Structural and rate studies of the formation of substituted benzynes

Article abstract of DOI:10.1021/ja0754655

The key elimination step for the formation of 3-substituted and 3,6-disubstituted benzynes from 2-haloaryllithiums displays a pronounced solvent-dependent regioselectivity. All 2-haloaryllithiums with electron withdrawing groups in the 6 position are show

Full text of DOI:10.1021/ja0754655

Published on Web 02/23/2008
Structural and Rate Studies of the Formation
of Substituted Benzynes
Jason C. Riggs,† Antonio Ramirez,† Matthew E. Cremeens,† Crystal G. Bashore,‡
John Candler,‡ Michael C. Wirtz,‡ Jotham W. Coe,*,‡ and David B. Collum*,†
Department of Chemistry and Chemical Biology, Baker Laboratory, Cornell UniVersity,
Ithaca, New York 14853-1301, and Pfizer Global Research and DeVelopment, Groton
Laboratories, Pfizer, Inc., Groton, Connecticut 06340
Received July 22, 2007; E-mail: dbc6@cornell.edu
Abstract: The key elimination step for the formation of 3-substituted and 3,6-disubstituted benzynes from
2-haloaryllithiums displays a pronounced solvent-dependent regioselectivity. All 2-haloaryllithiums with
6
electron withdrawing groups in the 6 position are shown by Li and 13C NMR spectroscopic studies to be
monomers in THF. DFT computational studies implicate trisolvates. Rate studies reveal that LiF eliminates
via monomer-based pathways requiring THF dissociation whereas LiCl eliminates via nondissociative
pathways. Elimination to form 3-chloro- and 3-fluorobenzyne from 2-chloro-6-fluorophenyllithium displays
a pronounced solvent-dependent regioselectivity that is traced to competing solvent-dissociative and
nondissociative dissociative pathways for the elimination of LiCl and LiF, respectively.
nicotinic receptor partial agonist4 currently marketed as an aid
to smoking cessation treatment under the names Chantix and
Introduction
Benzyne has historically been a molecule of largely theoretical
interest.1 One cannot help notice, however, that the use of
benzynes as reactive intermediates in synthesis has picked up
markedly.2,3 We recently reported investigations of bicyclic
derivatives of general structure 2 related to varenicline (1), a
Champix (eq 1).5 The initial syntheses of 2 relied on the
generation and subsequent trapping of benzyne.6 During the
course of these studies, we discovered a pronounced solvent-
dependent regioselectivity illustrated in eq 2.6a,7 Whereas
formation of benzyne from 1-chloro-3-fluorobenzene (3) in
hydrocarbons or hydrocarbons containing low concentrations
of ethereal solvents affords predominantly the product arising
from elimination of LiF, the analogous reaction in neat THF
solution affords a 50:1 preference for the elimination of LiCl.
† Cornell University.
‡ Pfizer, Inc.
(1) (a) Ozkan, I.; Kinal, A. J. Org. Chem. 2004, 69, 5390. (b) De Proft, F.;
Schleyer, P. v. R.; van Lenthe, J. H.; Stahl, F.; Geerlings, P. Chem.-Eur.
J. 2002, 8, 3402. (c) Johnson, W. T. G.; Cramer, C. J. J. Phys. Org. Chem.
2001, 14, 597. (d) Jiao, H.; Schleyer, P. v. R.; Beno, B. R.; Houk, K. N.;
Warmuth, R. Angew. Chem., Int. Ed. Engl. 1997, 36, 2761. (e) Radzisze-
wski, J. G.; Hess, B. A. Jr.; Zahradnik, R. J. Am. Chem. Soc. 1992, 114,
52, and references cited therein. (f) Wenthold, P. G.; Paulino, J. A.; Squires,
R. R. J. Am. Chem. Soc. 1991, 113, 7414. (g) Gronert, S.; DePuy, C. H. J.
Am. Chem. Soc. 1989, 111, 9253. (h) Hoffmann, R. W. Dehydrobenzene
and Cycloalkynes; Academic Press: New York, 1967.
We report herein structural studies of aryllithiums 6-10 and
(2) Recent reviews on benzyne: (a) Dyke, A. M.; Hester, A. J.; Lloyd-Jones,
G. C. Synthesis 2006, 4093. (b) Pellissier, H.; Santelli, M. Tetrahedron
2003, 59, 701. (c) Wenk, H. H.; Winkler, M.; Sander, W. Angew. Chem.,
Int. Ed. 2003, 42, 502. (d) Na´jera, C.; Sansano, J. M.; Yus, M. Tetrahedron
2003, 59, 9255. (e) Leroux, F.; Schlosser, M. Angew. Chem., Int. Ed. 2002,
41, 4272.
(3) Recent examples of generation of benzyne from haloaryllithiums: (i) via
LiF elimination: (a) Sanz, R.; Fernandez, Y.; Castroviejo, M. P.; Perez,
A.; Fan˜anas, F. J. J. Org. Chem. 2006, 71, 6291. (b) Heiss, C.; Leroux, F.;
Schlosser, M. Eur. J. Org. Chem. 2005, 24, 5242. (c) Kudzma, L. V.
Synthesis 2003, 1661. (d) Dyer, P. W.; Fawcett, J.; Hanton, M. J.; Kemmitt,
R. D. W.; Padda, R.; Singh, N. J. Chem. Soc., Dalton Trans. 2003, 104.
(e) Pawlas, J.; Begtrup, M. Org. Lett. 2002, 4, 2687. (f) Caster, K. C.;
Keck, C. G.; Walls, R. D. J. Org. Chem. 2001, 66, 2932. (ii) Via LiCl
elimination: (g) Heiss, C.; Cottet, F.; Schlosser, M. Eur. J. Org. Chem.
