Thermodynamic and kinetic effects of the double coordination of carbonyl groups by bidentate Lewis acids

Article abstract of DOI:10.1021/ja982725x

Treatment of 1,2,3,4-tetrahexylbenzene or 1,2,4,5-tetrahexylbenzene with Hg(OOCCF₃)₂ produces 1,2-phenylenedimercury bis(trifluoroacetate) 7 or its 1,4-isomer 9, which are bidentate Lewis acids. IR studies in CH₂Cl₂

Full text of DOI:10.1021/ja982725x

13016
J. Am. Chem. Soc. 1998, 120, 13016-13022
Thermodynamic and Kinetic Effects of the Double Coordination of
Carbonyl Groups by Bidentate Lewis Acids
Jean Vaugeois1 and James D. Wuest*
Contribution from the De´partement de Chimie, UniVersite´ de Montre´al, Montre´al,
Que´bec, H3C 3J7 Canada
ReceiVed July 31, 1998
Abstract: Treatment of 1,2,3,4-tetrahexylbenzene or 1,2,4,5-tetrahexylbenzene with Hg(OOCCF3)2 produces
1,2-phenylenedimercury bis(trifluoroacetate) 7 or its 1,4-isomer 9, which are bidentate Lewis acids. IR studies
in CH2Cl2 established that 1,2-isomer 7 forms doubly coordinated 1:1 complexes with amides in which both
atoms of mercury interact simultaneously with the carbonyl oxygen atom of the bound amide, whereas 1,4-
isomer 9 forms more weakly bound singly coordinated complexes. Variable-temperature IR studies revealed
that for the double coordination of diethylformamide (DEF) by 1,2-isomer 7, ∆H° ) -4.5 ( 0.2 kcal mol-1
and ∆S° ) -6.9 ( 1.0 eu, whereas for the single coordination of DEF by 1,4-isomer 9, ∆H° ) -3.4 ( 0.2
kcal mol-1 and ∆S° ) -8.1 ( 1.0 eu. These measurements establish that double coordination is significantly
more exothermic than single coordination. In addition, variable-temperature NMR studies showed that ∆G‡
for rotation around the N-C(O) bond of dimethylpivalamide (normally 12.2 ( 0.1 kcal mol-1 in CDCl3) is
increased more by the addition of an equimolar amount of 1,2-isomer 7 (12.9 ( 0.1 kcal mol-1) than by the
addition of an equimolar amount of 1,4-isomer 9 (12.3 ( 0.1 kcal mol-1). This demonstrates that the double
coordination of carbonyl compounds by bidentate Lewis acids can have measurable effects on the rates of
subsequent chemical transformations.
Introduction
Lewis acids play a major role as catalysts and stoichiometric
reagents in contemporary organic synthesis.
Lewis acids are useful because they can bind complementary
basic molecules and thereby modify their reactivity. An example
of particular practical importance is the activation of carbonyl
compounds by the formation of complexes 1 with Lewis acids.2
The precise geometries of adducts 1 vary, depending on the
Lewis acid and the substrate. In general, neutral main-group
Lewis acids favor η1(σ) complexes in which the Lewis acid
lies close to the carbonyl plane along the direction of one of
the formal sp2 lone pairs on oxygen.2,6 In contrast, Lewis acids
incorporating late transition metals tend to prefer η2(π) com-
plexes in which the metal lies above the carbonyl plane and
interacts simultaneously with both the carbon atom and the
oxygen atom of the carbonyl group.2 In principle, either mode
of complexation leaves one or more sp2 lone pairs free and
available for binding a second Lewis acid, which would produce
doubly coordinated termolecular complexes 2. Such adducts are
potentially valuable in synthesis because they promise to be
even more highly activated than their singly coordinated
analogues 1.
Unfortunately, doubly coordinated termolecular complexes
2 are presently unknown.7,8 However, we have shown in
previous work that closely related doubly coordinated bimo-
lecular complexes 3 can be formed when two sites of Lewis
acidity are joined to create single bidentate reagents.9,10
Furthermore, recent reports suggest that the double coordination
By reducing the electron density on the carbonyl carbon, further
polarizing the carbon-oxygen double bond, and lowering the
energy of the lowest unoccupied molecular orbital, the com-
plexation of carbonyl compounds by Lewis acids can have
dramatic kinetic effects. For example, the Diels-Alder reaction
of methyl acrylate with butadiene at 20 °C in benzene is 105
times faster in the presence of an equimolar amount of AlCl3.3,4
Furthermore, Lewis acids can also have beneficial effects on
regioselectivity and stereoselectivity.5 For all of these reasons,
(3) Inukai, T.; Kojima, T. J. Org. Chem. 1966, 32, 872.
(4) For related work, see: Angus, P. M.; Fairlie, D. P.; Jackson, W. G.
Inorg. Chem. 1993, 32, 450.
(5) For example, see: Narasaka, K. Synthesis 1991, 1.
(6) Wiberg, K. B.; Marquez, M.; Castejon, H. J. Org. Chem. 1994, 59,
681(7.) Related complexes are known in which carbonyl compounds are
doubly protonated or doubly hydrogen bonded. For references, see: Saied,
O.; Simard, M.; Wuest, J. D. J. Org. Chem. 1998, 63, 3756.
(8) Doubly coordinated termolecular complexes 2 have been postulated
as activated intermediates or transition states in additions of organometallic
reagents to carbonyl compounds. For example, see: Evans, D. A. Science
1988, 240, 420.
(1) Fellow of the Natural Sciences and Engineering Council of Canada,
1988-1992.
(2) Shambayati, S.; Crowe, W. E.; Schreiber, S. L. Angew. Chem., Int.
Ed. Engl. 1990, 29, 256. Hay, R. W. In ComprehensiVe Coordination
Chemistry; Wilkinson, G., Gillard, R. D., McCleverty, J. A., Eds.; Pergamon
Press: New York, 1987; Vol. 6, Chapter 61.4.
10.1021/ja982725x CCC: $15.00 © 1998 American Chemical Society
Published on Web 12/04/1998

Bidentate Lewis Acids
J. Am. Chem. Soc., Vol. 120, No. 50, 1998 13017
of carbonyl compounds by bidentate Lewis acids increases their
reactivity,11,12 although in none of the reported cases have the
structures of the intermediate complexes and the origin of the
activation been determined unambiguously.
