a-d have substituted indolyl (except position 3 of the indolyl) and an unsubstituted pyrrole on the other side of the ligand. For comparison, I have drawn catalysts 6a and 6b from our previous studies. All catalysts are shown in 1,1 configuration to show the substitution clearly, however, the ground state structure usually has an 5-pyrrolyl and an 1-indolyl. 116 Figure 3.4. Structure of all new precatalysts synthesized for this study. 3.3 Kinetics Data, Model Progression, and Discussions As mentioned before, we successfully synthesized 11 new 3UI-based titanium precatalysts (8 and 9). We used our standard kinetics conditions for alkyne hydroamination.2 The disappearance of 1-phenylpropyne over time was monitored to measure the rate of hydroamination of 1- phenylpropyne with an excess of aniline. The rate constants from each run of all precatalysts used in this study are shown in Table 3.1. The blue and green numbers next to each side of the bidentate ligand are values for LDP and %Vbur respectively. 117 Table 3.1. Kinetic data for catalysts all precatalysts used in this study with their respective LDP and %Vbur value for each side of the ligand. Catalyst Number Structure (LDP, %Vbur) Observed Rate Constants (10-4 s-1) 8a 8aʼ 8b 8c 8d 6.4 7.0 6.2 5.6 5.4 5.4 10.1 12.3 12.2 9.8 8.8 9.3 9.5 6.2 7.6 118 Table 3.1 (cont’d) 8e 8f 9a 9b 9c 9d 15.2 14.9 15.3 0.54 0.58 5.2 5.0 5.1 3.9 2.7 2.7 0.2 0.2 0.2 119 Table 3.1 (cont’d) 6a 6b 3.5 3.5 4.0 4.5 4.2 4.5 We first attempted to model the rate constants (Table 3.1) for the 3UI-containing ligands, 8 and 9, using our previous model on unsymmetrical ligands (Eq. 2.8). As we can see in Figure 3.3, the red diamonds are widely scattered in a horizontal pattern and don’t follow the model. Due to this poor fit, we went on to build a new model for 3UI-Ti precatalysts that would account for the differing mechanism. The new model is shown in Eq. 3.1, where LDP1 and %Vbur1 correspond to the pyrrolyl side of the bidentate ligand and LDP2 and %Vbur2 correspond to the 3UI side of the ligand. 𝑘 = 𝑎 + 𝑏(𝐿𝐷𝑃)1 + 𝑐(%𝑉𝑏𝑢𝑟)1 + 𝑑(𝐿𝐷𝑃)2 + 𝑒(%𝑉𝑏𝑢𝑟)2 Eq. 3.1 Utilizing the model specified in Eq. 3.1, we were able to discover a somewhat linear correlation between the experimental rate constant and the calculated rate constant for 3UI-Ti precatalysts. However, the fit is terrible with R2 = 0.78 (Figure 3.5), which tells us there are more factors involved with 3UI-ligands than just electronic and steric parameters of the ligands. 120 Figure 3.5. Plot for the experimentally observed (kobs) rate constant vs the model calculated rate constant for 3UI-Ti precatalysts using model Eq. 3.1. Let’s look at the hydroamination mechanism closely to understand the role of the 3-position of the indolyl ligand in increasing the reaction rate. The proposed mechanism for the hydroamination of alkynes involves the formation of titanium imide as the active catalyst. This titanium imido undergoes [2+2]-cycloaddition with alkyne to form a 4-membered azatitanacyclobutene ring. Protonation of the Ti‒C bond within the ring occurs when an additional amine binds to the metal center. Doye and coworkers have shown that the imide formation is the key step in this mechanism and Ti‒C protonolysis is slower than cyclo reversion.6 After the second protonolysis, enamine is released and the catalyst is regenerated. 121 Figure 3.6. Proposed mechanism for alkyne hydroamination. We believe the presence of hydrogens at the 3-position of indole is crucial for the mechanism change. In the catalytic cycle shown in Figure 3.6, coordinated amine aids the protonation of the Ti‒C bond. In the case of indoles, the 3rd position is an electron-rich site, and we hypothesized that this position could act as a proton shuttle in the new mechanism involving 3UI-Ti precatalysts (Figure 3.8) and facilitate protonation of the Ti‒C bond. The pairs 6a and 8a’ or 9b and 6b (Table 3.1) can be used to understand how the hydroamination reaction rate is affected by the presence of a methyl group at the 3-position of indole. In the case of 6a and 8a’ we noticed a 50% increase in reaction rate after removing the methyl group from the 3-position. However, for the pair 9b and 6b, the rate decreased after removing the 3-methyl group. Therefore, the fluctuation in rate constant value cannot be solely justified using electronic and steric properties of the ancillary ligand. Further, we investigated if the rate law of hydroamination changes while using 8a’ and 6a where the only difference is the presence of 3-methyl group in 6a. Initial rate measurements were carried out for catalysts 8a’ and 6a (details in the experimental section). The reaction conditions are shown 122 in Table 3.2 and 3.3 with the corresponding initial rate constants (Figure 3.7, slope of the [1- phenylpropyne] vs time plot). Table 3.2. Amount of reagents and catalyst used in initial rate method for Ti(NMe2)2(ind-CH2- pyr) (8a’). Expt No. 1 2 3 4 [Aniline] (mmol) 1 1 2 1 [1-phenylpropyne] (mmol) 1 1 1 2 [Catalyst] (mmol) 0.05 0.025 0.025 0.025 Initial rate (×10-4) 1.8 1.0 1.1 2.1 Table 3.3. Amount of reagents and catalyst used in initial rate method for Ti(NMe2)2(3-Meind- CH2-pyr) (6a). Expt No. 1 2 3 4 [Aniline] (mmol) 1 1 2 1 [1-phenylpropyne] (mmol) 1 1 1 2 [Catalyst] (mmol) 0.05 0.025 0.025 0.025 Initial rate (×10-4) 1.8 1.0 1.1 2.1 123 Figure 3.7. Overlay plot of [1-phenylpropyne] vs time for reaction conditions mentioned in Table 3.2 with Ti(NMe2)2(ind-CH2-pyr) (8a’) and Table 3.3 with Ti(NMe2)2(3-Meind-CH2-pyr) (6a). Using initial rates, we got the identical second-order rate law for both catalysts (1st order in alkyne and 1st order in catalyst). Naturally, having the same rate law may imply similar pathways, but there could be significant changes in mechanism that explain why 8a’ is roughly twice as rapid as 6a when the only difference is a 3-methyl group on the indolyl. Based on the experimental 124 evidence, we can postulate that the 3UI-ligands are increasing the reaction rate by affecting the rate of protonation (k1), the equilibrium to form titanium imide (K1), and/or the equilibrium reaction of the imide with the alkyne (K2) by acting as a proton shuttling agent through the indolyl (Figure 3.8). The cycloaddition of alkyne with titanium imide can be favored by the removal of protonated indoyl from the metal center, thus reducing the sterics around the Ti center. The hypothesized mechanism proposes a proton shuttling role for the indolyl's 3-position, similar to the mechanism proposed by Hao and Schafer for titanium amidates.7 Figure 3.8. Postulated mechanism using 3-unsubstituted indolyl where the indolyl acts as a proton shuttle. 125 We decided to explore the use of calculated (using DFT) proton affinity (PA) for the protonation of this site in the modeling. But before including PA in our model, we wanted to get some sort of experimental evidence for the protonation of this ligand’s site. Treatment of 8a with deuterated aniline followed by reaction with CD3OD resulted in the deuteration of the 3-position of the indolyl and all the positions of pyrrolyl as well (Scheme 3.3). The ratio of H:D from 1H NMR spectroscopy was about 1:1 for all these sites. As determined by proton NMR and GC-MS, the catalyst only underwent deuteration at the most acidic sites (i.e. the N‒H proton of the bidentate ligand).8 A control experiment was conducted to ascertain that deuterated aniline, rather than CD3OD, is accountable for the ligand’s deuteration. Scheme 3.3. Reaction of 8a with deuterated aniline followed by CD3OD to liberate the ligand. At this point, we were confident that the PA of the 3-position of the indolyl group should be considered while building a model for 3UI-Ti precatalysts. We propose that because the indolyl is likely the productive site of protonation, in order to be kinetically advantageous, this site should be protonated with energetic preference to the pyrrolyl side of the ligand (such as 8a). The PA was computed as the difference between the 3-positions of the heterocycles on each side of the bidentate ligand i.e., PA(pyrrolyl) – PA(indolyl) = ∆PA. The ∆PA is the difference in DFT calculated (M06/def2SVP) enthalpy for the two species shown below. Table 3.4 shows the ∆PA value of all the ligands used in this study (Figure 3.4). 126 Table 3.4. The difference in proton affinity between two sites of the ligand. M06/def2SVP 3H-pyrr 3H-Ind ∆PA (Hartrees) ∆PA (kcal/mol) 8a 8a' 8b 8c 8d 8e 8f 9a 9b 9c 9d -689.76926 -689.78642 0.017155 -611.31588 -611.33487 0.01899 -729.00489 -729.01761 0.012712 -920.4021 -920.41521 0.013118 -959.63282 -959.64389 0.011076 -1593.7472 -1593.7658 0.018618 -882.37803 -882.37704 -0.000992 -650.54382 -650.56482 0.020993 -809.59862 -809.6123 0.013685 -725.65745 -725.68186 0.02441 -840.00233 -840.02153 0.019201 10.8 11.9 8.0 8.2 7.0 11.7 -0.6 13.2 8.6 15.3 12.0 The model equation after the inclusion of the ∆PA term is shown below: 𝑘 = 𝑎 + 𝑏(𝐿𝐷𝑃)1 + 𝑐(%𝑉𝑏𝑢𝑟)1 + 𝑑(𝐿𝐷𝑃)2 + 𝑒(%𝑉𝑏𝑢𝑟)2 + 𝑓(∆𝑃𝐴) Eq. 3.2 The inclusion of the proton affinity difference (∆PA) led to a significant improvement in the linear model (R2 = 0.89) as compared to our initial model (Eq. 3.1, R2 = 0.77). 127 Figure 3.9. Plot for the experimentally observed (kobs) rate constant vs the model calculated rate constant for 3UI-Ti precatalysts using Eq. 3.2. We hypothesize that the 3-position of the indolyl acts as a proton shuttle. For any group to be effective in this role, the site must be protonated and deprotonated efficiently. In other words, it can’t be too basic or too acidic. As a result, we went on to investigate if there was a second-order effect in ∆PA (represented by ∆PA2). For new model, the data were scaled from –1 to +1, which gives much better results, especially for nonlinear models like the one here. This is because the natural values are of different absolute magnitudes, which can end up weighting natural value fits towards higher magnitude variables. As a result, one could make a variable more or less important in the model by, for example, changing the units, such as giving LDP in wavenumbers instead of kcal/mol. The scaling gives parameter weights that are only related to the importance of the variable in the model. By updating our model to Eq. 3.3, we saw an improved fit (R2 = 0.93). 128 𝑘 = 𝑎 + 𝑏(𝐿𝐷𝑃)1 + 𝑐(%𝑉𝑏𝑢𝑟)1 + 𝑑(𝐿𝐷𝑃)2 + 𝑒(%𝑉𝑏𝑢𝑟)2 + 𝑓(∆𝑃𝐴) + 𝑔(∆𝑃𝐴)2 Eq. 3.3 Figure 3.10. Plot for the experimentally observed (kobs) rate constant vs the model calculated rate constant for 3UI-Ti precatalysts using Eq. 3.3. Though the model has continued to improve, we still did not find it sufficient being so far from R2 = 1. In our previous studies, the linker between two sides of the ancillary ligand did not seem to have a noticeable effect.1,2 However, in the case of 3UI ligands, the rate of the reaction seemed to be dependent on the linker. The pair 8a and 8a’ were synthesized to understand the effect of the dimethyl linker between two sides of the ancillary ligand. Here we can see that 8a with a gem- dimethyl linker shows relatively faster catalysis than 8a’ which lacks this linker. The effect of the dimethyl could be inductive, or it could be that the gem-dimethyl substitution leads to faster reattachment of the indolyl ligand to the metal if it is removed during catalysis (Thorpe-Ingold effect).9 To incorporate the effect of a dimethyl linker we decided to add a dimethyl bridge and see 129 how our current model would respond to it. This was simply done as a noncontinuous variable since the only ligands in the set either had dimethylmethylene (assigned as “1”) or just methylene (assigned as “0”). The model then used is: 𝑘 = 𝑎 + 𝑏(𝐿𝐷𝑃)1 + 𝑐(%𝑉𝑏𝑢𝑟)1 + 𝑑(𝐿𝐷𝑃)2 + 𝑒(%𝑉𝑏𝑢𝑟)2 + 𝑓(∆𝑃𝐴) + 𝑔(∆𝑃𝐴)2 + ℎ(𝐵𝑟𝑑𝑔) Eq. 3.4 Figure 3.11. Plot for the experimentally observed (kobs) rate constant vs the model calculated rate constant for 3UI-Ti precatalysts using model Eq. 3.4. The addition of the bridging term improved the model to give an R2 = 0.97, making this addition seem necessary for a good fit. At this point we were quite satisfied with the model we had in hand (Eq. 3.4). With this we moved forward to analyze the residual plots which is a common practice in statistical and data analysis to assess the adequacy of the model’s fit to the data and to find any 130 pattern or trends that were overlooked during analysis. The residual plot (kobs ‒ kcalc) for the updated model (Eq. 3.4) is shown below. Besides the obvious problem of the slightly negative rate constants for the slow catalyst 9d, the model has something of a curvature below the zero line (Figure 3.12). In ideal conditions, if the model fits the data well, the residuals plot should exhibit a random scatter around zero. However, the curvature below the zero line in the residuals plot may indicate a systematic bias or trends in the model’s errors, especially when combined with slightly negative rate constants. This phenomenon in statistics implies that a second-order term is missing in the model.10 The left plot includes all the data points. The right plot shows the correlation between averaged rate constant for the catalysts vs residuals, this allows a better visual for detecting the curvature. Figure 3.12. Plot for the experimentally observed (kobs) rate constant vs residuals (from model Eq. 3.4). 131 As a result, we investigated every other possible second-order term that could come from the linear terms already described, mixed terms such as (LDP)(%Vbur) and squared terms such as (LDP)2 or (%Vbur)2. All of the terms were found to have coefficients of zero in the regression except for LDP1•%Vbur1. The full model is shown below in Eq 3.5, which gives R2 = 0.99. 𝑘 = 𝑎 + 𝑏(𝐿𝐷𝑃)1 + 𝑐(%𝑉𝑏𝑢𝑟)1 + 𝑑(𝐿𝐷𝑃)2 + 𝑒(%𝑉𝑏𝑢𝑟)2 + 𝑓(∆𝑃𝐴) + 𝑔(∆𝑃𝐴)2 + ℎ(𝐵𝑟𝑑𝑔) + 𝑖(𝐿𝐷𝑃)1 • (%𝑉𝑏𝑢𝑟)1 Eq. 3.5 Figure 3.13. Plot for the experimentally observed (kobs) rate constant vs the model calculated rate constant for 3UI-Ti precatalysts using model Eq. 3.5 The plot of residuals for the model shown above shows more evenly distributed data around zero. The plot below includes averaged rate constants with residuals. 132 Figure 3.14. Plot for the experimentally observed (kobs) averaged rate constant vs residuals (from model Eq. 3.5). In our final model shown in Eq. 3.5, the second order term required was the cross term (LDP)1•(%Vbur)1, which are the parameters for the pyrrolyl half of the ligand, suggesting some co- dependence of the electronic and steric term. On the one hand, the existence of the dependence between steric and electronic terms seems obvious as sterics can influence ligand approach and thus donor properties. In this case, the term in the model is perhaps important because the interrelation between the two descriptors is significantly different between the chromium test system used to measure them and the titanium catalytic system. As a result, a scaling factor is required for the interaction. The parameters from a-i are shown below. For the fit, R2 = 0.991, adjusted R2 = 0.987, n = 29, and F = 271 (p-value = 1.1 × 10–18). The p-values for all the individual parameters are 0.000, suggesting the significance of each variable in the model. The values shown 133 for each parameter (ui) in the table are ui = bi ± tcrit • si, where bi = parameter from regression, tcrit = critical t-value for the 95% probability level and 20 degrees of freedom (2.086), and si = standard error. Table 3.5. Parameters from the model shown in Eq. 3.5. Parameter (descriptor) a (intercept) b (LDP)1 c (%Vbur)1 d (LDP)2 e (%Vbur)2 f (∆PA g (∆PA)2 h (Brdg) i (LDP)1•(%Vbur)1 Scaled Parameters (–1 to +1) 9.3 95% Confidence Intervals ± tcrit • si ± 0.9 –20.2 6.13 3.8 –1.7 22.6 -9.5 1.3 7.8 ± 2.6 ± 1.5 ± 1.1 ± 0.5 ± 2.2 ± 2.0 ± 0.4 ± 2.5 From the parameters shown above the largest contributor is the parameter for proton affinity ∆PA, g = 22.6 ± 2.2. This suggests that the 3-position of indolyl ligands is the energetically preferred protonation site and contributes significantly to the rate. The second key variable is LDP1 (b = ‒ 20.2 ± 2.6), a negative sign indicates we require a lower LDP value i.e., electron-rich ligand on the opposite side (side 1) of 3UI. However, if we make side 1 of the ligand electron-rich, the favorability of the protonation at the 3-position of indolyl decreases. Unlike our previous models,1,2 the non-3UI side (side 1) wants to be bigger (c = 6.13 ± 1.5). The sterics of the side 2 134 (indolyl side) should be smaller (d = –1.7 ± 1.5). With this 9-parameter model, correlating all the variables together is relatively hard. Let’s look closely at the fastest catalyst we have: Catalyst 8e has the highest rate constant value (k (average) = 15.1 × 10-4s-1), although the proton affinity difference between side 2 (3UI) and side 1 is not the highest among all the catalysts reported. Further, according to the model side 1 should have an electron-donating ligand, but in the case of 8e the pyrrole is attached to an electron-deficient aryl ring. As mentioned before, if we introduce an electron-donating group on side 1, it could potentially impact ∆PA value. However, due to the complex interplay among these variables, synthesizing a “perfect” catalyst is challenging. Nonetheless, achieving a harmonious balance between these factors can augment the reaction rate effectively (catalyst 8e). The second order model has a rising ridge in proton affinity (Figure 3.15), plotted with rate constant and (LDP)2, f(x,y) = 3.8x + 22.2y – 9.5y2 with x = ∆PA, y = (LDP)2, and z = kcalc (x 10–4 s–1). The top of the ridge occurs at ∆PA = 1.2 in scaled parameters or ~16 kcal/mol in the natural units. The fastest catalyst developed here has ∆PA = 11.7. Consequently, one can surmise that faster catalysts may be accessible, but several competing factors make designing such a catalyst less intuitive. Most notably, a positive ∆PA implies a more electron-rich indolyl and/or electron- deficient other side (pyrrolyl) of the bidentate ligand. However, this is the opposite of what 135 negative (LDP)1 and positive (LDP)2 imply. It may be possible to decouple the proton affinity of the 3-positions and the donor ability towards the metal center, which perhaps defines an avenue for improved catalysis for this reaction. Figure 3.15. Surface plot of f(x,y) = 5.9x + 22.2y – 9.5y2 with x = ∆PA, y = (LDP)2, and z = kcalc (x10–4 s–1) of the model from the scaled parameters. 3.4 Efficiency 3-Unsubstitutedindolyl based Titanium Precatalyst In this study, we have isolated catalyst 8e, which is approximately 4 times faster than Ti(NMe2)2dpm (1a, Figure 2.4). I used catalyst 8e for the hydroamination of other substrates (Table 3.6 and 3.7 shown below) and compared the efficiency of 8e concerning previously reported 1a. The reaction conditions were kept similar to our previous report.11 The first reaction we tried was hydroamination of diphenylacetylene with aniline using 1a and 8e as a catalyst. The summarized reaction conditions are shown in the table below. 136 Table 3.6. Comparison of 1a and 8e for hydroamination of diphenylacetylene and aniline. Substrate 1 Substrate 2 Catalyst Conditionsa %Conversionb %Yield PhCCPh PhNH2 1a 8e 75 C, 24 h 65 C, 2 h - 100 84c 82c a Temperature, and reaction time. b Monitored by NMR. c Isolated yield after LAH reduction of imine into amine. Scheme 3.3. Hydroamination of diphenylacetylene and aniline followed by reduction of imine into amine using LAH. The reaction progress was monitored by GC and NMR spectroscopy. In 1H NMR spectroscopy, a 100% conversion of starting material into a product imine (shown below) was observed (a trace amount of enamine was also spotted in the NMR). Upon completion of the reaction, the reaction mixture was reduced to an amine and purified by column chromatography. As we can see here, 8e not only reduces the reaction time from 24 hours to just 2 hours for this reaction but also operates at slightly lower temperatures. 137 Figure 3.16. 1H NMR of a crude reaction mixture for hydroamination of diphenylacetylene and aniline. For the next synthetically challenging reaction, I tried using cyclohexylamine instead of aniline for hydroamination of diphenylacetylene. With 1a, the reaction required 48 hours at 100 °C to reach completion. However, with 8e, the same reaction can be done in just 9 hours with better yield at a significantly lower temperature of 75 °C. Additionally, in this instance, the major product is enamine instead of imine (by NMR spectroscopy). Consequently, hydrolyzing it rather than reducing it seemed a better choice. The NMR of the isolated purified product in each case is available in the experimental section. 138 Table 3.7. Comparison of 1a and 8e for hydroamination of diphenylacetylene and cyclohexyl amine. Substrate 1 Substrate 2 Catalyst Conditionsa %Conversionb %Yield PhCCPh CyNH2 1a 8e 100C, 48 h 75 C, 9 h - 100 68c 91c a Temperature, and reaction time. b Monitored by NMR. c Isolated yield after hydrolysis of imine into ketone. Scheme 3.4. Hydroamination of diphenylacetylene and cyclohexyl amine followed by hydrolysis of imine into ketone using SiO2. 139 Figure 3.17. 1H NMR of the crude reaction mixture for hydroamination of diphenylacetylene and cyclohexylamine. Catalyst 8e has proven to be quite effective, especially with these more challenging substrates. It decreased the reaction time and temperature by a significant margin. In another study, our group has shown the titanium-catalyzed hydroamination of 1,4-diynes by primary amines followed by in situ 5-exo dig cyclization results in the formation of pyrroles.12 I did the same reaction using 8e to show the efficiency of our catalyst towards the hydroamination of 1,4-diyne. A new benign synthesis of 1,4-diyne opposed to tosyloxyalkyne as starting material is reported here. In this modified synthesis, organocuprate (synthesized in situ) was treated with an alkyl halide to get the desired 1,4-diyne product. The hydroamination of 4-diyn-1-ylbenzene with aniline was carried out and the reaction progress was followed using GC. The final product (pyrrole) was isolated after 140 cyclization and purified by column chromatography. Using 8e as catalyst provided us with about twice the yield with milder reaction conditions than 1a. Scheme 3.5. (top) Modified synthesis of hexa-1,4-diyn-1-ylbenzene. (bottom) Hydroamination of diyne with aniline, followed by ring-closing reaction to generate pyrrole. Table 3.8. Comparison of 1a and 8e for hydroamination of hexa-1,4-diyn-1-ylbenzene and cyclohexyl amine. Substrate 1 Substrate 2 Catalyst Conditionsa %Yieldb PhNH2 1a 8e 110 C, 30 h 75 C, 6 h 35 64 a Temperature, and reaction time for hydroamination reaction. b Isolated yield after 5-exo dig cyclization with CuI. 3.5 Synthesis of Chromium Complexes for LDP and %Vbur Measurements We have been utilizing LDP and %Vbur from our previous study.1 For this study, which involves the use of 3UI ligands, I have synthesized chromium complexes to measure the LDP and %Vbur of 141 unprecedented 3UI ligands. The synthesis of NCr(NiPr2)2X is shown below (Scheme 3.6) where X is the ligand under investigation (adapted from literature),1 where NCr(NiPr2)2OPh is treated with freshly prepared lithium indolyl salt to yield a complex of interest NCr(NiPr2)2X (see experimental for more details). Along with 3UI-based ligands a chromium complex of 2- arylpyrrole (aryl = N,N-dimethylphenyl) was also synthesized to measure LDP of electron- donating aryl groups on pyrrole. According to our current model, the electron-rich nature of pyrrole may contribute to an enhanced reaction rate. To make the pyrrolyl derivative of the chromium complex, NCr(NiPr2)2I was treated with freshly prepared thallium salt of the 2-aryl pyrrolyl ligand. The reaction schemes are shown below. Scheme 3.6. Synthesis of chromium complexes. Let’s see briefly how the LDP value is measured. As we discussed in Chapter 2, The LDP value represents the enthalpic barrier to the rotation of the Cr−N bond of the diisopropylamide ligand in the NCr(NiPr)2X complex. To determine the rate of rotation, first, an appropriate temperature range must be identified to attain a suitable amide rotation rate. The T1’s of methine protons are calculated once the temperature has been carefully monitored and equilibrated. Finally, a 2D 1H NMR experiment is carried out to estimate the rate of rotation of the amide ligand using spin- saturation transfer NMR spectroscopy. As the rate of rotation increases, the protons of the amide 142 ligand syn to the nitride ligand become saturated, resulting in decreased integration of the protons anti to the nitride ligand (Figure 3.18). Figure 3.18. A pictorial representation of NCr(NiPr2)2X system, highlighting the hydrogen in different chemical environments. For each complex, a minimum of three trial runs are required within an error of 0.1 kcal/mol. An averaged LDP value with the temperature at which it was recorded is also mentioned in the table below. %Vbur was calculated using SambVca 2.0 program developed by Cavallo and coworkers.13 143 Table 3.9. LDP and %Vbur value of some new ligands used for this study. Complex Structure Temperature (K) G‡ (kcal/mol) H‡ (LDP) (kcal/mol) %Vbur 10a 10b 10c 10d 10e 10f 10g 240.74 15.18 13.04 22.6 298.03 16.66 13.96 22.4 236.78 15.29 13.15 22.7 262.5 15.36 12.99 22.5 262.62 16.57 14.21 25.4 254.1a 15.7a 13.39a 22.3b 264.7a 14.6a 12.22a 22.7b ameasured in previous studies, bmeasured by me for this study The collected LDP and %Vbur values for 3UI ligands generally align with the anticipated trend, suggesting consistency in their behavior within the scope of our study. The 3UI ligands are slightly less donating than their 3-methylindolyl counterparts. This could be attributed to the lack of methyl 144 groups in 3UI ligands. Intriguingly, 2-(N,N-dimethylphenyl)pyrrolyl has a higher LDP value than the other 2-arylpyrrolyl ligands,1 particularly considering that the attached group typically exhibits electron-donating behavior, which would conventionally suggest a lower LDP value (for reference LDP for 2-(4-MePh)pyrrolyl = 13.91). Analysis of the crystal structure reveals that the 2-aryl group is not planar (the dihedral angle N1‒C1‒C2‒C3 is 53.5°) with the pyrrole ring and is thus unable to get extended conjugation from the 2-aryl ring. However, in the case of 2-(4-MePh)pyrrolyl the dihedral angle N1‒C1‒C2‒C3 is 140.2°, which allows better conjugation with the pyrrole ring and consequently better donation to the chromium center (lower LDP). Figure 3.19. Crystal structure NCr(NiPr2)2(pyrrolyl4-NMe 2 Ph). We anticipated that 2-(N,N-dimethylphenyl)pyrrolyl would function as a strong donor ligand, however, it unexpectedly exhibited poorer donor characteristics than some of the other 2-aryl pyrroles. Plugging its LDP, %Vbur, and ∆PA parameters into the current model (Eq. 6.5) resulted 145 in a slower reaction rate compared to many catalysts synthesized in this study. Consequently, we opted not to proceed with the synthesis of the actual catalyst. This emphasizes the potential utility of our model for screening catalysts based on these parameters, which provides a more efficient way than synthesizing and evaluating each catalyst individually. 3.6 Conclusions In this study, we used 3UI-based unsymmetrical ligands for titanium-catalyzed hydroamination reaction. The use of 3UI ligands allowed us to find a faster catalyst than previously reported symmetrical and unsymmetrical ligand-based titanium precatalysts. The 3-position of indolyl is crucial for enhanced performance of the catalysts. The experiments suggested that aniline can protonate the 3-position of indolyl and all the positions of the pyrrolyl side of the ligand. However, the more basic site is the 3-position of indolyl ligand, and the protonation of pyrrole probably doesn’t affect the rate. Our previous studies with pyrrolyl-based ligands didn’t require the use of proton affinity terms in the model. The best fit in the model is where the PA for the indolyl is energetically favored over pyrrole. Based on experimental evidence, we believe that the proton present at the 3-position is acting as a proton shuttle and favoring the protonation step. The irreversible protonation step is considered to be a rate-determining step in the proposed catalytic cycle. The catalysts with 3-methyl indolyl (6a) and 3UI (8a) followed the same initial rate law but their rates of the hydroamination reaction were different. The inclusion of ∆PA allowed us to model the subtle changes in the mechanism that would otherwise be overlooked. Further, in the new model, the linker between two ligands affects the rate of the reaction. The catalysts with a dimethyl linker (8a) are generally faster than those with methylene (8a’). The nine-parameter model 146 couldn’t be simplified to fewer variables because the effects from each variable were above the error limit. The efficiency of the fastest 3UI-Ti precatalysts (8e) was compared with 1a for other hydroamination reactions. We found that using 8e led to a remarkable reduction in reaction time and temperature for some challenging hydroamination reactions. Further, for pyrrole synthesis 8e not only bolstered reaction efficiency but also doubled the yield achieved in comparison to established literature benchmarks. This study has paved the way for further exploration of catalytic processes, shedding light on the significant impact of ancillary ligands on catalyst performance. 147 3.7 Experimental Details Synthesis of Ligands Synthesis of 2-(1H-indol-2-yl)propan-2-ol A 250 mL Schlenk flask equipped with a stir bar was purged with dinitrogen and charged with ethyl indole-2-carboxylate (1.89 g, 10 mmol, 1.0 equiv) and dry THF (30 mL). The solution was cooled to –78 °C in a dry ice/acetone bath for 15 min. Methyllithium (3.1 M in DME, 16.1 mL, 5.0 equiv) was added dropwise over 15 min. The reaction was stirred at –78 °C for 3 h. Then, water (5 mL) was added dropwise to quench the reaction. The reaction mixture was allowed to warm to room temperature. Ethyl acetate (50 mL) was added to extract the product. The organic layer was washed with brine (50 mL) and separated from the aqueous layer. The organic layer was separated, dried over Na2SO4, and evaporated to afford the product as a light-yellow oil (1.44 g, 82%). 1H NMR (CDCl3, 500 MHz): δ 8.70 (s, 1H), 7.85 (s, 1H), 7.45-7.35 (m, 2H), 6.38 (s, 1H), 1.96 (s, 1H), 1.70 (s, 6H). LRMS (EI): calc’d: 175; found: 175. 1H NMR spectrum was consistent with those previously reported.14 Synthesis of Hind-CH2-Hpyr General procedure A: In a 25 mL Schlenk flask, (1H-indol-2-yl)methanol (500 mg, 1.0 equiv) was dissolved in an excess of pyrrole (1 mL). Slowly, BF3·Et2O (0.2 mL, 0.5 equiv) was added dropwise to the solution under a constant flow of N2. The reaction rapidly turned purple. The 148 mixture was allowed to stir at room temperature for 5 min. The heat was generated after the addition. The reaction was quenched by adding 20 mL of saturated sodium bicarbonate solution. The product was extracted by adding 20 mL of EtOAc. The organic layer was separated and dried over sodium sulfate. The crude product was purified by column chromatography (silica gel, hexanes: EtOAc, 20:1) to afford the desired products. Removal of the solvent afforded the product as a white solid (290 mg, 44%). 1H NMR (500 MHz, CDCl3): δ 7.90 (s, 1H), 7.85 (s, 1H), 7.56 (d, J = 7.7 Hz, 1H), 7.24 (s, 1H), 7.18 – 7.02 (m, 2H), 6.69 (s, 1H), 6.37 (s, 1H), 6.18 (d, J = 3.2 Hz, 1H), 6.12 (s, 1H), 4.15 (s, 2H). 13C{1H} NMR (126 MHz, CDCl3): δ 136.29, 136.16, 128.56, 128.16, 121.63, 120.10, 119.97, 117.68, 110.78, 108.46, 106.84, 100.94, 77.16, 26.91.14 Synthesis of Hind-C(CH3)2-Hpyr2-(3,5-CF3Ph) In an N2 glovebox, 2-(1H-indol-2-yl)propan-2-ol (400 mg, 2.3 mmol), (2-(3,5-bisCF3)phenyl)- 1H-pyrrole (701 mg, 2.2 mmol, 1.1 equiv), InCl3 (101 mg, 20 mol%), and toluene (8 mL) were added to a 35 mL Schlenk tube. The reaction was stirred for 18 h at 65 C. The reaction color changed from green to dark red. The crude product was purified by column chromatography (silica gel, hexanes: EtOAc 20:1) to afford the desired products. Removal of the solvent afforded the product as white solid (722 mg, 73%). 1H NMR (500 MHz, C6D6): δ 7.67 (dd, J = 7.0, 2.1 Hz, 1H), 7.61 (s, 1H), 7.55 (s, 1H), 7.50 (s, 2H), 7.25 – 7.17 (m, 2H), 7.15 (s, 1H), 6.97 (s, 1H), 6.32 – 6.28 (m, 1H), 6.26 (d, J = 2.3 Hz, 1H), 6.12 – 6.07 (m, 1H), 1.45 (s, 6H). 13C{1H} NMR (126 MHz, C6D6): δ 144.65, 141.88, 136.63, 135.22, 132.31 (q, J = 32.9 Hz), 129.04, 128.79, 124.94, 123.68 149 – 123.23 (m), 122.77, 122.34, 120.80, 120.50, 119.06 (p, J = 3.8 Hz), 111.06, 109.39, 107.46, 99.32, 36.02, 28.7. 19F NMR (470 MHz, C6D6): δ -62.85. LRMS (EI): calcd: 436; found: 436. Synthesis of Hind-C(CH3)-Hind3-Me General procedure A was followed using 2-(1H-indol-2-yl)propan-2-ol(279.6 mg, 1.6 mmol), 3-methylindole(230 mg, 1.75 mmol, 1.1equiv), BF3Et2O(113 mg, 0.5 equiv), and toluene(2 mL). The reaction was stirred for 5 minutes at room temperature. After workup removal of solvent afforded the product as tan solid (430 mg, 93%). 1H NMR (500 MHz, C6D6): δ 7.69 (d, J = 9.2 Hz, 1H), 7.54 (d, J = 7.5 Hz, 1H), 7.30 – 7.20 (m, 4H), 7.15 (d, J = 14.8 Hz, 3H), 7.08 – 6.88 (m, 3H), 6.34 (s, 1H), 1.98 (s, 3H), 1.46 (s, 6H). 13C{1H} NMR (126 MHz, C6D6): δ 145.74, 138.40, 136.73, 134.84, 130.90, 128.78, 128.59, 122.04, 120.80, 120.34, 119.84, 118.71, 111.16, 110.81, 107.50, 99.04, 37.03, 28.77, 9.23. LRMS (EI): calcd: 288; found: 288. Synthesis of Hind5-Me-CH2-Hpyr General procedure A was followed using (1H-indol-2-yl)methanol(143 mg, 0.9 mmol), pyrrole(1 mL, excess), BF3Et2O(63 mg, 0.5 equiv). The reaction was stirred for 5 minutes at room temperature. After workup removal of solvent afforded the product as white solid (98 mg, 46%). 