University of Groningen Formazanate As Redox-Active, Structurally

University of Groningen

Formazanate as redox-active, structurally versatile ligand platform Chang, Mu-Chieh

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record

Publication date: 2016

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA): Chang, M-C. (2016). Formazanate as redox-active, structurally versatile ligand platform: Zinc and boron chemistry. University of Groningen.

Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license. More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne- amendment.

Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

Download date: 07-10-2021

Formazanate as a Redox-Active,

Structurally Versatile Ligand Platform

Zinc and Boron Chemistry

Mu-Chieh Chang

The research described in this thesis was carried out at the Stratingh Institute for Chemistry, University of Groningen, The Netherlands.

The work was financially supported by the Netherlands Organization for Scientific Research (NWO).

Cover designed by Mu-Chieh Chang Printed by Ipskamp Drukkers, The Netherlands

ISBN: 978-90-367-8573-0 eISBN: 978-90-367-8572-3

Formazanate as a Redox-Active, Structurally Versatile Ligand Platform

Zinc and Boron Chemistry

PhD thesis

to obtain the degree of PhD at the University of Groningen on the authority of the Rector Magnificus Prof. E. Sterken and in accordance with the decision by the College of Deans.

This thesis will be defended in public on

Monday 18 January 2016 at 14.30 hours

Mu-Chieh Chang

born on 23 March 1985 in Tainan, Taiwan

Supervisor Prof. W.R. Browne

Co-supervisor Dr. E. Otten

Assessment Committee Prof. J.G. Roelfes Prof. J.G. de Vries Prof. F. Meyer

Contents

Chapter 1 Introduction ………………………………………………………………………… 1 1.1 General Introduction ………………………………………………………………… 2 1.2 Redox-Active Ligands in Nature ……………………………………………………. 3 1.3 Redox-Active Ligands in Laboratories ……………………………………………… 4 1.3.1 Dioxygen Activation by ZrIV Complex Bearing Diimine Ligands ……………….. 4 1.3.2 Catalytic Reactions Based on Redox-Active Ligands ……………………………. 5 1.3.2.1 Catalytic Cyclization of Enynes and Dienes …………………………………... 5 1.3.2.2 Catalytic Nitrene Transfer Based on ZrIV Complex ……………………………. 6 1.4 Formazanate Ligands: Nitrogen Rich Analogues of -Diketiminate Ligands ……… 7 1.4.1 Redox-Active Nature of Formazanate Ligands …………………………………... 8 1.4.2 Redox-Active Nature of -Diketiminate Ligands ………………………………... 9 1.4.3 Metal Complexes Bearing Fromazanate Ligands ………………………………… 11 1.5 Overview of Thesis ………………………………………………………………….. 11 1.6 References …………………………………………………………………………… 13

Chapter 2 Formazan Synthesis ………………………………………………………………... 15 2.1 Introduction ………………………………………………………………………….. 16 2.1.1 Method A1: Formazan Synthesis from Diazonium Salt and Arylhydrazone ………... 16 2.1.2 Method A2: Formazan Synthesis from Diazonium Salt and Active Methylene Group .. 18 2.1.3 Structure of Formazans and Formazanate Anions …………………………………... 18 2.2 Formazan Synthesis …………………………………………………………………. 19 2.2.1 Mono-Formazan Ligand Synthesis ………………………………………………….. 20 2.2.2 Phenylene-Linked di-Formazan Ligand Synthesis ………………………………….. 25 2.3 Conclusion …………………………………………………………………………... 26 2.4 Experimental Section ………………………………………………………………... 26 2.5 References …………………………………………………………………………… 36

Chapter 3 (Formazanate)Zinc Complexes ……………………………………………………. 37 3.1 Introduction ………………………………………………………………………….. 38 3.2 (Formazanate)Zinc Methyl complexes ……………………………………………… 39 3.2.1 Zinc Methyl Complex with Phenylene-Linked Diformazanate Ligand …………….. 40 3.2.2 Quantitative Description of Trigonal Pyramidal Zinc Center ……………………….. 42 3.3 Bis(Formazanate)Zinc Complexes ………………………………………………….. 44 3.3.1 Synthesis and Coordination Chemistry of Bis(Formazanate)Zinc Complexes ……... 44 3.3.2 UV-Vis Spectroscopy of Bis(Formazanate)Zinc Complexes ……………………….. 51 3.3.3 Cyclic Voltammetry of Bis(Formazanate)Zinc Complexes …………………………. 52 3.4 Chemical Reduction of Bis(Formazanate)Zinc complexes …………………………. 56 3.4.1 Synthesis and Characterization of 1-Electron Reduction Products …………………. 56 3.4.2 Synthesis and Characterization of 2-Electron Reduction Products …………………. 58 3.4.3 UV-Vis Spectroscopy of Reduced Products ………………………………………… 59 3.4.4 DFT Calculations ……………………………………………………………………. 60 3.4.4.1 DFT Calculations of 1-Electron Reduction Compound (5–) ……………………….. 60 3.4.4.2 DFT Calculations of 2-Electron Reduction Compound (5–2) ……………………….. 61 3.5 Conclusion …………………………………………………………………………... 62 3.6 Experimental Section ………………………………………………………………... 63 3.7 References …………………………………………………………………………… 72

Chapter 4 (Formazanate)Boron Difluoride Complexes Formation via Zinc to Boron Transmetallation …………………………………………………... 75 4.1 Introduction ………………………………………………………………………….. 76 4.2 Formazanate Transfer from Zinc to Boron ………………………………………….. 77 4.2.1 Formation of (Formazanate)Boron Difluoride via Transmetallation ………………... 77 4.2.2 Isolation of a Six-Coordinated Zinc Complex as a Key Intermediate ………………. 78 4.2.3 Proposed Mechanism of Transmetallation Reaction ………………………………... 81

4.2.4 Reaction of Heteroleptic Complex 5aj with BF3·Et2O ……………………………… 82 4.2.5 1,2,3-Triazole Formation ……………………………………………………………. 85 4.3 Conclusion …………………………………………………………………………... 86 4.4 Experimental Section ………………………………………………………………... 87 4.5 References …………………………………………………………………………… 89

Chapter 5 (Formazanate)Boron Complexes ………………………………………………….. 91 5.1 Introduction ………………………………………………………………………….. 92 5.2 Synthesis of (Formazanate)Boron Complexes ……………………………………… 94 5.2.1 (Formazanate)Boron Difluoride …………………………………………………….. 94 5.2.2 (Formazanate)Boron Diphenyl and (Formazanate)Boron Dihydride ……………….. 96 5.3 Characterization of Formazanate Boron Complexes ………………………………... 97 5.3.1 X-ray Crystallographic Analysis …………………………………………………….. 97 5.3.2 Absorption and Emission Spectroscopy …………………………………………….. 101 5.3.3 Redox Chemistry ……………………………………………………………………. 103 5.4 Reduced Products: mono-Formazanate Ligands ……………………………………. 106 5.4.1 Synthesis and Crystal Structures ……………………………………………………. 106 5.4.2 Absorption and Emission Spectroscopy …………………………………………….. 107 5.4.3 EPR Spectra and DFT Calculations …………………………………………………. 108 5.5 Reduced products: Phenylene-Linked Diformazanate Ligands …………………….. 110 5.5.1 Synthesis and Crystal Structures ……………………………………………………. 110 5.5.2 DFT calculations and VT-EPR Studies ……………………………………………… 112 5.6 Conclusion …………………………………………………………………………... 114 5.7 Experimental Section ………………………………………………………………... 114 5.8 References …………………………………………………………………………… 124

Chapter 6 Reduction of (Formazanate)Boron Difluoride Provides Evidence for an N-Heterocyclic B(I) Carbenoid Intermediate ……………………………………. 127 6.1 Introduction ………………………………………………………………………….. 128 6.2 Synthesis and Characterization ……………………………………………………… 129 6.2.1 Synthesis ……………………………………………………………………………………………………… 129 6.2.2 X-ray Crystallography ………………………………………………………………. 132 6.2.3 NMR Spectroscopy and UV-Vis Analysis …………………………………………... 134 6.2.4 Reduction Chemistry ……………………………………………………………………………………. 136 6.3 DFT Calculations …………………………………………………………………………………………... 140 6.4 Thermal Stability ……………………………………………………………………. 143 6.5 Trapping (Formazanate)Boron(I) by Acetylene …………………………………….. 146 6.6 Chemical Oxidation …………………………………………………………………. 148 6.7 Conclusion …………………………………………………………………………... 150 6.8 Experimental Section ………………………………………………………………... 150 6.9 References …………………………………………………………………………… 157

Chapter 7 Intramolecular Hydride Transfer Reactions in (Formazanate)Boron Dihydride Complexes ……………………………………………………………… 159 7.1 Introduction ………………………………………………………………………….. 160 7.2 Synthesis of (Formazanate)Boron Dihydride Complexes …………………………... 161 7.3 Thermally Induced Intramolecular Hydride Transfer ……………………………….. 161

7.3.1 [PhNNC(p-tolyl)NNPh]BH2 (10a) ………………………………………………….. 161

7.3.2 [PhNNC(p-tolyl)NNMes]BH2 (10c) ………………………………………………… 166

7.3.3 [C6F5NNC(p-tolyl)NNMes]BH2 (10f) ………………………………………………. 169 7.4 Kinetic Study and Proposed Mechanism ……………………………………………. 170

7.4.1 Kinetic Study of the Thermolysis of [PhNNC(p-tolyl)NNPh]BH2 (10a) …………… 170 7.4.2 Proposed Mechanism and DFT Calculations ……………………………………….. 171 7.5 Discussion …………………………………………………………………………… 173 7.6 Conclusion …………………………………………………………………………... 175 7.7 Experimental Section ………………………………………………………………... 175 7.8 References …………………………………………………………………………… 181

English Summary ……………………………………………………………………. 183 Nederlandse Samenvatting ………………………………………………………….. 189 Prospective ………………………………………………………………………….. 195 Acknowledgements …………………………………………………………………. 199 誌謝 …………………………………………………………………………………. 202

Chapter 1

Introduction Chapter 1

Chapter1 Introduction 1.1 General Introduction In the field of homogeneous catalysis, the bond breaking and the bond forming steps are often two-electron processes. In other words, the catalyst needs to donate or accept two electrons to or from the substrates in the catalytic cycle. For example, palladium-catalyzed coupling reactions have become very popular and useful catalytic reactions both in industry and academia, and the Nobel Prize in Chemistry was awarded in 2010 to Heck, Negishi, and Suzuki, who pioneered the development of these reactions.1 The key elementary steps of these palladium-catalyzed coupling reactions involve oxidative addition and reductive elimination (Scheme 1.1).2 In the catalytic cycle, the substrates are activated by oxidative addition in which the Pd center donates two electrons to the substrate to form new chemical bonds resulting in an increase in the oxidation state of the Pd center by 2. Subsequently, a second substrate is introduced by transmetallation or ligand exchange. The last step is product release and catalyst regeneration by reductive elimination, which results in a decrease in the oxidation state of the Pd center by 2.

R1 R 0 LnPd Reductive elimination RXOxidative R addition L PdII n R1 R L PdII n X MX

1 Transmetallation R M Scheme 1.1 The general mechanism of palladium-catalyzed coupling reactions In order to support the two-electron redox processes, most of the catalysts that are used today are based on precious metals.3 This is because precious metals, in general, have relatively stable and easily accessible (n) and (n+2) oxidation states. However, precious metals are usually scarce, expensive and toxic; most importantly, the natural abundance of precious metals is limited. All of these factors make the use of precious metals in catalysis increasingly expensive and this motivates chemists to develop more sustainable (catalytic) methods for organic synthesis and energy applications, in which the precious metal component of catalysts is replaced by more abundant elements (base metal or main group elements).

Introduction

A major challenge of using base metal catalysts is that the base metal is usually suitable for one-electron redox processes due to its relatively stable (n) and (n+1) oxidation states. In other words, the challenge of replacing precious metals by systems based on base metals is how to impart two-electron processes to these base metal complexes. A possible solution to meet this challenge is utilizing redox-active ligands to develop selective two-electron redox reactions using base metal catalysts.4 In this scenario, the redox-active organic ligand framework can actively participate in the redox changes required during chemical transformations by donating or accepting electrons to or from substrates. Utilizing this special feature, metal centers coordinated by redox-active ligands avoid reaching unstable oxidation states during chemical redox transformations. Thus, such complexes can perform catalytic redox reactions that are not possible with conventional ligands.

1.2 Redox-Active Ligands in Nature

Nature is definitely the pioneer of using redox-active ligands in chemical transformations or in catalysis.5 One of the famous examples from Nature is the active site of cytochrome P450 (Chart 1.1), which have been identified in all domains of life6 and is a catalyst for a monooxygenase reaction. The active site of cytochrome P450 comprises a heme cofactor constituted by a Fe+2 center and a porphyrin ligand.7 In 2010, a highly reactive intermediate, which is usually referred to as P450 Compound I (P450-I), of the catalytic cycle was isolated by Green and co-workers.8 The P450-I is an iron(IV)-oxo species bearing a singly oxidized porphyrin radical ligand and capable of conversion of hydrocarbons to alcohols. The P450-I promoted hydrocarbons to alcohols conversion is a 2-electron process, in which the redox equivalents are provided by the iron(IV) center and singly oxidized porphyrin radical ligand resulting in a formation of an iron(III) center and a porphyrin ligand.

Chart 1.1 Active site of Cytochrome P450 (left) and compound I (right)

Chapter 1

Another example of the redox-active ligand from Nature is an intradiol oxidative cleavage of catechols by catechol dioxygenase (Scheme 1.2).9 The first step of the reaction is the formation of a catecholate Fe(III, HS) complex from catechol and high-spin Fe(III) enzyme. Subsequently, the o-semiquinone Fe(II, HS) complex is generated via intramolecular electron transfer from the catecholate to the iron center. The o-semiquinone Fe(II, HS) complex then reacts with dioxygen generating muconic acid as the product.

Scheme 1.2 Intradiol cleavage of catechols by catechol dioxygenase 1.3 Redox-Active Ligands in Laboratories

Taking inspiration from the systems in Nature mentioned above, several redox-active ligands have been developed. Classical examples are given by oxygen-based catecholate type ligands10, or its nitrogen- or sulfur-analogue11, and the porphyrin type ligands (Chart 1.2).12 Besides the bio-inspired redox-active ligand systems, several artificial redox-active ligand systems have been reported. In the following paragraphs, some progress in the design and application of redox-active ligands will be presented.

Chart 1.2 General structure of catecholate type ligands (left) and porphyrin type ligands (right)

1.3.1 Dioxygen Activation by ZrIV Complex Bearing Diimine Ligands

Molecular oxygen is a very attractive oxidant for chemists due to its ready availability (from the air). In 2007, the multi-electron activation of dioxygen by a ZrIV complex, which was coordinated by two diamido ligands, leading to a stable ZrIV bisperoxo bis(diimine) complex 13 IV was reported by Abu-Omar and co-workers (Scheme 1.3). The (diamido)2Zr complex was synthesized from 2 equivalents of -diimine ligand, magnesium metal, and ZrCl4. The

Introduction

IV spectroscopic data and the crystal data shown that the isolated product is a (diamido)2Zr complex, in which each diamido ligand stores two electrons that come from the magnesium metal. The key features of this dioxygen activation reaction are that the electrons used to active the dioxygen were stored in the redox-active diamido ligands first, and the oxidation state of the ZrIV center did not change during the reaction with dioxygen. The reaction proves an important concept that redox equivalents can be stored in the redox-active ligand first and subsequently used for a chemical transformation (in this case, oxygen reduction). Importantly, the metal center is a binding site of substrates and does not change its oxidation state.

IV Scheme 1.3 O2 activation of (diamido)2Zr complex 1.3.2 Catalytic Reactions Based on Redox-Active Ligands

In a catalytic reaction, the catalyst must be regenerated at certain step(s) to close the catalytic cycle and to start the next turnover. The regeneration of the catalyst is one of the greatest challenges of designing catalytic reactions based on redox-active ligands. In the field of precious metal catalysts, the catalytic cycle is closed by reductive elimination resulting in the two-electron reduction of the metal center. In the field of redox-active ligand, the ligand must be reduced after the reductive elimination step to close the catalytic cycle. In the next two sections, two catalytic reactions are presented to illustrate the concept that base-metal complexes containing an artificial redox-active ligand are capable of catalyzing a multi- electron synthetic transformation.

1.3.2.1 Catalytic Cyclization of Enynes and Dienes

Chirik and co-workers reported several examples of iron catalyzed hydrogenative cyclization of enynes (Scheme 1.4) and diynes.10,14 These are beautiful examples showing the concept that base metal catalysts can catalyze multi-electron processes by the help of redox-active ligands. The catalyst they used is a Fe(PDI)(N2) complex (PDI: pyridine(diimine)), which has a FeII center coordinated by a triplet diradical [PDI]2- ligand and a dinitrogen molecule. In the presence of enynes, the dinitrogen molecule is replaced resulting in metallacycle formation and construction of a C-C bond. The enyne cyclization event forms an S = 1 iron compound

Chapter 1 with formal one-electron oxidation events occurring both at the pyridine(diimine) ligand ([PDI]2-→[PDI]1-) and iron center (FeII→FeIII). Upon hydrogenation, an iron(III) alkyl hydride complex is formed. The last step to close the catalytic cycle is the reductive elimination of the product and the regeneration of active catalyst FeII(PDI)2- followed by the coordination of new substrate. Even though this is a FeII/FeIII catalytic cycle, the iron complex still can catalyze the 2-electron process by the help of redox-active pyridine(diimine) ligand.

Scheme 1.4 Catalytic Cycle for the Fe-catalyzed Hydrogenative Cyclization of Enynes 1.3.2.2 Catalytic Nitrene Transfer Based on ZrIV Complex

The nitrene transfer reaction catalyzed by the ZrIV complex bearing the redox-active ligand bis(2-isopropylamido-4-methoxyphenyl)amide ([NNNcat]3-) was reported by Heyduk and co- workers in 2011 (Scheme 1.5).15 The [NNNcat]3- ligand is a nitrogen analogue of the catecholate type tridentate. Like the catecholate ligand, the [NNNcat]3- ligand can be oxidized by two electrons to its quinonate form ([NNNq]-). The active catalyst of the nitrene transfer was generated by ligand dissociation to open the binding site of the metal center. The second step is the transfer of nitrene from the organic azide to the zirconium(IV) center. In order to compensate the formation of the ZrIV-N multiple-bond, the [NNNcat]3- ligand is oxidized by two electrons to the [NNNq]- form. The next two steps are migratory insertion and reductive elimination leading to the formation of the carbodiimide C=N bond and the reduction of the [NNNq]- ligand back to the [NNNcat]3-. In this example, the oxidation state of ZrIV center does

Introduction not change in the catalytic cycle, and the redox equivalents of the catalytic cycle are provide by the [NNNcat]3- ligand.

Scheme 1.5 Heyduk’s ZrIV system of catalytic nitrene transfer 1.4 Formazanate Ligands: Nitrogen Rich Analogues of -Diketiminate Ligands

Even though the examples mentioned above show great potential of utilizing redox-active ligands in the field of small molecules activation and multi-electron (catalytic) processes, the diversity of redox-active ligands is still very limited. This is mainly due to a redox-inert nature of most organic compounds. In literature, reported examples of bidentate ligands having redox-active property are also very limited. In addition to the catecholate type ligands (Chart 1.2) mentioned above, other reported redox-active bidentate ligands are diimine (Scheme 1.3) or pyridine-imine type ligands.16 Therefore, there is a clear need for developing new types of redox-active bidentate ligands. From our point of view, formazanate ligands,

Chapter 1 which are deprotonated formazans, are good candidates for a new redox-active ligand platform due to the structural similarity between formazanate ligands and -diketiminate ligands, and an easily accessible redox-active nature of formazanate ligands.

Formazans, organic compounds that contain the N=N-C=N-N fragment, have a long history as dye molecules.17 In modern chemical research, formazans not only show application in the field of biochemical research18 but also are precursors towards heterocycles such as verdazyl radicals (Scheme 1.6).19 Verdazyls are an unusual class of organic compounds because they are radicals that have high thermodynamic stability without having to rely on bulky substituents. The stability of verdazyl radicals is due to their low-energy SOMO that is a * orbital that is delocalized over four nitrogen atoms (Chart 1.3). The verdazyl radicals are good building blocks for constructing multi-radical systems, such as di-, tri- or tetra-radicals (Chart 1.3).20 In addition, the verdazyl radicals are also used as ligand platforms for metal complexes synthesis due to their nitrogen-rich structure (Chart 1.3).21

Scheme 1.6 Verdazyl radical synthesis

Chart 1.3 SOMO of verdazyl radical (left), examples of di-radicals based on the verdazyl unit (middle), and metal complex bearing verdazyl radical ligands (right)

1.4.1 Redox-Active Nature of Formazanate Ligands The redox-active nature of formazanate ligands was first reported by Hicks and co-workers in 22 2007. They prepared a (formazanate)boron diacetate (LB(OAc)2) complex (Scheme 1.7), which shows one quasi-reversible22 and one irreversible reduction23 in cyclic voltammetry experiments. The singly reduced product ([LB(OAc)2][Cp2Co]) was synthesized by reacting

Introduction

LB(OAc)2 with Cp2Co resulting in a green powder. The ligand-based reduction of

[LB(OAc)2][Cp2Co] was confirmed by EPR and UV-Vis spectra. Later on the same group reported that a heteroleptic Pd complex (Scheme 1.7), which is coordinated by a formazanate and an acetylacetonate ligand, also shows a formazanate-based reduction, albeit irreversibly.24 These examples not only show the redox-active nature of formazanate ligands but also indicate that the ligand-based reduction of formazanate ligand is easily accessible by using mild reducing agents (such as Cp2Co in the first example).

Scheme 1.7 Synthesis and chemical reduction of the (formazanate)B(OAc)2 complex from Hicks and co-workers (left); and the structure of heteroleptic Pd complex (right) 1.4.2 Redox-Active Nature of -Diketiminate Ligands

Formazanate ligands are organic compounds structurally similar to well-known - diketiminate ligands (or NacNac-, Chart 1.4), which are nitrogen-based monoanionic bidentate ligands and have been studied extensively during the past decades.25 The - diketiminate ligands have found widespread use as a versatile auxiliary ligand. The large tunability of the steric parameters or electronic properties of the -diketiminate ligands is its major advantage in chemical research. By using different starting material, all the substituents (R1-R5) of -diketiminate ligands can be changed. In addition to the rich chemistry of the - diketiminate ligands, the redox-active property of -diketiminate ligands has been established in the past few years.

Chart 1.4 General structure (left), HOMO (middle), and LUMO (right) of -diketiminate ligands

Chapter 1

The ligand-based oxidation of the -diketiminate ligands was first reported by Khusniyarov, Wieghardt and co-workers based on a bis(-diketiminate)NiII complex (Scheme 1.8).26 The ligand-based oxidation of -diketiminate ligands was supported by EPR experiments and DFT calculations. The first crystallographic study of the singly oxidized bis(- diketiminate)NiII was reported by Itoh and co-workers in 2014.27 The metrical parameters of II + ligand backbone of the ([L]2Ni ) do not show much difference compared to the parent II complex [L]2Ni . This is due to the nonbonding character of the HOMO of -diketiminate ligands (Chart 1.4).26,27

Scheme 1.8 Ligand-based oxidation of bis(-diketiminate)Ni complex The ligand-based reduction of -diketiminate ligands was first mentioned and structurally characterized in 2002 by Lappert and co-workers in systems of Sm and Yb complexes.25c The - crystal structure of the reduced bis(-diketiminate)Sm complex ([L2Sm] ) clearly shows elongations of N-C bond lengths of one of the -diketiminate ligand indicating a dianionic ligand backbone (L-2, Scheme 1.9). In the next year, the ligand-based 2-electron reduction, which results in a trianionic ligand backbone (L-3, Scheme 1.9), was reported by the same group based on ytterbium and lithium complexes.28 The elongations of N-C bond lengths in reduced bis(-diketiminate) metal complexes are due population of orbitals that have anti- bonding character between C and N (the LUMO in Chart 1.4). Generally speaking, very strong reducing agents, such as Li metal or Yb-naphthalene complex, are necessary to synthesize reduced products of -diketiminate metal complexes.

Scheme 1.9 Consecutive two-electron reduction of a -diketiminate ligand

Introduction

1.4.3 Metal Complexes Bearing Fromazanate Ligands Even though -diketiminate ligands are very good supporting ligands, utilization of their redox-active properties for (catalytic) reactions is difficult, which is mainly due to the low stability of the oxidized forms26 and poor accessibility of reduced forms (which are only observed at very negative potentails)28. In the case of formazanate ligands, the four nitrogen atoms at the ligand backbone make its reduction chemistry occurs at much more accessible potentials and stabilizes the resulting products (similar with the case of the verdazyl radicals). In addition, the structural similarity with -diketiminate ligands suggests that formazanate ligands could also be good supporting ligands. These two features make formazanate ligands an attractive candidate for the development of a new redox-active ligand platform; however, the coordination chemistry of formazanate ligand is relatively unexplored and most of the reported literature focused on the late transition metal complexes such as Pd, Co, Fe, Ni, and Cu.29 Besides the limited examples of formazanate metal complexes, there are only few documented examples reporting chemical transformations based on formazanate metal 30 29b,31 complexes such as H2O2 decomposition, and ethylene oligomerization and none of them utilize (or describe) the redox-active nature of formazanate ligands.

1.5 Overview of Thesis The goal of the research described in this thesis is to establish the redox-active feature of formazanate ligands. To achieve this aim, synthetic procedures and characterization methods of the free ligand and metal complexes are developed. The redox properties of formazanate ligands were established by electrochemical methods, and the first examples of isolated and well-characterized ligand-based reduction products are described in this thesis.

In Chapter 2, the general method of formazan synthesis is introduced. The method was applied to synthesize a series of formazan ligands, which contain a large diversity of steric and electronic properties of the substituents. A synthetic procedure for the di-formazan system was developed, in which the two formazan fragments are linked by a phenylene linker at meta and para position.

In Chapter 3, mono- and bis(formazanate)zinc complexes are synthesized and characterized. The solid-state and solution-state studies of the zinc complexes reveal that formazanate ligands show a high flexibility in coordination chemistry. At room temperature, the equilibrium between six- and five-membered chelate rings was established. The bis(formazanate)zinc complexes show four steps of one-electron reduction in cyclic

Chapter 1 voltammetry experiments. The crystal structures and EPR spectra of the reduced (formazanate)zinc radical (or diradical) compounds conclusively establish formazanate ligands as a novel class of redox-active ligands.

In Chapter 4, an unexpected zinc-to-boron transmetallation from bis(formazanate)zinc complexes to mono(formazanate)boron difluoride complexes is described. The key intermediate of the transmetallation reaction, which is a six-coordinated zinc complex having 1 3 5 a composition of [R NNC(R )NN(BF3)R ]2Zn, was isolated and fully characterized. A key step of the reaction is an isomerization from six-membered chelate ring to five-member chelate ring of formazanate ligand to open space around the zinc center to accommodate incoming substrates (BF3 in this case).

In Chapter 5, the synthetic procedures for (formazanate)boron complexes (LBX2; X = F, Ph, and H) are described. All the LBX2 show two (quasi)reversible redox couples in their cyclic ·- voltammetry experiments. The 1-electron reduction products [LBF2] were isolated and characterized by EPR spectroscopy and x-ray crystallography. In addition, the absorption and emission spectra of the neutral (formazanate)boron complexes show that these are a new class of fluorescent dyes (structurally related to BODIPYs) that show strong absorption and large Stokes shifts but are only weakly emissive (low quantum yield).

In Chapter 6, attempts to synthesize 2-electron reduction products of (formazanate)boron difluoride (LBF2) complexes are described. While these products appear relatively stable based on electrochemical methods, the chemical reaction of 2 equivalents of Na/Hg and the

LBF2 complex results in a series of BN-heterocycles which were characterized by NMR and X-ray crystallography. The formation of the BN-heterocycles is shown to go through a reactive (formazanate)B(I) intermediate, which is stabilized by the low-lying * orbital of the formazanate ligand. The isolated BN-heterocycles can be chemically oxidized back to the

LBF2 complex by reacting with XeF2. The regeneration of LBF2 suggests that the (formazanate)B moiety that is incorporated in the BN-heterocycles shows reactivity that derives from a B(I) fragment.

In Chapter 7, the thermolysis reaction of the (formazanate)boron dihydride (LBH2) is discussed. Heating up the LBH2 complexes results in a series of intramolecular hydride transfer products. A series of intermediates was characterized by 1D and 2D NMR spectroscopy. One of the identified intermediates contains a cyclohexadiene substituent, which indicates that first step of the reaction is an unusual hydride transfer to the ortho

Introduction position of the N-Ar ring on the ligand. In the case of unsymmetrical formazanate ligands, the distribution of isomers of each intermediate is affected by the substituents of the formazanate ligand.

1.6 References (1) Colacot, T. J. Platinum Metals Rev., 2011, 55, 84–90. (2) Chen, X.; Engle, K. M.; Wang, D. H.; Yu, J. Q. Angew. Chem. Int. Ed., 2009, 48, 5094–5115. (3) (a) Spargo, P. L. Transition Metals for Organic Synthesis; 2nd Ed; WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany; 2004. (b) Diederich, F.; Stang, P. J. Metal-catalyzed cross-coupling reactions; 2nd Ed; WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany; 2008. (4) (a) Dzik, W. I.; van der Vlugt, J. I.; Reek, J. N. H.; de Bruin, B. Angew. Chem. Int. Ed., 2011, 50, 3356–3358. (b) Chirik, P. J.; Wieghardt, K. Science, 2010, 327, 794–795. (c) Lyaskovskyy, V.; de Bruin, B. ACS Catal., 2012, 2, 270–279. (d) van der Vlugt, J. I. Eur. J. Inorg. Chem., 2011, 2012, 363– 375. (5) Kaim, W.; Schwederski, B. Coord. Chem. Rev., 2010, 254, 1580–1588. (6) Lamb, D. C.; Lei, L.; Warrilow, A. G. S.; Lepesheva, G. I.; Mullins, J. G. L.; Waterman, M. R.; Kelly, S. L. J. Virol., 2009, 83, 8266–8269. (7) Poulos, T. L.; Finzel, B. C.; Howard, A. J. J. Mol. Biol., 1987, 195, 687–700. (8) Rittle, J.; Green, M. T. Science, 2010, 327, 933–937. (9) (a) Ito, M.; Que, L. Angew. Chem. Int. Ed., 1997, 36, 1342–1344. (b) Lin, G.; Reid, G.; Bugg, T. D. H. J. Am. Chem. Soc., 2001, 123, 5030–5039. (c) Borowski, T.; Bassan, A.; Siegbahn, P. E. M. Chem. Eur. J., 2004, 10, 1031–1041. (d) Borowski, T.; Bassan, A.; Siegbahn, P. E. M. Biochemistry, 2004, 43, 12331–12342. (e) Borowski, T.; Siegbahn, P. E. M. J. Am. Chem. Soc., 2006, 128, 12941–12953. (10) (a) Pierpont, C. G. Inorg. Chem., 2011, 50, 9766–9772. (b) Henthorn, J. T.; Lin, S.; Agapie, T. J. Am. Chem. Soc., 2015, 137, 1458–1464. (c) Kramer, W. W.; Cameron, L. A.; Zarkesh, R. A.; Ziller, J. W.; Heyduk, A. F. Inorg. Chem., 2014, 53, 8825–8837. (11) (a) Machata, P.; Herich, P.; Lušpai, K.; Bucinsky, L. Organometallics, 2014, 33, 4846–4859. (b) Petrenko, T.; Ray, K.; Wieghardt, K. E.; Neese, F. J. Am. Chem. Soc., 2006, 128, 4422–4436. (c) Cappillino, P. J.; Pratt, H. D.; Hudak, N. S.; Tomson, N. C.; Anderson, T. M.; Anstey, M. R. Adv. Energy Mater., 2014, 4, 1300566. (d) Lippert, C. A.; Arnstein, S. A.; Sherrill, C. D.; Soper, J. D. J. Am. Chem. Soc., 2010, 132, 3879–3892. (e) Lippert, C. A.; Hardcastle, K. I.; Soper, J. D. Inorg. Chem., 2011, 50, 9864–9878. (f) Blackmore, K. J.; Lal, N.; Ziller, J. W.; Heyduk, A. F. J. Am. Chem. Soc., 2008, 130, 2728–2729. (12) Blusch, L. K.; Craigo, K. E.; Martin-Diaconescu, V.; McQuarters, A. B.; Bill, E.; Dechert, S.; DeBeer, S.; Lehnert, N.; Meyer, F. J. Am. Chem. Soc., 2013, 135, 13892–13899. (13) Stanciu, C.; Jones, M. E.; Fanwick, P. E.; Abu-Omar, M. M. J. Am. Chem. Soc., 2007, 129, 12400– 12401. (14) (a) Sylvester, K. T.; Chirik, P. J. J. Am. Chem. Soc., 2009, 131, 8772–8774. (b) Bart, S. C.; Chłopek, K.; Bill, E.; Bouwkamp, M. W.; Lobkovsky, E.; Neese, F.; Wieghardt, K.; Chirik, P. J. J. Am. Chem. Soc., 2006, 128, 13901–13912. (c) Bouwkamp, M. W.; Bowman, A. C.; Lobkovsky, E.; Chirik, P. J. J. Am. Chem. Soc., 2006, 128, 13340–13341. (d) Russell, S. K.; Lobkovsky, E.; Chirik, P. J. J. Am. Chem. Soc., 2011, 133, 8858–8861. (15) Nguyen, A. I.; Zarkesh, R. A.; Lacy, D. C.; Thorson, M. K.; Heyduk, A. F. Chem. Sci., 2010, 2, 166- 169. (16) Myers, T. W.; Berben, L. A. Chem. Commun., 2013, 49, 4175–4177. (17) (a) Ullmann's Encyclopedia of Industrial Chemistry; Wiley-VCH Verlag GmbH & Co. KGaA, Ed.; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, Germany, 2000. (b) Bamberger, E.; Wheelwright, E. Ber. Dtsch. Chem. Ges., 1892, 25, 3201-3213. (c) Friese, P. Ber. Dtsch. Chem. Ges., 1875, 8, 1078– 1080. (18) (a) Moodley, S.; Koorbanally, N. A.; Moodley, T.; Ramjugernath, D.; Pillay, M. J. Microbiological Methods., 2014, 104, 72–78. (b) Martín, A.; Morcillo, N.; Lemus, D.; Montoro, E.; da Silva Telles, M. A.; Simboli, N.; Pontino, M.; Porras, T.; León, C.; Velasco, M.; Chacon, L.; Barrera, L.; Ritacco, V.; Portaels, F.; Palomino, J. C. Int. J. Tuberc. Lung Dis., 2005, 9, 901-906. (19) (a) Buzykin, B. I. Chem. Heterocycl. Comp., 2010, 46, 379–408. (b) Kuhn, R.; Munzing, W. Chem. Ber.-Recl. 1953, 86, 858–862. (c) Kuhn, R.; Neugebauer, F. A.; Trischmann, H. Monatsh. Chem., 1966, 97, 525–553. (d) Kuhn, R.; Trischmann, H. Monatsh. Chem., 1964, 95, 457–479. (e) Hausser, I.; Jerchel, D.; Kuhn, R. Chem. Ber., 1949, 82, 515–527.

Chapter 1

(20) Fico, R. M., Jr; Hay, M. F.; Reese, S.; Hammond, S.; Lambert, E.; Fox, M. A. J. Org. Chem., 1999, 64, 9386–9392. (21) (a) McKinnon, S.; Patrick, B. O.; Lever, A. Chem. Commun., 2010, 46, 773-775. (b) Barclay, T. M.; Hicks, R. G.; Lemaire, M. T.; Thompson, L. K. Inorg. Chem., 2003, 42, 2261–2267. (c) Brook, D. J R.; Fornell, S.; Stevens, J. E.; Noll, B.; Koch, T. H.; Eisfeld, W. Inorg. Chem., 2000, 39, 562-567. (d) Johnston, C. W.; McKinnon, S. D. J.; Patrick, B. O.; Hicks, R. G. Dalton Trans., 2013, 42, 16829– 16836. (e) McKinnon, S. D. J.; Patrick, B. O.; Lever, A. B. P.; Hicks, R. G. J. Am. Chem. Soc., 2011, 133, 13587–13603. (f) Sanz, C. A.; Ferguson, M. J.; McDonald, R.; Patrick, B. O.; Hicks, R. G. Chem. Commun., 2014, 50, 11676–11678. (g) Frolova, N. A.; Vatsadze, S. Z.; Stash, A. I.; Rakhimov, R. D.; Zyk, N. V. Chem. Heterocycl. Comp., 2006, 42, 1444–1456. (h) Myers, T. W.; Chavez, D. E.; Hanson, S. K.; Scharff, R. J.; Scott, B. L.; Veauthier, J. M.; Wu, R. Inorg. Chem., 2015, 54, 8077-8086. (22) Gilroy, J. B.; Ferguson, M. J.; McDonald, R.; Patrick, B. O.; Hicks, R. G. Chem. Commun., 2007, 43, 126–128. (23) Gilroy, J. B.; University of Victoria (Canada). The design, synthesis, and chemistry of stable verdazyl radicals and their precursors; ProQuest, 2008. (24) Gilroy, J. B.; Ferguson, M. J.; McDonald, R.; Hicks, R. G. Inorg. Chim. Acta., 2008, 361, 3388–3393. (25) (a) Tsai, Y.-C. Coord. Chem. Rev., 2012, 256, 722–758. (b) Asay, M.; Jones, C.; Driess, M. Chem. Rev., 2011, 111, 354–396. (c) Bourget-Merle, L.; Lappert, M. F.; Severn, J. R. Chem. Rev., 2002, 102, 3031–3066. (26) Khusniyarov, M. M.; Bill, E.; Weyhermueller, T.; Bothe, E.; Wieghardt, K. Angew. Chem. Int. Ed., 2011, 50, 1652–1655. (27) Takaichi, J.; Morimoto, Y.; Ohkubo, K.; Shimokawa, C.; Hojo, T.; Mori, S.; Asahara, H.; Sugimoto, H.; Fujieda, N.; Nishiwaki, N.; Fukuzumi, S.; Itoh, S. Inorg. Chem., 2014, 53, 6159–6169. (28) (a) Eisenstein, O.; Hitchcock, P. B.; Khvostov, A. V.; Lappert, M. F.; Maron, L.; Perrin, L.; Protchenko, A. V. J. Am. Chem. Soc., 2003, 125, 10790–10791. (b) Avent, A. G.; Hitchcock, P. B.; Khvostov, A. V.; Lappert, M. F.; Protchenko, A. V. Dalton Trans., 2004, 33, 2272–2280. (29) (a) Sigeikin, G. I.; Lipunova, G. N.; Pervova, I. G. Russ. Chem. Rev., 2006, 75, 885-900. (b) Siedle, A. R.; Pignolet, L. H. Inorg. Chem., 1980, 19, 2052–2056. (c) Zaĭdman, A. V.; Khasbiullin, I. I.; Belov, G. P.; Pervova, I. G.; Lipunov, I. N. Pet. Chem., 2012, 52, 28–34. (d) Zaidman, A.; Pervova, I.; Vilms, A.; Belov, G.; al, E. Inorg. Chim. Acta., 2011, 367, 29-34. (e) Gok, Y.; Senturk, H. Dyes and Pigments. 1991, 15, 279–287. (30) Gorbatenko, Y. A.; Pervova, I. G.; Lipunova, G. N.; Maslakova, T. I.; Lipunov, I. N.; Sigeikin, G. I. Russ. J. Appl. Chem., 2005, 78, 936–939. (31) (a) Rezinskikh, Z. G.; Pervova, I. G.; Lipunova, G. N.; Maslakova, T. I.; Gorbatenko, Y. A.; Lipunov, I. N.; Sigeikin, G. I. Russ. J. Coord. Chem., 2008, 34, 659–663. (b) Pavlova, I. S.; Pervova, I. G.; Belov, G. P.; Khasbiullin, I. I. Pet. Chem., 2013, 53, 127-133.

Chapter 2

Formazan Synthesis

The synthetic procedures for a series of formazan ligands, which have different electronic and steric properties, are described. The results show that suitable reaction conditions and isolation procedures are different for each formazan derivative. Besides the mono-formazan ligand, the synthesis of di-formazan compounds that are linked via a phenylene spacer is also described in this chapter. In some cases where formazan synthesis was unsuccessful, we were able to isolate and characterize other products from these reactions. Analysis of their structure provides additional insight in potential complications during formazan synthesis.

Parts of this chapter have been published:

M.-C. Chang, P. Roewen, R. Travieso-Puente, M. Lutz, and E. Otten* ”Formazanate Ligands as Structurally Versatile, Redox-Active Analogues of β-Diketiminates in Zinc Chemistry” Inorg. Chem., 2015, 54, 379-388.

R. Travieso-Puente, M.-C. Chang and E. Otten* “Alkali metal salts of formazanate ligands: diverse coordination modes as a result of the nitrogen-rich [NNCNN] ligand backbone” Dalton Trans., 2014, 43, 18035-18041. Chapter 2

Chapter 2 Formazan Synthesis 2.1 Introduction

Formazans, organic compounds that contain the N=N-C=N-N fragment, have a long history as dye molecules.1 The first formazan was synthesized more than a century ago by von Pechmann.2 Since then, several synthetic procedures for formazan synthesis have been developed.3 The most common synthetic procedure used today is using diazonium salts (Method A) to form the C-N bond of formazans. The formazan synthesis can be divided into two subgroups based on the starting materials (Method A1: arylhydrazone; Method A2: compounds having active methylene groups).

2.1.1 Method A1: Formazan Synthesis from Diazonium Salt and Arylhydrazone

Method A is the most common synthetic procedure of formazan synthesis and a majority of the known formazans were synthesized by this method, especially for triarylformazans or macrocyclic formazans.3ef,4 A key reagent used in Method A to synthesize formazans is an arylhydrazone (Method A1, Scheme 2.1), which can be synthesized by the condensation reaction of aldehyde and arylhydrazine in high yield (> 95%). The coupling reaction of the diazonium salt and arylhydrazone to generate a formazan occurs under basic reaction conditions. Pyridine, sodium hydroxide, sodium acetate and triethylamine are the common sources of the base of the reaction. This also means that for different combinations of starting materials, which includes arylhydrazone and diazonium salt, and solvents of the reaction, different sources of the base might be needed. In some of the case, a basic buffer system is used.5

Scheme 2.1 General reaction scheme of Method A1

The mechanism of the coupling reaction of the diazonium salt and the arylhydrazone is still not very clear, but likely mechanisms are shown in Scheme 2.2.3a The first step of the coupling reaction is the deprotonation of hydrazone NH by the base leading to the formation of a resonance stabilized azaenolate. The terminal nitrogen of the azaenolate anion will be a

Formazan Synthesis nucleophile attacks the diazonium cation forming the compound A. After the rearrangement of A, the desired formazan is formed. This mechanism was further supported by the isolation of the unstable compound B, which can isomerize to a formazan, by Busch in 1931.6

Scheme 2.2 Mechanisms of Method A1

Another possible mechanism after the deprotonation is that the diazonium cation attacks at the -carbon position of the azaenolate anion leading to a di-azo compound (C). The formazan can be isolated from C by deprotonation and protonation (tautomerization). The second pathway was supported by the isolation of compound D (Scheme 2.2).7 The importance of the hydrazone NH was proved by the work of Pechmann8 and Bush9. They used disubstituted hydrazone (E) as the starting material for the coupling reaction. Under a similar reaction condition with previous cases, the desired formazan was not isolated or detected but an unexpected azohydrazone (F) was identified (Scheme 2.3). While the experimental data on the mechanism of formazan formation using Method A1 does not lead to a conclusive picture (in fact could well differ from substrate to substrate), this represents the most versatile method for the synthesis of formazans today.

Scheme 2.3 Reaction of disubstituted hydrazone (E) with diazonum salt

Chapter 2

2.1.2 Method A2: Formazan Synthesis from Diazonium Salt and Active Methylene Group

The second commonly used strategy for formazan synthesis is a coupling between diazonium salts with compounds having active methylene groups, the activation of which is usually due to a carboxyl or nitro group on the -carbon atom. In this class of reaction, several functional groups, such as carboxyl, acetyl, ethoxycarbonyl and benzoyl substituents, can be replaced by diazonium cations under basic condition. The reaction of acetoacetic acid with diazonium cation is a beautiful example of this type of reaction (Scheme 2.4). Upon treatment of acetoacetic acid with one, two, or three equivalents of diazonium salt, the mono-, di- or tri- substitution product can be isolated, respectively,10 in which the di- and tri-substitution products contain the N=N-C=N-N formazan moiety.

Scheme 2.4 Reaction of acetoacetic acid with diazonium cation (Method A2)

2.1.3 Structure of Formazans and Formazanate Anions

The structures of formazans, which are shown in Chart 2.1, have two alternating double bonds in the backbone. The two double bonds in the ligand backbone makes that formazans can exist in four possible isomeric forms: syn, s-cis (closed form); syn, s-trans (open form); anti, s-cis; and anti, s-trans (linear form).11 Experimentally, only closed, open and linear forms have been observed, and they can be easily distinguished by 1H NMR spectroscopy. For example, the 1H NMR resonance of the NH group of the closed, open and linear form is located at 15-16, 12-10 and 10-8 ppm, respectively.3g Generally speaking, the closed form is the most usual and stable isomer of formazans; the open or linear form usually appears when the R3 substituent is a cyano or methylthio group.3g,12

The four nitrogen atoms of the formazan backbone and the four possible isomers lead to highly flexible coordination chemistry of formazanate ligands. The formazanate ligand can coordinate with metal ions in three different ways forming six-, five-, or four-membered chelating rings13 (Chart 2.1), which are not possible for the -diketiminate ligand system. In

Formazan Synthesis addition, the free formazanate anion prefers the linear isomer to reduce the repulsion between two lone pairs on the two terminal nitrogen atoms.12,13

Chart 2.1 Isomers of formazan, formazanate metal complex and formazanate anion

Despite that Method A1 shows great potential for synthesizing unsymmetrical formazans (R1 ≠ R5), the synthetic procedure has remained relatively little explored. In this chapter, several synthetic methods for unsymmetrical formazans, which contains steric and electronic asymmetry, were developed. Beside simple mono-formazan compounds (1a-1j), two examples of phenylene-linked di-formazan ligand systems (1k and 1l) were also achieved.

2.2 Formazan Synthesis

In this chapter, the formazans were synthesized by using Method A1 and Method A2 (Scheme 2.5). In Method A1, condensation of a monosubstituted hydrazine with an aldehyde generates a hydrazone, which then reacts with a diazonium salt under basic conditions to give the desired formazan. This synthetic approach can be applied to a broad range of substituents. By varying the substitution pattern on the starting materials, all the substituents on the resulting formazans can be changed. The potential limitations of this method are the accessibility of starting materials and stability of diazonium salts. In Method A2, the synthesis of formazan is achieved by treatment of acidic methylene compounds such as cyanoacetic acid with aryldiazonium salts under basic conditions. By using this method, formazans with sterically demanding N-aromatic groups can be prepared.3g One of the limitations of this method is that it produces only symmetrical formazan derivatives (R1 = R5). For Method A1, the major challenge is to find the suitable reaction conditions, which include the base source,

Chapter 2 temperature, and solvent mixture, to promote the desired coupling reaction and to slow down the competing decomposition of the diazonium salt.

Scheme 2.5 Formazan synthesis

2.2.1 Mono-Formazan Ligand Synthesis

As a starting point, we synthesized the known formazan PhNNC(p-tolyl)NNHPh (1a) in a biphasic reaction medium via the procedure published by Hicks and co-workers (Method A1).14 The coupling of trimethylacetaldehyde-phenylhydrazone with phenyldiazonium chloride afforded the bis(phenyldiazenyl)methane compound PhNNCH(tBu)NNPh (1b), which does not spontaneously tautomerize to the corresponding formazan.7 Deprotonation of 1b, however, is facile and results in a delocalized formazanate anion that may be reprotonated to give the formazan tautomer. Attempts to prepare the somewhat more sterically demanding asymmetric derivative MesNNC(p-tolyl)NNHPh (1c) by treatment of the hydrazone (p- + tolyl)C=N-NHPh with MesN2 (either prepared in situ as the chloride or isolated as the BF4- salt)15 resulted in intensely colored reaction mixtures from which we were unable to isolate the formazan. Changing the solvent in which the reaction was conducted to acetone/water with NaOH as the base yielded the product MesNNC(p-tolyl)NNHPh (1c) in moderate yield

Formazan Synthesis upon crystallization from CH2Cl2/MeOH. The symmetrical derivative MesNNC(p- tolyl)NNHMes (1d) required yet another solvent mixture: this sterically demanding formazan could be obtained in low yield (11%) from coupling of MesN2Cl with (p-tolyl)C=N-NHMes in MeOH with NaOH/NaOAc.5 Using similar synthetic procedures, electron-poor formazans with C6F5 substituents either at the terminal N atoms (PhNNC(p-tolyl)NNH(C6F5) (1e) and

MesNNC(p-tolyl)NNH(C6F5) (1f) or the backbone C atom (PhNNC(C6F5)NNHMes (1g) and

(C6F5NNC(C6F5)NNHMes (1h) were obtained. Formazan derivatives with an electron- withdrawing cyano group on the central carbon atom can be isolated from the direct coupling of cyanoacetic acid with two equivalents of aryl diazonium salts under basic conditions (Method A2).3a Using this strategy we prepared the formazan PhNNC(CN)NNHPh (1i) and MesNNC(CN)NNHMes (1j). In related -diketiminate chemistry, 2,6-disubstituted aromatic rings have become popular since they provide steric protection above and below the coordination plane; we anticipate that the mesityl groups in our formazanate ligands will behave similarly.

Scheme 2.6 Synthesis of 1j-B(C6F5)3

The cyano group in 1i and 1j is a potential site to tune the electronic properties of the ligand by coordination of a neutral Lewis acid. In order to prove this concept, 1j was reacted with tris(pentafluorophenyl)borane (B(C6F5)3) in toluene at room temperature (Scheme 2.6). The red crystalline material 1j-B(C6F5)3 can be isolated from DCM/hexane mixture with high 19 yield (77 %). In the F NMR spectrum, 1j-B(C6F5)3 shows three resonances at -133, -156 and

-163 ppm (in CDCl3), which are in a good agreement with the reported CH3CN-B(C6F5)3 16 system (-135, -155, -163 ppm, in C6D6), suggesting an acid-base interaction between the cyano group and B(C6F5)3. This acid-base interaction was further confirmed by single crystal X-ray crystallography (Figure 2.1, metrical parameters in Table 2.1). The B1-N5 bond length is

1.578(2) Å, which is similar to the reported B(C6F5)3-adduct of the cyano-substituted (- 17 diketiminate)ZrCl2Cp complex (1.586(4) Å) and shorter than the CH3CN-B(C6F5)3 system

Chapter 2

(1.616(3) Å).16 Comparing with 1j, the bond lengths of N3-N4 (1.309(2) Å) and C10-C11

(1.426(2) Å) of 1j-B(C6F5)3 are shorter, and the bond length of N3-C10 (1.321(2) Å) is elongated. The change of the bond lengths before and after the coordination of B(C6F5)3 to 1j is similar to the reported (-diketiminate)ZrCl2Cp complex (Figure 2.1 and metrical parameters in Table 2.1). The bond length of N1-N2, N2-C10, and C11-N5 of 1j-B(C6F5)3 do not change much compared to 1j. The bond lengths distribution described above suggest the contribution from different resonance structures (i and ii, Scheme 2.7).17b The resonance structure ii shows double bond character for the N3-N4 and C10-C11bonds and single bond character for the N3-C10 bond, which explains the shorter N3-N4 and C10-C11 bond and longer N3-C10 bond of 1j-B(C6F5)3 than 1j. In addition, the resonance structure ii has extra negative charge at N5, which makes N5 a better electron donor resulting in a relatively short N5-B1 bond.

Figure 2.1 Crystal structures of 1j-B(C6F5)3 (left) and (-diketiminate)ZrCl2Cp complex (right); Structures are showing 50% probability ellipsoids (all hydrogen atoms, chloride atoms and solvent molecules omitted for clarity).

a b Table 2.1 Selected bond lengths (Å) of 1j , 1j-B(C6F5)3, (nacnac)ZrCl2Cp and (nacnac)ZrCl2Cp- b B(C6F5)3 1j 1j-B(C6F5)3 (nacnac)ZrCl2Cp (nacnac)ZrCl2Cp-B(C6F5)3 N1-N2 1.269(2) 1.271(2) N1-C2 1.301(3) 1.296(4) N2-C10 1.392(2) 1.390(2) C2-C3 1.463(5) 1.478(5) N3-C10 1.304(2) 1.321(2) C3-C4 1.425(6) 1.447(5) N3-N4 1.325(2) 1.309(2) C4-N5 1.323(4) 1.310(3) C10-C11 1.444(3) 1.426(3) C3-C7 1.437(4) 1.409(4) C11-N5 1.137(3) 1.139(2) C7-N8 1.139(4) 1.135(4) N5-B1 - 1.578(2) N8-B1 - 1.586(5) a: Reported by Gilroy and co-workers.3g b: Reported by Rojas and co-workers.17

Formazan Synthesis

Scheme 2.7 Resonances structures of 1j-B(C6F5)3

Scheme 2.8 Synthesis of 1m and 2

Introducing a functional group at the para position of R3 substituents is a way to tune the electronic properties without influencing the steric properties of formazan. Therefore, 3-(4- bromophenyl)formazan (1m) was prepared (Scheme 2.8). A potential application of 1m is that other functional groups or fragments can be introduced at the bromo substituent to expand the research scope of the formazanate ligand.18 Compound 1m was synthesized by the same method as 1a. After recrystallization from DCM/MeOH mixture, two different crystalline products can be identified in the isolated product. One of them is the desired formazan (1m), which is a red crystal; the other one is a bis(diazo) compound (2), which is a colorless crystal. The formation of these two product was confirmed by single crystal X-ray crystallography (Figure 2.2). The N-N, N-C and C-C bond lengths of 2 are 1.239(2), 1.480(2), and 1.537(2) Å, respectively. The N-N bond length is very close to the normal azo compound, and the C-C bond length is close to normal C-C single bond. These metrical parameters suggest that 2 is the C-C coupling product of benzyl diazene radicals. The NNCCNN backbone of 2 has been reported in the 1960’s and 1970’s, but most of the compounds with this general atom connectivity that are known to date are di-hydrazone compounds instead of di-azo compounds.19 Compound 2 could have potential as being a bidentate ligand or photoswitchable molecule.

Chapter 2

Figure 2.2 Crystal structures of 1m (left) and 2 (right); Structures are showing 50% probability ellipsoids (all hydrogen atoms and solvent molecules omitted for clarity).

N N N + - PhNN Cl PhNN+Cl- N N N N NH N base NH base NH

1n 3 (not observed) Scheme 2.9 Synthesis of 3

In order to introduce an extra coordination site of the formazan ligand, we attempted to synthesize 3-(2-pyridyl)formazan from 2-Pyridinecarboxaldehyde. The desired hydrazone can be isolated with almost quantitative yield (97%), but the synthesis of PhNNC(2-pyridyl)NNPh (1n) by using the same reaction conditions as those employed in the synthesis of 1a was not successful (Scheme 2.9). Compound 3 was isolated after the reaction of hydrazone with phenyl diazonium salt in pyridine/methanol/acetic acid mixture (16%). The 1H-NMR of 3 clearly shows two distinct groups of resonances from phenyl groups and a singlet of an NH group at 7.79 ppm, which is outside the normal range of formazans (Figure 2.3). The formation of 3 was further confirmed by single crystal X-ray crystallography (Figure 2.3). The formation of 3 was also reported by Nikolay V. Zyk in 2006,20 in which a solvent mixture of pyridine, water and acetic acid was used.

Formazan Synthesis

1 Figure 2.3 H NMR spectrum of 3 (CDCl3, 400 MHz, left); crystal structure of 3 (right) showing 50% probability ellipsoids (solvent molecules omitted for clarity).

2.2.2 Phenylene-Linked di-Formazan Ligand Synthesis

Scheme 2.10 Synthesis of phenylene-linked di-formazan ligands (1k and 1l)

In order to study bimetallic systems incorporating formazan or formazanate ligands, two di- formazan ligands were prepared (Scheme 2.10).21 The synthesis of di-hydrazones is very similar to a mono-hydrazone system; condensation of a terephthalaldehyde or isophthalaldehyde with phenyl hydrazine affords the desired di-hydrazone derivatives. In the next step, reactions of di-hydrazone with phenyl diazonium chloride were conducted in pyridine/DMF mixture due to the low solubility of the di-hydrazones in more conventional solvents. The desired di-formazan ligands were further purified by recrystallization from 1 CH2Cl2/MeOH mixture to give the pure products in 15% (1k) and 50% (1l) yield. The H NMR spectra of 1k and 1l are shown in Figure 2.4. In both spectra, a peak at ~15 ppm that corresponds to 2 protons confirms the formation of a di-formazan compound. In the case of 1k, the highly symmetrical structure, which has D2h point group, results in 4 peaks with integrations 4:8:8:4 in the aromatic region. In the case of 1l, the singlet at 9.04 ppm, which corresponds to 1 proton, is a strong evidence for the presence of a 1,3-disubstutited benzene structure as the linker.

Chapter 2

1 Figure 2.4 H NMR spectra of 1k (THF-d8, 400 MHz, top) and 1l (CDCl3, 400 MHz, bottom)

2.3 Conclusion

In this chapter, several procedures for the synthesis of sterically or electronically unsymmetrical formazan ligands were successfully developed. Two examples of di-formazan compounds were also prepared and characterized. The results here show the generality of Method A for formazan synthesis. But the formation of 2 and 3 indicates that using Method A to synthesize formazan ligand is not straightforward all the time and competing side reactions can lead to significant amounts of other products through radical pathways. Even though in these cases the substituent we introduced is far away from the site where the desired C-N or N-N bond formation should take place (to give formazans), our results indicate that these remote positions can have a large influence on the products that are isolated.

2.4 Experimental Section

General Consideration. All manipulations were carried out under air atmosphere.

Deuterated solvents (CDCl3 Aldrich and C6D6 Apollo), pentafluorophenylhydrazine (Aldrich, 97%), 4-bromobenzaldehyde (Aldrich, 97%), 2-pyridinecarboxaldehyde (Acros 99%), terephthalaldehyde (Aldrich, 99%), isophthalaldehyde (Aldrich, 97%), NaOH (Acros), 2,4,6-

Formazan Synthesis trimethylaniline (Aldrich, 98%), aniline (Sigma-Aldrich, 99%), tetrabutylammonium bromide (Sigma-Aldrich, 99%), cyanoacetic acid (Aldrich, 99%), sodium nitrite (Sigma-Aldrich, 99%), sodium carbonate (Merck), phenylhydrazine (Aldrich, 97%), p-tolualdehyde (Aldrich, 97%) and tert-butyl nitrite (Aldrich, 90%) were used as received.

The compounds 1a,14 1b,7 1i,3g and 1j3g were synthesized according to published procedures. NMR spectra were recorded on Mercury 400 or Varian 500 spectrometers. The 1H and 13C NMR spectra were referenced internally using the residual solvent resonances and reported in ppm relative to TMS (0 ppm); J is reported in Hz. Assignment of NMR resonances was aided by gradient-selected COSY, NOESY, HSQC and/or HMBC experiments using standard pulse sequences. Elemental analyses were performed at the Microanalytical Department of the University of Groningen.

Hydrazone synthesis

[PhNNC(pTol)H]. Phenylhydrazine (2.70 g, 25 mmol) was combined with p-tolualdehyde (3.00 g, 25 mmol) and ethanol (40 mL). After the mixture was stirred at RT for 30 min at which time a yellow precipitate had formed, water (100 mL) was added and the mixture was stirred for additional 10 min. The light yellow solid was collected by filtration and washed 1 with ether (3 × 5 mL), yield 4.80 g (22.8 mmol, 92 %). H NMR (400 MHz, C6D6, 25 °C):  7.59 (d, 2H, J = 8 Hz, p-tolyl CH), 7.24 (t, 2H, J = 8 Hz, Ph m-CH), 7.11 (d, 2H, J = 8 Hz, Ph o-CH), 7.05 (d, 2H, J = 8 Hz, p-tolyl CH), 6.86 (t, 1H, J = 7 Hz, Ph p-CH), 6.80 (s, 1H, 13 N=CH), 6.69 (bs, 1H, NH), 2.12 (s, 3H, p-tolyl CH3). C NMR (100 MHz, C6D6, 25 °C): 

145.7, 138.7, 137.8, 133.8, 130.0, 129.9, 126.9, 120.5, 113.4, 21.7 (p-tolyl CH3) ppm.

