Hydrogen-bond symmetry in difluoromaleate monoanion

Article abstract of DOI:10.1021/ja2117848

The symmetry of the hydrogen bond in hydrogen difluoromaleate monoanion is probed by X-ray crystallography and by the NMR method of isotopic perturbation in water, in two aprotic organic solvents, and in an isotropic liquid crystal. The X-ray crystal structure of potassium hydrogen difluoromaleate shows a remarkably short O-O distance of 2.41 A and equal O-H distances of 1.206 A, consistent with a strong and symmetric hydrogen bond. Incorporation of ¹⁸O into one carboxyl group allows investigation of the symmetry of the H-bond in solution by the method of isotopic perturbation. The ¹⁹F NMR spectra of the mono-¹⁸O-substituted monoanion in water, CD₂Cl₂, and CD₃CN show an AB spin system, corresponding to fluorines in different environments. The difference is attributed to the perturbation of the acidity of a carboxylic acid by ¹⁸O, not to the mere presence of the ¹⁸O, because the mono-¹⁸O dianion shows equivalent fluorines. Therefore, it is concluded that the monoanion exists as an equilibrating pair of interconverting tautomers and not as a single symmetric structure not only in water but also in organic solvents. However, in the isotropic liquid crystal phase of 4-cyanophenyl 4-heptylbenzoate, tetrabutylammonium hydrogen difluoromaleate- ¹⁸O shows equivalent fluorines, consistent with a single symmetric structure. These results support earlier studies, which suggested that the symmetry of hydrogen bonds can be determined by the local environment.

Full text of DOI:10.1021/ja2117848

Article
pubs.acs.org/JACS
Hydrogen-Bond Symmetry in Difluoromaleate Monoanion
Charles L. Perrin,* Phaneendrasai Karri, Curtis Moore, and Arnold L. Rheingold
Department of Chemistry, University of CaliforniaSan Diego, La Jolla, California 92093-0358, United States
*
S Supporting Information
ABSTRACT: The symmetry of the hydrogen bond in
hydrogen difluoromaleate monoanion is probed by X-ray
crystallography and by the NMR method of isotopic
perturbation in water, in two aprotic organic solvents, and in
an isotropic liquid crystal. The X-ray crystal structure of
potassium hydrogen difluoromaleate shows a remarkably short
O−O distance of 2.41 Å and equal O−H distances of 1.206 Å,
consistent with a strong and symmetric hydrogen bond.
1
8
Incorporation of O into one carboxyl group allows investigation of the symmetry of the H-bond in solution by the method of
1
9
18
isotopic perturbation. The F NMR spectra of the mono- O-substituted monoanion in water, CD Cl , and CD CN show an AB
2
2
3
spin system, corresponding to fluorines in different environments. The difference is attributed to the perturbation of the acidity
1
8
18
18
of a carboxylic acid by O, not to the mere presence of the O, because the mono- O dianion shows equivalent fluorines.
Therefore, it is concluded that the monoanion exists as an equilibrating pair of interconverting tautomers and not as a single
symmetric structure not only in water but also in organic solvents. However, in the isotropic liquid crystal phase of
18
4
-cyanophenyl 4-heptylbenzoate, tetrabutylammonium hydrogen difluoromaleate- O shows equivalent fluorines, consistent with
a single symmetric structure. These results support earlier studies, which suggested that the symmetry of hydrogen bonds can be
determined by the local environment.
INTRODUCTION
Hydrogen-Bond Symmetry. A fundamental structural
question is whether the hydrogen in a hydrogen bond
H-bond) between two donor atoms of equal basicity is centered
between the two donors (“symmetric” H-bond) or is closer to
one and jumping between them. This is the distinction between
a single-well and a double-well potential. Centered hydrogens
are associated with short heavy-atom distances, and with “strong”
H-bonds, although these associations have been questioned.
Centered hydrogens can be observed in crystalline phases, but
they are elusive in solution, both protic and aprotic. Previous
studies suggested that the symmetry of the H-bonds can be
disrupted by the instantaneous local solvation environment,
Liquid Crystals. If the difference in H-bond symmetry is
due to the difference between the order of a crystal and the
disorder of solution, an intermediate environment might be a
good test of this interpretation. Such an intermediate
environment can be realized by dissolving the solute molecule
■
(
8
1
to be studied in a liquid crystalline solvent. Among different
types of liquid crystals, nematic phases are the least ordered and
the most mobile. A nematic phase is composed of rod-shaped
molecules that are aligned parallel to each other and that
2
9
orient solute molecules relative to their long axis. After some
3
exploration, 4-cyanophenyl 4-heptylbenzoate was chosen for
this study. It is a nematic liquid crystal between 40 and 53 °C,
10
4
and becomes isotropic above 53 °C.
