Spectroscopy of hydrothermal reactions 20: Experimental and DFT computational comparison of decarboxylation of dicarboxylic acids connected by single, double, and triple bonds

Article abstract of DOI:10.1021/jp020528l

The kinetics and pathways of decarboxylation of aqueous acetylenedicarboxylic acid at pH = 0.97 - 8.02 were studied in situ at 80-160°C and 275 bar by using an FT-IR spectroscopy flow reactor with sapphire windows. The first-order (or pseudo first-order) rate constants and corresponding Arrhenius parameters were obtained for the neutral acid, monoanion, and dianion. The decarboxylation rates are in the order: HO₂CC≡CCO₂^- > HO₂CC≡CCO₂H > ^-O₂CC≡CCO₂^-. The decarboxylation mechanisms of these reactants and the propiolic acid product were analyzed by B3LYP/6-31+G(d) density functional theory. The transition state structures were found for the neutral acids and monoanions. In gas phase the transition state structure is a four-member ring involving C-C(O)-O-H. In aqueous solution a cyclic structure incorporating at least one water molecule forms. A comparison of transition state structures for the decarboxylation of β-saturated (succinic) and β-unsaturated (maleic, fumaric, and acetylenedicarboxylic) aliphatic diacids was made with and without incorporating a water molecule. Consistent with experiment, the calculated activation energy for H-atom transfer to the α carbon atom in the decarboxylation step follows the order C≡C < C=C < C-C.

Full text of DOI:10.1021/jp020528l

J. Phys. Chem. A 2002, 106, 9491-9498
9491
Spectroscopy of Hydrothermal Reactions 20: Experimental and DFT Computational
Comparison of Decarboxylation of Dicarboxylic Acids Connected by Single, Double, and
Triple Bonds
Jun Li and Thomas B. Brill*
Department of Chemistry and Biochemistry, UniVersity of Delaware, Newark, Delaware 19716
ReceiVed: February 26, 2002; In Final Form: July 3, 2002
The kinetics and pathways of decarboxylation of aqueous acetylenedicarboxylic acid at pH ) 0.97-8.02
were studied in situ at 80-160 °C and 275 bar by using an FT-IR spectroscopy flow reactor with sapphire
windows. The first-order (or pseudo first-order) rate constants and corresponding Arrhenius parameters were
obtained for the neutral acid, monoanion, and dianion. The decarboxylation rates are in the order:
-
-
-
HO
2
CCtCCO
2
> HO
2
CCtCCO
2
2 2
H > O CCtCCO . The decarboxylation mechanisms of these reactants
and the propiolic acid product were analyzed by B3LYP/6-31+G(d) density functional theory. The transition
state structures were found for the neutral acids and monoanions. In gas phase the transition state structure
is a four-member ring involving C-C(O)-O-H. In aqueous solution a cyclic structure incorporating at least
one water molecule forms. A comparison of transition state structures for the decarboxylation of â-saturated
(
succinic) and â-unsaturated (maleic, fumaric, and acetylenedicarboxylic) aliphatic diacids was made with
and without incorporating a water molecule. Consistent with experiment, the calculated activation energy for
H-atom transfer to the R carbon atom in the decarboxylation step follows the order CtC < CdC < C-C.
Introduction
acids containing CdC and CtC bonds in the backbone. For
example, maleic acid and fumaric acid are difficult to decar-
Decarboxylation is an important practical reaction of car-
boxylic acids in organic chemistry, biochemistry and geochem-
istry. Decarboxylation can proceed via heterolytic cleavage or
homolytic cleavage of the C-C bond, although heterolytic
boxylate at hydrothermal conditions. They decompose instead
to simple acids such as formic, glyoxylic, and oxalic acids.20
Decarboxylation occurs with oxidative conditions, such as in
2
0
21
wet air oxidation and electroxidation. Another example of
complex decarboxylation occurs with itaconic, citraconic, and
1
cleavage is more common at hydrothermal conditions. Het-
erolysis with the loss of CO2 appears to occur by a variety of
mechanisms which depend on the structure of the parent acid
and the experimental conditions. The transition state structures
for decarboxylation include: (1) nucleophilic bimolecular attack
5
,22
mesaconic acids.
Decarboxylation of citraconic acid and
mesaconic acid occurs via the decarboxylation of itaconic acid,
i.e., citraconic acid and mesaconic acid initially isomerize to
itaconic acid, which then loses CO2. Thus, the position of the
CdC bond in the backbone influences the reaction rate and
mechanism.