1H NMR (500 MHz, CDCl3): δ 7.87 (s, 1H), 7.74 (s, 1H), 7.34 (d, J = 1.7 Hz, 1H), 7.13 (d, J = 8.2 Hz, 1H), 6.96 (dd, J = 8.2, 1.7 Hz, 1H), 6.67 (q, J = 2.2 Hz, 1H), 6.29 (dt, J = 1.9, 1.0 Hz, 1H), 6.18 (q, J = 2.9 Hz, 1H), 6.10 (dtd, J = 3.7, 1.8, 0.9 Hz, 1H), 4.12 (s, 2H), 2.43 (s, 3H). 13C{1H} 150 NMR (126 MHz, CDCl3): δ 136.37, 134.64, 129.20, 128.99, 128.26, 123.23, 119.87, 117.67, 110.40, 108.64, 106.87, 100.57, 77.16, 27.25, 21.58. Synthesis of (5,6-dimethoxy-1H-indol-2-yl)methanol In a 250 mL Schlenk flask, LAH (0.76 g, 5.0 equiv) was added to 40 mL of dry THF under a constant flow of N2. A solution of ethyl 5,6-dimethoxy-1H-indole-2-carboxylate (1.0 g, 1.0 equiv) in dry THF (15 mL) was added slowly. The reaction was stirred for 2 hours and 40 C. The reaction was slowly quenched with 1 M HCl and extracted with ethyl acetate. The organic layer was washed with brine, separated, and dried over sodium sulfate. The evaporation of the solvent afforded the crude product as grey solid (780 mg, 91%). 1H NMR (500 MHz, DMSO): δ 10.66 (s, 1H), 6.97 (s, 1H), 6.85 (s, 1H), 5.09 (t, J = 5.6 Hz, 1H), 4.53 (d, J = 5.5 Hz, 2H), 3.74 (s, 3H), 3.73 (s, 3H). 13C{1H} NMR (126 MHz, DMSO): δ 145.92, 144.19, 138.31, 130.58, 120.53, 102.53, 98.59, 95.09, 56.86, 55.92, 55.70. LRMS (EI): calcd: 207; found: 207. Synthesis of Hind5,6-diOMe-CH2-Hpyr General procedure A was followed using 5,6-dimethoxy-1H-indol-2-yl)methanol (289 mg, 1.0 equiv), pyrrole (1 mL), and BF3·Et2O (0.08 mL, 0.5 equiv). The crude product was purified by column chromatography (silica gel, 40% ethyl acetate in hexanes). Removal of the solvent afforded the product as a white solid (130 mg, 46%). 1H NMR (500 MHz, CDCl3): δ 7.81 (s, 1H), 7.58 (s, 1H), 7.02 (s, 1H), 6.70 (s, 1H), 6.63 (t, J = 3.4 Hz, 1H), 6.24 (d, J = 1.5 Hz, 1H), 6.18 (d, 151 J = 2.9 Hz, 1H), 6.10 (t, J = 3.7 Hz, 1H), 4.05 (s, 2H), 3.90 (s, 3H), 3.86 (s, 3H). 13C{1H} NMR (126 MHz, CDCl3): δ 146.66, 145.14, 134.82, 130.50, 128.57, 121.29, 117.64, 108.54, 106.76, 102.13, 100.65, 94.59, 56.47, 56.30, 27.12. LRMS (EI): calcd: 256; found: 256. Synthesis of hexa-1,4-diyn-1-ylbenzene In a 20 mL scintillation vial, a stir bar was loaded with phenylacetylene (1 g, 9.8 mmol) and THF (5 mL) in an N2 glovebox. To a cold solution (0 °C), ethylmagnesium bromide solution (3.2 mL, 3 M, 9.8 mmol) in diethylether was added dropwise for 20 minutes. If the solution gets hot cool it down and resume addition. The resulting solution was stirred at room temperature for an hour. To this tan color solution cuprous iodide (60 mg, 3 mol%) dissolved in THF (1 mL) was added and stirred for 10 minutes at room temperature. In a separate glass vial, a stir bar was loaded with 1-bromobut-2-yne (0.86 mL, 9.8 mmol) and pentanes (5 mL). This solution was cooled to 0 °C. Phenyl(ethynyl)magnesium bromide solution was added to the cold 1-bromobut-2-yne solution dropwise and stirred for 30 minutes at 5 – 15 °C. The resulting brown solution was brought outside glovebox, and poured over a solution of ammonium chloride (3 g) and sodium cyanide (100 mg) in water (15 mL). To this pentane was added (20 mL) and the solution was stirred vigorously to dissolve solids. The organic layer was sperated and aqueous layer was extracted with pentane thrice (3×10 mL). The combined layer was washed with 4 M ammonium chloride (30 mL) and subsequently, dried with magnesium sulphate. The solvent was evaporated and purified by column chromatography pentane and ethyl actetae (95:5). The product was isolated as pale yellow oil (0.98 g, 65%). 1H NMR (500 MHz, CDCl3) δ 7.49 – 7.38 (m, 2H), 7.35 – 7.26 (m, 3H), 3.36 (q, J = 2.6 Hz, 2H), 1.83 (s, 3H). 152 Synthesis of Titanium Catalysts General procedure B for titanium complexes In a 20 mL scintillation vial, a stir bar was loaded with Ti(NMe2)4 (1.1 equiv.) and ether (5 mL) in an N2 glovebox. A separate 20 mL scintillation vial was loaded with protic-free ligand (1.0 equiv.) and ether (5 mL). Then the solution of ligand was added dropwise to a vigorously stirred solution of Ti(NMe2)4 over 15 min. The reaction mixture was stirred at room temperature for 12 h. The volatiles were removed in vacuo to give a viscous orange/yellow oil. This oil was taken up by adding toluene (2 mL), followed by layering 2 mL of n-hexane. Standing in a –30 °C freezer overnight generally afforded orange/yellow crystals. Synthesis of Ti(NMe2)2(ind-C(CH3)2-pyr) (1a) General Procedure B was followed using H2pyr-C(CH3)2-ind (400 mg, 1.0 equiv), Ti(NMe2)4 (400 mg, 1.0 equiv), and dry diethyl ether (8 mL). Recrystallization from ether/n-hexane afforded the product as orange crystals (440 mg, 69%). 1H NMR (C6D6, 500 MHz): δ 7.77 (d, J = 8.6 Hz, 1H), 7.59 (d, J = 8.9 Hz, 1H), 7.42-7.36 (m, 1H), 7.32 (t, J = 7.9 Hz, 1H), 6.93 (s, 1H), 6.34 (dd, J = 2.7, 1.3 Hz, 1H), 6.32 (s, 1H), 5.89 (s, 1H), 2.93 (s, 12H), 1.83 (s, 6H). 13C{1H} NMR (C6D6, 126 MHz): δ 169.77, 162.08, 143.55, 130.25, 128.50, 120.43, 119.94, 119.30, 115.72, 115.37, 114.63, 97.49, 47.29, 40.06, 29.87. 153 Synthesis of Ti(NMe2)2(ind-CH2-pyr) (1a’) General Procedure B was followed using Hind-CH2-Hpyr (287 mg, 1.0 equiv), Ti(NMe2)4 (344 mg, 1.02 equiv), and dry diethyl ether (10 mL). Recrystallization from ether/n-hexane afforded the product as orange crystals (298 mg, 62%). 1H NMR (500 MHz, C6D6): δ 7.78 (dd, J = 7.6, 1.3 Hz, 1H), 7.61 (dd, J = 8.0, 0.9 Hz, 1H), 7.37 (ddd, J = 8.2, 7.1, 1.4 Hz, 1H), 7.31 (td, J = 7.4, 1.1 Hz, 1H), 6.93 (t, J = 1.4 Hz, 1H), 6.40 (q, J = 1.1 Hz, 1H), 6.09 (dd, J = 2.7, 1.3 Hz, 1H), 5.89 (dd, J = 2.6, 1.5 Hz, 1H), 4.14 (s, 2H), 2.96 (s, 12H). 13C{1H} NMR (126 MHz, C6D6): δ 159.55, 153.45, 144.35, 130.91, 128.49, 120.37, 120.01, 119.09, 116.95, 116.42, 115.46, 99.02, 47.29, 30.55. Synthesis of Ti(NMe2)2(ind-C(CH3)2-pyr2-Ph) (1c) General Procedure D was followed using Hind-C(CH3)2-Hpyr2-Ph (300 mg, 1.0 equiv), Ti(NMe2)4 (224 mg, 1.0 equiv), and dry diethyl ether (6 mL). Recrystallization from ether/n- hexane afforded product as orange crystals (321 mg, 74%). 1H NMR (C6D6, 500 MHz): δ 7.79 (d, J = 9.0 Hz, 1H), 7.70 (d, J = 8.1 Hz, 2H), 7.55 (d, J = 7.2 Hz, 1H), 7.36-7.27 (m, 2H), 7.09 (t, J = 7.6 Hz, 2H), 7.01 (t, J = 7.4 Hz, 1H), 6.48 (s, 2H), 6.37 (s, 1H), 2.75 (s, 12H), 1.88 (s, 6H). 13C{1H} NMR (C6D6, 126 MHz): δ 169.39, 160.80, 142.10, 130.41, 125.94, 120.37, 119.91, 119.34, 115.67, 111.30, 97.46, 46.79, 29.91. 154 Synthesis of Ti(NMe2)2(ind-C(CH3)2-pyr2-(3,5-CF3Ph) (1e) General Procedure B was followed using Hind-C(CH3)2-Hpyr2--(3,5-CF3Ph) (350 mg, 1.0 equiv), Ti(NMe2)4 (180 mg, 1.0 equiv), and dry diethyl ether (5 mL). Recrystallization from ether/n- hexane afforded product as orange crystals (420 mg, 92%). 1H NMR (500 MHz, C6D6): 1H NMR (500 MHz, C6D6) δ 8.11 (s, 2H), 7.76 (s, 1H), 7.64 (s, 1H), 7.38 (t, J = 4.5 Hz, 1H), 7.32 – 7.23 (m, 2H), 6.41 (s, 1H), 6.31 (s, 1H), 6.25 (d, J = 2.7 Hz, 1H), 2.60 (s, 12H), 1.79 (s, 6H). 13C NMR (126 MHz, C6D6) δ 168.51, 161.71, 143.00, 137.98, 136.54, 131.85 (q, J=3.8 Hz), 130.23, 125.58 – 125.52 (m) , 124.92, 122.75, 121.20 – 121.10 (m), 120.68, 120.24, 119.43, 116.14, 115.51, 111.90, 97.83, 46.64, 40.14, 29.83. 19F NMR (470 MHz, C6D6) δ -62.75. Synthesis of Ti(NMe2)2(ind-C(CH3)2-ind3-Me (1f) General Procedure B was followed using Hind-C(CH3)2-Hind3-Me (123 mg, 1.0 equiv), Ti(NMe2)4 (105 mg, 1.1 equiv), and dry diethyl ether (5 mL). Recrystallization from ether/n- hexane afforded product as orange crystals (105 mg, 58%). 1H NMR (500 MHz, C6D6): δ 7.72 (d, J = 5.8 Hz, 1H), 7.59 (d, J = 9.2 Hz, 1H), 7.25 (m, 5H), 7.18 (s, 1H), 6.42 (s, 1H), 2.93 (s, 12H), 2.46 (s, 3H), 1.93 (s, 6H). 13C{1H} NMR (126 MHz, C6D6): δ 158.42, 151.08, 144.30, 142.54, 155 131.57, 129.64, 122.25, 121.96, 120.98, 120.50, 120.17, 118.33, 114.61, 113.72, 107.21, 99.74, 43.65, 41.56, 31.70, 11.18. Synthesis of Ti(NMe2)2(ind5-Me-CH2-pyr) (2a) General Procedure B was followed using Hind5-Me-C(CH3)2-Hpyr (77.5 mg, 1.0 equiv), Ti(NMe2)4 (90.8 mg, 1.1 equiv), and dry diethyl ether (5 mL). Recrystallization from toluene/n- hexane afforded the product as orange crystals (75.5 mg, 60%). 1H NMR (500 MHz, C6D6): δ 7.53 (d, J = 7.6 Hz, 2H), 7.19 (d, J = 8.1 Hz, 1H), 6.94 (t, J = 1.4 Hz, 1H), 6.35 (d, J = 0.9 Hz, 1H), 6.12 – 6.05 (m, 1H), 5.93 – 5.85 (m, 1H), 4.14 (s, 2H), 2.98 (s, 12H), 2.55 (s, 3H). 13C{1H} NMR (126 MHz, C6D6): δ 159.70, 153.39, 142.83, 131.09, 128.54, 128.35, 121.86, 119.06, 116.76, 116.27, 115.17, 98.68, 47.29, 30.62, 21.94. Synthesis of Ti(NMe2)2(ind5,6-diOMe-CH2-pyr) (2d) General Procedure D was followed using Hind5,6-diOMe-CH2-Hpyr (130 mg, 1.0 equiv), Ti(NMe2)4 (125 mg, 1.1 equiv), and dry toluene (2 mL). The crude product was washed with cold toluene three times to give a pure dark brown solid as product (50 mg, 25%). 1H NMR (500 MHz, C6D6): δ 1H NMR (500 MHz, C6D6) δ 7.20 (s, 1H), 7.15 (s, 1H), 6.97 (t, J = 1.4 Hz, 1H), 6.33 (q, J = 1.2 Hz, 1H), 6.12 (dd, J = 2.7, 1.3 Hz, 1H), 5.91 (dd, J = 2.7, 1.5 Hz, 1H), 4.17 (s, 2H), 3.83 156 (s, 3H), 3.68 (s, 3H), 2.97 (s, 12H). 13C{1H} NMR (126 MHz, C6D6): δ 158.47, 153.95, 146.64, 138.85, 128.41, 128.35, 124.10, 116.89, 116.32, 102.82, 101.93, 98.76, 57.18, 56.63, 47.34, 30.64. General Procedure E of New Chromium Complexes for LDP and %Vbur Measurements The synthesis of new chromium complexes was adapted from the previously reported procedure.15 In an N2 glovebox, a scintillation vial was loaded with NCr(NiPr2)2OPh, a stir bar, and hexanes. The solution was frozen in a cold well using liquid N2 and transferred onto a stir plate. Once the solution started to stir, freshly prepared lithium indolide in toluene was added dropwise. The reaction was allowed to stir for 20 h at room temperature unless otherwise stated. The volatiles were removed in vacuo. The residue was extracted with pentane three times and filtered through Celite. The filtrate was concentrated and cooled to –35 C to yield purple crystals. Each sample is doubly recrystallized to yield pure chromium complex for LDP measurements. Synthesis of NCr(NiPr2)2(5-OMeInd) (10a) General procedure E was followed using NCr(NiPr2)2OPh (49.3 mg, 1.0 equiv), hexanes (2 mL), lithium 5-methoxyindolide (21.0 mg, 1 equiv), and toluene (2 mL). Recrystallization from pentanes afforded the product as purple crystals (31 mg, 58%). 1H NMR (500 MHz, CDCl3, −30 C) δ 8.00 (d, J = 8.8 Hz, 1H), 7.39 (d, J = 2.9 Hz, 1H), 7.07 (d, J = 2.6 Hz, 1H), 6.84 (dd, J = 8.9, 2.5 Hz, 1H), 6.53 (d, J = 2.8 Hz, 1H), 5.20 (hept, J = 6.5 Hz, 2H), 3.87 (s, 3H), 3.77 (h, J = 6.4 Hz, 2H), 1.78 (d, J = 6.3 Hz, 6H), 1.63 (d, J = 6.3 Hz, 6H), 1.23 (d, J = 6.4 Hz, 6H), 1.01 (d, J = 157 6.4 Hz, 6H). 13C{1H} NMR (126 MHz, CDCl3, −30 C) δ 153.79, 139.88, 133.98, 128.78, 116.69, 110.44, 102.11, 100.00, 58.18, 55.81, 55.67, 30.74, 30.33, 21.96, 21.49. Synthesis of NCr(NiPr2)2(4,6-diFInd) (10b) General procedure E was followed using NCr(NiPr2)2OPh (60.3 mg, 1 equiv), hexanes (3 mL), lithium 4,6-difluoroindolide (26.7 mg, 1 equiv), and toluene (3 mL). Recrystallization from pentanes afforded the product as purple crystals (35nmg, 50%). 1H NMR (500 MHz, CDCl3) δ 7.56 (s, 1H), 7.29 (s, 1H), 6.60 (d, J = 2.9 Hz, 1H), 6.56 (td, J = 10.0, 2.1 Hz, 1H), 5.22 (hept, J = 6.5 Hz, 2H), 3.82 (hept, J = 6.3 Hz, 2H), 1.81 (d, J = 6.3 Hz, 7H), 1.66 (d, J = 6.3 Hz, 7H), 1.25 (d, J = 6.4 Hz, 7H), 1.05 (d, J = 6.4 Hz, 7H). 13C NMR (126 MHz, CDCl3) δ 158.83 (dd, J = 234.6, 11.8 Hz), 155.08 (dd, J = 246.8, 15.4 Hz), 147.14 – 146.93(m), 133.41 (d, J = 3.6 Hz), 114.69 (d, J = 19.5 Hz), 98.48, 98.23 (dd, J = 25.4, 4.2 Hz), 94.32 (dd, J = 29.0, 23.7 Hz), 58.51, 56.45, 30.84, 30.35, 22.09, 21.69.19F NMR (470 MHz, CDCl3) δ -121.67 (td, J = 10.0, 3.0 Hz), -122.03 (dd, J = 10.4, 3.5 Hz). Synthesis of NCr(NiPr2)2(4,6-diFInd) (10c) General procedure E was followed using NCr(NiPr2)2OPh (39.2 mg, 1 equiv), hexanes (2 mL), lithium 5,6-dimethoxyindolide (20 mg, 1 equiv), and toluene (2 mL). Recrystallization from pentanes afforded the product as purple crystals (21 mg, 22.4%). 1H NMR (500 MHz, CDCl3, −35 158 C) δ 7.65 (s, 1H), 7.22 (d, J = 2.7 Hz, 1H), 7.01 (s, 1H), 6.47 (d, J = 2.9 Hz, 1H), 5.15 (hept, J = 6.5 Hz, 2H), 3.93 (s, 3H), 3.91 (s, 3H), 3.77 (hept, J = 6.4 Hz, 2H), 1.79 (d, J = 6.3 Hz, 6H), 1.60 (d, J = 6.3 Hz, 6H), 1.19 (d, J = 6.4 Hz, 6H), 0.97 (d, J = 6.4 Hz, 6H).13C{1H} NMR (126 MHz, CDCl3, −35 C) δ 145.37, 144.21, 138.95, 131.37, 120.74, 102.04, 99.47, 98.84, 58.28, 55.80, 30.83, 30.42, 21.95, 21.49. Synthesis of NCr(NiPr2)2(5-MeInd) (10d) General procedure E was followed using NCr(NiPr2)2OPh (52.4 mg, 1 equiv), hexanes (3 mL), lithium 5-methylindolide (20 mg, 1 equiv), and toluene (3 mL). Recrystallization from pentanes afforded the product as purple crystals (30 mg, 52%). 1H NMR (500 MHz, CDCl3, −10 C) δ 7.98 (d, J = 8.3 Hz, 1H), 7.36 (d, J = 2.9 Hz, 1H), 7.35 (s, 1H), 6.99 (dd, J = 8.4, 1.8 Hz, 1H), 6.47 (d, J = 2.8 Hz, 1H), 5.19 (hept, J = 6.4 Hz, 2H), 3.75 (hept, J = 6.4 Hz, 2H), 2.43 (s, 3H), 1.74 (d, J = 6.3 Hz, 6H), 1.62 (d, J = 6.3 Hz, 6H), 1.21 (d, J = 6.5 Hz, 7H), 0.99 (d, J = 6.4 Hz, 6H). 13C{1H} NMR (126 MHz, CDCl3, −10 C) δ 143.09, 133.60, 129.12, 128.61, 122.41, 118.65, 115.59, 101.92, 58.15, 55.82, 30.77, 30.28, 21.98, 21.61, 21.58. Synthesis of NCr(NiPr2)2(py4-NMe2Ph) (10e) In a 20 mL scintillation vial NCr(NiPr2)2I (6.2 mg, 1 equiv), toluene (3 mL), and a stir bar were loaded. To this reaction mixture, a freshly prepared thallium (4-NMe2Ph)pyrrole (6.2 mg, 1.1 159 equiv) in ether (3 mL) was added. The reaction was allowed to stir for 20 h at room temperature. The yellow-colored precipitate was removed by filtering through a celite plug. The volatiles were removed in vacuo and the residue was extracted in pentane. The concentrated pentane solution afforded bright orange crystals at ‒35 °C (3.6 mg, 50%). The starting material was found in NMR. Fortunately, the septet for methine peaks from the starting material and product has different chemical shifts. Due to the distinct separation between methine peaks from both chromium complexes, I performed an SST experiment with an impure sample (starting material left). 1H NMR (500 MHz, CDCl3) δ 7.49 (d, J = 8.7 Hz, 2H), 6.85 (s, 1H), 6.71 (d, J = 8.9 Hz, 2H), 6.30 (s, 1H), 6.16 (s, 1H), 5.09 (sept, J = 6.3 Hz, 2H), 3.68 (sept, J = 6.3 Hz, 3H), 2.91 (s, 6H), 1.54 (s, 6H), 1.47 (d, J = 6.3 Hz, 6H), 1.11 (d, J = 6.5 Hz, 6H), 1.05 (d, J = 6.4 Hz, 6H). 160 General Procedure for Kinetics The procedure from the previous study was followed.1 All manipulations were conducted in an N2 glovebox. All measurements are done using volumetric syringes for accurate volumes. A 2 mL volumetric flask was loaded with the titanium precatalyst (10 mol%, 0.1 mmol) and ferrocene (56 mg, 0.3 mmol). Toluene-d8 (~0.75 mL) was added to the volumetric flask to completely dissolve the solids. Next, aniline (911 μL, 10 mmol) and 1-phenylpropyne (125 μL, 1.0 mmol) were added to the volumetric flask. Lastly, the solution was diluted up to 2 mL with toluene-d8. The solution was mixed via a pipette, i.e., the solution was drawn up into the pipette and dispensed back into the volumetric flask, five times to ensure the homogeneity of the solution. Then, 0.75 mL solution was loaded into a threaded J. Young NMR tube and removed from the dry box. This solution was heated at 75 °C in the NMR spectrometer (Varian Inova 600). The relative 1-phenylpropyne versus ferrocene concentration was monitored as a function of time. The fits are exponential decay of the starting material. The expression used to fit the data was 𝑌𝑡 = 𝑌∞ + (𝑌0 − 𝑌∞)𝑒−𝑘𝑜𝑏𝑠𝑡,2 where Yt = concentration at time t, infinity (Y∞), or at the start of the reaction (Y0). Each kinetic experiment was done in triplicates and the average value was used to represent the rate constant for the catalyst under investigation. 161 Representative Plots from Kinetics Figure 3.20. Plot of [1-phenylpropyne] vs time with Ti(NMe2)2(ind-C(CH3)2-pyr) (8a). Figure 3.21. Plot of [1-phenylpropyne] vs time with Ti(NMe2)2(ind-CH2-pyr) (8a’). 162 Figure 3.22. Plot of [1-phenylpropyne] vs time with Ti(NMe2)2(ind-C(CH3)2-pyr2-Ph) (8c). Figure 3.23. Plot of [1-phenylpropyne] vs time with Ti(NMe2)2(ind-C(CH3)2-pyr2-(3,5-CF3Ph) (8e). 163 Figure 3.24. Plot of [1-phenylpropyne] vs time with Ti(NMe2)2(ind-C(CH3)2-ind3-Me (8f). Figure 3.25. Plot of [1-phenylpropyne] vs time with Ti(NMe2)2(ind5-Me-CH2-pyr) (9a). 164 Figure 3.26. Plot of [1-phenylpropyne] vs time with Ti(NMe2)2(ind5,6-diOMe-CH2-pyr) (8d). 165 Initial Rate Measurements To understand the role of the methyl group present at the 3-position of the indolyl side of the ligand, we studied the rate law for catalysts 1a’ and 3a. Similar to kinetic experiments all measurements are done using volumetric syringes for accurate volumes. The reaction conditions are mentioned below in the table. A 2 mL volumetric flask was loaded with titanium precatalyst and ferrocene. Toluene-d8 (~0.75 mL) was added to the volumetric flask to completely dissolve the solids. Next, aniline and 1-phenylpropyne were added to the volumetric flask. Lastly, the solution was diluted up to 2 mL with toluene-d8. The solution was mixed via a pipette, i.e., the solution was drawn up into the pipette and dispensed back into the volumetric flask, five times to ensure the homogeneity of the solution. Then, 0.75 mL solution was loaded into a threaded J. Young NMR tube and removed from the dry box. This solution was heated at 75 °C in the NMR spectrometer (Varian Inova 600). The relative 1-phenylpropyne versus ferrocene concentration was monitored as a function of time for the first 15 minutes of the reaction time. The fits are linear decay of the starting material. The expression used to fit the data was 𝑌𝑡 = 𝑌0 + 𝑘𝑡 where Yt = concentration at time t, Y0 = concentration at the start of the reaction, and k = initial rate constant. Rate Law for 1a’ and 3a The rate law for both compounds by initial rates studies shown in the previous section is Eq 3 from the manuscript. Naturally, (unless there is decomposition) the catalyst concentration is a constant, so first-order fits in alkyne are observed. − 𝑑[𝑎𝑙𝑘𝑦𝑛𝑒] 𝑑𝑡 = 𝑘𝑜𝑏𝑠[𝑎𝑙𝑘𝑦𝑛𝑒][𝑐𝑎𝑡𝑎𝑙𝑦𝑠𝑡] (3) 166 The conditions we are using has an excess of amine, which we found gave more reliable results across different catalysts, perhaps by increasing solubility of some catalysts and removing initiation times that were observed in some cases. Because an excess of amine (aniline) is present, it is thought that the resting state for the catalyst is likely XX’Ti(NHPh)2. A mechanistic scheme is shown below, with this likely dominate species as A. This would be in equilibrium (K1) with the titanium imide B through proton transfer and loss of amine. The imide could undergo cycloaddition with the alkyne in another equilibrium process, K2, to form metallacycle C. The metallacycle can pick up amine, K3, to give D, which could undergo irreversible (k4) Ti–C protonation. Loss of product (an amide) through protonation is likely rapid. This scheme ignores dimerization of B, which is observed at low amine concentrations with some catalysts.16 Because there is a large excess of amine, we will assume that the dominate species in solution will be A. In other words, [catalyst] = [A] + [B] + [C] + [D]. In the limit that [A] >> [B] + [C] + [D], [catalyst] ~ [A]. The rate of the reaction is then: 𝑅𝑎𝑡𝑒 = 𝑘4[𝑫] 167 Looking at the equilibria to find [D], starting with K1: rearranging Looking at K2: rearranging Substituting [B] Looking at K3: rearranging 𝐾1 = [𝑩][𝑁𝐻2𝑅] [𝑨] [𝑩] = 𝐾1[𝑨] [𝑁𝐻2𝑅] 𝐾2 = [𝑪] [𝑎𝑙𝑘𝑦𝑛𝑒][𝑩] [𝑪] = 𝐾2[𝑎𝑙𝑘𝑦𝑛𝑒][𝑩] [𝑪] = 𝐾1𝐾2[𝑎𝑙𝑘𝑦𝑛𝑒][𝑨] [𝑁𝐻2𝑅] 𝐾3 = [𝑫] [𝑪][𝑁𝐻2𝑅] [𝑫] = 𝐾3[𝑪][𝑁𝐻2𝑅] Substituting [C] and [A] ~ [catalyst] 168 [𝑫] = 𝐾1𝐾2𝐾3[𝑎𝑙𝑘𝑦𝑛𝑒][𝑐𝑎𝑡𝑎𝑙𝑦𝑠𝑡][𝑁𝐻2𝑅] [𝑁𝐻2𝑅] = 𝐾1𝐾2𝐾3[𝑎𝑙𝑘𝑦𝑛𝑒][𝑐𝑎𝑡𝑎𝑙𝑦𝑠𝑡] Substituting into gives 𝑅𝑎𝑡𝑒 = 𝑘4[𝑫] 𝑅𝑎𝑡𝑒 = 𝑘4𝐾1𝐾2𝐾3[𝑎𝑙𝑘𝑦𝑛𝑒][𝑐𝑎𝑡𝑎𝑙𝑦𝑠𝑡] This is consistent with the observed rate law from initial rates with kobs = k4K1K2K3 and explains the adequacy of first-order fits for the system, assuming [catalyst] = constant (excepting any catalyst decomposition in the sample). New LDP and %Vbur Values Used in This Study The measurement of LDP values was done by using Spin Saturation Transfer (SST) NMR experiments. The temperature chosen for each experiment was based on that required to reach the slow exchange limit of each chromium complex under investigation. T1 values were measured using the inversion recovery method. The concentration of each sample was between 0.02−0.03 M in CDCl3. S‡ was kept constant −9 cal•mol-1•K-1 based on our prior study.15 For calculating %Vbur, we used SambVca 2.0 program developed by Cavallo and coworkers.13 Details are available in our previous publication.1 169 Data Used in the Modeling The table below contains the data used in the modeling, which is from spin saturation transfer experiments (1H NMR) on relevant NCr(NiPr2)2X complexes, where X is the fragment of the bidentate ligand. In all cases, the LDP value was found for the fragment with the position of the bridge replaced by hydrogen in X. For example, as shown below, the electronic (LDP) and steric (%Vbur) descriptors for 1e were found experimentally, 1H NMR spectroscopy and single crystal X- ray diffraction, using the two chromium complexes shown on the right. The data for the pyrrolyl and 3-methyl indolyl fragments was found in the literature, LDP1 and %Vbur1.15 The 3-unsubstituted indolyls were previously unreported, LDP(unsub) and %Vbur(unsub). 170 Table 3.10. Raw data for 3-unsubstituted indolyl based titanium precatalysts. Catalyst 1a 1a 1a 1b 1b 1b 1c 1c 1c 1c 1d 1d 1e 1e 1e 1f 1f 2a 2a 2a 2b 2b 2c 2d 2d 2d 1a' 1a' 1a' LDP1 13.64 13.64 13.64 13.46 13.46 13.46 14.03 14.03 14.03 14.03 13.91 13.91 14.32 14.32 14.32 12.49 12.49 13.64 13.64 13.64 13.64 13.64 13.64 13.64 13.64 13.64 13.64 13.64 13.64 %Vbur1 20.4 20.4 20.4 23.7 23.7 23.7 27.1 27.1 27.1 27.1 26.7 26.7 27.9 27.9 27.9 22.6 22.6 20.4 20.4 20.4 20.4 20.4 20.4 20.4 20.4 20.4 20.4 20.4 20.4 LDP(unsub) 13.39 13.39 13.39 13.39 13.39 13.39 13.39 13.39 13.39 13.39 13.39 13.39 13.39 13.39 13.39 13.39 13.39 12.99 12.99 12.99 13.96 13.96 13.08 13.14 13.14 13.14 13.39 13.39 13.39 %Vbur(unsub) 22.3 22.3 22.3 22.3 22.3 22.3 22.3 22.3 22.3 22.3 22.3 22.3 22.3 22.3 22.3 22.3 22.3 22.5 22.5 22.5 22.4 22.4 22.6 22.7 22.7 22.7 22.3 22.3 22.3 kobs x 104 s–1 6.4 7 6.2 10.1 12.3 12.2 9.8 8.8 9.3 9.5 6.2 7.6 15.2 14.9 15.3 0.5 0.6 5.2 5.0 5.1 2.7 2.7 3.9 0.2 0.2 0.2 5.6 5.4 5.4 Proton affinity difference calculations were done on Gaussian16 on MSU’s High Performance Computing Center (HPCC). 171 Effect of Each Term in the Model Model Using LDP and %Vbur Alone We attempted to model the rate constants for the 3-unsubstituted indolyl-containing ligands, 1 and 2, using the same natural model (unweighted) found in our previous report on unsymmetrical ligand.2 This resulted in the plot shown in Fig. 3.2 in the article, where the blue diamonds are for catalysts 8 and 9. This 5-parameter model only includes an LDP and %Vbur descriptor for the ligand at each side of the chelate, a model that works relatively well for many ligands without 3- unsubstituted indolyl groups. The model is reproduced below. 𝑘 = 𝑎 + 𝑏(𝐿𝐷𝑃)1 + 𝑐(%𝑉𝑏𝑢𝑟)1 + 𝑑(𝐿𝐷𝑃)2 + 𝑒(%𝑉𝑏𝑢𝑟)2 If instead, we do regression just on this set of data with the LDP and %Vbur descriptors, a very poor fit is still obtained (R2 = 0.77), and the summary output copied and pasted from the regression analysis is below. Keep in mind that the standard error is multiplied by the t-value 2.086 to get the 95% confidence interval, i.e., the only parameters large enough to be considered nonzero are LDP1 and %Vbur2. 172 Table 3.11. Regression Statistics for Table 3.10. Regression Statistics Multiple R R Square 0.880758 0.775735 Adjusted R Square 0.738357 Standard Error 2.310382 Observations 29 ANOVA df SS MS F Significance F Regression 4 443.1292 110.7823 20.75404832 1.66859E-07 Residual 24 128.1087 5.337864 Total 28 571.2379 Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Intercept 397.9962 116.9066 3.404394 0.002332264 156.7128123 639.279614 X Variable 1 5.185999 1.287392 4.0283 0.000490532 2.528953629 7.84304495 X Variable 2 0.334585 0.201179 1.663119 0.109294637 -0.080628427 0.749797743 X Variable 3 -4.43031 2.317821 -1.91141 0.06796465 -9.214061099 0.353432091 X Variable 4 -18.3583 4.438725 -4.13595 0.000373506 -27.51942549 -9.197271028 In essence, if the same model as previously published gave a good fit with different coefficients, it would imply that a new mechanism was operative when 3-unsubstituted indolyls were present, but the same descriptors were not adequate. However, that doesn’t appear to be the case. 173 Model Using LDP, %Vbur, and Proton Affinity Terms Table 3.12. Inclusion of proton affinity in the data table for respective catalysts. ∆PA 0.43 0.43 0.43 0.08 0.08 0.08 0.11 0.11 0.11 0.11 -0.04 -0.04 0.55 0.55 0.55 -1 -1 0.74 0.74 0.74 0.16 0.16 1 0.58 0.58 0.58 0.57 0.57 0.57 ∆PA2 0.19 0.19 0.19 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0 0 0.3 0.3 0.3 1 1 0.54 0.54 0.54 0.02 0.02 1 0.34 0.34 0.34 0.33 0.33 0.33 Complex LDP1 %Vbur1 LDP(unsub) %Vbur(unsub) 8a 8a 8a 8b 8b 8b 8c 8c 8c 8c 8d 8d 8e 8e 8e 8f 8f 9a 9a 9a 9b 9b 9c 9d 9d 9d 8a' 8a' 8a' 0.26 0.26 0.26 0.06 0.06 0.06 0.68 0.68 0.68 0.68 0.55 0.55 1.00 1.00 1.00 -1.00 -1.00 0.26 0.26 0.26 0.26 0.26 0.26 0.26 0.26 0.26 0.26 0.26 0.26 -1.00 -1.00 -1.00 -0.12 -0.12 -0.12 0.79 0.79 0.79 0.79 0.68 0.68 1.00 1.00 1.00 -0.41 -0.41 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 0.00 0.00 0.00 -0.50 -0.50 0.50 1.00 1.00 1.00 -1.00 -1.00 -1.00 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -1.00 -1.00 -1.00 1.00 1.00 -0.81 -0.69 -0.69 -0.69 -0.18 -0.18 -0.18 174 Table 3.12 (cont’d) Regression Statistics Multiple R R Square 0.966259 0.933656 Adjusted R Square 0.915562 Standard Error 1.312499 Observations 29 ANOVA df SS MS F Significance F Regression 6 533.3396 88.88993 51.6005917 7.54133E-12 Residual 22 37.89837 1.722653 Total 28 571.2379 Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Intercept 455.4569 67.52275 6.745236 8.8613E-07 315.4232639 595.4905 X Variable 1 -18.1155 3.319885 -5.45668 1.7567E-05 -25.00056451 -11.2305 X Variable 2 2.548007 0.326719 7.798782 8.9869E-08 1.870434466 3.22558 X Variable 3 5.260455 2.018775 2.605766 0.016138 1.073772483 9.447137 X Variable 4 -14.9251 2.579697 -5.78562 8.0534E-06 -20.27510194 -9.57518 X Variable 5 19.16572 2.652341 7.225966 3.0625E-07 13.66510378 24.66634 X Variable 6 -4.97419 1.355036 -3.67089 0.00134123 -7.784362242 -2.16402 Inclusion of the proton affinity difference (PA) led to a significant improvement in the linear model (R2 = 0.93). Model Using LDP, %Vbur, Proton Affinity, and Term for Bridge 175 We added a term for dimethylation of the bridge because on checking the effect of these groups, it was found to be significant, c.f., 8a and 8a’. This was simply done as a noncontinuous variable since the only ligands in the set either had dimethylmethene (assigned as “1”) or just methylene (assigned as “0”). Again, this led to a significant improvement in the model (R2 = 0.97). Table 3.13. Addition of bridging term for dimethyl linker in the model. Complex LDP1 %Vbur1 LDP(unsub) %Vbur(unsub) ∆PA 0.43 0.43 0.43 0.08 0.08 0.08 0.11 0.11 0.11 0.11 -0.04 -0.04 0.55 0.55 0.55 -1.00 -1.00 0.74 0.74 0.74 0.16 0.16 1.00 0.58 0.58 0.58 0.57 0.57 0.57 -1.00 -1.00 -1.00 -0.12 -0.12 -0.12 0.79 0.79 0.79 0.79 0.68 0.68 1.00 1.00 1.00 -0.41 -0.41 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 0.00 0.00 0.00 -0.50 -0.50 0.50 1.00 1.00 1.00 -1.00 -1.00 -1.00 0.26 0.26 0.26 0.06 0.06 0.06 0.68 0.68 0.68 0.68 0.55 0.55 1.00 1.00 1.00 -1.00 -1.00 0.26 0.26 0.26 0.26 0.26 0.26 0.26 0.26 0.26 0.26 0.26 0.26 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -1.00 -1.00 -1.00 1.00 1.00 -0.81 -0.69 -0.69 -0.69 -0.18 -0.18 -0.18 8a 8a 8a 8b 8b 8b 8c 8c 8c 8c 8d 8d 8e 8e 8e 8f 8f 9a 9a 9a 9b 9b 9c 9d 9d 9d 8a' 8a' 8a' ∆PA2 Bridg 1.00 0.19 1.00 0.19 1.00 0.19 1.00 0.01 1.00 0.01 1.00 0.01 1.00 0.01 1.00 0.01 1.00 0.01 1.00 0.01 1.00 0 1.00 0 1.00 0.3 1.00 0.3 1.00 0.3 1.00 1 1.00 1 1.00 0.54 1.00 0.54 1.00 0.54 -1.00 0.02 -1.00 0.02 -1.00 1 -1.00 0.34 -1.00 0.34 -1.00 0.34 -1.00 0.33 -1.00 0.33 -1.00 0.33 176 Table 3.13. (cont’d) Regression Statistics Multiple R R Square 0.985752 0.971708 Adjusted R Square 0.962277 Standard Error 0.877269 Observations 29 ANOVA df SS MS F Significance F Regression 7 555.0763 79.29662 103.036 8.15E-15 Residual 21 16.16162 0.769601 Total 28 571.2379 Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Intercept 250.4801 59.36727 4.219162 0.000385 127.0191 373.9411 X Variable 1 -19.3975 2.232072 -8.69038 2.13E-08 -24.0394 -14.7557 X Variable 2 2.550812 0.218378 11.68071 1.2E-10 2.09667 3.004954 X Variable 3 10.61646 1.684159 6.303714 2.99E-06 7.114055 14.11886 X Variable 4 -8.24516 2.133762 -3.86414 0.000899 -12.6826 -3.80776 X Variable 5 21.27869 1.81685 11.71186 1.14E-10 17.50034 25.05703 X Variable 6 -4.27025 0.915335 -4.66523 0.000133 -6.1738 -2.36671 X Variable 7 1.666605 0.313594 5.314524 2.86E-05 1.01445 2.318761 177 Model Using LDP, %Vbur, Proton Affinity, Term for Bridge, and Crossterm For final modeling, the data were scaled from –1 to +1, which gives much better results, especially for nonlinear models like the one here. The scaling was done using the equation belowwhere xi = scaled variable, ui = natural variable, ui 0 = midpoint of the range of the natural variables, and Dui = the difference between the midpoint and the high value (half the full range). The equations for the calculation of ui 0 and Dui are shown below. Once the modeling is complete with the scaled values, the natural values can be obtained using the rearranged versions of the same equations. So far I have been using natural values for modeling however for direct comparisons of the coefficients we need to scale values from -1 to 1. 178 xi=ui-ui0DuiDui=uihigh-ui0ui0=uihigh+uilow2 Table 3.14. Data for final modeling with all the terms and rate constant values. Complex LDP1 %Vbur1 LDP(unsub) %Vbur(unsub) ∆PA ∆PA2 Bridg LDP1*Vbur kobs x 104 s–1 8a 8a 8a 8b 8b 8b 8c 8c 8c 8c 8d 8d 8e 8e 8e 8f 8f 9a 9a 9a 9b 9b 9c 9d 9d 9d 0.26 0.26 0.26 0.06 0.06 0.06 0.68 0.68 0.68 0.68 0.55 0.55 1.00 1.00 1.00 -1.00 -1.00 0.26 0.26 0.26 0.26 0.26 0.26 0.26 0.26 0.26 -1.00 -1.00 -1.00 -0.12 -0.12 -0.12 0.79 0.79 0.79 0.79 0.68 0.68 1.00 1.00 1.00 -0.41 -0.41 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -0.18 -1.00 -1.00 -1.00 1.00 1.00 -0.81 -0.69 -0.69 -0.69 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 0.00 0.00 0.00 -0.50 -0.50 0.50 1.00 1.00 1.00 179 0.43 0.19 0.43 0.19 0.43 0.19 0.08 0.01 0.08 0.01 0.08 0.01 0.11 0.01 0.11 0.01 0.11 0.01 0.11 0.01 -0.04 -0.04 0.55 0.55 0.55 -1.00 -1.00 0.74 0.54 0.74 0.54 0.74 0.54 0.16 0.02 0.16 0.02 1.00 0.58 0.34 0.58 0.34 0.58 0.34 0 0 0.3 0.3 0.3 1 1 1 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -1.00 -0.26 -0.26 -0.26 -0.01 -0.01 -0.01 0.54 0.54 0.54 0.54 0.38 0.38 1.00 1.00 1.00 0.41 0.41 -0.26 -0.26 -0.26 -0.26 -0.26 -0.26 -0.26 -0.26 -0.26 6.4 7 6.2 10.1 12.3 12.2 9.8 8.8 9.3 9.5 6.2 7.6 15.2 14.9 15.3 0.5 0.6 5.2 5 5.1 2.7 2.7 3.9 0.2 0.2 0.2 Table 3.14 (cont’d) 8a' 8a' 8a' 0.26 0.26 0.26 -1.00 -1.00 -1.00 -0.18 -0.18 -0.18 -1.00 -1.00 -1.00 0.57 0.33 0.57 0.33 0.57 0.33 -1.00 -1.00 -1.00 -0.26 -0.26 -0.26 5.6 5.4 5.4 180 Table 3.14 (cont’d) Regression Statistics Multiple R R Square 0.995431 0.990884 Adjusted R Square 0.987237 Standard Error 0.51027 Observations 29 ANOVA df SS MS F Significance F Regression 8 566.0304 70.7538 271.7378 1.11E-18 Residual 20 5.207505 0.260375 Total 28 571.2379 Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Intercept X Variable 1 X Variable 2 X Variable 3 X Variable 4 X Variable 5 X Variable 6 X Variable 7 X Variable 8 9.351 -20.251 6.128 3.814 -1.754 22.676 -9.535 1.380 7.810 0.454 1.249 0.713 0.518 0.249 1.079 0.971 0.188 1.204 20.596 -16.214 8.599 7.366 -7.051 21.025 -9.823 7.355 6.486 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 8.404 10.298 -22.857 -17.646 4.641 2.734 7.614 4.894 -2.273 -1.235 20.426 24.926 -11.560 -7.510 0.989 5.298 1.772 10.321 181 Other Substrates for Titanium Hydroamination Reaction In this study, we have isolated catalyst 8e, which is approximately 4 times faster than T(NMe2)2dpm (1a). We used catalyst 8e for hydroamination of other substrates (table shown below) and compared the efficiency of 8e with respect to previously reported 1a. The reaction conditions were kept similar as our previous report.11 Briefly, a 3 mL scintillation vial was loaded with catalyst 1e (17.8 mg; 5 mol% or 35.6 mg; 10 mol%) and dissolved in chlorobenzene (62 μL). To this mixture, aniline (62.7 μL, 0.687 mmol) or cyclohexylamine (79 μL, 0.687 mmol) and diphenylacetylene (110 mg, 0.625 mmol) were added. The vial was sealed and transferred to a hot plate at the required temperature. The reaction progress was monitored by NMR spectroscopy. After reaction completion, the imine was either reduced to an amine by LAH or hydrolyzed to the ketone using silica gel to get an isolated yield for the reaction.17 Catalyst 8e has proven to be quite effective, especially with these more challenging substrates. Decreasing reaction time and temperature by a significant margin. In another example, titanium- catalyzed hydroamination of 1,4-diynes by primary amines followed by in situ 5-exo dig cyclization results in the formation of pyrroles.12 Reaction progress was followed using GC and the final pyrrole was isolated and purified by column chromatography. Using 8e as catalyst provided us about twice the yield with milder reaction conditions than 1a. 182 NMR Spectra of Ligands and Catalysts Figure 3.27. 