Intra- and intermolecular C(sp2)-H...O hydrogen bonds in a series of isobenzofuranone derivatives: Manifestation and energetics

Article abstract of DOI:10.1002/ejoc.200800357

Derivatives 2-6 were prepared as models for studying intra- and intermolecular C(sp²)-H...O hydrogen bonding. Their X-ray structures confirm the presence of intramolecular hydrogen bonds in derivatives of the a series: the corresponding C...O distances vary between 2.91 and 2.97 A. The corresponding ¹³C-¹H coupling constants are increased by about 7.5 Hz, and the ¹H chemical shifts in CDCl₃ are 9.1-10.7 ppm. No intramolecular hydrogen bonds can form in derivatives of the isomeric b series. In this series, the chemical shifts of the corresponding aromatic protons exhibit strong solvent dependency; in particular, they are as sensitive as the proton in chloroform to the presence of DMSO. The vinylic protons activated by the electron-accepting COOR groups behave similarly. Quantum mechanical calculations in the gas phase and in DMSO reproduce the experimental observations. Energies of the intramolecular hydrogen bonds evaluated by two independent approaches vary between 3.7 and 4.4 kcal mol^-1 in the gas phase and still amount to at least 2.5 kcal mol^-1 in 2a in DMSO. These estimates are practically independent of the computational method (HF, MP2, and DFT B3LYP were employed for derivatives 2). We conclude that the behavior of both activated aromatic and vinylic C(sp²)-H atoms in the studied derivatives is qualitatively and quantitatively similar to the behavior of the C(sp ³)-H atom in chloroform. The existence of hydrogen bonds involving these atoms can easily be detected by NMR spectroscopy. Wiley-VCH Verlag GmbH & Co. KGaA, 2008.

Full text of DOI:10.1002/ejoc.200800357

FULL PAPER
DOI: 10.1002/ejoc.200800357
Intra- and Intermolecular C(sp2)–H···O Hydrogen Bonds in a Series of
Isobenzofuranone Derivatives: Manifestation and Energetics
Claude Niebel,[a] Vladimir Lokshin,[a] Mark Sigalov,[b] Pnina Krief,[b] and
Vladimir Khodorkovsky*[a]
Keywords: Hydrogen bonds / NMR spectroscopy / Density functional calculations
Derivatives 2–6 were prepared as models for studying intra-
and intermolecular C(sp2)–H···O hydrogen bonding. Their X-
ray structures confirm the presence of intramolecular hydro-
gen bonds in derivatives of the “a” series: the corresponding
C···O distances vary between 2.91 and 2.97 Å. The corre-
sponding 13C–1H coupling constants are increased by about
7.5 Hz, and the 1H chemical shifts in CDCl3 are 9.1–
10.7 ppm. No intramolecular hydrogen bonds can form in de-
rivatives of the isomeric “b” series. In this series, the chemi-
cal shifts of the corresponding aromatic protons exhibit
strong solvent dependency; in particular, they are as sensi-
tive as the proton in chloroform to the presence of DMSO.
The vinylic protons activated by the electron-accepting
COOR groups behave similarly. Quantum mechanical calcu-
lations in the gas phase and in DMSO reproduce the experi-
mental observations. Energies of the intramolecular hydro-
gen bonds evaluated by two independent approaches vary
between 3.7 and 4.4 kcalmol–1 in the gas phase and still
amount to at least 2.5 kcalmol–1 in 2a in DMSO. These esti-
mates are practically independent of the computational
method (HF, MP2, and DFT B3LYP were employed for deriv-
atives 2). We conclude that the behavior of both activated
aromatic and vinylic C(sp2)–H atoms in the studied deriva-
tives is qualitatively and quantitatively similar to the behav-
ior of the C(sp3)–H atom in chloroform. The existence of hy-
drogen bonds involving these atoms can easily be detected
by NMR spectroscopy.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2008)
value is more than 80% of the energy (4.51 kcalmol–1) cal-
culated for the HOH···OH2 dimer at the same level of
theory.[4] Further examples can be found in a recent re-
view.[5]
Introduction
Phenomena relating to hydrogen bonds (HBs) are the fo-
cus of a huge number of studies, owing to their importance
in chemistry, physics, and biology.[1] Whereas the roles of
“conventional” O–H···O HBs in determining the structure
and properties of water and functions of biomolecules have
been recognized for more than 50 years, C–H···O interac-
tions were categorized as true HBs only later, after sufficient
experimental data (IR, Raman, NMR spectra, gas-phase
and solid structures, and theoretical results) had been ac-
cumulated.[2] The interaction energy of CH4 and H2O was
calculated to be rather small: between 0.3 and
0.7 kcalmol–1, depending on the computational methods.[3]
This energy increases considerably when electronegative
substituents are attached to the C atom: the BSSE-cor-
rected MP2/6-31+G(d,p) values are 0.29 and 3.7 kcalmol–1
for CH3H···OH2 and CF3H···OH2, respectively.[4] The latter
IR spectroscopic data imply that proton-donating ability
increases in the series C(sp3)–H Ͻ C(sp2)–H Ͻ C(sp)–H,[2a]
which is corroborated by quantum mechanical calculations.
Thus, C–H···O binding energies for benzene as proton do-
nor were reported to amount to 1.1–1.7 kcalmol–1, de-
pending on the level of calculation and the proton ac-
ceptor.[6] These energies are considerably increased in poly-
fluorinated benzenes. However, relatively few examples of
vinylic or aromatic C–H groups involved in strong, easily
detectable HBs are known. Experimentally, formation of
complexes between 1,2,4,5-tetrafluorobenzene as proton
donor and various proton acceptors in the gas phase have
been observed by fluorescence-detected IR spectroscopy.[7]
Aromatic C–H···N and C–H···O HB formation was evi-
denced by low-frequency shifts and intensity enhancement
of the aromatic C–H stretching vibration. The stabilization
energy of the complex with dimethyl ether was calculated to
amount to 5.39 kcalmol–1 [MP2/6-31+G(d), without BSSE
correction], or to 4.27 and 3.16 kcalmol–1 with 50% and
100% BSSE correction, respectively.[7b] Similarly, super-
sonic-jet spectroscopy showed that fluorinated benzenes
form H-bonded dimers with 2-pyridone.[8] The C–H···O=C
[a] Centre Interdisciplinaire de Nanoscience de Marseille (CINaM)
(UPR CNRS 3118),
13288 Marseille Cedex 9, France
Fax: +33-4-91829301
E-mail: khodor@luminy.univ-mrs.fr
[b] Department of Chemistry, Ben Gurion University of the Negev,
Beer Sheva 84105, Israel
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
Eur. J. Org. Chem. 2008, 3689–3699
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3689

C. Niebel, V. Lokshin, M. Sigalov, P. Krief, V. Khodorkovsky
FULL PAPER
HB strengths were calculated as 2.8 kcalmol–1 for fluoro- condensation are a feature of HBs.[14] Indeed, the accepted
benzene and 4.0 kcalmol–1 for tetrafluorobenzenes at the geometric criteria of the existence of the HB as derived
PW91/6-311++G(d,p) level.[8b]
from crystallographic data[15] involve distances R and r, to-
Several theoretical investigations have been carried out gether with angles θ and φ
on model compounds to estimate possible HBs in solids,
especially those in proteins.[6c,9] C–H···O HBs, although
weaker than conventional O–H···O HBs, are certainly one
of the factors predetermining the structures of solids –
structures of proteins in particular. Calculations on N-
methylmaleimide complexes with acetone and water, in
which the both vinylic H atoms were involved in HBs,
yielded 1.4–1.7 kcalmol–1 per bond.[9c] The contribution of
where the relevant ranges are R = 3.0–4.0 Å, r = 2.0–2.8 Å,
θ = 110–180°, and φ = 120–140°.
