Diaryl Triazenides as Ligands for the and Related Studies

A thesis in partial fulfilment of the requirements for the degree of

Doctor of Philosophy by

Matthew Robert Gyton

Supervisor: Assoc. Prof. Marcus L. Cole

School of Chemistry

Faculty of Science

27th February 2015 Originality Statement

'I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception or in style, presentation and linguistic expression is acknowledged.'

Signed ··· ~ ····· · · ··· ········ · ·············

Date ... /<-If .. /?!/-£ ...... Acknowledgements

This thesis would no doubt be far less succesful without the support of many. My heartfelt gratitude is expressed to the following who have helped me to get this far.

To my supervisor Marcus Cole, I am grateful for the opportunity to work in this extremely challenging field, a pie, in which I intend to always have a few fingers. I am also thankful for the chance to learn the delicate art of X-ray crystallography, upon which this project was wholly reliant.

Secondly I would like to thank my co-supervisor Jason Harper for his invaluable mechanistic insights when my arrow pushing was found wanting, supervision when called upon at times to step-up and help with the initial stages of drafting this thesis.

I would like to thank the members of the Cole group both past and present for putting up with my, at times, cantankerous attitude and for keeping me company throughout this PhD. I will always have fond memories of my time with the group, especially the OZOM conference dinners!

My thanks must also be extended to the technical staff, who have been so helpful throughout my candidature. First and foremost to Dr Mohan Bhadbhade for his invaluable technical assistance with crystallography and for demonstrating just how truly useful Blu-Tack for repairing goniometer tips! My thanks must also be extended to the staff of the UNSW NMR facility, particularly Drs Don Thomas and Douglas Lawes for their technical and theoretical assistance and all-encompassing knowledge of NMR.

I would also like to extend my thanks to my friends who have had my back through thick and thin. In no particular order I would like to thank Steve, Brad, Jo, Sinead, Alasdair, Joana, Hamish, Milena, Rob and the Matts T and P. I will always consider you true friends and I will always look back fondly at UNSW as a result.

To my family I am forever thankful for your support. To Holly and Mum particularly: I win.

Finally I would like to thank my long suffering girlfriend Sam. I am forever in your debt for your unending support throughout this whole process. I am not sure that I could have done it without you. I am also forever thankful for all the times you didn’t strangle me, even if I did deserve it! Table of Contents

Abstract ...... iv

List of Abbreviations...... v

Table of Compounds by Number ...... xi

Chapter I Introduction ...... 1

1.1 Introduction to the Lanthanides ...... 1

1.2 Oxidation States ...... 3

1.3 Lanthanide Coordination Chemistry ...... 6

1.3.1 Trivalent Lanthanide Cyclopentadienyls ...... 7 1.3.2 Divalent Lanthanide Cyclopentadienyls and Heterocyclopentadienyls ... 11 1.3.3 Cyclopentadienyls, Cyclooctatetraenyls and Pentalenyls ...... 17 1.3.4 Trivalent Lanthanide Amides ...... 18 1.3.5 Divalent Lanthanide Amides ...... 20 1.3.6 Tetravalent Cerium Amides and Oxidation Chemistry ...... 23 1.4 Purpose of this Thesis ...... 24

Chapter II - Main-Group Metal Complexes of Triazenides and a Formamidinate. 26

2.1 Introduction ...... 26

2.1.1 Main Group Compounds as Synthetic Precursors for Organolanthanide Chemistry ...... 26 2.1.2 Triazenes and Related Amidines ...... 27 2.1.3 Alkali Metal Complexes ...... 30 2.1.4 Thallium Complexes ...... 34 2.1.5 Purpose of this Chapter ...... 35 2.2 Results and Discussion ...... 36

2.2.1 Synthesis of Triazene Precursors and Triazenes ...... 36

2.2.2 Attempts at the Synthesis of a Formamidine from Dipp*NH2...... 49 2.2.3 Synthesis of Alkali Metal Triazenides ...... 67 2.3 Conclusions and Future Work ...... 81

2.3.1 Conclusions ...... 81 2.3.2 Future Work ...... 82 i

2.4 Appendix I: Supplementary Crystal Structure ...... 84

2.5 Appendix II: Crystal Data for Structures Collected for this Chapter ...... 85

Chapter III Lanthanide Triazenides: Synthesis and Reactivity ...... 88

3.1 Introduction ...... 88

3.1.1 Trivalent Lanthanide Amidinates and Guanidinates ...... 88 3.1.2 Divalent Lanthanide Amidinates and Guanidinates ...... 89 3.1.3 Lanthanide and Actinide Triazenides ...... 91 3.1.4 Purpose of this Chapter ...... 92 3.2 Results and Discussion ...... 93

3.2.1 Synthesis of Trivalent Lanthanide Triazenides ...... 93 3.2.2 Reactivity of Trivalent Lanthanide Triazenides ...... 119 3.2.3 Synthesis of Divalent Lanthanide Triazenides ...... 125 3.2.4 Attempts to Synthesise a Tetravalent Cerium Triazenide ...... 134 3.3 Conclusions and Future Work ...... 145

3.3.1 Conclusions ...... 145 3.3.2 Future Work ...... 147 3.4 Appendix III: Supplementary Crystal Structures ...... 148

3.5 Appendix IV : Crystal Data for Structures Collected for this Chapter ...... 151

Chapter IV and Calcium Arenes: Unusual Rearrangements Leading to Solvent Separated Ion Pairs ...... 155

4.1 Introduction ...... 155

4.1.1 The Chemistry of Calcium Arenes ...... 155 4.1.2 The Chemistry of Ytterbium Arenes ...... 158 4.1.3 Solvent Separated Ion Pairs in Calcium and Ytterbium Chemistry ...... 161 4.1.4 Purpose of the Chapter ...... 163 4.2 Results and Discussion ...... 163

4.2.1 Synthesis of a Calcium Fluorenide Solvent Separated Ion Pair ...... 163 4.2.2 Synthesis of a Ytterbium Fluorenide Solvent Separated Ion Pair ...... 169 4.4 Appendix V: Crystal Data for Structures Collected for this Chapter ...... 176

Chapter V Experimental ...... 177

ii

5.1 General Procedures ...... 177

5.2 Characterisation ...... 178

References ...... 214

iii

Abstract This thesis outlines the synthesis and coordination chemistry of 1,3-diaryl triazenes, in particular, their alkali metal coordination chemistry and the utility of these complexes as precursors to lanthanide triazenide complexes.

Chapter II outlines the synthesis of the extremely sterically demanding 1,3-bis(2,6- dibenzhydryl-4-methylphenyl)triazene as well as a range of organic precursors to this triazene. The unsuccessful attempts to synthesise the analogous formamidine through many and varied routes are also outlined. The synthesis of the alkali metal complexes of 1,3-bis(2,6-diisopropylphenyl)triazene are described herein.

Chapter III describes the synthesis of a range of lanthanide triazenides in both the trivalent and divalent oxidation states. The utility of the common methods of synthesis of amidolanthanides to the triazenide complexes described herein is examined. The unsuccessful attempts to synthesise a cerium(IV) triazenide are also described in this chapter. Calculations performed to numerically describe the steric congestion at lanthanide triazenides and to rationalise their reactivity relative to previously described lanthanocenes are also outlined in this chapter.

Chapter IV describes the reactivity of 2,6-dibenzhydryl-4-methyliodobenzene with calcium and the lanthanide metals. An unusual fluorenide reaction product is identified with calcium and ytterbium metal and a proposed reaction mechanism is outlined. This reaction outcome is compared to pre-existing arylcalcium and arylytterbium chemistry.

iv

List of Abbreviations Å angstrom, 1 x 10-10 m

Ae Alkaline earth

An Actinide

9-BBN 9-Borabicyclo(3.3.1)nonane

[BnEt3N]Cl Benzyltriethylammonium chloride nBu n-Butyl nBuLi n-Butyllithium tBu tert-Butyl tBuLi tert-Butyllithium ca. circa, Latin for ‘about’

CIP Contact Ion Pair cf. confer, Latin for ‘compare’ cm-1 Wavenumber, unit of frequency (= ν/c) cm3 Cubic centimetre

Cp cyclopentadienyl ligand

Cpʺ 1,3-bis(trimethylsilyl)cyclopentadienyl ligand

Cp* 1,2,3,4,5-Pentamethylcyclopentadienyl ligand

COTʺ 1,4-Bis(trimethylsilyl)cyclooctatetraenyl ligand

Cy Cyclohexyl d Doublet dec. Decomposition temperature

v

δ NMR chemical shift in ppm

Δ Indicates heating

DFT Density Functional Theory

Diep 2,6-Diethylphenyl

Ditop* 2,6-Di(para-tert-butylphenyl)phenyl

Dipp 2,6-Diisopropylphenyl

DippForm 2,6-Diisopropylphenylformamidinate

Dipp*nBu 2,6-Dibenzhydryl-1-butyl-4-methylbenzene

Dipp*H 2,6-Dibenzhydryl-4-methylbenzene

Dipp*Li 2,6-Dibenzhydryl-1-lithio-4-methylbenzene

Dipp*NC (2,6-Dibenzhydryl-4-methylphenyl)isocyanide

Dipp*N3 2,6-Dibenzhydryl-4-methylazidobenzene

Dipp*N(CHO)(Et) 2,6-Dibenzhydryl-4-methylphenyl-N-ethylformamide

Dipp*NCH(OEt) Ethyl-N-(2,6-dibenzhydryl-4-methylphenyl)formimidate

Dipp*NCH(OMe) Methyl-N-(2,6-dibenzhydryl-4-methylphenyl)formimidate

Dipp*NCNDipp* Bis(2,6-dibenzhydryl-4-methylphenyl)carbodiimide

Dipp*NH(CHO) (2,6-Dibenzhydryl-4-methylphenyl)formamide

DippNH2 2,6-Diisopropylaniline

(Dipp*NH)2CO Bis(2,6-dibenzhydryl-4-methylphenyl)urea

Dipp2N3H 1,3-Bis(2,6-diisopropylphenyl)triazene

Dipp*2N3H 1,3-Bis(2,6-dibenzhydryl-4-methylphenyl)triazene

DMAP 4-Dimethylaminopyridine dme 1,2-Dimethoxyethane ligand vi

DME 1,2-Dimethoxyethane, free

Dmp 2,6-dimesitylphenyl e- Electron e.g. exempli gratia, Latin for ‘for example’

Et Ethyl

η Hapticity of a ligand et al. et alii, Latin for ‘and others’

Et2O Diethyl ether

EtOAc Ethyl acetate

EtOH Ethanol

Fluor* 4-Benzhydryl-2-methyl-9-phenylfluorenide fwhm Full Width at Half Maximum g Gram

HC(OEt)3 Triethylorthoformate

HC(OMe)3 Trimethylorthoformate

Hept nHeptyl

Hex nHexyl

HOAc Acetic acid i.e. id est, understood as ‘that is,’ Latin for ‘it is’ iPr Isopropyl

IR Infrared

κ Denotes binding mode for polydentate ligands

K Kelvin vii

KOtBu Potassium tert-butoxide lit. Literature value

Ln Lanthanide m Medium or Multiplet

Me Methyl

MeCN Acetonitrile

MeOH Methanol

Mes Mesityl, 2,4,6-trimethylphenyl

Mes* 2,4,6-tri(tert-butyl)phenyl mg Milligram mins Minutes

MO Molecular Orbital mol Moles mmol Millimoles m.p. Melting point (°C)

MRI Magnetic Resonance Imaging

µ Bridging ligand or micro

μmol Micromoles

NCy2 Dicyclohexylamide

NaOtBu Sodium tert-butoxide

NMR Nuclear Magnetic Resonance

Oct nOctyl

OTf Trifluoromethanesulfonate viii

O=PPh3 Triphenylphosphine oxide

[ox] Denotes and oxidant

Pip Piperidinyl pKa Acid dissociation constant pm Picometre, 1 x 10-12 m

PMDTA N,N,N′,N′,N′′-pentamethyldiethylenetriamine

Ph Phenyl

PhH Benzene

PhMe Toluene

PPh3 Triphenylphosphine ppm Parts per million py Pyridyl or pyridine r.t. Room temperature s Singlet or strong sh Sharp

SiMe3 Trimethylsilyl

SIR Sterically Induced Reduction solv Denotes a coordinated Lewis base that is also the reaction solvent

SSIP Solvent Separated Ion Pair t Triplet thf Tetrahydrofuran ligand

THF Tetrahydrofuran, free or solvent

TLC Thin Layer Chromatography ix

Tripp 2,4,6-Triisopropylphenyl

Trityl Triphenylmethyl

Tol Methylphenyl

UHP Ultra High Purity

ν Frequency in Hz

V Volt via Understood as ‘by way of’, Latin for ‘road’ vide infra Latin for ‘see below’ vide supra Latin for ‘see above’ w Weak

XRD X-Ray Diffraction

Xyl 2,6-Dimethylphenyl

x

Table of Compounds by Number

'Bu 'Bu~'Bu thf-Tm 'B~'Bu 'Bu 5

Ln=Dy6,Nd7 SiMe3 Me,Si

Me3Si4 ~ ~SiMe, ~Ln----"8J----Ln';/)]

M g·~SiM&a1 ~ 9a Me3Si SiMea

Ln = La9, Ce 10

11

12 13 14 15 16

~ ":<~ Dipp/ N '-::;;/ ~ K--tht thf- , o· I~/ IPP r,(:l!\}7\, M =Li22, Na23, K24 19 20 21

thf,Li-N lhr' ~ 25 At= Mes 26, Tripp 27 28 29

xi

30 31 32 33 34

p-N02Ph,N-::--N-':-N_...p-N02Ph Ar,N-:'-N-'::-N"Ar +1/f ~/f 1 I/~ l/1 N-- ,-N Ar/N-:_,N-;.-N,Ar p-N02Ph./ 'N__... 'p-N02Ph 36

Ph PX[hp~

N~N ""' I H tiYNH2 Ph Ph 41 Ph 42

Nex Ill R' / R' I " "'- " / / R' R1 Na R1 R' -~- -~- 1 2 -~- R =R ='Pr'6 R1 =R 2 =Me47 43 45 44 R1 =Me, R2= tau 48

Ph I Ph Ph Ph R' R' R' R' / / R1 =R2 =Me51 I I 1 2 Ph Ph "'- "'- R =R =H52 Ph~ Ph~ R1 =R2=/pr53 R' R' R1 ='Pr, R2 = H Sot.

49 50

55 56 57 56 Yll:~ sf:..~ PX(o Ph Ph 56 60 61 62 Yi~Ph _.rN~O/Et

Ph Ph 63 65

xii

c 'B u :sr.h-:/"' Ph PPhPh II :::::-.... I -;::.C-;:..N ~ N "'I Ph 'Bu Ph Ph "~" Ph 67 68 •• c II c c N II N ~

70 71 72 73

0 Dipp,N-;_N,;.N..--Dipp

" MI M=Ne75 ~~" K76 74

Dipp,N-;,N,:-N-"Dipp 78 " I ~~N,:~ /M'dme dme M=Na79 PhT'v'~Ph K80 Ph Ph 81

Ph Ph R PX[.hP~ I N~N ~ H Ph Ph Ph Ln =La 85, Ce 86, Nd 87, Ln = Nd 93, Tb 92 R=Me83,Ph8' Eu 88, Y 89, Er 90, Lu 91 R = lpr

R1 = Cy, R2 = 'Pr, Ln = Y 93, Lu 94 Ln = Sc 99, y 100, Lu 101 n = 1 105 R1 = Cy, R2 =Me, L.n = Nd 95, Y 98, Lu 97 Ln =La 102, Nd 103, Gd 104 n = 2 R1 =Ph, R2 = Me, Ln =Y98 Dipp Ph ' Dipp, ~ ,.._.SiMe3 FN N / ~ N Dipp-.N~N-Dipp Dipp-N...... _LI ,,,,thf n thf-Ln-thf"/ "I Dipp-N".. ...- I 'tht / '\ F-Ln-thf N ~ ~ N ~N Me Si / '( 'Dipp I " 3 Dlpp-N'v/N-olpp 'Dipp Ph 106 107 Ln = Sm 108, Eu 109 Ln = Sm 113, Eu 11•, Yb 115 Yb 110

xiii

NR:! Ph...._ ...-N, ...-Ph 'Bu N -· '-N Dlpp...._ ):...... - Dlpp N' 'N Ph...._PY~ ,?-PY...-Ph N-Ln~N \I 1: I\ :1 AN Ln N-- .-N ~/d ...-Ph I \ 'N N...- Ln-N /N'y/N, 1 I Dipp I Dipp Ph Ph ~\{:~ NR:! Ln = Er 123, Lu 124 I Ln = Sm 116, Eu 117, Yb 118 Ph R=Cy Ln =Er 125, Lu 128 Ln = Sm 119, Eu 120, Yb 121 R=;Pr

N(SiMe ) \f-o 3 2 Ph SiM&a Ad Xyk_ ( "ee~N(SiM~)2 \ I I n Ph-R=N N c, /'-' DIPP-NH/N~, ...- PPh ~l"' ,.IN Lu-N...- 2 (fJ~r;;'?•-.. Ph-P=N N / \ I I \ ' Dipp-NH N;;.N Ph SiM~ '--Ph 'oipp I 130 131 Mea 132 Ln=Lu133

Ln =Eu 135, Vb 138

141 p-Tol, N ~I'{ An/N-p-Tol I 'N ~ p-Tol ~~ ~ N I ~Tol An= Th, U 145 142 143 138

'Bu 'Bu "Bu I I neu YNI ...... _ Li / Nl ::-{ teu ~ N , [ / N- teu Li " / Li I 148

152 Ar =~ Tol 149, Xy 150 151 153 xiv

N Dipp,N~_N,:,-.N...-Dipp Dlpp~N~- -::-N-Dipp Dipp--.N~N-Dipp \/ \/ Dipp~ \ I / Dipp 1-Sm-thf 1-Sm-thf N-Ln-----N I \ I \ ~ .. /\.~ Dipp-N..:..-N-;.N-oipp Dipp-N~N-oipp ~N N• 154 155 Dip~ bipp

Ln = Ce 164, Nd 156, Sm 160, Yb 161

Dipp Dipp I I N=-::N N=-=N H Dipp-N'-...... ,, I ,,,,0,,, I /N-Dipp,, . /ca·, ~ee~ Dipp-N D N-Dipp ,, I H I lj N=-=N N=-=N Ln = Ce 165, Nd 188 I I Ln = Ce 167, Nd 168, Dipp Dlpp Sm 169, Yb 170 171 ?ipp ?ipp SiMe3 SiMe, N=-=N N=-::N Dipp-N'-....._,, I I /N-Dipp,, C-o-ce Me,Si~···''a''··.~SiMe3 Dipp-N J N-Dipp ,,/• I I'; Me,Si~ ,~,...... rSiMe3 N=-=N N=-=N I I Dlpp Dipp SiMe3 SIMe3 172 174

o-Tol o-Tol I I pN N~ o-Toi-N~ j .·'''g''·· \ ;N-o-Tol lh/Yb' ,....Ybc:::_lhf o-Toi-N \ 0 j ,N-o-Tol "'=N H Need I I o-Tol Dipp M = Ce 176, U 177 178 175

[Na(lh~sl Dipp--.N~-N-':-N-Dipp Dipp--.N-:,N,;.N-Dipp \I thf-M-thf l't.?thf /\ / \'thf Dipp--N..:..-N;,-N-oipp Dipp~N ..:._-N-;.. N-Dipp M = Yb 179, Ce 180, 185 Sm 182, Sr 184 181

[Li(dme),]

188 187 188 189 R = SiMe,, 190 R = Si'suM~, 191 lhf N(SIMa3h; lhf, ::- /lhf I Li M·' 'N(SIMe k 3 thf, t··hfj I thf .... I ""'thf M . Cl SIMea (MeaSikN ..-- "N(SiMeak 8 1 lhf 8a \ I N/ ~SiM&a l N-ee·' . M = Ce 192, U 193 1 "'N-S1Me3 194 MeaSi \ SiMe, 196

197 197

xv

M =Nd 201, Sm 202, La 203, 205 Yb 204, Pu 208

As= Sr 210, Ba 211

[Yb(dme),J Ph,JbhJJ ] Ph_..- I ' 0 thf· .J:.~ , thf I [ thftr' I ..,.thf Ph I 2 thf 218

Ln =Eu 218, Yb 217

Ph Cal Ph Ph I Ph ~Cal thk~ _ . , thf I Ph Ph y Ph thftr' I ~~ ""'thf Ph ·~~ thf

221 222 223

M = Ca 220, Yb 230

Ph I ~h c· Ph ' Ph

Cal" Ph Ph

224 M = Ca 227, Yb 228

~ thf thf ~-\~ cl-thf tht- 'ca::----... \;;' \ ' thf Ph ~~ca.-• thf I H.. I \ - thf Ph thf thf 229

231 xvi

Chapter I Introduction

1.1 Introduction to the Lanthanides

The lanthanide series comprises the elements from atomic numbers 57 through to 71. Incorporating the elements from to , they are often classified together with the remaining group 3 elements scandium and yttrium as the rare earth elements. This is a consequence of their similar chemical properties, such as a preference for the +3 oxidation state, and as such, scandium and yttrium are also frequently considered when discussing the lanthanide series.

The rare earths, in spite of their name, are relatively abundant in the earth’s crust (Table 1).1 The most common, cerium, occurs at approximately 66.5 ppm, in roughly the same concentration as copper. All of the rare earth elements occur in greater proportions than the platinum group metals and the remaining coinage metals. Even thulium, the least abundant naturally,i occurs at 0.52 ppm, which is similar to iodine.

The physical and chemical properties of the lanthanide (Ln) elements are dominated by several features, which are consistent across the series. In elemental form they are highly reducing, with E0 values (Ln3+/Ln0 -1.99 to -2.37 V) comparable to the alkaline earth metals (Ae2+/Ae0 -1.97 to -2.92). The ionic and atomic radii are distinguished by a phenomenon known as the (Table 1). The largest of these elements is the lightest of the series, lanthanum, with an ionic radius of 103.2 pm.2 As the group is traversed, the radii steadily decrease until lutetium, which has an ionic radius of 86.1 pm.1,2 This contraction is a product of the limited extension of the 4f orbitals. The 4f orbitals penetrate into the core orbitals and thus the outer electrons are pulled in by the increasing nuclear charge. There are two anomalies in the metallic radii, which occur at europium and ytterbium resulting in larger than expected radii. Europium and ytterbium have stable [Xe] 6s2 4f 7 and [Xe] 6s2 4f 14 electron 2+ - configurations and thus in elemental form, they likely exist as [Ln (e )2], causing an increase in the atomic radius.1

i Promethium exists in lower abundance but all isotopes are radioactive with short half-lives. 1

Sc Y La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Abundance/ 22 33 39 66.5 9.2 41.5 0 7.05 2.0 6.2 1.2 5.2 1.3 3.5 0.52 3.2 0.8 ppm3 Metallic 164.1 180.1 187.9 182.5 182.8 182.1 181.1 180.4 204.2 180.1 178.3 177.4 176.6 175.7 174.6 194.0 173.5 Radius/ pm3 Six Coordinate 74.5 90.0 103.2 101.0 99.0 98.3 97.0 95.8 94.7 93.8 92.3 91.2 90.1 89.0 88.0 86.8 86.1 Ionic Radius/ pm2 Redox potential/ V1 Ln+3 +3e- → Ln -2.03 -2.37 -2.37 -2.34 -2.35 -2.32 -2.29 -2.30 -1.99 -2.29 -2.30 -2.29 -2.33 -2.31 -2.31 -2.22 -2.30 Ln+3 +e- → Ln2+ -3.1 -3.2 -2.7 -2.6 -2.6 -1.55 -0.34 -3.9 -3.7 -2.5 -2.9 -3.1 -2.3 -1.05 -2.7

Ln+4 +e- → Ln3+ 1.70 Table 1 Selected physical properties of the lanthanides.

60

50

40 IE 1 30 IE 2 IE 3 20

IE 4 IonisationEnergy/ eV 10

0 La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

Figure 1 Ionisation energies for the lanthanides.1 2

1.2 Lanthanide Oxidation States

The +3 oxidation state is a dominating feature for much of the chemistry of the rare earths. This occurs as a result of the ease by which the electrons in the 6s and 5d orbitals are removed relative to those in the 4f orbitals. The fourth ionisation energy is typically very high, greater than the sum of the first three (Figure 1). This is a result of the stability of the electrons in the 4f orbitals, which are deeply embedded in the electronic core of the +3 ion.4 Therefore the lanthanide ions generally have the [Xe] 4f n configuration and favour the +3 oxidation state.

In contrast to the transition metals, the zerovalent state for molecular complexes of lanthanides is very rare. Arene and carbonyl complexes, which are common for the transition metals, are highly unusual for the lanthanides. Zerovalent lanthanide arene complexes (Figure 2) are synthesised by highly specialised metal vapour synthesis.5,6 Zerovalent homoleptic lanthanide carbonyls are extremely thermally unstable, decomposing above 20 K and have been observed only using IR spectroscopy in an argon matrix.7 It is noteworthy that coordinated CO is even readily lost from the more stable lanthanide and is only stabilised by an atmosphere of CO (Figure 2).8-10 The instability or difficulty in accessing these compounds prevents further chemistry and thus they will not be considered further.

Figure 2 Isolable lanthanide arenes and the calculated structure of a lanthanide carbonyl. 5,6,8-10

An examination of the trends in the cumulative ionisation energies shows that for europium and ytterbium the third ionisation is high in energy relative to the rest of the lanthanide series (cf. Table 1). The +2 oxidation state is therefore favourable for these elements. This can be seen as a consequence of the stability associated with the half- filled and fully filled 4f orbital in these states. The +2 oxidation state is also readily accessible for . It has a moderately high third ionisation energy and the oxidation state is stabilised by the near half-filled 4f orbital. Due to the reducing nature of these ions their chemistry is often restricted to non-aqueous and anaerobic conditions

3 as they tend to react with atmospheric oxygen and water.1 These three are conventionally described as the classical divalent lanthanides.11

Lanthanides in these oxidation states are traditionally encountered as the diiodides. They are readily synthesised from the reaction between the lanthanide element and an oxidant such as 1,2-diiodoethane or iodine in tetrahydrofuran solution (Scheme 1).12,13 Europium and ytterbium diiodide may also be synthesised from a solution of the corresponding metal and ammonium iodide in liquid .14 The stability of the classical lanthanide diiodides is best illustrated by an examination of their electrode potentials. These three have markedly smaller potentials for the Ln3+|Ln2+ redox couple (Table 1).

Scheme 1 Synthesis of lanthanide diiodides.12-14

The redox properties of low oxidation state samarium and ytterbium compounds causes them to find use in organic synthesis. Since the first report of Kagan of the use of samarium and ytterbium diiodide as one electron reducing agents,12 their popularity has rapidly grown,15-17 although less so for ytterbium due to its lower strength as a reductant.11,12,17 The combination of a strong reduction potential as well as a high degree of chemoselectivity makes them favourable reagents for organic synthesis. Samarium diiodide does not reduce carboxylic acids or esters.12,18 This can be seen in the use of the reagent for Reformatsky type coupling reactions with haloester substrates (Scheme 2a).17,18 Aldehydes are rapidly reduced relative to ketones allowing for their selective reduction.12,18,19 For example an equimolar mixture of octanal and 2-octanone treated with samarium diiodide gives almost exclusively 1-octanol (Scheme 2b). Samarium diiodide is also used as a reagent for the Barbier reaction.15,17 Under samarium Barbier conditions an alkyl halide is reductively coupled to a carbonyl carbon atom (Scheme 2c). This is a powerful complement to organomagnesium and organolithium reagents as it is often successful where these reagents are not and is selective for aldehydes and ketones over esters and amides.20

4

Scheme 2 Some reactions involving samarium diiodide.12,15,17-19

In recent years the divalent oxidation state has expanded beyond the classical divalent lanthanides of samarium, europium and ytterbium.11,21 In terms of redox potentials the next most accessible Ln2+ ions are thulium, dysprosium and neodymium (cf. Table 1).1 Tm2+ (Tm3+|Tm2+ E0 -2.3 V) is the only example of this triad that can be made using solution phase chemistry (Scheme 3). Thulium triiodide can be reduced to thulium diiodide in either tetrahydrofuran22 or 1,2-dimethoxyethane23 with a vast excess of thulium metal as the reducing agent. Neodymium diiodide and dysprosium diiodide are not accessible using solution phase methods and are synthesised in a solid phase reaction under inert conditions at high temperature between iodine and an excess of metal (Scheme 3).24-28 Extension of this solid phase method to the other lanthanides gives material of the empirical formula LnI2. From a solid state perspective these 3+ - materials are best described as Ln I2(e ) with the electron delocalised through the bulk material in a conduction band.1,4,11 They may therefore be described as trivalent materials. Lanthanocenes in the divalent state are also readily accessible by chemical reduction of trivalent precursors. This has allowed the isolation of divalent compounds for the entire lanthanide series with the exception of radioactive promethium (See Section 1.3.2).29

Scheme 3 Synthesis of thulium, dysprosium and neodymium diiodides.22-28

Examination of the electrode potentials (Table 1) for cerium reveals a low fourth ionisation energy and a favourable Ln4+|Ln3+ redox couple. Cerium(IV) is an extraordinarily common one electron oxidising agent in synthetic chemistry because of

5 this. The ceric ion is often encountered in the form of [(NH4)2Ce(NO3)6]. This reagent is capable of a large number of stoichiometric and catalytic oxidations in organic synthesis (Scheme 4).18,30-32 The oxidation of arenes to quinones is possible with 18 [(NH4)4Ce(SO4)4] (Scheme 4a). This is limited to symmetrical unsubstituted substrates as complicated mixtures can occur otherwise (vide infra). Alkyl substituents on arenes are also readily oxidised to the corresponding aldehydes for methyl groups (Scheme 4b) or ketones for benzylic methylene groups (Scheme 4c).31 Selective oxidation of secondary over primary alcohols is also possible using catalytic [(NH4)2Ce(NO3)6] in the presence of sodium bromate (Scheme 4d).1,18

Scheme 4 Some oxidation reactions using cerium(IV) compounds.

1.3 Lanthanide Coordination Chemistry

The coordination chemistry of the lanthanides has several dominating features. A large range of coordination numbers are possible, usually from 3-12.1,4 An excellent example of the desire for high coordination number is the six or seven aqua ligands usually bound to the metal centre in the aquated trihalides.4 The coordination environment about the metal is chiefly determined by steric factors.1 This occurs due to the shielded 4f orbitals playing no part in the bonding of ligands, meaning crystal field effects do not come into play.1 Bonding in the lanthanides is almost exclusively ionic, again due to the 6 limited radial extension of the valence 4f orbitals from the nucleus relative to the 6s or 5d orbitals. The lanthanides therefore prefer hard donor ligands with O- and N- donors to match their nature as hard Lewis acids.1 The aqueous chemistry of the lanthanides is dominated by carboxylate and amine donors.33 To satisfy the desire for high coordination numbers chelating ligands are often used with numerous ancillary donors with strongly electronegative functional groups. Good examples of these types of ligands for aqueous coordination chemistry are those used in gadolinium Magnetic Resonance Imaging (MRI) contrast agents (Figure 3).34,35

Figure 3 Ligands for gadolinium MRI contrast agents.34,35

1.3.1 Trivalent Lanthanide Cyclopentadienyls

Nonaqueous chemistry of the lanthanides has been, for a long time, dominated by a small number of ligand families. The leading example is the cyclopentadienyl ligand family.36,37 These anionic ligands satisfy the hard Lewis base requirement and have the distinction of being the first organolanthanides synthesised.38 The planar ring system fills three formal coordination sites and the variable hapticity makes for a flexible coordination environment about the metal centre.39 The volatility of lanthanide cyclopentadienyls is a big advantage to the synthetic chemist seeking a means of purification by sublimation.40 Perhaps most importantly, the ring may be readily functionalised by any number of different functional groups. Alkylated - cyclopentadienyls, particularly pentamethylcyclopentadienyl, (C5Me5 or Cp*) complexes are popular due to the high steric demand placed upon the metal centre and favourable solubility of the complexes in organic solvents relative to the non- functionalised ligands. Alkylation at the ring also makes for a more electron rich ligand, although the magnitude of the effect, considering the ionic nature of the bonding in lanthanide complexes, is debateable. (Poly)silylated cyclopentadienyls such as 40-44 C5H5-n(SiMe3)n are also a popular ligand family.

7

Steric demand is an important consideration for ligands in lanthanide chemistry. The tris(cyclopentadienyl)lanthanides are readily isolated from donor solvents as the monosolvent adduct, for example [LnCp3(thf)] and are essentially isostructural for the series. However when desolvated a range of coordination geometries result (Figure 4). The only monomeric complex of this series is tris(cyclopentadienyl)ytterbium, as determined by single crystal X-Ray diffraction structural determination. All other tris(cyclopentadienyl)lanthanides form oligomers bridged by cyclopentadienyl ligands in the solid state to satisfy the large coordination sphere of metal ion. In general the

Lewis base coordinated [LnCp3] monomers and the base free oligomers display low reactivity.1,45 As a contrast the tris(pentamethylcyclopentadienyl)lanthanides are all monomeric, solvent free complexes that exhibit high reactivity (vide infra).

1 Figure 4 Some structural motifs displayed by [LnCp3] complexes.

Monomeric, substituted lanthanide cyclopentadienyls have many practical applications in synthetic chemistry, some with direct relevance to industrial processes. In particular lanthanocene catalysts display high reactivity for the transformations of unsaturated carbon-carbon bonds. Some of the reactions catalysed by these complexes include polymerisation,46,47 hydrogenation (Scheme 5a),48 hydroboration (Scheme 5b),49,50 (Scheme 5c)49-51 and hydrosilylation (Scheme 5d).49,50,52

8

Scheme 5 Lanthanocene catalysed hydrogenation (a), hydroboration (b), hydroamination (c) and hydrosilylation (d).46-52

Lanthanocene catalysts (Figure 5) display high reactivity for all these reactions. The important structural features of these catalysts are a monomeric, coordinatively unsaturated metal centre. The more unsaturated the metal centre, the more free coordination sites available for substrate binding.53 Catalytically active lanthanocene hydrides may be prepared by hydrogenolysis of sterically bulky alkyl ligands. This facilitates the approach of the alkene or substrate, which generally inserts into the metal-hydride bond.50,53 Each of these reactions occurs via sigma bond metathesis, a mechanistic step that is principally the domain of d 0 and d 0 f n metals, rather than the oxidative-addition reductive-elimination cycles familiar in mid to late transition metal chemistry.50,54

Figure 5 Some common lanthanocene pre-catalysts for transformations of unsaturated bonds.50

The stoichiometric reactions of sterically saturated lanthanocenes with reducible organic substrates show another unique mechanism of reactivity. The best known example is 55-57 Sterically Induced Reduction (SIR) at [LnCp*3] complexes. This reaction mechanism utilises a trivalent oxidation state, thus any substrate reduction cannot derive from oxidation of the metal. (cf. Figure 1). The high degree of steric constraint at the metal forces the expulsion and oxidative coupling of two Cp* ligands via what is

9 presumed to be a radical process. This process releases one electron per Cp*, which are transferred to the organic substrate. Two equivalents of metal complex is therefore - necessary to couple one equivalent of (Cp*)2 per 2e reduced organic substrate (Scheme 6).

Scheme 6 The reductive dimerisation of Cp* in SIR.

This reductive reactivity exists throughout the series of [LnCp*3] complexes but is stronger for the smaller, more sterically congested lanthanides.57-59 This is illustrated with reference to the chemistry of tris(pentamethylcyclopentadienyl)samarium(III) 1,60 which displays reductive chemistry that often directly parallels the reductive chemistry of its divalent congener, bis(pentamethylcyclopentadienyl)samarium(II) 2 (Scheme 7).

Scheme 7 SIR at tris(pentamethylcyclopentadienyl)samarium(III) 1.55,57

Species accessed by reduction of tris(polyalkylcyclopentadienyl)lanthanides or bis(polyalkylcyclopentadienyl)lanthanide tetraphenylborates with potassium graphite have been shown to display similar reductive chemistry to that of SIR. When potassium reduction is conducted under a atmosphere a formal two-electron reduction of dinitrogen occurs to afford dinitrogen bridged lanthanide(III) compounds (Scheme 8).61- 64

10

2- 61-64 Scheme 8 Isolation of bridging N2 cyclopentadienyl lanthanides. These compounds act as 2e- reductants when their electrons are transferred from the 2- N2 fragment to the organic substrate, with the concomitant loss of dinitrogen. This is illustrated by the example of the lutetium tetramethylcyclopentadienyl complex 3 (Scheme 9). In the case of carbon dioxide, reductive dimerisation to oxalate occurs,65 whereas reaction with phenazine or anthracene gives the two electron reduction 65 products. These reactions have direct parallels to [SmCp*2] 2 chemistry (vide infra) and have therefore been termed “LnII-like” in the literature.

Scheme 9 Reactivity of reduced dinitrogen complexes with organic substrates.65

1.3.2 Divalent Lanthanide Cyclopentadienyls and Heterocyclopentadienyls

Unsubstituted cyclopentadienyls are a less common feature of divalent lanthanide chemistry. There are a few examples in the literature but these tend to be limited to donor solvent adducts. Progress is especially frustrated in the case of samarium(II) by the notorious insolubility of bis(cyclopentadienyl)samarium(II) and its solvates.66 This prevents further reactivity despite its strong reduction potential.67

The polyalkylated and polysilylated cyclopentadienyls (Figure 6) tend to dominate this field due to their favourable steric and solubility properties (vide supra). There is also an increased requirement for sterically bulky ligands due to the larger coordination sphere of the divalent metal centres (cf. Six-coordinate ionic radii Yb2+ 102 pm vs Yb3+ 86.8 pm).2 As an example bis(bis(trimethylsilyl)cyclopentadienyl)europium is polymeric in the solid state,68 whereas bis(tris(trimethylsilyl)cyclopentadienyl)europium

11 is monomeric.69 Even base-free bis(pentamethylcyclopentadienyl)ytterbium(II) 70 [YbCp*2] crystallises as a linear coordination polymer in the solid state. In the absence of ligands that induce high steric congestion, aggregation is often more pronounced than trivalent analogues in the solid state.

Figure 6 Some divalent lanthanide metallocenes.

A wealth of chemistry has emerged from these classes of complex, perhaps none more so than bis(pentamethylcyclopentadienyl)samarium(II) [SmCp*2] 2. First reported by Evans in 1981 as the tetrahydrofuran adduct, it represents a milestone in divalent lanthanide chemistry as the first reported, soluble organosamarium(II) species.71 When desolvated under high vacuum it displays an unexpected bent geometry72 and reacts with dinitrogen to yield what was the first dinitrogen bound lanthanide metal complex (Scheme 10).73 This occurs in a side-on fashion unlike that for most transition metal dinitrogen complexes.

Scheme 10 Some reductive chemistry of bis(pentamethylcyclopentadienyl)samarium(II) 2.54,56,73

The reductive chemistry of [SmCp*2] 2 parallels that of its trivalent congener’s SIR reactivity (Scheme 7), particularly towards phosphine chalcogenides, azobenzene and cyclooctatetraene (Scheme 10).55,57,74 In this case the metal centre acts as a one electron reductant and is oxidised to samarium(III).

The reductive chemistry of [EuCp*2] is not as well investigated, likely a result of its weaker reduction potential relative to the other members of the classical divalent triad

12

(Eu3+|Eu2+ E0 -0.34 V, Table 1). Thus far most reports are limited to the single electron reduction of redox active N,N’-diaryl 1,4-diazabutadienes and structurally related ligands.75-78

As would be expected on account of the stronger reduction potential (Yb3+|Yb2+ E0 -

1.05 V, See Table 1), [YbCp*2] displays a richer reductive chemistry. Phosphine chalcogenides are reduced to the corresponding phosphines and bimetallic ytterbium chalcogenides (Scheme 11a).79 Diaryl chalcogenides are also reductively cleft in twain 80 by two equivalents of [YbCp*2] This reactivity is similar to that observed for the samarium congener (vide supra). Exhaustive studies of the coordination of reducible nitrogenous Lewis bases such as bipyridines,81-85 1,4-diazabutadienes85,86 and many other heterocyclic bases87 have demonstrated highly complicated magnetic behaviour. Temperature dependent magnetic measurements suggest an intermediate valence between ytterbium(II) and ytterbium(III) with some cases displaying antiferromagnetic coupling at low temperature.81,82,85,87 Detailed analysis of the magnetism in these complexes yields the conclusion that they adopt a multiconfigurational open-shell singlet ground state with the ytterbium displaying intermediate valence tautomerism.83-85

Scheme 11 Some reductive chemistry of [YbCp*2].

13

In contrast to the readily accessible [LnCp*2] (Ln = Sm, Eu and Yb) complexes of the classical divalent triad, the more strongly reducing bis(pentamethylcyclopentadienyl)thulium(II) species cannot be isolated.88,89 The metathesis reaction of thulium diiodide and potassium pentamethylcyclopentadienide under an argon atmosphere in diethyl ether results in reductive cleavage of the solvent, 2- whereas a dinitrogen atmosphere is reduced by thulium(II) to yield the side-on N2 88 ligand as observed with [SmCp*2] (Scheme 12). The same metathesis reaction performed in tetrahydrofuran under a dinitrogen atmosphere affords a thulium(III) lanthanocene iodide complex.89 Attempts to reduce this material with elemental sodium in 1,2-dimethoxyethane results in reductive cleavage of the solvent (Scheme 12).89

Scheme 12 The reactivity of pentamethylcyclopentadienylthulium(II).

The (poly)silylated thulium(II) metallocenes display similar reactivity with dinitrogen, yielding dinuclear reduced dinitrogen complexes in high yield (Figure 7).88 Stability towards ether cleavage (cf. Scheme 12) is provided by the 1,3- bis(trimethylsilyl)cyclopentadienyl ligand, which gives a stable thulium(II) 4 under an argon atmosphere.90

Figure 7 Polysilylated thulium lanthanocenes.88,90

A thulium(II) lanthanocene 5 has also been isolated using the extremely sterically congested tris(tbutyl)cyclopentadienyl ligand.91,92 This compound was isolated by

14 chemical reduction of a thulium(III) metallocene with potassium graphite (Scheme 13). Extension of this same strategy to dysprosium and neodymium (Ln3+|Ln2+ E0 -2.5 and - 2.6 V respectively, Tm3+|Tm2+ -2.3V. cf. Table 1) was unsuccessful. However the addition of 18-crown-6 permitted the isolation of the highly reducing divalent species as the K(18-crown-6)(μ-I) adducts 6 and 7 (Scheme 13).93,94

Scheme 13 Chemical reduction of extremely sterically congested metallocenes leading to divalent thulium 5, dysprosium 6 and neodymium 7 metallocenes.91-94

Much of the non-classical divalent lanthanocene chemistry above, i.e. chemical reduction of lanthanide metallocenes, has a basis in the pioneering work of Lappert et al., who chemically reduced a range of lighter lanthanocenes with alkali metals. Early studies observed the high reactivity of these proposed unstable divalent intermediates towards ethers.95,96 For instance, dimeric methoxides were isolated from reaction of the 1,2-dimethoxyethane solvent with the putative highly reducing divalent lanthanocenes of lanthanum, cerium and neodymium (Ln3+|Ln2+ E0 -3.1, -3.2 and -2.6 V respectively, Table 1) (Scheme 14).

Scheme 14 Cleavage of 1,2-dimethoxyethane by divalent lanthanide intermediates.85,86

For lanthanum, further evidence of the lanthanum(II) intermediate was provided by EPR studies that observed signals consistent with lanthanum(II) ([Xe] 5d1)97 rather than a mixture of lanthanum(III) and a radical anion ligand.96 In arene solvents, reduction of the solvent was observed during the reduction of several lighter lanthanocenes, to afford dianionic 1,4-cyclohexa-2,5-dienyl ligands that bridge trivalent lanthanides (Scheme 15).98,99 Divalent lanthanide intermediates were also observed using EPR spectroscopy.99

15

Scheme 15 Synthesis of cyclohexadienyl lanthanocene complexes.

Final evidence of the intermediate divalent species was provided in the isolation of a subvalent lanthanum species with a bridging benzenide monoanion 8.100 This formulation was supported by a single crystal X-Ray study (Figure 8). This structural motif was also observed in the later isolation of toluenide complexes of both lanthanum(II) 9 and cerium(II) 10 (Figure 8).101

Figure 8 Products obtained from the reduction of some early lanthanocenes in arene solvents.

Recently it has been shown that the reduction of trivalent lanthanocenes with excess

KC8 in the presence of 18-crown-6 or 2.2.2-cryptand affords isolable, unambiguously divalent lanthanide tris(cyclopentadienyl) anions (Figure 9).41-43 This method has even been extended to afford a uranium(II) compound44 and gives complexes of moderate stability in the absence of donor solvents. Interestingly, Density Functional Theory Molecular Orbital (DFT MO) calculations describe a HOMO that has a high degree of 5d1 character.41-43 UV-Visible spectra of these compounds were consistent with this calculation. The conclusion has been made that unlike the lanthanide diiodides, which are all 4f n at the metal,11 these species exist in the 4f n 5d1 configuration. 41-43

16

Figure 9 Structurally characterised divalent lanthanide cyclopentadienyls from chemical reduction with excess KC8. Another small but growing class of divalent lanthanide complexes is those incorporating the heterocyclopentadienyl family of ligands. These ligands have allowed the isolation of many non-classical divalent lanthanide heterocyclopentadienyls (Figure 10).91,102,103 Thulium heterocyclopentadienyls seem to possess greater stability than the corresponding thulium cyclopentadienyls (vide supra). This has been attributed to the weak π-donor character of these ligands, which better stabilises the electron rich divalent lanthanide than the corresponding cyclopentadienyl ligands.102,103 This is exemplified by the use of triphospholyl ligands for the stabilisation of low oxidation state scandium complexes either through chemical reduction of trivalent precursors analogous to that discussed earlier or by metal vapour synthesis.104-106

Figure 10 A selection of rare-earth heterocyclopentadienyls.

1.3.3 Cerium Cyclopentadienyls, Cyclooctatetraenyls and Pentalenyls

Despite the readily accessed tetravalent oxidation state (Table 1), the organometallic chemistry of cerium is more common in the trivalent state. Cerium(IV) is strongly oxidising and this is typically incompatible with the reducing nature of organometallic ligands and their precursors. There are some examples that border on the tetravalent state (Figure 11). The best known example is bis(cyclooctatetraene)cerium 11107 and the substituted derivatives thereof.108-110 While formally cerium(IV) with each ligand having a -2 charge, there has been much speculation in the literature as to the true cerium oxidation state.111,112 The current theory classifies the cerium ion in these 17 compounds as possessing a multiconfigurational ground state. Cerocene is best described as an 80:20 mixture of the cerium(III) 4f 1 and cerium(IV) 4f 0 states. The pentalene relatives of the cerocenes display similar behaviour (Figure 11, right).113,114 Whatever the “true” oxidation state of the metal, chemically they behave as cerium(IV) compounds and their typical synthesis proceeds via oxidation of a trivalent cerate - precursor (e.g. [Ce(pentalene)2] ).

Figure 11 A selection of “tetravalent” cerium organometallics.

Some cerium(IV) cyclopentadienyls are known (Figure 11, far right). They are synthesised by the chemical oxidation of tris(cyclopentadienyl)cerium115 (R = Cl) or by the substitution of alkoxide116 or nitro ligands117 at cerium(IV) alkoxide precursors. A detailed examination of the oxidation state in these compounds has not been conducted t at this time although [CeCp3Cl] and [CeCp2(O Bu)2] have been reported as diamagnetic.115,117

1.3.4 Trivalent Lanthanide Amides

Lanthanide amides are widely used in non-aqueous coordination chemistry. The sterically demanding and readily available bis(trimethylsilyl)amide ligand is by far the most popular. The complete series of homoleptic lanthanide complexes of this ligand was first reported by Bradley in 1973.118 As low coordinate lanthanides with an unusual trigonal pyramidal geometry they were originally only of academic interest, yet have remained popular since due to synthetic versatility and ease of synthesis.

Scheme 16 The synthesis of tris(bis(trimethylsilylamide))lanthanides.

These complexes are readily synthesised from the lanthanide trihalides or pseudohalides119 in a metathesis reaction with the alkali metal amide (Scheme 16). Purification is simplified by the volatility of the silylamides allowing sublimation as an alternative to crystallisation. The high steric demand of the ligand only allows for a 18 coordination number of three at the metal centre and confers a high degree of solubility to the complex.

Scheme 17 Protolysis reactions of lanthanide silylamides.

Protolysis reactions from these complexes provide a convenient starting material for countless other lanthanide amides, cyclopentadienyls and alkoxides.120 Protic ligand precursors with a higher Brønsted acidity than the silylamine may be reacted in a stoichiometric fashion with the lanthanide silylamide precursor with the amine co- product removed readily in vacuo. Particularly useful is the ability to conduct these reactions in non-coordinating solvents, which eliminates the possibility of solvent adduct formation. A wide range of ligand precursors are suitable as shown above (Scheme 17). This list is by no means comprehensive but displays the synthetic versatility of these silylamide complexes. In some cases the bis(trimethylsilyl)amide ligand is too sterically demanding to allow the protolysis reaction to occur. In this instance the smaller bis(dimethylsilylamide) may be used (Scheme 18). This is generally encountered for the larger cyclopentadienyls121 and extremely sterically demanding alkoxides122 when coordinated to the smaller late lanthanides.

Scheme 18 Protolysis reactions for the introduction of extremely sterically demanding ligand precursors.

Treatment of tris(bis(trimethylsilyl)amide)lanthanides with a reducing alkali metal under a dinitrogen atmosphere demonstrates another example of “LnII-like” chemistry. 2- For the smaller metals of the lanthanide series a dimeric N2 complex is isolated with each metal present in the lanthanide(III) oxidation state (Scheme 19). In the case of the larger metals the tetrakis(bis(trimethylsilyl)amide)lanthanate compound is also formed and due to preferential crystallisation is the only product isolated. This product is not observed for the smaller metals, presumably due to steric constraints. For the larger metals, 15N NMR spectroscopy of the crude reaction mixtures is also used as evidence

19

2- of the respective dimeric N2 complexes. It is noteworthy that this chemistry fails under an argon atmosphere, indicating that the dinitrogen is mechanistically important.

Scheme 19 Products isolated from reduction of lanthanide amides under dinitrogen.

This reaction is more complicated for yttrium. Initial attempts isolated the same dimeric 2- N2 complex as those for the heavier rare-earths (Scheme 19). Closer investigation of these results revealed that a complicated series of reactions occur (Scheme 20).123-125 Importantly these provided evidence for a stable yttrium(II) oxidation state.125 At the time this oxidation state had previously not been observed but was later isolated as a metallocene (Figure 9).41 Yttrium(II) has been shown to be capable of reducing both 1- 1- 126 carbon monoxide and carbon dioxide to CO and CO2 respectively. The fully 3- 2- 127 desolvated N2 complex reacts with NO to give the radical dianion NO .

Scheme 20 The reduction chemistry of tris(bis(trimethylsilyl)amido)yttrium.

1.3.5 Divalent Lanthanide Amides

The chemistry of the divalent lanthanide amides in many ways parallels those of the cyclopentadienyl systems. The bis(trimethylsilyl)amide ligand remains popular amongst the classical divalent series since being pioneered by Andersen.128,129 Donor solvent adducts are extremely common. For example solvates are known for the three common ethereal solvents (tetrahydrofuran, 1,2-dimethoxyethane and diethyl ether) with all three of the classical divalent triad (Figure 12).128-131

20

Figure 12 Some divalent lanthanide amides.

The demand for high coordination numbers is also nicely illustrated with reference to these compounds. One formal equivalent of alkali metal amide may be incorporated into the molecule. These are distinguished from the lanthanide(III) amides in that amide bridging to the alkali metal enforces planarity at the lanthanide metal as opposed to the pyramidal geometry of the lanthanide(III) compounds (vide supra).129,132,133 Use of the less sterically demanding bis(dimethylsilyl)amide ligand, which has found some use in lanthanide(III) chemistry (vide supra) affords trimetallic oligomeric species in the solid state.134-136

Perhaps more interesting is the chemistry of the analogous, more strongly reducing divalent lanthanides. Attempts to synthesise thulium(II) and dysprosium(II) bis(trimethylsilyl)amide compounds exemplify the ease by which these metals are able to reduce dinitrogen (Scheme 21).137 These compounds are isostructural with those obtained by chemical reduction of the tris(amido)lanthanide(III) complexes under 2- dinitrogen (cf. Section 1.3.4). The dinitrogen ligand is formally assigned as N2 with one electron transferred from each metal centre. The putative divalent intermediates have not been isolated at this time.

Scheme 21 Metathesis reaction leading to dinitrogen reduction by dysprosium(II) and thulium(II) amides.

The reactivity of these divalent amides with organic substrates is less well examined than the corresponding metallocenes or the halides. Studies of ketone reduction and dehalogenation using [Sm(N(SiMe3)2)2(thf)2] demonstrated a rate enhancement over 138 SmI2 and SmI2/HMPA mixtures. Curiously, experimentally determined thermodynamic redox potentials showed that [Sm(N(SiMe3)2)2(thf)2] has a reduction potential between the other two systems.138 When considered with the temperature insensitivity of these reactions an inner sphere mechanism of reduction was concluded 21 in contrast to the outer sphere mechanism proposed for SmI2/HMPA mixtures. The authors noted that these data would prevent the use of this reagent in the Barbier 20 reaction, a typically cited strength of SmI2, due to the rapid rate of ketone reduction relative to alkyl halide reduction.

The one electron reduction of 9-fluorenone occurs readily with [Sm(N(SiMe3)2)2(thf)2] as observed using UV-Vis spectroscopy.139 Interestingly the ketyl product was unisolable due to a reversible, concentration dependent, pinacol coupling (Scheme 22).

This contrasted with [LnCp*2] (Ln = Sm, Yb) complexes, which allowed the isolation of the ketyl complexes in high yields.139

Scheme 22 The reduction and pinacol coupling of 9-fluorenone by [Sm(N(SiMe3)2)2(thf)2].

A detailed mechanistic investigation of the reduction of nitroarenes by

[Sm(N(SiMe3)2)2(thf)2] was able to fully characterise all the intermediates in this reaction (Scheme 23).140 The authors suggest analogous intermediates are also present in reductions using samarium(II) iodide. 6 equivalents of samarium diiodide are required for complete reduction of the nitroarene to the corresponding aniline. Most interestingly the reduction of azobenzene by this complex gives a bimetallic two electron reduction product with the dihydrazide bridging the two metal centres. This 141 contrasts with the one electron reduction that can be observed with [SmCp*2] (cf. Section 1.3.2).

Scheme 23 The products of the samarium(II) reduction of nitro arenes pre-hydrolysis.

Other reports of the reductive reactivity of divalent lanthanide amides with organic substrates include the reductive defluorination of aliphatic C-F bonds142 using

[Sm(N(SiMe3)2)2(thf)2], the Reformatsky type cyanation of chroman-4-ones using 143 mixtures of SmI2 and [Sm(N(SiMe3)2)2(thf)2] and the reductive coupling of imines 22

144 with [Sm(N(SiMe3)2)2(thf)2] or other samarium(II) compounds.

[Sm(N(SiMe3)2)2(thf)2] has also been used to catalyse the addition reactions to 145 carbodiimides and N-heterocyclic carbene adducts of [Yb(N(SiMe3)2)2] have shown promise as catalysts for cross-dehydrogenative coupling of silanes with amines.146

1.3.6 Tetravalent Cerium Amides and Oxidation Chemistry

The amide chemistry of tetravalent cerium suffers from the same difficulty observed for cerium metallocenes (cf. Section 1.3.3). In spite of this, there are several formally tetravalent cerium amides known (Figure 13). Interestingly many of these are diamagnetic in contrast to the cerocenes and cerium(IV) pentalenyls discussed earlier (cf. Section 1.3.3), which confirms the closed-shell configuration of cerium(IV).147-149

Figure 13 Some of the known cerium(IV) amido complexes.147-154

An examination of the synthesis of these compounds is a worthwhile exercise as it clearly demonstrates the difficulties faced in high oxidation state cerium chemistry (Scheme 24). In spite of the ready availability of ammonium cerium(IV) nitrate as a precursor, metathesis reactions often fail for non-alkoxide ligands due to the reducing nature of the ligand precursors as well as the redox activity of the nitrate ligands.117,150 It is for this reason that many of these compounds are prepared through chemical oxidation of trivalent precursors. Typically a neutral cerium(III) or alkali metal cerate(III) complex is treated with an oxidant that introduces a halide ligand. A wide 152 153 range of organic and inorganic oxidants have been used including TeCl4, PPh3Br2, molecular halogens,151 hexachloroethane,155 trityl halides147,149 and ferrocenium salts.149

Scheme 24 Synthetic routes to tetravalent cerium compounds.117,156 23

The reaction of dioxygen with cerium(III) amides has been demonstrated to have a range of reaction outcomes.157 This may explain why cerium(III) compounds, of all the lanthanide series, are often the most sensitive to oxygen. Treatment of [Ce(N(SiMe3)2)3] with molecular oxygen in excess at -27 °C affords either the peroxo or the bridging dioxide complex, depending on solution concentration (Scheme 25).157 All of these compounds are tentatively assigned as cerium(IV) by the authors and exhibit diamagnetism but have very low stability at room temperature in solution or the solid state. This precludes definitive assignment of the nature of the bridging oxygen unit as either a peroxide or a superoxide. Treatment of the same compound with molecular oxygen and an equivalent of an alkali metal amide gives the bimetallic oxy cerate compounds, again in low yield and room temperature stability.157

Scheme 25 The reactions of dioxygen with [Ce(N(SiMe3)3].

1.4 Purpose of this Thesis

The organometallic chemistry of the lanthanide metals has grown rapidly in the decades since its renaissance in the 1970s. A key factor in much of this chemistry has been the development of functionalised cyclopentadienyl ligands that have allowed for the isolation of coordinatively unsaturated monomeric metallocenes. Most relevant to this project is the reductive chemistry that has been developed in the past two decades of research. Metal and ligand derived electron transfer processes have revealed new bond formation mechanisms and chemical reactivity that have no parallels elsewhere in the periodic table.

The corresponding reduced lanthanide amide chemistry is a less developed field, both in terms of fundamental research and the corresponding reactivity with organic substrates. Lanthanide amides have the same, if not greater capability for steric and electronic modulation, as the planar cyclopentadienyls. To that end, Chapter II outlines the 24 synthesis of sterically demanding 1,3-diaryl triazenes and the alkali metal complexes of the resulting triazenides. These display interesting structural features as well as being useful precursors for coordination to lanthanide metals and the corresponding reductive chemistry thereof.

Chapter III reports the synthesis of sterically congested triazenido complexes of the lanthanide metals in both the trivalent and divalent oxidation states. The reductive chemistry is examined with a view to determining parallel reactivity to existing reactions with metallocenes. The organometallic and organoamido chemistry of cerium(IV) remains an underdeveloped field. There as yet exists no general route into non-aqueous complexes through metathesis or oxidation. In an attempt to further this research, some attempts towards cerium(IV) amides are outlined in this chapter.

Chapter IV outlines the reaction of an extremely sterically congested aryl iodide with elemental calcium and the lanthanide metals. This was conducted in an attempt to further the understanding of the reaction products of lanthanide metals with iodoarenes. It was anticipated that the steric demand of the iodoarene would kinetically stabilise the iodocalcium and -lanthanide aryls. The reaction of calcium and ytterbium did not give the anticipated metal aryls, instead reductively cyclised fluorenide Solvent Separated Ion Pairs (SSIPs) were observed as products in low-yield. This chapter proposes a mechanism of formation based upon the available evidence, as well as some of the observed decomposition products.

25

Chapter II - Main-Group Metal Complexes of Triazenides and a Formamidinate

2.1 Introduction 2.1.1 Main Group Compounds as Synthetic Precursors for Organolanthanide Chemistry

The synthesis of organolanthanide compounds can be considered reliant on two synthesis methods: salt metathesis or redox transmetallation.45,158 The former method requires an alkali metal, or thallium complex as a ligand precursor, which is then reacted with a lanthanide halide or trifluoromethanesulfonate (Scheme 26a and b). In general lithium complexes are not preferred as precursors due to the propensity for lithium halide incorporation in the final product1,45 and magnesium complexes are most reactive with the trifluoromethanesulfonates.85,99,159 Otherwise the variables in Scheme 26 are highly interchangeable, with the alkali metals the preferred metathesis reagents.45,120,158 The metathesis method is often used to synthesise lanthanide amides and alkyls, which will then react in an acid-base fashion with a desired ligand precursor. This synthesis is known as protolysis and is also very common. Some common 120 complexes synthesised in this fashion are lanthanide silylamides [Ln(N(SiMe3)2)n] 160 and trimethylsilylmethyls like [Ln(CH2(SiMe3))n]. Redox transmetallation requires a complex of an element that is less reducing than the lanthanide.45,158 Reaction with an elemental lanthanide occurs via a redox process with concomitant transfer of the ligand to the lanthanide ion (Scheme 26c). The more electronegative mercury and thallium compounds are generally used for this purpose due to their ease of synthesis and stability.161-164

26

Scheme 26 Metathesis (a and b) and redox transmetallation (c) reaction methods for the synthesis of organolanthanide complexes.

2.1.2 Triazenes and Related Amidines

Amidinates and guanidinates are extremely common ligands in coordination chemistry.165-169 They are part of the wider family of heteroallylic ligands.168,170 In main group chemistry, they are often used to kinetically stabilise coordinatively unsaturated, typically low-oxidation state complexes. Some of the most common ligand motifs are illustrated in Figure 14.

Figure 14 Some of the many amidines, guanidines and silylamidines known.

These ligand precursors are readily synthesised in bulk from commercially available anilines (Scheme 27a and b),170,171 carbodiimides (Scheme 27c)166,170 or aryl nitriles (Scheme 27d).172 As a result of their ready preparation, a large degree of steric and electronic tuning is possible based upon variation at the N- and C-substituents. These features have made them particularly popular ligand precursors for main group chemistry.167

27

Scheme 27 Synthesis of formamidines (a), amidines (b), guanidines (c) and silylamidines (d).

Triazenides, in contrast, have not been as popular as ligands for the main group metals. This is in spite of the fact that simple triazenes have been known173 and used as ligands174,175 for many decades (Figure 15). They have been quite popular for transition metal chemistry, particularly for generating early transition metal paddlewheel type complexes with metal-metal bonds (Figure 15).176-180 Figure 15 also shows some of the wide variety of coordination geometries that are observed with the triazenide ligand: the 2 1 1 η monometallic chelating 12-14 and μ:η :η bridging bimetallic 15-16.

Figure 15 A selection of transition metal triazenide complexes.

