On the novel function of the additive DBU. Catalytic stereoselective deprotonation by a mixed dimer of lithiated DBU and a chiral lithium amide

Article abstract of DOI:10.1039/b205618f

The additive DBU is used to increase the selectivity and reactivity of e.g. chiral lithium amides in both catalysed and non-catalysed asymmetric syntheses. This has been attributed to the coordinating ability of DBU favoring more reactive aggregates. However, we have found that LDA in THF deprotonates DBU to yield lithiated DBU (1) as shown by multinuclear NMR studies. Furthermore, compound 1 is found to form a mixed dimer (5) with e.g. the norephedrine-derived chiral lithium amide 2. Results of an investigation of the stereoselectivity of this novel reagent in the epoxide deprotonation are also reported. Computational studies using PM3 and DFT show possible structures of 1 and 5 in line with the NMR results. In addition, the role of THF and DBU in the solvation of the aggregates has been investigated by computational modelling and favoured complexes in the equilibria between homo- and heterocomplexes are also reported.

Full text of DOI:10.1039/b205618f

View Article Online / Journal Homepage / Table of Contents for this issue
On the novel function of the additive DBU. Catalytic
stereoselective deprotonation by a mixed dimer of lithiated DBU
and a chiral lithium amide†
2
Daniel Pettersen, Mohamed Amedjkouh, Sten O. Nilsson Lill and Per Ahlberg*
Organic Chemistry, Department of Chemistry, Göteborg University, SE-412 96 Göteborg,
Sweden. E-mail: Per.Ahlberg@oc.chalmers.se; Fax: ϩ46 31 772 2908; Tel: ϩ46 31 772 2899
Received (in Cambridge, UK) 11th June 2002, Accepted 27th June 2002
First published as an Advance Article on the web 12th July 2002
The additive DBU is used to increase the selectivity and reactivity of e.g. chiral lithium amides in both catalysed and
non-catalysed asymmetric syntheses. This has been attributed to the coordinating ability of DBU favoring more
reactive aggregates. However, we have found that LDA in THF deprotonates DBU to yield lithiated DBU (1) as
shown by multinuclear NMR studies. Furthermore, compound 1 is found to form a mixed dimer (5) with e.g. the
norephedrine-derived chiral lithium amide 2. Results of an investigation of the stereoselectivity of this novel reagent
in the epoxide deprotonation are also reported. Computational studies using PM3 and DFT show possible structures
of 1 and 5 in line with the NMR results. In addition, the role of THF and DBU in the solvation of the aggregates has
been investigated by computational modelling and favoured complexes in the equilibria between homo- and
heterocomplexes are also reported.
NMR spectra obtained at 20 ЊC and Ϫ80 ЊC showed that DBU
Introduction
had been deprotonated at C6 to afford 1 as the product.18
Chiral lithium amides are being developed and used for highly
1H,1H-COSY and 1H,13C-HMQC experiments were performed
stereoselective deprotonations such as rearrangement of
on both DBU and 1 and assigned NMR chemical shifts are
epoxides to yield enantiomers of allylic alcohols for use in syn-
given in Table 1. 1H,13C- and 1H,1H-COSY and 1H,13C-HMQC
thesis of e.g. biologically active compounds.1–9 There have been
spectra are to be found in the Supporting Information.†19
a number of attempts to run such stereoselective deproton-
In another experiment, DBU was added to a 0.3 M solution
ations under catalytic conditions.10–12 The chiral lithium amide
of [6Li]LDA in THF-d8 and found to be deprotonated at the
has been used in catalytic quantities and the less reactive
6-position yielding [6Li]1. In Fig. 2, the 6Li NMR spectra
lithium diisopropylamide (LDA) has commonly been used as
obtained from titration of LDA by DBU at Ϫ75 ЊC are shown,
a bulk base to regenerate the chiral lithium amide.13–17 The
clearly indicating the formation of [6Li]1 (δ 0.71–1.32) at the
enantioselectivity is usually lower under catalytic conditions
than under non-catalytic conditions, presumably due to non-
stereoselective deprotonation by LDA. However, addition of
1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) to the reaction
mixture has in some cases been found to increase the
stereoselectivity and influence the rate. It has generally been
assumed that this is caused by specific solvation by DBU
expense of LDA (δ 2.11). However, DBU was not completely
deprotonated by LDA. An equilibrium between the react-
ants DBU/LDA and products 1/diisopropylamine (DIPA) was
obtained [eqn. (1)].
favoring more reactive aggregates.16
(1)
Results in this paper will demonstrate a new and more
intricate role of the DBU. We report that DBU is lithiated by
LDA or n-butyllithium (n-BuLi) under the conditions used in
asymmetric deprotonations e.g. in THF-solutions. Lithiated
DBU (1) is found to form a mixed dimer with a chiral lithium
amide (Fig. 1). Results of an investigation of the stereo-
selectivity of this novel reagent in epoxide elimination are also
reported. Computational studies show possible structures of
the reagents and the role of THF and DBU in solvation of the
reagent complex.