1
on the carbon atom of carboxylate group; (2) formation of a
cyclic structure involving hydrogen-bonding between H of the
carboxylate group and an electronegative atom in the â position
2
,3
4,5
Among the saturated and unsaturated aliphatic diacids and
monoacids, acetylenedicarboxylic and propiolic acids, which
possess a CtC bond in the backbone, were found to decar-
boxylate the fastest at hydrothermal conditions. The reason for
this fact is addressed in this paper. In previous work Tommila
(
e.g., oxygen or carbon ); (3) proton shift in an R, â-unsatur-
6
ated acid to form a â,γ-unsaturated acid followed by (2); and
4) proton shift to form a zwitterionic structure (e.g., 2-amino-
formylacetic acids, orotic acid, or 4-pyridylacetic acid ). Low
molecular mass carboxylic acids, such as formic and acetic acid,
tend to have higher hydrothermal stability and require more
(
2,7
8
9
23
and Kivinen reported that acetylenedicarboxylic acid decom-
posed about 30% faster than its monoanion, but did not discuss
1
0
reactive conditions for decomposition, such as wet oxidation,
2
4
11
the dianion. Hsu and Huang determined the rate constant and
Arrhenius parameters for decarboxylation of the neutral acid in
acetophenone, and reported an unusually small preexponential
factor.
supercritical water oxidation, and heterogeneous surface
catalysis.1
1-14
15
A study of the structure-reactivity of acetic acid derivatives
revealed that both steric and electronic effects play a role in
the decarboxylation rate, but the steric effect is more important
The decarboxylation of carboxylic acids and derivatives
containing the CtC bond are well-suited for study with an FT-
IR spectroscopy flow reactor using sapphire windows because
both the CtC bond and CO2 intensely absorb in the available
band-pass. The mechanisms of decarboxylation of acetylene-
dicarboxylic acid, its monoanion, and propiolic acid were
analyzed by density functional theory. The transition state
structures with and without water catalysis were found. The
comparison of these transition state structures with those of other
monoacids (formic, acetic, acrylic, and 3-butenoic) and diacids
(oxalic, malonic, maleic, fumaric, and succinic) was made to
2,3
in most cases. For example, the cyclic transition state structure
as described by mechanism (2) above significantly reduces the
energy barrier toward decarboxylation. The R-hydroxy acids,
which lack this structure, have condition-dependent decomposi-
tion pathways1
6-19
involving competitive decarboxylation,
dehydration, and decarbonylation.
Another class of carboxylic acids whose decarboxylation
mechanisms are complicated are the unsaturated dicarboxylic
*
Corresponding author. E-mail: brill@Udel.edu.
1
0.1021/jp020528l CCC: $22.00 © 2002 American Chemical Society
Published on Web 09/14/2002

9492 J. Phys. Chem. A, Vol. 106, No. 41, 2002
Li and Brill
TABLE 1: Rate Constants and Arrhenius Parameters for
0
Decarboxylation of Acetylenedicarboxylic Acid (k ),
Acetylenedicarboxylate Monoanion (k
acetylenedicarboxylate dianion (k
1
), and
2
)
3
-1
3
-1
3
-1
T/°C
k
0
× 10 /s
k
1
× 10 /s
k
2
× 10 /s
8
9
1
1
1
1
1
1
1
0
1.56 ( 0.62
5.86 ( 2.47
14.5 ( 4.6
38.1 ( 12.6
92.0 ( 26.0
3.86 ( 0.51
0
10.6 ( 2.5
29.7 ( 4.4
78.8 ( 13.4
207 ( 32.3
00
10
20
30
40
50
60
1.55 ( 0.11
3.93 ( 0.22
10.8 ( 0.3
21.7 ( 0.8
37.7 ( 1.8
3
-1
3
-1
3
-1
k
0
× 10 /s
k
1
× 10 /s
k
2
× 10 /s
/kJ mol-
ln(A, s-1)
S‡/J K mol-1
1
115 ( 3.2
115 ( 1.2
(125.52 )
111 ( 8.6
27.6 ( 2.5
-24.8
E
a
a
a
(
(
126.36 )
b
99.16 )
32.0 ( 1.1
33.5 ( 0.4
a
a
(
(
35.63 )
(35.07 )
b
9.90 )
∆
c
18.7
22.6
a
Reference 23. b From ref 24 in acetophenone. c At 100 °C.
understand why the presence of the CtC bond has such an
acceleratory effect on the decarboxylation rate.
Figure 1. FT-IR spectra of 0.25 m monopotassium acetylenedicar-
boxylate at 275 bar and a residence time of 47 s as a function of
temperature.
Experimental Section
Acetylenedicarboxylic acid (HO2CCtCCO2H, 95%), mono-
potassium acetylenedicarboxylate (KO2CCtCCO2H, 98%),
maleic acid (cis-HO2CCHdCHCO2H, 99%), fumaric acid
CO2. The molar absorptivities of the 2113 cm-1 band and 2090
27
-1
cm band were determined to be 4.5 times and 9 times smaller
than that of CO2, respectively. The concentrations of propiolic
acid and propiolate anion were obtained by fitting the band area
(trans- HO2CCHdCHCO2H, 99%), succinic acid (HO2CCH2-
CH2CO2H, 99%), methacrylic acid (CH3CHdCHCO2H, 99%),
and KOH were purchased from Aldrich Chemical Co. and used
without further purification. Milli-Q deionized water was
sparged with compressed Ar before use to expel the atmospheric
gases. The solution of dipotassium acetylenedicarboxylate was
prepared by titrating an acetylenedicarboxylic acid solution with
KOH to the equivalent point. The other solutions with different
pH values (Table 2) were prepared by dissolving the diacid and
monopotassium salt in the required ratios. For example, the
solution with pH ) 2.17 was prepared by mixing KO2CCt
CCO2H and HO2CCtCCO2H in the ratio of 4:1 at 25 °C. The
pH values of the solutions were measured at room temperature
with an Orion model 330 pH meter. The concentrations of all
solutions were 0.25 molal.
