Spectroscopy of hydrothermal reactions 23: The effect of OH substitution on the rates and mechanisms of decarboxylation of benzoic acid

Article abstract of DOI:10.1021/jp0220231

The decarboxylation rates of aqueous benzoic acid and 12 mono-, di-, and trihydroxy derivatives of benzoic acid were compared by using spectra from a flow reactor FTIR spectroscopy cell operating at 275 bar in the temperature range of 120-330 °C. Each compound was investigated at its natural pH and as the neutral acid (pH = 1.3-1.5). The decarboxylation reactions followed the first-order (or pseudo-first-order) rate law enabling the rate constants and corresponding Arrhenius parameters of the undissociated acids to be obtained. Based on the half-lives of the reactions at 200 °C, the thermal stability of the OH substituted benzoic acids follow the order: 2,4,6 > 2,4 > 2,3,4 > 2,6 > 2,5 > 2,3 > 3,4,5 > 2 > 3,4 > 4. Solutions of 3,5-dihydroxybenzoic and 3-hydroxybenzoic acids and unsubstituted benzoic acid had the highest thermal stability, whereas no decarboxylation was observed up to 330 °C at a residence time of about 45s. In general, the rate order is multiple ortho, para-OH substitution > ortho substitution > para substitution > meta substitution. The range of activation energies for the decarboxylation of OH substituted benzoic acids is 90-97 kJ/mol, and the rate differences are controlled mainly by activation entropy. The transition state structures were determined using density functional theory. Starting from the anti carboxylic hydrogen conformers in the gas phase, the activation energies to the transition state structures having the four-member C-C(O)-O-H ring are 213-260 kJ/mol using B3LYP/6-31G//B3LYP/6-31G and 202-246 kJ/mol using B3LYP/6-31+G(d,p)//B3LYP/6-31G(d). Incorporation of one water molecule forms a six-member cyclic structure, which dramatically reduces the activation energy by about 120-130 kJ/mol using B3LYP/6-31G//B3LYP/6-31G and by about 75 kJ/mol using B3LYP/6-31+G(d,p)//B3LYP/6-31G(d). In the water-catalyzed transition state structure, the water molecule acts as a bridge linked by two hydrogen bonds which enables concerted proton transfer and C-(CO₂H) bond cleavage to occur. Although the calculated activation energy approximately follows the trend of the experimental half-lives, the experimental activation entropy appears to dominate in determining the rates.

Full text of DOI:10.1021/jp0220231

J. Phys. Chem. A 2003, 107, 2667-2673
2667
Spectroscopy of Hydrothermal Reactions 23: The Effect of OH Substitution on the Rates
and Mechanisms of Decarboxylation of Benzoic Acid
Jun Li and Thomas B. Brill*
Department of Chemistry and Biochemistry, UniVersity of Delaware, Newark, Delaware 19716
ReceiVed: September 9, 2002; In Final Form: January 15, 2003
The decarboxylation rates of aqueous benzoic acid and 12 mono-, di-, and trihydroxy derivatives of benzoic
acid were compared by using spectra from a flow reactor FTIR spectroscopy cell operating at 275 bar in the
temperature range of 120-330 °C. Each compound was investigated at its natural pH and as the neutral acid
(pH ) 1.3-1.5). The decarboxylation reactions followed the first-order (or pseudo-first-order) rate law enabling
the rate constants and corresponding Arrhenius parameters of the undissociated acids to be obtained. Based
on the half-lives of the reactions at 200 °C, the thermal stability of the OH substituted benzoic acids follow
the order: 2,4,6 > 2,4 > 2,3,4 > 2,6 > 2,5 > 2,3 > 3,4,5 > 2 > 3,4 > 4. Solutions of 3,5-dihydroxybenzoic
and 3-hydroxybenzoic acids and unsubstituted benzoic acid had the highest thermal stability, whereas no
decarboxylation was observed up to 330 °C at a residence time of about 45s. In general, the rate order is
multiple ortho, para-OH substitution > ortho substitution > para substitution > meta substitution. The range
of activation energies for the decarboxylation of OH substituted benzoic acids is 90-97 kJ/mol, and the rate
differences are controlled mainly by activation entropy. The transition state structures were determined using
density functional theory. Starting from the anti carboxylic hydrogen conformers in the gas phase, the activation
energies to the transition state structures having the four-member C-C(O)-O-H ring are 213-260 kJ/mol
using B3LYP/6-31G//B3LYP/6-31G and 202-246 kJ/mol using B3LYP/6-31+G(d,p)//B3LYP/6-31G(d).
Incorporation of one water molecule forms a six-member cyclic structure, which dramatically reduces the
activation energy by about 120-130 kJ/mol using B3LYP/6-31G//B3LYP/6-31G and by about 75 kJ/mol
using B3LYP/6-31+G(d,p)//B3LYP/6-31G(d). In the water-catalyzed transition state structure, the water
2
molecule acts as a bridge linked by two hydrogen bonds which enables concerted proton transfer and C-(CO H)
bond cleavage to occur. Although the calculated activation energy approximately follows the trend of the
experimental half-lives, the experimental activation entropy appears to dominate in determining the rates.