A general examination of triazenide bonding motifs is a worthwhile exercise; in general triazenides have a range of bonding modes, some of which are similar to those observed for amidinates and guanidinates.166,168,172 The nomenclature of some of these bonding modes is indentical to that used for metal amidinates (Figure 16). These include the E- syn and E-anti (Figure 16 A and B) as well as Z-syn and Z-anti (Figure 16C and D) designations respectively.171 These are all determined using conventional Cahn-Ingold- Prelog priority rules.181,182

28

Figure 16 Bonding designations for of metal triazenides.

The general structures of some triazenide bonding modes may be seen in Figure 17. Across the periodic table the chelating monometallic binding mode is the most common (Figure 17A). A μ:η1:η1 bridging bimetallic bonding mode (Figure 17B) is also very common, especially for metals that participate in metal-metal bonding such as chromium, molybdenum and tungsten,176,177,180 as well as copper, silver and gold.183-185 For the late transition metals the monodentate η1 mode is also common being observed for iridium,186 nickel,187 palladium,188 platinum189 and mercury190-192 complexes (Figure 17C). In all these examples the triazene adopts the E-syn configuration. There are also two bonding modes that are unique to triazenide ligands. The first of these is a rare bridging bimetallic (Figure 17D), with only three compounds reported, one nickel193 and two copper complexes.194,195 The final, unique bonding mode (Figure 17E) is observed principally for f-block and early transition metal triazenide complexes.196-199 In this mode the triazenide adopts a κ2-N1,N2 coordinated Z-syn geometry. In one instance this mode occurs in bridging fashion, spanning two metal centres in a μ-κ2-N1,N2, κ1- N3 fashion.200

Figure 17 The bonding modes commonly found for metal triazenide complexes.

Extensive studies have determined that some of these bonding modes may be distinguished by IR spectroscopy. Monodentate triazenide complexes (Figure 17C) have absorbances in several regions of the spectrum including 1150, 1190-1210, 1260-1300 and 1580-1600 cm-1.201 In contrast, chelating triazenide ligands (Figure 17A) have only the absorbances at 1260-1300 and 1580-1600 cm-1.201 It has been suggested that the latter absorbance arises from the aryl groups of the triazenide ligand not the central N3 moiety.201 Bridging triazenides (Figure 17B) are further distinguished by an absorbance in the range 1350-1375 cm-1.201

Triazenes may be prepared by one of two general synthetic methods. The older, more common method requires an aniline starting material that is readily diazotised under 29 aqueous conditions.173 Addition of second equivalent of aniline under basic conditions gives the desired triazene (Scheme 28a). More sterically demanding triazenes, which by extension are not readily diazotised in water due to low solubility, are prepared from aryl azides and carbanion nucleophiles, usually aryl lithium or aryl magnesium compounds.202-204 Hydrolysis permits the isolation of the desired triazene generally in good yield (Scheme 28b).

Scheme 28 The two general methods for synthesising triazenes.

2.1.3 Alkali Metal Complexes

Alkali metal amides are very common due to the ready availability and simple handling protocols of alkali metal bases, such as nbutyllithium and alkali metal bis(trimethylsilyl)amides, that are frequently used as starting materials.171,172 Alkali metal amidinates and guanidinates are no exception and have been well researched.165,166,170-172 The lithium complexes are the most common, especially amongst the amidinates.171,172

Some selected examples of alkali metal complexes of the sterically demanding 2,6-diisopropylphenylformamidinate, DippForm, ligand are shown in Figure 18. These provide a representative sample of the varying coordination modes and numbers for such systems. For example, with tetrahydrofuran as a Lewis base, monomeric complexes are observed with lithium 17 and sodium 18.205 The potassium complex 19 exists in the solid state as a linear polymer chain.206 This is likely a result of the potassium’s tendency to form metal to arene-π interactions.206,207 Some metal size effects are also observed. Lithium complexes 17 contain two tetrahydrofuran (thf) ligands or one 1,2-dimethoxyethane (dme) ligand 20.205 The corresponding sodium complexes 18 and 21 are able to coordinate one extra ligand due to the increased ionic radius relative to lithium.2 The use of a sterically demanding tridentate amine Lewis base such as pentamethyldiethylenetriamine (PMDTA) affords monomeric complexes

30 for lithium,208 sodium209 and potassium207 22-24. Similar comprehensive structural studies of the wider family of amidinates are not as prevalent.

Figure 18 Selected 2,6-diisopropylformamidinate (DippForm) alkali metal complexes.205-209

A recurring theme in this field is the change in coordination geometry and number with increasing steric demand of the amidinate ligand. For example the lithium complex of the small N,N’-diisopropylbenzamidinate 25 (Figure 19) exists as a dimer in the solid state with a bidentate bridging E-anti ligand as determined using X-ray crystallography.210 Unsymmetrical coordination geometries are enforced in the solid state for complexes of amidinate ligands with extremely sterically demanding backbone substituents such as 2,6-dimesitylphenyl 26211 or 2,6-bis(triisopropylphenyl)phenyl 27.212 The less sterically demanding 2,6(p-tbutylphenyl)benzamidinate 28 affords the bidentate chelating E-anti geometry212 and the 1-triptycyl amidinate 29 gives an unusual Z-syn coordination geometry.213 It should be noted that these lithium amidinates210,211 and many more have found use as metathesis precursors for lanthanide complexes. For a more in depth examination of the coordination chemistry of alkali metal amindinates, the interested reader is directed towards some of the many excellent reviews that have been written on the subject.165,170-172

Figure 19 Lithium complexes of extremely sterically demanding amidinate ligands.210-213

31

When compared to their amidinate cousins, guanidinate ligands are generally considered to be more electron-rich at the nitrogen donor atoms. This is a consequence of the three resonance forms that contribute to the overall electronic structure of the guanidinate ligand (Figure 20).166 In the two diazaallyl forms (Figure 20A and B) the electronic structure is akin to the amidinate ligands discussed above. In the iminium/diamide form (Figure 20C) a much greater negative charge density is localised at the N-donor atoms.214 These resonance contributors allow guanidinate ligands to be suitable for metals in a wide range of oxidation states.167,215,216

Figure 20 The three resonance contributors of the guanidinate ligand.

Alkali metal guanidinates show a similar degree of structural variety to that observed in amidinates. This includes the propensity for higher order nuclearity in the solid state for ligands with low steric demand, as may be seen in the selected examples in Figure 21. The addition of Lewis bases such as thf, as in sodium guanidinate 30, is insufficient to break up the strongly bonded dimeric structure.217 Sodium guanidinate 31 exists in the solid state as a base free trimer.218 The potassium complex 32 exists as a dimer218 with a similar bonding motif to the related lithium guanidinate 33 but with an increase in bonding interactions from the potassium to both guanidinate ligands due to the higher ionic radius of potassium relative to lithium.2 Lithium guanidinates 33-34 have found use as metathesis precursors for the synthesis of lanthanide guanidinates.219,220 The coordination chemistry of alkali metal guanidinates has been extensively reviewed.166,167,170

Figure 21 Selected compounds displaying some of the structural diversity in alkali metal guanidinates.217,218

32

In comparison to amidinates and guanidinates, triazenides are considered to be weaker donors. This has been demonstrated by: qualitative comparison of the pKa values of the 221 protonated acids, which showed that amidines have higher pKa values than triazenes; evaluation of the infrared spectrum of triazenides coordinated to transition metal 222 carbonyls, wherein each of the amidinates gave lower νCO stretches for the complexes than the corresponding triazenides and a Natural Bond Orbital analysis of a model triazenide in comparison to a model amidinate,204 which showed a lesser Natural Population Analysis charge at the donor nitrogen atoms in triazenides than in amidinates.

In contrast to the amidinates and guanidinates discussed above, there are only eight published alkali metal triazenides.185,221,223-225 These are illustrated in Figure 22. Interestingly the majority of these compounds are potassium complexes; this directly contrasts with amidinates and guanidinates, where the majority of characterised complexes are of lithium. There are only two lithium compounds221,224 and one sodium compound.185 One caesium triazenide complex has also been published.224 Given the ready availability and similar handling procedures for each of the alkali metal bases necessary to synthesise these compounds and the vast number of alkali metal amides, amidinates and guanidinates, the absence of a wider library of alkali metal triazenides is conspicuous. To date, only the potassium biphenyl triazenide 35 has been used as a metathesis precursor,226 the remaining compounds have been studied purely from a fundamental, coordination chemistry standpoint. Interestingly a large number of these compounds have been characterised as Lewis base free complexes. As a result a large number of metal arene-π interactions are observed in the solid state for the more sterically demanding triazenide ligands coordinated to potassium.

Figure 22 All of the known alkali metal triazenides.185,221,223-225

33

2.1.4 Thallium Complexes

Thallium triazenide complexes are somewhat common in comparison to the other s- and p-block elements. With that said, there are only 11 published thallium(I) or thallium(III) triazenide complexes (Figure 23). The crystallographically characterised examples number even less (9), however some structural trends may be observed from this limited sample. In each case the triazenido ligand adopts a chelating binding mode at thallium(I)225,227,228 and at thallium(III).229,230 A μ:η2:η2 mode is also observed with simple thallium(I) triazenides 37 and 38,227,228 as was observed for the potassium analogue 36. Extremely sterically demanding terphenyl triazenide ligands, as developed by Niemeyer, also show identical solid state behaviour to their potassium analogues (vide supra).225 One example displays a chelating bimetallic coordination mode, with the rest monomeric (Figure 23). These thallium compounds are sterically saturated by metal to arene-π interactions with the flanking aryl groups of the terphenyl ligand. Thus far no reports of the use of any of these compounds as ligand transfer agents in either redox transmetallation or metathesis have appeared in the literature.

Figure 23 The known thallium triazenide complexes. 225,227-230

Rarer still are thallium amidinates and guanidinates. In total there are six thallium amidinates and guanidinates in the chemical literature (Figure 24).215,216,231 Five of these are structurally characterised and are all thallium(I) compounds. The remaining, non- structurally authenticated compound was reported several decades ago and is the only thallium(III) amidinate known.232

34

Figure 24 All known thallium amidinates and guanidinates.215,216,231,232

2.1.5 Purpose of this Chapter

This chapter outlines the synthesis of a range of triazenide complexes of the alkali metals with different Lewis bases. The coordination chemistry and the solid state structures of these compounds at the alkali metals are compared to existing alkali metal amidinates. Also described is the synthesis and characterisation of the extremely sterically demanding N,N’-diaryl triazene Dipp*2N3H 39. (Figure 25) The coordination chemistry of this triazene is compared to the coordination chemistry of the previously known triazene Dipp2N3H 40. The attempts made to synthesise a structurally analogous formamidine Dipp*FormH 41 are also outlined. The effect of the different electronic character of triazenides on the coordination chemistry at the alkali metals, when compared to their amidinate and guanidinate cousins, is considered. It is anticipated this will further the chemistry of this underdeveloped field. The practicality and efficacy of the alkali metal triazenides as ligand transfer agents is examined as part of this thesis and as such their thermal stability is investigated and discussed in terms of their suitability for metathesis chemistry. The use of alkali metal triazenides as ligand transfer agents for amidolanthanide compounds is developed in Chapter III.

Figure 25 Ligand precursor targets.

35

2.2 Results and Discussion 2.2.1 Synthesis of Triazene Precursors and Triazenes

1,3-Bis(2,6-diisopropylphenyl)triazene, Dipp2N3H 40 was prepared by modification of 200 the method of Hill (Scheme 29). The reaction of 2,6-diisopropylaniline, DippNH2 42 with an excess of cold isoamyl nitrite gave the desired product in low yield 41% (lit. 52%) after storage at -25 °C for one week.200 This contrasts with the published method, in which the product was observed to crystallise after only two hours at room 200 1 temperature. The H NMR spectrum of Dipp2N3H 40 in C6D6 displays the corresponding downfield shift of the resonance attributable to NH proton of triazene 40 as well as a reduction in integral value relative to the resonances for the parent aniline 42. This resonance appears as an extremely broad signal at δ 9.01. A downfield shift of the methine resonance is also observed relative to that of aniline 42. This signal occurs at 3.31 ppm as a broad ill-defined multiplet rather than the septet in the aniline 42 starting material. These values compare favourably to the reported literature values.200 Bertrand and co-workers have commented that this reaction is susceptible to violent runaway thermal reactions.233 The reported reaction time of 2 hours has not been reproducible in the author’s hands.200 While visible signs of the reaction were no longer observed (cessation of gas evolution and no further colour changes) the product is only obtained as a crystalline precipitate after prolonged cooling to -25 °C. The variable purity of commercially available technical grade 2,6-diispropylaniline has also been observed to dramatically affect the yield during repeated preparations of this compound, even after re-distillation. It is likely that this is a root cause of the problems in reproducing the original reaction time.

Scheme 29 Preparation of 2,6-diisopropyltriazene, Dipp2N3H 40.

Despite three recrystallisations from npentane during purification, residual aniline 42 could not be removed from triazene 40, which has been reported in the literature to be contaminated with up to 15 percent of the starting aniline.200,233 Microanalysis returns values consistent with the incorporation of 12% of the aniline 42 starting material in the

36

1 final product by percent carbon composition. The H NMR spectrum in C6D6 also displays peaks corresponding to the minor constituent reported by Hill, which is 200 believed to be a bonded pair of DippNH2 42 and Dipp2N3H 40 (Figure 26). Consistent with the observations of Hill no single crystals that gave the structure 200 displayed in Figure 26 were observed. Contamination with DippNH2 was not observed to affect the purity of subsequent products synthesised from triazene 40.

Figure 26 The proposed hydrogen bonded aniline-triazene pair.

In addition to the solvent free solid state structure reported by Hill (triclinic, P1¯),200 triazene 40 crystallises from a saturated solution in toluene as large colourless blocks in the triclinic space group P1¯. A representation of the structure may be seen in Figure 27. Two full molecules are included in the asymmetric unit as a hydrogen bonded dimer as well as a heavily disordered toluene molecule, which could not be successfully modelled. The SQUEEZE234 function of PLATON235 was used to successfully refine the structure. Unlike the previously reported structure,200 the amino hydrogen atoms are non-locateable and as such have been modelled over two locations with equal occupancy. The intermolecular nitrogen to nitrogen distances (N(1)-N(4) 2.96(1) and N(3)-N(6) 2.98(1) Å ) are statistically equivalent in contrast with the non-equivalent distances in the reported structure (N(1)-N(4) 3.004(3) and N(3)-N(6) 2.931 (3) Å)200 but fall within the range of the reported distances. In addition to this the nitrogen- nitrogen bond lengths are statistically identical for both molecules (N(1)-N(2) 1.285(9) Å and N(2)-N(3) 1.294(9) Å, N(4)-N(5) 1.303(10) Å and N(5)-N(6) 1.324(10) Å). As such the location of the double and single bonds cannot be definitively assigned. This contrasts with the reported structure where clear localisation of the double and single bonds is observed (N(1)-N(2) 1.333(3), N(2)-N(3) 1.270(3), N(4)-N(5) 1.280(3), N(5)- N(6) 1.326(3) Å).200

37

Figure 27 Molecular structure of Dipp2N3H 40, POV-RAY illustration, 50% thermal ellipsoids, all non-amino hydrogen atoms omitted for clarity. The isopropyl carbon atoms are displayed as wireframes for clarity. Selected bond lengths (Å) and angles (°): N(1)-N(2) 1.285(9), N(2)-N(3) 1.294(9), N(4)-N(5) 1.303(10), N(5)-N(6) 1.324(10), N(4)∙∙∙∙N(1) 2.96(1), N(6)∙∙∙∙N(3) 2.98(1), N(1)–N(2)–N(3) 111.6(8), N(4)–N(5)–N(6) 110.8(7).

2,6-Dibenzhydryl-4-methylaniline (Dipp*NH2) 43 was prepared using the method of Markó with modification,236 4-methylaniline was dialkylated using a Friedel-Crafts type reaction with diphenylmethanol in the presence of and hydrochloric acid (Scheme 30). Rather than washing the crude product with ethyl acetate as reported, aniline 43 was obtained as a fine white powder after trituration with a boiling ethyl acetate/hexane mixture followed by cooling to room temperature. It has also been found to be unnecessary to conduct this reaction in a sealed tube as reported in the original preparation,236 thereby permitting larger scale preparations to be conducted in regular laboratory glassware. A follow-on benefit of the increased scale is the slightly improved yield of 86% (lit 75%).236

Scheme 30 The synthesis of 2,6-dibenzhydryl-4-methylaniline, Dipp*NH2.

1 The H NMR spectrum of Dipp*NH2 43 in C6D6 displays diagnostic singlet resonances at δ 5.46 and 6.73 in C6D6 that may be attributed to the benzylic CH and meta-aromatic

38

CH respectively. The signal corresponding to the amino group is shifted downfield to δ 3.15 upon alkylation. The resonance attributed to the methyl group is observed at δ 1.92. These are consistent with the values reported by Markó.236

Treatment of aniline 43 with excess sodium nitrite in a tetrahydrofuran water mixture in the presence of a seven-fold excess of sulfuric acid affords the intermediate diazonium salt 44. This was not isolated but treated with excess sodium azide in situ to give 2,6- dibenzhydryl-4-methylazidobenzene, Dipp*N3 45 in near quantitative yield. Dipp*N3 45 melts at 132-134 °C and readily crystallises from many common organic solvents as solvent-free fine colourless needles. Samples stored as dry powders at room temperature over the period of several months were observed to be indefinitely stable by 1H NMR spectroscopy in C6D6. This is a desirable property in an organic azide, which often have the potential to be explosive.237

* Scheme 31 Synthesis of aryl azide, Dipp N3 45

1 The H NMR spectrum of azide 45 in C6D6 shows the absence of the previously observed resonance assigned to the amino , confirming successful diazotisation. The resonance ascribed to the methyl group is shifted upfield relative to aniline 43 from δ 1.92 to 1.77. The singlet corresponding to the benzylic protons is observed to shift downfield to δ 6.06 from δ 5.46. The meta-aromatic CH proton resonance is also shifted upfield to δ 6.83 in comparison to aniline 43 at δ 6.73. The -1 infrared spectrum of Dipp*N3 displays an intense, sharp absorbance at 2126 cm . This is consistent with the presence of an aryl azide functional group,238-240 specifically the asymmetric N≡N stretch which is often observed in the range 2083-2114 cm-1.241

Single crystals of Dipp*N3 suitable for an X-ray diffraction structure determination were grown from a saturated solution of Dipp*N3 in 1,2-dimethoxyethane stored at room temperature over a period of several days. Dipp*N3 45 crystallises in the triclinic space group P1¯ with one full molecule in the asymmetric unit. An illustration of the structure may be seen in Figure 28. The azide functional group is bent out of the plane 39 of the central aromatic ring. This is shown by the angle N(2)-N(1)-C(1), which is measured as 119.0(4)°. This angle has been reported with a wide range in the literature for other aryl azides (Figure 29).242-245 For example, aryl azides that feature sterically demanding functional groups ortho to the azide functional group have a wide range of 242 reported values for this angle. Tripp2PhN3 46 exhibits a wide angle of 135.4(3)°, t 243 whereas DmpN3 47 and ( BuMe2Ph)PhN3 48 have angles of 122.1(1)° and 119.8(2)° 244 respectively. The N(3)-N(2)-N(1) of Dipp*N3 45 is slightly distorted away from linear at 170.3(5)°. This is also within the range reported for the azides in Figure 29, which range from 168.0(3)° to 173.5(7)°.242-244

Figure 28 Molecular structure of Dipp*N3 45, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms

omitted for clarity. Selected bond lengths (Å) and angles (°): N(1)-C(1) 1.463(5), N(1)-N(2) 1.252(5), N(2)-N(3)

1.139(5), N(3)-N(2)-N(1) 170.3(5), N(2)-N(1)-C(1) 119.0(4).

40

Figure 29 Relevant aryl azides for structural comparison.

Initial attempts at the synthesis of 2,6-dibenzhydryl-4-methyliodobenzene, Dipp*I 49 focused on the reaction conditions that were successful for Dipp*N3 45 (vide supra). These conditions gave low yields of aryl iodide 49 (Scheme 32) and equal amounts of deaminated product, arene 50. It is proposed that this arene co-product originates from a reaction between the Dipp* aryl radical, generated by the loss of dinitrogen from the diazonium salt 44 and the tetrahydrofuran solvent.246 The iodide nucleophile is also likely to be incompatible with the excess sulfuric acid used due to the documented ability of iodide to reduce sulfuric acid. Aryl iodide 49 and arene 50 were separated by multiple fractional recrystallisations from ethyl acetate/hexane mixtures for the purposes of characterisation but not in yields amenable to bulk scale preparations of Dipp*I 49. Separation of iodide 49 from arene 50 using flash column chromatography was also made difficult by the extreme acid sensitivity of iodide 49. This caused considerable deiodination to occur on silica, which was observed as a pink colouration of early column fractions, serving to increase the yield of arene 50. Yields of ca. 20% were typically obtained after chromatography on de-activated silica and two recrystallisations from ethyl acetate/hexane.

Scheme 32 Synthesis of aryl iodide, Dipp*I through diazotisation in THF

1 The H NMR spectrum of Dipp*H 50 in CDCl3 has a resonance at δ 2.23 that corresponds to the methyl group. The benzylic and 3,5- aromatic protons correspond to signals that are observed at δ 5.45 and 6.79. The signal corresponding to 2- aromatic proton is observed at δ 6.76. The IR spectrum of Dipp*H 50 is devoid of the absorbance from the amino functional group, which confirms deamination. The molecular structure

41 of Dipp*H 50 as determined from a single crystal X-ray structural determination is included in Section 2.4.

Aryl iodide 49 may instead be synthesised readily in an acetic acid/toluene mixture at 0 °C (Scheme 33). Complete reaction of aniline 43 was observed within 30 minutes rather than the 4 hours required in tetrahydrofuran-water mixtures. After aqueous work-up, the crude product was triturated with a boiling ethyl acetate/hexane mixture and washed with cold diethyl ether to afford Dipp*I 49 as a white powder in 56% yield. Iodide 49 gives satisfactory microanalysis and melts at 163-165 °C. No N-H stretch is observed in the IR spectrum of Dipp*I 49, which confirms the successful diazotisation. The 1H * NMR spectrum of Dipp I in C6D6 also shows the absence of the amino functional group of the aniline 43 at δ 3.15. The resonance ascribed to the para-methyl group is observed at δ 1.70 (Dipp*NH2 43, δ 1.92, Dipp*N3 45 δ 1.77). Singlet resonances attributed to the benzylic and meta-aromatic protons are observed at δ 6.24 and 6.85 respectively. Both signals are shifted downfield relative to aniline 43 by values of δ 0.78 and δ 0.12 respectively.

Scheme 33 The synthesis of Dipp*I in an acetic acid/toluene mixture.

Aryl iodide 49 crystallises as colourless square columns from an ethyl acetate hexane mixture in the monoclinic space group P21/n with one full molecule in the asymmetric unit. An illustration of the structure may be seen in Figure 30. The C(1)-I(1) (2.103(2) Å) distance is consistent with other structurally related aryl iodides 51-54.247-253 These distances have been observed to range from 2.105(2) to 2.135(5) Å.247,250 In the solid state Dipp*I 49 adopts the same head to tail zig-zag arrangement as iodobenzene 52. In that instance this was interpreted as intramolecular dipole-dipole head to tail interactions.247 The I(1)-H(8A) distance in Dipp*I is 3.384 Å. In iodobenzene 52 the equivalent distance is 3.155 Å,247 it is likely that this longer distance in the structure of Dipp*I is a consequence of the larger steric demand of the diphenylmethyl groups.

42

Figure 30 Molecular structure of Dipp*I 49, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. Selected bond length (Å): I(1) C(1) 2.103(2).

Figure 31 Structurally related aryl iodides.247,250

Reported syntheses of triazenes with similarly sterically demanding N-aryl substituents involve the treatment of an iodoarene with nbutyllithium and reacting the lithium aryl in situ with an aryl azide (Scheme 28b).203,204,225 In this instance the benzylic protons of iodide 49 are acidic enough to be deprotonated by alkyllithium bases at room temperature (Scheme 34, left). This undesired side-reaction is indicated by a deep red colour in solution, indicative of a trityl anion. Immediately quenching a red solution of trityl anion 55 prepared from Dipp*I 49 and nbutyllithium at room temperature with 1 D2O followed by analysis of the crude product by H NMR spectroscopy in C6D6 indicates deuteration at the benzylic position through a reduced integral value of the signal corresponding to this position. Instead aryl iodide 49 was converted to lithium aryl 56 using treatment with nbutyllithium at -80 °C in tetrahydrofuran solution and warming to 0 °C over three hours (Scheme 34, right).

Scheme 34 The reaction of Dipp*I with nbutyllithium at room temperature (left) and at -80 °C (right).

43

Treatment of a tetrahydrofuran solution of lithium aryl 56 as generated above at -80 °C with a tetrahydrofuran solution of aryl azide 45 at -80 °C followed by slow warming to room temperature and then stirring at room temperature or reflux and then quenching with water both failed to give triazene 39 after work-up (Scheme 35). Analysis of the 1 crude reaction products using TLC and H NMR spectroscopy in C6D6 was consistent with a mixture of Dipp*H 50 and Dipp*N3 45 as the major identifiable products.

Infrared spectroscopy also confirmed the presence of Dipp*N3 45 by the retention of the strong absorbance at 2126 cm-1.

Scheme 35 Attempted synthesis of aryl triazene Dipp*2N3H 39 from Dipp*Li 56 and Dipp*N3 45. The reaction of nBuLi with Dipp*I 49 was investigated further in order to gain some insight into the lack of reactivity of the Dipp*Li 56 intermediate with Dipp*N3. Accordingly nBuLi was added to a solution of Dipp*I 49 at -90 °C in tetrahydrofuran and the reaction quenched by the addition of water at -10 °C. Purification of the crude products by flash column chromatography afforded the anticipated Dipp*H 50 as the major product (45%) as well as Dipp*nBu 57 (27%) (Scheme 36). The 1H NMR n spectrum of Dipp* Bu 57 in CDCl3 is very similar to that of Dipp*H 50. The benzylic protons are observed at δ 5.77 and the meta-aromatic protons at δ 6.58. The butyl chain is observed as three ill-defined multiplets over the ranges δ 0.87-0.92, 1.31-1.51 and 2.49-2.54. These correspond to the terminal methyl group, two methylene groups and the aryl bound methylene group respectively. The para-methyl group corresponds to the signal observed at δ 2.11.

Scheme 36 The synthesis of Dipp*nBu 57 and Dipp*H 50 from Dipp*I 49 and nBuLi.

Upon re-examination of 1H NMR spectra of crude reaction mixtures from the attempted triazene syntheses (Scheme 35) Dipp*nBu 57 was also identified as one of the minor 44 products from this reaction. It is apparent that intermediate Dipp*Li 56 reacts with the iodobutane electrophile, generated by lithiation at the 2-position before the addition of the desired azide electrophile 45. This prevents the synthesis of the triazene by this method.

In light of this result, three alternative methods for the synthesis of Dipp*2N3H 39 were considered (Scheme 37). The first was the use of two equivalents of tBuLi as a replacement for nBuLi (Scheme 37a). The second equivalent of tBuLi is sufficiently reactive at low temperatures to induce of the elimination of lithium iodide from the tBuI by-product to give isobutane, isobutene and lithium iodide as the by-products of the exchange reaction.254 The second alternative was the direct reaction of Dipp*I 49 with lithium metal in tetrahydrofuran at room temperature (Scheme 37b). The final alternative considered was the reaction of Dipp*I 49 with excess magnesium to generate the magnesium aryl 58. Magnesium aryls have also been demonstrated to be efficient reagents for the synthesis of sterically demanding triazenes.202

Scheme 37 Proposed alternative methods of triazene synthesis.

Method a) was initially discounted due to the increased possibility of side-reactions between the alkyllithium base and the relatively acidic benzylic protons. Initial investigations into method b) did not prove fruitful. Sonication of a suspension of excess lithium powder and Dipp*I 49 in tetrahydrofuran for extended periods gave no evidence of conversion to the desired aryllithium species. The 1H NMR spectrum of the involatile materials obtained after work-up were consistent with the starting material This method was therefore abandoned in favour of method c).

45

A suspension of excess magnesium metal and Dipp*I 49 in tetrahydrofuran was heated for 18 hours, excess magnesium was removed by filtration. Dipp*N3 45 was added to the golden yellow solution of the magnesium aryl and the resulting solution was heated at reflux for a further 18 hours. The reaction was quenched by the addition of water and

Dipp*2N3H 39 isolated in moderate (58%) yield after removal of unconverted starting material by filtration and washing of the triazene that precipitates upon concentration (Scheme 38). Triazene 39 possesses high solubility in donor solvents such as tetrahydrofuran and moderate solubility in arenes but low solubility in other solvents such as ethyl acetate and . This simplifies the purification of preparations conducted in bulk.

Scheme 38 The synthesis of Dipp*2N3H 39 via an arylmagnesium iodide.

1 The H NMR spectrum of Dipp*2N3H 39 in C6D6 (Figure 32) is unexpectedly complex relative to the starting materials and all other structurally related compounds synthesised as part of this thesis. At least three isomers are believed to be observed at room temperature. One resonance is observed at δ 1.85 which corresponds to the para-methyl group. The N-H proton is observed as three signals, with the major isomers resonance observed at δ 8.51 and the two minor isomers resonances observed at δ 7.42 and δ 9.63. All aromatic protons, including the meta-aromatic protons, which are usually distinct in Dipp* compounds of this type (vide supra) correspond to a broad (fwhm = 23 Hz) singlet at δ 6.98. The benzylic protons of the major are observed to be chemically inequivalent and correspond to two broad (fwhm = 27 Hz and 23 Hz) resonances at δ 5.76 and 6.34. The analogous resonances of the minor isomers are observed at δ 5.61 and 6.09 as well as δ 5.37 and 6.06. The isomeric ratio is approximately 93:6:1 but is difficult to measure accurately due to the extremely low value integrals for the minor isomers.

46

1 Figure 32 The H NMR spectrum of Dipp*2N3H 39 in C6D6.

1 In stark contrast to this, the H NMR spectrum of Dipp*2N3H 39 in CDCl3 displays a single symmetrical isomer. The para-methyl group singlet resonance is observed at δ 2.17 and the N-H proton resonates at δ 8.26. The signals corresponding to the benzylic protons are observed at δ 5.70 as a single broad (fwhm = 8 Hz) singlet. This contrasts the magnetic and likely chemical inequivalency observed for Dipp*2N3H 39 in the spectrum recorded in C6D6 (both spectra collected at 298K and 300.30 MHz). The remaining aromatic protons are observed as three broad (fwhm = 11-21 Hz) singlets at δ

6.67, 6.86 and 7.11. The broad signals observed in the spectra recorded in both C6D6 and CDCl3 are consistent with a high degree of rotational and conformational movement in solution. Similar signal widths are observed when recording spectra at higher magnetic field strengths (600.13 MHz). The N-H stretching mode is observed as a weak absorbance at 3295 cm-1 in the infrared spectrum. The strong stretch at 2126 cm-1 that correlates to the azide functional group of Dipp*N3 45 is no longer observed. The strong absorbances observed at 1481 and 1493 cm-1 are consistent with characteristic modes of protonated N,N’-disubstituted triazenes.201 Triazene 39 gives satisfactory microanalytical results and melts at 176-188 °C with concomitant decomposition.

Triazene 39 crystallises as colourless long rectangular plates from a saturated solution in an ethyl acetate/hexane mixture in the monoclinic space group C2/c. Half a triazene 47 molecule is observed in the asymmetric unit, whereby a twofold rotation bisects the N(2) position. An illustration of the structure may be seen in Figure 33. In contrast to the solid state structure of Dipp2N3H 40 (vide supra), no hydrogen bonding interaction is observed. This is likely a consequence of the greatly increased steric demand of the diphenylmethyl groups relative to the isopropyl functional groups. In the solid state

Dipp*2N3H 39 adopts the E-anti configuration. This is the common conformation observed for triazenes in the solid state with only one E-syn structure reported, 202 DmpN3HMes 59 (Figure 34). The four reported aryl triazenes with ortho aryl groups on at least one of the N-aryl substituents crystallise with one full molecule in the asymmetric unit and exhibit localised double and single bonds in the solid state, with 255 the exception of one, Dmp2N3H 60 (Figure 34). This contrasts with Dipp*2N3H 39, which exhibits symmetry enforced averaging of the N-N bond lengths. The N(1)-N(2) bond length (1.2989(17) Å) is intermediate between reported nitrogen-nitrogen single and double bond lengths for sterically demanding aryl triazenes, which typically vary from 1.353-1.319 Å and 1.255-1.280 Å respectively.184,202,203,255 All other bond lengths and angles are consistent with those reported for other aryl triazenes such as Dmp2N3H 184,202,203,255 and DmpN3HMes 59-60 (Figure 34).

48

Figure 33 Molecular structure of Dipp*2N3H 39, POV-RAY illustration, 50% thermal ellipsoids, all non-amino or symmetry generated amino hydrogen atoms omitted for clarity. H(1) is included as 50% occupancy with its omitted symmetry equivalent H(1)# also 50% occupancy. The diphenylmethyl carbon atoms are represented as wireframes for clarity. Selected bond lengths (Å) and angles (°): N(1)-C(1) 1.434(2), N(1)-N(2) 1.2989(17), N(1)-N(2)-N(1)# 1 112.1(2), N(2)-N(1)-C(1) 116.19(14). Symmetry operation used to generate equivalent atoms: -x, y, /2-z.

202,255 Figure 34 Reported triazenes for comparison to the solid state structure of Dipp*2N3H.

2.2.2 Attempts at the Synthesis of a Formamidine from Dipp*NH2

High yielding single-step formamidine syntheses are typically conducted thermally in the absence of solvent as one or both starting materials are liquids and the alcohol co- product provides a further reaction medium (Scheme 27a).256 This is not suitable as a method for aniline 43 which has a melting point of 180 °C and a high molecular weight, both serving to limit solubility in a half equivalent of triethylorthformate.236 Moreover aniline 43 does not react with triethylorthoformate in ethanol, or in neat triethylorthoformate, at reflux for 48 hours in the presence of catalytic acetic acid 49

(Scheme 39). In both instances aniline 43 was observed to be the only non-volatile material after work-up as determined using TLC and 1H NMR spectroscopy (Scheme 39).

Scheme 39 The failed synthesis of Dipp*FormH 41 from Dipp*NH2 43 (For conditions see Table 2). Reasoning that more forcing conditions were required due to the large steric demand of aniline 43, a mixture of triethylorthoformate and two equivalents of aniline 43 was heated in the presence of catalytic acetic acid in a sealed tube at 200 °C for 24 hours. Under these conditions only trace quantities of complex mixtures other than the starting material were identified using 1H NMR spectroscopy. Some of the conditions attempted are summarised in Table 2

Solvent Temperature/ °C Catalyst Result Time/ h CH COOH 18 EtOH 78 3 No Reaction HC(OEt) CH COOH 48 3 140 3 No Reaction CH COOH 48 None/Melt 200 3 No Reaction H SO 72 o-Xylene 140 2 4 Formylation Table 2 Attempted formamidine synthesis conditions.

Aniline 43 reacts with triethylorthoformate in the presence of catalytic sulfuric acid in o-xylene at reflux to afford ethyl formamide Dipp*N(CHO)(Et) 61 in excellent yield (Scheme 40) without generation of the intended formamidine 41. In this instance the sulfuric acid acts as a dealcoholating agent as well as an acid catalyst, similar reactions have been observed previously with smaller aniline precursors.257

Scheme 40 The synthesis of ethyl formamide 61.

The infrared spectrum of ethyl formamide 61 does not display any absorbances above 3000 cm-1 attributable to N-H stretching modes in aniline 43. The infrared spectrum of

50 formamide 61 exhibits a strong absorbance at 1676 cm-1 that is attributed to the formyl 1 group C=O stretch. The H NMR spectrum of formamide 61 in C6D6 is unexpectedly complicated, exhibiting double the expected signals for a symmetrical system. The integral ratios support the presence of two isomers in a ratio of 4:1. It is proposed that these arise from restricted rotation about the formyl N-C bond (Figure 35). This results in two rotamers that do not readily interconvert and thus are observable on the 1H NMR time-scale. Formamide 61 melts at 215-217 °C and gives satisfactory microanalytical results.

Figure 35 The proposed equilibrium in ethyl formamide 61 leading to the rotamers observed on the 1H NMR timescale.

Ethyl formamide 61 crystallises as colourless cubes from o-xylene in the monoclinic space group P21/c with one full molecule in the asymmetric unit. An illustration of the molecular structure may be seen in Figure 36. The terminal carbon of the ethyl group is disordered across two locations C(35A) and C(35B) in a 63:37 ratio respectively. In the solid state the N-substituents are rotated orthogonal to the plane of the central arene ring. The planar nitrogen atom is consistent with the only other reported aryl ethyl formamide.258 In this orientation, the preferred rotation of the formyl group about the bond N(1)-C(36) is easily visualised as anti, presumably due to steric buttressing between the formyl group and the diphenylmethyl arene rings. This may be seen in Figure 37, which overlays the van der Waals radii of each atom and from which the formyl C=O is clearly hindered from assuming the syn conformation.

51

Figure 36 Molecular structure of Dipp*N(CHO)(Et) 61, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. Disordered atom of lower occupancy included with hatched bond (C35B). Selected bond lengths (Å) and angles (°): N(1)-C(1) 1.598(6), N(1)-C(34) 1.628(7) N(1)-C(36) 1.305(6), C(36)-O(1) 1.347(7), N(1)- C(36)-O(1) 120.7(6).

Figure 37 Molecular structure of Dipp*N(CHO)(Et) 61, with van der Waals radii overlayed.

Based upon the outcomes of the reactions summarised in Table 2 it was concluded that the high steric demand of Dipp*NH2 43 prevents the reaction of the amine nucleophile with the moderately congested methine carbon of triethylorthoformate. Thus trimethylorthoformate was considered owing to its reduced steric congestion at the electrophilic carbon. However, heating Dipp*NH2 43 at reflux in excess trimethylorthoformate in the presence of catalytic acetic acid afforded the methyl formimidate 62, in 98% yield and not the desired formamidine 41.

52

Scheme 41 The synthesis of Dipp*NCH(OMe) 62.

The absorbance characteristic N-H stretching modes as observed for Dipp*NH2 43 are absent in the infrared spectrum of methyl formimidate 62. The 1H NMR spectrum of

Dipp*NCH(OMe) 62 in C6D6 exhibits singlet resonances that correspond to the para- methyl and methyl ether groups at δ 1.92 and 3.52 respectively (cf. para-methyl resonance of Dipp*NH2 43 at δ 1.92). The benzylic protons and the aldimine proton correspond to signals observed at δ 5.78 and 6.35 respectively (Dipp*NH2 43 δ 5.46 and 6.73 respectively).

Methyl formimidate 62 crystallises from hexane as colourless rectangular blocks in the orthorhombic space group Pnma. One half of the molecule is included in the asymmetric unit with the remaining half generated by a mirror plane that bisects the formimidate group and the C(1), C(4) and C(18) atoms. An illustration of the molecular structure may be seen in Figure 38. The bond localisation is clearly evident between the single carbon-oxygen bond (O(1)-C(19) 1.340(3) Å) and carbon-nitrogen double bonds (N(1)-C(19) 1.425(3) Å).

Figure 38 Molecular structure of Dipp*NCH(OMe) 62, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. Selected bond lengths (Å) and angles (°): N(1)-C(19) 1.251(3), N(1)-C(1) 1.425(3), C(19)- O(1) 1.340(3), O(1)-C(20) 1.430(4), C(19)-N(1)-C(1) 118.1(2), C(19)-O(1)-C(20) 114.6(2). Symmetry operation 3 used to generate equivalent atoms: x, /2-y, z. 53

Alkyl formimidates like Dipp*NCH(OMe) 62 are usually isolated through the use of catalytic hydrochloric acid in the presence of excess triethylorthoformate. This kind of precursor is often used in order to synthesise N,N’-asymmetrically substituted formamidines.259 It was anticipated that such alkyl formimidates could allow the addition of a second equivalent of amine as a nucleophile and the desired formamidine to be synthesised. With this in mind Dipp*NH2 43 and excess triethylorthoformate were heated at reflux in the presence of catalytic hydrochloric acid in a distillation apparatus to remove ethanol and drive the reaction to completion. This afforded the ethyl formimidate Dipp*NCH(OEt) 63 in 83% yield (Scheme 42).

Scheme 42 The synthesis of ethyl formimidate 63.

Ethyl formimidate 63 is differentiated from Dipp*NH2 43 by the disappearance of absorbances that are attributed to the amino group in the starting material. The 1H NMR spectrum of formimidate 63 in C6D6 displays the expected triplet quartet pattern for an ethyl ether group at δ 1.28 and 4.09. The para-tolyl methyl resonance is observed at δ

1.92 (cf. Dipp*NH2 43 δ 1.92). The benzylic protons and the meta-aromatic protons are observed at δ 5.81 and 6.36 respectively. These are shifted downfield from δ 5.46 for the benzylic protons and δ 6.73 ppm for the meta-aromatic protons of the parent aniline 43. The aldimine proton is observed at δ 6.36 ppm, which is similar to the equivalent resonance for Dipp*NCH(OMe) 62 (δ 6.35).

Ethyl formimidate 63 crystallises directly from triethylorthoformate as colourless blocks in the triclinic space group P1¯. An illustration of the structure may be seen in Figure 39. In contrast with methyl formimidate 62, one full molecule is observed in the asymmetric unit. Like Dipp*NCH(OMe) 62 clear bond localisation is observed for the nitrogen carbon double bond (N(1)-C(34) 1.2636(18) Å, Dipp*NCH(OMe) 62 1.251(3) Å) and the carbon oxygen single bonds (C(34)-O(1) 1.3305(16) Å, Dipp*NCH(OMe) 62 1.340(3) Å).

54

Figure 39 Molecular structure of Dipp*NCH(OEt) 63, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. Selected bond lengths (Å) and angles (°): N(1)-C(1) 1.4276(17), N(1)-C(34) 1.2636(18), C(34)-O(1) 1.3305(16), O(1)-C(35) 1.429(2), C(34)-N(1)-C(1) 114.82(12), C(34)-O(1)-C(35) 117.11(12).

Ethyl formimidate 63 was heated at reflux in o-xylene for 48 hours with one equivalent * of Dipp NH2 43 (Scheme 43). No conversion of the starting materials was observed using TLC.

Scheme 43 The attempted reaction of ethyl formimidate 63 with Dipp*NH2 43.

* Reasoning that the steric demand of a second equivalent of Dipp NH2 43 may be too high for nucleophilic attack to occur at the desired position, smaller amine nucleophiles were examined for reactivity with formimidates 62 and 63. Mixtures of either 2,6- diisopropylaniline or isopropylamine with formimidate 63 were heated for 18 hours in ethanol at 50 °C (Scheme 44). Analyses of the involatile materials after work-up using 1H NMR spectroscopy afforded spectra consistent with either a mixture of the starting materials (2,6-diisopropylaniline) or just formimidate 63 (isopropylamine). It is presumed that the high volatility of isopropylamine (b.p. 33-34 °C/ 1 bar) caused it to be lost during work-up and thus no reaction occurred in either case.

55

Scheme 44 Attempted synthesis of asymmetrically substituted formamidines.

Formamidines may also be prepared by desulfurisation of symmetric thioureas using either nickel boride,260 nickel- alloys,261 or palladium catalysed reduction using triethylsilane and trimethylsilyltriflate.262 Thioureas may also be desulfurised over mercuric oxide and magnesium sulfate263 or with triphenylphosphine and carbon tetrachloride264 to afford carbodiimides, which may be hydrogenated over palladium on barium sulfate to give formamidines.265 Carbodiimides are also the precursor to 166 guanidinate ligands (Scheme 27c). As such, the reaction of Dipp*NH2 43 with one half equivalent of CS2 in the presence of Et3N was examined (Scheme 45). Heating in dioxane at reflux for 48 hours gave no conversion as determined using TLC and 1H NMR spectroscopy conducted on the reaction mixture.

Scheme 45 The attempted synthesis of a thiourea.

Structurally related ureas may also be dehydrated over a large excess of diphosphorus pentoxide in pyridine to give carbodiimides.266 During the preparation of this thesis the synthesis of a structurally related urea and carbodiimide was disclosed.267 In this instance the latter was employed as a precursor to an extremely sterically demanding guanidinate and the lithium, potassium and caesium complexes thereof (Scheme 46). It was anticipated that application of this method to Dipp*NH2 43 would yield the structurally analogous carbodiimide to that reported by Fortier267 and that the application of the reduction methods described above was reasonably anticipated to give the desired formamidine.

56

Scheme 46 The literature synthesis of an extremely sterically demanding guanidinate ligand.

Following the published procedure (Scheme 46)267 with modification, a suspension of

Dipp*NH2 43, 4-dimethylaminopyridine and a sub-stoichiometric amount of triphosgene was heated at 60 °C for 15 days in a pressure vessel in (Scheme 47). This procedure gives (Dipp*NH)2CO 64 in 80% yield. This yield is comparable to that reported by Fortier for the similar urea 65 (86%).267 Extended reaction times do not improve the yield, with the mass balance accounted for by the recovered aniline. It has also been found to be unnecessary to cool the reaction to -78 °C during the addition of triphosgene. Similar reaction outcomes are observed by adding triphosgene at -20 °C.

Scheme 47 The synthesis of urea (Dipp*NH)2CO 64.

(Dipp*NH)2CO 64, is isolated as a very fine white powder that is extremely susceptible to retention of static charge. (Dipp*NH)2CO 64 is extremely thermally stable and does 1 * not melt below 250 °C. The H NMR spectrum of (Dipp NH)2CO 64 in C6D6 displays a signal attributable to the methyl group at δ 1.82 (Dipp*NH2 43 δ 1.92). The carbamide protons resonate at δ 4.55, significantly downfield of the resonance for the amino protons in Dipp*NH2 (δ 3.15). The signal corresponding to the benzylic protons is observed substantially downfield at δ 6.28. This has shifted downfield from δ 5.46 in the starting material. The signal that corresponds to the meta-aromatic protons has

57

* shifted downfield from δ 6.73 in Dipp NH2 to δ 6.92. An extremely broad (fwhm = 60 Hz) unresolved multiplet is observed at low field (δ 7.25). This corresponds to some of the aromatic protons and is consistent with the spectra observed by Fortier et al. for 267 13 * their related urea 65. The C NMR spectrum of (Dipp NH)2CO 64 in C6D6 contains a resonance at δ 154.62 that may be attributed to the carbonyl carbon. The infrared spectrum of (Dipp*NH)2CO 64 displays two strong stretches that are also suggestive of a urea at 3381 and 1702 cm-1. The former is attributed to the carbamide N-H absorbance and the latter to the carbonyl C=O absorbance.

(Dipp*NH)2CO 64 crystallises as large colourless blocks from a tetrahydrofuran n solution layered with hexane over one week in the monoclinic space group P21/c. It exhibits one full urea molecule and four tetrahydrofuran molecules in the asymmetric unit. An illustration of the structure may be seen in Figure 40. The solvent molecules are heavily disordered and could not be successfully modelled. The SQUEEZE234 function of PLATON235 was used to successfully refine the structure. The unaccounted electron density and solvent accessible voids are consistent with four tetrahydrofuran molecules in the unit cell. This is supported by the 1H NMR spectrum of the isolated crystals, in which resonances due to free tetrahydrofuran may be observed in the correct integral ratios. Attempts to model the suggested disorder resulted in unsatisfactory thermal parameters and where therefore left as is. The structural parameters compare favourably with those in the reported analogue 63.267 The N(1)-C(1) and N(2)-C(1) bonds (1.348(4) and 1.338(4) Å) are shorter (1.3714(19) Å) than the equivalent bond in the reported analogue as may be expected given the substitution of a 4-tertbutyl group for a 4-methyl group. The carbon oxygen double bond (C(1)-O(1) 1.216(3) Å) is statistically the same as the analogous bond in 65 (1.214(3) Å). The N(1)-C(1)-N(2) angle (116.2(3)°) is slightly more obtuse than that of the reported urea 65 (113.1(2)°), which may again be a consequence of the change in respective principal arene 4-alkyl groups. These differences are minor and can be accounted for by the different spatial arrangement of the N-aryl substituents in the solid state structures of the two compounds.

58

Figure 40 Molecular structure of (Dipp*NH)2CO 64, POV-RAY illustration, 40% thermal ellipsoids, all hydrogen atoms omitted for clarity. The diphenylmethyl carbon atoms are represented as wireframes for clarity. Selected bond lengths (Å) and angles (°): N(1)-C(1) 1.348(4), N(2)-C(2) 1.338(4), C(1)-O(1) 1.216(3), C(1)-N(1)-C(2) 123.1(3), N(1)-C(1)-N(2) 116.2(3), C(1)-N(2)-C(35) 123.3(3).

Urea 64 may be dehydrated to give carbodiimide Dipp*NCNDipp* 66 under the forcing conditions reported by Fortier et al. with modification (Scheme 48),267 wherein the pressure vessel apparatus is substituted for a regular reflux apparatus. This potentially enables larger scale preparations. The yield of 73% attained for this procedure is comparable to the published procedure for carbodiimide 67 (81%).267 It has been found to be absolutely essential to use extremely vigorous stirring and the largest possible stirrer bars in this preparation due to the tendency of the suspended solids to form a solid mass at the bottom of the reaction vessel. In preparations where this occurred incomplete dehydration was observed.

Scheme 48 The synthesis of carbodiimide Dipp*NCNDipp* 66.

Carbodiimide 66 was isolated as an analytically pure white powder after recrystallisation from chloroform. The 1H NMR spectrum of Dipp*NCNDipp* 66 in 59

C6D6 is devoid of the protons that were previously attributable to the carbamide functional groups. The methyl group is observed as a singlet at δ 1.82 (cf.

(Dipp*NH)2CO 64 δ 1.82). The benzylic proton is matched to a signal that appears at δ 6.01. This is upfield from the equivalent resonance (δ 6.28 ppm) in urea 64. The meta- aromatic protons are assigned to the signal that is observed at δ 6.86, an upfield shift 13 from the starting material (δ 6.92). The C NMR spectrum of Dipp*NCNDipp* in C6D6 displays a signal that resonates at δ 143.82, which is consistent with a quaternary 267 carbodiimide carbon atom (cf. (Dipp*NH)2CO 64 δ 154.62). The infrared spectrum of Dipp*NCNDipp* 66 does not show absorbances attributable to the carbamide nor the carbonyl functional groups observed in the parent urea 64 (3381 and 1702 cm-1). Instead a strong sharp absorbance at 2164 cm-1 is observed. This is consistent with the presence of a carbodiimide functional group (cf. DippNCNDipp 2169 cm-1).268

1 It is worth noting that the H NMR spectra of both (Dipp*NH)2CO 64 and

Dipp*NCNDipp* 66 in C6D6 do not display the three isomers that were observed for

Dipp*2N3H 39. With the exception of the broad resonance observed for some of the arene protons in (Dipp*NH)2CO 64 the resonances for both these compounds are also sharp in C6D6, which contrasts with the broad resonances observed for Dipp*2N3H 39. It is believed that the greater rigidity of the RNH(C=O)NHR and N=C=N moieties in urea 64 and carbodiimide 66 in comparison to RN=N-NHR moiety in triazene 39 leads to a greatly reduced degree of conformational freedom in these systems.

Carbodiimide 66 crystallises as colourless rods from toluene in the monoclinic space group C2/c. One half of the molecule is included in the asymmetric unit (Figure 41) with the remaining half generated by a two-fold C2 rotational symmetry that bisects the C(1) carbon atom. The principal arenes are observed to twist orthogonal to each other in the solid state. This contrasts with the literature carbodiimide 67, in which the central arene rings lie in the same plane as each other in the solid state.267 The N(1)-C(1)-N(1)# bond is also not perfectly linear in the solid state (166.6(3)°). This contrasts with the aforementioned carbodiimide 67, which is linear (180.0°) but is similar to DippNCNDipp (169.3(2)°).263 The C(1) carbon is accessible to a suitable nucleophile as observed in a space-filling representation of carbodiimide 67 (Figure 42). The closest contact between the arene rings is between the meta- (6.517 Å); this is displayed as a red line in Figure 42. The distance between the arene centroids of the

60 associate rings is substantially longer (7.555 Å). It was anticipated that this void was large enough to allow the reduction of one of the carbon-nitrogen double bonds .

Figure 41 Molecular structure of Dipp*NCNDipp* 66, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. Selected bond lengths (Å) and angles (°): N(1)-C(1) 1.2104(16), N(1)-C(2) 1.414(2), C(1)- 1 N(1)-C(2) 136.75(17), N(1)-C(1)-N(1)# 166.6(3). Symmetry operation used to generate equivalent atoms: -x, y, /2-z.

Figure 42 Space filling representation of carbodiimide 66 with van der Waals radii overlayed showing the accessible NCN bonds.

It was anticipated that the treatment of Dipp*NCNDipp* 66 with one equivalent of a hydride would give the metal formamidinate intermediate, which upon alcoholysis

61 would give the desired formamidine Dipp*FormH 41 (

Scheme 49).

Scheme 49 The proposed synthesis of Dipp*FormH 41 by hydride reduction.

A range of borohydrides were examined for this reaction. The reaction of Dipp*NCNDipp* 66 with one equivalent of potassium or sodium triethylborohydride in tetrahydrofuran at room temperature for 18 hours afforded only the starting material after quenching as determined using TLC and 1H NMR spectroscopy on the crude reaction mixture. Heating this reaction to 60 °C for 4 days followed by methanolysis gave a mixture of Dipp*NCNDipp* 66 and Dipp*NH2 43 upon work-up as determined using TLC and 1H NMR spectroscopy in a roughly 1:2 ratio respectively. It was therefore reasoned that the triethylborohydride reagents may be too strong as a hydride addition reagents to selectively affect this reduction. Thus one equivalent of borane dimethyl sulfide complex was added to a solution of Dipp*NCNDipp* 66 in tetrahydrofuran at room temperature. Methanolysis after 18 hours gave similar results to the attempted reduction with triethylborohydride.

Scheme 50 The attempted reduction of Dipp*NCNDipp* with triethylborohydride or borane reagents.

At this point it was reasoned that over-reduction may be occurring due to the low steric demand of the hydride reagents used for this reaction. The highly sterically demanding borane 9-borabicyclo[3.3.1]nonane (9-BBN) has been successfully used for the selective hydroboration of dicyclohexylcarbodiimide 68 (Scheme 51).269 It was hoped 62 that this would allow the selective hydroboration of carbodiimide 66 to afford formamidine 41 in this instance. The reaction of one equivalent of 9-BBN with Dipp*NCNDipp* 66 in toluene at 100 °C for 48 hours followed by methanolysis gave a * mixture of Dipp*NCNDipp* 66 and Dipp NH2 43 in a roughly 1:2 ratio respectively as determined using TLC and 1H NMR spectroscopy. At this point attempts to selectively reduce carbodiimide 66 were abandoned.

Scheme 51 The hydroboration of dicyclohexylcarbodiimide with 9-BBN.269

As a final synthesis strategy towards formamidine 41 it was proposed that the insertion of an isonitrile into a metal amide bond could yield the desired compound (Scheme 52). The insertion of isonitriles into alkali metal-alkyl bonds has precedent270,271 and the insertion of nitriles into alkali metal-amides is the standard method for the synthesis of silylamidinates.172 It should be noted that there are stable alkali metal isonitrile complexes that have been reported272,273 however the limited number of reports were not taken as an indicator of the unviability of this proposed method. The synthesis of Dipp*NC 69 was therefore examined.

Scheme 52 Proposed reaction mechanism for the insertion of an isonitrile into an alkali metal-amide bond.

Initially the Hofmann isonitrile synthesis was considered due to the fact that it would 274,275 provide the desired isonitrile 69 in a single step. The reaction of Dipp*NH2 43 with dichlorocarbene, generated in situ from chloroform, under strongly basic phase- transfer conditions in a dichloromethane water mixture gives Dipp*NC 69 in low yield (Scheme 53). This reaction gives a large number of, thus far, unidentified side products and requires purification by flash column chromatography at high silica to compound

63 ratios (100:1 by mass). This prevents time-efficient large scale preparations from being conducted. It has also been observed to give highly variable results upon repetition, with some preparations giving much lower yields of the desired isonitrile 69. This variability in outcomes greatly hindered the effective optimisation of the reaction despite considerable efforts to do so.

Scheme 53 The synthesis of DippNC 69 under phase-transfer conditions.

1 The H NMR spectrum of Dipp*NC 69 in C6D6 is devoid of the amine proton resonance of Dipp*NH2 43. The singlet resonance attributable to the methyl group shifts upfield from δ 1.92 in Dipp*NH2 to δ 1.67 in Dipp*NC. The singlet resonance for the benzylic proton shifts downfield from δ 5.46 to 6.04 and the meta-aromatic proton singlet resonance shifts slightly from δ 6.82 to 6.73. In the infrared spectrum of Dipp*NC 69 the isonitrile functional group absorbance is observed as a strong sharp absorbance at 2109 cm-1. This absorbance is very similar to the reported values for related isocyanides -1 276 -1 240 such as DippNC 70 (2120 cm ), Dipp2PhNC 71 (2124 cm ) and Mes2PhNC 72 -1 277 13 (2120 cm ). In the C NMR spectrum of Dipp*NC 69 in C6D6 a resonance at low field (δ 173.36) can be attributed to the quaternary isonitrile carbon. This resonance is 277 further downfield from reported isonitriles Dipp2PhNC 71 (δ 171.9 in C6D6) and 240 Mes2PhNC 72 (δ 170.7 in C6D6).

Figure 43 Structurally related isocyanides.240,276-278

The low yield of Dipp*NC 69 under phase transfer conditions, as well as the difficulty of the chromatography required for purification, prompted the investigation of an alternative method of synthesis. Isonitriles like Dipp2PhNC 71 and Mes2PhNC 72 are readily prepared by dehydration of the corresponding N-formyl compounds.240,277 64

Following the conditions for the synthesis Mes2PhNH(CHO), Dipp*NH2 43 was heated at reflux in toluene with a vast excess of formic acid in a Dean-Stark apparatus for 60 h to yield Dipp*NH(CHO) 74 in good yield (Scheme 54). The reaction time is far longer 277 than the reported time (12 h) for PhMes2NH(CHO). This is suggestive of the fact that the amine functional group in Dipp*NH2 is more sterically protected than that of the reported terphenyl. This is supported by the fact that the synthesis of the more sterically hindered formyl precursor to Dipp2PhNC 71 requires acetic formic anhydride, a stronger electrophile than formic acid, in order to proceed.240

Scheme 54 The synthesis of DippNH(CHO) 74 using formic acid.

1 The H NMR spectrum of Dipp*NH(CHO) 74 in C6D6 displays similar complexity to that observed for Dipp*N(CHO)(Et) 61. This can be rationalised as a similar isomerism that arises from restricted rotation about the formyl N-C bond as was suggested for Dipp*N(CHO)(Et) 61. In this instance the solid state structure of Dipp*NH(CHO) 74 has not been determined and so this cannot be confirmed. However the sterically less congested compound DippNH(CHO) has a similar reported rotational isomerism observed in its 1H NMR spectrum.279 The resonance for the amine proton of the major isomer is observed at δ 5.26. This is substantially downfield of the signal of the starting material (δ 3.15). The formyl proton for the major isomer is observed as a doublet resonance at δ 7.51. This may be ascribed to a syn isomer on the basis of the smaller coupling constant (J = 1.3 Hz) for this signal relative to the minor anti isomer (δ 7.61, J 1 = 11.6 Hz) (Figure 44). The H NMR spectrum of Dipp*NH(CHO) 74 in CDCl3 displays a similar syn/anti isomerism, however the relative ratios of each isomeric component are different from the spectrum recorded in C6D6. In the former a ratio of 56:44 is observed, whereas in the latter the ratio is approximately 76:24. The infrared spectrum of Dipp*NH(CHO) 74 displays an absorbance at 1695 cm-1, which is consistent with the carbonyl stretch. The amide N-H bond corresponds to several absorbances in the range 3026-3163 cm-1.

65

Figure 44 The isomeric components observed in the 1H NMR spectrum of Dipp*NH(CHO) 74.

Formanilide 74 is readily dehydrated by treatment with excess POCl3 and diisopropylamine in dichloromethane at room temperature. This affords the desired isonitrile after hydrolysis in excellent yield (Scheme 55). The spectroscopic properties are identical with the material synthesised by phase transfer conditions (Scheme 53) but this route is much more convenient for large scale preparations due to the simple work- up. The overall yield from the two steps (58%) is also significantly higher. Unfortunately, this synthesis was developed late in the authors candidature and thus the reactivity of Dipp*NC 69 with metal amides, as per Scheme 52 has not been investigated to date.

Scheme 55 The synthesis of Dipp*NC 69 by dehydration.

Dipp*NC 69 crystallises from toluene as colourless parallelogram plates in the monoclinic space group Cc with one full molecule in the asymmetric unit. A representation of the solid state structure may be seen in Figure 45. The isonitrile functional group is almost perfectly linear in the solid state (C(7)-N(1)-C(1) 178.54(18)°). This is consistent with the reported angles for other uncoordinated aryl isonitriles, which have been observed to range from 176.7(1)° for Dipp2PhNC 71 to perfectly linear (180.000(1)°) in Mes*NC 73.278 The isonitrile multiple bond (N(1)-C(7) 1.155(2) Å) is also consistent with analogous bond lengths at the shorter end of the reported values for such compounds. For example it is the same within error as those of 240 277 Dipp2PhNC 71 (1.158(2) Å) and Mes2PhNC 72 (1.159(2) Å).

66

Figure 45 Molecular structure of Dipp*NC 69, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. Selected bond lengths (Å) and angles (°): N(1)-C(7) 1.155(2), N(1)-C(1) 1.406(2), C(7)-N(1)-C(1) 178.54(18).

2.2.3 Synthesis of Alkali Metal Triazenides

Triazenes are readily deprotonated with a range of alkali metal bases, generating the corresponding alkali metal triazenides. Treatment of Dipp2N3H 40 with sodium or potassium tert-butoxide in toluene afforded the Lewis base free alkali metal triazenides

[NaN3Dipp2] 75 and [KN3Dipp2] 76 in excellent yields (Scheme 56). Identical reaction outcomes were also achieved when sodium or potassium bis(trimethylsilyl)amide were used as the base. [NaN3Dipp2] 75 was isolated as a colourless powder and [KN3Dipp2] 76 as a pale pink powder. Both compounds had negligible solubility in non-donor solvents such as toluene or nhexane but dissolved readily in donor solvents to give the Lewis base adducts of those solvents (vide infra). The 1H NMR spectrum of triazenides

75 and 76 in THF-d8 includes a doublet at δ 1.09 for sodium triazenide 75 and at δ 1.09 for potassium triazenide 76 that is attributed to the methyl groups of the isopropyl functional groups. The resonance that may be attributed to the methine proton of the isopropyl group is observed at δ 3.55 for sodium triazenide 75 and at δ 3.54 for potassium triazenide 76. No resonance is observed that may be attributed to the amino proton of triazene 40 (δ 9.01 in C6D6) confirming successful deprotonation. Signals corresponding to either tert-butanol or the reaction solvents are also not observed. The infrared spectrum of triazenides 75 and 76 do not have significant absorbances above 3000 cm-1 supporting the absence of an NH functional group. The stretches that are 67

-1 -1 observed at 1294 cm for [NaN3Dipp2] 75 and 1293 cm for [KN3Dipp2] 76 are consistent with a chelating coordination mode of the triazene (See Section 2.1.2).201 Unfortunately this bonding mode of the triazenide ligand in both potassium and sodium triazenides 75 and 76 cannot be confirmed using X-ray crystallography as the low solubility of the base-free triazenides 75 and 76 in non-donor solvents frustrates the isolation of single crystals. Triazenides 75 and 76 are extremely moisture sensitive as solids. Samples submitted for microanalysis were repeatedly consistent with partial hydrolysis of the desired compound. Both are very thermally robust complexes and do not decompose below 250 °C.