1H NMR of Hind-C(CH3)2-Hpyr. 183 Figure 3.28. 13C NMR of Hind-C(CH3)2-Hpyr. 184 Figure 3.29. 1H NMR of Hind-CH2-Hpyr. 185 Figure 3.30. 13C NMR of Hind-CH2-Hpyr. 186 Figure 3.31. 1H NMR of Hind-C(CH3)2-Hpyr2-Ph. 187 Figure 3.32. 13C NMR of Hind-C(CH3)2-Hpyr2-Ph. 188 Figure 3.33. 1H NMR of Hind-C(CH3)2-Hpyr2-(3,5-CF3Ph). 189 Figure 3.34. 13C NMR of Hind-C(CH3)2-Hpyr2-(3,5-CF3Ph). 190 Figure 3.35. 19F NMR of Hind-C(CH3)2-Hpyr2-(3,5-CF3Ph). 191 Figure 3.36. 1H NMR of Hind-C(CH3)-Hind3-Me. 192 Figure 3.37. 13C NMR of Hind-C(CH3)-Hind3-Me. 193 Figure 3.38. 1H NMR of Hind5-Me-CH2-Hpyr. 194 Figure 3.39. 13C NMR of Hind5-Me-CH2-Hpyr. 195 Figure 3.40. 1H NMR of (5,6-dimethoxy-1H-indol-2-yl)Methanol. 196 Figure 3.41. 13C NMR of (5,6-dimethoxy-1H-indol-2-yl)Methanol. 197 Figure 3.42. 1H NMR of Hind5,6-diOMe-CH2-Hpyr. 198 Figure 3.43. 13C NMR of Hind5,6-diOMe-CH2-Hpyr. 199 Figure 3.44. 1H NMR of Ti(NMe2)2(ind-C(CH3)2-pyr) (8a). 200 Figure 3.45. 1H NMR of Ti(NMe2)2(ind-C(CH3)2-pyr) (8a). 201 Figure 3.46. 13C NMR of Ti(NMe2)2(ind-C(CH3)2-pyr) (8a). 202 Figure 3.47. 1H NMR of Ti(NMe2)2(ind-CH2-pyr) (8a’). 203 Figure 3.48. 13C NMR of Ti(NMe2)2(ind-CH2-pyr) (8a’). 204 Figure 3.49. 1H NMR of Ti(NMe2)2(ind-C(CH3)2-pyr2-Ph) (8c). 205 Figure 3.50. 13C NMR of Ti(NMe2)2(ind-C(CH3)2-pyr2-Ph) (8c). 206 Figure 3.51. 1H NMR of Ti(NMe2)2(ind-C(CH3)2-pyr2-(3,5-CF3Ph) (8e). 207 Figure 3.52. 13C NMR of Ti(NMe2)2(ind-C(CH3)2-pyr2-(3,5-CF3Ph) (8e). 208 Figure 3.53. 19F NMR of Ti(NMe2)2(ind-C(CH3)2-pyr2-(3,5-CF3Ph) (8e). 209 Figure 3.54. 1H NMR of Ti(NMe2)2(ind-C(CH3)2-ind3-Me (8f). 210 Figure 3.55. 13C NMR of Ti(NMe2)2(ind-C(CH3)2-ind3-Me (8f). 211 Figure 3.56. 1H NMR of Ti(NMe2)2(ind5-Me-CH2-pyr) (9a). 212 Figure 3.57. 13C NMR of Ti(NMe2)2(ind5-Me-CH2-pyr) (9a). 213 Figure 3.58. 1H NMR of Ti(NMe2)2(ind5,6-diOMe-CH2-pyr) (9d). 214 Figure 3.59. 13C NMR of Ti(NMe2)2(ind5,6-diOMe-CH2-pyr) (9d). 215 Figure 3.60. 1H NMR of NCr(NiPr2)2(5-OMeInd) (10a). 216 Figure 3.61. 13C NMR of NCr(NiPr2)2(5-OMeInd) (10a). 217 Figure 3.62. Stacked 1H NMR of NCr(NiPr2)2(5-OMeInd) (10a) pre and post SST. 218 Figure 3.63. 1H NMR of NCr(NiPr2)2(4,6-diFInd) (10b). 219 Figure 3.64. 13C NMR of NCr(NiPr2)2(4,6-diFInd) (10b). 220 Figure 3.65. 19F NMR of NCr(NiPr2)2(4,6-diFInd) (10b). 221 Figure 3.66. Stacked 1H NMR of NCr(NiPr2)2(4,6-diFInd) pre and post SST (10b). 222 Figure 3.67. 1H NMR of NCr(NiPr2)2(5,6-diOMeInd) (10c). 223 Figure 3.68. 13C NMR of NCr(NiPr2)2(5,6-diOMeInd) (10c). 224 Figure 3.69. Stacked 1H NMR of NCr(NiPr2)2(5,6-diOMeInd) pre and post SST (10c). 225 Figure 3.70. 1H NMR of NCr(NiPr2)2(5-MeInd) (10d). 226 Figure 3.71. 13C NMR of NCr(NiPr2)2(5-MeInd) (10d). 227 Figure 3.72. Stacked 1H NMR of NCr(NiPr2)2(5-MeInd) pre and post SST (10d). 228 Figure 3.73. 1H NMR of NCr(NiPr2)2(py4-NMe2ph) (10e). 229 Figure 3.74. Stacked 1H NMR of NCr(NiPr2)2(py4-NMe2ph) pre and post SST (10e). 230 Figure 3.75. 1H NMR of reduced amine product N-(1,2-diphenylethyl)aniline. 231 Figure 3.76. 1H NMR of purified hydrolyzed product 1,2-diphenylethan-1-one. 232 Figure 3.77. 1H NMR of pure 2-benzyl-5-methyl-1-phenyl-1H-pyrrole. 233 Single Crystal X-ray Diffraction Thermal ellipsoids of titanium precatalysts are drawn with 50% probability level. The violet, blue, green, red, grey, and white spheres represent titanium, nitrogen, fluorine, oxygen, carbon, and hydrogen atoms, respectively. Figure 3.78. Structure of Ti(NMe2)2(ind-(CH2)-pyr) (1a’) recrystallized from ether/n-hexane. 234 Table 3.15. Crystallographic data and structural refinement of Ti(NMe2)2(ind-(CH2)-pyr) (1a’). Empirical formula Formula weight Temperature/K Crystal system Space group a/Å b/Å c/Å α/° β/° γ/° Volume/Å3 Z ρcalcg/cm3 μ/mm-1 F(000) Crystal size/mm3 Radiation 2Θ range for data collection/° Index ranges Reflections collected Independent reflections Data/restraints/parameters Goodness-of-fit on F2 Final R indexes [I>=2σ (I)] Final R indexes [all data] Largest diff. peak/hole / e Å-3 C17H22N4Ti 330.28 100.01(10) monoclinic P21/n 8.01410(10) 19.4569(3) 11.0591(2) 90 110.769(2) 90 1612.38(5) 4 1.361 4.499 696.0 0.202 × 0.165 × 0.038 Cu Kα (λ = 1.54184) 9.09 to 155.308 -9 ≤ h ≤ 10, -24 ≤ k ≤ 19, -13 ≤ l ≤ 13 10014 3199 [Rint = 0.0294, Rsigma = 0.0296] 3199/0/203 1.106 R1 = 0.0293, wR2 = 0.0785 R1 = 0.0312, wR2 = 0.0800 0.31/-0.28 235 Figure 3.79. Structure of Ti(NMe2)2(ind5-Me-CH2-pyr) (2a) recrystallized from ether/n-hexane. 236 Table 3.16. Crystallographic data and structural refinement of Ti(NMe2)2(ind5-Me-CH2-pyr) (2a). Empirical formula Formula weight Temperature/K Crystal system Space group a/Å b/Å c/Å α/° β/° γ/° Volume/Å3 Z ρcalcg/cm3 μ/mm-1 F(000) Crystal size/mm3 Radiation 2Θ range for data collection/° Index ranges Reflections collected Independent reflections Data/restraints/parameters Goodness-of-fit on F2 Final R indexes [I>=2σ (I)] Final R indexes [all data] Largest diff. peak/hole / e Å-3 C18H24N4Ti 344.31 100.00(10) triclinic P-1 11.80050(10) 12.7050(2) 13.05400(10) 84.5170(10) 71.1680(10) 69.3520(10) 1733.00(4) 4 1.320 4.207 728.0 0.134 × 0.072 × 0.045 Cu Kα (λ = 1.54184) 7.156 to 155.408 -14 ≤ h ≤ 13, -15 ≤ k ≤ 16, -16 ≤ l ≤ 16 51368 7247 [Rint = 0.0368, Rsigma = 0.0210] 7247/0/310 1.080 R1 = 0.0338, wR2 = 0.0859 R1 = 0.0361, wR2 = 0.0870 0.57/-0.64 237 Figure 3.80. Unsymmetrical unit of C2 symmetric structure of NCr(NiPr2)2(5-OMeInd) (10a) recrystallized from pentanes. 238 Table 3.17. Crystallographic data and structural refinement of NCr(NiPr2)2(5-OMeInd). Empirical formula Formula weight Temperature/K Crystal system Space group a/Å b/Å c/Å α/° β/° γ/° Volume/Å3 Z ρcalcg/cm3 μ/mm-1 F(000) Crystal size/mm3 Radiation 2Θ range for data collection/° Index ranges Reflections collected Independent reflections Data/restraints/parameters Goodness-of-fit on F2 Final R indexes [I>=2σ (I)] Final R indexes [all data] Largest diff. peak/hole / e Å-3 C21H36CrN4O 412.54 100.00(10) orthorhombic Pnma 14.2422(3) 12.8612(6) 12.3121(3) 90 90 90 2255.23(13) 4 1.215 4.294 888.0 0.248 × 0.092 × 0.062 Cu Kα (λ = 1.54184) 9.496 to 155.024 -15 ≤ h ≤ 17, -15 ≤ k ≤ 16, -15 ≤ l ≤ 15 15005 2477 [Rint = 0.0347, Rsigma = 0.0261] 2477/0/220 1.076 R1 = 0.0289, wR2 = 0.0783 R1 = 0.0305, wR2 = 0.0795 0.53/-0.32 239 Figure 3.81. Structure of NCr(NiPr2)2(4,6-diFInd) (10b) recrystallized from pentane. 240 Table 3.18. Crystallographic data and structural refinement of NCr(NiPr2)2(4,6-diFInd). Empirical formula Formula weight Temperature/K Crystal system Space group a/Å b/Å c/Å α/° β/° γ/° Volume/Å3 Z ρcalcg/cm3 μ/mm-1 F(000) Crystal size/mm3 Radiation 2Θ range for data collection/° Index ranges Reflections collected Independent reflections Data/restraints/parameters Goodness-of-fit on F2 Final R indexes [I>=2σ (I)] Final R indexes [all data] Largest diff. peak/hole / e Å-3 Flack parameter C20H32CrF2N4 418.49 100.00(10) orthorhombic P212121 11.5215(3) 13.3080(3) 14.5734(6) 90 90 90 2234.51(12) 4 1.244 4.441 888.0 0.092 × 0.047 × 0.045 Cu Kα (λ = 1.54184) 8.998 to 155.368 -12 ≤ h ≤ 14, -16 ≤ k ≤ 8, -18 ≤ l ≤ 18 9807 4327 [Rint = 0.0455, Rsigma = 0.0552] 4327/0/253 1.077 R1 = 0.0469, wR2 = 0.0947 R1 = 0.0553, wR2 = 0.0975 0.68/-0.53 0.476(9) 241 Figure 3.82. Structure of NCr(NiPr2)2(5,6-diOMeInd) (10c) recrystallized from pentane. 242 Table 3.19. Crystallographic data and structural refinement of NCr(NiPr2)2(5,6-diOMeInd). Empirical formula Formula weight Temperature/K Crystal system Space group a/Å b/Å c/Å α/° β/° γ/° Volume/Å3 Z ρcalcg/cm3 μ/mm-1 F(000) Crystal size/mm3 Radiation 2Θ range for data collection/° Index ranges Reflections collected Independent reflections Data/restraints/parameters Goodness-of-fit on F2 Final R indexes [I>=2σ (I)] Final R indexes [all data] Largest diff. peak/hole / e Å-3 C22H38CrN4O2 442.56 100.00(10) monoclinic P21/n 13.21656(19) 13.11433(15) 15.3852(2) 90 114.9580(18) 90 2417.63(7) 4 1.216 4.071 952.0 0.11 × 0.082 × 0.032 Cu Kα (λ = 1.54184) 7.426 to 155.076 -16 ≤ h ≤ 15, -13 ≤ k ≤ 16, -15 ≤ l ≤ 18 18840 5015 [Rint = 0.0323, Rsigma = 0.0303] 5015/0/272 1.091 R1 = 0.0299, wR2 = 0.0800 R1 = 0.0332, wR2 = 0.0816 0.27/-0.32 243 Figure 3.83. Structure of NCr(NiPr2)2(5-MeInd) (10d) recrystallized from pentane. 244 Table 3.20. Crystallographic data and structural refinement of NCr(NiPr2)2(5,6-diOMeInd). Empirical formula Formula weight Temperature/K Crystal system Space group a/Å b/Å c/Å α/° β/° γ/° Volume/Å3 Z ρcalcg/cm3 μ/mm-1 F(000) Crystal size/mm3 Radiation 2Θ range for data collection/° Index ranges Reflections collected Independent reflections Data/restraints/parameters Goodness-of-fit on F2 Final R indexes [I>=2σ (I)] Final R indexes [all data] Largest diff. peak/hole / e Å-3 C21H36CrN4 396.54 100.00(10) monoclinic P21/n 7.71041(8) 24.8099(3) 11.71004(13) 90 102.8089(11) 90 2184.32(4) 4 1.206 4.377 856.0 0.197 × 0.065 × 0.057 Cu Kα (λ = 1.54184) 7.126 to 155.3 -9 ≤ h ≤ 9, -31 ≤ k ≤ 31, -14 ≤ l ≤ 14 28382 4599 [Rint = 0.0334, Rsigma = 0.0226] 4599/0/244 1.090 R1 = 0.0284, wR2 = 0.0735 R1 = 0.0306, wR2 = 0.0746 0.28/-0.29 245 Figure 3.84. Structure of NCr(NiPr2)2(py4-NMe2ph) (10e) recrystallized from pentane. 246 Table 3.21. Crystallographic data and structural refinement of NCr(NiPr2)2(py4-NMe2ph). Empirical formula Formula weight Temperature/K Crystal system Space group a/Å b/Å c/Å α/° β/° γ/° Volume/Å3 Z ρcalcg/cm3 μ/mm-1 F(000) Crystal size/mm3 Radiation 2Θ range for data collection/° Index ranges Reflections collected Independent reflections Data/restraints/parameters Goodness-of-fit on F2 Final R indexes [I>=2σ (I)] Final R indexes [all data] Largest diff. peak/hole / e Å-3 C24H41CrN5 451.62 100.00(10) monoclinic P21/c 9.43369(16) 30.3002(4) 9.95162(17) 90 116.733(2) 90 2540.54(8) 4 1.181 3.834 976.0 0.126 × 0.102 × 0.068 Cu Kα (λ = 1.54184) 5.834 to 155.196 -11 ≤ h ≤ 11, -38 ≤ k ≤ 35, -12 ≤ l ≤ 10 18393 5287 [Rint = 0.0559, Rsigma = 0.0368] 5287/0/282 1.042 R1 = 0.0512, wR2 = 0.1248 R1 = 0.0548, wR2 = 0.1271 1.53/-0.53 247 Figure 3.85. Structure of NCr(NiPr2)2(Ind) (10f) recrystallized from pentane. 248 Table 3.22. Crystallographic data and structural refinement of NCr(NiPr2)2(Ind) (10f). Empirical formula Formula weight Temperature/K Crystal system Space group a/Å b/Å c/Å α/° β/° γ/° Volume/Å3 Z ρcalcg/cm3 μ/mm-1 F(000) Crystal size/mm3 Radiation 2Θ range for data collection/° Index ranges Reflections collected Independent reflections Data/restraints/parameters Goodness-of-fit on F2 Final R indexes [I>=2σ (I)] Final R indexes [all data] Largest diff. peak/hole / e Å-3 C20H34CrN4 382.51 100(2) monoclinic P21/n 8.5381(2) 17.3549(3) 14.6630(3) 90 102.388(2) 90 2122.14(8) 4 1.197 4.488 824.0 0.205 × 0.047 × 0.039 CuKα (λ = 1.54184) 8.004 to 154.066 -10 ≤ h ≤ 10, -21 ≤ k ≤ 21, -18 ≤ l ≤ 15 19637 4262 [Rint = 0.0374, Rsigma = 0.0306] 4262/0/234 1.051 R1 = 0.0450, wR2 = 0.1093 R1 = 0.0486, wR2 = 0.1115 1.28/-0.59 249 Figure 3.86. Structure of NCr(NiPr2)2(3-MeInd5-OMe) (10g) recrystallized from pentane. 250 Table 3.23. Crystallographic data and structural refinement of NCr(NiPr2)2(3-MeInd5-OMe) (10g). Empirical formula Formula weight Temperature/K Crystal system Space group a/Å b/Å c/Å α/° β/° γ/° Volume/Å3 Z ρcalcg/cm3 μ/mm-1 F(000) Crystal size/mm3 Radiation 2Θ range for data collection/° Index ranges Reflections collected Independent reflections Data/restraints/parameters Goodness-of-fit on F2 Final R indexes [I>=2σ (I)] Final R indexes [all data] Largest diff. peak/hole / e Å-3 C22H38CrN4O 426.56 100.01(10) monoclinic P21/n 14.3059(2) 25.0018(3) 14.9403(2) 90 117.353(2) 90 4746.27(13) 8 1.194 4.096 1840.0 0.129 × 0.066 × 0.055 Cu Kα (λ = 1.54184) 7.072 to 153.662 -18 ≤ h ≤ 17, -31 ≤ k ≤ 26, -18 ≤ l ≤ 16 37019 9313 [Rint = 0.0308, Rsigma = 0.0279] 9313/0/533 1.027 R1 = 0.0316, wR2 = 0.0762 R1 = 0.0361, wR2 = 0.0781 0.27/-0.33 251 REFERENCES (1) Billow, B. S.; McDaniel, T. J.; Odom, A. L. Quantifying Ligand Effects in High-Oxidation- State Metal Catalysis. Nature Chem 2017, 9, 837–842. (2) Hou, Z.; Jena, R.; McDaniel, T. J.; Billow, B. S.; Lee, S.; Barr, H. I.; Odom, A. L. Modeling Complex Ligands for High Oxidation State Catalysis: Titanium Hydroamination with Unsymmetrical Ligands. ACS Catal. 2024, 5531–5538. (3) Ishiyama, T.; Takagi, J.; Yonekawa, Y.; Hartwig, J. F.; Miyaura, N. Iridium‐Catalyzed Direct Borylation of Five‐Membered Heteroarenes by Bis(Pinacolato)Diboron: Regioselective, Stoichiometric, and Room Temperature Reactions. Adv Synth Catal 2003, 345, 1103–1106. (4) Blakemore, D. Chapter 1. Suzuki–Miyaura Coupling; pp 1–69. (5) Rieth, R. D.; Mankad, N. P.; Calimano, E.; Sadighi, J. P. Palladium-Catalyzed Cross-Coupling of Pyrrole Anions with Aryl Chlorides, Bromides, and Iodides. Org. Lett. 2004, 6, 3981–3983. (6) Pohlki Dipl.-Chem., F.; Doye, S. The Mechanism of the [Cp2TiMe2]-Catalyzed Intermolecular Hydroamination of Alkynes. 2001, 40, 2305–2308. (7) Hao, H.; Schafer, L. L. Metal–Ligand Cooperativity in Titanium-Catalyzed Anti-Markovnikov Hydroamination. ACS Catal. 2020, 10, 7100–7111. (8) Hou, Z. Explorations of Heterocycles and Metal Complexes as Drug Candidates and Investigations of Ligand Effects in Titanium Catalysis., 2022. (9) Beesley, R. M.; Ingold, C. K.; Thorpe, J. F. CXIX.—The Formation and Stability of Spiro- Compounds. Part I. Spiro-Compounds from Cyclohexane. J. Chem. Soc., Trans. 1915, 107, 1080–1106. (10) Carlson, R.; Carlson, J. E. Design and Optimization in Organic Synthesis, 2nd rev. and enl. ed.; Else vier: Amsterdam, 2005. (11) Shi, Y.; Hall, C.; Ciszewski, J. T.; Cao, C.; Odom, A. L. Titanium Dipyrrolylmethane Derivatives: Rapid Intermolecular Alkyne Hydroamination. Chem. Commun. 2003, No. 5, 586–587. (12) Ramanathan, B.; Keith, A. J.; Armstrong, D.; Odom, A. L. Pyrrole Syntheses Based on Titanium-Catalyzed Hydroamination of Diynes. Org. Lett. 2004, 6, 2957–2960. (13) Falivene, L.; Credendino, R.; Poater, A.; Petta, A.; Serra, L.; Oliva, R.; Scarano, V.; Cavallo, L. SambVca 2. A Web Tool for Analyzing Catalytic Pockets with Topographic Steric Maps. Organometallics 2016, 35, 2286–2293. (14) Xu, J.; Rawal, V. H. Total Synthesis of (−)-Ambiguine P. J. Am. Chem. Soc. 2019, 141, 4820– 4823. 252 (15) DiFranco, S. A.; Maciulis, N. A.; Staples, R. J.; Batrice, R. J.; Odom, A. L. Evaluation of Donor and Steric Properties of Anionic Ligands on High Valent Transition Metals. Inorg. Chem. 2012, 51, 1187–1200. (16) Aldrich, K. E. An Exploration of Mid- to High-Valent Transition Metal Complexes for Application to Catalysis; 2019. (17) Haak, E.; Bytschkov, I.; Doye, S. Intermolecular Hydroamination of Alkynes Catalyzed by Dimethyltitanocene. Angew. Chem. Int. Ed. 1999, 38, 3389–3391. 253 CHAPTER 4: MECHANISTIC STUDIES OF PYRIDINE SYNTHESIS FROM ISOXAZOLE* *Reprinted in part with permission from the Royal Society of Chemistry. Lee, S.; Jena, R.; Odom, A. L. Substituted Pyridines from Isoxazoles: Scope and Mechanism. Org. Biomol. Chem., 2022, 20, 6630-6636. 4.1 Introduction An extensive range of applications can be found for pyridine derivatives in materials science, agrochemicals, medicines, pharmaceuticals, and other fields. Pyridines are extremely significant in medicinal chemistry because of their broad chemical properties and versatile structural nature.1 Thus, in organic synthesis and drug discovery endeavors, the development of efficient and reliable synthetic methods for the synthesis of substituted pyridines is crucial. The modification of a pharmacologically active heterocyclic core to synthesize novel molecules with biological activity has gained a lot of attention. In the field of medicinal chemistry, scaffold hopping is used to change one heterocyclic core to another by enhancing its substituents. This strategy allows the exploration of structural relationships related to biological activity.2 Traditionally, scientists follow a linear approach by gradually changing a core structure and building complexity upon it. Meanwhile, scaffold hopping could replace the core structure while maintaining the peripheral substituents. The method of core changing (ring expansion, contraction, or alteration) provides different strategies for rapid exploration of the structural relationships to biological activity. The inverse electron-demand hetero-Diels-Alder reaction between electron-rich olefins and isoxazole, which produces pyridines, is less explored. This reaction eliminates the oxygen atom 254 while substituting a pyridine for a core of substituted isoxazoles, or isoxazoles. Additionally, this method regioselectively adds additional substituents to the ring (Scheme 4.1). Despite being known for a long time, this reaction is still completely unexplored, offering a great chance for further exploration in organic synthesis.3 This is an appealing approach for obtaining structurally varied pyridine derivatives from readily accessible starting materials. In this study, we explored the synthesis of substituted pyridines from isoxazoles, delving into the nature of this transition and the underlying molecular mechanism. Scheme 4.1. Pyridine synthesis with isoxazole and 1-pyrrolidine-1-cyclohexene. The original idea put out by Ohta et al. involved reacting enamine with isoxazole to produce pyridine-N-oxide compounds that could be reduced by zinc and TiCl4.4 Dr. Seokjoo Lee tested different reducing reagents in our group and substituted fuming TiCl4 with TiCl4(THF)2. Remarkably, he found that Ti powder produced the best yield, matching or even exceeding those of the initial investigation, demonstrating the reaction’s reliability for pyridine synthesis.5 4.2 Insights into the Reaction Mechanism through DFT Analysis Seokjoo synthesized and characterized all the compounds in this study. We collaboratively carried out the Density Functional Theory (DFT) calculations to understand the mechanism involved with the pyridine synthesis from isoxazole and enamine. In this chapter, I will discuss the mechanistic studies elucidating the intricate details of the reaction pathways. The proposed mechanism of the pyridine synthesis (Scheme 4.2) involves [4+2]-cycloaddition of substituted isoxazole with the enamine to form a [2.2.1]-oxazabicyclic intermediate. The 255 observed product shows R5 in the α-position adjacent to the pyridine nitrogen, suggesting the alternative cycloaddition to be possibly unfavorable. There are two possible major pathways from the bicyclic intermediate: (A) initial loss of amine followed by ring opening and (B) initial ring opening followed by amine loss. Substituted pyridine-N-oxide is the resultant compound, which can be reduced in situ to produce pyridine directly. Scheme 4.2. The inverse-electron demand hetero-Diels Alder reaction isoxazole with enamine. The mechanism of the reaction was investigated using the Gaussian 16 B.01, B3LYPD3 with aug- cc-PVDZ.6–8 In the initial investigation, gas phase calculations were carried out. Path A and B share an initial [4 + 2]-cycloaddition, where our investigations into the mechanism began. The cycloaddition step gives us four possible isomers: endo and exo of I1-1 and I1-2 (Figure 4.1). The energy of isomers I1-1 exo/endo is ~1-2 kcal/mol lower than I1-2 (Figure 4.1). However, the activation energy for the formation of I1-1 is approximately 14 kcal/mol lower than I1-2; therefore, I1-1 would be the more favored regioisomer over I1-2. The energy barrier for the formation of endo isomers is always slightly lower than exo isomers; whereas, thermodynamically I1-1 exo is somewhat more stable than I1-1 endo. We can say that I1-1 exo is the thermodynamic 256 product, and I1-1 endo is the slightly favored kinetic product. In all likelihood, both endo and exo I1-1 are generated; however, both lead to the same observed final product after amine elimination. Figure 4.1. (top) Free energy profile for cycloaddition (CA) reaction, (bottom) diagram of the I1- 1 endo. According to the Natural Bond Orbital calculations, for I1-1, the lone pair on N1 donates its electron density into the C2-N3 antibonding orbital (Figure 4.1 bottom). Furthermore, there is a donation from N1 to the antibonding orbital of the C-H bond at C2. With another regioisomer I1- 2, there are no such secondary interactions possible. The cycloaddition step follows a concerted mechanism in the gas phase. The transition states for the cycloaddition step are symmetrical, i.e., two bonds are made to the electron-rich olefin, and 257 asynchronous, i.e., the asymmetric dienophile leads to a different amount of bond-making in the two new bonds. All the transition states were asynchronous, but the symmetry changes when incorporating solvent models into our calculations (vide infra). To quantify the trend in the symmetry of the transition state (TS), we used bond length ratios τ (Eq 1). This ratio τ can be used to discern the nature of the TS. Subjectively, when τ < ~0.7, the transition state looks asymmetric, and when τ > 0.7, the transition state looks symmetric. 𝜏 = (𝐶−𝑁 𝑏𝑜𝑛𝑑 𝑙𝑒𝑛𝑔𝑡ℎ 𝑖𝑛 𝑇𝑆)/(𝐶−𝑁 𝑏𝑜𝑛𝑑 𝑙𝑒𝑛𝑔𝑡ℎ 𝑖𝑛 𝑃𝑟𝑜𝑑𝑢𝑐𝑡) (𝐶−𝐶 𝑏𝑜𝑛𝑑 𝑙𝑒𝑛𝑔𝑡ℎ 𝑖𝑛 𝑇𝑆)/(𝐶−𝐶 𝑏𝑜𝑛𝑑 𝑙𝑒𝑛𝑔𝑡ℎ 𝑖𝑛 𝑃𝑟𝑜𝑑𝑢𝑐𝑡) Eq. 4.1 Table 4.1. The τ value for cycloaddition in different solvents with their dielectric constant (). No solvent 0 0.693 0.745 0.909 Heptane 1.92 0.678 0.708 0.896 1,4- Dioxane 2.21 0.676 0.699 0.893 THF 7.52 0.682 0.649 0.882 DCM EtOH H2O 8.93 0.684 0.649 0.878 24.6 0.692 0.591 0.885 78.54 0.695 0.588 0.845 0.873 0.865 0.864 0.857 0.854 0.844 0.823  TS1-I- endo TS1-I-exo TS1-II- endo TS1-II- exo After the cycloaddition step, this reaction mechanism divaricates into two pathways. In Path A, I1 (cycloaddition products, endo, and exo) eliminates ammonia and forms a double bond within the pyridine ring to make I2_α (Figure 4.2). Then, the C–O bond cleavage of norbornadiene-like I2_α leads to pyridine-N-oxide (I3). In Path B, C–O bond cleavage in I1 leads to dihydropyridine- N-oxide I2_β, and, losing ammonia, the pyridine ring aromatizes to give pyridine-N-oxide (I3). In Path A, norbornadiene-like I2_α (plus ammonia) is roughly 26 kcal/mol higher in energy than the initial cycloaddition products (I1), and the transition state energy is 61.5 kcal/mol (TS2_α). The 258 activation energy for the formation of the pyridine-N-oxide (TS3_α) is 60 kcal/mol, and this is an exergonic by 100 kcal/mol. Figure 4.2. Energy diagram for two pathway mechanisms for exo (right) and endo (left). In Path B, cyclohexadiene-like I2_β is 21.7 kcal/mol lower in energy than the initial cycloaddition product I1, and the activation energy is 38.7 kcal/mol. Elimination of the amine to make the pyridine-N-oxide (Step 2 in Path B) is also exergonic, with an activation barrier of 29.3 kcal/mol. 4.3 Effects of Solvent on the Mechanistic Pathway To get more realistic data, we added a solvent model and reoptimized the pathway. It was noticed that TS1-1 (Figure 4.3), both exo and endo, are asymmetric when a solvent model is used; whereas, TS1-2 endo/exo were both symmetric. The asymmetric transition states for TS1-1 endo/exo show more C–N bond formation than C–C bond. While the nature of the transition state is different with the inclusion of the solvent model, the energy of the transition state is similar to the gas phase calculation. To investigate the effect of solvent polarity on the energies and structures, the compounds and transition states were reoptimized in different solvents using the SMD model. The energies were recalculated in heptane, dioxane, tetrahydrofuran (THF), 259 dichloromethane (DCM), ethanol (EtOH), and water (H2O). Figure 4.3 shows the variation of τ value (Eq. 4.1) in different solvents. The only transition state structure showing structural changes (by this metric) as a function of solvent polarity is TS1-1 exo, which gets more asymmetric as the polarity increases. The other isomers of the transition state showed little energy change with solvent polarity. Figure 4.3. Correlation between τ (Eq. 4.1) and dielectric constant (κ) of different solvents. It was observed that I1-1 endo/exo are energetically favored products irrespective of the solvent polarity (Figure 4.4). In low polarity solvents, I1-1 exo is slightly less kinetically favored than endo isomer, but in high polarity solvents, I1-1 exo is both kinetically and thermodynamically favored product. 260 Figure 4.4. (top) Energy barrier ∆G‡ (kcal/mol) of cycloaddition transition states in different solvents (top), (bottom) free energy for cycloaddition products ∆Grxn (kcal/mol). Table 4.2. Transition state energy barrier (∆G‡) and free energy of cycloaddition products (∆Grxn) in different solvents. E (kcal/mol) Reactants TS1-I-endo TS1-I-exo TS1-II-endo TS1-II-exo I1-I-endo I1-I-exo I1-II-endo I1-II-exo Gas Phase 0.00 41.93 42.89 54.46 55.25 31.27 27.94 34.44 32.86 Heptane 0.00 40.69 42.12 54.97 55.92 30.70 27.66 34.23 32.82 THF 0.00 39.42 40.64 55.45 56.67 31.10 28.36 34.40 33.28 DCM 0.00 39.52 40.83 55.81 56.94 31.24 28.55 34.59 33.46 EtOH 0.00 37.13 36.56 54.50 55.67 29.01 26.57 32.45 31.31 H2O 0.00 36.17 35.22 54.02 54.86 26.94 24.57 30.35 28.95 Dioxane 0.00 39.51 40.90 54.10 55.13 29.70 26.70 33.18 31.84 261 After gas phase calculation we moved to a more realistic system where metal was involved in the reaction. In actuality, a variety of Lewis acids may be present in a solution like TiCl3. However, as a catalyst, we looked at TiCl4, the strongest Lewis acid in the solution. For this discussion, TiCl4 is abbreviated as [M] in Figure 4.5. The cycloaddition of enamine and isoxazole is crucial, and in the presence of Lewis acid, the pathway underwent a significant alteration (blue path, Figure 4.5). It was discovered that the enamine and isoxazole reacted sequentially, first forming a C–C bond and subsequently a C–N bond. This stepwise addition has a significantly lower energy barrier (than when the Lewis acid was absent. The Lewis acid stabilizes the transition state barriers (TS1A_M and TS1B_M) along with substantial stabilization of the cycloaddition product (I1_M, ⁓17 kcal/mol). Figure 4.5. Calculated pathway to the endo cycloaddition product between enamine and isoxazole with (blue) and without (black) [M] = TiCl4. 262 Reaction CoordinateD G (kcal/mol)NOH2N+[M]NONH2ONNH2ONNH2[M]051015202530354045-5ONNH2[M]ONNH2[M]ONNH2[M]TS1A_M+16.41I1A_M+16.2TS1B_M+18.1I1_M+17.3I1+31.3TS1+41.9SM0NOH2N+[M] As discussed above, there are two possible pathways from cycloaddition product (I1 or I1_M): (1) ring opening followed by amine loss and (2) initial amine loss followed by ring opening. In the presence of Lewis acid, the energy for each step has consistently reduced. In the case of initial amine loss followed by ring opening (Figure 4.6), the energy barrier for amine loss seems untenable (⁓60 kcal/mol). Figure 4.6. Calculated pathway starting with the cycloaddition product I1 for amine loss first, followed by ring opening with (blue) and without (black) TiCl4. [M] = TiCl4. The pathway involving ring opening first, followed by amine loss (Figure 4.7) has less than half energy barrier magnitude (35.9 and 27.3 kcal/mol with and without TiCl4) for the first step. After the ring opens, the aromatization of the pyridine ring may aid in the amine loss and give a lower energy barrier for this step. This suggests the likelihood of I1 or I1_M to undergo ring opening first than amine loss. We were unable to identify the transition state for amine loss at ring opening 263 Reaction CoordinateD G (kcal/mol)010203040506070-30-60ONNH2[M]ONHH2NONOd–NNOδ+I1_M0I10ONNH2ONHH2N[M]ON[M]Od–N[M]δ+NO[M]TS2_a63.8TS2_M_a62.0I2_a26.8I2_M_a19.0TS3_a59.8TS3_M_a41.2I3–40.5I3_M–56.0 with TiCl4 (TS2_M_β). However, as we can see in Figure 4.7 all transition states and intermediates with Lewis acid bound have lower energy, we would assume it has lower energy than TS2_ β. Figure 4.7. Calculated pathway starting with the cycloaddition product I1 for ring opening first followed by amine loss with (blue) and without (black) TiCl4. [M] = TiCl4. As a result, the calculations indicate that a Lewis acid can change the reaction pathways in the case of the cycloaddition step (stepwise addition of diene). Furthermore, it allowed for a lower energy pathway involving cleavage of the C–O bond first, followed by amine loss. Essentially, the aromatization of the pyridine ring helps with the high-energy amine loss. In I1, one would assume that N–O bond cleavage is more likely to happen than C–O due to the lone-pair bond weakening effect.9 We were curious why we observed heterolytic C–O bond cleavage over homolytic N–O cleavage. To answer this question, we performed calculations to calculate the energy barrier associated with each process. In the gas phase, it was shown that 264 Reaction CoordinateD G (kcal/mol)-50-40-30-20-10010203040-60ONNH2[M]–OdNH2N[M]ONHNH2ONHNH2NOOd–NH2NONHNH2[M]NO[M]I1_M0I10ONNH2I3–40.5I3_M–56.0d+TS3_b+35.9I2_M_b–27.3.0I2_b–19.4TS3_M_b27.3TS2_b+7.95TS2_M_bNFd+ heterolytic cleavage of the N-O bond in I2_ to create a diradical had a significantly lower barrier (+27.1 kcal/mol) than heterolytic cleavage of the C–O bond (Figure 4.8). However, in the case of 1,4-dioxane, the C–O heterolytic cleavage has a barrier of 33.5 kcal/mol, and the homolytic N–O cleavage has a higher energy barrier of 49.9 kcal/mol. This reversed stability of I3 in the solvent can be attributed to stabilizing the incipient charges in the solvent. Figure 4.8. (top) Calculation associated with homolytic C–O bond cleavage vs heterolytic N–O bond cleavage in gas phase, (bottom) in 1,4-dioxane. 265 4.4 Conclusions A novel approach uses an inverse-electron demand hetero-Diels Alder reactions to convert the isoxazole core into pyridines via enamines. This technique was transformed by replacing fuming TiCl4 with convenient TiCl4(THF)2. In this study, we explored the mechanistic pathway associated with this transformation. DFT calculations were performed using B3LYPD3 functional, aug-cc-PVDZ basis set, and SMD solvent model. The presence of TiCl4 made the concerted [4+2]-cycloaddition of enamine and isoxazole a stepwise addition of enamine with C–C bond formation first followed by C–N bond to the isoxazole. Using Lewis acidic metal, the energy barrier associated with the first step was greatly reduced (from 41.9 kcal/mol to 16.4 kcal/mol) making the reaction more feasible. After the cycloaddition step, the preferred pathway computationally involved the initial ring opening of the oxaza-[2.2.1]-bicycle formed to give an imine-N-oxide that could then lose amine to aromatize the pyridine ring (Figure 4.6). This pathway was energetically preferred over initial amine loss, followed by ring opening (Figure 4.5). Furthermore, the calculations suggest the formation of pyridine-N-oxide is more favorable than biradical (Figure 4.8) in 1,4-dioxane. This provides insights beyond what chemical intuition can alone offer because we would expect cleavage of the N–O bond over the C–O bond. Hence, calculations here offered us to understand the complexity of molecular behavior and the importance of relying on empirical evidence. 266 REFERENCES (1) Recent Developments in https://doi.org/10.1016/C2020-0-02397-2; Elsevier, 2023. Synthesis the and Applications of Pyridines, (2) Hu, Y.; Stumpfe, D.; Bajorath, J. Recent Advances in Scaffold Hopping: Miniperspective. J. Med. Chem. 2017, 60, 1238–1246. (3) Foster, R. A. A.; Willis, M. C. Tandem Inverse-Electron-Demand Hetero-/Retro-Diels–Alder Reactions for Aromatic Nitrogen Heterocycle Synthesis. Chem. Soc. Rev. 2013, 42, 63–76. (4) Ohta, K.; Iwaoka, J.; Kamijo, Y.; Okada, M.; Nomura, Y. Formation of Pyridines by the Reaction of Lsoxazoles with Enamines. NIPPON KAGAKU KAISHI 1989, No. 9, 1593–1600. (5) Lee, S.; Jena, R.; Odom, A. L. Substituted Pyridines from Isoxazoles: Scope and Mechanism. Org. Biomol. Chem. 2022, 20, 6630–6636. (6) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. A Consistent and Accurate Ab Initio Parametrization of Density Functional Dispersion Correction (DFT-D) for the 94 Elements H- Pu. The Journal of Chemical Physics 2010, 132, 154104. (7) Weigend, F.; Ahlrichs, R. Balanced Basis Sets of Split Valence, Triple Zeta Valence and Quadruple Zeta Valence Quality for H to Rn: Design and Assessment of Accuracy. Phys. Chem. Chem. Phys. 2005, 7, 3297. (8) Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. Universal Solvation Model Based on Solute Electron Density and on a Continuum Model of the Solvent Defined by the Bulk Dielectric Constant and Atomic Surface Tensions. J. Phys. Chem. B 2009, 113, 6378–6396. (9) Lauvergnat, D.; Maître, P.; Hiberty, P. C.; Volatron, F. Valence Bond Analysis of the Lone Pair Bond Weakening Effect for the X−H Bonds in the Series XH n = CH 4 , NH 3 , OH 2 , FH. J. Phys. Chem. 1996, 100, 6463–6468. 267 CHAPTER 5: SYNTHESIS, STRUCTURE, AND PROPERTIES OF A RARE YTTRIUM ISOCYANIDE COMPLEX FROM UNPRECEDENTED DIVALENT YTTRIUM COMPLEX* *Reprinted in part with permission from the Royal Society of Chemistry. Jena, R.; Benner, F.; Delano IV, F., Holmes, D.; McCracken, J.; Demir, S.; Odom, A. L. A rare isocyanide derived from an unprecedented neutral yttrium(II) bis(amide) complex. Chem. Sci. 2023, 14, 4257–4264. 