Conformational equilibria in formic acid and the adduct of formic acid and hexafluoroacetone, HCO2C(CF3)2OH

Article abstract of DOI:10.1021/jo061697v

Low-temperature ¹H and ¹³C NMR spectra of formic acid (1) showed separate signals for the E and Z conformations in solvents containing a hydrogen bond acceptor, dimethyl ether. The population of E-1 (6.2% in 3:1:1 CHClF₂/CHCl₂F/(CH₃)₂O) was larger than that for ¹³C-labeled methyl formate in the same solvent (0.2%), which indicated that the relative populations are not determined by steric effects. The free-energy difference between the E and Z conformations of 1 was 0.9 kcal/mol. In a 1:3 CD₂Cl₂/ (CH₃) ₂O solvent mixture, peaks for E and Z conformations were found at low temperatures by ¹H and ¹³C NMR for both formic acid and an adduct with hexafluoroacetone, HCO₂C(CF₃)₂OH (2). The population of E-1 in this solvent mixture was 4.3% by ¹³C NMR. The carbon spectrum showed two peaks in the carbonyl carbon region of nearly equal intensities at -151.6 °C, with E-2 (48%) absorbing downfield of the major Z-2 (52%). The large population of E-2 confirms that electron-withdrawing groups R′ in RCO₂R′ enhance the populations of the E-isomers. The free-energy barriers for 2 of 6.24 (E-to-Z) and 6.26 kcal/·mol (Z-to-E) were determined from rate constants obtained by line shape analysis at -143.2 °C.

Full text of DOI:10.1021/jo061697v

Conformational Equilibria in Formic Acid and the Adduct of
Formic Acid and Hexafluoroacetone, HCO C(CF ) OH
2
3 2
Diwakar M. Pawar, Dalephine Cain-Davis, and Eric A. Noe*
Department of Chemistry, Jackson State UniVersity, P.O. Box 17910, 1400 J. R. Lynch Street,
Jackson, Mississippi 39217-0510
ean12001@yahoo.com
ReceiVed August 15, 2006
Low-temperature H and 13C NMR spectra of formic acid (1) showed separate signals for the E and Z
1
conformations in solvents containing a hydrogen bond acceptor, dimethyl ether. The population of E-1
1
3
(
6.2% in 3:1:1 CHClF
same solvent (0.2%), which indicated that the relative populations are not determined by steric effects.
The free-energy difference between the E and Z conformations of 1 was 0.9 kcal/mol. In a 1:3 CD Cl
O solvent mixture, peaks for E and Z conformations were found at low temperatures by H and
C NMR for both formic acid and an adduct with hexafluoroacetone, HCO C(CF OH (2). The population
2 2 3 2
/CHCl F/(CH ) O) was larger than that for C-labeled methyl formate in the
2
2
/
1
3 2
(CH )
13
2
3 2
)
1
3
of E-1 in this solvent mixture was 4.3% by C NMR. The carbon spectrum showed two peaks in the
carbonyl carbon region of nearly equal intensities at -151.6 °C, with E-2 (48%) absorbing downfield of
the major Z-2 (52%). The large population of E-2 confirms that electron-withdrawing groups R′ in RCO R′
2
enhance the populations of the E-isomers. The free-energy barriers for 2 of 6.24 (E-to-Z) and 6.26 kcal/
mol (Z-to-E) were determined from rate constants obtained by line shape analysis at -143.2 °C.
Introduction
Dynamic NMR spectroscopy is often the most useful
experimental method for studying conformational equilibria in
solution. Rates and barriers for interconversion of conformations,
in addition to the populations and free-energy differences, can
generally be accurately determined if this method is applicable.