Scheme 56 The synthesis of Lewis base free alkali metal triazenides, [NaN3Dipp2] 75 and [KN3Dipp2] 76.

t Deprotonation of Dipp2N3H 40 with NaO Bu in tetrahydrofuran afforded

[Na(N3Dipp2)(thf)3] 77 as yellow-orange blocks in 66% yield (Scheme 57). In contrast to Lewis base-free [NaN3Dipp2] 75, triazenide 77 is extremely soluble in all common 1 laboratory solvents. The H NMR spectrum of [Na(N3Dipp2)(thf)3] in C6D6 does not display the signal that corresponds to the NH proton in triazene 40 (δ 9.01 in C6D6), confirming the successful deprotonation. The methyl and methine protons of the isopropyl functional group are observed at δ 1.33 and 3.76 respectively and have shifted upfield from δ 1.16 and 3.31 respectively in the parent triazene 40. The two signals that are attributed to the three coordinated thf ligands are observed at δ 3.46 ppm and 1.38 ppm as broad, unresolved multiplets, consistent with dynamic coordination behaviour in 280 solution (Free tetrahydrofuran in C6D6 δ 3.57 and 1.40). The infrared spectrum of -1 [Na(N3Dipp2)(thf)3] is distinguished by an absorbance at 1296 cm (cf.[NaN3Dipp2] 75 1294 cm-1). This is consistent with a chelating bonding mode of the triazenide ligand and supports that assignment in the base free complex 75 as well.201 No absorbances that may be attributed to an NH functional group are observed, again confirming successful deprotonation. Samples submitted for microanalysis were repeatedly inconsistent with the empirical formula due to the ease by which the tetrahydrofuran ligands may be lost (vide infra).

68

Scheme 57 The synthesis of [Na(N3Dipp2)(thf)3] 77.

In the solid state, the tetrahydrofuran ligands of [Na(N3Dipp2)(thf)3] 77 may be removed by heating. When samples of [Na(N3Dipp2)(thf)3] 77 were heated in sealed glass capillaries under argon to determine the solid state decomposition point, a colourless volatile liquid was observed to reflux in the capillary. This is first observed at temperatures above 70 °C, which is consistent with the boiling point of tetrahydrofuran (65-67 °C/ 1 bar). The pale yellow colour of triazenide 77 was also observed to fade upon heating to 70 °C, resulting in the formation of a colourless powder that is tentatively assigned as fully desolvated [NaN3Dipp2] 75. Slow decomposition was observed above 200 °C before violently melting above 300 °C. The discrepancy between this decomposition may be explained by the presence of a quantity of high temperature solvent vapour in the sealed capillary, which may present another mechanism of decomposition not available to desolvated [NaN3Dipp2] 75. Desolvation may also be achieved by extended (> 6 hrs) treatment of the tetrahydrofuran complex -1 with dynamic vacuum (ca. 1 x 10 mbar), which affords [NaN3Dipp2] 75 in quantitative yield.

Sodium triazenide 77 crystallises from nhexane as yellow plates in the monoclinic space group P21/c with one full molecule in the asymmetric unit. An illustration of the structure may be seen in Figure 46. Considering the triazenide ligand as a single donor located at the N(2) atom, the sodium of [Na(N3Dipp2)(thf)3] 77 is best described as adopting a heavily distorted tetrahedral geometry in the solid state. Only two of the angles associated with the geometry at the metal (N(2)-Na(1)-O(1) 108.43(13)°, O(1)- Na(1)-O(2) 106.91(13)°) approach the ideal tetrahedral value of 109.5°. The angle N(1)- N(2)-N(3) is 111.9(3)°. This is consistent with a triazenide ligand chelated to sodium. These angles have been observed to range from 109.2(8) to 110.8(9)° in the literature.185 This compound is essentially isostructural with the published sodium formamidinate 205 analogue, [Na(DippForm)(thf)3], 18 (Figure 18). The equal within error sodium to nitrogen bond lengths, Na(1)–N(1) (2.419(3) Å) and Na(1)–N(3) (2.411(4) Å) in triazenide complex 77 differ from those in the formamidinate analogue 18, which features a short sodium to nitrogen bond (Na(1)–N(1) 2.406(2) Å) and a long sodium to 69 nitrogen bond (Na(1)–N(2) 2.457(2) Å). The tetrahydrofuran ligands in triazenide 77 are bonded more closely to the sodium atom than in formamidinate 18. In triazenide 77 the distances are measured as Na(1)-O(1) 2.300(4) Å, Na(1)-O(2) 2.336(3) Å, Na(1)- O(3) 2.340(4) Å, whereas in formamidinate 18 they are Na(1)–O(2) 2.347(2) Å, Na(1)– O(1) 2.351(2) Å and Na(1)–O(3) 2.376(2) Å, which are significantly different. It is plausible, due to the reduced electron density of the triazenide ligand relative to the amidinate (vide supra), the tetrahydrofuran bond lengths are shorter to compensate. The same pseudo tetrahedral geometry is also observed for the formamidinate congener 18.

Figure 46 Molecular structure of [Na(N3Dipp2)(thf)3] 77, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. The tetrahydrofuran carbons are represented as wireframes for clarity. Selected bond lengths (Å) and angles (°): Na(1)-N(1) 2.419(3), Na(1)-N(3) 2.411(4), Na(1)-O(1) 2.300(4), Na(1)-O(2) 2.336(3), Na(1)-O(3) 2.340(4), N(1)-N(2)-N(3) 111.9(3), N(2)-Na(1)-O(1) 108.43(13), N(2)-Na(1)-O(2) 127.60(13), N(2)-Na(1)-O(3) 124.21(12), O(1)-Na(1)-O(2) 106.91(13), O(1)-Na(1)-O(3) 98.69(15), O(2)-Na(1)-O(3) 86.31(13).

t Deprotonation of Dipp2N3H 40 with KO Bu in tetrahydrofuran afforded

[K(N3Dipp2)(thf)] 78 as colourless plates in 70% yield. This compound has a much lower solubility in common non-donor solvents (toluene and hexane) than was observed 1 for sodium triazenide 77. The H NMR spectrum of [K(N3Dipp2)(thf)] 78 in C6D6 is therefore difficult to record. Attempts to dissolve [K(N3Dipp2)(thf)] 78 in C6D6 invariably cause the precipitation of copious amounts of colourless solid. The spectrum contains signals that may be attributed to the desired compound with a vast excess of

70 free tetrahydrofuran observed (ca. thf25:K1). It is believed that the lower solubility of this compound in arenes causes the desolvation and subsequent precipitation of large quantities of [KN3Dipp2] 76. In spite of this some data may be gathered: a doublet at δ 1.22 corresponding to the methyl groups and a broad unresolved multiplet at δ 3.38 for the methine protons are observed (cf. [Na(N3Dipp)2(thf)3] 77 δ 1.33 and 3.76 in C6D6). The tetrahydrofuran ligands correspond to signals observed at δ 1.40 and 3.56, also as broad unresolved multiplets. These are both almost identical to the values for free 280 tetrahydrofuran, which is observed δ 1.40 and 3.57 respectively in C6D6. In THF-d8 this behaviour is not observed. Instead an identical spectrum as to that recorded for base free [KN3Dipp2] 76 is observed, with the inclusion of signals for the tetrahydrofuran ligands. These are observed as unresolved multiplets at δ 1.76 and 3.63. This also allows quantification of the remaining tetrahydrofuran in the bulk material given the empirical formula [KN3Dipp2(thf)0.3]. This value is difficult to confirm with microanalysis due to the hydrolysis sensitivity of this compound as well as the extreme lability displayed by the tetrahydrofuran ligands (vide infra). The infrared spectrum of -1 201 [K(N3Dipp2)(thf)] 77 contains two important absorbances at 1293 and 1358 cm . The former is consistent with a chelating triazenide and the latter with a bridging triazenide.201 This suggests that a more complicated coordination behaviour in the solid state than is observed for the other alkali metal triazenides.

Scheme 58 The synthesis of [K(N3Dipp2)(thf)] 78.

As with [Na(N3Dipp2)(thf)3] 77, the tetrahydrofuran ligands of potassium complex 78 may be removed in vacuo over periods of greater than 6 hrs (ca. 1 x 10-1 mbar). When samples were heated under an argon atmosphere in sealed glass capillaries for solid state decomposition determination, tetrahydrofuran was observed to reflux at temperatures above 70 °C as was also observed with [Na(N3Dipp2)(thf)3] 77. A

71 colourless solid remained at the bottom of the tube until decomposition occurred at temperatures in excess of 240 °C.

[K(N3Dipp2)(thf)] 78 crystallises from a saturated tetrahydrofuran solution at -25 °C as colourless plates. Unfortunately, the quality of crystals obtained by this method is routinely low (R1 = 0.2457) due to a large number of disordered sites that could not be effectively modelled. As such connectivity is the only reliable information that can be determined by this method on this sample. Triazenide 78 crystallises in the triclinic space group P1¯ as a linear polymer consisting of alternating bis(tetrahydrofuran) and bis(triazenide) coordinated potassium atoms. The latter is divided over two unique sites that sit on special positions leading to an overall composition of

[{K(thf)2K(N3Dipp2)}n]. There is also one free tetrahydrofuran molecule per [K2L2thf2] unit. An isotropic representation of the solid state structure of [{K(thf)2K(N3Dipp2)}n] may be seen in Figure 47. This structure confirms the absence of a simple chelating triazenide ligand, as suggested by the infrared spectrum of triazenide 78. The base free potassium atom also has its coordination sphere saturated by two arene-π interactions from two separate triazenide ligands that adopt a Z-syn geometry. This is consistent with potassium’s propensity for arene-π interactions in the solid state,200,216,281-284 on account of its large ionic radius (Six coordinate ionic radius K 138 pm).2

Figure 47 Molecular structure of [{K(thf)2K(N3Dipp2)}n] 78 as determined from a crystal obtained from tetrahydrofuran solution. All atoms are refined isotropically. Lattice tetrahydrofuran and all hydrogen atoms omitted for clarity. The tetrahydrofuran carbons are represented as wireframes for clarity.

72

[K(N3Dipp2)(thf)] 78 may also be crystallised as colourless plates from toluene in sufficient quality for a full structural determination by single crystal X-ray diffraction. An illustration of the major polymer chain observed in the molecular structure may be seen in Figure 48. [K(N3Dipp2)(thf)] 78 crystallises from toluene in the triclinic space group P1¯ as from tetrahydrofuran. In this instance however, partial loss of tetrahydrofuran occurs upon concentrating the supernatant in vacuo (Scheme 59). Consequently a second tetrahydrofuran deficient polymer chain co-crystallises in the solid state structure. In this chain one tetrahydrofuran ligand has been lost from the bis(tetrahydrofuran) ligated potassium atom leading to a change in the N-N-N geometric isomerism to afford a new and as yet unreported triazenide bridging coordination mode. In order to satisfy the vacant coordination site the previously non-chelating triazenide ligand rotates and enters a chelating binding mode, to the potassium that bridges successive [KL2] units. This leads to [KL2] units, in which the central N2 nitrogen chelates the potassium with the π-arene rather than the non-π-arene coordinated potassium as per the tetrahydrofuran solvate. Representations of this polymer chain may be seen in Figure 49. The proportions of each within the structure are approximately equal as determined by the disorder modelling.

Scheme 59 The two distinct polymer chains present in [{K(thf)2K(N3Dipp2)}n] recrystallised from toluene.

73

Figure 48 Molecular structure of one polymer chain, [{K(thf)2K(N3Dipp2)}n] 78 recrystallised from toluene, POV- RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. The tetrahydrofuran and isopropyl carbons are represented as wireframes for clarity. Selected bond lengths (Å) and angles (°): K(1A)-N(3A) 2.812(4), K(1A)-Centroid(1)/(4) 2.96, K(1B)-N(1A) 2.778(2), K(1B)-N(2A) 2.654(4), K(1B)-O(1B) 2.675(2), K(1B)-O(2B) 2.913(6), K(1B)-N(1C) 2.730(2), K(1B)-N(2C) 2.628(3), K(1C)-N(3C) 2.856(3), K(1C)-Centroid(2)/(3) 2.97, N(3A)- K(1A)-Centroid(1)/(4) 77.9, N(3A)-K(1A)-N(3A)# 180, N(1A)-K(1B)-N(1C) 124.09(6), N(1A)-K(1B)-N(2C) 125.69(8), N(1A)-K(1B)-O(1B) 126.44(8), N(1A)-K(1B)-O(2B) 103.86(17), N(2A)-K(1B)-N(1C) 127.74(12), N(2A)-K(1B)-N(2C) 143.27(12), N(2A)-K(1B)-O(1B) 103.78(12), N(2A)-K(1B)-O(2B) 111.50(19), N(1C)-K(1B)- O(1B) 100.32(7), N(1C)-K(1B)-O(2B) 119.85(16), N(2C)-K(1B)-O(1B) 107.66(9), N(2C)-K(1B)-O(2B) 95.97(16), O(1B)-K(1B)-O(2B) 72.22(15), N(3C)-K(1C)-Centroid(2)/(3) 77.7, N(3C)-K(1C)-N(3C)# 180, N(1A)-N(2A)-N(3A) 117.4(4), N(1C)-N(2C)-N(3C) 116.9(3), C(1A)-N(1A)-N(2A) 123.5(3), C(13A)-N(3A)-N(2A) 109.7(4), C(1C)- N(1C)-N(2C) 122.8(2), C(13C)-N(3C)-N(2C) 109.6(3). Symmetry operations used for equivalent atoms: -x, 3-y, -z and -x, 2-y, 1-z.

Figure 49 Molecular structure of one polymer chain [{K(thf)K(N3Dipp2)}n] 78 recrystallised from toluene, POV- RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. The tetrahydrofuran and isopropyl carbons are represented as wireframes for clarity. Selected bond lengths (Å) and angles (°):K(1A)-N(2ZA) 3.215(10), K(1A)-Centroid(1)/(4) 2.96, K(1B)-N(1A) 2.778(2), K(1B)-N(3ZA) 2.773(7), K(1B)-O(1B) 2.675(2), K(1B)-N(1C) 2.730(2), K(1B)-N(2C) 2.628(3), K(1C)-N(3C) 2.856(3), K(1C)-Centroid(2)/(3) 2.97, N(2ZA)-K(1A)-Centroid(1)/(4) 67.3, N(2ZA)-K(1A)-N(2ZA)# 180, N(1A)-K(1B)-N(3ZA) 49.44(13), N(1A)-K(1B)-N(1C) 124.09(6), N(1A)- K(1B)-N(2C) 125.69(8), N(1A)-K(1B)-O(1B) 126.44(8), N(3ZA)-K(1B)-N(1C) 124.96(14), N(3ZA)-K(1B)-N(2C) 150.84(15), N(3ZA)-K(1B)-O(1B) 82.05(14), N(1C)-K(1B)-O(1B) 100.32(7), N(2C)-K(1B)-O(1B) 107.66(9), N(3C)-K(1C)-Centroid(2)/(3) 77.7, N(3C)-K(1C)-N(3C)# 180, N(1A)-N(2ZA)-N(3ZA) 110.0(8), N(1C)-N(2C)- N(3C) 116.9(3), C(1A)-N(1A)-N(2ZA) 98.3(4), C(1ZA)-N(3ZA)-N(2ZA) 111.2(8), C(1C)-N(1C)-N(2C) 122.8(2), C(13C)-N(3C)-N(2C) 109.6(3). Symmetry operation used for equivalent atoms: -x, 3-y, -z and -x, 2-y, 1-z. 74

t Deprotonation of Dipp2N3H 40 with NaO Bu in 1,2-dimethoxyethane afforded 1 [Na(N3Dipp2)(dme)2] 79 as yellow-orange blocks in good yield (Scheme 60). The H

NMR spectrum of [Na(N3Dipp2)(dme)2] 79 in C6D6 displays a doublet that corresponds to the methyl groups at δ 1.33 and the methine at δ 3.76 as a well resolved septet (cf.

[Na(N3Dipp2)(thf)3] 77 δ 1.33 and 3.76 in C6D6 respectively). The 1,2-dimethoxyethane ligands correspond to singlets that are observed at δ 2.97 (CH2) and 2.99 (CH3). The infrared spectrum displays strong absorbances at 1225 and 1272 cm-1. These are consistent with a triazenide in chelating binding mode but at lower wavenumbers than the tetrahydrofuran and base-free analogues 75-77 (1296 and 1294 cm-1 respectively).201

Scheme 60 The synthesis of Na(N3Dipp2)(dme)2 79 and K(N3Dipp2)(dme)2 80.

n [Na(N3Dipp2)(dme)2] 79 crystallises from hexane as yellow blocks in the monoclinic space group P21/n with two molecules and no lattice solvent in the asymmetric unit. Two 1,2-dimethoxyethane ligands of one molecule are heavily disordered, as is one diisopropylphenyl ring. These could not be refined anisotropically despite attempts to do so. Discussion of bonding parameters is therefore restricted to the other anisotropically refined molecule in the asymmetric unit, which does not exhibit any disorder. An illustration of one half of the asymmetric unit is found in Figure 50. In the solid state [Na(N3Dipp2)(dme)2] 79 is best described as a heavily distorted pseudo- octahedron. The triazenide is coordinated with the angle across the N(1A)-N(2A)- N(3A) unit measured as 111.9(4)°. This is the same within error as the equivalent measurement for [Na(N3Dipp2)(thf)3] 77 (N(1)-N(2)-N(3) 111.9(3) °). The bond distances for the triazenide to sodium are Na(1A)-N(1A) 2.357(4) Å and Na(1A)-N(3A) 2.518(4) Å (mean 2.44 Å), displaying a short and a long metal to nitrogen bond. This disparity contrasts the Na-N lengths of [Na(N3Dipp2)(thf)3] 77, which has roughly equal bond distances for Na(1)-N(1) 2.419(3) Å and Na(1)-N(3) 2.411(4) Å (mean 2.41 Å) despite the increased coordination number of the dimethoxyethane complex. The sodium to oxygen bond distances are longer for the dimethoxyethane ligands in

[Na(N3Dipp2)(dme)2] 79, with values of Na(1A)-O(1A) 2.412(4) Å and Na(1A)-O(2A) 2.430(5) Å for one ligand and Na(1A)-O(3A) 2.423(4) Å and Na(1A)-O(4A) 2.434(4) Å for the other. These are all longer than the tetrahydrofuran ligand sodium-oxygen bond 75 distances in [Na(N3Dipp2)(thf)3] 77 (Na(1)-O(1) 2.300(4) Å, Na(1)-O(2) 2.336(3) Å and Na(1)-O(3) 2.340(4) Å). It is likely that the relatively increased steric demand of the dimethoxyethane ligand, as well as the increased coordination number of the sodium atoms in [Na(N3Dipp2)(dme)2] 79 relative to [Na(N3Dipp2)(thf)3] 77, necessitates the lengthening of the Na-O bonds relative to those found for the tetrahydrofuran ligands in

[Na(N3Dipp2)(thf)3] 77 to alleviate steric strain.

Figure 50 Molecular structure of one unique molecule [Na(N3Dipp2)(dme)2] 79, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. The 1,2-dimethoxyethane carbons are represented as wireframes for clarity. Selected bond lengths (Å) and angles (°): Na(1A)-N(1A) 2.357(4), Na(1A)-N(3A) 2.518(4), Na(1A)-O(1A) 2.412(4), Na(1A)-O(2A) 2.430(5), Na(1A)-O(3A) 2.423(4), Na(1A)-O(4A) 2.434(4), N(1A)-N(2A)-N(3A) 111.9(4).

[K(N3Dipp2)(dme)2] 80 may be synthesised in a fashion identical to that employed for t [Na(N3Dipp2)(dme)2] 79 (Scheme 60). The reaction of KO Bu with Dipp2N3H 40 in 1,2- dimethoxyethane afforded only a single crop of [K(N3Dipp2)(dme)2] 80 in poor yield (28%) as yellow-orange blocks. A second crop was observed upon further concentration, however at this point the highly concentrated supernatant was extremely viscous and isolation through filtration or decantation of the supernatant was 1 impractical. The H NMR spectrum of [K(N3Dipp2)(dme)2] 80 in C6D6 shows the signals for the methyl groups of the isopropyl functional group at δ 1.35 and the methine at δ 3.75 (cf. [Na(N3Dipp2)(dme)2] 79 δ 1.33 and 3.76 respectively in C6D6). The 1,2-dimethoxyethane ligands are observed at δ 3.06 (methyl) and 3.25 (methylene), slightly downfield of those for the sodium triazenide 79 and in the reversed order (cf. 76

[Na(N3Dipp2)(dme)2] 79 δ 2.97 and 2.99 in C6D6 for the methylene and methyl respectively). A similar desolvation behaviour to that observed for [KN3Dipp2(thf)] 78 in C6D6 is also observed for [K(N3Dipp2)(dme)2] 80. This leads to precipitation of 1 [KN3Dipp2] 76 and the presence of a vast excess of 1,2-dimethoxyethane in the H

NMR spectrum in C6D6. This is alleviated by recording the spectrum in THF-d8. The resonances attributed to the 1,2-dimethoxyethane ligands are then observed at δ 3.27 and 3.43, in the correct integral ratios. It should be noted that long relaxation delays are required to ensure this integral value is correct. This behaviour is consistent with a dynamic coordination environment in solution. The IR spectrum of [K(N3Dipp2)(dme)2] 80 displays strong absorbances at 1293 and 1269 cm-1. As with the sodium congener these are considered to be indicative of a triazenide in chelating binding mode.201

[K(N3Dipp2)(dme)2] 80 crystallises from a saturated solution in 1,2-dimethoxyethane stored at 5 °C as large yellow orange blocks in P21/n, the same space group as the sodium congener. In contrast to [Na(N3Dipp2)(dme)2] 79, [K(N3Dipp2)(dme)2] 80 has only one molecule in the asymmetric unit, but unlike [K(N3Dipp2)(thf)] 78 it crystallises without lattice solvent. The angle across the N,N’-bonding unit of the triazenide (N(1)- N(2)-N(3) 112.59(14) °) is more open than that encountered for the sodium complex 79 (N(1A)-N(2A)-N(3A) 111.9(4) °), which is consistent with the increased radius of potassium relative to sodium (Six coordinate ionic radii Na 102 pm, K 138 pm).2 As such the ligand can be said to widen in order to accommodate the larger metal. The ligand to metal distances are also longer than sodium complex 79. For example in potassium complex 80 the values are measured as K(1)-N(1) 2.8323(15) Å and K(1)- N(3) 2.7648(15) Å, whereas in sodium complex 79 the values of the crystallographically well behaved molecule are Na(1A)-N(1A) 2.357(4) Å and Na(1A)- N(3A) 2.518(4) Å. The same heavily distorted octahedral geometry is observed for both sodium and potassium complexes 79 and 80 and as such they may be considered isostructural in the solid state.

77

Figure 51 Molecular structure of [K(N3Dipp2)(dme)2] 80, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. The 1,2-dimethoxyethane carbons are represented as wireframes for clarity. Selected bond lengths (Å) and angles (°): K(1)-N(1) 2.8323(15), K(1)-N(3) 2.7648(15), K(1)-O(1) 2.8834(14), K(1)- O(2) 2.7559(15), K(1)-O(3) 2.7708(15), K(1)-O(4) 2.8757(14), N(1)-N(2)-N(3) 112.59(14).

In a similar fashion to the smaller triazene Dipp2N3H 40, Dipp*2N3H 39 may be readily deprotonated by potassium tert-butoxide in toluene (Scheme 61). This affords

[KN3Dipp*2]∙3PhMe 81 in good yield as bright orange rods after crystallisation from toluene. Drying in vacuo (5 x 10-2 mbar, 18 h) reduces this content to an empirical 1 formula of [KN3Dipp*2]∙1.3PhMe 81 as determined using H NMR spectroscopy in

C6D6.

Scheme 61 The synthesis of [KN3Dipp*2] 81.

1 The H NMR spectrum of [KN3Dipp*2] 81 in C6D6 is devoid of the signal broadening that was observed for the parent triazene 39. This is consistent with a reduced degree of rotational freedom within the molecule upon coordination to the potassium metal. The singlet that corresponds to the para-methyl group shifts slightly downfield relative to the parent triazene to δ 1.96 (cf. Dipp*2N3H 39 δ 1.85 in C6D6). The benzylic protons

78 correspond to a sharp singlet observed at δ 6.46 (cf. Dipp*2N3H 39 δ 5.76 and 6.34). Of note, these protons are now chemically equivalent and correspond to a sharp signal where in the parent triazene in C6D6 they were inequivalent and broad. The remaining signals observed correspond to the aromatic protons, which manifest as unresolved multiplets over the range δ 6.84-7.21. By contrast the parent triazene 39 exhibits a single broad singlet in C6D6. Triazenide 81 is extremely thermally stable in the solid state and is observed to decompose over the range 205-215 °C. The infrared spectrum of triazenide 81 is devoid of the absorbance that corresponds to the N-H mode of -1 Dipp*2N3H 39 at 3295 cm . Three strong sharp absorbances in the range 1209-1250 -1 cm are consistent with a chelating triazenide ligand.

Single crystals suitable for an X-ray structure determination of [KN3Dipp*2]∙3PhMe 81 were grown from a saturated solution in toluene stored at -25 °C for 48 hours.

[KN3Dipp*2]∙3PhMe 81 crystallises in the monoclinic space group C2/c, the same as the parent triazene 39, with one full molecule in the asymmetric unit. The three lattice toluene molecules are heavily disordered and could not be successfully modelled, instead the SQUEEZE234 function of PLATON235 was used to successfully refine the structure. The residual electron density is consistent with three toluene molecules, as 1 can also be observed in the H NMR spectrum of the crystals of triazenide 81 in C6D6. The potassium is coordinatively saturated in the solid state not only by the chelate of the triazenide but also two η5 and one η3 metal to arene-π interactions from the flanking phenyl rings of the benzhydryl functional groups. The interactions were assigned using the method of Niemeyer, which clearly assigns the higher nuclearity interactions as η5 over η6 by calculating the K-centroid distances and comparing the angle between the K- centroid vector and the vector normal to plane of the arene. 204,224 Unfortunately, this method gives no clear answer for the lower nuclearity interaction. In this instance it is believed that the η3 classification is valid due to the longer distance from K(1) to the η5 centroid relative to the η3 centroid (3.278 and 3.328 Å respectively). The distances to the three atoms of this interaction (K(1)-C(9-11), 3.441(3) to 3.511(3) Å) are well within the range observed for K-arene interactions (2.819-4.116 Å)285,286 The high number of K-arene interactions is a consequence of the large ionic radius of potassium (7 coordinate ionic radius for K+ 146 pm).2 This coordination mode is similar to that observed for the related potassium guanidinate 82.267 In that instance however only two η6 interactions were observed, with K-C(arene) distances of 3.101(4)–3.284(5) and 79

3.035(4)–3.327(4) Å. In triazenide 81 the K-C(arene) distances are 3.215(3)-3.384(2) Å and 3.085(2)-3.502(3) Å. The angle across the triazenide binding motif (N(1)-N(2)-N(3) 110.95(14)°) is more acute than the equivalent angle in the parent triazene (N(1)-N(2)- N(1)# 112.1(2)°). This is consistent with chelation to a metal centre. The same angle is also noticeably more acute than other potassium complexes reported as part of this thesis, for example that of [K(N3Dipp2)(dme)2] 80 (N(1)-N(2)-N(3) 112.59(14)°).

Figure 52 Molecular structure of [KN3Dipp*2] 81, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. Selected bond lengths (Å) and angles (°): K(1)-N(1) 2.6483(16), K(1)-N(3) 2.8032(16), K(1)-C(9) 3.511(3), K(1)-C(10) 3.441(3), K(1)-C(11) 3.500(3), K(1)-Centroid(1) 2.997, K(1)-Centroid(2) 2.967, N(1)-N(2)-N(3) 110.95(14).

Figure 53 The reported potassium guanidinate 82 analogue of triazenide 81.267 80

2.3 Conclusions and Future Work 2.3.1 Conclusions

The aims of this chapter were to: i) synthesise an extremely sterically demanding triazene Dipp*2N3H 39; ii) the analogous formamidine Dipp*FormH 41 and iii) investigate the coordination chemistry of these ligands and Dipp2N3H 40 with the alkali metals with a view to their suitability as ligand transfer agents to the lanthanide metals.

The extremely sterically demanding triazene 39 required the preparation of the new compounds Dipp*N3 45 and Dipp*I 49. The latter compound has been demonstrated to n react efficiently with BuLi with a view to the synthesis of Dipp*2N3H 39. The resulting aryllithium species Dipp*Li 56 was demonstrated to react as an effective nucleophile towards the iodobutane generated in this reaction, preventing this method from being used as a method of synthesis for Dipp*2N3H 39. Dipp*2N3H 39 was instead synthesised through an aryl mangnesium compound Dipp*MgI 58 to meet the first aim.

In contrast to the above success the synthesis of the analogous formamidine Dipp*FormH 41, has not been achieved. Triethyl orthoformate was observed to react with Dipp*NH2 43 under forcing conditions in the presence of catalytic sulfuric acid to afford Dipp*N(CHO)(Et) 61. A thorough investigation of alternative reaction conditions has not yielded the desired product Dipp*FormH 41 in any case. Formimidates Dipp*NCH(OR) (R = Me and Et) 62 and 63 were synthesised and their i reactivity with amine nucleophiles (DippNH2 42, Dipp*NH2 43 and PrNH2) examined with the recovery of starting materials observed in all cases. The synthesis of

(Dipp*NH)2CO 64 and Dipp*NCNDipp* 66 were synthesised in anticipation that hydride reduction of the latter would provide a viable entry into Dipp*FormH 41.

Unfortunately, in all cases over-reduction to Dipp*NH2 43 was observed. Finally, in anticipation of the possibility of insertion into a metal amide (M-N) bond, the synthesis of isonitrile Dipp*NC 69 was investigated. Initially achieved through an as yet unoptimised and unreliable phase-transfer reaction with dichlorocarbene, this compound is more readily prepared through the dehydration of formanilide Dipp*NH(CHO) 74 with POCl3.

The coordination chemistry of triazene Dipp2N3H 40 with the alkali metals has been thoroughly investigated. Base free triazenides [MN3Dipp2] (M = Na, K) 75 and 76 have

81 been synthesised. These compounds display extremely low solubility in donor non- solvents and readily form adducts with donor solvents. The sodium complexes

[Na(N3Dipp2)(thf)3] 77 and [Na(N3Dipp2)(dme)2] 79 have been synthesised and their solid state structures investigated. Both these complexes were demonstrated to be monomeric in the solid state. The potassium complexes [{K(thf)2K(N3Dipp2)}n] 78 and

[K(N3Dipp2)(dme)2] 80 have been synthesised. 1,2-dimethoxyethane adduct 80 is monomeric in the solid state. In contrast with this tetrahydrofuran adduct 78 is a linear coordination polymer in the solid state that can display a range of different triazenide coordination modes depending on the level of desolvation it is subjected to. These complexes 77-80 are all thermally robust in the solid state and can be readily converted to the base free complexes 75 and 76 by treatment with vacuum or heat in the solid state. The lability of the Lewis bases coordinated to these alkalie metal triazenides as well as the greater thermal stability of the base free complexes suggests that these are the most useful reagents as ligand transfer agents. The potassium complex [KN3Dipp*2] 81 has been synthesised. The solid state structure displays a number of arene-π interactions (2 x η5 and 1 x η3) to the potassium.

2.3.2 Future Work

Several directions may be readily identified in order to build upon this work. The coordination chemistry of N3Dipp*2 to potassium has only been investigated thus far. The high degree of arene-π interactions observed in the solid state structure of this compound suggests that such interactions may be used as a method for coordinatively saturating metals with the need for other donor ligands such as the ethers used for the

Dipp2N3 complexes. It is plausible that extremely high coordination number caesium

Dipp*2N3 complexes for example could be synthesised.

Despite the extensive investigations into methods of synthesis outlined herein it is plausible that Dipp*FormH 41 may be synthesised by an alternative method. This thesis has outlined attempts using hydride reductants, which likely suffer due to preferential reduction of a formamidinate intermediate relative to stoichiometric single reduction of the carbodiimide. It is plausible that the use of single electron reductants such as samarium(II) would provide an alternative entry to Dipp*FormH 41.287 As a final alternative, the reaction of carbanion nucleophiles (MeLi, PhLi) with Dipp*NCNDipp*

82

66 can be reasonably anticipated to generate an acetamidine 83 or benzamidine 84 respectively (Scheme 62).

Scheme 62 The proposed synthesis of acetamidine 83 or benzamidine 84.

In addition to planned investigation of the proposed insertion of isonitrile Dipp*NC 69 into metal amide bonds (See Section 2.2.2) the coordination chemistry of isonitrile 69 is worth investigating. In terms of the scope of this thesis the coordination chemistry of Dipp*NC 69 with alkali metal amides is a readily achievable aim. Structurally authenticated isonitrile adducts of alkali metal amides number only three.272,273 It is pertinent to note isonitriles have also proven versatile ligands for the lanthanides.9,288-293 The coordination chemistry of this extremely sterically demanding isonitrile is also worth investigating with these metals.

Finally, the coordination chemistry of (Dipp*NH)2CO 64 has been identified as worth pursuing. Structurally characterised urea derivatives of lanthanides and actinides have previously been synthesised by the insertion of CO2 into a uranium imido (U=N) bond294 or the protolysis reaction between a urea and tris(cyclopentadienyl)ytterbium.295 The coordination mode versatility displayed by deprotonated ureas in these complexes (κ2-N1,N2 or κ2-N1,O1) prompts the investigation of the coordination chemistry of

(Dipp*NH)2CO 64 not only with the alkali metals but also with the lanthanides.

83

2.4 Appendix I: Supplementary Crystal Structure

The molecular structure of Dipp*H 50 was determined as part of this work. The structural parameters are of limited relevance to the discussion and are therefore not discussed in the body of the thesis. Dipp*H 50 crystallises from ethyl acetate hexane mixtures as colourless blocks in the monoclinic space group P21/n. One full molecule is included in the asymmetric unit, however severe whole molecule disorder is observed wherein two isomers co-crystallise in what has been modelled as a 56:44% ratio. The distinct forms represent rotamers about a single carbon to diphenylmethyl bond. This leads to a reduction in the overall structure quality of the collected dataset. Interestingly this is the only structure where the anti conformation of the diphenylmethyl group is observed, likely a consequence of the absence of a functional group at the ipso carbon of the principal arene. An illustration of the structures of both conformations overlayed as per their refined positions may be seen in Figure 54.

Figure 54 Molecular structure of Dipp*H 50, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. The major (anti) conformer is coloured red and the minor (syn) conformer in blue.

84

2.5 Appendix II: Crystal Data for Structures Collected for this Chapter

(Dipp N H) ·PhMe Dipp*N Dipp*H Dipp*I Dipp* N H 2 3 2 3 2 3 Chemical Moiety Formula C51.5H39N3 C33H27N3 C33H28 C33H27I C66H55N3

Chemical Sum Formula C55H78N6 C33H27N3 C33H28 C33H27I C33H22.5N1.5 Mol. Weight/ g mol-1 823.25 465.58 424.55 550.45 890.13 Temperature/ K 150(2) 150(2) 150(2) 173(2) 172(2) Crystal habit, colour Rectangular plate, colourless Blocks, pale yellow Rectangular, plate Columns, colourless Rectangular plate, colourless Crystal Size/ mm 0.05 x 0.05 x 0.01 0.1 x 0.05 x 0.01 0.12 x 0.05 x 0.02 0.3 x 0.24 x 0.24 0.18 x 0.08 x 0.04 Crystal System triclinic triclinic monoclinic Monoclinic monoclinic

Space Group P1¯ P1¯ P21/n P21/n C2/c a/ Å 11.7077(18) 5.4206(18) 8.0235(13) 9.6199(19) 29.816(4) b/ Å 13.282(2) 13.392(4) 11.6717(17) 11.703(2) 9.1001(9) c/ Å 17.179(3) 17.881(6) 26.254(5) 22.503(5) 22.918(3) α/ ° 78.383(5) 102.587(15) 90.00 90.00 90.00 β/ ° 78.383(5) 96.597(14) 94.876(4) 95.13(3) 127.544(4) γ/ ° 86.274(5) 97.095(13) 90.00 90.00 90.00 Volume/ Å3 2567.8(7) 1243.6(7) 2449.8(7) 2523.3(9) 4930.5(10) Z 2 2 4 4 4 Density calcd/ g cm-3 1.065 1.243 1.151 1.449 1.199 μ/ mm-1 0.062 0.073 0.065 1.288 0.069

F000 900 492 904 1112 1888 Reflections Collected 9926 3612 6663 7347 6232 Unique Reflections 2711 1690 2460 6061 3008 Parameters Varied 515 326 472 308 314 R(int) 0.0000 0.048 0.093 0.0361 0.0683 R1 0.082 0.0724 0.1408 0.1149 0.1120 wR1 (all data) 0.2348 0.1999 0.4124 0.1222 0.1399 Largest diff. peak and hole/ e Å-3 0.287, -0.266 0.216, -0.280 0.229, -0.232 0.98, −1.01 0.174, -0.197 GOOF 0.756 0.956 1.049 1.054 0.966

85

Dipp*N(CHO)(Et) Dipp*NCH(OMe) Dipp*NCH(OEt) (Dipp*NH) CO·4THF Dipp*NCNDipp* Dipp*NC 2 Chemical Moiety Formula C36H33NO C35H31NO C36H33NO C83H88N2O5 C67H54N2 C34H27N

Chemical Sum Formula C36H33NO C17.5H15.5N0.5O0.5 C36H33NO C83H88N2O5 C33.5H27N C34H27N Mol. Weight/ g mol-1 495.63 481.61 495.63 905.13 887.12 449.57 Temperature/ K 173(2) 150(2) 150(2) 150(2) 151(2) 150(2) Crystal habit, colour Cubic, colourless Block, colourless Plate, colourless Block, colourless Rod, colourless Plate, colourless Crystal Size/ mm 0.05 x 0.05 x 0.08 0.10 x 0.10 x 0.06 0.20 x 0.05 x 0.04 0.2 x 0.2 x 0.2 0.2 x 0.2 x 0.25 0.20 x 0.19 x 0.03 Crystal System monoclinic orthorhombic triclinic monoclinic monoclinic monoclinic

Space Group P21/c Pnma P1¯ P21/c C2/c Cc a/ Å 9.889(2) 11.8398(11) 8.682(2) 19.4859(13) 31.135(2) 10.1532(4) b/ Å 10.016(2) 24.516(4) 13.359(3) 12.8230(9) 9.0391(6) 15.9994(6) c/ Å 28.811(6) 9.6202(8) 13.788(4) 25.1870(16) 22.7079(15) 15.0113(5) α/ ° 90.00 90.00 111.065(11) 90.00 90.00 90.00 β/ ° 98.91(3) 90.00 96.604(11) 96.507(4) 130.403(5) 93.668(2) γ/ ° 90.00 90.00 107.266(10) 90.00 90.00 90.00 Volume/ Å3 2819.3(10) 2792.4(5) 1380.2(6) 6252.9(7) 4866.6(6) 2433.52(16) Z 4 4 2 4 4 4 Density calcd/ g cm-3 1.168 1.146 1.193 0.961 1.211 1.227 μ/ mm-1 0.069 0.068 0.071 0.056 0.069 0.07

F000 1056 1024 528 1920 1880 952 Reflections Collected 7302 3268 7168 18897 4738 4519 Unique Reflections 2070 1966 4763 5393 3268 4326 Parameters Varied 356 180 345 641 313 317 R(int) 0.0000 0.0994 0.047 0.0969 0.0618 0.0364 R1 0.0991 0.0545 0.0507 0.0993 0.0497 0.0371 wR1 (all data) 0.3058 0.1616 0.1505 0.3169 0.1353 0.1013 Largest diff. peak and hole/ e Å-3 0.251, -0.341 0.194, -0.292 0.241, -0.208 0.304, -0.285 0.201, 0.226 0.212, -0.280 GOOF 0.943 1.040 0.976 0.985 1.022 1.081

86

[Na(N Dipp )(thf) ] [K(N Dipp )(thf) ]·0.5PhMe [Na(N Dipp )(dme) ] [K(N Dipp )(dme) ] [K(N Dipp* )]·3PhMe 3 2 3 3 2 0.75 3 2 2 3 2 2 3 2 Chemical Moiety Formula C36H58N3NaO3 C57.5H84N6K2O1.5 C32H54N3NaO4 C32H52N3KO4 C87H78N3K

Chemical Sum Formula C36H58N3NaO3 C57.5H84N6K2O1.5 C64H108N6Na2O8 C32H52N3KO4 C87H78N3K Mol. Weight/ g mol-1 603.84 961.51 567.77 581.87 1204.70 Temperature/ K 150(2) 150(2) 150(2) 150(2) 151(2) Crystal habit, colour Plate, pale yellow Block, colourless Plate, pale yellow Block, yellow Rod, yellow-orange Crystal Size/ mm 0.15 x 0.1 x 0.03 0.14 x 0.09 x 0.03 0.15 x 0.1 x 0.02 0.1 x 0.1 x 0.1 0.2 x 0.2 x 0.2 Crystal System monoclinic triclinic monoclinic monoclinic monoclinic

Space Group P21/c P1¯ P21/n P21/n C2/c a/ Å 10.105(3) 13.1068(9) 20.0231(15) 14.2751(5) 37.8045(13) b/ Å 14.969(5) 14.9847(10) 10.9042(8) 15.4085(5) 17.0618(4) c/ Å 24.904(7) 15.9757(11) 33.135(3) 16.3889(6) 22.9880(7) α/ ° 90.00 96.923(4) 90.00 90.00 90.00 β/ ° 99.173(16) 110.851(3) 105.340(4) 108.5830(10) 119.351(2) γ/ ° 90.00 98.577(3) 90.00 90.00 90.00 Volume/ Å3 3719.0(19) 2847.7(3) 6976.8(9) 3416.9(2) 12924.2(7) Z 4 2 8 4 8 Density calcd/ g cm-3 1.078 1.121 1.081 1.131 1.2381 μ/ mm-1 0.078 0.209 0.081 0.192 0.118

F000 1320 1042 2480 1264 5120 Reflections Collected 10740 16413 13720 6705 14152 Unique Reflections 2468 8379 4597 5083 8545 Parameters Varied 416 772 747 373 633 R(int) 0.1828 0.0532 0.1597 0.0535 0.0827 R1 0.0851 0.0695 0.0884 0.0425 0.0583 wR1 (all data) 0.3420 0.2132 0.3065 0.1292 0.1721 Largest diff. peak and hole/ e Å-3 0.366, -0.273 0.611, -0.365 0.426, -0.282 0.447, -0.309 0.372, -0.711 GOOF 0.828 1.037 0.979 1.012 0.925

87

Chapter III Lanthanide Triazenides: Synthesis and Reactivity

3.1 Introduction 3.1.1 Trivalent Lanthanide Amidinates and Guanidinates

Trivalent lanthanide amidinates and guanidinates are rapidly growing in popularity as alternatives to trivalent lanthanocenes.296,297 A selection of some of the published homoleptic amidinates and guanidinates may be seen in Figure 55.210 Homoleptic trivalent lanthanide amidinates have typically been investigated from a fundamental, coordination chemistry perspective.210 There is also interest in volatile homoleptic amidinates and guanidinates for materials chemistry applications; for instance lanthanide amidinates 85-91,298 and guanidinates (Figure 55, far right) 299,300 have been used as precursors for oxide film deposition.298 Other applications of these complexes include terbium guanidinate 92 for luminescent materials301 and neodymium guanidinate 93 for the polymerisation of ε-caprolactone. 302 For a comprehensive overview of the coordination chemistry of these complexes the reader is directed towards some of the many reviews on the subject.165,169,170,303,304

Figure 55 A selection of homoleptic tris(amidinate) and tris(guanidinate) lanthanide complexes.210,298-302

Heteroleptic lanthanide mono(amidinate) alkyls have found use in polymerisation catalysis. Research in this area typically focuses on the changes in catalyst activity depending on the metal ionic radius.305-307 Amidinate alkyls 93-98 showed moderate 305 activity for polymerisation of L-lactide. The effects of the metal ionic radius on catalytic activity was well demonstrated for amidinate alkyls 99-105. Upon initiation

88 with [PhNMe2H][B(C6F5)4] these complexes showed a range of activity for the polymerisation of ethene over two orders of magnitude; with the largest and smallest radius metals 102, 99 and 105 showing low activity and the intermediate radius metals 100, 101, 103 and 104 showing much higher activity.306,307

Figure 56 Heteroleptic trivalent lanthanide amidinates for catalytic applications.305-307

Attempts to prepare extremely sterically congested tris(DippForm) lanthanide complexes through a tandem redox transmetallation-protolysis method did not give the expected product; instead bis(amidinate) lanthanide fluorides were isolated (Scheme 63).308,309 It was proposed in this instance that the sterically congested bis(amidinate) lanthanide perfluoroaryl 106 is incapable of further protolysis. Instead, intramolecular C-F activation occurs to give lanthanide fluorides (Scheme 63). The proposed tetrafluorobenzyne intermediate is then protonated by the third equivalent of formamidine followed by nucleophilic attack on a molecule of tetrahydrofuran to afford a ring opened tetrahydrofuran co-product 107 that incorporates the formamidine, tetrahydrofuran and tetrafluorobenzyne in a single co-product.

Scheme 63 The C-F activation reaction of sterically congested pentafluorophenyl lanthanide bis(formamidinate).308,309

3.1.2 Divalent Lanthanide Amidinates and Guanidinates

There are a number of divalent lanthanide amidinates (Figure 58).310-313 At this stage there are no divalent amidinates from the non-classical divalent triad (Ln = Nd, Dy and 11 Tm). An isostructural series of bis(formamidinates) [Ln(DippForm)2(thf)2] (Ln = Sm, Eu, Yb) 108-110 has been reported,310,311 all of which show reactivity towards chemical oxidants.311,314 For example, the strongly reducing samarium complex 108 was observed 89 to react with the carbodiimide DippNCNDipp, as well as other less sterically demanding carbodiimides to afford the very sterically congested tris(formamidinate) 287 [Sm(DippForm)3] 111. Compounds of this type proved inaccessible by the tandem redox transmetallation protolysis method (Scheme 63). The silylamidinates 112-115 also show similar reactivity with chemical oxidants such as diphenyldiselenide.312,313

Figure 57 Divalent lanthanide amidinates.310-313

Divalent lanthanide guanidinates are uncommon. Those known are also from the classical divalent triad (Figure 58).11 Sterically hindered guanidinates have been used to generate unusual low coordinate, Lewis-base free, square planar lanthanide complexes 116-121.315,316 The samarium guanidinate 116 reacts with carbon disulfide to yield the 2- 316 unusual C-S reductively coupled [SCSCS2] product 122 in preference to the 2- 317-319 conventional C-C coupled [S2CCS2] .

Figure 58 Divalent lanthanide guanidinates.315,316

90

3.1.3 Lanthanide and Actinide Triazenides

There a number of lanthanide and actinide triazenides. Unsurprisingly for the lanthanides, the trivalent oxidation state is more common; the structurally authenticated lanthanide triazenides are illustrated in Figure 59. Each of these features the triazenide in the chelating bonding geometry (See Figure 17 in Section 2.1.2). These are synthesised by a range of methods such as protolysis of lanthanide cyclopentadienyls for 123-126320 or metathesis from potassium triazenides for the tetramethylaluminates 127-129.226 The remaining complexes are synthesised through the insertion of an organic azide into a lanthanide carbon for 130-131321,322 or lanthanide phosphide bond for 132-133.323

Figure 59 All structurally characterised lanthanide triazenide complexes.320-323

There are four divalent lanthanide triazenide complexes. The synthesis of these compounds nicely illustrates the utility of the tandem redox transmetallation method of synthesis. Europium and ytterbium triazenides 134-136 are synthesised through a tandem redox transmetallation protolysis reaction method from [Hg(C6F5)2], the lanthanide element and the protonated triazene.226,324,325 Ytterbium triazenide dimer 137 is synthesised through a redox transmetallation with a mercury chloride triazenide and ytterbium metal.192 As with the more common trivalent lanthanide complexes, the triazenide ligand adopts the chelating geometry in all cases.

91

Figure 60 All structurally characterised divalent lanthanide triazenide complexes.192,226,324,325

There are several published actinide triazenides (Figure 61). Interestingly these are all synthesised by insertion of an organic azide to an actinide alkyl bond. This method is not unique to the actinides (vide supra) however the absence of metathetical entries is unusual. There is one example of a triazenide synthesised by protolysis, [UCp2(p-

Tol2N3)2] 138; however this has not been structurally authenticated and the monodentate coordination mode was inferred from 1H NMR and IR spectroscopy.326 The remainder are synthesised by the insertion of adamantyl- 139-143 or mesityl azide 144-145 into actinide alkyl bonds. Many of these compounds display the unusual κ2-N1,N2 bonding mode (See Figure 17 in Section 2.1.2), which has not been observed for the lanthanides to date (vide supra).

Figure 61 Actinide triazenide complexes synthesised by insertion of an azide into a metal alkyl bond.196-199,326,327

3.1.4 Purpose of this Chapter

There were several aims for this chapter. The first was to evaluate the most useful synthetic methods for the isolation of lanthanide triazenide complexes (See Section 2.1.1 in Chapter II). In situ generation of a lithium triazenide and use of the base-free thallium triazenide were targeted first in order to mitigate the complications of the

92 lability of the Lewis base coordinated triazenides synthesised in Chapter II. Secondly the base-free alkali metal triazenide complexes synthesised in Chapter II were evaluated for utility as metathesis precursors in reactions with lanthanide trihalides. Thirdly a thallium triazenide was evaluated as a precursor for redox transmetallation and metathesis synthesis protocols. Finally the protolysis synthetic method from the protonated ligand Dipp2N3H 40 and the tris(bis(trimethylsilyl)amide) lanthanide complexes was also examined as a method of synthesis of very sterically congested tris(triazenide)lanthanide complexes.

The second aim was to examine the reactivity of the trivalent lanthanide complexes in reaction types common to sterically congested lanthanocenes like [SmCp*3] 1. To achieve this aim, solid state steric calculations were performed on a range of lanthanocenes that have been used as starting materials for Sterically Induced Reduction (SIR) reactivity as well as precursors to non-classical divalent lanthanocenes. These results have been compared to the equivalent calculations performed on [Ln(N3Dipp2)3] complexes.

The third aim was to synthesize samarium and ytterbium triazenides in the divalent state to examine the effect of the electronics of the ligand relative to their amidinate congeners. It was anticipated that the reduced donicity of the triazenide relative to amidinates may complement the unusual oxidation state lanthanides. Preliminary forays into the application of Dipp2N3H 40 to non-classical divalent (thulium) triazenides were also made.

Finally extensive investigations targeting a cerium(IV) triazenide have been made. Some investigations regarding steric demand requirements at high oxidation state cerium compounds have also been undertaken.

3.2 Results and Discussion 3.2.1 Synthesis of Trivalent Lanthanide Triazenides

Metathetical synthetic strategies were investigated first as a method of synthesis of lanthanide triazenides. The solvent adducts synthesised in Chapter II were considered not as ideal as ligand transfer agents due to the lability demonstrated by the Lewis

93 bases. The base free complexes or complexes generated in situ were considered ideal to mitigate this lability. It was envisaged that the alkali metal by-products would be the easiest to remove during work-up. The ready availability of well characterised lanthanide triiodide tetrahydrofuran solvates,328 in contrast to the involved procedures329,330 required for the dehydration of commercially available lanthanide trichloride hydrates, also made this an attractive synthetic method.

Due to the possibility of later chemical reduction to access the divalent oxidation state (vide supra),95,99,331,332 the synthesis of neodymium triazenides was targeted first. It was believed that the slightly reduced steric profile of the triazenide ligand relative to the isostructural formamidine (See Section 2.2.3 in Chapter II) would allow the synthesis of homoleptic lanthanide tris(triazenides) where the analogous formamidine syntheses failed (Scheme 63). A metathesis method was considered advantageous as it eliminates the C-F activation side-reaction observed in those cases. A solution of three equivalents 255 of [LiN3Dipp2(thf)n] prepared in situ in tetrahydrofuran to ensure correct stoichiometry, was added to a suspension of one equivalent of [NdI3(thf)3.5], also in tetrahydrofuran. After work-up a mixture of two products was characterised by single crystal X-ray diffraction structural determination (Scheme 64). A lithium iodide adduct of the lithium triazenide, [Li2(μ-I)(μ-thf)(μ-N3Dipp2)(thf)2] 146 was crystallised from pentane. Subsequently, neodymium triazenide, [NdI2(N3Dipp2)(thf)3] 147 was crystallised from toluene. Attempts to fractionally recrystallise the mixture, in order to separate the two compounds or deduce the fate of the third equivalent of the lithium triazenide failed. This prevented further characterisation and illustrates the problems observed with metathesis reactions between lithium complexes and lanthanide halides: lithium halide inclusion and poor solubility of the lanthanide trihalide, which causes stoichiometric uncertainty. It became apparent that metathesis from lithium triazenides would give unpredictable results and was thus not investigated further. In spite of this, the reaction outcome is unusual and worthy of discussion as, while it is common place for lithium halides to be incorporated into lanthanide complexes, it is unusual for lithium halide by-products to be incorporated into the metathesis reagent precursor and in so doing prevent further reaction with the lanthanide halide as appears to be the case here.333-337

94

Scheme 64 The reaction of lithium triazenide with neodymium triiodide tetrahydrofuran solvate.

[Li2(μ-I)(μ-thf)(μ-N3Dipp2)(thf)2] 146 crystallises upon storage of a saturated solution of the crude reaction mixture in npentane at room temperature. An illustration of the molecular structure may be seen in Figure 62. Triazenide 146 crystallises as yellow square truncated plates in the monoclinic space group P21/n with one full molecule in the asymmetric unit. Triazenide 146 exhibits a unique structural motif, in that there are no published crystal structures of lithium iodide adducts of either N-aryl triazenides or N-aryl amidinates. A bis lithium N-alkyl amidinate lithium iodide adduct 148 is known, as well as other lithium halide adducts of lithium amidinates (Figure 63).338 However this displays a vastly different ladder-like structure due to the fact it is a 2:1 amidinate to halide adduct rather than the 1:1 stoichiometry observed here.

Figure 62 Molecular structure of [Li2(μ-I)(μ-thf)(μ-N3Dipp2)(thf)2] 146, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms and low occupancy partners of disordered atoms omitted for clarity. The tetrahydrofuran carbons are represented as wireframes for clarity. Selected bond lengths (Å) and angles (°): Li(1)- N(1) 1.991(4), Li(2)-N(3) 2.010(4), Li(1)-I(1) 2.794(4), Li(2)-I(1) 2.747(4), Li(1)-O(1) 2.042(4), Li(1)-O(2) 1.928(4), Li(2)-O(1) 2.057(5), Li(2)-O(3) 1.903(4), N(1)-N(2)-N(3) 113.56(17), Li(1)-I(1)-Li(2) 57.11(12), Li(1)-O(1)-Li(2) 80.51(17).

95

The N(1)-N(2)-N(3) angle of triazenide 146 (113.56(17)°) is very wide when compared to the same bond angle in the monometallic sodium, and potassium complexes 77-80 synthesised in Chapter II. These triazenide chelate complexes feature N(1)-N(2)-N(3) bond angles of 111.8(3)°, 111.9(3)° and 112.59(14)° for sodium and potassium triazenides 77, 79 and 80. The widening of this bond angle is due to the bridging bimetallic bonding mode (See Figure 17 in Section 2.1.2) exhibited by the triazenide ligand in complex 146. The iodide ligand is closer to Li(2) than Li(1) with bond lengths of 2.747(4) Å and 2.794(4) Å respectively. These distances are consistent with the lithium iodine bond distances reported for amidinate 148 which features bond lengths of 2.797(11) Å and 2.741(11) Å respectively.338 The iodide ligand is bent significantly out of the Li2N3 plane (out of plane distance 2.124 Å). The bridging tetrahydrofuran ligand observed in 146 is common in lithium amidinates and amidinate derivatives, for example compounds 149-151. The Li(1)-O(1)-Li(2) angle (80.51(17)°) is well within the range reported for other tetrahydrofuran ligands bridging two lithium atoms. For example the widest reported angle for this moiety is 87.6(2)° for the xylyl amidinate 150205 and the most acute is 77.4(2)° for the polysilylamidinate 151.339

Figure 63 Lithium amidinates 148-151 for comparison of structural features.205,338-340

Neodymium triazenide 147 crystallises from toluene in the orthorhombic space group

Pbca. The structure contains one full [NdI2(N3Dipp2)(thf)3] molecule in the asymmetric unit. However due to the extremely poor quality of the data a reasonable refinement beyond direct donor atom to neodymium contacts could not be achieved for this structure. The proposed composition of triazenide 147 is based on its isomorphous nature when compared to the published data for the samarium formamidinate 314 [Sm(DippForm)Br2(thf)3].

In addition to their utility as redox transmetallation reagents, thallium ligand transfer agents are attractive for lanthanide metathesis chemistry owing to the extremely low solubility of thallium(I) halides. In this instance [TlN3Dipp2] 152 was considered ideal as a ligand transfer agent due to the high solubility in common solvents as a base free

96 complex. The reactivity of the thallium(I) triazenide [TlN3Dipp2] 152 was therefore briefly investigated. [TlN3Dipp2] 152 had previously been synthesised by a colleague of 255 the author. Therefore CeCl3 was treated with three equivalents of thallium triazenide 152. It was anticipated that this would avoid the by-products observed when using tandem redox transmetallation-protolysis methods to try and synthesise tris(formamidinate) lanthanide complexes. (See Scheme 63 in Section 3.1.1).

Crystallisation of the product from toluene after multiple filtrations to remove a microfine powder suspension, gave the unusual, dimeric thallium hydroxide adduct of the thallium triazenide 153 (Scheme 65). This compound is extremely air and light sensitive and was obtained as part of a mixture of amorphous compounds, preventing purification and characterisation by other methods. No cerium containing products could be isolated from the reaction mixture. The most likely origin of the hydroxide in dimer 153 is thought to be the trichloride precursor, derived from thionyl chloride 329 dehydration of [CeCl3(H2O)6] by the literature method.

Scheme 65 The partial hydrolysis of thallium triazenide 153 by partially hydrated cerium trichloride.

[{Tl2(μ-OH)(μ-N3Dipp2)}2] 153 crystallises from toluene as yellow rods. As well as its sensitivity to light and air (vide supra) this compound was observed to decompose under the conventional immersion oil (Immersion oil type NVH) used for mounting crystals in house. Similar behaviour was observed in the commonly used alternative (Paratone N). Therefore the solid state structure was determined from a crystal mounted in silicone oil.

Thallium triazenide 153 crystallises in the triclinic space group P1¯ with one half of the molecule and one toluene molecule in the asymmetric unit. A representation of the structure may be seen in Figure 64. Thallium triazenide 153 adopts a ladder-like,

Tl4O2N2 structure. The two planes, N3Tl(2) and its symmetry equivalent sit parallel to one another with the central Tl2O2 plane oriented at 12.01° to the TlN3 planes. These three planes form a ladder when linked by the Tl2ON(3) planes, which are oriented at

85.96° to the aforementioned TlN3 planes. Each triazenide is bonded to two thallium 97 atoms, Tl(1) and Tl(2), which in turn bond to a single hydroxide unit (O(1)). The four N-arene rings lie in planes orthogonal to each other at angles of 48.12° and 48.96° to the triazenide N3 plane. This orientation is likely preferred due to the steric effects of the two large thallium atoms. Similar ladder-like structures have been observed for dimeric thallium triazenides previously (See Figure 23 in Section 2.1.4) however there are no equivalent thallium hydroxide adducts.

Figure 64 Molecular structure of [{Tl2(μ-OH)(μ-N3Dipp2)}2] 153, POV-RAY illustration, 50% thermal ellipsoids, all non-hydroxo hydrogen atoms and lattice toluene omitted for clarity. Selected bond lengths (Å) and angles (°): Tl(1)- N(1) 2.597(3), Tl(2)-N(3) 2.601(3), Tl(1)-O(1) 2.454(3), Tl(2)-O(1) 2.398(3), N(1)-N(2)-N(3) 110.8(3), N(2)-N(1)- C(1) 114.7(3), N(2)-N(3)-C(13) 116.1(3), Tl(1)-O(1)-Tl(2) 103.05(10), N(3)-Tl(2)-O(1) 91.42(9), N(1)-Tl(1)-O(1) 85.83(10).

Concurrent with the investigation of metathesis reactions with thallium triazenide, the reaction of the heavier alkali metal triazenides with lanthanide triiodides was investigated. It was hypothesised that the use of the heavier alkali metal triazenides would discourage the incorporation of alkali metal contaminants as per lithium triazenide [Li(N3Dipp2)(thf)n] (vide supra).

Given the difficulties observed in the attempts to synthesis homoleptic tris(triazenido)lanthanides, a 2:1 stoichiometry was considered. The reaction of two equivalents of [KN3Dipp2] 76 with [SmI3(thf)3.5] in tetrahydrofuran gave, after recrystallisation from toluene, the desired samarium (III) bis(triazenide) complex,

98

[SmI(N3Dipp2)2(thf)] 154 in moderate yield (Scheme 66). Despite the paramagnetic nature of this compound moderately instructive data could be obtained from its 1H

NMR spectrum in C6D6. For instance the aromatic protons are paramagnetically shifted and observed as multiplets over the range δ 7.12-7.15 and 7.43-7.54 and the methyl groups are observed as two resonances at δ 0.78 and 1.19. Unfortunately, it is not possible to distinguish which of the remaining signals correspond to the tetrahydrofuran protons or the isopropyl methine protons. The infrared spectrum displays an absorbance at 1256 cm-1. This is consistent with a triazenide ligand in a chelating monometallic binding mode and is supported by the molecular structure as determined by single crystal X-ray diffraction (Figure 65).192 Acceptable microanalyses could not be attained for this compound despite repeated attempts, during which the sample routinely gave anomalously low values for carbon.