5.1 Introduction In organometallic chemistry, ligands play a crucial role in facilitating the isolation of metal in unconventional oxidation states. Usually, all metals have some preferred oxidation states. For instance, group 3 metals and lanthanides(Ln) commonly show +3 formal oxidation states irrespective of the number of electrons in the f-orbital for lanthanides. Despite their most stable oxidation state of +3, lanthanides can also exhibit an oxidation state of +2 by proper design of the ligand environment.1–3 Over the past few decades, researchers have been engaged in the study of these unusual oxidation-state metal complexes and their the metal-ligand interactions, reactivity, and new electronic structures. By 2016, all lanthanides were reported in a +2 oxidation state except radioactive promethium.3 The choice of ligands is pivotal when trying to isolate rare, low-valent metal complexes.4 Due to the larger size of group 3 and lanthanide ions, they prefer to bind with a ligand system that provides steric saturation around the metal center to circumvent ligand reorganization. Lappert and coworkers isolated the first structurally characterized lanthanide in +2 oxidation state i.e., La(II) using the substituted cyclopentadienyl 1,3-(SiMe3)2C5H3 ligand (Figure 5.1).5 Further, the 268 substitution of a cyclopentadienyl ring also governs the reactivity of Ln(III) ions towards reduction of the metal center to Ln(II) or reduction of dinitrogen.5–7 After the remarkable discovery of the La(II) species, subsequent research has led to the discovery of yttrium and other lanthanides in the +2 oxidation state.1,2 Figure 5.1. First example of La(II) molecular complex. In 2012, Evans reported the first example of the Y(II) complex, [Y(C5H4SiMe3)3]–, which was only stable at −48 C.1 Ligands such as cyclopentadienyls,4,8,9 aryloxides,10 and hexamethyldisilazides11 have been used to stabilize low valent Y(II) metal complexes. Recently Evans and coworkers have shown the stability of the Y(II) complex is a function of the counterion.12 Among the handful of examples of structurally characterized Y(II) complexes (Figure 5.2), the most stable yttrium tris(aryloxide) complex has a lifetime of 48 h in THF solution.10 These unstable metal complexes pose a challenge when it comes to studying their reactivity and properties. In this study, we focused on accessing yttrium in a low oxidation state and studying its reactivity, which could provide valuable insights that can be extended to other lanthanide metals. 269 Figure 5.2. Reported Y(II) complexes. Ad = 1-adamantyl. In 1986, the first fully characterized example of 6-arene coordinating Sm(II) complex Sm(6- C6Me6)(µ2-AlCl4)3 was reported by Cotton and Schwotzer.13 Evans, Long, and Cloke took advantage of the π-interaction of the ligand system to stabilize a host of low-valent metal complexes.8 Recently, we reported the first example of a neutral divalent U(NHAr*)2 (Ar* = 2,6- (2,4,6-iPr3C6H2)C6H3) (synthesis is shown in Scheme 5.1)14 complex by using a terphenyl-based amide ligand system where the metal is sandwiched between two phenyl rings showing 6- coordination to the metal center. This ligand (NH2Ar*) system not only provides steric stabilization to the metal center but also introduces electronic effects by accepting electron density from the HOMO of the reduced metal center into the π*-orbital (LUMO) of the 6-coordinated ring system. A neodymium complex has already been reported utilizing the NHAr* ligand system.15 This ligand system seemed a good choice to us for accessing the Y(II) metal complex. 270 Scheme 5.1. Synthesis of first Neutral bis(amide) U(II) complex. (Ar = 2,4,6-iPr3C6H2). 5.2 Synthesis of Potassium Terphenyl Amide Ligand System As mentioned before, the terphenyl-based amide ligand is a suitable candidate for stabilizing low-coordination metal ions. These sterically encumbered ligands have been shown to have extensive utility in the synthesis of several species with unusual bonding and coordination numbers. One of the examples is shown in Scheme 5.1. Synthesis of the terphenyl amine ligand was carried out using the literature-reported process from the Power and Tilley groups.16,17 The ligand was deprotonated by using (trimethyl)silylmethyl potassium as a base (Scheme 5.2) which was easily synthesized from commercially available (trimethyl)silylmethyl lithium and potassium-tert-butoxide.18 The resulting amide was isolated pure after washing with cold n-hexane to be used in the next step. This procedure is simpler than the extensive synthesis reported by Liddle where terphenyl amine and KH were refluxed for 5 days in THF to make the potassium terphenyl amide ligand precursor.15 The proton NMR of the deprotonated terphenyl amide was consistent with the literature.15 We discovered that the deprotonated ligand (KNHAr*, Ar* = 2,6-(2,4,6-iPr3C6H2)C6H3) is best used immediately after preparation. The monopotassium salt likely undergoes an equilibrium to generate H2NAr + K2NAr, and the amine can be difficult to remove from the product if it is an impurity in the metal complex because of similar solubilities. 271 Scheme 5.2. Synthesis of potassium terphenyl amide ligand. 5.3 Synthesis and Properties of Y(NHAr*)2Cl The initial synthesis of Y(NHAr*)2Cl (1) was carried out by Florian Benner and Francis Delano from the Demir group at MSU. The salt metathesis of 2 equiv. of KNHAr* ligand with yttrium chloride in diethyl ether afforded Y(NHAr*)2Cl (1) in 64% recrystallized yield (Figure 5.3). Unlike the U(NHAr*)2I complex (Scheme 5.1), single-crystal X-ray diffraction on 1 showed only one of the ortho aryl rings of the terphenyl amide ligand system 6-coordinated to the metal center giving rise to a piano stool kind of geometry around the metal center. The extended radial extension of 5f orbitals in uranium might be responsible for the 6-interactions from both of the ortho-aryl ring of the ligand, which is not possible for Y due to the limited radial extension of the 4d orbital. There is literature precedent of a similar kind of bonding pattern with lanthanum, cerium and praseodymium, where 6-arene coordination was observed from one of the phenyl rings of tris(aryloxide) ligand.19 272 Figure 5.3. (Top) Synthesis of Y(NHAr*)2Cl (1) and (bottom) structure of 1 from single-crystal X-ray diffraction (thermal ellipsoids drawn at 50%). Solvent molecules (2 n-hexane) and hydrogen atoms on carbon are removed for clarity. The 6-arene (red) is discussed as Ar1 in the text. The distance between 6-arene centroid to the yttrium metal center was 2.493(3) Å, while the second closest distance with arene ring was 3.699(3) suggesting only one 6-arene coordination is present in 1. The centroid (Ar1)−Y1−N1 angle is 94.9(7)°, and the centroid (Ar1)−Y1−N2 angle is 101.8(6)°. In the solid-state the two amides are inequivalent in bond length Y1−N1 is 2.249(2) Å, and Y1−N2 is 2.210(2) Å. The bond angle Y1−N1−C1(ipso−Ar1) is 131.5(2)°, and Y1−N2−C36(ipso−Ar2) is 144.0(2)°. The largest disparity between the two amides is due to the 𝜂6-arene coordination involved only with Ar1, not Ar2. In proton NMR, N-H protons moved downfield to 4.57 ppm (contrasting with the free ligand resonance at 2.62 ppm) after coordination to the metal center as expected. However, 1H and 13C NMR spectroscopy suggest in solution all 273 the ortho-aryl resonances are equivalent on the timescale of these experiments, indicating fast exchange between bound and unbound ortho-aryl groups of 1. Complex 1 is stable for months in the freezer at ‒30 C. 5.4 Generation and Characterization of a Novel Neutral Y(II) Complex Figure 5.4. (Top) Synthesis of Y(NHAr*)2 (2), (bottom) structure of 2 from single-crystal X-ray diffraction (thermal ellipsoids drawn at 50%). The solvent molecule (1 THF) and hydrogen atoms on carbon are removed for clarity. Similar to the synthesis of the U(II) complex, reduction of Y(NHAr*)2Cl (1) was carried out by using the common reductant KC8 resulting in the formation of Y(NHAr*)2 (2) along with precipitation of easily removable KCl and graphite in the THF solution (Figure 5.4). The reaction color changed from pale yellow to black within an hour of reaction time. The dark solid was 274 recrystallized using n-hexane as solvent, which afforded dark yellow crystals overnight at −35 C. This Y(NHAr*)2 (2) complex is the first example of a neutral Y(II) complex. The geometries of 1 and 2 are significantly different in solid state. Y(NHAr*)2 (2) crystallizes as a C2 symmetric molecule with only half of the molecule occupying the asymmetric unit similar to previously reported U(NHAr*)2. The significant variations in the N–Y–N angles between the solid-state structures of 1 and 2 are worth mentioning in the comparison. The N–Y–N angle in complex 1 of 133.7(1) exhibits a greater degree of obtuseness compared to 2 120.4(1). This is due to an additional 𝜂6-arene interaction of the ortho aryl from 1 to 2. In the formally reduced complex 2, the Y1–Ar1(centroid) distance shortened slightly relative to 1 from 2.493(3) Å to 2.468(8) Å in 2. The perceived shrinkage of Y–Ar(centroid) distance in 2 is due to the increased backbonding in the formally reduced metal center to the π* of the Ar1 ring which is consistent with the previous U(NHAr*)2 complex.14 Furthermore, the analysis of the average C–C bond distance in Ar1 shows a subtle lengthening of bonds relative to 1 from 1.402(5) Å to 1.413(3) Å in 2. Naturally, the expectation is that the radius of the metal will increase on reduction, which is observed in the Y–N distance in 1 vs.2 of 2.230(2) vs. 2.261(1) Å, respectively. The electrochemical properties of Y(NHAr*)2 (2) were accessed by cyclic voltammetry using [NBu4][B(3,5-(CF3)2C6H3)4] in diethyl ether as an electrolyte and glassy carbon as a working electrode. The reduction potential of Y(NHAr*)2 +1/0 couple was observed at E1/2 = –1.16 ± 0.01 V w.r.t to FeCp2 +1/0 (left, Figure 5.5). A reversible feature was observed when a voltammogram of 2 was obtained using different scan rates (right, Figure 5.5) using 1.5 mM solution of 2 and the conditions mentioned above (for numerical values see the experimental section). For comparison, a recent study provided a quasi-reversible Y(C5H4SiMe3)3 −/0 potential of E1/2 = −3.06 V in THF with NBu4 + BPh4 − as the electrolyte.20 Approximately three times lower reduction potential value 275 suggests that complex Y(C5H4SiMe3)3 is quite difficult to reduce as compared to our complex Y(NHAr*)2Cl (1) suggesting the reduced complex Y(NHAr*)2 (2) is relatively more stable. As mentioned above, the stability of the complex 2 is due to the two 𝜂6-arene interactions between metal and the ortho aryl groups, which in this case are acting as electron acceptor ligands making a complex that is easier to reduce. On the other hand, Y(C5H4SiMe3)3 contains silyl-substituted cyclopentadienyl ligands, which are presumably stronger donors to the metal center that destabilize the reduced species. Figure 5.5. (left) Cyclic voltammogram of 2 (1.5 mM) in diethyl ether with [NBu4][B(3,5- (CF3)2C6H3)4] (100 mM) as supporting electrolyte and FeCp2 +1/0 as 0. (right) Overlay plot for scan rates 100 mV/s (black) and 50 mV/s (red). These voltammograms were scanned in the negative direction. The absorption spectrum of 2 was obtained in diethyl ether. In the Vis-NIR region, the yellow- brown solution showed a broad band at 752 nm with a molar extinction coefficient of ε = 605 cm- 1M-1 which is attributed to a charge transfer band (See experimental for more details). Complex 2 also shows two sharp and intense bands in the UV region at 244 nm (ε = 7300 cm-1M-1) and 297 nm (ε = 1040 cm-1M-1). These are also due to charge transfer transitions. Similar absorptions were 276 observed in the previously reported U(NHAr*)2 complex.14 The thermal stability of complex 2 was monitored by UV-Vis spectroscopy at room temperature in diethyl ether. After 100 hours of sample monitoring only 15% decomposition was observed at room temperature (more details are available in experimental section). The radical nature of the metal center in Y(NHAr*)2 (2) was confirmed by EPR spectroscopy, in collaboration with Prof. McCracken at Michigan State University. The continuous wave EPR spectrum of 2 in frozen diethyl ether shows an axial lineshape with hyperfine splitting that arises from an I = 1/2 nucleus, consistent with an 89Y-centered paramagnet. Analysis of this spectrum shows that it can be well-simulated (Figure 5.6, top, red trace) using an axial g-tensor with g‖ = 1.985 and g⊥ = 2.004, an axial 89Y hyperfine coupling with A‖ = 38.9 MHz and A⊥ = 40.9 MHz, and an intrinsic Gaussian lineshape with an average width of 1.06 mT (FWHM). The solution spectrum of 2 in diethyl ether (Figure 5.6, bottom, black trace) features a broad intrinsic lineshape with a modest inflection, also consistent with hyperfine splitting from an I = 1/2 nucleus. This spectrum was simulated using both fast- and slow-motion calculation models (Figure 5.6, bottom, red trace). The results of both approaches yield isotropic g- and 89Y-hyperfine coupling values of 1.995 and 49 MHz, respectively.21 For these simulations, a broad Lorentzian intrinsic lineshape of 2.5 mT was required, casting greater uncertainty on the spin Hamiltonian values obtained from the solution spectrum. 277 Figure 5.6. EPR spectra of Y(NHAr*)2 (2) in Et2O at 60 K (top) and 295 K (bottom). The black traces are the experimental spectra collected using the following conditions: microwave frequency, 9.3972 GHz (top), 9.3270 GHz (bottom); microwave power, 5 mW (top), 0.125 mW (bottom); field modulation amplitude, 0.1 mT (top), 0.4 mT (bottom). The red traces are spectra simulations offset for visualization purposes. Simulations were done with EasySpin using the spin Hamiltonian parameters given in the text. The spin Hamiltonian values obtained from analysis of the solid-state spectrum of 2 are most similar to those reported for the Y(C5H4SiMe3)3 − complexes reported by Evans and coworkers.12 These complexes also showed a modest g-anisotropy, g‖ = 2.00 and g⊥ = 1.99, but with g‖ > g⊥. For 2, our analysis shows g‖ < g⊥, most likely reflecting the different coordination geometry of the ligands around the metal center. Almost all reported examples of Y(II) complexes have trigonal geometry around the metal center (Figure 5.2) whereas complex 2 shows a slightly 278 bent sandwich geometry around the metal complex. For both 2 and Y(C5H4SiMe3)3 −, the presence of at least one g-value below 2.0023 and a resolved I = 1/2 isotropic hyperfine coupling is indicative of a d1-centered paramagnet. The isotropic hyperfine coupling attributed to 89Y of 2 is 40% of that reported for Y(C5H4SiMe3)3 −, consistent with more of the unpaired spin density being delocalized onto the NHAr* ligands. The spectra shown in Figure 5.6 lacks 14N hyperfine coupling in both liquid and solid-state EPR spectra suggesting the electron density residing only between the arene–Y–arene system (vertical plane of the molecule as drawn). Still, it is worth noting that in single crystal EPR studies of Y(II) in SrCl2, 89Y isotropic hyperfine couplings of 80.8 MHz were resolved.22 DFT calculations were performed to understand the bonding interactions between yttrium and ligand system in complex 2. The DFT calculations are consistent with the experimentally obtained structure (details are provided in the experimental section below). The highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital LUMO (Figure 5.7, top) of Y(NHAr*)2 (2) involved yttrium and arene rings indicating the movement of unpaired electron density throughout the metal-ligand backbone in the vertical plane of the molecule. Further, Mulliken spin density (Figure 5.7, bottom) confirmed the delocalized electron density on the ligand consistent with hyperfine splitting obtained from EPR spectroscopy. The isotopic hyperfine splitting constant Aiso (49 MHz) is significantly smaller than previously reported yttrium radical complexes suggesting that some of the electron density is delocalized on the NHAr* ligand backbone. Also, it can be observed this there is hardly any involvement of nitrogen atoms in the HOMO and LUMO thus justifying the absence of nitrogen hyperfine splitting in the EPR spectrum. 279 Figure 5.7. HOMO (top, left), LUMO (top, right), and spin density map (bottom) of Y(NHAr*)2 (2). Light blue, blue, grey, and white spheres represent yttrium, nitrogen, carbon, and hydrogen atoms, respectively. Solution-state magnetic susceptibility studies on 2 were carried out using the Evans method. At room temperature, the χMT value of 2 is 0.71 cm3 K mol-1 (μeff = 2.39 µB), which decreased to 0.57 cm3K mol-1 (μeff = 2.14 µB) at 183 K as expected by the Curie law. These values are slightly higher than the spin-only magnetic moment for one electron system (μs.o. = 1.73 µB) which can be attributed to the unquenched orbital angular momentum contribution to the magnetic moment. SQUID Magnetometry data collection and analysis was done with Florian Benner from Demir group at MSU. Complex 2 was also examined in the solid-state using SQUID magnetometry 280 (Figure 5.8). The χMT vs. T fit gave a slightly lower g value than the expected g value for an unpaired electron that is unaffected by spin-orbit coupling. Therefore, the isothermal field- dependent magnetization (M vs. H) data was collected between the temperature range of 2 and 10 K and at a field up to 7 T. The experiment data was analyzed using Brillouin functions, resulting in g values close to the anticipated range (1.9938(19)-2.3163), showing strong consistency with those obtained through EPR spectroscopy. The χMT value on the solid sample has a higher magnetic moment μeff = 2.14 µB than the relative spin-only value of an unpaired electron, attributed to TIP contribution. The magnetic moment values obtained in the solid and solution states are in agreement with each other. Figure 5.8. Temperature dependence of solution state effective magnetic moment for complex 2 (Evans method). 281 Figure 5.9. (top, left) Variable-temperature DC magnetic susceptibility data for a restrained polycrystalline sample of Y(NHAr*)2, 2, collected under 0.1 T, 0.5 T, and 1.0 T applied DC fields and at temperatures from 2 to 300 K. T (middle right) Corresponding Curie-Weiss plots (1/M vs. T) for 2 at 0.1 T, 0.5 T, and 1.0. (bottom left Explementary fits for M T vs. T) and 1/M vs. T (bottom right) plots at 0.5 T. Parameters for the M T vs. T fit: g = 1.7334(14), TIP = 9.586(51) x 10−4 and g = 1.7521(19), TIP = 9.297(45) x 10−4, zJ′ = −0.1195(102) cm−1. Parameters for the 1/M vs. T fit: C = 0.338(22) cm3•K mol−1,Θ = −1.298 K. 282 5.5 Reactivity of Y(NHAr*)2 (2) and Discovery of Yttrium Isocyanide Y(NHAr*)–NC (3) Relatively stable Y(NHAr*)2 (2) gave us an advantage in testing reactivity over similar complexes. The high-energy metallaradical was expected to react with tBuNC to liberate a tert- butyl radical which would disproportionate to give isobutene and butane gaseous byproducts. However, instead of producing more common metal cyanide (M–CN), an isocyanide (M–NC) complex was isolated. Irrespective of the reagent tBuNC or tBuCN, the same complex Y(NHAr*)– NC (3) was prepared (Figure 5.10). Figure 5.10. (Top) Synthesis of Y(NHAr*)2–NC (3), (bottom) structure of 3 from single-crystal X-ray diffraction (thermal ellipsoids drawn at 50%). Hydrogen atoms on carbon are removed for clarity. 283 Figure 5.10 (cont’d) To the best of our knowledge, this is the first structurally characterized yttrium isocyanide complex. In the diverse realm of isonitrile reactions with metal complexes, there are few examples known where nitriles and isonitriles have been employed to attach CN ligands to the f-block elements.23–26 Carmona, Andersen, and coworkers have isolated and characterized a uranium cyanide complex by using tris(cyclopentadienyl)uranium(IV) complex and isonitrile.23 Ephritikhine, Boucekkine, and Berthet have discussed different coordination modes of the CN (through N or C) based on the metals Ce and U in tris(hexamethyldisilazane) complexes.27 They have also extensively studied the effect of the ligand system on uranium towards the coordination mode of the CN moiety.28 Arnold and Maron have reported a thorium NHC–bpy complex where they have shown reductive cleavage of R–NC bond to make a thorium isocyanide complex.24 The structure from single crystal X-ray diffraction adopts a similar geometry around the metal center as Y(NHAr*)2Cl (2), where only one of the ortho-aryl rings is coordinated to the metal center 284 in a 6-binding mode. The Y1–N1 (2.348(2) Å) bond length is longer than Y1–N2 (2.221(1) Å), and Y1–N3 (2.221(1)Å) because the lone pair involved in this bonding has less s-character in the σ−bond with the metal. In 3, yttrium is coordinated to three electronegative nitrogen atoms rendering the metal more electropositive than in 1. This leads to an increase in Y1–Ar1 distance in 3 (2.518(2) Å) as compared to 1 (2.495(1) Å), attributed to a somewhat reduced backbonding effect to the arene system. After the surprising discovery of complex 3, to further solidify our findings and confirm their accuracy additional data is imperative. First, from single crystal X-ray diffraction complex 3 was refined with both the possibility of cyanide and isocyanide. The statistics for Y–NC structure R1 = 4.11% and wR2 = 9.53% are better than Y–CN structure R1 = 4.37% and wR2 = 10.61%. There is a slight but obvious variation in the refinement values due to the closely matched electron density of carbon and nitrogen. Further Y–NC structure gave us more realistic thermal displacement values (see details in the experimental section). Secondly, NMR spectroscopy proved to be valuable in determining the structure of 3. 1H and 13C NMR spectroscopy were consistent with the structure shown in Figure 5.11. The N–H protons have moved downfield 4.51 ppm in 1 to 4.65 ppm in 3 as expected by replacing the chloride with isocyanide on the metal center. Whereas the protons from the aromatic backbone were less affected and showed only a slight change in chemical shift. 285 Figure 5.11. (top) A portion of the 13C NMR spectrum for complex 3 shows a doublet peak due to yttrium coupling to isocyanide carbon. (bottom) A few examples of yttrium alkynyl complexes with the Y–C coupling constant to α-carbon and β-carbon. The absence of observable couplings between the proton and carbon in the CN ligand suggests that a “Y–NCH” structure is unlikely. Yttrium 89Y has a 100% natural abundance with I = ½. The 89Y–13C coupling constant for 3 was found to be 9.8 Hz in the 13C proton decoupled NMR spectrum. There are several examples known of yttrium alkynyl complexes that can be used to relate the coupling constant to α− and β− carbon atoms for a Y–C≡C–R moiety (some examples are shown in Figure 5.11).29–34 After conducting a thorough review of the literature, we can summarize that the typical range for a one-bond coupling constant to the α-carbon in yttrium alkynyl complexes is 286 50-75 Hz. The range for two bond coupling constants to the β-carbon of an alkynyl complex is 5- 15 Hz. For complex 3, the coupling constant is 9.8 Hz which is right in the middle of the range for two bond coupling constants to the β-carbon in yttrium alkynyl complexes. This piece of evidence suffices to substantiate that 3 is indeed an yttrium isocyanide complex, and not an yttrium cyanide complex. The 89Y NMR chemical shifts are generally observed over a wide range, such as for the metal centers in Y(C5H4Me)3(THF), (CH3C5H4)2YCl(THF) and Y{N(SiMe3)2}3 were found to be at −371(THF), −103(THF), and +570(CDCl3) ppm, respectively.35 The 89Y NMR spectra of 1 and 3 were collected in toluene-d8 whereas NMR of 2 was inconclusive due to the paramagnetic yttrium center. The 89Y NMR spectrum of complex Y(NHAr*)2Cl (1) exhibits its chemical shift at +427.7 ppm, and in Y(NHAr*)2–NC (3) the resonance is observed at +350.0 ppm. The higher chemical shift in 3 can be attributed to the shielding effect exerted on the yttrium by the two perpendicular π-electron clouds from the adjacent NC. In addition to the experimental findings, DFT calculations were performed to understand the preference for nitrogen binding over carbon to the yttrium center. Geometry optimization and frequency calculations showed that the ground state energy of the yttrium isocyanide is 4 kcal mol-1 more stable than the yttrium cyanide system. (More details in experimental section) Infrared spectroscopy was utilized to analyze the stretching frequency of the N≡C bond in 3. An intense band at 2053 cm-1 was observed in n-hexane at room temperature. For comparisons the CN bond stretched in starting materials CNtBu and NCtBu appears at 2025 and 2250 cm-1, respectively. The yttrium isocyanide bond stretch in 3 is quite similar to other known metal isocyanides.24,28 For instance, the thorium-NHC-bpy complex reported by Arnold and coworkers’ showed a metal isocyanide stretching band at 2046 cm-1, is comparable to similar features in Ephiritikhine and 287 coworkers’ U{N(SiMe3)2}3–NC complex at 2044 cm-1 and the NEt4 + U{N(SiMe3)2}3–NC salt at 2058 cm-1. Due to the lone pair bond weakening effect, the stretching frequency for CN doesn’t change significantly compared to free CN− isocyanide systems (The cyanide stretching band in the salt NBu4 +CN− occurs at 2050 cm-1).36 Thus there is a 100 cm-1 increase in stretching frequency when a metal is bound through the carbon in cyanide. In the cyanide ion (CN−), both carbon and nitrogen atoms possess a lone pair of electrons. According to Bent’s rule, a lone pair necessitates more s-character, while the only other orbital utilizing s-character in the ligand would be a C–N σ- bond.37,38 However, due to higher electronegativity of nitrogen, it can support the lone pair with less s-character compared to carbon. Natural Bond Orbital (NBO) calculations (M06, def2tzvp, NBO7) on cyanide ion confirm this, suggesting that the nitrogen’s and carbon’s lone pairs are held by sp1.1 and sp0.55 hybrids respectively. The C–N σ-bond consists of overlapping carbon-based sp1.5 and nitrogen-based sp0.9 orbitals. The hybridization difference between the lone pair and the σ- bonding orbital on the nitrogen atom is minimal (sp1.1 vs. sp0.9), while there is a considerable difference for the carbon atom (sp0.55 vs. sp1.5). Consequently, little additional C–N bonding occurs when a metal binds to the nitrogen atom, causing little change in CN stretching frequency of the isocyanide compared to free CN− ion (assuming no significant backbonding). However, when a metal is attached to the carbon atom, the CN bond strengthens significantly causing an increase in stretching frequency over the free cyanide ion. This is a common phenomenon in CO chemistry;39– 43 therefore, the slightly lower frequency of the isocyanide stretching band is likely due to the difference in lone pair weakening in the two bonding modes of the ambidentate ligand, and not the backbonding. 288 Figure 5.12. (top) Natural population Analysis gives a 75:25 charge distribution between N and C respectively. (middle) Carbon-based HOMO and LUMO from MO analysis. (bottom) Plot of enthalpy difference between the cyanide (e.g., HCN) and isocyanide (e.g., HNC) vs. the percentage of ionic bonding from NRT. While rare, there are well-characterized examples of metal isocyanides from the main group and f-elements.24–26,28,44,45 The scarcity of metal isocyanides over cyanides can be well explained by molecular orbital analysis. Both HOMO and LUMO are predominantly located on carbon (shown in the orbital picture Figure 5.12, M06/def2tzvp) therefore it is surprising that a stable molecule is formed by binding through N. Natural Population Analysis (NRT) shows that 75% of the ionic charge is situated on nitrogen in CN as expected of a more electronegative N atom. The correlation 289 between enthalpic favorability of the ground state of cyanide (X–CN)42,43,46–48 vs. isocyanide (X– NC) was correlated with %ionic character calculated using NRT (Here X–CN = MeCN, HCN, (F3C)3BCN−, Me3SiCN, and LiCN). It can be seen that a covalent system like MeCN prefers to bind with the carbon atom of CN whereas ionic systems like LiCN are inclined to N binding in CN. Harder and coworkers have extensively studied the Mg(dipyrrolylmethane)isocyanide system where they have shown that increased covalency in X–CN leads to binding through carbon whereas ionic systems prefer to bind through more electronegative nitrogen atoms.42,43,46–48 This is consistent with our finding where Y(NHAr*)2–NC (3) has the yttrium center in a +3 oxidation state, which is more of an ionic system therefore nitrogen binding is preferred over carbon with ambidentate ligand. 5.6 Conclusions In this study, we were able to prepare a Y(NHAr*)2Cl(1) complex, a half sandwich complex that upon reduction with KC8 afforded a thermally stable Y(II) complex. From single crystal X-ray diffraction, 2 has two 6-interactions with the ortho phenyl rings of the terphenyl amide ligand as compared to 1 where only one ring was 6-coordinated to the metal center. In comparison to other reported Y(II) complexes, 2 exhibits remarkable stability at room temperature. By UV-Vis spectroscopy, only 15% decomposition was observed in the sample after 100 hours. EPR spectroscopy and DFT suggested the unpaired electron density is delocalized on the ligand system in the π*arene−Y−π*arene direction along with a significant yttrium character. Cyclic voltammetry of 2 shows a reversible feature for Y(NHAr*)2 +/0 couple at −1.16  0.01 V with respect to FeCp2 +/0 in diethyl ether. The highly reactive metal radical 2 reacts with tBuNC and tBuCN to give a rare metal isocyanide Y(NHAr*)2−NC (3) complex after the loss of the tert-butyl radical. The crystal structure of 3 290 showed similar half sandwich coordination around the metal center as did the Y(NHAr*)2Cl(1) complex. Further by single crystal X-ray diffraction, the refinement and thermal parameters for yttrium isocyanide are better than yttrium cyanide complex. The 89Y−13C coupling constant in 13C NMR is consistent with two-bond coupling when compared with yttrium alkynyl complexes. The IR stretching frequency of isocyanide in 3 is consistent with previously reported uranium and thorium isocyanide complexes. Furthermore, DFT calculations also supported the preference for nitrogen binding to the metal center rather than carbon binding. The ground state structure of yttrium isocyanide is thermodynamically more favorable than the hypothetical yttrium cyanide complex by 4 kcal/mol. The terphenyl amide ligand system used in this work provides both steric and electronic stabilization to the yttrium in high and low oxidation states. Further studies may show if the amide hydrogen can be deprotonated in order to make yttrium imide complexes. Similarly, this chemistry can be extended to the lanthanide metals which will be discussed in the following chapters. 291 5.7 Experimental details All manipulations were performed under purified nitrogen atmosphere in either a glove box or using Schlenk techniques. Hexamethyldisiloxane (HDMSO) and n-hexane were dried with CaH2, distilled under nitrogen to remove oxygen and stored over 4 Å molecular sieves. Tetrahydrofuran was dried with Na/benzophenone, distilled under nitrogen to remove oxygen, and stored over 4 Å molecular sieves. 2,6-dichloroiodobenzene was purchased from Oakwood and, after freeze-pump- thawing three times to remove dissolved gases, was transferred into the glovebox. tert-Butyl isocyanide and trimethylacetonitrile were purchased from Sigma-Aldrich and, after freeze-pump- thawing three times to remove dissolved gases, were passed through activated alumina before use. YCl3 and trimethylsilylmethyllithium solution in pentane (0.1 M) were purchased from Sigma- Aldrich and used as received. Trimethylsilylmethylpotassium and tosyl azide were synthesized according to the literature procedure.18,49 H2NAr* was also synthesized according to the literature procedure,16,17 and 1H NMR spectra matched with reported data.15 NMR solvents C6D6 and C7D8 were purchased from Cambridge Isotope Laboratories, Inc. C6D6 and C7D8 are distilled over CaH2 and passed through activated alumina to ensure drying. Both NMR solvents were stored under an inert atmosphere. The NMR spectra were taken on Varian or Bruker instruments located in the Max T. Rogers Instrumentation facility at Michigan State University. 1H and 13C NMR spectra were recorded on a Bruker Avance Neo 600 MHz spectrometer. 89Y NMR spectra were collected on Bruker Avance III HD 500 MHz spectrometer operating at 24.5 MHz for 89Y. NMR chemical shifts are reported in ppm and reference to the solvent peaks for 1H NMR (C6D6, δ 7.16 ppm; C7D8, δ 2.08, 6.97, 7.01, 7.09 ppm), 13C NMR (C6D6, δ 128.06 ppm; C7D8, δ 20.43, 125.13, 127.96 ppm). 89Y NMR data was referenced using vendor supplied method of indirect referencing (Bruker TopSpin 3.6.2) relative to the lock solvent and aqueous Y(NO3)3. This 292 method is based on IUPAC recommendations for indirect referencing.50 Referencing derived from the 1D NMR data were used to calibrate the 2D 1H-89Y HMBC experiments. Single crystal data was collected on XtaLAB Synergy, Dualflex, Hypix diffractometer using CuKα radiation. Data collection was done at 100 K under a continuous flow of liquid nitrogen. In Olex2 program, crystal structures were solved with ShelXT solution using intrinsic phasing and refined with the SheXL refinement package using least squares minimization.51,52 All hydrogens are refined anisotropically. All crystals were stable at room temperature for mounting. Experimental details for IR spectroscopy, EPR Spectroscopy, and Cyclic Voltammetry are mentioned in their respective sections. Synthesis of Complexes Synthesis of KNHAr* A 20 mL scintillation vial charged with a stir bar was loaded with 2,6-Tripp2C6H3NH2 (500.3 mg, 1.0 mmol, 1.0 equiv.) and n-hexane (15 mL). This solution was kept in the freezer for 15 min. This solution was kept on a stir plate and then trimethylsilylmethyl potassium (133.3 mg, 1.0 mmol, 1.05 equiv.) was added slowly. The solution was left to stir for 12 h at room temperature. The reaction mixture was filtered (using a fritted funnel), washed with cold n-hexane, and dried under reduced pressure. The pure product was obtained as a colorless solid (433.0 mg, 0.80 mmol, 80% yield). 