the C–H···O interaction to stabilization of the adenine–ura-
cil pair has been estimated to reach 6% of the total interac-
tion energy.[9d] It is difficult to extract an estimation of the
HB energy from the total interaction energy either in this
case or from the results of other recently reviewed studies
relating to nucleic acid dimers.[10] C–H···O bonding has
been observed in several transition metal poly(azolyl)borato
complexes.[11a] Statistical analysis of the C–H···O interac-
tion in the doubly H-bonded C=CH–C=O dimer using data
available in the Cambridge Structural Database showed that
this pattern occurs frequently in crystals.[11b] The total bind-
ing energy within fragments approximating to H-bonded
1,4-benzoquinone dimers was calculated as 4.3 kcalmol–1,
of which about 1.4 kcalmol–1 was partitioned for each C–
H···O HB (MP2, EZPPB basis set). The interaction energy
in a two-dimensional slab of a 1,4-benzoquinone crystal
was calculated to reach 5.3 kcalmol–1 [HF/6-31 G(d,p),
BSSE-corrected].[12]
According to the X-ray structure, derivative 1 fits these
criteria, with R = 2.955 Å, θ = 138°, and φ = 117°. We
showed that of the 1.8 ppm low-field shift of the aromatic
H-7, only about 1.2 ppm can be explained by the magnetic
anisotropy of the carbonyl group, oxygen steric influence,
and electric field effect. The remaining 0.6 ppm can be re-
garded as a reasonable measure of the influence of the in-
tramolecular HB.[14] However, the rigid structure of deriva-
tive 1, involving two cyclic moieties connected by the
double bond, makes estimation of the HB energy also prob-
lematic in this case.
Here we report on our investigation of a series of 2-[3-
oxoisobenzofuran-1(3H)-ylidene]acetate derivatives 2–6.
The a series (E isomers) possesses a fragment isostructural
to bindone (1) in the vicinity of the ester carbonyl group
and exhibits similar 1H NMR features: the signals of the
protons potentially involved in intramolecular H-bonding
are strongly shifted to low field (9.1–10.7 ppm in CDCl3).
Unlike 1, however, derivatives 2 can also exist as Z isomers
(b series) and also, possibly, as aЈ and bЈ series. Whereas no
intramolecular HB can form either in the b or in the bЈ
series, the activated H atoms are free to participate in inter-
molecular HBs with the solvent molecules. Therefore, these
compounds might serve as very useful models for study of
HB-related phenomena. Indeed, quantum mechanical cal-
culations reproduce the observed 1H NMR features and
provide estimates of the HB energies.
C–H···O HBs with energies lower than 1.5 kcalmol–1 can
indeed be an important factor influencing crystal lattice
structures, but would hardly be expected to be observed at
room temperature in solution, as calculations predict that
HBs of this type weaken considerably with growing dielec-
tric constant.[9a] On the other hand, aromatic C–H···O in-
teractions can exceed 4 kcalmol–1 and in this case can be
studied in solution by conventional spectroscopic tech-
niques, of which NMR is one of the most informative. It
has thus been found that a pyridylurea–tetraazaanthracene-
dione complex involving three HBs is more stable than an
analogous complex with four HBs.[13] The X-ray structures
of the tetraazaanthracenedione dimer and the complex re-
vealed the presence of two N···H HBs and two C–H···O
contacts (2.941 Å) in the former and three N···H HBs and
one C–H···O contact (2.520 Å) in the latter. Monitoring of
Results and Discussion
Synthesis
1
the dimerization process by H NMR showed that the cor-
responding C–H signal is shifted by about 0.1 ppm. The
Derivatives 2 and their analogues are well known precur-
complexation process, which gave rise to the formation of sors for synthesis of natural and bioactive compounds, and
the intramolecular C–H···O contact (2.520 Å), was ac- several synthetic approaches have been elaborated. They
companied by a 1.73 ppm downfield shift. The C–H···O have been prepared by esterification of the acid,[16] treat-
short contacts were classified as “putative HBs” as they in- ment of o-iodobenzoic acid with ethyl propiolate in the
deed might be forced by the N···H HBs, and the observed presence of cuprous iodide in DMF,[17] Baylis–Hillman re-
downfield shifts may stem from the C–H groups’ proximity actions between o-carboxybenzaldehydes and acrylates,[18]
to the C=O anisotropic deshielding cones.[13] Recently, we palladium-catalyzed carbonylative cyclizations of Baylis–
demonstrated that unusually large low-field chemical shifts Hillman adducts,[19] and through Wittig reactions between
of one of the aromatic protons in the 1H NMR spectra phthalic anhydrides and [(alkoxycarbonyl)methylene]tri-
of bindone (1) and other products of indan-1,3-dione self- phenylphosphoranes.[20]
3690
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 3689–3699

C(sp2)–H···O Hydrogen Bonds
Except in the cases of the first two methods, which af- H-bonding cannot occur for steric reasons. Preparatively,
forded pure Z isomers, E isomers – characterized by un- the Z isomers can be isolated by dissolving the E isomers
usually large chemical shifts of 7-H (8.5–9 ppm) – were the in warm sulfuric acid and pouring onto ice.
sole or predominant products. Thus, Wittig reactions
Derivatives 3–6 were previously unknown and were pre-
yielded exclusively E isomers, in 60% yield when the reac- pared from commercially available pyromellitic anhydride
tion was carried out without a solvent[20a] or in 87% yield and the bis-anhydride of naphthalene-2,3,6,7-tetracarbox-
after 18 h in chloroform at reflux.[20b] Later, during a mech- ylic acid.[22]
anistic study of this reaction, it was found that in chloro-
form at room temperature the E isomer was formed in 52%
yield along with a 13% yield of the Z isomer.[20c] Non-cyclic
anhydrides are known to react with stabilized phosphoranes
to afford the acylation products, diketo phosphoranes,[21]
and the occurrence of the normal Wittig reaction and the
exclusive or predominant formation of the E isomers with
phthalic anhydride were explained in terms of formation of
a π–π complex between the C=O group of the phosphorane
and the electron-deficient benzenic ring of phthalic anhy-
dride.[20a,20c,20d] In the process of this investigation, when
we realized that 3,6-H atoms of phthalic anhydride and its
analogues can form relatively strong HBs, we assumed that
the H-bonded complexes with the phosphorane can precede
the Wittig reaction, and that chloroform, which is itself cap-
able of H-bonding, is therefore not the best choice as a sol-
vent. Indeed, the reaction in dry toluene afforded a mixture
of E and Z isomers 2a and 2b (R = Et) in 84% and 12%
yields, respectively. Formation of 2b could either be ex-
plained by the presence of acidic impurities in the commer-
cial ethoxycarbonylmethylenephosphorane or might result
It is worth noting that, although pyromellitic anhydride
from intermolecular reaction of the non-complexed compo-
is the strongest electron acceptor of the three anhydrides
nents.
employed here, the yield of the mixture of isomers in this
case was just 53%, whereas the mixture of isomers 5 and 6
was isolated in 95% yield. At the same time, the activated
aromatic C–H atoms are the most shielded (by the four car-
bonyl groups) in pyromellitic anhydride, which speaks more
in favor of the H-bonded complex intermediates than those
of the π–π complex. The full synthesis and isolation pro-
cedures are given in the Supporting Information.