Scheme 66 The synthesis of [SmI(N3Dipp2)2(thf)] 154 by metathesis. Samarium bis(triazenide) 154 crystallises from toluene as thin pale yellow rectangular plates in the triclinic space group P1¯. One full molecule is included in the asymmetric unit with no lattice solvent. An illustration of the structure may be seen in Figure 65. In the solid state the structure is best described as a distorted tetrahedron, if the triazenide ligands are considered as single point donors. The distorted geometry is best observed at I(1)-Sm(1)-O(1) (86.4(1)°) and N(5)-Sm(1)-N(2) (131.6(2)°), which deviate heavily from the ideal value of 109.5°. The structurally related samarium formamidinate complex [SmI(DippForm)2(thf)] 155 has been published (Figure 66) and it is instructive to compare the solid state structures of both due to the similar coordination geometries and steric demand of the ligands.314 The metal to nitrogen bond lengths are very similar when pairs of bond lengths in the formamidinate ligands are compared to the equivalent bond lengths in the triazenide ligand. In general, the metal to nitrogen bond lengths for the triazenides are slightly longer than for the formamidinate ligands.314 One formamidinate ligand has metal to nitrogen bond lengths of 2.3784(15) Å and 2.4751(15) Å. These compare favourably to the triazenide bond lengths of 2.389(6) Å and 2.479(6) Å. The other formamidinate has Sm-N bond lengths of 2.4238(16) Å and

99

2.4529(15) Å, which are similar in length to those observed in the other triazenide ligand, 2.429(6) Å and 2.447(5) Å. The nitrogen to nitrogen bond lengths are the same within error (N(1)-N(2) 1.323(7) Å, N(2)-N(3) 1.304(8) Å, N(4)-N(5) 1.316(8) Å, N(5)-

N(6) 1.307(9) Å), suggesting that the bonding is delocalised across the N3 unit. The differences in the Sm-N bond lengths are therefore not the result of charge localisation at the N3 unit and likely the result of steric buttressing between the ligands. The iodide and tetrahydrofuran ligands also have similar metal to ligand bond lengths with values of 2.9839(9) Å and 2.438(5) Å respectively, compared to the reported values in samarium formamidinate 155 of 2.9851(1) Å and 2.4535(13) Å for the iodide and tetrahydrofuran ligands respectively.314

Figure 65 Molecular structure of [SmI(N3Dipp2)2(thf)] 154, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. The tetrahydrofuran carbon atoms are depicted as a wireframe for clarity. Selected bond lengths (Å) and angles (°): Sm(1)-N(1) 2.429(6), Sm(1)-N(3) 2.447(5), Sm(1)-N(4) 2.389(6), Sm(1)- N(6) 2.479(6), Sm(1)-O(1) 2.438(5), Sm(1)-I(1) 2.9839(9), N(1)-N(2)-N(3) 110.1(6), N(3)-N(4)-N(5) 110.1(6), I(1)- Sm(1)-O(1) 86.4(1), I(1)-Sm(1)-N(2) 110.0(1), I(1)-Sm(1)-N(5) 106.9(1), N(5)-Sm(1)-O(1) 107.3(2), N(5)-Sm(1)- N(2) 131.6(2), N(2)-Sm(1)-O(1) 105.3(2).

Figure 66 The structurally related samarium bis(formamidinate) complex [SmI(DippForm)2(thf)] 155. Using identical reaction conditions that were successful in the synthesis of

[SmI(N3Dipp2)2(thf)] 154 proved unsuccessful for the isolation of [NdI(N3Dipp2)2(thf)].

100

The reaction of [KN3Dipp2] 76 and [NdI3(thf)3.5] in tetrahydrofuran gave a product, which unlike [SmI(N3Dipp2)2(thf)] 154, did not recrystallise from toluene readily. Recrystallisation of the reaction product from nhexane gave the homoleptic neodymium tris(triazenide) complex, [Nd(N3Dipp2)3] 156 in moderate yield (60%) instead of the 1 desired bis(triazenide) complex [NdI(N3Dipp2)2(thf)] 157 (Scheme 67). H NMR spectra of [Nd(N3Dipp2)3] 156 recorded in C6D6 are not particularly instructive due to paramagnetic shifting and broadening. [Nd(N3Dipp2)3] 156 is thermally unstable in the solid state and undergoes gradual decomposition over the range 109-121°C. The infrared spectrum has a range of intense absorbances around the most intense peak at 1258 cm-1, which is in the reported range for the N-N absorbances in chelating triazenide ligands.201

Scheme 67 The synthesis of [Nd(N3Dipp2)3] 156 by solvent induced ligand redistribution.

It is believed that this product is the result of solution phase redistribution of the ligands between metal centres driven by the insolubility of NdI3 in hexane (Scheme 67). Similar outcomes have been observed during preparation of lanthanide and actinide 2- aminopicolinates.341,342 This is consistent with the observation of precipitated insoluble material during recrystallisation of the reaction mixture from nhexane. This material must be neodymium triiiodide tetrahydrofuran solvate, which is known to be highly insoluble in alkane solvents.328 Analysis of this material was not undertaken due to the likely contamination with reaction byproducts such as potassium iodide, which would prevent sensible microanalytical values.

n [Nd(N3Dipp2)3] 156 crystallises from an hexane toluene mixture as long yellow rods in the monoclinic space group P21/n with one full molecule in the asymmetric unit as well as two toluene molecules. This is isostructural with the published structure for 310 [Sm(DippForm)3]. A representation of the structure may be seen in Figure 67. The coordination geometry of the complex may be described as a compressed and heavily distorted octahedron but is best described by counting the triazenide ligands as single point donors (The N2 atom as the central donor position), giving a trigonal planar 101 geometry overall. Slight distortion from ideal (120°) is observed (N(2)-Nd(1)-N(5) 122.6(2)°, N(2)-Nd(1)-N(8) 118.2(2)° and N(8)-Nd(1)-N(5) 119.2(2)°). The metal to nitrogen bond lengths are statistically similar to each other, both at the same ligand and between ligands, for each equivalent bonding nitrogen on each ligand. In contrast to the samarium complex 154 above, it is not as instructive to compare the solid state structures of the triazenides and formamidinates as the closest neodymium formamidinate complexes for comparison are the tris(formamidinate) complexes 309 [Nd(DiepForm)3] 158 and [Nd(MesForm)3] 159, which may be seen in Figure 68. Neither of these two includes formamidinates with a similar steric profile to the triazenide ligand in complex 156 and as such trends in bonding are not informative. In comparison to samarium bis(triazenide) 154, which has a longest metal to nitrogen distance of Sm(1)-N(6) 2.479(6) Å, the metal to nitrogen bonds in neodymium tris(triazenide) 156 are similar tending to longer, where only the shortest metal to nitrogen bond distance in 156, Nd(1)-N(6) 2.475(7) Å is comparable. It is likely that these differences are partly due to steric effects, that is neodymium complex 156 is more sterically congested and such congestion enforces longer metal to ligand bonds in order to accommodate each ligand. This phenomenon is also observed for sterically congested tris(pentamethylcyclopentadienyl) lanthanides where longer metal to ligand bonds are observed than for the Ln-Cp* bonds in mono- and bis(pentamethylcyclopentadienyl) complexes.55,57

102

Figure 67 Molecular structure of [Nd(N3Dipp2)3] 156, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. Selected bond lengths (Å) and angles (°): Nd(1)-N(1) 2.499(7), Nd(1)-N(3) 2.476(7), Nd(1)-N(4) 2.503(8), Nd(1)-N(6) 2.475(7), Nd(1)-N(7) 2.483(7), Nd(1)-N(9) 2.495(7), N(2)-Nd(1)-N(5) 122.6(2), N(2)-Nd(1)-N(8) 118.2(2), N(8)-Nd(1)-N(5) 119.2(2), N(1)-N(2)-N(3) 112.6(7), N(3)-N(4)-N(5) 110.8(8), N(7)- N(8)-N(9) 113.2(7).

309 Figure 68 Neodymium tris(formamidinate)s [Nd(DiepForm)3] and [Nd(MesForm)3].

This ligand redistribution equilibrium observed in the synthesis of [Nd(N3Dipp2)3] 156 was unanticipated due to the difficulties observed during the attempted preparations of similarly sized lanthanide formamidinates, which are observed to be sterically inhibited and difficult to isolate.287,308-310 It is plausible that this behaviour was not observed for the iodosamarium bis(triazenide) 154 for two reasons: firstly the greatly reduced solubility of lanthanide trihalides in straight chain alkanes relative to arene solvents328 and secondly the difference in the ionic radii.2 Six coordinate neodymium’s ionic radius is larger than samarium’s, with values of 98.3 pm and 95.8 pm respectively (See Table 1 in Section 1.1 in Chapter I). The tris(triazenide) samarium congener was therefore considered a viable synthetic target (vide infra) to test this hypothesis. This aim was supported by the fact that the related samarium formamidinate compound 111 had been previously synthesised (Scheme 68).310 103

287,310 Scheme 68 The preparation of literature samarium formamidinate [Sm(DippForm)3] 111. Upon appraising the metathesis reactions conducted, it became apparent that as a method of synthesis alkali metal and thallium triazenides would be prone to unpredictable, highly solvent dependent (vide supra) outcomes during the preparation of lanthanide triazenides. The availability of thallium triazenide 152 gave ready access to a redox transmetallation reagent (Scheme 26). This was investigated next. The samarium tris(triazenide) was targeted for the aforementioned reasons. Redox transmetallation methods have given access to many of the lanthanide formamidinates that are in the chemical literature.308-311,314 As highlighted in Scheme 63 these syntheses have used organomercury reagents to generate aryllanthanide complexes, which then react with a protonated ligand precursor to generate the desired lanthanide formamidinates. This tandem redox transmetallation protolysis method is required due to the instability of mercury and thallium formamidinates. It was therefore of interest to see if the readily accessible thallium triazenide 152 is as useful as cyclopentadienylthallium for the synthesis of lanthanide cyclopentadienyl complexes.69,162-164

The stoichiometric reaction of three equivalents of thallium triazenide 152 with one equivalent of samarium metal powder in tetrahydrofuran gave a mixture of starting materials and the desired samarium tris(triazenide) complex, [Sm(N3Dipp2)3] 160. Analysis of the rigorously dried solids isolated from the filtered reaction mixture

(Scheme 69) suggest 83% conversion of [TlN3Dipp2] 152 as determined by mass balance. During the reaction the suspended, freshly filed samarium metal powder was observed to aggregate into several large (>mm3) particles. It was believed that this degree of samarium aggregation hindered the overall reaction progress by reducing the amount of available samarium metal to the thallium triazenide. Due to similar, high solubilities in common laboratory solvents the thallium and samarium complexes could not be separated by fractional recrystallisation. It was therefore deemed necessary to use a large excess of samarium metal (Five equivalents) to ensure complete consumption of the thallium triazenide 152 (Scheme 69). This afforded samarium tris(triazenide) 160 in excellent yield after filtration and removal of volatiles in vacuo. The excess samarium 104 metal was added portion-wise to minimise the likelihood of reduction of the desired samarium(III) compound to a samarium(II) triazenide (vide infra). Indeed this reduction was observed to occur at the conclusion of the reaction when a faint green tinge was observed at the metal-solution interface (See Section 3.2.3 for the preparation of green

[Sm(N3Dipp2)2(thf)2]). This is in direct contrast to the reaction of thallium cyclopentadienide with vast excesses of samarium metal, which fails to yield the 162 corresponding samarium(II) metallocene. The infrared spectrum of [Sm(N3Dipp2)3] 160 exhibits a number of absorbances in the range 1200-1246 cm-1 with the most intense being observed at 1246 cm-1. This is consistent with the triazenide binding as a 201 chelating ligand. In addition to the anticipated hydrolysis instability, [Sm(N3Dipp2)3] 160 also has low thermal stability in the solid state like its neodymium analogue

([Nd(N3Dipp2)3] 156 dec. 109-120 °C), with gradual decomposition being observed 1 between 106-121 °C. In contrast to [SmI(N3Dipp2)2(thf)] 154, the H NMR spectrum of

[Sm(N3Dipp2)3] 160 in C6D6 is not instructive. Resonances are observed but do not provide useful characterisation data of the chemical environments in solution due to the high number of chemical environments observed for the high symmetry molecule.

Scheme 69 The reaction of [TlN3Dipp2] 152 with samarium metal yielding [Sm(N3Dipp2)3] 160. Samarium tris(triazenide) 160 crystallises from nhexane as deep yellow cubes in the trigonal space group R3¯c with one half of the molecule in the asymmetric unit and without lattice solvent. The same C3 molecular symmetry is observed as with

[Nd(N3Dipp2)3] 156 despite the different space group and asymmetric unit as well as the presence of lattice solvent in [Nd(N3Dipp2)3] 156. An illustration of the structure of

[Sm(N3Dipp2)3] 160 with the symmetry generated half of the molecule included may be seen in Figure 69, for the numbering scheme for the following metrical discussion see Figure 70. In the solid state the same pseudo-trigonal planar geometry is observed as in

105

[Nd(N3Dipp2)3] 156, with a similar range of angles (116.43(7)° and two times 121.78(4)°). The metal to nitrogen bond lengths for the reported samarium tris(formamidinate) analogue 111 (Scheme 68) compare favourably to those observed in

[Sm(N3Dipp2)3] 160. The Sm(1)-N(4) and Sm(1)-N(1) bond lengths (2.4427(19) Å and 2.450(2) Å respectively) in triazenide complex 160 are the same within error as Sm(1)- N(1) (2.448(6) Å) in formamidinate complex 111. The Sm(1)-N(3) bond (2.4597(18) Å) in triazenide complex 160 is also the same within error as the Sm(1)-N(5) and Sm(1)- N(6) bonds (2.467(6) Å and 2.460(6) Å respectively) in formamidinate complex 111. The triazenide has a much narrower bite angle (range 53.07-53.19°) relative to the formamidinate ligand (range 55.53-56.46°).310 This is a consequence of the differences in the angle across the NNN and NCN bonding units; for example in triazenide 160 (113.44(18)° and 112.3(3)°) this angle is much narrower than the equivalent angles in formamidinate 111 (120.1(8)°, 119.3(8)° and 119.5(7)°).310 The smaller bite angle leads to a wider CNN angle from the ipso carbon of the flanking arenes to the NNN central than the analogous CNC angle of the formamidinate ligands in

[Sm(DippForm)3] 111. The average of these angles is 112.1° in formamidinate 111 and 109.36° in triazenide 160.

Figure 69 Molecular structure of [Sm(N3Dipp2)3] 160, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. Selected bond lengths (Å) and angles (°): Sm(1)-N(1) 2.4427(19), Sm(1)-N(3) 2.4597(18), Sm(1)-N(4) 2.450(2), N(1)-N(2)-N(3) 113.44(18), N(4)-N(5)-N(4)# 112.3(3), C(1)–N(1)–N(2) 109.50(17), C(13)– N(3)–N(2) 109.21(18), N(1)-Sm(1)-N(3) 53.19(6), N(4)-Sm(1)-N(4)# 53.07(9), N(2)-Sm(1)-N(2)# 116.43(7), N(2)- 3 Sm(1)-N(5) 121.78(4). Symmetry operation used to generate equivalent atoms: x-y, -y, /2-z. 106

Figure 70 Numbering scheme of the relevant atoms in triazenide complex 160 and formamidinate complex 111.

The reaction of ytterbium metal with [TlN3Dipp2] 152 using similar conditions to those used in the synthesis of [Sm(N3Dipp2)3] 160 gave a small (<1%) crop of pink-red cubes after storage of a concentrated solution in hexane for several days (Scheme 70). These were confirmed to be [Yb(N3Dipp2)3] 161 by single crystal X-ray diffraction structure determination. Further concentration of the supernatant in vacuo gave an intractable oily residue which could not be readily purified by recrystallisation. The low isolated quantities of this compound 161 precluded its full characterisation. Attempts to prepare this compound by a protolysis reaction are detailed later in the thesis.

Scheme 70 The synthesis of [Yb(N3Dipp2)3] 161 from [Tl(N3Dipp2)] 152 and ytterbium metal.

n [Yb(N3Dipp2)3] 161 crystallises as red-pink cubes from hexane isomorphously to

[Sm(N3Dipp2)3] 160 in the trigonal space group R3¯c. There is one half of the molecule in the asymmetric unit and no lattice solvent. An illustration of the structure may be seen in Figure 71. The same pseudo-trigonal planar coordination geometry is observed as for tris(triazenides) 156 and 160. The metal to nitrogen bond lengths are sizably contracted when compared to those of samarium tris(triazenide) 160. For example in

[Yb(N3Dipp2)3] 161 these range 2.354(6)-2.367(5) Å whereas in [Sm(N3Dipp2)3] 160 the equivalent range is 2.4427(19)-2.4597(18) Å. This is undoubtedly a result of the greatly reduced ionic radius of ytterbium(III) relative to samarium(III) (Six coordinate ionic 86.8 pm and 95.8 pm respectively).2 Thus, in order to maintain the same level of electrostatic stabilisation that the triazenide coordination provides, the metal to ligand bond distance has to contract. The NNN unit of the ligand (N(1)-N(2)-N(3) 114.1(5)°) is expanded relative to the equivalent angle in samarium complex 160 (113.44(18)°) however the other crystallographically independent angle, N(4)-N(5)-N(4)# is the same within error with values of 112.3(3)° (Sm 160) and 112.8(8)° (Yb 161) respectively.

Some difference in ligand geometry is observed at the CipsoNN angles (vide supra). In 107 the samarium complex 160 the mean angle is 109.7° and in the ytterbium complex 161 the mean is 109.07°. The range is also larger in ytterbium complex 161 (107.9(6)- 109.7(6)°) than in samarium complex 160 (109.2(2)-110.4(2)°). It is presumed that the steric buttressing between the arene rings is greater at the smaller ytterbium metal centre. The arene rings are swept back from the metal centre to minimise the effects of this steric buttressing. As with neodymium tris(triazenide) 156, there is not an equivalent ytterbium formamidinate complex for comparison. The closest in terms of steric demand (Figure 72)309 do not provide good comparison of structural parameters, for example the range of Y-N bond lengths in [Yb(N3Dipp2)3] 161 (2.354(6)-2.367(5) 309 Å) is of similar magnitude to that in [Yb(MesForm)3] 162 (2.348(3)-2.375(3) Å) and 309 that in [Yb(DiepForm)3] 163 (2.328(4)-2.354(5) Å) but no clear trend is evident.

Figure 71 Molecular structure of [Yb(N3Dipp2)3] 161, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. Selected bond lengths (Å) and angles (°): Yb(1)-N(1) 2.366(6), Yb(1)-N(3) 2.367(5), Yb(1)-N(4) 2.354(6), N(1)-N(2)-N(3) 114.1(5), N(4)-N(5)-N(4)# 112.8(8), C(1)–N(1)–N(2) 109.6(5), C(13)–N(3)– N(2) 107.9(5), N(2)-Yb(1)-N(5) 121.11(11), N(2)-Yb(1)-N(2)# 117.8(2). Symmetry operation used to generate 3 equivalent atoms: x-y, -y, /2-z.

Figure 72 Some reported tris(formamidinate) ytterbium complexes.309

108

As a low overall yield with respect to the lanthanide metal in thallium triazenide redox transmetallation reactions was observed, protolysis at lanthanide tris(bis(trimethylsilyl)amide) complexes was considered as an alternative methodology. This pathway capitalises on the ease of synthesis,119 the strong basicity of the amide 343 ligand (pKa 25.8 in tetrahydrofuran) and the high solubility of [Ln(N(SiMe3)2)3] complexes in non-coordinating solvents.118 To test the applicability of this method, the synthesis of [Sm(N3Dipp2)3] 160 was used as a test case. A solution of three equivalents of Dipp2N3H 40 and [Sm(N(SiMe3)2)3] in hexane was stirred for three days. Filtration followed by concentration in vacuo to the point of incipient crystallisation afforded yellow-orange cubes of [Sm(N3Dipp2)3] 160 in good yield (70%) over three crops (Scheme 71). This yield was comparable to that obtained from redox transmetallation 1 (86%). The H NMR spectrum of [Sm(N3Dipp2)3] 160 in C6D6 and the infrared spectrum are identical to the material obtained through thallium redox transmetallation (Scheme 69) and the unit cell of a single crystal, as determined by X-ray diffraction, was identical within error to that collected during the aforementioned transmetallation studies.

Scheme 71 The synthesis of tris(triazenide) lanthanide complexes by protolysis.

Buoyed by this success, the method was extended to the neodymium tris(triazenido) complex 156, which had previously been isolated as the product of a metathesis reaction. Under identical reaction conditions to those outlined above (Scheme 71) neodymium tris(triazenide) 156 was isolated in moderate yield (55%) as lattice solvent free dichroic cubes (cf. toluene inclusion occurs upon recrystallisation from toluene, see Figure 67 and relevant discussion). This dichroic behaviour is common for neodymium(III) complexes and in this instance led to a graduation in colour from yellow-green to red-orange with increasing crystal size. The yield was comparable to that obtained earlier (60%) from solvent induced redistribution. Characterisation by 1H

NMR spectroscopy in C6D6 as well as infrared spectroscopy was consistent with the product that had previously been isolated by the metathesis protocol.

109

In addition to the monoclinic structure obtained from toluene reported earlier (Figure 67), neodymium tris(triazenide) 156 crystallises from nhexane as cubes in the trigonal space group R3¯c. This structure may be seen in Figure 73. This is the same space group as the samarium and ytterbium congeners 160-161 and as in those crystals, there is one half of the molecule in the asymmetric unit. The coordination geometries within the two solid state structures of [Nd(N3Dipp2)3] 156 compare favourably with both exhibiting a pseudo-trigonal planar geometry. The metal to nitrogen bond distances in the monoclinic structure discussed earlier are statistically similar to those in the trigonal space group. For example, the three longest metal to nitrogen bonds in the monoclinic structure ((Nd(1)-N(1) 2.499(7) Å, Nd(1)-N(4) 2.503(8) Å and Nd(1)-N(9) 2.495(7) Å) are the same within error as the equivalent bonds in the trigonal structure (Nd(1)-N(3) 2.485(3) Å and Nd(1)-N(4) 2.481(3) Å). This trend continues to the three shortest bonds in the monoclinic structure (Nd(1)-N(3) 2.476(7) Å, Nd(1)-N(6) 2.475(7) Å and Nd(1)- N(7) 2.483(7) Å), which are all the same within error as the short metal to nitrogen bond in the trigonal structure (Nd(1)-N(1) 2.469(3) Å). The remaining bonding parameters e.g. the angles across NNN unit of the triazenide ligands are essentially the same within error for the two structures: in the trigonal structure N(1)-N(2)-N(3) 113.3(2)° and N(4)-N(5)-N(4)# 112.5(4)° and for the monoclinic structure N(1)-N(2)- N(3) 112.6(7)° and N(7)-N(8)-N(9) 113.2(7)°, it is reasonable to conclude that any other minor changes between the two morphologies results from crystal packing effects that are invoked upon toluene inclusion.

110

Figure 73 Molecular structure of [Nd(N3Dipp2)3] 156, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. Selected bond lengths (Å) and angles (°): Nd(1)-N(1) 2.469(3), Nd(1)-N(3) 2.485(3), Nd(1)-N(4) 2.481(3), N(2)-Nd(1)-N(2) 116.35(10), N(2)-Nd(1)-N(5) 121.82(5), N(1)-N(2)-N(3) 113.3(2), N(4)-N(5)- N(4)# 112.5(4), C(1)–N(1)–N(2) 109.6(2), C(13)–N(3)–N(2) 109.1(2). Symmetry operation used to generate 3 equivalent atoms: x-y, -y, /2-z.

n The reaction of three equivalents of Dipp2N3H 40 and [Ce(N(SiMe3)2)3] in hexane afforded the tris(triazenide) cerium complex, [Ce(N3Dipp2)3] 164 in excellent yield (83%) as bright yellow cubes (Scheme 71). The infrared spectrum of triazenide 164 displays a series of absorbances as shoulders on an intense absorbance at 1245 cm-1. This is indicative of a triazenide ligand in the chelating binding mode.201 While all of the tris(triazenide) complexes synthesised as part of this thesis are extremely sensitive to hydrolysis (vide infra) cerium compound 164 is also extremely oxygen sensitive in both the solid and solution state. A brown colour is observed in both states upon exposure to trace quantities of oxygen. This behaviour is typically indicative of the accessibility of the cerium(IV) oxidation state109,113,148,157 and was thought to be a good indicator that this complex may be chemically oxidised to afford a cerium(IV) triazenide complex (vide infra). [Ce(N3Dipp2)3] 164 is also thermally sensitive with decomposition observed over a similar range (113-117°C) to the neodymium and samarium analogues 156 and 160.

Like its neodymium, samarium and ytterbium counterparts 156, 160 and 161, n [Ce(N3Dipp2)3] 164 crystallises solvent free from hexane in the trigonal space group 111

R3¯c with one half of the molecule in the asymmetric unit. An illustration of this complex, isomorphous with the other three tris(triazenide) complexes, may be seen in

Figure 74. If one views the N3Dipp2 ligands as single point donors at N2, cerium complex 164 exhibits the same pseudo-trigonal planar coordination geometry as the other tris(triazenides) (N(2)-Ce(1)-N(5) 121.93(4)° and N(2)-Ce(1)-N(5) 116.13(9)°). There are two metal to nitrogen bonds that are statistically identical (Ce(1)-N(3) and Ce(1)-N(4), 2.518(2) Å and 2.516(3) Å respectively) as well as a shorter metal to nitrogen bond (Ce(1)-N(1) 2.498(2) Å). This trend in bond lengths is consistent with the bonding observed in the other [Ln(N3Dipp2)3] complexes 156, 160-161 and 164. These bonds are the longest observed for the [Ln(N3Dipp2)3] series, which is consistent with the larger ionic radius of cerium (six-coordinate ionic radius Ce3+ 101 pm, see Table 1 in Section 1.1) relative to the other metals.

Figure 74 Molecular structure of [Ce(N3Dipp2)3] 164, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. Selected bond lengths (Å) and angles (°): Ce(1)-N(1) 2.498(2), Ce(1)-N(3) 2.518(2), Ce(1)- N(4) 2.516(3), N(2)-Ce(1)-N(2)# 116.13(9), N(2)-Ce(1)-N(5) 121.93(4), N(1)-N(2)-N(3) 113.2(2), N(4)-N(5)-N(4)# 112.7(4), C(1)–N(1)–N(2) 109.4(2), C(13)–N(3)–N(2) 109.2(2). Symmetry operation used to generate equivalent 3 atoms: x-y, -y, /2-z.

Attempts to extend this protolysis method to the synthesis of ytterbium triazenide 161 met with similar results as the redox transmetallation (Scheme 71). Similar small yields of the desired product was observed (<1%). Attempts to dry the gum-like material obtained after removal of volatiles in vacuo for extended periods(ca. 24 h) or under

112 diffusion pump vacuum (ca. 4 h) was unsuccessful. This has thus far prevented the complete characterisation of the product of this reaction.

A summary of the structural parameters for each of the four tris(triazenide) lanthanide complexes (in the R3¯c space group) may be seen in Table 3. Some general trends become clear upon consideration of these data as a whole. Firstly, the metal to nitrogen bond lengths decrease with increasing metal radius from an average value of 2.51 Å for cerium tris(triazenide) 164 (six coordinate ionic radius 101.0 pm) to an average of 2.36 Å for ytterbium tris(triazenide) 161 (six coordinate ionic radius 86.8 pm). There is no clear change in the NNN angle of the triazenide ligands across the series, wherein the N(1)–N(2)–N(3) and N(4)-N(5)-N(4)# angles are the same within error for all four complexes. A change in ligand geometry is observed for the position of one of the arene rings as measured across C(13)–N(3)–N(2) for ytterbium complex 161 only. Otherwise the tris(triazenides) 156, 160, 161 and 164 have internal angles at this position that are the same or close to the same within error across the series.

Ce Nd Sm Yb

Ln(1)–N(1) 2.498(2) 2.469(3) 2.4427(19) 2.366(6)

Ln(1)–N(3) 2.518(2) 2.485(3) 2.4597(18) 2.367(5)

Ln(1)–N(4) 2.516(3) 2.481(3) 2.450(2) 2.354(6)

N(1)–N(2)–N(3) 113.2(2) 113.3(2) 113.44(18) 114.1(5)

N(4)-N(5)-N(4)# 112.7(4) 112.5(4) 112.3(3) 112.8(8)

C(1)–N(1)–N(2) 109.4(2) 109.6(2) 109.50(17) 109.6(5)

C(13)–N(3)–N(2) 109.2(2) 109.1(2) 109.21(18) 107.9(5)

C(25)-N(4)-N(5) 111.0(3) 110.6(3) 110.4(2) 109.7(6)

Table 3 Summary of the bonding metrics of the tris(triazenide) lanthanide complexes in the solid state.

113

Figure 75 Numbering scheme for [Ln(N3Dipp2)3] complexes 156, 160, 161 and 164. It is also instructive to compare the steric demand of the triazenide ligands at complexes 156, 160, 161 and 164 with the steric demand of some common lanthanide metallocenes. The literature complexes chosen for comparison are illustrated in Figure 76.

Figure 76 Two lanthanocenes, a tris(formamidinate) and the tris(triazenide) complexes, for which ligand steric properties have been calculated.

Of the measures available for quantification of ligand steric demand, the cone angle is often used to describe the spatial demand of planar cyclopentadienyl ligands, especially when rationalising the reactivity of tris(pentamethylcyclopentadienyl) lanthanides.55,57 This is not directly applicable to ligand geometries that are not and therefore exhibit low rotational symmetry about the metal ligand bond vector.344-347 Recently the measurement of the percentage of the metal coordination sphere that is enshrouded by the ligand at the metal centre has gained traction as a useful steric parameter.348 This method begins by calculating the ligand’s solid angle, Ω in radians. Given a sphere surrounding the metal complex at an arbitrary radius, r, a ligand will eclipse an area of this sphere when viewed from the perspective of the metal centre. Ω is calculated by dividing the area obscured by the ligand on the surface of the sphere by the square of the radius (Equation 1).

퐴 훺 = 푟2

Equation 1 The calculation of the ligand solid angle, Ω. 114

From this value, Ω, the parameter, G may be calculated according to Equation 2. This parameter is defined as percentage of the metal’s coordination sphere that is shielded by a given ligand. It is therefore applicable to a range of ligand geometries and was chosen for the steric comparison of the tris(triazenide) lanthanide complexes 156, 160, 161 and 164 with trivalent lanthanocenes.

Ω 퐺 = 100 4휋

Equation 2 The calculation of the G parameter for a ligand.

The data below were calculated using the published molecular structures as determined from single crystal X-ray diffraction and deposited with the Cambridge Crystallographic Data Centre. Crystallographic information files (*.cif) were converted to atomic coordinate files (*.xyz) which were then used as inputs for the program Solid- G,348 which is used to calculate the parameter G. Solid-G is a powerful tool for ligand steric analysis and provides other output data that is included here for discussion. In particular it is unique in its calculation of steric parameters for an entire metal complex.

For some of the complexes discussed below ([CeCp″3] 168, [NdCp*3] 166 and

[NdCp″3] 169) the reported molecular structure includes disordered sites. Lower occupancy atomic positions, as modelled in the original refinements, were removed in order to avoid artificially inflated the ligand sterics. Full output files from the calculations for each metal complex are included with the *.cifs for this thesis.

The two important measurements included below are the Gmetal(L) value, which is the measure of the percentage of the metal’s coordination sphere that is shielded by a ligand and is expressed as a percentage value (vide supra) and the Gmetal(Complex) value, which is conceptually the same as the above value but treats all the ligands about a metal centre as if they were one ligand and thus in combination with Gmetal(L) values may be used to calculate ligand overlap at the metal centre. Relative to the accumulated individual ligand G values, this parameter can be considered a more relevant measure of the congestion at the metal as it accounts for the interleaving of adjacent ligands, which typically characterises steric strain, to be quantified. A significant difference between the values for Gmetal(Complex) and Σ Gmetal(L) indicates ligand buttressing and overlap and hence Σ Gmetal(L) is also included.

115

[CeCp* ] [CeCp″ ] [Ce(N Dipp ) ] 3 3 3 2 3 26.23 30.52 30.76 GCe(L)/% 26.26 27.71 31.32 26.26 30.70 30.79

Σ GCe(L)/% 78.76 88.93 92.87

GCe(Complex)/% 78.75 88.79 92.58

Table 4 G parameters for cerium complexes 165, 167 and 164.59,349

[NdCp* ] [NdCp″ ] [Nd(N Dipp ) ] 3 3 3 2 3 26.51 30.99 31.02 GNd(L)/% 26.26 30.01 31.54 26.27 29.57 31.02

Σ GNd(L) /% 79.98 90.56 93.58

GNd(Complex)/% 79.04 89.90 93.06

Table 5 G parameters for neodymium complexes 166 168 and 156.350,351

[SmCp*3] [SmCp″3] [Sm(DippForm)3] [Sm(N3Dipp2)3] 26.92 30.74 31.55 31.17 GSm(L)/% 26.91 29.58 31.61 31.16 26.88 31.21 31.66 31.76

Σ GSm(L) /% 80.71 91.53 94.82 94.10

GSm(Complex)/% 80.59 90.78 93.93 93.39

Table 6 G parameters for samarium complexes 1, 169, 111 and 160.310,352,353

[YbCp″3] [Yb(N3Dipp2)3] 31.31 31.94 GYb(L)/% 32.03 32.50 30.98 32.01

Σ GYb(L) /% 94.32 96.45

GSm(Complex)/% 92.98 94.46

Table 7 G parameters for ytterbium complexes 170 and 161.354

116

Selected data from the output files for the complexes in Figure 76 from the program Solid-G are displayed in Tables 4-7. Upon consideration of these data, three general trends emerge. Firstly in all series, the ligand occupies a larger percentage of the metal coordination sphere, as expressed by both the individual ligand values for Gmetal(L) and

Gmetal(Complex), with decreasing metal radius. For instance, the Gmetal(Complex) is

92.58% at [Ce(N3Dipp2)3] 164 (Table 4) and increases to 94.46% at [Yb(N3Dipp2)3] 161 2 (Table 7). These trends are roughly linear (R = 0.9929 for N3Dipp2, 0.9843 for Cp″ and 0.8493 for Cp*) and may be seen graphically in Figure 77. These data would benefit from inclusion of data for lanthanide ionic radii between ytterbium and samarium. Holmium, with a six-coordinate ionic radius of 90.1 pm (Table 1) would be ideal.

100 Cp* Cp" N3Dipp2 N3Dipp2

95 y = -0.129x + 105.69

R² = 0.9929

)/% 90 y = -0.286x + 117.92 R² = 0.9843

(Complex 85 metal G y = -0.3506x + 113.94 R² = 0.8493 80

75 86 88 90 92 94 96 98 100 102 Six Coordinate Ionic Radius/ pm

Figure 77 A plot of six coordinate ionic radius vs. Gmetal(Complex).

The second trend is that for a given homoleptic lanthanide complex the triazenide ligand is either equal to or exceeds the bis(trimethylsilyl)cyclopentadienyl ligand in steric demand at the metal centre as measured by Gmetal(L). When the values for

Gmetal(Complex) are considered, the tris(triazenide) complexes are more sterically congested than the corresponding tris(bis(trimethylsilyl)cyclopentadienyl) complexes

117 by margins of several percent. For example at cerium GCe(Complex) is 92.58% for

[Ce(N3Dipp2)3] 164 and 88.79% for CeCp″3 167 (Table 4). In other words, the cerium is more sterically congested by 3.79%. This margin gradually decreases until at ytterbium

Δ GYb(Complex) 1.48% with values of GYb(L) 94.46% for [Yb(N3Dipp2)3] 161 and

92.98% for YbCp″3 170 (Table 7). This is gratifying as the bis(trimethylsilyl)cyclopentadienyl ligand has proven to be an excellent support ligand for classical and non-classical divalent lanthanides68,88,90,95,97,99,355 and the tris(triazenide)complexes were synthesised with similar chemistry in mind. Thus on steric grounds alone this ligand choice is validated.

The third trend of note is that the well-known tris(pentamethylcyclopentadienyl) lanthanide complexes are actually the least sterically congested complexes by these calculations. Unfortunately the molecular structure of [YbCp*3] has not been determined and thus discussion of its steric parameters, while potentially the most interesting due to the small ionic radius of ytterbium, cannot be included here. In spite of this tentative predictions may still be made concerning such unknown complexes; for example [LnCp*3] (Ln = Yb, Lu) would be expected to have Gmetal(Complex) values of 83.5% and 83.8% respectively. It is doubtful that these complexes will ever be made due to the proposed high reactivity of these sterically congested metallocenes356,357 and thus the accuracy of this prediction is unlikely to be tested. It does suggest however that based upon this steric measure these metallocenes are sterically possible. In all cases the value for Gmetal(Complex) is at least 12% higher for the tris(triazenide) complexes than the pentamethylcyclopentadienyl complexes, ranging from Δ GCe(Complex) of 13.83%

(Table 4) to Δ GSm(Complex) of 12.8% (Table 6). This is also encouraging as it validates the choice of the tris(triazenide) complexes as potential targets for investigating the potential of sterically induced reduction (SIR) at complexes other than tris(pentamethylcyclopentadienyl) lanthanides, which dominate research in this field.55,57,74

It is also noteworthy that for samarium it is possible to directly compare the tris(triazenide) 160 and tris(formamidinate) 111 complexes (Table 6). The formamidinate ligand (Range GSm(DippForm) 31.55-31.66%) and its triazenide analogue (Range GSm(N3Dipp2) 31.16-31.76%) are very similar in steric demand. A

118 slightly more sterically congested metal centre is observed at [Sm(DippForm)3] 111

(GSm(Complex) 93.93%) than at [Sm(N3Dipp2)3] 160 (GSm(Complex) 93.39%).

3.2.2 Reactivity of Trivalent Lanthanide Triazenides

As previously mentioned the hydrolysis of the tris(triazenide)lanthanide compounds occurs very rapidly. This sensitivity was amplified for [Ce(N3Dipp2)3] 164. In spite of the usual rigorous precautions,358,359 hydrolysis and oxidation products will often occur 63,309,360-362 in non-aqueous f-block chemistry. In one preparation of [Ce(N3Dipp2)3] 164, after the removal of a crop of single crystals for an X-ray structure determination the supernatant was observed to darken upon standing at room temperature over the period of one week with concomitant re-dissolution of the deposited yellow cubes of

[Ce(N3Dipp2)3] 164 and deposition of a sizeable crop of large deep orange blocks. This reaction was allowed to occur in order to fully characterise the decomposition product. Like that of the parent compound, the 1H NMR spectrum of the isolated orange blocks in C6D6 was not instructive displaying a highly complicated spectrum with significant paramagnetic broadening, shifting and a high number of chemical environments. This suggested that a paramagnetic cerium(III) complex had been retained and that an oxidised cerium(IV) product was unlikely. The infrared spectrum has a strong sharp absorbance at 3064 cm-1. This is consistent with one of the absorbances that is observed to appear upon brief exposure of infrared samples of the tris(triazenide)lanthanide complexes 156, 160, 161 and 164 to the atmosphere and suggests that the same product is formed upon exposure of [Ce(N3Dipp2)3] 164 to adventitious water or oxygen in solution. Final confirmation of the identity of the decomposition product was provided by a single crystal X-ray diffraction experiment, which identified the product as a co- crystallised mixture of [{Ce(N3Dipp2)2(μ-OH)} 2] 171 and [{Ce(N3Dipp2)2}2(μ-O)] 172 (Scheme 72).

Scheme 72 The hydrolysis of [Ce(N3Dipp2)3] 164 by adventitious water.

119

Co-crystallisation of a μ-hydroxo and μ-oxo species is highly unusual and has not been observed to-date. However Mak has observed the crystallisation of multiple constitutional isomers of the formula Cp″2SmO2H2SmCp″2. A recent report of the co- 2– 2– crystallisation of yttrium (μ-S2) and (μ-S) amides with a similar arrangement of electron density peaks between the metals (Scheme 73) supports the possibility of a similar OH/O formulation (Scheme 73).363

2– 2– 363 Scheme 73 The synthesis of yttrium (μ-S2) and (μ-S) amides.

Investigations of the hydrolysis of lanthanocenes have also suggested that the hydrolysis reaction is a dynamic process that proceeds through a bis μ-hydroxo to mono μ-oxo mono μ-aqua and finally mono μ-oxo compounds.364 A possible pathway to decomposition is provided in Scheme 74. In contrast to the thermally frail parent compound [Ce(N3Dipp2)3] 164, which decomposes between 113-117 °C, this co- crystallised dimer is extremely thermally robust in the solid state, decomposing between 209-212 °C. In addition to the μ-hydroxide OH absorbance mentioned earlier, the infrared spectrum of the co-crystals 171 and 172 displays an absorbance at 1256 cm-1 that is consistent with a coordinated triazenide bonding mode.201

Scheme 74 The proposed pathway leading to the isolated hydrolysis products.

The mixture of 171 and 172 crystallises from nhexane as large orange blocks in the monoclinic space group P21/n. One full molecule is included in the asymmetric unit in the relative proportions of 58:42 for [{Ce(N3Dipp2)2(μ-OH)} 2] 171 to [{Ce(N3Dipp2)2}2 (μ-O)] 172. An illustration of the structure may be seen in Figure 78 and Figure 79, for clarity the structure has been separated into the two constituent components in the two figures. The closest structural relatives to both the μ-hydroxo and μ-oxo may be seen in Figure 80. The field of analogues for cerium(III) may be widened by the addition of uranium(III) metallocenes,360,365 which can be considered structurally related due to the similar ionic radii of cerium(III) and uranium(III) (Six coordinate ionic radii Ce3+ 101 pm, U3+ 102.5 pm).2 In the solid state the two triazenide ligands bonded to one cerium atom are oriented orthogonal to the triazenide ligands bonded to the opposite cerium

120 atom. The dihedral angles as defined by the three bond vectors are N(11)-Ce(2)-Ce(1)- N(5) (-81.8°), N(8)-Ce(2)-Ce(1)-N(2) (-81.4°), N(5)-Ce(1)-Ce(2)-N(8) (100.0°) and N(11)-Ce(2)-Ce(1)-N(2) (96.8°). This is likely to minimise steric buttressing in the solid state. The hydroxyl oxygen atoms and the two cerium atoms are co-planar. The distances Ce(1)–O(1) (2.375(12) Å), Ce(1)–O(2) (2.363(9) Å), Ce(2)–O(1) (2.472(10) Å), Ce(2)–O(2) (2.405(10) Å) are consistent with the formulation as a bridging μ- hydroxo dimer. The alternative bridging bis(μ-oxo) dimer formulation, would be expected to feature much shorter Ce-O bond lengths such as those observed for Ce(1)- O(3) (1.984(14) Å) and Ce(2)-O(3) (2.046(14) Å) in the mono(μ-oxo) component, in turn leading to larger O-Ce-O angles than those observed (Ce(1)-O(1)-Ce(2) 113.6(4)° and Ce(1)-O(2)-Ce(2) 113.8(4)°). These distances in themselves are shorter than has been previously observed for a bridging μ-oxo ligand bound to cerium(III) or uranium(III) however many of these compounds, particularly the cerium amides, contain cerium(IV) rather than cerium(III) observed here, which limits their utility as meaningful comparison. The bond lengths between cerium(III) or uranium(III) and a bridging μ-oxo ligand have been observed to range from 2.094-2.185 Å.63,360,366 Unfortunately there are no structurally characterised dimeric cerium μ-hydroxo compounds, however there is a μ-methoxy compound known 173.95 This compound has crystallographically independent Ce-O distances of 2.366(4) Å and 2.386(4) Å, both of which are the same within error as the Ce(1)–O(1) and Ce(1)–O(2) distances in triazenide complexes 171 and 172. The crystallographically independent uranium to oxygen distances in complex 174 are 2.295(3) Å and 2.299(3) Å, both of which are similar in magnitude but different outside of error from those observed in triazenide complexes 171 and 172. The closest amidinate congener is ytterbium formamidinate complex 175,309 which has a different coordination number relative to

[{Ce(N3Dipp2)2(μ-OH)}2] 171 and a much smaller ionic radius. As such ytterbium formamidinate 175 adopts a different spatial arrangement of the ligands in the solid state. This is also likely due to the reduced steric demand of the o-tolyl formamidinate ligands relative to the diisopropylphenyltriazenide herein. The single crystallographically independent Yb-OOH distance in formamidinate 175 is 2.202(6) Å, which is significantly shorter than the Ce-O distances found in [{Ce(N3Dipp2)2(μ-

OH)}2] 171 (2.363(9)-2.472(10) Å). This is unsurprising given the large difference in ionic radii between cerium(III) and ytterbium(III) (14.2 pm, Table 1). The angles within

121 the Ce2O2 metallacycle (67.9(4)° and 113.6(5)°) are similar but different outside of error to those found in the aforementioned formamidinate 175 (67.9(2)° and 112.1(2) °). The distances Ce(1)-O(3) (1.984(14) Å) and Ce(2)-O(3) (2.046(14) Å) in triazenide 172 are much shorter than the equivalent distances in the only reported dimeric cerium μ-oxo complexes: 2.1405(3) Å for complex 176;63 2.185(4) and 2.183(5) Å for complex 178.366 There are also much shorter than the same distance in uranium complex 177 (2.094(14) and 2.125(13) Å).360 The angle Ce(1)-O(3)-Ce(2) (175.9(7)°) is however within the range observed for cerium and uranium μ-oxo complexes 176-178 (171.5(6)- 180.00°)63,360,366

Figure 78 Molecular structure of [{Ce(N3Dipp2)2(μ-OH)}2] 171, POV-RAY illustration, 50% thermal ellipsoids, all non-hydroxo hydrogen atoms omitted for clarity. The isopropyl carbon atoms are depicted as wireframes for clarity. Selected bond lengths (Å) and angles (°): Ce(1)–N(1) 2.482(7), Ce(1)–N(3) 2.567(6), Ce(1)–N(4) 2.564(6), Ce(1)– N(6) 2.487(7), Ce(1)–O(1) 2.375(12), Ce(1)–O(2) 2.363(9), Ce(2)–O(1) 2.472(10), Ce(2)–O(2) 2.405(10), Ce(1)·· ·Ce(2) 4.0274(8), Ce(1)-O(1)-Ce(2) 113.6(4), Ce(1)-O(2)-Ce(2) 113.8(4), N(1)–N(2)–N(3) 110.2(7), N(4)–N(5)– N(6) 111.7(6).

122

Figure 79 Molecular structure of [{Ce(N3Dipp2)2(μ-O)}2] 172, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. The isopropyl carbon atoms are depicted as wireframes for clarity. Selected bond lengths (Å) and angles (°): Ce(1)-O(3) 1.984(14), Ce(2)-O(3) 2.046(14), Ce(1)-O(3)-Ce(2) 175.9(7).

Figure 80 Literature compounds for comparison to co-crystallised 171 and 172.63,95,309,360,365,366

It was anticipated that the high degree of steric congestion at the tris(triazenide) lanthanide complexes would provide for highly stable divalent triazenide analogues of the recently discovered divalent lanthanide metallocenes (See Figure 9 in Section 1.3.2), + 41-44,97 i.e. the formation of M [Ln(N3Dipp2)3] type complexes. A suspension of excess 91-94 potassium graphite, one equivalent of 18-crown-6 and [Nd(N3Dipp2)3] 156 in toluene was stirred at -25 °C, whereupon a colour change from yellow-green to red was observed. These conditions were modelled upon the successful isolation of non-classical divalent metallocenes by reduction of trivalent precursors (See Scheme 13 in Section

123

1.3.2).92-94 Filtration of the suspension and removal of the solvent in vacuo afforded no solid material. Attempts to extract the product from the precipitate with other common solvents were non-productive thus prohibiting the isolation of the reaction product. Analysis of the remaining insoluble material was not possible due to contamination with the excess potassium graphite and graphite. This was taken as indicative of decomposition of the lanthanide complex as solubility problems have not been encountered with other complexes of this ligand as part of this thesis. This was surprising given the favourable steric properties of [Nd(N3Dipp2)3] 156 (Table 5) when compared to the tris(bis(trimethylsilyl)cyclopentadienyl) lanthanide complexes, which were used so successfully for the isolation of the non-classical divalent lanthanides.97

Scheme 75 The attempted reaction of [Nd(N3Dipp2)3] 156 with KC8 and 18-crown-6. Using the calculated G parameters as a guide (Table 4-7) it was anticipated that the

[Ln(N3Dipp2)3] 156, 160, 161 and 164 isolated herein may also be suitable for sterically induced reduction (SIR) chemistry (See Scheme 7 in Section 1.3.1). Due to the large amount of published data on SIR at [SmCp*3] 1 for comparison it was decided that

[Sm(N3Dipp2)3] 160 presented a viable test case for this reactivity (Scheme 76). The reaction of O=PPh3 with two equivalents of [Sm(N3Dipp2)3] 160 was chosen due to the two possible spectroscopic handles available for characterisation of the reaction 1 31 products: H and P NMR spectroscopy, particularly of free PPh3 and O=PPh3. Unfortunately in this instance, no reaction was observed in toluene over 24 hours at room temperature as determined by examination of the 1H and 31P NMR spectra of the reaction mixture. This reaction outcome was considered puzzling due to the vastly greater steric congestion at [Sm(N3Dipp2)3] 160 when compared to [SmCp*3] 1 (Table

6). It is proposed that the key steric factor that drives SIR for [SmCp*3] 1 is simply the steric congestion at the metal centre as had been hypothesised55-57,74 but the distortion of the Cp* ligand away from its ideal geometry as a result of the steric congestion.350,352 In the instance of tris(pentamethylcyclopentadienyl)lanthanides such congestion is generally considered to induce elongation of the Ln-Cp* contact as well as distortion of the Cp* ligands away from planarity.55,60 It is conceivable that the greater coordinative

124 flexibility of the triazenide ligand relative to the pentamethylcyclopentadienide ligands of the [LnCp*3] complexes alleviates the electronic and steric destabilisation that triggers the SIR reactivity.

Scheme 76 The attempted reaction of Sm(N3Dipp2)3 160 with O=PPh3. 3.2.3 Synthesis of Divalent Lanthanide Triazenides

The failure to isolate a non-classical divalent lanthanide triazenide prompted the investigation of the chemistry of this ligand with classical divalent lanthanides, in this instance ytterbium and samarium. The reaction of two equivalents of [NaN3Dipp2] 75 with a solution of ytterbium(II) iodide in tetrahydrofuran afforded [Yb(N3Dipp2)2(thf)2] 179 in good yield (Scheme 77) after recrystallisation from nhexane. Unsurprisingly,

[Yb(N3Dipp2)2(thf)2] 179 is extremely sensitive to hydrolysis and oxidation. This fact prevented the attainment of satisfactory microanalytical data due to repeated sample decomposition pre-analysis.

Scheme 77 The synthesis of [Sm(N3Dipp2)2(thf)2] 182 and [Yb(N3Dipp2)2(thf)2] 179 from SmI2 and YbI2 respectively.

1 The H NMR spectrum of diamagnetic ytterbium(II) complex 179 in C6D6 displays a doublet for the methyl groups at δ 1.16 and a septet for the methine protons δ 3.46. The methylene signals of the tetrahydrofuran ligands are observed as extremely broad singlets in the 1H NMR spectrum. Given the diamagnetic nature of ytterbium(II) (Yb2+ [Xe] 4f14) rather than paramagnetic broadening this is suggestive of a labile coordination environment in C6D6 solution. For the signal at δ 3.73, which corresponds to the α protons of the tetrahydrofuran ligands the signal’s full width at half maximum is 124 Hz. The peak at δ 1.37, corresponding to the β protons of the tetrahydrofuran ligand, cannot be as readily measured due to signal overlap with the doublet at δ 1.16. The 13C

125

NMR of [Yb(N3Dipp2)2(thf)2] 179 in C6D6 spectrum is devoid of signals for the tetrahydrofuran ligands. Data collected over a considerable number of scans with extended relaxation delay times is also bereft of the expected methylene signals. In combination with the observed broadening of the methylene proton signals in the 1H NMR spectrum these data are indicative of either a ligand exchange process involving a 173 5 highly quadrupolar ytterbium nucleus, such as Yb (I = /2), that results in NMR silent methylene carbon atoms or a 13C relaxation process at ytterbium(II) that is vastly different from the time scale of the NMR experiment. The infrared spectrum of

[Yb(N3Dipp2)2(thf)2] 179 displays absorbances that can be attributed to the infrared active C-O modes of coordainted tetrahydrofuran at 865 and 1057 cm-1 as well as an -1 intense band at 1268 cm , which is consistent with the chelation of the triazenide ligands. Ytterbium bis(triazenide) 179 is observed to decompose in the solid state over the range 125-128 °C. This continues the trend of limited thermal stability of the lanthanide triazenides ([Ln(N3Dipp2)3] dec. range 106-121 °C) when compared to the alkali metal complexes (See Section 2.2.3 in Chapter II, dec. >200 °C).

n [Yb(N3Dipp2)2(thf)2] 179 crystallises from hexane as lattice solvent free carmine blocks in the monoclinic space group P21/c with two unique molecules in the asymmetric unit. Both molecules exhibit similar bonding parameters thus only one molecule will be discussed further. An illustration of one molecule from the asymmetric unit may be seen in Figure 81. In the solid state the two ligand pairs are bonded in a cis geometry about the ytterbium centre, if the triazenide ligands are considered as single point donors. The N3 units of the triazenide ligands orient such that the YbN3 metallacycles are close to orthogonal to one another to minimise steric repulsion between the sterically demanding 2,6-diisopropylphenyl N-substituents. This structure may be compared to not only the structurally related ytterbium bis(formamidinate) 311 complexes [Yb(DippForm)2(thf)2] 110 but also the calcium compound reported by 200,367 Hill et al [Ca(N3Dipp2)2(thf)2] 180. due to the similar ionic radii of calcium and ytterbium(II) (Six coordinate ionic radii: Ca2+ 100 pm, Yb2+ 102 pm).2 In this instance the calcium analogue crystallises in the trans geometry. Adopting the convention whereby the triazenides are considered as single point donors centred at N(2) and N(5), calcium complex 180 can be considered to be a pseudo square planar complex (e.g. Othf- 367 Ca-Othf 164.10(5)°). In contrast [Yb(N3Dipp2)2(thf)2] 179 is better described as heavily distorted pseudo tetrahedral under the same point donor convention (e.g. O(1)- 126

Yb(1)-O(2) 83.5(1)°, N(2)-Yb(1)-N(5) 137.6(1)°). The NNN angles in

[Yb(N3Dipp2)2(thf)2] 179 are statistically similar (N(1)-N(2)-N(3) 111.9(4)° and N(4)-

N(5)-N(6) 111.7(4)°) to those observed in [Ca(N3Dipp2)2(thf)2] 180 (112.4(2)° and 111.0(1)°) consistent with their similar ionic radii. The analogous calcium formamidinate complex 181 crystallises as a mono(tetrahydrofuran) complex (Figure 82).368 Thus the bis(2,6-diisopropylphenyl)formamidinate and -triazenide complexes of calcium and ytterbium(II) represent a rare instance where the conventional calcium:ytterbium structural analogy breaks down.

Figure 81 Molecular structure of one unique molecule of [Yb(N3Dipp2)2(thf)2] 179, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. The tetrahydrofuran carbon atoms are depicted as wireframes for clarity. Selected bond lengths (Å) and angles (°): Yb(1)-N(1) 2.454(4), Yb(1)-N(3) 2.546(4), Yb(1)- N(4) 2.475(4), Yb(1)-N(6) 2.550(4), Yb(1)-O(1) 2.442(4), Yb(1)-O(2) 2.438(4), N(1)-N(2)-N(3) 111.9(4), N(4)- N(5)-N(6) 111.7(4).

When compared to the ytterbium formamidinate analogue some differences also emerge. Firstly [Yb(DippForm)2(thf)2] 110 crystallises with a half molecule in the asymmetric unit due to increased symmetry but with the same overall cis geometry of the ligand pairs. Secondly in [Yb(DippForm)2(thf)2] 110 long and short metal to nitrogen bonds are observed: Yb(1)-N(1) 2.462(2) Å and Yb(1)-N(2) 2.496(2) Å. A similar long and short bond pattern is observed for each of the triazenide ligands in

[Yb(N3Dipp2)2(thf)2] 179, however overall the Yb-N bonds are significantly longer than those in the formamidinate complex 110. The reverse is true for the tetrahydrofuran 127 ligand bond lengths, Yb(1)-O(1) 2.442(4) Å and Yb(1)-O(2) 2.438(4) Å in

[Yb(N3Dipp2)2(thf)2] 179 versus Yb(1)-O(1) 2.461(2) Å in [Yb(DippForm)2(thf)2] 110. It is plausible that the longer bond lengths of the triazenide and its reduced donation vs the formamidinate lead to shortening of the Yb-O bonds. It is also likely that the reduced steric strain incurred by longer Yb-N bonds at the ytterbium metal centre invites shorter ytterbium to oxygen bonds.

Figure 82 Calcium and ytterbium complexes considered for comparisons of solid state structure to 200,311,368 [Yb(N3Dipp2)2(thf)2] 179. In a fashion analogous to ytterbium complex 179, the samarium bis(triazenide) complex

[Sm(N3Dipp2)2(thf)2] 182 is obtained as malachite plates from the reaction of a solution of samarium(II) iodide in tetrahydrofuran with two equivalents of [NaN3Dipp2] 75 followed by recrystallisation from nhexane (Scheme 77). Unlike ytterbium complex 179

(yield 66%), [Sm(N3Dipp2)2(thf)2] 182 was obtained in poor yield (35%). The remaining mass balance is retained by a hexane insoluble precipitate. It is likely that competing side reactions, such as those encountered during the synthesis of 310 [Sm(DippForm)2(thf)2] 108 account for the mass balance (Scheme 78). At this stage an analogue of the sodium samarate(III) 183 compound depicted in Scheme 78 has not 310 been isolated. In some preparations, a trace amount of [Sm(N3Dipp2)3] 160 was observed and identified by a single crystal X-ray unit cell determination. This lends credence to the proposal of complex redox equilibria during the reaction of SmI2 with

[NaN3Dipp2] 75. It is also possible however that a trivalent co-product results from oxidation by adventitious oxygen (According to published specifications, the purity of argon used for all experimental work, ALPHAGAZ™ 1 contains O2 < 2 ppm) as

[Sm(N3Dipp2)2(thf)2] 182 is visibly more sensitive to oxidation and hydrolysis than its ytterbium analogue (vide supra).

128

310 Scheme 78 By-products encountered in the synthesis of [Sm(DippForm)2(thf)2] 108. The infrared spectrum for samarium triazenide 182 contains a strong absorbance at 1257 cm-1, which is consistent with a chelating triazenide ligand as well as bands for the coordinated tetrahydrofuran ligands at 1056 cm-1 and 876 cm-1. Despite the strong paramagnetism of samarium(II) (Sm2+ [Xe] 4f 6) instructive NMR spectra can be recorded. This contrasts with the attempts to record spectra of [Sm(N3Dipp2)3] 160 (vide 1 supra). The H NMR spectrum of [Sm(N3Dipp2)2(thf)2] 182 in C6D6 displays broad (fwhm = 36 Hz) resonances that can be attributed to the coordinated tetrahydrofuran ligands at δ 0.53 and δ 3.53. The methyl groups correspond to a resonance at δ 2.84 and the methine protons are paramagnetically shifted far downfield to δ 5.05. This carbon is 13 also shifted far down field in the C NMR spectrum in C6D6, corresponding to the signal observed at δ 119.13. Like its ytterbium congener, samarium triazenide 182 also displays limited thermal stability in the solid state, decomposing over the range 120-124 °C.

n [Sm(N3Dipp2)2(thf)2] 182, crystallises from hexane as malachite plates in the triclinic space group P1¯ (cf. [Yb(N3Dipp2)(thf)2] P21/c) with one full molecule in the asymmetric unit. An illustration of the structure may be seen in Figure 83. One isopropyl group is observed to be disordered and has been successfully modelled across two locations. One of the tetrahydrofuran ligands is heavily disordered across multiple locations and could not be modelled successfully. To counter this it has been restrained during refinement, meaning that the bond lengths about this ligand cannot be considered as reliable. The remainder of the structure is crystallographically well behaved. Just as structural data for ytterbium complexes can be compared to calcium, it is considered equally valid to compare the structural data of strontium complexes to those of isostructural samarium complexes due to their similar ionic radii (Six coordinate ionic radii Sr2+ 118 pm, Sm2+ 115 pm).2,369 In this instance, the molecular structure of the isostructural strontium complex [Sr(N3Dipp2)2(thf)2] 184 has been published (Figure 200 84). Strontium triazenide 184 adopts the same cis geometry as [Sm(N3Dipp2)2(thf)2] 182 in the solid state but crystallises in a higher symmetry space group, with only one 129 half molecule in the asymmetric unit. Interestingly a similar pattern of disorder is observed in [Sr(N3Dipp2)2(thf)2] 184 as in [Sm(N3Dipp2)2(thf)2] 182. The metal to nitrogen bond lengths are longer at the strontium complex than for the samarium complex. For example in [Sr(N3Dipp2)2(thf)2] 184 the metal to nitrogen bond lengths are Sr(1)-N(1) 2.596(16) Å and Sr(1)-N(3) 2.604(2) Å. Both of these values are longer than three of the four samarium to nitrogen bonds with the exception of Sm(1)-N(3) 2.581(4) Å, which is the same as Sr(1)-N(1) (2.596(16) Å) within error. Samarium triazenide complex 182 also features the familiar long and short metal to nitrogen bond for each triazenide ligand. This contrasts with the strontium triazenide complex where both of the metal to nitrogen bond lengths are the same within error. In

[Sr(N3Dipp2)2(thf)2] 184 Sr(1)-O(1) (2.467(16) Å) is much shorter than the equivalent distance in [Sm(N3Dipp2)2(thf)2] 182, Sm(1)-O(1) 2.529(3) (Figure 83). It is likely this is a consequence of the reduced steric buttressing at the strontium metal centre due to the reduced proximity of the triazenide ligands. The angle across the triazene bonding unit is relatively consistent between the two complexes. For strontium triazenide 184 N(1)-N(2)-N(3) this is 111.0(3)°. The same angles for samarium triazenide 182 are 111.1(4)° and 111.9(3)°. With respect to the isostructural samarium formamidinate 310 complex 108 (Figure 84), surprisingly [Sm(N3Dipp)2(thf)2] 182, has shorter metal to nitrogen bonds overall. This is contrary to expectation owing to the reduced donicity of the triazenide ligands relative to the formamidinates.