The singlet (δ 0.71–1.32) originating from 1 shifted downfield
on addition of DBU, indicating that 1 is solvated by DBU.
Formation of heterodimers
Below results are reported that show that in mixtures of 1 and
the chiral lithium amide 2 in THF, heterodimers are formed.
The 6Li spectra of a solution prepared from 0.025 mmol
(i.e. 0.125 equiv/Li) of the 15N labelled amine isotopologue 3b
of 2a and 0.20 mmol [6Li]LDA and 0.25 mmol DBU (Fig. 3a)
show besides the singlets originating from the remaining
[6Li]LDA and [6Li]1, respectively, two new doublets of equal
intensity at δ 0.69 and δ 1.42 [J(6Li,15N] = 5.30 Hz and 4.19 Hz).
The two doublets indicate the presence of a species with two
non-equivalent lithiums and that each lithium is coupled to
only one 15N. 1H and 13C spectra show that 3b had been depro-
tonated to provide 2b. However, it was not the homodimer 4 of
2 that had formed, since the lithiums in 4 give rise to two triplets
(see below).20 Moreover, the two doublets cannot be due to a
Results and discussion
Deprotonation of DBU
Addition of 1.0 equiv n-BuLi to a 0.3 M THF-d8 solution of
DBU held at 20 ЊC or Ϫ78 ЊC gave a yellow solution. 1H and 13
C
1
† Electronic supplementary information (ESI) available: H, 13C, and
1H,1H-COSY and 1H,13C-HMQC spectra. See http://www.rsc.org/
suppdata/p2/b2/b205618f/
DOI: 10.1039/b205618f
J. Chem. Soc., Perkin Trans. 2, 2002, 1397–1405
This journal is © The Royal Society of Chemistry 2002
1397

View Article Online
Fig. 1 Structures of labelled and non-labelled precursors and their lithiated counterparts and dimeric complexes.
mixed dimer of LDA and 2 since such a species is not formed in
THF upon mixing LDA and 2. Therefore, we propose that the
two new doublets arise from the mixed heterodimer 5 formed
from a monomer of 2 and a monomer of 1 (Fig. 1). In 5,
each 6Li is bonded to a single 15N, which is consistent with the
results.
coupling constant [J(6Li,15N) = 5.81 Hz] had shifted to lower
fields than the triplet with smaller coupling constant [J(6Li,15N)
= 3.75 Hz]. Evidently, the homodimer 4 is not a strong enough
base to deprotonate DBU. In addition, 1H and 13C NMR
spectra clearly show that deprotonation of DBU does not occur
under these conditions. (The 1H spectra show 2 protons at
δ 2.27, and the 13C spectra show tertiary carbons at δ 164).
Upon addition of 0.5 equiv [6Li]n-BuLi two doublets of equal
intensity appeared [δ 1.57, J(6Li,15N) = 4.19 Hz and δ 0.83,
The 6Li NMR spectrum, after addition of another 0.025
mmol 3b (i.e. in total 0.25 equiv/Li) shows doublets with
increased intensity (Fig. 3b). In Fig. 3c, the 6Li spectrum
obtained after addition of 0.075 mmol 3b (0.375 equiv/Li) is
shown. Obviously, the concentration of the heterodimer 5 has
increased—mainly at the expense of 1 rather than LDA. The
carbanionic compound 1 is a stronger base than LDA, which is
also shown by the equilibrium spectra in Fig. 2. After addition
of 0.1 mmol of the [15N]-labelled diamine 3b (i.e. in total
0.5 equiv/lithium), the singlets from [6Li]LDA and [6Li]1 are
essentially absent and now the heterodimer 5 is observed as
the main lithiated species [J(6Li,15N) = 5.21 Hz and 4.19 Hz,
respectively] (Fig. 3d). Thus, the carbanionic compound 1 is
a strong enough base to deprotonate the diamine 3 to yield
the novel heterodimer 5 built from a monomer of 2 and a
monomer of 1. The heterodimer 5 is also formed from amine
3 and lithiated DBU 1 generated from n-BuLi in absence of
LDA.
