Fries rearrangement of aryl formates: A mechanistic study by means of 1H, 2H, and 11B NMR spectroscopy and DFT calculations

Article abstract of DOI:10.1021/jo061475x

(Figure Presented) ¹H, ²H, and ¹¹B NMR spectroscopy has been used to study the mechanism of the Fries rearrangement of aryl formates promoted by boron trichloride by monitoring both the substrate and the Lewis acid. DFT calculations were employed to investigate the energetics of several reaction paths and to calculate NMR chemical shifts of key intermediates and products. After the formation of a 1:1 substrate - Lewis acid adduct, the rearrangement proceeds in two steps, beginning with the cleavage of the ester bond and the release of formyl chloride in situ, which, in turn, acts as a formylating agent, introducing an aldehydic functionality into the aromatic ring. The high regioselectivity (only the ortho product is obtained) is also accounted for by the proposed intermolecular, Lewis acid-assisted mechanism.

Full text of DOI:10.1021/jo061475x

Fries Rearrangement of Aryl Formates: A Mechanistic Study by
1
2
11
Means of H, H, and B NMR Spectroscopy and DFT Calculations
Alessandro Bagno,† Willi Kantlehner,‡ Ralf Kress,‡ Giacomo Saielli,*,§ and
Edmont Stoyanov‡
Dipartimento di Scienze Chimiche, UniVersita` di PadoVa, Via Marzolo, 1 35131 PadoVa, Italy, Institut fu¨r
Organische Chemie der UniVersita¨t Stuttgart, Pfaffenwaldring 55, D-70569 Stuttgart, Germany, and
Istituto per la Tecnologia delle Membrane del CNR, Sezione di PadoVa Via Marzolo, 1 35131 PadoVa,
Italy
giacomo.saielli@unipd.it
ReceiVed July 17, 2006
1H, 2H, and 11B NMR spectroscopy has been used to study the mechanism of the Fries rearrangement of
aryl formates promoted by boron trichloride by monitoring both the substrate and the Lewis acid. DFT
calculations were employed to investigate the energetics of several reaction paths and to calculate NMR
chemical shifts of key intermediates and products. After the formation of a 1:1 substrate-Lewis acid
adduct, the rearrangement proceeds in two steps, beginning with the cleavage of the ester bond and the
release of formyl chloride in situ, which, in turn, acts as a formylating agent, introducing an aldehydic
functionality into the aromatic ring. The high regioselectivity (only the ortho product is obtained) is also
accounted for by the proposed intermolecular, Lewis acid-assisted mechanism.
Introduction
acid,3 methanesulfonic acid/POCl3,4 montmorillonite clays,5
hafnium trifluoromethanesulfonate,6 scandium trifluoromethane-
sulfonate,7 zirconium tetrachloride,8 and titanium tetrachloride9
have all been used to induce the rearrangement. More recently,
heteropoly acids10 and metal triflates, particularly yttrium and
The Fries rearrangement transforms an aryl ester into a
mixture of o- and p-hydroxycarbonyl compounds, the ratio
strongly depending on the reaction conditions, such as temper-
ature, medium, and, most importantly, Lewis acid catalyst used.1
In many cases, complete regioselectivity can be attained with a
careful choice of the catalyst and of its molar ratio with the
substrate.2 Thus, a variety of catalysts such as polyphosphoric
(3) Shargi, H.; Eshghi, E. Bull. Chem. Soc. Jpn. 1993, 66, 135.
(4) Kaboudin, B. Tetrahedron. 1999, 55, 12865.
(5) Venkatachalapathy, C.; Pitchumani, K. Tetrahedron 1997, 53, 17171.
(6) Kobayashi, S.; Moriwaki, M.; Hachiya, I. Tetrahedron Lett. 1996,
37, 2053.
(7) Kobayashi, S.; Moriwaki, M.; Hachiya, I. J. Chem. Soc., Chem.
Commun. 1995, 1527.
(8) Horrowven, D. C.; Dainty, R. R. Tetrahedron Lett. 1996, 37, 7659.
(9) Martin, R.; Demerseman, P. Synthesis 1989, 25.
(10) Kozhevnikova, E. F.; Derouane, E. G.; Kozhevnikov, I. V. Chem.
Commun. 2002, 1178; Kozhevnikova, E. F.; Quartararo, J.; Kozhevnikov,
I. V. App. Catal. A: Gen. 2003, 245, 69.
† Universita` di Padova.
‡ Universita¨t Stuttgart.
§ ITM-CNR, Sezione di Padova.
(1) March, J. AdVanced Organic Chemistry, 3rd ed., Wiley: New York,
1985; p 499.
(2) Martin, R. Org. Prep. Proc. Int. 1992, 24, 369.
10.1021/jo061475x CCC: $33.50 © 2006 American Chemical Society
Published on Web 11/15/2006
J. Org. Chem. 2006, 71, 9331-9340
9331

Bagno et al.
SCHEME 1. Fries Rearrangement of 3-Methoxyphenyl
Formate
SCHEME 2. 3-Methoxyphenyl Formate (1) and Its Possible
Coordination Modes with BCl3 (1a, 1b, and 1c)a
copper(II), together with methanesulfonic acid,11 have been
found to be efficient catalysts. Novel reaction methodologies
have also been applied; Fries rearrangements have also been
performed under microwave irradiation12 or in ionic liquids.13
Despite the wide use of the Fries rearrangement in organic
synthesis, also for preparation of intermediates in manufacturing
fine and specialty chemicals as well as pharmaceuticals,14 the
details of the mechanism are still unclear. In some cases an
intermolecular mechanism has been proposed;15 in others, an
intramolecular one has been put forward, at least for the ortho
isomer;16,17 for polymeric substrates, both mechanisms have been
suggested.18 In some cases, a mechanism that can be still
regarded as intramolecular has been proposed also for the
formation of the para isomer.19 A mechanistic investigation of
the rearrangement of phenyl benzoate promoted by AlBr3 was
reported by Hart and coworkers,20 where an active participation
a The resonance structure with a formal charge separation in 1a accounts
for the deshielding of the formyl carbon and proton upon complexation.
-
on the reaction mixture at the endpoint were not conclusive
concerning the nature of the postulated intermediates and
products.23 In this work we report on a systematic NMR study,
at various temperatures, of the reaction mechanism by monitor-
ing the substrate transformation via 1H and 2H NMR and through
11B NMR spectroscopy from the standpoint of the Lewis acid.