130

Figure 83 Molecular structure of [Sm(N3Dipp2)2(thf)2] 182, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms. The tetrahydrofuran carbon atoms are depicted as wireframes for clarity. Selected bond lengths (Å) and angles (°): Sm(1)-N(1) 2.508(4), Sm(1)-N(3) 2.581(4), Sm(1)-N(4) 2.571(3), Sm(1)-N(6) 2.538(4), Sm(1)-O(1) 2.529(3), Sm(1)-O(2) 2.368(4), N(1)-N(2)-N(3) 111.1(4), N(4)-N(5)-N(6) 111.9(3).

Figure 84 A strontium bis(triazenide) 184 and samarium bis(formamidinate) 108 selected for comparison to

[Sm(N3Dipp2)2(thf)2] 182.

Despite its strongly reducing nature, thulium(II) iodide may be synthesised by solution phase methods22,23 without the specialised equipment necessary for the dysprosium and neodymium diiodides (See Section 1.2 in Chapter I).28 A solution of thulium(II) iodide tetrahydrofuran complex was generated using the method of Bochkarev22 and then treated with two equivalents of [NaN3Dipp2], also in tetrahydrofuran at room temperature (Scheme 79). Instantaneous oxidation of the metal centre was observed with fading of the intense malachite of thulium(II) iodide to a pale yellow-brown colouration indicative of thulium(III). Repetition of the reaction at -30 °C gave rise to deep brown-green solution. Unfortunately this colour also faded to pale yellow-brown upon warming to room temperature indicating oxidation had again occurred. A similar deep brown-green colour was observed when a solution of two equivalents of

131

NaN(SiMe3)2 in tetrahydrofuran was added to thulium(II) iodide at -30 °C, consistent 137 with the reported colour of [Tm(N(SiMe3)2)2(thf)n] in solution. The addition of a solution of two equivalents of Dipp2N3H 40 in tetrahydrofuran to this intermediate followed by warming to room temperature also gave the same relatively weakly coloured oxidation product. Removal of volatiles in vacuo followed by extraction into and recrystallisation from nhexane in all cases afforded a large amount of amorphous material and several single crystals suitable for single crystal X-ray diffraction (vide infra). This allowed the crystalline product to be characterised as [TmI(N3Dipp2)2(thf)2] 185. This outcome is consistent with the observations of Bochkarev,89 who observed the formation of a thulium(III) mono iodide complex during attempts to synthesise bis(pentamethylcyclopentadienyl)thulium(II) (See Scheme 12 in section 1.3.2). Microanalysis of the bulk material of these reactions is inconclusive about the nature of the product and whether the single crystals give the identity of the only product. The 88,89,137 intense paramagnetism of thulium(III) (μeff (298K) 6.9-7.7 μB ) also prevents the meaningful application of NMR spectroscopy for characterisation of thulium complex 185.

Scheme 79 The reaction of TmI2 with an alkali metal triazenide.

n [TmI(N3Dipp2)2(thf)2] 185, crystallises from hexane as yellow blocks in the orthorhombic space group Pbca with one full molecule in the asymmetric unit. The thulium centre adopts a distorted pentagonal bipyramidal geometry in the solid state with an iodide ligand and a tetrahydrofuran at the apical positions of the bipyramid. The geometry is heavily distorted as a consequence of the chelating nature of and high steric repulsion between the two triazenide ligands.

132

Figure 85 Molecular structure of [TmI(N3Dipp2)2(thf)2] 185, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. The tetrahydrofuran carbon atoms are depicted as wireframes for clarity. Selected bond lengths (Å) and angles (°): Tm(1)-N(1) 2.486(5), Tm(1)-N(3) 2.382(5), Tm(1)-N(4) 2.441(5), Tm(1)-N(6) 2.409(5), Tm(1)-O(1) 2.366(4), Tm(1)-O(2) 2.367(4) Tm(1)-I(1) 2.9378(5), N(1)-N(2)-N(3) 111.7(5), N(4)-N(5)- N(6) 110.9(5).

By analogy with the work of Evans (See Scheme 21 in section 1.3.5),137 it was proposed that the reaction of the putative thulium(II) bis(triazenide) with dinitrogen would potentially yield the dimeric reduced dinitrogen dimer (Scheme 80). It was hoped that this would prove easier to isolate and characterise and thus allow confirmation that a thulium(II) bis(triazenide) complex was actually being formed. A solution of thulium(II) iodide tetrahydrofuran complex was treated with a tetrahydrofuran solution of two equivalents of [NaN3Dipp2] 75 also in tetrahydrofuran at -80 °C under an argon atmosphere. The brown solution was then frozen, the flask evacuated and dinitrogen introduced and allowed to warm to room temperature under a dinitrogen atmosphere. A slow colour change was observed at approximately 0 °C to pale yellow. Volatiles were removed in vacuo and the product extracted into hexane (Scheme 80). Unfortunately attempts to recrystallise this product from nhexane repeatedly gave material unsuitable for analysis by single crystal diffraction analysis. The inherent paramagnetism of 88,89,137 1 15 thulium(III) (μeff (298K) 6.9-7.7 μB ) prevented the use of H or N spectroscopy.63 Infrared spectroscopy of the reaction product gave a strong absorbance at 1260 cm-1 consistent with the incorporation of a chelating triazenide ligand. Absorbances consistent with coordinated tetrahydrofuran were not observed in the 133 infrared spectrum. The suggested formulation of this compound as depicted in Scheme 80 unfortunately cannot be confirmed.

Scheme 80 The proposed synthesis of a dinitrogen bridged thulium(III) bis(triazenide) dimer.

3.2.4 Attempts to Synthesise a Tetravalent Cerium Triazenide

The chemical oxidation of neutral cerium(III) complexes has proved to be a viable entry point into the synthesis of cerium(IV) compounds (See Scheme 24 in section 115,148,152,153,157 1.3.6). Bearing this in mind, a solution of [Ce(N3Dipp2)3] 164 in toluene was treated with one half equivalent of hexachloroethane (Scheme 81). This reagent has been previously demonstrated to be a synthetically useful oxidant for accessing 147,155 cerium(IV) compounds under mild conditions. Unfortunately, [Ce(N3Dipp2)3] 164 proved unreactive with no discernible change of colour observed, even in the presence of excess hexachloroethane. The colour change is a useful indicator of oxidation, as all reported cerium(IV) amides have noticeable colour changes, which give deep, highly coloured solutions. For example the following colour changes have been reported for cerium compounds on oxidation: yellow to deep blue ([Ce(NCy2)3(thf)] to 148 [CeN(Cy2)4]), “pale yellow to deep red” ([Li(thf)Ce(N(SiHMe2)2)4] to 147 i [Ce(N(SiHMe2)2)4]) and “orange to dark blue, almost black” ([Li(thf)Ce(N Pr2)4] to i 370 [Ce(N Pr2)4]). Addition of one half equivalent of 1,2-dibromoethane to a solution of

[Ce(N3Dipp2)3] 164 in toluene provided similar results (Scheme 81). Silver tetraphenylborate has also proved useful as a chemical oxidant for the synthesis of cerium(IV) compounds.113,371 A suspension of one equivalent of silver tetraphenylborate and [Ce(N3Dipp2)3] 164 in toluene (Scheme 81) also showed no evidence of a reaction at room temperature.

134

Scheme 81 The attempted reaction of [Ce(N3Dipp2)3] 164 with chemical oxidants. It is surprising that none of the chemical oxidants proved reactive towards

[Ce(N3Dipp2)3] 164 in view of its high reactivity towards air and moisture (See Section 3.2.2 in this chapter). Attempts to fractionally recrystallise partially oxidised material from nhexane or toluene proved unsuccessful due to the oxidation product having similar solubility. Although predictions regarding the mechanism of oxidation are difficult due to the paucity of data, it is reasonable to suppose that the extreme steric congestion of the cerium in complex 164 (See Table 4 in Section 3.2.3) precludes bond formation to the oxidant. As such the proposed cerium oxidation must proceed at a less hindered cerium(III) precursor. When considered in concert with the steric calculations performed earlier, it is plausible that [Ce(N3Dipp2)3] 164 is insensitive to chemical oxidation due to the inaccessibility of the metal centre. To test this hypothesis further steric calculations were performed to compare the accessibility of the cerium in tris(triazenide) 164 with cerium(III) amides that have been successfully oxidised to cerium(IV) amides (Figure 86).115,148,152,153,157 In addition some comparisons were also made with carbocyclic cerium complexes that are known precursors to published cerium(IV) metallocenes (Figure 86).115 The results of these calculations are listed in

Table 8 and Table 9 with the values for [Ce(N3Dipp2)3] 164 included for ease of comparison.

Figure 86 Published precursors to cerium(IV) compounds.

As can be seen in Table 8 [Ce(N3Dipp2)3] 164 is the most sterically congested metal centre by a reasonable margin relative to the other listed amidocerium(IV) precursors.

The closest is [Ce{N(SiMe3)2}3] 186 with a calculated GCe(Complex) value of 85.95%.

135

It is worth noting that oxidations of [Ce{N(SiMe3)2}3] 186 are somewhat more challenging than the other precursors listed in Table 8, requiring more esoteric oxidants 153 like TeCl4 and PPh3Br2 and are typically poorly yielding. [Li(thf)Ce(NCy2)4] 187, which has a similar calculated GCe(Complex) value of 84.86% to [Ce{N(SiMe3)2}3] 186 (85.95%), benefits from the higher flexibility of the dicyclohexylamide ligand when compared with the bis(diisopropylphenyl)triazenide and bis(trimethylsilylamide) ligands. This likely allows access of the oxidant to the metal centre. The presence of a lithium cation within the coordination sphere, as per [Li(thf)Ce{N(SiHMe2)2}4] 188 and

[Li(thf)Ce(NCy2)4] 187, also likely expedites the electron transfer mechanism by ensuring the co-ligands are in place about the metal centre and maintaining an exchangeable position for the oxidant to utilise for electron transfer. This mechanism is not available to [Ce(N3Dipp2)3] 164 (vide supra) based upon steric factors.

[Ce(N(SiMe3)2)3] [Li(thf)Ce(NCy2)4] 21.27 28.64 20.94 G (L)/% 28.65 Ce 22.50 28.65 20.24 G Ce 85.95 84.86 (Complex)/%

[Li(thf)Ce(N(SiHMe2)2)4] [Ce(N3Dipp2)3] 30.76 GCe(L)/% 20.12 31.32 30.79 G Ce 19.76 92.58 (Complex)/%

Table 8 G parameters for amidocerium(IV) precursors 186-188 and [Ce(N3Dipp2)3] 164. In contrast to the amidocerium(IV) precursors, a more complicated argument is presented when carbocyclic cerium(IV) precursors are considered. When considered as a monomer (there is a low hapticity contact to an adjacent molecule in the molecular structure, See Section 1.3.1 in Chapter I), tris(cyclopentadienyl)cerium(III) 189 has a

GCe(Complex) value of 65.66% (Table 9). This complex, while exceedingly sensitive to oxygen162, is also very difficult to oxidise cleanly and in high yield.115 In stark contrast, the cerium metal centres in the cycooctatetraenyl complexes 190 and 191 on steric grounds are almost as inaccessible as in [Ce(N3Dipp2)3] 164. It is plausible however, in these cases that the exceedingly complicated electronic configurations of the cerocenes111,112 and haptotropy have a greater effect on the ability of these molecules to 136 be oxidised. Under these circumstances outer sphere oxidation is reasonable. Thus, while the triazenide ligand displays considerable coordinative versatility, the sheer degree of cerium coverage and easily reduced ligand moiety likely frustrates access of an oxidant.

t [CeCp3] [Li(dme)3(Ce(COT″)2] [Li(dme)3(Ce(COT(Si BuMe2)2)2] 22.41 43.54 43.78 G (L)/% 21.97 Ce 44.32 43.88 21.28 G Ce 65.66 87.86 87.63 (Complex)/% Table 9 G parameters for cerocene(IV) precursors 189-191.

Given the failure to synthesise a cerium(IV) triazenide by chemical oxidation of

[Ce(N3Dipp2)3] 164, an alternative synthesis of a ceric triazenide was sought. The decrease in ionic radius of cerium(IV) relative to cerium(III) (six coordinate ionic radii Ce3+ 101 pm, Ce4+ 87 pm)2 suggests that oxidation of a highly sterically congested, homoleptic precursor is ambitious. Thus a precursor to a heteroleptic complex was sought. Buoyed by the success of the protolysis method (relative to metathesis and redox transmetallation) for the preparation of the trivalent triazenides (vide supra) a related pathway was conceived for the synthesis of a tetravalent cerium triazenide complex from an alkoxy precursor (Scheme 82).

Tetrakis(tert-butoxy)cerium bis(tetrahydrofuran)156 was initially considered to be an ideal precursor given the general utility of the tert-butoxide bases in the synthesis of the alkali metal triazenides (See Section 2.2.3). However stirring a solution of t [Ce(O Bu)4(thf)2] and two equivalents of Dipp2N3H 40 in hexane at room temperature produced a material, which is consistent with a mixture of starting materials and the t 372 1 known decomposition product [Ce3O(O Bu)10] by H NMR spectroscopy in C6D6 (Scheme 82).

Scheme 82 The attempted synthesis of a mixed triazenide alkoxide cerium(IV) compound.

t The failure of [Ce(O Bu)4(thf)2] to undergo protolysis with Dipp2N3H 40 prompted the pursuit of an alternative protolysis precursor with more basic amide ligands.373 Initial 137 attempts focused on a homoleptic bis(trimethylsilyl)amide cerium(IV) compound of the formulation [Ce{N((SiMe3)2}4] 192. This was considered a viable target for two reasons: firstly the recent isolation of [U{N((SiMe3)2}4] 193 by chemical oxidation of 374 [K(thf)6][U{N((SiMe3)2}4] 194 with copper(I) iodide (Scheme 83). The similar ionic radii of uranium and cerium in both oxidation states (Six coordinate ionic radii Ce3+ 101 pm, U3+ 102.5 pm and Ce4+ 87 pm, U4+89 pm)2 suggested that on steric grounds the cerium analogue [Ce{N((SiMe3)2}4] 192 was a viable synthetic target. It was hoped that despite the more favourable electrode potential for uranium (U3+|U4+ 0.52V, Ce3+|Ce4+ - 1.70V. cf. Table 1)3 the chemical oxidation would proceed as desired. The structure of the cerium compound [Na(thf)4(OEt2)][Ce{N((SiMe3)2}4] 195 has also been previously reported. Unfortunately this was synthesised as part of a mixture and does not provide a 375 reasonable route into [Ce{N((SiMe3)2}4] 192.

374 Scheme 83 The synthesis of [U{N((SiMe3)2}4] 193. The addition of four equivalents of lithium bis(trimethylsilylamide) to cerium(III) triflate in tetrahydrofuran, followed by treatment with half an equivalent of hexachloroethane gave a bright red solution. Removal of volatiles and recrystallisation n of the product from pentane gave [(thf)3Li(μ–Cl)Ce{N(SiMe3)2}3] 196 in low isolated yield instead of the desired product (Scheme 84).

Scheme 84 The synthesis of [(thf)3Li(μ–Cl)Ce{N(SiMe3)2}3] 196.

The 1H NMR spectrum of amide 196 displays three broad resonances on account of the paramagnetism of the cerium(III) nucleus. The trimethylsilyl groups correspond to the resonance at δ -3.14, this represents a slight upfield shift from the reported value for 118 [Ce(N(SiMe3)2)3] 186 (δ -3.10). The tetrahydrofuran methylene protons correspond to the resonances observed at δ 0.69 and 1.84. The infrared spectrum of amide 196 displays several strong absorbances for example at 834, 1001 and 1245 cm-1 that are

138 consistent with silylamide and coordinated tetrahydrofuran absorbances in compounds of this type, for example [(thf)3Li(μ–Cl)Nd{N(SiMe3)2}3] (828, 1017 and 1243 cm-1).333,334,337

n [(thf)3Li(μ–Cl)Ce(N(SiMe3)2)3] 196 crystallises from a saturated solution in pentane as large colourless plates in the monoclinic space group P21/c. Two molecules are observed in the asymmetric unit, which necessitated substantial restraints for satisfactory thermal parameters. Unfortunately the overall data quality is poor, which presents as a high degree of disorder in the tetrahydrofuran ligands, especially of one of the molecules in the asymmetric unit. Discussion of the structural parameters is therefore limited to the other molecule of the asymmetric unit, which displays more limited disorder. An illustration of the structure may be seen in Figure 87. Both metal centres in the solid state adopt pseudo tetrahedral geometry, with distortion away from ideal observed at the cerium metal centre due to steric buttressing between the bis(trimethylsilylamide) ligands. The μ-chloro ligand bridges between the two metal centres with the Li(1A)-Cl(1A)-Ce(1A) bond vector deviating slightly away from linear (172.7(4)°). Two interactions between trimethylsilyl groups and the cerium metal centre are observed in the solid state (Ce(1A)-Si(5A) 3.486(2) Å, Ce(1A)-Si(2A) 3.496(2) Å) These distances are longer than those observed in the solid state structure for 376 [Ce(N(SiMe3)2)3] 186 (3.322 Å).

139

Figure 87 Molecular structure of [(thf)3Li(μ–Cl)Ce{N(SiMe3)2}3] 196, POV-RAY illustration, 40% thermal ellipsoids, all hydrogen atoms omitted for clarity. The tetrahydrofuran carbon atoms are depicted as a wireframe for clarity. One half of the asymmetric unit is omitted for clarity. Selected bond lengths (Å) and angles (°): Ce(1A)- N(1A) 2.363(5), Ce(1A)-N(3A) 2.377(5), Ce(1A)-N(2A) 2.379(5), Ce(1A)-Cl(1A) 2.743(2), Ce(1A)-Si(5A) 3.486(2), Ce(1A)-Si(2A) 3.496(2), Cl(1A)-Li(1A) 2.296(15), O(1A)-Li(1A) 1.928(14), O(2A)-Li(1A) 1.926(17), O(3A)-Li(1A) 1.91(2), Li(1A)-Cl(1A)-Ce(1A) 172.7(4), N(1A)-Ce(1A)-Cl(1A) 101.04(14), N(2A)-Ce(1A)-Cl(1A) 99.23(16), N(3A)-Ce(1A)-Cl(1A) 99.02(16), N(1A)-Ce(1A)-N(2A) 117.46(18), N(2A)-Ce(1A)-N(3A) 118.97(19), N(1A)-Ce(1A)-N(3A) 115.1(2).

Lappert’s tetrakis(dicyclohexylamido)cerium(IV) 197 and Anwander’s tetrakis(bis(dimethylsilyl)amido)cerium(IV) were considered as alternative protolysis precursors. The former was selected as the ideal candidate due to the relatively high basicity of the dicyclohexylamido ligands in comparison to the silylamido ligands 377 343 378 (HNCy2 pKa = 35.7 in THF vs 25.8 for HN(SiMe3)2 and 22.8 for HN(SiHMe2)2 also in THF). It was proposed this would provide the maximum thermodynamic driving force for the desired protolysis.

The published preparative method for this complex is low yielding,148 likely due to the choice of molecular oxygen as the oxidant. This presumably leads to amidocerium oxo

140 by-products.379 To improve on this preparation oxygen was substituted for a stoichiometric oxidant, such as those used previously (See Section 1.3.6 in Chapter I). According to the published procedure,148 a solution of four equivalents of lithium dicyclohexylamide was added at 0 °C to a suspension of [CeCl3(thf)n] followed by stirring at room temperature. After removal of volatiles in vacuo and extraction into toluene, one half of an equivalent of hexachloroethane was added, instantaneously 148 giving the midnight blue colour of [Ce(NCy2)4] 197 (Scheme 85). Unfortunately the stability of the complex proved to be significantly lower than reported, with decomposition, as signified by a fading of the blue colour to the pale yellow of cerium(III) frequently encountered during handling or storage of solutions of

[Ce(NCy2)4] 197. To minimise these problems the cerium(IV) amide 197 was generated in situ at lower temperatures. This also guarded against ether cleavage reactions, as has i 377 been reported for lithium diisopropylamide (HN Pr2 and HNCy2 pKa = 35.7 in THF).

A solution of four equivalents of lithium dicyclohexylamide in tetrahydrofuran was added to a suspension a [CeCl3(thf)n] in tetrahydrofuran at -40°C and slowly warmed to 0 °C over 4 hours. Removal of volatiles in vacuo below 0 °C followed by extraction into toluene and immediate oxidation with half an equivalent of hexachloroethane afforded a midnight blue solution (Scheme 85). Addition of a solution of one or two equivalents of

Dipp2N3H 40 in toluene invariably lead to a slow fading of the colour of the solution to a pale yellow indicating that reduction had occurred. Removal of solvent gave a bright 1 yellow oil and some colourless blocks. The latter were identified as Dipp2N3H 40 by H

NMR spectroscopy (C6D6) and a single crystal X-ray diffraction unit cell determination. The yellow oil was determined to be paramagnetic due to significant broadening of a 1H

NMR spectrum collected in C6D6. This suggests the formation of a cerium(III) containing product. The yellow oil proved to be intractable in that extended treatment with vacuum did not cause it to solidify and thus this route was abandoned.

Scheme 85 The attempted reaction of HN3Dipp2 with [Ce(NCy2)4] 197.

An examination of the bond lengths published for [Ce(NCy2)4] 197 and 148 [Li(thf)Ce(NCy2)4] 187 reveals a dramatic shortening of the cerium to nitrogen bonds upon oxidation. Concluding that the substitution of a dicyclohexylamide ligand by a 141 more sterically demanding triazenide ligand induces an unstable cerium(IV) compound, attempts were made to achieve protolysis pre-oxidation. At this stage it was also considered that a bis(triazenide) would be too sterically congested about the smaller cerium(IV) metal centre. Thus efforts were focused on the synthesis of a mono(triazenide) cerium(III) complex derived from [Li(thf)Ce(NCy2)4] 187. The cerate complex was sought due to the anticipated ease of oxidation relative to the neutral complexes (vide supra). The reaction of Dipp2N3H 40 and [Li(thf)Ce(NCy2)4] 187 followed by extraction into and recrystallisation from nhexane did not give a cerium compound. Instead a dimeric lithium dicyclohexylamine adduct of the triazenide

[Li(N3Dipp2)(μ-N3Dipp2)(μ-Li)(HNCy2)] 198 was isolated in low yield (Scheme 86). It is believed in this instance that the low solubility of the cerium compound 1 [Ce(NCy2)3(thf)] 199 in alkanes causes its elimination by precipitation. The H NMR spectrum of dinuclear lithium triazenide 198 in C6D6 is consistent with the formulation. Curiously the aromatic protons of complex 198 are observed only as a broad singlet centred at δ 7.16, which overlaps with the solvent residual resonance of the deuterated solvent. The infrared spectrum of dinuclear 198 is consistent with the presence of the protonated amine with absorbances observed at 3060 and 3137 cm-1. The broad absorbance that is observed at 1257 cm-1 is consistent with a triazenide that is 201 chelating. In the solid state crystalline [Li(N3Dipp2)(μ-N3Dipp2)(μ-Li)(HNCy2)] 198 is observed to be thermochromic with a colour change from pale yellow to bright red observed at 85 °C. This is plausibly a result of conformational change at the triazenide ligands in the solid state.

n Scheme 86 The reactions of [Li(thf)Ce(NCy2)4] 187 with Dipp2N3H 40 in hexane and toluene.

[Li(N3Dipp2)(μ-N3Dipp2)(μ-Li)(HNCy2)] 198 crystallises from a saturated solution in nhexane as pale yellow plates in the triclinic space group P1¯ with one full molecule in the asymmetric unit. The triazenide ligands in the dinuclear unit adopt two very different geometries in the solid state; one in the common monometallic chelating

142 geometry and the second adopting an unusual κ2-N1,N2-κ1-N3 geometry, which acts as a bridging ligand between Li(1) and Li(2). As a result the NNN angle (N(6)-N(5)-N(4) 119.6(3)°) widens considerably relative to that observed at the monometallic chelating triazenide (N(3)-N(2)-N(1) 109.8(3)°) The former geometry is also distinguished by an η2 π-arene interaction to Li(2). The distance of the contacts Li(2)-C(26) (2.633(7) Å) is considerably longer than Li(2)-C(25) (2.456(7) Å), however it is still within the range expected for a lithium to arene-π interaction. This coordination mode is similar to that observed for [{K(thf)2K(N3Dipp2)}n] 78 (See Section 2.2.3 in Chapter II). In that instance a higher nuclearity, polymeric η6 arene-π to K interaction is observed, which is consistent with the larger ionic radius of the potassium cation relative to the lithium in this case.2 The metal to nitrogen bond lengths are all very similar, with the exception of Li(2)-N(6), which is much shorter (1.973(7) Å). This has plausibly been shortened as a consequence of the metal arene-π interactions to Li(2). The nitrogen-nitrogen bond lengths are paired into short and long bonds for each triazenide, however each bond is within the same range (1.284(4)-1.331(3) Å) as that observed in Dipp2N3H 40 in the solid state, where a degree of bond delocalisation was concluded.

Figure 88 Molecular structure of [Li(N3Dipp2)(μ-N3Dipp2)(μ-Li)(HNCy2)] 198, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. The cyclohexyl and isopropyl carbon atoms are depicted as wireframes for clarity. Selected bond lengths (Å) and angles (°): N(1)-N(2) 1.331(4), N(2)-N(3) 1.298(3), N(1)-Li(1) 2.005(7), Li(1)-N(3) 2.058(7), Li(1)-N(4) 1.992(7), Li(1)-N(5) 2.006(7), N(4)-N(5) 1.310(4), N(5)-N(6) 1.284(4), Li(2)-N(6) 1.973(7), Li(2)-N(7) 2.010(7), Li(2)-C(25) 2.456(7), Li(2)-C(26) 2.633(7), N(3)-N(2)-N(1) 109.8(3), N(6)-N(5)-N(4) 119.6(3).

Under similar reaction conditions to those that gave [Li(N3Dipp2)(μ-N3Dipp2)(μ-

Li)(HNCy2)] 198 but with Ce(OTf)3 as the initial cerium source and toluene as the 143 solvent did not give the desired cerate complex. Instead the neutral bis(dicyclohexylamide) tetrahydrofuran complex [Ce(NCy2)2(N3Dipp2)(thf)] 200 was isolated in moderate yield (Scheme 86). Unfortunately this preparation is often accompanied by a sizeable quantity of [Ce(NCy2)4] 197, which is formed at an intermediate stage of the reaction and is difficult to remove by washing or recrystallisation. It is plausible that that the loss of the extra equivalent of lithium dicyclohexylamide is the result of triazenide-dicyclohexylamide steric buttressing, which forces the lithium amide out of the coordination sphere with complete elimination occurring due to the low solubility of the lithium amide in toluene. The cause of the different reaction outcome due to different cerium source is at this stage unknown

[Ce(NCy2)2(N3Dipp2)(thf)] 200 crystallises as yellow-orange rods from a saturated solution in npentane in the triclinic space group P1¯. One full molecule and one half molecule of npentane are observed in the asymmetric unit. An illustration of

[Ce(NCy2)2(N3Dipp2)(thf)] 200, may be seen in Figure 89. Considering the triazenide ligand as a single point donor at the N2 position, a pseudo tetrahedral geometry is adopted in the solid state with all of the associated angles sizeably different outside of error from the ideal tetrahedral value of 109.5°. The N(1)-N(2)-N(3) triazenide bonding unit 112.9(3)° is the same within error as the N(1)-N(2)-N(3) (113.2(2)°) and N(4)-

N(5)-N(4)# (112.7(4) °) angles of [Ce(N3Dipp2)3] 164. The bond lengths in the mono(triazenide) 200 are, however, significantly longer outside of error when compared to [Ce(N3Dipp2)3] 164, presumably due to the increased steric congestion in the present instance. For example in [Ce(NCy2)2(N3Dipp2)(thf)] 200, these bond lengths are measured as Ce(1)-N(1) 2.582(3) and Ce(1)-N(3) 2.633(3) Å, while those in tris(triazenide) 164, range from 2.498(2) to 2.518(2) Å despite the smaller formal coordination number of the heteroleptic compound. The published molecular structure 148 of [Ce(NCy2)3(thf)] 199 compares favourably to [Ce(NCy2)2(N3Dipp2)(thf)] 200, wherein the tetrahydrofuran to cerium bond length is similar but shorter in

[Ce(NCy2)2(N3Dipp2)(thf)] 200 (2.582(2) vs. 2.531(3) Å respectively). A similar trend of shorter bonds is observed for the dicyclohexylamide ligands in

[Ce(NCy2)2(N3Dipp2)(thf)] 200. In [Ce(NCy2)3(thf)] 199 these distances range from

2.318(2) to 2.336(2) Å, vs. 2.271(3) and 2.291(3) Å in [Ce(NCy2)2(N3Dipp2)(thf)] 200.

144

Figure 89 Molecular structure of [Ce(N3Dipp2)(NCy2)2(thf)] 200, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms and lattice pentane molecule omitted for clarity. The tetrahydrofuran carbon atoms are depicted as a wireframe for clarity. Selected bond lengths (Å) and angles (°): Ce(1)-N(1) 2.582(3), Ce(1)-N(3) 2.633(3), Ce(1)- N(4) 2.271(3), Ce(1)-N(5) 2.291(3), Ce(1)-O(1) 2.531(3) N(1)-N(2)-N(3) 112.9(3), N(2)-Ce(1)-O(1) 105.40(9), N(2)- Ce(1)-N(4) 110.28(11), N(2)-Ce(1)-N(5) 124.01(10), O(1)-Ce(1)-N(4) 100.39(11), O(1)-Ce(1)-N(5) 91.36(11), N(4)- Ce(1)-N(5) 118.67(11).

3.3 Conclusions and Future Work 3.3.1 Conclusions

The chapter had four aims: i) to evaluate the metathesis, redox transmetallation and protolysis protocols for utility in the synthesis of trivalent lanthanide triazenides; ii) investigate the possibility of SIR at and alkali metal reduction of the sterically congested trivalent lanthanides synthesised earlier; iii) the synthesis of divalent lanthanide triazenides to investigate the ligand’s utility for the more reducing metals; iv) to investigate the accessibility of a ceric triazenide.

Metathesis protocols using the base-free triazenides synthesised in Chapter II were investigated. Lithium halide inclusion into the alkali metal triazenides was observed, yielding products such as [Li2(μ-I)(μ-thf)(μ-N3Dipp2)(thf)2] 198. Variable reaction outcomes have also been demonstrated, particularly with regard to solvent induced ligand redistribution mechanisms, which give homoleptic triazenides such as

[Nd(N3Dipp2)3] 156 in preference to heteroleptic complexes. The redox transmetallation

145 protocol has also been evaluated using thallium triazenide 152. While successful for the synthesis of homoleptic triazenides such as [Sm(N3Dipp2)3] 160, this method has been demonstrated to be highly inefficient, requiring large excesses of the lanthanide metal for complete conversion, a further disadvantage when a divalent oxidation state is readily accessible (Sm, Eu, Yb). The protolysis protocol between [Ln{N(SiMe3)2}3] (Ln = Ce, Nd, Sm and Yb) complexes has been evaluated and determined to be the most efficient method for the larger lanthanides of synthesising homoleptic trivalent triazenides, [Ln(N3Dipp2)3] 156, 160 and 164. The cause of the low isolable yield of

[Yb(N3Dipp2)3] 161 remains unknown.

The homoleptic trivalent triazenide [Nd(N3Dipp2)3] 156 was treated with excess potassium graphite with a view to synthesising a low-oxidation state neodymium triazenide. In this instance no isolable products were observed. The possibility of SIR was investigated between [Sm(N3Dipp2)3] 160 and O=PPh3. This was also not successful. In order to understand these outcomes, steric calculations were performed. By the measures outlined in this chapter the tris(triazenide) complexes have been demonstrated to be sterically more congested than the tris(pentamethylcyclopentadienyl)- and tris(1,3-bis(trimethylsilyl)cyclopentadienyl) lanthanide complexes. In light of these results an electronic rather than steric effect is implicated in the lack of SIR and reductive chemistry observed at these complexes.

The synthesis of classical divalent triazenides (Sm and Yb) has been achieved. Attempts to extend this to a divalent thulium compound were unsuccessful in all cases. It has been implied, based upon this result and the low isolated yield of the samarium complex that the reduced charge density at the N-donor atoms of a triazenide relative to amidinate and guanidinate ligands may make triazenides unsuitable ligands for the more reducing lanthanide oxidation states. The sterics of the triazenide ligand, while reduced compared with the formamidinate analogue are still comparable and as such this is unlikely to be a major effect.

Investigations into the synthesis of a tetravalent cerium triazenide have been carried out.

[Ce(N3Dipp2)3] 164 proved unreactive with common oxidants. Attempts to synthesise

[Ce{N(SiMe3)2}4] 192 for use as a protolysis precursor by oxidation of a trivalent cerium precursor instead yielded [(thf)3Li(μ–Cl)Ce{N(SiMe3)2}3] 196. The alternative protolysis precursor [Ce(NCy2)4] 197 did not give tractable cerium(IV) containing 146 products when treated with Dipp2N3H 40. Attempts to synthesise a heteroleptic cerium(III) triazenide, with a view to later oxidation yielded [Ce(NCy2)2(N3Dipp2)(thf)] 200. This compound was obtained only as part of a mixture and the plausibility of oxidation to cerium(IV) remains to be determined.

3.3.2 Future Work

This work may be built upon by extending to the synthesis of lanthanide triazenides derived from the more sterically demanding ligand Dipp*2N3H 39. It is anticipated that this ligand will be more suited to the more reducing divalent lanthanides than the

N3Dipp2 ligand as examined in this chapter. This assumption is based upon the observed high number of arene-π interactions in the potassium complex [KN3Dipp*2] 81. It is believed that this will sterically saturate the reducing metal centres and in turn stabilise them in this coordination environment.

147

3.4 Appendix III: Supplementary Crystal Structures

The lanthanide triiodide tetrahydrofuran solvate compounds used in this chapter were synthesised according to the method of Izod.328 These compounds were repeatedly found to be contaminated with material of a different colour to the reported triiodides. This material could be removed by careful fractional recrystallisation from tetrahydrofuran, which also enabled the structural characterisation of the impurities as unusual lanthanide diiodide triiodides [LnI2(I3)(thf)5] (Ln = Nd, Sm) 201 and 202. These compounds tend to have higher solubility in coordinating solvents than the triiodides, which facilitates their removal by washing. These are likely formed in the initial stages of synthesis, by the attack of an iodide ion on a molecule of iodine in solution (Scheme 87). The formation of these impurities was reproducibly formed in multiple batches of the desired lanthanide triiodides.

Scheme 87 The proposed reaction of lanthanide triiodides with excess iodine.

Compounds of this type have been isolated previously under similar conditions. Three are structurally characterised ([MI2(I3)(thf)5] Ln = La 203, Yb 204 and [PuI2(I3)(thf)4py] 205)380-382 and the formula of an unpublished, poor quality crystal structure has been 382 disclosed ([PuI2(I3)(thf)5] 206) (Figure 90).

Figure 90 Analogous diiodo triiodide compounds for comparison to complexes 201 and 202.380-382

Complexes 201 and 202 crystallise from concentrated solutions in tetrahydrofuran at - 25 °C as yellow-deep orange truncated square plates for neodymium 201 and deep orange prisms for samarium 202. Both crystallise in the orthorhombic space group Pbcn with one half of the ion pair in the asymmetric unit. This is isomorphous with the reported analogues.380-382 An illustration of the molecular structures may be seen in Figure 91 and the structural parameters for all relevant compounds are listed in Table 10.

148

Figure 91 Molecular structure of [LnI2(I3)(thf)5] 201 and 202, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. The tetrahydrofuran carbon atoms are depicted as wireframes for clarity. Selected bond lengths (Å) and angles (°) may be seen in Table 10. Symmetry operations used to generate equivalent atoms for 1 3 Nd: -x, y, /2-z and 1-x, 1-y, 1-z. Symmetry operations used to generate equivalent atoms for Sm: 2-x, y, /2-z and 1-x, 1-y, 1-z.

In the solid state both compounds adopt near-ideal pentagonal bipyramidal molecular geometries (e.g. I(1)-M(1)-I(1)# 179.54(3)° and 179.500(10)°, O(1)-M(1)-O(1)# 71.4(2)° and 71.56(10)° and O(1)-M(1)-O(2) 71.82(17)° and 72.00(7)° for Nd and Sm respectively). This is also consistent with the geometry of reported analogues, even for 382 [PuI2(I3)(thf)4py] 205. The ligand bond lengths are also consistent with the expected values with decreasing metal ionic radius. For example, for the neodymium complex 201 the Nd-O range is 2.450(5)-2.476(7) Å and for the samarium complex 202 the Sm- O range is 2.4245(18)-2.450(2) Å. By comparison the larger literature lanthanum compound 203 has a range of 2.511(5)-2.532(4) Å380 and the smaller ytterbium complex 204 2.341(4)-2.351(4) Å.381 This trend is also consistent across the range of metal to iodide bond lengths, which range from 3.130(1) Å, through 3.0613(6) and 3.0310(3) Å for complexes 203, 201 and 202 respectively to 2.9505(5) Å for the smallest metal, ytterbium 204.

149

Nd Sm La Yb Pu M(1)-I(1)/Å 3.0613(6) 3.0310(3) 3.130(1) 2.9505(5) 3.0721(5) M(1)-O(1)/Å 2.450(5) 2.4323(18) 2.530(4) 2.349(6) 2.462(4) M(1)-O(2)/Å 2.460(4) 2.4245(18) 2.511(5) 2.341(4) 2.486(4) M(1)-O(3)/Å or 2.476(7) 2.450(2) 2.532(4) 2.351(4) 2.596(7) Pu(1)-N(1) I(2)-I(3)/Å 2.9116(5) 2.9037(2) 2.920(1) 2.9058(5) 2.9113(6) I(1)-M(1)-I(1)#/° 179.54(3) 179.500(10) 179.57(2) 179.57(2) 179.79(2) O(1)-M(1)-O(2)/° 71.82(17) 72.00(7) 72.19(8) 72.10(10) 71.50(15) O(2)-M(1)-O(3)/° or 72.48(12) 72.23(4) 72.27(13) 71.99(14) 72.55(10) O(1)-Pu(1)-N(1) O(1)-M(1)-O(1)#/° 71.4(2) 71.56(10) 72.27(13) 71.9(2) 71.9(2)

Table 10 Structural parameters for diiodo triiodides.

150

3.5 Appendix IV : Crystal Data for Structures Collected for this Chapter

[Li (μ-I)(μ-thf)(μ-N Dipp )(thf) ] [Tl (μ-OH)(μ-N Dipp )] ·PhMe [Sm(N Dipp ) I(thf)] [Nd(N Dipp ) ]·1.5PhMe·0.5THF 2 3 2 2 2 3 2 2 3 2 2 3 2 3 Chemical Moiety Formula C36H58N3ILi2O3 C55H78N6O2Tl4 C52H76N6IOSm C86.50H122N9NdO

Chemical Sum Formula C36H58N3ILi2O3 C31H43N3OTl2 C52H76N6IOSm C86.50H122N9NdO Mol. Weight/ g mol-1 721.63 882.42 1078.43 1448.17 Temperature/ K 150(2) 150(2) 150(2) 150(2) Crystal habit, colour Plates, pale yellow Rod, yellow Plate, yellow Rod, yellow Crystal Size/ mm 0.25 x 0.15 x 0.03 0.2 x 0.03 x 0.03 0.1 x 0.1 x 0.01 0.05 x 0.15 x 0.01 Crystal System monoclinic triclinic triclinic monoclinic

Space Group P21/n P1¯ P1¯ P21/n a/ Å 9.3947(4) 10.2429(6) 11.747(3) 12.8870(8) b/ Å 18.4655(6) 12.2347(7) 14.612(3) 38.166(2) c/ Å 22.4774(9) 14.2028(8) 16.143(3) 16.4622(10) α/ ° 90.00 69.576(3) 82.610(9) 90.00 β/ ° 93.1970(10) 78.985(3) 77.276(10) 96.437(3) γ/ ° 90.00 71.944(3) 79.437(10) 90.00 Volume/ Å3 3893.3(3) 1578.97(16) 2645.8(11) 8045.8(8) Z 4 2 2 4 Density calcd/ g cm-3 1.231 1.856 1.354 1.196 μ/ mm-1 0.857 10.219 1.733 0.695

F000 1512 840 1106 3088 Reflections Collected 7548 8240 12379 15758 Unique Reflections 6335 6522 7675 8115 Parameters Varied 464 340 566 898 R(int) 0.031 0.0394 0.0897

R1 0.0278 0.0299 0.0687 0.0902 wR1 (all data) 0.0878 0.0626 0.2106 0.2369 Largest diff. peak and hole/ e Å-3 0.385, -0.346 1.088, -1.280 3.676, -3.807 1.791, -1.184 GOOF 0.670 1.051 0.979 1.086

151

[{Ce(N3Dipp2)2(μ-OH)}2]· [Sm(N3Dipp2)3] [Yb(N3Dipp2)3] [Nd(N3Dipp2)3] [Ce(N3Dipp2)3] [{Ce(N3Dipp2)2}2(μ-O)]

Chemical Moiety Formula C72H102N9Sm C72H102N9Yb C72H102N9Nd C72H102N9Ce C96H138Ce2N12O2

Chemical Sum Formula C36H51N4.5Sm0.5 C36H51N4.5Yb0.5 C36H51N4.5Nd0.5 C36H51N4.5Ce0.5 C96H138Ce2N12O2 Mol. Weight/ g mol-1 1243.99 1266.66 1237.87 1233.75 1764.9 Temperature/ K 150(2) 150(2) 150(2) 150(2) 150(2) Crystal habit, colour Cube, yellow Cube, pink Cube, yellow-green (dichroic) Cube, yellow Block, orange-yellow Crystal Size/ mm 0.1 x 0.1 x 0.1 0.4 x 0.4 x 0.4 0.05 x 0.05 x 0.05 0.05 x 0.05 x 0.05 0.15 x 0.15 x 0.15 Crystal System trigonal trigonal trigonal trigonal monoclinic

Space Group R3¯c R3¯c R3¯c R3¯c P21/n a/ Å 23.3609(7) 23.292(5) 23.3850(13) 23.3957(7) 17.9707(6) b/ Å 23.3609(7) 23.292(5) 23.3850(13) 23.3957(7) 27.5611(8) c/ Å 64.235(4) 63.773(14) 64.399(6) 64.736(4) 20.4577(7) α/ ° 90.00 90.00 90.00 90.00 90.00 β/ ° 90.00 90.00 90.00 90.00 97.1980(10) γ/ ° 120.00 120.00 120.00 120.00 90.00 Volume/ Å3 30358(2) 29961(11) 30499(4) 30687(2) 10052.7(6) Z 36 36 36 36 4 Density calcd/ g cm-3 1.225 1.264 1.213 1.202 1.171 μ/ mm-1 0.917 1.452 0.813 0.714 0.943

F000 11862 12006 11826 11790 3720 Reflections Collected 6644 6843 6684 6714 16324 Unique Reflections 5752 3623 5395 5433 10056 Parameters Varied 383 383 383 383 1041 R(int) 0.0547 0.3773 0.1015 0.0903 0.0485 R1 0.0241 0.0630 0.0305 0.0290 0.0716 wR1 (all data) 0.0858 0.1731 0.121 0.1202 0.2694 Largest diff. peak and hole/ e Å-3 1.223, -0.854 1.212, -3.702 1.419, -1.481 0.929, -1.448 1.065, -1.380 GOOF 1.255 0.988 0.984 1.219 1.101

152

[Yb(N Dipp ) (thf) ] [Sm(N Dipp ) (thf) ] [Tm(N Dipp2) I(thf) ] [(thf) Li(μ–Cl)Ce{N(SiMe ) } ] 3 2 2 2 3 2 2 2 3 2 2 3 3 2 3 Chemical Moiety Formula C56H84N6O2Yb C56H84N6O2Sm C56H84IN6O2Tm C30H78CeClLiN3O3Si6

Chemical Sum Formula C104H168N12O4Yb2 C56H84N6O2Sm C56H84IN6O2Tm C60H154Ce2Cl2Li2N6O6Si12 Mol. Weight/ g mol-1 1046.33 1023.64 1169.12 880.00 Temperature/ K 150(2) 150(2) 150(2) 150(2) Crystal habit, colour Plate, carmine Plate, malachite Block, yellow Block, colourless Crystal Size/ mm 0.2 x 0.1 x 0.02 0.1 x 0.05 x 0.01 0.15 x 0.15 x 0.10 0.15 x 0.15 x 0.15 Crystal System monoclinic triclinic orthorhombic monoclinic

Space Group P21/c P1¯ Pbca P21/c a/ Å 26.0360(13) 11.8969(6) 11.5086(4) 26.259(2) b/ Å 20.6939(11) 12.7597(7) 26.1043(11) 18.0047(12) c/ Å 22.2263(10) 19.7226(8) 40.7651(17) 23.1579(18) α/ ° 90.00 85.951(2) 90.00 90.00 β/ ° 110.466(2) 87.791(2) 90.00 110.356(3) γ/ ° 90.00 64.619(2) 90.00 90.00 Volume/ Å3 11219.3(10) 2698.1(2) 12246.8(8) 10264.8(13) Z 8 2 8 8 Density calcd/ g cm-3 1.239 1.260 1.268 1.139 μ/ mm-1 1.71 1.133 1.993 1.106

F000 4384 1080 4800 3720 Reflections Collected 22035 14398 12010 18799 Unique Reflections 14562 10322 8739 7414 Parameters Varied 1203 639 611 977 R(int) 0.0814 0.0783 0.0516 0.152 R1 0.0476 0.0555 0.0553 0.0567 wR1 (all data) 0.1455 0.1189 0.1512 0.1958 Largest diff. peak and hole/ e Å-3 2.326, -1.543 2.749, -1.281 1.261, -1.377 1.113, -0.759 GOOF 1.066 1.038 1.086 0.94

153

[Li(N Dipp )(μ-N Dipp )(μ-Li)(HNCy )] [Ce(NCy ) (N Dipp) (thf)]·0.5C H [NdI (I )(thf) ] [SmI (I )(thf) ] 3 2 3 2 2 2 2 3 2 5 12 2 3 5 2 3 5 Chemical Moiety Formula C60H90Li2N7 C54.50H91.50CeN5O C20H40I5NdO5 C20H40I5SmO5

Chemical Sum Formula C60H90Li2N7 C54.50H91.50CeN5O C10H20I2.5Nd0.5O2.5 C10H20I2.5Sm0.5O2.5 Mol. Weight/ g mol-1 923.26 972.95 1139.26 1145.36 Temperature/ K 150(2) 152(2) 150(2) 150(2) Crystal habit, colour Plate, pale yellow Rod, yellow-orange Plate, orange Prism, orange Crystal Size/ mm 0.1 x 0.09 x 0.02 0.1 x 0.03 x 0.02 0.10 x 0.10 x 0.02 0.10 x 0.10 x 0.09 Crystal System triclinic triclinic orthorhombic orthorhombic Space Group P1¯ P1¯ Pbcn Pbcn a/ Å 13.1405(12) 10.6937(11) 12.5159(13) 12.4214(6) b/ Å 13.8757(11) 11.4919(9) 22.011(3) 21.9151(11) c/ Å 18.2444(17) 23.050(2) 11.9210(18) 11.9023(7) α/ ° 73.227(3) 80.835(4) 90.00 90.00 β/ ° 78.363(3) 84.890(4) 90.00 90.00 γ/ ° 62.277(3) 73.259(4) 90.00 90.00 Volume/ Å3 2810.5(4) 2675.1(4) 3284.0(8) 3240.0(3) Z 2 2 8 8 Density calcd/ g cm-3 1.091 1.208 2.304 2.348 μ/ mm-1 0.063 0.891 6.305 6.601

F000 1010 1039 2100 2108 Reflections Collected 14079 10355 4881 3194 Unique Reflections 4467 8544 2548 3049 Parameters Varied 678 569 153 153 R(int) 0.1083 0.0662 0.076 0.0262 R1 0.089 0.0422 0.0472 0.0173 wR1 (all data) 0.259 0.1189 0.1106 0.0402 Largest diff. peak and hole/ e Å-3 0.379, -0.305 0.602, -0.813 1.27, −1.72 0.612, -0.822 GOOF 0.956 1.057 0.969 1.123

154

Chapter IV Ytterbium and Calcium Arenes: Unusual

Rearrangements Leading to Solvent Separated Ion Pairs

4.1 Introduction 4.1.1 The Chemistry of Calcium Arenes

While the chemistry of arylcalcium organometallics has experienced rapid growth over the last decade it is under developed relative to arylmagnesium chemistry.383-386 Investigations to date have focused on the more fundamental coordination chemistry of [ArAeX] (Ae = Ca, Sr, Ba) compounds with only a single report of Grignard-like reactivity in the literature.387 Thus far some major differences between these so called “Heavy-Grignard reagents” and the more common magnesium aryls have emerged: the synthesis of arylcalcium halides requires iodoarenes as starting materials in order to achieve reasonable yields and reaction rates; large amounts of Wurtz-type coupling products are observed when using chloro- and bromoarenes instead of insertion into the carbon-halogen bond.388,389 This contrasts with the synthesis of aryl magnesium compounds where even chloroarenes may be used as starting materials.390 This selective reactivity allows for a degree of functional group tolerance and selectivity in the synthesis of arylcalcium halides. For example meta-halo-iodobenzenes react once to yield the meta-halophenylcalcium iodide (Scheme 88) rather than a mixture of [m- 391 XC6H4CaI] and [m-IC6H4CaX] (X = Cl, Br). This can be considered a distinct advantage compared to arylmagnesium reagents, in which both the 1- and 3- positions are of similar reactivity.392

Scheme 88 The synthesis of meta-substituted arylcalcium iodides.391

A third major point of difference is the propensity of arylcalcium halide species to react with ethers at temperatures above -30 °C. A plausible decomposition mechanism of simple arylcalcium iodides (Scheme 89) has been reported for phenylcalcium iodide and

155 illustrates some of the intermediates commonly observed in this chemistry and the frequently observed products of decomposition.388,389 As illustrated (Scheme 89), the calcium aryl is sufficiently reactive to deprotonate the α proton of tetrahydrofuran to yield a secondary calcium alkyl, which then decomposes by a [3+2] cycloreversion to generate ethylene and a calcium enolate. Subsequent reaction of this enolate with further equivalents of phenylcalcium iodide generates calcium carbide, calcium diiodide and “calcium oxide”. Finally the calcium oxide reacts with three further equivalents of phenylcalcium iodide to yield the tetracalcium cluster 207. Such ether cleavage reactions limit this chemistry to low temperatures and diminish their synthetic utility.384- 386

Scheme 89 The decomposition of phenylcalcium iodide in tetrahydrofuran solution at room temperature.388,389

Another mechanism of decomposition is observed in arylcalcium halides with ortho- alkyl substituents. This is best illustrated by the example of mesitylcalcium iodide 208 (Scheme 90).389,393 The first step of decomposition is likely the same as that displayed in Scheme 89: mesityl calcium reacts with tetrahydrofuran in an acid-base reaction to yield mesitylene and the tetrahydrofuran derived calcium alkyl. This is followed by either cycloreversion (vide supra) or re-metallation of the mesitylene at a benzylic position. Under the reported conditions these two pathways occur in tandem at approximately a 5:1 ratio respectively.393

Scheme 90 The decomposition of mesitylcalcium iodide 208 to 3,5-dimethylbenzylcalcium iodide and binary calcium compounds.393 156

In spite of the above differences in solution phase stability, parallels between arylcalcium complexes and their lighter magnesium relatives may be found.383,385,386 For example Schlenk-equilibria are observed to exist. Manipulation of these can be used to synthesise diarylcalcium complexes394,395 in a similar fashion to magnesium relatives.396 In organomagnesium chemistry these equilibria are conventionally manipulated by precipitation of insoluble magnesium dihalide 1,4-dioxane complexes. This method does not translate to arylcalcium chemistry.394 Instead the addition of potassium tert- butoxide gives precipitates of insoluble potassium iodide and bis(tert-butoxy)calcium (Scheme 91) leaving diarylcalcium species in solution.394,395

Scheme 91 The synthesis of bis(aryl)calcium complexes.394,395

In some circumstances the diarylcalcium may be isolated by fractional crystallisation at extremely low temperatures. For example bis(mesityl)calcium may be separated from mesitylcalcium iodide through fractional recrystallisation of tetrahydrofuran solutions at -90 °C.397

Despite the aforementioned non-benign handling of aryl alkaline earths in ethereal solvents, thermally stable perfluoroaryl alkaline earth organometallics have been stabilised using extremely sterically demanding N-aryl triazenide ligands (Figure 92).204 The kinetic stability provided by these extremely demanding ligands allows for the isolation of these aryl alkaline earth complexes in tetrahydrofuran at room temperature. This contrasts with other arylcalcium complexes, which require sub-zero temperatures (vide supra). For example the calcium triazenide 209 was observed to decompose some 100 °C above phenylcalcium iodide despite coordination of tetrahydrofuran at calcium (Figure 92). The heavier strontium and barium analogues decomposed at the slightly higher temperatures of 95 °C and 144 °C respectively. It should be noted that these complexes 210 and 211 do not have tetrahydrofuran ligands directly coordinated to the metal (Figure 92). The authors attributed this high thermal stability to metal to N-aryl π- arene interactions that are observed in the solid state structures of triazenides 209-211. It is proposed that these interactions kinetically stabilise the arene complexes by steric shielding and electronic saturation of the alkaline earth metal centres. 157

Figure 92 Triazenide stabilised pentafluorophenylphenyl alkaline earth compounds.204

4.1.2 The Chemistry of Ytterbium Arenes

As has been previously discussed (See Section 3.2.3 in Chapter III), the chemistry of calcium and ytterbium(II) often have many parallels. It should come as no surprise then that the chemistry of aryllanthanide organometallics, in a similar fashion to arylcalcium chemistry, is underdeveloped relative to transition or main group metal analogues. This is in contrast to the related field of alkyllanthanide organometallics,160,398 which has been driven by the successful application of polysilylalkyl ligands.160,399 For the divalent lanthanides, bis(trimethylsilyl)methyl400-402 and tris(trimethylsilyl)methyl403,404 ligands have been used prodigiously but correspondingly simple homoleptic aryllanthides are rare. For example diphenyleuropium and diphenylsamarium are unknown and diphenylytterbium has been reported but is very poorly characterised by modern standards.405 A material reported as “dinaphthaleneytterbium” is known, however it has been suggested that this material is actually a π-coordinated ytterbium(0) complex rather than divalent ytterbium.406 Some unusual reaction chemistry has been demonstrated using this compound as a starting material. For instance, the reaction of dinaphthaleneytterbium with triphenylbismuth or diphenylmercury gave the unusual mixed valence complex pentaphenyldiytterbium 212 (Scheme 92).407 This was assigned as a mixture of one ytterbium(II) and one ytterbium(III) centre on the basis of magnetic moment measurements, which were consistent with a single diamagnetic ytterbium(II) centre and a paramagnetic ytterbium(III) centre.

Scheme 92 The synthesis of pentaphenyldiytterbium from dinaphthaleneytterbium.407 158

Polyfluoroaryl divalent lanthanide complexes have been known for some time.408,409 Both bis(pentafluorophenyl)europium and bis(pentafluorophenyl)ytterbium complexes have been crystallographically characterised. Due to the low thermal stability of these complexes409,410 they are often used as synthetic intermediates for protolysis reactions on the way to other organolanthanides.161 An example synthesis is the preparation of

[YbCp*(C6F5)(thf)] 213 (Scheme 93) from what is presumed to be a heteroleptic diaryl intermediate followed by a selective protolysis reaction.411 Like the parent diarylytterbium complex, [YbCp*(C6F5)(thf)3] 213 possesses low thermal stability in the solid state and in solution, requiring storage at -20 °C and is intolerant of non-polar solvents, which induce rapid decomposition.

* 411 Scheme 93 The synthesis of [YbCp (C6F5)(thf)3] 213 from a diphenylytterbium intermediate.

As was observed in the triazenido arylcalcium chemistry (Figure 92) sterically demanding N-terphenyl ligands can kinetically stabilise reactive ytterbium perfluorophenyl organometallics. This strategy has been successful in divalent ytterbium aryl chemistry.324 Both the tetrahydrofuran complex and the base-free ytterbium aryl triazenide 214 and 215 were isolated and characterised (Figure 93). Both of these display the same arene-π interactions as per those observed for the isostructural heavy alkaline earth congeners 209-211 (Figure 92).204 The same degree of solution phase stabilisation is also observed as both the mono-tetrahydrofuran and base free complexes are stable at room temperature with decomposition observed at temperatures above 100 °C in both instances.

159

Figure 93 Arylytterbium triazenide complexes.

Kinetically stabilised σ-bound terphenyl ytterbium complexes are also known. Interestingly in solution a Schlenk-type equilibrium was observed (Scheme 94).412,413 Unlike the non-fluxional triazenido aryl complexes discussed above, solution instability was observed for europium and ytterbium 2,6-terphenyl iodides 216 and 217 with a half-life of approximately one day in solution at ambient temperature. This is still a dramatic improvement on the thermal stabilities observed for other ytterbium(II) and europium(II) aryls.408-411,414,415 The solid state stabilities of complexes 216 and 217 are markedly greater than less hindered counterparts with decomposition of terphenylytterbium iodide 217 observed only above 110 °C in the solid state.

Scheme 94 The Schlenk-type equilibrium observed for europium(II) 216 and ytterbium(II) 217 terphenyls.412,413

In contrast to the arylcalcium iodides (vide supra) a reasonable amount of research has been conducted on the usage of arylytterbium halides in organic synthesis. Evans reported the reaction of alkyl and aryl iodides with ytterbium metal to give alkyl and aryl ytterbium iodide intermediates,414,416 which were then reacted in Grignard type fashion with common electrophiles such as benzophenone. Analysis of the “Grignard- like” intermediates was consistent with ytterbium(II) species as the dominant products according to magnetic measurements. Approximately equimolar ratios of iodine to metal were also observed upon titration. In related investigations both phenylytterbium and -europium iodides were demonstrated to be much more reactive towards N,N- dimethylbenzamide than the analogous Grignard reagent, phenylmagnesium iodide.417(Scheme 95). 160

Scheme 95 The reaction of phenylytterbium, -europium and -magnesium iodides with N,N-dimethylbenzamide.417

By modern standards, these arylytterbium ioidides were poorly characterised. It is highly probable that in solution these compounds form substantially more complex species than the empirical formula of YbPhI(thf)n suggests. For instance, the Solvent II III Separated Ion Pair (SSIP) (vide infra) complex [Yb (dme)4][Yb Ph4(dme)]2 218 was isolated in low yield from dimethoxyethane/hexane solutions of “YbPhI(thf)n” 418 subsequent to the isolation of the divalent complex [YbI2(dme)3]. In view of the aforementioned dinaphthaleneytterbium species (vide supra) it is therefore reasonable to suggest that in solution arylytterbium iodide complexes exist as ytterbium(0), ytterbium(II) and ytterbium(III) species.

Figure 94 Mixed valence phenylytterbium compound isolated from a solution of [YbPhI(dme)n] in dimethoxyethane/hexane.

4.1.3 Solvent Separated Ion Pairs in Calcium and Ytterbium Chemistry

Much of the chemistry of arylcalcium and –ytterbium compounds is dominated by the formation of contact ion pairs (CIPs) in the solid state.37,419 In recent years there has been an increase in the number of structures of ytterbium and calcium compounds that exist as Solvent Separated Ion Pairs (SSIPs), particularly when isolated from tetrahydrofuran solution. This is a structural phenomenon that arises due to a combination of several factors. Most important is the high oxophilicity420,421 of the metal cations. This leads to the coordination of a large number of tetrahydrofuran ligands, which serves to stabilise the charge separated metal cation in both solution and the solid state.

Thus far, a prediction of whether a compound will be a solvent separated or a contact ion pair cannot be made. However it has been observed that this structural phenomenon occurs readily for calcium and ytterbium cations with borate or similarly tetrahedral

161 counterions. Thus far a range of weakly coordinating calcium and ytterbium pentakis or hexakis tetrahydrofuran cations have been reported.387,422-426 A selection of these compounds may be seen in Figure 95. Importantly these display the isostructural relationship that often occurs with calcium and ytterbium compounds.

Figure 95 Some examples of calcium and ytterbium tetraarylmetallate SSIPs.387,422-426

In contrast to the weakly coordinating anions discussed above, there are also emerging examples of cyclopentadienide, fluorenide and triphospholide anions, which usually form strongly bound contact ion pairs,104-106,427-429 forming SSIPs with hexakis tetrahydrofuran calcium and ytterbium.430-433 Some examples of these may be seen in Figure 96. All are characterised by the inclusion of a bulky charge diffuse anion.

Figure 96 Planar anion SSIPs.430-433

One of the most unusual examples of this bonding format is the heterobimetallic 434 cerocene complex [Yb(thf)6][CeCOT‴2]2 219 (Scheme 96). This complex is synthesised from a tetravalent cerocene and excess ytterbium metal. The thermodynamic driving force for this reaction appears to be the strongly oxidising cerium(IV) (Ce4+|Ce3+ E0 1.70 V, Table 1) and the reducing ytterbium metal (Yb3+|Yb - 2.22 V and Yb3+|Yb2+ E0 -1.05 V, Table 1). As per the other SSIPs referred to above, the ytterbium metal in this compound is coordinatively saturated by six tetrahydrofuran ligands in an octahedral geometry and has no interactions with the cyclooctatetraene ligands. This contrasts with the multi-deckered sandwich complexes of lanthanide cycloctatetraenides, which are a common feature of this ligand’s coordination chemistry for the lanthanide elements.435-440 It is presumed that in this instance the high oxophilicity of the divalent ytterbium for the tetrahydrofuran ligands prevents the formation of contact ion pairs in solution and the solid state.

162

434 Scheme 96 The synthesis of heterobimetallic [Yb(thf)6][CeCOT‴2]2 219 from ytterbium metal and [CeCOT‴2].

4.1.4 Purpose of the Chapter

This chapter seeks to further the chemistry of ytterbium and calcium aryls. Specifically it was considered possible that the use of an extremely sterically demanding aryl iodide ligand precursor may kinetically stabilise both an arylytterbium and -calcium complex. To that end the reaction of Dipp*I 49 synthesised in Chapter II, with elemental ytterbium and calcium was investigated. The similarities in reactivity were examined in order to find parallels between the chemistry of each respective element’s aryl organometallic chemistry. The stability of the products was also examined relative to known arylcalcium and -ytterbium complexes.

