Spectroscopy of hydrothermal reactions 22. The effects of cations on the decarboxylation kinetics of trifluoroacetate, cyanoacetate, propiolate, and malonate ions

Article abstract of DOI:10.1021/jp020941t

The effect of Group 1 counterions on the decarboxylation rate of four carboxylate ions (trifluoroacetate, propiolate, cyanoacetate, and malonate) was determined in water at 140-280 °C and 275 bar. The decarboxylation kinetics were determined in real time by using IR spectroscopy and a titanium cell flow-reactor with sapphire windows. The complexity of the reaction of CF₃CO₂^- necessitated additional postreaction studies by ¹H and ¹⁹F NMR spectroscopy and F^- ion electrochemistry. Batch mode reactions in a titanium tube reactor were employed for the latter studies. The CF₃H product of CF₃CO₂^- decomposed, liberating F^- ions. The cation effect for CF₃CO₂^- is caused primarily by the reaction of F^- with the Group 1 cations in the order Li⁺ > Na⁺ > K⁺ ..., which changed the ionic strength. The rate of decaxboxylation of cyanoacetate was unaffected by the cation, whereas the Li⁺ ion slowed the rate of decarboxylation of propiolate, perhaps because of complexation with propiolate. These results were compared with the previously investigated malonate system. Where a cation effect was observed, the effect could be attributed to some form of ion pairing, but in each case the details of the interaction differ slightly.

Full text of DOI:10.1021/jp020941t

J. Phys. Chem. A 2002, 106, 11107-11114
11107
Spectroscopy of Hydrothermal Reactions 22. The Effects of Cations on the Decarboxylation
Kinetics of Trifluoroacetate, Cyanoacetate, Propiolate, and Malonate Ions
Davide Miksa, Jun Li, and Thomas B. Brill*
Department of Chemistry and Biochemistry, UniVersity of Delaware, Newark, Delaware 19716
ReceiVed: April 10, 2002; In Final Form: September 17, 2002
The effect of Group 1 counterions on the decarboxylation rate of four carboxylate ions (trifluoroacetate,
propiolate, cyanoacetate, and malonate) was determined in water at 140-280 °C and 275 bar. The
decarboxylation kinetics were determined in real time by using IR spectroscopy and a titanium cell flow-
reactor with sapphire windows. The complexity of the reaction of CF3CO2- necessitated additional postreaction
1
studies by H and 19F NMR spectroscopy and F- ion electrochemistry. Batch mode reactions in a titanium
tube reactor were employed for the latter studies. The CF3H product of CF3CO2- decomposed, liberating F-
ions. The cation effect for CF3CO2- is caused primarily by the reaction of F- with the Group 1 cations in the
order Li+ > Na+ > K+..., which changed the ionic strength. The rate of decarboxylation of cyanoacetate was
unaffected by the cation, whereas the Li+ ion slowed the rate of decarboxylation of propiolate, perhaps because
of complexation with propiolate. These results were compared with the previously investigated malonate
system. Where a cation effect was observed, the effect could be attributed to some form of ion pairing, but
in each case the details of the interaction differ slightly.
Introduction
cations affect the decarboxylation rate, but all of the effects seem
to fall under the umbrella of ion pairing.
One of the more important fundamental problems in hydro-
thermal chemistry is the effect of ionic substances on the
reactions of other organic and inorganic compounds. Whether
initially present or formed during a reaction, ionic substances
are frequently major components of hydrothermal reactions in
chemical processing and natural systems. Heavy emphasis has
been placed on the thermodynamic properties of metal-organic
systems,1-3 with less emphasis on kinetics.4,5 Thus, we have
attempted to cast more light on the effect of cations and anions
on the rates and pathways in selected hydrothermal reactions.
Experimental Section
The trifluoroacetate salts, propiolic acid HCtCCO2H, cyano-
acetic acid NtCCH2CO2H, LiOH, NaOH, KOH, and 50% by
wt aqueous RbOH and CsOH were obtained from Aldrich and
used as received. The Milli-Q deionized water was sparged with
compressed Ar to expel any gases before use. The aqueous
propiolate and cyanoacetate solutions of Li+, Na+, K+, Rb+,
and Cs+ were prepared by titrating the acids with metal
-
hydroxides in solution to the equivalence point. The CF3CO2
The intention was to determine the influence of the ionic
potential (charge/size) of Group 1 cations (M ) Li+, Na+, K+,
Rb+, Cs+) on the decarboxylation kinetics (reaction 1) of a series
of carboxylate anions, CF3CO2-, NCCH2COO-, HCCCO2-, and
HO2CCH2CO2-. Real-time IR spectral measurements with a
flow reactor were employed to compare the reaction rates.
salts were studied at 0.25 m concentration, whereas the
propiolate and cyanoacetate salts were studied at 0.5 m
concentration.