9
8
Introduction
buffer or a fairly strong base was present, however, the
decarboxylation of benzoic acid was accelerated. Electron-
1
0-13
14
Reaction chemistry of organic compounds in hydrothermal
environments has received much attention because of its
importance for understanding the formation of fossil fuels and
potential green chemical processes.1-6 Many classes of organic
compounds that are regarded to be unreactive in aqueous
solution undergo diverse chemical reactions when the temper-
ature is increased above the boiling temperature of water. Water
molecules can participate in the reactions in various ways such
withdrawing substituents, such as OH
the thermal stability of benzoic acid. The decarboxylation rate
of OH substituted benzoic acids at hydrothermal conditions is
and NO2, lower
11
pH dependent. Gallic acid (3,4,5-trihydroxybenzoic acid), a
naturally occurring aromatic acid, decarboxylates faster at pH
) 7.0 than at pH ) 4.3, which means that the anion has a faster
decarboxylation rate than the neutral acid. When a strong acid
was added to a solution of 2,4-dihydroxybenzoic acid or 2,4,6-
trihydroxybenzoic acid, the decarboxylation rate initially
13
7
12
as a catalyst, reactant, and solvent, as a result of its dramatic
changes in density, dielectric constant, and ionization constant
with increasing temperature, as well as its effect on the solubility
of the solute.
increased with increasing acidity and then remained constant.
1
5,16
17
Kinetic hydrogen
and carbon isotopic effects were used
to clarify the reaction mechanism. The rate-determining step
changed from slow proton transfer to C-C bond cleavage as
the solution was made more acidic. Segura et al.14 did not
observe a significant solvent effect in the decarboxylation rate
of dinitrobenzoate anions. Hydrolytic ionic reaction chemistry
dominates thermal (free-radical) routes at subcritical hydrother-
mal conditions. Generally, a carbanion reaction mechanism (SE1)
has been assumed in the decarboxylation of substituted and
Aromatic carboxylic acids are among the primary compounds
of interest in the formation of fossil fuels and in determining
the organic carbon distribution in natural aqueous systems.
Detailed information about the reaction rates and pathways of
aromatic carboxylic acids at hydrothermal conditions are
therefore needed for modeling of these systems. The decar-
boxylation of benzoic acid and several OH substituted benzoic
acids has been studied. Benzoic acid is more stable at hydro-
thermal conditions than the OH substituted benzoic acids.
1
8,19
unsubstituted benzoic acids.
1
8,19
Ruelle’s theoretical study
on the decarboxylation of
8
Katrizky et al. observed that only 0.1% of benzoic acid
benzoic acid and salicylic acid indicated how water molecules
can catalyze the reaction. Because of the low level of theory
and computation power at the time, his calculation was restricted
and did not agree well with experimental results. A recent
converted to benzene after 6 h at 350 °C. When a mineral redox
*
To whom correspondence should be addressed. brill@udel.edu.
1
0.1021/jp0220231 CCC: $25.00 © 2003 American Chemical Society
Published on Web 03/21/2003

2668 J. Phys. Chem. A, Vol. 107, No. 15, 2003
Li and Brill
computation study20 of aromatic nucleophilic substitution of
halobenzenes in the gas phase has shown that SNAr proceeds
via a single-step concerted mechanism without the formation
of a stable σ complex. Our experimental and computational
comparison of decarboxylation of aliphatic dicarboxylic acids
showed that decarboxylation proceeds via a concerted proton
transfer and C-C cleavage step in the transition state structure
in which one water molecule is involved. In the present paper,
we extend these studies on the decarboxylation of carboxylic
acids beyond aliphatic into aromatic acids. The reactivity and
kinetics of decarboxylation of benzoic acid and OH substituted
benzoic acids are measured in situ using a flow reactor with
FT-IR spectroscopy and are described by density functional
theory.
than were used in this work,26 but it forms an insoluble
precipitate that was not found here.
Conformational analyses of the acids and the transition state
structures for decarboxylation were performed by density
21
27
functional theory calculations using Gaussian 98 software at
2
8,29
the level of B3LYP theory.
The 6-31G basis set was used
for conformation analyses. The ground state and transition state
structures were first optimized with the 6-31G basis set and
then further optimized with the 6-31G(d) basis set. Vibrational
frequency analyses were conducted to confirm that the optimized
geometry was a local minimum or a transition state and provided
thermal corrections to the thermal energy, enthalpy, and free
energy. The effects of temperature and pressure were considered
at 200 °C and 275 bar at the level of B3LYP/6-31G(d) theory.
Finally, activation energies were calculated based on single-
point energy calculations at the level of B3LYP/6-31+G(d,p)
theory based on the thermal correction to 200 °C and 275 bar
at the level of B3LYP/6-31G(d) with the scaling factor 0.9804.
