Secondary deuterium isotope effects on the acidity of carboxylic acids and phenols

Article abstract of DOI:10.1021/ja069103t

Secondary deuterium isotope effects (IEs) on acidities have been accurately measured by an NMR titration method applicable to a mixture of isotopologues. Deuteration definitely decreases the acidity of carboxylic acids and phenols, by up to 0.031 in the ΔpK per D. For aliphatic acids, the IEs decrease as the site of deuteration becomes more distant from the OH, as expected, but a surprising result is that IEs in both phenol and benzoic acid do not decrease as the site of deuteration moves from ortho to meta to para. The experimental data are supported by ab initio computations, which, however, substantially overestimate the IEs. The discrepancy does not seem to be due to solvation. The IEs originate in isotope-sensitive vibrations whose frequencies and zero-point energies are lowered upon deprotonation. In the simplest case, formate, the key vibration can be recognized as the C-H stretch, which is weakened by delocalization of the oxygen lone pairs. For the aromatic acids, delocalization cannot account for the near constancy of IEs from ortho, meta, and para deuteriums, but the observed IEs are consistent with calculated vibrational frequencies and electron densities. Moreover, the ability of the frequency analysis to account for the IEs is evidence against an inductive origin.

Full text of DOI:10.1021/ja069103t

Published on Web 03/15/2007
Secondary Deuterium Isotope Effects on the Acidity of
Carboxylic Acids and Phenols
Charles L. Perrin* and Yanmei Dong
Contribution from the Department of Chemistry 0358, UniVersity of CaliforniasSan Diego, La
Jolla, California 92093-0358
Received December 19, 2006; E-mail: cperrin@ucsd.edu
Abstract: Secondary deuterium isotope effects (IEs) on acidities have been accurately measured by an
NMR titration method applicable to a mixture of isotopologues. Deuteration definitely decreases the acidity
of carboxylic acids and phenols, by up to 0.031 in the ∆pK per D. For aliphatic acids, the IEs decrease as
the site of deuteration becomes more distant from the OH, as expected, but a surprising result is that IEs
in both phenol and benzoic acid do not decrease as the site of deuteration moves from ortho to meta to
para. The experimental data are supported by ab initio computations, which, however, substantially
overestimate the IEs. The discrepancy does not seem to be due to solvation. The IEs originate in isotope-
sensitive vibrations whose frequencies and zero-point energies are lowered upon deprotonation. In the
simplest case, formate, the key vibration can be recognized as the C-H stretch, which is weakened by
delocalization of the oxygen lone pairs. For the aromatic acids, delocalization cannot account for the near
constancy of IEs from ortho, meta, and para deuteriums, but the observed IEs are consistent with calculated
vibrational frequencies and electron densities. Moreover, the ability of the frequency analysis to account
for the IEs is evidence against an inductive origin.
Introduction
In 1963, Streitwieser and Klein reported classic measurements
of the deuterium IEs on the acidity constants Ka of some
Isotope effects (IEs) are informative features of rates and
equilibria.1 Primary IEs are those where a bond to the isotope
is broken, in contrast to secondary IEs, where that bond remains
intact. Kinetic IEs continue to provide valuable insight into
mechanisms of organic,2 bioorganic,3 and enzymatic reactions.4
The best authenticated secondary IEs are kinetic IEs in SN1
reactions,5 where a carbon changes its hybridization from sp3
to sp2. This change is associated with a lowering of the C-H
out-of-plane bending frequency ν, which reduces the zero-point
energy ()1/2hν) in the transition state. The reduction for C-D
is smaller because zero-point energy is inversely proportional
to the square root of mass. Because the reduction for the protium
substrate is greater, it reacts faster.
carboxylic acids.6 These are equilibrium effects, not kinetic. The
acidity constants were measured by conductivities of the purified
acids, so they could be quite precise, unless there were a
systematic error due to an impurity in one of the samples. Table
1 lists the ∆pKa between some acids and their isotopologues
(species that differ in the number of isotopic substituents). The
data show that deuteration reduces the acidity, although the
reduction tends to decrease as the distance from the carboxyl
group increases, especially if expressed on a per-deuterium basis.
These cases differ from solvolyses in that there is no
rehybridization upon deprotonation. This cannot be a simple
inductive effect, whereby protium is more electron-withdrawing
than deuterium, because the Born-Oppenheimer Approximation
guarantees that the electronic wave function is independent of
nuclear mass.7 Besides, the decrease of acidity is opposite to
deuterium’s electron-withdrawing power in solvolyses.
(1) (a) Melander, L.; Saunders, W. H. Reaction Rates of Isotopic Molecules;
Wiley: New York, 1980. (b) Carroll, F. A. PerspectiVes on Structure and
Mechanism in Organic Chemistry: Brooks/Cole: Pacific Grove, CA, 1998,
pp 349-365. (c) Anslyn, E. V.; Dougherty, D. A. Modern Physical Organic
Chemistry; University Science Books: Sausalito, CA, 2006, p 421ff.
(2) (a) Lloyd-Jones, G. C.; Alder, R. W.; Owen-Smith, G. J. J. Chem.sEur. J.
2006, 12, 5361. (b) Fang, Y. R.; MacMillar, S.; Eriksson, J.; Kolodziejska-
Huben, M.; Dybala-Defratyka, A.; Paneth, P.; Matsson, O.; Westaway, K.
C. J. Org. Chem. 2006, 71, 4742. (c) Baldwin, J. E.; Singer, S. R. Can. J.
Chem. 2005, 83, 1510. (d) Hengge, A. C.; Onyido, I. Curr. Org. Chem.
2005, 9, 61.
For formic acid, Streitwieser and Klein accepted an origin
arising from zero-point energies, based on the agreement
between the observed IE and the IE calculated from the
-
experimental vibrational frequencies of HCOOH, HCO2
,
DCOOH, and DCO2-. For other acids, they invoked an inductive
effect on pK arising from an electrostatic interaction between
the negative charge of the carboxylate and the dipole moment,
µ, of a C-D bond, relative to a C-H bond. Since the dipole
moment is the product of charge separation and bond length
(3) Lewis, B. E.; Choytun, N.; Schramm, V. L.; Bennett, A. J. J. Am. Chem.
Soc. 2006, 128, 5049.
(4) (a) Pudney, C. R.; Hay, S.; Sutcliffe, M. J.; Scrutton, N. S. J. Am. Chem.
