TMP (2,2,6,6-tetramethylpiperidide)-aluminate bases: Lithium-mediated alumination or lithiation-alkylaluminium-trapping reagents?

Article abstract of DOI:10.1039/c4sc01108b

The lithium TMP-aluminate bases "LiTMP·Al(ⁱBu) ₃" 1 and "LiTMP·Al(TMP)(ⁱBu) ₂" 2, where TMP is 2,2,6,6-tetramethylpiperidide, have recently come under the spotlight as "aluminating" reagents in that they can perform aluminium-hydrogen exchange on a wide variety of aromatic substrates. Previous studies have intimated that 1 existed as a single species in THF solution formulated as [(THF)·Li(μ-TMP)(μ-ⁱBu)Al( ⁱBu)₂] 1·THF, having a contacted ion pair structure as evidenced by an X-ray crystallographic study of isolated crystals. But here using anisole as a case substrate it is revealed that pre-crystallised 1·THF cannot deprotonate anisole at all whether in hexane or THF solution contradicting earlier in situ applications of 1 which revealed near quantitative metallation of anisole. NMR spectroscopic studies of 1 made in situ in THF solution ascribe this reactivity distinction from 1·THF to complex equilibria involving five major species in LiTMP·THF, Al( ⁱBu)₃·THF, [{Li(THF)₄} ⁺{Al(TMP)(ⁱBu)₃}^-] 1·(THF)₄, [(THF)·Li(μ-TMP)(μ-OC₄H ₇)Al(ⁱBu)₂], 4, and (TMP)Al( ⁱBu)₂·THF. Reagent 2 in contrast is found to exist as only two separated homometallic species in LiTMP·THF and (TMP)Al( ⁱBu)₂·THF in THF solution. The constitutions of 1 and 2 in non-polar hexane solution are also revealed. With the aid of DFT calculations, discussion focuses on the fact that none of the aluminate species present in THF solutions of 1 or 2 can deprotonate/metallate anisole, instead the metallation processes appear to be LiTMP lithiations followed immediately by trapping by an alkylaluminium complex, in a metal exchange which drives the reaction to the product (arylaluminated) side. the Partner Organisations 2014.

Full text of DOI:10.1039/c4sc01108b

Chemical
Science
View Article Online
View Journal | View Issue
EDGE ARTICLE
TMP (2,2,6,6-tetramethylpiperidide)-aluminate
bases: lithium-mediated alumination or lithiation–
alkylaluminium-trapping reagents?†‡
Cite this: Chem. Sci., 2014, 5, 3031
David R. Armstrong, Elaine Crosbie, Eva Hevia, Robert E. Mulvey,* Donna L. Ramsay*
and Stuart D. Robertson
i
i
3 2
The lithium TMP-aluminate bases “LiTMP$Al( Bu) ” 1 and “LiTMP$Al(TMP)( Bu) ” 2, where TMP is 2,2,6,6-
tetramethylpiperidide, have recently come under the spotlight as “aluminating” reagents in that they can
perform aluminium–hydrogen exchange on a wide variety of aromatic substrates. Previous studies have
i
i
intimated that 1 existed as a single species in THF solution formulated as [(THF)$Li(m-TMP)(m- Bu)Al( Bu)
2
]
1$THF, having a contacted ion pair structure as evidenced by an X-ray crystallographic study of isolated
crystals. But here using anisole as a case substrate it is revealed that pre-crystallised 1$THF cannot
deprotonate anisole at all whether in hexane or THF solution contradicting earlier in situ applications of 1
which revealed near quantitative metallation of anisole. NMR spectroscopic studies of 1 made in situ in
THF solution ascribe this reactivity distinction from 1$THF to complex equilibria involving five major
i
+
i
ꢀ
3 4 4 4 7
species in LiTMP$THF, Al( Bu) $THF, [{Li(THF) } {Al(TMP)( Bu) } ] 1$(THF) , [(THF)$Li(m-TMP)(m-OC H )-
3
i
i
Al( Bu)
2
], 4, and (TMP)Al( Bu)
2
$THF. Reagent 2 in contrast is found to exist as only two separated
i
homometallic species in LiTMP$THF and (TMP)Al( Bu)
2
$THF in THF solution. The constitutions of 1 and 2
in non-polar hexane solution are also revealed. With the aid of DFT calculations, discussion focuses on
the fact that none of the aluminate species present in THF solutions of 1 or 2 can deprotonate/metallate
anisole, instead the metallation processes appear to be LiTMP lithiations followed immediately by
trapping by an alkylaluminium complex, in a metal exchange which drives the reaction to the product
(arylaluminated) side.
Received 15th April 2014
Accepted 7th May 2014
DOI: 10.1039/c4sc01108b
www.rsc.org/chemicalscience
these other metals form less polar and consequently less reac-
tive organometallic compounds than organolithium reagents so
Introduction
The changing landscape of metallation over the past decade or activation is required to adapt them for metallation applica-
so has seen the growth of a forest of new metallating agents. tions. Two types of activation are common. Stoichiometric
Now metals such as magnesium, zinc, and aluminium, in lithium chloride can be added to fashion mixed organometallic-
particular, and copper and manganese to a lesser extent, stand salt systems typied by the turbo-Hauser reagent (TMP)
3
tall beside lithium as capable of executing metal–hydrogen MgCl$LiCl (TMP is 2,2,6,6-tetramethylpiperidide). Though salt
exchange on a myriad of aromatic and heteroaromatic additive effects have long been recognised, Knochel has mas-
substrates. Moreover these fundamentally important reactions terminded a remarkable row of reagents of this type based upon
4
of this new set of metal reagents can oen offer general them. Activation can also be realised through mixed organo-
advantages (most signicantly, improved functional group metallic–organometallic systems where one metal is an alkali
tolerance, milder reaction conditions, greater compatibility metal and the second is one of the aforementioned nominally
5
with tandem transition metal catalysed bond forming strate- less reactive metals. Amido–alkyl combinations typied by
t
n
1
gies) over those executed by long established lithium alkyl and LiTMP$Zn( Bu)
(ref. 6) and (TMEDA)Na( Bu)$Mg(TMP) (ref. 7)
2
2
2
lithium amide reagents. Less electropositive than lithium, which can also be regarded as ates (zincate and magnesiate
respectively) have proved the best metallating agents in this
category – note all alkyl combinations show a greater tendency
8
WestCHEM, Department of Pure and Applied Chemistry, University of Strathclyde, for nucleophilic addition. While in metallation reactions
Glasgow, G1 1XL, UK. E-mail: r.e.mulvey@strath.ac.uk; Fax: +44(0) 141-548-4787
the efficiency and scope of these new activated systems
†
Electronic supplementary information (ESI) available. See DOI:
0.1039/c4sc01108b
have generally been well studied, by comparison denite
information on them in their own right has been rather thin on
the ground prompting some to be likened to black box
1
‡
Dedicated to the memory of Mike Lappert, an inorganic chemist extraordinaire,
whose remarkable record of research has led to numerous breakthroughs all over
the periodic table.
9
reagents.
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 3031–3045 | 3031

View Article Online
Chemical Science
Edge Article
A particularly attractive branch of this multicomponent ate
chemistry is alkali–metal-mediated alumination (AMMAl) due to
the high abundance, comparative cheapness, low toxicity and
recycling opportunities of the group 13 metal as well as the
10
documented halogen tolerance of lithium aluminates. It was in
the course of comparing the two reagents that dominate AMMAl
i
chemistry, Uchiyama and Wheatley's “LiTMP$Al( Bu)
3
” 1 (ref. 11)
” 2 (ref. 12)
i
and our own bis-TMP version “LiTMP$Al(TMP)( Bu)
2
13
in reactions with polydentate Lewis bases, and stimulated
further by Knochel's intriguing report of the turbo-Hauser base
analogue “TMPMgCl$LiCl$AlEt ”, that we had cause to revisit
14
3
these key reagents in their own right. Though both have been
studied previously the direct comprehensive comparison
between them made here through new NMR (including DOSY)
spectroscopic studies, reactivity observations and DFT calcula-
tions uncovers several remarkable and surprising ndings that
provide a more complete picture of these complicated multi-
Scheme 1 Lewis base mediated co-complexation reactions of LiTMP
i
and Al( Bu)
3
in bulk hexane solution.
component base mixtures. Specically doubt is cast on in situ 1 However the base mixture 1 is prepared and utilised in a vast
being a single species in THF solution and on the existence of 2 excess of THF in its synthetic applications so the question
as a mixed-metal species either in hexane or THF solution. The requiring an answer is “does 1$THF represent the experimental
most extraordinary revelation from this work is that neither 1 nor base in the THF solution mixture of 1?”
2
appears capable of “aluminating” substrates in THF solution
through a lithium aluminate species, so calling into question the prepared 1$THF in crystalline form following the aforemen-
term alkali–metal-mediated alumination (AMMAl) to describe tioned literature procedure and dissolved it in neat d -THF to
Towards answering this question in this new study we
8
their metallation applications in this medium. Instead evidence replicate the environment it is employed in during successful
1
points to these being lithiation reactions, the generated carban- AMMAl applications. Recording the H NMR spectrum at
ions of which are rapidly trapped by alkylaluminium species to ambient temperature revealed a simple pattern showing one set
15
i
form aluminate products.
of Bu and TMP resonances consistent with a single solution
species (Fig. 1). Supporting this single species assignment, the
7
Li NMR spectrum shows a sharp single resonance at 1.21 ppm.