DFT calculations have identified a viable route to the final
product, and calculated 1H and 11B chemical shifts of substrate,
intermediates, and final products agree very well with the
observed resonances.
of AlBr4 was observed. Such a variety of proposals partly
reflects also the large number of substrates, media, and Lewis
acids used. Nevertheless, some general mechanistic features
appear to be widely accepted and they involve the initial
formation of a 1:1 complex between substrate and Lewis acid
(but in some cases a 1:2 complex has been invoked21) which
then undergoes a dissociation with formation of an acylium ion;
this finally acts as the acylating agent.2
Recently, an extension of the Fries rearrangement to aryl
formates has been presented,22 which allows the introduction
of an aldehydic functionality into an aromatic ring (see Scheme
1).
Results and Discussion
A preliminary mechanistic study of the model systems phenyl
formate and 3-methoxyphenyl formate has been already re-
ported.23 DFT calculations of reaction paths of phenyl formate
allowed us to exclude some possible routes, like a direct
intramolecular rearrangement, and suggested that the formation
of formyl chloride as an intermediate is not prohibited on the
basis of the calculated energy barrier. On the other hand, 11B
NMR spectra recorded at low-temperature just after mixing and
Structure of the Initial Formate Ester-BCl3 Adduct. As
a first step we have investigated the nature of the initial adduct
formed when an excess (ca. 20%) of BCl3 is added to a solution
of 3-methoxyphenyl formate (see Scheme 2).
Three different coordination modes of the Lewis acid to the
formyl group can be conceived according to whether the
electron-deficient boron atom interacts with the acyl oxygen
(O-acyl, 1a), with the aryl ester oxygen (O-aryl, 1b), or with
the methoxy oxygen (O-methoxy, 1c). The optimized geometry,
for the three complexes, features a boron-oxygen distance and
dihedral B-Cl-Cl-Cl angle of 1.591 Å and 29.0° for 1a, 3.054
Å and 1.7° for 1b, and 1.646 Å and 27.9° for 1c. The
geometrical parameters indicate that a relatively strong interac-
tion occurs for complexes 1a and 1c. The interaction energies,
without BSSE correction, at the DFT level, are -7.1, -1.7,
and -0.9 kcal/mol, for 1a, 1b, and 1c, respectively. We also
note that the inclusion of the BSSE correction dramatically
changes the results (see Table S2 of Supporting Information)
still favoring complex 1a over 1c, while 1b appears very little
stabilized compared to free substrate and BCl3. Indeed, since
the geometry of BCl3 in the complexes 1a and 1c changes
significantly with respect to the free molecule, most of the
stabilization obtained including the BSSE correction is a
spurious outcome of the counterpoise method, while in the case
(11) Mouhtady, O.; Gaspard-Iloughmane, H.; Roques, N.; Le Roux, C.
Tetrahedron Lett. 2003, 44, 6379.
(12) Moghaddam, F. M.; Ghaffarzadeh, M.; Abdi-Oskoui, S. H. J. Chem.
Res. Synth. 1999, 574.
(13) Harjani, J. R.; Narja, S. J.; Salunkhe, M. Tetrahedron Lett. 2001,
42, 1779.
(14) Metivier, P. In Fine Chemicals through Heterogeneous Catalysis;
Sheldon, R. A., van Bekkum, H., Eds.; Wiley-Interscience: New York, 2001;
p 161.
(15) (a) Krausz, F.; Martin, R. Bull. Soc. Chim. Fr. 1965, 2192. (b)
Martin, R. Bull. Soc. Chim. Fr. 1974, 983. (c) Martin, R. Bull. Soc. Chim.
Fr. 1979, II-373.
(16) Ogata, Y.; Tabuchi, H. Tetrahedron 1964, 20, 1661.
(17) (a) Yamamoto, J.; Kisida, M.; Takenaka, Y.; Okamoto, Y. Nippon
Kagaku Kaishi 1988, 288. (b)Yamamoto, J.; Ishikawa, T.; Okamoto, Y.
Nippon Kagaku Kaishi 1989, 1870. (c) Isota, Y.; Ohkubo, N.; Takaoka,
M.; Yamamoto, J. Nippon Kagaku Kaishi 1999, 421.
(18) Warshawsky, A.; Kalir, R.; Patchornik, A. J. Am. Chem. Soc. 1978,
100, 4544.
(19) Oliveira, A. M. A. G.; Oliveira-Compas, A. M. F.; Raposo, M. M.
M.; Griffiths, J.; Machado, A. E. H. Tetrahedron 2004, 60, 6145.
(20) (a) Dawson, J. L.; Gibson, J. L.; Hart, L. S.; Waddington, C. R. J.
Chem. Soc. Perkin Trans. 1989, 2, 2133. (b) Gibson, J. L.; Hart, L. S. J.
Chem. Soc. Perkin Trans. 1991, 2, 1343.
(22) Ziegler, G.; Haug, E.; Frey, W.; Kantlehner, W. Z. Naturforsch. B
2001, 56, 1178.
(21) Cullinane, N. M.; Woolhouse, R. A.; Edwards, B. F. R. J. Chem.
Soc. 1961, 3842.
(23) Bagno, A.; Kantlehner, W.; Kress, R.; Saielli, G. Z. Naturforsch. B
2004, 59, 386.
9332 J. Org. Chem., Vol. 71, No. 25, 2006

Fries Rearrangement of Aryl Formates
of the 1b complex the BSSE correction appears more justified.
Thermal corrections (see the QM Calculations part of the
Experimental Section) amount to about 14 kcal/mol of desta-
bilization for 1a and 1c and about 8 kcal/mol of destabilization
for 1b. On the other hand, single-point calculations at the MP2
level, using the DFT-optimized geometries, provide a signifi-
cantly different picture: the electronic energy is largely in favor
of the complexes formation, particularly 1c, likely because of
dispersive interactions between chlorine atoms of BCl3 and the
aromatic ring. These data indicate that the energetics of this
complexation is very sensitive to the method used and would
be properly addressed only by recourse to very high levels of
theory for optimization and frequency calculation. However,
this is not strictly necessary for the purposes of this work,
because conclusive information can be obtained by analyzing
trends in NMR spectra, as follows.