1
J(6Li,15N) = 5.22 Hz] (Fig. 4b). Furthermore, the H and 13C
NMR spectra indicated that deprotonation of DBU had
yielded the heterodimer 5. Addition of another 0.5 equiv [6Li]-
n-BuLi resulted in the 6Li spectra shown in Fig. 4c. Apparently,
the heterodimer now is the dominating lithiated component in
the solution. The high field doublet (δ 0.90) being broader than
the other doublet (δ 1.60) indicates slow exchange involving
solvating DBU. Upon addition of up to 0.5 equiv DBU to the
above solution (Fig. 4d) the high field doublet became sharper
probably due to faster ligand exchange at the high concen-
tration of DBU. Further addition of 0.5 equiv DBU resulted in
6
the Li NMR spectrum shown in Fig. 4e consistent with the
proposed solvent effect. The coupling constants [J(6Li,15N)] for
the two doublets are 5.19 Hz and 4.19 Hz, respectively.
Thus these NMR findings show that DBU is deprotonated
both by LDA and n-BuLi in THF solutions and that lithiated
DBU (1) is involved in formation of a heterodimer together
with a chiral lithium amide.
6
The Li NMR spectrum of 0.15 M solution of 2b (Fig. 4a)
prepared from equal amounts of [6Li]n-BuLi and 3b displays
two triplets (δ 1.55 and δ 1.90) of equal intensity but with differ-
ent splittings [J(6Li,15N) = 5.81 Hz and 3.75 Hz, respectively]
originating from the homodimeric structure 4 of lithium amide
2. Addition of DBU to this solution shifted only the two triplets
to lower fields, presumably as a result of Li-solvation by DBU.
After addition of 1.0 equiv DBU the triplet with the larger
In a previous search for new useful bulk bases to be used
in catalytic asymmetric deprotonations we have investigated
1-methylimidazole (6) and 1,2-dimethylimidazole (7) as bulk
base precursors.21,22 These compounds were like DBU found to
be deprotonated by n-BuLi and yielded the carbenoid species
1398
J. Chem. Soc., Perkin Trans. 2, 2002, 1397–1405

View Article Online
Table 1 Assigned 1H and 13C NMR chemical shifts at Ϫ75 ЊC in THF-d8 for selected atoms in 2, 3, DBU, 1 and 5
Atom
δ
δ
δ
δ
δ
H at C2
H at C3
H at C4
H at C5
H at C6
H at C9
H at C10
H at C11
C2
C3
C4
C5
C6
3.15
1.53
1.60
1.55
2.27
3.12
1.70
3.17
53.36
29.89
31.00
27.50
38.15
3.14
1.60
1.45
2.00
2.68
3.06
1.68
3.10
54.76
30.66
30.19
28.17
59.89
60.72
165.4
C7
C9
C10
C11
160.1
49.45
24.09
49.19
164.9
45.33
28.34–30.51
50.97
C12
C13
C14
45.3
77.4
149.0
35.8
67.9
143.3
45.35
149.8
C15
C16
72.6
54.3–54.6
67.1
52.7–53.0
54.1–54.3
Table 2 Stoichiometric and catalytic deprotonation of cyclohexene oxide 14
Entry
[14]/M
[DBU]/M
[DIPA]/M
0.2
[6]/M
[7]/M
[3]/M
[13]/M
[n-BuLi]/M
Reaction time/h
Yield (%)a
ee (%)b
1
2
3
4
5
6
7
8
9
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.5
0.5
0.5
0.2
0.5
0.5
0.02
0.02
0.22
0.22
0.22
0.3
0.3
0.22
0.1
0.22
0.22
0.3
56
69
7
240
60
7
93
82
89
57
88
81
78
96
96
96
96
69
16
79
89
94
74
80
93
93
96
96
0.02
0.1
0.1
0.2
0.02
0.1
8
0.2
0.2
0.02
0.02
0.1
198
13
25
25
0.2
0.2
10
11
0.1
0.3
a The yield of 15 was determined by using internal standard. b Determined by chiral GC.
2-lithio-1-methylimidazole (8) and the carbanionic species
2-(lithiomethyl)-1-methylimidazole (9), respectively (Fig. 1).
The role of the bases 8 and 9 in the reactions was also found to
be more intricate than just acting as bulk bases. Together with
the chiral lithium amide catalyst they were found to form the
heterodimers 10 and 11, respectively.
in enantiomeric excess (Fig. 1).13,23 It has been used in stoichio-
metric amounts to deprotonate cyclohexene oxide 14 to yield
(S)-cyclohex-2-en-1-ol (S)-15 in 80% ee (Table 2, entry 7).
