New Lithium Cuprates
for the Promotion of
Directed Organic Transformations
Philip James Harford
Fitzwilliam College, University of Cambridge
This dissertation is submitted for the degree of Doctor of Philosophy
July 2014
Declaration
This dissertation is submitted to the Board of Graduate Studies in partial fulfilment
of the requirements for the qualification of Doctor of Philosophy at the University
of Cambridge. Research presented herein was carried out by the author at the
University Chemical Laboratories between October 2010 and April 2014. Except
where specific reference is made to the contrary, it is original work and contains
nothing that is the outcome of work done in collaboration. Neither the whole
nor any part of this work has been submitted before for a degree in any other
institution. This thesis does not exceed 60,000 words including tables, footnotes
and the bibliography.
Philip Harford
July 30th, 2014
i
You call me out upon the waters
The great unknown where feet may fail
And there I find You in the mystery
In oceans deep my faith will stand
Oceans - Matt Crocker, Joel Houston and Salomon Lighthelm (2013)
ii
Acknowledgements
Firstly, thanks must go to my supervisor Andrew Wheatley for his provision of
advice, expertise and encouragement throughout my time as a PhD student.
I’m very grateful to the members of the Wheatley group for all their friendship
and support, in particular Jo Haywood who was so helpful in pointing me in the
right direction when I started working on the project and for having the patience to
teach me how to solve crystal structures! There have also been several Part III and
summer vacation students who have contributed to this project so enthusiastically
and insightfully, Jack Rowbotham, Matt Smith, Andy Peel and Joe Taylor. There
have been many others from the Wright and Boss/Barker groups who have made
my time in Lab 301 so enjoyable, in particular Spud Page, Tom Scrase and Fran
Stokes. It really has been a lot of fun working (and, er, not working) with all of you.
A significant part of my project has involved X-ray crystallography, which would
not have been possible without the expertise and patience of John Davies, who
collected the majority of the data here in Cambridge, Paul Wood, Paul Raithby
(University of Bath) and Aggi Steiner (University of Liverpool). Peter Grice, An-
drew Mason and Duncan Howe from the NMR service have been a great help not
only running experiments but also providing training and arranging analysis on air-
sensitive samples. Thanks must also go to Ali Campbell Smith and Rebecca Melen
who helped collect NMR spectra. In addition, I am grateful to Patricia Irele and
Alan Dickerson for running elemental analysis on countless samples, particularly
the halide analyses which just did not want to play ball.
Much of the work presented in Chapter 5 was carried out in collaboration with Prof.
Florence Mongin (University of Rennes 1, France) who along with the members of
her group, in particular Tan Tai Nguyen, were very welcoming when I visited
their laboratories. Their expertise in organic chemistry and assistance in running
experiments has been an invaluable help.
iii
The theoretical calculations were carried in collaboration with Prof. Masanobu
Uchiyama and Dr. Shinsuke Komagawa (RIKEN, Japan) who have also contributed
to useful discussions and advised on organic reactions. I am also very thankful
to Prof. Yoshinori Kondo for inviting me to Sendai, Japan and the very kind
hospitality he and Misato Kobayashi provided for the duration of my trip.
Finally, I would like to thank my friends, parents and the rest of my family who
have been so supportive throughout.
iv
Publications
The work carried out during my PhD has contributed to the following papers:
Structural Effects in Lithiocuprate Chemistry: the Elucidation of Reactive Pen-
tametallic Complexes, P. J. Harford, A. J. Peel, J. P. Taylor, S. Komagawa, P.
R. Raithby, T. P. Robinson, M. Uchiyama and A. E. H. Wheatley, Chem Eur J
20(14); 3908–3912, (2014)
Synthesis of azafluorenones and related compounds using deprotocupration-aroyl-
ation followed by intramolecular direct arylation, N. Marquise, P. J. Harford, F.
Chevallier, T. Roisnel, V. Dorcet, A. L. Gagez, S. Sable´, L. Picot, V. Thie´ry, A. E.
H. Wheatley, P. C. Gros and F. Mongin, Tetrahedron, 69(47); 10123–10133, (2013)
Efficient two-step access to azafluorenones and related compounds, N. Marquise,
P. J. Harford, F. Chevallier, T. Roisnel, A. E. H. Wheatley, P. C. Gros and F.
Mongin, Tetrahedron Lett, 54(24); 3154–3157, (2013)
Expanding the tools available for direct ortho cupration-targeting lithium phosphi-
docuprates, P. J. Harford, J. Haywood, M. R. Smith, B. N. Bhawal, P. R. Raithby,
M. Uchiyama and A. E. H Wheatley, Dalton Trans, 41(20); 6148–6154, (2012)
Deprotonative Metalation of Chloro- and Bromopyridines Using Amido-Based Bi-
metallic Species and Regioselectivity-Computed CH Acidity Relationships, K. Sne´-
garoff, T. T. Nguyen, N. Marquise, Y. S. Halauko, P. J. Harford, T. Roisnel, V. E.
Matulis, O. A. Ivashkevich, F. Chevallier, A. E. H. Wheatley, P. C. Gros and F.
Mongin, Chem Eur J, 17(47); 13284–13297, (2011)
v
I have also contributed to the following papers and book chapters over the course
of my PhD:
Amidocuprates for Directed ortho Cupration: Structural Study, Mechanistic Investi-
gation, and Chemical Requirements, S. Komagawa, S. Usui, J. Haywood, P. J. Har-
ford, A. E. H. Wheatley, Y. Matsumoto, K. Hirano, R. Takita and M. Uchiyama,
Angew Chem Int Ed, 51(48); 12081–12085, (2012)
Lithiated Tertiary Carbanions Display Variable Coordination Modes: Evidence
from DFT and NMR Studies, M. A. Vincent, A. Campbell Smith, M. Donnard, P.
J. Harford, J. Haywood, I. H. Hillier, J. Clayden and A. E. H. Wheatley, Chem
Eur J, 18(35); 11036–11045, (2012)
The redox effect of the [1,2-(NH)
2
C6H4]
2– ligand in the formation of transition
metal compounds, F. A. Stokes, L. Kloo, Y. Lv, P. J. Harford, A. E. H. Wheatley
and D. S. Wright, Chem Commun, 48(92), 11298–11300, (2012)
New avenues in the directed deprotometallation of aromatics: recent advances in
directed cupration, A. E. H. Wheatley, P. J. Harford, A. J. Peel, F. Chevallier, R.
Takita, F Mongin and M. Uchiyama, Dalton Trans, in press
Alkali/coinage metals - organolithium, organocuprate chemistry, P. J. Harford and
A. E. H. Wheatley, SPR: Organometallic Chemistry, 38; 91–111, (2012)
Alkali/Coinage metals - organolithium, organocuprate chemistry, P. J. Harford, A.
J. Peel and A. E. H. Wheatley, SPR: Organometallic Chemistry, 39; 165–193 (2014)
vi
Abbreviations
12-c-4 12-crown-4 (C8H16O4)
Ar general aryl group
ax axial
nBu normal -butyl (CH2(CH2)2(CH3))
tBu tert-butyl (C(CH3)3)
CIPE Complex Induced Proximity Effect
CoSHH Control of Substances Hazardous to Health
Cp cyclopentadienyl (C5H5)
CSD Cambridge Structural Database
Cy cyclohexyl (C6H11)
DATMP diethylaluminium 2,2,6,6-tetramethylpiperidide
DIBA diisobutylamine
DFT Density Functional Theory
DG directing group
DMS dimethylsulfide (S(CH3)2)
DoAl Directed ortho Alumination
DoCu Directed ortho Cupration
DoLi Directed ortho Lithiation
DoM Directed ortho Metalation
DoMg Directed ortho Magnesiation
DoMn Directed ortho Manganation
DoZn Directed ortho Zincation
Et ethyl (C2H5)
vii
eq equatorial
FT-IR Fourier Transform Infra-Red
GIAO Gauge-Independent Atomic Orbital
GMF glass microfibre
Hal halide
HDMP cis-2,6-dimethylpiperidine
HMDS hexamethyldisiylyamine (HN[Si(CH3)3]2)
HTMP 2,2,6,6-tetramethylpiperidine
LDA lithium diisopropylamide (LiN(CH(CH3)2)2
LiDMP lithium cis-2,6-dimethylpiperidide
LiHMDS lithium hexamethyldisiylyamide
LiTMP lithium 2,2,6,6-tetramethylpiperidide
M general metal ion
Me methyl (CH3)
Mes mesityl, 2,4,6-trimethylphenyl (C9H11)
MNDO Modified Neglect of Differential Overlap
NHC N-heterocylic carbene
NMR nuclear magnetic resonance
Ph phenyl (C6H5)
Phen 1,10-phenanthroline
Phos 1-phenyl-2,5-bis(2-pyridyl)phosphole
PMDTA N,N,N′,N′′,N′′-pentamethyldiethylenetriamine
((CH3)2N(CH2)2N(CH3)(CH2)2N(CH3)2)
iPr iso-propyl (CH(CH3)2)
PPE personal protective equipment
PTFE polytetrafluoroethylene
R general organic group
S general solvent ligand
viii
Th thiolate (SR)
THF tetrahydrofuran (C4H8O)
THP tetrahydropyran (C5H10O)
THT tetrahydrothiophene (C4H8S)
TMEDA N,N,N′,N′-tetramethylethylenediamine ((CH3)2N(CH2)2N(CH3)2)
TMS trimethylsilyl (Si(CH3)3)
Trip 2,4,6-triisopropylphenyl
δ chemical shift
ppm parts per million
s singlet
d doublet
t triplet
q quartet
quint quintet
hept heptet
m multiplet
br broad
M molecular weight of asymmetric unit
a, b, c unit cell dimensions
α, β, γ unit cell angles
T temperature
λ wavelength
V unit cell volume
µ linear absorption coefficient
Z number of formula units per unit cell
ρcalcd calculated density
ix
e electron
R R-factor (unweighted)
wR2 R-factor (weighted)
S Goodness of Fit
θ angle of data collection
asu asymmetric unit
x
Definition of Terms
Throughout this dissertation the terms Gilman-type and Lipshutz-type have been
used in the following manner:
Gilman-type cuprate refers to a lower-order cuprate containing two organic groups
and a single lithium, exclusive of an LiX (X = R′′, CN, halide) unit.
Lipshutz-type cuprate refers to a higher-order cuprate, inclusive of an LiX unit,
forming a six- or seven-membered ring. The term Lipshutz-type cuprate includes
the group of compounds referred to as Lipshutz cuprates where, specifically, X =
CN.
R Cu R'
LiLi
X
(a) Gilman-type
R Cu R'
LiLi
X
(b) Lipshutz-type
xi
Definition of Numbered Compounds
1 [(TMP)2Cu(CN)Li2 · THF]2
2 [(TMP)2CuLi]2
3 [(TMP)2Cu(I)Li2 · THF]2
4 PhCu(µ-TMP)Li · 3THF
5 MeCu(µ-TMP)Li · TMEDA
6 2-iodo-N,N -diisopropylbenzamide
7 [(tBu2P)2Cu]
– [Li · 2THF]+
8 MeCuP(tBu)2Li · 3THF
9 [(Ph2P)6Cu4][Li · 4THF]2
10 [Ph2PCu(CN)Li · 2THF]
11 [(Cy2P)2CuLi · 2THF]
12 [(TMP)2Cu(Cl)Li2 · THF]2
13 [(TMP)2Cu(Br)Li2 · THF]2
14 [(TMP)2Cu(CN)Li2 ·OEt2]2
15 [(TMP)2Cu(I)Li2 ·OEt2]2
16 [(TMP)2Cu(Cl)Li2 ·OEt2]2
17 [(TMP)2Cu(Br)Li2 ·OEt2]2
18 [(DMP)2CuLi · Et2O]1.29 · [(DMP)2Li2 · Et2O]0.71LiCl
19 [(DMP)2CuLi ·OEt2]2LiCl
20 [(DMP)2CuLi ·OEt2]1.45[(DMP)2CuLi · HDMP]0.55LiBr
21 [(DMP)2CuLi ·OEt2]2LiBr
22 [(DMP)2CuLi ·OEt2]2LiI
23 [(DMP)2CuLi · 2THF]2LiBr
24 2-bromopyridine
25 2-chloropyridine
xii
Abstract
The use of bimetallic bases to effect highly stereoselective organic transformations
via Directed ortho Metalation (DoM) is an important synthetic tool and requires
thorough investigation to maximise efficiency and develop a mechanistic under-
standing. This thesis begins with a discussion of the history of DoM covering
firstly monometallic lithium bases before moving to bimetallic magnesiate, zin-
cate, aluminate, manganate bases and finally covering cuprate bases, which are
the focus of this thesis.
A description of the techniques employed in this project, which focuses on X-ray
crystallography as the primary characterisation tool, follows. The experimental
procedures and the associated characterisation data are also presented. These
demonstrate that the addition of 2 equivalents of an amido or phosphido lith-
ium to a copper(I) salt results in the crystallisation of a series of bis(amido)- and
bis(phosphido)cuprates, and phosphidocopper compounds, many of which are suit-
able Directed ortho Cupration (DoCu) reagents.
Syntheses involving phosphido ligands lead to the isolation of only one lithium
cuprate species, Gilman-type [(Cy2P)2CuLi · 2THF], 11, which forms a polymer
in the solid-state and is an ineffective DoCu agent. Alternative pathways which
yield interesting novel phosphidocopper compounds [(Ph2P)6Cu4][Li · 4THF]2, 9,
and Ph2PCu(CN)Li · 2THF, 10, are also investigated. The solid-state structures
of all three species are discussed in detail.
In situ preparations of bis(amido)cuprate bases from CuICl are shown to be effec-
tive in the deprotocupration of halopyridines. Reactions carried out with and with-
out LiCl present in the mixture demonstrate that Lipshutz-type formulation bases
are the active species. Analogous preparations from CuIBr are also effective in ortho
deprotonation although do not provide consistently high-yields. In the solid-state,
both bases are shown to form Lipshutz-type cuprates, [(TMP)2Cu(X)Li2 · THF]2
(X = Cl, Br).
The effects of changing the donor solvent to the weaker Lewis base diethyl ether on
the syntheses of Lipshutz-type cuprates are investigated, resulting in the isolation
of [(TMP)2Cu(X)Li2 · Et2O]2 (X = CN, Hal). The results show that the solid-
xiii
state structures are very similar to the analogous THF-solvated species. However,
when the recrystallisation step is carried in bulk Et2O Lipshutz-type cuprates form,
unlike when THF is employed which results in the formation of the Gilman-type
cuprate [(TMP)2CuLi]2, 2.
Changing the amido ligand to the less sterically demanding cis-2,6-dimethylpiperi-
dine (HDMP) results in the formation of a remarkable new pentametallic struc-
tural motif, [(DMP)2CuLi · OEt2]2LiX (X = Hal), which can be viewed as an
adduct between monomeric Lipshutz- and Gilman-type cuprates. The isolation of
[(DMP)2CuLi · 2THF]2LiBr, 23 shows that adduct-type cuprates form regardless
of the nature of the donor solvent and plausible explanations for the preference of
DMP-based cuprates to form adduct-type species rather than Lipshutz-type spe-
cies are presented. Reactivity studies demonstrate that 23 is an effective DoCu
agent and theoretical studies explore possible mechanisms as well as the relative
energies of the three cuprate structure types.
The thesis is completed with a summary of the conclusions drawn and suggestions
for further work. Ideas for further solid-state (including extensive investigations
into the effects of sterics on the formation of cuprates), solution-state, reactivity
and theoretical studies are put forward along with the rationale behind them.
xiv
Contents
Declaration i
Acknowledgements iii
Publications v
Abbreviations vii
Definition of Terms xi
Definition of Numbered Compounds xii
Abstract xiii
List of Figures xx
List of Schemes xxiii
1 Introduction 1
1.1 Directed ortho Metalation . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Directed ortho Lithiation . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2.1 Directed ortho Lithiation mechanisms . . . . . . . . . . . . . 4
1.2.2 Limitations of DoLi . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Directed ortho Magnesiation . . . . . . . . . . . . . . . . . . . . . . 8
1.4 Directed ortho Zincation . . . . . . . . . . . . . . . . . . . . . . . . 11
1.5 Directed ortho Alumination . . . . . . . . . . . . . . . . . . . . . . 22
1.6 Directed ortho Manganation . . . . . . . . . . . . . . . . . . . . . . 31
1.7 Limitations of Modern DoM Methods . . . . . . . . . . . . . . . . . 33
xv
1.8 Copper Based Reagents . . . . . . . . . . . . . . . . . . . . . . . . 34
1.8.1 Organocopper compounds . . . . . . . . . . . . . . . . . . . 34
1.8.2 Organocuprates . . . . . . . . . . . . . . . . . . . . . . . . . 38
1.8.2.1 Lower-order organocuprates . . . . . . . . . . . . . 40
1.8.2.2 Higher-order organocuprates . . . . . . . . . . . . . 41
1.8.2.3 Recent investigations into organocuprates . . . . . 44
1.8.2.4 Phosphidocuprates . . . . . . . . . . . . . . . . . . 54
1.9 Aims . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
2 General Experimental Techniques 58
2.1 CoSHH Considerations . . . . . . . . . . . . . . . . . . . . . . . . . 58
2.2 Inert Atmosphere Techniques . . . . . . . . . . . . . . . . . . . . . 58
2.3 Starting Materials and Solvents . . . . . . . . . . . . . . . . . . . . 60
2.4 Melting Point Determination . . . . . . . . . . . . . . . . . . . . . . 61
2.5 Elemental Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
2.6 Multinuclear Nuclear Magnetic Resonance (NMR) Spectroscopy . . 61
2.7 Single Crystal X-ray Diffractometry . . . . . . . . . . . . . . . . . . 62
2.8 Computational Calculations . . . . . . . . . . . . . . . . . . . . . . 63
3 Experimental Procedures and Results 64
3.1 Starting Material and Intermediate Reference Data . . . . . . . . . 64
3.2 Phosphidocuprate and Phosphidocopper Compounds . . . . . . . . 65
3.2.1 Attempted synthesis of dilithium cyano bis(diphenylphosphido)-
cuprate, [(Ph2P)2Cu(CN)Li2 · THF]2 . . . . . . . . . . . . . 65
3.2.2 Attempted synthesis of lithium bis(diphenylphosphido)cuprate,
[(Ph2P)2CuLi]2 . . . . . . . . . . . . . . . . . . . . . . . . . 67
3.2.3 Synthesis of lithium bis(dicyclohexylphosphido)cuprate,
(Cy2P)2CuLi · 2THF, 11 . . . . . . . . . . . . . . . . . . . . 69
3.3 TMP-based Bis(amido)cuprate Compounds . . . . . . . . . . . . . 70
3.3.1 Synthesis of lithium di(2,2,6,6-tetramethylpiperidido)cuprate,
[(TMP)2CuLi]2, 2 . . . . . . . . . . . . . . . . . . . . . . . . 70
3.3.2 Synthesis of dilithium chloro[di(2,2,6,6-tetramethylpiperidido)]-
cuprate, [(TMP)2Cu(Cl)Li2 · THF]2, 12 . . . . . . . . . . . . 71
3.3.3 Synthesis of dilithium bromo[di(2,2,6,6-tetramethylpiperidido)]-
cuprate, [(TMP)2Cu(Br)Li2 · THF]2, 13 . . . . . . . . . . . . 72
3.3.4 Synthesis of dilithium cyano[di(2,2,6,6-tetramethylpiperidido)]-
cuprate, [(TMP)2Cu(CN)Li2 ·OEt2]2, 14 . . . . . . . . . . . 73
xvi
3.3.5 Synthesis of dilithium iodo[di(2,2,6,6-tetramethylpiperidido)]-
cuprate, [(TMP)2Cu(I)Li2 ·OEt2]2, 15 . . . . . . . . . . . . 74
3.3.6 Synthesis of dilithium chloro[di(2,2,6,6-tetramethylpiperidido)]-
cuprate, [(TMP)2Cu(Cl)Li2 ·OEt2]2, 16 . . . . . . . . . . . 76
3.3.7 Synthesis of dilithium bromo[di(2,2,6,6-tetramethylpiperidido)]-
cuprate, [(TMP)2Cu(Br)Li2 ·OEt2]2, 17 . . . . . . . . . . . 77
3.4 DMP-based Amidocuprate Compounds . . . . . . . . . . . . . . . . 78
3.4.1 Attempted synthesis of dilithium chloro[di(cis-2,6,-dimethyl-
piperidido)]cuprate, [(DMP)2Cu(Cl)Li2 ·OEt2]2 . . . . . . . 78
3.4.2 Synthesis of trilithium chloro[tetra(cis-2,6,-dimethylpiperidido)]-
cuprate, [(DMP)2CuLi ·OEt2]2LiCl, 19 . . . . . . . . . . . . 79
3.4.3 Attempted synthesis of trilithium bromo[tetra(cis-2,6,-dimethyl-
piperidido)]cuprate, [(DMP)2CuLi ·OEt2]2LiBr . . . . . . . 81
3.4.4 Synthesis of trilithium bromo[tetra(cis-2,6,-dimethylpiperidido)]-
cuprate, [(DMP)2CuLi ·OEt2]2LiBr, 21 . . . . . . . . . . . . 82
3.4.5 Synthesis of trilithium iodo[tetra(cis-2,6,-dimethylpiperidido)]-
cuprate, [(DMP)2CuLi ·OEt2]2LiI, 22 . . . . . . . . . . . . . 84
3.4.6 Synthesis of trilithium bromo[tetra(cis-2,6,-dimethylpiperidido)]-
cuprate, [(DMP)2CuLi · 2THF]2LiBr, 23 . . . . . . . . . . . 86
3.5 Directed ortho Cupration Reactions . . . . . . . . . . . . . . . . . . 87
3.5.1 Preparation of 2-iodobenzonitrile using an in situ dicyclohexyl-
phospine-based Gilman-type formulation base . . . . . . . . 87
3.5.2 Preparation of 2-iodobenzonitrile using pre-isolated 11 . . . 88
3.5.3 General procedure for the preparation of 2-chloropyridin-3-yl
phenyl ketone using an in situ cuprate base prepared from
CuCl2 · TMEDA . . . . . . . . . . . . . . . . . . . . . . . . 89
3.5.4 General procedure for the preparation of 2-chloropyridin-3-yl
phenyl ketone using an in situ cuprate base prepared from
CuCl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
3.5.5 Preparation of 2-chloropyridin-3-yl phenyl ketone using pre-
isolated 2 in toluene . . . . . . . . . . . . . . . . . . . . . . 90
3.5.6 Preparation of 2-chloropyridin-3-yl phenyl ketone using pre-
isolated 2 in THF . . . . . . . . . . . . . . . . . . . . . . . . 90
3.5.7 Preparation of 2-iodo-N,N -diethylbenzamide using an in situ
DMP-based Lipshutz-type formulation base . . . . . . . . . 91
3.5.8 Preparation of 2-iodo-N,N -diethylbenzamide using pre-isolated
21 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
xvii
4 Phosphidocopper and Phosphidocuprate Compounds 93
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
4.2 [(Ph2P)6Cu4][Li · 4THF]2, 9 . . . . . . . . . . . . . . . . . . . . . . 94
4.3 [Ph2PCu(CN)Li · 2THF]∞, 10∞ . . . . . . . . . . . . . . . . . . . . 96
4.4 Lithium bis(dicyclohexylphosphido)cuprate,
(Cy2P)2CuLi · 2THF, 11 . . . . . . . . . . . . . . . . . . . . . . . . 101
4.5 Solution-state Studies of Phosphidocopper and Phosphidocuprate
Species . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
4.6 Directed ortho Cupration Abilities of Lithium bis(dicyclohexylphos-
phido)cuprate, (Cy2P)2CuLi · 2THF, 11 . . . . . . . . . . . . . . . 106
5 Investigating the Reactivities and Structures of THF-solvated Amido-
cuprates 107
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
5.2 Reactivity Studies of Chloride-Based Lithium Cuprates . . . . . . . 111
5.3 Dilithium chloro[di(2,2,6,6-tetramethylpiperidido)]cuprate,
[(TMP)2Cu(Cl)Li2 · THF]2, 12 . . . . . . . . . . . . . . . . . . . . . 114
5.4 Dilithium bromo[di(2,2,6,6-tetramethylpiperidido)]cuprate,
[(TMP)2Cu(Br)Li2 · THF]2, 13 . . . . . . . . . . . . . . . . . . . . . 116
5.5 Solution Studies of THF-solvated Lipshutz-type Cuprates . . . . . . 119
6 Investigating Solvent Effects in the Formation of Amidocuprates121
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
6.2 Dilithium cyano[di(2,2,6,6-tetramethylpiperidido)]cuprate,
[(TMP)2Cu(CN)Li2 ·OEt2]2, 14 . . . . . . . . . . . . . . . . . . . . 122
6.3 Dilithium iodo[di(2,2,6,6-tetramethylpiperidido)]cuprate,
[(TMP)2Cu(I)Li2 ·OEt2]2, 15 . . . . . . . . . . . . . . . . . . . . . 125
6.4 Diethyl Ether-solvated Lipshutz-type Cuprates Containing Chloride
and Bromide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
7 Pentametallic Adduct-type Amidocuprates 136
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
7.2 Attempted Synthesis of Dilithium chloro[di(cis-2,6-dimethylpiperidido)]-
cuprate, [(DMP)2Cu(Cl)Li2 ·OEt2]2 . . . . . . . . . . . . . . . . . . 137
7.3 Adduct-type Cuprate Structures . . . . . . . . . . . . . . . . . . . . 140
7.3.1 Trilithium chloro[tetra(cis-2,6,-dimethylpiperidido)]cuprate,
[(DMP)2CuLi ·OEt2]2LiCl, 19 . . . . . . . . . . . . . . . . . 140
7.3.2 Adduct-type cuprates containing bromide and iodide . . . . 142
xviii
7.3.3 Trilithium bromo[tetra(cis-2,6,-dimethylpiperidido)]cuprate,
[(DMP)2CuLi · 2THF]2LiBr, 23 . . . . . . . . . . . . . . . . 148
7.4 Solution Studies of Adduct-type Cuprates . . . . . . . . . . . . . . 151
7.5 Structural Differences Between Lipshutz-type and Adduct-type Cu-
prates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
7.6 Directed ortho Cupration Abilities of Adduct-type Cuprates . . . . 155
7.7 Theoretical Calculations . . . . . . . . . . . . . . . . . . . . . . . . 157
8 Conclusions 165
9 Further Work 169
9.1 Future Amidocuprate Species . . . . . . . . . . . . . . . . . . . . . 169
9.2 Further Phosphidocuprate Species . . . . . . . . . . . . . . . . . . . 172
9.3 Further Solution-state Studies . . . . . . . . . . . . . . . . . . . . . 174
9.4 Isolation of an ortho-cuprated Intermediate . . . . . . . . . . . . . . 175
9.5 Further Computational Studies . . . . . . . . . . . . . . . . . . . . 175
9.6 Applications in Directed ortho Metalation . . . . . . . . . . . . . . 176
Bibliography 178
xix
List of Figures
1.1 Classes of Directing Groups for ortho Lithiation . . . . . . . . . . . 3
1.2 Schematic representation of a DoLi on a singly functionalised aro-
matic ring via a CIPE mechanism . . . . . . . . . . . . . . . . . . . 5
1.3 Control of DoLi reactions . . . . . . . . . . . . . . . . . . . . . . . 7
1.4 Molecular structure of Me4MgLi2 · 2TMEDA . . . . . . . . . . . . . 10
1.5 Molecular Structure of the [Bn4MgLi ·TMEDA]
– anion of [Bn4MgLi ·
TMEDA]– [Li · 2TMEDA]+ . . . . . . . . . . . . . . . . . . . . . . 11
1.6 Molecular structures of tBu2Zn(TMP)Li·THF and
tBu2Zn(TMP)Li·
(O−)(Ph)CN(iPr)2 . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.7 Possible transition states for a Directed ortho Zincation reaction . . 14
1.8 Section of the polymeric [(LiHMDS)2 · (Me2Zn · TMEDA)]∞ . . . . 14
1.9 Molecular structure of tBuZn{µ-2-[1-C(O)N(iPr)2]C6H4}2Li ·TMEDA 15
1.10 Summary of deprotonative zincation pathways of the reaction be-
tween N,N -dimethylbenzamide and Me2Zn(Me2N)Li ·OMe2 . . . . 18
1.11 Molecular structure of Zn(µ-C6H4C(O)N-
iPr2-2)3Li · THF . . . . . . 19
1.12 Molecular structure of α-zincated THF . . . . . . . . . . . . . . . . 20
1.13 Molecular structure of tBu2Zn(µ-DMP)Li · TMEDA . . . . . . . . . 21
1.14 Et2Al(TMP)-promoted regioselective deprotonation of arylhydrazones 22
1.15 Deprotonative alumination followed by I2 quench for various func-
tionalised aromatics . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.16 Molecular structures of (iBu)2Al(µ-TMP)(µ-
iBu)Na · TMEDA and
2-iBu2Al{Me2NCH2CH2N(Me)CH2}{C6H4C(−O)N(
iPr)2}
Li · {PhC(−O)N(iPr)2} . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.17 Molecular structure of one of the rotamers of Me3Al(TMP)Li·N,N -
diisopropylbenzamide . . . . . . . . . . . . . . . . . . . . . . . . . . 27
1.18 Summary of deprotonative alumination pathways of the reaction be-
tween N,N -dimethylbenzamide and Me3Al(NMe2)Li ·OMe2 . . . . . 28
xx
1.19 Molecular structures of iBu3Al{C(O)N(
iPr)2(C10H6)}Li · THF and
iBu3Al{C(O)N(
iPr)2(C6H4)}Li · THF . . . . . . . . . . . . . . . . . 30
1.20 Comparison of mechanisms of DoAl and DoZn . . . . . . . . . . . . 30
1.21 Molecular structure of iBu2Zn(µ-
iBu)(µ-DMP)K · PMDTA . . . . . 31
1.22 Molecular structure of (CH2SiMe3)2Mn(µ-TMP)Li · TMEDA . . . . 32
1.23 Molecular structures of (TMP)Mn(µ-TMP)(2-C6H4OMe)Na·TMEDA
and (CH2SiMe3)Mn(µ-TMP){2-[C(O)N(
iPr)2]C6H4}Na · TMEDA . 33
1.24 Molecular structure of Et2NCS2Cu
III(CF3)2 . . . . . . . . . . . . . 35
1.25 Molecular structure of [Cu(CH2SiMe3)]4 . . . . . . . . . . . . . . . 36
1.26 Intramolecular stabilisation of an organocopper compound by a hetero-
atom-containing organic group . . . . . . . . . . . . . . . . . . . . . 37
1.27 Influence of solvent effects on mesitylcopper . . . . . . . . . . . . . 37
1.28 Influence of steric and solvent affects on organocopper aggregation . 38
1.29 Molecular structures of the anionic cages [Ph6Cu5]
– and [Ph6Cu3Li2]
– 40
1.30 Molecular structures of Gilman cuprate anions . . . . . . . . . . . . 41
1.31 Calculated possible structures of lower- and higher-order cuprates . 43
1.32 Suggested bonding in higher-order cuprates . . . . . . . . . . . . . . 43
1.33 Molecular structure of [tBu2Cu]
– [Li2CN · 2THF · 2PMDTA]
+ . . . 44
1.34 Molecular structure of (C6H4CH2NMe2)2Cu(CN)Li2 . . . . . . . . . 44
1.35 Molecular structures of (NPh2)2Cu(NPh2)Li2 · 2OEt2 and
[C6H4{CH2N(Me)CH2CH2NMe2}-2]2Cu(Br)Li2 . . . . . . . . . . . . 45
1.36 Molecular structure of [Mes2CuLi]2 . . . . . . . . . . . . . . . . . . 46
1.37 Molecular structure of [(TMP)2Cu(CN)Li2 · THF]2, 1 . . . . . . . . 48
1.38 Gilman- and Lipshutz-type cuprates prepared from HTMP and LiI 49
1.39 Examples of alkylamido Gilman cuprates . . . . . . . . . . . . . . . 50
1.40 Summary of deprotonative cupration pathways of N,N -dimethyl-
benzamide by MeCu(NMe2)Li · 2OMe2 . . . . . . . . . . . . . . . . 54
1.41 Molecular structure of [(tBu2P)2Cu]
– [Li · 2THF]+, 7 . . . . . . . . 55
1.42 Molecular structure of MeCu(tBu2P)Li · 3THF, 8 . . . . . . . . . . 55
4.1 Structure of 9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
4.2 Structures of the adamantyl cage cores of (Phos)2Cu4Cl4 and
[(Ph2P)10Cd4][Li · 4THF]2 . . . . . . . . . . . . . . . . . . . . . . . 96
4.3 Structure of the asymmetric unit of 10 . . . . . . . . . . . . . . . . 97
4.4 Structure of the extended network formed by 10 . . . . . . . . . . . 98
4.5 Structure of the extended network formed by 10 . . . . . . . . . . . 100
4.6 Structure of [(PPh3)4Cu2(CN)4Li4 · 10THF]
2+ [Ph2Cu]
–
2 . . . . . . . 101
xxi
4.7 Structure of 11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
5.1 Possible species present in cuprate base reaction mixtures prepared
in situ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
5.2 Structure of 12 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
5.3 Structure of 13 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
5.4 Side-on projections of 1, 12, 13, and 3 . . . . . . . . . . . . . . . . 118
6.1 Structure of 14 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
6.2 Structure of 15 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
6.3 Structures of the cores of 3 and 15 . . . . . . . . . . . . . . . . . . 128
6.4 Structure of 16 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
6.5 Structure of 17 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
6.6 Side-on projections of 16, 17 and 15 . . . . . . . . . . . . . . . . . 134
7.1 Structure of 18 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
7.2 Structure of 19 and the central core of 19 . . . . . . . . . . . . . . 141
7.3 Structure of 20 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
7.4 Structure of 21 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
7.5 Structure of 22 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
7.6 Structure of 23 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
7.7 Comparison of the structures of 16 and 19 . . . . . . . . . . . . . . 153
7.8 Structures of Lipshutz-type monomers incorporated in 16 and 19 . 154
7.9 Precluded arrangements of TMP ligands in higher-order cuprate
structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
7.10 Summary of the initial steps in the deprotonative cupration path-
ways of N,N -dimethylbenzamide by LMGM and MeCu(NMe2)(CN)Li2·
2OMe2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
9.1 Structure of [(2-methylpiperidine)2CuLi · 2THF]2LiBr . . . . . . . . 170
9.2 Possible orientations of trans-2,6-dimethylpiperidide ligands in Lipshutz-
and adduct-type cuprates . . . . . . . . . . . . . . . . . . . . . . . 171
9.3 Structure of [(TMP)Cu(DMP)(Br)Li2 · 2THF]2 . . . . . . . . . . . 172
9.4 Possible binding modes of a cyano group in adduct-type cuprate . . 176
xxii
List of Schemes
1.1 General scheme for a directed ortho metalation reaction . . . . . . . 1
1.2 Pathway to DoLi via dilithiation of secondary amines . . . . . . . . 7
1.3 A carbamate functional group undergoing a Fries rearrangement . . 8
1.4 Deprotonation of 3-fluoro-4-pyridine with Bu3MgLi . . . . . . . . . 9
1.5 Deprotonation of halopyridines with Bu3Mg(TMP)Li . . . . . . . . 9
1.6 Deprotonation of functionalised arenes using a TMP-zincate . . . . 12
1.7 Reaction pathways for the addition of an alkylzincate reagent to an
aromatic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.8 Calculated reaction pathway for kinetically unfavourable alkyl-mediated
ortho deprotonation . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.9 Calculated reaction pathway for step 1 (amine elimination) of amine-
mediated ortho deprotonation . . . . . . . . . . . . . . . . . . . . . 17
1.10 Calculated reaction pathway for step 2 (amine re-coordination/alkane
loss) of amine-mediated ortho deprotonation . . . . . . . . . . . . . 17
1.11 Modelled pathways for the deprotonative alumination of anisole by
Me3Al(NMe2)Li ·OMe2 . . . . . . . . . . . . . . . . . . . . . . . . . 29
1.12 Comparison of reactivity of organolithiums and organocuprates . . 41
1.13 Mongin’s in situ synthesis of a TMP-cuprate from CuCl2 · TMEDA 47
1.14 Ease of interconversion between Gilman and Lipshutz cuprates . . . 50
1.15 Modelled pathways for the deprotonation of N,N -dimethylbenzamide
by MeCu(NMe2)(CN)Li2 · 2OMe2 . . . . . . . . . . . . . . . . . . . 51
1.16 Modelled pathway for the coordination of the monomeric Gilman-
type complex MeCu(NMe2)Li ·OMe2 to N,N -dimethylbenzamide . 52
1.17 Modelled pathways for the deprotonation of N,N -dimethylbenzamide
by MeCu(NMe2)Li ·OMe2 . . . . . . . . . . . . . . . . . . . . . . . 53
4.1 General reaction scheme for syntheses of phosphidocuprates . . . . 94
xxiii
LIST OF SCHEMES
5.1 Conversion of 3-benzoyl-2-chloropyridine (25b2) to 5H -indeno[1,2-
b]pyridin-5-one . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
6.1 General reaction pathways used to investigate solvent effects on cu-
prates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
7.1 Synthetic strategy towards forming DMP-based bis(amido)cuprates 136
7.2 Synthetic strategy towards forming a pure adduct-type cuprate . . . 140
7.3 Synthetic strategy towards forming DMP-based Gilman-type bis-
(amido)cuprates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
7.4 Dissociation of Lipshutz- and adduct-type cuprates into Gilman- and
Lipshutz-type monomers . . . . . . . . . . . . . . . . . . . . . . . . 155
7.5 DoCu of N,N -diisopropylbenzamide by DMP-based cuprate bases . 157
7.6 The interconversion of a Lipshutz-type dimer and lithium amide-
coordinated Lipshutz-type monomer . . . . . . . . . . . . . . . . . . 158
7.7 The interconversion of Lipshutz- and Gilman-type dimers with se-
quential loss of LiCl . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
7.8 The interconversion of Lipshutz- and Gilman-type dimers in the
presence of varying amounts of Me2O . . . . . . . . . . . . . . . . . 160
7.9 The interconversion of Lipshutz- and Gilman-type dimers in the
presence of varying amounts of THF . . . . . . . . . . . . . . . . . 161
7.10 Modelled pathway for the deprotonation of N,N -dimethylbenzamide
by [(Me2N)2CuLi ·OMe2]2LiCl . . . . . . . . . . . . . . . . . . . . . 162
7.11 The interconversion of the Lipshutz-Gilman adduct and Gilman-
(N,N -dimethylbenzamide) complexes . . . . . . . . . . . . . . . . . 163
9.1 Possible routes to Lipshutz-type bis(phosphido)cuprates . . . . . . . 174
xxiv
Chapter 1
Introduction
1.1 Directed ortho Metalation
The synthetic chemist has a large array of organometallic reagents at their disposal,
many of which are in wide use on an industrial scale.1–3 One particular reaction
which organometallic reagents can be put to use in is Directed ortho Metalation
(DoM), which traces its routes back to the work of Gilman4 and Wittig5 nearly 75
years ago. The mechanism of DoM involves the regioselective C−H bond activation
of an aromatic compound containing at least one directing group by a metallic
base, resulting in the metalation of the ortho carbon atom. This is followed by
an electrophilic quenching step, often creating a C−C bond, to give the desired
product (see Figure 1.1).
DG
H
DG
MRn
metalating agent electrophile
DG
E
Scheme 1.1: General scheme for a directed ortho metalation reaction, DG = di-
recting group
As many of today’s synthetic targets contain aromatic or heteroaromatic rings,6–8
research into new DoM agents and the mechanism by which DoM occurs is cer-
tainly warranted. It should be noted that DoM is by no means the only tool
available to achieve aromatic substitution; well-known alternatives include elec-
trophilic Friedel-Crafts acylations and alkylations, which are still in use today.
1
1. INTRODUCTION
These present several drawbacks though, most notably their poor regioselectivity
leading to resource-heavy separation and purification steps.9,10 In large-scale in-
dustrial syntheses, the costs associated with these purification processes preclude
the wide use of such methods. DoM, however, provides a much more regioselec-
tive pathway, hence delivering the purity required by the pharmaceutical industry
is more straightforward.
Much research has focused on synthesising novel DoM agents in an effort to improve
yields under mild conditions and to understand the mechanisms involved. From the
initial discoveries where M = Li, Directed ortho Lithiation (DoLi), more recently
other metals have proved synthetically useful and research has moved from Di-
rected ortho Magnesiation (DoMg),11 through Directed ortho Zincation (DoZn),12
Directed ortho Manganation (DoMn),13 Directed ortho Alumination (DoAl)14 and
finally to Directed ortho Cupration (DoCu).15 This introduction will review both
the historical and recent advances in the field.
1.2 Directed ortho Lithiation
The first DoLi reaction to be reported was between nBuLi and anisole, work which,
coincidentally, was carried out independently and almost simultaneously in the labs
of Gilman4 and Wittig.5 Although yields were low, stereoselectivity for deprotona-
tion at the ortho position was noted by the researchers. This discovery prompted
further reactivity studies, initially by Gilman himself and later by Hauser,16 as the
scope of the reaction with a wide variety of directing groups began to be inves-
tigated. This research discovered that the electronic and steric effects conferred
by the directing group was a very important factor in the success of the DoLi re-
action, and enabled classification of substituents based on their directing ability
(see Figure 1.1).
2
1
.
IN
T
R
O
D
U
C
T
IO
N
O NR2
O NR2
tertiary
amides
secondary
amides
S
tBu
sulfones
O
O
S
NR2
tertiary
sulfonamides
O
O
S
sulfoxides
SH
NR2
secondary
sulfonamides
O
O
tBuO
O N
oxazolines
NR
imines
N
N-carbamates
OtBu
O
O
MOM ethers
O OR
ethers
NR2
anilines
Hal
halogens
CF3
trifluoromethyl
remote
amines
NR2
n
benzylic
alkoxide
O
decreasing ability to direct ortho lithiation
precise order varies with conditions
O
O-carbamates
NR2
O
N+O class O class
N class X classS+O class
Figure 1.1: Classes of Directing Groups for ortho Lithiation17
3
1. INTRODUCTION
The strongest directing groups fall into the ‘N+O’ class, where the aromatic sub-
stituent both confers a strong inducting effect, causing a significant rise in the
acidity of the ortho protons (and to a lesser extent, meta protons), and contains
a strongly Lewis basic atom (oxygen) which adeptly stabilises the incoming Li
atom. Such directing groups do have their drawbacks however, as the α-C atom
is strongly nucleophilic, and the attached carbonyl oxygen is also a strong Lewis
base. This leads to two possible competing reaction pathways, either nucleophilic
substitution, or deprotonation at the γ-C, with the carbonyl O atom affording sta-
bilisation to the lithiated intermediate. As such, the nature of the R groups on the
directing group is important in ortho substitution reactions. Taking benzamides
as an example, when R = Me no isolable product is synthesised, but when R =
iPr yields of >90 % have been reported.18
The next strongest directing groups belong to the ‘S+O’ class, which again contain
Lewis basic atoms able to coordinate to the Li atom, but a less nucleophilic α-atom.
While higher yields would thus be expected, due to complex work-up procedures
directing groups belonging to the ‘S+O’ class become less attractive options. In
addition, such sulfur-containing substituents are less synthetically relevant than
analogous carbon-containing substituents, and as such, expensive functional group
conversion procedures preclude their wide use in organic synthesis.
While compounds belonging to the ‘N’ and ‘O’ classes have less effective directing
groups, their ortho substituted products are synthetically useful so they are worth
consideration. For the ‘N’ class, the Lewis-base donor ability outweighs any induc-
tive effects, and hence compounds where the nitrogen is in the β- or γ-position have
stronger ortho-directing abilities. These factors reverse when the ‘O’ class is consid-
ered, and so compounds such as anisole where O is in the α-position are best suited
for DoLi. Finally, when the heteroatom is a halogen (a subset of the ‘X’ class) the
inductive effect now far outweighs any donor ability so the relative acidity of the
ortho proton will determine how successful the reaction is. A drawback here is the
possibility of lithium-halogen exchange, particularly relevant when X = Br or I.19
1.2.1 Directed ortho Lithiation mechanisms
Having carried out extensive reactivity studies, research focus moved on to inves-
tigating the mechanism of DoLi. Knowing that substituents with strong Lewis
base donor atoms were effective directing groups, a theory called the Complex
Induced Proximity Effect (CIPE) developed which involves pre-association of the
4
1. INTRODUCTION
lithium atom to the Lewis base prior to deprotonation, hence directing deproto-
nation to the ortho proton.
X
H
reactants
R Li+
X
H
Li
R
pre-reaction complex
X
H
Li
R
–RH
X
Li
lithiated producttransition state
Figure 1.2: Schematic representation of a DoLi on a singly functionalised aromatic
ring via a CIPE mechanism
Initially a pre-reaction complex is formed as the lithium from the organolithium
base is bound by an electronegative atom (X) in the directing group. This re-
sults in the R− group being in close proximity to one of the ortho protons which
forces regioselective deprotonation, the reaction proceeding via a six-membered
ring transition state. The conjugate acid is then released, leaving the lithium in
the ortho position, which depending on the nature of the directing group may still
be complexed (if sterics allow), conferring intramolecular stabilisation. The extent
of the intramolecular stabilisation has been probed, and it was found that the pro-
ton quench of ortho-anisyllithium was 3.6 kcal mol−1 less exothermic than that of
para-anisyllithium.20 There has been considerable research into providing evidence
for this mechanism, and while data from qualitative NMR spectroscopic studies,21
steric effect studies22 and kinetic isotope effect studies23 appear to support the
theory of a pre-reaction complex, it has yet to be proved beyond reasonable doubt.
The remaining uncertainty allowed other theories to be put forward, and in
1994 Schleyer, branding the data supporting CIPE as “circumstantial evidence”,
proposed an alternative mechanism, namely Kinetically Enhanced Metalation.24
Schleyer reasoned that if complexation was the key step, one would expect a slow
rate of reaction as the complexation energy would have to be overcome en route to
the transition state. Experimentally, however, rates were accelerated which demon-
strated that the transition state was lower in energy than the initial complex. The
Kinetically Enhanced Metalation theory fits this data as it relies only on the induc-
tive (in this case, electron withdrawing) properties of the directing group, the ortho
proton being closest to the DG and hence experiencing the effect to the greatest
extent relative to the other protons on the ring.
The debate continued as the arena moved from the experimental laboratory into the
computer rooms of the theoretical chemists. Using Modified Neglect of Differential
Overlap (MNDO) calculations, Saa et al. showed that agostic interactions in both
5
1. INTRODUCTION
the pre-reaction complex and intermediate are important, adding weight to the
argument for CIPE.25 It should be noted, however, that at around the same time,
work by Janes cast doubt upon the reliability of MNDO calculations, demonstrating
that computed structures of 1,2,4-triazine rings displayed significant errors when
compared with data from X-ray crystallography.26 This is due to the fact that
MNDO calculations are designed to give sensible answers when given vague ideas
of the expected structure as input. Using ab initio calculations, which are more
reliable as they are derived from first principles, Collum et al. determined that there
was little interaction between the lithium and the directing group in the transition
state.27 Moving back to the laboratory, new research discovered that unreactive, or
even hindering pre-reaction complexes may also form in DoLi reactions,28 overall
adding to the argument that Kinetically Enhanced Metalation is the most plausible
reaction mechanism.
1.2.2 Limitations of DoLi
While lithiating agents are very powerful synthetic tools, due to the highly nu-
cleophilic and basic nature of the R– group, they have a low functional group
tolerance and so the reaction has significant limitations. As briefly discussed in
Section 1.2.1, the combination of directing group and lithiating agent has to be
considered carefully. If the directing group contains a nucleophilic centre such as a
carbonyl bond, nucleophilic attack can occur in preference to deprotonation. The
result may be either the formation of a lithiated species, such as an alkoxylithium,
or, if there is a suitable leaving group, a substitution reaction. This means either
a bulky directing group, such as a diisopropylamide, or a bulky lithiating agent,
such as LDA, is required (Figure 1.3), both of which provide a steric block to
attack of the carbonyl bond.
Unfortunately, the use of a sterically bulky directing group can cause complications
later on in product synthesis, as they can be difficult to remove or transform. While
there are limited examples of bulky directing groups, for example N -tBu-N -Me-
amides30 and acyl oxazolidines31 which are readily cleaved by acids, most require
harsh work-up conditions which may also degrade other parts of the molecule. In
addition, if the directing group contains acidic hydrogens, for example β-hydrogens
adjacent to a carbonyl bond, these may be deprotonated preferentially to the ortho
proton. The creation of lithium enolates is used synthetically in the formation of
C−C bonds at an sp3 centre,32 and such reactions must be considered when at-
6
1. INTRODUCTION
nBu O
N O N O
Li
nBuLi LDA
N O N O
nBuLi
Li
Figure 1.3: Control of DoLi reactions. The treatment of N,N -diethylbenzamide
with nBuLi leads to nucleophilic addition at the carbonyl, whereas using LDA in-
stead causes ortho deprotonation. Conversely, treating N,N -diisopropylbenzamide
with nBuLi also results in ortho lithiation29
tempting DoLi reactions. Removal of a β-hydrogen has the effect of deactivating
the directing group and so on the addition of a second equivalent of a lithiat-
ing agent, further deprotonation or nucleophilic attack at the directing group is
disfavoured, and ortho lithiation can be achieved16 (Scheme 1.2). However, the
increased volume of lithiating agent and extra quench step required must be taken
into account if such a reaction is to be scaled-up.
N O
H
R
N O
Li
R
N O
R
N O
Li
R
Li
Li
RLi RLi
Scheme 1.2: Pathway to DoLi via dilithiation of secondary amines16
Another drawback to DoLi is that, depending on the functional group, anionic
ortho-Fries rearrangements may occur. A Fries rearrangement involves the trans-
formation of a phenyl ester to a ortho- or para-hydroxy aryl ketone, and in the case
of an ortho-lithiated phenyl ester, the ketone can (effectively) trade places with the
lithium to form an ortho-substituted lithium phenolate. This reaction pathway is
known as an “anionic ortho-Fries rearrangement”, and is a particular consideration
7
1. INTRODUCTION
when employing carbamate functional groups, which will undergo rearrangements
depending on the nature of the N -substituent.17 While N,N -diisopropylcarbamates
are stable after ortho-lithiation, N,N -dimethylcarbamates will undergo an anionic
ortho-Fries rearrangement. The reactivity of N,N -diethylcarbamates is tempera-
ture dependent; at –78 ◦C the ortho-lithiated carbamate is stable, but on warming
to room temperature it undergoes a rearrangement and a 2-hydroxybenzamide will
be isolated after an acidic work-up (Scheme 1.3).33
O O
NEt2
O O
NEt2
Li
OLiNEt2
O
OHNEt2
OsBuLi, –78 oC warm to rt H+
anionic ortho-Fries
rearrangement
Scheme 1.3: Lithiation of a N,N -diethylcarbamate at –78 ◦C, followed by anionic
ortho-Fries rearrangement after warming to ambient temperature and subsequent
trapping via an acidic work-up33
The substantial limitations of DoLi warranted research into alternative metalation
agents. Pioneering work by Wittig in the 1950s demonstrated the deprotonative
ability of bimetallic bases of the form Ph3LiM (M = Be, Zn, Cd, Mg),
34 termed
“’ate” complexes, but it has only been more recently that these methods have
been applied to DoM. The following sections focus on the development of DoM
agents which have shown the most promise: heterobimetallic magnesiates, zincates,
aluminates, manganates and latterly, cuprates.
1.3 Directed ortho Magnesiation
Preparative routes to lithium magnesiates have been known for many years, from
Wittig’s reactions between diorganomagnesium and organolithiums,34 through 3:1
and 4:1 mixtures of organolithiums and magnesium halides,35,36 to reduction of
dialkylmagnesium with lithium. Their synthetic applications, however, have not
been subject to extensive investigations until recently. Magnesiate compounds
have now found use in halogen-metal exchange reactions37 and in cyclisation re-
actions through modification of deprotonated substrates,38 as well as in ortho de-
protonation reactions.11 In 1990, Richey and Farkas published their research on
the reactivity of pyridine with alkyllithiums and alkylmagnesiums.39 The reaction
with standalone ethyllithium or diethylmagnesium led to the production of little
8
1. INTRODUCTION
or no 2-ethylpyridine. However, they were able to improve the yield by using a
mixture of both Et2Mg and EtLi, and in this case 4-ethylpyridine also formed in
low yields. As the behaviour of the mixture was different to that of either metal
individually, they proposed that a magnesiate species was responsible for the for-
mation of both alkylpyridines.39 Building on this work, Mongin and coworkers
demonstrated that when the pyridine incorporated a directing group, Bu3MgLi
could be used as an effective ortho deprotonation agent (Scheme 1.4). They noted
a favourable thermal stability of the arylmagnesiate, as reactions could be carried
out at –10 ◦C without a 3,4-pyridine forming, whereas 3-fluoro-4-lithiopyridine
must be kept below –60 ◦C.11
N
F
N
F1/3 eq. Bu3MgLi electrophile
N
E
F
THF, –10 °C, 2 h
MgR2Li
R = Bu or 3-fluoro-4-pyridyl
Scheme 1.4: Deprotonation of 3-fluoro-4-pyridine with Bu3MgLi and subsequent
electrophilic trapping11
To extend the study to other pyridyl species and in an attempt to improve yields,
research moved to magnesiates containing a 2,2,6,6-tetramethylpiperidido (TMP)
group, which had been successfully employed previously in cross-coupling reac-
tions.40 Due to the amido group’s greater migratory aptitude over that of the alkyl
group’s, deprotonation of a variety of flouro- and chloropyridines was successful,
and quenching with iodine afforded high yields of the 3-substituted product.