Lewis bases.9 In particular, cocrystallization of bidentate Lewis
acids 4 and 5 with various N,N-dialkylformamides gives doubly
coordinated complexes that can be represented by structure 6.
In these complexes, the two sites of Lewis acidity do not both
lie in the carbonyl plane but instead occupy widely varying
positions with respect to the plane, suggesting that carbonyl
oxygen atoms do not have very strong steric or electronic
preferences for dative bonding to mercury along specific
directions. This observation is consistent with the expectation
that the association of carbonyl compounds with Lewis acidic
organomercury compounds is inherently weak.13
This earlier work establishes that the double coordination of
carbonyl groups by Lewis acids is possible and may have useful
consequences. Nevertheless, many fundamental questions about
double coordination remain unanswered. In particular, no
detailed thermodynamic comparison of single and double
coordination has been made, and no values of ∆H° and ∆S°
for double coordination are available. Furthermore, in no
chemically well-defined case has double coordination been
shown to have special effects on the rates of subsequent
chemical transformations. We now provide this important
information by describing the thermodynamic and kinetic effects
of the double coordination of the carbonyl group of amides by
bidentate Lewis acids derived from 1,2-phenylenedimercury.
For this reason, the double coordination of carbonyl com-
pounds by bidentate Lewis acids derived from 1,2-phenylene-
dimercury is unlikely to find important applications in organic
synthesis. However, these reagents offer practical advantages
that compensate for their weak Lewis acidity and make them
attractive for thermodynamic and kinetic studies of double
coordination. In particular, (1) effective methods exist for their
synthesis;9 (2) unlike many other organometallic reagents, they
do not react with the carbonyl groups of potential substrates by
nucleophilic addition; (3) in general, they are not highly sensitive
to H2O or O2; and (4) the structures of their complexes with
N,N-dialkylformamides and other basic organic molecules are
well-established, at least in the solid state, and reveal a consistent
preference for double coordination. As a result, we decided to
use bidentate Lewis acids derived from 1,2-phenylenedimercury
in the first quantitative thermodynamic and kinetic comparison
of the single and double coordination of carbonyl compounds.
In making this choice, we were conscious of the specialized
nature and limited practical value of reagents derived from 1,2-
phenylenedimercury; however, we were optimistic that our
comparisons would be feasible and would provide fundamental
information of general significance.
Results and Discussion
Syntheses of 3,4,5,6-Tetrahexyl-1,2-phenylenedimercury
Bis(trifluoroacetate) (7), 2,3,5,6-Tetrahexyl-1,4-phenylene-
dimercury Bis(trifluoroacetate) (9), and Related Compounds.
Various derivatives of 1,2-phenylenedimercury, including 1,2-
phenylenedimercury dichloride (4) and 3,4,5,6-tetramethyl-1,2-
phenylenedimercury bis(trifluoroacetate) (5), are known to hold
Unfortunately, the solubilities of 1,2-phenylenedimercury
dichloride (4) and 3,4,5,6-tetramethyl-1,2-phenylenedimercury
bis(trifluoroacetate) (5) in common noncoordinating organic
solvents are too low to allow convenient study of double
coordination in solution. To make more soluble derivatives, we
prepared 3,4,5,6-tetrahexyl-1,2-phenylenedimercury bis(trifluo-
roacetate) (7) in 81% yield by treating 1,2,3,4-tetrahexylbenzene
(8) with Hg(OOCCF3)2. Its 1,4-phenylenedimercury isomer 9,
a closely related bidentate Lewis acid that cannot form a doubly
coordinated complex 6, was made similarly from 1,2,4,5-
tetrahexylbenzene (10) in 52% yield. The required tetrahexyl-
benzenes were synthesized from 1,2,4,5-tetraiodobenzene (11)14
according to Scheme 1. The synthesis of 1,2,3,4-tetrahexylben-
zene (8) features a modified version of the Jacobsen reaction,
in which sulfonation of highly substituted alkylbenzenes by
concentrated sulfuric acid normally induces the alkyl groups to
migrate to positions remote from the site of sulfonation.15,16
Although 1,2,4,5-tetrahexylbenzene (10) did not undergo the
normal Jacobsen reaction, we found that treatment with chloro-
sulfonic acid produced the rearranged 2,3,4,5-tetrahexylben-
zenesulfonyl chloride (13),17 which was subsequently converted
into 1,2,3,4-tetrahexylbenzene (8) according to Scheme 1.
two Lewis acidic atoms of mercury in an orientation that favors
the double coordination of carbonyl compounds and other simple
(9) Wuest, J. D. Acc. Chem. Res. 1998, in press. Vaugeois, J.; Simard,
M.; Wuest, J. D. Coord. Chem. ReV. 1995, 145, 55.
(10) For recent related studies of multidentate Lewis acids, see: Anto-
nisse, M. M. G.; Reinhoudt, D. N. J. Chem. Soc., Chem. Commun. 1998,
443. Tschinkl, M.; Schier, A.; Riede, J.; Mehltretter, G.; Gabba¨ı, F. P.
Organometallics 1998, 17, 2921. Hawthorne, M. F.; Zheng, Z. Acc. Chem.
Res. 1997, 30, 267. Beer, P. D.; Smith, D. K. Prog. Inorg. Chem. 1997,
46, 1. Schmidtchen, F. P.; Berger, M. Chem. ReV. 1997, 97, 1609.
Chistyakov, A. L.; Stankevich, I. V.; Gambaryan, N. P.; Struchkov, Yu.
T.; Yanovsky, A. I.; Tikhonova, I. A.; Shur, V. B. J. Organomet. Chem.
1997, 536-537, 413. Altmann, R.; Jurkschat, K.; Schu¨rmann, M.; Dak-
ternieks, D.; Duthie, A. Organometallics 1997, 16, 5716. Grdenic´, D.;
Korpar-Cˇ olig, B.; Matkovic´-Cˇ alogovic´, D. J. Organomet. Chem. 1996, 522,
297. Tamao, K.; Hayashi, T.; Ito, Y. J. Organomet. Chem. 1996, 506, 85.
(11) For example, see: Ooi, T.; Takahashi, M.; Maruoka, K. Angew.
Chem., Int. Ed. Engl. 1998, 37, 835. Reilly, M.; Oh, T. Tetrahedron Lett.
1995, 36, 221.
(13) Laszlo, P.; Teston, M. J. Am. Chem. Soc. 1990, 112, 8750.
