A study of BF3-promoted ortho lithiation of anilines and DFT calculations on the role of fluorine-lithium interactions

Article abstract of DOI:10.1002/anie.200705967

(Figure Presented) Director's cut: The poor directing ability of the dimethylamino group in directed ortho metalation is altered by complexation to BF3, which makes it more strongly activating than methoxy and chloro groups. DFTcalculations of simplified lithiated intermediates reveal a small role of the inductive effect, but a distinctive F...Li interaction that leads to a F-Li distance which is remarkably similar to the C-Li distance.

Full text of DOI:10.1002/anie.200705967

Angewandte
Chemie
DOI: 10.1002/anie.200705967
ꢀ
C H Activation
A Study of BF3-Promoted ortho Lithiation of Anilines and DFT
Calculations on the Role of Fluorine–Lithium Interactions**
Satinder V. Kessar,* Paramjit Singh,* Kamal N. Singh,* Prasad V. Bharatam,* Arvind K. Sharma,
Sneh Lata, and Amarjit Kaur
Synthetic and mechanistic aspects of heteroatom directed
ortho metalation (DoM) of aryl substrates (1!2!3!4,
Scheme 1) have fascinated organic chemists for more than
stereoselectivities. However, not much is known about its
mode of action, especially with respect to the role of
CIPE.[3a–c,4] Since CIPE is considered to be particularly
important in DoM reactions,[2,5] the effect of a Lewis acid
on the metalation of anilines could offer some insight into this
mechanism, which may also provide synthetic advantages.
Apriori, coordination of the nitrogen atom lone pair with a
Lewis acid should increase its inductive effect but may
compromise its CIPE.
In this work Lewis acid promoted lithiations were carried
out by adding one equivalent of BF3·Et2O to a solution of
N,N-dimethylanilines in THF prior to the addition of sBuLi
(2 equiv) at ꢀ788C. The electrophile was introduced after
1 hour of stirring at ꢀ788C. Under these conditions, ortho-
substituted products were obtained in 40–50% yield
(Table 1).[6a] The conditions were chosen so that product
formation was not observed if BF3 was omitted (Table 1,
entries 1 and 8). Benzophenone was used as the electrophile
Scheme 1. a) Directed ortho metallation of aryl substrates. b) Lithiation
ꢀ
of weakly acidic a-C H centers. A=heteroatom; X=F, H.
six decades.[1] Ahierarchy of DoM groups has been exper-
imentally established and rationalized in terms of the
inductive effect of the heteroatom and its ability to coordinate
with the metal, which is usually the Li atom of the base that is
used for deprotonation (complex induced proximity effect
(CIPE)).[2] Since a nitrogen atom occupies a very low position
in this established hierarchy, it was of interest to see if
ortho metalation of anilines could be facilitated by Lewis acid
complexation. This relatively recent methodology has been
Table 1: Lithiation of Lewis acid complexed N,N-dimethylanilines with
sBuLi.[a]
Entry Substrate Lewis acid E+
Product
Yield [%][b]
1
2
9a
9a
None
BF3
Ph2CO
Ph2CO
None
13a,
–
41
E=C(OH) Ph2
3
4
5
6
9a
9a
9a
9a
BF3
BF3
BF3
BF3
PhCHO 13a,
E=CH(OH)Ph
CH3OD 13a,
E=D
51
ꢀ
shown to be useful for lithiation of weakly acidic C -H
centers a to the heteroatoms of many aliphatic tertiary
amines and phosphines (5!6!7!8).[3] Lewis acid activation
of alkyl lithium reactions with a variety of other heteroatom-
containing substrates is also well documented;[4] for example,
the 1,2-additions to imines, oximes, carbonyl compounds, and
the cleavage of acetals, epoxides, and unstrained cyclic ethers.