X-ray Structures and Geometry Optimization
The NMR spectra of derivatives 2–6 (R = Et) showed
that each isolated pair of isomers is the E and Z isomer
with regard to one of two possible conformations of the
C=C–C=O fragment, but could not unequivocally distin-
guish between the cis and trans conformers a/aЈ and b/bЈ.
All signals are sharp, and no temperature dependence of
their shape and position was observed (in toluene between
20 and 100 °C, for instance), which excludes the possibility
of a ǞaЈ and bǞbЈ interconversion in solution.
X-ray structure determinations were carried out for de-
We found that all derivatives of the a series are indeed rivatives 2a, 2b, 3a, 4a, and 5a (R = Et), confirming the
sensitive to acids and partially isomerize into the Z isomers structural assignment. Single crystals of each compound
when heated in toluene in the presence of a catalytic were selected from batches of crystals of identical shape, the
amount of methanesulfonic acid. Isomerization of 2a into NMR spectra of which were identical to the spectra of the
2b had previously been observed to occur only on UV irra- compounds isolated from the reaction mixtures. All non-
diation.[20a] In all probability, isomerization proceeds hydrogen atoms and the aromatic and vinylic hydrogen
through intermediates of type 7, in which intramolecular atoms lie in one plane (within 1°), except for the carbonyls
Eur. J. Org. Chem. 2008, 3689–3699
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3691

C. Niebel, V. Lokshin, M. Sigalov, P. Krief, V. Khodorkovsky
FULL PAPER
of the ester groups, which form dihedral angles of 3.5–6.4°
(7.9° for 2b) with this plane. All derivatives of the a series
exhibit short C–H···O=C contacts: R varies between 2.91 Å
(3a) and 2.97 Å (4a), θ between 133° (3a) and 135° (5a),
and φ between 120° (5a) and 122° (3a) (see Figures 1, 2, 3,
and 4). The geometries of the C–H···O=C fragments in
these derivatives are very close to that in bindon (1) and
also perfectly fit the criteria for HBs. It is noteworthy that
of two R distances in 4a, one is one of the shortest (2.928 Å)
and another one the longest (2.965 Å) within the series.
Figure 2. Selected experimentally determined and calculated
(printed in italics) [B3LYP 6-31G(d,p)] distances [Å]. Compound
3a: C4–C4a, C8–C8a 1.388(2) (1.40); C8a–C1, C5a–C5 1.472(2)
(1.47); C1–C9, C5–C10 1.335(3) (1.35); C9–C11, C10–C12 1.467(3)
(1.47); C11–O1, C12–O4 1.206(2) (1.22); O1···C8, O4···C4 2.913
(2.94).
Figure 3. Selected experimentally determined and calculated
(printed in italics) [B3LYP 6-31G(d,p)] distances [Å]. Compound
4a: C8–C8a 1.390(3) (1.40), C8–C7a 1.394(3), C8a–C1 1.471(3)
(1.48), C7a–C7 1.469(3), C1–C9 1.332(3) (1.35), C7–C10 1.328(3),
C9–C11 1.463(3) (1.47), C10–C12 1.462(3).
Figure 1. Selected experimentally determined and calculated
(printed in italics) [B3LYP 6-31G(d,p)] distances [Å]. a) Compound
2a: C7–C7a 1.393(3) (1.40), C7a–C1 1.464(3) (1.47), C1–C8
1.329(3) (1.35), C8–C9 1.460(3) (1.47), C9–O1 1.196(3) (1.22),
O1···C7 2.957 (2.98); b) Compound 2b: C7–C7a 1.393(2) (1.40),
C7a–C1 1.458(2) (1.47), C1–C8 1.333(2) (1.35), C8–C9 1.460(3)
(1.47), C9–O1 1.207(2) (1.21), O1···O2 2.825 (2.869).
Since part of our later argument is based on the results
of quantum mechanical calculations, it is instructive at this
point to analyze how well the experimental geometries are
reproduced with different model chemistries.[23] Geometry
optimizations for 2–6 (R = Me) starting from nonplanar
molecules without symmetry constraints rapidly converged
to the planar systems in all cases (except for the crowded
4aЈ), which indicates the lack of appreciable steric hin-
drance. The experimentally determined R values for 2a are
Figure 4. Selected experimentally determined and calculated
(printed in italics) [B3LYP 6-31G(d,p)] distances [Å]. Compound
5a: C5–C5a, C10–C10a 1.372(3) (1.38); C5a–C6, C10a–C1 1.468(3)
(1.47); C1–C11, C6–C12 1.329(3) (1.35); C11–C13, C12–C14
1.460(3) (1.47); C13–O1, C14–O4 1.191(3) (1.22); O1···C10,
overestimated by about 0.02 Å by both HF/6-31G(d) and O4···C5 2.934 (2.96).
3692
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 3689–3699

C(sp2)–H···O Hydrogen Bonds
B3LYP/6-31G(d,p),[24] by 0.03 Å by B3LYP/6-31+G(d,p), Table 1. Relative energies (kcalmol–1) of the major isomers of de-
rivatives 2–6 calculated by the B3LYP/6-31G(d,p)//B3LYP/6-
31G(d,p) method (ZPE-corrected values given in parentheses).
by 0.01 Å by MP2/6-31G(d), and by 0.01 Å by MP2/6-
31+G(d,p) levels. The experimentally determined O1···O2
distance in 2b (2.83 Å) is reproduced on the MP2/6-31G(d)
level, but overestimated by 0.03, 0.04, 0.06, and 0.02 Å by
∆E [kcalmol–1]
∆E [kcalmol–1]
2a
0 (0); 0 (0)[a]
3.5 (3.2); –0.3 (–0.6)[a]
4.0 (3.9); 2.5 (2.4)[a]
3.4 (3.2); 0.8 (0.5)[a]
0 (0)
6.8 (6.5)
6.8 (6.8)
7.0 (6.7)
0 (0)
4aЈ
4bЈ
5a
7.5 (7.5)
5.0 (4.8)
0 (0)
8.3 (7.8)
8.5 (8.3)
8.2 (7.8)
0 (0)
8.0 (7.6)
8.7 (8.5)
7.8 (7.5)
the HF/6-31G(d), B3LYP/6-31G(d,p), B3LYP/6-31+G(d,p), 2b
2aЈ
2bЈ
3a
3b
3aЈ
3bЈ
4a
and MP2/6-31+G(d,p) levels, respectively. On the whole, for
both derivatives 2a and 2b, the HF/6-31G(d) model chemis-
try yielded the smallest mean deviation (about 0.005 Å),
5b
5aЈ
5bЈ
6a
6b
6aЈ
6bЈ
whereas the B3LYP and MP2 methods gave very similar
results with mean deviations of about –0.011 Å, both some-
what biased toward overestimation of the bond lengths. Ad-
dition of diffuse functions only increases the bias with re-
gard to the aromatic C–C and exocyclic C=C bonds. Fur-
ther calculations were therefore carried out with the HF/6-
31G(d) and B3LYP/6-31G(d,p) model chemistries as a rea-
sonable compromise. These two methods, applied to deriva-
tives 3a–5a, gave mean deviations of 0.009 and –0.007 Å,
respectively, which is close to the experimentally determined
3σ values. The details on the experimentally determined
and calculated geometries and statistical analysis of the re-
sults are given in the Supporting Information.