The flow reactor and Ti cell (0.0819 cm3) with sapphire
windows used in the experiments have been described in detail
elsewhere.7,8 Several improvements9,10 have automated the
operation. This apparatus was used to study hydrothermal
reactions in real time. The temperature range used was 140-
280 ( 1 °C at a constant pressure of 275 ( 1 bar. The flow
rates in the range of 0.10-1.00 ( 0.5% mL/min yielded
residence times of 4.4-44.5 s.
A Nicolet 560 Magna FTIR spectrometer with an MCT-A
detector was used for the transmission IR spectroscopy experi-
ments involving the flow cell. For kinetics, the IR spectral
measurements were recorded at 4 cm-1 resolution on 32
summed spectra. The total collection time for the summed
spectra was approximately 10 s. All spectra were normalized
with background spectra of pure water recorded at the same
conditions. The asymmetric stretch of aqueous CO2 at 2343
cm-1, CtC triple bond stretch at about 2090 cm-1, and CtN
triple bond stretch in the region of 2254 cm-1 were all within
the IR band-pass of sapphire and are useful for calculating the
kinetics of decarboxylation based on the absorbance change.
M[RCO2] + H2O f RH + CO2 + M+ + OH- (1)
-
The CF3CO2 ion was perhaps the most unusual because the
CF3H product engaged in further hydrothermal reactions that
produced F- ions. The presence of M+ and F- ions in solution
raised the ionic strength and caused an observable effect on the
rate of reaction 1, except in the cases where ion pairing was
the greatest.
The carboxylate ions containing the -CN and -CC- triple
bonds in the backbone were investigated to explore the effect
that the cation might have on possible resonance structures. The
malonate ion has been discussed before,6 and the effect of
chelation of the cation on the decarboxylation rate of the anion
was shown. In general, there appear to be many ways that the
* Corresponding author. E-mail: brill@udel.edu
10.1021/jp020941t CCC: $22.00 © 2002 American Chemical Society
Published on Web 10/24/2002

11108 J. Phys. Chem. A, Vol. 106, No. 46, 2002
Miksa et al.
The band areas of CO2 were converted into concentrations by
using the Beer-Lambert law with the previously determined
molar absorptivity of aqueous CO2.11 During decarboxylation
of the cyanoacetate ion, the CtN stretch of the CH3CN product
-
of decomposition overlapped that of the reactant NCCH2CO2
.
The contribution of CH3CN must, therefore, be subtracted from
the band area of NCCH2CO2- when determining the concentra-
tion of the anion. This procedure is described in the following
section. The weighted least-squares regression12 was performed
on the kinetics data in which the statistical weight was set to
be 1/σ2, where σ was the standard deviation of variables.
To calculate the total concentration of CO2 released from
decarboxylation of the anions, the equilibrium reactions 2 and
3 of CO2 in solution have to be taken into account.13
Ka1
CO2 + H2O y z H3O+ + HCO3
(2)
(3)
-
Ka2
Figure 1. Rate plot for the decarboxylation of 0.25 M K[CF3CO2] at
the temperatures shown and 275 bar pressure (reaction 5). The basis is
CO2 production.
HCO3- + H2O y z H3O+ + CO3
2-
The mass balance for the total CO2 concentration is given by
eq 4, where [CO2]free is the CO2 concentration based on the
symmetrical absorption band due to asymmetric stretching at
meter. NaF was used to prepare standards to create a ln [F-]
vs mV calibration plot. Total Ion Strength Adjuster Buffer
(ThermoOrion) was added to both the sample and standard and
an adjustment was made to account for the dilution of F-. All
measurements were made using polyethylene beakers, and the
solutions were stirred at the same rate using a Teflon coated
magnetic stirring bar.
2343 cm-1
.
-
2-
[CO2]t ) [CO2]free + [HCO3 ] + [CO3
]
(4)
Ka1
[H+]
Ka1Ka2
[H+]2
) [CO2]free 1 +
+
(
)
Results
Although Henry’s law guaranteed that all of the released CO2
was dissolved in the aqueous solution in hydrothermal solutions
at high temperature and pressure, this was further indicated by
the fact that we did not observe any gaseous CO2, which has
distinctive P and R rotational branches. The observed band
intensity (or concentration) of CO2 depends on the pH of the
solution. The equilibrium constants Ka1 and Ka2 were calculated
at high temperatures using the iso-Coulombic extrapolation
method.14 The ionization constant and specific volume of water
at high temperatures needed in the iso-Coulombic extrapolation
were calculated using the data of Marshall and Franck15 and
Uematsu and Franck.16
Hydrothermolysis of M[CF3CO2] Salts. The presence of
halocarbons in water is of concern because of their deleterious
effect on the environment. Removal of halocarbons from waste
streams by the use of hydrothermal methods has motivated
-
several studies.17-19 We chose to study CF3CO2 because of
its high solubility in water and the fact that few compounds
containing C-F bonds have been studied at hydrothermal
conditions. CF3CO2H is the strongest carboxylic acid and,
therefore, it exists mostly as CF3CO2- with counterions in water.