Esters, thiolesters, and amides have been studied by dynamic
1
NMR, but carboxylic acids are an important class of compounds
conformations were later observed for thioformic acid3 at
113 °C, but OH exchange in a carboxylic acid is generally a
faster process than SH exchange in a thioacid.
for which E and Z conformations have not been reported to be
observed by this method. This is due to rapid interconversion
of E and Z conformations, shown for formic acid (1), by
intermolecular exchange of the acidic proton. Separate NMR
signals for E- and Z-isomers of thioacetic acid in a ratio of 1:3
-
Formic acid in the gas phase includes monomer and the cyclic
dimer below, and the amount of cyclic dimer increases in the
4
2
liquid phase. The enthalpy change for dissociation of the dimer
were found at -150 °C, and nearly equal amounts of these
5
into two molecules of monomer is about 14 kcal/mol. In the
solid phase, the compound exists as long chains of monomer,
(
1) For example, see: Pawar, D. M.; Khalil, A. A.; Hooks, D. R.; Collins,
K.; Elliott, T.; Stafford, J.; Smith, L.; Noe, E. A. J. Am. Chem. Soc. 1998,
20, 2108 and references cited therein.
2) Noe, E. A. J. Am. Chem. Soc. 1977, 99, 2803.
1
(3) Kalinowski, H. O.; Hocking, W. H.; Winnewisser, B. P. J. Chem.
Res. 1978, 7, 260.
(
1
0.1021/jo061697v CCC: $37.00 © 2007 American Chemical Society
Published on Web 02/16/2007
J. Org. Chem. 2007, 72, 2003-2007
2003

Pawar et al.
The E-Z energy differences for formic acid esters decrease
1,15
in the order R′ ) methyl > vinyl > ethynyl. The hybridiza-
tion of carbon in the corresponding hydrocarbons R′H changes
3
2
from sp to sp to sp in the series, suggesting that the electron-
withdrawing abilities of the groups may be important in
enhancing the populations of the E-isomers. However, even with
very highly electronegative R′ groups such as trifluoromethyl,16
the Z conformations of esters are still generally preferred,
although by smaller amounts. The calculated dipole moments
for E- and Z-trifluoromethyl formate were 2.17 and 2.47 D at
the MP2/6-311G(df,pd) level, and the E-isomer was calculated
to have a free energy of 1.14 kcal per mol, relative to the Z.
Although the difference in dipole moments was calculated to
with individual molecules of the chain held together by CH‚‚‚O
6
and OH·‚‚O hydrogen bonds. Most carboxylic acids exist in
the solid state as cyclic hydrogen-bonded dimers, with structures
analogous to the one shown above for formic acid. A solid-
7
state T1 NMR spin-lattice relaxation study of p-toluic acid-
d7, completely deuterated except for the acidic proton, found
an apparent activation energy of only 1.15 kcal/mol for the
double proton transfer in this acid. Tunneling contributes
significantly to this process, particularly at low temperatures.
For observation of the E-isomer of a carboxylic acid, the
cyclic dimer needs to be avoided or minimized, as the
conformation is necessarily Z in the dimer, and intermolecular
exchange of the acidic proton is rapid in the dimer. Addition
of weak bases can disrupt the cyclic dimer by hydrogen bonding
1
7
be close to zero, with a slightly more favorable (lower) value
1
6
for the E-isomer, the Z-isomer was predicted to predominate.
In the present work, H and 13C NMR spectra of 1 were taken
for solutions containing dimethyl ether for hydrogen bonding
to the acid. Hexafluoroacetone was added to some samples to
react with the water and form the hydrate. Formic acid also
forms an adduct with hexafluoroacetone. E and Z conformations
of 1 were observed without the hexafluoride, but a lower
temperature was needed. In addition to lowering the amount of
water, formation of the adduct will lower the concentration of
the acid, which will tend to raise the coalescence temperature.
Decoalescence was also observed for the signals of the adduct.
1
1
to the OH proton, and H NMR spectra of 1 in CHClF2 with
concentrations of 0.02-0.05 M and a (5-8)-fold excess of THF
or HMPA showed doublets (J ) 11.6 Hz) for each of the protons
8
of the acid at temperatures from -163 to -123 °C. The
coupling constant observed is consistent with the trans relation-
ship between hydrogens of the Z conformation. Cis and trans
three-bond H-H coupling constants of ∼0.7 and 10.0 Hz were
9
found for the E,Z conformation of diformamide, (HCO)2NH.