Soc. 2006, 128, 14053. (b) Francis, K.; Gadda, G. Biochemistry 2006, 45,
13889. (c) Cleland, W. W. Arch. Biochem. Biophys. 2005, 433, 2. (d) Pu,
J. Z.; Ma, S. H.; Garcia-Viloca, M.; Gao, J. L.; Truhlar, D. G.; Kohen, A.,
J. Am. Chem. Soc. 2005, 127, 14879. (e) Vocadlo, D. J.; Withers, S. G.
Biochemistry 2005, 44, 12809.
(5) Streitwieser, A., Jr.; Jagow, R. H.; Fahey, R. C.; Suzuki, S. J. Am. Chem.
Soc. 1958, 80, 2326.
(6) Streitwieser, A., Jr.; Klein, H. S. J. Am. Chem. Soc. 1963, 85, 2759.
(7) Weston, R. E., Jr. Tetrahedron 1959, 6, 31.
9
4490
J. AM. CHEM. SOC. 2007, 129, 4490-4497
10.1021/ja069103t CCC: $37.00 © 2007 American Chemical Society

IEs on the Acidity of Carboxylic Acids and Phenols
A R T I C L E S
Table 1. Secondary Deuterium Isotope Effects on the Acidity of
Carboxylic Acids
Because of these questions, it is important that the experi-
mental data be incontrovertible. They do differ from earlier
values for acetic acid.15 Subsequent measurements, though, have
confirmed the value for formic acid16 and for other acids.17
These rely on NMR titrations that accurately fit the variation
of chemical shifts with the number of equivalents of base added.
The values obtained are included in Table 1. Although there
remains no uncertainty about these IEs, we have undertaken to
remeasure them, along with some additional IEs, by yet another
method.
a
acidD
pKD
−
pKH
pKD − pKH
DCO2H
CD3CO2H
(CD3)3CCO2H
2,6-C6H3D2CO2H
C6D5CO2H
0.030 ( 0.004
0.014 ( 0.001
0.018 ( 0.001
0.003 ( 0.001
0.010 ( 0.002
0.0342b
0.0134 ( 0.0005c
0.0040 ( 0.0002c
0.0099 ( 0.0001c
a Ref 6. b Ref 16. c Ref 17.
and since the C-H bond length is longer than that of C-D,
owing to anharmonicity, deuterium might be electron-donating.
Such an inductive effect remains consistent with the Born-
Oppenheimer Approximation. Indeed, the Mulliken charges on
the oxygens of the formate anion are calculated to be -0.729
in HCO2- and -0.731 in DCO2-, arising via the shorter C-D
NMR titration methods make it possible to measure relative
acidities with great precision. The procedure involves successive
additions of small aliquots of base to a mixture of acids. The
stronger acid will be deprotonated first. Its chemical shift will
then move ahead of that of the less acidic one, which lags
behind. The acidity constants, Ka, and chemical shifts, δ, of
both H and D acids can be related through eq 2,18 where δ0 or
δ- are for the acid or conjugate base, respectively, as measured
at the beginning or end of the titration. Therefore, a plot of the
bond.8 Such a greater negative charge might be sufficient to
-
account for a greater basicity for DCO2
.
Equation 1 expresses this inductive effect on pK. The
derivative, ∂pK/∂µ, was estimated from the effect of a C-Cl
dipole on acidity using the difference in the pKs of trichloro-
acetic acid (0.63) and acetic acid (4.75) and the difference in
the dipole moments of tert-butyl chloride (2.13 D) and isobutane
(-0.13 D). Next, ∆µ was estimated as 0.0086D, the difference
between the dipole moments of (CH3)3CD and (CH3)3CH.9
Then, eq 1 gave a ∆pK per D of 0.005, in excellent agreement
with the observed 0.014 for acetic-d3 acid.
0
quantity on the left versus (δH - δH-)(δD - δD) should be
h
linear, with a zero intercept. The ratio of acidity constants, Ka /
d
Ka , can then be evaluated as the slope of that plot.
(δH0 - δH)(δD - δD-) ) (Kah/Kad)(δH - δH-)(δD0 - δD)
(2)
This method is capable of exquisite precision because it is
based only on chemical-shift measurements, not on pH or
volume or molarity or equivalents of base added, as is usual in
pH titrations. The method is comparative. If there is no
difference in acidities, there is no lag of one chemical shift
behind the other. Therefore, minute imbalances of acidities can
be detected. Moreover, since the titration is performed on a
mixture of the two acids under conditions guaranteed identical
for both, it avoids systematic error due to impurities. A further
advantage is that it can be applied in any solvent, even the
nonaqueous ones here, where a pH electrode would be inopera-
tive. In addition to the above NMR titrations that fit to the
number of equivalents of base added,16,17 a variant method relied
on fitting 19F NMR chemical shifts to pH,19 but neither molarity
nor pH is as reliably measured as chemical shifts. Besides, none
of these utilizes the advantage of data analysis by linear least
squares. We now report IEs on acidities of a series of carboxylic
acids (1-4, 8-9) and phenols (5-7) measured by 1H, 13C, and
19F NMR titration using reporter nuclei indicated in boldface
in the structures.
∂pK
∂µ
∆pK ∆µ
/ ∆µ )
∆pK )
∆µ )
∆nCl ∆nCl
1 4.75 - 0.63
3 2.13 + 0.13
0.0086 ) 0.005 (1)
This estimate depends crucially on that difference (0.0086
D) between the dipole moments of isobutane and isobutane-d,
which is unusually large, amounting to 7% of their total dipole
moment. For comparison, the difference between HCl and DCl
is only 0.005 ( 0.002 D, or ∼0.5% of the total dipole moment
(1.08 D)10 of their much more polar bond. Indeed, it was
recognized that the observed difference in isobutane is rather
high and may be due to contributions from bending modes. An
alternative estimate of ∆µ is 1 × 10-4 D, based on dCH - dCD
) 0.5 pm, due to anharmonicity (for a Morse potential with a
dissociation energy of 100 kcal/mol)11 and a derivative of dipole
moment with respect to C-H distance of 0.004e from the
infrared intensities of methane.12 Application to eq 1 then leads
to a ∆pK per D of 0.00006, which is 2 orders of magnitude
lower than either the earlier estimate or the observed IE.
Further doubts have been expressed about the necessity of
invoking an inductive effect. The IE in formic acid agrees with
the totality of experimental vibrational frequencies and zero-
point energies.13 Likewise, the IE in acetic acid agrees with the
vibrational force field, but it is very sensitive to the choice of
force constants.14
Besides the experimental measurements of IEs, we have
turned to computation in order to help clarify their origin. In
1963, Streitwieser and Klein were not optimistic about such
detailed analysis for acids less simple than formic. Calculations
have succeeded in reproducing experimental secondary deuter-
ium kinetic IEs in the formation of several norbornyl cations.20
We now report computed IEs for these carboxylic acids and
phenols. However, they exceed the observed ones by a factor
(8) Reyes, A.; Pak, M. V.; Hammes-Schiffer, S. J. Chem. Phys. 2005, 123,
064104.