The effect of the Lewis acid is often evidenced by the
improved reaction rates or changes in the chemo-, regio-, or
40[c]
60[d]
59[e]
Ph2CO
13a,
E=C(OH)Ph2
13a,
Ph2CO
E=C(OH)Ph2
None
None
7
8
9
9a
9b
9b
BH3
None
BF3
Ph2CO
Ph2CO
Ph2CO
–
–
13b,
40[f,g]
E=C(OH)Ph2
14c,
E=C(OH) Ph2
13c,
E=C(OH) Ph2
13a,
E=C(OH) Ph2
[*] Prof. S. V. Kessar, Prof. P. Singh, Dr. K. N. Singh, A. K. Sharma,
S. Lata, Dr. A. Kaur
10
11
12
9c
9c
None
BF3
Ph2CO
Ph2CO
Ph2CO
15
30[h]
25[i]
Department of Chemistry
Panjab University
Chandigarh, 160014 (India)
Fax: (+91)172-254-5074
E-mail: svkessar@pu.ac.in
9a +
anisole
BF3
[a] Conditions: 2 equiv of sBuLi in THF at ꢀ788C. [b] Yields are for pure
Prof. P. V. Bharatam
products isolated after chromatography/recrystallization. [c] Yield esti-
Department of Medicinal Chemistry
NIPER, Mohali, 160062 (India)
Fax: (+91) 172-221-4692
1
mated from H NMR spectrum of the crude product mixture. [d] With
tBuLi as the base. [e] With Schlosser’s base. [f] 24% of 14b was also
formed. [g] No product formation withcorresponding o,m-anisidines.
[h] 3% of 14c was also formed. [i] Only one equivalent of base was used,
no anisole substitution product detected. a: R=H; b: R=OMe; c: R=
Cl.
[**] DST, INSA, and CSIR New Delhi, India supported this work.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2008, 47, 4703 –4706
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4703

Communications
because of the ease in NMR characterization of the ortho-
substitution products.
computations of the structures and the energies of the entities
involved in the lithiation were carried out at the B3LYP level
by using aniline and methyl lithium as model reactants
(Scheme 3).[7a]
The yield could be improved to 60% by using tBuLi or
Schlosserꢀs base (sBuLi + tBuOꢀK+, Table 1, entries 5 and 6).
Reactions using BH3 or BCl3 as the Lewis acid, or nBuLi,
MeLi, or PhLi as the base were unsuccessful. In an effort to
improve the yield, the time for the lithiation with sBuLi was
increased to 6 hours at ꢀ788, but product formation was not
observed.[6a,b] Product formation was not detected at ꢀ788C
when lithium 2,2,6,6-tetramethylpiperidide (LTMP) was used
as the base, but at 08C fragmentation to benzyne was
observed (11a!15 + Me2NBF3Li) which was trapped with
anthracene or PhSLi (Scheme 2).
Scheme 3. General structures used for computational studies: lithia-
tion of aniline (top), lithiation of aniline complexed to a boron atom
(middle), and regioisomers of lithiated ring (bottom). For structures:
a: R=H; b: R=OMe; c: R=Cl.
The computed energy values involved in the studied
transformations are presented in Table 2. It can be seen that
the heat of association of MeLi with aniline is about the same
as that with BX3-complexed aniline (Table 2, entries 1–3).
The BX3 coordination renders both DH and Eact more
favorable for lithiation, and to a greater extent in the case
of BF3 (Table 2, entries 4–6; 9–11). Another interesting
feature of the present DFT results is that in the cases where
BF3 is complexed with p-anisidine and p-chloroaniline, the
intermediates with the Li atom ortho to the amino group (25b
and c) are more stable than those with the Li atom ortho to
the methoxy or chloro groups (24b and c, Table 2, entries 12,
13); this result is in qualitative agreement with the exper-
imentally observed regioselectivity (Table 1, entries 9 and 11).
Scheme 2. General reaction scheme: generation of the lithiated aryl
ring (top), reaction of each of the lithiated regioisomers with an
electrophile (middle), and trapping of the in situ generated benzyne
species (bottom). E=electrophile. For structures: a) R=H,
b) R=OMe, c) R=Cl.