[C6F5NNC(pTol)H]. Pentafluorophenylhydrazine (3.96 g, 20 mmol) was combined with p- tolualdehyde (2.40 g, 20 mmol) and ethanol (20 mL). After the mixture was stirred at RT for 30 min at which time a yellow precipitate had formed, the reaction mixture was cooled slowly to -30 °C for 1 day. The light yellow solid was collected by filtration and washed with cold 1 ethanol (5 mL), yield 4.40 g (14.7 mmol, 73 %). H NMR (400 MHz, C6D6, 25 °C):  7.49 (d, 2H, J= 8 Hz, p-tolyl CH), 6.98 (d, 2H, J= 8 Hz, p-tolyl CH), 6.66 (s, 1H, N=CH), 6.40 (s, 1H, 13 NH), 2.07 (s, 3H). C NMR (100 MHz, C6D6, 25 °C):  143.4, 140.1, 138.9 (dm, J= 246 Hz), 138.5 (dm, J= 244 Hz), 135.7 (dtt, J= 245, 14, 4 Hz), 132.4, 130.0, 127.2, 121.5 (tm, J= 11 19 Hz), 21.6 (p-tolyl CH3). F NMR (375 MHz, C6D6, 25 °C):  -156.8 (d, 2F, J= 23 Hz, C6F5 o-CF), -164.4 (td, 2F, J= 22, 5 Hz, C6F5 m-CF), -168.9 (tt, 1F, J= 22, 5 Hz, C6F5 p-CF) ppm.

Anal. calcd for C14H9F5N2: C, 56.01; H, 3.02; N, 9.33. Found: C, 55.64; H, 3.00; N, 9.23.

Chapter 2

[PhNNC(C6F5)H]. 2,3,4,5,6-pentafluorobenzaldehyde (1.96 g, 10 mmol) and phenylhydrazine (1.08 g, 10 mmol) were stirred at room temperature in ethanol (30 mL) for 3 hours. After the reaction 60 mL water was added to the reaction mixture, and stirred for 1 hour. The light yellow solid that precipitated was collected and washed with water and hexane.

After drying under reduced pressure overnight 2.8 g light yellow solid of PhNHNC(C6F5)H 1 (0.96 mmol, 96 %) was obtained. H NMR (C6D6): δ 7.18 (t, 2H, J= 8 Hz, Ph m-H), 7.04 (d, 2H, J= 8 Hz, Ph o-H), 6.84 (t, 1H, J= 8 Hz, Ph p-H), 6.75 (s, 1H, NH), 6.62 (s, 1H, NHNCH). 13 C NMR (C6D6): δ 144.8 (dm, J= 252 Hz, C6F5), 144.3 (Ph i-C), 140.3 (dtt, J= 252, 14, 5 Hz,

C6F5 p-F), 138.2 (dm, J= 245 Hz, C6F5), 130.0 (Ph m-C), 123.6 (q, J= 3 Hz, NHNC), 121.9 19 (Ph p-C), 113.6(Ph o-C), 111.5 (td, J= 12, 4 Hz, C6F5 i-C). F NMR (C6D6): δ -144.1 (dd, 2F,

J= 22, 8 Hz, C6F5 m-F), -157.0 (t, 1F, J= 21 Hz, C6F5 p-F), -163.6 (td, 2F, J= 21, 8 Hz, , C6F5 o-F) ppm. Anal. calcd for C13H7N2F5: C, 54.56; H, 2.47; N, 9.79. Found: C, 54.59; H, 2.44; N, 9.75.

[C6F5NNC(C6F5)H]. Pentafluorophenylhydrazine (991 mg, 5 mmol) was combined with pentafluorobenzaldehyde ( 980 mg, 5 mmol) and ethanol (15 mL). 100 mL of deionized water was added after the mixture was stirred at room temperature (RT) for 3 hours. The light yellow solid was collected by filtration and washed with water and hexane,yield 1.779 g (4.73 1 13 mmol, 94 %). H NMR (400 MHz, C6D6, 25 °C): d 6.55 (s, 1H, N=CH), 6.45 (s, 1H, NH). C

NMR (100 MHz, C6D6, 25 °C):  145.2 (dm, J= 254 Hz), 141.3 (dtt, J= 255, 14, 5 Hz), 138.2 (dm, J= 248 Hz), 136.8 (dtt, J= 248, 14, 4 Hz), 129.4 (q, J= 3 Hz), 119.9 (tm, J= 10 Hz), 110.1 (td, J= 12, 4 Hz), In the range of 140.5-139.8 and 137.9-137.3 are two groups of 19 resonances containing two carbon. F NMR (375 MHz, C6D6, 25 °C): d -142.71(dd, 2F, J=

21, 8 Hz, C6F5 o-CF), -154.5 (t, 1F, J= 22 Hz, C6F5 p-CF), -155.7 (d, 2F, J= 23 Hz, C6F5 o-

CF), -163.0 (td, 2F, J= 21, 7 Hz, C6F5 m-CF), -163.7 (td, 2F, J= 22, 5 Hz, C6F5 m-CF), -166.0

(t, 1F, J= 22 Hz, C6F5 p-CF) ppm. Anal. calcd for C13H2F10N2: C, 41.51; H, 0.54; N, 7.45. Found: C, 41.37; H, 0.53; N, 7.46.

[PhNNC(4-BrC6H4)H]. Phenylhydrazine (1.1 mL, 11 mmol) was combined with 4- bromobenzaldehyde (1.96 g, 11 mmol) and ethanol (150 mL). After the mixture was stirred at RT for overnight at which time a yellow precipitate had formed, water (100 mL) was added and the mixture was stirred for additional 10 min. The light yellow solid was collected by filtration and washed with water and hexane, yield 2.63 g (9.6 mmol, 90 %). 1H NMR (400

MHz, C6D6, 25 °C):  7.27 (d, 2H, J = 8 Hz, Br-C6H4 CH), 7.22-7.15 (m, 4H, Ph o-CH and m-CH), 6.99 (d, 2H, J = 8 Hz, Br-C6H4 CH), 6.82 (t, 1H, J = 8 Hz, Ph p-CH), 6.65 (bs, 1H,

Formazan Synthesis

13 NH), 6.48 (s, 1H, N=CH) ppm. C NMR (100 MHz, C6D6, 25 °C):  144.5, 135.3, 134.6, 131.6, 129.2, 127.3, 121.9, 120.2, 112.8 ppm.

[PhNNC(C5H4N)H]. Phenylhydrazine (2.0 g mL, 18.5 mmol) was combined with 2- pyridinecarboxaldehyde (2.0 g, 18.7 mmol) and ethanol (50 mL). After the mixture was stirred at RT for 3 hours at which time a yellow precipitate had formed, water (100 mL) was added and the mixture was stirred for additional 30 min. The light yellow solid was collected by filtration and washed with water and hexane, yield 3.58 g (18.2 mmol, 97 %). 1H NMR

(400 MHz, CDCl3, 25 °C):  8.54 (d, 1H, J = 5 Hz, Py CH), 8.10 (s, 1H, NH), 8.01 (d, 1H, J = 8 Hz, Py CH), 7.82 (s, 1H, N=CH), 7.69 (td, 1H, J = 8, 2 Hz, Py CH), 7.29 (t, 2H, J = 8 Hz, Ph m-CH), 7.20-7.17 (m, 1H, Py CH), 7.15 (d, 2H, J = 8 Hz, Ph o-CH), 6.91 (t, 1H, J = 7 Hz, 13 Ph p-CH) ppm. C NMR (100 MHz, C6D6, 25 °C):  154.4, 148.7, 144.0, 136.9, 136.5, 129.3, 122.5, 120.8, 119.7, 113.0 ppm.

[(PhNNCH)2(para-C6H4)]. Phenylhydrazine (4.8 g, 44.7 mmol) was combined with terephthalaldehyde (3.0 g, 22.3 mmol) and ethanol (250 mL). After the mixture was stirred at RT for 3 hrs at which time a yellow precipitate had formed. The light yellow solid was collected by filtration and washed with ethanol, yield 5.87 g (19.7 mmol, 88 %). 1H NMR

(400 MHz, DMSO, 25 °C):  10.41 (s, 2H, NH), 7.88 (s, 2H, CH=N), 7.67 (s, 4H, C6H4), 7.25 (t, 4H, J = 7 Hz, Ph m-CH), 7.11 (d, 4H, J = 8 Hz, Ph o-CH), 6.77 (t, 2H, J = 7 Hz, Ph p-CH) 13 ppm. C NMR (100 MHz, DMSO, 25 °C):  145.7 (Ph i-C), 136.6 (C=N), 135.8 (C6H4 o-C),

129.6 (Ph m-C), 126.3 (C6H4 CH), 119.3 (Ph p-C), 112.5 (Ph o-C) ppm.

[(PhNNCH)2(meta-C6H4)]. Phenylhydrazine (3.2 g, 29.6 mmol) was combined with isophthalaldehyde (2.0 g, 14.9 mmol) and ethanol (50 mL). After the mixture was stirred at RT for 3 hrs at which time a yellow precipitate had formed. The light yellow solid was collected by filtration and washed with ethanol, yield 4.37 g (13.9 mmol, 93 %). 1H NMR

(400 MHz, DMSO, 25 °C):  10.40 (s, 2H, NH), 7.92 (s, 2H, CH=N), 7.86 (s, 1H, C6H4 CH),

7.61 (d, 2H, J = 8 Hz, C6H4 CH), 7.41 (t, 1H, J = 8 Hz, C6H4 CH), 7.25 (t, 4H, J = 8 Hz, Ph m-CH), 7.12 (d, 4H, J = 8 Hz, Ph m-CH), 7.68 (t, 2H, J = 8 Hz, Ph p-CH) ppm. 13C NMR

(100 MHz, DMSO, 25 °C):  145.7 (Ph i-C), 136.7 (C=N), 136.7 (C6H4), 129.6 (Ph o-C),

129.4 (C6H4), 125.4 (C6H4), 123.5 (C6H4), 119.3 (Ph p-C), 112.5 (Ph m-C) ppm.

Formazan synthesis

PhNNC(p-tol)NNHMes 1c. PhNNC(p-tol)H (2.10 g, 10 mmol), sodium hydroxide (4.00 g,

100 mmol), water (105 mL) and acetone (105 mL) were mixed at 0 °C. MesN2BF4 (2.34 g, 10

Chapter 2 mmol) was slowly added into the solution. After stirring for 1h at 0 °C the reaction mixture was slowly warmed up to RT and stirred overnight. The crude product was extracted into

CH2Cl2 and the volatiles were removed on the rotavap. The desired formazan [PhNNC(p- tol)NNMes] was purified by recrystallization (CH2Cl2/MeOH) to give 1.6 g of product (4.5 1 mmol, 46 %). H NMR (400 MHz, CDCl3, 25 °C):  15.10 (s, 1H, NH), 7.97 (d, 2H, J= 7.6 Hz, p-tolyl CH), 7.56 (d, 2H, J= 8.0 Hz, Ph o-CH), 7.39 (t, 2H, J= 8.0 Hz, Ph m-CH), 7.23 (d, 2H, J= 8.0 Hz, p-tolyl CH), 7.149 (t, 1H, J= 7.2 Hz, Ph p-CH), 6.98 (s, 2H, Mes m-CH), 2.54 13 (s, 6H, Mes o-CH3), 2.39 (s, 3H, p-tolyl CH3), 2.33 (s, 3H, Mes p-CH3). C NMR (100 MHz,

CDCl3, 25 °C):  146.8 (Ph ipso-C), 144.7 (Mes ipso-C), 142.0 (NCN), 138.2 (Mes p-C), 137.4 (p-tolyl p-C), 135.3 (p-tolyl ipso-C), 131.4 (Mes o-C), 130.8 (Mes m-C), 129.6 (Ph m- C), 129.4 (p-tolyl CH), 125.9 (p-tolyl CH), 125.6 (Ph p-C), 117.4 (Ph o-C), 21.5 (p-tolyl p-

CH3), 21.3 (Mes p-CH3), 21.2 (Mes o-CH3) ppm. Anal. calcd for C23H24N4: C, 77.50; H, 6.79; N, 15.72. Found: C, 77.30; H, 6.76; N, 15.51.

MesNNC(p-tol)NNHMes 1d. Mesityl hydrazine·HCl (835 mg, 4.47 mmol) was dissolved in methanol (50 ml), and Et3N (480 mg, 4.74 mmol) and p-tolyl-aldehyde (535 mg, 4.45 mmol) were added. The resulting mixture was stirred for 30 minutes to give a yellow solution. NaOH (806 mg, 20.15 mmol) and NaOAc (524 mg, 6.39 mmol) were added and the solution was cooled to 0 °C and stirred for 30 minutes. Mesityl diazonium chloride was prepared in a separate flask by mixing mesityl aniline (623 mg, 4.61 mmol) with water (10 mL) and HCl

(1.25 ml) forming a white suspension that was cooled to 0 °C. NaNO2 (313 mg, 4.54 mmol) was added to this in small portions while everything went into solution, and the mixture was stirred for 30 minutes at 0 °C. The mesityl diazonium salt was added drop wise at 0 °C to the first flask containing the hydrazone solution. The color turned to red immediately and gas evolution was observed. After stirring for 2 hours at 0 °C a black oily precipitate was formed that was filtered off and purified by column chromatography over silica using DCM/hexane (1:5) as eluent (r = 0.54). The fractions were collected and subsequent removal of the solvent in vacuo afforded the product as a dark red solid (188 mg, 0.472 mmol, 11 % yield). 1H NMR

(C6D6, 400 MHz): δ 14.86 (1H, s, NH), 8.28 (2H, d, J = 7.8 Hz, Ph o-H), 7.18 (2H, d, J = 7.8

Hz, p-tol m-H), 6.76 (4H, s, Mes o-H), 2.37 (12H, s, Mes o-CH3), 2.18 (3H, s, p-tol p-CH3), 13 2.12 (6H, s, Mes p-CH3). C NMR (C6D6, 126 MHz): δ 144.05 (Mes i-C), 142.98 (p-tol i-C), 137.45 (p-tol p-C), 136.88 (Mes p-C), 136.31 (NNCNN), 130.97 (Mes o-C), 130.90 (Mes m-

C), 129.98 (p-tol o- C), 126.23 (p-tol m-C), 21.60 (p-tol p-CH3), 21.31 (Mes p-CH3), 20.35

Formazan Synthesis

(Mes o-CH3). Elemental analysis calculated for C26H30N4: C 78.35% H 7.59% N 14.06%; found C 78.18% H 7.60% N 13.96%.

C6F5NNC(p-tol)NNHPh 1e. Pentafluorophenylhydrazine (9.91 g, 50 mmol) was combined with p-tolualdehyde (5.90 mL, 50 mmol) and ethanol (200 mL) at RT. The mixture was stirred for 30 min at which time a light yellow solid had formed. CH2Cl2 (250 mL) and the aqueous solution containing sodium carbonate (20 g, 161.5 mmol) and tetrabutylammonium bromide (1.5 g, 4.7 mmol) was added to reaction mixture before being stirred at 0 °C for 1 h. A solution of diazonium salt made from stirring aniline (4.6 mL, 50 mmol), sodium nitrite (3.80 g, 54.5 mmol), water (25 mL), and hydrochloric acid (12.5 mL) for 30 min at 0 °C. The diazonium solution was added dropwise into the first solution at 0 °C. After stirring for 2 hours at RT the organic layer was collected and the water layer was extracted by CH2Cl2 (3 × 100 mL). The organic layer was combined and reduced by rotavapor. The desired formazan

[C6F5NNC(pTol)NNHPh] was purified by recrystallization (CH2Cl2/MeOH), yield 6.3 g (15.6 1 mmol, 32 %). H NMR (400 MHz, CDCl3, 25 °C):  15.08 (s, 1H, NH), 7.93 (d, 2H, J= 8.4 Hz, p-tolyl CH), 7.73 (d, 2H, J= 7.6 Hz, Ph o-CH), 7.47 (t, 2H, J= 8.0 Hz, Ph m-CH), 7.36 (t, 19 1H, J= 7.6 Hz, Ph p-CH), 7.22 (d, 2H, J= 8.0 Hz, p-tolyl CH), 2.39 (s, 3H, p-tolyl CH3). F

NMR (375 MHz, CDCl3, 25 °C):  -152.3 (d, 2F, J= 16 Hz, C6F5 o-CF), -159.9 (t, 1F, J= 21 13 Hz, C6F5 p-CF), -162.5 (td, 2F, J= 21, 6 Hz, C6F5 m-CF). C NMR (100 MHz, CDCl3, 25 °C):

 148.1 (Ph ipso-C), 143.7 (NCN), 139.7 (dm, J= 256 Hz, C6F5), 138.5 (dm, J= 252 Hz, C6F5 and C6F5), 138.4 (p-tolyl ipso-C), 133.7 (p-tolyl p-C), 129.8 (Ph m-C), 129.7 (Ph p-C), 129.5

(p-tolyl CH), 126.1 (p-tolyl CH), 123.2 (C6F5 ipso-C), 120.2 (Ph o-C), 21.5 (p-tolyl CH3) ppm.

Anal. calcd for C20H13F5N4: C, 59.41; H, 3.24; N, 13.86. Found: C, 58.98; H, 3.29; N, 13.49.

C6F5NNC(p-tol)NNHMes 1f. C6F5NNC(pTol)H (1.86 g, 6.2 mmol), sodium hydroxide (4.00 g, 100 mmol), water (100 mL) and acetone (100 mL) were mixed at 0 °C. MesN2BF4 (1.45 g, 6.2 mmol) was slowly added into the solution. After stirring for 1 hour at 0 °C the reaction mixture was slowly warmed up to RT. The crude product was extracted into CH2Cl2 and the volatiles were subsequently removed on the rotavap. The desired formazan [C6F5NNC(p- tol)NNMes] was purified by recrystallization (CH2Cl2/MeOH) to give 1.1 g of product (2.5 1 mmol, 41%). H NMR (400 MHz, CDCl3, 25 °C):  13.99 (s, 1H, NH), 7.88 (d, 2H, J= 8.0 Hz, p-tolyl CH), 7.21 (d, 2H, J= 8.0 Hz, p-tolyl CH), 6.99 (s, 2H, Mes m-CH), 2.48 (s, 6H, Mes o- 13 CH3), 2.38 (s, 3H, p-tolyl CH3), 2.34 (s, 3H, Mes p-CH3). C NMR (100 MHz, CDCl3, 25

°C):  144.9 (Mes ipso-C), 144.3 (NCN), 140.0 (Mes o-C), 139.4 (dm, J= 251 Hz, C6F5),

138.5 (dm, J= 251 Hz, C6F5), 138.4 (p-tolyl ipso-C), 137.6 (dm, J= 248 Hz, C6F5), 133.8 (p-

Chapter 2 tolyl p-C), 132.3 (Mes p-C), 130.9 (Mes m-CH), 129.4 (p-tolyl CH), 126.4 (p-tolyl CH),

122.7 (td, J= 9, 4 Hz, C6F5 ipso-C), 21.5(p-tolyl p-CH3), 21.4 (Mes p-CH3), 20.7 (Mes o-CH3). 19 F NMR (375 MHz, CDCl3, 25 °C):  -153.2 (d, 2F, J= 18 Hz, C6F5 o-CF), -162.2 (t, 1F, J=

21 Hz, C6F5 p-CF), -162.8 (td, 2F, J= 21, 5 Hz, C6F5 m-CF) ppm. Anal. calcd for C23H19F5N4: C, 61.88; H, 4.29; N, 12.55. Found: C, 61.75; H, 4.25; N, 12.43.

PhNNC(C6F5)NNHMes 1g. A flask was charged with PhNHNC(C6F5)H (1.72 g, 6 mmol), sodium hydroxide (2.00 g, 50 mmol), water (100 mL) and acetone (160 mL) and the mixture + - cooled to 0 °C. At this temperature, [MesN2] [BF4] (1.40 g, 6 mmol) was added slowly with stirring. The reaction mixture was slowly warmed up to RT and stirred for an additional 30 mins. Acetic acid was added to the reaction mixture until pH = 7. The reaction mixture was stirred for another 2 hours. The crude organic product was extracted into CH2Cl2 and the solution was concentrated. The product was purified by recrystallization from CH2Cl2/MeOH 1 at -30 °C for 2 days to give 1.2 g of PhNNC(C6F5)NNHMes (2.7 mmol, 45%). H NMR (400

MHz, CDCl3, 25 °C) δ 12.21 (s, 1H, NH), 7.41- 7.35 (m, 4H, Ph o-H, Ph m-H), 7.31 (t, 1H,

J= 6.6, Ph p-H), 6.95 (s, 2H, Mes m-H), 2.41 (s, 6H, Mes o-CH3), 2.31 (s, 3H, Mes p-CH3). 19 F NMR (376.4 MHz, C6D6, 25 °C) δ -139.3 (dd, 2F, J= 23.3, 7.5, C6F5 m-F), -154.5 (t, 1F, 13 J= 20.9, C6F5 p-F), -162.9 (td, 2F, J= 23.0, 5.6, C6F5 o-F). C NMR (400 MHz, CDCl3, 25 °C)

δ 145.5 (dm, J= 251.8, C6F5), 145.2 (Ph i-C), 144.8 (Mes i-C), 141.4 (dm, J= 259.0, C6F5),

140.0 (Mes p-C), 137.8 (dm, J= 249.6, C6F5), 135.8 (NNCNN), 132.8 (Mes o-C), 130.8 (Ph m-C), 129.7 (Mes m-C), 125.3 (Ph p-C), 116.7(Ph o-C), 111.5-111.1 (m, C6F5), 21.4 (Mes p-

CH3), 20.6 (Mes o-CH3). Anal. calcd for C19H11N4F5: C, 61.11; H, 3.96; N, 12.96. Found: C, 61.23; H, 3.98; N, 12.80.

C6F5NNC(C6F5)NNHMes 1h. C6F5NNC(C6F5)H (1.13 g, 3 mmol), sodium hydroxide (1.00 g,

25 mmol), water (100 mL) and acetone (120 mL) were mixed at 0 °C. MesN2BF4 (0.7 g, 3 mmol) was slowly added into the mixed solution. The reaction mixture was slowly warmed up to RT after stirring for 1h at 0 °C. Then, the mixed solution was stirred overnight. The crude product was extracted into CH2Cl2 and the volatiles were removed on the rotavap. The desired formazan C6F5NNC(C6F5)NNHMes was purified by recrystallization (CH2Cl2/MeOH) 1 to give 0.89 g of product (1.7 mmol, 57%). H NMR (400 MHz, CDCl3, 25 °C):  10.62 (s,

1H, NH), 6.97 (s, 2H, Mes m-CH), 2.39 (s, 6H, Mes o-CH3), 2.33 (s, 3H, Mes p-CH3) ppm. 13 C NMR (100 MHz, CDCl3, 25 °C):  145.3 (dm, J= 249 Hz), 145.5 (bs), 142.7, 141.9 (dtt, J= 256, 13, 5Hz), 140.8 (bs), 139.0 (dm, J= 254 Hz), 139.0(overlapped) 138.8 (dm, J= 253 Hz),137.7 (dm, J= 253 Hz), 137.3 (dm, J= 250 Hz), 134.8, 131.2, 120.5 (bs), 109.2 (td, J= 19,

Formazan Synthesis

19 4 Hz), 21.6 (Mes p-CH3), 20.9 (Mes o-CH3) ppm. F NMR (375 MHz, CDCl3, 25 °C):  -

138.1 (dd, 2F, J= 22, 6 Hz, C6F5 o-CF), -152.7 (t, 1F, J= 21 Hz, C6F5 p-CF), -154.9 (bs, 2F,

C6F5 o-CF), -162.3 - -162.6 (m, 4FC6F5 m-CF), -163.2 (bs, 1F, C6F5 p-CF) ppm. Anal. calcd for C22H12F10N4: C, 50.59; H, 2.32; N, 10.73. Found: C, 50.49; H, 2.25; N, 10.66.

([1i]B(C6F5)3). 1i (200 mg, 0.6 mmol) and tris(pentafluorophenyl)borane (309 mg, 0.6 mmol) were stirred in toluene (15 mL) at room temperature for few hours. After which all the solvent was pumped out and the crude product was recrystallized from DCM/hexane to give 390 mg 1 of red crystalline product (0.46 mmol, 77 %). H NMR (400 MHz, CDCl3, 25 °C):  12.17 (bs,

1H, NH), 6.98 (s, 2H, Mes m-CH), 2.40 (s, 6H, Mes o-CH3), 2.34 (s, 3H, Mes p-CH3)ppm. 13 C NMR (100 MHz, CDCl3, 25 °C):  148.1 (dm, J= 244 Hz), 140.8, 140.3 (dm, J= 253 Hz), 137.2 (dm, J= 248 Hz), 132.5 (overlapped), 130.9, 122.6, 115.4 (bs), 111.2, 21.2, 19.8 ppm. 19 F NMR (375 MHz, CDCl3, 25 °C):  -133.8 (dd, 2F, J= 22, 6 Hz, C6F5 o-CF), -156.6 (t, 1F,

J= 20 Hz, C6F5 p-CF), -163.7 (td, 2F, J= 22, 6 Hz, C6F5 m-CF) ppm.

Ph2-Ph-Ph2 1k. [(PhNNCH)2(para-C6H4)] (2g, 6.4 mmol) was dissolve in a solvent mixture containing pyridine (150 mL) and DMF (150 mL). A solution of diazonium salt made from stirring aniline (1.18g, 12.7 mmol), sodium nitrite (0.96 g, 13.8 mmol), water (20 mL), and hydrochloric acid (3.2 mL) for 30 min at 0 °C. The diazonium solution was added dropwise into the first solution at 0 °C. After stirring for few hours at RT the colour of solution turned to red and some deep red solid precipitated out from solution. Some water was added into reaction mixture to precipitate more solid from solution. The red solid was collected by filtration and washed with methanol, water and ether. The product was purified by recrystallization from DCM/MeOH at -30 °C, yield 497.3 mg (0.95 mmol, 15 %). 1H NMR 8 (400 MHz, THF-d , 25 °C):  15.26 (s, 2H, NH), 8.23 (s, 4H, C6H4), 7.82 (d, 8H, J = 8 Hz, Ph o-CH),7.48 (t, 8H, J = 7 Hz, Ph m-CH), 6.29 (t, 4H, J = 8 Hz, Ph p-CH) ppm. 13C NMR (100 8 MHz, THF-d , 25 °C):  149.3 (Ph i-C), 142.3 (NCN), 137.7 (C6H4 o-C), 130.4 (Ph m-C),

128.5 (C6H4 CH), 126.8 (Ph p-C), 119.9 (Ph o-C) ppm.

Ph2-mPh-Ph2 1l. The synthetic procedure of 1l is the same as 1k. [(PhNNCH)2(meta-C6H4)] (2g, 6.4 mmol), pyridine (150 mL), DMF (110 mL), aniline (1.18g, 12.7 mmol), sodium nitrite (0.96 g, 13.8 mmol), water (20 mL), and hydrochloric acid (3.2 mL) were used. After recrystallization from DCM/MeOH at -30 °C, 1.7 g (3.3 mmol, 52 %) of L12H2 can be 1 isolated. H NMR (400 MHz, CDCl3, 25 °C):  15.44 (s, 2H, NH), 9.04 (s, 1H, C6H4 CH),

8.15 (dd, 2H, J = 8, 2 Hz, C6H4 CH), 7.77 (d, 8H, J = 8 Hz, Ph o-CH), 7.53 (t, 1H, J = 8 Hz,

Chapter 2

13 C6H4 CH), 7.48 (t, 8H, J = 8 Hz, Ph m-CH), 7.31 (t, 4H, J = 7 Hz, Ph p-CH) ppm. C NMR

(100 MHz, CDCl3, 25 °C):  148.0 (Ph i-C), 141.0 (NCN), 137.5 (C6H4), 129.4 (Ph o-C),

128.6 (C6H4), 127.4 (Ph p-C), 124.9 (C6H4), 123.4 (C6H4), 118.8 (Ph p-C) ppm.

PhNNC(4-BrC6H4)NNPh 1m and PhNNCH(4-BrC6H4)CH(4-BrC6H4)NNPh 2. A flask was charged with PhNNC(4-BrC6H4)H (1.36 g, 5 mmol), CH2Cl2 (150 mL) and the aqueous solution (150 mL) containing sodium carbonate (2 g, 16.2 mmol) and tetrabutylammonium bromide (0.125 g, 0.37 mmol). A solution of diazonium salt made from stirring aniline (0.46 mL, 5 mmol), sodium nitrite (0.375 g, 5.45 mmol), water (5 mL), and hydrochloric acid (1.3 mL) for 30 min at 0 °C. The diazonium solution was added dropwise into the first solution at 0 °C. After stirring for 2 hours at RT the organic layer was collected and the water layer was extracted by CH2Cl2. The organic solution was conbined and concentrated. The product mixture of 1m and 2 can be isolated by recrystallization from CH2Cl2/MeOH at -30 °C to give total amount of 50.1 mg.

PhNNC(Py)Ph 3. [PhNNC(C5H4N)H] (1g, 5.1 mmol) was dissolved in a solvent mixture containing pyridine (8 mL), MeOH (4 mL) and acetic acid (1 mL). A solution of diazonium salt made from stirring aniline (0.47 g, 5.1 mmol), sodium nitrite (0.35 g, 5.1 mmol), water (6 mL), and hydrochloric acid (1.5 mL) for 30 min at 0 °C. The diazonium solution was added dropwise into the first solution at 0 °C. The reaction mixture was stirred at 60°C for 10 mins and then a 2M NaOH solution was added until pH = 9. After stirring over a weekend at RT, the crude product was collected by filtration and was recrystallizd from ether. After which 1 233 mg (0.85 mmol, 17 %) of red crystal was obtained. H NMR (400 MHz, CDCl3, 25 °C):  8.51 (d, 1H, J = 5 Hz, Py CH), 8.18 (d, 1H, J = 8 Hz, Py CH), 7.79 (s, 1H, NH), 7.71 (td, 1H, J = 8, 2 Hz, Py CH), 7.59 (t, 2H, J = 8 Hz, Ph m-CH), 7.51 (t, 1H, J = 7 Hz, Ph p-CH), 7.38 (d, 2H, J = 8 Hz, Ph o-CH), 7.27 (t, 2H, J = 8 Hz, Ph m-CH), 7.15 (ddd, 1H, J = 7, 5, 1 Hz, Py CH), 7.12 (d, 1H, J = 8 Hz, Ph o-CH), 6.89 (t, 1H, J = 7 Hz, Ph p-CH) ppm. 13C NMR (100

MHz, CDCl3, 25 °C):  156.5, 148.9, 144.1, 144.0, 135.9, 132.0, 129.5, 129.3, 129.2, 129.2, 122.1, 120.7, 120.6, 113.2 ppm.

Crystallographic data

Suitable crystals of 1j-B(C6F5)3, 1m, 2, and 3 were mounted on a cryo-loop in a drybox and transferred, using inert-atmosphere handling techniques, into the cold nitrogen stream of a Bruker D8 Venture diffractometer. The final unit cell was obtained from the xyz centroids of

9724 (1j-B(C6F5)3), 9915 (1m), 9950 (2) and 9906 (3) reflections after integration. Intensity data were corrected for Lorentz and polarisation effects, scale variation, for decay and

Formazan Synthesis

absorption: a multiscan absorption correction was applied, based on the intensities of symmetry-related reflections measured at different angular settings (SADABS).22 The structures were solved by direct methods using the program SHELXS.23 The hydrogen atoms were generated by geometrical considerations and constrained to idealised geometries and allowed to ride on their carrier atoms with an isotropic displacement parameter related to the equivalent displacement parameter of their carrier atoms. Structure refinement was performed with the program package SHELXL.23 Crystal data and details on data collection and refinement are presented in following table. Crystallographic data 1j-B(C6F5)3 1m 2 3 chem formula C38H23BF15N5 C19H15BrN4 C30Br2N4 C18H15N3 Mr 845.42 379.26 576.16 273.33 cryst syst monoclinic orthorhombic triclinic monoclinic color, habit orange, plate red, block colourless, plate yellow, block size (mm) 0.25 x 0.90 x 0.10 0.17 x 0.15 x 0.06 0.18 x 0.11 x 0.02 0.33 x 0.20 x 0.15 space group P21/n Pbca P-1 P21/c a (Å) 13.2170(9) 7.8048(3) 6.1295(3) 10.6972(12) b (Å) 11.3150(7) 18.9675(8) 9.3415(4) 8.6263(10) c (Å) 23.5257(16) 22.1335(10) 10.4730(5) 15.7962(18)  (°) 83.9733(16) β (°) 90.469(2) 77.0380(17) 106.374(4)  (°) 84.8992(15) V (Å3) 3518.2(4) 3276.6(2) 579.84(5) 1398.5(3) Z 4 8 1 4 -3 calc, g.cm 1.596 1.538 1.570 1.298 Radiation [Å] Mo Kα 0.71073 Mo Kα 0.71073 Mo Kα 0.71073 Mo Kα 0.71073 -1 µ(Cu K), mm -1 µ(Mo K), mm 0.151 2.516 3.516 0.079 F(000) 1704 1536 274 576 temp (K) 100(2) 100(2) 100(2) 100(2)  range (°) 2.93-26.77 2.83-28.33 2.84-27.12 2.72-28.44 data collected (h,k,l) -16:16; -14:14; - -10:9; -25:25; -29:29 -7:7; -11:11; - -14:14; -11:11; - 29:29 13:13 20:21 min, max transm 0.6929, 0.7454 0.6515, 0.7457 0.6290, 0.7455 0.7104, 0.7457 rflns collected 114934 92987 17139 57374 indpndt reflns 7482 4087 2547 3514 observed reflns Fo  5696 3385 2333 2981 2.0 σ (Fo) R(F) (%) 4.14 2.88 2.39 4.07 wR(F2) (%) 8.56 6.70 5.39 10.3 GooF 1.034 1.039 1.066 1.028 weighting a,b 0.0382, 2.4179 0.034, 2.6501 0.0235, 0.4354 0.0579, 0.5229 params refined 542 221 145 194 min, max resid dens -0.258, 0.314 -0.228, 0.489 -0.346, 0.533 -0.212, 0.352

Chapter 2

2.5 References

(1) (a) Grychtol, K.; Mennicke, W. In Ullmann's Encyclopedia of Industrial Chemistry; Wiley-VCH Verlag GmbH & Co. KGaA: 2000. (b) Bamberger, E. and Wheelwright, E. Ber. Dtsch. Chem. Ges., 1892, 25, 3201-3213. (c) Friese, P. Ber. Dtsch. Chem. Ges., 1875, 8, 1078-1080. (2) v. Pechmann, H. Ber. Dtsch. Chem. Ges., 1892, 25, 3175-3190. (3) (a) Nineham, A. W. Chem. Rev., 1955, 55, 355–483. (b) Tezcan, H. Spectrochimica Acta Part A, 2008, 69, 971–979. (c) Ibrahim, Y. Tetrahedron, 1997, 53, 8507–8512. (d) Gok, Y. and Senturk, H. B. Dyes and Pigments. 1991, 15, 279–287. (e) Abbas, A. A. and Elwahy, A. H. M. ARKIVOC. 2009, 2009, 65– 70. (f) Katritzky, A. R.; Belyakov, S. A.; Durst, H. D. Synthesis, 1995, 1995, 577-581. (g) Gilroy, J. B.; Otieno, P. O.; Ferguson, M. J.; McDonald, R.; Hicks, R. G. Inorg. Chem., 2008, 47, 1279–1286. (4) (a) Turkoglu, G.; Berber, H.; Kani, I. New J. Chem., 2015, 39, 2728–2740. (b) Ibrahim, Y. A.; Abbas, A. A.; Elwahy, A. H. M. J. Heterocyclic Chem., 2004, 41, 135–149. (5) Di Zhu; Budzelaar, P. H. M. Dalton Trans., 2013, 42,11343–11354. (6) Busch, M.; Schmidt, R. J. Prakt. Chem., 1931, 131, 182-192. (7) Neugebauer, F. A.; Trischmann, H. Liebigs Ann. Chem., 1967, 706, 107-111. (8) Pechmann, H. V. Ber. Dtsch. Chem. Ges., 1894, 27, 1679-1693. (9) Busch, M.; Schmidt, R. Ber. Dtsch. Chem. Ges. A/B, 1930, 63, 1950-1952. (10) Bamberger, E. Ber. Dtsch. Chem. Ges., 1892, 25, 3547-3555. (11) Hausser, I.; Jerchel, D.; Kuhn, R. Chem. Ber., 1949, 82, 515-527. (12) Hutton, A. T.; Irving, H. M. N. H.; Nassimbeni, L. R. Acta Cryst., 1980, 36, 2071-2076. (13) Otten, E.; Travieso-Puente, R.; CHANG, M.-C. Dalton Trans., 2014, 43, 18035-18041. (14) Gilroy, J. B.; Ferguson, M. J.; McDonald, R.; Patrick, B. O.; Hicks, R. G. Chem. Commun., 2007, 126- 128. (15) Doyle, M. P.; Bryker, W. J. J. Org. Chem.,1979, 44, 1572-1574. (16) Jacobsen, H.; Berke, H.; Döring, S.; Kehr, G.; Erker, G. Organometallics, 1999, 18, 1724-1735. (17) (a) Rojas, R. S.; Peoples, B. C.; Cabrera, A. R.; Valderrama, M.; Fröhlich, R.; Kehr, G.; Erker, G.; Wiegand, T.; Eckert, H. Organometallics, 2011, 30, 6372-6382. (b) Cabrera, A. R.; Schneider, Y.; Valderrama, M.; Fröhlich, R.; Kehr, G.; Erker, G.; Rojas, R. S. Organometallics, 2010, 29, 6104-6110. (18) Leen, V.; Yuan, P.; Wang, L.; Boens, N.; Dehaen, W. Org. Lett., 2012, 14, 6150-6153. (19) Mirífico, M. V.; Caram, J. A.; Vasini, E. J. Tetrahedron Letters, 2006, 47, 6919-6922. (20) Frolova, N. A.; Vatsadze, S. Z.; Vetokhina, N. Y.; Zavodnik, V. E.; Zyk, N. V. Mendeleev Commun., 2006, 16, 251–254. (21) (a)Kuhn, R.; Neugebauer, F. A.; Trischmann, H. Monatshefte für Chemie und verwandte Teile anderer Wissenschaften. 1966, 97, 525-553. (b)Barbon, S. M.; Price, J. T.; Yogarajah, U.; Gilroy, J. B. RSC Adv., 2015, 5, 56316-56324. (22) Bruker. APEX2 (v2012.4-3), SAINT (Version 8.18C) and SADABS (Version 2012/1). Bruker AXS Inc., Madison, Wisconsin, USA. 2012. (23) Sheldrick, G. Acta Crystallographica Section A, 2008, 64, 112.

Chapter 3

(Formazanate)Zinc Complexes

The synthesis and characterization of mono(formazanate)ZnMe complexes (LZnMe; 4) and bis(formazanate)Zn complexes (L2Zn; 5) are presented. Single crystal structure determinations and VT-NMR studies of 5 show that the formazanate ligands have flexible coordination chemistry in both solid- and solution-state due to its nitrogen-rich structure. The cyclic voltammetry studies reveal that compounds 5 have five accessible oxidation states (50/- 1/-2/-3/-4), in which the first two reduced products ([5]- and [5]-2) were synthesized and characterized (5a-/-2 and 5b-/-2). The redox potentials of the compounds 5 can be altered in a straightforward manner over a relative wide range (~ 500 mV) by changing the steric or the electronic properties of the formazanate framework. The characterization data of the reduced products (5a-/-2 and 5b-/-2) prove the redox-active nature of the formazanate ligand.

Parts of this chapter have been published:

M.-C. Chang, T. Dann, Dr. D. P. Day, Dr. M. Lutz, Dr. G. G. Wildgoose and Dr. E. Otten* “The Formazanate Ligand as an Electron Reservoir: Bis(Formazanate) Zinc Complexes Isolated in Three Redox States” Angew. Chem. Int. Ed., 2014, 53, 4118–4122.

Chapter 3 (Formazanate)Zinc Complexes 3.1 Introduction

Despite the structural similarities between β-diketiminate,1 1,3,5-triazapentadienyl2 and formazanate, the coordination chemistry of formazanate ligands has remained relatively little explored. Most of the reported formazanate metal complex focused on late transition metals or noble metals. For example, Lipunov and co-workers reported bis(formazanate) nickel(II) complexes which were used to catalyze ethylene oligomerization.3 Poddel'sky and co-workers reported a heteroleptic cobalt complex, which is coordinated by a formazanate ligand and an o-semiquinonate radical anion, having a singlet ground state due to the antiferromagnetic exchange between low-spin cobalt(II) (S = 1/2) center and the o-semiquinonate radical anion (S = 1/2).4 Hicks and co-workers reported some well-characterized examples of late transition metal formazanate complexes,5 and boron compounds with formazanate ligands were shown to possess unusual redox properties.6

Unlike other late transition metal complex, the literature reports of (formazanate)zinc complexes are very limited and often poorly described. However, there are some reported zinc complexes bearing β-diketiminate ligands in the literature (Chart 3.1). Coates and co-workers reported a method for the synthesis of poly(ester-block-carbonate)s through a one-step, one- pot procedure with a (β-diketiminate)ZnX catalyst (A).7 Schaper and co-workers reported heteroleptic bis(β-diketiminate)zinc complexes (B) as the secondary building blocks of a 2D copper-zinc coordination polymer.8 The hydroamination of alkynes catalyzed by (β- diketiminate)ZnMe complexes (C) or bis(β-diketiminate)zinc complex (D) was reported by Roesky, Blechert and co-workers.9

Chart 3.1

NC Ar Py Py R3 R3 Ar N N N N N N Et N N Et Zn Zn Zn N Zn Et N Et N N Ar O O R2 R1 R1 R2 Ar N N

A B CD In this chapter, we synthesized a series of mono- and bis(formazanate) zinc complexes (compounds 4 and 5) to understand the coordination chemistry and redox-active property of the formazanate ligands. The reason why we chose Zn2+ complex as our starting point is because of the redox-inert nature of the Zn2+ ion, which has a stable +2 oxidation state and d10

(Formazanate)Zinc Complexes electron configuration. The stable +2 oxidation state (and lack of further Zn-based redox reactions) makes that any redox-chemistry that is observed in these compounds is due to the formazanate ligands. In addition, the d10 electron configuration results in a singlet ground state of the Zn complex, which allows us to use NMR to characterize the products or to follow the reactions.

3.2 (Formazanate) Zinc Methyl complexes

A preliminary study of metal complexation of formazanate ligands was carried out by synthesizing the mono(formazanate)ZnMe complexes (4). The compounds 4 were synthesized by reacting free formazan with 1equivlent of dimethyl zinc (Scheme 3.1). The 1H NMR spectrum of 4a shows a singlet resonance of the ZnMe group, which is located at -0.18 ppm, in addition to those expected for the formazanate ligand with a 1:1 ratio (Figure 3.1). Single crystals suitable for X-ray crystallography of 4a were obtained by slow diffusion of hexane into a solution of 4a in toluene (Figure 3.2, metrical parameters in Table 3.1). Even though 4a can be isolated as a crystalline material, 4a itself is unstable in the solution state. At room temperature, 4a establishes a Schlenk equilibrium with bis(formazanate)Zn complex (5a) in toluene or C6D6 solution with the color changing from deep purple to deep blue. The chemistry of bis(formazanate)Zn complexes (5) will be discussed later in this chapter.

Scheme 3.1 Synthesis of LZnMe (4) and formation of L2Zn (5)

1 Figure 3.1 H NMR spectrum of 4a (C6D6, 400 MHz)

Chapter 3

3.2.1 Zinc Methyl Complex with Phenylene-Linked Diformazanate Ligand

In the case of free di-formazan 1k, the corresponding zinc methyl complex 4k can be synthesized in high yield (93%) in THF solution using an excess of dimethyl zinc. The complex 4l can also be synthesized with the same method. Both compounds feature a 1 diagnostic resonance in the H NMR spectrum at -0.35 (4k, in THF-d8) and -0.08 (4l, in C6D6) ppm for the Zn-Me moiety. The crystal structure and metrical parameters of 4k are shown in Figure 3.2 and Table 3.1, respectively. Unlike 4a, the crystal structure of 4k shows two distorted trigonal pyramidal zinc centers, in which the formazanate and methyl groups occupy the equatorial positions, and one THF molecule sits at the axial position. All the metrical parameters in the formazanate ligand of 4k are very close to those in 4a. The Zn-N bonds of 4k are slightly longer than the Zn-N bonds of 4a. The difference is due to the different coordination number of zinc centers (4 for 4k and 3 for 4a). The Zn-O(thf) bond lengths of 4k are 2.190(3) and 2.220(3) Å, which are in a range (2.19-2.33 Å) of reported (- diketiminate)ZnEt(THF) complexes.10 The trigonal pyramidal metal center is very rare for the four coordinated complexes, in particular for the bidentate ligand system. In literature, only few of the reported (-diketiminate)ZnRX (R=alkyl; X=THF or pyridine) complexes show trigonal pyramidal structure.10,11 The common structures of four coordinated complexes are tetrahedral or square planar. In nature, trigonal pyramidal iron centers play a significant role in nitrogen fixation. For example, the active site of the Iron-Molybdenum cofactor of Nitrogenase is an iron belt constructed by six trigonal pyramidal iron centers (Chart 3.2).12 The vacant site of the trigonal pyramidal iron centers is a potential binding site for nitrogen molecule. Several iron complexes bearing -diketiminate ligands were prepared to mimic the trigonal pyramidal active site, but most of them show tetrahedral structure instead of the trigonal pyramidal structure. Another strategy to construct trigonal pyramidal metal centers is using specially designed tridentate ligands in which one of the donor atoms can occupy the axial position. For example, a trigonal pyramidal Ru(II) complex was synthesized by Turculet and co-workers by using a bis(phosphino)silyl ligand (Chart 3.2).13

(Formazanate)Zinc Complexes

Figure 3.2 Molecular structures of 4a (left) and 4k(THF)2 (right) showing 50% probability ellipsoids. The THF atoms except for O and all hydrogen atoms are omitted for clarity.

o Table 3.1 Selected bond length (Å) and bond angles ( ) of 4a and 4k(THF)2 4a 4k(THF)2 N1-N2 1.299(3) N1-N2 1.306(4) N5-N6 1.310(4) N2-C7 1.346(2) N2-C7 1.343(4) N6-C12 1.349(4) C7-N3 1.337(3) C7-N3 1.349(4) C12-N7 1.348(4) N3-N4 1.310(3) N3-N4 1.308(4) N7-N8 1.304(4) N1-Zn1 2.003(2) N1-Zn1 2.018(3) N5-Zn2 2.011(3) N4-Zn1 1.989(2) N4-Zn1 2.018(3) N8-Zn2 2.016(3) Zn1-C21 1.956(3) Zn1-C34 1.971(4) Zn2-C19 1.950(5) N1-Zn1-N4 90.80(7) N1-Zn1-N4 88.6(1) N5-Zn2-N8 89.0(1) N1-Zn1-C21 129.95(9) N1-Zn1-C34 136.3(1) N5-Zn2-C19 134.4(1) N4-Zn1-C21 138.62(9) N4-Zn1-C34 129.7(1) N8-Zn2-C19 126.3(1) Zn1-(N1N2N3N4)a 0.102 Zn1-( N1N2N3N4)a 0.355 Zn2-(N5N6N7N8)a 0.227 Zn1-(N1N4C21)b Zn1-(N1N4C34)b 0.256 Zn2-(N5N8C19)b 0.355 Zn1-O1 2.220(3) Zn2-O2 2.190(3) C34-Zn1-O2 103.4(1) C19-Zn1-O2 106.0(1) N1-Zn1-O2 90.7(1) N1-Zn1-O2 95.1(1) N4-Zn1-O2 95.6(1) N4-Zn1-O2 97.4(1)  0.72  0.57 S(T) 3.14 S(T) 2.45 S(TP) 3.41 S(TP) 3.67 S(bT) 2.12 S(bT) 1.42 S(bTP) 1.19 S(bTP) 1.73 a: Distance between Zn and N-N-N-N plane ; b: Distance between Zn and N-N-C(Me) plane.

Chart 3.2 Iron-Molybdenum cofactor of Nitrogenase (left) and bis(phosphino)silyl Ru(II) (right)

PCy2

Me Si H Si Cy2P Ru X Cy2P PCy2

Chapter 3

3.2.2 Quantitative Description of Trigonal Pyramidal Zinc Center

In order to quantitatively describe the rare trigonal pyramidal zinc centers of 4k, two parameters were used: angle-based parameter  and continuous shape measures S(G).14

(i) Angle-based parameter  :

In four coordinated complexes,  is the normalized difference between the sum of the basal ligand-basal ligand angles and the sum of the basal ligand-axial ligand angles (Chart 3.3). The angle-based parameter  is expressed in eq 3.1. The eq 3.1 is suitable to describe the structures that lie near the interconversion path between a perfect tetrahedron ( = 0) and a perfect trigonal pyramid ( = 1). The limitations of this parameter are that it can not describe the deviation from the C3 symmetry of the structures such as the bidentate ligand. In addition, it is not reliable in some extreme cases: the square planar ( = 0) and sawhorse ( = 1) structures. The  of two Zn centers of 4k are 0.72 and 0.57, which indicate that the first zinc center is very close to the trigonal pyramidal structure and the second zinc center is sitting between the tetrahedral and trigonal pyramidal structure. It is worth pointing out that  = 0.72 is the highest value we were able to find in the literatures.

Σ Σ 3.1 90

Chart 3.3

(ii) Continuous shape measures S(G):

The continuous shape measure was proposed by Avnir and co-workers to quantitatively describe the deviation of a set of atoms from a given ideal polyhedral shape G.14 The given ideal polyhedral shape is superimposed on the metal complexes to minimize the expression of eq 3.2. The vectors are the vectors constructed by the coordinates of N atoms of the investigated structure. The vectors are the vectors constructed by the coordinates for the perfect polyhedron closest in size and orientation. The vector is the coordinate vector of the geometrical center of the investigated structure. The value of S(G) is between 0 and 100;

(Formazanate)Zinc Complexes

S(G) = 0 means that the investigated structure is exactly the same as the given ideal polyhedral shape.

∑ 100 3.2 ∑

The S(tetrahedral: T) and S(trigonal pyramidal: TP) of the two zinc centers of 4k are 3.14/2.45 and 3.41/3.67 (Table 3.1), respectively, which seems to suggest that the two zinc centers are close to tetrahedral structure. However, the formazanate ligand is a bidentate ligand, which reduces the bond angle of N-Zn-N to 90° and increase the two bond angles of N-Zn-C. The bond angle of perfect tetrahedral and perfect trigonal pyramidal are 109.5° and 120° (at equatorial positions), respectively. The constraint of the bond angles from the bidentate ligand makes that the S(T) is systematically smaller than S(TP). In order to solve this problem, the S(bT) and S(bTP), in which one of the bond angles of the perfect polyhedrons was set to 90°, were used. The S(bT) and S(bTP) of two zinc centers of 4k now are 2.12/1.42 and 1.19/1.73, respectively, which suggest that the first zinc center is more close to the trigonal pyramidal structure and the second zinc center is sitting in between of tetrahedral and trigonal pyramidal structure.

The results from two different parameters leads to the same conclusion: one of the zinc centers of 4k is very close to the trigonal pyramidal structure, and another zinc center of 4k is sitting between tetrahedral and trigonal pyramidal structure. The structural difference between the two zinc centers can also be observed by the out of plane distance of the zinc centers from the NNC plane (Zn1: 0.256 Å and Zn2: 0.355 Å ), which was defined by the two terminal nitrogen atoms of the formazanate ligand and the carbon atom of the methyl group.

The (formazanate)zinc methyl compounds 4 discussed in this section provide important insight in the coordination chemistry of formazanate ligands, but the steric hindrance of the ligands is such that their stability in solution is rather limited: facile ligand exchange takes place in case of compound 4a, presumably because a bimolecular (Schlenk-type) exchange reaction is accessible that leads to bis(formazanate) zinc compounds 5. In the subsequent part of this chapter we will focus on a more detailed characterization of these L2Zn complexes.

Chapter 3

3.3 Bis(Formazanate)Zinc Complexes

3.3.1 Synthesis and Coordination Chemistry of Bis(Formazanate)Zinc Complexes

Treatment of Me2Zn with 2 equivalents of formazan 1a results in the complete conversion of the starting materials. The 1H NMR spectrum of the product is consistent with its formulation as the homoleptic bis(formazanate) complex [1a]2Zn (5a), with diagnostic resonances for the p-tolyl and Ph moieties in a 1:2 ratio. Compound 5a was isolated in 76% yield as dark violet crystals upon crystallization from toluene/hexane. The ligands 1b-1d reacted similarly, and good yields of the corresponding bis(formazanate)zinc complexes (5b-5d) were isolated as intensely colored crystalline material (Scheme 3.3).

Scheme 3.3 Synthesis of bis(formazanate) zinc compounds 5a-5d. Single crystal X-ray diffraction studies on compounds 5a-5d showed in all cases a central Zn atom surrounded by two formazanate ligands in a tetrahedral coordination environment (Figure 3.4, metrical parameters in Table 3.2). The formazanate bite angles differ little from the symmetric derivatives (R1 = R5; N-Zn-N av. 92.2°) while in the case of 5c the N-Zn-N angle is somewhat smaller at 87.98(4)°. The dihedral angle between the planes that contain the two N-Zn-N fragments deviates from the ideal tetrahedral value of 90° with values between 83.64(11)°-86.43(7)° for the symmetrical derivatives, and only 71.26(6)° for asymmetric complex 5c. Full delocalization within the formazanates in 5a-5d is indicated by the equivalent N-N and C-N bond lengths in the backbone of each ligand. Upon increasing the steric hindrance of the N-R substituents, the NNCNN backbone becomes increasingly puckered and the Zn atom is displaced out of the ligand plane, similar to what is observed in related β-diketiminate (nacnac) Zn complexes, where an envelope conformation of the (nacnac)Zn 6-membered ring is observed.15 The N-Ph rings in 5a-5c are found to be approximately coplanar with the ligand backbone (NNCNN / Ph dihedral angles < 20°), a situation which maximizes conjugation. However, steric interactions in 5c and 5d between the

2,6-Me2 groups on the N-Mes substituents and the ligand backbone cause these groups to

(Formazanate)Zinc Complexes

have much larger dihedral angles (65.80° in 5c and av. 74.64° in 5d). In the solid state structure of complex 5c the N-Mes rings are oriented nearly parallel (interplanar angle = 3.21° and distance between centroids = 3.660 Å) and stack in an off-center fashion as is usually observed for electron-rich aromatic rings.16

Table 3.2 Selected bond length (Å) and bond angles (o) of compounds 5a – 5d 5a 5bb 5c 5d (x = 1) (x = 2) (x = 1) (x = 2) (x = none) (x = 1) (x = 2) Zn(1)– N(1x) 1.9849(11) 1.9829(12) 1.9824(17) 1.9902(17) 2.0179(10) 1.9874(16) 2.0234(16) Zn(1)– N(4x) 1.9953(12) 2.0012(12) 1.9822(18) 1.9769(18) 1.9882(10) 2.0207(16) 1.9873(16) N(1x)–N(2x) 1.3067(16) 1.3120(16) 1.310(2) 1.307(2) 1.3026(14) 1.310(2) 1.303(2) N(3x)–N(4x) 1.3048(16) 1.3024(16) 1.307(2) 1.309(2) 1.3129(14) 1.304(2) 1.310(2) a N(2x)–Cbackbone 1.3462(18) 1.3464(18) 1.340(3) 1.353(3) 1.3504(15) 1.343(2) 1.357(3) a N(3x)–Cbackbone 1.3517(18) 1.3499(18) 1.343(3) 1.332(3) 1.3512(16) 1.353(2) 1.345(3) N(1x)–Zn(1)–N(4x) 92.21(5) 90.38(5) 93.18(7) 92.94(7) 87.98(4) 88.76(7) 88.03(7) (N-Zn-N)/(N-Zn-N)b 86.43(7) 83.64(11) 71.26(6) 85.07(9) 19.78 (Ph) av. Ar / NNCNN 15.02 10.24 74.6 65.80 (Mes) a b Cbackbone is the central C atom of the ligand backbone the angle between the planes defined by the central Zn atom and the coordinated N atoms. c Only one of two independent molecules is discussed.

Figure 3.4. Top: Molecular structure of 5a showing 50% probability ellipsoids (left). View along both NNCNN planes in 5a (right). Bottom: Molecular structures of 5c (left) and 5d (right) showing 50% probability ellipsoids (all hydrogen atoms and solvent molecules omitted for clarity). The NMR spectroscopic features for the symmetrical complexes 5a, 5b, and 5d are straightforward and consistent with equivalent N-Ar groups in both 1H and 13C spectra. The

Chapter 3 room temperature 1H NMR spectrum of 5c shows the CH and o-Me groups of the mesityl ring to be equivalent, which suggests that rotation around the N-Mes bond is facile. However, the resonances due to the o-Me groups are somewhat broadened, and cooling a toluene-d8 solution to – 25 °C results in decoalescence of these signals to afford an NMR spectrum that is consistent with the solid-state structure of 5c.

The observed structural deformations in 5c that result from the (sterically) asymmetric formazanate ligand prompted the synthesis of analogues with two electronically disparate N- 1 1 1 Ar substituents. Treatment of R NNC(p-tolyl)NNH(C6F5) (R = Ph, 1e; R = Mes, 1f) with

Me2Zn in a 2:1 molar ratio results in the formation of the desired complexes [1e]2Zn (5e) and

[1f]2Zn (5f) (Scheme 3.4). We were unable to crystallize compound 5e so that purification was difficult. Nevertheless, the reaction is clean and careful control of the reaction stoichiometry gave 5e in >95% purity as assessed by 1H NMR spectroscopy. The analogous mesityl-substituted compound 5f was isolated in 71% yield as dark violet crystalline material.

Scheme 3.4 Synthesis of compounds 5e and 5f

A single-crystal X-ray structure determination showed compound 5f to be an unusual bis(formazanate)zinc complex in which two distinct ligand binding modes are observed (Figure 3.5, metrical parameters in Table 3.3). One of the formazanates is bound in the ‘normal’ fashion to generate a 6-membered chelate ring, whereas the other ligand is

(Formazanate)Zinc Complexes coordinated through both the terminal N(11) and an internal N(31) atom resulting in a 5- membered chelate. As expected, the ligand bite angles differ considerably at 78.93(9) and 88.90(9)° for the 5 and 6-membered chelate ring, respectively. Balt and co-workers described the formation of an intermediate en route to Cu and Ni complexes with the formazan (2-OH- Ph)NNC(Ph)NNHPh which they tentatively identified as isomers with a 5-membered formazanate chelate ring,17 and a crystal structure was reported for the Ni-derivative.18 Furthermore, large structural diversity in alkali metal formazanate complexes was recently reported by us that includes 4- and 5-membered chelate rings.19 The structural relationship observed between the ubiquitous β-diketiminates and the formazanate ligands for compounds 5a-5d no longer holds in complex 5f: the presence of the additional nitrogens in the backbone that can function as donor atoms favor formation of a 5-membered chelate that is not accessible for β-diketiminates.1 The energetic balance, however, is quite subtle since 5- and 6- membered rings are observed simultaneously in 5f. A closer inspection of the N-N and N-C bond lengths within the formazanate frameworks (Table 3.3) shows that both ligands are best described with a localized negative charge at the C6F5-substituted nitrogen atom: the observed bond length alternation contrasts the delocalization in 5a-5d: the more localized nature is likely the result of the strong electron-withdrawing effect of the C6F5 ring. Whereas 5c shows off-center stacking of two mesityl rings, in compound 5f there is a face-centered interaction of an electron-poor C6F5 ring with an electron-rich mesityl group (interplanar angle = 5.31°, centroid-centroid distance = 3.375 Å), indicative of an aromatic donor-acceptor interaction that is commonly observed for combinations of electron-rich/poor rings.16

Figure 3.5 Molecular structure of 5f showing 50% probability ellipsoids. The hydrogen atoms are omitted for clarity.