2005, 24, 5236. (h) Shoji, Y.; Hari, Y.; Aoyama, T. Tetrahedron Lett. 2004,
45, 1769. (i) Bennet, M. A.; Kopp, M. R.; Wenger, E.; Willis, A. C. J.
Organomet. Chem. 2003, 667, 8. (j) Gohier, F.; Castanet, A.-S.; Mortier,
J. Org. Lett. 2003, 5, 1919. (k) Faigl, F.; Marzi, E.; Schlosser, M. Chem.-
Eur. J. 2000, 6, 771. (l) Wang, A.; Zhang, H.; Biehl, E. R. Heterocycles
2000, 52, 1133. (iii) From 3-haloanisoles: (m) Biland-Thommen, A. S.;
Raju, G. S.; Blagg, J.; White, A. J. P.; Barrett, A. G. M. Tetrahedron Lett.
2004, 45, 3181. (n) Bauta, W. E.; Lovett, D. P.; Cantrell, Jr. W. R.; Burke,
B. D. J. Org. Chem. 2003, 68, 5967. (o) Becht, J.-M.; Gissot, A.; Wagner,
A.; Mioskowski, C. Chem.-Eur. J. 2003, 9, 3209. (p) Kita, Y.; Higuchi,
K.; Yoshida, Y.; Iio, K.; Kitagaki, S.; Ueda, K.; Akai, S.; Fujioka, H. J.
Am. Chem. Soc. 2001, 123, 3214.
(4) Rollema, H.; Coe, J. W.; Chambers, L. K.; Hurst, R. S.; Stahl, S. M.;
Williams, K. E. Trends Pharmacol. Sci. 2007, 28, 316.
(5) (a) Coe, J. W.; Brooks, P. R. P. U.S. Cont.-in-part of Appl. No. PCT/
IB98/01813, US 6605610, 2003. (b) Coe, J. W.; et al. J. Med. Chem. 2005,
48, 3474.
9
3406
J. AM. CHEM. SOC. 2008, 130, 3406-3412
10.1021/ja0754655 CCC: $40.75 © 2008 American Chemical Society

Formation of Substituted Benzynes
A R T I C L E S
Table 1. 6Li and 13C Spectral Dataa
compd
6Li,
δ
(m, 3JLiF
)
13C, (m, 2JFC 1JLiC
δ , )
6b
0.81 (d, 1.2)
0.49 (s)
1.20 (t, 1.4)
0.72 (s)
164.5 (dt, 130.3, 10.0)c
183.4 (t, -, 11.3)
7b
8b
145.5 (tt, 127.1, 10.1)
175.5 (t, -, 10.6)
9d
10d
1.75 (d, 0.7)
154.6 (dt, 132.1, 9.6)e
a 6Li NMR spectra were recorded on samples containing 0.20 M
aryllithium, whereas 13C NMR spectra were recorded on samples containing
0.30 M aryllithium. Coupling constants were measured after resolution
enhancement. Multiplicities are denoted as follows: s ) singlet, d ) doublet,
t ) triplet. The chemical shifts are reported relative to 0.30 M 6LiCl/MeOH
at -90 °C (0.0 ppm) and neat THF (25.6 and 67.6 ppm). All J values are
reported in Hz. b Recorded at -100 °C. c Sample containing 0.40 M
aryllithium. d Recorded at -110 °C. e Obscured by another resonance.
rate studies of their conversion to the corresponding benzyne
intermediates. We show that the THF concentration-dependent
regioselectivities illustrated in eq 2 derive from fundamental
differences between the elimination of LiCl and LiF. Studies
of the structure and reactivity of 6 were communicated previ-
ously.7
Figure 1. (A) 13C NMR spectrum of 0.40 M [6Li]3-chloro-6-fluorophe-
nyllithium (6) in neat THF at -100 °C, and (B) 6Li NMR spectrum of
0.20 M [6Li]-(6) in THF (10.3 M) with toluene cosolvent at -100 °C.
Table 2. Approximate Relative Rate Constants (krel) for Benzyne
Formation from Aryllithiums 6-10 in THF Corrected to -35 °C
ArLi
1.0 M THF
9.0 M THF
6a
25
35
7
115
140
15
-
1
-
10
Results
7a
8a
Structures of Haloaryllithiums. Haloaryllithiums [6Li]6-
10 used for NMR spectroscopic analysis were prepared by
treatment of their arene precursors in THF at -196 °C with
recrystallized [6Li]n-BuLi in pentane.8 13C NMR spectra
9b
10c
a Measured at -25 °C. Measured at -50 °C. Measured at -35 °C.
b
c
6
recorded at -100 °C contain a single resonance displaying -
inductive stabilization of the charge and modest (less conse-
quential) steric inhibition of aggregation.13
Li-13C coupling of the 6Li-bearing carbon consistent with
aryllithium monomers. The C-1 resonances of [6Li]6, [6Li]8,
Solvent-Dependent Relative Rate Constants. Relative rate
constants for the elimination of lithium halide from aryllithiums
6-10 were measured using protocols described below and are
listed in Table 2. The values should be viewed qualitatively in
that adjustments to -35 °C were calculated assuming a 2-fold
change in rate with each 10 °C change.