In the absence of a hydrogen-bonding solvent, a triplet was
found for the OH proton of 1 at low temperatures, and rapid
double proton exchange within the dimer was assumed to
Results and Discussion
8
1
Low-temperature H NMR spectra of a solution containing
account for the averaged coupling (JH-H ) 5.9 Hz, average of
2% formic acid and ca. 3% hexafluoroacetone in 1:3 CD2Cl2/
dimethyl ether are shown in Figure 1, and larger spectra are
shown in the Supporting Information. The proton NMR
spectrum at -90.0 °C shows the expected four peaks for formic
acid and adduct in the region of δ 11 to 7, as shown in Table
1. Both signals for formic acid decoalesce at lower temperatures
into separate peaks for E- and Z-isomers, and the proton
chemical shifts at -118.3 °C are summarized in Table 1. As
expected, the signals for OH protons moved to lower field with
decreasing temperature. For 1, assignments are based on
intensities, with the Z-isomer expected to predominate, based
1
1.8 and 0.0 Hz).
Signals for the E-isomer of 1 were observed in the gas phase
by microwave spectroscopy,10 and this conformation was found
to be 3.9 kcal/mol higher in energy than Z-1. A barrier of 13.8
kcal/mol was estimated10 for conversion of the Z conformation
to the E. Ab initio calculations at the MP3/6-311+G**//6-31G*
level11 predict E-Z energy differences of 4.61 and 5.59 kcal/
mol for formic acid and methyl formate, respectively, and the
Z-to-E barrier for formic acid was calculated to be 12.4 kcal/
mol. The dipole moments10 of E and Z formic acid are 3.79
and 1.42 D, and the energy difference between these conforma-
tions is expected to decrease in solution. A free-energy
difference of 2.15 kcal/mol was reported12 for methyl formate
in 1:1 acetone-d6/DMF, and the E-Z free-energy difference for
this ester was predicted13 by ab initio calculations at the MP2/
11
10
on the results of gas-phase calculations, the microwave study,
and the effect of solvent on the E-Z free-energy difference for
13
methyl formate. The formyl protons of E-1 and Z-1 have the
expected relative positions, but they are reversed for the carbonyl
carbon peaks of 1, although the Z-E shift difference is small.
At still lower temperatures, both the OH and the CH signals of
the adduct broadened, and separate signals for E and Z
conformations were observed by -153.8 °C. The chemical shifts
are listed in Table 1. Assignments to E and Z conformations
are based on the assumption that the formyl hydrogen and
carbonyl carbon peaks (see below) of E-2 will be downfield of
the corresponding peaks for Z-2, as is often observed for esters
6
-31+G*//6-31G* level to decrease from 5.16 kcal/mol in the
gas phase to 1.66 kcal/mol for a solution in acetonitrile.
Evidence for observable amounts of E-1 in solution has been
1
4
obtained from the vibrational spectra of formic acid and
HMPA in several solvents.
(
4) Minary, P.; Jedlovszky, P.; Mezei, M.; Turi, L. J. Phys. Chem. B
2
000, 104, 8287. An earlier structural investigation of formic acid by X-ray
1
and related compounds. The chemical shift for the OH protons
and neutron diffraction and reverse Monte-Carlo study found that no
significant amount of cyclic hydrogen bonded dimer was present. Jed-
lovszky, P.; Bako, I.; Palinkas; Dore, J. C. Mol. Phys. 1995, 86, 87.
(14) Golubev, N. S.; Pushkareva, E. G. Vestn. Leningr. UniV., Ser. 4,
Fiz., Khim. 1985, 2, 87.
(15) Langley, C. H.; Pawar, D. M.; Noe, E. A. THEOCHEM 2005, 732,
99.
(16) Pawar, D. M.; Sims, Y. S.; Moton, D. M.; Noe, E. A. THEOCHEM
2003, 626, 159.
(
(
(
(
5) Kim, Y. J. Am. Chem. Soc. 1996, 118, 1522.
6) Nahringbauer, I. Acta Crystallogr. 1978, B34, 315.
7) Meier, B. H.; Graf, F.; Earnst, R. R. J. Chem. Phys. 1982, 76, 767.
8) Golubev. N. S.; Denisov, G. S.; Koltsov, A. I. J. Mol. Struct. 1981,
7
5, 333.