Results and discussion
Has the active base of 1 been crystallographically
1
Re-running the H NMR spectra at elevated temperatures (from
1
18
3
00 to 330 K) did not change this pattern. A DOSY H NMR
spectrum (Fig. 2) of this d -THF solution of crystalline 1$THF
showed that all of the resonances of the aluminate moiety (that
characterised?
8
1
1a
In the original preparation of 1 reported by Uchiyama in 2004,
i
LiTMP prepared in situ by the action of n-butyllithium on is Bu, CH
2
, ꢀ0.22 ppm; CH
3
, 0.88 ppm; CH, 1.88 ppm: TMP,
, 1.51 ppm) lie along the
ꢁ
TMP(H) at ꢀ78 C was subsequently treated with triisobutyla- CH
3
, 1.21 ppm; b-CH
2
, 1.20 ppm; g-CH
2
ꢁ
luminium and the mixture was warmed to 0 C. The bulk same line on the y-axis with essentially the same diffusion
ꢀ
10
ꢀ10
2
ꢀ1
solvent employed was THF (in an approximate 25 molar excess coefficient (6.95 ꢂ 10
ꢃ 0.09 ꢂ 10
m s ) implying that
on a 2 mmol scale reaction) though the mixture also contained the two distinct ligands belong to the same compound/struc-
hexane from the lithium and aluminium reagent solutions ture. Only the THF resonances (at 1.78 and 3.61 ppm) of the
employed. Evidence that LiTMP and triisobutylaluminium can 1$THF formulation lie outside of this line. The obvious expla-
interact with each other under the mediation of a Lewis base L nation is that THF is labile and hence could be undergoing a
i
i
to forge co-complexes of the type [L$Li(m-TMP)(m- Bu)Al( Bu)
came from our crystallographic characterisation of three librium in d
examples where L is N,N-diisopropylbenzamide, TMP(H) or down the y-axis as the THF has a smaller molecular weight
2
]
metal-attached coordinative, metal-free decoordinative equi-
8
-THF solution. These resonances appear lower
16
ꢀ9
triethylamine (Scheme 1). Going one step better, Naka, and thus a higher diffusion coefficient (2.05 ꢂ 10 ꢃ 0.01 ꢂ
ꢀ10
2
ꢀ1
i
Uchiyama and Wheatley subsequently crystallographically 10
m s ) than the “LiTMP$Al( Bu) ” portion of crystalline
3
characterised an aluminate compound containing all the 1$THF. The obvious assignment for this single species in
components of the base reaction mixture 1 in the mono-THF a vast pool of excess THF is the solvent-separated ion
i
i
11b
+
i
ꢀ
complex [(THF)$Li(m-TMP)(m- Bu)Al( Bu)
2
] 1$THF (Scheme 1).
pair [{Li(THF) } {Al(TMP)( Bu) } ] 1$(THF) . Supporting this
4 3 4
These Lewis base stabilised aluminates have in common con- assignment a comparatively broad resonance at 139.8 ppm is
2
7
tacted ion pair structures where Li and Al connect through found in the Al NMR spectrum in d
8
-THF solution consistent
} ] ion and as aforemen-
i
ꢀ
ligand bridges. Signicantly the crystals of all of these with an asymmetrical [{Al(TMP)( Bu)
compounds were grown in bulk hydrocarbon solutions. The tioned its Li NMR spectrum shows a singlet resonance at 1.21
most experimentally relevant set of 1$THF were crystallised ppm which coincides exactly with the Li NMR spectrum of the
4
from a bulk hexane solution containing a stoichiometric de- ate compound [{Li(THF) } {Al( Bu) } ] implying that the sepa-
3
7
7
+
i
ꢀ
4
i
17
+
ciency of THF [0.625 mmol per 1 mmol of “LiTMP$Al( Bu)
3
”].
rated {Li(THF)
4
} cation is common to both ates [note though
3032 | Chem. Sci., 2014, 5, 3031–3045
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
1
i
Fig. 1 H NMR spectrum of crystalline THF$LiTMP$Al( Bu)
3
1$THF in d
8
-THF solution in which it forms solvent-separated [{Li(TH-
+
i
ꢀ
F)
4
3
4
} {Al(TMP)( Bu) } ] 1$(THF) .
Scheme
LiTMP$Al( Bu)
2
The previously postulated dismutation of putative
1 in bulk THF solution.
i
3
1
i
Fig. 2 DOSY H NMR spectrum of crystalline THF$LiTMP$Al( Bu)
3
species unequivocally identied from the mixture. Arrived at by
simply balancing the stoichiometry of the equilibrium reaction,
1
8
$THF in d -THF solution. Note that a trace amount of TMP(H) is
evident at 1.04 ppm.
+
i
ꢀ
the putative co-product “[{Li(THF)
inspired us to the idea of employing the bis-TMP species
4 2 2 4
} {Al(TMP) ( Bu) } ]”, 2$(THF)
+
i
4
2
that the chemical shi for the {Li(THF) } cation is highly “LiTMP$Al(TMP)( Bu) ”, 2 in AMMAl reactions (see later). Kno-
sensitive to changes in concentration – see Fig. S1(b) in ESI† for chel's subsequent discovery that a closely related equilibrium may
an example]. It is well known that low local symmetry around Al be operating in THF solution mixtures of TMPMgCl$LiCl and
1
9
centres in general, and indeed specically in TMP attached Al(Et) leading to the tetraalkylaluminate “MgCl$Al(Et) ” (charac-
3
4
27
systems leads to broad signals [in (TMP) AlX systems they can terised in part by a sharp resonance at 159 ppm in Al NMR
2
20
14
be hundreds or even thousands of Hz broad].
Intriguingly these NMR spectroscopic results from d
solutions of crystalline 1$THF are in discordance with our earlier the THF solutions of crystalline 1$THF and its in situ form 1. Our
studies of “LiTMP$Al( Bu) ” prepared in situ in bulk THF solu- ndings detailed below were unexpected.
3
tion. A combination of H, Li and C NMR data pointed strongly
to the existence of a dismutation process (Scheme 2) in contrast to
the single species implicated in the d -THF solution of crystalline
$THF. The key piece of evidence towards this dismutation was the
characterisation of the tetraalkylaluminate [{Li(THF)
a solvent-separated ion pair structure though this was the only allation (DoM) chemistry in the classical studies of Wittig
2
spectra) and the alkylaluminium amide (TMP)Al(Et) $THF
8
-THF motivated us to revisit in greater detail the comparison between
i
21
1
7
13
Comparative reactivity studies of in situ 1 and crystalline
$THF with anisole
} ], A pivotal molecule in the development of directed ortho met-
8
1
1
+
i
ꢀ
4
} {Al( Bu)
4
22
23
3
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 3031–3045 | 3033

View Article Online
Chemical Science
Edge Article
24
and Gilman, anisole is a benchmark substrate for measuring a major component of in situ 1 came from the observation of a
2
7
the performance of metallating agents. Hence there are sharp resonance at 152.5 ppm in its Al NMR spectrum in
numerous reports of the ortho metallation of anisole by d8-THF solution that matches that of an authentic sample of 3.
different metallating reagents. Lithium mono-TMP aluminate 1 The sharpness of this resonance is consistent with the high
is included in this number as in fact anisole was utilised as the degree of symmetry in the tetrahedral Al centre in 3. This sharp
25
model substrate by Uchiyama when this reagent was rst resonance (reported at 153.4 ppm) appears to have been
11a
introduced. This original study found that a THF solution of wrongly assigned as belonging to putative contacted ion-pair
i
11b
3
in situ 1 gave a 99% yield of o-iodoanisole following quenching “LiTMP$Al( Bu) ” 1 in an earlier paper. A highly asymmetrical
i
ꢀ
3
] centre would give rise to a broader resonance
of the metallated intermediate with iodine (Table 1: this table [Al(TMP)( Bu)
gives the yields of the reactions of anisole with various Li–Al and which as mentioned earlier appears in our spectrum at 139.8
Al reagents mentioned in this paper). This “AMMAl” was carried ppm. Note as mentioned above that a similar Al chemical shi
out at room temperature for three hours and most signicantly is found in the related highly symmetrical tetraethylaluminate
2
7
+
ꢀ
14
the base:anisole stoichiometry employed to achieve this yield “[(MgCl) (AlEt ) ]” (at 159 ppm) as reported by Knochel.