(14) Mattern, D. L. J. Org. Chem. 1984, 49, 3051.
(15) Suzuki, H. Bull. Chem. Soc. Jpn. 1963, 36, 1642. Smith, L. I. Org.
React. 1942, 1, 370.
(16) Koeberg-Telder, A.; Cerfontain, H. Recl. TraV. Chim. Pays-Bas
1987, 106, 85.
(17) It is noteworthy that the Jacobsen reaction has not previously been
used to synthesize simple tetraalkylbenzenes with substituents other than
methyl or ethyl, and chlorosulfonic acid has been reported to be incapable
of promoting the reaction.16
(12) For an earlier related study of the activation of thiocarbonyl
compounds by bidentate Lewis acids, see: Wuest, J. D.; Zacharie, B. J.
Am. Chem. Soc. 1985, 107, 6121.

13018 J. Am. Chem. Soc., Vol. 120, No. 50, 1998
Vaugeois and Wuest
Scheme 1
Figure 1. van’t Hoff plots for the association of bidentate Lewis acids
7 (b) and 9 (9) with N,N-diethylformamide in CH2Cl2. The lines
provide the best fit to the van’t Hoff equation (ln Ka ) -∆H°/RT +
∆S°/R), with ∆H° ) -4.5 ( 0.2 kcal mol-1 and ∆S° ) -6.9 ( 1.0 eu
for 1,2-isomer 7, and with ∆H° ) -3.4 ( 0.2 kcal mol-1 and ∆S° )
-8.1 ( 1.0 eu for 1,4-isomer 9.
bis(trifluoroacetate) (5) in the solid state suggests that tetrahexyl
analogue 7 behaves similarly in solution and favors doubly
coordinated adducts 6 in which one or more atoms of mercury
lie out of the carbonyl plane.
As expected, hexyl-substituted phenylenedimercury bis(tri-
fluoroacetates) 7 and 9 both showed much higher solubilities
in organic solvents such as CH2Cl2 than their methyl-substituted
analogue 5. The corresponding 3,4,5,6-tetrakis(tetradecyl)-1,2-
phenylenedimercury bis(trifluoroacetate) (16), prepared in an
analogous way via 1,2,3,4-tetrakis(tetradecyl)benzene (15),
proved to have similar solubility, and thus we decided to use
hexyl-substituted derivatives in subsequent thermodynamic and
kinetic studies of the coordination of carbonyl compounds.
Thermodynamic Effects of the Single and Double Coor-
dination of Carbonyl Compounds by Lewis Acids. For the
following reasons, we elected to use N,N-dialkylformamides as
the carbonyl substrates in our studies: (1) Amides are among
the most basic carbonyl compounds,18 allowing them to associate
with Lewis acids that are intrinsically weak; and (2) simple N,N-
dialkylformamides are known to form well-characterized mul-
tiply coordinated adducts with 3,4,5,6-tetramethyl-1,2-phenyl-
enedimercury bis(trifluoroacetate) (5) in the solid state, and we
could reasonably expect its tetrahexyl analogue 7 to behave
similarly, possibly even in solution.
Quantitative analysis of IR spectra recorded at -40, -20, 0,
and 20 °C yielded association constants (Ka) for the binding of
DEF by bidentate Lewis acids 7 and 9,19 and van’t Hoff plots
(Figure 1) revealed that for the double coordination of DEF by
1,2-isomer 7, ∆H° ) -4.5 ( 0.2 kcal mol-1 and ∆S° ) -6.9
( 1.0 eu, whereas for the single coordination of DEF by 1,4-
isomer 9, ∆H° ) -3.4 ( 0.2 kcal mol-1 and ∆S° ) -8.1 (
1.0 eu. As expected, both values of |∆H°| are very small, and
they provide quantitative confirmation that the association of
carbonyl compounds with Lewis acidic organomercury com-
pounds is inherently weak.13 Nevertheless, our measurements
are noteworthy because they provide the first quantitative
demonstration that double coordination is significantly more
exothermic than single coordination. Specifically, our results
suggest that conversion of a singly coordinated complex into a
doubly coordinated complex is exothermic by 1.1 ( 0.2 kcal
mol-1 in this particular system, or approximately 32% as
exothermic as initial formation of the singly coordinated adduct
itself. This provides experimental support for earlier ab initio
calculations of ∆H° for formation of the singly and doubly
coordinated 1:1 and 1:2 σ complexes of formaldehyde with BH3
in the gas phase, which suggested that ∆H° for addition of the
second equivalent of BH3 (-2.3 kcal mol-1) is 32% of the value
of ∆H° for addition of the first (-7.3 kcal mol-1).20 It is
noteworthy that a similar thermodynamic relationship between
single and double coordination governs two widely different
Lewis acids that form complexes likely to have distinctly
different geometries. This encourages us to believe that our
thermodynamic and kinetic studies of the effects of double
coordination in a special system may have broader significance.
The double coordination of carbonyl compounds by two sites
of Lewis acidity may be enthalpically feasible in other systems
as well, and ∆H° for coordination of the second site may be a
significant fraction of ∆H° for coordination of the first.
IR spectroscopy permitted an initial assessment of the
association of bidentate Lewis acid 7 or its 1,4-isomer 9 with
DEF in CH2Cl2 at various concentrations. The carbonyl regions
of typical spectra showed peaks at 1696 and 1668 cm-1
corresponding to terminal trifluoroacetate and free amide,
respectively. In addition, characteristic bands appeared at 1633
cm-1 for DEF bound by 1,2-isomer 7 and at 1647 cm-1 for
DEF bound by 1,4-isomer 9. The distinctly different stretching
frequencies observed for bound DEF immediately suggested that
1,2-isomer 7 forms a doubly coordinated complex in solution,
as related compounds do in the solid state, whereas 1,4-isomer
9 forms a singly coordinated complex. Double coordination
causes a significantly larger shift relative to free DEF (35 cm-1
)
than single coordination does (21 cm-1), suggesting that double
coordination is stronger than single coordination and that it
weakens the carbon-oxygen double bond more substantially.
In neither case do our data provide information about the precise
geometry of coordination, but the consistent behavior of the
closely related bidentate Lewis acids 1,2-phenylenedimercury
dichloride (4) and 3,4,5,6-tetramethyl-1,2-phenylenedimercury
(19) For an instructive related study that used variable-temperature IR
spectroscopy for the quantitative evaluation of weak association, see:
Gellman, S. H.; Dado, G. P.; Liang, G.-B.; Adams, B. R. J. Am. Chem.