For the purposes of this study, the isotropic phase of this
which is disordered, leading to solvatomers (isomers that differ
5
material was used at 55 °C. It avoids the dipolar couplings that
in solvation). The disorder of solvation renders the two donor
11
appear in the nematic phase. Although the long-range order
of the nematic phase is absent (as is the alignment relative to
the applied magnetic field in an NMR spectrometer), the
atoms instantaneously unequal, whereupon the hydrogen
attaches to the less well solvated donor.
Recently, the results of those studies were reinterpreted in
terms of the desymmetrizing effect of isotopic substitution on an
6
12
isotropic phase retains short-range order. The hope of this
study is that the short-range order is sufficient that the two
carboxyls are in the same environment, similar to the situation
in a crystal.
7
otherwise symmetrical potential-energy surface. The desym-
metrization arises from anharmonic coupling between a
desymmetrizing vibrational mode and other modes whose
zero-point energies are isotope-dependent. Calculated trajecto-
ries of hydrogen motion then lead to H-bond asymmetries and
time-averaged NMR chemical shifts for the ipso carbon of
phthalate monoanion that are comparable to experimental ones,
at least in organic solvents. Therefore, it was concluded that the
asymmetry of an aprotic organic solvent environment is not
strong enough to disrupt the symmetry of the H-bond, although
such a disruption was accepted for aqueous solution.
To avoid submerging the NMR signals of the substrate under
1
13
the dominant signals of the H or C nuclei of a liquid crystal
1
19
solvent, and to avoid the necessity of decoupling H, F NMR
is especially convenient. Besides, 19F NMR chemical shifts are
quite sensitive to chemical environment, a property that was
Received: December 27, 2011
Published: April 21, 2012
©
2012 American Chemical Society
7766
dx.doi.org/10.1021/ja2117848 | J. Am. Chem. Soc. 2012, 134, 7766−7772

Journal of the American Chemical Society
Article
1
3
can be estimated from the 19F NMR chemical shifts of
difluoromaleic acid 1 (or less reliably of the monoanion 3) and
difluoromaleate dianion 4, respectively. These are very crude
estimates because fluorine adjacent to carboxylate is also
opposite to a carboxylic acid and vice versa, and because the
useful for measuring secondary deuterium isotope effects.
Therefore, difluoromaleic acid (1) has been chosen for this
study. It has an exceedingly simple NMR spectrum. The avai-
lability of the anhydride (2) makes possible the incorpora-
tion of one 18O into the diacid, for the purpose of isotopic
perturbation.
14
−
H-bond between the OH and the O affects the chemical shifts.
It is also necessary to measure Δ , the intrinsic isotope shift due
0
18
to the proximity to O, but this can be modeled by diacid 1 or
dianion 4. Nevertheless, the signature of a tautomeric mixture
is a difference between the chemical shifts of the upper
1
8
16
fluorine, adjacent to O, and the lower fluorine, adjacent to O.
13
Indeed, such a difference was seen in the C NMR spectrum
1
8
19
of aqueous mono- O-labeled succinic acid, and the dipolar
couplings of this monoanion in a liquid crystal were measured
Isotope Shifts and Isotopic Perturbation of Equili-
brium. Asymmetric structures can be distinguished exper-
imentally from symmetric ones by the NMR method of isotopic
20
and used to evaluate the conformation.
15
K − 1
perturbation. This was originally applied to carbocations, and
it succeeds even when signals are coalesced by rapid exchange.
The method depends on measuring the isotope shift (isotope
Δeq
=
D
K + 1
(3)
We therefore have undertaken a study of the symmetry of
the H-bond in difluoromaleate monoanion by X-ray diffraction
n
effect on chemical shift) Δ, due to a heavier isotope n atoms
16
away from the reporter nucleus (eq 1). There are two
contributions to the observed isotope shift (eq 2), an intrinsic
18
in a crystalline solid and by the method of O-induced isotopic
19
perturbation of F NMR chemical shifts in water, in two
shift, Δ , owing simply to the presence of an isotope, and a
0
organic solvents, and in a nematic liquid crystal as solvent. We
now report that this H-bond is symmetric in the crystal and
in the isotropic liquid crystal, but not in aqueous or organic
solution.
shift, Δ , induced by isotopic perturbation of an equilibrium.
eq
n
Δ = δheavy − δlight
(1)
Δobs = Δ + Δ
0
eq
(2)
EXPERIMENTAL SECTION
Synthesis. Difluoromaleic anhydride (2) was obtained by the
literature procedure, starting from 1,1,2-trichloro-2,3,3-trifluorocyclo-
butane or from trifluorosuccinic acid.