4
was 2.2 : 1.0 molar equivalents, that is the base was in a slightly Interestingly when we repeated the original reaction carried out
greater than twofold excess. Hence this implies that at least 50% by Uchiyama using a 2.2 : 1.0 stoichiometric ratio of in situ 1 to
of the base 1 is inactive towards anisole under the conditions anisole in THF solution and recorded the NMR spectrum of the
8
studied. For comparison in this work we repeated this original metallated intermediate in d -THF solvent we see lithiated
reaction but this time using a 1 : 1, base:anisole stoichiometry anisole (conrmed by comparison to a spectrum of an authentic
in bulk THF solution. That in situ 1 could deprotonate anisole sample) as well as aluminated anisole through diagnostic
effectively was conrmed by this repeat reaction though doublet of doublet resonances for the anisolyl meta C–H adja-
1
signicantly the yield of deprotonated anisole observed in a H cent to the ortho site of metallation at 7.65 and 7.48 ppm
6
NMR spectrum of a d -benzene solution of the reaction mixture respectively in an appropriate 1 : 4 ratio (Fig. 3). This provided
only approached 50% conversion of anisole starting material. the rst strong hint that the reactions of in situ 1 are not merely,
This loss of about 50% of base activity is explicable if the dis- if at all, aluminium–hydrogen exchange reactions.
mutation equilibrium in Scheme 2 lies to the right hand side
Surprisingly, contrasting with the previous straightforward
and if one of the two components, the tetrabutylaluminate 3 metallation of anisole using in situ prepared 1, on dissolving
was inactive towards anisole. Previously we had reported that 3 crystalline 1$THF in THF solution mixed with anisole and stir-
21
failed to react with N,N-diisopropylbenzamide which carries a ring the mixture for several hours to replicate the reaction with in
much stronger ortho-deprotonating directing group than the situ 1 no reaction was observed to take place (Table 1) as deter-
methoxy substituent of anisole so it was anticipated that 3 mined from the recovered anisole seen in NMR spectra. The
would be inert towards anisole in bulk THF solution and a implication of this nding is that once the aluminate structure of
+
i
ꢀ
4 4
control reaction between them conrmed this view. Moreover 1$THF, presumably as [{Li(THF) } {Al(TMP)( Bu) } ], 1$(THF) , is
3
we found that 3 even failed to deprotonate the acidic N–H bond formed all deprotonative reactivity of the mixture is lost. To probe
of TMP(H), the co-product obtained when the TMP anion this idea further 1$THF was also prepared in situ in hexane
i
functions as a base. Further proof that tetrabutylaluminate 3 is solution by combining its component compounds [Al( Bu) ,
3
LiTMP and THF in a 1 : 1 : 1 ratio] but even this mixture proved
inert towards anisole. Under these poorly solvating conditions
the aluminate will almost certainly be in its contacted ion pair
Table 1 Comparative reactivities of various Li–Al or Al reagents
towards anisole
i
i
2
form [(THF)$Li(m-TMP)(m- Bu)Al( Bu) ], 1$THF. On the basis of
these pieces of evidence we can conclude with some certainty that
1
1
$THF is not the active experimental base in the solution mixture
, that crystalline 1$THF does not undergo a dismutation equi-
librium in the THF solution akin to that shown for in situ 1 in
Scheme 2 but remains as the solvent-separated species 1$(THF)4,
and in answer to the question posed the actual active base of 1
Metal reagent
Solvent
Yield (%)
i
LiTMP$Al( Bu)
3
(1) (in situ 2.2 equiv.)
THF
THF
THF or hexane
Hexane
99
50
0
i
LiTMP$Al( Bu)
3
(1) (in situ 1 equiv.)
i
(
(
(
(
(
THF)LiTMP$Al( Bu)
THF)LiTMP$Al( Bu)
in situ 1 equiv.)
THF)LiTMP$Al(TMP)( Bu)
in situ 1 equiv.)
3
(1$THF) (crystals)
(1$THF)
i
3
0
i
2
(2$THF)
Hexane
77
1
Fig. 3 Part of the aryl region of the H NMR spectrum of the reaction
+
i
ꢀ
i
[
{Li(THF)
4
} {Al( Bu)
4
} ] (3)
THF
THF
0
0
of in situ LiTMP$Al( Bu) 1 with anisole in d -THF solution showing both
3
8
i
Al( Bu)
3
lithiated anisole and aluminated anisole.
3034 | Chem. Sci., 2014, 5, 3031–3045
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
has therefore seemingly not been crystallographically charac-
Turning to reactivity issues, earlier studies showed 2$(THF)
n
i
i
terised or more accurately [(THF)$Li(m-TMP)(m- Bu)Al( Bu) ] was an effective base as it executed AMMAl on a range of organic
2
13
1$THF is not the active base (but see qualication later).
substrates (Scheme 3). Most signicantly, 2$(THF) was found
1
to metallate THF in bulk hexane solution as evidenced by the
slow appearance of resonances attributed to ortho-deprotonated
i
Towards solving the puzzle of “LiTMP$Al(TMP)( Bu) ”
ꢀ
2
THF (o-OC
4
H
7
), while addition of a second THF molecule led
i
For reasons that will become clear later in the discussion we have to the formation of [(THF)$Li(m-TMP)(m-OC H )Al( Bu) ], 4,
4
7
2
been unsuccessful in our several attempts to isolate a solid form
i
let alone a crystalline form of “LiTMP$Al(TMP)( Bu) ”, 2, the
2
putative co-product of the hypothesised dismutation portrayed in
Scheme 2. However it was the postulated presence of 2 in a THF
n
solvated form 2$(THF) within this equilibrium having the
attraction of seemingly possessing extra TMP power (as it is the
single TMP ligand in 1 that is its active base component) that
encouraged us to make a reagent of this twofold TMP stoichi-
1
ometry. To begin the study of 2 here, we recorded the H NMR
26
spectra of its individual constituent compounds LiTMP and
i
12
(
TMP)Al( Bu)
2
,
2 itself, as well as a 1 : 1 mixture of 2 and THF in
d
14-hexane solution (Fig. 4). Interestingly this comparison
i
2
revealed that LiTMP and (TMP)Al( Bu) remain separated as
homometallic species. When stoichiometric THF is introduced it
appears to interact preferentially with the Al reagent to afford
i
27
[
(TMP)Al( Bu) $THF] which we have previously characterised as
2
i
the chemical shis of the Bu resonances move (most diagnos-
tically the CH attached to the metal moves from 0.28 to 0.15
2
ppm) together with those for the TMP anion; whereas those of
LiTMP remain unchanged. While these species appear to stay
separated, it should be noted that there is a precedent for donor-
28
3 N
free co-complexation in polymeric [{Li(TMP)Al(Me) } ], though
signicantly this crystalline compound was formed under much
harsher reux conditions in toluene than the room temperature
conditions of our NMR comparison coupled with the fact that its
alkyl groups are much smaller than those in 2.
Scheme 3 Selection of deprotonation reactions executed by bis-TMP
reagent 2$(THF)
n
.
1
Fig. 4 Overlay of H NMR spectra of 2 and its component parts in d14-hexane solution.
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 3031–3045 | 3035

View Article Online
Chemical Science
Edge Article
Scheme 4 Previously hypothesised open ring structure pathway for intramolecular AMMAl reaction of TMEDA (top) and new proposed two-step
i
2 2 2 2 2
mechanism for formation of “aluminated” TMEDA complex Li[CH N(Me)CH CH NMe ](TMP)Al( Bu) (6) (bottom).
which was crystallographically authenticated (see ESI,† lack of available LiTMP in the reaction mixture as it would be
29
1
Fig. S9). Anisole was similarly ortho-aluminated by 2$(THF) to locked within a closed contacted structure with a strong
i
i
generate crystalline [(THF)$Li(m-TMP)(o-C
6
H
4
OMe)Al( Bu)
2
], 5, Li(m-TMP)(m- Bu)Al bridge less sterically congested than an
Al bridge. Interestingly our initial empirical
which in turn could be intercepted by the electrophile iodine to unstable Li(m-TMP)
2
12
produce o-iodoanisole in a 77% yield (Table 1). As mentioned reasoning that installing two TMP ligands within 2$THF would
previously this behaviour contrasts with that of in situ 1$THF, boost reactivity levels compared to that of the mono-TMP base
which fails to metallate anisole at all in hexane solution under 1$THF appears correct but for the wrong reason: in no example
the conditions studied.
yet has 2$THF functioned as a di-TMP reacting species, instead
0
0
Since the diamine TMEDA (N,N,N ,N -tetramethylethylene- it appears to be the “free” LiTMP present in the hexane solution
diamine), the methyl groups of which are only weakly acidic, mixture that boosts its reactivity compared to that of 1$THF.
could also be “aluminated” at one of these terminal methyl sites Unlike other bimetallic reagents which can show unusual
by in situ 2 in hexane solution we originally proposed an regioselective orientations, the regioselectivities observed here
intramolecular mechanism through a contacted but open for 2$THF are the same as those found for LiTMP (but in
30
structure as depicted in Scheme 4. However DFT calculations improved yields through the subsequent generation of Al–C
performed here (see below) indicate that such a twofold TMP bonds which lead to greater stability).
structure would be unstable with the Al bound TMP ligand
Until the present study no comparable reactivity study of 2
under geometry optimisation moving across to the Li centre in a had been carried out in bulk THF solution. Therefore we dis-
i
i
non-solvated (TMP)Li(m-TMP)Al( Bu)
2 2
structure which breaks solved the components of 2, LiTMP and (TMP)Al( Bu) in THF
31
apart to the homometallic components THF$LiTMP and solution at room temperature and added one molar equivalent
i
1
(
TMP)Al( Bu)
envisioned that LiTMP does the metallation (lithiation) of spectrum of the reaction mixture in d
TMEDA to give Li[CH N(Me)CH CH NMe ], the reduced steric conrmed that 2, as anticipated, also deprotonates anisole in
prole of which compared to that of LiTMP allows its this bulk polar medium.
2
on addition of a single THF ligand. It is therefore of anisole then stirred the mixture for several hours. A H NMR
6
-benzene solution
2
2
2
2
trapping via co-complexation (trans-metal-trapping is probably
What about the reactivity of 2 in bulk THF solution in the
i
more apt here, see later) with carbophilic (TMP)Al( Bu)
generate the observed heterobimetallic product Li[CH
2
to absence of anisole? As mentioned above, previously we estab-
N(Me)- lished that 2 readily deprotonates a stoichiometric quantity of
2
i
CH
2
CH
2
NMe
2
](TMP)Al( Bu)
2
(6) (Scheme 4).