The effect of metal counterions on the decarboxylation rate of
-
CF3CO2 might therefore be an important phenomenon.
The hydrothermal behavior of CF3CO2- proved to be a two-
pronged problem. The decarboxylation reaction 5 occurred but
was followed by the decomposition of CF3H according to
reactions 6 and 7.
In addition to the real-time flow-reactor kinetics experiments,
batch mode experiments of hydrothermolysis were conducted
in order to characterize the reactions. A Ti tube with an internal
volume of 12 cm3 was sealed with the same aqueous solutions
prepared for the flow reactor and then heated in a fluidized sand
bath for the desired length of time. The amount of solution used
was based on the specific volume of water, in order that the
solution completely fill the tube at the reaction temperature.
M[CF3CO2] + H2O f CF3H + CO2 + M+ + OH- (5)
CF3H + 3OH- f HCO2- + 3F- + H+ + H2O (6)
The tube was then cooled in a water bath to quench the reaction,
and opened. In the case of CF3CO2
HCO2- + H2O f CO2 + H2 + OH-
(7)
,
19F and 1H NMR
-
spectroscopy and F- electrochemistry of the cooled solution
were also conducted. Because of the basicity of solutions of
lithium propiolate and cyanoacetate, a white precipitate (Li2CO3)
was found to form, which would plug the flow reactor. A similar
problem arose with lithium trifluoroacetate, but in this case the
precipitate was LiF. For this reason, lower temperatures and
shorter residence times were selected for kinetics experiments
involving these compounds in order to reduce the potential for
plugging of the reactor.
To address the relative roles of reactions 5-7 in the overall
behavior of CF3CO2-, we measured the rate of CO2 formation
in real time using a flow-reactor with IR spectroscopy and
1
designed a series of supporting H and 19F NMR spectral, F-
electrochemical, and pH measurements on the products from
postreaction solutions from the tube reactor.
In situ measurements with IR spectroscopy using the Ti flow
reactor were used to determine the initial decarboxylation rates
(reaction 5) of M[CF3CO2], where M ) Group 1 cations.
Conversion up to 40% based on the rate of formation of CO2
was used and was found to follow a first-order (or pseudo first-
order) rate law as the data in Figure 1 show. The rate of CO2
The concentration of F- was measured in quenched solutions
of 0.25 m CF3CO2M (M ) Li, Na) that had been heated to
200-260 °C for 5 to 30 min by means of a ThermoOrion model
94-09 fluoride electrode and an Accumet model 25 pH/ion

Spectroscopy of Hydrothermal Reactions
J. Phys. Chem. A, Vol. 106, No. 46, 2002 11109
-
TABLE 1: Rate Constants and Arrhenius Parameters for
the corresponding amount of CF3CO2 that had to have
Decarboxylation of M[CF3CO2] (M ) Li, Na, K, Rb, Cs)
decomposed, and these values are shown in Table 2. Table 2
-
temp
°C
k
Ea
ln(A)
s-1
∆S‡
also shows the resulting rate constants for removal of CF3CO2
s-1 (×104)
kJ‚mol-1
J‚K-1‚mol-1
based on F- production. By comparing the 260 °C data in Tables
-
1 and 2, it can be determined that only about 4% of the CF3CO2
Li trifluoroacetate
(0.25 m)
179.7 ( 16.8
37.4 ( 4.1
53
has produced F- (through reaction 6 in the Ti tube) after 5 min,
(210-250 °C)a
-
200
210
5.1 ( 1.2
7.2 ( 0.6
whereas 40% of the CF3CO2 has converted to CO2 in 40 s.
Thus reaction 5 is much faster than reaction 6. For a more
general view, the rates of disappearance of CF3CO2- based on
CO2 and F- production from Na[CF3CO2] are compared in the
Arrhenius plot shown in Figure 2. Since F- production requires
both reactions 5 and 6, and reaction 6 occurs more slowly, the
overall rate of F- formation is slower than CO2 production,
which requires only reaction 5.
220
230
240
250
15.2 ( 0.4
38.2 ( 0.3
85.5 ( 0.3
170.1 ( 0.6
Na trifluoroacetate
(0.25 m)
159.7 ( 6.0
32.43 ( 1.4
12
(200-260 °C)a
200
210
3.6 ( 0.4
5.7 ( 0.3
The pH of the solutions from the tube reactions gave
additional support for reactions 5-7. Figure 3 shows that the
solution is more basic at shorter reaction times, as might be
expected from the OH- production by reaction 5, but became
less basic as the reaction progressed because of the gradual
consumption of OH- by reaction 6. When the F- concentration
in the Na[CF3CO2] solution was followed, reaction 6 slowed at
longer times in part because the H+ formed decreased the pH.
