Acidity of ortho-substituted benzoic acids: An infrared and theoretical study of the intramolecular hydrogen bonds

Article abstract of DOI:10.1039/b601875k

The structures of ortho-substituted benzoic acids with substituents bearing hydrogen atoms (OH, NH₂, COOH and SO₂NH₂) were investigated by means of IR spectroscopy and of density functional theory at the B3LYP/6-311 + G(d,p) level. All possible conformations, hydrogen bonds, tautomeric forms and zwitterions were taken into consideration and particular attention was given to intramolecular H-bonds and their effect on acidity. Strong H-bonds in the anions of all four acids, were revealed by calculations. In three cases they were confirmed by the IR spectra of the tetrabutylammonium salts in tetrachloromethane solution, while the salt of 1,2-benzenedicarboxylic acid was not sufficiently soluble. The H-bonds are of different strengths but in all cases they are the main cause of the strengthened acidity of these acids in the gas phase and also in solution, although their effect is opposed by weaker H-bonds present in the undissociated acid molecules. The substituent effect on the acidity was evaluated in terms of isodesmic reactions, separately in the acid molecules and in the anions. While the acidity of the 2-OH and 2-NH ₂ acids is determined essentially by the H-bonds, that of the 2-COOH and 2-SO₂NH₂ acids is strengthened by the polar effect operating in the undissociated molecule in addition to the H-bond in the anion. The steric inhibition of resonance (SIR), estimated from model conformations with fixed torsional angles, is of little importance. This analysis goes significantly beyond the classical explanation obtained from the acidities in solution but essentially conforms with it. The Royal Society of Chemistry 2006.

Full text of DOI:10.1039/b601875k

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Acidity of ortho-substituted benzoic acids: an infrared and theoretical study of
the intramolecular hydrogen bonds†
Pavel Fiedler,a Stanislav Bo¨hm,b Jirˇ´ı Kulha´nekc and Otto Exner*a
Received 7th February 2006, Accepted 2nd March 2006
First published as an Advance Article on the web 7th April 2006
DOI: 10.1039/b601875k
The structures of ortho-substituted benzoic acids with substituents bearing hydrogen atoms (OH, NH2,
COOH and SO2NH2) were investigated by means of IR spectroscopy and of density functional theory
at the B3LYP/6-311 + G(d,p) level. All possible conformations, hydrogen bonds, tautomeric forms and
zwitterions were taken into consideration and particular attention was given to intramolecular H-bonds
and their effect on acidity. Strong H-bonds in the anions of all four acids, were revealed by calculations.
In three cases they were confirmed by the IR spectra of the tetrabutylammonium salts in
tetrachloromethane solution, while the salt of 1,2-benzenedicarboxylic acid was not sufficiently soluble.
The H-bonds are of different strengths but in all cases they are the main cause of the strengthened
acidity of these acids in the gas phase and also in solution, although their effect is opposed by weaker
H-bonds present in the undissociated acid molecules. The substituent effect on the acidity was
evaluated in terms of isodesmic reactions, separately in the acid molecules and in the anions. While the
acidity of the 2-OH and 2-NH2 acids is determined essentially by the H-bonds, that of the 2-COOH and
2-SO2NH2 acids is strengthened by the polar effect operating in the undissociated molecule in addition
to the H-bond in the anion. The steric inhibition of resonance (SIR), estimated from model
conformations with fixed torsional angles, is of little importance. This analysis goes significantly
beyond the classical explanation obtained from the acidities in solution but essentially conforms with it.
either by a blend of inductive and resonance effects8a,b or simply
by correlation with the reactivities of para derivatives.8c In our
Introduction
previous analysis,9 we tried to distinguish theoretical concepts of
two kinds. Some of them are purely models, sometimes poorly
defined: for instance one can hardly distinguish what is a short-
range field effect and what is a primary steric effect. On the
other hand, there are theoretical concepts immediately related to
observable facts: for instance the steric inhibition of resonance4,7
(SIR) assumes a nonplanar minimum-energy conformation and
this assumption can be tested.10,11 Of particular importance is the
effect of hydrogen bonds, involving either a hydrogen atom of
the substituent or of the functional group;1–3,5,7 the presence of
both can be checked directly by experiment. Note that most of
the named terms were defined and experimentally proven on the
model of 2-substituted benzoic acids. One reason for this may be
that, besides the availability of these compounds, their meta and
para isomers served as a standard model in connection with the
Hammett equation.
Benzene ortho derivatives, represented very often by 2-substituted
benzoic acids, have been studied less extensively than their meta
and para isomers, nevertheless they are still objects of interest.1,2
The main problem is that their reactivity and physical properties
are controlled by more complex substituent effects,3,4 which cannot
be described by a simple relationship like the Hammett equation.