Lithiation/substitution of p-N,N-dimethylaminoanisole
(9b) is reported to occur exclusively at the position ortho to
the methoxy substituent with sBuLi in ether at 358C.[5] At
ꢀ788C in THF reaction was not observed with 9b (Table 1,
entry 8), however, upon the addition of one equivalent of
BF3·Et2O prior to treatment with the base the metalation was
accelerated and occurred to a greater extent at the position
ortho to the amino group (Table 1, entry 9). The results with
p-chloro-N,N-dimethylaniline were even more striking. In the
absence of BF3 some substitution at the ortho position to the
chloro group was observed, but with prior BF3 complexation
substitution was almost exclusively at the position ortho to
dimethylamino group (Table 1, entries 10 and 11). In an
intermolecular competition experiment it was also found that
in a 1:1 mixture of anisole and N,N-dimethylaniline, lithiation
can be completely directed to the amine by prior addition of
one equivalent of BF3·Et2O (Table 1, entry 12).
Table 2: Computed heats of reaction and activation energies.
Entry Reaction
DH/Eact [kcal
molꢀ1
]
1
2
3
4
5
6
7
8
16 + 17!18
DH=ꢀ16.56
DH=ꢀ16.12
DH=ꢀ16.03
DH=ꢀ15.94
DH=ꢀ31.94
DH=ꢀ30.93
21 + 17!22 (X=F)
21 + 17!22 (X=H)
16 + 17!19 + 20
21 + 17!23 + 20 (X=F)
21 +17!23 + 20 (X=H)
21 + 17·2Me2O!23·2Me2O + 20 (X=F) DH=ꢀ29.22
21 + 17·2Me2O!23·2Me2O + 20 (X=H) DH=ꢀ24.80
9
18!19 + 20
Eact =22.60
Eact =11.20
Eact =14.20
10
11
12
13
22!23 + 20 (X=F)
22!23 + 20 (X=H)
24!25 (R=OMe)
24!25 (R=Cl)
DH=ꢀ10.67
DH=ꢀ13.01
Absolute energy values calculated at B3LYP/6-31+G* level + ZPE
To delineate the respective roles of the inductive and the
CIPE effects in Lewis acid promoted metalations, DFT
scaled by 0.9806.
4704 www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 4703 –4706

Angewandte
Chemie
Calculations of the natural bond orbital (NBO)-based
partial atomic charges,[8] show a negative charge of ꢀ0.54 on
the lithiated carbon atom at the ortho position of aniline, and
there is no significant change in this charge upon coordination
of the nitrogen center with BH3 or BF3 (Figure 1).[7b] Even
charged hydrogen atoms (Li···H1 = 1.92 Š, Li···H2 = 1.92 Š),
ꢀ
which results in high ellipticity of the Li B bond (e = 2.74).
From the NBO and AIM analyses described above, and
the geometry shown in Figure 1, we infer that in case of BH3
the primary electrostatic interaction of Li+ is with the
negatively charged boron atom, which brings two hydrogen
atoms close to the Li center causing a greater spatial require-
ꢀ
ment for this bond. In contrast, with BF3 even a single Li F
bond can, because of its strength, provide effective stabiliza-
tion and this results in a six-membered chelation ring with less
crowding around lithium (cf. X = F and X = H; Figure 1); this
may also be additionally relevant for the solvated moieties.