Higher [OH-] may facilitate the decomposition reaction of CF3H
according to reactions 8 and 9. Nucleophilic attack by OH- at
220
230
240
250
15.6 ( 0.3
34.0 ( 0.2
74.6 ( 0.2
156.8 ( 0.4
253.9 ( 0.4
260
K trifluoroacetate
(0.25 m)
210
220
230
240
250
260
157.9 ( 15.7
32.5 ( 3.8
13
(220-240 °C)a
10.4 ( 0.2
27.1 ( 0.3
60.7 ( 0.4
113.1 ( 0.3
216.2 ( 0.2
274.3 ( 0.3
404.0 ( 0.2
488.5 ( 0.3
-
the carbon atom of CF3 is plausible based on the fact that a
Hartree-Fock SCF level calculation of the CF3- ion indicated
that the charge on the carbon atom is +0.5e.
270
280
Rb trifluoroacetate
(0.25 m)
220
230
240
250
260
Cs trifluoroacetate
(0.25 m)
220
230
240
250
260
270
215.6 ( 33.5
45.8 ( 8.0
39.7 ( 7.5
122
(220-240 °C)a
CF3H +OH- f CF3- + H2O
(8)
(9)
12.7 ( 0.6
39.4 ( 0.6
90.8 ( 0.5
157.6 ( 0.4
248.1 ( 0.4
CF3- + 2OH- f HCO2- + H+ + 3F-
The rate of reaction 7 is known from previous measurements
in a Ti flow cell.20 The rate of reaction 7 was extrapolated from
the higher temperature regime of measurement into that
determined here for reactions 5 and 6. Considered together, the
rates of reactions 5-7 follow the order 5>6>7, as is shown in
Figure 2. At longer reaction times the solution becomes slightly
more basic (Figure 3), perhaps because of reaction 7.
190.0 ( 31.3
72
(220-240 °C)a
13.4 ( 0.4
39.1 ( 0.4
84.8 ( 0.3
152.9 ( 0.3
247.8 ( 0.3
295.4 ( 0.2
Whether the F- might result from direct defluorination of
a Temperature range considered for calculation of Arrhenius param-
eters.
-
CF3CO2 rather than CF3H is partially answered by the faster
rate of formation of CO2 compared to F-. Additional evidence
that CF3H is the source of F- was obtained by comparing the
formation was converted into the rate of decomposition of
CF3CO2- by subtraction of the CO2 concentration at each time
from the initial CF3CO2 concentration. Table 1 gives the
resulting rate constants. Discussed below is the fact that CO2
formed in reaction 7 appears much more slowly and can be
ignored at this stage of the reaction.
-
integrated intensities of F- and CF3CO2 in the 19F NMR
-
spectrum, which are shown in Figure 4. Notice that most of the
-
CF3CO2 has disappeared before the F- ion is observed.
Conversion of the rates for reaction 5 into an Arrhenius plot
-
for the salts of CF3CO2 yielded Figure 5. Within the 95%
The CF3H product of reaction 5 slowly reacts at 200-260
°C and 275 bar according to reaction 6. The presence of CF3H
was clearly confirmed by 19F NMR spectroscopy, but quanti-
tation of CF3H is not reliable because of loss due to its high
vapor pressure and poor solubility in water. The presence of
HCO2- was confirmed by the use of 1H NMR spectroscopy of
the aqueous solution of CF3CO2- that had been heated in a tube
reactor at 260 °C for 15 min. F- was also clearly detected by
19F NMR spectroscopy. There were no 19F signals for other
species such as CH2F2 or CH3F. The rate of reaction 6 could
not be determined in the flow reactor because none of the species
absorb in the IR band-pass of the sapphire windows. Therefore,
the rate of reaction 6 was obtained by measuring the concentra-
tion of F- by the use of electrochemistry on solutions of
CF3CO2- heated at 200-260 °C in the tube reactor for the times
shown in Table 2. The F- concentrations can be converted to
confidence interval shown, all of the salts behave in the same
manner. However, the individual Arrhenius plots shown in
Figure 5 reveal that the Li+ and Na+ salts are linear in the
temperature range studied, whereas those of K+, Rb+, and Cs+
are curved and become less dependent on the temperature at
higher temperatures. The hypothesized explanation for the
curvature in the Na+, K+, and Rb+ salt data is as follows: The
production of F- by reaction 6 presents the opportunity to form
insoluble MF salts because the bulk dielectric constant of H2O
is lower at higher temperature. Both ion pair formation and
precipitation of MF are expected in the order Li+ > Na+ >...
>Cs+, that is, in the order of increasing ionic potential of the
cation. However, only LiF was observed to precipitate in the
batch reaction work. LiF also precipitated in the flow reactor
resulting in plugging at higher temperatures. While NaF
precipitate was not observed, considerable ion pairing is

11110 J. Phys. Chem. A, Vol. 106, No. 46, 2002
Miksa et al.