Hydrolysis of anhydride 2 in water produced difluoromaleic acid
■
Our task is to distinguish whether difluoromaleate
monoanion has a centered hydrogen (3) or is a mixture
of two tautomers, with an asymmetric H-bond between the
2
1
−
18
OH and the O . To do so, we use mono- O-substituted di-
1
8
fluoromaleate monoanion. Either this has a centered hydrogen
(1), free of the stereoisomeric fumaric acid. Incorporation of O could
be achieved by stirring 0.1−0.2 mmol of anhydride with 10−15 μL of
1
8
(
3
3- O) or it is an equilibrium mixture of two tautomers,
18
1
8
18
H
O plus 50−100 μL of tetrahydrofuran for solubility. The extent
a- O and 3b- O. (The presence of additional isotopomers
2
18
18
of mono- O-labeling was monitored by mass spectrometry, and a
reaction time of 2 h at room temperature was found suitable. At longer
with O in the carbonyl instead of in the H-bond halves the
observed isotope effect but can be ignored.). Owing to
vibrational frequencies and zero-point energies, an 18O-labeled
carboxylic acid is ∼1% less acidic than an ordinary carboxylic
acid. Consequently, the tautomeric equilibrium is shifted
slightly toward 3b- O, with an equilibrium constant K ∼ 1.01.
1
8
times, further equilibration occurred, and additional O could be
incorporated. Alternatively, to favor a high yield of mono- O-acid, the
18
18
16
anhydride was hydrolyzed in H2 O containing >2 equiv of Na OH,
1
7
18
18
such that the O content was ∼85%, to produce the dianion (4- O).
The excess hydroxide is necessary because with only one equivalent
the resulting monoanion would neutralize another equivalent of
hydroxide.
1
8
Incorporation of 18O was confirmed by mass spectrometry. The
1
8
diacid isolated from one reaction of the anhydride with H O was 1:1
2
1
8
unlabeled and mono- O-labeled, plus a small amount of doubly labeled
Figure S1, Supporting Information). Diacid from another reaction
was a 1:2 mixture (Figures S3, S4). Mass spectrometric analysis was
(
18
unsuccessful for the nonvolatile dianion (4- O), even after conversion
18
to the diacid, owing to O exchange in 20% H SO . A more reliable
2
4
This perturbation of the equilibrium by an isotope then
results in a chemical-shift inequivalence between the two
measure of the 18O content is 85%, from intensities in the F NMR
19
spectrum shown and analyzed below.
1
8
fluorines. In tautomer 3a- O, the upper F is adjacent to a
The potassium salt of the unlabeled monoanion (3) was generated
by treating the diacid with 1 equiv of potassium hydroxide. Potassium
hydrogen difluoromaleate was then isolated by concentrating the
solution to dryness and recrystallizing the salt from water. The 50%-
labeled diacid served as a sample of the monoanion because this is
such a strong acid that it is fully ionized in dilute aqueous solution.
carboxylate, and the lower F is adjacent to a carboxylic acid. In
1
8
19
tautomer 3b- O, these are reversed. If the F NMR chemical
‑
shift of a fluorine adjacent to a carboxylate is designated as δCO2
and the chemical shift of a fluorine adjacent to a carboxylic acid
group as δCOOH, then the chemical shifts of the upper and lower
18
‑
‑
The sodium salt of the mono- O-difluoromaleate monoanion was
obtained from the solution of the dianion by adding small amounts
of acetic acid until the chemical shifts remained constant. The
tetrabutylammonium salt of the partially labeled monoanion (1:1 or
fluorines are X δ
a CO2
+ X3bδCOOH and X3aδCOOH + X3bδCO2 ,
3
respectively, where X3a and X3b are mole fractions. The
difference between these two chemical shifts is then an
equilibrium isotope shift, given by eq 3, where D = δ
−
18
COOH
1:2 3 + 3- O) was prepared by dissolving the partially labeled diacid
1
8
3
4
‑
δCO2
.
Strictly, this isotope shift is Δ − Δ, because both
in water containing 1 equiv of tetrabutylammonium hydroxide and
removing the water under high vacuum to give a white solid. To
confirm that the sample was the monoanion, the intensity of the H
18
fluorines are subject to the influence of the O, but their
1
8
1
‑
relationship to the O differs. The parameters δ
and δCO2
COOH
7
767
dx.doi.org/10.1021/ja2117848 | J. Am. Chem. Soc. 2012, 134, 7766−7772

Journal of the American Chemical Society
Article
signal at 20.08 ppm in CD CN was compared with the intensities of
irregularity with proton number may be due to uncertainty in
the chemical shift of the BF4− standard.
3
the n-butyl signals.
18
The tetrabutylammonium salt of 1:2 3 + 3- O was dissolved in
-cyanophenyl 4-heptylbenzoate to produce a 2−3 wt % solution. The
If the parameters δCOOH and δCO2 are estimated from the 19F
‑
4
NMR chemical shifts of 1 and 4, respectively, and if the
sample was homogenized by repeated heating to the isotropic phase
and then cooling back to the nematic phase, with physical shaking or
1
8
18
equilibrium constant between tautomers 3a- O and 3b- O is
vortex mixing. The 19F NMR spectrum was then obtained for the
taken as 1.01, then eq 3 predicts a chemical-shift difference bet-
18
ween the two fluorines of mono- O-difluoromaleate mono-
anion of 0.07 ppm. Of course, this is a crude estimate, inasmuch
as the models for K and D lack the H-bond, which is likely to
decrease the values of these parameters. Nevertheless, the
isotropic phase.