N
1) 1/3 eq. Bu3Mg(TMP)Li2
THF, –10 °C, 2 h
2) I2
X1X3 N X1X3
I
X2 X2
a 73 %
b 84 %
c 84 %
a X1 = X3 = H, X2 = Cl
b X1 = X3 = F, X2 = H
c X1 = F, X2 = X3 = H
Scheme 1.5: Deprotonation of halopyridines with Bu3Mg(TMP)Li
11
The first crystal structure of a magnesiate was published in 1981, the reaction
9
1. INTRODUCTION
between Me2Mg, (MeLi)4 · 2TMEDA and TMEDA in a 2:1:2 ratio in diethyl ether
yielding Me4MgLi2 · 2TMEDA. The structure displays both four-coordinate Mg
and Li, with two (MgCLiC) rings approximately perpendicular to one another due
to the tetrahedral arrangement of the methyl groups around the Mg centre. A
TMEDA molecule caps the two Li atoms at either end, acting as a Lewis base
donor to the highly electropositive metal atoms, filling their coordination spheres
and preventing polymerisation41 (Figure 1.4).
Figure 1.4: Molecular structure of Me4MgLi2 · 2TMEDA
41
Weiss later extended this study to include an ion-separated magnesiate with
the characterisation of a monomeric benzylmagnesiate [Bn4MgLi · TMEDA]
–
[Li · 2TMEDA]+ (Figure 1.5). This species was formed from the reaction between
ethyllithium and diethylmagnesium in an excess of toluene, in which the ethyl
groups were exchanged for the more strongly σ-donating benzyl group. Again the
(MgCLiC) ring motif is noted, with two benzyl groups coordinating to both the
Mg and Li atoms via the meso-carbon and the other two terminally capping the
Mg (Figure 1.5). The species is anionic, with [Li · 2TMEDA]+ acting as the coun-
terion, but parallels can be drawn with the synthetic metalating agents discussed
above where the active species is likely to have a similar structure except with a
three-coordinate Mg centre.
10
1. INTRODUCTION
Figure 1.5: Molecular Structure of the [Bn4MgLi · TMEDA]
– anion of [Bn4MgLi ·
TMEDA]– [Li · 2TMEDA]+42
1.4 Directed ortho Zincation
The use of zinc in organometallic compounds stretches back to the 1850’s, when
Frankland synthesised dimethyl- and diethylzinc from a simple reaction between
finely granulated zinc and ethyliodide.43,44 Like their magnesiate counterparts,
lithium zincates were in synthetic use before their potential application in DoM
was realised. Owing to their soft nature, they are ideal for use in halogen-metal
exchange reactions, for example in alkylation of alkenes.45 Harada et al. developed
an efficient one-pot pathway to propargylic derivatives via allenic zinc compounds
formed from the reaction between triorganozincates and propargylic substrates46
and demonstrated that a wide range of nucleophiles and electrophiles could be in-
troduced at the 1,3-positions of a propargylic moiety.47 Kondo and coworkers were
the first to attempt DoMwith zincates, by adding arylcarboxylic esters to a mixture
of di-tert-butylzinc and LiTMP. tert-Butyl groups were chosen as they displayed
the lowest preference for migration when compared with other alkyl groups,48 and
LiTMP as it had been previously used for DoLi of arylcarboxylic esters.49 After
probing the zincate mixture by NMR spectroscopy, they proposed an empirical
formula of tBu2Zn(TMP)Li for the active species, and it proved to be an efficient
DoM agent; reactions were carried out at room temperature and high yields were
recorded after trapping the arylzincate with an electrophile.12 The chemistry was
also found to be applicable to chemoselective deprotonation of heterocycles such
as pyridine and furan, which had previously proved difficult due to the asymmet-
rical electron distribution in the ring caused by the electronegative heteroatoms.
As a further advantage, no intermolecular condensation products formed in the
11
1. INTRODUCTION
reactions, a clear benefit over corresponding aryllithium intermediates.
DG DG
ZntBu2Li
DG
E1) E-X
2) hydrolysis
tBu2Zn(TMP)Li
THF, RT, 3 h
Scheme 1.6: Deprotonation of functionalised arenes using a TMP-zincate and sub-
sequent electrophilic trapping reactions12
Following on from this work, Kondo was able to apply DoZn to the formation
of reactive asymmetric benzyne species from 1,3-substituted benzenes. Previ-
ous methods required 1,2,3- and 1,2,4-substituted benzene precursors, which were
difficult to synthesise, and had a low tolerance to functional groups. Initially,
Kondo’s group were unable to form benzyne species, as their choice of zincate
base, tBu2Zn(TMP)Li, led exclusively to ortho metalation, the bulky R-group pre-
venting elimination of R2ZnBr. However, by replacing the tert-butyl groups with
sterically undemanding methyl groups, benzyne formation proceeded smoothly in
high yields. This reactive intermediate could then undergo an electrophilic quench
or a Diels-Alder cyclisation to form a wide variety of products.50
R
Br
1) tBu2Zn(TMP)Li
2) I2
1) Me2Zn(TMP)Li
–HTMP, –Me2ZnBr
O
Ph
Ph
2)
R
RR
Br
I
O Ph
Ph
Scheme 1.7: Reaction pathways for the addition of an alkylzincate reagent to an
aromatic50
Kondo’s proposed composition of the zincate base was later confirmed by X-ray
crystallography.51 The base was prepared in a hydrocarbon solvent and one equiv-
alent of THF and displays a heterocyclic (C2ZnNLi) core containing the α- and β-
carbons from one of the tBu substituents (Figure 1.6(a)), unlike the four-membered
12
1. INTRODUCTION
rings observed in the structures of magnesiates discussed in Section 1.3. Again, the
highly electropositive lithium is stabilised both by agostic donation from an alkyl
carbon (Li−C bond length 2.410(6) A˚) and also from a Lewis base, in this case
a molecule of the solvent, THF. To try and gain insights into the DoZn reaction,
the synthesis was repeated with N,N -diisopropylbenzamide in place of THF, and
the product analysed by X-ray crystallography. As the structure contains the same
heterocyclic core with an agostic interaction between a β-carbon and the Li, it is
structurally similar to the THF-solvated zincate (Figure 1.6(b)).
(a) (b)
Figure 1.6: Molecular structures of (a) tBu2Zn(TMP)Li · THF and (b)
tBu2Zn(TMP)Li · (O−)(Ph)CN(
iPr)2
51
The latter structure was interpreted as a precomplex to DoM, and it was used to
propose two possible transition states of the ortho deprotonation reaction. The
first involves deprotonation by one of the tBu groups, which is then eliminated via
a nine-membered ring transition state. In the second pathway, TMP carries out
the deprotonation and is subsequently eliminated, this time via a seven-membered
ring (Figure 1.7). The structure of tBu2Zn(TMP)Li · (O−)(Ph)CN(
iPr)2 shows
that the µ-tBu group lies closer to the ortho-proton than the amido N centre. The
µ-tBu group is oriented towards the aromatic ring due to the agostic interaction
with the Li centre and is only bound strongly to the Zn centre therefore making
it more sterically accessible. For this reason it was thought that alkyl-mediated
deprotonation was the most likely mechanism.
Mulvey et al. were also able to fully characterise a dimethyl(TMP)zincate, using
TMEDA as an aid to crystallisation. The zincate forms a monomer in the solid-
state, analogous to the tert-butyl zincate base, although as there are no β-carbons
present, a four-membered (LiNZnC) ring forms with a weak agostic bond between
the bridging methyl carbon and lithium (2.603(5) A˚). The agostic interaction is
evidenced by the distortion around the Li centre to a slightly pyramidal geometry
13
1. INTRODUCTION
(iPr)2N O
H
Li
N
Zn
tBu
tBu
CH3
CH3
C
H3
(a)
(iPr)2N O
H
Li
N
H3
C
Zn
CH3
CH3
tBu
tBu
(b)
Figure 1.7: Possible transition states for a DoZn reaction, via either (a) elimination
of isobutane or (b) elimination of HTMP51 (N = TMP)
(sum of N−Li−N angles 353.06◦) and the disparity between the two N−Zn−C bond
angles (111.93(11)◦ where the carbon is involved in agostic bonding, cf. 126.55(10)◦
otherwise). They also extended the study to other amines, finding a surprising
result with hexamethyldisilylamide. LiHMDS is known to form dimers in the solid-
state in the presence of simple donors, and this motif was retained on addition to
dimethylzinc in the presence of TMEDA. The structure of the resulting compound
is polymeric and can best be described as an adduct, with Lewis basic −CH3(Zn ·
TMEDA)CH3− bridges donating to the Li centre through methyl groups linking
Lewis acidic LiHMDS dimers (Figure 1.8). NMR spectroscopic studies showed
that in a non-coordinating solvent (C6D6) the LiHMDS and Me2Zn ·TMEDA units
remain linked, suggesting that the choice of amine will have a bearing on the DoM
ability of the zincate base.52
Figure 1.8: Section of the polymeric [(LiHMDS)2 · (Me2Zn · TMEDA)]∞
52
14
1. INTRODUCTION
To probe the mechanism of zincation further, the procedure used to form
tBu2Zn(TMP)Li · (O−)(Ph)CN(
iPr)2 was repeated, but this time the mixture was
allowed to stir for 30 minutes after the addition of the benzamide. The product
that crystallised was an unusual diarylzincate species, which formed because the
reactant zincate base deprotonated two equivalents of the amide but remained co-
ordinated to both (in the expected ortho position). HTMP and isobutane had both
been eliminated during the reaction (Figure 1.9). The results suggested that both
alkyl- and amido-mediated deprotonations were possible, although it raised many
questions as to the pathway by which the reaction was occurring.53
Figure 1.9: Molecular structure of tBuZn{µ-2-[1-C(O)N(iPr)2]C6H4}2Li·TMEDA
53
While structural studies were ongoing, Uchiyama et al. were conducting detailed
computational calculations on the zincation mechanism. Coincidentally, the group
also solved the crystal structure of tBu2Zn(TMP)Li · THF, which they were then
able to use as a starting point for their theoretical models. Taking Me2Zn(Me2N)Li·
OMe as the base, anisole as the substrate and employing the B3LYP/6–31+G*
basis set, the DFT calculations indicated that although alkyl deprotonation is
favoured thermodynamically, deprotonation by the TMP ligand is the most kinet-
ically favourable route.54 Further work demonstrated that while it is energetically
favourable for alkyllithiums to perform 1,2-addition at the functional groups of
benzonitrile and methyl benzoate in preference to DoM, the reverse is true for
lithium zincates, which was in agreement with the experimental data and added
further evidence for the DoZn mechanism.55
15
1. INTRODUCTION
The known dibasic nature of zincate bases was probed further theoretically by
Wheatley et al. Initially, a pathway involving transfer of the bridging-alkyl ligand
was calculated (Scheme 1.8). The overall reaction was found to be exothermic
(–23.3 kcal mol−1) although it has to proceed via a very high energy transition
state, TS1 (∆G‡ = 37.8 kcal mol−1).
Me
Zn
Li
Me2N
Me
S
Me
Zn
Li
Me2N
Me
S O
NMe2
CP1
RT1
PD1
Me2N
Zn
Li
Me
S
O
NMe2
CH4
NMe2
OH
Me
Zn
Li
Me2N
Me
S
NMe2
H
O
TS1
2.01
2.08
2.83
2.06
1.9685.3
105.7
65.4
85.0 1.95–1.7 +37.8
+
2.00
2.12
1.91
2.11
2.17
72.3
102.6
77.3
107.7
–59.4
PD2
2.01
2.09
1.96
2.08
2.50
69.7
107.0
81.3
96.82.03 1.93
1.91
1.40
1.09
1.98
1.97
2.00
2.54
2.38
68.7
96.4
88.9
101.32.01 1.89
1.46
1.48
1.91
+
51.9
Scheme 1.8: Calculated structures of the reactants, transition states, products, and
Gibbs free energy changes (∆G) for kinetically unfavourable alkyl-mediated ortho
deprotonation. Bond lengths, angles and energy changes at the B3LYP/6–31+G*
level are shown in A˚, degrees and kcal mol−1, respectively (S = OMe2)56
The high kinetic penalty imposed by the mechanism led to a second, stepwise
pathway being postulated. This pathway involves amido-ligand exchange (Scheme
1.9) followed by re-coordination of in situ generated amine to the lithium and
subsequent alkane loss (Scheme 1.10). The activation barrier to the deprotonation
step is significantly lower (∆G‡ = +24.0 kcal mol−1) compared to the alkyl-transfer
pathway, although the process is nominally endothermic (∆G = +2.2 kcal mol−1).
The second step, elimination of the alkane mediated by re-coordination of Me2NH,
is significantly exothermic (∆G = –25.5 kcal mol−1) with a relatively uninhibitory
activation barrier (∆G‡ = +21.6 kcal mol−1). Overall, both pathways result in
the same, most thermodynamically stable products, but the latter does not require
direct alkyl elimination (Figure 1.10) and provides a rational explanation for the
observed experimental data.
16
1. INTRODUCTION
NMe2
Zn
Li
Me
Me
S O
Me2N
CP2
(S = Me2O)
PD3
Me
Zn
Li
Me
S
O
NMe2
HNMe2
NMe2
OH
TS2
NMe2
Zn
Li
Me
Me
S O
H
NMe2
2.01
2.15
2.39
2.10
2.1971.7
114.6
67.0
102.4 1.94–2.3 +24.0
+
–19.5
PD4
2.01
2.06
2.44
2.12
1.98
79.9
104.7
70.9
96.52.00 1.93
1.40
1.09
1.91
2.01
2.07
2.25
2.09109.6
81.0
1.97 1.91
2.97 1.31
1.43
+ 102.5
RT1
Scheme 1.9: Calculated structures and energy changes for step 1 (amine elimina-
tion) in stepwise amido-mediated ortho deprotonation. Bond lengths, angles and
energy changes (∆G) at the B3LYP/6–31+G* level are shown in A˚, degrees and
kcal mol−1 respectively56
(S = Me2O)
H
Me Zn
Li
Me
S
O NMe2
Me2N
CP3
Me Zn
Li
Me
S
O NMe2
Me2N
H
TS3
PD1
Me2N
Zn
Li
Me
S
O
NMe2
CH4
+4.2 +21.6 –51.3
PD4
+
2.04
2.18
2.33
2.03
4.04
111.6
77.0
1.98
1.922.09
2.00
2.06
2.80
2.32 115.2
72.6
106.8
1.94
1.91
1.38
1.41
1.97
59.7
99.3
2.01
2.08
2.83
2.06
1.9685.3
105.7
65.4
85.0 1.95
+
PD2
1.91PD3
Scheme 1.10: Calculated structures and energy changes for step 2 (amine re-
coordination/alkane loss) in stepwise amido-mediated ortho deprotonation. Bond
lengths, angles and energy changes (∆G) at the B3LYP/6–31+G* level are shown
in A˚, degrees and kcal mol−1 respectively56
17
1. INTRODUCTION
CP1
–1.7
RT1
TS1
+36.1
TS2
+21.7
CP2
–2.3
PD2
PD4
+2.2
∆G (kcal mol-1)
0.0
10.0
20.0
30.0
–10.0
Me2N
H
+
O
Deprotonation with
amido ligand
Deprotonation with
alkyl ligand
+37.8
+24.0
–23.3
TS3
+28.0
CP3
Deprotonation of Me2N-H
with alkyl ligand
+21.6
PD1
+
–20.0
PD3
+
+6.4
40.0
PD2
–23.3
PD1
+
Figure 1.10: Summary of deprotonative zincation pathways of the reaction between
N,N -dimethylbenzamide and Me2Zn(Me2N)Li·OMe2 at the B3LYP/6–31+G* level
of theory. Energy changes are shown in kcal mol−1 and are relative to (RT1 +
N,N -dimethylbenzamide) (see Scheme 1.8)56
18
1. INTRODUCTION
Although only bis(amidoaryl)zincates had been observed experimentally, the theo-
retical data suggested that zincates should be able to exhibit tribasicity, if amine/
alkyl exchange (Scheme 1.10) occurs each time after the first two deprotonations.
Using a sterically undemanding aromatic reagent, N,N -diisopropylbenzamide, and
THF as a donor to allow more facile coordination of amine to the lithium, the
crystallisation of a tris(amidoaryl)zincate, Zn(µ-C6H4C(O)N-
iPr2-2)3Li · THF was
achieved (Figure 1.11). To accommodate the steric bulk of three aryl groups around
the zinc, the amine-arene twist angle (mean torsional angle 52.8◦) in the struc-
ture is significantly distorted from the ideal near-perpendicular geometry. How-
ever, when a sterically demanding aromatic reagent, N,N -diisopropylnaphthamide,
was employed, the zincate base only exhibited dibasicity in a similar manner to
the previously discussed behaviour observed by Mulvey et al. The reaction be-
tween the napthamide and EtZn(µ-Et)(µ-TMP)Li, which was shown to be poly-
meric in the solid-state, formed EtZn(µ-C10H6C(O)N
iPr2-2)2Li · 2THF. This spe-
cies has a mean amide-arene twist angle of 86.2◦, much higher compared to the
tris(amidoaryl)zincate species, as the steric demands of the napthamide ring makes
further reduction in the twist angle, and hence a third napthamide deprotonation,
unfavourable.56
Figure 1.11: Molecular structure of Zn(µ-C6H4C(O)N-
iPr2-2)3Li · THF
56
Subsequent NMR spectroscopic studies by Hevia et al. added to the evidence
for the two-step deprotonation mechanism. Two zincated anisole intermediates,
R2Zn(C6H4-OMe)Li · 2THF (R = Me,
tBu) were isolated and characterised be-
fore and after addition of HTMP. The size of the R group and the solvent system
proved to be key factors of the outcome of the reaction. When R = Me, the reac-
tants reformed almost exclusively, whereas when R = tBu, in benzene amido/alkyl
19
1. INTRODUCTION
substitution occurred almost exclusively, although in THF a mixture of reactants,
an amidoalkylzincated aromatic and butane, was observed.57
Recently, lithium zincates have been employed in the synergic transformation of
N,N ′-diisopropylethylenediamine to tBuZn(iPrNCHCHNiPr)Li · TMEDA by reac-
tion with one equivalent each of nBuLi and tBu2Zn in the presence of TMEDA.
While tBu2Zn on its own was merely coordinated by the amine and
nBuLi affected
an amino α-deprotonation as expected, the combination of both reagents led to the
isolation of a metallocycloalkene involving Zn with the Li η4-coordinated asym-
metrically to the −N−C−C−N− unit. DFT calculations showed that plausible
mechanisms exist going via either a mono- or dihydride intermediate species.58,59
In addition, a sodium zincate, (Me3SiCH2)Zn(µ-CH2SiMe3)(µ-TMP)Na · TMEDA
has been shown to have the remarkable capability of functionalising cyclic ethers.
Trimethylsilyl-based ligands (Me3SiCH
–
2) were chosen as a more gentle alternative
to aggressive purely alkyl ligands, as the α-silyl substituent stabilises the carbanion.
Reacting the base in THF or THP led to regioselective zincation of the α-carbon
as demonstrated by the X-ray crystallographic data (Figure 1.12). The zincation
rendered the metalated α-carbons in each structure chiral with the cyclic ether
fragments stabilised by coordination to both Zn and Na (via the O centre), in an
analogous manner to other characterised DoZn intermediates (vide supra). Elec-
trophilic quenching of the zincated species demonstrated that the method provides
a viable route to cyclic ether-functionalised ketones.60
Figure 1.12: Molecular structure of (Me3SiCH2)Zn(µ-TMP)(µ-C4H7O)Na ·
TMEDA60
Structural studies of bimetallic zincates have latterly focused on the incorporation
of alternative amido ligands to the widely used TMP in order to reduce synthetic
20
1. INTRODUCTION
costs. O’Hara et al. successfully characterised a series of lithium and sodium
zincates containing DMP and DIBA.61 The choice of DMP as a ligand was sup-
ported by the fact that LiDMP had previously been utilised as a bulky base in
various applications, for example the synthesis of annulated nicotine analogues62
and exocyclic allylic alcohols.63 The reaction of LiDMP and NaDMP with alkyl-
zinc reagents resulted in the formation of monomeric zincate species analogous to
the TMP-zincates (vide supra) in the solid-state. tBu2Zn(µ-DMP)Li · TMEDA
(Figure 1.13) contains an open, curved −Li−N−Zn−C− motif. When the larger
alkali metal Na was employed, however, the structure of the resulting species
tBu2Zn(µ-DMP)Na ·TMEDA includes a weak agostic interaction to the metal from
a β-carbon on the bridging tBu group. However, the geometry around the Zn cen-
tre was found to be almost identical in both species (range of angles and total angle
around the Zn centre: 116.19(6)–122.61(6)◦ and 359.86◦ (Li-analogue); 117.8(2)–
123.7(2)◦ and 359.9◦ (Na-analogue)). Treatment of tBu2Zn with NaDIBA led to the
formation of a diamido zincate (DIBA)2Zn(
tBu)Na ·TMEDA containing two bridg-
ing amido ligands. By comparing the Na−NTMEDA bond lengths in the latter two
species with those in tBuZn(µ-iPr2N)2Na · TMEDA
64 it was suggested that DMP
imposes a larger steric demand than diisopropylamide, which in turn is bulkier
than DIBA. However, as the differences are small and the solid-state structures are
similar, the new species are potentially useful as DoZn agents.
Figure 1.13: Molecular structure of tBu2Zn(µ-DMP)Li · TMEDA
61
21
1. INTRODUCTION
1.5 Directed ortho Alumination
Much like organozinc compounds, the origins of organoaluminiums can be traced
back to the mid-nineteenth century, when the first synthesis of (C2H5)3Al2I3
65 was
achieved. It took almost 100 years before they began to be put to use, however,
starting with Ziegler’s Nobel prize-winning work synthesising trialkylaluminiums
and utilising them to catalyse olefin polymerisation reactions.66 Unlike organozin-
cates, however, organoaluminates have been used in directed metalation for many
years, stretching back to 1974 when Yamamoto et al. achieved the highly re-
gioselective deprotonation of epoxides using DATMP.67 The base was prepared
simply by mixing Et2AlCl and LiTMP in benzene at 0
◦C for 30 minutes and af-
forded (E )-cyclododecen-1-ol in a yield of 90 % when added to 0.25 equivalents of
(E )-cyclododecene oxide and stirred for 3 hours at room temperature. The research
showed that the nature of the amine was a very important factor in the effective-
ness of the base, recording low yields when TMP was replaced by diethylamide
(< 5 %), dicyclohexylamide (36 %), or diisopropylamide (45 %).∗ Yamamoto was
also able to apply this chemistry to the Fischer-Indole synthesis, demonstrating
that DATMP preferentially abstracts the anti-α-methylene proton of arylhydra-
zones (Figure 1.14), with much higher regioselectivity than had been recorded
when traditional catalysts were employed.68
N
N
R1
R2
R3
H
Al
N
Et Et
Figure 1.14: Et2Al(TMP)-promoted regioselective deprotonation of arylhydra-
zones68
While the chemistry of aliphatic aluminiums has been well developed, the synthesis
of aromatic aluminium species presented many challenges and so investigations into
their uses had been very limited. Conventional routes to aryl aluminiums typically
involved transmetalation with aryl lithium or aryl Grignard reagents,69 although
these have low tolerance towards many functional groups on aromatic rings (see
Section 1.2.2). In addition, hydro- and carboalumination, widely used in aliphatic
∗ Reactions stirred for 1 hour, 80 % yield with TMP under same conditions
22
1. INTRODUCTION
chemistry, are not suitable for aromatic systems due to the structural limitations
of benzene rings.70 This meant that a simple, effective route to arylaluminium
species would provide an opportunity to study their modes of action and allow a
targeted approach towards synthesising new reagents.
Uchiyama and coworkers, noting the failure of Et2Al(TMP) to deprotonate func-
tionalised aromatic compounds but recognising the success of tri-coordinated zin-
cate reagents in DoZn, looked to synthesise analogous tetra-coordinated aluminate
reagents. Following methods employed in lithium zincate synthesis (see Section
1.4), a 1:1 mixture of LiTMP and iBu3Al was prepared and reacted with a range
of functionalised aromatics, proving to be an effective DoM agent71 (Figure 1.15).
iBu3Al(TMP)Li
anisole
OMe
(99 %)
I
benzonitrile
CN
(100 %)
I
CON(iPr)2
(94 %)
I
PhCON(iPr)2
4-iodo-
anisole
(83 %)
OMe
II
(94 %)
CN
II
4-iodo-
benzonitrile
OMe
1,3-dimethoxybenzene
Cl
I
Cl
1,3-dichlorobenzene
(74 %)
(92 %)
I
OMe
Figure 1.15: Deprotonative alumination followed by I2 quench for various function-
alised aromatics (reaction conditions vary)71
The resulting aluminate base displayed high functional group tolerance in the
screening experiments, as no nucleophilic addition to nitrile or carbonyl groups,
halogen-metal exchange with iodo-substituents, or benzyne formation was ob-
served. While benzyne formation could be controlled by varying the alkyl ligand
on zincate bases, research on aluminates found the same control could be achieved
23
1. INTRODUCTION
with iBu3Al(TMP)Li just by varying the temperature. At –78
◦C N,N -diisopropyl-
3-bromobenzamide was converted via DoAl to the 2-iodo product in a 65 % yield,
while at room temperature, the reaction proceeded via a 3-functionalised benzyne
to a Diels-Alder adduct with 1,3-diphenylisobenzofuran forming in a quantitative
yield.
After the initial reactivity studies with lithium aluminates, the first crystal struc-
ture to be solved was from the work of Mulvey et al. which was the sodium
aluminate (iBu)2Al(µ-TMP)(µ-
iBu)Na · TMEDA (Figure 1.16(a)).72 Containing
the same ligands as the base Uchiyama et al. had been working with, it once again
suggested that the proposed formulation of the active species was correct. In the
solid-state, the TMP ligand and one of the alkyl ligands bridge between the two
metal centres, forming an (AlCLiN) metallocycle. Interestingly, it is the α-carbon
which forms an agostic bond to the Al centre and the bridging and terminal Al−C
bonds are essentially identical in length (2.038(2) and (mean) 2.037 A˚ respectively).
The geometry around the Al centre is distorted tetrahedral, with subtending bond
angles varying between 99.85(11) and 120.57(10)◦. Due to Uchiyama’s success
using lithium as the second metal, a reaction mixture containing iBu3Al, LiTMP
and TMEDA was prepared, to which N,N -diisopropylbenzamide was added, result-
ing in the crystallisation of a remarkable heterobimetallic-heterotrianionic species,
2-iBu2Al{Me2NCH2CH2N(Me)CH2}{C6H4C(−O)N(
iPr)2}Li · {PhC(−O)N(
iPr)2}
(Figure 1.16(b)).
(a) (b)
Figure 1.16: Molecular structures of (a) (iBu)2Al(µ-TMP)(µ-
iBu)Na ·
TMEDA and (b) 2-iBu2Al{Me2NCH2CH2N(Me)CH2}{C6H4C(−O)N(
iPr)2}Li ·
{PhC(−O)N(iPr)2}
72
The structure contains an Al centre bonded to the ortho-carbon of the N,N -
24
1. INTRODUCTION
diisopropylbenzamide with loss of TMP, suggesting a DoAl mechanism analogous
with DoZn (see Section 1.4). However, the Al is only bonded to two iBu groups,
the third having been eliminated, and to a CH2 formed by the deprotonation of
a TMEDA methyl group. Both N centres of the TMEDA are donating to the Li
centre, which completes its coordination sphere with bonds to oxygens both on
the deprotonated benzamide and non-metalated benzamide. As both an alkyl and
amido group were lost it was not clear whether the deprotonation was alkyl- or
amido-mediated, although the presence of deprotonated TMEDA suggested that
the aluminate base exhibits dual alkyl/amido basicity as in DoZn. Even though
the second equivalent of benzamide had not been deprotonated, it is possible that
the presence of TMEDA means that the ortho proton is sterically inaccessible.72
At the same time as Mulvey was carrying out solid-state investigations, Uchiyama
was conducting further studies of aluminate bases, and with Wheatley and cowork-
ers, crystallised iBu3Al(TMP)Li·THF. The structure of this aluminate is analogous
to Mulvey’s species, although the Li centre is only three coordinate, which is most
likely due to the absence of a chelate effect present when TMEDA (containing two
Lewis basic nitrogens) is coordinating to the Li centre. While NMR spectroscopic
studies showed that, in solution, there was no reformation of LiTMP or iBu3Al,
it was unclear whether the solid-state structure remained intact, as all three iBu
groups appeared in the same position. DFT studies agreed with the solid-state
data, predicting a more stable structure with Li and Al centres bridged by one
alkyl and one amido group, rather than two alkyl groups and a three-coordinate
Li centre. Using this structure, GIAO calculations agreed well with the experi-
mental NMR spectroscopic values. The choice to study the combination of LiTMP
and iBu3Al had come from initial screening studies which identified three key con-
stituents of aluminate complexes, the amido group, the alkyl substituent and the
co-metal. The inflexible and bulky TMP proved more effective than more flexible,
less sterically demanding candidates iPr2N and TMS2N while K, MgCl,
tBuMg and
tBuZn all gave complex mixtures of products. To find the best alkyl substituent, a
systematic test was undertaken with a series of alkyl aluminium compounds which
were combined with LiTMP and reacted with two different aromatics, one con-
taining an electron-withdrawing directing group (benzonitrile) and the other an
electron-donating directing group (anisole), followed by a quench with I2. With
anisole both Et3Al and
iBu3Al were promising candidates, while when benzoni-
trile was the substrate only iBu3Al led to the isolation of the required product,
2-iodobenzonitrile (see Table 1.1).14
25
1. INTRODUCTION
Compound Reagent Yield (%)
PhOMe Me3Al(TMP)Li 12
PhOMe Et3Al(TMP)Li 87
PhOMe iBu3Al(TMP)Li 88
PhCN Me3Al(TMP)Li 0
PhCN Et3Al(TMP)Li 0
PhCN tBu3Al(TMP)Li 0
PhCN iBu3Al(TMP)Li 100
Table 1.1: Screening of various reagents for DoAl of functionalised benzene com-
pounds14
Next followed a detailed mechanistic study of the reaction pathway, starting with
NMR spectroscopic characterisation of aluminated aromatics. A representative
aromatic, anisole, was chosen and reacted sequentially with tBuLi and iBu3Al and
an NMR spectrum was recorded after each step. The movement of the ortho-
carbon at each stage could be clearly tracked in the 13C NMR spectra, showing
that the lithium was being displaced by the aluminium. When anisole was reacted
directly with the aluminate base, the NMR spectrum recorded matched the final
spectrum from the former experiment, meaning that the aromatic species was being
aluminated. DFT calculations on the reaction between anisole and a representa-
tive base, Me3Al(N2Me)Li ·OMe2 (Scheme 1.11), found that all plausible pathways
involved the formation of a relatively stable first intermediate, a coordination com-
plex between the Li and anisole oxygen centre (IM1), a required step for any
CIPE-mediated process. To study this intermediate further, in situ FT-IR spec-
troscopy was used to follow the interaction between N,N -diisopropylbenzamide,
chosen due to the presence of a carbonyl bond, and iBu3Al(TMP)Li. The ben-
zamide carbonyl stretch appeared at 1636 cm−1, and when the aluminate base
was added at –25 ◦C a transient peak at 1585 cm−1 appeared followed by a peak
at 1600 cm−1. As no peaks indicating the presence of lithiated aromatic species
(which would be expected to appear at 1578 cm−1) were observed it suggested the
presence of an interaction between the Li centre and the benzamide carbonyl. The
transient peak was attributed to the intermediate, and the peak at 1600 cm−1 to
an aluminated aromatic. X-ray crystallographic data for an initial electrostatic
adduct between N,N -diisopropylbenzamide and Me3Al(TMP)Li (IM1) was later
obtained (Figure 1.17), consistent with these findings.14 Two crystallographically
independent rotamers were observed in the asymmetric unit of the adduct, with
26
1. INTRODUCTION
the Li centre in contact with the amide O centre in both cases. The only notable
difference between the rotamers are the torsional angles between the C−O and
Li−N bonds (53.0(10)◦ (for the rotamer shown in Figure 1.17) and 40.6(13)◦) in-
dicating that there is some flexibility in the arrangement of the two components
in the other rotamer forming the adduct.
Figure 1.17: Molecular structure of one of the rotamers of Me3Al(TMP)Li·N,N -
diisopropylbenzamide14
As with the zincate bases, two possible DoAl mechanisms were postulated, alkyl-
and amido-mediated deprotonation. The amido-mediated pathway goes via a tran-
sition state with a much lower activation energy, +29.8 kcal mol−1, compared with
+47.9 kcal mol−1 for the alkyl-mediated pathway, agreeing with Yamamoto’s work
which highlighted the higher reactivity of amido ligands over alkyl ligands in such
systems.68
The parallels with the calculations on zincate species are clear; while the alkyl-
mediated pathway gives the most thermodynamically stable product, the amido-
mediated pathway is kinetically favourable and still gives products which are ther-
modynamically favourable compared with the starting materials (Scheme 1.11). It
was therefore logical to find out if the re-coordination and subsequent deprotonation
of the amine was a plausible pathway. On this occasion, N,N -dimethylbenzamide
was employed as the substrate and the calculations shown in Scheme 1.11 were
modified to include a third transition state, TS3, to provide a pathway between
IM3 and IM4 (Figure 1.18).
The change from anisole to N,N -dimethylbenzamide has only a minor effect on the
relative energy levels of the intermediates and transition states (except IM1/3, as
the benzamide is a better Lewis donor). Interestingly though, the kinetic barrier
27
1. INTRODUCTION
IM1
RT
TS1
TS2
IM3
IM2
IM4
TS3
IM4
IM2
+0.0
+30.7
+54.8
+2.2
–8.4
+24.6
∆E (kcal mol-1)
0.0
10.0
20.0
30.0
–10.0
+30.7
+52.6
+13.5
40.0
50.0 +51.2
+37.7
HH
O
Me2
Al
NMe2Li
S
H
H
Me2N
H H
O
AlMe2
NMe2Li
S
H
H
Me2N
O
AlMe2
Me
N
Me2
Li
S
H
Me2N
Me2N
H
O
AlMe
N
Me
Me
Li
S
Me
Me
S = Me2O
+
Deprotonation with
alkyl ligand
Deprotonation with
amido ligand
Figure 1.18: Summary of deprotonative alumination pathways of the reaction
between N,N -dimethylbenzamide and Me3Al(NMe2)Li · OMe2 at the B3LYP/6–
31+G* level of theory. Energy changes are shown in kcal mol−1 and are relative to
IM173
28
1. INTRODUCTION
IM1
IM3 TS2
AlMe2
NMe2
Li
S
H
H
H
AlMe2
Me2
N
Li
S
H
H
O
AlMe2
NMe2
Li
S
O
AlMe2Me
NMe2Li
S
O
H AlMe2
Me2
N
Me
Li
S
O
IM2TS1
IM4
O
H
AlMe
N
Me
Me
Li
S
Me
Me
RT
S = Me2O
+ CH4
Deprotonation with
amido ligand
Deprotonation with
alkyl ligand
+
H
–15.8 +29.8 –16.2
+0.5
+47.9 –55.1
AlMe2Me2N
MeLi
S
O
H
O
Me
Scheme 1.11: Modelled pathways for the deprotonative alumination of anisole by
Me3Al(NMe2)Li ·OMe2; deprotonation by amido ligand (IM1-TS1-IM2) and de-
protonation by an alkyl ligand (IM3-TS2-IM4). Energy changes (∆E) at the
B3LYP/6–31+G* level are shown in kcal mol−114
to alkyl-mediated deprotonation increases proportionally more than the barrier to
amido-mediated deprotonation. The energetic barrier to alkyl deprotonation of the
amine was found to be 37.7 kcal mol−1, a kinetic penalty most likely too high for
a multi-deprotonation pathway to be viable.73 These findings were inconsistent
with Mulvey’s structural studies (Figure 1.16(b)), so it was clear further research
would be necessary to elucidate an answer.
Wheatley et al. were able to crystallise two aluminated species via a transmeta-
lation reaction, by treating N,N -diisopropylbenzamide and N,N -diisopropylnapth-
amide sequentially with tBuLi and iBu3Al in THF. An identical aluminated benza-
mide was also crystallised by Mulvey et al. by treating the substrate directly with
iBu3Al(TMP)Li, and in both cases the products were tris-THF solvated, mono-aryl
aluminates (Figure 1.19). Cryoscopic measurements on the two species showed par-
tial desolvation occurring in both cases, but after addition of excess HTMP to either
pre-isolated or in situ generated 2-(iBu3Al)-C6H4C(−O)N(
iPr)2Li ·3THF followed
by heating at reflux, crystalline deposits of the unreacted aluminated-aryl species
formed at –30 ◦C. Spectroscopic data supported these findings as no change in
shifts was observed after the addition of HTMP, diethylamine or diisopropylamine
29
1. INTRODUCTION
to either the aluminated benzamide or napthamide in either Lewis donor (THF)
or non-polar (benzene) solvents.
(a) (b)
Figure 1.19: Molecular structures of (a) iBu3Al{C(O)N(
iPr)2(C10H6)}Li · THF
73
and (b) iBu3Al{C(O)N(
iPr)2(C6H4)}Li · THF
74
The failure of aluminate reagents to exhibit polybasicity can be explained by the
electronic nature of Al in such species. As the Al atom is coordinatively saturated,
it is unwilling to act as a Lewis acid and accept a pair of electrons from the incoming
amine ligand, reflected in an Al−N bond length of 3.56 A˚ in TS3 (Figure 1.18).
This is compared to zincate bases, where the analogous transition state complex has
a Zn−N bond length of only 2.77 A˚ (Scheme 1.10, TS3; bond length not shown),
and as the Zn atom has just 16e− in its outer shell (except in the transition state),
it will readily accept a pair of electrons from the incoming amine56,73 (Figure 1.20).
Me2N O
Al
H
H
H
H
Li
S
N
(a)
Me2N O
Zn
H
H
H
H
Li
S
N
(b)
Figure 1.20: Differing abilities of metalated complexes to undergo methyl mediated
amido deprotonation; unfavoured in (a) DoAl, but favoured in (b) DoZn73
More recently, much as cheaper routes to zincates have been investigated (see
30
1. INTRODUCTION
Section 1.4), the same methods have been applied to aluminate bases. To make
a comparison between TMP, DMP and HMDS ligands, Mulvey et al. synthesised
a series of potassium aluminates of general formula iBu2Zn(µ-
iBu)(µ-amide)K ·
PMDTA (amide = TMP, DMP, HDMS). While the central (KNAlC) ring in the
TMP-based species was found to be essentially planar, the lower steric demand of
the DMP ligand led to a puckered ring forming in the solid-state (Figure 1.21). The
ring puckering creates a gap in the coordination sphere of the K centre which is filled
by a weak agostic interaction between the alkali metal and the terminal alkyl group
(K· · ·C) separation 3.567(2) A˚). This agostic interaction is precluded in the TMP-
based species due to the extra methyl groups, when compared to DMP, sterically
blocking such an approach. While the HMDS-based species deviates slightly from
having a planar ring, in this case the deviation was attributed to weak agostic
interactions between the carbons from the amido ligand and the K centre.75
(a) (b)
Figure 1.21: (a) Molecular structure of iBu2Zn(µ-
iBu)(µ-DMP)K · PMDTA and
(b) projection of the central (KNAlC) core displaying a puckered conformation75
1.6 Directed ortho Manganation
While the DoM agents already discussed contain main-group or pseudo-main-group
metals, in recent years the chemistry has been extended to include a transition
metal, manganese, due to the greater range of properties (e.g. redox, catalytic,
magnetic) that it could potentially provide. Initial attempts to form a manganate
base required the researchers to be careful with their choice of ligands and sol-
vents, due the instability of MnII compounds and their preference for forming
tetra-coordinated species. To this end, [Mn(CH2SiMe3)]∞ was chosen as a reagent
31
1. INTRODUCTION
due to its thermal stability, TMP as a ligand due to its sterically demanding nature
and strong basicity and TMEDA as the coordinating solvent to prevent polymeri-
sation at the Li centre. From the reaction between the three components a lithium
manganate base very similar in structure to previously discussed ’ate complexes,
(CH2SiMe3)2Mn(µ-TMP)Li · TMEDA (Figure 1.22), was crystallised. The struc-
ture displays distorted trigonal planar geometry around the Mn centre (bond angle
sum = 358.09◦) and the TMP ligand bridging between the two metal centres. How-
ever, whereas other ’ate complexes contain a second bridging centre, the relevant
Li· · ·C distance in the manganate structure is large (2.742(6) A˚) which, coupled
with the fact that the two Mn−C bonds are very similar in length (2.159(3)and
2.140(3) A˚), suggests that any Li· · ·C interaction is minimal. When the manganate
was reacted with ferrocene (Cp2Fe), the base exhibited dialkyl/monoamido triba-
sicity, forming a product, {Fe(C5H4)2}3Mn2Li2 · 2TMEDA in which the ferrocene
had been dideprotonated.76
Figure 1.22: Molecular structure of (CH2SiMe3)2Mn(µ-TMP)Li · TMEDA
76
Replacing LiTMP with NaTMP in the reaction mixture led to the unexpected
formation of a bisamido, monoalkyl manganate base, (TMP)Mn(µ-CH2SiMe3)(µ-
TMP)Na · TMEDA, which was capable of deprotonating benzene77 and selec-
tively ortho deprotonating anisole, N,N -diisopropylbenzamide13 and methoxynaph-
thalenes.78 Intermediates of the ortho deprotonation of both anisole and N,N -di-
isopropylbenzamide were isolated (Figure 1.23) and their structures showed that
while the deprotonation of anisole was alkyl-mediated, amido-mediated deprotona-
tion accounted for the reaction with N,N -diisopropylbenzamide. In each case, the
deprotonated α-carbon centre is bound to the Mn centre but is not interacting with
the Na centre, which instead is bound to the oxygen, forming six-membered (NaN-
MnCCO) and seven-membered (NaNMnCCCO) rings (Figures 1.23(a) and 1.23(b)
32
1. INTRODUCTION
respectively). When the species were coupled with iodobenzene in the presence of
a PdII catalyst, 2-methoxybiphenyl formed in a yield of 98 %, although the isolated
yield of N,N -diisopropyl-2-phenylbenzamide was much lower at 66 %.13
(a) (b)
Figure 1.23: Molecular structures of (a) (TMP)Mn(µ-TMP)(2-C6H4OMe)Na ·
TMEDA and (b) (CH2SiMe3)Mn(µ-TMP){2-[C(O)N(
iPr)2]C6H4}Na · TMEDA
13
1.7 Limitations of Modern DoM Methods
The advances in DoM achieved through the employment of magnesium, zinc, alu-
minium and manganese ’ate complexes have been invaluable to the scope of re-
actions which are now possible. Through careful control of alkyl and amido sub-
stituents, high regioselectivity has been achieved for a wide variety of functional
groups, allowing highly selective and efficient syntheses of complicated organic
molecules which were previously only achievable in low yields, if at all. The metals
each present their own nuances to the chemistry; for example while DoZn operates
under thermodynamic control, allowing zincates to exhibit tri-basicity meaning
they can be employed in substoichiometric amounts, DoAl operates under kinetic
control and hence avoids potentially problematic ligand scrambling effects. Un-
fortunately, these ’ate complexes still present one key drawback, as the formation
of carbon-carbon and carbon-heteroatom bonds often requires the addition of a
metal-based catalyst, usually palladium or copper-based.79,80 The requirement for
a catalyst increases the cost of syntheses, decreases atom efficiency and comes
with the possibility of unwanted side-reactions such as homo cross-couplings. Re-
cently, development has focused on the synthesis of DoM agents which are active
33
1. INTRODUCTION
in C−C and C−X formation without the need for a catalytic second metal. Cop-
per was chosen as the most promising candidate metal as it was already known
to facilitate the coupling reactions, and is inexpensive compared to palladium, the
other obvious choice. The syntheses, characterisation and reactivities of prospec-
tive copper-based DoM agents are the focus of this project.
In the following sections, a brief overview of organocopper chemistry will be pre-
sented, followed by comprehensive detailing of the research which provides the
basis for the work carried out towards this thesis.
1.8 Copper Based Reagents
Copper is a highly versatile metal, known for its useful properties of malleability,
ductility and high conductivity, making it ideal for use in electrical wires. Certain
copper alloys are employed for their biostatic properties, which act on a wide variety
of microorganisms such as E. coli81 and, given their high corrosion resistivity,
make excellent netting materials for the aquaculture industry.82 Liquid copper
compounds can be used as wood preservatives83 and specific isotopes of copper
(e.g. copper-60 and copper-62) can be used as radio-tracers in positron emission
spectroscopy.84 Copper nanoparticles are inexpensive to form and have been shown
to be effective catalysts, for example in multicomponent azide-alkyne 1,3-dipolar
cycloadditions.85 Chemically, copper is a late transition metal found in Group 11,
meaning a number of oxidation states are accessible. Of these, CuII is by far the
most stable meaning that most copper salts are limited to this oxidation state,
although in organocopper chemistry, CuI compounds are most common.
1.8.1 Organocopper compounds
The first synthesis of an organocopper compound was achieved in 1859, when
Bo¨ttger passed acetylene gas through copper chloride (CuCl) to form copper acetyl-
ide (Cu−C−−C−Cu).
86 Unfortunately, the explosive nature of this compound de-
terred chemists from further research in the area for over 60 years, until Reich
successfully prepared phenylcopper from copper iodide (CuI) and a phenyl Grig-
nard reagent in 1923.87 This paved the way for much of the current research in the
field, which has diversified to the extent that the classical definition of organocopper
compounds containing copper-carbon bonds is now normally extended to include
34
1. INTRODUCTION
copper-heteroatom bonds, for example copper amides and phosphides.
Examples of isolated and characterised organo-CuIII and CuII compounds are rare,
and the first reported example of the former was published in 1989 by Burton et
al. The oxidation of a CuI species, [(CF3)2Cu]
– [CdI]+, with thiuramdisulphide
formed Et2NCS2Cu
III(CF3)2 (Figure 1.24), which displays distorted quadratic pla-
nar geometry around the central Cu centre and short, asymmetric Cu−C bond
distances (mean 1.945 A˚).
Figure 1.24: Molecular structure of Et2NCS2Cu
III(CF3)2
88
Recently, evidence for the formation of an organo-CuIII intermediate in conjugate
addition to α,β-unsaturated carbonyls (see Section 1.8.2.1 for a more detailed dis-
cussion of this reaction) has been presented. While theoretical calculations had
predicted that the isolation of such an intermediate should be possible,89 it took
several years before experimental work by Bertz et al. was able to prove its exis-
tence. Rapid-injection NMR (RI-NMR) showed the presence of Me peaks shifted
downfield with respect to the starting materials and other isolated intermediates
only attributable to a CuIII species.90 The instability of organo-CuII compounds in
the solid-state is most likely due to their susceptibility to undergo reduction to a
Cu0 species and an oxidatively coupled product. This characteristic is utilised in
coupling reactions to form complex organic molecules, for example biaryls.91,92
In the +1 oxidation state, copper has a full shell (d10) electron configuration and
so there is no electron preference for any particular geometry. This means steric
factors determine the structures of organo-CuI compounds, and they usually limit
the coordination number to two (linear) or three (trigonal planar). In most circum-
stances, aggregation states greater than one are observed, in which ligands form
3-centre-2-electron bonds and bridge between two copper centres,93 for example in
the structure of [Cu(CH2SiMe3)]4
94 (Figure 1.25). The exception to this are com-
35
1. INTRODUCTION
pounds with bulky ligands, such as Cu-NHCs (NHC = N-heterocyclic carbene),
which have attracted considerable interest in recent years and are used as catalysts
in a variety of organic transformations.95,96
Figure 1.25: Molecular structure of [Cu(CH2SiMe3)]4
94
One disadvantage posed by organocopper compounds is their thermal instability,
which is caused by the weak nature of Cu−C and Cu−X bonds (X = N, P, S etc.)93
This is best demonstrated by methylcopper, which is explosive above –15 ◦C. If
minor colloidal Cu0 impurities are present in a sample even lower decomposition
temperatures are observed, only serving to worsen the issue.97 One solution is to use
multiply-bonded carbon ligands, as thermal stability was found to increase going
from sp3- to sp-hybridised carbons, although even then it is still relatively poor.98
As Cu−X bonds are stronger than Cu−C bonds, heteroatom-based ligands acting
as external or intramolecular donor groups can alleviate this problem. For example,
CuMe(PPh3)3 has a decomposition temperature of around +75
◦C.99 Intramolecu-
lar stabilisation is well demonstrated in the crystal structure of [Cu2(SAr-2)(Mes)]2
(Ar = C6H4(CH2NMe2)), where two of the Cu centres are dative covalently bound
by a tertiary amine connected to an aryl ring (Figure 1.26).
The four copper atoms adopt a puckered “butterfly” conformation with symmetri-
cally bridging mesityl and thioarene ligands, typical of electron-deficient 3-centre-2-
electron Cu−C−Cu bonds.100 The Cu−Cu distances of coppers bridged by mesityls
are very short (2.3805(7) and 2.4075(7) A˚) while the much larger separation be-
tween copper atoms bridged by thioarenes (2.7952(7) and 2.6983(7) A˚) and acute
Cu−S−Cu angles of 74.82(2) and 78.09(2)◦ are in the ranges expected for electron-
deficient Cu2S bonding.
94
36
1. INTRODUCTION
S
Cu
Cu
S
Cu
Cu
N
N
Figure 1.26: Intramolecular stabilisation of an organocopper compound by a hetero-
atom-containing organic group
Any inter- or intramolecular stabilisation afforded to organocoppers also serves
to deaggregate the species, which increases reactivity as the charge becomes more
localised on the R– or X– moiety. In addition, the extra coordination means that the
compounds become more soluble in a wide variety of solvents. This deaggregation
effect can clearly be seen on the addition of THT to mesitylcopper (MesCu). In
non-donor solvent, MesCu forms a pentameric aggregate in the solid-state, but
upon addition of THT, deaggregation to a tetramer occurs (Figure 1.27).101,102
(a) (b)
Figure 1.27: Differing aggregation states of mesitylcopper forming (a) a pentameric
species in a non-donor solvent and (b) a tetrameric species in the presence of a donor
solvent (THT)101,102
The deaggregation occurs as the addition of a donor solvent provides both electronic
stabilisation to and increases steric demand around the copper centre. When the
use of donor solvents is combined with sterically bulky ligands, it is possible to
37
1. INTRODUCTION
observe monomeric organocopper species in the solid-state. While copper 2,4,6-
triisopropylbenzene forms a tetramer in hydrocarbon media (Figure 1.28(a)), Power
et al. found that by treating CuBr with Li(C6H2-2, 4, 6-
tBu3) in the presence of
DMS affords monomeric Me2SCu(C6H2-2, 4, 6-
tBu3) (Figure 1.28(b)). While the
Cu−C bond in the monomer is comparable in length to those in other organocopper
species, the Cu−S bond is notably shorter (by ca. 0.2 A˚) than in other Cu-SMe2
complexes, indicating that the electronic stabilisation afforded by the donor solvent
is increased on deaggregation.
(a) Molecular structure of tetrameric spe-
cies Cu
4
(C
6
H
2
iPr3-2, 4, 6)4
(b) Molecular structure of monomeric
species Me
2
SCu(C
6
H
2
-2,4,6-tBu3)
Figure 1.28: Influence of steric and solvent affects on organocopper aggrega-
tion103,104
From the highly unstable methylcopper, advances in the field provided routes to
stable, soluble and reactive organocopper species. The chemistry has been utilised
to provide synthetically useful R– anions, in an analogous manner to to the pre-
viously discussed lithium- (Section 1.2), magnesium- (Section 1.3), zinc- (Section
1.4), aluminium- (Section 1.5) and manganese- (Section 1.6) based reagents.
1.8.2 Organocuprates
The most prominent method by which organocopper compounds can provide re-
active R– anions is via organocuprates, and they have been put to use in a wide
variety of organic syntheses. Organocuprates were first observed by Gilman in 1952,
during investigations into methylcopper. He discovered that addition of one equiv-
alent of MeLi to CuI resulted in the formation of a highly reactive, ether-insoluble
yellow product, but on addition of a further equivalent of MeLi an ether-soluble
38
1. INTRODUCTION
clear product formed.105 Later research showed that on replacement of CuI with
CuSCN the same products were observed, suggesting that the products were not
dependent on nature of the copper salt. The products were therefore proposed
to be ether-insoluble (MeCu)∞ and ether-soluble Me2CuLi · nS.
106,107 Significant
contributions to the field have also been made by Lipshutz, who in 1981 noticed
that substituting CuI for the cheaper, more stable CuCN in the formation of cu-
prates led to much higher reactivities in halide substitutions. He proposed that
the cyanide anion remained attached to the Cu centre and suggested that this
accounted for the increased reactivity.92
As cuprates began to be widely employed as synthetic reagents, interest grew in
their structure and mechanism of action. While NMR spectroscopic studies pro-
vided evidence for the presence of cuprate species in solution,106,107 solid-state
isolation and characterisation initially proved difficult. The first crystallographic
evidence was provided in 1982 by the structure of [Ph6Cu5]
– [Li · 4THF]+ (Figure
1.29(a)), an ion-separated species with an anionic copper cluster, prepared from
the reaction of CuBr with a slight excess of PhLi.108 This was swiftly followed
by a similar species, [Ph6Cu3Li2]
–
2 [Li4Cl2 · 10OEt2]
2+ (Figure 1.29(b)), prepared
from the reaction of CuCN with two equivalents of PhLi∗. This contained the first
example of an anionic organocopper-lithium cluster.109 The Cu5 cage in [Ph6Cu5]
–
is best described as a “squashed” pentagonal bipyramid with short Cu(ax)−Cu(eq)
distances (2.449 A˚av.) compared to the Cu(eq)−Cu(eq) distances (3.151 A˚av.) and
phenyl groups bridging each of the edges. The Cu3Li2 cage in [Ph6Cu3Li2]
– is
structurally very similar, maintaining the same geometry with Li replacing Cu in
the axial positions (Li(ax)−Cu(eq) 2.639 A˚av., Cu(eq)−Cu(eq) 3.318 A˚av.).