Kinetics of Benzyne Formation. Aryllithiums 6-10 used
for rate studies were generated in situ by treatment of the
corresponding haloarenes at e-25 °C with freshly recrystallized8
commercial n-BuLi. Ortholithiations to form 6-8 were es-
sentially instantaneous. The halogens at the 4-position of
aryllithiums 9 and 10 were required to accelerate the ortholithia-
tion relative to the benzyne formation so as to avoid n-Bu-
and [6Li]10 also display two-bond 13C-19F coupling (2JF-C
≈
130-140 Hz).9,10 Thus, the 13C NMR spectra of [6Li]6 and [6-
Li]10 display a doublet of triplets, and the spectrum of [6Li]8
displays a triplet of triplets (Table 1, Figure 1A). The 6Li NMR
spectra reveal a doublet for [6Li]6 and [6Li]10 and a triplet for
[6Li]8. All show nearly identical Li-19F coupling constants
6
(3JLi-F ≈ 1 Hz, Figure 1B). We did not ascertain the solvation
numbers of the aryllithiums spectroscopically, but DFT calcula-
tions (below) and analogy with crystal structures11 provide
support for trisolvated monomers.12 We presume that the
exclusive formation of monomer stems from a combination of
(6) (i) Synthesis via formation of 3-halobenzyne: (a) Coe, J. W.; Wirtz, M.
C.; Bashore, C. G.; Candler, J. Org. Lett. 2004, 6, 1589. (ii) For alternative
routes, see: (b) Brooks, P. R.; Caron, S.; Coe, J. W.; Ng, K. K.; Singer, R.
A.; Vazquez, E.; Vetelino, M. G.; Watson, Jr. H. H.; Whritenour, D. C.;
Wirtz, M. C. Synthesis 2004, 1755. (c) Singer, R. A.; McKinley, J. D.;
Barbe, G.; Farlow, R. A. Org. Lett. 2004, 6, 2357.
(11) (a) X-ray of monomeric 2,3,4,5-tetrafluorophenyllithium: Kottke, T.; Sung,
K.; Lagow, R. J. Angew. Chem., Int. Ed. Engl. 1995, 34, 1517. (b) For a
related monomer, see: Bosold, F.; Zulauf, P.; Marsch, M.; Harms, K.;
Lohrenz, J.; Boche, G. Angew. Chem., Int. Ed. Engl. 1991, 30, 1455. (c)
2,3,5,6-Tetrafluorophenyllithium exists as a monomer in THF: Stratakis,
M.; Wang, P. G.; Streitwieser, A. J. Org. Chem. 1996, 61, 3145.
(12) (a) Schlosser, M. In Organometallics in Synthesis: A Manual, 2nd ed.;
Schlosser, M., Ed.; John Wiley & Sons: Chichester, 2002; Chapter 1. (b)
Reich, H. J.; Green, D. P.; Medina, M. A.; Goldenberg, W. S.; Gudmunds-
son, B. O¨ .; Dykstra, R. R.; Phillips, N. H. J. Am. Chem. Soc. 1998, 120,
7201. (c) Seebach, D. Angew. Chem., Int. Ed. Engl. 1988, 22, 1624.
(13) (a) Reich, H. J.; Goldenberg, W. S.; Sanders, A. W.; Jantzi, K. L.;
Tzschucke, C. C. J. Am. Chem. Soc. 2003, 125, 3509. (b) Reich, H. J.;
Goldenberg, W. S.; Gudmundsson, B. O¨ .; Sanders, A. W.; Kulicke, K. J.;
Simon, K.; Guzei, I. A. J. Am. Chem. Soc. 2001, 123, 8067. (c) Reich, H.
J.; Sikorski, W. H.; Gudmundsson, B. O¨ .; Dykstra, R. R. J. Am. Chem.
Soc. 1998, 120, 4035.
(7) Ramirez, A.; Candler, J.; Bashore, C. G.; Wirtz, M. C.; Coe, J. W.; Collum,
D. B. J. Am. Chem. Soc. 2004, 126, 14700.
(8) (a) Hoffmann, D.; Collum, D. B. J. Am. Chem. Soc. 1998, 120, 5810. (b)
Kottke, T.; Stalke, D. Angew. Chem., Int. Ed. Engl. 1993, 32, 580.
(9) Spin quantum numbers: 6Li has I ) 1; 19F has I ) 1/2.
2
(10) (i) Similar JF-C values have been observed for related 2-fluorophenyl-
lithiums: (a) Singh, K. J.; Collum, D. B. J. Am. Chem. Soc. 2006, 128,
13753. (b) Menzel, K.; Fisher, E. L.; DiMichele, L.; Frantz, D. E.; Nelson,
2
T. D.; Kress, M. H. J. Org. Chem. 2006, 71, 2188. (ii) JC-F values have
been correlated with π-bond orders and total electronic charge at the 13C
atom: (c) Doddrell, D.; Jordan, D.; Riggs, N. V. J. Chem. Soc., Chem.
Commun. 1972, 1158. (d) Doddrell, D.; Barfield, M.; Adcock, W.;
Aurangzeb, M.; Jordan, D. J. Chem. Soc., Perkin Trans. 2 1976, 402.
9
J. AM. CHEM. SOC. VOL. 130, NO. 11, 2008 3407

A R T I C L E S
Riggs et al.
Table 3. Summary of Rate Studies for the Elimination of Lithium
Halide from Aryllithiums 6-10
ArLi
temp (
°
C)
ArLia order
THF order
6
7
8
9
10
-25
-25
-25
-50
-35
1.05 ( 0.04b
0.98 ( 0.04b
1.06 ( 0.07b
1.1 ( 0.1c
-1.12 ( 0.09, 0
0
-0.98 ( 0.07
0
-1.10 ( 0.06
1.00 ( 0.09c
a Values derived from the best fit of initial rates.15 b [THF] ) 6.8 M.
c [THF] ) 5.0 M.
Figure 3. Plot of kobsd vs [THF] in toluene cosolvent for the elimination
of LiCl from 7 at -25 °C. The curve depicts the result of an unweighted
least-squares fit to kobsd ) k[THF] + k′ (k ) (-4 ( 9) × 10-6, k′ ) (1.5
( 0.1) × 10-3).