Running the reaction using 12 in catalytic amounts together
with 1 as bulk base generated from DBU and n-BuLi, (S)-15 is
formed in 79% ee (entry 3). On the other hand, when lithium
amide 12 is used in catalytic amounts in presence of excess
LDA and DBU, the ee decreased to 74% ee (entry 6). The
decrease in selectivity in the presence of LDA is probably due to
the background reaction by LDA. These findings indicate that 1
is functioning as catalyst-regenerating base and participates in a
catalytic cycle.
Stereoselective deprotonation of cyclohexene oxide
The base lithium (S)-2-(pyrrolidin-1-ylmethyl)pyrrolidinide 12
derived from the amine 13 reported by Asami et al. has been
used in several synthetic procedures to generate allylic alcohols
J. Chem. Soc., Perkin Trans. 2, 2002, 1397–1405
1399

View Article Online
Fig. 3 6Li NMR spectra obtained at Ϫ75 ЊC of a THF-d8 solution
prepared from 0.20 mmol [6Li]LDA and 0.25 mmol of DBU titrated
with diamine 3b: a) 0.025 mmol diamine 3b added; b) 0.05 mmol
diamine 3b added; c) 0.075 mmol diamine 3b added; and d) 0.1 mmol
diamine 3b added.
the choice of bulk base for a particular chiral ligand is signifi-
cant. Amide 2 is most stereoselective in the presence of bulk
base 9 and amide 12 is most stereoselective in the presence of
bulk base 1. Furthermore, the use of LDA is not required to
obtain a catalytic cycle when employing 1, 8 or 9.
The generality of this behavior is currently being
investigated.24
Computational investigations
Fig. 2 6Li NMR spectra obtained at Ϫ75 ЊC of a THF-d8 solution of
0.3 M [6Li]LDA in the absence and presence of added DBU: a) 0.3 M
[6Li]LDA; b) 0.025 equiv DBU added; c) 0.5 equiv DBU added; d) 0.75
equiv DBU added; e) 1.0 equiv DBU added; and f ) 1.25 equiv DBU
added.
Monomer and homodimer structures of lithiated DBU
The PM3 and DFT optimised structures of the monomer
of C6-lithiated DBU 1 are given in Fig. 5. The structures
of homodimer 16 of 1 are shown in Fig. 6 together with the
THF- or DBU-solvated structures. The role of the solvent was
examined by specific THF- and DBU-solvation. Conform-
ational searches were performed to find the most stable con-
formation of the rings of DBU together with searches with
respect to rotation around Li-solvent bonds. Atom numbering
is given in Table 1.
Upon deprotonation–lithiation of DBU at C6, the monomer
1 is formed. DFT-optimisation (B3LYP/6–31ϩG(d)) of the
structure results in lithium-coordination to N8 (1.9 Å), C6
(2.0 Å) and C7 (2.3 Å), thus tricoordination to the azaallylic
system. Optimisation of 1 using PM3 results in weaker
coordination by lithium to the azaallylic system than at DFT-
level as seen from the longer bond distances. Deprotonation
Employing the homodimer 4 as reagent in the deprotonation
of 14 yields, as previously reported, (S)-15 in 93% ee (entries
8–9).20 Running the reaction using 2 in catalytic amounts
together with 1 as bulk base generated from DBU and n-BuLi,
(S)-15 is formed in 69% ee (entry 1) due to change of reagent
i.e. to the heterodimer 5. If LDA is used as bulk base instead of
1, in the reaction described above, the ee is drastically decreased
to 16% (entry 2). When employing stoichiometric amounts of
lithium amide 2 and bulk base 1 in THF the selectivity increases
to 89% ee in the absence of excess DBU (entry 4) and to 94% ee
in the presence of excess DBU (entry 5). Employing com-
pounds 8 or 9, higher enantiomeric excesses were obtained as
previously reported (entries 10–11).21,22 These results show that
1400
J. Chem. Soc., Perkin Trans. 2, 2002, 1397–1405

View Article Online
Fig. 5 PM3 and DFT (B3LYP/6–31ϩG(d)) optimised structures of
DBU and lithiated DBU 1.
Homodimer solvation
Specific solvation of the dimer 16 by one THF (16ؒTHF) at
PM3-level resulted in loss of the allylic carbon–lithium
coordination in one of the monomeric units (Fig. 6). Li1 is
solvated by THF (2.0 Å) and coordinates also to an imine
nitrogen (2.1 Å) and a carbanionic carbon (2.3 Å). This makes
the carbanionic carbon pentavalent. The other lithium (Li2) is
tetracoordinated via the azaallylic moiety and the imine nitro-
gen (2.0 Å). In the azaallylic unit not having carbanion–lithium
coordination, the bond distances are similar to those found
in the monomer 1. The solvation enthalpy is calculated to be
Ϫ5.2 kcal molϪ1
.