Chapter 3

Characterization of the structure of 5e and 5f by solution NMR spectroscopy reveals remarkable differences. Compound 5e shows one set of resonances for the NPh and N(C6F5) groups, consistent with a coordination geometry around the Zn center similar to that found in 5a-5d. For 5f, a larger number of resonances than expected was observed in the 1H NMR spectrum, independent of the batch of material and despite the fact that the isolated crystals seemed homogeneous. Invariably, three distinct sets of formazanate resonances were observed in both the 1H and 19F NMR spectrum at room temperature in an approximate 1:0.7:0.7 ratio (Figure 3.6).

2,00

1,50 y = 0,0143x ‐ 3,3631 R² = 0,9754

1,00 y = 0,0135x ‐ 3,2904 R² = 0,9805 G(KJ/mol)

 0,50

0,00 200 250 300 350 400

T (K) ‐0,50

19 19 Figure 3.6 F NMR of compound 5f (C6D6, 375 MHz, up), F EXSY of compound 3.6 1 19 (Toluene-d8, 562.5 MHz, bottom left) and VT NMR data of compound 5f (x: H NMR, o: F NMR, bottom right)

The bis(formazanate) zinc complex (5g) with a C6F5-substituent at the central carbon atom of the ligand framework was also synthesized (Scheme 3.5). While the compound could not be crystallized, it is likely that 5g contains 2 six-membered ring chelates: the steric properties of the ligand are very similar to that in 5c and both formazanates are equivalent in the 1H NMR.

(Formazanate)Zinc Complexes

Bis(formazanate) zinc complexes with two different formazanate ligands are accessible via the compound 4a, which is obtained from a 1:1 reaction of 1a with Me2Zn. Stirring 4a with an equivalent of MesNNC(CN)NNHMes (1j) results in the formation of the mixed complex [1a][1j]Zn (5aj, Scheme 3.5). In the solid state structure of 5aj (Figure 3.7, metrical parameters in Table 3.3), two different formazanate binding motifs are found that are similar to those in 5f: the formazanate ligands 1a and 1j form 6- and 5-membered chelate rings, respectively, that have quite different bite angles (92.63(6) and 79.85(5)°). The mesityl- substituted ligand 1j adopts a 5-membered chelate ring to minimize steric interactions. The N- N and N-C bond lengths within the 5-membered chelate are indicative of localized single and 1 double bonds. The H NMR spectrum of a crystalline sample of 5aj in C6D6 solution shows the presence of two isomers in approximately 1:0.2 ratio (Figure 3.8). For the major isomer (5aj-i), the ligand [1a]- appears as a symmetric set of resonances with equivalent N-Ph groups while ligand [1j]- is asymmetrically bound as indicated by inequivalent N-Mes moieties. The minor isomer (5aj-ii) is fully symmetric. Evidence for exchange between the isomers comes from the observation of exchange crosspeaks in the EXSY spectrum between the two species at 100 °C in toluene-d8 solution (no exchange is observed at room temperature). The intensity of the signals due to the minor isomer change only slightly when the temperature is increased from 0 °C (major: minor = 1:0.17) to 100 °C (1:0.29). Evaluation of the equilibrium constant at 4 different temperatures between 0 and 100 °C allows the enthalpy and entropy terms to be determined as ΔH = +4.7 kJ·mol-1 and ΔS = +2.5 J·mol-1·K-1. In contrast to complex 5f, two 6-membered formazanate chelate rings are enthalpically disfavored for 5aj, while there is hardly any preference based on entropy.

Scheme 3.5 Synthesis of compounds 5g and 5aj

Chapter 3

Figure 3.7. Molecular structure of 5aj showing 50% probability ellipsoids. The hydrogen atoms are omitted for clarity.

Table 3.3 Selected bond length (Å) and bond angles (o) of 5f, 5aj. 5f 5aj (x = 1) (x = 2) (x = 1) (x = 2) Zn(1) – N(1x) 1.988(2) 1.998(2) 1.9877(13) 1.9550(14) Zn(1) – N(3x) 2.094(2) 2.0890(15) Zn(1) – N(4x) 2.001(2) 1.9711(14) N(1x) – N(2x) 1.322(4) 1.321(3) 1.310(2) 1.3035(19) N(3x) – N(4x) 1.284(3) 1.287(3) 1.277(2) 1.3042(19) a N(2x) – Cbackbone 1.323(4) 1.324(3) 1.338(2) 1.345(2) a N(3x) – Cbackbone 1.395(4) 1.379(3) 1.378(2) 1.351(2) N(1x) – Zn(1) – N(3x) 78.92(10) 79.85(5) N(1x) – Zn(1) – N(4x) 88.90(9) 92.63(6) (N-Zn-N) / (N-Zn-N) 59.44(14) 78.43(8) a b Cbackbone is the central C atom of the ligand backbone. one of the two independent molecules is discussed.

(Formazanate)Zinc Complexes

1 Figure 3.8 H NMR spectra of compound 5aj (C6D6, 400 MHz).

3.3.2 UV-Vis Spectroscopy of Bis(Formazanate)Zinc Complexes

The optical properties of the bis(formazanate)zinc complexes (5) were measured by UV-Vis absorption spectroscopy in THF solution. All compounds show intense absorption bands with extinction coefficients between 17000-50000 L mol−1 cm−1 in the visible range of the spectrum (between 450 and 600 nm), which are due to π-π* transitions within the formazanate framework (Figure 3.9). The symmetrical derivatives absorb most strongly, and introducing an electron-donating tBu group instead of a p-tolyl moiety results in a blue-shift in the absorption maximum (578 and 536 nm for 5a and 5b, respectively). Upon substitution of NPh for NMes groups, λmax is progressively blue-shifted from 578 (5a) to 520 in 5c to 465 nm in 5d. This likely results from the perpendicular orientation of the NMes rings which prevents conjugation between the formazanate backbone and the NAr substituent, thereby limiting the length of the π-conjugated system. A similar trend is observed for the compounds 5e and 5f, but this comparison is complicated by the presence of an isomer mixture for 5f. Complex 5g, with a C6F6-substituent on the central C-atom of the ligand, has the lowest absorption wavelength of the series with λmax = 463 nm.

Chapter 3

Figure 3.9. Absorption spectra for compounds 5a-5g and 5aj

3.3.3 Cyclic Voltammetry of Bis(Formazanate)Zinc Complexes

In order to establish the detail redox-active nature of the formazanate ligands in compounds 5a and 5b, we ran cyclic voltammetry (CV) experiments in THF solution with 20 [Bu4N][B(C6F5)4] electrolyte. Upon scanning in a reductive direction, the CV shows two quasi-reversible, single-electron redox processes, the labelled system I/I’ and II/II’ (Figure 3.10). These correspond to the reversible formation of the radical anions of 5a and 5b (5a– or 5b–; system I/I’) and the corresponding dianions (5a2– or 5b2–; system II/II’) respectively. If the scan direction is reversed after the reductive peak I but before peak II, then the reoxidation of the radical anion (5a– or 5b–) is once again observed as peak I’ indicating that each redox process is sequential and independent. When the scan rate is varied between 100 and 1000 mVs-1, all processes exhibited a linear relationship between peak current and the square root of the voltage scan rate, indicative of diffusion-controlled redox processes. Excellent fits between experiment and digital simulation of the cyclic voltammetry of 5a and 5b (Figure 3.10) yielded optimized values of formal potentials, E0, and electron transfer rate constants, k0 listed in Table 3.4. Replacing the inductive electron donating tert–butyl group with the electron withdrawing p–tolyl group on the formazanate ligands has the expected effect on the reduction potentials (Table 3.4). Interestingly, the one-electron reduction of both 0 bis(formazanate)zinc complexes 5a and 5b occurs at more negative potentials (E I/I’ = –1.31 V vs. Fc0/+, 5a; –1.57 V vs. Fc0/+, 5b) than Hicks and co-workers have reported for boron mono(formazanate) compounds (E0 ~ -0.9 V vs. Fc/Fc+).6 This likely reflects the different Lewis acidity of the boron and zinc centers together with a different degree of covalency in

(Formazanate)Zinc Complexes the metal-ligand bonding. Cyclic voltammetric characterization indicates that both the singly– reduced radical anion and the doubly–reduced dianionic states of 5a and 5b are synthetically accessible.

Figure 3.10 Cyclic voltammograms of a 2.5 mM solution of 5a in THF (0.1 M -1 [Bu4N][B(C6F5)4]) recorded at 100, 200, 400, 600, 800, and 1000 mVs . Solid lines = experimental data; open circles = simulated data.

Table 3.4 Optimised electron transfer parameters determined from digital simulation of experimental voltammetry 5a E0 vs Fc0/+ / V k0 / 10-2 cms-1 System I/I’ –1.31± 0.01 1.25 ± 0.05 System II/II’ –1.55 ± 0.01 0.90 ± 0.05 5b System I/I’ –1.57 ± 0.01 1.30 ± 0.05 System II/II’ –1.85 ± 0.02 0.75 ± 0.05

For practical reason, the CV data of all bis(formazanate) zinc complexes were recorded in

THF with [Bu4N][PF6] electrolyte (Table 3.5 and Figure 3.11). By comparison the CV data of

5a and 5b in THF with two different electrolyte ([Bu4N][B(C5F5)4] and [Bu4N][PF6]), it shows that the influence from different electrolyte is very small. In all cases, quasi-reversible redox-processes were observed with two independent, sequential reductions taking place to transform 5 to [5]– and subsequently to [5]-2 (Figure 3.11). For all compounds, the difference between first and second reduction potential range between 0.29 V for 5a and 0.48 V for 5f. As expected, changing the electron-withdrawing p-tolyl substituent on the C atom of the formazanate backbone (5a: -1.38/-1.68 V) for an electron-donating tBu group (5c: -1.52/-1.84

Chapter 3

V) shifts the redox potentials to more negative values by ~ 0.15 V. A similar effect is observed for the N-substituents: a donating mesityl group shifts the reduction potential to more negative values in comparison to phenyl (for 5c: -1.60 and -1.99 V vs. Fc0/+). Two N- Mes groups, as in 5d, show redox-chemistry that occurs at even more negative potential (- 1.86 and -2.30 V vs. Fc0/+). Conversely, electron-withdrawing N-R groups shift the observed redox-potentials in the positive direction, and a similar effect is observed for a C-C6F5 group such as in 5g. Although the substituent on the central carbon atom of the formazanate ligand does not contribute to the redox-active molecular orbital,21 inductive effects are apparently equally important in modulating the redox-potentials of these compounds. The cyclic voltammograms indicate that for all compounds studied here, two separate redox events occur corresponding to sequential ligand-based reductions from 5 to the radical anion [5]– and the dianion [5]-2, all of which are quite stable based on the quasi-reversible nature of the CV peaks. For some compounds (most prominently in 5d), features in addition to that of simple reversible electron-transfer were observed. We have not investigated this in detail, but we tentatively assign it to the occurrence of a subsequent chemical step that could be isomerization from 6- to 5-membered chelate ring(s).

Table 3.5 Electrochemical parameters for compounds 5a E0 vs. Fc0/+ [V] 1 5 3 0/-1 -1/-2 R/R R L2Zn L2Zn Δ [V] (I/I') (II/II') 5a Ph2 p-tol -1.39 -1.68 0.29 t 5b Ph2 Bu -1.52 -1.84 0.32 5c Ph/Mes p-tol -1.60 -1.99 0.39 5d Mes2 p-tol -1.86 -2.30 0.44 5e Ph/C6F5 p-tol -1.17 -1.57 0.40 5f Mes/C6F5 p-tol -1.34 -1.82 0.48 5g Ph/Mes C6F5 -1.41 -1.87 0.46 b 5g Ph/Mes C6F5 -1.41 -1.81 0.40 Mes CN 5aj 2 -1.33 -1.82 0.49 Ph2 p-tol a 1.5 mM solution of zinc complex in THF, 0.1 M [Bu4N][PF6] electrolyte, scan rate 100 mV·s-1. b in dichloroethane solution.

(Formazanate)Zinc Complexes

Figure 3.11 Cyclic voltammograms of compounds 5a-5g and 5aj (ca. 1.5 mM solution of -1 zinc complex in THF, 0.1 M [Bu4N][PF6] electrolyte, scan rate 100 mV·s )

Figure 3.12 Cyclic voltammograms of compounds 5a (line) and 5b (dotted line) showing the presence of reduction waves corresponding to 2-electron reduction of a single formazanate ligand (L3-, reductions III and IV). By scanning the CV towards more negative potentials for compound 5a, two additional redox processes were found (Figure 3.12). For compound 5a, two additional redox-events were recorded at ca. -2.55 and -2.95 V vs. Fc0/+ to give a total of 5 accessible oxidation states for this compound (5a0/-1/-2/-3/-4). For the related compound 5b, only one additional reduction wave was observed between -2.5 and -3.5 V vs. Fc0/+ (peak potential -2.84 V), corresponding to the 5b-2/-3 couple and this reduction is much less reversible on the basis of the voltammetric

Chapter 3 response. Nevertheless, these data indicate that 2-electron reduction of a single formazanate ligand to give L3--type structures could be general.

3.4 Chemical Reduction of Bis(Formazanate)Zinc complexes

3.4.1 Synthesis and Characterization of 1-Electron Reduction Products

In accordance with the CV data, the chemical reduction of neutral bis(formazanate) complexes 5a– and 5b – could be accomplished by treatment with 1.0 equivalent of Na(Hg) – in THF. The resulting radical species [(PhNNC(R)NNPh)2Zn][Na(THF)3] (R = p-tolyl, 5a ; R = tBu, 5b –) could be isolated as crystalline material by slow diffusion of hexane into the THF solution (Scheme 3.6). Compounds 5a– and 5b – are NMR silent but show broad EPR signals (g-value ~ 2) both in THF and the solid state (298 and 77K) devoid of observable hyperfine coupling (Figure 3.13).

Ph Ph N N N N R Zn R N N N N Ph Ph 5a/5b 2 Na(Hg) 1 equiv Na(Hg) THF THF

2 Ph Ph Ph Ph N N 1 equiv N N N N R Zn R Na(Hg) N N N N R Zn R N N THF N N N N Ph Ph Ph Ph (THF)3Na Na(THF)3 Na(THF)3 - - 5a-2/5b-2 5a /5b Scheme 3.6 Synthesis of monoanionic and dianionic bis(formazanate)zinc complexes

Figure 3.13 EPR of 5a– (left) and 5b – (right)

(Formazanate)Zinc Complexes

The crystallographically determined structures (Figures 3.14, metrical parameters in Table 3.6) – + show that the tetrahedral [5] radical anion interacts with a Na(THF)3 cation through one (5a–) or two (5b–) nitrogen atoms of a formazanate ligand. A closer inspection of the metrical parameters within the formazanate backbone reveals that there are two distinctly different ligands in the [5]– fragment. One of the formazanates is very similar to those in the neutral precursors 5a and 5b (av. Zn-N: 2.014 Å; N-N: 1.304 Å), while the formazanate that + binds the Na(THF)3 has shortened Zn-N (av 1.957 Å) and elongated N-N bond lengths (av 1.363 Å). Thus, the [5]– anion is best described as a Zn2+ center coordinated by a ‘normal’ monoanionic formazanate (L-) and a reduced dianionic ligand (L2-). The Zn-coordinated L2- fragment can be considered an inorganic analogue of a verdazyl radical.22 The observation of two distinct ligand redox states is likely related to electrostatic interactions with the cation, which localizes the additional negative charge. A similar situation is observed in related t bis(ligand) complexes: in the case of neutral radical species [PhB(μ-N Bu)2]2M (M = Al, 23 24 Ga) and (β-diketiminate)2Al the unpaired electron is fully delocalized over the spirocyclic t - structure, while for the radical anions [PhB(μ-N Bu)2]2M (M = Mg, Zn) localized spin density is observed due to interaction with the cation.25

Table 3.6 Selected bond lengths (Å) of 5a– and 5b –. 5a– 5b –

(x = 1) (x = 2) (x = 1) (x = 2) Zn(1) – N(1x) 2.0226(15) 1.9542(15) 1.9857(16) 1.9447(16) Zn(1) – N(4x) 2.0272(15) 1.9625(15) 2.0207(16) 1.9657(16) N(1x) – N(2x) 1.299(2) 1.357(2) 1.313(2) 1.360(2) N(3x) – N(4x) 1.308(2) 1.364(2) 1.295(2) 1.370(2)

Figure 3.14 Molecular structure of 5a– showing 50% probability ellipsoids. The hydrogen atoms are omitted for clarity.

Chapter 3

3.4.2 Synthesis and Characterization of 2-Electron Reduction Products

As suggested by the CV measurements, compounds 5a– and 5b– react with an additional equivalent of Na amalgam to give the dianionic tetrahedral Zn complexes –2 t –2 –2 [(PhNNC(R)NNPh)2Zn][Na(THF)3]2 (R = p-tolyl, 5a ; R = Bu, 5b ), of which 5b was crystallographically characterized (Figure 3.15, metrical parameters in Table 3.7). It contains two sodium cations that both interact with two N-atoms of a different formazanate ligand. The presence of an additional electron in both ligands (L2-) is evidenced by the similar bond lengths observed (Table 3.7). Specifically, all N-N bonds in 5b–2 are elongated in comparison to those in the neutral precursor while they differ little from the L2- fragment in compounds 5a– and 5b–.

Figure 3.15 Molecular structure of [5b]Na2(THF)5 showing 50% probability ellipsoids. The carbon atoms of the THF moieties and all hydrogen atoms are omitted for clarity.

Table 3.7 Selected bond lengths (Å) of [5b]Na2(THF)5. (x = 1) (x = 2) Zn(1) – N(1x) 1.9793(14) 1.9696(14) Zn(1) – N(4x) 1.9839(14) 1.9952(15) N(1x) – N(2x) 1.359(2) 1.355(2) N(3x) – N(4x) 1.376(2) 1.378(2)

26 EPR spectra of diradicals [5a]Na2 and [5b]Na2 in frozen THF solution (77K) are very similar and show features indicative of randomly oriented triplets (g = 2.0028) with characteristic half-field (Δms = 2) signals (Figure 3.16). The zero-field splitting parameters D

(Formazanate)Zinc Complexes

-3 -3 -1 = 11.6 × 10 and 11.5 × 10 cm from the EPR spectra for [5a]Na2 and [5b]Na2, respectively, t are somewhat smaller than those in the neutral triplet diazabutadiene complex [ Bu2DAB]2Zn (23.1 × 10-3 cm-1),27 and comparable to D-values found for purely organic phenylene-linked bis(radical) compounds (radical = semiquinone;28 verdazyl21a,29), which range between ca. 4- 10 × 10-3 cm-1. It should be noted that although metal complexes with coordinated verdazyl 30 radicals have been prepared, compounds [5a]Na2 and [5b]Na2 present the first examples of diradical ‘metallaverdazyl’ compounds.

Figure 3.16 EPR spectra (frozen THF solution) of [5a]Na2 and [5b]Na2 (asterisk denotes a doublet impurity). Inset: half-field region. 3.4.3 UV-Vis Spectroscopy of Reduced Products

UV-Vis spectroscopy of neutral, anionic and dianionic bis(formazanate) zinc compounds (Figure 3.17) provides additional evidence for ligand-based reduction and formation of verdazyl-type (L2-) ligands. The neutral compounds 5a and 5b show single broad absorptions in the visible range at 578 and 536 nm, respectively. The 1-electron reduction compounds 5a– and 5b– have both formazanate (L-) and one-electron reduced, verdazyl-type (L2-) ligands. As a consequence, they feature new absorption bands at longer (5a–/5b–: 769/755nm) and shorter wavelengths (5a–/5b–: 508/462 nm) due to L2-; the position of these bands is similar to that observed in organic (Kuhn-type) triarylverdazyls.31 In addition, a weakened and bathochromically shifted absorption is observed which is attributed to the L- fragment in 5a–and 5b–. For compounds 5a–2 and 5b–2, the intensity of the low and high energy absorptions due to the L2- fragment is increased relative to the singly reduced species 5a–and 5b–. The most prominent absorptions in the visible range are at 510/798 (5a–2) and

Chapter 3

436/755 nm (5b–2) in agreement with the presence of only reduced formazanate (verdazyl- type) ligands.

Figure 3.17 UV-Vis of 5a, [5a]Na, [5a]Na2 (left), and 5b, [5b]Na, [5b]Na2 (right)

3.4.4 DFT Calculations

In recent years, more and more experimental chemists start to incorporate theoretical calculations in their research due to the easy access of computational resources, such as software and high-performance computing clusters. The density functional theory (DFT), which was developed by Walter Kohn in 1960’s,32 is one of the most widely used methods. DFT is a very powerful tool to predict a great variety of molecular properties33: molecular structure, vibrational frequency, electronic structure and magnetic properties. DFT is an approximation to solve the Schrodinger equation for the many-body system by using an electron density to simulate the total electron properties. Using DFT approximation, the number of spatial coordinates of an N-electrons system can be reduced from 3N to 3 resulting in a reduction of computational cost.

3.4.4.1 DFT Calculations of 1-Electron Reduction Compound (5–)

In order to understand the electronic structure of the complexes described above, DFT calculations were used. The crystallographically determined bond lengths and angles are reproduced accurately by (unrestricted) B3LYP/6-31G(d) calculations using Gaussian09 starting from the X-ray coordinates. However, geometry optimization of the ‘free’ radical anions 5a–and 5b– at the UB3LYP/6-31G(d) level of theory resulted in structures in which – the SOMO is delocalized over both ligands. For example, in 5a calc the diagnostic N-N bond lengths are all equivalent at ~ 1.322 Å, in between the short (L-: av 1.304 Å) and long N-N

(Formazanate)Zinc Complexes

2- + bonds (L : av 1.361 Å) observed experimentally. When the countercation [Na(THF)3] that is present in the crystal structure determination is included in the computations, the unpaired – electron is localized (see Figure 3.18 for 5a calc). This is in agreement with the experimental data and suggests that electrostatic effects are responsible for this localization. The calculated 14 – hyperfine interactions with the N nuclei are small in 5a calc (< 2.1 G), which likely accounts for the broad, featureless EPR signals observed experimentally.

– Figure 3.18 Top: SOMO (left) and spin density plot (right) for 5a calc. Bottom: Two ligand- centered SOMOs for the BS(1,1) solution (truncated structures, left) and spin density plot –2 (right) for 5a calc. 3.4.4.2 DFT Calculations of 2-Electron Reduction Compound (5–2)

For the diradicals 5a–2 and 5b–2, geometry optimizations of the 5–2 fragment in the absence of countercations converges at structures that have two (virtually) identical L-2 ligands with elongated N-N bond lengths of ca. 1.346 Å, which is somewhat shorter than those observed experimentally for 5b–2 (av. 1.367 Å). DFT calculations of singlet and triplet states in compounds 5a–2 and 5b–2 suggest that the triplet is favoured by ~ 14-16 kcal·mol-1 for both UB3LYP and UM06 calculations, but a broken-symmetry (BS)33a singlet diradical solution was found to have approximately equal energy as the triplet. For this BS(1,1) solution (Figure 3.18), the two ligand-based unpaired electron spins are antiferromagnetically coupled with -1 Jcalcd = - 7.9 cm to give a singlet diradical ground state. The calculated spin density in –2 diradical 5a calc indicates that the unpaired electrons are located at the nitrogen atoms of the ligands, with some contribution of the aromatic substituents of the ligand. To verify

Chapter 3 experimentally the ground state of 5a–2, preliminary EPR studies were carried out in the temperature range of 6-60K (THF glass) to determine the temperature dependence of the EPR signal intensity (I). A plot of I×T vs. T shows that I×T decreases upon lowering the temperature (Figure 3.19). This behavior is indicative of a singlet diradical ground state,26 thus corroborating our computational results. Although singlet biradical species like 5a–2 and 5b–2 are rare, Roesky and co-workers recently reported a dicarbene zinc compound for which the singlet diradical was calculated to be lower in energy than the triplet by ~ 4 kcal/mol.34

2,E+08

1,E+08

1,E+08 I x T 8,E+07

6,E+07

4,E+07

2,E+07

0,E+00 0 102030405060 T (K)

Figure 3.19 The plot of I x T vs. T for compound 5b–2 (I = EPR intensity).

3.5 Conclusion

In this chapter, several bis(formazanate)zinc complexes (L2Zn, 5) were successfully synthesized and fully characterized. Based on the crystal structures and NMR analysis, formazanate ligands are capable of isomerization between six-membered and five-membered chelate rings. This high flexibility of coordination chemistry allows the formazanate ligand to protect metal centers by using six-membered chelate ring in the absence of substrates. When substrates were introduced, the formazanate ligand can isomerize to form five-membered chelates and open space around the metal center to accommodate incoming substrates (please see Chapter 4). The cyclic voltammograms of 5 show two quasi-reversible redox-couples, the potential of which are tunable by introducing different substituents. In the case of 5a, two additional redox-couples can be located upon scanning to more negative potential. These two additional redox-couples indicate that each formazanate ligands can store up to two extra electrons. The crystal structures and theoretical calculations of the reduced bis(formazanate)zinc complexes (5- and 5-2) reveal the ligand-based radical property. The results presented here prove that the formazanate ligand is a new redox-active ligand platform.

(Formazanate)Zinc Complexes

3.6 Experimental section

General Considerations. All manipulations were carried out under nitrogen atmosphere using standard glovebox, Schlenk, and vacuum-line techniques. Toluene, hexane, and pentane

(Aldrich, anhydrous, 99.8%) were passed over columns of Al2O3 (Fluka), BASF R3-11- supported Cu oxygen scavenger, and molecular sieves (Aldrich, 4 Å). Diethyl ether and THF

(Aldrich, anhydrous, 99.8%) were dried by percolation over columns of Al2O3 (Fluka).

Deuterated solvents were vacuum transferred from Na/K alloy (C6D6, toluene-d8, Aldrich) and stored under nitrogen.

Dimethylzinc (Aldrich, 2.0 M in toluene) was used as received. NMR spectra were recorded on Varian Gemini 200, VXR 300, Mercury 400 or Varian 500 spectrometers. The 1H and 13C NMR spectra were referenced internally using the residual solvent resonances and reported in ppm relative to TMS (0 ppm); J is reported in Hz. Assignment of NMR resonances was aided by gradient-selected COSY, NOESY, HSQC and/or HMBC experiments using standard pulse sequences. All electrochemical measurements were performed under an inert N2 atmosphere in a glove box using an Autolab PGSTAT 100 (or PGSTAT 302N) computer-controlled potentiostat. Cyclic voltammetry (CV) was performed using a three-electrode configuration comprising of a Pt wire counter electrode, a Ag wire pseudoreference electrode and a Pt disk working electrode (CHI102, CH Instruments, diameter = 2 mm, or GoodFellow, Cambridge, UK; 99.99%; area = 1.25 × 10-3 ± 0.05 cm2). The Pt working electrode was polished before experiment using alumina slurry (0.05 μm), rinsed with distilled water and subjected to brief ultrasonication to remove any adhered alumina microparticles. The electrodes were then dried in an oven at 75 °C overnight to remove any residual traces of water. The CV data was calibrated by adding ferrocene in THF solution at the end of experiments. In all cases, there is no indication that addition of ferrocene influences the electrochemical behaviour of products.

All electrochemical measurements were performed at ambient temperatures under an inert N2 n n atmosphere in THF containing 0.1 M [ Bu4N][PF6] (or [ Bu4N][B(C6F5)4]) as the supporting electrolyte. Data were recorded with Autolab NOVA software (v.1.8). UV-Vis spectra were recorded in THF solution (~ 10-5 M) using a Avantes AvaSpec 3648 spectrometer and AvaLight-DHS lightsource inside a N2 atmosphere glovebox. Elemental analyses were performed at the Microanalytical Department of the University of Groningen or Kolbe Microanalytical Laboratory (Mülheim an der Ruhr, Germany).

Chapter 3

Product Synthesis

[PhNNC(p-tolyl)NNPh]ZnMe 4a. PhNNC(p-tolyl)NNHPh (1a) (0.715 g, 2.27 mmol) was dissolved in toluene (20 mL) and a 1.2 M solution of dimethyl zinc in toluene (2.1 mL, 2.52 mmol) was added. The solution turned from dark red to dark violet and gas evolution was observed. The solution was stirred for 30 minutes and removing the solvent in vacuo afforded the product (0.705 g, 1.79 mmol, yield: 79%). Crystals suitable for x-ray analysis were 1 obtained from toluene cooled to -30°C. H NMR (C6D6, 400 MHz): δ 8.33 (2H, d, J = 7.8 Hz, p-tol m-H), 7.83 (4H, d, J = 8.0 Hz, Ph o-H), 7.29 (2H, d, J = 7.8 Hz, p-tol o-H), 7.20 (2H, t,

J = 7.7 Hz, Ph m-H), 7.04 (2H, t, J = 7.2 Hz, Ph p-H), 2.26 (3H, s, p-tol p-CH3), -0.13 (3H, s, 13 ZnMe) ppm. C NMR (C6D6, 126 MHz): δ 154.33 (Ph i-C), 144.37 (p-tol i-C), 138.24 (NNCNN), 137.25 (p-tol p-C), 130.02 (p-tol o-C), 129.77 (Ph m-C), 128.05 (Ph p-C), 126.52

(p-tol m-C), 121.45 (Ph o-C), 21.64 (p-tol p-CH3), -8.60 (ZnMe) ppm. Anal. calcd for

C21H20N4Zn: C, 64.05; H, 5.12; N, 14.23. Found: C, 64.30; H, 5.16; N, 13.90.

[Ph2-Ph-Ph2](ZnMe)2(THF)2 4k. A vial was charged with 1k (105.8 mg, 0.20 mmol), 1.2 M solution of Me2Zn in toluene (2.5 mL, 3.0 mmol) and THF (5 mL). The reaction mixture was stirred at RT for overnight and some red solid was formed during stirring. The reaction mixture was then separated into two vials and hexane was add into both vials. After which 155.8 mg (0.19 mmol, 93 %) of red solid was obtained. The crystal suitable for x-ray crystallography was obtained by mixing 1k with 2 eqs of ZnMe2 solution in THF. 1H NMR 8 (400 MHz, THF-d , 25 °C):  8.14 (s, 4H, C6H4), 7.99 (d, 8H, J = 8 Hz, Ph o-CH),7.44 (t, 8H, 13 J = 7 Hz, Ph m-CH), 7.25 (t, 4H, J = 7 Hz, Ph p-CH), -0.35 (s, 6H, CH3) ppm. C NMR (100 8 MHz, THF-d , 25 °C):  154.1 (Ph i-C), 143.3 (NCN), 138.5 (C6H4 i-C), 128.9 (Ph m-C),

126.8 (C6H4 o-CH), 124.9 (Ph p-C), 120.3 (Ph o-C), -13.7 (CH3) ppm.

[Ph2-mPh-Ph2](ZnMe)2 4l. The procedure is the same as 4k; 1l (150.6 mg, 0.29 mmol), 1.2

M solution of Me2Zn in toluene (3.6 mL, 4.3 mmol) and THF (5 mL) were used. The product was quite soluble in THF/hexane or toluene/hexane mixture even at -30 °C; therefore, the 1 product was used directly without any further purification. H NMR (400 MHz, C6D6, 25 °C):

 9.50 (s, 1H, C6H4 CH), 8.51 (dd, 2H, J = 8, 1 Hz, C6H4 CH), 7.99 (d, 8H, J = 8 Hz, Ph o-

CH), 7.72 (t, 1H, J = 8 Hz, C6H4 CH), 7.27 (t, 8H, J = 8 Hz, Ph m-CH), 7.09 (t, 4H, J = 7 Hz, 13 Ph p-CH), -0.08 (s, 6H, CH3) ppm. C NMR (100 MHz, C6D6, 25 °C):  154.6, 144.5, 141.3, 129.9, 129.6, 125.4, 124.3, 121.5, -8.8 ppm.

(Formazanate)Zinc Complexes

[PhNNC(p-tolyl)NNPh]2Zn 5a. A 1.2 M solution of Me2Zn in toluene (0.82 mL, 0.98 mmol) was added slowly to a suspension of PhNNC(p-tolyl)NNHPh (1a) (620 mg, 1.97 mmol) in 10 mL of toluene at room temperature. The mixture was stirred for 2 h after which the colour had changed to intense blue. The volatiles were removed in vacuo and the residue was subsequently extracted into a hot 3:1 hexane/toluene mixture. Slow cooling of the clear dark blue solution to -30 °C for 2 days afforded 552 mg dark violet crystals of (PhNNC(p- 1 tolyl)NNPh)2Zn·(toluene)0.5 (5a) (0.75 mmol, 76%). H NMR (200 MHz, C6D6, 25 °C) δ 8.42 (d, 2H, J = 7.9, p-tolyl CH), 7.70 (d, 4H, J = 7.8, Ph o-H), 7.29 (d, 2H, J = 7.9, p-tolyl CH), 6.81 (t, 4H, J = 7.7, Ph m-H), 6.66 (t, 2H, J = 7.4, Ph p-H), 2.25 (s, 3H, p-tolyl CH3). 13C

NMR (50.4 MHz, C6D6, 25 °C) δ 152.8 (Ph ipso-C), 144.0 (NCN), 137.6 (p-tolyl ipso-C), 137.1 (p-tolyl CMe), 129.7 (Ph m-CH), 129.6 (p-tolyl CH), ~127.5 (overlapped, Ph p-CH),

126.4 (p-tolyl CH), 120.2 (Ph o-CH), 21.2 (p-tolyl CH3). Anal. Calcd for C43.5H38N8Zn: C, 70.78; H, 5.19; N, 15.18. Found: C, 70.74; H, 5.21; N, 15.13.

[PhNNC(tBu)NNPh]2Zn 5b. A 1.2M solution of Me2Zn in toluene (0.37 mL, 0.44 mmol) was diluted with 5 mL additional toluene. To this was added a solution of PhNNCH(tBu)NNPh (1b) (250 mg, 0.89 mmol) in 5 mL of toluene and the resulting mixture was stirred at 50 °C overnight, after which the reaction mixture was dried under vacuum. Extraction of the residue into hot hexane and subsequent crystallization at -30 °C yielded 227 mg of (PhNNC(tBu)NNPh)2Zn (5b) (as violet crystalline material (0.36 mmol, 82%). Crystals suitable for x-ray analysis were grown by slow evaporation of a pentane solution at room 1 temperature. H NMR (200 MHz, C6D6, 25 °C) δ 7.62 (d, 4H, J = 7.6, Ph o-H), 6.87 (t, 4H, J 13 = 7.5, Ph m-H), 6.68 (t, 2H, J = 7.4, Ph p-H), 1.71 (s, 9H, tBu). C NMR (50.4 MHz, C6D6, 25 °C) δ 153.2 (Ph ipso-C), 152.2 (NCN), 129.6 (Ph m-CH), 127.5 (Ph p-CH), 119.9 (Ph o-

CH), 39.7 (CMe3), 31.4 (CMe3). Anal. Calcd for C34H38N8Zn: C, 65.43; H, 6.14; N, 17.95. Found: C, 65.49; H, 6.07; N, 17.89.

[PhNNC(p-tolyl)NNMes]2Zn 5c. A 1.2 M solution of ZnMe2 in toluene (1.0 mL, 1.2 mmol) was added slowly to a solution of 1c (855.5 mg, 2.4 mmol) in toluene (20 mL) at room temperature. The mixture was stirred for 5 hours after which the volatiles were removed in vacuo. The residue was subsequently extracted into a hot hexane (15 mL). Slow cooling of the clear dark purple solution to -30°C for 2 days afforded 609 mg dark crystals of 5c (0.78 1 mmol, 65 %). H NMR (400 MHz, C6D6, 25 °C):  8.25 (d, 2H, J= 8.4 Hz, p-tolyl CH), 7.70 (dd, 2H, J= 8.4, 1.2 Hz, Ph o-H), 7.20 (d, 2H, J= 8.0 Hz, p-tolyl CH), 6.92 (t, 2H, J= 8.0 Hz,

Ph m-H), 6.76 (t, 1H, J= 7.4 Hz, Ph p-H), 6.38 (s, 2H, Mes CH), 2.20 (s, 3H, p-tolyl CH3),

Chapter 3

13 1.97 (s, 3H, Mes p-CH3), 1.88 (bs, 6H, Mes o-CH3). C NMR (100 MHz, C6D6, 25 °C): δ 152.9 (Ph ipso-C), 148.7 (Mes ipso-C), 144.4 (NCN), 137.4 (p-tolyl ipso-C), 137.1 (p-tolyl CMe), 130.0 (Ph m-CH), 129.8 (p-tolyl CH), 128.0 (Ph p-CH), 126.5 (p-tolyl CH), 121.5 (Ph o-CH), 21.6 (p-tolyl CH3). Anal. calcd for C46H46N8Zn: C, 71.17; H, 5.97; N, 14.43. Found: C, 71.16; H,6.06; N, 13.82.

[MesNNC(p-tolyl)NNMes]2Zn 5d. A 1.2 M solution of ZnMe2 (101.9 µl, 0.122 mmol) in toluene was added to an orange solution containing 1d (97.5 mg, 0.245 mmol) in toluene. The mixture was stirred overnight at 80 °C, after which the solvent was removed in vacuo to give 57.3 mg of compound 5d as a sticky powder (0.067 mmol, 54% yield). Recrystallization by slow diffusion of pentane into a toluene solution afforded orange crystalline material (16.9 mg, 1 0.020 mmol, 16% yield). H NMR (C6D6, 25 ºC, 500 MHz): δ 8.04 (d, 4H, J = 7.8, p-tol o-H),

7.08 (d, 4H, J = 7.8, p-tol m-H), 6.59 (s, 8H, Mes m-H), 2.11 (s, 6H, p-tol CH3), 2.08 (s, 12H, 13 Mes p-CH3), 1.88 (s, 24H, Mes o-CH3). C NMR (C6D6, 25 ºC, 126 MHz): δ 149.08 (Mes ipso-C), 145.31 (NCN), 137.28 (p-tol p-C), 136.69 (Mes p-C), 136.66 (p-tol ipso-C), 132.64

(Mes o-C), 130.20 (Mes m-CH), 129.85 (p-tol m-CH), 125.65 (p-tol o-CH), 21.48 (p-tol CH3),

21.15 (Mes p-CH3), 18.29 (Mes o-CH3). Anal. calcd. for C52H58N8Zn: C, 72.58; H, 6.79; N, 13.02; found: C,73.13; H, 6.87; N, 13.17.

[PhNNC(p-tolyl)NNC6F5]2Zn 5e. A 1.2 M solution of ZnMe2 in toluene (0.29 mL, 0.35 mmol) was added slowly to a solution of 1e (273.8 mg, 0.68 mmol) in toluene (15 mL) at room temperature. The mixture was stirred overnight after which the volatiles were removed in vacuo to afford 237.2 mg dark crystals of 5e (0.27 mmol, 80 %). The product thus obtained was ca. 95% pure. Further purification by recrystallization was not successful. 1H NMR (400

MHz, C6D6 25 °C):  8.22 (d, 2H, J = 8.4 Hz, p-tolyl CH), 7.50 (d, 2H, J = 7.6 Hz, Ph o-CH), 7.19 (d, 2H, J = 8.0 Hz, p-tolyl CH), 6.75 (t, 2H, J = 7.6 Hz, Ph m-CH), 6.67 (t, 1H, J = 7.6 19 Hz, Ph p-CH), 2.16 (s, 3H, p-tolyl CH3). F NMR (375 MHz, C6D6, 25 °C):  -154.4 (dd, 2F,

J = 22.5, 5.6 Hz, C6F5 o-CF), -158.3 (t, 1F, J = 21.5 Hz, C6F5 p-CF), -162.4 (td, 2F, J = 22.1, 13 5.2 Hz, C6F5 m-CF). C NMR (100 MHz, C6D6, 25 °C):  151.7 (Ph ipso-C), 145.7 (NCN),

141.1 (dm, J = 255 Hz, C6F5), 138.6 (dm, J = 256 Hz, C6F5), 137.8 (p-tolyl ipso-C), 137.7 6

(dm, J = 255 Hz, C6F5), 135.6 (p-tolyl p-C), 129.6( Ph m-C), 129.5 (Ph p-C), 129.4 (p-tolyl

CH), 125.8 (p-tolyl CH), 120.5 (Ph o-C), 120.4 (C6F5 ipso-C), 20.8 (p-tolyl CH3).

[MesNNC(p-tolyl)NNC6F5]2Zn 5f. A 1.2 M solution of ZnMe2 in toluene (0.5 mL, 0.6 mmol) was added slowly to a suspension of 1f (536 mg, 1.2 mmol) in toluene (20 mL) at room temperature. The mixture was stirred overnight after which the color had changed to intense

(Formazanate)Zinc Complexes purple. The volatiles were removed in vacuo and the residue was subsequently extracted into a hot 3:1 hexane/toluene mixture. Slow cooling of the clear dark purple solution to -30 °C for 2 days afforded 407 mg dark violet crystals of 5f (0.43 mmol, 71 %). NMR analysis shows the presence of 2 predominant isomers in solution; an additional isomer (< 5%) was observed in the 19F NMR, but resonances due to this species could not be separately observed in the 1H NMR. The 13C NMR spectrum is too complicated to do a full assignment of both isomers separately.

1 Major isomer (5f-i): H NMR (C6D6, 400 MHz, 25 °C)  8.09 (d, 2H, J = 8.0 Hz, p-tolyl CH),

7.22-7.09 (overlapped m, 2H, p-tolyl CH), 6.42 (s, 2H, Mes CH), 2.14 (s, 3H, p-tolyl p-CH3), 19 1.98 (s, 9H, Mes o-CH3 and p-CH3). F NMR (375 MHz, C6D6, 25 °C -152.7 (d, 2F, J =

18.0 Hz, C6F5 o-CF), -160.7 (t, 1F, J = 21.9 Hz, C6F5 p-CF), -163.3 (td, 2F, J = 21.9, 4.1 Hz,

C6F5 m-CF).

1 Minor isomer (5f-ii): H NMR (C6D6, 25 °C)  8.25 (d, 2H, J = 8.0 Hz, p-tolyl CH), 7.87 (d, 2H, J = 8.0 Hz, p-tolyl CH), 7.22-7.09 (overlapped m, 4H, 2 x p-tolyl CH), 6.37 (bs, 2H, Mes

CH), 6.31 (bs, 2H, Mes CH), 2.20 (s, 3H, p-tolyl p-CH3), 2.14 (s, 3H, p-tolyl p-CH3), 1.98 (s,

6H, Mes o-CH3), 1.94 (s, 3H, Mes p-CH3), 1.89 (s, 3H, Mes p-CH3), 1.87 (s, 6H, Mes o-CH3). 19 F NMR (375 MHz, C6D6, 25 °C) -153.0 (d, 2F, J = 19.3 Hz, C6F5 o-CF), -156.3 (d, 2F, J =

21.9 Hz, C6F5 o-CF), -160.7 (t, 1F, J = 21.9 Hz, C6F5 p-CF), -162.8 (td, 2F, J = 22.0, 4.7 Hz,

C6F5 m-CF), -165.3 (td, 2F, J = 21.6, 4.1 Hz, C6F5 m-CF), -167.6 (t, 1F, J = 21.7 Hz, C6F5 p- CF).

19 Trace isomer (5f-iii): F NMR (375 MHz, C6D6, 25 °C)  -155.2 (bs, 2F, C6F5 o-CF), -164.2

(t, 2F, J = 20.3 Hz, C6F5 m-CF), -167.3 (t, 1F, J = 23.5 Hz, C6F5 p-CF).

13 C NMR (100 MHz, C6D6, 25 °C) 152.3, 149.7 (p-tolyl ipso-C, 3.6b), 148.1 (Mes ipso-C, 3.6a), 147.8, 146.5 (p-tolyl ipso-C, 3.6b), 146.0 (p-tolyl ipso-C, 3.6a), 141.9 (dm, J = 250.4

Hz, C6F5), 141.4 (dm, J = 250.7 Hz, C6F5), 139.0 (p-tolyl p-C, 3.6b), 138.9, 138.7 (p-tolyl p-C,

3.6b), 138.7, 138.6 (dm, J = 249.3 Hz, C6F5), 138.5 (p-tolyl p-C, 3.6a), 135.6, 135.1, 131.9, 131.5, 131.5, 130.5 (Mes m-C, 3.6a), 130.1 (Mes m-C, 3.6b), 129.8, 129.7, 129.5, 128.9 (p- tolyl CH, 3.6b), 128.5, 128.3, 126.4 (p-tolyl CH, 3.6b), 126.1 (p-tolyl CH, 3.6a), 126.0, 21.6,

21.6 (p-tolyl p-CH3, 3.6b), 21.5 (Mes p-CH3, 3.6b), 21.3, 21.0, 20.9, 20.7 (Mes p-CH3, 3.6b),

18.3, 17.9 (Mes o-CH3, 3.6b).

Anal. calcd for C46H36F10N8Zn: C, 57.78; H, 3.79; N, 11.72. Found: C, 58.66; H, 4.00; N, 11.13.

Chapter 3

[PhNNC(C6F5)NNMes]2Zn 5g. A schlenk flask was charged with PhNNC(C6F5)NNHMes

(1g) (258.7 mg, 0.60 mmol), 1.2M solution of Me2Zn in toluene (0.25 mL, 0.60 mmol) and toluene 10 mL. The reaction mixture was stirred at RT overnight after which all volatiles were removed under vacuo. The crude product was dissolved in 10 mL of hexane and all volatiles was removed under vacuo again. After drying under vacuo, 210.5 mg deep orange solid of

(PhNHNC(C6F5)NNMes)2Zn (0.55 mmol, 76%) could be obtained. The isolated product is about 95 % pure with some unidentified impurity and free ligand present. Because of good solubility in all the common organic solvents, 5g is very hard to purify further and was used 1 as such in subsequent reactions. H NMR (400 MHz, C6D6, 25 °C) δ 7.75 (d, 2H, J = 7.9, Ph o-H), 7.21 (t, 2H, J = 7.9, Ph m-H), 6.92 (t, 1H, J = 7.6, Ph p-H), 6.33 (s, 2H, Mes m-H), 1.95 19 (s, 3H, Mes p-CH3), 1.82 (bs, 6H, Mes o-CH3). F NMR (376.4 MHz, C6D6, 25 °C) δ -143.67

(dd, 2F, J = 24.7, 7.9, C6F5 m-F), -156.4 (t, 1F, J = 21.5, C6F5 p-F), -163.0 (td, 2F, J = 21.7, 13 8.1, C6F5 o-F). C NMR (100.6 MHz, C6D6, 25 °C) δ 151.8 (Ph i-C), 148.0 (Mes i-C), 148.1-

147.7 (m, C6F5), 145.7-145.3 (m, C6F5), 143.0-142.4 (m, C6F5), 140.4-139.9 (m, C6F5), 139.9-

139.3 (m, C6F5), 137.4 (Mes p-C), 133.7 (NNCNN), 130.4 (Ph o-C), 130.1 (bs, Mes o-C),

129.3 (Mes m-C), ~128.5 (overlapped, Ph p-C), 120.9 (Ph o-C), 115.8 (td, J = 17.3, 4.0, C6F5 i-C), 21.1 (Mes p-CH3), 18.5 (bs, Mes o-CH3).

[PhNNC(p-tolyl)NNPh][MesNNC(CN)NNMes]Zn 5aj. A flask was charged with [PhNNC(p-tol)NNPh]ZnMe (4a) (783 mg, 2.0 mmol), 1j (663 mg, 2.0 mmol) and toluene (20 mL). The reaction mixture was stirred at room temperature for 2 hours after which the volatiles were removed in vacuo and the residue was subsequently extracted into a hot 1:1 hexane/toluene mixture. Slow cooling of the clear dark purple solution to -30 °C for 2 days afforded 796 mg dark violet crystals of 5aj (1.12 mmol, 56 %). NMR analysis shows the presence of 2 isomers in solution. The 13C NMR spectrum (see Supporting Information) is too complicated to do a full assignment of both isomers separately.

1 Major isomer (5aj-i): H NMR (400 MHz, C6D6, 25 °C) 7.98 (d, 2H, J = 8.4 Hz, p-tolyl CH), 7.51 (d, 4H, J = 8.4 Hz, Ph o-H), 7.21 (d, 2H, J = 8.0 Hz, p-tolyl CH), 6.95 (t, 4H, J = 8.0 Hz, Ph m-H), 6.85 (t, 2H, J= 7.6 Hz, Ph p-H), 6.55 (s, 2H, Mes m-CH), 6.32 (s, 2H, Mes m-CH), 2.24 (s, 3H, Mes p-CH3), 2.21 (s, 6H, Mes o-CH3), 1.94 (s, 3H, p-tolyl CH3), 1.59 (s,

3H, Mes p-CH3), 1.56 (s, 6H, Mes o-CH3).

1 Minor isomer (5aj-ii): H NMR (400 MHz, C6D6, 25 °C) 8.14 (d, 2H, J = 8.4 Hz, p-tolyl CH), 7.33 (d, 4H, J = 7.6 Hz, Ph o-CH), 7.19-7.13 (m, 2H, p-tolyl CH), 6.98-6.92 (m, 4H, Ph

(Formazanate)Zinc Complexes m-CH), 6.80 (t, 2H, J = 7.4 Hz, Ph p-H), 2.18 (s, 3H, Mes p-CH3), 1.94 (s, 3H, p-tolyl CH3),

1.84 (s, 6H, Mes o-CH3).

13 C NMR (100 MHz, C6D6, 25 °C)  152.4 (Ph ipso-C, 5aj-ii), 151.9 (Ph ipso-C, 5aj-i), 150.9 (Mes ipso-C, 5aj-i), 146.4 (Mes ipso-C, 5aj-ii), 145.5 (Mes ipso-C, 5aj-i), 143.3, 142.4 (p-tolyl ipso-C, 5aj-i), 138.3 (Mes p-C, 5aj-i), 137.6 (Mes ipso-C, 5aj-ii), 136.8, 136.6 (Mes p-C, 5aj-i), 136.0 (p-tolyl p-C, 5aj-i), 131.4, 130.9 (Mes p-C, 5aj-i), 130.6 (Mes m-CH, 5aj-i), 130.0 (Mes m-CH, 5aj-ii), 129.8 (Mes m-CH, 5aj-i), 129.4 (Ph m-H, 5aj-i), 129.3, 129.0 (p- tolyl CH, 5aj-ii), 128.8, 128.2 (Ph p-H, 5aj-ii), 127.8, 127.5, 126.4, 125.8 (p-tolyl CH, 5aj-i), 125.8 (p-tolyl CH, 5aj-ii), 125.5, 120.0 (p-tolyl CH, 5aj-ii), 119.6 (Ph o-H, 5aj-i), 117.9,

113.0, 20.8 (Mes p-CH3, 5aj-i), 20.3, 20.3 (Mes o-CH3, 5aj-ii), 20.1 (Mes p-CH3, 5aj-i), 19.5

(Mes o-CH3, 5aj-i), 18.0 (Mes o-CH3, 5aj-ii), 17.0 (Mes o-CH3, 5aj-i).

Anal. calcd for C40H39N9Zn: C, 67.55; H, 5.53; N, 17.73. Found: C, 67.79; H, 5.58; N, 17.24.

[PhNNC(p-tolyl)NNPh]2Zn[Na(THF)3] [5a]Na. One leg of a double-schlenk flask was charged with 5a (750 mg, 1.02 mmol), Na(Hg) (2.447 wt% Na, 1222 mg, 1.30 mmol) and 15 mL of THF. The reaction mixture was stirred for overnight, filtered and reduced to half of the original volume. Slow diffusion of hexane (15 mL) into the THF solution precipitated 658 mg + - of [Na(THF)3] [(PhNNC(p-tolyl)NNPh)2Zn] as brown crystalline material (0.707 mmol,

69%). Anal. Calcd for C52H58N8NaO3Zn: C, 67.05; H, 6.28; N, 12.03. Found: C, 66.84; H, 6.25; N, 11.90.

[PhNNC(p-tolyl)NNPh]2Zn[Na(THF)3]2 [5a]Na2. A mixture of solid 5a (50 mg, 0.068 mmol) and 149.5 mg Na/Hg (2.447 wt%, 3.7 mg Na, 0.159 mmol) was prepared. With stirring, 7 mL of THF was added at room temperature. After stirring the reaction mixture one week, 28 mL of pentane was added at room temperature to precipitate the product. The supernatant was decanted and the crystalline product washed with pentane to give 30.0 mg of + 2- [Na(THF)3] 2[(PhNNC(p-toyl)NNPh)2Zn] as green crystalline material (0.026 mmol, 39%).

Anal. Calcd for C64H82N8Na2O6Zn: C, 65.55; H, 7.06; N, 9.57. Found: C, 65.53; H, 6.89; N, 9.87.

[PhNNC(tBu)NNPh]2Zn[Na(THF)3] [5b]Na. A mixture of solid 5b (500 mg, 0.801 mmol) and 715 mg Na/Hg (2.447 wt%, 17.5 mg Na, 0.761 mmol) was prepared. With stirring, 8 mL of THF was added at room temperature. After stirring the reaction mixture overnight, 10 mL of pentane was added at -30 °C to precipitate the product. The supernatant was decanted and the crystalline product washed with pentane to give 365 mg of dark violet

Chapter 3

+ - [Na(THF)3] [(PhNNC(tBu)NNPh)2Zn] (0.432 mmol, 54%). Anal. Calcd for

C46H62N8NaO3Zn: C, 63.99; H, 7.24; N, 12.98. Found: C, 63.61; H, 7.24; N, 12.66.

.- [PhNNC(tBu)NNPh]2Zn[Na(THF)3]2 [5b]Na2. A solution of [5b] Na (49.5 mg, 57.3 μmol) in 2 mL THF was added to 53.9 mg Na/Hg (2.447 wt%, 1.32 mg Na, 57.4 μmol). After stirring for 6h at room temperature, the resulting green solution was filtered and the product precipitated by addition of hexane at -30 °C. A dark oil was obtained, which was redissolved in 2 mL THF. Slow diffusion of hexane (5 mL) into the THF solution at room temperature + 2- afforded 26.2 mg [Na(THF)3] 2[(PhNNC(tBu)NNPh)2Zn] as dark green crystalline material (25.4 μmol, 44%).

Crystal structure determination Suitable crystals reported in this chapter were mounted on a cryo-loop in a drybox and transferred, using inert-atmosphere handling techniques, into the cold nitrogen stream of a Bruker D8 Venture diffractometer. The final unit cell was obtained from the xyz centroids of 9964 (4a), 9951 (4k), 70044 (5a), 12978 (5b), 26123 (5c), 9969 (5d), 33453 (5f), 9144 (5aj),

9756 ([5a]Na), 9538 ([5b]Na), and 9258 ([5b]Na2) reflections after integration. Intensity data were corrected for Lorentz and polarisation effects, scale variation, for decay and absorption: a multiscan absorption correction was applied, based on the intensities of symmetry-related reflections measured at different angular settings (SADABS).35 The structures were solved by direct methods using the program SHELXS.36 The hydrogen atoms were generated by geometrical considerations and constrained to idealised geometries and allowed to ride on their carrier atoms with an isotropic displacement parameter related to the equivalent displacement parameter of their carrier atoms. Structure refinement was performed with the program package SHELXL.36 Crystal data and details on data collection and refinement are presented in the following tables.

Crystallographic data 4a 4k(THF)2 5a 5b chem formula C21H20N4Zn C42H46N8O2Zn2 C40H34N8Zn(C7H8)0.5 C34H38N8Zn Mr 393.78 825.61 738.19 624.09 cryst syst monoclinic monoclinic monoclinic triclinic color, habit blue, platelet purple, block dark brown, needle dark purple, block size (mm) 0.14 x 0.13 x 0.03 0.33 x 0.12 x 0.10 0.71 x 0.20 x 0.17 0.79 x 0.25 x 0.25 space group P21/c P21/c C2/c P-1 a (Å) 17.3788(4) 13.4061(7) 31.8770(12) 9.9902(4) b (Å) 15.6932(3) 16.5157(10) 13.0561(4) 18.0401(10) c (Å) 21.3006(5) 18.1655(10) 18.6713(5) 18.7642(8)  (°) 76.312(3) β (°) 111.7590(10) 108.149(2) 106.316(1) 77.040(2)  (°) 87.463(3) V (Å3) 5395.4(2) 3821.9(4) 7457.9(4) 3201.9(3)

(Formazanate)Zinc Complexes

Z 12 4 8 4 -3 calc, g.cm 1.454 1.435 1.315 1.295 Radiation [Å] Cu Kα 1.54178 Mo Kα 0.71073 Mo Kα 0.71073 Mo Kα 0.71073 -1 µ(Mo K), mm 1.303 0.702 0.803 -1 µ(Cu K), mm 1.976 F(000) 2448 1720 3080 1312 temp (K) 100(2) 100(2) 150(2) 110(2)  range (°) 2.74-72.98 2.70-27.18 1.70–27.49 1.81–27.73 data collected (h,k,l) -21:20; -19:17; - -17:17; -21:21; - -41:40, -16:16, - -12:12, -23:23, - 19: 26 23:23 24:24 24:24 min, max transm 0.7694, 0.9523 0.6631, 0.7455 0.845, 0.888 0.786,0.818 rflns collected 67652 84674 95132 43315 indpndt reflns 10411 8466 8570 15086 observed reflns Fo  8756 5707 7422 13971 2.0 σ (Fo) R(F) (%) 3.57 4.89 2.90 3.88 wR(F2) (%) 8.49 9.87 7.38 10.21 GooF 1.045 1.042 1.056 1.044 weighting a,b 0.0419, 5.1816 0.0332, 12.5761 0.0351, 6.7906 0.0539, 2.25352 params refined 709 487 508 789 min, max resid dens -0.477, 1.178 -0.842, 1.609 -0.330, 0.338 -0.624, 1.634

5c 5d 5f 5aj chem formula C46H46N8Zn C52H58N8Zn C46H36F10N8Zn C40H39N9Zn Mr 776.28 860.43 956.20 711.17 cryst syst Monoclinic Triclinic Monoclinic Monoclinic color, habit violet, block dark red, needle purple, block violet, block size (mm) 0.28 x 0.40 x 0.46 0.08 x 0.09 x 0.21 0.23 x 0.52 x 0.80 0.19 x 0.16 x 0.09 space group C2/c P-1 P2(1)/n P2(1)/c a (Å) 17.9988(6) 11.1353(6) 11.1564(4) 15.070(4) b (Å) 11.4774(5) 12.0886(7) 16.0507(5) 12.864(3) c (Å) 19.3454(5) 17.9233(9) 24.0955(10) 19.235(5)  (°) 75.379(2) β (°) 96.906(1) 87.426(2) 101.165(3) 104.010(8)  (°) 76.677(2) V (Å3) 2271.5(2) 4233.1(3) 3618.2(15) Z 2 4 -3 calc, g.cm 1.300 1.258 1.500 1.306 Radiation [Å] Mo Kα 0.71073 Mo Kα 0.71073 Mo Kα 0.71073 Mo Kα 0.71073 -1 µ(Mo K), mm 0.663 0.586 0.669 0.721 -1 µ(Cu K), mm F(000) 1632 912 1952 1488 temp (K) 150(2) 100(2) 150(2) 100(2)  range (°) 2.1-27.5 2.75-27.13 1.5-27.5 2.1- 27.9 data collected (h,k,l) -23:22; -14:14; - -14:14; -15:15; - -14:14; -20:20; - -19:19; -16:16; - 25:23 22:21 30:31 25:25 min, max transm rflns collected 28246 75805 48788 99698 indpndt reflns 4542 10034 9672 8651 observed reflns Fo  2.0 4357 8037 7477 7215 σ (Fo) R(F) (%) 2.54 4.04 5.08 3.39 wR(F2) (%) 7.33 9.41 14.28 8.69 GooF 1.04 1.043 1.04 1.04 weighting a,b 0.0391, 3.9002 0.0368, 1.7988 0.0677, 4.5821 0.0375 2.0127 params refined 253 564 594 458 min, max resid dens -0.41, 0.39 -0.39, 0.42 -0.52, 0.88 -0.38, 0.36

Chapter 3

[5a]Na [5b]Na [5b]Na2 chem formula C52H58N8NaO3Zn C46H62N8NaO3Zn C54H78N8Na2O5Zn Mr 931.42 863.40 1030.59 cryst syst Monoclinic Triclinic Monoclinic color, habit dark violet, needle dark brown, block dark green, block size (mm) 0.22 x 0.13 x 0.10 0.31 x 0.24 x 0.22 0.39 x 0.36 x 0.17 space group P21/n P-1 P21/c a (Å) 10.0100(4) 9.8856(3) 17.4494(10) b (Å) 24.9939(10) 11.7402(3) 30.9957(18) c (Å) 19.3574(7) 20.6200(6) 20.2854(11)  (°) 76.8064(15) β (°) 104.4800(10) 85.8121(15) 93.569(2)  (°) 76.8111(15) V (Å3) 4689.2(3) 2267.96(11) 10950.2(11) Z 4 2 8 -3 calc, g.cm 1.319 1.264 1.250 Radiation [Å] Mo Kα 0.71073 Mo Kα 0.71073 Mo Kα 0.71073 -1 µ(Mo K), mm 0.586 0.599 0.517 -1 µ(Cu K), mm F(000) 1964 918 4400 temp (K) 100(2) 100(2) 100(2)  range (°) 3.07-28.32 2.84-26.37 2.68-28.28 data collected (h,k,l) -13:13, -33:33, -22:25 -12:12;-14:14; -25:25 -23:23, -41:41, -27:27 min, max transm 0.882,0.944 0.8360, 0.8794 0.8237, 0.9172 rflns collected 146989 119332 214654 indpndt reflns 11624 9275 27150

observed reflns Fo  2.0 σ (Fo) 8762 8722 19678 R(F) (%) 4.06 4.39 4.13 wR(F2) (%) 7.74 10.13 9.26 GooF 1.034 1.039 1.010 weighting a,b 0.0264, 4.3585 0.0431, 3.6733 0.0435, 8.1093 params refined 588 625 1372 min, max resid dens -0.469, 0.064 -1.569, 1.224 -0.767, 0.927

3.7 References

(1) (a) Bourget-Merle, L.; Lappert, M. F.; Severn, J. R. Chem. Rev., 2002, 102, 3031–3066. (b) Tsai, Y.-C. Coord. Chem. Rev., 2012, 256, 722–758. (2) Kopylovich, M. N.; Pombeiro, A. Coord. Chem. Rev., 2011, 255, 339-355. (3) Zaidman, A.; Khasbiullin, I.; Belov, G.; Pervova, I. G.; Lipunov, I. N. Pet. Chem., 2012, 52, 28-34. (4) Protasenko, N. A.; Poddel’sky, A. I.; Bogomyakov, A. S.; Fukin, G. K.; Cherkasov, V. K. Inorg. Chem., 2015, 54, 6078–6080. (5) Gilroy, J. B.; Ferguson, M. J.; McDonald, R.; Hicks, R. G. Inorg. Chim. Acta, 2008, 361, 3388–3393. (6) Gilroy, J. B.; Ferguson, M. J.; McDonald, R.; Patrick, B. O.; Hicks, R. G. Chem. Commun., 2007, 126–128. (7) Jeske, R. C.; Rowley, J. M.; Coates, G. W. Angew. Chem. Int. Ed., 2008, 47, 6041–6044. (8) Rousset, É.; Whitehorne, T. J. J.; Baslon, V.; Reber, C.; Schaper, F. Eur. J. Inorg. Chem., 2011, 331– 335. (9) Biyikal, M.; Löhnwitz, K.; Meyer, N.; al, E. Eur. J. Inorg. Chem. 2010, 1070–1081. (10) (a) Pilz, M. F.; Limberg, C.; Lazarov, B. B; Hultzsch, K. C; Ziemer, B. Organometallics, 2007, 26, 3668–3676. (b) Farwell, J. D.; Hitchcock, P. B.; Lappert, M. F.; Luinstra, G. A.; Protchenko, A. V.; Wei, X.-H. J. Organ. Chem., 2008, 693, 1861–1869.