Figure 2. Plot of the initial slope ∆[C6H4Cl2]/∆t vs initial [7] in THF (0.3
M) and toluene cosolvent for the elimination of LiCl from 7 at -25 °C.
The curve depicts the result of an unweighted least-squares fit to ∆[C6H4-
Cl2]/∆t ) k[7]n (k ) (1.6 ( 0.1) × 10-3, n ) 0.97 ( 0.04).
containing byproducts.6a Aryllithiums 9 and 10 also required a
slight excess of starting arene to ensure the metalation was
complete prior to the onset of benzyne formation.14
The rates of benzyne formation were determined by the
method of initial rates15 using an aqueous quench and subsequent
GC analysis of the haloarenes relative to an n-octane or n-decane
internal standard.16 The reaction temperatures were chosen to
obtain convenient time scales (t1/2 ) 45-300 min in 2.0 M
THF). The resulting benzynes were trapped with excess spiro-
[2.4]hepta-4,6-diene17 as the corresponding cycloadducts 5a-
d. In the absence of diene, the benzynes reacted with the
aryllithiums, affording spurious rate data and complex product
mixtures containing biphenyls. Representative data are plotted
in Figures 2-5. The rate studies are summarized in Table 3.
Additional graphical and tabular data are included in Supporting
Information.
Figure 4. Plot of kobsd vs [THF] in toluene cosolvent for the elimination
of LiF from 8 at -25 °C. The curve depicts the result of an unweighted
least-squares fit to kobsd ) k[THF]n + k′ (k ) 2.4 ( 0.2) × 10-4, n ) -1.0
( 0.1, k′ ) (0.1 ( 2.0) × 10-5).
slopes versus initial aryllithium concentrations are all nearly
linear; Figure 2 is representative. In conjunction with the
assignments of aryllithiums 7-10 as monomers, the rates are
consistent with monomer-based eliminations.19 In the case of
the mixed halogenated derivative 6, contributions of the two
pathways could be separated (as described in detail below) and
shown to follow first order behaviors. In addition, the indepen-
dence of the product ratio (5a:5b) on both the percent conversion
and the initial concentration of ArLi confirm that the competitive
pathways proceed via transition structures of equivalent (mon-
omeric) aggregation states.
Reaction Orders in THF. The dependencies of the decay
of 7-10 on THF concentration (in toluene) show two distinct
behaviors. Plots of kobsd versus THF concentration for the
decomposition of dichloro derivatives 7 and 9 show THF-
Reaction Orders in ArLi. The values of kobsd are independent
of the initial concentrations of aryllithiums 7-10, consistent
with a first-order dependence.18 Alternatively, plots of the initial
concentration-independent rates (rate
[THF]0, Figure 3)
consistent with a mechanism requiring neither association nor
dissociation of THF. In contrast, plots of kobsd versus THF
concentration for the decomposition of difluoro derivatives 8
and 10 follow inverse-first-order THF concentration depend-
encies (Figure 4). Eliminations of 8 and 10 may also have
(14) Rates of benzyne formation from aryllithiums 9 and 10 showed inverse
dependencies at low arene concentrations due to incomplete metalation at
low temperatures.
(15) The initial rates (∆[arene]/∆t) were converted to their corresponding
observed first-order rate constants (kobsd, see Supporting Information).
(16) (a) Remenar, J. F.; Collum, D. B. J. Am. Chem. Soc. 1998, 120, 4081. (b)
Remenar, J. F.; Collum, D. B. J. Am. Chem. Soc. 1997, 119, 5573.
(17) Hart, H.; Lai, C.; Chukuemeka, G.; Shamouilian, S. Tetrahedron 1987,
43, 5203.
(19) A single report indicates that the elimination of LiCl from 2-chloro-3-
dimethylamino-6-phenylsulfonylphenyllithium is a first-order process:
Zieger, H. E.; Wittig, G. J. Org. Chem. 1962, 27, 3270.
(18) Espenson, J. H. Chemical Kinetics and Reaction Mechanisms, 2nd ed.;
McGraw-Hill: New York, 1995.
9
3408 J. AM. CHEM. SOC. VOL. 130, NO. 11, 2008

Formation of Substituted Benzynes
A R T I C L E S
contributions from a zeroth-order dependency evidenced by
small, but nonzero, values of kobsd in the high THF concentration
limit (rate C1/[THF] + C2[THF]0).
A plot of kobsd versus THF concentration for the decomposi-
tion of the mixed halogenated aryllithium 6 is shown in Figure
5. kobsd is a composite rate constant described as kobsd ) kLiF
+
kLiCl, such that kLiF and kLiCl correspond to the rate constants
for the elimination of LiF and LiCl, respectively. The product
ratio, 5a/5b, provides kLiF/kLiCl. Consequently, kLiF and kLiCl can
be monitored independently (Figure 5). It is evident that the
competitive elimination of LiF and LiCl from 6 is a composite
of the zeroth-order and inverse-order pathways observed for 7
and 8, respectively.
Overall Rate Laws. The structural and rate studies combine
to provide the idealized20 rate laws and mechanisms described
by eqs 3-12. The rate data are consistent with generic transition
structures 11-13. The solvation numbers of aryllithiums 6-10
and, by inference from the rate laws, transition structures 11-
13 are based on DFT computations described below.