THF in bulk hexane solution to afford the crystalline THF anion
ꢀ
If the trans-metal-trapping by the aluminium reagent is not (C
4
H
7
O ) contact ion pair complex 4 in a novel example of
32
100% efficient then lithiated substrates could persist, which “cleave and capture chemistry”. Here in this work we allowed 2
might explain the presence of lithiated anisole as well as alu- on its own to be stirred in THF solution for 24 hours at room
1
minated anisole in the aforementioned reaction with in situ 1 temperature before recording a H NMR spectrum of the
and anisole (this inefficient trapping was proven directly by resulting mixture in d -THF solution. Resonances characteristic
8
+
i
ꢀ
mixing lithiated anisole and the salt [{Li(THF)
4
} {Al( Bu)
-THF solution and recording the H NMR spectrum which 2.90, 3.42 and 3.74 ppm, see Fig. S14 in ESI†) consistent with 4,
revealed no reaction had taken place – in contrast to the neutral but signicantly these were only visible on magnifying the
4
} ], 3, of the deprotonated but intact THF ring were observed (e.g., at
1
in d
8
i
species (TMP)Al( Bu)
2
which proved an excellent trapping spectrum. A substantially larger resonance was seen for TMP(H)
reagent for the lithiated anisole, as detailed below). Applying at 1.06 ppm, much greater in relative integration terms than
ꢀ
this same train of thought, the failure of 1$THF to likewise could be accounted for by the TMP consumed in generating
metallate anisole in hexane solution can be attributed to the the trace amount of 4 witnessed in the spectrum. While
3036 | Chem. Sci., 2014, 5, 3031–3045
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
hydrolysis can never be ruled out completely as a contributing time lead to the deprotonation/decomposition of THF] LiTMP
i
factor (though we scrupulously dried the THF solvent before and (TMP)Al( Bu) exist as separate species each solvated by
2
employing it in the reaction), it seems more likely that the THF. This viewpoint is supported by a DOSY spectrum, which
ꢀ
generated THF anion (C H O ) is unstable in the bulk polar shows a signicant difference in the diffusion coefficients for
4
7
ꢀ
9
2
ꢀ1
i
medium. It could exist initially either as the lithium derivative each compound [LiTMP ¼ 1.22 ꢂ 10 m s ; (TMP)Al( Bu)
2
¼
ꢀ10
2
ꢀ1
[(THF)
x
(LiOC
4
H
7
)
n
]
or the solvent-separated aluminate 8.37 ꢂ 10
m s ]. Thus clearly whether in hexane or THF
} ] (contrast this with the solution there is no interaction between LiTMP and (TMP)-
due to a formidable steric barrier, but once LiTMP
solution) but would then decompose presumably via a [3 + 2] metallates a substrate (e.g., THF or anisole) the new lithiated
cycloreversion to the enolate of acetaldehyde and ethene. To substrate species being of reduced steric prole and greater
investigate what effect this formation and break down of 4 nucleophilicity through Li–C bond formation can join together
would exert on the Brønsted basic properties of 2 we stirred a (be trapped) with the neutral aluminium complex. Evidence
THF solution of 2 for 24 hours before introducing anisole as the that such fragments can join together comes from the previ-
+
i
ꢀ
2
[{Li(THF)
4
} {(TMP)(OC
4
H
7
)Al( Bu)
i
bimetallic-stabilised contacted ion pair form 4 found in hexane Al( Bu)
2
29
33
4 7
metallation probe. As expected no metallation of anisole took ously reported crystal structures of [(THF)$Li(m-TMP)(m-OC H )-
1
i
29
i
12
place as a H NMR spectrum of the reaction mixture revealed Al( Bu)
free anisole as well as THF anions and a substantial amount of respectively (see also the DFT study detailed below).
TMP(H). From these observations we conclude that if le to stir Taking stock of these ndings the picture of 2 in bulk THF
2 6 4 2
], 4, and [(THF)$Li(m-TMP)(o-C H OMe)Al( Bu) ], 5,
for a period of time in THF solution, 2 will deprotonate THF solution emerging now is that only LiTMP (solvated by THF),
i
releasing TMP(H) and be consumed. To check whether all base (TMP)Al( Bu) (solvated by THF), and depending on the age
2
activity is lost under such circumstances, we crystallised 4 from and history of the solution variable amounts of the THF
hexane solution, isolated it and dissolved it in bulk THF solu- degradation products 4 (presumably in its THF-separated form
+
i
ꢀ
2 x 4 7 n
} ]), [(THF) (LiOC H ) ],
tion. Anisole was added subsequently and the solution was [{Li(THF)
stirred for 24 hours. NMR analysis of the resulting mixture [(THF) LiO–C]CH
revealed that again no metallation of anisole had occurred spectroscopic interrogation. Signicantly we can nd no
4
} {(TMP)(OC
4
H
7
)Al( Bu)
34
n
2
] and ethene are observable by this NMR
i
conrming that aluminate 4, probably present in the modied evidence for a bimetallic cocomplex “LiTMP$Al(TMP)( Bu)
solvent-separated form [{Li(THF) } {(TMP)(OC H )Al( Bu) } ] is which falls into line with our DFT computational analysis (see
2
” 2
+
i
ꢀ
4
4
7
2
inactive as a base even though it contains a TMP ligand.
below) that questions the thermodynamic feasibility of such a
Curious about the constitution of 2 in THF solution we di-TMP contacted or solvent-separated bimetallic structure.
1
compared its H NMR spectrum with those of its constituent
Having tested all of the metal species within this 2 mixture
i
parts LiTMP and (TMP)Al( Bu)
2
(see composite spectra in Fig. 5). for their metallating ability the only possible candidate to
Close examination of these spectra show that the principal emerge is the aforementioned LiTMP, which in bulk THF
resonances of LiTMP (Me of TMP at 1.07 ppm) and (TMP)- solution exists in solvated form as deduced by Renaud and Fox
i
i
7
Al( Bu)
2 2
(Me of TMP at 1.21 ppm; CH of Bu at 0.03 ppm) match who observed both dimeric and monomeric forms through Li
35
almost exactly with corresponding resonances in 2 (1.04, 1.21 NMR spectroscopic studies. Wheatley et al. conrmed these
11b
and 0.03 ppm, respectively) though it is noticeable that the assignments in a later paper.
resonances associated with LiTMP are extremely sensitive to To establish whether LiTMP was the active Brønsted base
even small changes in concentration. Therefore it appears component in 2 we dissolved freshly prepared LiTMP in
certain that under the conditions studied [longer periods of
6
d -benzene solution in an NMR tube to which a few drops of
1
Fig. 5 Overlay of H NMR spectra of 2 and its component parts in d
8
-THF solution.
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 3031–3045 | 3037

View Article Online
Chemical Science
Edge Article
1
THF were added. A H NMR spectrum of this mixture was depicted in Scheme 6, this insertion of the aluminium reagent
recorded aer 30 minutes and again aer 24 hours (see Fig. S17 into the Li–C(anisolyl) bond should drive the equilibrium
in ESI†). Ethene was revealed in both spectra through a reso- between anisole and lithiated anisole towards the lithiated
nance at 5.25 ppm, which increased with time, consistent with species thus increasing the overall metallation yield of the
the metallation, ring opening and cleavage of THF. Signicantly reaction. We established that such an equilibrium exists
when 2 is le to stir in bulk THF solution for 24 hours a small between lithiated anisole and LiTMP by taking a freshly
amount of aluminate 4 is observed as mentioned previously, prepared sample of the former and mixing it with an equimolar
the implication being that LiTMP is lithiating THF to amount of TMP(H) in d
6
-benzene solution and stirring the
ꢀ
generate “C
4
i
H
7
O ” anions a small amount of which is trapped solution for 10 minutes. At this point LiTMP was observed via a
i
ꢀ
1
i
by (TMP)Al( Bu) to generate [(TMP)(OC H )Al( Bu) ] while the
H NMR spectrum. It is worthy of comment that (TMP)Al( Bu) ,
2
2
4
7
2
remainder decompose to ethene and lithium enolate. We also though not a metallating agent itself, contributes to the success
examined the lithiation of anisole (Scheme 5). Uchiyama, of the metallation reactions of 2 in two key ways: rstly, it traps
Mongin et al. previously reported that subjecting anisole to one the lithium carbanion and stabilises the carbanion moiety by
molar equivalent of LiTMP in THF solution over two hours reducing the polarity of the metal–carbon bond; secondly, by
produced aer iodine quenching only 9% of ortho-iodoani- not engaging at all with LiTMP on the le hand side of the
36
sole. To ascertain how much lithiated anisole was present equation (Scheme 6) the equilibrium can shi towards the
prior to any quenching step we reacted LiTMP with anisole in desired anisolyl aluminium product. This hypothesis of non-
i
the same stoichiometry in THF solution, but found the reaction cocomplexed LiTMP and (TMP)Al( Bu) homometallic species
2
afforded only about a 5% yield of lithiated anisole. Hence swimming separately in a pool of THF runs counter to any
i
LiTMP can denitely metallate/lithiate anisole, unlike any of the thinking that a “LiTMP$Al(TMP)( Bu) ” cocomplex was respon-
2
other species identied within the mixture of 2, albeit in a sible for these “AMMAl” reactions. Therefore the weight of
meagre yield. Since 2 furnishes excellent yields of metallated evidence from this work suggests these reactions are not in fact
anisole following iodine quenching the implication is that once direct aluminations (aluminium–hydrogen exchanges) at all but
formed any lithiated anisole will be quickly trapped by the rather two step lithiation/trans-metal-trapping processes (trans-
i
strongly carbophilic (TMP)Al( Bu)
2
.