(15) Halevi, E. A.; Nussim, M.; Ron, A. J. Chem Soc. 1963, 866.
(16) Ellison, S. L. R.; Robinson, M. J. T. J. Chem. Soc., Chem. Commun. 1983,
745.
(9) Lide, D. R. J. Chem. Phys. 1960, 33, 1314.
(10) Bell, R. P.; Coop, I. E. Trans. Faraday Soc. 1958, 34, 1209.
(11) Bell, R. P.; Guggenheim, E. A. Trans. Faraday Soc. 1936, 32, 1013.
(12) Oliveira, A. E.; Guadagnini, P. H.; Custo´dio, R.; Bruns, R. E. J. Phys.
Chem. A 1998, 102, 4615.
(13) Bell, R. P.; Crooks, J. E. Trans. Faraday Soc. 1962, 68, 1407.
(14) Bron, J. J. Chem. Soc., Faraday Trans. 2 1975, 71, 1772.
(17) Pehk, T.; Kiirend, E.; Lippmaa, E.; Ragnarsson, U.; Grehn, L. J. Chem.
Soc., Perkin Trans. 2 1997, 445.
(18) (a) Perrin, C. L.; Fabian, M. A.; Armstrong, K. B. J. Org. Chem. 1994, 59,
5246. (b) Perrin, C. L.; Fabian, M. A. Anal. Chem. 1996, 68, 2127.
(19) Forsyth, D. A.; Yang, J.-R. J. Am. Chem. Soc. 1986, 108, 2157.
(20) Muchall, H. M.; Werstiuk, N. H. Can. J. Chem. 1998, 76, 1926.
9
J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007 4491

A R T I C L E S
Perrin and Dong
NMR Titrations. The NMR sample prepared for each titration
contained 0.5-0.7 mL of D2O or organic solvent and appropriate
concentrations of internal standard and isotopologues of the acid,
adjusted to ensure nearly equal peak heights. Unequal heights facilitate
continuity of peak assignments, especially in case of a crossover during
the titration.
Tables S1 and S2 list the initial concentrations of all substances in
each sample. Tables S3 and S4 list the base used for each titration and
the starting and ending chemical shifts for deuterated and nondeuterated
acids. For all acids, deuterium produces an upfield shift, as is usual
for isotope shifts,21 except for acetic-d3 in D2O and THF, for aqueous
4-fluorophenol, where no isotope shift is resolvable, and for 4-fluoro-
2-methylphenol, where the isotope shift of meta-CD3 is not resolvable
in the phenoxide. Another exception is that deuterium produces a
downfield shift in 4-fluorophenol in CD3CN, but the signals cross and
recross during the titration. All acids move upfield upon deprotonation,
except for formic and acetic-d3, which are the only ones monitored by
13C NMR. For 4-fluorophenol, the isotopologues are seen to separate
during the titration, and they could be distinguished by their intensities.
For 4-fluoro-2-methylphenol, the outermost component of each doublet
of triplets was sufficient to define the chemical shift of each isotopomer.
In DMSO, the 19F signals are broadened in the middle of the titration,
owing to slow proton transfer between an acid and its conjugate base.22
Consequently, chemical shifts could not be measured accurately, and
unreliable values of KH/KD for 8 in DMSO are excluded from the results.
Other unsuccessful titrations occurred with 1-d with KOtBu in CD3CN,
2-d2 in DMSO-d6 and CD3CN, 2-d3 with Bu4NCN in CD3CN, 4 in
CD3OD, and 8 with Bu4NCN in CD3CN, where the reporter nucleus
was not sufficiently sensitive to the state of protonation.
of >6. Nevertheless, they support an origin in zero-point
energies and make an inductive explanation unnecessary.
Experimental Section
1
NMR Spectroscopy. H and 13C NMR spectra were recorded on a
Varian Mercury 300 or 400 or Unity 500 spectrometer. The 19F NMR
spectra were recorded on a Varian Mercury 300 spectrometer, without
1
1H decoupling. The spectral window was reduced to 700 Hz for H
NMR at 500 MHz, to 1700 Hz for 13C at 100 MHz, and to 20 kHz for
19F at 282.3 MHz, and the data were zero-filled to increase digital
resolution. Proton chemical shifts in aqueous solutions are relative to
tert-butyl alcohol (δ 1.23) or dioxane (δ 3.75) as the internal standard.
Carbon chemical shifts in aqueous solutions are relative to N,N-
dimethylformamide (δ 165.53) or N,N-dimethylacetamide (δ 174.57)
as the internal standard. In organic solvents, the internal standards are
cyclohexane (δ 1.42 in DMSO-d6, 1.44 in CD3CN), N,N-dimethylform-
amide (δ 162.29 in DMSO-d6), NaBF4 (δ -154 in CD3CN), and 1,3-
C6H4(CF3)2 (δ -65). Fluorine chemical shifts in aqueous solutions are
relative to sodium tetrafluoroborate (δ -154) as the internal standard.
All of the chemical shifts were read as Hz. To help distinguish the
peaks, the spectrum of a mixture of deuterated species and the internal
standard was obtained, and then the undeuterated compound was added.
Syntheses. Formic acid (1), formic-d (1-d) acid-d, acetic acid (2),
acetic-d3 (2-d3) acid-d, hydroxyacetic acid (3), pivalic acid (4),
4-fluorophenol (5), 3,5-difluorophenol (6), 4-fluorobenzoic acid (8),
4-fluorobenzoic-d4 acid (8-d4), 3,5-difluorobenzoic acid (9), and other
reagents were commercially available and used without purification.
Anhydrous formic acid was obtained by reduced-pressure distillation
of 97% formic acid that had been mixed with excess phthalic anhydride
and refluxed for 6 h. The center distillate was collected and stored at
4 °C. Acetic-d2 acid (2-d2) was obtained by heating a mixture of
malonic-d2 acid-d2 and H2O to 150 °C. The amount of H2O was adjusted
empirically to maximize the 1H NMR intensity of the acetic-d2 signals,
because acetic-d acid interferes, but acetic-d3 acid is invisible. Hy-
droxyacetic-d acid (3-d) was prepared by reducing glyoxylic acid with
zinc dust in D2O. Pivalic-d8 acid (4-d8) was prepared by carboxylation
of the Grignard reagent obtained from commercially available “98%”
tert-butyl-d9 chloride, which contained ∼2% tert-butyl-d8 chloride and
other impurities. The major product, pivalic-d9 acid (4-d9), is invisible
in the CH region of the 1H NMR spectrum, which shows a quintet due
to a CHD2 group, well separated from CH3 and a trace of CH2D.