Geometry optimizations and energy comparisons in the
following sections were performed without taking the basis
set superposition error (BSSE) into account, leading to
overestimation of the interaction energy, since the usual
procedure of counterpoise correction[25] cannot be used for
single molecules. However, with the relatively small basis
set employed in this study, the incomplete recovery of the
correlation energy giving rise to underestimation of the in-
teraction energy (i.e., acting in the opposite direction) is
also pronounced, and the BSSE correction is known to
underestimate the energies in many cases even at higher
levels of theory.[26] Only comparison of computational re-
sults with available experimental data can answer the ques-
tion of whether the two opposite factors cancel at the se-
lected level of theory. It is shown in the following sections
that canceling errors indeed occur to a considerable degree
at the B3LYP/6-31G(d,p) level.
4b
5.1 (4.8)
[a] Calculated in DMSO.
cepting effect of the neighboring carbonyl group). This pos-
sibility makes our estimation the minimum at the selected
level of calculations. Analogous considerations applied for
the other derivatives of the a series (except for 4a, as 4aЈ is
nonplanar as a result of steric hindrance) yield 3.3 (3.3), 4.3
(4.2), and 4.5 (4.3) kcalmol–1 per HB for 3a, 5a, and 6a,
respectively. Interestingly, according to this estimation, the
HB strength increases in the series 3a Ͻ 2a Ͻ 5a Ͻ 6a,
although the aromatic protons in 3a are the most activated
by the accepting substituents.
Scheme 1. Relative energies of derivative 2 isomers.
Energetics of Intramolecular HBs
An alternative estimation of HB energies can be achieved
The calculated relative energies of the major possible iso- if we consider the transition state energies for, for example,
mers of derivatives 2–6 are given in Table 1. the conversions 2a Ǟ2aЈ and 2bǞ2bЈ. The relaxed poten-
In the gas phase, the derivatives of the a series are the tial energy scans for rotation over the CH–COOR bonds
most stable in all cases and those of the aЈ series are the are instructive and are shown in Figure 5. Unexpectedly
least stable. Derivatives of the b and bЈ series have approxi- large barriers to rotation (about 5 kcalmol–1 for conver-
mately the same stability. As an initial approximation, we sions 2aЈǞ2a and 2bǞ2bЈ) show that interconversion be-
can assume that through-space interaction (namely, the tween these species, at least in nonpolar solvents, is hardly
HB) exists only in the a series, as shown in Scheme 1 for possible. Apparently, the conjugation between the alkoxy-
derivatives 2. In such case, ∆Ecis/trans = –0.1 kcalmol–1 carbonyl group and the phthalide moiety contributes con-
(0 kcalmol–1 ZPE-corrected) and the energy of the HB in siderably to stabilization of each species. The Z isomer 2bЈ
2a is E2a – ∆Ecis/trans = 4.1 (3.9) kcalmol–1. The energy is more stable than the E isomer 2aЈ with the same trans
differences in terms of enthalpies [∆H°(298 K)] are very configuration by ∆E2 = 0.6 kcalmol–1. The difference in en-
close. Of course, derivatives of the aЈ series may also involve ergies between the transition states ∆E3, when both the con-
non-negligible stabilizing C–H···O-bonding (weakened by jugation and possible through-space interactions between
the unfavorable planar conformation and electron-ac- the two moieties are “switched off”, is about 0.3 kcalmol–1,
Eur. J. Org. Chem. 2008, 3689–3699
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3693

C. Niebel, V. Lokshin, M. Sigalov, P. Krief, V. Khodorkovsky
FULL PAPER
also in favor of the Z isomers. Both ∆E2 and ∆E3 are of the the first step and 3.7 kcalmol–1 for the second. This is a
same order of magnitude, and the difference between them very reasonable result showing that the ability of the aro-
might originate either from destabilization of 2aЈ or from matic H atom to form the HB is almost (but not fully)
stabilization of 2bЈ by 0.3 kcalmol–1. Therefore, in the ab- exhausted by one HB. It is also in qualitative agreement
sence of the HB in 2a, that isomer should be less stable with the experimental findings (see discussion in the next
than 2b by the same 0.3 kcalmol–1 (∆E3), since both are in section).
the same cis configuration. This provides us with another
Analogous calculations performed for the pairs 2a Ǟ2aЈ
independent estimation of the HB energy in 2a: E2b + ∆E3 and 2bǞ2bЈ with the IEF-PCM model in DMSO (ε = 46.7)
= 3.5 + 0.3 = 3.8 kcalmol–1. This value is close to the pre- revealed a different picture (Figure 6). In this case, the most
vious estimation from ∆E1 (4.1 kcalmol–1) and is also the stable isomer is 2b, which is more stable than 2a by
minimum value, as we assumed that there are no additional 0.3 kcalmol–1. The reason for such considerable stabiliza-
stabilizing effects in 2b, such as through-space –O···O= in- tion of 2b is the increase in the –O···O= ↔ –O+···–O–
teraction. The correction of 0.3 kcalmol–1 should be larger through-space interaction induced by the reaction field of
if such an interaction is present. Interestingly, this estima- the polar solvent. Indeed, whereas the optimized geometries
tion involving the energies of the transition states is almost of 2bЈ in the gas phase and DMSO are similar and the –
independent of the calculation method: we obtained 3.8 and O···O– distance remains practically the same, the –O···O=
3.7 kcalmol–1 from the analogous calculations on the HF/ distance in 2b in DMSO is shortened by 0.021 Å. At the
6-31G(d) and MP2/6-31(d) levels, respectively.
same time, the H···O distance in 2a in DMSO is 0.04 Å
longer than in the gas phase. As above, we can estimate the
HB energy by comparing the energies of 2a and 2aЈ [i.e., as
∆E1, which is 2.5 (2.4) kcalmol–1]. If it is taken into account
that in the gas phase 2b was less stable than 2bЈ by
0.1 kcalmol–1, the extra stabilization of 2b in DMSO can
be estimated as 1.1 kcalmol–1. The difference in energies
between 2aЈ and 2bЈ (∆E2) is close to the difference in ener-
gies of the two excited states (∆E3), as was the case in the
gas phase. Therefore, another independent estimation of the
HB energy is 1.1+1.4 = 2.5 kcalmol–1 (i.e., the same value
as above).
Figure 5. Relaxed potential energy scans for conversions 2a Ǟ2aЈ
and 2bǞ2bЈ calculated by B3LYP/6-31G(d,p) method in the gas
phase. The dihedral angle H–C–C=O was varied between 0° and
180°.
The transition state energies were also calculated at the
same level of theory for derivatives 3 and 5. These deriva-
tives each involve two CH–COOR bonds, so two transition
states corresponding to consecutive rotation over each bond
were found. A scheme used for estimation of HB energies
and the energies of the corresponding isomers is given in
the Supporting Information. Breaking the first HB in 3a
produced 3.7 kcalmol–1 as the minimum estimation of the
HB energy at the B3LYP/6-31G(d,p) level, and the same
value was calculated for the second step. The HF/6-31G(d)
method yielded 3.7 and 3.8 kcalmol–1 for the first and the
Figure 6. Relaxed potential energy scans for conversions 2a Ǟ2aЈ
and 2bǞ2bЈ in DMSO, calculated by the B3LYP/6-31G(d,p)
method. The dihedral angle H–C–C=O was varied between 0° and
180°.
