Transition metal-dinitrogen complexes : preparation and reactivity 9783527344260, 3527344268

Citation preview

Transition Metal-Dinitrogen Complexes

Transition Metal-Dinitrogen Complexes Preparation and Reactivity

Edited by Yoshiaki Nishibayashi

Editor Yoshiaki Nishibayashi

The University of Tokyo School of Engineering Hongo, Bunkyo-ku 113-8656 Tokyo Japan

All books published by Wiley-VCH are carefully produced. Nevertheless, authors, editors, and publisher do not warrant the information contained in these books, including this book, to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate. Library of Congress Card No.: applied for British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library. Bibliographic information published by the Deutsche Nationalbibliothek

The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at . © 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Boschstr. 12, 69469 Weinheim, Germany All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form – by photoprinting, microfilm, or any other means – nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Print ISBN: 978-3-527-34425-3 ePDF ISBN: 978-3-527-34429-1 ePub ISBN: 978-3-527-34427-7 oBook ISBN: 978-3-527-34426-0 Typesetting SPi Global, Chennai, India Printing and Binding

Printed on acid-free paper 10 9 8 7 6 5 4 3 2 1

v

Contents Preface xi About the Editor xiii 1

Overviews of the Preparation and Reactivity of Transition Metal–Dinitrogen Complexes 1 Yoshiaki Tanabe and Yoshiaki Nishibayashi

1.1 1.2 1.3 1.4

Introduction 1 Biological Nitrogen Fixation 4 Historical Background of Transition Metal–Dinitrogen Complexes 9 Coordination Chemistry of Transition Metal–Dinitrogen Complexes 13 Coordination Patterns of Dinitrogen and Mononuclear Transition Metal–Dinitrogen Complexes 13 Multinuclear Transition Metal–Dinitrogen Complexes 16 Chemical Activation and Reactivity of Dinitrogen Using Transition Metal Complexes 21 Protonation of Transition Metal-bound Dinitrogen 21 Cleavage of Transition Metal-bound Dinitrogen 25 Reaction of Transition Metal-bound Dinitrogen with Dihydrogen 26 Functionalization of Transition Metal-bound Dinitrogen 29 Electrochemical and Photochemical Conversion of Dinitrogen Using Transition Metal Complexes 31 Catalytic Conversion of Dinitrogen into Ammonia Using Transition Metal Complexes 34 Catalytic Formation of Ammonia or Hydrazine Using Molybdenum Complexes 34 Catalytic Formation of Ammonia or Hydrazine Using Transition Metal Other than Molybdenum (Iron, Ruthenium, Osmium, Cobalt, and Vanadium) Complexes 40 Catalytic Transformation of Hydrazine into Ammonia 45 Catalytic Formation of Silylamine 47 Conclusion and Perspectives 50 References 51

1.4.1 1.4.2 1.5 1.5.1 1.5.2 1.5.3 1.5.4 1.5.5 1.6 1.6.1 1.6.2

1.6.3 1.6.4 1.7

vi

Contents

2

Group 4 Transition Metal–Dinitrogen Complexes 79 Hidetake Seino and Yuji Kajita

2.1 2.2

Introduction 79 Preparation of Group 4 Transition Metal–Dinitrogen Complexes 80 Dinitrogen Complexes of Bis(cyclopentadienyl)titanium Derivatives 80 Dinitrogen Complexes of Bis(cyclopentadienyl)zirconium and Bis(cyclopentadienyl)hafnium Derivatives 89 Other Dinitrogen Complexes Based on Cyclopentadienyl Ligands 98 Dinitrogen Complexes Supported by σ-donor Ligands 100 Heterobimetallic Dinitrogen Complexes 109 Reactions of Group 4 Transition Metal–Dinitrogen Complexes 112 Protonation 112 Reduction 115 Reactions with Hydrogen 120 Reactions with Si—H and B—H Bonds 129 Reactions with Alkyl Halides and Their Equivalents 131 Reactions with Alkynes 136 Reactions with Carbon Dioxide and Cumulenes 138 Reactions with Carbon Monoxide 142 Dinitrogen Ligand Substitution 148 Conclusion and Perspectives 151 Addition After Acceptance of this Manuscript 151 References 152

2.2.1 2.2.2 2.2.3 2.2.4 2.2.5 2.3 2.3.1 2.3.2 2.3.3 2.3.4 2.3.5 2.3.6 2.3.7 2.3.8 2.3.9 2.4 2.5

3

Group 5 Transition Metal-Dinitrogen Complexes 159 Leila M. Duman and Lawrence R. Sita

3.1 3.2 3.2.1 3.2.2 3.2.3 3.3 3.3.1 3.3.2 3.3.3 3.4

Introduction 159 Preparation of Group 5 Metal N2 Complexes 160 Vanadium 160 Niobium 174 Tantalum 178 N≡N Bond Cleavage Within Group 5 Metal N2 Complexes 187 Vanadium 188 Niobium 192 Tantalum 197 Nitrogen Fixation Mediated by Group 5 Transition-metal N2 Complexes 201 Vanadium 202 Niobium 204 Tantalum 206 CPAM Group 5 Bimetallic (μ-η1 :η1-N2 ) Complexes 206 Conclusions and Perspectives 212 References 214

3.4.1 3.4.2 3.4.3 3.5 3.6

Contents

4

Group 6 Transition Metal–Dinitrogen Complexes 221 Nicolas Mézailles

4.1 4.2 4.2.1 4.2.1.1 4.2.1.2 4.2.1.3 4.2.2

Introduction 221 Preparation of Group 6 Transition Metal–Dinitrogen Complexes 222 End-on Dinitrogen Complexes from N2 222 Arene and Phosphine Ligands 222 Thioether Ligands 226 Nitrogen and Cp Ligands 226 End-on Bridging Dinitrogen Complexes from N2 : Synthesis and N2 Splitting 228 Stoichiometric Reactions of Group 6 Transition Metal–Dinitrogen and Metal–Nitrido Complexes 234 N—H Bond Formation 234 N—C Bond Formation 238 N-element Bond Formation 241 Catalytic Reactions of Group 6 Transition Metal–Dinitrogen Complexes 247 Catalytic Formation of N2 H4 /NH3 from Nonisolated M–N2 Complexes 247 Catalytic Formation of N(SiMe3 )3 247 Catalytic Formation of NH3 251 Chemistry of Cr Complexes 259 Conclusion and Perspectives 261 References 263

4.3 4.3.1 4.3.2 4.3.3 4.4 4.4.1 4.4.2 4.4.3 4.5 4.6

5

Toward N—N Bond Cleavage: Synthesis and Reactivity of Group 7 Dinitrogen Complexes 271 Elon A. Ison

5.1 5.1.1 5.1.2 5.1.2.1 5.1.2.2 5.2 5.2.1 5.2.2 5.3

Synthesis of Group VII N2 Complexes 271 Syntheses of Terminal N2 Complexes 271 Reactivity of Terminal N2 Complexes 275 Synthesis of Bridged N2 Complexes by Reaction with Lewis Acids 276 Alternative Syntheses of Bridged N2 Complexes 279 Cleavage and Functionalization of N2 Bonds 280 Generation of Diazomethane from CpMn(CO)2 N2 280 Cleavage of N2 in the Coordination Sphere of Rhenium 281 Conclusions and Future Outlook 281 References 282

6

Group 8 Transition Metal–Dinitrogen Complexes 285 Adam D. Piascik and Andrew E. Ashley

6.1 6.2 6.2.1 6.2.2 6.2.3

Introduction 285 Preparation of Group 8 Transition Metal–Dinitrogen Complexes Ligand Substitution 288 Precursor Reduction 292 Other Methods 296

288

vii

viii

Contents

6.3 6.3.1 6.3.2 6.3.3 6.4 6.4.1 6.4.2 6.4.3 6.4.4 6.5

Stoichiometric Reactions of Group 8 Transition Metal–Dinitrogen Complexes 297 Substitution Reactions and Lability of Bound N2 297 Cleavage and Functionalization of Coordinated N2 301 Other Stoichiometric Reactivity 309 Catalytic Reactions of Group 8 Transition Metal–Dinitrogen Complexes 311 Early Results and Fe Bis(diphosphine) Systems for Catalytic N2 Fixation 311 Catalytic NH3 Production by EPR 3 -supported Systems 313 Catalytic N2 Fixation by Other Systems 317 Other Catalytic Reactions of Group 8 M–N2 Complexes 319 Conclusion and Perspectives 327 References 328

7

Group 9 Transition Metal–Dinitrogen Complexes 337 Connie C. Lu and Steven D. Prinslow

7.1 7.1.1 7.1.1.1 7.1.1.2 7.1.2 7.1.2.1 7.1.2.2 7.1.2.3 7.1.2.4

Cobalt–Dinitrogen Complexes 337 Monodentate Phosphine Donors 338 CoH(N2 )(PR3 )3 and Related Co(I) Complexes 338 Cobaltate Complexes: [Co(N2 )(PR3 )3 ]− 342 Tripodal Polyphosphine Ligands 345 Tris(phosphine) Ligands 345 Tris(phosphino)borate Ligands 346 Trisphosphine Systems with an Apical Main Group Donor 347 Trisphosphine Systems with an Apical Transition Metalloligand Donor 350 Ligands with Exclusively Nitrogen Donors 355 Tris(pyrazoyl)borate (Tp) Ligands 355 β-diketiminate Ligands 356 Bis(α-imino)pyridine Ligands 358 N-heterocyclic Carbene Ligands 359 Pincer Ligands 360 Monoanionic PNP-Type and PBP-Type Ligands 361 Pincer Ligands with N/P Donors 363 N-heterocyclic Carbene-Based Pincer Ligands 365 Other Assorted Ligands 367 Analysis and Summary of Cobalt–Dinitrogen Complexes 369 Rhodium–Dinitrogen Complexes 370 Early Rh–N2 Complexes 370 Phosphine Ligands 372 Ligands with Exclusively Nitrogen Donors 374 Bis(α-imino)pyridine Ligands 374 β-diketiminate Ligands 375 Pincer Ligands 375 PCP Pincer Ligands 376 PNP Pincer Ligands 378

7.1.3 7.1.3.1 7.1.3.2 7.1.3.3 7.1.4 7.1.5 7.1.5.1 7.1.5.2 7.1.5.3 7.1.6 7.1.7 7.2 7.2.1 7.2.2 7.2.3 7.2.3.1 7.2.3.2 7.2.4 7.2.4.1 7.2.4.2

Contents

7.2.4.3 7.2.5 7.2.6 7.3 7.3.1 7.3.2 7.3.3 7.3.3.1 7.3.3.2 7.3.4 7.3.4.1 7.3.4.2 7.3.5 7.3.6 7.3.7 7.4 7.4.1 7.4.1.1 7.4.1.2 7.4.2

Other Pincer Ligands 380 N-heterocyclic Carbene Ligands 380 Summary of Rhodium–Dinitrogen Complexes 381 Iridium–Dinitrogen Complexes 381 Early Ir–N2 Complexes 382 Phosphine Ligands 383 Ligands with Exclusively Nitrogen Donors 385 Tris(pyrazoyl)borate (Tp) Ligands 385 β-diketiminate Ligands 386 Pincer Ligands 386 PNP-Type Pincer Ligands 386 PCP- and PSiP-Type Pincer Ligands 388 N-heterocyclic Carbene Ligands 390 Miscellaneous 391 Summary of Iridium–Dinitrogen Complexes 391 Group 9 Catalysts for N2 Functionalization 392 Cobalt-Based Catalysts 392 Dinitrogen Silylation 393 Dinitrogen Fixation 395 Outlook for Rhodium and Iridium Catalysts 396 Acknowledgments 396 References 396

8

Group 10 and 11 Transition Metal–Dinitrogen Complexes 403 Ricardo B. Ferreira and Leslie J. Murray

8.1 8.2 8.2.1 8.2.1.1 8.2.1.2 8.2.1.3 8.2.1.4 8.2.2 8.3 8.3.1 8.3.1.1 8.3.1.2 8.3.1.3 8.3.2 8.4

Introduction 403 Group 10 Transition Metal–Dinitrogen Complexes 405 Nickel 405 Interaction of Dinitrogen with Nickel Surfaces 406 Matrix-Assisted Isolation of Binary or Ternary Compounds 406 Coordination Compounds 408 Structural Relationships and Comparisons 420 Palladium and Platinum 422 Group 11 Transition Metal–Dinitrogen Complexes 423 Copper 423 Matrix-Assisted Isolation of Binary or Ternary Compounds 423 Coordination Compounds 425 Structural Relationships and Comparisons 427 Silver and Gold 429 Conclusion and Perspectives 430 References 431

9

Group 3 Transition Metal, Lanthanide, and Actinide–Dinitrogen Complexes 441 Yoshiaki Tanabe

9.1

Introduction 441

ix

x

Contents

9.2 9.2.1

9.2.2

9.2.3 9.2.4

9.2.5 9.3 9.3.1

9.3.2 9.3.3 9.4

Preparation and Characterization of Group 3 Transition Metal, Lanthanide, and Actinide–Dinitrogen Complexes 443 Overviews of Preparation, Structures, and Characterization of Group 3 Transition Metal, Lanthanide, and Actinide–Dinitrogen Complexes 443 Preparation and Structures of Side-on-Bound {(N2 )2− }-Bridged Dinuclear Group 3 Transition Metal, Lanthanide, and Actinide–Dinitrogen Complexes 443 Preparation and Structures of Side-on-bound {(N2 )3− }-Bridged Dinuclear Group 3 Transition Metal and Lanthanide Complexes 456 Preparation and Structures of {(N2 )4− }-Bridged Dinuclear, Trinuclear, and Tetranuclear Lanthanide and Actinide–Dinitrogen Complexes 457 Preparation and Structures of End-on-Bound Group 3 Transition Metal, Lanthanide, and Actinide–Dinitrogen Complexes 460 Reactivity and Property of Group 3 Transition Metal, Lanthanide, and Actinide–Dinitrogen Complexes 462 Cleavage, Protonation, and Functionalization of Dinitrogen upon Group 3 Transition Metal, Lanthanide, and Actinide–Dinitrogen Complexes 462 Group 3 Transition Metal–Dinitrogen Complexes as Mediators for the Transformation of Small Molecules 466 {(N2 )3− }-Bridged Dinuclear Group 3 Transition Metal and Lanthanide Complexes as Single-Molecule Magnets 468 Conclusion and Perspectives 469 References 470 Index 475

xi

Preface A great progress has recently been achieved in the research area of Nitrogen Fixation, as one of the most important subjects in chemistry. Especially, the development of catalytic ammonia formation from nitrogen gas under mild reaction conditions has been repowered by several research groups. Although nitrogen fixation chemistry is one of the most notable fields of research, books provide comprehensive knowledge of the relevant fields are rather limited until now. I believe that the latest research results by researchers engaged in state-of-the-art research on synthesis of transition metal–dinitrogen complexes and their reactivity in this book will give very useful information to researchers, teachers, and students who are interested in the research filed of nitrogen fixation by using transition metal–dinitrogen complexes. I would like to thank all the contributors for their chapters in this book and their enthusiastic efforts to present recent advances of Nitrogen Fixation by using transition metal–dinitrogen complexes. I anticipate that their contributions will stimulate further study in Nitrogen Fixation. I would like also to offer my warm thanks to the Wiley-VCH team for their continuous support. Finally, I deeply appreciate staffs and students in my research group for their valuable assistances. May 2018

Yoshiaki Nishibayashi The University of Tokyo School of Engineering Tokyo, Japan

xiii

About the Editor Yoshiaki Nishibayashi is a full professor at the University of Tokyo, since 2016. He received his PhD in 1995 from Kyoto University under the supervision of Professor Sakae Uemura. He became an assistant professor at the University of Tokyo in 1995 and moved to Kyoto University in 2000. In 2005, he became an associate professor at the University of Tokyo as PI. Since 2016, he has been a full professor at the University of Tokyo. He received the Chemical Society of Japan Award for Distinguished Young Chemists in 2001, the Minister Award for Distinguished Young Scientists Japan in 2005, the JSPS Prize in 2012, the Green & Sustainable Chemistry Honorable Award in 2012, the Nissan Chemical Industries Award for Novel Reaction & Method from the Society of Synthetic Organic Chemistry, Japan in 2016, the Japan Society of Coordination Chemistry Award for Creative Work in 2017, the Inoue Prize for Science in 2018, and the Prizes for Science and Technology (Research Category) in the Commendation for Science and Technology by the Minister of Education, Culture, Sports, Science, and Technology in 2018. His current research interests are focused on organic and organometallic chemistry. He is the author of more than 200 publications and review articles.

1

1 Overviews of the Preparation and Reactivity of Transition Metal–Dinitrogen Complexes Yoshiaki Tanabe and Yoshiaki Nishibayashi Department of Systems Innovation, School of Engineering, The University of Tokyo, Hongo, Bunkyo-ku, Tokyo 113-8656, Japan

1.1 Introduction Nitrogen, the fifth most abundant element in the solar system, is the most abundant element in the atmosphere of Earth [1] as well as the fourth most abundant element in cellular biomass [2]. However, it is rather a trace element in the lithosphere of Earth [3]. Thus, utilization of chemically inert gaseous molecular dinitrogen (N2 ) that exists in the atmosphere of Earth as the primary nitrogen source is inevitable in both biogeography and industry. Indeed, fixation of atmospheric nitrogen can be achieved by the conversion of molecular dinitrogen into ammonia (NH3 ) containing the most reduced form of nitrogen (−3) that can be a convenient precursor for several nitrogen-containing compounds and has been the most fundamental reaction pathway of the global nitrogen cycle [4, 5]. Industrially, NH3 is one of the 10 largest commodity chemical products and has been produced by the Haber–Bosch process in which atmospheric dinitrogen reacts with gaseous dihydrogen (N2 + 3 H2 → 2 NH3 ) since the early twentieth century [6–14]. Haber and van Oordt in 1904 first succeeded in the conversion of the mixture of N2 and H2 into NH3 in the presence of transition metal catalyst (Fe or Ni) at a high temperature in a laboratory [15–17]. Later, modification of the reactors and catalysts was achieved, and 90 g of ammonia was shown to be obtained every hour by using an osmium-based catalyst with the total yield of ammonia up to 8 vol% at 550 ∘ C and a total pressure of 175 atm of a stoichiometric mixture of dinitrogen and dihydrogen (1 : 3) in an experimental lecture held in Karlsruhe on 18 March 1909 [18–20]. Further modification of the catalysts for industrialization was investigated by Mittasch and coworkers in BASF, leading to the discovery of the combination of iron, K2 O, and Al2 O3 as one of the most active catalysts by 1910 [6, 21]. The first commercial plant for ammonia synthesis at Oppau began its operation by 1913 in collaboration with Bosch and coworkers at BASF, while the earlier commercial methods to fix atmospheric nitrogen such as Frank–Caro cyanamide process (CaC2 + N2 → CaCN2 + C) and Birkeland–Eyde electric arc process (N2 + O2 → 2 NO) were gradually replaced by the Haber–Bosch ammonia process [6–14]. Typical reaction conditions of the Haber–Bosch process are Transition Metal-Dinitrogen Complexes: Preparation and Reactivity, First Edition. Edited by Yoshiaki Nishibayashi. © 2019 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2019 by Wiley-VCH Verlag GmbH & Co. KGaA.

2

1 Overviews of the Preparation and Reactivity of Transition Metal–Dinitrogen Complexes

(a)

N2 + 3 H2

cat. Fe3O4/K2O/Al2O3 100–200 atm, 300–500 °C

N2 + 0.2682 O2 + 0.8841 CH4 + 1.2318 H2O Air (b) (N2 : O2 = 78.084 : 20.946)

cat. Ba–Ru/C 50–100 atm 370–400 °C

2 NH3

2 NH3 + 0.8841 CO2

Figure 1.1 (a) Prototype Haber–Bosch process operated at the first BASF’s Oppau plant. H2 is originally obtained from steam reforming of coal. (b) Kellogg advanced ammonia process with methane steam reforming.

shown in Figure 1.1a [6], where the reaction is carried out under high temperature and high pressure in the presence of heterogeneous solid-state catalysts prepared from magnetite (Fe3 O4 ) with the addition of alumina (Al2 O3 ), silica (SiO2 ), or alkaline earth metal oxide (CaO) as a “structural” promoter and alkaline metal oxide (K2 O) as an “electronic” promoter. Although formation of NH3 from N2 and H2 is thermodynamically favored under standard conditions (Δr H ∘ = −45.90 kJ mol−1 , Δr G∘ = −16.37 kJ mol−1 at 1 bar and 25 ∘ C), this conversion can hardly occur at ambient reaction conditions because the dissociation energy of the dinitrogen triple bond is high (D0 ∘ = 945.37 kJ mol−1 ) [22]. To lower and surmount the activation energy of this conversion, elevated pressure and temperature as well as heterogeneous solid-state catalysts are necessary, where bond-breakings upon chemisorption on the surface of solid-state catalysts were experimentally observed by Ertl and coworkers, who clarified the surface reaction pathway of the Haber–Bosch process as shown in Figure 1.2 [23–29]. Activation energy and turnover frequency of the catalytic ammonia synthesis are highly dependent not only on the catalyst but also on temperature, pressure, and the ratio of the substances and products, where the logarithm of the equilibrium constant for the reaction of N2 + 3 H2 = 2 NH3 at 1 bar becomes zero theoretically at 456 K [22]. For example, the apparent activation energy for the catalytic ammonia synthesis on the Fe(111) surface of an iron single crystal at around 748 K and a total pressure of 20 atm of a stoichiometric mixture of dinitrogen and dihydrogen (1 : 3) was determined by Somorjai and coworkers as 81.2 kJ mol−1 with an initial turnover frequency of 12.7 ± 2.0 molecules of ammonia per C4 surface iron atom per second [30]. A more improved method such as Kellogg advanced ammonia process (KAAP) uses ruthenium-based catalyst supported on graphite-containing carbon copromoted with barium, cesium, or rubidium performed at comparably lower pressure and temperature, the stoichiometry of which can be expressed as Figure 1.1b, when natural gas steam reforming is applied to ammonia production without the separation of dinitrogen from air [8–14, 31–35]. In this reaction, methane is the main hydrogen source of ammonia, and the gaseous ammonia obtained from the stoichiometry in Figure 1.1b theoretically contains 20.8 GJ per metric ton or 355 kJ mol−1 as chemical energy calculated based on the heat of combustion of methane in the lower heating value (LHV) (Δc H ∘ = −802.3 kJ mol−1 , Δc G∘ = −800.8 kJ mol−1 ) or 18.6 GJ per metric ton

1.1 Introduction

N+3H +1127 Radical pathway NH + 2 H +813

NH2 + H +408 Stepwise hydrogenation pathway 1/2 cis-N2H2 (g) + H2 +106 1/2 N2 1/2 N 1/2 N2H4 (l) + 1/2H2 2,ad +25 + 3/2 H2 +25 + 3/2 H2 0 NH3 Surface pathway −23 − NH2,ad −86 ΔH NH NH3,ad −46 ad −96 Nad −106 + Had kJ mol−1 + 2 Had −96 + 3 Had −106 −116 −136

Figure 1.2 Potential energy diagram for ammonia synthesis on the surface of iron, via stepwise hydrogenation or via formation of radicals.

based on that of ammonia (Δc H ∘ = −316.8 kJ mol−1 , Δc G∘ = −326.5 kJ mol−1 ) if full recovery of the reaction heat is assumed (Δr H ∘ = −37.8 kJ mol−1 and Δr G∘ = −27.5 kJ mol−1 per NH3 for Figure 1.1b) [22]. A classical BASF-type Haber–Bosch process that uses coke consumes chemical energy of 100 GJ per metric ton of NH3 in 1920 [6], which is much more efficient than the Birkeland–Eyde electric arc process (600 GJ per metric ton of fixed nitrogen) or the Frank–Caro cyanamide process (190 GJ per metric ton of NH3 derived from the decomposition of CaCN2 with H2 O) [12], whereas the most efficient ammonia plant with the ruthenium-based catalyst and methane steam reforming consumes as low as 27.2 GJ per metric ton or 463 kJ mol−1 of NH3 , where energy efficiency of around 75% with respect to the stoichiometric methane demand is achieved, which also means that additional chemical energy of 108 kJ mol−1 is required for the industrial synthesis of NH3 as represented in Figure 1.1b [13]. In an exergy analysis of a low-energy ammonia process to obtain the liquefied ammonia at −33 ∘ C (20.14 GJ per metric ton or 343 kJ mol−1 ) by Dybkjaer under a model reaction at 140 kgf cm−2 in an indirectly cooled two-bed radial converter using pure methane, cooling water available at 30 ∘ C, a steam to the carbon ratio of 2.5, and so forth, a total exergy of 30.69 GJ per metric ton or 523 kJ mol−1 is consumed with an exergy loss of 10.55 GJ per metric ton or 180 kJ mol−1 corresponding to a thermodynamic efficiency of 66% for the production of NH3 , where the biggest loss of exergy occurs at methane steam reforming sections with rather a slight loss made during the actual ammonia synthesis (1.70 GJ per metric ton or 29 kJ mol−1 ) [13, 36]. Further improvement of Haber–Bosch catalysts is still in progress, especially in the development of electronic and structural promoters. For example, Hosono and coworkers have

3

4

1 Overviews of the Preparation and Reactivity of Transition Metal–Dinitrogen Complexes

developed ruthenium-loaded electrode catalysts, which show higher catalytic performance than the conventional ruthenium catalysts at lower temperatures and pressures [37–40]. In total, the Haber–Bosch process annually produces more than 170 million metric tons of NH3 [41], consumes fossil fuels as the hydrogen source of NH3 , corresponding to 1–2% of the world’s annual primary energy supply, and is responsible for the emission of more than 450 million metric tons of CO2 [42–44]. This pollution can be reduced by using renewable energy sources for producing dihydrogen from water, but it should be more convenient to use water as a proton source for ammonia without using dihydrogen gas in high pressure and temperature. It must be noted that ammonia is attracting attention as a possible hydrogen carrier in the future, as well as a fuel for vehicles [45–49], which can minimize the use of fossil fuels. The present Haber–Bosch process requires a lot of reactors to obtain high pressure and temperature; thus, biological nitrogen fixation that can be carried out in small cells at ambient reaction conditions by using water as a proton source has been investigated as a model of an alternative method for the Haber–Bosch process [50–54].

1.2 Biological Nitrogen Fixation Atmospheric molecular dinitrogen has been fixed as ammonia via biological nitrogen fixation using electron carriers (ferredoxins or flavodoxins) as reducing reagents and water as a proton source under ambient pressure and temperature by some specific bacterial and archaeal organisms that possess nitrogen-fixing enzyme called nitrogenase [51, 52]. Based on the difference in transition metal (Mo, V, or Fe) included in its key cofactor (iron–molybdenum cofactor (FeMo-co), iron–vanadium cofactor (FeV-co), or iron–iron cofactor (FeFe-co)) consisting of an iron–sulfur cluster, nitrogenase can be classified into molybdenum nitrogenase, vanadium nitrogenase, or iron-only nitrogenase, among which molybdenum nitrogenase, the canonical form of this enzyme, works most efficiently, where 8 equiv of electrons and protons is consumed for reducing 1 equiv of dinitrogen to form 2 equiv of ammonia together with the formation of an equimolar amount of dihydrogen gas (Figure 1.3a), whereas vanadium nitrogenase (Figure 1.3b) or iron-only nitrogenase (Figure 1.3c) is less effective requiring more protons and electrons wasted to form more dihydrogen molecules [55, 56]. All the diazotrophic bacteria known to date encode molybdenum nitrogenase, whereas some diazotrophic bacteria especially living in soils possess the genes for alternative vanadium or iron-only nitrogenase. Few species such as Azotobacter vinelandii, an aerobic free-living microorganism in soils, are known to contain all the three types of nitrogenases, but utilization of alternative vanadium or iron-only nitrogenase occurs under molybdenum limitation or both molybdenum and vanadium limitations, respectively [57–59]. Structures of FeMo-co (Figure 1.4a) and FeV-co (Figure 1.4b) are determined both crystallographically and spectroscopically, where Fe4 S3 and Fe3 MS3 (M = Mo or V) cuboidal units share one central carbon atom, and are further

1.2 Biological Nitrogen Fixation

(a)

Mo nitrogenase 2 NH3 + H2 N2 + 8 e– + 8 H+ rt (1 atm) 16 Mg∙ATP + 16 H2O 16 Mg∙ADP + 16 H3PO4 V nitrogenase

N2 + 12 e– + 12 H+ (b)

(c)

rt 24 Mg∙ATP + 24 H2O

(1 atm)

2 NH3 + 3 H2 24 Mg∙ADP + 24 H3PO4

Fe-only nitrogenase 2 NH3 + 7.5 H2 N2 + 21 e– + 21 H+ rt (1 atm) 42 Mg∙ATP + 42 H2O 42 Mg∙ADP + 42 H3PO4

Figure 1.3 Proposed stoichiometry of biological nitrogen fixation by three types of nitrogenases: (a) molybdenum nitrogenase, (b) vanadium nitrogenase, and (c) iron-only nitrogenase. S S

Cys

C S Fe S Fe S Cys Fe S

(a) Fe

S Fe S Fe Cys Fe S O

Fe

(b) Cys

Cys S

S

NH

Fe S

O V

O

NH

O His

S

Fe S

CH2COO– CH2CH2COO

S Ser H S Fe O Fe Fe S S Cys S Fe

(c)

Cys S

S Fe S S Fe

S Fe Fe S

S Cys

S

HN Cys

–

Cys S S Cys S S S Ser H Fe Fe O Fe Fe S Fe Fe S S S S Fe Fe S S Cys Cys S Cys (d) Cys

N

S

S S Cys Fe S S Fe S

CH2CH2COO–

O

Fe

Cys

CH2COO–

N His

C S

Fe

O

S

Fe

C O

O

Fe S Mo

S S S

(e)

S

Fe

Fe

O

Cys

Figure 1.4 Structures of (a) FeMo-co in MoFe protein, (b) FeV-co in VFe protein, (c) P-cluster in the oxidized state in MoFe protein, (d) P-cluster in the reduced state in MoFe protein, and (e) [4Fe–4S] cluster in Fe protein.

bridged by three sulfur atoms for FeMo-co [60–62] or a combination of two sulfur atoms and one carboxylate for FeV-co [63], respectively. The structure of FeFe-co has not yet been determined crystallographically but has been spectroscopically supposed to have a similar structure to FeMo-co or FeV-co, where molybdenum or vanadium atom is substituted for the corresponding iron atom [51, 52]. As shown in Figure 1.4a,b, molybdenum and vanadium atoms are coordinatively saturated by the chelation of homocitrate, whereas the iron atoms surrounding the carbon atom have vacant sites. Thus, recent theories on the reaction mechanism of nitrogen fixation prefer coordinatively unsaturated iron atoms to molybdenum or vanadium atom where conversion of dinitrogen

5

6

1 Overviews of the Preparation and Reactivity of Transition Metal–Dinitrogen Complexes

into ammonia occurs, whereas the precise reaction pathways for the conversion of dinitrogen into ammonia remain arguable [64–72]. Thermodynamic favorability of the formation of ammonia in aqueous solution changes depending on the pH of the solution because proton transfers are involved in the reaction, and ammonia exists as an ammonium cation (pK a = 9.25) in acidic or neutral conditions. Standard transformed Gibbs energy of the reaction of dinitrogen, electrons, and protons to form ammonium cation and dihydrogen in a ratio of 2 : 1 in an aqueous solution is given as −159.7 kJ mol−1 per dinitrogen at pH 0 (a+H = 1) and zero ionic strength, which corresponds to standard electrode potential of +0.276 V. On the other hand, standard transformed Gibbs energy at pH 7 shifts to +239.8 kJ mol−1 , corresponding to standard apparent reduction potential of −0.311 V vs. SHE (standard hydrogen electrode) (Figure 1.5a) [73]. Thus, the reaction requires the introduction of appropriate reducing reagents such as ferredoxin (E′ ∘ value varies from −0.377 to −0.434 V at pH 7 from different biological sources) (Figure 1.5b) [74, 75] and hydrolysis of several ATPs (ATP = adenosine triphosphate; Figure 1.5c) [75–77]. The schematic shown in Figure 1.6 summarizes the key metabolic pathways related to nitrogen fixation by molybdenum nitrogenase, which consists of two component proteins: molybdenum–iron protein also called dinitrogenase or nitrogenase component 1 containing FeMo-co and P-cluster whose structures in different oxidation states are shown in Figure 1.4c,d [78] and iron protein also called dinitrogenase reductase or nitorgenase component 2 containing [4Fe–4S] cluster whose structure is shown in Figure 1.4e [79, 80]. An electron is transferred from ferredoxin or flavodoxin to the [4Fe–4S] cluster in iron protein, which docks with the aid of 2 M amount of ATP to molybdenum–iron protein to N2 (aq) + 8 e– + 10 H+ (a)

2 NH4+ + H2 (aq)

ΔrG′° = +239.8 kJ mol–1 E′° = –0.311 V 8 Fdred–

(b)

(c)

8 Fdox + 8 e– ΔrG′° = –304.6 kJ mol–1 E′° = –0.395 V

16 ADP3– +16 HPO42– + 16 H+ 16 ATP4– + 16 H2O (l) ΔrG′° = –602.2 kJ mol–1 N2 (aq) + 8 Fdred– + 16 ATP4– + 16 H2O (l)

2 NH4+ + H2 (aq) + 8 Fdox + 16 ADP3– + 16 HPO42– + 6 H+

(d)

ΔrG′° = –667.0 kJ mol–1

Figure 1.5 Standard transformed Gibbs energies and standard apparent reduction potentials of reactions in molybdenum nitrogenase at 25 ∘ C, 0 ionic strength, and pH 7: (a) nitrogen fixation, (b) reduction of ferredoxin (reduction potential based on the data obtained from Clostridium pasteurianum), (c) hydrolysis of ATP, and (d) total reactions. Stoichiometry in (c) and (d) is shown ignoring HATP3− (pK a = 7.60), HADP2− (pK a = 7.18), and H2 PO4 − (pK a = 7.22), but thermodynamic data in (c) and (d) are calculated considering these equilibria at pH 7 (not 16 H+ but 11.9 H+ for (c), not 6 H+ but 1.9 H+ for (d)).

1.2 Biological Nitrogen Fixation Glucose

Nitrogenase complex

Dinitrogenase Dinitrogenase O O reductase (MoFe protein) CH3C COH (Fe protein) Fdox HSCoA Ferredoxin Pyruvate:ferredoxin P-cluster FeMo-co or flavodoxin [4Fe-4S] oxidoreductase e– e– CO2 Fdred O CH3C SCoA Mg∙ADP + H3PO4 2 Mg∙ATP + 2 H2O 2 Mg∙ADP + 2 H3PO4 NAD(P)H Mg∙ATP + HSCoA Ferredoxin:NAD(P)+ oxidoreductase CH3COOH NAD(P)+

Fdox

N 2 + 8 H+

2 NH3 + H2

Fdox

Hydrogenase

QH∙ or QH2

2 H+ Fdred

Photosynthesis I Fdred

Q (quinones)

Nitrogenase

Nitrogenase

Figure 1.6 Metabolic relationship between nitrogen fixation by Mo nitrogenase and electron transfers from pyruvate degradation, hydrogen uptake, respiration, or photosynthesis by ferredoxin/flavodoxin.

transfer an electron from the [4Fe–4S] cluster to the P-cluster, from which the FeMo-co obtains electrons [81, 82]. The rate-determining step is the dissociation of iron protein from molybdenum–iron protein (6 s−1 at 25 ∘ C, pH = 7.4) [83], whereas the turnover of the formation of 1 M ammonia per molybdenum nitrogenase has been measured to be 1.5 seconds at 23 ∘ C by Thorneley and Lowe [84], who proposed a kinetic model of the catalytic cycle of nitrogenase reaction, where eight steps of reduction and protonation against dinitrogen occur for molybdenum nitrogenase (Figure 1.7) [64, 66, 69]. Although the amount of ATPs required for the reduction of 1 M dinitrogen has not been precisely determined by experiments, 16 ATPs are at least consumed by molybdenum nitrogenase (Figure 1.3a) based on the assumption that 2 ATPs are hydrolyzed for the transfer of one electron, whereas vanadium and iron-only nitrogenases consume at least 24 and 42 ATPs, respectively, based on the same assumption (Figure 1.3b,c) [51, 52, 55, 56]. In a typical stoichiometry by molybdenum nitrogenase, the standard transformed Gibbs energy of the reduction of dinitrogen is given as −667 kJ mol−1 at zero ionic strength (Figure 1.5d) [73, 75–77]. It must be noted that both diazene (HN=NH) and hydrazine (H2 N—NH2 ) are the substrates of nitrogenase to afford ammonia and that hydrazine is obtained as a minor product from the reduction of dinitrogen in appropriate reaction conditions [85, 86]. Without dinitrogen, protons can work as substrates to afford only dihydrogen [64]. In addition, other substrates such as ethylene, cyclopropene, acetylene, propyne, 1- or 2-butyne, allene, propargyl alcohol or amine, cyanide, cyanamide, several nitriles or isocyanides, diazirine, dimethyldiazenze, carbon monoxide, carbon dioxide, carbon disulfide, carbonyl sulfide, thiocyanate, cyanate [87], nitrite, hydroxylamine [88], or azide have been known to be reduced by nitrogenase [64]. Figure 1.7 denotes the Lowe–Thorneley kinetic model modified by Hoffman and coworkers [66, 69], where formation of at least an equimolar amount of dihydrogen is inevitable for the reduction

7

8

1 Overviews of the Preparation and Reactivity of Transition Metal–Dinitrogen Complexes

H2 N N H2

e–/H+

H H

M

M H H

N2 H H

H2

H H M

M

N2 H H

H E3

Dinitrogen complex

H

e–/H+

e–/H+

M

E2

M Resting state Eo

E1

M = CFe7MoS9 (FeMo-co)

NH

NH

N

NH

M H Diazenido complex E4

e–/H+

H2

e–/H+

Diazene complex

Distal pathway

NH M

M

Janus intermediate

NH3

Alternating Pathway NH2 NH2 e–/H+

NH2 M

NH3 e–/H+

M

Hydrazido(1–) Hydrazine complex complex E5

NH2

e–/H+ NH3

Amido complex

E6

E7

M Ammonia complex E8

NH2 N M

e–/H+

N M

e–/H+

Hydrazido(2–) Nitrido complex complex

NH M

e–/H+

Imido complex

Figure 1.7 Modified Lowe–Thorneley kinetic model of the conversion of N2 into NH3 and H2 on FeMo-co. Coordination of nitrogen- or hydrogen-containing ligands is shown as if they form mononuclear complexes, although M can be multimetallic centers and sulfur atoms where ligands may bridge or coordinate to different atoms.

of dinitrogen [89]. The first four reduction/protonation steps from E0 , the resting state of FeMo-co (CFe7 MoS9 ), give a (CFe7 MoS9 )(H+ )2 (H− )2 species, where hydrido can bridge several transition metal centers in FeMo-co, whereas protonation likely occurs on bridging sulfur atoms. Reductive elimination of dihydrogen and coordination of dinitrogen occur in the E4 “Janus” intermediate and then pairs of reduction/protonation on dinitrogen take place to afford 2 M amounts of ammonia and the starting resting E0 state. Here, the “alternating” reaction pathway where both distal and proximal nitrogen atoms are protonated stepwise and the “distal” reaction pathway where the first three protonation reactions occur at the distal nitrogen atom to give the nitrido intermediate can be drawn as shown in Figure 1.7, but the “alternating” pathway is highly likely because similar intermediates are spectroscopically observed when diazene or hydrazine is used as a reactant, and formation of hydrazine as an intermediary product is also detected. Ferredoxin or flavodoxin, the reducing reagent of nitrogenase, transfers electrons from several metabolites, but the main source of electrons is the degradation of pyruvate for both anaerobic and aerobic microorganisms (Figure 1.6). Hydrogenase can further recycle the dihydrogen produced in nitrogen fixation, thereby minimizing the loss of energy during nitrogenase catalysis. Ferredoxin or

1.3 Historical Background of Transition Metal–Dinitrogen Complexes

flavodoxin can also be reduced by NADH (nicotinamide adenine dinucleotide), NADPH (nicotinamide adenine dinucleotide phosphate), or quinones, which are produced by several metabolic pathways including both anaerobic and aerobic respiration or photosynthesis (Figure 1.6) [57–59, 90]. Cyanobacteria such as Anabaena variabilis perform oxygen-evolving photosynthesis and oxygen-inhibited nitrogen fixation in different cells (vegetable cells and heterocysts), or the former during day and the latter during night in the same cells, preventing the inactivation of nitrogenase by dioxygen gas [91, 92].

1.3 Historical Background of Transition Metal–Dinitrogen Complexes Biological nitrogen fixation was experimentally confirmed by 1888 [93–95], and lithium was reported to react with dinitrogen at room temperature and an atmospheric pressure to form lithium nitride (LiN3 ) that can be easily converted to ammonia in 1892 [96–99]. However, formation of other nitrido complexes from the reaction dinitrogen requires higher temperature [100, 101], and further reactivities of metals with molecular dinitrogen under ambient reaction conditions have been limited in number. In 1964, Haight and Scott have reported the detection of a small amount of ammonia on prolonged cathodic reduction of dinitrogen or reduction by stannous chloride in the presence of aqueous solution of molybdate and tungstate at room temperature, although the pressure of dinitrogen gas is not well documented in the literature [102]. Conversion of dinitrogen into ammonia using transition metal complexes under ambient reaction conditions has been first reported in 1964 by Vol’pin and Shur, who obtained a small amount of ammonia when dinitrogen gas at atmospheric pressure was passed through a mixture of anhydrous CrCl3 and LiAlH4 or EtMgBr in ether at room temperature [103]. Other transition metal complexes such as [Cp2 TiCl2 ] (Cp = η5 -C5 H5 ) or TiCl4 in combination with EtMgBr or i Pr3 Al also fixes dinitrogen [104–106]. Formation of aniline, p-toluidine, or aliphatic amines as a dinitrogen-derived nitrogen-containing compound has also been reported by bubbling dinitrogen through a mixture of [Cp2 TiCl2 ] or [Cp2 TiPh2 ], with PhLi, p-TolLi, EtMgBr, or n BuLi at an atmospheric pressure and room temperature, followed by further hydrolysis [107, 108]. Isolation of a series of transition metal–dinitrogen complexes where a molecular dinitrogen is coordinated to a transition metal was first reported in 1965 by Allen and Senoff [109, 110], who performed the reduction of [RuCl3 (H2 O)3 ] with hydrazine hydrate in water at room temperature to afford a ruthenium–dinitrogen complex [Ru(NH3 )5 (N2 )]2+ in the late 1963 (Figure 1.8a) [111]. At first, they mistakenly identified that they obtained a ruthenium–hydrido complex but later found that the compound was diamagnetic with a strong infrared band around 2170–2100 cm−1 attributable to the coordinated N≡N stretching, liberating dinitrogen gas on treatment with sulfuric acid. [Ru(NH3 )5 (N2 )]Cl2 also became the first transition metal–dinitrogen complex whose molecular structure was determined by a single-crystal X-ray analysis in 1966 [112].

9

10

1 Overviews of the Preparation and Reactivity of Transition Metal–Dinitrogen Complexes

[RuCl3(H2O)3]

(1) N2H4·H2O (2) aq HCl H2O, 25 °C

H3N

trans-[IrCl(CO)(PPh3)2]

CHCl3, 0 °C Ar = α-furyl or Ph

(b)

Ph3P

[Co(acac)3] (c)

Ether or toluene 0 °C

NH3 NH3

NH3 N N Ir

O PPh3 + ArCNCO

Cl N N

N2 (1 atm) AlEt2OEt, PPh3

Ru

H3N

(a) O ArCN3

2+

N N

Ph3P

Co H

PPh3 PPh3

Figure 1.8 Early reports of the preparation of transition metal–dinitrogen complexes by the reduction of metal centers in the presence of (a) hydrazine, (b) azide, and (c) dinitrogen.

The second example of transition metal–dinitrogen complexes was reported in 1966 by Collman and Kang, who obtained the iridium–dinitrogen complex trans-[IrCl(N2 )(PPh3 )2 ] by the reaction of Vaska’s iridium complex trans-[IrCl(CO)(PPh3 )2 ] with a variety of aromatic acyl azides in chloroform at 0 ∘ C (Figure 1.8b) [113–115]. The first transition metal–dinitrogen complex with the direct fixation of gaseous molecular dinitrogen was reported in 1967 by Yamamoto et al. who obtained the cobalt–dinitrogen complex [CoH(N2 )(PPh3 )3 ] by the reduction of [Co(acac)3 ] with AlEt2 OEt under atmospheric pressure of dinitrogen in the presence of PPh3 in ether or toluene (Figure 1.8c), which became the third example of isolated transition metal–dinitrogen complexes [116–118]. There was a confusion in the identification of its structure whether the compound contained a hydrido ligand or not, but it was later confirmed as a (hydrido)(dinitrogen) complex [119–124]. All the above three complexes are mononuclear complexes with a dinitrogen ligand coordinated to a metal center in an “end-on” manner. On the other hand, the binuclear transition metal–dinitrogen complex with a bridging dinitrogen was first reported in 1968 by Taube and coworkers, who prepared the diruthenium–dinitrogen complex trans-[{Ru(NH3 )5 }2 (μ-N2 )]4+ by the reduction of trans-[Ru(NH3 )5 Cl]2+ with zinc amalgam in water under an atmospheric pressure of dinitrogen (Figure 1.9a) [125]. This compound was first identified in 1967 as the same complex with Allen and Senoff’s complex [Ru(NH3 )5 (N2 )][BF4 ]2 , which shows a strong IR absorption band at 2154 cm−1 [109, 126], whereas a Raman band at 2100 cm−1 was observed for [{Ru(NH3 )5 }2 (N2 )][BF4 ]4 [127], whose molecular structure was determined crystallographically [128]. Preparation of the heterobimetallic dinitrogen-bridged transition metal– dinitrogen complex [(PMe2 Ph)4 ClRe(μ-N2 )MoCl4 (PEtPh2 )] was reported by Chatt et al. in 1969 via the ligand exchange reaction of a molybdenum phosphine

1.3 Historical Background of Transition Metal–Dinitrogen Complexes

[Ru(NH3)5Cl]2+

(a)

Zn–Hg aq H2SO4, rt Ar (1 atm)

[Ru(NH3)5(H2O)]2+

N2 (1 atm) H2O, rt

4+ NH3 NH3 NH3 NH3 H3N Ru N N Ru NH2 H3N

NH3 H3N

NH2 N

O PhCNHNH2·HCl Ph3P PPh3 [ReOCl3(PPh3)2] Ph3P Benzene/EtOH reflux

Cl Re Cl

N PMe2Ph (excess) PhMe2P N N C Ph Benzene/MeOH PhMe2P O reflux – PhCOOMe, HCl, 2 PPh3

[MoCl4(thf)2] Cl CH2Cl2/MeOH

(b) (1) PEtPh2 (3 equiv) Ph2EtP (2) NaBH4, 0 °C EtOH

(c)

PMe2Ph PMe2Ph

Cl

Cl PMe2Ph PMe2Ph Cl Re N N Mo OMe Cl PMe2Ph

Cl N

H2(1 atm) [FeCl2(H2O)2]2+

PhMe2P

Re

H H Fe

N PEtPh2 N2 (1 atm) H

Ph2EtP H

EtOH, rt

Ph2EtP

Fe

PEtPh2 H

Ph2EtP H N

N2 (1 atm) AlR3, dppe (2 equiv) [Mo(acac)3]

(d)

Toluene –40 °C to rt R = Et or iPr

Ph2 N P Mo P Ph2 N

Ph2 P P Ph2

N

Figure 1.9 Early reports of the preparation of transition metal–dinitrogen complexes of (a) dinuclear with bridging dinitrogen, (b) heterobimetallic dinuclear with bridging dinitrogen, (c) iron, and (d) molybdenum.

complex [MoCl4 (PEtPh2 )2 ] with the mononuclear rhenium–dinitrogen complex trans-[ReCl(N2 )(PMe2 Ph)4 ] [129–131], the dinitrogen ligand of which is originated from benzoylhydrazine [132–134]. The IR band attributable to the N≡N triple bond shifts from 1922 cm−1 for mononuclear rhenium complex to 1810 cm−1 for the heterobimetallic complex [129]. The molecular structure of its analogous complex [(PMe2 Ph)4 ClRe(μ-N2 )MoCl4 (OMe)] was later confirmed by an X-ray analysis (Figure 1.9b) [135, 136]. Preparation of another heterobimetallic dinitrogen-bridged transition metal–dinitrogen complex [(NH3 )5 Os(μ-N2 )Ru(NH3 )5 ]4+ was also reported in 1969 [137–140]. Preparation of iron– [141–145], molybdenum– [146–152], or vanadium– dinitrogen complexes [153–156] has been of great interest from the viewpoint of a model for the active site of nitrogenase. Sacco and Aresta reported the formation of the first iron–dinitrogen complex cis,mer-[FeH2 (N2 )(PEtPh2 )3 ] in 1968 by the reaction of dinitrogen with the dihydrogen complex cis,mer-[FeH2 (H2 )(PEtPh2 )3 ], which was first formulated as a dihydrido complex [FeH2 (PEtPh2 )3 ] [157], then reformulated as a tetrahydrido complex [FeH4 (PEtPh2 )3 ] [158, 159], but later identified as the dihydrogen complex based on the T 1 relaxation time measurement (Figure 1.9c) [160]. Thus, coordination of dinitrogen occurs by the

11

12

1 Overviews of the Preparation and Reactivity of Transition Metal–Dinitrogen Complexes

ligand exchange with a dihydrogen ligand rather than by the reductive elimination of two dihydrido ligands. The molecular structures of these complexes were later determined by X-ray and neutron diffraction studies [161]. Preparation of molybdenum–dinitrogen complex was first reported by Hidai et al. in 1969, who obtained the molybdenum–dinitrogen complex trans[Mo(N2 )2 (dppe)2 ] by the reaction of [Mo(acac)3 ] with aluminum-reducing reagents in the presence of dppe under atmospheric pressure of dinitrogen (Figure 1.9d) [162–164]. The structure of this compound was later determined by an X-ray crystallographic analysis [165]. It is very surprising that several transition metal–dinitrogen complexes have been prepared in the late 1960s within a few years since the first discovery of transition metal–dinitrogen complexes [166–168]. Preparation and identification of dinitrogen complex of vanadium, another important transition metal of nitrogenase, was reported comparably later, when Ihmels and Rehder have reported the preparation of the anionic vanadium–dinitrogen complex [V(CO)5 (N2 )]− by UV irradiation of [V(CO)6 ]− or [V(CO)5 (acetone)]− in 2-methylteterahydrofuran at 200 K in the presence of atmospheric dinitrogen in 1985 [169, 170]. The first vanadium–dinitrogen complex crystallographically analyzed was reported in 1989 by Gambarotta and coworkers, who succeeded in the preparation of dinitrogen-bridged divanadium complex [(V{o-(Me2 NCH2 )C6 H4 }2 (py))2 (μ-N2 )] (Figure 1.10a) [171]. For construction of biomimetic reactions based on the metal–sulfur clusters in metalloenzymes, a lot of sulfur-bridged transition metal clusters have been synthesized as models of nitrogenase [53, 141, 172–179], but the first dinitrogen complex [(Cp*Ir)3 {Ru(tmeda)(N2 )}(μ3 -S)4 ] (Cp* = η5 -C5 Me5 ) where dinitrogen is coordinated to sulfur-bridged transition metal cluster has been isolated rather recently by Mizobe and coworkers (Figure 1.10b) [180, 181]. N N

V py (a)

i

N Me2

py

N N

Pr2 P Fe i Pr2 P Si

(c)

N N

V N Me2

N N S

Me2 S N Ru S N Ir Me2 Ir S

NMe2

NMe2

Fe Si

Ir S

(b) iPr

i

i

Pr2 P i P Pr2

i i

Pr

Pr S

(d)

F

N N

Pr

iPr

Fe S

2–

iPr

F

Figure 1.10 Selected examples of transition metal–dinitrogen complexes as models for cofactors in nitrogenase: (a) vanadium–dinitrogen complex, (b) cubane–dinitrogen complex, (c) multimetallic iron–dinitrogen complex with Fe—S bonds, and (d) iron–dinitrogen complex with Fe—S and Fe—C bonds.

1.4 Coordination Chemistry of Transition Metal–Dinitrogen Complexes

Isolation of the sulfur-supported multimetallic iron complex [{Fe(N2 )}2 (μ-SAr)]− (Ar = 2,5-C6 H4 {Si(C6 H4 Pi Pr2 -o)2 }2 ) has also been reported more recently by Creutz and Peters (Figure 1.10c) [182]. Recent analyses of nitrogenase have clarified that FeMo-co contains a carbide atom that constitutes the edge of two cuboidal clusters ([Fe4 S3 ] and [Fe3 MoS3 ]) [60–62], where the carbon atom has been transferred from the methyl radical originated from S-adenosylmethionine to bridge the two clusters [183]. The iron–dinitrogen complex [Fe(N2 )(L)]2− (LH2 = 6,6′′ -F2 -3,3′′ -(2,4,6-i Pr3 C6 H2 )2 m-terphenyl-2,2′′ -(SH)2 ) bearing both Fe—S and Fe—C bonds has been prepared by Holland and coworkers in 2015 (Figure 1.10d) [179, 184].

1.4 Coordination Chemistry of Transition Metal–Dinitrogen Complexes 1.4.1 Coordination Patterns of Dinitrogen and Mononuclear Transition Metal–Dinitrogen Complexes Dinitrogen is a diatomic molecule with a Raman band at 2330, 2291, or 2252 cm−1 for gaseous 14 N2 , 14 N15 N, or 15 N2 , respectively, because of the stretching vibration of the N≡N triple bond [185, 186]. The interatomic distance between two nitrogen atoms has been measured to be ranging from 1.09 to 1.11 Å by X-ray analyses of several different phases of solid-state dinitrogen (α-, β-, γ-, and δ-N2 ) at very low temperatures or at extremely high pressures [187–198], whereas that of gaseous molecular dinitrogen calculated based on the spectroscopic data for the electronic ground state is 1.0977 Å (Figure 1.11a) [199, 200]. Three isomers are known for diazene or diimine: trans-diazene (Figure 1.11b), cis-diazene (Figure 1.11c), and isodiazene (H2 N+ =N− ) [201, 202]. trans-Diazene is the most stable isomer among them, but cis-diazene, only 21 kJ mol−1 higher in enthalpy than trans-diazene [202, 203], works as an hydrogenation reagent against unsaturated compounds with stereoselective syn addition of H2 [204] and is also regarded as an intermediary structure of the reduction of dinitrogen in nitrogenase reactions [205]. The interatomic distance between two nitrogen atoms in trans-diazene has been determined to be 1.247 Å based on the UV and Figure 1.11 Molecular structures of (a) N2 , (b) trans-N2 H2 , (c) cis-N2 H2 , and (d) N2 H4 .

N

N

H

1.0977 Å

N

1.247 Å 2330 cm–1 (Raman)

(d)

1.02–1.05 Å

112–113°

N

H

N

1.24–1.27 Å

H

1529 cm–1 (IR stretch for N N of trans isomer)

(a)

(b)

(c)

HA 106°

112°

N

N

HB

H

106.3°

N

1.029 Å

1.447 Å 1.015 Å

HB 107°

HA

1098 cm–1 (IR stretch for N–N in liquid phase)

91°

HA

HB N HB

HA

13

14

1 Overviews of the Preparation and Reactivity of Transition Metal–Dinitrogen Complexes

IR spectroscopies (Figure 1.11b) [206, 207], whereas spectroscopic observations for cis-diazene that has not been isolated in the pure form have been problematic [208]. The bond lengths and angles of cis-diazene shown in Figure 1.11c are those taken from theoretical calculations [203, 209, 210]. The melting point of free hydrazine is not low (+1.4 ∘ C), and the solid-state structure was analyzed by an X-ray analysis, which gave the N—N bond length at 1.46 Å at −15 ∘ C [211], whereas the electron diffraction studies and microwave spectroscopies gave the N—N bond distance at 1.447 Å (Figure 1.11d) [212–215]. The Raman and IR spectra give the stretching vibration for N–N in the range of 1076–1126 cm−1 , which can vary according to the phases of hydrazine (gas, liquid, or solid) [216–220]. It must be noted that the dihedral angle of the H–N–N–H is almost 90∘ because of the existence of lone pairs of nitrogen atoms, suggesting that the bond order of N—N in hydrazine is one. Based on the crystallographic data of compounds containing N—N bonds, bond distances of 1.10, 1.22, and 1.46 Å as reference values for triple-, double-, and single-bond orders are proposed [221]. Since 1965, a lot of transition metal–dinitrogen complexes have been prepared [166–168, 221–231], including both mononuclear dinitrogen complexes and dinitrogen-bridged multinuclear complexes. General bonding modes of dinitrogen in mononuclear and dinuclear transition metal–dinitrogen complexes are summarized in Figure 1.12a [221, 227]. End-on Mononuclear

M N

M

N

N

N

M

N M

N N

M

N

Dinuclear M N

M

N

N M

M

N M

N N

M

(a)

N N N

1πg*

N N

M

N

N

M

N M

M M

M N

N M

3σg*

N M

M

N

M

M

End-on Side-on

Side-on

N

M M

M N

N–N: 1.01–1.95 Å N···N: 2.38–2.95 Å

(b)

(c)

Figure 1.12 (a) General bonding modes of dinitrogen in mononuclear and dinuclear transition metal–dinitrogen complexes. (b) Schematics of molecular orbital interactions of mononuclear end-on-bound transition metal–dinitrogen complex. (c) Metric difference of side-on-bridged dinuclear transition metal–dinitrogen complexes and bis(nitrido)-bridged complexes.

1.4 Coordination Chemistry of Transition Metal–Dinitrogen Complexes

Dinitrogen is isoelectric with carbon monoxide, and the structure of dinitrogen complexes is closely similar to that expected for the corresponding carbonyl complexes, where dinitrogen works as a comparably weaker σ donor as well as a poorer π acceptor than carbon monoxide does [232, 233]. As shown in Figure 1.12b, coordination of dinitrogen to the transition metal center in an end-on manner can be described according to the Dewar–Chatt–Duncanson model where σ donation from the slightly antibonding filled 3σg orbital (HOMO, highest occupied molecular orbital) of dinitrogen to the suitable empty d orbital of the transition metal, and the backbonding from the suitable filled d orbital of the transition metal to the empty 1πg * orbital (LUMO, lowest unoccupied molecular orbital) of the dinitrogen [166–168]. End-on (or η1 )-coordinated dinitrogen in mononuclear dinitrogen complexes is roughly linear (M–N–N wider than 164∘ ) with an averaged bond distance of 1.11(5) Å in general. Dinitrogen is a weak Lewis base; thus, the π-accepting character plays an important role in the variety of property and reactivity of the dinitrogen ligand [221]. By coordinating to the transition metal complexes in an end-on manner, the stretching frequency of the N≡N triple bond always shifts from 2330 cm−1 [185] downward to the region of 2250–1800 cm−1 , depending on the π-donating character of the transition metal centers [143, 222–224]. The most investigated mononuclear transition metal–dinitrogen complexes have a d6 configuration in the metals (V(I–), Cr(0), Mo(0), W(0), Mn(I), Tc(I), Re(I), Fe(II), Ru(II), Os(II), and Ir(III)), whereas other mononuclear transition metal complexes include d2 (Ti(II), Zr(II)), d3 (Mo(III), W(III)), d4 (Mo(II), W(II), Re(III)), d5 (Mo(I), W(I), Re(II), Os(III)), d7 (Fe(I), Ru(I), Os(I)), d8 (Mn(I−), Fe(0), Ru(0), Os(0), Co(I), Rh(I), Ir(I), Ni(II)), d9 (Co(0), Rh(0)), and d10 (Mn(III−), Co(I−), Ni(0), Cu(I)) configurations [143, 221–224, 226]. In general, mononuclear transition metal– dinitrogen complexes have been well isolated and investigated for rather mid to late transition metals (groups 6–9) with rather low oxidation states enough to support backbonding with dinitrogen. Counting the molecular orbitals of dinitrogen, side-on (η2 ) coordination of dinitrogen to the mononuclear transition metal center is possible in addition to the end-on (η1 ) coordination, whereas most of the reported mononuclear transition metal–dinitrogen complexes have the end-on-bound dinitrogen ligand. There was an early claim in which the side-on-bound dinitrogen was crystallographically identified for a Vaska-type rhodium–dinitrogen complex [RhCl(η2 -N2 )(Pi Pr3 )2 ] [234], but a later study has suggested that the former X-ray analysis was in error and that the dinitrogen is actually end-on bound as [RhCl(η1 -N2 )(Pi Pr3 )2 ] [235]. Similarly, a paramagnetic zirconium–dinitrogen complex with the side-on coordination of dinitrogen [Cp2 ZrCH(SiMe3 )2 (η2 -N2 )] has been proposed based on its ESR (electron spin resonance) measurement [236, 237], but the actual structure has been recently revised by Chirik and coworkers as the anionic mixed valence dizirconium–dinitrogen complexes [Na(thf )6 ][(Cp2 Zr{CH(SiMe3 )2 })2 (μ-η1 :η1 -N2 )] where a dinitrogen ligand bridges two zirconium atoms both in an end-on manner [238]. On the other hand, Armor and Taube have already reported a linkage isomerism for the 15 N-labelled Allen and Senoff’s complex [Ru(NH3 )5 (N2 )]2+ and have proposed an end-to-end rotation through the intermediacy of a side-on

15

16

1 Overviews of the Preparation and Reactivity of Transition Metal–Dinitrogen Complexes

O (a)

C C O

Re 15N 14

N

N Os

H3N (b)

NH3

Re

15

N

14N

O

C C O

Re 14N 15N

2+

2+

N H3N

C O C O

N N

NH3

He/Cd laser (λ = 325 nm, 75 mW cm–2)

H 3N

NH3

1 hour, 110 K

H 3N

Os

NH3 NH3

NH3

Figure 1.13 Intramolecular linkage isomerism to form side-on-bound transition metal–dinitrogen complexes analyzed by (a) NMR and (b) X-ray crystallography.

dinitrogen complex in 1970 [239]. Cusanelli and Sutton have more firmly confirmed nondissociative and intramolecular linkage isomerization of the dinitrogen ligand for a series of rhenium–dinitrogen complexes (Figure 1.13a) [240]. More recently, side-on coordination of the dinitrogen to a single-metal center has been confirmed spectroscopically or crystallographically for Allen and Senoff’s complex [Ru(NH3 )5 (N2 )]2+ [241] or its osmium analog [Os(NH3 )5 (N2 )]2+ [242], where irradiation of UV light to the solid-state crystals of [M(NH3 )5 (N2 )]2+ (M = Ru [109–112] or Os [243, 244]) gave rise to the formation of metastable states of [M(NH3 )5 (η2 -N2 )]2+ , where the vibration of dinitrogen has shifted from 2025 to 1838 cm−1 for the osmium–dinitrogen complex for instance (Figure 1.13b) [241, 242]. Anyway, both experimental and theoretical studies have confirmed that the side-on coordination is always higher in energy than end-on in mononuclear dinitrogen complexes [245–251]. 1.4.2

Multinuclear Transition Metal–Dinitrogen Complexes

In contrast to the mononuclear dinitrogen complexes, bridging dinitrogen complexes can be prepared for both early and late transition metals (group 4–10). The most investigated dinuclear transition metal–dinitrogen complexes with the end-on-bound bridging μ-η1 :η1 -dinitrogen have a d2 (Ti(II), Zr(II), V(III), Nb(III), Ta(III), Mo(IV), W(IV)) or d6 (Cr(0), Mo(0), W(0), Mn(I), Tc(I), Re(I), Fe(II), Ru(II), Os(II), Ir(III)) configuration in the metals, whereas the other dinuclear transition metal complexes include d1 (Ti(III), Hf(III), Nb(IV), Ta(IV), Mo(V)), d3 (Ti(I), V(II), Nb(II), Ta(II), Mo(III), W(III)), d4 (V(I), Ta(I), Cr(II), Mo(II), W(II), Fe(IV)), d5 (V(0), Cr(I), W(I), Os(III)), d7 (Fe(I), Ru(I)), d8 (Fe(0), Ru(0), Co(I), Rh(I), Ir(I)), d9 (Co(0), Rh(0), Ni(I)), and d10 (Ni(0)) configurations [221–224]. Bridging end-on-coordinated dinitrogen in dinuclear dinitrogen complexes is roughly linear or slightly bent (M–N–N usually wider than 160∘ with exceptions [252]) and an averaged bond distance of 1.19(7) Å, corresponding to rather a double bond, and the bond lengths of bridging dinitrogen that tend to be elongated in early transition metal complexes compared to those in late transition metal complexes [221]. The longest bong lengths up to 1.40 Å almost identical to those of a single bond were observed by X-ray analyses for dinitrogen-bridged tungsten(III)

1.4 Coordination Chemistry of Transition Metal–Dinitrogen Complexes

complexes (Figure 1.14a) [253, 254]. Indeed, stretching vibration attributable to the NN bond is shifted sometimes drastically downward to the range of 2150 to 800 cm−1 [143, 222–224], which is usually observed in Raman spectroscopies, for symmetric structures found in the dinitrogen-bridged dinuclear complexes that tend to make the molecules Raman active but IR inactive. CH2tBu CH2tBu N N W

N

N

CH2tBu

i

Pr

N N

iPr

N

W

N

N N BuCH 2 CH2 CH2tBu 1.39(2) Å

W N N

N

tBu

(a)

t

i

N

N

N

W

iPr

Pr

1.277(8)–1.402(17) Å +

i

Pr

N iPr N Mo i Pr N N

N

i

Pr N Pr N

Mo

i

(b)

i

N

i

Pr

– e– e–

Pr

i

N Pr N

Fe

Mo –e

e–

Mes Mes

Mo

Fe – e–

iPr

2+

Mo

Fe

Mo

Fe

P Et2 1.256(9) Å –

2–

1.225(7)–1.233(8) Å

N

P Et2N Et N P2

e–

1.226(4) Å

Mes V N N V Mes Mes Mes

N Pr N

Mo

1.265(5) Å 1349 cm–1

P Et2

1.182(5) Å

Pr

Et2 P

Mo

Fe

N iPr N i Pr

N

i

i

Pr

P Et2N Et N P2

–

Mo N

+

P Et2

(d)

i

N

Et2 P

Mo

(c)

e–

Mo

N

– e–

1.239(4) Å 1503 cm–1

Et2 P P Et2N Et N P2

iPr

N

iPr

1.212(2)–1.217(2) Å 1630 cm–1

Fe

N iPr N Mo i Pr N

N

2+

–e– e–

Mes Mes

Mes V N N V Mes Mes Mes 1.222(4) Å

Figure 1.14 Metric features of end-on-bridged dinuclear transition metal–dinitrogen complexes: (a) longest bond lengths found by X-ray analyses and (b–d) change in bond lengths of dinuclear end-on-bridged transition metal–dinitrogen complexes by redox.