NMR Spectroscopy. NMR spectra were recorded at 18 °C on a
1
9
JEOL-500 FT-NMR spectrometer (470 MHz F). In aqueous
solutions, NaBF was used as internal standard, with two sets of
4
2
2
peaks in a 1:4 ratio at −151.04 ppm (br) and −151.10 ppm (q). In
organic solvents and in liquid crystal, CFCl (0 ppm) was used as
19
sensitivity of F chemical shifts to the state of protonation is
3
what makes this experiment feasible.
internal standard.
19
F NMR Spectra and Signal Assignments. The NMR
18
18
spectrum of mono- O-labeled difluoromaleate dianion (4- O)
RESULTS
■
in D O is shown in Figure S2. The NMR spectrum of
2
18
18
Crystal Structure. Crystallographic data for potassium
hydrogen difluoromaleate are summarized in Table S1. The
structure of hydrogen difluoromaleate monoanion (3) within a
crystal of potassium hydrogen difluoromaleate is shown in
Figure 1. According to the X-ray data, the O−O distance in the
mono- O-labeled difluoromaleate monoanion (3- O) in D O
2
is shown in Figure 2. For comparison, the NMR spectrum of a
18
1
:2 mixture of unlabeled and mono- O-labeled difluoromaleate
monoanion in D O is shown in Figure S3. The NMR spec-
2
trum of the tetrabutylammonium salt of a 1:2 mixture of
18
unlabeled and mono- O-labeled difluoromaleate monoanion
in CH CN is shown in Figure 3. The NMR spectrum of the
3
1
8
tetrabutylammonium salt of a 2:1 mixture of mono- O-labeled
and unlabeled difluoromaleate monoanion in CD Cl is shown
2
2
in Figure S4. The NMR spectrum at 55 °C of a 1:2 mixture of
18
unlabeled and mono- O-labeled tetrabutylammonium hydro-
1
8
18
gen difluoromaleate- O (3 + 3- O) in 4-cyanophenyl
4
-heptylbenzoate liquid crystal as solvent is shown in Figure 4.
The simplest 19F NMR spectra are those of the mono- O-
18
labeled difluoromaleate dianion in D O (Figure S2) and of the
2
18
tetrabutylammonium salt of the partially O-labeled monoanion
in 4-cyanophenyl 4-heptylbenzoate liquid crystal (Figure 4).
Those spectra show only singlets. The diacid (1), produced in
Figure 1. X-ray structure of hydrogen difluoromaleate monoanion in
the crystal of its potassium salt.
18
20% H SO from the O-labeled difluoromaleic acid, also
2 4
1
8
showed a singlet, but it is likely that the H SO catalyzed
O
2
4
hydrogen difluoromaleate monoanion is 2.41 Å, and the
hydrogen is placed exactly in the middle of two oxygen atoms,
with equal O−H distances of 1.206 Å. The centering of the
hydrogen is a consequence of its position on a crystallographic
mirror plane. The alternative of a hydrogen disordered across
the mirror plane is less likely, in view of the neutron-diffraction
data that exclude this possibility for the similar salt, potassium
exchange with solvent, so that this spectrum is inconclusive.
The most complicated spectra are those of the mono- O-
1
8
labeled difluoromaleate monoanion in D O (Figure 2) and of
2
18
the 2:1 mixture of mono- O-labeled and unlabeled difluor-
omaleate monoanion in D O (Figure S3). Each of them is six
2
18
lines. Because they differ only in the extent of O incorporation,
comparison permits the strong signals at 125.263 ppm to
be assigned to unlabeled difluoromaleate monoanion (3).
Unlabeled material is a necessary consequence of the presence
23
hydrogen chloromaleate.
Chemical Shifts. The 19F NMR chemical shift of the
difluoromaleate dianion (4) could be obtained from the
product of hydrolysis of the anhydride in excess NaOH. The
chemical shift of the monoanion (3) could be obtained from
the aqueous solution of the diacid because this is such a strong
acid that it is fully ionized in dilute aqueous solution. The
chemical shift of difluoromaleic acid (1) could be obtained only
in 20% H SO , where it is likely that the chemical shift of the
!
2
6
of H O in the hydrolysis.
19
The four F NMR signals at −125.240, −125.245, −125.286,
and −125.291 in Figure 2, two upfield and two downfield
of the signal due to 3, can then be recognized as a pair of
doublets constituting an AB spin system, arising from the
1
8
18
mono- O-labeled difluoromaleate monoanion (3- O). Like-
wise, the outer four signals in Figure S3 are an AB pattern.