Elimination of 7 and 9:
Figure 5. (a) Plot of kobsd vs [THF] in toluene cosolvent for the elimination
of LiCl and LiF from 6 (0.20 M) at -25 °C. The curve depicts the result
of an unweighted least-squares fit to kobsd ) k[THF]n + k′ (k ) (8 ( 1) ×
10-5, n ) (-1.1 ( 0.1), k′ ) (1.5 ( 0.1) × 10-4). (b) Plot of kLiCl vs
[THF]. The curve depicts the result of an unweighted least-squares fit to
kLiCl ) k[THF] + k′ (k ) (1 ( 9) × 10-7, k′ ) (1.4 ( 0.1) × 10-4). (c)
Plot of kLiF vs [THF]. The curve depicts the result of an unweighted least-
squares fit to kLiF ) k[THF]n + k′ (k ) (8 ( 1) × 10-5, n ) -1.1 ( 0.1,
k′ ) (1 ( 8) × 10-6).
-d[ArLi]/dt ) kCl[ArLi][THF]0
(3)
(4)
kCl
3 98
[ArLi(THF) ]q
ArLi(THF)
3
(7 or 9)
(11)
Elimination of 8 and 10:
-d[ArLi]/dt ) kFKeq[ArLi][THF]-1 + kF′[ArLi][THF]0 (5)
Vibrational frequencies calculated at the same level character-
ized these stationary points as minima. Me2O was used as a
model for THF. Fully optimized structures were obtained for
aryllithiums 6-10. Structures with idealized C(2)-C(1)-Li
angles were examined in addition to those containing Li-
halogen and Li-OMe interactions detected during the compu-
tational studies of aryllithiums with low solvation numbers.23
Selected geometric characteristics of the calculated structures
are archived in Supporting Information.
Calculated ArLi Structures. Serial solvation of monomeric
aryllithiums 6-10 was investigated as summarized in eq 13
and Table 4. PhLi is included for comparison.24 The relative
free energies as defined by eq 14 are also listed in Table 4.
Keq
ArLi(THF)
(8 or 10)
3 y z ArLi(THF)2 + THF
(6)
(7)
(8)
kF
8
[(ArLi(THF) ]q
ArLi(THF)2
9
2
(12)
kF′
3 98
[(ArLi(THF) ]q
ArLi(THF)
3
(8 or 10)
Elimination of 6:
(13)
-d[ArLi]/dt ) kFKeq[ArLi][THF]-1 + kCl[ArLi][THF]0 (9)
Keq
ArLi(THF)
3 y z ArLi(THF)2 + THF
(10)
(11)
(12)
(6)
kF
98
[ArLi(THF) ]q
ArLi(THF)2
2
∆G4
(12)
PhLi(Me2O)3 + ArH y z ArLi(Me2O)3 + PhH (14)
kCl
3 98
[ArLi(THF) ]q
ArLi(THF)
The optimized aryllithiums 6-10 routinely display Li-C(1)
and Li-O bond distances that increase with increasing solvation
3
(6)
(11)
(23) (a) Li-F contacts have been found with ab initio calculations and associated
to the formation of benzyne: Streitwieser, A. Abu-Hasanyan, F.; Neuhaus,
A.; Brown, F. J. Org. Chem. 1996, 61, 3151. (b) DFT computations indicate
that Li-F interactions are stronger than Li-Cl contacts and control structure
and solvation of carbenoid precursors: Pratt, L. M.; Ramachandran, B.;
Xidos, J. D.; Cramer, C. J.; Truhlar, D. G. J. Org. Chem. 2002, 67, 7607.
(c) For a discussion on anti-MeO-Li interactions, see: Bauer, W.; Schleyer,
P. v. R. J. Am. Chem. Soc. 1989, 111, 7191.
DFT Calculations. All calculations were performed with
Gaussian 0321 using the hybrid density functional theory B3LYP
method22 and 6-31G(d) basis set for geometry optimizations.
(20) We define the idealized rate law as that obtained by rounding the observed
reaction orders to the nearest rational order.
(21) Frisch, M. J.; et al. Gaussian 03, revision B.04; Gaussian, Inc.: Pittsburgh,
PA, 2003.
(22) For related B3LYP computations of aryllithiums, see: (a) Wiberg, K. B.;
Sklenak, S.; Bailey, W. F. J. Org. Chem. 2000, 65, 2014. (b) Kremer, T.;
Junge, M.; Schleyer, P. v. R. Organometallics 1996, 15, 3345. See also
ref 10a.
(24) (i) In THF, PhLi is a mixture of monomer and dimer.12b For a recent DFT
calculation of PhLi(Me2O)3, see: (a) Krasovsky, A.; Straub, B. F.; Knochel,
P. Angew. Chem., Int. Ed. 2006, 45, 159. (ii) For computational discussions
of lithiated benzenes, see: (b) Bachrach, S. M.; Chamberlin, A. C. J. Org.
Chem. 2004, 69, 2111. (c) Kwon, O.; Sevin, F.; McKee, M. L. J. Phys.
Chem. A 2001, 105, 913.
9
J. AM. CHEM. SOC. VOL. 130, NO. 11, 2008 3409

A R T I C L E S
Riggs et al.
Table 4. Calculated Free Energies (∆G, kcal/mol) of Aryllithiums
6-10 Solvated by Me2O (eqs 13 and 14)
ArLia
∆G1
∆G2
∆G3
∆G4
PhLi
6
7
8
9
-14.0
-13.2
-13.1
-14.3
-12.2
-13.0
-4.3
-4.3
-4.9
-2.9
-3.1
-2.9
+3.1
+3.8
+5.3
+2.5
+5.3
+4.5
0.0
-19.8
-16.4
-21.4
-15.3
-17.9
10
a ArLi refers to unsolvated analogues of aryllithiums 6-10 and phenyl-
lithium.
number. Attempts to locate minima corresponding to tetrasol-
vates resulted in desolvation. Trisolvation of aryllithiums 7, 9,
and 10 substituted by ortho Cl or MeO groups is less favored
(∼2 kcal/mol) than trisolvation of PhLi and aryllithiums 6 and
8, which are ortho substituted by sterically undemanding H and
F atoms.25 Although the free energies of the third solvation step
(∆G3) are positive, this is not unusual10a,26 because DFT
computations on microsolvated models tend to overestimate
steric effects.27 The discussion focuses on the trisolvated forms.28
Figure 6. Potential energy (∆E, kcal/mol) as a function of C-Cl bond
lengthening (Å) relative to fully optimized 7-(Me2O)3.
contrast, the lower solvates of chloro analogue 9 prefer the anti-
oriented isomer with Li-OMe contacts.