(
(
(
(
(
9) Noe, E. A.; Raban, M. J. Am. Chem. Soc. 1975, 97, 5811.
10) Hocking, W. H. Z. Naturforsch. A 1976, 31, 1113.
11) Wiberg, K. B.; Laidig, K. E. J. Am. Chem. Soc. 1987, 109, 5935.
12) Grindley, T. B. Tetrahedron Lett. 1982, 23, 1757.
13) Wiberg, K. B.; Wong, M. W. J. Am. Chem. Soc. 1993, 115, 1078.
(17) Dipole moments are often assumed to be a measure of dipole-
dipole interactions, but Perrin and Young18 have shown that this is not
always valid, and have described two examples for which this assumption
leads to the wrong conclusion. However, the relationship was described as
applicable to esters.
2004 J. Org. Chem., Vol. 72, No. 6, 2007

Conformational Equilibria in Formic Acid
TABLE 1. Chemical Shifts (in ppm) Obtained from Low-Temperature H and 13C NMR Spectra of a Solution Containing 2% Formic Acid
1
and Ca. 3% Hexafluoroacetone in 1:3 CD2Cl2, (CH3)2O (Internal Reference TMS)
chemical shifts (ppm)
proton
carbon
formic acid
adduct
formic acid
OH
CH
OH
CH
CdO
adduct CdO
temp (°C)
90
E
Z
E
Z
E
Z
E
Z
E
Z
E
Z
-
10.87
7.93
9.18
9.77
8.42
8.58
-118.2
-120.0
-151.6
-153.8
11.75
11.45
8.46
7.98
163.34
163.76
158.52
159.33 158.70
9.85
9.90
9.02
8.20
1
FIGURE 1. Low-temperature H NMR spectra of a mixture of 2% formic acid and 3% hexafluoroacetone in 1:3 CD
2
Cl
2
/dimethyl ether.
J. Org. Chem, Vol. 72, No. 6, 2007 2005

Pawar et al.
δ 11.77 and 12.22 at -107 °C for the Z- and E-isomers,
respectively. Integration of the spectrum gave populations of
0
.938 and 0.062. These values are somewhat larger than those
observed for 1:3 CD2Cl2/(CH3)2O as solvent, as noted above.
The higher population for this solution can be attributed to a
higher dielectric constant due to the smaller amount of dimethyl
ether and/or the higher temperature. The free-energy barriers
for methyl formate at -70.8 °C in 1:1 acetone-d6/DMF are 9.85
2
0
and 11.91 kcal/mol, and the population for the E-isomer of
this ester at -82.5 °C in 3:1:1 CHClF2/CHCl2F/dimethyl ether
is 0.002. On the basis of the barriers for methyl formate, a
coalescence temperature of g-90.3 °C is reasonable for
intramolecular exchange in 1. As long as exchange of the OH
proton between the E and Z conformations of 1 is slow, it is
not necessary for the corresponding E-E and Z-Z exchanges
to also be slow to see separate signals for the E and Z
conformations, although we expect E-E exchange to also be
slow (see below). The absence of splitting at -107 °C for the
OH signal of Z-1 at δ 11.77 indicates that exchange between Z
conformations is still rapid at this temperature, but the absence
of splitting for the E-isomer does not provide any information
about the rate of E-E exchange because the coupling constant
FIGURE 2. Low-temperature 13C NMR spectrum for the carbonyl
region of a mixture of formic acid and hexafluoroacetone in 1:3 CD
Cl /dimethyl ether.
2
-
2
8,9
is expected to be close to zero. We believe that Z-Z exchange
occurs by way of the cyclic hydrogen-bonded dimer after loss
of the solvating dimethyl ether. This pathway is not available
for E-E and E-Z exchange, and the barriers for interconversion
of the E and Z conformations by rotation about the C-OH bond
are then responsible for the slower E-E and E-Z exchanges.
To obtain further information about the rate of proton
exchange between Z-isomers, a 0.5% solution of a mixture of
formic acid and acetic acid in a ratio of 1:1 by volume was
studied at low temperatures in a 3:1:1 mixture of CHClF2/
CHCl2F/(CH3)2O containing a small amount of acetic anhydride.