(Formazanate)Zinc Complexes

(11) (a) Vela, J.; Zhu, L.; Flaschenriem, C. J.; Brennessel, W. W.; Lachicotte, R. J.; Holland, P. L. Organometallics, 2007, 26, 3416–3423. (b) Lee, S. Y.; Na, S. J.; Kwon, H. Y.; Lee, B. Y.; Kang, S. O. 2004, 23, 5382–5385. (12) (a) McWilliams, S. F.; Holland, P. L. Acc. Chem. Res., 2015, 48, 2059–2065. (b) Burgess, B. K. Chem. Rev., 1990, 90, 1377–1406. (13) MacInnis, M. C.; McDonald, R.; Ferguson, M. J.; Tobisch, S.; Turculet, L. J. Am. Chem. Soc., 2011, 133,13622–13633. (14) (a) Pinsky, M.; Avnir, D. Inorg. Chem., 1998, 37, 5575-5582. (b) Cirera, J.; Alemany, P.; Alvarez, S. Chem. Eur. J., 2004, 10, 190–207. (15) (a) Cheng, M.; Moore, D. R.; Reczek, J. J.; Chamberlain, B. M.; Lobkovsky, E. B.; Coates, G. W. J. Am. Chem. Soc., 2011, 123, 8738–8749. (b) Peng, Y. L.; Hsueh, M. L.; Lin, C. C. Acta Cryst., 2007, E63, m2388. (c) Schulz, S.; Eisenmann, T.; Bläser, D.; Boese, R. Z. Anorg. Allg. Chem., 2009, 635, 995-1000. (d) Vaughan, B. A.; Wetherby, A. E. Acta Cryst., 2012, E68, m343. (16) (a) Martinez, C. R.; Iverson, B. L. Chem. Sci., 2012, 3, 2191–2201. (b) Salonen, L. M.; Ellermann, M.; Diederich, F. Angew. Chem. Int. Ed., 2011, 50, 4808–4842. (17) (a) Balt, S.; Renkema, W. E. Inorg. Chim. Acta, 1977, 330, 161–168. (b) Balt, S.; Meuldijk, J.; Renkema, W. E. Inorg. Chim. Acta, 1980, 333, 173–178. (18) Meuldijk, J.; Renkema, W. E.; Van Herk, A. M.; Stam, C. H. Acta Cryst., 1983, C39, 1536–1538. (19) Travieso-Puente, R.; Chang, M.-C.; Otten, E.; Dalton Trans., 2014, 43, 18035-18041. (20) Geiger, W. E.; Barrière, F. Acc. Chem. Res., 2010, 43, 1030–1039. (21) (a) Gilroy, J. B.; McKinnon, S. D.; Kennepohl, P.; Zsombor, M. S.; Ferguson, M. J.; Thompson, L. K.; Hicks, R. G. J. Org. Chem., 2007, 72, 8062–8069. (b) Robin G Hicks, 1.; Lars Öhrström, 2. A.; Patenaude1, G. W. Inorg. Chem., 2001, 40, 1865–1870. (22) "Verdazyls and Related Radicals Containing the Hydrazyl [R2NNR] Group": Hicks, R. G. in Stable Radicals, Wiley,Hoboken, 2010, p. 245. (23) Chivers, T.; Eisler, D. J.; Fedorchuk, C.; Schatte, G.; Tuononen, H. M.; Boeré, R. T. Chem. Commun., 2005, 3930–3932. (24) Moilanen, J.; Borau-Garcia, J.; Roesler, R.; Tuononen, H. M. Chem. Commun., 2012, 48, 8949–8951. (25) Chivers, T.; Eisler, D. J; Fedorchuk, C.; Schatte, G.; Tuononen, H. M.; Boeré, R. T. Inorg. Chem., 2006, 45, 2119–2131. (26) Abe, M. Chem. Rev., 2013, 113, 7011–7088. (27) Gardiner, M. G.; Hanson, G. R.; Henderson, M. J.; Lee, F. C.; Raston, C. L. Inorg. Chem., 1994, 33, 2456–2461. (28) (a) Shultz, D. A.; Boal, A. K.; Driscoll, D. J.; Kitchin, J. R.; Tew, G. N. J. Org. Chem., 1995, 60, 3578–3579. (b) David A Shultz; Andrew K Boal, A.; Farmer, G. T. J. Org. Chem., 1998, 63, 9462– 9469. (29) Fico, R. M., Jr; Hay, M. F.; Reese, S.; Hammond, S.; Lambert, E.; Fox, M. A. J. Org. Chem., 1999, 64, 9386–9392. (30) (a) Tosha M Barclay; Robin G Hicks; Martin T Lemaire, A.; Thompson, L. K. Inorg. Chem., 2001, 40, 6521–6524. (b) Brook, D. J. R.; Yee, G. T.; Hundley, M.; Rogow, D.; Wong, J.; Van-Tu, K. Inorg. Chem., 2010, 49, 8573–8577. (c) Anderson, K. J.; Gilroy, J. B.; Patrick, B. O.; McDonald, R.; Ferguson, M. J.; Hicks, R. G. Inorg. Chim. Acta, 2011, 364, 480–488. (31) Kuhn, R.; Trischmann, H. Monatsh. Chem., 1964, 95, 457–479. (32) (a) Hohenberg, P.; Kohn, W. Phys. Rev., 1964, 136, B864–B871. (b) Kohn, W.; Sham, L. J. Phys. Rev., 1965, 137, A1133–A1138. (33) (a) Neese, F. Coord. Chem. Rev., 2009, 253, 526–563. (b) Ciofini, I. Chem. Rev., 2003, 247, 187–209. (34) Singh, A. P.; Samuel, P. P.; Roesky, H. W.; Schwarzer, M. C.; Frenking, G.; Sidhu, N. S.; Dittrich, B. J. Am. Chem. Soc., 2013, 135, 7324–7329. (35) Bruker. APEX2 (v2012.4-3), SAINT (Version 8.18C) and SADABS (Version 2012/1). Bruker AXS Inc., Madison, Wisconsin, USA. 2012. (36) Sheldrick, G. Acta Crystallographica Section A 2008, 64, 112.

Chapter 3

Chapter 4

(Formazanate)Boron Difluoride Formation via Zinc to Boron Transmetallation

A high-yield synthetic route to obtain mono(formazanate)boron difluoride complexes (LBF2; compounds 6) via an uncommon zinc to boron transmetallation reaction is presented. Two six-coordinated zinc complexes (compounds 7), which are key intermediates in the transmetallation reaction, were isolated and characterized by multi-nuclear NMR spectroscopy and X-ray crystallography. A mechanism was proposed that utilizes the flexible coordination ability of formazanate ligands, involving interconversion between 6- and 5- membered chelate binding modes. The observed transmetallation demonstrates that the ability of these ligands to rearrange to a 5-membered chelate ring leads to decreased steric hindrance around the central metal atom, which can open up new pathways for reactivity.

Parts of this chapter have been published:

M.-C. Chang and E. Otten* “Synthesis and ligand-based reduction chemistry of boron difluoride complexes with redox-active formazanate ligands” Chem. Commun., 2014, 50, 7431-7433. Chapter 4

Chapter 4 (Formazanate)Boron Difluoride Formation via Zinc to Boron Transmetallation

4.1 Introduction

Transmetallation is a key elementary reaction in many important (catalytic) reactions, such as Suzuki coupling1 and Negishi coupling2 (Scheme 4.1), both of which are powerful tools to make C-C bonds.3 Both of these reactions form a R-Pd-X intermediate (X = halide) resulting from oxidative addiction (OA) of an organic halide (R-X). In a subsequent step, transmetallation (TM) takes place, transferring a second organic fragment from organoboron (Suzuki coupling) or organozinc (Negishi coupling) reagents to the palladium center to give a palladium diorganyl intermediate. The catalytic cycle is closed by a reductive elimination (RE) resulting in regeneration of active catalyst and product dissociation. In addition to these well- known coupling reactions, transmetallation also plays an important role in preparing a wide range of organozinc reagents, most of which are not commercially available or expensive, via boron-to-zinc transmetallation (Scheme 4.1).4 Organozinc reagents and chiral ligands are a very useful combination for asymmetric arylation of aldehydes resulting in enantiopure diaryl methanols5, which are valuable compounds in the pharmaceutical field as active ingredients or as synthetic intermediates.6 Besides applications in organic synthesis, organozinc reagents are also frequently used to transfer aryl or alkyl groups to transition metal centers, such as Ru, Nb, and Ta.7 The use of organozinc reagents is often advantageous because these are milder alkylating agents and less prone to give side products resulting from 1e-reduction than the corresponding organolithium or Grignard reagents.

In all of the examples mentioned above the transferable groups from either organoboron or organozinc reagents are usually reactive anionic monodentate ligands, more specifically aryl and alkyl groups. In the field of inorganic/organometallic chemistry, transmetallation reactions also are frequently used to transfer multidentate ancillary ('spectator') ligands from alkali metal salts to the target metal centers. Examples of commonly used reagents in this area are the alkali metal salts MCp8 and M(nacnac)9 (M = Li, Na, K), which are able to transfer the anionic organic moiety to less electropositive (transition) metal halides. These ionic reagents show high ligand exchange rates (lability), which results in kinetically facile transmetallation (metathesis) reactions.

(Formazanate)Boron Difluoride Formation via Zinc to Boron Transmetallation

Scheme 4.1 General mechanism of Suzuki coupling (top left), Negishi coupling (top right), and preparation of organozinc reagents via boron-to-zinc transmetallation (bottom). A key feature of transmetallation reaction is that the two metal centers (B and Pd;10 Zn and Pd;2b,11 B and Zn;4b or alkali metal and metal halide) should be able to access a (hetero)bimetallic dimeric intermediate from which ligand exchange takes place.12 Therefore, it is relatively uncommon to observe a transmetallation reaction occurring at a four- coordinated zinc center bearing two bidentate ligands due to the steric hindrance around the metal center. Here we present an unusual zinc to boron transmetallation between bis(formazanate)zinc complexes (L2Zn, 5) and boron trifluoride (BF3) resulting in formation of (formazanate)boron difluoride compounds (LBF2, 6).

4.2 Formazanate Transfer from Zinc to Boron

4.2.1 Formation of (Formazanate)Boron Difluoride via Transmetallation

In Chapter 3, we showed that bis(formazanate) zinc complexes such as

[PhNNC(R)NNPh]2Zn (5a: R = p-tolyl; 5b:R = tBu) are capable of storing one or two electrons in the ligand frameworks to form stable 1- and 2-electron reduced complexes (5a– /5b– and 5a–2/5b–2). The crystal structures of the reduced complexes show (weak)

Chapter 4 interactions between the internal nitrogen atoms of the formazanate ligands and sodium counter cations (Chart 4.1, also see Figure 3.14 and Figure 3.15 in Chapter 3).

Chart 4.1

Based on this observation we anticipated that the reduction potential of bis(formazanate) zinc compounds could be altered by coordination to neutral Lewis acids. In the course of testing a range of Lewis acids for binding to 5a, we found that BF3 reacts cleanly via salt metathesis, opening a high-yield synthetic route to obtain mono(formazanate)boron difluoride complexes

(LBF2; compound 6) (Scheme 4.2). The chemical and physical properties of complexes 6 will be discussed in Chapter 5. In the following parts of this chapter, the unexpected products and intermediates isolated from the reaction of compounds 5 with BF3·Et2O will be described.

Scheme 4.2 Synthesis of mono(formazanate)boron difluoride complexes (LBF2: 6) from bis(formazanate)zinc complexes (L2Zn: 5). 4.2.2 Isolation of a Six-Coordinated Zinc Complex as a Key Intermediate

Compound 6a can be easily prepared by reacting the bis(formazanate) zinc compound 5a with

3 equivalents of BF3·Et2O at 70°C overnight. During the reaction, the color of the reaction mixture changes from deep blue to red and a white precipitate, which could be ZnF2, can be observed in the reaction flask. In contrast, in the case of (PhNNC(C6F5)NNMes)2Zn (5g) the color of the reaction mixture fades from deep to light orange upon heating 5g in the presence of 3 eq BF3·Et2O at 70°C overnight, but precipitation of ZnF2 was not observed. From this reaction mixture, orange crystals of the product 7g were obtained by slow diffusion of hexane into a toluene solution at -30°C in 85% yield (Figure 4.1, metrical parameters in Table 4.1).

(Formazanate)Boron Difluoride Formation via Zinc to Boron Transmetallation

The crystal structure of 7g shows a distorted octahedral zinc center, with two BF3 units incorporated in the compound. In the crystal structure, there are two tridentate

(PhNNC(C6F5)NNMes(BF3)) units coordinated to the Zn center in a meridional fashion via two nitrogens and a fluorine atom to give a [NNF]2Zn compound. This unusual binding motif results from the interaction of BF3 with the terminal nitrogen of the formazanate fragment (the NMes group), which gives rise to 2 five-membered chelate rings upon coordination to the Zn center. To the best of our knowledge, the structural characterization of this BF3 binding mode has no precedent in the literature, although the ‘frustrated Lewis pair’ (tmp)MgCl/BF3 has been postulated to contain a B-F fragment appended to a Mg-N(tmp) bond.13 The formation of compound 7g suggests that the formazanate ligand has the ability to isomerize from a binding mode that involves the two terminal nitrogen atoms of the ligand (6-membered chelate) to a five-membered chelating ring, in which the Zn center binds both a terminal and an internal N atom. This ligand rearrangement opens sufficient space around the Zn center to allow the BF3 substrate to bind.

B1_1 F11_1 N4_1 N6_1 N5_1 C29_1 N3_1 C7_1 Zn1_1 N7_1 F14_1 N1_1 B2_1 N2_1 N8_1

Figure 4.1 Molecular structure of 7g showing 50% probability ellipsoids. One of the two independent molecules in 7g is shown; hydrogen atoms are omitted for clarity.

The metrical parameters of 7g (Table 4.1) show that the bond lengths of N1-N2 and N5-N6 are shorter than N3-N4 and N7-N8 (1.272(3)/1.266(3) Å vs. 1.319(3)/1.327(3) Å); in addition, the bond lengths of N2-C7 and N6-C29 are longer than C7-N3 and C29-N7 (1.397(3)/1.407(3) Å vs. 1.317(3)/1.311(3) Å). These bond length distributions suggest that the negative charge

Chapter 4 of the formazanate ligand [1g]- is partially localized at the terminal nitrogen atoms coordinated to boron centers due to the strong electron-withdrawing ability of the BF3 units.

In compound 7g there is a face-centered interaction of an electron-poor C6F5 ring with an electron-rich mesityl group (interplanar angle = 19.00° and 25.08°, centroid-centroid distance = 3.375 Å and 3.575 Å), indicative of an aromatic donor-acceptor interaction that is commonly observed for combinations of electron-rich/poor rings.14

Table 4.1 Selected bond length (Å) and bond angles (o) of 7ga N1-N2 1.272(3) N5-N6 1.266(3) N2-C7 1.397(3) N6-C29 1.407(3) C7-N3 1.317(3) C29-N7 1.311(3) N3-N4 1.327(3) N7-N8 1.319(3) N4-B1 1.584(5) N8-B2 1.584(4) B1-F11 1.349(4) B2-F14 1.415(4) Zn1-N1 2.066(2) Zn1-N5 2.067(2) Zn1-N3 2.088(2) Zn1-N7 2.109(2) Zn1-F11 2.203(2) Zn1-F14 2.156(2) N1-Zn1-N3 77.26(9) N5-Zn1-N7 76.66(8) N3-Zn1-F11 75.76(8) N7-Zn1-F14 76.54(8) b b C6F5(C7)-Mes(N4) 3.375 C6F5(C29)-Mes(N8) 3.575 NNCNNZn1/NNCNNZn1c 85.24 a the _1 labels on atoms are omitted; b the centroid to centroid distance of c C6F5 ring and Mes ring; the angle between the planes defined by the central Zn atom and the N atoms of the formazanate backbone.

The 19F NMR of 7g shows six distinct resonances with integration ratio of 1:1:3:1:1:1 (Figure

4.2). Five resonances with the same integration (1F) suggest that all F substituents of the C6F5 ring are inequivalent due to hindered rotation around the C-C6F5 bond. The resonance 11 10 integrating as 3F shows B and B coupling features and can be assigned to the BF3 unit. 19 The appearance of the BF3 group in the F NMR does not change upon cooling to -55 °C, which suggests that the barrier to rotation around the N-BF3 bond is low. Conversely, heating an NMR sample of 7g to moderate temperature (65 °C for 24 h) does not result in changes in the spectroscopy, confirming that the octahedral [NNF]2Zn complex is quite stable. Upon heating the NMR tube to 130 °C overnight, full conversion to the corresponding LBF2 complex 6g is obtained.

(Formazanate)Boron Difluoride Formation via Zinc to Boron Transmetallation

19 Figure 4.2 F NMR spectrum of compound 7g (C6D6, 375 MHz)

4.2.3 Proposed Mechanism of Transmetallation Reaction

The sequential transformation 5g→7g→6g suggests that a six-coordinate species related to 7g is likely also involved in the formation of 7g. Based on these observations, we propose the following mechanism for the transmetallation leading to compound 7g (Scheme 4.3): (i) formazanate rearrangement from a 6- to a 5-membered chelate ring liberates the terminal N- atom, (ii) BF3 binds to this terminal N-atom and brings a B-F group in proximity of the Zn centre, and (iii) the F atom binds to the Lewis acidic Zn(II) centre to from a tridentate [NNF] ligand with two 5-membered chelate rings. Repeating this sequence for the second formazanate ligand results in the formation of the [NNF]2Zn complex 7g. Elimination of ZnF2 from this complex either occurs rapidly (in case of 5a→6a) or requires heating to proceed so that the intermediate can be isolated (5g→7g→6g). A reason for the increased stability of 7g vs that of putative intermediate in the case of 5a could be the favorable -interactions between 14 the electron-rich N-Mes and the electron-poor C-C6F5 substituents which are present only when the formazanate ligands adopt a 5-membered chelate ring.

Chapter 4

Scheme 4.3 Proposed mechanism of transmetallation from bis(formazanate) zinc complex to mono(formazanate) boron difluoride complex.

4.2.4 Reaction of Heteroleptic Complex 5aj with BF3·Et2O

When the heteroleptic complex 5aj ([1a][1j]Zn) was reacted with BF3·Et2O (1 or 3 eq) at room temperature, there was no indication of formation of boron difluoride complexes (6a and 6j). Instead, 1H NMR spectroscopy allows identification of the homoleptic bis(formazanate) complex 5a as one of the major species as a result of ligand redistribution (Scheme 4.4). The resonances in the aliphatic region indicate that the two mesityl groups of the remaining formazanate ligand [1j]- become inequivalent. Based on the formation of 5a, it is likely that this product is a homoleptic zinc complex containing [1j]-. From the reaction mixture, two types of crystals with different shapes can be isolated. One of these shows a similar shape as 5a; the other crystals are black diamonds, which were shown by X-ray crystallography to be compound [MesNNC(CN)NNMes(BF3)]2Zn (7j, Figure 4.3, metrical parameters in Table 4.2). The data indicate that 7j has very similar structure as 7g; both of them have an octahedral zinc center bearing two formazanate(BF3) units, which in the case of

7j is a MesNNC(CN)NNMes(BF3) unit. All the metrical parameters of 7j are very similar with 7g. In the solid state structure of complex 7j the N-Mes rings are oriented nearly parallel (interplanar angle = 2.45° and distance between centroids = 3.641 Å) and stack in an off- center fashion as is usually observed for electron-rich aromatic rings.14 The formation of 5a and 7j from compound 5aj by treatment with BF3·Et2O indicates that formazanate ligands (in this case [1a]- and [1j]-) are capable of intermolecular ligand exchange.

(Formazanate)Boron Difluoride Formation via Zinc to Boron Transmetallation

Figure 4.3 Molecular structure of 7j showing 50% probability ellipsoids.

Table 4.2 Selected bond length (Å) and bond angles (o) of 7j N1-N2 1.277(2) N6-N7 1.274(3) N2-C10 1.397(2) N7-C30 1.397(3) C10-N3 1.322(2) C30-N8 1.325(2) N3-N4 1.305(2) N8-N9 1.307(2) N4-B1 1.599(3) N9-B2 1.595(3) B1-F1 1.418(3) B2-F4 1.416(3) Zn1-N1 2.082(2) Zn1-N6 2.078(2) Zn1-N3 2.108(2) Zn1-N8 2.098(2) Zn1-F1 2.203(1) Zn1-F4 2.234(1) N1-Zn1-N3 76.59(6) N6-Zn1-N8 77.35(6) N3-Zn1-F1 74.47(5) N8-Zn1-F4 73.90(5) Mes(N1)-Mes(N6)a 3.641 NNCNNZn1/NNCNNZn1b 73.06 a the centroid to centroid distance of Mes rings b the angle between the planes defined by the central Zn atom and the N atoms of the formazanate backbone.

Even though the isolation of 7j as a pure compound proved to be difficult, the 1H and 19F NMR characterization data (as a mixture with 5a and some unidentified side products) can be 1 collected by treatment of 5aj with 1.5 eq of BF3·Et2O in C6D6 (Figure 4.4). The H NMR spectrum of 7j shows two sets of resonances for the mesityl groups. In the aliphatic region, five distinct resonances with integration ratio 3:3:3:3:6 indicate that the free rotation of one of the mesityl substituents was blocked, similar to what was observed for 7g.

Chapter 4

Scheme 4.4 Summary of reaction of compound 5aj with BF3

Figure 4.4 Reaction of compound 5aj with 1.5 eq of BF3·Et2O in C6D6 on NMR scale

(Formazanate)Boron Difluoride Formation via Zinc to Boron Transmetallation

4.2.5 1,2,3-Triazole Formation

In the course of optimizing the conditions for generating 6j from 5aj, we obtained an unexpected side product 9 (Scheme 4.4). After comfirming the formation of 5a and 7j by 1H

NMR spectroscopy, the NMR solution was treated with a few drops of d8-THF and the mixture was allowed to stand at room temperature for overnight. Subsequent 1H NMR analysis did not show the resonances of 5a or 7j. Workup of the reaction mixture by crystallization from toluene/hexane afforded single crystals of the new compound 8-toluene, which contains a toluene molecule in the lattice. The solvomorph 8-THF (Figure 4.5, metrical parameters in Table 4.3), which has a THF molecule in the lattice, was obtained from another independent reaction followed by recrystallization from THF/hexane mixture.

Figure 4.5 Molecular structure of 8-THF showing 50% probability ellipsoids. The THF solvent molecule and all hydrogen atoms are omitted; the formazanate ligands [1a]- are shown as wireframe for clarity. Table 4.3 Selected bond length (Å) and bond angles (o) of 8-THF N1-N2 1.302(2) N6-C30 1.408(2) N3-C7 1.352(2) N3-N4 1.304(2) C30-N7 1.347(2) Zn1-N1 2.016(2) N5-N6 1.263(3) C30-C31 1.409(3) Zn1-N4 2.012(1) N7-N8 1.318(2) C31-N9 1.351(2) Zn1-N9 2.081(2) N8-N9 1.362(2) N2-C7 1.345(2) Zn1-C31 1.998(2) N4-Zn1-N1 90.25(6) N9-Zn1-C31 105.28(7) C31-N9-N8 106.3(1) N8-N7-C30 103.3(1) N9-C31-C30 104.6(2) C31-N9-N9 111.7(2) N9-N8-N7 114.1(1) Zn1-C31-N9 129.3(1) Zn1-N9-C31 124.6(1)

A crystal structure determination shows that 8 contains two mono(formazanate)zinc units ([1a]Zn) and two anionic triazole units, which are bridging between two zinc centers by using

Chapter 4

N9 and C31 as donor atoms (Figure 4.5). The triazole has one mesityl substituent at the 2- position and one mesityldiazenyl group at the 4-position. The composition of the triazole unit is exactly the same as [1j]-; in addition, all atom connections are retained. The only additional connection is between the N atom in the cyano group and one of the terminal N atoms in the formazanate backbone. The formation of this triazole unit is unlikely to follow a traditional 1,3-dipolar cycloaddition pathway directly.15 We propose that the formation of the triazole - unit in 8 is initiated by Lewis acid (BF3) activation of [1j] , followed by dissociation of the - - [NC-BF3] unit from activated [1j] to generate the neutral carbene fragment X (Scheme 4.5). Fragment X is a resonance structure of the nitrile imine 1,3-dipole Y, which is able to undergo - a dipolar cycloaddition with [NC-BF3] to generate the final 1,2,3-triazole unit. The 1,2,3- triazole is a useful building block for more complex chemical compounds and usually synthesized form azide-alkyne Huisgen cycloaddition. The Reaction we presented here could be a potential route for 1,2,3-triazole synthesis.

Scheme 4.5 Proposed mechanism of [1j]- cyclization

4.3 Conclusion In this chapter, a high-yield synthetic route to obtain mono(formazanate)boron difluoride complexes (LBF2; compounds 6) via an uncommon zinc to boron transmetallation reaction is developed. The isolation of the six-coordinated zinc complex ([L(BF3)]2Zn, 7) from the reaction of 5 and BF3 proved the concept mentioned at the end of Chapter 3 that when substrates were introduced, the formazanate ligand can isomerize to form five-membered chelates and open space around the metal center to accommodate incoming substrates. The reaction of heteroleptic complex 5aj with BF3·Et2O not only show a intermolecular ligand exchange but also open a potential route for 1,2,3-triazole synthesis.

(Formazanate)Boron Difluoride Formation via Zinc to Boron Transmetallation

4.4 Experimental section

General Considerations. All manipulations were carried out under nitrogen atmosphere using standard glovebox, Schlenk, and vacuum-line techniques. Toluene, hexane, and pentane (Aldrich, anhydrous, 99.8%) were passed over columns of Al2O3 (Fluka), BASF R3-11- supported Cu oxygen scavenger, and molecular sieves (Aldrich, 4 Å). Diethyl ether and THF (Aldrich, anhydrous, 99.8%) were dried by percolation over columns of Al2O3 (Fluka).

Deuterated solvents were vacuum transferred from Na/K alloy (C6D6, toluene-d8, Aldrich) and stored under nitrogen.

BF3(ether) was used as received from Aldrich. NMR spectra were recorded on Varian Gemini 200, VXR 300, Mercury 400 or Varian 500 spectrometers. The 1H and 13C NMR spectra were referenced internally using the residual solvent resonances and reported in ppm relative to TMS (0 ppm); J is reported in Hz. Assignment of NMR resonances was aided by gradient- selected COSY, NOESY, HSQC and/or HMBC experiments using standard pulse sequences. Elemental analyses were performed at the Microanalytical Department of the University of Groningen.

Synthesis and Characterization

[PhNNC(C6F5)NN(BF3)Mes]2Zn 7g. A mixture of 5g (100.0 mg, 0.108 mmol), BF3·Et2O (0.04 mL, 0.32 mmol) and 10 mL of toluene was prepared. The reaction mixture was stirred at 70°C for 2 hours after which the color had changed to orange. Slow diffusion of 5 mL of hexane into the toluene solution at -30 °C for 4 days resulted in precipitation of 98 mg orange 1 crystals of 7g (0.092 mmol, 85%). H NMR (400 MHz, C6D6, 25 °C) δ 7.76 (d, 2H, J = 8.1, Ph o-H), 7.01 (t, 2H, J = 7.7, Ph m-H), 6.89 (t, 1H, J = 7.4, Ph p-H), 6.35 (s, 1H, Mes m-H),

6.14 (s, 1H, Mes m-H), 2.48 (s, 3H, Mes o-CH3), 2.29 (s, 3H, Mes o-CH3), 1.79 (s, 3H, Mes 11 19 p-CH3), B NMR (376.4 MHz, C6D6, 25 °C) δ 0.78 (s, 1B, BF3). F NMR (128.3 MHz, C6D6,

25 °C) δ -130.3 (d, 1F, J = 23.9, C6F5 m-F), -136.7 (d, 1F, J = 23.9, C6F5 m-F), -148.4 (s, 3F,

BF3), -152.7 (t, 1F, J = 21.8, C6F5 p-F), -161.6 (td, 1F, J = 22.6, 7.3, C6F5 o-F), -163.0 (td, 1F, 13 J = 22.7, 7.7, C6F5 o-F). C NMR (100.6 MHz, C6D6, 25 °C) δ 148.6 (Ph i-C), 146.0-145.0

(m, C6F5), 143.5-142.6 (m, C6F5), 140.8-140.1 (m, C6F5), 139.2 (Mes i-C), 138.6-137.6 (m,

C6F5), 136.3 (Mes o-C), 136.0 (Mes o-C), 135.9-135.2 (m, C6F5), 134.5 (Ph p-C), 132.4 (Mes p-C), 130.5 (NNCNN), 129.7 (Ph m-C), 129.1 (Mes m-C), ~128.5 (overlapped, Mes m-C),

122.4 (Ph o-C), 108.9 (td, J = 19.4, 4.0, C6F5 i-C), 19.9 (Mes p-CH3), 17.9 (Mes o-CH3). Anal.

Calcd for C44H32B2F16N8Zn: C, 49.68; H, 3.03; N, 10.53. Found: C, 50.18; H, 3.15; N, 10.18.

Chapter 4

1 NMR data of [MesNNC(CN)NN(BF3)Mes]2Zn 7j. H NMR (400 MHz, C6D6, 25 °C): δ 6.77 (s, 1H, Mes m-CH, overlapped with compound 5a), 6.70 (s, 1H, Mes m-CH), 6.15 6.70

(s, 2H, Mes m-CH), 2.54 (s, 3H, Mes CH3), 2.43 (s, 3H, Mes CH3), 1.99 (s, 3H, Mes CH3), 19 1.82 (s, 3H, Mes CH3), 1.72 (s, 6H, Mes o-CH3). F NMR (128.3 MHz, C6D6, 25 °C): δ - 149.1 ppm.

Crystal structure determination

Suitable crystals reported in this chapter were mounted on a cryo-loop in a drybox and transferred, using inert-atmosphere handling techniques, into the cold nitrogen stream of a Bruker D8 Venture diffractometer. The final unit cell was obtained from the xyz centroids of 9950 (7g), 9328 (7j), and 9677 (8-THF) reflections after integration. Intensity data were corrected for Lorentz and polarisation effects, scale variation, for decay and absorption: a multiscan absorption correction was applied, based on the intensities of symmetry-related reflections measured at different angular settings (SADABS).16 The structures were solved by direct methods using the program SHELXS.17 The hydrogen atoms were generated by geometrical considerations and constrained to idealised geometries and allowed to ride on their carrier atoms with an isotropic displacement parameter related to the equivalent displacement parameter of their carrier atoms. Structure refinement was performed with the program package SHELXL.17 Crystal data and details on data collection and refinement are presented in the following tables.

Crystallographic data 7g 7j 8-THF chem formula C44H32B2F16N8Zn C40H44B2F6N10Zn C88H94N18O2Zn2 Mr 1063.76 865.84 1566.55 cryst syst monoclinic triclinic monoclinic color, habit orange, platelet black, block red, block size (mm) 0.20 x 0.16 x 0.03 0.16 x 0.12 x 0.07 0.22 x 0.19 x 0.06 space group P21/c P-1 P21/c a (Å) 28.1225(14) 11.6518(8) 14.5130(5) b (Å) 20.6482(10) 13.2253(8) 19.2848(7) c (Å) 15.3728(7) 14.0756(9) 14.5017(4)  (°) 85.167(2) β (°) 93.868(2) 85.244(2) 107.0814(11)  (°) 74.634(2) V (Å3) 8906.3(7) 2079.9(2) 3879.7(2) Z 8 2 2 -3 calc, g.cm 1.587 1.383 1.341 Radiation [Å] Mo Kα 0.71073 Mo Kα 0.71073 Mo Kα 0.71073 -1 µ(Mo K), mm 0.663 0.66 0.681 -1 µ(Cu K), mm F(000) 4288 896 1648 temp (K) 100(2) 100(2) 100(2)  range (°) 5.878-52.737 2.91-27.19 2.94-27.14

(Formazanate)Boron Difluoride Formation via Zinc to Boron Transmetallation

data collected (h,k,l) -36:36, -26:26, -20:19 -14:14; -16:16: -18:18 -18:18; -24:24; -18:18 min, max transm 0.6950, 0.7456 0.6773, 0.7455 0.5155, 0.7455 rflns collected 285779 70841 132574 indpndt reflns 20514 9196 8585

observed reflns Fo  2.0 σ (Fo) 15776 7344 7295 R(F) (%) 5.26 3.53 3.56 wR(F2) (%) 11.07 6.93 8.39 GooF 1.083 0.983 1.051 weighting a,b 0.0422, 16.6302 0.0211, 2.1321 0.0355, 4.2240 params refined 1353 544 503 min, max resid dens -0.914, 2.757 -0.394, 0.419 -0.561, 0.668

4.5 Reference (1) Miyaura, N.; Suzuki, A. Chem. Rev., 1995, 95, 2457–2483. (2) (a) Baba, S.; Negishi, E. J. Am. Chem. Soc., 1976, 98, 6729–6731. (b) Phapale, V. B.; Cárdenas, D. J. Chem. Soc. Rev., 2009, 38, 1598–1607. (c) Zhou, J.; Fu, G. C. J. Am. Chem. Soc., 2003, 125, 12527– 12530. (3) Diederich, F.; Stang, P. J. Metal-catalyzed cross-coupling reactions; 2nd Ed; WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany; 2008. (4) (a) Rappoport, Z., Marek, I., The Chemistry of Organozinc Compounds; John Wiley & Sons, Ltd: Chichester, UK, 2006. (b) Paixão, M. W.; Braga, A. L.; Lüdtke, D. S. J. Braz. Chem. Soc., 2008, 19, 813-830. (5) (a) Bolm, C.; Rudolph, J. J. Am. Chem. Soc., 2002, 124, 14850–14851. (b) Schmidt, F.; Stemmler, R. T.; Rudolph, J.; Bolm, C. Chem. Soc. Rev., 2006, 35, 454–470. (c) Jimeno, C.; Sayalero, S.; Fjermestad, T.; Colet, G.; Maseras, F.; Pericàs, M. A. Angew. Chem. Int. Ed., 2008, 47, 1098–1101. (d) Bedford, R. B.; Gower, N. J.; Haddow, M. F.; Harvey, J. N.; Nunn, J.; Okopie, R. A.; Sankey, R. F. Angew. Chem. Int. Ed., 2012, 51, 5435–5438. (e) Bolm, C.; Hildebrand, J. P.; Muñiz, K.; Hermanns, N. Angew. Chem. Int. Ed., 2001, 40, 3284–3308. (6) (a) Bolshan, Y.; Chen, C.-Y.; Chilenski, J. R.; Gosselin, F.; Mathre, D. J.; O'Shea, P. D.; Roy, A.; Tillyer, R. D. Org. Lett., 2003, 6, 111–114. (b) Sperber, N.; Papa, D.; Schwenk, E.; Sherlock, M. J. Am. Chem. Soc., 1949, 71, 887–890. (c) Harms, A. F.; Nauta, W. T. J. Med. Chem., 1960, 2, 57–77. (7) (a) Lee, J.-D.; Han, W.-S.; Kim, T.-J.; Kim, S. H.; Kang, S. O. Chem. Commun., 2010, 47, 1018–1020. (b) Bott, S. G.; Hoffman, D. M.; Rangarajan, P. J. Chem. Soc. Dalton Trans., 1996, 9 1979–1982. (c) Schrock, R. R.; Parshall, G. W. Chem. Rev., 1976, 76, 243–268. (d) Juvinall, G. L. J. Am. Chem. Soc., 1964, 86, 4202–4203. (e) Moorhouse, S.; Wilkinson, G. J. Chem. Soc. Dalton Trans., 1974, 20, 2187– 2190. (8) (a) Singh, N.; Elias, A. J. Organometallics, 2012, 31, 2059-2065. (b) Tao, X.; Gao, W.; Huo, H.; Mu, Y. Organometallics, 2013, 32, 1287–1294. (c) Schnöckelborg, E.; Hartl, F.; Langer, T.; al, E. Eur. J. Inorg. Chem., 2012, 2012, 1632-1638. (d) Tsoureas, N.; Summerscales, O. T.; Cloke, F. G. N.; Roe, S. M. Organometallics, 2013, 32, 1353–1362. (9) (a) Wooles, A. J.; Lewis, W.; Blake, A. J.; Liddle, S. T. Organometallics, 2013, 32, 5058-5070. (b) Vidović, D.; Findlater, M.; Cowley, A. H. J. Am. Chem. Soc., 2007, 129, 8436–8437. (c) Trofymchuk, O. S.; Gutsulyak, D. V.; Quintero, C.; Parvez, M.; Daniliuc, C. G.; Piers, W. E.; Rojas, R. S. Organometallics, 2013, 32, 7323– 7333 (d) Tsai, Y.-C. Coord. Chem. Rev., 2012, 256, 722–758. (10) (a) Lennox, A. J. J.; Lloyd-Jones, G. C. Angew. Chem. Int. Ed., 2013, 52, 7362–7370. (b) Braga, A. A. C.; Morgon, N. H.; Ujaque, G.; Maseras, F. J. Am. Chem. Soc., 2005, 127, 9298–9307. (11) Liu, Q.; Lan, Y.; Liu, J.; Li, G.; Wu, Y.-D.; Lei, A. J. Am. Chem. Soc., 2009, 131, 10201–10210. (12) (a) Ariafard, A.; Yates, B. F. J. Am. Chem. Soc., 2009, 131, 13981–13991. (b) Clarke, M. L.; Heydt, M. Organometallics, 2005, 24, 6475–6478. (c) Westerhausen, M.; Bollwein, T.; Makropoulos, N.; Piotrowski, H. Inorg. Chem., 2005, 44, 6439–6444. (13) Jaric, M.; Haag, B. A.; Unsinn, A.; Karaghiosoff, K.; Knochel, P. Angew. Chem. Int. Ed., 2010, 49, 5451–5455. (14) Martinez, C. R.; Iverson, B. L. Chem. Sci., 2012, 3, 2191–2201. (15) (a) Meldal, M.; Tornøe, C. W. Chem. Rev., 2008, 108, 2952–3015. (b) Liang, L.; Astruc, D. Coord. Chem. Rev., 2011, 255, 2933–2945. (16) Bruker. APEX2 (v2012.4-3), SAINT (Version 8.18C) and SADABS (Version 2012/1). Bruker AXS Inc., Madison, Wisconsin, USA. 2012. (17) Sheldrick, G. Acta Crystallographica Section A 2008, 64, 112.

Chapter 4

Chapter 5

(Formazanate)Boron Complexes

A series of (formazanate)boron difluoride complexes (LBF2; 6), which have a large variety of steric and electronic properties, were synthesized and fully characterized by multinuclear NMR, UV-Vis spectroscopy, cyclic voltammetry, and X-ray crystallography. The compounds 6 show strong absorption ( = 9700 - 21300), large Stokes shift (130 – 240 nm) and long wavelength emission (> 600 nm). In addition, all the compounds 6 have two reversible ligand- based reductions to their radical anion ([6]-) and dianion ([6]-2) forms. The redox potential of the LBF2 complexes can be altered in a straightforward manner over a relative wide range (~ 530 mV) by changing the steric or the electronic properties of the formazanate framework.

Examples of (formazanate)boron diphenyl (LBPh2; 9) and (formazanate)boron dihydride

(LBH2; 10) were prepared for comparison.

Parts of this chapter have been published:

M.-C. Chang and E. Otten* “Synthesis and ligand-based reduction chemistry of boron difluoride complexes with redox-active formazanate ligands” Chem. Commun., 2014, 50, 7431-7433.

Chapter 5

Chapter 5 (Formazanate)Boron Complexes 5.1 Introduction Fluorescent dyes have attracted increasing research interest in various fields of modern research, including biological imaging1, molecular probes2, electroluminescent devices3, and photosensitizers4. An ideal fluorescent dye contains high absorption coefficients and quantum yields, a large Stokes shift, tunable absorption/emission profiles, and high chemical and photochemical stability. Fluorescent dyes with large Stokes shift have a higher chance of showing emission in the long wavelength region (> 600 nm), which is desirable for applications in laser printing, information storage, displays and solar power conversion.5 Most importantly, the fluorescent dyes emitting within the biological window (650-900 nm) are useful in bioimaging applications due to the small light scattering6, low background emission, and deep penetration into cells and tissues. Fluorescent dyes with small Stokes shifts usually suffer from self-quenching7, which limits the application in high concentration conditions and as solid state materials. A very popular family of fluorescent dyes is the boron difluoride fragment (BF2) bearing nitrogen donor ligands, which have been widely studied: these include dipyrrins (A)8, -diketiminates (B)9, anilido-pyridines (C)10, hydrazine−Schiff base linked bispyrrole (D)11, and indigo-N,N’-diarylamines (E)12 (Chart 5.1).

Chart 5.1

Boron difluoride complexes bearing dipyrrin ligands, commonly known as BODIPYs (A), are the most popular systems in this family due to their high quantum yield, tunable absorption/emission profiles, and great photo and chemical stability. The optical properties of the BODIPYs are tunable by changing the R1-R7 substituents of the ligand backbone.8,13 In

(Formazanate)Boron Complexes addition, the F- substituents of the boron center can also be replaced by aryl, alkyl, and alkoxide groups.14 Even though the largest documented Stokes shift of the mono-BODIPYs is 185 nm15, for most of the BODIPYs, the Stokes shift is usually less than 40 nm. One of the useful strategies to increase the Stokes shift of fluorescent dyes is to incorporate BODIPYs into an energy transfer fragment in which the energy absorbed by BODIPY unit is transferred to a second fluorophore.1a,4a,14c,1 A drawback, however, is that this strategy usually needs large synthetic efforts.

In order to develop new fluorescent dye systems, modifications of the BODIPY fragment have been developed. Boron difluoride complexes bearing -diketiminates ligands (B), which have the same ligand backbones as BODIPYs, are exhibit strong absorption. In addition, compounds B show larger Stokes shifts of around 80nm in comparison to the BODIPYs. A desymmetrized -diketiminates analogue, the anilido-pyridine (C) framework, reported by the group of Piers and Heyne in 2011 and shows large Stokes shifts of 90-120 nm.10

In recent years, the ligand design for boron difluoride complexes has become more and more elaborate. Several ligand systems are capable of coordinating two BF2 fragments such as the hydrazine−Schiff base linked bispyrrole (D) and the indigo-N,N’-diarylamines (E). The compound D is highly fluorescent ( > 0.99) in the solution state with Stokes shifts around 40 nm.11 In addition, compound D shows emission in thin film and solid powder. The compounds E are a class of redox-active and near-infrared dyes that show Stokes shifts of 30- 70 nm.12

While our work on the synthesis and characterization of formazanate boron complexes was in progress, Gilroy and co-workers reported a series of very similar compounds.16 Their work focused on the electronic effect of the formazanate ligand and indicated that the (formazanate)boron difluoride complexes usually have larger Stokes shifts (80-150 nm) than other BF2 complexes with N-donor ligands. In addition, the asymmetrical (formazanate)boron difluoride complexes have stronger emission intensities than the symmetrical system. In 2015, the application of the (formazanate)boron difluoride complexes in cell imaging17 and electrochemiluminescence16b were reported by the same group.

Here the synthesis and characterization of three types of (formazanate)boron complexes

(LBF2, 6; LBPh2, 9; LBH2, 10;) are reported. The formazanate ligands we used here contain large varieties of electronic and steric effects. The optical and redox properties of the (formazanate)boron complexes are discussed.

Chapter 5

5.2 Synthesis of Formazanate Boron Complexes

5.2.1 Formazanate Boron Difluoride

At the end of Chapter 3, we described a new method for the synthesis of (formazanate)boron difluoride complexes (LBF2, 6) via a transmetallation reaction starting from bis(formazanate)zinc complexes (L2Zn, 5) (Scheme 5.1, Path A). In 2014, a direct method to access (formazanate)boron difluoride complexes was developed by Gilroy and co-workers (Scheme 5.1, Path B).16c-d Both methods can be used to obtain the desired products with moderate yield (60-90%); therefore, both methods were used in this research. In addition, the direct method was used to synthesize compound 6k and 6l from the di-formazan ligand system (1k and 1l). While this thesis was being prepared, a report describing the same synthesis of 6k and 6l was published.18 The formation of complexes 6 was confirmed by 11B NMR and 19F NMR spectroscopy (as an example the spectra for 6a are shown in Figure 5.1). In the 11B NMR spectra, the diagnostic 1:2:1 triplets with a chemical shift between -0.7 and - 2.3 ppm indicate that the boron centers are four-coordinated and bound with two fluorides.19 19 In the F NMR spectra, 1:1:1:1 quartets between -145.6 and -159.4 ppm with JB-F = 20 – 30

Hz are observed that are consistent with the presence of a BF2 moiety.

Scheme 5.1 Indirect (Path A) and direct (Path B) methods of (formazanate)boron difluoride (6) synthesis.

(Formazanate)Boron Complexes

Figure 5.1 1H NMR (top), 19F NMR (bottom left) and 11B NMR (bottom right) spectra of 6a

(in C6D6). In order to probe the influence of Lewis acids on the optical and electrochemical properties of (3-cyanoformazanate)boron difluoride complexes, compound 6i was reacted with tris(pentafluorophenyl)borane (B(C6F5)3) in toluene solution. The formation of the desired 19 19 acid-base product (6i-B(C6F5)3) was confirmed by F NMR spectroscopy. The F NMR spectrum shows four resonances at -125.7, -134, -155, and -163 ppm with integration ratio of 19 2:6:3:6, which are assigned to BF2, m-C6F5, p-C6F5, and o-C6F5, respectively. The F resonances of the B(C6F5)3 fragment are located at similar positions as seen in 1j-B(C6F5)3, which was described in Chapter 2, and support the formation of the desired acid-base adduct

6i-B(C6F5)3. This acid-base interaction was further confirmed by the single crystal X-ray crystallography (Figure 5.2, metrical parameters in Table 5.1). Even though the crystal of 6i-

B(C6F5)3 we obtained were only of low quality, and as a consequence there are relatively large errors associated with the metrical parameters, the difference of the bond lengths between 6i and 6i-B(C6F5)3 are similar with the case of the free formazan ligand 1j and 1j-

B(C6F5)3. Upon coordination to B(C6F5)3, the bond length of N5-C8 and C8-C7 are shortened and elongated, respectively.

Chapter 5

Figure 5.2 Molecular structure of 6i-B(C6F5)3. Thermal ellipsoids are shown at 50% probability, and hydrogen atoms are removed for clarity. Table 5.1 Selected bond lengths (Å) of 6i and 6i- B(C6F5)3 a 6i 6i-B(C6F5)3 N1-N2 1.293(1) 1.308(6) N2-C7 1.341(1) 1.344(7) N3-C7 1.338(2) 1.339(7) N3-N4 1.298(1) 1.302(6) C7-C8 1.444(2) 1.421(8) C8-N5 1.145(2) 1.153(7) N5-B2 - 1.614(7) N1-B1 1.579(2) 1.561(8) N4-B1 1.573(2) 1.560(8) a: Reported data from Gilroy and co-workers.16c

5.2.2 Formazanate Boron Diphenyl and Formazanate Boron Dihydride

The synthesis of (formazanate)boron diphenyl complexes (LBPh2, 9a) and

(formazanate)boron dihydride complexes (LBH2, 10a) was achieved by reacting formazan ligand with triphenyl borane (BPh3) and borane dimethylsulfide complex (BH3 ∙ SMe2), respectively (Scheme 5.2). The formation of 9a was confirmed by the 1H NMR spectrum, which shows a disappearance of the NH resonance of the formazan 1a and a clean formation of a single product (based on the CH3 resonance). Unlike the case of 9a, the formation of 10a from a reaction of 1a and borane dimethylsulfide complex is not clean. The 1H NMR spectrum of the crude product shows resonances of 10a, unreacted 1a and several unidentified species (based on the CH3 resonance). Fortunately, the pure compound 10a can be isolated by column chromatography with a yield of 30 %. The 1H NMR spectrum of 10a is very similar

(Formazanate)Boron Complexes to the spectrum of 6a, but the spectrum of 10a shows an additional broad resonance at 3.67 ppm, which is assigned to the BH2 fragment (Figure 5.3).

Scheme 5.2 Synthesis of LBPh2 (9a) and LBH2 (10a)

1 Figure 5.3 H NMR spectrum of 10a in C6D6

5.3 Characterization of Formazanate Boron Complexes 5.3.1 X-ray Crystallographic Analysis The single crystals suitable for X-ray diffraction analysis of 6a, 6b, 6e, 6i, and 6j were obtained by recrystallization in heptane or hexane solution. The metrical parameters and crystal structure of 6e are shown in Table 5.2 and Figure 5.4, respectively. The solid-state structures of all compounds show four-coordinate boron centers, which are bound to a formazanate ligand through two terminal nitrogen atoms to form a six-membered chelate ring. In all the structures, the C-N and N-N bonds are in the range of 1.33-1.36 Å and 1.29-1.31 Å, respectively, which are similar to the bond lengths of the bis(formazanate)zinc complexes (compounds 5 in Chapter 3). The C-N and N-N bond lengths in the symmetrical derivatives (with R1 = R5) suggest that the formazanate backbone is fully delocalized. In the case of 6e 1 5 (with R = Ph and R = C6F5), the N1-N2 bond length (1.299(2) Å) is shorter than N3-N4 5 (1.329(2) Å). This is due to the electron-withdrawing C6F5- group at the R position, which

Chapter 5 makes the negative charge of the formazanate anion is partially localized at N4. A similar effect can be found in the zinc complex (MesNNC(p-tolyl)NNC6F5)2Zn (5f), which was described in Chapter 3. The partially localized negative charge at the N4 position also makes the bonding between N4 and B1 (1.548(2) Å) is stronger than bonding between N1 and B1 (1.577(3) Å). All the boron centers of 6a, 6b, 6e, 6i and 6j show a distorted tetrahedral geometry, and the boron center is slightly displaced from the formazanate backbone (N1-N2- N3-N4). The displacements are in the range of 0.17-0.60 Å, which is similar to the data (0.02- 0.58 Å) reported by Gilroy and co-workers.16c-d The displacement of the boron atom from formazanate backbone makes the two fluoride atoms inequivalent in the solid state. However, the 19F NMR spectra only show one type of resonance, which suggests that a flip of the boron center in formazanate six-membered chelate ring is facile. The N-aryl substituents are slightly twisted (around 23 in 6i), with respect to the N1-N2-N3-N4 plane. The solid-state structure of 6a, 6b and 6e exhibit larger twisting of the N-aryl substituents with a dihedral angle of 42 to 50. In the case of 6j, due to the methyl groups at the ortho positions of the N-aryl substituents, the dihedral angle is around 88, which means that the mesityl group is virtually perpendicular to the plane of the formazanate ligand backbone.

Single crystals suitable for X-ray diffraction analysis of 9a and 10a were obtained by recrystallization from hexane solution (Figure 5.4, metrical parameters in Table 5.2). The metrical parameters of the formazanate ligand ([1a]-) of 9a and 10a are very similar to those seen in 6a. The N-N and N-C bond lengths within formazanate ligand show that both 9a and 10a have a fully delocalized backbone. The major difference between these three structures (6a, 9a and 10a) is the B-N bond lengths. The B-N bond length of 6a (av. 1.556 Å) is shorter than the B-N bond lengths of 9a (av. 1.598 Å) and 10a (av. 1.568 Å). This is due to the strong electron-withdrawing F- groups make the boron center of 6a more electron-deficient than 9a and 10a and then strengthen the bonding between electron-deficient boron and electron-rich nitrogen. The longer B-N bond lengths of 9a than 10a is due to the steric repulsion between the phenyl group and the formazanate ligand.

(Formazanate)Boron Complexes

Figure 5.4 Molecular structure of 6e (left), 9a (middle) and 10a (right). Thermal ellipsoids are shown at 50% probability, and hydrogen atoms are removed for clarity.

o Table 5.2 Selected bond length (Å) and bond angles ( ) of LBF2 (6), LBPh2 (9a) and LBH2 (10a) complexes 6a 6b 6e 6i 6j 9a 10a N1-N2 1.308(1) 1.299(2) 1.294(1) 1.306(1) 1.312(2) 1.318(1) 1.293(1) N3-N4 1.308(1) 1.329(2) 1.295(1) 1.309(1) 1.304(2) C7-N2 1.346(1) 1.361(2) 1.344(1) 1.345(1) 1.350(3) 1.339(1) 1.344(1) C7-N3 1.343(1) 1.329(2) 1.340(2) 1.347(1) 1.346(3) N1-B1 1.559(2) 1.577(3) 1.573(2) 1.599(2) 1.568(3) 1.552(1) 1.566(2) N4-B1 1.552(2) 1.548(2) 1.580(2) 1.596(1) 1.567(3) N2-C7-N3 124.2(1) 123.4(1) 124.6(1) 129.4(1) 127.6(1) 123.1(1) 123.1(2) N1-B1-N4 102.40(9) 100.2(1) 101.5(1) 105.87(9) 103.8(1) 97.92(8) 99.5(2) Boron 0.500 0.588 0.511 0.171 0.294 0.685 0.681 displacementa 47.98, 47.54, 46.06, 37.24, Dihedral 24.16, 42.00, 47.97 50.09, 88.11 52.04, 35.82, anglesb 22.35 7.01 13.05 15.36 14.04 aDistance between B1 and N1-N2-N3-N4 plane. bAngles between the plane defined by N1, N4, and C7 aryl substituents and the N1-N2-N3-N4 plane.

The single crystals suitable for X-ray diffraction analysis of 6k and 6l were obtained by slow evaporation of the DCM solution. The crystal structures and metrical parameters are shown in Figure 5.5 and Table 5.3, respectively. The formazanate backbone of 6k and 6l shows very similar metrical parameters to 6a. The N-N and N-C bond lengths of 6k and 6l are in the range of 1.30 to 1.32 Å and 1.33 to 1.34 Å, respectively, which suggest that the negative charges are fully delocalized in the formazanate backbone.

Chapter 5

Figure 5.5 Top view (up) and side view (bottom, Ph groups are shown as wireframe for clarity) of the molecular structures of 6k (left) and 6l (right). The structures are shown at 50% probability, and hydrogen atoms are removed for clarity.

Table 5.3 Selected bond length (Å) and bond angles (o) of 6k and 6l 6k 6l N1-N2, N3-N4, 1.307(4), 1.305(4), N1-N2, N3-N4, 1.307(2), 1.309(2), N5-N6, N7-N8 1.324 (5), 1.316 (5) N5-N6, N7-N8 1.316(2), 1.305(2) C7-N2, C7-N3, 1.336(4), 1.342(4), C7-N2, C7-N3, 1.342(2), 1.338(2), C20-N6, C20-N7 1.333(4), 1.339(4) C26-N6, C26-N7 1.341(2), 1.345(2) N1-B1, N4-B1, 1.547(4), 1.556(5), N1-B1, N4-B1, 1.564(2), 1.557(2), N5-B2, N8-B2 1.547(5), 1.537(4) N5-B2, N8-B2 1.559(2), 1.561(2) C7-C14, C17-C20 1.481(5), 1.486(5) C7-C8, C12-C26 1.485(2), 1.478(2) C14-C15, C15-C16, 1.392(5), 1.377(5), C8-C9, C9-C10, 1.396(2), 1.392(3), C16-C17, C17-C18, 1.388(4), 1.386(5), C10-C11, C11-C12, 1.390(2), 1.394(2), C18-C19, C19-C14 1.388(5), 1.387(4) C12-C13, C13-C8 1.395(2), 1.387(2) N2-C7-N3, 126.6(3), N2-C7-N3, 126.3(1), N6-C20-N7 126.0(3) N6-C26-N7 124.3(1) N1-B1-N4, 105.4(3), N1-B1-N4, 104.6(1), N5-B2-N8 103.7(3) N5-B2-N8 102.3(1) Boron Boron displacementb 0.163, 0.493 0.308, 0.580 displacementb 31.68, 24.81, 10.34, 38.81, 18.74, 8.82, Dihedral anglesc Dihedral anglesd 49.46, 27.52 17.41 26.98, 40.15, 11.81 aOnly one of the two independent molecules was shown. bDistance between B1/B2 and N1- N2-N3-N4/ N5-N6-N7-N8 plane. cAngles between the plane defined by N1, N4, and C7 aryl substituents and the N1-N2-N3-N4 plane; N5, N8, and C26 aryl substituents and the N5-N6- N7-N8 plane.