Dichloroaryllithium 7 shows significant distortion by Li-Cl
interactions (14) but only in the chemically unrealistic unsol-
vated and monosolvated forms; the higher solvates show no
Li-Cl interaction. By contrast, the difluoro derivative 8 shows
significant distortion and possibly affiliated stabilization at-
tributable to a Li-F interaction even in the di- and trisolvated
forms (15). The mixed halogenated derivative 6 is a composite
of 7 and 8 and converges to structure 16 displaying a Li-F
interaction regardless of starting geometry.
Evaluation of the Reaction Pathways. We examined the
reaction coordinates for the decomposition of aryllithiums 6-10
using DFT computations. Exhaustive unconstrained searches
failed to converge.29 In short, weak benzyne-lithium halide
bonding affords a flat potential energy surface with many floppy
structures displaying significant geometrical variations separated
by trivial differences in stability. The alternative protocol of
gradually lengthening the C(2)-halide bond and optimizing all
remaining parameterssthe so-called relaxed-scan approximations
affords potential energy profiles that are generated from
nonstationary points.30 With caution and the hope that it may
be instructive, we offer a representative relaxed scan (Figure
6). Lengthening of the C(2)-halide bond is accompanied by
simultaneous lengthening of the Li-C(1) distance and contrac-
tion of both Li-halide and C(1)-C(2) distance for all solvation
numbers represented emblematically by structures 17 and 18.
Arene and C(1)-C(2)-Li-halogen substructures are located in
almost co-incident planes. In all cases, the structures defining
the limits of benzyne-lithium halide dissociation display residual
C(2)-halide contacts (∼2.5 Å) along with clear Li-C(1)
disconnection (g3.0 Å) suggestive of a E1cb-type mecha-
nism.31,32 In general, the affiliated energies approach 20-35
Haloanisole-derived aryllithiums 9 and 10 afforded optimized
structures with the methoxy oriented either syn or anti relative
to lithium (see below).23c The most stable trisolvates are anti-9
and anti-10, in which the lone pairs on oxygen are directed
toward the lithium, but neither display Li-OMe contacts.
Trisolvated anti-9 does not show a Li-Cl interaction, whereas
the analogous anti-10 trisolvate displays a Li-F contact that is
stabilizing when compared with the trisolvated anti-10 without
Li-heteroatom interactions (∼0.5 kcal/mol).23a At lower solva-
tion numbers, the syn-oriented isomers of the lower solvates of
fluoroanisole 10 show a distinct Li-F contact and are more
stable than the anti isomers (∼2.5 kcal/mol), even though the
anti forms display potentially stabilizing Li-OMe contacts. By
(25) For a discussion of the ortho-substituent effect on reducing solvent access
to halophenylcesiums, see: Streitwieser, A. Abu-Hasanyan, F.; Neuhaus,
A.; Brown, F. J. Org. Chem. 1996, 61, 3151.
(26) Recent example of a small positive ∆G for the third solvation of an
organolithium monomer calculated using DFT: Pratt, L. M.; Mu, R.; Jones,
D. R. J. Org. Chem. 2005, 70, 101.
(27) (i) Examples of a large positive ∆G for the third solvation of an
organolithium monomer calculated using DFT: (a) Zuend, S. J.; Ramirez,
A.; Lobkovsky, E.; Collum, D. B. J. Am. Chem. Soc. 2006, 128, 13753.
(b) Mogali, S.; Darville, K.; Pratt, L. M. J. Org. Chem. 2001, 66, 2368.
(ii) For a discussion on the limitations of DFT microsolvation models in
organolithium chemistry, see: (c) Pratt, L. M.; Mogali, S.; Glinton, K. J.
Org. Chem. 2003, 68, 6484.
(28) Serial solvation of aryllithiums reveals negative enthalpies up to trisolvation
(in kcal/mol): ∆H3(6) ) -8.0, ∆H3(7) ) -7.6, ∆H3(8) ) -8.7, ∆H3(9)
) -6.5, ∆H3(10) ) -7.6, ∆H3(PhLi) ) -9.3.
(29) Carlqvist, P.; O¨ stmark, H.; Brinck, T. J. Org. Chem. 2004, 69, 3222.
(30) (a) Fressigne´, C.; Lautrette, A.; Maddaluno, J. J. Org. Chem. 2005, 70,
7816. (b) Pomelli, C. S.; Bianucci, A. M.; Crotti, P.; Favero, L. J. Org.
Chem. 2004, 69, 150. (c) Gillies, M. B.; Tønder, J. E.; Tanner, D.; Norrby,
P. J. Org. Chem. 2002, 67, 7378. (d) Fressigne´, C.; Maddaluno, J.; Marquez,
A.; Giessner-Prettre, C. J. Org. Chem. 2000, 65, 8899.
(31) (a) Saunders, W. H., Jr.; Cockerill, A. F. Mechanisms of Elimination
Reactions; John Wiley & Sons: New York, 1973; Vol. II, pp 60-68. (b)
Baciocchi, E. In The Chemistry of Halides, Pseudo Halides and Azides,
Supplement D; Patai, S., Rappoport, Z., Eds.; Wiley: Chichester, U. K.,
1983.