As a formal description, the effects of ortho substituents were
either assumed to be controlled by similar components to those
present in the meta and para isomers (inductive effects and reso-
nance) but in a different proportion, or alternatively, additional
factors were considered, often not exactly defined:3–6 primary
steric effects, steric inhibition to resonance, short-range field
effects, hydrogen bonds, steric inhibition to solvation. In numerous
treatments of this subject, certain of these effects received more or
less attention, sometimes they were even expressed by an empirical
equation with a limited success.5–7 Alternatively, separation of the
effects is abandoned and they are brought together under the
term ‘ortho effect’. As an extreme opinion, reactivities of ortho
derivatives were explained without referring to any steric effects,8
In our opinion, the main shortcoming of the outlined treatment
of the ortho effect is that it has been mainly based on the
dissociation constant (less frequently on the rate constant) in water
or in aqueous solvents. In the basic model of 2-substituted benzoic
acids, the substituent effect is defined by the Gibbs energy (or
enthalpy) of the isodesmic reaction:
aInstitute of Organic Chemistry and Biochemistry, Academy of Sciences of
the Czech Republic, 166 10, Praha 6, Czech Republic
bDepartment of Organic Chemistry, Prague Institute of Chemical Technol-
ogy, 166 28, Praha 6, Czech Republic
2-XC6H4COOH + C6H5COO− ꢀ 2-XC6H4COO−
cDepartment of Organic Chemistry, University of Pardubice, 532 10,
Pardubice, Czech Republic
+ C6H5COOH
(1)
† Electronic supplementary information (ESI) available: IR spectral data
of the ortho-substituted benzoic acids and of their tetrabutylammonium
salts, details of their DFT calculated geometry and DFT energies of some
structures with fixed conformation. See DOI: 10.1039/b601875k
More recently, this reaction12 and other acidobasic equilibria13
were also investigated in the gas phase. The results can be
correlated with the structures of isolated molecules more closely
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 2003–2011 | 2003
©

View Article Online
than the acidities in solution. However, there still remains a
fundamental shortcoming: that the substituent effect is given
by a difference between the effect in the acid molecule and in
the anion. It may happen that the two effects are of similar
magnitude and their difference is difficult to interpret.10,11 In
order to separate the two effects on ionization, we used the
principle of isodesmic reactions14 and applied it systematically to
inductive,15 resonance,16 and other10,11 effects. When this principle
is applied to ortho substituted benzoic acids, the isodesmic (and
also homodesmotic17) reactions have the form of eqns (2) and (3).
Wide intervals of concentrations were used, from 10−1 to 10−6 mol
l−1, depending on the solubility of individual compounds. Nev-
ertheless, some bands of the carboxylic acid dimers could not be
completely eliminated by dilution. The cell thickness was changed,
depending on concentration, from 0.1 mm to 10 cm. Several solid
state spectra were recorded in KBr pellets, spectra of the partially
deuterated samples of 3 in paraffin oil mulls.
The spectral data are collected in Tables S1 to S6 (supplementary
information†).
Calculations
C6H5COOH + C6H5X ꢀ 2-XC6H4COOH + C6H6
C6H5COO− + C6H5X ꢀ 2-XC6H4COO− + C6H6
(2)
Energies of 2-substituted benzoic acids and of their anions were
calculated by the DFT method at a B3LYP/6-311 + G(d,p) level
with the GAUSSIAN 03 program.19 Planarity or any symmetry
was never anticipated. Geometry optimization was carried out
with redundant internal coordinates. All reasonable conforma-
tions were taken into consideration and calculations were started
from the pertinent near structures; tautomeric and zwitterionic
forms were also considered as given by the formulas 1–8. All
optimized structures were checked by vibrational analysis and
behaved as energy minima.
The calculated energies E(DFT), without corrections to zero-
point energies, are listed in Table 1 together with the most
important geometric parameters, further geometric parameters
are given in Table S7 (supplementary information), the parameters
connected directly to H-bonds are given also in Table 2. Population
of conformers was estimated for 298 K in two ways: either from
(3)
The reaction enthalpies of these reactions, D2H◦ or D3H◦, are
a measure of the substituent effects in the acid molecule and in
the anion, respectively. The values of D2H◦ could be obtained, in
principle, from the enthalpies of formation of all species involved,
but the available values18 are not sufficiently accurate.9 Therefore,
we preferred9,16b calculations within the framework of the density
functional theory19 (DFT).
In our previous work, we divided 2-substituted benzoic acids
into three subsets according to the character of the substituent:
alkyl groups, polar substituents without an acidic hydrogen atom
and substituents bearing such H atoms. Alkyl groups exhibited
steric effects10–including SIR11–of similar magnitude in the acid
and in its anion; acidity was controlled mainly by polarization in
the anion.11 Polar substituents without hydrogen represent a more
complex problem. We tentatively divided the substituent effects
into four parts:9 H-bond in the acid molecule, SIR, polar effect
mediated through the benzene ring (PE) and steric van der Waals
effect (SE); the two last effects were merely fictive.
In this article, we extended our investigation to the last group of
substituents, those bearing an acidic hydrogen atom. We dealt with
four 2-substituted benzoic acids: 2-OH, 2-NH2, 2-COOH and 2-
SO2NH2. The H-bonds, assumed to be the most important factor
influencing the acidity, were examined both theoretically and from
the IR spectra of the acids (in part already well described) and of
their tetrabutylammonium salts; this was possible only with three
of the named acids. The DFT calculations were carried out at the
same level as previously;9 great attention was given to all possible
conformations, H-bonds and zwitterions. By these methods it was
possible to clearly distinguish strong and weak H-bonds and to
estimate roughly their relative energies. In all cases the H-bonds
were found both in the acid molecule and in the anion, even when
they were of different strengths or when their effects on the acidity
partly compensated.
the calculated DG◦(298) or with the assumption DG◦(298)
∼
=
DE(DFT). Since the calculated vibrational frequencies sometimes
differed rather a lot from the experimental values, we were of the
opinion that the latter estimates are more reliable. Nevertheless, the
differences between the two sets of values were mostly negligible.
The only case when they were not is given in Table 1, footnote
f, nevertheless even in this case the impact on the final values
is minute. The effective energies were then evaluated for the
equilibrium mixture. The energies D2E◦ and D3E◦ of the isodesmic
reactions of eqns (2) and (3), further the relative acidities D1E◦
(Table 3) relate always to this equilibrium mixture at 298 K.
The infrared frequencies were calculated using the same
program;19 they were scaled by a factor20 of 0.9679 but still
remained significantly greater than the experimental values. The
calculated frequencies are given in Table S7.