The dichotomy in BH3 and BF3 bonding to Li can not only
help to explain the superior effectiveness of BF3 in promoting
lithiations, but also explain the gross variations in crystal
structure of LiBH4 (Li···B = 2.47-2.54 Š, Li···H = 1.98, 2.02,
2.15 Š) and LiBF4 (F-Li = 1.85 Š).[10–12]
In conclusion we have shown that, in principle, BF3
complexation methodology can be extended to directed
ortho metalations and thereby improve the poor ability of
the dimethyl amino group in DoMꢀs to exceed the directing
ability of even chloro and methoxy groups. The DFT
computations are in line with the experimental findings on
the comparative reactivity and regioselectivity of the reaction,
and clearly indicate that chelation of lithium by BX3
coordinated to a nitrogen center (CIPE) has an important
role; an aspect largely ignored in earlier mechanistic consid-
erations on BX3 acceleration of alkyl lithium reactions.[3,4]
Interestingly, in the lithiated intermediates complexed to
BH3, tridentate chelation of Li+ is indicated, whereas a two-
ꢀ
center F Li bond, with characteristics remarkably similar to
Figure 1. 3D-structures of the lithiated intermediates (23) and the
transition states (22TS) that lead to their formation. The important
partial atomic charges are also shown. a) 23 (X=H): the Li atom is
bonded to the B atom with a bond length of 2.21 ꢀ, b) 23 (X=F): the
Li atom is bonded to the F atom with a bond length of 1.80-ꢀ,
c) transition-state 22 (X=H): the Li–B bond length is 2.27 ꢀ, and
d) transition-state 22 (X=F): the Li–F bond length is 1.82 ꢀ.
ꢀ
those of the C Li bond, is seen in the corresponding BF3
complexes. In view of the recent demonstration of the
acceleration of metathesis by chelation of ruthenium to an
aryl fluorine,[13] the effect of BF3 on transition-metal-medi-
ated reactions of heteroatom-containing substrates also needs
to be explored, and if successful it can have an impact on the
scope of this chemistry.[14]
within the known limitations of the method,[8] this result is
surprising, as well as contrary to apriori assumptions;[3] this
suggests a more important role for dipole interactions and
lithium chelation. In contrast, there is clear indication of a
strong interaction between the fluorine and lithium atoms
based on the computations of lithiated intermediate 23 (X =
F). Although one may tend to consider the lithium–carbon
bond as the primary bond, there is, in terms of an atoms in
molecule (AIM) analysis,[9] a six-membered ring critical point
Received: December 28, 2007
Revised: March 16, 2008
Published online: May 16, 2008
Keywords: anilines · boron · density functional calculations ·
.
lithiation · metalation
[1] a) H. Gilman, W. Langham, A. L. Jacoby, J. Am. Chem. Soc.
1939, 61, 106; b) G. Wittig, G. Fuhrmann, Chem. Ber. 1940, 73,
1197.
ꢀ
ꢀ
in 23 and a remarkable similarity between the Li F and Li C
bonds in terms of charge (C = ꢀ0.56, F = ꢀ0.62, Li = 0.90),
ꢀ
ꢀ
ꢀ
bond length (C Li = 2.01 Š, F Li = 1.80 Š), 1 (C Li = 0.039,
[2] a) H. W. Gschwend, H. R. Rodriguez, Org. React. 1979, 26, 1;
b) V. Snieckus, Chem. Rev. 1990, 90, 879; c) M. C. Whisler, S.
MackNeil, V. Snieckus, P. Beak, Angew. Chem. 2004, 116, 2256;
Angew. Chem. Int. Ed. 2004, 43, 2206; d) for early computations
on ortho lithiation see: N. J. R. Von Eikema Hommes, P. v. R.
Schleyer, Tetrahedron 1994, 50, 5903; e) A.-M. Sapre, P. v. R.
Schleyer, Lithium Chemistry; A theoretical and experimental
overview, Wiley, New York, 1995.
2
ꢀ
ꢀ
ꢀ
ꢀ
F Li = 0.036), 5 1 (C Li = 0.19, F Li = 0.29), and e (C Li =
ꢀ
0.02, F Li = 0.02). AIM computations on the solvate,
23·2Me2O, give essentially similar results, including a six-
membered ring critical point with the expected decrease in
electron density of the Li bonds to F and C (see the
Supporting Information). In comparison, in the lithiated
BH3 complex of 23 (X = H) there is a five-membered ring
critical point with Li···B distance of 2.21 Š (charge Li = 0.89,
B = ꢀ0.23). Lithium is also in contact with two weakly
[3] a) S. V. Kessar, P. Singh, Chem. Rev. 1997, 97, 721; b) S. V.
Kessar, P. Singh, K. N. Singh, A. Kaur, P. Venugopalan, P. V.