While the structure of [Ph6Cu3Li2]
–
2 did not appear to agree with proposed formu-
lations for either Gilman or Lipshutz cuprates, their formation can be explained
by the presence of excess organolithium in the reaction mixture being incorporated
into the main cage. Recent research has shed more light on Gilman and Lipshutz’s
observations, with work showing that higher-order cuprates analogous to Lipshutz
cuprates but containing halides or R groups can be synthesised, suggesting that
the broader term “Lipshutz-type” is more appropriate to describe the wider group
of comparable cuprates. This means that organocuprates can be classified into one
of two distinct families: either lower-order Gilman-type cuprates (which can be
neutral (RCuR′Li ·nS) or charge-separated ([RCuR′]– [Li ·nS]+) species), or higher-
order Lipshutz-type cuprates of the form RCuR′Li2(X) · nS (X = R
′′, CN, halide).
∗ The presence of chloride in the crystal structure was due to impurities in the CuCN
39
1. INTRODUCTION
(a) (b)
Figure 1.29: Molecular structures of the anionic cages of (a) [Ph6Cu5]
– [Li ·
4THF]+108 and (b) [Ph6Cu3Li2]
–
2 [Li4Cl2 · 10OEt2]
+
2
109 showing the strong simi-
larities between the two species
These two families will be discussed in more detail in the following sections.
1.8.2.1 Lower-order organocuprates
In attempting to synthesise lithium cuprates from CuI and (halide-free) organolithi-
ums in the presence of a crown ether, 12-c-4, Power et al. successfully abstracted the
lithium ion with the crown ether, and crystallised two charge-separated Gilman-
type species, [Me-Cu-Me]– [Li · 2 (12-c-4)]+ and [Ph-Cu-Ph]– [Li · 2 (12-c-4)]+.110
The first examples of mononuclear cuprate species, both structures display two-
coordinate, linear geometry across the C−Cu−C bond and contain C−Cu bonds
(1.935 and 1.925 A˚av. respectively) consistent with previously known organocopper
compounds.
While 12-c-4 was an important aid to crystallisation, the charge-separated species
described above exhibited poor solubility compared to their oligomeric counter-
parts. This led to considerable research into solvent effects on the structures of
cuprates. When THF and DME were employed as solvents in reactions involving
TMS3C
– and Me– anions respectively, charge-separated species were isolated in
40
1. INTRODUCTION
(a) (b)
Figure 1.30: Molecular structures of the anionic moieties of (a) [Me2Cu]
– [Li ·2 (12-
c-4)]+ and (b) [Ph2Cu]
– [Li · 2 (12-c-4)]+110
both cases: [(TMS3C)2Cu]
– [Li · 4THF]+ containing a four-coordinate lithium111
and [Me2Cu]
– [Li · 3DME]+ containing a six-coordinate lithium.112 However, using
more weakly donating solvents, the isolation of associated structures (Ph2CuLi·nS)2
(S = Et2O, n = 1.5;
113 S = DMS, n = 1114) was achieved. The use of DMS was
particularly of note, as not only was it a good solvent for copper halides but had
also been shown to confer greater stability and reactivity to cuprate species.115
One application of lower-order organocuprates is in conjugate addition to α,β-
unsaturated carbonyls, as shown in Scheme 1.12. While hard organolithium com-
pounds undertake nucleophilic addition, as discussed in Section 1.2.2, soft Gilman-
type cuprates attack the β-carbon forming the 1,4-substituted product.116 To un-
derstand the mechanism of this reaction, knowledge of the behaviour of cuprates
in solution is required. However, NMR spectroscopic and molecular weight studies
of cuprate systems indicated that RCuLi, R3Cu2Li, R3Cu2Li2 and R5Cu3Li2 were
all present in solution.106,107,117 In a one-pot synthesis utilising a cuprate, any LiX
that forms will still be present, meaning that a Gilman-type cuprate could perhaps
be viewed as R2CuLi · nS + LiX, or if the LiX is interacting with the cuprate then
the reactive species is, in fact, a higher-order cuprate.
O
R1
R2 O
R1
R2
R3
(i) R32CuLi
OH
R1
R2
R3
(ii) H+(ii) H+
(i) R3Li
Scheme 1.12: Comparison of reactivity of organolithiums and organocuprates116
1.8.2.2 Higher-order organocuprates
As briefly discussed in Section 1.8.2, increased reactivities of cuprates had been
observed when replacing CuI with CuCN92 and also when using a 3:1 stoichiometry
41
1. INTRODUCTION
of organolithium:copper halide.118 It was suggested that this could be explained
by the formation of a higher-order cuprate, although while the body of evidence
of Gilman-type cuprates was growing, the evidence for the existence of higher-
order cuprates was sparse. In 1990, Bertz published a paper disputing Lipshutz’s
proposal that CN– was incorporated into the structure of the cuprate. The data he
provided were detailed NMR spectroscopic studies on aryl and alkyl cuprates of the
formulation R2Cu(X)Li2 (X = I, CN) deliberately picking anions from opposite ends
of the spectrochemical series. The 1H, 13C or 6Li spectra were all identical regardless
of the nature of X, contrary to what would be expected if the cyano group was
bound to the Cu centre, in which case electron density would be donated to the Cu
centre and so affect the electronic properties of the compound.119 Although Bertz
entertained the possibility that the cyano group may not be exerting any effect on
the NMR shifts, he was much more confident that he was not observing higher-order
cuprates in solution. In an extraordinary turn of events, the article following Bertz’s
was authored by Lipshutz on the very same topic, providing contrary evidence to
Bertz’s studies and pointing towards the existence of higher-order organocuprates
in solution. Lipshutz carried out similar NMR spectroscopic experiments, and
observed changes in shifts on addition of LiCN to Me2Cu(I)Li2, which he concluded
supported his proposal that the CN– binds directly to the Cu centre.120 However,
while he suggested a structural change on addition of LiCN, there are alternative
explanations; for example nitrile coordination to the R group via a lithium centre.
Theoretical studies by Snyder indicated a wide variety of stable structures for cupra-
tes were plausible (Figure 1.31) and he concluded that three-coordinate copper was
neither obligatory nor sufficiently stable and that viable alternatives based on the
known structure of Gilman-type cuprates existed.121 Coupled with other evidence,
these results caused Lipshutz to rethink his previous assumptions, and after car-
rying out further NMR spectroscopic experiments and collecting IR data, he came
up with a new proposal for a bonding model. This described the nitrile ligand both
σ-bound to an Li+ and pi-bound to the Cu centre (Figure 1.32), a bonding mode
that was well-precedented in manganese, rhodium and molybdenum chemistry.122
Snyder continued the debate with further theoretical studies concluding that the
nitrile must be bound to the lithium centres, the most likely structures be-
ing a or b in Figure 1.31,123 which was confirmed by subsequent solid-state
studies. These studies discovered the structures of two cyanocuprates, the ion-
separated [tBu2Cu]
– [(CN)Li2 · 2THF · 2PMDTA]
+ (Figure 1.33)124 and polymeric
[(C6H4CH2NMe2)2Cu(CN)Li2]∞ (Figure 1.34).
125 In the former, the use of two
42
1. INTRODUCTION
Me Cu Me
Li Li
N
C
OH2H2O
Me Cu Me
Li Li
OH2H2O
N
Me
Li Li
N
C
Cu
Me
OH2H2O
Me
Cu
Me
Li
N
Li
C
OH2H2O
N Cu Me
Li Li
Me OH2H2O
C
Li
Me Me
Cu
C
N
Li(OH2)
OH2
Me Cu Me
Li Li
OH2H2O
C
N
Me
Cu
Me
Li Li
N
C
OH2H2O
a b c d
e f g h
Figure 1.31: Calculated possible structures of lower- and higher-order cuprates121
CN
Cu R
Li
R
O
Figure 1.32: Bonding diagram for higher-order cuprates as suggested by Lipshutz122
strong donor solvents prevented the tert-butyl groups bridging between the Cu
and Li centres and an ion-separated species containing two five-coordinate lithiums
crystallised, which cryoscopy measurements showed to remain intact in THF solu-
tions. In the latter structure, a similar motif is observed (Figure 1.34). In this case
the lithium centres are four-coordinate, bound externally by two THF molecules
and intramolecularly by one aryl NMe2 group, but each Li−C≡N−Li unit is bridg-
ing two Ar2Cu units, which causes the species to polymerise. While the structures
matched Snyder’s rather than Lipshutz’s predictions more closely, demonstrating
that the nitrile will bind preferentially to lithium over copper, due to Lipshutz’s
pioneering work in the field, higher-order cyanocuprates still bear his name.
43
1. INTRODUCTION
Figure 1.33: Molecular structure of [tBu2Cu]
– [(CN)Li2 · 2THF · 2PMDTA]
+,
displaying an ion-separated species with a [Li−C≡N−Li]+ unit containing five-
coordinate lithium solvated by PMDTA and THF124
Figure 1.34: Molecular structure of (C6H4CH2NMe2)2Cu(CN)Li2, showing a poly-
meric species with four-coordinate lithium intramolecularly bonded to bridging
NMe2 substituents
125
1.8.2.3 Recent investigations into organocuprates
Recent work on Gilman- and Lipshutz-type cuprates has provided information on
solid-state and solution behaviour, as well their abilities as Directed ortho Cupra-
tion (DoCu) reagents. Examples of cuprate complexes are numerous, with both
homo- and heteroleptic structures known, containing a wide range of substituents
including alkynyls,126 alkoxides,127 amides,128 phosphides129 and halides.130 For ex-
ample, the 3:1 reaction of lithium diphenylamide and copper chloride led to the iso-
lation of a Lipshutz-type monomer, (NPh2)2Cu(NPh2)Li2 ·2OEt2 (Figure 1.35(a)).
44
1. INTRODUCTION
The structure contains a six-membered cyclic core, with a NPh2 group bridging
between two lithiums and almost linear geometry across the Cu centre (N−Cu−N
177.4(2)◦), proving nitrogen to be an effective bridge between copper and lith-
ium centres.131 The use of chiral ligands was investigated by van Koten et al.,
who crystallised the intramolecularly stabilised [C6H4{CH2N(Me)CH2CH2NMe2}-
2]2Cu(Br)Li2 (Figure 1.35(b)) which exhibits chirality on coordination of NMe to
the Li centre. By introducing a second, enantiomerically pure, chiral centre at the
benzylic carbon of the ligand, an enantiomerically pure product was observed in
solution, which did not undergo interchange even at high temperatures.
(a) (b)
Figure 1.35: Molecular structures of (a) (NPh2)2Cu(NPh2)Li2 · 2OEt2
131 and (b)
[C6H4{CH2N(Me)CH2CH2NMe2}-2]2Cu(Br)Li2
132
Preparations of cuprates excluding any external or intramolecular donor groups
were explored by Davies et al., resulting in the formation of a Gilman-type dimer,
(Mes2CuLi)2 containing unusual 2-centre-2-electron Cu−C bonds and η
1,η6-co-
ordinated lithium (Figure 1.36). This is indicated by Cu−C distances (1.925(2) and
1.936(2) A˚) which are shorter than those usually observed in 3-centre-2-electron
cuprates, and the observation of two distinct Li−ligand distances (2.316 A˚av. for η
6-
Li, 2.129(4) A˚ for η1-Li). The C−Cu−C bond angle is near linear, consistent with
previously characterised cuprates, although the mesityl groups are twisted relative
to each other by approximately 9.6◦ to maximise bonding interactions to the Li.
DFT studies confirmed that η1,η6-coordination is the most stable conformation for
the cuprate (cf. η1,η1 and η6,η6 conformers) and NMR spectroscopic experiments
showed that the dimeric η1,η6 structure is retained in solution.
In order to explore the synthetic applications of lithium cuprates, a series of screen-
ing experiments investigating their ability to regioselectively deprotonate aromatic
compounds (DoCu) was carried out by Uchiyama et al. in 2007 (Table 1.2). Amido
45
1. INTRODUCTION
Figure 1.36: Molecular structure of [Mes2CuLi]2 showing η
1,η6-coordinated lith-
ium133
ligands were chosen due to their ability to act both as dummy (non-transferable)
ligands in C−C bond formation, which had previously been evidenced through
DFT calculations,134 and also as chiral auxiliaries. Initial studies indicated that
TMP would be particularly effective, hence experiments focused on reagents con-
taining this ligand. In all cases, benzonitrile was employed as the substrate, and
the products were trapped with I2.
15
CN CN
I1) Cuprate, THF, 0 oC, 3 h
2) I2, rt, 16 h
Cuprate Yield (%)i Cuprate Yield (%)i
a MeCu(CN)Li 0 i MeCu(NMe2)(CN)Li2 0
b TMPCu(CN)Li 0 j MeCu(NiPr2)(CN)Li2 53
c TMPCu · LiI 0 k MeCu(TMP)(CN)Li2 46
ii/91
d MeCu(TMP)(I)Li2 51 l
nBuCu(TMP)(CN)Li2 83
e (TMP)2Cu(I)Li2 0
ii/75 m tBuCu(TMP)(CN)Li2 70
f (NMe2)2Cu(CN)Li2 0
ii n PhCu(TMP)(CN)Li2 93
g (NiPr2)2Cu(CN)Li2 37 o ThCu(TMP)(CN)Li2 58
h (TMP)2Cu(CN)Li2 85
ii/74 p TMSCu(TMP)(CN)Li2 54
Table 1.2: Screening of cuprates for directed ortho cupration of benzonitrile
(iIsolated yield of ortho iodinated product, ii1.1 eq. cuprate used, all others 2.0
eq.)15
While cuprates with a Gilman-type formulation proved to be ineffective, where
46
1. INTRODUCTION
TMP was employed in tandem with another amido, alkyl or aryl ligand, Lipshutz-
type cuprates containing a cyano group gave high yields of iodinated product. The
success with an array of ligand combinations was a significant improvement from
DoZn and DoAl, both of which require strict structural arrangements to achieve
successful ortho deprotonation. Further synthetic studies showed that the reagents
were tolerant of a variety of functional groups, including benzamides, anisoles and
functionalised heteroaromatics, giving yields of ≥70 % of ortho iodinated prod-
uct in all cases.
The cuprates employed by Uchiyama et al. were synthesised in situ by the addition
of lithiated alkyl or amido species to CuCN. In testing the deprotonative metalation
ability of cuprates towards pyridyl species, Mongin et al. took a different approach
by using CuCl2·TMEDA as the reagent. Chloride reagents had previously proved to
be useful precusors from which to form reactive bases, for example TMPZnCl·LiCl79
and Al(TMP)3 · 3LiCl
80 prepared by Knochel and coworkers and Mongin’s own
species (TMP)3CdLi.
135 The CuII salt was chosen due to its higher air stability
compared to CuICl, and TMEDA was employed to improve moisture stability. To
create a reactive CuI reagent, the CuCl2 was initially reduced with an equivalent
of nBuLi before the addition of two equivalents of LiTMP136 (Scheme 1.13).
CuCl2
nBuLi
–½Bu-Bu
–LiCl
CuCl
LiTMP LiTMP
TMPCu
–LiCl
(TMP)2CuLi
Scheme 1.13: Mongin’s in situ synthesis of a TMP-cuprate from CuCl2 · TMEDA
and LiTMP136
While this preparation was relatively ineffective if I2 was used as the electrophile, it
proved to be capable of ortho deprotonating 1,4-dimethoxybenzene effectively and
benzoylations and cross-couplings of 1-chloro- and 1-fluoropyridine were achieved
in high yields (>75 %).136 Extending this study, it was found that a prepara-
tion starting from CuCl2 was also most effective in a bis-TMP formulation, which
was successfully used to deprotonate a variety of anisoles, furans, thiophenes and
methoxypyridines.137
Attempts to isolate a reactive amidocuprate were successful when LiTMP and
CuCN were mixed in a 2:1 ratio in THF; recrystallisation from toluene affording,
as predicted, a Lipshutz-type cuprate, [(TMP)2Cu(CN)Li2 ·THF]2, 1 (Figure 1.37).
The structure of 1 displays a near linear geometry across the N−Cu−N bond
(177.51(7)◦) with the CN bridging between the two Li centres, as had been observed
47
1. INTRODUCTION
previously in Lipshutz-type cuprate structures. However in this case, a dimer
formed, linked by a central (N2Li2) ring, which could possibly be attributed to a
deficiency of available Lewis-basic donors or the steric demand of the TMP ligands.
Figure 1.37: Molecular structure of [(TMP)2Cu(CN)Li2 · THF]2, 1
15
Further research into the synthesis of amidocuprates showed that the nature of
the cuprate species that forms is highly dependent on the solvent conditions. A
bis(amido) Gilman-type cuprate, [(TMP)2CuLi]2, 2 (Figure 1.38(a)) was isolated
from the 2:1 reaction of LiTMP and CuI in hydrocarbon media, whether recrys-
tallisation was carried out in the same solvent or from a mixture containing bulk
THF.138 In the latter case, the strong Lewis basic nature of THF abstracts the LiI
which remains in solution. However, the same reaction in toluene containing two
equivalents of THF (with respect to CuI) led to the formation of a Lipshutz-type
cuprate, (TMP)2Cu(I)Li2 · THF, 3 (Figure 1.38(b)), which is structurally analo-
gous to the cyanocuprate discussed above.138 Complex 2 represents an unusual
example of a lithium cuprate where the Li centres remain two-coordinate, even in
an excess of THF, and this can be explained by the large steric demand of the
TMP which prevents any further coordination to the Li centres. Interestingly, this
is not compensated for by the amido ligands, as the Li−N bonds in 2 are in fact
slightly longer (2.061 A˚av.) than those in 1 (1.965 A˚av.), which can be explained
by the steric demand of the ligands.
As alkyl- and arylamidocuprates were also promising candidates for DoCu, at-
tempts were made to characterise these species, in an effort to probe the identi-
ties of the putative phenyl- and methyl(TMP)cuprates, PhCu(CN)(TMP)Li2 and
MeCu(CN)(TMP)Li2. Reactions between 1:1:1 mixtures of RLi (R = Ph, Me),
48
1. INTRODUCTION
(a) (b)
Figure 1.38: Molecular structures of (a) [(TMP)2CuLi]2, 2, and (b)
[(TMP)2Cu(I)Li2]2, 3, which were both prepared from a mixture of HTMP and
LiI138
LiTMP and CuCN in donor solvents in fact led to the isolation of two Gilman-type
cuprates, PhCu(µ-TMP)Li · 3THF, 4, and MeCu(µ-TMP)Li ·TMEDA, 5, respec-
tively (Figure 1.39). The structures are both monomers in the solid-state and both
are unusual as the only terminal C−Cu bonds that had been observed previously
in lithium cuprates were those in ion-separated species (for example, [Me2Cu]
–
[Li · (12-c-4)2]
+, Figure 1.30). The structures suggested that to observe Li(µ-CN)Li
bonding in the solid-state strong inter-metal (Cu−R−Li) bonding is required for
both ligands. As screening experiments had suggested that Gilman cuprates were
unreactive in DoCu reactions, a comparison was conducted of pre-isolated 5 and
a solution of 5 + LiCN to gauge their relative abilities to ortho deprotonate N,N -
diisopropylbenzamide. Trapping the product with I2, preisolated 5 only afforded
2-iodo-N,N -diisopropylbenzamide, 6 in a yield of 37 %, while a solution from which
5 had deposited at –30 ◦C (and which therefore logically contained LiCN) prior to
its re-dissolution at ambient temperature afforded 6 in a yield of 89 %.139
Extensive theoretical studies have also been carried out on lithium cuprate sys-
tems and the possible mechanism of DoCu. Firstly, the interconversion between
Gilman- and Lipshutz-type states of a bis(amido)cuprate and an alkylamidocuprate
was probed (Scheme 1.14). In the case of a representative alkylamido Lipshutz-
type cuprate MeCu(NMe2)(CN)Li2 · 2OMe2, in the presence of an excess of donor
solvent Me2O, a ∆E value of –3.5 kcal mol
−1 for the conversion to the Gilman-type
analogue, MeCu(NMe2)Li · 3OMe2, suggested that the two forms exist in equilib-
rium. However, in the case of the representative homoleptic bis(amido) cuprate,
(Me2N)2Cu(CN)Li2 · 2OMe2, a ∆E value of +9.1 kcal mol
−1 for the conversion
49
1. INTRODUCTION
(a) (b)
Figure 1.39: Molecular structures of (a) PhCu(µ-TMP)Li · 3THF, 4 and (b)
MeCu(µ-TMP)Li · TMEDA, 5139
to Me2NCu(NMe2)Li · 3OMe2 shows that the Lipshutz-type form dominates the
equilibrium. This is consistent with the observation of 1 in the solid-state, and
the ability of a mixture of 5 + LiCN to give a high yield in ortho deprotonation
reactions, even though a Gilman-type cuprate crystallised preferentially.
Me2N Cu Me
Li Li
C NS S
+4S
Me2N Cu Me
Li
S
C NLi
S
S
SS
S
+
–3.5 kcal mol-1
Me2N Cu NMe2
Li Li
C NS S
+4S
Me2N Cu NMe2
Li
S
C NLi
S
S
SS
S
+
+9.1 kcal mol-1
Lipshutz-type Gilman-type S = Et2O
Scheme 1.14: Ease of interconversion between Gilman and Lipshutz states for alky-
lamido (top) and bis(amido) (bottom) cuprates, calculations carried out using the
B3LYP method, SVP basis set for Cu and 6-31+G* basis set for other atoms139
Secondly, the mechanism by which DoCu occurs was probed, and in an analogous
manner to the calculations for DoZn (Section 1.4) and DoAl (Section 1.5), two plau-
sible pathways were explored. These were alkyl- or amido-mediated deprotonation
by the representative Lipshutz-type cuprate MeCu(NMe2)(CN)Li2 ·2OMe (Scheme
1.15). After an initial electrostatic interaction between the cuprate and the rep-
resentative substrate, N,N -dimethylbenzamide, deprotonation by the alkyl ligand
50
1. INTRODUCTION
leads to the most thermodynamically stable product (IM4L), although the path-
way IM3L-TS2L-IM4L has a high activation energy (+40.8 kcal mol
−1) relative
to that of IM3L meaning it is kinetically unfavourable. However, deprotonation
by the amido ligand has a much lower activation energy (+27.7 kcal mol−1) and so
represents the most favourable pathway even if the product IM2L does not afford
a significant thermodynamic gain relative to the starting materials. Subsequent
conversion of IM2L to the thermodynamically stable IM4L via an amido-alkyl
ligand exchange process was considered, although the high barrier to activation
(+31.3 kcal mol−1) means the process would be disfavoured if it occurred at all.138
TS1L+
H
Me2N O
RTL
IM1L
IM2L
H
N
Me2
Cu
Me
Li
NC
LiOMe2N
S1.89
1.46
1.30
1.97
2.06
‡
N
Me2
Cu
MeLiN
C
Li
OMe2N
S
1.88
1.99
1.94
2.11
H
H N
Me2
Cu Me
Li
NCLi
OMe2N
S
1.90
2.00
2.09
2.12
+27.7 –3.1–3.9
Deprotonation with
amido ligand
O
LiC
NMe2
H
Me
Cu
N
LiS
NMe2
‡
TS3L
+31.3 –46.0O
NMe2
Cu Me
LiCN
LiS
NMe2
H
OMe2N
Cu
Li
CN
LiS
NMe2
+ CH4
Ligand exchange
IM5L
IM4L
–16.8
OMe2N
H
Me
Cu
NMe2
Li
CNLi
S
TS2L
‡
+40.8
–57.0
H
Me2N O
Li
LiN
C
NMe2
Cu
MeS
Deprotonation with
alkyl ligand
IM3L
+5.4
2.38
1.98 1.96
2.05
2.17
1.32
1.52
1.98
2.68
1.95 1.94
2.04
1.59 1.39
2.06
2.32 1.96
1.09
1.96
2.05
2.19
2.09
OMe2N
Cu
Li
CN
Li
S
NMe2
+ CH4
IM4L
1.95 1.94
2.04
IM2L
ortho-lithiated ortho-cuprated
4.55
H N
Me2
Cu Me
Li
NCLi
OMe2N
S
1.90
2.00
2.09
2.12
4.55
Li
N
MeCuMe2N
C
LiS
S
1.971.96
1.96
1.97
2.18
– S
Scheme 1.15: Modelled pathways for the deprotonation of N,N -dimethylbenzamide
by MeCu(NMe2)(CN)Li2 · 2OMe2: deprotonation by the NMe2 ligand (IM1L-
TS1L-IM2L), deprotonation by the Me ligand (IM3L-TS2L-IM4L), and quench-
ing of HNMe2 by the Me ligand (IM5L-TS3L-IM4L). Bond lengths and energy
changes (∆G) at the B3LYP/6–31+G* level are shown in A˚ and kcal mol−1 re-
spectively, S = OMe2138
Further investigations into plausible mechanistic pathways identified a favourable
route to forming an ortho cuprated aromatic species via a monomeric Gilman-
type complex, in turn formed from the elimination of LiCN · Me2O from the
parent Lipshutz-type complex (Scheme 1.17). As was found in the Lipshutz-
type pathway, the alkyl-mediated deprotonation is thermodynamically favourable,
but the activation barrier to the transition state TS2G is kinetically prohibitive
51
1. INTRODUCTION
(+40.3 kcal mol−1 relative to the pre-reaction complex IM3G). In the amido-
mediated deprotonation pathway, while the overall activation energy from RTL
to TS1G is +26.0 kcal mol
−1, neither of the two steps (RTL-IM1G and IM1G-
TS1G) has a prohibitively high activation energy. For the first step (RTL-IM1G)
the dissociation of the reactants into the reactive species was also taken into con-
sideration (Scheme 1.16) finding that while each step involves a positive change in
∆G (RTL-RSG +9.5 kcal mol
−1, RSG-IM1G +1.6 kcal mol
−1), the overall total
(∆G = +11.1 kcal mol−1) is still small.
H
OMe2N
N
Me2
Cu
Li Me
S
Li
N
MeCuMe2N
C
LiS
S
Me2N Cu
Li
Me
S
RTL
+16.8
[+9.5]
–10.2
[+1.6]
+
O NMe2
RSG
IM1G
–LiCN·S
Scheme 1.16: Modelled pathway for the dissociation of MeCu(NMe2)(CN)Li2 ·
2OMe2 into the monomeric Gilman-type complex MeCu(NMe2)Li · OMe2 and
LiCN·S and subsequent coordination to N,N -dimethylbenzamide. Energy changes
(∆E [∆G]) at the B3LYP/6–31+G* level are shown in kcal mol−1, S = OMe2
Although the analogous Lipshutz-type pathway (Scheme 1.15) has a slightly lower
overall activation energy (+23.8 kcal mol−1 for RTL-TS1L), the second step
(IM1L-TS1L) has an activation barrier of +27.7 kcal mol
−1. While quenching
of the amine for the monomeric Gilman pathway (IM5G-TS3G-IM4G) has an
entropically permissible activation energy of +22.4 kcal mol−1, the overall energy
(+28.4 kcal mol−1) of TS3G relative to RTL means that, once again, it is unlikely
to occur. To complete the study, a DoCu pathway via a Gilman-type dimer was
calculated, which was found to have an activation energy barrier of +28.6 kcal
mol−1 to forming the transition state, which as it means this pathway is unlikely
to occur, agrees well with the experimental results.138
52
1. INTRODUCTION
H
OMe2N
N
Me2
Cu
Li Me
S
+14.9
H
O
Me2N
N
Me2
Cu
Li Me
S
TS1G
‡
–2.9
H
O
Me2N
Me
Cu
Li NMe2
S
H
O
Me2N
Me
Cu
Li
NMe2
S
‡
+40.3 –58.1
TS3G
H
O
Me2N
NMe2
Cu
Li
Me
S
‡
TS2G
O
Me2N
Cu
Li
NMe2
S
–44.4
+ CH4
H
O
Me2N
NMe2
Cu
Li
Me
S
+22.4
+11.1
+1.8
H
OMe2N
NMe2
Cu
Li Me
S
1.10
2.30
1.89
1.86
1.47
1.28 1.96
2.67
3.79
2.66
2.24
2.16 2.01
3.21
1.91
Deprotonation with
amido ligand
1.97
1.97 2.49
2.08
1.90
2.13 1.33
1.52
2.86
1.96
1.93
1.95
2.00
1.88
Ligand exchange
–17.3
1.09
2.83 1.94
1.88
1.94
Deprotonation with
alkyl ligand
1.59
1.39
1.97
2.15
ortho-cupratedortho-lithiated
H
OMe2N
NMe2
Cu
Li Me
S
3.792.16 2.01
3.21
1.91
O
Me2N
Cu
Li
NMe2
S
+ CH4
1.93
1.95
2.00
1.88
– LiCN•S+
H
Me2N O
RTL
2.09
Li
N
MeCuMe2N
C
LiS
S
1.971.96
1.96
1.97
2.18
1
IM5G
IM4G
IM1G
IM3G
IM2G
IM2G IM4G
Scheme 1.17: Modelled pathways for the deprotonation of N,N -dimethylbenzamide
by MeCu(NMe2)Li · 2OMe2: deprotonation by the NMe2 ligand (IM1G-TS1G-
IM2G), deprotonation by a Me ligand (IM3G-TS2G-IM4G), and ligand exchange
with the resultant HNMe2 by the Me ligand (IM5G-TS3G-IM4G). Bond lengths
and energy changes (∆G) at the B3LYP/6–31+G* level are shown in A˚ and kcal
mol−1 respectively, S = OMe2138
53
1. INTRODUCTION
50.0
40.0
30.0
20.0
10.0
0.0
–10.0
–20.0
∆G (kcal mol-1)
RTL
0.0
+11.1
IM1G
–3.9
IM1L
+23.8
TS1L
+26.1
TS1G
+23.2
IM2G
+20.7
IM2L
+5.4
IM3L
+1.8
IM3G
+42.1
TS2G
+46.2
TS2L
–10.8
IM4L
–16.0
IM4G
+3.9
IM5L
+5.9
IM5G
+28.4
TS3G
+35.2
TS3L
–10.8
IM4L
–16.0
IM4G
+14.9
+27.7
+40.8
+40.3
+31.3
+22.4
Deprotonation with
amido ligand
Ligand 
exchange
Deprotonation with
alkyl ligand
Figure 1.40: Summary of deprotonative cupration pathways of N,N -
dimethylbenzamide by MeCu(NMe2)Li · 2OMe2 at the B3LYP/6–31+G* level of
theory. Energy changes are shown in kcal mol−1 and are relative to RTL
138
1.8.2.4 Phosphidocuprates
While much work has focused on amidocuprates, the area of phosphido cupra-
tes is, as yet, relatively unexplored. In the solid-state, very few species have
been characterised, although encouragingly, the structures of polymeric bispho-
sphido cuprate [(tBu2P)2Cu]
– [Li · 2THF]+, 7140 (Figure 1.41) and heteroleptic
Gilman-type cuprate MeCuP(tBu)2Li ·3THF, 8
129 (Figure 1.42) suggest that phos-
phidocuprates will exhibit similar behaviour to alkyl- and amidocuprates. 8 was
formed from a 1:1 mixture of CuPtBu2 (in turn prepared from Me3SiP(
tBu)2 and
CuCl) and MeLi, and the structure displays linear geometry across the copper
centre and a C−Cu bond distance (1.940(6) A˚) very similar to those observed
in [Ph2Cu]
– and [Me2Cu]
– (see Section 1.8.2.1).129 By contrast, when CuPtBu2
was reacted with LiPtBu2 the ion-separated homoleptic bis(phosphido) cuprate 7
formed which polymerised into one-dimensional chains in the solid-state.140 The
P−Cu bond lengths in 7 (2.256 A˚av.) are slightly longer than the one in 8 (2.217(2)
A˚) and the approximate angle between phophido lone pairs of 180◦ allows the
polymeric chains to form.
54
1. INTRODUCTION
Figure 1.41: Molecular structure of [(tBu2P)2Cu]
– [Li · 2THF]+, 7140
Figure 1.42: Molecular structure of MeCu(tBu2P)Li · 3THF, 8
129
Moving to solution studies, in 1982 Bertz and Dabbagh set out to investigate the
relative stabilities of amido- and phosphidocuprates. Their experiments involved
reacting a lithium reagent with various CuI precusors at –50 ◦C and then quench-
ing samples with benzoyl chloride at that temperature and after warming to 0 ◦C
and +25 ◦C. While dicyclohexyl-phosphidocuprate and -amidocuprate displayed
identical stabilities at +25 ◦C (89 % yield of PhCOR), diphenylphosphidocuprate
displayed a far greater stability (95 %) at +25 ◦C compared to the amido analogue
(1 %).141 The increased stability was attributed to backdonation from the Cu into
P d-orbitals, which are not available to N. Synthetic studies have shown that phos-
phidocuprates are useful reagents for 1,4-addition to α,β-unsaturated carbonyls129
and in particular for 3,5,5,-trimethylcyclohex-2-enone, for which BuCu(NCy2)Li
proved to be an ineffective reagent. Employing BuCu(PPh2)Li returned high yields
(≥75 %) of the alkylated product even at high temperatures.129,142
55
1. INTRODUCTION
1.9 Aims
The aims of this thesis can now be set out in context.
As discussed in Section 1.8.2.4 phosphidocuprates are synthetically useful and form
similar solid-state structures to known reactive amidocuprates. Therefore, it has
been postulated that due to their greater thermal stability, phosphidocuprates will
offer some benefits over amidocuprates for DoCu reactions. In addition the presence
of phosphorus provides a 31P NMR handle meaning that further NMR spectroscopic
studies can be carried out, including experiments to probe the equilibrium between
Gilman-type and Lipshutz-type conformations in solution. This has previously not
been possible with amidocuprates, as the amido-carbons do not exhibit significant
signal shifts upon structural changes in 13C NMR experiments. This project aims
to synthesise and characterise bis(phosphido)cuprates incorporating simple phos-
phide ligands, dicyclohexylphosphide and diphenylphosphide, and also investigate
their DoCu ability. If both Gilman- and Lipshutz-type phosphidocuprates can be
successfully characterised, this work also aims to probe the equilibrium between
the two structure types in solution using NMR spectroscopic methods.
With the use of cuprate bases recently expanding to include those formed from
copper chloride reagents,136,137 research into the nature of the active species is
warranted. The work in this thesis aims to crystallise and fully characterise a
Lipshutz-type chloride-based bis(TMP)cuprate, and to further study its efficacy
towards directed ortho deprotonation of representative aromatic species. This work
also aims to investigate a previously unexplored route to bis(TMP)cuprates from
CuBr, in order to discover what effect the nature of the halide has on the DoCu
ability of cuprate species in an attempt to maximise the efficiency of ortho de-
protonation reactions.
As discussed in Section 1.8.2.3, to determine whether a Gilman-type or Lipshutz-
type cuprate will be isolated from solution the solvent conditions must be chosen
carefully. To summarise our knowledge so far, a Lipshutz-type cuprate will form if
a stoichiometric amount of THF is employed in the reaction, or if the reaction is
carried out in THF but then removed in vacuo and replaced by a toluene. However,
a Gilman-type cuprate will form if the reaction is carried out in a hydrocarbon
media, such as hexane or toluene, or if the recrystallisation is carried out in a
mixture of hydrocarbon and THF solvents, both in bulk quantities. To crystallise
a heteroleptic Gilman-type cuprate, bulk donor solvent is required in order to fill
the coordination sphere of the Li centre. This thesis aims to investigate the effects
56
1. INTRODUCTION
of using a weaker donating solvent, diethyl ether (Et2O), in place of THF. Et2O
will solvate the Li centres less effectively, which could mean that a LiX (X = e.g.
I, CN) unit is more likely to be incorporated into the complex resulting in the
formation of Lipshutz-type cuprates, or that LiX will precipitate out of solution
resulting in the formation of Gilman-type cuprates.
In recent years the cost of HTMP has greatly increased. This means, in turn,
that carrying out DoCu reactions is becoming more expensive. This thesis seeks
to discover whether an alternative, cheaper amine will also form reactive Lipshutz-
type cuprates and carry out DoCu efficiently. As bulky amido ligands have proven
to be most effective, HDMP is an ideal candidate as it has already been proven as a
suitable ligand in alkali metal zincates and aluminates and especially as it offers a
massive cost saving when the retail prices of the two amines are considered.∗ This
thesis aims to synthesise Gilman- and Lipshutz-type lithium cuprate bases using
the new ligand and, if successful, to test the species for efficiency in DoCu reactions.
∗ The cheapest available HTMP costs £2.62 g−1, the cheapest available HDMP £0.20 g−1.
Data from http://www.sigmaaldrich.com, accessed 19th March 2014
57
Chapter 2
General Experimental Techniques
2.1 CoSHH Considerations
The experiments described in this thesis involved handling compounds which were
potentially hazardous to health. CoSHH assessments were carried out before each
new experiment and standard PPE (safety spectacles, gloves and laboratory coat)
were worn at all times. Experiments were carried out in a fumehood to reduce
the risk of inhalation or contact (e.g. through the skin) with the hazardous com-
pounds. When a syringe and needle were required to transport dry solvents from
the solvent stills to the fumehood, the needle was capped or sheathed to prevent
injury. Cyanide reagents were stored in a designated poisons cupboard and were
handled entirely in the fumehood, and all contaminated or potentially contami-
nated equipment was washed with a 1:1 solution of bleach (sodium hypochlorite)
and water. A specially trained first aider was notified at all times when cyanide
reagents were being used, and an oxygen cylinder and face mask were stored in
the laboratory in case of emergency.
2.2 Inert Atmosphere Techniques
Several of the starting materials and the majority of the products handled in this
project were both air-sensitive and hygroscopic. For this reason, all experimental
work, syntheses, isolations and analyses described in the following sections were
carried out under a dry, inert atmosphere (Ar or N2). In addition, the phosphine
reagents that were used were also photosensitive, so needed to be stored in sealed,
58
2. GENERAL EXPERIMENTAL TECHNIQUES
covered containers under a dry, inert atmosphere.
The apparatus used were a glove box and a vacuum line. Syntheses were carried
out using a bench-top vacuum line. Standard inert-atmosphere Schlenk techniques
were employed for the syntheses.143,144 The reactions were all carried out in Schlenk
tubes, which were dried overnight at +80 ◦C, and all joints were then lubricated
using silicone grease (Dow-Corning) prior to use. These were then evacuated three
times to less than 0.1 Torr, being refilled with N2 from the house supply each
time. Solid reagents that were not air-sensitive were added directly to a Schlenk
tube before evacuation, which served to dry and remove any oxygen from these
reagents. Air-sensitive solid reagents were stored in the glove box, allowing them
to be accurately weighed and transferred to a Schlenk tube under an inert atmo-
sphere. The Schlenk tubes were then sealed and removed from the glove box for
immediate transfer to the vacuum line ensuring the contents were not exposed to
the atmosphere. Liquid reagents for use in sensitive reactions were added directly
to the nitrogen-filled Schlenk via a dry syringe, the needle and syringe having being
flushed with dry N2 on removal from the packaging.
If filtration was required, either a glass filter stick or plastic syringe filter tip were
employed, depending on the nature of the precipitate. For coarse particulates,
the former method was used. A Schlenk tube was connected to a high porosity
glass filter stick, capped by a glass stopper and an inert atmosphere was created
in the manner described above. The whole apparatus was then inverted, and
after removing the stoppers, quickly inserted into the Schlenk tube containing the
reaction mixture. This whole setup was again inverted and the filtrate collected,
under vacuum if necessary. For finer particulates which were not effectively removed
by filtration through the highest porosity glass filter stick, the plastic syringe filter
tip method was required. The filter tip was selected depending on the amount and
nature of the precipitate and volume of solution being filtered. PTFE filter tips of
size (diameter, pore size) 30 mm, 1.0 µm; 25 mm, 0.45 µm; and 17 mm, 0.45µm and
glass microfibre (GMF) filter tips of size 30 mm, 1.2 µm were purchased from Fisher
Scientific and Gilson Scientific. Before use, the filter tip was attached to a syringe
and needle and then flushed with dry N2 three times. A separate syringe and needle
were then flushed with dry N2 three times and used to collect the reaction mixture
to be filtered. This second syringe was quickly transferred to the filter tip and the
filtrate transferred to a clean, dry Schlenk tube under an inert atmosphere. In both
cases, the reaction was then brought to completion in the second Schlenk tube.
Crystallisations were carried out at either room temperature, +5, –5, or –27 ◦C
59
2. GENERAL EXPERIMENTAL TECHNIQUES
under an inert atmosphere in the Schlenk tube used to carry out the reaction.
Where necessary, isolation of the products was achieved by returning the Schlenk
tube to the double manifold line and removing the mother liquor with a syringe
under a constant positive pressure of dry N2. The Schlenk tube was then evacuated
to ensure the product was devoid of any remaining solution and transferred to the
glove box for storage and preparation for analysis.
The majority of characterised products were handled in the glove box after isola-
tion. Samples and equipment entered the glove box via the port. The port was
evacuated three times, each time refilling with N2. The same procedure was em-
ployed for the gloves, in order to ensure that any air which may have leaked in
was removed. Opening the glove doors allowed access to the glove box, and the
samples and equipment were accessed using the inner port door. The atmosphere
in the glove box was kept pure by constant recirculation by a rotary pump through
four purification columns: one containing a molecular sieve (BDH 3 A˚, 16”) which
removed moisture, and three containing a Cu catalyst (BASF Cu catalyst R11)
which removed oxygen as copper oxide. Any faults, for example leaking gloves
or seals, were readily detected by moisture, oxygen and pressure meters which
monitor the glove box.
2.3 Starting Materials and Solvents
Reagents were obtained from Sigma-Aldrich, were of the highest purity available
and were used as received, except for DMPH and TMPH which were distilled if
necessary to remove trace H2O. Phosphine reagents were stored under an inert
atmosphere at +5 ◦C and other liquid reagents were stored over molecular sieves
(4 A˚) for at least 24 hours prior to use. nButyllithium (nBuLi) was obtained as
a standard solution from Sigma-Aldrich and stored under N2 at +5
◦C in Aldrich
SureSealTM glass bottles. All dry solvents used were obtained from the house sol-
vent stills and freshly distilled prior to use: toluene over sodium, and tetrahydro-
furan (THF), diethyl ether (Et2O) and hexane over a sodium-potassium amalgam.
Solvents used for work-up steps and purification of non air-sensitive products were
purchased from Fisher Scientific and used as received.
60
2. GENERAL EXPERIMENTAL TECHNIQUES
2.4 Melting Point Determination
A melting point tube was loaded with 2–3 mg of the sample in the glove box,
and then sealed with vacuum grease and ParafilmTM prior to removal from the
inert atmosphere. The melting point was then determined using standard Griffin
melting point apparatus.
2.5 Elemental Analyses
Elemental analysis was performed on a Perkin-Elmer 240 Elemental Analyser to
find the carbon, hydrogen, nitrogen and phosphorous percentages by mass of the
characterised products. Into each of two pre-weighed aluminium capsules was
placed 1–2 mg of the product, which were then sealed using a press. This process
was carried out in the glove box to prevent any unwanted oxidation or degrada-
tion of the products. The capsules were removed from the glove box and then
accurately weighed before analysis.
For analysis of halogen elements, 5–10 mg of product was placed into a vial and
sealed in the glove box, from which a 1–2 mg sample was removed and accurately
weighed. In the case of Br, the sample was immediately burned and then hy-
drolysed. In the case of I and Cl, the sample was immediately hydrolysed. The
resulting solution was then titrated against AgNO3.
2.6 Multinuclear Nuclear Magnetic Resonance
(NMR) Spectroscopy
NMR spectroscopy samples of non-air sensitive materials were prepared in the
fumehood using deuterated benzene (C6D6), d8-toluene or d8-THF solvents and
then transferred to a thin-walled NMR tube (Wilmad, 528-PP). NMR spectroscopy
samples for air-sensitive compounds were prepared in the glove box under an inert
atmosphere of N2. A known mass (5–10 mg for
1H NMR spectroscopy, 20–30 mg for
13C, 7Li and 31P NMR spectroscopy) was dissolved in a suitable deuterated solvent
which had been dried over a sodium mirror for at least 24 hours prior to use, and
transferred to an Young’s tap NMR tube. The tubes were sealed with ParafilmTM
before being removed from the glove box to prevent atmospheric degradation. The
61
2. GENERAL EXPERIMENTAL TECHNIQUES
choice of solvent was dependent on the solubility of the material to be analysed,
the expected nature of the spectrum and the temperature at which the spectrum
was recorded. General NMR spectroscopic experiments were carried out between
+22 and +27 ◦C as specified in the experimental data. Variable temperature NMR
spectroscopic studies were conducted between –20 and +60 ◦C with spectra being
recorded at regular intervals as specified.
All 1H, 13C, 7Li and 31P spectra were recorded using a either a Bruker DPX 400 FT-
NMR spectrometer or a Bruker DRX 500 FT-NMR spectrometer. The chemical
shifts were internally referenced against the deuterated solvents and plotted relative
to TMS (CDCl3, 1 %) for
1H and 13C, LiCl (D2O, 9.7 M) for
7Li and H3PO4
(CDCl3, 85 %) for
31P.
2.7 Single Crystal X-ray Diffractometry
Single crystal X-ray diffractometry was the primary method used to determine the
solid state structures of the compounds synthesised. Crystals grown were ideally no
greater in dimensions than 0.5 x 0.5 x 0.5 mm3. The crystals were removed from
the Schlenk tube under a positive pressure of Ar and were immediately coated
in a layer of per-fluorinated polyether oil to ensure minimal degradation due to
atmospheric oxygen or moisture. A microscope with a polarising lens was then
used to examine the oil-coated crystals in order to select a suitable crystal. This
was mounted on the glass fibre attached to the diffractometer goniometer head,
which was then attached to the diffractometer. The oil coating the crystal was then
frozen using a cold, dry stream of N2 gas, which served several purposes: it fixed
the orientation of the crystal, further limited degradation of the crystal through
oxidation, hydrolysis or drying of the sample and limited the lattice vibrations,
resulting in clearer diffraction spots.145 In order to limit lattice vibrations in the
structures reported here the diffraction patterns were collected at between 140 and
180 K. Care was taken to avoid cooling the crystal too quickly, as this can cause
the crystal to crack, precluding collection of a useful dataset.
Data were collected on a Nonius Kappa CCD or an Agilent Gemini A-Ultra diffrac-
tometer equipped with an Oxford Cryostream low-temperature device. Collection
time depended on the lattice type, size of unit cell and the intensity of reflections.
The SHELXL-97 program was employed to solve the structures, first locating the
heavy atoms via direct methods.146 Refinement was carried out on F2 by full matrix
62
2. GENERAL EXPERIMENTAL TECHNIQUES
least squares techniques147 with non-hydrogen atoms being treated anisotropically
and hydrogen atoms being added in calculated positions and allowed to ride on
their parent atom (unless otherwise stated).
2.8 Computational Calculations
The computational calculations were carried out using the RIKEN supercomputer
cluster at RIKEN, Japan by Dr. Shinsuke Komagawa and the group of Prof.
Masanobu Uchiyama.
All calculations were carried out with the Gaussian 09 program package.148 The
molecular structures and harmonic vibrational frequencies were obtained using
the hybrid density functional method based on Becke’s three-parameter exchange
function and the Lee-Yang-Parr nonlocal correlation functional (B3LYP).149–152
Ahlrichs’ SVP153 all-electron basis set was used for Cu atoms and 6–31+G* for all
other atoms (denoted as 631SVP). Geometry optimisation and vibrational analysis
were performed at the same level. All stationary points were optimised without
any symmetry assumptions, and characterised by normal coordinate analysis at the
same level of theory (number of imaginary frequencies (NIMAG), 0 for minima and
1 for transition states). The intrinsic reaction coordinate (IRC) method was used
to track minimum energy paths from transition structures to the corresponding
local minima.154–157
63
Chapter 3
Experimental Procedures and
Results
3.1 Starting Material and Intermediate
Reference Data
Characterisation of HPPh
2
1H NMR spectroscopy (400 MHz, 22 ◦C, C6D6): δ 7.45 (m, 4H; Ph), 7.08 (m,
6H; Ph), 5.27 (d, 1JPH = 216 Hz, 1H; PH)
{1H}31P NMR spectroscopy (162 MHz, 22 ◦C, C6D6): δ –39.9 (s; Ph2PH)
31P NMR spectroscopy (162 MHz, 22 ◦C, C6D6): δ –39.9 (dquint,
1JPH = 216
Hz, 3JPH = 7 Hz; Ph2PH)
Characterisation of HPCy
2
1H NMR spectroscopy (400 MHz, 22 ◦C, C6D6): δ 2.97 (dt,
1JPH = 193 Hz,
3JHH = 6 Hz, 1H; PH), 1.85 – 1.63 (br m, 11H; Cy), 1.23 (br m, 11H; Cy)
{1H}31P NMR spectroscopy (162 MHz, 22 ◦C, C6D6): δ –27.4 (s; Cy2PH)
31P NMR spectroscopy (162 MHz, 22 ◦C, C6D6): δ –27.4 (d,
1JPH = 193
Hz; Cy2PH)
64
3. EXPERIMENTAL PROCEDURES AND RESULTS
Characterisation of HDMP
1H NMR spectroscopy (400 MHz, 25 ◦C, C6D6): δ 2.52 (tq,
3JHH = 3 Hz,
3JHH
= 6 Hz, 2H; HDMP-2,6), 1.72 (dquint, 2JHH = 13 Hz,
3JHH = 3 Hz, 1H; HDMP-4),
1.50 (dq, 2JHH = 12 Hz,
3JHH = 3 Hz, 2H; HDMP-3,5), 1.30 (qt,
2JHH =
3JHH
= 13 Hz, 3JHH = 4 Hz, 1H; HDMP-4), 1.10 (dm,
2JHH = 12 Hz, 2H; HDMP-3,5),
1.01 (d, 3JHH = 6 Hz, 6H; HDMP-Me), 0.83 (br s, 1H; HDMP-NH)
13C NMR spectroscopy (100 MHz, 25 ◦C C6D6): δ 52.6 (HDMP-2,6), 34.5
(HDMP-3,5), 25.5 (HDMP-4), 23.3 (HDMP-Me)
Characterisation of LiDMP
7Li NMR spectroscopy (194 MHz, 25 ◦C, C6D6): δ 3.15 (0.3Li), 2.19 (1Li),
1.32 (0.3Li)
3.2 Phosphidocuprate and Phosphidocopper
Compounds
3.2.1 Attempted synthesis of dilithium cyano bis(diphenyl-
phosphido)cuprate, [(Ph
2
P)
2
Cu(CN)Li
2
· THF]
2
a) nButyllithium (1.25 mL, 1.6 M in hexanes, 2 mmol) was added dropwise
to a solution of diphenylphosphine (0.36 mL, 2 mmol) in toluene (3 mL)
under N2 at –78
◦C. The resulting solution of lithium diphenylphosphide was
allowed to reach room temperature whereupon it was added to a suspension
of copper(I) cyanide (0.089 g, 1 mmol) in toluene (3 mL) under N2 at –
78 ◦C. The resulting orange slurry was allowed to reach room temperature
whereupon it was filtered. The solvent was removed in vacuo and replaced
with THF (10 mL) yielding a bright orange solution from which 9 deposited
as orange blocks after storage at room temperature for 24 hours.
X-ray crystallographic data for 9
C104H124Cu4LiO2P6, M = 1955.89, trigonal, space group P31c, a = b =
15.4214(2), c = 25.1056(3) A˚, V = 5156.0(3)A˚3, Z = 2, ρcalcd = 1.260 g
65
3. EXPERIMENTAL PROCEDURES AND RESULTS
cm−3, Mo-Kα radiation, λ = 0.71070 A˚, µ = 0.959 mm−1, T = 150(2) K.
26194 data (5271 unique, Rint = 0.0576, θ < 30.44
◦) were collected. wR2
= 0.1772, conventional R = 0.0674 on F values of 5027 reflections with F 2
> 2σ(F 2), S = 1.200, 363 parameters. Residual electron density extrema
±0.928 eA˚−3.
The crystallographic data revealed that the experiment had not yielded the
expected product, but instead an adamantane-type cage structure contain-
ing four copper atoms and six diphenylphosphine units, [(Ph2P)6Cu4]
2−[Li ·
4THF]2
+, 9.
Refinement of the main adamantane-type P6Cu4 structure proved facile; how-
ever the THF-solvated lithium cations were less simple. In each of the
Li · 4THF moieties, a three-fold axis passes through the lithium and one
of the THF molecules. The lithium and oxygen centres that lie on the three
fold axis were each modelled with 1/3 occupancy, and the carbons making up
the THF that is completed by the symmetry operations were modelled with
2/3 occupancy. The oxygen and these carbons were modelled isotropically as
the model was not good enough to support anisotropic refinement. In this
THF the alpha hydrogens were placed riding on the alpha carbons with 2/3
occupancy. The beta hydrogens were modelled in two positions correspond-
ing to the two THF molecules utilising each carbon. Each of these hydrogens
were modelled with 1/3 occupancy. In one of the Li · 4THF units geometric
constraints were applied to the symmetry-completed THF molecules. The
three remaining THF molecules were generated from a single whole THF
molecule in the asu.
Unfortunately, the yield from this preparation was too low to allow collec-
tion of any supporting data. Subsequently, an alternative pathway to 9 was
identified which afforded 9 in higher yields following the method detailed
below.
b) nButyllithium (1.25 mL, 1.6 M in hexanes, 2 mmol) was added dropwise to a
solution of diphenylphosphine (0.36 mL, 2 mmol) in THF (3 mL) under N2 at
–78 ◦C. The resulting solution of lithium diphenylphosphine was allowed to
reach room temperature whereupon it was added to a suspension of copper(I)
iodide (0.190 g, 1 mmol) in THF (3 mL) under N2 at –78
◦C. The resulting
orange slurry was then allowed to reach room temperature whereupon THF (5
mL) was added and the mixture gently warmed. The resulting bright orange
66
3. EXPERIMENTAL PROCEDURES AND RESULTS
solution deposited 9 as orange blocks after storage at room temperature for
24 hours. A crystallographic cell check verified the crystal data obtained by
route (a).