The frozen conformations of some 2- and 4-substituted benzoic
acids with a fixed dihedral angle U were calculated in such a
way that all remaining geometry parameters were optimized with
internal coordinates. The results are given in Table S8.
Experimental
Results and discussion
Tetrabutylammonium salts 2, 4, 6 and 8 were prepared by titration
of the acids 1, 3, 5 and 7, respectively, with the solution of
tetrabutylammonium hydroxide exactly to the equivalence point.
The solution was evaporated to dryness and the glassy residue
subjected to spectroscopy. The purity was checked by NMR
spectra; no signals of impurities were detected.
Infrared absorption spectra were recorded on a Bruker Equinox
55 FTIR spectrometer in the region 400–3800 cm−1 in tetra-
chloromethane solutions and, in a few cases, also in chloroform.
The problems of structure and acidity are rather different with the
four acids. Therefore, the results will be reported separately.
2-Hydroxybenzoic acid (1)
This acid has been studied extensively. There seems to be no
contradiction in the literature that the strong acidity1a (pK = 2.99
in water) is attributed1,4,7 to the H-bond in the anion (2b, Scheme 1)
but both spectroscopic experiments21 and calculations at different
2004 | Org. Biomol. Chem., 2006, 4, 2003–2011
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
Table 1 DFT energies and some geometric parameters of ortho-substituted benzoic acids
Dihedral anglesb
Substituent
2-OH
Conformation a
u1
u2
DFT energy/a. u.
Relat. Ec/kJ mol−1
Populationd (%)
1a
1b
1c
1d
1e
2a
2c
2d
2b
2e
3a
3b
3c
4a
5a
5b
5c
5d
6a
6b
7a
7b
7c
8a
sp,sp-HB
ap,sp-HB
ap,ap-HB
ap,ap
0.0
−496.2056825
−496.2002301
−496.1902134
−496.1885893
−496.1879379
−495.6731258
−495.6435371
−495.6412893
−495.6720633
−495.6320693
−476.3311607
−476.3265439
−476.3178081
−475.7813756
−609.5709966
−609.5707100
−609.5705234
−609.5701600
−609.0577479
−609.0300078
−1024.9565881
−1024.9534333
−1024.9509626
−1024.4337535
0
99.7
0.3
180.0
176.7
180.0
0.0
180.0
180.0
0.0
14.30
40.60
44.86
46.57
0
77.66
83.56
2.76
107.77
0
sp,ap
Anion-Phe
ap-HB
ap
sp
sp,sp-HB
sc,ap
75.3
24.7
Anion-Cae
2-NH2
0.0
76.4
0.8
176.7
166.1
6.4
22.0
141.3
21.9
33.4
1.3
sp,sp-HB
ap,sp-HB
ap,ac-HB
sp,sp
7.0
99.2
0.8
12.8
15.0
12.10
35.04
Anion
2-COOH
sp,ac
234.7
141.3
200.7
44.0
181.6
111.3
48.5
222.2
55.7
57.4
0
36.3f
26.8f
22.0f
14.9f
ac+,ac+
sp,ap-HB
sc+,sc+
0.75
1.24
2.20
0
72.83
0
Anion
sp,ap-HB
sp,ac
100
0.0
2-SO2NH2
sc+,sc+-HB
ac+, ac−
ac+,sc+-HB
sc+,sc+-HB
35.6
135.4
142.3
39.2
96.3
3.4
0.3
8.28
14.77
Anion
a Conformation on the bonds C1–C(O) and C2–X; the hydrogen bond is indicated in all conformations where it exists; Zw stands for zwitterion. b Dihedral
angles C2–C1–C O and C1–C2–X–Y, respectively. c Relative energy with respect to the most stable form. d Approximate population in the equilibrium
=
mixture at 298 K calculated from DE values. e Phenolate and carboxylate anion, respectively. f From DG◦ we got the population 39.8, 34.2, 20.8 and 5.2%.
Table 2 Quantities characterizing the hydrogen bond
∠X–H. . .Y/◦
DEa/kJ mol−1
Dm calc./cm−1
Dm exp./cm−1
˚
˚
Compound
Structure
X–H/A
H. . .Y/A
=
1a
1b
1c
2a
2b
3c
5c
6a
3a
3b
4a
7a
8a
O–H. . .O
O–H. . .O
O–H. . .O
0.971
0.970
0.971
1.038
1.052
0.983
0.986
1.184
1.010
1.006
1.032
1.018
1.054
1.76
1.80
1.79
1.48
1.45
1.83
1.62
1.19
1.95
1.97
1.76
2.03
1.66
145
141
145
158
158
143
162
179
129
126
141
139
160
−26
−12
359b
141
68
not reliable
not reliable
343
395b
O–H. . .O−
O–H. . .O−
O–H. . .N
−73
342c
264b
—
=
O–H. . .O
408
O–H. . .O−
not reliable
=
N–H. . .O
N–H. . .O
−10
−14
158d
133d
—
120d
=
N–H. . .O−
393d
245d
119e
201e
=
N–H. . .O
134e
N–H. . .O−
not reliable
a Estimated energy change when the two groups connected by a H-bond are brought together, see text. b Related to m(O–H) of phenol. c Related to
m(O–H) of to benzoic acid. d Difference of mas and ms of the NH2 group, it is to be compared with Dm of aniline, calculated 97 cm−1, experimental 80 cm−1
e Difference of mas and ms of the NH2 group, compared with Dm of benzenesulfonamide, calculated 109 cm−1, experimental 201 cm−1
.