Upon addition of BCl3 to 1, the only significant change in
the proton resonances is a large deshielding of the formyl proton
from 8.31 ppm of pure 1 (at room temperature) to 8.87 ppm at
room temperature and up to 9.15 ppm at -5 °C. The other
resonances are also deshielded, compared to the pure substrate,
but only by about 0.05-0.1 ppm, depending on the temperature
(note that a systematic small shift of all resonances is observed
simply as a solvent effect when the solution of BCl3 in heptane
is added to the solution of substrate in CDCl3). The 1H spectrum
at three different temperatures after the addition of BCl3 is
shown in Figure S1 of the Supporting Information (except the
methoxy signal at 3.85 ppm). The aromatic pattern of the adduct
is the same as that of 1, featuring, for increasing shielding, the
triplet of proton H5, two doublets (H4 and H6), and the singlet
of H2. The large changes that the formyl proton resonance
undergoes upon addition of BCl3, and its temperature depen-
dence, indicate the formation of a complex involving the formyl
moiety that can be attributed to the coordination of the Lewis
acid to the acyl oxygen. The large deshielding of the formyl
proton is consistent with a reduced electron density on the
formyl group in 1a (Scheme 2).
Strong support to this proposal is provided by DFT calcula-
tions of the relevant chemical shifts.24 The calculated results
for the free substrate are 8.36 ppm (formyl proton) and 3.77
ppm (methoxy protons), quite in good agreement with experi-
ment. For the O-acyl complex 1a the resonance of the formyl
proton is calculated to be largely deshielded (∆δ ) 1.02 ppm),
as observed experimentally, while the methoxy resonance is
deshielded by only 0.06 ppm, also in very good agreement with
experiments. In contrast, calculated chemical shift variations
for 1b and 1c are totally in disagreement with experiments: for
1b the formyl proton is slightly shielded (∆δ ) -0.05 ppm)
and the methoxy protons are essentially unchanged (∆δ ) -0.01
ppm), while for 1c the formyl proton is again slightly shielded
(∆δ ) -0.03 ppm) and the methoxy protons significantly
deshielded (∆δ ) 0.62 ppm).
A significant deshielding is also observed for the resonance
of the formyl carbon (see Figures S3-S5 of Supporting
Information), which is shifted from 159.2 ppm in the pure
substrate to 173.6 ppm after the Lewis acid is added at -10 °C
(∆δ ) 14.4 ppm), while the methoxy carbon changes only by
0.3 ppm (again a small solvent effect is expected). The
corresponding calculated chemical shift variations are the
following, for the formyl and methoxy carbons respectively:
18.7 and 0.6 ppm for 1a, -1.6 and 0.1 ppm for 1b, and -1.5
and 13.7 ppm for 1c. As with the proton resonances, the
calculated results for 1b and 1c are in complete disagreement
with the experiment.
The 11B spectrum of the initial adduct is also in agreement
with O-acyl complexation. After mixing with 1, the resonance
of BCl3 in heptane (46.5 ppm) disappears from the spectrum,
which now consists of a single resonance at 27.3 ppm at 13 °C
(25.6 ppm at -5 °C) (Figure S2). Since BCl3 is in 20% excess,
the presence of a single signal indicates that free BCl3 and the
initial adduct are in the fast exchange regime. A comparison
with calculated 11B chemical shifts is again useful: for BCl3,
δ(11B) ) 52.8 ppm (the calculated value is too deshielded,
compared to the experimental value, mainly because of the
neglect of spin-orbit coupling23). The calculated 11B chemical
shifts of the initial adduct are, respectively, 18.8, 23.7, and 50.3
ppm for the O-acyl, O-methoxy, and O-aryl complex; the latter
result is very close to that of free BCl3, as expected. Therefore,
the experimental value is somewhat higher than the expected
chemical shift of the O-acyl complex. This is consistent with
the rapid equilibrium of BCl3 between the bound form and the
free Lewis acid.
Thus, all available information consistently indicates that the
substrate is converted to the O-acyl complex 1a and that the
latter is in fast equilibrium with excess BCl3.
Kinetics of the Fries Rearrangement. The kinetics has been
qualitatively studied by means of 1H, 2H and 11B NMR
spectroscopy at several temperatures. In Figure 1a-c we report
1
the time evolution at +7 °C of the H spectra, subdivided for
clarity in three regions (aldehydic, methoxy, aromatic).
In the 1H NMR spectra of Figure 1, three sets of signals can
be clearly identified: those of the substrate which are decreasing
to zero; those of a final product, which are increasing; and those
of an intermediate, reaching a maximum concentration after
about 4 h (at 7 °C). In ref 23 it was postulated that the Fries
rearrangement of aryl formates promoted by BCl3 proceeds
through the formation of formyl chloride as an intermediate,
followed by acylation of the aromatic ring in the ortho position.
The transient signal that appears at 9.70 ppm in the 1H NMR
spectra (Figure 1a) can indeed be attributed to formyl chloride.
In fact, a DFT calculation of the proton chemical shift of formyl
chloride gives a value of 9.82 ppm, in very close agreement
with experiment. However, the existence of the proposed
intermediate needs to be further supported.
Formyl chloride has been isolated and characterized at low
temperature.25 It has been found to decompose into CO and
HCl26 and its lifetime has been found to range from tens or
hundreds of minutes in the gas phase27 to less than a millisecond
in water.26 Recent ab initio calculations suggest a strong catalytic
effect of water on the decomposition reaction and indicate that
at least four water molecules are necessary to increase the rate
of decomposition in water with respect to the gas phase.28 The
(25) (a) Staab, H. A.; Datta, A. P. Angew. Chem., Int. Ed. 1964, 3, 132.
(b) Devos, A.; Remion, J.; Frisque-Hesbain, A. M.; Colens, A.; Ghosez, L.
J. Chem. Soc. Chem. Commun. 1979, 1180. (c) Villeneuve, G. B.; Chan,
T. H. Tetrahedron Lett. 1997, 38, 6489.
(26) Dowideit, P.; Mertens, R.; von Sonntag, C. J. Am. Chem. Soc. 1996,
118, 11288.
(27) (a) Hisatsune, C.; Heicklen, J. Can. J. Spectrosc. 1973, 18, 77-81.
(b) Libuda, H. G.; Zabel, F.; Fink, E. H.; Becker, K. H. J. Phys. Chem.
1990, 94, 5860.
(28) Phillips, D. L.; Zhao, C.; Wang, D. J. Phys. Chem. A. 2005, 109,
9653.
(24) (a) Bagno, A. Chem. Eur. J. 2001, 7, 1652. (b) Bagno, A.; Rastrelli,
F.; Saielli, G. J. Phys. Chem. A. 2003, 107, 9964. (c) Bagno, A.; Rastrelli,
F.; Saielli, G. Chem. Eur. J. 2006, 12, 5514.
J. Org. Chem, Vol. 71, No. 25, 2006 9333

Bagno et al.