1H NMR (500 MHz, C6D6) δ 7.12 (s, 4H), 7.04 (d, J = 7.1 Hz, 2H), 6.56 (t, J = 7.1 Hz, 1H), 3.52-3.38 (m, 4H), 2.88-2.73 (m, 2H), 2.62 (s, 1H), 1.27 (t, J = 7.3 Hz, 24H), 1.23 (d, J = 6.8 Hz, 12H). The compound should not be recrystallized as this leads to contamination with H2NAr* and formation of an insoluble solid, which is likely K2NAr*. 293 Synthesis of Y(NHAr*)2Cl (1) A 20 mL scintillation vial charged with a stir bar was loaded with YCl3 (27.30 mg, 0.14 mmol, 1.0 equiv.) and Et2O (5 mL). A separate 20 mL scintillation vial was loaded with KNHAr* (150.0 mg, 0.28 mmol, 2.0 equiv.) and Et2O (5 mL). Both solutions were cooled in a dry ice/acetone cold well for 20 minutes and then the YCl3 solution was suspended above a magnetic stir bar. When the solution had thawed enough to stir, a cold solution of KNHAr* was added dropwise. The solution was left to stir for 12 h at room temperature. The reaction color changed from pale yellow to dark yellow over the period of 12 h. The volatiles were removed in vacuo. The resulting yellow solid was extracted with n-hexane and the solvent was removed in vacuo. A concentrated solution in n- hexane resulted in the formation of yellow-colored X-ray quality single crystals overnight at –35 °C in the freezer (100.0 mg, 0.090 mmol, 64% yield). Anal. Calcd for C72H100N2YCl: C, 77.35; H, 9.01; N, 2.50. Found: C, 77.07; H, 9.46; N, 2.46. 1H NMR (600 MHz, C7D8) δ 7.25 (s, 8H), 6.92 (d, J = 7.4 Hz, 4H), 6.67 (t, J = 7.4 Hz, 2H), 4.51 (s, 2H), 2.97 (hept, J = 6.8 Hz, 8H), 2.82 (p, J = 6.9 Hz, 4H), 1.32 (d, J = 6.9 Hz, 24H), 1.28 (d, J = 7.0 Hz, 24H), 1.10 (d, J = 6.8 Hz, 24H). 13C NMR (151 MHz, C7D8) δ 156.54, 156.51, 149.47, 148.80, 137.45, 130.67, 129.00, 128.84, 128.68, 128.09, 127.92, 127.77, 126.14, 125.26, 125.10, 124.93, 123.49, 120.45, 114.65, 34.48, 31.12, 25.51, 24.56, 24.29, 23.91, 20.78, 20.65, 20.40, 20.15, 20.02. 89Y NMR (25 MHz, C7D8) δ 427.73. Decomposition temperature: 190 °C. Synthesis of Y(NHAr*)2 (2) A 20 mL scintillation vial charged with crystals of 1 (60.0 mg, 0.05 mmol, 1.0 equiv.), 5 mL THF, and a magnetic stir bar. The vial was placed in a liquid nitrogen-cooled cold well until the solution froze. Once frozen, the vial was removed from the cold well and suspended above a magnetic stir plate. When the solution had thawed enough to stir, a suspension of KC8 (14.5 mg, 294 0.10 mmol, 2.0 equiv.) in THF (2 mL) was added. The solution turned color from yellow to black rapidly. The solution was stirred for 1 h at room temperature. The volatiles were then removed under reduced pressure, and the remaining residue dissolved in 3 mL of diethyl ether. The ether solution was stirred for ~5 min and filtered using Celite. The filtrate was then dried in vacuo to remove volatiles. The remaining black residue was dissolved in n-hexane and filtered using Celite twice. X-ray quality single crystals were produced by chilling a concentrated n-hexane solution of 1 in a –35 °C freezer overnight (30.7 mg, 0.02 mmol, 52% yield). Anal. Calcd for C72H100N2Y: C, 77.35; H, 9.01; N, 2.50. C72H100N2Y·C4H8O: C, 79.05; H, 9.43; N, 2.43. Found: C, 78.90; H, 9.77; N, 2.33. 1H NMR and 13C NMR spectra did not show any distinctive peaks for the paramagnetic complex. EPR (in the manuscript) and UV-vis/NIR spectra (below) were recorded. Synthesis of Y(NHAr*)2NC (3) A 20 mL scintillation vial was charged with crystals of 2 (49.5 mg, 0.04 mmol, 1.0 equiv.), 3 mL n-hexane, and a magnetic stir bar. A separate 20 mL scintillation vial was loaded with tert- butylisonitrile (3.60 mg, 0.04 mmol, 1.0 equiv.) and 2 mL n-hexane. Both solutions were cooled in a dry ice/acetone cold well for 20 minutes and then tBuNC solution was added dropwise into the solution of 2. The reaction color changed from black to pale yellow rapidly. The reaction mixture was allowed to stir for 30 minutes. The solvent was removed in vacuo and the resulting yellow solid was extracted with n-hexane and solvent was removed. A concentrated solution in n-hexane resulted in the formation of yellow-colored X-ray quality single crystals overnight at –35 °C in the freezer (31.6 mg, 0.03 mmol, 67% yield). Anal. Calcd for C73H100N3Y: C, 79.02; H, 9.09; N, 3.79. Found: C, 78.50; H, 9.28; N, 3.62.1H NMR (600 MHz, C7D8) δ 7.29 (s, 2H), 6.91 (d, J = 7.4 Hz, 2H), 6.67 (t, J = 7.4 Hz, 1H), 4.65 (d, J = 1.6 Hz, 1H), 3.05-2.86 (m, 3H), 1.36 (d, J = 7.0 Hz, 9H), 1.26 (d, J = 5.0 Hz, 1H), 1.09 (d, J = 6.8 Hz, 8H). 13C NMR (151 MHz, C7D8) δ 181.18 (d, 2JY-C = 295 9.8 Hz), 156.55 (d, 2JY-Car = 3.3 Hz), 149.31, 149.09, 137.46, 130.40, 126.15, 115.03, 34.57, 31.33, 25.07, 24.25, 24.09. 89Y NMR (25 MHz, C6D6) δ 350.28. Decomposition temperature: 184 °C. An NMR scale reaction was set up using trimethylacetonitrile as starting material instead of tert- butylisonitrile, and the same product was observed by 1H NMR labeled as 3a. 296 NMR Spectra of Complexes 1 and 3 Figure 5.13. 1H NMR spectrum of Y(NHAr*)2Cl (1) in C7D8. 297 Figure 5.14. 13C NMR spectrum of Y(NHAr*)2Cl (1) in C7D8. 298 Figure 5.15. 89Y NMR spectrum of Y(NHAr*)2Cl (1) (18 mM) in C7D8. 299 Figure 5.16. HMBC for 1H‒89Y spectrum of Y(NHAr*)2Cl (1) in C7D8. This shows a cross peak between N‒H and Y which are two-bond separated. A prominent cross peak was observed with relaxation delay of 2s and 4 scans. 300 Figure 5.17. 1H NMR spectrum of Y(NHAr*)2‒NC (3) in C7D8. 301 Figure 5.18. 1H NMR spectrum of Y(NHAr*)2‒NC (3a) in C7D8 (From tBuCN). 302 Figure 5.19. 13C NMR spectrum of Y(NHAr*)2‒NC (3) in C7D8. 303 Figure 5.20. 89Y NMR spectrum of Y(NHAr*)2‒NC (3) (12 mM) in C7D8. (This spectrum was recorded with 60 s relaxation delay and 935 scans.) 304 Figure 5.21. 89Y NMR spectrum of Y(NHAr*)2‒NC (3) (16 mM) in C6D6. (This spectrum was recorded with 10 s relaxation delay and 299 scans.) 305 Figure 5.22. HMBC for 1H‒89Y spectrum of Y(NHAr*)2‒NC (3) in C7D8. This shows a cross peak between N‒H and Y, which are two bonds away, as expected. A prominent cross peak is observed with a relaxation delay of 2 s and 4 scans. 306 Figure 5.23. HMBC for 1H‒13C spectrum of Y(NHAr*)2‒NC (3) in C7D8. The N≡C carbon does not correlate with the N‒H hydrogens, which suggests that they are at least four bonds away to not appear in this HMBC spectrum. This spectrum further supports the presence of an isocyanide rather than cyanide. If it was cyanide instead of isocyanide, a cross peak might have been observed for three bond correlation, i.e., H‒N‒Y‒CN, if the dihedral angle is not near 90°. 307 Figure 5.24. HSQC for 1H‒13C spectrum of Y(NHAr*)2‒NC (3) in C7D8. Peak picking for residual protio-toluene omitted for clarity. The N≡C carbon does not correlate with any hydrogen, which shows there is no proton on the isocyanide, i.e., Y‒NCH. 308 UV-Vis-NIR Absorption Spectrum of Y(NHAr*)2 (2) These electronic absorption spectra were recorded on a double-beam PerkinElmer 1050 spectrophotometer. The measurement was done in a 1 cm pathlength quartz cell at a ⁓1 mM concentration of 2. Preparation of samples was performed in the N2 glovebox using dry diethyl ether. The raw data were fitted with OriginPro 9.0 software to obtain accurate maxima assuming Gaussian peak shapes. All spectra were baseline corrected for diethyl ether and collected at ambient temperature. A broadband feature was observed at 752 nm with a molar extinction coefficient of 604.6 M-1 cm-1. This broadband feature can be attributed to a charge transfer band in Y(NHAr*)2 2. We were unable to locate d-d band which could be easily hidden under this charge transfer band. A similar broad band was observed for previously reported U(NHAr*)2 14 complex at ⁓600 nm with molar extinction coefficient of ⁓1200 M-1cm-1. Figure 5.25. UV-vis spectra at 1 mM concentration of Y(NHAr*)2 2 in diethyl ether. To observe charge transfer bands low concentrations of 2 were monitored. UV-Vis spectra were 309 collected using an Ocean Optics DH-mini UV-Vis spectrophotometer in an N2 glovebox. In order to obtain molar extinction coefficients for charge transfer bands, multiple concentrations of 125 μM, 100 μM, 75 μM, 50 μM, and 25 μM solutions were prepared after dilutions. Figure 5.26. UV-vis spectra at low concentrations of Y(NHAr*)2 2 in diethyl ether. Two absorption bands were observed between 220‒350 nm. Two peak maxima are at 244 nm and 297 nm with ε = 7.3 x 103 M-1 cm-1 and 1.04 x 104 M-1cm-1, respectively. The transition in this high-energy region is mostly due to metal-to-ligand charge transfer. 310 Figure 5.27. Concentration vs absorbance plot to calculate molar extinction coefficient at 297 nm for Y(NHAr*)2 2 in diethyl ether. The slope of the line provides ε = 1.04 x 104 M-1cm-1 for the electronic transition occurring at 297 nm. Thermal Stability of Y(NHAr*)2 (2) The thermal stability of complex 2 at room temperature was monitored by UV-vis spectroscopy. These electronic absorption spectra were recorded on a double-beam PerkinElmer 1050 spectrophotometer. The measurement was done in a 1 cm pathlength Teflon-capped quartz cell. Preparation of samples was performed in the N2 glovebox using dry diethyl ether. The raw data were fit with OriginPro 9.0 software to obtain accurate maxima assuming Gaussian peak shapes. All spectra were baseline corrected for diethyl ether and collected at ambient temperature for up to 2 weeks. As shown in Figure 5.28 (top), the decay in absorption peak at 294 nm shows a slow decomposition of the sample over time (100 h). After 100 hours, there was a slight increase in the concentration, which we believe is due to a leak in the Schlenk tube over the prolonged experiment 311 time. The initial concentration was ⁓110 μM which was decreased to ⁓93 μM (85% of the initial concentration) after approximately 100 hours. The first-order plot between concentration and time does not show any correlation between the decomposition rate of complex 2 at room temperature (Figure 5.28, bottom). On average, 3.8% of the sample was decomposed per day which is very impressive considering the high reactivity of complex 2. Figure 5.28. (top) Overlay of the absorption spectrum of Y(NHAr*)2 over time. (bottom) The logarithm of yttrium complex 2 concentration vs time plot. 312 Figure 5.28 (cont’d) Solution-State Magnetism (Evans’ Method) Solution-state magnetic susceptibility studies were carried out on 2 using the Evans method. The resulting temperature-dependent paramagnetism was measured in 10 K intervals from 298-183 K. In a Teflon capped J-Young NMR tube 20 mM sample of 2 in toluene-d8 and hexamethyldisiloxane was transferred with a sealed capillary tube containing 20 mM hexamethyldisiloxane in toluene-d8 as a reference. The difference between HMDSO peak value with and without paramagnetic species was analyzed to measure effective magnetic moment of 2.53 At room temperature the MT value of 2 is 0.71 cm3 K mol-1 (μeff = 2.39), which decreased to 0.57 cm3 K mol-1 (μeff = 2.14) at 183 K as expected by the Curie law. These values are slightly higher than the spin-only magnetic moment for one electron system (μs.o. = 1.73) which can be attributed to unquenched orbital angular momentum contribution to the magnetic moment. Data points were corrected for diamagnetic contributions using Pascal’s constant.53 313 Solid-State Magnetism (SQUID) “The data collection and analysis were done by Florian Benner from Demir group” Direct current (dc) magnetic susceptibility measurements were carried out on a polycrystalline sample of 2 via SQUID magnetometry in applied dc magnetic fields of 0.1 T, 0.5 T, and 1.0 T between 2 and 300 K, Figure 5.9. The temperature dependence of the MT product exhibits a gradual downturn in MT with decreasing temperature indicating a deviation from the Curie-Weiss behavior expected for an ideal paramagnet. Such deviations can be a result of a) thermal depopulation of low-lying excited states, b) intermolecular through-space coupling or c) temperature-independent paramagnetism. To explain this variance from ideal behavior, the MT vs. T curves were fit by freely refining the g value taking into account a weak intermolecular coupling term (zJ′) alone. However, these fits yielded unrealistic high intermolecular coupling constants and precluded a satisfactorily description of the experimental values. Instead, the gradual decline in MT required the consideration of a small contribution from temperature-independent paramagnetism (TIP, 10.400(42) – 9.212(60) x 10−4 cm3 mol−1 (0.1 - 1.0 T)), affording a satisfying agreement with the experimental values. The inclusion of TIP and zJ′ contributions gave rise to small antiferromagnetic zJ′ values (−0.0890(91) - −0.1651(104) cm−1) while the fitted g- and TIP-values remained largely unaltered. This weak intermolecular antiferromagnetic coupling is also reflected in the fitted values of the Curie-Weiss plots (1/M vs. T) below 70 K, where small negative Weiss constants were found (Θ = −1.3141 - −1.2980 K). Notably, the 1/M vs. T plots deviate from ideal Curie-Weiss behavior, indicated by the substantial positive curvatures at temperatures > 60 K for all applied dc fields. This is largely attributed to the TIP contribution apparent in the MT vs. T curves. The room temperature MT value of 0.534 cm3 K mol−1 at 1.0 T (0.581 cm3 K mol−1 at 0.1 T, and 0.549 cm3 K mol−1 at 0.5 T) is higher than the expected value of 0.375 cm3 K mol−1 for the 314 corresponding free 4d15s0 Y(II) ion. Such discrepancy between experimental and theoretical value is not uncommon and slightly higher room temperature MT values were reported for other Ln(II) complexes such as [K(crypt-222)][Cp′3Ln].54 The deviation is considerably lower for 2 and closer to the expected value when subtracting the fitted temperature-independent contributions for each field, resulting in room temperature MT values of 0.318, 0.315 and 0.312 cm3 K mol−1 for 0.1 T, 0.5 T and 1.0 T. The fits of the MT vs. T data (Figure 5.29) engendered slightly lower g values than the expected g value of 2.0023 for an unpaired electron that is unaffected by spin-orbit coupling. Hence, the isothermal field-dependent magnetization (M vs. H) data were collected between 2 and 10 K up and at fields up to 7 T. The resulting experimental data was fit to a set of Brillouin functions to afford g values near the expected value (1.9938(19) - 2.3163(36)), which are in excellent agreement with the values attained from EPR spectroscopy. The magnetic properties of 2 were also probed through measuring a toluene solution of 2 employing Evans’ method between temperatures of 183 and 298 K, Figure 5.8. Similar to the determined MT values on the solid sample, a higher magnetic moment meff = 2.39 μB was obtained relative to the spin-only value of 1.73 μB for an unpaired electron, which is likely ascribed to the aforementioned TIP contribution. 315 Figure 5.29. Temperature dependence of the product of magnetic susceptibility and temperature, χMT, for a restrained polycrystalline sample of Y(NHAr*)2, 2, with fits to g and TIP, and under consideration of additional zJ′ terms, collected under 0.1 T (left) and 1.0 T (right) applied dc fields and at temperatures between 2 to 300 K. Fit parameters 0.1 T: g = 1.7483(12), TIP = 10.4(42) x 10−4, residue = 14.1 x 10−4; g = 1.7627(18), TIP = 10.400(42) x 10−4, zJ′ = −0.0890(91) cm−1 (0.1 T), residue: 8.6 x 10−4. Fit parameters 1.0 T g = 1.7236(17), TIP = 9.212(60) x 10−4, residue: 26.3 x 10−4; g = 1.7491(19), TIP = 8.819(44) x 10−4, zJ′ = −0.1651(104) cm−1, residue: 9.6 x 10−4. 316 Figure 5.30. (top) Plots of the parameters employed to fit the MT vs. T of Y(NHAr*)2, 2: the g and TIP values are plotted against the field, and (right) the zJ′ and fit residues are plotted against the field. In both plots, the red symbols represent the fit parameters without a zJ′ contribution and the blue symbols consider a zJ′ term. 317 Figure 5.31. (top) Curie-Weiss plots of the inverse magnetic susceptibility (1/M) versus temperature (T) for Y(NHAr*)2, collected under 0.1 T, 0.5 T and 1.0 T applied dc fields and at temperatures between 2 and 300 K: superimposed spectra and corresponding fits, and (bottom left) 1/M vs. T plots at 0.1 T and (bottom right) 1.0 T. Fit parameters 0.1 T: C = 0.345(22) cm3 K mol−1, Θ = −1.314 K. Fit parameters 1.0 T: C = 0.334(22) cm3 K mol−1, Θ = −1.298 K. 318 Figure 5.32. (left) Variable-temperature dc magnetic susceptibility data for a restrained polycrystalline sample of Y(NHAr*)2, 2, collected under 0.1 T, 0.5 T and 1.0 T applied dc fields and at temperatures from 2 to 300 K and (right) MT vs. T curves with subtracted TIP contributions. Figure 5.33. Variable temperature M(H) curves for 2 collected from 0 to 7 T. The black lines represent fits to the Brillouin function for each temperature. Fit parameters: 2 K: g = 1.9938(19), N = 0.6087(8); 4 K: g = 2.0491(8), N = 0.5804(4); 6 K: g = 2.1229(14), N = 0.5417(6); 8 K: g = 2.2117(23), N = 0.4987(9); 10 K: g = 2.3163(36), N = 0.4534(13). 319 FT-IR Spectroscopy IR Spectra were recorded using Varian 3100 FT-IR spectrometer with a spectral resolution of 2 cm-1 using a CaF2 air-free cell. Sample preparation was done in the glove box under an N2 atmosphere and sealed with Teflon caps before measurements. All spectra were recorded in the solution phase using dry n-hexane as solvent. The IR transmittance spectra were collected at 298 K and baseline corrected. The νCN stretch for tert-butyl isocyanide is 2127 cm-1 in n-hexane matches with literature reported value.23 Since complex 3 can also be synthesized from trimethylacetonitrile, we recoded its IR spectrum (2237 cm-1) for comparison. The νCN stretch for complex 3 is 2053 cm- 1 is lower in energy than both starting materials and is also in agreement with previously reported metal isocyanides. Figure 5.34. IR spectrum for trimethylacetonitrile in n-hexane. 320 Figure 5.35. IR spectrum for complex Y(NHAr*)2–NC 3 (⁓5 μM) in n-hexane. EPR Spectroscopy EPR spectra were collected on a Bruker E-680X spectrometer operating at X-band and fitted with a Bruker SHQ-E cavity. For experiments done over the range of 130 – 330 K, a Bruker B-VT- 2000 temperature control system was used. Experiments done at 20 – 80 K made use of an Oxford ESR-900 cryostat and an ITC – 301 temperature controller. The magnetic field was calibrated using weak pitch as standard and the microwave frequency was monitored with an EIP – 25B counter. Spectral simulations were done using EasySpin 5.2.3521 and fit using the “fminsearch” function of MATLAB R2020A. The quality of the fits was judged by calculating a normalized 2 value based on using 1% of the maximum spectral amplitude as the standard deviation. For the fits, the normalized 2 values were 0.9 (60 K) and 0.4 (295 K). Samples were prepared for EPR spectroscopy in the dry box using dry degassed diethyl ether. The samples were loaded as a solution into a quartz EPR tube and sealed under vacuum. 321 Cyclic Voltammetry Cyclic voltammetry was recorded using a PGSTAT204 potentiostat from Metrohm with a glassy carbon working electrode, platinum wire pseudo reference electrode, and platinum wire as a counter electrode. All measurements were in the glovebox under argon atmosphere. Y(NHAr*)2 2 (1.5 mM) was dissolved in diethyl ether with [NBu4]+[B(3,5-(CF3)2C6H3)4]– (100 mM) as supporting electrolyte. Cyclic voltammetry of ferrocene was performed six times to get standard deviation in potential shifts. The reduction potential for Fc+/0 redox couple was observed at 1.2 ± 0.2 V. All voltammograms were externally referenced to ferrocene with the same electrolyte concentration. All measurements were done in triplicate to get an average value of half-cell potential with scan rates of 50 mV/s and 100 mV/s (Figure 5.36). A reproducible quasi-reversible feature was observed at –1.16 ± 0.01 V vs. Fc+/Fc0 on the time scale of the electrochemical experiments assigned to the 0/+1 redox couple. Figure 5.36. Overlay plot for scan rates 100 mV/s (black) and 50 mV/s (red). 322 Table 5.1. Summarized results from scan rate measurements to study reversibility of 0/+1 redox couple. Scan 100 mV/s 50 mV/s Ec‒Ea ic/ia ip/√(scan rate) 0.22 1.09 36.8 0.27 1.08 33.1 Single Crystal X-ray Diffraction Thermal ellipsoids of all structures are drawn with a 50% probability level. Pink, green, blue, grey, and white ellipsoids represent yttrium, chlorine, nitrogen, carbon, and hydrogen atoms respectively. Figure 5.37. Structure of Y(NHAr*)2Cl (1) from X-ray diffraction recrystallized from n-hexane. There are two disordered molecules of n-hexane in a lattice per molecule of 1. 323 Table 5.3. Crystallographic data and structural refinement of Y(NHAr*)2Cl (1). Complex number Empirical formula Formula weight Temperature/K Crystal system Space group a/Å b/Å c/Å α/° β/° γ/° Volume/Å3 Z ρcalcg/cm3 μ/mm-1 F(000) Crystal size/mm3 Radiation 2Θ range for data collection/° Index ranges Reflections collected Independent reflections Data/restraints/parameters Goodness-of-fit on F2 Final R indexes [I>=2σ (I)] Final R indexes [all data] Largest diff. peak/hole / e Å-3 Y(NHAr*)2Cl (1) C84H125ClN2Y 1287.21 99.99(10) monoclinic C2/c 39.6030(5) 16.4949(2) 25.0473(3) 90 91.0170(10) 90 16359.5(3) 8 1.045 1.590 5592.0 0.196 × 0.12 × 0.101 Cu Kα (λ = 1.54178) 5.804 to 154.574 -48 ≤ h ≤ 40, -20 ≤ k ≤ 15, -28 ≤ l ≤ 31 59902 15940 [Rint = 0.0751, Rsigma = 0.0692] 15940/287/1044 1.032 R1 = 0.0598, wR2 = 0.1564 R1 = 0.0765, wR2 = 0.1727 0.80/-0.75 324 Figure 5.38. Asymmetric unit of Y(NHAr*)2 (2) recrystallized from n-hexane. 325 Table 5.4. Crystallographic data and structural refinement of Y(NHAr*)2 (2). Complex number Empirical formula Formula weight Temperature/K Crystal system Space group a/Å b/Å c/Å α/° β/° γ/° Volume/Å3 Z ρcalcg/cm3 μ/mm-1 F(000) Crystal size/mm3 Radiation 2Θ range for data collection/° Index ranges Reflections collected Independent reflections Data/restraints/parameters Goodness-of-fit on F2 Final R indexes [I>=2σ (I)] Final R indexes [all data] Largest diff. peak/hole / e Å-3 Y(NHAr*)2 (2) C76H108N2OY 1154.55 100.00(10) monoclinic C2/c 18.0163(3) 17.0386(3) 22.7773(4) 90 106.062(2) 90 6719.1(2) 4 1.141 1.539 2500.0 0.076 × 0.059 × 0.024 Cu Kα (λ = 1.54184) 7.28 to 153.926 -22 ≤ h ≤ 21, -20 ≤ k ≤ 19, -27 ≤ l ≤ 27 23782 6674 [Rint = 0.0295, Rsigma = 0.0278] 6674/198/427 1.033 R1 = 0.0417, wR2 = 0.1130 R1 = 0.0446, wR2 = 0.1150 0.59/-0.70 326 Figure 5.39. Structure of CN‒Y(NHAr*)2 (3) recrystallized from n-hexane. 327 Table 5.5. Crystallographic data and structural refinement of Y(NHAr*)2Cl (1), Y(NHAr*)2 (2), and Y(NHAr*)2‒NC (3). Complex number Empirical formula Formula weight Temperature/K Crystal system Space group a/Å b/Å c/Å α/° β/° γ/° Volume/Å3 Z ρcalcg/cm3 μ/mm-1 F(000) Crystal size/mm3 Radiation 2Θ range for data collection/° Index ranges Reflections collected Independent reflections Data/restraints/parameters Goodness-of-fit on F2 Final R indexes [I>=2σ (I)] Final R indexes [all data] Largest diff. peak/hole / e Å-3 Y(NHAr*)2NC (3) C73H100N3Y 1108.46 100.00(10) monoclinic P21/n 13.82181(13) 28.6811(3) 17.11754(15) 90 106.3762(9) 90 6510.53(11) 4 1.131 1.563 2392.0 0.087 × 0.07 × 0.039 Cu Kα (λ = 1.54184) 6.202 to 154.464 -16 ≤ h ≤ 17, -33 ≤ k ≤ 36, -21 ≤ l ≤ 11 50292 12920 [Rint = 0.0433, Rsigma = 0.0388] 12920/0/726 1.035 R1 = 0.0411, wR2 = 0.0911 R1 = 0.0508, wR2 = 0.0951 0.61/-0.55 Figure 5.39 shows the structure obtained from X-ray diffraction. Along with structure the presence of Y(NHAr*)2‒NC is reasoned by using the 89Y‒13C coupling constant in the text. We tried to solve the structure by exchanging positions of C1 to N1 and N1 to C1. Analysis of the theoretical cyanide structure gave us poorer refinement values R1 = 4.37 and wR2 = 10.61 with unrealistic non-positive definite thermal displacement parameters. 328 Table 5.6. Thermal parameters for carbon and nitrogen atoms in Y(NHAr*)2‒NC (3) vs Y(NHAr*)2‒CN crystal structure. Y(NHAr*)2‒NC Y(NHAr*)2‒CN U11 U22 U33 U11 U22 U33 C 0.03700 0.04300 0.03890 0.01068 0.00373 0.02139 N 0.02170 0.03470 0.02610 0.05490 0.02060 0.02495 Figure 5.40. Structure of Y(NHAr*)2‒CN. The CN moiety thermal ellipsoids become non-positive definite in the structure. 329 Table 5.7. Metric data from the crystal structures of Y(NHAr*)2Cl (1), Y(NHAr*)2 (2), and Y(NHAr*)2‒NC (3). All distances are in Å and angles are in (˚). Complex Y(NHAr*)2Cl (1) Y(NHAr*)2 (2) 2.249(2) 2.213(2) 2.5071(8) 131.5(2) 144.2(2) 133.76(1) 104.41(7) 107.70(1) 94.98(7) 101.99(7) 2.495(1) 2.781(3), 2.821(3), 2.871(3), 2.921(3), 2.903(3), 2.868(3) 1.416(5), 1.423(4), 1.400(5), 1.400(5), 1.386(5), 1.388(4) 2.2628(7) 130.24(13) 102.37(9) 95.91(4) 112.84(4) 2.46770(8) 133.87(1) 2.729(2), 2.787(2), 2.953(2), 2.802(2), 3.013(2), 2.766(2) 1.444(3), 1.407(3), 1.409(3), 1.392(3), 1.418(3), 1.412(3) Y‒N Y‒Cl N‒C1 Y‒N‒Cipso N‒Y‒N N‒Y‒Cl Arcent‒Y‒N Y‒N‒Cterminal Y-ArCent Cent‒Y‒Cent Y‒CAr η6-CAr‒CAr Average CAr‒ CAr bond Y(NHAr*)2NC (3) 2.2206(17) 2.2205(17) 2.3481(18) Y1‒N1 1.039(4) 145.40(15) 133.20(14) 112.33(7) N3‒Y1‒N2 106.92(6) N3‒Y1‒N1 126.54(6) N2‒Y1‒N1 93.70(5) 106.63(5) 170.38(19) 2.5178(2) 2.800(2), 2.893(2), 2.939(2), 2.958(2), 2.880(2), 2.827(2) 1.424(3), 1.413(3), 1.402(3), 1.402(3), 1.391(3), 1.403(3) 1.402(5) 1.413(3) 1.406(3) 330 DFT Calculations DFT calculations were carried out using Gaussian 16 (B01). 55 The starting coordinates for the geometry optimization were taken from the structure found from X-ray diffraction. Geometry optimization and frequency calculations were performed using the B3LYP functional and def2- SV(P)56 basis set on all atoms with the ECP28MDF57 pseudopotential on the yttrium atom with Grimme’s dispersion correction GD3.58,59 The NBO calculations were carried out using NBO7. DFT calculations were performed to understand the bonding interactions between yttrium and ligand system in complex 2. In comparison to the initial structure obtained crystallographically, the DFT calculations provided a consistent structure to experimentally obtained structure (shown in Figure 5.38 and Table 5.3). Figure 5.41. Optimized structure of Y(NHAr*)2 (2) using B3LYP/def2-SV(P). All hydrogens are removed for clarity except N‒H hydrogens. Light blue, blue, grey, and white spheres represent yttrium, nitrogen, carbon, and hydrogen atoms, respectively. 331 Table 5.8. Structural comparisons between crystallographically obtained geometry of 2 and optimized geometry using DFT calculations. Shown here are some characteristic lengths (Å) and angles (˚). Atoms Y‒N Y-ArCent Cent‒Y‒Cent Y‒N‒Cipso Arcent‒Y‒N Average CAr‒CAr bond Distance (Å)/Angle (˚) Experimental 2.2628(7) 2.46770(8) 133.87(1) 130.24(13) 95.91(4) 112.84(4) 1.413(3) Distance (Å)/Angle (˚) Calculated 2.277 2.465 137.24 130.11 95.82 110.85 1.419 The yttrium isocyanide Y(NHAr*)2‒NC (3) and hypothetical yttrium cyanide Y(NHAr*)2‒CN (3’) were also studied computationally. The structure obtained from DFT calculations (Figure 5.41) was consistent with the experimental results from X-ray diffraction (Table 5.9). The NBO analysis showed triple bonds between terminal N‒C bond with a lone pair residing on the carbon atom further proves presence of isocyanide over cyanide linkage. 332 Figure 5.42. Optimized structures of Y(NHAr*)2‒NC (3) (left) and Y(NHAr*)2‒CN (3’) (right) using B3LYP/def2-SV(P). All hydrogens are removed for clarity except N‒H hydrogens. Light blue, blue, grey, and white spheres represent yttrium, nitrogen, carbon, and hydrogen atoms, respectively. Table 5.9. Structural comparisons between crystallographically obtained geometry of Y(NHAr*)2‒NC (3) and optimized geometry using DFT calculations. Shown here are some characteristic lengths (Å) and angles (˚). Atoms Y‒N2, N3 Y-N1 Y-ArCent N1‒C1 Y‒N1‒C1 Y‒N‒Cipso Average CAr‒ CAr bond Distance (Å)/Angle (˚) Experimental 2.2206(17) 2.2205(17 2.3481(18) 2.5178(2) 1.039(4) 145.40(15) 133.20(14) 1.406(3) Distance (Å)/Angle (˚) Calculated 2.2245 2.2134 2.2785 2.603 1.179 170.33 139.85 134.65 1.412 333 Figure 5.43. NBO of N‒C bond, N‒C (σ-bond; top left), N‒C (π-bond; top right), N‒C (π-bond; bottom left), and lone pair on C (bottom right) of Y(NHAr*)2‒NC (3). Light blue, dark blue, grey, and white spheres represent yttrium, nitrogen, carbon, and hydrogen atoms, respectively. The frequency calculations were carried out to obtain thermochemical data on yttrium isocyanide Y(NHAr*)2‒NC (3) and hypothetical yttrium cyanide Y(NHAr*)2‒CN (3’) structures. The sum of electronic and thermal free energies for Y(NHAr*)2‒NC (3) and Y(NHAr*)2‒CN (3’) 334 are –3041.40663 and –3041.400367 Hartree/particle, respectively. This suggests that the Y(NHAr*)2‒NC (3) structure is thermodynamically more stable (–4 kcal/mol) than the hypothetical Y(NHAr*)2‒CN (3’) structure. 335 REFERENCES (1) MacDonald, M. R.; Ziller, J. W.; Evans, W. J. Synthesis of a Crystalline Molecular Complex of Y 2+ , [(18-Crown-6)K][(C 5 H 4 SiMe 3 ) 3 Y]. J. Am. Chem. Soc. 2011, 133, 15914–15917. (2) MacDonald, M. R.; Bates, J. E.; Fieser, M. E.; Ziller, J. W.; Furche, F.; Evans, W. J. Expanding Rare-Earth Oxidation State Chemistry to Molecular Complexes of Holmium(II) and Erbium(II). J. Am. Chem. Soc. 2012, 134, 8420–8423. (3) MacDonald, M. R.; Bates, J. E.; Ziller, J. W.; Furche, F.; Evans, W. J. Completing the Series of +2 Ions for the Lanthanide Elements: Synthesis of Molecular Complexes of Pr 2+ , Gd 2+ , Tb 2+ , and Lu 2+. J. Am. Chem. Soc. 2013, 135, 9857–9868. (4) Evans, W. J. Tutorial on the Role of Cyclopentadienyl Ligands in the Discovery of Molecular Complexes of the Rare-Earth and Actinide Metals in New Oxidation States. Organometallics 2016, 35, 3088–3100. (5) Hitchcock, P. B.; Lappert, M. F.; Maron, L.; Protchenko, A. V. Lanthanum Does Form Stable Molecular Compounds in the +2 Oxidation State. Angew. Chem. Int. Ed. 2008, 47, 1488–1491. (6) Evans, W. J.; Lee, D. S.; Ziller, J. W. Reduction of Dinitrogen to Planar Bimetallic M 2 (μ - η 2 : η 2 -N 2 ) Complexes of Y, Ho, Tm, and Lu Using the K/Ln[N(SiMe 3 ) 2 ] 3 Reduction System. J. Am. Chem. Soc. 2004, 126, 454–455. (7) Evans, W. J.; Lee, D. S.; Rego, D. B.; Perotti, J. M.; Kozimor, S. A.; Moore, E. K.; Ziller, J. W. Expanding Dinitrogen Reduction Chemistry to Trivalent Lanthanides via the LnZ 3 /Alkali Metal Reduction System: Evaluation of the Generality of Forming Ln 2 (μ - η 2 : η 2 -N 2 ) Complexes via LnZ 3 /K. J. Am. Chem. Soc. 2004, 126, 14574–14582. (8) Sun, J.; Berg, D. J.; Twamley, B. Yttrium Complexes of a Phenanthrene-Fused Cyclopentadienyl: Synthetic, Structural, and Reactivity Studies. Organometallics 2008, 27, 683–690. (9) Corbey, J. F.; Woen, D. H.; Palumbo, C. T.; Fieser, M. E.; Ziller, J. W.; Furche, F.; Evans, W. J. Ligand Effects in the Synthesis of Ln 2+ Complexes by Reduction of Tris(Cyclopentadienyl) Precursors Including C–H Bond Activation of an Indenyl Anion. Organometallics 2015, 34, 3909–3921. (10) Moehring, S. A.; Miehlich, M.; Hoerger, C. J.; Meyer, K.; Ziller, J. W.; Evans, W. J. A Room-Temperature Stable Y(II) Aryloxide: Using Steric Saturation to Kinetically Stabilize Y(II) Complexes. Inorg. Chem. 2020, 59, 3207–3214. (11) Jenkins, T. F.; Bekoe, S.; Ziller, J. W.; Furche, F.; Evans, W. J. Synthesis of a Heteroleptic Pentamethylcyclopentadienyl Yttrium(II) Complex, [K(2.2.2-Cryptand)]{(C 5 Me 5 ) 2 Y II [N(SiMe 3 ) 2 ]}, and Its C–H Bond Activated Y(III) Derivative. Organometallics 2021, 40, 3917–3925. 336 (12) Moore, W. N. G.; Ziller, J. W.; Evans, W. J. Optimizing Alkali Metal (M) and Chelate (L) Combinations for the Synthesis and Stability of [M(L)][(C 5 H 4 SiMe 3 ) 3 Y] Yttrium(II) Complexes. Organometallics 2021, 40, 3170–3176. (13) Cotton, F. Albert.; Schwotzer, Willi. Sm(.Eta.6-C6Me6)(.Eta.2-AlCl4)3: The First Structure of a Rare Earth Complex with a Neutral .Pi.-Ligand. J. Am. Chem. Soc. 1986, 108, 4657–4658. (14) Billow, B. S.; Livesay, B. N.; Mokhtarzadeh, C. C.; McCracken, J.; Shores, M. P.; Boncella, J. M.; Odom, A. L. Synthesis and Characterization of a Neutral U(II) Arene Sandwich Complex. J. Am. Chem. Soc. 2018, 140, 17369–17373. (15) Arnold, P. L.; Liddle, S. T. Synthesis and Reactivity of Neodymium(III) Amido-Tethered N-Heterocyclic Carbene Complexes. Comptes Rendus Chim. 2008, 11, 603–611. (16) Barnett, B. R.; Mokhtarzadeh, C. C.; Figueroa, J. S.; Lummis, P.; Wang, S.; Queen, J. D.; Gavenonis, J.; Schüwer, N.; Tilley, T. D.; Boynton, J. N.; Power, P. P.; Ditri, T. B.; Weidemann, N.; Barnett, B. R.; Agnew, D. W.; Figueroa, J. S.; Smith, P. W.; Ditri, T. B.; Barnett, B. R.; Carpenter, A. E.; Mokhtarzadeh, C. C.; Agnew, D. W.; Figueroa, J. S.; Smith, P. W.; Pratt, J. K.; Power, P. P.; Mendelson, N. D.; Figueroa, J. S.; Queen, J. D.; Power, P. P.; Agnew, D. W.; Carpenter, A. E.; Figueroa, J. S. TERPHENYL LIGANDS AND COMPLEXES. In Inorganic Syntheses; Power, P. P., Ed.; Wiley, 2018; Vol. 37, pp 85–122. (17) Gavenonis, J.; Tilley, T. D. Synthesis and Reactivity of Alkyl, Hydride, and Silyl Derivatives of the (Terphenyl)Imido Fragments Cp*(Ar Mes N)Ta (Cp* = η 5 -C 5 Me 5 ; Ar Mes = 2,6-(2,4,6-Me 3 C 6 H 2 ) 2 C 6 H 3 ) and Cp*(Ar Trip N)Ta (Ar Trip = 2,6-(2,4,6- i Pr 3 C 6 H 2 ) 2 C 6 H 3 ). Organometallics 2004, 23, 31–43. (18) Clegg, W.; Conway, B.; Kennedy, A. R.; Klett, J.; Mulvey, R. E.; Russo, L. Synthesis and Structures of [(Trimethylsilyl)Methyl]Sodium and ‐potassium with Bi‐ and Tridentate N‐ Donor Ligands. Eur. J. Inorg. Chem. 2011, 2011, 721–726. (19) Palumbo, C. T.; Halter, D. P.; Voora, V. K.; Chen, G. P.; Ziller, J. W.; Gembicky, M.; Rheingold, A. L.; Furche, F.; Meyer, K.; Evans, W. J. Using Diamagnetic Yttrium and Lanthanum Complexes to Explore Ligand Reduction and C–H Bond Activation in a Tris(Aryloxide)Mesitylene Ligand System. Inorg. Chem. 2018, 57, 12876–12884. (20) Trinh, M. T.; Wedal, J. C.; Evans, W. J. Evaluating Electrochemical Accessibility of 4f n 5d 1 and 4f n +1 Ln( II ) Ions in (C 5 H 4 SiMe 3 ) 3 Ln and (C 5 Me 4 H) 3 Ln Complexes. Dalton Trans. 