It was found that these temperature and pressure corrections
had a negligible effect.
Experimental Section
All acids (benzoic, 2-hydroxybenzoic, 3-hydroxybenzoic,
4
2
-hydroxybenzoic, 2,3-dihydroxybenzoic, 2,4-dihydroxybenzoic,
,5-dihydroxybenzoic, 2,6-dihydroxybenzoic, 3,4-dihydroxy-
benzoic, 3,5-dihydroxybenzoic, 2,3,4-trihydroxybenzoic, 2,4,6-
trihydroxybenzoic, and 3,4,5-trihydroxybenzoic acids) were
purchased from Sigma-Aldrich and used without further puri-
fication. Milli-Q deionized water was sparged with compressed
Ar before use to expel the atmospheric gases. Solution pH values
of aqueous OH substituted benzoic acids and the acidified
solutions produced by titration with HCl were recorded with
an Orion model 330 pH meter. Because of the relatively low
solubility of the acids in pure water, the concentrations of the
solutions were 0.025m, except for 2-hydroxybenzoic acid
20
A comparison by Bach et al. has shown that the activation
barriers were overestimated at the MP2 level and underestimated
at the B3LYP level, but the difference at these levels is small.
The results are least accurate at the Hartree-Fock level which
was not used here. The absolute value of the barrier is not vital
for this paper because the focus is on the relative comparison
of the activation energies.
Results and Discussion
Decarboxylation Kinetics. The overall decarboxylation reac-
tion of OH substituted benzoic acids followed eq 1
(0.01m) and 2,3,4-trihydroxybenzoic acid (0.015m).
The flow reactor FT-IR spectroscopy cell constructed from
titanium with sapphire windows and gold foil seals has been
22,23
(HO) ArCO H f (HO) Ar + CO
2
(1)
described in detail elsewhere.
The temperature and pressure
n
2
n
were controlled within (1 °C and (1 bar, respectively. The
chosen flow rate (0.1-1.0 mL/min range) was controlled with
an accuracy of 1% by the use of an Isco syringe pump.
Correction of the flow rate was made to account for the density
change with temperature. Data were collected in the 120-280°C
range, although 3-hydroxybenzoic acid, 3,5-hydroxybenzoic
acid, and unsubstituted benzoic acid were subjected to temper-
atures up to 330 °C and found not to react at a residence time
A large difference in the hydrothermal stability exists however
as a result of the position of the OH substituents on the benzene
ring. For example, 3-hydroxybenzoic and 3,5-dihydroxybenzoic
acids did not decarboxylate below 330 °C with a residence time
of 45 s, whereas 2,4,6-trihydroxybenzoic acid began to decom-
pose at 50 °C.
Reaction 1 was clearly indicated by in situ FT-IR spectros-
copy. As an example, Figure 1 presents FT-IR spectra of the
aqueous 2,4-dihydoxybenzoic acid solution (0.025m) at different
temperatures, 275 bar, and a residence time of 45 s. Unfortu-
nately, no absorbance for aqueous CO2 was observed for
aqueous neutral and basic solutions because of the low solubility
of the OH substituted benzoic acids and hydrolysis of CO2.30
The hydrolysis of CO2 in the acidic solutions (either natural or
obtained by adding HCl) was calculated to be negligible in this
work, even through these benzoic acids are weak.
-
1
of 45 s. Transmission IR spectra were recorded at 4 cm
resolution with a Nicolet 560 Magna FTIR spectrometer and
an MCT-A detector. Background spectra recorded for pure water
at the same conditions were subtracted. Thirty-two spectra were
summed at each condition and the rate data reported herein are
the average of three replicated measurements.
During the decarboxylation reaction, only the asymmetric
stretching mode of aqueous CO2 centered at 2343 cm- was
observed in the band-pass of the sapphire windows. To obtain
the kinetic parameters, the band area of CO2 was converted into
concentration at each condition by using the Beer-Lambert Law
and the previously determined molar absorptivity of aqueous
CO2.24 Weighted least-squares regression with a 95% confi-
dence interval was performed in which the statistical weight
1
The rate constants of the undissociated acids were determined
in acidified solutions where the predominant species is the
neutral acid. The solution pH25 values of 1.3 or 1.5 prepared
by adding HCl were suitable for real time FT-IR spectroscopy
at hydrothermal conditions because 99% of the acid is in the
neutral form at 25 °C. The ratio is even higher when extrapola-
tion is made to higher temperatures, because the ionization
constant of H2O increases and dissociation constant of the acids
decreases with increasing temperature. The dissociation con-
stants of the acids needed for the charge balance analysis were
25
2
was set to be 1/σ , where σ is the standard deviation of the
variables.