TABLE 2: Calculated Values for [CF3CO2Na]t and Rate Constants Based on Electrode Measurements of [F-]t
[CF3CO2Na]ta
230 °C
0.2489
0.2469
0.2448
0.2431
0.2418
0.24
11.97 ( 0.83
reaction
time (min)
200 °C
0.25
0.25
0.2499
0.2499
0.2498
0.2497
0.20 ( 0.49
210 °C
0.2499
0.2497
0.2496
0.2493
0.2491
0.2487
1.31 ( 0.72
220 °C
0.2501
0.2496
0.2489
0.2479
0.2478
0.2465
3.66 ( 0.65
240 °C
0.248
0.2444
0.2396
0.2378
0.2363
250 °C
0.2449
0.2388
0.2357
0.2335
0.2323
260 °C
5
10
15
20
25
30
0.2398
0.2322
0.2295
0.2276
0.2248
0.2251
21.33 ( 0.70
0.2358
12.44 ( 0.60
0.2305
15.84 ( 0.64
k, s-1 (×104)
a [CF3CO2Na]t ) {[CF3CO2Na]0 - ([F-]t/3)}.
Figure 4. The integrated 19F NMR signal intensity for aqueous Li-
Figure 2. Arrhenius plots comparing the rate constants of decomposi-
tion of Na[CF3CO2] based on F- (electrochemistry), which is indicative
of reaction 6, and CO2 (flow reactor) production, which is related to
reaction 5. Also shown are extrapolated rate constants for reaction 7
(HCO2-).
-
[CF3CO2] showing that most of the CF3CO2 has disappeared before
the F- appears. This is evidence that F- does not come directly from
CF3CO2-. The error bars and F- are within the size of the data points.
Figure 3. The change in pH (measured at 25 °C) of the quenched
solution from the Ti tube reaction of aqueous Na[CF3CO2] undergoing
reactions 5-7. The trend observed is consistent with the progress of
the reaction.
Figure 5. Arrhenius plots for the decarboxylation of the Group 1 cation
salts of CF3CO2- based on CO2 production showing the 95% confidence
interval.
expected at the hydrothermal conditions used. Thus, kinetic data
for the Li+ and Na+ salts could not be obtained in the highest
temperature range. The rates in Figure 5 are seen to have
decreasing temperature dependence at higher temperature for
the K+, Rb+, and Cs+ salts, perhaps because less ion pairing
with F- enables the effective ionic strength to increase faster
during the reaction. This hypothesis was supported by separate
experiments that showed that higher ionic strengths obtained
by the addition of excess KF to the reaction mixture reduced
the decarboxylation rate of aqueous K[CF3CO2] (Figure 6). Even
clearer evidence that removal of F- ions from the solution occurs
in the order Li+>Na+ was obtained by comparing the F-
-
concentration from electrochemical analysis of the CF3CO2
salts of Li+ and Na+ obtained from the batch mode decomposi-
tion. These results are shown in Figure 7. At higher temperatures
the LiF begins to precipitate, thus lowering the free F-
concentration, whereas the concentration of F- from the Na+
salt increases with temperature but at a decreasing rate. The
reduction of the decarboxylation rate as the ionic strength of
the solution is increased has also been observed with other acids,
such as the amino acid alanine.21

Spectroscopy of Hydrothermal Reactions
J. Phys. Chem. A, Vol. 106, No. 46, 2002 11111
Figure 6. Arrhenius plot showing that the addition of 0.25 m KF to
0.25 m K[CF3CO2] slows down the decarboxlyation rate. The dashed
line fits the four lowest temperature data points to emphasize the
curvature at the higher temperatures.
Figure 8. Plot based on the Kirkwood eq 9. All five alkali metal salts
of CF3CO2- exhibit a negative slope which is indicative of a transition
state that is less polar than the reactants.
Figure 7. The F- concentration from the electrochemical measurements
of aqueous solutions of the Li+ and Na+ salts of CF3CO2 showing
-
that F- is removed at higher temperatures in the order Li+>Na+ due
to MF association.
Based on the observations above, the transition state for
decarboxylation would appear to be less polar than the reactant.
A more theory-based assessment on the polarity of the transition
state can be made based on the Kirkwood analysis (eq 10),
which takes into account Coulombic interactions, dipole mo-
ments, radii, and the dielectric constant of the medium as they
apply to kinetics.22
Figure 9. First-order rate plot for decarboxylation of 0.5 m
K[HCCCO2] at 275 bar.
µ2TS µ2A µ2B
less polar than the reactants, the slopes should be negative. If
the slope is negative, the (µ2TS/rT3S - µA2 /rA3 - µB2 /r3B) portion of
eq 10 must also be negative. This can be possible only if the
dipole moments of reactants A and B are larger than that of the
transition state. Figure 8 is a plot of ln kH vs (ꢀ - 1/2ꢀ + 1) for
the five alkali metal salts of CF3CO2-. In all five cases the slopes
are negative, confirming that the transition state formed during
the heterolytic cleavage of the C-C bond is less polar than the
reactants.
Hydrothermolysis of M[HCCCO2] Salts. The hydrother-
molysis of propiolic acid has been investigated recently.24
Propiolic acid is a weaker acid than trifluoroacetic acid.