With monosolvation of 16 by DBU, the bond distances to
lithium are all similar compared to the THF-solvated dimer.
The distance between Li1 and the solvent is 2.1 Å, somewhat
longer than for THF. The solvation enthalpy is found to be
Ϫ14.7 kcal molϪ1 at PM3//PM3-level.
In the THF-disolvated dimer 16ؒ(THF)2 the lithiums have
only coordination to the nitrogens in the azaallylic moieties.
Each lithium coordinates to two imine nitrogens at 2.1 Å, and a
solvent molecule at 2.0 Å. In the azaallylic system the C6–C7
bonds are shortened by 0.14 Å to 1.36 Å upon loss of coordin-
ation to lithium and thus have more double bond character,
while the C7–N8 bonds are elongated by 0.11 Å to 1.44 Å,
i.e. have less double bond character.
Disolvation by DBU has a similar effect on the dimer as
found for disolvation by THF: only lithium–nitrogen coord-
inations could be detected. The lithium–DBU distances are
calculated to be 2.1 Å.
The solvation enthalpy from a second solvating THF is
found to be Ϫ4.6 kcal molϪ1 and the corresponding enthalpy
for the second DBU is Ϫ13.2 kcal molϪ1. From results pre-
sented in a previous study we expect the solvation enthalpies for
THF to be underestimated using PM3, while the solvation
enthalpies for DBU are probably overestimated.22
Fig. 4 6Li NMR spectra obtained at Ϫ75 ЊC of a THF-d8 solution
prepared from a solution of 0.15 M of 2b and added DBU and [6Li]n-
BuLi: a) 0.15 M 2b; b) 1.0 equiv DBU and 0.5 equiv [6Li]n-BuLi added;
c) 1.0 equiv DBU and 1.0 equiv [6Li]n-BuLi added; d) 1.5 equiv DBU
and 1.0 equiv [6Li]n-BuLi added; e) 2.0 equiv DBU and 1 equiv [6Li]n-
BuLi added.
followed by lithiation at other positions than at C6 in DBU
was also examined by PM3. All such isomers were found to
have higher enthalpy than 1, which is in agreement with the
experimental findings.
Formation of the homodimer 16 from two monomers of 1 is
strongly exothermic, Ϫ31.8 kcal molϪ1 at pBP86/DN*//PM3-
level and Ϫ26.7 kcal molϪ1 at PM3//PM3-level. The most stable
isomer at PM3//PM3-level is 16a, while at pBP86/DN*//PM3-
level isomer 16b is calculated to be 8.1 kcal molϪ1 more stable
than 16a. These two isomers differ in the conformations of
the ring system of DBU. Optimisation of 16a at B3LYP/6–
31ϩG(d)-level results in shorter bonds between lithium and the
azaallylic system than found at PM3-level. In the dimer each
lithium is tricoordinated to the azaallylic system of one mono-
mer and also coordinates the carbanionic carbon in the other
monomer. Each lithium in the dimer coordinates to two car-
banionic centers at 2.2–2.3 Å, to C7 at 2.4 Å and to an imine
nitrogen at 1.9 Å at DFT-level. In the dimeric azaallylic moiety
the Li–C7 bonds are 0.1 Å longer at DFT-level than in the
monomer, while the Li–C6 bonds in the dimer are 0.3 Å longer
than in the monomer using DFT.
Heterodimer structures
PM3- and DFT-optimised structures of the heterodimer 5 are
given in Fig. 7. The structures represent the most stable isomers.
Formation of heterodimer 5 from a monomer of lithium
amide 2 and a monomer of lithiated DBU (1) is exothermic by
Ϫ36.8 kcal molϪ1 at PM3//PM3-level. In 5, one lithium (Li1) is
dicoordinated while the other (Li2) is tricoordinated. The
N8–C7-bond is calculated to be 1.34 Å using DFT, similar to
J. Chem. Soc., Perkin Trans. 2, 2002, 1397–1405
1401

View Article Online
Fig. 6 PM3 and DFT (B3LYP/6–31ϩG(d)) optimised structures of homodimer 16 of lithiated DBU with and without solvation by THF and DBU.
the distance found in the homodimer 16. The C6–C7-bond is
slightly shorter in 5 than in the homodimer, 1.42 Å and 1.45 Å,
respectively. The imine N8–Li2 bond is found to be 1.93 Å,
while the carbanionic carbon coordinates to Li1 at a distance of
2.13 Å. No lithium coordination to C7 in the azaallylic system
was found in the heterodimer 5. The amide nitrogen originating
from the monomer of 2 coordinates to both lithiums at 1.88
and 1.98 Å, respectively. Li2 has a third nitrogen coordination
to the pyrrolidine nitrogen (2.1 Å).
and pBP86/DN*//PM3-level is Ϫ6.8 kcal molϪ1. The bonds to
the solvated lithium are somewhat elongated upon solvation.