.
levels22 deal only with the H-bond in the acid molecule and
exclusively prefer the form 1a. Acidity is commonly discussed in
comparison with the weaker 4-hydroxybenzoic acid.4,22 However,
the latter behaves in the gas phase as a substituted phenol and its
deprotonation proceeds on the phenolic OH group.12,16b Therefore,
one must also take into consideration the phenolic OH acidity
in the case of 1. We calculated the energies of the acid and
perhaps of the anion in all possible conformations, 1a–e and 2a–
e, respectively; the relative energies and populations are given in
Table 1, the last two columns.
IR investigation and with previous spectral data.21 A strong H-
bond was observed at all concentrations, independently of the
equilibrium of 1a with its dimer (Table S1, ESI†). The form 1b
can be neglected since the energy difference between 1b and 1a
is 14.3 kJ mol−1 (in satisfactory agreement with calculations22a at
a slightly lower level, 12.1 kJ mol−1.) Concerning the structure of
the anion, we obtained probably the first direct experimental proof
of the H-bond from the IR spectrum of the tetrabutylammonium
salt in tetrachloromethane. Moreover, we found two structures
with a H-bond. Calculations predicted the planar structures 2a
and 2b, differing only by the position of the hydrogen atom.
Within the framework of our theoretical model they should be
As expected, the planar structure 1a is the only detectable form
of the acid, in agreement with previous calculations,22 with our
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 2003–2011 | 2005
©

View Article Online
Scheme 1 Conformations of 2-hydroxybenzoic acid 1 and of its anion 2 in the sequence of increasing energy. (Conformers populated in the equilibrium
mixture by more than one percent are underlined; structures for which no energy minimum was obtained are in square brackets).
much better fit for meta- and para-substituted benzoic acids16b (s =
Table 3 Estimated substituent effects in 2-substituted benzoic acids
(in kJ mol−1
)
2.69 kJ mol−1) and even for ortho-substituted acids without hydro-
gen bonds.9 Moreover, when a disagreement of DFT calculated
and experimental substituent effects was observed, these effects
were always overestimated (in absolute values) by calculations,
never underestimated.15a,16,23 We examined the possibility that the
difference could be due to the entropy24 and calculated D1G◦ of
eqn. (1) in spite of our rather bad experience23 with calculation
by statistical thermodynamics within the framework of DFT. We
obtained only a slight improvement to −51.1 kJ mol−1.
Substituent
2-OH
HB + SE SIR PE
Total DEa
Acid (1)
−23
0
0
0
0
0
0
14
0
−4
−1
3
−26.7
−75.6
−48.9
−16.1
−21.7
−5.6
Anion (2) −75
Acidity
Acid (3)
−52
−8
2-NH2
−8
Anion (4) −29
7
Acidity
Acid (5)
−21
15
4
−19
2-COOH
2-SO2NH2
20
37.5
Anion (6) −44
−63.4
−100.9
21.1
The H-bonds in 1a, 2a and 2b are strong according to common
criteria25 (Table 2) but quantitatively these criteria need not give
concordant results.25 According to the distance H. . .O, the H-
bonds in 2a and 2b are much stronger than in 1a, according to
the shift Dm of the m(O–H) frequency, there is little difference.
Values of m (O–H) were related to benzoic acid or to phenol for
the sake of simplicity and these relative values were predicted by
calculations in fair agreement with the IR spectra (Table 2). For
evaluating the effect on acidity, the deciding property of the H-
bond would be its energy defined as an energy difference between
the real structure and any reference structure without the H-bond.
However, the latter cannot be evidently defined in an unambiguous
way.22a We used a rather sophisticated model which can also be
applied to the acid 3 and to the anions 2 and 4. When the carboxyl
group was rotated out by 90◦ and fixed in this conformation (1f),
the H-bond was broken but the conjugation was also inhibited
(SIR). The latter component was estimated from the isomeric 4-
hydroxybenzoic acid (Table S8) and subtracted. In the same way,
the energy of the H-bond in the anion 2a ꢀ 2b was also estimated
(2f). According to Table 2 the H-bond is rather strong in the
acid 1a but much stronger in the anion and makes the acid 1
stronger.
Acidity
Acid (7)
−64
−14 −23
3
13
21 −32
−37
5
Anion (8) −44
−54.9
Acidity
−47
8
−76.0
a Substituent effect in the acid molecule (D2E), or in the anion (D3E), or
the effect on the acidity (D1E), as the case may be.
in equilibrium but separated only by a very low energy barrier of
1.75 kJ mol−1. In agreement with this, two IR bands were observed
(Table S1) differing by 150 cm−1; calculation yielded a difference of
119 cm−1 but the individual frequencies were calculated wrongly.
The agreement of calculated and experimental frequencies was
generally bad, even with the recently suggested correction factor;20
relative values agreed better. Qualitatively, calculations and IR
spectra agree well: Calculations predict that the phenolate anion
2a is more populated and its m(O–H) is at higher frequencies; in
the spectra the higher-frequency band is more intense.
When we calculated the relative acidity D1E, eqn. (1), from
the forms 1a, 2a and 2b present in equilibrium, we got
−48.9 kJ mol−1 compared with the experimental D1G◦ in the
gas phase −58.2 kJ mol−1. The fit is not good compared to the
2006 | Org. Biomol. Chem., 2006, 4, 2003–2011
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
2-Aminobenzoic acid (3)
compared to 97 cm−1 in aniline; this H-bond is weaker in the dimer
of 3a and still weaker in the solid state. For the stronger H-bond in
4a we got Dm 393 cm−1. The values of Dm were also calculated with a
fair fit to experiments (Table 2), although the fit for the frequencies
themselves was much worse. In any case, the calculations predict
the great difference between the weak H-bond in 3a and strong
bond in 4a well, which is further confirmed by the geometric
parameters (Table 2). The energy of the H-bonds was estimated in
the same way as above with the species 1 and 2. In the fixed perpen-
dicular conformations 3g and 4c the H-bonds are broken; for SIR
a correction is calculated from the isomeric 4-aminobenzoic acid
(Table S8) and subtracted. Remarkably only a small difference was
found between these energies for 3a and 4a (Table 2); consequently
the H-bond should have only small effect on the acidity of 3.
