Lithium choreography: Intramolecular arylations of carbamate-stabilised carbanions and their mechanisms probed by in situ IR spectroscopy and DFT calculations

Article abstract of DOI:10.1002/chem.201201761

Deprotonation of O-allyl, O-propargyl or O-benzyl carbamates in the presence of a lithium counterion leads to carbamate-stabilised organolithium compounds that may be quenched with electrophiles. We now report that when the allylic, propargylic or benzylic carbamate bears an N-aryl substituent, an aryl migration takes place, leading to stereochemical inversion and C-arylation of the carbamate α to oxygen. The aryl migration is an intramolecular S _NAr reaction, despite the lack of anion-stabilising aryl substituents. Our in situ IR studies reveal a number of intermediates along the rearrangement pathway, including a "pre-lithiation complex," the deprotonated carbamate, the rearranged anion, and the final arylated carbamate. No evidence was obtained for a dearomatised intermediate during the aryl migration. DFT calculations predict that during the reaction the solvated Li cation moves from the carbanion centre, thus freeing its lone pair for nucleophilic attack on the remote phenyl ring. This charge separation leads to several alternative conformations. The one having Li⁺ bound to the carbamate oxygen gives rise to the lowest-energy transition structure, and also leads to inversion of the configuration. In agreement with the IR studies, the DFT calculations fail to locate a dearomatised intermediate. Dancing to lithium's tune: In situ IR spectroscopy and DFT calculations allow the detailed pathways of aryl migrations taking place in lithiated carbamates to be characterised (see scheme).

Full text of DOI:10.1002/chem.201201761

DOI: 10.1002/chem.201201761
Lithium Choreography: Intramolecular Arylations of Carbamate-Stabilised
Carbanions and Their Mechanisms Probed by In Situ IR Spectroscopy and
DFT Calculations
Anne M. Fournier,[a] Christopher J. Nichols,[b] Mark A. Vincent,[a] Ian H. Hillier,*[a] and
Jonathan Clayden*[a]
Abstract: Deprotonation of O-allyl, O-
propargyl or O-benzyl carbamates in
the presence of a lithium counterion
leads to carbamate-stabilised organo-
lithium compounds that may be
quenched with electrophiles. We now
report that when the allylic, propargylic
or benzylic carbamate bears an N-aryl
substituent, an aryl migration takes
place, leading to stereochemical inver-
sion and C-arylation of the carbamate
a to oxygen. The aryl migration is an
intramolecular SNAr reaction, despite
the lack of anion-stabilising aryl sub-
stituents. Our in situ IR studies reveal
a number of intermediates along the
gration. DFT calculations predict that
during the reaction the solvated Li
cation moves from the carbanion
centre, thus freeing its lone pair for nu-
cleophilic attack on the remote phenyl
ring. This charge separation leads to
several alternative conformations. The
one having Li+ bound to the carba-
mate oxygen gives rise to the lowest-
energy transition structure, and also
leads to inversion of the configuration.
In agreement with the IR studies, the
DFT calculations fail to locate a dearo-
matised intermediate.
rearrangement pathway, including
a
“pre-lithiation complex,” the deproto-
nated carbamate, the rearranged anion,
and the final arylated carbamate. No
evidence was obtained for a dearoma-
tised intermediate during the aryl mi-
Keywords: carbamates
·
density
functional calculations · IR spectro-
scopy · lithium · organolithium · re-
action mechanisms · rearrangement
Introduction
The synthesis of compounds 1, containing quaternary stereo-
genic centres bearing a heteroatom, is a challenge that has
been met in a number of ways,[1–4] most of which fall into
two classes: 1) the enantioselective addition of a nucleophile
to a prochiral p system (a ketone or imine, for example),[1]
or 2) the stereospecific elaboration of an existing tertiary
stereogenic centre (Scheme 1).[4] Both methods offer advan-
tages, with the latter allowing the formation of stereogenic
centres in an enantiomerically pure fashion even when the
centre carries groups that are sterically and electronically
similar. We have shown that benzylic carbamates,[5,6]
ureas[7–9] and thiocarbamates[10] 2, containing N-aryl substitu-
ents, undergo an N to C aryl migration reaction, allowing in-
[a] A. M. Fournier, Dr. M. A. Vincent, Prof. I. H. Hillier,
Prof. J. Clayden
Scheme 1. Quaternary centres carrying heteroatoms: aryl migration in
benzyllithium compounds.
School of Chemistry, University of Manchester
Oxford Road, Manchester, M13 9PL (UK)
Fax : (+44) 161-275-4939
E-mail: clayden@man.ac.uk
troduction of an aryl substituent a to the benzylic heteroa-
tom (Scheme 1). With derivatives of enantiomerically pure
secondary alcohols or thiols, or amines with secondary sub-
stituents, the migration may be stereospecific, generating a
product 3 containing a new quaternary stereogenic centre
bearing an acylated O, N or S substituent. The stereospeci-
ficity is dependent on the configurational stability of the in-
[b] Dr. C. J. Nichols
GlaxoSmithKline, Gunnels Wood Road
Stevenage, SG1 2NY (UK)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201201761. It contains experi-
mental data for preparation and characterisation of all new com-
pounds, including copies of H and 13C NMR spectra.
1
16478
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 16478 – 16490