Solvation of 5 by DBU results in a longer solvent–Li bond
(2.1 Å) compared to solvation by THF (2.0 Å). Other bonds are
almost unaffected by the solvent change. The solvation
enthalpy is found to be Ϫ16.0 kcal molϪ1 at PM3//PM3 level
and Ϫ8.3 at pBP86/DN*//PM3-level.
Disolvation by THF results in a structure without pyrrolidine
nitrogen–lithium coordination. The bond distance between Li2
and the second THF is 2.0 Å and the second solvation enthalpy
is calculated to be only Ϫ1.7 kcal molϪ1 at PM3 and Ϫ1.6 kcal
molϪ1 at pBP86/DN*//PM3-level.
Heterodimer solvation
Upon monosolvation of 5 by THF, Li1 rather than Li2 is
solvated. The Li–O(THF) distance at PM3-level is calculated to
be 2.0 Å and the solvation enthalpy at both PM3//PM3-level
Disolvation with DBU also results in a structure without
pyrrolidine nitrogen–lithium coordination. The solvent–lithium
bond distances are calculated to be 2.11–2.12 Å at PM3-
1402
J. Chem. Soc., Perkin Trans. 2, 2002, 1397–1405

View Article Online
Fig. 7 PM3 and DFT (B3LYP/6–31ϩG(d)) optimised structures of the heterodimer 5 and monomer of 2 unsolvated and solvated by THF.
Hydrogens are omitted for clarity.
level and the calculated second solvation enthalpy for DBU is
Ϫ8.0 kcal molϪ1 using PM3//PM3 but only Ϫ0.6 kcal molϪ1 at
pBP86/DN//PM3-level.
solvent effect upon the NMR-spectra as described previously.
Similar conclusions can be drawn for states B and C. The
disolvated homodimers, part of state C, are only 0.1 kcal molϪ1
less stable than state G.
In Scheme 2, the basicity of the homodimer 16 of lithiated
DBU 1 and homodimer 14 of lithium amide 2 is compared. It is
found that thermodynamically it is favourable to deprotonate
the diamine 3 with 16 rather than to deprotonate DBU with 4.
This correlates with the experimental observations.
In Scheme 1, studies of equilibria between homo- and
heterodimers with and without solvation by THF or DBU are
shown. Relative energies are calculated at pBP86/DN*//PM3
and pBP86/DN//PM3 levels of theory. State G, involving the
DBU-monosolvated heterodimer 5ؒDBU is used as the refer-
ence state. The DBU-disolvated heterodimer 5ؒ(DBU)2, part
of state H, is only 1.1 kcal molϪ1 more stable than state G.
This low solvation enthalpy indicates that the mono-solvated
heterodimer is preferred. State E, involving the THF-
monosolvated heterodimer, is found to be 2.9 kcal molϪ1
less stable than the reference state. Thus, DBU solvates the
heterodimer better than THF, in accordance with the observed
Conclusion
DBU is deprotonated by LDA in THF and the product lithiated
DBU (1) forms heterodimers with chiral lithium amides. Such
heterodimers are the new reagents in the present stereoselective
J. Chem. Soc., Perkin Trans. 2, 2002, 1397–1405
1403

View Article Online
at C2 (δ 1.73) and from C2 carbon (δ 25.57), respectively, in
the solvent THF-d8. Lithium resonances were referenced to
external [6Li] in a 0.3 M [6Li]Cl in MeOH-d4 (δ 0.0) in a separate
NMR tube. The probe temperature was measured using a
calibrated methanol thermometer from Varian Inc.
Typical NMR experiment
To THF-d8 (650 µl) in an NMR tube was added diiso-
propylamine (28 µl, 0.2 mmol). [6Li]LDA was prepared by titra-
tion of diisopropylamine with [6Li]n-BuLi (ca. 10 M, ca. 20 µl)
by monitoring the disappearance of the signal from the methine
protons in diisopropylamine at δ 2.85 and the appearance of the
signal from methine protons in [6Li]LDA at δ 3.05 at Ϫ75 ЊC.