17

18

1 Overviews of the Preparation and Reactivity of Transition Metal–Dinitrogen Complexes

Change of bond lengths in bridging dinitrogen by redox is significantly observed for molybdenum–dinitrogen complexes [(Mo{Nt Bu(C6 H3 Me2 -3,5)}3 )2 (μ-N2 )] (Figure 1.14b) [255] and [{Cp*Mo(depf )}2 (μ-N2 )] (depf, 1,1′ -bis(diethylphosphino)ferrocene) (Figure 1.14c) [256], where stepwise oxidations can result in the elongation of bridging dinitrogen (1.21–1.27 Å or 1.18–1.26 Å) as well as the lowering of Raman band because of the NN bond stretching (1630 to 1349 cm−1 ) [255–257]. These complexes are also noteworthy that they give the corresponding nitrido complexes by thermal [258, 259] or photochemical methods [256]. On the other hand, bond lengths are shortened upon oxidation in cases for vanadium (Figure 1.14d) [260] or tantalum complexes [261, 262], thus are quite sensitive to the nature of the transition metal centers. Side-on-bridged transition metal–dinitrogen complex was first characterized crystallographically by Jonas, Krüger, and Tsay for a nickel complex [{(PhLi)6 Ni2 (N2 )(OEt2 )2 }2 ], where each of the two dinitrogen ligands bridges two nickel atoms in side-on manners perpendicular to the Ni—Ni bond with two or three lithium atoms additionally coordinating to each of the nitrogen atoms (Figure 1.15a) [263, 264]. The same group has also revealed the similar structure for [{Ph(NaOEt2 )2 Ph2 Ni2 (N2 )NaLi6 (OEt4 )OEt2 }2 ] [265], and the N—N bond lengths of these side-on-bound dinitrogen with end-on coordination to lithium atoms range between 1.35 and 1.36 Å [263–266]. In 1988, the side-on-bridged planar dinuclear dinitrogen complex analyzed crystallographically was first reported by Evans et al. for [(Cp*2 Sm)2 (μ-η2 :η2 -N2 )], which is also the first crystallographically analyzed f-block dinitrogen complex (Figure 1.15b) [267]. The bond length of the dinitrogen ligand in this compound is 1.09 Å, almost identical to a triple bond, whereas that of the second example of the side-on-bridged planar dinuclear dinitrogen complex analyzed crystallographically reported by Fryzuk et al. in 1990 for [({N(SiMe2 CH2 Pi Pr2 )2 }ZrCl)2 (μ-η2 :η2 -N2 )] is 1.55 Å with a Raman band at 731 cm−1 , even longer than the expected N—N single bond and enough to be assigned as (N2 )4− (Figure 1.15c) [268–270]. In 1991, Gambarotta and coworkers reported the isolation of the unique bis(side-on)-bridged transition metal–dinitrogen complex [({(Me2 Si)2 N}2 Ti)2 (μ-η2 :η2 -N2 )2 ] where two dinitrogen ligands bridge two titanium atoms in side-on manners (Figure 1.15d) [271]. Since then, several side-on-bridged dinitrogen complexes have been prepared and characterized, especially for f-block lanthanide (Ce, Pr, Nd, Sm, Gd, Tb, Dy, Ho, Er, Tm, and Lu) and actinide (U) complexes [272–274]. Except for f-block elements, d1 (Sc(II), Y(II), La(II), Zr(III)), d2 (Y(I), Ti(II), Zr(II), Hf(II)), d3 (Nb(II)), d5 (Cr(I)), and d10 (Ni(0)) configurations are known by counting dinitrogen as a neutral ligand, whereas the bond length of the dinitrogen ligand in the side-on-bound dinitrogen-bridged transition metal–dinitrogen complexes including f-block element–dinitrogen complexes ranges from 1.01 to 1.63 Å (Figure 1.12c) [221, 224, 229, 274]. Elongation of the bond length of the side-on-bound bridging dinitrogen ligand has been well observed for rare earth element–dinitrogen complexes, where the one-electron reduction of the dinuclear species assigned as with an (N2 )2− ligand results in the formation of an anionic species assigned as with an (N2 )3− ligand (Figure 1.15e) [275–277]. For the side-on-bridged dinuclear transition metal– dinitrogen complexes, the oxidation states of neutral (N2 ) to more reduced

1.4 Coordination Chemistry of Transition Metal–Dinitrogen Complexes

Ph Li Ph Ph Ni Li Ph

Ph

Li Ni

N

Li

Li

OEt2

Sm

Ph

N

Ph

N Li

Li

Li

N

Ph

Ph

Ni

(Me3Si)2N

Y

Li

Y

1.088(12) Å

thf

e–

N(SiMe3)2 – e– thf

1.285(4) Å 1425 cm–1

(e) tBu

t

Y

t

Bu tBu

O O

Bu

2– tBu

O

Nb N

O

tBu

(f)

t

Si Me2

PiPr2

N

(c)

Si Me2

1.548(7) Å 731 cm–1 –

tBu

Bu Bu N–N: 1.390(17) Å

tBu

Nb

O

t

Bu

O

O O

O

Na N

N Na O

Nb O

O

O Na

O

tBu t

Cl

Zr

thf

O

O (1) 2 Na, THF (2) DME

N Nb O O

N

1.401(6)–1.405(3) Å 989 cm–1

tBu

O

N

P

iPr 2

N(SiMe3)2

Y N

(Me3Si)2N

Me2 Si

N(SiMe3)2

N

(Me3Si)2N

N

Zr

iPr P 2

N (Me3Si)2N N(SiMe3)2 N Ti N Ti (Me3Si)2N N N(SiMe3)2 (d) 1.379(21) Å

N(SiMe3)2

N

(Me3Si)2N

(b)

Ph

N–N: 1.35 Å

N

Pr2P

Li Ni

Ph OEt2

thf

i

Sm

Cl

Li Ph

N N

Li

Et2O

(a)

Me2 Si

OEt2

O

O

O

t tBu

O

Na

Bu

tBu

O

N···N: 2.598(8) Å

Figure 1.15 Metric features of side-on-bridged dinitrogen complexes: (a) first crystallographically characterized side-on-bridged transition metal–dinitrogen complex, (b) first crystallographically characterized side-on-bridged dinuclear f-block element–dinitrogen complex, (c) first crystallographically characterized side-on-bridged dinuclear d-block element–dinitrogen complex, (d) first crystallographically characterized bis(side-on)-bridged dinuclear dinitrogen complex, (e) change in bond lengths of dinuclear bis(side-on)-bridged transition metal–dinitrogen complexes by redox, and (f ) further reductive cleavage of the dinitrogen ligand.

forms ((N2 )2− , (N2 )3− , and (N2 )4− ) have been proposed based on the elongation of N—N bond lengths or vibration spectroscopies [221, 229, 274], whereas the N—N bond lengths determined by X-ray diffraction experiments can be sometimes underestimated and may not be appropriate to assess the level of dinitrogen reduction [278, 279]. It must be noteworthy that further reduction of dinitrogen by bimetallic systems can result in the formation of bis(nitrido)-bridged complexes with two (μ-N)3− bridging ligands [280–285]. For example, Floriani and coworkers have

19

20

1 Overviews of the Preparation and Reactivity of Transition Metal–Dinitrogen Complexes

reported that the two-electron reduction of the dianionic dinitrogen-bridged dinibobium complex [Na(L)x ]2 [{(p-t Bu-calix[4]arene)Nb}2 (μ-N2 )] (L = diglyme or thf ) affords the corresponding bis(nitrido)-bridged complex [280] (Figure 1.15f ). Bis(nitrido)-bridged complexes can be significantly distinguishable from the side-on-bridged dinitrogen complexes by metric features, for the two nitrido nitrogen atoms separated by 2.4–3.0 Å (Figure 1.12c) [221]. The third coordination mode of dinitrogen (μ-η1 :η2 -N2 , Figure 1.12a) for a dinuclear complexes was reported in 1998 by Fryzuk et al. for a ditantalum complex [Ta2 (μ-H)2 {PhP(CH2 SiMe2 NPh)2 }2 (μ-η1 :η2 -N2 )], where the bridging dinitrogen coordinates to the metal centers both in end-on and side-on manners (Figure 1.16a) [286, 287]. A similar structure was also found for a dizirconium complex (Figure 1.16b) [252], where the N—N bond lengths (1.20–1.32 Å) correspond to double or between double and single bonds. However, a similar coordination has been already reported by Pez et al. in 1982 for the fulvalene-bridged tetranuclear titanium complex [{Ti2 (η1 :η5 -C5 H4 )Cp3 }{Ti2 (η5 :η5 -C10 H8 )Cp2 }(μ3 -η1 :η1 :η2 -N2 )], where the dinitrogen ligand bridges three titanium atoms in end-on, end-on, and side-on manners (Figure 1.16c) [288]. Dinitrogen as a bridging ligand coordinating to three transition metal centers has also been recently reported by Chirik’s and Murray’s groups for titanium (Figure 1.16d) [289] and copper (Figure 1.16e) [290] complexes, respectively. Furthermore, dinitrogen ligand coordinating to six transition metal centers has been reported for hexanuclear gold complex [(AuPPh3 )6 (μ6 -N2 )]2+ , iPr

Ph

Ph Ph N

H H

Ti

P

SiMe2 Me2Si N Ta Ta N SiMe 2 Me2Si N N P N Ph

Zr N

Ti

Cl

N

Zr N

N

Ti

Ti

Ph

Ph 1.319(4) Å

(a)

(b)

1.196(4)–1.220(10) Å 1583 cm–1

1.301(12) Å 1282 cm–1

(c)

2+ N Ti N Ti

N Cu N

N Ti

N

Cu N N N Cu N

PPh3

Ph3P Au Ph3P Au

Au N N

Au

Au PPh3

Ph3P 1.320(3) Å (d)

(e)

1.0854(1)–1.0956(1) Å 1968–1952 cm–1

Au PPh3

1.45 Å (f)

Figure 1.16 Metric features of transition metal complexes with side-on–end-on-bridged or multidentate dinitrogen ligand: bidentate side-on–end-on-bridged dinuclear transition metal–dinitrogen complexes of (a) tantalum and (b) zirconium; tridentate side-on–end-on-bridged transition metal–dinitrogen complexes of (c) tetranuclear titanium, (d) trinuclear titanium, and (e) trinuclear copper; and (f ) hexanuclear transition metal–dinitrogen complex of gold.

1.5 Chemical Activation and Reactivity of Dinitrogen Using Transition Metal Complexes

Na(thf)3 N Ad Ad

N N

N Ad

Mo

Mes Mes

N

1.152(8) Å (a) 1757–1783 cm–1

Mes V N N V Mes Na

i tBuPr

N

Na

iPr t

i

Pr N

Bu

Fe N N Fe tBu

1.28(2) Å (b)

iPr

(c)

N i

Pr

Pr Na iPr

i

N t iPr Bu

1.213(4) Å 1583 cm–1

Figure 1.17 Selected examples of transition metal–dinitrogen complexes containing (a) end-on, (b) side-on, and (c) bis(side-on) interactions with sodium.

which was prepared from the reaction with hydrazine [291]. However, the N—N bond length (1.45 Å) is long enough to be identified as a hydrazido(4−) complex (N2 )4− with the oxidation states of two gold atoms at +1. As already shown in Figure 1.15a [263–266], alkaline metal and alkaline earth metal elements can bind even to dinitrogen ligands coordinated to transition metal centers in both end-on and side-on manners, and its coordination chemistry is more complicated than that of transition metal–dinitrogen complexes [292, 293]. Selected examples of transition metal–dinitrogen complexes where sodium is coordinated in end-on [294], side-on [295], or bis(side-on) manners [296] are shown in Figure 1.17.

1.5 Chemical Activation and Reactivity of Dinitrogen Using Transition Metal Complexes 1.5.1

Protonation of Transition Metal-bound Dinitrogen

Since the first discovery of transition metal–dinitrogen complex [109], activation and transformation of dinitrogen into ammonia or other nitrogen-containing organic or inorganic compounds under ambient conditions using dinitrogen complexes have been a frontline topic in synthetic chemistry [50–54, 72, 141– 156, 221–225, 227–231, 297–313]. Formation of ammonia from dinitrogen by using transition metal complexes has already been reported by Vol’pin and Shur in 1960s [103–106], and several transition metal complexes where dinitrogen is not coordinated have been found to afford ammonia or hydrazine [314–322]. Ammonia formation from the decomposition of transition metal–dinitrogen complexes was first confirmed for titanium complexes in early 1970s, where small amounts of ammonia were obtained on treatment of reducing reagents or acids [320, 323–327]. In 1975, Chatt et al. have reported that the reaction of zero-valent group 6 transition metal (Mo and W)–dinitrogen complexes with sulfuric acid affords stoichiometric amount of ammonia (Figure 1.18) [328, 329], whereas Brûlet and van Tamelen have also reported the formation of ammonia for the molybdenum dinitrogen

21

22

1 Overviews of the Preparation and Reactivity of Transition Metal–Dinitrogen Complexes

N N PhMe2P PhMe2P

M

PMe2Ph H2SO4 (excess)

Dinitrogen complex M = Mo and W HX X = Cl, Br, I NH N PhMe2P PhMe2P

rt

N N PMe2Ph

M

up to 1.98 equiv / W

2 HX

PMe2Ph HX

X PMe2Ph

Diazenido complex

2 NH3 + N2 + M(VI) species

NH2 PhMe2P

N M

H2SO4 PMe2Ph or KOH X

PhMe2P

NH3 + N2H4

X Hydrazido complex

Figure 1.18 Protonation of dinitrogen bound to Mo or W centers to afford ammonia or hydrazine.

complex trans-[Mo(N2 )2 (dppe)2 ] on treatment with hydrobromic acid in the same year [330]. Since then, stoichiometric conversion of dinitrogen toward ammonia, hydrazine, or other nitrogen-containing compounds by the reaction of group 6 molybdenum– and tungsten–dinitrogen complexes with Brønsted acids, alcohols, or water has been surveyed by Chatt’s [222] and Hidai’s groups [223, 331–335]. As shown in Figure 1.12b, the dinitrogen ligand coordinated to the electron-rich transition metal centers is activated toward the attack by electrophiles, which leads to the formation of ammonia or hydrazine (Figure 1.18) [328–330, 336–344]. Stepwise protonation of the dinitrogen complexes leads to the formation of diazenido (MN=NH), hydrazido (MN—NH2 ), and hydrazidium (MN—N+ H3 ) complexes, which gives ammonia as a major product sometimes with the formation of hydrazine as a minor product on further treatment with acids or bases (Figure 1.18) [345–360]. A hypothetical catalytic cycle called “Chatt cycle” (Figure 1.19) has been proposed based on these isolated intermediates in addition to nitrido (MN), imido (MNH), amido (MNH2 ), and ammonia (ammine) complexes (MNH3 ), whereas the definition of the Chatt cycle, especially explication of oxidation states of the transition metal centers, differs among researchers [150, 222, 361–366]. In a typical Chatt cycle based on the “distal” reaction pathway as drawn in Figure 1.19, an equimolar amount of ammonia is produced by protonation at the distal nitrogen atom of the end-on-coordinated dinitrogen and nitrogen–nitrogen bond split to afford nitrido complexes originated from the proximal nitrogen atom of the dinitrogen ligand, where the following protonation steps occur to afford another equimolar amount of ammonia. Thus, 2 M amount of dinitrogen is obtained per cycle, whereas six-electron reduction is necessary to recover the starting dinitrogen complex; thus, an appropriate choice of proton sources and reducing reagents is required. In the original paper by Chatt, this catalytic cycle was proposed for biological nitrogen fixation upon FeMo-co, while “alternating” pathway leading

1.5 Chemical Activation and Reactivity of Dinitrogen Using Transition Metal Complexes

N2

NH3

H H H N P

P

N N

P P M H+ e– H P P H H P P H+ N N M(0) M(0) P P N Dinitrogen P M + P M P P complex P P M(I) H+ Ammonia M(II) H+ H H Diazenido complexes N complex P P H H N M + P P M(I) N+ P P Amido 2 e– M complexes H +H P M(IV) P N P P + 2 e– Hydrazido M H H P P complexes N M(III) Imido N H H+ P P complexes M N Hydrazidium P + P H P P H +H complex M M(II) N P P H+ H Nitrido M(III) N + P + P complex e– N M P + P M P P N P P M(IV) P + P M(IV) M H+ P P M(IV) NH M

3

Figure 1.19 An example of proposed Chatt cycle.

to the formation of hydrazine was concluded to be rather a side reaction [222], although the alternating pathway is now more probable for nitrogenase reactions (Figure 1.7) [66, 69]. Other mononuclear dinitrogen complexes of vanadium [153–156, 367–370], chromium [371, 372], iron [142–145, 373–390], and cobalt [390–393] also gave but comparably lower yields of ammonia or hydrazine when treated with Brønsted acids or other proton sources. For instance, Tyler and coworkers have reported that the reaction of an iron–dinitrogen complex with trifluoromethanesulfonic acid (HOTf ) affords a mixture of ammonia and hydrazine. Three possible reaction pathways (symmetric H-addition pathway, asymmetric H-addition pathway, and bridging-N2 pathway) can be drawn by connecting all the isolated intermediary complexes as shown in Figure 1.20, where the asymmetric H-addition pathway is the most likely with both proximal and distal nitrogen atoms protonated stepwise based on the DFT calculations [143, 373–381].

23

24

1 Overviews of the Preparation and Reactivity of Transition Metal–Dinitrogen Complexes

P P

HOTf Et2O/THF

N N

Fe P

NH3 + N2H4 + Fe(II) species 17% 2%

P

Asymmetric H-addition pathway

Symmetric 2 H+, 2 e– H-addition pathway

H+ +

N N

P

P

P

P Fe

NH Fe

P

P

NH

P

H

P 2 H+, 2 e–

H+

Bridging-N2 pathway

+

NH HN

Fe N N Fe

Fe P

P

P

NH2

P P

P

P

NH Fe

+ H2 + N2 P

P

H

P

P

P

P

+

2+ P

P

P

H+

2 H+, 2 e– NH2

H+

+

2+ P

H2N P

P Fe P

P H

P Fe

NH3, N2H4

NH2

P

+

NH2

P

Fe(II) species

P

P =

MeOH2CH2CH2C MeOH2CH2CH2C

P

P

CH2CH2CH2OMe CH2CH2CH2OMe

Figure 1.20 Three possible reaction pathways for the formation of ammonia and hydrazine by using an iron–dinitrogen complex.

Treatment of several dinitrogen-bridged early transition metal–dinitrogen complexes of zirconium [268, 269, 327, 394–397], niobium, or tantalum [398–402] with acids gives hydrazine as a major product. For example, the reaction of [{Cp*2 Zr(N2 )}2 (μ-N2 )] with hydrochloric acid was reported by Bercaw and coworkers to afford an equimolar amount of hydrazine and 2 M amount of [Cp*2 ZrCl2 ] (Figure 1.21a) [327, 394–397]. A similar reactivity has also been reported for tantalum– or niobium–dinitrogen complexes [{M(N2 S2 CNEt2 )3 }2 (μ-N2 )] (M = Nb, Ta) where stepwise protonation reactions of the bridging dinitrogen ligand with hydrogen halide to afford an equivalents of hydrazine have been confirmed (Figure 1.21b) [398–401]. On the other hand, a mixture of ammonia and hydrazine is obtained for other dinitrogen-bridged early transition metal–dinitrogen (titanium [288] and

1.5 Chemical Activation and Reactivity of Dinitrogen Using Transition Metal Complexes

N N Zr N N

4 HCl

Zr

Toluene –80 °C

N (a)

N2H4 + 2 N2 + 2 [Cp*2ZrCl2] 86%

N NEt2

Et2N

S

S M

S S

NEt2 S S

N N Et2N

S

S S

Et2N

M S

4 HX

S S

rt NEt2

NEt2 HX

(b)

S M S S N2H4 + 2 S S >90% Et2N X

NEt2

HX H [M N NH2]+ + [MX]+

M = Nb, Ta X = Cl, Br M = M(S2CNEt2)3

H H H HX [M N N M]+ [M N N

X S

M]2+

HX

H H2 [M N N MX]2+

Figure 1.21 Reaction of dinitrogen-bridged early transition metal–dinitrogen complexes with hydrogen halide to afford hydrazine for (a) zirconium and (b) niobium or tantalum complexes.

vanadium [260, 295, 403]) or heterobimetallic dinitrogen complexes [404, 405], whereas the dinitrogen-bridged dinitrogen complexes of middle-to-late transition metals (molybdenum, tungsten [406, 407], or nickel [266]) afford ammonia as the major product. 1.5.2

Cleavage of Transition Metal-bound Dinitrogen

In 1995, Laplaza and Cummins have reported that the dinitrogen-bridged molybdenum complex [(Mo{Nt Bu(C6 H3 Me2 -3,5)}3 )2 (μ-N2 )] as already shown in Figure 1.14b [255] affords a pair of nitrido complexes with thermal cleavage of dinitrogen (Figure 1.22a) [258, 259]. As shown in Figure 1.15f, dinitrogen can be cleaved to afford the nitrido complexes by reduction [280]. Furthermore, photochemical cleavage of dinitrogen has been reported by Floriani and coworkers, where cleavage of the dinitrogen-bridged dimolybdenum complex [{Mo(Mes)3 }2 (μ-N2 )] occurs under irradiation of UV light to give the corresponding nitrido complex [Mo(N)(Mes)3 ] [408], which reacts with [Mo(Mes)3 ] species to afford the nitrido-bridged dimolybdenum complex [{Mo(Mes)3 }2 (μ-N2 )] (Figure 1.22b) [409]. Since then, several transition metal– nitrido complexes [410, 411] have been prepared by thermal, reductive, or oxidative cleavage of the nitrogen ligand coordinated to transition metals [254–256, 261, 280–285, 294, 389, 412–435]. Further treatment of the nitrido complexes with acids leads to the formation of ammonia [256, 284, 285, 389, 412–416, 436, 437]. Reversely, coupling of nitrido complexes can lead to the regeneration of dinitrogen complexes [256, 438–448]. For example, irradiation of visible light to the neutral dinitrogen-bridged dimolybdenum complex [{Cp*Mo(depf )}2 (μ-N2 )] leads to the formation of the nitrido

25

26

1 Overviews of the Preparation and Reactivity of Transition Metal–Dinitrogen Complexes

tBu tBu

N Mo N

N

tBu

tBu

Mo N N

N2 (1 atm)

N

–35 °C

t

N tBu N tBu

N Bu Mo tBu tBuN N N

t

30 °C t1/2 = 35 min Bu Mo t Bu tBuN N N

(a) Mes Mo N N Mes Mes (b)

Mes Mes

– N2 N

hν (λ = 365 nm) 2

Mo

Mes Mo Mes Mes

Mes Mes

Mo

Mes Mo N

Mes Mes Mes

Mes

Mo Mes 2+

Et2 P

Et2 P Fe

Mes Mes

Mo P Et2N Et2 N P Mo

Et2 P

hν (λ > 400 nm) 2 Fe

Fe

Mo P N Et2

Fe 2 Fc+ rt

Mo P N Et2 Et N P2 Mo

Fe

P Et2

P Et2 – 2 e– (c)

2 e–

Figure 1.22 Cleavage of dinitrogen by (a) thermal or (b) photochemical methods. (c) Regeneration of the dinitrogen ligand by coupling of the nitrido complexes.

complex [Cp*Mo(N)(depf )], which can be coupled to afford the dicationic dinitrogen-bridged dimolybdenum complex [{Cp*Mo(depf )}2 (μ-N2 )]2+ by the oxidation with ferrocenium cation (Figure 1.22c) [256]. As already shown in Figure 1.14c, the electron configurations of the dinitrogen-bridged core can be changed stepwise by redox processes; thus, the starting neutral dinitrogen complex can be regenerated by redox processes from the cleaved nitrido complex (Figure 1.22c) [256, 449]. Thus, both cleavage and reformation of molecular dinitrogen are induced by a pair of two different external stimuli (photochemistry and redox) under ambient reaction conditions. 1.5.3

Reaction of Transition Metal-bound Dinitrogen with Dihydrogen

Reaction of gaseous dihydrogen with dinitrogen complexes has been apt to lead to the loss of dinitrogen to form hydride complexes [450–452]; thus,

1.5 Chemical Activation and Reactivity of Dinitrogen Using Transition Metal Complexes N Ph2 N P W P Ph2 N N

Ph2 P P Ph2

BF4 +2

Ar2 Ru P P Ar2

BF4 Ru

H

H Ar2P

H

H PAr2

Two tautomers in the ratio 1 : 1.6 Ar = 4-CF3C6H4 NH2 Ph2 N P W P Ph2 F

THF, rt (a)

BF4 Ph2 P

+2

P Ph2

Ar2 Ru P P Ar2

+ BF3·(thf) H

+ PPh2 Ph2P

Ru

Cl

Ph2P PPh2

N N PhMe2P PhMe2P

(b)

PMe2Ph +6 N N PMe2Ph

W

H2 (1 atm)

+

H H Ph2 P Ru P Ph2 Cl

Ph2 P P Ph2

2 NH3 + 6 55 °C 24 hours 0.55 equiv W−1

Ph2 H P Ru P Ph2 Cl

Ph2 P P Ph2

+ W(VI) + N2 species

Figure 1.23 Reaction of tungsten–dinitrogen complex with acidic ruthenium–η2 -dihydrogen complexes to form (a) hydrazido complex or (b) ammonia.

appropriate activation of dihydrogen was required for the direct reaction of transition metal-bound dinitrogen and dihydrogen under ambient reaction conditions. Indeed, Morris and coworkers have reported that the dinitrogen ligand of the tungsten–dinitrogen complexes trans-[W(N2 )2 (dppe)2 ] can be protonated by the acidic ruthenium–η2 -dihydrogen complex to afford the hydrazido complex (Figure 1.23a) [453]. When the tungsten–dinitrogen complex cis-[W(N2 )2 (PMe2 Ph)4 ] was treated with an excess amount of the cationic ruthenium–η2 -dihydrogen complex trans-[RuCl(η2 -H2 )(dppp)2 ]+ under the atmospheric pressure of dihydrogen at 55 ∘ C, 0.55 equiv of ammonia was obtained based on the tungsten atom, with the dinitrogen ligand likely protonated by the proton (H+ ) formed from the heterolytic cleavage of dihydrogen on the ruthenium center, affording the stoichiometric amount of hydrido ruthenium complex trans-[RuHCl(dppp)2 ] (Figure 1.23b) [454, 455]. Ammonia has also been obtained by the reaction of the tungsten–dinitrogen complexes and their derivatives with the sulfido-bridged dinuclear molybdenum complex [(Cp2 Mo)2 (μ-S)(μ-SH)(μ-S2 CH2 )]+ under the atmospheric pressure of dihydrogen [456, 457]. Reaction of dihydrogen gas with the side-on-coordinated dinitrogen-bridged dizirconium complex to afford the (diazenido)(hydrido)-bridged dizirconium

27

28

1 Overviews of the Preparation and Reactivity of Transition Metal–Dinitrogen Complexes Me2 Me2 Si Si Si N N Si Me2 PhP Me2 PPh Zr N

N

Me2 Me2 Si Si Si N N Si Me2 PhP Me2 PPh Zr

H2 (1–4 atm)

N

N

H

H

Toluene

Zr Zr PhP Me Me2 PPh PhP Me 2 Me2 PPh 2 Si Me2 Me2 N N Si Si N N Si H2(1 atm) Si Si Si Si Si Si Me2 Me2 Me2 Me2 Si N N Si Me2 PhP Me2 PPh Zr Hexane THF N N H H Zr PhP Me Me2 PPh 2 Si N N Si Si Si Me2 Me2

(a)

N Zr – H2

Zr

N H2

85 °C

Zr

N N

Zr

H2 (1 atm) 22 °C

H H N H Zr N Zr H

H2 (1 atm)

H

Zr

2

85 °C

H

+ 2 NH3 10–15%

4 HCl 2

Zr

Cl + 2 NH + H 3 2

Cl

(b)

Figure 1.24 Reaction of dinitrogen-bridged zirconium complex with gaseous dihydrogen to afford (a) diazenido-bridged complex or (b) ammonia.

complex has been reported by Fryzuk and coworkers, where the (dinitrogen)(dihydrogen)-bridged complex has been isolated as an intermediary complex (Figure 1.24a) [458–460], whereas formation of ammonia has not been successful. Similar diazenido-bridged dizirconium or dihafnium complexes have been isolated by Sita and coworkers from the reaction of dinitrogen-bridged complexes with dihydrogen gas [461]. The reaction of side-on-coordinated dinitrogen-bridged dizirconium complex [{(η5 -C5 Me4 H)Zr}2 (μ-η2 :η2 -N2 )] with gaseous dihydrogen has also been reported by Chirik and coworkers, where 10–15% yield of ammonia is obtained via the formation of the bis(imido)-bridged dizirconium complexes [{(η5 -C5 Me4 H)Zr}2 (μ-η2 :η2 -N2 )], from which stoichiometric amount of ammonia is obtained on treatment with hydrochloric acid (Figure 1.24b) [462–464]. This result is perfectly in contrast to that obtained for [(Cp*Zr)2 (μ-η2 :η2 -N2 )],

1.5 Chemical Activation and Reactivity of Dinitrogen Using Transition Metal Complexes

which affords a stoichiometric amount of hydrazine on treatment with hydrogen chloride (Figure 1.21a) [394–397], demonstrating that the small difference in the design of auxiliary ligands can drastically affect the reactivity of the coordinated dinitrogen [465–468]. Stepwise hydrogenation of dinitrogen by using poly(titanium) system has been reported by Hou and coworkers [416]. Hydrogenation of nitride to afford ammonia has also been recently reported by using a ruthenium complex with a PNP-type pincer ligand [469, 470]. 1.5.4

Functionalization of Transition Metal-bound Dinitrogen

Functionalization of dinitrogen into nitrogen-containing compounds has been first reported by Vol’pin and Shur, who obtained amines by reaction with electrophiles [107, 108]. However, most of the conversion of dinitrogen into nitrogen-containing compounds has been remaining substoichiometric based on the amount of transition metals except for the catalytic formation of ammonia [471–473], hydrazine [474, 475], and silylamines [476–479]. As the dinitrogen ligand coordinated to the electron-rich transition metal centers is activated toward the electrophiles (Figure 1.12b), the reaction of dinitrogen with electrophiles has been well investigated [147, 222, 223, 227, 331– 334, 465, 480–484]. Acylation of dinitrogen complex to afford acyldiazenido or acylhydrazido complexes was first reported by Chatt et al. in 1972 [485, 486], whereas alkylation of the dinitrogen complexes to afford alkyldiazenido or alkylhydrazido complexes has also been reported by the groups of Chatt and George (Figure 1.25a) [487–490]. Further destructive decomposition of the dialkylhydrazido complexes with reducing reagents, acids, or bases affords a stoichiometric amount of amines or alkylhydrazines (Figure 1.25a) [491–493]. Hidai and coworkers have revealed that the hydrazido complexes obtained from the corresponding dinitrogen complexes give diazoalkane complexes on treatment with ketones or aldehydes, which liberate a stoichiometric amount of ammonia, hydrazine, alkylamines, or ketone azines via decomposition by reducing reagents, acids, or bases (Figure 1.25b) [494–497]. Isolation of silyldiazenido complexes by silylation of dinitrogen complexes has also been successful for the group 6 transition metals, where a stoichiometric amount of silylamines or ammonia is obtained by further treatment with reducing reagents [477, 498–501] (Figure 1.25b). Similarly, boryldiazenido complexes have been isolated by the borylation reaction of the group 6 transition metal–dinitrogen complexes (Figure 1.25a) [502], where alumination [503–505], gallation [506], or germylation [501] of the dinitrogen ligand are also possible to afford the corresponding isolable group 6 transition metal complexes. Hydrosilylation, hydroboration, and hydroalumination of transition metal–dinitrogen complexes have been investigated by Fryzuk and coworkers for tantalum complexes [507–509]. The reaction of carbon monoxide with dinitrogen to afford isocyanate (N− =C=O) was first reported by Sobota and Janas [510], who obtained N,Ndimethylacetamide on the treatment of titanium complex with iodomethane (Figure 1.26a) [510–512]. Chirik and coworkers have achieved dinitrogen functionalization with carbon dioxide or carbon monoxide by using dinitrogen–

29

30

1 Overviews of the Preparation and Reactivity of Transition Metal–Dinitrogen Complexes + R C

R = Me, Et, Ph, etc.

N Ph2 N P M P Ph2 N N

Ph2 P

RCOX

P Ph2

N

O

Ph2 N P M P Ph2 X

Ph2 P

X = Cl, Br, I 1

R C

H+

P Ph2

R1

R , R = Me, Et, etc. hν R1X

Ph2 N P M P Ph2 X

M = Mo, W

O

Ph2 N P M P Ph2 X

Ph2 P P Ph2 +

R1

R2

N

2

HN

N Ph2 P

Ph2 N P M P Ph2 X

R2X

P Ph2

Ph2 P P Ph2

LiAlH4 85 °C

Me2NH 0.95 equiv / W

B N Ph2 N P W P Ph2 OTf

[9-BBN]OTf M=W

(a)

Ph2 P P Ph2

R NH2 PhMe2P

N M

N

O PMe2Ph

R

R′

X

PhMe2P X H+

PhMe2P

PhMe2P

(b) M = Mo, W

M

M X

R = Me, Ph, etc. R′ = H, Me, Ph, etc.

N PMe2Ph Me3SiI

N N PMe2Ph

PhMe2P

PMe2Ph

LiAlH4

NH3 + iPrNH2 rt 60–55 93–95 equiv / W

X

PhMe2P

N N PhMe2P

N

R′

N M

HBr N2H4 + Me2C=NN=CMe2 R = R′ = Me 25–30% 57–64% M=W

SiMe3 PMe2Ph

NaH (Me3Si)2NH + NH3 + N2

PMe2Ph

PhMe2P I

M=W

0.06

0.65

0.52

equiv / W

Figure 1.25 Functionalization of dinitrogen: (a) acylation, alkylation, and borylation and (b) condensation with ketones/aldehydes and silylation.

bridged dihafnium complexes [430, 465–467, 484]. For example, insertion of carbon dioxide into the dinitrogen-bridged dihafnium complex occurs to afford the hydrazido-1,1-dicarboxylato-bridged dihafnium complex, which liberates dicarboxyl silylsubstituted hydrazine by the reaction with trimethylsilyl chloride (Figure 1.26b) [513, 514]. On the other hand, insertion of carbon monoxide gives the nitrido-bridged isocyanato complex, where insertion of isocyanates, nitriles, or alkynes occurs to afford the corresponding ligand-coordinated complexes (Figure 1.26b) [430, 464, 515, 516]. Formation of nitriles from nitrido complexes has been investigated, where nitrido ligands are not derived from dinitrogen [517, 518]. On the other hand,

1.5 Chemical Activation and Reactivity of Dinitrogen Using Transition Metal Complexes

[TiCl3(thf)3]

(a)

N2 (1 atm) Mg THF, rt

CO (1 atm)

MeI

[TiNMg2Cl2(thf)]

Me2NCOMe 73%

[Ti(NCO)Mg2Cl2(thf)]

R' N R'

R' 2 CO2

Hf

Hf

R'

N

Hf

N

O

N

R' = Me

O

CO (1 equiv) R' = H

O C N Hf N Hf

Me3SiCl

Hf

O

O C N Hf O Hf

N Hf

N

Hf

O

120 °C

H N N

N

O N Hf

Hf

C

O

TfO t

BuN=C=NMe

CyN

CyCN

SiMe3

O C N Hf O Hf

MeOTf

C

Bu

O C N Hf N Hf

N N

N

N t

CyCN

O

SiMe3SiMe3

O BuN=C=O

Me3Si

O

C

t

Hf II +

2

H N

Me3SiCl Hf

N

Cy N C

O N Hf Cl Cy

SiMe3

Cy

CN

O C Hf

N N

Hf

Hf 65 °C

N

O

Hf

Me3SiCl 1/2

N C N Hf Hf N C N

+

Hf

OSiMe3 Cl

C (b)

N

Figure 1.26 Functionalization of dinitrogen via the reaction with carbon monoxide or carbon dioxide by using (a) titanium complex or (b) hafnium complexes.

formation of nitriles from the nitrido complex derived from molecular dinitrogen has been investigated by Cummins and coworkers for niobium and molybdenum–nitrido complexes [424, 519], and more recently by Schneider and coworkers for a rhenium–nitrido complex [520]. 1.5.5 Electrochemical and Photochemical Conversion of Dinitrogen Using Transition Metal Complexes The electrochemical synthesis of ammonia- or nitrogen-containing compounds from dinitrogen under ambient reaction conditions has been investigated since the late 1960s [102, 521–523], around the same years that the transition metal–dinitrogen complexes were first isolated. Theoretically, cathodic reduction of dinitrogen in aqueous solution is dependent on pH, where ammonium cation (NH+4 ) in acidic conditions or ammonia (NH3 ) in basic conditions is obtained (Figure 1.27a). In the Haber–Bosch process, this reaction is coupled with the oxidation of dihydrogen (Figure 1.27b), whereas the more attractive method to obtain protons and electrons to reduce dinitrogen is the oxidation of water (Figure 1.27c). A schematic of the representative cell for ammonia synthesis from dinitrogen is shown in Figure 1.27d, where water or dihydrogen can be used as the proton sources for ammonia [306, 524–529].

31

32

1 Overviews of the Preparation and Reactivity of Transition Metal–Dinitrogen Complexes Acidic conditions 2 NH4+ (aq) N2 (g) + 8 H+ + 6 e– Basic conditions + – 2 NH3 (aq) (a) N2 (g) + 6 H + 6 e

(b)

Acidic conditions H2 2 H+ + 2 e– Basic conditions H2 + 2 OH– 2 H2O + 2 e– Acidic conditions 2 H2O O2(g) + 4 H+ + 4 e–

(c)

e–

e– H 2 , H2 O

N3– or OH–

N2

H+ NH3, O2 (d)

Electrolyte

NH3

Basic conditions 4 OH– 2 H2O + O2 (g) + 4 e–

Figure 1.27 Basic half reactions for (a) the reduction of dinitrogen into ammonia, (b) oxidation of water into dioxygen, and (c) oxidation of dihydrogen. (d) Schematic ammonia synthesis cell from dinitrogen and dihydrogen or water under ambient reaction conditions.

The calculated standard apparent reduction potential of dinitrogen into ammonia in water (V vs. SHE) plotted against pH is shown in Figure 1.28 together with the Pourbaix diagram for water, including equilibrium regions for water, dioxygen, and dihydrogen under the standard conditions (1 bar, 25 ∘ C, 0 ionic strength) [73]. As shown in Figure 1.28, reduction of dinitrogen to form ammonia is thermodynamically favored over the production of dihydrogen, whereas usage of water as a proton source for the production of ammonia requires power supply from the outside. However, both oxidation and reduction of dinitrogen are kinetically hindered, and most of the electrode reactions of nitrogen compounds are practically irreversible [530]. Indeed, formation of hydrazine (N2 H4 ) or hydrazidium cation (N2 H+5 ) is thermodynamically disfavored over the production of dihydrogen [531], whereas formation of diazene (N2 H2 ) is unlikely to be observed under electrocatalytic conditions because of its extremely unfavorable reduction thermodynamics and rapid disproportionation to more easily reduced hydrazine (Figure 1.28) [22, 532, 533]. Thus, the electrochemical synthesis of ammonia by the reduction of dinitrogen under ambient reaction ambient conditions requires appropriate catalysts, such as transition metal complexes [102, 521–523, 534–553], transition metal–dinitrogen complexes and their derivatives [554–560], solid-state heterogeneous catalysts [561–571], or nonmetal catalysts [572, 573], whereas electrical efficiency or yield of ammonia have not yet been high enough. It must be noted that the reverse reaction of the electrosynthesis of ammonia from water and dinitrogen compromises the basis of the ammonia fuel cells; i.e. ammonia and dioxygen are supplied into anode and cathode, respectively, to afford dinitrogen and water as exhausted reactants (4 NH3 + 3 O2 → 2 N2 + 6 H2 O) [45–49, 574–577]. Theoretically, the ammonia fuel cell has an electrical potential of 1.172 V at 25 ∘ C and 1 bar, corresponding to the thermal efficiency of 88.6% (HHV, higher heating value) [22], whereas the most efficient ammonia-fed solid oxide fuel cells are operated at much higher temperatures [576, 577]. Indeed, the idea of using ammonia as a fuel for engines goes back to

1.5 Chemical Activation and Reactivity of Dinitrogen Using Transition Metal Complexes 1.5

1.0 O2 (g) + 4 H+ + 4 e– = 2 H2O E′° = +1.229 – 0.0592 pH N2 (g) + 8 H+ + 6 e– = 2 NH4+ (aq)

0.5

E′° = +0.274 – 0.0789 pH 2 H+ + 2 e– = H2 (g) E′° = –0.0592 pH

E′° (V vs SHE)

0.0 N2 (g) + 6 H+ + 6 e– = 2 NH3 (aq) E′° = +0.092 – 0.0592 pH

–0.5

pKa = +9.25

N2 (g) + 5 H+ + 4 e– = 2 N2H5+ (aq) E ′° = –0.214 – 0.0739 pH –1.0

pKa = +7.99 N2 (g) + 4 H+ + 4 e– = 2 N2H4 (aq) E′° = –0.332 – 0.0592 pH

–1.5 N2 (g) + 2 H+ + 2 e– = 2 trans-N2H2 (aq) E′° = –1.198 – 0.0592 pH –2.0

–2.5 0

7

14

pH

Figure 1.28 Pourbaix diagram for water with theoretical apparent standard reduction potentials for dinitrogen reduction to form ammonia, hydrazine, and trans-diazene under standard conditions (1 bar, 298.15 K, 0 ionic strength).

the nineteenth century [578], and several preproduction ammonia-fed motors have been constructed. For example, Norway’s Norsk Hydro has constructed a small truck with an ammonia reformer that extracted dihydrogen from ammonia to burn in its internal combustion engines in 1933 [579]. Later, ammonia was commercially utilized as a fuel for motor buses in Belgium during WWII [580]. Ammonia is now gaining attention as a possible candidate for hydrogen storage and as a direct fuel that does not exhaust carbon dioxide [45–49, 575]. In 1977, Schrauzer and Guth have reported the photoinduced reaction of dinitrogen equilibrated with water vapor to afford a mixture of ammonia, hydrazine, and dioxygen by using transition metal-doped titanium dioxide powder as a photosensitizer as well as a catalyst for the dinitrogen reduction under the irradiation of UV or sunlight (Figure 1.29) [581]. Typical titanium dioxide materials are known to absorb UV light to give a bandgap of 3.0–3.2 eV, where the valence band at around +3 V vs. NHE (normal hydrogen electrode) is positive enough to oxidize water, whereas the conductance band edge at −0.2 to +0.1 V vs. NHE is

33

34

1 Overviews of the Preparation and Reactivity of Transition Metal–Dinitrogen Complexes

360-W Hg-arc lamp 0.20 wt%Fe2O3-doped TiO2 (0.2 g) NH3 + N2H4 + O2 N2 + H2O (g) 40 °C, 4 hours 6.6 μmol 0.14 μmol (1 atm)

Figure 1.29 Photoinduced reduction of dinitrogen and oxidation of water to afford ammonia, hydrazine, and dioxygen.

comparable to or slightly negative than the theoretical reduction potential of dinitrogen into ammonia (Figure 1.28) [22, 582, 583]. Since then, several works have been reported for the photochemical reduction of dinitrogen into ammonia or hydrazine using transition metal compounds [534, 584–598] or nonmetal compounds as catalysts [599, 600], where decomposition of ammonia to give dinitrogen is sometimes rather preferred, and yields of dinitrogen and hydrazine are still low [601, 602]. Photoinduced catalytic conversion of dinitrogen into ammonia can be achieved in combination of catalytic water oxidation and catalytic dinitrogen fixation [603]. As already shown in Figure 1.22b,c, direct photolytic splitting of dinitrogen to nitrides [255, 256, 409, 427, 428, 431, 435] has been clarified, which may be the key route to realize photosynthesis of ammonia from dinitrogen [449].

1.6 Catalytic Conversion of Dinitrogen into Ammonia Using Transition Metal Complexes 1.6.1 Catalytic Formation of Ammonia or Hydrazine Using Molybdenum Complexes Early examples of catalytic conversion of dinitrogen or hydrazine in solution by using transition metal complexes were reported by groups of Vol’pin [314] and Shilov [318], although the reactions required higher temperatures or higher pressures. Catalytic conversion of dinitrogen into hydrazine under ambient reaction conditions (atmospheric pressure and room temperature) was first reported by Shilov and coworkers, who obtained 20.5 equiv of hydrazine and 2.0 equiv of ammonia based on the molybdenum atom of the polynuclear mixed valence Mo(V)–Mo(VI) molybdenum–magnesium complex as a catalyst, when the reaction was carried out in the presence of sodium amalgam, l-α-dipalmitoylphosphatidylcholine (PC), Et2 PhP, and the catalyst at room temperature under 1 atm of dinitrogen (Figure 1.30a) [474]. The structure of the catalyst was later identified as the anionic octanuclear molybdenum complex [Mg(MeOH)6 ][Mg2 Mo8 O22 (OMe)6 (MeOH)4 ]⋅6MeOH [604, 605], where up to 1600 equiv of hydrazine based on the molybdenum atom was reported to be obtained when the reaction was carried out in the presence of sodium amalgam, PC, n Bu3 P, and the catalyst at room temperature under 1 atm of dinitrogen (Figure 1.30b) [475]. Ammonia was also reported to be formed together with hydrazine in the case of Figure 1.30b, although the precise amount ammonia was not reported in later papers [475, 606–610]. Catalytic reduction of dinitrogen into ammonia under ambient reaction conditions by using transition metal–dinitrogen complex as a catalyst was first

1.6 Catalytic Conversion of Dinitrogen into Ammonia Using Transition Metal Complexes

N2 + 4 Na–Hg + 4 MeOH (a) (1 atm)

cat. ([Mo] = 2.0 × 10–5 mol L−1) PC (4 × 10–4 mol L−1) Et2PhP (6.8 × 10–3 mol L−1) MeOH/H2O

20.5

20 °C, 1 hours

N2 + 4 Na–Hg + 4 MeOH (1 atm) (1 mL)

cat. (6.2 × 10–7 mol L−1) PC (4 × 10–4 mol L−1) nBu P (3 × 10–4 mol L−1) 3

N2H4 + 4 NaOMe

MeOH/H2O (12 ml)

(b)

H OMe MeO H O O O Mg O O O Mo O Mo Mo O MeO O O O MeO O O Mo Mo O OMe O O O Mo O Mg O O

2.0

equiv / Mo

1600

18 °C, 140 min

cat.:

N2H4 + NH3 + 4 NaOMe

equiv / Mo 2–

Me O

O

Mo OMe

OMe

Mo O

HO Me Me OH Molecular structure of the anionic part of the catalyst.

Figure 1.30 Catalytic formation of ammonia or hydrazine under ambient reaction conditions using anionic octanuclear molybdenum complex as a catalyst.

reported in 2003 by Yandulov and Schrock, who used decamethylchromocene (CrCp2 *) as a reducing agent and 2,6-lutidinium tetrakis[3,5-bis(trifluoromethyl)phenyl]borate ([LutH]BArF4 ) as a proton source to afford 7.56 equiv of ammonia based on the molybdenum dinitrogen complex bearing a tetradentate hexaisopropylterphenyl (HIPT)-substituted triamidoamine ligand, or in 63% yield based on CrCp2 * (Figure 1.31) [471]. Reactive intermediates such as diazenido, nitrido, ammonia, and hydrido complexes have also been isolated, where almost the same catalytic activities are reproduced (c. 8 equiv of ammonia based on the molybdenum atom of the catalyst, corresponding to 63–66% yield based on CrCp*2 ) (Figure 1.31) [471, 611–614], whereas analogous dinitrogen, imido, and nitrido complexes of vanadium, chromium, and tungsten have only given stoichiometric amounts of ammonia [370, 615, 616]. On the other hand, similar molybdenum complexes bearing tetradentate triamidoamine ligands with different substituents on the amido nitrogen atoms have shown less catalytic activities [471, 613, 617–619], whereas some nitrido complexes have shown catalytic activities comparable to those of the molybdenum complexes bearing HIPT-substituted triamidoamine ligands (Figure 1.31) [613]. Both experimental and theoretical studies have confirmed the reaction pathway of the catalytic transformation of dinitrogen toward ammonia, which is called

35

36

1 Overviews of the Preparation and Reactivity of Transition Metal–Dinitrogen Complexes

BArF4 N2 + 6 (1 atm)

cat.:

HIPT

+6

Cr

Me

rt, 7 hours

Me

N H

N N

NH HIPT

HIPT

i HIPT Pr

N

HIPT N Mo N N

N

i

Diazenido complex

7.56 equiv 63% yield

7.73 equiv 63% yield

N

HIPT

HIPT N

NH3

HIPT N Mo N N

HIPT

Ammonia complex

7.97 equiv 66% yield

8.06 equiv 64% yield

N N pBrHIP Mo N N

pBrHIP

iPr

Pr

H N HIPT Mo N N

HIPT N

Hydrido complex 7.65 equiv 66% yield

HMT i

i

Pr iPr Br

Pr

i

Pr

HIPT

N

DPP N Mo N N

iPr

BArF4 HIPT

iPr

N

6.4–7.0 equiv

DPP

iPr

HIPT

N

Nitrido complex

pBrHIP

iPr

N

Dinitrogen complex

HIPT N Mo N N

2 NH3

c. 8 equiv / Mo

(36 equiv) (48 equiv) Reducing agent Proton source

HIPT N Mo N N

HIPT

cat.

DPP N

N iPr HIPT N Mo N N N

i

Pr

iPr

iPr

pBrHIP Pr iPr Br N N i

2.0 equiv iPr

2.53 equiv Nitrido complexes

iPr

DPP

Figure 1.31 Catalytic nitrogen fixation using Schrock’s catalysts.

“Schrock cycle,” where the addition of protons and electrons occurs stepwise in the “distal” way to release 2 M amount of ammonia molecules with the retention of higher oxidation states of the molybdenum center (Figure 1.32) [471, 611, 612, 614, 620–622]. The second example of the catalytic formation of ammonia by using transition metal–dinitrogen complexes has been reported in 2010 by Nishibayashi and coworkers, who obtained 12 equiv of ammonia based on the Mo atom of the dinitrogen-bridged dimolybdenum–dinitrogen complex bearing a tridentate PNP-type pincer ligand [{Mo(N2 )2 (t BuPNP)}2 (μ-N2 )] (t BuPNP = 2,6bis[(di-tert-butylphosphino)methyl]pyridine) on treatment with cobaltocene (CoCp2 ) as a reducing agent and 2,6-lutidinium trifluoromethanesulfonate

1.6 Catalytic Conversion of Dinitrogen into Ammonia Using Transition Metal Complexes N2 H

H

+

H

N

H H

H H N+

e–

N Mo N N N

NH3 N N

H H N

N Mo N N N

N Mo N N N

Mo(III)

Mo(III)

N N Mo N N

Mo(IV)

Mo(IV)

N N

N

N N Mo N N

Diazenido complex

Amido complexes

H

H+ N H

Dinitrogen complex

Ammonia complexes

e–

N

Mo(IV)

N Mo N N N

H

N–

e–

Mo(IV) H+ H

Mo+ N

H

Hydrazido complexes

H

Imido complexes

N

Nitrido complex

H e–

N N

H

Hydrazidium complex

Mo+ N

H

Mo(VI)

H+

N N Mo N N Mo(VI)

H H N+

N N Mo N N

N

NH3

e–

N

H

e–

N N Mo N N N

N

N

H

Mo(VI)

N Mo N N N Mo(V)

N

N N Mo+ N N N

N Mo(V) +

N

H+

Mo(V)

N

Mo(V)

Figure 1.32 Proposed Schrock cycle for the catalytic formation of ammonia. Compounds shown on light gray backgrounds are isolated or spectroscopically observed intermediates.

([LutH]OTf) as a proton source in the presence of the catalyst at room temperature for 20 hours under the atmospheric pressure of dinitrogen, where the yield of ammonia was 49% based on the amount of CoCp2 (Figure 1.33a) [472]. Introduction of electron-donating groups at the 4-position of the pyridine ring of the PNP-type pincer ligands has increased the catalytic activity where up to 26 equiv of ammonia has been obtained based on the molybdenum atom when methoxy group is introduced to the 4-position of the pyridine ring [623], whereas introduction of ferrocenyl or ruthenocenyl moieties has also been effective (Figure 1.33a) [624]. Changing substituents on the phosphorous atom, for example, introduction of adamantly (Ad) group instead of tert-butyl group to the phosphorous atom, has been less effective [625], whereas the corresponding complex of tungsten [626], dinitrogen-bridged tetrachloride complex [627], or

37

38

1 Overviews of the Preparation and Reactivity of Transition Metal–Dinitrogen Complexes OTf

N2

+

(1 atm) cat.:

R

Fc =

cat. 2 NH3 Me Toluene, rt, 20 hours N H equiv / Mo (108 equiv / Mo) (144 equiv /Mo) Co +

6

6

Me

N t N Bu2 P

t

Bu2 N P N N Mo N N Mo N N P N t N tP Bu2 Bu2 N EtFc =

Fe

18 (R = Fc) 12 (R = H) 11 (R = Ph) 15 (R = EtFc) 12 (R = SiMe3) 5 (R = PhFc) 14 (R = tBu) 13 (R = Rc) 16 (R = Me) 17 (R = OMe)

R

Fe up to 26 equiv / Mo (R = OMe) up to 23 equiv / Mo (R = Fc)

Rc =

PhFc =

Ru

Fe

(a) N t N Bu2 P

t

Bu2 N P N N Mo N N Mo N N P N t N tP Bu2 Bu2 N Dinitrogen complex t Bu2 N P

NH3, OTf –

e– N2

N N N Mo N N NH3 N tP Bu2 P tBu N

N

P

tBu OTf 2

(b)

Ammonia complex

NH

N

P

tBu 2

P tBu2

OTf

Nitrido complex

2 H+ e–

t

Bu2 N P N N Mo N N Mo N N P tBu OTf N tP 2 Bu2 Hydrazidium complex N P tBu 2

N Mo 3 H+ 3 e– N2

t

Bu2 N P N t P Bu2 N N Mo N N Mo N N P P tBu OTf N tBu 2 2 Diazenido complex

N2

NH3+

2

N Mo

H+ OTf –

NH3

e–

Figure 1.33 (a) Catalytic formation of ammonia using dinitrogen-bridged dimolybdenum– dinitrogen complexes as catalysts and (b) its proposed catalytic reaction pathway.

the corresponding molybdenum complex with the analogous pincer-type ligand containing arsenic [628] has not worked as catalysts. The hydrazido complex obtained by stoichiometric protonation of the dinitrogen complex also has not worked as a catalyst for the reduction of dinitrogen to afford ammonia [626], whereas several nitrido complexes have worked as catalysts [629, 630]. Based on the experimental results and DFT calculations, a catalytic cycle during which the structure of dinitrogen-bridged core is maintained for the protonation of terminally coordinated dinitrogen and reduction of the molybdenum center to afford 1 M amount of ammonia and the corresponding nitrido complex, which is further protonated and reduced to afford ammonia and the starting dinitrogen-bridged dimolybdenum–dinitrogen complex via ligand exchange of ammonia with dinitrogen (Figure 1.33b) [629, 631, 632].

1.6 Catalytic Conversion of Dinitrogen into Ammonia Using Transition Metal Complexes

Me N2 + 6

Co

Ph P

Mo

t N Bu2 P

Ph P

Cl

Mo

tBu

2

Me

Toluene, rt, 20 hours

6.4 equiv 53% yield

F t N Bu2 BAr 4 P

Ph P t

2

11.0 equiv 92% yield

(a)

2 NH3

up to 11 equiv / Mo

Cl

P

P Cl

tBu

N H

(48 equiv)

t

Cl Bu2 P

cat.

+6 Me

(1 atm) (36 equiv)

OTf

Mo

P Bu2

Cl

9.6 equiv 80% yield up to 63 equiv OTf

Cr

N2 + 6

+6 Me

(1 atm) (720 equiv) N t N Bu2 P

(b)

t

Bu2 N P N N N Mo N N Mo N N NP P N t Bu2N N tBu2 100 equiv 42% yield up to 73% yield

cat. N H

Me

(960 equiv)

Toluene, rt, 20 hours

2 NH3

up to 115 equiv / Mo N tBu 2 N P

t

Bu2 N P N N N Mo N N Mo N N P N NP tBu t 2N N Bu2 115 equiv 48% yield

Figure 1.34 Catalytic formation of ammonia using molybdenum complexes with (a) PPP-type or (b) PCP-type pincer ligands.

Ligation of PNNNP-, PNN-, or PNP-type pincer ligands to the molybdenum moieties has failed to reduce dinitrogen catalytically [630, 633], whereas the molybdenum nitrido complexes with PPP-type pincer ligands (Figure 1.34a) [634] or the dinitrogen-bridged dimolybdenum–dinitrogen complexes with N-heterocyclic carbene-based PCP-type pincer ligands (Figure 1.34b) [635] have been successful, where up to 63 or 115 equiv of ammonia is obtained based on the molybdenum atom of the catalyst, respectively. Furthermore, the molybdenum iodide complexes bearing a PNP-type pincer ligand have been found to have higher catalytic activities than the corresponding dinitrogen complexes, where up to 415 equiv of ammonia is produced based on the molybdenum atom of the catalyst (Figure 1.35a) [636]. Experimental details and DFT calculations have suggested that the reaction pathway includes the formation of a dinitrogen-bridged dimolybdenum iodide complex, where the splitting of the bridging dinitrogen ligand occurs to afford the nitrido complex (Figure 1.35b) [636]. This is the first example of the catalytic conversion of

39

40

1 Overviews of the Preparation and Reactivity of Transition Metal–Dinitrogen Complexes Me N2 + 6

cat.

+6 Me

Co

(1 atm) (36 equiv)

OTf

Me

N H

up to 12.2 equiv / Mo N t P Bu2

(48 equiv) I

cat.:

P tBu2

N Mo I

N Mo tBu 2

10.9 equiv 91% yield up to 415 equiv

12.2 equiv 96% yield t

I

Pt

1/2 N2 I

Bu2

N Mo I

1/2

N2, 2 e–

N N

P tBu2

Mo N

N

N

I

I

P

NH3 P tBu2

NH3

P tBu2

N Mo

N Mo P t Bu2

I

tBu 2

N Mo N2

(b)

N

Bu2 P

N HP t tBu t Bu2 2 P Mo P Bu2

P I

tBu 2

I

P

P I

tBu 2

(a)

2 NH3

Toluene, rt, 20 hours

P

tBu

I

3 H+, 3 e–

2

Figure 1.35 (a) Catalytic formation of ammonia using dinitrogen-bridged dimolybdenum–dinitrogen complexes as catalysts and (b) its proposed catalytic reaction pathway.

molecular dinitrogen into ammonia via the direct cleavage of dinitrogen on a molecular catalyst as the key reaction step under ambient reaction conditions. The molybdenum nitrido complex bearing pyridine- and diimido-based NNN-type pincer ligand has also been reported by Schrock and coworkers to work as a catalyst, where up to 10.3 equiv of ammonia is obtained based on the catalyst [637]. 1.6.2 Catalytic Formation of Ammonia or Hydrazine Using Transition Metal Other than Molybdenum (Iron, Ruthenium, Osmium, Cobalt, and Vanadium) Complexes In 2013, Peters and coworkers have reported the catalytic system producing ammonia by using iron–dinitrogen complexes as catalysts, where KC8 and the Brookhart’s acid ([H(OEt2 )2 ]BArF4 ) are employed as the reducing reagent and the proton source, respectively, to afford 7.0 equiv of ammonia based on the iron atom of the anionic iron–dinitrogen complex bearing a tris(phosphine)borane ligand used as the catalyst (Figure 1.36a) [473]. In order to prevent the formation of dihydrogen gas by the direct reaction of KC8 with Brookhart’s acid, the reaction must be carried out at sufficiently low temperatures. The yield of

1.6 Catalytic Conversion of Dinitrogen into Ammonia Using Transition Metal Complexes cat.

N2 + 6 KC8 + 6 [H(Et2O)2]BArF4

2 NH3 equiv / M Et up to 7.0 equiv / Fe up to 2.4 equiv / Co i Et Pr N [cation]+ N N iPr iPr N N 2 Fe N N M P Fe Fe Et P iPr2 S i Et Pr Si Si N

–78 or –95 °C

(1 atm) (45–50 equiv) (37–50 equiv)

iPr 2

cat.:

P iPr 2 P

[cation]+

N N

iPr

Fe

PiPr2

iPr

2

P 2 P

X

iPr M = K(18-crown-6) 2.6 equiv

X = B, Si: [cation] = Na(12-crown-4)2 [cation] = Na(12-crown-4)2 X = C: [cation] = K(Et2O)0.5 1.8 equiv 7.0 equiv / up to 59 equiv / up to 88.1 equiv (Hg light) (X=B) 4.6 equiv / up to 36 equiv (X = C) 0.8 equiv / up to 3.8 equiv (X = Si) BArF4 iPr iPr Et Et N N Et Et H i i Pr Pr H Fe Fe iPr P PiPr2 Fe 2 i iPr Pr P N Et Et N N N Et Et Ph i i Pr Pr up to 8.9 equiv 3.4 equiv 3.3 equiv (Hg light) [cation]+ N N iPr 2 iPr 2

(a)

P P

N Co

iPr

PiPr2

iPr

P 2 P

Co

[cation] = Na(12-crown-4)2 2.4 equiv

0.8 equiv

+6 [H2NPh2]OTf

PiPr2

iPr

12.8 equiv up to 84 equiv

2

H

Fe N N

iPr P 2

Fe P

H

PiPr2

Ph up to 66.7 equiv (Hg light)

P P

N N M

[K(thf)2]

PiPr2

Si

4.3 equiv (M = Ru) 1.6 equiv (M = Os) cat. −78 °C

N (1 atm) (46–54 equiv) (50–108 equiv) F ] N i [BAr iPr 4 Pr2 2 P i P Fe M P Pr2 iPr P iPr P cat.: 2 2 Si B

(b)

P Ph

iPr 2

2

B

Co

P PiPr2

N

B

N2 + 6

iPr 2

2 NH3 equiv / M [K(thf)2] PiPr2

1.4 equiv (M = Fe) 0.8 equiv (M = Ru) 7.1 equiv up to 120 equiv (M = Os)

Figure 1.36 Catalytic formation of ammonia using iron, ruthenium, osmium, and cobalt complexes by using (a) Brookhart’s acid or (b) diphenylammonium trifluoromethanesulfonate as the proton source.

41

42

1 Overviews of the Preparation and Reactivity of Transition Metal–Dinitrogen Complexes

ammonia has been found to increase up to 59 equiv based on the iron atom with an estimate of the initial turnover frequency of 1.2 ± 0.1 min−1 by increasing the amount of KC8 and the acid [638], or up to 88 equiv by further irradiation of UV light [639], whereas formation of catalytic amounts of ammonia has been found for other iron complexes [638–642], cobalt–dinitrogen complexes [643], or ruthenium– and osmium–dinitrogen complexes [644] as shown in Figure 1.36a. Spectroscopic observation and isolation of reaction intermediates or their derivatives lead to the conclusion that the conversion by using these iron complexes proceeds rather via the Chatt-type distal mechanism [645–649]. Switching the proton source from Brookhart’s acid to diphenylammonium trifluoromethanesulfonate has been effective especially for the osmium complex, where up to 120 equiv of ammonia has been obtained [644, 650]. Catalytic formation of hydrazine in addition to ammonia has been recently reported by Nishibayashi and coworkers, up to 22.7 equiv of ammonia and 1.7 equiv of hydrazine, corresponding to 26.1 equiv of nitrogen atoms based on the iron atom of the iron–dinitrogen complex bearing the anionic pyrrole-based, PNP-type pincer ligand [651, 652] (Figure 1.37a). The cobalt–dinitrogen complex bearing the anionic pyrrole-based, PNP-type pincer ligand has also been found to be effective in fixing up to 17.9 equiv of nitrogen (15.9 equiv of ammonia and 1.0 equiv of hydrazine) based on the cobalt atom of the catalyst (Figure 1.37a) [653]. Several iron and cobalt complexes bearing anionic pyrrole-based, PNP-type pincer ligands or an anionic carbazole-based, PNP-type pincer ligand are active toward the catalytic fixation of dinitrogen into ammonia and hydrazine (Figure 1.37a) [651–654], whereas iron complexes bearing azaferrocene-based, PNP-type pincer ligands [655] or iron– and cobalt–dinitrogen complexes bearing PSiP-type pincer ligands [656] are not active toward the catalytic reduction of dinitrogen into ammonia or hydrazine. A catalytic cycle is proposed where alternating protonation at the distal and proximal positions of dinitrogen occurs to afford an intermediary hydrazine complex, where both hydrazine and ammonia can be produced (Figure 1.37b) [651]. More selective production of hydrazine has been recently achieved by Ashley and coworkers who have fixed up to 50 equiv of nitrogen atom in the forms of 0.95 equiv of ammonia and 24.5 equiv of hydrazine based on the iron atom, when CoCp∗2 , diphenylammonium trifluoromethanesulfonate, and the iron–dinitrogen complex [Fe(N2 )(depe)2 ] are employed as a reducing agent, a proton source, and a catalyst, respectively (Figure 1.38) [657]. Related iron complexes also give stoichiometric amounts of ammonia and hydrazine on treatment with acids, but not all of them work as catalysts [657, 658]. In addition to molybdenum– and late transition metal–dinitrogen complexes, earlier transition metal–dinitrogen complexes (vanadium and titanium) have been found to work as effective catalysts for the reduction of dinitrogen into ammonia [659, 660]. Indeed, Nishibayashi and coworkers have reported that up to 16 equiv of fixed nitrogen (12 equiv of ammonia and 1.8 equiv of hydrazine) has been obtained based on the vanadium atom of the catalyst, when a mononuclear vanadium complex bearing an anionic pyrrole-based, PNP-type pincer ligand and an aryloxy ligand in equilibrium with the formation of dinitrogen-bridged divanadium complex under atmospheric dinitrogen as a catalyst (Figure 1.39a) [659].