An advantage of the use of the small-scale flow reactor is
that any precipitation of the reactant or products is readily
detected in the form of disruption of the flow rate. The relatively
low solubility of benzoic acid and its derivatives in water might
result in such an event. However, no irregularity was found in
the flow rates in any of the experiments. Decomposition by the
loss of -OH is a reaction known to occur at higher temperatures
31
obtained by extrapolation of the ionization constant and
3
2
33
specific volume of water from room temperature to higher
34
temperatures using the iso-Coulombic method. Corrosion and
leakage are the two leading factors which limit the use of flow

Spectroscopy of Hydrothermal Reactions 23
J. Phys. Chem. A, Vol. 107, No. 15, 2003 2669
Figure 1. In situ FT-IR spectra of the 0.025m 2,4-dihydroxybenzoic
acid solution at different temperatures, a pressure of 275 bar, and a
residence time of about 45 s.
Figure 2. Rate plot for the decarboxylation of the 0.025m 2,4-
dihydroxybenzoic acid solution at different temperatures.
reactor spectroscopy at even more acidic solutions at high
temperature and pressure.
Based on the rate of formation of CO2, a plot of ln([acid]0-
reason that, at fast flow rate and high temperature, the cell is
probably at or beyond the limit of the assumption of uniform
2
3
temperature. The first-order rate law applied to the data in
Figure 2 and the resulting rate constants of OH substituted
benzoic acids at their natural pH and more acidic conditions
are listed in Table 1. The variation of the rate constants with
temperature for the undissociated acids followed the Arrhenius
equation. The Arrhenius plots and resulting parameters, as well
[CO2]T) vs residence time (Figure 2 is an example) provides
the decarboxylation rates, where [acid]0 is the initial concentra-
tion of the benzoic acid. At low temperature (low extent of
reaction), the rates extrapolate well to the initial concentration.
At high temperature, this extrapolation is not accurate for the
3
TABLE 1: Observed First-Order or Pseudo-First-Order Rate Constants (kobs × 10 ) of Mono-, Di-, and Trihydroxybenzoic
Acids at Different Values of the Solution pH
2
-hydroxybenzoic acid(0.01m)
4-hydroxybenzoic acid(0.025m)
2,3-dihydroxybenzoic acid(0.025m)
temp/°C
pH25 ) 1.34
pH25 ) 2.68a
temp/°C
pH25 ) 1.52
pH25 ) 3.23a
temp/°C
pH25 ) 1.42
pH25 ) 2.50a
210
220
230
240
250
260
270
7.09(0.19
10.61(0.37
17.75(0.59
27.35(0.51
40.11(0.74
210
220
230
240
250
260
270
1.42(0.08
2.44(0.09
3.68(0.08
5.91(0.17
8.15(0.30
190
200
210
220
230
240
250
6.498(0.21
10.97(0.27
17.74(0.38
26.95(0.49
40.01(0.90
9.67(0.19
15.82(0.30
27.93(0.45
38.89(1.23
52.91(0.88
77.23(3.04
3.67(0.08
5.19(0.20
8.41(0.21
12.81(0.31
19.21(0.50
28.21(0.68
7.46(0.33
12.39(0.22
20.20(0.33
35.48(0.86
52.35(1.23
75.56(1.68
2
,4-dihydroxybenzoic acid(0.025m)
2,5-dihydroxybenzoic acid(0.025m)
2,6-dihydroxybenzoic acid(0.025m)
temp/°C
pH25 ) 1.32
pH25 ) 2.57a
temp/°C
pH25 ) 1.35
pH25 ) 2.52a
temp/°C
pH25 ) 1.40
pH25 ) 2.04a
130
140
150
160
170
180
10.99(0.26
22.63(0.37
47.69(1.21
77.53(1.60
141.91(5.70
223.2(10.5
7.36(0.08
14.45(0.16
27.98(0.31
54.75(1.11
86.79(1.81
148.3(3.8
190
200
210
220
230
240
4.60(0.19
7.67(0.14
13.82(0.38
20.70(0.37
120
130
140
150
160
170
180
2.62(0.04
6.13(0.13
12.86(0.15
21.99(0.19
38.28(0.39
66.26(0.65
3.26(0.05
6.60(0.11
13.74(0.17
26.19(0.37
51.09(0.77
86.73(2.55
8.5(0.11
13.83(0.17
22.39(0.61
37.86(1.49
51.57(1.92
250
3
,4- dihydroxybenzoic acid(0.025m)
3,4,5-triihydroxybenzoic acid(0.025m)
2,3,4-triihydroxybenzoic acid(0.015m)
temp/°C
pH25 ) 1.45
pH25 ) 3.13a
temp/°C
pH25 ) 1.51
pH25 ) 3.11a
temp/°C
pH25 ) 1.36
pH25 ) 2.68a
210
220
230
240
250
260
3.76(0.22
5.29(0.20
8.6(0.21
13.65(0.21
20.29(0.52
200
210
220
230
240
250
3.24(0.22
5.51(0.22
9.40(0.21
14.97(0.19
23.48(0.55
36.05(0.65
130
140
150
160
170
180
190
200
8.12(0.06
16.17(0.17
30.67(0.19
55.91(0.71
94.82(1.95
133.6(2.9
6.16(0.11
10.02(0.28
17.51(0.50
22.43(0.49
38.22(1.85
20.24(0.26
36.14(0.40
61.39(0.63
101.9(2.0
160.9(4.9
240.4(8.2
17.23(0.52
30.06(1.01
44.68(1.24
62.06(1.30
93.74(2.49
145.5(4.3
2
2
2
60
70
80
a
Natural solution pH.