4-Fluorobenzoic-3,5-d2 acid (8-d2) and 3,5-difluorobenzoic-2,4,6-d3 acid
(9-d3) were prepared by treating either 4-fluorobenzoic acid or 3,5-
difluorobenzoic acid with 98% D2SO4 at 140 °C and repeating the
procedure. 4-Fluorophenol-2,6-d2 (5-d2) was prepared by treating
4-fluorophenol with 50% D2SO4 at 75 °C and repeating the procedure.
3,5-Difluorophenol-2,4,6-d3 (6-d3) was prepared by treating 3,5-
difluorophenol with 60% D2SO4 at 75 °C and repeating the procedure.
The exchange of the aromatic hydrogens in each of these substrates
was monitored by 1H and 19F NMR, and samples were quenched when
equilibrium was established. 4-Fluoro-2-d3-methylphenol (7-d3) was
synthesized by activating 5-fluoro-2-hydroxybenzoic acid with CH3-
OCOCl and reducing with NaBD4. Details of all syntheses are available
in the Supporting Information, along with spectral characterizations.
Aqueous samples were first acidified with 2µL of 0.1 M DCl to
ensure complete protonation and then titrated with as many as twenty
5- or 10-µL aliquots of NaOD in D2O. Samples in DMSO-d6, CD3CN,
and THF-d8 were acidified with 2µL of 0.1 M trifluoroacetic acid, and
then 5µL aliquots of KOtBu, LiN[Si(CH3)3]2, or Bu4NCN were used
1
as the base. Acetic-d2 acid was also subjected to a reverse H NMR
titration, where a mixture of potassium acetate-d0, -d2, and -d3 in DMSO-
d6 was titrated with trifluoroacetic acid.
NMR spectra were recorded after each addition. At least 10 aliquots
were used for each titration. Titrations were assumed to be complete
when no peak movement was observed upon further addition of titrant.
Chemical shifts of appropriate reporter nuclei were extracted from the
spectrum after adding each aliquot, and the data were fit to eq 2. Figure
1 shows such a plot from 19F chemical shifts of fluorobenzoic-d0 and
-d4 acids. As an indicator of the linearity, the correlation coefficient is
0.99998.
Computations. Ab initio density functional calculations on car-
boxylic acids and phenols were performed at the B3LYP/6-31* or
B3LYP/6-31G level. For formic acid, various higher-level calculations
were undertaken, but these gave nearly the same results.
Harmonic vibrational frequencies were calculated at optimized
geometries. The double difference, ∆∆Σν, of sums of all of the
-
frequencies of the four species, (ΣνR
- ΣνR COOH) - (ΣνR
COOH
H
CO
H 2
D
-
- ΣνR
), was calculated. This difference was then converted through
CO
D
2
zero-point energies to ∆pK ()pKD - pKH) per D.
Results
Experimental. Table 2 lists the secondary deuterium IEs on
the acidities of carboxylic acids and phenols. The values are
presented in several forms, as KH/KD, pKD - pKH, and ∆pK
per D. They are sequenced in order of increasing n, the number
(21) (a) Batiz-Hernandez, H.; Bernheim, R. A. Prog. Nucl. Magn. Reson.
Spectrosc. 1967, 3, 63. (b) Jameson, C. J.; Osten, H. J. Annu. Rep. NMR
Spectrosc. 1986, 17, 1. (c) Hansen, P. E. Prog. Nucl. Magn. Reson.
Spectrosc. 1988, 20, 207. (d) Dziembowska, T.; Hansen, P. E.; Rozwad-
owski, Z. Prog. Nucl. Magn. Reson. Spectrosc. 2004, 45, 1.
(22) Ritchie, C. D.; Lu, S. J. Am. Chem. Soc. 1990, 112, 7748.
9
4492 J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007

IEs on the Acidity of Carboxylic Acids and Phenols
A R T I C L E S
Figure 2. Deuterium IE on acidity versus number of bonds between D
and O. (x) Observed ∆pK per D. (O) Calculated ∆pK per D, scaled down
6-fold. For clarity, the dashed lines connect the averages for each integer
n but have no significance.
Figure 1. Linearized plot (eq 2) for titration of fluorobenzoic-d0 and -d4
acids (8) in THF-d8 with Bu4NCN as the base: (x) first half of the titration,
(+) second half.
Table 3. Calculated Secondary Deuterium Isotope Effects on
Acidities
Table 2. Secondary Deuterium Isotope Effects on Acidities
acidD
solvent
KH/KD
pKD
−
pKH
∆
pK per D na
1
acidD
method
∆∆Σν, cm-
∆pK per D
na
1-db
D2O
1.0743 ( 0.0014 0.0311 ( 0.0005 0.031
2
2
2
2
3
3
3
3
3
3
3
c
1-db
1-db
1-db
2-d2
2-d3
2-d3
2-d3
3-de
DMSO-d6 1.0530 ( 0.0016 0.0224 ( 0.0006 0.022
DCOOH
DCOOH
DCOOH
DCOOH
HF 6-311+G(d,p)
B3LYP/6-311+G(d,p)
MP2/6-31G(d)
MP2/6-31G(2d,p)
B3LYP/6-311+G(d,p)
HF 6-311+G(d,p)
B3LYP/6-31*
B3LYP/6-31G
B3LYP/6-31*
B3LYP/6-31G
140
158
160
158
164
16
104
60
182
75
0.150
0.168
0.171
0.168
0.174
0.017
0.037
0.032
0.022
0.040
0.017
0.023
0.026
0.056
0.026
2
2
2
2
2
2
3
3
4
4
4
4
5
5
6
CD3CNc
CD3CNd
D2O
1.0086 ( 0.0011 0.0037 ( 0.0005 0.0037
0.021
1.049 ( 0.004
0.021 ( 0.002
e
b
b
b
1.025 ( 0.003 0.0107 ( 0.0012 0.0053
1.0304 ( 0.0016 0.0130 ( 0.0007 0.0043
DCOOD
D2O
DCOOHb
CD3COOH
phenol-2,6-d2
pivalic-d9
phenol-3,5-d2
d
THF-d8
1.025 ( 0.006
0.011 ( 0.003
0.0036
CD3CNd
D2O
1.0285 ( 0.0020 0.0122 ( 0.0008 0.0041
1.008 ( 0.0014 0.0035 ( 0.0006 0.0035
1.0259 ( 0.0011 0.0111 ( 0.0005 0.0055
1.0188 ( 0.0020 0.0081 ( 0.0008 0.0041
f
5-2,6-d2 D2O
f
5-2,6-d2 CD3CNc
f
benzoic-2,6-d2 B3LYP/6-31G
o-CD3phenol B3LYP/6-31G
benzoic-3,5-d2 B3LYP/6-31G
phenol-4-d
benzoic-4-d
31
65
49
53
6-d3
4-d8
4-d8
7-d3
8-d4
8-d4
8-d4
9-d3
D2O
1.0442 ( 0.0005 0.0188 ( 0.0002 0.0062 3,5
e
e
f
D2O
1.0530 ( 0.0015 0.0224 ( 0.0006 0.003
1.039 ( 0.005 0.016 ( 0.002 0.002
1.0195 ( 0.0012 0.0084 ( 0.0005 0.0028
4
4
4
CD3CNg
D2O
B3LYP/6-31G
B3LYP/6-31G
f
24
D2O
1.0159 ( 0.0016 0.0069 ( 0.0007 0.0017 4,5
1.0155 ( 0.0020 0.0067 ( 0.0008 0.0017 4,5
1.0209 ( 0.0016 0.0090 ( 0.0007 0.0022 4,5
1.0133 ( 0.0019 0.0058 ( 0.0008 0.0019 4,6
f
CD3CNd
f
d
THF-d8
D2O
a Number of bonds between D and OH (or OD). b To DCO2-Li+ as a
conjugate base.