The chemical-shift separation between the two fluorines is
2
4
−
standard has changed, inasmuch as the chemical shift of BF is
quite sensitive to solvent. Table 1 lists F chemical shifts for
4
24
19
3
0.046 ppm, and the scalar coupling constant J between them is
2
.4 Hz. This coupling constant is unusually small, even for cis
Table 1. 19F Chemical Shifts (ppm) for Difluoromaleic Acid
fluorines, but a 5-Hz coupling constant was observed for the
25
26
(
1), Its Monoanion (3), and Its Dianion (4)
monobenzyl ester of difluoromaleic acid. The weak signal
at −125.268 ppm in Figure 2 or at −125.272 ppm in Figure S3
can then be assigned to doubly O-labeled difluoromaleate
monoanion (3- O ), with one O in each carboxyl.
Incorporation of a second O is a consequence of the further
1
3
4
18
−
126.43
−125.26
−139.60
18
18
2
18
difluoromaleic acid and its conjugate bases in aqueous media.
As expected, F chemical shifts are quite sensitive to the state of
protonation, especially for the dianion. However, the apparent
equilibration of the acid formed from hydrolysis of the
19
18
18
anhydride in H O. However, incorporation of a second
O
2
into the dianion formed by hydrolysis of the anhydride with
7
768
dx.doi.org/10.1021/ja2117848 | J. Am. Chem. Soc. 2012, 134, 7766−7772

Journal of the American Chemical Society
Article
Figure 2. 19F NMR spectrum of mono- O-labeled difluoromaleate monoanion (3- O) in D O.
18
18
2
1
9
18
Figure 3. F NMR spectrum of the tetrabutylammonium salt of a 1:1 mixture of unlabeled and mono- O-labeled difluoromaleate monoanion
1
8
(
3 + 3- O) in CD CN.
3
NaOH is more problematic. The mechanism for this
incorporation is proposed below.
signal for unlabeled 3 is more prominent. The coupling
constant is still 2 Hz, but the chemical-shift separation is lower,
0.029 ppm. In CD Cl (Figure S4), the AB pattern is still
The 19F NMR spectrum of the 18O-labeled monoanion in
2
2
CH CN (Figure 3) also shows an AB pattern, although the
present, with the same lower chemical-shift separation, along
3
7
769
dx.doi.org/10.1021/ja2117848 | J. Am. Chem. Soc. 2012, 134, 7766−7772

Journal of the American Chemical Society
Article
Figure 4. 19F NMR spectrum at 55 °C of a 1:2 mixture of tetrabutylammonium hydrogen difluoromaleate and difluoromaleate- O (3 + 3- O) in
18
18
4-cyanophenyl 4-heptylbenzoate isotropic liquid crystal.
with a prominent signal for unlabeled 3, but the coupling
isotope shifts, Δ , due to perturbation of an equilibrium, as
eq
constant is not resolvable.
Isotope Shifts. The F NMR singlet for the mono- O-
labeled difluoromaleate dianion (4- O) (Figure S2) shows that
there is no resolvable intrinsic isotope shift Δ due to
substitution per se. The two middle F NMR peaks of the
in eq 3.
1
9
18
Values of Δeq are summarized in Table 2. According to the
estimate based on eq 3, Δeq ought to be >0. This means that
1
8
18
O
0
1
9
Table 2. Equilibrium 19F Isotope Shifts (ppm) for
Mono- O-Labeled Difluoromaleate Monoanion (3- O)
1
8
18
difluoromaleate monoanion in Figures 2 and S3, due to 3 and
1
8
3
- O , are offset slightly upfield and downfield relative to the
2
solvent
Δ , ppm
eq
1
8
center of the AB pattern of 3- O. This offset represents a very
D O
0.046
0.029
0.028
18
2
small intrinsic isotope shift Δ of 0.0025 ppm per O, and it
0
CD CN
3
seems to be the same (or unresolvably different) regardless of
2
CD Cl2
whether the 1 O is separated from the F by 3 or 4 bonds.
8
19
1
8
The key result is the AB pattern for the O-labeled difluo-
romaleate monoanion (3- O) in D O (Figure 2), in CD CN
18
the fluorine adjacent to O is predicted to be downfield of the
1
8
2
3
16
fluorine adjacent to O. However, it is not possible to verify
(
Figure 3), and in CD Cl (Figure S3). The two fluorines are
2 2
this because the fluorines cannot be assigned relative to the
inequivalent, with different chemical shifts, and they can split each
other. The chemical-shift separations are 0.046, 0.029, and
18
position of the O.
A further result is that no isotope shift is seen in Figure 4 for
0.028 ppm, respectively. These are not due to an intrinsic
18
the tetrabutylammonium salt of difluoromaleate- O mono-
isotope shift Δ because no Δ was resolvable in the dianion (4).