Soc. 1991, 113, 1164.
(20) LePage, T. J.; Wiberg, K. B. J. Am. Chem. Soc. 1988, 110, 6642.
Deschatelets, P.; Wuest, J. D., unpublished results.
(18) Gal, J.-F.; Maria, P.-C. Prog. Phys. Org. Chem. 1990, 17, 159.

Bidentate Lewis Acids
J. Am. Chem. Soc., Vol. 120, No. 50, 1998 13019
Scheme 2
carefully matched and fully suitable for assessing the thermo-
dynamic and kinetic effects of single and double coordination.
Because the two bidentate Lewis acids are structurally similar
and have two ortho or para Hg(OOCCF3) groups, any com-
parisons of their reactivity are likely to be free of significant
steric or electronic bias.
If it is true that the coordination chemistry of bidentate Lewis
acids derived from 1,2-phenylenedimercury is insensitive to
steric effects, then amides more highly substituted than DEF
and DMF should also form similar doubly coordinated com-
plexes 6. This was confirmed by an X-ray crystallographic study
of the 1:1 complex of 3,4,5,6-tetramethyl-1,2-phenylenedimer-
cury bis(trifluoroacetate) (5) with N,N-diethylacetamide (DEA).
The results of this study, which are summarized in Figure 2,
establish that both Lewis acidic atoms of mercury bind the
carbonyl oxygen atom of DEA and that both lie out of the
carbonyl plane, creating the partial structures shown in Figure
2. All four carbonyl O‚‚‚Hg distances are much shorter than
the sum of the van der Waals radii of oxygen (1.40 Å)21 and
mercury (1.73 Å),22 and the average distance (2.68(2) Å) is
similar to the one measured for doubly coordinated DEF in its
3:2 complex with bis(trifluoroacetate) 5 (2.75(2) Å).23 Simul-
taneously, the two atoms of mercury in the 1:1 complex of
bis(trifluoroacetate) 5 with DEA retain enough Lewis acidity
to doubly bind the carbonyl oxygen atom of a trifluoroacetate
group in a neighboring molecule.24 However, this additional
double coordination is much weaker than that of DEA, and the
average carbonyl O‚‚‚Hg distance is 2.94(2) Å.
In the 1:1 complex of bis(trifluoroacetate) 5 with DEA, the
average carbonyl C-O bond length in the bound amide (1.27(2)
Å) is longer than the generally accepted value for free tertiary
amides (1.23(1) Å), and the N-C(O) bond length (1.32(3) Å)
is slightly shorter than the normal value (1.35(1) Å).25 These
differences are consistent with the hypothesis that double
coordination weakens the carbon-oxygen double bond but they
are not large enough to be considered crystallographically
significant.
Highly substituted amides are complexed by bidentate Lewis
acids derived from 1,2-phenylenedimercury in solution, as well
as in the solid state. For example, we found that variable-
temperature IR spectra of solutions of N,N-dimethylpivalamide
(DMP) in CHCl3 (0.093 M) containing equimolar amounts of
either 3,4,5,6-tetrahexyl-1,2-phenylenedimercury bis(trifluoro-
acetate) (7) or its 1,4-isomer 9 resemble analogous spectra of
N,N-dialkylformamides such as DEF. Specifically, the carbonyl
regions show a band at 1609 cm-1 characteristic of free DMP
and an additional band at either 1569 or 1577 cm-1, corre-
sponding to DMP that is doubly or singly coordinated,
respectively. In contrast, similar IR studies showed that ketones
and other less basic carbonyl compounds are bound more weakly
or not at all by bidentate Lewis acid 7 or its 1,4isomer 9.
Kinetic Effects of the Single and Double Coordination of
Carbonyl Compounds by Lewis Acids. Collectively, our
structural, spectroscopic, and thermodynamic studies suggest
that the double coordination of amides by bidentate Lewis acids
derived from 1,2-phenylenedimercury is a general phenomenon
and that it weakens the carbon-oxygen double bond more than
single coordination does. If so, double coordination should have
As expected, the observed values of ∆S° for single and double
coordination are both consistent with the formation of 1:1
complexes. The two values are not significantly different, even
though it is likely that the doubly coordinated adduct is
somewhat more highly ordered. In part, the similarity of the
values is presumably due to the rigid orientation of the Lewis
acidic atoms of mercury in 3,4,5,6-tetrahexyl-1,2-phenylene-
dimercury bis(trifluoroacetate) (7). Furthermore, previously
reported structural data indicate that dative bonds between
oxygen and mercury are characteristically long, weak, and
deformable,9 so the values of |∆S°| associated with their
formation should be unusually small. In such cases, desolvation
may make an important contribution to the observed values of
∆S°.
We believe that the observed values of ∆H° are different
primarily because 3,4,5,6-tetrahexyl-1,2-phenylenedimercury
bis(trifluoroacetate) (7) forms a doubly coordinated complex
with DEF, whereas its 1,4-isomer 9 forms a singly coordinated
complex. For the following reasons, we doubt that subtle steric
differences between bidentate Lewis acid 7 and its 1,4-isomer
9 have important effects on ∆(∆H°). Steric effects are likely to
be negligible because of the deformable nature of dative bonds
between oxygen and mercury; in addition, the characteristically
long Hg-C bonds in phenylmercury compounds make the sites
of coordination remote from the potential steric influence of
adjacent alkyl substituents.
To assess possible electronic differences between 3,4,5,6-
tetrahexyl-1,2-phenylenedimercury bis(trifluoroacetate) (7) and
its 1,4-isomer 9, we synthesized a soluble monodentate analogue,
(2,3,4,5,6-pentahexylphenyl)mercuric trifluoroacetate (17), by
the route summarized in Scheme 2. The Lewis acidities of
bidentate Lewis acid 9 and its monodentate analogue 17 were
evaluated qualitatively by recording their 199Hg NMR spectra
in CD2Cl2 (70 mM, 20 °C) and then by comparing the values
of ∆(δ 199Hg) caused by the addition of equimolar amounts of
DEF. As expected, these experiments indicated that monodentate
Lewis acid 17 is intrinsically less Lewis acidic than either of
its bidentate analogues. This confirms that electronic factors
increase the Lewis acidity of a phenylmercuric trifluoroacetate
when a second Hg(OOCCF3) group is introduced, at least in
the ortho or para positions. We conclude that 3,4,5,6-tetrahexyl-
1,2-phenylenedimercury bis(trifluoroacetate) (7) and its 1,4-
isomer 9 provide a pair of bidentate Lewis acids that are
(21) Pauling, L. The Nature of the Chemical Bond; Cornell University
Press: Ithaca, NY, 1960; p 514.