1.6 Catalytic Conversion of Dinitrogen into Ammonia Using Transition Metal Complexes

cat.

N2 + KC8 + [H(Et2O)2]BArF4 (1 atm) PR12

R2

cat.:

L

N M R2

PR12

Et2O, –78 °C, 1 hour

M = Fe, R1 = tBu, R2 = H, L = N2 M = Fe, R1 = tBu, R2 = Me, L = N2 M = Fe, R1 = tBu, R2 = Ph, L = N2 M = Fe, R1 = tBu, R2 = H, L = H M = Fe, R1 = tBu, R2 = Me, L = H M = Fe, R1 = tBu, R2 = H, L = Me M = Co, R1 = tBu, R2 = H, L = N2 M = Co, R1 = Cy, R2 = H, L = N2 M = Co, R1 = tBu, R2 = H, L = H M = Co, R1 = tBu, R2 = H, L = Me

NH3 + N2H4 equiv / M 14.3 1.8 22.7 1.7 4.7 1.2 3.0 0.1 18 2.0 3.7 95% yield) or even with ethanol. This N2 ligand is sufficiently basic to accept transfer of NH protons, as demonstrated in the reactions with Me2 NH or Me2 NNH2 to produce a bridging N2 H2 ligand. The strongly activated end-on dinitrogen complex [(η5 -C9 H5 -1-iPr-3-Me)2 Zr(NaCl)]2 (μ-η1 :η1 -N2 ) (see Scheme 2.21) can liberate N2 H4 in 96% yield through protonation with excess H2 O (45% yield with HCl) [87]. Production of N2 H4 from other indenyl complexes is described in Section 2.2.3 (Scheme 2.22) [88]. Addition of equimolar H2 O to the side-on dinitrogen complex supported by a macrocyclic ligand {[P2 N2 ]Zr}2 (μ-η2 :η2 -N2 ) ([P2 N2 ] = (PhP(CH2 SiMe2 NSiMe2 CH2 )2 PPh)) results in the formation of {[P2 N2 ]Zr}2 (μ-η2 :η2 -N2 H2 )(μ-O), in which the N2 unit remains as a side-on HNNH ligand (Scheme 2.33) [101]. The N2 ligand in this complex is also protonated with N2 H4 to generate {[P2 N2 ]Zr}2 (μ-η2 :η2 -N2 H2 )2 [25].

113

114

2 Group 4 Transition Metal–Dinitrogen Complexes

OZ

Excess ZOH

N Zr

2

Zr

N

Zr

(Z = H, Et)

OZ

+ N2H4

2 HNR2 Excess EtOH

Trace H2O

– 2 HNR2 (Z = Et)

H Zr

Zr

Zr

O

N NR2 H

H

HR N 2 N Zr

+ N2H4 NR2 = NMe2, NHNMe2

Scheme 2.32 Protonation of a side-on dinitrogen complex of zirconocene derivative.

N N

P Ph

Ph

Zr

H Ph P

N

H2 O

N N

Ph N

Ph P

THF 50% yield Ph

Interatomic distances (Å): N–N, 1.434(9); Zr–Zr: 3.1922(10); Zr–N, 2.180(7)–2.238(6) 1H NMR: δ = 3.93 (NH)

P Ph

Zr

(1 equiv)

N O H Zr N N

P

P

N2H4 N THF

Zr P

N N

Ph P

P P N N Ph Zr Ph H H N NN N H H Ph Ph Zr N P P N

d(N–N) = 1.438(3) Å

Scheme 2.33 Protonation of the side-on dinitrogen complex of zirconium bearing a macrocyclic P2 N2 ligand.

It has been briefly reported that the treatment of [(η5 -C5 Me5 )2 Ti(N2 )]2 (μ-η1 : η1 -N2 ) with HCl gives nearly quantitative yield of N2 H4 , as observed for the Zr congener. This Ti complex loses terminal N2 ligands at the temperatures higher than −10 ∘ C to afford [(η5 -C5 Me5 )2 Ti]2 (μ-η1 :η1 -N2 ), which on addition of HCl liberates only minor amounts of reduced N2 (5–20%) [52]. Protonation of (Cp2 Ti)2 (N2 ), whose structure has not yet been definitively characterized, is mentioned in Section 2.2.1.

2.3 Reactions of Group 4 Transition Metal–Dinitrogen Complexes

2.3.2

Reduction

Teuben and van der Waij investigated the liberation of nitrogen compounds from the well-defined dinitrogen complexes (Cp2 TiR)2 (μ-η1 :η1 -N2 ) (R = aryl, benzyl). Treatment of these TiIII complexes with gaseous HCl or Br2 merely leads to the evolution of N2 accompanied by quantitative cleavage of the Ti—R bond (Scheme 2.34) [40]. Thermolysis of the solid (Cp2 TiPh)2 (μ-η1 :η1 -N2 ) at 150 ∘ C released benzene and the major part of N2 (∼90%), whereas a small amount of NH3 (0.14 equiv to Ti) was also detected after decomposition with H2 SO4 [40, 120]. Although thermal reactions were examined also in solution under high N2 pressures (up to 40 atm), the yields of NH3 were still low (3 days) evolve toward the trans complex. With phenyl-substituted phosphines, the stepwise reduction provides a better yield (Scheme 4.3). A variety of stable dinitrogen complexes with tridentate [32, 38–43], tetradentate, and pentadentate phosphines have been developed, in particular by the groups of George and Tuczek, following similar strategies (Scheme 4.4) [7, 9, 44–50]. In these works, saturation of the coordination sphere of the metal center was searched in order to favor the coordination of either one- or two-terminal N2 ligands. Thus, R3 P, alkene, or bidentate PP ligands were used in conjunction with mainly the “PPP” ligand PPh PPh PPh [9, 40, 47, 51] but also with the “PEP” ligands (P = PPh2 , E = O, NH, S) [32, 43, 50, 52]. In turn, this would allow for fine adjustments of the reactivity developed with [M(PP)2 (N2 )2 ] and [M(PR3 )4 (N2 )2 ] (vide infra). It is to be noted that the stability of these complexes depended strongly on the “E” ligand as well as on the PR3 , and most significant results in terms of reactivity were subsequently obtained with the “PPP” tridentate ligand. Dahlenburg and Pietsch have obtained similar complexes with more strongly donating tridentate phosphine PMe PMe PMe and PMe3 [53]. Neutral tridentate ligands (PEP; P = PtBu2 ; E = N, P, C) featuring electron-donating phosphines have been used from 2011 by Nishibayashi and coworkers (Scheme 4.5). The Na/Hg reduction of the [MoCl3 (PEP)] complex under N2 , without additional phosphine, was optimized to yield the [Mo(PEP)(N2 )2 ]2 (μ-N2 ) complexes featuring both terminal and bridging N2 ligands [54–57]. The infrared spectrum exhibits a strong 𝜈 NN band between 1910 and 1978 cm−1 (THF) for the terminal N2 ligand of the PEP complex and a strong 𝜈 NN band at 1890 cm−1 in the Raman spectrum for the bridging N2 ligand. Notably, if the PNP and PCP complexes could be obtained efficiently, the PPP (PPP = PtBu PPh PtBu ) complex could not

4.2 Preparation of Group 6 Transition Metal–Dinitrogen Complexes N2

Cl E

PPh2

Mo

P Ph2

3 Na/Hg

E

N2

Cl

P Ph2

PR3

Cl

E = PPh, O, NH

PR3 N2

PR3 = PPh3, PMePh2, PMe2Ph

Cl E

PPh2

N2

3 Na/Hg

PPh2

Mo

PPh2

Mo

E

N2

Cl

PPh2

PR3

Cl

Mo

PPh2 PR3

N2

E = PPh, S N2

N2

Cl P

3 Na

PMe2

Mo

PMe2

N2

Cl

P

PMe2

PMe3

Cl

Mo

PMe2

PMe3

PMe3

–N2

P

PMe2

Mo

PMe2

PMe3 PMe3

N2

mer, cis

cis + trans

Scheme 4.4 Cl P

E M

2

Cl

P

N2 3 Na/Hg N2

Cl PtBu2 PEP = R′

N PtBu2

R′ = H, Me Ph, tBu, MeO, SiMe3

E

Mo P

P

P N

N Mo N2 P

N2

IR: (THF) Ph R′ = H; 1944 cm–1 P R′ = Me; 1939 cm–1 –1 R′ = Ph; 1950 cm PR2 R′ = tBu; 1939 cm–1 R′ = MeO 1932 cm–1 R′ = SiMe3; 1947 cm–1 Raman: R′ = H; 1890 cm–1

PR2

N

PtBu2 (solid) IR: R' = H; 1978 cm–1 R' = Me; 1969 cm–1

C R′

N PtBu2

(solid) IR: 1968 cm–1 Raman: 1946 cm–1

R = tBu, Cy

PtBu2 R′

N2 E

N

(solid) IR: 1910 cm–1

C N PtBu2

R′ = H, Me

Scheme 4.5

be obtained in the pure form, although strong IR bands for terminal N2 ligand were observed [56]. A related PPP (PCy PPh PCy ) complex was isolated by Mézailles following a similar strategy [58]. The dimer [Mo(PCy PPh PCy )(N2 )2 ]2 (μ-N2 ) was shown to be formed upon crystallization (Nujol mull, terminal N2 𝜈 NN at 1968 cm−1 in IR, bridging N2 𝜈 NN at 1946 cm−1 ), whereas mixtures of monomeric paramagnetic [Mo(PCy PPh PCy )(N2 )2 ] and diamagnetic [Mo(PCy PPh PCy )(N2 )3 ] are seen in solution (𝜈 NN at 2070, 2002, and 1969 cm−1 in IR). Nishibayashi and

225

226

4 Group 6 Transition Metal–Dinitrogen Complexes ′

coworkers have shown that the electronic nature of R′ in the PNR P ligands allows a fine modulation of the redox properties of the complex [55]. 4.2.1.2

Thioether Ligands

In light of the importance of S-type ligands in nitrogenases, efforts have been paid to synthesize Mo(0) dinitrogen complexes with thioether ligands. With bidentate PhSCH2 CH2 SPh, the complex cis-[Mo(PhSCH2 CH2 SPh)(PMe2 Ph)2 (N2 )2 ] was synthesized but proved unstable at room temperature, via N2 evolution [1]. Later on, Morris et al. used the mixed phosphine–thioether ligand PPh2 CH2 CH2 SMe to form trans-[Mo(PPh2 CH2 CH2 SMe)(PMePh2 )2 (N2 )2 ], whereas the bis-thioether ligands RSCH2 CH2 SR (R = Me, Et) resulted in immediate displacement of all N2 [59]. Use of the macrocyclic crown thioether ligand Me8 [16]aneS4 resulted in increased stability. Reduction of the corresponding Mo dibromide complex by a large excess of Na/Hg under 5 atm of N2 resulted in the formation of the desired trans-[Mo(Me8 [16]aneS4 )(N2 )2 ] in 36% yield [60], whose chemistry was subsequently studied (vide infra) [61]. The NN bond distances measured in the above-mentioned group 6 metal (0) end-on dinitrogen complexes are virtually identical to those for free N2 . The measure of the NN bond distance is thus not an important parameter to judge the degree of activation of N2 upon coordination in this class of complexes. 4.2.1.3

Nitrogen and Cp Ligands

Schrock and coworkers have devised (NR 3 N)3− triamido-amine ligands to enforce tetradentate coordination at a metal center, as well as to provide a sterically protected threefold symmetry in order to favor the binding and activation of N2 . With the first ligand, (NC6 F5 3 N)3−, they showed that the reduction of the [(NC6 F5 3 N)MoCl] complex by two electrons resulted in the formation of the diazenido complex (NC6 F5 3 N)Mo—N=N—Na(ether)x , whereas the reduction by one electron yielded the homo-bimetallic diazenido [(NC6 F5 3 N)Mo—N=N—Mo(NC6 F5 3 N)] [62]. After careful optimization, they developed the much more bulky derivative [((HIPT)NCH2 CH2 )3 N]3− ligand, where HIPT = 3,5-(2,4,6-i-Pr3 C6 H2 )2 (C6 H3 )) to (i) maximize steric protection of a metal coordination site in a monometallic species, (ii) prevent bimetallic complex formation, and (iii) provide increased solubility of intermediates of N2 functionalization at Mo. With this system also, reduction of the MoCl precursor with two electrons results in the formation of diamagnetic Mo–N2 − (Scheme 4.6) [63, 64]. One electron oxidation of this complex by ZnCl2 produces N Cl

R N R

Mo

R N

N

Mg R

N

Mo

R N

N

IR nNN cm–1

ZnCl2

1855 (NBu4+)

N

N

R N R

N

N

Scheme 4.6

N

N

R

Mo

N N

1990

R N

FeCp2

CoCp*2

N

R N R

Mo

N N

2225

R N

4.2 Preparation of Group 6 Transition Metal–Dinitrogen Complexes

the paramagnetic Mo–N2 complex in good yield. If the Mo15 N2 complex (92 : 8 ratio of 15 N2 to 14 N2 ) is further oxidized by Cp2 Fe+ under 14 N2 , and reduced back with CoCp*2 after 15 minutes, a mixture of the Mo14 N2 and Mo15 N2 complexes is formed in a ratio of 82 : 12, which proved the lability of the N2 ligand in the oxidized Mo(IV) complex [65]. Subsequently, other ligands with modulated electronics and sterics were developed [66]. Chirik and coworkers have used neutral tridentate NNN-type ligands (PDI, pyridine-di-imine) to stabilize Mo dinitrogen complexes (Scheme 4.7). Thus, the bridging [Mo(PDI)(N2 )]2 (μ-N2 ) complex was isolated upon reduction of the corresponding trichloride with excess of Na/Hg under N2 . They have studied the reactivity of this complex toward H2 , PrNH2 , and NH3 , which resulted in the displacement of bridging and/or terminal N2 . With the amines, NH bond insertions and ligand rearrangements were observed as fast and efficient reactions [67]. Ph

Ph

Ph N iPr N iPr

Mo Cl Cl

N Ar

N

Ph

Mo

Cl N

xs Na/Hg

iPr

Ph

N Ar

iPr

N

N Ph

–N2

N

Ph

NH3 or N2H4

N

H N Ph Ar N N H Mo N H Ar Ar

–N2

NH 1 hour

Mo H

Ph

H2

N

N Ar N N Mo N N N Ar

N Ar

N

–N2 Ph Ar

N N

Mo

Ph

Ph N C H 2 N Mo N Ar N N Ar

Ph

Ph

N NH Ph

Ar HN

N Ar

Mo

N N

Ph

Ar

Scheme 4.7

The strongly electron-donating and bulky 1,2,3,4,5-pentamethylcyclopentadienyl ligand Cp* has been used to stabilize unsaturated Mo centers and thus to promote N2 coordination. If the complex [Cp*Mo(Cl)(PMe3 )2 (N2 )] can be synthesized, N2 is readily lost because of the greater stability of the 16-electron triplet complex vs. the singlet 18-electron N2 -coordinated species [68]. The use of bidentate depf ligand was proposed to favor the singlet state and thus the stability of the N2 coordinated species (Scheme 4.8). Accordingly, a ca 2 : 1 mixture of [Cp*Mo(depf )(μ-N2 )]Na(THF)3 and [Cp*Mo(depf )(H)(N2 )] was obtained via the reduction of the [(Cp*MoCl4 )2 (μ-depf )] precursor with a large excess of Na/Hg under N2 . Notably, the analogous W complexes could not be obtained, and instead of N2 coordination, CH bond activation of the Cp* ligand was observed [69]. Hydride abstraction from [Cp*Mo(depf )(H)(N2 )] resulted in the formation of [Cp*Mo(depf )](μ-N2 )+ cationic complex. One-electron

227

228

4 Group 6 Transition Metal–Dinitrogen Complexes MoCp*Cl4

Et2 P

PEt2

PEt2

Fe

N2

20 Na/Hg

P Et2

Et2 P

MoCp*Cl4

Ph2C –H2 –Ph2CH

Mo

H

N

P Et2

P Et2

N N

Et2 P Fe

Mo P Et2

Et2 P Fe P Et2

–2e

+2e

–1e

N

Et2 P

Fe Mo

P Et2

N

+1e

+

Fe

N

Mo

Mo Et2 P Fe

Mo

Fe

Fe

N N

Et2 P Mo P Et2

2

N N

Et2 P Fe

Mo

Na(THF)3

P Et2

Scheme 4.8

oxidation and reduction allowed the isolation of the series of bridging end-on N2 Mo dimers. X-ray structures revealed an almost linear geometry of the Mo–N–N–Mo. 4.2.2 End-on Bridging Dinitrogen Complexes from N2 : Synthesis and N2 Splitting As readily understood following the Dewar–Chatt–Duncanson model, the more backdonation to the N2 moiety from the metal, the more the NN bond is reduced. Accordingly, examination of the N—N bond length has been associated with the degree of N2 activation. However, as mentioned above, in all the terminal end-on N2 complexes of group 6 metals, which is the vast majority of the complexes, such activation is rather minimal/negligible. Greater activation can be achieved if the N2 ligand bridges two (reducing) metal centers. Moreover, in hetero-bimetallic bridging N2 complexes, one metal can render N2 a better Lewis base for the second metal center. Indeed, as early as 1974, the complex [Re(Cl)(PMe2 Ph)4 (N2 )] was used as a “ligand” for Mo(IV) to form the N2 bridging dimer [ReCl(PMe2 Ph)4 (μ-N2 )MoCl4 (MeOH)] [70]. Trimetallic Re–N2 –M–N2 –Re (M = Mo, W) complexes were subsequently isolated with the same Re precursor and [MCl4 (PPh3 )2 ] complexes in which phosphines are weakly bound [71]. Hidai and coworkers prepared heterobimetallic complexes with bridging N2 ligand starting from cis-[W(PMe2 Ph)4 (N2 )2 ] [72]. These complexes result formally from the insertion of the W–N2 fragment into M—X bonds of Cp2 MX2 (M = Ti, Zr, Hf; X = Cl, I) (Scheme 4.9). The W/Zr complex was crystallized and showed the W–N–N–Zr geometry to be close to linear (W–N–N angle = 172∘ ; Zr–N–N angle = 175∘ ), and the NN bond (1.24(2) Å) is consistent with a diazenido (N2 2− ) moiety, consistent with the 𝜈 NN band at 1541 cm−1 [73].

4.2 Preparation of Group 6 Transition Metal–Dinitrogen Complexes

N PhMe2P

N2

W

PMe2Ph

X

N

W

Cl

N

Cp M Cp

–N2

PMe2Ph PMe2Ph

PhMe2P

PhMe2P

Cp2TiCl2 or Cp2MCl2/NaI

N

N2

PhMe2P

M = Zr, Hf

Scheme 4.9

Schrock and coworkers have studied N2 coordination to high-oxidation group 6 metals in the 1990s. In the homo-bimetallic W complexes [W(Cp*)(Me)3 ]2 (μ-N2 ) generated by the reduction of [W(Cp*)(Me)3 (OTf )], they have evidenced the M to N2 electron transfer that results in the reduced N—N bond order. Bond distances of ca 1.27 Å are similar to N=N double bonds (Scheme 4.10) [74]. In a following article, the analogous homo-bimetallic [Mo(Cp*)(Me)3 ]2 (μ-N2 ) and hetero-bimetallic [W(Cp′ )(Me)3 ](μ-N2 )[Mo(Cp*)(Me)3 ] were made [75]. In a series of articles, Schrock and coworkers reported the use of the platform [(N3 N)Mo—N=N—Mg(THF)2 ] to generate hetero-polymetallic complexes for which the “(N3 N)Mo—N=N− ” anion is an X-type ligand, similar to N3 − . With FeCl2 , the trigonal planar [{(N3 N)Mo(N=N)}3 Fe] is obtained (Scheme 4.11) [76], whereas the [{(N3 N)Mo(N=N)}2 VCl(THF)] and [{(N3 N)Mo(N=N)}2 ZrCl2 ] complexes are formed from VCl3 (THF)3 and ZrCl4 (THF)2 , respectively [77].

W N H3C H3C CH3

Na/Hg

W H3C OTf H3C CH3

N2

H3C N

CH3 CH3 W

H3C N

W N X H3C CH3

HX

CH3 X W

X = Cl, OTf, O2CC6F5, OC6F5, SC6F5

Scheme 4.10

N N

R

N

THF

R

N N N

Mg N

THF N

N

N

N

N

Mo R

R

R = SiMe3

Scheme 4.11

R

R

N Mo

N

R

R N N

FeCl2 –MgCl2

2/3

N N N

R

N

R R

N

Mo

R

Fe N

N

N

N

N

N

Mo

Mo R

R

R

N N N

229

230

4 Group 6 Transition Metal–Dinitrogen Complexes

Mixed Cp*/amidinate (or guanidinate) complexes of Mo and W(V) were used as precursors for bridged N2 complexes by Sita and coworkers in 2010. Reduction of these precursors by three electrons under N2 yielded the amidinate complexes in an excellent yield (77% and 92% for Mo and W complexes, respectively) and guanidinate complex in a poor yield (38%, M = Mo) as diamagnetic species ([Cp*M(N(iPr)C(X)N(iPr))]2 (μ-N2 )). Here also, NN bond distances of ca 1.27 Å were measured. Interestingly, reduction of the Mo(IV) complex with BuLi resulted in the formation of the dinuclear end-on-bridged μ-η1 :η1 -N2 dihydride meso complex, also as a diamagnetic species. In this complex, the NN bond distance is shorter at 1.189(4) Å. Evolution of H2 was observed upon mild heating, to form the [Cp*Mo(N(iPr)C(Me)N(iPr))]2 (μ-N2 ) complex (Scheme 4.12) [78].

Cl M Cl N Cl N

THF, 25 °C 3 Na/Hg

iPr

iPr N

N2

N

N

N

iPr

X

X

X N

M

M

N

iPr

X = Me or NMe M = Mo, W

tol, 50 °C – H2

N

Mo Cl N

Cl

Et2O 2 BuLi N2

iPr N

M N

iPr H N

N H iPr

N

M

N

iPr

Scheme 4.12

In the bridging dimers, the question of the electronic ground state can be difficult to address. Indeed, it will depend on the extent of π backdonation from the two metals into the π* orbitals of N2 . Thus, two extreme resonance forms can be written, the first featuring two Mn metals with neutral N2 ligand and the second with two M(n + 2) metal centers with reduced N2 4− ligand. In reality, the true electronic configuration lies in between these two extremes. Accordingly, even with bridging N2 featuring elongated NN bond consistent with N2 2− , displacement of N2 by neutral π acceptor ligands has been evidenced [78]. These works were motivated by the belief that end-on bridging N2 ligands are ideal for functionalization (protonation in particular) because of a formal activation, i.e. reduced NN bond. This proved to be rather deceptive, and the most interesting transformation of bridging N2 ligand is the reductive cleavage to yield terminal nitride complexes. In fact, this requires a transfer of a total of six electrons into the π* and σ* manifold of N2 , thus three electrons per M center of the M–N2 –M moiety. The first example of such process was given by Cummins and coworker. The coordinatively unsaturated Mo(III) complex [Mo(N(R)Ar)3 ] (R = C(CD3 )2 CH3 ) reacts with N2 at −35 ∘ C to first form a monometallic complex, which dimerizes to [Mo(N(R)Ar)3 ](μ-N2 ) (Scheme 4.13). This latter

4.2 Preparation of Group 6 Transition Metal–Dinitrogen Complexes

Na(THF)x N Ar(R)N

N2 Mo

N(R)Ar

Ar(R)N

N

N Ar(R)N

Mo

Ar(R)N

N(R)Ar

N(R)Ar

Ar(R)N

A

N

Na/Hg

Mo

Ar(R)N A

–A

R = C(CD3)2CH3 N(R)Ar

N(R)Ar

Ar(R)N

Ar(R)N Mo

2 Ar(R)N

N

N

Mo

N

Ar(R)N

N(R)Ar

Ar(R)N

N(R)Ar

Mo N

e transfer

N

Mo

Ar(R)N

Na(THF)x

Mo N(R)Ar

N(R)Ar Ar(R)N

N(R)Ar

Ar(R)N

Scheme 4.13

complex has a half-life of 35 minutes at 30 ∘ C and evolves to the terminal nitrido complex [Mo(N)(N(R)Ar)3 ] [79]. The mechanism of this splitting has been investigated in detail, and it was shown to involve a zig-zag transition state to yield the nitrido complex. This zig-zag transition state is indeed required for the electron flow (two net electrons) from the π symmetry orbitals into the σ symmetry orbitals. This electron flow results in the population of the orbital featuring N—N σ antibonding interaction [80]. In a subsequent article, they reported that not only the binding of N2 can be accelerated by the addition of reducing agent but also that the splitting of N2 by the monomeric [Mo(N(R)Ar)3 ] complex can be redox catalyzed [81]. Access to the monomeric anionic [Mo(N(R)Ar)3 (N2 )]− complex allowed them to also study the functionalization with electrophiles. Related reactivity was later observed with the “Mo(Mes)3 ” fragment (Mes, 2,4,6-trimethylphenyl). Indeed, the thermally stable dimer [Mo(Mes)3 ]2 (μ-N2 ) was formed by the reaction between MoCl4 ⋅DME (DME, 1,2-dimethoxyethane) and MesMgBr. Irradiation with UV light results in N2 splitting, yet the terminal nitrido complex was not observed, but a mixed-valent nitrido dimer [{Mo(Mes)3 }(μ-N)]. The authors propose that the intermediate terminal nitrido complex reacts with the starting complex [Mo(Mes)3 ]2 (μ-N2 ) to generate an intermediate tetra Mo complex, which finally breaks apart by N2 elimination to yield the bridging nitrido [82]. In 2012, Schrock and coworkers synthesized the [Mo(PCP)(I)2 ] pincer complex (PCP = C6 H3 -1,3-[OP(t-Bu)2 ]2 ), which upon an overall two-electron reduction yields the anionic [Mo(PCP)(N)(I)]− complex in 57% (Scheme 4.14) [83]. The proposed sequence involves a one-electron reduction to favor N2 coordination and dimer formation en route to splitting before a second reduction. Protonation of this anionic complex results in a rare example of a bridging H between Mo and P rather than the expected formation of the parent imido species.

231

232

4 Group 6 Transition Metal–Dinitrogen Complexes

O 2

O I

I

O

4 Na/Hg

2

P*

Mo

*P

O

I

P* = PtBu2

2 Na/Hg

O

Et3NH

2

P*

Mo

*P

O H

N

P*

Mo

*P

I

N

2 Na/Hg

O P*

*P O I

Mo

O 2

N N Mo I

O P*

O P*

Mo

*P

I

*P O

N

Scheme 4.14

Another example of the photochemical splitting of N2 was reported in 2014 by Nishibayashi and coworkers. Of the three dinitrogen-bridged dimers presented in Scheme 4.8, only the least activated complex provided the desired nitrido complex upon photolysis with visible light (>400 nm) in 33% yield. The one electron oxidation of the Mo(IV) nitrido forms the unstable and unobserved [Cp*Mo(depf )(N)]+ complex, which spontaneously reforms the N—N bond, and the known dicationic dimer [Cp*Mo(depf )](μ-N2 )2+ [69]. In 2015, Sita and coworkers reported the photochemical splitting of [M(Cp*)(N(iPr)C(CH3 )N(iPr))]2 (μ-N2 ) (M = Mo, W). In the case of Mo, they observed a mixture of bridging nitrido complexes (bis-μ and mono-μ nitrido): the Mo(V)–Mo(V) [Mo(Cp*)(N(iPr)C(CH3 )N(iPr))]2 (μ-N) and Mo(III)-Mo(IV) [Mo(Cp*)(N(iPr)C(CH3 )N(iPr))(μ-N)]2 complexes (Scheme 4.15) [84]. In the case of W, the reaction appeared cleaner, yet the long reaction times required for full conversion did not allow them to isolate large-scale quantities of the desired [{Cp*W(N(iPr)C(Me)N(iPr))(μ-N)}2 ] species. An alternative synthesis of the bis(μ-N) complexes was devised to further probe their reactivity. In a subsequent work, they looked for the parameters to switch between the photochemical process and the more competitive thermal N2 splitting. Based on their extensive work on cyclopentadienyl/amidinate early transition metal dinuclear μ-N2 complexes, they reasoned that decreasing the steric crowding would lower the energy

R2 N

M N

R1

R1

R2 N N

N

R2

M

N

(A) hν, 25–40 °C M = Mo : 48 hours 2 M = W : 72 hours R

R2 (B) thermal 60 °C – 70 hours M = Mo and W

N N

M M N

R1

2

R

N

N

R2

+

R2 N

Mo N

1

R

N

Mo

R2

Observed path A R1 = Me or Ph R2 = iPr or Et M = Mo, W

Scheme 4.15

R1

R2 N

R1

R2 N

N

R2

4.2 Preparation of Group 6 Transition Metal–Dinitrogen Complexes

2

Cl

Ph P

4 Na/Hg, PCy2 2 NaI

Mo P Cy2

P

N2

Cl

I

Ph

PCy2 Mo N N P Cy2

Cl

Cy2P P

Mo I

N

Ph

Ph

2

Mo

P

PCy2 I

PCy2

P Cy2

60%

Scheme 4.16

for the transition state, leading to the bridging nitrido complex. Not only was this strategy successful but it also provided the desired complexes in excellent isolated yields (Mo = 80%; W = 77%), opening ways for further functionalization of nitrido (vide infra) [85]. The same year, Mézailles and coworkers observed the spontaneous N2 splitting upon a two-electron reduction of the [Mo(PCy PPh PCy )(Cl)3 ] complex by two electrons (Na/Hg) in the presence of NaI. It is proposed that 2 equiv of the unobserved unsaturated Mo(I) intermediate coordinate N2 to favor the overall six-electron transfer. The nitrido–iodide complex [Mo(PCy PPh PCy )(N)(I)] is formed in a good 60% yield (80% NMR) (Scheme 4.16) [86]. It is interesting to note that the corresponding [Mo(PCy PPh PCy )(N2 )2 ]2 (μ-N2 ), although being more reduced, does not split N2 , which emphasizes again the importance of the geometrical constraints imposed to allow the two-electron transfer between the π orbital into the σ orbital (antibonding N—N). In 2017, the [Mo(PNP)(Cl)3 ] complex (PNP = N(CH2 CH2 PtBu2 )2 ) could be reduced by two electrons (Na/Hg) by Schneider and coworkers to form the bridged dimer [Mo(PNP)]2 (μ-N2 ) in 70% yield as a diamagnetic species (Scheme 4.17). The short Mo—N2 bond (1.799(4) Å) and long N—N bond (1.258(9) Å) are indicative of a strong activation of N2 with a double-bond character, also corroborated by the band at 1343 cm−1 in the Raman spectrum [87]. This complex is thermally stable toward N2 splitting (several hours at 60 ∘ C). Double protonation at the amido substituent of the tridentate ligand triggered the N2 cleavage, and the [Mo(PNH P)(N)(Cl)]+ X− (X = OTf, BArF 4 )

P

P

N Mo

N

N

P Cl

N

Mo P

N

H

2 HOTf 2

P = PtBu2

N

Mo t

P Bu2

Cl

OTf PtBu2 Cl 80%

HOTf OTf P

N H Mo

P N

P Cl

Scheme 4.17

N

N

Mo P

Cl

HOTf

P

(OTf)2

N H Mo

P Cl

P N

N

N

Mo P

Cl

H

233

234

4 Group 6 Transition Metal–Dinitrogen Complexes

complex was obtained in 80% yield. This proton induced N2 splitting being quite counterintuitive, its mechanism was probed experimentally and by density functional theory (DFT) calculations. It was linked to a change in ground-state multiplicity between the neutral, protonated intermediate, and doubly protonated complexes. Indeed, the latter is a quintet for which the barrier for N2 cleavage is reduced significantly. Together, these works show that N2 cleavage at Mo can be a facile process. It necessitates crucial conditions to be met: (i) formation of a dimeric Mo intermediate with a bridging N2 ligand (or a Mo fragment associated to another metallic fragment), and (ii) an energetically accessible zig-zag transition state to allow (by symmetry) the transfer of two electrons from the π system of N2 into the higher energy σ* system. This latter condition can be met either thermally or by photolysis. Functionalization of the nitrido complexes thereby obtained will be presented in a subsequent part.

4.3 Stoichiometric Reactions of Group 6 Transition Metal–Dinitrogen and Metal–Nitrido Complexes 4.3.1

N—H Bond Formation

Chatt et al. reported the first reactivity at Mo and W coordinated dinitrogen in 1972. They showed that the acid chloride RCOCl (R = Me, Et, Ph, p-MeOPh) reacted with trans-[M(dppe)2 (N2 )2 ] in THF to form [M(Cl)2 (dppe)2 (N2 HCOR)] (Scheme 4.18) [88, 89]. The H atom presumably comes from partial hydrolysis of RCOCl. They also studied the reaction toward acids [25, 90]. Addition of anhydrous HX (X = Cl or Br) to trans-[M(dppe)2 (N2 )2 ] and trans-[Mo(depe)2 (N2 )2 ] H

NH Ph2 NH P M P Ph X X 2

Ph2 P P Ph2

–X

N

Ph2 N P M P Ph2 X

H Ph2 P P Ph2

–N2 HX, THF

H

N Ph2 P

N

Ph2 P

M P Ph2

N N

Scheme 4.18

P Ph2

RCOCl, THF –N2

N

COR

Ph2 Ph2 N P P M P P Ph2Cl Cl Ph2

4.3 Stoichiometric Reactions of Group 6 Transition Metal–Dinitrogen

(depe, bis-diethylphosphinoethane) led to the diazene complexes [M(X)2 (dppe)2 (N2 H2 )] and [Mo(Br)2 (depe)2 (N2 H2 )]. Interestingly, the diazene moiety in [M(X)2 (dppe)2 (N2 H2 )] isomerizes to hydrazido(2–) [M(X)(dppe)2 (N2 H2 )]X upon halide decoordination. X-ray crystal study of the hydrazido derivatives show the linear M–N–N arrangement, trigonal planar geometry at the terminal N, and N—N bond order greater than one [91, 92]. The protonation of complexes featuring dppe classically stopped at the hydrazido(2–) stage. In one of these seminal articles, Chatt et al. also studied the reactivity of cis-[W(N2 )2 (PMe2 Ph)4 ] toward acid [25]. Addition of excess amounts of HCl in THF did result in the formation of diazene species cis-[W(Cl)(PMe2 Ph)4 (N2 H2 )]Cl. They later reported that quantitative formation of ammonia (1.98 equiv per W center) was achieved when H2 SO4 in MeOH was added to complex cis-[W(N2 )2 (PMe2 Ph)4 ] [93, 94]. The use of MeOH was crucial for the reaction, as other solvents provided mixtures of NH3 and N2 H4 . Similar reactivity with the Mo analog did result in a lower yield of 36% of NH3 . The sequence of N2 reduction all the way to NH3 was thus clarified by Chatt et al. and involves diazenido (MNNH), hydrazido(2–) (MNNH2 ), hydrazinium (MNNH3 + ) [95], and nitrido (MN) intermediates formed by stepwise protonations and electron transfers from the metal center. These formed the basis of the so-called Chatt cycle (Scheme 4.19). The intermediates involved have been investigated, and the protonation of M–N2 complexes has become a H H H

P

P

M

P

P

P

P

M

H

N

N

P

P

H

N

H

N

N

M

H

N

e P

N2

N

N2 + 6H

P

P

N P

e

P

H

P

N

P

N 2 P P

P

P

P

M P

Scheme 4.19

N

P N

M

H

N

P

P

P H

P

M P

H

P

M H

H

H

P

H

M P

P

P

N NH3

2

M

H

H

H

P 2 NH3 2e

P

N N

+ 6e

M P

H

Chatt cycle 2e

H 2

P

P

H

H

P

P

NH3

H

M

H

M

H

P

P

P

P

P

H

N

P

P

P

P

235

236

4 Group 6 Transition Metal–Dinitrogen Complexes

benchmark reaction to probe the reactivity of a newly developed complex. It is to be noted that protonation of the hydrazido intermediate at the proximal N atom yields hydrazine rather than NH3 . These two pathways are therefore in kinetic competition, typically in favor of NH3 for group 6 metals. Accordingly, both the hydrazido(2–) derivative and extensive protonation to NH3 /NH2 NH2 are searched. The first complex only requires two protons from the M–N2 species but leads to a formal four-electron oxidation of the metal center posing strong constraints on the ligand system that has to efficiently stabilize low- and high-oxidation states of the metal. The extensive protonation, leading either to NH3 or to hydrazine derivatives, provides additional indications on the efficiency of electron transfer. The metal complexes stabilized by various polydentate systems were thus tested in these benchmark reactions [9, 25, 33–35, 90–92, 96–101]. Interestingly, Mock and coworkers have studied in 2015 the site of protonation for a family of four related complexes featuring bidentate dialkylphosphines (Scheme 4.20). When an additional pendant arm is present in the diphosphine, protonation can occur in a competitive manner at the three possible sites Mo, N2 , and pendant amine, proving again the dramatic electronic and steric influences of the ligands [102]. Heterobimetallic complexes featuring bridging N2 ligands were also tested for NH3 /NH2 NH2 formation [72–75]. N

N Et2 P

N Mo

Ar

Ph

N N

PMe2Ph PMe2Ph

P

N

N Ph N

N

N Mo

Ph

PMe2Ph

P N

Et2 P

PMe2Ph

N

Ph

Ar H

Et2 P

N

N

N

N

Mo

PMe2Ph PMe2Ph

P N N

Et2 P

iPr iPr

N

iPr

Mo

PMe2Ph PMe2Ph

P N N

N

iPr

Scheme 4.20

Pickett and Talarmin built on these discoveries to develop an electrochemical process by which an external source of electrons is used rather than the metal center. Most importantly, starting from the [W(OTs)(dppe)2 (NNH2 )]+ complex, the parent trans-[W(dppe)2 (N2 )2 ] complex is regenerated upon NH3 formation. After three cycles, the yield of NH3 was 73% (vs. W complex) (Scheme 4.21) [103]. Metal hydrides were subsequently used to yield low-to-moderate amounts of NH3 from complex cis-[W(PMe2 Ph)4 (N2 )2 ] [104, 105]. The bis(dinitrogen)–M

4.3 Stoichiometric Reactions of Group 6 Transition Metal–Dinitrogen

N2

NH3 THF

H

H

N

Ph2 N P W P Ph2 OTs

N –2.6 V vs. Fc/Fc+

Ph2 P

Ph2 P

N

W P Ph2 N

P Ph2

Ph2 P P Ph2

N

TsOH·H2O

N2

Scheme 4.21 (dppp)2RuCl H2 N2 Me2PhP Me2PhP

W

PPhMe2 +6

N2 PPhMe2

Ph2 H H P Ru P Ph2 Cl

Ph2 P

H2, 1 atm 55 °C, 24 hours

P Ph2

Ph2 H P Ru P Ph2 Cl

2 NH3 + 6 ~55%

+ N2

N

W P Ph2 N

P Ph2

+ W(VI) species (?)

NH2

N Ph2 P

Ph2 P

Ph2 P +2

N H

HB(C6F5)3

P Ph2

Toluene rt 30 hours

Ph2 N P W P Ph2 H

Ph2 P

HB(C6F5)3

P Ph2

H2

N

N

B(C6F5)3

Scheme 4.22

complexes being stable toward H2 were envisaged as the source of H to form NH3 , much like in the Haber–Bosch process. Morris and coworkers first showed that the complex [Ru(Cp)(PP)(H2 )] could be used to protonate trans-[W(dppe)2 (N2 )2 ] to form the corresponding hydrazido(2–) derivative [106]. Few years later, Hidai and coworkers further developed this approach using [Ru(dppp)(Cl)(η2 -H2 )]X as the proton source. With trans-[W(dppe)2 (N2 )2 ], the reaction stopped at the hydrazido(2–) derivative, whereas with cis-[W(PMe2 Ph)4 (N2 )2 ], 55% yield of NH3 was obtained (Scheme 4.22) [107]. Mechanistically, this reaction is believed to involve the direct attack of the basic terminal N atom on the coordinated dihydrogen ligand, leading to N—H bond formation and Ru hydride. The cis-[W(PMe2 Ph)4 (N2 )2 ] complex being more

237

238

4 Group 6 Transition Metal–Dinitrogen Complexes

basic than the trans-[W(dppe)2 (N2 )2 ] corroborates the experimental formation of NH3 only with the former. Accordingly, no reactivity with the Mo analogs was observed. In parallel to these results, hydrosulfide-bridged complexes were used as proton transfer agents. The hydrosulfido complexes can be used directly [108], or generated in situ from H2 . Thus, when cis-[W(PMe2 Ph)4 (N2 )2 ] was reacted with 10 equiv of [(CpMo)2 (μ-S2 CH2 )(μ-S)(μ-SH)][OTf ] under H2 , ammonia was formed in 38% yield. It is interesting to note that the efficient generation of the hydrazido(2–) intermediate appears to be crucial, as if it is used as the starting material, near quantitative yield of NH3 is formed. If similar reaction is carried out with the dppe analog, the hydrazido(2–) complex Mo=NNH2 + is again observed as the final species [109]. Hidai and coworkers also studied the use of a variety of bridging hydrosulfido complexes of Mo, Ru, Rh, and Ir as proton donors for terminal N2 complex cis-[W(PMe2 Ph)4 (N2 )2 ], mimicking the postulated mode of action of enzymes. Upon addition of 2 equiv of these complexes, the hydrazido(2–) complex was observed and variable amounts of NH3 were measured upon addition of 10 equiv, with Ir being the most efficient (78% yield of NH3 per W) [108]. In 2017, Sivasankar and coworker used the concept of FLP (frustrated Lewis Pair) to activate H2 and efficiently transfer H+ to terminal W–N2 of trans-[W(dppe)2 (N2 )2 ]. The double protonation to form the trans-[(H)W(dppe)2 (N2 H2 )][HB(C6 F5 )3 ] complex in 70% yield was observed. This complex evolves within days to the corresponding trans-[W(dppe)2 (F)(N2 H2 )][HB(C6 F5 )3 ] in 36% yield (Scheme 4.22) [110]. Finally, protonation studies with two families of M–N2 complexes developed by Schrock (featuring trisamido-amine (NN3 )Mo fragment) and Nishibayashi (featuring neutral tridentate ligands (PNP, PPP, and PCP)) have been done. They have led to the discovery of the only known catalyzed N2 to NH3 processes with group 6 metals and will be presented in Section 4.4.3. 4.3.2

N—C Bond Formation

The first CN bond formation from a terminal W–N2 complex resulted from the reactions with acyl and aroylchlorides [88, 89]. Chatt et al. and George and coworkers have extended this nucleophilic reactivity toward alkyl halides (typically bromides) [111, 112]. However, in this case, at least, the first step proceeds via the attack of the alkyl radical generated under irradiation with a W lamp at the terminal N (Scheme 4.23). The second addition at the same N center involves SN 2 mechanism when MeI is used [112]. It is important to note that this chemistry relies on the nucleophilic character of the terminal N center and is thus limited to the Mo and W featuring bidentate phosphines (dppe or depe1). These reactions do not occur with the monophosphine complexes. Dialkylhydrazido(2–) complexes trans-[M(dppe)2 (Br)(NN(CH2 )n CH2 )] have been obtained with α,ω-dibromide [113], whereas gem-dibromides react to form diazoalkane complexes [114]. Based on these results, Pickett and Leigh developed an electrochemical synthetic cycle to produce hydrazines from dinitrogen. The key point in the strategy envisioned is the reduction of the [XM(NNR2 )]X M(IV) intermediate efficiently into the M(II) derivative. The hydrazines can be liberated from the

4.3 Stoichiometric Reactions of Group 6 Transition Metal–Dinitrogen

R′ R

N Ph2 P

N

Ph2 P

M P Ph2 N

hn

Ph2 P

Br(CH2)nBr –N2

M P Ph2 Br

P Ph2

1,2

Br

Ph2 N P M P Ph2 Br

Ph2 P P Ph2 Br

R

N

Ph2 P

N hν

RBr – N2

Ph2 N P M P Ph2 Br

–N2

P Ph2

N

N

R, R′ = H, Me hν RR′CBr2

Br N

N Ph2 P P Ph2

hν MeBr R = Me

Ph2 N P M P Ph2 Br

Ph2 P P Ph2

Scheme 4.23

metal center upon double protonation and two-electron reduction, which regenerates the “M(N2 )2 ” complex. Starting from 1,5-dibromopentane and the complex trans-[Mo(dppe)2 (N2 )2 ], N-aminopyrrolidine was formed in 60–70% yield. Most interestingly, the starting complex can be reformed in up to 45% yield, although the yield is said to be variable (15–45%) (Scheme 4.24) [115]. In line with these results, the Mo tetra-thioether complex developed by Yoshida et al. appeared much more reactive toward organic halides. Not only does trans-[Mo(Me8 [16]aneS4 )(N2 )2 ] react with MeI but it also reacted with PhCH2 Br or even ArX (X = Br, I; yields 17–32%) (Scheme 4.25) [61]. Formation of C—N bonds has been most extensively studied by Hidai and coworkers taking advantage of the reactivity of the hydrazido(2–) intermediates. Indeed, these stable compounds, readily made by double protonation, react as “classical” amines in condensation reactions with carbonyl derivatives. In this way, stoichiometric formation of various N derivatives can be achieved from N2 [2]. For example, diazoalkane complexes can be obtained from hydrazido(2–) complexes and 1,3-diketones [116]. Subsequent reaction with KOH in EtOH yields pyrazole derivatives (Scheme 4.26) [117]. Similarly, reaction with phthalaldehyde in the presence of catalytic HCl generated the diazo intermediate, which cyclized to form the phthalimidine ring. The W–N cleavage can then be achieved using either HBr or KOH [118]. A synthetic cycle to form pyrrole derivatives was devised. It involves the double protonation to from the hydrazido(2–) derivative in a first step. Then,

239

240

4 Group 6 Transition Metal–Dinitrogen Complexes

1,2

Br N Ph2 P

Ph2 N P Mo P Ph2 Br

hν

2e P Ph2 2 Br

Br(CH2)nBr –N2

1,2

N Ph2 P

N

N

Ph2 P

Ph2 N P Mo P Ph2

Mo P Ph2

P Ph2 N N

2 N2

1,2

2e

N

Ph2 P P Ph2

+ 2H

NH2

Scheme 4.24

H3C

: CH2–C(Me)2–CH2

N

H3C

CH2Ph

I +

M

MeI/ toluene

S

S S

M S

S I

I

–N2

M

S

S

S

S

N N

CH3

N

N

S

N

I

S

S N N

ArX –N2 PhCH2Br – N2

Ar

Ar

N

N

Br N S

S M

S

S Br

Scheme 4.25

S

S

CH2Ph

CH3 I

N PhH2C

N

M X

S

S M

S

S

MeI

N S

S I

4.3 Stoichiometric Reactions of Group 6 Transition Metal–Dinitrogen R3

R1 = Me, tBu; R2 = Me, H; R3 = Me, Ph, tBu

R2

O

R2

N R1 /H

R3

N L W

O

O

N L W

PPhMe2

N

EtOH

N

R1

H

Cl PPhMe2

R2

R3 KOH

Cl

Me2PhP

NH2

R1

L = PMe2Ph, CO

Cl

Me2PhP Cl

CHO O CHO HBr

N NH2

OHC O

N N L W

PPhMe2 Cl

Me2PhP

H

N N

L W

Cl

Me2PhP

Cl

KOH PPhMe2

Cl

O

N

+ NH3

H

Scheme 4.26

condensation with a cyclic acetal of succinaldehyde occurs to form the dialkyhydrazido(2–). This isolable complex reacts with hydride sources to liberate the pyrrole ring and form the W tetrahydride complex (Scheme 4.27). Finally, the latter complex is known to be reduced to the starting bis-dinitrogen complex under photolytic conditions [119]. Metal nitrido complexes, obtained from N2 splitting, are appealing as N1 synthons. Cummins and coworkers have studied the reactivity of the [(N3 )MoN] Mo(VI) complex toward various electrophiles. It appeared as a weak nucleophile and could only be reacted with strongly activated substrates, such as RCO(OTf ), SiMe3 OTf, or MeI to yield the corresponding imido complexes. Functionalization of the latter complex can be done, and a synthetic cycle to produce nitrile derivatives was developed (Scheme 4.28) [120]. 4.3.3

N-element Bond Formation

Hidai and coworkers confirmed the nucleophilic character of the terminal N center of trans-[M(N2 )2 (dppe)2 ] and cis-[M(N2 )2 (PMe2 Ph)4 ] via silylation reactions with R3 SiI or R3 SiCl/NaI, at 50 ∘ C under rigorously dry conditions, to form the corresponding diazenido complexes trans-[M(I)(PMe2 Ph)4 (N2 SiR3 )] and trans-[M(I)(dppe)2 (N2 SiR3 )] (M = Mo, W) [121, 122]. The second, intermolecular, addition to form the hydrazido(2–) “M=N—N(SiR3 )2 ” complex did not occur even with excess of R3 SiCl/NaI. However, it can be achieved in an intramolecular

241

242

4 Group 6 Transition Metal–Dinitrogen Complexes

NH2 N2

N

P

P W

aq HBF4

P

–N2

P

P

P

P

W P F

N2

OMe O

N2 hn

– 2 H2

OMe

H H

P

P W P

N

P

H H

N

+

, N

N

NH2

H

P

P

LiAlH4

W P

P F

+ NH3

Scheme 4.27 N Mo

*N

N2 NaH

N*

RCOCl R′3SiOTf

O

*N R

N Mo

*N *N

N*

Mo N*

N* =

*N

*N

N

OTf

Mg(anthracene) SiMe3OTf

Mg R

Cl *N

Mo

+

*N R

N

R = Ph, tBu, Me

Scheme 4.28

OSiMe3

N

N*

*N

*N 1/2 SnCl2 or 1 ZnCl2

Mo N*

4.3 Stoichiometric Reactions of Group 6 Transition Metal–Dinitrogen

N N Me2PhP

W

SiMe3

Me2Si

N

PPhMe2

Me2PhP

PPhMe2

Me2PhP

SiMe2

N

PPhMe2

W

I

Me2PhP

I

I N

SiMe3Cl/NaI or SiMe3I

N Me2PhP

50 °C

W

Me2Si PPhMe2

/ NaI

Cl

PPhMe2

Me2PhP N2

GaCl3 or AlCl3∙Py

R3P

L

Cl2 M

PR3 W

Cl PR3 = PPhMe2 L = Py or PPhMe2

SiMe2

Cl

N PR3

N

PR3

R3P N

M Cl2

N R3P

W

Cl L

M = Al, Ga

Scheme 4.29

manner, using the ClSiMe2 CH2 CH2 SiMe2 Cl/2NaI couple (Scheme 4.29). Most importantly, reduction of the diazenido complex trans-[M(I)(PMe2 Ph)4 (N2 SiR3 )] with Na was shown to give mixtures of silylamines, which was the basis for the first efficient catalytic process of N2 reduction (vide infra) [122]. The related germylation was reported later with the W monodentate phosphine complex to form trans-[W(I)(PMe2 Ph)4 (N2 GePh3 )] [123]. The Lewis basic character of the terminal N2 complexes had been studied early on, and adduct with AlMe3 was formed with the dppe complexes of Mo and W, but was not stable with monodentate phosphines [124]. On the other hand, reactivity toward AlCl3 and GaCl3 resulted in the formation of bridging dimers featuring the hydrazido(2–) moiety [125, 126]. Silylation of neutral and anionic dinitrogen complexes featuring depf ligand has subsequently been used successfully to synthesize the silyldiazenido complex (Scheme 4.30) [69, 127]. Mézailles and coworkers have studied the stepwise silylation/reduction of terminal N2 complexes relevant to the N2 -to-N(SiMe3 )3 catalytic process. The silylhydrazido(2–) and imido intermediates have been isolated with a tetradentate phosphine “PP3 ” ligand (Scheme 4.31) [128]. In a following study with a tridentate ligand “PP2 ,” the stepwise reduction of a mixture of [Mo(PCy PPh PCy )(N2 )x ] (x = 2, 3) in the presence of chlorosilanes and NaI allowed isolation of not only the silylhydrazido(2–) and silylimido complexes but also the nitrido derivative [Mo(PCy PPh PCy )(N)(I)] (Scheme 4.32) [58]. This

243

244

4 Group 6 Transition Metal–Dinitrogen Complexes

MoCp*Cl4

Et2 P

N2 20 Na/Hg

PEt2 Fe

Mo

Fe

SiMe3Cl

Mo

Fe N

P Et2

PEt2

Et2 P

N2 SiMe3Cl

N

N

P Et2

SiMe3

N Na(15-crown-5)

MoCp*Cl4

Scheme 4.30

+

PCy2 P

Mo

BPh4– Cl

PCy2

PCy2

4 Na/Hg 2 Cl R 6 Na/Hg 4 Cl R PCy2 P

Mo PCy2

2 Na/Hg 2 Me2SiR′Cl

R N

PCy2

N

P R

PCy2

Si

PCy2

Mo

N TMS

R′ N Si CH2 CH3

PCy2

, N(TMS)3

Si Si NR2 = N

TMS , N

Si

or

TMS TMS

TMS

R′ = CH3, Ph

N SiMe2Ph

Scheme 4.31

N2

Ph P

PCy2

Mo P Cy2 Ph P P Cy2

Me2Si

SiMe2 Ph

Cl

Cl

N2 Mo

+

Me2Si

SiMe2 N

N2

P PCy2

NaI

N

P Cl Cy2

N2

PCy2

Mo I

– NaCl

Ph P

N Mo

PCy2

I

Me2Si

+

SiMe2

N

PCy2

Na

PMe3

RSiMe2Cl

Me2Si

80 °C 12 hours

rt

SiMe2 H

H

I

P Cl Cy2 R = CH3, Ph, CH2CH2SiMe2Cl

I Ph P

PCy2 Mo P H Cy2

Scheme 4.32

PCy2

Mo

P

N2 SiMe2R

N

Ph

2 Na/Hg

+ PMe3

Me2Si N H

SiMe2

4.3 Stoichiometric Reactions of Group 6 Transition Metal–Dinitrogen

terminal nitrido complex appeared to be reactive toward Si—H bonds providing a synthesis of bis-silylamine in one pot under mild conditions [86]. In 2015, Sita and coworkers have proposed a synthetic cycle to produce SiMe3 NCO. It involves N2 splitting under photolytic conditions and subsequent functionalization of the nitrido intermediate with SiMe3 Cl to form both the silylimido and the dichloro intermediates (Scheme 4.33). The silylimido complex is nucleophilic and reacts with CO2 to yield the corresponding isocyanate species and metal oxo complex. Interestingly, the metal oxo can react with excess SiMe3 Cl to form the dichloro intermediate and TMS–O–TMS as a by-product. Regeneration of the bridging dimer can be achieved with Na [84].

iPr N

3

M N

R1

R2 N N

N

iPr

M

N

hν 4 Me3SiCl

4

R2

iPr N

M

N

M = Mo, W

M N

–N2

12 Na + 3 N2 –12 NaCl

6

iPr

N

SiMe3

iPr

4 CO2 4 Me3SiNCO

8 Me3SiCl

Cl Cl N iPr

4

iPr N

M N

O iPr

4 Me3SiOSiMe3

Scheme 4.33

In a subsequent study, with the related system that splits N2 thermally, they could achieve the protonation of the silylimido derivative with various alcohols (iPrOH, tBuOH, Me3 SiOH, and PhOH) and thus generate the bis-silylamine HN(SiMe3 )2 as the N-containing derivative (Scheme 4.34) [129]. BH bond addition of boranes across the M–N2 fragment was initially studied by Hidai and coworkers, who showed that it could work only with the more nucleophilic W complex and after activation, i.e. replacing one N2 moiety by NCS− [130]. Thus, trans-[W(dppe)2 (NCS)(N2 )]− complex reacted with (CH3 )2 CHC(CH3 )2 BH2 to form trans-[W(dppe)2 (NCS)(N2 BHC(CH3 )2 CH(CH3 )2 )]. In 2017, Simonneau et al. have used BCF (tris-pentafluorophenyl borane) to activate coordinated N2 toward BH and SiH functionalization [131] (Scheme 4.35). In conclusion, the seminal works by Chatt and coworkers and Hidai and coworkers pertaining to the protonation of the terminal dinitrogen complexes of group 6 laid the ground for the development of the catalytic reduction of N2 into NH3 or silylamines, which are presented below. The direct N2 splitting by two trisamido–Mo fragments reported by Cummins in the 1990s has witnessed a surge of interest in the 2010s. Although impressive developments in terms of

245

246

4 Group 6 Transition Metal–Dinitrogen Complexes

Ph

Et N

Et

3

N

Mo

N

Mo N

Ph

12 Na +3 N2

6

N

Et N Ph

Et

4 Me3SiCl

N

Et

4

Et

Mo

N

N

–N2

Ph

N

4 Me3SiCl 4 ROH

–12 NaCl

4 Me3SiCl

Mo

Cl Cl N Et

4

Et N

Me3Si 4

N

Ph

SiMe3

4 ROSiMe3

Mo

Cl N

N

SiMe3

Et

Et

H

SiMe3

H

Scheme 4.34

B(C6F5)2 N Ph2 N P M P Ph2

N Ph2 P

N

Ph2 P

HB(C6F5)2

M P Ph2 N

B(C6F5)3

P Ph2 tol.

N

N

Ph2 P

HSiEt3

P Ph2

Ph2 P

NCS

Scheme 4.35

P Ph2

B BH2

Ph2 P P Ph2

HB(C6F5)3

N Ph2 N P W P Ph2 NCS

SiEt3 N Ph2 N P M P Ph2

H

W P Ph2

HB(C6F5)3

B(C6F5)3

N Ph2 P

P Ph2

N

–N2

Ph2 N P M P Ph2

NCS

Ph2 P

Ph2 P P Ph2

4.4 Catalytic Reactions of Group 6 Transition Metal–Dinitrogen Complexes

functionalization of the nitrido complex have been seen in the past years, so far no catalytic process has emerged from this approach.