The acidity of 3 in water (pK2 = 4.85) and aqueous solvents
was the object of much speculation but the crucial problem
was the presence and population of the zwitterion, 3e or 3f. Its
population in water was estimated differently,26 in any case it
diminishes in less polar solvents.27 Even the X-ray structure is
partly controversial.28 Acidity in the gas phase was interpreted
by a H-bond in the anion,12 not directly proven, but no proof in
favour of a H-bond was found in the chemometrical analysis of
reactivities in various solvents.7 The H-bond in the acid molecule
found by IR spectroscopy in the gas phase was classified as
weak;29 nevertheless it was claimed even in various solutions.21b,30
Semiempirical calculations31 predicted two conformers of 3.
We tried to calculate energies of all possible forms (Scheme 2)
but we reached no energy minima for the zwitterions 3e or 3f,
and also no minima for a conformation near to 3d, or for a form
of anion near to 4b. The ionization is sufficiently represented by
the acid in the conformation 3a and anion in 4a; the forms 3b
and 3c are negligibly populated (Table 1). Both species contain
H-bonds, observed also from the IR spectra, but these H-bonds
are of different strengths as seen from various criteria (Table 2).
The relative acidity D1E of 3 was calculated from the energies
of 3a and 4a as −4.4 kJ mol−1 compared to the experimental12
D1G◦(298) −11.7. The difference may have similar grounds as with
the acid 1. Note that the deviations, although significant for the
individual acids, are of little importance in the perspective of the
whole series.9,16b
Benzene-1,2-dicarboxylic acid (5)
Discussion of the acidity of 5 was almost always restricted4 to
the ratio of the two dissociation constants K1 and K2. (in water
pK1 = 2.94, pK2 = 5.40). A possible H-bond in the monoanion,
requiring a seven-membered ring, was rarely taken into consid-
eration. On the basis of correlation analysis it was assumed in
dimethylformamide but not in water;1b no proof for it was found
in kinetics in various solvents.7 From direct comparison of K1 and
K2, this H-bond was anticipated to be very weak in water, stronger
in dimethyl sulfoxide.32 Conformation of the undissociated acid
5 was investigated in tetrachloromethane on the basis of the
experimental dipole moment and the Cotton–Mouton effect33
with the result that several nonplanar conformers are populated.34
Even in crystal form, the conformation is unexpected35 (somewhat
distorted nonplanar 5d, Scheme 3).
The strength of the H-bonds in 3a and 4a is best characterized
in the IR spectra by the difference of the antisymmetrical and sym-
metrical NH2 frequencies, Dm = mas(NH2) − ms(NH2). When one of
the N–H bonds is engaged in a H-bond, the coupling is weaker and
Dm increases. For the weak H-bond in 3a we found Dm = 133 cm−1
IR investigation was in this case prevented by the low solubility
of the acid in tetrachloromethane or chloroform; the acid salt
Scheme 2 Conformations of 2-aminobenzoic acid and of its anion in the sequence of increasing energy. (See Scheme 1.)
This journal is The Royal Society of Chemistry 2006 Org. Biomol. Chem., 2006, 4, 2003–2011 | 2007
©

View Article Online
Scheme 3 Conformations of phthalic acid and of its monoanion in the sequence of increasing energy. (See Scheme 1.)
could not be investigated because of its disproportionation into
neutral salt and acid. In our opinion, we can rely on the
calculations due to good agreement with the experiments in the
preceding cases. We obtained four forms of the acid, 5a–d, all
of comparable stability, with a H-bond only in 5c. Although this
bond could be strong according to the calculated Dm, its energy is
counterbalanced by unfavourable conformation on the C–O bond.
Let us consider the equilibrium of the four forms 5a to 5d: when 5c
is removed from the equilibrium mixture, the energy would prac-
tically not change, only by some hundredths of kJ mol−1. We thus
take for granted that the H-bond in 5c has no effect on the acidity.
For the monoanion, only one stable form was found (6a) with
a remarkably strong H-bond. The seven-membered ring is almost
planar and the H-bond is symmetrical, with the H-atom situated in
the middle between the two oxygen atoms (Table 2); the geometric
parameters are much more favourable than in 5c. Unfortunately,
experimental determination of Dm was not possible and its calcu-
lation was not reliable similarly as with the preceding H-bonds
in the anions (2a and 2b). We also found no reliable model for
estimating the energy of the hydrogen bonds in 5c and 6a; the fixed
conformers 5f and 6c were of no use since it was not possible to
take into account further simultaneous changes of conformation.
According to the geometry and calculated IR frequency, the H-
bond in 6a is strong and contributes significantly to the acidity of 5.
restricted to dimethylformamide solution suggested a H-bond in
the anion making the acid stronger.1b We searched for various
conformations of 7, which all are nonplanar and somewhat less
exactly defined. Particular attention was given to various possible
conformations of the SO2NH2 group but they have small effect
on the energy and are mostly unstable. Ultimately we found three
conformers, 7a–7c (Scheme 4), forming an equilibrium mixture
but 7a containing a H-bond represents 96% (Table 1). The H-
bond is weak according to its geometry and to calculated Dm. For
the anion only one minimum-energy form was found, 8a, with a
much stronger H-bond.
The calculations were in accord with the spectroscopic results.