(22) Canty, A. J.; Deacon, G. B. Inorg. Chim. Acta 1980, 45, L225.
(23) Simard, M.; Vaugeois, J.; Wuest, J. D. J. Am. Chem. Soc. 1993,
115, 370.
(24) For a similar observation, see: Deguire, S.; Beauchamp, A. L. Acta
Crystallogr. 1990, C46, 27.
(25) Chakrabarti, P.; Dunitz, J. D. HelV. Chim. Acta 1982, 65, 1555.

13020 J. Am. Chem. Soc., Vol. 120, No. 50, 1998
Vaugeois and Wuest
Figure 2. ORTEP views of the structure of the 1:1 complex of µ-(3,4,5,6-tetramethyl-1,2-phenylene)bis(trifluoroacetato-O)dimercury (5) with
N,N-diethylacetamide (DEA). (a) The upper view provides the atomic-numbering scheme. Non-hydrogen atoms are represented by ellipsoids
corresponding to 30% probability, hydrogen atoms are omitted for clarity, dative bonds to mercury are indicated using narrow lines, and bonds in
DEA are represented by solid lines. Important bond lengths include Hg(11)-O(31) ) 2.63(2), Hg(12)-O(31) ) 2.76(2), Hg(21)-O(41) ) 2.73(2),
Hg(22)-O(41) ) 2.61(2), C(31)-O(31) ) 1.26(2), C(41)-O(41) ) 1.28(2), C(31)-N(31) ) 1.31(3), C(41)-N(41) ) 1.33(3), Hg(11)-O(212)
) 2.82(2), Hg(12)-O(212) ) 2.89(2), Hg(21)-O(112) ) 2.82(2), and Hg(22)-O(112) ) 3.21(2) Å. (b) The lower views show only the bound
amides and the two associated atoms of mercury. To facilitate comparison, each structure is viewed along the carbonyl CdO axis from slightly
above the carbonyl plane. All hydrogen atoms are omitted for simplicity. The three geometric parameters that appear next to each mercury atom
correspond to the carbonyl O‚‚‚Hg distance (Å)/CdO‚‚‚Hg angle (deg)/N-CdO‚‚‚Hg dihedral angle (deg).
unique effects on the reactivity of bound amides. Because
bidentate Lewis acids derived from 1,2-phenylenedimercury are
weak, any special kinetic effects of double coordination will
be small and hard to detect. For this reason, initial comparisons
of the effects of 3,4,5,6-tetrahexyl-1,2-phenylenedimercury
bis(trifluoroacetate) (7) and its 1,4-isomer 9 on various bimo-
lecular reactions of amides proved inconclusive. As an alterna-
tive, we decided to compare the effects of bidentate Lewis acids
7 and 9 on barriers for rotation around the N-C(O) bond in
amides,26 using values of ∆G‡ determined by variable-temper-
kcal mol-1 27
. As expected, the differences are small; neverthe-
less, they are significant, and they establish for the first time
that the double coordination of carbonyl compounds can have
measurable effects on the rates of subsequent chemical trans-
formations. In this particular case, the effects presumably arise
because the association constant for double coordination is
higher than that for single coordination, and because double
coordination further weakens the carbon-oxygen double bond
of amides, further strengthens the N-C(O) bond, and thereby
inhibits rotation more effectively.
1
ature H NMR spectroscopy. To maximize the percentage of
Conclusions
amide bound by the weakly Lewis acidic reagents 7 and 9, we
needed to conduct these experiments at the lowest possible
temperatures, and therefore we required an amide substrate in
which rotation is especially fast. Highly substituted amides such
as DMP satisfy this condition; in addition, we had already
established that DMP is doubly coordinated in solution by
bidentate Lewis acid 7 and singly coordinated by 1,4-isomer 9.
Our results establish that (1) the double coordination of
carbonyl compounds by two sites of Lewis acidity is enthalpi-
cally feasible in the solid state and in solution, (2) ∆H° for
coordination of the second site is a significant fraction of ∆H°
for coordination of the first, and (3) the double coordination of
carbonyl compounds by Lewis acids can have measurable effects
on the rates of subsequent chemical transformations. Our results
are based on the special case of the complexation of amides by
weak Lewis acids derived from phenylenedimercury; neverthe-
less, our conclusions may prove to have general validity, and
significant thermodynamic and kinetic effects may be observed
in other cases when basic molecules are bound by suitable
bidentate Lewis acids.
Coalescence of the methyl signals of DMP occurred at Tc )
-1.5, -14, or -16 ( 1 °C, respectively, in solutions containing
an equimolar amount of 1,2-isomer 7, an equimolar amount of
1,4-isomer 9, or DMP alone (0.093 M in CDCl3). The
corresponding values of ∆G‡ are 12.9, 12.3, and 12.2 ( 0.1
(26) Laidig, K. E.; Cameron, L. M. J. Am. Chem. Soc. 1996, 118, 1737.
Wiberg, K. B.; Rablen, P. R.; Rush, D. J.; Keith, T. A. J. Am. Chem. Soc.
1995, 117, 4261.
(27) Sandstro¨m, J. Dynamic NMR Spectroscopy; Academic Press:
London, 1982; p 96.

Bidentate Lewis Acids
J. Am. Chem. Soc., Vol. 120, No. 50, 1998 13021
combined extracts by evaporation under reduced pressure. Flash
chromatography (silica, hexane (90%)/CHCl3 (10%)) of the residue
provided an analytically pure sample of 2,3,4,5-tetrahexylbenzene-
sulfonyl chloride (13; 0.60 g, 1.2 mmol, 50%) as a pale yellow oil: IR
Bidentate Lewis acids remain poorly studied, but it is
increasingly difficult to dismiss them as esoteric reagents of
mere academic interest. Recent studies have shown that they
can bind and activate organic substrates in genuinely useful
ways,11,12 and our present results provide fundamental new
information that should accelerate development of this field of
research. Moreover, strongly encouraging analogies exist be-
tween the activation induced by bidentate Lewis acids and
catalytic processes that are already known to play important
roles in nature. In particular, double coordination of the oxygen
atom of carbonyl groups is similar to double coordination of
the oxygen atom of phosphoryl groups by two metal cations,
which is believed to be a crucial step in various enzyme-
catalyzed phosphoryl transfers.28,29 As a result, we believe that
bidentate Lewis acids will eventually find broad applications
in chemistry, particularly as catalytic or stoichiometric reagents
in organic synthesis.