Yield: 0.239 g, 49 % with respect to CuI
Melting Point: 122–124 ◦C
Elemental Analysis: C72H60Cu4LiO2P6 requires C 63.86 %, H 6.39 %, P
9.50 %, found C 62.81 %, H 6.61 %, P 9.06 %
1H NMR spectroscopy (500 MHz, 25 ◦C, C6D6): δ 7.78 (br m, 28H; Ph),
7.05 (br m, 42H; Ph), 5.72 (d, trace; 1JPH = 290 Hz; PH), 3.51 (m, 32H;
THF), 1.37 (m, 32H; THF)
13C NMR spectroscopy (100 MHz, 27 ◦C, C6D6): δ 144.5 (d,
1JCP = 231
Hz; ipso-Ph), 135.2, 134.4, 124.3 (Ph-CH), 68.2 (THF), 25.6 (THF)
{1H}31P NMR spectroscopy (202 MHz, 25 ◦C, C6D6): δ –22.1 (br s;
Ph2PCu), –32.6 (s; Ph2PH)
31P NMR spectroscopy (202 MHz, 25 ◦C, C6D6): δ –22.1 (br s; Ph2PCu),
–32.6 (d, 1JPH = 290 Hz; Ph2PH)
7Li NMR spectroscopy (194 MHz, 25 ◦C, C6D6): δ –0.13 (br s)
3.2.2 Attempted synthesis of lithium bis(diphenylphos-
phido)cuprate, [(Ph
2
P)
2
CuLi]
2
nButyllithium (1.25 mL, 1.6 M in hexanes, 2 mmol) was added dropwise to a
solution of diphenylphosphine (0.36 mL, 2 mmol) in THF (3 mL) under N2 at –
78 ◦C. The resulting solution of lithium diphenylphosphine was allowed to reach
room temperature whereupon it was added to a suspension of copper(I) cyanide
(0.089 g, 1 mmol) in THF (2 mL) under N2 at –78
◦C. The resulting mixture was
allowed to reach room temperature, yielding a dark orange slurry. Gentle heating
afforded a bright orange solution from which [(Ph2P)Cu(CN)Li · 2THF]∞, 10∞,
was deposited as colourless blocks after storage at room temperature for 12 hours.
Yield: 0.229 g, 54 % with respect to CuCN
Melting Point: decomposed from ca. 210 ◦C
Elemental Analysis: C42H52Cu2Li2N2O4P2 requires C 59.22 %, H 6.15 %, N 3.29
%, P 7.27 %, found C 58.93 %, H 6.03 %, N 3.25 %, P 7.59 %
1H NMR spectroscopy (500 MHz, 27 ◦C, d8-THF): δ 7.79 – 6.57 (m, 5H; Ph),
3.49 (m, 4H; THF), 1.70 (m, 4H; THF)
67
3. EXPERIMENTAL PROCEDURES AND RESULTS
13C NMR spectroscopy (100 MHz, 27 ◦C, d8-THF): δ 145.0 – 124.8 (Ph-CH +
CN), 68.2 (THF), 26.4 (THF)
7Li NMR spectroscopy (194 MHz, 27 ◦C, d8-THF): δ –2.72 (br s)
1H NMR spectroscopy (500 MHz, 25 ◦C, C6D6): δ 7.77 – 6.77 (m, 5H; Ph),
5.43 (d, 1JPH = 239 Hz, trace; PH), 3.68 (m, 4H; THF), 1.53 (m, 4H; THF)
13C NMR spectroscopy (100 MHz, 25 ◦C, C6D6): δ 144.5 (br d,
1JCP = 231 Hz;
ipso-Ph), 135.2, 134.1, 124.1 (Ph-CH + CN), 68.2 (THF), 26.4 (THF)
{1H}31P NMR spectroscopy (202 MHz, 25 ◦C, C6D6): δ –14.9, –21.6, –28.1 (s;
Ph2PCu), –38.3 (s; Ph2PH)
31P NMR spectroscopy (202 MHz, 25 ◦C, C6D6): δ –14.9, –21.6, –28.1 (s;
Ph2PCu), –38.3 (d,
1JPH = 233 Hz; Ph2PH)
7Li NMR spectroscopy (194 MHz, 25 ◦C, C6D6): δ 0.89 (br s, 6Li), –0.92 (s, 1Li)
{1H}31P NMR spectroscopy (202 MHz, 40 ◦C, d8-toluene):
δ –13.3 (s, ∆1/2 = 6 Hz; Ph2PCu), –38.4 (s, ∆1/2 = 72 Hz; Ph2PH)
{1H}31P NMR spectroscopy (202 MHz, 27 ◦C, d8-toluene):
δ –13.6 (s, ∆1/2 = 4 Hz; Ph2PCu), –38.4 (s, ∆1/2 = 52 Hz; Ph2PH)
{1H}31P NMR spectroscopy (202 MHz, 0 ◦C, d8-toluene):
δ –14.1 (s, ∆1/2 = 7 Hz; Ph2PCu), –38.3 (s, ∆1/2 = 116 Hz; Ph2PH)
{1H}31P NMR spectroscopy (202 MHz, –20 ◦C, d8-toluene):
δ –14.3 (s, ∆1/2 = 4 Hz; Ph2PCu), –38.5 (s, ∆1/2 = 223 Hz; Ph2PH)
X-ray crystallographic data for 10
C42H52Cu2Li2N2O4P2, M = 851.76, trigonal, space group P3, a = b = 18.3517(3),
c = 14.0350(4) A˚, V = 4093.50(15) A˚3, Z = 3, ρcalcd = 1.037 g cm
−3, Mo-Kα
radiation, λ = 0.71073 A˚, µ = 0.870 mm−1, T = 180(2) K. 17083 data (8002
unique, Rint = 0.0318, θ < 26.34
◦) were collected. wR2 = 0.1894, conventional
R = 0.0690 on F values of 5957 reflections with F 2 > 2σ(F 2), S = 1.065, 463
parameters. Residual electron density extrema ±0.805 eA˚−3.
The crystallographic data revealed that the experiment had not yielded the ex-
pected product, but instead a polymeric species, [Ph2P)Cu(CN)Li ·2THF]∞, 10∞.
While refinement of the structure was, on the whole, straightforward, two of the
THF molecules within the structure of 10 were found to show significant crystal-
lographic disorder. These THF molecules were refined as conformers each with 1/2
occupancy (entirely in one case and for only the carbon atoms in the second case)
68
3. EXPERIMENTAL PROCEDURES AND RESULTS
and geometric constraints were applied. Due to these constraints isotropic dis-
placement parameters were utilised, which were constrained to be equal between
the two conformers.
3.2.3 Synthesis of lithium bis(dicyclohexylphosphido)-
cuprate, (Cy
2
P)
2
CuLi · 2THF, 11
nButyllithium (1.25 mL, 1.6 M in hexanes, 2 mmol) was added dropwise to a
solution of dicylohexyphosphine (0.4 mL, 2 mmol) in THF (3 mL) under N2 at –78
◦C. The resulting solution of lithium dicyclohexyl phosphine was allowed to reach
room temperature whereupon it was added to a suspension of copper(I) cyanide
(0.089 g, 1 mmol) in THF (2 mL) under N2 at –78
◦C. The resulting mixture was
allowed to reach room temperature whereupon it was allowed to stir for 10 minutes.
The solvent was removed in vacuo and replaced with toluene (6 mL) and THF (6
mL) yielding a pale yellow solution from which 11 deposited as colourless blocks
after storage at –27 ◦C for 24 hours.
Yield: 0.088 g, 15 % with respect to CuCN
Melting Point: 151–153 ◦C
Elemental Analysis: C64H120Cu2Li2O4P4 requires C 63.08 %, H 9.93 %, P 10.17
%, found C 62.89 %, H 9.85 %, P 9.92 %
1H NMR spectroscopy (500 MHz, 27 ◦C, C6D6): δ 3.63 (m, 8H; THF), 3.09
(dt, 1JPH = 200 Hz,
3JHH = 6 Hz, trace; Cy2PH), 2.21 (br m, 4H; Cy-CH2), 1.94,
1.78 (br m, 10H; Cy-CH + Cy-CH2), 1.47 (m, 8H; THF), 1.24 – 1.06 (br m, 8H;
Cy-CH2)
13C NMR spectroscopy (125 MHz, 27 ◦C, C6D6): δ 67.8 (THF), 35.2 (d,
1JCP
= 255 Hz; Cy-CH), 26.7, 26.4 (Cy-CH2), 25.8 (THF), 25.3 (Cy-CH2)
{1H}31P NMR spectroscopy (162 MHz, 22 ◦C, C6D6): δ –14.2, –16.4, –19.7 (s;
Cy2PCu), –26.6 (s; Cy2PH)
31P NMR spectroscopy (162 MHz, 22 ◦C, C6D6): δ –14.2, –16.4, –19.7 (s;
Cy2PCu), –26.6 (d,
1JPH = 200 Hz; Cy2PH)
7Li NMR spectroscopy (194 MHz, 25 ◦C, C6D6): δ 2.34 (br s)
{1H}31P NMR spectroscopy (202 MHz, 60 ◦C, d8-toluene):
δ –12.1 (s, ∆1/2 = 408 Hz; Cy2PCu), –26.3 (s, ∆1/2 = 13 Hz; Cy2PH)
{1H}31P NMR spectroscopy (202 MHz, 40 ◦C, d8-toluene):
δ –13.5 (s, ∆1/2 = 679 Hz; Cy2PCu), –26.4 (s, ∆1/2 = 24 Hz; Cy2PH)
69
3. EXPERIMENTAL PROCEDURES AND RESULTS
{1H}31P NMR spectroscopy (202 MHz, 27 ◦C, d8-toluene):
δ –14.0 (s, ∆1/2 = 520 Hz; Cy2PCu), –26.5 (s, ∆1/2 = 28 Hz; Cy2PH)
{1H}31P NMR spectroscopy (202 MHz, 0 ◦C, d8-toluene):
δ –15.6 (s, ∆1/2 = 381 Hz; Cy2PCu), –26.3 (s, ∆1/2 = 127 Hz; Cy2PH)
{1H}31P NMR spectroscopy (202 MHz, –20 ◦C, d8-toluene):
δ –15.8 (s, ∆1/2 = 537 Hz; Cy2PCu), –25.3 (s, ∆1/2 = 530 Hz; Cy2PH)
X-ray crystallographic data for 11
C64H120Cu2Li2O4P4, M = 1218.44, monoclinic, space group P 1¯, a = 11.03790(10),
b = 16.0222(2), c = 20.8778(3) A˚, α = 77.1870(10), β = 82.6640(10), γ =
75.7050(10) ◦, V = 3478.45(7) A˚3, Z = 2, ρcalcd = 1.163 g cm
−3, Mo-Kα radi-
ation, λ = 0.71073 A˚, µ = 0.744 mm−1, T = 180(2) K. 53328 data (16555 unique,
Rint = 0.0627, θ < 27.87
◦) were collected. wR2 = 0.0957, conventional R = 0.0526
on F values of 10332 reflections with F 2 > 2σ(F 2), S = 1.009, 685 parameters.
Residual electron density extrema ±0.438 eA˚−3.
3.3 TMP-based Bis(amido)cuprate Compounds
3.3.1 Synthesis of lithium di(2,2,6,6-tetramethylpiperi-
dido)cuprate, [(TMP)
2
CuLi]
2
, 2
nButyllithium (1.25 mL, 1.6 M in hexanes, 2 mmol) was added dropwise to a so-
lution of 2,2,6,6-tetramethylpiperidine (0.34 mL, 2 mmol) in toluene (3 mL) and
THF (0.08 mL, 1 mmol) under N2 at –78
◦C. The resulting solution of lithium
2,2,6,6-tetramethylpiperidide was allowed to reach room temperature. It was then
added to a suspension of copper(I) chloride (99 mg, 1 mmol) in toluene (3 mL) un-
der N2 at –78
◦C. The mixture was allowed to reach room temperature whereupon
it was filtered. The solvent was reduced in volume to approximately 2 mL giving
a pale orange solution from which 2 deposited as colourless needles after storage
at +5 ◦C for 24 hours. This preparation was scaled up as required.
Melting point: 178–179 ◦C
1H NMR spectroscopy (500 MHz, 27 ◦C, C6D6): δ 1.86 (s, 6H; TMP-Me), 1.70
(s, 6H; TMP-Me), 1.50 (br s, 2H; TMP-3,5), 1.35 (br m, 2H, TMP-3,5), 1.17 (s,
70
3. EXPERIMENTAL PROCEDURES AND RESULTS
2H; TMP-4).
7Li NMR spectroscopy (194 MHz, 27 ◦C, C6D6): δ 1.28 (s)
3.3.2 Synthesis of dilithium chloro[di(2,2,6,6-tetramethyl-
piperidido)]cuprate, [(TMP)
2
Cu(Cl)Li
2
·THF]
2
, 12
nButyllithium (2.5 mL, 1.6 M in hexanes, 4 mmol) was added dropwise to a so-
lution of 2,2,6,6-tetramethylpiperidine (0.68 mL, 4 mmol) in toluene (4 mL) and
THF (0.32 mL, 4 mmol) under N2 at –78
◦C. The resulting solution of lithium
2,2,6,6-tetramethylpiperidide was allowed to reach room temperature. This was
then added to a suspension of copper(I) chloride (198 mg, 2 mmol) in toluene (3
mL) under N2 at –78
◦C. The mixture was allowed to reach room temperature
whereupon it was filtered. The solvent was reduced in volume to approximately
4 mL giving a pale orange solution from which 12 deposited as colourless blocks
after storage at –27 ◦C for 12 hours.
Yield: 0.195 g, 21 % with respect to CuCl
Melting Point: 173–175 ◦C
Elemental Analysis: C44H88Cu2Li4N4O2Cl2 requires C 56.77 %, H 9.53 %, N
6.02 %, Cl 7.62 %, found C 56.34 %, H 9.69 %, N 6.03 %, Cl 7.92 %
1H NMR spectroscopy (500 MHz, 25 ◦C, C6D6): δ 3.67 (br m, 4H; THF), 2.05
(br m, 2H; TMP-3,5), 1.81 (s, 6H; TMP-Me), 1.68 – 1.63 (m, 18H; TMP-Me), 1.46
(br m, 6H; TMP-3,5), 1.39 (br m, 4H; THF), 1.13 (m, 4H; TMP-4)
13C NMR spectroscopy (125 MHz, 25 ◦C, C6D6, G = TMP2CuLi,
138 L = 12): δ
68.7 (THF), 54.6 (TMP-2, G), 54.0 (TMP-2, L), 52.1 (TMP-2, L), 42.6 (TMP-3,5,
G), 42.4 (TMP-3,5, L), 41.1 (TMP-3,5, L), 40.6 (TMP-Me, G), 39.0 (TMP-Me,
L), 38.6 (TMP-Me, L), 35.0 (TMP-Me, G), 34.6 (TMP-Me, L), 32.0 (TMP-Me,
L), 25.3 (THF), 20.6 (TMP-4, L), 19.8 (TMP-4, L), 19.7 (TMP-4, G)
7Li NMR spectroscopy (194 MHz, 25 ◦C, C6D6, G = TMP2CuLi,
138 L = 12):
δ 1.92 (s, 6Li; L), 1.27 (s, 1Li; G)
1H NMR spectroscopy (500 MHz, 25 ◦C, d8-THF): δ 3.65 (br m, 4H; THF),
1.81 (br m, 4H; THF), 1.68 (br m, 4H; TMP-4), 1.38, 1.09 (s, 24H, TMP-Me), 1.33
(br m, 8H, TMP-3,5)
1H NMR spectroscopy (500 MHz, 0 ◦C, d8-THF): δ 3.65 (br m, 4H; THF),
1.81 (br m, 4H; THF), 1.67 (br m, 4H; TMP-4), 1.34, 1.08 (s, 24H, TMP-Me),
1.31 (br m, 8H, TMP-3,5)
71
3. EXPERIMENTAL PROCEDURES AND RESULTS
13C NMR spectroscopy (125 MHz, 25 ◦C, d8-THF, G = TMP2CuLi,
138 L =
12): δ 68.2 (THF), 54.2 (TMP-2, L), 50.0 (TMP-2, G), 41.9 (TMP-3,5, L), 38.6
(TMP-3,5, G), 37.3 (TMP-Me, L), 32.2, 32.1 (TMP-Me, G), 26.4 (THF), 20.5
(TMP-4, L), 19.3 (TMP-4, G)
13C NMR spectroscopy (125 MHz, 0 ◦C, d8-THF, G = TMP2CuLi,
138 L =
12): δ 65.1 (THF), 51.3 (TMP-2, L), 47.1 (TMP-2, G), 39.1 (TMP-3,5, L), 35.7
(TMP-3,5, G), 34.8 (TMP-Me, L), 29.4, 29.3 (TMP-Me, G), 23.0 (THF), 18.0
(TMP-4, L), 17.5 (TMP-4, G)
7Li NMR spectroscopy (194 MHz, 25 ◦C, d8-THF): δ 0.70 (s)
7Li NMR spectroscopy (194 MHz, 0 ◦C, d8-THF): δ 0.22 (s)
X-ray crystallographic data for 12
C44H88Cu2Li4N4O2Cl2, M = 930.92, monoclinic, space group P21/c, a =
14.3830(3), b = 8.1954(2), c = 22.1569(5) A˚, β = 97.2250(10) ◦, V = 2590.99(10)
A˚3, Z = 2, ρcalcd = 1.193 g cm
−3, Mo-Kα radiation, λ = 0.71070 A˚, µ = 0.959
mm−1, T = 180(2) K. 16499 data (6110 unique, Rint = 0.0763, θ < 27.88
◦) were
collected. wR2 = 0.1450, conventional R = 0.0569 on F values of 4478 reflec-
tions with F 2 > 2σ(F 2), S = 1.091, 269 parameters. Residual electron density
extrema ±1.259 eA˚−3.
3.3.3 Synthesis of dilithium bromo[di(2,2,6,6-tetramethyl-
piperidido)]cuprate, [(TMP)
2
Cu(Br)Li
2
·THF]
2
, 13
nButyllithium (2.5 mL, 1.6 M in hexanes, 4 mmol) was added dropwise to a solution
of 2,2,6,6-tetramethylpiperidine (0.68 mL, 4 mmol) in toluene (4 mL) and THF
(0.32 mL, 2 mmol) under N2 at –78
◦C. The resulting solution of lithium 2,2,6,6-
tetramethylpiperidide was allowed to reach room temperature. It was then added
to a suspension of copper(I) bromide (246 mg, 2 mmol) in toluene (3 mL) under
N2 at –78
◦C. The mixture was allowed to reach room temperature whereupon it
was filtered. The solvent was reduced in volume to approximately 4 mL giving
a pale orange solution from which 13 deposited as colourless blocks after storage
at –27 ◦C for 12 hours.
Yield: 0.353 g, 35 % with respect to CuBr
Melting Point: decomposed from ca. 174 ◦C
72
3. EXPERIMENTAL PROCEDURES AND RESULTS
Elemental Analysis: C44H88Cu2Li4N4O2Br2 requires C 51.80 %, H 8.70 %, N
5.49 %, Br 15.67 %, found C 52.45 %, H 8.72 %, N 5.71 %, Br 14.97 %
1H NMR spectroscopy (500 MHz, 25 ◦C, C6D6): δ 3.61 (br s, 4H; THF), 2.00
(br m, 2H; TMP-3,5), 1.74 (br s, 12H; TMP-Me), 1.61 (br m, 12H; TMP-Me), 1.37
(br m, 4H; THF), 1.34 (br m, 6H; TMP-3,5), 1.07 (m, 4H; TMP-4)
13C NMR spectroscopy (125 MHz, 25 ◦C, C6D6, G = TMP2CuLi,
138 L = 13):
δ 68.5 (THF), 54.5 (TMP-2, G), 54.0 (TMP-2, L), 49.5 (TMP-2, L), 42.4 (TMP-
3,5, G), 40.9 (TMP-3,5, L), 40.4 (TMP-3,5 L), 38.9 (TMP-Me, L), 38.4 (TMP-Me,
L), 34.8 (TMP-Me, G), 34.4 (TMP-Me, L), 31.9 (TMP-Me, L), 25.2 (THF), 19.7
(TMP-4, L), 19.5 (TMP-4, L), 18.6 (TMP-4, G)
7Li NMR spectroscopy (194 MHz, 25 ◦C, C6D6): δ 1.34 (br s)
X-ray crystallographic data for 13
C44H88Cu2Li4N4O2Br2, M = 1019.84, monoclinic, space group P21/c, a =
14.5737(5), b = 8.1912(3), c = 22.0868(9) A˚, β = 96.244(2) ◦, V = 2617.79(17)
A˚3, Z = 2, ρcalcd = 1.294 g cm
−3, Mo-Kα radiation, λ = 0.71073 A˚, µ = 2.374
mm−1, T = 180(2) K. 15970 data (7212 unique, Rint = 0.0627, θ < 30.03
◦) were
collected. wR2 = 0.1061, conventional R = 0.0391 on F values of 4802 reflec-
tions with F 2 > 2σ(F 2), S = 0.951, 270 parameters. Residual electron density
extrema ±1.172 eA˚−3.
3.3.4 Synthesis of dilithium cyano[di(2,2,6,6-tetramethyl-
piperidido)]cuprate, [(TMP)
2
Cu(CN)Li
2
·OEt
2
]
2
, 14
nButyllithium (1.25 mL, 1.6 M in hexanes, 2 mmol) was added dropwise to a
solution of 2,2,6,6-tetramethylpiperidine (0.34 mL, 2 mmol) in toluene (2 mL)
under N2 at –78
◦C. The resulting solution of lithium 2,2,6,6-tetramethylpiperidide
was allowed to reach room temperature. It was then added to a suspension of
copper(I) cyanide (90 mg, 1 mmol) in toluene (2 mL) under N2 at –78
◦C. After
allowing the resulting slurry to warm to room temperature, the solvent was removed
and diethyl ether (8 mL) was added. The resulting solution was reduced to ca. 3
mL volume and 15 mL hexane was added, whereupon the mixture was filtered.
The resulting yellow-white solution deposited 14 as colourless blocks after storage
at –27 ◦C for 24 hours.
73
3. EXPERIMENTAL PROCEDURES AND RESULTS
Yield: 0.190 g, 41 % with respect to CuCN
Melting Point: decomposed from ca. 191 ◦C
Elemental Analysis: C46H92Cu2Li4N6O2 requires C 60.31 %, H 10.12 %, N
9.17 %, found C 60.01 %, H 9.97 %, N 9.22 %
1H NMR spectroscopy (500 MHz, 25 ◦C, d8-THF): δ 3.42 (q,
3JHH = 7 Hz,
4H; Et2O), 1.65 (br m, 4H; TMP-3,5), 1.33 (br m, 4H; TMP-4), 1.22 (br m, 4H;
TMP-3,5) 1.15 (t, 3JHH = 7 Hz, 6H; Et2O), 1.12 (br m, 18H, TMP-Me), 1.09 (m,
6H, TMP-Me)
13C NMR spectroscopy (100 MHz, 25 ◦C, d8-THF): δ 66.3 (Et2O), 53.6 (TMP-
2), 39.2 (TMP-3,5), 37.8 (TMP-Me), 32.2 (TMP-Me), 19.3 (TMP-4), 15.7 (Et2O)
7Li NMR spectroscopy (194 MHz, 25 ◦C, d8-THF): δ –0.73 (br s, 1Li), –2.28
(br s, 2Li)
X-ray crystallographic data for 14
C46H92Cu2Li4N6O2, M = 916.10, monoclinic, space group P21/c, a = 8.1679(2), b
= 26.2989(6), c = 13.2527 A˚, β = 106.5320(10) ◦, V = 2729.09(12) A˚3, Z = 2, ρcalcd
= 1.115 g cm−3, Mo-Kα radiation, λ = 0.71073 A˚, µ = 0.816 mm−1, T = 180(2)
K. 19055 data (6349 unique, Rint = 0.1031, θ < 27.88
◦) were collected. wR2 =
0.3413, conventional R = 0.1241 on F values of 2997 reflections with F 2 > 2σ(F 2),
S = 1.055, 281 parameters. Residual electron density extrema ±2.526 eA˚−3.
It should be noted that the quality of the crystallographic data collected on 14
was poor despite repeated attempts. However, the connectivity was unambiguous
and the structure was analogous to [(TMP)2Cu(CN)Li2 · THF]2, 1, which was
synthesised employing THF as the donor solvent.15
3.3.5 Synthesis of dilithium iodo[di(2,2,6,6-tetramethyl-
piperidido)]cuprate, [(TMP)
2
Cu(I)Li
2
·OEt
2
]
2
, 15
a) nButyllithium (2.5 mL, 1.6 M in hexanes, 4 mmol) was added dropwise to a
solution of 2,2,6,6-tetramethylpiperidine (0.68 ml, 4 mmol) in toluene (2 mL)
under N2 at –78
◦C. The resulting solution of lithium 2,2,6,6-tetramethyl-
piperidide was allowed to reach room temperature. It was then added to a
suspension of copper(I) iodide (380 mg, 2 mmol) in toluene (3 mL) under N2
at –78 ◦C and the resulting mixture was allowed to reach room temperature.
74
3. EXPERIMENTAL PROCEDURES AND RESULTS
The solvent was removed and diethyl ether (8 mL) was added, whereupon
the mixture was filtered. The resulting solution was reduced to ca. 3 mL
volume and 12 mL hexane was added. The resulting solution deposited 15
as yellow blocks after storage at +5 ◦C for 24 hours.
Yield: 0.303 g, 27 % with respect to CuI
Melting Point: decomposed from ca. 185 ◦C
Elemental Analysis: C44H92I2Cu2Li4N4O2 requires C 47.27 %, H 8.30 %,
N 5.01 %, I 22.70 %, found C 46.53 %, H 8.23 %, N 5.11 %, I 22.80 %
1H NMR spectroscopy (500 MHz, 25 ◦C, C6D6): δ 3.45 (q,
3JHH = 7
Hz, 4H; Et2O), 1.90 (br m, 2H; TMP-3,5), 1.84 – 1.51 (m, 24H; TMP-Me),
1.31 (m, 4H; TMP-4), 1.23 (br m, 6H; TMP-3,5), 1.23 (t, 3JHH = 7 Hz, 6H;
Et2O)
13C NMR spectroscopy (100 MHz, 25 ◦C, C6D6): δ 65.8 (Et2O), 54.3
(TMP-2,6), 41.0 (TMP-3,5), 39.2 (TMP-Me), 34.3 (TMP-Me), 19.7 (TMP-
4), 15.1 (Et2O)
7Li NMR spectroscopy (194 MHz, 25 ◦C, C6D6): δ 0.95 (s, 1Li), 0.09 (s,
1Li)
X-ray crystallographic data for 15
C44H92I2Cu2Li4N4O2,M = 1117.86, triclinic, space group P 1¯, a = 12.8770(1),
b = 15.0455(2), c = 16.1130(2) A˚, α = 87.0190(10), β = 70.4260(10), γ =
70.2500(10) ◦, V = 2761.72(5) A˚3, Z = 2, ρcalcd = 1.344 g cm
−3, Mo-Kα
radiation, λ = 0.71073 A˚, µ = 1.922 mm−1, T = 180(2) K. 38687 data
(15904 unique, Rint = 0.0412, θ < 30.03
◦) were collected. wR2 = 0.0859,
conventional R = 0.0300 on F values of 12468 reflections with F 2 > 2σ(F 2),
S = 1.036, 561 parameters. Residual electron density extrema ±0.651 eA˚−3.
The majority of the structure was straightforward to refine, although the cen-
tral iodine atoms were split over two sites, each modelled with 1/2 occupancy.
b) nButyllithium (2.5 mL, 1.6 M in hexanes, 4 mmol) was added dropwise to a
solution of 2,2,6,6-tetramethylpiperidine (0.68 ml, 4 mmol) in toluene (3 mL)
and diethyl ether (0.42 mL) under N2 at –78
◦C. The resulting solution of
lithium 2,2,6,6-tetramethylpiperidide was allowed to reach room temperature.
It was then added to a suspension of copper(I) iodide (380 mg, 2 mmol) in
toluene (3 mL) under N2 at –78
◦C and the resulting mixture was allowed to
75
3. EXPERIMENTAL PROCEDURES AND RESULTS
reach room temperature. The mixture was filtered and the resulting yellow
solution deposited 15 as colourless blocks after storage at –27 ◦C for 36 hours.
Yield: 188 mg (17 %) with respect to CuI
Melting Point: decomposed from ca. 185 ◦C
A crystallographic cell check verified the crystal data obtained by route (a).
3.3.6 Synthesis of dilithium chloro[di(2,2,6,6-tetramethyl-
piperidido)]cuprate, [(TMP)
2
Cu(Cl)Li
2
·OEt
2
]
2
, 16
A room temperature solution of lithium 2,2,6,6-tetramethylpiperidide was prepared
as for 14 and was added to a suspension of copper(I) chloride (99 mg, 1 mmol) in
toluene (2 mL) under N2 at –78
◦C. The mixture was allowed to reach room temper-
ature. The solvent was removed and diethyl ether (8 mL) was added, whereupon
the mixture was filtered. The resulting solution was reduced to ca. 3 mL volume
and 10 mL hexane was added. The resulting pale yellow solution deposited 16 as
colourless needles after storage at –27 ◦C for 72 hours.
Yield: 0.060 g, 13 % with respect to CuCl
Melting Point: decomposed from ca. 190 ◦C
Elemental Analysis: C44H92Cl2Cu2Li4N4O2 requires C 56.52 %, H 9.92 %, N
5.99 %, Cl 7.48 % found C 55.71 %, H 9.68 %, N 6.95 %, Cl 7.96 %
1H NMR spectroscopy (500 MHz, 25 ◦C, C6D6): δ 3.46 (q,
3JHH = 7 Hz, 4H;
Et2O), 2.08 (br m, 2H; TMP-3,5), 1.82 (br m, 12H; TMP-Me), 1.71 – 1.67 (br m,
12H; TMP-Me), 1.37 (br m, 4H; TMP-4), 1.19 (br m, 6H; TMP-3,5), 1.18 (t, 6H,
3JHH = 7 Hz; Et2O)
13C NMR spectroscopy (100 MHz, 25 ◦C, C6D6): δ 65.8 (Et2O), 54.6 (TMP-
2,6), 42.5 (TMP-3,5), 40.6 (TMP-Me), 35.0 (TMP-Me), 19.6 (TMP-4), 15.5 (Et2O)
7Li NMR spectroscopy (194 MHz, 25 ◦C, C6D6): δ 1.89 (s, 6Li), 1.83 (s, 1Li)
X-ray crystallographic data for 16
C44H92Cl2Cu2Li4N4O2,M = 934.96, monoclinic, space group P21/n, a = 8.1295(2),
b = 21.4601(4), c = 15.3005(3) A˚, β = 92.3120(10) ◦, V = 2667.15(10) A˚3, Z =
2, ρcalcd = 1.164 g cm
−3, Mo-Kα radiation, λ = 0.71070 A˚, µ = 0.932 mm−1,
T = 180(2) K. 30021 data (6088 unique, Rint = 0.0892, θ < 27.49
◦) were col-
lected. wR2 = 0.1365, conventional R = 0.0582 on F values of 4397 reflections
76
3. EXPERIMENTAL PROCEDURES AND RESULTS
with F 2 > 2σ(F 2), S = 1.062, 272 parameters. Residual electron density ex-
trema ±0.908 eA˚−3.
3.3.7 Synthesis of dilithium bromo[di(2,2,6,6-tetramethyl-
piperidido)]cuprate, [(TMP)
2
Cu(Br)Li
2
·OEt
2
]
2
, 17
a) A room temperature solution of lithium 2,2,6,6-tetramethylpiperidide was
prepared as for 14 and was added to a suspension of copper(I) bromide (143
mg, 1 mmol) in toluene (1 mL) under N2 at –78
◦C. The resulting mixture was
allowed to reach room temperature. The solvent was removed and diethyl
ether (8 mL) was added, whereupon the mixture was filtered. The resulting
solution was reduced to ca. 3 mL volume and 12 mL hexane was added,
whereupon the mixture was filtered. The resulting yellow solution deposited
17 as colourless blocks after storage at –27 ◦C for 48 hours.
Yield: 0.091 g, 18 % with respect to CuBr
Melting Point: decomposed from ca. 193 ◦C
Elemental Analysis: C44H92Br2Cu2Li4N4O2 requires C 51.61 %, H 9.06 %,
N 5.47 %, Br 15.61 %, found C 50.77 %, H 8.74 %, N 5.47, Br 15.49 %
1H NMR spectroscopy (500 MHz, 25 ◦C, d8-THF): δ 3.42 (q, 4H,
3JHH =
7 Hz; Et2O), 1.71 – 1.62 (br m, 12H; TMP-3,4,5), 1.32 (br m, 12H; TMP-Me),
1.14 (t, 3JHH = 7 Hz, 6H; Et2O), 1.08 (s, 12H; TMP-Me)
13C NMR spectroscopy (125 MHz, 25 ◦C, d8-THF): δ 66.3 (Et2O), 50.1,
50.0 (TMP-2,6), 39.2 (TMP-3,5), 32.2 (TMP-Me), 31.2 (TMP-Me), 19.3
(TMP-4), 15.7 (Et2O)
7Li NMR spectroscopy (194 MHz, 25 ◦C, d8-THF): δ –1.40 (s)
X-ray crystallographic data for 17
C44H92Br2Cu2Li4N4O2, M = 1023.88, triclinic, space group P 1¯, a =
12.7782(2), b = 15.0503(2), c = 15.7127(2) A˚, α = 86.357(5), β = 69.795(5),
γ = 69.874(5) ◦, V = 2657.44(7) A˚3, Z = 2, ρcalcd = 1.280 g cm
−3, Mo-Kα
radiation, λ = 0.71073 A˚, µ = 2.339 mm−1, T = 180(2) K. 46172 data (12143
unique, Rint = 0.0601, θ < 27.47
◦) were collected. wR2 = 0.1573, conven-
tional R = 0.0609 on F values of 9074 reflections with F 2 > 2σ(F 2), S =
1.062, 543 parameters. Residual electron density extrema ±1.205 eA˚−3.
77
3. EXPERIMENTAL PROCEDURES AND RESULTS
b) nButyllithium (2.5 mL, 1.6 M in hexanes, 4 mmol) was added dropwise to a
solution of 2,2,6,6-tetramethylpiperidine (0.68 mL, 4 mmol) in toluene (3 mL)
and diethyl ether (0.42 mL) under N2 at –78
◦C. The resulting solution of
lithium 2,2,6,6-tetramethylpiperidide was allowed to reach room temperature.
It was then added to a suspension of copper(I) bromide (286 mg, 2 mmol)
in toluene (2 mL) under N2 at –78
◦C. The mixture was allowed to reach
room temperature whereupon it was filtered. The resulting yellow solution
deposited 17 as colourless blocks after storage at –27 ◦C for 36 hours. A
crystallographic cell check verified the crystal data obtained by route (a).
Yield: 0.242 g, 24 % with respect to CuBr
Melting Point: decomposed from ca. 194 ◦C
3.4 DMP-based Amidocuprate Compounds
3.4.1 Attempted synthesis of dilithium chloro[di(cis-2,6,-
dimethylpiperidido)]cuprate,
[(DMP)
2
Cu(Cl)Li
2
· OEt
2
]
2
nButyllithium (1.25 mL, 1.6 M in hexanes, 4 mmol) was added dropwise to a
solution of cis-2,6-dimethylpiperidine (0.54 mL, 4 mmol) in hexane (5 mL) and
diethyl ether (0.42 mL) under N2 at –78
◦C. The resulting solution of lithium cis-
2,6-dimethylpiperidide was allowed to reach room temperature. It was then added
to a suspension of copper(I) chloride (198 mg, 2 mmol) in hexane (3 mL) under
N2 at –78
◦C. The mixture was allowed to reach room temperature whereupon it
was filtered and the resulting golden yellow solution deposited 18 as yellow blocks
after storage at –27 ◦C for 12 hours.
Yield: 0.165 g, 22 % Melting Point: decomposed from ca. 87 ◦C
Elemental Analysis: C36H76ClCu1.29Li3.71N4O2 requires C 58.42 %, H 10.35 %,
N 7.57 %, Cl 4.79 %, found C 54.64 %, H 9.66 %, N 7.74 %, Cl 4.05 %
X-ray crystallographic data for 18
C36H76ClCu1.29Li3.71N4O2,M = 740.17, monoclinic, space group C2, a = 25.157(5),
b = 12.689(3), c = 15.149(3) A˚, β = 112.82(3) ◦, V = 4457.2(15) A˚3, Z = 4, ρcalcd
= 1.103 g cm−3, Mo-Kα radiation, λ = 0.71073 A˚, µ = 0.716 mm−1, T = 180(2)
78
3. EXPERIMENTAL PROCEDURES AND RESULTS
K. 20569 data (7930 unique, Rint = 0.0525, θ < 25.35
◦) were collected. wR2 =
0.2035, conventional R = 0.0807 on F values of 5683 reflections with F 2 > 2σ(F 2),
S = 1.047, 423 parameters. Residual electron density extrema ±0.655 eA˚−3.
X-ray crystallographic data indicated that the expected product had not formed,
but instead a mixture of two novel cuprate species, 18. Copper/lithium disorder
was observed for one atomic site in the structure, with a model of 29 % copper
and 71 % lithium providing a best fit for the data after the occupancies were
allowed to refine. In addition, disorder was modelled in the two coordinated Et2O
molecules. Each carbon was modelled split over two sites and the occupancies were
allowed to refine separately for each Et2O molecule, with occupancy summing to
unity in both cases.
On careful observation of the copper/lithium disorder, when lithium is present at
the disordered site, the structure can best be described as a Lipshutz-type monomer
coordinated by a lithium amide dimer, (DMP)2CuLi·LiCl·2(LiDMP)·2Et2O. When
copper is present at the disordered site, the structure is a hitherto unidentified
cuprate species, [(DMP)4Cu2(Li ·Et2O)]LiCl, best described as an adduct between
a Lipshutz-type monomer and a Gilman-type monomer. Efforts from herein focused
on isolating pure samples of the newly discovered cuprate species.
3.4.2 Synthesis of trilithium chloro[tetra(cis-2,6,-dimethyl-
piperidido)]cuprate, [(DMP)
2
CuLi ·OEt
2
]
2
LiCl, 19
nButyllithium (1.25 mL, 1.6 M in hexanes, 2 mmol) was added dropwise to a
solution of cis-2,6,-dimethylpiperidine (0.27 mL, 2 mmol) in diethyl ether (2 mL)
under N2 at –78
◦C. The resulting solution of lithium cis-2,6,-dimethylpiperidide
was allowed to reach room temperature. It was then added to a suspension of
copper(I) chloride (99 mg, 1 mmol) in diethyl ether (2 mL) under N2 at –78
◦C.
The mixture was allowed to reach room temperature whereupon it was filtered and
further diethyl ether (1 mL) was added. The resulting yellow solution deposited
19 as colourless blocks after storage at –27 ◦C for 12 hours.
Yield: 0.046 g, 12 % with respect to CuCl
Melting Point: decomposed from ca. 121 ◦C
Elemental Analysis: C36H76ClCu2Li3N4O2 requires C 55.41 %, H 9.82 %, N 7.18
%, Cl 4.54 %, found C 51.88 %, H 9.61 %, N 7.45 %, Cl 4.38 %
Despite repeated attempts it was not possible to obtain accurate elemental analysis
79
3. EXPERIMENTAL PROCEDURES AND RESULTS
data for this compound, likely due to the highly air and moisture sensitive nature
of the crystalline material.
1H NMR spectroscopy (500 MHz, 25 ◦C, C6D6): δ 3.38 (q,
3JHH = 7 Hz, 8H;
Et2O), 3.08 – 2.55 (br m, 8H; DMP-2,6), 1.88 – 1.58 (br m, 24H; DMP-3,4,5), 1.52
(m, 24H; DMP-Me), 1.22 (t, 3JHH = 7 Hz, 12H; Et2O), 1.07 (br s, 4H; HDMP-Me),
0.85 (br s, 1.5H; HDMP-NH)
13C NMR spectroscopy (100 MHz, 25 ◦C, C6D6): δ 65.6 (Et2O ), 58.7 (DMP-
2,6), 52.4 (HDMP-2,6), 39.4, 38.8 (DMP-3,5), 34.3 (HDMP-3,5), 28.5 (DMP-Me),
28.1, 27.3 (DMP-4), 25.4 (HDMP-4), 23.1 (HDMP-Me), 15.3 (Et2O )
7Li NMR spectroscopy (194 MHz, 25 ◦C, C6D6): δ 2.16 (br s, 0.4Li; LiDMP),
1.83 (br s, 1Li; 19), 1.48 (br s, 2Li ·OEt2; 19), –0.51 (br s, 0.2Li; unidentified)
The peak at 0.85 ppm in the 1H NMR spectrum suggested that some cis-2,6,-
dimethylpiperidine had re-formed during the process of preparing a sample to per-
form NMR spectroscopy experiments, and can be attributed the highly sensitive
nature of the compound. It has not been possible to identify the source of the very
minor signal at δ –0.51 ppm in the 7Li NMR spectrum although it is suspected
that it can be attributed to a solvate of LiCl, for example (LiCl · OEt2)4.
X-ray crystallographic data for 19
C36H76ClCu2Li3N4O2, M = 780.36, orthorhombic, space group Pna21, a =
14.877(5), b = 23.510(5), c = 12.600(5) A˚, V = 4407(2) A˚3, Z = 4, ρcalcd =
1.176 g cm−3, Mo-Kα radiation, λ = 0.71070 A˚, µ = 1.057 mm−1, T = 173.15
K. 90225 data (13112 unique, Rint = 0.0567, θ < 31.029
◦) were collected. wR2 =
0.1429, conventional R = 0.0551 on F values of 11345 reflections with F 2 > 2σ(F 2),
S = 1.111, 457 parameters. Residual electron density extrema ±1.574 eA˚−3.
The crystallographic data was modelled with a small amount of disorder in the
coordinated Et2O molecules. One methyl carbon on each solvent molecule was
refined in separate positions with occupancy summing to unity.
80
3. EXPERIMENTAL PROCEDURES AND RESULTS
3.4.3 Attempted synthesis of trilithium bromo[tetra(cis-
2,6,-dimethylpiperidido)]cuprate,
[(DMP)
2
CuLi · OEt
2
]
2
LiBr
nButyllithium (2.5 mL, 1.6 M in hexanes, 4 mmol) was added dropwise to a solution
of cis-2,6,-dimethylpiperidine (0.54 ml, 4 mmol) in hexane (6 mL) and diethyl
ether (0.42 mL) under N2 at –78
◦C. The resulting solution of lithium cis-2,6,-
dimethylpiperidide was allowed to reach room temperature. It was then added
to a suspension of copper(I) bromide (286 mg, 2 mmol) in hexane (3 mL) under
N2 at –78
◦C. The mixture was allowed to reach room temperature whereupon it
was filtered. The resulting orange solution deposited 20 as orange blocks after
storage at –27 ◦C for 48 hours.
X-ray crystallographic data for 20
C37.65H78.75BrCu2Li3N4.55O1.45, M = 846.44, triclinic, space group P 1¯, a =
12.6429(3), b = 12.9514(3), c = 16.6788(5) A˚, α = 101.5700(10), β = 98.1560(10),
γ = 114.4840(10) ◦, V = 2356.15(11) A˚3, Z = 2, ρcalcd = 1.216 g cm
−3, Mo-Kα
radiation, λ = 0.71070 A˚, µ = 1.782 mm−1, T = 180(2) K. 39087 data (10228
unique, Rint = 0.0637, θ < 27.04
◦) were collected. wR2 = 0.1586, conventional
R = 0.0655 on F values of 5156 reflections with F 2 > 2σ(F 2), S = 1.031, 530
parameters. Residual electron density extrema ±0.622 eA˚−3.
X-ray crystallographic data revealed that an adduct-type cuprate species had
formed, however, some cis-2,6-dimethylpiperidine (HDMP) had re-formed and co-
ordinated to one of the lithium centres in lieu of Et2O in 55 % of the molecules,
giving an overall formula of [(DMP)2CuLi ·OEt2]1.45[(DMP)2CuLi ·HDMP]0.55LiBr.
This disorder proved difficult to model as there were three atomic sites where the
Et2O and HDMP were overlapping entirely. In these cases the atoms were split
and the parts constrained to stay in the same position. The other two atomic
sites in which both Et2O and HDMP were contributing to were refined over two
different sites, and the occupancy refined to the aforementioned ratio of 55:45 %
HDMP:Et2O. The amino proton was a assigned to a q peak and geometric con-
straints were applied to this and the other disordered atomic sites. In addition,
the atoms of the second coordinated Et2O molecule were refined over two sites
with occupancy summing to unity.
Yield: 0.171 g, 40 % with respect to CuBr
81
3. EXPERIMENTAL PROCEDURES AND RESULTS
Melting Point: decomposed from ca. 93 ◦C
Elemental Analysis: C37.65H78.75BrCu2Li3N4.55O1.45 requires C 53.43 %, H 9.38
%, N 7.53 %, Br 9.44 %, found C 55.08 %, H 9.76 %, N 7.92 %, Br 9.94 %
1H NMR spectroscopy (500 MHz, 25 ◦C, C6D6): δ 3.38 (q,
3JHH = 7 Hz, 5.8H;
Et2O), 3.26 – 2.65 (br m, 8H; DMP-2,6), 2.52 (br m, 1.1H; HDMP-2,6), 2.02-1.58
(m, 27.3H; DMP-3,4,5), 1.50 (br m, 22.8H; DMP-Me), 1.18 (t, 3JHH = 7 Hz, 8.4H;
Et2O), 1.10 (d, 4.5H, DMP-Me), 1.10 (d, 4.5H, DMP-Me), 0.87 (br s, 1.6H; HDMP-
NH)
13C NMR spectroscopy (100 MHz, 25 ◦C, C6D6): δ 65.8 (Et2O), 58.9, 57.7
(DMP-2,6), 52.7 (HDMP-2,6), 39.8, 39.5, 38.9 (DMP-3,5), 34.0 (HDMP-3,5), 28.9,
27.5, 27.0 (DMP-Me), 26.6 (DMP-4), 24.8 (HDMP-4), 23.0 (HDMP-Me), 15.2
(Et2O)
7Li NMR spectroscopy (194 MHz, 25 ◦C, C6D6): δ 2.16 (br s, 0.4Li; LiTMP),
1.83 (br s, 1Li; 20), 1.66 and 1.48 (s + sh, 2Li · L, L = 0.5HDMP + 1.5Et2O; 20)
As was observed in the 1H NMR spectrum of 19, the peak at 0.87 ppm again sug-
gested that some cis-2,6,-dimethylpiperidine had re-formed during the process of
preparing a sample to perform NMR spectroscopy experiments, indicating that the
adduct-type species are highly sensitive regardless of the identity of the central
halide anion.
3.4.4 Synthesis of trilithium bromo[tetra(cis-2,6,-dimethyl-
piperidido)]cuprate, [(DMP)
2
CuLi ·OEt
2
]
2
LiBr, 21
a) A room temperature solution of lithium cis-2,6,-dimethylpiperidide was pre-
pared as for 19 and was added to a suspension of copper(I) bromide (143
mg, 1 mmol) in diethyl ether (2 mL) under N2 at –78
◦C. The mixture was
allowed to reach room temperature whereupon it was filtered. The resulting
pale orange solution deposited 21 as yellow blocks after storage at –27 ◦C for
12 hours.
Yield: 0.088 g, 21 % with respect to CuBr
Melting Point: decomposed from ca. 95 ◦C
Elemental Analysis: C36H76BrCu2Li3N4O2 requires C 52.42 %, H 9.29 %,
N 6.79 %, Br 9.69; found C 50.43 %, H 8.93 %, N 6.56 %, Br 10.25 %
1H NMR spectroscopy (500 MHz, 25 ◦C, C6D6): δ 3.38 (q,
3JHH = 7
Hz, 8H; Et2O), 3.08 – 2.76 (br m, 8H; DMP-2,6), 1.92 – 1.60 (br m, 24H;
82
3. EXPERIMENTAL PROCEDURES AND RESULTS
DMP-3,4,5), 1.52 (m, 24H; DMP-Me), 1.21 (t, 3JHH = 7 Hz, 12H; Et2O),
1.09 (br s, 2H; DMP-Me), 0.85 (br s, 1.3H; HDMP-NH)
13C NMR spectroscopy (100 MHz, 25 ◦C, C6D6): δ 65.7 (Et2O), 59.4
(DMP-2,6), 52.5 (HDMP-2,6), 40.9, 40.2, 38.8 (DMP-3,5), 34.1 (HDMP-3,5),
28.7, 28.0 (DMP-Me), 27.3 (DMP-4), 26.5 (HDMP-4), 23.0 (HDMP-Me),
15.2 (Et2O)
7Li NMR spectroscopy (194 MHz, 25 ◦C, C6D6): δ 2.18 (br s, 0.4Li;
LiTMP), 1.84 (br s, 1Li; 21), 1.48 (br s, 2Li · Et2O, 21)
The peak at 0.85 ppm in the 1H NMR spectrum again suggests that some
cis-2,6,-dimethylpiperidine had re-formed during the process of preparing a
sample to perform NMR spectroscopy experiments.
X-ray crystallographic data for 21 route (a)
C36H76BrCu2Li3N4O2, M = 824.82, orthorhombic, space group Pna21, a =
14.8880(2), b = 23.7813(5), c = 12.6698(2) A˚, V = 4485.82(13) A˚3, Z = 4,
ρcalcd = 1.221 g cm
−3, Mo-Kα radiation, λ = 0.71073 A˚, µ = 1.870 mm−1, T =
180(2) K. 16869 data (8233 unique, Rint = 0.0329, θ < 27.47
◦) were collected.
wR2 = 0.0950, conventional R = 0.0420 on F values of 6987 reflections with
F 2 > 2σ(F 2), S = 1.015, 480 parameters. Residual electron density extrema
±0.661 eA˚−3.
The crystallographic data revealed disorder in one Et2O molecule, and so
each atom was split into two separate positions with the occupancy summing
to unity.
b) nButyllithium (4.14 mL, 1.6 M in hexanes, 6.6 mmol) was added dropwise to
a solution of cis-2,6,-dimethylpiperidine (0.81 mL, 6 mmol) in hexane (6 mL)
and diethyl ether (0.63 mL) under N2 at –78
◦C. The resulting solution of
lithium cis-2,6,-dimethylpiperidide was allowed to reach room temperature.
It was then added to a suspension of copper(I) bromide (429 mg, 3 mmol)
in hexane (3 mL) under N2 at –78
◦C. The mixture was allowed to reach
room temperature whereupon it was filtered. The resulting orange solution
deposited 21 as orange/brown blocks after storage at –27 ◦C for 72 hours.
Yield: 0.584 g, 47 % with respect to CuBr
Melting Point: decomposed from ca. 95 ◦C
Elemental Analysis: C36H76BrCu2Li3N4O2 requires C 52.42 %, H 9.29 %,
83
3. EXPERIMENTAL PROCEDURES AND RESULTS
N 6.79 %, Br 9.69 %, found C 52.32 %, H 9.16 %, N 7.75 %, Br 10.62 %
1H NMR spectroscopy (500 MHz, 25 ◦C, C6D6): δ 3.35 (q,
3JHH = 7 Hz,
8H; Et2O), 3.28 – 2.66 (m, 8H; DMP-2,6), 2.20 (br m, 4H; DMP-4), 2.02 –
1.70 (m, 10H; DMP-3,5), 1.62 (br m, 10H DMP-3,4,5) 1.50 (d, 22H; DMP-
Me), 1.12 (t, 3JHH = 7 Hz, 12H; Et2O), 1.08 (d, 2H, DMP-Me), 0.85 (br s,
0.6H; HDMP-NH)
13C NMR spectroscopy (100 MHz, 25 ◦C, C6D6): δ 65.7 (Et2O), 58.6, 57.4
(DMP-2,6), 52.7 (HDMP-2,6) 39.6, 38.5 (DMP-3,5), 33.7 (HDMP-3,5), 27.4
(DMP-Me), 26.8 (DMP-4), 24.5 (HDMP-4), 22.7 (HDMP-Me), 14.7 (Et2O)
7Li NMR spectroscopy (194 MHz, 25 ◦C, C6D6): δ 2.15 (br s, 0.6Li;
LiTMP), 1.83 (br s, 1Li; 21), 1.50 (br s, 2Li ·OEt2; 21)
The peak at 0.85 ppm in the 1H NMR spectrum again suggested that some
cis-2,6,-dimethylpiperidine had re-formed during the process of preparing a
sample to perform NMR spectroscopy experiments.
X-ray crystallographic data for 21 route (b)
C36H76BrCu2Li3N4O2,M = 824.82, triclinic, space group P 1¯, a = 12.4266(4),
b = 12.7166(4), c = 16.5053(7) A˚, α = 102.634(2), β = 95.762(2), γ =
113.775(2) ◦, V = 2276.57(14) A˚3, Z = 2, ρcalcd = 1.203 g cm
−3, Mo-Kα
radiation, λ = 0.71073 A˚, µ = 1.842 mm−1, T = 180(2) K. 12561 data
(6862 unique, Rint = 0.0388, θ < 25.24
◦) were collected. wR2 = 0.2631,
conventional R = 0.0842 on F values of 3859 reflections with F 2 > 2σ(F 2),
S = 1.087, 387 parameters. Residual electron density extrema ±1.242 eA˚−3.
Significant disorder was observed in the coordinated Et2O groups and so
geometric constraints were applied. In addition, all Et2O carbon and oxygen
atoms were refined only isotropically, since the model was not strong enough
to support anisotropic refinement.
3.4.5 Synthesis of trilithium iodo[tetra(cis-2,6,-dimethyl-
piperidido)]cuprate, [(DMP)
2
CuLi ·OEt
2
]
2
LiI, 22
nButyllithium (1.25 mL, 1.6 M in hexanes, 2 mmol) was added dropwise to a
solution of cis-2,6,-dimethylpiperidine (0.27 mL, 2 mmol) in hexane (2 mL) and
84
3. EXPERIMENTAL PROCEDURES AND RESULTS
diethyl ether (0.1 mL) under N2 at –78
◦C. The resulting solution of lithium cis-
2,6,-dimethylpiperidide was allowed to reach room temperature. It was then added
to a suspension of copper(I) iodide (180 mg, 1 mmol) in hexane (2 mL) under N2
at –78 ◦C. The mixture was allowed to reach room temperature whereupon it was
filtered. The resulting pale yellow solution deposited 22 as colourless blocks after
storage at –27 ◦C for 24 hours.