100

(Formazanate)Boron Complexes

5.3.2 Absorption and Emission Spectroscopy The optical properties of the formazanate boron complexes (compounds 6, 9 and 10) were studied by UV-Vis absorption and emission spectroscopy in THF solution (Figure 5.6 and Table 5.4). All compounds show medium to strong absorption bands with extinction coefficients between 9700 and 30200 L·mol-1·cm-1 in the visible range of the spectrum (between 410 and 520 nm). The spectral trends observed here are very similar to the bis(formazanate)zinc complexes (compounds 5 in Chapter 3). Introducing an electron- donating tBu- group (6b) instead of a p-tolyl moiety (6a) results in a blue shift in the absorption maximum (517 and 473 nm for 6a and 6b, respectively). In the case of introducing 3 an electron-withdrawing C6F5- group (6g, 6h) at R position instead of a p-tolyl moiety (6c, 6f), the absorption maximum shows blue shift also (431 nm for 6g vs. 464 nm for 6c and 414 1 5 nm for 6h vs. 460 m for 6f). Upon substitution of phenyl for mesityl at R or R position, max is blue-shifted. For example, 6i shows a max at 489 nm and 6j has a max at 428 nm. This likely results from the perpendicular orientation of the mestiyl group, which doesn’t allow conjugation between the formazanate backbone and the N-aryl substituent, thereby limiting the length of the -conjugated system.16cd Introducing tris(pentafluorophenyl)borane to compound 6i results in no change of max and lowers the extinction coefficient of the -1 -1 characteristic absorption band in toluene (6i: 502 nm(30400 M cm ); 6i(B(C6F5)3): 502 nm(20582 M-1 cm -1) ). The UV-Vis absorption spectra of the di-formazanate system (6k and 6l) are very similar to 6a but with higher extinction coefficients. Especially in the case of 6k, the extinction coefficient of 6k is two times larger than 6a. Replacing the fluoride substituents with phenyl anions (9a: 505 nm) or hydrides (10a: 545 nm) results in blue- and redshift of the absorption spectra.

Similar to other reported boron difluoride complexes (A-E), compounds 6 were shown to be emissive, with emission wavelengths em in the range of 580 nm to 670 nm in THF solution (Figure 5.6). The reported (formazanate)boron difluoride complexes usually show large Stokes shifts (100 – 150 nm).16abd For most of our compounds, the Stokes shift is even larger

(> 150 nm). For example, 6g shows max at 414 nm and em at 650 nm, a Stokes shift of 236 nm, which, to the best of our knowledge, is the highest value reported to date for the class of compounds. The large Stokes shift of compounds 6c-h might be due to the unsymmetrical formazanate backbone. Introducing a tris(pentafluorophenyl)borane to compound 6i results in no shift of max (502 nm) but redshift of em (from 586 nm to 632 nm), which increases the Stokes shift from 84 nm to 130 nm in toluene solution. These results suggest that the emission

101

Chapter 5 profile of (3-cyanoformazanate)boron difluoride complexes is probably tunable by introducing a Lewis acid. Comparing with 6a, the emission spectra of 6k and 6l are red- and blue-shifted, respectively. In addition, the quantum yields of 6k and 6l are 3 and 6.5 times higher than 6a. Replacing the fluoride substituents of 6a with phenyl (9a) or hydride (10a) substituents results in a weak emission at around 690 nm and larger Stokes shifts (9a: 182; 10a: 143 nm). The larger Stokes shift of 9a and 10a than 6a is consistent with the results of the BODIPY system.14a The long-wavelength (more specifically NIR: 650-900 nm) emission of 9a and 10a is particularly useful for in vivo imaging application.

Figure 5.6 UV-Vis absorption spectra (left) and normalized emission spectra (right) of 6c, 6f, and 6g. Data were collected in [6] = 10-5 M dry THF solution. The excitation wavelength of emission spectra is at 473 nm. Table 5.4 Optical Properties of formazan boron complexes in THF    c Quantum Stokes R1 R3 R5 max em (nm) (M-1 cm -1) (nm) yield (%) shift (nm) 6a Ph p-Tol Ph 517 13736 641 0.2 124 6b Ph tBu Ph 473 13220 643 < 0.1 170 6c Ph p-Tol Mes 464 21306 617 < 0.1 153 6e C6F5 p-Tol Ph 482 12523 640 < 0.1 158 6f C6F5 p-Tol Mes 460 12139 610 0.1 150 6g Ph C6F5 Mes 431 13702 612 < 0.1 181 6h C6F5 C6F5 Mes 414 11852 650 < 0.1 236 6ia Ph CN Ph 489 25400 585 5 96 6iab Ph CN Ph 502 30400 586 15 84 6j Mes CN Mes 428 9662 643 0.2 215 b 6i(B(C6F5)3) Ph CN(B(C6F5)3) Ph 502 20582 632 0.5 130 6k Ph2-Ph-Ph2 523 30227 664 0.6 141 6l Ph2-mPh-Ph2 508 16768 631 1.3 123 9a Ph p-Tol Ph 505 17788 687 - 182 10a Ph p-Tol Ph 545 15051 688 - 143 a: Reported data from Gilroy and co-workers.16c b: Data was collected in Toluene. c: The samples were excited at 473 nm.

102

(Formazanate)Boron Complexes

5.3.3 Redox Chemistry The redox chemistry of (formazanate)boron difluoride complexes (6) was studied by cyclic voltammetry (CV) in THF or dichloroethane (DCE) solution under nitrogen atmosphere. The results of cyclic voltammetry studies are summarized in Table 5.5 and Figure 5.7. All the (formazanate)boron difluoride complexes have two (quasi)reversible redox processes, which indicates that complexes 6 are capable of accepting one or two electrons to form the 1- or 2- - - -2 -2 electron reduction products (radical anion (LBF2 ; [6] ) and dianion (LBF2 ; [6] )). It is also very clear to see that the redox potentials of LBF2 complexes can be altered over a wide range (up to 530 mV) by changing substituents. The trend of redox potentials of 6 can be rationalized by the various substituents on the formazanate ligands. When the substituent at the R3 position was changed from p-tolyl (6a) to t-butyl (6b) or CN- (6i), the redox potential shifts to negative or positive direction, respectively. That is because the t-butyl (6b) is an electron-donating group, and the CN- (6i) is an electron-withdrawing group in comparison with the p-tolyl (6a) group. A similar result can be found in the case of 6c (R3 = p-tolyl) and 3 6g (R = C6F5). The former has a relatively more negative redox potential than the latter. 1 When the R substituents was changed from Ph- to a more electron-withdrawing -C6F5 group (6a vs. 6e, 6c vs. 6f and 6g vs. 6h), the redox potentials also shift in the positive direction by around 130-180 mV. Changing a phenyl group to mesityl group in formazanate ligands (6a vs. 6c, 6f vs. 6e and 6i vs. 6j) will shift redox potentials to more negative values by around 240 mV. In comparison to the phenyl substituent, the mesityl substituent is not coplanar with the backbone of the formazanate ligand. This will break the conjugation with the formazanate backbone, making the (formazanate)boron difluoride complexes harder to be reduced. Surprisingly, introducing a tris(pentafluorophenyl)borane to compound 6i results in a very small shift of the redox potentials from -0.65 V and -1.76 V (6i) to -0.67 V and -1.75 V

(6i(B(C6F5)3)). The little influence from the Lewis acid on redox potentials might be due to the high concentration of electrolyte: these data may suggest that the B(C6F5)3 group does not interact strongly anymore in the high ionic strength solvent system used for collecting CV data. Replacing the BF2 fragment (6a) with a BPh2 (9a) or BH2 (10a) unit results in the redox potential shifting to more negative values. That is because the strong electron-withdrawing fluoride substituents were replaced by electron-donating hydride or phenyl substituents. In addition, the similar redox potentials for 9a and 10a indicates that phenyl and hydride have a similar electron donating ability in our system.

103

Chapter 5

Figure 5.7 Cyclic voltammogram of 6 recorded at 0.1 V/s in 1.5 mM THF solution containing 0.1 M tetrabutylammonium hexafluorophosphate.

Figure 5.8 Cyclic voltammograms of 6k, and 6l recorded at 0.1 V/s in 1.5 mM THF solution containing 0.1 M tetrabutylammonium hexafluorophosphate.

104

(Formazanate)Boron Complexes

+ a Table 5.5 Electrochemical Data (V vs Fc/Fc ) of formazan boron complexes (LBX2)  1 3 5 0/-1 -1/-2 R R R LBX2 LBX2 V 6a Ph p-Tol Ph -0.98 -2.06 1.08 6b Ph tBu Ph -1.08 -2.21 1.13 6c Ph p-Tol Mes -1.19 -2.34 1.15 6e C6F5 p-Tol Ph -0.85 -1.99 1.14 6f C6F5 p-Tol Mes -1.01 -2.26 1.25 6g Ph C6F5 Mes -1.02 -2.25 1.23 6h C6F5 C6F5 Mes -0.84 -2.17 1.33 6i Ph CN Ph -0.66 -1.83 1.17 6ib Ph CN Ph -0.65 -1.76 1.10 b 6i(B(C6F5)3) Ph CN(B(C6F5)3) Ph -0.67 -1.75 1.08 6j Mes CN Mes -0.90 -2.41 1.51 6k Ph2-Ph-Ph2 -0.95 -2.0~-2.3 - 6l Ph2-mPh-Ph2 -0.94 -2.0~-2.4 - 9a Ph p-Tol Ph -1.35 -2.26 0.91 10a Ph p-Tol Ph -1.26 -2.20 0.94 Cyclic voltammetry experiments were collected in THF solution containing 1.5 mM LBR2 and 0.1 M tetrabutylammonium hexafluorophosphate as supporting electrolyte at a scan rate if 0.1 V/s. All voltammograms were referenced internally against the ferrocene/ferrocenium redox couple. a: X = F, Ph and H for compound 6, 9 and 10, respectively. b: Data was collected in DCE solution.

The cyclic voltammogram of compounds 6k and 6l show similar redox couples to 6a (Figure 5.8). The first redox couple of 6k and 6l are two overlapped one-electron processes, which are located at around -0.95 and -0.94 V (vs. Fc/Fc+), respectively. These two very close one- electron processes of the first redox couple of 6k and 6l are supported by the slightly larger peak to peak separations of 6k (0.287 V) and 6l (0.268 V) than 6a (0.222 V). If the first redox couple of 6k and 6l are two-electron processes, a smaller peak to peak separation is expected. The second redox couple of 6k and 6l are two very close one-electron processes, which are located in the range of -2.0 V- -2.4 V (vs. Fc/Fc+). In the case of 6l, the difference between the two one-electron processes of the second redox couple is larger than the case of 6k.

Another interesting influence from substituents of the formazanate ligand is the potential difference between the 1st and 2nd redox-couple. The potential difference between the 1st and nd 2 redox couple is around 1.0~1.5 volt for all LBF2 complexes. In general, complexes 6 with smaller conjugation length in the formazanate ligand have the larger potential difference. Take 6i and 6j as examples, the mesityl substitution in 6j is perpendicular to the formazanate backbone; therefore it has a shorter conjugation system than 6i which results in a larger potential difference.

105

Chapter 5

5.4 Reduced Products: mono-Formazanate Ligands 5.4.1 Synthesis and Crystal Structures

According to the results of cyclic voltammetry experiments, all the singly reduced LBF2 complexes ([6]X; X: counter cation) can be synthesized by using cobaltocene (Cp2Co) as the reducing agent. However, not all products from these reactions proved easy to obtain in crystalline form, and therefore reactions with other reductants (decamethylcobaltocene * (Co 2Co) or Na/Hg alloy) were also attempted. It was possible to obtain crystals suitable for X-ray diffraction for a representative series of compounds ([6a]-, [6c]-, and [6g]-): selected metrical parameters in Table 5.6 and molecular structure of [6c][Na(15-crown-5)] in Figure 5.9). While some of the crystal structure determinations suffered from disorder in the cationic moiety, the anionic (formazanate)boron fragments are well-defined. Significant changes were observed in the metrical parameters of the radical anions in comparison to the neutral precursors. The most obvious change is the elongation of the N-N bond (from 1.30 - 1.31 Å to 1.34 - 1.37 Å) and shortening of the N-C bond (from 1.34 - 1.36 Å to 1.33 - 1.34 Å ) in formazanate backbone. Based on the bis(formazanate)zinc complexes (5) reported in the Chapter 3, the similar changes in the N-N and N-C bond lengths observed for compounds [6]- suggesting that the extra electron is localized in the formazanate backbone and occupies an N-N anti-bonding orbital. After reduction, the electron density (charge) of the formazanate ligand is increased which makes the bonding between the boron center and the terminal nitrogens of the ligand stronger. This is supported by shortening of the B-N bond lengths; for example, the B-N bond lengths of 6a (1.559(2) and 1.552(2)) are longer than [6a]Cp2Co (1.532(3), 1.536(3)).

In addition to the changes of the N-N and B-N bond lengths, the reduction of formazanate ligand decreases the boron displacement to less than 0.1 Å (for [6a][Cp2Co], and [6g][Cp2Co]) and decreases the twisting of N-phenyl substituents resulting in a more flat (or co-planar) geometry, which increases the conjugation system of ligand backbone and stabilizes the reduced product. The large boron displacement of [6c][Na(15-Crown-5)] (0.376 Å) is due to the interaction between the BF2 moiety and the Na(15-crown-5) counter cation (F-Na: 2.383(4) and 2.419(4) Å) which increases the steric hindrance around the boron center in comparison to other cases such as [6a][Cp2Co]. For those compounds that have N-C6F5 or N-Mes substituents ([6c][Na(15-crown-5)], and [6g][Cp2Co]), the twisting is still up to 60-80.

106

(Formazanate)Boron Complexes

3240 3290 3340 3390 3440 Magnetic Field Strength (G)

Figure 5.9 Left: The molecular structure of [6c][Na(15-Crown-5)] are showing 50% probability ellipsoids. All the all hydrogen atoms are omitted, and the 15-crown-5 group is shown as wireframe for clarity. Right: EPR of [6a][Cp2Co] in frozen THF solution (20K).

o Table 5.6 Selected bond length (Å) and bond angles ( ) of singly reduced LBF2 complexes ([6]X)

[6a][Cp2Co] [6c][Na(15-Crown-5)] [6g][Cp2Co] N1-N2, N3-N4 1.362(2), 1.359(2) 1.361(7), 1.374(7) 1.374(2), 1.356(2) C7-N2, C7-N3 1.336(2), 1.330(2) 1.337(7), 1.336(8) 1.329(3), 1.333(2) N1-B1, N4-B1 1.532(3), 1.536(3) 1.535(8), 1.505(7) 1.533(3), 1.512(3) F-Na1 - 2.419(4), 2.383(4) - N1-N2-C7, N4-N3-C7 116.8(2), 117.2(2) 117.1(5), 115.0(5) 116.0(2), 114.7(2) N2-C7-N3 129.3(2) 128.9(5) 131.2(2) N1-B1-N4 108.4(2) 107.1(5) 108.5(2) Boron displacementa 0.090 0.376 0.075 Dihedral anglesb 4.96, 0.72, 5.59 27.29, 65.10, 27.35 12.64, 83.44, 62.62 aDistance between B1 and N1-N2-N3-N4 plane. bAngles between the plane defined by N1, N4, and C7 aryl substituents and the N1-N2-N3-N4 plane.

5.4.2 Absorption and Emission Spectroscopy

UV-Vis absorption spectroscopy of anionic mono(formazanate) difluoride compounds ([6a]-, [6b]-, [6c]-, [6g]- and [6j]-) provides additional evidence for ligand-based reduction and formation of the dianionic ligand radical (L2-) (Figure 5.10, Table 5.7). In all cases of anionic mono(formazanate) difluoride compounds ([6a]-, [6b]-, [6c]-, [6g]- and [6j]-), the major absorptions locate at shorter wavelength than the neutral parent complexes with an additional (also weaker) absorptions at longer wavelength than the neutral parent complexes. For example, the neutral compound 6a shows a single broad absorption in the visible at 521 nm. In the case of [6a]-, absorption is at longer wavelength (716 nm) and at 454 nm with shoulders at 674 and 431 nm, respectively. These absorption bands are in agreement with the

107

Chapter 5 presence of reduced formazanate ligands reported in Chapter 3. In addition to the absorption spectrum, emission spectrum of [6a]-, which is very similar with the neutral parent complex 6a, was also collected. The similar emission spectra of [6a]- and 6a might suggest that [6a]- is not emitting at all and the collected emission spectrum of [6a]- is due to decomposition of [6a]- back to 6a.

Figure 5.10 UV-Vis spectra of 6a and [6a]Cp2Co in THF

Table 5.7 UV-Vis data of compounds [6]- in THF 1 3 5   R R R max (nm) (M-1 cm -1) 6a Ph p-Tol Ph 454, 716 20893, 7372 6b Ph tBu Ph 465, 668 15164, 6620 6c Ph p-Tol Mes 448, 646 19283, 7178 6g Ph C6F5 Mes 399, 596 10580, 4056 6j Mes CN Mes 378, 517 14657, 1069

5.4.3 EPR Spectra and DFT Calculations

The EPR spectrum of [6a][Cp2Co] shows a broad EPR signal in frozen THF solution (77K) with g-value of ~2, which suggests a ligand-based radical in [6]- (Figure 5.9). Unlike the similar verdazyl radical system, which usually shows a clearly resolved hyperfine structure in the EPR spectra20, the EPR spectrum of [6]- lacks hyperfine structure. This might due to the more diffuse spin distribution of [6]- than in the related verdazyl radical system. The calculated spin density of model compounds F (for [6]-), G and H (for verdazyl radical) are summarized in Chart 5.2 and Table 5.8. The DFT calculations were carried out on the anionic fragment [6]- (the cation was omitted) at the B3LYP/6-31G(d) level in the gas phase. The calculated metric parameters, such as N-N and B-N bond lengths, of the model compound F

108

(Formazanate)Boron Complexes are in good agreement with the crystal structure of [6a][Cp2Co]. The highest Mulliken atomic spin density of compound F is found on the four nitrogen atoms with a total amount of 1.01. The verdazyl radical compound G and H show similar spin distribution, but with larger spin density at the nitrogen atoms (1.21 for G and 1.07 for H) than calculated for F. In addition, the spin densities at the C6 and C7 position of F (0.040 and 0.041) are larger than the spin density of G (0.012 and 0.012). The larger spin density at the C6 and C7 position of F than G and smaller spin density at the nitrogen atoms of F than G and H suggest that the spin distribution of [6]- is more diffuse than the verdazyl radical species. The more diffuse spin density of F than G and H also influences the hyperfine coupling constant (hfcc) of the system: compound F shows smaller calculated hfcc than compound G and H (Table 5.9).

Chart 5.2 Top: Structure (left), SOMO (middle) and spin density (right) of model compound F. Bottom: Structures of verdazyl radical G and H.

Table 5.8 Mulliken atomic spin densitya for F, G, and H F G H F G H N1 0.179 0.191 0.167 C8 0.028 0.036 0.032 N2 0.322 0.412 0.364 C9 -0.027 -0.032 -0.029 C3 -0.145 -0.169 -0.159 C10 0.018 0.020 0.018 N4 0.331 0.414 0.372 C11 -0.028 -0.030 -0.028 N5 0.182 0.191 0.170 C12 0.018 0.020 0.018 C6 -0.040 -0.012 -0.043 C13 -0.027 -0.032 -0.029 C7 -0.041 -0.012 -0.045 a: DFT calculation at the B3LYP/6-31G(d) level of theory.

109

Chapter 5

Table 5.9 Calculated isotropic hyperfine coupling constant (hfcc) in Ga for F, G and H F G H N1 2.79 3.98 3.78 N2 4.59 5.82 5.09 N4 4.59 5.82 5.09 N5 2.79 3.98 3.79 a: DFT calculation at the B3LYP/EPR-III level of theory.

5.5 Reduced products: Phenylene-Linked Diformazanate Ligands

5.5.1 Synthesis and Crystal Structures

The cyclic voltammetry data of 6k and 6l show that the first redox couples are two one- electron reductions and located at around -0.95 V (vs. Fc/Fc+). Unfortunately, the attempted chemical synthesis of the 1-electron reduction products was not successful. Even though only one equivalent of reducing agent (Cp2Co or Cp*2Co) was used, the isolated product is the mixture of 2-electron reduction product and unreacted starting material, which presumably precipitates from the reaction mixture and drives the equilibrium to the observed products.

Single crystals suitable for X-ray diffraction analysis of [6k][Cp*2Co]2 (Figure 5.11, metrical parameters in Table 5.10) were obtained by using decamethylcobaltocene (Cp*2Co) as reducing agent in the solvent mixture of DMSO/THF/Toluene. The centroid of the phenylene spacer between two formazanate fragments is sitting at an inversion center; therefore, only one formazanate unit will be discussed. Unlike the slightly twist structure of the [6k], which has the dihedral angles between the phenylene spacer and the formazanate backbone of 10o and 17o, the crystal structure of the [6k]-2 is almost flat. The two formazanate units of [6k]-2 are almost coplanar with the phenylene spacer with the dihedral angle of 1.79o, and the boron centers are sitting in the plane of the formazanate backbones with the displacement of 0.05 Å. In addition, the four phenyl substituents are nearly coplanar with the formazanate backbones with dihedral angles of 4o and 17o, which are similar to the case of [6b]-. The metric parameters within the formazanate units of [6k][Cp*2Co]2 are very similar with the parameters of [6a][Cp2Co]: the N-N bond lengths of [6k][Cp*2Co]2 elongate to 1.360(3) and 1.362(3) Å and the N-B bond lengths shorten to 1.539(4) and 1.536(4) Å. The N-N and N-B bond lengths of [6k][Cp*2Co]2 again suggest the presence of a reduced formazanate ligand.

110

(Formazanate)Boron Complexes

Figure 5.11 Solid-state structure of [6k][Cp*2Co]2 (right; up: top view; bottom: side view) and calculated structure of [6k]-BS (left; up: top view; bottom: side view). For the crystal structure: Thermal ellipsoids are shown at 50% probability, and hydrogen atoms and decamethylcobaltocenium cations are removed; Ph groups (bottom) are shown as wireframe for clarity. o Table 5.10 Selected bond length (Å) and bond angles ( ) of [6k][Cp*2Co]2 [6k][Cp*2Co]2 N1-N2, N3-N4 1.360(3), 1.362(3) C7-N2, C7-N3 1.335(3), 1.339 (4) N1-B1, N4-B1 1.539(4), 1.536(3) C7-C8 1.487(3) C8-C9, 1.403(3), C8-C10, 1.404(4), C9-C10’ 1.388(3) N2-C7-N3 129.63 N1-B1-N4 108.45 Boron displacementa 0.048 Dihedral anglesb 4.04, 13.27, 1.79 aDistance between B1/B2 and N1-N2-N3-N4/ N5-N6-N7-N8 plane. bAngles between the plane defined by N1, N4, and C7 aryl substituents and the N1- N2-N3-N4 plane.

In addition to the expected changes of the N-N and N-B bond lengths after the reduction, the average bond length of the phenylene spacer of [6k][Cp*2Co]2 is slightly longer than 6k (1.398 Å vs. 1.386 Å). The elongation of the C-C bond lengths after the reduction suggests some perturbation of the central phenylene spacer. We also notice that the bond lengths of C8-C9 (1.403(3) Å) and C8-C10 (1.404(4) Å) are slightly longer that the bond length of C9- C10’ (1.388(3) Å). But the bond length difference is very small (~ 0.01 Å) in comparison to

111

Chapter 5 the famous Thiele’s hydrocarbon (Chart 5.3), which shows significant C-C bond length alternation of the phenylene spacer (1.449 (3) vs. 1.346 (3) Å) and a quinoidal character.21

The C-C bond length alternation similar to that bserved in [6k][Cp*2Co]2 has been reported for phenylene-bridged diverdazyl radical system20a, which were reported to not have quinoidal character. In addition, the C7-C8 bond length of [6k][Cp*2Co]2 (1.487(3) Å) is statistically indistinguishable with the case of 6k (1.481(5) and 1.486(5)Å). If [6k][Cp*2Co]2 would have quinoidal character, the C7-C8 bond length of [6k][Cp*2Co]2 should be shorter than the corresponding bonds in 6k. All the information mentioned above suggests that [6k][Cp*2Co]2 has no (or very small) quinoidal character and [6k][Cp*2Co]2 is best described as two reduced

(formazanate)BF2 units bridging by a phenylene spacer.

Chart 5.3 Resonance structure of Thiele’s hydrocarbon

5.5.2 DFT calculations and VT-EPR Studies In order to understand the ground state electronic configuration of diradical species of [6k]-2 and [6l]-2, the open-shell triplet (T) and broken-symmetry (BS) state of [6k]-2 and [6l]-2 were subjected to DFT calculations at the B3LYP/6-31G(d) level of theory. In the case of [6k]-2, two local minima can be located. The local minimum with lower energy shows a slightly twisted structure (Figure 5.11). The other local minimum, which has higher calculated energy, shows similar flat geometry with the crystal structure of the [6k][Cp*2Co]2. The flat geometry of the [6k][Cp*2Co]2 observed in the crystal structure might be due to packing effects.

The calculated ground state of [6k]-2 is the broken-symmetry state [6k]-2-BS) with a nearly degenerate triplet state (G = 0.21 kcal/mol), and the ground state of [6l]-2 is a triplet state ([6l]-2-T) with a nearly degenerate broken-symmetry state (G = 0.13 kcal/mol). These calculated ground states of [6k]-2 and [6l]-2 are the same to those reported for phenylene- linked verdazyl (and other) radicals.22 The nearly degenerate singlet and triplet state of [6k]-2 and [6l]-2 suggest that exchange coupling, which is quantified by the coupling constant J,

112

(Formazanate)Boron Complexes between two radical centers is very weak. This weak interaction was also shown in the reported verdazyl radical systems.23 The SOMO of the reduced formazanate ligand shows a nodal plane passing through the meso position of formazanate ligand (Chart 5.2). As a result, there is small spin density at the meso position of the formazanate ligand (C3 in Chart 5.2) leading to the weak interaction between two radical centers.

The ground state electronic configurations of [6k][Cp2Co]2 and [6l][Cp2Co]2 were experimentally studied by VT-EPR experiments (Figure 5.12) in the solvent mixture of

MeTHF/DMSO due to the low solubility of the [6k][Cp2Co]2 and [6l][Cp2Co]2. The EPR spectra of [6k][Cp2Co]2 (at 40K) and [6l][Cp2Co]2 (at 19K) show similar feature as

[6a][Cp2Co], which has a doublet ground state with g-value of ~2. The doublet EPR spectra of [6k][Cp2Co]2 and [6l][Cp2Co]2 suggests that there is no (or very weak) interaction between the two radical centers of both [6k][Cp2Co]2 and [6l][Cp2Co]2.

Lowering the temperature, the major change of EPR spectrum was found in the compound

[6k][Cp2Co]2. The intensity of the EPR resonance of [6k][Cp2Co]2 is decreasing when lowering the temperature from 40K to 5K. This temperature dependent EPR feature suggests antiferromagnetic coupling (J < 0) between radical centers at low temperature. The origin of the antiferromagnetic interaction is very likely an intramolecular interaction instead of an intermolecular interaction due to the long intermolecular distance (> 11 Å) between spin centers in the crystal structure of [6k][Cp*2Co]2. The intermolecular antiferromagnetic interaction between two reduced formazanate units provides experimental support for the (broken-symmetry) singlet ground state of [6k]-2 that was found by DFT calculations.

Figure 5.12 VT-EPR spectra of [6k][Cp2Co]2 and [6l][Cp2Co]2

113

Chapter 5

The EPR feature of [6l][Cp2Co]2 only has a small influence from the temperature. Lowering the temperature from 19K to 5K results in slightly broadening of the signal, and the intensity of the EPR signal does not obviously decrease. Even though the change of the VT-EPR spectra of [6l][Cp2Co]2 is small, it is very likely that the two radical centers of [6l][Cp2Co]2 have the ferromagnetic coupling (J > 0), which is normal for the 1,3-benzene bridged diradical system, resulting in a triplet ground state. The less obvious change of the VT-EPR of

[6l][Cp2Co]2 than [6k][Cp2Co]2 suggest that the radical-radical interaction of [6l][Cp2Co]2 is weaker than [6k][Cp2Co]2. The weaker interaction (smaller J in magnitude) of [6l][Cp2Co]2 47 than [6k][Cp2Co]2 is similar to the reported verdazyl-based diradical systems and in a good agreement with the results of DFT calculations.

5.6 Conclusion

In conclusion, a series of (formazanate)boron complexes (LBX2, 6: X=F, 9: X=Ph, 10: X=H) were synthesized and fully characterized. The optical properties of the LBX2 complexes include strong absorption, long wavelength emission, and large Stokes shift. The large Stokes shift and long wavelength emission of the (formazanate)boron complexes suggests that they have great potential application as the fluorescent dyes. The electrochemical studies of the

LBX2 reveal two one-electron reductions of the formazanate boron complexes. The crystal structure and the EPR spectrum of the reduced product ([6]-) show the organic radical property, which suggest the redox-active feature of the formazanate ligand. In addition, the optical and redox-active properties of the LBF2 complexes (6) are tunable by the substituents of the formazanate ligand. In the case of the di-formazan system (6k and 6l), the optical and redox-active properties are similar to 6a but with stronger emission due to the two (formazanate)boron difluoride units in the structure. The reduced products of 6k and 6l show biradical charter. At the low-temperature region, [6k]-2 and [6l]-2 have weak antiferromagnetic and ferromagnetic coupling, respectively.

5.7 Experimental section General Considerations. All manipulations were carried out under nitrogen atmosphere using standard glovebox, Schlenk, and vacuum-line techniques. Toluene, hexane, and pentane

(Aldrich, anhydrous, 99.8%) were passed over columns of Al2O3 (Fluka), BASF R3-11- supported Cu oxygen scavenger, and molecular sieves (Aldrich, 4 Å). Diethyl ether and THF

(Aldrich, anhydrous, 99.8%) were dried by percolation over columns of Al2O3 (Fluka).

Deuterated solvents were vacuum transferred from Na/K alloy (C6D6, THF-d8, Aldrich) and stored under nitrogen. The compound [PhNNC(CN)NNPh]BF2 (6i) was synthesized

114

(Formazanate)Boron Complexes

16c according to a published procedure. BF3(ether), BH3(SMe2) and BPh3 were purchased from Aldrich and used as received. NMR spectra were recorded on Mercury 400, Inova 500 or Agilent 400 MR spectrometers. The 1H and 13C NMR spectra were referenced internally using the residual solvent resonances and reported in ppm relative to TMS (0 ppm); J is reported in Hz. Assignment of NMR resonances was aided by gradient-selected gCOSY, NOESY, gHSQCAD and/or gHMBCAD experiments using standard pulse sequences. 11B NMR spectra were recorded in quartz (or normal glass) NMR tubes using a OneNMR probe on an Agilent 400 MR system. UV-Vis spectra were recorded in THF solution (~ 10-5 M) using a

Avantes AvaSpec 3648 spectrometer and AvaLight-DHS lightsource inside a N2 atmosphere glovebox. The photoluminescence quantum yield of 6, 9a and 10a was determined using an optically dilute solution in THF (λex = 473 nm) with optically dilute Rubipy aqueous solution as reference. Spectra were recorded using a 75W Xenon lamp coupled to a Zolix 150 monochromator coupled directly to a Qpod cuvette holder (Quantum Northwest) and emission was collected through a fibre optic connected Shamrock 163 spectrograph and iDUS-420A-OE CCD detector. Spectra are uncorrected for instrument response. Elemental analyses were performed at the Microanalytical Departement of the University of Groningen or Kolbe Microanalytical Laboratory (Mülheim an der Ruhr, Germany).

General Procedure for (formazanate)BF2 complexes

Indirect Method. A schlenk flask was charged with formazan, 0.5 equivalents of dimethylzinc (1.2M in toluene) and dry toluene. The reaction mixture was stirred overnight at RT, and then all volatile were removed under vacuum. At this step, bisformazanate zinc complex could be formed. The same schlenk flask was charged with boron trifluoride diethyl etherate (1.5-2.0 eq vs. formazan) and dry toluene. After the reaction mixture had been stirred overnight at 70 °C, all volatiles was removed under vacuum. The product was purified by chromatography (DCM/hexane, silica gel) or recrystallization from alkane (hexane or heptane). This procedure also can start from isolated bisformazanate zinc complexes and boron trifluoride diethyl etherate directly.

Direct Method.16cd The schlenk flask was charged with formazan and dry toluene. Triethylamine (3eq vs. formazan) was then added slowly at RT, and the reaction mixture was stirred for 10-20 min. Then, boron trifluoride diethyl etherate (5 eq vs. formazan) was added, and the solution was stirred at 80 °C for 24 hours. After the solution had been cooled to room temperature, deionized water was added to quench the reaction. The toluene solution was washed with deionized water three times, dried over MgSO4, gravity filtered and concentrated

115

Chapter 5 by rotavapor. The crude product was purified by flash chromatography (DCM, silica gel). The product can be further purified by recrystallization from hexane or heptane.

(PhNNC(p-tol)NNPh)BF2 6a. A schlenk flask was charged with [PhNNC(p-tolyl)NNHPh]

(1a) (610 mg, 0.83 mmol), BF3·Et2O (0.4 mL, 3.24 mmol) and 10 mL of toluene. The reaction mixture was stirred at 70°C overnight after which the color had changed from blue to purple and all volatiles were removed under reduced pressure. To the residue was added 10 mL of toluene and BF3·Et2O (0.4 mL, 3.24 mmol), and the mixture was stirred at 70 °C overnight, upon which the color had changed to red. The volatiles were removed in vacuo and the residue was taken up into hexane and recrystallized by slowly cooling the clear red solution to

-30 °C to afford 517 mg of (PhNNC(p-tol)NNPh)BF2 (1.43 mmol, 86%) as red crystalline 1 material. H NMR (400 MHz, C6D6, 25 °C) δ 8.06 (d, 2H, J = 8.2, p-tolyl CH), 7.91 (d, 4H, J = 8.4, Ph o-H), 7.09 (d, 2H, J = 8.2, p-tolyl CH), 7.01 (t, 4H, J = 7.1, Ph m-H), 6.95 (t, 2H, J 19 = 7.2, Ph p-H), 2.13 (s, 3H, p-tolyl CH3). F NMR (376.4 MHz, C6D6, 25 °C) δ -144.4 (q, 2F, 11 13 J = 28.8, BF2). B NMR (128.3 MHz, C6D6, 25 °C) δ -0.06 (t, 1B, J = 28.7, BF2). C NMR

(100.6 MHz, C6D6, 25 °C) δ 150.3 (Ph i-C), 144.8 (NCN), 139.8 (p-tolyl i-C), 131.8 (p-tolyl CMe), 130.1 (Ph m-CH), 130.0 (p-tolyl CH), 129.6 (Ph p-CH), 126.4 (p-tolyl CH), 124.2 (Ph o-CH), 21.6 (p-tolyl CH3). Anal. Calcd for C20H17BF2N4: C, 66.32; H, 4.73; N, 15.47. Found: C, 66.39; H, 4.75; N, 15.30.

t t (PhNNC( Bu)NNPh)BF2 6b. [PhNNC( Bu)NNHPh] (1b) (336.5 mg, 1.2 mmol), dimethylzinc 1.2M solution in toluene (0.5 mL, 0.6 mmol) and boron trifluoride diethyl etherate (0.25 mL, 2.0 mmol) were used (indirect method). After recrystallization from hexane, 260.9 mg of orange crystal of [PhNNC(tBu)NNPh]BF2 was isolated (66 %). 1H

NMR (400 MHz, CDCl3, 25 °C): 7.79 (d, 4H, J = 6.0 Hz, Ph o-H), 7.42 (t, 4H, J = 6.4 Hz, Ph 13 m-H), 7.36 (t, 2H, J = 6.0 Hz, Ph p-H), 1.45 (s, 9H, tBu). C NMR (100.6 MHz, CDCl3, 25 °C): 158.9 (NCN), 144.1 (Ph i-C), 129.4 (Ph p-CH), 129.2 (Ph m-CH), 123.4 (Ph o-CH), 37.0 19 (tBu C(CH3)3), 29.8 (tBu CH3). F NMR (376.4 MHz, CDCl3, 25 °C): -145.7 (q, 2F, J = 28.9 11 Hz, BF2). B NMR (128.3 MHz, CDCl3, 25 °C): -0.70 (t, 1B, J = 28.7 Hz, BF2). Anal. calcd for C17H19BF2N4: C, 62.22; H, 5.84; N, 17.07. Found: C, 62.37; H, 5.82; N, 16.90.

[PhNNC(p-tolyl)NNMes]BF2 6c. [PhNNC(p-tolyl)NNHMes] (1c) (255.2 mg, 0.72 mmol), dimethylzinc 1.2M solution in toluene (0.3 mL, 0.36 mmol) and boron trifluoride diethyl etherate (0.15 mL, 1.22 mmol) were used (indirect method). After chromatography

(DCM/hexane: 2/5, silicagel, Rf: 0.4), 182.0 mg of deep red crystal of [PhNNC(p- 1 tolyl)NNMes]BF2 was isolated (63 %). H NMR (400 MHz, CDCl3, 25 °C): 7.94 (d, 2H, J =

116

(Formazanate)Boron Complexes

8.4 Hz, p-tolyl CH), 7.83 (d, 2H, J = 7.2 Hz, Ph o-CH), 7.49-7.38 (m, 3H, Ph m- and p-CH),

7.24 (d, 2H, J = 8.0 Hz, p-tolyl CH), 6.92 (s, 2H, Mes m-CH), 2.41 (s, 3H, p-tolyl CH3), 2.30 13 (s, 3H, Mes p-CH3), 2.05 (s, 6H, Mes o-CH3). C NMR (100.6 MHz, CDCl3, 25 °C): 151.4 (NCN), 143.3 (Ph i-C), 139.8 (p-tolyl p-C), 139.8 (Mes o-C), 139.3 (Mes p-C), 134.7 (Mes i- C), 130.4 (p-tolyl i-C), 129.7 (Ph p-CH), 129.6 (p-tolyl CH), 129.5 (Mes m-CH), 129.3 (Ph m-CH), 125.8 (p-tolyl CH), 123.9 (Ph o-CH), 21.6 (p-tolyl p-CH3), 21.3 (Mes p-CH3), 18.0 19 11 (Mes o-CH3). F NMR (376.4 MHz, CDCl3, 25 °C): -156.9 (q, 2F, J = 23.5 Hz, BF2). B

NMR (128.3 MHz, CDCl3, 25 °C): -1.07 (t, 1B, J = 23.6 Hz, BF2). Anal. calcd for

C23H23BF2N4: C, 68.33; H, 5.73; N, 13.86. Found: C, 68.65; H, 5.83; N, 13.57.

[PhNNC(p-tolyl)NNC6F5]BF2 6e. [PhNNC(p-tolyl)NNHC6F5] (1e) (536 mg, 1.33 mmol), dimethylzinc 1.2M solution in toluene (0.5 mL, 0.6 mmol) and boron trifluoride diethyl etherate (0.3 mL, 2.45 mmol) were used (indirect method). After recrystallization from 1 hexane, 525.7 mg of orange crystal of [PhNNC(p-tolyl)NNC6F5]BF2 was isolated (88 %). H

NMR (400 MHz, CDCl3, 25 °C): 8.01 (d, 2H, J = 8.4 Hz, p-tolyl CH), 7.76-7.74 (m, 2H, Ph o-H), 7.04 8.01 (d, 2H, J = 8.0 Hz, p-tolyl CH), 6.94-6.90 (m, 3H, Ph m-H and p-H), 2.09 (s, 13 3H, p-tolyl CH3). C NMR (100.6 MHz, CDCl3, 25 °C): 153.1 (NCN), 144.2 (dm, J = 255.5

Hz, C6F5), 143.9 (m, Ph i-C), 141.9 (dm, J = 257.5 Hz, C6F5), 140.9 (p-tolyl i-C), 138.2 (dm,

J = 250.9 Hz, C6F5), 131.7 (Ph m-CH), 130.4 (p-tolyl CMe), 130.3 (p-tolyl CH), 129.8 (Ph p-

CH), 126.6 (p-tolyl CH), 124.9 (t, J = 2.0 Hz, Ph o-CH), 119.4 (m, C6F5 i-C), 21.6 (p-tolyl 19 CH3). F NMR (376.4 MHz, CDCl3, 25 °C): -147.8 – 147.9 (m, 2F, C6F5 m-CF), -153.0 (t, 1F,

J = 22.5 Hz, C6F5 p-CF), -154.4 (q, 2F, J = 24.0 Hz, BF2), -161.5 - -161.6 (m, 2F, C6F5 o-CF). 11 B NMR (128.3 MHz, CDCl3, 25 °C): -1.06 (t, 1B, J = 23.7 Hz, BF2). Anal. calcd for

C20H12BF7N4: C, 53.13; H, 2.68; N, 12.39. Found: C, 53.50; H, 2.72; N, 12.14.

[C6F5NNC(p-tolyl)NNMes]BF2 6f. [C6F5NNC(p-tolyl)NNMes]2Zn (5f) (535 mg, 0.56 mmol) and boron trifluoride diethyl etherate (0.24 mL, 1.9 mmol) were used (indirect method). The product was purified by flash chromatography (DCM, silicagel). After which 439.6 mg of 1 orange solid of [C6F5NNC(p-tolyl)NNMes]BF2 was isolated (80 %). H NMR (400 MHz,

CDCl3, 25 °C): 7.83 (d, 2H, J = 8.0 Hz, p-tolyl CH), 7.25 (d, 2H, J = 7.7 Hz, p-tolyl CH, overlap with CDCl3), 6.96 (s, 2H, Mes m-CH), 2.40 (s, 3H, p-tolyl CH3), 2.32 (s, 3H, Mes p- 13 CH3), 2.13 (s, 6H, Mes o-CH3). C NMR (100.6 MHz, CDCl3, 25 °C): 152.6-152.2 (m,

NCN), 143.8 (dm, J = 255.8 Hz, C6F5), 141.9 (dm, J = 246.2 Hz, C6F5), 140.7 (Mes p-C),

140.4 (p-tolyl i-C), 139.9 (Mes i-C), 138.2 (dm, J = 251.2 Hz, C6F5), 134.4 (Mes o-C), 129.9 (Mes m-CH), 129.8 (p-tolyl CH), 129.4 (p-tolyl i-C), 126.0 (p-tolyl CH), 119.0-118.5 (m,

117

Chapter 5

19 C6F5 i-C), 21.6 (p-tolyl CH3), 21.3 (Mes p-CH3), 18.0 (Mes o-CH3). F NMR (376.4 MHz,

CDCl3, 25 °C): -146.4 – 146.6 (m, 2F, C6F5 m-CF), -152.2 (t, 1F, J = 21.3 Hz, C6F5 p-CF), - 11 159.4 (qt, 2F, J = 21.3, 6.1 Hz, BF2), -160.9 - -161.1 (m, 2F, C6F5 o-CF). B NMR (128.3

MHz, CDCl3, 25 °C): -2.06 (t, 1B, J = 20.9 Hz, BF2). Anal. calcd for C23H18BF7N4: C, 55.90; H, 3.67; N, 11.34. Found: C, 56.01; H, 3.64; N, 11.03.

[PhNNC(C6F5)NNMes]BF2 6g. A solution of [PhNNC(C6F5)NN(BF3)Mes]2Zn (7g) (103.1 mg, 0.097 mmol) in toluene (10 mL) was stirred at 130°C for 12 hours after which all volatiles were removed under vacuo. The crude product was dissolved in hexane and separated from solid by filtration. The product was further purified by silica column chromatography with CH2Cl2/hexane (1:10)(r = 0.2). The fractions were collected and removing solvent in vacuo afforded the product as red solid (51 mg, 0.104 mmol, 54%). 1H

NMR (400 MHz, CDCl3, 25 °C) δ 7.85-7.78 (m, 2H, Ph o-H), 7.52-7.43 (m, 3H, Ph m-H and 11 p-H), 6.90 (s, 2H, Mes m-H), 2.27 (s, 3H, Mes p-CH3), 2.05 (s, 6H, Mes o-CH3). B NMR 19 (376.4 MHz, CDCl3, 25 °C) δ -1.34 (t, 1B, J = 23.8, BF2). F NMR (128.3 MHz, CDCl3, 25

°C) δ -140.9 - -141.1 (m, 2F, C6F5 m-F), -151.8 (tt, 1F, J = 21.0, 2.2, C6F5 p-F), -153.2 (q, 2F, 13 J = 23.7, BF2), -161.2 - -161.4 (m, 2F, C6F5 o-F). C NMR (100.6 MHz, CDCl3, 25 °C) δ

147.2-146.7 (m, C6F5), 144.7-144.2 (m, C6F5), 143.8-143.2 (m, C6F5), 142.8 (Ph, i-C), 141.1-

140.7 (m, C6F5), 140.7-140.4 (m, C6F5), 139.7 (NNCNN), 139.3 (Mes i-C). 139.5-139.1 (m,

C6F5), 137.1-136.6 (m, C6F5), 134.4 (Mes o-C), 130.7 (Ph p-C), 129.6 (Mes m-C and Ph m-C),

123.9 (Ph o-C), 109.8 (td, J = 15.6, 4.1, C6F5 i-C), 21.3 (Mes p-CH3), 17.8 (Mes o-CH3). Anal.

Calcd for C19H10BF7N4: C, 52.09; H, 2.30; N, 12.79.

[C6F5NNC(C6F5)NNMes]BF2 6h. [C6F5NNC(C6F5)NNHMes] (1h) (301.3 mg, 0.58 mmol), triethylamine (0.24 mL, 1.72 mmol) and boron trifluoride diethyl etherate (0.35 mL, 2.84 mmol) were used (direct method). After chromatography (DCM/hexane: 1/5, silicagel, Rf: 1 0.5), 290.5 mg of orange crystal of [C6F5NNC(C6F5)NNMes]BF2was isolated (88 %). H

NMR (400 MHz, CDCl3, 25 °C): 6.95 (s, 4H, Mes m-CH), 2.30 (s, 6H, Mes p-CH3), 2.13 (s, 13 12H, Mes o-CH3). C NMR (100.6 MHz, CDCl3, 25 °C): 146.9 (m, NCN), 143.7 (dm, J =

256.5 Hz, C6F5), 143.2 (dm, J = 250.1 Hz, C6F5), 143.6 (dm, J = 255.4 Hz, C6F5), 142.4 (dm,

J = 258.8 Hz, C6F5), 141.0(Mes p-C), 139.5 (Mes i-C), 138.1 (dm, J = 251.5 H, C6F5), 134.3

(Mes o-C), 130.0 (Mes m-CH), 118.1 (m, C6F5 i-C), 108.7 (m, C6F5 i-C), 21.3 (Mes p-CH3), 19 17.8 (Mes o-CH3). F NMR (376.4 MHz, CDCl3, 25 °C): -140.6 (dd, 2F, J = 22.8, 7.4 Hz,

C6F5 o-CF), -146.4 (m, 2F, C6F5 o-CF), -150.3 - -150.5 (m, 2F, C6F5 p-CF and C6F5 p-CF), -

158.1 (qt, 2F, J = 20.9, 6.1 Hz, BF2), -160.2 - -160.4 (m, 2F, C6F5 m-CF), -160.7 - -160.9 (m,

118

(Formazanate)Boron Complexes

11 2F, C6F5 m-CF). B NMR (128.3 MHz, CDCl3, 25 °C): -2.28 (t, 1B, J = 20.3 Hz, BF2). Anal. calcd for C22H11BF12N4: C, 46.35; H, 1.94; N, 9.83. Found: C, 46.77; H, 2.00; N, 9.55.

[PhNNC(CN(B(C6F5)3))NNPh]BF2 6i(B(C6F5)3). A vial was charged with

[PhNNC(CN)NNPh]BF2 (6i) (99.5 mg, 0.33 mmol), B(C6F5)3 (178.2 mg, 0.35 mmol) and toluene and then the reaction mixture was stirred at room temperature for 1 hour. After which the crude product was recrystallized from toluene/hexane mixture at -30 °C and then the 1 orange crystalline material of 6i(B(C6F5)3) (237.1 mg, 89 %) was obtained. H NMR (400

MHz, CDCl3, 25 °C): 7.70-7.64 (m, 4H, Ph o-CH), 6.85-6.79 (m, 6H, Ph m-CH and p-CH). 13 C NMR (100.6 MHz, CDCl3, 25 °C): 148.2 (dm, J = 238.4 Hz, C6F5), 142.6 (Ph i-C), 140.7

(dm, J = 253.6 Hz, C6F5), 137.5 (dm, J = 249.6 Hz, C6F5), 132.2 (Ph p-CH)., 129.4 (Ph m- 19 CH)., 122.6 (Ph o-CH). F NMR (376.4 MHz, CDCl3, 25 °C): -125.8 (q, 2F, J = 29.4 Hz,

BF2), -134.1 (dd, 6F, J = 23.5, 9.9 Hz, C6F5 o-CF), -154.8 (t, 3F, J = 21.0 Hz, C6F5 p-CF), - 11 162.8 (td, 6F, J = 22.2, 8.6 Hz, C6F5 m-CF). B NMR (128.3 MHz, CDCl3, 25 °C): -0.76 (t,

1B, J = 31.6 Hz, BF2).

[MesNNC(CN)NNMes]BF2 6j. [MesNNC(CN)NNHMes] (1j) (200.6 mg, 0.60 mmol), triethylamine (0.25 mL, 1.80 mmol) and boron trifluoride diethyl etherate (0.39 mL, 3.16 mmol) were used (direct method). After flash chromatography (DCM, silicagel) and recrystallization from heptane, 188.8 mg of yellow crystal of [MesNNC(CN)NNMes]BF2 was 1 isolated (82 %). H NMR (400 MHz, CDCl3, 25 °C): 6.94 (s, 4H, Mes m-CH), 2.30 (s, 6H, 13 Mes p-CH3), 2.17 (s, 12H, Mes o-CH3). C NMR (100.6 MHz, CDCl3, 25 °C): 140.4 (Mes p- C), 139.0 (Mes i-C), 134.1 (Mes o-C), 130.2 (Mes m-CH), 129.1 (NCN), 113.8 (CN), 21.2 19 (Mes p-CH3), 18.5 (Mes o-CH3). F NMR (376.4 MHz, CDCl3, 25 °C): -145.6 (q, 2F, J = 11 23.8 Hz, BF2). B NMR (128.3 MHz, CDCl3, 25 °C): -2.28 (t, 1B, J = 23.7 Hz, BF2). Anal. calcd for C20H22BF2N5: C, 63.01; H, 5.82; N, 18.37. Found: C, 63.02; H, 5.79; N, 18.17.

[Ph2-Ph-Ph2](BF2)2 6k. 1k (40.9 mg, 0.08 mmol), triethylamine (0.65 mL, 4.68 mmol) and boron trifluoride diethyl etherate (1.00 mL, 8.10 mmol) were used (direct method). After flash chromatography (DCM, silicagel) and recrystallization from DCM/hexane, 37.5 mg of deep 1 red crystal of 6k was isolated (78 %). H NMR (400 MHz, CDCl3, 25 °C): 8.22 (s, 4H, C6H4), 7.94 (d, 8H, J = 7.7, Ph o-CH), 7.54-7.42 (m, 12H, Ph m-CH and p-CH). 19F NMR (376.4 11 MHz, CDCl3, 25 °C): -144.1 (q, 2F, J = 28.5 Hz, BF2). B NMR (128.3 MHz, CDCl3, 25 °C): 13 -0.53 (t, 1B, J = 28.9 Hz, BF2). C NMR (100.6 MHz, CDCl3, 25 °C): 148.7 (C6H4 i-C),

143.8 (Ph i-C), 134.3 (NCN), 129.8 (Ph p-CH), 129.1 (Ph m-CH), 125.7 (C6H4 CH), 123.5

119

Chapter 5

(Ph, o-CH). Anal. calcd for C32H24B2F4N8: C, 62.17; H, 3.91; N, 18.13. Found: C, 61.93; H, 3.83; N, 17.94.

[Ph2-mPh-Ph2](BF2)2 6l.

1l (39.8 mg, 0.08 mmol), triethylamine (0.65 mL, 4.68 mmol) and boron trifluoride diethyl etherate (1.00 mL, 8.10 mmol) were used (direct method). After flash chromatography (DCM, silicagel) and recrystallization from DCM/hexane, 25.2 mg of deep red crystal of 6l was 1 isolated (54 %). H NMR (400 MHz, CDCl3, 25 °C): 8.81 (s, 1H, H2), 8.19 (dd, 2H, J = 7.8, 1.6 Hz, H4 and H6), 7.94 (d, 8H, J = 7.7 Hz, Ph o-CH), 7.60 (t, 1H, J = 7.9 Hz, H5), 7.50- 19 7.41 (m, 12H, Ph m-CH and p-CH). F NMR (376.4 MHz, CDCl3, 25 °C): -143.7 (q, 2F, J = 11 13 28.3 Hz, BF2). B NMR (128.3 MHz, CDCl3, 25 °C): -0.50 (t, 1B, J = 29.0 Hz, BF2). C

NMR (100.6 MHz, CDCl3, 25 °C): 148.7 (NCN), 143.8 (Ph i-C), 134.3 (C1 and C3), 129.8 (Ph p-CH), 129.2 (C5), 129.1 (Ph m-CH), 126.2 (C4 and C6), 123.4 (Ph o-CH), 122.6 (C2).

Anal. calcd for C32H24B2F4N8: C, 62.17; H, 3.91; N, 18.13. Found: C, 61.59; H, 3.94; N, 17.70.

[PhNNC(p-tol)NNPh]BPh2 9a. A schlenk flask was charged with [PhNNC(p-tolyl)NNHPh]

(1a) (250.8 mg, 0.8 mmol), BPh3 (230.4 mg, 0.95 mmol) and dry toluene (35 mL). The reaction mixture was heated to 100 °C under a nitrogen atmosphere for 3 days. After which the solvents were evaporated and the yield of the product 9a was 94 %. 1H NMR (400 MHz,

CDCl3, 25 °C): 7.79 (d, 2H, J = 8.2 Hz, p-tolyl CH), 7.30 (d, 4H, J = 7.5 Hz, Ph), 7.25 (tt, 2H, J = 7.3, 1.0 Hz, Ph p-CH), 7.19 (d, 2H, J = 8.0 Hz, p-tolyl CH), 7.20-7.16 (m, 4H, Ph o-CH), 11 7.13-7.06 (m, 10H, Ph), 2.37 (s, 3H, p-tolyl CH3). B NMR (128.3 MHz, C6D6, 25 °C): 13 2.04. C NMR (125 MHz, CDCl3, 25 °C): 153.4, 146.3, 138.8 (p-tolyl p-C), 134.7, 129.1 (p- tolyl CH), 128.3 (Ph p-CH), 127.7 (Ph), 127.0, 126.7, 126.3 (Ph o-CH), 125.2 (p-tolyl CH),

21.3 (p-tolyl CH3).

[PhNNC(p-tol)NNPh]BH2 10a. A schlenk flask was charged with [PhNNC(p-tolyl)NNHPh]

(1a) (601.2 mg, 1.91 mmol), BH3(SMe2) (0.18 mL, 1.90 mmol) and dry toluene. The reaction mixture was stirred overnight at RT, and then all volatile were removed under vacuum. The product was purified by chromatography (DCM/hexane = 1/2, silica gel, Rf = 0.71). After

120

(Formazanate)Boron Complexes

1 which 178.2 mg (0.46 mmol, 29 %) of 10a can be obtained. H NMR (400 MHz, C6D6, 25 °C): 8.16 (d, 2H, J = 8.2, p-tolyl CH), 7.88 (d, 4H, J = 8.1, Ph o-CH), 7.10 (d, 2H, p-tolyl CH, overlap with C6D6), 6.97 (t, 4H, J = 8.3, Ph m-CH), 6.88 (tt, 4H, J = 7.3, 1.8, Ph p-CH), 3.67 11 13 (bs, 2H, BH2), 2.11 (p-tolyl CH3). B NMR (128.3 MHz, C6D6, 25 °C): -10.8. C NMR

(100.6 MHz, C6D6, 25 °C): 153.3 (NCN), 145.9 (Ph i-C), 138.7 (p-tolyl p-C), 131.7 (p-tolyl i- C), 129.3 (p-tolyl CH), 128.9 (Ph m-CH), 128.2 (Ph p-CH), 125.4 (p-tolyl CH), 122.4 (Ph o-

CH), 20.9 (p-tolyl CH3).

[PhNNC(p-tolyl)NNPh]BF2[Cp2Co] [6a][Cp2Co]. A mixture of solid (PhNNC(p- tol)NNPh)BF2 (24.9 mg, 0.069 mmol) and Cp2Co (14.9 mg, 0.079 mmol) was prepared. After addition of 2 mL THF, the color of the reaction mixture changed to green. Slow diffusion of hexane (4 mL) into the THF solution precipitated 42 mg of [(PhNNC(p- tol)NNPh)BF2][Cp2Co](THF) as green crystalline material (0.067 mmol, 98%). Anal. Calcd for C34H35BCoF2N4O: C, 65.50; H, 5.66; N, 8.99. Found: C, 65.47; H, 5.71; N, 8.99.

[PhNNC(t-Bu)NNPh]BF2[Cp2Co] [6b][Cp2Co]. A vial was charged with 30.0 mg of [PhNNC(t-Bu)NNPh]BF2 (0.09 mmol) and THF 2 mL. After which, 2 mL THF solution of

17.5 mg of Cp2Co (0.09 mmol) was added into flask slowly. After the reaction mixture stands at RT for 6 hours, 10 mL hexane was slowly added into the vial. After overnight, 47.0 mg of deep green crystal/powder of [PhNNC(t-Bu)NNPh]BF2[Cp2Co] can be isolated (0.09 mmol, 99 %).

[PhNNC(p-tolyl)NNMes]BF2[Na(15-crown-5)] [6c][Na(15-crown-5)]. A flask was charged with 12.1 mg of [PhNNC(p-tolyl)NNMes]BF2 (0.03 mmol), 28.4 mg of Na/Hg (2.447 % of Na, 0.03 mmol), 6.1 L of 15-crown-5 (0.03 mmol) and 1.5 mL of THF. After the reaction mixture had been stirred at RT for 6 hours, Hg(l) was removed by filtration. After which hexane was added slowly to form a second layer on the top of THF solution. Green crystal of

[PhNNC(p-tolyl)NNMes]BF2[Na(15-crown-5)] can be obtained overnight (17 mg, 0.026 mmol, 88%).

[PhNNC(C6F5)NNMes]BF2[Cp2Co] [6g][Cp2Co]. The procedure is the same as [PhNNC(p- tolyl)NNPh]BF2[Cp2Co]. 29.8 mg of [PhNNC(C6F5)NNMes]BF2 (0.06 mmol) and 13.4 mg of

Cp2Co (0.07 mmol) were used. After reaction, 41.1 mg of [PhNNC(C6F5)NNMes]BF2[Cp2Co] (0.06 mmol, 99 %) was obtained.

121

Chapter 5

[6k][Cp2Co]2. The procedure is the same as [PhNNC(p-tolyl)NNPh]BF2[Cp2Co]. 19.7 mg of

6k (0.03 mmol) and 14.4 mg of Cp2Co (0.08 mmol) were used. After reaction, 30.1 mg of

[6k][Cp2Co]2 (0.03 mmol, 99 %) was obtained.

[6l][Cp2Co]2. The procedure is the same as [PhNNC(p-tolyl)NNPh]BF2[Cp2Co]. 19.8 mg of

6l (0.03 mmol) and 14.6 mg of Cp2Co (0.08 mmol) were used. After reaction, 32.4 mg of

[6l][Cp2Co]2 (0.03 mmol, 99 %) was obtained.