4.4 Catalytic Reactions of Group 6 Transition Metal–Dinitrogen Complexes 4.4.1 Catalytic Formation of N2 H4 /NH3 from Nonisolated M–N2 Complexes In 1964, Haight and Scott reported that the reduction of N2 in aqueous solutions of molybdic acid or tungstic acid at Pt or porous graphite electrodes generates 10−4 –10−3 mol l−1 of NH4 + after 24 hours of bubbling [132]. This work prompted Vol′ pin and Shur to test the reduction of several metal salts or complexes by EtMgCl under high pressure of N2 (150 atm) [133]. Substoichiometric amounts of NH3 were formed (mol/mol metal): 0.17 equiv for CrCl3 , 0.08 equiv for MoCl5 , 0.15 equiv for WCl6 , 0.09 equiv for FeCl3 , and 0.10 equiv for TiCl4 . Interestingly, the use of Cp2 TiCl2 resulted in the formation of 0.9–1.0 equiv of NH3 . Few years later, Shilov studied mainly heterogeneous systems featuring M–OH moieties that are able to reduce N2 (PN2 = 100 atm) into N2 H4 and/or NH3 in protic media. The system could become catalytic when Mo was used. Thus, when the compound MgTi2 O2 was used as the reductant and Mo(OH)3 as the catalyst, 170 equiv of N2 H4 /NH3 were formed at 110 ∘ C. It is proposed that the Mo–N2 complex is adsorbed at the surface of active sites of MgTi2 O2 , and hydrazine/ammonia is formed via electron transfer from MgTi2 O2 and protonation by the solvent. Further refining of this model, based on electroconductive reductant, has led Shilov and coworker to use Na/Hg as the reducing agent, being “reasonably stable” with H2 O/MeOH [134]. After careful optimization, a highly active catalyst was developed: Mg[Mg2 Mo8 O22 (OMe)6 (MeOH)4 ] [6]. It was prepared by the reaction of MoCl5 with MeOH in the presence of MgCl2 . This complex, supported on the Na/Hg surface using a phospholipid surfactant and phosphines, proved very active in hydrazine synthesis under atmospheric pressure of N2 , i.e. reaching up to 1000 equiv/Mo. When the pressure was increased to 100 atm, impressive turnovers of ca 10 000 are reported. Although the use of phosphines allowed a drastic increase in the yields, their precise role was more difficult to elucidate. Indeed, it is stated that neither the ratio N2 H4 /NH3 nor the initial rate of reduction are changed in the presence of phosphines, which suggests that the nature of the catalyst is not changed. 4.4.2

Catalytic Formation of N(SiMe3 )3

Apart from the N2 -to-NH3 , there is only one catalytic process that reduces N2 into N-containing derivatives: the N2 -to-N(SiR3 )3 . In a seminal work, Shiina reported in 1972 that N2 could be transformed to N(SiMe3 )3 at room temperature using Li as the reducing agent and various metal salts as catalysts. The best result was obtained with CrCl3 , which allowed the formation of 5.4 equiv of silylamine [135]. Turning to phosphine-stabilized complexes, Hidai

247

248

4 Group 6 Transition Metal–Dinitrogen Complexes

and coworkers improved the efficiency to a large extent, obtaining 24 equiv of silylamine with cis-[Mo(PMe2 Ph)4 (N2 )2 ], using Na dispersion as the reducing agent in 1989 [122]. The trans-[Mo(dppe)2 (N2 )2 ] complex and the W analog cis-[W(PMe2 Ph)4 (N2 )2 ] provided only ca stoichiometric amounts of silylamine (1–3 equiv, respectively). Based on the fact that trans-[Mo(PMe2 Ph)(I) (NNSiMe3 )] yields N(SiMe3 )3 in the presence of excess Na and SiMe3 Cl, as well as the radical reactivity of the Mo–N2 complexes toward alkyl halides, Hidai and coworker proposed a radical mechanism for the catalytic process that goes through a diazenido intermediate and stated that “further investigation to elucidate the detailed mechanism” is needed (Scheme 4.36) [2]. N N Me2PhP N2

Me2PhP

SiMe3Cl

Mo

PPhMe2

Na

ClSiMe3 PPhMe2

NaCl

N

N N Me2PhP Me2PhP

Mo

N

PPhMe2

Me2PhP

N N PPhMe2

Me2PhP

N

N(SiMe3)3 NaCl Na, N2, SiMe3Cl

SiMe3

N Me2PhP Me2PhP

Mo

Mo

PPhMe2

∙

SiMe3 PPhMe2

SiMe3

SiMe3∙ Dimerization (competitive)

PPhMe2

ClSiMe3 PPhMe2

SiMe3Cl

Scheme 4.36

Using their trans-[Mo(depf )2 (N2 )2 ] complex, in 2011, Nishibayashi and coworkers improved this catalytic process tremendously. Indeed, when SiMe3 Cl (4000 equiv/Mo) and Na (4000 equiv) were reacted with N2 in THF for 100 hours, 150 equiv of N(SiMe3 )3 /Mo were obtained. Most impressively, the catalyst remains active after these 100 hours, as if the flask is recharged with 4000 equiv of Na and SiMe3 Cl, an additional 76 equiv of silylamine is formed for a total of 226 equiv, making the trans-[Mo(depf )2 (N2 )2 ] complex the most efficient catalyst for this reaction to date [127]. Other complexes, such as trans-[Mo(depf )(PMe2 Ph)2 (N2 )2 ] and the W analog trans-[W(depf )2 (N2 )2 ] also appeared as competent catalysts, but lower turnover numbers (TONs) were obtained (54 and 60, respectively) under the same conditions. Interestingly, the hydrazido complex resulting from the addition of SiMe3 OTf on the trans-[Mo(depf )2 (N2 )2 ] complex, trans-[Mo(depf )2 (N2 SiMe3 )(OTf )] was much less efficient with 44 equiv of silylamine. cis-[Mo(PMe2 Ph)4 (N2 )2 ] provided 32 equiv of N(SiMe3 )3 under these conditions, compared to the 24 equiv obtained by Hidai with only 100 equiv of Na and SiMe3 Cl. Using the

4.4 Catalytic Reactions of Group 6 Transition Metal–Dinitrogen Complexes

same ligand depf, the same authors later reported that the anionic complex [Cp*Mo(μ-N2 ){Na(15-crown-5)}(depf )]− was less efficient (TON = 22) [14]. These contrasting results highlight the sensitivity of the overall process to the nature of all ligands in the coordination sphere. DFT calculations have been used to propose a catalytic cycle consistent with the experimental data (Scheme 4.37). Here also, the active species is proposed to be the SiMe3 radical, generated by the reaction of SiMe3 Cl with Na in THF. This is corroborated by the experimental observation of Me3 Si–SiMe3 as a competitive by-product. The initial step involves attack at the terminal N by the Si radical to form the silyldiazenido intermediate. From that intermediate, N N

Et2 P

Et2 P

Mo

Fe

Fe Et2P

PEt2 N N

˙

SiMe3 –N2 SiMe3 N N

Et2 P

PEt2

N2

˙

Fe Et2P

PEt2

N N Et2 P Mo P Et2

Fe Et2P

Me3Si Me3Si

N

SiMe3 Me3Si Na

SiMe3 Me3Si

N

Et2 P P Et2

N

SiMe3

N Mo

Fe PEt2

˙

SiMe3 P Et2 Fe PEt2

SiMe3 Trissilyl-hydrazido intermediate

Scheme 4.37

P Et2

Bissilyl-hydrazido intermediate

Me3Si

Na

SiMe3

Me3Si

Et2 P Mo

N

N2

Silyl-diazenido intermediate

Et2 P Fe

˙

SiMe3

Mo

Fe

SiMe3

Et2 P

249

250

4 Group 6 Transition Metal–Dinitrogen Complexes

three different pathways have been evaluated to lead to the bis-silylhydrazido intermediate. They imply N2 decoordination/Mo–P cleavage/Si radical attack at the terminal N following the different combinations. All combinations are kinetically accessible. The rate-determining step is the third attack of the silyl radical. The two sites of attack have been envisaged: at terminal N, which leads to direct elimination of N(SiMe3 )3 , or at the proximal N, which forms the Mo–N(SiMe3 )–N(SiMe3 )2 intermediate. Most importantly, although the addition at the terminal N is thermodynamically much more favored than at the proximal N (ca −82 kcal mol−1 vs. −11 kcal mol−1 ), the pathway leading to amine formation was found to be kinetically significantly higher in energy (28.7 kcal mol−1 vs. 21.6 kcal mol−1 ), thereby providing a kinetic preference for the tris-silylhydrazido intermediate. Addition of one electron to this intermediate is a very facile and thermodynamically highly favored process that results in the elimination of the (SiMe3 )2 N–N(SiMe3 )− anionic derivative from the Mo center. The tetra-coordinated Mo center, coordinatively unsaturated, readily coordinates N2 to start a new cycle by the SiMe3 radical attack. A key feature of the present depf–Mo system lies in the capacity of the ligand to adopt flexible bonding modes to Mo in addition to their bulkiness and electron richness. Indeed, before the rate-determining step, one of the phosphine arms of one depf ligand dissociates from the Mo center, thereby providing the appropriate space for SiMe3 radical approach. In 2013, Masuda and coworkers have reported several Mo–N2 complexes featuring various bidentate bis(diphenylphosphino)amine RN(PPh2 )2 ligands, as well as their behavior in the same silylamine catalytic formation. Unfortunately, the performances of the catalysts were quite limited as 0.3–4.0 equiv of N(SiMe3 )3 were obtained (by GC/MS) [136]. In 2014, as mentioned above, Mézailles and coworkers reported the isolation of bis-silylhydrazido as well as silylimido complexes of Mo(II) stabilized by a tetradentate PP3 ligand by stepwise functionalization of coordinated N2 . They studied the catalytic N2 -to-N(SiMe3 )3 process. After optimization of the conditions (200 equiv of Si–Cl/200 equiv K, 50 ∘ C), moderate activity of the complexes was demonstrated (TON: 11.7–15.0). Interestingly, however, the bissilylhydrazido and silylimido complexes are similarly competent, which suggests their involvement in the catalytic process, and thus N—N bond splitting occurs in the coordination of Mo (Scheme 4.38). The use of Na/Hg as a reducing agent only provided stoichiometric amounts of silylamine, which is a proof in this case that Si radicals are also involved during catalysis, at least in one elementary step, as SiMe3 Cl is not reduced to Si radicals by Na/Hg [128]. In a subsequent study, the same authors have shown that the tridentate ligand “PP2 Cy ” provided more active catalysts. Indeed, using the 1,2-bis(chlorodimethylsilyl)ethane and K as the reducing agent (400 equiv of Si–Cl/400 equiv of K, 50 ∘ C), 33–39 equiv of NH4 + were obtained upon hydrolysis of the crude silylamine mixture. Again, they showed that all isolated intermediates, i.e. [(PP2 Cy )MoCl3 ], [(PP2 Cy )Mo(N2 )2 ]2 (μ-N2 ), [(PP2 Cy )Mo(=N—NSiMe2 CH2 CH2

4.4 Catalytic Reactions of Group 6 Transition Metal–Dinitrogen Complexes

+

PCy2 P

Cl

Mo PCy2

BPh4–

PCy2

2 e–, N2

Me3Si N SiMe3 Me3Si

2 TMSCl 2 e–

[Mo–N2]

2 TMSCl 2 e– N2 PCy2

PCy2 Mo N TMS

P

PCy2

PCy2

Me3Si N

SiMe3

P

Mo PCy2

TMS N

N TMS

PCy2

2 TMSCl 2 e–

Me3Si

Scheme 4.38

SiMe2 )(I)(Cl)], [(PP2 Cy )Mo(N)(I)(Cl)], and [(PP2 Cy )Mo(=NSiMe3 )(I)(Cl)] resulting from stepwise functionalization of N2 , are equally competent catalysts, pointing their involvement as intermediates (Scheme 4.39). Mechanistically, they showed that the ionic functionalization (Si–Cl + electrons) could be kinetically competitive with the Si radical one [58]. 4.4.3

Catalytic Formation of NH3

In 2003, Yandulov and Schrock reported the first N2 -to-NH3 process involving well-defined Mo complex as the catalyst [137]. The Mo(III) catalytic precursor features a sterically highly demanding ligand HIPTN3 N scaffold (HIPTN3 N = [{3,5-(2,4,6-iPr3 C6 H2 )2 (C6 H3 )NCH2 CH2 }3 N]3− ), which proved to be crucial as it also prevents the formation of bimetallic Mo–N2 –Mo complex and provides good solubility in nonpolar solvents. Addition of 48 equiv of [LutH][BArF 4 ] (sparingly soluble proton source in heptane) and 36 equiv of

251

252

4 Group 6 Transition Metal–Dinitrogen Complexes

R3Si N

N2

Ph P

SiR3

PCy2

Mo P Cy2

SiR3

2 SiR3Cl

N2 N2

–2 N2

Dinitrogen

4e + 2 SiR3Cl

R3Si

SiR3 N

SiR3 N

Ph

Mo

P

Cl

N

Ph

PCy2 Silylimido

Bissilyl-hydrazido

P Cy2Cl

P Cy2 Cl

SiR3Cl

P

N

Cl

2e

Nitrido Ph

PCy2

Mo

P

PCy2

Mo PCy2

Cl R3Si N

–Cl R3Si N

SiR3

SiR3

SiR3Cl

SiR3

Scheme 4.39

decamethylcobaltocene Cp*2 Co (Cp* = η5 -C5 Me5 ) as the reducing agent (slow addition using syringe pump) to a solution of [(HIPTN3 N)Mo(N2 )] under a N2 atmosphere yields ca 8 equiv of NH3 . This corresponds to an efficiency of ca 65%. The use of sparingly soluble source of H+ is important to limit the competitive proton reduction to H2 , as is the slow addition of the reducing agent. Most importantly, several intermediates, mainly obtained by stepwise electron and proton additions, have been isolated (in black in the scheme) and perform nearly identically in the catalytic process [64, 65, 138–140]. This has allowed Schrock to propose a catalytic cycle (Scheme 4.40), later also corroborated by DFT calculations [141–144]. Extensive studies on this system revealed that the limited activity of the catalyst is due to protonation at the amide moiety, leading to catalyst decomposition. Subsequent studies have been dedicated to improve the catalytic performance via tuning of the steric bulk and electronics of the triamido-amine ligand (Scheme 4.41) [66, 145, 146]. However, they did not meet success. The group of Nishibayashi reported in 2011 an improved catalytic system based on less basic (Brønsted) ligand scaffold, i.e. the neutral tridentate PNP

4.4 Catalytic Reactions of Group 6 Transition Metal–Dinitrogen Complexes NH3

R

R

N H

H R

N

R N

N

N

R

N N N

N2

NH3

R

N

e

R H

Schrock cycle

e N2 + 6 H

R

N N

N

Mo

R N

NH3

e R H

R

N

R N

N

H

R

Mo

N

R

N

N

R N

N Mo

R N

R

N

R N

N

iPr

R

R

e

H

N

N

H N H

Mo

N

R N N

:

N

R

Mo

N

N

R N

N

R N

N

iPr iPr

iPr and

R= iPr

H

N

H

N Mo

R N

N Mo

N

R

N with

H H

N

N

R N

e

H

N H

H

R N

Mo

R N

+6e

N

Mo

R

N

H

2 NH3

H

N N

H R

R N

N

Mo

H

N N

N

R N

e

R

N

H R

N

N

Mo

R

Mo

N

H

H

N

R N

N

R

Mo

R N R

R N

H

N

N

R

N

Mo

N N

= CH2CH2

iPr

Complexes in bold: isolated

Scheme 4.40

system (Scheme 4.42). Treatment of the Mo dinitrogen dimer [(PtBu NPtBu )Mo (N2 )2 ]2 (μ-N2 ) (with PtBu NPtBu = 2,6-bis[(di-tert-butylphosphino)methyl] pyridine) complex under N2 (atm pressure) with 72 equiv of CoCp2 (reducing agent) and 96 equiv of [LutH][OTf ] (proton source) for 20 hours in toluene yielded a total of 12 equiv of NH3 . This corresponds to an efficiency of 49%. A maximum of 23 equiv of ammonia was obtained with larger amounts of reducing agent and proton source (216 and 288 equiv, respectively) [54]. Here also, slow addition of the reducing agent to a suspension of the proton source is required to limit the formation of H2 . In the same study, the authors clearly evidenced the importance of having both the bridging and terminal N2 moieties. Indeed, starting from [(PtBu NPtBu )Mo(CO)2 ]2 (μ-N2 ) or [(PtBu NPtBu )Mo(PMe2 Ph)(N2 )2 ], only 0.4 and 0.2 equiv of NH3 was produced in the same experimental conditions. It was later confirmed as [(PtBu NPtBu )Mo(Cl)2 ]2 (μ-N2 ) and [(PtBu NPtBu )W(PMe2 Ph)(N2 )2 ]

253

254

4 Group 6 Transition Metal–Dinitrogen Complexes

BArF4

N2 + 6

+6

Cr

N

(1 atm)

2 NH3 heptane, rt 7 hours

H (36 equiv)

cat.

(48 equiv) R

N

R

N

N

Mo

cat.

R N N

= CH2CH2 tBu R′

Me

R′

R=

R′ =

R′ = tBu

tBu

NH3 : 1.49 equiv

NH3 : 1.06 equiv iPr

Br R′

Me

Me

R′ R′ =

R′ =

R=

iPr

iPr

NH3 : 7.0 equiv

iPr

N

iPr

NH3 : 2.53 equiv

iPr tBu R=

tBu

R= iPr

iPr

NH3 : 2.0 equiv

NH3 : 1.02 equiv

Scheme 4.41

did not yield NH3 production [147, 148]. Efforts were then paid to modulate the sterics of the ligand via the introduction of unsymmetrical PNP ligands of the form PtBu NPR where R = Ph, i Pr, Ad, and Cy [149]. Only the ligand with the bulky adamantyl moiety allowed a moderate improvement of the catalytic performance (14 equiv of NH3 vs. 12 for the tBu analog), and others showed a drastic decrease (between 1 and 3 equiv of NH3 ), as was also observed for the AstBu NAstBu ligand (2 equiv of NH3 ) [150]. A series of Mo–nitrido complexes, which are postulated intermediates in the NH3 synthesis, were synthesized from SiMe3 N3 and tested [151]. Quite interestingly, both the pentacoordinated Mo(IV) [(PtBu NPtBu )Mo(N)(Cl)] and cationic Mo(V) [(PtBu NPtBu )Mo(N)(Cl)]+ are as efficient as the Mo(0) dinitrogen dimer (6.8–7.1 equiv of NH3 vs. 5.9), but

OTf N2 + 6

Co

+6

(1 atm)

toluene, r.t. 20 hours

H (72 equiv)

P = PtBu2

cat. 2 NH3

N

TON: equiv/Mo

(96 equiv)

N N P N Mo

P

P

N2

N2

N

N2 Mo N

N N

N Mo P

P

TON = 5.9

N

P

TON = 1.2

TON = 6.6

N Mo P Cl TON = 1.2

Scheme 4.42

TON = 7.1

Py HN

P

Cl

TON = 1.3

N

N P O C Mo

OTf

O

P

PAd2

N Mo Cl

TON = 0.3

PAd2 Cl

TON = 1.8

Cl

P

TON = 6.8

BArF4

Na(15-crown-5) H N P N N Mo

F P BAr 4

F P BAr 4

N Mo Cl

P

N

OTf

P

N Mo Cl

P

Cl

H N

N

P

N Mo

Cl

N

P

N Mo NEt2 Cl

TON = 1.0

OTf

256

4 Group 6 Transition Metal–Dinitrogen Complexes

the hexacoordinated Mo(V) [(PtBu NPtBu )Mo(N)(Cl)2 ] is not efficient (1.2 equiv of NH3 ). Surprisingly, the Mo(V) [(PAd NPAd )Mo(N)(Cl)]+ cationic complex with symmetrical very bulky ligand (P-adamantyl moieties) performed poorly compared to the related Mo(0) dimer featuring the mixed PtBu NPAd ligand (1.8 equiv of NH3 vs. 7.0). Other modulations of the ligand resulted in drastic lowering of the efficiency of the process, becoming stoichiometric. In 2017, Schrock and coworkers also found catalytic reduction of N2 into NH3 using [(N2 N)Mo(N)(X)] complexes, where N2 N is a diamido-pyridine ligand and X is an alkoxide/aryloxide or chloride ligand. With the optimized conditions (108 equiv of Cp*2 Co and 140 equiv of Ph2 NH2 OTf ), up to 10.3 equiv of ammonia is formed, whereas with KC8 /H(Et2 O)BArF 4 , only 2.4–2.8 equiv is produced [152]. DFT calculations of the process catalyzed by [(PtBu NPtBu )Mo(N2 )2 ]2 (μ-N2 ), done independently by Nishibayashi and coworkers and Batista and coworkers, provided very significant information for further ligand optimization [151, 153]. The most important result is that, although the dimer can be in equilibrium with the monomer [(PtBu NPtBu )Mo(N2 )3 ] in solution, the double protonation at terminal N2 to form the hydrazido intermediate is only exergonic for the dimer. A strong kinetic preference for the pathway involving the dimeric species is thus evidenced for the first crucial steps. It was rationalized by the fact that the Mo fragment ((PNP)Mo(N2 )3 ) can act as intramolecular (via N2 bridge) electron donor to the doubly protonated hydrazido intermediate, which cannot occur at the monometallic complex (Scheme 4.43). P = PtBu2 H

N N P N Mo P N2

P

N

N2 Mo N

N N N2

2H

N Mo N2

H P P

P

H

N N

2H N Mo

N N

P

2

N2 Mo N

N N N2

∆G ca –15 kcal/mol

P

– N2

P

N Mo

– N2

N N P

N

P

Intramolecular stabilization

H 2

vs

P N N

No stabilization

∆G ca +20 kcal/mol

Scheme 4.43

Subsequent protonation and electron transfer lead to the spontaneous NN bond splitting and formation of the first equivalent of NH3 together with the nitrido intermediate. It is proposed by Nishibayashi and coworkers, based on the very small BDE (bond dissociation energy) of the Mo–nitrido fragment and the bridging N2 in the dimer, that splitting of the dimer occurs at this point, and functionalization of monometallic nitrido leads to the second

4.4 Catalytic Reactions of Group 6 Transition Metal–Dinitrogen Complexes

equivalent of NH3 . This is consistent with the fact that isolated monometallic nitrido complexes are as efficient as the Mo dinitrogen dimer [151]. On the other hand, Batista and coworkers propose that the integrity of the dimer is retained till the end of functionalization [153]. Again, protonation at N in the dimeric form is strongly kinetically favored over the monomeric form. These two approaches differ in the fact that the counterion was taken into account by Nishibayashi and coworkers but not by Batista. In fact, both monomeric [(PNP)Mo(N)(OTf )] and dimeric [(PNP)Mo(N)(OTf )](μ-N2 )[(PNP)Mo] complexes have been observed simultaneously by mass spectrometry in a stoichiometric reaction and [(PNP)Mo(NH3 )](μ-N2 )[(PNP)Mo(N2 )2 ] or [(PNP)Mo (NH3 )(N2 )](μ-N2 )[(PNP)Mo(N2 )] in a catalytic reaction, which are consistent with the easy Mo–N2 splitting and formation in the reaction conditions. Overall, a catalytic scheme consistent with the experiments can be proposed (Scheme 4.44). N

H H

N Mo

H N

N P

P = PtBu2

N Mo

N N

P TfO

N2

e NH3

2H

3H

e

3e N

N N

P TfO

N N Mo P

P N2 Mo N

P

N2

P

+ OTf

N N N2

H NH3

N

P TfO H

H

H N N

P

N

N Mo e

P N2 Mo N

P

P

P N2

N Mo

N2

N2

N2 Mo N

N2 Mo N

N N

P

P

P

P

N Mo P TfO

H P N2 Mo N

P N N N2

P

H P N2 Mo N

P N N N2

P

Scheme 4.44

Subsequent electronic modulation at the pyridine ring allowed fine-tuning of the catalysis. Thus, electron-donating substituents at the para position of the pyridine increased the yield of NH3 up to 34 equiv (for R = OMe vs. 23 for H). Larger amounts of reductant/H+ source yielded 52 equiv of NH3 per catalyst (or 26 per Mo center) [55]. Interestingly, time profiles for the formation of NH3 for different catalysts show that the complex [(PtBu NOMe PtBu )Mo(N2 )2 ]2 (μ-N2 ) is slower than the [(PtBu NH/Me PtBu )Mo(N2 )2 ]2 (μ-N2 ) complexes, but apparently more robust under catalytic conditions. Indeed, for the latter complexes, the maximum TON is obtained within one hour, whereas it requires 20 hours for the

257

258

4 Group 6 Transition Metal–Dinitrogen Complexes

former. It was hypothesized that redox noninnocent moieties, such as ferrocene or ruthenocene, would favor at the same time during the important protonation step but also the subsequent reduction steps. Accordingly, the ferrocene substituent was shown to be the fastest catalyst (rate for the first hour) as well as slightly more efficient than the complex featuring the PtBu NOMe PtBu ligand (37 equiv of NH3 vs 34: conditions: N2 , 216 equiv of Cp2 Co, and 288 equiv of LutHOTf in toluene for 20 hours) [154]. Modification of the central moiety of the ligand to phosphine and carbene ligand provided even more efficient catalysts. These ligands possess lower Brønsted character compared to the pyridine moiety, thereby preventing a possible decomposition pathway. The Mo(0) dinitrogen dimer was not isolated when using the PtBu PPh PtBu tridentate phosphine ligand because of a weak Mo—N2 (—Mo) bond, although it was clearly observed by IR. Alternatively, the Mo(V) cationic nitrido complex [(PtBu PPh PtBu )Mo(N)(Cl)]+ and the [(PtBu PPh PtBu )Mo(Cl)3 ] precursor were used. Thus, using the tandem Cp*2 Co as the reducing agent (36 equiv) and collidinium triflate as the proton source (48 equiv), 11 equiv of NH3 was obtained, which corresponds to a 92% yield (based on the reducing agent). A maximum of 63 equiv of NH3 was obtained, and the time profile of the reaction again pointed a fast reaction in the first two hours followed by a plateau indicating catalyst decomposition [56]. In 2017, the most efficient catalytic systems for NH3 synthesis from N2 to date were developed by Nishibayashi and coworkers (Scheme 4.45). In a first system, the central moiety of the tridentate ligand is a carbene center, and the improved activity (up to 115 equiv of NH3 ) is rationalized in terms of strong σ donation of the carbene as well as π acceptor capabilities [57]. The most efficient system to date was obtained with the known PNP tridentate ligand, in conjunction with iodide in the precatalyst: [Mo(PNP)(I)3 ]. The halide effect is impressive as 415 equiv of NH3 is formed in this case, whereas 12 equiv is measured with the [Mo(PNP)(Cl)3 ] precatalyst under optimized conditions (2880 [ColH]OTf/2160 CoCp*2 ) [155]. A similar effect, yet much less pronounced, was observed starting from the corresponding nitrido complexes [Mo(PNP)(N)(X)], X = Cl, Br, and I, with 8.4, 11.1, and 12.2 equiv of NH3 formed when 48 equiv of [ColH]OTf/36 CoCp*2 was used. Fc = ferrocenyl

Fc

P = PtBu2

N N P P N2 N Mo N N Mo N P N2 N2 P 2011 max TON Rate (h–1) in the first hour per Mo center

Scheme 4.45

N N P P N2 N P R Ph N Fc P Mo Mo N N Mo R N Cl N P P N2 N2 P R = H, Me 2015 2017 N

R R

I P N Mo I P I 2017

53

63

115

415

11.5

15.3

27

28

4.5 Chemistry of Cr Complexes

4.5 Chemistry of Cr Complexes By comparison with the chemistry of the heavier Mo and W analogs, the chemistry of Cr dinitrogen complexes has been much less studied. In fact, these complexes appeared to be less stable, and the synthesis of [Cr(PMe3 )4 (N2 )2 ] (decomposes at room temperature by N2 and PMe3 loss) [156] and [Cr(dppe)2 ]2 (μ-N2 ) [157] was only reported several years after the first Mo and W dinitrogen complexes. With a PDI ligand, Budzelaar and coworkers reported the synthesis of a bridged N2 complex via the one-electron reduction (using Na or NaH) of the CrI precursor under N2 (Scheme 4.46) [158]. The NN bond distance of 1.241(6) Å suggests a double reduction of N2 to the N2 2− fragment. Reduction/functionalization of this N2 fragment was then achieved via further addition of NaH. X-ray analysis revealed a formal proton transfer from one of the imine moieties to the bridging N. Addition of an excess of NaH resulted in the N—N bond breaking and the formation of the anionic imido complex. The final treatment of this imido complex with HCl releases NH4 + . The group of Mock has recently focused on Cr dinitrogen complexes stabilized by eight-membered ring diphosphine as well as 12- and 16-membered P macrocycles. With the diphosphine PPh 2 NBn 2 , both the cis and trans isomers could be obtained, although in poor yields (ca 15%) (Scheme 4.47) [159]. They later showed that the macrocyclic diphosphine ligand rearranges in situ to the 12 (triphosphine)- and 16 (tetraphosphine)-membered macrocycles. The 16-membered ring allows the formation of trans-[Cr(N2 )2 (PPh 4 NBn 4 )] upon reduction of CrCl2 by Mg under N2 atmosphere in a good yield (63%) [160]. The trans geometry is evidenced by the strong and weak 𝜈 NN bands at 1918 and at 2072 cm−1 , respectively, as well as X-ray diffraction. Two different Cr–N (1.930(2) and 1.884(2) Å) and N—N (1.112(3) and 1.120(3) Å) bonds Ar

Ar

N N

2

N

Cr

H

xs NaH

Cl

N

N2

N

Cr N

Ar

H

NH4

N iPr Na

THF

iPr

N2 2 NaH

NaH 3 NaH

Ar

Ar

N

Cr

THF N N Cr

N N

2 NaH

N

Cr

THF N

N Ar

Scheme 4.46

Ar

Ar

Ar N

N

N

H N

THF

Na N THF Na

Ar

N

Cr N Ar

N

259

260

4 Group 6 Transition Metal–Dinitrogen Complexes Bn Bn

Bn

N

N N2

Ph P N Bn

Cr

P Ph

N2

Bn

Bn N

N

Ph P

KC8, N2

P Ph

PPh2NBn2

N

Ph P N Bn

Bn

P Ph

Cl

Cr

Ph P

Mg, N2

Cl PPh2NBn2

P Ph

N Bn

Cl

trans

N Ph P Cr

N

Bn

P Ph N2

N2 cis

Scheme 4.47 N Ph Bn

N Ph P

Cl Bn Ph N P Ph Cr

P N Bn Cl

N Bn Ph N P Ph Cr

Ph N

Bn

Mg, N2

Bn

N Ph P P

P

N Bn N N

N

Bn

P

8 equiv –50 °C

HOTf

N H Bn

Ph

N Ph P P

N Bn Ph N P Ph Cr N Bn N N

H N

Bn

P

+ NH2–NH3 and NH4 + ox. Cr complexes

Scheme 4.48

are measured because of contrasting steric environment above and below the equatorial plane. Protonation at room temperature resulted in decomposition, whereas at −50 ∘ C, protonation at N moieties of the ligand is observed initially, followed by transfer to the β N of N2 . Finally, evolution of hydrazinium N2 H5 + is observed as the major N-containing species rather than NH4 + (traces) (Scheme 4.48). In 2015, using the 12-membered P macrocycle, they reported the reduction of fac-[CrCl3 (PPh 3 NBn 3 )] by Mg, under N2 atmosphere in the presence of dmpe [161]. The Cr(0) dinitrogen complex again features a weakly activated N2 (N–N of 1.132(3) Å and 𝜈 NN at 1918 cm−1 ). Reaction with HBArF at −78 ∘ C resulted in immediate and quantitative protonation at Cr, in accordance with the electron richness of the metal center imparted by the pentaphosphorus coordination. Addition of an excess of acid only resulted in the formation of traces of NH4 + in addition to free N2 . In 2016, they studied a family of group 6 complexes. If the analogous bi-dinitrogen complexes could be obtained for Cr, Mo, and W, only

4.6 Conclusion and Perspectives

Et2 P

F N F

Et2P M

P Et2

M = Cr

F PEt2 N

N

F

N

100 HOTf

NH4

: 0.08 equiv

–40 °C N

N

N2H5 : 0.22 equiv Not detected for Mo, W

Scheme 4.49

the protonation of the former allowed observation of both NH4 + (0.22 equiv) and N2 H5 + (0.08 equiv) (Scheme 4.49) [30]. Finally, in 2018, more than 45 years after the seminal work of Shiina, they showed that several of the complexes they synthesized, mentioned above, are catalysts of the N2 -to-N(SiMe3 )3 transformation. In the optimized conditions (100 equiv of Na and SiMe3 Cl, 1 atm N2 , 23 ∘ C, 16 hours, THF), between 4.8 and 6.8 equiv of N(SiMe3 )3 were obtained for the bidentate and tridentate ligands. Tetraphosphine, featuring the 16-membered macrocycle, provided the best results (10.6 equiv). However, more importantly, with 10 000 equiv of Na and chlorosilane, the TON of silylamine reached 21.1 and 34.1 after reloading, proving that catalytically active species are still present after the first 16 hours [162]. This ligand is therefore so far unique in the Cr chemistry that it reduces ligand lability and allows the catalytically active species to remain molecular.

4.6 Conclusion and Perspectives The seminal works of Hidai and Chatt and coworkers have shown that stoichiometric functionalization at nitrogen of M–N2 complexes could be achieved under mild conditions. They led the way to the development of the only known catalytic reductions of N2 : chronologically, the N2 -to-N(SiMe3 )3 and the N2 -to-NH3 processes. Several group 6 metal–dinitrogen complexes are efficient catalysts for the N2 -to-N(SiMe3 )3 process. The most efficient ones involve polydentate phosphine ligands, either two bidentate phosphines or a tridentate phosphine at Mo centers, or a macrocyclic tetradentate phosphine at Cr. In these best cases, the reaction is slow at room temperature (typically requiring days), which is linked to the kinetics of generation and reactivity of the Si radicals. One of the crucial features of the complexes is therefore to be stable for extended periods under strongly reducing media, as well as react fast with the silyl radical to limitate the competitive pathways: radical dimerization and radical attack of the solvent. One of the advantages of this transformation is that it relies on the use of simple, yet very strong, reducing agents (Na, K, …) and that the final silylamines can be transformed readily into NH3 upon hydrolysis. For the N2 -to-NH3 -catalyzed process, the mechanistic investigation performed by Schrock and coworkers on their tetradentate trianionic “NN3 ” ligand/ Mo system has led to the use of bulky neutral tridentate ligands/Mo systems. Impressive TONs of NH3 (115 and 415) have been obtained in 2017 with a

261

262

4 Group 6 Transition Metal–Dinitrogen Complexes

phosphine–carbene–phosphine and a phosphine–pyridine–phosphine ligand by Nishibayashi and coworkers. In both cases, the reaction is typically fast at room temperature, being essentially over within two hours. The major drawback of the first system is the use of stoichiometric Cr derivative as the reducing agent. In the second system, a dramatic halide effect was noticed, with the activity of the [Mo(PNP)(I)3 ] precatalyst being 35 times higher than the [Mo(PNP)(N2 )2 ](μ-N2 ) precursor. It is to be noted that the corresponding nitrido derivative [Mo(PNP)(N)(I)] is as efficient as the [Mo(PNP)(I)3 ] precatalyst showing its intermediacy in the catalytic cycle. Although there have been breakthroughs in the catalytic functionalization of N2 , considerable efforts have still to be made to find conditions to improve the efficiency of the processes (the nature of the reducing agent and the functionalizing moieties), in order to render them applicable in the future (Scheme 4.50). - Fast kinetics

xH xe

N

2 NH3 + (x–6)/2 H2

N

Further improvements: - Milder H+ source - Milder reducing agent - Decrease H2 formation - More robust catalysts

Catalytic processes

Mo(L)n x SiR3∙

- Slow kinetics 2 N(SiR3)3 + (x–6)/2 R3Si–SiR3

Further improvements: - Milder reducing agent - Decrease Si–Si formation - More robust catalysts

Scheme 4.50

Another approach to N2 functionalization relied on initial splitting of N2 between two metal centers (Mo(III), Mo(I), and Re(II)). This requires a total of six electrons to yield the M–nitrido complex, and thus three electrons per M center. Subsequent functionalization by strong electrophiles and Si—H bonds allowed the formation of “N1 ” compounds, nitriles, and silylamine, respectively. So far, these transformations (splitting of N2 + functionalization) are not catalytic. Rendering the N2 to amine “NA3 ” (using A–B substrate) transformation catalytic, following this approach is far from being trivial, as it involves an odd number of electrons: three per M center for the splitting of N2 , followed by successive two-electron transfers per N—A bond formation. These two electrons come from the A—B bond. Thus, an external source of electrons will have to be used to close the cycle, to formally eliminate three equivalents of B− . A hypothetical cycle is given in Scheme 4.51. Alternatively, A–B could be chosen so as to favor B–B elimination, in which case, only one external electron is needed to eliminate B− . Obviously, a lot remains to be discovered to render this process catalytic.

References

N N

Mo(L)n

Mo(L)n

N

“N1” compounds

2 Mo(L) n

: RCN, N(SiR3)2H

- Non catalytic - Stepwise functionalization of N Further improvements: - Catalytic A

M N 1/2

N

N2 splitting

N

M

3 e process

M

N

A–B

M

2 e process

B

per M

2 e process A–B

3e

N2

–3 B B

M

B

A

A

A

B

A–B B

+ N A

A

A

M B

B

A

A

N

2 e process

N M

B

B

Scheme 4.51

As a final conclusion, it can be said that major developments, especially in the catalytic reduction of N2 into either NH3 or N(SiMe3 )3 , have been achieved in the current decade. They have been possible because of the precise understanding of the reactivity of coordinated N2 provided by earlier mechanistic works. Continual research is expected to result in the development of even more efficient catalysts as well as new transformations of N2 in the coordination of metal centers.

References 1 2 3 4 5 6 7 8

Chatt, J., Dilworth, J.R., and Richards, R.L. (1978). Chem. Rev. 78 (6): 589. Hidai, M. and Mizobe, Y. (1995). Chem. Rev. 95 (4): 1115. Hidai, M. and Ishii, Y. (1996). J. Mol. Catal. A: Chem. 107 (1–3): 105. MacKay, B.A. and Fryzuk, M.D. (2004). Chem. Rev. 104 (2): 385. Hidai, M. and Mizobe, Y. (2005). Can. J. Chem. 83 (4): 358. Shilov, A.E. (2003). Russ. Chem. Bull. 52 (12): 2555. Dreher, A., Stephan, G., and Tuczek, F. (2009). Adv. Inorg. Chem. 61: 367. Tanabe, Y. and Nishibayashi, Y. (2013). Coord. Chem. Rev. 257 (17–18): 2551.

263

264

4 Group 6 Transition Metal–Dinitrogen Complexes

9 Broda, H., Hinrichsen, S., and Tuczek, F. (2013). Coord. Chem. Rev. 257 (2):

587. 10 van der Ham, C.J.M., Koper, M.T.M., and Hetterscheid, D.G.H. (2014).

Chem. Soc. Rev. 43 (15): 5183. 11 Jia, H.-P. and Quadrelli, E.A. (2014). Chem. Soc. Rev. 43 (2): 547. 12 Khoenkhoen, N., de Bruin, B., Reek, J.N.H., and Dzik, W.I. (2015). Eur. J.

Inorg. Chem. 4: 567. 13 Tanaka, H., Nishibayashi, Y., and Yoshizawa, K. (2016). Acc. Chem. Res. 49

(5): 987. 14 Tanabe, Y. and Nishibayashi, Y. (2016). Chem. Rec. 16 (3): 1549. 15 Burford, R.J. and Fryzuk, M.D. (2017). Nat. Rev. Chem. 1 (4): 0026. 16 Bhattacharya, P., Prokopchuk, D.E., and Mock, M.T. (2017). Coord. Chem.

Rev. 334: 67. 17 Hidai, M., Tominari, K., Uchida, Y., and Misono, A. (1969). J. Chem. Soc. D 5

(14): 814. 18 Green, M.L.H. and Silverthorn, W.E. (1973). J. Chem. Soc., Dalton Trans. 2

(3): 301. 19 Green, M.L.H. and Silverthorn, W.E. (1974). J. Chem. Soc., Dalton Trans. 3

(20): 2164. 20 Hidai, M., Tominari, K., and Uchida, Y. (1972). J. Am. Chem. Soc. 94 (1): 110. 21 Hidai, M., Tominari, K., Uchida, Y., and Misono, A. (1969). J. Chem. Soc. D 5

(23): 1392. 22 Atkinson, L.K., Mawby, A.H., and Smith, D.C. (1971). J. Chem. Soc. D 2 (3):

157. 23 George, T.A. and Seibold, C.D. (1973). Inorg. Chem. 12 (11): 2548. 24 Bell, B., Chatt, J., and Leigh, G.J. (1972). J. Chem. Soc., Dalton Trans. 1 (22):

2492. 25 Chatt, J., Heath, G.A., and Richards, R.L. (1974). J. Chem. Soc., Dalton Trans.

3 (19): 2074. 26 Galindo, A., Gutiérrez, E., Monge, A. et al. (1995). J. Chem. Soc., Dalton

Trans. 24 (23): 3801. 27 Carmona, E., Galindo, A., Poveda, M.L., and Rogers, R.D. (1985). Inorg.

Chem. 24 (24): 4033. 28 Chatt, J., Pearman, A.J., and Richards, R.L. (1977). J. Chem. Soc., Dalton

Trans. 6 (21): 2139. 29 Anderson, S.N., Richards, R.L., and Hughes, D.L. (1986). J. Chem. Soc.,

Dalton Trans. (2): 245. 30 Egbert, J.D., O’Hagan, M., Wiedner, E.S. et al. (2016). Chem. Commun. 52

(60): 9343. 31 Kubas, G.J., Burns, C.J., Eckert, J. et al. (1993). J. Am. Chem. Soc. 115 (2):

569. 32 George, T.A. and Kovar, R.A. (1981). Inorg. Chem. 20 (1): 285. 33 Yuki, M., Miyake, Y., Nishibayashi, Y. et al. (2008). Organometallics 27 (15):

3947. 34 Yuki, M., Midorikawa, T., Miyake, Y., and Nishibayashi, Y. (2009).

Organometallics 28 (16): 4741.

References

35 Yuki, M., Miyake, Y., and Nishibayashi, Y. (2009). Organometallics 28 (19):

5821. 36 Labois, L.A., Weiss, C.J., Egbert, J.D. et al. (2015). Z. Anorg. Allg. Chem. 641

(1): 105. 37 Weiss, C.J., Groves, A.N., Mock, M.T. et al. (2012). Dalton Trans. 41 (15): 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62

4517. George, T.A. and Tisdale, R.C. (1988). Inorg. Chem. 27 (17): 2909. George, T.A., Rose, D.J., Chang, Y. et al. (1995). Inorg. Chem. 34 (5): 1295. George, T.A. and Tisdale, R.C. (1985). J. Am. Chem. Soc. 107 (18): 5157. George, T.A., Ma, L., Shailh, S.N. et al. (1990). Inorg. Chem. 29 (23): 4789. George, T.A., Jackson, M.A., and Kaul, B.B. (1991). Polyhedron 10 (4–5): 467. George, T.A. and Jackson, M.A. (1988). Inorg. Chem. 27 (5): 924. Hinrichsen, S., Kindjajev, A., Adomeit, S. et al. (2016). Inorg. Chem. 55 (17): 8712. Sivasankar, C., Böres, N., Peters, G. et al. (2005). Organometallics 24 (22): 5393. Broda, H., Hinrichsen, S., Krahmer, J. et al. (2014). Dalton Trans. 43 (5): 2007. Stephan, G.C., Peters, G., Lehnert, N. et al. (2005). Can. J. Chem. 83 (4): 385. Römer, R., Stephan, G., Habeck, C. et al. (2008). Eur. J. Inorg. Chem. (21): 3258. Römer, R., Gradert, C., Bannwarth, A. et al. (2011). Dalton Trans. 40 (13): 3229. Hinrichsen, S., Schnoor, A.-C., Grund, K. et al. (2016). Dalton Trans. 45 (37): 14801. Hanna, B.S., MacIntosh, A.D., Ahn, S. et al. (2014). Organometallics 33 (13): 3425. Söncksen, L., Gradert, C., Krahmer, J. et al. (2013). Inorg. Chem. 52 (11): 6576. Dahlenburg, L. and Pietsch, B. (1989). Chem. Ber. 122 (11): 2085. Arashiba, K., Miyake, Y., and Nishibayashi, Y. (2011). Nat. Chem. 3 (2): 120. Kuriyama, S., Arashiba, K., Nakajima, K. et al. (2014). J. Am. Chem. Soc. 136 (27): 9719. Arashiba, K., Kinoshita, E., Kuriyama, S. et al. (2015). J. Am. Chem. Soc. 137 (17): 5666. Eizawa, A., Arashiba, K., Tanaka, H. et al. (2017). Nat. Commun. 8: 14874. Liao, Q., Saffon-Merceron, N., and Mézailles, N. (2015). ACS Catal. 5 (11): 6902. Morris, R.H., Ressner, J.M., Sawyer, J.F., and Shiralian, M. (1984). J. Am. Chem. Soc. 106 (12): 3683. Yoshida, T., Adachi, T., Kaminaka, M. et al. (1988). J. Am. Chem. Soc. 110 (14): 4872. Yoshida, T., Adachi, T., Ueda, T. et al. (1989). Angew. Chem. Int. Ed. Engl. 28 (8): 1040. Kol, M., Schrock, R.R., Kempe, R., and Davis, W.M. (1994). J. Am. Chem. Soc. 116 (10): 4382.

265

266

4 Group 6 Transition Metal–Dinitrogen Complexes

63 Yandulov, D.V., Schrock, R.R., Rheingold, A.L. et al. (2003). Inorg. Chem. 42

(3): 796. 64 Yandulov, D.V. and Schrock, R.R. (2002). J. Am. Chem. Soc. 124 (22): 6252. 65 Weare, W.W., Dai, X., Byrnes, M.J. et al. Proc. Natl. Acad. Sci. U.S.A. 103

(46): 17099. 66 Ritleng, V., Yandulov, D.V., Weare, W.W. et al. (2004). J. Am. Chem. Soc. 126

(19): 6150. 67 Margulieux, G.W., Turner, Z.R., and Chirik, P.J. (2014). Angew. Chem. Int. 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88

Ed. 53 (51): 14211. Baker, R.T., Morton, J.R., Preston, K.F. et al. (1991). Inorg. Chem. 30 (1): 113. Miyazaki, T., Tanabe, Y., Yuki, M. et al. (2013). Chem. Eur. J. 19 (36): 11874. Mercer, M. (1974). J. Chem. Soc., Dalton Trans. 3 (16): 1637. Cradwick, P.D., Chatt, J., Crabtree, R.H., and Richards, R.L. (1975). J. Chem. Soc., Chem. Commun. (9): 351. Ishino, H., Nagano, T., Kuwata, S. et al. (2001). Organometallics 20 (18): 188. Mizobe, Y., Yokobayashi, Y., Oshita, H. et al. (1994). Organometallics 13 (21): 3764. O’Regan, M.B., Liu, A.H., Finch, W.C. et al. (1990). J. Am. Chem. Soc. 112 (11): 4331. Schrock, R.R., Kolodziej, R.M., Liu, A.H. et al. (1990). J. Am. Chem. Soc. 112 (11): 4338. O’Donoghue, M.B., Zanetti, N.C., Davis, W.M., and Schrock, R.R. (1997). J. Am. Chem. Soc. 119 (11): 2753. O’Donoghue, M.B., Davis, W.M., Schrock, R.R., and Reiff, W.M. (1999). Inorg. Chem. 38 (2): 243. Fontaine, P.P., Yonke, B.L., Zavalij, P.Y., and Sita, L.R. (2010). J. Am. Chem. Soc. 132 (35): 12273. Laplaza, C.E. and Cummins, C.C. (1995). Science 268 (5212): 861. Laplaza, C.E., Johnson, M.J.A., Peters, J.C. et al. (1996). J. Am. Chem. Soc. 118 (36): 8623. Peters, J.C., Cherry, J.-P.F., Thomas, J.C. et al. (1999). J. Am. Chem. Soc. 121 (43): 10053. Solari, E., Da Silva, C., Iacono, B. et al. (2001). Angew. Chem. Int. Ed. 40: 3907. Hebden, T.J., Schrock, R.R., Takase, M.K., and Müller, P. (2012). Chem. Commun. 48 (13): 1851. Keane, A.J., Farrell, W.S., Yonke, B.L. et al. (2015). Angew. Chem. Int. Ed. 54 (35): 10220. Duman, L.M., Farrell, W.S., Zavalij, P.Y., and Sita, L.R. (2016). J. Am. Chem. Soc. 138 (45): 14856. Liao, Q., Cavaillé, A., Saffon-Merceron, N., and Mézailles, N. (2016). Angew. Chem. Int. Ed. 55 (37): 11212. Silantyev, G.A., Förster, M., Schluschaß, B. et al. (2017). Angew. Chem. Int. Ed. 56 (21): 5872. Chatt, J., Heath, G.A., and Leigh, G.J. (1972). J. Chem. Soc., Chem. Commun. 8 (8): 444.

References

89 Chatt, J., Diamantis, A.A., Heath, G.A. et al. (1977). J. Chem. Soc., Dalton

Trans. 6 (7): 688. 90 Chatt, J., Heath, G.A., and Richards, R.L. (1972). J. Chem. Soc., Chem. Com-

mun. 8 (18): 1010. 91 Heath, G.A., Mason, R., and Thomas, K.M. (1974). J. Am. Chem. Soc. 96 (1):

259. 92 Hidai, M., Kodama, T., Sato, M. et al. (1976). Inorg. Chem. 15 (11): 2694. 93 Chatt, J., Pearman, A.J., and Richards, R.L. (1975). Nature 253 (5486): 39. 94 Chatt, J., Pearman, A.J., and Richards, R.L. (1977). J. Chem. Soc., Dalton

Trans. 6 (19): 1852. 95 Galindo, A., Hills, A., Hughes, D.L. et al. (1990). J. Chem. Soc., Dalton Trans.

19 (1): 283. 96 Chatt, J., Pearman, A.J., and Richards, R.L. (1976). J. Chem. Soc., Dalton

Trans. 5 (15): 1520. 97 Chatt, J., Pearman, A.J., and Richards, R.L. (1978). J. Chem. Soc., Dalton

Trans. 7 (12): 1766. 98 Kaul, B.B., Hayes, R.K., and George, T.A. (1990). J. Am. Chem. Soc. 112 (5):

2002. 99 Anderson, S.N., Fakley, M.E., Richards, R.L., and Chatt, J. (1981). J. Chem.

Soc., Dalton Trans. 10 (9): 1973. 100 Chatt, J., Fakley, M.E., Hitchcock, P.B. et al. (1982). J. Chem. Soc., Dalton

Trans. 11 (2): 345. 101 Hinrichsen, S., Broda, H., Gradert, C. et al. (2012). Annu. Rep. Prog. Chem.

Sect. A: Inorg. Chem. 108: 17. 102 Labios, L.A., Heiden, Z.M., and Mock, M.T. (2015). Inorg. Chem. 54 (9):

4409. 103 Pickett, C.J. and Talarmin, J. (1985). Nature 317 (6038): 652. 104 Nishihara, H., Mori, T., Tsurita, Y. et al. (1982). J. Am. Chem. Soc. 104 (16):

4367. 105 Hidai, M., Takahashi, T., Yokotake, I., and Uchida, Y. (1980). Chem. Lett. 9

(6): 645. 106 Jia, G., Morris, R.H., and Schweitzer, C.T. (1991). Inorg. Chem. 30 (4): 593. 107 Nishibayashi, Y., Iwai, S., and Hidai, M. (1998). Science 279 (5350): 540. 108 Nishibayashi, Y., Iwai, S., and Hidai, M. (1998). J. Am. Chem. Soc. 120 (40):

10559. 109 Nishibayashi, Y., Wakiji, I., Hirata, K. et al. (2001). Inorg. Chem. 40 (3): 578. 110 Tamizmani, M. and Sivasankar, C. (2017). Eur. J. Inorg. Chem. (37): 4239. 111 Bossard, G.E., Busby, D.C., Chang, M. et al. (1980). J. Am. Chem. Soc. 102

(3): 1001. 112 Hussain, W., Leigh, G.J., Modh-Ali, H., and Pickett, C.J. (1986). J. Chem. Soc.,

Dalton Trans. 15 (7): 1473. 113 Abu Bakar, M., Hughes, D.L., and Leigh, G.J. (1988). J. Chem. Soc., Dalton

Trans. 17 (10): 2525. 114 Ben-Shoshan, R., Chatt, J., Leigh, G.J., and Hussain, W. (1980). J. Chem. Soc.,

Dalton Trans. 9 (5): 771. 115 Pickett, C.J. and Leigh, G.J. (1981). J. Chem. Soc., Chem. Commun. (3): 1033.

267

268

4 Group 6 Transition Metal–Dinitrogen Complexes

116 Aoshima, T., Tamura, T., Mizobe, Y., and Hidai, M. (1992). J. Organomet.

Chem. 435 (1–2): 85. 117 Harada, Y., Mizobe, Y., Ishii, Y., and Hidai, M. (1998). Bull. Chem. Soc. Jpn.

71 (11): 2701. 118 Seino, H., Ishii, Y., and Hidai, M. (1997). Inorg. Chem. 36 (2): 161. 119 Seino, H., Ishii, Y., Sasagawa, T., and Hidai, M. (1995). J. Am. Chem. Soc. 117

(49): 12181. 120 Curley, J.J., Sceats, E.L., and Cummins, C.C. (2006). J. Am. Chem. Soc. 128

(43): 14036. 121 Hidai, M., Komori, K., Kodama, T. et al. (1984). J. Organomet. Chem. 272

(2): 155. 122 Komori, K., Oshita, H., Mizobe, Y., and Hidai, M. (1989). J. Am. Chem. Soc.

111 (5): 1939. 123 Oshita, H., Mizobe, Y., and Hidai, M. (1993). J. Organomet. Chem. 456 (2):

213. 124 Chatt, J., Crabtree, R.H., Jeffery, E.A., and Richards, R.L. (1973). J. Chem.

Soc., Dalton Trans. 2 (11): 1167. 125 Takahashi, T., Kodama, T., Watakabe, A. et al. (1983). J. Am. Chem. Soc. 105

(6): 1680. 126 Takagahara, K., Ishino, H., Ishii, Y., and Hidai, M. (1998). Chem. Lett. 27 (9):

897. 127 Tanaka, H., Sasada, A., Kouno, T. et al. (2011). J. Am. Chem. Soc. 133 (10):

3498. 128 Liao, Q., Saffon-Merceron, N., and Mézailles, N. (2014). Angew. Chem. Int.

Ed. 53 (51): 14206. 129 Duman, L.M. and Sita, L.R. (2017). J. Am. Chem. Soc. 139 (48): 17241. 130 Ishino, H., Ishii, Y., and Hidai, M. (1998). Chem. Lett. 27 (7): 677. 131 Simonneau, A., Turrel, R., Vendier, L., and Etienne, M. (2017). Angew. Chem. 132 133 134 135 136 137 138 139 140 141 142 143 144

Int. Ed. 56 (40): 12268. Haight, G.P. Jr. and Scott, R. (1964). J. Am. Chem. Soc. 86 (4): 743. Vol’pin, M.E. and Shur, V.B. (1966). Nature 209 (5029): 1236. Bazhenova, T.A. and Shilov, A.E. (1995). Coord. Chem. Rev. 144: 69. Shiina, K. (1972). J. Am. Chem. Soc. 94 (26): 9266. Ogawa, T., Kajita, Y., Wasada-Tsutsui, Y. et al. (2013). Inorg. Chem. 52 (1): 182. Yandulov, D.V. and Schrock, R.R. (2003). Science 301 (5629): 76. Yandulov, D.V. and Schrock, R.R. (2005). Inorg. Chem. 44 (4): 1103. Hetterscheid, D.G.H., Hanna, B.S., and Schrock, R.R. (2009). Inorg. Chem. 48 (17): 8569. Kinney, R.A., McNaughton, R.L., Chin, J.M. et al. (2011). Inorg. Chem. 50 (2): 418. Schrock, R.R. (2008). Angew. Chem. Int. Ed. 47 (30): 5512. Schenk, S., Le Guennic, B., Kirchner, B., and Reiher, M. (2008). Inorg. Chem. 47 (9): 3634. Schenk, S., Kirchner, B., and Reiher, M. (2009). Chem. Eur. J. 15 (20): 5073. Thimm, W., Gradert, C., Broda, H. et al. (2015). Inorg. Chem. 54 (19): 9248.

References

145 Weare, W.W., Schrock, R.R., Hock, A.S., and Müller, P. (2006). Inorg. Chem.

45 (23): 9185. 146 Reithofer, M.R., Schrock, R.R., and Müller, P. (2010). J. Am. Chem. Soc. 132

(24): 8349. 147 Arashiba, K., Kuriyama, S., Nakajima, K., and Nishibayashi, Y. (2013). Chem.

Commun. 49 (95): 11215. 148 Arashiba, K., Sasaki, K., Kuriyama, S. et al. (2012). Organometallics 31 (5):

2035. 149 Kinoshita, E., Arashiba, K., Kuriyama, S. et al. (2012). Organometallics 31

(23): 8437. 150 Tanabe, Y., Kuriyama, S., Arashiba, K. et al. (2013). Chem. Commun. 49 (81):

9290. 151 Tanaka, H., Arashiba, K., Kuriyama, S. et al. (2014). Nat. Commun. 5: 3737. 152 Wickramasinghe, L.A., Ogawa, T., Schrock, R.R., and Müller, P. (2017). J.

Am. Chem. Soc. 139 (27): 9132. 153 Tian, Y.-H., Pierpont, A.W., and Batista, E.R. (2014). Inorg. Chem. 53 (8):

4177. 154 Kuriyama, S., Arashiba, K., Nakajima, K. et al. (2015). Chem. Sci. 6 (7): 3940. 155 Arashiba, K., Eizawa, A., Tanaka, H. et al. (2017). Bull. Chem. Soc. Jpn. 90

(10): 1111. 156 Karsch, H.H. (1977). Angew. Chem. Int. Ed. Engl. 16 (1): 56. 157 Sobota, P. and Jezowska-Trzebiatowska, ̇ B. (1977). J. Organomet. Chem. 131

(3): 341. 158 Vidyaratne, I., Scott, J., Gambarotta, S., and Budzelaar, P.H.M. (2007). Inorg.

Chem. 46 (17): 7040. 159 Mock, M.T., Chen, S., Rousseau, R. et al. (2011). Chem. Commun. 47 (44):

12212. 160 Mock, M.T., Chen, S., O’Hagan, M. et al. (2013). J. Am. Chem. Soc. 135 (31):

11493. 161 Mock, M.T., Pierpont, A.W., Egbert, J.D. et al. (2015). Inorg. Chem. 54 (10):

4827. 162 Kendall, A.J., Johnson, S.I., Bullock, R.M., and Mock, M.T. (2018). J. Am.

Chem. Soc. 140 (7): 2528.

269

271

5 Toward N—N Bond Cleavage: Synthesis and Reactivity of Group 7 Dinitrogen Complexes Elon A. Ison North Carolina State University, Department of Chemistry, 2620 Yarbrough Drive, Raleigh, NC 27695-8204, USA

Compared to groups 6 and 8, relatively few dinitrogen complexes of group 7 metals have been reported [1]. Further, the reactivity of these metal complexes has only been demonstrated in a few cases. Nonetheless, the reactivity of group 7 dinitrogen complexes is of high interests, given that many elements in this group have access to a variety of stable oxidation states and thus have the potential to provide the six electrons necessary to reduce N2 [2]. In recent years, the chemistry of group 7 complexes toward N2 activation and functionalization has seen renewed interests, and there has been one example to date of a rhenium complex that has been capable of reducing dinitrogen [3]. In this chapter, we review a variety of methods for the synthesis of group 7 dinitrogen complexes. The reactivity of these complexes is then discussed. Finally, we discuss efforts to exploit the reactivity toward cleavage and functionalization of the N2 bond.

5.1 Synthesis of Group VII N2 Complexes 5.1.1

Syntheses of Terminal N2 Complexes

By far, the most commonly encountered group 7 dinitrogen complexes involve terminal N2 ligands. The majority of these molecules incorporate rhenium, with sporadic examples of complexes with manganese and to a lesser extent technetium. As detailed below, in the majority of these complexes, the N2 ligand is weakly activated with the N2 stretching vibration of ∼1900–2100 cm−1 and N—N bonds of ∼1.1–1.2 Å [4]. Perhaps the most common synthetic strategy involves the reduction of the transition metal complex with a strong reducing agent. One of the earliest examples of this strategy involves the synthesis of rhenium complexes of the form ClRe(N2 )(PMe2 Ph)4 , 1 (Figure 5.1) [5]. This complex has been structurally characterized, and an N—N bond of 1.06 Å was observed. The N2 stretching vibration was observed at 1922 cm−1 . Similar complexes, ClRe(N2 )(PEt2 Ph)4 , 2, and HRe(N2 )(PEt2 Ph)4 , 3, were also synthesized by this method [6]. Complex 3 Transition Metal-Dinitrogen Complexes: Preparation and Reactivity, First Edition. Edited by Yoshiaki Nishibayashi. © 2019 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2019 by Wiley-VCH Verlag GmbH & Co. KGaA.

272

5 Toward N—N Bond Cleavage: Synthesis and Reactivity of Group 7 Dinitrogen Complexes

CI

CI PhMe2P

N2 PMe2Ph Na/Hg PMe2Ph PhMe2P Re Re CI PMe2Ph THF/N2 PhMe2P PMe2Ph 1 CI

Figure 5.1 Syntheses of group 7 terminal dinitrogen complexes by reduction with strong reducing agents.

N2 PhEt2P Na/Hg

THF CI

CI

Re

CI

PEt2Ph N2

THF

PEt2Ph PEt2Ph 2

+

THF

CI

CI

PhEt2P

Re

N2 PhEt2P PhEt2P

Re

PEt2Ph PEt2Ph

H

3

N

Ph PMe3

Ph N CI Ph3P

Re CI

PMe3 PPh3 Na/Hg CI

THF/N2

H N2 Me3P

Re

PMe3 PMe3 4

was structurally characterized with this strategy with similar N—N bond lengths (1.018 Å) and N2 stretching frequencies (1945 cm−1 ). The imido complex, Re(NPh)(PPh3 )2 (Cl)2 , is reduced by sodium amalgam in the presence of PMe3 in tetrahydrofuran (THF) to produce 4, which displays IR stretches for the N—N bond at 2000 cm−1 [7]. A structural analysis of this compound revealed an N—N bond of 1.101 Å. Complexes incorporating cyclopentadienyl (Cp) [8, 9] and its derivatives were also synthesized from strong reducing agents (Figure 5.2). For example, 5 was synthesized by the reduction of the aryldiazenido complex, Cp*Re(PR3 )(N2 C6 H4 OMe) (R = Me, OMe), with sodium amalgam [10]. For these Cp* complexes, the N—N stretching frequency was observed by IR at 1975 and 2014 cm−1 for the methyl and methoxy derivatives, respectively. Complex 6, incorporating a chelating phosphine, as well as a Cp* ligand, is also synthesized by this method and displays an N–N stretch by IR of 1977 cm−1 [10]. More recently, the group of Mösch-Zanetti et al. described the synthesis of 7 (IR: 2010 cm−1 , X-ray: 1.087 Å), a rhenium complex incorporating triamidoamine ligands, from sodium amalgam [11]. Finally, this synthetic method was also used for the synthesis of a technetium N2 complex [12]. Reduction of Tc(Cl)4 (PPh3 )2 with sodium amalgam in benzene results in the formation of a rare technetium dinitrogen complex, 8. An X-ray crystal structure for this molecule was obtained with a reported N–N distance of 0.98 Å [13]. Another popular strategy for the synthesis of terminal group 7 dinitrogen complexes involves the degradation of chelated benzoyldiazenido complexes (Figure 5.3). This strategy was originally employed by Chatt et al. as well as Pombeiro and coworkers, for the synthesis of the rhenium dinitrogen complexes, 9 and 10, with ancillary phosphine, phosphite, or carbonyl ligands [4g, 5b, 14]. In later years, Harman and coworkers (Figure 5.4) have extended this strategy for

5.1 Synthesis of Group VII N2 Complexes

Figure 5.2 Syntheses of group 7 terminal dinitrogen complexes incorporating Cp ligands and their derivatives, by reduction with strong reducing agents.

Cp*

+

Cp*

Na/Hg Re

R3P

N2C6H4OMe

R3P

Re

R3P

THF

R3P

N2 5

R = Me, OMe +

Cp*

Cp* Na/Hg

Re

Me2P

PMe2

R

N2C6H4OMe

CI

R N

Me2P

THF

Re PMe2

N2

6

R

Re N

N

R R N

Na/Hg, N2

N

R = C6F5 N

N2

R

Re N N 7

Na/Hg PPh3 CI

Tc

CI

Ph Ph Ph H P P Tc P P N2 Ph Ph Ph Ph 8

N2

CI

dppe

CI

Benzene

Ph

PPh3 PPh2

dppe = Ph2P

Ph PPh3 N O N Re CI CI PPh3

N2

L

L CI

PPh3 NNCOPh Re CI

L L

Re

Ph3P CI or

CH3OH

PPh3

PPh3 L 9 L = CO, PF3, phosphites

N2 L′ L

Re Cl

L L 10 L = aryl phosphines

+ HCI + MeOCOPh (+ 2 PPh3)

Figure 5.3 Syntheses of group 7 terminal dinitrogen complexes by the degradation of chelated benzoyldiazenido ligands in the presence of phosphines, phosphites, and carbonyls.

the synthesis complexes, 11 and 12, with amines and polypyridines [15]. Complexes 9 and 10 exhibit N–N stretching vibrations in the range 1925–2090 cm−1 , whereas 11 and 12 exhibit vibrations in the range 1872–2033 cm−1 . A somewhat related strategy involves the degradation of aryldiazenido complexes in the presence of a nucleophile. This method was employed to generate the manganese dinitrogen complex bearing Cp′ (Cp′ = MeC5 H4 ), 13b (Figure 5.5). Complex 13a

273

274

5 Toward N—N Bond Cleavage: Synthesis and Reactivity of Group 7 Dinitrogen Complexes

L′ (en, bpy) NNCOPh NaOTf, MeOH

PPh3 L (etpb, PF3)

Ph Ph3P Re

Cl

Re

Cl

N N

O

L

Cl

PPh3

Cl

–L

L′ (dien) NaOTf, MeOH

– PPh3

Ph3P

L′ (ampy) NaOTf, MeOH [ReCl2(N2)(PPh3)(L)]

L (tbpy) – PPh3

N

N

N

N

tBu

bpy

NH2 tBu

bpy

NH2

12 + HCl + NaCl + MeOCOPh

Et

N

OO P

en

H N

[Re(N2)(PPh3)(L)(L′)]+

dien

NH2 NH2

H2N

[Re(N2)(PPh3)2(L′)]+ 11

O

etpb

ampy

Figure 5.4 Syntheses of group 7 terminal dinitrogen complexes by the degradation of chelated benzoyldiazenido ligands in the presence of amines and polypyridines.

Cp′

Cp′

+ KX

OC

Mn

OC

N2C6H4R

–XC6H4R

Mn OC N2 OC

Acetone

Figure 5.5 Syntheses of manganese terminal dinitrogen complexes by the degradation of chelated aryldiazenido ligands with nucleophiles.

13b

R = CF3, F, H X = halide Cp

Cp 2+

OC

Mn

OC

+ 2 H 2O 2 N2H4

Cu

THF, –40 °C Cp′ = MeC5H4

OC

Mn

OC

N2 13a

related to complex 13b was also synthesized by oxidation of a hydrazine ligand in CpMn(CO)2 (N2 H4 ) [9, 16]. Terminal N2 complexes have also been synthesized by the reduction of group 7 complexes in a higher oxidation state with relatively mild reducing agents or via reductive elimination. An example of the latter was shown by Jones and coworkers where the manganese complex, 14, was synthesized by reductive elimination of H2 from MnH3 (dmpe)2 under a nitrogen atmosphere [17]. Spectroscopic data included an IR stretch at 1947 cm−1 and a N—N bond length of 1.127 Å. Similarly, thialato rhenium complexes, 15, were reported by Hughes and coworkers that resulted from the reaction of rhenium hydrides with Htipt (2,4,6-trisopropylthiophenol) or Hdipbt (HSC6 H2 -2,6-Pri 2 ,4-Br) ligands under a

5.1 Synthesis of Group VII N2 Complexes

H N2

MnH3(dmpe)2

dmpe =

P

C6D6

Me2P

P

Mn

P

N2

P

14

PMe2

PPh3 L

HL H7Re(PPh3)2 or H5Re(PPh3)2 S

–

S

N2 Toluene

–

Re PPh3 15

L= Br (dipbt)

(tipt) (Et4N)ReH9

+ dppe

dppe =

L

L

Ph Ph Ph H P P Re P P N2 Ph Ph Ph Ph

Ph N2

2-Propanol, 25 °C Ph2P

16

PPh2

Figure 5.6 Additional syntheses of manganese terminal dinitrogen complexes. R

R N O CH3 Re N N R

Me2PhSiH Toluene, N2, 80 °C

R

N2 N

Re

N

N Re N2

N

R = C6F5 17

N N

R

R

Figure 5.7 Synthesis of a bimetallic dinitrogen rhenium complex.

nitrogen atmosphere (Figure 5.6) [18]. In the IR spectrum of 15, a strong band at 2130 cm−1 was assigned to the N—N bond for L = tipt. A bond length for this bond of 1.01 Å of was also reported by X-ray crystallography. Similarly, the treatment of (Et4 N)2 ReH9 with 1,2-bis(diphenyphosphino)ethane, under a nitrogen atmosphere, resulted in 16. The IR absorption band for 16 was observed at 2006 cm−1 [19]. A bimetallic complex 17, with terminal N2 ligands, was synthesized by the Ison group (Figure 5.7) by the reduction of a rhenium(V) precursor with Me2 PhSiH under a nitrogen atmosphere [20]. N–N stretches were observed at 2015 and 2031 cm−1 , while the N—N bond of 1.120 and 1.127 Å was reported. 5.1.2

Reactivity of Terminal N2 Complexes

Pioneering studies on the reactivity of rhenium(I) dinitrogen complexes were performed by Chatt et al. [14a, b, 21, 22]. It was demonstrated that complexes

275

276

5 Toward N—N Bond Cleavage: Synthesis and Reactivity of Group 7 Dinitrogen Complexes

9 and 10 could be oxidized further with oxidizing transition metal salts such as silver nitrate, iron (III) chloride, and copper (I) chloride to yield rhenium (II) dinitrogen complexes [14b, 21b]. As expected, π-backbonding to the N2 ligand in these complexes is attenuated, and N–N vibrations were observed at 2010–2065 cm−1 [14b]. Perhaps more relevant to N2 bond cleavage is the observation that rhenium (I) dinitrogen complexes are nucleophilic and will react with Lewis acids to form adducts [23]. As discussed below, the dinitrogen ligand in most of these cases is weakly activated, and thus far, none of these adducts have been shown to directly lead to N2 cleavage, although this type of adduct has been implied in one recent study (vide infra). 5.1.2.1

Synthesis of Bridged N2 Complexes by Reaction with Lewis Acids

As suggested above, Chatt et al. first demonstrated that bridged N2 group 7 complexes could be synthesized by exploiting the nucleophilic nature of terminal dinitrogen complexes [22, 24]. As shown in Figure 5.8, the complex trans-[ReCl(N2 )(PMe2 Ph)4 ] will react with Lewis acids to form adducts of the form 17 and 18 [22, 24a, 25]. Of note, Lewis acids such as BCl3 and AlCl3 , which do not possess d electrons, do not lead to adduct formation but instead lead to nitrogen evolution. Adduct formation generally leads to a decrease in the N–N stretching vibration with frequencies observed from 1630 to 1890 cm−1 [24a].