1
(CHCl3) 1365, 1161 cm-1; H NMR (300 MHz, CDCl3) δ 0.85-0.95
(m, 12H), 1.25-1.70 (m, 32H), 2.55-2.70 (m, 6H), 2.95-3.05 (m,
2H), 7.75 (s, 1H); 13C NMR (75.4 MHz, CDCl3) δ 14.0 (4C), 22.5,
22.5, 22.6 (2C), 29.0, 29.4, 29.6, 29.7, 29.9, 30.0 (2C), 30.7, 31.0,
31.1, 31.3, 31.3, 31.4 (2C), 31.5, 32.9, 127.5, 137.9, 139.8, 140.9, 142.7,
147.8. Anal. Calcd for C30H53ClO2S C, 70.20; H, 10.41. Found C, 70.37;
H, 10.36.
2,3,4,5-Tetrahexylbenzenethiol (14). A solution of 2,3,4,5-tetra-
hexylbenzenesulfonyl chloride (13; 3.1 g, 6.0 mmol) in THF (50 mL)
was stirred at 25 °C under dry Ar, solid LiAlH4 (1.9 g, 50 mmol) was
added in portions, and the mixture was heated at reflux for 12 h. The
resulting mixture was cooled to 0 °C and treated dropwise with H2O.
Solids were then dissolved by adding 6 N aqueous HCl, and volatiles
were partially removed by evaporation under reduced pressure. The
aqueous residue was extracted with CHCl3, solvent was removed from
the combined extracts by evaporation under reduced pressure, and the
residue was purified by flash chromatography (silica, hexane). This
yielded 2,3,4,5-tetrahexylbenzenethiol (14; 1.6 g, 3.6 mmol, 60%) as
a colorless oil: 1H NMR (300 MHz, CDCl3) δ 0.90-1.00 (m, 12H),
1.30-1.64 (m, 32H), 2.45-2.62 (m, 6H), 2.62-2.71 (m, 2H), 3.21 (s,
1H), 6.97 (s, 1H); 13C NMR (75.4 MHz, CDCl3) δ 14.0 (4C), 22.5,
22.6 (3C), 29.0, 29.5, 29.6, 29.7, 29.8, 30.0, 30.0, 31.3, 31.4, 31.5 (5C),
31.6, 32.8, 127.3, 129.4, 136.8, 137.0, 139.2, 139.8. HRMS (FAB,
thioglycerol) calcd for C30H54S m/e 446.3973, found 446.3946.
1,2,3,4-Tetrahexylbenzene (8). A solution of 2,3,4,5-tetrahexyl-
benzenethiol (14; 12 g, 27 mmol) in a mixture of hexane (250 mL)
and C2H5OH (250 mL) was stirred at 25 °C and treated with solid
KOH (1.8 g, 32 mmol) and then with an aqueous slurry of Raney nickel
(55 g, 50 wt %). After 12 h, a second portion of Raney nickel (12 g)
was added. After an additional period of 24 h, the mixture was filtered,
the filtered solid was rinsed with hexane, and volatiles were removed
from the combined filtrates by evaporation under reduced pressure.
Kugelrohr distillation (200 °C/0.5 Torr) of the residue provided an
analytically pure sample of 1,2,3,4-tetrahexylbenzene (8; 8.5 g, 20
mmol, 74%) as a colorless oil: 1H NMR (300 MHz, CDCl3) δ 0.86-
0.96 (m, 12H), 1.24-1.69 (m, 32H), 2.50-2.60 (m, 8H), 6.94 (s, 2H);
13C NMR (75.4 MHz, CDCl3) δ 14.0, 14.0, 22.6, 22.6, 29.1, 29.6, 30.1,
31.3, 31.5, 31.6, 31.7, 33.0, 126.5, 138.3, 138.6. Anal. Calcd for C30H54
C, 86.88; H, 13.12. Found C, 86.80; H, 12.79.
Experimental Section
Tetrahydrofuran (THF) and ether were dried by distillation from the
sodium ketyl of benzophenone, and piperidine, CS2, and CH2Cl2 were
dried by distillation from CaH2. N,N-Diethylformamide and N,N-
dimethylformamide were purified by distillation at 0.2 Torr and then
dried over 4 Å molecular sieves. PdCl2(PPh3)2 was prepared in the
normal way.30 All other reagents were commercial products that were
used without further purification. Flash chromatography was performed
in the usual manner.31
1,2,4,5-Tetra-1-hexynylbenzene (12). A solution of 1,2,4,5-tet-
raiodobenzene (11; 30 g, 52 mmol),14 PdCl2(PPh3)2 (2.6 g, 3.7 mmol),
and CuI (3.5 g, 18 mmol) in piperidine (1.0 L) was stirred at 25 °C
under dry Ar, 1-hexyne (21 g, 260 mmol) was added, and the resulting
mixture was stirred at 25 °C for 12 h. Volatiles were then removed by
evaporation at 50 °C under reduced pressure, and the solid residue was
extracted with hexane. The extracts were filtered, solvent was removed
by evaporation under reduced pressure, and the residue was purified
by flash chromatography (silica, hexane (98%)/ethyl acetate (2%)) to
give 1,2,4,5-tetra-1-hexynylbenzene (12; 14 g, 35 mmol, 67%) as an
1
analytically pure yellow solid: IR (melt) 2220 cm-1; H NMR (300
MHz, CDCl3) δ 0.95 (t, 12H, 3J ) 7.1 Hz), 1.45-1.65 (m, 16H), 2.45
3
(t, 8H, J ) 6.8 Hz), 7.38 (s, 2H); 13C NMR (75.4 MHz, CDCl3) δ
13.6, 19.2, 21.8, 30.6, 78.8, 95.3, 125.0, 135.0. HRMS (EI) calcd for
C30H38 m/e 398.2974, found 398.2970. Anal. Calcd for C30H38 C, 90.39;
H, 9.61. Found C, 90.16; H, 9.75.