4.2 Results and Discussion 4.2.1 Synthesis of a Calcium Fluorenide Solvent Separated Ion Pair

Dipp*I 49 reacts readily with unactivated calcium metal powder with sonication to afford a deep, blood-red suspension. Filtration, concentration in vacuo followed by storage at -25 °C overnight afforded not the desired arylcalcium iodide but an unexpected bright orange-red iodocalcium fluorenide SSIP, [CaI(thf)5][Fluor*]∙THF 220 in low yield (Scheme 97) as determined by single crystal X-ray structure determination. Other products are also observed in this reaction including [CaI2(thf)4], as determined through comparison of the unit cell of an isolated crystal with the published data,441 as well as Dipp*H 50, using 1H NMR spectroscopy (vide infra). SSIP 220 is an extremely thermally sensitive compound in the solid state and decomposes readily at room temperature once isolated. Brief exposure to moisture or vacuum led to rapid loss of the bright orange colouration of the solid. It is presumed that this is caused by the lability of the tetrahydrofuran ligands in the coordination sphere of the calcium as well as the extreme hydrolysis sensitivity of the fluorenide counter ion.

163

Scheme 97 The synthesis of [CaI(thf)5][Fluor*] 220 from Dipp*I 49 and calcium metal. The formation of the fluorenide component of 220 from the presumed arylcalcium iodide intermediate has no precedent. An examination of the decomposition mechanisms of arylcalcium iodides (vide supra) suggests several by-products that should be observable if decomposition occurs via a recognised mechanism. Firstly, the predominant decomposition reaction for arylcalcium iodides in tetrahydrofuran is the cleavage of ethereal solvents by deprotonation at the α position of tetrahydrofuran (See Scheme 89 in Section 4.1.1). For phenylcalcium iodide, this reaction yields an arylcalcium oxide cluster as the final calcium-containing product, as well as large amounts of protonated arene and ethylene gas from the α-deprotonation of tetrahydrofuran and [3+2] cycloreversion of the resulting alkylcalcium.

The reaction in Scheme 97 was repeated on the NMR scale at room temperature in

THF-d8 over 18 hours at room temperature in order to observe any of the possible intermediates and gain insight on the reaction mechanism. A plausible mechanism was determined based upon these observations (Scheme 98). At high concentrations the SSIP 220 precipitates readily from solution and as such characterisation is difficult by 1H NMR spectroscopy. Under these conditions the major solution phase product is Dipp*H 50. At lower concentrations all products remain soluble at room temperature however the extremely complicated nature of the 1H NMR spectrum prevents meaningful spectroscopic characterisation of the SSIP 220.

164

Scheme 98 The proposed mechanism for the formation of fluorenide SSIP 220.

Ethylene was not observed in the 1H NMR spectrum. This is unsurprising given this 1 would generate C2D4 from THF-d8.As previously stated, the major product by H NMR spectroscopy is Dipp*H 50, with no evidence of deuterium incorporation at the ipso- carbon. This is inconsistent with tetrahydrofuran deuteron extraction, by the aryl calcium anion via an acid-base mechanism, suggesting a proton source other than the solvent. It is proposed that the benzyhydryl functional group serves as a viable alternative as a proton source. Triphenylmethane has a reported pKa in tetrahydrofuran of 34.6442 and may be considered as a model compound for the acidity of the benzyhydryl functional group in these compounds. It is noteworthy that the pKa of the arene protons of Dipp*H 50 is likely to be approximately 40 based upon the reported 443,444 Ar-H pKa of benzene and toluene in tetrahydrofuran, 41 and 40.7 respectively. Thus in the absence of α-tetrahydrofuran deuteron extraction the likely proton source for the generation of Dipp*H 50 is the benzhydryl functional group of another Dipp* molecule, e.g. Dipp*I 49, Dipp*H 50 or [Dipp*CaI] 221.

The proposed product of this initial acid-base reaction 222, is believed to be highly reactive towards a metal-halogen exchange reaction as an alkylcalcium iodide. This is based upon the observations made by Westerhausen regarding the reactivity of 2,4,6- tri(tert-butyl)phenylcalcium iodide.386 It is proposed that the diiodinated Dipp* product of this reaction intermediate 223 reacts with elemental calcium to form a transient

165 triphenylmethyl radical. The reaction of triphenylmethyl halides with reducing metals has long been known as a method for the generation of triphenylmethyl radicals.445 For example, the reaction of bromotriphenylmethane with half an equivalent of magnesium metal is known to generate the triphenylmethyl radical.446 Alkaline earth radicals have also long been implied in the mechanism of formation of Grignard reagents at bulk magnesium metal.447-452 The generation of a radical or anion at this position without the initial metal-halogen exchange was considered unlikely given the observation that the alkaline earth metals do not react with triphenylmethane without the addition of ammonia.453 A radical mechanism is proposed due to the well characterised stability of the triphenylmethanides of the alkaline earth metals,453-455 suggesting that an anionic intermediate is implausible here. Moreover, the triphenylmethyl radical is known to cyclise to give 9-phenylfluorene, which is analogous to what is observed here.456 Bond formation then occurs between the calcium and organic two radical at the ortho position of a phenyl substituent to give intermediate 226. Cyclisation then occurs at the ipso carbon with loss of calcium(II) diiodide. Calcium iodide tetrahydrofuran complex has been isolated from the reaction mixtures and identified by comparison to the published unit cell441 from a single crystal X-ray experiment. It is proposed that re-aromatisation then occurs upon the deprotonation of the fluorene by a final molecule of [CaDipp*I] to afford SSIP 220. The overall reaction outcome may be seen in Scheme 99.

Scheme 99 The overall reaction between Dipp*I 49 and calcium metal.

[CaI(thf)5][Fluor*] 220 crystallises from a saturated solution in tetrahydrofuran as deep orange-red rods in the monoclinic space group P21/n. One ion pair and one heavily disordered lattice molecule of tetrahydrofuran are included in the asymmetric unit. An illustration of the structure may be seen in Figure 97. The ion pair is well separated in the solid state with the distance measured from Ca(1) to the centroid of the closest fluorenide five-membered ring of 6.086 Å. Analogous distances to the two closest arene centroids are 5.855 Å and 6.387 Å. The geometry at the calcium atom is slightly distorted from ideal octahedral geometry. This is particularly evident for the trans angles O(5)-Ca(1)-O(3) and O(2)-Ca(1)-I(1) which are ca. 10° away from linear 166

(170.94(14)° and 171.07(9)° respectively). This distorted octahedral geometry is consistent with other published structures containing the [CaI(thf)5] cation as part of a SSIP.387,425,457,458 The central five membered ring of the fluorenide has a geometry that is consistent with that of the two reported uncoordinated fluorenide SSIPs that feature calcium or ytterbium 227 and 228 (Figure 98).430,431 The bonds C(28)-C(29) 1.435(6) Å and C(1)-C(6) 1.441(6) Å are the same within error as the equivalent bonds in SSIPs 227 and 228. However C(1)-C(28) (where the bond length is 1.448(6) Å) is statisically similar to the equivalent distance in 227 (1.43(1) Å) but statisically different from the equivalent distance in 228 (1.424(4) Å). It is plausible that this any differences observed are a consequence of the greatly increased steric demand of the substituents in calcium SSIP 220 relative to the reported compounds 227 and 228. The remaining three bond lengths for the five membered ring are different outside of error but are to those seen in SSIPs 227 and 228.

Figure 97 Molecular structure of [CaI(thf)5][Fluor*]∙THF 220, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms and lattice tetrahydrofuran molecule omitted for clarity. Coordinated tetrahydrofuran carbon atoms are depicted as wireframes for clarity. Selected bond lengths (Å) and angles (°): Ca(1)-I(1) 3.0753(10), Ca(1)-O(1) 2.360(3), Ca(1)-O(2), 2.371(3), Ca(1)-O(3) 2.360(3), Ca(1)-O(4) 2.360(3), Ca(1)-O(5) 2.349(3), C(9)-C(29) 1.422(6), C(9)-C(6) 1.425(6), C(1)-C(28) 1.448(6), C(28)-C(29) 1.435(6), C(1)-C(6) 1.441(6), Ca(1)-(Fluor* centroid) 6.086, O(5)-Ca(1)-O(4) 91.54(12), O(5)-Ca(1)-O(3) 170.94(14), O(4)-Ca(1)-O(3) 84.00(12), O(5)-Ca(1)-O(1) 89.65(11), O(4)-Ca(1)-O(1) 175.92(13), O(3)-Ca(1)-O(1) 94.27(12), O(5)-Ca(1)-O(2) 85.53(12), O(4)-Ca(1)-O(2) 90.58(12), O(3)-Ca(1)-O(2) 86.63(13), O(1)-Ca(1)-O(2) 85.62(11), O(5)-Ca(1)-I(1) 92.57(9), O(4)-Ca(1)-I(1) 98.20(10), O(3)- Ca(1)-I(1) 95.86(11), O(1)-Ca(1)-I(1) 85.65(8), O(2)-Ca(1)-I(1) 171.07(9).

167

Figure 98 The two published fluorenide SSIPs 227 and 228.430,431

In contrast to the “metal in a box” supramolecular structure suggested by Harder,430,431 which attributed the formation of the calcium and ytterbium 9-trimethylsilyfluorenide SSIPs to the ordering of a two dimensional coordination polymer around the metal centres of two neighbouring SSIPs in the crystal structure, [CaI(thf)5][Fluor*] consists of planes of cation and anion pairs that are separated by disordered tetrahydrofuran molecules that readily fill the voids between these sheets. These planes are oriented parallel with the hypotenuse of the a and c axes in the cell and an illustration of this macro structure is provided in Figure 99.

Figure 99 Extended packing diagram of [CaI(thf)5][Fluor*]∙THF 220 viewed along the b axis. Note the packing of cation (green) and anion (red) planes and the layer of disordered tetrahydrofuran (purple) between each plane. The unit cell is overlaid in blue.

The reaction of calcium metal with Dipp*I 49 on the preparative scale followed by hydrolysis gives the Dipp*H 50 as the major product after chromatography (Scheme 100). Other products are observed but cannot be readily separated by flash column chromatography, instead being co-eluted after flushing the column with strongly polar 1 solvents. They are also difficult to identify by H NMR spectroscopy in C6D6 as there are multiple chemical environments observed in regions of the spectrum consistent with 168 a compound containing the Dipp* moiety. In spite of this the isolated yield of Dipp*H 50 from this reaction (72%) is consistent with the theoretical amount calculated (66%) from the overall reaction of the proposed mechanism (Schemes 98 and 99)

Scheme 100 The synthesis of Dipp*H 50 from Dipp*I 49 and calcium metal.

Long term storage of preparative tetrahydrofuran solutions at -25 °C invariably lead to decomposition and deposition of colourless crystals in several different morphologies.

From these [CaI2(thf)4] was readily identified as part of this mixture by comparison of the unit cell to the published values.441 A second colourless crystal morphology was identified as the novel tricalcium triiodide bis-hydroxyl cluster

[Ca3(μ-I)3(μ-OH)2(thf)9]I 229. Attempts to refine this structure were complicated by the 1 high symmetry (Asymmetric Unit = /12 of an ion pair) and substantial disorder in the tetrahydrofuran hydrocarbyl chain. The crystallographic information file for this structure is included with all others as part of this thesis. It is clear however that long term storage in solution is not feasible.

4.2.2 Synthesis of a Ytterbium Fluorenide Solvent Separated Ion Pair

The reaction of excess elemental ytterbium with a solution of Dipp*I 49 with sonication in tetrahydrofuran over 12 hours gave the analogous ytterbium SSIP,

[YbI(thf)5][Fluor*] 230 in similar low yield. In tetrahydrofuran solution the same blood red colour as that seen during the preparation of calcium SSIP 220 is observed, providing a clear visual indicator of the initiation of the reaction. Concentration in vacuo and storage at -25 °C afforded the SSIP [YbI(thf)5][Fluor*] 230 as a crop of yellow-orange blocks (Scheme 101) free of lattice solvent (cf. lattice tetrahydrofuran observed in the solid state structure of calcium SSIP 220). This reaction outcome is consistent with a similar mechanism to that proposed above (Scheme 98).

169

Scheme 101 The synthesis of [YbI(thf)5][Fluor*] 230 from elemental ytterbium and Dipp*I 49.

[YbI(thf)5][Fluor*] 230 crystallises in the orthorhombic space group P 212121 with one ion pair in the asymmetric unit. In contrast to the structure of the calcium analogue 220 there is no lattice solvent included in the lattice of SSIP 230. The distance to the fluorenide centroid is measured in [YbI(thf)5][Fluor*] 230 as 6.259 Å whereas in

[CaI(thf)5][Fluor*] 220 this distance is shorter at 6.086 Å. This distance is still much longer than the average bonding distance observed for reported fluorenides coordinated in an η5 fashion to ytterbium(II), which have a range of 2.414-2.519 Å.428,459,460 The

[YbI(thf)5] cation adopts the same distorted octahedral geometry that was observed for the calcium analogue 220. This distortion is particularly evident at O(4)-Yb(1)-O(2), O(5)-Yb(1)-I(1) and O3-Yb(1)-O(1), which are measured as 168.06(17)°, 171.43(10)° and 176.21(16)° respectively. Interestingly this is the first report of this cation in a structure.

170

Figure 100 Molecular structure of [YbI(thf)5][Fluor*] 230 , POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. Coordinated tetrahydrofuran carbon atoms are depicted as wireframes for clarity. Selected bond lengths (Å) and angles (°): Yb(1)-I(1) 3.0802(6), Yb(1)-O(1) 2.382(4), Yb(1)-O(2) 2.399(4), Yb(1)- O(3) 2.379(5), Yb(1)-O(4) 2.414(5), Yb(1)-O(5) 2.411(5), C(9)-C(29) 1.483(9), C(9)-C(6) 1.375(8), C(1)-C(28) 1.479(9), C(28)-C(29) 1.399(9), C(1)-C(6) 1.461(9), Yb(1)-(Centroid) 6.259, O(5)-Yb(1)-O(4) 84.20(17), O(5)- Yb(1)-O(3) 93.42(16), O(4)-Yb(1)-O(3) 85.54(18), O(5)-Yb(1)-O(1) 83.15(14), O(4)-Yb(1)-O(1) 92.48(17), O(3)- Yb(1)-O(1) 176.21(16), O(5)-Yb(1)-O(2) 84.60(15), O(4)-Yb(1)-O(2) 168.06(17), O(3)-Yb(1)-O(2)- 90.98(16), O(1)-Yb(1)-O(2) 90.31(16), O(5)-Yb(1)-I(1) 171.43(10), O(4)-Yb(1)-I(1) 97.84(13), O(3)-Yb(1)-I(1) 95.04(13), O(1)-Yb(1)-I(1) 88.43(11), O(2)-Yb(1)-I(1) 93.84(10).

The packing diagrams seen in Figures 101 and 102 displays the different ion packing structure of [YbI(thf)5][Fluor*] 230 compared to that of [CaI(thf)5][Fluor*] 220. In the absence of lattice tetrahydrofuran the ion pairs of SSIP 230 pack much more tightly relative to calcium SSIP 220. This may be seen in Figure 102, which shows the view looking down the c axis. Unlike [CaI(thf)5][Fluor*] 230, the cation and anion pairs are not organised into planes. In Figure 102 the view looking down the a axis shows an overall herringbone-like structure of the cation and anion pairs. The supramolecular structure may be conceived as zig-zagging chains of fluorenide anions, with the voids between filled by the cation.

171

Figure 101 The supramolecular structure of [YbI(thf)5][Fluor*] 230 looking down the c axis demonstrating the tightly packed anion (red) and cation(green) pairs. The unit cell is overlaid in blue

Figure 102 Extended supramolecular structure of [YbI(thf)5][Fluor*] 230 looking down the a axis demonstrating the packing of the herringbone-like structure. For clarity the anions are coloured red and the unit cell overlaid in blue.

172

In a fashion similar to the calcium analogue, [YbI(thf)5][Fluor*] 230 possesses limited stability in the solution phase at -25 °C. In this instance ytterbium containing decomposition products, as per the calcium analogue (vide supra), have not been forthcoming. However an oxidatively coupled difluorene (Fluor*)2 231 is observed to form in trace quantities and has been characterised by a single crystal X-ray diffraction experiment. Although this oxidation may be caused by the leakage of trace oxygen into the Schlenk flask containing solutions of [YbI(thf)5][Fluor*] 230 (Scheme 102), a redox process involving a ytterbium(II)/ytterbium(III) or ytterbium(0)/ytterbium(II) redox couple cannot be discounted at this stage. Similar, albeit reductive, coupling processes have been observed previously with 9-fluorenenone at divalent samarium and ytterbium (See Scheme 22 in Section 1.3.5).139

Scheme 102 The proposed formation of (Fluor*)2 231 by oxidation of SSIP 230. Difluorene 231 crystallises from tetrahydrofuran as pale yellow parallelograms in the triclinic space group P1¯ with two and a half tetrahydrofuran molecules in the lattice. The molecule is highly strained about the C(7A)-C(7B) bond, which manifests as an elongation of this bond length to 1.641(4) Å. The dihedral angle (63.8°) across this bond vector (C(8A)C(7A)-C(7B)C(8B)) is limited by the steric buttressing between opposing benzhydryl functional groups leading to the conformation observed in the structure.

173

Figure 103 Molecular structure of [YbI(thf)5][Fluor*]∙2.5THF 231, POV-RAY illustration, 50% thermal ellipsoids, all hydrogen atoms omitted for clarity. Lattice tetrahydrofuran molecules are omitted for clarity. Selected bond lengths (Å) and angles (°): C(7A)-C(7B) 1.641(4), C(2A)-C(7A) 1.544(4), C(1A)-C(2A) 1.417(4), C(1A)-C(9A) 1.467(4), C(8A)-C(9A) 1.416(4), C(7A)-C(8A) 1.525(4), C(2B)-C(7B) 1.546(3), C(7B)-C(8B) 1.524(4), C(1B)- C(2B) 1.400(4), C(1B)-C(9B) 1.475(4), C(8B)-C(9B) 1.406(4), C(8A)-C(7A)-C(2A) 100.8(2), C(8B)-C(7B)-C(2B) 100.5(2).

Attempts to further extend this chemistry to other lanthanide metals were not as successful. Under the same or more forcing conditions, neither neodymium nor lanthanum metal were observed to react with Dipp*I 49. Samarium does react with Dipp*I 49 under similar conditions to give an intense brown solution however attempts to isolate the product yielded only amorphous brown gums.

4.3 Conclusions and Future Work 4.3.1 Conclusions

This chapter aimed to extend the chemistry of calcium and ytterbium aryls. It was anticipated the extremely sterically demanding Dipp* moiety would allow the isolation of thermally robust metal aryls like the synthetically useful Dipp*MgI 58 used in Chapter II. In that regard this aim was not met. Instead of the desired metal aryls solvent separated fluorenide ion pairs 220 and 230 were observed in low yields, as crystalline products. A mechanism of formation based upon the observations herein was proposed. The possibility of extending this chemistry to other lanthanides was examined but no 174 tractable products were isolable for samarium and other lanthanides failed to react. The low solid state and solution stability prevented detailed characterisation of these compounds 220 and 230. The complicated nature of the other organic products of this reaction also prevented the isolation of the fluorene component.

4.3.2 Future Work

The fluorene component of the SSIPs 220 and 230, [Fluor*H] is a desirable target. A plausible synthesis by another route is the logical extension of this chapter. It is anticipated that the successful synthesis of this fluorene would allow the further investigation and characterisation of SSIPs of the lanthanide and alkaline earth metals and whether CIPs can be synthesised using this ligand. It is of specific interest if this behaviour is observed for the larger metals of the lanthanides and alkaline earths.

175

4.4 Appendix V: Crystal Data for Structures Collected for this Chapter

[CaI(thf) ][Fluor*] [YbI(thf) ][Fluor*] (Fluor*) 5 5 2 Chemical Moiety Formula C57H73CaIO6 C53H65IO5Yb C76H70O2.5

Chemical Sum Formula C57H73CaIO6 C53H65IO5Yb C76H70O2.5 Mol. Weight/ g mol-1 1021.13 1081.99 1023.32 Temperature/ K 150(2) 150(2) 150(2) Crystal habit, colour Rod, orange-red Block, yellow-orange Parallelogram, pale yellow Crystal Size/ mm 0.2 x 0.19 x 0.03 0.04 x 0.03 x 0.03 0.2 x 0.18 x 0.05 Crystal System monoclinic orthorhombic triclinic

Space Group P21/n P212121 P1¯ a/ Å 9.7232(6) 9.6705(4) 11.9873(8) b/ Å 22.4996(9) 22.6383(10) 14.3476(9) c/ Å 24.5552(14) 22.6552(9) 18.8460(12) α/ ° 90.00 90.00 105.591(3) β/ ° 94.366(2) 90.00 97.793(3) γ/ ° 90.00 90.00 111.961(3) Volume/ Å3 5356.3(5) 4959.8(4) 2792.5(3) Z 4 4 2 Density calcd/ g cm-3 1.266 1.449 1.217 μ/ mm-1 0.741 2.552 0.072

F000 2144 2184 1092 Reflections Collected 14040 9684 13705 Unique Reflections 5417 6323 6074 Parameters Varied 672 543 734 R(int) 0.0936 0.068 0.0665 R1 0.0578 0.0394 0.0805 wR1 (all data) 0.1535 0.0754 0.2369 Largest diff. peak and hole/ e Å-3 0.59, −0.63 0.887, -1.166 0.486, -0.398 GOOF 0.931 1.026 0.97

176

Chapter V Experimental

5.1 General Procedures

Unless otherwise noted, all synthetic procedures were conducted using standard Schlenk and glovebox techniques under an atmosphere of ultra-high purity (UHP) argon. All glassware for Schlenk operations was flame-dried under vacuum prior to use. 3 Å molecular sieves were activated by heating to 300 °C for 24 h and then cooling to room temperature in vacuo. Diethyl ether, tetrahydrofuran and toluene were collected from an Innovative Technology PureSolv MD-5 solvent purification system, distilled under an atmosphere of dinitrogen from sodium diphenyl ketyl once a persistent purple colour was observed. These solvents were stored in flasks fitted with J. Youngs high vacuum stopcocks over activated 3 Å molecular sieves under UHP argon and freeze-thaw degassed immediately before use. 1,2-Dimethoxyethane was distilled under an atmosphere of UHP argon from sodium diphenyl ketyl once a persistent purple colour was observed, then stored and handled as above. nHexane was dried over sodium wire, sparged with dinitrogen and distilled from sodium diphenyl ketyl solubilised by the addition of tetraglyme under dinitrogen once a persistent purple colour was observed, then stored in flasks fitted with J. Youngs high vacuum stopcocks over Na/K alloy under UHP argon and freeze-thaw degassed immediately before use. nPentane was collected from an Innovative Technology PureSolv MD-5 solvent purification system, stored in flasks fitted with J. Youngs high vacuum stopcocks over Na/K alloy under UHP argon and freeze-thaw degassed immediately before use. Anhydrous dichloromethane was collected from an Innovative Technology PureSolv MD-7 solvent purification system and stored in flasks fitted with J. Youngs high vacuum stopcocks over activated 3 Å molecular sieves under UHP argon. Pyridine was distilled from CaH2 under an atmosphere of UHP argon and stored in flasks fitted with J. Youngs high vacuum stopcocks over activated 3 Å molecular sieves under UHP argon. All air- sensitive reagents were stored in a Saffron Scientific Alpha glovebox under an atmosphere of argon. Lanthanide and alkaline earth metals were stored in the glovebox as ingots and filed to powders immediately before use. Solutions of nbutyllithium in hexanes were decanted from commercial bottles and stored in flasks fitted with glass J.

177

Youngs high vacuum stopcocks under UHP argon at 5 °C and standardised using 1H NMR spectroscopy according to the method of Hoye.461 All reaction temperatures were measured in the cooling or heating bath fluid. Flash column chromatography was performed on Davisil 60 Å, 40-63 μm silica according to Leonard et al..462 Reduced pressure, where referred to in the text, is delivered by a single stage diaphragm pump. In vacuo refers to a vacuum delivered by a dual stage, oil sealed, rotary vane pump through a liquid nitrogen cooled trap. The following compounds were prepared by the 328 255 literature methods [LnI3(thf)n] (Ln = Nd, Sm), [TlN3Dipp2], [Ln{N(SiMe3)2}3] (Ln 119 463 = Ce, Nd, Sm, Yb) and LiNCy2. All other materials were purchased from commercial suppliers (Sigma-Aldrich, Alfa Aesar, Ajax Finechem, Baotou Rare Earth and Chem-Supply) and used as received.

5.2 Characterisation

1H and 13C NMR spectra were recorded on either a Bruker Avance DPX 300 spectrometer at 300.30 and 75.5 MHz respectively, an Avance III Spectrometer at 400.13 and 100.62 MHz respectively, or an Avance III spectrometer at 500.13 and 125.76 MHz respectively. The probe temperature was 298 K unless otherwise specified.

Spectra were recorded in the solution state with either benzene-d6, chloroform-d1 or tetrahydrofuran-d8 as lock solvents. Chemical shifts were recorded on the δ scale and referenced to a residual protonated solvent resonance at the values reported by 280 Fulmer. Chloroform-d1 was used as received. Benzene-d6 and tetrahydrofuran-d8 were dried over sodium metal and freeze-thaw degassed prior to use. Signal multiplicities are denoted as either singlet (s), doublet (d), triplet (t), quartet (q), quintet (quin), sextet (sext), septet (sept) or multiplet (m) where appropriate. Infrared spectra were recorded on either a Nicolet Avatar 320 or a iS10 FT-IR spectrometer over the range 4000 to 400 cm-1. Samples were prepared on sodium chloride plates as Nujol mulls, in an argon filled glovebox where necessary. Spectra are reported in wavenumbers (cm-1) and the intensities of the absorbances noted thus: strong (s), medium (m), or weak (w) and labelled broad (br) or sharp (sh) as appropriate. Elemental analyses were carried out either by Mr Bob McAllister at the Campbell Microanalytical Laboratory, University of Otago, Dunedin, New Zealand or by Ms Sasha Melnitchenko at the Research School of Chemistry Microanalytical Unit, Australia National 178

University, Canberra, Australia. Melting points were determined in sealed glass capillaries under argon on a Gallenkamp MPD350 instrument and are uncorrected. Single crystal X-ray diffraction (XRD) data collection was carried out on a Bruker- Nonius Kappa diffractometer with either a Mo source fitted with a triumph monochromators or a Mo micro-focus X-ray source, Quazar Multilayer optics and an APEX II CCD detector array. Structures were processed and refined using the SHELX- 97 suite of programs464 with the X-Seed GUI.465

5.1 Synthesis

1,3-Bis(2,6-diisopropylphenyl)triazene, Dipp2N3H 40

In air, iAmONO (148.0 g, 834.8 mmol) cooled to -20 °C was added to 2,6-diisopropylaniline (71.47 g, 403.1 mmol) at room temperature and the mixture stirred for 30 seconds. The flask was allowed to stand undisturbed for 2 hours at room temperature, during which time vigorous gas evolution was observed and the yellow solution darkened to red and then black-brown. This solution was stored at -20 °C for 7 days before the crude crystalline product was isolated using filtration and washed with the minimum amount (ca. 3 x 25 cm3) of sub-zero methanol. The filtrate and washings were combined and allowed to evaporate at room temperature and atmospheric pressure to ca. 25 cm3. Filtration of the resulting black tar like suspension and washing of the isolated solid with the minimum amount of sub-zero methanol afforded a second crop of crude product. The combined crops were then thrice recrystallised from npentane (ca. 3 100 cm ) to afford Dipp2N3H as colourless blocks (30.39 g, 83.14 mmol, 41%). Spectroscopic data was consistent with the reported values.200 1H NMR (300.30 MHz,

C6D6) δ 1.16 (d, J = 6.9 Hz, 24H, CH(CH3)2), 3.31 (m, 4H, CH(CH3)2), 7.01-7.16 (m,

6H, ArH), 9.01 (s br, 1H NH). Elemental analysis calculated (%) for C24H35N3: C, 78.85; H, 9.65; N, 11.49. Found: C, 79.00; H, 9.87; N, 11.37.

179

2,6-Dibenzhydryl-4-methylaniline, Dipp*NH2 43

In air, a freshly prepared solution of ZnCl2 (18.56 g, 136.3 mmol) in 3 HCl(aq.) (32% w/v, 26.8 cm , 273 mmol) was added dropwise over 5 mins to a hot (80 °C) stirred mixture of molten p-toluidine (29.120 g, 272.5 mmol) and diphenylmethanol (100.27 g, 544.27 mmol) in a 1 litre round bottom flask. A thick white fog in the reaction vessel was observed during addition. The resulting mixture was heated at 160 °C for 4 hours during which time the melt solidified. The solid mass was cooled to room temperature, dichloromethane (700 cm3) 3 was added and the resulting solution washed with NH4Cl(aq.) (10% w/w, 300 cm ) and 3 sat. NaCl(aq.) (300 cm ). The organic phase was dried over K2CO3, filtered and volatiles were removed under reduced pressure. nHexane (225 cm3) and ethyl acetate (75 cm3) were added and the resulting suspension heated at reflux with vigorous stirring for 30 mins. The suspension was cooled to room temperature and filtered. The residue was n 3 washed with further hexane (ca. 2 x 50 cm ) and dried by suction to afford Dipp*NH2 as a fine white powder (103.13 g, 234.61 mmol, 86%). Spectroscopic data was 236 1 consistent with the reported values. H NMR (300.30 MHz, C6D6) δ 1.92 (s, 3H, p-

CH3), 3.15 (s, 2H, NH2), 5.46 (s, 2H, CHPh2), 6.73 (s, 2H, m-ArH), 6.99-7.06 (m, 20H, ArH).

2,6-Dibenzhydryl-4-methylazidobenzene, Dipp*N3 45

In air, H2SO4 (98%, 18.4 g, 188 mmol) was added dropwise over 5 mins to a stirred, finely divided suspension of Dipp*NH2 (11.05 g, 25.14 mmol) in tetrahydrofuran:water (2.5:1, 400 cm3) cooled to 0

°C. A cold (ca. 5 °C) solution of NaNO2 (5.24 g, 75.9 mmol) in water (10 cm3) was added dropwise over a period of 5 mins and the resulting red-orange suspension was stirred for 4 hours at 0 °C. A cold (ca. 5 °C) solution of NaN3 (3.27 g, 50.3 mmol) in water (10 cm3) was then added dropwise over 10 mins. Vigorous gas evolution was observed. The resulting off-white suspension was allowed to slowly 3 warm to room temperature and then poured into sat. Na2CO3(aq.) (500 cm ). The product was extracted into ethyl acetate (3 x 100 cm3). The combined organic phases were 3 3 washed with sat. Na2CO3(aq.) (500 cm ), water (500 cm ), dried over MgSO4 and

180 filtered. Removal of volatiles under reduced pressure gave the title compound as an off- white fine powder (11.23 g, 24.12 mmol, 96%). m.p. 132-134 °C. 1H NMR (300.30

MHz, C6D6) δ 1.77 (s, 3H, p-CH3), 6.06 (s, 2H, CHPh2), 6.84 (s, 2H, m-ArH), 6.94-7.10 13 (m, 20H, ArH). C NMR (75.52 MHz, C6D6) δ 21.11 (p-CH3), 53.02 (CHPh2), 126.51

(CHAr), 126.88 (CHAr) 128.61 (CHAr), 128.78 (CHAr), 129.89 (CHAr), 130.07 (CHAr), -1 135.52 (CAr), 136.26 (CAr), 140.21 (CAr), 143.53 (CAr). IR (Nujol) ν/cm ; 666 (sh, w), 678 (sh, w), 699 (sh, s), 722 (sh, w), 738 (sh, w), 747 (sh, w), 766 (sh, w), 797 (sh, w), 831 (sh, w), 860 (sh, w), 874 (sh, w), 911 (sh, w), 1003 (sh, w), 1031 (sh, w), 1079 (sh, w), 1121 (sh, w), 1156 (sh, w), 1253 (sh, w), 1280 (sh, w), 1306 (sh, w), 1494 (sh, m), 1599 (sh, w), 1804 (sh, w), 1882 (sh, w), 1953 (sh, w), 2126 (sh, s), 2728 (sh, w), 3025

(sh, w), 3060 (sh, w), 3083 (sh, w). Elemental analysis calculated (%) for C33H27N3: C, 85.13; H, 5.85; N, 9.03. Found: C, 85.02; H, 5.85; N, 9.09.

2,6-Dibenzhydryl-4-methyliodobenzene, Dipp*I 49

In air, a cold (ca. 5 °C), freshly prepared solution of NaNO2 (2.85 g, 41.3 mmol) in water (5 cm3) was added dropwise to a suspension of 3 Dipp*NH2 (8.94 g, 20.3 mmol) in glacial acetic acid (75 cm ) and toluene (20 cm3) at 1 °C over five mins. A red solution slowly formed over the period of one hour as the white solid dissolved. A cold (ca. 5 °C), freshly prepared solution of KI (10.10 g, 60.84 mmol) in water (10 cm3) was then added dropwise over 10 mins. Vigorous gas evolution and formation of a brown precipitate was observed. The suspension was held at 1 °C for 2 hours, then at room temperature 3 for 2 hours with stirring before it was cautiously poured into sat. Na2CO3(aq.) (500 cm ). The product was extracted into ethyl acetate (3 x 100 cm3). The combined organic 3 3 phases were washed with sat. Na2CO3(aq.) (500 cm ), Na2S2O3(aq.) (10% w/w, 500 cm ) 3 and water (500 cm ). The organic phase was dried over Na2SO4, filtered and volatiles were removed under reduced pressure. The crude product was recrystallised from ethyl acetate:hexane (15:85, ca. 125 cm3) to afford the title compound as colourless rods. 1 (5.81 g, 10.6 mmol, 56%). m.p. 163-165 °C. H NMR (300.30 MHz, CDCl3) δ 2.17 (s,

3H, p-CH3), 6.11 (s, 2H, CHPh2), 6.71 (s, 2H, m-ArH), 7.14-7.17 (m, 8H, ArH), 7.27- 13 7.39 (m, 12H, ArH). C NMR (75.52 MHz, CDCl3) δ 21.26 (p-CH3), 62.11 (CHPh2),

108.15 (I-CAr), 126.44 (CHAr), 128.37 (CHAr), 129.95 (CHAr), 130.39 (CHAr), 137.07 181

-1 (CAr), 143.31 (CAr), 147.27 (CAr). IR (Nujol) ν/cm ; 605 (sh, w), 621 (sh, w), 641 (sh, w), 700 (sh, s), 736 (sh, w), 750 (sh, w), 766 (sh, w), 784 (sh, w), 826 (sh, w), 857 (sh, w), 910 (sh, w), 1002 (sh, m), 1078 (sh, w), 1153 (sh, w), 1176 (sh, w), 1243 (sh, w), 1409 (sh, w), 1492 (sh, m), 1556 (sh, w), 1581 (sh, w), 1597 (sh, w), 3023 (sh, w), 3057

(sh, w). Elemental analysis calculated (%) for C33H27I: C, 72.00; H, 4.94. Found: C, 72.36; H, 5.00.

1,3-Bis(2,6-dibenzhydryl-4-methylphenyl)triazene, Dipp*2N3H 39

Attempted preparation: A solution of nBuLi (1.56 mol L-1, 1.85 cm3, 2.89 mmol) in hexanes was added dropwise to a solution of Dipp*I (1.59 g, 2.89 mmol) in tetrahydrofuran (ca. 50 cm3) cooled to -100 °C and the mixture was slowly warmed over four hours to 0 °C. The resulting solution was then cooled to -80 °C and a solution of 3 Dipp*N3 (1.34 g, 2.88 mmol) also in tetrahydrofuran (ca. 50 cm ) was added dropwise over 10 mins. The resulting solution allowed to slowly warm to room temperature over three hours and then heated at reflux for 24 h. After cooling to room temperature the 3 reaction was quenched by slow addition of water (10 cm ) and then sat. NaCl(aq.) (100 cm3). The product was extracted into ethyl acetate (3 x 50 cm3), the organic phases were combined, dried over Na2SO4, filtered and volatiles were removed under reduced pressure. Analysis of the crude material using TLC and 1H NMR spectroscopy gave

Dipp*N3 and Dipp*H as the only identifiable products in an approximately equimolar proportion.

Method B: A suspension of magnesium turnings (0.750 g, 31.0 mmol) and Dipp*I (11.03 g, 20.04 mmol) in tetrahydrofuran (ca. 150 cm3) was heated at reflux overnight. The suspension was cooled to room temperature, filtered and the metal washed with further tetrahydrofuran (3 x 10 cm3). The combined filtrate and washings were added to 3 a solution of Dipp*N3 (8.95 g, 19.2 mmol) in tetrahydrofuran (ca. 150 cm ) and the resulting yellow-orange solution heated at reflux overnight. The solution was cooled to room temperature, water (10 cm3) was added and the resulting suspension poured into 3 3 sat. NaCl(aq.) (400 cm ). The product was extracted into ethyl acetate (3 x 150 cm ), the combined organic phases dried over Na2SO4, filtered and then concentrated under

182 reduced pressure to ca. 300 cm3. The resulting suspension was filtered and the precipitate washed with ethyl acetate:hexane (5:95, 100 cm3). The combined filtrate and washings were concentrated to ca. 150 cm3 and the resulting suspension filtered. The precipitate was washed with ethyl acetate:hexane (5:95, 50 cm3) and the filtrate and washings discarded. The combined precipitates were triturated with boiling ethyl acetate (ca. 400 cm3) for 30 mins with vigorous stirring, cooled to room temperature with stirring and the suspension filtered. The precipitate was washed with ethyl 3 acetate:hexane (5:95, 2 x 100 cm ) and then dried in vacuo to afford Dipp*2N3H as a fine colourless powder. (9.94 g, 11.2 mmol, 58%). 176-178 °C (dec.) 1H NMR (300.30

MHz, C6D6) Major isomer δ 1.85 (s, 6H, p-CH3), 5.76 (s br, 2H,CHPh2, fwhm = 27 Hz),

6.34 (s br, 2H, CHPh2, fwhm = 23 Hz), 6.98 (s br, 44H, ArH, fwhm = 23 Hz), 8.51 (s br, 1H, NH, fwhm = 72 Hz). First minor isomer, overlaps with the major except at δ

5.61 (s, 2H, CHPh2), 6.09 (s, 2H, CHPh2), 7.42 (s, 1H, NH). Second minor isomer, overlaps with the major except at δ 5.37 (s, 2H, CHPh2), 6.06 (s, 2H, CHPh2), 9.63 (s, 1 1H, NH) H NMR (300.30 MHz, CDCl3) δ 2.17 (s, 6H, pCH3), 5.70 (s br, 4H,CHPh2, fwhm = 95 Hz), 6.67 (s br, 4H, ArH, fwhm = 8 Hz), 6.86 (s, 16H, ArH, fwhm = 18 Hz), 7.11 (s, 24H, ArH, fwhm = 11 Hz), 8.26 (s, 1H, NH, fwhm = 21 Hz). 13C NMR (75.52

MHz, CDCl3) δ 21.72 (p-CH3), 51.40 (br, CHPh2), 126.20 (CHAr), 128.24 (CHAr),

129.62 (CHAr), 129.71 (CHAr), 143.59 (br, CAr). No other signals are observed. IR (Nujol) ν/cm-1; 606 (sh, w), 701 (sh, m), 717 (sh, w), 744 (sh, w), 762 (sh, w), 776 (sh, w), 832 (sh, w), 854 (sh, w), 1002 (sh, w), 1030 (sh, w), 1078 (sh, w), 1160 (br, w), 1215 (sh, w), 1249 (sh, w), 1456 (br, s), 1481 (sh, w), 1493 (sh, m), 1599 (sh, w), 3024 (sh, w), 3295(sh, w). Elemental analysis calculated (%) for C66H55N3: C, 89.05; H, 6.23; N, 4.72. Found: C, 89.26; H, 6.38; N, 4.70.

2,6-Dibenzhydryl-1-butyl-4-methylbenzene, Dipp*nBu 57 and 2,6-Dibenzhydryl-4- methylbenzene, Dipp*H 50

A solution of nBuLi (1.35 mol L-1, 7.40 cm3, 10.0 mmol) in hexanes was added dropwise to a solution of Dipp*I (5.502 g, 10.0 mmol) in tetrahydrofuran (ca. 50 cm3) cooled to -90 °C. The resulting yellow-orange suspension was warmed to -10 °C over the period of two hours, during which time the pale yellow 183 precipitate slowly dissolved. Water (10 cm3) was added dropwise over the period of two 3 mins with stirring. The resulting suspension was poured into sat. NaCl(aq.) (ca. 100 cm ) and the products extracted into ethyl acetate (3 x 50 cm3). The combined organic phases were dried over Na2SO4, filtered and then volatiles removed under reduced pressure. The residue was purified by flash column chromatography (dichloromethane:hexane, n 10:90, 20:1 SiO2:Compound) to afford Dipp* Bu (1.31 g, 2.73 mmol, 27%) and Dipp*H (1.92 g, 4.52 mmol, 45%).

n 1 Characterisation data for Dipp* Bu: m.p. 160-163 °C. H NMR (300.30 MHz, CDCl3) δ

0.87-0.92 (m, 3H, CH2CH3), 1.31-1.51 (m, 4H, CH2CH2CH3), 2.11 (s, 3H, p-CH3),

2.49-2.54 (m, 2H, ArCCH2CH2), 5.77 (s, 2H, CHPh2), 6.58 (s, 2H, m-ArH), 7.06-7.08 13 (m, 8H, ArH), 7.19-7.32 (m, 12H, ArH). C NMR (75.52 MHz, C6D6) δ 13.87

(CH2CH3), 21.46 (p-CH3), 23.46 (CH2CH2CH3), 28.30 (CH2CH2CH2), 33.53

(ArCCH2CH2), 52.92 (CHPh2), 126.13 (CHAr), 128.19 (CHAr), 129.37 (CHAr), 129.64

(CHAr), 134.23 (CAr), 136.83 (CAr), 142.00 (CAr), 144.28 (CAr). Elemental analysis calculated (%) for C37H36: C, 92.45; H, 7.55. Found: C, 92.40; H, 7.64.

1 Characterisation data for Dipp*H: m.p. 93-95 °C. H NMR (300.30 MHz, CDCl3) δ

2.27 (s, 3H, p-CH3), 5.50 (s, 2H,CHPh2), 6.81 (s, 1H, i-ArH), 6.84 (s, 2H, m-ArH), 13 7.11-7.14 (m, 8H, ArH), 7.21-7.33 (m, 12H, ArH). C NMR (75.52 MHz, C6D6) δ

21.54 (pCH3), 56.71 (CHPh2), 126.14 (CHAr), 128.09 (CHAr), 128.18 (CHAr), 129.39

(CHAr), 143.70 (CAr), 144.04 (CAr), two aromatic carbons are not observed. IR (Nujol) ν/cm-1; 700 (sh, m), 722 (sh, w), 740 (sh, w), 763 (sh, w), 1030 (sh, w), 1077 (sh, w),

1155 (sh, w), 1494 (sh, w), 1597 (sh, w). Elemental analysis calculated (%) for C33H28: C, 93.35; H, 6.65. Found: C, 93.34; H, 6.66.

Attempted reaction of Dipp*I with elemental lithium

A suspension of lithium powder (42 mg, 6.1 mmol) and Dipp*I (303 mg, 550 μmol) in tetrahydrofuran (ca. 10 cm3) was sonicated at room temperature for three days. Neither consumption of the metal nor colour change of the solution was observed during this time. The suspension was filtered and volatiles removed in vacuo. Analysis of the 1 residue using H NMR spectroscopy showed only unchanged Dipp*I.

184

Attempted syntheses of N,N’-bis(2,6-dibenzhydryl-4-methylphenyl)formamidine 41

Method A: In air, a suspension of Dipp*NH2 (2.17 g, 4.93 mmol), triethylorthoformate (367 mg, 2.48 mmol) and catalytic acetic acid in absolute ethanol (70 cm3) was heated at reflux for 48 hours. Volatiles were removed under reduced pressure. Analysis of the crude material thus obtained using TLC and 1H NMR spectroscopy was consistent with a mixture of starting materials.

Method B: In air, a suspension of Dipp*NH2 (1.09 g, 2.48 mmol) and catalytic acetic acid in triethylorthoformate (733 mg, 4.95 mmol) was heated at reflux for 24 hours. Volatiles were removed under reduced pressure. Analysis of the crude material thus obtained using TLC and 1H NMR spectroscopy was consistent with a mixture of starting materials.

Method C: In air, Dipp*NH2 (2.54 g, 5.79 mmol), triethylorthoformate (432 mg, 2.91 mmol) and catalytic acetic acid were heated at 200 °C in a sealed pressure tube for 48 hours. The tube was cooled to room temperature and washed out with dichloromethane (ca. 2 x 25 cm3). Volatiles were removed under reduced pressure. Analysis of the crude material thus obtained using TLC and 1H NMR spectroscopy was consistent with a mixture of starting materials.

2,6-Dibenzhydryl-4-methylphenyl-N-ethylformamide, Dipp*N(CHO)(Et) 61

In air, a suspension of Dipp*NH2 (2.17 g, 4.93 mmol) in 3 triethylorthoformate (25 cm , 0.15 mol) and H2SO4 (98%, 2 drops) was heated at reflux for 72 hours. Cooling the reaction mixture to room temperature afforded a precipitate, which was isolated through filtration, washed with nhexane (3 x 15 cm3) and air dried to afford the title compound as an off-white solid (2.19 g, 4.36 mmol, 88%). m.p. 215-217 °C. 1H NMR (300.30

MHz, CDCl3) δ 1.20 (t, J = 7.2 Hz, 3H, NCH2CH3), 2.26 (s, 3H, p-CH3), 3.66 (q, J =

7.2 Hz, 3H, NCH2CH3), 5.63 (s, 2H, CHPh2), 6.88 (s, 2H, m-ArH), 7.01-7.12 (m, 8H, 13 ArH), 7.17 (s, 1H, (N(CHO)), 7.2-7.36 (m, 12H, ArH). C NMR (75.52 MHz, C6D6) δ

13.00 (NCH2CH3), 21.76 (p-CH3), 42.39 (NCH2CH3) 51.91 (CHPh2), 126.67 (CHAr),

126.87 (CHAr), 128.51 (CHAr), 128.69 (CHAr), 129.48 (CHAr), 130.44 (CHAr), 135.93

185

(CAr), 138.27 (CAr), 142.83 (CAr), 143.20 (CAr), 143.49 (CAr), 163.86 (N(CHO)). IR (Nujol) ν/cm-1; 606 (sh, w), 702 (sh, m), 728 (sh, w), 750 (sh, w), 765 (sh, w), 1030 (sh, w), 1078 (sh, w), 1110 (sh, w), 1154 (sh, w), 1288 (br, w), 1494 (sh, m), 1599 (sh, w),

1676 (sh, s), 3024 (sh, w). Elemental analysis calculated (%) for C36H33NO: C, 87.24; H, 6.71; N, 2.83. Found: C, 86.91; H, 6.86; N, 2.66.

Methyl-N-(2,6-dibenzhydryl-4-methylphenyl)formimidate, Dipp*NCH(OMe) 62

In air, a suspension of Dipp*NH2 (4.39 g, 9.99 mmol) in trimethylorthoformate (25 cm3) and glacial acetic acid (4 drops) was heated at 110 °C for 18 hours in a distillation apparatus. During this time methanol was removed by distillation. The temperature was then raised to 150 °C and heated at this temperature until trimethylorthoformate began to distil as determined from the boiling point of the distillate. After cooling to room temperature volatiles were removed under reduced pressure and the solid dried in vacuo to afford Dipp*NCH(OMe) as a white powder (4.73 g, 9.82 mmol, 98%). m.p. 136-138 1 °C. H NMR (300.30 MHz, C6D6) δ 1.92 (s, 3H, p-CH3), 3.52 (s, 3H, NCHOCH3), 5.78 13 (s, 2H, CHPh2), 6.35 (s, 1H, NCHOCH3), 6.92-7.13 (m, 22H, ArH). C NMR (75.52

MHz, C6D6) δ 21.20 (p-CH3), 52.91 (CHPh2 and NCHOCH3), 126.52 (CHAr), 128.35

(CHAr), 128.60 (CHAr), 129.40 (CHAr), 130.11 (CHAr), 132.38 (CAr), 136.15 (CAr), -1 143.79 (CAr), 144.63 (CAr), 156.33 (NCHOCH3). IR (Nujol) ν/cm ; 606 (sh, w), 666 (sh, w), 703 (sh, s), 747 (sh, w), 760 (sh, w), 769 (sh, w), 797 (sh, w), 829 (sh, w), 871 (sh, w), 912 (sh, w), 937 (sh, w), 980 (sh, w), 1032 (sh, w), 1078 (sh, w), 1125 (sh, w), 1176 (sh, w), 1197 (sh, s), 1231 (sh, m), 1277 (sh, w), 1422 (sh, w), 1455 (sh, s), 1494 (sh, m), 1597 (sh, w), 1641 (sh, s), 3024 (sh, w), 3082 (sh, w), 3062 (sh, w). Elemental analysis calculated (%) for C35H31NO: C, 87.28; H, 6.49; N, 2.91. Found: C, 87.33; H, 6.32; N, 2.89.

186

Ethyl-N-(2,6-dibenzhydryl-4-methylphenyl)formimidate, Dipp*NCH(OEt) 63

In air, a suspension of Dipp*NH2 (2.176 g, 4.950 mmol) in 3 triethylorthoformate (15 cm , 90 mmol) and 1M HCl(aq.) (3 drops) was heated in a distillation apparatus at 135 °C for 36 hours, during which time ethanol was removed through distillation. Cooling to room temperature afforded a precipitate, which was isolated through filtration, washed with nhexane (3 x 15 cm3) and air dried to afford the title compound as colourless 1 blocks (2.036 g, 4.108 mmol, 83%). m.p. 156-158°C. H NMR (300.30 MHz, C6D6) δ

1.28 (t, J = 7.1 Hz, 3H, OCH2CH3), 1.92 (s, 3H, p-CH3), 4.09 (q, J = 7.1 Hz, 3H, 13 OCH2CH3), 5.81 (s, 2H, CHPh2), 6.36 (s, 1H, NCHO), 6.92-7.14 (m, 22H, ArH). C

NMR (75.52 MHz, C6D6) δ 14.48 (OCH2CH3), 21.22 (p-CH3), 52.90 (CHPh2), 61.71

(OCH2CH3), 126.51 (CHAr), 128.58 (CHAr), 129.42 (CHAr), 130.14 (CHAr), 132.30 -1 (CAr), 136.18 (CAr), 144.10 (CAr), 144.70 (CAr), 156.02 (NCHO). IR (Nujol) ν/cm ; 561 (sh, w), 606 (sh, w), 622 (sh, w), 648 (sh, w), 659 (sh, w), 704 (sh, s), 747 (sh, w), 759 (sh, w), 772 (sh, w), 800 (sh, w), 827 (sh, s), 876 (sh, w), 847 (sh, w), 911 (sh, w), 982 (sh, w), 1015 (sh, w), 1031 (sh, w), 1078 (sh, w), 1107 (sh, w), 1129 (sh, w), 1177 (sh, w), 1195 (sh, s), 1231 (sh, s), 1247 (sh, w), 1277 (sh, m), 1463 (sh, m), 1496 (sh, m), 1582 (sh, w), 1598 (sh, w), 1644 (sh, s), 1686 (sh, w), 3023 (sh, w), 3053 (sh, w), 3082

(sh, w). Elemental analysis calculated (%) for C36H33NO: C, 87.24; H, 6.71; N, 2.83. Found: C, 87.21; H, 6.73; N, 2.78.

Attempted Reaction of Dipp*NCH(OEt) with Dipp*NH2

A suspension of Dipp*NCH(OEt) (1.05 g, 2.11 mmol) and Dipp*NH2 (928 mg, 2.11 mmol) in o-xylene (10 cm3) was heated at 135 °C for 48 hours and then reflux for five hours. Volatiles were removed under reduced pressure after cooling to room temperature. The solid was triturated with nhexane (2 x 10 cm3) and the solid dried under reduced pressure. Analysis of the crude material using 1H NMR spectroscopy and TLC displayed only evidence of starting materials.

187

General Procedure for the Attempted Reaction of Ethyl Formimidate with amines

In air, a mixture of Dipp*NCH(OEt) (1 eq.) and the alkyl or aryl amine (1 eq.) was heated at reflux in ethanol (ca. 100 cm3) overnight. Volatiles were removed under reduced pressure. Analysis of the crude material using 1H NMR spectroscopy and TLC displayed only evidence of starting materials.

The following quantities were used according to the general procedure with similar results:

Dipp*NCHOEt (2.50 g, 5.04 mmol) and DippNH2 (0.89 g, 5.04 mmol)

i 3 Dipp*NCHOEt (2.50 g, 5.04 mmol) and PrNH2 (0.43 cm , 5.0 mmol).

Bis(2,6-dibenzhydryl-4-methylphenyl)urea, (Dipp*NH)2CO 64

In air, triphosgene (1.62 g, 5.46 mmol) was added to a solution of Dipp*NH2 (14.35 g, 32.64 mmol) and 4- dimethylaminopyridine (4.00 g, 32.7 mmol) in dichloromethane (95 cm3) at -20 °C in a 350 cm3 pressure vessel; a white precipitate immediately formed. The vessel was sealed and the reaction mixture heated at 70 °C for 15 days. The resulting suspension was cooled to room temperature, water (100 cm3) was added and the suspension stirred for 30 mins. The organic phase was then separated and washed with further water (2 x 100 cm3), dried over anhydrous Na2SO4, filtered and volatiles were removed under reduced pressure. The crude product was washed with toluene (3 x 25 cm3), hexane (50 cm 3) and then dried in vacuo overnight to afford (Dipp*NH)2CO as a fine white powder (11.87 g, 1 13.11 mmol, 80%). m.p. >250 °C. H NMR (300.30 MHz, C6D6) δ 1.82 (s, 6H, p-CH3),

4.55 (s, 2H, (HN)2CO), 6.28 (s, 4H, CHPh2), 6.92 (s, 4H, m-ArH), 6.95-7.08 (m, 24H, 13 ArH), 7.25 (s br, 16H, ArH, fwhm = 60 Hz). C NMR (75.52 MHz, C6D6) δ 20.91 (p-

CH3), 52.02 (CHPh2), 126.3 (CHAr), 126.93 (CAr), 127.98 (CHAr), 128.31 (CHAr),

129.02 (CAr), 129.90 (CHAr), 132.17 (CAr), 143.17 (CAr), 154.62 (N(C=O)N). IR (Nujol) ν/cm-1; 606 (sh, w), 700 (sh, s), 727 (sh, m), 746 (sh, w), 769 (sh, w), 882 (sh, w), 1031 (sh, w), 1078 (sh, w), 1215 (sh, m), 1264 (sh, w), 1489 (sh, s), 1598 (sh, w), 1702 (sh, s),

188

3022 (sh, w), 3055 (sh, w), 3381 (sh, s). Elemental analysis calculated (%) for

C67H56N2O: C, 88.90; H, 6.24; N, 3.09. Found: C, 88.79; H, 6.16; N, 3.09.

Bis(2,6-dibenzhydryl-4-methylphenyl)carbodiimide, Dipp*NCNDipp* 66

A suspension of basic Al2O3 (14.59 g, 143.1 mmol), P2O5 (20.8 g, 146.5 mmol) and (Dipp*NH)2CO (12.94 g, 14.30 mmol) in anhydrous pyridine (175 cm3) was heated at reflux with vigorous stirring for 2 days. The colourless suspended solids were observed to form a thick orange tar within 30 mins. The suspension was cooled to room temperature and filtered. The precipitate was washed with toluene (4 x 100 cm3), the filtrate and the washings combined and volatiles were removed under reduced pressure. The resulting white solid was redissolved in dichloromethane (300 cm3), 3 3 washed with 1 M HCl(aq.) (2 x 100 cm ) and sat. NaCl(aq.) (100 cm ). The organic phase was dried over anhydrous MgSO4, filtered and volatiles were removed under reduced pressure. The residue was recrystallised from CHCl3 to afford Dipp*NCNDipp* as colourless needles over three crops (9.286 g, 10.47 mmol, 73%). m.p. 239-242 °C. 1H

NMR (300.30 MHz, C6D6) δ 1.82 (s, 6H, p-CH3), 6.01 (s, 4H, CHPh2), 6.86 (s, 4H, m- 13 ArH), 6.94-7.06 (m, 40H, ArH). C NMR (75.52 MHz, C6D6) δ 21.16 (p-CH3), 52.72

(CHPh2), 126.64 (CHAr), 128.35 (CHAr), 128.62 (CHAr), 129.97 (CHAr), 130.11 (CHAr), -1 132.91 (CAr), 133.98 (CAr), 139.78 (CAr), 143.82 (N=C=N). IR (Nujol) ν/cm ; 573 (sh, w), 605 (sh, w), 700 (sh, s), 721 (sh, w), 747 (sh, m), 760 (sh, w), 833 (sh, w), 855 (sh, w), 1030 (sh, w), 1078 (sh, w), 1121 (sh, w), 1157 (sh, w), 1205 (sh, w), 1244 (sh, w), 1262 (sh, w), 1492 (sh, m), 1566 (sh, w), 1599 (sh, w), 2102 (sh, w), 2164 (sh, s), 3024

(sh, w). Elemental analysis calculated (%) for C67H54N2: C, 90.71; H, 6.14; N, 3.16. Found: C, 90.58; H, 6.24; N, 3.18.

General Procedure for the Reaction of Dipp*NCNDipp* with Hydride Reductants.

A solution of Dipp*NCNDipp* (1 eq.) and the hydride reductant (1 eq.) was stirred in either tetrahydrofuran or toluene at 60 or 100 °C respectively for 48 hours. Methanol was then added and the resulting suspension stirred for 30 mins. The product was then extracted into ethyl acetate (3 x 50 cm3), the combined organic phases washed with 3 water (ca. 100 cm ), dried over Na2SO4, filtered and then volatiles removed under 189 reduced pressure. Analysis of the crude material in all cases gave a mixture of

Dipp*NH2 and Dipp*NCNDipp* in approximately a 2:1 ratio as determined using TLC 1 and H NMR spectroscopy in C6D6.

The following quantities were used according to the general procedure:

Dipp*NCNDipp* (884 mg, 996 μmol) and [KHBEt3] (139 mg, 1.00 mmol) in tetrahydrofuran (ca. 50 cm3).

-1 Dipp*NCNDipp* (889 mg, 1.00 mmol) and [NaHBEt3] (1.09 mol L in toluene, 0.95 cm3, 1.0 mmol) in tetrahydrofuran (ca. 50 cm3).

-1 Dipp*NCNDipp* (440 mg, 496 μmol) and [BH3∙SMe2] (2.0 mol L in tetrahydrofuran, 0.25 cm3, 0.50 mmol) in tetrahydrofuran (ca. 25 cm3)

Dipp*NCNDipp* (4.44 g, 5.00 mmol) and 9-Borabicyclo(3.3.1)nonane (611 mg, 2.50 mmol) in toluene (ca. 100 cm3).

(2,6-Dibenzhydryl-4-methylphenyl)formamide, Dipp*NH(CHO) 74

In air, a mixture of Dipp*NH2 (13.20 g, 30.03 mmol) and formic acid (88%, 20.43 g, 390.6 mmol) in toluene (75 cm3) was heated at reflux in a flask fitted with a Dean-Stark apparatus for 60 hours. Condensed formic acid was returned to the reaction vessel approximately every 12 hours. Volatiles were removed under reduced pressure after cooling to room temperature. The residue was loaded onto a short column of silica and eluted with nhexane:toluene (10:90, ca. 500 cm3) and this solution discarded. Elution with dichloromethane (ca. 500 cm3) followed by removal of volatiles under reduced pressure, washing with diethyl ether (ca. 50 cm3) and drying by suction afforded the product as a white powder. Yield (9.95 g, 21.3 mmol, 71%). m.p. 196-198 °C. 1H NMR

(300.30 MHz, C6D6) Major (syn) isomer δ 1.77 (s, 3H, p-CH3), 5.26 (s, 1H, NH(CHO)),

5.75 (s, 2H, CHPh2), 6.81 (s, 2H, m-ArH), 6.99-7.09 (m, 20H, ArH), 7.51 (d, J = 1.3 Hz, 1H, (N(CHO)). The minor (anti) isomer overlaps with the major except at δ 5.72 (s, 1 2H, CHPh2), 6.84 (s, 2H, m-ArH), 7.61 (d, J = 11.6 Hz, 1H, (N(CHO)). H NMR

(300.30 MHz, CDCl3) Major (syn) isomer δ 2.17 (s, 3H, p-CH3), 5.70 (s, 2H, CHPh2),

190

6.06 (s, 1H, NH(CHO)), 6.66 (s, 2H, m-ArH), 7.08-7.11 (m, 8H, ArH), 7.22-7.36 (m, 12H, ArH), 8.18 (s, 1H, (N(CHO)). The minor (anti) isomer overlaps with the major except at δ 2.22 (s, 3H, p-CH3) 5.74 (s, 2H, CHPh2), 6.61 (s, 1H, NH(CHO)), 6.76 (s, 13 2H, m-ArH), 7.47 (d, J = 11.9 Hz, 1H, (N(CHO)). C NMR (75.52 MHz, C6D6) Peaks that can be definitively assigned to one isomer are denoted where possible. δ 21.23 (p-

CH3), 52.53 (CHPh2, major), 52.93 (CHPh2 minor) 126.75 (CHAr), 126.93 (CHAr),

128.35 (CHAr), 128.70 (CHAr), 128.85 (CHAr), 129.86 (CHAr), 129.95 (CHAr), 130.02

(CHAr), 137.74 (CAr), 142.57 (CAr), 143.05 (CAr), 143.83 (CAr), 144.15 (CAr), 159.49 13 (N(CHO) major), 163.53 (N(CHO) minor). C NMR (75.52 MHz, CDCl3) Peaks that can be definitively assigned to one isomer are denoted where possible. δ 21.70 (p-CH3, major), 21.72 (p-CH3, minor), 52.38 (CHPh2, major) 52.49 (CHPh2, minor), 126.57

(CHAr), 126.80 (CHAr), 128.51 (CHAr), 128.67 (CHAr), 129.46 (CHAr), 129.54 (CHAr),

129.59 (CHAr), 129.64 (CHAr), 129.79 (CAr), 130.24 (CAr), 137.71 (CAr), 138.01 (CAr),

141.83 (CAr), 142.49 (CAr), 143.02 (CAr), 143.13 (CAr), 160.07 (N(CHO) major), 164.38 (N(CHO) minor). ( IR (Nujol) ν/cm-1; 569 (sh, w), 607 (sh, w), 644 (sh, w), 701 (sh, s), 752 (sh, w), 764 (sh, w), 832 (sh, w), 852 (sh, w), 889 (sh, w), 912 (sh, w), 1003 (sh, w), 1031 (sh, w), 1078 (sh, w), 1156 (sh, w), 1180 (sh, w), 1245 (sh, w), 1288 (sh, w), 1330 (sh, w), 1450 (sh, m), 1495 (sh, w), 1581 (sh, w), 1599 (sh, w), 1695 (sh, s), 1816 (sh, w), 1954 (sh, w), 2738 (sh, w), 3026 (sh, w), 3060 (sh, w), 3163 (sh, w). Elemental analysis calculated (%) for C34H29NO: C, 87.33; H, 6.25; N, 3.00. Found: C, 87.02; H, 6.30; N, 3.00.