No hexafluoroacetone was present. By approximately -144 °C,21
a small peak for E-1 and a large averaged OH signal for Z-1
and acetic acid were observed at δ 13.5 and 12.0, respectively,
and the latter peak split into two broad peaks at δ 12.2 and
11.5 by about -173 °C, as exchange between Z-1 and acetic
acid becomes slow. This experiment provides information about
the temperature required for the slow exchange between
molecules of Z-formic acid under these conditions. Slowing Z-Z
exchange probably requires slow dissociation of the hydrogen
bonds between the acids and dimethyl ether, and this requires
a lower temperature than for E-Z exchange, which requires
rotation about the C-OH bond of the acid for conversion of E
to Z, followed by rapid proton exchange by way of the dimer.
With the large ratio of dimethyl ether to carboxylic acids used,
nearly all of the acids are expected to be monomeric and
hydrogen bonded to the dimethyl ether.
FIGURE 3. Low-temperature 13C NMR spectrum for the carbonyl
region of the adduct, HCO C(CF OH, for a solution of 2% formic
acid and 3% hexafluoroacetone dissolved in 1:3 CD Cl /(CH O.
2
3 2
)
2
2
3 2
)
of the hydrate, (CF3)2C(OH)2, was δ 7.23 and 7.85 at -90 and
-
118.2 °C, respectively.
The carbon spectrum for the carbonyl region at -120 °C is
shown in Figure 2. The peaks for the Z-1 and E-1 isomers appear
at δ 163.76 and 163.34, respectively. The peak at δ 158.52 is
assigned to the adduct. At -120.0 °C, the population of E-1
1
3
from the C signals was 4.3%, and at -118.2 °C, the
1
populations of this conformation from the H spectrum were
4
.5% (OH) and 4.2% (CH). Decoalescence of the carbonyl
carbon peak of 2 at lower temperatures is shown in Figure 3.
The chemical shifts at -151.6 °C were δ 159.33 (E) and 158.70
(Z).
Assignments of peaks in Figure 3 to conformations were
based on the assumption that E absorbs downfield of Z, as found
1
previously for formate esters, but the chemical-shift difference
is small, and the assignments could be reversed. Electronic
A 2% formic acid solution in 3:1:1 CHClF2/CHCl2F/(CH3)2O,
but without hexafluoroacetone, also showed peaks for the E and
Z conformations in the OH region at about -144 °C. The
presence of hexafluoroacetone does not markedly change the
1
3
integration of the C spectrum at -151.6 °C gave populations
of 48% and 52% for the E and Z conformations, respectively.
The finding of a large population for E-2 indicates that the
electron-withdrawing groups tend to favor the E-isomer. A first-
-
1
order rate constant of 84.6 s was obtained by line shape
matching program1 for the E to Z conversion at the coalescence
temperature of -143.2 °C (Figure 3), corresponding to a free-
energy barrier of 6.24 kcal/mol. The rate constant for the reverse
(18) Perrin, C. L.; Young, D. B. Tetrahedron Lett. 1995, 36, 7185.
9
(19) Calculated spectra were obtained with the program DNMR-SIM
written by G. Hegele and R. Fuhler; Heinrich-Hein University, Dussel-
dorf Institute of Inorganic and Structural Chemistry: Dusseldorf, FRG,
1
994.
20) Cain, D.; Pawar, D. M.; Stewart, M.; Billings, H., Jr.; Noe, E. A. J.
Org. Chem. 2001, 66, 6092.
21) The temperatures for this experiment were not calibrated promptly
-
1
process was 84.6 × (48/52) ) 78.1 s , and the Z-to-E free-
(
energy barrier at this temperature was 6.26 kcal/mol.
(
For a solution containing 2% formic acid and 3% hexafluo-
roacetone in 3:1:1 CHClF2/CHCl2F/(CH3)2O, coalescence was
observed at -90.3 °C, and the OH proton signals absorbed at
after the spectra were taken, and were estimated from many other
calibrations. The error limits (ca. (5 °C) are larger than those for the other
spectra.