Two conformers of 7 were observed, 7a and 7b, but the H-bond in
7a is very weak according to the frequencies mas(NH2) and ms(NH2)
(Table 2). At higher concentrations an intermolecular bond is
observed in addition, probably involving the SO2 group. On the
contrary, the H-bond observed with the tetrabutylammonium salt
is strong and the difference between the NH2 frequencies is striking
(Table 2). Also the intermolecular bond at higher concentrations
is more apparent. We were unable to estimate the energy of the
H-bonds in 7a and 8a. The nonplanar conformers 7d or 8b could
not be fixed in their simple perpendicular conformation since the
H-bonds persisted even at the cost of other deformation of the
molecule. We had to be satisfied with the qualitative conclusion
that the acidity of 7 is strengthened by the H-bond in the anion 8a
while the opposing effect of the H-bond in the acid is negligible.
2-Sulfonamidobenzoic acid (7)
This acid has been studied very little due to its low solubility.
A weak H-bond in the anion was deduced from chemometrical
analysis of the relatively strong acidity in various solvents (pK
in water1a 3.14) but it was not evident from the graphical
representation.1a A more specified chemometrical investigation
Comparison of substituent effects
Previously we attempted to divide the ortho effect observed in
2-substituted benzoic acids into real or fictive components;9 this
analysis was restricted to substituents without acidic hydrogen
2008 | Org. Biomol. Chem., 2006, 4, 2003–2011
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
Scheme 4 Conformations of 2-sulfonamidobenzoic acid and of its monoanion in the sequence of increasing energy. (See Scheme 1.)
atoms. Extension to substituents with a hydrogen atom is not
straightforward since the contribution of the H-bonds is difficult
to estimate. Let us start with the values of substituent effects
separated into the effect in the anion, D3E [eqn. (3)], and in
the acid molecule D2E [eqn. (2)]. These values (Table 3, last
column) are, in our opinion, reliable and can serve as a firm
basis for the analysis. In Fig. 1 D2E are plotted vs. D3E together
with other 2-substituted benzoic acids and with the 3- and 4-
isomers; the purpose of this figure is to show the fundamental
difference between substituent effects in the ortho position and in
the meta and para positions. A linear dependence for meta and
para derivatives is evident (the Hammett line). Its large negative
slope means that the acidity is controlled by a strong polar effect
in the anion and slightly supported by the opposite effect in the
acid molecule.16b Deviations of certain para substituents (NH2 and
N(CH3)2) has been discussed16b,36 and will not be dealt with here.
The ortho derivatives without a hydrogen atom (noted ꢁ) deviate
strongly to the right from the Hammett line; the energy of both
undissociated molecules and anions is raised (steric hindrance)
and the effects on the acidity may partly compensate. Our four
acids with acidic hydrogen atoms deviate still more strongly. Their
anions are stabilized by H-bonds, while in the acid molecules one
observes both destabilizing (2-COOH, 2-SO2NH2) and stabilizing
interactions (2-OH, 2-NH2).
Fig. 1 gives a true picture of substituent effects, which was
not obtainable from the dissociation constants in solution.4–8 We
attempted a still more detailed analysis along the lines given
previously9 but the results are only approximate. Firstly we
defined the polar effect PE (Table 3, the last but one column) as
proportional to the substituent effect in the pertinent 4-isomers.
The proportionality coefficient was subjected to discussion1a,9,37
but is near to unity and the exact value is of little consequence.
We used the value9 of 0.81. The term PE has no exact meaning
in the real world but can serve as a reference; it is evidently
important for the anions 6 and 8. The difference of the total
effect DE and PE may be called the ortho effect and attributed
to at least three components: steric (van der Waals) effects SE,
SIR and the effect of H-bonds. SIR is evidently absent in the
planar molecules of 1, 2, 3, 4 and 6; in the remaining species it was
Fig.
1
Calculated substituent effects with the ionization of or-
tho-substituted benzoic acids: x-axis effects in the unionized acid, eqn (2),
y-axis effects in the anion, eqn (3); ᭜ substituents with an acidic hydrogen
atom, ꢀ substituent without an acidic hydrogen atom, ᭺ meta and ꢀ para
derivatives shown as reference standards.
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 2003–2011 | 2009
©

View Article Online
estimated by the approach developed previously,9,15c referring to
the fixed conformation of the pertinent para isomer (the energies
see Table S8). Their values (Table 3, column 3) have only a limited
impact on the acidity.38 The remaining two components, HB and
SE, cannot, in our opinion, be reliably separated. One can imagine
that when the two parts of the molecule are approaching, they need
some energy to overcome SE (and/or SIR); once the H-bond is
formed, a greater energy is gained back; these steps cannot be
divided. We obtained the assumed sum of HB + SE (Table 3,
column 3) as the difference DE − PE − SIR. Some of these values
may be reliable: the estimates for 1 and 2 are virtually equal to
the estimates for 1a and 2a (Table 2) made in a quite different
way: any SE is negligible. Even for 3 and 4 the agreement is not so
bad. The destabilizing effect in 5 may be classified as steric, while
the stabilizing effect in the anions 6 and 8 is certainly due to H-
bonds overcoming any SE. The substituent effects on the acidity
are perhaps somewhat more reliable than those in the anions and
in the acid molecules, since some components partly compensate.
Summarizing, Table 3 correctly reflects the difference between the
acidity of 1 or 3, controlled by the difference of the H-bonds, and
5 or 7 given by the sum of HB and PE in the anions with smaller
contributions from the acid molecules.
speculate about steric hindrance of solvation of the acid molecules
or specific hydration of the anions but direct proof is not available
to date.