1
FIGURE 1. Time evolution over 17 h of the H NMR spectra of the 1-BCl3 reacting system at 7 °C. Top to bottom: (a) aldehydic region, (b)
methoxy region, and (c) aromatic region. The 32 spectra, from front to rear, are separated by 15 min (1-9), 30 min (10-26), and 60 min (27-32).
lifetime of the intermediate in our reaction mixtures is longer
than the acquisition time, which is of the order of 1 s, which
seems compatible with the above findings, since only traces of
water are present in solution, if any.
To further support our hypothesis we ran the reaction using
BBr3 as Lewis acid. Assuming that the reaction proceeds by
the same mechanism,22 we should now observe formyl bromide
as intermediate, whose proton resonance is predicted to be 10.72
ppm by DFT calculation. Indeed, the signal of the intermediate
is observed at 10.70 ppm when BBr3 is used. The “heavy-atom
effect” (due to spin-orbit coupling, which is responsible for
an additional contribution to the shielding constant, σSO) has
been calculated to be -0.17 ppm at the BP86/TZ2P level using
the software ADF;29 therefore, it does not alter our conclusion.
However, BBr3 is known to behave as a dealkylating agent;30
therefore, secondary products are observed at the end of the
reaction, together with the expected product (4-methoxysalicy-
laldehyde). We did not further investigate this aspect and we
proceeded only with BCl3 as Lewis acid.
(29) te Velde, G.; Bickelhaupt, F. M.; Baerends, E. J.; Fonseca Guerra,
C.; van Gisbergen, S. J. A.; Snijders, J. G.; Ziegler, T. J. Comput. Chem.
2001, 22, 931.
(30) (a) Brindaban, R.; Bhar, S. Org. Prep. Proc. Int. 1996, 28, 371. (b)
Chem. Abstr. 1996, 125, 8934. (c) Williard, P. W.; Fryhle, C. B. Tetrahedron
Lett. 1980, 21, 3771.(d) Press, J. B. Synth. Commun. 1979, 9, 407-440.
(e) Bonher, F. G.; Boume, E. J.; Mc Nally, S. J. Chem. Soc. 1960, 2929.
9334 J. Org. Chem., Vol. 71, No. 25, 2006

Fries Rearrangement of Aryl Formates
FIGURE 2. (top) 2H NMR spectrum of 1-d reaction mixture and
FIGURE 3. Time evolution over 17.5 h of the 11B NMR spectra of
the 1-BCl3 reacting system at 13 °C. The 29 spectra, from front to
rear, are separated by 15 min (1-9), 30 min (10-22), 60 min (23-
29).
1
(bottom) H NMR spectrum of 1 reaction mixture after 4 h at 7 °C.
2
The H spectrum has been shifted so that the formyl signal is exactly
1
superimposed on that of the H spectrum.
The dependence of the formyl resonance of the intermediate
on the halide of the Lewis acid also allows us to discard the
hypothesis of the intermediate being the formylium cation,
HCO+, for which the calculation predicts a significantly shielded
value of the proton resonance (7.28 ppm). Moreover, HCO+
has been the most elusive of the acylium cations: it has been
observed in condensed phase only in the presence of superacids,
like HF-SbF5, under high CO pressure.31 Its high instability,
compared to other substituted acylium cations, R-CO+, is
related with the easy loss of a proton to give the neutral CO
molecule.32 The high instability also means a high reactivity,
which is inconsistent with the observed regioselectivity of the
Fries rearrangement.
a more complicated time evolution: the decrease of intensity
is accompanied by a shift to higher values until, at the end of
the reaction, we observe again the signal of free BCl3 at 46.5
ppm. This behavior may depend on the fact that the initial signal
does not correspond to a single species but is the weighted
average of free and complexed BCl3. In contrast, the final
product does not seem to be involved in an exchange equilib-
rium. Other resonances, at 31.8 and 26.4 ppm, also appear in
the final spectrum.
We have run the reaction also in the presence of an equimolar
quantity of toluene to test whether tolualdehyde is formed and
we have found no evidence, at least in the 1H NMR, of
tolualdehydes. We note, however, that this result is not
conclusive regarding the intra/intermolecularity of the mecha-
nism: in fact, the para position of the substrate 1 is certainly
more reactive than any position of toluene. Nevertheless, the Fries
rearrangement only produces the ortho isomer. As we will see
in the computational section, such regioselectivity is nicely
accounted for by the intermolecular mechanism. Likewise, the
fact the toluene is not acylated can be explained. This observa-
tion supports the hypothesis that the acylating agent is a rather
weak electrophile, like formyl chloride, rather that a strong one,
like the formylium cation.
As a final test we also studied the kinetics at 7 °C of a sample
2
of 1 deuterated at the formyl position, 1-d, using H NMR. In
the deuterium spectra we could only observe, at the beginning,
the signal of the formyl deuterium of the reactant, and, as the
reaction proceeded, the signals of the intermediate and of the
final product as observed in the aldehydic region of the 1H NMR
spectra of the nondeuterated sample in Figure 1a. In Figure 2
1
2
we show the superposition of the H and H NMR spectra, of
the 1 and 1-d reaction mixtures, after 4 h, as an example (see
Figure S6 for the full kinetics). This clearly demonstrates that
the intermediate at about 9.7 ppm is indeed originating from
the formyl group of the reactant (9 ppm) and is transformed
into the aldehydic group of the product (8.7 ppm) as discussed
further below.
Structure of the Final Product in the Reaction Mixture.
We now analyze the nature of the final products that are formed
in the reaction. In Figure 4 we show the 1H spectrum after the
reaction has reached its endpoint at room temperature overnight,
without any workup.
With regard to the time evolution of the spectra of the
methoxy region, the signals of the substrate completely disappear
after 20 h; thus, e.g., the methoxy group at 3.86 ppm in Figure
1b is replaced by the signal of the main product at 4.02 ppm,
together with a secondary product whose methoxy protons
resonate at 3.81 ppm. The same considerations apply to the
aromatic region (Figure 1c). The integrated intensity of the
secondary product amounts to about 15% of the main product
at 7 °C, but it is lower at lower temperatures. We will come
back to this point later.
We note the presence of an aldehyde signal at 8.70 ppm,
rather broad and poorly resolved in a doublet; its peculiar shape
will be discussed later. The aromatic pattern is typical of a 1,2,4-
trisubstituted benzene ring, which is consistent with the expected
4-methoxysalicylaldehyde. As already noted, there is also a
methoxy resonance at 4.02 ppm. A secondary product is
1
observed in the H spectrum, with the characteristic pattern of
a 1,3-disubstituted benzene ring: the H4 triplet, overlapped with
the solvent resonance at 7.26 ppm; two doublets (6.77 and 6.62
ppm); and an unresolved triplet at 6.58 ppm. These signals are
marked with an asterisk in Figure 4. As already mentioned, the
methoxy protons of this secondary product resonate at 3.81 ppm.