Bharatam, A. K. Sharma, J. Am. Chem. Soc. 2007, 129, 4506;
Angew. Chem. Int. Ed. 2008, 47, 4703 –4706
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 4705

Communications
c) E. Vedejs, J. T. Kendall, J. Am. Chem. Soc. 1997, 119, 6941;
[8] A. E. Reed, L. A. Curtiss, F. Weinhold, Chem. Rev. 1988, 88, 899.
[9] R. F. W. Bader, Chem. Rev. 1991, 91, 893.
[10] J.-Ph. Soulie, G. Renaudin, R. Cerny, K. Yvon, J. Alloys Compd.
2002, 346, 200 – 205.
d) A. Ariffin, A. J. Blake, M. R. Ebden, W.-S. Li, N. S. Simpkins,
D. N. A. Fox, J. Chem. Soc. Perkin Trans. 1 1999, 2439; e) D.
Kuck, Angew. Chem. 2000, 112, 129; Angew. Chem. Int. Ed. 2000,
39, 125; f) V. Ferey, L. Toupet, T. L. Gall, C. Mioskowski, Angew.
Chem. 1996, 108, 475; Angew. Chem. Int. Ed. Engl. 1996, 35, 430;
g) J. M. Concellón, J. R. Saurez, S. G. Granda, M. R. Diaza,
Angew. Chem. 2004, 116, 4433; Angew. Chem. Int. Ed. 2004, 43,
4333; h) X. Sun, K. Manabe, W. W.-L. Lam, N. Shiraishi, J.
Kobayashi, M. Shiro, H. Utsumi, S. Kobayashi, Chem. Eur. J.
2005, 11, 361, and references therein.
[11] Monodentate BF4ꢀ bonding with Li+ leads to high formula unit
volume of the LiBF4 crystal which has been linked to higher
standard entropy and decomposition temperature, an advantage
for the use in lithium batteries: a) K. Matsumoto, R. Hagiwara,
Z. Mazej, E. Goreshnik, B. Zemva, J. Phys. Chem. B 2006, 110,
2138 – 2141; b) R. Dedryvere, S. Leroy, H. Martinez, F. Blan-
chard, D. Lemordant, D. Gonbeau, J. Phys. Chem. B 2006, 110,
12986.
[4] K. B. Aubrecht, M. D. Winemiller, D. B. Collum, J. Am. Chem.
Soc. 2000, 122, 11084.
[12] The DFT computations on monomeric LiBF4 favor C2v
structure with two fluorine atoms near the lithium center;
a) X. Xuan, H. Zhang, J. Wang, H. Wang, J. Phys. Chem. A 2004,
108, 7513; b) Q. Ge, J. Phys. Chem. A 2004, 108, 8682.
[13] a) T. Ritter, M. W. Day, R. H. Grubbs, J. Am. Chem. Soc. 2006,
128, 11768; b) For the role of CF3 in lithiation, see: K. J. Singh,
D. B. Collum, J. Am. Chem. Soc. 2006, 128, 13753.
[5] D. W. Slocum, C. A. Jennings, J. Org. Chem. 1976, 41, 3653.
[6] a) See the supporting information; b) This could be because of
the instability of 11a as the presence of BF3 on the nitrogen
center can make it a suitable leaving group (R. W. Hofmann,
Dehydrobenzene and Cycloalkynes, Acadmic Press, New York,
1967, pp. 68).
[7] a) For seminal computational work on Lewis acid promoted
deprotonation chemistry ignoring the lithium counter cation,
see: J. Ren, D. B. Workman, R. R. Squires, J. Am. Chem. Soc.
1998, 120, 10511; b) NBO analysis of a-lithiated aliphatic amines
also reveals no decrease in the carbanionic charge upon
coordination with BX3 (see the Supporting Information).
[14] J. P. Collman, L. S. Hegedus, J. R. Norton, R. G. Finke, Principles
and Applications of Organotransition Metal Chemistry, Oxford
University Press, 1987, p. 727.
[15] We thank the referees for the suggestions on widening the
context (reference [4] and additional computations on solvates).
4706 www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 4703 –4706