Conclusions
2-Substituted benzoic acids are certainly not a suitable model for
evaluating steric and other substituent effects since these effects
may be dissimilar for various substituents and also different in
the anions and in the acid molecules. Previous analysis based
on the acidities in aqueous solution1,4–8 could meet only with
very limited success. We obtained additional information from
quantum chemical calculations of energies when utilizing their
values as if they were experimental quantities. The main result
was separating the substituent effects in the anion and in the
undissociated molecule. Of the individual factors influencing the
acidity, the H-bonds are very important and can be reliably proven
by a combination of calculations and experiments. Several other
factors remain more or less theoretical constructions since they
refer to fixed conformations (as SIR) or to comparison with
isomers or other molecules (PE and SE).
The whole foregoing analysis was restricted to the behaviour
of isolated molecules while all previous discussions were based
on the dissociation constants in water.4–8 We have some indirect
proof that the strong H-bonds persist even in water. In Fig. 2 the
acidities of substituted benzoic acids in water are compared with
the calculated acidities of isolated molecules; the ortho effect is
again, very clearly, apparent. As expected the meta derivatives
determine the Hammett line; its slope of 0.1 is a measure of
attenuation.12 Some para derivatives (4-NH2, 4-NMe2, Fig. 2 at
the top) deviate—the acids are weaker in water (zwitterions). The
acids 2-COOH and 2-SO2NH2 are also weaker since the H-bond
in the anion is weakened; nevertheless it still exists since these
acids are much stronger than their 4-isomers. In the big deviation
of 2-NH2 the two grounds are combined, zwitterion and H-bond.
The substituent 2-OH deviates to the opposite side, the acid is
stronger in water. Probably the H-bond in the acid molecule has
been more weakened in water than the stronger H-bond in the
anion. All other ortho derivatives deviate in the sense that the
acids are stronger or much stronger (2-NO2) in water. One can
Acknowledgements
This work was supported by the projects Z4 055 0506 of the
Academy of Sciences of the Czech Republic (to P. F. and O. E.)
and MSM0021627501 of the Ministry of Education, Youth and
Sports of the Czech Republic (to J. K.).
References
1 (a) O. Pytela and O. Prusek, Collect. Czech. Chem. Commun., 1999,
64, 1617–1628; (b) O. Pytela and J. Kulha´nek, Collect. Czech. Chem.
Commun., 2002, 67, 596–608.
2 I. Bauerova´ and M. Ludwig, Collect. Czech. Chem. Commun., 2001, 66,
770–784.
3 O. Exner, Correlation Analysis of Chemical Data, Plenum Press, New
York, 1988, pp. 135–137.
4 (a) H. C. Brown, D. H. McDaniel and O. Ha¨fliger, in Determination of
Organic Structures by Physical Methods, E. A. Braude and F. C. Na-
chod, eds, Academic Press, New York, 1955, pp. 567–662; (b) G. S.
Hammond, in Steric Effects in Organic Chemistry, M. S. Newman, ed.,
Wiley, New York, 1956, pp. 425–470; (c) J. Hine, Structural Effects on
Equilibria in Organic Chemistry, Wiley, New York, 1975, Chapter 3–4.
5 (a) J. Shorter, in Advances in Linear Free Energy Relationships,
N. B. Chapman and J. Shorter, eds, Plenum Press, London, 1972, pp.
71–118; (b) T. Fujita and T. Nishioka, Prog. Phys. Org. Chem., 1976,
12, 49–89; (c) U. Berg, R. Gallo, G. Klatte and J. Metzger, J. Chem.
Soc., Perkin Trans. 2, 1980, 1350–1355; (d) A. C. Farthing and B. Nam,
in Steric Effects in Conjugated Systems, G. W. Gray, ed., Butterworths,
London, 1958, pp. 131–140; (e) M. T. Tribble and J. G. Traynham,
J. Am. Chem. Soc., 1969, 91, 379–388.
6 M. Charton, in Similarity Models in Organic Chemistry, Biochemistry
and Related Fields, R. I. Zalewski, T. M. Krygowski and J. Shorter, eds,
Elsevier, Amsterdam, 1991, p. 629–687.
7 M. H. Aslam, A. G. Burden, N. B. Chapman, J. Shorter and M.
Charton, J. Chem. Soc., Perkin Trans. 2, 1981, 500–508.
8 (a) M. Charton, Prog. Phys. Org. Chem., 1971, 8, 235–317; (b) V.
Nummert and M. Pirsalu, Collect. Czech. Chem. Commun., 2002, 67,
1833–1857; (c) K. Bowden and G. E. Manser, Can. J. Chem., 1968, 46,
2941–2944.
9 S. Bo¨hm, P. Fiedler and O. Exner, New J. Chem., 2004, 28, 67–74.
10 (a) M. Decouzon, P. Ertl, O. Exner, J.-F. Gal and P.-C. Maria, J. Am.
Chem. Soc., 1993, 115, 12071–12078; (b) M. Decouzon, J.-F. Gal, P.-C.
Maria, S. Bhm, P. Jime´nez, M. V. Roux and O. Exner, New J. Chem.,
1997, 21, 561–573.
Fig. 2 Substituent effects on the acidity of substituted benzoic acids:
x-axis calculated energies of isolated molecules, y-axis experimental
Gibbs energies in water; ᭜ substituents with an acidic hydrogen atom,
ꢀ substituent without an acidic hydrogen atom, ᭺ meta and ꢀ para
derivatives given as reference.
2010 | Org. Biomol. Chem., 2006, 4, 2003–2011
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
11 (a) S. Bo¨hm and O. Exner, Chem.–Eur. J., 2000, 6, 3391–3398; (b) S.
Bo¨hm and O. Exner, New J. Chem., 2001, 25, 250–254.
12 T. B. McMahon and P. Kebarle, J. Am. Chem. Soc., 1997, 99, 2222–
2230.