-
1
of 2113 and 2090 cm with a four-parameter Voigt function.
2
8
Weighted least-squares regression with a 95% confidence
interval was performed for both the rate constants and the
Arrhenius parameters in which the statistical weight was set to
2
be 1/σ , where σ is the standard deviation of the variables.
Succinic, maleic, fumaric, and methacrylic acids did not
decompose even when the flow reactor was run at its limit (330
°
C and 275 bar) with a residence time of 60 s.
The density functional theory calculations on the transition
state structures for decarboxylation were performed using
29
30,31
Gaussian 98 software at the level of theory of B3LYP
with
the basis set of 6-31G and 6-31+G(d). The geometries of
reactants and transition state structures were optimized and
vibrational frequency analyses were conducted to confirm that
the optimized geometry was a local minimum or a transition
state. The frequency analyses also provide thermal energy
corrections to the total energy. All calculations were made at
The flow reactor-IR spectroscopy cell constructed from
titanium with sapphire windows and gold foil seals has been
described in detail elsewhere.
25,26
The temperature and pressure
were controlled within ( 1 °C and ( 1 bar, respectively. The
chosen flow rate in the 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. Transmission IR spectra were recorded
2
98.15 K and 1 bar. The effect of temperature and pressure
changes on the calculation was found to be negligible relative
to the uncertainty in the computational method. In any event
the comparisons made herein are best viewed as relative rather
than absolute.
-
1
at 4 cm resolution with a Nicolet 560 Magna FTIR spec-
trometer and an MCT-A detector. Background spectra recorded
on 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, the asymmetric stretches
Kinetics
Reaction Pathways. Acetylenedicarboxylic acid decarboxy-
lates to produce CO2 and propiolic acid. Simultaneously,
propiolic acid decarboxylates to form CO2 and acetylene. In
hydrothermal solutions, these reaction pathways were clearly
indicated in situ using FT-IR spectroscopy and confirmed by
comparing the spectra with those of the pure compounds. Real-
time FT-IR spectra during the decarboxylation of 0.25 m
monopotassium acetylenedicarboxylate are shown in Figure 1
-1
of aqueous CO2 centered at 2343 cm and the CtC triple bond
-1
-1
at 2113 cm for propiolic acid and 2090 cm for the propiolate
anion were observed in the band-pass of sapphire. 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

Decarboxylation of Dicarboxylic Acids
J. Phys. Chem. A, Vol. 106, No. 41, 2002 9493
SCHEME 1
the hydrolysis of CO2 in basic solution. Equation 5 was
parametrized by the use of eqs 6 and 7.
3
7
C ) [CO ] + [HCO ] + [CO 2-]
-
(4)
(5)
C
2 obs
3
3
Ka1
K K
a1 a2
)
[CO2]obs 1 +
+
(
+
+ 2
)
[
H ] [H ]
Ka1
-
+
CO + H O w x HCO + H
(6)
2
2
3
Ka2
HCO3- w x CO3 + H
2-
+
(7)
SCHEME 2
-
-
This procedure enables the rate constant k2 for O2CCtCCO2
to be obtained by fitting eq 2. The calculated values of the rate
constant k2 are given in Table 1 at pH ) 8.02.
The determination of the rate constants for HO2CCtCCO2H
-
-
-
and HO2CCtCCO2 is different from that of O2CCtCCO2 .
3
8
From experiment, propiolic acid and propiolate anion were
observed to require higher temperatures to obtain the same
decarboxylation rate as the acetylenedicarboxylic acid species.
Hence, at lower temperatures in the early stage of decarboxy-
lation, the CO2 from the propiolic acid species is very small
and can be ignored. The rate of formation of CO2 follows eq 8
assuming that the first-order rate law applies to the decomposi-
tion of acetylenedicarboxylic acid species.
at different temperatures with a pressure of 275 bar and a
residence time of about 47 s. The existence of aqueous phase
-
1
-1
CO2 (2343 cm ), propiolic acid (2113 cm ), and propiolate
-
1
anion (2090 cm ) is clearly indicated. Two additional weak
peaks were not identified.
Scheme 1 summarizes the reaction pathway in which K1, K2,
and Kp are the first and second dissociation constants of
acetylenedicarboxylic acid and the dissociation constant of
propiolic acid, respectively; k0, k1, and k2 are the rate constants
of the neutral diacid, monoanion, and dianion of acetylenedi-
carboxylic acid; and k3 and k4 are the rate constants of propiolic
acid and the propiolate anion, respectively. Acetylenedicar-
boxylic acid (pK1 ) 1.23 and pK2 ) 2.53) and propiolic acid
pKp ) 1.89) are rather strong acids which are mostly dissociated
into the corresponding anions. Therefore, before and during
decarboxylation, the dissociation equilibration of acetylenedi-
carboxylic acid and propiolic acid must be taken into account,
and CO2 potentially forms from all of the acetylenedicarboxylic
acid and propiolic acid species.