(2,6-Dibenzhydryl-4-methylphenyl)isocyanide, Dipp*NC 69

Method A: In air, a solution of NaOH (19.97 g, 499.3 mmol) in water 3 (40 cm ) was slowly added to a solution of Dipp*NH2 (6.594 g, 15.00 mmol) and benzyltriethylammonium chloride (301 mg, 1.32 mmol) 3 3 in dichloromethane (30 cm ) and CHCl3 (6 cm , 75 mmol). The mixture was heated at reflux for 2 hours, cooled to room temperature and then transferred to a separatory funnel. The reaction vessel was washed out with further dichloromethane (10 cm3). The phases were separated and the aqueous phase extracted with further dichloromethane (3 x 20 cm3). The combined organic phases were washed with water (100 cm3), sat. 3 NaCl(aq.) (2 x 100 cm ), dried over K2CO3, filtered and volatiles removed under reduced 191 pressure. Purification by flash column chromatography (toluene:hexane, 45:55, 100:1

SiO2:Compound) afforded Dipp*NC as an off-white powder. Yield (1.58 g, 3.51 mmol, 23%).

Method B: Anhydrous diisopropylamine (14.8 cm3, 107 mmol) was added to a solution of Dipp*NH(CHO) (7.00 g, 15.0 mmol) in anhydrous dichloromethane (ca. 75 cm3) and 3 the solution then cooled in an ice bath. Freshly distilled POCl3 (3.50 cm , 38.0 mmol) was added dropwise over the period of five mins via syringe. A white suspension rapidly formed. The suspension was stirred at room temperature for 3 days. Saturated 3 Na2CO3(aq.) (100 cm ) was added and the mixture stirred for 20 mins. The mixture was transferred to a separatory funnel, the two phases separated and the aqueous phase extracted with further dichloromethane (2 x 50 cm3). The combined organic phases 3 were washed with saturated Na2CO3(aq.) (100 cm ), dried over Na2SO4, filtered and volatiles removed under reduced pressure. The crude product was suspended in nhexane (100 cm3) and stirred for 30 mins. The suspension was filtered and the precipitate washed with further nhexane (2 x 50 cm3). The precipitate was dried in vacuo to give the title compound as a white powder. (5.54 g, 12.3 mmol, 82%).

1 m.p. 192-193 °C H NMR (300.30 MHz, C6D6) δ 1.67 (s, 3H, p-CH3), 6.04 (s, 2H, 13 CHPh2), 6.82 (s, 2H, m-ArH), 6.82-7.11 (m, 20H, ArH). C NMR (75.52 MHz, C6D6) δ

20.94 (p-CH3), 53.14 (CHPh2), 126.68 (CHAr), 128.45 (CHAr), 129.09 (CHAr), 129.59

(CHAr), 138.59 (CAr), 141.22 (CAr), 141.83 (CAr), 173.36 (ArNC), one carbon atom is not observed. IR (Nujol) ν/cm-1: 666 (sh, w), 698 (sh, s), 733 (sh, m), 753 (sh, m), 1031 (sh, m), 1077 (sh, m), 1492 (sh, m), 1598 (sh, m), 2109 (sh, s), 3025 (sh, w), 3056 (sh, w). Elemental analysis calculated (%) for C34H27N: C, 90.83; H, 6.05; N, 3.12. Found: C, 90.82; H, 6.05; N, 3.13.

1,3-Bis(2,6-diisopropylphenyl)triazenidosodium, [NaN3Dipp2] 75

Method A: A solution of Dipp2N3H (17.54 g, 47.98 mmol) and NaOtBu (4.62 mg, 48.1 mmol) in tetrahydrofuran (ca. 150 cm3) was stirred for 18 hours at room temperature. The resulting cloudy solution was filtered and volatiles removed in vacuo. Coordinated tetrahydrofuran was removed by drying in vacuo at room temperature for 6 hours. 192 nPentane (ca. 50 cm3) was added to the resulting colourless solid and the resulting suspension was stirred for 30 mins. The suspension was filtered, the solid washed with n 3 further pentane (2 x 30 cm ) and then dried in vacuo for 24 h to afford [NaN3Dipp2] as a pale yellow-brown powder. Yield (14.3 g, 37.0 mmol, 77%).

Method B: A solution of NaOtBu (482 mg, 5.02 mmol) in toluene (ca. 50 cm3) was 3 added to a solution of Dipp2N3H (1.83 g, 5.00 mmol) also in toluene (ca. 15 cm ). The resulting thick suspension was stirred for 2 hours and then filtered. The precipitate was washed with nhexane (3 x 10 cm3) and then dried in vacuo at room temperature for overnight to afford [NaN3Dipp2] as a colourless powder. Yield (1.82 g, 4.71 mmol, 94%).

1 m.p. >250 °C. H NMR (300.30 MHz, THF-d8) δ 1.10 (d, J = 6.9 Hz, 24H, CH(CH3)2),

3.55 (sept, J = 6.9 Hz, 4H, CH(CH3)2 ), 6.76-6.81 (m, 2H, p-CH), 6.93-6.95 (m, 4H, m- 13 CH). C NMR (75.5 MHz, THF-d8) δ 24.30 CH(CH3)2, 27.87 (CH(CH3)2, 122.09 -1 (CHAr), 122.45 (CHAr), 142.44 (CAr), 151.75 (CAr). IR (Nujol) ν/cm ; 666 (sh, w), 723 (sh, w), 757 (sh, m), 777 (sh, m), 797 (sh, w), 934 (sh, w), 1057 (sh, w), 1096 (sh, w), 1186 (sh, w), 1232 (s, w), 1294 (br, s), 1336 (s, w), 1588 (sh, w). Elemental analysis calculated (%) for C24H34N3Na: C, 74.38; H, 8.84; N, 10.84. Found: C, 70.20; H, 8.87; N, 10.03.

1,3-Bis(2,6-diisopropylphenyl)triazenidotris(tetrahydrofuran)sodium,

[Na(N3Dipp2)(thf)3] 77

t A solution of Dipp2N3H (2.92 g, 8.00 mmol) and NaO Bu (774 mg, 8.05 mmol) in tetrahydrofuran (ca. 50 cm3) was stirred for 18 hours at room temperature. The resulting cloudy solution was filtered and concentrated in vacuo to the point of incipient crystallisation (ca. 10 cm3) and then slowly cooled over several hours to -25 °C. Storage at this temperature for 48 h afforded a crop of yellow orange blocks, which were isolated through decantation of the supernatant. Further concentration of the supernatant (ca. 5 cm3) and storage at -25 °C did not afford a second crop. Yield (3.17 g, 5.25 mmol, 66%). m.p. 1 200-210 °C (dec.). H NMR (300.30 MHz, C6D6) δ 1.33 (d, J = 6.9 Hz, 24H,

CH(CH3)2), 1.38 (s br, 12H, OCH2CH2), 3.46 (s br, 12H, OCH2CH2), 3.76 (sept, J = 6.8 193

13 Hz, 4H, CH(CH3)2), 7.10-7.15 (m, 2H, p-ArCH), 7.24-7.26 (m, 4H, m-ArCH). C

NMR (75.5 MHz, C6D6) δ 20.43 CH(CH3)2, 25.26 (OCH2CH2), 27.59 (CH(CH3)2),

67.62 (OCH2CH2), 122.40 (CHAr), 122.45 (CHAr), 142.08 (CAr), 150.74 (CAr). IR (Nujol) ν/cm-1; 723 (sh, w), 757 (sh, m), 777 (sh, m), 797 (sh, w), 916 (sh, w), 1055 (sh, w), 1074 (sh, w), 1097 (sh, w), 1187 (s, w), 1232 (sh, m), 1296 (br, s), 1357 (sh, w), 1435 (sh, w), 1367 (sh, m), 1376 (sh, m), 1586 (sh, s), 1694 (sh, w), 1784 (sh, w), 1838 (sh, w), 1892 (sh, w), 2472 (br, w), 2720 (br, w). Elemental analysis calculated (%) for

C36H58N3NaO3: C, 71.60; H, 9.68; N, 6.96. Found: C, 71.77; H, 9.89; N, 9.60.

1,3-Bis(2,6-diisopropylphenyl)triazenidobis(1,2-dimethoxyethane)sodium,

[Na(N3Dipp2)(dme)2] 79

t A solution of Dipp2N3H (2.91 g, 7.96 mmol) and NaO Bu (770 mg, 8.01 mmol) in 1,2-dimethoxyethane (ca. 30 cm3) was stirred for 18 hours at room temperature. The resulting cloudy solution was filtered and concentrated in vacuo to the point of incipient crystallisation (ca. 10 cm3) and slowly cooled over two hours to -25 °C. Storage at this temperature overnight afforded a crop of yellow blocks, which were isolated through decantation of the supernatant. Further concentration of the supernatant to the point of incipient crystallisation (ca. 3-4 cm3) afforded a second crop of yellow-orange blocks of equivalent purity. Yield (3.56 g, 6.28 mmol, 79%). m.p. 146-148 °C (with dec.). 1H

NMR (300.30 MHz, C6D6) δ 1.33 (d, J = 6.8 Hz, 24H, CH(CH3)2), 2.97 (s, 8H,

CH3OCH2), 2.99 (s, 12H, CH3OCH2), 3.76 (sept, J = 6.7 Hz, 4H, CH(CH3)2), 7.05-7.16 13 (m, 2H, p-ArCH), 7.18-7.30 (m, 4H, m-ArCH). C NMR (75.5 MHz, C6D6) δ 24.67

CH((CH3)2), 27.91 (CH(CH3)2), 58.75 (CH3OCH2), 71.08 (CH3OCH2), 122.58 (CHAr), -1 122.83 (CHAr), 142.33 (CAr), 151.13 (CAr). IR (Nujol) ν/cm ; 724 (sh, w), 553 (sh, s), 765 (sh, s), 797 (sh, w), 843 (sh, w), 863 (sh, s), 937 (sh, w), 1031 (sh, m), 1090 (br, s), 1130 (sh, s), 1225 (br, s), 1272 (br, s), 1357 (sh, m), 1367 (sh, m), 1376 (sh, m), 1586 (sh, s), 1694 (sh, w), 1784 (sh, w), 1838 (sh, w), 1892 (sh, w), 2472 (br, w), 2720 (br, w). Elemental analysis calculated (%) for C32H54N3NaO4: C, 67.69; H, 9.59; N, 7.40. Found: C, 67.51; H, 9.70; N, 7.54.

194

1,3-Bis(2,6-diisopropylphenyl)triazenidopotassium, [KN3Dipp2] 76

Method A: A solution of Dipp2N3H (1.86 g, 5.08 mmol) and KOtBu (571 mg, 5.09 mmol) in tetrahydrofuran (ca. 20 cm3) was stirred for 18 hours at room temperature. The resulting cloudy solution was filtered and volatiles removed in vacuo. Coordinated tetrahydrofuran was removed by drying in vacuo at room temperature for 18 hours. nPentane (ca. 30 cm3) was added to the resulting colourless solid and the resulting suspension was stirred for 30 mins. The suspension was filtered, the solid washed with n 3 further pentane (2 x 5 cm ) and then dried in vacuo for 4 h to afford [KN3Dipp2] as a colourless powder. Yield (1.79 g, 4.43 mmol, 87%).

Method B: A solution of KOtBu (567 mg, 5.05 mmol) in toluene (ca. 80 cm3) was 3 added to a solution of Dipp2N3H (1.831 g, 5.009 mmol) also in toluene (ca. 15 cm ). A suspension gradually formed over 30 mins. The resulting suspension was stirred for 2 hours and then filtered. The precipitate was washed with nhexane (2 x 10 cm3) and then dried in vacuo at room temperature overnight to afford [KN3Dipp2] as a pale pink, fine powder. Yield (1.821 g, 4.511 mmol, 90%).

1 m.p. >250 °C. H NMR (300.30 MHz, THF-d8) δ 1.09 (d, J = 6.9 Hz, 24H, CH(CH3)2),

3.54 (sept, J = 6.9 Hz, 4H, CH(CH3)2), 6.70-6.75 (m, 2H, p-CH), 6.90-6.92 (m, 4H, m- 13 CH). C NMR (75.5 MHz, THF-d8) δ 24.64 (CH(CH3)2), 28.27 (CH(CH3)2), 121.50 -1 (CHAr), 122.53 (CHAr), 142.08 (CAr), 153.07 (CAr). IR (Nujol) ν/cm ; 727 (sh, w), 754 (sh, m), 781 (sh, m), 801 (sh, m) 934 (sh, s), 954 (sh, w), 969 (sh, w), 1057 (sh, w), 1097 (sh, m), 1159 (sh, w), 1186 (sh, w), 1235 (sh, m), 1293 (br, s), 1319 (sh, w), 1358 (sh, m). Elemental analysis calculated (%) for C24H34N3K: C, 71.41; H, 8.49; N, 10.41. Found: C, 69.69; H, 8.92; N, 10.09.

195

1,3-Bis(2,6-diisopropylphenyl)triazenidotetrahydrofuranpotassium, [K(N3Dipp2)(thf)0.3] 78

t A solution of Dipp2N3H (2.92 g, 7.99 mmol) and KO Bu (899 mg, 8.01 mmol) in tetrahydrofuran (ca. 50 cm3) was stirred for 18 hours at room temperature. The resulting cloudy solution was filtered and concentrated in vacuo to the point of incipient crystallisation (ca. 8 cm3) and slowly cooled over two hours to -25 °C. Storage at this temperature for 2 hours afforded a crop of colourless plates, which were isolated through decantation of the supernatant and briefly (ca. 10 s ) dried in vacuo. Storage of the supernatant at -25 °C overnight afforded a second crop of colourless plates, which were again isolated through decantation of the supernatant and dried as above. The supernatant was concentrated in vacuo to ca. 3 cm3 and stored at -25 °C overnight to afford a third and final crop of colourless plates, which were isolated and dried as above. Yield (2.67 g, 5.61 mmol, 70%). m.p. 240-250 °C 1 (dec.). H NMR (300.30 MHz, C6D6) δ 1.22 (d, J = 6.9 Hz, 24H, CH(CH3)2), 1.40 (m,

OCH2CH2), 3.38 (s br, 4H, CH(CH3)2 ), 3.56 (m, OCH2CH2), 6.86-6.92 (m, 2H, p-CH), 1 7.04-7.07 (m, 4H, m-CH). H NMR (300.30 MHz, THF-d8) δ 1.10 (d, J = 6.9 Hz, 24H,

CH(CH3)2), 1.76 (m, 1.3H, OCH2CH2), 3.54 (sept, J = 6.9 Hz, 4H, CH(CH3)2), 3.62 (m, 13 1.3H, OCH2CH2), 6.71-6.76 (m, 2H, p-CH), 6.90-6.93. C NMR (75.5 MHz, C6D6) δ

24.65 (CH(CH3)2, 26.37 (OCH2CH2), 28.27 (CH(CH3)2, 68.22 (OCH2CH2), 121.50 -1 (CHAr), 122.53 (CHAr), 142.09 (CAr), 153.07 (CAr). IR (Nujol) ν/cm ; 646 (sh, w), 666 (sh, w), 728 (sh, w), 757 (sh, s), 779 (sh, m), 804 (sh, s), 915 (br, w), 933 (sh, w), 952 (s, w), 1038 (sh, w), 1059 (sh, m), 1095 (sh, w), 1107 (sh, w), 1157 (sh, w), 1185 (sh, w), 1235 (sh, w), 1257 (sh, w), 1293 (sh, m), 1323 (sh, w), 1357 (sh, m), 1430 (sh, m), 1585

(sh, w). Elemental analysis calculated (%) for C28H42N3KO: C, 70.69; H, 8.90; N, 8.83. Found: C, 69.95; H, 9.39; N, 9.10.

196

1,3-Bis(2,6-diisopropylphenyl)triazenidobis(1,2-dimethoxyethane)potassium,

[K(N3Dipp2)(dme)2] 80

t A solution of Dipp2N3H (2.92 g, 7.99 mmol) and KO Bu (990 mg, 8.02 mmol) in 1,2-dimethoxyethane was stirred for 18 hours at room temperature. The resulting cloudy solution was filtered and concentrated in vacuo to the point of incipient crystallisation (ca. 5 cm3) and slowly cooled over two hours to 5 °C. Storage at this temperature overnight afforded a crop of yellow-orange blocks, which were isolated through decantation of the supernatant. Further concentration of the supernatant (ca. 5 cm3) and storage at -25 °C did afford a second crop, which could not be isolated due to the highly viscous nature of the supernatant. Yield (1.30 g, 2.22 mmol, 28%). m.p. 230-240 °C (dec.). 1H NMR

(300.30 MHz, C6D6) δ 1.35 (d, J = 6.9 Hz, 24H, CH(CH3)2), 3.06 (s, CH3OCH2), 3.25

(s, CH3OCH2), 3.75 (m br, 4H, CH(CH3)2 ), 7.11-7.14 (m, 2H, p-CH), 7.26-7.29 (m, 1 4H, m-CH). H NMR (300.30 MHz, THF-d8) δ 1.10 (d, J = 6.9 Hz, 24H, CH(CH3)2),

3.27 (s, 12H, CH3OCH2), 3.43 (s, 8H CH3OCH2), 3.90 (sept, J = 6.9 Hz, 4H, CH(CH3)2 13 ), 6.71-6.76 (m, 2H, p-CH), 6.90-6.93 (m, 4H, m-CH). C NMR (75.5 MHz, THF-d8) δ

24.65 CH(CH3)2, 28.26 (CH(CH3)2, 58.90 (CH3OCH2), 72.67 (CH3OCH2), 121.49 -1 (CHAr), 122.52 (CHAr), 142.08 (CAr), 153.09 (CAr). IR (Nujol) ν/cm ; 723 (sh, w), 754 (sh, w), 762 (sh, w), 779 (sh, w), 803 (sh, w), 858 (sh, w), 936 (sh, w), 1032 (br, w), 1095 (sh, m), 1159 (sh, w), 1187 (sh, w), 1225 (br, m), 1269 (sh, m), 1293 (sh, m), 1358 (sh, w), 1586 (sh, s), 1694 (sh, w), 1784 (sh, w), 1838 (sh, w), 1892 (sh, w), 2472 (br, w), 2720 (br, w). Elemental analysis calculated (%) for C32H54N3KO4: C, 65.83; H, 9.32; N, 7.20. Found: C, 62.18; H, 9.10; N, 9.18.

1,3-Bis(2,6-dibenzhydryl-4-methylphenyl)triazenidopotassium toluene solvate,

[KN3Dipp*2]∙3PhMe 81

Toluene (ca. 80 cm3) was added to a Schlenk flask charged t with Dipp*2N3H (3.56 g, 4.00 mmol) and KO Bu (450 mg, 4.01 mmol) and the suspension was sonicated at room temperature for 15 mins to dissolve all the suspended solids. The resulting orange solution was stirred overnight at room temperature. The resulting suspension

197 was warmed to 70 °C to re-dissolve the precipitated orange solid and slowly cooled overnight to afford a crop of microcrystalline product. The precipitate was isolated through filtration and then dried in vacuo for 30 mins. Storage of the supernatant at -25 °C for 48 hours afforded further product as orange rods, suitable for a single crystal X- ray structural determination. The crystals were isolated through filtration and dried in vacuo for 30 mins. The supernatant was concentrated in vacuo to the point of incipient crystallisation. Storage overnight at -25 °C afforded a final crop of microcrystalline orange rods, which were isolated and dried as above. Yield (3.389 g, 3.23 mmol, 81%). 1 m.p. 205-215 °C (dec.). H NMR (300.30 MHz, C6D6) δ 1.96 (s, 6H, p-CH3), 2.10 (s

4H, p-CH3 Tol), 6.46 (s, 4H, CHPh2), 6.84-6.89 (m, 13H, ArH), 6.95-7.02 (m, 22H, 13 ArH), 7.12-7.21 (m, 29H, ArH). C NMR (75.52 MHz, C6D6) δ 21.40 (p-CH3), 52.37

(CHPh2), 125.59 (CHAr), 125.70 (CAr), 128.13 (CHAr), 128.35 (CHAr), 128.57 (CAr),

129.34 (CHAr), 129.64 (CHAr), 130.71 (CHAr), 135.92 (CAr), 136.47 (CAr), 146.45 (CAr). IR (Nujol) ν/cm-1; 609 (sh, w), 701 (sh, m), 731 (sh, w), 746 (sh, w), 764 (sh, w), 771 (sh, w), 802 (br, w), 862 (sh, w), 881 (sh, w), 1031 (sh, w), 1076 (sh, w), 1178 (sh, w), 1209 (sh, m), 1237 (sh, m), 1250 (sh, m), 1433 (sh, w), 1450 (sh, w), 1493 (sh, w), 1597

(sh, w), 3022 (sh, w), 3058 (sh, w). Elemental analysis calculated (%) for C66H64N3K: C, 85.76; H, 6.04; N, 4.25. Found: C, 84.10; H, 5.43; N, 4.08.

Attempted reaction of [LiN3Dipp2(thf)n] with [NdI3(thf)3.5]

A solution of nBuLi (1.10 cm3, 1.54 mmol, 1.40 molL-1) in hexanes was added to a 3 solution of Dipp2N3H (548 mg, 1.50 mmol) in tetrahydrofuran (ca. 10 cm ) and stirred at room temperature for 2 hours. This was then added to a suspension of [NdI3(thf)3.5] (388 mg, 499 μmol) in tetrahydrofuran (ca. 10 cm3). The resulting solution was stirred overnight and volatiles removed in vacuo. nPentane (20 cm3) was added and the resulting suspension filtered. The precipitate was washed with further npentane (4 x 20cm3) and the filtrate and washings concentrated in vacuo to ca. 10 cm3. Storage overnight afforded a crop of yellow plates of [Li2(μ-I)(μ-thf)(μ-N3Dipp2)(thf)2] (73 mg, 0.11 mmol). The original precipitate was washed with toluene (3 x 20 cm3) and the 3 filtrate and washing concentrated in vacuo to ca. 10 cm . Storage overnight at room temperature gave several single crystals of [NdI2(N3Dipp2)(thf)] and large amounts of amorphous solids. 198

Attempted reaction of [TlN3Dipp2] with CeCl3

3 A solution of [TlN3Dipp2] (0.85 g, 1.5 mmol) in tetrahydrofuran (ca. 15 cm ) was added 3 to a suspension of CeCl3 (0.12 g, 0.49 mmol) also in tetrahydrofuran (ca. 15 cm ). The resulting colourless suspension was stirred at room temperature overnight, filtered twice to remove a fine precipitate and volatiles removed in vacuo. The residue was redissolved in toluene (ca. 10 cm3) and concentrated to the point of incipient crystallisation in vacuo. The crystals once isolated exhibited extreme light and temperature sensitivity that precluded further characterisation other than a single crystal

X-ray diffraction study, which identified a product as [{Tl2(μ-OH)(μ-N3Dipp2)}2]. Further characterisation was not possible.

Bis(1,3-bis(2,6-diisopropylphenyl)triazenido)iodotetrahydrofuransamarium(III),

[SmI(N3Dipp2)2(thf)] 154

A solution of [KN3Dipp2] (0.80 g, 2.0 mmol) in tetrahydrofuran (15 3 cm ) was added to a suspension of SmI3(thf)3.5 (0.78 g, 1.0 mmol) also in tetrahydrofuran. The suspension was stirred for 18 hours before volatiles were removed in vacuo. The residue was suspended in toluene (ca. 30 cm3) and sonicated for 1.5 hours at 50 °C. The suspension was allowed to settle, filtered and the precipitate washed with toluene (2 x 5 cm3). The filtrate and washings were concentrated to the point of incipient crystallisation and stored at 5 °C overnight to afford a crop of yellow plates that were isolated through decantation of the supernatant and then dried in vacuo to afford the title compound as a pale yellow powder. Yield 1 (0.43 g, 0.40 mmol, 40%). m.p. 101-105 °C (dec.). H NMR (300.30 MHz, C6D6). δ 0.78 (s br, 24H, fwhm = 18 Hz), 1.19 (d, J = 6.8 Hz, 24H), 1.57 (s br, 4H fwhm = 60 Hz), 3.15 (s br, 4H, fwhm = 85 Hz), 3.33 (m, 4H), 4.67 (s br, 2H, fwhm = 105 Hz), 7.12-7.15 (m, 4H), 7.43-7.54 (m, 6H). IR (Nujol) ν/cm-1; 666 (sh, w), 723 (sh, w), 755 (sh, m), 777 (sh, w), 799 (sh, m), 835 (sh, w), 935 (sh, w), 1056 (sh, w), 1098 (br, w), 1200 (sh, w), 1256 (br, s), 1294 (s, w), 1330 (sh, w), 1362 (sh, w). Elemental analysis calculated (%) for C52H76N6IOSm: C, 57.91; H, 7.10; N, 7.79. Found: C, 56.87; H, 7.25; N, 7.37. 199

Tris(bis(2,6-diisopropylphenyl)triazenido)cerium(III), [Ce(N3Dipp2)3] 164

A solution of Dipp2N3H (3.289 g, 8.997 mmol) and [Ce(N(SiMe3)2)3] (1.863 g, 2.999 mmol) in nhexane (ca. 50 cm3) was stirred for 3 days at room temperature. The cloudy solution was filtered and the supernatant was concentrated to the point of incipient crystallisation in vacuo (ca. 10 cm3). Storage at room temperature overnight afforded a crop of crystalline material. The crystals were isolated through decantation of the supernatant, washed with an equal volume of sub-zero npentane and then dried in vacuo. Further crops were isolated by concentrating the filtrate and washings to the point of incipient crystallisation and repeating the above isolation and washing procedure. The title compound was isolated as microcrystalline yellow cubes over 4 crops. Yield (3.086 g, 1 2.501 mmol, 83%). m.p. 113-117 °C (dec.). H NMR (300.30 MHz, C6D6). δ -11.32, 1.18, 1.20 3.33. IR (Nujol) ν/cm-1; 658 (sh, w), 677 (sh, w), 724 (sh, w), 754 (sh, w), 771 (sh, w), 799 (sh, w), 814 (sh, w), 828 (sh, w), 863 (sh, w), 898 (sh, w), 935 (sh, w), 983 (br, m), 1044 (sh, w), 1054 (sh, w), 1097 (sh, w), 1162 (sh, w), 1179 (sh, w), 1206 (sh, w), 1245 (br, m), 1319 (sh, w), 1362 (sh, m), 1584 (sh, w), 1620 (sh, w), 1861 (sh, w),

1928 (sh, w). Elemental analysis calculated (%) for C72H102N9Ce: C, 70.09; H, 8.33; N, 10.22. Results were repeatedly inconsistent with calculated values due to sample decomposition pre-analysis.

Tetrakis((2,6-diisopropylphenyl)triazenido)bis(hydroxo)dicerium(III),

[{Ce(N3Dipp2)2(μ-OH)}2] 171 and tetrakis((2,6-diisopropylphenyl)triazenido)oxo dicerium(III), [{Ce(N3Dipp2)2}2(μ-OH)] 172

A solution of Dipp2N3H (3.286 g, 8.989 mmol) and [Ce(N(SiMe3)2)3] (1.861 g, 2.996 mmol) in nhexane (ca. 50 cm3) was stirred for 3 days at room temperature. The cloudy solution was filtered and the supernatant was concentrated to the point of incipient crystallisation in vacuo (ca. 10 cm3). Storage at room temperature overnight afforded a crop of yellow crystalline material. Some of these crystals were removed for single crystal X-ray structure determination of [Ce(N3Dipp2)3] without decantation of the supernatant. Overnight the solution was observed to darken in colour. The yellow cubes 200 were observed to dissolve over the period of several days with concomitant deposition of large orange blocks. The crystals were isolated through decantation of the supernatant, washed with an equal volume of sub-zero npentane and dried in vacuo. Yield (1.844 g, 1.045 mmol, 70%). m.p. 209-212 °C (dec.). IR (Nujol) ν/cm-1; 681 (sh, w), 723 (sh, w), 753 (sh, s), 766 (sh, s), 796 (sh, s), 835 (sh, m), 884 (sh, w), 899 (sh, w), 921 (sh, w), 935 (sh, m), 954 (sh, w), 964 (sh, w), 1045 (sh, w), 1056 (sh, m), 1095 (sh, m), 1104 (sh, w), 1118 (sh, w), 1146 (sh, w), 1161 (sh, w), 1179 (sh, w), 1209 (sh, w), 1266 (br, s), 1311 (sh, w), 1336 (sh, w), 1381 (sh, s), 1432 (sh, s), 1582 (sh, w), 1856 (sh, w), 1920 (sh, w), 2455 (br, w), 2725 (sh, w), 2754 (sh, w), 3020 (br, w), 3064 (sh, s), 3162 (br, w). Elemental analysis calculated (%) for C96H138N12Ce2O2: C, 65.05; H, 7.85; N, 9.48. Results were repeatedly inconsistent with calculated values due to sample decomposition pre-analysis.

Tris(bis(2,6-diisopropylphenyl)triazenido)neodymium(III), [Nd(N3Dipp2)3] 156

Method A: A solution of Dipp2N3H (3.291 g, 9.003 mmol) and n 3 [Nd(N(SiMe3)2)3] (1.877 g, 3.001 mmol) in hexane (ca. 50 cm ) was stirred for 3 days at room temperature. The cloudy solution was filtered and the supernatant was concentrated to the point of incipient crystallisation in vacuo (ca. 10 cm3). Storage at room temperature overnight afforded a crop of crystalline material. The crystals were isolated through decantation of the supernatant, washed with an equal volume of sub-zero npentane and then dried in vacuo. Further crops were isolated by concentrating the filtrate and washings to the point of incipient crystallisation and repeating the above isolation and washing procedure. The title compound was isolated as size dependant dichroic (yellow-green to red-orange with increasing crystal size) cubes over 5 crops. Yield (2.054 g, 1.659 mmol, 55%).

Method B: A solution of [KN3Dipp2] (1.63 g, 4.04 mmol) in tetrahydrofuran (ca. 30 3 cm ) was added to a suspension of [NdI3(thf)3.5] (1.56 g, 2.01 mmol) also in tetrahydrofuran and stirred overnight. Volatiles were removed in vacuo and the residue

201 suspended in toluene (ca. 30 cm3) and sonicated at 50 °C for 1.5 h. The suspension was filtered and volatiles were removed in vacuo. The residue was redissolved in nhexane (ca. 25 cm3), the suspension filtered and concentrated in vacuo to ca. 5 cm3. Storage at 5 °C afforded a crop of yellow needles, which were isolated through decantation of the supernatant and dried in vacuo. A second crop was isolated by precipitating the product through the addition of npentane (ca. 10 cm3). The resulting suspension was allowed to settle, filtered and the solid dried in vacuo. The supernatant was concentrated in vacuo to ca. 1-2 cm3. nPentane (ca. 5 cm3) was added, the suspension allowed to settle, filtered and the precipitate dried in vacuo. Yield over all three crops as a pale yellow powder (1.00 g, 807 μmol, 60%).

1 m.p. 109-121 °C (dec.). H NMR (300.30 MHz, C6D6). δ -13.24, -3.64, 1.18, 1.20, 2.80, 3.33, 5.59, 6.23, 8.77, 10.16. IR (Nujol) ν/cm-1; 721 (sh, w), 752 (sh, w), 767 (sh, w), 798 (sh, w), 835 (sh, w), 935 (sh, w), 1057 (sh, w), 1097 (sh, w), 1190 (sh, w), 1257 (br, m), 1332 (br, w), 1362 (br, m), 1518 (sh, w), 1054 (sh, w), 1097 (sh, w), 1162 (sh, w), 1179 (sh, w), 1206 (sh, w), 1245 (br, m), 1319 (sh, w), 1362 (sh, m), 1584 (sh, w), 1620

(sh, w), 1861 (sh, w),1928 (sh, w). Elemental analysis calculated (%) for C72H102N9Nd: C, 69.86; H, 8.31; N, 10.18. Found: C, 69.53; H, 8.53; N, 9.99.

Tris(bis(2,6-diisopropylphenyl)triazenido)samarium(III), [Sm(N3Dipp2)3] 160

Method A: A solution of Dipp2N3H (3.290 g, 9.000 mmol) and n 3 [Sm(N(SiMe3)2)3] (1.895 g, 3.001 mmol) in hexane (ca. 50 cm ) was stirred for 3 days at room temperature. The cloudy solution was filtered and the supernatant was concentrated to the point of incipient crystallisation in vacuo (ca. 10 cm3). Storage at room temperature overnight afforded a crop of crystalline material. The crystals were isolated through decantation of the supernatant, washed with an equal volume of sub-zero npentane and then dried in vacuo. Further crops were isolated by concentrating the filtrate and washings to the point of incipient crystallisation and repeating the above isolation and washing procedure. The title compound was isolated as yellow-orange microcrystalline cubes over 3 crops. Yield (2.607 g, 2.096 mmol, 70%).

202

Method B: A suspension of freshly filed samarium powder (0.14 g, 0.93 mmol) and 3 [TlN3Dipp2] (1.59 g, 2.80 mmol) in tetrahydrofuran (ca. 15 cm ) was sonicated for 2 hours at 50 °C. The transformation of the metal powder to large metal particles was observed over this time. The suspension was filtered and the metal washed with tetrahydrofuran (ca. 2 x 5 cm3). Further samarium powder (0.06 g, 0.40 mmol) was then added to the supernatant and the suspension stirred overnight. The metal was again observed to aggregate into large particles. The filtration and washing was repeated as above and further samarium powder (0.10 g, 0.66 mmol) added. The suspension was stirred again overnight at room temperature. The suspension was re-filtered and further samarium (0.41 g, 2.7 mmol) added. The suspension was again stirred overnight to afford a yellow solution. Filtration followed by removal of volatiles in vacuo afforded the title compound as a yellow-orange powder (997 mg, 801 μmol, 86%).

1 m.p. 106-121 °C (dec.). H NMR (300.30 MHz, C6D6). δ -1.30, -0.1, 1.01, 1.08, 1.10, 1.12, 1.15, 1.18, 1.20, 1.52, 1.61, 3.33, 4.87, 6.65, 7.04, 7.06, 7.09, 7.15, 7.18, 7.22. IR (Nujol) ν/cm-1; 665 (br, w), 724 (sh, w), 754 (sh, m), 770 (sh, m), 799 (sh, m), 829 (sh, w), 845 (sh, w), 863 (sh, w), 935 (sh, w), 970 (sh, m), 1044 (sh, w), 1056 (sh, w), 1097 (sh, w), 1246 (br, s), 1361 (sh, w), 1516 (sh, w), 1586 (sh, w), 1620 (sh, w). Elemental analysis calculated (%) for C72H102N9Sm: C, 69.52; H, 8.26; N, 10.13. Found: C, 69.19; H, 8.50; N, 10.18.

Tris(bis(2,6-diisopropylphenyl)triazenido)ytterbium(III), [Yb(N3Dipp2)3] 161

Method A: A solution of Dipp2N3H (3.262 g, 8.924 mmol) and n 3 [Yb(N(SiMe3)2)3] (1.947 g, 2.976 mmol) in hexane (ca. 50 cm ) was stirred for 4 days at 60 °C . The cloudy solution was filtered and the supernatant was concentrated to the point of incipient crystallisation in vacuo (ca. 5 cm3). Storage at -25 °C overnight afforded a small crop of pink-red cubes. The crystals were isolated through decantation of the supernatant and then dried in vacuo. The supernatant was evaporated in vacuo and the residue redissolved in npentane (ca. 10 cm3). Storage at -25 °C for one day afforded a further crop. Yield (20 mg, 16 μmol,

<1%). Removal of all volatiles in vacuo and drying of the red gum thus obtained in

203 vacuo did not yield a dry material suitable for bulk analysis. IR and 1H NMR spectroscopy were consistent with this material containing the title compound however.

Method B: A suspension of freshly filed ytterbium powder (0.50 g, 2.9 mmol) and 3 [TlN3Dipp2] (0.59 g, 1.0 mmol) in tetrahydrofuran (ca. 10 cm ) was stirred for four days at room temperature. The transformation of the majority of the metal powder to large metal particles was observed over this time. The suspension was filtered and the metal washed with tetrahydrofuran (ca. 2 x 5 cm3). Volatiles were removed in vacuo and the residue redissolved in nhexane (ca. 5 cm3). Storage at room temperature afforded a small crop of red-pink cubes (ca. 5 mg) in sufficient quantity for an X-ray single crystal structure determination, which were isolated by decantation of the supernatant. Removal of volatiles in vacuo, gave a greasy residue from which further tractable materials could not be isolated by recrystallisation. Yield (ca. 5 mg, ca. 4 μmol, <1%).

IR (Nujol) ν/cm-1; 627 (br, w), 682 (br, w), 707 (br, w), 727 (sh, w), 754 (sh, m), 770 (sh, m), 800 (sh, w), 838 (br, w), 868 (sh, w), 936 (br, m), 1045 (sh, w), 1056 (sh, w), 1099 (br, w), 1179 (sh, w), 1252 (br, s), 1383 (sh, m), 1596 (sh, w), 1620 (sh, w).

Attempted reduction of [Nd(N3Dipp2)3] with potassium graphite and 18-crown-6

A suspension of KC8 (665 mg, 4.92 mmol), 18-crown-6 (132 mg, 499 μmol) and

[Nd(N3Dipp2)3] (547 mg, 500 μmol) was stirred at -20 °C for 2 hours, during which time a colour change from pale-yellow green to orange-red was observed. The suspension was stored at -25 °C overnight and then filtered. Volatiles were removed in vacuo to afford only trace amounts of solid materials.

Attempted reaction between [Sm(N3Dipp2)3] and triphenylphosphine oxide

A solution of O=PPh3 (11 mg, 40 μmol) and [Sm(N3Dipp2)3] (113 mg, 90.8 μmol) in toluene (ca. 5 cm3) was stirred overnight at room temperature. Filtration and removal of volatiles in vacuo gave a mixture of starting materials as determined using 1H and 31P NMR spectroscopy of the filtrate and precipitate.

204

Bis(1,3-bis(2,6-diisopropylphenyl)triazenido)bis(tetrahydrofuran)ytterbium(II),

[Yb(N3Dipp2)2(thf)2] 179

A solution of [YbI2(thf)n] was prepared by the addition of iodine (510 mg, 2.01 mmol) to a suspension of freshly filed ytterbium metal (1.04 g, 6.00 mmol) in tetrahydrofuran (ca. 25 cm3) and stirred overnight. The suspension was filtered and the precipitate washed with tetrahydrofuran (2 x 5 3 3 cm ). A solution of [NaN3Dipp2] (1.55 g, 4.02 mmol) in tetrahydrofuran (ca. 30 cm ) was added to the yellow solution prepared above, which immediately turned blood red and stirred for 18 hours at room temperature. Volatiles were removed in vacuo, nhexane (ca. 25 cm3) added and the resulting suspension stirred for 2 hours at room temperature. The suspension was filtered, the precipitate washed with further nhexane (2 x 5 cm3) and the combined filtrate and washings concentrated in vacuo to ca. 10 cm3. Storage overnight at room temperature afforded a crop of dried blood red blocks, which were isolated through decantation of the supernatant and briefly (ca. 5 s) dried in vacuo. The supernatant was concentrated in vacuo and stored at room temperature to afford a second crop of blocks, which were isolated and dried as above. A final crop was isolated as above by concentrating the supernatant to ca. 1-2 cm3 and storing at -25 °C overnight. Yield over three crops (1.38 g, 1.32 mmol, 66%). m.p. 125-128 °C (dec.). 1H

NMR (400.13 MHz, C6D6) δ 1.16 (d, J = 6.9 Hz, 48H, CH(CH3)2), 1.36 (s br, 8H,

OCH2CH2), 3.46 (sept, J = 6.8 Hz, 8H, CH(CH3)2 ), 3.73 (s br, 8H, OCH2CH2), 7.11- 13 7.19 (m, 12H, ArH). C NMR (100.62 MHz, C6D6) δ 24.77 CH(CH3)2, 28.43 - (CH(CH3)2, 123.54 (CHAr), 124.62 (CHAr), 142.33 (CAr), 148.22 (CAr). IR (Nujol) ν/cm 1; 666 (sh, w), 724 (sh, w), 754 (sh, w), 767 (sh, s), 798 (sh, m), 865 (sh, w), 883 (sh, w), 936 (sh, w), 961 (br, w), 1057 (sh, w), 1095 (sh, w), 1179 (sh, w), 1208 (sh, w), 1247 (sh, w), 1268 (br, s), 1315 (sh, w), 1337 (sh, w), 1432 (sh, w), 1583 (sh, w), 2726 (sh, w). Elemental analysis calculated (%) for C56H84N6O2Yb: C, 64.28; H, 8.09; N, 8.03. Found: C, 66.58; H, 8.63; N, 9.55.

205

Bis(1,3-bis(2,6-diisopropylphenyl)triazenido)bis(tetrahydrofuran)samarium(II),

[Sm(N3Dipp2)2(thf)2] 182

A solution of [SmI2(thf)n] was prepared through the addition of iodine (509 mg, 2.01 mmol) to a suspension of freshly filed samarium metal (0.53 g, 3.5 mmol) in tetrahydrofuran (ca. 25 cm3) and stirred overnight. A solution of [NaN3Dipp2] (1.55 g, 4.00 mmol) in tetrahydrofuran (ca. 10 cm3) was added to the blue solution prepared above, which immediately turned emerald-green and stirred for 18 hours at room temperature. Volatiles were removed in vacuo to a volume of ca. 5 cm3, nhexane (ca. 30 cm3) was added and the resulting suspension stirred for 30 mins at room temperature. The suspension was filtered, the precipitate washed with further nhexane (2 x 5 cm3) and the supernatant concentrated in vacuo to ca. 25 cm3. Storage at room temperature for 3 hours gave a greasy precipitate, which was removed by filtration of the suspension. Storage of the supernatant overnight at room temperature afforded a crop of emerald green blocks, which were isolated through decantation of the supernatant and briefly (ca. 5 s) dried in vacuo. The supernatant was concentrated in vacuo to ca. 5 cm3 and stored at room temperature to afford a second crop of blocks, which were isolated and dried as above. Yield over two 1 crops (738 mg, 721 μmol, 36%). m.p. 120-124 °C (dec.). H NMR (300.30 MHz, C6D6)

δ 0.53 (s br, 8H, OCH2CH2 or OCH2CH2, fwhm = 36 Hz), 2.84 (s, 48H, CH(CH3)2),

3.35 (s br, 8H, OCH2CH2 or OCH2CH2, fwhm = 36 Hz), 5.05 (d, J = 7.8 Hz, 8H, 13 CH(CH3)2), 6.28 (m, 4H, ArH), 10.80 (s br, 8H, ArH, fwhm = 108 Hz). C NMR

(75.52 MHz, C6D6) δ 28.22 (CH(CH3)2), 30.36 (OCH2CH2 or OCH2CH2), 50.63

(OCH2CH2 or OCH2CH2), 119.13 (CH(CH3)2), 122.13 (CHAr), 128.34 (CHAr). No other peaks are observed. IR (Nujol) ν/cm-1; 707 (sh, w), 724 (sh, w), 754 (sh, m), 771 (sh, w), 800 (sh, w), 837 (br, w), 876 (br, w), 936 (sh, w), 1030 (sh, m), 1056 (sh, w), 1096 (sh, w), 1177 (sh, w), 1200 (sh, w), 1257 (br, s), 1294 (sh, w), 1332 (sh, w), 1360 (sh, w),

1437 (sh, w). Elemental analysis calculated (%) for C56H84N6O2Sm: C, 65.71; H, 8.27; N, 8.21. Found: C, 62.08; H, 6.76; N, 7.93.

206

Attempted reaction of [NaN3Dipp2] with [TmI2(thf)n]

I2 (506 mg, 1.99 mmol) was added to a suspension of freshly filed thulium metal (2.37 g, 14.0 mmol) in tetrahydrofuran (ca. 20 cm3) and the resulting suspension stirred overnight. The colourless suspension was sonicated at 50 °C whereupon the white precipitate was observed to dissolve to give a malachite coloured suspension. Further tetrahydrofuran (ca. 40 cm3) was added, the suspension filtered and the metal powder 3 washed with further tetrahydrofuran (2 x 5 cm ). A solution of [NaN3Dipp2] (1.53 g, 3.95 mmol) in tetrahydrofuran (ca. 15 cm3) was added and the solution colour was observed to change to yellow brown and stirred overnight. Volatiles were removed in vacuo and nhexane (ca. 40 cm3) added. The suspension was filtered and the precipitate washed with further nhexane (2 x 5 cm3). The filtrate and washings were concentrated in vacuo to ca. 25 cm3. Storage overnight afforded a crop of several single crystals of

[TmI(N3Dipp2)2(thf)2] as determined by a single crystal X-ray diffraction study and a large quantity of an amorphous white, henceforth insoluble precipitate.

General Procedure for attempted reactions of [Ce(N3Dipp2)3] with organic or inorganic oxidants

3 To a solution of [Ce(N3Dipp2)3] (1 eq.) in toluene (ca. 10cm ) was treated with the oxidant (0.5 or 1 eq.) at room temperature. The mixture was stirred at room temperature overnight with no colour change of the solution observed.

The following quantities were used according to the general procedure with similar results:

[Ce(N3Dipp2)3] (0.62 g, 0.50 mmol), hexachloroethane (59 mg, 0.25 mmol)

[Ce(N3Dipp2)3] (0.62 g, 0.50 mmol), 1,2-dibromoethane (47 mg, 0.25 mmol)

[Ce(N3Dipp2)3] (0.62 g, 0.50 mmol), 1,2-dibromoethane (0.21 g, 0.25 mmol)

207

t Attempted reaction of [Ce(O Bu)4(thf)2] with Dipp2N3H

n 3 A solution of Dipp2N3H (731 mg, 2.00 mmol) in hexane (ca. 10 cm ) was added to a t n 3 solution of [Ce(O Bu)4(thf)2] (0.580 g, 1.00 mmol) also in hexane (ca. 25 cm ). The solution was stirred overnight before volatiles were removed in vacuo. Analysis of the crude material using 1H NMR spectroscopy gave a material consistent with a mixture of starting materials.

Tris(tetrahydrofuran)lithium(μ-chloro)tris(bis(trimethylsilyl)amido)cerium(III),

[(thf)3Li(μ–Cl)Ce{N(SiMe3)2}3] 196

Tetrahydrofuran (ca. 20 cm3) was added to a Schlenk flask charged with Ce(OTf)3 (1.17 g, 1.99 mmol) and LiN(SiMe3)2 (1.34 g, 8.01 mmol) at 0 °C. The resulting solution was stirred at 55 °C for 36 hours. The yellow solution was cooled to room temperature and a solution of hexachloroethane (240 mg, 1.01 mmol) in tetrahydrofuran (ca. 10 cm3) was added and the resulting red solution stirred at room temperature for 3 days. Volatiles were removed in vacuo and the product extracted into npentane (2 x 25 cm3). Storage of this solution overnight gave a fine white suspension, which was filtered and the solution again stored overnight at room temperature during which time further precipitate developed. The suspension was again filtered and then concentrated in vacuo to ca. 10 cm3. Storage overnight at room temperature gave a crop of the title compound as pale yellow plates. The crystals were isolated through filtration, dried in vacuo and the supernatant concentrated in vacuo to ca. 5 cm3. Storage overnight at room temperature gave another crop of colourless plates, which were isolated and dried as above. Yield (500 mg, 568 μmol, 29%)%). m.p. 100-110 °C (dec. with some sublimation of 1 [Ce(N(SiMe3)2)3]). H NMR (300.30 MHz, C6D6). δ -3.14 (s br, 54H, SiCH3, fwhm =

80 Hz), 0.69 (s br, 12H, OCH2CH2 or OCH2CH2, fwhm = 21 Hz), 1.84 (s br, 12H, -1 OCH2CH2 or OCH2CH2, fwhm = 39 Hz). IR (Nujol) ν/cm ; 596 (sh, s), 664 (sh, s), 765 (sh, s), 834 (br, s), 865 (sh, s), 1001 (br, s), 1046 (sh, m), 1245 (sh, s). Elemental analysis calculated (%) for C30H78N3O3Si6ClLiCe: C, 40.95; H, 8.93; N, 4.78. Results were repeatedly inconsistent with calculated values due to sample decomposition pre- analysis.

208

General Procedure for Attempted Reactions of Dipp2N3H with [Ce(NCy2)4]

3 A solution of LiNCy2 (4 eq.) in tetrahydrofuran (ca. 40 cm ) was added at -40 °C to a 3 suspension of [CeCl3(thf)n] (1 eq.) also in tetrahydrofuran (ca. 10 cm ). The suspension was warmed to 0 °C over two hours. Volatiles were removed in vacuo below 0 °C. Toluene (ca. 40 cm3) was added and the suspension filtered and the precipitate washed with further toluene (ca. 2 x 5 cm3). A solution of hexachloroethane (0.5 eq.) in toluene (ca. 5 cm3) was added immediately yielding a midnight blue suspension. A solution of 3 Dipp2N3H (1 or 2 eq.) in toluene (ca. 10 cm ) was added and the suspension stirred overnight at room temperature, during which time a gradual colour change to pale yellow was observed. The suspension was then filtered and volatiles removed in vacuo 1 to afford a mixture of Dipp2N3H as identified by single crystal X-ray and H NMR spectroscopy and an as yet unidentified yellow oil.

The following quantities were used according to the general procedure with similar results:

CeCl3 (501 mg, 2.03 mmol), LiNCy2 (1.52 g, 8.12 mmol), hexachloroethane (241 mg,

1.02 mmol) and Dipp2N3H (1.484 g, 4.06 mmol).

CeCl3 (503 mg, 2.04 mmol), LiNCy2 (1.53 g, 8.17 mmol), hexachloroethane (243 mg,

1.02 mmol) and Dipp2N3H (746 g, 2.04 mmol).

(1,3-bis(2,6-diisopropylphenyl)triazenidolithium(μ-1,3-bis(2,6- diisopropylphenyl)triazenido)dicyclohexylamine (μ-lithium), [Li(N3Dipp2)(μ-

N3Dipp2)(μ-Li)(HNCy2)] 198

A solution of LiNCy2 (1.50 g, 8.00 mmol) in tetrahydrofuran (ca. 40 cm3) was added at -40 °C to a suspension of

[CeCl3(thf)n] (493 mg, 2.00 mmol) also in tetrahydrofuran (ca. 10 cm3). The yellow orange solution was slowly warmed to 0 °C over the period of 2 hours before volatiles were removed in vacuo. The residue was resuspended in toluene (ca. 40 cm3). A

209

3 solution of Dipp2N3H (732 mg, 2.00 mmol) also in toluene (ca. 10 cm ) was then added and the resulting suspension stirred overnight. Volatiles were then removed in vacuo, nhexane (ca. 50 cm3) was added and the resulting suspension filtered. Concentration of the supernatant in vacuo to ca. 10 cm3 and storage overnight at room temperature gave the title compound as colourless crystals, which were isolated through decantation of the supernatant and then dried in vacuo. The supernatant was further concentrated in vacuo to ca. 3 cm3 to afford a second crop of colourless crystals, which were isolated and dried as above. Yield (344 mg, 372 μmol, 37%). m.p. 85-95 °C (colour change to 1 bright red), 103-106 °C (dec.). H NMR (300.30 MHz, C6D6) δ 0.49-1.62 (m, 21H,

NCHCH2), 1.23 (d, J = 6.7 Hz, 48H, CH(CH3)2), 2.26 (s br, 1H, NCHCH2), 3.45 (s br,

8H, CH(CH3)2 ), 3.73 (s br, 8H, OCH2CH2), 7.16 (s, 12H, ArH), 9.30 (s br, 1H, NH). 13 C NMR (75.52 MHz, C6D6) δ 23.63 CH(CH3)2, 24.95 (CH2CH2CH2), 25.54

(CH2CH2CH2), 28.07 (CH(CH3)2, 33.71 (NCHCH2), 52.94 (NCH), 123.19 (CHAr), -1 126.21 (br CHAr), 142.87 (br CAr). One CAr peak is not observed. IR (Nujol) ν/cm ; 665 (sh, w), 721 (sh, w), 754 (sh, w), 751 (sh, m), 799 (sh, m), 837 (sh, w), 935 (sh, w), 1058 (sh, w), 1097 (sh, w), 1189 (sh, w), 1219 (sh, w), 1257 (br, m), 1308 (sh, w), 1327 (sh, w), 1361 (sh, w), 1518 (sh, w), 1588 (sh, w), 3060 (sh, w), 3137 (br, w). Elemental analysis calculated (%) for C60H91N7Li2: C, 77.97; H, 9.92; N, 10.61. Results were repeatedly inconsistent with calculated values due to sample decomposition pre- analysis.

Bis(dicyclohexylamido)(1,3-bis(2,6-diisopropylphenyl)triazenidotetrahydrofuran cerium(III), [Ce(NCy2)(N3Dipp2)(thf)] 200

A solution of LiNCy2 (1.50 g, 8.03 mmol) in tetrahydrofuran (ca. 3 40 cm ) was added to a suspension of Ce(OTf)3 (1.18 g, 2.01 mmol) also in tetrahydrofuran (ca. 10 cm3) at -40 °C and warmed to 0 °C over two hours. Volatiles were removed in vacuo and the 3 residue suspended in toluene (ca. 30 cm ). A solution of Dipp2N3H (735 mg, 2.01 mmol) also in toluene (ca. 10 cm3) was added and the suspension stirred overnight. Volatiles were removed in vacuo and nhexane (ca. 40 cm3) was added. The suspension was filtered and the precipitate washed with further nhexane (2 x 5 cm3). The filtrate and 3 washings were concentrated in vacuo to ca. 5 cm and stored at room temperature 210 overnight to afford a mixed crop of black rods, which were confirmed as [Ce(NCy2)4] by examination of the unit cell of a single crystal and comparison to published data and orange rods of the title compound. Attempts to re-crystallise the title compound by n extraction into hexane caused the oxidation of the cerium compound to [Ce(NCy2)4].

Iodopentakis(tetrahydrofuran)calcium 4-benzhydryl-2-methyl-9-phenylfluorenide,

[CaI(thf)5][Fluor*] 220

A suspension of calcium powder (244 mg, 6.09 mmol) and Dipp*I (1.67 g, 3.04 mmol) in tetrahydrofuran (35 cm3) was stirred at room temperature for 24 hours. The resulting blood-red suspension was filtered and the precipitate washed with further tetrahydrofuran (2 x 5 cm3). The combined filtrate and washings were stored at -20 °C overnight to afford the title compound as orange-red blocks suitable for X-ray single crystal structural determination. Attempts to further manipulate the isolated compound in the solid state invariably led to decomposition, which prohibited further characterisation.

2,6-Dibenzhydryl-4-methylbenzene, Dipp*H 50

A suspension of calcium powder (2.492 g, 62.18 mmol) and Dipp*I (8.266 g, 15.02 mmol) in tetrahydrofuran (ca. 100 cm3) was sonicated at room temperature for 30 mins and then stirred at room temperature for 24 hours. The resulting blood-red suspension was filtered and the precipitate washed with further tetrahydrofuran (2 x 10 cm3) to afford calcium powder (1.858 g, 46.36 mmol, 95%). The tetrahydrofuran solution was 3 quenched by dropwise addition of sat. NH4Cl(aq.)(10 cm ) and then partitioned between 3 further sat. NH4Cl(aq.) (200 cm ). The product was extracted into ethyl acetate (3 x 50 3 cm ), the combined organic phases dried over Na2SO4, filtered and then volatiles removed under reduced pressure. The crude product was purified by flash column chromatography (dichloromethane:hexane, 15:85, 20:1 SiO2:Compound) to afford Dipp*H (4.56 g, 10.7 mmol, 72%) as the only identifiable product. Characterisation was identical to previously obtained data (vide supra).

211

Iodopentakis(tetrahydrofuran)ytterbium(II) 4-benzhydryl-2-methyl-9-phenylfluorenide,

[YbI(thf)5][Fluor*] 230

A suspension of ytterbium powder (545 mg, 3.15 mmol) and Dipp*I (689 mg, 1.25 mmol) in tetrahydrofuran (15 cm3) was sonicated overnight. The resulting blood-red suspension was filtered and the precipitate washed with further tetrahydrofuran (3 x 5 cm3). The combined filtrate and washings were concentrated to the point of incipient crystallisation (ca. 10 cm3) and stored at -20 °C overnight to afford the title compound as yellow-orange blocks (215 mg, 199 μmol, 16%).

Reaction of Dipp*I with elemental samarium.

A suspension of samarium powder (1.713 g, 11.39 mmol) and Dipp*I (2.743 g, 4.983 mmol) in tetrahydrofuran (ca. 50 cm3) was sonicated for 2 hours at room temperature and then stirred overnight also at room temperature. The deep-brown suspension was filtered and the precipitate washed with further tetrahydrofuran. The combined filtrate and washings were concentrated in vacuo to ca. 25 cm3 and then stored at -25 °C for 48 hours. An amorphous brown gum was observed to precipitate, which proved resistant to further characterisation.

Attempted reaction of Dipp*I with elemental neodymium or lanthanum.

2 A suspension of the elemental lanthanide ( /3 eq.) and Dipp*I (1 eq.) in tetrahydrofuran (ca. 25 cm3) was sonicated overnight and then heated at reflux for 24 hours. No consumption of metal was observed at this stage. Volatiles were removed in vacuo and 1,2-dimethoxyethane (ca. 15 cm3) was added. The suspension was heated at reflux for 48 hours. No consumption of the metal was observed at this stage.

The following quantities were used according to the general procedure with identical results:

212

Lanthanum powder (191 mg, 1.37 mmol) and Dipp*I (1.13 g, 2.06 mmol).

Neodymium powder (249 mg, 1.73 mmol) and Dipp*I (1.42 g, 2.59 mmol.

Crystallographic information files for all structures collected during the preparation of this thesis are included on the accompanying Special Features DVD.