2670 J. Phys. Chem. A, Vol. 107, No. 15, 2003
Li and Brill
TABLE 2: Experimentally Determined Arrhenius Parameters and Calculated Activation Parameters with and withoutH
2
O
Incorporation for OH Substituted and Unsubstituted Benzoic Acids
calc E
a
/kJ mol-
1
B3LYP/6-31G//
B3LYP/6-31Gb
B3LYP/6-31+G(d,p)//
c
experimental
∆S‡/J K mol-1
B3LYP/6-31G(d)
-1
‡
-1
-1
t
1/2/s
without
with
without
with
∆S /J K mol
1
ln(A,s-1
d
acid
benzoic acid
E
a
/kJ mol-
)
200 °C
200 °C
H
2
O
H
2
O
H
2
O
H
2
O
200 °C
259.18 131.97 245.94 175.24
172.34 225.26 95.56 217.89 138.71
260.32 133.74 247.79 176.38
493.93 247.46 117.49 234.11 158.40
91.06 227.23 102.93 219.57 144.78
-105.80
-108.20
-104.35
-110.33
-106.66
-108.22
-106.53
-102.82
-110.05
-106.07
-110.72
-107.03
-107.14
2
3
4
2
2
2
2
3
3
3
2
2
-hydroxybenzoic acid
-hydroxybenzoic acid
-hydroxybenzoic acid
,3-dihydroxybenzoic acid
,4-dihydroxybenzoic acid
,5-dihydroxybenzoic acid
,6-dihydroxybenzoic acid
,4-dihydroxybenzoic acid
,5-dihydroxybenzoic acid
92.04(5.32 17.88(1.26
-108.41
91.54(1.50 16.70(0.34
96.69(2.58 19.70(0.63
93.20(2.83 23.34(0.81
96.13(3.95 19.58(0.98
92.12(3.16 22.07(0.89
95.86(5.40 18.31(1.28
-118.22
-93.27
-63.01
-94.27
-73.57
-104.83
0.98 213.50
88.86 230.36 104.76 221.19 146.08
2.66 228.86 98.93 212.60 130.49
89.34 204.99 128.44
296.04 246.66 118.29 234.92 159.89
259.02 131.46 246.31 174.01
156.70 243.05 115.06 231.97 156.45
,4,5-trihydroxybenzoic acid 93.20(.99
18.27(0.92
-105.16
-73.82
-54.95
,3,4-trihydroxybenzoic acid 89.87(2.33 22.04(0.68
1.55 217.87
0.02 219.93
94.05 208.94 133.09
89.02 202.20 118.12
a
a
,4,6-trihydroxybenzoic acid 82.01
24.31
a
From Schubert, W. M.; Gardner, J. D. J. Am. Chem. Soc. 1953, 75, 1401. b At 25 °C and 1 bar. c At 200 °C and 275 bar. d Incorporating one
water molecule at the level of B3LYP/6-31G(d) theory.
stronger hydrogen bonding between the CO2H group and NO2
group compared to an OH substituent. The similarity of the
activation energies for OH substituted benzoic acids suggests
that decarboxylation proceeds via an analogous transition state
structure in all cases. The difference in the decarboxylation rates
originates primarily from the contribution of the activation
entropy.
The effectiveness of o- and p-OH substitution in promoting
decarboxylation can be explained by the resonance contribution.
The OH substituent is electron-withdrawing. The developing
negative charge on the ipso carbon atom in the transition state
can be transferred to the ortho or para position via the resonance
effect, thereby stabilizing the transition state structure. The
resonance effect is indicated by the following transition state
structure calculations using density functional theory.
Geometries and Energetics. The conformational analyses
in gas phase of benzoic acid and OH substituted benzoic acids
were performed at the level of B3LYP/6-31G theory. The
number of conformers found for each of the following acids is
given parenthetically: benzoic (2), 2-hydroxybenzoic (7), 3-hy-
droxybenzoic (8), 4-hydroxybenzoic (4), 2,3-dihydroxybenzoic
Figure 3. Arrhenius plot for the decarboxylation of OH substituted
undissociated benzoic acids.