V ) kobs[HO CCtCCO H]
T,t
(8)
2
2
The observed first-order rate constant is kobs, and the total
acetylenedicarboxylic acid concentration as a function of time
is given by eq 9.
[
HO CCtCCO H] ) [HO CCtCCO H] +
2 2 T,t 2 2 t
-
-
-
[
HO CCtCCO ] + [ O CCtCCO ] )
2 2 t 2 2 t
[
HO CCtCCO H] -[CO ] (9)
2 2 0 2 T,t
[HO2CCtCCO2H]0 is the initial concentration of acetylenedi-
Reaction Rates. The dissociation constants needed for
calculation of the species distribution at high temperatures were
carboxylic acid. The calculation of the total concentration of
CO2 (eq 5) requires that the solution pH be known at each
temperature and residence time during decarboxylation. At low
32
extrapolated from room temperature by using the iso-Cou-
3
3
34
lombic method employing the specific volume and ionization
-
2-
pH, the hydrolysis of CO2 to HCO3 and CO3 is entirely
negligible. In fact, except for KO2CCtCCO2K at pH ) 8.02,
all of the experiments were conducted in this pH range starting
from a solution pH of 0.97 (the natural pH of acetylenedicar-
boxylic acid) to the upper limit of 2.17. That the observed rate
constants were first-order was confirmed as illustrated by the
rate plot in Figure 2 for decarboxylation of 0.25 m KO2CCt
CCO2H. The observed first-order rate constants at different pH
values in the range of 0.97-2.17 are listed in Table 2.
35
constants of water. The prepared dipotassium acetylenedicar-
boxylate solution contains only 0.05% of the monoanion and
0.0004% of the diacid at 100 °C. It is difficult to obtain analytic
rate expressions for all species because of the multiple reaction
pathways shown in Scheme 1, but the analytic rate expressions
can be derived when the acetylenedicarboxylate dianion pre-
dominates and Scheme 2 applies. The rate expressions 1-3 are
appropriate when calculating k2, where CA0 is the initial
-
-
concentration of O2CCtCCO2 , CB is the concentration of
When the decarboxylation of
HO2CCtCCO2H,
HO2CCtCCO2 , and O2CCtCCO2 follow first-order, eqs
0 and 11 apply.
-
HCtCCO2 , CC is the total concentration of CO2 in aqueous
-
-
-
solution, i.e., eq 4, and k4 is our previously determined rate
1
36
constant for the propiolate anion. CC,
-
-
k2t
V ) k [HO CCtCCO H] + k [HO CCtCCO ] +
0
2
2
t
1
2
2
t
C ) C e
(1)
(2)
A
A0
-
-
k [ O CCtCCO ] (10)
2
2
2
t
A0
C k2
(e-k2t - e-k4t)
C )
+
2
+
B
k [H ] + k [H ]K + k K K
2 1 2
k4 - k2
0
1
1
kobs
)
(11)
+
2
+
[
H ] + [H ]K + K K
1 1 2
CA0
-k2t
- k2e-k4t] (3)
C ) 2C
-
[(2k - k )e
4 2
C
A0
k4 - k2
Fitting of the nonlinear eq 11 with the observed rate constants,
ionization constants K1 and K2, the rate constant k2, and solution
pH at high-temperature yields rate constant k0 for HO2CCt
which is given by eq 4, was calculated by eq 5, which describes

9494 J. Phys. Chem. A, Vol. 106, No. 41, 2002
Li and Brill
TABLE 2: Observed First-Order Rate Constants (kobs × 10 /s-1) for Decarboxylation of 0.25 m Acetylenedicarboxylic Acid at
3
Different pH Values
pH25
T/°C
0.97
1.36
1.52
1.66
1.91
2.17
8
9
1
1
1
0
0
00
10
20
1.43 ( 0.022
4.36 ( 0.04
14.2 ( 0.2
36.9 ( 0.5
97.6 ( 2.7
2.90 ( 0.05
9.38 ( 0.14
18.9 ( 0.3
2.53 ( 0.04
7.28 ( 0.08
23.1 ( 0.4
2.60 ( 0.05
6.84 ( 0.06
23.2 ( 0.3
2.70 ( 0.02
6.19 ( 0.03
16.3 ( 0.1
1.83 ( 0.02
5.24 ( 0.05
13.9 ( 0.4
33.0 ( 0.8
61.6 ( 1.2
47.6 ( 0.4
61.5 ( 1.0
59.4 ( 0.7
44.1 ( 0.2
110.1 ( 1.1
153.7 ( 2.8
143.9 ( 3.5
123.4 ( 2.0
Figure 3. Arrhenius plot for the decarboxylation of neutral acetylene-
diacrboxylic acid, acetylenedicarboxylate monoanion, and acetylene-
dicarboxylate dianion.
Figure 2. Rate plot for the decarboxylation of 0.25 m monopotassium
acetylenedicarboxylate at different temperatures and 275 bar.
phenylmalonic acid41 when the rate constants for the dianions
are inferred from those of the acid as a function of the pH.
CCO2H and k1 for HO2CCtCCO2H. The solution pH at high
temperatures can be calculated fortunately from the concentra-
tion ratio of propiolic acid and propiolate anion (eq 12).