It is noteworthy that the barriers for the conversions
second steps, respectively. The same method applied for 5a 2aЈǞ2a and 2bǞ2bЈ calculated in DMSO, although lower
gave an estimation of 4.4 kcalmol–1 for each HB in this than those calculated in the gas phase, still exceed
molecule, which is close to the 4.3 kcalmol–1 estimated from 4 kcalmol–1. All energy differences computed in DMSO are
the energy difference between 5a and 5aЈ. The approach practically the same in terms of ∆H°(298 K) as in the case
involving transition state energies can be applied to 4a, as of the gas-phase calculations. Calculated ZPE-corrected en-
the energy of the sterically crowded 4aЈ is not involved. The ergies, enthalpies, and free energies of all isomers and tran-
corresponding calculations afforded only 0.8 kcalmol–1 for sition states are given in the Supporting Information.
3694
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 3689–3699

C(sp2)–H···O Hydrogen Bonds
NMR Spectra of Derivatives 2–6 and Intermolecular H-
Bonding
Table 2. Experimentally determined and calculated [GIAO B3LYP/
6-31G(2d,p)] 1H chemical shifts (ppm) for derivatives 2–6.
[D8]Toluene CDCl3 [D6]Acetone [D6]DMSO Calculated
1
The H signals (δ) of all aromatic and vinylic protons
2a H-4
H-6
7.62
7.29
9.29
6.12
7.54
7.01
6.90
5.69
–
–
–
–
–
–
–
–
–
–
–
–
–
8.01
7.86
9.09
6.19
8.01
7.82
7.79
5.92
9.70
6.35
8.32
6.08
8.55
10.71
6.39
8.83
9.95
6.32
8.77
10.00
6.34
8.16
8.12
9.19
6.28
8.17
8.10
8.33
6.32
–
–
–
–
–
–
–
–
–
–
–
–
–
8.12
8.07
8.94
6.33
8.11
8.02
8.34
6.51
9.34
6.50
8.92
6.71
8.55
10.15
6.50
8.92
9.58
6.30
9.11
9.74
6.41
8.05
7.88 (8.19)[a]
9.97 (9.49)[a]
5.95
8.08
7.74
7.77 (8.42)[a]
5.67 (6.46)[a]
9.93
(except for H-5 in 2) in derivatives 2–6 undergo unusually
large solvent shifts (∆δ). The corresponding experimentally H-7
determined chemical shifts and those calculated by the
GIAO method[27] are given in Table 2. The calculated values
agree reasonably well with the experimentally determined
shifts observed in chloroform. However, the chemical shifts
of the protons participating in the HBs are overestimated
as in the case of derivative 1,[14] for possible reasons dis-
cussed below. The most solvent-sensitive protons in deriva-
tives 2 are those most activated by the electron-accepting
H-8
2b H-4
H-6
H-7
H-8
3a H-4,8
H-9,10
3b H-4,8
H-9,10
4a H-4
5.72
7.93
5.52
8.23
11.28
5.79
8.84
10.81
6.13
8.70
10.93
6.13
substituents. In both compounds, the δ value of the vinylic H-8
H-9,10
5a H-4,9
H-5,10
H-11,12
6a H-4,5
H-9,10
H-8 gradually shifts towards lower field when moving from
toluene (neither a proton donor nor acceptor) to chloro-
form (a proton donor), acetone (a proton acceptor), and
DMSO (both a proton donor and acceptor). The overall ∆δ
of this proton in 2b is 0.82 ppm and is almost four times
larger than in 2a. The overall solvent shift of aromatic H-4
is about 0.5 ppm in both compounds, reaching its maxi-
mum in acetone. These large shifts can only partially be
explained by the known Aromatic Solvent Induced Shifts
(ASISs), as the ∆δ values are still unusually large for the
H-11,12
[a] Calculated for 1:1 complexes with DMSO for the protons in-
volved in H-bonding.
1
The H NMR spectra of 2a and 2b in CDCl3/DMSO
chloroform ǞDMSO solvent pair. The most remarkable mixtures were also measured, with different molar ratios of
from this point of view is the behavior of H-7. Whereas ∆δ these solvents and constant concentration (4.6ϫ10–2 ) of
in 2b amounts to 1.64 ppm (tolueneǞDMSO), the trend the solutes. The changes in the chemical shifts of the most
observed in 2a is the opposite, and ∆δ is –0.31 ppm. The sensitive protons of 2a and 2b, along with those of chloro-
aromatic H-4 and H-8 in 3a and 3b show the same pattern: form, are shown in Figure 7. The initial rapid changes be-
∆δ (CDCl3 ǞDMSO) is –0.36 and 0.60 ppm, respectively. come smaller, but increase again after the molar ratio of
Although the NMR spectra calculations in solvent give CDCl3/DMSO becomes approximately 1:1. An additional
generally good results (see[28] and refs. therein), our sharp shift to lower field was observed when the 5:95%
attempts to reproduce the chemical shifts of 2a and 2b by mixture was replaced by pure DMSO. Although the initial
calculations using the IEF-PCM model in DMSO did not changes of the chloroform proton shift were more pro-
yield even qualitative agreement with the experiment.
nounced, qualitatively the behavior of all protons, except
In general, these experimental observations and the failure for H-7 of 2a, is the same, so a more detailed consideration
of the computational approach to predict the chemical shifts of the HB in Cl3C–H···O=SMe2 may be helpful here.
of the H-bonded protons reasonably without taking specific
The formation of 2:1 and 1:1 chloroform/DMSO com-
interactions into account can be explained as follows. With plexes has been a subject of several experimental investi-
regard to the relaxed potential scan results (see Figures 5 and gations, including NMR and thermodynamic data analy-
6), we can see that at room temperature (about 0.6 kcalmol–1) sis.[30] The H-bonding natures of these complexes have been
the dihedral angle in the H–C–C=O fragment can reach 20°, demonstrated by Raman and IR spectra measurements of
opening room for the solvent to form an intermolecular HB their mixtures.[31] The reported enthalpy of 1:1 complex for-
in addition to the intramolecular one. No intermolecular HB mation was determined to be about –2.6,[30c] –2.7,[30d] and
form with toluene is possible, and the observed δ value for 2a –3.3[30a] kcalmol–1. Neat DMSO is known to exist associ-
is lower than that calculated as the result of ASIS. Acetone ated in oligomers and dimers.[30d,32] Assuming that all
can form a weaker HB with H-7 in addition to the existing DMSO molecules are dimerized, we calculated ∆H°(298 K)
intramolecular bond, only slightly lowering the overall effect, for the reaction (DMSO)2 + CHCl3 + CHCl3 Ǟ 2(CHCl3/
whereas chloroform will partially bind the C=O group, weak- DMSO). The HF method yielded –2.2 kcalmol–1, whereas
ening the existing intramolecular HB more noticeably. The the B3LYP method gave –2.6 kcalmol–1 as a minimum esti-
variations of ∆δ (CDCl3ǞDMSO) and the different signs for mation, in good agreement with the experimentally mea-
the 2a, 3a and the 2b, 3b couples are more difficult to inter- sured values. As discussed above, no BSSE corrections were
pret. They are considerably larger than usually observed for carried out either for geometry optimization or in the final
polar aprotic compounds (Ϯ0.1 ppm)[29] and comparable to frequency calculations. Using the counterpoise correction
the shifts of dichloromethane and chloroform (0.49 and procedure yielded an unreasonably large ∆H°(298 K) value
1.06 ppm, respectively), explained by formation of HBs with of –4.9 kcalmol–1 with the B3LYP method.[33] The calcu-
DMSO.[29]
lated C–H stretching frequency of 2964 cm–1 for the com-
Eur. J. Org. Chem. 2008, 3689–3699
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3695

C. Niebel, V. Lokshin, M. Sigalov, P. Krief, V. Khodorkovsky
FULL PAPER
Figure 7. 1H chemical shifts vs. molar ratio of [D6]DMSO in
CDCl3/[D6]DMSO mixtures. Triangles: protons of residual CHCl3,
circles: H-7 (2b), rhombs: H-8 (2b), squares: H-7 (2a).
plex (scale factor 0.961) is also in good agreement with the
experimentally measured value of 2977 cm–1.[31] The HB en-
ergy, which corresponds to the electronic energy difference
between the complex (CHCl3/DMSO) and the free compo-
nents, was 6.5 (HF) and 7.4 kcalmol–1 (B3LYP) (ZPE-cor-
rected).