FULL PAPER
termediate organolithium 2Li over the timescale of the rear-
rangement, and is high for ureas, slightly lower for thiocar-
bamates and significantly lower for carbamates. Deprotec-
tion in basic[6] or neutral conditions[8] reveals an a-arylated
amine,[2] alcohol or thiol[3] product. Aggarwal et al. have re-
ported[4] a method for the synthesis of tertiary alcohols
which also proceeds through lithiation of a benzylic carba-
mate. Trapping with a borane or a boronic ester proceeds
with stereochemical retention or inversion, and a concomi-
tant 1,2-shift, entailing departure of the carbamate, leads to
formation of a chiral borane or boronate that may be oxi-
dised to a tertiary alcohol.
Prior to this work, it was well established that not only
benzylic[11] but also allylic and propargylic carbamates may
be deprotonated to form synthetically versatile organolithi-
um compounds of defined stereochemistry.[12] Hoppe et al.
used a-metallated 2-alkenyl N,N-diisopropylcarbamates 4Li
as homoenolate equivalents in regio- and stereospecific re-
actions with aldehydes and ketones (Scheme 2).[13] Enan-
ate, which is stable in solution and adds to aldehydes and
ketones to form homoaldol products 5 with >95:5 d.r.[16]
Deprotonation of (E)-cinnamyl N,N-diisopropylcarbamate
(4c) with n-butyllithium/(À)-sparteine in toluene shows that
the cinnamyllithium 4cLi compound formed upon deproto-
nation epimerises within thirty minutes at À788C to a ther-
modynamically controlled ratio of diastereoisomeric allyl-
lithium–(À)sparteine complexes. Reactions with several
electrophiles proceed regio- and stereospecifically at the a-
or g-position but at different rates for the two diastereom-
ers, leading to a decrease in enantioselectivity compared to
the diastereomeric ratio of the complexes.[17a] Lithiation of
cinnamyl carbamates and their 1-naphthyl analogues in the
presence of chiral bisoxazoline (BOX) ligands results in a
greater discrimination between the diastereoisomers than
with (À)-sparteine, and gives enantioenriched substitution
products with ee values up to 94%.[17b]
(Z)-1-Alken-1-yl N,N-diisopropylcarbamates 6 bearing an
anion-stabilising group in the 1-position are deprotonated
by n-butyllithium/(À)-sparteine with a high degree of selec-
tivity between the enantiotopic protons at the g-position to
form allyllithium intermediates 6Li that are configurational-
ly stable at À708C (Scheme 3). These combine with ketones
and aldehydes with high regio- and stereoselectivities (up to
97% ee), allowing a simple and efficient approach to enan-
tiomerically enriched homoenolate reagents starting from
achiral precursors.[18]
In the alkyne series, enantiomerically enriched, configura-
tionally stable propargyllithium compounds may be pro-
duced by deprotonation of enantiomerically pure secondary
Scheme 2. Allylic carbamates in homoaldol reactions.
tioenriched (a-lithio-2-alkenyl)-
carbamates derived from secon-
dary carbamates, such as (E)-2-
pentenyl N,N-diisopropylcarba-
mate (4a), are configurationally
stable below À708C in non-
polar solvents. They can be gen-
erated either by deprotonation
Scheme 3. Enantiotope-differentiating g-deprotonation of 1-alkenyl carbamates.
of the optically active precur-
sors by n-butyllithium/tetrame-
thylethylenediamine (TMEDA) in diethyl ether/hexane[14] or
by kinetic resolution of racemic precursors, deprotonating
with sec-butyllithium/(À)-sparteine in pentane. The allyl-
lithium is formed with greater than 80% ee, as estimated
from trapping experiments.[15]
In contrast, (a-lithio-2-alkenyl)carbamates 4Li derived
from carbamate derivatives of primary alcohols, such as (E)-
2-butenyl N,N-diisopropylcarbamate (4b)[16] or (E)-cinnamyl
N,N-diisopropylcarbamate (4c),[17] turn out to be configura-
tionally unstable in solution with butyllithium/(À)-sparteine
even at À788C. In the case of (E)-2-butenyl N,N-diisopro-
pylcarbamate (4b), however, the sparteine complex crystalli-
ses from a pentane/cyclohexane solution with concomitant
dynamic kinetic resolution, resulting in up to 96:4 d.r. in the
2-alkynyl N,N-diisopropylcarbamates with n-butyllithium/
TMEDA in hexane at À788C.[19] Deprotonation of primary
2-alkynyl N,N-diisopropylcarbamates by n-butyllithium/(À)-
sparteine in toluene at À788C and trapping the intermediate
ion pairs by carboxylation or silylation leads to slightly
enantioenriched products. However, when the deprotona-
tion is carried out in pentane, dynamic resolution of the lith-
ium–(À)-sparteine complexes by selective crystallisation
gives much higher levels of enantioselectivity (up to
96:4 e.r.).[20]
Lithiated allylic and propargylic carbamates have thus
proved their worth in stereoselective synthesis in which the
organolithium is trapped by an electrophile such as a car-
bonyl compound or alkylating agent. However, there are
currently no methods allowing the stereoselective arylation
of lithiated allyl or propargyl carbamates. We therefore set
solid. Reaction of the solid with Ti
ACHTUNGTERN(NNUG OiPr)4 proceeds with
configurational inversion to give the allyltitanium intermedi-
Chem. Eur. J. 2012, 18, 16478 – 16490
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16479

I. H. Hillier, J. Clayden et al.
out to explore the potential of
aryl migration within the lithiat-
ed carbamate as a means to
achieve this aim, bearing in
mind that aryl migration in
lithiated ureas is an effective
method for the synthesis of hin-
dered benzylic amines.[2,7] In
this paper, we describe the
scope and limitations of the
aryl migration reactions as a
means of making tertiary allylic
or propargylic alcohols. We also
describe spectroscopic (in situ
IR) and computational (DFT)
investigations into the mecha-
nism of the reaction, including
its stereospecificity for inver-
sion.
Scheme 4. Aryl migration in lithiated allyl carbamates: preliminary experiments
Results and Discussion
Scope of the rearrangement: In
preliminary experiments to
assess the reactivity of lithiated
O-allyl-N-aryl carbamates to-
wards rearrangement, we treat-
ed a series of substituted O-
allyl-N-aryl carbamates 7[21]
with a base (Scheme 4). The
simplest member of the set, 7a
(N-methyl aniline protected as
Scheme 5. Proposed pathway to rearrangement products 9–12. Solvation of the lithium cation is assumed but
not explicitly shown.
its
N-allyloxycarbonyl
(N-
Alloc) derivative), decomposed under basic conditions, but
7b underwent lithiation followed by a 1,2-acyl shift from ni-
trogen to carbon to give compound 9—a known reaction of
lithiated carbamates,[22] and one we have previously ob-
served as a side reaction in N to C aryl migrations.[23] More
encouraging results were obtained from O-cinnamyl carba-
mate 7c on treatment with sBuLi (2.5 equiv) in THF at
À788C. After 1 h, the principal product of the reaction
(41% yield) was the aryl ketone 11, a compound that can
only arise by migration of the N-phenyl ring from N to C,
and a minor product was enol carbamate 10, from which 11
can be derived by hydrolysis.
Scheme 6. Rearrangement of an a-methylated allyl carbamate.
A plausible pathway from 7c to 10, and hence 11, is
shown in Scheme 5. Deprotonation of 7 to give the allyllithi-
um 7Li is followed by aryl migration to yield 12Li. A
second deprotonation gives 10Li, and g-reprotonation
during the workup yields the vinyl carbamate 10.
To capture the initial product of rearrangement, 12Li,
without the complicating features of the second deprotona-
tion, we replaced the cinnamyl derivative (7c) with its a-me-
thylated analogue 7d. On deprotonation with sBuLi
(Scheme 6), the principal isolated product 14 (23%) con-
tained a sec-butyl group (Table 1, entry 1). This compound
presumably arises from aryl migration to give 12 followed
by carbolithiation[24] of the styrenic double bond. Traces of a
second product 13, formed in 6% yield, suggested rear-
rangement to 12 (Scheme 5), followed by loss of the carba-
mate.
Deprotonation of 7d with lithium diisoproylamide (LDA)
in THF and quenching with MeOH gave, after 30 min, a
compound identified as 12 in the NMR spectrum of the
crude reaction mixture (Table 1, entry 2). However, this ma-
terial proved extremely unstable under the conditions re-
16480
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 16478 – 16490