Mechanisms by Density Functional Theory
+
-
[
H ] ) K [HCtCCO H]/[HCtCCO ]
(12)
Geometries and Energetics. The optimized geometries and
p
2
2
energetics of the neutral acids and their anions are shown in
-
The pH values obtained are a little higher than the initial solution
pH values calculated from the charge balance equation at high
temperatures. This is not surprising because the reactant
acetylenedicarboxylic acid and the product propiolic acid are
relatively strong acids and the decomposition yields are not high.
In fact, the pH at the mean value of the different residence times
was used to represent the real solution pH at a given temperature.
The obtained rate constants k0 and k1 and Arrhenius parameters
are given in Table 1. The Arrhenius plots are displayed in Figure
Figure 4 (ACDC represents HO2CCtCCO2H, ACDC is
-
HO2CCtCCO2 , and Prop is HCtCCO2H). From Figure 4 it
can be seen that the conformers in which the hydrogen atom is
anti to the carbonyl oxygen (hereinafter the anti carboxylic
hydrogen) are higher in energy than the syn conformers. Nagy42
also observed this pattern. The C-C single bond lengths
adjacent to the CtC triple bond are about 0.07 Å shorter than
that of a normal C-C bond (ca. 1.51 Å) as a result of resonance
of the lone pair atomic orbitals on the oxygen atoms with the
CtC bond. This significant delocalization of excess negative
charge for acetylenedicarboxylate anion was evident to Skurski
3
. Clearly, the decarboxylation rates of acetylenedicarboxylic
-
acid species are in the order: HO2CCtCCO2 > HO2CCt
-
-
43
CCO2H > O2CCtCCO2 . For comparison, the previously
reported activation energies and preexponential factors of neutral
acid and monoanion in aqueous solution, as well as that of
neutral acid in acetophenone are also listed in Table 1. The
et al., when the acetylenedicarboxylate and succinate dianions
were compared, and was proposed to be the reason for why
succinate anion was not easily observed by photoelectron
spectroscopy. Another finding in the calculations of Figure 4
is that the carbon chain backbone increases in length by 0.04
Å and 0.07 Å, respectively, when the neutral acid dissociates
into the monoanion and the dianion. Dissociation produces
negative charge on the terminal carboxylate groups, and the
resulting electrostatic repulsion stretches the carbon chain and
forces the dihedral angle of the two carboxylate groups to
approach 90°.
2
1
Arrhenius parameters determined by Tommila and Kivinen
are close to those of this work, but the preexponential factor
for the neutral acid in acetophenone reported by Hsu and
Huang is probably too small even when the effect of the
solvent is taken into account. The decarboxylation rate of the
dianion is the slowest, which is also the case with malonic acid.
2
2
39
4
0
The same order of rates is expected for oxalic acid and

Decarboxylation of Dicarboxylic Acids
J. Phys. Chem. A, Vol. 106, No. 41, 2002 9495
Figure 4. Geometries and relative energetics of acetylenedicarboxylic acid and propiolic acid species at the level of B3LYP/6-31+G(d). The bond
lengths are given in angstroms.
Transition State Structures. The global decarboxylation
reaction of aliphatic acids is given by eq 13:
energy barriers increased because of the additional energy of
the hydrogen bonds between reactant and the water molecule.
We compared the energy barrier of the transition state with water
present but without participation of water in the proton transfer
step. Four water molecules were positioned to interact with the
external regions of the carboxylate groups. The results are shown
in Figure 6. There was no difference in energy barrier for the
intramolecular proton transfer process when 1, 2, 3, or 4 water
molecules surrounded the reactant in this manner. A similar
result has recently been obtained for carbonic acid. The effect
of more extensive solvation on the energy barrier was not
considered because there are many possible orientations of water
molecules when surrounding the reactant. However, a conclusion
that may be drawn from these calculations, as well as those for
carbonic acid, is that small changes in solvent density do not
affect the transition state.
R-CO H f RH + CO2
(13)
2
The proton-transfer step is a key element of the decarboxylation
process. Ab initio and DFT calculations have been carried out
for the decarboxylation process of many acids including
4
4-48
49-52
4
53
54,55
formic,
acetic,
3-butenoic, acrylic, oxalic,
ma-
56
lonic,2 and itaconic. Among the transition state structures
,3
5
obtained, the six-member ring in malonic acid (involving
2
,3
4
intramolecular hydrogen bonding), 3-butenoic acid, and
5
itaconic acid leads to proton transfer as a low energy barrier
process. The similar transition state consisting of a six-member
ring structure is impossible for acetylenedicarboxylic and
propiolic acid, because the CtC bond forces linearity. Hence
the transition state structure in which the carboxylate hydrogen
atom starts in the anti orientation and forms a four-member ring
structure is the only reasonable structure, but has a high energy
barrier because of the strain energy. However, when one water
molecule participates to form a six-member ring structure, the
The difference in the calculated and experimental activation
-
energy for HO2CCtCCO2 may arise from at least three
factors: The choice of the reference state; the fact that no
countercation was incorporated; and the fact that no solvation
shell was included in the calculation. Also, the experimental
value of Ea for HCtCCO2H may have been affected by slight
energy barrier is significantly reduced. The involvement of two
water molecules reduces the energy even more,4
4,48,49,53,56
but
38
curvature in the rate plots making the value somewhat
the incremental difference is smaller than that produced by the
first water molecule. The resulting transition states with and
without participation of one water molecule are shown in Figure
. The energy barrier for decarboxylation of HO2CCtCCO2H
is 118.7 kJ/mol when one water molecule acts as a catalyst,
uncertain.