Cl Cl

THF Cl Mo Cl

Figure 5.8 Syntheses of bridging dinitrogen group 7 complexes by reactions with transition metal Lewis acids.

N2 +

PhMe2P PhMe2P

PMe2Ph

Re

PMe2Ph

CI

THF

PhMe2P PhMe2P Re CH2CI2 MeOH

Cl

N

Cl

N

CI

PMe2Ph PMe2Ph 17

Mo MeO

Cl Cl N2

THF THF CI

Cr

CI CI

PhMe2P

Re

PMe2Ph PMe2Ph

PhMe2P CI

THF

PhMe2P PhMe2P Re THF

N

CI

N

Cr THF

CI CI

18

CI

PMe2Ph PMe2Ph

+ THF

5.1 Synthesis of Group VII N2 Complexes

Figure 5.9 Syntheses of bridging dinitrogen group 7 complexes by reactions with iron porphyrin complexes.

Et

Et

OTf N2

N

Et N Et

Et

Fe

PMe2Ph PhMe2P Re PMe2Ph PhMe2P

+

N Et

N Et

CI +

Et Et

Et

N

N

Et N Et

PhMe2P PhMe2P Re

Fe

Et

Et

Et

Tol N N

Fe

19

N

N

Tol

PMe2Ph PMe2Ph

N Et

CI

N2

OTf +

N

PhMe2P PhMe2P

N

PMe2Ph

Re

PMe2Ph

CI Tol

Tol

+ PhMe2P PhMe2P Re

Tol N N

Tol N N

CI

PMe2Ph PMe2Ph 20

Fe N N

Tol

Tol

Complex 18 reported by Crabtree and coworkers has an unusually long N—N bond (1.21 Å) for a group 7 μ-N2 ligand [25a]. The N–N stretch for this complex was reported at 1660 cm−1 . Despite the unusually long bond reported here, it is worth noting that the N—N bond is not cleaved. This synthetic strategy was employed later by Leung and coworkers for the synthesis of bridged dinitrogen complexes of iron (Figure 5.9), 19 and 20, and chromium, 21, porphyrins (Figure 5.10) [26]. The N—N bond of 1.17 Å was observed for 19 along with N–N vibrations of 1803–1820 cm−1 for 19 and 20, respectively. In order to test the hypothesis that asymmetric binding in a heterobimetallic bridging dinitrogen complex may lead to unsymmetrical nitride cleavage, Brown and a coworker examined the chemistry of the heterobimetallic complex, 22 (Figure 5.11) [27]. The N—N bond for 22 was reported at 1.167 Å, whereas

277

278

5 Toward N—N Bond Cleavage: Synthesis and Reactivity of Group 7 Dinitrogen Complexes

Et

Et

OTf N2

N

Et N

Et

Cr

Et

PhMe2P

+

N

PMe2Ph

PhMe2P

Et

N

PMe2Ph

Re CI

+ Et

PhMe2P PhMe2P Re

Et Et

Et

PMe2Ph PMe2Ph

N N

N

Et N Et

21

Et

N

Cr

Et

N

Et

CI

Et

Figure 5.10 Syntheses of bridging dinitrogen group 7 complexes by reactions with chromium porphyrin complexes.

Et

N

S S

Et N 1/2

Et

S Mo S

S

N

2+

S

S

S

Et Et

Et

Mo S

Et N Et

N2 +

PhMe2P

N

PMe2Ph

Re

PhMe2P

PMe2Ph

CI

S

S Et

S

Et N Et

Et

Et

N

S Mo

N Et

+ S

S

Et

Et

S

S

Et N

S

Et

N N PhMe2P PhMe2P

Re

PMe2Ph PMe2Ph

CI 22

Figure 5.11 Synthesis of heterobimetallic group 7 dinitrogen complex.

5.1 Synthesis of Group VII N2 Complexes

N–N vibrations were observed at 1818 cm−1 . Most importantly, the N—N bond is not cleaved in this molecule and the corresponding nitrides (isolated independently) do not couple to produce 22. The authors conclude that it is kinetic rather than thermodynamic factors that govern the stability of the bimolecular molecule 22. 5.1.2.2

Alternative Syntheses of Bridged N2 Complexes

Group 7 complexes with bridging N2 ligands have also been synthesized by other methods. One of the simplest methods involves trapping of a low-valent group 7 species with N2 . This method was used by Harman and coworkers to produce 23 supported by hydridotris(pyrazolyl)borate (Tp) (Figure 5.12) [28]. A N—N bond length of 1.15 Å was reported for this molecule. An alternative method was demonstrated by Herrmann and coworkers (Figure 5.13) who showed that treatment of Cp/Cp′ Mn(CO)2 (THF) (Cp = C5 H5 ) with 1,1,1-trifluorodiazoethane resulted in the bridged nitrogen manganese(I) complex 24 [29]. The N–N vibration was observed by Raman spectroscopy at 1971 cm−1 for [Cp′ Mn(CO)2 ]2 (N2 ), whereas a N—N bond of 1.118 Å was observed by X-ray crystallography. For this system, the intermediacy and nucleophilicity of the dinitrogen ligand in CpMn(CO)2 (N2 ) [15] is strongly implied, as this complex was shown to react with Cr(CO)5 (THF) to generate 25 [30]. A strategy analogous to methods described earlier where a low-valent group 7 metal was reduced with a strong reducing agent was employed by the Arnold group to synthesize manganese bridged dinitrogen complexes (Figure 5.14). Manganese precursors with P2 N2 ligands were treated with sodium naphthalide at −40 ∘ C under a nitrogen atmosphere to generate 26 [31]. Bridged N2 complexes can also be synthesized by photolysis. This is exemplified in Figure 5.15, where a rare technetium complex bearing hydridotris(3,5Figure 5.12 Synthesis of bimetallic group 7 dinitrogen complex by trapping a rhenium(I) precursor with N2 .

Tp Re

O

Benzene OC

OC OC

CO

O

Mn

–1,1,1-Trifluoroethane

Cp′/Cp

N

Re

23

Cp/Cp′

1,1,1-Trifluorodiazoethane CO

Mn

N

Re OC

Cp/Cp′ OC OC

Tp

Tp

N2

N N 24

Mn OC

CO

Cp′ = MeC5H4 CO

Cp Toluene/THF CO + Cr(CO)5THF Mn OC N2 OC

Mn –50 °C Cp

OC CO O N N Cr 25

OC

CO

Figure 5.13 Synthesis of bimetallic group 7 dinitrogen complex by decomposition of 1,1,1-trifluorodiazomethane.

CO CO

279

280

5 Toward N—N Bond Cleavage: Synthesis and Reactivity of Group 7 Dinitrogen Complexes i

Pr

iPr CI

i

Pr

i

P

i

Pr

NaC10H8 N2

tBu

Pr Mn

N

P

N

THF, –40 °C N

Me2Si

Si Me2

i

t

Pr

Bu

iPr P Pi Pr Mn N N iPr

N tBu

N

SiMe2

Mn iPr

iPr

P

N

P

iPr

26

Figure 5.14 Synthesis of bimetallic manganese dinitrogen complex with P2 N2 ligands. Tp′ 2

OC

hν N2

Tc

CO

OC

THF

Tp′ OC

CO

Tc

N N 27

OC

Tc

CO Tp′

Figure 5.15 Synthesis of bimetallic technetium dinitrogen complex by photolysis.

dimethyl-1-pyrazolyl)borate (Tp′ ), 27, was synthesized from the photolysis of Tp′ Tc(CO)3 in THF under a nitrogen atmosphere [32]. This molecule was structurally characterized, and a 1.16 Å N—N bond was reported.

5.2 Cleavage and Functionalization of N2 Bonds As outlined above, a few group 7 dinitrogen complexes have been reported. Further, terminal nitrogen complexes have been shown to react with transition metals to form bridged dinitrogen complexes. However, there have been only two examples where a coordinated N2 ligand has been reduced. These examples are outlined below. 5.2.1

Generation of Diazomethane from CpMn(CO)2 N2

Sellmann and Weiss [33] reported the formation of diazomethane by the initial reaction of methyllithium in THF with CpMn(CO)2 N2 at −30 ∘ C to generate intermediate 28, followed by the reaction with Meerwein reagent Me3 OBF4 to Cp

Cp LiCH3

Mn OC N2 –30 °C, THF OC

Mn OC N – OC N

N N

28

100 bar 20 °C, Et2O

N2

Cp (CH3)3OBF4

Mn

OC OC 29

N N

–30 °C, THF

Figure 5.16 Formation of diazomethane by the reaction of methyllithium with CpMn(CO)2 N2 .

5.3 Conclusions and Future Outlook

tBu

tBu P

tBu

1 equiv Na/Hg or Co(Cp*)2

CI N Re P

N Re P

P

CI N

N

Re N

CI

P

tBu

tBu

tBu

tBu

tBu

P

N2

CI tBu

tBu

tBu

tBu

30 tBu

tBu

tBu P

P

N

N Re

XOTf CI

N Re

P

P

tBu

tBu

tBu

tBu

31

X N

+

CI tBu

X = Me, H

Figure 5.17 Cleavage of N2 in the coordination sphere of rhenium.

generate the second intermediate 29 (Figure 5.16). Diazomethane is ultimately released by pressurizing 29 with N2 . This reaction could not be made catalytic because of the side reactions of CpMn(CO)2 . However, the most notable aspect of this reaction is that the N—N bond is ultimately not cleaved. 5.2.2

Cleavage of N2 in the Coordination Sphere of Rhenium

The first example of N2 cleavage in the coordination sphere of a group 7 element was reported by Schneider and coworkers in 2014 [3a]. As shown in Figure 5.17, reduction of a rhenium(III) complex under a nitrogen atmosphere resulted in the rhenium(V) species 31. Although not directly detected, the reaction presumably proceeds through a putative dirhenium(II) complex, 30, with a bridged dinitrogen molecule. For N2 cleavage, the density functional theory (DFT) calculations suggest that the transition state for this reaction contains a zigzag structure similar to the molybdenum complex reported by Cummins and coworkers [34]. The resulting nitrido species 31 undergoes expected nucleophilic reactivity with MeOTf and triflic acid. This reactivity was exploited later by the same group to develop a synthetic cycle for the synthesis of acetonitrile [3b].

5.3 Conclusions and Future Outlook Clearly compared to groups 6 and 8, N2 activation with group 7 complexes is still in its infancy. However, given that both terminal and bridged N2 complexes are known, and at least in one case, a bridged dinitrogen complex is proposed to lead to N2 cleavage, the prospect for catalytic N2 activation with group 7 is still promising. It is noteworthy that the intermediate in Schneider’s chemistry features two

281

282

5 Toward N—N Bond Cleavage: Synthesis and Reactivity of Group 7 Dinitrogen Complexes

d5 Re(II) centers bridged by a nitrogen molecule. It is not clear at this time why, or if, this particular electronic arrangement is critical. An understanding of this particular electronic structure is thus important and may provide useful insights for future catalytic designs.

References 1 Khoenkhoen, N., de Bruin, B., Reek, J.N.H., and Dzik, W.I. (2015). Eur. J.

Inorg. Chem. 567–598. 2 Bezdek, M.J. and Chirik, P.J. (2016). Angew. Chem. Int. Ed. 55: 7892–7896. 3 (a) Klopsch, I., Finger, M., Würtele, C. et al. (2014). J. Am. Chem. Soc. 136:

4

5

6 7 8

9 10 11

6881–6883. (b) Klopsch, I., Kinauer, M., Finger, M. et al. (2016). Angew. Chem. Int. Ed. 55: 4786–4789. (a) Kirillov, A.M., Haukka, M., Guedes, da Silva, M.F.C. et al. (2006). J. Organomet. Chem. 691: 4153–4158. (b) Albertin, G., Antoniutti, S., Bacchi, A. et al. (2000). Inorg. Chem. 39: 3283–3293. (c) Pombeiro, A.J.L., Hitchcock, P.B., and Richards, R.L. (1987). J. Chem. Soc., Dalton Trans. 319–325. (d) Carvalho, F.N.N., Pombeiro, A.J.L., Orama, O. et al. (1982). J. Organomet. Chem. 240: C18–C22. (e) Carvalho, M.F.N.N., Pombeiro, A.J.L., Schubert, U. et al. (1985). J. Chem. Soc., Dalton Trans. 2079–2084. (f ) Wang, Y., Fraústo, Da Silva, J.J.R., Pombeiro, A.J.L. et al. (1993). J. Organomet. Chem. 454: 211–216. (g) Guedes da Silva, M.F.C., Fraústo da Silva, J.J.R., Pombeiro, A.J.L. et al. (1996). J. Chem. Soc., Dalton Trans. 2763–2772. (h) Heinekey, D.M., Voges, M.H., and Barnhart, D.M. (1996). J. Am. Chem. Soc. 118: 10792–10802. (i) Albertin, G., Antoniutti, S., Bordignon, E., and Visentin, E. (2001). Inorg. Chem. 40: 5465–5467. (j) Darensbourg, D.J. and Madrid, D. (1974). Inorg. Chem. 13: 1532–1534. (a) Ibers, J.A. and Davis, B.R. (1971). Inorg. Chem. 10: 578–585. (b) Chatt, J., Dilworth, J.R., Richards, R.L., and Sanders, J.R. (1969). Nature 224: 1201–1202. (a) Chiu, K.W., Howard, C.G., Rzepa, H.S. et al. (1982). Polyhedron 1: 441–451. (b) Komiya, S. and Baba, A. (1991). Organometallics 10: 3105–3110. Chiu, K.W., Wong, W.-K., Wilkinson, G. et al. (1982). Polyhedron 1: 37–44. (a) Sellmann, D. (1972). J. Organomet. Chem. 36: C27–C28. (b)Sellmann, D. and Kleinschmidt, E. (1977). Z. Naturforsch., B: Anorg. Chem., Org. Chem. 32: 795–801. (c) Klahn, A.H. and Sutton, D. (1989). Organometallics 8: 198–206. (d) Cusanelli, A. and Sutton, D. (1989). J. Chem. Soc., Chem. Commun. 1719–1720. (e) Cusanelli, A. and Sutton, D. (1996). Organometallics 15: 1457–1464. (f ) Cusanelli, A. and Sutton, D. (1996). Organometallics 15: 3717–3722. Barrientos-Penna, C.F., Einstein, F.W.B., Sutton, D., and Willis, A.C. (1980). Inorg. Chem. 19: 2740–2749. Cusanelli, A. and Sutton, D. (1995). Organometallics 14: 4651–4660. (a) Mösch-Zanetti, N.C., Köpke, S., Herbst-Irmer, R., and Hewitt, M. (2002). Inorg. Chem. 41: 3513–3520. (b) Neuner, B. and Schrock, R.R. (1996).

References

12 13 14

15

16 17 18 19 20 21

22 23 24

25

26 27 28

Organometallics 15: 5–6. (c) Reid, S.M., Neuner, B., Schrock, R.R., and Davis, W.M. (1998). Organometallics 17: 4077–4089. Kaden, L., Lorenz, B., Schmidt, K. et al. (1979). Z. Chem. 19: 305–306. Struchkov, Yu.T., Bazanov, A.S., Kaden, L. et al. (1982). Z. Anorg. Allg. Chem. 494: 91–97. (a) Chatt, J., Dilworth, J.R., and Leigh, G.J. (1969). J. Chem. Soc., Chem. Commun. 687–688. (b) Chatt, J., Dilworth, J.R., and Leigh, G.J. (1973). J. Chem. Soc., Dalton Trans. 612–618. (c) Fernanda, M., Carvalho, N.N., and Pombeiro, A.J.L. (1989). Polyhedron 8: 1778–1779. (d) Carvalho, M.F.N.N., and Pombeiro, A.J.L. (1990). J. Organomet. Chem. 384: 121–131. (e) Carvalho, M.F.N.N., Herrmann, R., and Pombeiro, A.J.L. (1993). Monatsh. Chem. 124: 739–749. (f ) Costa, M.T.A.R.S., Fraústo, da Silva, J.J.R., Pombeiro, A.J.L. et al. (1998). Inorg. Chim. Acta 280: 308–315. (g) Pombeiro, A.J.L., Costa, M.T.A.R.S., Wang, Y., and Nixon, J.F. (1999). J. Chem. Soc., Dalton Trans. 3755–3758. (h) Smole´nski, P. and Pombeiro, A.J.L. (2008). Dalton Trans. 87–91. (i) Chatt, J., Hussain, W., Leigh, G.J. et al. (1985). J. Chem. Soc., Dalton Trans. 1131–1136. (a) Chin, R.M., Barrera, J., Dubois, R.H. et al. (1997). Inorg. Chem. 36: 3553–3558. (b) Orth, S.D., Barrera, J., Sabat, M., and Harman, W.D. (1994). Inorg. Chem. 33: 3026–3027. (a) Sellmann, D. (1971). Angew. Chem. Int. Ed. Engl. 10: 919. (b) Sellmann, D. (1972). Angew. Chem. Int. Ed. Engl. 11: 534. Perthuisot, C., Fan, M., and Jones, W.D. (1992). Organometallics 11: 3622–3629. Dilworth, J.R., Hu, J., Thompson, R.M., and Hughes, D.L. (1992). J. Chem. Soc., Chem. Commun. 551–553. Tully, M.E. and Ginsberg, A.P. (1973). J. Am. Chem. Soc. 95: 2042–2044. Smeltz, J.L., Boyle, P.D., and Ison, E.A. (2012). Organometallics 31: 5994–5997. (a) Chatt, J., Dilworth, J.R., and Leigh, G.J. (1970). J. Organomet. Chem. 21: P49–P50. (b) Chatt, J., Dilworth, J.R., Gunz, H.P. et al. (1970). J. Chem. Soc. D 90–91. (c) Chatt, J., Crabtree, R.H., Dilworth, J.R., and Richards, R.L. (1974). J. Chem. Soc., Dalton Trans. 2358–2362. Chatt, J., Fay, R.C., and Richards, R.L. (1971). J. Chem. Soc. A. 702–704. Robson, R. (1974). Inorg. Chem. 13: 475–479. (a) Chatt, J., Dilworth, J.R., Leigh, G.J., and Richards, R.L. (1970). J. Chem. Soc. D 955–956. (b) Cradwick, P.D., Chatt, J., Crabtree, R.H., and Richards, R.L. (1975). J. Chem. Soc., Chem. Commun. 351–352. (c) Cradwick, P.D. (1976). J. Chem. Soc., Dalton Trans. 1934–1936. (a) Mercer, M., Crabtree, R.H., and Richards, R.L. (1973). J. Chem. Soc., Chem. Commun. 808–809. (b) Mercer, M. (1974). J. Chem. Soc., Dalton Trans. 1637–1640. Zhang, Q.-F., Chim, J.L.C., Lai, W. et al. (2001). Inorg. Chem. 40: 2470–2471. Seymore, S.B. and Brown, S.N. (2006). Inorg. Chem. 45: 9540–9550. Gunnoe, T.B., Sabat, M., and Harman, W.D. (1998). J. Am. Chem. Soc. 120: 8747–8754.

283

284

5 Toward N—N Bond Cleavage: Synthesis and Reactivity of Group 7 Dinitrogen Complexes

29 (a) Weidenhammer, K., Herrmann, W.A., and Ziegler, M.L. (1979). Z. Anorg.

30 31 32 33 34

Allg. Chem. 457: 183–188. (b) Ziegler, M.L., Weidenhammer, K., Zeiner, H. et al. (1976). Angew. Chem. Int. Ed. 15: 695–696. Sellman, D., Gerlach, R., and Jödden, K. (1979). J. Organomet. Chem. 178: 433–447. (a) Chomitz, W.A. and Arnold, J. (2007). Chem. Commun. 4797–4799. (b) Chomitz, W.A. and Arnold, J. (2009). Dalton Trans. 1714–1720. Joachim, J.E., Apostolidis, C., Kanellakopolus, B. et al. (1993). J. Organomet. Chem. 455: 137–141. Sellmann, D. and Weiss, W. (1978). Angew. Chem. Int. Ed. Engl. 17: 269–270. Laplaza, C.E., Johnson, M.J.A., Peters, J.C. et al. (1996). J. Am. Chem. Soc. 118: 8623–8638.

285

6 Group 8 Transition Metal–Dinitrogen Complexes Adam D. Piascik and Andrew E. Ashley Imperial College London, Department of Chemistry, Molecular Science Research Hub, White City Campus 80 Wood Lane, W12 0BZ, UK

6.1 Introduction The group 8 transition metals (TMs), particularly Fe and Ru, play a key role in some of the most important systems for biological and industrial dinitrogen fixation. The Haber–Bosch (H–B) process for ammonia production is one of the largest industrial chemical processes globally, with an annual output of approximately 150 MT, and uses a Fe3 O4 catalyst doped with several additives that promote activity (Figure 6.1a) [1, 2]. Biological nitrogen fixation is achieved by the use of nitrogenases, which are metalloenzymes featuring a TM-containing cofactor (Figure 6.2). To date, all known nitrogenases contain Fe in this cofactor, either as the sole transition metal component or in tandem with Mo or V [3, 4]. In 1965, Allen and Senoff reported the first transition metal–dinitrogen (TM–N2 ) complexes of the type [Ru(NH3 )5 (N2 )]X2 (X = Br− , I− , BF4 − , or PF6 − ) [5], and Ru-based heterogenous catalysts have proved highly effective for ammonia production on an industrial scale, e.g. the Kellogg advanced ammonia process (KAAP), which employs much milder operational conditions than the H–B process (Figure 6.1b) [6–8]. Osmium, in contrast, has seen comparatively little attention from the industrial and research communities because of its expense and toxicity. Nevertheless, Haber’s first ammonia-producing reactors did in fact employ a metallic osmium catalyst, the scarcity of which prompted the subsequent search for a cheaper, more abundant replacement [9]. cat. Fe (a)

400–500 °C, 100–250 atm

N2 + 3 H2

2 NH3 (b) cat. Ru

260– 280 °C, 90 –140 atm

Figure 6.1 (a) The Haber–Bosch process and (b) the KAAP process for NH3 production. Transition Metal-Dinitrogen Complexes: Preparation and Reactivity, First Edition. Edited by Yoshiaki Nishibayashi. © 2019 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2019 by Wiley-VCH Verlag GmbH & Co. KGaA.

286

6 Group 8 Transition Metal–Dinitrogen Complexes

Nitrogenase

N2 + 8 H+ + 8 e–

16 ATP

2 NH3 + H2

16 ADP + Pi

Figure 6.2 Schematic for the biological fixation of N2 by nitrogenase enzymes (FeMo cofactor pictured).

S S Fe

(Cys)S

Fe C

S Fe Fe

S

Fe

S

S

Fe S Mo Fe

S

N(His) O O O

S

CO2H

HO2C

FeMo-cofactor

Because of this importance in biology and industry, much research in homogeneous TM–N2 chemistry has focused on the use of Fe and Ru, and to a lesser extent Os. As already mentioned, the complexes of the group 8 metals were among the first N2 -containing complexes isolated (Figure 6.3) [5, 10, 11]; however, early investigations into the reactivity of group 8 TM–N2 complexes found little success in functionalizing the coordinated N2 ligand. In their original publication, Allen and Senoff reported that reaction of [Ru(NH3 )5 (N2 )]I2 with aqueous NaBH4 and NaOH produced NH3 in a 24% yield; later studies disproved this finding and showed the NH3 to arise from the reduction of a hydrazine-containing impurity, a side product in the synthesis of the Ru–N2 complex [12]. A greater promise was shown in other bond activation reactions; early examples include the insertion of CO2 and C2 H4 into the Fe—H bond of FeH2 N2 L3 (L = PEtPh2 ) and the use of [(PP3 )Fe(H)(N2 )]BPh4 (PP3 = P(C2 H4 PPh2 )3 ) as an effective precatalyst for the hydrogenation of alkynes to alkenes [13, 14]. Interest in the use of group 8 TM–N2 complexes to fix dinitrogen was rekindled in 1991 with the suggestion by Leigh and coworker that NH3 could be produced from a simple Fe bis(diphosphine) complex [15]. They reported that the treatment of [Fe(H)(N2 )(dmpe)2 ]BPh4 (dmpe = Me2 PC2 H4 PMe2 ) with KOt Bu in tetrahydrofuran (THF), followed by the acidification of the resulting solution with HCl, produced NH3 in the yields of up to 12%. No intermediate species were 1965

H3N H3N

N2 Ru NH3

NH3

1968

1967 2+

2+

–

2X

NH3

(1) X = Br, I, BF4, or PF6

H3N H3N

N2 Os NH3

NH3

–

2X

NH3

(2) X = Cl, Br, I, ClO4, BF4, or BPh4

Figure 6.3 The first group 8 M–N2 complexes synthesized.

RPh2P

N2 Fe

H H

PPh2R PPh2R

(3) R = Et or Bu

6.1 Introduction

P

N2

P

Fe P

H

KOtBu

P

Fe P

P

P N2

P

xs. HCl

P P

Cl Fe

P

+ NH4Cl ( +30 kcal mol–1 (b)

N

–n

N H P Fe P H Si O

P Fe P Si

+ N2 ΔG°(N2)

N

N

N H P Fe P H Si O

N Fe

–n

P P

Si

n = 0, ΔG°(N2)= –0.06 kcal mol–1 n = 1, ΔG°(N2)= –8.8 kcal mol–1

Figure 6.27 Variation of N2 binding affinity upon redox changes in (a) an Fe tetraphosphine complex and (b) an Fe μ-(H)2 system. All data are reported at 1 atm N2 and 298 K; counterions omitted for clarity.

6.3.2

Cleavage and Functionalization of Coordinated N2

Examples of the cleavage of a coordinated N2 ligand have, for the most part, been limited to complexes of Fe; only two other examples have been reported, both involving Os. Upon photolysis, the Os(II)/Os(III) mixed valence dimer

301

302

6 Group 8 Transition Metal–Dinitrogen Complexes

[(NH3 )5 Os(μ-N2 )Os(NH3 )5 ]5+ was found to cleave the N—N bond to primarily yield the Os(VI) nitride [Os(≡N)(NH3 )4 ]3+ ; the related Os(II) monomer [Os(N2 )(NH3 )5 ]2+ was found to react photolytically with O2 to give the same Os(VI) nitride and NO−2 , formally a disproportionation of the N2 ligand to N(III) and N(−III) [147, 148]. The first example of stepwise reduction of the N—N bond order by an Fe complex came in 2001, reported by Holland and coworkers [78]. Reduction of a β-diketimate Fe chloride under an atmosphere of N2 resulted in the formation of a binuclear μ-N2 Fe(I) complex (Figure 6.28) with a substantially weakened N—N bond, as evidenced by a resonance Raman band of 1778 cm−1 for the N—N stretch (cf. 2331 cm−1 in free N2 ). A further reduction with Na or K metal yielded the dianionic Fe(0) congeners, wherein the alkali metal ions are coordinated to the N2 ligand in a side-on manner. The N2 unit itself exhibits Raman-active stretches at 1123 and 1589 cm−1 , and a solid-state structure revealed an N–N distance between 1.23 and 1.24 Å; these data are consistent with an N=N double bond. A computational study of this system found that the N—N bond is weakened upon reduction by an increased population of Fe–N π and N–N π* orbitals, and coordination of the alkali metal cation enhances the mixing of the involved N–N π* and Fe dπ orbitals [149]. A reduction in the steric bulk of the β-diketimate backbone causes a profound change in the extent of N2 reduction achievable; treatment of an analogous but less bulky Fe chloride precursor with KC8 under N2 furnished a multinuclear complex in which the N2 unit is completely cleaved (Figure 6.29) [26]. The complex contains extensively coordinated nitride fragments, and a detailed computational investigation revealed both the necessity of multiple Fe centers and the

tBu

tBu

N

KNap

Fe Cl tBu

Fe N

N2 atm

N

N

N

tBu

N

N

tBu

Fe N

tBu

Na or K metal

N Fe N

tBu

tBu

M N N

tBu

Fe

N

N M

tBu

M = Na, K

Figure 6.28 Formation of an Fe μ-N2 complex and subsequent reductive cleavage of the N≡N bond.

N

N

Ar Ar N Fe N N N Fe Ar Ar N N Fe

N Fe N

Cl Cl

N

2 KC8

N

N

N2 atm

N

Fe

Fe

Cl

K

Cl

K

Ar = 2,6-(Me)2C6H3

Figure 6.29 Complete cleavage of the N≡N bond by a multimetallic Fe/K complex.

6.3 Stoichiometric Reactions of Group 8 Transition Metal–Dinitrogen Complexes

N N Fe N

Cl Cl

N Ar

Ar N N N Fe N Fe Ar Ar N N

K K

Fe

Ar = 2,6-(Me)2C6H3

2 18-crown-6 –

N N

Cl Fe

N

Cl

K(18-crown-6)

+

Fe

HN N

N

Ar N

Fe

N

+ [K(18-crown-6)]

Fe

N Ar

N Ar

Figure 6.30 N—H bond formation via ligand deprotonation following K+ sequestration from an Fe bis(nitride) complex.

N N Fe N

Cl Cl

K K

N Ar

Ar N N N Fe N Fe Ar Ar N N Fe

CO (1 atm) 4 –N2

O C N Fe C O N C O

Ar = 2,6-(Me)2C6H3

Figure 6.31 N2 formation from an Fe bis(nitride) complex induced by addition of CO.

importance of coordinating K+ ions, in achieving full cleavage of the N2 unit [150, 151]. Indeed, sequestration of the K+ ions in the Fe bis(nitride) complex by 18-crown-6 causes decomposition of the complex by proton abstraction from ligand backbone (Figure 6.30), presumably via a transient anionic nitride fragment [152]. Remarkably, cleavage of N2 to nitrides by this system was found to be reversible, with the addition of CO, ArNC (Ar = 2,6-(CH3 )3 C6 H3 ), or catalytic pyridine in benzene inducing oxidation of the nitride fragments and concomitant expulsion of N2 (Figure 6.31) [27]. In an initial report, the above Fe bis(nitride) complex was claimed to yield substoichiometric quantities of NH3 upon addition of H2 ; this was later disproven, the NH3 instead being formed upon addition of ethereal HCl during the workup procedure. Subsequent mechanistic investigation identified several

303

304

6 Group 8 Transition Metal–Dinitrogen Complexes (a) N N Fe N

Cl Cl

K K

N Ar

Ar N N N Fe N Fe Ar Ar N N Fe

N Ar′-C CH

Fe

HN

Ar′ = 3,5-(CF3)2C6H2

N

N

N Fe

N Fe

Ar

Ar

N

Ar

N Ar

Ar = 2,6-(Me)2C6H3 (b) N

Fe

HN N

N

N Fe

N Fe

N

N

Ar

Ar

2,4,6-tBu3C6H2OH

N

Ar

H2 N N

Ar

Fe

N

N Fe

N Fe

Ar

Ar

N

N

Ar

Ar

(c) N

Fe

H2N N

N

N Fe

N Fe

N

Ar

N

Ar

Ar

N Ar

2 equiv 2,4,6-tBu3C6H2OH

N Fe

H2N N

Ar

NH2 Fe

N

Ar

Figure 6.32 Sequential N—H bond formation using an Fe bis(nitride) complex.

partially protonated intermediates in the H+ -mediated production of NH3 ; an imido/nitride species, an amido/nitride species, and a bis(amido) species were all independently isolated (Figure 6.32) [28]. Potassium is not unique in facilitating the cleavage of N2 to nitrides at β-diketimate-supported Fe centers; both sodium and rubidium reductants have been shown to affect cleavage, although the product of reduction with Na metal possesses a slightly different structure (Figure 6.33) [153, 154]. By contrast, CsC8 proved ineffective for N2 cleavage, although reaction of 2 equiv of CsC8 with an Fe chloride precursor under an N2 atmosphere did yield a novel trinuclear complex with two bridging N2 units (Figure 6.34). Use of Mg(0) as a reductant was attempted but did not lead to an isolable product [155]. A related β-diketimate Fe system has also been shown to bind and cleave N2 upon reduction of an Fe bromide precursor (Figure 6.35) [156]. The N2 -cleaved product contains N–H fragments rather than nitrides (although the exact nature of these fragments was not identified), and the source of the protons in this species is unclear; treatment of this N2 -cleaved product with HCl generated substoichiometric NH3 .

6.3 Stoichiometric Reactions of Group 8 Transition Metal–Dinitrogen Complexes

(a) N

N Fe N

Cl

N

2 Na metal

N

N2 atm

Fe

Cl

THF Na THF

N Ar

Ar N N N Fe N Fe Ar Ar N N Fe

Ar = 2,6-(Me)2C6H3 (b) N N Fe N

Cl

N

2 RbC8

N

N

N2 atm

N

Fe

Cl

Fe

N Ar

Ar N N Fe N Fe Ar Ar N N Fe

Rb N Rb

Cl Cl

Ar = 2,6-(Me)2C6H3

Figure 6.33 Formation of Fe nitride complexes. (a) a triiron bis(nitride) complex using a Na reductant and (b) a tetrairon bis(nitride) complex employing a Rb reductant.

Ar N N

Cl

Fe

Cl

N Fe N

2 CsC8

Ar N N

N2 atm Ar

Fe

N Cs N N Cs Fe Fe Cl N Ar N N N N

Ar

Ar

Ar = 2,6-(Me)2C6H3

Figure 6.34 CsC8 reduction of an Fe chloride precursor to a trinuclear Fe bis(μ-N2 ) species.

N N

Br Fe

N

N

Fe Br Br

N

N Fe N

6 KC8 N2 atm

N H xN Fe N

Fe N NHx

NHx N Fe N

Figure 6.35 Formation of a trinuclear Fe complex with (NHx ) bridging ligands by reduction of an Fe bromide precursor.

305

306

6 Group 8 Transition Metal–Dinitrogen Complexes

Stoichiometric functionalization of Fe-bound N2 was first reported by Peters and coworker who found that addition of MeOTs to an anionic Fe(0) dinitrogen species resulted in the methylation of terminal nitrogen of the N2 unit (Figure 6.36) [96]. Subsequent studies of N2 functionalization on Fe have tend to focus on silylation, silylium ions (R3 Si+ ) particularly being popular mimics for H+ , but whose N2 -functionalized products often display greater kinetic stability than their protonated counterparts [157]. The tripodal ligand systems used by Peters and coworkers have proven adept at stoichiometric silylation of Fe-bound N2 . The anion [(SiPiPr 3 )Fe(N2 )]− reacted with Me3 SiCl to give an isolable silyldiazenido species (Figure 6.37); a similar silylation was achieved for the [(BPiPr 3 )Fe(N2 )]− anion, and further reduction of the resultant diazenido product yielded an unusual mono-silylated hydrazido species (Figure 6.38a) [119, 158]. An analogous bis(silyl) hydrazido complex could be generated on the same framework (Figure 6.38b); collectively, this series of related species indicates that stepwise functionalization of the N2 unit causes sequential weakening of the N—N bond with concomitant strengthening of the Fe—N bond. Modification of the supporting ligand system has allowed access to a series of unusual species derived from molecular N2 [92]. Reaction of (BPR2 Ph)Fe(Br) (R = i Pr, Ph) with Na/Hg and 1,2-bis(chlorodimethylsilyl)ethane under an atmosphere of N2 furnished a bis(silyl) hydrazido species in which the hapticity of the phenyl ligand shifts from η1 to η3 (Figure 6.39) [159]. Addition of PhSiH3 to the isopropyl-substituted hydrazido complex caused hydrosilylation across the Fe≡N MgCl(THF)2

Cl

N

N

N

N MeOTs

Me

Fe

Mg

Fe

P P P

N2 atm

P P P

P P P

B

B

B

Ph

Ph

Ph

P=

Fe

PiPr2

Figure 6.36 Synthesis of an anionic Fe–N2 complex followed by methylation at the terminal nitrogen. Na(THF)3

SiMe3

N

N

N P Fe P Si

P

P = PiPr2

Me3SiCl

N P Fe P Si

P

Figure 6.37 Monosilylation of [(SiPiPr 3 )Fe(N2 )]− to generate a silyldiazenido species.

6.3 Stoichiometric Reactions of Group 8 Transition Metal–Dinitrogen Complexes

Na

(a)

SiMe3

N

N

N

N

P Fe P B

Me3SiCl

P

P Fe P B

Me3Si xs. Na/Hg

P

N Na(THF)

N P Fe P B i

Pr2P

P = PiPr2 (b)

Si

Si N Br P Fe P B

xs. Na/Hg N2 atm

P

Si

N P Fe P B

Si

P

Cl Cl P = PiPr2

Figure 6.38 Functionalization of N2 by a (BPiPr 3 )Fe system. (a) silyldiazenido and mono(silyl) hydrazido species and (b) a bis(silyl) hydrazido species.

P

Fe

B PhSiH3

P B

3 Na/Hg N2 atm

P Fe

Br

P

PiPr2

P

B Si

P=

Si

P N

H

N Si

SiH2Ph

Si Fe

N N Si

Si

Cl Cl P = PPh2, PiPr2 H2 atm

P = PPh2

Si P

N

P B

Fe H

Si N H

Figure 6.39 Coordination and reaction of N2 to yield unusual N2 -functionalized products supported by an Fe bis(phosphino)borane framework.

307

308

6 Group 8 Transition Metal–Dinitrogen Complexes

(a)

O

N Ar

N O

O K

Fe N N

O

O

Ar

R3SiCl

Fe N

R = Me, Et

N SiR3

O N Ar

N Ar

Ar = 2,6-C6H3(iPr)2 +

(b) P

P Fe

P

N

N

P

Me3SiCl, KBArF4 ArF =

3,5-C6H3(CF3)2

P

P Fe

P

P

N

BArF4–

N SiMe3

P = PMe2, PEt2

Figure 6.40 Monosilylation of (a) an Fe(−I) bis(carbene) species and (b) Fe(0)–N2 bis(diphosphine) complexes.

bond and formation of a bridging hydride between the Fe and B atoms. Addition of H2 to the analogous phenyl derivative was proposed to hydrogenate the Fe≡N bond and generate the same bridging hydride, but this rapidly underwent a complex rearrangement to a species in which the N—N bond has been completely cleaved. Monosilylation of N2 from a two-coordinate Fe(−I) complex bearing bulky carbene ligands (Figure 6.40a) has been reported; this Fe(−I) species also proved capable of the catalytic conversion of N2 to N(SiMe3 )3 [93]. Ashley and coworkers have achieved analogous monosilylation of N2 using Fe bis(diphosphine) complexes [160]. Treatment of Fe(N2 )(pp)2 (pp = dmpe/depe) with a mixture of Me3 SiCl and KBArF4 (ArF = 3,5-C6 H3 (CF3 )2 ) furnished the corresponding [Fe(NN-SiMe3 )(pp)2 ]+ species (Figure 6.40b), the first examples of cationic Fe diazenido species. Detailed theoretical calculations showed these to be useful electronic surrogates for their more elusive protonated analogs, [Fe(NN-H)(pp)2 ]+ . Despite their noted instability, a handful of species bearing a protonated N2 ligand derived from molecular N2 has also been documented. Peters and − coworker reported that addition of triflic acid to the anion [(SiPiPr 3 )FeN2 ] afforded an isolable, cationic hydrazido species (Figure 6.41a), which was stable in the solid state and in solution below 0 ∘ C [161]. Reduction to a neutral hydrazido analog was possible using CoCp∗2 , but the instability of this compound above −78 ∘ C limited spectroscopic studies. Likewise, the analogous + ∘ cationic hydrazido species [(BPiPr 3 )Fe≡NNH2 ] was stable only below −78 C; despite this, a detailed characterization by EXAFS (Extended X-Ray Absorption

6.3 Stoichiometric Reactions of Group 8 Transition Metal–Dinitrogen Complexes (a)

K(THF)x

+ NH2

N N P Fe P Si

P

3 TfOH –135 °C

P = PiPr2 (b)

N P Fe P Si

NH2

P

OTf

CoCp*2 –135 °C

Stable below 0 °C

N P Fe P Si

P

Stable below –78 °C

+

K(THF)2 NH2

N N P Os P Si

–

3 TfOH P –135 °C

N P Os P Si

P

OTf

–

P = PiPr2

Figure 6.41 (a) The synthesis of an isolable, cationic Fe hydrazido species and subsequent reduction and (b) the synthesis of an isolable Os hydrazido species.

Fine Structure), EPR (Electron Paramagnetic Resonance), and Mössbauer spectroscopy was achieved [162]. Recently, the generation of an Os hydrazido species was reported (Figure 6.41b); not only is this a rare example of N2 functionalization mediated by Os but this complex was shown to be capable of catalyzing the conversion of N2 to NH3 , albeit with a modest turnover (2.6 equiv NH3 per Os center) [20]. In addition to cationic species, neutral Lewis acids can weaken the N—N bond in group 8 TM–N2 complexes through coordination to the N2 unit; early work found the frequency of the N–N stretching absorption in the IR spectrum of Os(Cl)2 (N2 )(PEt2 Ph)3 to decrease by 81 cm−1 upon addition of AlMe3 , indicative of a weakened N—N bond [163]. A systematic investigation into the effect of a range of exogenous Lewis acids on the N—N bond strength (as judged by IR spectroscopy) in Fe(N2 )(depe)2 has been conducted by Szymczak and coworkers, who found N—N bond strength to inversely correlate with Lewis acid strength, as quantified by acceptor number [164]. Furthermore, an N2 -bound adduct of the neutral Lewis acid B(C6 F5 )3 based on this framework was isolated and found to promote kinetic pronation at the terminal nitrogen (rather than thermodynamically preferred protonation at Fe) because of the polarization of the N—N bond (Figure 6.42). 6.3.3

Other Stoichiometric Reactivity

Stoichiometric reactions of group 8 TM–N2 complexes in which the bound N2 acts merely as a spectator ligand are rare. Nevertheless, AlEt3 was found to abstract a phosphine ligand from the six-coordinate hydrido–dinitrogen complex

309

310

6 Group 8 Transition Metal–Dinitrogen Complexes

+

P

P Fe

P

N

P

P

B(C6F5)3

N

Fe P

P

N

N

B(C6F5)3 HBArF4

P

–30 °C

P

P

P Fe

N

B(C6F5)3

N

H

P

BArF4–

ArF = 3,5-C6H3(CF3)

P = PEt2

Figure 6.42 Generation of an N2 -bound adduct of B(C6 F5 )3 and subsequent protonation at the terminal nitrogen.

O

H +CO2

–

–CO2

N + N Ph3P Ru PPh3 H

–

Figure 6.43 Insertion of CO2 into a ligand C—H bond in an Ru–N2 complex.

N + N Ph3P Ru PPh3 H

N2

OH

N2

+

(a)

2+ H

N

N P P

H

1 equiv H(OEt2)2BArF4

P P

H

Fe P

Fe P

N2

N2

P

P

N

N

P = PEt2

(b) N

(THF)3 Na N

N

N P′ P′

Fe

P

1 equiv H(OEt2)2BArF4

P′ P′

Fe Si

Si

P′ = P

N

N P H

NMe

Figure 6.44 Protonation studies on Fe–N2 complexes containing pendant amines.

6.4 Catalytic Reactions of Group 8 Transition Metal–Dinitrogen Complexes

Fe(H)2 (N2 )(PEtPh2 )3 to give the five-coordinate Fe(H)2 (N2 )(PEtPh2 )2 [11]. This provided an early example of the influence of coordination number of a metal center on the degree of activation of bound N2 : in the six-coordinate complex, an IR absorption at 2055 cm−1 attributed to an N–N stretch was shifted to 1989 cm−1 in the five-coordinate analog, indicative of a weaker N—N bond. The same complex was also found to insert CO2 into both Fe—H bonds to generate a diformate complex Fe(O2 CH)2 (PEtPh2 )2 with the loss of N2 . Finally, an unusual example of reversible CO2 insertion into a remote C—H bond of a ligand has been reported for a Ru dinitrogen species (Figure 6.43); in neither of these last cases was further reduction of the formate ligand reported [14, 165]. The nitrogenase enzyme contains a histidine-195 residue, possibly serving as a proton shuttle during the fixation of N2 ; in work pertinent to this, protonation of Fe–N2 complexes containing pendant amines has been investigated [54, 166, 167]. Strong acids were found to protonate such complexes at either the amine group or the Fe center (Figure 6.44), although the production of fixed-N products was not observed [46, 102].

6.4 Catalytic Reactions of Group 8 Transition Metal–Dinitrogen Complexes 6.4.1 Early Results and Fe Bis(diphosphine) Systems for Catalytic N2 Fixation Inspired by the role of Fe in the Haber process and the action of nitrogenase enzymes, the catalytic fixation of N2 to NH3 by homogenous Fe complexes has long been a goal in inorganic chemistry. A milestone in the pursuit of this goal appeared to come in 1991, with a report by Leigh and Jimenez-Tenorio claiming the production of up to 0.12 equiv of NH3 from [Fe(H)(N2 )(dmpe)2 ]BPh4 (dmpe = Me2 PC2 H4 PMe2 ) upon deprotonation with KOt Bu, followed by treatment with an excess of HCl (the Leigh cycle, Figure 6.45) [15]. Later work improved the reported yield of NH3 to 20%, but in neither case was the proposed intermediate Fe(N2 )(dmpe)2 isolated [168]. Komiya and coworkers later reported the isolation of the analogous compound Fe(N2 )(depe)2 (depe = Et2 PC2 H4 PEt2 )

P P

N2 Fe

+

P

KOtBu

P

P

Fe P

H

P N2

P

xs. HCl

P

Cl

P

Fe P

P Cl

P = PMe2 H2 P N2

H2 Fe

P

NaBH4 in EtOH P P

H

Figure 6.45 The Leigh reaction cycle proposed for NH3 formation.

+ NH4Cl (< 13%)

311

312

6 Group 8 Transition Metal–Dinitrogen Complexes

P

Cl

P

NaNap

P

N2 atm

Fe P Cl

P

P

xs. HCl or H2SO4

Fe P

N2

H2 + N2 + Fe(II)by-products

P

P = PEt2

Figure 6.46 Acidification of isolated Fe(N2 )(depe)2 yields N2 and H2 only. + P

P

NH

Fe P

P

NH

P

P

Fe P

P

+2

NH

P

NH2

P

P Fe P

NH2 NH2

P = PEt2 P(CH2CH2CH2OCH3)2

Figure 6.47 Representative examples of synthesized species relevant to N2 fixation by Fe bis(diphosphine) complexes. Counterions omitted for clarity.

and found no evidence of NH3 formation upon treatment of this complex with either HCl or H2 SO4 (Figure 6.46), which is contrary to the original claims of Leigh [16, 75]. This inconsistency was resolved by a recent investigation by Field et al. who unfortunately found the apparent formation of NH3 in Leigh’s original report to arise from a flaw in the procedure used for NH3 quantification, the so-called indophenol method [17]. They found that this spectrophotometric methodology was subject to interference by volatile phosphine ligands produced in the reaction workup, giving rise to a false positive for NH3 production; an NH3 quantification protocol using quantitative 1 H NMR was subsequently developed, which is not susceptible to such an interference, and confirmed the absence of NH3 in the procedure reported by Leigh. Despite this setback, the groups of Tyler and Field succeeded in synthesizing a variety of Fe bis(diphosphine) complexes bearing potential intermediates in N2 to NH3 conversion as ligands (some examples of which are shown in Figure 6.47), hinting at the ability of such platforms to support N2 fixation [116, 169–176]. Indeed, since Leigh’s work, it has been shown that complexes of the type Fe(N2 )(pp)2 (pp = dmpe, depe, DMeOPrPE) (DMeOPrPE = R2 PC2 H4 PR2 , with R = CH2 CH2 CH2 OMe) can mediate the substoichiometric conversion of N2 to fixed products using the strong acid TfOH (Figure 6.48) [94, 115], and Ashley and coworkers have recently reported the ability of Fe(N2 )(depe)2 to produce N2 H4 in a catalytic manner from N2 upon treatment with an excess of [Ph2 NH2 ][OTf ] and CoCp∗2 as H+ and e− sources, respectively (Figure 6.49) [25]. This marked the highest production of N2 H4 from N2 observed for any molecular catalyst, and the system’s remarkable selectivity for N2 H4 over NH3 is unique to date, distinguishing it from other catalytic N2 fixation systems (vide infra).

6.4 Catalytic Reactions of Group 8 Transition Metal–Dinitrogen Complexes

Figure 6.48 Generation of substoichiometric yields of NH3 and N2 H4 from Fe bis(diphosphine) complexes.

P

P Fe

P

N2

P

N2

P

25 °C, N2 atm

NH3 + N2H4

Ligand

Yield of NH3 (%)

Yield of N2H4 (%)

depe

0

9

dmpe

5

24

DMeOPrPE

15

2

360 equiv [Ph2NH2][OTf] 270 equiv CoCp*2 –78 °C

N2

10 equiv TfOH

Fe

P

P = PEt2

P

P

N2H4 (24.5 equiv) + NH3 (95% conversion in just 1.5 hours (Figure 6.66) [77, 97, 187]. Despite this

319

320

6 Group 8 Transition Metal–Dinitrogen Complexes

+

R

O

1 mol% [Fe]

O

22 °C, 15 minutes

HB

> 98% yield R = n-hexyl i-butyl t-butyl CH2SiMe3 CH2NMe2

N N Ar

N N

[Fe] =

BPin

R

Fe N

N

N

Ar Ar = 2,6-C6H3(iPr)2

Figure 6.63 Olefin hydroboration catalyzed by a bis(imino)pyridine Fe–N2 complex. 0.3 mol% [Fe] +

R

PhSiH3

> 98% Substrate

N N N

Ph

Fe N

N

Ar

Time (h)a 1

Ar

N

[Fe] =

SiH2Ph

R

22 °C

1.5 70

N

Ar = 2,6-C6H3(iPr)2

15.5 a

Time to reach 98% conversion

Figure 6.64 Some examples of olefin hydrosilylation catalyzed by a bis(imino)pyridine Fe–N2 complex.

increased activity, the catalytic hydrogenation of sterically hindered, unfunctionalized olefins still proved challenging with these catalysts. However, a modified system wherein the ligand imine groups are replaced by NHCs was found to be highly competent at achieving this transformation for a variety of olefins (Figure 6.67) [79]. The improved activity of the NHC complexes compared with their bis(imino)pyridine counterparts was attributed to the more electron-rich nature of the former, although differences in the redox behavior of the ligands was not ruled out.

6.4 Catalytic Reactions of Group 8 Transition Metal–Dinitrogen Complexes

C6H13

0.004 mol% [Fe]

+ Me(MeO)2SiH

23 °C, 15 minutes

Me(MeO)2Si

C6H13 > 98% yield

N N N

Ar

N

[Fe] =

Ar

Fe N

N

N N

Ar

Fe N

N N Ar

N

Ar = 2,6-C6H3(Me)2

Figure 6.65 Formation of an industrially important silane, used as an agricultural adjuvant, in a hydrosilylation reaction catalyzed by an Fe–N2 complex at very low loadings.

Beyond olefin hydrofunctionalization, bis(imino)pyridine Fe–N2 complexes have been reported to mediate the catalytic hydrogenation of aryl azides (Figure 6.68a), the selective 3 H isotopic labeling of aromatic substrates (Figure 6.68b), and the reductive cyclization of 1,6 enynes and diynes (Figure 6.68c) [188–191]. It is noteworthy that it is the ability of this family of complexes to catalyze the [2π+2π] cycloaddition of alkenes, a powerful synthetic method difficult to achieve without metal catalysis [192]. Both the intramolecular cycloaddition of dienes (Figure 6.69a) and the intermolecular cycloaddition of alkenes (Figure 6.69b) have been achieved, with the tolerance of a moderate range of functional groups demonstrated [193–195]. A detailed mechanistic study of the cycloaddition of diallyl amines revealed a likely mechanism involving the coordination of each alkene fragment to the metal center, followed by metallacycle formation and reductive elimination to furnish a ring-closed cyclobutane product (Figure 6.70) [100]. Group 8 TM–N2 complexes bearing ancillary ligand sets other than a bis (imino)pyridine framework have proven active precatalysts for other transformations. Peters and coworkers have utilized the tris(phosphino)silyl-supported M–N2 complexes extensively explored for N2 functionalization (vide supra) to affect an unusual nitrene coupling. Aryl azides could be converted to the corresponding azoarenes by either [(SiPi3Pr )Fe(N2 )] (Figure 6.71a) or [(SiPi3Pr )Ru(N2 )] (Figure 6.71b), albeit with high catalyst loadings [196, 197]. Meanwhile, the related complex [(BPi Pr 3 )Fe(N2 )] is a moderately active precatalyst for the hydrogenation of ethylene, with the resting state of the system proposed to be an Fe–borohydrido–hydride species containing a labile H2 ligand (Figure 6.72) [198].

321

322

6 Group 8 Transition Metal–Dinitrogen Complexes

(a)

5 mol% [Fe]

O

O 50% yield

23 °C, 4 atm H2 24 hours

OEt

OEt

N N Ar

N N

[Fe] =

Fe N

N

N

Ar Ar= 2,6-C6H3(iPr)2

(b)

5 mol% [Fe]

O

O > 95% yield

23 °C, 4 atm H2 1.5 hours

OEt

OEt

N N N

Ar

N

[Fe] =

Ar

Fe N

N

N

N

Ar

Fe N N Ar

N

N

Ar = 2,6-C6H3(Me)2

Figure 6.66 Differing rates in the hydrogenation of an α,β-unsaturated ester by (a) a mononuclear Fe–N2 complex; and (b) a dinuclear Fe–N2 complex containing a less bulky ligand. 5 mol% [Fe] R

> 95%

R

23 °C, 4 atm H2

Substrate N

N Ar

N

[Fe] =

Ph

Fe

N

N N Ar

Time (h)a

Ph

1

N N

1

N

Ar = 2,6-C6H3(Me)2

12

Figure 6.67 Some examples of catalytic olefin hydrogenation by an Fe–N2 NHC complex.

6.4 Catalytic Reactions of Group 8 Transition Metal–Dinitrogen Complexes

(a)

NH2

N3 10 mol% [Fe] 23 °C, 4 atm H2 6 hours

iPr

iPr

100% conversion i

i

Pr

Pr

N N Ar

N N

[Fe] =

Fe N

N

N

Ar Ar = 2,6-C6H3(iPr)2 CF3

(b)

CF3

1 mol% [Fe] 45 °C, 4 atm 3H2 24 hours

90% yield 3H

3H 3

N

H

N Ar

N

[Fe] =

Fe

N

N

N N

N

N

Ar

Ar= 2,6-C6H3(iPr)2 (c)

H

5 mol% [Fe]

NtBu

23 °C, 4 atm H2 3 hours

NtBu

68% yield

N N N

Ar

N

[Fe] =

Fe N

N

Ar

N

Ar = 2,6-C6H3(iPr)2

Figure 6.68 Representative examples of some transformations catalyzed by bis(imino)pyridine and NHC-supported Fe–N2 complexes. (a) Hydrogenation of aryl azides; (b) isotopic labeling of aromatic substrates; and (c) reductive cyclization of 1,6 enynes.

323

324

6 Group 8 Transition Metal–Dinitrogen Complexes

(a) 10 mol% [Fe]

NtBu

NtBu

23 °C, 5 minutes >95% conversion N N Ar

N N

[Fe] =

Fe N

N

N

Ar Ar = 2,6-C6H3(iPr)2

(b)

2.5 mol% [Fe]

+

95% yield

23 °C, 16 hours N N N

Ar

N

[Fe] =

Ar

Fe N

N

N N

Ar

Fe N N Ar

N

N

Ar = 2,6-C6H3(Me)2

Figure 6.69 Representative examples of intramolecular (a) and intermolecular (b) cycloaddition reactions catalyzed by bis(imino)pyridine Fe–N2 complexes.

A phosphinothiolate-supported Fe–N2 complex can catalyze the selective 1,2 hydroboration of a wide range of substituted pyridines under mild conditions (some examples illustrated in Figure 6.73) [70]. Finally, a Rh/Ru tetranuclear heterobimetallic complex featuring a Ru–N2 unit has proven capable of catalyzing the addition of CCl4 to alkenes at very low catalyst loadings and with remarkable moisture tolerance (Figure 6.74) [65]. The extremely high activity and robustness of this complex distinguishes it from most other group 8 TM–N2 complexes employed for catalysis to date and points toward the possibility of the wider application of such species in the research community.

6.4 Catalytic Reactions of Group 8 Transition Metal–Dinitrogen Complexes

N

Ar

N

NtBu

Fe N

N

– 2N2

N

NtBu

Ar

Ar

N N Fe N

N N

Ar Ar = 2,6-C6H3(iPr)2

+ 2N2 N

Ar

N Fe

tBu

N

N Ar

NtBu

Figure 6.70 Proposed catalytic cycle for the cyclization of dienes mediated by a bis(imino)pyridine Fe–N2 complex. (a) 5 mol% [Fe] MeO

N3

MeO

N

[Fe] =

P

(b) MeO

25 mol% [Ru]

N3

23 °C, N2 atm < 1 minutes

[Ru] =

MeO

N N

OMe

93% yield

N N P Ru P Si

OMe

50% yield

N N P Fe P Si

N

70 °C, N2 atm < 3 hours

P

Figure 6.71 Catalytic conversion of aryl azides to azoarenes by Fe–N2 and Ru–N2 complexes.

325

326

6 Group 8 Transition Metal–Dinitrogen Complexes

(a) H2C

0.03 mol% [Fe]

CH2

CH3

H3C

23 °C, 1 atm H2 < 5 minutes

(b)

[Fe] =

P Fe P

+ C2H4 N N P Fe P P B

H H P Fe H P H P B

H2 atm –N2

B

– H2

H

H P

Migratory insertion + 2H2 P Fe P H P B

– C2H6

Figure 6.72 (a) Catalytic ethylene hydrogenation with [SiPiPr 3 ]Fe(N2 ) and (b) a proposed catalytic mechanism. R

R

1 mol% [Fe] 50 °C, 24 hours

N

H

N Bpin R

Yield (%)

H

61

p-Me

91

p-CF3

99

m-F

86

m-Ph

98

Cp*

[Fe] =

Fe

S

Ph2P

N PPh2

N S

Fe Cp*

Figure 6.73 Representative examples of catalytic hydroboration of pyridines by a phosphinothiolate Fe–N2 complex.

R

+

CCl4

60 °C

R Substrate

O Ph

Cl P Ph Cl P Rh Ru Cl N cat. = Ph Ph N Ru P P

Cl

0.003 mol% cat.

Ph Ph Cl Rh Cl Cl Ph

CCl3

Time (h) Yield (%) 1

98

1

93

24

75

O

O Ph

Ph

MeO C6H17

Figure 6.74 Addition of CCl4 to alkenes catalyzed by a Rh/Ru–N2 heterobimetallic complex at very low loadings.

6.5 Conclusion and Perspectives

6.5 Conclusion and Perspectives This chapter has hopefully provided the reader with a thorough overview of the chemistry of group 8 TM–N2 complexes. As has been established, the group 8 metals collectively have particular importance in the fixation of dinitrogen, both in the context of historical and contemporary research. Biologically, Fe plays a key role in the action of the nitrogenase enzyme, whereas catalysts based on both Fe and Ru are used for the industrial production of NH3 . The first documented TM–N2 complex was based on Ru, and complexes based on Fe have yielded some of the most important advances in the understanding (a)

–

Ar

PtBu2

R Fe

Fe R

3.3 equiv NH3

Fe

N

N

P P

R

N P Fe P E

N

14.3 equiv NH3 (R = H) 22.7 equiv NH3 (R = Me)

PtBu2

R

N

N N P M P Si

P

P

PtBu2

N Ar

(b)

–

N

N

PtBu2

P = PiPr2

P = PiPr2

88 equiv NH3 (E = B) 47 equiv NH3 (E = C)

4.3 equiv NH3 (M = Ru) 9.3 equiv NH3 (M = Fe) 120 equiv NH3 (M = Os)

P Fe

N

N

P P = PEt2

2.4 equiv N2H4 (R = H) 1.9 equiv N2H4 (R = Me)

24.5 equiv N2H4

(c)

Ar

N

N P′

N

PMe3 R

Fe

Si

Fe N

N

Ph N

P P

H N Ar

P′

Fe

Ph Ph P Ph P Ph

N Ph

P′ = PCy2 24.4 equiv N(SiMe3)3

15 equiv N(SiMe3)3 (R = Me) 26 equiv N(SiMe3)3 (R = Ph)

65 equiv N(SiMe3)3

Figure 6.75 Group 8 complexes for catalytic N2 fixation. (a) NH3 production catalysts; (b) N2 H4 production catalysts; and (c) N(SiMe3 )3 production catalysts. Counterions omitted throughout for clarity.

327

328

6 Group 8 Transition Metal–Dinitrogen Complexes

of homogeneous N2 fixation. Although Os has been a rarer focus of interest, an Os–N2 complex has recently emerged as a pre-eminent catalyst for N2 to NH3 conversion (see Figure 6.75 for a summary of group 8 systems capable of catalytic N2 fixation) [20, 25, 29, 63, 93, 95, 111, 178, 199]. In the wider field of catalysis, group 8 TM–N2 complexes, particularly bis(imino)pyridine-supported Fe–N2 species, have proven adept at performing a number of transformations. The diversity of catalysis achieved with group 8 TM–N2 complexes, as well as the ever-improving performance of these metals in the fixation of N2 , is testament to the vibrancy of research in this field. Many important advancements in the synthesis, utility, and understanding of group 8 TM–N2 species have emerged within the last decade and undoubtedly produce many more exciting developments that will be seen in the future.

References 1 Ober, J.A. (2016). Mineral Commodity Summaries 2016. Washington, D.C.:

U.S. Department of the Interior and U.S. Geological Survey. 2 Rayment, T., Schlögl, R., Thomas, J.M., and Ertl, G. (1985). Nature 315:

311–313. 3 Hu, Y. and Ribbe, M.W. (2015). J. Biol. Inorg. Chem. 20: 435–445. 4 Patil, B.S., Hessel, V., Seefeldt, L.C. et al. (2000). Nitrogen fixation. In: Ull-

5 6 7

8 9 10 11 12 13 14 15 16 17 18

mann’s Encyclopedia of Industrial Chemistry (ed. in chief (ed. B. Elvers). Weinheim: Wiley-VCH https://doi.org/10.1002/14356007.a17_471.pub2. Allen, A.D. and Senoff, C.V. (1965). Chem. Commun. 1: 621–622. Strait, R. (1999). Nitrogen & Methanol 238: 37–43. Nielsen, S.E. (2009). Ammonia synthesis: catalyst and technologies. In: Innovations in Industrial and Engineering Chemistry, ACS Symposium Series 1000, Chapter 2. , 15–39. Washington, D.C.: American Chemical Society. Dow Jones & Company (2002). Nitrogen & Methanol 257: 34–49. Leigh, G.J. (2004). The World’s Greatest Fix: A History of Nitrogen and Agriculture. Oxford: Oxford University Press. Allen, A.D. and Stevens, J.R. (1967). Chem. Commun. 3: 1147. Sacco, A. and Aresta, M. (1968). Chem. Commun. 4: 1223–1224. Chatt, J., Nikolsky, A.B., Richards, R.L. et al. (1970). J. Chem. Soc. A 1479–1483. Bianco, V.D., Doronzo, S., and Aresta, M. (1972). J. Organomet. Chem. 42: C63–C64. Bianco, V.D., Doronzo, S., and Rossi, M. (1972). J. Organomet. Chem. 35: 337–339. Leigh, G.J. and Jimenez-Tenorio, M. (1991). J. Am. Chem. Soc. 113: 5862–5863. Hirano, M., Akita, M., Morikita, T. et al. (1997). J. Chem. Soc., Dalton Trans. 26: 3453–3458. Field, L.D., Hazari, N., and Li, H.L. (2015). Inorg. Chem. 54: 4768–4776. Bianchini, C., Peruzzini, M., Zanobini, F. et al. (1991). J. Am. Chem. Soc. 113: 5453–5454.

References

19 Barbaro, P., Bianchini, C., Peruzzini, M. et al. (1994). Inorg. Chim. Acta 220: 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

5–19. Fajardo, J. Jr., and Peters, J.C. (2017). J. Am. Chem. Soc. 139: 16105–16108. Anderson, J.S., Rittle, J., and Peters, J.C. (2013). Nature 501: 84–87. Creutz, S.E. and Peters, J.C. (2014). J. Am. Chem. Soc. 136: 1105–1115. Chalkley, M.J., Del Castillo, T.J., Matson, B.D. et al. (2017). ACS Cent. Sci. 3: 217–223. Kuriyama, S., Arashiba, K., Nakajima, K. et al. (2016). Nat. Commun. 7: 12181. Hill, P.J., Doyle, L.R., Crawford, A.D. et al. (2016). J. Am. Chem. Soc. 138: 13521–13524. Rodriguez, M.M., Bill, E., Brennessel, W.W., and Holland, P.L. (2011). Science 334: 780–783. MacLeod, K.C., Vinyard, D.J., and Holland, P.L. (2014). J. Am. Chem. Soc. 136: 10226–10229. MacLeod, K.C., McWilliams, S.F., Mercado, B.Q., and Holland, P.L. (2016). Chem. Sci. 7: 5736–5746. Prokopchuk, D.E., Wiedner, E.S., Walter, E.D. et al. (2017). J. Am. Chem. Soc. 139: 9291–9301. Giannoccaro, P., Rossi, M., and Sacco, A. (1972). Coord. Chem. Rev. 8: 77–79. Liang, Q. and Song, D. (2017). Inorg. Chem. 56: 11956–11970. Saouma, C.T., Moore, C.E., Rheingold, A.L., and Peters, J.C. (2011). Inorg. Chem. 50: 11285–11287. Ohki, Y., Hatanaka, T., and Tatsumi, K. (2008). J. Am. Chem. Soc. 130: 17174–17186. Aresta, M., Giannoccaro, P., Rossi, M., and Sacco, A. (1971). Inorg. Chim. Acta 5: 203–206. Aresta, M., Giannoccaro, P., Rossi, M., and Sacco, A. (1971). Inorg. Chim. Acta 5: 115–118. Eliades, T.I., Harris, R.O., and Zia, M.C. (1970). J. Chem. Soc. D 6: 1709. Harris, R.O., Hota, N.K., Sadavoy, L., and Yuen, J.M.C. (1973). J. Organomet. Chem. 54: 259–264. Christ, M.L., Sabo-Etienne, S., Chung, G., and Chaudret, B. (1994). Inorg. Chem. 33: 5316–5319. Abdur-Rashid, K., Gusev, D.G., Lough, A.J., and Morris, R.H. (2000). Organometallics 19: 1652–1660. Chaudret, B., Devillers, J., and Poilblanc, R. (1985). Organometallics 4: 1727–1732. Lau, S., Ward, B., Zhou, X. et al. (2017). Organometallics 36: 3654–3663. Bianchini, C., Perez, P.J., Peruzzini, M. et al. (1991). Inorg. Chem. 30: 279–287. Namura, K. and Suzuki, H. (2014). Organometallics 33: 2968–2983. Hills, A., Hughes, D.L., Jimenez-Tenorio, M., and Leigh, G.J. (1990). J. Organomet. Chem. 391: C41–C44. Lee, J.-H., Pink, M., Tomaszewski, J. et al. (2007). J. Am. Chem. Soc. 129: 8706–8707.