Lithium Choreography
FULL PAPER
Table 1. Choice of base for rearrangement of allyl and propargyl carbamates.
Entry
S.M.
Base
Product, yield [%]
e.r.[a]
1
2
3
4
5
6
7
8
7d
7d
sBuLi (2.5 equiv), THF, À788C, 1 h
13, 6; 14, 23
–
–
LDA (2 equiv), THF, À788C, 30 min
12[b]
7d or (R)-7d
(R)-7d
(R)-7d
8b or (S)-8b
8c or (S)-8c
8b
8b
8b
8d
LDA (2 equiv), THF, À788C, 30 min; then tBuONO (6 equiv), RT, 16 h
LDA (2 equiv), Et2O, À788C, 2 h; then tBuONO (6 equiv), RT, 16 h
LDA (2 equiv), Et2O, (À)-sparteine, À458C, 2 h; then tBuONO (6 equiv), RT, 16 h
LDA (2 equiv), THF, À788C, 15 min
13, 68
50:50
54:46[d]
62:38
50:50
50:50
–
–
–
–
(À)-(S)-13[c]
(À)-(S)-13[c]
(E)-16b, 45
LDA (2 equiv), THF, À788C, 15 min
(E)-16c, 70
LiHMDS (2 equiv), THF, 08C, 15 min
(E)-16b, 14; (Z)-16b, 31
(E)-16b, 7; (Z)-16b, 44
(E)-16b, 16; (Z)-16b, 16
15d, 66
9
10
11
NaHMDS (2 equiv), THF, 08C, 15 min
KHMDS (2 equiv), THF, 08C, 15 min
LDA (2.5 equiv), THF, DMPU, À788C, 1 h
[a] The e.r. was determined from the crude reaction mixture. [b] Observed in the crude reaction mixture. [c] Configuration assigned by comparison with
a reported optical rotation: [a]2D5 =À1.5 (c=1.0 in CHCl3); literature value [a]D20 =À5.33 (c=0.6 in Et2O] (see ref. [26]). [d] The remaining starting materi-
al was recovered with 77:23 e.r. HMDS=hexamethyldisilazide, DMPU=1,3-dimethyltetrahydropyrimidin-2(1H)-one.
quired for purification, presumably due to the ready forma-
tion of a highly stabilised cation under acidic conditions. By
adding instead tert-butyl nitrite to form an N-nitroso deriva-
tive in situ[25] and stirring the resulting basic reaction mix-
ture for 24 h at room temperature, the alcohol 13 could be
obtained in 68% yield (Table 1, entry 3). Under these condi-
tions the aryl migration clearly takes place cleanly and in
good yield, the applicability of the reaction being limited
only by the instability of the product.
The related propargyl carbamates behaved in a compara-
ble manner. As with 7a, primary O-propargylcarbamate 8a,
on treatment with base, produced products of attack directly
on the carbamate C=O group. However, with the a-methy-
lated analogues of this compound 8b–d, treatment with base
led to migration of the aryl ring from N to C (Scheme 7),
Figure 1. X-ray crystal structure of (E)-16b.
structure of the cyclised product (E)-16b (Figure 1). The
cyclisation, although interesting, detracts from the possible
utility of the rearrangement. Nevertheless, the formation of
these compounds also indicates that the aryl migration can
be initiated by organometallic intermediates of potassium
and sodium, something not observed previously.
By starting with enantiomerically pure 7d, we hoped to
be able to form the allyllithium compound 7dLi as a single
enantiomer that would rearrange to 12dLi faster than it can
racemise. Enantiomerically enriched benzylcarbamates may
be lithiated and rearranged stereospecifically, provided con-
ditions are carefully chosen,[6] and carbamates closely relat-
ed to 7d have some degree of configurational stability.[14,28]
Compound (R)-7d was made by kinetic resolution[29] of its
Scheme 7. Rearrangement of O-propargylcarbamates.
allylic alcohol precursor. Treatment with LDA in THF
under the conditions used for the reaction in Scheme 6,
and in the case of 8d the rearranged carbamate 15 was iso-
lated in 66% yield. With the phenyl-substituted alkynes 8b
and 8c, the migration was followed by cyclisation of the re-
sulting carbamate anion 15Li onto the triple bond[27] to
yield the benzylidene oxazolidinones (E)- and (Z)-16
(Table 1, entries 6–10) in a ratio dependent upon the base
used in the cyclisation. X-ray crystallography confirmed the
however, returned the racemic product (Table 1, entry 3).
Since rearrangements in less coordinating solvents tend to
return higher enantiomeric excesses,[6,10] the reaction was re-
peated in diethyl ether. However, this reaction did not reach
completion in 2 h, and gave a product with a low enantio-
meric excess (Table 1, entry 4). Tellingly, the recovered start-
ing material was also partially racemised, suggesting that the
Chem. Eur. J. 2012, 18, 16478 – 16490
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16481

I. H. Hillier, J. Clayden et al.
lithiated intermediate is not
configurationally stable under
the reaction conditions. Dia-
mines such as (À)-sparteine or
TMEDA may affect the rate at
which
organolithium
com-
pounds racemise,[30] and indeed
adding either (À)-sparteine or
TMEDA (1 equiv) to the rear-
rangement in Et2O resulted in a
product retaining some degree
of enantiomeric enrichment
(Table 1, entry 5). The maxi-
mum achievable e.r. was little
more than 60:40, but the sign of
the optical rotation of the prod-
uct (À)-(S)-13 was consistent
with the literature value for this
compound[26] and confirmed the
absolute configuration of the
major enantiomer.[31] It appears
therefore that the rearrange-
ment of allyl carbamates, like
that of benzyl carbamates,[5,6]
proceeds with inversion of con-
Scheme 8. Effect of (À)-sparteine on the rearrangement of lithiated benzylcarbamates. [a] The absolute config-
uration of (S)-17 remains unassigned.
provides yet another mechanistic possibility for carrying out
figuration. This is in stark contrast with the related rear-
rangements of ureas[7,32] (including allyl ureas[23]) and thio-
carbamates.[10]
an enantioselective rearrangement—asymmetric deprotona-
tion. We had previously shown that asymmetric g-deproto-
nation of an N-carbamoyl enamine (an N-vinyl urea) with
chiral lithium amides generates N-substituted allyllithium
compounds enantioselectively.[23] The corresponding N-aryl-
(Z)-enol carbamates 19 were made by the method reported
by Feringa et al.:[35] the sodium enolates of the ketones 18
were O-acylated by addition of the carbamoyl chloride de-
rivatives of N-methyl aniline or N-methyl-4-chloroaniline
(Scheme 9).
Treatment of the O-vinylcarbamates 19 with LDA
(2 equiv) in THF gave the rearranged carbamates 20, deriv-
atives of tertiary doubly benzylic alcohols, in excellent yield
(Scheme 9), indicating that the reactivity towards migration
of the aryl ring from N to C of lithiated O-allyl, O-benzyl
and O-vinyl carbamates bearing N-aryl groups provides a
valuable method for arylating an organolithium centre adja-
cent to an oxygen atom from very easily made substrates.
Replacing LDA with chiral analogues 21 or 22[23] led to a
slower reaction, and levels of enantiomeric enrichment were
relatively poor, reaching a maximum of 68:32 (Table 2).
Again it seems likely that racemisation competes with rear-
rangement, and no further enantioselective reactions were
pursued with these compounds.
Similarly, enantiomerically enriched O-propargyl carba-
mates (S)-8b and (S)-8c were made by reduction of 4-phe-
nylbut-3-yn-2-one by using the method reported by Noyori
and co-workers.[33] As expected, in THF at À788C, the rear-
rangement/cyclisation of (S)-8b or (S)-8c with LDA gave
only racemic product (Table 1, entries 6 and 7). Replacing
THF with Et2O diverted the course of the reaction of 8b to-
wards attack directly on the carbamate C=O group, whereas,
with 8c, Et2O completely shut down the reaction. At various
temperatures, with or without addition of (À)-sparteine,
decay of the ee of the starting material, by racemisation of
its lithium derivative, was observed.
The use of (À)-sparteine as a means of reducing the rate
of racemisation[30] relative to the rate of rearrangement was
explored as a way of improving the results obtained from
the rearrangements of benzyl carbamates 2 described previ-
ously (Scheme 8).[5] Addition of (À)-sparteine to the rear-
rangement of (R)-2b improved the e.r. of the product, (S)-
3b, from 90:10 to 97:3, but at the expense of the yield,
whereas addition of (À)-sparteine to the less reactive (S)-2c
induced a remarkable change in the mechanism of the reac-
tion, which produced the 2-hydroxyamide (S)-17, by a 1,2-
acyl shift, in 80:20 e.r., instead of the aryl migration product
(S)-3c, generated in the absence of (À)-sparteine.[34]
Identification of reaction intermediates by in situ infrared
spectroscopy: The presence of the carbonyl group makes
carbamates particularly suitable for mechanistic studies by
in situ IR spectroscopy. Recently, OꢁBrien et al.[36] reported
that the lithiation of N-tert-butoxycarbonyl (N-Boc) piperi-
dine with sBuLi/TMEDA, (À)-sparteine or (+)-sparteine
could be followed by in situ infrared spectroscopy. A “pre-
N- or O-Substituted allyllithium compounds may be
formed either by a-deprotonation of an allylamine/allyl al-
cohol derivative or alternatively by g-deprotonation of an
enamine/enol derivative.[28] The advantage of the latter
method is the lack of chirality in the starting material, which
16482
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 16478 – 16490