(11), 2,4-dihydroxybenzoic (14), 2,5-dihydroxybenzoic (14), 2,6-
dihydroxybenzoic (6), 3,4-dihydroxybenzoic (12), 3,5-dihy-
droxybenzoic (8), 2,3,4-trihydroxybenzoic (16), 2,4,6-trihydroxy-
benzoic (6), and 3,4,5-trihydroxybenzoic acids (10). (Details
can be found in Table S, in the Supporting Information). Nagy
as half-lives of the decarboxylation reactions, are shown in
Figure 3 and Table 2, respectively. The following conclusions
were drawn from these data: (1) the effect of a single OH
substitutent on the decarboxylation rate follows the order: o-
35,36
et al.
found 2, 8, and 4 conformers for benzoic, 2-hydroxy-
>
p > m-; (2) when a second OH substituent exists, if the first
benzoic, and 4-hydroxybenzoic acid, respectively. Using density
functional theory, we found that one of their eight conformers
for 2-hydroxybenzoic acid (structure 1), in which all of the atoms
lie in one plane, is not a stationary conformer but is a second-
order saddle point on the potential energy surface. Shapley et
substitution occurs at the ortho position, the decarboxylation
rate order is: o-,p- > o-,o > o-,m-; if the first OH substituent
is at the meta position, the order is the same as in (1), i.e., m-,o
>
m-,p- > m-,m-; (3) for three OH substituents, substitution at
3
7
the ortho and para positions causes the fastest decarboxylation
rate; (4) meta substitution (both single and double) has only a
very small effect on the decarboxylation rate and leads to
stability comparable to that of benzoic acid.
al. compared the hydrogen bonding and solvation energies of
o-, m-, and p-hydroxybenzoic acids.
From our conformational analyses, we have drawn the
following conclusions: (1) The anti carboxylic conformers are
energetically higher than the syn carboxylic conformers. (2) Anti
carboxylic conformers have a nonplanar arrangement of -OH
groups except when the substitution occurs at single ortho
position. The intramolecular hydrogen bonding between the
carboxylate group and one ortho hydroxyl group leads to a
planar conformation for mono ortho OH substituted benzoic
acids. The carboxylate group shares the same plane as the
The activation energies of OH substituted benzoic acids fall
into a narrow range of 90-97 kJ/mol with an average of 93.5
kJ/mol, which resembles the previously reported value for gallic
acid of 116 kJ/mol at pH ) 4.3,1 but are smaller than those of
NO2 substituted benzoic acids. 4 The larger activation energy
of o-NO2 substituted benzoic acids, for example, can be
explained by the larger steric effect of the NO2 group and
11

Spectroscopy of Hydrothermal Reactions 23
J. Phys. Chem. A, Vol. 107, No. 15, 2003 2671
Figure 4. Transition state structures of benzoic acid and monohydroxybenzoic acids with and without one water molecule (B3LYP/6-31G//B3LYP/
-31G). Bond distances are given in angstroms.
6
benzene ring when two OH groups are placed at the ortho
positions (2,6-dihydroxybenzoic and 2,4,6-trihydroxybenzioc
acids) at the level of B3LYP/6-31G theory; however, this
conformation was a first-order saddle point at the B3LYP/6-
bond shortens by 0.01-0.02 Å, whereas OH substitution on
the meta position has no effect. When one water molecule is
involved in the formation of the transition state structure, the
activation energies are reduced by about 120-130 kJ/mol with
B3LYP/6-31G//B3LYP/6-31G or by about 75 kJ/mol with
B3LYP/6-31+G(d,p)//B3LYP/6-31G(d) compared to the free
acid, and a six-member cyclic structure results, which is
common for the decarboxylation of aliphatic and aromatic
carboxylic acids. Incorporation of two water molecules thus
forming an eight-member cyclic structure or three water
molecules forming ten-member cyclic structure is possible, but
involvement of more than one water molecule in the transition
state structure is not as effective in reducing the activation barrier
toward decarboxylation as is the initial water molecule.3 The
effect of the basis set on the activation barriers is different for
the transition state structures with and without inclusion of a
water molecule. The larger basis set decreases the activation
barrier of the transition state structure by about 10 kJ/mol when
no water molecule is present and increases it by about 40 kJ/
mol with inclusion of the water molecule. The exceptions are
31G(d) level. (3) The anti carboxylic conformations with the
dihedral angles of 90° and 0° between the carboxylate group
and benzene ring are transition state structures for the rotation
of carboxylate group. The potential energies at 0° are even lower
than those of the stationary conformers. (4) The differences in
the potential energies for the anti carboxylic conformations with
dihedral angles of the carboxylate group and benzene ring of
0
-90° are small at the level of B3LYP/6-31G theory, which
means that the rotary energy barriers for the anti carboxylate
group are small. (5) Two of the conformers for 3,4,5-trihy-
droxybenzoic acid have structures 2 and 3 because of the steric
effect of the OH groups on the ring. (6) The carboxylate group
and the benzene ring of the syn conformer of 2,6-dihydroxy-
benzoic acid are not coplanar, as is shown by structure 4.
8,39
2,6-dihydroxybenzoic and 2,4,6-trihydroxybenzoic acids where
the ground state conformation depends on the basis set used.