µ-(3,4,5,6-Tetrahexyl-1,2-phenylene)bis(trifluoroacetato-O)dimer-
cury (7). A solution of 1,2,3,4-tetrahexylbenzene (8; 1.1 g, 2.7 mmol)
in CH2Cl2 (30 mL) was stirred at 25 °C and treated with Hg(OOCCF3)2
(2.8 g, 6.6 mmol) and then with CF3COOH (20 mL). The resulting
mixture was stirred at 25 °C for 12 h, and then volatiles were removed
by evaporation under reduced pressure. The residue was dried in vacuo
for 12 h and then extracted with hexane, and the extracts were passed
through a 0.2-µm Millipore filter. Cooling of the filtrate to -78 °C
caused the precipitation of µ-(3,4,5,6-tetrahexyl-1,2-phenylene)bis(tri-
fluoroacetato-O)dimercury (7; 2.3 g, 2.2 mmol, 81%), which was
isolated as an analytically pure colorless solid: mp 149-150 °C; IR
1,2,4,5-Tetrahexylbenzene (10).32 A suspension containing 10%
Pd/C (22 g) and 1,2,4,5-tetra-1-hexynylbenzene (12; 36 g, 90 mmol)
in a mixture of hexane (600 mL) and C2H5OH (600 mL) was stirred at
25 °C under H2 (6.9 atm) in a Parr reactor. After 96 h, the mixture was
filtered, and volatiles were removed from the filtrate by evaporation
under reduced pressure. Kugelrohr distillation (195 °C/0.5 Torr) of the
residue gave 1,2,4,5-tetrahexylbenzene (10; 29 g, 70 mmol, 78%) as
an analytically pure colorless oil: 1H NMR (300 MHz, CDCl3) δ 0.92
3
(t, 12H, J ) 6.9 Hz), 1.30-1.50 (m, 24H), 1.50-1.65 (m, 8H), 2.53
1
(CH2Cl2) 1695 cm-1; H NMR (300 MHz, CDCl3) δ 0.82-0.98 (m,
3
(t, 8H, J ) 7.7 Hz), 6.92 (s, 2H); 13C NMR (75.4 MHz, CDCl3) δ
12H), 1.25-1.65 (m, 32H), 2.50-2.67 (m, 8H); 13C NMR (75.4 MHz,
CDCl3) δ 13.8, 14.0, 22.4, 22.5, 29.5, 30.0, 30.2, 31.3, 31.4 (2C), 33.3,
14.0, 22.6, 29.5, 31.4, 31.7, 32.3, 129.7, 137.6. Anal. Calcd for C30H54
C, 86.88; H, 13.12. Found C, 86.53; H, 12.58.
1
2
38.9, 118.0 (q, JC-F ) 288 Hz), 141.9, 143.8, 152.0, 162.0 (q, JC-F
) 40 Hz); 199Hg NMR (71.6 MHz, CD2Cl2) δ -1535. Anal. Calcd for
C30H52F6Hg2O4 C, 39.27; H, 5.04; F, 10.96. Found C, 38.49; H, 5.15;
F, 10.75.
2,3,4,5-Tetrahexylbenzenesulfonyl Chloride (13). A solution of
1,2,4,5-tetrahexylbenzene (10; 1.0 g, 2.4 mmol) in CHCl3 (7 mL) was
stirred at 25 °C and treated cautiously with neat chlorosulfonic acid
(15 mL). After a vigorous reaction, the mixture was stirred at 25 °C
for 1 h, and then H2O (25 mL) was added slowly. The resulting mixture
was extracted with CHCl3, and volatiles were removed from the
µ-(2,3,5,6-Tetrahexyl-1,4-phenylene)bis(trifluoroacetato-O)dimer-
cury (9). A solution of 1,2,4,5-tetrahexylbenzene (10; 0.12 g, 0.29
mmol) in CH2Cl2 (10 mL) was stirred at 25 °C and treated with
Hg(OOCCF3)2 (0.32 g, 0.75 mmol) and then with CF3COOH (10 mL).
The resulting mixture was stirred at 25 °C for 3 h, and then volatiles
were removed by evaporation under reduced pressure. The residue was
dried at 70 °C in vacuo for 12 h and then extracted with hot hexane,
and the extracts were passed through a 0.2-µm Millipore filter. Cooling
of the filtrate to -78 °C caused the precipitation of µ-(2,3,5,6-tetrahexyl-
1,4-phenylene)bis(trifluoroacetato-O)dimercury (9; 0.16 g, 0.15 mmol,
(28) Kim, Y.; Eom, S. H.; Wang, J.; Lee, D.-S.; Suh, S. W.; Steitz, T.
A. Nature 1995, 376, 612.
(29) Hurst, P.; Takasaki, B. K.; Chin, J. J. Am. Chem. Soc. 1996, 118,
9982. Tsubouchi, A.; Bruice, T. C. J. Am. Chem. Soc. 1995, 117, 7399.
(30) Hartley, F. R. Organomet. Chem. ReV., Sect. A 1970, 6, 119.
(31) Still, W. C.; Kahn, M.; Mitra, A. J. Org. Chem. 1978, 43, 2923.
(32) For a previous synthesis, see: Eapen, K. C.; Dua, S. S.; Tamborski,
C. J. Org. Chem. 1984, 49, 478.

13022 J. Am. Chem. Soc., Vol. 120, No. 50, 1998
Vaugeois and Wuest
52%), which was isolated as an analytically pure colorless solid: mp
mixture was shaken vigorously, and the phases were separated. The
1:1 complex of µ-(3,4,5,6-tetramethyl-1,2-phenylene)bis(trifluoroac-
etato-O)dimercury (5) with N,N-diethylacetamide separated from the
pentane phase as colorless crystals suitable for X-ray crystallographic
105-107 °C; IR (KBr) 1714, 1693 cm-1; 1H NMR (300 MHz, CDCl3)
3
δ 0.91 (t, 12H, J ) 6.9 Hz), 1.28-1.40 (m, 16H), 1.40-1.55 (m,
3
8H), 1.55-1.69 (m, 8H), 2.69 (t, 8H, J ) 7.9 Hz); 13C NMR (100.6
1
3
MHz, CDCl3) d 13.8, 22.4, 29.4, 31.4, 33.4, 39.3, 118.9 (q, JC-F
)
study: 1H NMR (300 MHz, DMSO-d6) δ 0.99 (t, 3H, J ) 7.2 Hz),
288 Hz), 143.5, 151.2, 161.7 (q, JC-F ) 39 Hz); 199Hg NMR (71.6
MHz, CD2Cl2) δ -1366. Anal. Calcd for C30H52F6Hg2O4 C, 39.27; H,
5.04. Found C, 40.17; H, 5.38.
1.08 (t, 3H, J ) 7.2 Hz), 1.95 (s, 3H), 2.17 (s, 6H), 2.37 (s, 6H),
2
3
3.16-3.30 (m, 4H).