Lithium Choreography
FULL PAPER
the rearranged product. To ex-
plore the pathway from starting
material to product in lithiated
carbamates, we therefore un-
dertook detailed in situ IR
studies, in a range of solvents,
of the rearrangements of repre-
sentative allylic, vinylic and
benzylic carbamates, 7d, 19b
and 2c.
In a flask equipped with the
probe of a React IR machine,
7d was treated with LDA
(1.5 equiv) at À608C in THF
(Scheme 10). On addition of
LDA, the C=O stretching fre-
quency of n˜ =1709 cmÀ1, corre-
sponding to the starting carba-
Scheme 9. g-Deprotonation-mediated N to C aryl transfer in enolcarbamates.
Table 2. Lithiation and rearrangement of (Z)-19a and (Z)-19b.
Starting
material
Base
T
[8C]
t
Product
Yield
[%]
e.r.
[h]
(Z)-19a
(Z)-19b
(Z)-19a
(Z)-19a
(Z)-19a
(Z)-19b
(Z)-19b
LDA[a]
LDA[a]
21[b]
À78
À45
À78
À78
À60
À78
À60
1
1
3
5
3
5
3
20a
20b
20a
20a
20a
20b
20b
80
–
–
80
28[d]
72
68:32
56:44
54:46
52:48
52:48
22[c]
21[b]
50[e]
18[d]
62
21[b]
21[b]
[a] In THF. [b] From the amine and nBuLi. [c] From the amine hydro-
chloride and 2ꢂnBuLi. [d] Starting material remaining. [e] Formation of
a side-product.
Scheme 10. Proposed intermediates in the rearrangement of 7d. Solva-
tion of the lithium cation is assumed but not explicitly shown.
lithiated complex” (n˜C O =1675 cmÀ1) between the substrate
=
and the lithiating agent was observed in tert-butyl methyl
ether (TBME) at À788C prior to formation of the lithiated
piperidine (n˜C O =1644 cmÀ1).[36] Beak et al. also monitored
mate 7d, was replaced over 2 min by a band at n˜ =
1660 cmÀ1, which persisted until, after 45 min, the reaction
was quenched with methanol. At this point a peak at n˜ =
1768 cmÀ1 appeared, which corresponds to the product car-
bamate 12 (Figure 2). In this reaction, and under these con-
ditions, we saw no evidence for any other intermediates, and
we assume that the band at n˜ =1660 cmÀ1 corresponds to the
product anion 12Li, which forms by instantaneous rear-
rangement of the lithium derivative 7dLi.
The rearrangement of 19b (Scheme 11) provided more
detail about the mechanistic pathway, presumably because
the allyl anion generated by deprotonation is more stable
than that formed from 7d. Immediately upon addition of
LDA to 19b in THF at À608C, a new band at n˜ =1578 cmÀ1
appeared and lasted for only a few minutes, before giving
way to a band at n˜ =1649 cmÀ1 (Figure 3). This band, which
corresponds to the band at n˜ =1660 cmÀ1 in Figure 2, persist-
ed until the reaction is quenched, which generated the prod-
uct peak at n˜ =1735 cmÀ1. The band at n˜ =1649 cmÀ1 is
again presumably that corresponding to the product anion
20bLi, with the band at n˜ =1578 cmÀ1 presumably arising
from the lithiated starting material 19bLi.
=
the lithiation of an N-Boc allylamine through in situ IR
spectroscopy with nBuLi/(À)-sparteine in toluene at
À738C,[37] reporting a similar reaction pathway involving a
pre-lithiated complex with an IR peak at n˜ =1675 cmÀ1 and
a lithiated complex that appears at n˜ =1640 cmÀ1. Addition
of nBuLi to N-Boc allylamine generated a stable carba-
mate–nBuLi complex with an IR peak at n˜ =1675 cmÀ1 and
addition of (À)-sparteine led to deprotonation, as seen by
the appearance of the peak at n˜ =1640 cmÀ1. Our own previ-
ous studies in this area initially explored the rearrangements
of ureas in THF by NMR spectroscopy[38a] and by in situ IR
spectroscopy,[38] and failed to find evidence of either a pre-
lithiation complex or of the presence of a postulated dearo-
matised intermediate with
a benzenoid ring migrating
(though a dearomatised intermediate was observed, and
indeed trapped, with a migrating naphthyl ring). However,
in THF, the rearrangement is fast. We recently found[38b]
that a pre-lithiation complex can be detected by in situ IR
spectroscopy during a rearrangement involving vinyl migra-
tion, but only in Et2O or toluene. Again, no intermediate
could be found between the lithiated starting material and
Chem. Eur. J. 2012, 18, 16478 – 16490
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16483

I. H. Hillier, J. Clayden et al.
carbamate 2c was the most informative about the detailed
reaction pathway (Figure 4 and Scheme 12). In THF or
TBME, compound 2c showed a C=O absorbance at n˜ =
1698 cmÀ1. Treatment with sBuLi (1.1 equiv) in THF trans-
Figure 2. The in situ IR spectra of the rearrangement of 7d during the re-
action using LDA in THF at À608C (a three-dimensional plot of absorb-
ance versus wavenumber versus time).
Scheme 11. Intermediates detected in the rearrangement of 19b (Ar=p-
ClC6H4). Solvation of the lithium cation is assumed but not explicitly
shown.
Figure 4. The in situ IR spectra of the rearrangement of 2c during the re-
action using s-BuLi at À608C. a) A three-dimensional plot of absorbance
versus wavenumber versus time and b) a difference spectrum, created by
subtracting the initial spectrum of 2c.
formed this signal directly into one at n˜ =1642 cmÀ1, which
we assign to 3cLi, and addition of MeOH returned the car-
bamate 3c with
a C=O stretching frequency of n˜ =
1735 cmÀ1. A similar peak, at n˜ =1646 cmÀ1 was observed
when LDA (1.5 equiv) was used as the base. Evidence that
the absorption at n˜ =1642 cmÀ1 arises from the product
anion 3cLi was obtained by treating 3c with sBuLi in THF;
the spectrum assigned to 3cLi reappears under these condi-
tions (see the Supporting Information).
In the absence of THF, the rearrangement was much
slower (Figure 4a). Treatment of 2c in TBME at À608C
with sBuLi (2 equiv) gave a transient absorbance at n˜ =
Figure 3. A three-dimensional infrared profile for the rearrangement of
19b by using LDA in THF at À608C.
Although the evolution of the IR spectra of 7c and 19b
gave useful information about the lithiated intermediates in-
volved in the reaction, and their relative lifetimes, we found
that in situ IR monitoring of the rearrangement of benzyl
16484
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 16478 – 16490