213

References 1. S. Cotton, Lanthanide and Actinide Chemistry, John Wiley & Sons Ltd, Chichester, 2006. 2. R. D. Shannon, Acta Crystallogr., Sect. A, 1976, 32, 751-767. 3. W. M. Haynes, ed., CRC Handbook of Chemistry and Physics, 94 edn., CRC Press, Boca Raton, 2013. 4. N. N. Greenwood and A. Earnshaw, Chemistry of the Elements, 2nd edn., Elsevier, 1997. 5. F. G. N. Cloke, Chem. Soc. Rev., 1993, 22, 17-24. 6. M. N. Bochkarev, Chem. Rev., 2002, 102, 2089-2118. 7. R. K. Sheline and J. L. Slater, Angew. Chem., Int. Ed. Engl., 1975, 14, 309-313. 8. P. Selg, H. H. Brintzinger, M. Schultz and R. A. Andersen, Organometallics, 2002, 21, 3100-3107. 9. M. Schultz, C. J. Burns, D. J. Schwartz and R. A. Andersen, Organometallics, 2001, 20, 5690-5699. 10. L. Maron, L. Perrin, O. Eisenstein and R. A. Andersen, J. Am. Chem. Soc., 2002, 124, 5614-5615. 11. F. Nief, Dalton Trans., 2010, 39, 6589-6598. 12. P. Girard, J. L. Namy and H. B. Kagan, J. Am. Chem. Soc., 1980, 102, 2693- 2698. 13. M. Szostak, M. Spain and D. J. Procter, J. Org. Chem., 2012, 77, 3049-3059. 14. J. K. Howell and L. L. Pytlewski, J. Less-Common Met., 1969, 18, 437-439. 15. G. A. Molander and C. R. Harris, Chem. Rev., 1996, 96, 307-338. 16. D. J. Edmonds, D. Johnston and D. J. Procter, Chem. Rev., 2004, 104, 3371- 3404. 17. H. B. Kagan, Tetrahedron, 2003, 59, 10351-10372. 18. G. A. Molander, Chem. Rev., 1992, 92, 29-68. 19. M. Szostak and D. J. Procter, Angew. Chem., Int. Ed., 2012, 51, 9238-9256. 20. A. Krief and A.-M. Laval, Chem. Rev., 1999, 99, 745-778. 21. M. N. Bochkarev, Coord. Chem. Rev., 2004, 248, 835-851. 22. I. L. Fedushkin, M. N. Bochkarev, S. Dechert and H. Schumann, Chem. Eur. J., 2001, 7, 3558-3563. 23. M. N. Bochkarev, I. L. Fedushkin, A. A. Fagin, T. V. Petrovskaya, J. W. Ziller, R. N. R. Broomhall-Dillard and W. J. Evans, Angew. Chem., Int. Ed. Engl., 1997, 36, 133-135. 24. M. N. Bochkarev, I. L. Fedushkin, S. Dechert, A. A. Fagin and H. Schumann, Angew. Chem., Int. Ed., 2001, 40, 3176-3178. 25. M. N. Bochkarev and A. A. Fagin, Chem. Eur. J., 1999, 5, 2990-2992. 26. M. A. Katkova, G. K. Fukin, A. A. Fagin and M. N. Bochkarev, J. Organomet. Chem., 2003, 682, 218-223. 27. W. J. Evans, N. T. Allen and J. W. Ziller, J. Am. Chem. Soc., 2000, 122, 11749- 11750. 28. W. J. Evans, N. T. Allen, P. S. Workman and J. C. Meyer, Inorg. Chem., 2003, 42, 3097-3099. 29. M. E. Fieser, M. R. MacDonald, B. T. Krull, J. E. Bates, J. W. Ziller, F. Furche and W. J. Evans, J. Am. Chem. Soc., 2015, 137, 369-382. 30. V. Nair, L. Balagopal, R. Rajan and J. Mathew, Acc. Chem. Res., 2003, 37, 21- 30. 214

31. V. Nair and A. Deepthi, Chem. Rev., 2007, 107, 1862-1891. 32. V. Sridharan and J. C. Menéndez, Chem. Rev., 2010, 110, 3805-3849. 33. V. S. Sastri, J.-C. Bünzli, V. R. Rao, G. V. S. Rayudu and J. R. Perumareddi, Modern Aspects of Rare Earths and Their Complexes, Elsevier, Amsterdam, 2003. 34. P. Hermann, J. Kotek, V. Kubicek and I. Lukes, Dalton Trans., 2008, 3027- 3047. 35. P. Caravan, J. J. Ellison, T. J. McMurry and R. B. Lauffer, Chem. Rev., 1999, 99, 2293-2352. 36. S. Arndt and J. Okuda, Chem. Rev., 2002, 102, 1953-1976. 37. F. T. Edelmann, in Comprehensive Organometallic Chemistry III, eds. D. M. P. Mingos and R. H. Crabtree, Elsevier, Oxford, 2007, pp. 1-190. 38. G. Wilkinson and J. M. Birmingham, J. Am. Chem. Soc., 1954, 76, 6210-6210. 39. R. H. Crabtree, The Organometallic Chemistry of the Transition Metals, 4th edn., John Wiley & Sons, Inc., 2005. 40. M. F. Lappert, A. Singh and J. Takats, in Inorganic Syntheses, ed. A. P. Ginsberg, John Wiley & Sons, Inc., New York, 1990, vol. 27, pp. 168-172. 41. M. R. MacDonald, J. W. Ziller and W. J. Evans, J. Am. Chem. Soc., 2011, 133, 15914-15917. 42. M. R. MacDonald, J. E. Bates, M. E. Fieser, J. W. Ziller, F. Furche and W. J. Evans, J. Am. Chem. Soc., 2012, 134, 8420-8423. 43. M. R. MacDonald, J. E. Bates, J. W. Ziller, F. Furche and W. J. Evans, J. Am. Chem. Soc., 2013, 135, 9857-9868. 44. M. R. MacDonald, M. E. Fieser, J. E. Bates, J. W. Ziller, F. Furche and W. J. Evans, J. Am. Chem. Soc., 2013, 135, 13310-13313. 45. R. Anwander, in Lanthanides: Chemistry and Use in Organic Synthesis, ed. S. Kobayashi, Springer Berlin Heidelberg, 1999, vol. 2, pp. 1-61. 46. G. Jeske, H. Lauke, H. Mauermann, P. N. Swepston, H. Schumann and T. J. Marks, J. Am. Chem. Soc., 1985, 107, 8091-8103. 47. Z. Hou and Y. Wakatsuki, Coord. Chem. Rev., 2002, 231, 1-22. 48. G. Jeske, H. Lauke, H. Mauermann, H. Schumann and T. J. Marks, J. Am. Chem. Soc., 1985, 107, 8111-8118. 49. G. A. Molander and J. A. C. Romero, Chem. Rev., 2002, 102, 2161-2186. 50. A. L. Reznichenko and K. C. Hultzsch, in Molecular Catalysis of Rare-Earth Elements, ed. P. W. Roesky, Springer Berlin Heidelberg, 2010, vol. 137, pp. 1- 48. 51. S. Hong and T. J. Marks, Acc. Chem. Res., 2004, 37, 673-686. 52. G. A. Molander, Pure Appl. Chem., 2000, 72, 1757-1761. 53. G. A. Molander and E. D. Dowdy, in Lanthanides: Chemistry and Use in Organic Synthesis, ed. S. Kobayashi, Springer Berlin Heidelberg, 1999, vol. 2, pp. 119-154. 54. R. Waterman, Organometallics, 2013, 32, 7249-7263. 55. W. J. Evans, Inorg. Chem., 2007, 46, 3435-3449. 56. W. J. Evans, J. Organomet. Chem., 2002, 647, 2-11. 57. W. J. Evans and B. L. Davis, Chem. Rev., 2002, 102, 2119-2136. 58. W. J. Evans, T. J. Mueller and J. W. Ziller, J. Am. Chem. Soc., 2009, 131, 2678- 2686. 59. W. J. Evans, J. M. Perotti, S. A. Kozimor, T. M. Champagne, B. L. Davis, G. W. Nyce, C. H. Fujimoto, R. D. Clark, M. A. Johnston and J. W. Ziller, Organometallics, 2005, 24, 3916-3931. 215

60. W. J. Evans, K. J. Forrestal and J. W. Ziller, J. Am. Chem. Soc., 1998, 120, 9273-9282. 61. W. J. Evans, D. S. Lee, C. Lie and J. W. Ziller, Angew. Chem., Int. Ed., 2004, 43, 5517-5519. 62. W. J. Evans, D. S. Lee, M. A. Johnston and J. W. Ziller, Organometallics, 2005, 24, 6393-6397. 63. W. J. Evans, D. B. Rego and J. W. Ziller, Inorg. Chem., 2006, 45, 10790-10798. 64. S. Demir, S. E. Lorenz, M. Fang, F. Furche, G. Meyer, J. W. Ziller and W. J. Evans, J. Am. Chem. Soc., 2010, 132, 11151-11158. 65. W. J. Evans, S. E. Lorenz and J. W. Ziller, Inorg. Chem., 2009, 48, 2001-2009. 66. G. W. Watt and E. W. Gillow, J. Am. Chem. Soc., 1969, 91, 775-776. 67. A. M. Bond, G. B. Deacon and R. H. Newnham, Organometallics, 1986, 5, 2312-2316. 68. P. B. Hitchcock, J. A. K. Howard, M. F. Lappert and S. Prashar, J. Organomet. Chem., 1992, 437, 177-189. 69. M. H. Kuiper and H. Lueken, Z. Anorg. Allg. Chem., 2007, 633, 1407-1409. 70. M. Schultz, C. J. Burns, D. J. Schwartz and R. A. Andersen, Organometallics, 2000, 19, 781-789. 71. W. J. Evans, I. Bloom, W. E. Hunter and J. L. Atwood, J. Am. Chem. Soc., 1981, 103, 6507-6508. 72. W. J. Evans, L. A. Hughes and T. P. Hanusa, J. Am. Chem. Soc., 1984, 106, 4270-4272. 73. W. J. Evans, T. A. Ulibarri and J. W. Ziller, J. Am. Chem. Soc., 1988, 110, 6877- 6879. 74. W. J. Evans, Coord. Chem. Rev., 2000, 206–207, 263-283. 75. L. Gehman, K. Vasudevan, R. Butorac, S. Al-Deyab and A. Cowley, J. Chem. Crystallogr., 2011, 41, 998-1001. 76. K. V. Vasudevan, I. Vargas-Baca and A. H. Cowley, Angew. Chem., Int. Ed., 2009, 48, 8369-8371. 77. J. A. Moore, A. H. Cowley and J. C. Gordon, Organometallics, 2006, 25, 5207- 5209. 78. K. Vasudevan and A. H. Cowley, Chem. Commun., 2007, 3464-3466. 79. D. J. Berg, C. J. Burns, R. A. Andersen and A. Zalkin, Organometallics, 1989, 8, 1865-1870. 80. D. J. Berg, R. A. Andersen and A. Zalkin, Organometallics, 1988, 7, 1858-1863. 81. M. Schultz, J. M. Boncella, D. J. Berg, T. D. Tilley and R. A. Andersen, Organometallics, 2001, 21, 460-472. 82. M. D. Walter, D. J. Berg and R. A. Andersen, Organometallics, 2006, 25, 3228- 3237. 83. G. Nocton, C. H. Booth, L. Maron and R. A. Andersen, Organometallics, 2013, 32, 5305-5312. 84. C. H. Booth, D. Kazhdan, E. L. Werkema, M. D. Walter, W. W. Lukens, E. D. Bauer, Y.-J. Hu, L. Maron, O. Eisenstein, M. Head-Gordon and R. A. Andersen, J. Am. Chem. Soc., 2010, 132, 17537-17549. 85. C. H. Booth, M. D. Walter, D. Kazhdan, Y.-J. Hu, W. W. Lukens, E. D. Bauer, L. Maron, O. Eisenstein and R. A. Andersen, J. Am. Chem. Soc., 2009, 131, 6480-6491. 86. M. D. Walter, D. J. Berg and R. A. Andersen, Organometallics, 2007, 26, 2296- 2307.

216

87. D. J. Berg, J. M. Boncella and R. A. Andersen, Organometallics, 2002, 21, 4622-4631. 88. W. J. Evans, N. T. Allen and J. W. Ziller, J. Am. Chem. Soc., 2001, 123, 7927- 7928. 89. I. L. Fedushkin, F. Girgsdies, H. Schumann and M. N. Bochkarev, Eur. J. Inorg. Chem., 2001, 2001, 2405-2410. 90. W. J. Evans, N. T. Allen and J. W. Ziller, Angew. Chem., Int. Ed., 2002, 41, 359-361. 91. F. Jaroschik, F. Nief, X.-F. Le Goff and L. Ricard, Organometallics, 2007, 26, 3552-3558. 92. F. Jaroschik, F. Nief and L. Ricard, Chem. Commun., 2006, 426-428. 93. F. Jaroschik, F. Nief, X.-F. Le Goff and L. Ricard, Organometallics, 2007, 26, 1123-1125. 94. F. Jaroschik, A. Momin, F. Nief, X.-F. Le Goff, G. B. Deacon and P. C. Junk, Angew. Chem., Int. Ed., 2009, 48, 1117-1121. 95. Y. K. Gun'ko, P. B. Hitchcock and M. F. Lappert, J. Organomet. Chem., 1995, 499, 213-219. 96. M. Cristina Cassani, M. F. Lappert and F. Laschi, Chem. Commun., 1997, 1563- 1564. 97. P. B. Hitchcock, M. F. Lappert, L. Maron and A. V. Protchenko, Angew. Chem., Int. Ed., 2008, 47, 1488-1491. 98. M. C. Cassani, Y. K. Gun'ko, P. B. Hitchcock and M. F. Lappert, Chem. Commun., 1996, 1987-1988. 99. M. C. Cassani, Y. K. Gun'ko, P. B. Hitchcock, M. F. Lappert and F. Laschi, Organometallics, 1999, 18, 5539-5547. 100. M. C. Cassani, D. J. Duncalf and M. F. Lappert, J. Am. Chem. Soc., 1998, 120, 12958-12959. 101. Y. K. Gun'ko, P. B. Hitchcock and M. F. Lappert, Organometallics, 2000, 19, 2832-2834. 102. F. Nief, D. Turcitu and L. Ricard, Chem. Commun., 2002, 1646-1647. 103. D. Turcitu, F. Nief and L. Ricard, Chem. Eur. J., 2003, 9, 4916-4923. 104. P. L. Arnold, F. G. N. Cloke, P. B. Hitchcock and J. F. Nixon, J. Am. Chem. Soc., 1996, 118, 7630-7631. 105. P. L. Arnold, F. Geoffrey N. Cloke and J. F. Nixon, Chem. Commun., 1998, 797-798. 106. G. K. B. Clentsmith, F. G. N. Cloke, J. C. Green, J. Hanks, P. B. Hitchcock and J. F. Nixon, Angew. Chem., Int. Ed., 2003, 42, 1038-1041. 107. A. Greco, S. Cesca and W. Bertolini, J. Organomet. Chem., 1976, 113, 321-330. 108. V. Lorenz, B. M. Schmiege, C. G. Hrib, J. W. Ziller, A. Edelmann, S. Blaurock, W. J. Evans and F. T. Edelmann, J. Am. Chem. Soc., 2011, 133, 1257-1259. 109. U. Kilimann, R. Herbst-Irmer, D. Stalke and F. T. Edelmann, Angew. Chem., Int. Ed. Engl., 1994, 33, 1618-1621. 110. T. R. Boussie, D. C. Eisenberg, J. Rigsbee, A. Streitwieser and A. Zalkin, Organometallics, 1991, 10, 1922-1928. 111. M. D. Walter, C. H. Booth, W. W. Lukens and R. A. Andersen, Organometallics, 2009, 28, 698-707. 112. A. Kerridge, R. Coates and N. Kaltsoyannis, J. Phys. Chem. A, 2009, 113, 2896- 2905. 113. G. Balazs, F. G. N. Cloke, J. C. Green, R. M. Harker, A. Harrison, P. B. Hitchcock, C. N. Jardine and R. Walton, Organometallics, 2007, 26, 3111-3119. 217

114. A. Ashley, G. Balazs, A. Cowley, J. Green, C. H. Booth and D. O'Hare, Chem. Commun., 2007, 1515-1517. 115. P. Dröse, A. R. Crozier, S. Lashkari, J. Gottfriedsen, S. Blaurock, C. G. Hrib, C. Maichle-Mössmer, C. Schädle, R. Anwander and F. T. Edelmann, J. Am. Chem. Soc., 2010, 132, 14046-14047. 116. A. Gulino, M. Casarin, V. P. Conticello, J. G. Gaudiello, H. Mauermann, I. Fragala and T. J. Marks, Organometallics, 1988, 7, 2360-2364. 117. W. J. Evans, T. J. Deming and J. W. Ziller, Organometallics, 1989, 8, 1581- 1583. 118. D. C. Bradley, J. S. Ghotra and F. A. Hart, J. Chem. Soc., Dalton Trans., 1973, 1021-1023. 119. S. A. Schuetz, V. W. Day, R. D. Sommer, A. L. Rheingold and J. A. Belot, Inorg. Chem., 2001, 40, 5292-5295. 120. R. Anwander, in Organolanthanoid Chemistry: Synthesis, Structure, Catalysis, Springer Berlin, 1996, vol. 179, pp. 33-112. 121. M. G. Klimpel, H. W. Görlitzer, M. Tafipolsky, M. Spiegler, W. Scherer and R. Anwander, J. Organomet. Chem., 2002, 647, 236-244. 122. W. A. Herrmann, R. Anwander, F. C. Munck, W. Scherer, V. Dufaud, N. W. Huber and G. R. J. Artus, Z. Naturforsch., B: Chem. Sci., 1994, 49, 1789-1797. 123. W. J. Evans, M. Fang, G. l. Zucchi, F. Furche, J. W. Ziller, R. M. Hoekstra and J. I. Zink, J. Am. Chem. Soc., 2009, 131, 11195-11202. 124. J. H. Farnaby, M. Fang, J. W. Ziller and W. J. Evans, Inorg. Chem., 2012, 51, 11168-11176. 125. M. Fang, D. S. Lee, J. W. Ziller, R. J. Doedens, J. E. Bates, F. Furche and W. J. Evans, J. Am. Chem. Soc., 2011, 133, 3784-3787. 126. M. Fang, J. H. Farnaby, J. W. Ziller, J. E. Bates, F. Furche and W. J. Evans, J. Am. Chem. Soc., 2012, 134, 6064-6067. 127. W. J. Evans, M. Fang, J. E. Bates, F. Furche, J. W. Ziller, M. D. Kiesz and J. I. Zink, Nature Chem., 2010, 2, 644-647. 128. T. D. Tilley, A. Zalkin, R. A. Andersen and D. H. Templeton, Inorg. Chem., 1981, 20, 551-554. 129. T. D. Tilley, R. A. Andersen and A. Zalkin, Inorg. Chem., 1984, 23, 2271-2276. 130. W. J. Evans, D. K. Drummond, H. Zhang and J. L. Atwood, Inorg. Chem., 1988, 27, 575-579. 131. G. B. Deacon, G. D. Fallon, C. M. Forsyth, H. Schumann and R. Weimann, Chem. Ber., 1997, 130, 409-415. 132. M. Niemeyer, Inorg. Chem., 2006, 45, 9085-9095. 133. W. J. Evans, M. A. Johnston, R. D. Clark, R. Anwander and J. W. Ziller, Polyhedron, 2001, 20, 2483–2490. 134. I. Nagl, W. Scherer, M. Tafipolsky and R. Anwander, Eur. J. Inorg. Chem., 1999, 1999, 1405-1407. 135. G. W. Rabe, A. L. Rheingold and C. D. Incarvito, Z. Kristallogr. -New Cryst. Struct., 2000, 215, 560. 136. A. M. Bienfait, C. Schadle, C. Maichle-Mossmer, K. W. Tornroos and R. Anwander, Dalton Trans., 2014, 43, 17324-17332. 137. W. J. Evans, G. Zucchi and J. W. Ziller, J. Am. Chem. Soc., 2002, 125, 10-11. 138. E. Prasad, B. W. Knettle and R. A. Flowers, J. Am. Chem. Soc., 2002, 124, 14663-14667. 139. Z. Hou, A. Fujita, Y. Zhang, T. Miyano, H. Yamazaki and Y. Wakatsuki, J. Am. Chem. Soc., 1998, 120, 754-766. 218

140. E. D. Brady, D. L. Clark, D. W. Keogh, B. L. Scott and J. G. Watkin, J. Am. Chem. Soc., 2002, 124, 7007-7015. 141. W. J. Evans, D. K. Drummond, L. R. Chamberlain, R. J. Doedens, S. G. Bott, H. Zhang and J. L. Atwood, J. Am. Chem. Soc., 1988, 110, 4983-4994. 142. M. Janjetovic, A. M. Traff, T. Ankner, J. Wettergren and G. Hilmersson, Chem. Commun., 2013, 49, 1826-1828. 143. T. Ankner, M. Fridén-Saxin, N. Pemberton, T. Seifert, M. Grøtli, K. Luthman and G. Hilmersson, Org. Lett., 2010, 12, 2210-2213. 144. M. Kim, B. W. Knettle, A. Dahlén, G. Hilmersson and R. A. Flowers II, Tetrahedron, 2003, 59, 10397-10402. 145. Z. Du, W. Li, X. Zhu, F. Xu and Q. Shen, J. Org. Chem., 2008, 73, 8966-8972. 146. W. Xie, H. Hu and C. Cui, Angew. Chem., Int. Ed., 2012, 51, 11141-11144. 147. A. R. Crozier, A. M. Bienfait, C. Maichle-Mossmer, K. W. Tornroos and R. Anwander, Chem. Commun., 2013, 49, 87-89. 148. P. B. Hitchcock, M. F. Lappert and A. V. Protchenko, Chem. Commun., 2006, 3546-3548. 149. U. J. Williams, J. R. Robinson, A. J. Lewis, P. J. Carroll, P. J. Walsh and E. J. Schelter, Inorg. Chem., 2013, 53, 27-29. 150. A. F. England, Ph.D. Thesis, Massachusetts Institute of Technology, 1995. 151. C. Morton, N. W. Alcock, M. R. Lees, I. J. Munslow, C. J. Sanders and P. Scott, J. Am. Chem. Soc., 1999, 121, 11255-11256. 152. O. Eisenstein, P. B. Hitchcock, A. G. Hulkes, M. F. Lappert and L. Maron, Chem. Commun., 2001, 1560-1561. 153. P. B. Hitchcock, A. G. Hulkes and M. F. Lappert, Inorg. Chem., 2003, 43, 1031- 1038. 154. A. R. Crozier, C. Schadle, C. Maichle-Mossmer, K. W. Tornroos and R. Anwander, Dalton Trans., 2013, 42, 5491-5499. 155. D. Werner, G. B. Deacon, P. C. Junk and R. Anwander, Chem. Eur. J., 2014, 20, 4426-4438. 156. W. J. Evans, T. J. Deming, J. M. Olofson and J. W. Ziller, Inorg. Chem., 1989, 28, 4027-4034. 157. M. P. Coles, P. B. Hitchcock, A. V. Khvostov, M. F. Lappert, Z. Li and A. V. Protchenko, Dalton Trans., 2010, 39, 6780-6788. 158. M. N. Bochkarev, L. N. Zakharov and G. S. Kalinina, Organoderivatives of Rare Earth Elements, Springer Netherlands, Dordrecht, 1995. 159. C. D. Sofield and R. A. Andersen, J. Organomet. Chem., 1995, 501, 271-276. 160. M. Zimmermann and R. Anwander, Chem. Rev., 2010, 110, 6194-6259. 161. G. B. Deacon, C. M. Forsyth and S. Nickel, J. Organomet. Chem., 2002, 647, 50-60. 162. G. Deacon, A. Koplick and T. Tuong, Aust. J. Chem., 1984, 37, 517-525. 163. G. B. Deacon, C. M. Forsyth, R. H. Newnham and T. D. Tuong, Aust. J. Chem., 1987, 40, 895-906. 164. G. B. Deacon, G. N. Pain, T. D. Tuong, W. J. Evans, K. R. Levan and R. Dominguez, in Inorganic Syntheses, John Wiley & Sons, Inc., New York, 1989, vol. 26, pp. 17-23. 165. F. T. Edelmann, in Advances in Organometallic Chemistry, eds. A. F. Hill and M. J. Fink, Academic Press, 2008, vol. Volume 57, pp. 183-352. 166. P. J. Bailey and S. Pace, Coord. Chem. Rev., 2001, 214, 91-141. 167. C. Jones, Coord. Chem. Rev., 2010, 254, 1273-1289. 168. J. Barker and M. Kilner, Coord. Chem. Rev., 1994, 133, 219-300. 219

169. F. Edelmann, in Molecular Catalysis of Rare-Earth Elements, ed. P. W. Roesky, Springer Berlin Heidelberg, 2010, vol. 137, pp. 109-163. 170. F. T. Edelmann, in Advances in Organometallic Chemistry, eds. F. H. Anthony and J. F. Mark, Academic Press, 2013, vol. Volume 61, pp. 55-374. 171. P. C. Junk and M. L. Cole, Chem. Commun., 2007, 1579-1590. 172. F. T. Edelmann, Coord. Chem. Rev., 1994, 137, 403-481. 173. W. W. Hartman and J. B. Dickey, in Organic Syntheses, John Wiley & Sons, Inc., 1934, vol. 14, pp. 24-27. 174. F. P. Dwyer and D. P. Mellor, J. Am. Chem. Soc., 1941, 63, 81-83. 175. F. P. Dwyer, J. Am. Chem. Soc., 1941, 63, 78-81. 176. F. A. Cotton, G. W. Rice and J. C. Sekutowski, Inorg. Chem., 1979, 18, 1143- 1149. 177. F. A. Cotton, W. H. Ilsley and W. Kaim, Inorg. Chem., 1980, 19, 1450-1452. 178. M. H. Chisholm, D. A. Haitko, J. C. Huffman and K. Folting, Inorg. Chem., 1981, 20, 2211-2215. 179. M. J. Chetcuti, M. H. Chisholm, K. Folting, D. A. Haitko and J. C. Huffman, J. Am. Chem. Soc., 1982, 104, 2138-2146. 180. M. H. Chisholm, H. T. Chiu, J. C. Huffman and R. J. Wang, Inorg. Chem., 1986, 25, 1092-1096. 181. R. S. Cahn, C. Ingold and V. Prelog, Angew. Chem., Int. Ed. Engl., 1966, 5, 385- 415. 182. V. Prelog and G. Helmchen, Angew. Chem., Int. Ed. Engl., 1982, 21, 567-583. 183. A. L. Johnson, A. M. Willcocks and S. P. Richards, Inorg. Chem., 2009, 48, 8613-8622. 184. H. S. Lee and M. Niemeyer, Inorg. Chim. Acta, 2011, 374, 163-170. 185. C. Lego and B. Neumüller, Z. Anorg. Allg. Chem., 2011, 637, 1784-1789. 186. A. Immirzi, W. Porzio, G. Bombieri and L. Toniolo, J. Chem. Soc., Dalton Trans., 1980, 1098-1100. 187. A. S. Peregudov, D. N. Kravtsov, G. I. Drogunova, Z. A. Starikova and A. I. Yanovsky, J. Organomet. Chem., 2000, 597, 164-174. 188. D. F. Back, M. Hörner, F. Broch and G. Manzoni de Oliveira, Polyhedron, 2012, 31, 558-564. 189. G. Bombieri, A. Immirzi and L. Toniolo, Inorg. Chem., 1976, 15, 2428-2432. 190. M. Hörner, G. M. de Oliveira, E. G. Koehler and L. do Canto Visentin, J. Organomet. Chem., 2006, 691, 1311-1314. 191. M. Hörner, F. Broch and L. do Canto Visentin, Z. Anorg. Allg. Chem., 2007, 633, 1779-1782. 192. S.-O. Hauber, J. Woo Seo and M. Niemeyer, Z. Anorg. Allg. Chem., 2010, 636, 750-757. 193. W. Li, J.-Y. Chen, W.-Q. Xu, E.-X. He, S.-Z. Zhan and D.-R. Cao, Inorg. Chem. Commun., 2011, 14, 916-919. 194. W. Li, X.-W. Tan, W.-J. Lei, S.-Z. Zhan and D.-R. Cao, Transition Met. Chem., 2010, 35, 835-839. 195. W. Li, Y.-L. Liu, J. Liu, H.-t. Zhang, S.-Z. Zhan and D.-R. Cao, Transition Met. Chem., 2011, 36, 255-260. 196. S. J. Kraft, P. E. Fanwick and S. C. Bart, Organometallics, 2013, 32, 3279-3285. 197. E. M. Matson, W. P. Forrest, P. E. Fanwick and S. C. Bart, Organometallics, 2013, 32, 1484-1492. 198. W. J. Evans, J. R. Walensky, J. W. Ziller and A. L. Rheingold, Organometallics, 2009, 28, 3350-3357. 220

199. W. J. Evans, C. A. Traina and J. W. Ziller, J. Am. Chem. Soc., 2009, 131, 17473- 17481. 200. A. G. M. Barrett, M. R. Crimmin, M. S. Hill, P. B. Hitchcock, G. Kociok-Köhn and P. A. Procopiou, Inorg. Chem., 2008, 47, 7366-7376. 201. D. S. Moore and S. D. Robinson, in Advances in Inorganic Chemistry, ed. H. J. Emeléus, Academic Press, 1986, vol. 30, pp. 1-68. 202. S. G. Alexander, M. L. Cole, C. M. Forsyth, S. K. Furfari and K. Konstas, Dalton Trans., 2009, 2326-2336. 203. S. Balireddi and M. Niemeyer, Acta Crystallogr,. Sect. E.-Struct Rep. Online, 2007, 63, o3525-o3525. 204. S.-O. Hauber, F. Lissner, G. B. Deacon and M. Niemeyer, Angew. Chem., Int. Ed., 2005, 44, 5871-5875. 205. M. L. Cole, A. J. Davies, C. Jones and P. C. Junk, J. Organomet. Chem., 2004, 689, 3093-3107. 206. M. L. Cole and P. C. Junk, J. Organomet. Chem., 2003, 666, 55-62. 207. M. L. Cole, A. J. Davies, C. Jones and P. C. Junk, J. Organomet. Chem., 2007, 692, 2508-2518. 208. M. L. Cole, A. J. Davies, C. Jones and P. C. Junk, New J. Chem., 2005, 29, 1404-1408. 209. M. L. Cole, A. J. Davies, C. Jones and P. C. Junk, Z. Anorg. Allg. Chem., 2011, 637, 50-55. 210. J. Richter, J. Feiling, H.-G. Schmidt, M. Noltemeyer, W. Brüser and F. T. Edelmann, Z. Anorg. Allg. Chem., 2004, 630, 1269-1275. 211. J. A. R. Schmidt and J. Arnold, Chem. Commun., 1999, 2149-2150. 212. J. A. R. Schmidt and J. Arnold, J. Chem. Soc., Dalton Trans., 2002, 2890-2899. 213. R. J. Baker and C. Jones, J. Organomet. Chem., 2006, 691, 65-71. 214. Wood, G. P. A. Yap and D. S. Richeson, Inorg. Chem., 1999, 38, 5788-5794. 215. C. Jones, P. C. Junk, J. A. Platts and A. Stasch, J. Am. Chem. Soc., 2006, 128, 2206-2207. 216. G. Jin, C. Jones, P. C. Junk, A. Stasch and W. D. Woodul, New J. Chem., 2008, 32, 835-842. 217. Y.-J. Luo, Y. Zhang and Q. Shen, Chin. J. Chem., 2007, 25, 562-565. 218. G. R. Giesbrecht, A. Shafir and J. Arnold, J. Chem. Soc., Dalton Trans., 1999, 3601-3604. 219. Y. Luo, Y. Yao, Q. Shen, K. Yu and L. Weng, Eur. J. Inorg. Chem., 2003, 2003, 318-323. 220. Y. Zhou, G. P. A. Yap and D. S. Richeson, Organometallics, 1998, 17, 4387- 4391. 221. P. Gantzel and P. J. Walsh, Inorg. Chem., 1998, 37, 3450-3451. 222. R. Rossi, A. Duatti, L. Magon and L. Toniolo, Inorg. Chim. Acta, 1981, 48, 243- 246. 223. M. Hörner, A. Marques da Silva and H. Fenner, Z. Anorg. Allg. Chem., 2007, 633, 776-779. 224. H. S. Lee and M. Niemeyer, Inorg. Chem., 2006, 45, 6126-6128. 225. H. S. Lee, S.-O. Hauber, D. Vinduš and M. Niemeyer, Inorg. Chem., 2008, 47, 4401-4412. 226. R. Litlabo, H. S. Lee, M. Niemeyer, K. W. Tornroos and R. Anwander, Dalton Trans., 2010, 39, 6815-6825. 227. J. Beck and J. Strähle, Z. Naturforsch., B: Chem. Sci., 1986, 41b, 1381-1386.

221

228. M. Hörner, G. M. de Oliveira, L. Bresolin and A. B. de Oliveira, Inorg. Chim. Acta, 2006, 359, 4631-4634. 229. M. Hörner, H. Fenner, J. Beck and W. Hiller, Z. Anorg. Allg. Chem., 1989, 571, 69-74. 230. M. Hörner, A. B. de Oliveira and J. Beck, Z. Anorg. Allg. Chem., 1997, 623, 65- 68. 231. C. Jones, P. C. Junk, J. A. Platts, D. Rathmann and A. Stasch, Dalton Trans., 2005, 2497-2499. 232. D. Kottmair-Maieron, R. Lechler and J. Weidlein, Z. Anorg. Allg. Chem., 1991, 593, 111-123. 233. J. Bouffard, B. K. Keitz, R. Tonner, G. Guisado-Barrios, G. Frenking, R. H. Grubbs and G. Bertrand, Organometallics, 2011, 30, 2617-2627. 234. P. Van Der Sluis and A. L. Spek, Acta Crystallogr., Sect. A, 1990, 46, 194-201. 235. A. L. Spek, J. Appl. Crystallogr., 2003, 36, 7-13. 236. G. Berthon-Gelloz, M. A. Siegler, A. L. Spek, B. Tinant, J. N. H. Reek and I. E. Marko, Dalton Trans., 2010, 39, 1444-1446. 237. S. Bräse, C. Gil, K. Knepper and V. Zimmermann, Angew. Chem., Int. Ed., 2005, 44, 5188-5240. 238. J. Gavenonis and T. D. Tilley, Organometallics, 2002, 21, 5549-5563. 239. J. Gavenonis and T. D. Tilley, Organometallics, 2003, 23, 31-43. 240. T. B. Ditri, B. J. Fox, C. E. Moore, A. L. Rheingold and J. S. Figueroa, Inorg. Chem., 2009, 48, 8362-8375. 241. E. Lieber, C. N. R. Rao, T. S. Chao and C. W. W. Hoffman, Anal. Chem., 1957, 29, 916-918. 242. B. Twamley, C.-S. Hwang, N. J. Hardman and P. P. Power, J. Organomet. Chem., 2000, 609, 152-160. 243. Rei, A. Schulz, A. Villinger and N. Weding, Dalton Trans., 2010, 39, 9962- 9972. 244. R. J. Wright, A. D. Phillips, T. L. Allen, W. H. Fink and P. P. Power, J. Am. Chem. Soc., 2003, 125, 1694-1695. 245. T. Takayama, T. Mitsumori, M. Kawano, A. Sekine, H. Uekusa, Y. Ohashi and T. Sugawara, Acta Crystallogr., Sect. B-Struct. Sci., 2010, 66, 639-646. 246. J. I. G. Cadogan and G. A. Molina, J. Chem. Soc., Perkin Trans. 1, 1973, 541- 542. 247. D. A. Dickie, D. Abeysekera, I. D. McKenzie, H. A. Jenkins and J. A. C. Clyburne, Cryst. Eng., 2003, 6, 79-86. 248. A. Sattler and G. Parkin, J. Am. Chem. Soc., 2012, 134, 2355-2366. 249. M. Niemeyer, Organometallics, 1998, 17, 4649-4656. 250. C. Stanciu, A. R. Fox, A. F. Richards, J. C. Fettinger and P. P. Power, J. Organomet. Chem., 2006, 691, 2546-2553. 251. M. Saito, Y. Okuyama, T. Tajima, D. Kato and M. Yoshioka, Appl. Organomet. Chem., 2007, 21, 604-611. 252. M. Olaru, S. Rosca and C. I. Rat, Acta Crystallogr,. Sect. E.-Struct Rep. Online, 2011, 67, o213. 253. B. Twamley, N. J. Hardman and P. P. Power, Acta Crystallogr., Sect. C-Cryst. Struct. Commun., 2000, 56, e514-e515. 254. F. Leroux, M. Schlosser, E. Zohar and I. Marek, in The Chemistry of Organolithium Compounds, John Wiley & Sons, Ltd, 2009, pp. 435-493. 255. A. I. McKay, Ph.D. Thesis, The University of New South Wales, 2015.

222

256. E. L. Kolychev, I. A. Portnyagin, V. V. Shuntikov, V. N. Khrustalev and M. S. Nechaev, J. Organomet. Chem., 2009, 694, 2454-2462. 257. R. M. Roberts, T. D. Higgins and P. R. Noyes, J. Am. Chem. Soc., 1955, 77, 3801-3805. 258. B. A. Trofimov, L. V. Andriyankova, A. G. Mal'kina, K. V. Belyaeva, L. P. Nikitina, O. A. Dyachenko, O. N. Kazheva, A. N. Chekhlov, G. V. Shilov, A. V. Afonin, I. A. Ushakov and L. V. Baikalova, Eur. J. Org. Chem., 2007, 2007, 1018-1025. 259. A. Binobaid, M. Iglesias, D. J. Beetstra, B. Kariuki, A. Dervisi, I. A. Fallis and K. J. Cavell, Dalton Trans., 2009, 7099-7112. 260. J. M. Khurana, G. Kukreja and G. Bansal, J. Chem. Soc., Perkin Trans. 1, 2002, 2520-2524. 261. C. Kashima, M. Shimizu, T. Eto and Y. Omote, Bull. Chem. Soc. Jpn., 1986, 59, 3317-3319. 262. T. Matsumura and M. Nakada, Tetrahedron Lett., 2014, 55, 1412-1415. 263. M. Findlater, N. J. Hill and A. H. Cowley, Dalton Trans., 2008, 4419-4423. 264. J. Obenauf, W. P. Kretschmer and R. Kempe, Eur. J. Inorg. Chem., 2014, 2014, 1446-1453. 265. J. C. Jochims, Chem. Ber., 1965, 98, 2128-2133. 266. C. L. Stevens, G. H. Singhal and A. B. Ash, J. Org. Chem., 1967, 32, 2895- 2895. 267. A. K. Maity, S. Fortier, L. Griego and A. J. Metta-Magaña, Inorg. Chem., 2014, 53, 8155-8164. 268. A.-L. Chen, K.-L. Wei, R.-J. Jeng, J.-J. Lin and S. A. Dai, Macromolecules, 2010, 44, 46-59. 269. R. Boese, R. Kösterb and M. Yalpani, Z. Naturforsch., B: Chem. Sci., 1994, 49b, 1453-1458. 270. P. B. Hitchcock, M. F. Lappert and M. Layh, Chem. Commun., 1998, 201-202. 271. P. B. Hitchcock, M. F. Lappert and M. Layh, J. Chem. Soc., Dalton Trans., 2001, 2409-2416. 272. C. F. Caro, P. B. Hitchcock, M. F. Lappert and M. Layh, Chem. Commun., 1998, 1297-1298. 273. B. Ledig, M. Marsch, K. Harms and G. Boche, Angew. Chem., Int. Ed. Engl., 1992, 31, 79-80. 274. G. W. Gokel, R. P. Widera and W. P. Weber, in Organic Syntheses, John Wiley & Sons, Inc., 1976, vol. 55, pp. 96-98. 275. W. P. Weber and G. W. Gokel, Tetrahedron Lett., 1972, 13, 1637-1640. 276. P. C. J. Kamer, R. J. M. Nolte and W. Drenth, J. Am. Chem. Soc., 1988, 110, 6818-6825. 277. B. J. Fox, Q. Y. Sun, A. G. DiPasquale, A. R. Fox, A. L. Rheingold and J. S. Figueroa, Inorg. Chem., 2008, 47, 9010-9020. 278. A. S. Ionkin, W. J. Marshall, B. M. Fish and L. A. Howe, Organometallics, 2010, 29, 4154-4158. 279. W. Sattler, M. E. Ener, J. D. Blakemore, A. A. Rachford, P. J. LaBeaume, J. W. Thackeray, J. F. Cameron, J. R. Winkler and H. B. Gray, J. Am. Chem. Soc., 2013, 135, 10614-10617. 280. G. R. Fulmer, A. J. M. Miller, N. H. Sherden, H. E. Gottlieb, A. Nudelman, B. M. Stoltz, J. E. Bercaw and K. I. Goldberg, Organometallics, 2010, 29, 2176- 2179. 281. C. Glock, H. Görls and M. Westerhausen, Inorg. Chem., 2008, 48, 394-399. 223

282. C. Schädle, C. Meermann, K. W. Törnroos and R. Anwander, Eur. J. Inorg. Chem., 2010, 2010, 2841-2852. 283. A. Noor, T. Bauer, T. K. Todorova, B. Weber, L. Gagliardi and R. Kempe, Chem. Eur. J., 2013, 19, 9825-9832. 284. C. Loh, S. Seupel, H. Görls, S. Krieck and M. Westerhausen, Eur. J. Inorg. Chem., 2014, 2014, 1312-1321. 285. K. Izod, C. M. Dixon, E. McMeekin, L. Rodgers, R. W. Harrington and U. Baisch, Organometallics, 2013, 33, 378-386. 286. B. Horn, S. Pfirrmann, C. Limberg, C. Herwig, B. Braun, S. Mebs and R. Metzinger, Z. Anorg. Allg. Chem., 2011, 637, 1169-1174. 287. M. L. Cole, G. B. Deacon, C. M. Forsyth, P. C. Junk, D. Polo-Ceron and J. Wang, Dalton Trans., 2010, 39, 6732-6738. 288. F. Weber, M. Schultz, C. D. Sofield and R. A. Andersen, Organometallics, 2002, 21, 3139-3146. 289. M. Schultz, C. D. Sofield, M. D. Walter and R. A. Andersen, New J. Chem., 2005, 29, 919-927. 290. S. Jank, J. Hanss, H. Reddmann, H.-D. Amberger and N. M. Edelstein, Z. Anorg. Allg. Chem., 2002, 628, 1355-1365. 291. S. Jank, C. Guttenberger, H. Reddmann, J. Hanss and H.-D. Amberger, Z. Anorg. Allg. Chem., 2006, 632, 2429-2438. 292. H. A. Stecher, A. Sen and A. L. Rheingold, Inorg. Chem., 1988, 27, 1130-1132. 293. W. J. Evans, T. J. Mueller and J. W. Ziller, Chem. Eur. J., 2010, 16, 964-975. 294. S. C. Bart, C. Anthon, F. W. Heinemann, E. Bill, N. M. Edelstein and K. Meyer, J. Am. Chem. Soc., 2008, 130, 12536-12546. 295. Y. Sun, Z. Zhang, X. Wang, X. Li, L. Weng and X. Zhou, Organometallics, 2009, 28, 6320-6330. 296. F. T. Edelmann, D. M. M. Freckmann and H. Schumann, Chem. Rev., 2002, 102, 1851-1896. 297. W. E. Piers and D. J. H. Emslie, Coord. Chem. Rev., 2002, 233–234, 131-155. 298. J. Paivasaari, I. V. C. L. Dezelah, D. Back, H. M. El-Kaderi, M. J. Heeg, M. Putkonen, L. Niinisto and C. H. Winter, J. Mater. Chem., 2005, 15, 4224-4233. 299. A. P. Milanov, R. A. Fischer and A. Devi, Inorg. Chem., 2008, 47, 11405- 11416. 300. A. P. Milanov, K. Xu, S. Cwik, H. Parala, T. de los Arcos, H.-W. Becker, D. Rogalla, R. Cross, S. Paul and A. Devi, Dalton Trans., 2012, 41, 13936-13947. 301. X. Pang, H. Sun, Y. Zhang, Q. Shen and H. Zhang, Eur. J. Inorg. Chem., 2005, 2005, 1487-1491. 302. C. Jing-Lei, Y. Ying-Ming, L. Yun-Jie, Z. Li-Ying, Z. Yong and S. Qi, J. Organomet. Chem., 2004, 689, 1019-1024. 303. F. T. Edelmann, Chem. Soc. Rev., 2012, 41, 7657-7672. 304. F. T. Edelmann, Chem. Soc. Rev., 2009, 38, 2253-2268. 305. Y. Luo, X. Wang, J. Chen, C. Luo, Y. Zhang and Y. Yao, J. Organomet. Chem., 2009, 694, 1289-1296. 306. S. Bambirra, M. W. Bouwkamp, A. Meetsma and B. Hessen, J. Am. Chem. Soc., 2004, 126, 9182-9183. 307. S. Bambirra, E. Otten, D. van Leusen, A. Meetsma and B. Hessen, Z. Anorg. Allg. Chem., 2006, 632, 1950-1952. 308. M. L. Cole, G. B. Deacon, P. C. Junk and K. Konstas, Chem. Commun., 2005, 1581-1583.

224

309. M. L. Cole, G. B. Deacon, C. M. Forsyth, P. C. Junk, K. Konstas and J. Wang, Chem. Eur. J., 2007, 13, 8092-8110. 310. M. L. Cole and P. C. Junk, Chem. Commun., 2005, 2695-2697. 311. M. L. Cole, G. B. Deacon, C. M. Forsyth, P. C. Junk, K. Konstas, J. Wang, H. Bittig and D. Werner, Chem. Eur. J., 2013, 19, 1410-1420. 312. S. Yao, H.-S. Chan, C.-K. Lam and H. K. Lee, Inorg. Chem., 2009, 48, 9936- 9946. 313. M. Wedler, M. Noltemeyer, U. Pieper, H.-G. Schmidt, D. Stalke and F. T. Edelmann, Angew. Chem., Int. Ed. Engl., 1990, 29, 894-896. 314. M. L. Cole, G. B. Deacon, P. C. Junk and J. Wang, Organometallics, 2013, 32, 1370-1378. 315. D. Heitmann, C. Jones, P. C. Junk, K.-A. Lippert and A. Stasch, Dalton Trans., 2007, 187-189. 316. D. Heitmann, C. Jones, D. P. Mills and A. Stasch, Dalton Trans., 2010, 39, 1877-1882. 317. O. P. Lam, F. W. Heinemann and K. Meyer, Angew. Chem., Int. Ed., 2011, 50, 5965-5968. 318. V. Christou, S. P. Wuller and J. Arnold, J. Am. Chem. Soc., 1993, 115, 10545- 10552. 319. N. L. Cromhout, A. R. Manning, C. J. McAdam, A. J. Palmer, A. L. Rieger, P. H. Rieger, B. H. Robinson and J. Simpson, Dalton Trans., 2003, 2224-2230. 320. D. Pfeiffer, I. A. Guzei, L. M. Liable-Sands, M. J. Heeg, A. L. Rheingold and C. H. Winter, J. Organomet. Chem., 1999, 588, 167-175. 321. D. P. Mills, L. Soutar, O. J. Cooper, W. Lewis, A. J. Blake and S. T. Liddle, Organometallics, 2012, 32, 1251-1264. 322. Z. R. Turner, R. Bellabarba, R. P. Tooze and P. L. Arnold, J. Am. Chem. Soc., 2010, 132, 4050-4051. 323. B. Liu and D. Cui, Dalton Trans., 2009, 550-556. 324. S.-O. Hauber and M. Niemeyer, Inorg. Chem., 2005, 44, 8644-8646. 325. H. S. Lee and M. Niemeyer, Inorg. Chem., 2010, 49, 730-735. 326. G. Paolucci, P. Zanella and A. Berton, J. Organomet. Chem., 1985, 295, 317- 326. 327. W. J. Evans, J. R. Walensky and J. W. Ziller, Organometallics, 2009, 29, 101- 107. 328. K. Izod, S. T. Liddle and W. Clegg, Inorg. Chem., 2003, 43, 214-218. 329. J. H. Freeman and M. L. Smith, J. Inorg. Nucl. Chem., 1958, 7, 224-227. 330. M. D. Taylor and C. P. Carter, J. Inorg. Nucl. Chem., 1962, 24, 387-391. 331. C. M. Cassani, Y. K. Gun'ko, P. B. Hitchcock, A. G. Hulkes, A. V. Khvostov, M. F. Lappert and A. V. Protchenko, J. Organomet. Chem., 2002, 647, 71-83. 332. S. Al-Juaid, Y. K. Gun’ko, P. B. Hitchcock, M. F. Lappert and S. Tian, J. Organomet. Chem., 1999, 582, 143-152. 333. E. Sheng, S. Wang, G. Yang, S. Zhou, L. Cheng, K. Zhang and Z. Huang, Organometallics, 2003, 22, 684-692. 334. S.-L. Zhou, S.-W. Wang, G.-S. Yang, X.-Y. Liu, E.-H. Sheng, K.-H. Zhang, L. Cheng and Z.-X. Huang, Polyhedron, 2003, 22, 1019-1024. 335. H.-X. Li, Q.-F. Xu, J.-X. Chen, M.-L. Cheng, Y. Zhang, W.-H. Zhang, J.-P. Lang and Q. Shen, J. Organomet. Chem., 2004, 689, 3438-3448. 336. B. Martı́n-Vaca, A. Dumitrescu, H. Gornitzka, D. Bourissou and G. Bertrand, J. Organomet. Chem., 2003, 682, 263-266.

225

337. F. T. Edelmann, A. Steiner, D. Stalke, J. W. Gilje, S. Jagner and M. Håkansson, Polyhedron, 1994, 13, 539-546. 338. T. Chivers, A. Downard and M. Parvez, Inorg. Chem., 1999, 38, 4347-4353. 339. W. M. Boesveld and M. F. Lappert, Chem. Commun., 1997, 2091-2092. 340. M. L. Cole, P. C. Junk and L. M. Louis, J. Chem. Soc., Dalton Trans., 2002, 3906-3914. 341. J. Baldamus, M. L. Cole, U. Helmstedt, E.-M. Hey-Hawkins, C. Jones, P. C. Junk, F. Lange and N. A. Smithies, J. Organomet. Chem., 2003, 665, 33-42. 342. D. M. King, W. Lewis and S. T. Liddle, Inorg. Chim. Acta, 2012, 380, 167-173. 343. R. R. Fraser, T. S. Mansour and S. Savard, J. Org. Chem., 1985, 50, 3232-3234. 344. C. A. Tolman, J. Am. Chem. Soc., 1970, 92, 2956-2965. 345. K. A. Bunten, L. Chen, A. L. Fernandez and A. J. Poë, Coord. Chem. Rev., 2002, 233–234, 41-51. 346. J. A. Bilbrey, A. H. Kazez, J. Locklin and W. D. Allen, J. Comput. Chem., 2013, 34, 1189-1197. 347. T. L. Brown and K. J. Lee, Coord. Chem. Rev., 1993, 128, 89-116. 348. I. A. Guzei and M. Wendt, Dalton Trans., 2006, 3991-3999. 349. S. D. Stults, R. A. Andersen and A. Zalkin, Organometallics, 1990, 9, 115-122. 350. W. J. Evans, C. A. Seibel and J. W. Ziller, J. Am. Chem. Soc., 1998, 120, 6745- 6752. 351. Z. Xie, K. Chui, Z. Liu, F. Xue, Z. Zhang, T. C. W. Mak and J. Sun, J. Organomet. Chem., 1997, 549, 239-244. 352. W. J. Evans, S. L. Gonzales and J. W. Ziller, J. Am. Chem. Soc., 1991, 113, 7423-7424. 353. W. J. Evans, R. A. Keyer and J. W. Ziller, J. Organomet. Chem., 1990, 394, 87- 97. 354. M. P. Coles, P. B. Hitchcock, M. F. Lappert and A. V. Protchenko, Organometallics, 2011, 31, 2682-2690. 355. Y. K. Gun'ko and P. B. Hitchcock, Chem. Commun., 1998, 1843-1844. 356. H. Schumann, M. Glanz, H. Hemling and F. H. Görlitz, J. Organomet. Chem., 1993, 462, 155-161. 357. W. J. Evans, T. M. Champagne and J. W. Ziller, J. Am. Chem. Soc., 2006, 128, 14270-14271. 358. R. J. Errington, Advanced Practical Inorganic and Metalorganic Chemistry, CRC Press, Boca Raton, 1997. 359. D. F. Shriver and M. A. Drezdzon, The Manipulation of Air-Sensitive Compounds, 2nd edn., John Wiley & Sons, New York, 1986. 360. W. J. Evans, S. A. Kozimor and J. W. Ziller, Polyhedron, 2004, 23, 2689-2694. 361. G. Lin and W.-T. Wong, Polyhedron, 1995, 14, 3167-3169. 362. W. J. Evans, B. L. Davis, G. W. Nyce, J. M. Perotti and J. W. Ziller, J. Organomet. Chem., 2003, 677, 89-95. 363. J. F. Corbey, M. Fang, J. W. Ziller and W. J. Evans, Inorg. Chem., 2014, 54, 801-807. 364. Z. Xie, Z. Liu, K. Chui, F. Xue and T. C. W. Mak, J. Organomet. Chem., 1999, 588, 78-82. 365. W. W. Lukens, S. M. Beshouri, L. L. Blosch and R. A. Andersen, J. Am. Chem. Soc., 1996, 118, 901-902. 366. B.-J. Deelman, M. Booij, A. Meetsma, J. H. Teuben, H. Kooijman and A. L. Spek, Organometallics, 1995, 14, 2306-2317.

226

367. A. G. M. Barrett, M. R. Crimmin, M. S. Hill, P. B. Hitchcock, S. L. Lomas, P. A. Procopiou and K. Suntharalingam, Chem. Commun., 2009, 2299-2301. 368. M. L. Cole and P. C. Junk, New J. Chem., 2005, 29, 135-140. 369. Y. Q. Jia, J. Solid State Chem., 1991, 95, 184-187. 370. D. Schneider, T. Spallek, C. Maichle-Mossmer, K. W. Tornroos and R. Anwander, Chem. Commun., 2014, 50, 14763-14766. 371. M. Gregson, E. Lu, J. McMaster, W. Lewis, A. J. Blake and S. T. Liddle, Angew. Chem., Int. Ed., 2013, 52, 13016-13019. 372. J. Schläfer, S. Stucky, W. Tyrra and S. Mathur, Inorg. Chem., 2013, 52, 4002- 4010. 373. J. Schläfer, W. Tyrra and S. Mathur, Inorg. Chem., 2014, 53, 2751-2753. 374. A. J. Lewis, U. J. Williams, P. J. Carroll and E. J. Schelter, Inorg. Chem., 2013, 52, 7326-7328. 375. W. J. Evans, D. S. Lee, D. B. Rego, J. M. Perotti, S. A. Kozimor, E. K. Moore and J. W. Ziller, J. Am. Chem. Soc., 2004, 126, 14574-14582. 376. J. W. S. Rees, O. Just and D. S. Van Derveer, J. Mater. Chem., 1999, 9, 249- 252. 377. R. R. Fraser and T. S. Mansour, J. Org. Chem., 1984, 49, 3442-3443. 378. J. Eppinger, M. Spiegler, W. Hieringer, W. A. Herrmann and R. Anwander, J. Am. Chem. Soc., 2000, 122, 3080-3096. 379. M. Lappert, A. Protchenko, P. Power and A. Seeber, in Metal Amide Chemistry, John Wiley & Sons, Ltd, 2008, pp. 79-120. 380. S. Anfang, M. Karl, N. Faza, W. Massa, K. Dehnicke and J. Magull, Z. Anorg. Allg. Chem., 1997, 623, 1425-1432. 381. L. Huebner, A. Kornienko, T. J. Emge and J. G. Brennan, Inorg. Chem., 2004, 43, 5659-5664. 382. A. J. Gaunt, S. D. Reilly, A. E. Enriquez, T. W. Hayton, J. M. Boncella, B. L. Scott and M. P. Neu, Inorg. Chem., 2008, 47, 8412-8419. 383. M. Westerhausen, M. Gärtner, R. Fischer and J. Langer, Angew. Chem., Int. Ed., 2007, 46, 1950-1956. 384. M. Westerhausen, M. Gärtner, R. Fischer, J. Langer, L. Yu and M. Reiher, Chem. Eur. J., 2007, 13, 6292-6306. 385. M. Westerhausen, Coord. Chem. Rev., 2008, 252, 1516-1531. 386. M. Westerhausen, Z. Anorg. Allg. Chem., 2009, 635, 13-32. 387. J. Langer, S. Krieck, H. Gorls, G. Kreisel, W. Seidel and M. Westerhausen, New J. Chem., 2010, 34, 1667-1677. 388. R. Fischer, H. Görls and M. Westerhausen, Inorg. Chem. Commun., 2005, 8, 1159-1161. 389. R. Fischer, M. Gärtner, H. Görls and M. Westerhausen, Organometallics, 2006, 25, 3496-3500. 390. H. Gilman and N. B. St. John, Recl. Trav. Chim. Pays-Bas, 1930, 49, 717-723. 391. M. Köhler, J. Langer, H. Görls and M. Westerhausen, Organometallics, 2012, 31, 8647-8653. 392. G. Bruhat and V. Thomas, C. R. Acad. Sci., 1926, 183, 297-299. 393. R. Fischer, M. Gärtner, H. Görls and M. Westerhausen, Angew. Chem., Int. Ed., 2006, 45, 609-612. 394. J. Langer, M. Köhler, H. Görls and M. Westerhausen, Chem. Eur. J., 2014, 20, 3154-3161. 395. J. Langer, S. Krieck, H. Görls and M. Westerhausen, Angew. Chem., Int. Ed., 2009, 48, 5741-5744. 227

396. W. Schlenk and W. Schlenk, Ber. Dtsch. Chem. Ges., 1929, 62, 920-924. 397. R. Fischer, M. Gärtner, H. Görls, L. Yu, M. Reiher and M. Westerhausen, Angew. Chem., Int. Ed., 2007, 46, 1618-1623. 398. S. A. Cotton, Coord. Chem. Rev., 1997, 160, 93-127. 399. C. Eaborn, K. Izod and J. D. Smith, J. Organomet. Chem., 1995, 500, 89-99. 400. P. B. Hitchcock, A. V. Khvostov and M. F. Lappert, J. Organomet. Chem., 2002, 663, 263-268. 401. P. B. Hitchcock, S. A. Holmes, M. F. Lappert and S. Tian, J. Chem. Soc., Chem. Commun., 1994, 2691-2692. 402. J. R. van den Hende, P. B. Hitchcock, S. A. Holmes, M. F. Lappert and S. Tian, J. Chem. Soc., Dalton Trans., 1995, 3933-3939. 403. C. Eaborn, P. B. Hitchcock, K. Izod, Z.-R. Lu and J. D. Smith, Organometallics, 1996, 15, 4783-4790. 404. C. Eaborn, P. B. Hitchcock, K. Izod and J. D. Smith, J. Am. Chem. Soc., 1994, 116, 12071-12072. 405. T. A. Starostina, R. R. Shifrina, L. F. Rybakova and E. S. Petrov, Zh. Obshch. Khim., 1987, 57, 2402. 406. M. N. Bochkarev, A. A. Trifonov, E. A. Fedorova, N. S. Emelyanova, T. A. Basalgina, G. S. Kalinina and G. A. Razuvaev, J. Organomet. Chem., 1989, 372, 217-224. 407. M. N. Bochkarev, V. V. Khramenkov, Y. F. Rad'kov, L. N. Zakharov and Y. T. Struchkov, J. Organomet. Chem., 1992, 429, 27-39. 408. G. B. Deacon and D. G. Vince, J. Organomet. Chem., 1976, 112, C1-C2. 409. G. B. Deacon, W. D. Raverty and D. G. Vince, J. Organomet. Chem., 1977, 135, 103-114. 410. C. M. Forsyth and G. B. Deacon, Organometallics, 2000, 19, 1205-1207. 411. G. B. Deacon and C. M. Forsyth, Organometallics, 2003, 22, 1349-1352. 412. G. Heckmann and M. Niemeyer, J. Am. Chem. Soc., 2000, 122, 4227-4228. 413. M. Niemeyer, Acta Crystallogr,. Sect. E.-Struct Rep. Online, 2001, 57, m578- m580. 414. D. F. Evans, G. V. Fazakerley and R. F. Phillips, Journal of the Chemical Society A: Inorganic, Physical, Theoretical, 1971, 1931-1934. 415. G. B. Deacon and C. M. Forsyth, Chem. Eur. J., 2004, 10, 1798-1804. 416. D. F. Evans, G. V. Fazakerley and R. F. Phillips, J. Chem. Soc. D, Chem. Commun., 1970, 244-244. 417. K. Yokoo, Y. Fujiwara, T. Fukagawa and H. Taniguchi, Polyhedron, 1983, 2, 1101-1102. 418. M. Wiecko, G. B. Deacon and P. C. Junk, Chem. Commun., 2010, 46, 5076- 5078. 419. T. P. Hanusa, in Comprehensive Organometallic Chemistry III, ed. D. M. P. M. H. Crabtree, Elsevier, Oxford, 2007, pp. 67-152. 420. J. K. Gibson, The Journal of Physical Chemistry A, 2003, 107, 7891-7899. 421. M. Vasiliu, D. Feller, J. L. Gole and D. A. Dixon, J. Phys. Chem. A, 2010, 114, 9349-9358. 422. S. Marks, J. G. Heck, M. H. Habicht, P. Oña-Burgos, C. Feldmann and P. W. Roesky, J. Am. Chem. Soc., 2012, 134, 16983-16986. 423. B. Liu, V. Dorcet, L. Maron, J.-F. Carpentier and Y. Sarazin, Eur. J. Inorg. Chem., 2012, 2012, 3023-3031. 424. M. G. Cushion and P. Mountford, Chem. Commun., 2011, 47, 2276-2278.

228

425. S. Krieck, H. Görls and M. Westerhausen, Organometallics, 2008, 27, 5052- 5057. 426. W. J. Evans, M. A. Johnston, M. A. Greci, T. S. Gummersheimer and J. W. Ziller, Polyhedron, 2003, 22, 119-126. 427. S. Harder, M. Lutz and A. W. G. Straub, Organometallics, 1997, 16, 107-113. 428. A. A. Trifonov, E. N. Kirillov, S. Dechert, H. Schumann and M. N. Bochkarev, Eur. J. Inorg. Chem., 2001, 2001, 2509-2514. 429. G. K. B. Clentsmith, F. G. N. Cloke, M. D. Francis, J. R. Hanks, P. B. Hitchcock and J. F. Nixon, J. Organomet. Chem., 2008, 693, 2287-2292. 430. S. Harder, F. Feil and T. Repo, Chem. Eur. J., 2002, 8, 1991-1999. 431. S. Harder, Angew. Chem., Int. Ed., 2004, 43, 2714-2718. 432. G. B. Deacon, C. M. Forsyth, F. Jaroschik, P. C. Junk, D. L. Kay, T. Maschmeyer, A. F. Masters, J. Wang and L. D. Field, Organometallics, 2008, 27, 4772-4778. 433. S. M. Mansell, M. Green, R. J. Kilby, M. Murray and C. A. Russell, C. R. Chim., 2010, 13, 1073-1081. 434. Rei, P. Poremba, L. Lameyer, D. Stalke and F. T. Edelmann, Chem. Commun., 1999, 1865-1866. 435. L. Ungur, J. J. Le Roy, I. Korobkov, M. Murugesu and L. F. Chibotaru, Angew. Chem., Int. Ed., 2014, 53, 4413-4417. 436. J. J. Le Roy, L. Ungur, I. Korobkov, L. F. Chibotaru and M. Murugesu, J. Am. Chem. Soc., 2014, 136, 8003-8010. 437. V. Lorenz, S. Blaurock, C. G. Hrib and F. T. Edelmann, Organometallics, 2010, 29, 4787-4789. 438. A. Edelmann, V. Lorenz, C. G. Hrib, L. Hilfert, S. Blaurock and F. T. Edelmann, Organometallics, 2012, 32, 1435-1444. 439. J. J. Le Roy, M. Jeletic, S. I. Gorelsky, I. Korobkov, L. Ungur, L. F. Chibotaru and M. Murugesu, J. Am. Chem. Soc., 2013, 135, 3502-3510. 440. A. Edelmann, S. Blaurock, V. Lorenz, L. Hilfert and F. T. Edelmann, Angew. Chem., Int. Ed., 2007, 46, 6732-6734. 441. K. F. Tesh, D. J. Burkey and T. P. Hanusa, J. Am. Chem. Soc., 1994, 116, 2409- 2417. 442. R. R. Fraser, M. Bresse and T. S. Mansour, J. Chem. Soc., Chem. Commun., 1983, 620-621. 443. R. R. Fraser and T. S. Mansour, J. Organomet. Chem., 1986, 310, C60-C62. 444. H. Ahlbrecht and G. Scheneider, Tetrahedron, 1986, 42, 4729-4741. 445. M. Gomberg, J. Am. Chem. Soc., 1900, 22, 757-771. 446. M. Gomberg and W. E. Bachmann, J. Am. Chem. Soc., 1930, 52, 2455-2461. 447. H. M. Walborsky and M. Topolski, J. Am. Chem. Soc., 1992, 114, 3455-3459. 448. E. C. Ashby and J. Oswald, J. Org. Chem., 1988, 53, 6068-6076. 449. K. Maruyama and T. Katagiri, J. Phys. Org. Chem., 1989, 2, 205-213. 450. K. S. Root, C. L. Hill, L. M. Lawrence and G. M. Whitesides, J. Am. Chem. Soc., 1989, 111, 5405-5412. 451. L. M. Lawrence and G. M. Whitesides, J. Am. Chem. Soc., 1980, 102, 2493- 2494. 452. E. A. Vogler, R. L. Stein and J. M. Hayes, J. Am. Chem. Soc., 1978, 100, 3163- 3166. 453. S. R. Drake and D. J. Otway, J. Chem. Soc., Chem. Commun., 1991, 517-519. 454. S. Harder, S. Müller and E. Hübner, Organometallics, 2003, 23, 178-183.

229

455. M. A. Guino-o, A. Torvisco, W. Teng and K. Ruhlandt-Senge, Inorg. Chim. Acta, 2012, 389, 122-130. 456. R. L. Letsinger, R. Collat and M. Magnusson, J. Am. Chem. Soc., 1954, 76, 4185-4186. 457. S. Krieck, H. Görls and M. Westerhausen, Chem. Asian J., 2010, 5, 272-277. 458. E. D. Brady, T. P. Hanusa, M. Pink and V. G. Young, Inorg. Chem., 2000, 39, 6028-6037. 459. A. A. Trifonov, E. A. Fedorova, G. K. Fukin, N. O. Druzhkov and M. N. Bochkarev, Angew. Chem., Int. Ed., 2004, 43, 5045-5048. 460. H. Nakamura, Y. Nakayama, H. Yasuda, T. Maruo, N. Kanehisa and Y. Kai, Organometallics, 2000, 19, 5392-5399. 461. T. R. Hoye, B. M. Eklov and M. Voloshin, Org. Lett., 2004, 6, 2567-2570. 462. J. Leonard, B. Lygo and G. Procter, Advanced Practical Organic Chemistry, 2nd edn., CRC Press, Boca Raton, 1994. 463. J. Jubb, P. Berno, S. Hao and S. Gambarotta, Inorg. Chem., 1995, 34, 3563- 3566. 464. G. M. Sheldrick, Acta Crystallogr., Sect. A, 2008, 64, 112-122. 465. L. J. Barbour, J. Supramol. Chem., 2001, 1, 189-191.

230