To confirm that the main product observed at the end of the
reaction is the expected 4-methoxysalicylaldehyde, possibly in
In the 11B spectra of Figure 3 we clearly see the growth of a
signal at 8.3 ppm, while the initial resonance at 27.3 ppm has
(31) de Rege, P. J. F.; Gladysz, J. A.; Horva´th, I. T. Science 1997, 276,
776.
(32) Sorensen, T. S. Angew. Chem., Int. Ed. 1998, 37, 603.
J. Org. Chem, Vol. 71, No. 25, 2006 9335

Bagno et al.
SCHEME 3. The Three Main Steps of the Fries
Rearrangement as Indicated by NMR Spectroscopy
1
FIGURE 4. Final H NMR spectrum of the reaction solution after 1
day at room temperature. The resonances of a secondary product are
marked with an asterisk. The signal at 7.26 ppm is the residual solvent
overlapped with a triplet of the secondary product. The aldehyde
resonance is enlarged in the inset.
HCl and CO. Therefore, a corresponding amount of 2 cannot
be acylated and remains in the reaction mixture.
The proposed steps are consistent with the above results and
account for the observed products. However, this scheme leaves
some major points unanswered. First of all, of course, the
intermolecular mechanism proposed does not provide a cogent
explanation for the high regioselectivity of the reaction. Second,
the changes in the 11B spectra need to be accounted for.
As far as the 11B spectrum is concerned, calculated 11B
chemical shifts of chloroborate esters B(OR)3-nCln (n ) 0, 1,
2)23 showed that dichloroborate monoesters (B(OR)Cl2), chlo-
roborate diesters (B(OR)2Cl), and borate triesters (B(OR)3) have
11B chemical shifts of about 30, 20, and 10 ppm, respectively,
almost regardless of the organic substituent R. Considering the
systematic error of +2-3 ppm due to the neglect of spin-orbit
coupling, the resonance at 31.8 ppm can then be assigned to
the intermediate 3-methoxyphenyl dichloroborate 2. However,
it is difficult to assign the main product, at 8.3 ppm, to the
similar boron monoester of the rearranged product, 3. In fact,
in ref 23 we suggested the possibility that the main product
was the borate triester of the rearranged product. However, this
hypothesis does not easily fit the mechanism proposed: in
particular, there is no evidence in the 1H and 11B spectra of the
presence of mono- and diesters of the rearranged product, which
are expected to be precursors of the proposed triester.
The three main steps reported in Scheme 3 will be analyzed
in more detail in the next section with the aid of DFT
calculations; we will then show that the 11B spectrum can be
easily assigned to the species reported in Scheme 3 and that
the proposed mechanism can also explain the high regioselec-
tivity.
the form of a borate ester, we have measured the 1H spectra of
a solution of 4-methoxysalicylaldehyde after the addition of a
20% excess of BCl3. After allowing the system to react at room
temperature, we obtained a spectrum identical to that of Figure
4, except for the absence of the secondary product. The
secondary product has been identified as a borate ester of
3-methoxyphenol by comparing the proton resonances of the
secondary product of Figure 4 with those of a solution of
3-methoxyphenol after the addition of a 20% excess of BCl3,
which showed the same proton resonances observed for the
secondary product. A close inspection of the 1H NMR spectrum
also reveals the presence of a low intensity (less than 3% of
the total) triplet at 7.86 ppm, a doublet at 6.45 ppm (both visible
in Figure 4), a singlet at 4.09 ppm (not shown), and a broad
signal at 9.30 ppm. These resonances, together with the weak
signal in the 11B spectrum at 26.4 ppm, have not been assigned.
They are, however, compatible with the other ortho product,
2-hydroxy-6-methoxybenzaldehyde. We believe that the steric
bulk of the methoxy group renders the rearrangement of the
formyl group to position 2 of the substrate 1 largely disfavored
compared to position 6.
We remark at this point that only the ortho isomer is obtained;
1
no traces, at least detectable in the H spectrum, of the para
isomer are found. This is a sensitive point that of course has a
bearing on all mechanistic issues and will be further dealt with
below.
Proposed Main Steps of the Fries Rearrangement. On the
basis of this evidence, we propose the following three steps for
the Fries rearrangement of 3-methoxyphenyl formate, reported
in Scheme 3: (a) first the substrate is predominantly converted
to the complex at the acyl oxygen, 1a; (b) 1a then reacts to
form HCOCl and a borate ester of 3-methoxyphenol (3-
methoxyphenyl dichloroborate 2), which is finally (c) formylated
at the ortho position to give the borate ester of the rearranged
product 4-methoxysalicylaldehyde as its dichloroborate ester 3.
We note that while the signal of the intermediate HCOCl
actually decreases to zero at the end of the reaction, those of
the other intermediate (2) persist in the mixture. This is probably
due to the fact that during the course of the reaction a small
amount (ca. 15% at +7 °C) of formyl chloride decomposes to
DFT Calculations of the Reaction Mechanism. In ref 23
we investigated several pathways for the Fries rearrangement
of a model aryl formate (phenyl formate). Here we will discuss
in more detail some of them for the system investigated
experimentally, 3-methoxyphenyl formate 1. They are presented
in Schemes 4 and 5. We can safely assume that O-acyl
complexation yielding 1a is not the rate-determining step;
therefore, we turn our attention to the second and third steps in
Scheme 3.