Lithium Choreography
FULL PAPER
peak at n˜ =1735 cmÀ1, corresponding to the rearranged car-
bamate 3c. The entire course of the reaction is illustrated in
Figure 4a, with the intermediates highlighted more clearly
in Figure 4b, which shows a difference spectrum, created by
subtracting the initial spectrum of 2c.
It is notable that n˜C for lithiated carbamate 2cLi and
=
O
the product carbamate anion 3cLi are very similar. Howev-
er, the spectrum of 3cLi also shows a characteristic peak at
n˜ =1333 cmÀ1, which is absent from the spectrum of 2cLi.
As shown in the Supporting Information, this peak grew
progressively with time, allowing us to monitor the progress
of the rearrangement pathway. No other intermediates were
observed along the reaction pathway.
Computational strategy and details: To shed light on the
mechanistic features of the aryl migrations described herein
and building on our previous work,[5] appropriate models
were constructed involving the benzylic carbamate 2a in its
metallated form 2aLi and the base LDA, with THF as the
solvent. The potential-energy surface for the reaction was
obtained by using electronic-structure calculations employ-
ing DFT. Properly characterised minima and transition
structures (TSs) were located at the B3LYP/6-31G level,
giving both vibrational frequencies and starting structures
for optimisation at the B3LYP/6-31+ +G** level. The struc-
tures and energetics quoted herein are for the larger basis,
with thermodynamic and zero-point corrections from the
smaller basis. All calculations were carried out with the
Gaussian suite of programmes.[39] In our previous study of
the stereospecific intramolecular arylation of lithiated
ureas,[38] we found that the major features of the potential-
energy surfaces for both retention and inversion of stereo-
chemistry are unchanged when the M06-2X, rather than the
B3LYP, functional is employed. For this reason, we have not
investigated the effects of different functionals in this case.
We first note that a crystal structure of an LDA–THF
complex[40] shows a dimer with each Li+ ion being coordi-
nated to a single THF molecule and sharing two amide li-
gands with the other Li+ ion. In the model used here, the
carbonyl oxygen of the carbamate 2a replaces one of the
THF molecules of this dimer. Thus, our model (Scheme 13)
involves one explicit LDA dimer coordinated to the carbon-
yl oxygen of the carbamate 2a (A), a structure correspond-
ing to the pre-lithiated complex indicated by the in situ IR
studies. Furthermore, we assume that deprotonation to form
2aLi initially gives the reactant structure B in which the car-
banion is bound to a solvated lithium ion. In view of the ex-
perimental observation that excess base facilitates the reac-
tion our model has a dimer of LDA coordinated to the car-
bonyl oxygen atom. In our corresponding study of lithiated
ureas,[38] we found that LDA facilitated the breaking of the
intramolecular Li-O(=C) bond in B, a change in structure
necessary for the substrate to adopt the reactive conforma-
tion. As the precise degree of solvation of the lithium ion is
unknown, we have considered a model in which the ion is
solvated by three THF molecules. As in our previous study
of lithiated ureas,[38] we do not consider further implicit sol-
Scheme 12. Intermediates proposed in the rearrangement of 2c. Solvation
of the lithium cation is assumed but not explicitly shown.
1686 cmÀ1, which then gave way over a period of minutes to
a band at n˜ =1649 cmÀ1. These bands were assigned by
quenching the reaction with MeOH once the band at n˜ =
1649 cmÀ1 had appeared. Since this quenching reaction re-
generated the starting material 2c and not product 3c, we
assign the absorbance at n˜ =1649 cmÀ1 to a structure corre-
sponding to 2cLi, and propose that the transient absorption
at n˜ =1686 cmÀ1 is the pre-lithiated complex 2c·RLi.[36,37]
When 2c was treated with sBuLi in toluene at À608C, only
the absorbance at n˜ =1686 cmÀ1 was observed, indicating
that under these conditions, complexation may take place,
but not deprotonation. After a reaction time of 30 min, ad-
dition of TMEDA (1 equiv) indicated regeneration of the
starting material 2c, probably by decomplexation of the spe-
cies 2·RLi. A similar outcome was obtained by using LDA
(3.5 equiv) in TBME at À608C, which produced a band at
n˜ =1690 cmÀ1. After stirring for 1 h, addition of CD3OD to
the reaction mixture resulted in the recovery of the starting
material 2c with no deuterium incorporation (see the Sup-
porting Information). Computational results presented
below indicate that the absorption at n˜ =1649 cmÀ1 is best
modelled by a complex [2Li·RLi]b in which the intramolec-
ular interaction in 2Li or the alternative structure
[2Li·RLi]a is absent (Scheme 12). This unexpected result
appears to arise from steric interactions between the second
equivalent of LDA and the solvated lithium bond to the car-
banion centre.
Conversion of the solution of 2cLi formed from 2c with
sBuLi in TBME to 3cLi was achieved by adding 4 equiva-
lents of THF to the reaction mixture in order to promote
the rearrangement. Little change in the IR spectrum was ob-
served, but adding MeOH after 90 min generates a new
Chem. Eur. J. 2012, 18, 16478 – 16490
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16485

I. H. Hillier, J. Clayden et al.
to three different reactant con-
formations that were higher in
energy than the lowest-energy
reactant we have located, (R4,
Figure 5d). In R4, Li+ is clearly
bound to the carbanion centre
(at a distance of 2.20 ꢃ). In the
three reactant configurations
this distance is increased con-
siderably to free up the carban-
ion lone pair for nucleophilic
attack. This charge separation
is facilitated by solvation of the
lithium cation, and explains
why coordinating solvents, such
as THF and DMPU, accelerate
the reaction. In Figure 5, we
show both the reactants and the
corresponding transition struc-
tures and it is clear how the dif-
ferent reactant conformations
give rise to the different transi-
tion structures. The reactant
structure leading to retention of
configuration (R3, Figure 5c) is
of considerably higher energy
than the lowest energy mini-
Scheme 13. Structures along the computed reaction pathways.
mum (R4, Figure 5d; by
82 kJmolÀ1), whilst those lead-
vation in our model, in line with the findings of Streitwieser
et al.,[41] who found that the integral equation polarisable
continuum model (IEPCM) was poor at predicting the sol-
vation energies of a range of organic species in THF, and in
the absence of bulk solvation, pK(Li) values for a range of
lithium compounds were well predicted.
ing to inversion (R2, Figure 5b and R1, Figure 5a) are
higher by 36 and 46 kJmolÀ1, respectively (Table 3). The sta-
bility of R4 relative to intramolecularly coordinated struc-
tures such as B, which are usually assumed for such lithiated
carbamates (Scheme 12), results from the reduction in the
steric interaction between the LDA coordinated to the car-
bonyl group and the solvated lithium cation bound to the
carbanion, and the associated increase in entropy.
We now discuss the three mechanisms involving these re-
actant and transition structures. For retention of configura-
Our strategy to understand the mechanism of the 1,4-aryl
shift was to search for the transition structures (TSs), and to
follow these back to the reactants, and forward to the prod-
ucts. This revealed the corresponding reactant structures
and the changes that occur as the TSs are formed from
them. Transition structures for the reactions leading to aryl
migration were located that gave products in which either
retention or inversion of configuration at the carbanion
centre had occurred. We found two types of structure lead-
ing to inversion of stereochemistry. In one of the two struc-
Table 3. Relative energies [kJmolÀ1] of stationary structures.
Reaction and stereospecificity
Structure
Free
energy
lowest-energy minimum
retention[a]
R4
R3
TS3
P3
R1
TS1
P1
R2
TS2
P2
R5
TS5
P5
0.0
81.5
81.3
tures, {LiACHTUNGTRENNUNG
(thf)3}+ was bound to the carbamate oxygen atom
À66.0
(Figure 5a, TS1), whereas in the other it was bound to the
phenyl ring adjacent to the carbanion centre (Figure 5b,
TS2). A third transition structure (Figure 5c, TS3) in which
the solvated lithium species was located between the two
phenyl groups was found that led to retention of stereo-
chemistry. We have previously shown[38] that in the case of
the corresponding reaction of lithiated ureas, retention of
stereochemistry occurs via a transition structure analogous
to TS3.
inversion (with Li+ migrating to O)
inversion (with Li+ migrating to the benzylic ring)
attack on C=O
45.7
62.4
À67.7
36.0
70.7
À48.3
60.7
75.1
À93.1
[a] TS3 is of marginally lower free energy than R3 due to zero-point and
thermodynamic corrections, but is of higher energy (by 16 kJmolÀ1) if
electronic energy alone is considered.
Our strategy of determining the transition structures, and
then following these back to the corresponding reactant, led
16486
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 16478 – 16490