However, by using the iso-Coulombic extrapolation method,
only 0.02% of the 0.5 m propiolate ion is converted to the
N
ꢀ - 1
ln kH ) ln k0 +
-
-
(10)
3
rTS r3A r3B
(
)(
)
RT 2ꢀ + 1
where kH is the hydrolysis constant at temperature T, k0 is the
rate constant in a hypothetical medium having ꢀ ) 1, and µ
and r are the dipole moment and radii of reactants A, B, and
the transition state (TS). Equation 10 has been treated in a way
that does not require knowledge of the dipole moments and radii
of the reactants and the transition state.23 Equation 10 has the
form of the equation of a straight line. Therefore, if ln k0 is
taken to be the y intercept and N/RT is constant, then a plot of
ln kH vs (ꢀ - 1/2ꢀ + 1) will result in a line with slope (µ2 /r3
TS TS
- µ2A/r3A - µB2 /r3B). For reactions with transition states that are

11112 J. Phys. Chem. A, Vol. 106, No. 46, 2002
Miksa et al.
TABLE 3: Rate Constants and Arrhenius Parameters for
Decarboxylation of Propiolate Salts
temp
°C
k
Ea
kJ‚mol-1
ln(A)
s-1
∆S‡
s-1 (×103)
J‚K-1‚mol-1 a
Li propiolate
(0.5 m)
140
150
160
146.8 ( 2.9 37.1 ( 0.8
52
3.6 ( 0.3
9.6 ( 0.3
26.7 ( 0.7
64.1 ( 1.5
170
Na propiolate
(0.5 m)
140
150
160
133.9 ( 2.8 33.7 ( 0.8
23
5.0 ( 0.1
13.4 ( 0.3
31.7 ( 0.7
68.7 ( 1.1
169.8 ( 4.8
170
180
K propiolate
(0.5 m)
140
150
160
170
180
Rb propiolate
(0.5 m)
140
150
160
170
180
Cs propiolate
(0.5 m)
140
150
160
136.6 ( 1.5 34.6 ( 0.5
133.8 ( 2.3 33.9 ( 0.6
134.5 ( 1.6 34.1 ( 0.5
134.70 ( 1.1 34.1 ( 0.3
31
25
26
5.5 ( 0.1
14.8 ( 0.2
36.5 ( 0.3
83.9 ( 0.6
183.6 ( 4.8
Figure 10. Effect of the Group 1 cations on the decarboxylation rate
5.5 ( 0.1
15.7 ( 0.2
38.4 ( 0.4
84.3 ( 1.2
191.1 ( 1.9
-
of HCCCO2 at various temperatures.
5.8 ( 0.1
15.8 ( 0.2
37.1 ( 0.5
85.8 ( 1.0
192.1 ( 2.9
170
180
propiolate anion b
a Calculated at average temperature. b Excludes Li+ data.
neutral propiolic acid at temperature 200 °C, and therefore
propiolic acid need not be considered further in the calculation.
The IR spectra for the decarboxylation of propiolate anion
at 275 bar and temperature range from 140 to 190 °C were based
on the disappearance of the 2090 cm-1 band due to the CtC
of propiolate. The CtC stretching mode of the C2H2 product
of decarboxylaton is of course IR inactive. Figure 9 shows the
rate plot for the decarboxylation of 0.5 m potassium propiolate
solution at 275 bar, which indicates that the decarboxylation
process is pseudo-first-order. The rate constants with different
counterions and Arrhenius parameters are given in Table 3. The
effect of the counterions on the decarboxylation rate is displayed
in Figure 10, where a plot of the rate constants versus the ion
radii25 in aqueous solution is shown. It can be seen that the rate
constant of decarboxylation increases with the enlargement of
ion radius of counterion, but Table 3 reveals that the mean value
activation energy and frequency factor for the Na+, K+, Rb+,
and Cs+ salts are not significantly different. The Li+ salt is,
however, significantly different and produces a lower rate of
decomposition.
Figure 11. IR spectra for Na[NCCH2CO2] from the flow reactor at
275 bar showing the nitrile overlap region and CO2 formation.
The kinetics of reaction 11 were calculated from the observed
CtN band of NCCH2CO2-. However, the product CH3CN has
two bands located at 2297 and 2257 cm-1. The band at 2257
cm-1 overlapped the 2254 cm-1 band of the reactant nitrile.
Therefore, the CH3CN contribution to the band area of 2254
cm-1 must be removed to obtain the concentration of
NCCH2CO2-. The hydrolysis of CH3CN to acetamide and then
to acetic acid is slow in neutral hydrothermal solution even at
300 °C and can be assumed to be negligible26 within the error
limits of the kinetic experiments. Nevertheless, the absorptivity
of CH3CN decreased with increasing temperature in the 100-
330 °C range at 275 bar according to flow reactor IR spectros-
copy. This changing absorptivity, which is not due to reaction
of CH3CN, was taken into account to reduce the uncertainty in
M[NCCH2CO2]. The decarboxylation kinetics of cyanoacetic
acid and its anion have been studied in our previous work,9
where the decarboxylation reaction 11 was clearly seen from
the spectra at different temperatures (Figure 11). Less than 1%
of the cyanoacetate ion hydrolyzes to the acid at 200 °C so that
the acid contribution can be ignored.