2021, 50, 14384–14389. (21) Stoll, S.; Schweiger, A. EasySpin, a Comprehensive Software Package for Spectral Simulation and Analysis in EPR. J. Magn. Reson. 2006, 178, 42–55. (22) Herrington, J. R.; Estle, T. L.; Boatner, L. A. Electron-Paramagnetic-Resonance Investigation of the Dynamic Jahn—Teller Effect in Sr Cl 2 : Y 2 + and Sr Cl 2 : Sc 2 +. Phys. Rev. B 1973, 7, 3003–3013. 337 (23) Del Mar Conejo, M.; Parry, J. S.; Carmona, E.; Schultz, M.; Brennann, J. G.; Beshouri, S. M.; Andersen, R. A.; Rogers, R. D.; Coles, S.; Hursthouse, M. B. Carbon Monoxide and Isocyanide Complexes of Trivalent Uranium Metallocenes. Chem. - Eur. J. 1999, 5, 3000– 3009. (24) Garner, M. E.; Hohloch, S.; Maron, L.; Arnold, J. Carbon–Nitrogen Bond Cleavage by a Thorium‐NHC‐bpy Complex. Angew. Chem. Int. Ed. 2016, 55, 13789–13792. (25) Chen, X.; Li, Q.; Gong, Y.; Andrews, L.; Liebov, B. K.; Fang, Z.; Dixon, D. A. Formation and Characterization of Homoleptic Thorium Isocyanide Complexes. Inorg. Chem. 2017, 56, 5060–5068. (26) Tarlton, M. L.; Yu, X.; Ward, R. J.; Kelley, S. P.; Autschbach, J.; Walensky, J. R. Backbonding in Thorium(IV) and Uranium(IV) Diarsenido Complexes with t BuNC and CO. Chem. – Eur. J. 2021, 27, 14396–14400. (27) Bouzidi, Y.; Belkhiri, L.; Ephritikhine, M.; Halet, J.-F.; Boucekkine, A. Cyanide Linkage Isomerism in Cerium(III) and Uranium(III) Complexes. A Relativistic DFT Study. J. Organomet. Chem. 2017, 847, 82–89. (28) Hervé, A.; Bouzidi, Y.; Berthet, J.-C.; Belkhiri, L.; Thuéry, P.; Boucekkine, A.; Ephritikhine, M. U III –CN versus U IV –NC Coordination in Tris(Silylamide) Complexes. Inorg. Chem. 2015, 54, 2474–2490. (29) Robert, D.; Voth, P.; Spaniol, T. P.; Okuda, J. Rare‐Earth Metal Alkyl and Hydride Complexes Supported by a Linked Anilido–Cyclopentadienyl Ligand: Synthesis, Structure, and Reactivity. Eur. J. Inorg. Chem. 2008, 2008, 2810–2819. (30) Deelman, B.-J.; Stevels, W. M.; Teuben, J. H.; Lakin, M. T.; Spek, A. L. Insertion Chemistry of Yttrium Complex Cp*2Y(2-Pyridyl) and Molecular Structure of an Unexpected CO Insertion Product (Cp*2Y)2(.Mu.-.Eta.2:.Eta.2-OC(NC5H4)2). Organometallics 1994, 13, 3881–3891. (31) Den Haan, K. H.; Wielstra, Ytsen.; Teuben, J. H. Reactions of Yttrium-Carbon Bonds with Active Hydrogen-Containing Molecules. A Useful Synthetic Method for Permethylyttrocene Derivatives. Organometallics 1987, 6, 2053–2060. (32) Casely, I. J.; Ziller, J. W.; Evans, W. J. C–H Activation via Carbodiimide Insertion into Yttrium–Carbon Alkynide Bonds: An Organometallic Alder-Ene Reaction. Organometallics 2011, 30, 4873–4881. (33) Karpov, A. V.; Shavyrin, A. S.; Cherkasov, A. V.; Fukin, G. K.; Trifonov, A. A. Reactions of Bis(Alkyl)Yttrium Complexes Supported by Bulky N,N Ligands with 2,6- Diisopropylaniline and Phenylacetylene. Organometallics 2012, 31, 5349–5357. (34) Duchateau, R.; Brussee, E. A. C.; Meetsma, A.; Teuben, J. H. Synthesis and Reactivity of Bis(Alkoxysilylamido)Yttrium η 2 -Pyridyl and η 2 -α-Picolyl Compounds. Organometallics 1997, 16, 5506–5516. 338 (35) Evans, W. J.; Meadows, J. H.; Kostka, A. G.; Closs, G. L. Yttrium-89 NMR Spectra of Organoyttrium Complexes. Organometallics 1985, 4, 324–326. (36) Lauvergnat, D.; Maître, P.; Hiberty, P. C.; Volatron, F. Valence Bond Analysis of the Lone Pair Bond Weakening Effect for the X−H Bonds in the Series XH n = CH 4 , NH 3 , OH 2 , FH. J. Phys. Chem. 1996, 100, 6463–6468. (37) Bent, H. A. An Appraisal of Valence-Bond Structures and Hybridization in Compounds of the First-Row Elements. Chem. Rev. 1961, 61, 275–311. (38) Weinhold, F.; Landis, C. R. Valency and Bonding: A Natural Bond Orbital Donor-Acceptor Perspective; Cambridge University Press: Cambridge, UK ; New York, 2005. (39) Strauss, S. H. Copper(I) and Silver(I) Carbonyls. To Be or Not to Be Nonclassical. J. Chem. Soc. Dalton Trans. 2000, No. 1, 1–6. (40) Lupinetti, A. J.; Frenking, G.; Strauss, S. H. Prog. Inorg. Chem. 2001. (41) Hurlburt, P. K.; Rack, J. J.; Luck, J. S.; Dec, S. F.; Webb, J. D.; Anderson, O. P.; Strauss, S. H. Nonclassical Metal Carbonyls: [Ag(CO)]+ and [Ag(CO)2]+. J. Am. Chem. Soc. 1994, 116, 10003–10014. (42) Finze, M.; Bernhardt, E.; Willner, H.; Lehmann, C. W. Cyano- and Isocyanotris(Trifluoromethyl)Borates: Syntheses, Spectroscopic Properties, and Solid State Structures of K[(CF 3 ) 3 BCN] and K[(CF 3 ) 3 BNC]. J. Am. Chem. Soc. 2005, 127, 10712– 10722. (43) Rao, V. S.; Vijay, A.; Chandra, A. K. A Comparative Study of the Energetics, Structures, and Mechanisms of the HCN ↔ HNC and LiCN ↔ LiNC Isomerizations. Can. J. Chem. 1996, 74, 1072–1077. (44) Ballmann, G.; Elsen, H.; Harder, S. Magnesium Cyanide or Isocyanide? Angew. Chem. Int. Ed. 2019, 58, 15736–15741. (45) Coutsolelos, A. G.; Tsapara, A.; Daphnomili, D.; Ward, D. L. Pseudohalogeno-Bonding of Thallium(III) Porphyrins, Stabilisation of Isocyano Bonding. Crystal Structure of Isocyano(5,10,15,20-Tetraphenylporphyrinato)Thallium(III). J. Chem. Soc. Dalton Trans. 1991, No. 12, 3413. (46) Baghal-Vayjooee, M. H.; Collister, J. L.; Pritchard, H. O. The Enthalpy of Isomerisation of Methyl Isocyanide. Can. J. Chem. 1977, 55, 2634–2636. (47) Booth, M. R.; Frankiss, S. G. Trimethylsilyl Isocyanide. Chem. Commun. Lond. 1968, No. 21, 1347. (48) Booth, M. R.; Frankiss, S. G. The Constitution, Vibrational Spectra and Proton Resonance Spectra of Trimethylsilyl Cyanide and Isocyanide. Spectrochim. Acta Part Mol. Spectrosc. 1970, 26, 859–869. 339 (49) Górski, K.; Mech-Piskorz, J.; Leśniewska, B.; Pietraszkiewicz, O.; Pietraszkiewicz, M. Synthesis and Reactivity of 5-Heterotruxenes Containing Sulfur or Nitrogen as the Heteroatom. J. Org. Chem. 2019, 84, 11553–11561. (50) Harris, R. K.; Becker, E. D.; Cabral De Menezes, S. M.; Goodfellow, R.; Granger, P. NMR Nomenclature: Nuclear Spin Properties and Conventions for Chemical Shifts. IUPAC Recommendations 2001. International Union of Pure and Applied Chemistry. Physical Chemistry Division. Commission on Molecular Structure and Spectroscopy. Magn. Reson. Chem. 2002, 40, 489–505. (51) Dolomanov, O. V.; Bourhis, L. J.; Gildea, R. J.; Howard, J. A. K.; Puschmann, H. OLEX2 : A Complete Structure Solution, Refinement and Analysis Program. J. Appl. Crystallogr. 2009, 42, 339–341. (52) Sheldrick, G. M. Crystal Structure Refinement with SHELXL. Acta Crystallogr. Sect. C Struct. Chem. 2015, 71, 3–8. (53) Bain, G. A.; Berry, J. F. Diamagnetic Corrections and Pascal’s Constants. J. Chem. Educ. 2008, 85, 532. (54) Meihaus, K. R.; Fieser, M. E.; Corbey, J. F.; Evans, W. J.; Long, J. R. Record High Single- Ion Magnetic Moments Through 4fn5d1 Electron Configurations in the Divalent Lanthanide Complexes [(C5H4SiMe3)3Ln]-. J. Am. Chem. Soc. 2015, 137, 9855–9860. (55) Twamley, B.; Hardman, N. J.; Power, P. P. A Terphenyl Halide Series: 2,6-Trip 2 H 3 C 6 X (Trip = 2,4,6-Triisopropylphenyl; X = Cl, Br, I). Acta Crystallogr. C 2000, 56, e514–e515. (56) Weigend, F.; Ahlrichs, R. Balanced Basis Sets of Split Valence, Triple Zeta Valence and Quadruple Zeta Valence Quality for H to Rn: Design and Assessment of Accuracy. Phys. Chem. Chem. Phys. 2005, 7, 3297. (57) Peterson, K. A.; Figgen, D.; Dolg, M.; Stoll, H. Energy-Consistent Relativistic Pseudopotentials and Correlation Consistent Basis Sets for the 4d Elements Y–Pd. J. Chem. Phys. 2007, 126, 124101. (58) Smith, D. G. A.; Burns, L. A.; Patkowski, K.; Sherrill, C. D. Revised Damping Parameters for the D3 Dispersion Correction to Density Functional Theory. J. Phys. Chem. Lett. 2016, 7, 2197–2203. (59) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. A Consistent and Accurate Ab Initio Parametrization of Density Functional Dispersion Correction (DFT-D) for the 94 Elements H- Pu. J. Chem. Phys. 2010, 132, 154104. 340 CHAPTER 6: SYNTHESIS, STRUCTURE, AND REACTIVITY OF DYSPROSIUM COMPLEXES* *Manuscript submitted. 1. Benner, F.; Jena, R.; Odom, A. L.; Demir, S. Magnetic Hysteresis in a Pseudo Low-Coordinate Bisamide Dysprosium Complex. Manuscript Submitted. 2. Jena, R.; Benner, F.; Staples, R. J.; Demir, S.; Odom, A. L. A Neutral Du(II) bisamide: Synthesis, Magnetism, and a butterfly-P4 2- complex. Manuscript in preparation. 6.1 Introduction Exploration of lanthanides in the realm of Single-Molecule Magnets (SMMs) has gained much attention after the discovery of their unique potential for high-density data storage, spintronics, and quantum computing applications.1–4 SMMs have the remarkable property to keep magnetization even when an external magnetic field is removed, in stark contrast to ordinary paramagnets. The magnetic anisotropy and unpaired spin states lead to magnetization retention below a certain threshold temperature unique to each SMM. Currently, the field of SMM is dominated by lanthanide ions due to their high Spin-Orbit Coupling (SOC) and extremely anisotropic 4f electron densities. The limited radial extension of the 4f orbitals allows weak interaction with the ligand’s crystal field leading to higher unquenched orbital angular momentum contribution to the total magnetic moment, distinguishing them from their transition metal counterparts.5 Dysprosium stands out among all lanthanides due to its significantly high magnetic anisotropy arising from large unquenched orbital angular momentum. Secondly, Dy3+ is a Kramer ion meaning it has a bistable ground state which is essential for SMM behavior. The magnetic anisotropy of the Dy3+ ion can be enhanced by optimal ligand geometry around the metal center. A Dy3+ ion has an 341 oblate-shaped electron density that is a sphere flattened from poles (Figure 6.1). Thus, a strong axial field perpendicular to the plane of the spheroidal axes and a minimal equatorial crystal field would increase the magnetic anisotropy. The fundamental parameters to quantify SMM properties of molecules include a barrier to spin reversal (Ueff), blocking temperature (TB), and coercivity (Hc). Current research focuses on improving the TB of SMM for advancement in practical applications. Layfield and coworkers isolated a [(CpCp*)Dy]+ cation as a salt of non-coordinating anion, which showed tremendous SMM behavior where Cp = pentaisopropylcyclopentadienyl, and Cp* = pentamethylcyclopentadienyl. This state-of-the-art molecule shows magnetic hysteresis up to liquid N2 temperature.6 Figure 6.1. (left) Ideal design for an SMM molecule. (right) Best SMM known to date by Layfield and coworkers.6 Ligand design is crucial for providing a strong axial field to the metal center. Historically, bulky cyclopentadienyl-based ligands have been employed to yield metal sandwich complexes with high TB.6,7 Recent developments in the field of SMM have highlighted the use of bulky amide ligands, 342 which provide a stronger axial field with minimum equatorial interactions.8,9 Mills and coworkers have isolated a bent dysprosium bis(amide) SMM, however, they got 3 times lower Ueff than the predicted value for the linear bis-amide complex (Figure 6.2) due to rapid quantum tunneling.10 Murugesu and coworkers have shown the utility of terphenyl bisaniline ligand for the isolation of see-saw shaped dysprosium SMM, which shows magnetic hysteresis up to 5.8 K (Figure 6.2).9 Figure 6.2. Reported examples of dysprosium amide complexes. Gao and coworkers reported the first dysprosium 6-arene piano-stool complex)Dy(k2- AlCl4)3].11 There are only a few examples in the literature showcasing the use of amide ligands in conjunction with arene ligands (one shown above) to develop dysprosium SMMs. The goal of this project is to synthesize dysprosium molecules using terphenyl amide ligand (used in the previous chapter) and study their magnetic properties in collaboration with the Demir group. We believe that this ligand system can give rise to a strong axial field with minimal equatorial interactions and, consequently, could enhance magnetic anisotropy. Here I will be discussing synthesis and structural analysis in detail. Florian Benner from the Demir group has collected and analyzed magnetism data herein. 6.2 Synthesis and Characterization of Dy(NHAr*)2Cl Complex For discussion purposes, I have continued the numbering of metal complexes from Chapter 5. As dysprosium is a paramagnetic ion, understanding its reaction chemistry has always been challenging because we cannot simply check NMR spectroscopy to confirm the reaction progress. 343 Researchers have used yttrium as a surrogate to establish reaction protocol first and then move to dysprosium chemistry, as yttrium has similar reaction chemistry to the rare earth elements. Further, the atomic radius of yttrium (180 pm) and dysprosium (177 pm) is comparable, which contributes to similar reactivity.12 Synthesis of Dy(NHAr*)2Cl (4) was carried out by salt metathesis of KNHAr* with dysprosium chloride resembling the synthesis of Y(NHAr*)2Cl (1). After 16 h of reaction time yellow crystals of 4 were isolated in 58% yield. Figure 6.3. (Top) Synthesis of Dy(NHAr*)2Cl (4) and (bottom) structure of 4 from single-crystal X-ray diffraction (thermal ellipsoids drawn at 50%). The solvent molecule (1 n-hexane) and hydrogen atoms on carbon are removed for clarity. The 6-arene (red) is discussed as Ar1’ in the text. From single-crystal X-ray diffraction complexes 4 and 1 are isostructural with only one of the ortho aryl rings of the ligand coordinated to the metal center. The Dy–Cnt (Ar1’) distance is 344 2.5417(2) Å, while the second closest ortho aryl ring is 3.4740(2) Å from the metal center suggesting no interaction with the metal center. A congener bulky phenoxide ligand with dysprosium (ArO)Dy(OAr’)] (Ar’ = 6-Dipp-2-(2’-iPr-6’-CHMe(CH2 -)C6H3)C6H3O-, has a Dy–Cnt distance of 2.649(3) Å (disordered structure).13 As we know amides show superior donor ability compared to phenoxides, consequently leading to relatively more backbonding into π* of the Ar1’ ring, thereby resulting in a smaller metal-centroid distance. The Dy–Cnt distance for 4 is consistent with Dy bisanilide-based system (2.56 Å).9 In the solid-state the two amides are different in bond length, with Dy1−N1 being 2.258(3) Å, and Dy1−N2 being 2.235(3) Å. The bond angle Dy1−N1−C1(ipso−Ar1’) is 131.9(7)°, and Dy1−N2−C36(ipso−Ar2) is 137.9(3)°. Similar behavior was observed with yttrium complex 1, where the presence of only one 6-arene ring interaction with the metal center gives rise to variance in the amide bonds. The N–Dy–N angle is 140.9(1)°, highlighting a substantially bent nitrogen coordination to the metal center in the equatorial plane. The absorption spectra of 4 were recorded in diethyl ether. In the Vis-NIR region, the yellow solution showed multiple narrow bands between 550 to 1400 nm with molar extinction coefficients of ε = 1.5–2.5 M-1cm-1, which were attributed to Laporte forbidden 4f–4f transitions. Complex 4 also shows two sharp and intense bands in the UV region at 244 nm (ε = 7300 M-1cm-1) and 297 nm (ε = 11081 M-1cm-1). These intense bands are due to ligand-based π–π* charge transfer transitions. Similar absorptions were observed in the previously reported U(NHAr*)2I.14 Solution-state magnetic susceptibility studies of 4 were conducted using the Evans method. At room temperature, the χMT value of 4 is 14.38 cm3Kmol-1 (μeff = 10.73 µB), which is in excellent agreement with the expected theoretical value for Dy3+ ion at room temperature (14.17 cm3Kmol- 1).1,15 As expected, the χMT value of 4 decreased to 8.35 cm3Kmol-1 (μeff = 8.27 µB) at 181 K.16 345 The solid-state SQUID measurements were carried out using a polycrystalline sample of 4. The room-temperature χMT value 13.86 cm3Kmol-1 is in close agreement with the Evans method. The Ueff for 4 is 601 cm-1 which is consistent with the highest reported Ueff for formally two coordinate amide and arene stabilized Dy complex.17 The hysteresis loops measured for 4 between 1.8 and 8.0 K are all waist constricted and thus, closed at zero field. At higher fields, the hysteresis loops open to a butterfly shape below 8 K. At 8 K, the hysteresis curves are fully closed (Figure 6.4). Figure 6.4. Variable-field magnetization (M) data collected for Dy(NHAr*)2Cl (4). 6.3 Synthesis and Characterization of a Neutral Dy(II) Complex Similar to the synthesis of the Y(II) complex, the reduction of Dy(NHAr*)2Cl (4) was carried out by using the common reductant KC8 resulting in the formation of Dy(NHAr*)2 (5). The reaction color changed from pale yellow to black within an hour of reaction time. The dark solid was recrystallized using n-hexane as a solvent, which afforded black crystals overnight at −35 C in 32% crystalline yield. 346 Figure 6.5. (Top) Synthesis of Dy(NHAr*)2 (5) and (bottom) structure of 5 from single-crystal X- ray diffraction (thermal ellipsoids drawn at 50%). The solvent molecule (1 THF) and hydrogen atoms on carbon are removed for clarity. From single-crystal X-ray diffraction, complex 5 is isostructural to 2 and the U(II) complex reported by our group previously18,14, where 5 crystallizes as a C2 symmetric molecule with only half of the molecule occupying an asymmetric unit of the cell with a molecule of disordered THF. The significant variations in the N–M–N angles (M = Dy, Y) between the solid-state structures of 5 and 2 are worth mentioning in this comparison. The N–Dy–N angle in complex 5 is 102.08(11) which is more consistent with the angle N–U–N in the U(II) complex than the Y(II) N–Y–N angle of 120.4(1). This large disparity in the N–M–N could be due to the presence of f-orbitals in 347 dysprosium and uranium as opposed to yttrium. However, the Cnt–Dy–Cnt angle in 5 134.40(7) is consistent with the Cnt–Y–Cnt angle in 2 133.87(1). The metal centroid distance between the reduced dysprosium center 5 and ortho aryl rings has shrunk to 2.4607(9) Å relative to complex 4 (2.541(2) Å). Naturally, the size of a reduced metal center is larger, in lanthanides an increase of 0.16–0.19 Å (difference in the ionic radii) is observed for a formally reduced metal center (+2 oxidation state).19 Consequently, lengthening of metal centroid distance has been observed for [LnCp’3]- based systems to minimize electronic repulsion between metal and ligand systems.19,20 However, in our case, the Dy–Cnt distance has reduced by 0.08 Å from complex 4 to 5. Evans and coworkers have extensively discussed the preference for the 4f95d1 electronic ground state due to the crystal field induced by the ligand’s symmetry around the metal center.21 They showed that the perceived C3 symmetry of triscyclopentadienyl ligand around the metal center drops the energy of 5dz2 orbital and thus is suitable for populating it compared to 4f orbital. Based on the same argument however, Long and coworkers reported a rare example of low valent C2 symmetric linear metallocene dysprosium complex Dy(CpiPr5)2 showing a similar trend for the metal centroid distances.22 This suggested that the ground state electronic configuration for 5 may have a considerable 4f95d1 electronic configuration. The unpaired electron in the 5d1 could introduce covalency between metal and aryl systems due to the extended radial extension of 5d orbitals as opposed to 4f.22,23 Similar to our previous Y(II) and U(II) systems, this will lead to an increased back donation of electron density from the formally reduced metal center to the π* of the arene system. Consequently, contraction in Dy–Cnt should be observed, and indeed is observed in 5. Further, the presence of isopropyl groups hinders ortho-aryl rings to approach the metal center, thereby only slight contraction in Dy–Cnt (0.08 Å) is observed. 348 The absorption spectrum of 5 features a broad band between 600 nm to 1600 nm resembling the spectrum of Y(II) complex 2 and with the most recent example of Dy(II) bis-amidinate complex.18,24 This further confirms the occupancy of an unpaired electron in 5d. The molar extinction coefficient (ε) for the Laporte-allowed 4f–5d electronic transition in 5 is ⁓780 M-1cm-1 (at 755 nm) which is enormously higher than the weak 4f–4f transitions in complex 4. Analogous to 4, a strong charge transfer band for 5 is observed in the UV region with ε = 19307 M-1cm-1 (at 297 nm). The Evans method gave us room temperature magnetic susceptibility value for 5 as χMT = 15.28 cm3•Kmol-1 (μeff = 11.08 µB). The theoretical magnetic moment values Dy3+ and Dy2+ ions have similar effective magnetic moments 25,20. For lanthanides with 4fn5d1 electron configuration, the ground state can be described by Russell-Saunders coupling (LS coupling) rules. Inspired by Cloke and coworkers,26 the Long and Evans group extensively studied the impact of LS coupling on magnetic susceptibility values. There are two scenarios to consider, and the summary of their findings is tabulated below. Coupled Scheme (f and d spin coupling >> LS coupling) • Strong spin-spin coupling between the f and Uncoupled Scheme (LS coupling >> f electron spin coupling) • LS coupling for f electrons is stronger. d electrons. • Spins of f electrons and d-electrons Stot = S4f +1/2. • The total spin calculated above couples with the orbital angular momentum of the f electrons to yield an overall J value (Jtot = Stot + L4f). • Theoretical χMT values are calculated using these couplings. • The χMT value for d electron (0.355 for S = ½) is added to χMT value of LS coupling of the f electrons. For Dy(II) ion with 4f95d1 electron configuration, the coupled scheme and uncoupled scheme will give us χMT values of 17.01 and 14.51 cm3Kmol-1 respectively. For complex 5, the χMT = 15.28 349 cm3•Kmol-1) which is in close agreement with the Dy(II) bis-amidinate complex reported by Zhen and coworkers.24 Their in-depth analysis of Dy(II)bis-amidinate complex using Complete Active Space Self-Consistent Field Spin-Orbit (CASSCF-SO) calculations showed that f‒d coupling is stronger than LS coupling and further supported the existence of 4f95d1 configuration over 4f10 for Dy(II). However, the experimental χMT value (15.28 cm3•Kmol-1) for 5 is lower than the theoretical value (17.01 cm3•Kmol-1) which could be attributed to the 5d and 6s orbital mixing which leads to weaker f‒d coupling in linear metallocene complexes.24,27 The solid-state SQUID measurements were carried out using a polycrystalline sample of 5. A closed hysteresis loop was obtained at zero field, as expected from a non-Kramer Dy(II) ion. The electrochemical properties of Dy(NHAr*)2 (5) were accessed by cyclic voltammetry using 1.2 mM of 5 in ether with [NBu4][B(3,5-(CF3)2C6H3)4] as an electrolyte and glassy carbon as a working electrode. The redox potential of the Dy(NHAr*)2 +1/0 couple was observed at E1/2 = –1.094 ± 0.001 V w.r.t to FeCp2 +1/0 (Figure 6.6). A reversible feature was observed when a voltammogram of 5 was obtained using different scan rates (see experimental section). 350 Figure 6.6. Cyclic Voltammogram of 2 in Et2O with [NBu4]+[B(3,5-(CF3)2C6H3)4]– (100 mM) as supporting electrolyte and Fc+/0 as 0 V. This voltammogram was scanned in the negative direction with 100 mV/s scan rate. DFT calculations were performed to understand the bonding interactions between reduced Dy(II) complex 5 and the ligand system. The HOMO, LUMO, and Mulliken spin density map encompass the metal and the arene rings suggesting the delocalization of the electron density throughout the metal-ligand backbone (Figure 6.7). This is also consistent with our divalent yttrium complex (2).18 351 Figure 6.7. (top left) HOMO, (top right) LUMO, and (bottom)spin density map of Dy(NHAr*)2 (5). Hydrogens are removed for clarity. 6.4. Synthesis of Characterization of Dysprosium(III) Cation As mentioned before, the ideal design for SMMs exhibits strong axial ligand donors to increase single-ion magnetic anisotropy. Complex 4 shows poor SMM properties, therefore, we decided to remove the chloride from the molecule's equatorial plane to eliminate the transverse field provided by the chloride anion. This was achieved by using two synthetic routes: (A) oxidation of Dy(NHAr*)2 (5) by [FeCp2][BArF24] or (B) abstraction of chloride anion from the Dy(NHAr*)2Cl (4) using TlBArF24. Although route A has a literature precedence, it adds an extra step to our 352 synthesis. Treatment of Dy(NHAr*)2 with TlBArF24 resulted in a drastic color change from black to orange and the formation of a cationic complex [(NHAr*)2Dy][BArF24] (6) in 22% crystalline yield. Figure 6.8. (Top) Synthesis of [(NHAr*)2Dy][ BArF24] (6) and (bottom) structure of 6 from single-crystal X-ray diffraction (thermal ellipsoids drawn at 50%). Counter anion BArF24 and hydrogen atoms on carbon are removed for clarity. 353 Figure 6.8 (cont’d) Single-crystal X-ray diffraction studies revealed that complex 6 recrystallizes as a C2 symmetric molecule with only half of the asymmetric unit in the unit cell. Similar to 5, we observe both ortho- aryl rings coordinated to the metal center with 6-binding mode. The Dy–Cnt distance is 2.4967 Å, approximately 0.05 Å shorter than the corresponding distance in 4, which can be ascribed to less equatorial steric hindrance due to the lack of chloride. Consequently, the Dy–N is slightly shorter in 6 as opposed to 4. As Dy–Cnt distance has increased by 0.036 Å in 6 compared to 5, this was ascribed to reduced backbonding from the metal center to the π* of the aryl ligand. The N–Dy–N angle is 121.71(13) which is more consistent with 5 and almost 19 more acute than its Dy3+ congener 4 (140.9). This could be attributed to the additional 6-arene coordination in 6 as compared to 5, thus more space is required around the metal center. Consequently, N–Dy–N will contract owing to the smaller angle. The easy axis of magnetization for these molecules passes through N–Dy–N, thus this 19 deduction in the angle could drastically affect the SMM behavior 354 between complexes 5 and 6. The absorption spectra of 6 were recorded in diethyl ether. In the Vis-NIR region, the yellow solution showed multiple narrow bands between 500 to 1400 nm with molar extinction coefficients of ε = 4–12 M-1cm-1 which could be attributed to Laporte forbidden 4f–4f transitions. Complex 6 also shows two sharp and intense bands in the UV region at 304 nm and 415 nm (ε = 4618 M-1cm- 1). These intense bands are likely due to ligand-based π–π* charge transfer transitions. Similar features are observed in arene-capped bisamide complexes such as [K(DME)n][LDyX2] (L ={C6H4[(2,6-iPrC6H3)NC6H4]2}2−, X = Cl, I).17 The solid-state SQUID measurements were carried out using a polycrystalline sample of 6. The room-temperature χMT value 14.19 cm3Kmol-1 is in excellent agreement with the expected value of Dy3+ ion (14.17 cm3Kmol-1). The Ueff for 6 is 597 cm-1 which is similar to 4. However, the hysteresis loops measured for 6 are open at zero field below 19.0 K (Figure 6.9). At 1.8 K, the coercive field is HC = 1.03 T which is the largest found for a mononuclear dysprosium complex with arene or amide ligands in the first coordination sphere.10,17 355 Figure 6.9. Variable-field magnetization (M) data was collected for [Dy(NHAr*)2]+ (4). 6.5 Reactivity of Dy(NHAr*)2 (5) and Discovery of Dysprosium Isocyanide Dy(NHAr*)–NC (7) Similar to our study of the reactivity of Y(NHAr*)2 (2), we aimed to test the reactivity of Dy(NHAr*)2 (5) to determine if it would yield the corresponding Dy(NHAr*)–NC (7). Indeed, we saw similar reactivity when high energy metallaradical was treated with either tBuNC or tBuCN, and the reaction color immediately turned from black to pale yellow. The yellow crystals of Dy(NHAr*)–NC (7) were isolated in 34% yield. While there are some metal isocyanide complexes known in the literature,18,28,29 this is the first example of a dysprosium isocyanide complex. 356 Figure 6.10. Synthesis of Dy(NHAr*)–NC (7) and (bottom) structure of 7 from single-crystal X- ray diffraction (thermal ellipsoids drawn at 50%). Hydrogen atoms on carbon are removed for clarity. Analysis of structure obtained from single-crystal X-ray diffraction reveals that complex 7 is isostructural to 4 with only one of the ortho-aryl rings being 6-coordinated to the metal center. The Dy1–N1 (2.372(2) Å) bond length is longer than Dy1–N2 (2.236(1) Å) and Dy1–N3 (2.218(1) Å). The N2–Dy1–N3 angle is 136.80(7) similar to 4. Further, the N1–C1 bond length is 1.025(4) Å which is consistent with N≡C moiety. Due to the paramagnetic dysprosium center, complex 7 is challenging to characterize by NMR spectroscopy, and was even found to be NMR silent. I tried refining complex 7 with both 357 possibilities of cyanide and isocyanide. The statistics for Dy–NC structure R1 = 3.64% and wR2 = 8.64% are better than Dy–CN structure R1 = 3.78% and wR2 = 9.14%. Additionally, we tried assigning variable percentages of atom occupancy to N and C; however, all attempts increased refinement parameters (see details in the experimental section). After encountering obvious limitations with NMR spectroscopy for paramagnetic ions, we turned to DFT calculations to understand the preference of isocyanide vs cyanide for the dysprosium center. The ground state energy and frequency calculations were executed using TPSSH functional with SVP basis set on all atoms and ECP55MWB core potential on dysprosium. The Dy–NC structure is ⁓2.5 kcal/mol more stable than the Dy–CN system in the ground state (more details in the experimental section). Infrared spectroscopy was utilized to analyze the stretching frequency of the N≡C bond in 7. An intense band at 2052 cm-1 was observed in n-hexane. The dysprosium isocyanide bond stretch in 7 is in excellent agreement with the other metal isocyanides.30,31 Our yttrium congener has almost the same νNC 2053 cm-1, further confirming the formation of dysprosium isocyanide. The theoretically calculated (by DFT) νNC for Dy–NC is 2131.5 cm-1, which is in better agreement with the experimental value than Dy–CN (2217.6 cm-1). 6.6 Reactivity of Dy(NHAr*)2 (5) with Elemental Phosphorous (P4) Phosphorous is one of the biogenic elements and a key component in industrial chemistry.32 White phosphorous is known for its extreme reactivity and toxicity among the variety of allotropic forms of phophorous. Organophosphorus compounds are widely used as ligands for homogeneous catalysis.33,34 The precursors for organophosphorus compounds are traditionally synthesized from the oxidation of elemental P4 with chlorine gas followed by halogen exchange with fluoride. However, using chlorine gas poses a challenge to use it on an industrial scale efficiently. By 358 leveraging the coordination chemistry of metals, scientists have been modifying the reactivity and selectivity of phosphorous, enabling the synthesis of novel phosphorous-containing compounds.35– 39 Here we wanted to study the reactivity of our high-energy metallaradical Dy(II) complex 5 with P4. Phosphorous is isolobal with methine (CH), which means P4 is isolobal with C4H4. Consequently, P4 undergoes 2 electron reduction to yield a P4 2- system; either in a cyclotetraphosphide (cyclo P4 2-) or [1.1.0]bicyclotetraphosphane-1,4-diide (butterfly P4 2-) form. The expected reactivity was that the two equivalents of complex 5 would react with one P4 molecule to form a bimetallic (NHAr*)2Dy–P4–Dy(NHAr*)2 sandwich complex, where the P4 unit is doubly reduced (P4 2-) by accepting an electron from each dysprosium (Figure 6.11). Figure 6.11. (top) Reduction of elemental P4. (bottom) Expected reaction of 5 with P4. Treatment of 0.5 equiv. of P4 with complex 5 resulted in a sudden reaction color change from black to orange in benzene. Orange-colored X-ray quality crystals were obtained after 48 hours at ‒35 °C. In contrast to the expected reactivity, the structure obtained from single crystal X-ray diffraction showed a butterfly binding mode of the P4 moiety to only one dysprosium metal center. However, we observed another dysprosium molecule with two amide ligands in the crystal lattice. 359 At first glance, the crystal structure suggested that P4 is undergoing only one electron reduction. Based on chemical intuition, I conducted a stoichiometric reaction between complex 5 and P4 to determine if I could drive the reaction to completion and isolate (NHAr*)2Dy–P4. Interestingly, this also resulted in the same crystal structure where the (NHAr*)2Dy–P4 is co-crystallized with (NHAr*)2Dy and three disordered diethyl ether molecules. Upon close analysis of bond lengths and angles, we determined that the obtained product is a salt [(NHAr*)2Dy][(NHAr*)2Dy–P4] (8), with (NHAr*)2Dy as the countercation (Figure 6.12). Figure 6.12. (top) Synthesis of [(NHAr*)2Dy][(NHAr*)2Dy–P4] (8) and (bottom) structure of 8 from single-crystal X-ray diffraction (thermal ellipsoids drawn at 50%). Counter cation [(NHAr*)2Dy]+ and hydrogen atoms on carbon are removed for clarity. 