This was established metal-trapping seems a more apt description here than the
unequivocally by taking a 1 : 1 mixture of lithiated anisole usual applied “transmetallation” for although aluminium is
t
(
0
prepared separately by reaction of anisole and BuLi in THF at replacing lithium in binding to the carbanion C atom the
ꢁ
37
i
C) and (TMP)Al( Bu) in d -THF solution in an NMR tube lithium may not necessarily leave the aluminium system but
2 8
1
13
and recording the H and C spectra (see Fig. S11 and S12 could remain part of a contacted ion pair or solvent separated
in ESI†). The most diagnostic resonance in the former, the meta ion pair compound). Similar reactivities to this one have
C–H adjacent to the metallated C–M, shows a signicant upeld recently been reported for the homoleptic bimetallic mixtures
shi (from 7.66 to 7.49 ppm) signifying the attached metal M of LiTMP and Zn(TMP)
2
(ref. 38) or Cd(TMP)
appear not to be “LiZn(TMP) ” or “LiCd(TMP)
spectra for the metallated C–M atom from 159.2 to 154.4 ppm but instead remain separated components in which LiTMP is
2
(ref. 39) as they
1
3
has switched from Li to Al; with a similar shi seen in the
C
3
3
” tris-TMP ates,
9,40
[
note that Uchiyama reported a similar but not identical also the active metallating base.
t
upeld Li to Al shi on treating lithiated anisole with BuLi with
Despite this evaluation of the reactions of 2 in bulk THF
the C–Al resonance appearing at 152.91 ppm, the main solutions not being AMMAls the essential point remains the
i
distinction being the trapping Al reagent used was Al( Bu)
3
]. same that these reactions are still synergistic in origin (for
From integration ratios this trans-metal trapping of the anisolyl
i
2
carbanion by (TMP)Al( Bu) seems essentially quantitative. As
Scheme 5 Capture of “aluminated” anisole by direct (RHS) and indirect Scheme 6 Proposed two-step pathway for the “alumination” of
LHS) approaches. anisole.
(
3038 | Chem. Sci., 2014, 5, 3031–3045
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
efficiency but not for any special selectivity) for without partic-
ipation of the aluminium reagent, quenching of the lithiated
+
substrates with electrophiles E would be unsatisfactory leading
+
ꢀ
to poor yields of the desired E (substrate) products. This
reects the non-selective nature of iodine quenching as it would
quench both lithiated anisole and LiTMP to prevent the equi-
librium in Scheme 6 shiing towards lithiated anisole; whereas
the aluminium reagent selectively targets lithiated anisole and
ignores the bulkier LiTMP.
Re-evaluating the composition and active base component of
in situ 1 in THF solution
Now that the picture of 2 in THF solution is much more
transparent following these new ndings, the composition of 1 Fig. 6 H NMR spectrum of (TMP)Al( Bu)
in THF solution needs to be re-considered. Taking into account
the surprising discovery that LiTMP is the active base compo-
nent within 2 we can propose a more complete composition for
1
i
+
i
ꢀ
2
+ [{Li(THF)
4
} {Al( Bu)
4
} ] 3 in
8
d -THF solution.
When TMP is attached to Al this Me resonance moves
downeld to 1.21 ppm in (TMP)Al( Bu)
{Li(THF)
1
(Scheme 7). Far removed from the original idea that it existed
i
2
and 1.20 ppm in
i
i
as a single species of formula [(THF)$Li(TMP)( Bu)Al( Bu)
2
], in
+
i
ꢀ
[
4
} {Al(TMP)( Bu)
3
} ] though these signals cannot be
this proposal 1 contains no less than ve species in two inter-
connected equilibria including most signicantly the separated
monometallic species LiTMP, which we have already estab-
lished can perform metallation of a substrate. Convincing
evidence for the second equilibrium came from mixing
differentiated in the combined spectrum. Since the equilibrium
under these ambient temperature conditions greatly favours
i
(
TMP)Al( Bu) its Me(TMP) resonance is much larger than that
2
+
i
ꢀ
4
of [{Li(THF) } {Al(TMP)( Bu) } ]. Measurement of the relative
3
+
i
ꢀ
i
i
integrals of [{Li(THF)
hampered by the broadness of the Al–CH
former species. Note that the corresponding resonance for the
4
} {Al(TMP)( Bu)
3
} ] and Al( Bu)
3
is
+
i
ꢀ
authentic samples of the salt [{Li(THF) } {Al( Bu) } ], 3, and
4
4
2
( Bu) resonance of the
i
1
(
2 8
TMP)Al( Bu) in d -THF solution and recording the H NMR
spectrum at room temperature (Fig. 6). The low frequency
+
i
ꢀ
21
homoleptic ate [{Li(THF) } {Al( Bu) } ] is similarly broad. In
4 4
region about 0 ppm is extremely informative as each species
both cases the broadness can be attributed to the quadrupolar
i
exhibits a well-dened Al–CH
2
( Bu) resonance within it.
27
effect of the Al centre (spin 5/2). In the symmetrical species
Four such resonances observed at (0.02, ꢀ0.10, ꢀ0.22 and
+
i
ꢀ
i
[
{Li(THF) } {Al( Bu) } ] the Al–CH ( Bu) resonance is a doublet
4 4 2
i
i
ꢀ
0.37 ppm) can be assigned to (TMP)Al( Bu) , Al( Bu) ,
2
3
ꢀ
due to coupling with the adjacent CH but this is further split by
the Al into a doublet of sextets though as the environment is not
perfectly symmetrical some merging of the lines occurs and the
resonance observed appears wide and broad (see Fig. S24 and
+
i
ꢀ
+
i
[
{Li(THF) } {Al(TMP)( Bu) } ] and [{Li(THF) } {Al( Bu) } ]
respectively. Four CH ( Bu) resonances were also observed for
4
3
4
4
i
3
the four distinct species though the doublet of doublets for
i
+
i
ꢀ
Al( Bu)
3
and [{Li(THF)
4
} {Al(TMP)( Bu)
3
} ] overlap. All assign-
1
27
S25 in the ESI† for decoupling and 2D [ H, Al] HSQC and
ments were veried by comparison with the spectra of authentic
i
HSQC-TOCSY experiments which support the Al and Bu
samples of the individual components. The trialkyl-amido
assignments within 1$(THF) and 3). The equilibria are also
4
+
i
ꢀ
aluminate [{Li(THF)
4
i
} {Al(TMP)( Bu)
3
} ] was prepared by react-
implicated on mixing equimolar proportions of LiTMP and
i
ing neutral (TMP)Al( Bu) with an equimolar amount of ( Bu)Li
i
2
Al( Bu)
3
in d
8
-THF solution (see the spectra comparison in
in THF solution (Fig. S20 and S21 in ESI,† note the corre-
1
ꢁ
Fig. 7). On recording this H NMR spectrum at 0 C, the reso-
nances for LiTMP (most diagnostically the Me resonance at 1.05
ppm, though this overlaps with a TMPH resonance the presence
i
sponding resonance for ( Bu)Li comes more upeld at
ꢀ
0.98 ppm) and this aluminate gave an identical spectrum to
that of crystalline 1$THF dissolved in d -THF solution,
8
of which is unavoidable due to attack of THF by LiTMP) and
which we discovered was inactive as a base. It is also
i
Al( Bu) (at ꢀ0.25 ppm) are the most prominent. A smaller
3
signicant that no LiTMP was found in the spectrum of the
+
i
ꢀ
4
extremely broad resonance for [{Li(THF) } {Al( Bu) } ] is clearly
4
+
i
ꢀ
i
[{Li(THF)
4
} {Al( Bu)
4
} ], 3 and (TMP)Al( Bu)
2
mixture as evi-
+
i
ꢀ
4
seen too. The presence of [{Li(THF) } {Al( Bu) } ] is also clearly
4
denced by the absence of a Me resonance at 1.05 ppm.
i
distinguishable in the Me region of the Bu group as a doublet at
i
0
.84 ppm, though the analogous doublets for the other Bu
containing species overlap into a complex multiplet at about
.89 ppm consistent with there being multiple species present
0
i
3
rather than simply Al( Bu) . Due to the fact that we are seeing
some metallation in the solution due to the presence of TMPH
i
then it is possible that the amount of (TMP)Al( Bu)2 in
solution is being decreased as this species will trap any carb-
Scheme 7 Re-evaluated composition of “single-species” THF$Li(TMP)
i
anion formed upon metallation (see earlier discussion of effi-
$
Al( Bu)
3
showing its existence as two connected equilibria involving
i
five distinct species.
cient trapping of anisolyl anions by (TMP)Al( Bu)
). Clearly the
2
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 3031–3045 | 3039

View Article Online
Chemical Science
Edge Article
1
i
+
3 4 3 3 8
i
ꢀ
i
Fig. 7 Overlay of H NMR spectra of LiTMP, Al( Bu) , [{Li(THF) } {Al(TMP)( Bu) } ] and [LiTMP + Al( Bu) ] in d -THF solution.
i
(
Bu)Me region at 0.80–0.95 ppm in containing more than the excess of 1$THF (2.2 molar equivalents) used by Uchiyama et al.
four overlapping species you would expect in the equilibrium in their synthetic applications to ensure maximum yields of the
proposed supports this suggestion. Other ates of formula metallated/quenched substrates were obtained. Moreover
i
-
ꢀ
ꢀ
[(TMP)Al( Bu)
2
X] where X is for example C
4
H
7
O
or C
2
H
3
O
1$THF is never going to be a good base in hexane solution for if
i
could also be present. To check whether LiTMP was partici- the LiTMP : Al( Bu)
pating in an equilibrium with the salt [{Li(THF)
we added both to a d -THF solution and monitored it by H However, there are at least two qualications. First, at higher
3
ratio in the starting mixture is exactly 1 : 1
ꢀ
4 4
} {Al( Bu) } ] there will be no free LiTMP available to perform the metallation.