The second step of Scheme 3 is detailed in Scheme 4: we
assume a pre-equilibrium where the activated carbonyl carbon
is attacked by a chloride anion lost by the Lewis acid
coordinated to the oxygen, yielding the chloroalkyl dichlorobo-
rate 4. We have not attempted a computational investigation of
9336 J. Org. Chem., Vol. 71, No. 25, 2006

Fries Rearrangement of Aryl Formates
TABLE 1. Calculated Enthalpies and Free Energies in kcal/mol
for the Reaction Path Shown in Figure 7a
SCHEME 4. Details of the Second Step of the Fries
Rearrangement Shown in Scheme 3
species
∆Hcorr, DFT
∆Gcorr, DFT
MP2
1a
4
0.0
2.8
0.0
3.5
0.0
4.1
TS42a
2 + HCOCl
TS23
25.7
4.4
17.8
-17.8
27.7
-8.4
20.6
22.2
-9.6b
16.2
3 + HCl
-25.6
-22.0b
a DFT, MPW1K/6-311+G**//MPW1K/6-311+G**; MP2, MP2/6-
311+G**//MPW1K/6-311+G**. ZPE and thermal corrections refer to the
gas phase at 298 K and 1 atm and are calculated at the DFT level. b For
the bimolecular states 2 + HCOCl and 3 + HCl, corrections to free energies
in order to change the standard state from gas at 1 atm to concentrations
units and accounting for the condensed phase amount to 3.1 kcal/mol, thus
reducing the reaction barrier of TS23.34
SCHEME 5. Details of the Third Step of the Fries
Rearrangement Shown in Scheme 3
FIGURE 5. The transition states for the boron migration (second step
of Schemes 3 and 4) and ArO-COHCl bond cleavage: left, TS42a;
right, an additional BCl3 molecule is involved (TS42b).
important to note that if we calculate the electronic energy at
the MP2 level of theory (using the DFT-optimized geometries
with ZPE and thermal corrections) the calculated enthalpies and
free energies are noticeably lowered, as reported in Tables 1
and S3. This approach may be questionable, since a different
theoretical level may alter not only the energetics but also the
nature of a sensitive point on the potential energy surface, like
a transition state, and therefore it is not free from pitfalls (for
example, it overestimates the stabilization of transition state
TS42b). Nonetheless, it gives an indication that a proper account
of electron correlation, also in optimizing the geometries and
in calculating vibrational frequencies, would be needed for an
accurate quantitative result. This is, however, outside the scope
of the present work.
Then we have investigated the third step of Scheme 3 in more
detail, considering the acylation both in the ortho and para
positions (Scheme 5). We have found that the transition state
for the ortho substitution (TS23) is relatively low in energy
because of a strong stabilizing interaction of the oxygen of
formyl chloride with the boron atom: the activation enthalpy
is 13.4 kcal/mol and the Gibbs free energy 29.0 kcal/mol; the
latter is again reduced to 25.8 kcal/mol using MP2 electronic
energies and further reduced to 22.7 kcal/mol including the
corrections for changing the reference state from the ideal gas
at 1 atm to concentration units and including the correction for
a condensed phase.34 Such corrections do not alter the energy
of the other stationary points, which are unimolecular. In the
transition state (TS23, Figure 6), the distance between the boron
this step (1a f 4), since it may be intermolecular, through the
participation of HCl present in the system. We remark that
reacting species 4 in Scheme 4 is not detected in the NMR
spectrum. We can even exclude that 4 is the one observed at
1
the beginning of the reaction in the H spectrum, since the
1
calculated H chemical shift of the ArOCHOCl proton is 7.32
ppm, significantly more shielded than the resonance observed
at about 9 ppm. Even so, its intermediacy is a conceivable and
logical step in the reaction sequence. Moreover, a very similar
nucleophilic attack of chloride anion to a carbonyl of a
benzaldehyde has been found to be promoted by BCl3 in hexane,
leading to the formation of an intermediate strictly related to
4.33
Subsequently the boron atom migrates from the acyl oxygen
to the aryl oxygen together with cleavage of the ArO-C bond
(4) and formation of HCOCl and 3-methoxyphenyl dichlorobo-
rate 2 as intermediates; the involved transition states are labeled
TS42a,b in Scheme 4. The calculations (see Table 1 and Table
S3 of Supporting Information) suggest that this pathway is viable
for the first step of the rearrangement. The activation enthalpy
for TS42a, at the DFT level, is 25.7 kcal/mol and decreases to
12.7 kcal/mol if we include an additional BCl3 molecule
assisting the migration of boron (TS42b). The calculated
activation Gibbs free energies are, instead, 27.7 and 28.0 kcal/
mol, for TS42a and TS42b, respectively. The geometry of the
two transition states is reported in Figure 5. However, it is
(34) (a) Benson, S. W. Thermochemical Kinetics; J. Wiley & Sons: New
York, 1976; p 9. (b) Benson, S. W. The Foundation of Chemical Kinetics;
McGraw Hill: New York, 1960; p 507.
(33) (a) Kabalka, G. W.; Wu, Z. Tetrahedron Lett. 2000, 41, 579. (b)
Kabalka, G. W.; Wu, Z.; Ju, Y. J. Organomet. Chem. 2003, 680, 12.
J. Org. Chem, Vol. 71, No. 25, 2006 9337

Bagno et al.
FIGURE 6. Transition state TS23 (left), Wheland intermediate 5 (middle), and rearranged product 3 + HCl (right) for the ortho acylation.
in quite good agreement with the experimental result of 8.3 ppm,
considering the neglect of spin-orbit effects. We note that the
11B chemical shift of this compound is in the same range of
borate triesters that we investigated in our previous work.23
Nevertheless, the tetracoordination of boron, with salicylalde-
hyde acting as a chelating agent, causes a relatively strong
shielding of the boron nucleus.
Inspection of the aldehyde resonance in Figure 4 reveals that
the proton signal appears as a broad doublet with a separation
of about 6 Hz and clearly exhibiting side shoulders, for a total
line width of about 15 Hz. Since no 1H-1H correlation involving
the resonance at 8.70 ppm has been observed in the COSY
spectrum, this splitting can only arise from coupling to boron,
whose 11B signal is fairly narrow. A DFT calculation on 3
predicts a 1H-11B coupling constant (Fermi contact term only)35
of about 5 Hz, in reasonable agreement with the observed
splitting. However, since 11B has I ) 3/2, one would expect a
1:1:1:1 quartet broadened by the quadrupolar 11B. The discrep-
ancy can be reconciled by recalling that 11B has a natural
abundance of 80%, the remainder being 10B (I ) 3) with a
magnetogyric ratio 1/3 of that of 11B. Therefore, a proton
coupled to 10B is split into a 1:1:1:1:1:1:1 septet with a
separation of ca. 1.6-2 Hz, possibly with a small isotope shift.
A simulation (see Figure S9) reveals that the peculiar shape of
the resonance at 8.70 ppm indeed derives from the superposition
of a poorly resolved quartet and septet with the given coupling
constants, further confirming the structure of the raw product
as 3. We recall that such large broadening is only observed in
the proton resonance of the final product and not, for example,
in the proton resonance of the intermediate. This is an indication
that in the intermediate the boron atom is not bound to the
formyl oxygen, confirming the hypothesis that such an inter-
mediate is formyl chloride. In contrast, we do not expect any
broadening in 1a because of the fast exchange regime.