f
f
8-3,5-d2 D2O
1.0088 ( 0.0016 0.0038 ( 0.0007 0.0019
5
Table 4. Calculated Average C-D Stretching Frequencies (cm-1
in ortho-, meta-, and para- Di- or Monodeuterated Phenol,
Phenoxide, Benzoic Acid, and Benzoate and the Calculated
Increase in Electron Density at ortho-, meta-, and para-C and H
upon Deprotonating Phenol or Benzoic Acid
)
a Number of bonds between D and O. b By 13C NMR. c KOtBu base.
1
d Bu4NCN base. e By H NMR. f By 19F NMR. g LiN[Si(CH3)3]2 base.
of bonds between the deuterium substituent and the oxygen.
Figure 2 shows a graph of ∆pK per D versus n (or the average
n if there are two distances). Notice, though, that the IE per
D is lower for hydroxyacetic acid (3) in D2O than for acetic
(2); therefore, n is not the only parameter determining the
IEs.
IEs are slightly lower in the aprotic organic solvents DMSO-
d6 and CD3CN. For several acids, the difference is hardly beyond
experimental error, but for formic acid (1), the difference is
real. For acetic-d3 acid (2-d3) in organic solvents, the IE,
measured by 13C NMR, is nearly the same as that in D2O.
Computational. Table 3 lists calculated secondary deuterium
IEs on the acidities of carboxylic acids and phenols. They too
are sequenced in order of increasing n, the number of bonds
between the deuterium substituent and the OH (or the OD in
HCOOD/DCOOD). The data are presented as ∆∆Σν, the double
difference of sums of all the vibrational frequencies of the four
species, and as ∆pK ()pKD - pKH) per D. Figure 2 includes
the calculated ∆pK per D versus n but scaled down by a factor
of 6. For further comparison, Table 4 lists average C-D
ortho
meta
para
νphenol
2374
2333
41
0.068
0.090
2397
2377
20
2373
2297
75
-0.011
0.089
2377
2335
42
2378
2334
45
0.052
0.092
2374
2340
34
νphenoxide
∆νphenoxide
∆qC,phenoxide
∆qH,phenoxide
νbenzoic acid
νbenzoate
∆νbenzoate
∆qC,benzoate
∆qH,benzoate
0.042
0.039
0.000
0.061
0.025
0.064
stretching frequencies of ortho-, meta-, and para-deuteriums
in deuterated phenol, phenoxide, benzoic acid, and benzoate,
along with the calculated increases in electron density at ortho-,
meta-, and para-carbons and hydrogens upon deprotonating
phenol or benzoic acid.
The dependence on the basis set and computational level was
probed for formic acid. The computed values of ∆pK in Table
3 range only from 0.150 to 0.171. Therefore, B3LYP/6-31* or
B3LYP/6-31G were judged as adequate for other acids.
9
J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007 4493

A R T I C L E S
Perrin and Dong
There is hardly any difference in the calculated ∆pK between
DCOOH and DCOOD, 0.168 versus 0.174. Therefore, no effort
was made to compare the IEs in H2O and D2O, inasmuch as
experimental error would be expected to overwhelm any
difference. There is a large reduction in the calculated IE of
formic acid when the conjugate base is lithium formate (with
both Li-O distances equal to 1.86 Å), rather than formate itself.
This is a model for the behavior of an intimate ion pair in a
nonpolar aprotic solvent. Consistent with this calculated reduc-
tion is the considerably lower ∆pK, 0.0037, observed for formic
acid in CD3CN when K+ is the counterion, compared to 0.021
with Bu4N+.
comparing Tables 1 and 2. This agreement is a testament to
the purity of their materials, a requirement that could be
circumvented with our NMR titration method.
Fluorine is used as a reporter nucleus for phenols and benzoic
acids. This is more sensitive than 13C because the shift of the
19F signal upon deprotonation even of a remote OH is often
>1000 Hz. Besides, it was not possible to resolve the isotope
shift at the carboxyl carbon of benzoic-d5 acid. Although fluorine
exerts its own inductive effect, it is the same inductive effect
for both H and D; therefore, it does not introduce complications.
Indeed, no regular variation of IEs with inductive substituents
was seen.24
Distance Dependence. The IE is expected to decrease as n
increases, that is, as the site of deuteration recedes from the
acidic OH. This is almost what is found. According to the data
in Tables 2 and 3, the decreases in both the experimental and
the calculated ∆pK per D are considerable from n ) 2 to 3.
Beyond n ) 3, the decrease of the experimental ∆pK per D is
quite gradual, but there is no drop in the calculated values. This
contrast between n < 3 and n > 3 can be seen in Figure 2.
The individual contributions of ortho-, meta-, and para-
deuteriums can be extracted from the observed IEs for the three
different fluoro- and difluorobenzoic acids. From 4-fluoroben-
zoic-3,5-d2 acid (8-d2), the contribution of one meta-deuterium
to pKD - pKH is 0.0019 ( 0.0003. Subtracting this from the IE
of 4-fluorobenzoic-d4 acid (8-d4) gives 0.0015 ( 0.0005 as the
contribution of one ortho-deuterium. Subtracting this from the
IE of 3,5-difluorobenzoic-d3 acid (9-d3) gives 0.0027 ( 0.0013
as the contribution of one para-deuterium. This value is included
in Figure 2. These results agree, but not exactly, with those of
ref 17, where the effect of an ortho-, meta-, or para-deuterium
on ∆pK was found to be 0.0020, 0.0019, or 0.0018, respectively.