Crystallographic data

Suitable crystals of 6a, 6b, 6e, 6j, 6k, 6l, 9a, 10a, [6a]Cp2Co, [6c][(15-C-5)Na] , [6g]Cp2Co, and [6k][Cp*2Co]2 were mounted on a cryo-loop in a drybox and transferred, using inert- atmosphere handling techniques, into the cold nitrogen stream of a Bruker D8 Venture diffractometer. The final unit cell was obtained from the xyz centroids of 9981 (6a), 9403

(6b), 1241 (6e), 9954 (6j), 6441 (6k), 9926 (6l), 9867 (9a), 9915 (10a), 9687 ([6a]Cp2Co),

9745 ([6c][(15-C-5)Na]), 9965 ([6g]Cp2Co), and 9958 ([6k][Cp*2Co]2) reflections after integration. Intensity data were corrected for Lorentz and polarisation effects, scale variation, for decay and absorption: a multiscan absorption correction was applied, based on the intensities of symmetry-related reflections measured at different angular settings (SADABS).24 The structures were solved by direct methods using the program SHELXS.25 The hydrogen atoms were generated by geometrical considerations and constrained to idealised geometries and allowed to ride on their carrier atoms with an isotropic displacement parameter related to the equivalent displacement parameter of their carrier atoms. Structure refinement was performed with the program package SHELXL.25 Crystal data and details on data collection and refinement are presented in following tables.

Crystallographic data 6a 6b 6e 6j chem formula C20H17BF2N4 C17H19BF2N4 C20H12BF7N4 C20H22BF2N5 Mr 362.19 328.17 452.15 381.23 cryst syst monoclinic orthorhombic monoclinic orthorhombic color, habit red, needle orange, block red, block Yellow, needle size (mm) 0.34 x 0.24 x 0.05 0.30 x 0.10 x 0.08 0.35 x 0.31 x 0.04 0.33 x 0.04 x 0.03 space group P21/c Pnma P21/c Pnma a (Å) 10.4160(6) 16.6601(5) 18.5945(8) 5.9716(2) b (Å) 18.8213(11) 17.4301(5) 7.3570(3) 17.0399(6) c (Å) 9.2633(5) 5.6620(2) 14.5283 18.8991(6)  (°) β (°) 101.539(2) 108.9680(10)  (°) V (Å3) 1779.30(17) 1644.17(9) 1879.55(14) 1923.09(11) Z 4 4 4 4 -3 calc, g.cm 1.352 1.326 1.598 1.317 Radiation [Å] Mo Kα 0.71073 Mo Kα 0.71073 Mo Kα 0.71073 Mo Kα 0.71073

122

(Formazanate)Boron Complexes

-1 µ(Mo K), mm 0.096 0.096 0.145 0.094 -1 µ(Cu K), mm F(000) 752 688 912 800 temp (K) 100(2) 100(2) 100(2) 100(2)  range (°) 5.80–55.82 3.38-27.14 2.81-27.13 3.22-27.15 data collected -13:13, -24:24, - -21:21; -22:22; - -23:23; -9:9; - -7:7; -21:21; - (h,k,l) 11:12 7:7 18:18 24:24 min, max transm 0.9680, 0.9952 0.6954, 0.7455 0.6733, 0.7455 0.7013, 0.7455 rflns collected 49978 24656 32581 29576 indpndt reflns 4282 1877 4156 2201 observed reflns Fo 3564 1551 3175 1796  2.0 σ (Fo) R(F) (%) 3.74 3.42 3.94 3.70 wR(F2) (%) 8.92 7.77 9.33 8.52 GooF 1.037 1.089 1.020 1.020 weighting a,b 0.0416, 0.7967 0.0363, 0.5435 0.0507, 0.9616 0.0395, 1.1881 params refined 245 120 290 139 min, max resid -0.219, 0.379 -0.224, 0.298 -0.214, 0.308 -0.255, 0.317 dens

6k 6l 9a 10a chem formula C48H36B3F6N12 C32H24B2F4N8 C32H27BN4 C20H18BN4 Mr 927.32 618.21 478.38 326.20 cryst syst triclinic monoclinic monoclinic monoclinic color, habit red, plate red, needle red, block purple, plate size (mm) 0.38 x 0.03 x 0.01 0.25 x 0.14 x 0.05 0.32 x 0.18 x 0.08 0.28 x 0.12 x 0.03 space group P-1 C2/c P21c Cc a (Å) 11.5188(11) 18.4424(8) 9.5375(3) 4.14730(10) b (Å) 14.1771(10) 17.9458(8) 15.5183(5) 20.9249(7) c (Å) 15.2919(12) 18.5978(8) 17.4792(6) 19.0650(6)  (°) 65.001(2) β (°) 69.085(2) 110.9447(14) 99.2996(12) 94.7599(12)  (°) 82.791(3) V (Å3) 2113.2(3) 5748.5(4) 2553.02(14) 1648.79(9) Z 2 8 4 4 -3 calc, g.cm 1.457 1.429 1.245 1.314 Radiation [Å] Cu Kα 1.54178 Mo Kα 0.71073 Mo Kα 0.71073 Mo Kα 0.71073 -1 µ(Mo K), mm 0.106 0.074 0.079 -1 µ(Cu K), mm 0.902 F(000) 954 2544 1008 688 temp (K) 141(2) 100(2) 100(2) 100(2)  range (°) 3.39-54.35 3.19-27.15 2.88-27.15 3.76-27.10 data collected -11:12; -14:14; - -23:23; -22:20; - -12:12; -15:19; - -5:5; -26:26; - (h,k,l) 14:15 23:23 22:22 24:23 min, max transm 0.6726, 0.7507 0.7101, 0.7455 0.7180, 0.7455 0.6987, 0.7455 rflns collected 13063 54862 31641 16573 indpndt reflns 4998 6370 5633 3350 observed reflns Fo 3876 5095 4698 3193  2.0 σ (Fo) R(F) (%) 4.55 3.95 3.80 3.25 wR(F2) (%) 9.83 9.07 9.04 8.32 GooF 1.045 1.017 1.041 1.070 weighting a,b 0.0430, 1.7775 0.0416, 6.1438 0.0421, 1.0988 0.0496, 0.6185 params refined 622 415 335 235 min, max resid -0.201, 0.368 -0.214, 0.332 -0.219, 0.307 -0.216, 0.181 dens

123

Chapter 5

[Cp2Co][6a] [Na(15-C-5)][6c] [Cp2Co][6g] [Cp*2Co]2[6k] chem formula C33.25H33BCoF2N4 C33H43BF2N4NaO5 C32H26BCoF7N4 C72H84B2Co2F4N8 O Mr 612.88 647.51 669.31 1276.95 cryst syst monoclinic monoclinic monoclinic triclinic color, habit green, platelet green, block green, plate green, block size (mm) 0.30 x 0.18 x 0.06 0.16 x 0.16 x 0.10 0.16 x 0.11 x 0.01 0.17 x 0.12 x 0.05 space group P21/n P21 P21/c P-1 a (Å) 12.2031(7) 11.0275(9) 15.6779(3) 11.6834(7) b (Å) 17.4363(9) 12.8604(9) 24.3649(5) 12.0343(7) c (Å) 14.4788(7) 23.3615(18) 15.2499(3) 16.1553(10)  (°) 90.699(2) β (°) 97.446(2) 93.306(2) 93.9593(9) 105.390(2)  (°) 117.4381(17) V (Å3) 3054.8(3) 3307.6(4) 5811.4(2) 1919.5(2) Z 4 4 8 1 -3 calc, g.cm 1.333 1.300 1.530 1.105 Radiation [Å] Mo Kα 0.71073 Mo Kα 0.71073 Cu Kα 1.54178 Mo Kα 0.71073 -1 µ(Mo K), mm 0.607 0.106 0.483 -1 µ(Cu K), mm 5.302 F(000) 1276 1372 2728 672 temp (K) 200(2) 100(2) 100(2) 100(2)  range (°) 5.675-53.32 2.93-27.14 3.36-65.30 2.83-27.33 data collected -15:15, -22:22, - -14:14; -13:16; - -18:18; -28:28; - -15:15; -15:15; - (h,k,l) 18:17 29:29 17:17 20:20 min, max transm 0.8389, 0.9645 0.6742, 0.7455 0.5549, 0.7526 0.700, 0.746 rflns collected 105601 64630 54498 51828 indpndt reflns 6775 13766 9908 8571 observed reflns Fo 5657 11734 8398 7022  2.0 σ (Fo) R(F) (%) 4.37 8.04 3.42 5.49 wR(F2) (%) 12.67 19.71 7.41 14.77 GooF 1.070 1.048 1.027 1.073 weighting a,b 0.0770, 1.6064 0.0948, 8.6049 0.0320, 3.9578 0.0716, 3.1127 params refined 455 837 817 411 min, max resid -0.512, 0.960 -0.856, 2.314 -0.231, 0.340 -0.747, 0.637 dens

5.8 References (1) (a)Martin, A.; Moriarty, R. D.; Long, C.; Forster, R. J.; Keyes, T. E. Asian J. Org. Chem., 2013, 2, 763–778. (b)Sreenath, K.; Yuan, Z.; Allen, J. R.; Davidson, M. W.; Zhu, L. Chem. Eur. J., 2015, 21, 867–874. (c)Yamada, A.; Hiruta, Y.; Wang, J.; Ayano, E.; Kanazawa, H. Biomacromolecules, 2015, 16, 2356–2362. (d)Agarwal, P.; Beahm, B. J.; Shieh, P.; Bertozzi, C. R. Angew. Chem. Int. Ed., 2015, 54, ASAP. (2) (a)Hudnall, T. W.; Gabbaï, F. P. Chem. Commun., 2008, 38, 4596–4597. (b)Kuperman, M. V.; Chernii, S. V.; Losytskyy, M. Y.; Kryvorotenko, D. V.; Derevyanko, N. O.; Slominskii, Y. L.; Kovalska, V. B.; Yarmoluk, S. M. Anal. Biochem., 2015, 484, 9–17. (3) (a)Prokhorov, A. M.; Hofbeck, T.; Czerwieniec, R.; Suleymanova, A. F.; Kozhevnikov, D. N.; Yersin, H., J. Am. Chem. Soc. 2014, 136, 9637–9642. (b)Cao, X.; Miao, J.; Zhu, M.; Zhong, C.; Yang, C.; Wu, H.; Qin, J.; Cao, Y. Chem. Mater., 2014, 27, 96-104. (c)Yao, L.; Zhang, S.; Wang, R.; Li, W.; Shen, F.; Yang, B.; Ma, Y. Angew. Chem. Int. Ed., 2014, 53, 2119-2123. (4) (a)Topel, S. D.; Cin, G. T.; Akkaya, E. U. Chem. Commun., 2014, 50, 8896-8899. (b)Takatoshi Yogo; Yasuteru Urano; Yukiko Ishitsuka; Fumio Maniwa, A.; Tetsuo Nagano. J. Am. Chem. Soc., 2005, 127, 12162-12163.

124

(Formazanate)Boron Complexes

(5) Erten-Ela, S.; Yilmaz, M. D.; Icli, B.; Dede, Y.; Icli, S.; Akkaya, E. U. Org. Lett., 2008, 10, 3299-3302. (6) Ntziachristos, V.; Ripoll, J.; Weissleder, R. Opt. Lett., 2002, 27, 333–335. (7) (a)Farley, S. J.; Rochester, D. L.; Thompson, A. L.; Howard, J. A. K.; Williams, J. A. G. Inorg. Chem., 2005, 44, 9690–9703. (b)Wenwu Qin; Mukulesh Baruah; Mark Van der Auweraer; Frans C De Schryver, A.; Boens, N. J. Phys. Chem. A., 2005, 109, 7371-7384. (c) Viger, M. L.; Live, L. S.; Therrien, O. D.; Boudreau, D. Plasmonics, 2008, 3, 33–40. (8) Loudet, A.; Burgess, K. Chem. Rev., 2007, 107, 4891–4932. (9) (a)Barbon, S. M.; Staroverov, V. N.; Boyle, P. D.; Gilroy, J. B., Dalton Trans., 2013, 43, 240–250. (b)Macedo, F. P.; Gwengo, C.; Lindeman, S. V.; Smith, M. D.; Gardinier, J. R., Eur. J. Inorg. Chem., 2008, 20, 3200–3211. (c)Qian, B.; Baek, S. W.; Smith, M. R., III. Polyhedron, 1999, 18, 2405–2414. (10) Araneda, J. F.; Piers, W. E.; Heyne, B.; Parvez, M.; McDonald, R. Angew. Chem. Int. Ed., 2011, 50, 12214–12217. (11) Yu, C.; Jiao, L.; Zhang, P.; Feng, Z.; Cheng, C.; Wei, Y.; Mu, X.; Hao, E. Opt. Lett., 2014, 39, 3048– 3051. (12) (a)Nawn, G.; Oakley, S. R.; Majewski, M. B.; McDonald, R.; Patrick, B. O.; Hicks, R. G. Chem. Sci., 2013, 4, 612-621. (b)Nawn, G.; Waldie, K. M.; Oakley, S. R.; Peters, B. D.; Mandel, D.; Patrick, B. O.; McDonald, R.; Hicks, R. G. Inorg. Chem., 2011, 50, 9826–9837. (13) (a) Shie, J.-J.; Liu, Y.-C.; Lee, Y.-M.; Lim, C.; Fang, J.-M.; Wong, C.-H. J. Am. Chem. Soc., 2014, 136, 9953–9961. (b) Rosenthal, J.; Lippard, S. J. J. Am. Chem. Soc., 2010, 132, 5536–5537. (14) (a)Duran-Sampedro, G.; Esnal, I.; Agarrabeitia, A. R.; Bañuelos Prieto, J.; Cerdán, L.; Garcia-Moreno, I.; Costela, A.; López-Arbeloa, I.; Ortiz, M. J. Chem. Eur. J., 2014, 20, 2646–2653. (b)Lundrigan, T.; Crawford, S. M.; Cameron, T. S.; Thompson, A. Chem. Commun., 2012, 48, 1003–1005. (c)More, A. B.; Mula, S.; Thakare, S.; Sekar, N.; Ray, A. K.; Chattopadhyay, S. J. Org. Chem., 2014, 79, 10981– 10987. (d)Lundrigan, T.; Thompson, A. J. Org. Chem., 2013, 78, 757–761. (e)Sampedro, G. D.; Agarrabeitia, A. R. Adv. Funct. Mater., 2013, 23, 4195-4205. (15) Martin, A.; Long, C.; Forster, R. J.; Keyes, T. E. Chem. Commun., 2012, 48, 5617–5619. (16) (a)Barbon, S. M.; Staroverov, V. N.; Gilroy, J. B. J. Org. Chem., 2015, 80, 5226–5235. (b)Hesari, M.; Barbon, S. M.; Staroverov, V. N.; Ding, Z.; Gilroy, J. B. Chem. Commun., 2013, 51, 3766–3769. (c)Barbon, S. M.; Reinkeluers, P. A.; Price, J. T.; Staroverov, V. N.; Gilroy, J. B. Chem. Eur. J., 2014, 20, 11340–11344. (d)Barbon, S. M.; Price, J. T.; Reinkeluers, P. A.; Gilroy, J. B. Inorg. Chem., 2014, 53, 10585–10593. (17) Maar, R.R.; Barbon, S.M.; Sharma, N.; Groom, H.; Luyt, L.G.; Gilroy, J.B. Chem. Eur. J. 2015, 36, In Press. (18) Barbon, S. M.; Price, J. T.; Yogarajah, U.; Gilroy, J. B. RSC Adv., 2015, 5, 56316-56324. (19) Hermanek, S. Chem. Rev., 1992, 92, 325–362. (20) (a)Gilroy, J. B.; McKinnon, S. D.; Kennepohl, P.; Zsombor, M. S.; Ferguson, M. J.; Thompson, L. K.; Hicks, R. G. J. Org. Chem., 2007, 72, 8062–8069. (b)Anderson, K. J.; Gilroy, J. B.; Patrick, B. O.; McDonald, R.; Ferguson, M. J.; Hicks, R. G. Inorg. Chim. Acta., 2011, 374, 480–488. (c)Brook, D. J. R.; Fornell, S.; Stevens, J. E.; Noll, B.; Koch, T. H., Eisfeld, W. Inorg. Chem., 2000, 39, 562–567. (21) Montgomery, L. K.; Huffman, J. C.; Jurczak, E. A.; Grendze, M. P. J. Am. Chem. Soc., 2002, 124, 6004–6011. (22) (a)Fico, R. M., Jr; Hay, M. F.; Reese, S.; Hammond, S.; Lambert, E.; Fox, M. A. J. Org. Chem., 1999, 64, 9386–9392. (b)Nakazono, S.; Karasawa, S.; Koga, N.; Iwamura, H. Angew. Chem. Int. Ed., 1998, 37, 1550–1552. (c)Ko, K. C.; Park, Y. G.; Cho, D.; Lee, J. Y. J. Phys. Chem. A., 2014, 118, 9596– 9606. (d)Bhattacharya, D.; Misra, A. J. Phys. Chem. A., 2009, 113, 5470–5475. (23) Koivisto, B. D.; Hicks, R. G. Coord. Chem. Rev., 2005, 249, 2612–2630. (24) Bruker. APEX2 (v2012.4-3), SAINT (Version 8.18C) and SADABS (Version 2012/1). Bruker AXS Inc., Madison, Wisconsin, USA. 2012. (25) Sheldrick, G. Acta Crystallographica Section A 2008, 64, 112.

125

Chapter 5

126

Chapter 6

Reduction of (Formazanate)Boron Difluoride Provides Evidence for an N- Heterocyclic B(I) Carbenoid Intermediate

Despite the current interest in structure and reactivity of sub-valent main group compounds, neutral boron analogues of N-heterocyclic carbenes have been elusive due to their high reactivity. Here we provide evidence that 2-electron reduction of a (formazanate)BF2 (6) precursor leads to formation of an N-heterocyclic boron carbenoid, and describe the formation of a series of unusual BN heterocycles that result from trapping of this fragment. Subsequent chemical oxidation by XeF2 demonstrates that the trapped (formazanate)B fragment retains carbenoid character and regenerates the boron difluoride starting material in good yield. These results indicate that the formazanate ligand framework provides a unique entry into sub-valent boron chemistry.

Parts of this chapter have been published:

M.-C. Chang and E. Otten* “Reduction of (Formazanate)boron Difluoride Provides Evidence for an N‑Heterocyclic B(I) Carbenoid Intermediate” Inorg. Chem., 2015, 54, 8656-8664 Chapter 6

Chapter 6 Reduction of (Formazanate)Boron Difluoride Provides Evidence for an N-Heterocyclic B(I) Carbenoid Intermediate

6.1 Introduction

The synthesis of reactive compounds with novel or unusual bonding motifs has fascinated chemists for centuries and has led to new insight into the nature and stability of chemical bonds. For example, despite the (perceived) high reactivity of carbon in its divalent state, the successful stabilization of such compounds via substituent effects has led to the now ubiquitous N-heterocyclic carbenes (NHCs).1 Recently, much work has focused on the synthesis of related low-valent compounds of heavier group 14 elements, which show reactivity profiles beyond that of the carbon analogs.2 In contrast, sub-valent compounds of the group 13 elements have received considerably less attention despite their involvement in a variety of chemical transformations. While low-valent species of the heavier group 13 elements are stabilized due to the 'inert-pair' effect, molecular boron(I) compounds are especially rare due to their high reactivity. Nevertheless, in the coordination sphere of transition metal centers borylenes can be stable and show very rich coordination chemistry as well as reactivity.4 Early work on free monomeric borylenes prepared by thermolysis of boron halides,5 reduction of organodihaloboranes6 or photolysis of triarylboranes7 showed these species to be highly reactive and undergo C-H insertion and C=C addition reactions. More recently, matrix isolation studies have allowed spectroscopic identification of PhB,8 and high- level theoretical calculations on the reactivity of free borylenes have been reported.9 In the past decade, it has been shown that Lewis bases, NHCs in particular, can stabilize boron in unusual coordination environments. For example, neutral diborene10 and diboryne11 compounds have been prepared that are stable at room temperature. In a similar fashion, attempted preparation of base-stabilized borylenes has been reported, but these also are often still highly reactive towards C-H and C=C bonds,12 and isolable carbene-stabilized borylenes have only been reported recently (e.g., A and B in Chart 6.1).13 Attempts to obtain monomeric B(I) compounds as their N-heterocyclic derivatives (analogous to NHCs) have mostly been thwarted by their high reactivity, but in 2006 the group of Nozaki and co-workers described the isolation of the first nucleophilic N-heterocyclic boryl compound (C).14 Neutral sub-valent compounds with -diketiminate ligands have been prepared for the heavier group 13 elements (D),15 but the boron analogues are unknown. The absence of isolable boron compounds of this type likely reflects the small singlet-triplet separation that is calculated to be ca. 3.5

128

Reduction of (Formazanate)Boron Difluoride Provides Evidence for an N‐Heterocyclic B(I) Carbenoid Intermediate kcal/mol.16 Aldridge and co-workers recently described the assembly of the -diketiminate boron fragment HC[(CMe)(NMes)]2B in the coordination sphere of an iron complex by a metal-templated approach,17 but the direct synthesis of a neutral boron-analogue of NHCs remains elusive.

Chart 6.1

In this chapter we report that Na/Hg reduction of a (formazanate)BF2 compound (6) results in a series of trimeric products that incorporate an intact N-heterocyclic (formazanate)B fragment, from which the starting material may be regenerated upon treatment with XeF2. The experimental data is complemented by a computational study, which suggests the involvement of a relatively stable (formazanate)B carbenoid intermediate.

6.2 Synthesis and Characterization

6.2.1 Synthesis

The synthesis and characterization of (formazanate)boron difluoride complexes (LBF2, 6) have been described in Chapter 4. Cyclic voltammetry (CV) of 6a showed two quasi- reversible reductions at -0.98 and -2.06 V vs. Fc0/+, indicative of ligand-based redox- chemistry. Compound 6a can be converted to a green radical anion [(PhNNC(p- - - tolyl)NNPh)BF2] ([6a] ) upon chemical reduction using cobaltocene (Cp2Co) as the reducing agent, and the spectroscopic and structural changes accompanying this conversion indicate a 'borataverdazyl'-type structure for [6a]-. Accessing the second reduction product observed by

129

Chapter 6

CV requires a reducing agent stronger than Cp2Co. To explore the possibility of isolating the putative 2-electron reduction product, experiments were carried out in which 6a was reacted with 2 equivalents of sodium amalgam (Na/Hg) in a THF/toluene solvent mixture, which resulted in a gradual color change from red to purple. NMR analysis of the crude reaction mixture indicated the presence of several diamagnetic products, none of which contained fluorine (on the basis of 19F NMR spectroscopy). Fractional crystallization allowed the isolation and structural characterization of the 4 main components in the mixture, accounting in total for ca. 45 % of the starting material. The major component, compound 11, was isolated in 23 % yield and characterized by single-crystal X-ray crystallography. The structure determination of 11 established that it contains a central B3N3 heterocyclic core (Scheme 6.1). Except for the fluorine atoms, all other atoms originally present in the boron starting material are retained in 11: its composition suggests it is a trimer of (formazanate)boron-derived products. On the basis of the observed atom connectivity, we propose compound 11 to be formed by the assembly of the putative reduction intermediates X and Y in a 1:2 ratio (Scheme 6.1).

Compound 11 does not contain fluorine, suggesting that the initial 2-electron reduction -2 + product [LBF2] is unstable in the presence of Na cations (from the Na/Hg reducing agent). Although the B-F bond is among the most stable chemical bonds (B-F BDE = 732 kJ/mol,18 19 18 calculated for BF3 = 720-724 kJ/mol) the lattice energy of NaF (910 kJ/mol) provides the - 2- driving force for the loss of 2 equivalents of F from [LBF2] to (transiently) generate fragment X. Unfortunately, only alkali metal-based reducing agents are sufficiently reducing 2- 20 to provide access to [LBF2] . It is likely that these will all result in fluoride abstraction, 2- resulting in products that are different from [LBF2] which is observed by electrochemical methods. Of the fragments that constitute 11, one contains an intact boron formazanate moiety (fragment X) while for the other N-N bond cleavage has occurred to give an imidoborane (fragment Y). The latter transformation has precedent in β-diketiminate chemistry, where reductive cleavage of the ligand backbone has been observed both in transition metal21 and main group chemistry.22

130

Reduction of (Formazanate)Boron Difluoride Provides Evidence for an N‐Heterocyclic B(I) Carbenoid Intermediate

Ph N N F p-tol B NN F Ph 6a

Na/Hg THF/Toluene (2 eq)

p-tol p-tol p-tol p-tol N N N N N N N Ph NN NN NN N N Ph Ph Ph N p-tol Ph Ph p-tol Ph p-tol Ph Ph B Ph Ph B B Ph B N N N N N N N N N + + + N p-tol Ph N Ph p-tol N B N B B N B B N B N N Ph N N N N B N N N B N N Ph Ph N N Ph p-tol N Ph N Ph p-tol Ph p-tol N N Ph [Na(DME) ] Ph 3

11 (X + 2Y) 12 (2X + Y) 13 (X + Y +Z) 14 (X-X + Y)

22.7% 6.1% 11.3% 4.0%

Ph Ph N Ph Ph p-tol N N N N N N N N p-tol B p-tol BN N B BN N B NN N Ph N N p-tol Ph N Ph N p-tol Ph Ph Ph fragment X fragment Y fragment Z

Scheme 6.1 Synthesis of compounds 11-14 with their constituent fragments between brackets. The yields given correspond to isolated, crystalline material. A second species, compound 12, was separated from the reaction mixture in 6 % yield as crystalline material, and this was also subjected to X-ray diffraction analysis. The molecular structure of 12 contains a central 7-membered [B3N4] ring. Similar to 11, compound 12 is also a hetero-trimer but it consists of two intact (formazanate)boron moieties (X) and one fragment (Y) that results from N-N bond cleavage (Scheme 6.1). For the intact formazanates in 12, one is bound to the boron center via the two terminal nitrogen atoms of the NNCNN backbone while the other forms a 5-membered chelate ring in which one of the terminal nitrogen atoms is available to bridge to a second B center.

Two additional products could reproducibly be isolated by fractional crystallization. Compound 13, which was obtained in 11 % yield, was shown by X-ray crystallography to contain one intact (formazanate)B fragment incorporated into a B3N3 core that is similar to 11 (Scheme 6.1). However, the atom connectivity in the 6-membered chelate ring around B(3) is unexpected: it contains a NNNCN backbone in which a terminal N-Ph group has migrated onto the other N-Ph moiety of the formazanate ligand to result in a triazenyl heterocycle (fragment Z, vide infra). Finally, a fourth component (compound 14) could be obtained from

131

Chapter 6

the reaction mixture in 4 % yield and shown to be an ionic product in which the anionic part is composed of 3 (formazanate)boron-derived fragments that are assembled into a 7- + membered B3N4 core similar to 12 (Scheme 6.1). The countercation is Na , which is coordinated by 3 molecules of DME (the solvent of crystallization). The salient feature of 14 is the presence of a phenylene linker that connects two formazanate fragments. While the pathway to formation of the ortho-disubstituted C6H4 fragment in 14 is not known in detail, it is likely that it involves addition of a nitrogen-centered radical to the C6H5 ring via homolytic aromatic substitution.23

6.2.2 X-ray Crystallography

Due to the weak diffraction of crystals of 11 and 12, for which only small platelets could be obtained, the crystallographic data for these compounds are of low quality (Figure 6.1, metrical parameters in Table 6.1). Nevertheless, the atom connectivity is clearly established, and a brief discussion of the metrical parameters is included below. The crystal structures of 11-13 show similar metrical parameters for the intact 6-membered formazanate chelate ring. The N-N bond distances vary little (11: 1.317(6)/1.316(6) Å; 12: 1.302(8)/1.322(8) Å; 13: 1.308(2)/1.298(2) Å) and are similar to the values found in the starting material 6a (1.3080(13)/1.3078(13) Å). This indicates that the formazanate is retained as a delocalized monoanionic ligand (L-), with little or no contribution from the radical dianionic form (L2-) that we characterized in the Chapter 3 and Chapter 4.

The B(1) boron atom in compounds 11-14 is tetracoordinate, while the boron centers B(2) and B(3) are tricoordinate. Of the 4 B-N bonds around the tetracoordinate B(1) atom in compound

11, those to the nitrogen atoms in the central B3N3 ring are the shortest (B(1)-N(5) = 1.510(8) Å and B(1)-N(10) = 1.501(8) Å), but still significantly elongated in comparison to those for the tricoordinate B atoms (B-N distances 1.409(8)-1.460(9) Å). Similar values are found for the other compounds. The B3N3 core in 11 is somewhat reminiscent of borazine (HN=BH)3, often referred to as the inorganic analogue of benzene,24 but in contrast to 11, borazine shows equivalent (delocalized) B-N bonds of 1.430 Å.24a While recent calculations on donor/acceptor complexes of borazine and its derivatives suggested that binding of an 25 external Lewis base (NH3) is not energetically favorable, the presence of 4-coordinate boron atoms around the B3N3 core in 11 and 13 suggests these to have significant acceptor properties. The metrical parameters for the B3N3 core in 13 are similar to those in 11, and both are virtually planar. In addition to the central B3N3 ring, compound 13 contains a highly

132

Reduction of (Formazanate)Boron Difluoride Provides Evidence for an N‐Heterocyclic B(I) Carbenoid Intermediate puckered BNNNCN 6-membered heterocycle in which the triazenyl chain shows long consecutive N-N bonds in the N(9)-N(10)-N(11) fragment of 1.413(2) and 1.403(2) Å. A similarly long N-N bond is found in the puckered B3N4 core of 12 (N(11)-N(12) = 1.411(9) Å). These N-N bonds are significantly elongated in comparison to those in delocalized formazanate ligands and are indicative of N-N single bond character. The [B3N4] core in 14 is similar to that in 12 with both compounds adopting a pseudo-boat conformation of the 7- membered heterocycle.

Figure 6.1 Molecular structures of compounds 11-14 are showing 50% probability ellipsoids, + and selected bond distances (Å). The Na(DME)3 counter-ion and hexane solvent molecule in 14, the toluene molecule in 13 and all hydrogen atoms are omitted; Ph and p-tol groups are shown as wireframe for clarity, except for the NPh groups in 14 that are coupled.

133

Chapter 6

Table 6.1 Selected bond length (Å) of compound 11-14 11 12 13 14 N1-N2 1.317(6) N1-N2 1.302(8) N1-N2 1.308(2) N1-N2 1.397(3) N3-N4 1.316(6) N3-N4 1.322(8) N3-N4 1.298(2) N3-N4 1.418(3) B1-N1 1.589(9) B1-N1 1.614(11) B1-N1 1.582(3) N5-N6 1.411(3) B1-N4 1.576(9) B1-N4 1.539(11) B1-N4 1.602(3) N7-N8 1.417(4) B1-N5 1.510(8) B1-N5 1.522(11) B1-N7 1.509(3) B1-N1 1.544(4) N5-B2 1.410(9) N5-B2 1.441(11) N7-B2 1.437(3) B1-N3 1.547(4) B2-N9 1.461(9) B2-N6 1.451(11) B2-N8 1.442(3) B1-N5 1.557(4) N9-B3 1.425(9) N6-B3 1.444(11) N8-B3 1.419(3) N5-N6 1.411(3) B3-N10 1.444(9) B3-N11 1.411(11) B3-N12 1.453(3) N6-B2 1.422(4) N10-B1 1.501(8) N11-N12 1.411(11) N12-B1 1.536(3) B2-N11 1.433(4) N12-B1 1.524(12) N11-B3 1.463(4) B3-N12 1.444(4) N12-B1 1.539(4)

6.2.3 NMR Spectroscopy and UV-Vis Analysis

Although compounds 11-14 are diamagnetic, their NMR spectra are not very informative (Figure 6.2). The 1H NMR spectra contain several overlapping sets of resonances in the aromatic region. More diagnostic is the aliphatic region: as expected, each compound shows 3 separate singlets in the range of  2.0-2.5 ppm, consistent with 3 different p-tolyl CH3 environments as required for the C1 symmetric trimers observed in the solid state. Furthermore, the 11B NMR spectrum for each compound shows two resonances: a broad signal around  26 ppm (FWHH > 700 Hz) and a somewhat sharper one around  0 ppm (FWHH < 220 Hz) that are attributed to the 3- and 4-coordinate B centers, respectively.

134

Reduction of (Formazanate)Boron Difluoride Provides Evidence for an N‐Heterocyclic B(I) Carbenoid Intermediate

Compound 11

8.4 8.2 8.0 7.8 7.6 7.4 7.2 7.0 6.8 6.6 6.4 6.2 6.0 5.8 5.6 5.4 5.2 5.0 4.8 4.6 4.4 4.2 4.0 3.8 3.6 3.4 3.2 3.0 2.8 2.6 2.4 2.2 2.0 1.8 1.6 f1 (ppm)

Compound 12

8.4 8.2 8.0 7.8 7.6 7.4 7.2 7.0 6.8 6.6 6.4 6.2 6.0 5.8 5.6 5.4 5.2 5.0 4.8 4.6 4.4 4.2 4.0 3.8 3.6 3.4 3.2 3.0 2.8 2.6 2.4 2.2 2.0 1.8 1.6 f1 (ppm)

Compound 13

8.4 8.2 8.0 7.8 7.6 7.4 7.2 7.0 6.8 6.6 6.4 6.2 6.0 5.8 5.6 5.4 5.2 5.0 4.8 4.6 4.4 4.2 4.0 3.8 3.6 3.4 3.2 3.0 2.8 2.6 2.4 2.2 2.0 1.8 1.6 f1 (ppm)

Compound 14

8.4 8.2 8.0 7.8 7.6 7.4 7.2 7.0 6.8 6.6 6.4 6.2 6.0 5.8 5.6 5.4 5.2 5.0 4.8 4.6 4.4 4.2 4.0 3.8 3.6 3.4 3.2 3.0 2.8 2.6 2.4 2.2 2.0 1.8 1.6 f1 (ppm) 1 Figure 6.2 H NMR spectra of 11-13 (in CD2Cl2, RT) and 14 (in d8-THF, -60Ԩ)

135

Chapter 6

UV-Vis spectroscopy for 11-13 in THF solution (Figure 6.3) shows broad absorption features in the visible range of the spectrum that account for their intense color ( = 19000 - 30000 L·mol−1·cm−1). Compound 11 has a maximum absorption at 557 nm that is due to the * 26 formazanate π-π transition, similar to that observed in 6a (max = 521 nm, in Chapter 4). Compounds 12 and 13 show absorption maxima at lower and higher wavelengths (527 and 601 nm, respectively). Conversely, the appearance of the spectrum of 14 is quite different −1 −1 with a much less intense absorption at max = 408 nm ( = 6400 L·mol ·cm ). The difference between 11-13 and 14 is due to the absence of a delocalized 6-membered formazanate [NNCNN] chelate ring in 14. Formazanate boron difluorides (6) have recently been investigated as analogues of BODIPY dyes and showed tunable luminescence properties, with quantum yields of up to 77% for compounds with a 3-cyanoformazanate ligand.26a,27 We thus investigated the emission spectra of 11-14 in THF solution. Whereas the neutral compounds 11-13 are only weakly emissive, compound 14 shows a relatively intense emission band at 477 nm (Stokes shift of 69 nm) upon excitation at 400 nm with a quantum yield of 6 %.

Figure 6.3 UV-Vis spectra for compounds 11-14 in THF solution. Inset: Emission spectrum of 14 showing normalized intensities.

6.2.4 Reduction Chemistry

The BN-heterocycles 11-14 contain an intact formazanate unit that can be expected to show (reversible) redox-chemistry that is typical of the formazanate NNCNN ligand backbone. To test this hypothesis, cyclic voltammetry was recorded in THF solution using [Bu4N][PF6] as

136

Reduction of (Formazanate)Boron Difluoride Provides Evidence for an N‐Heterocyclic B(I) Carbenoid Intermediate the supporting electrolyte. The data for 11 show a complicated electrochemical response, but two quasi-reversible 1-electron redox processes can be located by scanning in smaller regions (Figure 6.4, left). Conversely, compounds 12 and 13 show quasi-reversible, sequential 1- electron redox processes (Figure 6.4, right) that are reminiscent of those observed for 6a. These correspond to the reversible formation of the radical anions of 11–, 12– or 13– (-1.25, -1.13 and -1.06 V vs Fc0/+) and the corresponding dianions (112–, 122– or 132–, -2.61, -2.26 and -2.35 V vs Fc0/+), respectively. The first reduction occurs at more negative potential than that in the boron difluoride starting material 6a (-0.98 V vs Fc0/+, in Chapter 4), and 11 is harder to be reduced than 12 and 13. For the second reduction to 122- and 132- the order is reversed, and 112- is still the hardest one to be reduced.

Figure 6.4 Cyclic voltammograms of 11 (top, green line), 12 (bottom, red line) and 13 -1 (bottom, blue line) in THF (0.1M [Bu4N][PF6]) recorded at 100 mVs (potential in V vs. Fc0/+).

137

Chapter 6

Although the cyclic voltammetry data suggest that the radical anion 11– might not be stable, we nevertheless attempted its synthesis. Chemical reduction of compound 11 was carried out in THF solution using a stoichiometric amount of Na/Hg, which resulted in a color change to very dark green. Upon diffusion of hexane into the THF solution, a black powder is precipitated together with a small amount of green crystalline material. X-ray analysis of the crystals confirmed it to be the expected reduction product [11][Na(THF)3]. Unfortunately, the desired product was always obtained together with the (unidentified) black powder and an analytically pure sample has not been obtained. The X-ray structure determination of + [11][Na(THF)3] shows that a Na countercation (with 3 coordinated THF molecules) is bound to N(11) of the BNNCN 5-membered ring of 11– (Figure 6.5). The radical anion 11– is virtually isostructural to the neutral precursor 11, but the N-N bonds within the intact formazanate NNCNN fragment are elongated significantly (11–: 1.359(4)/1.355(4); 11: 1.317(6)/1.316(6) Å), as expected for a formazanate-based reduction.

Similarly, the reduction of 13 with Na/Hg in THF solution resulted in a color change from blue to deep green, and the radical anion 13– was obtained as its Na+ salt in quantitative yield by recrystallization in the presence of 15-crown-5. The structure determination shows two 13– fragments, one of which contains a Na+(15-crown-5) cation bound to a triazenyl N atom (Figure 6.5). A second Na+(15-crown-5) bridges between the two 13– units via the BNNCN 5-membered rings (Figure 6.5). The metrical parameters for the two independent [13]– moieties are very similar, and only one of them will be discussed. Although the quality of the structure determination of [13][Na(15-c-5)] is negatively affected by the disorder observed in the bridging Na+(15-crown-5), it is clear that the integrity of the BN-heterocycle 13 is retained upon reduction to the radical anion 13–. The formazanate N-N bond lengths range between 1.355(4) and 1.362(4) Å, indicating that the formazanate ligand backbone is the electron acceptor. In addition, the B-N(formazanate) bonds around the 4-coordinate B center are contracted from 1.582(3)/1.602(3) Å in 13 to 1.538(5)/1.540(5) Å in 13–, with concomitant elongation of the B-N bonds to the central 6-membered heterocyclic ring (B(1)-N(7)/B(1)- N(12) in 13: 1.509(3)/1.536(3) Å and 13–: 1.537(5)/1.598(5) Å).

138

Reduction of (Formazanate)Boron Difluoride Provides Evidence for an N‐Heterocyclic B(I) Carbenoid Intermediate

Figure 6.5 Molecular structures of compounds [11][Na(THF)3] (top left), [13][Na(15-c-5)] (one of the independent molecules, top right) and [13][Na(15-c-5)] (showing the disordered Na(15-crown-5) fragment, bottom) showing 50% probability ellipsoids. The Ph/p-tol groups and the C atoms of the THF/15-crown-5 molecules are shown as wireframe for clarity.

Geometry optimizations of the radical anions [11]– and [13]– (gas phase calculations in the absence of the Na+ countercation) at the UB3LYP/6-31G(d) level of theory converged on minima that have very similar metrical parameters as the crystallographically determined structures. In agreement with the experimental data, the SOMO in both radical anions is localized on the formazanate backbone and has N-N anti-bonding character (Figure 6.6).

139

Chapter 6

Figure 6.6 DFT models of [11]– (left) and [13]– (right) showing the SOMO.

6.3 DFT Calculations

To garner insight in the pathway that leads to compounds 11-13, DFT calculations were carried out at the B3LYP/6-31G(d) level in the gas phase. For these calculations, the p-tolyl group in 11-13 was replaced by Ph for computational efficiency. We first evaluated the relative stabilities of the possible intermediates X and Y and compared those to the isolated products. The two-electron reduction of 6a was observed by cyclic voltammetry to occur at 0/+ 2- E½ = -2.06 V vs. Fc , presumably forming the dianionic species [LBF2] . In the presence of + 2- Na , the dianionic species [LBF2] is unstable and rapidly eliminates 2 equivalent of NaF(s), forming the carbenoid intermediate X. The geometry of X was optimized in the gas phase in the closed-shell singlet (Xs), triplet (Xt) and singlet diradical state (XBS). The optimized geometry on the singlet potential energy surface shows a puckered formazanate boron chelate ring for Xs, which is calculated to be higher in energy than the triplet Xt (∆G = 6.8 kcal/mol).

A broken-symmetry, singlet diradical solution (XBS) is shown to be slightly higher in energy than the triplet Xt. Thus, these calculations indicate that the ground state of X contains a ligand-based unpaired electron spin, which is ferromagnetically coupled to a boron-based -1 unpaired electron (Jcalcd = -40.8 cm ). The ground state structure Xt is virtually planar and shows elongated N-N bonds of 1.380 Å, indicative of population of the N-N π* orbitals and the presence of a reduced 'verdazyl'-type radical dianionic ligand (L2-). Thus, the electronic structure of X is different than that of a borylene, with the unpaired electrons in X occupying a sp2 orbital on B and a ligand-based π* orbital which leads to a (triplet) diradical ground state (Figure 6.7). This is in agreement with calculations on related boron compounds.16,28 While

140

Reduction of (Formazanate)Boron Difluoride Provides Evidence for an N‐Heterocyclic B(I) Carbenoid Intermediate the small singlet-triplet gap in neutral N-heterocyclic boron(I) compounds is suggested as the reason for their high reactivity (and absence from the literature),16b our results indicate that the increased stability of the biradical state due to the low-lying formazanate π*-orbital can in fact be used advantageously to allow isolation of several trapped (formazanate)B species.

Figure 6.7 Singly occupied frontier molecular orbitals (left, middle) and spin density plot

(right) for the ground state triplet Xt. Next, we computationally evaluated the formation of imidoborane fragment Y. A transition state for the interconversion of X and Y could be located on the singlet potential energy surface with a Gibbs free energy that is only 8.61 kcal/mol higher than Xs (Scheme 6.2). Conversion to the imidoborane product Y is calculated to be exergonic (∆G = -61.6 kcal/mol) for the most stable E-isomer, and formation of the less stable Z-isomer (which is incorporated into the isolated products) is also favorable with ∆G = -51.42 kcal/mol. The 5-membered ring in fragment Y is planar, and the compound is calculated to have an exocyclic B-N bond that is longer (1.374 Å) than that observed in the related compound [CtBuCHCtBuNAr]B=NMes (1.3396(19) Å).22d Incorporation of fragment Y into the heterocyclic trimers 11 and 12 results in pronounced elongation of the N-N bonds (1.297 Å calculated for Y vs. 1.406(7)/1.412(8) Å in the experimentally observed structures 11/12) and contraction of the B-N(N) bond indicative of a delocalized bonding situation upon formation of the trimers 11/12. Finally, we calculated fragment Z, which is one of the constituents in trimer 13. Z is shown to have a singlet ground state that is slightly more stable than Xs (∆G = -5.39 kcal/mol). It is at present unclear what the mechanism of formation of fragment Z is, but given the relative energies discussed above it seems unlikely that fragment Y is involved as an intermediate. Instead, we tentatively propose that formation of the NCNNN moiety in fragment Z (and thus in 13) occurs in a dimeric or trimeric precursor to 13.

141

Chapter 6

Scheme 6.2 DFT calculated energies G/[H] kcal/mol of fragments X, Y, Z and transition state TS(Xs-Y)

Geometry optimizations of the final products 11, 12 and 13 converged on minima (11calc -

13calc) that are in good agreement with the experimentally determined structures. Specifically, the characteristic N-N bonds in the 6-membered formazanate chelate rings range between

1.296-1.304 Å in the DFT models, and the metrical parameters of the central B3N3 and B3N4 heterocycles are reproduced accurately. Formation of the experimentally observed products is calculated to be very exergonic from the respective fragments, with ∆G = -118.4, -163.4 and -

155.7 kcal/mol for 11calc, 12calc and 13calc respectively (all energies relative to the fragments

Xs, E-Y and Z). Although a more detailed analysis of the reaction mechanisms is beyond the scope of the present contribution, it seems plausible that initial formation of a dimeric species is involved and thus the reaction X + Y was evaluated computationally. Geometry optimizations identified a minimum on the potential energy surface in which fragment X binds to the endocyclic N-atom of fragment Y. A second minimum was located that is 16.4 kcal/mol lower in energy, with fragment Y bound via both endo- and exocyclic N-atoms to result in a 4-membered B2N2 ring (XY, Scheme 6.3). Subsequent insertion of another fragment Y to give compounds 11 is energetically favorable (∆G = -36.4 kcal/mol) and, similarly, formation of products 12calc and 13calc is also downhill. A comparison of the total energy of the constitutional isomers 11calc -13calc shows that 11calc is thermodynamically the most stable: 12calc and 13calc are higher in energy by ∆G = 16.7 and 18.9 kcal/mol, respectively (Scheme 6.3).

142

Reduction of (Formazanate)Boron Difluoride Provides Evidence for an N‐Heterocyclic B(I) Carbenoid Intermediate

Scheme 6.3 DFT calculated energies of fragments X, Y and Z and the final products 11-13. Values are relative Gibbs free energies in kcal/mol. 6.4 Thermal Stability

The thermal stability of compound 11-14 was tested by NMR experiments. Heating NMR tubes to 130 °C (11, 12 and 13 in C6D6 or d8-toluene) or 80 °C (14 in d8-THF) resulted in no significant changes in the 1H NMR spectra of 11, 12 and 14. In the case of 13, the resonances of a new species (compound 15, Scheme 6.4) can be identified in the 1H NMR spectrum 1 (Figure 6.8). The H NMR spectrum of 15 shows two types of CH3 resonances with integration ratio of 1:2. In addition, the resonances of 15 in the aromatic region become simpler than the starting material of 13. The 1H NMR feature of 15 suggests that compound 15 has a higher symmetry than compound 13. In addition to resonances for 15, the 1H NMR spectrum of the thermolysis mixture shows signals that can be attributed to aniline and azobenzene (based on comparison to authentic samples). Based on the information collected from the NMR experiment, we assume that a NPh group was released from the fragment Z of 13 leading to the formation of 15 at high temperature. In the crystal structure of 13, fragment Z is a highly puckered BNNNCN 6-membered heterocycle in which N9 is clearly out of the heterocycle with long N(9)-N(10) (1.413(2) Å) and N(9)-B(3) (1.478(4) Å) bonds in comparison with other N-N and N-B(2) bonds in the structure. In addition, the distance between N(10) and B(3) is only 2.187(4) Å. The long N(9)-N(10) and N(9)-B(3) bond length and the short distance between N(10) and B(3) make the formation of compound 15 from compound 13 is possible at high temperature. The formation of aniline and azobenzene suggest hemolytic cleavage of N(9)-N(10) and N(9)-B(3) bands leading to the release of NPh group and formation of a new bond between N(10) and B(3). In a preparative scale reaction compound 15 can be synthesized from 13 in toluene solution with an isolated yield of 44 %.

143

Chapter 6

Scheme 6.4 Formation of 15 from 13 7.86 7.84 7.48 7.42 7.38 7.29 7.16 6.86 6.85 6.81 6.79 6.76 6.74 6.73 2.40 1.91 3.90 4.06 2.10 2.14 2.34 6.53 11.03 5.99 3.16 6.00

1 Figure 6.8 H NMR spectrum of 15 (CD2Cl2, 400 MHz)

Single crystals of 15 suitable for X-ray crystallography were obtained by recrystallization from toluene/hexane solvent mixture (Figure 6.9, metrical parameters in Table 6.2). The crystal structure determination shows that 15 is constructed from one (formazanate)boron unit (fragment X) and two imidoborane units (fragment Y), which share a NPh group. Specifically, it confirms that loss of a NPh unit from compound 13 occurs at high temperature, and matches the features observed in the 1H NMR spectrum. All the metrical parameters of the B3N3 core and imidoborane in compound 15 are very similar to those in compound 13. Cyclic voltammetry (CV) of 15 shows two quasi-reversible redox couples at -1.07 and -2.29 V vs. Fc0/+, which are very similar to the redox, potentials of 13. The singly reduced product of compound 15 ([15][Cp2Co]) can be synthesized and isolated as a crystalline material

(Figure 6.9, metrical parameters in Table 6.2) by using Cp2Co as reducing agent. The metrical parameters of 15- are very similar to 11- and 13-. Upon reduction, the N-N bond lengths of the formazanate ligand elongated significantly (15-: 1.350(7)/1.367(7); 15: 1.299(5)/1.312(5) Å), which are as expected for a formazanate-based reduction. In addition, the B-

144

Reduction of (Formazanate)Boron Difluoride Provides Evidence for an N‐Heterocyclic B(I) Carbenoid Intermediate

N(formazanate) bonds around the 4-coordinate B center are contracted from 1.607(5)/1.577(6) Å in 15 to 1.533(5) Å in 15-, with concomitant elongation of the B-N bonds to the central 6- membered heterocyclic ring (B(1)-N(5)/B(1)-N(7) in 15: 1.516(5)/ 1.509(5) Å and B(1)- N(5)/B(1)-N(7) in 15-: 1.552(8)/ 1.538(7) Å).

Figure 6.9 Crystal structures of 15 (left) and [15][Cp2Co] (right) showing 50% probability ellipsoids. All hydrogen atoms are omitted and the p-tolyl, Ph and Cp2Co are shown as wireframe for clarity. Table 6.2 Selected bond length (Å) and of compound 15 and [15][Cp2Co] 15 [15][Cp2Co] N1-N2 1.299(5) N1-N2 1.350(7) N3-N4 1.312(5) N3-N4 1.367(7) N1-B1 1.607(5) N1-B1 1.533(8) N4-B1 1.577(6) N4-B1 1.533(8) B1-N5 1.516(5) B1-N11 1.552(8) N5-B2 1.435(6) N11-B3 1.437(8) B2-N6 1,434(5) B3-N8 1.437(8) N6-B3 1.441(5) N8-B2 1.433(8) B3-N7 1.447(6) B2-N7 1.435(7) N7-B1 1.509(5) B1-N7 1.538(7) N5-C21 1.397(4) N11-C47 1.382(6) C21-N8 1.296(5) C47-N10 1.302(6) N8-N9 1.408(5) N10-N9 1.415(7) N9-B2 1.430(5) N9-B3 1.444(8) N7-C61 1.394(4) N7-C27 1.385(6) C61-N10 1.298(5) C27-N6 1.301(7) N10-N11 1.422(4) N6-N5 1.416(6) N11-B3 1.420(5) N5-B2 1.425(8)

145

Chapter 6

6.5 Trapping (Formazanate)Boron(I) Unit by Acetylene

Bis(trimethylsilyl)acetylene is one of the commonly used trapping agents for reactive B(I) species resulting in a formation of borirene (compound F, Scheme 6.5).7b In order to trap the (formazanate)boron fragment X, the 2-electron reduction of 6a was performed in the presence of 4 equivalents of bis(trimethylsilyl)acetylene at low temperature. Unfortunately, the isolated product is not the desired borirene product F but compound 16 (Scheme 6.5). Single crystals suitable for X-ray crystallography of 16 were obtained from toluene/hexane solvent mixture. Even though the quality of the crystal structure is not very good, the connection of atoms is still reliable (Figure 6.10). The crystal structure of 16 has two (formazanate)boron fluoride fragments ([1a]BF) and two amidoborane fragments, which share one NPh group. The boron center of [1a]BF fragment is coordinated by a endocyclic N-atom of amidoborane fragment forming a four-coordinated boron center. The mechanism of the formation of compound 16 is still not clear but the presence of B-F unit in the structure of 16 suggests that (formazanate)boron monofluoride radial ([1a]BF) was formed after the first reduction of

LBF2 by Na/Hg. The (formazanate)boron monofluoride radial can be further reduced by sodium amalgam to generate (formazanate)boron fragment X. The metrical parameters of 16 will not be discussed here due to the low quality of the data set.

Scheme 6.5 Formation of 16 from 6a (left) and the expected borirene product F (right).

146

Reduction of (Formazanate)Boron Difluoride Provides Evidence for an N‐Heterocyclic B(I) Carbenoid Intermediate

Figure 6.10 Crystal structures of 16 showing 50% probability ellipsoids. All hydrogen atoms are omitted and the p-tolyl and Ph groups are shown as wireframe for clarity) The NMR spectra of 16 (Figure 6.11) are more informative than the NMR spectra of 11-14 1 due to the C2 axis in 16. The H NMR spectrum of 16 shows 2 separate singlets at  2.43 and 2.23 ppm, consistent with 2 different p-tolyl environments, which were observed in the solid state. The resonances of one of the -CH of the p-tolyl group can be located at  7.49 ppm. The resonances at  7.90 and 7.38 ppm are assigned to the o-CH and m-CH of the phenyl group of the formazanate ligand. More importantly, the resonances of the bridging NPh group between two amidoborane fragments are located at  5.92, 5.75 and 5.55 ppm, which are outside an expected range of aromatic rings and are broader than other resonances. The unusual chemical shifts and broad shape of resonances of the bridging NPh group might due to some radical property, which could be resulted from radical impurities or a diradical property of compound 16. The 19F NMR spectrum shows one broad resonance at  -155.1 ppm which correspond to one type of fluorine environment in the structure. Furthermore, the 11B NMR spectrum for 16 shows two broad resonances: a resonance at  29.0 ppm (very broad) and a somewhat sharper one around  4.4 ppm that are attributed to the 3- and 4-coordinate B centers, respectively.

147

Chapter 6

Figure 6.11 1H NMR (top), 19F NMR (bottom left) and 11B NMR (bottom right) spectra of 16

(in d8-THF, RT)

6.6 Chemical Oxidation

In attempts to elicit reactivity that stems from the trapped (formazanate)B carbenoid fragment, we treated compounds 11-13 with the oxidizing agent XeF2. The reaction between 11 and

XeF2 was monitored in an NMR tube (C6D6 solution). The addition of 1 or 2 equivalent of

XeF2 relative to 11 (XeF2:B ratio < 1) resulted in rapid disappearance of the starting material, but an intractable mixture was obtained.

19 Surprisingly, reaction with 3 equivalent of XeF2 resulted in a signal in the F NMR spectrum that is diagnostic for the boron difluoride starting material (PhNNC(p- tolyl)NNPh)BF2 (6a). Addition of a larger excess of XeF2 to 11 (12 equivalent, XeF2:B ratio = 4) converted 90% of the (formazanate)B fragment X present in 11 to the difluoride 6a after 30 min, with quantitative conversion after standing overnight (based on 19F NMR integration relative to an internal standard). As may be anticipated, only the moiety with the intact

148

Reduction of (Formazanate)Boron Difluoride Provides Evidence for an N‐Heterocyclic B(I) Carbenoid Intermediate

NNCNN formazanate backbone (indicated in bold in Scheme 6.6) is able to regenerate 6a. The fate of the remaining B-containing fragments is unclear at present. While we were unable to observe similar reactivity between 13 and XeF2 (3 or 12 equivalent), also 12 reacted with excess XeF2 to give the boron difluoride 6a (Scheme 6.6). Compound 12 contains two intact (formazanate)B fragments (X, as 5- and 6-membered chelates) which are both converted to 6a, leading to a total of 1.48 equivalents (74% yield based on available (formazanate)B) after 16h.

It is at present unclear why compounds 11 and 12 give good yields of 6a upon oxidation with XeF2, but 13 does not. Also, the mechanism for the reaction with XeF2 is not known and needs further investigation. Nevertheless, these results demonstrate that the 'trapped' (formazanate)B species is able to show carbenoid reactivity at the boron center.28b

Scheme 6.6 Chemical oxidation of compounds 11 and 12 to regenerate the boron difluoride 6a (yields based on the amount of (formazanate)B available).

149

Chapter 6

6.7 Conclusion

Although the high reactivity of triplet boron carbenoids has been suggested as the reason for their conspicuous absence from the literature, our results indicate that judicious ligand design imparts sufficient stability to allow isolation of novel BN-heterocycles derived from the carbenoid fragment (PhNNC(p-tolyl)NNPh)B (X). Fragment X is calculated to have a triplet biradical ground state that is stabilized due to the low-lying formazanate π*-orbitals. Reductive cleavage of a N-N bond in X generates the iminoborane species Y, and these two intermediates combine under the reaction conditions to form the final trimeric products (11- 14). The thermolysis of 13 results in the formation of a new BN-heterocyclic product 15 and releasing a NPh fragment. The (formazanate)B fragment that is incorporated in the BN- heterocyclic products 11-13 and 15 remains available as a redox-active moiety as shown by cyclic voltammetry and X-ray structural characterization of the radical anions [11]–, [13]– and [15]–. Significantly, the (formazanate)B moiety retains carbenoid character and is able to perform a net 2-electron oxidative addition reaction with XeF2 that regenerates the boron difluoride starting material 6a. The results presented here highlight that formazanate ligands have considerable potential in stabilizing highly reactive B compounds, and demonstrate a novel design strategy towards the synthesis of isolable N-heterocyclic boron carbenoids.

6.8 Experimental Section

General Considerations. All manipulations were carried out under nitrogen atmosphere using standard glovebox, Schlenk, and vacuum-line techniques. Toluene, hexane, and pentane

(Aldrich, anhydrous, 99.8%) were passed over columns of Al2O3 (Fluka), BASF R3-11- supported Cu oxygen scavenger, and molecular sieves (Aldrich, 4 Å). Diethyl ether and THF

(Aldrich, anhydrous, 99.8%) were dried by percolation over columns of Al2O3 (Fluka).

Deuterated solvents were vacuum transferred from Na/K alloy (C6D6, THF-d8, Aldrich) or

CaH2 (CD2Cl2) and stored under nitrogen. XeF2 was purchased from Alfa Aesar and used as received. NMR spectra were recorded on Varian Gemini 200, VXR 300, Mercury 400, Inova 500 or Agilent 400 MR spectrometers. The 1H and 13C NMR spectra were referenced internally using the residual solvent resonances and reported in ppm relative to TMS (0 ppm); J is reported in Hz. Assignment of NMR resonances was aided by gradient-selected gCOSY, NOESY, gHSQCAD and/or gHMBCAD experiments using standard pulse sequences. 11B NMR spectra were recorded in quartz NMR tubes using a OneNMR probe on an Agilent 400

150

Reduction of (Formazanate)Boron Difluoride Provides Evidence for an N‐Heterocyclic B(I) Carbenoid Intermediate

MR system. UV-Vis spectra were recorded in THF solution (~ 10-5 M) using a Avantes

AvaSpec 3648 spectrometer and AvaLight-DHS lightsource inside a N2 atmosphere glovebox. The photoluminescence quantum yield of 15 was determined using an optically dilute solution in THF (λex = 400 nm) with optically dilute fluorescein (0.5M NaOH) as reference. Spectra were recorded using a 75W Xenon lamp coupled to a Zolix 150 monochromator coupled directly to a Qpod cuvette holder (Quantum Northwest) and emission was collected through a fibre optic connected Shamrock 163 spectrograph and iDUS-420A-OE CCD detector. Spectra are uncorrected for instrument response. Elemental analyses were performed at the Microanalytical Departement of the University of Groningen or Kolbe Microanalytical Laboratory (Mülheim an der Ruhr, Germany).

Computational studies. Calculations were performed with the Gaussian09 program using density functional theory (DFT). In order to increase computational efficiency, the p-tolyl group was replaced by Ph. Geometries were fully optimised starting from the X-ray structures using the B3LYP exchange-correlation functional with the 6-31G(d) basis set. Geometry optimisations were performed without (symmetry) constraints, and the resulting structures were confirmed to be minima on the potential energy surface by frequency calculations (number of imaginary frequencies = 0). Transition state calculations were performed with the QST3 algorithm in Gaussian09. The calculated transition states were confirmed by frequency calculations, which show one imaginary frequency, and IRC calculations in both directions.