DBU (7.5 µl, 0.05 mmol, 0.25 equiv) was added and the sol-
ution was allowed to equilibrate for 45 minutes before spectra
were recorded at Ϫ75 ЊC.
Scheme 1 pBP86/DN*//PM3 and pBP86/DN//PM3 calculated relative
enthalpies (kcal molϪ1) of states in equilibrium involving dimers with
varying degree of solvation.
Typical deprotonation experiment of 14
Amine 3a (4.4 µl, 0.02 mmol) and DBU (30 µl, 0.20 mmol) were
added to THF (867 µl) in a reaction vessel in the glove box.
After transfer of the vessel out of the glove box, n-BuLi (89 µl,
2.47 M in hexanes, 0.22 mmol) was added in a nitrogen atmos-
phere. The yellow solution was allowed to equilibrate at 20.00
0.05 ЊC for 10 minutes prior to addition of cyclohexene oxide 14
(10 µl, 0.10 mmol). To follow the reaction, 50 µl samples were
withdrawn from the reaction vessel at different intervals and
diluted with diethyl ether (500 µl). The solutions were quenched
in saturated NH4Cl (250 µl) and washed with brine (250 µl). The
reaction yield of 15 was determined by GC using a standard
added after the quenching as previously described.20
Scheme 2 pBP86/DN*//PM3 calculated relative enthalpies in kcal
molϪ1
.
deprotonations and presumably also in previous similar appli-
cations of DBU. The match between the bulk base and the
chiral lithium amide determines the stereoselectivity. Com-
pound 1 also functions as a bulk base in catalytic stereoselective
deprotonations. The additive DBU is also functioning as a
solvating ligand. Computational studies have given insights
into structures of reagents and solvation.
Computational details
Geometries were optimised at PM325,26 and B3LYP/6–31ϩG(d)
levels of theory.27–31 In Spartan32 the option HHON was used
to correct for hydrogens in close contact.33 All geometries were
characterised as minima on the potential energy surface (PES)
by the sign of the eigenvalues of the force constant matrix
obtained from a frequency calculation. Reaction energies were
calculated at pBP86/DN*//PM3,32,34,35 pBP86/DN//PM3, and
PM3//PM3-levels of theory.
Experimental
General
All syringes and glass vessels used were dried overnight in a
vacuum oven (50 ЊC) before being transferred into a glove box
(Mecaplex GB 80 equipped with a gas purification system that
removes oxygen and moisture) containing a nitrogen atmos-
phere. Typical moisture content was less than 0.5 ppm. All hand-
ling of the compounds were carried out with gas-tight syringes.
THF was distilled from sodium and benzophenone. The
concentration of the commercially available n-BuLi (ca. 2.5 M
in hexanes, Acros) was determined by titration.20 THF-d8 was
distilled in a vacuum line and stored over molecular sieves (4 Å)
in the glove box. Diisopropylamine, DBU, and cyclohexene
oxide were purified by distillation from CaH2 and were found to
be >99.5% pure by NMR and GC. [6Li]n-BuLi,21 amine 3a,20
3b,21 and 1323 were prepared as described. All GC analysis was
run on a chiral stationary phase column (CP-Chirasil-DEX CB,
25 m, 0.32 µm) from Chrompack. The column was held at 90 ЊC
(injector 225 ЊC, detector 250 ЊC) using helium (2 ml minϪ1) as a
carrier gas. tR(14) = 3.25 min, tR((S)-15) = 7.45 min, tR((R)-15) =
7.90 min.
Acknowledgements
We are grateful to the Swedish Natural Science Research
Council for support and we thank Senior Research Engineer
Charlotta Damberg and Docent Göran Hilmersson for
assistance with the NMR experiments.
References
1 J. K. Whitesell and S. W. Felman, J. Org. Chem., 1980, 45, 755.
2 P. J. Cox and N. S. Simpkins, Tetrahedron: Asymmetry, 1991, 2, 1.
3 N. S. Simpkins, Pure Appl. Chem., 1996, 68, 691.
4 P. O’Brien, J. Chem. Soc., Perkin Trans. 1, 1998, 1439.
5 P. O’Brien, J. Chem. Soc., Perkin Trans. 1, 2001, 95.
6 S. K. Bertilsson and P. G. Andersson, Tetrahedron, 2002, 58, 4665.
7 P. C. Brookes, D. J. Milne, P. J. Murphy and B. Spolaore,
Tetrahedron, 2002, 58, 4675.
8 A. Seki and M. Asami, Tetrahedron, 2002, 58, 4655.
9 S. E. De Sousa, P. O’Brien and C. D. Pilgram, Tetrahedron, 2002, 58,
4643.