13 (a) R. T. McIver, Jr. and J. H. Silvers, J. Am. Chem. Soc., 1973, 95,
8462–8464; (b) S. Bhm, M. Decouzon, O. Exner, J.-F. Gal and P.-C.
Maria, J. Org. Chem., 1994, 59, 8127–8131.
14 (a) A. Pross, L. Radom and R. W. Taft, J. Org. Chem., 1980, 45, 818–
826; (b) O. Exner, Org. Reactivity (Tartu), 1995, 29, 1–6.
15 (a) O. Exner and P. Nausˇ, J. Phys. Org. Chem., 2000, 13, 693–698; (b) O.
Exner and S. Bo¨hm, Chem.–Eur. J., 2002, 8, 5147–5152; (c) O. Exner
and S. Bo¨hm, J. Phys. Org. Chem., 2004, 17, 124–130.
16 (a) O. Exner and S. Bo¨hm, J. Chem. Soc., Perkin Trans. 2, 2000, 1994–
1999; (b) O. Exner and S. Bo¨hm, J. Org. Chem., 2002, 67, 6320–6327.
17 P. George, M. Trachtman, C. W. Bock and A. M. Brett, J. Chem. Soc.,
Perkin Trans. 2, 1976, 1222–1227.
22 (a) P. I. Nagy, W. J. Dunn, III, G. Alagona and C. Ghio, J. Phys.
Chem., 1993, 97, 4628–4642; (b) H. Lampert, W. Mikenda and A.
Karpfen, J. Phys. Chem., 1996, 100, 7418–7425; (c) J. Catala´n and
J. I. Fernandez-Alonso, J. Mol. Struct., 1975, 27, 59–65; (d) F. T. T.
Huque and J. A. Platts, Org. Biomol. Chem., 2003, 1, 1419–
1424.
23 S. Bo¨hm and O. Exner, J. Comput. Chem., 2006, 27, 571–577.
24 The experimental gas-phase acidities are obtained as DG◦ and their
transformation into D1H◦ is only approximate: S. G. Lias, J. E.
Bartmess, J. F. Liebman, J. L. Holmes, R. D. Levin and W. G. Mallard,
J. Phys. Chem., 1988, 17, suppl. 1.
25 S. J. Grabowski, J. Phys. Org. Chem., 2004, 17, 18–31.
26 (a) L. M. Schwartz, R. I. Gelb, J. Mumford-Zisk and D. A. Laufer,
J. Chem. Soc., Perkin Trans. 2, 1987, 453–460; (b) V. I. Zaionts and
E. Sh. Gutshabash, Zh. Org. Khim., 1982, 18, 1912–1917.
27 I. Mentre, Ann. Chim. (Paris), 1973, 8, 115–122.
18 G. Pilcher, in The Chemistry of Functional Groups. Supplement B: The
Chemistry of Acid Derivatives, Vol. 2, S. Patai, ed., Wiley, New York,
1992, pp. 51–94.
28 (a) C. J. Brown and M. Ehrenberg, Acta Crystallogr., Sect. C: Cryst.
Struct. Commun., 1985, 41, 441–443; (b) H. Takazawa, S. Ohba and
Y. Saito, Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 1986, 42,
1880–1881.
19 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A.
Robb, J. R. Cheeseman, J. A. Montgomery, Jr., T. Vreven, K. N.
Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone,
B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A. Petersson, H.
Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa,
M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X.
Li, J. E. Knox, H. P. Hratchian, J. B. Cross, C. Adamo, J. Jaramillo,
R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi,
C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth,
P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D.
Daniels, M. C. Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K.
Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S.
Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz,
I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y.
Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, C. Gonzalez and J. A. Pople, Gaussian 03, Revision
B.03, Gaussian, Inc., Pittsburgh, PA, 2003.
29 Yu. Ya. Kharitonov and I. I. Oleinik, Dokl. Akad. Nauk SSSR, 1987,
294, 151–154.
30 Yu. G. Kondrashina, S. G. Vul’fson and A. P. Timosheva, Izv. Akad.
Nauk. SSSR, Ser. Khim., 1994, 859–861.
31 M. A. Palafox, M. Gil and J. L. Nunez, Vib. Spectrosc., 1993, 6, 95–
105.
32 P. J. Choi, K. A. Petterson and J. D. Roberts, J. Phys. Org. Chem., 2002,
15, 278–286.
33 A. P. Timosheva, S. G. Abdullina, S. G. Vul’fson and L. V. Ermolaeva,
Izv. Akad. Nauk. SSSR, Ser. Khim., 1993, 1568–1571.
34 In ref. 33 the experimental data were complemented by semiempirical
calculations but it is not clear which conformations were actually con-
sidered. The only safe conclusion seems to be that several conformers
are in equilibrium and that they are not planar; this is in agreement
with our results.
35 O. Ermer, Helv. Chim. Acta, 1981, 64, 1902–1909.
36 S. Bo¨hm, P. Parˇ´ık and O. Exner, New J. Chem., 2006, 30, 384–391.
37 M. Charton, Can. J. Chem., 1960, 38, 2493–2499.
38 The generally small importance of SIR was analyzed in detail: J.
Kulha´nek, S. Bo¨hm, K. Pala´t, Jr. and O. Exner, J. Phys. Org. Chem.,
2004, 17, 686–693 and references therein.
20 M. P. Andersson and P. Uvdal, J. Phys. Chem. A, 2005, 109, 2937–2941.
21 (a) A. E. Lutskii and Yu. I. Dolzhenko, Zh. Obshch. Khim., 1980,
50, 440–446; (b) A. Ya. Kaminskii, V. M. Ivanova, S. S. Gitis, E. G.
Kaminskaya and L. I. Khabarova, Reakts. Sposobn. Org. Soedin., 1971,
8, 343–362.
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 2003–2011 | 2011
©