The calculated activation entropies at 200 °C and 275 bar using
the B3LYP/6-31G(d) level are relatively similar (about -107 J
-
1
-1
K
mol ). Therefore, by computation of the structures in the
gas phase, the activation barrier, rather than the activation
entropy, appears to be mainly responsible for the difference in
the rate constants.
Transition State Structures. Transition state structures for
benzoic acid and OH substituted benzoic acids calculated at
the level of B3LYP/6-31G theory with and without incorporation
of one water molecule are shown in Figures 4-6. The starting
structures are anti carboxylic conformers in all cases, but the
anti carboxylic conformers for 2,6-dihydroxybenzoic and 2,4,6-
trihydroxybenzoic acids are nonplanar at the level of B3LYP/
The experimental data were measured in aqueous solution
where a complex solvent field is present. In contrast to the gas-
phase DFT model, the experimental Ea values are relatively
similar, whereas the differences in activation entropy appear to
play the major role in the rate differences (Table 2). It is difficult
to identify a specific reason for why the activation entropy is
so important, but the most obvious factor appears to be solvation
differences of the reactants and transition states. The calculated
dipole moments for benzoic acid and the OH substituted benzoic
acids are in the order: anti carboxylic conformer reactant <
transition state structure with one water molecule < transition
state structure without one water molecule. It seems that the
transition state structures are more solvated than the reactants
and so the high polarity of the water solvent would accelerate
the decarboxylation reaction. In contrast to the ground state of
the reactants, preferential solvation or desolvation of the
6
-31G(d). The calculated activation energies are very high in
the gas phase but less than those where the proton acceptor is
2
1
a â-C-C single bond or a âsCdC double bond. The
capability of benzene ring to act as a proton acceptor during
the decarboxylation process is comparable to a CtC triple bond
2
1
at the â position. In the transition state structures of benzoic
acid and OH substituted benzoic acids, the developing negative
charge on the ipso carbon atom was dispersed throughout the
benzene ring. OH substitution at the ortho or para positions
results in more effective charge dispersion and makes the C-OH

2672 J. Phys. Chem. A, Vol. 107, No. 15, 2003
Li and Brill
Figure 5. Transition state structures of dihydroxybenzoic acids with and without one water molecule (B3LYP/6-31G//B3LYP/6-31G). Bond distances
are given in angstroms.
Figure 6. Transition state structures of trihydroxybenzoic acids with and without one water molecule (B3LYP/6-31G//B3LYP/6-31G). Bond distances
are given in angstroms.
transition state could reduce or increase the activation energy.
An example where preferential solvation reduces the activation
energy is the decarboxylation of carbonic acid.38 An example
where the activation energy is increased by desolvation is the
a number of conformers of the acids and the variety of ways in
which H2O is able to interact with the benzene ring42 and
hydroxyl groups and (2) none of implicit solvation models or
the hybrid discrete-continuum model could exactly describe the
4
0,41
41
43
decarboxylation of benzisoxazole-3-carboxylic acid.
Gao
effect of solvation at present time.
reported that the primary contribution of desolvation comes from
the difference in the intrinsic charge distribution (i.e., dipole)
for the reactant and transition state. When there is no preferential
solvation or desolvation in the transition state, the calculated
activation energy agrees very well with the experimental result.
Another example of this occurrence is the decarboxylation of
undissociated acetylenedicarboxylic acid.21 Unfortunately, sol-
vation differences are probably complex and very difficult to
calculate for these hydroxybenzoic acids because (1) there are
It is interesting to note that the calculated activation energies
follow the trend of the experimental half-lives for decarboxyl-
ation of the OH substituted benzoic acids. Benzoic acid and
the OH substituted benzoic acids having the highest activation
energies did not decarboxylate in the present study. The 2,4,6-
trihydroxybenzoic acid has the lowest activation energy and
decarboxylates easily. It can be seen that the OH substituent
effect on the decarboxylation rate is more precisely reflected
in the calculated activation energies in the gas phase than in

Spectroscopy of Hydrothermal Reactions 23
J. Phys. Chem. A, Vol. 107, No. 15, 2003 2673
the experimental results. By comparing the experimental and
calculated activation parameters, there is evidence that the
decarboxylation rate changes from control by the activation
energy in gas phase to control by the activation entropy in
aqueous solution.
(12) Shubert, W. M.; Gardner, J. D. J. Am. Chem. Soc. 1953, 75, 1401.
(
63
(
13) Willi, A. V.; Cho, M. H.; Won, C. M. HelV. Chim. Acta 1970, 53,
6
14) Segura, P.; Bunnett, J. F.; Villanova, L. J. Org. Chem. 1985, 50,
1041.
(
(
(
(
(
(
15) Willi, A. V.; Cho, M. H.; Chen, W. Z. Phys. Chem. 1974, 91, 193.
16) Willi, A. V. Z. Naturfrorsch. 1958, 13A, 997.