Lithium Choreography
FULL PAPER
phenyl ring. This migration is facilitated by the conjugation
of the carbanion lone pair with the phenyl group, further en-
hanced by interaction of the phenyl ring with the lithium
cation. Such conjugation restricts the free rotation of the
phenyl ring and thus provides a “memory” of the configura-
À
tion at the carbanion centre, despite the lack of a C Li
bond. Thus, although the stereogenic carbanion centre has
become planar by the migration of the Li+ ion, the original
stereochemical disposition of the groups around it is main-
tained.
Both transition structures TS3 and TS2, leading to reten-
tion and inversion, respectively, involve the interaction of
the solvated Li+ species with the benzyl ring. Structure TS2
(Figure 5b) differs from TS3, which was previously dis-
cussed, in that the lithium ion lies on the other face of the
phenyl group, and thus is unable to stabilise the developing
negative charge on the ring suffering nucleophilic attack. In
structure TS2, the length of the forming C···C bond is
1.93 ꢃ, which is shorter than in the transition structure that
leads to retention of stereochemistry (TS3). The transition
structure TS2 leads to inversion of stereochemistry (TS2,
À
Figure 5b) because the forming C C bond is orientated in
À
the opposite direction to the original C Li bond. In transi-
À
tion structure TS3, C C bond formation is in the same di-
rection as the original Li carbanion bond and leads to re-
À
tention of configuration.
In the transition structure TS1 and the corresponding re-
actant (R1, Figure 5a), the {LiACHTUNGTRNEUNG
(thf)3}+ is bound to and in the
plane of the carbamate group, requiring a 1,2-shift of the
Li+ cation. This Li+ shift from the carbon atom to the in-
plane oxygen lone pair requires conformational changes
from the minimum-energy reactant structure so that the Li+
is now in the plane of the carbamate and adjacent to the
oxygen lone pair. We note that the other lone pair on
oxygen is not available, being conjugated with the carbonyl
group. Following the transfer of Li+ to the oxygen atom, the
carbanion is neatly arranged for reaction. Indeed, the car-
banion is slightly pyramidal, as is the phenyl carbon atom
that is attacked, and thus the bond has started to form. This
mechanism gives rise to inversion because moving the lithi-
um into the plane of the carbamate, prior to its transfer to
À
the oxygen atom, results in the Li C bond pointing away
from the remote phenyl group.
We now consider the energetics of the three reactions we
have identified (Table 3). We find that the transition struc-
tures for retention (TS3) and inversion (TS1 and TS2) lie,
respectively, 81, 62 and 71 kJmolÀ1 above the lowest energy
reactant structure (R4). However, the corresponding three
reactant structures (R1, R2 and R3) are of considerably
higher energy than this lowest energy structure (R4) so that
the barriers for reaction from these structures are considera-
bly lower, all being below 35 kJmolÀ1. Thus, our calculations
clearly predict that inversion is favoured. TS3, leading to re-
tention, is higher in energy by 19 kJmolÀ1 than the more
favoured route (via TS1) for inversion, which involves a Li+
Figure 5. Optimal structures along the reaction pathways. For clarity
LDA and THF have been omitted (except in R4). a) Reactant R1 and
transition structure TS1 leading to inversion of stereochemistry; b) reac-
tant R2 and transition structure TS2 leading to inversion of stereochemis-
try; c) reactant R3 and transition structure TS3 leading to retention of
stereochemistry; and d) the lowest energy reactant, R4.
tion to take place, via TS3 (Figure 5c), the solvated Li+ ion
in the reactant structure (R4) has to migrate from the car-
banion to the adjacent ring (R3), thereby freeing the car-
banion lone pair for nucleophilic attack on the remote
À
O interaction. However, we note that the alternative tran-
sition structure for inversion, TS2, having a Li+ phenyl in-
À
Chem. Eur. J. 2012, 18, 16478 – 16490
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16487

I. H. Hillier, J. Clayden et al.
Table 4. Harmonic vibrational frequencies [cmÀ1] and intensities [kmmolÀ1] for stationary structures.
Species
Mode
Frequency
Intensity
carbamate 2a
C=O stretch
1693
344
ring modes
mixed C=O and ring modes
ring modes
mixed C=O and ring C H modes
mixed C=O stretch and ring modes
ring modes
N=C(-O)-N modes
phenyl ring stretches
C=O
1643–1666
1652, 1672
1644–1665
1638, 1642, 1650
1619, 1624
1578–1659
1702,1294
1636–1661
1745
3–105
363, 373
1-75
259, 263, 167
518, 274
23–201
350, 244
1–8
carbamate with LDA, 2a·RLi
À
R4
R1
P1
N-protonated product 3a
383
product anion coordinated to Li
(thf)3 through carbamate N 3Li
N-C(=O)-O
1335, 1654
237, 502
teraction, is only 8 kJmolÀ1 higher in energy, so that inver-
sion by both mechanisms is clearly feasible. Although this
energy difference is quite small, we can speculate as to the
origin of the preferred pathway, bearing in mind that SNAr
reactions usually need anion stabilising groups to facilitate
reaction. The key factor would appear to be the lack of sta-
bility of the anion centre once the Li+ is bound to the car-
bamate oxygen atom. This is in line with a comparison of
the stationary structures on the alternative pathways, and in
the two reactive conformers (R1 and R2), the bond-forming
data are for compound 2c, having a chlorine substituent on
one phenyl ring, which is absent in our model employing 2a.
We focus on the changes in the C=O stretching frequency
that we have observed experimentally during the rearrange-
ment of 2c. In the neutral reactant, the C=O stretch is com-
puted to occur at n˜ =1693 cmÀ1, close to the experimentally
determined band at n˜ =1698 cmÀ1. This good agreement
gives us confidence that we can understand the variation in
the C=O frequency along the reaction pathway. The addi-
tion of LDA leads to a peak at n˜ =1686 cmÀ1, which corre-
lates with the computed peak at n˜ =1672 cmÀ1 for structure
A (Scheme 13), corresponding to a pre-lithiated complex
having the LDA dimer coordinated to the carbonyl oxygen
atom. We predict a second nearby peak at n˜ =1652 cmÀ1, in
À
C C separation for the preferred pathway is shorter in the
reactant (R1) and longer in the transition structure (TS1).
Thus, in the more stable reactant structure (R2), the charge
on the conjugated carbanion is stabilised by the Li+ over
the phenyl ring, whilst the lower-energy transition structure
(TS1), which lacks this stabilisation and is earlier than the
alternative structure, readily undergoes nucleophilic attack
to yield a stable anion through aryl transfer.
À
which the C=O stretching band is mixed with the ring C H
modes. We have previously discussed that our calculations
have identified a low-energy lithiated carbamate structure
(R4), which, following conformational changes, gives the
structures that actually undergo reaction. This structure
In a corresponding study of the rearrangement of a lithiat-
ed urea, a dearomatised naphthyl intermediate was charac-
terised by NMR spectroscopy, and was predicted to be
stable by DFT calculations.[38] In the case of the carbamates
studied herein, no such intermediates were detected experi-
mentally, and they could not be located on our computed re-
action pathways. We therefore propose that this SNAr reac-
tion proceeds, unusually, without a dearomatised intermedi-
ate.
We have also studied the corresponding 1,2-aryl transfer
of the benzyl carbamate, which proceeds via a cyclic three-
membered transition structure, and may occur with N-alkyl,
rather than N-aryl, carbamates.[22] We find that this TS is
75 kJmolÀ1 above our lowest reactant minimum and is thus
13 kJmolÀ1 above our calculated barrier for the observed
1,4-aryl shift reaction. Thus, it would appear that the ener-
getic penalty of forming this strained cyclic transition struc-
ture is decisive in favouring the observed attack on the p
system of the phenyl group.
(R4) is predicted to have strong peaks at n˜ =1638, 1642 and
À1
À
1650 cm , which are mixtures of C=O and C H ring modes.
Experimentally, a peak at n˜ =1649 cmÀ1 is seen to grow fol-
lowing the addition of LDA to the benzyl carbamate 2c,
which we thus assume corresponds to our computed struc-
ture R4. We have found other structures, similar to R4, in-
À
cluding those with intramolecular Li O coordination as in
B, and computed their harmonic frequencies, as well as
those for the higher-energy structures (R1, R2, R3) that
lead to reaction. However, we find that none of these struc-
tures give the excellent prediction of the observed IR spec-
trum of lithiated 2c that is provided by structure R4.
For the product anion in which the solvated Li+ is coordi-
nated to the nitrogen atom of the carbamate, we predict in-
tense bands at n˜ =1335 and 1654 cmÀ1, values which are in
excellent agreement with the peaks observed experimentally
at n˜ =1333 and 1642 cmÀ1 for 3cLi. In an alternative struc-
ture (P1, the product from TS1, Scheme 13 and Figure 5), in
which the solvated cation remains bound to the carbamate
oxygen after aryl transfer, the corresponding computed fre-
quencies are n˜ =1294 and 1702 cmÀ1, indicating that, in the
final species generated on completion of the rearrangement,
lithium has migrated to the nitrogen atom of the carbamate
anion. After protonation of the anion, the computed carbon-
We now discuss the harmonic vibrational frequencies for
the various minimum-energy species that our computations
have identified along the reaction pathway of the benzyl car-
bamate (Table 4). Our aim is to use these values to aid in
the identification of the species found in the IR spectra we
have previously described. We note that the experimental
16488
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 16478 – 16490