Similarly, the IE of a para deuterium on phenol can be
assigned as 0.0077 ( 0.0005 by subtracting the IE in 4-fluoro-
phenol-2,6-d2 from that in 3,5-difluorophenol-2,4,6-d3. This
value too is included in Figure 2. It is larger than the IE (per
D) of 0.0055 for an ortho-deuterium. This is consistent with
the calculated IEs per D in Table 3, which increase from ortho
to meta to para. It is inconsistent with experimental ortho-,
meta-, and para-deuterium kinetic IEs of -2.4, -0.4, and
-1.8%, respectively, in the R deprotonation of toluene.23
These comparisons highlight a further contrast between
aliphatic and aromatic acids. Aliphatic acids do show a drop
beyond n ) 3 for both the experimental and calculated IEs, at
least for the comparison of acetic acid (2) with pivalic (4) or
ortho-methylphenol (7, which is mixed aliphatic-aromatic). It
is the aromatic acids whose IEs show no drop beyond n ) 3.
Although the uncertainties are large, owing to propagation of
errors upon subtraction, the values increase from ortho to meta
to para. Likewise, the calculated IEs in both phenol and benzoic
acid generally increase (from 0.032 to 0.040 to 0.056 and from
0.017 to 0.026 to 0.026) along this series.
The observed IEs of ortho-D and ortho-CD3 correspond to a
reduction of acidity, just as those of meta- and para-D. This is
evidence against an inductive origin, arising from the field effect
of a dipole, because the angular dependence of an ortho
substituent’s dipole field is reversed and would have produced
an opposite IE.
Isotope-Sensitive Vibrational Frequencies. The calculated
IEs in Table 3 corroborate the observed decreases in the acidity
of carboxylic acids and phenols upon deuteration. These
calculated IEs ignore anharmonicity, upon which the estimate
of eq 1 depends. Therefore, the ability of these calculations to
reflect IEs provides no support for an inductive origin, as was
concluded from the estimated (eq 1) inductive effect on ∆pK
of only 0.00006 per D. Instead, these calculations involve only
zero-point energies. There are key isotope-sensitive vibrations
whose frequencies, ν, and zero-point energies (1/2hν) decrease
upon deprotonating the acid. Because the zero-point energy of
a CH vibration is greater than that of a CD, the decrease of
frequency is more stabilizing for H. These vibrations thus
contribute positively to the double difference, ∆∆Σν, or
-
-
(ΣνR COOH - ΣνR COOH) - (ΣνR CO - ΣνR CO ), in Table 3.
H
D
H
2
D
2
Which vibrations are these? For formic acid, it is easy to
recognize that the key isotope-sensitive vibration is the C-H
stretch. According to the B3LYP/6-311+G(d,p) calculations,
which are representative, its frequency decreases from 3057 in
HCOOH to 2571 cm-1 in HCO2-. The corresponding decrease
for DCOOH is from 2274 to 1876 cm-1. This contributes 88
cm-1 to ∆∆Σν. This is only slightly more than half the total,
and other modes, such as the CH bend and the CdO stretch,
also contribute.
For other acids, key isotope-sensitive vibrations are less
readily identifiable. Because of variations in the way individual
bond vibrations couple into normal modes, it is not unambiguous
to correlate a frequency in the acid with one in the deuterated
acid or the carboxylate. Prime candidates, though, are the C-H
modes. For acetic acid, the average of the three C-H stretching
frequencies is calculated to decrease from 3085 to 3007 cm-1
in the acetate anion. For pivalic acid, the average of the nine
C-H stretching frequencies is calculated to decrease from 3105
to 3056 cm-1 in the anion. For ortho-methylphenol, deproton-
ation reduces the average methyl C-H stretching frequency by
14 cm-1. For all three of these acids, the total decrease of the
C-H stretching frequencies upon deprotonation is sufficient to
account for slightly more than half of the calculated IEs, but
other modes contribute, especially CCH bends. For selectively
ring-deuterated phenols and benzoic acids, it is easier to isolate
Discussion
Experimental Results. Our results confirm secondary deu-
terium IEs on acidity. Deuteration definitely decreases the
acidity of carboxylic acids and phenols. For those acids that
overlap, the agreement with the values of Streitwieser and Klein
and with subsequent studies is very close, as can be seen by
(24) Barnes, D. J.; Golding, P. D.; Scott, J. M. W. Can. J. Chem. 1974, 52,
(23) Streitwieser, A., Jr.; Humphrey, J. S., Jr. J. Am. Chem. Soc. 1967, 89, 3767.
1966.
9
4494 J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007

IEs on the Acidity of Carboxylic Acids and Phenols
A R T I C L E S
Scheme 2. Delocalization of π Electrons to or from ortho and
para Positions in (a) Phenoxide and (b) Benzoic Acid and of σ
Electrons to ortho, meta, and para Positions in (c) Phenoxide and
(d) Benzoate
Scheme 1. n-σ* Delocalization of Lone Pairs in Formate Anion,
Alcoholates, Amines, and Acetate
the average C-D stretching frequency. Values are in Table 4,
along with the frequency decreases upon deprotonation. These
stretching modes account for about half of the calculated IEs
of meta-deuterium. Other modes dominate for ortho and para,
but it has not been possible to discern which, except that out-
of-plane bends account for more than half of the calculated IEs
of phenol-4-d.
Delocalization. The reduction of the C-H frequency upon
deprotonation of formic acid can be ascribed to electron
delocalization, as in Scheme 1. In the formate anion, there are
increased electron densities at ortho- and para-carbons are as
expected from resonance theory, owing to delocalization either
of negative charge from the oxyanion of the phenoxide or of
electron density to the carbonyl of the benzoic acid, as shown
in Scheme 2a,b. Yet this delocalization is of π electron density,
whereas the hydrogens are bonded by σ electrons. Sigma
delocalization can increase the electron density only at the meta-
hydrogens of phenoxide and at the ortho- and para-hydrogens
of benzoate, as shown in Scheme 2c,d. This is not the pattern
of changes seen. Therefore, n-σ* delocalization cannot account
for these electron densities or for the calculated IEs, reasonable
though it is for formic acid (Scheme 1).
oxygen lone pairs that are antiperiplanar to the antibonding σCH
*
orbital. The consequent delocalization of the lone-pair electrons
reduces the C-H bond order and its frequency and zero-point
energy. This delocalization is also manifested in a lengthening
of the calculated C-H distance, from 1.097 to 1.138 Å, and a
greater Mulliken electron density on the H.