Analogous calculations for 1:1 complexes 2a/DMSO and
2b/DMSO were also performed with the B3LYP method.
Whereas the ∆H°(298 K) value for the first complex was
only –0.4 kcalmol–1, for the second one it amounted to
–6.3 kcalmol–1. The optimized structures of both com-
plexes involve several shortened O···H distances characteris-
Figure 8. The equilibrium geometries of 2a/DMSO and 2b/DMSO
complexes.
tic of HBs, as shown in Figure 8. In the 2a/DMSO complex, 1:1 after 16 h of heating. Derivative 3a is more reactive and
the carbonyl group interacts not only with the aromatic H7 had already isomerized at 100 °C, in agreement with the
hydrogen atom, but also with both of the Me groups in weaker HBs predicted by calculations. Derivative 4a had
DMSO. As a result, the aromatic C–H···O=C becomes already undergone isomerization upon warming of its
0.072 Å longer than in 2a, accounting for the high-field DMSO solution, in agreement with the weakest first HB
shift observed in the NMR spectrum in DMSO. Both aro- energy predicted by calculations. In DMSO it may proceed
matic and vinylic H atoms interact with the O atom of via intermediates of type 8 resulting from the nucleophilic
DMSO in 2b; these protons are indeed the most sensitive attack of the DMSO oxygen atom on the positively polar-
to the presence of DMSO. The above qualitative conclu- ized C-1 atoms of 2a–6a. Delocalization of the negative
sions are also supported quantitatively: the calculated charge within the C–COOR moiety prevents rotation about
chemical shifts for the protons involved in H-bonding in the C–C bond so that after the intramolecular HB is
complexes 2a/DMSO and 2b/DMSO given in Table 2 are in broken, neither 2aЈ nor 2bЈ can form.
better agreement with the experimentally determined val-
The isomerization of derivatives of the a series into the
ues. Apparently, calculations should be carried out for com- b series, stabilized in the solvents of high dielectric con-
plexes involving two or even three molecules of DMSO to stants, can thus proceed both by electrophilic (sulfuric acid,
reproduce the chemical shifts of all protons.
ε = 100) and by nucleophilic (DMSO, ε = 46.7) pathways.
Thus, in addition to stabilization of 2b in the reaction However, the NMR spectra of the crude reaction mixtures
field of the polar solvent (DMSO) and the corresponding did not reveal the presence of derivatives of the aЈ and bЈ
weakening of the intramolecular HB in 2a, the specific in- series, which can be regarded as true isomers owing to the
teraction with this solvent increases these effects and gives high isomerization barriers.
1
rise to the characteristic H NMR spectral changes. This
The 13C chemical shifts, along with the 13C–1H coupling
conclusion is also confirmed experimentally. We found that constants for 2a, 2b, 3a, and 4a, are collated in Tables 3 and
although heating of 2a and 5a in dry DMSO at 100 °C for 4. The example of chloroform is also helpful for analysis of
1
more than 20 h did not affect their NMR spectra, the ap- these data. It is well known that not only are the H and
pearance of 2b and the isomer of 5b involving only one HB 13C chemical shifts of chloroform changed in the solvents
was observed at 140 °C. The ratio of isomers became about capable of forming the HB, but the 13C–1H coupling con-
3696
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 3689–3699

C(sp2)–H···O Hydrogen Bonds
Table 3. 13C Chemical shifts [ppm] and 1J(13C-1H) [Hz] for 2a and 2b.
C-1
C-3
C-4
C-5
C-6
C-7
C-8
C-9
C-4a
C-7a
2a[a]
2a[b]
2b[a]
2b[b]
157.79
165.68
125.31
(167.0)
125.44
132.44
(165.1)
133.18
(167.0)
132.27
(165.1)
132.74
(165.1)
135.23
(163.1)
135.68
(165.1)
135.00
(165.1)
135.54
(165.1)
128.18
(172.7)
127.32
(172.7)
121.16
(165.1)
122.52
(167.0)
102.43
(163.1)
101.83
(165.1)
96.00
(163.1)
95.86
(167.0)
165.61
126.52
136.10
157.28
154.04
154.07
165.18
165.57
165.82
164.09
163.61
163.09
126.08
124.75
124.01
135.30
138.96
138.64
125.98
(167.0)
125.49
(167.0)
[a] In CDCl3. [b] In [D6]DMSO.
Table 4. 13C Chemical shifts [ppm] and 1J(13C-1H) [Hz] of 3a and 4a in CDCl3.
C-1
C-3
C-4
C-5
C-7
C-8
C-9
C-4a
C-8a
3a
4a
155.75
163.71
126.05
(178.5)
122.76
(174.8)
155.75
163.71
126.05
(178.5)
128.26
(182.3)
105.35
(163.1)
106.10
(163.1)
138.65
129.71
132.13
141.40
155.78
163.35
163.35
155.78
stant also increases considerably.[34] Thus, the value of are strikingly insensitive to the calculation methods and, if
1J(13C-1H) increases by 8.2 Hz for chloroform solution in our assumptions are wrong, will only increase. B3LYP/6-
DMSO.[34b]
31G(d,p) not only appeared to be optimal for describing the
For 2a and 2b in chloroform, the major variation is seen geometries, but also fortunately makes the BSSE correction
in C-7, directly attached to the hydrogen atom involved in unnecessary. The estimated HB energy in 2a is diminished
the HB in 2a. Both the chemical shift and the one-bond by about 35% in DMSO solution, because of the solvent
coupling constant of C-7 are larger – by 7 ppm and 7.6 Hz, reaction field, but still amounts to 2.5 kcalmol–1 in this sol-
respectively – in 2a than in 2b. Similar, albeit smaller, varia- vent. Moreover, specific interactions with DMSO giving
tions can be seen in the spectra of 3a and 4a. The difference rise to the formation of H-bonded complexes, are especially
1
in the one-bond coupling constant of C-8 and C-4 in 4a is evident in the H NMR spectra of the derivatives of the b
still about the same, at 7.5 Hz. The observed changes in the series, in which the activated H atom, locked in the a series,
one-bond coupling constants are close to those detected for can now form intermolecular HBs with the solvent.
derivative 1 and its analogues.[14] In DMSO, the variations
of the one-bond coupling constants are smaller, but never-
Experimental Section
theless instructive. Whereas all coupling constants, except
for C-7, are increased by about 2 Hz in 2a, only those of C-
7 and C-8 are increased – by 2 and 5 Hz, respectively – in
2b, in agreement with the proposed model.