Lithium Choreography
FULL PAPER
yl stretching frequency is n˜ =1745 cmÀ1, which is close to the
experimentally determined value of n˜ =1735 cmÀ1
Chem. Int. Ed. 1998, 37, 388; c) B. Goldfuss, Synthesis 2005, 2271;
d) F. Schmidt, R. T. Stemmler, J. Rudolph, C. Bolm, Chem. Soc. Rev.
2006, 35, 454; e) O. Riant, J. Hannedouche, Org. Biomol. Chem.
2007, 5, 873; f) M. Shibasaki, M. Kanai, Chem. Rev. 2008, 108, 2853;
g) S. Kobayashi, H. Ishitani, Chem. Rev. 1999, 99, 1069; h) D. A.
Cogan, J. A. Ellman, J. Am. Chem. Soc. 1999, 121, 268; i) R. Bloch,
Chem. Rev. 1998, 98, 1407.
Conclusion
[2] J. Clayden, M. Donnard, J. Lefranc, D. J. Tetlow, Chem. Commun.
2011, 47, 4624.
The rearrangement of lithiated carbamates carrying N-aryl
substituents, already reported for O-benzyl carbamates, is a
characteristic reactivity also of O-allyl, O-vinyl and O-prop-
argyl carbamates. The products of the rearrangements con-
tain quaternary oxygen-substituted centres, and the product
carbamates may be converted into tertiary alcohols. Racemi-
sation of the organolithium compounds means that products
with high enantiomeric ratios are difficult to obtain, but in
cases in which enantiomeric enrichment is preserved, inver-
sion of configuration during the rearrangement is observed.
A combination of in situ IR measurements and electronic-
structure calculations has allowed us to identify a number of
intermediates in the stereospecific intramolecular arylation
of lithiated O-benzyl carbamates. Coordination of the start-
ing carbamate with the base generates a pre-lithiated com-
plex, detectable by IR spectroscopy in the absence of THF
that has the LDA dimer coordinated to the carbonyl oxygen
atom. Deprotonation in coordinating solvents gives a lithiat-
ed carbamate, for which we propose a non-intramolecularly
coordinated structure on the basis of its modelled IR spec-
trum. The computations predict that movement of the sol-
vated Li+ is required to give new reactant structures that
differ in the location of the Li+ ion and can lead to different
stereochemical outcomes. The most energetically favoured
pathway is 1,2-migration of Li+ from the carbanion centre
to the adjacent carbamate oxygen atom, followed by an aryl
shift from N to C with inversion of configuration at the car-
banion centre. An alternative reactant structure, generated
by migration of the Li+ bound to the face of the adjacent
phenyl ring, also results in inversion of configuration but
with a slightly higher barrier. A reactant structure in which
the solvated Li+ ion is found between the two phenyl rings
gives rise to retention of configuration but with a rather
higher barrier. A comparison of the IR data with frequency
calculations gives good agreement and in two ambiguous
cases allows structures to be assigned to intermediates. Thus,
the lithiated starting material (in the presence of excess
[3] J. Clayden, P. MacLellan, Beilstein J. Org. Chem. 2011, 7, 582.
[4] a) J. L. Stymiest, V. Bagutski, R. M. French, V. K. Aggarwal, Nature
2008, 456, 778; b) H. K. Scott, V. K. Aggarwal, Chem. Eur. J. 2011,
17, 13124; c) M. Shimizu, Angew. Chem. 2011, 123, 6122; Angew.
Chem. Int. Ed. 2011, 50, 5998.
[5] J. Clayden, W. Farnaby, D. M. Grainger, U. Hennecke, M. Mancinel-
li, D. J. Tetlow, I. H. Hillier, M. A. Vincent, J. Am. Chem. Soc. 2009,
131, 3410.
[6] A. M. Fournier, R. A. Brown, W. Farnaby, H. Miyatake-Ondozabal,
J. Clayden, Org. Lett. 2010, 12, 2222.
[7] J. Clayden, J. Dufour, D. Grainger, M. Helliwell, J. Am. Chem. Soc.
2007, 129, 7488.
[8] J. Clayden, U. Hennecke, Org. Lett. 2008, 10, 3567.
[9] R. Bach, J. Clayden, U. Hennecke, Synlett 2009, 421.
[10] P. MacLellan, J. Clayden, Chem. Commun. 2011, 47, 3395.
[11] A. Carstens, D. Hoppe, Tetrahedron 1994, 50, 6097.
[12] Topics in Organometallic Chemistry, Vol. 5, Organolithiums in Enan-
tioselective Synthesis (Ed.: D. M. Hodgson), Springer, Berlin, 2003.
[13] a) D. Hoppe, R. Hanko, A. Brçnneke, F. Lichtenberg, Angew.
Chem. 1981, 93, 1106; Angew. Chem. Int. Ed. Engl. 1981, 20, 1024;
b) D. Hoppe, T. Krꢄmer, J.-R. Schwark, O. Zschage, Pure Appl.
Chem. 1990, 62, 1999; c) D. Hoppe, T. Hense, Angew. Chem. 1997,
109, 2376; Angew. Chem. Int. Ed. Engl. 1997, 36, 2282.
[14] a) D. Hoppe, T. Krꢄmer, Angew. Chem. 1986, 98, 171; Angew.
Chem. Int. Ed. Engl. 1986, 25, 160; b) T. Krꢄmer, D. Hoppe, Tetrahe-
dron Lett. 1987, 28, 5149.
[15] a) O. Zschage, J.-R. Schwark, D. Hoppe, Angew. Chem. 1990, 102,
336; Angew. Chem. Int. Ed. Engl. 1990, 29, 296; b) O. Zschage, J.-R.
Schwark, T. Krꢄmer, D. Hoppe, Tetrahedron 1992, 48, 8377.
[16] a) D. Hoppe, O. Zschage, Angew. Chem. 1989, 101, 67; Angew.
Chem. Int. Ed. Engl. 1989, 28, 69; b) O. Zschage, D. Hoppe, Tetrahe-
dron 1992, 48, 5657; c) H. Paulsen, C. Graeve, D. Hoppe, Synthesis
1996, 141.
[17] a) K. Behrens, R. Frçhlich, O. Meyer, D. Hoppe, Eur. J. Org. Chem.
1998, 2397; b) T. Hꢅmery, R. Huenerbein, R. Frohlich, S. Grimme,
D. Hoppe, J. Org. Chem. 2010, 75, 5716.
[18] a) M. Seppi, R. Kalkofen, J. Reupohl, R. Frçhlich, D. Hoppe,
Angew. Chem. 2004, 116, 1447; Angew. Chem. Int. Ed. 2004, 43,
1423; b) J. Reuber, R. Frçhlich, D. Hoppe, Org. Lett. 2004, 6, 783–
786; c) E.-U. Wꢆrthwein, D. Hoppe, J. Org. Chem. 2008, 73, 9055.
[19] S. Dreller, M. Dyrbusch, D. Hoppe, Synlett 1991, 397.
[20] C. Schultz-Fademrecht, B. Wibbeling, R. Frçhlich, D. Hoppe, Org.
Lett. 2001, 3, 1221.
[21] Carbamates were made by treating allylic alcohols with aryl isocya-
nates and methylating the products under standard conditions. See
the Supporting Information for further details.
[22] a) P. Zhang, R. E. Gawley, J. Org. Chem. 1993, 58, 3223; b) S. Super-
chi, N. Sotomayor, G. Miao, J. Babu, M. G. Campbell, V. Snieckus,
Tetrahedron Lett. 1996, 37, 6061; c) G. B. C. A. Slana, M. S. de Aze-
vedo, R. S. C. Lopes, C. C. Lopes, J. N. Cardoso, Beilstein J. Org.
Chem. 2006, 2, 1.
À
base) apparently lacks the intramolecular Li O interaction
generally proposed in lithiated carbamates, and the product
carbamate anion is an N-lithio, rather than an O-lithio, spe-
cies.
Acknowledgements
[23] D. J. Tetlow, U. Hennecke, J. Raftery, M. J. Waring, D. S. Clarke, J.
Clayden, Org. Lett. 2010, 12, 5442.
We are grateful to GlaxoSmithKline and the EPSRC for a studentship
(to A.M.F.) and to Dr. Nick Wooster for assistance with developing the
React IR methods.
[24] a) X. Wei, R. J. K. Taylor, Chem. Commun. 1996, 187; b) X. Wei, P.
Johnson, R. J. K. Taylor, J. Chem. Soc. Perkin Trans. 1 2000, 1109;
c) X. Wei, R. J. K. Taylor, Tetrahedron Lett. 1997, 38, 6467; d) X.
Wei, R. J. K. Taylor, Tetrahedron Lett. 1996, 37, 4209; e) H. Felkin,
G. Swierczewski, A. Tambutꢅ, Tetrahedron Lett. 1969, 10, 707; f) T.
Kato, S. Marumoto, T. Sato, I. Kuwajima, Synlett 1990, 671; g) S.
[1] a) P. I. Dosa, G. C. Fu, J. Am. Chem. Soc. 1998, 120, 445; b) E. J.
Corey, A. Guzman-Perez, Angew. Chem. 1998, 110, 2092; Angew.
Chem. Eur. J. 2012, 18, 16478 – 16490
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
16489