0
0
anion in 4-cyanophenyl 4-heptylbenzoate liquid crystal solvent
at 55 °C. Although the signal is broader than the signals in
isotropic solutions, it is a singlet, with no sign of separation into
an AB multiplet. For comparison, Figure S5 is an adaptation of
Figure S4, with the same 1:2 composition as that of Figure 4,
but with an additional line-broadening of 6 Hz to match the
Moreover, if these were intrinsic, they would be the difference
between three-bond and four-bond isotope shifts, whereas each
1
8
of these in the monoanion is only 0.0025 ppm per O and
3
4
the same for Δ and Δ , so that the difference would be far
0
0
less than what is observed. Therefore, these are equilibrium
7
770
dx.doi.org/10.1021/ja2117848 | J. Am. Chem. Soc. 2012, 134, 7766−7772

Journal of the American Chemical Society
Article
line-width of Figure 4. This is graphic evidence that
difluoromaleate- O monoanion in liquid crystal does not
display an AB pattern within the broadened signal.
This chemical-shift difference is thus an isotope shift attribu-
table to the perturbation of an equilibrium by the isotopic label
and indicates that the H-bonded monoanion exists as a pair of
tautomeric structures. Its H-bond is not symmetric.
18
DISCUSSION
In contrast, in the isotropic phase of the liquid crystal
4-cyanophenyl 4-heptylbenzoate as solvent, the tetrabutylam-
■
Mechanism for Incorporation of Two 18O into Difluo-
romaleate Dianion (4). The weakest signals in Figure 2
18
18
monium salt of hydrogen difluoromaleate- O (3- O) appears
as a single peak, albeit broadened, and does not show any
isotope shift. In principle, this result could be due to a smaller
D (eq 3) than in the other two aprotic organic solvents.
However, the near identity of the observed isotope shifts in
those two solvents (Table 2) suggests that a similar AB pattern
would have been seen in this liquid crystal. Subject to this
uncertainty in D, the absence of an isotope shift indicates that
in this environment the monoanion exists as a single structure,
with a symmetric H-bond. We conclude that the partial
ordering of this isotropic liquid crystal is sufficiently close to the
order in a crystal, where the H-bond is also symmetric.
1
8
and Figure S3 are assigned to doubly O-labeled difluo-
romaleate monoanion (3- O ), with one O in each car-
boxyl group. The second O in the sample of Figure S3 is a
consequence of the further equilibration of the acid formed
1
8
18
2
1
8
1
2
8
from hydrolysis of the anhydride in H O. However, the
second 18O in the sample of Figure 2 is more problematic
because in excess NaOH the dianion ought to be inert to
further exchange. Consequently, it is still necessary to account
for the second 18O, in order to justify the signal assignment.
We propose the mechanism in Scheme 1 to account for this
Scheme 1. Mechanism for Incorporation of a Second 18O
into Difluoromaleate Dianion (4)
These results are evidence that the symmetry of H-bonds can
be determined by the disorder of the instantaneous solvation
environment. They counter the conclusions that the asymmetry
of an aprotic organic solvent environment is not strong enough
to disrupt the symmetry of the H-bond and that observed
isotope shifts are due to isotopic desymmetrization on an
7
otherwise symmetrical potential-energy surface. That isotopic
desymmetrization ought to have been the same in a liquid
crystal as in an aprotic organic solvent. Yet the observed isotope
shift disappears in the liquid crystal, consistent with an H-bond
that has become symmetric because the partial ordering of this
liquid crystal allows the two donor oxygens to be in the same
environment.
incorporation. It requires dianion 4 to be weakly competitive
with OH− as a nucleophile toward the anhydride (2), a
reactivity that is not impossible.
It is also possible, although half as probable, for both 18O to
appear in the same carboxylate. Upon neutralization with acetic
Would the calculated trajectories of hydrogen motion on a
acid, this isotopomer would show double the Δ of the
symmetrical potential-energy surface again lead to unequal
eq
18
18
18
mono- O-labeled difluoromaleate monoanion (3- O) and
would appear as an AB pattern with twice the separation of the
prominent AB pattern in Figure 2 and with half the total
intensity of the weak singlet of 3- O. Some samples did show
this additional AB pattern.
time-averaged NMR chemical shifts for difluoromaleate- O
1
8
7
monoanion (3- O), as are seen experimentally in Table 2?
Moreover, would the calculated isotope shifts depend on
solvent? Here the distance between the reporter nucleus and
the isotope is one bond greater than that for the ipso carbon in
phthalic acid. This is a challenge to computation.