(32) Theoretical studies of benzyne-halide ion complexes: (a) Morgon, N. H.
J. Phys. Chem. 1995, 99, 17832. (b) Wong, M. W. J. Chem. Soc., Chem.
Commun. 1995, 2227. (c) Linnert, H. V.; Riveros, J. M. J. Chem. Soc.,
Chem. Commun. 1993, 48.
9
3410 J. AM. CHEM. SOC. VOL. 130, NO. 11, 2008

Formation of Substituted Benzynes
A R T I C L E S
kcal/mol as the C(2)-halide bond significantly exceeds normal
of LiF at low THF concentrations and elimination of LiCl at
high THF concentrations (eq 15). To be more precise, it is the
dependence of the LiF elimination on the THF concentration
that causes regiochemical variation; the rate of LiCl elimination
is independent of THF concentration. Rate studies of the
competitive elimination of aryllithium 6 show that LiCl and
LiF eliminate via 19 and 20, respectively.
bond lengths.33,34
Discussion
Aryllithium Reactivities. Two aspects deriving from experi-
ment attracted our attention: (1) the solvent-dependent elimina-
tion of LiCl versus LiF and (2) the high reactivity of lithiated
haloanisoles 9 and 10. A number of literature precedents note
that haloanisoles are especially reactive toward LiX elimina-
tion.36 Comparing the calculated aryllithium energies (eq 14,
Table 4) with measured relative rate constants qualitatively
shows that greater stability correlates with lower reactivity
(Table 2). Unfortunately, computational studies of the LiX
elimination to form benzynes have been thwarted on several
occasions. Brinck and co-workers studied the decomposition
of 2-halophenyl carbanions (no counterion) using B3LYP and
MP2 computational methods.29 Although strong solvent effects
were implicated, the computations failed to find gas-phase
transition structures. We also failed to find legitimate transition
structures starting from lithium counterions and the full range
of solvation numbers.37 The relaxed-scan approximation showed
that stretching the Li-C(1) bond causes more extensive
lengthening of the Li-Cl bond than Li-F bond, suggesting a
greater E1cb (carbanion) flavor to the LiF elimination.
An interest at Pfizer in the synthesis of varenicline (Champix
and Chantix; 1) and structurally related modulators of nicotinic
receptors led to extensive investigations of 2-haloaryllithium
eliminations, which form benzynes.5,6 It was the pronounced
solvent-dependent regioselectivity described in eq 2 that led to
a Pfizer-Cornell collaboration and the results described herein.
Aryllithium Structures. Structural and computational studies
afford a reasonably coherent description of the structures of
2-haloaryllithiums 6-10. 6Li and 13C NMR spectroscopic
studies reveal that 6-10 are exclusively monomeric in THF
and THF/hydrocarbon mixtures. The solvation numbers were
evaluated using DFT methods. Serial solvation of aryllithiums
reveal negative enthalpies up to trisolvation and negative free
energies for all but the third solvation. Because DFT methods
often overestimate steric effects in congested systems,27 it seems
quite likely that 6-10 are trisolvated.
Spectroscopic studies offered indirect evidence of strong
Li-F interactions in monomers 6, 8, and 10 in the form of
unusually large two-bond 13C-19F coupling.10 The computations
implicate distorted geometries (see 15 or 16) even at the highest
solvation numbers. It would seem, however, that the difluoro
derivative 8 ought to display an attenuated per-fluorine 19F-
13C coupling because of a 50% statistical weighting;27a,35 both
fluorines in 15 cannot interact with Li concurrently. Nonetheless,
8 shows no diminished 19F-13C coupling when compared with
6 and 10.
Mechanism of Benzyne Formation. Rate studies reveal
distinct patterns in the lithium halide eliminations en route to
benzyne. Eliminations of LiF from 8 and 10 occur predomi-
nantly via pathways requiring dissociation of a single THF prior
to the rate-limiting steps. Assignment of the aryllithium
monomers as trisolvates implicates disolvated monomer-based
transition structure 12. The trisolvated transition structure 13
may also contribute to a limited extent, but the evidence is not
compelling. By contrast, the LiCl eliminations from 7 and 9
require no such dissociations, suggesting that direct elimination
via trisolvated monomer-based transition structure 11 is the
exclusive pathway.
Conclusions
We described herein structural and mechanistic studies of
2-haloaryllithiums and their conversion to benzynes. A number
of facets of this work would have been difficult to predict only
a few years ago. Certainly using benzynes for the synthesis of
the pharmacologically important varenicline in a pharmaceutical
setting is unorthodox. Also, the rates and mechanisms of the
lithium halide eliminations to form benzynes are unusual. It
seems unlikely, for example, that many would have predicted
that elimination of LiF could be more facile than elimination
of LiCl.38 By probing the role of solvation as a decidedly
(36) (i) Reports of facile formation of benzyne from a 2-halo-6-methoxyaryl-
lithiums: (a) Dabrowski, M.; Kubicka, J.; Lulinski, S.; Serwatowski, J.
Tetrahedron Lett. 2005, 46, 4175. (b) Barluenga, J.; Fan˜ana´s, F. J.; Sanz,
R.; Ferna´ndez, Y. Chem.-Eur. J. 2002, 8, 2034. (c) Slocum, D. W.; Dietzel,
P. Tetrahedron Lett. 1999, 40, 1823. (d) Iwao, M. J. Org. Chem. 1990, 55,
3622. (e) Adejare, A.; Miller, D. D. Tetrahedron Lett. 1984, 25, 5597. (ii)
For the role of the ortho-substituent on the stabilization of 2-haloaryllithiums
toward benzyne formation, see: (f) Mongin, F.; Rebstock, A.-S.; Tre´court,
F.; Que´guiner, G.; Marsais, F. J. Org. Chem. 2004, 69, 6766. (g) Gohier,
F.; Mortier, J. J. Org. Chem. 2003, 68, 2030.