+
i
1
8
NMR spectroscopy. The spectra revealed that the resonances temperatures the contacted ion pair structure of 1$THF could
associated with the two individual compounds remain break up and release LiTMP making metallations of suitably
unchanged. We thus conclude that these two species are not in thermally stable substrates a possibility. In contrast free LiTMP
equilibrium with each other. Neither is LiTMP in equilibrium is always available in hexane solutions of 2$THF. Second, Lewis
i
with (TMP)Al( Bu)
2
as we established through the aforemen- bases can coordinate to the Lewis acidic lithium centre and
i
tioned studies of 2$THF.
3
generate a contacted ion pair aluminate with Al( Bu) that can
Based on these new observations we can nd no evidence at subsequently metallate a C–H bond in the Lewis base in a
+
i
ꢀ
2
n 2
all for a species of composition “[{Li(THF) } {Al(TMP) ( Bu) } ]” genuine example of alkali–metal-mediated alumination. This
that we had originally presumed in the dismutation process possibility has already been demonstrated in the formation of
i
shown in Scheme 2. DFT calculations (see below) support the [Li{Me NCH CH N(Me)CH } Al( Bu) ] (6) by reaction of 1 with
2
2
2
2 2
2
30
non-existence of such a heteroleptic aluminate species. It two equivalents of TMEDA in hexane solution.
transpires that 1$THF is much more complicated existing in at
least ve distinct species in THF solution. The complexity can
i
DFT calculations
Theoretical calculations. The structure of 1 “LiTMP$Al( Bu)
ꢀ
4 2
] or [Al( Bu) ] containing species; was previously interrogated by DFT calculations using the
3
be attributed to the lability of Al( Bu) which can add a TMP
i
ꢀ
i
i
ligand to generate [{(TMP)Al( Bu)
3
} ] and gain or lose an Bu
3
”
i
i
+
group to form [Al( Bu)
whereas by comparison the solution chemistry of 2$THF is B3LYP/6-31+G* level of theory. This comprehensive theoret-
11b
i
much simpler due to the relative poor lability of (TMP)Al( Bu)
ical study by Naka, Uchiyama, and Wheatley et al. was per-
2
and specically its inability to form a co-complex with LiTMP on formed on [S$LiNMe $Al(Me) ] (where S ¼ the donor solvent
2
3
steric grounds. Where 1$THF and 2$THF do coincide is in the Me O) for calculational simplicity and its metallation reaction
2
fact that the active base ingredient in both mixtures in THF with anisole was modelled. Rigorous in its detail and broad in
solution is LiTMP. Ironically, revisiting the original question, scope, this study uncovered possible intermediates and transi-
“
has the active base of 1 been crystallographically charac- tion states along the reaction coordinate as well as quantifying
terised?”, the revised answer is yes, as LiTMP has been crys- the energy differences involved. However, its starting point of a
tallographically characterised in two different polymorphic contacted ion pair structure [S$Li(m-Me)(m-NMe )$Al(Me) ] and
2
2
26
31
forms as well as in a THF-solvated form. 2$THF is the the subsequent pre-metallation complex it forms with anisole
strongest base because it would always have the largest [Ph(Me)O$Li(m-Me)(m-NMe )$Al(Me) ], a dative coordination
proportion of LiTMP present in a solution of the same molarity; leading to an energy saving of ꢀ15.8 kcal mol , are, on the
whereas in 1$THF some LiTMP will always be lost due to the basis of the new information established here, not relevant to
equilibria in operation. This last point is in agreement with the the actual experimental reagent 1 employed in bulk THF
2
2
ꢀ
1
3040 | Chem. Sci., 2014, 5, 3031–3045
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
solution. This is because all the mixed lithium-aluminium
species present in bulk THF solution are solvent separated and
so the Li cannot cooperate with Al by providing the anisole
with a coordination point adjacent to the amido ligand
attached to the Al. In other words no complex induced
proximity effect would be possible. We conrmed this experi-
mentally by showing that the solvent-separated aluminate
Scheme 9 ChemDraw representation of the rearrangement of Li(m-
i
i
TMP)
2
Al( Bu)
2
2
closed into (TMP)Li(m-TMP)Al( Bu)
2
2open as predicted by
DFT calculations.
+
i
ꢀ
[{Li(THF)
} {Al(TMP)( Bu) } ] 1$(THF) is incapable of metal-
4 3 4
lating anisole in bulk THF solution.
In earlier work, we also performed DFT calculations
considering the feasibility of a structure of putative 2$THF of
i
12
formula [THF$Li(m-TMP) Al( Bu) ]. These calculations used
2
2
the Gaussian 03 package with geometry optimisation using the
B3LYP density functionals and the 6-311(d, p) basis set with
zero point energy corrections. While this study conrmed the
most energetically stable arrangement of 2$THF has two
i
bridging TMP ligands with the two terminal Bu ligands on Al,
signicantly it also exposed the relative instability of 2$THF
with respect to either its homometallic components LiTMP and
i
Scheme 10 Fragmentation of (THF)(TMP)Li(m-TMP)Al( Bu)
2
into its
i
(TMP)Al( Bu)2 or THF solvates thereof. Depending on the
homometallic components as predicted by DFT calculations.
homometallic components employed this instability ranged
ꢀ
1
from +14.16 to +20.60 kcal mol (Scheme 8).
To shed more light on 2 having accrued much more knowl- 2closed. That notwithstanding, on introducing a THF ligand to
edge on the experimental system through this study we the lithium centre to mimic the experimental stoichiometry of
have carried out additional calculations using the same 2$THF, the structure fragmented under geometry optimisation
parameters. To start we modelled a THF-free version of 2, into the homometallic components THF$LiTMP and (TMP)
i
i
Li(m-TMP) Al( Bu) , 2closed, having a closed four-membered Al( Bu) (Scheme 10). The energy given by the sum of these two
2
2
2
(
LiNAlN) ring, comparing it against an open version 2open, to separated homometallic components is ꢀ1614.811704 a.u.
ascertain the effect that relaxing the steric strain by opening the compared to ꢀ1614.805526 a.u. for 2
$THF, equating to the
closed
ring might have on the stability of 2 (Scheme 9). Our rst model
of 2open was derived by breaking one of the Li–N(TMP) bonds in
2closed to leave a single Li–N(TMP)–Al bridge with the remaining
three ligands bonded solely to Al. However, under geometry
optimisation this 1-coordinate Li/4-coordinate Al model rear-
ranged through the migration of the terminal Al-attached TMP
ligand to a terminal position on Li to generate a more realistic 2-
coordinate Li/3-coordinate Al structure, isoelectronic with
4
1
crystallographically characterized (TMEDA)Li(m-TMP)Li(TMP)
42
and (PMDETA)Na(m-TMP)Li(TMP). Relieving the steric strain
by opening the LiNAlN ring in this way does indeed increase the
ꢀ
1
stability with 2open being ꢀ5.53 kcal mol more stable than
Scheme 8 Energies obtained from DFT calculations for the possible Scheme 11 Energies of the modelled metallations and subsequent
i
reactions in which putative THF$Li(m-TMP)
2
Al( Bu)
2
is formed.
trapping reactions of anisole.
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 3031–3045 | 3041

View Article Online
Chemical Science
Edge Article
Scheme 12 Summary of the compositions of “aluminate” reagents 1 and 2. Note depending on the age of solutions, THF degradation products
will also be present.
3042 | Chem. Sci., 2014, 5, 3031–3045
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
ꢀ1
former being more stable by ꢀ3.87 kcal mol [or by a more quickly trapped by an alkylaluminium species (we term this
ꢀ
1
realistic ꢀ14.19 kcal mol
if the dimeric aggregation of trans-metal-trapping), which drives the reaction forward to a
(
THF$LiTMP) is taken into account]. Collectively these results high yield of “aluminated” anisole. Reagent 2 is more simple
2
suggest that a THF solvate of the contacted ion pair in solution remaining as its added components LiTMP and
i
i
LiTMP$Al(TMP)( Bu)
2 2
, whether in a closed or open ring (TMP)Al( Bu) in hexane or as THF solvates thereof when stoi-
arrangement would be too high in energy to exist, supporting chiometric THF is added or in bulk THF solution. The lack of
i
the experimental NMR investigations which failed to detect any complexity can be attributed to the extra bulk of (TMP)Al( Bu)
2
i
such species.