This is analogous to the delocalization of antiperiplanar lone
pairs that was proposed to explain â secondary IEs in alcohols
and amines.25 These too are illustrated in Scheme 1. Indeed, it
was shown that the IE in amines is of stereoelectronic origin.
This conclusion depended on the ability to measure relative
acidities of two isotopomers, stereoisomers that differ only in
the position of an isotope (as distinguished from the isotopo-
logues here, which differ in the number of isotopic substitutions).
Unlike formic acid, the frequency changes in higher acids
are not readily attributed to a delocalization that is clearly
stereoelectronic. Upon deprotonation of acetic acid, the C-C
distance is calculated to lengthen from 1.508 to 1.576 Å,
consistent with n-σ* delocalization as in Scheme 1, and the
calculated electron densities at all three hydrogens increase. The
three C-H distances increase, and the increase is greater for
the one that is in the molecular plane. This is consistent with
the result in Table 2 that the IE per D is lower for hydroxyacetic
acid than it is for acetic, because the OH of the former lies in
the molecular plane,26 relegating deuterium to a position where
it affects the IE to a lesser extent. Upon deprotonation of pivalic
acid, the C-CR distance is calculated to lengthen, but all of the
CR-Câ and C-H distances remain nearly constant. The methyl
carbons lose electron density, whereas all nine H atoms gain,
especially those three that are antiperiplanar to the C-CR bond.
This is consistent with relayed n-σ* delocalization.
Distance Dependence. Although the IE is expected to
decrease as the site of deuteration becomes more distant from
the acidic OH, the data in Tables 2 and 3 and in Figure 2 show
that it is not so simple. Both the experimental and the calculated
∆pK per D decrease considerably from n ) 2 to 3. Beyond n
) 3 there is a further decrease for the aliphatic acids. However,
for deuteration of benzoic acid, the experimental IE is nearly
constant, and for deuteration of phenol, the IE increases from
ortho to para. For both benzoic acid and phenol, the calculated
IEs increase from ortho to meta to para. The apparent peak at
n ) 5 is due to the large IE of a para-D in phenol, both
experimentally and computationally. Such behavior is opposite
to the usual falloff of IEs and other substituent effects with
distance.
There is no difficulty in accounting for the falloff from n )
2 to 3 or for a further falloff for n > 3 in aliphatic acids. The
IE in formic acid is due to n-σ* delocalization, which is
attenuated in the higher acids by the additional sp3 carbons that
are interposed. Indeed, the deprotonation-induced changes in
the C-H stretching frequencies and in the calculated electron
densities at hydrogen decrease with n. However, it should be
noted that the attenuation of IEs is not by a constant factor.
For the aromatic acids, the role of electron delocalization is
less clear. Table 4 includes the calculated increases of electron
density upon deprotonation. For phenol, the ortho- and para-
carbons acquire electron density, but all five hydrogens gain
nearly equally. For benzoic acid, the ortho- and para-carbons
again acquire electron density upon deprotonation but to a lesser
extent than for phenol. Yet it is the meta- and para-hydrogens
that acquire more electron density than does the ortho-. The
The near constancy of the IEs in the aromatic acids, both
experimental and calculated, is surprising. A clue to this lies in
the average C-D stretching frequency in selectively deuterated
phenols and benzoic acids, given in Table 4. These frequencies
decrease upon deprotonation, but the decreases are largest at
the meta-C-D, and in phenol-3,5-d2, the decrease at meta is
nearly twice as large (per C-D) as that at the ortho or para.
This is quite unexpected and contrary to the general principle
that substituent effects are transmitted most effectively to and
from ortho and para positions, as illustrated in Scheme 2a. It
is certainly not consistent with an inductive effect of the
(25) (a) Gawlita, E.; Lantz, M.; Paneth, P.; Bell, A. F.; Tonge, P. J.; Anderson,
V. E. J. Am. Chem. Soc. 2000, 122, 11660. (b) Perrin, C. L.; Ohta, B. K.;
Kuperman, J. J. Am. Chem. Soc. 2003, 125, 15008. (c) Perrin, C. L.; Ohta,
B. K.; Kuperman, J.; Liberman, J.; Erde´lyi, M. J. Am. Chem. Soc. 2005,
127, 9641.
(26) Newton, M. D.; Jeffrey, G. A. J. Am. Chem. Soc. 1977, 99, 2413.
9
J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007 4495

A R T I C L E S
Perrin and Dong
Table 5. Comparison of Experimental and Calculated IEs (∆pK
also seems too large, from 1.508 to 1.576 Å, or 4%. Yet,
according to an extensive survey of crystal structures,30 the
average C-C bond lengths in carboxylic acids and carboxylates
are 1.502 ( 0.014 and 1.520 ( 0.011 Å, respectively,
corresponding to only a 1% lengthening. Thus, we judge that
the discrepancy between calculated and observed IEs is due to
unrealistically large changes in calculated vibrational frequen-
cies.
per D)
acid
D2O
B3LYP
calcd/exptl
formic
acetic
pivalic
phenol-2,6-d2
phenol-4-d
cresol
benzoic-2,6-d2
benzoic-3,5-d2
benzoic-4-d
0.031
0.0053
0.003
0.0055
0.0077
0.0028
0.0015
0.0019
0.0027
0.168
0.037
0.022
0.032
0.056
0.023
0.017
0.026
0.026
5.4
7.0
7.3
5.8
7.3
8.2
11.3
13.7
9.6
Of course all of the experimental data are in solution or
crystals, whereas the calculations are for the gas phase. This is
the usual explanation for such discrepancies in frequencies and
bond lengths.31 Lone-pair delocalization in the carboxylates
(Scheme 1) is responsible for reduction of vibrational frequen-
cies, and that delocalization is maximum in the gas phase. In
water, hydrogen bonding to the lone pairs reduces their
delocalization and reduces the IE. Yet it is doubtful that this
can account for the overestimates of the IEs by at least 6-fold.
To test this influence of solvation, IEs were measured in
solvents of lower polarity, specifically aprotic ones. These
resemble the gas phase more, and the IE would be expected to
be higher in these solvents and closer to the calculated values.