329

330

6 Group 8 Transition Metal–Dinitrogen Complexes

46 Heiden, Z.M., Chen, S., Labios, L.A. et al. (2014). Organometallics 33:

1333–1336. 47 Bancroft, G.M., Mays, M.J., Prater, B.E., and Stefanini, F.P. (1970). J. Chem.

Soc. A 2146–2149. 48 Stepowska, E., Jiang, H., and Song, D. (2010). Chem. Commun. 46: 556–558. 49 Beach, M.T., Walker, J.M., Wang, R., and Spivak, G.J. (2011). J. Organomet.

Chem. 696: 3198–3205. 50 Bellerby, J.M., Mays, M.J., and Sears, P.L. (1976). J. Chem. Soc., Dalton Trans.

5: 1232–1236. 51 Hughes, D.L., Leigh, G.J., Jimenez-Tenorio, M., and Rowley, A.T. (1993). J.

Chem. Soc., Dalton Trans. 22: 75–82. 52 de la Jara Leal, A., Jiménez Tenorio, M., Puerta, M.C., and Valerga, P. (1995).

Organometallics 14: 3839–3847. 53 Kandler, H., Gauss, C., Bidell, W. et al. (1995). Chem. Eur. J. 1: 541–548. 54 Heiden, Z.M., Chen, S., Mock, M.T. et al. (2013). Inorg. Chem. 52:

4026–4039. 55 Gilbertson, J.D., Szymczak, N.K., Crossland, J.L. et al. (2007). Inorg. Chem.

46: 1205–1214. 56 Burling, S., Häller, L.J.L., Mas-Marzá, E. et al. (2009). Chem. Eur. J. 15:

10912–10923. 57 Fernández, F.E., Puerta, M.C., and Valerga, P. (2013). Inorg. Chem. 52:

4396–4410. 58 Sellmann, D., Hautsch, B., Rösler, A., and Heinemann, F.W. (2001). Angew.

Chem. Int. Ed. 40: 1553–1555. 59 Sellmann, D., Hille, A., Rösler, A. et al. (2004). Chem. Eur. J. 10: 819–830. 60 Ledger, A.E.W., Moreno, A., Ellul, C.E. et al. (2010). Inorg. Chem. 49:

7244–7256. 61 Amoroso, D., Yap, G.P.A., and Fogg, D.E. (2001). Can. J. Chem. 79: 958–963. 62 Tskhovrebov, A.G., Solari, E., Scopelliti, R., and Severin, K. (2012).

Organometallics 31: 7235–7240. 63 Imayoshi, R., Nakajima, K., Takaya, J. et al. (2017). Eur. J. Inorg. Chem. 32:

3769–3778. 64 Bianchini, C., Linn, K., Masi, D. et al. (1993). Inorg. Chem. 32: 2366–2376. 65 Quebatte, L., Scopelliti, R., and Severin, K. (2004). Angew. Chem. Int. Ed. 43:

1520–1524. 66 Silverthorn, W.E. (1971). J. Chem. Soc. D 7: 1310–1311. 67 Sellmann, D. and Kleinschmidt, E. (1975). Angew. Chem. Int. Ed. Engl. 14:

571. 68 Yamamoto, A., Miura, Y., Ito, T. et al. (1983). Organometallics

2: 1429–1436. 69 Sunada, Y., Imaoka, T., and Nagashima, H. (2010). Organometallics 29:

6157–6160. 70 Zhang, F., Song, H., Zhuang, X. et al. (2017). J. Am. Chem. Soc. 139:

17775–17778. 71 Kizaki, T., Matsumoto, T., and Ogo, S. (2010). Dalton Trans. 39: 1339–1344. 72 Takahashi, T., Hiratani, K., and Kimura, E. (1993). Chem. Lett. 22:

1329–1332.

References

73 Camenzind, M.J., James, B.R., and Dolphin, D. (1986). J. Chem. Soc., Chem.

Commun. 22: 1137–1139. 74 Camenzind, M.J., James, B.R., Dolphin, D. et al. (1988). Inorg. Chem. 27:

3054–3057. 75 Komiya, S., Akita, M., Yoza, A. et al. (1993). J. Chem. Soc., Chem. Commun.

29: 787–788. 76 Whited, M.T., Mankad, N.P., Lee, Y. et al. (2009). Inorg. Chem. 48:

2507–2517. 77 Russell, S.K., Darmon, J.M., Lobkovsky, E., and Chirik, P.J. (2010). Inorg.

Chem. 49: 2782–2792. 78 Smith, J.M., Lachicotte, R.J., Pittard, K.A. et al. (2001). J. Am. Chem. Soc.

123: 9222–9223. 79 Yu, R.P., Darmon, J.M., Hoyt, J.M. et al. (2012). ACS Catal. 2: 1760–1764. 80 Moret, M.-E. and Peters, J.C. (2011). Angew. Chem. Int. Ed. 50: 2063–2067. 81 van de Watering, F.F., van der Vlugt, J.I., Dzik, W.I. et al. (2017). Chem. Eur.

J. 23: 12709–12713. 82 Gilbert-Wilson, R., Field, L.D., Colbran, S.B., and Bhadbhade, M. (2013).

Inorg. Chem. 52: 3043–3053. 83 Gilbert-Wilson, R., Field, L.D., and Bhadbhade, M. (2014). Inorg. Chem. 53:

12469–12479. 84 Takaoka, A., Gerber, L.C.H., and Peters, J.C. (2010). Angew. Chem. Int. Ed.

49: 4088–4091. 85 Suzuki, T., Wasada-Tsutsui, Y., Ogawa, T. et al. (2015). Inorg. Chem. 54:

9271–9281. 86 McSkimming, A. and Harman, W.H. (2015). J. Am. Chem. Soc. 137:

8940–8943. 87 Ouyang, Z., Cheng, J., Li, L. et al. (2016). Chem. Eur. J. 22: 14162–14165. 88 Cummins, D.C., Yap, G.P.A., and Theopold, K.H. (2016). Eur. J. Inorg. Chem. 89 90 91 92 93 94 95 96 97 98 99

2349–2356. ˇ c, I., Mercado, B.Q., Bill, E. et al. (2015). Nature 526: 96. Cori´ ˇ c, I. and Holland, P.L. (2016). J. Am. Chem. Soc. 138: 7200–7211. Cori´ Higuchi, J., Kuriyama, S., Eizawa, A. et al. (2017). Dalton Trans. https://doi .org/10.1039/C7DT04327a. Rittle, J. and Peters, J.C. (2013). Proc. Natl. Acad. Sci. U.S.A. 110: 15898–15903. Ung, G. and Peters, J.C. (2015). Angew. Chem. Int. Ed. 54: 532–535. Doyle, L.R., Hill, P.J., Wildgoose, G.G., and Ashley, A.E. (2016). Dalton Trans. 45: 7550–7554. Sekiguchi, Y., Kuriyama, S., Eizawa, A. et al. (2017). Chem. Commun. 53: 12040–12043. Betley, T.A. and Peters, J.C. (2003). J. Am. Chem. Soc. 125: 10782–10783. Bart, S.C., Lobkovsky, E., and Chirik, P.J. (2004). J. Am. Chem. Soc. 126: 13794–13807. Mankad, N.P., Whited, M.T., and Peters, J.C. (2007). Angew. Chem. Int. Ed. 46: 5768–5771. Stieber, S.C.E., Milsmann, C., Hoyt, J.M. et al. (2012). Inorg. Chem. 51: 3770–3785.

331

332

6 Group 8 Transition Metal–Dinitrogen Complexes

100 Hoyt, J.M., Sylvester, K.T., Semproni, S.P., and Chirik, P.J. (2013). J. Am.

Chem. Soc. 135: 4862–4877. 101 Danopoulos, A.A., Wright, J.A., and Motherwell, W.B. (2005). Chem. Com-

mun. 41: 784–786. 102 Creutz, S.E. and Peters, J.C. (2017). Chem. Sci. 8: 2321–2328. 103 Rittle, J., McCrory, C.C.L., and Peters, J.C. (2014). J. Am. Chem. Soc. 136:

13853–13862. 104 Creutz, S.E. and Peters, J.C. (2015). J. Am. Chem. Soc. 137: 7310–7313. 105 Chatt, J., Leigh, G.J., and Richards, R.L. (1969). J. Chem. Soc. D 5:

515b–516b. 106 Chatt, J., Leigh, G.J., and Richards, R.L. (1970). J. Chem. Soc. A 2243–2246. 107 Richardson, D.E., Sen, J.P., Buhr, J.D., and Taube, H. (1982). Inorg. Chem. 21:

3136–3140. 108 Stoppioni, P., Mani, F., and Sacconi, L. (1974). Inorg. Chim. Acta 11:

227–230. 109 MacBeth, C.E., Harkins, S.B., and Peters, J.C. (2005). Can. J. Chem. 83:

332–340. 110 Takaoka, A., Mankad, N.P., and Peters, J.C. (2011). J. Am. Chem. Soc. 133:

8440–8443. 111 Buscagan, T.M., Oyala, P.H., and Peters, J.C. (2017). Angew. Chem. Int. Ed.

56: 6921–6926. 112 Ghilardi, C.A., Midollini, S., Sacconi, L., and Stoppioni, P. (1981). J.

Organomet. Chem. 205: 193–202. 113 Hounjet, L.J., Adhikari, D., Pink, M. et al. (2015). Z. Anorg. Allg. Chem. 641:

45–48. 114 Gusev, D.G., Dolgushin, F.M., and Antipin, M.Y. (2000). Organometallics 19:

3429–3434. 115 Gilbertson, J.D., Szymczak, N.K., and Tyler, D.R. (2005). J. Am. Chem. Soc.

127: 10184–10185. 116 Field, L.D., Li, H.L., Dalgarno, S.J., and McIntosh, R.D. (2013). Inorg. Chem.

52: 1570–1583. 117 Scott, J., Vidyaratne, I., Korobkov, I. et al. (2008). Inorg. Chem. 47: 896–911. 118 Demadis, K.D., El-Samanody, E.-S., Meyer, T.J., and White, P.S. (1998). Inorg.

Chem. 37: 838–839. 119 Lee, Y., Mankad, N.P., and Peters, J.C. (2010). Nat. Chem. 2: 558–565. 120 Demadis, K.D., Meyer, T.J., and White, P.S. (1997). Inorg. Chem. 36:

5678–5679. 121 Knoth, W.H. Jr., (1968). J. Am. Chem. Soc. 90: 7172–7173. 122 Buhr, J.D. and Taube, H. (1980). Inorg. Chem. 19: 2425–2434. 123 Huynh, M.H.V., Meyer, T.J., Hiskey, M.A., and Jameson, D.L. (2004). J. Am.

Chem. Soc. 126: 3608–3615. 124 Allen, A.D., Bottomley, F., Harris, R.O. et al. (1967). J. Am. Chem. Soc. 89:

5595–5599. 125 Harrison, D.F., Weissberger, E., and Taube, H. (1968). Science 159: 320–322. 126 Diamantis, A.A. and Dubrawski, J.V. (1981). Inorg. Chem. 20: 1142–1150. 127 Buchler, J.W. and Smith, P.D. (1974). Angew. Chem. Int. Ed. Engl. 13: 745.

References

128 Osman, R., Pattison, D.I., Perutz, R.N. et al. (1997). J. Am. Chem. Soc. 119:

8459–8473. 129 Whittlesey, M.K., Mawby, R.J., Osman, R. et al. (1993). J. Am. Chem. Soc.

115: 8627–8637. 130 Yu, Y., Sadique, A.R., Smith, J.M. et al. (2008). J. Am. Chem. Soc. 130:

6624–6638. 131 Buhr, J.D. and Taube, H. (1979). Inorg. Chem. 18: 2208–2212. 132 Douglas, P.G., Feltham, R.D., and Metzger, H.G. (1970). J. Chem. Soc. D 6:

889–890. 133 Coia, G.M., Devenney, M., White, P.S. et al. (1997). Inorg. Chem. 36:

2341–2351. 134 Bart, S.C., Bowman, A.C., Lobkovsky, E., and Chirik, P.J. (2007). J. Am. 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154

Chem. Soc. 129: 7212–7213. Betley, T.A. and Peters, J.C. (2004). J. Am. Chem. Soc. 126: 6252–6254. Henderson, R.A. (1988). J. Chem. Soc., Dalton Trans. 17: 515–520. Bart, S.C., Lobkovsky, E., Bill, E. et al. (2007). Inorg. Chem. 46: 7055–7063. Darmon, J.M., Yu, R.P., Semproni, S.P. et al. (2014). Organometallics 33: 5423–5433. Sadique, A.R., Brennessel, W.W., and Holland, P.L. (2008). Inorg. Chem. 47: 784–786. Bowman, A.C., Bart, S.C., Heinemann, F.W. et al. (2009). Inorg. Chem. 48: 5587–5589. Russell, S.K., Hoyt, J.M., Bart, S.C. et al. (2014). Chem. Sci. 5: 1168–1174. Darmon, J.M., Stieber, S.C.E., Sylvester, K.T. et al. (2012). J. Am. Chem. Soc. 134: 17125–17137. Cowley, R.E., Elhaïk, J., Eckert, N.A. et al. (2008). J. Am. Chem. Soc. 130: 6074–6075. Eckert, N.A., Vaddadi, S., Stoian, S. et al. (2006). Angew. Chem. Int. Ed. 45: 6868–6871. Bart, S.C., Chłopek, K., Bill, E. et al. (2006). J. Am. Chem. Soc. 128: 13901–13912. Brown, S.D. and Peters, J.C. (2004). J. Am. Chem. Soc. 126: 4538–4539. Kunkely, H. and Vogler, A. (2010). Angew. Chem. Int. Ed. 49: 1591–1593. Kunkely, H. and Vogler, A. (2012). Inorg. Chim. Acta 391: 229–231. Smith, J.M., Sadique, A.R., Cundari, T.R. et al. (2006). J. Am. Chem. Soc. 128: 756–769. Figg, T.M., Holland, P.L., and Cundari, T.R. (2012). Inorg. Chem. 51: 7546–7550. Chiang, K.P., Bellows, S.M., Brennessel, W.W., and Holland, P.L. (2014). Chem. Sci. 5: 267–274. MacLeod, K.C., Menges, F.S., McWilliams, S.F. et al. (2016). J. Am. Chem. Soc. 138: 11185–11191. Grubel, K., Brennessel, W.W., Mercado, B.Q., and Holland, P.L. (2014). J. Am. Chem. Soc. 136: 16807–16816. McWilliams, S.F., Rodgers, K.R., Lukat-Rodgers, G. et al. (2016). Inorg. Chem. 55: 2960–2968.

333

334

6 Group 8 Transition Metal–Dinitrogen Complexes

155 Dugan, T.R., MacLeod, K.C., Brennessel, W.W., and Holland, P.L. (2013). Eur.

J. Inorg. Chem. 3891–3897. 156 Lee, Y., Sloane, F.T., Blondin, G. et al. (2015). Angew. Chem. Int. Ed. 54:

1499–1503. 157 Hidai, M., Komori, K., Kodama, T. et al. (1984). J. Organomet. Chem. 272:

155–167. 158 Moret, M.-E. and Peters, J.C. (2011). J. Am. Chem. Soc. 133: 18118–18121. 159 Suess, D.L.M. and Peters, J.C. (2013). J. Am. Chem. Soc. 135: 4938–4941. 160 Piascik, A.D., Hill, P.J., Crawford, A.D. et al. (2017). Chem. Commun. 53:

7657–7660. 161 Rittle, J. and Peters, J.C. (2016). J. Am. Chem. Soc. 138: 4243–4248. 162 Anderson, J.S., Cutsail, G.E., Rittle, J. et al. (2015). J. Am. Chem. Soc. 137:

7803–7809. 163 Chatt, J., Crabtree, R.H., and Richards, R.L. (1972). J. Chem. Soc., Chem.

Commun. 8: 534. 164 Geri, J.B., Shanahan, J.P., and Szymczak, N.K. (2017). J. Am. Chem. Soc. 139:

5952–5956. 165 Annibale, V.T. and Song, D. (2012). Chem. Commun. 48: 5416–5418. 166 Hoffman, B.M., Lukoyanov, D., Dean, D.R., and Seefeldt, L.C. (2013). Acc. 167 168 169 170 171 172 173 174 175 176 177 178 179 180

Chem. Res. 46: 587–595. Dilworth, M.J., Fisher, K., Kim, C.-H., and Newton, W.E. (1998). Biochemistry 37: 17495–17505. Hall, D.A. and Leigh, G.J. (1996). J. Chem. Soc., Dalton Trans. 3539–3541. Crossland, J.L., Zakharov, L.N., and Tyler, D.R. (2007). Inorg. Chem. 46: 10476–10478. Field, L.D., Li, H.L., Dalgarno, S.J., and Turner, P. (2008). Chem. Commun. 44: 1680–1682. Field, L.D., Li, H.L., and Magill, A.M. (2009). Inorg. Chem. 48: 5–7. Crossland, J.L., Balesdent, C.G., and Tyler, D.R. (2009). Dalton Trans. 38: 4420–4422. Yelle, R.B., Crossland, J.L., Szymczak, N.K., and Tyler, D.R. (2009). Inorg. Chem. 48: 861–871. Field, L.D., Li, H.L., Dalgarno, S.J. et al. (2011). Inorg. Chem. 50: 5468–5476. Crossland, J.L., Balesdent, C.G., and Tyler, D.R. (2012). Inorg. Chem. 51: 439–445. Balesdent, C.G., Crossland, J.L., Regan, D.T. et al. (2013). Inorg. Chem. 52: 14178–14187. Thompson, N.B., Green, M.T., and Peters, J.C. (2017). J. Am. Chem. Soc. 139: 15312–15315. Del Castillo, T.J., Thompson, N.B., and Peters, J.C. (2016). J. Am. Chem. Soc. 138: 5341–5350. Rittle, J. and Peters, J.C. (2017). J. Am. Chem. Soc. 139: 3161–3170. Bezdek, M.J., Pappas, I., and Chirik, P.J. (2017). Determining and understanding N—H bond strengths in synthetic nitrogen fixation cycles. In: Nitrogen Fixation, Topics in Organometallic Chemistry, vol. 60 (ed. Y. Nishibayashi), 1–21. Cham: Springer International Publishing.

References

181 Anderson, J.S., Moret, M.-E., and Peters, J.C. (2013). J. Am. Chem. Soc. 135:

534–537. 182 Tanabe, Y. and Nishibayashi, Y. (2013). Coord. Chem. Rev. 257: 2551–2564. 183 Bianchini, C., Meli, A., Peruzzini, M. et al. (1989). Organometallics 8:

2080–2082. 184 Obligacion, J.V. and Chirik, P.J. (2013). Org. Lett. 15: 2680–2683. 185 Atienza, C.C.H., Tondreau, A.M., Weller, K.J. et al. (2012). ACS Catal. 2:

2169–2172. 186 Tondreau, A.M., Atienza, C.C.H., Weller, K.J. et al. (2012). Science 335:

567–570. 187 Trovitch, R.J., Lobkovsky, E., Bill, E., and Chirik, P.J. (2008). Organometallics

27: 1470–1478. 188 Bart, S.C., Lobkovsky, E., Bill, E., and Chirik, P.J. (2006). J. Am. Chem. Soc.

128: 5302–5303. 189 Yu, R.P., Hesk, D., Rivera, N. et al. (2016). Nature 529: 195–199. 190 Sylvester, K.T. and Chirik, P.J. (2009). J. Am. Chem. Soc. 131: 8772–8774. 191 Yu, R.P., Darmon, J.M., Semproni, S.P. et al. (2017). Organometallics 36:

4341–4343. 192 Lautens, M., Klute, W., and Tam, W. (1996). Chem. Rev. 96: 49–92. 193 Bouwkamp, M.W., Bowman, A.C., Lobkovsky, E., and Chirik, P.J. (2006). J.

Am. Chem. Soc. 128: 13340–13341. 194 Russell, S.K., Lobkovsky, E., and Chirik, P.J. (2011). J. Am. Chem. Soc. 133:

8858–8861. 195 Hoyt, J.M., Schmidt, V.A., Tondreau, A.M., and Chirik, P.J. (2015). Science

349: 960–963. 196 Mankad, N.P., Müller, P., and Peters, J.C. (2010). J. Am. Chem. Soc. 132:

4083–4085. 197 Takaoka, A., Moret, M.-E., and Peters, J.C. (2012). J. Am. Chem. Soc. 134:

6695–6706. 198 Fong, H., Moret, M.-E., Lee, Y., and Peters, J.C. (2013). Organometallics 32:

3053–3062. 199 Tanabe, Y. and Nishibayashi, Y. (2016). Chem. Rec. 16: 1549–1577.

335

337

7 Group 9 Transition Metal–Dinitrogen Complexes Connie C. Lu and Steven D. Prinslow Department of Chemistry, University of Minnesota – Twin Cities, 207 Pleasant Street SE, Minneapolis, MN 55455, USA

This chapter provides comprehensive coverage of the N2 coordination chemistry of group 9 metal centers. Among the triad, and even among other transition metals, cobalt is emerging as a prominent player in both N2 coordination chemistry and N2 catalysis. The chapter is organized by the metal element (Co, Rh, and Ir) and then by various ligands. The contents span the synthesis, structure, and reactivity of the group 9 N2 complexes. The vast collection of the group 9 N2 compounds allows for interesting comparisons to be drawn and thereby provides critical insights into the underlying basis for the varying extents of N2 activation. In the last section, the group 9 catalysts for N2 functionalization are summarized.

7.1 Cobalt–Dinitrogen Complexes The cobalt–N2 complex, CoH(N2 )(PPh3 )3 , was reported in 1967 [1–3], shortly after the first N2 coordination complex, [Ru(N2 )(NH3 )5 ]2+ [4]. Although the use of simple phosphine ligands would drive Co–N2 coordination chemistry until the early 1980s, the subsequent two decades were relatively quiet for the field. Cobalt–N2 coordination chemistry gained momentum in the early 2000s as new ligand platforms fueled the discoveries of different coordination donor environments and geometries. Today, there are over a hundred isolated Co–N2 compounds. The vast majority of these compounds require a multidentate ligand, which is shown collectively in Figure 7.1. The discussion of Co–N2 compounds is organized according to the various ligand types. Compared to iron, cobalt is often considered a less reactive metal for N2 activation. Iron features prominently in the active sites of nitrogenases and is the most widely used metal catalyst in the Haber–Bosch process. When isostructural Fe– and Co–N2 complexes are known, the Co counterparts typically exhibit a lesser extent of N2 activation. This trend has been explained by cobalt’s greater electronegativity, which makes cobalt worse at π-backbonding into the N2 π*-orbitals. Consequently, stoichiometric N—H, N—Si, or N—C bond forming reactions of Co–N2 complexes were rare. However, since 2015, discoveries of Transition Metal-Dinitrogen Complexes: Preparation and Reactivity, First Edition. Edited by Yoshiaki Nishibayashi. © 2019 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2019 by Wiley-VCH Verlag GmbH & Co. KGaA.

338

7 Group 9 Transition Metal–Dinitrogen Complexes

cobalt-based catalysts for the reductive silylation of N2 to N(SiMe3 )3 and for the conversion of N2 to NH3 have propelled cobalt out from iron’s long shadow. One clear lesson is emerging from these cobalt systems: N2 catalytic activity does not correlate with the extent of N2 activation. Although it remains to be seen whether cobalt will become an important metal in future N2 catalysis, understanding how these catalysts function will benefit the development of even more efficient molecular catalysts for N2 conversion. 7.1.1

Monodentate Phosphine Donors

7.1.1.1

CoH(N2 )(PR3 )3 and Related Co(I) Complexes

The first cobalt–N2 complex was reported in 1965 using simple monodentate phosphine ligands. Yamamoto et al. isolated the Co(I) N2 adduct, CoH(N2 )(PPh3 )3 , from the reaction of Co(acac)3 , AlEt2 OEt, and PPh3 under a N2 atmosphere (Figure 7.2) [1]. The presence of N2 ligand was established by 7.1.2 Tripodal polyphosphine ligands R2P

R2P

iPr

PR2

2P

tppme, R = Ph tmpme, R = Me iPr

2P

iPr

M

N

N

iPr P 2

PiPr2

B

P

Ph PhBiPrP3

PP3

PiPr2

2P

PR2

R2P

N N

N

N M

R′

2P

iPr

2P

PiPr2

E

TPB, E = B SiP3, E = Si CP3, E = C

M = Ti, Zr, Hf, Nb, Ta i 2PNXyl, R = Pr, R′ = 3,5-Me2(C6H3) iPr PNMes, R = iPr, R′ = 2,4,6-Me (C H ) 2 3 6 2 iPr PNiPr, R = iPr, R′ = iPr 2 Ph2PNiPr, R = Ph, R′ = iPr iPr

R′ PR2 N R′

L/X

iPr

PPh2 Ph P PPh2 2

M(o-(iPr2PCH2N)C6H4)3N) M = Al, Ti, V, Cr, Co 7.1.3 Ligands with exclusively N-donors R R

N N

R′

iPr

iPr

R

iPr

N

N

N N R B H R′

N

iPr

iPr

N

R

TptBu,Me, R = tBu, R′ = Me R = Np, R′ = H TpNp, TpiPr,Me, R = iPr, R′ = Me TptBu, R = tBu, R′ = H

iPr iPr

Me

N

N N

iPr

Me

R iPrPDI

LtBu,dipp, R = tBu LMe,dipp, R = Me

7.1.4 N-heterocyclic carbene ligands N

N IMes

Cy

N

N

Cy

ICy

Figure 7.1 Ligands found in cobalt–N2 coordination complexes, excluding monodentate phosphines.

7.1 Cobalt–Dinitrogen Complexes 7.1.5 Pincer and pincer-type ligands iPr

PtBu2

PiPr2

2P

PiPr2

Me2Si

N

N

N

N

Me2Si

PiPr2

PtBu2 iPr,Ar

iPr

tBu,Si

PNP

iPr,Et

PNP

N

tBu,pyr

PNP

PiPr2 tBu2P

PiPr2 iPr2P

2P

PR2

R2P

Cy,pyr

PtBu2

PNP, R = tBu PNP, R = Cy PiPr2

B

N

N

iPr

N

N

N iPr py

R

P NP

iPr

H

4-R-iPrPpyNP

PBP

iPrPNpN

R iPr

iPr iPr

N N

N

R N

N

iPr

2P

N

iPr

N

N

N dipp

CCC, R = 2,6-iPr2(C6H3) CCC, R = 2,4,6-Me3(C6H2)

dipp

CpyNC

mes

Me

Me N

N

Me

Me

Me Ph

iPr

Si

N

Cy2P

Me

Me

Me

N

iPr

N

N

N Me

Me CSiC

PNC

7.1.6 Assorted ligands iPr

R

2P

B

PiPr2

DPB

S P

O CO2

S

SPS

C

N

O2C R O CNArmes,R = 2,4,6-Me3(C6H2) dipp i dobdc ,R = 2,6- Pr2(C6H3) CNAr

Figure 7.1 (Continued)

pyrolysis of the complex, which released 1.0 equiv of N2 , and by a characteristic IR band at 2088 cm−1 , which corresponded to the N–N stretching frequency (cf. 2359 cm−1 for free N2 [6]). Of interest, the hydride ligand was initially dismissed because no Co–H vibration band was detected, and hence, the complex was incorrectly assigned as “Co(N2 )(PPh3 )3 .” Soon thereafter, Misono et al. suggested the presence of the hydride ligand based on the compound’s diamagnetism and the evolution of 0.5 equiv of H2 during thermal decomposition [2]. The solid-state

339

340

7 Group 9 Transition Metal–Dinitrogen Complexes

N2 1.101(12) 178(2)°

N

1.784(13)

N Ph3P

Co H

(a)

N1

PPh3 PPh3

2.185(4) P1

(b)

a

b 2.167(4) c

P2 2.202(4)

Co1 1.6

P3 H

Figure 7.2 (a) CoH(N2 )(PPh3 )3 and (b) its molecular structure. One of two independent molecules is shown without phenyl groups. Source: Davis et al. 1969 [5]. Adapted with permission from the American Chemical Society.

structure by Ibers and coworkers further revealed an end-on, terminal N2 ligand that is positioned trans to the hydride with the N—N bond lengths of 1.10(1) and 1.12(1) Å (Figure 7.2, cf. 1.10 Å for free N2 ) [5, 7]. Also of relevance, Sacco and Rossi were able to prepare CoH(N2 )(PPh3 )3 by simply exposing CoH3 (PPh3 )3 to N2 [3]. More recently, CoH(N2 )(PPh3 )3 was synthesized from CoCl2 (PPh3 )2 , 2 equiv of NaBEt3 H, and 1 equiv PPh3 under a N2 atmosphere [8]. The latter route could be preferable as it avoids the use of ill-defined organoaluminum reagents. The lability of the N2 ligand in CoH(N2 )(PPh3 )3 facilitates reactivity by opening a coordination site at a Co(I) center [9]. The N2 ligand can be reversibly exchanged with both H2 and NH3 . Indeed, the reverse exchange of a NH3 ligand for N2 is a critical step in catalytic N2 fixation schemes because release of NH3 and rebinding of N2 are necessary for turnover. Attempts to initiate N—H bond formation by adding strong acids to CoH(N2 )(PPh3 )3 were unsuccessful and only resulted in the loss of N2 [10]. Arguably, the weak activation of N2 in CoH(N2 )(PPh3 )3 makes it a poor precursor for N2 functionalization processes. Nonetheless, the displacement of N2 from CoH(N2 )(PPh3 )3 generates a reactive species, “CoH(PPh3 )3 ,” which can undergo bond activation processes that are relevant to other catalytic applications. For example, CoH(N2 )(PPh3 )3 catalyzes the reduction of N2 O to N2 in the presence of PPh3 , which becomes transformed to the by-product (O)PPh3 [9]. CoH(N2 )(PPh3 )3 readily activates H—H, Si—H, and B—H bonds and further mediates their transfer to olefin substrates [8, 9, 11]. CoH(N2 )(PPh3 )3 has been employed as a precatalyst in various applications, including olefin hydrogenation, dimerization, isomerization, hydrosilylation, and hydroboration. An interesting feature of the CoH(N2 )(PPh3 )3 system is that terminal olefins are rapidly isomerized to internal olefins before the addition of HBpin (pinacolborane) in a reagent across the C=C bond [8]. Hence, the stereoselectivity of CoH(N2 )(PPh3 )3 is different from that of the most base metal catalysts, which typically affect hydroboration at the terminal position. The PMe3 analog, CoH(N2 )(PMe3 )3 , is significantly less stable than the PPh3 congener. A new synthetic route was devised, whereby the Co(–I) N2 adduct,

7.1 Cobalt–Dinitrogen Complexes C

C

C P

H Me3P Me3P Co

C

PMe3

Me3P

P C

P

P

N N

C

C

N

C

Co

C

N

Co

H

C

C C

P

C

Co

C

C

H

PMe3 PMe3

P

C

H C

(a)

(b)

C

Figure 7.3 (a) {CoH(PMe3 )3 }2 (μ-N2 ) and (b) its molecular structure (protons omitted). Source: Klein et al. 1991 [12]. Adapted with permission from https://creativecommons.org/licenses/by/ 3.0/.

K(Co(N2 )(PMe3 )3 ) (vide infra), is protonated with a weak acid, such as methanol, at −78 ∘ C in diethyl ether [12]. The isolated complex was identified by elemental analysis and X-ray crystallography to be the dicobalt–N2 -sandwich complex, {CoH(PMe3 )3 }2 (μ-N2 ), where the N2 ligand is bridging two Co(I) centers in an end-on manner and is in the trans position relative to hydrides (Figure 7.3). The absence of a N–N vibration in the IR spectrum is also consistent with a symmetric bridging N2 ligand. The N—N bond distance of 1.117(4) Å indicates weak activation. In solution, the dicobalt species is in equilibrium with the monocobalt complex, CoH(N2 )(PMe3 )3 . The latter exhibits a diagnostic N–N stretch at 2068 cm−1 . Hence, exchanging the PPh3 donors for the more electron-donating PMe3 effects a lowering of the N–N stretching frequency by 20 cm−1 , which is consistent with increased electron density at Co π-backbonding into the π*-orbitals of N2 . Of note, the instability of CoH(N2 )(PMe3 )3 likely stems from the decreased sterics of PMe3 , as the P(n-Bu)3 analog, CoH(N2 )(PBu3 )3 , was readily generated by adding P(n-Bu)3 to a benzene solution of CoH(N2 )(PPh3 )3 [9]. The formation of CoH(N2 )(PBu3 )3 was confirmed by a diagnostic IR band at 2048 cm−1 . Despite the instability of CoH(N2 )(PMe3 )3 , a related Co(I) N2 complex was generated from the CoCl2 (PMe3 )2 precursor. Jones et al. added t-Bu2 PLi to CoCl2 (PMe3 )2 and isolated the dicobalt(I,I) complex, {Co(μ-t-Bu2 P)N2 (PMe3 )}2 [13]. The X-ray structure reveals a Co2 P2 diamond core with terminal, end-on N2 ligands. The N—N bond length of 1.092(5) Å puts {Co(μ-t-Bu2 P)N2 (PMe3 )}2 on par with most of the Co(I) N2 adducts; however, the solution N–N frequency of 1910 cm−1 would indicate considerably stronger activation. An interesting structural feature is the short Co—Co bond distance of 2.414(1) Å, which the authors propose is reflective of a “long Co—Co double bond.” {Co(μ-t-Bu2 P)N2 (PMe3 )}2 is unstable in solution and decomposes to an ill-defined mixture of species. The only other reactivity is ligand exchange of N2 with CO to yield isostructural {Co(μ-t-Bu2 P)CO(PMe3 )}2 .

341

342

7 Group 9 Transition Metal–Dinitrogen Complexes

7.1.1.2

Cobaltate Complexes: [Co(N2 )(PR3 )3 ]−

The formally Co(–I) N2 adducts, [Co(N2 )(PR3 )3 ]− , have been isolated for both R = Me and Ph, as well as, with group 1 and 2 countercations. In general, their molecular structures reveal an end-on N2 ligand that bridges the Co center to the group 1/group 2 monatomic ion, as in the structural formula Co(PR3 )3 (μ-N2 )M(solv)n (where M is a monatomic cation). The N2 ligand in these formally cobaltate complexes is considerably more activated than in the HCo(N2 )(PR3 )3 counterparts. This trend is consistent with increased backbonding into the N2 π*-orbitals from a more reduced Co center, i.e. Co(–I). Alternatively, these complexes can be viewed from the other extreme, where the N2 ligand has been formally reduced by two electrons to (N=N)2− , which, for charge balance, would require consideration of the Co(I) oxidation state. In this latter interpretation, the ligation of an electropositive cation at the β-N-atom of N2 should facilitate charge transfer from Co to N2 by further stabilizing the anionic charge of (N2 )2− . Without detailed spectroscopic and/or electronic structural data, the correct assignment remains unclear, as many of these complexes have N—N bond lengths and N–N vibrational frequencies that are exactly intermediate between those expected for N2 (cf. 1.10 Å; 2359 cm−1 ) and (N2 )2− (cf. diazene: 1.25 Å, 1583, 1529 cm−1 ) [6, 14]. Another possible explanation is that the two distinct electronic forms represent valence tautomers, which are in equilibrium and rapidly interconvert with one other. Even though the debate on their electronic structures continue, the formally Co(–I) N2 adducts have shown great promise as synthons in N2 functionalization reactions. The synthesis of a Co(–I) N2 complex was first reported by Huttner and coworkers in 1976 [15]. By mixing CoCl2 , PMe3 , and excess Mg turnings for three days, an orange-colored diamagnetic complex was isolated. The complex was identified by X-ray crystallography as the dicobalt complex, {Co(PMe3 )3 (μ-N2 )}2 Mg(THF)4 (THF, tetrahydrofuran), where the N2 ligand bridges each Co center to the central Mg2+ ion (Figure 7.4). As an aside, the reported synthesis of {Co(PMe3 )3 (μ-N2 )}2 Mg(THF)4 predated that of CoH(N2 )(PMe3 )3 by nearly 15 years. The Co—N—N and Mg—N—N bond angles are 172∘ and 158∘ , respectively. The N—N frequency and bond length of 1830 cm−1 and 1.18 Å, respectively, indicate significantly stronger activation compared to the Co(I) hydride complexes, CoH(N2 )(PR3 )3 , and {CoH(PMe3 )3 }2 (μ-N2 ) [10, 15]. Of interest, small modifications to the synthesis of {Co(PMe3 )3 (μ-N2 )}2 Mg(THF)4 , e.g. using less Mg and/or shorter reaction times, leads to the Co(0) complex, Co(PMe3 )4 , which does not further react with N2 [15]. Co(PMe3 )4 is also observed as a primary decomposition product of {Co(PMe3 )3 (μ-N2 )}2 Mg(THF)4 . Specifically, solutions of {Co(PMe3 )3 (μ-N2 )}2 Mg(THF)4 decompose to a transient electron paramagnetic resonance (EPR)-active compound, which was proposed as “Co(N2 )(PMe3 )3 ” based on an IR band at 2064 cm−1 . Subsequent loss of the N2 and phosphine rearrangement would then generate Co(PMe3 )4 and Co metal. Further reaction of {Co(PMe3 )3 (μ-N2 )}2 Mg(THF)4 with potassium metal leads to the formation of Co(PMe3 )3 (N2 )K, which crystallizes as the hexameric cluster, [Co(PMe3 )3 (N2 )K]6 [16]. In the structure, the six KN2 units form a double cubane with six outer “Co(PMe3 )3 ” groups (Figure 7.5). Each N2 ligand bridges a

7.1 Cobalt–Dinitrogen Complexes

C C

C

C

C C C

C P

C

C

C

C C

C

C

O

P

O

P Co

172

N

20

C

8

C

N

Mg 204

158° P

C

C C

C

O C

N N 118

Co

213

C

C P C

P

O C

C C

C C C

C

C

C

Z504.1

Figure 7.4 Molecular structure of {Co(PMe3 )3 (μ-N2 )}2 Mg(THF)4 with bond lengths (pm) and angles (degrees). Source: Hammer et al. 1976 [15]. Reproduced with permission from John Wiley & Sons.

C

K3

C

P

N22 N21

C

Co2

K2

Co3 N12

N31 N32 K1

N11

Co1

Figure 7.5 Molecular structure of the hexameric cluster [Co(PMe3 )3 (N2 )K]6 . Source: Klein et al. 1978 [16]. Reproduced with permission from https://creativecommons.org/licenses/by/3.0/.

343

344

7 Group 9 Transition Metal–Dinitrogen Complexes

pair of Co and K+ centers in an end-on manner, while also interacting with adjacent K+ ions in a side-on manner. The N—N bond lengths range from 1.16 to 1.18 Å, which are similar to those in {Co(PMe3 )3 (μ-N2 )}2 Mg(THF)4 . The reactivity of Co(PMe3 )3 (N2 )K has been investigated with organic halides. In all cases, the N2 ligand is lost, and the initial Co product is the corresponding organometallic complex, which then undergoes further reactivity. For example, the reaction of Co(PMe3 )3 (N2 )K and MeI presumably generates “CoMe(PMe3 )3 ,” which then rearranges to the final products, CoMe(PMe3 )4 and Co metal. For n-propyl bromide, the end product is the allyl complex, Co(η3 -C3 H5 )(PMe3 )3 , with the elimination of H2 . Lastly, for ethyl bromide, the Co-ethyl intermediate undergoes β-hydride elimination to yield the Co–H ethylene adduct, CoH(C2 H4 )(PMe3 )3 . The formally Co(–I) N2 adducts supported by PPh3 ligands were reported by Yamamoto et al. in 1983, nearly two decades after the initial report of CoH(N2 )(PPh3 )3 [10]. The limited reactivity of the latter in N2 functionalization stands in contrast to that of its reduced analog, [Co(N2 )(PPh3 )3 ]− (vide infra). The reaction of CoH(N2 )(PPh3 )3 and MgEt2 generated {Co(PPh3 )3 (μ-N2 )}2 Mg(THF)4 , which is isostructural to the PMe3 derivative described above. The Li+ variant was prepared by reacting either CoH(N2 )(PPh3 )3 or {Co(PPh3 )3 (μ-N2 )}2 Mg(THF)4 with n-BuLi. The solid-state structures of Li(solvate)3 [Co(PPh3 )3 (μ-N2 )] were obtained for both the Et2 O and THF solvates. Their corresponding N—N bond distances of 1.17(2) and 1.19(4) Å are similar to those of the [Co(N2 )(PMe3 )3 ]− complexes. The Na+ salt was also isolated from the reaction of {Co(PPh3 )3 (μ-N2 )}2 Mg(THF)4 and Na metal. The N–N vibrational frequencies for the [Co(N2 )(PPh3 )3 ]− series vary by 80 cm−1 , from 1840 to 1920 cm−1 . Specifically, the 𝜈N2 energies of [Co(N2 )(PPh3 )3 ]− increase in order for the following cations (𝜈N2 , cm−1 ): Mg2+ (THF)4 (1840) < Li+ (THF)3 (1890) < Li+ (Et2 O)3 (1900) < Na+ (THF)3 (1910) < Li+ (12-crown-4)1.5 (1920). Hence, the extent of N2 activation strongly depends on the identity of the group 1/group 2 ion, as well as the solvate molecules. Also, the stronger Lewis acids such as Mg2+ result in greater activation. On an average, the 𝜈N2 values for [Co(N2 )(PPh3 )3 ]− are drastically lower by ∼200 cm−1 than those of CoH(N2 )(PPh3 )3 (cf. 2088 cm−1 ). Lastly, the N–N frequency for {Co(PPh3 )3 (μ-N2 )}2 Mg(THF)4 increased by only 10 cm−1 relative to the isostructural PMe3 analog. In a seminal study, Yamamoto et al. investigated the [Co(N2 )(PPh3 )3 ]− series for facilitating N2 conversion to NH3 and/or hydrazine upon hydrolysis with strong acids such as H2 SO4 or HCl in Et2 O [10]. Through careful quantification of the N2 -derived products, the authors concluded that a significant fraction of the original N2 ligand (20–30%) is reduced to ammonia and hydrazine. The study was the first demonstration of converting Co–N2 complexes into NH3 and N2 H4 . Moreover, by stark contrast, CoH(N2 )(PPh3 )3 releases 1.0 equiv of N2 upon exposure to strong acids. The authors also uncovered different selectivities for ammonia vs. hydrazine that is affected by the differing coordination environments of the end-on N2 ligand, or in other words, by the differing countercations that are ligated to the β-N-atom of N2 . For the Mg2+ (THF)4 complex, hydrazine is the favored product with a N2 H4 :NH3 molar ratio of 4.3. The Li+ (THF)3 and Na+ (THF)3 complexes also showed a preference for producing hydrazine, albeit with smaller N2 H4 :NH3 ratios of ∼1.7 and 1.6, respectively. The Li+ (Et2 O)3

7.1 Cobalt–Dinitrogen Complexes

species exhibited no preference, as both N2 H4 and NH3 were generated in nearly equivalent molar amounts. Unexpectedly, for the Li+ (12-crown-4)1.5 analog, the selectivity completely inverted, and NH3 was the dominant product, being preferred over N2 H4 by ∼16 : 1. The authors speculated that the presence of the crown ether profoundly alters the protonolysis pathways. Perhaps the more significant conclusion was that the extent of N2 activation is critical to engaging N–H functionalization schemes via protonation. Presumably, the β-N-atom of N2 must compete with the Co center as the site of protonation, where the latter case would result in N2 loss. The alternative possibility, proton transfer to the β-N-atom, yields a Co—N=N—H intermediate that can be further protonated to provide NH3 and/or N2 H4 . This desired pathway would be favored for complexes where the β-N-atom is rendered more electron rich because of significant π-backbonding from the transition metal to the N2 fragment or alternatively because of the reduced nature of an (N2 )2− ligand. 7.1.2

Tripodal Polyphosphine Ligands

Although the beginning of Co–N2 coordination chemistry featured simple monodentate phosphine ligands, the vast majority of Co–N2 complexes contain multidentate ligands. In hindsight, it seemed only natural that Co–N2 complexes would move from the early CoH(N2 )(PR3 )3 and Co(PR3 )3 (μ-N2 )M(solv)n -type complexes toward those with polyphosphine ligands. The simple chelate effect of tethering multiple phosphine donors, however, did not by itself expand the landscape of Co–N2 coordination chemistry. During the past decade or so, the rapid progress in developing molecular complexes for N2 functionalization has been intimately linked to the novel ligand designs and their adoption for Co–N2 chemistry. In this section, the evolution of tris(phosphine)-type ligands is described. 7.1.2.1

Tris(phosphine) Ligands

The first Co–N2 complex using a chelating ligand was reported as early as 1985. Cecconi et al. used the ligand 1,1,1-tris(diphenylphosphinomethyl)ethane (abbreviated as tppme) in a reaction mixture with CoCl2 and Na/Hg (4 equiv) under a N2 atmosphere [17]. The resulting product was the formally zero-valent dicobalt complex, {Co(tppme)}2 (μ-N2 ). The end-on, bridging N2 ligand, based on a N—N bond distance of 1.18(2) Å, is as activated as the bridging N2 ligands in the formally cobaltate(–I) complexes, Co(PR3 )3 (μ-N2 )M(solv)n (where M is a monatomic cation), as well as the dicobalt {CoH(PMe3 )3 }2 (μ-N2 ) compound. {Co(tppme)}2 (μ-N2 ) is paramagnetic, and a magnetic moment of 4.36 μB (Bohr magneton) was measured at room temperature using the Guoy method. At a lower temperature of 92 K, the magnetic moment decreased to 3.80 μB . The temperature dependency of the magnetic moment was preliminarily attributed to a triplet-quintet spin equilibrium. In a related work, the alkyl phosphine derivative, 1,1,1-tris(dimethylphosphinomethyl)ethane (abbreviated as tmpme), was also used to isolate the dinuclear Co(I) complex, {CoH(tmpme)3 }2 (μ-N2 ), whose structure is proposed to be isostructural to {CoH(PMe3 )3 }2 (μ-N2 ) (vide supra) [12]. The lack of a N–N vibration band by IR spectroscopy and a 1 : 1 Co:N elemental ratio further

345

346

7 Group 9 Transition Metal–Dinitrogen Complexes

supported the formulation of a N2 -bridged, symmetric species. Exchanging the three PMe3 donors with the tmpme chelate appears to offer little benefit: the only difference that was reported was a modest boost in the thermal stability. 7.1.2.2

Tris(phosphino)borate Ligands

In 2003, Betley and Peters reported a series of Co(0) and Co(I) N2 complexes that were stabilized by the tris(phosphino)borate ligand, [PhB(CH2 Pi-Pr2 )3 ]− (abbreviated as PhBiPr P3 ) [18]. The PhBiPr P3 ligand presents three diisopropyl phosphine donors in a facial coordination geometry that is similar to the classic triphos ligands. An important distinction of PhBiPr P3 is the tethering of an anionic borate functionality in the ligand backbone, which confers an overall anionic charge to the ligand and significant electron richness to the phosphine donors [19]. Starting from the Co(II) iodide precursor, CoI(PhBiPr P3 ), the addition of a strong group 1/group 2 metal reductant under a N2 atmosphere resulted in reduced Co complexes that spontaneously bind N2 [18]. The choice of the reductant is critical. For example, the addition of 1 equiv of Na/Hg generated the dicobalt(I,I) complex, {Co(PhBiPr P3 )}2 (μ-N2 ), whose connectivity was established by a low-quality X-ray structure (Figure 7.6). This compound can be further reduced with an additional equivalent of Na/Hg to provide the formally mixed-valent dicobalt (I,0) complex, Na(THF)6 [{Co(PhBiPr P3 )}2 (μ-N2 )]. An X-ray structure revealed a N—N bond length of 1.147(4) Å. Although no EPR data were provided for Na(THF)6 [{Co(PhBiPr P3 )}2 (μ-N2 )], the cobalt sites are likely equivalent, Co(0.5)(μ-N2 )Co(0.5), because the unpaired spin could readily delocalize through the Co–N–N–Co π-system. The reduction of CoI(PhBiPr P3 ) with excess magnesium powder leads to the two-electron reduced product, {Co(PhBiPr P3 )(μ-N2 )}2 Mg(THF)4 , with formally zero-valent Co centers. The N–N vibrational frequency of 1863 cm−1 compares well with the values for the isostructural, formally Co(–I) N2 adducts with PR3 donors (cf. 1830 cm−1 for R = Me; 1840 cm−1 for R = Ph). The comparatively lesser extent of N2 activation in {Co(PhBiPr P3 )(μ-N2 )}2 Mg(THF)4 vs. {Co(PR3 )(μ-N2 )}2 Mg(THF)4 is consistent with the higher Co oxidation state in the former, although the oxidation-state change appears to be significantly compensated by the greater electron richness of the phosphine donors in PhBiPr P3 . Encapsulation of Mg2+ ion with 18-crown-6 (18-c-6) stabilizes the terminal end-on N2 adducts of Co(0), or [Co(PhBiPr P3 )(N2 )]− . The N–N vibration shifts to a higher energy by 33 cm−1 (𝜈N2 = 1896 cm−1 ), which confirms the role of the Mg2+ ion in Co(μ-N2 )Mg-type complexes to facilitate charge transfer from Co to N2 and thereby promote further activation of the N2 ligand. The N2 functionalization reactivity of {Co(PhBiPr P3 )(μ-N2 )}2 Mg(THF)4 was investigated with the electrophilic reagents, MeOTs and Me3 SiCl. In both cases,

P Pr2 Co P

i

Ph

B

P Pr2

i

0/–

iPr 2

iPr 2

P

N

N

Co

P Pr2 P iPr 2 i

B

Ph

Figure 7.6 Neutral and monoanionic {Co(PhBiPr P3 )}2 (μ-N2 ).

7.1 Cobalt–Dinitrogen Complexes

N—C/N—Si bond formation was affected, and the diazenido Co(II) complexes were isolated, Co(N=NR)(PhBiPr P3 ), where R = Me or SiMe3 . The corresponding N–N vibrational frequencies of 1599 and 1654 cm−1 , respectively, are fully consistent with the presence of a N=N double bond. Using Evans’ method, an S = 1/2 ground state was determined. Pseudo-tetrahedral Co(II) centers are typically high spin with S = 3/2 ground states. Hence, the PhBiPr P3 and diazenido ligands confer sufficiently strong ligand field to render the Co(II) center low spin. 7.1.2.3

Trisphosphine Systems with an Apical Main Group Donor

One of the emerging ligand motifs in N2 coordination chemistry is a tetradentate ligand with three equatorial phosphine donors that are tethered to an apical donor. Such ligands tend to coordinate the transition metal center in a trigonal pyramidal geometry, positioning the apical donor trans to the N2 binding pocket. A large range of apical “donors” have been investigated, including σ-donors and σ-acceptors. In an early report, Bianchini et al. stabilized a cationic Co(I) N2 complex using the tetradentate polyphosphine, P(CH2 CH2 PPh2 )3 , which is abbreviated as PP3 [20]. The compound, [Co(N2 )(PP3 )]BPh4 , has an end-on terminal N2 ligand with a N—N bond length of 1.08(2) Å and a corresponding vibrational frequency of 2125 cm−1 . The N–N frequency is higher than that of CoH(N2 )(PPh3 )3 , which can be attributed to the difference in the overall molecular charge (cationic vs. neutral) and/or the different trans donor strengths (phosphine vs. hydride), which are both less favorable for N2 activation. Reactivity of [Co(N2 )(PP3 )]BPh4 was explored in the solid–gas interface by exposing crystals of the complex to organic vapors at 90 ∘ C. With N2 as a labile ligand, the complex mediates the activation of weak C—H bonds, including those of acetylene, formaldehyde, and acetaldehyde. For acetylene, the final product is the cobalt vinylidene, [Co=C=CH2 (PP3 )]+ , wherein C–H oxidative addition of acetylene is followed by a 1,3-hydride shift. In the case of formaldehyde and acetaldehyde, the products are the carbonyl adduct, [Co(CO)(PP3 )]+ , and the acyl complex, [Co(COMe)(PP3 )]+ , respectively, both of which occur with formal H2 elimination (1 and 0.5 equiv, respectively) after the initial C—H bond activation(s). Because N2 is formally displaced in these products, one mechanistic possibility involves initial N2 dissociation before association of the organic substrate. However, crystals of [Co(N2 )(PP3 )]BPh4 are quite stable and do not lose N2 either under argon or under reduced pressure (0.1 Torr) at 100 ∘ C. Based on these observations, the authors propose a mechanistic sequence where the initial step is dechelation of a phosphine donor. A subsequent association of the organic substrate then results in N2 expulsion and recoordination of the phosphine donor. Peters et al. have prepared a series of isostructural EP3 cobalt complexes, where E represents the apical borane, silyl, or a carbanion donor that is positioned trans to the N2 binding pocket [21–23]. The coordination of N2 is in the terminal, end-on mode. The ligands, which adhere to the molecular formula, (o-(i-Pr2 P)C6 H4 )3 E, are referred to as tris(phosphino)borane, or TPB for E = B; SiP3 for E = Si; and CP3 for E = C. The cobalt complexes differ in their trans donor strength, overall molecular charge, and the adaptability or flexibility of the Co–E linkage. All of these factors impact the extent of N2 activation (vide infra) [23].

347

348

7 Group 9 Transition Metal–Dinitrogen Complexes

0/–

N N iPr P 2

Co B

PiPr

PiPr2

TPB

2

iPr

2P

N

N

N

N

Co Si

PiPr2 PiPr2

iPr P 2

Co C

N iPr P 2

PiPr2 PiPr2

CP3

SiP3

0/–

N

0/+

N

Co PiPr2 PiPr2 Al N N N

Al(o-(iPr2PCH2N)C6H4)3N

Figure 7.7 Co(EP3 ) and related complexes with the variation of the apical donor/acceptor (E = B, Si, C, Al). Table 7.1 The extent of N2 activation in Co(N2 )(EP3 ) and related complexes where the apical donor/acceptor E is B, Al, C, or Si. Complex

Co(n)

N—N (Å)

𝝂N2 (cm−1 )

[Co(N2 )(TPB)]−

−I

1.129(4)

1978

[Co(N2 ){Al(o-(i-Pr2 PCH2 N)C6 H4 )3 N}]−

−I

1.110(8)

1995

Co(N2 )(CP3 )

I

1.024(6)

2057 2063

Co(N2 )(SiP3 )

I

1.123(3)

Co(N2 ){Al(o-(i-Pr2 PCH2 N)C6 H4 )3 N}

0

1.107(4)

2081

Co(N2 )(TPB)

0

1.057(2)

2089

Co(N2 )(SiPPh 3 )

I

—

2095

[Co(N2 )(CP3 )]+

II

1.105(4)

2182

Data include the formal Co(n) oxidation state, N—N bond distances, and 𝜈N2 energies.

These complexes and closely related analogs are shown in Figure 7.7, and their 𝜈N2 are energies listed in Table 7.1. For example, the diphenylphosphino variant, Co(N2 )(o-(Ph2 P)C6 H4 )3 Si, or Co(N2 )(SiPPh 3 ), is included for comparison [21]. Lu and coworkers have also studied isostructural Co(N2 ) complexes featuring the alumatrane ligand, Al(o-(i-Pr2 PCH2 N)C6 H4 )3 N [24, 25]. Collectively, these complexes allow the effect of the varying E apical donor to be closely examined for the light B and C donors, as well as their heavier congeners, Al and Si, respectively. The syntheses of the Co(N2 )(EP3 )-type complexes often involve the addition of a strong reductant to the halide precursor, CoX(EP3 ), where X = Cl or Br [21, 22]. In specific cases, a high-yielding synthesis of Co(N2 )(EP3 ) is possible from a one-pot reaction of CoCl2 (THF)1.5 /CoBr2 , ligand, and 2–3 equiv of the reductant [23, 24]. Typically, strong reductants are employed, such as Na/Hg, Na(C10 H8 ), KC8 , or MeMgCl. The Co oxidation state in the neutral Co(N2 )(EP3 )-type complexes is low spin, either S = 1/2 Co(0) or S = 0 Co(I), depending on the charge of the E donor being neutral (E = B, Al) or anionic (E = C, Si), respectively. For the group 13 donors (E = B, Al), the Co(N2 )(EP3 ) complexes exhibit a Co(0/−I) redox couple at −2.0 and −2.1 V vs. [FeCp2 ]+/0 ,

7.1 Cobalt–Dinitrogen Complexes

respectively [23, 26]. Chemical reduction with another equivalent of a strong reductant in the presence of an encapsulating crown ether reagent provides the monoanionic [Co(N2 )(EP3 )]− analogs, where N2 remains coordinated in a terminal, end-on manner [23, 25]. The Co center is formally d10 Co(–I), which is consistent with the diamagnetism of these monoanionic compounds. In the case of Co(N2 )(CP3 ), which exhibits a Co(I/II) redox couple at −1.1 V vs. [FeCp2 ]+/0 , chemical oxidation with [FeCp2 ]BArF 4 (where ArF = 3,5-(CF3 )2 C6 H3 ) leads to the cationic N2 adduct, [Co(N2 )(CP3 )]+ [23]. In comparison, chemical oxidation of Co(N2 )(TPB) can be effected with [H(Et2 O)2 ]BArF 4 but results in N2 loss in the cationic state, [Co(TPB)]+ . Of note, the one-electron oxidation of Co(N2 )(TPB) at −0.2 V is reversible on the timescale of the cyclic voltammetry experiment. The extent of N2 activation can be evaluated by the elongation of the N—N bond and by a corresponding decrease in the N–N stretching frequency. Because the latter data are more reliable, the Co(N2 )(EP3 )-type complexes are ranked in the order of increasing 𝜈N2 energies in Table 7.1 to facilitate the comparisons. Generally, as one might expect, greater N2 activation is associated with more electron-rich Co centers, e.g. overall more negative charges, lower Co formal oxidation states, stronger trans donors, and alkyl vs. aryl phosphine substituents (cf. Co(N2 )(SiP3 ) vs. Co(N2 )(SiPPh 3 ), where Δ(𝜈N2 ) = 32 cm−1 ). Within the neutral Co(N2 )(EP3 )-type subset featuring diisopropylphosphine donors, the extent of N2 activation does not adhere to the expectations based on formal Co oxidation states, but rather more to the different donor strengths of E, where C > Si > Al > B. Although the choice of E does tune N2 activation, the full range of 𝜈N2 is within a moderate window of 32 cm−1 . More significant increases in N2 activation are affected by the addition of an electron. From [Co(N2 )(CP3 )]+ to Co(N2 )(CP3 ), 𝜈N2 decreases by 125 cm−1 . A slightly smaller change of 111 cm−1 is observed in 𝜈N2 between Co(N2 )(TBP) and [Co(N2 )(TBP)]− [23]. By contrast, the Al analogs, [Co(N2 ){Al(o-(i-Pr2 PCH2 N)C6 H4 )3 N}]0/− , result in an even smaller decrease in 𝜈N2 of 86 cm−1 [25]. The smaller perturbation in N2 activation was attributed to a significant increase in the Co → Al σ-back donation. In support, the Co—Al bond distance contracts by 0.12 Å upon one-electron reduction, whereas the corresponding Co—B bond length shortens by only 0.02 Å. In a comprehensive 2015 study, the isostructural Co(N2 )(EP3 ) complexes were evaluated for the catalytic reduction of N2 to NH3 . In addition to the identity of the apical donor, the study also examined different charge states as well as other Co precursors. The standard catalytic conditions involved the sequential addition of [H(Et2 O)2 ]BArF 4 (47 equiv) and KC8 (60 equiv) to a solution of the Co complex in Et2 O at −78 ∘ C under 1 atm N2 . Under these conditions, both Co(N2 )(SiP3 ) and Co(N2 )(CP3 ) were ineffective catalysts, generating ≤0.1 equiv of NH3 per Co complex. The Co(TPB) complexes were somewhat promising in that NH3 formation was more significant from 0.7 to 2.4 equiv. The amount of NH3 produced strongly depended on the exact precursor complex, despite sharing an identical “Co(TPB)” core. Only [Co(N2 )(TBP)]− reliably generated more than 2 equiv of NH3 and thus was the only Co(N2 )(EP3 ) complex in this study that could be regarded as a true N2 -to-NH3 catalyst. The neutral species, Co(N2 )(TBP), only generated 0.8 equiv, akin to CoBr(TBP) (0.7 equiv), and worse than [Co(TBP)]+

349

350

7 Group 9 Transition Metal–Dinitrogen Complexes

(1.6 equiv). The authors concluded that NH3 activity did not solely depend on the different extents of N2 activation across the Co(N2 )(EP3 ) systems. Rather, the factors that better correlate with higher NH3 production are (i) the overall charge state, which transfers more negative charge to the Nβ -atom of the bound N2 and (ii) the geometric flexibility of the Co–E linkage, which is directly positioned trans to the N2 binding pocket. The rigidity of the Co–C and Co–Si cores appear to hinder activity, whereas the more flexible Co–B core promotes turnover. For the Al analog, Co(N2 ){Al(o-(i-Pr2 PCH2 N)C6 H4 )3 N}, catalytic N2 functionalization was also investigated, but in the context of reductive N2 silylation to N(SiMe3 )3 [27]. Under 1 atm N2 at ambient temperature, the Co–Al complex was stirred with excess KC8 (2000 equiv) and Me3 SiCl (2000 equiv) in THF. The Co(N2 ){Al(o-(i-Pr2 PCH2 N)C6 H4 )3 N} catalyst generated 30 equiv of N(SiMe3 )3 . The identity of the E donor was also discovered to be highly critical for imparting catalytic activity. For example, replacing Al with a second Co ion results in superior catalytic activity (vide infra). 7.1.2.4

Trisphosphine Systems with an Apical Transition Metalloligand Donor

In continuation of the previous section on Co(N2 )(EP3 ) complexes, the selection of the apical donor has extended beyond the main group elements (E) and toward the transition metal block (M). By moving to transition metal-based ligands, the apical donor can also be redox active. One promise of “transition metalloligands” is their potential to harbor additional electrons, which could benefit N2 functionalization processes by providing additional reducing equivalents. The presence of a direct metal–metal interaction may further facilitate electron transfer from the metalloligand to the active metal where N2 is terminally bound and undergoes stepwise reduction to NH3 [26]. Alternatively, the N2 substrate may engage a metal–metal bond by binding both metals simultaneously in a side-on manner, although this type of bonding has yet to be shown. Akin to the hemilability of the apical main group donor to a transition metal center described above, the dynamic range of metal–metal bonding interactions can also tune the extent of N2 activation and facilitate catalysis [26]. Lastly, the sheer number of possible metal–metal pairings may translate into an extensive, electronically diverse library for catalyst discovery [28]. Building such a library will require versatile organic scaffolds that allow for the modular synthesis of diverse metal–metal pairings. In 2009, Thomas and coworkers reported the first N2 –Co–M complex featuring a strong Co–M metal–metal interaction [29]. To date, the N2 —Co—M bonding motif has been shown to be quite general as different transition metals across the d-block have been employed. The compounds are primarily C 3 symmetric and feature a triphosphine donor set that is tethered to three amido donors, which not only ligate the second transition metal (M) but also place it in bonding proximity to the Co center. Lu’s heptadentate ligand, [(o-(i-Pr2 PCH2 N)C6 H4 )3 N]3− , described previously for the alumatrane ligand [24], was also used to isolate N2 –Co–M complexes, where the M center varied across the first-row period from early to late transition metals [27, 30]. Using phosphinoamide ligands [R2 PNR′ ]− , Thomas and coworkers isolated more than two dozen (N2 )Co–M tris(phosphinoamide) complexes, where M is a group

7.1 Cobalt–Dinitrogen Complexes

4 or 5 transition metal. Among all the Co–M complexes, a couple of noteworthy cases of catalysis involving N—N bond cleavage have been reported. First, the Co–Co complexes, [(N2 )Co{Co(o-(i-Pr2 PCH2 N)C6 H4 )3 N}]−/0 , catalyze the reductive silylation of N2 to N(SiMe3 )3 in relatively high yields [27]. Second, the (N2 )Co(μ-i-Pr2 PNXyl)3 Ti(L) complexes are catalysts for the disproportionation of hydrazine into N2 and NH3 [31]. The isostructural Co{M(o-(i-Pr2 PCH2 N)C6 H4 )3 N} complexes, where M = Ti, V, Cr, and Co, are synthesized in a two-step metalation process. First, the deprotonated ligand is metalated with the supporting metal, MCl3 , and then the isolated metalloligand, M(o-(i-Pr2 PCH2 N)C6 H4 )3 N, is subsequently metalated with CoBr2 and 2 equiv of KC8 [27, 28, 30, 32, 33]. Of interest, all these complexes exhibit a one-electron reduction process that is irreversible under N2 but becomes reversible under Ar. Consistent with this electrochemical behavior, the complexes were found to bind N2 upon one-electron reduction, as another equivalent of KC8 generates the corresponding [(N2 )Co{M(o-(i-Pr2 PCH2 N)C6 H4 )3 N}]− complexes for M = V, Cr, and Co. Notably, in the case of Ti, the Co(N2 ) adduct could not be isolated, presumably because of the extremely harsh potential of −3.2 V vs. [FeCp2 ]+/0 required to reduce the Co–Ti precursor complex. The data for the [(N2 )Co{M(o-(i-Pr2 PCH2 N)C6 H4 )3 N}]− series is provided in Table 7.2. Greater N2 activation was found to correlate with more negative reduction potentials, which is consistent with more electron-rich Co centers both having more electron density to donate into N2 and having more negative reduction potentials. The impact of the identity of M on the N–N stretching frequencies is subtle (∼25 cm−1 range), although an interesting periodic trend was elucidated, where the extent of N2 activation decreases as M is varied across (left to right) the first-row period. Of note, Table 7.2 The extent of N2 activation in bimetallic Co–N2 complexes featuring a metalloligand, ML, where L is the heptadentate ligand, [(o-(i-Pr2 PCH2 N)C6 H4 )3 N]3− . Complex

Co(n)

E∘ (V)a)

[(N2 )CoV(L)]− b)

−I

[(N2 )CoCr(L)]− b)

N—N (Å)

𝝂N2 (cm−1 )

−2.48

1.130(4) 1.135(4)

1971

−I

−2.32

1.120(7) 1.135(6)

1990

[(N2 )CoCo(L)]−

0

−2.11

1.114(4)

1994

[(N2 )CoAl(L)]−

−I

−2.12

1.110(8)

1995

(N2 )CoAl(L)

0

−2.12

1.107(4)

2081

The N2 –Co–Al analogs are also included for comparison. Data include the formal Co(n) oxidation state, reduction potentials, N—N bond distances, and 𝜈N2 energies. a) E∘ for [CoML]0/− (M = Co, Cr, V) and for [(N2 )CoAlL]0/− vs. [FeCp2 ]+/0 under Ar in [n-Bu4 N]PF6 /THF. b) Two unique molecules per asymmetric unit.