1
8
H-Bond Symmetry in Crystal. According to the X-ray
structure in Figure 1, the O−O distance in difluoromaleate
monoanion (3) is 2.41 Å, which qualifies it as having a “short”
H-bond. Moreover, structure refinement places the hydrogen
exactly in the middle of the two oxygen atoms, with equal O−H
distances of 1.206 Å. These distances fit exactly the correlation
of OHO distances in so-called “strong” H-bonds, as corrected
CONCLUSIONS
■
The short O−O distance of 2.41 Å and equal O−H distances in
the X-ray structure of potassium hydrogen difluoromaleate
K 3) support a symmetric, short, H-bond in the crystal. The
F NMR spectra of the difluoromaleate- O monoanion (3- O)
+
(
19
18
18
27
for zero-point vibrational effects. Therefore, in this salt, where
the crystalline environment appears to be symmetric, the
H-bond of the monoanion is symmetric. This contrasts with the
X-ray structures of several salts of (±)-α,α′-di-tert-butylsucci-
nate monoanion, where the hydrogen is closer to one oxygen
in water and in two aprotic organic solvents show an AB
pattern, indicating that the two fluorines are inequivalent, with
different NMR chemical shifts. The inequivalence is not seen
1
8
in the diacid (1- O), so it is not due simply to the presence of
28
18
O. Instead the chemical-shift difference is attributed to
than to the other.
1
8
Isotope Shifts and H-Bond Symmetry. Not only in
perturbation of the acidity constant of a carboxylic acid by O
substitution. Therefore, hydrogen difluoromaleate monoanion
3 exists in solution as a pair of interconverting tautomers rather
than as a single symmetric structure. However, the absence of
any isotope shift for 3 in 4-cyanophenyl 4-heptylbenzoate is
taken as evidence for a symmetric H-bond in this isotropic
liquid crystal.
aqueous solution but also in the aprotic organic solvents
1
9
CD CN and CD Cl the F NMR spectrum of difluoromaleate
3
2
2
1
8
18
monoanion (3- O) exhibits an AB spin system. The
O
renders the two fluorines inequivalent, with chemical shifts that
differ by 0.046 or 0.028 ppm. This is adequately close to the
0
.07 ppm predicted by eq 3, which is recognized as an over-
estimate, owing to the crudeness of the models for evaluating
We therefore affirm the role of solvation in determining the
symmetry of H-bonds. In water and in an aprotic organic
solvent, the disorder of the local environment renders the two
oxygens of 3 inequivalent, leading to an asymmetric H-bond.
In contrast, the order in the crystal and in the isotropic liquid
K and D. The chemical-shift difference is not merely due to the
1
8
presence of the O, because it is not seen in the dianion
1
8
18
(
4- O) or in the diacid (1- O), although the spectrum of
1
8
the latter is inconclusive, owing to O exchange with solvent.
7
771
dx.doi.org/10.1021/ja2117848 | J. Am. Chem. Soc. 2012, 134, 7766−7772

Journal of the American Chemical Society
Article
crystal allows the two donor oxygens to be in the same
environment, permitting a symmetric H-bond.
(14) Jacobs, C. A.; Brahms, J. C.; Dailey, W. P.; Beran, K.; Harmony,
M. D. J. Am. Chem. Soc. 1992, 114, 115−121. Abdo, B. T.; Amer, H.;
Banks, R. E.; Brain, P. T.; Cox, A. P.; Dunning, O. J.; Murtagh, V.;
Rankin, D. W. H.; Robertson, H. E.; Smart, B. A. J. Phys. Chem. A
ASSOCIATED CONTENT
Supporting Information
Instrumentation. Diffraction Techniques. Materials. Figure S1:
■
1
(
9
999, 103, 1758−1767.
15) Saunders, M.; Telkowski, L.; Kates, M. J. Am. Chem. Soc. 1977,
9, 8070−8072.
(16) Jameson, C. J.; Osten, H. J. Annu. Rep. NMR Spectrosc. 1986, 17,
*
S
mass spectrum of the diacid isolated from reaction of
1
8
19
difluoromaleic anhydride (2) with H O. Figure S2:
F
1. Hansen, P. E. J. Labelled Compd. Radiopharm. 2007, 50, 967−981.
(17) Ellison, S. L. R.; Robinson, M. J. T. J. Chem. Soc., Chem Commun.
2
1
8
NMR spectrum of mono- O-labeled difluoromaleate dianion
4- O) in D O. Figure S3: F NMR spectrum of a 1:2 mixture
of unlabeled and mono- O-labeled difluoromaleate monoanion
3 + 3- O) in D O. Figure S4: F NMR spectrum of the
18
19
1
983, 745−746. Tanaka, N.; Araki, M. J. Am. Chem. Soc. 1985, 107,
(
2
1
8
7780−7781.
18
19
(18) Siehl, H.-U. Adv. Phys. Org. Chem. 1987, 23, 63−163.
(
2
(
19) Perrin, C. L.; Thoburn, J. D. J. Am. Chem. Soc. 1989, 111, 8010−
tetrabutylammonium salt of a 1:2 mixture of unlabeled and
8
(
012.
20) Smith, A. A.; Drake, M. D.; Rahim, A. K.; Roberts, J. D. J. Phys.
Chem. A 2008, 112, 12367−12371.