2006 J. Org. Chem., Vol. 72, No. 6, 2007

Conformational Equilibria in Formic Acid
populations of the E and Z conformations for 1, but does raise
the coalescence temperature, probably by removal of water and
also by decreasing the concentration of the acid by formation
roacetone in 1:3 CD
hydrogen-bonding solvent. The NMR samples were prepared in
-mm thin-walled screw-capped NMR tubes, and a small amount
2 2 3 2
Cl /(CH ) O. Dimethyl ether was used as a
5
of TMS was added for an internal reference. Caution: High
pressure. The sample tubes were stored and handled below 0 °C
most of the time. Spectra were recorded on a wide-bore NMR
spectrometer operating at 300.52 MHz for proton and 75.57 MHz
of the adduct. E-Z free-energy differences of 2.15, 1.67, 1.36,
and 0.48 kcal/mol were reported12 for solutions of methyl, ethyl,
isopropyl, and tert-butyl formate, and the trend is expected,
based on the sizes of groups attached to oxygen. The finding
of a larger population for the E-isomer of formic acid than for
methyl formate indicates that the relative populations are not
determined by steric effects.
1
13
for carbon. The H and C spectra were obtained with a 5-mm
dual probe. For proton spectra, a pulse width of 8.2 µs, corre-
sponding to a tip angle of 83°, and a pulse repetition period of 3 s
were used. A sweep width of (3200 Hz, data size of 64 K, and
800 pulses were used. Carbon spectra were recorded with a line
broadening of 3.0 Hz. Spinning was discontinued below about
Conclusions
-
140 °C.
Low-temperature H and 13C NMR spectra of solutions of
formic acid (1) showed separate signals for the E and Z
conformations. Dimethyl ether was used to hydrogen bond with
the acid and break up the cyclic dimer. A 2% solution of 1 in
1
Because of the difficulty in ejecting the samples at lower
temperatures, the temperature calibrations were performed sepa-
rately, using a copper-constantan thermocouple immersed in the
same solvents contained in a nonspinning sample tube and under
conditions as nearly identical as possible. The emf’s were measured
with a potentiometer. The uncertainty in temperatures was estimated
to be (2 °C.
1
:3 CD2Cl2/(CH3)2O, which also contained 3% hexafluoroac-
etone, gave populations of 4.3%, 4.5%, and 4.2% E-1 from
integration of the signals from the carbonyl carbons, hydroxyl
protons, and formyl protons of 1 at -120 °C (carbon) or
Acknowledgment. We thank the National Institutes of
Health for financial support (NIH-SCORE grant no. S06GM-
-
118 °C (proton). The population of E-1 in 3:1:1 CHClF2/
1
3
CHCl2F/(CH3)2O was larger (6.2%) than that for C-labeled
methyl formate (0.2%), which indicated that the relative
populations are not determined entirely by steric effects. The
presence of electron-withdrawing groups on the ether oxygen
of ester 2, HCO2C(CF3)2OH, markedly increased the population
of the E-isomer (48%).
0
084047).
Supporting Information Available: The NMR spectra for (1)
a mixture of formic acid and acetic acid in a ratio of 1:1 by volume
in a 3:1:1 mixture of CHClF /CHCl F/(CH O containing a small
2
2
3 2
)
1
amount of acetic anhydride, (2) low-temperature H NMR spectra
of 2% formic acid + 3% hexafluoroacetone in 3:1 dimethyl ether/
dichloromethane solvent mixture, for eight temperatures, (3) low-
temperature 13C NMR spectra of 2% formic acid + 3% hexaflu-
oroacetone in 3:1 dimethyl ether/dichloromethane solvent mixture.
This material is available free of charge via the Internet at
http://pubs.acs.org.
Experimental Section
According to the label, the formic acid contained about 2% or
less water, which was not removed before use. Three solvent
systems were used: (1) 2% formic acid + 3% hexafluoroacetone
in 3:1:1 CHClF
2
/CHCl
2
F/(CH
3 2
) O; (2) 2% formic acid in 3:1:1
CHClF /CHCl F/(CH
2
2
3 2
) O; and (3) 2% formic acid + 3% hexafluo-
JO061697V
J. Org. Chem, Vol. 72, No. 6, 2007 2007