Experimental Procedures: Toluene was dried by azeotropic distil-
lation and 1,4-dioxane by heating at reflux over sodium and distil-
lation under argon. Commercially available phthalic anhydride and
pyromellitic dianhydride were both purified by crystallization in
acetic anhydride.
Conclusions
Melting points were recorded on a Büchi 510 capillary melting
point apparatus. IR spectra were measured on a Nicolet Avatar 320
FT-IR spectrometer. NMR spectra were recorded on a Bruker
250 MHz spectrometer. The NMR parameters (δ in ppm, J in Hz)
were determined by simulation and subsequent optimization by use
of the gNMR 5.0 program (P. H. M. Budzelaar, ©1995–2006 Ivory-
Soft).
The X-ray structure determinations, selective formation,
high thermal stability in solution, and the NMR spectral
features demonstrate the presence of relatively strong intra-
molecular HBs in derivatives of the a series. We used two
independent approaches to evaluate the HB energies in
these compounds: one based on the assumption that there
are no through-space stabilizing interactions in derivatives
of the aЈ series and the second involving the differences in
energies of the transition states. The two approaches yielded
similar gas-phase energy estimations, which vary between
3.5 and 4.4 kcalmol–1. The only exception was derivative
4a, in which the aromatic H atom is bonded with two oxy-
gen atoms and the energy corresponding to breaking the
first of two HBs is only 0.8 kcalmol–1, showing that the H-
bonding capability of the H atom is already almost fully
saturated by one HB, the energy of which is the same as in
Reaction between Phthalic Anhydride and Ethyl (Triphenylphosphor-
anylidene)acetate: Ethyl (triphenylphosphoranylidene)acetate (1.18 g,
3.38 mmol) was added in portions over 30 min at room temperature
to a solution of phthalic dianhydride (0.5 g, 3.38 mmol) in toluene
(5 mL). After stirring for 3 h at room temperature and heating at
85 °C for 15 min, the reaction mixture was concentrated under re-
duced pressure. The crude pink solid obtained was purified by
chromatography on silica gel (Et2O/cyclohexane, 3:7). The major
isomer 2a was eluted first (0.61 g, 84%), followed by the minor
isomer 2b (0.09 g, 12%). The overall yield 96%.
Ethyl [(E)-3-Oxo-1,3-dihydroisobenzofuran-1-ylidene]acetate (2a):
derivative 3a (3.7 kcalmol–1). The HB energy estimations Colorless crystals, m.p. 71.0–72.0 °C. H NMR (CDCl3): δ = 9.09
1
Eur. J. Org. Chem. 2008, 3689–3699
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3697

C. Niebel, V. Lokshin, M. Sigalov, P. Krief, V. Khodorkovsky
FULL PAPER
3
4
5
3
[7]
[8]
[9]
a) V. Venkatesan, A. Fujii, T. Ebata, N. Mikami, J. Phys. Chem.
A 2005, 109, 915–921; b) V. Venkatesan, A. Fujii, N. Mikami,
Chem. Phys. Lett. 2005, 409, 57–62.
a) R. Leist, J. A. Frey, S. Leutwyler, J. Phys. Chem. A 2006,
110, 4180–4187; b) J. A. Frey, R. Leist, S. Leutwyler, J. Phys.
Chem. A 2006, 110, 4188–4195.
a) T. Kar, S. Scheiner, Int. J. Quantum Chem. 2006, 106, 843–
851; b) D. Wu, L. Liu, G. Liu, D. Jia, J. Phys. Chem. A 2007,
111, 5244–5252; c) B. Wang, J. F. Hinton, P. Pulay, J. Phys.
Chem. A 2003, 107, 4683–4687; d) E. B. Starikov, T. Steiner,
Acta Crystallogr., Sect. D 1997, 53, 345–347.
(m, J6,7 = 8.0, J5,7 = 0.9, J4,7 = 0.8 Hz, 1 H, H-7), 8.01 (m, J4,5
4
3
= 7.6, J4,6 = 1.2 Hz, 1 H, H-4), 7.86 (m, J5,6 = 7.5 Hz, 1 H, H-
5
6
6), 7.74 (m, 1 H, H-5), 6.19 (br., J7,8 = 0.3, J6,8 = 0.1 Hz, 1 H,
H-8), 4.26 (q, 2 H, 3J = 7.1 Hz, CH2), 1.33 (t, 3 H, 3J = 7.1,
CH3) ppm. 13C NMR (CDCl3): δ = 165.68 (C-3), 165.61 (C-9),
157.79 (C-1), 136.10 (C-7a), 135.23 (C-6), 132.44 (C-5), 128.18 (C-
7), 126.52 (C-4a), 125.31 (C-4), 102.43 (C-8) ppm. IR (Nujol): ν
1805, 1787 (C=O phthalide), 1710 (C=O ester), 1652
(C=C) cm–1
˜
max
=
.
Ethyl [(Z)-3-Oxo-1,3-dihydroisobenzofuran-1-ylidene]acetate (2b):
Colorless crystals, m.p. 133.0–134.0 °C. 1H NMR (CDCl3): δ = 8.01
[10]
[11]
S. Scheiner in ref. [1d], p. 263–292.
a) C. Janiak, T. G. Scharmann, J. C. Green, R. P. G. Parkin,
M. J. Kolm, E. Riedel, W. Mickler, J. Elguero, R. M.
Claramunt, D. Sanz, Chem. Eur. J. 1996, 2, 992–1000; b) J.
Van de Bovenkamp, J. M. Matxain, F. B. van Duijneveldt, T.
Steiner, J. Phys. Chem. A 1999, 103, 2784–2792.
3
4
5
3
(m, J4,5 = 7.7, J4,6 = 0.8, J4,7 = 1 Hz, 1 H, H-4), 7.82 (m, J5,6
= 7.5, 3J6,7 = 7.9 Hz, 1 H, H-6), 5.88 (br., 5J7,8 = 0.4, 6J6,8 = 0.2 Hz,
1 H, H-8), 4.30 (q, 3J = 7.1 Hz, 2 H, CH2), 1.35 (t, 3J = 7.1 Hz,
CH3) ppm. 13C NMR (CDCl3): δ = 165.57 (C-3), 163.61 (C-9),
154.04 (C-1), 138.96 (C-7a), 135.00 (C-6), 132.27 (C-5), 125.98 (C-
[12]
G. I. Cárdenas-Girón, A. Masunov, J. J. Dannenberg, J. Phys.
Chem. A 1999, 103, 7042–7046.
4), 124.75 (C-4a), 121.16 (C-7), 96.00 (C-8) ppm. IR (Nujol): ν
˜
max
[13]
[14]
= 1805 (sh), 1792 (C=O phthalide), 1718 (C=O ester), 1676
(C=C) cm–1
J. R. Quinn, S. C. Zimmerman, Org. Lett. 2004, 6, 1649–1652.
M. Sigalov, A. Vashchenko, V. Khodorkovsky, J. Org. Chem.
2005, 70, 92–100.
G. R. Desiraju, Acc. Chem. Res. 1996, 29, 441–449.
A. Gabriel, Ber. Dtsch. Chem. Ges. 1924, 57, 301–302.
C. E. Castro, E. J. Gaughan, D. C. Owsley, J. Org. Chem. 1966,
31, 4071–4078.