21) Raasch, M. S.; Miegel, R. E.; Castle, J. E. J. Am. Chem. Soc. 1959,
1
8
18
mono- O-labeled difluoromaleate monoanion (3 + 3- O) in
1
9
CD Cl . Figure S5: Simulated F NMR spectrum of a 1:2
2
2
1
8
mixture of unlabeled and mono- O-labeled difluoromaleate
monoanion (3 + 3- O) in CD Cl , as in Figure S4, but with 6
(
1
8
2
2
81, 2678.
Hz additional line broadening. Table S1 summarizing crystallo-
graphic data for potassium hydrogen difluoromaleate (K 3).
X-ray crystallographic file in CIF format for potassium hydrogen
difluoromaleate. This material is available free of charge via the
Internet at http://pubs.acs.org.
(22) Okabe, A.; Fukushima, T.; Ariga, K.; Niki, M.; Aida, T. J. Am.
Chem. Soc. 2004, 126, 9013.
+
(23) Ellison, R. D.; Levy, H. A. Acta Crystallogr. 1965, 19, 260.
(24) Haque, R.; Reeves, L. W. J. Phys. Chem. 1966, 70, 2753.
(25) Emsley, J. W.; Phillips, L.; Wray, V. Prog. Nucl. Magn. Reson.
Spectrosc. 1977, 10, 83−756.
(
(
26) Hudlicky, M. J. Fluorine Chem. 1988, 40, 99−108.
27) Limbach, H.-H.; Tolstoy, P. M.; Perez-Hernandez, N.; Guo, J.;
AUTHOR INFORMATION
Corresponding Author
cperrin@ucsd.edu
■
́
́
Shenderovich, I. G.; Denisov, G. S. Isr. J. Chem. 2009, 49, 199.
(28) Perrin, C. L.; Lau, J. S.; Kim, Y.-J.; Karri, P.; Moore, C.;
Rheingold, A. L. J. Am. Chem. Soc. 2009, 131, 13548−13554.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This research was supported by NSF Grant CHE07-42801 to
CLP and by Instrumentation Grant CHE97-09183. A.L.R.
gratefully acknowledges NSF Grant CHE-0634989 for purchase
of the X-ray diffractometer.
REFERENCES
■
(
1) Hibbert, F.; Emsley, J. Adv. Phys. Org. Chem. 1990, 26, 255−379.
Perrin, C. L.; Nielson, J. B. Annu. Rev. Phys. Chem. 1997, 48, 511.
(
(
2) Perrin, C. L. Acc. Chem. Res. 2010, 43, 1550−1557.
3) Perrin, C. L.; Ohta, B. K. J. Am. Chem. Soc. 2001, 123, 6520−
6
526. Perrin, C. L; Ohta, B. K. Bioorg. Chem. 2002, 30, 3−15. Perrin,
C. L; Ohta, B. K. J. Mol. Struct. 2003, 644, 1−12. Perrin, C. L.; Karri, P.
Chem. Commun. 2010, 46, 481−483.
(
4) Perrin, C. L.; Thoburn, J. D. J. Am. Chem. Soc. 1992, 114, 8559−
8565. Perrin, C. L. Science 1994, 226, 1665−1668. Perrin, C. L.;
Nielson, J. B. J. Am. Chem. Soc. 1997, 119, 12734. Perrin, C. L. Pure
Appl. Chem. 2009, 81, 571−583.
(
5) Perrin, C. L.; Lau, J. S. J. Am. Chem. Soc. 2006, 128, 11820−
1
1824.
6) Dopieralski, P.; Perrin, C. L.; Latajka, Z. J. Chem. Theory Comput.
011, 7, 3505−3513.
7) Bogle, X. S.; Singleton, D. A. J. Am. Chem. Soc. 2011, 133,
7172−17175.
8) Saupe, A. Angew. Chem., Int. Ed. 1968, 7, 97−112. Suryaprakash,
N. Concepts Magn. Reson. 1998, 10, 167.
9) Robert, J. C.; Yim, C. J.; Gilson, F. R. Can. J. Chem. 1971, 49,
345.
(
2
(
1
(
(
2
(
(
10) Sun, H.; Fung, B. M. Liq. Cryst. 2000, 27, 755−761.
11) de Lange, C. A.; Burnell, E. E. In NMR of Ordered Liquids;
Burnell, E. E., de Lange, C. A., Eds.; Kluwer: Dordrecht, 2003; pp 3−
6.
12) Chandrasekhar, S. Liquid Crystals; Cambridge University Press:
Cambridge, 1977; p 64. Dutt, G. B. J. Chem. Phys. 2003, 119, 11971.
13) Perrin, C. L.; Dong, Y. J. Am. Chem. Soc. 2007, 129, 4490−4497.
Perrin, C. L.; Flach, A. Angew. Chem. 2011, 50, 7674−7676.
2
(
(
7
772
dx.doi.org/10.1021/ja2117848 | J. Am. Chem. Soc. 2012, 134, 7766−7772