Another finding is that the dihedral angle of the two
carboxylate groups is close to 0° when one water molecule is
incorporated into neutral acetylenedicarboxylic acid and its
monoanion (see Figure 5), whereas the angle is 90° in the
unsolvated structure. However, this angle is retained at about
90° when solvation of the solute by up to four water molecules
occurs and two water molecules coordinate each carboxylate
group to form a hydrogen bond net (Figure 6). This is an
extreme example showing that solvation differences can cause
a conformational change.
Why Do Acetylenedicarboxylic and Propiolic Acids De-
carboxylate So Easily? Acetylenedicarboxylic and propiolic
acids decarboxylate readily at 80-160 °C. On the other hand,
there is no evidence of decarboxylation of succinic, maleic,
fumaric, and methacrylic acids when the flow reactor was run
5
which is in excellent agreement with the experimental value of
-
1
15.5 kJ/mol in Table 1. For HO2CCtCCO2 , the calculated
value of 66.6 kJ/mol is about half the experimental value of
14.6 kJ/mol, and the value for HCtCCO2H of 124.7 kJ/mol
1
38
is larger than the experimental value of 88 kJ/mol. The origin
of the differences for the latter two compounds is discussed
further below.
It should be pointed out that the reference state for the
calculation of the energy barrier is that in which the reactant
molecule and water molecule are separated at infinite distance.
When solvation by one water molecule was considered, the

9496 J. Phys. Chem. A, Vol. 106, No. 41, 2002
Li and Brill
Figure 5. Transition state structures leading to decarboxylation with and without the participation of one water molecule and the starting structures
-
2 2 2 2 , 2
CCtCCO H, HO CCtCCO and HCtCCO H at the level of B3LYP/6-31+G(d). The values in parentheses are the activation energies
for HO
calculated starting from the solvated structures.
2 2 2
H and HO CCtCCO
-.
Figure 6. The effect of solvation of the carboxylate groups on the energy barrier of decarboxylation for HO
2
CCtCCO
at its limit (330 °C and 275 bar) for 60 s. As discussed above,
the six-member cyclic structure with the carboxylate hydrogen
atom transferring to the â-position oxygen atom2 is energeti-
cally the lowest pathway for malonic acid. When such a structure
5
7,58
was applied to maleic
acid and itaconic acid (dCH2 in the
,3
59
â-position), decarboxylation did not occur. Instead, the transi-

Decarboxylation of Dicarboxylic Acids
J. Phys. Chem. A, Vol. 106, No. 41, 2002 9497
Conclusions
The decarboxylation rates of acetylenedicarboxylic species
-
follow the order: HO2CCtCCO2 > HO2CCtCCO2H >
-
-
O2CCtCCO2 . The experimentally determined activation
energies for these three species are approximately 113 kJ/mol.
The transition state structures were found for the reactant acid
species (neutral acid and monoanion) and product acid species
(propiolic acid) by using density function theory at the level of
B3LYP/6-31+G(d). In gas phase, the transition state structure
is a four-member ring involving C-C(O)-OH with the proton
transferring from the carboxylate group to the R-carbon. In
aqueous solution, a cyclic structure incorporating at least one
water molecule forms. The difference in the calculated activation
energies for HO2CCtCCO2H and HCtCCO2H is consistent
with their relative hydrothermal reactivity based on the experi-
mental data.
A comparison of the calculated activation energy for the
decarboxylation of â-saturated and â-unsaturated aliphatic
diacids revealed that the order is C-C > CdC > CtC.
Incorporation one water molecule in the transition state structure
reduced the energy barrier to about half that without the water
molecule, but did not change the ordering. The same order for
decarboxylation rate is found in the experimental data.
Figure 7. Transition state structures at the level of B3LYP/6-31G for
decarboxylation with and without the participation of one water
molecule and the starting structures of maleic, fumaric, and succinic
acids.
tion state structures found corresponded to that for a low barrier
Acknowledgment. We are grateful to the National Science
Foundation for support of this work on Grant CHE-9807370.
60,61
hydrogen bond,
in which the distance between two hydrogen
atom acceptors is short enough to reduce the energy barrier for
H transfer.
Alternatively, proton transfer leading to the release of CO2
can occur via a four-member ring structure starting from the
anti carboxylic hydrogen conformer. The energy barriers for
References and Notes
(1) Clark, C. L. In The Chemistry of Carboxylic Acids and Esters. The
Chemistry of Functional Groups; Patai, S., Ed.; Wiley: New York, 1969;
p 589.
4
4-48
52
54
3
this process in formic,
acetic, oxalic, malonic, 3-oxo-
(
2) Bach, R. D.; Canapa, C. J. Org. Chem. 1996, 61, 6346.