Synthesis and characterization

The following procedure is representative for the synthesis and sequential isolation of pure samples of 11-14:

A flask was charged with [PhNNC(p-tolyl)NNPh]BF2 (6a, 300 mg, 0.828 mmol), Na/Hg (2.447 %, 1.558 g, 1.657 mmol of Na), THF (5 mL) and toluene (10 mL). The reaction mixture was stirred at room temperature for 3 days and the color changed from red to purple. After removal of all the volatiles under vacuum, the crude reaction mixture was analyzed by 1 H NMR spectroscopy in C6D6 and shown to contain compounds 11, 12 and 13 in a 1:0.6:0.5 ratio based on the integration of the p-tolyl CH3 resonances and comparison to isolated, pure materials (see below).

For further workup, the crude product was dissolved in dimethoxyethane (4 mL).

Isolation of 14. Slow diffusion of hexane into the dimethoxyethane solution -30 °C precipitated compound 14. Analytically pure material was obtained by dissolving the

151

Chapter 6

precipitate again in dimethoxyethane, followed by diffusion of hexane into the solution to yield 14.8 mg of 14 as light yellow crysalline material (0.014 mmol, 4 %).

Isolation of 11. Upon separation of 14, the supernatant was pumped to dryness and the solid residue was dissolved in a toluene/THF mixture (ratio 1:2, total 4 mL). Hexane was allowed to diffuse into the solution at -30 °C, which resulted in the precipitation of 61.8 mg of 11 as dark purple crystalline material (0.064 mmol, 23 %).

Isolation of 13. All volatiles were removed from the supernatant that was left upon isolation of 12. Dissolving the residue in toluene (4 mL) and allowing diffusion of hexane into the toluene layer at -30 °C afforded 33.0 mg of 13 as deep blue crystals (0.031 mmol, 11 %).

Isolation of 12. Further concentration of the supernatant and cooling to -30 °C precipitated 16.5 mg of 12 as deep red crystalline material (0.017 mmol, 6 %).

1 Characterization data for compound 11. H NMR (CD2Cl2, 400 MHz, 25 °C): 8.03 (d, 2H, J = 8.6 Hz, p-tolyl CH), 7.76 (dm, 2H, J = 8.4 Hz, p-tolyl CH), 7.58-7.43 (m, 7H), 7.46 (t, 1H, J = 7.2 Hz, Ph p-CH), 7.39 (d, 2H, J = 8.0 Hz), 7.35 (d, 2H, J = 8.0 Hz), 7.22 (t, 1H, J = 7.4 Hz, Ph p-CH), 7.08-7.03 (m, 9H), 6.93 (t, 2H, J = 7.6 Hz, Ph m-CH), 6.89-6.77 (m, 11H), 6.63 (t, 1H, J = 7.2 Hz, Ph p-CH), 6.41 (d, 2H, J = 8.4 Hz, Ph o-CH), 2.32 (s, 3H, p-tolyl 13 CH3), 2.28 (s, 3H, p-tolyl CH3), 2.18 (s, 3H, p-tolyl CH3). C-NMR (CD2Cl2, 100 MHz, 25 °C): 158.1 (NCN), 152.8 (NCN), 150.8 (NCN), 147.5, 147.0, 146.7, 146.1, 144.8, 143.5, 141.7 (p-tolyl p-C), 138.4 (p-tolyl p-C), 138.3 (p-tolyl p-C), 132.7, 132.6, 132.1, 129.7, 129.6, 129.5, 129.4, 129.2, 129.1, 129.0, 128.9, 128.7, 128.5, 128.4, 128.3, 127.8, 127.8, 124.6 (p- tolyl CH), 124.2, 124.2 (Ph p-CH), 123.8 (p-tolyl CH), 123.5, 123.4, 21.6 (p-tolyl CH3), 21.3 11 (p-tolyl CH3), 21.2 (p-tolyl CH3). B-NMR (C6D6, 128 MHz, 25 °C): 26.2 (FWHH = 1050

Hz), 0.5 (FWHH = 220 Hz). Anal. calcd for C60H51B3N12(C6H14)0.5: C, 74.50; H, 5.76; N, 16.55. Found: C, 74.42; H, 5.49; N, 16.94.

1 Characterization data for compound 12. H NMR (CD2Cl2, 400 MHz, 25 °C): 7.83 (d, 2H, J= 8.4 Hz, p-tolyl o-CH), 7.67 (d, 2H, J= 8.0 Hz, p-tolyl o-CH), 7.62-7.59 (m, 2H), 7.33 (d, 2H, J= 8.0 Hz, p-tolyl m-CH), 7.22 (d, 2H, J= 8.0 Hz, Ph o-CH), 7.14-7.10 (m, 9H), 7.05 (d, 2H, J= 8.4 Hz, p-tolyl o-CH), 7.00-6.87 (m, 3H), 6.97 (d, 2H, J= 8.0 Hz, p-tolyl m-CH), 6.93 (d, 2H, J= 7.6 Hz, Ph o-CH), 6.83 (d, 2H, J= 7.2 Hz, Ph m-CH), 6.68 (d, 2H, J= 7.6 Hz, p- tolyl m-CH), 6.58-6.50 (m, 3H, Ph p-C, Ph o-CH), 6.37-6.31 (m, 4H, Ph m-CH), 6.19 (t, 1H,

J= 7.2 Hz, Ph p-CH), 6.06 (d, 2H, J= 8.0 Hz, Ph o-CH), 2.48 (s, 3H, p-tolyl CH3), 2.25 (s, 3H, 13 p-tolyl CH3), 2.11 (s, 3H, p-tolyl CH3). C NMR (CD2Cl2, 100 MHz, 25 °C): 151.2 (NCN),

152

Reduction of (Formazanate)Boron Difluoride Provides Evidence for an N‐Heterocyclic B(I) Carbenoid Intermediate

149.3 (NCN), 148.9 (NCN), 148.1(Ph i-C), 147.7 (Ph i-C), 146.8 (Ph i-C), 146.3 (Ph i-C), 143.2 (Ph i-C), 142.8 (Ph i-C), 139.5 (p-tolyl p-C), 139.4 (p-tolyl p-C), 139.2 (p-tolyl p-C), 131.7, 129.8 (p-tolyl m-C), 129.3 (p-tolyl m-C), 129.3, 129.2, 129.1, 128.9, 128.8 (p-tolyl m- C), 128.4 (p-tolyl o-C), 128.4 (Ph m-C), 127.9 (Ph m-C), 127.3 (Ph o-C), 126.7 (p-tolyl o-C), 125.8 (p-tolyl i-C), 125.6, 125.6 (p-tolyl o-C), 125.0 (Ph o-C), 124.9 (Ph o-C), 123.5, 123.4,

122.4 (Ph p-C), 122.1 (Ph o-C), 120.6 (Ph p-C), 120.2 (Ph m-C), 21.5 (p-tolyl CH3), 21.4 (p- 11 tolyl CH3), 21.3 (p-tolyl CH3). B NMR (THF-d8, 128 MHz, -60 °C): 28.2 (FWHH = 1420 Hz), 5.4 (FWHH = 150 Hz).26.3 (FWHH = 970 Hz), 2.3 (FWHH = 220 Hz). Anal. calcd for

C60H51B3N12: C, 74.10; H, 5.29; N, 17.28. Found: C, 74.02; H, 5.36; N, 16.75.

1 Characterization data for compound 13. H NMR (CD2Cl2, 400 MHz, 25 °C): 7.91 (d, 4H, J = 7.6 Hz), 7.47-7.39 (m, 6H), 7.24-7.19 (m, 6H, Ph o-CH), 7.15 (d, 4H, J = 8.0 Hz, Ph m- CH), 7.08 (d, 1H, J = 6.4 Hz), 7.01 (d, 2H, J = 8.4 Hz, Ph o-CH), 6.92-6.88 (m, 6H, Ph o- CH), 6.77-6.71 (m, 6H, Ph m-CH), 6.66 (t, 2H, J = 7.2 Hz), 6.65-6.60 (m, 3H), 6.57 (d, 2H, J

= 8.4 Hz, Ph m-CH), 2.39 (s, 3H, p-tolyl CH3), 1.80 (s, 3H, p-tolyl CH3), 1.77 (s, 3H, p-tolyl 13 CH3). C NMR (CD2Cl2, 100 MHz, 25 °C): 155.1, 151.5, 147.6, 146.3, 145.5, 144.1, 142.7, 142.2, 139.4, 139.2, 138.1, 131.8, 131.7, 129.6, 129.6, 129.4, 129.2, 129.0, 129.0, 128.9, 128.8, 128.7, 128.7, 128.6, 128.4, 128.3, 128.2, 128.0, 125.6, 125.1, 124.4, 124.2, 123.3,

120.9, 120.4, 116.1, 114.0 (Ph o-C), 21.3 (p-tolyl CH3), 20.8 (p-tolyl CH3), 20.8 (p-tolyl 11 CH3). B NMR (C6D6, 128 MHz, 25 °C): 25.1 (FWHH = 720 Hz), -0.3 (FWHH = 122 Hz).

Anal. calcd for C60H51B3N12(C7H8): C, 75.58; H, 5.59; N, 15.79. Found: C, 75.18; H, 5.59; N, 15.89.

1 Characterization data for compound 14. H NMR (THF-d8, 400 MHz, -60 °C):  8.15 (d,

2H, J = 7.6 Hz, p-tolyl o-CH), 7.91 (d, 1H, J = 7.2 Hz, C6H4), 7.79 (d, 2H, J = 7.6 Hz, p-tolyl o-CH), 7.50 (br, 2H, Ph), 7.41 (d, 2H, J = 7.6 Hz, p-tolyl o-CH), 7.20 (d, 2H, J = 7.6 Hz, Ph o-CH), 7.10 (d, 2H, J = 7.6 Hz, p-tolyl m-CH), 7.06 (br, 2H, Ph), 6.97 (t, 2H, J = 7.2 Hz, Ph m-CH), 6.94 (br, 1H, Ph), 6.89 (d, 1H, J = 7.2 Hz, Ph), 6.83 (d, 2H, J = 8.0 Hz, p-tolyl m-

CH), 6.78 (d, 2H, J = 7.6 Hz, p-tolyl m-CH), 6.76-6.67 (overlapped, 3H, 2  C6H4 + 1  Ph),

6.63-6.51 (overlapped, 3H, Ph), 6.51-6.39 (overlapped, 5H, 4  Ph + 1  C6H4), 6.38-6.28 (overlapped, 2H, Ph), 6.23-6.07 (overlapped, 3H, Ph), 6.02 (t, 1H, J = 7.5 Hz, Ph), 5.67 (t,

1H, J = 7.5 Hz, Ph), 3.40 (s, 12H, O-CH2 DME), 3.24 (s, 18H, O-CH3 DME), 2.32 (s, 3H, p- 13 tolyl), 2.20 (s, 3H, p-tolyl), 2.08 (s, 3H, p-tolyl). C NMR (THF-d8, 100 MHz, -60 °C): 149.6, 148.7, 148.0, 147.9, 147.7, 143.6, 143.4, 142.9, 136.9, 136.5, 135.4, 130.2, 130.2, 128.2, 128.1, 128.0, 127.7, 127.6, 127.3, 127.2, 126.8, 126.7, 126.4, 126.2, 126.1, 126.0,

153

Chapter 6

125.7, 123.4, 123.2, 122.8, 122.1, 121.3, 120.5, 119.7, 119.2, 118.4, 117.5, 115.1, 112.8,

112.5, 112.0, 111.8, 109.5, 71.7 (O-CH2 DME), 58.0 (O-CH3 DME), 20.7 (p-tolyl CH3), 20.6 11 (p-tolyl CH3), 20.3 (p-tolyl CH3). B NMR (THF-d8, 128 MHz, -60 °C): 28.2 (FWHH = 1420

Hz), 5.4 (FWHH = 150 Hz). Anal. calcd for C75H87B3N12NaO6: C, 68.87; H, 6.70; N, 12.85. Found: C, 68.13; H, 6.66; N, 12.67.

Synthesis of 15. A flask was charged with 13 (29.9 mg, 0.028 mmol) and toluene (10 mL). The reaction mixture was stirred at 130 Ԩ for overnight and the color remain as blue. The product 15 was purified by recrystallization from toluene/hexane mixture (10.9 mg, 0.012 1 mmol, 44 %). H NMR (CD2Cl2, 400 MHz, 25 °C): 7.85 (d, 4H, J = 8.0 Hz, Ph o-CH), 7.48 (t, 4H, J = 8.0 Hz, Ph m-CH), 7.40 (t, 2H, J = 7.2 Hz, Ph p-CH), 7.30 (d, 2H, J = 8.1 Hz, p-tolyl

CH), 7.17 (d, 2H, J = 8.1 Hz, p-tolyl CH), 6.91-6.70 (m, 23 H), 2.40 (s, 3H, CH3), 1.91 (s, 6H, 13 CH3). C NMR (CD2Cl2, 100 MHz, 25 °C): 151.5, 146.5 (Ph i-C), 144.6 (NCN), 143.3, 142.4, 139.2, 138.2 (p-tolyl p-C), 132.0 (p-tolyl i-C), 129.6 (Ph p-C), 129.5 (Ph m-C), 129.4, 129.1 (p-tolyl CH), 129.0, 128.9, 128.6, 127.8, 127.5, 125.2 (p-tolyl CH), 124.5, 124.3, 124.0, 123.4 11 (Ph o-C), 21.4 (CH3), 20.9 (CH3). B NMR (CD2Cl2, 128 MHz, 25 °C): 24.5 (FWHH = 400 Hz), -1.6 (FWHH = 100 Hz).

Synthesis of 16. A flask was charged with 6a (29.8 mg, 0.082 mmol), Na/Hg (2.447 %, 151.4 mg, 0.161 mmol) bis(trimethylsilyl)acetylene (57.7 mg, 0.339 mmol), toluene (1.5 mL) and THF (4 mL). The reaction mixture was stirred at -63Ԩ and slowly warmed back to room temperature. After stirred at room temperature for overnight, the solvent was removed and the 1 product 16 was isolated by recrystallization from toluene/hexane mixture. H NMR (CD2Cl2, 500 MHz, 25 °C): 7.94 (d, 8H, J = 6.8 Hz, p-tolyl CH and p-tolyl CH), 7.53 (d, 4H, J = 7.6 Hz, Ph o-CH), 7.44-7.39 (m, 8H, Ph m-CH and p-tolyl CH), 7.12 (t, 2H, J = 7.0 Hz, Ph p-CH), 7.08 (d, 4H, J = 7.9 Hz, Ph o-CH), 7.05 (d, 4H, J = 7.7 Hz, p-tolyl CH), 6.75 (t, 2H, J = 6.8 Hz, Ph p-CH), 6.68 (t, 4H, J = 7.4 Hz, Ph m-CH), 6.63 (d, 4H, J = 6.4 Hz, Ph o-CH), 6.50 (t, 4H, J = 7.6 Hz, Ph m-CH), 6.39 (t, 2H, J = 7.0 Hz, Ph p-CH), 5.96 (bs, 1H, Ph p-CH), 5.77

(bs, 2H, Ph m-CH), 5.57 (bs, 2H, Ph o-CH), 2.46 (s, 6H, p-tolyl CH3), 2.27 (s, 6H, p-tolyl 13 CH3). C NMR (CD2Cl2, 125 MHz, 25 °C): 152.5, 145.0, 142.8, 140.1, 137.3, 132.2, 129.6, 129.0, 128.7, 128.2, 128.2, 127.9, 127.0, 126.9, 126.4, 125.6, 124.8, 124.3, 124.2, 121.9, 11 19 121.3, 120.0, 119.8, 20.5, 20.4. B NMR (CD2Cl2, 128 MHz, 25 °C): 28.5 (bs), 4.5 (bs) F

NMR (CD2Cl2, 375 MHz, 25 °C): -155.1 (bs)

154

Reduction of (Formazanate)Boron Difluoride Provides Evidence for an N‐Heterocyclic B(I) Carbenoid Intermediate

Synthesis of [13][Na(15-c-5)]. To a solution of 13 (21.5mg, 0.020 mmol) in 1.5 mL of THF was added Na/Hg (2.447 % of Na, 22.8 mg, 0.024 mmol) and 15-crown-5 (4 L, 0.020 mmol). The mixture was stirred for 5 h, after which the supernatant solution was separated from Hg(l) using a pipette. Addition of hexane to the THF solution precipitated 24.3 mg of [13][Na(15-c-

5)] as green crystalline material (0.019 mmol, 99%). Anal. calcd for C70H71B3N12NaO5: C, 69.15; H, 5.89; N, 13.82. Found: C, 69.09; H, 5.90; N, 13.59.

Synthesis of [15][Cp2Co]. Compound 15 (3.8 mg, 0.0043 mmol) and Cp2Co (1.1 mg, 0.0058 mmol) were dissolved in THF (1.5 mL). The solution was stirred at RT for overnight, after which the product [15][Cp2Co] was isolated by recrystallization from THF/hexane mixture as green crystalline material (4.2 mg, 0.0039 mmol, 91 %).

Crystallographic data

Suitable crystals of 11-15, [11][Na(THF)3], [13][Na(15-c-5)] and [15][Cp2Co] were mounted on a cryo-loop in a drybox and transferred, using inert-atmosphere handling techniques, into the cold nitrogen stream of a Bruker D8 Venture diffractometer. The final unit cell was obtained from the xyz centroids of 9250 (11), 6880 (12), 9937 (13), 9958 (14), 9918 (15),

9750 ([11][Na(THF)3]) 9319 ([13][Na(15-c-5)]), and 9890 ([15][Cp2Co]) reflections after integration. Intensity data were corrected for Lorentz and polarisation effects, scale variation, for decay and absorption: a multiscan absorption correction was applied, based on the intensities of symmetry-related reflections measured at different angular settings (SADABS).29 The structures were solved by direct methods using the program SHELXS.30 The hydrogen atoms were generated by geometrical considerations and constrained to idealised geometries and allowed to ride on their carrier atoms with an isotropic displacement parameter related to the equivalent displacement parameter of their carrier atoms. Structure refinement was performed with the program package SHELXL.30 Crystal data and details on data collection and refinement are presented in following tables.

Crystallographic data 11 12 13 14 chem formula C60H51B3N12 C60H51B3N12 C60H51B3N12 C75H87B3N12NaO6 Mr 972.55 972.55 972.55 1307.98 cryst syst monoclinic orthorhombic monoclinic monoclinic color, habit red, platelet purple, platelet purple, platelet yellow, block size (mm) 0.240.130.02 0.120.070.01 0.280.060.01 0.200.070.04 space group C2/c P212121 P21/n P21/n a (Å) 32.1139(13) 11.0453(10) 10.5968(4) 13.1785(6) b (Å) 14.4081(7) 19.3899(17) 17.7087(7) 21.1594(9) c (Å) 25.2549(10) 22.940(2) 30.8111(10) 26.1619(10)  (°)

155

Chapter 6

β (°) 109.612(2) 99.332(2) 94.1169(13)  (°) V (Å3) 11007.5(8) 4913.1(8) 5705.4(4) 7276.4(5) Z 8 4 4 4 -3 calc, g.cm 1.174 1.315 1.132 1.194 -1 µ(Cu K), mm 0.554 0.621 0.535 -1 µ(Mo K), mm 0.082 F(000) 4080 2040 2040 2780 temp (K) 100(2) 100(2) 100(2) 100(2)  range (°) 3.397-56.447 2.984-54.597 2.888-65.282 2.899- 24.756 data collected (h,k,l) -34:34, -14:16, - -11:10, -15:20, - -12:12, -20:18, - -15:15, -24:24, - 28:27 23:23 36:36 30:30 min, max transm 0.917, 0.989 0.949,0.994 0.962,0.995 0.993,0.997 rflns collected 50659 24631 55095 94891 indpndt reflns 7272 5883 9693 12418 observed reflns Fo  4069 4522 6687 7523 2.0 σ (Fo) R(F) (%) 9.89 7.34 5.12 6.12 wR(F2) (%) 26.49 12.51 12.05 13.98 GooF 1.036 1.120 1.007 1.023 weighting a,b 0.1031, 72.3279 0, 11.9205 0.0512, 2.6001 0.0442, 6.4770 params refined 679 680 679 915 min, max resid dens -0.286, 0.457 -0.340, 0.321 -0.223, 0.401 -0.373, 0.400

15 [11][Na(THF)3] [13][Na(15-c-5)] [15][Cp2Co] chem formula C61H53B3N11 C76H83B3N12NaO4 C70H71B3N12NaO5 C64H56B3CoN11 Mr 972.57 1283.96 1215.80 1070.55 cryst syst triclinic monoclinic triclinic orthorhombic color, habit blue, needle green, platelet green, platelet green, platelet size (mm) 0.33 x 0.10 x 0.08 0.18 x 0.14 x 0.02 0.25 x 0.20 x 0.04 0.13 x 0.09 x 0.02 space group P-1 P21/n P-1 P21212 a (Å) 10.7880(6) 11.1938(4) 11.8029(5) 21.5649(6) b (Å) 11.9037(8) 20.4602(7) 22.3034(8) 22.7448(6) c (Å) 21.8310(14) 31.0805(11) 25.7940(10) 11.9160(3) α (°) 98.5634(19) 97.5364(17) β (°) 90.3164(19) 100.366(2) 100.8958(18) γ (°) 111.8180(17) 94.8132(17) V (Å3) 2568(3) 7002.1(4) 6568.8(4) 5844.7(3) Z 2 4 4 4 -3 calc, g.cm 1.167 1.218 1.229 1.217 -1 µ(Cu Kα), mm 0.654 2.683 -1 µ(Mo Kα), mm 0.07 0.084 F(000) 954 2724 2564 2236 temp (K) 100(2) 100(2) 100(2) 100(2)  range (°) 3.00-26.37 3.61 - 59.21 2.78 - 26.02 3.71-59.58 data collected (h,k,l) -13:12; -8:14; - -12:12; -22:22; - -14:14; -27:26; - -24:18; -25:25; - 22:27 34:34 31:31 12:13 min, max transm 0.6798, 0.7454 0.891, 0.987 0.6433, 0.7455 0.5752, 0.7146 rflns collected 13984 64589 76760 26511 indpndt reflns 9074 9451 22580 8131 observed reflns Fo  5754 7747 11530 6262 2.0 σ (Fo) R(F) (%) 7.85 7.57 8.58 5.33 wR(F2) (%) 17.79 19.42 24.60 11.53 GooF 1.015 1.131 1.012 1.021 weighting a,b 0.0826, 4.0966 0.0604, 19.7970 0.1038, 7.8864 0.0576 params refined 679 878 1690 715

156

Reduction of (Formazanate)Boron Difluoride Provides Evidence for an N‐Heterocyclic B(I) Carbenoid Intermediate min, max resid dens -0.605, 1.142 -0.403, 0.557 -0.353, 0.682 -0.202, 0.367

6.9 References

(1) (a) Bourissou, D.; Guerret, O.; Gabbaï, F. P.; Bertrand, G. Chem. Rev. 2000, 100, 39–92. (b) Hahn, F. E.; Jahnke, M. C. Angew. Chem. Int. Ed. 2008, 47, 3122–3172. (c) Hopkinson, M. N.; Richter, C.; Schedler, M.; Glorius, F. Nature 2014, 510, 485–496. (2) (a) Mizuhata, Y.; Sasamori, T.; Tokitoh, N. Chem. Rev. 2009, 109, 3479–3511. (b) Blom, B.; Stoelzel, M.; Driess, M. Chem. Eur. J. 2013, 19, 40–62. (c) Blom, B.; Gallego, D.; Driess, M. Inorg. Chem. Front. 2014, 1, 134–148. (3) Fischer, R. C.; Power, P. P. Chem. Rev. 2010, 110, 3877-3923. (4) (a) Vidovic, D.; Pierce, G. A.; Aldridge, S. Chem. Commun. 2009, 1157-1171 (b) Vidovic, D.; Aldridge, S. Chem. Sci. 2011, 2, 601–608. (c) Brand, J.; Braunschweig, H.; Sen, S. S. Acc. Chem. Res. 2013, 47, 180–191. (d) Braunschweig, H.; Dewhurst, R. D.; Gessner, V. H. Chem. Soc. Rev. 2013, 42, 3197–3208. (5) Timms, P. L. Acc. Chem. Res. 1973, 6, 118–123. (6) van der Kerk, S. M.; Boersma, J.; van der Kerk, G. J. M. Tetrahedron Lett. 1976, 17, 4765–4766. (7) (a) Ramsey, B. G.; Anjo, D. M. J. Am. Chem. Soc. 1977, 99, 3182–3183. (b) Pachaly, B.; West, R. Angew. Chem. Int. Ed. 1984, 23, 454-455. (8) Bettinger, H. F. J. Am. Chem. Soc. 2006, 128, 2534–2535. (9) Krasowska, M.; Bettinger, H. F. J. Am. Chem. Soc. 2012, 134, 17094–17103. (10) (a) Wang, Y.; Quillian, B.; Wei, P.; Wannere, C. S.; Xie, Y.; King, R. B.; Schaefer, H. F.; Schleyer, P. v. R.; Robinson, G. H. J. Am. Chem. Soc. 2007, 129, 12412–12413. (b) Wang, Y.; Quillian, B.; Wei, P.; Xie, Y.; Wannere, C. S.; King, R. B.; Schaefer, H. F.; Schleyer, P. v. R.; Robinson, G. H. J. Am. Chem. Soc. 2008, 130, 3298–3299. (c) Bissinger, P.; Braunschweig, H.; Damme, A.; Kupfer, T.; Vargas, A. Angew. Chem. Int. Ed. 2012, 51, 9931–9934. (11) (a) Braunschweig, H.; Dewhurst, R. D.; Hammond, K.; Mies, J.; Radacki, K.; Vargas, A. Science 2012, 336, 1420–1422. (b) Koppe, R.; Schnockel, H. Chem. Sci. 2015, 6, 1199–1205. (c) Böhnke, J.; Braunschweig, H.; Constantinidis, P.; Dellermann, T.; Ewing, W. C.; Fischer, I.; Hammond, K.; Hupp, F.; Mies, J.; Schmitt, H.-C.; Vargas, A. J. Am. Chem. Soc. 2015, 137, 1766-1769. (12) (a) Bissinger, P.; Braunschweig, H.; Damme, A.; Dewhurst, R. D.; Kupfer, T.; Radacki, K.; Wagner, K. J. Am. Chem. Soc. 2011, 133, 19044–19047. (b) Bissinger, P.; Braunschweig, H.; Kraft, K.; Kupfer, T. Angew. Chem. Int. Ed. 2011, 50, 4704–4707. (c) Curran, D. P.; Boussonnière, A.; Geib, S. J.; Lacôte, E. Angew. Chem. Int. Ed. 2012, 51, 1602–1605. (13) (a) Kinjo, R.; Donnadieu, B.; Celik, M. A.; Frenking, G.; Bertrand, G. Science 2011, 333, 610–613. (b) Kong, L.; Li, Y.; Ganguly, R.; Vidovic, D.; Kinjo, R. Angew. Chem. Int. Ed. 2014, 53, 9280–9283. (c) Dahcheh, F.; Martin, D.; Stephan, D. W.; Bertrand, G. Angew. Chem. Int. Ed. 2014, 53, 13159–13163. (14) Segawa, Y.; Yamashita, M.; Nozaki, K. Science 2006, 314, 113–115. (15) Asay, M.; Jones, C.; Driess, M. Chem. Rev. 2011, 111, 354–396. (16) (a) Reiher, M.; Sundermann, A. Eur. J. Inorg. Chem. 2002, 2002, 1854-1863. (b) Chen, C.-H.; Tsai, M.-L.; Su, M.-D. Organometallics 2006, 25, 2766-2773. (17) Firinci, E.; Bates, J. I.; Riddlestone, I. M.; Phillips, N.; Aldridge, S. Chem. Commun. 2013, 49, 1509– 1511. (18) CRC Handbook of Chemistry and Physics; 96 ed.; CRC Press: Boca Raton, FL, 2015-2016. (19) Rablen, P. R.; Hartwig, J. F. J. Am. Chem. Soc. 1996, 118, 4648 (20) Connelly, N. G.; Geiger, W. E. Chem. Rev. 1996, 96, 877–910. (21) (a) Bai, G.; Wei, P.; Stephan, D. W. Organometallics 2006, 25, 2649–2655. (b) Basuli, F.; Kilgore, U. J.; Brown, D.; Huffman, J. C.; Mindiola, D. J. Organometallics 2004, 23, 6166–6175. (c) Obenhuber, A. H.; Gianetti, T. L.; Berrebi, X.; Bergman, R. G.; Arnold, J. J. Am. Chem. Soc. 2014, 136, 2994- 2997. (22) (a) Cramer, C. J.; Tolman, W. B. Acc. Chem. Res. 2007, 40, 601–608. (b) Yamashita, M.; Aramaki, Y.; Nozaki, K. New J. Chem. 2010, 34, 1774–1782. (c) Li, J.; Li, X.; Huang, W.; Hu, H.; Zhang, J.; Cui, C. Chem. Eur. J. 2012, 18, 15263–15266. (d) Xie, L.; Zhang, J.; Hu, H.; Cui, C. Organometallics, 2013, 32, 6875-6878. (23) (a) Bowman, W. R.; Storey, J. M. D. Chem. Soc. Rev. 2007, 36, 1803–1822. (b) Studer, A.; Bossart, M. In Radicals in Organic Synthesis; Renaud, P., Sibi, M. P., Eds.; Wiley-VCH Verlag GmbH: 2008; Vol. 2, p 62. (c) Beaume, A.; Courillon, C.; Derat, E.; Malacria, M. Chem. Eur. J. 2008, 14, 1238-1252.

157

Chapter 6

(24) (a) Boese, R.; Maulitz, A. H.; Stellberg, P. Chem. Ber. 1994, 127, 1887-1889. (b) Kiran, B.; Phukan, A. K.; Jemmis, E. D. Inorg. Chem. 2001, 40, 3615–3618. (c) Anand, B.; Nöth, H.; Schwenk-Kircher, H.; Troll, A. Eur. J. Inorg. Chem. 2008, 2008, 3186–3199. (25) Lisovenko, A. S.; Timoshkin, A. Y. Inorg. Chem. 2010, 49, 10357–10369. (26) (a) Barbon, S. M.; Reinkeluers, P. A.; Price, J. T.; Staroverov, V. N.; Gilroy, J. B. Chem. Eur. J. 2014, 20, 11340–11344. (b) Buemi, G.; Zuccarello, F.; Venuvanalingam, P.; Ramalingam, M.; Salai Cheettu Ammal, S. J. Chem. Soc., Faraday Trans. 1998, 94, 3313–3319. (27) Barbon, S. M.; Price, J. T.; Reinkeluers, P. A.; Gilroy, J. B. Inorg. Chem. 2014, 53, 10585–10593. (28) (a) Findlater, M.; Hill, N. J.; Cowley, A. H. Dalton Trans. 2008, 4419–4423. (b) Aramaki, Y.; Omiya, H.; Yamashita, M.; Nakabayashi, K.; Ohkoshi, S.-i.; Nozaki, K. J. Am. Chem. Soc. 2012, 134, 19989– 19992. (29) Bruker. APEX2 (v2012.4-3), SAINT (Version 8.18C) and SADABS (Version 2012/1). Bruker AXS Inc., Madison, Wisconsin, USA. 2012. (30) Sheldrick, G. Acta Crystallographica Section A 2008, 64, 112.

158

Chapter 7

Intramolecular Hydride Transfer Reactions in (Formazanate)Boron Dihydride Complexes

The thermolysis reaction of (formazanate)boron dihydrides (LBH2; 10) results in the formation of aminoborane compounds (18) via a series of intramolecular hydride transfer reactions. Based on the NMR analysis, several intermediates of the hydride transfer reaction were identified, and a mechanism of the reaction was proposed. The key steps of the proposed mechanism include the isomerization from the six-membered chelate ring to the five- membered chelate ring followed by hydride transfer to the formazanate ligand and N-N bond cleavage. Importantly, in case of a ligand with an N-Mes substituent it was possible to characterize an intermediate (21c-i) in this transformation that shows an unexpected cyclohexadiene moiety, which results from hydride transfer to the ortho position of the mesityl substituent. The results presented here show a new type of ligand-based 2e-reduction of the formazanate ligand.

Parts of this chapter will be submitted for publication::

M.-C. Chang and E. Otten*, submitted

Chapter 7

Chapter 7 Intramolecular Hydride Transfer Reactions in (Formazanate)Boron Dihydride Complexes 7.1 Introduction Group 13 metal hydrides and their Lewis base adducts have been extensively studied due to their great application potential in the field of organic synthesis and material science. For example, LiAlH4 and NaBH4 are very common reducing agents, and the Lewis base stabilized

LMH3 complexes (M: B, Al, Ga, In; L: Lewis base) show chemo- and stereo-selective reduction of unsaturated organic substrates.1 In the field of material science, aluminum and gallium hydride complexes are used as precursors for metal thin films, nanocrystalline metal, and group 13-15 semiconductors.2 In recent years, aluminum or boron hydride compounds 3 such as LiAlH4 and H3NBH3 are investigated as hydrogen storage materials. When we surveyed the literature, we surprisingly noticed that the research of boron hydride compounds bearing multidentate ligands is very limited. Besides the well-known catecholborane (HBcat) and pinacolborane (HBpin), which are commonly used in hydroboration4, borylation of arylhalides5 and catalytic C-H activation of hydrocarbons6, there are only a few examples that have been synthesized and fully characterized (Chart 7.1). In 2004, Hey-Hawkins and co- workers reported a series of intramolecularly base-stabilized borane compounds with six- and seven-membered chelate rings (A).7 In 2012, several three-coordinate boron monohydride complexes ligated by the bis(3-methylindolyl)methanes (B) were reported by Mason and co- workers.8 Piers and co-workers indicated the formation of an unstable dipyrrinato boron dihydride complex (C) by UV-Vis absorption and emission spectroscopy.9 Complex C is too reactive to be isolated, and it decomposes to a dipyrromethane-coordinated borane (D) via hydride transfer to the meso position of the ligand. In 2014, a boron dihydride complex bearing bis(imidazolin-2-imine) ligand (E), which can be converted to a thioxoboran salt (F), was reported by Inoue and co-workers.10

In Chapter 6, we reported the 2-electron reduction chemistry of the boron formazanate compound [PhNNC(p-tolyl)NNPh]BF2, which suggested (transient) formation of a reactive low-valent (formazanate)boron species that ultimately forms BN heterocyclic products

(compounds 11-14 in Chapter 6). The heavier group 13 element hydrides (MXHn, M = Ga or In, X = Cl or Br) are known to give low-valent compounds via reductive dehydrogenation,11 but to the best of our knowledge similar chemistry with B or Al compounds is unknown. We thus hypothesized that dehydrogenation of formazanate boron hydrides could provide an alternative entry into low-valent boron chemistry. Here we describe the synthesis and

160

Intramolecular Hydride Transfer Reactions in (Formazanate)Boron Dihydride Complexes characterization of three (formazanate)boron hydride (LBH2, 2) compounds and evaluate their reactivity.

Chart 7.1

7.2 Synthesis of Formazanate Boron Dihydride Complexes The synthesis of (formazanate)boron dihydride complexes (10a,c,f) was achieved by reacting free formazan with borane dimethylsulfide complex (BH3 ∙ SMe2) at room temperature (Scheme 7.1). The products were purified by column chromatography to give the pure compounds in a yield of around 10-30 %. The purity of the products was assessed by NMR spectroscopy and elemental analysis.

Scheme 7.1 Synthetic procedure of LBH2 (10) 7.3 Thermally Induced Intramolecular Hydride Transfer

7.3.1 [PhNNC(p-tolyl)NNPh]BH2 (10a)

In order to test the possibility to promote H2 reductive elimination from a molecular boron 1 complex, the thermolysis of [PhNNC(p-tolyl)NNPh]BH2 (10a) was monitored by H NMR.

After a solution of 10a was heated in an NMR tube (C6D6, 100 °C, 4 hours), the formation of hydrogen gas or B-N heterocyclic products such as those formed upon reduction of the LBF2

161

Chapter 7 analogues (see Chapter 5) was not observed. Nevertheless, the reaction proceeded cleanly to give two new products (17a and 18a) resulting from hydride transfer, which were identified by the 1H NMR spectra (Scheme 7.2 and Figure 7.1). It is worth mentioning that the analogous LBF2 (6) and LBPh2 (9) complexes are stable at high temperature.

Scheme 7.2 Thermally induced intramolecular hydride transfer reaction of 10a (C6D6, 100 °C). In the 1H NMR spectrum of the thermolysis experiment taken after 4 hours at 100 °C (Figure 7.1, middle), a new group of resonances containing all the resonances of the formazanate ligand ([1a]-) can be identified in addition to starting material. The triplet resonance, which has the integration corresponding to one proton at  6.72 ppm, indicates that the two phenyl substituents of the formazanate ligand are no longer equivalent. The most likely explanation for the asymmetry of the formazanate ligand is that it coordinates to the boron center with one internal and one terminal nitrogen atom forming a five-membered chelate ring. The singlet located at  5.23 ppm has the integration corresponding to one proton and does not link to any carbon atoms, which was confirmed by the gHSQCAD spectrum. This singlet resonance indicates that the new compound has an NH functional group. In addition, a very broad resonance at around  5.08 ppm is observed. Measurement of the 1H{11B} spectrum results in sharpening of this signal, which indicates that this is a B-H functional group. Based on all the information collected from the NMR experiments, the formation of the new complex 17a is postulated (Scheme 7.2). Attempts to obtain 17a as a pure compound were not successful due to the formation of a subsequent product 18a that occurs simultaneously with conversion of 10a. The 1H NMR spectrum of 17a is always mixed with either the starting material 10a or the final thermolysis product 18a. In addition, the attempt to isolate complex 17a from the reaction mixture was also not successful. Even though we never got a spectrum of pure 17a in any of the NMR experiments, all the NMR characterizations support the structure of 17a.

162

Intramolecular Hydride Transfer Reactions in (Formazanate)Boron Dihydride Complexes

10a 3.72

4.5 4.0 3.5 3.0 f1 (ppm) 1.05 3.97 1.98 3.30

8.4 8.0 7.6 7.2 6.8 6.4 6.0 5.6 5.2 4.8 4.4 4.0 3.6 3.2 2.8 2.4 2.0 f1 (ppm)

17a

6 5 4 18a 17a 17a 4.0 3.5 17a f1 (ppm) 2.08 7.18 4.00 1.00 1.94 3.38 3.00

8.4 8.0 7.6 7.2 6.8 6.4 6.0 5.6 5.2 4.8 4.4 4.0 3.6 3.2 2.8 2.4 2.0 f1 (ppm)

18a 1.78 2.02 2.03 1.08 0.99 2.54 1.05 1.81 0.55 0.64 1.00 3.05

8.4 8.0 7.6 7.2 6.8 6.4 6.0 5.6 5.2 4.8 4.4 4.0 3.6 3.2 2.8 2.4 2.0 f1 (ppm) 1 Figure 7.1 H NMR spectra of thermolysis reaction of 10a (400 MHz in C6D6); top: starting material 10a; middle: mixture of 10a and 17a; bottom: final product of thermolysis reaction 18a An alternative formulation of the initial hydride transfer product that is consistent with the observed NMR data is 17a', in which the hydride is transferred to the internal nitrogen resulting in a boron-analogue of leuco-verdazyls (Chart 7.2), which are often involved in synthesis of verdazyl radicals and formazans.12 To evaluate which of the two possibilities is most likely, DFT calculation of both five- (17a) and six-membered chelate (17a') isomers were carried out at the B3LYP/6-31G(d) level in the gas phase (Figure 7.2). The calculations show that isomer 17a is more stable than the six-membered chelate ring 17a' (G = 11.9

163

Chapter 7 kcal/mol). In the optimized structure of 17a, the N3-C6 bond (1.308 Å) is shorter than the N1- C6 bond (1.399 Å); in addition, the N-N bond lengths (1.401 Å and 1.391 Å) of the optimized structure of 17a are close to bond lengths of N-N single bonds. This bond length distribution of 17a suggests that it is better described as a boron(III) monohydride complex bearing a doubly reduced formazanate ligand ([1a]-2BIIIH; Chart 7.2) instead of a boron(I) monohydride complex coordinated by a (neutral) formazan ligand ([1a]BIH). In other words, the formation of 17a can be treated as a 2e-reduction of the formazanate ligand. To the best of our knowledge, this is the first reported 2e-reduction of formazanate ligands by a chemical method (hydride reduction in this case).

Chart 7.2 Structure of 17a' (left), and 17 showing the possible resonance structures of [1a]- 2BIIIH and [1a]BIH (right)

H H H N N N N N N N N N N NB NB B H H H

17a' -2 III I 17a: [LH] B H 17a : [LH]B H

Figure 7.2 Optimized structures of 17a (left; N1-N2: 1.401; N1-C6: 1.399; C6-N3: 1.308; N3- N4: 1.391) and 17a' (right; N1-N2: 1.431; N1-C6: 1.415; C6-N3: 1.285; N3-N4: 1.400).

Keeping the same NMR tube at 100 for several hours results in the disappearance of 10a and 17a with the appearance of a new compound 18a (Figure 7.1: bottom). The 1H NMR spectrum of 18a shows two broad singlets at  6.35 and 4.62 ppm, which do not show any

164

Intramolecular Hydride Transfer Reactions in (Formazanate)Boron Dihydride Complexes crosspeaks in the gHSQCAD spectrum. These two broad singlets indicate that the new compound has two inequivalent NH functional groups. In the NOESY spectrum of 18a, the resonances at  6.35 and 4.62 ppm show crosspeaks with the o-CH resonance of the p-tolyl substituent and the o-CH resonances of two phenyl substituents, respectively. In addition, the rest of the resonances of 18a indicate that the formazanate ligand does not have C2 symmetry. Based on the full NMR analysis of 18a, the new compound was assigned as an aminoborane complex (Scheme 7.2).

Even though the isolation of 18a as a pure product was not successful, the formation of 18a was indirectly confirmed by hydrolysis of the reaction mixture. Hydrolysis of 18a in the NMR tube results in the formation of a new compound (19a) and aniline (Scheme 7.3). The formation of aniline was confirmed by overlapping a 1H NMR spectrum of pure aniline with the 1H NMR spectrum of the reaction mixture. The 1H NMR spectrum of 19a shows a 1:1 ratio of the p-tolyl group and the phenyl group. In addition, an N-H resonance having an integration corresponding to one proton can be located at 5.85 ppm. Based on the NMR features mentioned above, we postulate that compound 19a is a borinic acid derivative.13 While we were unable to obtain 19a as a pure compound and characterize it directly, attempted workup of the reaction mixture afforded crystals of the borinic anhydride 20a (Figure 7.3), the formation of which likely goes through 19a.14

Scheme 7.3 Hydrolysis of 18a and the formation of borinic acid (19a), and borinic acid anhydride (20a)

165

Chapter 7

Figure 7.3 Molecular structure of borinic acid anhydride (20a) showing 50% probability ellipsoids. All hydrogen atoms (except for H100) are omitted for clarity. Selected bond length: B1-N1: 1.439(2); N1-N2: 1.401(2); N2-C7: 1.308(2); C7-N3: 1.384(2); N3-B1: 1.430(2); N3- H100: 0.91(2); O1-B1: 1.366.

In order to get more mechanistic insights of the hydride transfer reaction, two LBH2 complexes (10c and 10f), of which the formazanate ligands have different electronic and steric properties of the substituents, were heated under similar conditions in an NMR tube. The R1 and R5 substituents of 10c and 10f are different; therefore, two isomers, which result from hydride transfer to either the R1 or R5 sides of the ligand, can be expected for the intermediate 17 and the product 18 of the reaction.

7.3.2 [PhNNC(p-tolyl)NNMes]BH2 (10c)

The initial result of the thermolysis study (C6D6, 100 °C, 5 hours) of 10c shows the formation of the expected intermediate 17c, which is formed as two isomers 17c-i and 17c-ii with a ratio of 1:2 (Scheme 7.4, Figure 7.4). Besides the expected intermediates 17c, a new intermediate 21c-i was observed in the 1H NMR spectrum. The 1H NMR spectrum of 21c-i features several characteristic resonances: a quintet at  5.83 ppm, a doublet at  5.25 ppm, a quintet at  3.89 ppm, a resonance at  2.07 ppm (overlapped), a triplet at  1.51 ppm and a doublet at  0.83 ppm with the integration ratio of 1:1:1:3:3:3 (Figure 7.4, middle). Even though the 1H NMR spectrum of reaction mixture contains at least four species (10c, 17c-i, 17c-ii, and 21c-i), the full 1H NMR assignment of 21c-i is still achieved by the help of 2D NMR spectra such as gCOSY (Figure 7.5) and gHSQCAD. In the gHSQCAD spectrum, the six resonances mentioned above show crosspeaks, which means that all protons connect to carbon atoms directly, and none of them is an N-H group. The integration ratio of these six resonances suggests that two methyl groups at ortho positions and two hydrogen at meta positions of mesityl substituent are no longer equivalent in 21c-i.

166

Intramolecular Hydride Transfer Reactions in (Formazanate)Boron Dihydride Complexes

Scheme 7.4 The intermediates and products of the thermolysis reaction of 10c

1 Figure 7.4 H NMR spectra of thermolysis reaction of 10c (400 MHz in C6D6); top: starting material 10c; middle: mixture of 10c, 17c-i, 17c-ii, and 21c-i; bottom: mixture of 17c-i and 17c-ii.

167

Chapter 7

The most characteristic resonances of 21c-i are those at 0.83 (3H), 3.89 (1H) and 5.25 (1H) ppm. The three protons doublet at 0.83 ppm is coupling with the quintet at 3.89 ppm with a 3 coupling constant of 7 Hz, which is in the normal range of a JH-H coupling. This feature indicates a (CH3)CH fragment in the structure of 21c-i; therefore, the resonances at 0.83 ppm and 3.89 ppm are assigned to H9 and H6 (Figure 7.5), respectively. The gCOSY spectrum shows the coupling between H6 and the resonance at 5.25 ppm, which is in a typical range of a C-C double bond which allows its assignment as H5.

H8 H9

H3 H5 H6

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0

Figure 7.5 gCOSY spectrum of a mixture of 10c, 17c-i, 17c-ii, 21c-i, (500 MHz, C6D6)

These NMR features suggest that 21c-i has a mesityl-derived 2,4,6-trimethylcyclohexadiene substituent. Based on these NMR features of 21c-i, we can conclude that one of the hydrides of the BH2 unit shifts to the ortho position of mesityl substituent resulting in dearomatization of the mesityl substituent to a cyclohexadiene substituent. A similar dearomatization reaction was reported by Barclay and co-workers in 1973,15 who showed intermolecular hydride transfer from vitride, which is a comparable hydride reducing agent with LiAlH4, to 2,4,6-tri- t-butylnitrobenzene to give 2,4,6-tri-t-butyl-2,4-cyclohexadienone oxime at room

168

Intramolecular Hydride Transfer Reactions in (Formazanate)Boron Dihydride Complexes temperature (Scheme 7.5). The 2,4,6-tri-t-butyl-2,4-cyclohexadienone oxime was further transformed to 2,4,6-tri-t-butylaniline and 2,4,6-tri-t-butylnitrosobenzene at 170. To the best of our knowledge, the intramolecular hydride shift from a boron hydride to an aromatic ring to form compound 21c-i is without precedent in the literature.

Scheme 7.5 Hydride reduction of 2,4,6-tri-t-butylnitrobenzene by vitride at room temperature

Keeping the same NMR tube at room temperature for several days, the intensity of 21c-i decreased, the intensity of the 1H NMR signal 17c-i increased while the intensity of 10c and 17c-ii remained unchanged. The change of the intensities suggests that compound 21c-i is an intermediate in the conversion of 10c to 17c-i. It is also reasonable to assume that complex 21c-ii, which is not observed in the NMR experiment, is the intermediate from 10c to 17c-ii. Heating the same NMR tube at 100 for several hours shows similar reactivity as observed for 17a: the disappearance of 17c and the appearance of 18c, which is formed as two isomers 18c-i and 18c-ii. At the end of the thermolysis reaction of 10c, the ratio of two isomers (18c- i:18c-ii) is close to 1:1 (Figure 7.4, bottom).

7.3.3 [C6F5NNC(p-tolyl)NNMes]BH2 (10f)

Complex 10f is an interesting subject for the thermolysis experiment due to its electronic and steric asymmetry in the formazanate ligand ([1f]-), from which both isomer i and ii are - expected to form. The formazanate ligand [1f] has one –C6F5 substituent, which is a strong electron-withdrawing group, and one mesityl substituent, which is an electron-donating group with steric hindrance. The result of the thermolysis of 10f can give us some hint about the influence from the –C6F5 substituent on reactivity and selectivity of the intramolecular hydride transfer reaction. Applying similar reaction conditions (C6D6, 100 ) of 10a/c to 10f, the thermolysis of 10f is much faster than 10a/c. At 100 , all starting material 10f was consumed within 1 hour, and only a single product was formed (17f-i), without any intermediates observable by NMR. Keeping the NMR tube at 130 for few hours leads to the formation of compound 18f-i. This example shows that introducting an electron-

169

Chapter 7 withdrawing group results in a change in reactivity (faster reaction) and selectivity (single isomer formation) of the reaction (Scheme 7.6).

 H  H F N N N N N N N o F o F F N N 100 C 130 C NB B NB H F HN H H F F F F F F F F F F 10f 17f-i 18f-i

Scheme 7.6 Thermolysis reaction of 10f 7.4 Kinetic Study and Proposed Mechanism

7.4.1 Kinetic Study of the Thermolysis of [PhNNC(p-tolyl)NNPh]BH2 (10a)

The kinetic study of the transformation from 10a to 17a was carried out in NMR tubes at 100 °C with two different concentrations (EXP 1: 3.5 mM; EXP 2: 17.2 mM). The kinetic data of both EXP 1 and EXP 2 were followed to 2 half-lifes as shown in Figure 7.5 (left). The data show that the transformation from 10a to 17a is first-order in 10a with a rate constant of 0.004 min-1 at 100 °C. The first-order transformation from 10a to 17a suggests that the hydride transfer reaction is an intramolecular reaction. After following the reaction from 10a to 17a for 2 half-lifes, the NMR tubes of both EXP 1 and EXP 2 were heated up to 130 °C for 30 minutes to complete the transformation from 10a to 17a. When all 10a is consumed, the subsequent disappearance of 17a to form 18a was monitored at 130 °C for 2 half-lifes as shown in Figure 7.5 (right). These data suggest that also the transformation from 17a to 18a is a first-order reaction with a rate constant of 0.009 min-1 at 130 °C.

Figure 7.5 Kinetic study of transformations 10a→17a (left, 100 °C) and 17a→18a (right, 130 °C). Exp 1: [10a] = 3.5 mM; Exp 2: [10a] = 17.2 mM.

170

Intramolecular Hydride Transfer Reactions in (Formazanate)Boron Dihydride Complexes

7.4.2 Proposed Mechanism and DFT Calculations

Based on the information mentioned above, which include the identified intermediates and results from kinetic studies, a mechanism for the thermally induced hydride transfer reaction is proposed in Scheme 7.7. A summary of the steps that are proposed to take place in the transformation from 10 to 18 is as follows:

Step 1 (10→22): Reversible isomerization from six-membered chelate ring to five-membered chelate ring. The similar reversible isomerization of formazanate ligand was shown to occur in formazanate Zn complexes (Chapter 3).

Step 2 (22→21): Hydride transfers to the ortho position of the aromatic ring, which results in the dearomatization of the aromatic ring and the formation of a cyclohexadiene substituent.

Step 3 (21→17): Hydride shifts from the cyclohexadiene substituent to the terminal nitrogen and the cyclohexadiene substituent is aromatized back to an aromatic ring.

Step 4 (17→18): N-N bond cleavage and formation of an aminoborane product.

Scheme 7.7 Proposed Mechanism of thermally induced intramolecular hydride transfer reaction of (formazanate)boron dihydride complex (10c)

All the proposed intermediates were subjected to DFT calculations at the B3LYP/6-31G(d) level in the gas phase (Figure 7.6). The calculated energies of the intermediates reveal that the hydride transfer reaction of compound 10 is energetically uphill for the first step and downhill for all the following steps.

171

Chapter 7

Figure 7.6 Energy diagram of the thermally induced intramolecular hydride transfer reaction of (formazanate)boron dihydride complex (10c). (Red: isomer i; blue: isomer ii)

For the thermolysis of 10c, the intermediate 21c-i and 21c-ii have similar calculated energies (-7.5 vs. -6.2 kcal/mol), which suggest that both intermediates are possible to form. The reason for the formation of only a single isomer (21c-i) might due to a fast subsequent transformation (21→17) in the case of isomer ii. In other words, the consumption of 21c-ii is faster than its generation. The fast transformation from 21 to 17 might also the reason why we don’t see intermediate 21a in the thermolysis reaction of 10a. It is worth pointing out that based on the experimental data we have, the possibility of a direct hydride transfer resulting in the formation from 10 to 17, in which the intermediate 21 is not involved, can not be completely ruled out.

172

Intramolecular Hydride Transfer Reactions in (Formazanate)Boron Dihydride Complexes

In the case of 10f, the calculated energies of all possible intermediates also suggest that formation of isomers ii is possible. The reason for single isomers (17f-i, and 18f-i) formation is likely due to the strong electron-withdrawing –C6F5 substituent, which localizes the - negative charge of formazanate ligand [1f] at the terminal nitrogen close to the –C6F5 substituent resulting in resonance structure G being the dominant contributor (Chart 7.3, see also the structure of 5f in Chapter 3). The resonance structure G makes the isomerization from G to 22f-i is more favorable than the isomerization from H to 22f-ii and leads to a single product (isomer 18f-i in this case). In addition, the resonance structure G also favors the hydride transfer to the terminal nitrogen close to the mesityl substituent due to the less electron density at that nitrogen in the case of the direct hydride transfer pathway.

+ Chart 7.3 Resonance structures of 10f, the [BH2] unit was omitted for clarity.

7.5 Discussion The thermally induced boron to ligand hydride transfer reaction of compounds 10 presented here is shown to take place in two steps. A first hydride (2e/H+) transfer from the boron center to the ligand backbone forms the (LH)BH intermediate 17. Subsequently, this is converted to the final product 18, in which transfer of the remaining borohydride is accompanied by N-N bond cleavage to form a boron tri(amido) complex. For the initial hydride transfer, an intermediate can be intercepted in which a borohydride has reacted with the N-Ar ring to result in dearomatization, which is very rare in the literature. Most of the reported examples of metal to arenes hydride transfer are based on mononuclear transition metal hydride complexes, such as Nb16, W17, Ta18, Zr19, Co20, and Fe21. In these examples, an anionic cyclohexadienyl moiety is formed upon hydride shifts from the metal center to an arene, which is further stabilized by coordination to a metal center. In 2014, examples of C-C bond cleavage and rearrangement of benzene and toluene by a trinuclear titanium hydride were reported by Hou and co-workers.22 These reactions are promoted by highly reactive multinuclear metal-hydride clusters.23 The only non-transition metal example we were able to find in the literature is based on an aluminum hydride complex (Scheme 7.5), which was reported by Barclay and co-workers in 1973.15 To the best of our knowledge, the

173

Chapter 7 transformation from compound 10c to 21c-i is the first example of the dearomatization of an arene by using boron hydride species. In general, boron hydride reagents are not sufficient enough to dearomatize an arene via hydride transfer. The major reason for the observation of an intramolecular boron to arene hydride transfer reactions in compounds 10 is due to the redox-active nature of formazanate ligands: they behave as good electron reservoirs. Once the hydride shifts to an ortho position of an N-Ar substituent, the electron density introduced by the hydride can be released into the ligand backbone instead of being accumulated at the C6 ring resulting in a formation of a neutral cyclohexadiene imine moiety (I in Scheme 7.8) and two anionic N-donor groups bound to the boron center. The cyclohexadiene moiety of structure I will then rearomatize back to an aromatic ring resulting in formation of structure J. J is related to the monoanionic formazanate ligand in the starting material via 2e/H+ transfer and is thus equivalent to a 2e-reducted neutral formazan.

Scheme 7.8 Hydride reduction of formazanate ligand; the [BH]+2 moiety was omitted for clarity The second half of the hydride transfer reaction is a N-N bond cleavage of the formazanate backbone resulting in an aminoborane product (compounds 18). The structure of 18 is similar to the imidoborane described in Chapter 6 (fragment X, in Chart 7.4), both of which are boron-containing heterocycles having a very rare BN3C core structure. The BN3C core of compounds 18 is a triazaborole, which makes compounds 18 a potential candidate for bioisosteric replacement of imidazoles and pyrazoles due to the isoelectronic relationships between the B-N and C=C units.24 It is worth pointing out that since the first triazaborole has been prepared by Dewar and co-worker in 1971,25 the synthetic methods of triazaboroles are still very limited. The most common synthetic procedure for preparing triazaboroles is based on reactions of amidrazones and boronic acid derivatives (RBX2; X= Cl, Br, OMe, OEt, OH and NMe2) (Chart 7.4).26 Therefore, the formation of compounds 18 from (formazanate)boron dihydride (10) opens a potential route for preparing triazaboroles.

174

Intramolecular Hydride Transfer Reactions in (Formazanate)Boron Dihydride Complexes

Chart 7.4

7.6 Conclusion The formation of the complexes 17, 18 and 21, which resulted from the thermally induced boron to formazanate hydride transfer reaction, shows a new type of formazanate-based 2e- reduction. The results of the thermolysis reaction of three different starting materials (10a, 10c, and 10f) suggests that the reactivity and selectivity of the hydride transfer reaction are tunable by different substituents on the formazanate ligands. The intermediates identified here expand our understanding of the reactivity of boron hydrides and form a starting point for further investigations on 2-electron reduction chemistry of complexes bearing formazanate ligands

7.7 Experimental Section General Considerations. All manipulations were carried out under nitrogen atmosphere using standard glovebox, Schlenk, and vacuum-line techniques. Toluene, hexane, and pentane

(Aldrich, anhydrous, 99.8%) were passed over columns of Al2O3 (Fluka), BASF R3-11- supported Cu oxygen scavenger, and molecular sieves (Aldrich, 4 Å). Deuterated solvent

(C6D6) was vacuum transferred from Na/K alloy and stored under nitrogen. NMR spectra were recorded on Mercury 400, Inova 500 or Agilent 400 MR spectrometers. The 1H and 13C NMR spectra were referenced internally using the residual solvent resonances and reported in ppm relative to TMS (0 ppm); J is reported in Hz. Assignment of NMR resonances was aided by gradient-selected gCOSY, NOESY, gHSQCAD and/or gHMBCAD experiments using standard pulse sequences. 11B NMR spectra were recorded in quartz (or normal glass) NMR tubes using a OneNMR probe on an Agilent 400 MR system. Elemental analyses were performed at the Mi-croanalytical Departement of the University of Groningen.

Kinetic study

EXP 1: A NMR young tube was charged with [PhNNC(p-tolyl)NNPh]BH2 (10a) 4.4 mg, hexane (1.6 L), and C6D6 0.45 mL.

175

Chapter 7

EXP 2: A NMR young tube was charged with [PhNNC(p-tolyl)NNPh]BH2 (10a) 20.3 mg, hexane (8.0 L), and C6D6 0.45 mL.

In order to follow the transformation from 10a to 17a, the tubes of both EXP1 and EXP 2 were heated at 100 °C and monitored by NMR spectroscopy for every 40 to 60 mins. The concentration of 10a in each 1H NMR spectra was determined by the integrations of the resonance located at 8.18 ppm (2H of 10a) and 0.87 ppm (6H of hexane).

After followed the transformation from 10a to 17a for 2 half-lifes, the NMR tubes were heated at 130 °C for 30 mins to convert all of the 10a to 17a. After which, the transformation from 17a to 18a was promoted at 130 °C and followed by NMR spectroscopy for every 15 to 40 mins. The concentration of 17a in each 1H NMR spectra was determined by the integrations of the resonance located at 7.95 ppm (2H of 17a) and 0.87 ppm (6H of hexane).

DFT Calculation

Calculations were performed with the Gaussian09 program using density functional theory (DFT). Geometries were fully optimised starting from the X-ray structures using the B3LYP exchange-correlation functional with the 6-31G(d) basis set. Geometry optimisations were performed without (symmetry) constraints, and the resulting structures were confirmed to be minima on the potential energy surface by frequency calculations (number of imaginary frequencies = 0).