I. H. Hillier, J. Clayden et al.
Marumoto, I. Kuwajima, Chem. Lett. 1992, 1421; h) S. Marumoto, I.
Kuwajima, J. Am. Chem. Soc. 1993, 115, 9021.
[35] X.-B. Jiang, M. van den Berg, A. J. Minnaard, B. L. Feringa, J. G.
de Vries, Tetrahedron: Asymmetry 2004, 15, 2223.
[36] D. Stead, G. Carbone, P. OꢁBrien, K. R. Campos, I. Coldham, A.
Sanderson, J. Am. Chem. Soc. 2010, 132, 7260; see also: N. S.
Sheikh, D. Leonori, G. Barker, J. D. Firth, K. R. Campos, A. J. H. M.
Meijer, P. OꢁBrien, I. Coldham, J. Am. Chem. Soc. 2012, 134, 5300.
[37] D. J. Pippel, G. A. Weisenburger, N. C. Faibish, P. Beak, J. Am.
Chem. Soc. 2001, 123, 4919.
[25] A. M. Fournier, J. Clayden, Org. Lett. 2012, 14, 142.
[26] H. Li, P. J. Walsh, J. Am. Chem. Soc. 2005, 127, 8355.
[27] R. Ramesh, Y. Chandrasekaran, R. Megha, S. Chandrasekaran, Tet-
rahedron 2007, 63, 9153.
[28] J. Clayden, Organolithiums: Selectivity for Synthesis, Pergamon,
Oxford, 2002.
[38] a) D. M. Grainger, A. Campbell Smith, M. A. Vincent, I. H. Hillier,
A. E. H. Wheatley, J. Clayden, Eur. J. Org. Chem. 2012, 731; b) J.
Lefranc, A. M. Fournier, G. Mingat, S. Herbert, T. Marcelli, J. Clay-
den, J. Am. Chem. Soc. 2012, 134, 7286.
[29] V. S. Martin, S. S. Woodward, T. Katsuki, Y. Yamada, M. Ikeda,
K. B. Sharpless, J. Am. Chem. Soc. 1981, 103, 6237.
[30] T. I. Yousaf, R. L. Williams, I. Coldham, R. E. Gawley, Chem.
Commun. 2008, 97.
[39] Gaussian 09, Revision B.01, M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V.
Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X.
Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Son-
nenberg, 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.
Stratmann, 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. Dannenberg, S. Dapprich, A. D. Daniels, ꢇ.
Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, D. J. Fox, Gaussian,
Inc., Wallingford CT, 2009.
[31] Racemic products were also formed in the presence of OꢁBrienꢁs
(+)-sparteine surrogate [P. OꢁBrien, Chem. Commun. 2008, 655], or
when racemic starting materials were deprotonated in the presence
of an alkyllithium/sparteine complex, reactions carried out in an at-
tempt to induce kinetic resolution of the chiral starting material or
dynamic kinetic or thermodynamic resolution of the intermediate
organolithium [A. Basu, S. Thayumanavan, Angew. Chem. 2002,
114, 740; Angew. Chem. Int. Ed. 2002, 41, 716]. A recent publication
[M. Sasaki, H. Ikemoto, M. Kawahata, K. Yamagushi, K. Takeda,
Chem. Eur. J. 2009, 15, 4663] suggested that electron-donating
groups also reduce racemisation rates in benzyl- and allylithium
compounds, but rearrangement of the para-methoxy analogue of 7d
gave no improvement in the results.
[32] J. Clayden, M. Donnard, J. Lefranc, A. Minassi, D. J. Tetlow, J. Am.
Chem. Soc. 2010, 132, 6624.
[33] K. Matsumura, S. Hashigushi, T. Ikariya, R. Noyori, J. Am. Chem.
Soc. 1997, 119, 8738.
[34] The absolute configuration of the product of the 1,2-acyl shift is un-
known, but such 1,2-acyl shifts are reported to proceed with reten-
tion of configuration: K. Tomooka, H. Shimizu, T. Inoue, H. Shiba-
ta, T. Nakai, Chem. Lett. 1999, 759.
[40] P. G. Williard, J. M. Salvino, J. Org. Chem. 1993, 58, 1.
[41] A. Streitwieser, J. R. Reyes, T. Singhapricha, S. Vu, R. Shah, J. Org.
Chem. 2010, 75, 3821.
Received: May 18, 2012
Revised: August 29, 2012
Published online: October 23, 2012
16490
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 16478 – 16490