Yield: 0.056 g, 14 % with respect to CuI
Melting Point: decomposed from ca. 103 ◦C
Elemental Analysis: C36H76ICu2Li3N4O2 requires C 49.60 %, H 8.62 %, N 6.17
%, I 14.56 % found C 48.59 %, H 8.62 %, N 6.43 %, I 14.83 %
1H NMR spectroscopy (500 MHz, 25 ◦C, C6D6): δ 3.40 (q,
3JHH = 7 Hz, 8H;
Et2O), 3.05 – 2.70 (br m, 8H; DMP-2,6), 2.01 – 1.60 (br m, 24H; DMP-3,4,5),
1.60 – 1.31 (br m, 18H; DMP-Me), 1.18 (t, 3JHH = 7 Hz, 12H; Et2O), 1.16 (s, 6H;
DMP-Me), 0.93 (br s, 4H; HDMP-NH)
13C NMR spectroscopy (100 MHz, 25 ◦C, C6D6): δ 65.9 (Et2O), 58.8 (DMP-
2,6), 52.8 (HDMP-2,6), 39.6 (DMP-3,5), 33.9 (HDMP-3,5), 27.4 (DMP-Me), 26.4
(HDMP-4), 24.7 (DMP-4), 23.1 (HDMP-Me), 15.1 (Et2O)
7Li NMR spectroscopy (194 MHz, 25 ◦C, C6D6): δ 2.17 (br s, 0.2Li; LiTMP),
1.84 (br s, 1Li; 22), 1.41 (br s, 2Li · OEt2; 22)
The peak at 0.93 ppm in the 1H NMR spectrum again suggested that some cis-
2,6,-dimethylpiperidine had re-formed during the process of preparing a sample to
perform NMR spectroscopy experiments.
X-ray crystallographic data for 22
C36H76ICu2Li3N4O2, M = 851.65, triclinic, space group P 1¯, a = 12.733(3), b =
12.961(3), c = 15.777(3) A˚, α = 78.34(2), β = 88.76(3), γ = 63.77(3) ◦, V =
2280.5(8) A˚3, Z = 2, ρcalcd = 1.240 g cm
−3, Mo-Kα radiation, λ = 0.71073 A˚,
µ = 1.639 mm−1, T = 173(2) K. 22384 data (8031 unique, Rint = 0.0570, θ <
25.38 ◦) were collected. wR2 = 0.2205, conventional R = 0.0789 on F values of
5723 reflections with F 2 > 2σ(F 2), S = 1.057, 455 parameters. Residual electron
density extrema ±1.572 eA˚−3.
While refinement of the structure proved straightforward, the refinement required
the central iodine atom to be split over two sites, and each was modelled with
1/2 occupancy. In addition, displacement constraints were applied to the carbon
85
3. EXPERIMENTAL PROCEDURES AND RESULTS
atoms in one of the Et2O molecules.
3.4.6 Synthesis of trilithium bromo[tetra(cis-2,6,-dimethyl-
piperidido)]cuprate, [(DMP)
2
CuLi · 2THF]
2
LiBr, 23
a) nButyllithium (1.25 mL, 1.6 M in hexanes, 4 mmol) was added dropwise
to a solution of cis-2,6,-dimethylpiperidine (0.54 mL, 4 mmol) in toluene (5
mL) under N2 at –78
◦C. The resulting solution of lithium cis-2,6,-dimethyl-
piperidide was allowed to reach room temperature. It was then added to a
suspension of copper(I) bromide (286 mg, 2 mmol) in toluene (2 mL) under N2
at –78 ◦C. The solvent was removed and THF (8 mL) was added, whereupon
the mixture was filtered. The resulting solution was reduced to ca. 3 mL
volume and 12 mL hexane was added. The resulting orange-red solution
deposited 23 as fine, colourless needles after storage at –27 ◦C for 24 hours.
Yield: 0.141 g, 15 % with respect to CuBr
Melting Point: decomposed from ca. 81 ◦C
Elemental Analysis: C44H88BrCu2Li3N4O4 requires Br 8.28 %, found Br
8.13 %
It was not possible to collect satisfactory C, H and N data for this compound
despite repeated attempts.
1H NMR spectroscopy (500 MHz, 25 ◦C, C6D6): δ 3.67 (m, 16H; THF),
3.19 – 2.74 (br m, 8H; DMP-2,6), 2.25 (br m, 4H; DMP-3), 2.07 – 1.76 (br
m, 16H; DMP-3,4,5), 1.63 (br m, 23H; DMP-Me), 1.50 (m, 4H; DMP-4,5),
1.42 (m, 16H; THF), 1.09 (d, 1H; DMP-Me)
13C NMR spectroscopy (125 MHz, 25 ◦C, C6D6): δ 68.2 (THF), 59.0, 52.4
(DMP-2,6), 39.7 (DMP-3,5), 34.2 (DMP-4), 27.5 (DMP-Me), 27.3 (DMP-4),
25.3 (THF), 23.1 (DMP-Me)
7Li NMR spectroscopy (194 MHz, 25 ◦C, C6D6): δ 1.52 (br s)
X-ray crystallographic data for 23
C14.67H29.33Br0.33Cu0.67LiN1.33O1.33, M = 321.66, triclinic, space group P 1¯, a
= 12.234(2), b = 14.798(3), c = 17.745(4) A˚, α = 112.84(3), β = 91.90(3),
γ = 113.92(3) ◦, V = 2636.4(9) A˚3, Z = 6, ρcalcd = 1.216 g cm
−3, Mo-
Kα radiation, λ = 0.71073 A˚, µ = 1.603 mm−1, T = 180(2) K. 38698 data
86
3. EXPERIMENTAL PROCEDURES AND RESULTS
(9608 unique, Rint = 0.1630, θ < 25.44
◦) were collected. wR2 = 0.1162,
conventional R = 0.0734 on F values of 4462 reflections with F 2 > 2σ(F 2),
S = 1.016, 531 parameters. Residual electron density extrema ±0.397 eA˚−3.
b) nButyllithium (1.25 mL, 1.6 M in hexanes, 4 mmol) was added dropwise
to a solution of cis-2,6,-dimethylpiperidine (0.54 mL, 4 mmol) in toluene (5
mL) under N2 at –78
◦C. The resulting solution of lithium cis-2,6,-dimethyl-
piperidide was allowed to reach room temperature. It was then added to a
suspension of copper(I) bromide (286 mg, 2 mmol) in toluene (2 mL) under
N2 at –78
◦C. The solvent was removed and THF (8 mL) was added, where-
upon the mixture was filtered and the solution reduced to ca. 2 mL volume.
The resulting orange-brown solution deposited 23 as colourless blocks after
storage at –27 ◦C for 48 hours. A crystallographic cell check verified the
crystallographic data obtained by route (a).
Yield: 0.246 g, 26 % with respect to CuBr
Melting Point: decomposed from ca. 80 ◦C
Elemental Analysis: C44H88BrCu2Li3N4O4 requires C 54.76 %, H 9.19 %,
N 5.81 %, Br 8.28 %, found C 53.93 %, H 8.97 %, N 5.85 %, Br 7.94 %
1H NMR spectroscopy (500 MHz, 25 ◦C, C6D6): δ 3.68 (m, 16H; THF),
3.27 – 2.65 (br m, 8H; DMP-2,6), 2.25 (br m, 4H; DMP-3), 2.07 – 1.76 (br
m, 16H; DMP-3,4,5), 1.66 (br m, 23H; DMP-Me), 1.50 (m, 4H; DMP-4,5),
1.42 (m, 16H; THF), 1.09 (d, 1H; DMP-Me)
13C NMR spectroscopy (125 MHz, 25 ◦C, C6D6): δ 68.2 (THF), 59.1, 52.5
(DMP-2,6), 39.7 (DMP-3,5), 34.2 (DMP-4), 27.5 (DMP-Me), 27.3 (DMP-4),
25.3 (THF), 23.1 (DMP-Me)
7Li NMR spectroscopy (194 MHz, 25 ◦C, C6D6): δ 1.55 (br s)
3.5 Directed ortho Cupration Reactions
3.5.1 Preparation of 2-iodobenzonitrile using an in situ
dicyclohexylphospine-based Gilman-type formula-
tion base
nButyllithium (5 mL, 1.6 M in hexanes, 8 mmol) was added to a solution of di-
cyclohexylphosphine (1.6 mL, 8 mmol) in THF (20 mL) under N2 at –78
◦C. The
87
3. EXPERIMENTAL PROCEDURES AND RESULTS
resulting solution of LiPCy2 was then added to a suspension of copper(I) cyanide
(0.36 g, 4 mmol) in THF (12 mL) under N2 at –78
◦C. The resulting yellow solution
was allowed to reach room temperature whereupon it was returned to –78 ◦C and
treated with benzonitrile (0.21 mL, 2 mmol). The reaction vessel was returned to
0 ◦C and stirred for 3 hours whereupon it was cooled once again to –78 ◦C and
I2 (3.55 g, 14 mmol) was introduced as a THF solution (8 mL). The mixture was
allowed to reach room temperature at which point it was stirred for 16 hours. The
reaction mixture was then poured into sat. NaHCO3 and sat. NH4Cl (50 mL each)
and extracted with AcOEt. The AcOEt layer was washed with brine, dried over
(MgSO4), filtered, and the solvent removed in vacuo. The product was purified
by silica gel flash chromatography.
1H NMR spectroscopy of the crude product suggested that 2-iodobenzonitrile was
present in low yields, however, further purification yielded only a trace quantity of
the expected product and it was not possible to carry out further characterisation.
3.5.2 Preparation of 2-iodobenzonitrile using pre-isolated
11
Crystals of 11 were prepared as described in Section 3.2.3 and subsequently isolated
after storage at –27 ◦C for at least 24 hours. 1.57 g (2.8 mmol) of 11 was placed in
a Schlenk tube and dissolved in THF (20 mL), whereupon the solution was treated
with benzonitrile (0.148 mL, 1.4 mmol) under N2 at –78
◦C. The reaction vessel
was returned to 0 ◦C and stirred for 3 hours whereupon it was cooled once again
to –78 ◦C and I2 (2.54 g, 10 mmol) was introduced as a THF solution (6 mL). The
mixture was allowed to reach room temperature at which point it was stirred for
16 hours. Work-up was as above and yielded 2-iodobenzonitrile as a yellow solid.
Yield: 0.029 g, 9 %
1H NMR spectroscopy (500 MHz, 25 ◦C, C6D6): 8.03 (d, J = 7 Hz, 1H; Ar),
7.62 (dd, J = 7 Hz, 1 Hz, 1H; Ar), 7.48 (t, J = 8 Hz, 1H; Ar), 7.41 (m, 1H; Ar)
While the signals in the 1H NMR spectrum match those reported in the literature,15
the spectrum also shows that there is a high level of impurities present in the sample
(although this does not include unreacted starting material158) suggesting that the
product is very difficult to separate from the reaction mixture.
88
3. EXPERIMENTAL PROCEDURES AND RESULTS
3.5.3 General procedure for the preparation of 2-chloro-
pyridin-3-yl phenyl ketone using an in situ cuprate
base prepared from CuCl
2
· TMEDA
nButyllithium (1.25 mL, 1.6 M in hexanes, 2 mmol) and a solution of lithium
2,2,6,6-tetramethylpiperide (4 mmol) prepared in toluene (4 mL) at 0 ◦C from
2,2,6,6-tetramethylpiperidine (0.68 mL, 4 mmol) and nbutyllithium (2.5 mL, 1.6 M
in hexanes, 4 mmol) were successively added to a suspension of CuCl2 · TMEDA
(0.5 g, 2.0 mmol) in toluene (4 mL) at 0 ◦C. The resulting mixture was stirred for 15
minutes at this temperature before being treated with 2-chloropyridine (0.16 mL,
2 mmol). The reaction vessel was returned to room temperature and stirred for 2
hours whereupon it was cooled once again to 0 ◦C and benzoyl chloride (0.48 mL,
4 mmol) was added. The mixture was allowed to reach room temperature at which
point it was stirred for 16 hours. Sat. ammonium chloride (5 mL) was then added
to the solution and the organic phase extracted with Et2O. The combined organic
layers were washed with brine (10 mL), dried over Na2SO4 or MgSO4, filtered,
and the solvent removed in vacuo. Purification by silica gel flash chromatography
yielded 2-chloropyridin-3-yl phenyl ketone as a yellow oil.
1H NMR spectroscopy (300 MHz, CDCl3): δ 8.52 ppm (dd, J = 5 Hz, 2 Hz,
1H; Ar), 7.81 – 7.75 (m, 2H; Ar), 7.72 (dd, J = 8 Hz, 2 Hz, 1H; Ar), 7.61 (m, 1H;
Ar), 7.50 – 7.42 (m, 2H; Ar), 7.38 (dd, J = 8 Hz, 5 Hz, 1H; Ar)
13C NMR spectroscopy (75 MHz, CDCl3): δ 193.3 (CO), 150.9, 147.7, 138.0,
135.7, 134.9, 134.3, 130.0, 128.9, 122.3 (Ar)
3.5.4 General procedure for the preparation of 2-chloro-
pyridin-3-yl phenyl ketone using an in situ cuprate
base prepared from CuCl
A solution of lithium 2,2,6,6-tetramethylpiperide (4 mmol) prepared from 2,2,6,6-
tetramethylpiperidine (0.68 mL, 4 mmol) and nbutyllithium (2.5 mL, 1.6 M in
hexanes, 4 mmol) was added to a suspension of CuCl (0.198 g, 2 mmol) at 0 ◦C.
The resulting mixture was stirred for 15 minutes at this temperature or allowed to
warm to room temperature and filtered. The resulting mixture or solution was then
treated with 2-chloropyridine (0.16 mL, 2 mmol). The reaction vessel was returned
to room temperature and stirred for 2 hours whereupon it was cooled once again to
89
3. EXPERIMENTAL PROCEDURES AND RESULTS
0 ◦C and benzoyl chloride (0.48 mL, 4 mmol) was added. The mixture was allowed
to reach room temperature at which point it was stirred for 16 hours. Work-up
was as above and yielded 2-chloropyridin-3-yl phenyl ketone as a yellow oil, as
verified by NMR spectroscopy.
3.5.5 Preparation of 2-chloropyridin-3-yl phenyl ketone us-
ing pre-isolated 2 in toluene
Crystals of 2 were prepared as above and subsequently isolated after storage at +5
◦C for at least 24 hours. 0.176 g (0.5 mmol) of 2 was dissolved in toluene (4 mL).
This was treated with a solution of 2-chloropyridine (0.04 mL, 0.5 mmol) under
N2 at 0
◦C. The reaction vessel was returned to room temperature and stirred for
3 hours whereupon it was cooled once again to 0 ◦C and benzoyl chloride (0.06
mL, 0.5 mmol) was added. The mixture was allowed to reach room temperature
at which point it was stirred for 16 hours. Work-up was as above and yielded 2-
chloropyridin-3-yl phenyl ketone as a yellow oil, as verified by NMR spectroscopy.
Yield: 0.016 g, 15 %
3.5.6 Preparation of 2-chloropyridin-3-yl phenyl ketone us-
ing pre-isolated 2 in THF
Crystals of 2 were prepared as above and subsequently isolated after storage at
+5 ◦C for at least 24 hours. 0.352 g (1 mmol) of 2 was dissolved in THF (2 mL)
and treated with a solution of 2-chloropyridine (0.08 mL, 1 mmol) under N2 at 0
◦C. The reaction vessel was returned to room temperature and stirred for 3 hours
whereupon it was cooled once again to 0 ◦C and benzoyl chloride (0.12 mL, 1 mmol)
was added. The mixture was allowed to reach room temperature at which point it
was stirred for 16 hours. Work-up was as above and yielded 2-chloropyridin-3-yl
phenyl ketone as a yellow oil, as verified by NMR spectroscopy.
Yield: 0.054 g, 25 %
90
3. EXPERIMENTAL PROCEDURES AND RESULTS
3.5.7 Preparation of 2-iodo-N,N -diethylbenzamide using
an in situ DMP-based Lipshutz-type formulation
base
nButyllithium (10 mL, 1.6 M in hexanes, 16 mmol) was added to a solution of 2,6-
cis-dimethylpiperidine (2.16 mL, 16 mmol) in THF (6 mL) under N2 at –78
◦C.
The resulting solution of LiDMP was then added to a suspension of CuBr (1.14 g, 8
mmol) in THF (4 mL) under N2 at –78
◦C. The resulting black slurry was allowed
to reach room temperature whereupon it was filtered. This resulting dark orange
solution was treated with a solution of N,N -diethylbenzamide (0.82 g, 4 mmol) in
THF (5 mL) under N2 at –78
◦C. The reaction vessel was returned to 0 ◦C and
stirred for 3 hours whereupon it was cooled once again to –78 ◦C and I2 (2.02 g, 8
mmol) was introduced as a THF solution (10 mL). The mixture was allowed to reach
room temperature whereupon it was stirred for 16 hours. The reaction mixture was
then poured into sat. NaHCO3, sat. NH4Cl (50 mL each) and sat. Na2S2O3 (200
mL) and extracted with AcOEt. The AcOEt layer was washed with brine, dried
over MgSO4, filtered, and the solvent removed in vacuo. Purification by silica gel
flash chromatography yielded 2-iodo-N,N -diethylbenzamide as a yellow-white solid.
Yield: 1.06 g, 80 %
Melting Point: 188–189 ◦C
1H NMR spectroscopy (400 MHz, 27 ◦C, CDCl3): δ 7.81 (dd, J = 8 Hz, 1 Hz,
1H; Ar), 7.35 (dt, J = 7 Hz, 1 Hz, 1H; Ar), 7.13 (dd, J = 7 Hz, 2 Hz, 1H; Ar),
7.02 (dt, J = 8 Hz, 2 Hz, 1H; Ar), 3.58 (sept, J = 7 Hz, 1H; NCH), 3.51 (sept, J
= 7 Hz, 1H; NCH), 1.60 (d, J = 7 Hz, 3H; Me), 1.56 (d, J = 7 Hz, 3H; Me), 1.27
(d, J = 7 Hz, 3H; Me), 1.06 (d, J = 7 Hz, 3H; Me)
13C NMR spectroscopy (100 MHz, 27 ◦C, CDCl3): δ 169.8 (CO), 144.3 (1-Ar),
139.3, 129.4, 128.2, 125.8 (Ar), 92.3 (2-Ar), 51.2, 46.0 (NC), 20.8, 20.7, 20.1 (Me)
3.5.8 Preparation of 2-iodo-N,N -diethylbenzamide using
pre-isolated 21
Crystals of 21 were prepared via route (b) (Section 3.4.4) and subsequently isolated
after storage at –27 ◦C for at least 24 hours. 1.15 g (1.4 mmol) of 21 was dissolved
in THF (7 mL). This was treated with a solution of N,N -diethylbenzamide (0.29
g, 1.4 mmol) in THF (5 mL) at –78 ◦C. The reaction vessel was returned to 0 ◦C
91
3. EXPERIMENTAL PROCEDURES AND RESULTS
and stirred for 3 hours whereupon it was cooled once again to –78 ◦C and I2 (0.72
g, 2.8 mmol) was introduced as a THF solution (10 mL). The mixture was allowed
to reach room temperature at which point it was stirred for 16 hours. Work-up
was as above and yielded 2-iodo-N,N -diethylbenzamide as a yellow-white solid, as
verified by NMR spectroscopy.
Yield: 0.38 g, 82 %
Melting Point: 185–186 ◦C
92
Chapter 4
Phosphidocopper and
Phosphidocuprate Compounds
4.1 Introduction
With the aim of creating stable lithium cuprates in order to study the equilib-
rium between Gilman-type and Lipshutz-type forms, syntheses of bis(phosphido)-
cuprates were attempted. Phosphidocuprates contain a 31P NMR spectroscopic
handle, which would allow the equilibrium in the solution-state to be studied.
Phosphido ligands share many properties with their amido counterparts; phos-
phorus can be found just below nitrogen in Group 15 of the periodic table and
hence their chemistries are expected to be similar. For example, theoretical gas-
phase acidity studies have shown that the acidities of secondary phosphines and
amines are comparable (∆Gf
◦ = 343.7 (PhP2H), 364.0 (HPCy2), 382.2 (HTMP)
kcal mol−1)159∗ and as the phosphine acidities are slightly lower they should react
more readily with lithiating agents such as nBuLi. For these reasons, attempts
to synthesise phosphidocuprates followed methodologies previously developed to
form amidocuprates,15,139 namely mixing 0.5 equivalents of a CuI salt with a lith-
ium phosphide, in turn prepared from a 1:1 mixture between the phosphine and
nBuLi (Scheme 4.1).
∗ The measured pK a values for HPPh2 and HPCy2 were 22.9 and 34.6 respectively
93
4. PHOSPHIDOCOPPER AND PHOSPHIDOCUPRATE
COMPOUNDS
2 R2PH
Tol or THF, –78 oC
2 R2PLi
 2 nBuLi CuCN
Tol or THF, –78 oC
 (R2P)2CuLi.nS or  
(R2P)2Cu(CN)Li2.nS
Scheme 4.1: General reaction scheme for syntheses of phosphidocuprates. The
choice of solvent was consistent throughout each step displayed
4.2 [(Ph2P)6Cu4][Li · 4THF]2, 9
Sterically demanding amide ligands are a key factor in the formation of stable,
reactive bis(amido)cuprates, and so a bulky secondary phosphine, diphenylphos-
phine, was chosen as a suitable substrate to form lithium bis(phosphido)cuprates.
The first attempt to fabricate a phenylphosphidocuprate of either Gilman- or
Lipshutz-type formulation involved generation of LiPPh2 in a hydrocarbon medium
before addition to 0.5 equivalents of CuCN. The resulting solution was concen-
trated, followed by the addition of THF, allowing X-ray grade crystals to be iso-
lated in a low yield. Crystallographic analysis revealed an unexpected product,
[(Ph2P)6Cu4][Li · 4THF]2, 9 (Figure 4.1). While this result proved problematic
to reliably reproduce, a similar preparation employing CuI as the copper source
carried out solely in THF led to the isolation of 9 in a moderate yield. As copper(I)
halides had proven to be suitable starting materials to form bis(amido)cuprates,138
it was postulated that if the synthesis of 9 was repeated replacing CuCN with
CuI, the absence of a cyano group would prevent a polymer from forming, instead
allowing a lithium bis(phosphido)cuprate to be isolated. The addition of LiPPh2
to 0.5 equivalents of CuI in THF solvent in fact led to the crystallisation of 9, as
confirmed by an X-ray crystallographic cell check. While elemental analysis of 9
confirmed the purity of the material, 31P NMR spectroscopic data of redissolved
crystals revealed a dominant peak attributable to the species at δ –22.1 ppm along-
side a minor doublet at δ –32.6 ppm attributable to HPPh2. This suggests that
high moisture-sensitivity causes decomposition of 9 in solution.
The structure of 9 contains an interesting dianionic adamantoid cage made up of
six Ph2P ligands and four Cu centres, with two four-fold THF-solvated Li cations
acting as the counterions. The P−Cu bond lengths in 9 (2.287 A˚av., range 0.01 A˚)
are comparable with those in known phosphidocopper compounds (2.279 A˚, range
0.16 A˚)∗. In typical adamantane structures160,161 each vertex exhibits a regular
tetrahedral geometry. However, as sterics dictate the arrangement of the three
ligands bound to each Cu centre in 9 they exhibit approximately trigonal planar
∗ Taken from a search of the CSD, which found 6 compounds
94
4. PHOSPHIDOCOPPER AND PHOSPHIDOCUPRATE
COMPOUNDS
Figure 4.1: Structure of the [(Ph2P)6Cu4]
2− core of [(Ph2P)6Cu4][Li · 4THF]2, 9,
H and C atoms (except for C-ipso) omitted for clarity
geometry (the angles around Cu(1) and Cu(2) sum to 355.77 and 354.86◦ respec-
tively), while the (µ-P)−Cu−(µ-P) angles are acute (Cu(1/2)−P(1/2)−Cu(2/2A)
84.85(8)/85.80(8)◦) so allowing the adamantoid cage to form.
The literature contains examples of characterised copper-based adamantoid clusters
incorporating both group 16162–165 and 17166–169 inorganic ligands. However, such
clusters incorporating phosphorus-based donor ligands are limited to the mixed-
ligand systems (Phos)2Cu4Cl4 (Figure 4.2(a)), which forms a distorted adamantyl
cage in the solid-state,170 and (odpf)2Cu4X4 (X = Cl, Br), which contains both
tri- and tetra-coordinated Cu centres.171 The most closely related structure to
9, [(Ph2P)10Cd4]
2– (Figure 4.2(b)), was synthesised by Wright et al. and con-
tains a cadmium-phosphide core.172 The cadmium ions have a formal oxidation
state of +2 and hence an identical valence electronic configuration to CuI. Al-
though, as each Cd centre is bound to an exo Ph2P ligand, the Cd cations are four-
coordinate and so exhibit approximate tetrahedral geometry (P−Cd−P 108.7◦av.).
However, it is noteworthy that the geometry at the Cd centres is far from uni-
form ((µ-P)−Cu−(µ-P) 99.11(8) – 109.43(9)◦) unlike in 9, and the Cd−P−Cd
angles (115.97(9) – 121.62(9)◦) are significantly more obtuse than their Cu−P−Cu
analogues in 9.
The isolation of 9 from reactions carried out with different copper sources and under
varying conditions suggests that the product is thermodynamically highly favoured.
In the preparation from CuCN in toluene a grey precipitate forms, strongly sug-
gesting that LiCN precipitates out of solution. This is removed by filtration and the
95
4. PHOSPHIDOCOPPER AND PHOSPHIDOCUPRATE
COMPOUNDS
(a) (b)
Figure 4.2: Structures of (a) the distorted adamantyl P2Cu4Cl4 core of
(Phos)2Cu4Cl4
170 and (b) the P10Cd4 core of [(Ph2P)10Cd4][Li · 4THF]2,
172 H, C
and N atoms omitted for clarity
remaining building blocks evidently then favour assembly into an adamantoid cage
over a cuprate. When CuI is employed in THF, the resulting mixture is heated into
solution without filtration and hence the lithium salt (in this case LiI) remains in
the reaction mixture. As 9 crystallises in favour of a solvated LiI salt, it indicates
that 9 has a very high lattice enthalpy.
P(1)−Cu(1) 2.289(2) P(1)−Cu(1)−P(1A) 118.59(3)
P(1)−Cu(2) 2.289(2) P(1)−Cu(2)−P(2) 117.32(9)
P(2)−Cu(2) 2.280(2) P(1)−Cu(2)−P(2A) 118.80(9)
P(2A)−Cu(2) 2.290(2) P(2)−Cu(2)−P(2A) 118.74(11)
Cu(1)−P(1)−Cu(2) 84.85(8)
Cu(2)−P(2)−Cu(2A) 85.80(8)
Table 4.1: Selected bond lengths, A˚, and angles, ◦, for 9
4.3 [Ph2PCu(CN)Li · 2THF]∞, 10∞
As reactions with lithium diphenylphosphide in hydrocarbon media had proved
unsuccessful, the focus moved to attempting to form monomeric Gilman- and
Lipshutz-type cuprates in the presence of donor solvents. To this end, CuCN
was treated with 2 equivalents of LiPPh2 in a THF medium. Analysis of the X-ray
grade crystals obtained on repeated attempts revealed the species to be a poly-
meric array with empirical formula Ph2PCu(CN)Li · 2THF, 10 (Figure 4.3). The
96
4. PHOSPHIDOCOPPER AND PHOSPHIDOCUPRATE
COMPOUNDS
{1H}31P NMR spectrum of a redissolved sample of 10 displayed a major signal
attributable to the species at δ –14.9 ppm and, amongst other minor signals, a
peak at δ –38.3 ppm which split into a doublet on removal of proton decoupling.
The latter signal matches the dominant signal from a reference spectrum of HPPh2
(δ –39.9 ppm) and so can be attributed to solvent decomposition of the sample, al-
though as the elemental analysis data matched the theoretical values very closely,
the crystalline material was clearly very pure.
Figure 4.3: Structure of the asymmetric unit of 10, [(Ph2P)Cu(CN)Li · 2THF]2, H
atoms omitted for clarity
The dimer displayed in Figure 4.3 clearly shows a central 4-membered (NLi)2
metallocycle at the centre of the asymmetric unit, which forms from the inter-
action between the cyanide ligands and doubly THF-solvated Li centres. The
(NLi)2 metallocycles link the PCu units, which aggregate to form a highly unusual
supramolecular array containing six-membered (PCu)3 rings acting as trigonal-
symmetry nodes, sitting in the core of [(Ph2P)3(CuCN)3]
3– anions (Figure 4.4).
Each Cu centre is formally bound to just one diphenylphosphido unit and three
such PCu units interact with each other to form cyclohexane-type (PCu)3 nodes
within the supramolecular array.
The P−Cu bond lengths (2.262 A˚av.) are consistent with those observed in 9
(2.287 A˚av.) and do not vary within each (PCu)3 ring (e.g. P(1A/2A)−Cu(1/2)
2.265(2)/2.256(2) A˚). The latter data indicate that electron density is distributed
equally around the (PCu)3 ring. This means the structure cannot be viewed as a
simple aggregation of dimeric [Ph2PCu(CN)Li · 2THF]2 units, as such a structure
would be expected to contain alternating covalent and dative-covalent P−Cu bonds
with non-uniform bond lengths. The Cu centres are 3-coordinate trigonal planar
(P−Cu−P 119.7◦av., P−Cu−C 119.9
◦
av.), while the P centres are, as expected,
approximately tetrahedral, causing the rings to pucker (Cu−P−Cu 95.4◦av.) and
so adopt a cyclohexane-like chair shape. The cyano groups at the core of the
97
4. PHOSPHIDOCOPPER AND PHOSPHIDOCUPRATE
COMPOUNDS
dimer displayed in Figure 4.3 are not cleaved from copper during the synthesis as
would be expected, but instead remain σ-bonded to the metal (C(1/2)−Cu(1/2)
1.919(9)/1.945(9) A˚) in a near-linear fashion (Cu(1/2)−C(1/2)−N(1/2) 173.3(9)/
172.0(9) A˚). The three interactions between cyano groups and Cu centres be-
longing to a single node give rise to a basket motif (Figure 4.4, one such trigo-
nal node (−Cu(2)−P(2)−Cu(2B)−P(2B)−Cu(2A)−P(2A)−) is clearly displayed
at the foot of the image).
Figure 4.4: Structure of the trigonal symmetry (PCu)3 node (composed of Cu(2),
P(2) and symmetry equivalents) in the extended network formed by 10 in the
solid-state, H and C atoms (except for CN and C-ipso) omitted for clarity
In the supramolecular array, each of the (PCu)3 nodes is linked to three others via
the (NLi)2 metallocyclic linkers. As mentioned, these linker rings are formed from
the interactions between units of a solvated Li centre bonded to a cyanide group
(CNLi · 2THF). They can therefore be viewed as being created by the dimerisa-
tion of two crystallographically independent monomers of Ph2PCu(CN) · 2THF,
with the 4-membered (NLi)2 ring forming alternating long and short N−Li bonds
(2.17 and 1.98 A˚av. respectively) unlike the P−Cu bonds in the (PCu)3 node which
are regular. The geometry across the C−−N−Li units from each monomer is close
to linear (C(1/2)−N(1/2)−Li(1/2) 176.3(11)/177.0(11)◦) and each cyano group
lies approximately trans to the shorter of the N−Li bonds. Turning attention
to the longer N−Li interactions, the concomitantly short distances between the
cyano carbon centres and alkali metals (C(2/1)· · ·Li(1/2) 2.69(3)/2.74(3) A˚) sug-
gest that the Li centres are pi-stabilised by the cyano groups. These structural
98
4. PHOSPHIDOCOPPER AND PHOSPHIDOCUPRATE
COMPOUNDS
observations are consistent with the (NLi)2 rings at the centre of the bis(amido)
Lipshutz-type cuprate [(TMP)2Cu(CN)Li2·THF]2, 1,
15 although interestingly while
the metallocycle in 1 is essentially rectangular (N−Li−N 90.67(18)◦, Li−N−Li
89.33(18)◦), in 10 the metallocycle is best described as a diamond (N−Li−N
102.1◦av., Li−N−Li 75.9
◦
av.). The diamond motif is not unique in cuprates in-
corporating RCuCNLi moieties (R = organo) although in most cases the differ-
ence between the two angles is not as pronounced as in 10 (selected examples
are listed in Table 4.2). One possible explanation for the pronounced diamond
shape is that the formation of the extended network with (PCu)3 hexamers at ei-
ther end of the (CNLi)2 linker requires the N(1)−N(2) interatomic distance to be
lengthened, effectively “pinching” the Li−N−Li bonds and distorting the (NLi)2
ring from a rectangular shape. Finally, the cyano groups are σ-bonded to the
respective Cu centres (C(1/2)−Cu(1/2) 1.919(9)/1.945(9) A˚) with near-linear ge-
ometry (Cu(1/2)−C(1/2)−N(1/2) 173.3(9)/172.0(9)◦) such as to join the (PCu)3
and (NLi)2 metallocycles.
N−Li−N/Li−N−Li (◦)
[{C6H3-2, 6-Trip2}Cu(CN)Li · 2THF]2 97.5(5)/82.5(5)
173
[{(SiMe2Ph)3C}Cu(CN)Li · 2THF]2 97.1(4)/82.9(4)
174
[{(SiMe3)3C}Cu(CN)Li · 2THF]2 (average) 96.3/83.4
175
[{(SiMe3)2(SiMe2NMe2)C}Cu(CN)}Li · 2THF]2 (average) 95.8/83.9
175
Table 4.2: Examples of lithium cyanocuprates containing N2Li2 diamond motifs
and related N−Li−N and Li−N−Li bond angles173–175
The cyclohexane-like shape of the (PCu)3 nodes means that the (CNLi · 2THF)
linkers are oriented at approximately 120◦ to each other. This causes the poly-
mer to aggregate into an extended network of six-fold symmetric hexamers of
Ph2PCu(CN)Li · 2THF trimers as displayed in Figure 4.5. After the extended
network has formed, it is likely that the product gets trapped in a thermodynamic
sink. Even though there is a second equivalent of LiPPh2 present in the reaction
mixture, since the product is trapped in a thermodynamic sink, LiPPh2 is unable to
react with 10 and will instead be solvated by THF and stabilised.176 The stability of
10 is also evidenced by its very high melting point (decomposition observed above
210 ◦C), which is particularly high due to the lack of β-hydrogens in the phosphine
ligand, preventing decomposition via β-hydrogen elimination pathways.133
A search of the CSD revealed that 10 is the first crystallographically recorded
species containing a R2PCu(CN)Li motif. Davies et al. have synthesised a related
compound incorporating neutral PPh3 ligands, [(PPh3)4Cu2(CN)4Li4 · 10THF]
2+
99
4. PHOSPHIDOCOPPER AND PHOSPHIDOCUPRATE
COMPOUNDS
Figure 4.5: Structure of the extended network formed by 10 in the solid-state, H
atoms omitted for clarity
100
4. PHOSPHIDOCOPPER AND PHOSPHIDOCUPRATE
COMPOUNDS
[Ph2Cu]
–
2 (Figure 4.6), although since the Cu centres are four-coordinate and the
PPh3 groups act as donor ligands it is difficult to draw direct comparisons.
177
Cu2N
Li
N
Li
C CuCCu
C
PPh3
PPh3
N
Li
THF
THF
THF
Ph3P
C
Ph3P
N
Li
THF
THF THF
2+
Figure 4.6: Schematic representation of the structure of [(PPh3)4Cu2(CN)4Li4 ·
10THF]2+ [Ph2Cu]
–
2, displaying four-coordinate Cu centres in the doubly-charged
cation177
N(1)−Li(1) 2.00(2) N(1)−Li(1)−N(2) 101.8(14)
N(2)−Li(1) 2.17(3) N(1)−Li(2)−N(2) 102.4(15)
N(1)−Li(2) 2.17(3) Li(1)−N(1)−Li(2) 75.2(11)
N(2)−Li(2) 1.96(2) C(1)−N(1)−Li(1) 176.3(11)
O(1)−Li(1) 1.89(3) C(1)−N(1)−Li(2) 106.6(11)
O(2)−Li(1) 1.99(2) Li(1)−N(2)−Li(2) 76.6(11)
O(3)−Li(2) 1.96(3) C(2)−N(2)−Li(2) 177.0(13)
O(4)−Li(2) 1.92(3) C(2)−N(2)−Li(1) 106.0(10)
C(2)−−−Li(1) 2.69(3) Cu(1)−C(1)−N(1) 173.3(9)
C(1)−−−Li(2) 2.74(3) Cu(2)−C(2)−N(2) 172.0(9)
P(1)−Cu(1) 2.260(2) P(1)−Cu(1)−C(1) 120.4(3)
C(1)−Cu(1) 1.919(9) P(2)−Cu(2)−C(2) 119.4(2)
P(2)−Cu(2) 2.264(2) P(1)−Cu(1)−P(1A) 119.69(10)
C(2)−Cu(2) 1.945(9) P(2)−Cu(2)−P(2A) 119.74(10)
C(1)−N(1) 1.157(11) Cu(1)−P(1)−Cu(1A) 95.31(8)
C(2)−N(2) 1.131(10) Cu(2)−P(2)−Cu(2A) 95.40(8)
Table 4.3: Selected bond lengths, A˚, and angles, ◦, for 10
4.4 Lithium bis(dicyclohexylphosphido)cuprate,
(Cy2P)2CuLi · 2THF, 11
One possible explanation for why attempts to form lithium bis(phosphido)cuprates
incorporating diphenylphosphido ligands were unsuccessful is that the ligand does
101
4. PHOSPHIDOCOPPER AND PHOSPHIDOCUPRATE
COMPOUNDS
not confer a high enough steric demand. For this reason, an alternative, more ster-
ically bulky ligand was sought and as dicyclohexylphosphine is readily available
it was an ideal candidate. The addition of LiPCy2 to a slurry of CuCN in bulk
THF yielded a pale yellow solution which was subsequently stored at –27 ◦C. The
cooled solution yielded X-ray grade crystals, which were expected to be of either a
monomeric or dimeric Gilman-type phosphido cuprate in line with known structures
of Gilman-type amidocuprates (see Section 1.8.2.3). However, X-ray crystallogra-
phy showed that while the species indeed had a Gilman-type cuprate formulation,
it had aggregated to give a polymer in the solid-state, [(Cy2P)2CuLi·2THF]∞, 11∞
(Figure 4.7). The 31P NMR spectrum of 11 displayed one dominant, broad peak
at δ –14.2 ppm and consistent with the solution behaviour of 9 and 10, a small
doublet at δ –26.6 ppm attributable to the starting material, HPCy2. The suggests
that high moisture-sensitivity is a general feature of phosphidocuprate species and
is not limited those incorporating Ph2P
– ligands.
Figure 4.7: Structure of the asymmetric unit of [(Cy2P)2CuLi · 2THF]∞, 11∞, H
atoms omitted for clarity
Figure 4.7 displays the asymmetric unit of 11∞, which consists of two crystallo-
graphically independent monomers of [(Cy2P)2CuLi ·2THF]. The overall structure
can be viewed as a chain of (R2P)2Cu moieties bridged by Li · 2THF units. As
observed in all other Gilman-type cuprates, the structure displays a near-linear
geometry across each Cu centre (P(1/3)−Cu(1/2)−P(2/4) 176.11(3)/179.35(3)◦).
The P−Cu bond lengths (2.230 A˚av.) are consistent within the structure (range <
0.01 A˚), with the compounds described above (2.287 A˚av. (9), 2.262 A˚av.(10)), and
with the species discussed in Section 1.8.2.4 (2.256 A˚av. (7),
140 2.217(2) A˚ (8)129).
As expected, the P−Cu bond lengths are significantly longer than the N−Cu bonds
102
4. PHOSPHIDOCOPPER AND PHOSPHIDOCUPRATE
COMPOUNDS
in cuprates containing amido ligands (1.914 A˚av.
∗), which can easily be rationalised
due to the larger valence orbitals of phosphorus compared with nitrogen. These
data are mirrored by the P−Li bonds (2.587 A˚av.) which are highly uniform (range
= 0.01 A˚), similar in length to that in known Gilman-type phosphidocuprate spe-
cies MeCu(tBu2P)Li · 3THF (2.54(1) A˚) and much longer than N−Li bonds in
analogous amidocuprate structures (2.030 A˚av.
†).
The lithium centres in 11 are both four-coordinate, which is unusual for lithium
cuprates with aggregation states higher than one. In bis(amido)cuprates the coor-
dination number of the lithium can be as low as two, for example in [(TMP)2CuLi]2,
2,138 and is not normally higher than three as is observed for Lipshutz-type bis-
(amido)cuprates such as [(TMP)2Cu(CN)Li2 ·THF]2, 1.
15 A Cy2P
– ligand confers
a much lower steric demand than a TMP– ligand, which can be demonstrated by
measuring the “angle of steric obstruction” between a metal centre and hydrogens
on an adjoining ligand. This measurement can be compared to the Tolman cone
angle, originally devised to measure the size of neutral phosphine ligands.178 The
largest angle between hydrogens on the cyclohexyl rings bound to P(4) and Li(2)
in 11 is 112.22◦ which compares to the analogous largest H-Li-H “cone” in 2 of
135.17◦. This data combined with the relatively long P−Li bonds and chain ar-
rangement of the polymer results in a large area around each lithium centre for
THF molecules to bind and fill the coordination sphere.
The polymeric chain structure of 11 is able to form by virtue of the obtuse
inter-metal Cu−P−Li bridges (108.28(10), 110.53(10), 114.74(10) and 118.35(10)◦).
These cover a large range (>10◦) and add a significant degree of irregularity to
the asymmetric unit. This is in contrast to known bis(amido)cuprates where the
Cu−N−Li angles are all very similar and relatively close to 90◦.138 The P−Li−P
bridges which link [(Cy2P)2Cu]
– units together are, however, much more regu-
lar (P(1/2)−Li(2/1)−P(4/3) 125.85(18)/124.95(18)◦). The orientations and steric
demands of the THF molecules and cyclohexyl rings provide the most plausible
explanation for these structural features. From Figure 4.7 it is clear that the
THF molecules bound to Li(2) are oriented towards one of the cyclohexyl rings
belonging to P(3). This gives rise to the largest of the four Cu−P−Li angles
(Cu(2)−P(4)−Li(2) 118.35(10)◦), which prevents any steric clashing between the
cyclohexyl and THF rings. The THF molecules bound to Li(1) are oriented perpen-
dicular to those on Li(2) allowing the adjacent inter-metal bridges to be more acute
∗ Taken from a CSD search sampling 10 compounds. The range was 0.092 A˚
† Taken from the same CSD search as above. The range was 0.239 A˚
103
4. PHOSPHIDOCOPPER AND PHOSPHIDOCUPRATE
COMPOUNDS
(Cu(1/2)−P(2/3)−Li(1) 110.53(10)/114.74◦) without creating any steric clashes.
The solvation of 11 is an interesting point of note. While the use of bulk Lewis
basic solvent leads to the isolation of either monomeric Gilman-type organo(amido)-
cuprates or unsolvated dimeric Gilman-type bis(amido)cuprates, in the case of 11
the same conditions led to a solvated, aggregated Gilman-type species.
The polymeric structure of 11 represents an unusual example of a Gilman-type
cuprate. In the case of heteroleptic organo(amido)cuprates, theoretical stud-
ies have suggested a preference for head-to-tail dimerisation179,180 which has re-
cently been confirmed experimentally by, for example, the dimeric structure of
MesCu[N(CH2Ph)2]Li.
128 Similar dimers have also been observed in the case
of homoleptic bis(organo)cuprates, for example [Mes2CuLi]2
133 (Figure 1.36),
and bis(amido)cuprates such as 2138 (Figure 1.38(a)). Research into phosphi-
docuprates is still in a relatively early stage and there is only one reported species,
[(tBu2P)2Cu]
– [Li · 2THF]+, 7 (Figure 1.41), which is similar to 11.140 This struc-
ture also displays linear geometry across the Cu centre and has comparable P−Cu
bond lengths (P(1/2)−Cu(1) 2.266(4)/2.246(5) A˚). However, as implied by the
formula and as described in Section 1.8.2.4, this cuprate is best viewed as an ion-
separated species due to the large P−Li separations (2.835 A˚av.) when compared
with those of 11 (2.59 A˚av.). In addition, the synthesis of [(
tBu2P)2Cu]
– [Li·2THF]+
initially required the use of tBu2PSiMe3 in place of
tBu2PLi. This was reported to
be necessary in order to render the phosphine active towards the CuI reagent. The
reaction proceeded via desilylation of the phosphine to form tBu2PCu to which
tBu2PLi could then be added to form the product. In contrast, the formation of
11 does not demand such an involved, multi-step synthesis. Instead, it shows that
following the simple methodologies employed to form amidocuprates, 2:1 mixtures
of R2PLi and CuX offer a potentially general route to lithium phosphidocuprates.
104
4. PHOSPHIDOCOPPER AND PHOSPHIDOCUPRATE
COMPOUNDS
P(1)−Cu(1) 2.2308(7) P(1)−Cu(1)−P(2) 176.11(3)
P(2)−Cu(1) 2.2327(7) P(3)−Cu(2)−P(4) 179.35(3)
P(3)−Cu(2) 2.2286(7) Cu(1)−P(1)−Li(2) 108.28(10)
P(4)−Cu(2) 2.2284(7) Cu(1)−P(2)−Li(1) 110.53(10)
P(1)−Li(2) 2.594(5) Cu(2)−P(3)−Li(1) 114.74(10)
P(2)−Li(1) 2.591(4) Cu(2)−P(4)−Li(2) 118.35(10)
P(3)−Li(1) 2.581(4) P(1)−Li(2)−P(4) 125.85(18)
P(4)−Li(2) 2.583(5) P(2)−Li(1)−P(3) 124.95(18)
O(1)−Li(1) 2.011(5)
O(2)−Li(1) 2.008(5)
O(3)−Li(2) 2.013(5)
O(4)−Li(2) 1.996(5)
Table 4.4: Selected bond lengths, A˚, and angles, ◦, for 11
4.5 Solution-state Studies of Phosphidocopper
and Phosphidocuprate Species
The solution-state behaviours of all three compounds 9, 10, and 11 were found
to be very similar. {1H}31P NMR spectroscopic studies on samples of crystalline
material in C6D6 revealed all three to have a dominant low-field signal and a weak
high-field singlet peak (δ –22.1 and –32.6 ppm (9), δ –14.9 and –38.3 ppm (10),
δ –14.2 and –26.6 ppm (11)). Minor signals were also observed in the spectra of
10 (δ –21.6, –28.1 ppm) and 11 (δ –16.4, –19.7 ppm). Proton-coupled 31P NMR
spectra were also recorded, and while the majority of signals remained unchanged
(δ –22.1 ppm (9); δ –14.9, –21.6 and –28.1 ppm (10); δ –14.2, –21.6 and –28.1 ppm
(11)), each of the weak high-field signals split into doublets (δ –26.6 ppm, J = 200
Hz (11), δ –38.3 ppm, J = 233 Hz (10), δ –32.6 ppm, J = 290 Hz, (9)). These data
compare with 31P NMR reference spectra recorded for HPCy2 and HPPh2 (δ –27.4
ppm, 1JPH = 193 Hz and δ –39.9 ppm,
1JPH = 216 Hz, respectively) indicating
that reformation of the starting material had occurred.
1H NMR spectroscopy in each case revealed the presence of cyclohexyl or phenyl
groups and THF in the expected ratios (3:2 (9), 1:1 (10), 2:1 (11)) and a weak
signal attributable to PH (δ 5.72 ppm (9), δ 5.49 ppm (10), δ 3.09 ppm (11)),
confirming that the reformation of starting material had only occurred in trace
quantities. The reformation of starting materials is most likely to be caused by
trace quantities of H2O present in the deuterated solvents and is due to the high
moisture-sensitivity of each of the compounds.
105
4. PHOSPHIDOCOPPER AND PHOSPHIDOCUPRATE
COMPOUNDS
To probe whether 10 or 11 will undergo interconversion between different aggrega-
tion states, variable temperature {1H}31P NMR spectroscopic studies were carried
out in d8-toluene. In both cases, at all temperatures only the major signals de-
scribed above were observed (δ –14.9 and –38.3 ppm (10), δ –14.2 and –26.6 ppm
(11)) with the low-field signal of the two remaining essentially unchanged apart
from a small migration (δ –13.3 ppm at +40 ◦C to δ –14.3 ppm at –20 ◦C (10), δ
–12.1 ppm at +60 ◦C to δ –15.8 ppm at –20 ◦C (11)) in each case.
4.6 Directed ortho Cupration Abilities of Lith-
ium bis(dicyclohexylphosphido)cuprate,
(Cy2P)2CuLi · 2THF, 11
As phosphidocuprate compounds have a higher thermal stability than their ami-
docuprate counterparts, they are potentially effective DoCu agents. When amido-
cuprates are employed in DoCu reactions, Lipshutz-type species are required, which
act as sources of the active monomeric Gilman-type species. Experiments carried
out here set out to discover whether the Gilman-type phosphidocuprate 11 could
act as such a source. Two strategies were employed: firstly a solution of 11 was
prepared in situ and secondly crystals of 11 were isolated prior to reaction with
the substrate, benzonitrile. The in situ preparation of 11 yielded the expected
product, 2-iodobenzonitrile, only in trace amounts and while use of pre-isolated 11
resulted in a slightly higher yield of 9 %, characterisation of the product by NMR
spectroscopy revealed that all the impurities had not been removed. While it is
possible that 11 remains highly aggregated in solution when redissolved, the results
suggest that this particular phosphidocuprate is not an effective DoCu agent.
106
Chapter 5
Investigating the Reactivities and
Structures of THF-solvated
Amidocuprates
5.1 Introduction
To gain insight into the deprotonative ability of “’ate” complexes towards hete-
rocyclic aromatics,11,39,136 Mongin and coworkers carried out a series of reactions
with halopyridines, comparing lithium cuprate bases to lithium zincate bases. 2-
chloro- and 2-bromopyridine were selected as suitable substrates and were treated
with in situ preparations of alkyl, amido and alkylamido zincate bases, and a
bis(amido)cuprate base. The zincate bases were synthesised from a reaction be-
tween ZnCl2 ·TMEDA and either
nBuLi (3 equivalents) or LiTMP (3 equivalents),
or a combination of the two (3 equivalents of lithium in total). To synthesise the
bis(amido)cuprate base, firstly CuCl2 · TMEDA was reduced with
nBuLi followed
by the addition of 2 equivalents of LiTMP. For both preparations, after the addition
of the substrate, the mixture was allowed to stir for 2 hours before introduction of
the electrophile. The results of these reactions are summarised in Table 5.1.
107
5
.
IN
V
E
S
T
IG
A
T
IN
G
T
H
E
R
E
A
C
T
IV
IT
IE
S
A
N
D
S
T
R
U
C
T
U
R
E
S
O
F
T
H
F
-S
O
L
V
A
T
E
D
A
M
ID
O
C
U
P
R
A
T
E
S
N X N X N X N X N X
E
E
E E
1) ZnCl2·TMEDA (n eq.) + RLi (2n eq.) + R'Li (n eq.)
2) I2
1) CuCl2·TMEDA (n eq.) + nBuLi (n eq.) + LiTMP (2n eq.)
2) PhC(O)Cl or 4-Cl-C6H4C(O)Cl
:
:
X = Br
X = Cl
24a
25a
24b
25b
24c
25c
24d
25d
24e
25e
Substrate (X) M, n (eq.) R, R’ Products (E), yields (%)
a 24a (Br) Zn, 0.5 TMP, TMP 24b1 (I), 36i 24c (I), 16i 24d (I), 22i 24e (I), 8i
b Zn, 0.3 TMP, Bu 24b1 (I), 42i 24c (I), 13i 24d (I), 35i 24e (I), 1i
c Zn, 0.7 TMP, Bu 24b1 (I), 65i 24c (I), 32i 24d (I), tracei 24e (I), -
d Zn, 1 TMP, Bu 24b1 (I), 37i 24c (I), 15i 24d (I), 34i 24e (I), 7i
e Zn, 0.7 Bu, TMP 24b1 (I), 20i 24c (I), 18i 24d (I), - 24e (I), -
f Zn, 1 Bu, TMP 24b1 (I), 34i 24c (I), 23i 24d (I), - 24e (I), -
giii Zn, 1 Bu, Bu 24b1 (I), - 24c (I), - 24d (I), - 24e (I), -
h Cu, 1 - 24b2 (C(O)Ph), 20ii
i 25a (Cl) Zn, 0.5 TMP, TMP 25b1 (I), 43i 25c (I), 9i 25d (I), 30ii 25e (I), 15ii
j Zn, 0.7 TMP, Bu 25b1 (I), 63i 25c (I), 11i 25d (I), 4ii 25e (I), -
k Zn, 1 TMP, Bu 25b1 (I), 41i 25c (I), 6i 25d (I), 10ii 25e (I), 3ii
l Zn, 0.7 Bu, TMP 25b1 (I), 58i 25c (I), 19i 25d (I), - 25e (I), -
m Zn, 1 Bu, TMP 25b1 (I), 66i 25c (I), 17i 25d (I), - 25e (I), -
n Cu, 1 - 25b2 (C(O)Ph), 83ii
o Cu, 1 - 25b3 (C(O)C6H4-4-Cl), 79
ii
Table 5.1: Deprotonative metalation of 2-halopyridines 24a and 25a followed by electrophilic trapping. iYields estimated by 1H
NMR spectroscopy for inseparable mixtures, ii Yields after isolation by column chromatography, iii2-iodopyridine was isolated in
72 % yield181
108
5. INVESTIGATING THE REACTIVITIES AND STRUCTURES OF
THF-SOLVATED AMIDOCUPRATES
The results showed that the lithium zincate bases display similar reactivity to-
wards the chloro- and bromopyridine, forming a mixture of products in almost
all cases (some including disubstituted pyridines), although substitution at the 3-
position (adjacent to the halogen) was recorded in the highest yield. The only
exception was the sterically non-demanding tributylzincate base, which did not
deprotonate at any position, but instead carried out a halogen-exchange reaction
and formed 2-iodopyridine, which was isolated in a 72 % yield. While the lithium
cuprate base displayed poor reactivity with 2-bromopyridine due to the sensitivity
of the metalated species under reaction conditions, it was still completely regiose-
lective. When 2-chloropyridine was employed as the substrate, high yields of the
3-substituted products were recorded (79 – 83 %).