X-ray Crystallographic Study of the 1:1 Complex of µ-(3,4,5,6-
Tetramethyl-1,2-phenylene)bis(trifluoroacetato-O)dimercury (5) with
N,N-Diethylacetamide. Crystals of the 1:1 complex of µ-(3,4,5,6-
tetramethyl-1,2-phenylene)bis(trifluoroacetato-O)dimercury (5) with
N,N-diethylacetamide belong to the orthorhombic space group Pbca
with a ) 15.098(5) Å, b ) 17.281(7) Å, c ) 39.52(2) Å, a ) 90°, b
) 90°, g ) 90°, V ) 10311(8) Å3, Dcalcd ) 2.254 g cm-3, and Z ) 8.
Data were collected at 225 K, and the structure was refined to R1 )
0.066, wR2 ) 0.148 for 5295 reflections with I > 2.00 σ(I). A detailed
description of the structure and its determination is provided in the
Supporting Information.
Variable-Temperature IR Studies of the Binding of N,N-
Diethylformamide by µ-(3,4,5,6-Tetrahexyl-1,2-phenylene)bis(tri-
fluoroacetato-O)dimercury (7) and µ-(2,3,5,6-Tetrahexyl-1,4-phe-
nylene)bis(trifluoroacetato-O)dimercury (9). These studies employed
a Perkin-Elmer model 1650 FT-IR spectrometer equipped with a Specac
model 21500-6 variable-temperature cell and a Specac model 20426
temperature controller. All cell windows in contact with solutions of
bidentate Lewis acids 7 and 9 were made of CaF2.33 At each
temperature, values of Ka were determined using multiple nominal
concentrations of DEF and either bidentate Lewis acid 7 or 9. Each
experiment was performed in triplicate. In studies of the association
of 1,2-isomer 7 with DEF, the nominal concentrations of DEF ranged
from 4.5 mM to 18 mM, and the nominal concentrations of compound
7 ranged from 4.8 mM to 9.6 mM. In studies involving the 1,4-isomer
9, the nominal concentrations of DEF varied from 31 mM to 94 mM,
and the nominal concentrations of compound 9 varied from 31 mM to
62 mM. At each temperature, the amount of free DEF was estimated
using calibration curves measured at the same temperature.
A Job plot confirmed that 1,2-isomer 7 forms a 1:1 complex, and
for both isomers the formation of higher-order aggregates is negligible
under the conditions described. At temperatures below -40 °C and at
higher concentrations (19 mM), spectra of solutions containing DEF
and 1,2-isomer 7 revealed a minor peak at 1645 cm-1 in addition to
those at 1633 and 1668 cm-1 associated with doubly coordinated DEF
and free DEF, respectively. It is possible that the additional peak
corresponds to a 1:2 complex in which bidentate Lewis acid 7 doubly
coordinates two molecules of DEF.34 To minimize complications caused
by higher-order association, we avoided temperatures below -40 °C
and concentrations of bidentate Lewis acid 7 above 9.6 mM.
Acknowledgment. We are grateful to the Natural Sciences
and Engineering Research Council of Canada, the Ministe`re de
l′EÄ ducation du Que´bec, and Merck Frosst for financial support.
In addition, acknowledgment is made to the donors of the
Petroleum Research Fund, administered by the American
Chemical Society, for support of this research. We also thank
Professor Samuel H. Gellman and Dr. Gregory P. Dado for
helpful advice about variable-temperature IR spectroscopy. In
addition, we are grateful to Dr. Michel Simard of the Laboratoire
de Diffraction des Rayons-X, De´partement de Chimie, Univer-
site´ de Montre´al, for having determined the structure of the 1:1
complex of µ-(3,4,5,6-tetramethyl-1,2-phenylene)bis(trifluoro-
acetato-O)dimercury (5) with N,N-diethylacetamide.
Supporting Information Available: Full experimental
details for the synthesis and characterization of µ-(3,4,5,6-
tetrakis(tetradecyl)-1,2-phenylene)bis(trifluoroacetato-O)dimer-
cury (16), (2,3,4,5,6-pentahexylphenyl)(trifluoroacetato-O)mer-
cury (17), and their precursors; tables of atomic coordinates and
isotropic thermal parameters, complete bond lengths and angles,
anisotropic thermal parameters, hydrogen atom coordinates and
thermal parameters, and distances to weighted least-squares
planes for the 1:1 complex of µ-(3,4,5,6-tetramethyl-1,2-
phenylene)bis(trifluoroacetato-O)dimercury (5) with N,N-di-
ethylacetamide (29 pages, print/PDF). See any current masthead
page for ordering information and Web access instructions.
1:1 Complex of µ-(3,4,5,6-Tetramethyl-1,2-phenylene)bis(tri-
fluoroacetato-O)dimercury (5) with N,N-Diethylacetamide. µ-(3,4,5,6-
Tetramethyl-1,2-phenylene)bis(trifluoroacetato-O)dimercury (5; 185 mg,
0.244 mmol)35 was dissolved in hot N,N-diethylacetamide (100 mL).
The solution was cooled to 60 °C, pentane (1.0 mL) was added, the
(33) Derivatives of 1,2-phenylenedimercury are known to bind chloride.
Beauchamp, A. L.; Olivier, M. J.; Wuest, J. D.; Zacharie, B. J. Am. Chem.
Soc. 1986, 108, 73.
(34) Complexes in which derivatives of 1,2-phenylenedimercury doubly
coordinate the carbonyl oxygen atoms of two amides at the same time have
been characterized.9,23
(35) Nadeau, F.; Simard, M.; Wuest, J. D. Organometallics 1990, 9, 1311.
JA982725X