Synthesis of LBH2

[PhNNC(p-tol)NNPh]BH2 10a. A schlenk flask was charged with [PhNNC(p-tolyl)NNHPh]

[PhNNC(p-tol)NNMes]BH2 10c. The procedure is similar with 10a. [PhNNC(p- tolyl)NNHMes] (1c) (302.0 mg, 0.85 mmol) and BH3(SMe2) (0.08 mL, 0.85 mmol) was used.

The product was purified by chromatography (DCM/hexane = 1/2, silica gel, Rf = 0.68). After which 64.8 mg (0.18 mmol, 21 %) of 10c was obtained.

[MesNNC(p-tol)NNC6F5]BH2 10f. The procedure is similar with 10a. [MesNNC(p- tolyl)NNHC6F5] (1f) (324.5 mg, 0.73 mmol) and BH3(SMe2) (0.07 mL, 0.74 mmol) was used.

176

Intramolecular Hydride Transfer Reactions in (Formazanate)Boron Dihydride Complexes

The product was purified by chromatography (DCM/hexane = 1/2, silica gel). After which 40.9 mg (0.09 mmol, 12 %) of 10f was obtained.

Characterization Data of products and intermediates

1 [PhNNC(p-tolyl)NNPh]BH2 10a. H NMR (400 MHz, C6D6, 25 °C): 8.16 (d, 2H, J = 8.2, p- tolyl CH), 7.88 (d, 4H, J = 8.1, Ph o-CH), 7.10 (d, 2H, p-tolyl CH, overlap with C6D6), 6.97 (t,

4H, J = 8.3, Ph m-CH), 6.88 (tt, 4H, J = 7.3, 1.8, Ph p-CH), 3.67 (bs, 2H, BH2), 2.11 (p-tolyl 11 13 CH3). B NMR (128.3 MHz, C6D6, 25 °C): -10.8. C NMR (100.6 MHz, C6D6, 25 °C): 153.3 (NCN), 145.9 (Ph i-C), 138.7 (p-tolyl p-C), 131.7 (p-tolyl i-C), 129.3 (p-tolyl CH),

128.9 (Ph m-CH), 128.2 (Ph p-CH), 125.4 (p-tolyl CH), 122.4 (Ph o-CH), 20.9 (p-tolyl CH3).

1 [PhNNC(p-tolyl)NNHPh]BH 17a. H NMR (C6D6, 500 MHz, 25 °C): 7.96-7.94 (m, 4H, p- Tolyl CH, Ph o-CH), 7.25 (t, J = 8.0 Hz, 2H, Ph m-CH), 7.02-6.97 (m, 3H, Ph m-CH, Ph p- CH, overlap with Ph m-CH of 10a), 6.94-6.89 (m, 2H, p-Tolyl CH, overlap with Ph p-CH of 10a), 6.72 (t, 7.4 Hz, 1H, Ph p-CH), 6.39 (d, 7.9 Hz, 2H, Ph o-CH), 5.26 (s, 1H, NH), 5.08(bs, 13 1H, BH), 2.00 (s, 3H, p-Tolyl CH3). C NMR (C6D6, 125 MHz, 25 °C): 151.7 (NCN), 148.4 (Ph i-C), 143.5 (Ph i-C), 139.0 (p-Tolyl i-C), 129.3 (Ph m-C), 129.2 (Ph m-C), 128.9 (p-Tolyl CH), 128.4 (p-Tolyl CH), 128.0 (p-Tolyl p-C), 124.0 (Ph p-C), 120.7 (Ph p-C), 117.7 (Ph o- 11 C), 112.8 (Ph o-C), 20.9(p-Tolyl CH3). B NMR (C6D6, 128 MHz, 25 °C): 23.82

1 [PhNNC(p-tolyl)NH]BNHPh 18a. H NMR (C6D6, 400 MHz, 25 °C): 7.68 (d, 8.0 Hz, 2H ,Ph o-CH), 7.52 (d, 8.0 Hz, 2H , p-tolyl CH), 7.23 (t, 8.1 Hz, 2H ,Ph m-CH), 7.08 (t, 7.6 Hz, 2H ,Ph m-CH), 6.95 (t, 7.6 Hz, 1H ,Ph p-CH), 6.93 (d, 8.0 Hz, 2H , p-tolyl CH), 6.82 (t, 7.6 Hz, 1H ,Ph p-CH), 6.57 (d, 7.6 Hz, 2H ,Ph o-CH), 6.34 (bs, NH), 4.62 (bs, NH), 2.05 (s, 13 3H, p-Tolyl CH3). C NMR (C6D6, 100 MHz, 25 °C): 147.0 (p-tolyl i-C), 144.4 (Ph i-C), 144.3 (Ph i-C), 138.5 (p-Tolyl p-C), 129.4 (Ph m-C), 129.2 (Ph m-C), 129.2 (p-tolyl CH), 125.0 (p-Tolyl CH), 123.0 (Ph p-C), 120.7 (Ph p-C), 119.5 (Ph o-C), 118.3 (Ph p-C), 112.8 11 (NCN), 20.9 (p-Tolyl CH3). B NMR (C6D6, 128 MHz, 25 °C): 23.06

1 [PhNNC(p-tolyl)NH]BOH 19a + aniline. H NMR (C6D6, 500 MHz, 25 °C): 8.15 (dd, 8.6, 1.0 Hz, 2H ,Ph o-CH), 7.58 (d, 8.3 Hz, 2H , p-tolyl CH), 7.30 (dd, 8.5, 7.0 Hz, 2H ,Ph m-CH), 6.99 (d, 8.5 Hz, 2H , p-tolyl CH), 6.98 (tt, 7.5, 1.0 Hz, 1H ,Ph p-CH), 2.09 (s, 3H, p-Tolyl

CH3). aniline: 7.06 (dd, 8.5, 7.5 Hz, 2H ,Ph m-CH), 6.71 (tt, 7.4, 0.9 Hz, 1H ,Ph p-CH), 6.34

(dd, 8.4, 1.0 Hz, 2H ,Ph o-CH), 2.75 (bs, 2H, NH2)

1 [PhNNC(p-tolyl)NNMes]BH2 10c. H NMR (C6D6, 500 MHz, 25 °C): 8.18 (d, J = 8.5 Hz, 2H, p-tolyl CH), 7.92 (d, J = 8.8 Hz, 2H, Ph o-CH), 7.08 (d, J = 8.0 Hz, 2H, p-tolyl CH), 7.01

177

Chapter 7

(t, J = 8.1 Hz, 2H, Ph m-CH), 6.91 (tt, J = 7.4, 1.7 Hz, 1H, Ph p-CH), 6.54 (s, 2H, Mes m-CH),

3.53 (bs, 2H, BH2), 2.10 (s, 3H, p-tolyl CH3), 2.00 (s, 6H, Mes o-CH3), 1.98 (s, 3H, Mes p- 13 CH3). C NMR (C6D6, 125 MHz, 25 °C): 153.1(NCN), 145.8 (Ph i-C), 144.0 (Mes i-C), 138.6 (p-tolyl i-C), 137.9 (Mes p-C), 134.2 (Mes o-C), 131.6 (p-tolyl p-C), 129.3 (p-tolyl CH), 129.1(Ph m-CH), 129.0 (Mes m-CH), 128. (Ph p-CH), 125.3 (p-tolyl CH), 122.5 (Ph o-CH), 11 20.9 (p-tolyl CH3), 20.5 (Mes p-CH3), 17.9 (Mes o-CH3). B NMR (C6D6, 128 MHz, 25 °C):

-8.42 (295 Hz). Anal. Calcd for C23H25BN4: C, 75.01; H, 6.84; N, 15.21. Found: C, 75.24; H, 6.92; N, 15.06.

1 [PhNNC(p-tolyl)NNHMes]BH 17c-i. H NMR (C6D6, 500 MHz, 25 °C): 8.22 (d, J = 8.4 Hz, 2H, p-tolyl CH), 8.37 (dm, J = 8.2 Hz, 2H, Ph o-CH), 7.17 (dd, J = 8.6, 7.4 Hz, 2H, Ph m-CH),

7.15 (2H, p-tolyl CH, overlap with C6D6), 6.21 (tt, J = 6.2, 1.1 Hz, 2H, Ph p-CH), 6.62 (s, 2H,

Mes m-CH), 5.46 (s, 1H, NH), 2.14 (s, 3H, p-tolyl CH3), 2.05 (s, 3H, Mes p-CH3), 1.92 (s, 13 6H, Mes o-CH3). C NMR (C6D6, 125 MHz, 25 °C): 129.9 (Mes m-CH), 128.9 (p-tolyl CH),

123.8 (Ph p-CH), 117.6 (Ph o-CH), 21.0 (p-tolyl CH3), 20.9 (Mes p-CH3), 17.9 (Mes o-CH3).

1 [MesNNC(p-tolyl)NNHPh]BH 17c-ii. H NMR (C6D6, 500 MHz, 25 °C): 8.01 (d, J = 8.2 Hz, 2H, p-tolyl CH), 7.05 (d, J = 7.3 Hz, 2H, p-tolyl CH), 7.04 (m, 2H, Ph m-CH), 6.85 (s, 2H, Mes m-CH), 6.73 (tt, J = 7.4, 1.1 Hz, 1H, Ph p-CH), 6.57 (dd, J = 7.7, 0.9 Hz, 2H, Ph o-CH),

5.38 (s, 1H, NH), 2.34 (s, 6H, Mes o-CH3), 2.16 (s, 3H, Mes p-CH3), 1.97 (s, 3H, p-tolyl CH3). 13 C NMR (C6D6, 125 MHz, 25 °C):129.0 (p-tolyl CH), 129.0 (Mes m-CH), 128.2 (p-tolyl CH),

120.6 (Ph p-CH), 112.7 (Ph o-CH), 20.8 (p-tolyl CH3), 20.7 (Mes p-CH3), 18.3 (Mes o-CH3).

1 [PhNNC(p-tolyl)NH]BNHMes 18c-i. H NMR (C6D6, 500 MHz, 25 °C): 7.79 (dm, J = 8.8 Hz, 2H, Ph o-CH), 7.33 (d, J = 8.2 Hz, 2H, p-tolyl CH), 7.31 (dd, J = 8.4, 7.4 Hz, 2H, Ph m- CH), 6.99 (tt, J = 7.3, 1.1 Hz, 1H, Ph p-CH), 6.89 (d, J = 8.0 Hz, 2H, p-tolyl CH), 6.83 (s, 2H,

Mes m-CH), 5.84 (s, 1H, NH), 3.86 (s, 1H, NH), 2.16 (s, 3H, Mes p-CH3), 2.07 (s, 6H, Mes o- 13 CH3), 2.03 (s, 3H, p-tolyl CH3). C NMR (C6D6, 125 MHz, 25 °C): 147.1 (NCN), 145.1(Ph i- C), 138.2 (p-tolyl p-C), 133.8 (Mes o-C),129.2 (Ph m-CH), 124.9 (p-tolyl CH), 122.4 (Ph p-

CH), 118.6 (Ph o-CH), 20.8 (p-tolyl CH3), 18.5 (Mes o-CH).

1 [MesNNC(p-tolyl)NH]BNHPh 18c-ii. H NMR (C6D6, 500 MHz, 25 °C): 7.58 (dm, J = 8.2 Hz, 2H, p-tolyl CH), 7.08 (dd, J = 7.8, 7.4 Hz, 1H, Ph m-CH), 6.94 (d, J = 8.1 Hz, 2H, p-tolyl CH), 6.88 (s, 2H, Mes m-CH), 6.81 (tt, J = 7.4, 1.0 Hz, 1H, Ph p-CH), 6.51 (s, 1H, NH), 6.48

(dd, J = 7.5, 1.1 Hz, 2H, Ph o-CH), 4.21 (s, 1H, NH), 2.32 (s, 6H, Mes o-CH3), 2.16 (s, 3H, 13 Mes p-CH3), 2.07 (s, 3H, p-tolyl CH3). C NMR (C6D6, 125 MHz, 25 °C): 146.7 (NCN),

178

Intramolecular Hydride Transfer Reactions in (Formazanate)Boron Dihydride Complexes

144.5(Ph i-C), 138.0 (p-tolyl i-C), 137.7 (Mes o-C), 133.8 (p-tolyl p-C), 124.8 (p-tolyl CH),

120.0 (Ph p-CH), 117.5 (Ph o-CH), 20.7 (p-tolyl CH3), 18.1 (Mes o-CH).

1 21c-i. H NMR (C6D6, 500 MHz, 25 °C): 8.13 (d, J = 8.2 Hz, 2H, p-tolyl CH), 7.95 (dm, J = 8.7 Hz, 2H, Ph o-CH), 7.24 (dd, J = 8.6, 7.4 Hz, 2H, Ph m-CH), 7.03 (d, J = 8.6 Hz, 2H, p-tolyl CH), 6.98 (m, 1H, Ph p-CH), 5.83 (quintet, J = 1.5 Hz, 1H, H3), 5.26 (d, J = 5.4 Hz, 1H, H5), 3.89 (quintet, J = 6.6 Hz, 1H, H6), 2.07

(bs, 6H, p-tolyl CH3 + H7), 1.51 (t, J = 1.4 Hz, 3H, H8), 0.83 (d, 13 J = 7.2 Hz, 3H, H9). C NMR (C6D6, 125 MHz, 25 °C): 171.7 (C1), 117.9 (Ph o-CH), 32.94 (C6), 134.2 (C3), 129.0 (C5), 20.6(C8), 20.3 (p-tolyl), 17.9(C7), 17.1 (C9).

1 [C6F5NNC(p-tolyl)NNMes]BH2 10f. H NMR (C6D6, 500 MHz, 25 °C): 8.10 (d, J = 8.0 Hz, 2H, p-tolyl o-CH), 7.05 (d, J = 8.0 Hz, 2H, p-tolyl o-CH), 6.60 (s, 2H, Mes m-CH), 3.31 (bs, 13 2H, BH2), 2.16 (s, 6H, Mes o-CH3), 2.08 (s, 3H, p-tolyl CH3), 1.97 (s, 3H, Mes p-CH3). C

NMR (C6D6, 125 MHz, 25 °C): 153.8 (NCN), 143.6 (Mes i-C), 143.3 (dm, J = 251.4 Hz, C6F5

CF), 140.2 (dm, J = 251.5 Hz, C6F5 CF), 139.4 (p-tolyl i-C), 139.2 (Mes p-C), 137.5 (dm, J =

254.3 Hz, C6F5 CF), 134.2 (Mes o-C), 130.6 (p-tolyl p-C), 129.5 (Mes m-CH), 129.5 (p-tolyl

CH), 125.3 (p-tolyl CH), 122.2 (td, J = 12.0, 4.5 Hz, C6F5 i-C), 20.9 (p-tolyl CH3), 20.5 (Mes 19 p-CH3), 18.0 (Mes o-CH3). F NMR (C6D6, 375 MHz, 25 °C): -148.3 (d, J = 19.7 Hz, 2F,

C6F5 o-CF), -155.5 (t, J = 22.1 Hz, 1F, C6F5 p-CF), -162.3 (td, J = 22.4, 5.4 Hz, 2F, C6F5 m- 11 CF). B NMR (C6D6, 128 MHz, 25 °C): -7.11(442 Hz).

1 [C6F5NNC(p-tolyl)NNHMes]BH 17f-i. H NMR (C6D6, 500 MHz, 25 °C): 8.28 (d, J = 8.4 Hz, 2H, p-tolyl o-CH), 7.10 (d, J = 8.0 Hz, 2H, p-tolyl o-CH), 6.63 (s, 2H, Mes m-CH), 5.51

(s, 1H, NH), 4.65 (bs, 1H, BH), 2.12 (s, 3H, p-tolyl CH3), 2.05 (s, 3H, Mes p-CH3), 1.95 (s, 13 6H, Mes o-CH3). C NMR (C6D6, 125 MHz, 25 °C): 152.4 (NCN), 142.6 (dm, J = 250.2 Hz,

C6F5 CF), 141.9 (Mes o-C), 139.4 (p-tolyl p-C), 139.2 (dm, J = 252.3 Hz, C6F5 CF), 137.6

(dm, J = 252.0 Hz, C6F5 CF), 133.4 (Mes p-C), 129.9 (Mes m-C), 129.0 (p-tolyl CH), 128.9

(Mes i-C), 128.8 (p-tolyl CH), 126.4 (p-tolyl i-C), 119.3 (td, J = 12.5, 4.5 Hz, C6F5 i-C), 20.9 19 (p-tolyl CH3), 20.3 (Mes p-CH3), 17.8 (Mes o-CH3). F NMR (C6D6, 375 MHz, 25 °C): -

149.7 (dm, J = 23.7 Hz, 2F, C6F5 o-CF), -159.0 (t, J = 22.0 Hz, 1F, C6F5 p-CF), -163.5 (tdm, J 11 = 22.2, 5.5 Hz, 2F, C6F5 m-CF). B NMR (C6D6, 128 MHz, 25 °C): 23.4 (546 Hz).

1 [C6F5NNC(p-tolyl)NH]BNHMes 18f-i. H NMR (C6D6, 500 MHz, 25 °C): 7.33 (d, J = 8.2 Hz, 2H, p-tolyl o-CH), 6.86 (d, J = 7.9 Hz, 2H, p-tolyl o-CH), 6.77 (s, 2H, Mes m-CH), 5.77

(s, 1H, NH), 3.54 (s, 1H, NH), 2.19 (s, 6H, Mes o-CH3), 2.14 (s, 6H, Mes p-CH3), 2.03 (s, 3H,

179

Chapter 7

13 p-tolyl CH3). C NMR (C6D6, 125 MHz, 25 °C): 149.2 (NCN), 143.1 (dm, J = 251.8 Hz,

C6F5 CF), 138.9 (p-tolyl p-C), 136.4 (Mes o-C), 134.2 (Mes i-C), 129.1 (p-tolyl CH), 128.9

(Mes m-C), 128.9 (Mes p-C), 126.9 (p-tolyl i-C), 125.0 (p-tolyl CH), 119.0 (m, C6F5 i-C), 19 20.8 (p-tolyl CH3), 20.4 (Mes p-CH3), 18.3 (Mes o-CH3). F NMR (C6D6, 375 MHz, 25 °C):

-147.7 (dd, J = 22.9, 5.7 Hz, 2F, C6F5 o-CF), -160.4 (t, J = 22.1 Hz, 1F, C6F5 p-CF), -163.9 11 (td, J = 22.5, 5.5 Hz, 2F, C6F5 m-CF). B NMR (C6D6, 128 MHz, 25 °C): 22.9 (422 Hz).

Crystallographic Data

Suitable crystals of 20a was mounted on a cryo-loop in a drybox and transferred, using inert- atmosphere handling techniques, into the cold nitrogen stream of a Bruker D8 Venture diffractometer. The final unit cell was obtained from the xyz centroids of 9974 (20a) reflections after integration. Intensity data were corrected for Lorentz and polarisation effects, scale variation, for decay and absorption: a multiscan absorption correction was applied, based on the intensities of symmetry-related reflections measured at different angular settings (SADABS).27 The structures were solved by direct methods using the program SHELXS.28 The hydrogen atoms were generated by geometrical considerations and constrained to idealised geometries and allowed to ride on their carrier atoms with an isotropic displacement parameter related to the equivalent displacement parameter of their carrier atoms. Structure refinement was performed with the program package SHELXL.28 Crystal data and details on data collection and refinement are presented in following table.

Crystallographic data 20a(THF) chem formula C32H34B2N6O2 Mr 556.27 cryst syst monoclinic color, habit colourless, block size (mm) 0.37 x 0.31 x 0.28 space group C2/c a (Å) 12.2629(5) b (Å) 19.8344(9) c (Å) 14.5639(8)  (°) β (°) 113.0854(13)  (°) V (Å3) 3258.7(3) Z 4 -3 calc, g.cm 1.134 -1 µ(Cu K), mm -1 µ(Mo K), mm 0.072 F(000) 1176 temp (K) 100(2)  range (°) 2.963-28.44 data collected -16:16; -26:26; -

180

Intramolecular Hydride Transfer Reactions in (Formazanate)Boron Dihydride Complexes

(h,k,l) 19:19 min, max transm 0.7156, 0.7457 rflns collected 69661 indpndt reflns 4090 observed reflns 3678

Fo  2.0 σ (Fo) R(F) (%) 4.90 wR(F2) (%) 14.85 GooF 1.076 weighting a,b 0.0863, 4.2507 params refined 196 min, max resid -0.285, 0.4 dens

7.8 Reference (1) (a) Downs, A. J.; Pulham, C. R. Chem. Soc. Rev., 1994,23, 175-184. (b) Aldridge. S.; Downs, A. J. Chem. Rev., 2001, 101, 3305–3366. (c) Raston, C. L.; Siu, A. F. H.; Tranter, C. J.; Young, D. J. Tetrahedron Lett., 1994, 35, 5915–5918. (d) Abernethy, C. D.; Cole, M. L.; Davies, A. J.; Jones, C. Tetrahedron Lett., 2000, 41, 7567–7570. (2) Jegier, J. A.; Gladfelter, W. L. Coord. Chem. Rev., 2000, 206-207, 631–650. (3) (a) Ley, M. B.; Jepsen, L. H.; Lee, Y.-S.; Cho, Y. W.; Bellosta von Colbe, J. M.; Dornheim, M.; Rokni, M.; Jensen, J. O.; Sloth, M.; Filinchuk, Y.; Jørgensen, J. E.; Besenbacher, F.; Jensen, T. R. Mater. Today, 2014, 17, 122– 128. (b) Sakintuna, B.; Lamari-Darkrim, F.; Hirscher, M. Int. J. Hydrogen Energy, 2007, 32, 1121-1140. (c) Graetz, J.; Wegrzyn, J.; Reilly, J. J. J. Am. Chem. Soc., 2008, 130, 17790–17794. (d) Staubitz, A.; Robertson, A. P. M.; Manners, I. Chem. Rev., 2010, 110, 4079–4124. (4) (a) Obligacion, J. V.; Neely, J. M.; Yazdani, A. N.; Pappas, I.; Chirik, P. J. J. Am. Chem. Soc., 2015, 137, 5855-5858. (b) Zhang, L.; Peng, D.; Leng, X.; Huang, Z. Angew. Chem. Int. Ed., 2013, 52, 3676 - 3680. (c) Obligacion, J. V.; Chirik, P. J. Org. Lett., 2013, 15, 2680-2683. (d) Greenhalgh, M. D.; Thomas, S. P. Chem. Commun., 2013, 49, 11230-11232. (e) Palmer, W. N.; Diao, T.; Pappas, I.; Chirik, P. J. ACS Catal., 2015, 5, 622-626. (f) Mazet, C.; Gerard, D. Chem. Commun., 2011, 47, 298-300 (g) Crudden, C. M.; Edwards, D. Eur. J. Org. Chem., 2003, 4695-4712. (h) Burgess, K.; Ohlmeyer, M. J. Chem. Rev., 1991, 91, 1179-1191. (i) Männig, D.; Nöth, H. Angew. Chem. Int. Ed., 1985, 24, 878-879. (j) Brown, H. C.; Chandrasekharan, J. J. Org. Chem., 1983, 48, 5080-5082. (k) Lane, C. F.; Kabalka, G. W. Tetrahedron, 1976, 32, 981-990. (5) (a) Billingsley, K. L. ; Buchwald, S. L. J. Org. Chem., 2008, 73, 5589-5591. (b) Broutin, P.-E.; Čerňa, I.; Campaniello, M.; Leroux, F.; Colobert, F. Org. Lett., 2004, 6, 4419-4422. (c) Murata, M.; Oyama, T.; Watanabe, S.; Masuda, Y. J. Org. Chem., 2000, 65, 164-168. (d) Baudoin, O.; Guénard, D.; Guéritte, F. J. Org. Chem., 2000, 65, 9268-9271. (e) Murata, M.; Watanabe, S.; Masuda, Y. J. Org. Chem., 1997, 62, 6458-6459. (f) Pintaric, C.; Olivero, S.; Gimbert, Y.; Chavant, P. Y., Duñach, E. J. Am. Chem. Soc., 2010, 132, 11825-11827. (g) Clary, J. W.; Rettenmaier, T. J.; Snelling, R.; Bryks, W.; Banwell, J.; Wipke, W. T.; Singaram, B. J. Org. Chem., 2011, 76, 9602-9610. (6) (a) Iverson, C. N.; Smith III, M. R.; J. Am. Chem. Soc., 1999, 121, 7696-7697. (b) Shimada, S.; Batsanov, A. S.; Howard, J. A. K.; Marder, T. B. Angew. Chem. Int. Ed., 2001, 40, 2168-2171. (c) Cho, J. Y.; Tse , M. K.; Holmes, D.; Maleczka , R. E.; Smith III, M. R . Science, 2002, 295, 305-308. (d) Mkhalid, I. A. I.; Barnard, J. H.; Marder, T. B.; Murphy, J. M.; Hartwig, J. F. Chem. Rev., 2010, 110, 890-931. (e) Lee, C.-I; Zhou, J.; Ozerov, O. V. J. Am. Chem. Soc., 2013, 135, 3560-3566. (f) Lee, C.-I.; DeMott, J.C.; Pell, C. J.; Christopher, A.; Zhou, J.; Bhuvanesh, N.; Ozerov, O. V. Chem. Sci., 2015, 6, 6572-6582. (7) Al-Masri, H. T.; Sieler, J.; Lönnecke, P.; Junk, P. C.; Hey-Hawkins, E. Inorg. Chem., 2004, 43, 7162– 7169. (8) Song, B.; Kirschbaum, K.; Mason, M. R. Organometallics, 2011, 30, 191–196. (9) (a) Bonnier, C.; Piers, W. E.; Parvez, M.; Sorensen, T. S. Chem. Commun., 2008, 24, 4593–4595. (b) Bonnier, C.; Piers, W. E.; Parvez, M. Organometallics, 2011, 30, 1067–1072. (10) Franz, D.; Irran, E.; Inoue, S. Angew. Chem. Int. Ed., 2014, 53, 14264–14268. (11) (a) Ball, G. E.; Cole, M. L.; McKay, A. I. Dalton Trans., 2012, 41, 946-952. (b) Nogai, S. D.; Schmidbaur, H. Organometallics, 2004, 23, 5877–5880. (c) Marcus L Cole; Cameron Jones, A.; Kloth, M. Inorg. Chem., 2005, 44, 4909–4911.

181

Chapter 7

(12) (a) Nineham, A. W. Chem. Rev., 1955, 55, 355–483. (b) Barr, C. L.; Chase, P. A.; Hicks, R. G.; Lemaire, M. T.; Stevens, C. L. J. Am. Chem. Soc., 1999; 121, 8893–8897. (c) Hicks, R. G.; Hooper, R. Inorg. Chem., 1999, 38, 284–286. (13) Loh, Y. K.; Chong, C. C.; Ganguly, R.; Li, Y.; Vidović, D.; Kinjo, R. Chem. Commun., 2014, 50, 8561–8564. (14) Dewar, M. J. S.; Kubba, V. P. J. Org. Chem., 1960; 25, 1722–1724. (15) Barclay, L.; McMaster, I. T.; Burgess, J. K. Tetrahedron Lett., 1973, 14, 3947-3950. (16) Michael D Fryzuk; Christopher M Kozak; Michael R Bowdridge, A.; Patrick, B. O. Organometallics, 2002, 21, 5047–5054. (17) Margaret R Lentz; Phillip E Fanwick, A.; Rothwell, I. P. Organometallics, 2003, 22, 2259–2266. (18) (a) Gavenonis, J.; Tilley, T. D. Organometallics, 2002, 21, 5549–5563. (b) Gavenonis, J.; Tilley, T. D. J. Am. Chem. Soc., 2002; 124, 8536–8537. (19) Bazinet, P.; Tilley, T. D. Organometallics, 2009, 28, 2285–2293. (20) (a) Jonas, K. Angew. Chem. Int. Ed., 1985, 24, 295–311. (b) Jonas, K. J. Organomet. Chem., 1990, 400, 165–184. (21) Jonas, K. Pure Appl. Chem., 1990, 62, 1169–1174. (22) Hu, S.; Shima, T.; Hou, Z. Nature, 2014, 512, 413–415. (23) (a) Tardif, O.; Hashizume, D.; Hou, Z. J. Am. Chem. Soc., 2004; 126, 8080–8081. (b) Cui, P.; Spaniol, T. P.; Okuda, J. Organometallics, 2013, 32, 1176–1182. (c) Shima, T.; Luo, Y.; Stewart, T.; Bau, R.; McIntyre, G. J.; Mason, S. A.; Hou, Z. Nat. Chem., 2011, 3, 814–820. (d) Shima, T.; Hu, S.; Luo, G.; Kang, X.; Luo, Y.; Hou, Z. Science, 2013, 340, 1549–1552. (24) Zurwerra, D.; Quetglas, V.; Kloer, D. P.; Renold, P.; Pitterna, T. Org. Lett., 2015, 17, 74–77. (25) Dewar, M. J. S.; Golden, R.; Spanninger, P. A. J. Am. Chem. Soc., 1971, 93, 3298–3299. (26) (a) Weber, L.; Schnieder, M.; Stammler, H.-G.; Neumann, B.; Schoeller, W. W. Eur. J. Inorg. Chem., 1999, 1999, 1193–1198. (b) Dewar, M. J. S.; Spanninger, P. A. Tetrahedron, 1972, 28, 959–961. (27) Bruker. APEX2 (v2012.4-3), SAINT (Version 8.18C) and SADABS (Version 2012/1). Bruker AXS Inc., Madison, Wisconsin, USA. 2012. (28) Sheldrick, G. Acta Crystallographica Section A, 2008, 64, 112.

182

English Summary Redox-active ligands are ligand platforms that can actively participate in the redox changes during chemical reactions by accepting or donating electrons from or to substrates. Utilizing this special feature, metal centers coordinated by redox-active ligands avoid reaching unstable oxidation states during chemical redox processes. Thus, such complexes can perform catalytic redox reactions that are not possible with conventional ligands. The research of redox-active ligands is getting more and more attention today; this is due to its potential of replacing precious metals in (homogeneous) catalysis with base metals. The ligand systems presented in this thesis are based on the formazan framework, which is a nitrogen-rich analogue of well-known -diketimine ligands. The thesis focuses on Zn and B complexes and uses them to understand the chemical and physical properties of formazanate ligands. The results presented in this thesis establish a strong basis for future research into new (catalytic) chemistry mediated by complexes with redox-active formazanate ligands.

In Chapter 1, an introduction into redox-active ligands from both nature and artificial system is presented. Examples of such systems in nature include cytochrome P450 and catechol dioxygenase, both of which promote the development of bio-inspired redox-active ligands such as porphyrin and catecholate type ligands. In the case of artificial redox-active ligands, several selected examples show an important concept that redox equivalents can be stored in redox-active ligands first and subsequently used for a chemical transformation. In addition, two catalytic reactions are presented to illustrate the concept that base-metal complexes containing an artificial redox-active ligand are capable of catalyzing a multi-electron transformation.

In Chapter 2, the synthetic methods for the synthesis of formazan ligands (R1NHNC(R3)NNR5, 1), which contain large diversity of substituents, were developed. By using different starting materials, the electronic and steric properties at R1, R3, and R5 position are tunable. For synthesizing different formazan derivatives, different reaction conditions and purification procedures are necessary. In addition to the mono-formazan ligands, two examples of phenylene-bridged di-formazan ligands were also prepared.

183

Scheme 1. Formazan synthesis In Chapter 3, mono(formazanate)ZnMe (LZnMe, 4) complexes and a series of bis(formazanate)Zn (L2Zn, 5) complexes were synthetized and fully characterized. The crystal structures and VT-NMR analysis of compounds 5 reveal that the formazanate ligands have flexible coordination chemistry, and can isomerize between six-membered chelate ring and five-membered chelate ring. The cyclic voltammetry of compounds 5 shows up to four (quasi)reversible redox-couples, which suggests that each formazanate ligand can be reduced to its dianion radical ([1]-2) and trianion ([1]-3) forms. The crystal structures, EPR spectra and DFT calculations of the reduced products ([5]- or [5]-2) prove the redox-active property of the formazanate ligands.

R3 R3 R5 R1 2 eq ZnMe N N - ZnMe N N N N N N 2 2 3 3 2 2 NN R Zn R NH N 1 R5 NN 1 5 R Zn NN R R 1 5 Me R R

LH (1)LZnMe (4) L2Zn (5) Scheme 2. Synthesis of compounds 4 and 5

184

Figure 1. Cyclic voltammograms of compounds 5a (line) and 5b (dotted line) showing the presence of reduction waves corresponding to 2-electron reduction of a single formazanate ligand (L3-, reductions III and IV).

In Chapter 4, an unusual zinc to boron formazanate transfer resulting in formation of

(formazanate)boron difluoride complexes (LBF2, 6) is presented. This zinc to boron transmetallation reaction opens a new synthetic route to access compounds 6 with high yield. An important intermediate of the reaction (compound 7), which is a six-coordinated zinc 1 3 5 complex bearing two [R NNC(R )NNR (BF3)] units, is isolated and fully characterized. The transformation of 5→7→6 shows the importance of the flexible coordination chemistry of formazanate ligand and proves the concept that when substrates were introduced, the formazanate ligands can isomerize to form five-membered chelates and open space around the metal center to accommodate incoming substrates.

5 F F R 5 F F 5 R1 R B R N B 5 N N BF3 F F N R -ZnF N N N N 2 3 F R3 Zn R3 N Zn N R B R3 3 F NN N N R NN 5 N N 1 R N N R1 R 1 R1 R

6 5 7

Scheme 3. Zinc to boron formazanate transfer reaction

In Chapter 5, a series of (formazanate)boron difluoride complexes (LBF2, 6) was synthesized by two different methods and fully characterized. The cyclic voltammetry of 6 shows two (quasi)reversible one-electron processes, which indicates that each formazanate ligand is

185 capable of storing two electrons leading to its di- and tri-anion state ([1]-2 and [1]-3), similar to that observed for bis(formazanate)zinc complexes. In addition to the redox-active property of the formazanate ligand, compounds 6 show strong absorption bands in the visible range of the electromagnetic spectrum and are emissive with large Stokes shifts. The large Stokes shift of 6 shows its great potential as fluorescence dye. For comparative purposes, the analogous

(formazanate)boron diphenyl (LBPh2, 9) and (formazanate)boron dihydride complexes (LBH2, 10) were prepared. The optical and electrochemical properties of 9 and 10 are very similar to that of compound 6: these also show strong absorption, large Stokes shift and ligand-based reduction.

Figure 2. UV-Vis absorption spectra (left) and normalized emission spectra (right) of 6c, 6f, and 6g. Data were collected in 10-5 M dry THF solution. The excitation wavelength of emission spectra is at 473 nm.

In Chapter 6, a series of BN-heterocyclic products (11-14) were isolated from the reaction of the [PhNNC(p-tolyl)NNPh]BF2 (6a) with 2 equivalents of Na/Hg. Compound 11-14 are constituted by (formazanate)B(I) (fragment X) and imidoborane (fragment Y) (or their isomers). Compound 12-13 show two (quasi)reversible one-electron processes in the cyclic voltammetry due to the six-membered chelate ring of the fragment X. The formation of 11-14 indicates that the formazanate ligand is capable of stabilizing reactive B(I) center by using its low-lying * orbital. The oxidation of 11 and 13 back to the difluoride starting material 6a by excess XeF2 reveals the B(I) property of the fragment X in the BN-heterocycles.

186

Scheme 4. Formation of BN-heterocyclic products via 2-electron reduction of (formazanate)boron difluoride complex

In Chapter 7, a thermally induced hydride transfer of the (formazanate)boron dihydride

(LBH2, 10) is presented. Two key intermediates (compound 17, and 21) and final product aminoborane (18) were identified and characterized by NMR spectroscopy. The DFT calculations of compounds 17 suggest that it is better described as a boron(III) monohydride complex bearing a doubly reduced formazanate ligand ([1a]-2BIIIH). To the best of our knowledge, compounds 17 are the first reported examples of 2-electron reduction of the formazanate ligand by a chemical method (hydride reduction in this case). The reactivity and selectivity of the thermally induced hydride transfer reaction are tunable by the ligand substitution pattern.

Scheme 5. Proposed mechanism of thermally induced intramolecular hydride transfer reaction of (formazanate)boron dihydride complex (10c)

187

188

Nederlandse Samenvatting Redox-actieve liganden zijn liganden die in staat zijn actief deel te nemen aan het maken/breken van bindingen met behulp van redox-reacties, bijvoorbeeld door te fungeren als electron-donor of -acceptor. Wanneer dit type liganden gebonden is aan metaalcentra, biedt dit de mogelijkheid om ongebruikelijke redox-reacties te bewerkstelligen doordat instabiele oxidatietoestanden van het metaalcentrum worden vermeden. Dit geeft aanleiding tot nieuwe (katalytische) reactiviteit die met conventionele liganden niet toegankelijk is. De ontwikkeling van nieuwe liganden met redox-actieve eigenschappen kan zich verheugen op toenemende belangstelling, vanwege de kansen die dit biedt om edele metalen te vervangen door goedkopere en minder schaarse elementen voor toepassing in (homogene) katalyse. Het ligand systeem dat in dit proefschrift wordt beschreven is gebaseerd op de NNCNN-structuur van formazan verbindingen, wat grote overeenkomsten heeft met de veelvuldig toegepaste -diketimine liganden. In dit proefschrift wordt de synthese van nieuwe coördinatiecomplexen van dit ligand met de elementen Zn en B beschreven, alsmede de chemische en fysische eigenschappen van deze verbindingen. De resultaten laten zien dat deze formazanaat anionen unieke redox-eigenschappen hebben en leggen een brede basis voor de verdere ontwikkeling van de (katalytische) reactiviteit van organometaalcomplexen met dit type ligand.

In Hoofdstuk 1 wordt ter introductie een aantal voorbeelden beschreven van redox-actieve liganden, zowel systemen die voorkomen in de natuur (enzymen) alsook voorbeelden van dergelijke liganden die in laboratoria ontwikkeld zijn. De enzymen cytochroom P450 en catechol dioxygenase bieden inspiratie voor de ontwikkeling van nieuwe type liganden gebaseerd op porphyrines en catecholaten. Een aantal voorbeelden van niet-natuurlijke ligand systemen dient ter illustratie van de mogelijkheden die dit type liganden biedt voor (tijdelijke) opslag van redox-equivalenten, die vervolgens kunnen worden gebruikt voor het bewerkstelligen van een chemische omzetting. Bovendien worden voorbeelden van katalytische cycli besproken waaruit blijkt dat metaalcomplexen met redox-actieve liganden in staat zijn multi-electronen reacties te katalyseren die zonder deze liganden niet mogelijk zijn.

189

Hoofdstuk 2 beschrijft de ontwikkeling van synthesemethoden om de formazan liganden (R1NHNC(R3)NNR5, 1) te maken met verschillende substitutiepatronen. De sterische en elektronische eigenschappen van dit type liganden is te variëren door de groepen R1, R3 en R5 te veranderen, uitgaande van startmaterialen met verschillende substituenten op die posities. Hoewel het mogelijk is om de gewenste formazan producten te synthetiseren, is een algemeen protocol niet voorhanden: iedere variatie vereist optimalisatie van reactiecondities en zuiveringsmethoden. Naast een serie mono-formazan liganden zijn er ook twee voorbeelden van di-formazan verbindingen gemaakt die verbonden zijn via een C6H4-brug.

Schema 1. Synthese van formazan verbindgen beschreven in Hoofdstuk 2

In Hoofdstuk 3 wordt de synthese en karakterisatie van mono(formazanaat)zink methyl

(LZMe, 4) en bis(formazanaat)zink complexen (L2Zn, 5) beschreven. De kristalstructuren en VT-NMR data van verbindingen 5 laten zien dat de formazanaat liganden aanleiding geven tot flexibele coördinatiechemie, doordat de aanwezigheid van meerder N-atomen in de NNCNN backbone van deze liganden isomerisatie tussen 5- en 6-ring chelaat complexen mogelijk

190 maakt. In de cyclische voltammetrie van de complexen 5 is te zien dat reversibele, ligand-gebaseerde redox-reacties plaatsvinden (in sommige gevallen zelfs 4 afzonderlijke redox-koppels), wat suggereert dat ieder formazanaat ligand gereduceerd kan worden van het anion ([1]-) tot het overeenkomstige radicaal dianion ([1]-2) en trianion ([1]-3). De kristalstructuren, EPR spectra en DFT berekeningen van de gereduceerde producten laten zien dat formazanaten een nieuwe klasse van redox-actieve liganden zijn.

R3 R3 R5 R1 2 eq ZnMe N N - ZnMe N N N N N N 2 2 3 3 2 2 NN R Zn R NH N 1 R5 NN 1 5 R Zn NN R R 1 5 Me R R

LH (1)LZnMe (4) L2Zn (5) Schema 2. Synthese van de complexen 4 and 5

Figuur 1. CV data van verbindingen 5a (dichte lijn) en 5b (stippellijn) geven aan dat 2-electronen reductie van een formazanaat ligand mogelijk is (L3-, reductions III and IV).

Hoofdstuk 4 gaat in op een reactie waarbij een formazanaat ligand wordt overgedragen van zink in de complexen 5 naar een boor atoom, waarbij (formazanaat)boor difluoride verbindingen (6) worden gevormd. Een belangrijk intermediair in deze reactie, een 6-voudig gecoordineerd zink complex (7) waarbij een BF3 fragment is gebonden aan het ligand (i.e, 1 3 5 [R NNC(R )NNR (BF3)] is geïsoleerd en kristallografisch gekarakteriseerd. De omzetting 5→7→6 laat zien dat de flexibele coördinatiechemie van formazaat liganden kan leiden tot nieuwe mogelijkheden om substraten te binden; isomerisatie tot een 5-ring chelaat ligand maakt dat er twee extra coördinatieplaatsen beschikbaar komen.

191

5 F F R 5 F F 5 R1 R B R N B 5 N N BF3 F F N R -ZnF N N N N 2 3 F R3 Zn R3 N Zn N R B R3 3 F NN N N R NN 5 N N 1 R N N R1 R 1 R1 R

6 5 7

Schema 3. Reactie waarbij een formazanaat ligand wordt overgedragen van zink naar boor

Hoofdstuk 5 laat zien dat de hiervoor beschreven (formazanaat)boor difluoride complexen

(LBF2, 6) unieke ligand-gebaseerde redox-reacties hebben. Het cyclisch voltammogram van 6 bevat twee (quasi)reversibele processen, waaruit blijkt dat een formazanaat ligand in staat is om twee elektronen op te nemen waarbij het radicaal dianion en trianion wordt gevormd ([1]-2 en [1]-3), vergelijkbaar met de zink complexen beschreven in hoofdstuk 3. Bovendien hebben de verbindingen 6 intense absorpties in het zichtbare gebied van het elektromagnetisch spectrum, en laten ze emissie zien waarbij er sprake is van een grote Stokes shift. Deze eigenschappen maken deze verbindingen potentieel interessant als fluorescerende kleurstoffen. Ter vergelijking met de difluoride complexen (6) zijn de analoge

(formazanaat)boor diphenyl (LBPh2, 9) en dihydride (LBH2, 10) verbindingen gemaakt. De optische en elektochemische eigenschappen van 9 en 10 zijn vergelijkbaar met die van 6: ook deze complexen hebben intense absorptiebanden, grote Stokes shifts en ligand-gebaseerde redox-chemie.

Figuur 2. UV-Vis absorptie spectra (links) and genormaliseerde emissie spectra (rechts) van 6c, 6f, en 6g. Data zijn gemeten als 10-5 M oplossing in droge THF. De emmissie spectra zijn opgenomen door excitatie met een 473 nm laser.

192

In Hoofdstuk 6 wordt in detail gekeken naar de producten die verkregen worden uit de preparatieve 2-electronen reductie van [PhNNC(p-tolyl)NNPh]BF2 (6a) met 2 equivalenten Na/Hg. De karaterisatie van produkten die uit deze reactiemengels verkregen kan worden laat zien dat er een serie BN-heterocyclische produkten (11-14) wordt gevormd die bestaan uit combinaties van een (formazanaat)B(I) fragment (X) en een imidoborane (fragment Y, Schema 4). Het intacte formazanaat fragment in 12 en 13 geeft aanleiding tot (quasi)reversibele redox-chemie, zoals in dit proefschrift beschreven voor andere formazanaat complexen. Het feit dat een bijzonder reactief (formazanaat)B fragment wordt ingebouwd in de producten suggereert dat het ligand in staat is het laag-valente B(I) intermediair te stabiliseren. DFT berekeningen laten zien dat dit mogelijk is doordat een ongewone (triplet) elektronische structuur toegankelijk is waarbij een ongepaard elektron zich in de * orbitaal van het ligand bevindt. Na oxidatie reactie van 11 en 13 met XeF2 wordt het difluoride startmateriaal weer gevormd, waaruit blijkt dat het (formazanaat)B fragment in deze heterocylische verbindingen reageert als B(I).

Schema 4. Vorming van BN-heterocyclische producten via 2-elektronen reductie van een (formazanaat)boor difluoride complex

In Hoofdstuk 7 wordt gekeken naar de reactiveit van de (formazanaat)boor dihydride verbindingen (10). Thermolyse van deze complexen leidt tot transfer van de B-hydrides naar het formazanaat ligand. Twee intermediaren die gevormd worden in deze reacties (verbindingen 17 en 21) zijn geidentificeerd met behulp van NMR spectroscopie. In combinatie met DFT berekeningen wordt duidelijk dat 17 het beste te beschrijven is als een boor(III) hydride met een formazan ligand dat gereduceerd is met 2 elektronen ([1a]-2BIIIH). Variatie in het substitutie patroon van het ligand leidt tot verschillen in reactiviteit/selectiviteit voor deze hydride transfer reacties.

193

Schema 5. Voorgesteld mechanisme (gekarateriseerde intermediairen) voor de hydride transfer reacties die het gevolg zijn van thermolyse van verbinding 10c

194

Perspective In this thesis, synthetic procedures of formazan ligand and corresponding zinc and boron complexes have been developed. The redox-active nature of formazanate ligand was established by electrochemical and chemical methods. The results presented in this thesis provide a strong basis for future research using formazanate ligands. In the following sections, the perspective of my research and potential research directions that make use of the unique properties of formazanate ligands are described.

The first challenge encountered in this work is the formazan synthesis. Even though several synthetic procedures for preparing formazan ligands have been reported, the outcome of synthetic methods for new formazan derivatives is still hard to predict. The results in Chapter 2 show that using different starting materials the reaction conditions and purification procedures to obtain the final product can be different. For example, in order to introduce a mesityl substituent on the formazan N atom (1c), a solvent mixture of acetone/water is used. The reason for using acetone/water mixture is to find a balance between dissolving all necessary chemicals (NaOH, hydrazone, and diazonium salt) and the stability of diazonium salt. In addition to the unpredictable reaction conditions, another general challenge of formazan synthesis is to introduce sterically demanding substituents to the ligand backbone, which is a frequently used strategy to control reactivity and selectivity of catalysts. Unfortunately, for formazan ligands, bulky substituents, such as 2,6-di-iso-propylphenyl groups, are not easy to introduce due to the low stability of the corresponding diazonium salt. Therefore, it is necessary to develop new synthetic procedures to access formazan ligands having bulky substituents.

In Chapter 3, bis(formazanate)zinc complexes (compound 5) were synthesized and fully characterized. The synthetic method to obtain the heteroleptic bis(formazanate)zinc complex (5aj) has a great potential to combine two different ligand system into one metal complex, in which the redox-active formazanate ligand can be used to store redox equivalents and the second ligand system can be used to control the reactivity and selectivity of the metal complex. At the end of Chapter 3, we discovered a potential method to synthesize a triazole from the reaction of the heteroleptic bis(formazanate)zinc complex (5aj) with BF3. However,

195 more work is needed to understand the mechanism, establish substrate scope and to develop the observation into a proper synthetic procedure for triazole synthesis starting from 3-cyanoformazan ligands.

The challenge encountered in Chapter 6 is to isolate and characterize all the BN-heterocycles that are obtained upon 2-electron reduction of the (formazanate)boron difluoride precursor (compound 6a). The formation of a series of BN-heterocyclic products suggests that formazanate ligands are capable of stabilizing reactive B(I) centers. Unfortunately, the initial attempts of trapping the (formazanate)B(I) species were not successful. However, it is still worth trying to isolate the (formazanate)B(I) species and compare its reactivity with known B(I) species in literature. In addition, the results of this chapter indicate that redox-active ligands is not simply being reservoirs for redox equivalents; the ligand-based reduction can lead to interesting structural rearrangements.

In Chapter 7, the thermally induced intramolecular hydride transfer of the (formazanate)boron dihydride complexes results in the formation of aminoborane products, which are a bioisosteric replacement of imidazoles and pyrazoles. The potential challenges of studying aminoborane products are the synthetic procedures of both (formazanate)boron dihydride complexes (compounds 10) and aminoborane products (compounds 18). The yield of the current synthetic procedure of (formazanate)boron dihydride complexes are very low (10-30%) and the purifications of the aminoborane products are not established.

Potential research directions of formazanate complexes: Ligand-Based Oxidation In this thesis, the ligand-based reduction of formazanate ligand has been established; but unlike the research of the -diketiminate ligand system, which shows both ligand-based reduction and oxidation reactions, the ligand-based oxidation reactions of formazanate complexes are still not clear. The ligand-based oxidation reaction will make formazanate ligands a promising candidate for participating 2-electron process by changing its oxidation states between the 1-electron reduction product (L-2) and the 1-electron oxidation product (L0). The advantages of this process are that the 1-electron reduction product is easily accessible, and there is no need to reach the reactive (or unstable) 2-electron reduction product. Therefore,

196 there is clear motivation to study the formazanate-based oxidation reactions.

The cyclic voltammogram of both zinc and boron complexes bearing formazanate ligands do not show any evidences of oxidation reactions under measuring conditions. By synthesizing metal complexes having electron-rich metal centers and easily access d-electrons, both of which can stabilize oxidation products, the ligand-based oxidation of formazanate complexes is possible to reach. The preliminary results from Dr. Barbara Milani prove this concept. The cyclic voltammogram of a bis(formazanate)palladium complex (L2Pd) shows a clear evidence of a (quasi)reversible ligand-based oxidation. The next steps in this direction are to synthesize and characterize the oxidation products as well as to develop suitable catalytic reaction based on this system.

Oxidative Addition of (Formazanate)Al(I) Complexes The low-valent group 13 complexes, especially Al(I), bearing formazanate is an attractive complex to made. Based on the results of the boron chemistry presented in Chapter 6, the formazanate ligands are capable of stabilizing reactive B(I) center. We can expect that the (formazanate)Al(I) complexes will be more stable than the B(I) complexes due to the inert pair effect. The higher stability of (formaznate)Al(I) complexes makes it as a potential candidate for the study of oxidative addition reactions.

Base-Metal Complexes Replacing precious metals by base metals from catalysts is one of the ultimate goals for researches of redox-active ligands. Therefore, preparing base metal complexes bearing formazanate ligands is a research direction worthy studying. From my point of view, Fe and Co complexes are very attractive targets for the future research of the formazanate ligand. The successful researches of Fe and Co complexes bearing tridentate redox-active ligands have been reported by Paul and co-workers (please see Chapter 1), but the bidentate analogue of Fe and Co complexes are still limited. The desired Fe and Co complexes bearing formazanate ligands will give us not only comparisons with tridentate ligand system but also more insight about the role of redox-active ligand in catalytic reactions.

197

198

Acknowledgements

I’m very happy that I had a great PhD life in Groningen, which is a lovely city and full of kind people. Everyone says that doing PhD is not an easy journey, I totally agree with that, especially when you do the PhD in a foreign country. But, I’m very lucky. There are many people beside me and they are willing to go through this unforgettable journey with me.

I would like to start with the person who contributed the most to the success of the research described in this thesis: Dr. Edwin Otten, my promoter. Four years ago, I sent an email to you to apply the PhD position in your startup group. After a Skype interview, you decided to take a big risk for accepting me as your first PhD student, and I decided to take the risk, too. I appreciate the opportunity of joining your group to do my PhD research. I love my research project and the atmosphere in the lab. In the past four years, you gave me a lot of freedom of research directions and working pace. I could try almost everything I wanted to try, and I could decide when to take a vacation freely. I also want to thank you for giving me chances to go to many conferences, summer schools, and workshops. I learned a lot from you, such as experimental skills, tastes of science and working attitude. Working with you is more like working with a postdoc or a friend than with an advisor, and I enjoyed it very much. I’m very happy having you as my promoter during my PhD. I hope you feel the same way as I do. Look the results we have created in the past four years! The risk we took four years ago now turns into high profits.

I also want to thank Dr. Wesley Brown. Since the very early stage of my research, you had given me a lot of help and suggestions on spectroscopy. In the last year of my PhD program, you became one of my promoters. I appreciate for having you as my promoter in the last stage of my PhD. You gave me a lot of useful advice about my research project and the thesis.

Of course, I want to thank my reading committee, Prof. Franc Meyer, Prof. Gerard Roelfes and Prof. Hans de Vries, for willing to read through my thesis, provide valuable suggestions for the improvement, and most importantly, approve it for defense.

Seb, I’m very happy having you be my side during my PhD. You are my first and the best friend in the lab. In the past few years, we worked not only in the same lab but also in the same office. I still remember my first day in the lab; you gave me a worm welcome and

199 helped me to start my work. In the very first months after I joined lab, we almost could not understand each other. But after sharing the office for months, we started to talk about our research and interests such as sports, movies, cartoons, and most importantly, computer games. We had so much fun inside and outside the lab. I still remember that Friday afternoon when we played angry bird and candy crush in the office and then were caught by Edwin. Now, I want to thank you for all of the great memory we had in the past four years. It is very sad that you cannot join my promotion, and I cannot join yours either.

Raquel, you joined the lab only six months after me. We worked together on formazanate chemistry in the past few years. It is great to have you in the group. You have better working style than me. You always plan your experiments ahead and know all the details in the lab. Maybe you think those are just small and unimportant things in the lab, but from my point of view, you have great contributions to the lab. Not surprised, you do a great job in your projects; especially the iron chemistry, which gives me a lot of fun in my last year. I also want to thank you for taking good care of gloveboxes in the past few years. Without gloveboxes in good conditions, all the chemistry I presented in the thesis would not happen. Of course, you played an important role in all of the lab activities, for example, setting up a new fume hood. In the end, I also want to thank you for being my paranymph for the thesis defense.

Ranaijt, I’m very happy working with you in the last few months of my PhD. You are very enthusiastic about research and willing to learn everything in the lab. You are doing a good job on the boron chemistry. Keep going like this, I wish you a great PhD journey in this wonderful group.

Besides the PhD students mentioned above, the master and bachelor students in the group are also very important in my PhD journey. Peter, Marco, Douwe, Francesca, Onno, Folkert, Linda, and Daan, each of you has some overlapping with my PhD journey, and all of you are unforgettable to me. Thank you all for creating so many funny moments and interesting events in the lab. I want to thank Linda as my paranymph of the defense. In addition, I want to thank our technician Oetze for taking care all the equipment in the lab.

I want to thank the people from Harder’s group, Julia, Johanne, Harmen, and Cedric, for participating the first year of my PhD. Watching football, having BBQ, wine testing, Christmas market, and wedding, all of them are special memories and experience for me.

200

I also want to thank the people from Wesley’s group, Shaghayegh, Davide, Tjalling, Luuk, Francesco, Sandeep, Pattama and Duenpen. Because of you, the MOLAN group becomes stronger and we have more chances to discuss our research and share our experience.

Thanks my Taiwanese friends in lovely Groningen, Kuang-Yen, Wayne, Belinda, Judy, Larry, Bin-Yan, Wen-Hao, and Chewing. Because of you, the life in Groningen is not lonely anymore. Of course, I have Chinese friends here, such as Jia Jia, Yanxi, and Juan. Kuang-Yen, thank you very much for helping me a lot of things at the beginning of this journey. Larry, we started our PhD almost at the same time and it was good to have you as my housemate for few months. We also had a lot of conversation in the past four years. Wen-Hao, thank you for helping me move to my current place and many helps in my daily life. Bin-Yan, thank you for showing me the bird-house in the university, that is really cool. Chewing, it was very nice to meet you in my last year of PhD. You are a very nice girl, and you add so much fun to the Taiwanese group. Judy, thank you for help cutting my hair many times and sharing delicious Taiwanese foods with me. Wayne and Belinda, thank both of you for giving me many suggestions on both the career and personal life. Yanxi, thank you for spending so much time with me at the corridor and on the way back to home. Juan, thank you for arranging the first MOLAN activity together with me.

Actually, it is impossible to thank all the people helping me during this journey, and I need to finish this section at some point. But before I go to the end, I want to give special thanks to my family. Even though they were in Taiwan, could not be here with me in the past four years, but without their support I can’t finish this journey. Especially my sister and her husband, thank both of you for taking care of our parents, arranging a family trip to Japan last year, bring our mother here for my promotion and sending Taiwanese foods to me. The last person I want to thank here is my girlfriend, Yu-Shih. You gave me a lot of support in the past four years and helped me to face challenges during this journey. Also, thank you for arranging so many unforgettable trips during our vacation in the past four years. Now, I am going to finish my PhD, and it’s time for us to start our new adventure together.

201

誌謝

終於到了動手寫這部分的時候了，這表示四年的博班生活準備要告一個段落，也即將開始下一段的旅程。博班的過程中有感謝的人很多，如果沒有這些大家的幫忙，這段路不會如此順利與開心。

最首先要感謝的就是我的指導教授 Dr. Edwin Otten，如果沒有 Edwin 四年前的把賭注押在我身上（畢竟對於一個剛開始的實驗室，要收一個遠在 9000 公里外的學生可是很大的賭注），這一切都不會發生；從結果看來，當初的賭注對 Edwin 跟我來說都是個好結局，我們一起完成了許多讓人興奮的成果。非常謝謝 Edwin 在不管是工作上或是生活上給我的許多幫助。實驗上，Edwin 給了我很大的空間，讓我可以嘗試許多有趣的實驗；在工作節奏上，Edwin 也給了我很大的自由，讓我能隨心所欲地安排假期。也很感謝 Edwin 給我在實驗上的許多幫助，給我很多學習新技術的機會，當然還有幫忙修改我的英文寫作。和 Edwin 一起工作其實不怎麼像是老師跟學生的感覺，有點像是跟實驗室的博士後一起工作，更多的是像朋友的感覺。我真的非常開心與感激和 Edwin 一起工作的四年時光，另外也要謝謝 Edwin 兩個可愛的小孩，他們的參與讓實驗室的聚餐多了許多樂趣。

再來要感謝的就是我的法國同事 Seb，從我開始工作的第一天一直到他結束博班的工作我們都一起在同一個辦公室。兩個人從一開始的雞同鴨講慢慢演變到無話不談，也一起經歷過許多難忘的時光，一起出去開會，一起做實驗，一起看足球，一起在實驗室內外玩耍（還有在辦公室偷打小遊戲）。這許許多多都是珍貴難忘的回憶，也讓我的博班生活增添許多樂趣。

我也想感謝在我加入實驗室半年後從西班牙來的 Raquel，相對於我的粗枝大葉，Raquel 則細心許多，對於實驗室大大小小的事情都盡心盡力。和 Seb 不同的是，Raquel 跟我的研究方向比較接近，所以兩個人有較多的機會討論研究，也很感謝 Raquel 在研究上給我的幫助，以及在後期一起擔負管理實驗室的責任。當然，不只在工作上，Raquel 在工作以外的活動也扮演非常重要的角色。

除了實驗室的人以外，最需要感謝的就是我的家人。雖然這不是第一次長時間不在家（上

202

次是當兵），但不在家裡四年並加上相隔 9000 公里的距離，感覺跟當兵非常不一樣，對於家裡的種種變化都無法親身參與。很謝謝我的家人讓我能沒有後顧之憂的專注在這邊的工作，特別是要謝謝姊姊還有姐夫幫忙照顧爸媽，寄補給品給我還有接送我往返機場，更在去年安排了去日本的家族旅遊。最後要感謝的就是我女朋友佑蒔，感謝佑蒔當初支持我來荷蘭唸書以至於兩個人要分隔兩地四年之久，也感謝佑蒔規劃了許多假日的旅遊，讓我們在少少的見面日子中充實快樂，荷蘭的冒險要告個段落了，接下來就是一起在美國冒險的日子，一起努力吧。

203