Turning to 2,6-dihalopyridines, again significant differences were noted between the
reactivities of the zincate and cuprate reaction mixtures. Employing (TMP)3ZnLi
when 2,6-dibromopyridine was the substrate led to substitution at the 4-position
in high yields (≥85 %) under most conditions, although when 2,6-dichloropyridine
was the substrate, substitution at the 4-position occurred in a much lower yield
and a disubstituted product, 3,4-diiodo-2,6-dichloropyridine, made up a significant
proportion of the product mixture. Due to the sensitivity of cuprates towards
bromopyridines, the bis(TMP)cuprate base was only tested for ortho-deprotonation
ability with 2,6-chloropyridine. Subsequent to trapping with benzoyl chloride and
purification, the product mixture was found to contain both 3- and 4-benzoyl-
2,6-chloropyridine in yields of 41 and 24 %, respectively. While the cuprate base
showed a much lower regioselectivity than when 2-chloropyridine was employed as
the substrate, its preference was to metalate at the 3-position while the zincate
displayed a strong preference for deprotonation at the 4-position. Moving to 2,3-
dichloropyridine, (TMP)2CuLi once again displayed a preference for deprotonation
adjacent to the halogen and this time displayed complete regioselectivity, leading
to the isolation of 4-benzoyl-2,3-dichloropyridine in a 70 % yield.181
These data suggested that the lithium cuprate formed in situ was acting as a
monometallic base, for example akin to LDA182 or PhLi (in the presence of a cat-
alytic amount of diisopropylamine),183 both of which were found to be capable
of lithiating 2-chloropyridine at the 3-position. However, while these monometal-
lic lithium reagents require low temperatures to work effectively, the bimetallic
lithium cuprate bases can be employed under ambient conditions, which would
provide a significant cost saving when carrying out ortho deprotonations on an
industrial scale. It was postulated that the lithium cuprate’s “monometallic base-
109
5. INVESTIGATING THE REACTIVITIES AND STRUCTURES OF
THF-SOLVATED AMIDOCUPRATES
like” reactivity with halopyridines could be used in the synthesis of azafluorenones,
derivatives of which have been shown to have antifungal,184 antibacterial185 and an-
timalarial186 properties. Starting from ketone 25b2, an intramolecular cyclisation
condensing HCl is required to form 4-azafluorenone 5H -indeno[1,2-b]pyridin-5-one.
After optimisation of the conditions it was found that 4-azafluorenone could be
formed in an 87 % yield in the presence of catalytic amounts of Pd(OAc)2 and
Cy3P · HBF4
187 (Scheme 5.1). This represents a significant improvement from ex-
isting azafluorenone syntheses which require harsh conditions, for example heating
2-phenylnicotinic acid at 190 ◦C in phosphoric acid.188
N Cl
O
N
O
Pd(OAc)2 (5 mol %)
Cy3P·HBF4 (10 mol %)
K2CO3 (2 eq.)
DMF, 130 °C, 48 h
Scheme 5.1: Optimal conditions for the conversion of 3-benzoyl-2-chloropyridine
(25b2) to 5H -indeno[1,2-b]pyridin-5-one187
As data on the reactivity of lithium cuprates prepared from copper chloride (vide
supra) and copper iodide (Table 1.2) had been collected, studies were then extended
to compare these copper sources with copper bromide. While it would be of interest
to compare all the halides, copper(I) fluoride is unstable with respect to oxidation
due to the high electronegativity of F,189 and as such would be unsuitable as a
copper source for these investigations. Research undertaken by Mongin’s group
compared lithium cuprate bases prepared in situ from either CuCl or CuBr in
an otherwise identical manner by reacting them with 2-chloropyridine, followed
by trapping with a benzoylic electrophile (Table 5.2). The results showed that
while the reactivity of the two halide species is comparable, higher yields can be
obtained more reproducibly with the chloride species.
110
5. INVESTIGATING THE REACTIVITIES AND STRUCTURES OF
THF-SOLVATED AMIDOCUPRATES
N Cl N Cl
COAr1) CuX + LiTMP (2 eq.)
TMEDA (1 eq.)
THF, RT, 2 h 2) ArC(O)Cl
25a 25b
ClCOAr CuHal Yieldi (%)
a PhC(O)Cl CuCl 78–90
b CuBr 60–70
c
4-OMe-C6H4C(O)Cl
CuCl 60–80
d CuBr 49
e 4-Cl-C6H4C(O)Cl
CuCl 56–80
f CuBr 55–61
Table 5.2: DoCu of 2-chloropyridine using in situ prepared chloride- or bromide-
based amidocuprate bases followed by benzoylation. iYield after purification by
column chromatography, the remaining amount was starting material
5.2 Reactivity Studies of Chloride-Based Lith-
ium Cuprates
In order to probe the reactivity of lithium cuprate bases further a series of reactions
involving 2-chloropyridine as the substrate were carried out in collaboration with
Mongin’s research group. Various reaction conditions were investigated, including
varying the solvent conditions and using CuCl instead of CuCl2 as the copper
source to reduce the amount of LiCl present in the reaction mixture (Table 5.3).
Initially, employing CuCl as the copper source, reactions were carried out in bulk
THF (entries a and b), as it is capable of dissolving LiCl. This would be expected
to result in the formation of a cuprate with a Lipshutz-type stoichiometry (e.g.
(TMP)2Cu(Cl)Li2) in solution which would be able to act as a source of the re-
active monomeric Gilman-type cuprate (see Section 1.8.2.3). The reaction with
2-chloropyridine, after quenching, afforded 3-benzoyl-2-chloropyridine (25b2) in
high yields (>85 %) whether or not TMEDA was present in the reaction mixture.
Next, to compare these results to a situation where LiCl is more likely to precip-
itate out of the reaction mixture, the solvent was changed to toluene. Starting
with CuCl2 · TMEDA and successively adding one equivalent of
nBuLi and two
equivalents of LiTMP to form the cuprate base, the reaction with 2-chloropyridine
afforded 25b2 in a yield of 61 % (entry c). However, starting from CuCl2 ·TMEDA
without introducing a second equivalent of TMEDA afforded 25b2 in a yield of
111
5. INVESTIGATING THE REACTIVITIES AND STRUCTURES OF
THF-SOLVATED AMIDOCUPRATES
Solvent
Cu
source
n eq. TMEDA,
n’ eq. nBuLi
TMEDA
(eq.) present
LiCl (eq.)
present
Yield
(%)i
a THF CuCl 1, 0 1 1 90
b CuCl 0, 0 0 1 86
c toluene CuCl2
ii 0, 1 1 2 60
d CuCl2 0, 1 0 2 19
e toluene CuCl 1, 0 1 iii 37
f CuCl 0, 0 0 iii 37
g CuCl 1, 0 1 1 61
h CuCl 0, 0 0 1 50
iiv THF CuCl 0, 0 0 0 25
jiv toluene CuCl 0, 0 0 0 15
Table 5.3: Deprotonative metalation of 2-chloropyridine (25a) by TMP-based cu-
prates formed from copper chlorides in toluene or THF followed by benzoylation.
iYields after isolation by column chromatography, iichelate CuCl2 · TMEDA was
employed as the starting material, iiiinsoluble LiCl removed by filtration, ivreactions
performed using a base prepared from CuCl in toluene containing 1 eq. of THF,
and isolated by filtration
only 19 % (entry d), meaning that the amount of TMEDA present under these
conditions had a significant bearing on the effectiveness of the cuprate base. Given
this, it is likely that the coordinating TMEDA allowed a higher proportion of a
cuprate base incorporating LiCl, for example (TMP)2Cu(Cl)Li2 ·2TMEDA (Figure
5.1(a)), to form in solution. In the absence of TMEDA it is likely that LiCl precipi-
tated out of the solution, causing a higher proportion of an unreactive Gilman-type
species, e.g. [(TMP)2CuLi]2, 2, (Figure 5.1(b)), to be present in the reaction mix-
ture. The dimerisation is likely to occur since, in the absence of a coordinating
solvent, the electropositive lithium ions will look to fill their coordination spheres
as far as possible, which can be achieved by bonding to nitrogen atoms in the
TMP ligands. As Gilman-type complexes have been shown to be ineffective DoCu
agents both experimentally15 and theoretically,138 it provides an explanation for
the low reactivity observed here.
Next, attention moved to investigating what bearing the presence of LiCl in the
reaction mixture had on effectiveness of the cuprate. By removing LiCl it should
prevent a reactive Lipshutz-type formulation cuprate forming and this would be
expected to have the effect of significantly decreasing the yield of 25b2. Two
strategies were employed in an attempt to create a reaction mixture in which only
the Gilman-type species would be present. Firstly, starting from CuCl and carry-
ing out the synthesis in a non-polar solvent, insoluble LiCl was removed from the
112
5. INVESTIGATING THE REACTIVITIES AND STRUCTURES OF
THF-SOLVATED AMIDOCUPRATES
TMP Cu TMP
Li Li
Cl
N
NN
N
(a) (TMP)
2
Cu(Cl)Li
2
· 2TMEDA
TMP Cu TMP
Li Li
TMP TMPCu
(b) [(TMP)
2
CuLi]
2
Figure 5.1: Possible dominant species of cuprate base reaction mixtures prepared
in situ from CuCl2 ·TMEDA and HTMP when (a) a further equivalent of TMEDA
was added to the mixture and (b) no extra TMEDA was added
reaction mixture by filtration before 2-chloropyridine was added (entries e and f).
This afforded 25b2 in a yield of 37 % whether or not TMEDA was subsequently
added. To compare this to a control, the reactions were repeated without remov-
ing insoluble LiCl, affording 25b2 in a yield of 61 % when TMEDA was present
(entry g) and 50 % when no TMEDA was added (entry h). Thus, the presence of
TMEDA has a positive effect on the metalating ability of the cuprate when LiCl
is present, but does not make a difference when LiCl has been removed from the
reaction mixture, as would be expected. Taking the two situations where either
TMEDA is present (entries e and g) or not (entries f and h), as the yields within
these pairs of reactions do not match it is clear the presence of LiCl does have a
bearing on the reactivity of the cuprate base. This indicates that a Lipshutz-type
formulation base is required for efficient DoCu activity, consistent with previous
observations. Secondly, as removal of LiCl by filtration may not be 100 % efficient,
it was necessary to validate our ideas by crystallising, isolating and redissolving
the Gilman-type cuprate, [(TMP)2CuLi]2, 2, before reaction with the substrate
(entries i and j). To continue the investigations into solvent choice, reactions were
carried out in both THF and toluene, affording 25b2 in yields of 25 and 15 %,
respectively. As the yield for entry j is lower than that for entry f it suggests that
even though the majority of the LiCl is filtered off, a small amount remains in
solution and thus promotes reaction. As the yield of entry i is higher than that for
entry j, it suggests that in THF the Gilman-type dimer is more capable of acting as
a source of a reactive monomeric Gilman-type species, which is likely to be because
it will deaggregate more readily due to the presence of a donor solvent.
113
5. INVESTIGATING THE REACTIVITIES AND STRUCTURES OF
THF-SOLVATED AMIDOCUPRATES
5.3 Dilithium chloro[di(2,2,6,6-tetramethylpipe-
ridido)]cuprate, [(TMP)2Cu(Cl)Li2 · THF]2,
12
As the chloride-based lithium cuprate base prepared in bulk THF had proven to be
highly reactive and regioselective in the deprotocupration of 2-chloropyridine, in-
sight into the structure of a chloride-based Lipshutz-type cuprate was of great inter-
est. Successful synthetic strategies that had previously led to the characterisation
of Lipshutz-type cuprates [(TMP)2Cu(CN)Li2 ·THF]2, 1,
15 and [(TMP)2Cu(I)Li2 ·
THF]2, 3,
138 were useful references and crystallisation of [(TMP)2Cu(Cl)Li2·THF]2,
12 was achieved by treating CuCl with a preformed solution of LiTMP in toluene in
the presence of two equivalents of THF, followed by storage of the resulting filtrate
at –27 ◦C. 1H NMR spectroscopic analysis of a redissolved sample of 12 showed the
presence of signals attributable to both TMP and THF in the expected 2:1 ratio.
X-ray grade crystals of 12 were obtained from the solution and crystallographic
analysis showed the structure to be a dimeric Lipshutz-type cuprate (Figure 5.2),
analogous to the structures found for 1 and 3.
The structure can be viewed as resulting from the aggregation of two six-membered
(NCuNLiClLi) rings, containing TMP bridges between Li and Cu centres on either
side, joined via a central (LiCl)2 metallocyclic core. The Cu−N bond lengths in 12
are consistent with those observed in both Lipshutz- and Gilman-type bis(TMP)-
cuprates 1, 3 and [(TMP)2CuLi]2, 2, which have been previously characterised
(Table 5.4).
12 2 1 3
1.908(2) 1.926(2) 1.9258(17) 1.920(2)
1.918(2) 1.924(2) 1.9200(17) 1.925(2)
1.928(2)
1.922(2)
Table 5.4: Comparison of Cu−N bond lengths, A˚, in the bis(amido)cuprates 12,
2, 1, and 3
These bond lengths are slightly longer than the average Cu−N bond length of
1.872 A˚ seen for other bis(amido)cuprates: [{(MePh2Si)2N}2Cu][Li · 4THF],
190
(NPh2)2Cu(NPh2)Li2 · 2OEt2 (Figure 1.35(a)),
131 [(HPhN)Cu(NHMes)Li ·
DME]2,
131 [(Ph2N)2Cu][K · 3 (Phen)]
191 and [(Ph2N)2Cu][Na · 3 (Phen)].
192∗ The
∗ A search of the CSD found the 5 compounds listed. The range was 0.047 A˚
114
5. INVESTIGATING THE REACTIVITIES AND STRUCTURES OF
THF-SOLVATED AMIDOCUPRATES
Figure 5.2: Structure of [(TMP)2Cu(Cl)Li2 ·THF]2, 12, H atoms omitted for clarity
steric demand of the TMP groups is the likely cause for the long Cu−N bond lengths
in the bis(TMP)cuprates 12, 2, 1, and 3. The significantly unequal Li−N−Cu an-
gles (Li(1/2)−N(1/2)−Cu(1) 92.1(2)/84.77(19)◦), although both close to an ideal
of 90◦, create asymmetry in the six-membered rings. The geometry across the Cu
centre is approximately linear (N(1)−Cu(1)−N(2) 174.45(11)◦) and although the
deviation from 180◦ is larger than has previously been observed in bis(TMP)cuprate
compounds (lowest N−Cu−N angle 176.94(10)◦ (2)), it is not large enough to have
a significant impact on the overall structure. The Li−Cl bonds in the six-membered
ring are, of course, much shorter than the Li−I bonds in 3, and although more dif-
ficult to draw direct comparisons to 1, the Li−Cl bonds are longer than the Li−N
bond within the 7-membered ring in 1 (2.168(5) A˚), which can be explained the
strong hard-hard lithium-nitrogen interactions in the latter species. The short
Li−Cl contacts are a possible explanation for the relatively small N−Cu−N bond
angle, as the relatively large deviation from 180◦ allows the N centres to form more
effective inter-metal bridges.
115
5. INVESTIGATING THE REACTIVITIES AND STRUCTURES OF
THF-SOLVATED AMIDOCUPRATES
Cl(1)−Li(1) 2.344(6) N(1)−Cu(1)−N(2) 174.45(11)
Cl(1)−Li(2) 2.371(6) Cl(1)−Li(1)−Cl(1A) 98.7(2)
Cl(1)−Li(1A) 2.328(6) Cl(1)−Li(1)−N(1) 128.5(3)
Cu(1)−N(1) 1.908(2) Cl(1)−Li(2)−N(2) 128.2(3)
Cu(1)−N(2) 1.918(2) Li(1)−N(1)−Cu(1) 92.1(2)
N(1)−Li(1) 1.966(6) Li(2)−N(2)−Cu(1) 84.77(19)
N(2)−Li(2) 2.016(6)
Table 5.5: Selected bond lengths, A˚, and angles, ◦, for 12
5.4 Dilithium bromo[di(2,2,6,6-tetramethylpipe-
ridido)]cuprate, [(TMP)2Cu(Br)Li2 · THF]2,
13
As discussed in Section 5.1, cuprate base mixtures prepared from CuBr and LiTMP
are reactive in the deprotocupration of 2-chloropyridine. However, as the yields
were not as consistently high as when chloride-base cuprate bases were employed,
insights into the structure of a bromide-based cuprate were of interest in order to at-
tempt to explain the differences in reactivity. The 2:1 reaction of LiTMP and CuBr
in a predominantly hydrocarbon medium containing 2 equivalents of THF deposited
crystalline material after storage at –27 ◦C. 1H NMR spectroscopy displayed the
presence of several signals attributable to the TMP ligands as well as two signals (δ
3.61 and 1.37 ppm) attributable to THF. Integration of the peaks gave a ratio of 2:1
TMP:THF which strongly suggested that a mono-solvated Lipshutz-type cuprate
had crystallised. X-ray diffraction data confirmed that [(TMP)2Cu(Cl)Li2 ·THF]2,
13 (Figure 5.3) had been isolated from the reaction mixture.
The crystallographic data shows that the structure of 13 is very similar to that
of 12. It also dimerises in the solid-state, joined by a central (Li2Br2) ring with
one THF molecule coordinating to each of the other two Li centres (Li(1) and
Li(3)). Again, linear geometry is observed across the N−Cu−N moiety, and there is
also significant asymmetry across the six-membered ring (Li(1/2)−N(1/2)−Cu(1)
85.45(13)/92.81◦). The Li−Br bond lengths in the six-membered ring are larger
than Li−Cl bond lengths in 12 as would be expected for a larger halide (Table
5.6), which is mirrored in the N−Cu−N bond angle being closer to the ideal 180◦
(175.07(8)◦ (13), 174.45(11) (12)), meaning that less strain is present in the struc-
ture. This data may explain why 12 is more reactive in DoCu reactions as the
increased ring strain compared to 13 could make it more favourable to release
116
5. INVESTIGATING THE REACTIVITIES AND STRUCTURES OF
THF-SOLVATED AMIDOCUPRATES
Figure 5.3: Structure of [(TMP)2Cu(Br)Li2 ·THF]2, 13, H atoms omitted for clarity
LiHal and form the Gilman-type monomer.
12 13
Li(1)−Hal(1) 2.344(6) 2.545(4)
Li(2)−Hal(1) 2.371(6) 2.485(4)
Li(2A/1A)−Hal(1) 2.328(6) 2.488(4)
Table 5.6: Comparison of Li−Hal bond lengths, A˚, in 12 and 13 (Hal = Cl (12),
Br (13))
Comparing the four Lipshutz-type cuprates 1, 3, 12, and 13, the solid-state data
suggests that there is a link between the nature of the central X group (X = CN,
Hal) and the planarity of the central cores. In [(TMP)2Cu(CN)Li2 · THF]2, 1, the
small, highly charge-dense cyano-N centres form strong hard-hard interactions with
the Li centres in the (LiN)2 metallocyclic ring, with the result that the remaining
two THF-solvated Li centres (Li(2) and Li(2A)) lie close to the plane defined by the
aforementioned ring (within ±0.14 A˚, Figure 5.4(a)). On replacement of the CN
with a halide, the LiX interactions (X = N, halide) in the central metallocycles be-
117
5. INVESTIGATING THE REACTIVITIES AND STRUCTURES OF
THF-SOLVATED AMIDOCUPRATES
come weaker, due to the softer nature of the halides when compared with nitrogen.
Descending Group 17 the halides become larger and softer, with the result that the
deviations of the THF-solvated lithium centres from the (LiHal)2 rings increase
(±0.49 A˚, Hal = Cl (12), Figure 5.4(b); ±0.80 A˚, Hal = Br (13), Figure 5.4(c)).
Although the analogous deviation from the (LiI)2 ring in [(TMP)2Cu(I)Li2 ·THF]2,
3, (±0.08 A˚, Figure 5.4(d)) is not consistent with the trend described here, the
crystallographic refinements must be taken into account. For 3 it was necessary
to model the I centres split over two positions (each with 1/2 occupancy) which
suggests that the halide is so large and soft that the central core is unable to ac-
commodate it sufficiently with the result that either the I atoms move between two
positions or the structure takes up one of two metastable conformations.
(a) 1 (b) 12 (c) 13 (d) 3
Figure 5.4: Side-on projections of THF-solvated Lipshutz-type cuprates 1, 12, 13,
and 3 showing displacement of unsolvated Li centre from central metallocyclic core
(Li2X2) which is only very minor for 1 (X = N) but increasingly large for 12 and 13
(X = Cl, Br respectively). 3 (X = I) appears to shown only a minor deviation
although the requirement to model positional disorder in the crystallographic data
must be taken into account. H and THF-C atoms omitted for clarity
Br(1)−Li(1) 2.545(4) N(1)-Cu(1)−N(2) 175.07(8)
Br(1)−Li(2) 2.485(4) Br(1)−Li(2)−Br(1A) 99.46(14)
Br(1)−Li(2A) 2.488(4) Br(1)−Li(1)−N(1) 127.29(19)
Cu(1)−N(1) 1.9190(17) Br(1)−Li(2)−N(2) 130.4(2)
Cu(1)−N(2) 1.9114(17) Li(1)−N(1)−Cu(1) 85.45(13)
N(1)−Li(1) 2.002(5) Li(2)−N(2)−Cu(1) 92.81(14)
N(2)−Li(2) 1.963(4)
Table 5.7: Selected bond lengths, A˚, and angles, ◦, for 13
118
5. INVESTIGATING THE REACTIVITIES AND STRUCTURES OF
THF-SOLVATED AMIDOCUPRATES
The structures of 12 and 13 show that the behaviour of bis(TMP)cuprates in
the solid-state is consistent regardless of the nature of X (X = CN, Hal) as they
all form Lipshutz-type structures. Although minor differences are noted in the
relative DoCu abilities, consideration of availability and cost of CuX reagents or
by-products formed (for example the presence of cyano-containing compounds) may
far outweigh any detriment to the overall yield. CuCN and CuCl are the cheapest
CuI salts to source∗ although as reactions involving CuCN require the removal
of any potentially harmful cyano-based by-products, cuprate mixtures synthesised
from CuCl represent the most commercially viable reagents.
5.5 Solution Studies of THF-solvated Lipshutz-
type Cuprates
Elemental analysis of bulk samples of [(TMP)2Cu(Cl)Li2 · THF]2, 12, and
[(TMP)2Cu(Br)Li2 · THF]2, 13, proved the purity of the isolated crystals, and
in particular, the percentage values for the halogen elements (Cl 7.92 % (12); Br
14.97 % (13)) were very close to the expected values (Cl 7.62 % (12); Br 15.67 %
(13)). This indicates that the the Gilman-type cuprate [(TMP)2CuLi]2, 2 was not
co-crystallising with the Lipshutz-type cuprates in either case, as this would have
the effect of lowering the percentage values for the halide elements. In solution,
however, a large number of signals were observed in the 13C NMR spectra for both
12 and 13 which could not all be assigned to environments in the Lipshutz-type
cuprates. On comparison with NMR spectra recorded from samples of 2138 it was
clear that a Gilman-type species was also present, indicating that a small amount
of LiHal (Hal = Cl, Br) was being eliminated on dissolution of the compounds.
Evidence of the extent of the LiCl elimination from 12 was provided by the 7Li
NMR spectroscopic data which displayed two signals, one at δ 1.27 ppm, which
can be attributed to Gilman-type 2 (see Section 3.3.1), and one at δ 1.91 ppm
attributable to 12. Integrating the area under the peaks gave a ratio of 6:1 12:2,
indicating that the extent of LiCl elimination was limited and so the Lipshutz-type
structure was almost fully retained when 12 was dissolved in C6D6. Only one sig-
nal (at δ 1.34 ppm) was present in the 7Li NMR spectrum of 13 which suggests
signals produced by the Gilman-type and Lipshutz-type species occur at the same
∗ The cheapest available CuCN and CuCl costs £0.04 g−1. For comparison, the cheapest
available CuBr costs £0.07 g−1, and the cheapest available CuI costs £0.15 g−1. Data from
http://www.alfa.com, accessed 19th March 2014
119
5. INVESTIGATING THE REACTIVITIES AND STRUCTURES OF
THF-SOLVATED AMIDOCUPRATES
frequency. Some insight can be gained into the extent of LiBr elimination, how-
ever, from the 13C NMR spectrum as the signals attributable to 2 displayed a lower
intensity than the signals attributable to 13, suggesting that the LiBr elimination
was limited, as was the case for 12.
In order investigate the interconversion between the Gilman- and Lipshutz-type
species in solution further, low temperature (0 ◦C) and room temperature (+25
◦C) NMR spectra of 12 in d8-THF were recorded. Only very minor differences
were noted in the 1H and 13C NMR spectra at either temperature, and as the 7Li
NMR spectra only displayed one signal at both room and low temperatures (which
migrated from δ 0.70 ppm at +25 ◦C to δ 0.22 ppm at 0 ◦C) it was not possible
to draw any firm conclusions about the solution behaviour of 12.
120
Chapter 6
Investigating Solvent Effects in
the Formation of Amidocuprates
6.1 Introduction
The results presented in Chapter 5 have added to the growing body of research
on the solid-state structures, syntheses and reactivities of THF-solvated Lipshutz-
type cuprates (introduced in Section 1.8.2.2). However, the influence of solvent
on the structures of lithium cuprates is, as yet, largely unexplored. To this end,
an aprotic, Lewis basic solvent was required and diethyl ether (Et2O) was chosen
as it not only meets this criterion but is also readily available. Et2O and THF
are almost identical in structure; THF can be viewed as Et2O fused into a five-
membered ring. However, this small structural difference does have a significant
impact on their Lewis basicity, with the result that THF is a stronger donor than
Et2O.
∗ In the synthesis of lithium cuprates, if the reaction is carried out in bulk
THF, or the product is recrystallised from bulk THF, the solvent abstracts LiX
(X = CN, Hal) from the mixture and a Gilman-type cuprate can be isolated. If,
however, stoichiometric amounts (2 equivalents) of THF are added to the reaction
mixture, the donor solvent is unable to effectively stabilise LiX but is able to
support LiX units incorporated into the cuprate. In this case a Lipshutz-type
species is the favoured product. As Et2O is a weak donor it is not able to stabilise
LiX as effectively as THF meaning that LiX is more likely to precipitate out of
solution. If stoichiometric amounts of Et2O are employed this may mean that the
donor solvent is unable to effectively stabilise a Lipshutz-type cuprate, although if
∗ Dipole moments: 1.75 ±0.04 D (THF), 1.098 ±0.001 D (Et
2
O)193
121
6. INVESTIGATING SOLVENT EFFECTS IN THE FORMATION
OF AMIDOCUPRATES
Et2O is present in bulk quantities it is likely to act in a similar manner to THF by
abstracting LiX from the mixture and resulting in the formation of a Gilman-type
cuprate. To test these hypotheses, a series of reactions was carried out following
known preparations of bis(TMP)cuprates described in the literature138 and in this
thesis, replacing THF with Et2O (Scheme 6.1).
NH
4 nBuLi, –78 °C
NLi
toluene
2 CuX, –78 °C – toluene
+ hexane/Et2O
X = Cl, Br, I, CN
NH 4 nBuLi, –78 °C NLi
toluene + 2eq. Et2O
2 CuX, –78 °C
(b)
(a)
TMP Cu TMP
Li
TMP Cu TMP
Li
TMP Cu TMP
Li Li
XEt2O
2
4
4
4
4
Scheme 6.1: General reaction pathways followed to investigate the effect of em-
ploying Et2O in place of THF in the syntheses of cuprate bases
6.2 Dilithium cyano[di(2,2,6,6-tetramethylpiper-
idido)]cuprate, [(TMP)2Cu(CN)Li2 · OEt2]2,
14
Following pathway (b) in Scheme 6.1 was expected to result in the formation of
Gilman-type cuprate [(TMP)2CuLi]2, 2, as had previously been observed when
utilising THF as the donor solvent.138 Employing CuCN as the copper source
resulted in the formation of crystalline material after storage of the solution at
–27 ◦C. In the expectation of confirming the formation of 2, NMR spectroscopic
experiments in C6D6 were attempted. However, the isolated material was only
slightly soluble in the deuterated solvent and the resulting sample was too dilute
to record interpretable NMR spectra from. This observation was surprising as
crystalline material of 2 isolated previously from the different preparation (using
THF) was highly soluble in C6D6 and indicated that the expected product had not
formed. Subsequently, a second NMR spectroscopic sample was prepared in a more
polar deuterated solvent, d8-THF, allowing NMR spectroscopic data to be recorded.
The 1H NMR spectrum contained signals that could be assigned to the TMP ligand
protons and also two signals (at δ 3.42 and 1.15 ppm) which could be attributed to
Et2O. The
7Li NMR spectrum revealed two distinct lithium environments (at δ –
122
6. INVESTIGATING SOLVENT EFFECTS IN THE FORMATION
OF AMIDOCUPRATES
0.73 and –2.28 ppm). As only one 7Li signal would be expected to be observed for 2,
and coupled with the Et2O signals in the
1H NMR spectrum, it was difficult to draw
firm conclusions, although the data was consistent with the formation of a higher-
order cuprate. X-ray crystallographic analysis confirmed that 2 had not formed in
the reaction, and revealed that a novel Lipshutz-type cuprate, [(TMP)2Cu(CN)Li2 ·
OEt2]2, 14, had been synthesised (Figure 6.1). Unfortunately, even after repeated
attempts, only low quality X-ray data (Rint >10 %) could be collected. However,
the connectivity could be determined unambiguously, showing the structure to be
a dimer analogous to the THF-solvated species [(TMP)2Cu(CN)Li2 · THF]2, 1.
15
Figure 6.1: Structure of [(TMP)2Cu(CN)Li2 · OEt2]2, 14, H atoms omitted for
clarity
Although the poor quality of the data means that it is not possible to draw
firm conclusions about the exact structural parameters of 14, the bond lengths
and angles appear to theclosely match those found in 1, with a N−Cu−N
angle slightly deviating from a linear geometry and an approximately planar
square N2Li2 ring bridging between the two monomers. This suggests that
123
6. INVESTIGATING SOLVENT EFFECTS IN THE FORMATION
OF AMIDOCUPRATES
the replacement of THF with Et2O opens up the possibility of using Et2O
to carry out Lipshutz-type chemistry, for example in DoCu reactions. Di-
ethyl ether-solvated ion-associated lithium cuprates have been previously re-
ported in the literature, for example the dimeric Gilman-type species [Ph2CuLi ·
OEt2]2
194 and [({Me3Si}CH2)2CuLi · OEt2][({Me3Si}CH2)2CuLi · 2OEt2],
112 the
higher-order (PhC−−C)9Cu3Li6 · 3OEt2 which contains two linked structural
units (PhC−−C)3Cu3(Li · OEt2)3 and (PhC−−C)6Cu3Li3,
195 and the Lipshutz-type
(NPh2)2Cu(NPh2)Li2 · 2OEt2
131 (Figure 1.35(a)). 14, however, represents the first
diethyl ether-solvated Lipshutz-type cyanocuprate to be observed.
N(3)−Li(1) 1.991(15) N(1)−Cu(1)−N(2) 178.9(2)
N(3A)−Li(1) 2.134(14) N(3)−Li(1)−N(3A) 92.4(6)
Cu(1)−N(1) 1.926(5) Li(1)−N(3)−Li(1A) 87.6(6)
Cu(1)−N(2) 1.909(5) N(3A)−Li(1)−N(2) 137.7(7)
N(1)−Li(2) 2.014(18) C(5A)−Li(2)−N(1) 122.6(7)
N(2)−Li(1) 2.009(13) Li(1)−N(2)−Cu(1) 94.6(5)
C(5)−Li(2A) 2.089(17) Li(2)−N(1)−Cu(1) 89.7(4)
C(5)−−−Li(1A) 2.572(15)
C(5)−N(3) 1.133(11)
Table 6.1: Selected bond lengths, A˚, and angles, ◦, for 14
124
6. INVESTIGATING SOLVENT EFFECTS IN THE FORMATION
OF AMIDOCUPRATES
6.3 Dilithium iodo[di(2,2,6,6-tetramethylpiperi-
dido)]cuprate, [(TMP)2Cu(I)Li2· OEt2]2, 15
The formation of Lipshutz-type 14 could be due to one of two variables or a com-
bination of both: either the solvent employed (Et2O) or the nature of the X group
(CN). To discover if the result reported above was limited to cyano-based species or
not, CuI was treated with 0.5 equivalents of LiTMP in toluene and after filtration
the product was recrystallised from a mixture of hexane and diethyl ether (Scheme
6.1, pathway (b)). This synthesis is known to result in the formation of Gilman-
type cuprate 2 when THF is the donor solvent.138 1H NMR spectroscopic data
revealed the presence of Et2O in a sample of redissolved crystalline material (sig-
nals at δ 3.45 and 1.23 ppm), suggesting that a Lipshutz-type cuprate had formed
once again. In addition, unlike 2 which in 7Li NMR spectroscopic experiments
produces a single signal (at δ 1.28 ppm), the crystals obtained here produced two
distinct lithium environments in a 1:1 ratio (at δ 0.95 and 0.09 ppm). X-ray grade
crystals were obtained which provided a high quality dataset, and on refinement,
the data confirmed that a Lipshutz-type cuprate, [(TMP)2Cu(I)Li2 · OEt2]2, 15,
had formed. (Figure 6.2).
Comparing the solid-state structures of 15 and [(TMP)2Cu(I)Li2 · THF]2, 3,
138 it
is clear that a change in the coordinating solvent from THF to Et2O has only very
minor effects. In both cases, the central Li2I2 metallocycle has a degree of instability
due to the size of the halide anion. This is evidenced by the crystallographic data,
where a model in which each iodine is equally split over two distinct sites provides
the best fit for the data. The solvated Li centres in 15 are only three-coordinate, as
would be expected due to the larger steric demand of Et2O compared to THF. The
difference in the Lewis basic character of the solvents is noticeable in the Li−O
bond lengths (1.943 A˚av. (3), 1.982 A˚av. (15)), although it is only slight. The
strength of the Li−O interactions does not have an effect on the bond lengths in
the core of compound, however, which to within the calculated error, are essentially
identical for both compounds. This is best exemplified by the Li−I and Cu−N
bond lengths, which are summarised in Table 6.2.
125
6. INVESTIGATING SOLVENT EFFECTS IN THE FORMATION
OF AMIDOCUPRATES
Figure 6.2: Structure of [(TMP)2Cu(I)Li2 ·OEt2]2, 15, H atoms and minor iodide
disorder omitted for clarity
126
6. INVESTIGATING SOLVENT EFFECTS IN THE FORMATION
OF AMIDOCUPRATES
3 15
Li−I (intramonomer) 2.734(6) 2.730(4)
2.763(6) 2.743(4)
2.825(7) 2.866(5)
2.689(7) 2.715(5)
Li−I (intermonomer) 2.798(6) 2.747(4)
2.805(7) 2.776(5)
Li−I average 2.769 2.763
Cu−N 1.920(2) 1.9191(15)
1.925(2) 1.9193(14)
1.928(2) 1.9264(15)
1.922(2) 1.9272(15)
Cu−N average 1.924 1.923
Table 6.2: Comparison of the Li−I and Cu−N bond lengths, A˚, in 3 and 15,
including both Li−I bonds within individual six-membered rings (intramonomer)
and between the rings (intermonomer)
The bond angles, however, reveal more significant differences between the two
structures. The geometry across the Cu centres is far closer to an ideal linear
arrangement in 3 (179.0◦av.) than in 15 (176.9
◦
av.), and while in 3 the Li−N−Cu
inter-metal bridges are all above 93◦, the six-membered rings here are far more
symmetrical than is the case for 15 (Table 6.3). There is also difference in the
Li2I2 metallocycles, where the ring in 15 exhibits a puckered “butterfly” confor-
mation whereas the same ring in 3 forms a planar rectangle (Figure 6.3, Table
6.4). The puckered ring in 15 is the most likely cause for the asymmetrical inter-
metal bridges which will alter in order to maintain strong Li−N and Li−I contacts
within each six-membered ring.
3 15
Li(1)−N(1)−Cu(2/1) 94.13(17) 92.26(12)
Li(2)−N(2)−Cu(2/1) 93.11(18) 89.93(11)
Table 6.3: Comparison of the inter-metal bridge (Li−N−Cu) bond angles, ◦, in
3 and 15
3 15
I−Li−I 101.7(2) 96.74(15)
100.4(2) 95.71(14)
Li−I−Li 78.63(17) 80.03(13)
79.3(2) 79.78(13)
Table 6.4: Comparison of the I−Li−I and Li−I−Li bond angles, ◦, in 3 and 15
127
6. INVESTIGATING SOLVENT EFFECTS IN THE FORMATION
OF AMIDOCUPRATES
(a) 3 (b) 15
Figure 6.3: Structures of the cores of 3 and 15 displaying a planar rectangular
Li2I2 ring in 3 and a puckered “butterfly” Li2I2 ring in 15
As discussed in Section 6.1, syntheses of lithium cuprates in solvent systems con-
taining stoichiometric amounts of diethyl ether were also of interest (Scheme 6.1,
pathway (a)). This synthetic pathway is more straightforward as it does not in-
volve the removal and replacement of solvent, and importantly when THF is the
donor solvent, gives a different product (Lipshutz-type cuprates) than when fol-
lowing pathway (b) (Gilman-type cuprates). For this reason CuI was treated with
LiTMP in a solvent mixture containing toluene and 2 equivalents of Et2O and,
following filtration, the solution deposited crystalline material after storage at –27
◦C. The parameters calculated from a crystallographic cell check matched those
for crystals of 15 prepared via pathway (b) in Scheme 6.1. The formation of a
Lipshutz-type cuprate in this case is consistent with the behaviour observed when
stoichiometric amounts of THF were employed in an analogous reaction.138 This
contrasts with the behaviour observed so far of reactions following pathway (b),
establishing that this route is dependent on the nature of the donor solvent; the use
of bulk THF leads to Gilman-type cuprates, whereas employing bulk Et2O results
in the crystallisation of Lipshutz-type cuprates.
128
6. INVESTIGATING SOLVENT EFFECTS IN THE FORMATION
OF AMIDOCUPRATES
I(1)−Li(1) 2.747(4) N(1)−Cu(1)−N(2) 176.72(6)
I(1)−Li(3) 2.730(4) N(3)−Cu(2)−N(4) 177.07(6)
I(1)−Li(4) 2.743(4) I(1)−Li(2)−I(2) 96.74(15)
I(2)−Li(1) 2.715(5) I(1)−Li(3)−I(2) 95.71(14)
I(2)−Li(2) 2.866(5) Li(1)−I(1)−Li(3) 80.03(13)
I(2)−Li(3) 2.776(5) Li(1)−I(2)−Li(3) 79.78(13)
Cu(1)−N(1) 1.9191(15) I(1)−Li(3)−N(3) 130.52(17)
Cu(1)−N(2) 1.9193(14) I(1)−Li(4)−N(4) 123.68(16)
Cu(2)−N(3) 1.9264(15) I(2)−Li(1)−N(1) 134.07(18)
Cu(2)−N(4) 1.9272(15) I(2)−Li(2)−N(2) 125.53(18)
N(1)−Li(1) 1.964(4) Li(1)−N(1)−Cu(1) 92.26(12)
N(2)−Li(2) 1.998(4) Li(2)−N(2)−Cu(1) 89.93(11)
N(3)−Li(3) 1.966(4) Li(3)−N(3)−Cu(2) 93.18(12)
N(4)−Li(4) 1.994(3) Li(4)−N(4)−Cu(2) 89.62(11)
O(1)−Li(2) 1.988(4)
O(1)−Li(4) 1.975(3)
Table 6.5: Selected bond lengths, A˚, and angles, ◦, for 15
6.4 Diethyl Ether-solvated Lipshutz-type Cupra-
tes Containing Chloride and Bromide
Experiments moved to investigate syntheses of Lipshutz-type bis(TMP)cuprates
where CuCl and CuBr were employed as the copper source. Both pathways in
Scheme 6.1 were followed in order to assess whether changing the halide affects
the nature of the compounds formed via either pathway. Given that when THF
was employed as the donor solvent the nature of the halide did not affect the solid-
state structures significantly (see Chapter 5), it was expected that all permutations
described here (reactions starting from either CuCl or CuBr and recrystallised in
either bulk or stoichiometric amounts of Et2O) would lead to the formation of
Lipshutz-type cuprates. In the case of CuCl, crystals of 16 were obtained when
carrying out the recrystallisation from hexane (∼10 mL) and diethyl ether (∼3 mL)
and for CuBr, crystals of 17 were obtained in this manner as well as from a reaction
incorporating 2 equivalents of Et2O (see Scheme 6.1). As 17 was insoluble in C6D6,
NMR spectroscopy was carried out in d8-THF, and the
7Li spectrum contained
only one signal (at δ –1.40 ppm). Fortunately, 16 was soluble in C6D6 although
while the 7Li spectrum revealed two distinct signals (at δ 1.89 and 1.83 ppm),
the close similarity in frequencies meant it was difficult to draw firm conclusions.
1H NMR spectra were much clearer, however, revealing the presence of a quartet
and a triplet attributable to Et2O in each case (δ 3.46 and 1.18 ppm (16, C6D6),
129
6. INVESTIGATING SOLVENT EFFECTS IN THE FORMATION
OF AMIDOCUPRATES
δ 3.42 and 1.14 ppm (17, d8-THF)), suggesting that Lipshutz-type cuprates had
formed. X-ray crystallography confirmed that each set of crystals were of Lipshutz-
type cuprates, [(TMP)2Cu(Cl)Li2 · THF]2, 16, and [(TMP)2Cu(Br)Li2 · THF]2,
17 (Figures 6.4 and 6.5 respectively), which are analogous to the iodide-based
structure, [(TMP)2Cu(I)Li2 · OEt2]2, 15.
Figure 6.4: Structure of [(TMP)2Cu(Cl)Li2 ·OEt2]2, 16, H atoms omitted for clarity
130
6. INVESTIGATING SOLVENT EFFECTS IN THE FORMATION
OF AMIDOCUPRATES
Figure 6.5: Structure of [(TMP)2Cu(Br)Li2·OEt2]2, 17, H atoms omitted for clarity
131
6. INVESTIGATING SOLVENT EFFECTS IN THE FORMATION
OF AMIDOCUPRATES
Comparing the three species 15, 16, and 17, it is clear that the nature of the
halide does not, beside the length of the Li−X bonds in the core, significantly
affect the bond lengths in the structures, but it does have subtle effects on their
three-dimensional shape. This is evidenced firstly by the consistent Cu−N and
Li−N bond lengths across the three structures (Table 6.6).
16 17 15
Cu(1)−N(1) 1.925(3) 1.920(4) 1.9191(15)
Cu(1)−N(2) 1.923(2) 1.914(4) 1.9193(14)
Cu(2)−N(3) 1.927(4) 1.9264(15)
Cu(2)−N(4) 1.926(4) 1.9272(15)
Cu−N (average) 1.924 1.922 1.923
N(1)−Li(1) 2.024(6) 1.976(9) 1.964(4)
N(2)−Li(2) 1.953(7) 2.010(9) 1.998(4)
N(3)−Li(3) 1.972(8) 1.966(4)
N(4)−Li(4) 1.994(9) 1.994(3)
Li−N (average) 1.989 1.988 1.981
Table 6.6: Comparison of Cu−N and Li−N bond lengths, A˚, in 16, 17, and 15,
which are almost identical across the three structures
As expected, the Li−Hal bonds increase as the size of the halide increases (Table
6.7) and this, in turn, has an effect on the N−Cu−N bonds which cap either
end of the structure. For the THF-solvated cuprates (Chapter 5) a decrease in
the N−Cu−N bond angles was observed as the size of the central halide atom
decreased. Even though the geometry across the Cu centre in 15 deviates much
further from the ideal linear arrangement than in the analogous THF-solvated
species, 3 (see Section 6.3), the same trend is mirrored in the series of diethyl ether-
solvated cuprates. The N−Cu−N bond angles in 16 and 17 are then lower than
the same angles in their THF-solvated counterparts [(TMP)2Cu(Cl)Li2 ·THF]2, 12,
and [(TMP)2Cu(Br)Li2 ·THF]2, 13, most noticeable in 16 which has a particularly
small N−Cu−N bond angle (172.42(12)◦).
The asymmetry across the six-membered rings noted for 15 is also observed for
the bromide-based analogue 17, where the inter-metal bridge angles match those
in 15 very closely (Table 6.8). Interestingly, in 16 it appears that the smaller
halide confers a more regular structure as both inter-metal bridges are close to
90◦, although this is at odds with the analogous THF-solvated 12 which contains
a highly asymmetrical six-membered ring (Table 6.8).
As discussed in Section 5.3, where a trend was noted for the THF-solvated Lipshutz-
132
6. INVESTIGATING SOLVENT EFFECTS IN THE FORMATION
OF AMIDOCUPRATES
16 17 15
Li−Hal (intramonomer) 2.344(6) 2.485(9) 2.730(4)
2.332(7) 2.556(8) 2.743(4)
2.498(8) 2.715(5)
2.546(8) 2.866(5)
Li−Hal (intermonomer) 2.330(6) 2.543(8) 2.747(4)
2.498(9) 2.776(5)
Li−Hal (average) 2.335 2.521 2.763
Table 6.7: Comparison of the Li−Hal bond lengths, A˚, in 16, 17, and 15, showing
the expected increase in length with increasing halide size
15 17 16 12
Li(1)−N(1)−Cu(1) 92.26(12) 92.1(3) 90.4(2) 92.1(2)
Li(2)−N(2)−Cu(1) 89.93(11) 89.1(3) 91.8(2) 84.77(19)
Li(3)−N(3)−Cu(2) 93.18(12)
Li(4)−N(4)−Cu(2) 89.62(11)
Table 6.8: Comparison of the inter-metal bridges (Li−N−Cu bond angles), ◦, in
15, 17, 16, and 12, highlighting the similarity between Et2O-solvated 15 and 17,
but difference between the chloride-base species 16 (Et2O-solvated) and 12 (THF-
solvated)
type species where a larger, softer X group (X = CN, Hal) resulted in a larger
deviation of the solvated Li centres from the central Li2X2 core (X = N, halide).
In the case of the diethyl ether-solvated analogues, although an exact comparison
cannot be made with [(TMP)2Cu(CN)Li2·OEt2]2, 14, due to the poor quality of the
X-ray crystallographic data, comparing the other three compounds 15, 16, and 17,
the same trend holds true (Figure 6.6). For 16 and 17, while the X−Li−X and
Li−X−Li angles match each other closely (X−Li−X/Li−X−Li 100.4(2)/79.6(2)◦
(16, X = Cl), 98.9/80.0◦av. (17, X = Br)), there is an increase in the deviation of the
diethyl ether-solvated Li centres (Li(1) and Li(1A) (16); Li(2) and Li(4) (17)) from
the mean plane of the metallocyclic core as the halide becomes larger and softer
(±0.71 A˚ (16); ±0.27 A˚ (17)). The puckered “butterfly” conformation of the Li2I2
ring in 15 means that there is a large deviation of the diethyl ether-solvated Li
centres (Li(2) and Li(4)) from a mean plane of the metallocyclic core (±0.93 A˚).
133
6. INVESTIGATING SOLVENT EFFECTS IN THE FORMATION
OF AMIDOCUPRATES
(a) 16 (b) 17 (c) 15
Figure 6.6: Side-on projections of diethyl ether-solvated Lipshutz-type cuprates
16, 17 and 15 highlighting the trend in deviations from planar structures. H and
diethyl ether carbon atoms omitted for clarity
Cl(1)−Li(1) 2.344(6) N(1)−Cu(1)−N(2) 172.42(12)
Cl(1)−Li(2) 2.332(7) Cl(1)−Li(2)−Cl(1A) 100.4(2)
Cl(1)−Li(2A) 2.330(6) Li(2)−Cl(1)−Li(2A) 79.6(2)
Cu(1)−N(1) 1.925(3) Cl(1)−Li(1)−N(1) 125.2(3)
Cu(1)−N(2) 1.923(2) Cl(1)−Li(2)−N(2) 127.3(3)
N(1)−Li(1) 2.024(6) Li(1)−N(1)−Cu(1) 90.4(2)
N(2)−Li(2) 1.953(7) Li(2)−N(2)−Cu(1) 91.8(2)
Table 6.9: Selected bond lengths, A˚, and angles, ◦, for 16
134
6. INVESTIGATING SOLVENT EFFECTS IN THE FORMATION
OF AMIDOCUPRATES
Br(1)−Li(1) 2.485(9) N(1)−Cu(1)−N(2) 175.34(16)
Br(1)−Li(2) 2.556(8) N(3)−Cu(2)−N(4) 175.91(15)
Br(1)−Li(3) 2.543(8) Br(1)−Li(1)−Br(2) 99.6(3)
Br(2)−Li(1) 2.498(9) Br(1)−Li(3)−Br(2) 98.1(3)
Br(2)−Li(3) 2.498(8) Li(1)−Br(1)−Li(3) 79.7(3)
Br(2)−Li(4) 2.546(8) Li(1)−Br(2)−Li(3) 80.3(3)
Cu(1)−N(1) 1.920(4) Br(1)−Li(1)−N(1) 130.6(4)
Cu(1)−N(2) 1.914(4) Br(1)−Li(2)−N(2) 126.9(4)
Cu(2)−N(3) 1.927(4) Br(2)−Li(3)−N(3) 127.4(4)
Cu(2)−N(4) 1.926(4) Br(2)−Li(4)−N(4) 124.7(4)
N(1)−Li(1) 1.976(9) Li(1)−N(1)−Cu(1) 92.1(3)
N(2)−Li(2) 2.010(9) Li(2)−N(2)−Cu(1) 89.1(3)
N(3)−Li(3) 1.972(8) Li(3)−N(3)−Cu(2) 92.3(3)
N(4)−Li(4) 1.994(9) Li(4)−N(4)−Cu(2) 88.4(3)
Table 6.10: Selected bond lengths, A˚, and angles, ◦, for 17
135
Chapter 7
Pentametallic Adduct-type
Amidocuprates
7.1 Introduction
In an attempt to create alternative, more cost-effective DoCu agents than the
bis(TMP)cuprates [(TMP)2Cu(X)Li2 · THF]2 (X = CN (1), I (3), Cl (12)), inves-
tigations focused on synthesising cuprates from amines cheaper than HTMP. As a
sterically bulky amido ligand has been identified as an important feature in efficient
DoCu agents15 (see Table 1.2), cis-2,6-dimethylpiperidine (HDMP) was an ideal
choice given that it retails for a fraction of the price of HTMP. As discussed in
Chapter 6, employing diethyl ether as the donor solvent in lithium cuprate synthe-
sis provides a versatility not offered by THF. For this reason, reactions with HDMP
were, in the first instance, carried out using Et2O as the donor solvent. Syntheses
were attempted in the presence of stoichiometric amounts (1 or 2 equivalents) of
Et2O and in bulk Et2O (Scheme 7.1).
NLi
2 CuX
–78 °C
4
NH
4 4 nBuLi, –78 °C
Et2O or 
Hex + Et2O (1/2 eq.)
N Cu N
Li LiX
Li X
Et2O
Li
NN Cu
OEt2
Scheme 7.1: Synthetic strategy towards forming DMP-based bis(amido)cuprates
136
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
7.2 Attempted Synthesis of Dilithium chloro[di-
(cis-2,6-dimethylpiperidido)]cuprate,
[(DMP)2Cu(Cl)Li2 · OEt2]2
In an attempt to synthesise a DMP-based Lipshutz-type cuprate, CuCl was treated
with 2 equivalents of LiDMP. The reaction was run in hexane with 2 equivalents
of Et2O, as this was found to be the most straightforward pathway to forming
diethyl ether-solvated Lipshutz-cuprates containing TMP ligands, and led to the
isolation of crystalline material. Elemental analysis revealed the presence of chlo-
rine in a much smaller proportion (4.05 %) than required for the expected prod-
uct, the Lipshutz-type cuprate (DMP)2Cu(Cl)Li2 ·OEt2 (8.62 %). This suggested
that an approximately 1:1 mixture of the expected Lipshutz-type cuprate and a
Gilman-type cuprate, (DMP)2CuLi had co-crystallised in the reaction vessel. This
theory was supported by the proportion of carbon (54.64 %) and nitrogen (7.74 %)
found by elemental analysis, which were both between the values expected for the
aforementioned putative Lipshutz-type (52.55 % (C), 6.81 % (N)) and Gilman-type
(56.33 % (C), 9.73 % (N)) cuprates. X-ray grade crystals were obtained from the
reaction mixture and a sample was examined under a microscope in the expecta-
tion of finding two types of crystal which could then be separated prior to X-ray
diffraction studies. However, the sample of crystals were indistinguishable from one
other and so a single crystal of suitable size and quality was selected to undergo
X-ray crystallographic analysis. After initial refinement of the data it was clear
that the structure was neither a conventional Lipshutz- nor Gilman-type cuprate.
However, the data did suggest that there was a mixture of two species present with
very similar structural parameters. These species contain either a Cu or Li centre
at a particular atomic site, but are otherwise identical (Figure 7.1).
After refinement of the occupancies, the model showed that the majority of
molecules contained a Li centre at the disordered site (Li(2A)). In this case, the
structure could be viewed as a Lipshutz-type monomer (incorporating the six-
membered N(1)−Cu(1)−N(2)−Li(2)−Cl(1)−Li(1) ring) coordinated by a lithium
amide (LiDMP) dimer. The LiDMP dimer exhibits an interesting structural motif
as it forms a chain rather than a ring as would usually be expected.196 Given
this, the structure can be viewed as formally resulting from the insertion of a
Lipshutz-type monomer into an Li−N bond of the cyclic amide dimer. The re-
maining molecules, which contain Cu at the disordered site (Cu(2)), can be viewed
137
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
Figure 7.1: Structure of [(DMP)2CuLi · Et2O]1.29 · [(DMP)2Li2 · Et2O]0.71LiCl, 18,
H atoms and minor Et2O disorder omitted for clarity
138
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
as an adduct between Lipshutz- and Gilman-type monomers, which represents
an unprecedented lithium cuprate structural motif. The adduct is approximately
symmetrical across the Li(2)−Cl(1) bond and contains only 3-coordinate lithium
centres. Focussing on the six-membered ring containing Cu(1), this Lipshutz-type
monomer is structurally similar to those in [(TMP)2Cu(Cl)Li2 · THF]2, 12, and
[(TMP)2Cu(Cl)Li2 · THF]2, 16, as it has comparable Cu−N bond lengths (Ta-
ble 7.1) and the N−Cu−N bond angle deviates significantly from the expected
linear geometry.