Instead, IEs are slightly lower in DMSO-d6 and CD3CN. The
difference is not always beyond experimental error, but the IE
is definitely not higher. Moreover, although n-σ* delocalization
is reasonable for formic acid (Scheme 1), it cannot account for
the detailed electron densities around the aromatic rings of
phenol and benzoic acid, where the discrepancy between
calculated and experimental IEs is thus not readily attributed
to a solvent effect upon delocalization.
deuteriums. It is consistent with the changes of electron density,
which do not fall off with distance. It is also consistent with
the observed IEs in fluorinated benzoic acids, which increase
from 0.015 to 0.0019 to 0.0027 (but with large errors) as
deuteration moves from ortho to meta to para, six atoms away
from the OH. Thus, the frequencies and the electron densities
are consistent with IEs that arise even from deuteration at more
distant carbons.
Comparison of Experiment and Computation. The com-
putations are certainly adequate to account for the observed
decreases in acidity upon deuteration. For the simplest, formic
acid, the computations assign responsibility to a decrease of
the C-H stretching frequency upon deprotonation. For the other
acids, the observed IEs parallel changes in key stretching
frequencies that are associated with changes in electron densities
and bond lengths upon deprotonation. The computations are also
successful in reflecting the near constancy of IEs in phenol and
benzoic acid even as the site of deuteration becomes more
distant, from ortho to meta to para.
However, there is a quantitative discrepancy between calcula-
tion and experiment, as summarized in Table 5. On an energetic
basis, the calculated IEs are at least 5 times as large as the
experimental ones, and the average ratio of calculated to
experimental IE is ∼9. Thus, this discrepancy is substantial and
serious. It is not obviously due to a too simplistic computational
level because the data in Table 3 show that the results for formic
acid are nearly independent of level. It is too large to be
corrected by any scaling factor typically used for vibrational
frequencies.27
A related explanation for an experimental IE that is lower
than the calculated even in aprotic solvents is that untitrated
acid RCOOH serves to hydrogen-bond to the anion. This is a
reasonable possibility because the acid would then mimic a
protic solvent. In this case, the IE would be lower during the
first half of the titration, while RCOOHOCOR- is the product
-
of deprotonation. During the second half, while RCO2 is the
product, the IE would be higher, characteristic of an aprotic
medium and approaching the gas phase. This behavior would
be characterized by a linearized titration curve (eq 2) that reliably
follows that lower slope in the first half of the titration but shows
distinct curvature in the second half, with a corresponding
deterioration of the correlation coefficient. This would be easily
recognized, but it was not seen. Figure 1 confirms this, with
the points from the first half of the titration distinguished as ×
and those from the second half as +.
Although the results for formic acid in Table 3 are nearly
independent of level, all of the levels used must be too simplistic.
Certainly at a sufficiently high level the computation ought to
agree with the experiment. The appropriate level and the source
of the discrepancy at lower levels are being sought.32
For formic acid, where the key isotope-sensitive vibration
can be identified, the discrepancy can be inspected. According
to the B3LYP/6-311+G(d,p) calculations, the frequency of the
C-H stretch decreases from 3057 in HCOOH to 2571 cm-1 in
HCO2-. Other levels of calculation, even with larger basis sets,
give comparable decreases of 15-20% in the C-H stretch upon
deprotonating formic acid. This is too extreme. Such a low
frequency is certainly not seen in the infrared or Raman
spectrum of sodium formate, whose C-H stretch is at 2830 or
2825 cm-1 28
Others have experienced similar difficulty in
.
calculating this particular frequency accurately.29 Moreover,
although the lengthening of the C-H bond length upon
deprotonation is consistent with delocalization, the calculated
change, from 1.097 to 1.138 Å, seems too large. The calculated
lengthening of the C-C bond upon deprotonation of acetic acid
Conclusions
These results confirm that secondary deuterium IEs decrease
the acidity of carboxylic acids and phenols. For aliphatic acids,
the IEs, per D, are lower when the deuterium is moved farther
from the OH, as expected. However, it is surprising that IEs in
(27) Scott, A. P.; Radom, L. J. Phys. Chem. 1996, 100, 16502.
(28) (a) Kidd, K. G.; Mantsch, H. H. J. Mol. Spectrosc. 1981, 85, 375. (b)
Newman, R. J. Chem. Phys. 1952, 20, 1663.
(29) (a) Magalha˜es, A. L.; Mada´ıl, S. R. R. S.; Ramos, M. J. Theor. Chem.
Acc. 2000, 105, 68. (b) Dixon, D. A.; Feller, D.; Francisco, J. S. J. Phys.
Chem. A 2003, 107, 186.
(30) Allen, F. H.; Kennard, O.; Watson, D. G.; Brammer, L.; Orpen, A. G.;
Taylor, R. J. Chem. Soc., Perkin Trans. 2 1987, S1.
(31) Masamura, M. Theor. Chim. Acta 1989, 75, 433.
(32) Baldridge, K. K. Personal communication.
9
4496 J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007

IEs on the Acidity of Carboxylic Acids and Phenols
A R T I C L E S
the aromatic acids do not decrease with distance. The experi-
mental data, including this lack of distance dependence, can be
reproduced by ab initio computations, but these greatly over-
estimate the values. The overestimate does not seem to be due
to neglect of solvation, which might have reduced the experi-
mental IE, because the IEs are nearly solvent-independent.
According to the computations, the IEs originate from
vibrations whose frequencies and zero-point energies are
lowered upon deprotonation, but it is not always easy to
distinguish which vibrations are responsible nor to understand
how all of the calculated changes in frequencies, bond lengths,
and electron distributions are manifested in IEs. Besides, the
changes in electron distribution do not always follow the patterns
expected from delocalization of σ or π electrons.
responsible. Moreover, the failure of the IEs to diminish from
ortho to meta to para is not consistent with an inductive effect.
Acknowledgment. This research was supported by NSF Grant
CHE03-53091. Purchase of the spectrometers was made possible
by grants from NIH and NSF. Computations at the San Diego
Supercomputer Center were supported by the W. M. Keck
Foundation through resources at the Keck Laboratory for
Integrated Biology.
Supporting Information Available: Synthesis and spectral
characterizations. Tables S1-S4 (sample preparation). Com-
putational methods. Tables S5-S6 (calculated energies, opti-
mized geometries, atomic charges, and interatomic distances
of all structures). Table S7 (calculated vibrational frequencies
for all structures, including deuterated isotopologues). This
material is available free of charge via the Internet at
http://pubs.acs.org.
A key conclusion is that the computations support an
interpretation of these IEs in terms of deprotonation-induced
changes in vibrational frequencies and zero-point energies.
Consequently, there is no need to invoke an inductive effect as
JA069103T
9
J. AM. CHEM. SOC. VOL. 129, NO. 14, 2007 4497