351

352

7 Group 9 Transition Metal–Dinitrogen Complexes

the 𝜈N2 energies compare well to those of the Co(–I)/group 13 compounds, [Co(N2 )(TBP)]− and [(N2 )Co{Al(o-(i-Pr2 PCH2 N)C6 H4 )3 N}]− . Theoretical studies of the [(N2 )Co{M(o-(i-Pr2 PCH2 N)C6 H4 )3 N}]− series support the oxidation state assignment of Co(–I) for the early-to-mid transition metals, Ti, V, and Cr [30]. However, in the case of the Co–Co complex, the Co oxidation states are II and 0, corresponding to the Co metalloligand and the CoP3 site, respectively [30]. An unexpected finding was that the 𝜈N2 energies and reduction potentials are nearly identical for the Co–Co and Co–Al analogs. The similarities suggest that despite differences in formal oxidation states, Co(0) vs. Co(–I), the electron-donating effects of an electron-rich Co(II) metalloligand on a Co(0)N2 moiety are electronically equivalent to the electron-withdrawing effects of a Lewis acidic Al(III) ligand on a subvalent Co(–I)N2 moiety. Moving toward N2 catalysis, both the Co–Co complexes, [(N2 )Co{Co(o-(iPr2 PCH2 N)C6 H4 )3 N}]−/0 , catalyze the reductive silylation of N2 to N(SiMe3 )3 [27]. At ambient temperature and under 1 atm N2 , the neutral Co–Co complex was stirred with excess KC8 (2000 equiv) and Me3 SiCl (2000 equiv) in THF. The Co–Co catalyst generated 195 equiv of N(SiMe3 )3 in a single catalytic run, and 320 equiv of N(SiMe3 )3 in a double catalytic run. The monoanionic Co–Co analog showed a similar activity, with 178 equiv of N(SiMe3 )3 in a single run. Perhaps the more striking lesson is that under identical catalytic conditions, the Co–Al catalyst performed relatively dismally (cf. 30 equiv). Hence, the extent of N2 activation appears to have no bearing on the overall activity in N2 functionalization schemes. To date, the underlying cause for the differences in activity between the Co–Co and Co–Al systems is unknown. The superior performance of the Co–Co catalyst could be attributed to lower activation barriers along the reaction pathway, to being more robust to decomposition, and/or to avoiding off-pathway species that ultimately lower the concentration of active species. To disentangle these and other possibilities, detailed kinetic/mechanistic studies will be necessary. One of the best studied N2 –Co–M compounds is (N2 )Co(μ-i-Pr2 PNMes)3 Zr(THF), where Mes is 2,4,6-Me3 (C6 H2 ), as shown in Figure 7.8 [34]. Table 7.3 lists the relevant characterization data for this complex and related tris(phosphinoamide) analogs that share a total valence d-electron count of 10. The synthesis of (N2 )Co(μ-i-Pr2 PNMes)3 Zr(THF) begins with a two-electron reduction of the dihalide precursor, ICo(μ-i-Pr2 PNMes)3 ZrCl, using excess Na/Hg in THF to afford the cobaltate complex, Na(THF)5 [(N2 )Co(μ-iPr2 PNMes)3 Zr(X)], where X is a mixture of I− and Cl− [29]. Dissolution of the latter complex into benzene results in facile extrusion of NaX and generation of (N2 )Co(μ-i-Pr2 PNMes)3 Zr(THF) [34]. An X-ray absorption spectroscopic study revealed that the complex is best described as a Co(–I)/Zr(IV) zwitterion [35]. Mes

O

Figure 7.8 Heterobimetallic Co–Zr complex, (N2 )Co(μ-i-Pr2 PNMes)3 Zr(THF).

i

Pr2 P

N Zr N

N Mes Mes

Co P iPr P 2 iPr 2

N

N

7.1 Cobalt–Dinitrogen Complexes

Table 7.3 The extent of N2 activation in bimetallic Co–N2 tris(phosphinoamide) complexes. Entry

General molecular formula

[R2 PNR′ ]−

M(L/X)

{CoM}n

𝝂N2 (cm−1 )

1

(N2 )Co(μ-R2 PNR′ )3 M(L)

i-Pr2 PNMes

Zr(THF)

10

2026

2

i-Pr2 PNXyl

Zr(THF)

10

2045

3

i-Pr2 PNi-Pr

Hfa)

10

2046

4

i-Pr2 PNi-Pr

Zr(THF)

10

2056

5

i-Pr2 PNXyl

Ti(THF)

10

2084 1940

6b)

[(N2 )Co(μ-R2 PNR′ )3 M(X)]−

Ph2 PNi-Pr

Nb(Nt-Bu)

10

7b)

Ph2 PNi-Pr

Ta(Nt-Bu)

10

1940

8

i-Pr2 PNMes

Hf(Cl)

10

1992

9

i-Pr2 PNMes

Zr(I/Cl)c)

10

2023

[R2 PNR′ ]− is the phosphinoamide, M represents a second transition metal, and L/X refers to any additional L- or X-type ligands. The complexes are grouped into two general subclasses of complexes. Data include the total valence electron count {CoM}n and 𝜈N2 energies. a) No additional ligand is coordinated. b) The N2 ligand bridges Co and the Na countercation. c) X-type ligand is a mixture of I− and Cl− .

The relatively contracted Co—Zr bond of 2.36 Å was attributed to the presence of three (1σ + 2π) donor–acceptor, or Co → Zr dative, interactions, as elucidated by a natural bond order analysis. The presence of these strong Co → Zr dative interactions also explains the relatively low degree of N2 activation (Table 7.3, entry 1, 𝜈N2 = 2026 cm−1 ) for a formally Co(–I) center (cf. 1971–1995 for Co(–I) complexes with an apical E or M donor). For (N2 )Co(μ-i-Pr2 PNR′ )3 Zr(THF) compounds, changing the R′ substituent groups on the amido donors to Xyl (where Xyl = 3,5-Me2 (C6 H3 )) or i-Pr shifts 𝜈N2 higher to 2045 and 2056 cm−1 , respectively [34, 36]. By changing the group 4 center affects the same magnitude of change in the N—N bond stretching frequency. For example, moving up the group 4 metal triad from Zr to Ti results in a 39 cm−1 increase (cf. Table 7.3, entries 2 and 5), whereas moving down from Zr to Hf results in a 10 cm−1 decrease (cf. Table 7.3, entries 4 and 3) [31, 37]. These trends are consistent with decreasing Co → group 4 interactions down the triad, and with the group 4 M(IV) ion outcompeting N2 for electron density from the Co(–I) center. The related [(N2 )Co(μ-R2 PNR′ )3 M(X)]− complexes also share a total d-electron count of 10. In some cases, the presence of an X ligand on the M center can attenuate the Co → M interactions and thereby perturb the extent of N2 activation at the Co site. For example, the (N2 )Co–Hf(Cl) complex, Na(THF)5 [(N2 )Co(μ-i-Pr2 PNMes)3 HfCl], has a significantly more activated N2 ligand with 𝜈N2 = 1992 cm−1 (Table 7.3, entry 8) [37]. Moreover, the 𝜈N2 of [(N2 )Co(μ-i-Pr2 PNMes)3 HfCl]− compares well with the values of other Co(–I) complexes featuring an apical E or M donor. Even more striking are [(N2 )Co(μ-R2 PNR′ )3 M(X)]− complexes containing group 5 imido groups, Ta/Nb=Nt-Bu (Table 7.3, entries 6 and 7) for which 𝜈N2 values of 1940 cm−1

353

354

7 Group 9 Transition Metal–Dinitrogen Complexes

suggest a strong N2 activation [38]. Of note, the Co-group 5 complexes are distinct in that N2 is not an end-on terminal ligand as is observed for entries 8 and 9, but rather the N2 bridges between Co and the solvated Na(THF)5 ion. The Na+ ion, akin to the Mg2+ ions discussed previously, promotes increased charge transfer from Co to N2 by stabilizing the negative charge that accumulates on the N2 ligand. Lastly, of note, the Na(THF)5 [(N2 )Co(μ-i-Pr2 PNMes)3 Zr(X)] complex (entry 9) is a good counterexample, where the X ligand does not perturb the extent of N2 activation [29]. Presumably, the Zr–X interaction in Na(THF)5 [(N2 )Co(μ-i-Pr2 PNMes)3 Zr(X)] is so weak that there is only a 3 cm−1 difference in 𝜈N2 between it and (N2 )Co(μ-i-Pr2 PNMes)3 Zr(THF) (cf. entries 9 and 1). The reactivity of (N2 )Co(μ-i-Pr2 PNMes)3 Zr(THF) has been extensively studied. The N2 ligand is either released or remains bound to Co as a spectator ligand. In the latter case, N2 serves as a reporter ligand for the electronic changes at Co. Despite possessing a bonafide Co(–I)/Zr(IV) core [35], all the reactivity of (N2 )Co(μ-i-Pr2 PNMes)3 Zr(THF) is consistent with one-electron reactivity at the Zr site, as one might expect for a Zr(III) center. The bimetallic complex is a catalyst for ketone hydrosilylation and Kumada coupling [39, 40]. It also reacts with benzophenone to provide an initial Co(0)/Zr(IV) ketyl radical adduct, (N2 )Co(μ-i-Pr2 PNMes)3 Zr(O–C⋅Ph2 ), which then undergoes reversible intermolecular C—C bond coupling, and ultimately yields a Co carbene product, Ph2 C=Co(μ-O)(μ-i-Pr2 PNMes)2 Zr(η2 -i-Pr2 PNMes), where the C=O bond is fully cleaved [41, 42]. Stoichiometric O—H/N—H bond activations of alcohols and hydrazines result in one-electron oxidations to the Co(0)−N2 products, (N2 )Co(μ-i-Pr2 PNMes)3 Zr(OR), and (N2 )Co(μ-i-Pr2 PNMes)3 Zr(η2 -NH2 NR′ ), respectively, where R = H, Me, Et, and R′ = Me, Ph, H [43, 44]. Consistent with a more oxidized Co(0) center, the 𝜈N2 energies of the products shift to higher energies in the range of 2038–2052 cm−1 . Akin to the early CoH(N2 )(PR3 )3 complexes, the Co(I) hydride species, H(N2 )Co(μ-i-Pr2 PNMes)3 Zr(OSiPh3 ), has also been reported [45]. The N2 ligand is weakly activated, and the N–N stretching frequency of 2094 cm−1 is slightly higher than that of CoH(N2 )(PPh3 )3 (cf. 2088 cm−1 ). From the perspective of N2 functionalization, the Co–Ti complexes are more promising than the Co–Zr congeners, especially for mediating N—N and N=N bond activations. The Co–Ti complex, (N2 )Co(μ-i-Pr2 PNXyl)3 Ti(THF), catalyzes the disproportionation of hydrazine: 3 N2 H4 → 4 NH3 + N2 [31]. A standard catalytic run involved addition of hydrazine (20 equiv) to an Et2 O solution of the Co–Ti catalyst at ambient temperature for 30 minutes. The formation of NH3 maximizes in the range of 16–18 equiv because of catalyst decomposition in the presence of excess NH3 . The related ammine adduct, (N2 )Co(μ-i-Pr2 PNXyl)3 Ti(NH3 ), also catalyzes hydrazine disproportionation with similar yields. However, overexposure to NH3 results in the generation of the protonated ligand, i-Pr2 PNHXyl. The same complex also stoichiometrically cleaves the N=N bond of azobenzene to provide the Co(I)−N2 –Co(I) sandwich product, {Co(μ-NPh)(μ-i-Pr2 PNXyl)2 Ti(𝜅 2 -XylNPi-Pr2 NPh)}2 (μ-N2 ). Overall, the four-electron reduction of azobenzene is supplied by the two initial Co(–I) centers, and the resulting “NPh” fragments separately insert between Co and Ti,

7.1 Cobalt–Dinitrogen Complexes

and between Ti and P [31]. Of note, substituting azobenzene with PhNH–NHPh leads to the same product. In closing, the reactivity differences between the Co–Ti and Co–Zr complexes, (N2 )Co(μ-i-Pr2 PNXyl)3 Ti(THF) and (N2 )Co(μ-i-Pr2 PNMes)3 Zr(THF), respectively, have been rationalized on the basis of greater Co → M interactions for M = Ti compared to Zr. The Co → M interactions directly impact N2 activation with an inverse correlation (vide supra), and the N–N stretching frequencies between the Co–Ti and Co–Zr complexes differ by as much as 58 cm−1 (cf. Table 7.3, entries 1 and 5). In the case of Co–Ti, the weaker Co → N2 π-backbonding interaction means a more labile N2 that ultimately allows for reactivity at the Co site, whereas in the case of Co–Zr, the stronger Co → N2 π-backbonding interaction hinders Co-based reactivity, and as a result, nearly all the reactivity occurs at the Zr site. 7.1.3

Ligands with Exclusively Nitrogen Donors

With the gaining momentum of phosphine-ligated Co–N2 complexes, the ligands also underwent diversification toward exclusively nitrogen donors. In 1990, Theopold and coworkers reported the first Co–N2 compound featuring nitrogen-only chelates: an end-on Co(I) N2 complex with a variant of the hydridotris(pyrazolyl)borate ligand (abbreviated as Tp) [46]. The tetrahedral geometry of the Tp Co(I) center bears some resemblance to the early tris(phosphine) Co(0) N2 complexes and are closely related to the Co(I) N2 tris(phosphino)borate analogs. Two other exclusively N-donor chelates, the β-diketiminate and the bis(α-imino)pyridine ligands, have led to approximately a dozen more Co–N2 compounds. Unlike Tp, these ligands stabilize planar Co–N2 complexes, which were not known before 2000. The first planar Co–N2 complex was supported by the bis(α-imino)pyridine ligand [47], which would set the stage for an even larger diversity of pincer-type ligands. 7.1.3.1

Tris(pyrazoyl)borate (Tp) Ligands

The hydridotris(pyrazolyl)borate, or Tp, ligands developed by Trofimenko present a facial triad of pyrazolyl donors [48]. In common with the tris(phosphino)borate ligand, the anionic borate group in the ligand backbone confers an overall anionic charge to the Tp ligand. Electronically, the Tp ligands are less electron donating than their phosphine counterparts. On the other hand, the pyrazolyl groups are more easily functionalized, and substituents at the 3-position of the ring can confer steric protection to the metal center. This feature allows the ligand to act as a “tetrahedral enforcer” and/or hinder dimerization processes [49]. Theopold and coworkers developed the hydrotris(3-tert-butyl-5-methylpyrazolyl)borate variant (abbreviated as TptBu,Me ) [46]. The additional methyl substituents shield the B—N bonds from undesired borotropic rearrangements. As a result, TptBu,Me has become a workhorse ligand for studying small-molecule activation, in particular for O2 . The end-on Co(I) N2 complex, Co(N2 )(TptBu,Me ), was prepared by reducing the Co(II) halide, CoI(TptBu,Me ), with Mg metal [46]. The N2 ligand is weakly activated based on a N–N stretching frequency of 2046 cm−1 . The magnetic moment of 3.87 μB is consistent with an S = 1 ground state.

355

356

7 Group 9 Transition Metal–Dinitrogen Complexes

Upon dissolution in CH2 Cl2 , a full equivalent of N2 is released and CoCl(TptBu,Me ) is formed. The N2 ligand can be displaced by O2 to yield the side-on superoxo species, Co(η2 -O2 )(TptBu,Me ). Subjecting Co(N2 )(TptBu,Me ) to repeated cycles of freeze–pump–thaw, a half equivalent of N2 can be cleanly removed to form the dicobalt–N2 -sandwich complexes, {Co(TptBu,Me )}2 (μ-N2 ) [49]. Other Tp ligand variants can directly generate the dicobalt(I,I) N2 -sandwich ′ complexes, {Co(TpR,R )}2 (μ-N2 ), where R,R′ = neopentyl (Np),H; i-Pr,Me; and t-Bu,H [49–51]. Akin to the synthesis of Co(N2 )(TptBu,Me ), the preparation of {Co(TpNp )}2 (μ-N2 ) proceeds via the reduction of CoI(TpNp ) with Mg [50]. In the molecular structure, the end-on N2 ligand is bent away from the Co–B axis by ∼36∘ and the N—N bond length of 1.14(3) Å. The complex displayed a weak IR band at 2056 cm−1 , which is assigned to the N–N vibration. Presumably, this band is IR active because of the lowered symmetry [50]. In the related complexes, {Co(TpiPr,Me )}2 (μ-N2 ) and {Co(TptBu )}2 (μ-N2 ), the N2 ligand also deviates from the Co–B vector [49]. Hence, this effect was attributed to a Jahn–Teller distortion of a d8 ion in a pseudotetrahedral ligand field [50]. The N—N bond distance of 1.154(9) in {Co(TpiPr,Me )}2 (μ-N2 ) is close to that of {Co(TpNp )}2 (μ-N2 ) [49]. 7.1.3.2

𝛃-diketiminate Ligands

One of the hallmarks of bulky β-diketiminate ligands is their ability to stabilize a trigonal planar coordination geometry, where the metal center is both low coordinate and less valent. For the late first-row transition metals, such as Fe and Co, the β-diketiminate ligand field destabilizes a singly occupied d-orbital that has the right symmetry match to π-backbond into a N2 π*-orbital [6]. This unusual ligand field may account for the ability of Co to achieve the full two-electron reduction of N2 to (N2 )2− . Holland and coworkers have expanded the coordination chemistry of Co–N2 complexes bearing the bulky β-diketiminate ligand, [CH{C(t-Bu)N(2,6i-Pr2 C6 H3 )}2 ]− , abbreviated as LtBu,dipp . Reduction of CoCl(LtBu,dipp ) with 1 equiv. of KC8 provided the dinuclear N2 -sandwich complex, {Co(LtBu,dipp )}2 (μ-N2 ) [52]. The molecular structure reveals a three-coordinate Co center. The metal is severely distorted into a T-shaped geometry because of a rather large N—Co—N bond angle of 163∘ (Figure 7.9), as opposed to the trigonal, or Y-shaped, alternative. The N—N bond length of 1.139(2) Å reveals a weakly activated N2 ligand. In comparison, the isostructural Fe analog has a significantly more activated N2 ligand based on its N—N bond length of 1.189(4) Å. The magnetic susceptibility measurement of {Co(LtBu,dipp )}2 (μ-N2 ) is consistent with an S = 2 ground state, which arises from ferromagnetic coupling of two S = 1 Co(I) centers. The reduction of CoCl(LtBu,dipp ) with 2 equiv of KC8 provided the dianionic N2 -sandwich complex, K2 {Co(LtBu,dipp )}2 (μ-N2 ), which is formally Co(0) [6]. The molecular structure is similar to {Co(LtBu,dipp )}2 (μ-N2 ) except for two key differences (Figure 7.9). First, the Co centers adopt the typical Y-shaped geometry. Second, the potassium ions are bonded to the N2 ligand in a side-on manner and to a ligand arene ring from each monocobalt fragment. The N—N bond length of 1.220(2) Å and corresponding vibrational energy of 1599 cm−1 (Raman spectroscopy) is fully consistent with a N=N double bond (cf. diazene: 1.25 Å, 1583, 1529 cm−1 ). Hence, the formal Co(0) oxidation state in K2 {Co(LtBu,dipp )}2 (μ-N2 ) is

7.1 Cobalt–Dinitrogen Complexes

N11

N11 N21

N21 Co1

Co1

N1

N1 K1

N1

K2 N1

Co2 Co2

N11

N11 N21

N21

(a)

(b)

Figure 7.9 Molecular structures of (a) {Co(LtBu,dipp )}2 (μ-N2 ) and (b) K2 {Co(LtBu,dipp )}2 (μ-N2 ). Source: Ding et al. 2009 [52]. Reproduced with permission from ACS.

misleading because the presence of (N2 )2− requires the metals’ oxidation states to be Co(I). Comparison to the isostructural Fe analog reveals a similar extent of N2 reduction. The magnetic susceptibility measurement of K2 {Co(LtBu,dipp )}2 (μ-N2 ) reveals an S = 1 ground state, which may arise two S = 1 Co(I) centers that antiferromagnetically couple to an S = 1 (N2 )2− ligand, i.e. Co(↑↑)−N2 2− (↓↓)−Co(↑↑). Scheer et al. recently reported another dianionic dicobalt–N2 -sandwich complex with the β-diketiminate ligand, [CH{C(Me)N(2,6-i-Pr2 C6 H3 )}2 ]− , or LMe,dipp [53]. The product, K2 {Co(LMe,dipp )}2 (μ-N2 ), was crystallized as two different solvomorphs, hexane and Et2 O. Both solvomorphs possess long N—N bond lengths of 1.215(3) and 1.220(4) Å, which are consistent with a N=N double bond. In support, the Raman study of the hexane solvomorph elucidated a N–N vibration band at 1568 cm−1 . Among all the Co–N2 complexes, the β-diketiminate complexes, K2 {Co(LR,dipp )}2 (μ-N2 ), demonstrate the greatest extent of N2 activation by far. Their N–N vibrational energies are even slightly lower than bona fide Co diazenido, Co(N=NR), complexes (Section 7.1.2.2). What factors explain the strong activation in the β-diketiminate systems? In addition to their unique ligand field, these systems exploit bimetallic cooperativity for N2 activation via the bridging Co–N2 –Co motif. Furthermore, the side-on chelation of the potassium ions to the N2 ligand should further enhance charge transfer from Co to N2 . To unravel the effect of the potassium ions, Holland and coworkers conducted a density functional theory (DFT) study of a truncated molecule in the presence and absence of the potassium ions: K2 {Co(C3 N2 H5 )}2 (μ-N2 ) and [{Co(C3 N2 H5 )}2 (μ-N2 )]2− [52]. The corresponding N–N vibrational energies of 1603 and 1742 cm−1 , respectively, suggest that a significant portion of the overall

357

358

7 Group 9 Transition Metal–Dinitrogen Complexes

N2 reduction stems from the addition of reducing equivalents to the Co–N2 –Co core. To date, no N2 functionalization reactions involving K2 {Co(LR,dipp )}2 (μ-N2 ) have been reported. Although the assembly of the K2 Co2 N2 core is beneficial for N2 activation, it may also sterically preclude reagents from approaching the (N2 )2− ligand. 7.1.3.3

Bis(𝛂-imino)pyridine Ligands

In 1990, Gibson et al. reported the first square-planar Co–N2 complex, which was stabilized by the bis(α-imino)pyridine ligand, 2,6-{MeC=N(2,6-iPr2 C6 H3 )}2 C5 H3 N (abbreviated as iPr PDI) [47]. The cationic Co(I) N2 adduct, [Co(N2 )(iPr PDI)]+ , was prepared by abstracting the methylide group in Co(Me)(iPr PDI) with B(C6 F5 )3 under a N2 atmosphere. The resulting N—N bond length of 1.112(6) Å and a N–N vibrational frequency of 2184 cm−1 are consistent with weak activation. The latter is the highest frequency reported for any Co–N2 complex. Gamboratta and coworkers reported a related complex starting from the doubly deprotonated state of the iPr PDI ligand, Li2 [2,6-{CH2 =C—N(2,6-i-Pr2 C6 H3 )}2 C5 H3 N], which comprises two enamido groups. Metalation of this deprotonated ligand with CoCl2 (THF)1.5 generated a linked “dimer” of the pyridine-di-imine (PDI) Co(I) N2 complex, where two enamido carbons undergo intermolecular, radical C—C bond coupling. The product’s N—N bond length of 1.095(6) Å and a N–N vibrational frequency of 2153 cm−1 are comparable to those of [Co(N2 )(iPr PDI)]+ . In their pursuit of lower valent Co complexes, formally Co(0) and Co(–I), bearing the bis(α-imino)pyridine ligands, Chirik and coworkers discovered several square-planar Co–N2 complexes, where the PDI π-system is the primary site of reduction (Figure 7.10) [54]. For instance, the reaction of CoCl2 (iPr PDI) with excess Na/Hg generates the Co(I) N2 complex, Co(N2 )(iPr PDI•− ). The EPR signal has an isotropic g-value of 2.003, which is consistent with an S = 1/2 ligand radical. Substituting Na/Hg with Na(C10 H8 ) results in the doubly reduced product, Na(solv)x [Co(N2 )(iPr PDI2− )], wherein the metal center is Co(I) and the ligand is reduced to the dianionic state, (iPr PDI)2− . The modest changes in the N–N vibrational frequencies for this series further support the conclusion that the Co(I) oxidation state is invariant. Specifically, the N–N band energies shift modestly from 2184 to 2093 to 2046 cm−1 for [Co(N2 )(iPr PDI0 )]+ , Co(N2 )(iPr PDI•− ), and Na(solv)x [Co(N2 )(iPr PDI2− )], respectively. Interestingly, the 𝜈N2 value for Na(solv)x [Co(N2 )(iPr PDI2− )] lowers from 2046 to 1981 cm−1 upon changing the solvent from toluene to THF. Presumably, the latter solvent favors the formation + N

N iPr

+ 1e

iPr

N

N

N i

Pr

[(iPrPDI0)ColN2]+

N

N

–

Co

N

–

Ar

Ar + 1e

Co

i

Pr N

N N

iPr

[(iPrPDI•1–)ColN2]

N

N

–

Ar

Co N N

iPr

[(iPrPDI2–)ColN2]–

Figure 7.10 Co(N2 )(iPr PDI) series showcasing ligand-based reduction. Source: Bowman et al. 2010 [54]. Reproduced with permission from ACS.

7.1 Cobalt–Dinitrogen Complexes

of a bridging N2 interaction between Co and Na+ , i.e. Co–N2 –Na+ . Overall, the generality of the Co(I) N2 electronic structure upholds for other PDI ligand variants, and the range of 𝜈N2 is modest [54, 55]. The redox noninnocence of the PDI class of ligands appears to limit the extent of N2 activation. 7.1.4

N-heterocyclic Carbene Ligands

N-heterocyclic carbenes (NHCs), which are carbenes that are stabilized within a N-heterocycle, increasingly rival the ubiquitous phosphine ligands in organometallic catalysis. The first Co–N2 complex bearing an NHC ligand only appeared in 2012 and was synthesized using a classical NHC ligand, 1,3-dimesitylimidazol-2-ylidene (abbreviated as IMes). The reaction of MeLi and CoCl(IMes)2 led to the formation of Co(N2 )(IMes)(𝜅 2 -IMes’), wherein IMes’ represent a cyclometalated IMes ligand at the benzylic carbon of the mesityl group [56]. The Co(I) N2 complex is diamagnetic and exhibits a N–N vibration at 2006 cm−1 , which is consistent with weak activation. Based on the molecular structure, N2 has a N–N distance of 1.092(3) Å and binds end-on to Co positioned trans to the cyclometalated benzyl group. Hence, the overall geometry is analogs to that of a square-planar Co(I) N2 compound. Lower valent Co(0) and Co(–I) N2 complexes were stabilized using the 1,3-dicyclohexylimidazol-2-ylidene scaffold (abbreviated as ICy) [57]. In a remarkable study, Deng and coworkers reported a highly efficient Co(ICy) catalyst system for the reductive silylation of N2 and isolated potential intermediates in the catalysis. The Co(0) precatalyst, Co(N2 )(ICy)3 , was synthesized by reducing CoCl(ICy)3 with KC8 in a N2 atmosphere. The N2 ligand is characterized by a bond length of 1.051(5) Å and a stretching frequency of 1917 cm−1 . Cyclic voltammetry studies of Co(N2 )(ICy)3 revealed a one-electron reduction at −2.10 V vs. SCE (saturated calomel electrode). Chemical reduction using an alkali metal (M) generated unusual bis(dinitrogen) coordination complexes, Co(μ-N2 )2 (ICy)2 M, where Co is distorted tetrahedral and the oxidation state is formally −I. Whether the alkali reductant used was K, Rb, or Cs, all the resulting Co bis(dinitrogen) products crystallized as isostructural one-dimensional polymers, where the M+ cations interlink the “Co(μ-N2 )2 (ICy)2 ” units via interactions with the N2 ligands and the imidazole N-atoms (Figure 7.11). Although the N2 ligands are each terminally bound (η1 ) to Co, their interaction with the M+ ion can be either η1 or η2 . The N—N bond lengths range from 1.145(6) to 1.162(5) Å. Infrared spectra of the Co(μ-N2 )2 (ICy)2 M complexes display two N–N vibrations (cm−1 ): 1807, 1881 (M = K); 1804, 1888 (M = Rb); and 1811, 1882 (M = Cs). Hence, the group 1 cation has a little effect on the extent of N2 activation. Encapsulation of the potassium ion with 18-crown-6 generates dimeric {Co(N2 )(μ-N2 )(ICy)2 K(18-c-6)}2 , where one N2 ligand is terminally bound only to Co, and the other N2 ligand bridges between Co and K+ . Despite possessing a slightly different N2 binding mode, the N—N bond lengths (1.148(4), 1.161(4) Å) and the stretching frequencies (1812, 1892 cm−1 ) are very similar to their nonencapsulated group 1 analogs. Deng and coworkers have further investigated the reactivity of Co(N2 )(ICy)3 , Co(μ-N2 )2 (ICy)2 M, and Co(N2 )(μ-N2 )(ICy)2 K(18-c-6) with strong acids, such

359

360

7 Group 9 Transition Metal–Dinitrogen Complexes

(a) N3

N4 N1

N2 Co1 C3 N12 N9 C1

Co1 N2

C2

Co2

C1

C4 N10

C2

N11 K1

N3 N1 N4

(b)

K2

(c)

Figure 7.11 Snapshots of the solid-state structure of “KCo(μ-N2 )2 (ICy)2 ” emphasizing (a) the one-dimensional polymer chain; (b) a close-up of the two different K+ interactions with N2 and/or imidazole N-atom; and (c) the tetrahedral Co(N2 )2 center. Source: Go et al. 2018 [57]. Reproduced with permission from ACS.

as triflic acid or HCl in Et2 O [57]. The Co(0) N2 species reacted with strong acids (16 equiv) to give a detectable yield of hydrazine (5%). On the other hand, the Co(–I) bis(dinitrogen) complexes all gave significantly higher yields of hydrazine (23–31%). The Co compounds were also screened for catalytic N2 fixation using KC8 as the reductant and H(Et2 O)[BArF 4 ] or [Ph2 NH2 ]OTf as the proton source. None of the complexes were effective, and only trace amounts of NH3 were observed. Stoichiometric silylation of the N2 adducts was also investigated. Adding 1 equiv of Me3 SiCl or Et3 SiCl to Co(μ-N2 )2 (ICy)2 K resulted in the formation of the Co(II) disilyl diazene adduct, Co(η2 -R3 SiNNSiR3 )(ICy)2 , in a moderate yield. The reaction also generates an unknown Co product, which is speculated to be the Co(0) bis(dinitrogen) species, Co(N2 )2 (ICy)2 . The N—N bond length of 1.457(3) Å in the disilyl diazene adduct is consistent with a single N—N bond. Lastly, the authors demonstrate that all the aforementioned [Co(N2 )1−2 (ICy)3−2 ]0/− complexes are active N2 silylation catalysts, yielding N(SiMe3 )3 to similar extents of 103–125 equiv. The intermediacy of Co(η2 -R3 SiNNSiR3 )(ICy)2 was also supported by its similarly high activity in the catalysis. 7.1.5

Pincer Ligands

In the past decade, the number of Co–N2 complexes bearing pincer and pincer-like ligands has boomed. Pincer ligands are tridentate and typically

7.1 Cobalt–Dinitrogen Complexes

coordinate metal centers in a meridional manner. They are considered privileged ligands because they confer strong thermodynamic stability to the ligated metal complex, which translates into broad applicability in metal-mediated reaction chemistry and catalysis [58, 59]. The pincer ligand donors can comprise vastly different elements, including P, N, C, Si, B, or S. The most ubiquitous pincer ligands are the symmetric variants where the two flanking donors are equivalent. A large majority of the Co–N2 pincer complexes are Co(I), diamagnetic, and square planar with slight distortions. In rare cases, other Co formal oxidation or spin states, such as an S = 1 Co(I), Co(–I), Co(0), and Co(III), have been reported. In 2016, Nishibayashi and coworkers reported a Co PNP-type pincer complex that catalyzes the reduction of N2 to NH3 [60]. The catalyst has good selectivity for NH3 , generating 16 equiv, with only 1 equiv of NH2 NH2 . The central N-pyrrolide donor, which is positioned trans to the N2 binding pocket, plays a crucial role in catalysis. Masuda et al. developed a notable Co pincer catalyst featuring an exotic enamidoiminophosphorane donor [61]. This Co catalyst mediates the reductive silylation of N2 to N(SiMe3 )3 with as high efficiency (200 equiv) as the most active catalysts to date. 7.1.5.1

Monoanionic PNP-Type and PBP-Type Ligands

Early examples of pincer Co–N2 complexes featured PNP-type ligands with two phosphine donors and an anionic amide functionality (Table 7.4) [62, 63]. Using [N(2-Pi-Pr2 -4-MeC6 H3 )2 ]− , which is abbreviated as iPr,Ar PNP, Mindiola and coworkers isolated the dicobalt(I,I) N2 -sandwich complex, {Co(iPr,Ar PNP)}2 (μ-N2 ), by reducing CoCl(iPr,Ar PNP) with t-BuLi under a N2 atmosphere [62]. The molecular structure of {Co(iPr,Ar PNP)}2 (μ-N2 ) shows two square-planar Co centers bridged by an end-on N2 ligand and a staggered disposition of the two pincer ligands. The N–N vibrational frequency of 2024 cm−1 indicates weak activation. Upon reducing CoCl(iPr,Ar PNP) with excess Na(C8 H10 ), a diamagnetic dianionic Co(–I) “Co(N2 )(iPr,Ar PNP)” species was generated [62]. The molecular structure revealed a structurally complex dimer, Table 7.4 The extent of N2 activation in Co–N2 complexes with PNP-type or PBP-type pincer ligands. Complex

Co(n)

N—N (Å)

𝝂N2 (cm−1 )

{Co(iPr,Ar PNP)}2 (μ-N2 )a)

I

1.144(3)

2024

Co(N2 )(Cy Ppyr NP)

I

1.090(8)

2020

Co(N2 )(tBu Ppyr NP)

I

1.117(5)

2016

Co(N2 )(PBP)

I

1.119(2)

2013

Co(N2 )(tBu,Si PNP)

I

—

2004

Co(N2 )(iPr,Et PNP)

I

1.124(2)

1999

{Co(iPr,Ar PNP)(μ-N2 )Na2 (THF)3 }2

−I

1.18a)

1784

Data include the formal Co(n) oxidation state, N—N bond distances, and 𝜈N2 energies (in decreasing order). a) Datum may not be reliable because of the poor quality of the X-ray diffraction data.

361

362

7 Group 9 Transition Metal–Dinitrogen Complexes

Na33 N10 P18

N32 Col

N31 P2

Na34 Col N31

Na34

N32

Na34A

Na33 (a)

(b)

Figure 7.12 Molecular structure of dimeric {Co(iPr,Ar PNP)(μ-N2 )Na2 (THF)3 }2 emphasizing (a) the coordination environment of a single Co site; and (b) the bonding in the {Co(μ-N2 )Na}2 core. Source: Fout et al. 2006 [62]. Reproduced with permission from John Wiley & Sons.

{Co(iPr,Ar PNP)(μ-N2 )Na2 (THF)3 }2 , where two linear Co(N2 ) units are linked via two bridging Na ions (Figure 7.12). Each of these Na ions interacts with an Nβ of one Co(N2 ) unit and with the Co—Nα bond of the other. The dimerization also results in a severe distortion of the Co centers to pseudotetrahedral. The N–N vibrational frequency of 1784 cm−1 indicates moderate N2 activation (cf. 1583, 1529 cm−1 for diazene) and is substantially lower than those of Co(–I) complexes in a tris(phosphine) environment with a main group/transition metal apical donor (cf. 1971–1995, Sections 7.1.2.3 and 7.1.2.4). The greater N2 activation in {Co(iPr,Ar PNP)(μ-N2 )Na2 (THF)3 }2 may be attributed to a combination of the overall dianionic charge on each Co(N2 ) unit, the chelation of Na ions to N2 , and the strongly donating amide group. Other Co–N2 PNP-type systems are known, where the central amido donors and phosphine substituents are different. Although iPr,Ar PNP featured a diarylamido group, other pincer ligands, [N(SiMe2 CH2 Pt-Bu2 )2 ]− (abbreviated as tBu,Si PNP) and [N(CH2 CH2 Pi-Pr2 )2 ]− (abbreviated as iPr,Et PNP), have stronger disilylamido and dialkylamido donors, respectively, and consequently afford Co(I) N2 compounds with greater N2 activation [63, 64]. Other Co(N2 )(PNP) compounds that feature PNP-type ligands have a relatively weak pyrrolide group, [NC4 H2 (CH2 Pt-Bu2 )2 ]− (abbreviated as tBu Ppyr NP) and [NC4 H2 (CH2 PCy2 )2 ]− (abbreviated as Cy Ppyr NP). The N–N stretching frequencies of Co(N2 )(tBu Ppyr NP) and Co(N2 )(Cy Ppyr NP) differ by only a few wavenumbers (cf. 2016 and 2020 cm−1 , respectively), so that one may surmise that the extent of N2 activation is determined primarily by the nature of the trans amido donor and, to a lesser extent, the cis-phosphine substituents [60]. Indeed, the N–N vibrational frequencies of the Co(N2 )(PNP) compounds decrease by 25 cm−1 with increasing amide basicity (Table 7.4). Of relevance, the diazaborole PBP-type pincer ligand, [B(1,2-(NCH2 Pt-Bu2 )2 C6 H4 )]− , also leads to a Co(I) N2 complex with a 𝜈N2 value of 2013 cm−1 [65]. Hence, the strongly σ-donating boryl group is less effective

7.1 Cobalt–Dinitrogen Complexes

Figure 7.13 Co(N2 )(tBu Ppyr NP) and its molecular structure. Source: Kuriyama et al. 2016 [60]. Adapted with permission from John Wiley & Sons.

PtBu2 P2

N

Co

N

N

N1

Co1

N2 N3

P1

PtBu2

at promoting N2 activation than the highly π-basic amido donors. Lastly, the N2 ligands in both Co(N2 )(tBu,Si PNP) and Co(N2 )(PBP) are labile, leading to the isolation of rare three-coordinate, T-shaped Co(I) compounds [63, 65]. To date, the most efficient Co catalyst for the transformation of N2 to NH3 is Co(N2 )(tBu Ppyr NP), which is shown in Figure 7.13 [60]. The optimized catalytic runs were performed at −78 ∘ C under a N2 atmosphere in Et2 O for one hour with KC8 (200 equiv) as the reductant and H(OEt2 )2 [BArF 4 ] (184 equiv, ArF = 3,5-bis(trifluoromethyl)phenyl) as the proton source. Under these conditions, Co(N2 )(tBu Ppyr NP) generated 15.9(2) equiv of NH3 and 1.0(4) equiv of NH2 NH2 , resulting in 17.9 equiv of fixed N atoms. Under slightly modified conditions where the reductant/proton equivalents are reduced to 40/30, respectively, the Co(N2 )(tBu Ppyr NP) catalyst generated only 4.2(1) equiv NH3 , but with no hydrazine. The related Co PNP species, Co(N2 )(Cy Ppyr NP), fared slightly worse with 3.1(1) equiv NH3 , and trace hydrazine. Before this report, [Co(N2 )(TBP)]− was the only complex that had yielded more than 2 equiv of NH3 under similar conditions [23]. Another interesting comparison with Co(N2 )(tBu Ppyr NP) is the reported inactivity of Co(N2 )(PBP), as these Co complexes are isostructural and isoelectronic [23]. Hence, the choice of the donor that is positioned trans to the N2 ligand plays a critical role in catalysis [60]. Intriguingly, both systems show similar extents of N2 activation (Table 7.4). To investigate the differences between a central pyrrolide vs. boryl donor, DFT calculations were performed [60]. Evaluation of Co—Npyr and Co—B Mayer bond orders indicates that the former is slightly greater than the latter. The Mayer bond order for the Co–N2 interaction is also greater for the pyrrolide vs. the boryl group, which is consistent with the greater trans influence of the boryl ligand. The authors proposed that a moderate σ-donor positioned trans to the N2 binding pocket is vital to catalysis, as demonstrated by Co(N2 )(tBu Ppyr NP). 7.1.5.2

Pincer Ligands with N/P Donors

Cobalt(I) N2 complexes are also known for the bis(phosphino)pyridine pincer ligand, 2,6-{CH2 Pi-Pr2 }2 C5 H3 N (abbreviated as iPr Ppy NP). Unlike the amide-based PNP ligands described in Section 7.1.5.1, the iPr Ppy NP ligand is charge neutral, and hence, the corresponding Co(I) N2 complex must bind an additional X-type ligand. Chirik and coworkers reported a series of Co(N2 )(Bpin)(iPr Ppy NP)-type complexes (pin, pinacolate), where the Pi-Pr2 donors are fixed, and only the para substituents on the central pyridine ring are varied [66]. In addition to the parent pyridine donor, the other para substituents include Me, Bpin, and

363

364

7 Group 9 Transition Metal–Dinitrogen Complexes

pyrrole. With the exception of the 4-pyrrole-pyridine system, the remaining Co(N2 )(Bpin)(iPr Ppy NP)-type complexes are unstable to vacuum. The N–N stretching frequencies of 2055 and 2065 cm−1 for the parent system and its 4-Bpin variant, respectively, indicate weak activation. Although IR data were not reported for the 4-pyrrole analog, a N—N bond length of 1.12(2) Å was gleaned from the single-crystal X-ray diffraction study (Figure 7.14) [66]. The pentacoordinate Co(I) center adopts a pseudo-trigonal bipyramidal geometry, where the end-on N2 ligand resides in the equatorial plane along with the two phosphines. These complexes catalyze arene borylation, and the mechanistic proposal involves initial dissociation of N2 to generate the active species, “Co(Bpin)(iPr Ppy NP).” Another related Co(I) N2 complex features a tridentate PNN donor set, comprising a single phosphine donor that is tethered to an enamidoiminophosphorane. The latter is analogs to the β-diketiminate functionality, except that a single phosphorus atom replaces one of the imino carbon atom in the ligand backbone. The ligand as shown in Figure 7.1 is abbreviated as iPr PNpN. The Co(N2 )(iPr PNpN) compound is paramagnetic, which sharply contrasts the other diamagnetic Co(I) N2 pincer complexes [61]. Its molecular structure reveals a distorted tetrahedral Co(I) center, which is also distinct from the typical square-planar geometries of other Co(I) pincer compounds (Figure 7.15). A N–N stretching frequency of 2071 cm−1 indicates even weaker activation by comparison, and indeed, the N2 adduct is in equilibrium with the naked Co counterpart in solution. This trend can be explained by the decrease in π-backbonding from the Co(I) center when the spin state is S = 1 compared to S = 0. Of interest, Co(N2 )(iPr PNpN) catalyzes the reductive silylation of N2 . Using 2000 equiv of Me3 SiCl and 1800 equiv of KC8 under 1 atm N2 in THF at −40 ∘ C, the catalyst generated 200(20) equiv of N(SiMe3 )3 . Hence, Co(N2 )(iPr PNpN) is among the most efficient N2 reductive silylation catalysts to date. ′

Figure 7.14 Co(N2 )(Bpin)(iPr Ppy NP) and its molecular structure. Source: Obligacion et al. 2016 [66]. Adapted with permission from ACS.

N2

N

N4 N1

N3 P2

Co1

N iPr P 2

P1

N2

Co

B1

PiPr2

02

01

PinB

Figure 7.15 Co(N2 )(iPr PNpN) and its molecular structure. Source: Suzuki et al. 2018 [61]. Adapted with permission from ACS.

i

Pr2 P N

– i

Pr

P1 C17

N

Co

PiPr2

N2

C13 N1 Co2 N3

i

Pr

N2

N4

P2

7.1 Cobalt–Dinitrogen Complexes

7.1.5.3

N-heterocyclic Carbene-Based Pincer Ligands

The incorporation of NHCs into the pincer motif has the potential to be a “powerful amalgamation of modern ligand design concepts” [67]. In Co–N2 coordination chemistry, the two most prevalent pincer ligands containing NHC donors are CNC and CCC motifs, where the central donor group is pyridine or arene, respectively. In evaluating Co(dipp Cpy NC) catalysts for olefin hydrogenation (where dipp refers to the 2,6-i-Pr2 (C6 H3 ) substituent on the imidazole ring), Chirik and coworkers uncovered several N2 adducts where the overall molecular charge, Co oxidation state, and the ligand redox states are varied. The compound, [Co(N2 )(dipp Cpy NC)]BArF 4 , adheres to the typical Co(I) end-on N2 square-planar complex, wherein N2 is very weakly activated (2141 cm−1 ) [68]. The N2 activation is slightly worse than in Co(I) N2 pincer complexes but compares well with the Co(I) PDI compounds, which also have pyridine as the central donor. Reduction of [Co(N2 )(dipp Cpy NC)]BArF 4 yields the Co(0) N2 compound, Co(N2 )(dipp Cpy NC), where the 𝜈N2 decreases substantially to 2047 cm−1 . Previously, reduction of Co(I) PDI complexes only resulted in ligand-based reduction, and hence, this result signals one significant difference between the two ligand systems. To be cautious, the dipp CNC ligand is also redox active, although typically, only one electron is stored within the pyridine ring (instead of full delocalization in PDI), leading to radical-based reactivity at the para-position. For example, a unique Co(I) N2 complex was observed upon decomposition of CoH(dipp Cpy NC) in benzene at 22 ∘ C [68]. The product is proposed to be Co(N2 )(4-H-dipp Cpy NC) where the hydride ligand formally migrates to the para-carbon of the pyridine ring, which ultimately transforms the central donor from a neutral σ-donor to an anionic diene-amido functionality. Hence, its 𝜈N2 value of 2048 cm−1 begins to approach that of the other Co(I) N2 pincer complexes listed in Table 7.4. Of interest, in the presence of 1,1-diphenylethylene, CoH(dipp Cpy NC) decomposes in benzene to Co(N2 )(4-CPh2 Me-dipp Cpy NC). Presumably, olefin insertion into the Co—H bond is followed by alkyl migration to the para-carbon of the pyridine ring. Identification of this product was confirmed by a single-crystal X-ray diffraction study. The N—N bond length is 1.112(2) Å, and the 𝜈N2 value of 2050 cm−1 is nearly identical to that of Co(N2 )(4-H-dipp Cpy NC) [68]. Fout and coworkers have investigated the NHC–aryl–NHC pincer ligands (abbreviated as CCC) for supporting Co-based hydrogenation and hydrosilylation catalysis. The flanking carbene donors of the CCC pincer ligands are based on benzimidazole, which are linked to a central phenylene group in a meta relationship. The Co(I) N2 complex, Co(N2 )(dipp CCC), is another example of the square-planar Co(I) N2 coordination motif [69]. As is typical, the complex is diamagnetic, and the N2 ligand is weakly activated (2063 cm−1 ). Coordination of a PPh3 ligand results in a square monopyramidal complex, Co(N2 )(PPh3 )(dipp CCC), where PPh3 occupies the apical position [69]. Despite the additional donor, the 𝜈N2 significantly increases by 50 cm−1 to 2117 cm−1 . The related complexes, Co(N2 )(PPh3 )(mes CCC) and Co(N2 )(PMe3 )(mes CCC) [69, 70], where the aryl substituent is 2,6-Me2 (C6 H3 ), show similar 𝜈N2 values of 2112 and 2114 cm−1 , respectively (Figure 7.16). Hence, the nature of the substituents at the phosphine and at the aryl ring do not significantly perturb the electronic density

365

366

7 Group 9 Transition Metal–Dinitrogen Complexes

C13

N

N

C1

P1

C20 Co1

Co

N Mes

Me3P

N

N5 N6

N

Mes

N

(a)

(b)

Figure 7.16 (a) Co(N2 )(PMe3 )(mes CCC) and (b) its molecular structure. Source: Tokmic et al. 2016 [70]. Adapted with permission from ACS.

at the Co(I) center. Notably, these diamagnetic Co(I) N2 pincer ligands are atypical because of their pentacoordination. In elucidating the reaction chemistry with Si—H bonds in the context of hydrosilylation catalysis, Fout and coworkers observed even more unusual Co(III) N2 complexes: CoH(N2 )(SiHPh2 )(dipp CCC) and Co(N2 )(SiHPh2 )2 (dipp CCC) [71]. Unfortunately, the data for these complexes in regard to N2 activation is scant. The latter species was reported to have a N—N bond distance of 1.102(3) Å. The design of pincer ligands containing an NHC donor is increasingly sophisticated, where NHCs are combined with exotic donors and/or different donor types. Deng et al. prepared an NHC–silyl–NHC pincer ligand (abbreviated as CSiC, Figure 7.1) to investigate the effect of a silyl donor. The Co(I) N2 complex, Co(N2 )(CSiC), displayed a 𝜈N2 band at 2004 cm−1 , which is remarkably low compared to that of other NHC-containing Co pincer complexes (Table 7.5) [72]. The 𝜈N2 frequency compares well with the PNP-type pincer systems, Table 7.5 The extent of N2 activation in Co–N2 complexes featuring NHC donors, group by related ligands. Complex

Co(n)

N—N (Å)

𝝂N2 (cm−1 )

Co(N2 )(IMes)(𝜅 2 -IMes′ )

I

1.092(3)

2006

Co(N2 )(dipp Cpy NC)

0

—

2047

Co(N2 )(4-H-dipp Cpy NC)

I

—

2048

I

1.112(2)

2050

Co(N2 )(4-CPh2 Me-dipp Cpy NC) [Co(N2 )(

dipp

py

F

I

1.076(4)

2141

Co(N2 )(dipp CCC)

C NC)]BAr

I

1.111(3)

2063

Co(N2 )(PPh3 )(mes CCC)

I

1.101(2)

2112

Co(N2 )(PMe3 )(mes CCC)

I

1.022(2)

2114

Co(N2 )(PPh3 )(

dipp

4

I

—

2117

CoH(N2 )(SiHPh2 )(dipp CCC)

CCC)

III

—

—

Co(N2 )(SiHPh2 )2 (dipp CCC)

III

1.102(3)

—

Co(N2 )(CSiC)

I

1.119(3)

2004

Co(N2 )(PNC)

I

0.987(3)a)

2057

Data include the formal Co(n) oxidation state, N—N bond distances, and 𝜈N2 energies. a) The bond length is unrealistically short and inconsistent with 𝜈N2 .

7.1 Cobalt–Dinitrogen Complexes

where N is a strongly donating amido donor, such as Co(N2 )(tBu,Si PNP). The molecular structure shows a N—N bond distance is 1.119(3) Å. More notably, the geometry is strongly distorted from square planar toward a see–saw geometry, where the N2 ligand is bent significantly out of the Co(CSiC) plane with a Si—Co—N bond angle of 131∘ [72]. Presumably, the strongly σ-donating silyl donor is the basis for the geometrical distortion as well as the greater N2 activation across various Co(N2 )(NHC)-type systems. A pincer ligand comprising phosphine–pyridine–NHC donors (abbreviated as Ppy NC, Figure 7.1) has also been investigated in Co–N2 coordination chemistry. The complex, Co(N2 )(Ppy NC), is a typical square-planar Co(I) species with a terminally bound N2 [73]. The N–N stretching frequency of 2057 cm−1 is one of the higher values in Table 7.5 and is closer to the frequencies measured for Co(N2 )(4-H-dipp Cpy NC) and Co(N2 )(4-CPh2 Me-dipp Cpy NC); in common, all three of these Co–N2 pincer complexes have a dearomatized pyridine ring. 7.1.6

Other Assorted Ligands

For bis(o-diisopropylphosphinophenyl)phenylborane (abbreviated as DPB), a Co(0) end-on N2 complex and a dicobalt(0,0) N2 -sandwich complex were cocrystallized as revealed by a single-crystal X-ray diffraction study [74]. In both structures, each Co center is pentacoordinate because of an additional interaction with the ipso-carbon of the phenyl borane group. The monocobalt species, Co(N2 )(DPB), is the dominant species in solution and exhibits a N–N stretching frequency of 2098 cm−1 . In comparison, the tris(phosphine)borane analog, Co(N2 )(TPB) has a slightly lower energy band at 2089 cm−1 [22]. A dicobalt–N2 -sandwich complex was uncovered featuring octahedral Co(II) sites using two supporting ligands: bipyridine (bpy) and the dianionic SPS ligand, [PhP(C6 H4 S)]2− [75]. The SPS ligand adopts a fac-coordination where the central phosphine donor is positioned trans to N2 . The N—N bond distance is 1.156(7) Å. The complex was prepared electrochemically, using an electrolyte solution with the ligands, H2 (SPS) and bpy, in the presence of a sacrificial Co anode. Sterically encumbering meta-terphenyl isocyanides, 2,6-Ar2 (C6 H3 )N≡C, where Ar is mesityl (CNArmes ) or 2,6-diisopropylphenyl (CNArdipp ), have also enabled access to organometallic Co(I) complexes featuring a terminally bound N2 ligand. The trigonal bipyramidal complex, Co(N2 )(SiMe3 )(CNArdipp )3 , wherein N2 is positioned trans to the SiMe3 group, has a N—N bond length of 1.046(7) Å and a corresponding vibrational band at 2231 cm−1 (Figure 7.17) [76]. The latter has the distinction of being the highest frequency reported for any Co–N2 compound to date. It is structurally related to Co(N2 )(SiP3 ) (𝜈N2 = 2063 cm−1 ), although the exchange of the equatorial phosphines with strongly π-backbonding isocyanides largely depletes the electron density at the Co(I) center such that the 𝜈N2 increases by nearly 170 cm−1 . Even with only one isocyanide ligand, the organometallic complex, Co(N2 )(Cp* )(CNArdipp ) where Cp* = C5 Me5 − , has a relatively high 𝜈N2 value of 2110 cm−1 [77]. This complex is unusual in that it remains the only N2 adduct for the “Co(Cp* )” fragment. The binding of small molecules has been actively explored with metal–organic frameworks (MOFs). An in situ X-ray crystallographic study of a single

367

368

7 Group 9 Transition Metal–Dinitrogen Complexes

N3

N

N2

N NC

Co

N1b C1b

C N

N1c

NC

C1c Co1

Si

(a)

C1a N1a Si1

(b)

Figure 7.17 (a) Co(N2 )(SiMe3 )(CNArdipp )3 and (b) its molecular structure. Source: Carpenter et al. 2012 [76]. Adapted with permission from John Wiley & Sons.

N N 2.236(6)Å Co

O (a)

(b)

Figure 7.18 Solid-state structure of Co2 (dobdc)3 ⋅3.8N2 with (a) the view down the channel and (b) zoom of a single Co site. Source: Gonzalez et al. 2012 [78]. Adapted with permission from John Wiley & Sons.

MOF crystal, Co2 (dobdc)3 , where (dobdc)4− is the 2,5-dioxido-1,4-benzenedicarboxylate linker, provided structural evidence for an unusual N2 adduct of a high-spin Co(II) center [78]. The X-ray experiment was performed on a single crystal of Co2 (dobdc)3 under 0.8 bar N2 at 100 K and yielded a structure of Co2 (dobdc)3 ⋅3.8N2 (Figure 7.18). Two types of N2 environments were elucidated: a terminally bound N2 at the CoO5 site and a second N2 near the linker O-atoms that are within van der Waals contact (N· · ·O from 3.44 to 3.77 Å). For the former, the Co—NN2 bond length is exceptionally long at 2.236(6) Å, indicative of an interaction so weak that there is no precedent in molecular systems. In support, a search of the Cambridge Structural Database [79] reveals that the current longest Co—NN2 bond length for a molecular Co end-on N2 complex is 1.933(5) Å, which was reported for Co(N2 )(SiMe3 )(CNArdipp )3 (Co—NN2 : avg. 1.819 Å, SD = 0.039 Å, range 1.73–1.93 Å).

7.1 Cobalt–Dinitrogen Complexes

7.1.7

Analysis and Summary of Cobalt–Dinitrogen Complexes

An analysis of 84 cobalt–N2 complexes with reported 𝜈N2 values was performed to gain deeper insights into N2 activation. The 𝜈N2 values range from 1568 to 2231 cm−1 , with an average value of 2032 cm−1 , and a standard deviation of 106 cm−1 . With one close exception, all the cobalt–N2 complexes have a 𝜈N2 value between that of free N2 (2359 cm−1 ) and diazene (1583, 1529 cm−1 ). Figure 7.19 shows the frequency distribution of the Co–N2 complexes as a function of 𝜈N2 , where a large majority has 𝜈N2 values between 2000 and 2100 cm−1 . At the extreme of greatest N2 activation are the β-diketiminate complexes, K2 {Co(LR,dipp )}2 (μ-N2 ), whose 𝜈N2 values of 1568 and 1599 cm−1 indicate two-electron reduction of N2 to the diazenido ligand, (N=N)2− . The next lowest 𝜈N2 value of 1784 cm−1 corresponds to the dimer, {Co(iPr,Ar PNP)(μ-N2 )Na2 (THF)3 }2 , which comprises two dianionic Co(–I) PNP-type pincer fragments, where the N2 ligand bridges between the Co and Na+ centers. The next compounds with low 𝜈N2 values feature Co(μ-N2 )Mn+ linkages, where M is either a group 1 or 2 cation. These include the Co(–I) bis(dinitrogen) bis(NHC) complexes, Co(μ-N2 )2 (ICy)2 M (∼1809 and 1886 cm−1 ), where M is a group 1 cation, and the Co(μ-N2 )Mg2+ species stabilized in a triphosphine coordination environment, {Co(PR3 )3 (μ-N2 )}2 Mg(THF)4 (1830, 1840 cm−1 ) and {Co(PhBiPr P3 )(μ-N2 )}2 Mg(THF)4 (1863 cm−1 ). In common, all the cobalt compounds with 𝜈N2 < 1900 cm−1 contain an electropositive group 1/group 2 ion that interacts directly with the N2 ligand in an end-on and/or side-on manner. Another interesting finding is that nearly all of the Co complexes with 𝜈N2 values between 1900 and 2000 cm−1 are bimetallic complexes, where the cobalt site is coordinated by phosphine donors. The dicobalt species {Co(μ-t-Bu2 P)N2 (PMe3 )}2 exhibits the lowest 𝜈N2 value (1910 cm−1 ) for a terminally bound η1 -N2 at cobalt with no additional interactions to a group 1/group 2 ion. The Co/group 5 complexes, Na(μ-N2 )Co(μ-R2 PNR′ )3 M=Nt-Bu, where M is Ta or Nb, also feature a Co(μ-N2 )Na+ linkage with a correspondingly low 𝜈N2 value of 1940 cm−1 . The monoanionic [Co(N2 )(EP3 )]− -type compounds, where EP3 represents a tris(phosphine) ligand with an apical donor such as a group 13 (B, Al) or a transition metal (V, Cr, Co) ion, have characteristic 𝜈N2 Figure 7.19 Distribution of cobalt–N2 complexes as a function of 𝜈N2 (in units of 50 cm−1 ).

Co N2 complexes

30

20 15 10 5

N–N bond stretching frequencies (cm–1)

2150

2200

2100

2050

2000

1950

1850

1900

1800

1750

1700

1650

1600

0 1550

Number

25

369

370

7 Group 9 Transition Metal–Dinitrogen Complexes

values between 1970 and 1995 cm−1 . These are largely Co(–I) compounds with an η1 -N2 ligand; the alkali metal ion, which is necessary for charge balance, is encapsulated by crown ethers or cryptands and do not interact with the N2 ligand. There is a notable monometallic exception. The low-coordinate and low-valent Co(N2 )(ICy)3 also has a low 𝜈N2 value of 1917 cm−1 . At the other extreme of N2 activation, Co(N2 )(SiMe3 )(CNArdipp )3 has the largest 𝜈N2 value (2231 cm−1 ) that has been reported for any Co complex. The three aryl isonitrile ligands outcompete the N2 ligand for π-backbonding from the Co(I) center. Other Co complexes with large 𝜈N2 values greater than 2100 cm−1 include the cationic Co(II) compound, [Co(N2 )(CP3 )]+ ; several Co(I) PDI complexes; the cationic Co(I) NHC-type pincer complex with a central pyridine donor, [Co(N2 )(dipp Cpy NC)]+ ; and the five-coordinate Co(I) NHC–pincer/phosphine compounds, Co(N2 )(PR3 )(Ar CCC).

7.2 Rhodium–Dinitrogen Complexes The history of Rh–N2 coordination chemistry spans over approximately 50 years, beginning with RhCl(N2 )(PPh3 )2 in 1967. Although diverse ligand platforms have been investigated, the N2 coordination chemistry of rhodium is quite limited in comparison with cobalt. Nearly, all Rh–N2 compounds contain a Rh(I) center and are square planar. The energetic preference for this geometry may explain the dominance of pincer scaffolds among the ligands known to stabilize Rh–N2 compounds. The discussion is organized by ligand type, and many of the uncommon ligand platforms are shown in Figure 7.20. 7.2.1

Early Rh–N2 Complexes

The first Rh–N2 complex was successfully synthesized in 1967 via the reaction of RhCl(CO)(PPh3 )2 with benzoyl azide to generate the desired Rh(I) product, RhCl(N2 )(PPh3 )2 , with benzoyl isocyanate [80, 81]. The N–N vibrational stretch at 2152 cm−1 suggested a terminally bound N2 ligand that is barely activated. The synthetic procedure was adapted by Kozub and Karpov for making dirhodium–N2 -sandwich complexes that feature a bridging, wide bite-angle bis(phosphine) ligand, such as 1,4-bis(diphenylphosphinomethyl)benzene (referred to as P^P) [82]. As an example procedure, the dirhodium precursor, {Rh(acac)(CO)}2 (μ-P^P) (where acac is acetylacetonate), is mixed with benzoyl azide to form a product that was formulated as {Rh(acac)}2 (μ-N2 )(μ-P^P)•PhC(O)NCO. The presence of N2 in the complex was established by elemental analysis, and displacement of N2 was observed upon treatment of the complex with N,N-dimethylformamide. Moreover, the lack of a N–N vibration by IR spectroscopy suggested a linear Rh–N2 –Rh motif. The authors claimed that the two Rh centers are colinked via the P^P and N2 ligands to generate the first examples of a cyclic N2 complex. The proposed cyclic structure, however, still needs to be confirmed.

7.2 Rhodium–Dinitrogen Complexes

7.2.1 & 7.2.2 Polyphosphine ligands PPh2

Ph2P

t

PtBu2

Bu2P dtbpm

P^P 7.2.3 Ligands with exclusively N-donors R

R N

R

N N

Me

R R

N

R

Me

R N

Me iPrPDI, MePDI,

R = iPr R = Me

R Me

LMe,dmp, R = Me LMe,dipp, R = iPr

7.2.4 Pincer and pincer-type ligands PtBu2 R

PtBu2

C

O

PiPr2

PR2

O

PiPr2

C PtBu2

PtBu2 P*CP

PCP, R = H MePCP, R = Me i

PR2

PtBu2

PiPr2

Pr2P

PARCP, R = tBu iPrPArCP, R = iPr PtBu2

Me2Si

N

N

tBu,Si

tBu,ene

PNP

tBu P 2

PtBu2

PNP

tBu P 2

N

O

PONOP PtBu2

PtBu2

N

PtBu2

Bu2P

PtBu2

PtBu2 PNP

t

O

N

Me2Si iPr,Ar

POCOP

N

N

B N

tBu py t

Bu2P

PBP

tBu py*

P NP

P

t

NEt2

NP

Bu2P

t

NEt2

N

N

PpyNN

Ppy*NN

Bu2P

O

PtBu2 PtBu2

PArOP

7.2.5 N-Heterocyclic carbene ligands iPr

iPr

N

N

N iPr

iPr

IPr

iPr

iPr

N iPr

iPr

SIPr

Figure 7.20 Ligands found in rhodium–N2 coordination complexes, excluding monodentate phosphines.