(37) For analogous results of shallow potentials during the evaluation of transition
structures involving C-halogen lengthening, see: (a) Xie, J.; Feng, D.; Feng,
S.; Zhang, J. Chem. Phys. 2006, 323, 185. (b) Flock, M.; Marschner, C.
Chem.-Eur. J. 2005, 11, 4635. (c) Gronert, S.; Kass, S. R. J. Org. Chem.
1997, 62, 7991. (d) Merrill, G. N.; Gronert, S.; Kass, S. R. J. Phys. Chem.
A 1997, 101, 208.
Regioselectivity of Benzyne Formation. The solvent-de-
pendent regioselectivity described by eq 2 favors elimination
(33) As the C(2)-halide bond was lengthened, achieving wave function
convergence became more difficult, a common limitation of DFT calcula-
tions.
(34) Calculations of separate benzyne and XLiS3 products result in the following
(38) For examples of preferential elimination of F versus Cl, see: (a) Bach, R.
D.; Evans, J. C. J. Am. Chem. Soc. 1986, 108, 1374. (b) Baciocchi, E.;
Ruzziconi, R. J. Org. Chem. 1984, 49, 3395. (c) Baciocchi, E.; Ruzziconi,
R.; Sebastiani, G. V. J. Am. Chem. Soc. 1983, 105, 6114. (d) Baciocchi,
E.; Ruzziconi, R.; Sebastiani, G. V. J. Org. Chem. 1982, 47, 3237. (e)
Okuhara, K. J. Org. Chem. 1976, 41, 1487.
∆G values relative to the reactants (kcal/mol): ∆G(6LiF) ) +33.8, ∆G(6LiCl
)
) +17.3, ∆G(7) ) +15.1, ∆G(8) ) +35.6, ∆G(9) ) +8.9, ∆G(10) )
+26.5.
(35) De la Lande, A.; Fressigne´, C.; Ge´rard, H.; Maddaluno, J.; Parisel, O.
Chem.-Eur. J. 2007, 13, 3459.
9
J. AM. CHEM. SOC. VOL. 130, NO. 11, 2008 3411

A R T I C L E S
Riggs et al.
series and titrated a second time. A series of oven-dried, argon-flushed
5 mL serum vials (8-10 per rate constant) fitted with stir bars were
charged with a stock solution containing the corresponding haloarene,
spiro[2.4]hepta-4,6-diene (in excess relative to the aryllithium),17 THF/
toluene, and n-octane (0.15 M; GC standard). The reaction vials were
maintained at the indicated temperature (0.2 °C (Table 3). The
reactions were initiated by adding aliquots of n-BuLi in toluene, also
maintained at the same temperature as the vials, to afford the desired
overall ArLi concentration with a 0.2 M excess of ArH. The vessels
were serially quenched with 1:1 H2O-THF at intervals chosen to ensure
an adequate sampling of the first 10% conversion. The quenched
aliquots were extracted into Et2O, and the Et2O extracts were washed
with saturated aqueous solution of NH4Cl. The extracts were analyzed
using an autoinjecting GC fitted with a 60 meter DB-5 column. The
metalations were monitored by following the decrease of haloarene
relative to the internal n-octane standard. The observed rate constants
were determined by nonlinear least-squares analyses and shown to be
reproducible within (10%. Following the formation of the correspond-
ing cycloadducts, 5a-d afforded initial slopes that were converted to
rate constants. The reported errors correspond to one standard deviation.
molecular phenomenon, we found that LiF and LiCl eliminations
are distinguished by the dissociation of a THF ligand in the
former. We suspect that the importance of solvent dissociation
from organolithiums (rather than association) may not be widely
appreciated by the community at large. One of the most
pedestrian observations, the unusually large two-bond 13C-19F
coupling first mentioned in the preliminary communication,7
has found a niche. Menzel and co-workers at Merck exploited
the large coupling to confirm the ortholithiation of an aryl
fluoride.10b One never knows a priori what will pique the interest
of the community.
Experimental Section
Reagents and Solvents. Arenes used to form 6-10 are commercially
available. n-BuLi used in the rate studies was obtained from Acros
and recrystallized.8 [6Li]n-BuLi was prepared and recrystallized as
described previously.8 THF, n-pentane, and toluene were distilled from
sodium/benzophenone. The pentane still contained 1% tetraglyme to
dissolve the ketyl. The diphenylacetic acid used to check solution titers39
was recrystallized from methanol and sublimed at 120 °C under full
vacuum. Air- and moisture-sensitive materials were manipulated under
argon or nitrogen using standard glove box, vacuum line, and syringe
techniques.
Acknowledgment. D.B.C., J.C.R., and A.R. thank the Na-
tional Science Foundation for direct support of this work as
well as Pfizer for indirect support.
Kinetics. For a kinetic run corresponding to a single rate constant,
a relatively concentrated (2.2 M) solution of n-BuLi in toluene at -78
°C was prepared and titrated to determine the precise concentration.
The solution was diluted to a concentration appropriate for the particular
Supporting Information Available: NMR, rate, and compu-
tational data. Complete refs 5b and 21. This material is available
free of charge via the Internet at http://pubs.acs.org.
(39) Kofron, W. G.; Baclawski, L. M. J. Org. Chem. 1976, 41, 1879.
JA0754655
9
3412 J. AM. CHEM. SOC. VOL. 130, NO. 11, 2008