NtCCH2CO2- + H2O f CH3CtN + CO2 + OH- (11)

Spectroscopy of Hydrothermal Reactions
J. Phys. Chem. A, Vol. 106, No. 46, 2002 11113
TABLE 4: Rate Constants and Arrhenius Parameters for
Decarboxylation of Cyanoacetate Salts
temp
°C
k
Ea
kJ‚mol-1
ln(A)
s-1
∆S‡
s-1 (×103)
J‚K-1‚mol-1
Li cyanoacetate
(0.25 m)
160
170
180
190
200
210
Na cyanoacetate
(0.25 m)
160
170
180
190
200
210
K cyanoacetate
(0.25 m)
160
170
180
190
200
210
Rb cyanoacetate
(0.25 m)
160
170
180
190
200
210
143.5 ( 8.5 32.5 ( 2.2
145.8 ( 4.1 33.1 ( 1.1
141.9 ( 4.5 32.1 ( 1.2
141.3 ( 6.3 31.9 ( 1.6
143.5 ( 4.4 32.5( 1.2
143.2 ( 2.6 32.2 ( 0.7
13a
17
10
8
0.6 ( 0.1
1.3 ( 0.1
3.7 ( 0.1
9.2 ( 0.2
19.3 ( 0.5
32.8 ( 1.0
0.5 ( 0.1
1.4 ( 0.04
3.8 ( 0.1
8.9 ( 0.2
18.3 ( 0.5
36.6 ( 1.3
0.51 ( 0.03
1.7 ( 0.1
3.8 ( 0.1
9.1 ( 0.1
18.4 ( 0.4
37.4 ( 0.8
Figure 12. Effect of the Group 1 cations on the decarboxylation rate
-
of NCCH2CO2 at various temperatures.
shown in Figure 5 to have the same rate within the 95%
confidence interval at the conditions used.
0.56 ( 0.03
1.39 ( 0.05
4.0 ( 0.1
9.7 ( 0.2
18.4 ( 0.3
35.6 ( 0.8
Discussion of Counterion Effects. Careful studies of the rates
of decarboxylation of carboxylate ions uncover the fact that the
Group 1 cations can influence the rate of decomposition in a
variety of ways. Although the differences are not large, each of
the four anions studied behaves slightly differently. Therefore,
the counterion effects would appear to have many origins and
therefore be difficult to predict without guidance from experi-
mental data.
Cs cyanoacetate
(0.25 m)
160
170
13
0.53 ( 0.04
1.50 ( 0.04
3.8 ( 0.1
9.0 ( 0.1
18.4 ( 0.4
36.1 ( 1.1
180
190
200
210
The mechanism of decarboxylation of carboxylic acids and
carboxylates in aqueous solution is often accepted28 to be similar
to that of SN2 nucleophilic bimolecular substitution. In aqueous
solution, water molecules act as the nucleophile, making the
observed rate constant pseudo-first-order. First, the nucleophile
attacks the carbonyl carbon atom, followed by the loss of carbon
dioxide and the formation of the carbanion by heterolytic
cleavage. The activated complex is formed by initial nucleophilic
attack on the carbonyl carbon atom. The carbanion picks up
H+ to form the neutral species, in this case CF3H. An analogous
mechanism involving ion pairing29 that may be more relevant
to the carboxylate salts involves two transition states. The first
is the formation of contact ion pairs and the second is the
formation of an activated complex by nucleophilic attack of
water molecules on the ion pair. If formation of the first
transition state is rate-determining, then the global decarboxyl-
ation process is a unimolecular reaction. In accordance, the
entropies of activation in Tables 1, 3, and 4 tend to be positive.
If formation of the second transition state is rate-determining,
then the reaction should be bimolecular. Countercations having
the highest ionic potential might stabilize the carbanions formed
by heterolytic cleavage in the decarboxylation process involving
the ion pair mechanism because they form contact ion pairs or
solvent separated ion pairs more efficiently at hydrothermal
conditions. As a result, smaller counterions might be expected
to increase the rate constant.
cyanoacetate anionc
a Calculated at average value of temperature range.