(3) Huang, C. L.; Wu, C., C.; Lien, M. H. J. Phys. Chem. A 1997,
01, 7867.
4) Wang, Y. L.; Poirier, R. A. Can. J. Chem. 1994, 72, 1338.
5) Li, J.; Brill, T. B. J. Phys. Chem. A 2001, 105, 10839.
3
3
propanoic, and acetoacetic acids are in the vicinity of 290 kJ/
mol (except for 3-oxopropanoic acid which is 375.7 kJ/mol),
which are much higher than those for acetylenedicarboxylic and
propiolic acids. This is the case irrespective of the use of
different theories and basis sets (e.g., HF, MP2, MP4, B3LYP,
1
(
(
(6) Arnold, R. T.; Elmer, O. C.; Dodson, R. M. J. Am. Chem. Soc.
950, 72, 4350.
1
(7) Bach, R. D.; Canapa, C. J. Am. Chem. Soc. 1997, 119, 11725.
(8) Phillips, L. M.; Lee, J. K J. Am. Chem. Soc. 2001, 123, 12067.
(9) Sicinska, D.; Truhlar, D. G.; Paneth, P. J. Am. Chem. Soc. 2001,
6-31G, and 6-311G(d)) for these acids. For further clarification,
the transition state structures were calculated at the level of
B3LYP/6-31G for maleic (â-cis CdC), fumaric (â-trans Cd
C), and succinic (C-C) acids and are displayed in Figure 7.
123, 7683.
(
(
(
10) Shende, R. V.; Levec, J. Ind. Eng. Chem. Res. 1999, 38, 3830.
11) Yu, J. L.; Savage, P. E. Ind. Eng. Chem. Res. 1998, 37, 2.
12) Maiella, P. G.; Brill, T. B. J. Phys. Chem. A 1998, 102, 5886.
62
62
The conformational analyses of maleic, fumaric, and suc-
cinic63 acids have been performed previously resulting in five
conformations for maleic and fumaric acids, and fifteen
conformations for succinic acid. Two anti conformations not
(13) Palmer, D. A.; Drummond, S. E. Geochim. Cosmochim. Acta 1986,
50, 813.
14) Bell, J. L. S.; Palmer, D. A. Organic Acids in Geological Processes;
Springer-Verlag: Berlin, 1994; p 227.
15) Belsky, A. J.; Maiella, P. G.; Brill, T. B. J. Phys. Chem. A 1999,
103, 4253.
(
6
2
previously found were found in the present study. One of the
anti conformers was chosen here for each acid to conduct
detailed calculations. These calculations reveal the probable
reason for the decarboxylation rate differences. Compared to
the ease of the carboxylic hydrogen atom transferring to the
carbon atom of C-C bond, there is a reduction of 30 kJ/mol in
the energy barrier when this hydrogen atom transfers to the
carbon atom of CdC bond, and a reduction of 73 kJ/mol when
this hydrogen atom transfers to the carbon atom of CtC bond.
Therefore, if the hydrogen atom acceptor is a CtC bond, the
energy barrier to proton transfer is the smallest, with or without
participation of water molecules. The inclusion of one water
molecule in the transition state structure reduces energy barrier
to about half that without the water molecule. Succinic acid
with and without water molecules has the highest calculated
activation energy among the diacids. The energy barriers toward
decarboxylation of maleic and fumaric acids (Figure 7) are
(
(16) Domingo, L. R.; Andres, J.; Moliner, V.; Safont, V. S.J. Am. Chem.
Soc. 1997, 119, 6415.
(17) Li, L. X.; Portela, J. R.; Vallejo, D.; Gloyna, E. T. Ind. Eng. Chem.
Res. 1999, 38, 2599.
(18) Mok, W. S.-L.; Antal, M. J. J. Org. Chem. 1989, 54, 4596.
(19) Lira, C. T.; McCrackin, P. J. Ind. Eng. Chem. Res. 1993, 32, 2608.
(20) Shende, R. V.; Levec, J. Ind. Eng. Chem. Res. 2000, 39, 40.
(21) Cataldo F.; Progega, S. N. C. Croat. Chem. Acta 2000, 73, 435.
(22) Carlsson, M.; Habenicht, C.; Kam L. C.; Antal, M. J. Jr.; Bian, N.;
Cunningham, R. J.; Jones M. Jr. Ind. Eng. Chem. Res. 1994, 33, 1989.
(23) Tommila, E.; Kivinen, A. Suomen Kemistilehti 1951, B24, 46.
(24) Hsu, S. U.; Huang, T. T-S. Int. J. Chem. Kinet. 1974, 5, 567.
(25) Kieke, M. L.; Schoppelrei, J. W.; Brill, T. B. J. Phys. Chem. 1996,
100, 7455.
(26) Schoppelrei, J. W.; Kieke, M. L.; Wang, X.; Klein. M. T.; Brill, T.
B. J. Phys. Chem. 1996, 100, 14343.
(27) Maiella, P. G.; Schoppelrei, J. W.; Brill, T. B. Appl. Spectrosc.
1999, 53, 351.