10 J. P. Tierney, A. Alexakis and P. Mangeney, Tetrahedron:
Asymmetry, 1997, 8, 1019.
11 M. Amadji, J. Vadecard, J.-C. Plaquevent, L. Duhamel and
P. Duhamel, J. Am. Chem. Soc., 1996, 118, 12483.
12 T. Yamashita, D. Sato, T. Kiyoto, A. Kumar and K. Koga,
Tetrahedron, 1997, 53, 16987.
13 M. Asami, T. Ishizaki and S. Inoue, Tetrahedron: Asymmetry, 1994,
5, 793.
NMR spectroscopy
All NMR experiments were performed in Wilmad tubes (5 mm)
fitted with a Wilmad/Omnifit Teflon valve assembly (OFV) and
a Teflon/Silicon septum. 1D NMR spectra were recorded with a
Varian Unity 500 spectrometer equipped with a 5 mm triple-
resonance probe head custom built by Nalorac. Measuring
frequencies were 499.9 MHz (1H), 125.7 MHz (13C) and
73.57 MHz (6Li). 1H,1H-COSY and 1H,13C-HMQC NMR
1
spectra were recorded with a Varian Unity 600 MHz. The H
14 M. Asami, T. Suga, K. Honda and S. Inoue, Tetrahedron Lett., 1997,
38, 6425.
and 13C spectra were referenced to signals from residual protons
1404
J. Chem. Soc., Perkin Trans. 2, 2002, 1397–1405

View Article Online
15 M. J. Södergren and P. G. Andersson, J. Am. Chem. Soc., 1998, 120,
10760.
16 M. J. Södergren, S. K. Bertilsson and P. G. Andersson, J. Am. Chem.
Soc., 2000, 122, 6610.
17 S. K. Bertilsson, M. J. Södergren and P. G. Andersson, J. Org.
Chem., 2002, 67, 1567.
18 N. Matsumura, T. Ohba and S. Yoneda, Chem. Lett., 1983, 317.
19 J. W. Wiench, L. Stefaniak, E. Grech and E. b. Bednarek, J. Chem.
Soc., Perkin Trans. 2, 1999, 885.
20 D. Pettersen, M. Amedjkouh, S. O. Nilsson Lill, K. Dahlén and
P. Ahlberg, J. Chem. Soc., Perkin Trans. 2, 2001, 1654.
21 M. Amedjkouh, D. Pettersen, S. O. Nilsson Lill, Ö. Davidsson and
P. Ahlberg, Chem. Eur. J., 2001, 7, 4368.
22 S. O. Nilsson Lill, D. Pettersen, M. Amedjkouh and P. Ahlberg,
J. Chem. Soc., Perkin Trans. 1, 2001, 3054.
29 R. Krishnan, J. S. Binkley, R. Seeger and J. A. Pople, J. Chem. Phys.,
1980, 72, 650.
30 M. J. Frisch, J. A. Pople and J. S. Binkley, J. Chem. Phys., 1984, 80, 3265.
31 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery
Jr., R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam,
A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi,
V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli,
C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala,
Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck,
K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz,
B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi,
R. Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham,
C. Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe,
P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, J. L. Andres,
C. Gonzalez, M. Head-Gordon, E. S. Replogle, and J. A. Pople, in
GAUSSIAN 98, Rev. A7, Pittsburgh, PA, 1998.
23 M. Asami, Chem. Lett., 1984, 829.
24 D. Pettersen, M. Amedjkouh and P. Ahlberg, Tetrahedron, 2002, 58,
4669.
32 W. J. Hehre, B. J. Deppmeier, A. J. Driessen, J. A. Johnson, P. E.
Klunzinger, L. Lou, J. Yu, J. Baker, J. E. Carpenter, R. W. Dixon,
S. S. Fielder, H. C. Johnson, S. D. Kahn, J. M. Leonard and
W. J. Pietro, in Spartan v. 5.0.1, Irvine, CA, 1997.
33 W. Huang, Personal communication, 980513.
34 A. D. Becke, Phys. Rev. A: Gen. Phys., 1988, 38, 3098.
35 J. P. Perdew, Phys. Rev. B: Condens. Matter, 1986, 33, 8822.
25 E. Anders, R. Koch and P. Freunscht, J. Comput. Chem., 1993, 14,
1301.
26 J. J. P. Stewart, J. Comput. Chem., 1989, 10, 209.
27 A. D. Becke, J. Chem. Phys., 1993, 98, 5648.
28 C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter,
1988, 37, 785.
J. Chem. Soc., Perkin Trans. 2, 2002, 1397–1405
1405