FIGURE 7. Energy profile of the Fries rearrangement. Calculated
Gibbs free energies at the DFT level (open squares) and MP2 level
(solid squares) include ZPE and thermal corrections at 298 K, gas phase,
1 atm. See Table S3 of Supporting Information for more details. For
bimolecular states contributions are calculated by adding those of the
two isolated molecules.
atom and the oxygen of formyl chloride is only 1.553 Å, slightly
longer than the bond length between boron and the aryl oxygen
(1.440 Å). In contrast, the C-C bond between the acyl group
and the benzene ring is not yet formed, the distance being 2.192
Å. Such a strong boron-oxygen interaction also stabilizes the
ortho Wheland intermediate 5 and the final borate ester of the
rearranged product 3, as shown in Figure 6.
In contrast, at the level of theory used, the para Wheland
intermediate is not even a minimum on the potential energy
surface. Therefore, the interaction of formyl chloride with the
Lewis acid coordinated to 3-methoxyphenol lowers the activa-
tion barrier, thereby driving the reaction through the ortho
substitution pathway.
Relative Reactivity of Phenyl and 3-Methoxyphenyl For-
mates. Useful information can be obtained also by comparing
the reactivity of phenyl and 3-methoxyphenyl formates. It is
known that the reaction of phenyl formate with BCl3 does not
proceed to the rearranged product.22,23 Generally speaking, this
result is consistent with the activating effect of the 3-methoxy
group with respect to the acylation in ortho position. However,
the first and second step of the Fries rearrangement reported in
Scheme 3 should not be influenced by the presence of a
substituent on the phenyl ring. Therefore, this comparison offers
The calculated activation enthalpies, although affected by the
approximations due to the level of theory used, are in the typical
range for organic reactions occurring at or below room tem-
perature, as our experimental case. Therefore, the paths inves-
tigated are viable routes connecting the intermediates that we
have identified by comparing experimental and computed
chemical shifts. In Figure 7 we summarize the energy profile
of the mechanism proposed; full data are reported in Table S3.
1
We can now comment on the 11B and H chemical shifts of
the final rearranged product shown in Figure 6 (right panel).
The calculated value of the 11B chemical shift of the dichlo-
roborate monoester of 4-methoxysalicylaldehyde is 11.1 ppm,
(35) Onak, T.; Jaballas, J.; Barfield, M. J. Am. Chem. Soc. 1999, 121,
2850.
9338 J. Org. Chem., Vol. 71, No. 25, 2006

Fries Rearrangement of Aryl Formates
Conclusion
We have investigated the mechanism of the Fries rearrange-
ment of aryl formates promoted by the Lewis acid boron
trichloride. H, H, and 11B NMR spectroscopy have given
essential insights on the key steps of the reaction. The reaction
proceeds in two main steps, after the formation of an initial 1:1
substrate-Lewis acid adduct: in the first major step, formyl
chloride is produced in the reaction mixture; the aldehyde
functionality is then introduced in the aromatic ring through an
electrophilic acylation. The mechanism in this case is intermo-
lecular. The ortho position is strongly favored because the
corresponding transition state benefits from a chelating effect
of the two ortho oxygen atoms on the boron atom. Such an
interaction significantly lowers the energy barrier, in contrast
to the case of the para isomer, where it is not effective. DFT
calculations of the energetics of several reaction paths have
confirmed that the proposed mechanism is indeed a viable route
1
2
1
FIGURE 8. (Top) H NMR spectrum of phenyl formate at 273 K
1
just after the addition of BCl3 in 1:1.2 ratio. (Bottom) the same sample
after 1 day at room temperature.
to the ortho rearranged product. DFT calculations of H and
11B chemical shifts agree with the observed resonances in the
NMR spectra.
The general validity of this mechanistic proposal for the Fries
rearrangement at large remains, however, still an open question.
Clearly, in all cases where the para isomer is the major or the
only product, a different scheme has to be invoked. Indeed, even
in the reaction investigated herein, if the rearrangement of aryl
formates is carried out with a protic superacid such as CF3-
SO3H in place of the Lewis acid, the para isomer is also
obtained.22 In this case, several features can be expected to be
different, like the strength of the aryl formate-acid interaction
and the structure of all intermediates; thus, for example, one
would not expect any favorable interaction of the ortho oxygen
atoms with the catalyst. Hence, whereas this reaction probably
possesses a multifaceted course depending on the precise nature
of reactant and catalyst, we hope to have at least clarified most
factors that underlie the final outcome in this instance.
FIGURE 9. 11B spectrum of the equilibrium (after 2 days at room
temperature) reaction mixture of phenyl formate and BCl3.
the opportunity to investigate the early events of the reaction,
without intervention of the rearrangement step.
Experimental Section
Preparation of Samples. The nondeuterated formates have been
prepared according to ref 36. The reaction mixtures were prepared
directly in the NMR tube with a 0.082 M solution of 3-methox-
yphenyl formate in CDCl3 (0.6 mL, 0.05 mol) and adding ca. 0.06
mL of a 1 M solution of BCl3 in heptane. The reaction mixture
was kept in dry ice/acetone during the first few minutes before
inserting the tube into the probehead. Usually the tube was allowed
to equilibrate for a few minutes before starting the acquisition. The
solution turns to a pale orange/yellow color as soon as BCl3 is added
and eventually it becomes dark red as the substrate has completely
reacted.
Preparation of 3-Methoxyphenyl Formate-1-d (1-d). 1-d was
prepared according to a new synthetic procedure that employs
triformamide as formylating reagent, prepared in situ from meth-
anesulfonyl chloride and sodium diformamide.36 This method allows
us to use formic acid (99% D, 95% in D2O) as deuterated starting
material according to the following procedure.
Methyl Formate-1-d (DCOOCH3). A mixture of DCOOD (5.10
g, 95% D, 0.1 mol), calcium chloride (1.20 g), and dry methanol
(30 mL) is heated at 60-65 °C, under distillative removal of the
product methyl formate-1-d through a 25-cm Vigreux column (∼3
h) to yield 3.51 g (57%); bp 31 °C/720 Torr.