K. Y. Lee, J. M. Kim, J. N. Kim, Synlett 2003, 357–360.
F. Coelho, D. Veronese, C. H. Pavam, V. I. de Paula, R. Buffon,
Tetrahedron 2006, 62, 4563–4572.
a) P. A. Chopard, R. F. Hudson, R. J. G. Searle, Tetr. Lett.
1965, 2357–2360; b) A. Allahdad, D. W. Knight, J. Chem. Soc.
Perkin Trans. 1 1982, 1855–1863; c) M. M. Kayser, K. L. Hatt,
D. L. Hooper, Can. J. Chem. 1992, 70, 1985–1996; d) M. M.
Kayser, K. L. Hatt, H. S. Yu, D. L. Hooper, Can. J. Chem.
1993, 71, 1010–1021.
.
Isomerization of Derivatives from the a-Series into the b-Series: A
derivative from the a-series (0.10 g) was dissolved in concentrated
sulfuric acid (1 mL). After stirring for 20 h at room temperature,
the reaction mixture was poured into ice. The precipitate was fil-
tered and washed with water. The crude esters were purified by
crystallization to remove the admixture of the corresponding acid.
[15]
[16]
[17]
[18]
[19]
Compound 3b: This compound was obtained in 60% yield as a col-
1
[20]
orless powder, m.p. Ͼ 250 °C. H NMR (CDCl3): δ = 8.37 (s, 2 H,
3
H-4,8), 6.13 (s, 2 H, H-8,10), 4.38 (q, J = 7.1 Hz, 4 H, CH2), 1.42
(t, 3J = 7.1 Hz, 6 H, CH3) ppm. 13C NMR (CDCl3): δ = 163.34
(C-3,7), 162.70 (C-10,11), 151.94 (C-1,5), 141.53 (C-4a,7a), 130.46
(C-5a,8a), 119.32 (C-4,8), 99.22 (C-9,10), 61.47 (2ϫCH2), 14.21
(2ϫCH ) ppm. IR (Nujol): ν
= 1813 (C=O phthalide), 1697
˜
3
max
[21]
[22]
[23]
P. A. Chopard, R. J. G. Searle, F. H. Devitt, J. Org. Chem.
1965, 30, 1015–1019.
O. W. Webster, W. H. Sharkey, J. Org. Chem. 1962, 27, 3354–
3355.
(br., C=O ester + C=C) cm–1
.
Supporting Information (see also the footnote on the first page of
this article): Experimental procedures, NMR spectroscopic data,
Cartesian coordinates and energies for computed structures, X-ray
structure experimental details.
M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, J. A. Montgomery Jr, T. Vreven,
K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tom-
asi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega,
G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota,
R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratch-
ian, J. B. Cross, C. Adamo, J. Jaramillo, R. Gomperts, R. E.
Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli,
J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Sal-
vador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D.
Daniels, M. C. Strain, O. Farkas, D. K. Malick, A. D. Rabuck,
K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Ba-
boul, S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Lia-
shenko, P. Piskorz, I. Komaromi, R. L. Martin, D. J. Fox, T.
Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, M.
Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W.
Wong, C. Gonzalez, J. A. Pople, Gaussian 03, Revision B.05,
Gaussian, Inc., Pittsburgh PA, 2003.
R. G. Parr, W. Yang, Density-functional theory of atoms and
molecules. Oxford Univ. Press, Oxford, 1989.
S. F. Boys, F. Bernardi, Mol. Phys. 1970, 19, 553–556.
K. S. Kim, P. Tarakeshwar, Y. Y. Lee, Chem. Rev. 2000, 100,
4145–4186.
K. Wolinski, J. F. Hilton, P. Pulay, J. Am. Chem. Soc. 1990,
112, 8251–8260.
C. Benzi, O. Grescenzi, M. Pavone, V. Barone, Magn. Reson.
Chem. 2004, 42, S57–S67.
Acknowledgments
M. S. and P. K. are thankful to the Israel Science Foundation for
partial financial support (Grant 135/06).
[1] a) G. A. Jeffrey, An Introduction to Hydrogen Bonding, Oxford
University Press, New York, 1997; b) S. Scheiner, Hydrogen
Bonding, Oxford University Press, New York, 1997; c) G. R.
Desiraju, T. Steiner, The Weak Hydrogen Bond, Oxford Univer-
sity Press, Oxford, 1999; d) Hydrogen Bonding – New Insights
(Ed.: S. J. Grabowski), Springer, Berlin, 2006.
[2] a) A. Allerhand, P. v. R. Schleyer, J. Am. Chem. Soc. 1963, 85,
1715–1723; b) R. D. Green, Hydrogen Bonding by C–H Groups,
Wiley Interscience, New York, 1973.
[3] a) D. E. Woon, P. Zeng, D. R. Beck, J. Chem. Phys. 1990, 93,
7808–7812; b) M. M. Szczesniak, G. Chalasinski, S. M. Cybul-
ski, P. Cieplak, J. Chem. Phys. 1993, 98, 3078–3089; c) J. J. No-
voa, F. Mota, Chem. Phys. Lett. 1997, 266, 23–30.
[4] Y. Gu, T. Kar, S. Scheiner, J. Am. Chem. Soc. 1999, 121, 9411–
9422.
[24]
[25]
[26]
[27]
[28]
[29]
[5] P. Hobza, Z. Havlas, Chem. Rev. 2000, 100, 4253–4264.
[6] a) R. L. Ornstein, Y. Zheng, J. Biomol. Struct. Dyn. 1997, 14,
657–665; b) K. Kim, R. A. Friesner, J. Am. Chem. Soc. 1997,
119, 12952–12961; c) S. Scheiner, T. Kar, J. Pattanyak, J. Am.
Chem. Soc. 2002, 124, 13257–13264.
R. J. Abraham, J. J. Byrne, L. Griffiths, M. Perez, Magn. Re-
son. Chem. 2006, 44, 491–509.
3698
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 3689–3699

C(sp2)–H···O Hydrogen Bonds
[30] a) A. L. McClellan, S. W. Nicksic, J. C. Guffy, J. Mol. Spec-
trosc. 1963, 11, 340–348; b) W.-C. Lin, S. J. Tsay, J. Phys. Chem.
1970, 74, 1037–1041; c) D. V. Fenby, L. G. Helper, J. Chem.
Thermodyn. 1974, 6, 185–189; d) V. A. Durov, O. G. Tereshin,
I. Yu. Shilov, J. Mol. Liquids 2004, 110, 69–79.
[31] D. C. Daniel, J. L. McHale, J. Phys. Chem. A 1997, 101, 3070–
3077.
[32] R. L. Frost, J. Kristof, E. Horvath, J. T. Kloprogge, J. Phys.
Chem. A 1999, 103, 9654–9660.
dissociation energy of 5.4 kcalmol–1, discussed in detail in W.
Koch, M. C. Holthausen, A Chemist’s Guide to Density Func-
tional Theory, Wiley-VCH, 2001, chapter 12. We calculated the
binding energy of the dimer at the same B3LYP/6-31G(d,p)
level with and without BSSE correction to obtain 3.1 and
5.3 kcalmol–1, respectively (ZPE-corrected).
[34] a) R. L. Lichter, J. D. Roberts, J. Phys. Chem. 1970, 74, 912–
916; b) D. F. Evans, J. Chem. Soc. 1963, 5575–5577.
Received: April 7, 2008
[33] The water dimer is one of the most intensively studied intermo-
lecular H-bonded systems, with an experimentally determined
Published Online: June 3, 2008
Eur. J. Org. Chem. 2008, 3689–3699
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3699