3
compared to Al( Bu) which prevents its association with LiTMP
We have also modelled the reaction of LiTMP with anisole, and thus preventing the complicated equilibria witnessed for 1.
which experimentally produced less than 10% of either lithiated On the basis of these ndings one must advise caution against
anisole or its 2-iodo derivative following quenching with iodine. assuming that a crystalline bimetallic species grown from
In the calculations where lithiated anisole was modelled solution is the active reagent in AMMAl reactions; but on the
somewhat unrealistically as an unsolvated monomer or a mono- other hand, it was only through the isolation of such a metallo
THF-solvated monomer where the Li atoms have low coordi- intermediate that its inactivity as a base could be unequivocally
nation numbers the DE values for the reactions were highly exposed. Looking more generally, while the complexities
endergonic (Scheme 11). However even when the lithiated ani- apparent in alkali–metal-mediated metallation reactions have
43
sole was modelled more realistically as a tetramer or tri-THF- recently been cause to tag them as “black box chemistry”, shas
solvated monomer starting from (LiTMP)4 or (THF$LiTMP)2 of light are now becoming visible through it. In this specic case
respectively as the base, the reactions are close to thermoneu- the message seems to be that unless these aluminate species are
ꢀ
1
ꢀ1
tral (DE is +1.73 kcal mol or ꢀ1.18 kcal mol respectively) in contacted ion pair form where the alkali metal can act as a
though the latter one is marginally exergonic. The thermody- Lewis acidic coordination point for an incoming substrate to
namics changed signicantly when (TMP)Al( Bu)
i
2
was intro- closely approach the anionic aluminium moiety, AMMAl will
duced to the lithiated anisole. Depending on what form of not generally occur; otherwise any observed metallation may be
lithiated anisole (tetramer, mono-THF-solvated monomer or tri- effected by the separated lithium reagent followed by rapid
THF-solvated monomer) was employed the DE values ranged trapping of the newly formed lithium carbanion via an alkyla-
ꢀ
1
ꢀ1
from ꢀ9.39 kcal mol to ꢀ28.45 kcal mol , so in all three luminium reagent.
cases the reaction proved exergonic (Scheme 11). These calcu-
lations are therefore fully consistent with our experimental
observations that LiTMP can lithiate anisole to only a limited
Experimental section
extent, but that introducing the aluminium trapping reagent General methods
makes the C–H to C–metal transformation much more
All reactions and manipulations were carried out under a
favourable.
protective dry pure argon atmosphere using standard Schlenk
techniques. Products were isolated and NMR samples prepared
within an argon-lled glovebox. Hexane was dried by heating to
reux over sodium-benzophenone and distilled under nitrogen
Conclusions
This study has examined in detail the constitutions of the two
most important alkali metal aluminate reagents made to date in
n
i
prior to use. ( Bu)Li (1.6 M in hexanes) and Al( Bu)
3
(1.0 M in
hexanes) were purchased from Aldrich and used as received.
i
i
“LiTMP$Al( Bu) ” 1 and “LiTMP$Al(TMP)( Bu) ” 2. In contrast to
i
3
2
(
Bu)Li (1.6 M in heptane) was purchased from ACROS and used
previous investigations that viewed 1 as a single species in THF
solution, this study uncovers ve distinct species, which
appear to exist simultaneously in two connected equilibria in
THF solution. Scheme 12 gives a pictorial summary of the
species that exist in both hexane and THF solution as well as
those of 2 in the same media. One striking observation is
that the single species previously identied in crystal form
as received. TMP(H) was obtained from Aldrich and dried over
26b
i
12
˚
4
A molecular sieves prior to use. LiTMP and (TMP)Al( Bu)2
were prepared according to literature procedures. NMR spectra
were recorded on a Bruker AVANCE 400 NMR spectrometer,
1
7
operating at 400.13 MHZ for H, 155.50 MHz for Li, 104.25
27
13
13
MHz for Al and 100.62 MHz for C. All C NMR spectra were
proton decoupled. H and C spectra were referenced to the
appropriate solvent signal, Li NMR spectra were referenced
2
against LiCl in D O at 0.00 ppm and Al NMR spectra were
1
13
i
i
[
(THF)$Li(m-TMP)(m- Bu)Al( Bu) ], 1$THF, is not an active base in
7
2
either hexane or THF solution using anisole as a test Brønsted
acid. We conrmed that these crystals do indeed form a single
species when dissolved in THF solution in the solvent-separated
27
referenced against AlCl in D O at 0.00 ppm.
3
2
+
i
ꢀ
modication [{Li(THF)
ably, however, on making up 1 in situ by adding LiTMP and
4 3 4
} {Al(TMP)( Bu) } ] 1$(THF) . Remark-
Theoretical calculations
i
44
3
Al( Bu) to THF solution, four other species in addition to DFT calculations were performed using the Gaussian
+
i
ꢀ
[{Li(THF)
4
} {Al(TMP)( Bu)
3
} ] 1$(THF)
4
are produced as identi- computational package G03. In this series of calculations the
45
46

ed from NMR data. Preparing authentic samples of all of these B3LYP density functional and the 6-311(d, p) basis set were
species and testing them all individually with anisole, we found used. Aer each geometry optimization a frequency analysis
that only the lithium amide LiTMP was capable of metallating was performed. The energy values quoted include the zero point
anisole. Though the yield of lithiated anisole was low, it can be energy contribution.
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 3031–3045 | 3043

View Article Online
Chemical Science
Edge Article
+
i
ꢀ
4
Synthesis of [{Li(THF)
4
} {Al( Bu)
} ] (3)
113, 7470–7562; (e) F. Mongin and A. Harrison-Marchand,
Chem. Rev., 2013, 113, 7563–7727.
n
(
Bu)Li (3.13 mL, 1.6 M in hexanes, 5 mmol) was added to a
ꢁ
6 (a) Y. Kondo, M. Shilai, M. Uchiyama and T. Sakamoto, J. Am.
Chem. Soc., 1999, 121, 3539–3540; (b) W. Clegg, S. H. Dale,
E. Hevia, G. W. Honeyman and R. E. Mulvey, Angew. Chem.,
Int. Ed., 2006, 45, 2370–2374; (c) M. Uchiyama,
Y. Matsumoto, D. Nobuto, T. Furuyama, K. Yamaguchi and
K. Morokuma, J. Am. Chem. Soc., 2006, 128, 8748–8750.
mixture of THF (4 mL) and TMPH (0.85 mL, 5 mmol) at ꢀ78 C
ꢁ
i
and the mixture was stirred for 10 min at 0 C. Al( Bu)
3
(5 mL,
ꢁ
1
M in hexanes, 5 mmol) was then added at ꢀ78 C and the
ꢁ
mixture stirred for 30 min at 0 C to give a pale yellow solution
and a white solid. The reaction mixture was then heated to
reuxing temperature to obtain a clear solution and subsequent
7
8
9
E. Hevia, D. J. Gallagher, A. R. Kennedy, R. E. Mulvey,
C. T. O'Hara and C. Talmard, Chem. Commun., 2004, 2422–
bench cooling of this solution afforded colourless crystals of 3
0.55 g, 20%). H NMR (C D , 400.13 MHz, 300 K): d 3.46
1
(
6
6
2423.
(
m, 16H, OCH
4H, CH CH(CH
CH CH(CH
OCH THF), 29.55 (CH
ppm (CH THF) [note that the resonance for {CH
could not be observed in either C D or [D ]THF solution
2
THF), 2.37 (sept, 4H, CH
), 1.32 (m, 16H, CH THF), 0.07 ppm (d, 8H,
, 100.62 MHz, 300 K): d 67.75
), 28.03 (CH CH(CH ), 25.48
3 2
CH(CH ) }
2 3 2
CH(CH ) ), 1.38 (d,
(a) C. T. O'Hara, Organomet. Chem., 2011, 37, 1–26; (b)
R. E. Mulvey and S. D. Robertson, Top. Organomet. Chem.,
2
2
3
)
2
2
1
3
2
3
)
2
); C NMR (C
6 6
D
2013, 45, 103–140.
(
2
2
CH(CH
3
)
2
2
3 2
)
D. R. Armstrong, A. R. Kennedy, R. E. Mulvey, J. A. Parkinson
and S. D. Robertson, Chem. Sci., 2012, 3, 2700–2707.
2
2
6
6
8
1
13
10 (a) S. H. Wunderlich and P. Knochel, Angew. Chem., Int. Ed.,
2009, 48, 1501–1504; (b) B. Conway, E. Crosbie,
A. R. Kennedy, R. E. Mulvey and S. D. Robertson, Chem.
Commun., 2012, 48, 4674–4676.
however its existence was conrmed by a H- C HSQC experi-
ment (see Fig. S22 and S23 in ESI†)]; Li NMR (C D , 155.50
MHz, 300 K): d ꢀ1.19 ppm.
7
6
6
1
1 (a) M. Uchiyama, H. Naka, Y. Matsumoto and T. Ohwada,
J. Am. Chem. Soc., 2004, 126, 10526–10527; (b) H. Naka,
M. Uchiyama, Y. Matsumoto, A. E. H. Wheatley,
M. McPartlin, J. V. Morey and Y. Kondo, J. Am. Chem. Soc.,
Acknowledgements
We gratefully acknowledge the Royal Society (Wolfson Merit
Award to R. E. M.), the Royal Society of Edinburgh (BP Trust
Fellowship to S. D. R.), the European Research Council (ERC
Starting Grant to E. H.), the University of Strathclyde/EPSRC
2
007, 129, 1921–1930; (c) H. Naka, J. V. Morey, J. Haywood,
D. J. Eisler, M. McPartlin, F. Garcia, H. Kudo, Y. Kondo,
M. Uchiyama and A. E. H. Wheatley, J. Am. Chem. Soc.,
2
2 R. E. Mulvey, D. R. Armstrong, B. Conway, E. Crosbie,
A. R. Kennedy and S. D. Robertson, Inorg. Chem., 2011, 50,
1
3 R. Campbell, E. Crosbie, A. R. Kennedy, R. E. Mulvey,
R. A. Naismith and S. D. Robertson, Aust. J. Chem., 2013,
6
008, 130, 16193–16200.