(28) Cvetanovic, R. J.; Singleton, D. L. Int. J. Chem. Kinet. 1977, 9,
6
4
481.
higher than those needed for isomerization (66 kJ/mol),
(29) Frisch M. J.; Trucks G. W.; Schlegel H. B.; Scuseria G. E.; Robb
65
hydration (94 kJ/mol), or decomposition to simple acids (56.6
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.
kJ/mol for maleic and 71 kJ/mol for fumaric).2
0

9498 J. Phys. Chem. A, Vol. 106, No. 41, 2002
Li and Brill
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.; Chen W.; Wong
M. W.; Andres J. L.; Gonzalez C.; Head-Gordon M.; Replogle E. S.; Pople
J. A. Gaussian 98, version A.9; Gaussian, Inc.: Pittsburgh, PA, 1998.
(45) Ruelle, P. J. Am. Chem. Soc. 1987, 109, 1722.
(46) Goddard, J. D.; Yamaguchi, Y.; Schaefer, H. F. J. Chem. Phys.
1992, 96, 1158.
(
(
(
(
(
47) Francisco, J. S. J. Chem. Phys. 1992, 96, 1167.
48) Akiya, N.; Savage, P. E. AIChE J. 1998, 44, 405.
49) Ruelle, P. Chem. Phys. 1986, 110, 263.
50) Nguyen, M. T.; Ruelle, P. Chem. Phys. Lett. 1987, 138, 486.
51) Nguyen, M. T.; Sengupta, D.; Raspoet, G.; Vanquickenborne, L.
G J. Phys. Chem. 1995, 99, 11883.
(
(
(
30) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
31) Lee, C.; Yang, W., Parr, R. G. Phys. ReV. B 1988, 37, 785.
32) Serjeant, E. P.; Dempsey, B. Ionization Constants of Organic Acids
(
(
52) Duan X. F.; Page M. J. Am. Chem. Soc. 1995, 117, 5114.
53) Ruelle, P. J. Comput. Chem. 1987, 8, 158.
in Aqueous Solutions; Pergamon: Oxford, 1979.
33) Lindsay, W. T. Proc. Int. Water Conf. Eng. Soc. W. Pa. 1980, 41,
84.
(54) Bock, C. W.; Redington, R. L. J. Chem. Phys. 1986, 85, 5391.
(55) Higgins, J.; Zhou, X. F.; Liu, R. F.; Huang, T. T-S. J. Phys. Chem.
A 1997, 101, 2702.
(56) Tautermann, C. S.; Voegle, A. F.; Loerting, T.; Kohl, I.; Hall-
brucker, A.; Mayer, E.; Liedle, K. R. Chem. Eur. J. 2002, 8, 66.
(
2
1
2
(34) Uematsu, M.; Franck, E. U. J. Phys. Chem. Ref. Data 1980, 9,
(57) Bach, R. D.; Dmitrenko, O.; Glukhovtsev, M. N. J. Am. Chem.
95.
Soc. 2001, 123, 7134.
(36) Miksa, D.; Li, J.; Brill, T. B. J. Phys. Chem., in press.
37) Butler, J. N. Ion Equilibrium, Solubility and pH Calculation; John
(58) McAllister, M. A. Can. J. Chem. 1997, 75, 1195.
(
(59) Chen, J. G.; McAllister, M. A.; Lee, J. K.; Houk, K. N. J. Org.
Wiley & Sons: New York, 1998; p 368.
Chem. 1998, 63, 4611.
(
(
(
(
(
38) Li, J.; Brill, T. B. J. Phys. Chem. A 2001, 105, 6171.
(
(
11.
60) Frey, P. A.; Cleland, W. W. Bioorg. Chem. 1998, 26, 175.
61) Perrin, C. A.; Nielson, J. B. Annu. ReV. Phys. Chem. 1997, 48,
39) Gunawardena, N. R.; Brill, T. B. J. Phys. Chem. A 2001, 105, 1876.
40) Crossey, L. J. Geochim. Cosmochim. Acta 1991, 55, 1515.
41) Hall, G. A.; Hanrahan, E. S. J. Phys. Chem. 1965, 69, 2402.
42) Nagy, P. I.; Smith, D. A.; Alagona, G.; Ghio, C. J. Phys. Chem.
5
(
62) Macoas, E. M. S.; Fausto, R.; Lundell, J.; Pettersson, M.; Khri-
achtchev, L.; Masanen, M. J. Phys. Chem. A 2001, 105, 3922.
63) Price D. J.; Roberts, J. D.; Jorgenson, W. L. J. Am. Chem. Soc.
1
994, 98, 486.
43) Skurski, P.; Simons, J.; Wang X. B.; Wang, L. Sh. J. Am. Chem.
Soc. 2000, 122, 4499.
44) Ruelle, P.; Kesselring, U. K.; Nam-Tran, H. J. Am. Chem. Soc.
986, 108, 371.
(
(
1
998, 120, 9672.
(
(64) Hojendahl, K. J. Phys. Chem. 1924, 28, 758.
(65) Rozelle, L. T.; Alberty, R. A.J. Am. Chem. Soc. 1957, 61, 1637.
1