17) Lynn, K. R.; Bourns, A. N. Chem. Ind. 1963, 19, 782.
18) Ruelle, P. J. Comput. Chem. 1987, 8, 158.
19) Ruelle, P. J. Chem. Soc., Perkin Trans. 2 1986, 12, 1953.
20) Glukhovtsev, M. N.; Bach, R. D.; Laiter, S. J. Org. Chem. 1997,
Conclusions
The experimental decarboxylation rates of hydroxybenzoic
acids in solution essentially follow the trend anticipated for
conventional aromatic ring resonance effects. The Arrhenius
parameters suggest however that the activation entropy as
manifested in the preexponential factor is more important in
determining the relative rates than is the activation energy.
Density functional theory supports that at least one H2O
molecule acts as a bridge and participates in the proton-transfer
step of the decarboxylation reaction and that the trend in the
calculated activation energies qualitatively matches the experi-
mentally found half-lives. However, an apparently more im-
portant role in the experimental data is played by the activation
entropy. It might be hypothesized therefore that differences in
the solvation shell of the reactant and the transition state
structures are responsible.
6
1
B. J. Phys. Chem. 1996, 100, 14343.
(24) Maiella, P. G.; Schoppelrei, J. W.; Brill, T. B. Appl. Spectroscopy
1999, 53, 351.
25) Cvetanovic, R. J.; Singleton, D. L. Int. J. Chem. Kinetics 1977, 9,
81.
26) La Marca, C.; Libanati, C.; Klein, M. T.; Cotter, R. J.; Andrews,
S. M. Preprints of Papers, ACS Fuel DiV., 1991, 36, 676.
27) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
4
M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.;
Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A.
D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi,
M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.;
Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick,
D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.;
Ortiz, J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi,
I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.;
Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M.
W.; Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon,
M.; Replogle, E. S.; Pople, J. A. Gaussian 98, revision A.9; Gaussian,
Inc.: Pittsburgh, PA, 1998.
Acknowledgment. We are grateful to the National Science
Foundation for support of this work on Grant CHE-9807370.
Supporting Information Available: The Z matrixes for the
optimized geometries of all benzoic acid derivatives studied in
the work using B3LYP/6-31G (Gaussian 98) are compiled in
Table S (41 pages). This material is available free of charge
via the Internet at http://pubs.acs.org.
Wiley & Sons: New York, 1998; p 368.
2
1
95.
291.
References and Notes
1) Siskin, M.; Katritzky, A. R. Chem. ReV. 2001, 101, 825.
in Aqueous Solutions; Pergamon: Oxford, 1979.
(34) Lindsay, W. T. Proc. Int. Water Conf. Eng. Soc. W. Pa. 1980, 41,
284.
(35) Nagy, P. I.; Dunn, W. J. III.; Alagona, G.; Ghio, C. J. Phys. Chem.
1993, 97, 4628.
W. J. Org. Chem. 1997, 62, 2505.
5) Katrizky, A. R.; Allin, S. A.; Siskin, M. Acc. Chem. Res. 1996,
9, 399.
6) Li, C. J.; Chan, T. H. Organic Reactions in Aqueous Media;
Wiley: New York, 1997.
(36) Nagy, P. I.; Smith, D. A.; Alagona, G.; Ghio, C. J. Phys. Chem.
1994, 98, 486.
(37) Shapley, W. A.; Bacskay, G. B.; Warr, G. G. J. Phys. Chem. B.
1998, 102, 1938.
(38) Tautermann, C.; Voegele, A. F.; Loeting, T.; Kohl, I.; Hallbrucker,
A.; Mayer, E.; Liedl, K. Chem. Eur. J. 2002, 8, 66.
(39) Akiya, N.; Savage, P. E. AIChE J. 1998, 44, 405.
(40) Zipe, H.; Apaydin, G.; Houk, K. N. J. Am. Chem. Soc. 1995, 117,
8606.
(
2
(
(
7) Akiya, N.; Savage, P. E. Chem. ReV. 2002, 102, 2725.
(8) Katritzky, A. R.; Balasubramanian, M.; Siskin, M. Energy Fuel
1
990, 4, 499.
(9) Manion, J. A.; McMillen, D. F.; Malhotra, R. Energy Fuel 1996,
1
0, 776.
(
(41) Gao, J. J. Am. Chem. Soc. 1995, 117, 8600.
(42) Fredricks, S. Y.; Jordan, K. D.; Zwier, T. S. J. Phys. Chem. 1996,
100, 7810.
(43) Sicinska, D.; Paneth, P.; Truhlar, D. G. J. Phys. Chem. B 2002,
106, 2708.
10) McCollom, T. M.; Seewald, J. S.; Simoneit, B. R. T. Geochim.
Cosmochim. Acta 2001, 65, 455.
11) Boles, J. S.; Crerar, D. A.; Grissom, G.; Key, T. C. Geochim.
Cosmochim. Acta 1988, 52, 341.
(