(
(
Doctoral Training Grants to D. L. R. and E. C.) and EPSRC
grant award no. EP/K001183/1) for their generous sponsorship
1
1
of this research. Dr C. O'Hara is also thanked for his useful
insights into the work as well as Dr J. A. Parkinson and Craig
Irving for their help and advice regarding the NMR spectro-
scopic studies.
2241–12251.
6, 1189–1201.
1
4 A. Unsinn, S. H. Wunderlich, A. Jana, K. Karaghiosoff and
P. Knochel, Chem.–Eur. J., 2013, 19, 14687–14696.
5 T. Klatt, K. Groll and P. Knochel, Chem. Commun., 2013, 49,
6953–6955.
References
1
1
(a) J. Clayden, Organolithiums: Selectivity for Synthesis,
Elsevier, Oxford, 2002; (b) M. Schlosser, Organometallics in 16 J. Garcia-Alvarez, E. Hevia, A. R. Kennedy, J. Klett and
Synthesis, Third Manual, Wiley, Hoboken, NJ, 2013. R. E. Mulvey, Chem. Commun., 2007, 2402–2404.
2
(a) M. Lappert, P. Power, A. Protchenko and A. Seeber, Metal 17 For full details see page S27 of ref. 11b.
Amide Chemistry, Wiley, Chichester, 2009; (b) R. E. Mulvey 18 (a) A. Macchioni, G. Ciancaleoni, C. Zuccaccia and
and S. D. Robertson, Angew. Chem., Int. Ed., 2013, 52,
1470–11487.
(a) A. Krasovskiy and P. Knochel, Angew. Chem., Int. Ed.,
D. Zuccaccia, Chem. Soc. Rev., 2008, 37, 479–489; (b) D. Li,
I. Keresztes, R. Hopson and P. G. Williard, Acc. Chem. Res.,
2009, 42, 270–280.
1
3
2004, 43, 3333–3336; (b) A. Krasovskiy, V. Krasovskaya and 19 (a) J. Mason, Multinuclear NMR, Plenum Press, New York and
P. Knochel, Angew. Chem., Int. Ed., 2006, 45, 2958–2961; (c)
P. Garcia-Alvarez, D. V. Graham, E. Hevia, A. R. Kennedy,
J. Klett, R. E. Mulvey, C. T. O'Hara and S. Weatherstone,
Angew. Chem., Int. Ed., 2008, 47, 8079–8081.
London, 1987; (b) R. Benn, E. Janssen, H. Lehmkuhl and
A. Runska, J. Organomet. Chem., 1987, 333, 155–168; (c)
R. J. Less, H. R. Simmonds and D. S. Wright, Dalton Trans.,
2014, 43, 5785–5792.
4
5
B. Haag, M. Mosrin, H. Ila, V. Malakhov and P. Knochel, 20 (a) I. Krossing, H. N ¨o th, C. Tacke, M. Schmidt and
Angew. Chem., Int. Ed., 2011, 50, 9794–9825.
(a) R. E. Mulvey, Organometallics, 2006, 25, 1060–1075; (b)
R. E. Mulvey, F. Mongin, M. Uchiyama and Y. Kondo,
H. Schwenk, Chem. Ber./Recueil, 1997, 130, 1047–1052; (b)
I. Krossing, H. N ¨o th and H. Schwenk-Kircher, Eur. J. Inorg.
Chem., 1998, 927–939.
Angew. Chem., Int. Ed., 2007, 46, 3802–3824; (c) 21 B. Conway, E. Hevia, J. Garcia-Alvarez, D. V. Graham,
R. E. Mulvey, Acc. Chem. Res., 2009, 42, 743–755; (d)
A. Harrison-Marchand and F. Mongin, Chem. Rev., 2013,
A. R. Kennedy and R. E. Mulvey, Chem. Commun., 2007,
5241–5243.
3044 | Chem. Sci., 2014, 5, 3031–3045
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
22 (a) M. C. Whisler, S. MacNeil, V. Snieckus and P. Beak, 39 For prime examples see the following paper and others
Angew. Chem., Int. Ed., 2004, 43, 2206–2225; (b)
V. Snieckus, Chem. Rev., 1990, 90, 879–933.
3 G. Wittig and G. Fuhrmann, Ber. Dtsch. Chem. Ges., 1940, 73,
referenced therein: F. Chevallier, T. Blin, E. Nagaradja,
F. Lassagne, T. Roisnel, Y. S. Halauko, V. E. Matulis,
O. A. Ivashkevich and F. Mongin, Org. Biomol. Chem., 2012,
10, 4878–4885.
2
1197–1218.
24 H. Gilman and R. L. Bebb, J. Am. Chem. Soc., 1939, 61, 109– 40 P. Garcia-Alvarez, R. E. Mulvey and J. A. Parkinson, Angew.
1
12.
Chem., Int. Ed., 2011, 50, 9668–9671.
41 P. G. Williard and Q.-Y. Liu, J. Am. Chem. Soc., 1993, 115,
3380–3381.
2
2
5 See page S22 of ref. 11b.
6 (a) M. F. Lappert, M. J. Slade, A. Singh, J. L. Atwood,
R. D. Rogers and R. Shakir, J. Am. Chem. Soc., 1983, 105, 42 D. R. Armstrong, A. R. Kennedy, R. E. Mulvey and
02–304; (b) E. Hevia, A. R. Kennedy, R. E. Mulvey, S. D. Robertson, Chem.–Eur. J., 2011, 17, 8820–8831.
D. L. Ramsay and S. D. Robertson, Chem.–Eur. J., 2013, 19, 43 S. Harder, J. Boersma, L. Brandsma, G. P. M. van Mier and
4069–14075. J. A. Kanters, J. Organomet. Chem., 1989, 364, 1–15.
7 E. Crosbie, A. R. Kennedy, R. E. Mulvey and S. D. Robertson, 44 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
3
1
2
Dalton Trans., 2012, 41, 1832–1839.
8 For full details see page S23 of ref. 11b.
9 E. Crosbie, P. Garcia-Alvarez, A. R. Kennedy, J. Klett,
R. E. Mulvey and S. D. Robertson, Angew. Chem., Int. Ed.,
M. A. Robb, J. R. Cheeseman, J. A. Montgomery Jr,
T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam,
S. S. Iyengar, V. Tomasi, B. Barone, B. Mennucci, M. Cossi,
G. Scalmani, N. Rega, G. A. Petersson, H. Nakatsuji,
M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa,
M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai,
M. Klene, X. Li, J. E. Knox, H. P. Hratchian, J. B. Cross,
V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts,
R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi,
C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morokuma,
G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzewski,
S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas,
D. K. Malick, A. D. Rabuck, K. Raghavachari,
J. B. Foresman, J. V. Ortiz, Q. B. Cui, A. G. Baboul,
S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu,
A. Liashenko, P. Piskorz, I. Komaromi, R. L. Martin,
D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng,
A. Nanayakkara, M. Challacombe, P. M. W. Gill,
B. Johnson, W. Chen, M. W. Wong, C. Gonzalez and
J. A. Pople, Gaussian, Inc., Wallingford, CT, 2004.
2
2
2010, 49, 9388–9391.
3
0 B. Conway, J. Garcia-Alvarez, E. Hevia, A. R. Kennedy,
R. E. Mulvey and S. D. Robertson, Organometallics, 2009,
28, 6462–6468.
3
1 D. R. Armstrong, P. Garcia-Alvarez, A. R. Kennedy,
R. E. Mulvey and S. D. Robertson, Chem.–Eur. J., 2011, 17,
6725–6730.
3
3
2 R. E. Mulvey, Dalton Trans., 2013, 42, 6676–6693.
3 R. B. Bates, L. M. Kroposki and D. E. Potter, J. Org. Chem.,
1
972, 37, 560–562.
4 J. Q. Wen and J. B. Grutzner, J. Org. Chem., 1986, 51, 4220–
224.
5 P. Renaud and M. A. Fox, J. Am. Chem. Soc., 1988, 110, 5702–
705.
3
3
3
4
5
6 K. Snegaroff, J.-M. L'Helgoual'ch, G. Bentabed-Ababsa,
T. T. Nguyen, F. Chevallier, M. Yonehara, M. Uchiyama,
A. Derdour and F. Mongin, Chem.–Eur. J., 2009, 15, 10280– 45 (a) W. Kohn, A. D. Becke and R. G. Parr, J. Phys. Chem., 1996,
1
0290.
100, 12974–12980; (b) A. D. Becke, Phys. Rev. A: At., Mol., Opt.
Phys., 1988, 38, 3098–3100.
46 (a) A. D. McLean and G. S. Chandler, J. Chem. Phys., 1980, 72,
5639–5648; (b) R. Krishnan, J. S. Binkley, R. Seeger and
J. A. Pople, J. Chem. Phys., 1980, 72, 650–654.
3
7 W. Clegg, B. Conway, E. Hevia, M. D. McCall, L. Russo and
R. E. Mulvey, J. Am. Chem. Soc., 2009, 131, 2375–2384.
8 For prime examples see the following paper and others
referenced therein: E. Nagaradja, F. Chevallier, T. Roisnel,
V. Dorcet, Y. S. Halauko, O. A. Ivashkevich, V. E. Matulis
and F. Mongin, Org. Biomol. Chem., 2014, 12, 1475–1487.
3
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 3031–3045 | 3045