360 Figure 6.12 (cont’d) The orange color of the compound is consistent with the salt complex [(NHAr*)2Dy][BArF24] (8). Further, 8 has poor solubility in n-hexane suggesting the formation of a salt complex. Based on the above analysis we can conclude that it is indeed a double reduction of the P4 to yield a P4 2‒ unit. The structure of [(NHAr*)2Dy–P4]- has a disorder at the Dy and P4 unit. Interestingly, modeling the disorders showed that the binding preference of P4 2- is distributed between different modes. Specifically, P4 2‒ prefers to bind through the butterfly mode ⁓70% of the time and through the pseudo cyclo mode ⁓30% of the time. This further explains the larger thermal ellipsoids in crystal structure due to the fluxional P4 2- unit. 361 Bond Lengths (Å)/ Angles (°) Dy‒N1 Dy‒N2 Dy‒P2 Dy‒P4 Dy‒P1 Dy‒P3 P2‒P3 P3‒P4 P4‒P1 P1‒P2 P1‒P3 P2‒P4 Dy‒Cent Dy‒P2‒P1 Dy‒P4‒P1 P3‒P2‒P1 P3‒P4‒P1 P4‒P3‒P2 P2‒P1‒P4 N1‒Dy‒N2 Butterfly P4 2- Pseudo planar P4 2- 2.195(9) 2.218(8) 2.824(4) 2.672(12) 3.269(8) 3.323(4) 2.197(6) 2.181(9) 2.166(10) 2.146(7) 2.294(6) 3.250(6) 3.6262(7), 3.3164(6) 81.0(2) 84.3(4) 63.8(2) 63.7(3) 95.9(4) 97.8(3) 136.16(18) 2.522(9) 2.259(8) 2.961(8) 2.933(19) 3.06(2) 3.002(7) 2.114(10) 2.055(19) 2.170(17) 2.015(12) 2.501(16) 3.287(11) 4.1198(13), 2.9703(17) 73.2(7) 72.0(7) 74.5(4) 72.5(6) 104.0(6) 103.5(8) 117.7(2) Figure 6.13. (top) Zoom-in for P4 2- butterfly (left) and planar (right) form. (bottom) Structural metrics from Single-crystal x-ray diffraction. The solid-state structure of [(NHAr*)2Dy–P4]- shows no 6-arene ring interaction with ortho- aryl rings. This absence of 6-arene interactions can be attributed to the significant steric hindrance 362 around the metal center due to the additional P4 unit, which effectively caps dysprosium and obstructs the ortho-aryl ring’s approach. The Dy‒N distance for [(NHAr*)2Dy–P4]- is consistent with 4, 6, and 7. The Dy‒N distances in counter cation [(NHAr*)2Dy]+ are 2.224(6) and 2.235(8) Å. The Dy‒Cnt distance (2.521 Å) is slightly longer than in cationic structure 6 (2.497 Å). The Cnt‒ Dy‒Cnt and N‒Dy‒N angles are 140.13° and 108.40°, respectively, which are intermediaries between complexes 5 and 6. In the butterfly binding mode two phosphorous atoms P2 (2.824(4) Å) and P3 (2.672(12) Å) are closer to the metal as compared to P1 (3.269(8) Å) and P4 (3.323(4) Å). In pseudo-cyclo binding mode Dy‒P1, P2, P3, and P4 bond lengths are 2.961(8), 2.933(19), 3.06(2), and 3.002(7) Å respectively. This suggests that all the phosphorous atoms are almost equidistant from the metal center which is consistent with a cyclo form. As expected, Dy‒P bond lengths in both forms are longer than the transition metal analogs40,41 due to the higher electronic repulsion between the bigger lanthanide ion and P4 2-. The average P‒P distance in butterfly and cyclo P4 2- structures is 2.173 Å and 2.089 Å respectively. These are between typical P‒P single (2.22 Å) and P=P (2.04 Å) double bond lengths, further indicating the presence of P4 2- ligand. Further, the recrystallization of Complex 8 from toluene and n-hexane resulted in visibly bigger crystals, so I collected data again. This time, we again observed a predominance of butterfly- structure (77.5%) over pseudo-planar-P4 2- (22.5%) at 100 K. This made us think that if we change the temperature, can we change the percentages of these forms? We collected data at 200 K, where the butterfly:pseudo-planar is almost the same (79:21). Further, increasing the temperature to 300 K led to the decomposition of the crystal, which could be due to loss of crystallization solvent at room temperature under N2 flow. However, the pre-experiment gives the same unit cells as previous experiments. 363 Let us understand the preference for different binding modes of P4 2- unit using DFT calculations. The gas phase ground state energy calculation suggested that the P4 2‒ unit is ⁓10 kcal/mol more stable when the charge is delocalized in a planar ring compared to when it is localized on two phosphorous atoms. However, when the P4 2‒ unit is bound to a metal then the butterfly form is ⁓18 kcal/mol more stable than the planar form. This could be attributed to the fact that the highly electropositive nature of the metal could induce uneven charge distribution in the P4 unit and make two phosphorous atoms (bound to metal) more negatively charged leading to a butterfly binding mode. Figure 6.14 Relative enthalpies (M06/aug-cc-PVTZ) of P4 2‒ units in different binding modes. 364 Interestingly, planar mode is energetically more favorable when the P4 2‒ unit is sandwiched between two metals, however in our case the sterically encumbered amide ligand restricts the approach of two metals towards each other to make an (NHAr*)2Dy–P4–Dy(NHAr*)2 sandwich complex. The frequency calculations were done to obtain thermochemical data on both isomers for [(NHAr*)2Dy–P4]- structure. The sum of electronic and thermal free energies for butterfly and pseudo planar [(NHAr*)2Dy–P4]- are –4312.944645 and –4312.946343 Hartree/particle, respectively. This suggests that the planar binding mode of the P4 unit is thermodynamically more stable (⁓1 kcal/mol) than the butterfly binding mode. This small energy difference indicates that both forms will coexist as obtained from the crystal structure. As expected, HOMO and LUMO are heavily located on anionic phosphorus for both forms of [(NHAr*)2Dy–P4]- structure. Figure 6.15. Frontier molecular orbitals for butterfly [(NHAr*)2Dy–P4]- structure. 365 Figure 6.16. Frontier molecular orbitals for pseudo planar [(NHAr*)2Dy–P4]- structure. 6.7 Conclusions In this chapter, we synthesized five novel dysprosium complexes by leveraging the chemical information from the yttrium chapter. I successfully isolated the Dy(NHAr*)2Cl (4) and Dy(NHAr*)2 (5) which are isostructural to their yttrium congener. The Vis-NIR spectroscopy, room-temperature magnetic susceptibility value (χMT), and DFT calculations suggest that the ground state electronic configuration for 5 is 4f95d1. The reduction potential for Dy(NHAr*)2 +/0 pair is –1.094 ± 0.001 V with respect to FeCp2 +/0 in diethyl ether. The magnetic hysteresis of Kramer ion 4 is closed at zero field and is thus not a good magnet. For the SMM application, we removed chloride from the first coordination sphere of complex 4 to make Dy(NHAr*)2 + (6). Although this pseudo low-coordinate complex has a similar barrier for spin reversal (Ueff) as 4, it features an open hysteresis at zero field with a blocking temperature (TB) of 19 K. This is a record temperature for any mononuclear dysprosium complexes with amide and 366 arene ligands. The highly reactive metal radical 5 reacts with tBuNC and tBuCN to give a rare metal isocyanide Dy(NHAr*)2−NC (7) complex after the loss of the tert-butyl radical. DFT calculations suggest that the metal isocyanide is more stable than cyanide. This reaction shows charge-controlled reactivity like the yttrium analog (3). Further treatment of 5 with elemental P4 resulted in the formation of the anionic [Dy(NHAr*)2−P4]- (8) with a Dy(NHAr*)2 + as a counterion. The doubly reduced P4 showed fluxional behavior at the metal center with butterfly geometry 70% of the time and 30% cyclo geometry. The use of bulky terphenyl amide ligands has led to the discovery of rare metal complexes, which underscores the importance of ligand design for the isolation of these complexes. Further, the exciting discovery of novel isocyanide and P4 coordinated metal complexes has shown that there is still so much to learn about the reactivity of organometallic complexes. 367 6.8 Experimental Details All manipulations were done under a purified nitrogen atmosphere in either a glove box or using either a glove box or Schlenk techniques. n-Hexane was dried with CaH2, distilled under nitrogen to remove oxygen, and stored over 4 Å molecular sieves for 12 h prior to use. Diethyl ether was purified by passing through alumina columns to remove water after being sparged with dry nitrogen to remove oxygen n-Hexane was dried over CaH2 and, after sparging with dinitrogen, was passed through alumina. The melting point or decomposition temperature was measured by a capillary packed inside a nitrogen glovebox, sealed, and transferred to the instrument. 2,6- dichloroiodobenzene was purchased from Oakwood and, after freeze-pump-thawing three times to remove dissolved gases, was transferred into the glovebox. DyCl3 and trimethylsilylmethyllithium solution in pentane (0.1 M) were purchased from Sigma-Aldrich and used as received. Trimethylsilylmethylpotassium and tosyl azide were synthesized according to the literature procedure.42,43 H2NAr* was also synthesized according to the literature.44,45 Synthesis of Metal Complexes Synthesis of (NHAr*)2DyCl (4) A 20 mL scintillation vial charged with a stir bar was loaded with DyCl3 (108.6 mg, 0.4 mmol, 1 equiv.) and diethyl ether (8 mL). A separate 20 mL scintillation vial was loaded with KNHAr* (433.0 mg, 0.8 mmol, 2 equiv.) and diethyl ether (8 mL). Both solutions were cooled in a dry ice/acetone cold well for 20 min, and then the DyCl3 solution was moved to stir plate to stir. When the solution had thawed enough to stir, a cold solution of KNHAr* was added dropwise over 10 min. The solution was left to stir for 16 h at room temperature. The volatiles were removed in vacuo. The resulting yellow solid was extracted with n-hexane, and the solvent was removed in vacuo. A concentrated solution in n-hexane was kept overnight at –35 °C in the freezer, resulting in the 368 formation of yellow-colored X-ray quality single crystals (277 mg, 58% yield). Complex 4 is mostly NMR silent, with some of the isopropyl peaks showing up. 1H NMR (500 MHz, C6D6) δ 3.26, 3.04, 2.87, 2.11, 1.29, 1.23, 1.11, 0.88. Anal. Cald for C72H100N2DyCl: C, 72.57; H, 8.46; N, 2.35. Found: C, 72.32; H, 8.67; N, 2.26. λmax (nm) 1205, 1154, 987, 866, 771, 633, 516 (at higher concentration), 297 (at lower concentration). Decomposition temperature 82 °C. Synthesis of (NHAr*)2Dy (5) A 20 mL scintillation vial charged with crystals of Dy(NHAr*)2Cl (4) (272.8 mg, 0.22 mmol, 1 equiv.), THF (5 mL), and a magnetic stir bar. The vial was placed in a liquid nitrogen-cooled cold well until the solution froze. Once frozen, the vial was removed from the cold well to a magnetic stir plate. When the solution had thawed enough to stir, a suspension of KC8 (61.9 mg, 0.45 mmol, 2 equiv.) in THF (2 mL) was added. The solution rapidly turned color from yellow to black. The reaction was stirred for 1 h at room temperature. The volatiles were removed in vacuo, and the remaining residue was dissolved in n-hexane and filtered twice through Celite using a pipette filter. Black-colored X-ray quality single crystals were grown by chilling a concentrated n-hexane solution of 2 in a –35 °C freezer overnight (85.7 mg, 32.4% yield). The complex is stable for months at –35 °C. The crystals contain one THF molecule per metal complex. Anal. Cald for C72H100N2Dy: C, 74.80; H, 8.72; N, 2.42. C72H100N2Dy•C4H8O: C, 74.30; H, 8.87; N, 2.28. Found: C, 73.99; H, 9.45; N, 2.14. Decomposition temperature 110 °C. . λmax (nm) 755 (at higher concentrations), 297 (at lower concentrations). Synthesis of [(NHAr*)2Dy][BArF24] (6) A 20 mL scintillation vial was charged with 4 (106.9 mg, 0.09 mmol, 1 equiv.), diethyl ether (3 mL), and a magnetic stir bar. A separate 20 mL scintillation vial was loaded with TlBArF24 46 (95.7 mg, 0.09 mmol, 1 equiv.) and diethyl ether (2 mL). The TlBArF24 solution was added dropwise to 369 the solution of 1 about a minute. The solution color rapidly changed from yellow to orange. The solution was stirred for 2 h at room temperature. The volatiles were then removed in vacuo, and the remaining residue was dissolved in diethyl ether (2 mL) and filtered through Celite. Orange X-ray quality single crystals were produced by chilling a concentrated diethyl ether solution of 3 in a –35 °C freezer overnight (40.0 mg, 22.0% yield). The complex is stable up to 3 d at room temperature and for months at –35 °C. 1H NMR (500 MHz, THF-d8) δ 7.17 (s), 6.90 (s), 6.80 (s), 6.69 (s), 6.17 (s), 5.10 (br), 4.96 (s), 3.52 (s), 3.47, 3.41, 2.99, 2.87, 2.34, 1.73, 1.33, 1.20, 1.19, 1.15, 1.14, 1.12, 0.92, -1.48, -4.34, -17.21, -17.87, -18.97. (br). 13C NMR (126 MHz, THF-d8) δ 160.84, 160.45, 160.03, 159.65, 149.29, 148.68, 144.23, 134.80, 133.03, 130.34, 129.77, 127.64, 127.42, 126.29, 126.10, 124.10, 122.06, 119.74, 117.81, 115.79, 111.53, 67.57, 35.67, 32.81, 31.63, 31.41, 25.51, 23.80, 15.92, 14.68. 19F NMR (470 MHz, THF-d8) δ -66.01. Anal. Cald for C104H112 N2DyBF24: C, 61.85; H, 5.59; N, 1.39. Found: C, 60.89; H, 5.38; N, 1.31. λmax (nm) 1218, 1189, 1048, 863, 781 (at higher concentration), 418, 293 (at lower concentration). Decomposition temperature 94 °C. Synthesis of (NHAr*)2DyNC (7) A 20 mL scintillation vial was charged with 5(12.7 mg, 0.01 mmol, 1 equiv.), n-hexane (1 mL), and a magnetic stir bar. A separate 20 mL scintillation vial was loaded with tert-butylisonitrile (5.1 mg, 0.01 mmol, 1 equiv.) and n-hexane (1 mL). Both solutions were cooled in a dry ice/acetone cold well for 20 min, and then the CNtBu solution was added dropwise into the solution of 5. The solution color changed rapidly from black to pale yellow. The reaction mixture was allowed to stir [Dy(NHAr*)2][Dy(NHAr*)2P4] for 30 min. The solvent was removed in vacuo. The resulting yellow solid was extracted with n-hexane and filtered through Celite. The solvent was removed in vacuo. A concentrated solution of the product in n-hexane resulted in the formation of yellow- colored X-ray quality single crystals overnight at –35 °C in the freezer (4.2 mg, 0.003 mmol, 34% 370 yield). Anal. Calcd for C73H100N3Dy: C, 74.22; H, 8.53; N, 3.56. Found: C, 73.95; H, 8.72; N, 3.47. Decomposition temperature 98 °C. λmax (nm) 1439, 1402, 1316, 1249, 1192 (at higher concentrations), 297 (at lower concentrations). νNC = 2052 cm-1. Synthesis of [Dy(NHAr*)2][Dy(NHAr*)2P4] (8) A 20 mL scintillation vial was charged with 5 (31.5 mg, 0.025 mmol, 1 equiv.), benzene (2 mL), and a stir bar and cooled in a –35 °C freezer for 15 mins. A separate 20 mL scintillation vial was loaded with white phosphorous (1.5 mg, 0.012 mmol, 0.5 equiv.) and benzene (1 mL). To a frozen solution of 2, a solution of white phosphorus was added. The reaction color changed from black to orange after 1 hour of stirring at room temperature. The solvent was removed in vacuo. The orange solid was dissolved in diethyl ether and filtered through celite. A concentrated solution of the product in diethyl ether resulted in the formation of orange-colored X-ray quality single crystals overnight at –35 °C in the freezer (10.8 mg, 0.008 mmol, 29% yield). Anal. Calcd for C144H200N4Dy2P4: C, 70.96; H, 8.21; N, 2.29. Found: C, 70.86; H, 8.69; N, 2.23. λmax (nm) (at higher concentrations), 297 (at lower concentrations). 371 NMR Spectra of Dysprosium Complexes Figure 6.17. 1H NMR spectrum of (NHAr*)2DyCl (4) in C6D6 at room temperature. 372 Figure 6.18. 1H NMR spectrum of [(NHAr*)2Dy][BArF24] (2) in THF-d8 at room temperature. 373 Figure 6.19. 1H NMR spectrum (zoom in) of [(NHAr*)2Dy][BArF24] (2) in THF-d8 at room temperature. 374 Figure 6.20. 13C NMR spectrum of [(NHAr*)2Dy][BArF24] (2) in THF-d8 at room temperature. 375 Figure 6.21. 19F NMR spectrum of [(NHAr*)2Dy][BArF24] (2) in THF-d8 at room temperature. 376 UV-Vis-Near IR Spectra The electronic absorption spectra were recorded on a double-beam PerkinElmer 1050 spectrophotometer. The measurement was done in a 1 cm pathlength quartz cell. Preparation of samples was performed in the N2 glovebox using dry ether. The raw data were fitted with OriginPro 9.0 software to obtain accurate maxima assuming Gaussian peak shapes. Low concentrations solutions of complexes 4-8 were monitored to observe charge transfer bands. UV-Vis spectra were collected using an Ocean Optics DH-mini UV-Vis spectrophotometer in an N2 glovebox. To determine the molar extinction coefficients for charge transfer bands, a series of solutions were prepared with varying concentrations through serial dilutions. All spectra were baseline corrected for diethyl ether and collected at ambient temperature. . Figure 6.22. Vis-NIR spectrum of (NHAr*)2DyCl (4) obtained from 25 μM solution in diethyl ether. 377 Figure 6.23. UV-Vis spectrum of (NHAr*)2DyCl (4) obtained from serial dilution of 100 μM solution. Figure 6.24. Concentration vs absorbance plot to calculate molar extinction coefficient at 297 nm for (NHAr*)2DyCl (4) in diethyl ether. The slope of the line provides ε = 1108.1 M−1 cm−1 for the electronic transition occurring at 297 nm. 378 Figure 6.25. Vis-NIR spectrum of Dy(NHAr*)2 (5) obtained from 1 mM solution in ether. The values of molar absorptivity are approximated from one measurement. Figure 6.26. UV-vis spectrum of Dy(NHAr*)2 (5) obtained from serial dilutions. 379 Figure 6.27. Concentration vs absorbance plot to calculate molar extinction coefficient at 297 nm for Dy(NHAr*)2 (5) in diethyl ether. The slope of the line provides ε = 19307 M-1 cm-1. Figure 6.28. Vis-near IR spectrum of [Dy(NHAr*)2][BArF24] (6) obtained from 4.26 mM solution in ether. Values of molar absorptivity are approximated from one measurement. 380 Figure 6.29. UV-Vis spectrum of [(NHAr*)2Dy][BArF24] (6) obtained from serial dilution of 570 μM solution. Figure 6.30. Concentration vs absorbance plot to calculate molar extinction coefficient at 415 nm for [(NHAr*)2Dy][BArF24] (6) in diethyl ether. The slope of the line provides ε = 4618 M−1 cm−1 for the electronic transition occurring at 415 nm. 381 Figure 6.31. Vis-NIR spectrum of Dy(NHAr*)2NC (7) obtained from 7.4 mM solution in ether. The values of molar absorptivity are approximated from one measurement. Figure 6.32. UV-vis spectrum of Dy(NHAr*)2NC(3) obtained from serial dilutions. 382 Figure 6.33. Concentration vs absorbance plot to calculate molar extinction coefficient at 297 nm for Dy(NHAr*)2NC (3) in diethyl ether. The slope of the line provides ε = 7518 M-1 cm-1. Figure 6.34. Vis-NIR spectrum [Dy(NHAr*)2][Dy(NHAr*)2P4] (8) obtained from 7.4 mM solution in ether. The values of molar absorptivity are approximated from one measurement. 383 Figure 6.35. UV-vis spectrum of [Dy(NHAr*)2][Dy(NHAr*)2P4] (4) obtained from serial dilutions. Figure 6.36. Concentration vs absorbance plot to calculate molar extinction coefficient at 297 nm for [Dy(NHAr*)2][Dy(NHAr*)2P4] (8) in diethyl ether. The slope of the line provides ε = 40425 M-1 cm-1. 384 Figure 6.37. Concentration vs absorbance plot to calculate molar extinction coefficient at 400 nm for [Dy(NHAr*)2][Dy(NHAr*)2P4] (8) in diethyl ether. The slope of the line provides ε = 18687 M-1 cm-1. Solution-State Magnetism (Evans’ Method) Solution state magnetic susceptibility studies on Dy(NHAr*)2Cl (4) and Dy(NHAr*)2 (5) were conducted using the Evans method. Since we have Dy(III) center in Dy(NHAr*)2Cl (4), [Dy(NHAr*)2][BArF24] (6), Dy(NHAr*)2NC (7), [Dy(NHAr*)2][Dy(NHAr*)2P4] (8) and solution- state magnetism was only measured for Dy(NHAr*)2Cl (4). The temperature-dependent paramagnetism was measured in 10 K intervals from 298-181 K. In a Teflon capped J-Young NMR tube 23 mM sample of 1 and 2 in toluene-d8, and hexamethyldisiloxane was transferred with a sealed capillary tube containing 20 mM hexamethyldisiloxane in toluene-d8 as a reference. The difference between HMDSO peak value with and without paramagnetic species was analyzed to measure the effective magnetic moment of complexes 4 and 5. 385 Figure 6.38. Temperature dependence of solution state effective magnetic moment for complex Dy(NHAr*)2Cl (4). Figure 6.39. Temperature dependence of solution state effective magnetic moment for complex Dy(NHAr*)2 (5). 386 FT-IR Spectroscopy IR Spectra were recorded using Varian 3100 FT-IR spectrometer with a spectral resolution of 2 cm-1 using a CaF2 air-free cell. Sample preparation was done in the glove box under an N2 atmosphere and sealed with Teflon caps before measurements. The spectrum was recorded in the solution phase using dry n-hexane as solvent. The IR transmittance spectra were collected at 298 K and the baseline was corrected. The νNC stretch for Dy(NHAr*)2NC (7) is 2051.9 cm-1 lies in the range of previously reported value metal isocyanides.18,28,29,47 Figure 6.40. FT-IR spectrum for complex Dy(NHAr*)2 (4) in the solid state. 387 Figure 6.41. FT-IR spectrum for complex Dy(NHAr*)2 (5) in the solid state. Figure 6.42. FT-IR spectrum for complex [(NHAr*)2Dy][BArF24] (6) in the solid state. 388 Figure 6.43. FT-IR spectrum for complex Dy(NHAr*)2NC (7) in the solid state. Figure 6.44. Zoomed in FT-IR spectrum for complex Dy(NHAr*)2NC (7) (⁓5 μM) in n-hexane. 389 Figure 6.45. FT-IR spectrum for complex [Dy(NHAr*)2][Dy(NHAr*)2P4] 8 in the solid state. Cyclic Voltammetry Cyclic voltammetry was recorded using a PGSTAT204 potentiostat from Metrohm with a glassy carbon working electrode, platinum wire pseudo reference electrode, and platinum wire as a counter electrode. All measurements were in the glovebox under argon atmosphere. Dy(NHAr*)2 5 (1.2 mM) was dissolved in diethyl ether with [NBu4]+[B(3,5-(CF3)2C6H3)4]– (100 mM) as supporting electrolyte. Cyclic voltammetry of ferrocene was performed six times to get standard deviation in potential shifts. The reduction potential for Fc+/Fc0 redox couple was observed at 1.14 ± 0.4 V. All voltammograms were externally referenced to ferrocene with the same electrolyte concentration. All measurements were done in triplicate to get an average value of half-cell potential. The redox potential for Dy(II)/Dy(III) couple is 1.089 ± 0.001 V with respect to Fc+/Fc0 couple. The reversibility of this one-electron redox process was studied at different scan rates from 90 mV/s to 60 mV/s with 10 mV/s intervals. 390 Figure 6.46. Overlay plot for scan rates in diethyl ether with [NBu4]+[B(3,5-(CF3)2C6H3)4]– (100 mM) as supporting electrolyte and Fc+/0 as 0 V. Single Crystal X-ray Diffraction Single crystal data was collected on XtaLAB Synergy, Dualflex, Hypix diffractometer using CuKα or MoKα radiation. Data collection was done at 100 K under a continuous flow of liquid nitrogen. In Olex2 program,48 crystal structures were solved with ShelXT solution using intrinsic phasing and refined with the SheXL refinement package using least squares minimization.49 All hydrogens are refined anisotropically. All crystals were stable at room temperature for mounting. Complex Dy(NHAr*)2Cl (1) shows a “B Alert” due to a residual peak between metal and chloride. This is due to the presence of heavy metal ion and therefore cannot be removed after refinement. Another B Alert is present in the Dy(NHAr*)2NC (4) checkcif file due to a terminal carbon in isocyanide. Thermal ellipsoids are drawn with a 50% probability level. Dark green, green, blue, gray, and 391 white spheres represent dysprosium, chlorine, nitrogen, carbon, and hydrogen atoms respectively. Figure 6.47. Structure of Dy(NHAr*)2Cl (4) recrystallized from n-hexane. One molecule of n- hexane is in the lattice per molecule of 4. 392 Table 6.1. Metric data from the crystal structures of Dy(NHAr*)2Cl (4). Empirical formula Formula weight Temperature/K Crystal system Space group a/Å b/Å c/Å α/° β/° γ/° Volume/Å3 Z ρcalcg/cm3 μ/mm-1 F(000) Crystal size/mm3 Radiation 2Θ range for data collection/° Index ranges Reflections collected Independent reflections Data/restraints/parameters Goodness-of-fit on F2 Final R indexes [I>=2σ (I)] Final R indexes [all data] Largest diff. peak/hole / e Å-3 C78H114ClDyN2 1277.66 99.96(16) triclinic P-1 9.54318(14) 16.7779(2) 22.9544(3) 88.7134(10) 78.3872(12) 73.9747(13) 3458.07(9) 2 1.227 6.441 1358.0 0.107 × 0.103 × 0.04 Cu Kα (λ = 1.54184) 5.484 to 153.926 -11 ≤ h ≤ 12, -21 ≤ k ≤ 19, -28 ≤ l ≤ 25 40403 13657 [Rint = 0.0519, Rsigma = 0.0592] 13657/28/782 1.124 R1 = 0.0512, wR2 = 0.1318 R1 = 0.0545, wR2 = 0.1335 2.31/-0.79 393 Figure 6.48. Asymmetric unit of Dy(NHAr*)2 (5) recrystallized from n-hexane. There is one molecule of disordered THF in the lattice per molecule. 394 Table 6.2. Metric data from the crystal structures of Dy(NHAr*)2 (5). Empirical formula Formula weight Temperature/K Crystal system Space group a/Å b/Å c/Å α/° β/° γ/° Volume/Å3 Z ρcalcg/cm3 μ/mm-1 F(000) Crystal size/mm3 Radiation 2Θ range for data collection/° Index ranges Reflections collected Independent reflections Data/restraints/parameters Goodness-of-fit on F2 Final R indexes [I>=2σ (I)] Final R indexes [all data] Largest diff. peak/hole / e Å-3 C76H108DyN2O 1228.14 100.0(2) monoclinic C2/c 17.9866(4) 17.0426(3) 22.7758(5) 90 106.132(2) 90 6706.7(2) 4 1.216 1.158 2608.0 0.144 × 0.079 × 0.063 Mo Kα (λ = 0.71073) 4.5 to 61.774 -24 ≤ h ≤ 24, -22 ≤ k ≤ 24, -31 ≤ l ≤ 30 46009 8982 [Rint = 0.0658, Rsigma = 0.0529] 8982/54/391 1.042 R1 = 0.0398, wR2 = 0.1016 R1 = 0.0518, wR2 = 0.1070 1.43/-0.76 395 Figure 6.49. The asymmetric unit of [Dy(NHAr*)2][BArF24] (6), recrystallized from n-hexane. 396 Table 6.3. Metric data from the crystal structures of [Dy(NHAr*)2][BArF24] (6). Empirical formula Formula weight Temperature/K Crystal system Space group a/Å b/Å c/Å α/° β/° γ/° Volume/Å3 Z ρcalcg/cm3 μ/mm-1 F(000) Crystal size/mm3 Radiation 2Θ range for data collection/° Index ranges Reflections collected Independent reflections Data/restraints/parameters Goodness-of-fit on F2 Final R indexes [I>=2σ (I)] Final R indexes [all data] Largest diff. peak/hole / e Å-3 Flack parameter C104H112BDyF24N2 2019.26 99.99(10) monoclinic C2 21.2676(3) 15.6856(2) 14.7701(2) 90 92.7530(10) 90 4921.55(11) 2 1.363 0.851 2074.0 0.135 × 0.098 × 0.086 Mo Kα (λ = 0.71073) 4.188 to 61.866 -29 ≤ h ≤ 29, -21 ≤ k ≤ 21, -20 ≤ l ≤ 20 44781 12201 [Rint = 0.0299, Rsigma = 0.0293] 12201/1/617 1.033 R1 = 0.0252, wR2 = 0.0613 R1 = 0.0254, wR2 = 0.0614 0.95/-0.60 0.041(5) 397 Figure 6.50. Structure of Dy(NHAr*)2NC (7) recrystallized from n-hexane. Half a molecule of n- hexane is in the lattice per molecule of 7. 398 Table 6.4. Metric data from the crystal structures of Dy(NHAr*)2NC (7). Empirical formula Formula weight Temperature/K Crystal system Space group a/Å b/Å c/Å α/° β/° γ/° Volume/Å3 Z ρcalcg/cm3 μ/mm-1 F(000) Crystal size/mm3 Radiation 2Θ range for data collection/° Index ranges Reflections collected Independent reflections Data/restraints/parameters Goodness-of-fit on F2 Final R indexes [I>=2σ (I)] Final R indexes [all data] Largest diff. peak/hole / e Å-3 C76H107DyN3 1225.14 100(2) monoclinic P21/n 14.89576(12) 23.41339(13) 19.38462(16) 90 97.9385(7) 90 6695.80(9) 4 1.215 6.280 2600.0 0.202 × 0.062 × 0.03 CuKα (λ = 1.54184) 5.954 to 160.388 -19 ≤ h ≤ 18, -22 ≤ k ≤ 29, -24 ≤ l ≤ 24 59167 14316 [Rint = 0.0371, Rsigma = 0.0334] 14316/0/755 1.049 R1 = 0.0364, wR2 = 0.0864 R1 = 0.0407, wR2 = 0.0884 0.97/-0.86 Structural Analysis of [Dy(NHAr*)2][Dy(NHAr*)2P4] (4) A metallic intense orange cube-shaped crystal with dimensions 0.26×0.14×0.11 mm3 was mounted on a nylon loop with paratone oil. Data were collected using a XtaLAB Synergy, Dualflex, HyPix diffractometer equipped with an Oxford Cryosystems 800 low-temperature device, operating at T = 100.00(10) K. Data were measured using w scans using Cu Ka radiation (micro-focus sealed X-ray tube, 50 kV, 1 mA). The total number of runs and images was based on the strategy calculation from the 399 program CrysAlisPro system (CCD 43.119a 64-bit (release 08-04-2024)). The achieved resolution was Q = 80.287. Cell parameters were retrieved using the CrysAlisPro 1.171.43.119a (Rigaku OD, 2024) software and refined using CrysAlisPro 1.171.43.119a (Rigaku OD, 2024) on 23351 reflections, 22% of the observed reflections. Data reduction was performed using the CrysAlisPro 1.171.43.119a (Rigaku OD, 2024) software, which corrects for Lorentz polarization. The final completeness is 100.00 out to 80.287 in Q CrysAlisPro 1.171.43.119a (Rigaku Oxford Diffraction, 2024). Numerical absorption correction is based on Gaussian integration over a multifaceted crystal model. Empirical absorption correction using spherical harmonics is implemented in the SCALE3 ABSPACK scaling algorithm. The structure was solved in the space group Pca21 (# 29) by using dual methods using the ShelXT 2018/2 structure solution program.49 The structure was refined by Least Squares ShelXL incorporated in the Olex2 software program.48 All non-hydrogen atoms were refined anisotropically. Hydrogen atom positions were calculated geometrically and refined using the riding model. The ether molecules were not clearly identified by the residual electron density. Program BYPASS in Olex2 suggested there are three ether molecules per asymmetric unit cell. Therefore, we did our best to model and refine these molecules isotropically. Once in location, hydrogens were added, and the molecules were restrained with the AFIX 1 command, as these molecules would not converge. Using the BYPASS method in Olex2, we calculated the solvent as three ether molecules per asymmetric unit cell. _smtbx_masks_void_probe_radius 1.2 _smtbx_masks_void_truncation_radius 1.2 loop_ 400 _smtbx_masks_void_nr _smtbx_masks_void_average_x _smtbx_masks_void_average_y _smtbx_masks_void_average_z _smtbx_masks_void_volume _smtbx_masks_void_count_electrons _smtbx_masks_void_content 1 -0.192 -0.417 0.769 869.9 150.8 '3 C5H10O' 2 0.192 -0.416 0.269 869.9 143.7 '3 C5H10O' 3 0.308 -0.597 0.769 869.9 150.8 '3 C5H10O' 4 0.692 -0.289 0.269 869.9 143.7 '3 C5H10O' 401 Figure 6.51. Structure of [Dy(NHAr*)2][Dy(NHAr*)2P4] 8 recrystallized from diethyl ether. Three molecules of diethyl ether are in the lattice per molecule of 8. This is the best of crystals grown from various solvents and various data collections. All crystals had poor diffraction, and they must be used immediately once removed from the glove box in oil. This crystal was a racemic twin and showed significant disorder. Initial trials attempted to put the two P4 units in a butterfly orientation, but this proved to be very unsuccessful. The Dy atom was modeled disordered, and the occupancy was refined to that of the P4 unit best, where the largest occupancy went with the P4 in the butterfly orientation. These were then locked as one free variable (70:30). It was anticipated that the rest of the ligand would have some portion of disorder, but so far attempts to model this molecule closer to a full molecule disorder have failed, leaving what may be some distorted Dy‒N bond lengths. 402 There is a single formula unit in the asymmetric unit, which is represented by the reported sum formula. In other words, Z is 4, and Z' is 1. The moiety formula is C72H100DyN2P4, C72H100DyN2, 3(C4H10O). The Flack parameter was refined to 0.513(12). Determination of absolute structure using Bayesian statistics on Bijvoet differences using the Olex2 results in 0.043(4). The chiral atoms in this structure are: C7A(R), C8A(R), C9A(R), C11A(S), C12A(S), C43A(S), C44A(R), C45A(R), C47A(S), C48A(S). Note: The Flack parameter is used to determine the chirality of the crystal studied, and the value should be near 0. A value of 1 means that the stereochemistry is wrong, and the model should be inverted. A value of 0.5 means that the crystal consists of a racemic mixture of the two enantiomers. 