12 16 18
N(1)−Cu(1)−N(2) 174.45(11) 172.42(12) 171.4(2)
Cu(1)−N(1) 1.908(2) 1.925(3) 1.886(5)
Cu(1)−N(2) 1.918(2) 1.923(2) 1.922(5)
Table 7.1: Comparison of N−Cu−N bond angles, ◦, and Cu−N bond lengths, A˚,
in 12 and 16 and the Lipshutz-type monomer incorporated into 18 highlighting
the similarity of all three structures
However, while in 12 and 16 all of the Li−Cl bonds in the six-membered rings are
very similar in length (Li(1/2)−Cl(1) 2.344(6)/2.371(6) A˚ (12), 2.344(6)/2.332(7)
A˚ (16)), in 18, the Li(1)−Cl(1) bond length is similar (2.323(11) A˚) but the
Li(2)−Cl(1) bond is relatively extended (2.556(15) A˚). Due to the metal disor-
der it is difficult to draw any firm conclusions about the structure of the second
six-membered ring, but interestingly the metal centre does not sit centrally be-
tween N(3) and N(4) (M−N(3/4) 1.876(6)/2.031(7) A˚) and the Li−N−M angles
are both significantly greater than 90◦ (Li(2/3)−N(3/4)−M 99.9(5)/97.5(5)◦). This
creates a large separation between Li(2) and Li(3) (4.328 A˚) when compared with
the Li(2)· · ·Li(1) separation (3.974 A˚), both of which are large compared to the
Li(1)· · ·Li(2) separations in the Lipshutz-type species 12 (3.612 A˚) and 16 (3.771
A˚). These data provide a likely explanation for the long Li(2)−Cl(1) bond as
these atoms are forced to weaken their interaction with each other in order to
form bonds to N(3) and Li(3) respectively. Overall, these weak interactions mean
18 is very unstable, which is evidenced by its low melting point (decomposition
observed above 87 ◦C under gentle heating).
While the Lipshutz-type monomer coordinated by a LiDMP dimer represents an
interesting structure, it is essentially a Lipshutz-type cuprate and as it contains
a 4:1 ratio of Li:Cu, the synthetic pathway employed is a very inefficient route
to its formation. The adduct between the Lipshutz- and Gilman-type monomers,
139
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
however, mirrors the stoichiometry of the reactants added (with loss of 1 equivalent
of LiCl) and represents a remarkable lithium cuprate structure, of a type previously
unreported in the literature. Hence this new class of cuprate, herein referred to
as “adduct-type”, is of significant interest. Efforts from this point focussed on
crystallising a pure form of an adduct-type species containing no metal disorder.
Cl(1)−Li(1) 2.323(11) N(1)−Cu(1)−N(2) 171.4(2)
Cl(1)−Li(2) 2.556(15) N(3)−Cu(2)/Li(2A)−N(4) 172.4(3)
Cl(1)−Li(3) 2.313(12) N(2)−Li(2)−N(3) 141.1(8)
Cu(1)−N(1) 1.886(5) Li(1)−Cl(1)−Li(2) 109.0(3)
Cu(1)−N(2) 1.922(5) Li(3)−Cl(1)−Li(2) 125.4(4)
Cu(2)/Li(2A)−N(3) 1.876(6) Cl(1)−Li(1)−N(1) 121.6(5)
Cu(2)/Li(2A)−N(4) 2.031(7) Cl(1)−Li(3)−N(4) 112.7(6)
N(1)−Li(1) 2.033(11) Li(1)−N(1)−Cu(1) 89.4(4)
N(2)−Li(2) 2.003(11) Li(2)−N(2)−Cu(1) 95.6(4)
N(3)−Li(2) 1.969(11) Li(2)−N(3)−Cu(2)/Li(2A) 99.9(5)
N(4)−Li(3) 1.936(14) Li(3)−N(4)−Cu(2)/Li(2A) 97.5(5)
Table 7.2: Selected bond lengths, A˚, and angles, ◦, for 18
7.3 Adduct-type Cuprate Structures
7.3.1 Trilithium chloro[tetra(cis-2,6,-dimethylpiperidido)]-
cuprate, [(DMP)
2
CuLi · OEt
2
]
2
LiCl, 19
The first attempt to synthesise a pure adduct-type cuprate involved the addition
of LiDMP (2 equivalents) to CuCl in bulk Et2O Scheme 7.2) and the resulting
solution yielded crystalline material after storage at –27 ◦C. The material was
analysed spectroscopically, and the 1H NMR spectrum revealed peaks attributable
to the DMP ligands and Et2O in a 2:1 ratio, consistent with both Lipshutz-type
and adduct-type structures.
4 HDMP
4 nBuLi, –78 °C
Et2O
4 DMPLi
2 CuCl
–78 °C
[(DMP)2CuLi·OEt2]2LiCl + LiCl
Scheme 7.2: Synthetic strategy towards forming a pure adduct-type cuprate
The 7Li NMR spectrum contained four signals (at δ 2.16, 1.83, 1.48 and –0.51 ppm)
which, even though one could be attributed to LiDMP (δ 2.16 ppm), were more
than the two signals that might be expected to be produced by a Lipshutz-type
140
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
cuprate. The combination of NMR spectroscopic data suggested that the synthesis
had been successful, and X-ray crystallography confirmed the formation of the
adduct-type cuprate [(DMP)2CuLi · OEt2]2LiCl, 19 (Figure 7.2(a)). The solid-
state structure of 19 contains a triangularly coordinated Cl atom and exhibits
approximate C 2 symmetry around a central Li(2)−Cl(1) bond, which is shared
between two six-membered Lipshutz-type rings.
Figure 7.2: Structure of [(DMP)2CuLi ·OEt2]2LiCl, 19 (H atoms and minor Et2O
disorder omitted for clarity) and the central core viewed along the Li(2)−Cl(1)
bond
Both N−Cu−N units in 19 exhibit approximately linear geometry, and each N cen-
tre bridges between a Cu and a Li centre (Li−N−Cu bond angles 91.5◦av.). How-
ever, there is asymmetry noted across each six-membered ring with the Cu−N−Li
angles distinctly larger where the Li centre is bridging between two amido groups
(Cu(1/2)−N(2/3)−Li(2) 94.1(2)/94.2(2)◦) than where it is bridging between an
amido group and the Cl centre (Cu(1/2)−N(1/4)−Li(1/3) 87.6(2)/90.1◦). Com-
pared to 18 the Li(2)−Cl(1) bond in 19 is significantly shorter (2.556(15) A˚ (18),
2.412(7) A˚ (19)) although it is still relatively extended compared to the other two
Li−Cl bonds (Li(1/3)−Cl(1) 2.354(6)/2.301(7) A˚) and to the average Li−Cl bond
length of 2.365 A˚av.
∗, which suggests that factors other than metal disorder, such
as the strength of the Li−N and Li−Cl bonds must be taken into account when
explaining the disparity in the Li−Cl bond lengths (vide infra).
∗ Taken from a CSD search sampling 50 compounds, the range was 0.339 A˚
141
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
Cl(1)−Li(1) 2.354(6) N(1)−Cu(1)−N(2) 172.42(12)
Cl(1)−Li(2) 2.412(7) N(3)−Cu(2)−N(4) 174.43(14)
Cl(1)−Li(3) 2.301(7) N(2)−Li(2)−N(3) 129.2(4)
Cu(1)−N(1) 1.905(3) Li(1)−Cl(1)−Li(2) 109.1(2)
Cu(1)−N(2) 1.902(3) Li(3)−Cl(1)−Li(2) 111.3(2)
Cu(2)−N(3) 1.891(3) Cl(1)−Li(1)−N(1) 121.9(3)
Cu(2)−N(4) 1.914(3) Cl(1)−Li(3)−N(4) 124.5(4)
N(1)−Li(1) 1.986(8) Li(1)−N(1)−Cu(1) 87.6(2)
N(2)−Li(2) 2.035(7) Li(2)−N(2)−Cu(1) 94.1(2)
N(3)−Li(2) 2.054(6) Li(2)−N(3)−Cu(2) 94.2(2)
N(4)−Li(3) 1.970(9) Li(3)−N(4)−Cu(2) 90.1(2)
Table 7.3: Selected bond lengths, A˚, and angles, ◦, for 19
7.3.2 Adduct-type cuprates containing bromide and iodide
Examples of planar triangulated LiHal structural motifs in the CSD are limited to
a single species where Hal = F, and three ion-separated species where Hal = Cl
containing four-coordinate Li centres that fill their coordination spheres via solva-
tion.197,198 However, such a motif has not previously been observed in cuprates. For
this reason, investigations next sought to discover if adduct-type cuprate species
could be formed incorporating larger halides, and CuBr and CuI were the obvious
choices as copper sources. Initial reactions with LiDMP and CuBr in a solvent
system containing 2 equivalents of Et2O were successful in forming adduct-type
cuprates of the general formula [(DMP)2CuLi ·OEt2]x[(DMP)2CuLi ·HDMP]yLiBr
(where x + y = 2). However, as is made clear by the empirical formula, in each
attempt a small amount of protonated amide (HDMP) was found to be coordinated
to one of the lithium centres in the solid-state, having either remained unreacted
during the synthesis or re-formed prior to crystallisation. The best quality crys-
tallographic dataset gave 20 (Figure 7.3), where HDMP had “replaced” one Et2O
group in 55 % of the molecules (i.e. x = 1.45, y = 0.55).
To overcome this issue two strategies were attempted, both of which were successful.
Firstly a slight excess of nBuLi (1.1 equivalents) was added to the HDMP when
forming the LiDMP solution to ensure all the amine was deprotonated during
the first step. Secondly, the reaction was run in bulk Et2O, which if present in
excess should bind preferentially to the Li centres (as it is a stronger Lewis base
than HDMP). However, the second method also had the effect of significantly
affecting the isolated yield (21 % compared to 47 % from the first method) as
the recrystallisation was carried out in purely donor solvent and the complex was
142
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
Figure 7.3: Structure of [(DMP)2CuLi · OEt2]1.45[(DMP)2CuLi · HDMP]0.55LiBr,
20, H atoms (except NH) and minor Et2O disorder omitted for clarity
143
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
highly soluble in this. Both pathways led to the isolation of only the adduct-type
cuprate [(DMP)2CuLi · OEt2]2LiBr, 21 (Figure 7.4).
Figure 7.4: Structure of [(DMP)2CuLi ·OEt2]2LiBr, 21, H atoms and minor Et2O
disorder omitted for clarity
Changing the copper source to CuI also presented synthetic challenges to forming
an adduct-type lithium cuprate. While the use of bulk Et2O did not result in
the formation of any isolable material, employing 2 equivalents of donor solvent
(a method successfully applied in the formation of 21) resulted only in the crys-
tallisation of LiI · Et2O.
199 Given the high solubility of LiI it is likely that it has a
preference for remaining in solution during the filtration step. As the adduct-type
cuprate structures crystallised so far only contained 1 equivalent of Et2O per Cu
atom, the amount of Et2O was restricted further in an effort to encourage precipi-
tation of LiI prior to the filtration step without affecting the nature of the product.
This strategy afforded crystalline material and the 1H NMR spectrum of a sample
redissolved in C6D6 revealed several signals attributable to DMP ligand protons,
providing evidence that the formation of pure LiI · Et2O had been avoided. The
7Li NMR spectrum of the same sample revealed a signal attributable to LiDMP
(at δ 2.17) alongside two other signals (at δ 1.84 and 1.41 ppm) in a 1:2 ratio
144
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
which would be expected for an adduct-type cuprate. X-ray grade crystals from
the sample were obtained and diffraction experiments confirmed the formation of
an adduct-type cuprate, [(DMP)2CuLi · OEt2]2LiI, 22, the structure of which is
shown in Figure 7.5. Relevant bond lengths and angles of 20, 21, and 22 are
summarised in Tables 7.6 and 7.7.
Figure 7.5: Structure of [(DMP)2CuLi · OEt2]2LiI, 22, H atoms and minor iodide
disorder omitted for clarity
The structures of 20, 21, and 22 are all structurally similar to the chloride-
containing adduct-type cuprate 19; in each case the halide sits centrally in the
molecule bonded to three lithium centres and exhibits trigonal planar geometry,
demonstrating that the adduct-type structure is capable of incorporating halide
ions of differing sizes, much like Lipshutz-type cuprates. The size of the halide
does have an affect on the structure of the cuprate however, which although is
not obvious from the images of the structures, can be seen by comparison of the
bond lengths and angles (Tables 7.3, 7.6 and 7.7). As the size of the halide in-
creases the lithium-halide bond lengths increase as expected (Li−Cl 2.356 A˚av. (19);
Li−Br 2.558 A˚av. (20), 2.527 A˚av. (21); Li−I 2.786 A˚av. (22)) which in turn means
the geometry across the N−Cu−N moiety relaxes and nears the idealised 180◦
(N(1/3)−Cu(1/2)−N(2/4) 173.4◦av. (19), 173.7
◦
av. (20), 174.7
◦
av. (21), 176.5
◦
av.
(22)). As was observed in 19, in each of 20, 21, and 22 the Li(2)−Hal bond
145
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
is relatively extended compared to Li(1/3)−Hal, although the increase in bond
lengths on descending Group VII is not consistent. The ratio between Li(2)−Hal
and the mean of Li(1/3)−Hal is much larger when Hal = I (1.10) than when Hal
= Br (1.07 (20), 1.04 (21)) and Hal = Cl (1.04). This suggests that the bond
extension cannot be attributed only to the increased ionic radius of the anion, but
that the hard/soft nature of the lithium-halide interactions must also be taken into
account. The interaction between lithium and iodide is a weak hard-soft one, and
so the lithium ion exhibits a preference for forming hard-hard interactions with
the anionic N centres (N(2/3)) in the DMP rings. The increase in stabilisation
by the DMP N centres is demonstrated by a decrease in the Li(2)−N(2/3) bond
lengths across the series of adduct-type species 19, 20, 21, and 22 as the central
halide becomes softer (Table 7.4).
19 20 21 22
Li(2)−N(2) 2.035(7) 2.032(9) 2.048(8) 2.022(13)
Li(2)−N(3) 2.054(6) 2.046(9) 2.011(8) 2.000(14)
Table 7.4: Comparison of the Li(2)−N(2/3) bond lengths, A˚, in 19, 20, 21, and
22, which decrease as the halide becomes larger and softer (Hal = Cl (19), Br
(20 and 21), I (22))
To accommodate this preference the Li(2)−N−Cu bond angles widen as the halide
increases in size, and the Li(1/3)−N−Cu bond angles follow the same trend,
although consistent with the shorter Li(1/3)−Hal bonds, they are significantly
smaller than the respective N(2/3) inter-metal bridges (Table 7.5).
19 20 21 22
Li(2)−N(2)−Cu(1) 94.1(2) 96.4(3) 97.3(2) 100.1(5)
Li(2)−N(3)−Cu(2) 94.2(2) 95.5(3) 97.5(3) 101.2(5)
Li(1)−N(1)−Cu(1) 87.6(2) 90.6(3) 92.5(3) 96.1(5)
Li(3)−N(4)−Cu(2) 90.1(2) 92.4(3) 89.0(3) 96.8(5)
Table 7.5: Comparison of Li(2)−N(2/3)−Cu(1/2) and Li(1/3)−N(1/4)−Cu(1/2)
bond angles, ◦, in 19, 20, 21, and 22, highlighting the increase as the size of the
central halide increases
Interestingly, unlike the ratio between the Li(2)−Hal and mean Li(1/3)−Hal
distances, the ratio between the mean Li(2)−N(2/3)−Cu(1/2) and mean
Li(1/3)−N(1/4)−Cu(1/2) inter-metal bridges does not follow the same trend and
is essentially consistent throughout the series of compounds (1.06 (19), 1.05 (20),
1.07 (21), 1.04 (22)). This data also shows that the asymmetry across each six-
146
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
membered ring noted for 19 in Section 7.3.1 is a feature shared by all of the adduct-
type species. While similar to the THF-solvated Lipshutz-type cuprates 12 and 13
(where the difference between Cu−N−Li and Cu−N−Li · S angles is greater than
7◦ in both cases) it contrasts with the diethyl ether-solvated Lipshutz-type species
14, 15, 16, and 17 discussed in Chapter 6 where the analogous difference does not
exceed 3◦. Notably, none of these structural features have any effect on the Cu−N
bonds, which are essentially consistent across all of the adduct-type species (1.903
A˚av. (19), 1.897 A˚av. (20), 1.902 A˚av. (21), 1.915 A˚av. (22)) and consistent with
expected Cu−N bond lengths in amidocuprate compounds (see Section 5.3).
20 21 22
Li(1)-Hal(1/1A) 2.499(9) 2.474(8) 2.720(13)
Li(2)-Hal(1/1A) 2.679(10) 2.592(7) 2.971(16)
Li(3)-Hal(1/1A) 2.496(9) 2.515(8) 2.667(14)
Cu(1)−N(1) 1.895(4) 1.916(3) 1.922(6)
Cu(1)−N(2) 1.894(4) 1.887(3) 1.910(6)
Cu(2)−N(3) 1.896(4) 1.892(3) 1.908(6)
Cu(2)−N(4) 1.901(4) 1.912(3) 1.918(6)
N(1)−Li(1) 1.991(10) 1.955(9) 1.916(13)
N(2)−Li(2) 2.032(9) 2.048(8) 2.022(13)
N(3)−Li(2) 2.046(9) 2.011(8) 2.000(14)
N(4)−Li(3) 2.004(10) 1.995(8) 1.934(14)
Table 7.6: Selected bond lengths, A˚, for 20, 21 and 22. Hal(1) = Br(1) (20 and
21), Hal(1A) = I(1A) (22)
20 21 22
N(1)−Cu(1)−N(2) 174.17(18) 175.37(16) 177.0(2)
N(3)−Cu(2)−N(4) 173.28(17) 173.98(14) 176.0(3)
N(2)−Li(2)−N(3) 132.2(5) 133.0(4) 133.5(8)
Li(1)−Hal(1/1A)−Li(2) 104.4(3) 108.9(2) 102.0(4)
Li(3)−Hal(1/1A)−Li(2) 104.7(3) 106.5(2) 102.1(4)
N(1)−Li(1)−Hal(1/1A) 122.1(4) 122.8(4) 124.6(6)
N(4)−Li(3)−Hal(1/1A) 118.8(5) 120.3(4) 120.3(6)
Li(1)−N(1)−Cu(1) 90.6(3) 92.5(3) 96.1(5)
Li(2)−N(2)−Cu(1) 96.4(3) 97.5(3) 101.2(5)
Li(2)−N(3)−Cu(2) 95.5(3) 97.3(2) 100.1(5)
Li(3)−N(4)−Cu(2) 92.4(3) 89.0(3) 96.8(5)
Table 7.7: Selected bond angles, ◦, for 20, 21, and 22. Hal(1) = Br(1) (20 and 21),
Hal(1A) = I(1A) (22)
147
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
7.3.3 Trilithium bromo[tetra(cis-2,6,-dimethylpiperidido)]-
cuprate, [(DMP)
2
CuLi · 2THF]
2
LiBr, 23
In an attempt to form a Gilman-type cuprate incorporating DMP ligands, vari-
ous syntheses were attempted, including a reaction between LiDMP and CuBr in
toluene followed by recrystallisation from a mixture of THF and hexane (Scheme
7.3).
NLi
2 CuBr
tol,  –78 °C
4
1) – tol, + THF
2) filter
3) reduce THF, + hex
N Cu N
Li Li
NN Cu
Scheme 7.3: Synthetic strategy towards forming DMP-based Gilman-type
bis(amido)cuprates
The synthesis outlined in Scheme 7.3 yielded isolable crystalline material after
storage at –27 ◦C. 7Li NMR spectroscopy revealed a single peak (at δ 1.52 ppm),
which as the 7Li NMR spectrum of [(TMP)2CuLi]2, 2, also produces one signal (at
δ 1.28 ppm), suggested that the synthesis had been successful. However, the 1H
NMR spectrum revealed two peaks attributable to THF (at δ 3.67 and 1.42 ppm)
as well as several peaks attributable to the DMP ligand protons, and integration of
peaks indicated a 1:1 ratio. In addition, elemental analysis revealed the presence
of bromide in the sample (8.28 %) which coupled with the 1H NMR spectroscopic
data suggested that a higher-order cuprate had crystallised, which appeared to
be contrary to the 7Li NMR spectroscopic data. X-ray crystallographic studies
were able to provide conclusive data, revealing that the expected Gilman-type
structure had not formed, but instead that a THF-solvated adduct-type cuprate,
[(DMP)2CuLi · 2THF]2LiBr, 23 (Figure 7.6) had been synthesised. This result
could be reliably reproduced if the product was crystallised from bulk THF (in
low volumes) without the need to add a large volume of hexane in the final step.
The structure of 23 is highly analogous to that of [(DMP)2CuLi ·OEt2]2LiBr, 21,
although each Li centre is now pseudo-tetrahedral four-coordinate and therefore
has a full coordination sphere, which is possible due to the lower steric demand
imposed by THF as compared to Et2O.
Although the core of 23 appears unchanged relative to the cores of 19, 20, 21, and
22 on the replacement of Et2O with THF, on close inspection the effect of using
148
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
Figure 7.6: Structure of [(DMP)2CuLi · 2THF]2LiBr, 23, H atoms omitted for
clarity
149
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
a stronger Lewis donor can be noted. While the N(2/3)−Li(2) bond lengths are
unaffected (2.045(11)/2.029(11) A˚), the N(1/4)−Li(1/3) bond lengths are relatively
extended (2.067(13)/2.094(15) A˚) when compared not only to 21 but also to the
other adduct-type cuprates and the Li−N bond lengths in the previously discussed
Lipshutz-type cuprates containing both Et2O and THF donors (Chapters 5 and
6). This is coupled with a significant increase in all of the Li−Br bond lengths
when compared with 21 (Table 7.8).
21 23
Li(1)−Br(1) 2.474(8) 2.609(11)
Li(1)−Br(2) 2.592(7) 2.677(11)
Li(1)−Br(3) 2.515(8) 2.602(12)
Table 7.8: Comparison of the Li−Br bond lengths, A˚, in 21 and 23, which are
longer in the THF-solvated 23 than the analogous Et2O-solvated 21
The similar increase in length for each of the three bonds means that the ratio
between the Li(2)−Br(1) and mean Li(1/3)−Br(1) bond lengths (1.03) remains
essentially unchanged in 23 when compared to 21 (1.04). The increase in Li−N and
Li−Br bond lengths does affect the N−Cu−N bond angles though, which are much
more similar to those for [(DMP)2CuLi·OEt2]2LiI, 22, than those for 21 (Table 7.9).
21 22 23
N(1)−Cu(1)−N(2) 175.37(16) 177.0(2) 177.2(2)
N(3)−Cu(2)−N(4) 173.28(17) 176.0(3) 176.1(2)
Table 7.9: Comparison of the N−Cu−N bond angles in 21, 22, and 23, show-
ing that, surprisingly, the values for the THF-solvated bromide-based adduct-
type cuprate [(DMP)2CuLi · THF]2LiBr, 23, match those for the iodide-based
[(DMP)2CuLi · OEt2]2LiI, 22, more closely than the values for the Et2O-solvated
bromide-based [(DMP)2CuLi ·OEt2]2LiBr, 21
This data suggests that both the donor solvent and the nature of the halide play
a role in determining the structure of the core of the cuprate. While the inter-
metal bridges in 21 are highly unsymmetrical and obtuse, those in 23 are both
symmetrical and close to 90◦ (93.6◦av., range = 2.3
◦), possibly because in 23 the
Li(1/3)−N(1/4) bonds are longer since Li(1) and Li(3) are more heavily solvated.
150
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
Br(1)−Li(1) 2.609(11) N(1)−Cu(1)−N(2) 177.2(2)
Br(1)−Li(2) 2.677(11) N(3)−Cu(2)−N(4) 176.1(2)
Br(1)−Li(3) 2.602(12) N(2)−Li(2)−N(3) 132.2(6)
Cu(1)−N(1) 1.903(5) Li(1)−Br(1)−Li(2) 107.5(3)
Cu(1)−N(2) 1.904(5) Li(3)−Br(1)−Li(2) 105.2(4)
Cu(2)−N(3) 1.882(5) Br(1)−Li(1)−N(1) 115.4(5)
Cu(2)−N(4) 1.904(6) Br(1)−Li(3)−N(4) 116.6(6)
N(1)−Li(1) 2.067(13) Li(1)−N(1)−Cu(1) 94.3(4)
N(2)−Li(2) 2.045(11) Li(2)−N(2)−Cu(1) 93.9(4)
N(3)−Li(2) 2.029(11) Li(2)−N(3)−Cu(2) 94.2(4)
N(4)−Li(3) 2.094(15) Li(3)−N(4)−Cu(2) 92.0(4)
Table 7.10: Selected bond lengths, A˚, and angles, ◦, for 23
7.4 Solution Studies of Adduct-type Cuprates
1H NMR spectroscopic studies on redissolved crystalline samples of each of the
adduct-type cuprates 19, 20, 21, 22, and 23 in C6D6 revealed the presence of DMP
groups and Et2O (19, 20, 21, and 22) or THF (23) molecules in the expected ratios
given the crystallographic data. However, each spectrum also contained a signal
at δ ∼0.85 ppm, which can be attributed to the NH signal from HDMP (δ 0.83
ppm, see Section 3.1 for further details). With the exception of 20, any HDMP
present would have to have re-formed during preparation of the NMR samples. As
in all cases the deuterated solvents in which the samples were dissolved were stored
over fresh sodium mirrors, it indicates that these compounds all exhibit extreme
moisture sensitivity. This is corroborated by the 13C NMR data, where weak sig-
nals attributable to each carbon environment in HDMP are observed in each case.
In the 1H spectrum of THF-solvated 23 the HDMP signal is significantly weaker
when compared with that produced by compounds 19, 21, and 22, suggesting that
the THF-solvation confers superior stability over the etherated species. However,
hydrolysis effect notwithstanding, analysis of the 7Li NMR spectra indicate a sig-
nificant level of solid-state structure retention in the solution-state for all of these
compounds. Each spectrum exhibits a low-field signal (at δ 2.16–2.18 ppm) which
matches the major signal (at δ 2.19 ppm) in a reference spectrum of LiDMP (see
Section 3.1). For compounds 19, 21 and 22 there are two dominant signals at
δ 1.83–1.84 and 1.41–1.50 ppm in a 1:2 ratio, consistent with the presence of two
distinct lithium environments (Li(2) and Li(1/3), respectively) in the crystal struc-
tures. The 7Li NMR data for 20 is more complicated but also consistent with the
crystallographic data. Along with the LiDMP signal at δ 2.16 ppm, the presence
of Li ·HDMP introduces a third major peak at δ 1.66 ppm. The proximity of this
151
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
peak to the Li · OEt2 peak at δ 1.48 ppm causes it to appear as a shoulder on
the former, making conclusive integration of the three dominant peaks difficult.
However, they are approximately in the expected ratio of 1:1:1. In 23 only a single
signal is observed in the 7Li NMR spectrum. This can be interpreted both in terms
of the previously noted greater stability (substantially eliminating HDMP forma-
tion) and the more polar environment created by the four THF molecules over the
Et2O/HDMP molecules in compounds 19, 20, 21, and 22. This suggests that for
23 the solid-state structure is not retained in solution.
7.5 Structural Differences Between Lipshutz-
type and Adduct-type Cuprates
The discovery of the new cuprate structure-type reported in Section 7.3 clearly
led to an important question: what causes the adduct-type structure to form in
preference to the Lipshutz-type structure? The results above show that neither the
volume nor identity of the donor solvent can be a determining factor for several
reasons. Firstly, the methods employed to synthesise the Lipshutz-type cuprates
[(TMP)2Cu(Hal)Li2 · OEt2]2 (Hal = Cl (16), Br (17), I (15)) and the adduct-
type cuprates [(DMP)2Cu(Hal)Li · OEt2]2LiHal (Hal = Cl (18 and 19), Br (20
and 21), I (22)) show clear similarities in terms of the amount of donor solvent
employed (see Chapter 3 for full details). Secondly, adduct-type cuprates formed
both from reaction mixtures containing stoichiometric and bulk amounts of donor
solvent. Finally, adduct-type cuprates formed with both Et2O and THF acting
as coordinating solvents, so in the reactions containing LiDMP clearly the adduct-
type structure is the most energetically favourable product regardless of the solvent
conditions. Eliminating these factors leaves only one variable in the reactions:
the choice of amine reagent. This is therefore presumed to determine which of
the structure types forms. As a representative example, comparing 16 with 19
(Figure 7.7) it is clear that the orientation of the amido ligands with respect to
the structure cores are dramatically different.
In 16, the two TMP ligands connected to a given Cu centre lie endo,endo with
respect to the central core, pointing away from one another (Figure 7.8(a)). Lig-
ands in an endo orientation can be thought of as lying “flat”, meaning they are
approximately parallel to the N−Cu−N unit, with obtuse Cu−N−C-4 angles (C-4
= carbon in the 4-position on the relevant amido ligand, e.g. Cu(1)−N(1)−C(10)
152
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
(a) 16 (b) 19
Figure 7.7: Structures of (a) [(TMP)2Cu(Cl)Li2 ·OEt2]2, 16, and (b) [(DMP)2CuLi·
OEt2]2LiCl, 19
158.1◦). This can be contrasted with the amido orientations in 19 where each set
of DMP-ligands lie face-on to one another projecting away from the metallocyclic
structure in an exo,exo orientation (Figure 7.8(b)). Ligands in an exo orientation
lie almost perpendicular to the N−Cu−N units (i.e. “upright”) with Cu−N−C-4
bond angles much closer to 90◦ (e.g. Cu(1)−N(1)−C(4) 106.8◦). This is possible
as both of the methyl groups in each DMP ligand lie equatorial and so avoid any
steric clash between two amido ligands bridging the same Cu centre (e.g. N(1) and
N(2) in Figure 7.7(b)). However, the TMP ligands in 16 would not be able to take
up an exo orientation as this would create a steric interaction between the axial
methyl groups (Figure 7.9(a)). The steric interaction between amido ligands either
side of the unsolvated Li centre should also be considered. To form an adduct-type
cuprate, the two ligands bridging the unsolvated Li centre (N(2) and N(3) in Fig-
ure 7.7(b)) must reside face-on to each other. TMP ligands would be precluded
from taking up this orientation as of the four methyl groups, two will be axial.
These axial methyl groups would interact with each other so imposing a highly
unfavourable steric penalty (Figure 7.9(b)).
153
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
(a) (b)
Figure 7.8: Structures of the Lipshutz-type monomers incorporated in (a) TMP-
based 16 and (b) DMP-based 19, highlighting the endo,endo and exo,exo amido
orientations in either
N Cu N
(a)
N
Li
N
(b)
Figure 7.9: Schematic representations of arrangements of TMP ligands lying
exo,exo which are precluded due to steric clashing between the axial methyl groups
in (a) Lipshutz-type and (b) adduct-type structures
154
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
7.6 Directed ortho Cupration Abilities of Add-
uct-type Cuprates
It has been previously established that DoCu reactivity requires Lipshutz-type
cuprates which act as a source of Gilman-type monomer. This was achieved via
theoretical calculations which probed several plausible reaction pathways and found
that deprotocupration by a monomeric Gilman-type cuprate, formed from the
dissociation of a dimeric Lipshutz-type cuprate, provided the most energetically
favourable route138 (see Section 1.8.2.3). As such, investigations were undertaken
seeking to find out if adduct-type cuprates can also act as a source of reactive
Gilman-type monomer and therefore be viable DoCu reagents. This is plausible
as, like Lipshutz-type cuprates, adduct-type cuprates incorporate a Gilman-type
monomer into their structure and in fact can be viewed as resulting from the asso-
ciation of Lipshutz- and Gilman-type monomers. To act as a source of Gilman-type
monomers all that is required, as is also the case for Lipshutz-type cuprates, is for
the adduct-type cuprate to dissociate into its constituent parts (Scheme 7.4).
Li
NR2
Cu
R2
N
Cl
Li
Li
R2N
Cu
R2
N
Cl
Li Li
R2N Cu NR2
Li
N
R2
Cu
NR2
Cl
Li
Li
R2N
Cu
R2
N
Cl
Li
Li
R2N Cu NR2
Lipshutz-type dimer
Adduct-type Gilman-type monomer
Gilman-type monomer
Lipshutz-type monomer
Li
R2N Cu NR2
+
+
Cl
Li Li
R2N Cu NR2
Lipshutz-type monomer
+ LiCl
Scheme 7.4: Dissociation of Lipshutz- and adduct-type cuprates yielding DoCu-
active Gilman-type monomers. In each case, the resulting Lipshutz-type monomer
could also dissociate further, yielding a second equivalent of the Gilman-type
monomer and LiCl
To this end, N,N -diisopropylbenzamide was reacted with both an in situ prepara-
tion of a “Lipshutz-type formulation” cuprate and with pre-isolated crystals of an
adduct-type cuprate followed by trapping of the product with I2. It was hoped that
this data could be compared to the reaction between N,N -diisopropylbenzamide
and a DMP-based Gilman-type cuprate. However, despite repeated attempts to
synthesise such a species, crystals that were unambiguously a Gilman-type cuprate
155
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
could not be obtained.∗ In each case where X-ray grade crystals were isolated, re-
finements of the crystallographic data indicated that both [(DMP)2CuLi·THF]2 and
[DMPCu]4 were present in varying proportions, although a satisfactory model could
not be found for any of the datasets. However, Gilman-type cuprates based on TMP
have been shown to be unreactive in DoCu reactions, which has been evidenced
both theoretically and experimentally: Theoretical calculations demonstrated that
a pathway where a Gilman-type dimer is the active species is highly unfavourable.15
Experimentally, it was shown that in reaction with N,N -diisopropylbenzamide, the
Gilman-type monomer MeCu(µ-TMP)Li·TMEDA, 5, returned a much poorer yield
of the 2-substituted product than a reaction mixture containing 5 and LiCN.139
Given this evidence, a DMP-based Gilman-type species would be expected to
be similarly unreactive. As the results presented in Chapter 5 indicated that
the nature of the halide should not have a significant effect on the DoCu abil-
ity of the cuprate, initial studies sought to investigate which of the syntheses of
adduct-type cuprates was most viable for scaling-up in order to produce a suitable
quantity of crystals to redissolve and react with the benzamide. As the synthe-
sis of [(DMP)2CuLi · OEt2]2LiBr, 21, carried out in hexane with 2 equivalents
of Et2O (Section 3.4.4, route (b)), proved to be most efficient, this was chosen
as the representative adduct-type cuprate. To ensure consistency, CuBr was em-
ployed as the copper source when preparing the in situ mixture of Lipshutz-type
formulation cuprate. The reaction between a mixture of LiDMP, CuBr and N,N -
diisopropylbenzamide in a 4:2:1 ratio (2 equivalents of Lipshutz-type cuprate per
arene) in THF afforded 2-iodo-N,N -diisopropylbenzamide, 6, in a yield of 80 %. A
1:1 mixture of redissolved 21 and N,N -diisopropylbenzamide (2 equivalents of Cu
per arene) in THF afforded 6 in an 82 % yield (Scheme 7.5).
∗ This will be the subject of future work, see Section 9.1
156
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
I
O NiPr2
H
O NiPr2
I2, rt, 16 h
O NiPr2
Cu(DMP)Li
THFTHF
4 DMPLi + 2 CuBr
O NiPr2
H
I2, rt, 16 h
O NiPr2
Cu(DMP)Li
THFTHF
(80 %)
(82 %)
(a)
(b) 21 6
6
Scheme 7.5: DoCu of N,N -diisopropylbenzamide by (a) a “Lipshutz-type formula-
tion” base prepared in situ from LiDMP and CuBr and (b) pre-isolated 21, with
subsequent functionalisation by I2
Importantly, the high yields prove that the adduct-type cuprates are efficient
DoCu agents. As the aforementioned theoretical studies have shown that the most
favourable pathway to deprotocupration by Lipshutz-type cuprates is via reactive
Gilman-type monomers, and it is plausible that adduct-type cuprates can dissoci-
ate to form Gilman-type monomers, it is likely that the reaction pathways are the
same regardless of the cuprate structure-type (as supported by theoretical data, see
Section 7.7). The substantial cost differential between HDMP and HTMP means
that adduct-type cuprates are highly-promising DoCu agents. As the yields from
the two routes are essentially identical it means that an in situ preparation con-
tains a reactive species and no benefits are gained by purification of the cuprate
base via recrystallisation.
7.7 Theoretical Calculations
To investigate the relationship between Gilman-, Lipshutz- and adduct-type species
further, theoretical calculations were carried out employing DFT methods. Pre-
vious calculations (discussed in Section 1.8.2.3) showed that the interconversion
between monomeric Gilman- and Lipshutz-type bis(amido)cuprates containing a
cyano group displays a slight enthalpic preference for the Lipshutz-type form (by
157
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
9.1 kcal mol−1),139 and these further studies aimed to investigate the relative en-
ergy of adduct-type cuprates.
To make a fair comparison with previous theoretical studies of cuprate compounds
(see Section 1.8.2.3), a simplified system involving NMe2 ligands was employed.
Firstly, calculations were carried out on the interconversion between a Lipshutz-
type dimer, [(Me2N)2Cu(Cl)Li2 · OMe2]2 (LD), and [(Me2N)2Cu(Cl)Li2 · OMe2] ·
[(Me2N)2Li2 · OMe2] (LM·(LiNMe2)2), which represents a solvate of the lithium
amide-coordinated Lipshutz-type monomer and was found to make up 71 % of
the composition of [(DMP)2CuLi · Et2O]1.29 · [(DMP)2Li2 · Et2O]0.71LiCl, 18 by
crystallography (Scheme 7.6). The loss of CuCl from the Lipshutz-type dimer,
LD is clearly highly disfavoured (∆G = +45.9 kcal mol
−1) which suggests that
other factors play a key role in the formation of 18. These might include poor
dissolution of CuCl, which forms a slurry in both neat hydrocarbon media and
solvent systems containing donor solvents such as THF and Et2O. This would
mean that the ratio of lithium amide to copper present in solution is much greater
than the intended 2:1 ratio.
Cl
Li
Li
NMe2
Cu
N
Me2
Cl
Li
Li
Me2N
Cu
Me2
N
S
S
Li
NMe2
Li
Me2
N
Cl
Li
Li
Me2N
Cu
Me2
N
SS
+54.5
[+45.9]
– CuCl
LD
S = OMe2
LM·(LiNMe2)2
Scheme 7.6: The interconversion of a Lipshutz-type dimer (LD) and lithium amide-
coordinated Lipshutz-type monomer (LM·(LiNMe2)2) at the B3LYP/631SVP
level of theory (∆E [∆G] are in kcal mol−1)
Next, LM·(LiNMe2)2 was replaced with a simplified adduct-type cuprate
(LMGM) and the relative energies of LD, LMGM and the corresponding Gilman-
type dimer (Me2N)2CuLi · OMe2 (GD) were calculated. Following the previous
calculation, starting at LD one equivalent of LiCl was removed sequentially to
form LMGM and finally GD (Scheme 7.7). The results show that the Lipshutz-
type dimer is highly favoured over the other two species. Although, as an isolated
monomeric LiCl lattice fragment is highly unstable, the enthalpy changes are large
and positive (∆E = +45.3 kcal mol−1 (LD-LMGM) and +41.6 kcal mol
−1 (LMGM-
GD)) and dominate the Gibbs free energy changes (∆G = +36.2 and +32.1 kcal
mol−1 respectively) even though the processes are entropically favourable.
158
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
Cl
Li
Li
NMe2
Cu
N
Me2
Cl
Li
Li
Me2N
Cu
Me2
N
S
S
Li
NMe2
Cu
Me2
N
Cl
Li
Li
Me2N
Cu
Me2
N
SS
+45.3
[+36.2]
– LiCl
LD
S = OMe2
+41.6
[+32.1]
Me2N
Li
Cu NMe2
Li
Me2N Cu NMe2
S S
– LiCl
GMLM GD
Scheme 7.7: The interconversion of Lipshutz- and Gilman-type dimers with se-
quential loss of LiCl at the B3LYP/631SVP level of theory (∆E [∆G] are in kcal
mol−1)
In a real system, LiCl would be unlikely to exist as discrete monomeric units and
when a donor solvent is present, in all probability the Lewis base will act to stabilise
the lithium centres. Hence, further calculations were carried out which took this
hypothesis into account. Accordingly, “LiCl” was modelled as a cubane tetramer,
to represent a system in predominantly hydrocarbon media (save for the donor
solvent coordinated to the cuprate species), and as LiCl ·3OMe2 and [LiCl ·OMe2]4
to represent systems where a donor solvent is present in either bulk or stoichiometric
quantities, respectively (Scheme 7.8). In the case of elimination of unsolvated
(LiCl)4 lattice fragments (Scheme 7.8, top) there continues to be an enthalpic
preference for the Lipshutz-type dimer LD over LMGM (∆E = 9.9 kcal mol
−1)
although now LMGM is much closer in energy toGD (∆E = 6.2 kcal mol
−1). When
a Lewis basic donor is introduced, the solvation of LiCl means that the enthalpic
preference switches in favour of GD. When there is an excess of Me2O (Scheme 7.8,
middle), the enthalpic preference forGD over LD is small (∆E = 8.5 kcal mol
−1) but
the entropic loss associated with forming LiCl · 3OMe2 means that LD is the most
favourable of the three complexes. Experimental observations have indicated that
when LiHal is excluded from a cuprate species in a reaction mixture containing
Et2O (and crystallises favourably or concomitantly with cuprate compounds) it
forms tetrameric (LiHal·OEt2)4 in the solid-state, which can particularly be an issue
when Hal = I.200 When this scenario is modelled theoretically (Scheme 7.8, bottom),
the enthalpic preference for GD (∆E = 13.7 kcal mol
−1) is again small although
significantly larger than in the previous case. However, as would be expected,
entropically the conversion of LD to GD is substantially less unfavourable than
when “LiCl” = LiCl · 3OMe2, and for both steps the change in ∆G is small (+4.3
kcal mol−1 (LD-LMGM), +0.2 kcal mol
−1 (LMGM-GD)). Overall, the difference
in energy between the various cuprate forms is now low enough to allow the whole
system to be considered as being in equilibrium.
159
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
Cl
Li
Li
NMe2
Cu
N
Me2
Cl
Li
Li
Me2N
Cu
Me2
N
S
S
Li
NMe2
Cu
Me2
N
Cl
Li
Li
Me2N
Cu
Me2
N
SS
+9.9
[+9.1]
+6.2
[+5.0]
Me2N
Li
Cu NMe2
Li
Me2N Cu NMe2
S S
– 1/4(LiCl)4 – 1/4(LiCl)4
Cl
Li
Li
NMe2
Cu
N
Me2
Cl
Li
Li
Me2N
Cu
Me2
N
S
S
Li
NMe2
Cu
Me2
N
Cl
Li
Li
Me2N
Cu
Me2
N
SS
–2.4
[+21.1]
–6.1
[+17.0]
Me2N
Li
Cu NMe2
Li
Me2N Cu NMe2
S S
– LiCl·S3
+ 3S
– 1/4(LiCl·S)4
Cl
Li
Li
NMe2
Cu
N
Me2
Cl
Li
Li
Me2N
Cu
Me2
N
S
S
Li
NMe2
Cu
Me2
N
Cl
Li
Li
Me2N
Cu
Me2
N
SS
–5.0
[+4.3]
Me2N
Li
Cu NMe2
Li
Me2N Cu NMe2
S S
+ S
– 1/4(LiCl·S)4
–8.7
[+0.2]
+ S
– LiCl·S3
+ 3S
LMGM GDLD
Scheme 7.8: The interconversion of Lipshutz- and Gilman-type dimers in the pres-
ence of varying amounts of Me2O at the B3LYP/631SVP level of theory (∆E [∆G]
are in kcal mol−1, S = OMe2)
160
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
In order to investigate solvent effects theoretically, the calculations outlined in
Scheme 7.8 were repeated replacing Me2O with THF (Scheme 7.9). When consid-
ering the elimination of (LiCl)4 (Scheme 7.9, top), the entropy changes are slightly
smaller with THF acting as the donor solvent. However, the values of ∆E and
∆G do not change significantly. In the presence of an excess of THF (Scheme 7.9,
middle), solvation of the Li ions by the strong Lewis base means that elimination
of LiCl · 3THF is enthalpically favourable, giving a large value of ∆E for the con-
version of LD to GD (–17.6 kcal mol
−1). However, this does still not compensate
for the high entropic penalty for ordering the solvent around the LiCl and LD is
still strongly favoured overall (∆G = 30.4 kcal mol−1). Finally, when “LiCl” =
[LiCl · THF]2 (Scheme 7.9, bottom), the nature of the solvent has only a small
effect on the values of ∆E and ∆G for the interconversions calculated. Once again,
GD is favoured enthalpically (∆E = 16.8 kcal mol
−1) but the entropic effects mean
that the conversions of LD to LMGM (∆G = +2.7 kcal mol
−1) and LMGM to GD
(∆G = +1.7 kcal mol−1) are very close to being in equilibrium.
Cl
Li
Li
NMe2
Cu
N
Me2
Cl
Li
Li
Me2N
Cu
Me2
N
S
S
Li
NMe2
Cu
Me2
N
Cl
Li
Li
Me2N
Cu
Me2
N
SS
+9.8
[+7.7]
+5.7
[+6.7]
Me2N
Li
Cu NMe2
Li
Me2N Cu NMe2
S S
– 1/4(LiCl)4 – 1/4(LiCl)4
Cl
Li
Li
NMe2
Cu
N
Me2
Cl
Li
Li
Me2N
Cu
Me2
N
S
S
Li
NMe2
Cu
Me2
N
Cl
Li
Li
Me2N
Cu
Me2
N
SS
–6.7
[+15.7]
–10.9
[+14.7]
Me2N
Li
Cu NMe2
Li
Me2N Cu NMe2
S S
– LiCl·S3
+ 3S
– 1/4(LiCl·S)4
Cl
Li
Li
NMe2
Cu
N
Me2
Cl
Li
Li
Me2N
Cu
Me2
N
S
S
Li
NMe2
Cu
Me2
N
Cl
Li
Li
Me2N
Cu
Me2
N
SS
–6.3
[+2.7]
Me2N
Li
Cu NMe2
Li
Me2N Cu NMe2
S S
+ S
– 1/4(LiCl·S)4
–10.5
[+1.7]
+ S
– LiCl·S3
+ 3S
LMGM GDLD
Scheme 7.9: The interconversion of Lipshutz- and Gilman-type dimers in the pres-
ence of varying amounts of THF at the B3LYP/631SVP level of theory (∆E [∆G]
are in kcal mol−1, S = THF)
Directed ortho Cupration reaction pathways where an adduct-type cuprate,
[(Me2N)2CuLi(OMe)2]2LiCl, is the active species have also been probed theoreti-
161
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
cally, employing N,N -dimethylbenzamide as a representative substrate. Initially,
a pathway where an intact adduct-type cuprate molecule, LMGM carries out the
deprotonation was considered (Scheme 7.10). Coordination of the substrate to
LMGM and concomitant loss of a solvent molecule to form IM1G-L is favourable
enthalpically (∆E = –4.3 kcal mol−1) while the change in entropy is, as expected,
negligible. However, the deprotonation step to form TSG-L has a prohibitively high
activation energy (∆G‡ = +26.5 kcal mol−1) and forms a high energy intermediate,
IM2G-L (∆G = +19.4 kcal mol
−1, relative to RTG-L).
Li
NMe2
Cu
Me2
N
Cl
Li
Li
Me2N
Cu
Me2
N
SS –5.7
[–2.7]Li
NMe2
Cu
Me2
N
Cl
Li
Li
Me2N
Cu
Me2
N
SO
NMe2H
Li
NMe2
Cu
Me2
N
ClLi
Li
Me2N
Cu
Me2
N
SO
NMe2
H Li
NMe2
Cu
Me2
N
Cl
Li
Li
Me2N
Cu
Me2
N
SO
NMe2
H
IM2G-L
+27.0
[+26.5]
TS1G-LIM1G-L
RTG-L
–4.3
[–4.4]
+
NMe2
O
– S
2.33
2.60
1.96 2.55
2.36
1.96
1.86
2.32
1.93
1.93
2.47
2.27
2.33
1.95
1.29
1.46
1.88
2.33
2.48
2.29
1.88
2.11
1.97
Scheme 7.10: Modelled pathway for the deprotonation of N,N -dimethylbenzamide
by [(Me2N)2CuLi(OMe)2]2LiCl (IM1G-L-TS1G-L-IM2G-L) at the B3LYP/631SVP
level of theory (Bond lengths and ∆E [∆G] are in A˚ and kcal mol−1 respectively,
S = OMe2)
Given that Gilman-type monomers are considered to be required for effective
DoCu,138 a more plausible pathway for reaction was felt to involve the adduct-
type species LMGM dissociating into constituent Gilman-type and Lipshutz-type
monomers (GM and LM) followed by coordination of the Gilman-type monomer
to the substrate (IM1G) and association of Lipshutz-type monomers to form a
dimer (LD) (Scheme 7.11). The dissociation step is energetically unfavourable but
not considered to be prohibitive (∆G = +15.4 kcal mol−1), given that the most
favourable pathway for deprotocupration via dissociation of a monomeric Lipshutz-
type cuprate MeCu(NMe2)(CN)Li2 · 2OMe2 involves a similar-sized energy change
(∆G‡ = +14.9 kcal mol−1) between the first intermediate and the transition state138
(Figure 7.10). The favourable coordination step (∆G = –9.3 kcal mol−1) means
that the overall change in ∆G for the pathway is just +6.1 kcal mol−1. These values
can be compared with those calculated for the DoCu of N,N -dimethylbenzamide
by the monomeric Lipshutz-type species, MeCu(NMe2)(CN)Li2 ·2OMe2. Although
this is an alkyl(amido)cuprate, for the pathway where deprotonation is mediated by
the amido group, a reasonable comparison can be drawn (Figure 7.10). In this case
the dissociation step, which results in the formation of a Gilman-type monomer and
162
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
LiCN ·S (RSG), is less unfavourable (∆G = +9.5 kcal mol
−1) than is the transition
from LMGM to RSG(A). However, the coordination step (RSG-IM1G) is enthalpi-
cally much less favourable (∆E = –10.2 kcal mol−1) than for the adduct-mediated
pathway (RSG(A)-IM1G(A)), meaning that the entropic loss makes the coordina-
tion step slightly energetically unfavourable (∆G = +1.6 kcal mol−1). Overall, the
formation of IM1G from RTL involves a greater energy change (∆G = +11.1 kcal
mol−1) than for the pathway starting with the dissociation of LMGM (RTG-L-
IM1G(A) ∆G = +6.1 kcal mol
−1). This data suggests that the calculated route
to IM1G(A) (starting with LMGM) is very viable and that it compares favourably
with the pathway calculated using a Lipshutz-type monomer. Overall it reinforces
the evidence that LMGM adduct-type cuprates are effective DoCu agents.
Me2N
Cu
Me2
N
Li
Me2
N
Cl
Li
Cu
NMe2
Li
RTG-L
OMe2N
H
O
Me2N
N
Me2
Li
NMe2
Cu
Me2N
Cu
Me2
N
Li
Cl
Li
2
++
IM1G(A) LD
½
+2.8
[+6.8]
S S
S
S
Me2N
Li
NMe2
Cu
S
Me2N
Cu
Me2
N
Li
Cl
Li
2
½
S
–20.3
[–9.3]
–23.1
[–15.4]
+
Me2N O
RSG(A)
Scheme 7.11: The interconversion of Lipshutz-Gilman adduct LMGM and the
monomeric Gilman-(N,N -dimethylbenzamide) complex (IM1) for DoCu at the
B3LYP/631SVP level of theory (∆E [∆G] are in kcal mol−1, S = OMe2)
163
7. PENTAMETALLIC ADDUCT-TYPE AMIDOCUPRATES
30.0
20.0
10.0
0.0
∆G (kcal mol-1)
0.0
+11.1
IM1G
+26.1
TS1G
+14.9
RTL
+9.5
RSG
30.0
20.0
10.0
0.0
0.0
+15.4
RSG(A)
RTG-L
+15.4
+6.1
IM1G(A)
Figure 7.10: Summary of the initial steps in the deprotonative cupration pathways
of N,N -dimethylbenzamide by MeCu(NMe2)(CN)Li2 · 2OMe2 (left) and LMGM
(right) at the B3LYP/631SVP level of theory. Energy changes are shown in kcal
mol−1 and are relative to RTL and RTG-L respectively
164
Chapter 8
Conclusions
In Chapter 4 it was demonstrated that the behaviour of phosphido ligands in
the formation of lithium cuprates is very different to that of analogous amido
ligands. Targeting Gilman- and Lipshutz-type cuprates with diphenylphosphido
ligands, the treatment of CuCN with LiPPh2 (2 equivalents) in THF in fact
only afforded Ph2PCu(CN)Li · 2THF, 10, as isolable crystalline material. In the
solid-state this species agglomerated into a novel network based on cyclohexyl-
type [(Ph2P)3(CuCN)3]
3– nodes. Each exocyclic CN ligand interacts with two bis-
solvated Li centres which link the resulting puckered 2D sheets together. In con-
trast, the treatment of CuI with LiPPh2 (2 equivalents) in a hydrocarbon medium
resulted in the incomplete conversion of the lithium phosphide and reproducibly
generated the unusual ion-separated [(Ph2P)6Cu4][Li · 4THF]2, 9, where the di-
anionic phosphidocopper component forms an adamantoid cage. Results from the
employment of dicyclohexylphosphido ligands were far more promising, although
unlike lithium amidocuprate structure-types reported elsewhere in this thesis (and
those previously observed), [(Cy2P)2CuLi·2THF], 11, aggregates into a linear poly-
mer in the solid-state. The phosphido ligands act as inter-metal bridges not only
within the monomeric unit but also as part of the polymeric structure, favouring
the exclusion of LiCN over the formation of a Lipshutz-type cuprate.