Formamide-1-d (DCONH2). Ammonium chloride (10 g, 0.18
mol) is suspended in dry methanol (30 mL). A 30% solution of
sodium methoxide (32.40 g, 0.18 mol) in methanol is added
We thus consider the effect of adding BCl3 (again in the ratio
1:1.2) to phenyl formate. In Figure 8 we report the 1H spectrum
of the overnight equilibrated reaction mixture of phenyl formate
and BCl3 (see Figure S10 for a time evolution of the NMR
spectra). We note that, compared to the spectrum acquired just
after the addition of BCl3, the aldehydic resonance has disap-
peared. According to the second step of Scheme 3, the Lewis
acid should remove the formyl group producing phenyl dichlo-
roborate and formyl chloride. The latter, at room temperature,
is completely decomposed into HCl and CO and therefore it is
not detected in the spectrum. If this is correct, the low-intensity
resonance at 31.8 ppm observed in the final 11B spectrum in
Figure 3, which corresponds to the monoester intermediate,
should now be the only one observed, together with the excess
free BCl3 at 46.5 ppm. This is indeed the case as we can see in
Figure 9.
Therefore, only the first and second steps of Scheme 3 take
place if the activating 3-methoxy group is not present, strongly
supporting the proposed mechanism. Analogous lack of reactiv-
ity is found for 4-methoxyphenyl formate, where the methoxy
group cannot act as activating group toward ortho electrophilic
substitution. In contrast, the Fries rearrangement does proceed
with 3,5-dimethoxyphenyl formate.
(36) Ziegler, G.; Kantlehner, W. Z. Naturforsch. B. 2001, 56, 1172.
J. Org. Chem, Vol. 71, No. 25, 2006 9339

Bagno et al.
dropwise with stirring at 0 °C. After the addition is terminated, the
mixture is stirred for 30 min and then filtered. The methanolic
ammonia solution thus obtained is mixed with methyl formate-1-d
(3.51 g, 57 mmol). After 30 min of stirring at room temperature,
the mixture is heated for 1 h under reflux and then left standing
for 12 h. Methanol is distilled off and the residue is distilled in
vacuum to give 1.90 g (74%) of formamide-1-d; bp 58-60 °C/0.1
Torr.
The combined cyclohexane extracts are evaporated. The residue
(156 mg, 20%) consists of 3-methoxyphenyl formate-1-d with an
88% deuterium content (NMR, see Figure S7).
1
2
NMR Spectroscopy. All H and H NMR measurements have
been run at 400 and 61 MHz, respectively, with a 5-mm multi-
2
nuclear inverse probehead. H measurements were run unlocked,
with lock cables unplugged to prevent RF interference. The magnet
stability was preliminarily checked to be sufficient within the time
frame of the experiments. 11B and 13C NMR measurements have
been run at 96 and 75 MHz, respectively, with a 5-mm multinuclear
probehead. Kinetics measurements have been run in the range of
temperatures between 263 and 298 K over a time period of 20 h
following the time evolution of 1H and 11B spectra. For 2H kinetics
experiments only a single temperature was selected.
Sodium Diformamide-1,1′-d2 [(DCO)2N- Na+]. A solution of
sodium ethoxide in ethanol is prepared from sodium (0.60 g, 26
mmol) and dry ethanol (15 mL). After the addition of formamide-
1-d (1.90 g, 41 mmol) in 10 mL of absolute ethanol, the mixture
is left standing for 18 h at room temperature. The product is filtered
off, washed with absolute ethanol, and dried in vacuum to yield
1.6 g (84%); mp 234-236 °C (dec).
QM Calculations. Density functional theory calculations have
been run using the software Gaussian 03.37 Following the work of
Truhlar and coworkers38 and Gilbert,39 we have used the MPW1K
functional40 with the 6-311+G** basis set for energy calculations
and for geometry optimization of reactants, products, and transitions
states, the latter ones using the QST3 algorithm. The MPW1K
functional has been found to have a better performance than other
functionals, particularly B3LYP, for the calculation of transition
barriers;38 it was also found to give very accurate geometries in
cases where dative bonds with boron are involved.39 Therefore, it
appears to be the method of choice for our system. In addition to
this, we have also calculated single point energies at the MP2/6-
311+G** level for the structures optimized at the DFT level. ZPE
and thermal corrections to free energy are calculated at the DFT
level and referred to the gas phase at 1 atm and 298 K. Transition
states have been checked by frequency calculations and intrinsic
reaction coordinate calculations to confirm that the observed
transition state is characterized by a single negative vibrational
frequency and that the corresponding normal mode connects
reactants and products. NMR properties calculations were carried
out at the B3LYP/cc-pVTZ//B3LYP/6-31G** level as previously
described.23,25 The chemical shift is calculated as δ ) σref - σ,
where σ is the shielding constant of the nucleus investigated and
σref is the shielding constant in TMS.
3-Methoxyphenyl Formate-1-d (1-d). Methanesulfonyl chloride
(0.57 g, 5 mmol) is added at -10 °C dropwise with stirring to a
suspension of sodium diformamide-1,1′-d2 in 30 mL of absolute
acetonitrile (see Scheme S1). The temperature is left to increase to
0 °C within 20-22 h under stirring. At this temperature, 3-meth-
oxyphenol (0.62 g, 5 mmol) is added. The mixture is heated under
reflux for 90 min. After removal of the acetonitrile by distillation
the residue is cooled to 0-2 °C and 30 mL of absolute ether is
added with stirring. After 12 h the ethereal solution is separated.
The ether is removed by distillation and the residue is distilled
through a 15-cm Vigreux column in vacuum to yield 0.46 g (61%)
of 3-methoxyphenyl formate-1-d; bp 41 °C/10-3 Torr. The product
contains small amounts of formamide. To obtain a sample free from
formamide, the product is extracted three times with cyclohexane.
(37) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K.
N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.;
Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.;
Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.;
Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li,
X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.;
Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.;
Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.;
Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich,
S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A.
D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A.
G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.;
Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham,
M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.;
Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian
03, Revision C.02; Gaussian, Inc., Wallingford CT, 2004.
Acknowledgment. We thank Prof. G. Scorrano (Universita`
di Padova) for useful comments. We also acknowledge a referee
and Associate Editor Professor D. A. Singleton for their
insightful suggestions.
Supporting Information Available: NMR spectra of com-
pounds, Cartesian coordinates, and computed energies of optimized
structures. This material is available free of charge via the Internet
at http://pubs.acs.org.
(38) Lynch, B. J.; Fast, P. L.; Harris, M.; Truhlar, D. G. J. Phys. Chem.
A. 2000, 104, 4811.
(39) Gilbert, T. M. J. Phys. Chem. A. 2004, 108, 2550.
(40) The route section of Gaussian03 input files was # mpwpw91 IOp-
(3/76)1000004280) IOp(3/77)0572005720) IOp(3/78)1000010000).
JO061475X
9340 J. Org. Chem., Vol. 71, No. 25, 2006