403 Table 6.5. Metric data from the crystal structures of [Dy(NHAr*)2][Dy(NHAr*)2P4] (8) Compound Formula Dcalc./ g cm-3 m/mm-1 Formula Weight Color Shape Size/mm3 T/K Crystal System Flack Parameter Hooft Parameter Space Group a/Å b/Å c/Å a/° b/° g/° V/Å3 Z Z' Wavelength/Å Radiation type Qmin/° Qmax/° Measured Refl's. Indep't Refl's Refl's I≥2 s(I) Rint Parameters Restraints Largest Peak Deepest Hole GooF wR2 (all data) wR2 R1 (all data) R1 [Dy(NHAr*)2][Dy(NHAr*)2P4] C156H230Dy2N4O3P4 1.169 6.006 2658.31 metallic intense orange cube-shaped 0.26×0.14×0.11 100.00(10) orthorhombic 0.513(12) 0.043(4) Pca21 25.2032(3) 21.9756(3) 27.2612(4) 90 90 90 15098.8(4) 4 1 1.54184 Cu Ka 2.668 80.287 108558 28739 19142 0.0938 1419 604 1.844 -2.125 1.080 0.2785 0.2427 0.1207 0.0857 404 DFT Calculations DFT calculations were carried out using Gaussian 16 (B01 and C01).50,51 The starting coordinates for the geometry optimizations were taken from the structure found from X-ray diffraction. Geometry optimization and frequency calculations were performed using the TPSSH hybrid functional52,53 and def2-SVP basis set54 on C, N, H, and ECP55MWB pseudopotential55,56 on the dysprosium atom with Grimme’s dispersion correction GD3.57,58 The colors Turquoise, blue, grey, and white spheres represent dysprosium, nitrogen, carbon, and hydrogen atoms, respectively. First, the DFT calculations were performed Dy(NHAr*)2 (2) to identify the distribution of the unpaired electron density. The comparison of optimized bond lengths and bond angles is shown below. The theoretically obtained structure is consistent with the crystal structure. Further, the HOMO, LUMO, and electron charge density map looks similar to our previously reported Y(II) complex.18 This suggests that the ground state electronic configuration for 2 is 4f95d1, where the unpaired electron is delocalized majorly over the arene system. Table 6.6. Structural comparisons between crystallographically obtained geometry of 2 and optimized geometry using DFT calculations. Shown here are some characteristic lengths (Å) and angles (˚). Atoms Dy‒N Dy-ArCent Cent‒Dy‒Cent Dy‒N‒Cipso Arcent‒Dy‒N Average CAr‒ CAr bond Experimental Distance (Å)/Angle (˚) 2.275(2) 2.460(1) 134.399(7 130.38(16) 95.48(5) 113.07(5) 1.411(4) Theoretical Distance (Å)/Angle (˚) 2.316 2.476 135.17 130.01 95.40 112.31 1.423 The preference for the formation of dysprosium isocyanide over hypothetical cyanide is 405 discussed here. The structure obtained from DFT calculations is consistent with the single-crystal structure 3 (Table S6). The optimized geometries of 3 and 3’ are shown below. Figure 6.52. Optimized structures of Dy(NHAr*)2NC (3) (left) and Dy(NHAr*)2CN (3’) (right). All hydrogens except N‒H hydrogens are removed for clarity. Table 6.7. Structural comparisons between crystallographically obtained geometry of Dy(NHAr*)2NC (3) and optimized geometry using DFT calculations. Shown here are some characteristic bond lengths (Å) and angles (˚). Atoms Dy‒N2, N3 Dy-N1 Dy-ArCent N1‒C1 Dy‒N1‒C1 Dy‒N‒Cipso Experimental Distance (Å)/Angle (˚) 2.2182(19) 2.2366(19) 2.372(2) 2.5178(2) 1.025(5) 174.6 147.73(16) 133.10(15) Theoretical Distance (Å)/Angle (˚) 2.27562 2.28537 2.332 2.570 1.184 169.86 136.97 134.68 The ground energy difference between Dy(NHAr*)2NC (3) and hypothetical Dy(NHAr*)2CN (3’) is ⁓2.5 kcal/mol. Further frequency calculation gave us νNC stretch for 3 at 2131.5 cm-1, which 406 is relatively I n better agreement with the experimental value (2052 cm-1). The calculated νNC for the hypothetical Dy(NHAr*)2CN (3’) is 2217.6 cm-1 higher than isocyanide due to the lone pair bond weakening effect discussed in our previous study.18 Additionally, ground state energy calculations were performed on [(NHAr*)2Dy–P4]- to confirm the charge distribution. The [(NHAr*)2Dy]+ was excluded from the calculations to improve computational efficiency and convergence time. The optimized geometry for butterfly and pseudo planar dysprosium P4 2- structures are shown in Figure 6.53. The optimized geometry is in close approximation with the crystal structure. Figure 6.53. Optimized structures of butterfly-[(NHAr*)2Dy–P4]- (left) and pseudo-planar [(NHAr*)2Dy–P4]- (right). All hydrogens except N‒H hydrogens are removed for clarity. 407 Table 6.8. Structural comparisons between crystallographically obtained geometry of butterfly- and pseudo planar-[(NHAr*)2Dy–P4]-, and optimized geometry using DFT calculations. Shown here are some characteristic bond lengths (Å) and angles (˚). Bond Experimenta l Lengths (Å)/ Angles (°) Dy‒N1 Dy‒N2 Dy‒P2 Dy‒P4 Dy‒P1 Dy‒P3 P2‒P3 P3‒P4 P4‒P1 P1‒P2 P1‒P3 P2‒P4 Dy‒Cent P3‒P2‒P1 P3‒P4‒P1 P4‒P3‒P2 P2‒P1‒P4 N1‒Dy‒N2 2- Butterfly P4 2.190 (13) 2.213(14) 2.811(11) 2.659(18) 3.279(15) 3.320(7) 2.182(13) 2.17(2) 2.182(19) 2.112(16) 2.296(9) 3.21(2) 3.638(12), 3.304(12) 64.6(5) 63.7(6) 95.2(5) 96.9(6) 136.5(5) Theoretical 2- Butterfly P4 Experimental 2- Pseudo planar P4 Theoretical Pseudo planar P4 2- 2.537(16) 2.254(12) 2.965(12) 3.01(3) 3.01(4) 3.006(16) 2.13(2) 2.12(3) 2.14(4) 2.01 (3) 2.45(3) 3.37(13) 4.109(12), 2.984(11) 72.4(10) 70.1(11) 104.9(10) 108.3(14) 117.1(5) 2.331 2.2331 2.989 2.994 2.966 2.95 2.17 2.18 2.17 2.18 3.07 3.06 3.96, 2.88 89.9 90.1 89.9 90.1 123.3 2.319 2.315 2.779 2.775 3.467 3.466 2.27 2.27 2.27 2.27 2.16 3.36 3.513 3.485 56.67 56.69 95.38 95.43 128.20 408 REFERENCES (1) Benelli, C.; Gatteschi, D. Introduction to Molecular Magnetism: From Transition Metals to Lanthanides, 1st ed.https://doi.org/10.1002/9783527690541; Wiley, 2015. (2) Coronado, E. Molecular Magnetism: From Chemical Design to Spin Control in Molecules, Materials and Devices. Nat Rev Mater 2019, 5, 87–104. (3) Moreno-Pineda, E.; Wernsdorfer, W. Measuring Molecular Magnets for Quantum Technologies. Nat Rev Phys 2021, 3, 645–659. (4) Pei, T.; Thomas, J. O.; Sopp, S.; Tsang, M.-Y.; Dotti, N.; Baugh, J.; Chilton, N. F.; Cardona- Serra, S.; Gaita-Ariño, A.; Anderson, H. L.; Bogani, L. Exchange-Induced Spin Polarization in a Single Magnetic Molecule Junction. Nat Commun 2022, 13, 4506. (5) Kurzen, H.; Bovigny, L.; Bulloni, C.; Daul, C. Electronic Structure and Magnetic Properties of Lanthanide 3+ Cations. Chemical Physics Letters 2013, 574, 129–132. (6) Guo, F.-S.; Day, B. M.; Chen, Y.-C.; Tong, M.-L.; Mansikkamäki, A.; Layfield, R. A. Magnetic Hysteresis up to 80 Kelvin in a Dysprosium Metallocene Single-Molecule Magnet. Science 2018, 362, 1400–1403. (7) Goodwin, C. A. P.; Ortu, F.; Reta, D.; Chilton, N. F.; Mills, D. P. Molecular Magnetic Hysteresis at 60 Kelvin in Dysprosocenium. Nature 2017, 548, 439–442. (8) Chilton, N. F.; Goodwin, C. A. P.; Mills, D. P.; Winpenny, R. E. P. The First Near-Linear Bis(Amide) f-Block Complex: A Blueprint for a High Temperature Single Molecule Magnet. Chem. Commun. 2015, 51, 101–103. (9) Harriman, K. L. M.; Brosmer, J. L.; Ungur, L.; Diaconescu, P. L.; Murugesu, M. Pursuit of Record Breaking Energy Barriers: A Study of Magnetic Axiality in Diamide Ligated Dy III Single-Molecule Magnets. J. Am. Chem. Soc. 2017, 139, 1420–1423. (10) Emerson-King, J.; Gransbury, G. K.; Whitehead, G. F. S.; Vitorica-Yrezabal, I. J.; Rouzières, M.; Clérac, R.; Chilton, N. F.; Mills, D. P. Isolation of a Bent Dysprosium Bis(Amide) Single- Molecule Magnet. J. Am. Chem. Soc. 2024, jacs.3c12427. (11) Liu, S.-S.; Ziller, J. W.; Zhang, Y.-Q.; Wang, B.-W.; Evans, W. J.; Gao, S. A Half-Sandwich Organometallic Single-Ion Magnet with Hexamethylbenzene Coordinated to the Dy( iii ) Ion. Chem. Commun. 2014, 50, 11418–11420. (12) Cotton, S. A. Scandium, Yttrium & the Lanthanides: Inorganic & Coordination Chemistry. In Encyclopedia of Inorganic Chemistry; King, R. B., Crabtree, R. H., Lukehart, C. M., Atwood, D. A., Scott, R. A., Eds.; Wiley, 2005. (13) Meng, Y.; Xu, L.; Xiong, J.; Yuan, Q.; Liu, T.; Wang, B.; Gao, S. Low‐Coordinate Single‐ Ion Magnets by Intercalation of Lanthanides into a Phenol Matrix. Angew Chem Int Ed 2018, 409 57, 4673–4676. (14) Billow, B. S.; Livesay, B. N.; Mokhtarzadeh, C. C.; McCracken, J.; Shores, M. P.; Boncella, J. M.; Odom, A. L. Synthesis and Characterization of a Neutral U(II) Arene Sandwich Complex. J. Am. Chem. Soc. 2018, 140, 17369–17373. (15) Guo, F.-S.; He, M.; Huang, G.-Z.; Giblin, S. R.; Billington, D.; Heinemann, F. W.; Tong, M.- L.; Mansikkamäki, A.; Layfield, R. A. Discovery of a Dysprosium Metallocene Single- Molecule Magnet with Two High-Temperature Orbach Processes. Inorg. Chem. 2022, 61, 6017–6025. (16) Mugiraneza, S.; Hallas, A. M. Tutorial: A Beginner’s Guide to Interpreting Magnetic Susceptibility Data with the Curie-Weiss Law. Commun Phys 2022, 5, 95. (17) Harriman, K. L. M.; Murillo, J.; Suturina, E. A.; Fortier, S.; Murugesu, M. Relaxation Dynamics in See-Saw Shaped Dy( iii ) Single-Molecule Magnets. Inorg. Chem. Front. 2020, 7, 4805–4812. (18) Jena, R.; Benner, F.; Delano, F.; Holmes, D.; McCracken, J.; Demir, S.; Odom, A. L. A Rare Isocyanide Derived from an Unprecedented Neutral Yttrium( ii ) Bis(Amide) Complex. Chem. Sci. 2023, 14, 4257–4264. (19) Evans, W. J. Tutorial on the Role of Cyclopentadienyl Ligands in the Discovery of Molecular Complexes of the Rare-Earth and Actinide Metals in New Oxidation States. Organometallics 2016, 35, 3088–3100. (20) Jaroschik, F.; Nief, F.; Le Goff, X.-F.; Ricard, L. Isolation of Stable Organodysprosium(II) Complexes by Chemical Reduction of Dysprosium(III) Precursors. Organometallics 2007, 26, 1123–1125. (21) Fieser, M. E.; MacDonald, M. R.; Krull, B. T.; Bates, J. E.; Ziller, J. W.; Furche, F.; Evans, W. J. Structural, Spectroscopic, and Theoretical Comparison of Traditional vs Recently Discovered Ln 2+ Ions in the [K(2.2.2-Cryptand)][(C 5 H 4 SiMe 3 ) 3 Ln] Complexes: The Variable Nature of Dy 2+ and Nd 2+. J. Am. Chem. Soc. 2015, 137, 369–382. (22) Gould, C. A.; McClain, K. R.; Yu, J. M.; Groshens, T. J.; Furche, F.; Harvey, B. G.; Long, J. R. Synthesis and Magnetism of Neutral, Linear Metallocene Complexes of Terbium(II) and Dysprosium(II). J. Am. Chem. Soc. 2019, 141, 12967–12973. (23) MacDonald, M. R.; Bates, J. E.; Fieser, M. E.; Ziller, J. W.; Furche, F.; Evans, W. J. Expanding Rare-Earth Oxidation State Chemistry to Molecular Complexes of Holmium(II) and Erbium(II). J. Am. Chem. Soc. 2012, 134, 8420–8423. (24) Jin, P.-B.; Luo, Q.-C.; Gransbury, G. K.; Vitorica-Yrezabal, I. J.; Hajdu, T.; Strashnov, I.; McInnes, E. J. L.; Winpenny, R. E. P.; Chilton, N. F.; Mills, D. P.; Zheng, Y.-Z. Thermally Stable Terbium(II) and Dysprosium(II) Bis-Amidinate Complexes. J. Am. Chem. Soc. 2023, 145, 27993–28009. 410 (25) Bochkarev, M. N.; Fagin, A. A. A New Route to Neodymium(II) and Dysprosium(II) Iodides. 1999, 5. (26) Anderson, D. M.; Cloke, F. G. N.; Cox, P. A.; Edelstein, N.; Green, J. C.; Pang, T.; Sameh, A. A.; Shalimoff, G. On the Stability and Bonding in Bis(η-Arene)Ianthanide Complexes. J. Chem. Soc., Chem. Commun. 1989, No. 1, 53–55. (27) McClain, K. R.; Gould, C. A.; Marchiori, D. A.; Kwon, H.; Nguyen, T. T.; Rosenkoetter, K. E.; Kuzmina, D.; Tuna, F.; Britt, R. D.; Long, J. R.; Harvey, B. G. Divalent Lanthanide Metallocene Complexes with a Linear Coordination Geometry and Pronounced 6s–5d Orbital Mixing. J. Am. Chem. Soc. 2022, 144, 22193–22201. (28) Chen, X.; Li, Q.; Gong, Y.; Andrews, L.; Liebov, B. K.; Fang, Z.; Dixon, D. A. Formation and Characterization of Homoleptic Thorium Isocyanide Complexes. Inorg. Chem. 2017, 56, 5060–5068. (29) Bouzidi, Y.; Belkhiri, L.; Ephritikhine, M.; Halet, J.-F.; Boucekkine, A. Cyanide Linkage Isomerism in Cerium(III) and Uranium(III) Complexes. A Relativistic DFT Study. Journal of Organometallic Chemistry 2017, 847, 82–89. (30) Garner, M. E.; Hohloch, S.; Maron, L.; Arnold, J. Carbon–Nitrogen Bond Cleavage by a Thorium‐NHC‐bpy Complex. Angew Chem Int Ed 2016, 55, 13789–13792. (31) Hervé, A.; Bouzidi, Y.; Berthet, J.-C.; Belkhiri, L.; Thuéry, P.; Boucekkine, A.; Ephritikhine, M. U–CN versus Ce–NC Coordination in Trivalent Complexes Derived from M[N(SiMe 3 ) 2 ] 3 (M = Ce, U). Inorg. Chem. 2014, 53, 6995–7013. (32) Cummins, C. C. Phosphorus: From the Stars to Land & Sea. Daedalus 2014, 143, 9–20. (33) Gillespie, J. A.; Zuidema, E.; Van Leeuwen, P. W. N. M.; Kamer, P. C. J. Phosphorus Ligand Effects in Homogeneous Catalysis and Rational Catalyst Design. In Phosphorus(III) Ligands in Homogeneous Catalysis: Design and Synthesis; Kamer, P. C. J., Van Leeuwen, P. W. N. M., Eds.; Wiley, 2012; pp 1–26. (34) Ung, S. P.-M.; Li, C.-J. From Rocks to Bioactive Compounds: A Journey through the Global P( v ) Organophosphorus Industry and Its Sustainability. RSC Sustain. 2023, 1, 11–37. (35) Tofan, D.; Cossairt, B. M.; Cummins, C. C. White Phosphorus Activation at a Metal– Phosphorus Triple Bond: A New Route to Cyclo -Triphosphorus or Cyclo -Pentaphosphorus Complexes of Niobium. Inorg. Chem. 2011, 50, 12349–12358. (36) Schwarzmaier, C.; Noor, A.; Glatz, G.; Zabel, M.; Timoshkin, A. Y.; Cossairt, B. M.; Cummins, C. C.; Kempe, R.; Scheer, M. Formation of Cyclo ‐E 4 2− Units (E 4 =P 4 , As 4 , AsP 3 ) by a Complex with a Cr Cr Quintuple Bond. Angew Chem Int Ed 2011, 50, 7283–7286. (37) Riu, M.-L. Y.; Jones, R. L.; Transue, W. J.; Müller, P.; Cummins, C. C. Isolation of an Elusive Phosphatetrahedrane. Sci. Adv. 2020, 6, eaaz3168. 411 (38) Figueroa, J. S.; Cummins, C. C. A Niobaziridine Hydride System for White Phosphorus or Dinitrogen Activation and N- or P-Atom Transfer. Dalton Trans. 2006, No. 18, 2161. (39) Cossairt, B. M.; Cummins, C. C. Radical Synthesis of Trialkyl, Triaryl, Trisilyl and Tristannyl Phosphines from P4. New J. Chem. 2010, 34, 1533. (40) Hauer, S.; Horsley Downie, T. M.; Balázs, G.; Schwedtmann, K.; Weigand, J. J.; Wolf, R. to Cobalt‐Mediated Acylcyanophosphanides. Angew Chem Int Ed 2024, 63, e202317170. Phosphorus: Access Fragmentation of White [3+1] (41) Scherer, O. J.; Winter, R.; Wolmershäuser, G. Niob‐ Und Tantalkomplexe Mit P 2 ‐ Und P 4 ‐Liganden. Zeitschrift anorg allge chemie 1993, 619, 827–835. (42) Clegg, W.; Conway, B.; Kennedy, A. R.; Klett, J.; Mulvey, R. E.; Russo, L. Synthesis and Structures of [(Trimethylsilyl)Methyl]Sodium and ‐potassium with Bi‐ and Tridentate N‐ Donor Ligands. Eur J Inorg Chem 2011, 2011, 721–726. (43) Górski, K.; Mech-Piskorz, J.; Leśniewska, B.; Pietraszkiewicz, O.; Pietraszkiewicz, M. Toward Soluble 5,10-Diheterotruxenes: Synthesis and Reactivity of 5,10-Dioxatruxenes, 5,10-Dithiatruxenes, and 5,10-Diazatruxenes. J. Org. Chem. 2020, 85, 4672–4681. (44) Barnett, B. R.; Mokhtarzadeh, C. C.; Figueroa, J. S.; Lummis, P.; Wang, S.; Queen, J. D.; Gavenonis, J.; Schüwer, N.; Tilley, T. D.; Boynton, J. N.; Power, P. P.; Ditri, T. B.; Weidemann, N.; Barnett, B. R.; Agnew, D. W.; Figueroa, J. S.; Smith, P. W.; Ditri, T. B.; Barnett, B. R.; Carpenter, A. E.; Mokhtarzadeh, C. C.; Agnew, D. W.; Figueroa, J. S.; Smith, P. W.; Pratt, J. K.; Power, P. P.; Mendelson, N. D.; Figueroa, J. S.; Queen, J. D.; Power, P. P.; Agnew, D. W.; Carpenter, A. E.; Figueroa, J. S. TERPHENYL LIGANDS AND COMPLEXES. In Inorganic Syntheses; Power, P. P., Ed.; Wiley, 2018; Vol. 37, pp 85–122. (45) Gavenonis, J.; Tilley, T. D. Synthesis and Reactivity of Alkyl, Hydride, and Silyl Derivatives of the (Terphenyl)Imido Fragments Cp*(Ar Mes N)Ta (Cp* = η 5 -C 5 Me 5 ; Ar Mes = 2,6-(2,4,6- Me 3 C 6 H 2 ) 2 C 6 H 3 ) and Cp*(Ar Trip N)Ta (Ar Trip = 2,6-(2,4,6- i Pr 3 C 6 H 2 ) 2 C 6 H 3 ). Organometallics 2004, 23, 31–43. (46) Park, J. G.; Jeon, I.-R.; Harris, T. D. Electronic Effects of Ligand Substitution on Spin Crossover in a Series of Diiminoquinonoid-Bridged FeII2 Complexes. Inorg. Chem. 2015, 54, 359–369. (47) Tarlton, M. L.; Yu, X.; Ward, R. J.; Kelley, S. P.; Autschbach, J.; Walensky, J. R. Backbonding in Thorium(IV) and Uranium(IV) Diarsenido Complexes with t BuNC and CO. Chemistry A European J 2021, 27, 14396–14400. (48) Dolomanov, O. V.; Bourhis, L. J.; Gildea, R. J.; Howard, J. A. K.; Puschmann, H. OLEX2 : A Complete Structure Solution, Refinement and Analysis Program. J Appl Crystallogr 2009, 42, 339–341. (49) Sheldrick, G. M. Crystal Structure Refinement with SHELXL. Acta Crystallogr C Struct Chem 2015, 71, 3–8. 412 (50) M. J.Frisch, G. W.Trucks, H. B.Schlegel, G. E.Scuseria, M. A.Robb, J. R.Cheeseman, G.Scalmani, V.Barone, G. A.Petersson, H.Nakatsuji, X.Li, M.Caricato, A. V.Marenich, J.Bloino, B. G.Janesko, R.Gomperts, B.Mennucci, H. P.Hratchian, J. V.Ortiz, A. F.Izmaylov, J. L.Sonnenberg, D.WilliamsYoung, F.Ding, F.Lipparini, F.Egidi, J.Goings, B.Peng, A.Petrone, T.Henderson, D.Ranasinghe, V. G.Zakrzewski, J.Gao, N.Rega, G.Zheng, W.Liang, M.Hada, M.Ehara, K.Toyota, R.Fukuda, J.Hasegawa, M.Ishida, T.Nakajima, Y.Honda, O.Kitao, H.Nakai, T.Vreven, K.Throssell, J. A.Montgomery, Jr, J. E.Peralta, F.Ogliaro, M. J.Bearpark, J. J.Heyd, E. N.Brothers, K. N.Kudin, V. N.Staroverov, T. A.Keith, R.Kobayashi, J.Normand, K.Raghavachari, A. P.Rendell, J. C.Burant, S. S.Iyengar, J.Tomasi, M.Cossi, J. M.Millam, M.Klene, C.Adamo, R.Cammi, J. W.Ochterski, R. L.Martin, K.Morokuma, O.Farkas, J. B.Foresman and D. J.Fox, Gaussian Program Suite (Revision B01), 2016. (51) Gaussian 16, Revision C.01, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, G. A. Petersson, H. Nakatsuji, X. Li, M. Caricato, A. V. Marenich, J. Bloino, B. G. Janesko, R. Gomperts, B. Mennucci, H. P. Hratchian, J. V. Ortiz, A. F. Izmaylov, J. L. Sonnenberg, D. Williams-Young, F. Ding, F. Lipparini, F. Egidi, J. Goings, B. Peng, A. Petrone, T. Henderson, D. Ranasinghe, V. G. Zakrzewski, J. Gao, N. Rega, G. Zheng, W. Liang, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, K. Throssell, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. J. Bearpark, J. J. Heyd, E. N. Brothers, K. N. Kudin, V. N. Staroverov, T. A. Keith, R. Kobayashi, J. Normand, K. Raghavachari, A. P. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, J. M. Millam, M. Klene, C. Adamo, R. Cammi, J. W. Ochterski, R. L. Martin, K. Morokuma, O. Farkas, J. B. Foresman, and D. J. Fox, Gaussian, Inc., Wallingford CT, 2019.Sonnenberg, D. Williams-Young, F. Ding, F. Lipparini, F. Egidi, J. Goings, B. Peng, A. Petrone, T. Henderson, D. Ranasinghe, V. G. Zakrzewski, J. Gao, N. Rega, G. Zheng, W. Liang, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, K. Throssell, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. J. Bearpark, J. J. Heyd, E. N. Brothers, K. N. Kudin, V. N. Staroverov, T. A. Keith, R. Kobayashi, J. Normand, K. Raghavachari, A. P. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, J. M. Millam, M. Klene, C. Adamo, R. Cammi, J. W. Ochterski, R. L. Martin, K. Morokuma, O. Farkas, J. B. Foresman, and D. J. Fox, Gaussian, Inc., Wallingford CT, 2019. (52) Humphrey, W.; Dalke, A.; Schulten, K. VMD: Visual Molecular Dynamics. Journal of Molecular Graphics 1996, 14, 33–38. (53) Staroverov, V. N.; Scuseria, G. E.; Tao, J.; Perdew, J. P. Comparative Assessment of a New Nonempirical Density Functional: Molecules and Hydrogen-Bonded Complexes. The Journal of Chemical Physics 2003, 119, 12129–12137. (54) Weigend, F.; Ahlrichs, R. Balanced Basis Sets of Split Valence, Triple Zeta Valence and Quadruple Zeta Valence Quality for H to Rn: Design and Assessment of Accuracy. Phys. Chem. Chem. Phys. 2005, 7, 3297. (55) Dolg, M.; Stoll, H.; Savin, A.; Preuss, H. Energy-Adjusted Pseudopotentials for the Rare Earth Elements. Theoret. Chim. Acta 1989, 75, 173–194. 413 (56) Dolg, M.; Stoll, H.; Preuss, H. A Combination of Quasirelativistic Pseudopotential and Ligand Field Calculations for Lanthanoid Compounds. Theoret. Chim. Acta 1993, 85, 441–450. (57) Smith, D. G. A.; Burns, L. A.; Patkowski, K.; Sherrill, C. D. Revised Damping Parameters for the D3 Dispersion Correction to Density Functional Theory. J. Phys. Chem. Lett. 2016, 7, 2197–2203. (58) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. A Consistent and Accurate Ab Initio Parametrization of Density Functional Dispersion Correction (DFT-D) for the 94 Elements H- Pu. The Journal of Chemical Physics 2010, 132, 154104. 414 CHAPTER 7: CHARGE-CONTROLLED REACTIVITY OF YTTRIUM COMPLEXES 7.1 Motivation and Background As we embarked on our research, we were confronted with the challenge of enhancing the magnetic properties of dysprosium complexes. Drawing from the insights of Chapter 6, we recognized the importance of good donor ligands in creating a strong axial crystal field around the dysprosium metal. In our pursuit of a solution, we propose a unique way to improve the donor potential of the ligand while preserving its backbone. This chapter unveils our methods and the discovery of some unusual molecular structures that piqued our curiosity. 7.2 Approach to Modification in Metal Complexes For discussion purposes, the numbering for metal complexes is continued from Chapters 5 and 6. Nitrogens are responsible for providing the axial field to the metal center; however, they are not the best donors. Our initial thought was to treat Dy(NHAr*)2Cl (4) with one equivalent of (trimethylsilyl)methylpotassium to give the product shown in Scheme 7.1 and KCl and Si(CH3)4 as byproducts. Scheme 7.1. Proposed synthesis of dysprosium imide complex. 415 The yellow-colored crystals were grown overnight from the concentrated n-hexane solution at ‒35 °C. However, the crystal structure showed two amide ligands encapsulating the potassium ion owing to the ligand dissociation from the dysprosium complex. Besides the obtained structure, we could not get NMR data due to the presence of paramagnetic species in the solution. We decided to do this by using Y(NHAr*)2Cl (1) complex to first establish a reaction protocol for dysprosium. Perhaps (trimethylsilyl)methylpotassium is too strong a base for this purpose, leading to the dissociation of ligands from the metal, so we turned to lithium congener of it. Treatment of Y(NHAr*)2Cl (1) with one equivalent of (trimethylsilyl)methyllithium gave complex 9 after the overnight reaction. Figure 7.1. Reactivity of Y(NHAr*)2Cl (1) with one equivalent of (trimethylsilyl)methyllithium. 416 Figure 7.2. Crystal structure of 9. The single-crystal X-ray diffraction suggests the formation of a bimetallic yttrium complex with two chlorides as bridging atoms (Figure 7.2). Each yttrium atom is coordinated with one amide ligand, two bridging chlorides, and one methyl(trimethylsilyl)group, satisfying the yttrium valency. The Y‒Cnt distances are 2.5652(8) and 2.5590(8) Å, slightly longer than complex 1 (2.493(3) Å). This elongation is due to the accommodation of ligands to form a pseudo 4- coordinate complex as opposed to a 3-coordinate complex 1. The Y1‒N1 and Y2‒N2 distances are 2.212(2) and 2.226(2) Å respectively, consistent with 1. Further, Y1‒Cl1 (2.691 Å) and Y2‒Cl2 (2.705 Å) are shorter than Y1‒Cl2 (2.731 Å) and Y2‒Cl1 (2.725 Å), suggesting ionic and dative interactions of a chloride with two metal centers. The formation of complex 9 indicates that it is easier to remove amide ligand than chloride and, consequently, the formation of a bimetallic complex. 417 After this unexpected discovery of complex 9, I decided to treat complex 1 with two equivalents of (trimethylsilyl)methyllithium and potentially drive out the chloride ions. The treatment of two equivalents of the base with complex 1 resulted in the formation of mononuclear Y(NHAr*)(CH2SiMe3)2 10 as a pale-yellow solid. After removing LiCl salt, the X-ray quality crystals were grown overnight from a concentrated solution of 10 in n-hexane at -35 °C. The Y‒ Cnt and Y‒N distances are 2.4859(1) and 2.2297(1) Å, respectively, consistent with complex 1. Figure 7.3. (top) Reactivity of Y(NHAr*)2Cl (1) with two equivalents of (trimethylsilyl)methyllithium, (bottom) crystal structure of 10. 418 Attempts to replace chlorides were carried out using different reagents such as thallium triflate, silver triflate, thallium tosylate, and silver tosylate. All these reactions resulted in ligand dissociation from the metal; therefore, multiple attempts to displace chloride have failed. This unexpected discovery of complexes 9 and 10 made us think that we need to understand the charge distribution in 1 to understand the reactivity. We used Natural Population Analysis (NPA) to calculate charge occupancy in Complex 1. As shown in Table 7.1, nitrogen has more negative electron density than chlorine; therefore, it is easier to remove amide than chloride. In other words, a more ionic Y‒N bond will break first than a Y‒Cl bond Table 7.1. Natural population analysis of Y(NHAr*)2Cl (1). Atom Y N1 N2 Cl H1 H2 Natural Charge 2.20965 -1.19665 -1.18525 -0.74508 0.38626 0.39661 The reactivity of complex 1 is very similar to its dysprosium congener 4, where chloride can be abstracted by using TlBArF24 salt. Treatment of 1 with TlBArF24 resulted in an immediate color change from black to orange. The orange-colored crystals of [(NHAr*)2Y][BArF24] (11) in 43% crystalline yield. Single-crystal X-ray diffraction studies revealed that complex 11 recrystallizes as a C2 symmetric molecule with only half of the asymmetric unit in the unit cell. The Y–Cnt and Y–N distances are 2.4941(8) and 2.216(3) Å, respectively, similar to starting material 1 and dysprosium congener 6. The N–Y–N angle is 120.05(16)°, which is 13° more acute than 1 but consistent with Dy congener 6.1 419 Figure 7.4. (top) Synthesis of [(NHAr*)2Y][ BArF24] (11) and (bottom) structure of 11 from single-crystal X-ray diffraction (thermal ellipsoids drawn at 50%). Counter anion BArF24 and hydrogen atoms on carbon are removed for clarity. The reactivity of complex 1 is charged controlled; therefore, we can't exchange chloride with any other functionality or ‒R group. A different approach is proposed here where yttrium triflate or yttrium alkoxides can be used instead of yttrium chloride to synthesize synthetic congener of 1. 420 Scheme 7.2. Proposed synthetic route to access yttrium alkyl complexes. However, no isolable product has been obtained yet. Perhaps a different synthetic route is required for this transformation. Once two amide ligands are installed on either of these starting materials, the alkoxide or triflate group could be exchanged with different functionalities. Additionally, NPA suggests that oxygen has more negative charges than chlorine. This advises that the reactivity of 12 could be different from 1. Surprisingly, the negative charge on oxygen is still lower than nitrogen; thus, studying the reactivity of 12 could be quite enriching to understand the chemistry of these complexes. Table 7.2. Natural population analysis of Y(NHAr*)2OTf (12). Atom Y N1 N2 O H1 H2 Natural Charge 2.28125 -1.21245 -1.20701 -1.16991 0.38787 0.39692 421 7.3 Conclusions In the quest to convert an amide ligand into imide, we discovered the reactivity of Y(NHAr*)2Cl (1) complex is charge-controlled and it is easier to displace chloride than the amide ligands. It is expected the reactivity of 1 can be modulated by using Y(OTf)3 or Y(OAr)3 starting material instead of YCl3. From this study, we wish to explore the reactivity of Y(NHAr*)2(OR) towards nucleophilic reagents such as lithium alkynyls and lithium alkyls. Further treatment with a base could deprotonate the N‒H proton and thus drive out the ‒OR group from the metal complex resulting in the formation of a metal imide complex. 422 7.4 Experimental Details Synthesis of [Y(NHAr*)(CH2SiMe3)(μ-Cl)]2 (9) A 20 mL scintillation vial charged with crystals of Y(NHAr*)2Cl (1) (24.1 mg, 0.02 mmol, 1 equiv.), hexane (2 mL), and a magnetic stir bar. The vial was placed in a liquid nitrogen-cooled cold well until the solution froze. Once frozen, the vial was removed from the cold well and placed on a magnetic stir plate. When the solution had thawed enough to stir, (trimethylsilyl)methyllithium (2 mg, 0.02 mmol, 1 equiv.) in hexane (1 mL) was added. The reaction was stirred for 12 hours at room temperature. The volatiles were removed in vacuo, and the remaining residue was dissolved in n-hexane and filtered twice through Celite using a pipette filter. Yellow-colored X-ray quality single crystals were grown by chilling a concentrated n-hexane solution of 9 in a –35 °C freezer overnight (6 mg, 39.3% yield). Y(NHAr*)2Cl (1) was synthesized using previously reported procedure.2 Synthesis of Y(NHAr*)(CH2SiMe3)2 (10) A 20 mL scintillation vial charged with crystals of Y(NHAr*)2Cl (1) (24.2 mg, 0.02 mmol, 1 equiv.), hexane (2 mL), and a magnetic stir bar. The vial was placed in a liquid nitrogen-cooled cold well until the solution froze. Once frozen, the vial was removed from the cold well and placed on a magnetic stir plate. When the solution had thawed enough to stir, (trimethylsilyl)methyllithium (4 mg, 0.04 mmol, 2 equiv.) in hexane (1 mL) was added. The reaction was stirred for 12 hours at room temperature. The volatiles were removed in vacuo, and the remaining residue was dissolved in n-hexane and filtered twice through Celite using a pipette filter. Yellow-colored X-ray quality single crystals were grown by chilling a concentrated n-hexane solution of 10 in a –35 °C freezer overnight (12.1 mg, 73.8% yield). 1H NMR (500 MHz, C6D6) δ 423 7.41 (s, 2H), 7.29 (s, 2H), 7.20 (dd, J = 7.3, 1.7 Hz, 1H), 6.96 (d, J = 7.3 Hz, 1H), 6.78 (t, J = 7.3 Hz, 1H), 5.46 (s, 1H), 3.13 – 2.96 (m, 4H), 2.90 (hept, J = 7.0 Hz, 1H), 2.73 (hept, J = 6.9 Hz, 1H), 1.41 (d, J = 7.0 Hz, 6H), 1.31 (d, J = 7.0 Hz, 6H), 1.25 (dd, J = 6.9, 2.7 Hz, 13H), 1.11 (d, J = 7.0 Hz, 6H), 0.94 (d, J = 6.9 Hz, 6H), 0.22 (s, 18H), -0.45 (dd, J = 11.3, 2.9 Hz, 2H), -0.70 (dd, J = 11.3, 3.2 Hz, 2H). 13C NMR (126 MHz, C6D6) δ 158.76 (d, J = 3.2 Hz), 158.75, 157.64, 151.64, 148.98, 148.06, 147.83, 133.76, 131.20, 127.50 – 127.36 (m), 127.05, 123.52, 122.97, 121.35, 114.49, 36.22, 35.88, 35.03, 34.37, 31.97, 31.25, 31.06, 25.30, 25.17, 24.70, 24.49, 23.59, 23.06, 22.43, 14.35, 4.24. Synthesis of [(NHAr*)2Y][BArF24] (11) A 20 mL scintillation vial was charged with 4 (120 mg, 0.1 mmol, 1 equiv.), diethyl ether (5 mL), and a magnetic stir bar. A separate 20 mL scintillation vial was loaded with TlBArF24 (114.6 mg, 0.1 mmol, 1 equiv.) and diethyl ether (2 mL). The TlBArF24 solution was added dropwise to the solution of 1 for about a minute. The solution color rapidly changed from yellow to orange. The solution was stirred for 2 h at room temperature. The volatiles were then removed in vacuo, and the remaining residue was dissolved in diethyl ether (2 mL) and filtered through Celite. Orange X-ray quality single crystals were produced by chilling a concentrated diethyl ether solution of 3 in a –35 °C freezer overnight (90 mg, 43.0% yield). 424 Figure 7.5. 1H NMR Spectrum of Y(NHAr*)(CH2SiMe3)2 (10). 425 Figure 7.6. 13C NMR Spectrum of Y(NHAr*)(CH2SiMe3)2 (10). 426 Figure 7.7. HSQC Spectrum of Y(NHAr*)(CH2SiMe3)2 (10). Single Crystal X-ray Diffraction Single crystal data was collected on XtaLAB Synergy, Dualflex, Hypix diffractometer using CuKα radiation. Data collection was done at 100 K under a continuous flow of liquid nitrogen. In Olex2 program,3 Crystal structures were solved with the ShelXT solution using intrinsic phasing and refined with the SheXL refinement package using least squares minimization.4 All hydrogens are refined anisotropically. All crystals were stable at room temperature for mounting. 427 Figure 7.8. Structure of [Y(NHAr*)(CH2SiMe3)(μ-Cl)]2 (9) recrystallized from n-hexane. Each molecule of 9 recrystallizes with 0.75 molecules of n-hexane. 428 Table 7.3. Crystallographic data and structural refinement of [Y(NHAr*)(CH2SiMe3)(μ-Cl)]2 (9). Identification code Empirical formula Formula weight Temperature/K Crystal system Space group a/Å b/Å c/Å α/° β/° γ/° Volume/Å3 Z ρcalcg/cm3 μ/mm-1 F(000) Crystal size/mm3 Radiation 2Θ range for data collection/° Index ranges Reflections collected Independent reflections Data/restraints/parameters Goodness-of-fit on F2 Final R indexes [I>=2σ (I)] Final R indexes [all data] Largest diff. peak/hole / e Å-3 RJ5-102 C84.38H130.88Cl2N2Si2Y2 1478.18 100(2) monoclinic P21/c 13.15349(16) 16.34541(18) 43.6150(5) 90 93.1826(12) 90 9362.72(19) 4 1.049 2.693 3164.0 0.097 × 0.082 × 0.051 CuKα (λ = 1.54184) 5.776 to 160.596 -15 ≤ h ≤ 16, -20 ≤ k ≤ 20, -55 ≤ l ≤ 48 80459 20026 [Rint = 0.0857, Rsigma = 0.0572] 20026/0/969 1.228 R1 = 0.1239, wR2 = 0.2824 R1 = 0.1300, wR2 = 0.2851 1.39/-1.47 429 Figure 7.9. Structure of Y(NHAr*)(CH2SiMe3)2 (10) recrystallized from n-hexane. 430 Table 7.4. Crystallographic data and structural refinement of Y(NHAr*)(CH2SiMe3)2 (10). Empirical formula Formula weight Temperature/K Crystal system Space group a/Å b/Å c/Å α/° β/° γ/° Volume/Å3 Z ρcalcg/cm3 μ/mm-1 F(000) Crystal size/mm3 Radiation 2Θ range for data collection/° Index ranges Reflections collected Independent reflections Data/restraints/parameters Goodness-of-fit on F2 Final R indexes [I>=2σ (I)] Final R indexes [all data] Largest diff. peak/hole / e Å-3 C88H144N2Si4Y2 1520.22 100(2) triclinic P-1 9.80501(15) 13.3179(2) 18.4337(3) 84.7472(13) 87.0318(12) 69.7687(15) 2248.59(7) 1 1.123 2.527 820.0 0.143 × 0.118 × 0.063 CuKα (λ = 1.54184) 4.816 to 160.184 -12 ≤ h ≤ 12, -16 ≤ k ≤ 16, -23 ≤ l ≤ 22 53158 9198 [Rint = 0.0474, Rsigma = 0.0285] 9198/0/463 1.050 R1 = 0.0331, wR2 = 0.0856 R1 = 0.0380, wR2 = 0.0895 0.63/-0.73 431 Figure 7.10. Structure of [(NHAr*)2Y][BArF24] (11) recrystallized from ether. 432 Table 7.5. Crystallographic data and structural refinement of [(NHAr*)2Y][BArF24] (11). Empirical formula Formula weight Temperature/K Crystal system Space group a/Å b/Å c/Å α/° β/° γ/° Volume/Å3 Z ρcalcg/cm3 μ/mm-1 F(000) Crystal size/mm3 Radiation 2Θ range for data collection/° Index ranges Reflections collected Independent reflections Data/restraints/parameters Goodness-of-fit on F2 Final R indexes [I>=2σ (I)] Final R indexes [all data] Largest diff. peak/hole / e Å-3 Flack parameter C104H112BF24N2Y 1945.67 100(2) monoclinic C2 21.21247(13) 15.73375(12) 14.74716(10) 90 92.5538(6) 90 4917.00(6) 2 1.314 1.630 2020.0 0.139 × 0.1 × 0.063 CuKα (λ = 1.54184) 6 to 160.88 -18 ≤ h ≤ 26, -19 ≤ k ≤ 19, -18 ≤ l ≤ 18 50510 10435 [Rint = 0.0334, Rsigma = 0.0249] 10435/1/612 1.035 R1 = 0.0393, wR2 = 0.1049 R1 = 0.0399, wR2 = 0.1054 0.75/-0.54 -0.020(4) 433 Table 7.6. Metric data from the crystal structures of [Y(NHAr*)(CH2SiMe3)(μ-Cl)]2 (9), Y(NHAr*)(CH2SiMe3)2 (10), and [(NHAr*)2Y][BArF24] (11). All distances are in Å and angles are in (˚). Complex [Y(NHAr*)(CH2SiMe Y(NHAr*)(CH2SiMe3)2 Y‒N Y‒Cl Y‒CH2 Y‒N‒Cipso N‒Y‒N N‒Y‒Cl Arcent‒Y‒N Y-ArCent Cent‒Y‒ Cent Y‒CAr Average CAr‒ CAr bond 3)(μ-Cl)]2 (9) 2.211(7), 2.226(6) 2.705(2), 2.725(2), 2.7304(19), 2.690(2) 2.355(9), 2.358(10) 135.4(6) (10) 2.2279(16) 2.371(2), 2.382(2) 131.40(13) 104.41(7) 107.70(1) 91.59(18) 91.61(8) 2.5655(8) 95.91(4) 2.48444(19) [(NHAr*)2Y][BArF24] (11) 2.216(3) 131.3(3) 120.05 99.30(8) 94.54(8) 2.49412(8) 152.114(14) 2.975(8), 2.971(9), 2.951(9), 2.864(8), 2.867(9), 2.912(8) 1.402(5) 2.8513(19), 2.8407(18), 2.7903(18), 2.8501(19), 2.8944(19), 2.9044(18) 1.408(1) 2.910(3), 2.950(3), 2.874(4), 2.842(3), 2.764, 2.838(3) 1.407(5) 434 REFERENCES (1) Benner, F.; Jena, R.; Odom, A. L.; Demir, S. Magnetic Hysteresis in a Pseudo Low-Coordinate Bisamide Dysprosium Complex. Submitted manuscript. (2) Jena, R.; Benner, F.; Delano, F.; Holmes, D.; McCracken, J.; Demir, S.; Odom, A. L. A Rare Isocyanide Derived from an Unprecedented Neutral Yttrium( II ) Bis(Amide) Complex. Chem. Sci. 2023, 14, 4257–4264. (3) Dolomanov, O. V.; Bourhis, L. J.; Gildea, R. J.; Howard, J. A. K.; Puschmann, H. OLEX2 : A Complete Structure Solution, Refinement and Analysis Program. J Appl Crystallogr 2009, 42, 339–341. (4) Sheldrick, G. M. Crystal Structure Refinement with SHELXL. Acta Crystallogr C Struct Chem 2015, 71, 3–8. 435