Chapter 5 presented the continuation of reactivity and solid-state studies of bis-
(amido)cuprates incorporating TMP. Previous research had concluded that the
most effective DoCu agents were Lipshutz-type cuprates (which act as a source
of the reactive monomeric Gilman-type cuprates) and that these required bulky
amides,15,138,139 meaning that TMP was an ideal choice of ligand. While much
was known about cyano- and iodide-based cuprates, this thesis extended the study
165
8. CONCLUSIONS
to other halides, namely chloride and bromide. Reactivity studies demonstrated
that lithium cuprates synthesised from CuCl and HTMP are highly efficient in the
DoCu of 2-chloropyridine and that reaction conditions are optimal when the base is
prepared in situ in bulk THF. As expected, by precipitating or removing LiCl from
the reaction mixture, or by carrying out reactions with a pre-isolated Gilman-type
base, yields of 3-benzoyl-2-chloropyridine were significantly reduced. A lithium
cuprate mixture synthesised from CuBr and HTMP was also found to be an effi-
cient DoCu agent in reactions with 2-chloropyridine, although it was not able to
furnish the substrate with various benzoyls at the 3-position in as consistently high
yields as its chloride-based counterpart. These results were consistent with previ-
ous reactivity studies and suggested that a reaction mixture containing a cuprate
of Lipshutz-type formulation is required to carry out DoCu effectively. Solid-state
studies found that the reaction between HTMP (2 equivalents) and CuX (X = Cl,
Br) in a mainly hydrocarbon medium containing 2 equivalents of THF resulted
in the isolation of the Lipshutz-type cuprates [(TMP)2Cu(Cl)Li2 · THF]2, 12, and
[(TMP)2Cu(Br)Li2 ·THF]2, 13, which both have structures highly analogous to pre-
viously isolated Lipshutz-type bis(TMP)cuprates [(TMP)2Cu(CN)Li2 ·THF]2, 1,
15
and [(TMP)2Cu(I)Li2 ·THF]2, 3.
138 These results showed that the nature of the X
group (X = CN, Hal) does not have a significant effect on the solid-state structure of
THF-solvated Lipshutz-type bis(TMP)cuprates. However, the increased ring strain
in 12 compared to 13, which is due to the smaller halide in the central metallocycle,
provides a possible explanation for the greater DoCu reactivity of 12, as it is more
favourable to exclude LiX and form a reactive monomeric Gilman-type species.
In Chapter 6 the synthesis of a series of Lipshutz-type cuprates [(TMP)2Cu(X)Li2 ·
OEt2]2 (X = CN, Hal) was achieved. The results established that the identity of the
donor solvent coordinated to a Lipshutz-type bis(TMP)cuprate in the solid-state
does not necessarily affect the structure. However, it was found that the nature of
the product resulting from a particular synthetic pathway can be affected by the
identity of the donor solvent. Specifically, studies here discovered that if bulk Et2O
was added prior to recrystallisation, the abstraction of LiX was disfavoured and
dimeric Lipshutz-type cuprates formed. This method was successfully employed
in the synthesis of [(TMP)2Cu(CN)Li2 · OEt2]2, 14, [(TMP)2Cu(I)Li2 · OEt2]2,
15, [(TMP)2Cu(Cl)Li2 · OEt2]2, 16, and [(TMP)2Cu(Br)Li2 · OEt2]2, 17. This
contrasts with the observed behaviour of THF when employed as a recrystallisation
solvent in bulk quantities. In this case, the use of THF caused dimeric Gilman-
type cuprates to crystallise from the solution as the donor solvent was able to
effectively stabilise LiX units. If a stoichiometric amount (2 equivalents) of Et2O
166
8. CONCLUSIONS
was included in the reaction mixture, dimeric Lipshutz-type cuprates formed, as
demonstrated by the synthesis of 15 and 17 via this pathway. This behaviour
matched that observed when 2 equivalents of THF were employed in the synthesis
of Lipshutz-type cuprates (vide supra). With both Et2O and THF a stoichiometric
quantity of solvent was not able to effectively solvate LiX units (X = CN, Hal).
However, LiX is more likely to precipitate out when the solvent system contains
limited quantities of the weaker Lewis base Et2O, resulting in low yields of the
Lipshutz-type dimer. Although a dimeric Gilman-type cuprate may be expected
to crystallise preferentially if a proportion of LiX has precipitated out of solution,
theoretical work has suggested that the Lipshutz-type form is more energetically
stable than the Gilman-type form.138 Overall, the results suggest that the use of
diethyl ether provides a level of versatility not available from THF, as the amount
of donor solvent added does not have to be carefully controlled depending on the
required cuprate formation.
Taking the results presented in Chapters 5 and 6 as a whole, understanding of
higher-order cuprates has been developed and it has been demonstrated that CN
can be replaced by Cl, Br, or I without significantly affecting the nature of the spe-
cies that forms. Therefore, Lipshutz cuprates, RCu(R′)(CN)Li2, should be thought
of more generally as Lipshutz-type cuprates, RCu(R′)(X)Li2, where X = CN is just
one example of a wide class of complexes with related structures and reactivities.
In Chapter 7, the effects of changing the amido ligand from TMP to the cheaper but
still bulky DMP in the synthesis of bis(amido)cuprates were investigated. With the
aim of forming DoCu-active Lipshutz-type bis(DMP)cuprates, LiDMP (2 equiva-
lents) was added to CuHal (Hal = Cl, Br, I). However, solid-state studies revealed
a previously unreported class of triangulated amidocuprate best viewed as a 1:1
adduct between Gilman- and Lipshutz-type monomers. These adduct-type cupra-
tes have the general formula [(DMP)2CuLi ·nS]2LiHal (S = Et2O, n = 1; S = THF,
n = 2) and can be thought of as resulting from the abstraction of LiHal from a
Lipshutz-type dimer or the insertion of LiHal into a Gilman-type dimer. The dif-
ference between the Lipshutz-type and adduct-type cuprates can be rationalised by
the relative steric demands of the relevant ligands (TMP and DMP respectively).
The efficiency of adduct-type cuprates in DoCu reactions reinforces the importance
of LiX-containing systems (X = CN, Hal), which are believed to be required for
facile formation of the reactive monomeric Gilman-type species, in amidocuprate
reactivity. It also provides a cost-effective route to forming carbon-carbon bonds
to aromatic substrates, which on an industrial scale means the viability of util-
167
8. CONCLUSIONS
ising DoCu in organic syntheses, for example in the pharmaceutical industry, is
greatly improved.
168
Chapter 9
Further Work
The work presented in this thesis has helped to improve knowledge of the solid-
state structures of cuprate species and how they can be utilised in Directed ortho
Cupration reactions. It has also opened up new avenues of research, most notably
the discovery of a novel cuprate-structure type, which merit further investigation.
The work has also raised several questions to which future studies can be directed
or for which alternative solutions should be sought.
9.1 Future Amidocuprate Species
The structures of the adduct-type cuprate species [(DMP)2CuLi · nS]2LiHal (S =
Et2O, n = 1; S = THF, n = 2) synthesised using HDMP contrast with the struc-
tures of the Lipshutz-type species [(TMP)2Cu(X)Li2 ·S]2 (X = CN, Hal; S = Et2O,
THF), synthesised from the related but bulkier HTMP. The hypothesis that the
steric demand of the ligand plays a key role in determining which cuprate structure-
type forms in the solid-state clearly warrants further research. Work is now ongoing
to collect solid-state data for cuprates formed from a series of piperidine-based lig-
ands which confer varying steric demands. 2-methylpiperidine has a lower steric
demand than HDMP and so would be expected to form an adduct-type cuprate in
the solid-state. As it is readily available it makes an ideal candidate, and initial
reactions have proved promising. X-ray crystallographic data has been collected
of a higher-order cuprate in which, although the quality is poor, the connectiv-
ity is unambiguous and shows that, indeed, an adduct-type structure is favoured
(Figure 9.1). 2-methylpiperidine can be thought of as a “tied-back” analogue of
N -methylpropan-2-amine which, due to its low steric demand, would have previ-
169
9. FURTHER WORK
ously not been considered as a suitable amine ligand to form cuprates either to
employ in DoCu reactions or to study in the solid-state. This was because initial
DoCu reactivity studies concluded that a bulky amide ligand was a key feature
of an effective cuprate base. However, as cuprates incorporating ligands with a
low steric demand can be synthesised, this study could be extended further, for
example to piperidide, which is sterically analogous to dimethylamide. If prepa-
rations of piperidide-based lithium cuprates are successful, the species should be
tested for their efficiency in DoCu reactions, as importantly, piperidine is cheaper
to source than cis-2,6-dimethylpiperidine∗ and so the cost of deprotocupration re-
actions could be lowered even further.
Figure 9.1: Structure of [(2-methylpiperidine)2CuLi · 2THF]2LiBr
Ligands which confer steric demand intermediate between those of HDMP and
HTMP are of considerable interest for solid-state studies. There are two possible
candidates, trans-2,6-dimethylpiperidine and 2,2,6-trimethylpiperidine which, al-
though not commercially available, can be synthesised from 2-methylpiperidine201
and 2,2,6-trimethyl-2,3,4,5-tetrahydropyridine202 respectively. If higher-order cu-
prates can be synthesised incorporating these ligands, in the solid-state the ligands
would have at least one axial methyl group, which gives rise to several possibilities.
Lipshutz-type cuprates may result as adduct-type cuprate formation may well be
∗ The cheapest available piperidine costs £0.03 g−1. Data from http://www.alfa.com, accessed
19th March 2014
170
9. FURTHER WORK
precluded by steric hindrance between the axial methyl groups (Figure 9.2(a)). If
the ligands are able to arrange themselves in such a way as to avoid axial methyl
group steric clash (i.e. the axial methyl groups are situated on opposite faces of
the central core, Figure 9.2(b)) an adduct-type cuprate may be the most ther-
modynamically favourable form. The final possibility is that neither adduct- nor
Lipshutz-type cuprates form as the ligands favour Gilman-type structures or even
novel cuprate structure-types.
N Cu N
(a)
N
Li
N
(b)
Figure 9.2: Schematic representation of how trans-2,6-dimethylpiperidide ligands
may form either (a) a Lipshutz-type cuprate if the axial methyl groups are oriented
on the same side of the central core, precluding adduct-type cuprate formation, or
(b) an adduct-type cuprate if the axial methyl groups are oriented on opposite faces
of the central core
Although it has not been possible as yet to prove conclusively that a Gilman-
type cuprate can be formed containing DMP ligands, the data collected so far
has found a mixture of [(DMP)2CuLi · THF]2 and [(DMP)Cu]4 in each of the
individual crystals selected for X-ray diffraction studies. This suggests that it
should be possible to synthesise pure samples of [(DMP)2CuLi · THF]2 as there
appears to be a degree of Cu/Li disorder at the “lithium sites” (i.e. where one would
expect to find lithium in a Gilman-type cuprate). Hence current work is looking
into finding a synthetic pathway that can reproducibly deposit pure crystalline
material of [(DMP)2CuLi · THF]2. This species could then be tested for its ability
to ortho deprotonate N,N -diisopropylbenzamide and the results compared with the
analogous TMP-based cuprate, [(TMP)2CuLi]2, 2. In addition, the isolation of a
cyano-containing adduct-type cuprate would be of interest in order to ascertain
whether the inclusion of a C−−N group triangulated between three lithium centres
is possible (see Section 9.5).
As this work has demonstrated that higher-order homoleptic amidocuprates incor-
porating any of three halides (chloride, bromide and iodide) can be synthesised with
171
9. FURTHER WORK
both DMP and TMP as the ligand, it opens up the possibility of forming heterolep-
tic amidocuprates. Preliminary work conducted by our group suggests that this is
achievable, as findings have shown a 1:1:1 mixture of LiTMP, LiDMP and CuBr
will deposit a crystalline Lipshutz-type cuprate [(TMP)Cu(DMP)(Br)Li2 · 2THF]2
(Figure 9.3). These studies could be extended to include 2-methylpiperidine and,
if attempts to form homoleptic amidocuprates are successful, piperidine, trans-2,6-
dimethylpiperidine and 2,2,6-trimethylpiperidine. This would present an opportu-
nity to gain the ability to tune the properties of amidocuprates in order to improve
their efficiency in DoCu reactions.
Figure 9.3: Structure of [(TMP)Cu(DMP)(Br)Li2 · 2THF]2
9.2 Further Phosphidocuprate Species
The Gilman-type cuprate [(Cy2P)2CuLi · 2THF], 11, represents the most promis-
ing phosphidocuprate compound to result from the work presented here, and one
possibility for further research would be to investigate the generality with which
analogues can be synthesised from a variety of phosphine substrates. In addition,
the propensity for 11 to include LiX (X = CN, Hal) in the solid-state is of interest.
As the synthesis of 11 was carried out in bulk THF, to isolate a Lipshutz-type
172
9. FURTHER WORK
analogue of 11, strategies could include employing diethyl ether as the donor sol-
vent since it is less likely than THF to abstract LiCN from the reaction mixture, or
using only a stoichiometric amount of donor solvent in the reaction mixture. Both
of these strategies have led to the formation of Lipshutz-type bis(amido)cuprate
compounds. Even though 11 was found to be an ineffective DoCu agent, an iso-
lated Lipshutz-type phosphidocuprate may be a superior source of a monomeric
Gilman-type species and therefore any Lipshutz-type phosphidocuprates that are
characterised in the solid-state should be tested for their DoCu ability.
The structures of Ph2PCu(CN)Li · 2THF, 10, and [(Ph2P)6Cu4][Li · 4THF]2, 9,
are interesting examples of phosphidocopper species and it would be helpful to
optimise their syntheses in order that they become more elegant and less wasteful.
However, the tendency of reaction mixtures to which LiPPh2 has been added to
deposit 10 and 9 highlights some of the challenges faced in isolating Gilman- and
Lipshutz-type phosphidocuprates. In taking this research forward there are two
clear obstacles to overcome. The first is the tendency for the products to form
in stoichiometries disproportionate to the ratio of reagents added to the reaction
mixture. This stoichiometric disparity could be explained by the poor atmospheric-
and photostability of diphenylphosphine. To check the validity of this theory, reac-
tions could be carried out with both one and two equivalents of the substrate from
a fresh source, or the degradation of the substrate could be regularly checked by
NMR spectroscopy and the volume adjusted accordingly. Another solution could be
to add the CuX (X = CN, Hal) slurry to the lithiated phosphine solution, ensuring
that LiPR2 is always in excess which may prevent a 1:1 reaction occurring initially.
The second obstacle is that the products get trapped in a thermodynamic sink. It
was found that the reaction mixtures were often insoluble at room temperature,
and required gentle heating to form a solution prior to crystallisation. Strategies
to explore would be the careful choice of solvent systems, using alternative copper
sources (CuCl, CuBr) or filtering the reaction mixture in preference to heating.
An alternative method to forming bis(phosphido)cuprates would be to combine a
solution of pre-isolated crystals of 10 with a further equivalent of LiPPh2, which
also suggests a possible route to heteroleptic bis(phosphido)cuprates by adding
LiPR2 (R 6= Ph) (Scheme 9.1). On a similar theme, a more radical route of in-
vestigation would be to revert to previous known synthetic methods employed by
Martin129 and Cowley.140 This would involve the formation of a phosphidocopper
species followed by reaction with an equivalent of LiPR2. The resource intensive na-
ture of these syntheses contrasts with the facile syntheses presented in this thesis
173
9. FURTHER WORK
which has precluded their use thus far.
+ LiPPh2 (Ph2P)2Cu(CN)Li2.nTHF
+ LiPR2 (Ph2P)Cu(PR2)(CN)Li2.nTHF
(a)
(b) 10
10
Scheme 9.1: Possible routes to (a) the homoleptic bis(phosphido)cuprate
(Ph2P)2Cu(CN)Li · nTHF and (b) heteroleptic bis(phosphido)cuprates starting
from pre-isolated crystals of 10
9.3 Further Solution-state Studies
Due to the impressive ability of adduct-type cuprates to effect ortho deprotona-
tion (see Section 7.6), an improvement of the mechanistic understanding of the
process(es) occurring is warranted and a detailed study of the solution behaviour
of the compounds should be carried out. This could involve variable temperature
NMR spectroscopic studies of Lipshutz- and adduct-type cuprates. Another possi-
bility would be to record NMR spectra at set intervals over a period of several hours
in an attempt to discover how stable the different cuprate species are in solution.
While this project was not able to achieve the isolation of a Lipshutz-type bis(phos-
phido)cuprate, solution-state behaviour of phosphidocuprates remains of interest.
By preparing reaction mixtures of 11 + LiCN in various solvent systems (with and
without donor solvents present) in situ NMR spectroscopic studies could be carried
out. These would initially determine the propensity for 11 to include LiCN in the
solution-state which could possibly allow the interconversion between Gilman- and
Lipshutz-type states to be probed.
The formation of amido(phosphido)cuprates would also be of interest for the same
purpose. The use of an amido ligand may help to favour the formation of a Gilman-
or Lipshutz-type cuprate, while the presence of the phosphido ligand would provide
an NMR spectroscopic handle. From the results presented here, the use of HPCy2
and HTMP as the substrates would be a sensible starting point. In addition,
the solid-state structures of such species would provide an interesting avenue of
research in of themselves.
14N and 15N NMR spectroscopic studies on homo- and heteroleptic bis(amido)cu-
prates could also provide useful solution-state data, although they do each have
their drawbacks. The large quadropole moment of 14N and low abundance of 15N
174
9. FURTHER WORK
means that either very symmetrical environments (14N) or high concentrations
and long experiment times (15N) would be required to record a spectrum. To
circumvent these issues 15N-enriched substrates could be employed, although the
high cost involved must be taken into consideration.∗
9.4 Isolation of an ortho-cuprated Intermediate
To help provide solid-state evidence of any intermediates that are involved in DoCu
reactions, the isolation of an ortho-cuprated aryl species is highly desirable. Nu-
merous unsuccessful attempts, involving the use of both Lipshutz-type formulation
cuprate reaction mixtures and pre-isolated crystals of [(TMP)2Cu(Cl)Li2 · THF]2,
12, have been made to isolate such a compound. When X-ray grade crystals were
deposited from the reaction mixtures, they proved to be solvated lithium salts
such as [LiCl ·THF]4 or cross-coupled aromatics such as 2,2
′-dimethoxybiphenyl.200
Reactions carried out in less polar donor solvents, such as diethyl ether, or with
largely non-polar aromatic reagents in non-polar solvents such as toluene or hex-
ane may encourage the precipitation of LiX salts (X = CN, Hal) which could be
removed by filtration prior to crystallisation. Also, attempts thus far have involved
addition of a neat aromatic substrate such as 2-chloropyridine or anisole to the
cuprate base, which was difficult to control as the reactions were run on a 1 or
2 mmol scale, meaning that only very small volumes (<100 µL) of the substrates
were required. Therefore, if the substrate is pre-dissolved and added dropwise to
the cuprate base it may prove to be a more successful method as the addition step
could be more carefully controlled.
9.5 Further Computational Studies
Experimental work has suggested that the solid-state structures of bis(amido)-
cuprates are dependent on the size and nature of the amine ligand. Results have
shown that the use of the bulky TMP ligand affords Lipshutz-type species, whereas
the less sterically demanding DMP ligand affords adduct-type cuprate species. The-
oretical models which probed this phenomenon would be of interest. Calculations
mirroring those displayed in Scheme 7.8 could be repeated, changing the amine
∗ For example, the cheapest available 15N-labelled CuCN costs £626 g−1. Data from
http://www.sigmaaldrich.com, accessed 19th March 2014
175
9. FURTHER WORK
ligand from Me2N to TMP and DMP. If these calculations prove to be too resource
intensive, tBu2N and
iPrN would be possible representatives for TMP and DMP,
respectively, although the fact that the alkyl groups are not “tied-back” into a ring
means that the steric demands may not match well enough. In addition, a com-
parison between the interconversion of structure-types (Lipshutz-adduct-Gilman)
incorporating different X groups (X = CN, Hal) would provide useful informa-
tion. As yet, an adduct-type cuprate containing a cyano group has not as yet
been observed in the solid-state, and these calculations would suggest whether it
is indeed plausible. In this case, several possible binding modes of the cyano group
may need to be considered, for example either σ- or pi-bonded to the unsolvated
Li centre (Figure 9.4). Finally, comparing the relative energies of Lipshutz-type
cuprates containing the range of halides may provide explanations for the different
reactivities observed (as discussed in Chapter 5).
R2N
Cu
R2
N
Li
R2
N
C
Li
Cu
NR2
Li
SS
N
(a)
R2N
Cu
R2
N
Li
R2
N
Li
Cu
NR2
Li
SS
C N
(b)
Figure 9.4: Schematic representation of how a cyano group may be incorporated
into an adduct-type cuprate via (a) σ-bonding and (b) pi-bonding to the unsolvated
Li centre
9.6 Applications in Directed ortho Metalation
The results in Section 7.6 prove that DMP-based cuprate species are effective DoCu
agents and warrant further research to find out if they represent a feasible alterna-
tive over their TMP-based analogues. Initially, if [(DMP)2CuLi ·THF]2 can be syn-
thesised reproducibly it could be isolated and reacted with N,N -diisopropylbenz-
amide following the preparation described in Section 3.5. This is predicted to yield
2-iodo-N,N -diisopropylbenzamide, 6, only in low yields in a similar manner to that
observed when [(TMP)2CuLi]2, 2, was employed as the base.
138 From there, a range
176
9. FURTHER WORK
of aromatic substrates can be treated with DMP-based Lipshutz-type formulation
cuprate mixtures (or pre-isolated adduct-type cuprates) to discover the species’
tolerance to a variety of functional groups.
Expanding the horizon beyond cuprates, several zincate and aluminate species con-
taining DMP ligands are already known61,75 (see Sections 1.4 and 1.5) and recently,
two such zincate base mixtures were tested for their DoZn ability. These stud-
ies involved treating N,N -diisopropylbenzamide with tBu2Zn(µ-DMP)Li ·TMEDA
and Et2Zn(µ-DMP)Li · TMEDA. While the former zincate did not furnish N,N -
diisopropylbenzamide with an electrophile at the 2-position in high yields, reactions
with the latter were more promising.203 Although yields were lower than when
an analogous TMP-zincate base was employed, which was also true for analogous
DoCu reactions carried out in this project (see Section 7.6), the cost saving far out-
weighs the small decrease in the yield. This reactivity could be explored further,
and in addition the DoAl ability of DMP-based aluminate bases could also be in-
vestigated. Although no solid-state structures of either aluminate or zincate bases
incorporating piperidide are known, the further cost saving associated with this
amide means that its employment in DoZn and DoAl may be of significant interest.
177
Bibliography
[1] P. M. Maitlis, A. Haynes, G. J. Sunley and M. J. Howard, J Chem Soc,
Dalton Trans, pages 2187–2196, (1996).
[2] F. E. Herkes, Catalysis of Organic Reactions, CRC Press, (1998).
[3] W. Keim, F. H. Kowalt, R. Goddard and C. Kru¨ger, Angew Chem, Int Ed
Engl, 17(6); 466–467, (1978).
[4] H. Gilman and R. L. Bebb, J Am Chem Soc, 61(1); 109–112, (1939).
[5] G. Wittig and G. Fuhrmann, Chem Ber, 73(11); 1197–1218, (1940).
[6] A. F. G. Goldberg and B. M. Stoltz, Org Lett, 13(16); 4474–4476, (2011).
[7] O. V. Kulikov and A. D. Hamilton, RSC Adv, 2(6); 2454–2461, (2012).
[8] I. R. Baxendale, S. Cheung, M. O. Kitching, S. V. Ley and J. W. Shearman,
Bioorg Med Chem, 21(14); 4378–4387, (2013).
[9] P. Wyatt and S. Warren, Organic Synthesis: Strategy and Control, John
Wiley & Sons, Chichester, England; Hoboken, NJ, (2007).
[10] G. D. Yadav and B. Kundu, Can J Chem Eng, 79(5); 805–811, (2001).
[11] H. Awad, F. Mongin, F. Tre´court, G. Que´guiner, F. Marsais, F. Blanco,
B. Abarca and R. Ballesteros, Tetrahedron Lett, 45(36); 6697–6701, (2004).
[12] Y. Kondo, M. Shilai, M. Uchiyama and T. Sakamoto, J Am Chem Soc,
121(14); 3539–3540, (1999).
[13] V. L. Blair, W. Clegg, B. Conway, E. Hevia, A. R. Kennedy, J. Klett, R. E.
Mulvey and L. Russo, Chem Eur J, 14(1); 65–72, (2008).
[14] H. Naka, M. Uchiyama, Y. Matsumoto, A. E. H. Wheatley, M. McPartlin,
J. V. Morey and Y. Kondo, J Am Chem Soc, 129(7); 1921–1930, (2007).
178
[15] S. Usui, Y. Hashimoto, J. V. Morey, A. E. H. Wheatley and M. Uchiyama,
J Am Chem Soc, 129(49); 15102–15103, (2007).
[16] W. H. Puterbaugh and C. R. Hauser, J Org Chem, 29(4); 853–856, (1964).
[17] J. Clayden, Organolithiums: Selectivity for Synthesis, Tetrahedron Organic
Chemistry Series 23, Pergamon Press, Oxford, (2002).
[18] P. Beak and R. A. Brown, J Org Chem, 47(1); 34–46, (1982).
[19] W. Langham, R. Q. Brewster and H. Gilman, J Am Chem Soc, 63(2); 545–
549, (1941).
[20] G. W. Klumpp and M. J. Sinnige, Tetrahedron Lett, 27(20); 2247–2250,
(1986).
[21] R. R. Fraser, M. Bresse and T. S. Mansour, J Am Chem Soc, 105(26); 7790–
7791, (1983).
[22] D. A. Shirley and J. P. Hendrix, J Organomet Chem, 11; 217–226, (1968).
[23] D. A. Shirley, T. E. Harmon and C. C. Fong, J Organomet Chem, 69(3);
327–344, (1974).
[24] N. J. R. van Eikema Hommes and P. von Rague´ Schleyer, Angew Chem, Int
Ed Engl, 31(6); 755–758, (1992).
[25] J. M. Saa, J. Morey, A. Frontera and P. M. Deya, J Am Chem Soc, 117(3);
1105–1116, (1995).
[26] R. W. Janes and R. A. Palmer, J Mol Struct (Theochem), 339(13); 95–101,
(1995).
[27] S. T. Chadwick, R. A. Rennels, J. L. Rutherford and D. B. Collum, J Am
Chem Soc, 122(36); 8640–8647, (2000).
[28] M. C. Whisler, S. MacNeil, V. Snieckus and P. Beak, Angew Chem, Int Ed,
43(17); 2206–2225, (2004).
[29] R. E. Ludt, J. S. Griffiths, K. N. McGrath and C. R. Hauser, J Org Chem,
38(9); 1668–1674, (1973).
[30] D. B. Reitz and S. M. Massey, J Org Chem, 55(4); 1375–1379, (1990).
[31] M. Anstiss, J. Clayden, A. Grube and L. H. Youssef, Synlett, pages 290–294,
(2002).
179
[32] J. Eames, G. S. Coumbarides, M. J. Suggate and N. Weerasooriya, Eur J
Org Chem, 2003(4); 634–641, (2003).
[33] M. P. Sibi and V. Snieckus, J Org Chem, 48(11); 1935–1937, (1983).
[34] G. Wittig, Angew Chem, 70(3); 65–71, (1958).
[35] C. W. Kamienski and J. F. Eastham, J Org Chem, 34(4); 1116–1121, (1969).
[36] C. W. Kamienski and J. F. Eastham, “Diorganomagnesium Reagents and
Methods of Preparing Same”, (1973).
[37] K. Kitagawa, A. Inoue, H. Shinokubo and K. Oshima, Angew Chem, Int Ed,
39(14); 2481–2483, (2000).
[38] M. Yasuda, M. Ide, Y. Matsumoto and M. Nakata, Synlett, pages 899–902,
(1997).
[39] H. G. Richey and J. Farkas, Organometallics, 9(6); 1778–1784, (1990).
[40] S. Dumouchel, F. Mongin, F. Tre´court and G. Que´guiner, Tetrahedron Lett,
44(19); 3877–3880, (2003).
[41] T. Greiser, J. Kopf, D. Thoennes and E. Weiss, Chem Ber, 114(1); 209–213,
(1981).
[42] B. Schubert and E. Weiss, Chem Ber, 117(1); 366–375, (1984).
[43] E. Frankland, Q J Chem Soc, 2(3); 263–296, (1850).
[44] D. Seyferth, Organometallics, 20(14); 2940–2955, (2001).
[45] T. Harada, D. Hara, K. Hattori and A. Oku, Tetrahedron Lett, 29(31); 3821–
3824, (1988).
[46] T. Katsuhira, T. Harada, K. Maejima, A. Osada and A. Oku, J Org Chem,
58(23); 6166–6168, (1993).
[47] T. Harada, T. Katsuhira, A. Osada, K. Iwazaki, K. Maejima and A. Oku, J
Am Chem Soc, 118(45); 11377–11390, (1996).
[48] Y. Kondo, M. Fujinami, M. Uchiyama and T. Sakamoto, J Chem Soc, Perkin
Trans 1, pages 799–800, (1997).
[49] C. J. Upton and P. Beak, J Org Chem, 40(8); 1094–1098, (1975).
[50] M. Uchiyama, T. Miyoshi, Y. Kajihara, T. Sakamoto, Y. Otani, T. Ohwada
and Y. Kondo, J Am Chem Soc, 124(29); 8514–8515, (2002).
180
[51] W. Clegg, S. H. Dale, E. Hevia, G. W. Honeyman and R. E. Mulvey, Angew
Chem, Int Ed, 45(15); 2370–2374, (2006).
[52] D. V. Graham, E. Hevia, A. Kennedy and R. E. Mulvey, Organometallics,
25(14); 3297–3300, (2006).
[53] W. Clegg, S. H. Dale, R. W. Harrington, E. Hevia, G. W. Honeyman and
R. E. Mulvey, Angew Chem, Int Ed, 45(15); 2374–2377, (2006).
[54] M. Uchiyama, Y. Matsumoto, D. Nobuto, T. Furuyama, K. Yamaguchi and
K. Morokuma, J Am Chem Soc, 128(27); 8748–8750, (2006).
[55] M. Uchiyama, Y. Matsumoto, S. Usui, Y. Hashimoto and K. Morokuma,
Angew Chem, Int Ed, 46(6); 926–929, (2007).
[56] Y. Kondo, J. V. Morey, J. Morgan, H. Naka, D. Nobuto, P. R. Raithby,
M. Uchiyama and A. E. H. Wheatley, J Am Chem Soc, 129(42); 12734–
12738, (2007).
[57] W. Clegg, B. Conway, E. Hevia, M. D. McCall, L. Russo and R. E. Mulvey,
J Am Chem Soc, 131(6); 2375–2384, (2009).
[58] R. Campbell, P. Garc´ıa-A´lvarez, A. R. Kennedy and R. E. Mulvey, Chem
Eur J, 16(33); 9964–9968, (2010).
[59] R. Campbell, D. Cannon, P. Garc´ıa-A´lvarez, A. R. Kennedy, R. E. Mulvey,
S. D. Robertson, J. Saßmannshausen and T. Tuttle, J Am Chem Soc, 133(34);
13706–13717, (2011).
[60] A. R. Kennedy, J. Klett, R. E. Mulvey and D. S. Wright, Science, 326(5953);
706–708, (2009).
[61] R. Campbell, B. Conway, G. S. Fairweather, P. Garc´ıa-A´lvarez, A. R.
Kennedy, J. Klett, R. E. Mulvey, C. T. O’Hara and G. M. Robertson, Dalton
Trans, 39(2); 511–519, (2010).
[62] J. R. Lennox, S. C. Turner and H. Rapoport, J Org Chem, 66(21); 7078–7083,
(2001).
[63] D. R. Williams and S. Y. Sit, J Am Chem Soc, 106(10); 2949–2954, (1984).
[64] D. R. Armstrong, W. Clegg, S. H. Dale, J. Garc´ıa-A´lvarez, R. W. Harrington,
E. Hevia, G. W. Honeyman, A. R. Kennedy, R. E. Mulvey and C. T. O’Hara,
Chem Commun, pages 187–189, (2008).
181
[65] W. Hallwachs and A. Schafarik, Liebigs Ann, 109(2); 206–209, (1859).
[66] J. Boor, Ziegler Natta Catalysts and Polymerisations, Academic Press, New
York, (1979).
[67] A. Yasuda, S. Tanaka, K. Oshima, H. Yamamoto and H. Nozaki, J Am Chem
Soc, 96(20); 6513–6514, (1974).
[68] K. Maruoka, M. Oishi and H. Yamamoto, J Org Chem, 58(27); 7638–7639,
(1993).
[69] J. J. Eisch, Comprehensive Organometallic Chemistry, Vol. 6, Pergamon,
Oxford, (1982).
[70] H. Hart, “Chapter 18 - Arynes and Heteroarynes”, in S. Patai (Editor), The
Chemistry of Triply-Bonded Functional Groups, Supplement C2, volume 2,
Wiley, Chichester, UK, (1994).
[71] M. Uchiyama, H. Naka, Y. Matsumoto and T. Ohwada, J Am Chem Soc,
126(33); 10256–10257, (2004).
[72] J. Garc´ıa-A´lvarez, D. V. Graham, A. R. Kennedy, R. E. Mulvey and
S. Weatherstone, Chem Commun, pages 3208–3210, (2006).
[73] H. Naka, J. V. Morey, J. Haywood, D. Eisler, M. McPartlin, F. Garcia,
H. Kudo, Y. Kondo, M. Uchiyama and A. E. H. Wheatley, J Am Chem Soc,
130(48); 16193–16200, (2008).
[74] B. Conway, E. Hevia, J. Garc´ıa-A´lvarez, D. V. Graham, A. R. Kennedy and
R. E. Mulvey, Chem Commun, pages 5241–5243, (2007).
[75] B. Conway, P. Garc´ıa-A´lvarez, A. R. Kennedy, J. Klett, R. E. Mulvey and
S. D. Robertson, New J Chem, 34(8); 1707–1712, (2010).
[76] J. Garc´ıa-A´lvarez, A. R. Kennedy, J. Klett and R. E. Mulvey, Angew Chem,
Int Ed, 46(7); 1105–1108, (2007).
[77] L. M. Carrella, W. Clegg, D. V. Graham, L. M. Hogg, A. R. Kennedy,
J. Klett, R. E. Mulvey, E. Rentschler and L. Russo, Angew Chem, Int Ed,
46(25); 4662–4666, (2007).
[78] V. L. Blair, W. Clegg, R. E. Mulvey and L. Russo, Inorg Chem, 48(18);
8863–8870, (2009).
[79] M. Mosrin and P. Knochel, Org Lett, 11(8); 1837–1840, (2009).
182
[80] S. H. Wunderlich and P. Knochel, Angew Chem, Int Ed, 48(8); 1501–1504,
(2009).
[81] S. A. Wilks, H. Michels and C. W. Keevil, Int J Food Microbiol, 105(3);
445–454, (2005).
[82] C. Powell and H. Stillman, “Corrosion Behaviour of Copper Alloys used
in Marine Aquaculture”, http://www.copper.org/applications/marine/cuni/
pdf/marine aquaculture.pdf, (2009), Accessed 24th October, 2013.
[83] T. P. Schultz, D. D. Nicholas and W. P. Henry, Holzforschung, 59(3); 370–
373, (2005).
[84] J. S. Lewis, J. L. J. Dearling, J. K. Sosabowski, J. Zweit, P. Carnochan,
L. R. Kelland, H. M. Coley and P. J. Blower, Eur J Nucl Med, 27(6); 638–
646, (2000).
[85] P. Abdulkin, Y. Moglie, B. R. Knappett, D. A. Jefferson, M. Yus, F. Alonso
and A. E. H. Wheatley, Nanoscale, 5(1); 342–350, (2013).
[86] R. Bo¨ttiger, Liebigs Ann, 109(3); 351–362, (1859).
[87] R. Reich, C R Acad Sci, 177; 322–324, (1923).
[88] M. A. Willert-Porada, D. J. Burton and N. C. Baenziger, J Chem Soc, Chem
Commun, pages 1633–1634, (1989).
[89] J. P. Snyder, J Am Chem Soc, 117(44); 11025–11026, (1995).
[90] S. H. Bertz, S. Cope, M. Murphy, C. A. Ogle and B. J. Taylor, J Am Chem
Soc, 129(23); 7208–7209, (2007).
[91] G. M. Whitesides, J. San Filippo, C. P. Casey and E. J. Panek, J Am Chem
Soc, 89(20); 5302–5303, (1967).
[92] B. H. Lipshutz, R. S. Wilhelm and D. M. Floyd, J Am Chem Soc, 103(25);
7672–7674, (1981).
[93] N. Krause, Modern Organocopper Chemistry, Wiley-VCH Verlag GmbH,
(2002).
[94] J. A. J. Jarvis, R. Pearce and M. F. Lappert, J Chem Soc, Dalton Trans,
pages 999–1003, (1977).
[95] T. Fujihara, T. Xu, K. Semba, J. Terao and Y. Tsuji, Angew Chem, Int Ed,
50(2); 523–527, (2011).
183
[96] T. Vergote, F. Nahra, A. Welle, M. Luhmer, J. Wouters, N. Mager, O. Riant
and T. Leyssens, Chem Eur J, 18(3); 793–798, (2012).
[97] A. Miyashita, T. Yamomoto and A. Yamomoto, Bull Chem Soc Jpn, 50;
1109–1117, (1977).
[98] G. van Koten, S. L. James and J. T. B. H. Jastrzebski, “2 - Copper and
Silver”, in E. W. Abel, F. G. A. Stone and G. Wilkinson (Editors), Com-
prehensive Organometallic Chemistry II, volume 3, pages 57–133, Elsevier,
Oxford, (1995).
[99] T. Ikariya and A. Yamamoto, J Organomet Chem, 74(1); 145–151, (1974).
[100] G. van Koten and J. T. B. H. Jastrzebski, Tetrahedron, 45(2); 569–578,
(1989).
[101] S. Gambarotta, C. Floriani, A. Chise-Villa and C. Guastini, J Chem Soc,
Chem Commun, pages 1156–1158, (1983).
[102] E. M. Meyer, S. Gambarotta, C. Floriani, A. Chise-Villa and C. Guastini,
Organometallics, 8(4); 1067–1079, (1989).
[103] D. Nobel, G. van Koten and A. L. Spek, Angew Chem, Int Ed Engl, 28(2);
208–210, (1989).
[104] X. He, M. M. Olmstead and P. P. Power, J Am Chem Soc, 114(24); 9668–
9670, (1992).
[105] H. Gilman, R. G. Jones and L. A. Woods, J Org Chem, 17(12); 1630–1634,
(1952).
[106] H. O. House, W. L. Respess and G. M. Whitesides, J Org Chem, 31(10);
3128–3141, (1966).
[107] G. M. Whitesides, W. F. Fisher, J. San Filipo, R. W. Bashe and H. O. House,
J Am Chem Soc, 91(17); 4871–4882, (1969).
[108] P. G. Edwards, R. W. Gellert, M. W. Marks and R. Bau, J Am Chem Soc,
104(7); 2072–2073, (1982).
[109] H. Hope, D. Oram and P. P. Power, J Am Chem Soc, 106(4); 1149–1150,
(1984).
[110] H. Hope, M. M. Olmstead, P. P. Power, J. Sandell and X. Xu, J Am Chem
Soc, 107(14); 4337–4338, (1985).
184
[111] C. Eaborn, P. B. Hitchcock, J. Smith and A. C. Sullivan, J Organomet Chem,
263(2); C23–C25, (1984).
[112] M. John, C. Auel, C. Behrens, M. Marsch, K. Harms, F. Bosold, R. M.
Gschwind, P. R. Rajamohanan and G. Boche, Chem Eur J, 6(16); 3060–
3068, (2000).
[113] M. M. Olmstead and P. P. Power, J Am Chem Soc, 112(22); 8008–8014,
(1990).
[114] N. P. Lorenzen and E. Weiss, Angew Chem, Int Ed Engl, 29(3); 300–302,
(1990).
[115] S. H. Bertz and G. Dabbagh, Tetrahedron, 45(2); 425–434, (1989).
[116] M. S. Kharasch and P. O. Tawney, J Am Chem Soc, 63(9); 2308–2316, (1941).
[117] G. van Koten, J. T. B. H. Jastrzebski and J. G. Noltes, J Organomet Chem,
140(2); C23–C27, (1977).
[118] H. O. House, D. G. Koepsell and W. J. Campbell, J Org Chem, 37(7); 1003–
1011, (1972).
[119] S. H. Bertz, J Am Chem Soc, 112(10); 4031–4032, (1990).
[120] B. H. Lipshutz, S. Sharma and E. L. Ellsworth, J Am Chem Soc, 112(10);
4032–4034, (1990).
[121] J. P. Snyder, D. P. Spangler, J. R. Behling and B. E. Rossiter, J Org Chem,
59(10); 2665–2667, (1994).
[122] B. H. Lipshutz and B. James, J Org Chem, 59(25); 7585–7587, (1994).
[123] J. P. Snyder and S. H. Bertz, J Org Chem, 60(14); 4312–4313, (1995).
[124] C. M. P. Kronenberg, J. T. B. H. Jastrzebski, A. L. Spek and G. van Koten,
J Am Chem Soc, 120(37); 9688–9689, (1998).
[125] G. Boche, F. Bosold, M. Marsch and K. Harms, Angew Chem, Int Ed, 37(12);
1684–1686, (1998).
[126] M. Ito, D. Hashizume, T. Fukunaga, T. Matsuo and K. Tamao, J Am Chem
Soc, 131(50); 18024–18025, (2009).
[127] A. P. Purdy and C. F. George, Polyhedron, 14(6); 761–769, (1995).
185
[128] R. P. Davies, S. Hornauer and P. B. Hitchcock, Angew Chem, Int Ed, 46(27);
5191–5194, (2007).
[129] S. F. Martin, J. R. Fishpaugh, J. M. Power, D. M. Giolando, R. A. Jones,
C. M. Nunn and A. H. Cowley, J Am Chem Soc, 110(21); 7226–7228, (1988).
[130] C. S. Hwang and P. P. Power, Organometallics, 18(4); 697–700, (1999).
[131] P. Reiß and D. Fenske, Z Anorg Allg Chem, 626(6); 1317–1331, (2000).
[132] C. M. P. Kronenburg, C. H. M. Amijs, J. T. B. H. Jastrzebski, M. Lutz, A. L.
Spek and G. van Koten, Organometallics, 21(22); 4662–4671, (2002).
[133] R. Davies, S. Hornauer and A. White, Chem Commun, pages 304–306, (2007).
[134] M. Yamanaka and E. Nakamura, J Am Chem Soc, 127(13); 4697–4706,
(2005).
[135] J.-M. L’Helgoual’ch, G. Bentabed-Ababsa, F. Chevallier, M. Yonehara,
M. Uchiyama, A. Derdour and F. Mongin, Chem Commun, pages 5375–5377,
(2008).
[136] T. T. Nguyen, F. Chevallier, V. Jouikov and F. Mongin, Tetrahedron Lett,
50(49); 6787–6790, (2009).
[137] T. T. Nguyen, N. Marquise, F. Chevallier and F. Mongin, Chem Eur J,
17(37); 10405–10416, (2011).
[138] S. Komagawa, S. Usui, J. Haywood, P. J. Harford, A. E. H. Wheatley,
Y. Matsumoto, K. Hirano, R. Takita and M. Uchiyama, Angew Chem, Int
Ed, 51(48); 12081–12085, (2012).
[139] J. Haywood, J. V. Morey, A. E. H. Wheatley, C. Y. Liu, S. Yasuike, J. Kurita,
M. Uchiyama and P. R. Raithby, Organometallics, 28(1); 38–39, (2009).
[140] A. H. Cowley, D. M. Giolando, R. A. Jones, C. M. Nunn and J. M. Power, J
Chem Soc, Chem Commun, pages 208–209, (1988).
[141] S. H. Bertz and G. Dabbagh, J Chem Soc, Chem Commun, pages 1030–1032,
(1982).
[142] S. H. Bertz, G. Dabbagh and G. Villacorta, J Am Chem Soc, 104(21); 5824–
5826, (1982).
[143] J. J. Eisch, Organometallic Syntheses Vol. 2: Nontransition-metal com-
pounds, Academic Press Inc., New York, 1st edition, (1981).
186
[144] D. F. Schriver and M. A. Drezdzon, The Manipulation of Air-Sensitive Com-
pounds, John Wiley & Sons, New York, 2nd edition, (1986).
[145] T. Kottke and D. Stalke, J Appl Cryst, 26; 615–619, (1993).
[146] G. Sheldrick, Acta Crystallogr, Sect A, 46; 467–473, (1990).
[147] G. Sheldrick, “SHELXL-97 Program for Crystal Structure Refinement”,
(1997).
[148] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb,
J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Peters-
son, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov,
J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota,
R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao,
H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro,
M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov,
R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S.
Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox,
J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Strat-
mann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L.
Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dan-
nenberg, S. Dapprich, A. D. Daniels, O¨. Farkas, J. B. Foresman, J. V. Ortiz,
J. Cioslowski and D. J. Fox, “Gaussian 09 Revision C.01”, Gaussian Inc.
Wallingford CT (2010).
[149] A. D. Becke, Phys Rev, A38(6); 3098–3100, (1988).
[150] A. D. Becke, J Chem Phys, 98(2); 1372–1377, (1993).
[151] A. D. Becke, J Chem Phys, 98(7); 5648–5652, (1993).
[152] C. Lee, W. Yang and R. G. Parr, Phys Rev, B37(2); 785–788, (1988).
[153] A. Scha¨fer, H. Horn and R. Ahlrichs, J Chem Phys, 97(4); 2571–2577, (1992).
[154] K. Fukui, Acc Chem Res, 14(12); 363–368, (1981).
[155] K. Ishida, K. Morokuma and A. Komornicki, J Chem Phys, 66(5); 2153–2156,
(1977).
[156] C. Gonzalez and H. B. Schlegel, J Chem Phys, 90(4); 2154–2161, (1989).
[157] H. B. Schlegel and C. Gonzalez, J Phys Chem, 94(14); 5523–5527, (1990).
187
[158] Y. Suzuki, T. Yoshino, K. Moriyama and H. Togo, Tetrahedron, 67(21); 3809–
3814, (2011).
[159] J.-N. Li, L. Liu, Y. Fu and Q.-X. Guo, Tetrahedron, 62(18); 4453–4462,
(2006).
[160] R. A. Alden, J. Kraut and T. G. Traglor, J Am Chem Soc, 90(1); 74–82,
(1968).
[161] G. Rindorf, F. S. Jørgensen and J. P. Snyder, J Org Chem, 45(26); 5343–5347,
(1980).
[162] I. G. Dance, Polyhedron, 2(10); 1031–1043, (1983).
[163] S. S. Dhingra and R. C. Haushalter, J Am Chem Soc, 116(8); 3651–3652,
(1994).
[164] S. Zeevi and E. Y. Tshuva, Eur J Inorg Chem, pages 5369–5376, (2007).
[165] A. Banerjee, R. Singh, P. Mondal, E. Colacio and K. K. Rajak, Eur J Inorg
Chem, pages 790–798, (2010).
[166] G. A. Bowmaker, G. R. Clark and D. K. P. Yuen, J Chem Soc, Dalton Trans,
pages 2329–2334, (1976).
[167] N. P. Rath and E. M. Holt, J Chem Soc, Chem Commun, pages 665–667,
(1985).
[168] C. Hasselgren and S. Jagner, Inorg Chim Acta, 336; 137–141, (2002).
[169] H. Zhao, Z.-R. Qu, Q. Ye, B. Abrahams, Y.-P. Wang, Z.-G. Liu, Z. Xue,
R.-G. Xiong and X.-Z. You, Chem Mater, 15(22); 4166–4168, (2003).
[170] B. Nohra, E. Rodriguez-Sanz, C. Lescop and R. Re´au, Chem Eur J, 14(11);
3391–3403, (2008).
[171] J. Schno¨dt, R. Winter and M. Zabel, Z Anorg Allg Chem, 636(7); 1242–1248,
(2010).
[172] M. A. Beswick, P. R. Raithby, C. A. Russell, A. Steiner, K. L. Verhorevoort,
K. L. Ward and D. S. Wright, Angew Chem, Int Ed Engl, 34(23-24); 2662–
2664, (1996).
[173] C.-S. Hwang and P. P. Power, J Am Chem Soc, 120(17); 6409–6410, (1990).
[174] C. Eaborn, M. S. Hill, P. B. Hitchcock and J. D. Smith, Organometallics,
19(26); 5780–5783, (2000).
188
[175] C. Eaborn, S. M. El-Hamruni, M. S. Hill, P. B. Hitchcock and J. D. Smith,
J Chem Soc, Dalton Trans, pages 3975–3979, (2002).
[176] R. A. Bartlett, M. M. Olmstead and P. P. Power, Inorg Chem, 25(8); 1243–
1247, (1986).
[177] R. P. Davies and S. Hornauer, Eur J Inorg Chem, pages 51–54, (2005).
[178] C. A. Tolman, J Am Chem Soc, 92(10); 2956–2965, (1970).
[179] R. K. Dieter and M. Tokles, J Am Chem Soc, 109(7); 2040–2046, (1987).
[180] B. E. Rossiter, M. Eguchi, G. Miao, N. M. Swingle, A. E. Herna´ndez, D. Vick-
ers, E. Fluckiger, R. G. Patterson and K. V. Reddy, Tetrahedron, 49(5);
965–986, (1993).
[181] K. Sne´garoff, T. T. Nguyen, N. Marquise, Y. S. Halauko, P. J. Harford,
T. Roisnel, V. E. Matulis, O. A. Ivashkevich, F. Chevallier, A. E. H. Wheat-
ley, P. C. Gros and F. Mongin, Chem Eur J, 17(47); 13284–13297, (2011).
[182] F. Tre´court, F. Marsais, T. Gu¨ngo¨r and G. Que´guiner, J Chem Soc, Perkin
Trans 1, pages 2409–2415, (1990).
[183] M. Mallet, J Organomet Chem, 406(12); 49–56, (1991).
[184] Y.-C. Wu, Heterocycles, 29(3); 463–475, (1989).
[185] S. Prachayasittikul, P. Manam, M. Chinworrungsee, C. Isarankura-Na-
Ayudhya, S. Ruchirawat and V. Prachayasittikul, Molecules, 14(11); 4414–
4424, (2009).
[186] D. Mueller, R. A. Davis, S. Duffy, V. M. Avery, D. Camp and R. J. Quinn,
J Nat Prod, 72(8); 1538–1540, (2009).
[187] N. Marquise, P. J. Harford, F. Chevallier, T. Roisnel, A. E. H. Wheatley,
P. C. Gros and F. Mongin, Tetrahedron Lett, 54(24); 3154–3157, (2013).
[188] K. J. Stauffer, P. D. Williams, H. G. Selnick, P. G. Nantermet, C. L. Newton,
C. F. Homnick, M. M. Zrada, S. D. Lewis, B. J. Lucas, J. A. Krueger, B. L.
Pietrak, E. A. Lyle, R. Singh, C. Miller-Stein, R. B. White, B. Wong, A. A.
Wallace, G. R. Sitko, J. J. Cook, M. A. Holahan, M. Stranieri-Michener,
Y. M. Leonard, J. J. Lynch, D. R. McMasters and Y. Yan, J Med Chem,
48(7); 2282–2293, (2005).
189
[189] N. N. Greenwood, Chemistry of the Elements, Butterworth-Heinemann, Ox-
ford, 2nd edition, (1997).
[190] P. P. Power, K. Ruhlandt-Senge and S. C. Shoner, Inorg Chem, 30(26); 5013–
5015, (1991).
[191] R. Giri and J. F. Hartwig, J Am Chem Soc, 132(45); 15860–15863, (2010).
[192] C.-K. Tseng, C.-R. Lee, C.-C. Han and S.-G. Shyu, Chem Eur J, 17(9);
2716–2723, (2011).
[193] “CRC Handbook of Chemistry and Physics 94th Ed.”, http://www.
hbcpnetbase.com, (2013)–(2014), Accessed 21st February, 2014.
[194] N. P. Lorenzen and E. Weiss, Angew Chem, Int Ed Engl, 29(3); 300–302,
(1990).
[195] F. Olbrich, J. Kopf and E. Weiss, Angew Chem, Int Ed Engl, 32(7); 1077–
1079, (1993).
[196] F. Pauer and P. P. Power, “Structures of Lithium Salts of Heteroatom Com-
pounds: Lithium Compounds of the Main Group 5 Elements”, in A.-M. Sapse
and P. von Rague´ Schleyer (Editors), Lithium Chemistry: A Theoretical and
Experimental Overview, pages 328–340, Wiley, New York; Chichester, (1995).
[197] N. Sievert, A. Fischer, U. Klingebiel, A. Pal and M. Noltemeyer, Z Anorg
Allg Chem, 633(8); 1223–1232, (2007).
[198] P. Rosa, N. Me´zailles, L. Ricard, F. Mathey and P. Le Floch, Angew Chem,
Int Ed, 39(10); 1823–1826, (2000).
[199] M. Brym, C. Jones, P. C. Junk and M. Kloth, Z Anorg Allg Chem, 632(6);
1402–1404, (2006).
[200] P. J. Harford, J. P. Taylor, A. J. Peel and A. E. H. Wheatley, unpublished
data.
[201] M. F. A. Adamo, V. K. Aggarwal and M. A. Sage, Synth Commun, 29(10);
1747–1756, (1999).
[202] D. M. Hodgson and N. S. Kaka, Angew Chem, Int Ed, 47(51); 9958–9960,
(2008).
190
[203] D. R. Armstrong, J. A. Garden, A. R. Kennedy, S. M. Leenhouts, R. E.
Mulvey, P. O’Keefe, C. T. O’Hara and A. Steven, Chem Eur J, 19(40); 13492–
13503, (2013).
191