371

372

7 Group 9 Transition Metal–Dinitrogen Complexes

A unique study by Ozin and Voet investigated the bonding between Rh and N2 in inert gas matrices [83]. A series of Rh(N2 )n species, where n = 1–4, were generated in situ by co-condensing Rh atom vapor and N2 in dilute N2 /Ar matrices and then deposited onto a cryotip cooled to 10 K for IR characterization. The mono-N2 species, Rh(N2 ), showed a band at 2154 cm−1 , which is consistent with an end-on N2 ligand. Based on the total number of IR bands, the Rh(N2 )2 and Rh(N2 )3 analogs were proposed to possess linear (2188 cm−1 ) and trigonal planar (2196 cm−1 ) geometries, respectively. The tetrakis(N2 ) species, Rh(N2 )4 , exhibited two N–N vibrational stretches at 2203 and 2179 cm−1 , which indicates a distortion from an ideal tetrahedral geometry. 7.2.2

Phosphine Ligands

The first crystal structure of a Rh–N2 complex was determined for RhH(N2 ){P(tBu)2 Ph}2 by Ibers and coworkers in 1976 (Figure 7.21) [84]. The Rh–N2 complex was synthesized by the reduction of RhCl3 ⋅3H2 O with sodium amalgam in the presence of P(t-Bu)2 Ph in THF under N2 . The N–N stretching frequency of 2155 cm−1 is consistent with a terminally bound N2 ligand, which was validated by the crystal structure. The structure revealed a square-planar geometry, where the N2 ligand is positioned trans to the hydride. An unexpectedly short N—N bond length of 1.074(7) Å was attributed to disorder in the N2 unit (cf. 1.10 Å for free N2 ). By changing P(t-Bu)2 Ph to a trialkylphosphine ligand, the analogous transRhH(N2 )(PR3 )2 compounds become highly unstable and readily extrude N2 in solution [84, 85]. Because of their instability, these trans-RhH(N2 )(PR3 )2 species were prepared from their trihydride precursors, RhH3 (PR3 )2 , and N2 . The 𝜈N2 energies of 2145, 2130, and 2140 cm−1 for R = t-Bu, Cy, and i-Pr, respectively, are slightly lower than those of RhH(N2 ){P(t-Bu)2 Ph}2 (cf. 2155 cm−1 ). Facile N2 loss from trans-RhH(N2 )(Pt-Bu3 )2 results in the three-coordinate complex, RhH(Pt-Bu3 )2 . On the other hand, N2 extrusion from the PCy3 and the Pi-Pr3 variants proceeds to the dirhodium–N2 -sandwich Figure 7.21 (a) RhH(N2 ) {P(t-Bu)2 Ph}2 and (b) its molecular structure. Source: Hoffman et al. 1976 [84]. Figure adapted with permission from ACS.

P′ tBu

H

2PPh

Rh

H

N

Rh

N(1)

N(2)

N C(3)

tBu

C(6)

P C(2)

2PPh

C(7)

C(5)

C(11) C(14)

C(4)

(a)

(b)

C(8)

C(12) C(13)

C(1)

C(16) C(15)

7.2 Rhodium–Dinitrogen Complexes

complex, {RhH(PR3 )}2 (μ-N2 ), where R = Cy or i-Pr. A molecular structure of {RhH(Pi-Pr3 )}2 (μ-N2 ) shows a linear H–Rh–N2 –Rh–H motif with a N—N bond length of 1.134(5) Å. Building on the precedent of RhCl(N2 )(PPh3 )2 , the halide compounds, trans-RhX(N2 )(PR3 )2 , have been successfully isolated and exhibit increased stability relative to their hydride counterparts. van Gaal et al. prepared trans-RhCl(N2 )(PCy3 )2 by storing RhCl(PCy3 )2 (solv) under 1 atm N2 for five days [86]. Limited reactivity has been reported for RhCl(N2 )(PCy3 )2 : the N2 ligand is displaced by CO, but not by ethylene or H2 . In a subsequent study, van Gaal and van den Bekerom measured the rates of N2 uptake for the halide series, RhX(PCy3 )2 , where X = F, Cl, Br, and I [87]. The N2 -addition rates increase for the RhX(PCy3 )2 complexes in the following order: F < Cl < Br < I. Notably, no N2 uptake was observed for the fluoro complex. On the other hand, the chloro, bromo, and iodo derivatives reacted with N2 fully in 4–5 days, 3 hours, and 15 minutes, respectively. Based on the trend, the authors proposed that increasing electron density at Rh facilitates N2 binding. The resulting N2 complexes, RhX(N2 )(PCy3 )2 , have similar N–N vibrational frequencies of 2100, 2103, and 2108 cm−1 for X = Cl, Br, and I, respectively. In a related work, Perego and coworkers isolated RhCl(N2 )(Pi-Pr3 )2 by refluxing {Rh(μ-Cl)(cyclooctene)2 }2 and Pi-Pr3 in benzene under a N2 atmosphere. The corresponding N–N stretching frequency of 2100 cm−1 is nearly identical to that of the PCy3 analog [88]. The N2 ligand was highly disordered in the single crystal, but the structure refinement confirmed the trans disposition of the Cl and N2 ligands. The N2 ligand was initially proposed to be bound in a side-on manner. However, a careful reinvestigation by Ibers and coworkers concluded that the end-on structure was more feasible. To further prove the end-on binding mode, the authors conducted a 15 N-NMR study of the labeled complex, RhCl(15 N2 )(Pi-Pr3 )2 . The observation of two 15 N peaks at 269 and 302 ppm, with corresponding JRh –N values of 28 and 4 Hz, respectively, provides strong support for a terminally bound N2 at the Rh center [89]. Although further reactivity at the N2 moiety has not been reported in this system, several diazenido compounds have been isolated by the addition of diazo reagents to {Rh(μ-Cl)(Pi-Pr3 )2 }2 , affording trans-RhCl(N2 CRR′ )(Pi-Pr3 )2 , where R = R′ = Ph, p-C6 H4 Me, p-C6 H4 Cl, and R = Ph, R′ = p-C6 H4 Me, o-C6 H4 -Me, CH3 , CH2 Ph, CF3 [90, 91]. The N–N stretching frequencies lie in the range of 1935–1940 cm−1 , which is consistent with end-on coordination of the diazoalkane and shows enhanced activation relative to free N2 CPh2 (cf. 2041 cm−1 ) [92]. The cationic Rh(I) diazoadduct, [trans-Rh(acetone)(N2 CPh2 )(Pi-Pr3 )2 ]PF6 , catalyzes the C—C bond forming processes between ethylene and diazodiarylmethane with extrusion of N2 [93]. Lastly, using the cis-chelating bis(phosphine) ligand, t-Bu2 PCH2 Pt-Bu2 (abbreviated as dtbpm), Hofmann and coworkers isolated a dirhodium–N2 -sandwich compound, {RhNp(𝜅 2 -dtbpm)}2 (μ-N2 ), where Np is neopentyl [94]. The complex was prepared by exposing a pentane solution of RhNp(𝜅 2 -dtbpm) to N2 . The molecular structure shows that the N2 ligand is positioned trans to two phosphine donors and has a N—N bond length of 1.106(6) Å.

373

374

7 Group 9 Transition Metal–Dinitrogen Complexes

7.2.3

Ligands with Exclusively Nitrogen Donors

7.2.3.1

Bis(𝛂-imino)pyridine Ligands

The bis(α-imino)pyridine scaffold has also been used to synthesize square-planar Rh–N2 complexes. Burger and coworkers isolated both a terminally bound N2 complex and a dirhodium–N2 -sandwich complex by varying the size of the PDI ligand (Figure 7.22) [95]. The more sterically encumbered ligand, 2,6-{MeC=N(2,6-i-Pr2 C6 H3 )}2 C5 H3 N (abbreviated as iPr PDI), stabilizes the monorhodium product, Rh(N2 )(iPr PDI). On the other hand, the less bulky methyl variant, 2,6-{MeC=N(2,6-Me2 C6 H3 )}2 C5 H3 N (abbreviated as Me PDI), results in the sandwich species, {Rh(Me PDI)}2 (μ-N2 ), where the planar Rh(Me PDI) units are staggered. The N—N bond length in the sandwich complex (1.130(7) Å) is substantially elongated relative to that of the terminal N2 complex (1.087(3) Å). The N–N vibrational frequencies of 2139 and 2059 cm−1 for Rh(N2 )(iPr PDI) and {Rh(Me PDI)}2 (μ-N2 ), respectively, further support a greater extent of N2 activation in the sandwich species. Of interest, both Rh compounds were reported to have an EPR signal with a g-value of 2. Considering the propensity of PDI scaffolds to undergo redox reactions, the EPR data are consistent with an S = 0 d8 Rh(I) center and a ligand-based radical (PDI•− ), as was described for the Co PDI complexes (Section 7.1.3.3). In the same study, the authors also targeted a Rh PDI nitride species via the thermal decomposition of the azido compound, Rh(N3 )(iPr PDI) [95]. In a thermogravimetric study of the azide complex, N2 loss was observed at temperatures above 200 ∘ C. However, the putative Rh nitride is too unstable and rapidly decomposes by two C—H bond activation processes involving the isopropyl groups of the ligand. No evidence for a bimolecular coupling of two Rh≡N species to generate a dirhodium–N2 -sandwich compound was found. Such a transformation is highly intriguing to researchers in the field of N2 activation, as it would provide mechanistic plausibility for N2 splitting by two metal centers to generate two metal nitrides, via the principle of microscopic reversibility [96]. So far, the potential of Rh PDI systems to mediate either N—C or N— Si bond formation at the N2 ligand is undemonstrated. In theory, the resulting products are likely to be stable, as shown by the isolation of a diazomethane adduct, [(iPr PDI)Rh(N2 CHSiMe3 )][BArF 4 ], where BArF 4 = tetrakis(3,5-

C16

N3 N1

C5

Rh1 N4 N4

N1

C6

Rh1′

N2

C1 C6

(a)

C5

C1

C8 Rh1 N3

N4 N5

N2 (b)

Figure 7.22 Molecular structures of (a) {Rh(Me PDI)}2 (μ-N2 ) and (b) Rh(N2 )(iPr PDI). Source: Schöffel et al. 2010 [95]. Reproduced with permission John Wiley & Sons.

7.2 Rhodium–Dinitrogen Complexes

bis(trifluoromethyl)phenyl)borate [97]. The molecular structure and IR data of [(iPr PDI)Rh(N2 CHSiMe3 )]+ reveal a N—N bond length of 1.067(5) Å and corresponding vibrational energy of 2030 cm−1 (cf. 2069 cm−1 for free N2 CHSiMe3 ). Hence, the complex possesses a weakly activated N≡N triple bond within the Rh—N≡N—CHSiMe3 fragment. This complex was further shown to react with aromatic aldehydes via N2 extrusion and homologation of the carbene fragment with the aldehyde functionality to form the corresponding trimethylsilyloxy ethylene product, ArCH=CHOSiMe3 , with higher selectivity for the cis-isomer. 7.2.3.2

𝛃-diketiminate Ligands

Despite the rich Co–N2 coordination chemistry with β-diketiminate platforms, only a couple Rh–N2 complexes bearing the β-diketiminate scaffold have been reported. Using the ligand, [CH{C(Me)N(2,6-Me2 C6 H3 )}2 ]− (abbreviated as LMe,dmp ), Budzelaar and coworkers isolated the three-coordinate Rh(COE)(L) compound (where COE is cyclooctene), which reacts reversibly with N2 . Under a N2 atmosphere, the N2 adduct, Rh(N2 )(COE)(L), persists and characterized by a N–N stretching frequency of 2172 cm−1 , which is indicative of end-on binding. Later, Stephan and coworker isolated and structurally characterized a related Rh–N2 species with the more bulky LMe,diip variant that contains diisopropylphenyl substituents. By mixing Li(Et2 O)[LMe,dipp ] and 0.5 equiv {Rh(COE)2 (μ-Cl)}2 , the end-on N2 product, Rh(N2 )(COE)(LMe,dipp ), was isolated with a N–N vibrational frequency of 2175 cm−1 [98]. The molecular structure shows a square-planar geometry with a N—N bond length of 1.091(6) Å. 7.2.4

Pincer Ligands

The coordination chemistry of Rh–N2 compounds gained momentum in the mid-1990s with the rise of pincer ligands. Utilizing a bis(phosphine) pincer platform, Milstein and coworkers isolated and investigated the first pincer-supported Rh(I) N2 complexes [99]. To date, Rh–N2 complexes bearing a pincer ligand account for nearly half of all structurally characterized Rh–N2 complexes, which is a testament to the stability engendered by the pincer scaffold. The PNP-type pincer ligands have been successfully used in the N2 coordination chemistry of all group 9 metals. However, the PCP pincer scaffold, which is commonly seen for the heavier group 9 metals, is notably absent among Co–N2 complexes. Table 7.6 collects relevant data for the Rh–N2 pincer complexes. Many of the Rh–N2 pincer complexes equilibrate in solution between a monorhodium end-on N2 species and a dirhodium bridging N2 complex. Another interesting facet of pincer ligands is their ability to support both low-valent Rh–N2 complexes and high-valent Rh nitrides. This duality is noteworthy and has inspired creative efforts to chemically connect the N2 and nitride species via N—N bond scission and to incorporate these steps in a catalytic cycle. Although no catalysis has been demonstrated yet, the Rh nitride species has been successfully isolated and shown to bimolecularly couple to generate a transient dirhodium–N2 -sandwich complex, which then reacts with additional N2 to generate two discrete Rh–N2 compounds. These chemical transformations collectively represent the microscopic reverse of N2 scission, and intense

375

376

7 Group 9 Transition Metal–Dinitrogen Complexes

Table 7.6 The extent of N2 activation in Rh–N2 complexes for different types of pincer ligands. Complex

Rh(n)

N—N (Å)

𝝂N2 (cm−1 )

Rh(N2 )(PCP)

I

—

2108

Rh(N2 )(PMe CP)

I

0.963(14)a)

2110

Rh(N2 )(PC* P)

I

—

2124

Rh(N2 )(PAr CP)

I

—

2133

Rh(N2 )(iPr PAr CP)

I

1.108(3)a)

2165

Rh(N2 )(POCOP)

I

1.122(11)a)

2162

Rh(N2 )(iPr,Ar PNP)

I

1.1191(17)a)

2115

Rh(N2 )(tBu,Si PNP)

I

1.160(13)b)

2092 2120 2153

I

1.091(4)b)

[Rh(N2 )(tBu Ppy NP)]OTf

I

1.116(4)b)

Rh(N2 )(tBu Ppy* NP)

I

—

2122

[Rh(N2 )(PONOP)]BArF 4

I

1.063(5)b)

2202

{Rh(PBP)}2 (μ-N2 )

I

1.106(4)a)

2063c)

[Rh(N2 )(P NN)]BF4

I

1.124(3)a)

2157

Rh(N2 )(Ppy* NN)

I

—

2110

Rh(N2 )(PAr OP)

I

1.129(4)a)

2095

Rh(N2 )(

tBu,ene

PNP)

py

Data include the formal Rh(n) oxidation state, N—N bond distances,a) and 𝜈N2 energies. a) The N—N bond lengths are taken from the corresponding dirhodium–N2 -bridging complexes. b) This N—N bond length is taken from the mononuclear Rh–N2 complex. c) Determined by Raman spectroscopy.

research efforts are underway to promote a favorable energy landscape for N2 scission. Similar efforts are also being pursued with related Ir pincer systems. 7.2.4.1

PCP Pincer Ligands

The first pincer-supported Rh–N2 complex featured a PCP-type pincer ligand, [HC(CH2 CH2 P(t-Bu)2 )2 ]− (abbreviated as PCP), where the central donor is an anionic sp3 carbanion (Figure 7.23) [99]. The complex, Rh(N2 )(PCP), was synthesized by reducing the Rh(III) precursor, RhHCl(PCP), with excess NaH under a N2 atmosphere. The N–N stretching frequency of 2108 cm−1 is consistent with a terminally bound N2 complex. The N2 ligand can be displaced by CO, CO2 , H2 , and ethyt P Bu2 lene. Except for CO, the binding of the other substrates is reversible, and the N2 adduct can be reformed. Milstein H C Rh N N and coworkers have also investigated the PMe CP ligand variant, where the proton on the carbide is swapped for PtBu2 a methyl group [100]. The corresponding N2 complex, Me Figure 7.23 Rh(N2 )(PCP). Rh(N2 )(P CP), showed a nearly identical energy for the

7.2 Rhodium–Dinitrogen Complexes

N–N vibration of 2110 cm−1 and has a N—N bond length of 0.963(14) Å, both of which are indicative of weak activation (Table 7.6) [100]. Also of note, the PCP ligand backbone can be noninnocent and undergoes Rh-mediated C—H bond activations that ultimately lead to another isostructural Rh(I) N2 complex where, the PCP ligand has been desymmetrized to [(t-Bu)2 PCH2 CH2 C=CHCH2 P(t-Bu)2 ]− (abbreviated as PC* P), and the central donor is transformed into a vinylic sp2 carbanion [101]. The Rh(N2 )(PC* P) compound has a slightly higher N–N stretching frequency of 2124 cm−1 . Another PCP-type scaffold features a central arene donor, [2,6-(CH2 PR2 )2 C6 H3 ]− (abbreviated as PAr CP for R = t-Bu; and iPr PAr CP for R = i-Pr). The former ligand was found to stabilize a Rh(N2 )(PAr CP) species, where the N–N stretching frequency of 2133 cm−1 is slightly higher in energy than that of the PCP and PMe CP analogs [102]. The lower extent of N2 activation is consistent with the less electron-donating nature of the trans sp2 -carbanion donor in PAr CP, compared to the sp3 -carbanion in PCP and PMe CP. The Pr PAr CP ligand, which bears less bulky phosphine substituents, enables more dynamic N2 binding. The equilibrium between the mononuclear Rh(I) N2 species, Rh(N2 )(iPr PAr CP), and the dirhodium–N2 -sandwich complex, {Rh(iPr PAr CP)}2 (μ-N2 ), is sensitive to concentration, temperature, and N2 pressure (Figure 7.24) [103]. At ∼0.4 M, room temperature under 1 atm N2 , both the mono- and dirhodium species are observed in a ratio of 8 : 3. Lowering the temperature to −50 ∘ C evens the ratio to 1 : 1. Rh(N2 )(iPr PAr CP) displays a N–N vibration band at 2165 cm−1 , which is slightly higher in energy than that of the PCP and PMe CP analogs. The lower extent of N2 activation is consistent with the less electron-donating nature of the trans sp2 -carbanion donor in PAr CP, compared to the sp3 -carbanion in PCP and PMe CP. The molecular structure of {Rh(iPr PAr CP)}2 (μ-N2 ) reveals a N—N bond length of 1.108(3) Å and a perpendicular disposition of the two “Rh(iPr PAr CP)” fragments. The N2 ligand of Rh(N2 )(iPr PAr CP) can be displaced by diazodiphenylmethane to form a transient Rh diazo adduct, which then undergoes further reactivity to give a mixture of the Rh carbene complex, Rh(CPh2 )(iPr PAr CP), Rh(N2 )(iPr PAr CP), and stilbene. The reactivity of the Rh(N2 )(iPr PAr CP) system with benzalazine, PhHC=N—N=CHPh, has also been investigated [104]. The asymmetric cleavage of the azine substrate into nitrile and imine molecules is supported by the isolation of their adducts, Rh(N≡CPh)(iPr PAr CP) and Rh(HN=CHPh)(iPr PAr CP), respectively (Figure 7.25). Despite the overall cleavage of the N—N bond, the mechanism involving direct N—N bond cleavage was ruled out based on DFT calculations. Rather, the authors speculate that the Rh azine adduct, PiPr2

PiPr2

2

iPr 2

– N2 Rh PiPr2

N

N

Rh PiPr2

N

N

P Rh

P Pr2

i

Figure 7.24 Equilibrium between Rh(N2 )(iPr PAr CP) and {Rh(iPr PAr CP)}2 (μ-N2 ).

377

378

7 Group 9 Transition Metal–Dinitrogen Complexes

PiPr2 Ph

PiPr2

2

2 Benzalazine Rh

N

N

– N2

CH

2

Rh

N N

PiPr2 HC

PiPr2

Ph – Benzalazine PiPr2 Ph

PiPr2

CH Rh

N

C

Ph

+

Rh

N H

PiPr2

PiPr2

Figure 7.25 Catalytic benzalazine cleavage mediated by Rh(N2 )(iPr PAr CP). Source: Cohen et al. 2003 [104]. Adapted with permission from John Wiley & Sons.

which is detected spectroscopically, undergoes C—H bond activation involving another Rh complex. With 10 equiv of benzalazine, the Rh(N2 )(iPr PAr CP) system generates benzonitrile catalytically (7 equiv). However, product inhibition of the catalyst precludes higher turnovers even with 100 equiv of the azine substrate. Bis(phosphinite) pincer ligands with a central arene donor have also yielded similar coordination chemistry as their bis(phosphine) counterparts. The [2,6-(OPi-Pr2 )2 C6 H3 ]− ligand (abbreviated as POCOP) also enables a dynamic equilibrium between Rh(N2 )(POCOP) and {Rh(POCOP)}2 (μ-N2 ) [102]. Unlike the isostructural bis(phosphine) iPr PAr CP system, the POCOP system exhibits a stronger thermodynamic preference (9 : 1 ratio) for the bridging N2 dirhodium complex in solution at ambient temperature. The molecular structure of {Rh(POCOP)}2 (μ-N2 ) shows a N—N bond length of 1.122(11) Å [102]. A weak IR band at 2162 cm−1 , which likely corresponds to Rh(N2 )(POCOP), is close in value to that for Rh(N2 )(iPr PAr CP), despite the expectation that phosphinite ligands are more π-acidic than their phosphine counterparts. Rh(N2 )(POCOP) complex reacts with CO and C2 H4 , which displace the N2 ligand. Treatment of MeI or benzyl chloride to the Rh(I) species results in oxidative addition of substrates to generate Rh(III) organometallic compounds with loss of N2 . 7.2.4.2

PNP Pincer Ligands

Rh–N2 compounds are well known for various PNP pincer ligands, where the central N-donor is either an anionic amide or a neutral pyridine donor. The amido ligands, iPr,Ar PNP and tBu,Si PNP (Figure 7.1), have been used to support Rh(I) N2 complexes. The 𝜈N2 energies for Rh(N2 )(iPr,Ar PNP) and Rh(N2 )(tBu,Si PNP) of 2115 and 2092 cm−1 , respectively, indicate a terminally bound N2 ligand that is weakly activated [105, 106]. Of interest, the isostructural Co systems are known; for each ligand scaffold, the Co counterparts exhibit a greater N2 activation by ∼90 cm−1 (Section 7.1.5.1). Rh(N2 )(iPr,Ar PNP) also exists in a solution equilibrium with the dirhodium–N2 -sandwich complex, {Rh(iPr,Ar PNP)}2 (μ-N2 ), which contains a N—N bond length of 1.1191(17) Å [105]. A PNP platform with

7.2 Rhodium–Dinitrogen Complexes tBu

P

N

Rh

N3

0.5

P

hν N – N2

P

tBu

tBu 2

2

Rh

t

N

Bu2P

P

tBu

2

2

P tBu 2

PtBu2

N

P

Rh

Rh N

tBu 2

N

N

N2

N2

tBu P 2

Rh

PtBu2

N

Figure 7.26 Rh–N2 complexes from bimolecular coupling of an in situ Rh nitride. Source: Scheibel et al. 2013 [107]. Adapted with permission from ACS.

a diene-amido donor, [N(CH=CHPt-Bu2 )2 ]− (abbreviated as tBu,ene PNP), also provides both the mono and dirhodium–N2 complexes. Rh(N2 )(tBu,ene PNP) has a N—N bond length of 1.091(4) Å and a corresponding vibrational frequency of 2120 cm−1 [107]. The Rh(N2 )(tBu,ene PNP) system is highly unique in that both Rh–N2 species are observed as decomposition products of an unstable Rh(IV) nitride (Figure 7.26). By irradiating the azido precursor, Rh(N3 )(tBu,ene PNP), in a frozen solution, the nitrido species was generated and confirmed by in situ EPR and IR spectroscopy. Warming of the irradiated solution generates a mixture of the mono- and dirhodium–N2 compounds, providing evidence for a bimolecular coupling of Rh≡N to a dirhodium–N2 -bridging complex, which can further react with N2 to generate 2 equiv of the monorhodium–N2 adducts. The use of a central pyridine donor in the PNP platform can also stabilize the square-planar Rh(I) N2 complex, despite the difference in charge. The PNP variant, 2,6-(t-Bu2 PCH2 )2 C5 H3 N (abbreviated as tBu Ppy NP), provides a reactive precursor, Rh(OTf )(tBu Ppy NP), where the triflate ligand is highly labile and can be partly displaced by N2 to form [Rh(N2 )(tBu Ppy NP)]OTf [108]. The cationic charge of the molecular ion is likely responsible for the slightly lesser N2 activation in this species (2153 cm−1 , 1.116(4) Å) relative to other Rh PNP systems (Table 7.6). Treating [Rh(N2 )(tBu Ppy NP)]OTf with a strong base results in the deprotonation of a methylene proton in the ligand backbone, which transforms the neutral pyridine donor into an anionic, asymmetric diene amide [108]. The product, Rh(N2 )(tBu Ppy N* P), has a 30 cm−1 lower N–N stretching frequency of 2122 cm−1 , which is nearly identical to that of Rh(N2 )(tBu,ene PNP). Bis(phosphinite) pincer ligands with a central pyridine donor have also yielded similar coordination chemistry as their bis(phosphine) counterparts. The PONOP ligand, 2,6-(t Bu2 PO)2 C5 H3 N, was used to support a reactive Rh(I) complex, [Rh(𝜅 1 -ClCH2 Cl)(PONOP)]BArF 4 , with a labile CH2 Cl2 ligand [109]. Crystallization of this molecule under N2 also resulted in the serendipitous discovery of the terminally bound N2 adduct, [Rh(N2 )(PONOP)]BArF 4 , which has a N–N distance of 1.063(5) Å and a N–N vibration of 2202 cm−1 [109]. The cationic charge of the molecular ion and the more electron-withdrawing phosphinite donors can account for the uncharacteristically high 𝜈N2 energy for a Rh(I) N2 pincer complex (Table 7.6). Indeed, the N–N vibration for

379

380

7 Group 9 Transition Metal–Dinitrogen Complexes

the isostructural bis(phosphine) complex, [Rh(N2 )(tBu Ppy NP)]+ , is lower by nearly 50 cm−1 . 7.2.4.3

Other Pincer Ligands

The diazaborole PBP-pincer ligand, [B(1,2-(NCH2 Pt-Bu2 )2 C6 H4 )]− , stabilizes the N2 -bridging dirhodium complex, {Rh(PBP)}2 (μ-N2 ), where the N–N stretching frequency of 2063 cm−1 was determined by Raman spectroscopy [110]. The molecular structure shows a N—N bond length of 1.106(4) Å, which compares well with the Rh–N2 –Rh motifs featuring the PCP- or PNP-type ligands. The central pyridine donor has also been combined with a mixed phosphineamine donor set. The Ppy NN ligand, which corresponds to 2-(t-Bu2 PCH2 )6-(Et2 NCH2 )C5 H3 N, is a neutral pincer ligand that also supports the Rh(I) N2 species. The terminally bound N2 complex, [Rh(N2 )(Ppy NN)]BF4 , exists in equilibrium with the dirhodium–N2 -bridged species [111]. The former gives rise to an IR-active N–N vibration at 2157 cm−1 , and the latter is characterized by a N—N bond length of 1.124(3) Å. Akin to the tBu Ppy NP analog, the addition of a strong base to [Rh(N2 )(Ppy NN)]BF4 leads to deprotonation of a methylene proton in the ligand backbone, which transforms the neutral pyridine donor into an anionic, asymmetric diene amide [111]. The product, Rh(N2 )(Ppy* NN), shows a 47 cm−1 lower N–N stretching frequency of 2110 cm−1 . Another pincer-type variation is a bis(phosphine) ligand with a central aryloxide donor, [2,6-(CH2 Pt-Bu2 )2 -4-CH3 (C6 H3 O)]− (abbreviated as PAr OP). An equilibrium between the monorhodium–N2 adduct and the dirhodium–N2 -bridging complex was observed by the solution state [112]. The former gives rise to an IR-active N–N vibration at 2095 cm−1 , and the latter is characterized by a N—N bond length of 1.129(4) Å [112]. The formal exchange of an arene donor in Rh(N2 )(PAr CP) (cf. 2133 cm−1 ) for an anionic p-cresol donor results in a 38 cm−1 red shift of the N–N stretching frequency. 7.2.5

N-heterocyclic Carbene Ligands

NHCs are strong σ-donors and have gained popularity as better alternatives to the ubiquitous phosphine ligand family. Crudden and coworkers isolated the NHC counterparts of the Rh(N2 )Cl(PR3 )2 compounds: Rh(N2 )Cl(IPr)2 and Rh(N2 )Cl(SIPr)2 , where IPr = 1,3-bis(2,6-diisopropylphenyl)imidazole-2-ylidene and SIPr = 1,3-bis(2,6-diisopropylphenyl)imidazolidine [113, 114]. These complexes were prepared by metalating the NHC ligand with {Rh(C2 H4 )2 Cl}2 as the Rh(I) source. Like their bis(phosphine) cousins, the N2 ligands in Rh(N2 )Cl(NHC)2 are positioned trans to Cl. Rh(N2 )Cl(IPr)2 has a N—N bond length of 1.100(6) Å and a N–N vibration of 2103 cm−1 [113]. Similarly, Rh(N2 )Cl(SIPr)2 has a N—N bond length of 1.11(1) Å and a N–N vibration of 2107 cm−1 [114]. The extent of N2 activation is equivalent to that observed in Rh(N2 )Cl(PR3 )2 species. The N2 ligand can be displaced by H2 , O2 , and CO. [113] Exposure of single crystals of Rh(N2 )Cl(SIPr)2 to O2 results in full conversion to single crystals of Rh(η2 −O2 )Cl(SIPr)2 , where the O2 ligand is bound in a side-on manner [114].

7.3 Iridium–Dinitrogen Complexes

Rh N2 complexes 16

Number

12 8 4

2200

2150

2100

2050

2000

0

N–N bond stretching frequencies (cm–1)

Figure 7.27 Distribution of rhodium–N2 complexes as a function of 𝜈N2 (in units of 50 cm−1 ).

7.2.6

Summary of Rhodium–Dinitrogen Complexes

An analysis of the 36 rhodium–N2 complexes with reported 𝜈N2 values provides insights into N2 activation, especially in comparison with the Co analogs. Excluding the diazo adducts, the 𝜈N2 values of Rh–N2 compounds range from 2013 to 2203 cm−1 , with an average value of 2133 cm−1 and a standard deviation of 40 cm−1 . Figure 7.27 shows the frequency distribution of the Rh–N2 complexes as a function of 𝜈N2 , where the large majority has 𝜈N2 values between 2100 and 2200 cm−1 . Hence, not only are Rh–N2 compounds fewer in number than their Co counterparts, but also the range of N2 activation at Rh is much more limited. This is a consequence of the favorable Rh(I) square-planar motif, which dominates nearly all the structures. Notable exceptions are the homoleptic Rh(N2 )n species, where n = 1–4, which were generated and studied in a gas matrix. These species are formally Rh(0), and their lack of strong donors rationalizes their possession of nearly all the high N–N stretching frequencies in this set.

7.3 Iridium–Dinitrogen Complexes The first isolable dinitrogen complex using a group 9 metal was the iridium complex, trans-IrCl(N2 )(PPh3 )2 [115]. Despite the promising beginning, Ir–N2 coordination compounds currently number the fewest among the group 9 triad. Many of the ligand platforms are shown in Figure 7.28. Although a multitude of aryldiazo adducts of Ir has been reported, none of these complexes are derived from N2 functionalization at the Ir center. The redox flexibility of Ir could be advantageous for N2 functionalization catalysis that necessitates a metal center to traverse multiple oxidation states. Of relevance, the complete scission of N2 into metal nitrides would formally involve a +3 change in the metal’s oxidation state. The reverse reaction, or the coupling of two metal nitrides to form a N2 -sandwich complex, has been elucidated for Ir [116]. Understanding the mechanistic and electronic details or this reaction may lend unique insights into designing new metal compounds for the forward reaction.

381

382

7 Group 9 Transition Metal–Dinitrogen Complexes 7.3.2 Phosphine ligands

i

Pr2P

PiPr2

iPr P 2

Si PiPr2 PCH2Ad(iPr)2

SiP3

7.3.3 Ligands with exclusively N-donors

R R

N N

R

i

N N N B R′ H R′

R′

i

Pr

i

N

N

Pr

Me

Me,Me

Tp Tptol,

Pr

N

, R = Me, R′ = Me R = 4-Me-C6H4, R′ = H

i

Pr

Me LMe,dipp

7.3.4 Pincer and pincer-type ligands

PiPr2

iPr P 2

N

PtBu2

PtBu2

PtBu2

PtBu2

O

PR2

O

PR2

N

iPr,ArPNP

tBu,enePNP

PArCP

tBu

POCOP, R = tBu R = 2,4,6-(CF3)3-C6H2

ArPOCOP,

PtBu2

R2P

PR2 Si

H

Me

PtBu2

iPr,Ar

PSiP, R = iPr R = Cy

PCyCP

Cy,ArPSiP,

7.3.5 N-heterocyclic carbene ligands

N

N IMes

7.3.6 Miscellaneous

N

N,C-phenylquinolyl

Figure 7.28 Ligands found in iridium–N2 coordination complexes, excluding simple phosphines.

7.3.1

Early Ir–N2 Complexes

The iridium–dinitrogen complex, trans-IrCl(N2 )(PPh3 )2 , was reported by Collman and Kang in 1966 [115] and has the historical distinction of being the second isolated metal–N2 complex after [Ru(N2 )(NH3 )5 ]2+ [4]. By reacting Vaska’s complex IrCl(CO)(PPh3 )2 with various organic azides (RN3 ), IrCl(N2 )(PPh3 )2 was isolated as a yellow solid (Figure 7.29). The infrared spectrum of trans-IrCl(N2 )(PPh3 )2 showed a band at 2105 cm−1 , which was assigned

7.3 Iridium–Dinitrogen Complexes

CO

L

+

Ir

O

N

L

NCAr Ir

ArCN3

L

Cl

N

O

Cl

L

C O

O L Cl

+ N

C Ar

N

L Ir

Ir C

O–

L L = (C6H5)3P

Cl

O

N +

ArCNCO

L ROH

O O

O

CNH COR

Figure 7.29 Synthesis of trans-IrCl(N2 )(PPh3 )2 . Source: Collman et al. 1968 [117]. Adapted with permission from ACS.

to the N–N vibrational mode [117, 118]. The reaction was serendipitous in that the isocyanate by-product, which would normally bind the Ir center and affect the N2 loss, was intercepted by EtOH, which was present in the solvent as an inhibitor, to generate carbamate ester. In support, the Ir isocyanate adduct, IrCl(η2 -RN=C=O)(PPh3 )2 , becomes the product of the reaction when alcohol is absent. The compound, IrCl(N2 )(PPh3 )2 , was found to react with alkyl halides, carbon tetrachloride, and acetylenes to drive off the labile N2 ligand and form the corresponding new complexes [115, 117, 119]. With diethyl maleate (cis-EtO2 CCH=CHCO2 Et), an Ir product was isolated where N2 remained bound. The product was proposed as the formal Ir(III) diethyl maleate adduct, IrCl(N2 )(η2 -(CHCO2 Et)2 )(PPh3 )2 . The higher formal oxidation state and coordination number potentially account for the substantial increase in the N–N vibrational frequency to 2190 cm−1 [117]. Oxidative addition reactions of MeOTf or HX with IrCl(N2 )(PPh3 )2 also provide six-coordinate Ir(III)–N2 adducts [120, 121]. The Ir(III) product of MeOTf addition, IrCl(N2 )(PPh3 )2 (CH3 )(OTf ), displays an even higher N–N frequency of 2215 cm−1 . 7.3.2

Phosphine Ligands

Goldman and Halpern trapped the unsaturated polyhydride complex, “Ir(H)3 (Pi-Pr3 )2 ,” at low temperatures as a mixture of the N2 adducts, Ir(N2 )(H)3 (Pi-Pr3 )2 , and {Ir(H)3 (Pi-Pr3 )2 }2 (μ-N2 ) [122]. The binding of N2 was confirmed by the NMR spectroscopic observation of 15 N–P coupling (1.8 Hz) for the corresponding 15 N2 -labeled species. Although no IR data were reported, an equilibrium constant of ∼1.5 was measured between the two Ir–N2 compounds at −35 ∘ C, according to the reaction: Ir(N2 )(H)3 (Pi-Pr3 )2 + N2 = 2 Ir(N2 )(H)3 (Pi-Pr3 )2 .

383

384

7 Group 9 Transition Metal–Dinitrogen Complexes

P H

Ir

N

N

RT –N2

P

P

H

Ir

N2 P

H

Figure 7.30 Equilibrium reaction of IrH(N2 )(P(CH2 Ad)i-Pr2 )2 and Ir(H)2 (P(CH2 Ad)i-Pr2 ) (𝜅 2 -P(CH2 Ad* )i-Pr2 ). Source: Millard et al. 2010 [123]. Adapted with permission from ACS.

Using the unique monodentate phosphine ligand, P(CH2 Ad)i-Pr2 (Ad is adamantyl), Figueroa and coworkers isolated and crystalized IrH(N2 )(P(CHAd)iPr2 )2 , where the N2 ligand is positioned trans to the hydride [123]. The N—N bond length is 1.108(4) Å, and the N–N stretching frequency is 2127 cm−1 . Dissolution of the complex led to partial N2 loss. The subsequent intramolecular C—H bond activation of one Ad group generated the cyclometalated Ir(III) product, Ir(H)2 (P(CH2 Ad)i-Pr2 )(𝜅 2 -P(CH2 Ad* )i-Pr2 ) (Figure 7.30). Remarkably, the C—H bond of the Ad group could be reversibly formed upon exposure to a N2 atmosphere. The tris(phosphine) ligand with an apical silyl donor, [(o-(i-Pr2 P)C6 H4 )3 Si]− (abbreviated as SiP3 ), was employed to stabilize the trigonal bipyramidal Ir(I) compound, Ir(N2 )(SiP3 ) [21]. The N–N stretching frequency of 2122 cm−1 is 59 cm−1 less activated than the Co analog, Co(N2 )(SiP3 ). The N2 ligand in Ir(N2 )(SiP3 ) is quite labile, and it can be removed by exposure to vacuum. Although the N—C and N—Si bonds forming reactions of Ir–N2 adducts remain elusive, Ir–aryldiazo complexes are well known. Adapting Collman’s synthesis, Ibers and coworker isolated several [IrCl(N2 Ar)(PPh3 )2 ]+ compounds by adding aryldiazonium salts to the reaction of Vaska’s complex IrCl(CO)(PPh3 )2 with organic azides [124]. The N–N vibrational frequency of [IrCl(N2 Ph)(PPh3 )2 ]+ is 1868 cm−1 . Addition of an exogeneous ligand (L = PPh3 , C≡NEt, CO) generated the five-coordinate complexes, [IrCl(N2 Ph) (PPh3 )2 L]+ , where the N–N stretching frequencies are significantly reduced to 1499–1651 cm−1 and are comparable to diazene (cf. diazene: 1.25 Å, 1583, 1529 cm−1 ). Specifically, the molecular structure of five-coordinate [IrCl(N2 Ph) (PPh2 Me)3 ]+ revealed a N—N bond length of 1.241(11) Å (𝜈N2 = 1644 cm−1 ) and a bent phenyldiazo ligand where the N—N—C bond angle is 118.8(8)∘ [125]. The neutral four-coordinate Ir diazo adduct, trans-IrCl(N2 C5 Cl4 )(PPh3 )2 , where C5 Cl4 is tetrachlorocyclopentadiene, has also been structurally characterized with a N—N bond length of 1.171(8) Å (𝜈N2 = 1858 cm−1 ) and a bent N—N—C bond angle of 141∘ [126, 127]. Six-coordinate Ir diazo complexes are also possible. As an example, IrCl2 (N2 -o-NO2 C6 H4 )(CO)(PPh3 )2 has a bent diazo group and a N—N bond distance of 1.19(4) Å [128]. The structure also confirms the presence of two trans-phosphines and two cis-chlorides [128]. An

7.3 Iridium–Dinitrogen Complexes

intriguing synthesis of Ir diazo complexes was reported: trans-IrCl(N2 )(PPh3 )2 reacts with sulfonated benzotriazole via an unusual ring-opening reaction to provide the aryldiazo trans-IrCl(N2 -o-NSO2 R-C6 H4 )(PPh3 )2 [129]. Diazonium adducts are also known for Cp*Ir platforms, although they are not covered here [130, 131]. 7.3.3 7.3.3.1

Ligands with Exclusively Nitrogen Donors Tris(pyrazoyl)borate (Tp) Ligands

Poveda and coworkers isolated a Ir(III)–N2 complex using the hydrotris(3,5dimethylpyrazolyl)borate ligand (abbreviated as TpMe,Me ). The Ir(III) precursor, IrH(CH=CH2 )(CH2 =CH2 )TpMe,Me , was heated in benzene under a N2 atmosphere to yield Ir(N2 )(C6 H5 )2 TpMe,Me , where two benzene molecules were activated [132, 133]. The mono-iridium–N2 complex exists in equilibrium with the N2 -sandwich species, {Ir(C6 H5 )2 TpMe,Me }(μ-N2 ). Vibrational spectroscopy determined their corresponding N–N stretching frequencies to be 2190 cm−1 (IR) and 2130 cm−1 (Raman), respectively. The N—N bond distance in {Ir(C6 H5 )2 TpMe,Me }(μ-N2 ) is 1.13(3) Å. This compound was also the first X-ray structure of a Ir–N2 complex. Switching to hydrotris(3-p-tolylpyrazolyl)borate ligand (abbreviated as Tptol ), Carmona and coworkers studied the related Ir(III) precursor, IrH(CH=CH2 )(CH2 =CH2 )Tptol . UV irradiation of the precursor in benzene under A N2 atmosphere gave the six-coordinate Ir(III) product, Ir(N2 ) (C6 H5 )Tptol* , where both an intermolecular and an intramolecular arene C—H bond activation have occurred (Figure 7.31). The latter involves one of the tolyl substituents, which becomes cyclometalated at the Ir center. The N2 ligand has a N—N bond length of 1.089(5) Å and a N–N stretching frequency of 2205 cm−1 [134]. Tellers and Bergman extended the Tp platform to support cationic Ir(III)–N2 compounds, such as [IrMe(N2 )(PMe3 )TpMe,Me ]+ [135]. This complex was prepared by treating IrMe(OTf )(PMe3 )TpMe,Me with NaBArF 4 in CH2 Cl2 under N2 .

N5 N3

H

N1 C13

B N N tol

N N N Ir

C31

Ir

C14

N N7

tol

N8

C19

N2 (a)

Me

(b)

Figure 7.31 (a) Ir(N2 )(C6 H5 )Tptol* and (b) molecular structure. Source: Conejero et al. 2011 [134]. Adapted with permission from ACS.

385

386

7 Group 9 Transition Metal–Dinitrogen Complexes

The N–N vibrational band at 2225 cm−1 indicated very weak activation. The methylide complex reacts with benzene to generate [IrPh(N2 )(PMe3 )TpMe,Me ]+ , which has a similar 𝜈N2 value of 2236 cm−1 [136]. Of interest, the isoelectronic complex, [IrMe(N2 )(PMe3 )Cp* ]+ , where Cp* is C5 Me5 , was detected in situ only under high N2 pressure (≥10 atm), and the N2 ligand was confirmed by a 𝜈N2 value of 2207 cm−1 [136]. 7.3.3.2

𝛃-diketiminate Ligands

Employing the β-diketiminate ligand, [CH{C(Me)N(2,6-i-Pr2 C6 H3 )}2 ]− (abbreviated as LMe,dipp ), Chirik and coworkers isolated the Ir(I)–N2 complex, Ir(N2 )(COE)LMe,dipp , where COE is cyclooctene [137]. The N—N bond length of 1.107(4) Å and the 𝜈N2 value of 2126 cm−1 are consistent with a weakly activated N2 ligand. Exposure of Ir(N2 )(COE)LMe,dipp to 4 atm H2 results in the Ir tetrahydride product, Ir(H)4 LMe,dipp , where the COE ligand was hydrogenated and N2 was released. 7.3.4

Pincer Ligands

7.3.4.1

PNP-Type Pincer Ligands

Using [N(2-Pi-Pr2 -4-MeC6 H3 )2 ]− (abbreviated as iPr,Ar PNP), Whited and Grubbs isolated the square-planar Ir(I)–N2 adduct, Ir(N2 )(iPr,Ar PNP) [138]. The complex has a N—N bond length of 1.128(7) Å and has a 𝜈N2 energy of 2067 cm−1 . Although no functionalization of the N2 ligand was reported, Ir(N2 )(PNP) catalyzes the oxidation of t-butyl methyl ether (t-BuOMe) with an organic azide (RN3 ) to produce formimidate, HC=NR(Ot-Bu). In the proposed mechanism (Figure 7.32), the N2 -labile compound, Ir(N2 )(iPr,Ar PNP), reacts with t-BuOMe to undergo two successive C—H bond activations with H2 transfer to a sacrificial olefin, generating the Fischer carbene intermediate, Ir=CHOtBu(iPr,Ar PNP), which was isolated. In the next step, the Ir carbene reacts with organic azide to release formimidate as the product while regenerating Ir(N2 )(iPr,Ar PNP). The PNP platform with an diene-amido donor, [N(CH=CHPt-Bu2 )2 ]− (abbreviated as tBu,ene PNP), was used by Schneider and coworkers to isolate several Ir–N2 complexes: the terminally bound Ir(I)–N2 adduct, Ir(N2 )(tBu,ene PNP),

PiPr2 N Ir

N2

hν (1) NBE MTBE

PiPr2 N

(2) AdN3

PiPr2

Ir PiPr2

OtBu

NAd H

OtBu

Figure 7.32 Ir(N2 )(iPr,Ar PNP) catalyzes the conversion of t-BuOMe (MTBE) and adamantyl azide to generate formimidate. The reaction requires a sacrificial H2 acceptor, norbornene (NBE). Source: Whited and Grubbs [138]. Adapted with permission from ACS.

7.3 Iridium–Dinitrogen Complexes

and a triad of diiridium–N2 -sandwich compounds in different charge states, [{Ir(tBu,ene PNP)}2 (μ-N2 )]0/+/2+ [116, 139]. Mononuclear Ir(N2 )(tBu,ene PNP) is less stable than its dinuclear analog and has a weakly activated N2 ligand as evidenced by a 𝜈N2 energy of 2077 cm−1 . The 𝜈N2 value of the isostructural monorhodium analog, Rh(N2 )(tBu,ene PNP), is 43 cm−1 lower, suggesting that Ir(tBu,ene PNP) is worse at π-backbonding than Rh(tBu,ene PNP) (cf. 2120 cm−1 , Section 7.2.4.2). Also analogous to its Rh counterpart, photolysis of the azide complex, Ir(N3 )(tBu,ene PNP), generates a formal Ir(IV) nitride that undergoes bimolecular coupling to give a mixture of {Ir(tBu,ene PNP)}2 (μ-N2 ) and Ir(N2 )(tBu,ene PNP). The reaction proceeds with a second-order rate law with respect to the Ir nitride species. The kinetic data indicate that the primary reaction product is {Ir(tBu,ene PNP)}2 (μ-N2 ), which is further trapped by additional N2 to generate 2 equiv of the mononuclear Ir(N2 )(tBu,ene PNP) [116]. Chemical oxidation of {Ir(tBu,ene PNP)}2 (μ-N2 ) with [FeCp2 ]PF6 or 2 equiv of AgSbF6 provided the mono- and dicationic analogs, [{Ir(tBu,ene PNP)}2 (μ-N2 )]PF6 , and [{Ir(tBu,ene PNP)}2 (μ-N2 )](SbF6 )2 , respectively. The triad of diiridium–N2 sandwich complexes has been characterized by X-ray crystallography. The compounds are isostructural with a perpendicular disposition of the Ir(tBu,ene PNP) fragments. The N—N bond length (Å) of the bridging N2 ligand is unperturbed across the three different charge states (n): 1.135(4) for n = 0, 1.136(6) for n = +1, and 1.138(6) for n = +2. Hence, the redox changes are likely localized at the Ir center(s), and the Ir oxidation states were assigned as Ir(I)/Ir(II), Ir(I)/Ir(II), and Ir(II)/Ir(II) for n = 0, +1, and +2, respectively. Of note, the mixed-valent Ir(I)/Ir(II) N2 -sandwich complex exhibits an IR-active N–N vibration at 1960 cm−1 . The observation of the vibration is consistent with an asymmetric structure where the mixed-valent Ir(I)/Ir(II) electronic state is strongly localized. In support, close examination of the molecular structure shows two distinct Ir sites with different metal–ligand bond lengths. Of note, only the neutral dirhodium–N2 species is known for this ligand, which attests to the greater redox flexibility of the diiridium system. Considering the significance of N2 scission to nitride in potential N2 functionalization schemes, and its reverse process, Schneider and coworkers further investigated the bimolecular coupling reactivity of two charge-distinct Ir nitrides, [Ir(N)(tBu,ene PNP)]0/+ [139]. The neutral nitride undergoes fast bimolecular coupling to form neutral {Ir(tBu,ene PNP)}2 (μ-N2 ) (Figure 7.33). By contrast, the cationic nitride does not react to yield the dicationic [{Ir(tBu,ene PNP)}2 (μ-N2 )]2+ . The lack of reactivity was attributed in part to the charge repulsion and to the closed-shell nature of the cationic Ir(V) nitride. On the other hand, the neutral Ir nitride is proposed to have a substantial nitridyl radical character based on DFT calculations, allowing for a diradical coupling of two reactive Ir =̇ N groups with minimum reorganization. Mixing the neutral and cationic nitride species in a 1 : 1 ratio generates the monocationic [{Ir(tBu,ene PNP)}2 (μ-N2 )]+ compound even more rapidly. This bimolecular coupling is predicted to be essentially barrierless and is the most facile among the three coupling reactions in the study. The low barrier is attributed to the excellent energy and symmetry match between the donor (SOMO of neutral Ir =̇ N) and acceptor (LUMO of cationic Ir(V) nitride) orbitals (SOMO = singly occupied molecular orbital; LUMO = lowest unoccupied MO).

387

388

7 Group 9 Transition Metal–Dinitrogen Complexes tBu 2 t

2

N

P Bu2 Ir N

kA Reaction A

P tBu 2

P

t

N

P Bu2 Ir N N

P tBu 2

N

Ir P

tBu

2

k B > kA t

Pt N

Ir

Bu2 N

PtBu +

N

P t Bu2

Ir

2

+

N

kB Reaction B

P t Bu2

Pt N

Ir

Bu2 N N

P t Bu2

Ir t

2

N

Ir

P t Bu2

N

2

+

kC Reaction C

P

PtBu2 N

Ir

P tBu 2

N N

N

P Bu2

tBu

PtBu

+

Bu2 P

2+

2

Ir

N

P Bu2

t

Figure 7.33 Bimolecular coupling reactions of [Ir(N)(tBu,ene PNP)]0/+ . Source: Abbenseth et al. 2016 [139]. Adapted with permission from https://creativecommons.org/licenses/by/3.0/.

7.3.4.2

PCP- and PSiP-Type Pincer Ligands

The Ir coordination chemistry of the PCP ligand featuring a central arene donor, [2,6-(CH2 Pt-Bu2 )2 C6 H3 ]− (abbreviated as PAr CP), has been investigated by Jensen and coworkers [140–142]. During the investigation of the Ir(PAr CP) catalyst for alkane dehydrogenation, Ir–N2 complexes were isolated and characterized. The diiridium–N2 -sandwich complex, {Ir(PAr CP)}2 (μ-N2 ), was synthesized from the reaction of Ir dihydride, Ir(H)2 (PAr CP), and t-butyl ethylene under a N2 atmosphere [140]. The molecular structure confirmed the bridging nature of the N2 ligand and the perpendicular orientation of the two Ir(PAr CP) fragments. The N2 ligand is labile and is readily displaced by CO2 or CO [141]. A re-examination of this system revealed that the monoiridium complex, Ir(N2 )(PAr CP), is also present in equilibrium with {Ir(PAr CP)}2 (μ-N2 ) [142]. The monoiridium species, Ir(N2 )(PAr CP), displays an IR-active band at 2076 cm−1 and has a N—N bond length of 1.107(3) Å, consistent with weak activation [142]. The diiridium–N2 -bridged complex, {Ir(PAr CP)}2 (μ-N2 ), was characterized by a Raman-active N–N stretching frequency of 1979 cm−1 and a N—N bond length of 1.134(2) Å, showing a marked increase in N2 activation upon binding of two Ir centers [142]. Comparing the isostructural Ir and Rh analogs, M(N2 )(PAr CP), the Ir species has a lower 𝜈N2 energy by 57 cm−1 (Section 7.2.4.1, cf. 2133 cm−1 ). Notably, the reverse trend was observed for the Rh/Ir(tBu,ene PNP) systems (vide supra), which suggests the importance of the central donor in tuning the Rh/Ir d-orbitals.

7.3 Iridium–Dinitrogen Complexes

In exploring the Ir coordination chemistry of the phosphinite-based POCOP platform, Brookhart and coworkers isolated Ir–N2 complexes and investigated their reactivity. With [2,6-(OPt-Bu2 )2 -4-ArF (C6 H3 )]− (where ArF = 3,5(CF3 )2 C6 H3 , abbreviated as tBu POCOP), the diiridium–N2 -bridged compound, {Ir(tBu POCOP)}2 (μ-N2 ), was isolated from a hydrodechlorination reaction of IrHCl(tBu POCOP) with NaOt-Bu [143]. The molecular structure further revealed a N—N bond length of 1.119(6) Å. A N–N stretching frequency of 2118 cm−1 was also reported, which may correspond to the terminally bound N2 species, Ir(N2 )(tBu POCOP). The Ir(tBu POCOP) system catalyzes transfer dehydrogenation reactions, akin to the Ir(PAr CP) system. The {Ir(tBu POCOP)}2 (μ-N2 ) catalyst showed no inhibition from N2 binding, whereas catalysis is hindered by N2 binding in the Ir(PAr CP) system. Switching to a POCOP variant with diaryl phosphinite donors, [2,6-{OP(2′ ,4′ ,6′ -(CF3 )3 C6 H2 )2 }2 (C6 H3 )]− (abbreviated as Ar POCOP), the end-on N2 compound, Ir(N2 )(Ar POCOP), was crystallized, revealing a N—N bond length of 1.106(5) Å. The lability of the N2 ligand was exploited for single-crystal to single-crystal transformations with CO, NH3 , C2 H4 , H2 , and O2 [144]. In times ranging from one minute to one day, these small molecules all successfully displaced the N2 ligand while maintaining the integrity of the single crystal. Even more intriguing is that the Ir(N2 )(Ar POCOP) catalyzed the alkene hydrogenation as a single crystal, where the large preference (25 : 1) for ethylene hydrogenation over propylene hydrogenation is consistent with the catalysis occurring within the crystal. Wendt and coworkers have also studied PCP-type pincer ligands featuring a central cyclohexyl-based sp3 -C donor, [2,6-(CH2 Pt-Bu2 )2 C6 H11 ]− (abbreviated as PCy CP) [145]. The mononuclear N2 -end-on complex, Ir(N2 )(PCy CP), was characterized by IR spectroscopy (𝜈N2 = 2065 cm−1 ) and X-ray crystallography (N—N bond length of 1.064(7) Å). Shimada and coworkers used a unique bis(phosphino)silyl pincer ligand to synthesize rare examples of Ir(III)–N2 pincer complexes [146]. The PSiP-type pincer ligand, [MeSi(2-PR2 -C6 H4 )2 ]− , where R = i-Pr or Cy (abbreviated as iPr,Ar PSiP and Cy,Ar PSiP, respectively), is coordinatively flexible as it can coordinate in both mer and fac geometries. The Ir(III) precursor, IrH(Cl)(mer-R,Ar PSiP), reacts with [NMe4 ]BH4 to generate Ir(H)2 (N2 )(R,Ar PSiP) as a mixture of the mer and fac isomers (Figure 7.34), where the mer isomer is the dominant product (85–90% yield). In the mer isomer, the N2 ligand is positioned trans to a hydride ligand, and the N–N stretching frequency for both R = i-Pr and Cy phosphine substituents is identical at 2091 cm−1 . In the minor fac isomer, the N2 ligand is positioned trans to the silyl donor. Addition of PMe3 to Ir(H)2 (N2 )(mer-Cy,Ar PSiP) gave a mixture of Ir(H)2 (PMe3 )(mer-Cy,Ar PSiP) and Ir(N2 )(PMe3 )(fac-Cy,Ar PSiP). The latter is a trigonal bipyramidal Ir(I)–N2 complex, where the N2 ligand is positioned trans to the silyl donor, akin to Ir(N2 )(SiP3 ) (Section 7.3.2). Despite the similar coordination environment of Ir(N2 )(PMe3 )(fac-Cy,Ar PSiP) and Ir(N2 )(SiP3 ), the former shows a greater extent of N2 activation. The N–N stretching frequency of 1926 cm−1 for Ir(N2 )(PMe3 )(fac-Cy,Ar PSiP) is the lowest reported for a terminally bound N2 ligand at Ir and is nearly 200 cm−1 lower than that of Ir(N2 )(SiP3 ) (cf. 2122 cm−1 ).

389

390

7 Group 9 Transition Metal–Dinitrogen Complexes

P2

P1 H57 Si1

H56

Ir1 H56

Si1

N2

Ir1 H57

N1 P2

N2 N1

P1

(a)

(b)

Figure 7.34 Molecular structures of the (a) mer and (b) fac isomers of Ir(H)2 (N2 )(Cy,Ar PSiP). Source: Fang et al. 2011 [146]. Reproduced with permission from John Wiley & Sons. Figure 7.35 Molecular structure of cationic [trans-Ir(N2 )2 (IMes)2 ]+ . Source: Tang et al. 2010 [147]. Reproduced with permission from ACS.

N(4) N(5)

7.3.5

Ir(1) N(2)

N(3)

N-heterocyclic Carbene Ligands

NHC-supported square-planar Ir(I)–N2 complexes are also known. Using the IMes ligand, where IMes is 1,3-dimesitylimidazol-2-ylidene, Aldridge and coworkers isolated the first example of a bis(dinitrogen) complex of a group 9 metal [147]. The complex, [Ir(N2 )2 (IMes)2 ]BArF 4 , has a square-planar geometry and two trans-disposed N2 ligands (Figure 7.35). A coordinating solvent such as THF can displace one of the two N2 ligands to generate [Ir(N2 )(THF) (IMes)2 ]BArF 4 . The N–N stretching frequencies of [Ir(N2 )2 (IMes)2 ]+ and [Ir(N2 )(THF)(IMes)2 ]+ are quite similar at 1987 and 1985 cm−1 , respectively. The latter species has a N—N bond length of 1.109(8) Å. Another NHC-supported Ir(I)–N2 compound is Ir(N2 )Cl(COE)(IMes), where the N2 ligand is positioned trans to Cl and has a slightly higher 𝜈N2 energy of 1993 cm−1 [148].

7.3 Iridium–Dinitrogen Complexes

C9 C10

C8 C15

C11 C18

C17

N5

C14

C7 C6

C5 C4

C19 N4

C20

C12

N1

C16 C21

Ir2

C13

C3 C1

Ir1

C22

C2

C45 C31 N2 C32

C23

C44 C43

C38 N3

C29 C24 C30

C33 C28

C37 C40

C42 C41

C34 C39

C36 C35

C25 C27 C26

Figure 7.36 Molecular structure of a Ir(II)/Ir(II) metal–metal-bonded–N2 complex. Source: Yang et al. 2011 [149]. Reproduced with permission from Royal Society of Chemistry.

7.3.6

Miscellaneous

An Ir(II)/Ir(II) metal–metal-bonded complex with a single end-on N2 ligand was serendipitously isolated from a reaction of {Ir(𝜅 2 -N,C-2-phenylquinolyl)2 (μ-Cl)}2 with NaOMe (Figure 7.36) [149]. The diiridium complex is asymmetric, and the N2 ligand is terminally bound to a five-coordinate Ir center that is ligated to two 2-phenylquinoline ligands via 𝜅 2 -N,C and 𝜅 1 -C linkages. The N2 ligand is positioned trans to an arene sp2 -C donor. The N–N vibration of 2014 cm−1 and the N—N bond distance of 1.106(7) Å indicate weak activation. 7.3.7

Summary of Iridium–Dinitrogen Complexes

Iridium–dinitrogen complexes, although lacking in number, are a fairly diverse group of complexes. An analysis of the 26 iridium–N2 complexes with reported 𝜈N2 values was conducted. Excluding the diazo adducts, the 𝜈N2 values of Ir–N2 compounds range from 1926 to 2236 cm−1 , with an average value of 2096 cm−1 and a standard deviation of 88 cm−1 . Figure 7.37 shows the frequency distribution of the Ir–N2 complexes as a function of 𝜈N2 , where nearly half of these compounds have 𝜈N2 values between 2050 and 2150 cm−1 . The range of 𝜈N2 energies is larger for Ir than for Rh, but relatively limited compared to that for Co. As the case for Rh, Ir(I) square-planar complexes are common. However, Ir demonstrates much greater redox flexibility as evidenced by several Ir(III)– and Ir(II)–N2 complexes.

391

7 Group 9 Transition Metal–Dinitrogen Complexes

Ir N2 complexes 6

4

2200

2150

2100

2050

2000

0

1950

2

1900

Number

392

N–N bond stretching frequencies (cm–1)

Figure 7.37 Distribution of iridium–N2 complexes as a function of 𝜈N2 (in units of 50 cm−1 ).

The lowest N–N stretching frequency of 1926 cm−1 corresponds to the Ir(I) trigonal bipyramidal complex, Ir(N2 )(PMe3 )(fac-Cy,Ar PSiP). The diiridium– N2 -sandwich complexes are also characterized by low N–N stretching frequencies, e.g. the mixed-valent Ir(I)/Ir(II) [{Ir(tBu,ene PNP)}2 (μ-N2 )]PF6 and Ir(I)/Ir(I) {Ir(PAr CP)}2 (μ-N2 ). Of note, these diiridium compounds have known monoiridium–N2 analogs, and the coordination of a second Ir(PNP) fragment to the N2 ligand is estimated to provide ∼100 cm−1 lowering, or enhanced N2 activation. The next lowest 𝜈N2 energies correspond to Ir(I)–NHC complexes, [Ir(N2 )2 (IMes)2 ]+ and [Ir(N2 )(THF)(IMes)2 ]+ , despite their cationic charge. At the opposite end, showing the least N2 activation, are the Ir(III)–Tp complexes, e.g. [IrPh(N2 )(PMe3 )TpMe,Me ]+ and [IrMe(N2 )(PMe3 )TpMe,Me ]+ . The former has the highest N–N stretching frequency of 2236 cm−1 in this set, which is quite close to that of free N2 (cf. 2359 cm−1 ).

7.4 Group 9 Catalysts for N2 Functionalization 7.4.1

Cobalt-Based Catalysts

Heterogeneous iron catalysts promoted by Al2 O3 , K2 O, and/or CaO are widely used in the Haber–Bosch process. As an alternative catalyst, pure Co metal suffers from low N2 chemisorption and exhibits poor activity [150]. However, ammonia production is greatly enhanced when the Co-active phase is mixed with promoters, such as Ba and/or La [151, 152]. In molecular N2 functionalization catalysis, cobalt was an unknown metal center before 2015 [153]. In that banner year, several Co compounds were reported for the reductive silylation of N2 to N(SiMe3 )3 [27, 154]. The discovery of Co-based catalysts for converting N2 to NH3 would follow closely in 2016 [60]. The recent momentum suggests that Co may play a prominent role in N2 catalysis in the future.

7.4 Group 9 Catalysts for N2 Functionalization

7.4.1.1

Dinitrogen Silylation

In 1972, Shiina screened various metal salts for mediating reductive N2 silylation. The catalysis tests were performed using Me3 SiCl, Li metal, and 2 mol% catalyst (defined with respect to Me3 SiCl) under 1 atm N2 in THF [155]. The best performing metal salt, CrCl3 , yielded 5.4 equiv of N(SiMe3 )3 . As a catalyst, CoCl2 produced a substoichiometric amount of silylamine (1.2 equiv). From this modest beginning, Mo-based compounds were developed by several research groups that demonstrated significant activity in reductive N2 silylation [156–159]. Up to 2015, the Mo complex, trans-Mo(N2 )2 (depf )2 , where depf is 1,10-bis(diethylphosphino)ferrocene, was the only highly active catalyst, producing 150 equiv of N(SiMe3 )3 in a single catalytic run [157]. The catalysis used a low catalyst loading (0.025 mol%), Na as the reductant, and 1 atm N2 at ambient temperature [157]. A further increase in N(SiMe3 )3 production to 226 equiv was observed after replenishing the catalytic reaction with a second batch of Me3 SiCl and Na. Later catalyst screening tests would employ similar conditions: a low catalyst loading (