the concentration calculation. The process was divided into three
steps. First, the total CO2 concentration was calculated from
the observed CO2 concentration using the [H+] of initial
solutions at each temperature and residence time. Second,
according to eq 11, the concentration of CH3CN is equal to
that of the total CO2 concentration, which enabled the band areas
of CH3CN to be converted to concentrations at each temperature
and residence time. Finally, the band areas of reactant
-
NCCH2CO2 were obtained by subtracting the CH3CN band
area from the total band area centered at 2254 cm-1. After this
correction, the calculated rate constants were found to be the
same within the uncertainty of the measurement. The obtained
rate constants are given in Table 4 for the decarboxylation of
the cyanoacetate ion with different counterions. The effect of
the ion radius on the decarboxylation rate is shown in Figure
12. No significant effect was observed.
Hydrothermolysis of M[O2CCH2CO2H] Salts. A previous
kinetic study of the rate of decarboxylation of M[O2CCH2CO2H]
salts M ) Li, Na, K, Rb, Cs also revealed some curvature in
the Arrhenius plots at higher temperatures.6 The data show that
the rate increases with increasing cation size, with the possible
exception of Rb+.6,27 The sensitivity of the rate of the malonate
salts to the cation contrasts with the CF3CO2- salts, which were
There are several problems with the present data in the
scenerio given above. For example, the rate of decarboxylation
of the cyanoacetate anion with different counterions revealed

11114 J. Phys. Chem. A, Vol. 106, No. 46, 2002
Miksa et al.
only small variation with the choice of the counterion. Thus, in
the trifluoroacetate anion, where the ion pairing between CF3
fastest and any species that competes with H+ for the position
in the ring will reduce the decarboxylation rate.6 When M is a
Group 1 cation, M+ competes with H+ for the position in the
ring in the order Li+ > Na+ > K+. Hence, the rate of
decarboxylation of the Li+ salt is the lowest.
-
and M+ should be greatest, Li+ instead retards the rate of
decarboxylation. The decarboxylation rate of the trifluoroacetate
anion is mostly affected by the secondary ionic strength effect
in which F- is complexed by the cation. The addition of KF to
the trifluoroacetate retarded the reaction rate (Figure 6). In the
propiolate anion, only the Li+ ion affects the rate differently
from the other ions and retards the decarboxylation reaction.
In the malonate anion, there is a more or less systematic trend
of decreasing rate with increasing ionic potential of the cation.
These differences can be understood better by hypothesizing
slightly different reasons for the counterion effect in accordance
with the experimental findings. Indeed, the ion association
concept can be retained in all cases, but with detailed differences
that depend on the carboxylate ions. The behavior of the
CF3CO2- ion is a somewhat special case in that the cation effect
appears to result primarily from the change of ionic strength of
the solution during the reaction. As the F- ions form as a result
of secondary decomposition of the CF3H product, they remove
the Li+ ions by precipitation and the Na+ ions by ion pairing.
This leads to less influence by the salt effect for the Li+ and
Na+ salts compared to the K+, Rb+, and Cs+ salts. The
decarboxylation rate of trifluoroacetate salts of the latter three
ions therefore increases less rapidly at high temperature. In the
case of propiolate, only the Li+ ion affects the rate compared
to the other Group 1 cations. Since Li+ has the highest ionic
potential, it can associate most strongly with the HCCCO2- ion
as in structure I. Stabilization of this ion pair may involve the
resonance form II, which strengthens the C-CO2 bond and
slightly reduces the decarboxylation rate.
Acknowledgment. We are grateful for support of this work
by the National Science Foundation on CHE-9807370 and the
Army Research Office on DAAG55-98-1-0253.
References and Notes
(1) Shock, E. L.; Korestky, C. M. Geochim. Cosmochim. Acta 1993,
57, 4899.
(2) Shock, E. L.; Korestky, C. M. Geochim. Cosmochim. Acta 1995,
59, 1497.
(3) Sverjensky, D. A.; Shodi, E. L.; Helgeson, H. C. Geochim.
Cosmochim. Acta 1997, 61, 1359.
(4) Bell, J. L. S.; Palmer, D. A.; Barnes, H. L.; Drummond, S. E.
Geochim. Cosmochim. Acta 1994, 58, 4155.
(5) Belsky, A. J.; Maiella, P. J.; Brill, T. B. J. Phys. Chem. A 1999,
103, 4253.
(6) Gunawardena, N. R.; Brill, T. B. J. Phys. Chem. A 2001, 105,
1876.
(7) Kieke, M. L.; Schoppelrei, J. W.; Brill, T. B. J. Phys. Chem. 1996,
100, 7455.
(8) Schoppelrei, J. W.; Kieke, M. L.; Wang, X.; Klein, M. T.; Brill, T.
B. J. Phys. Chem. 1996, 100, 14343.
(9) Belsky, A. J.; Maiella, P. G.; Brill, T. B. J. Phys. Chem. A 1999
103, 4253.
(10) Brill, T. B. J. Phys. Chem. A 2000, 104, 4343.
(11) Maiella, P. G.; Schoppelrei, J. W.; Brill, T. B. Appl. Spectrosc.
1999, 53, 351.
(12) Cvetanovic, R. J.; Singleton, D. L. Int. J. Chem. Kinet. 1977, 9,
481.
(13) Butler, J. N. Ion Equilibrium, Solubility and pH Calculation; John
Wiley & Sons: New York, 1998; p 368.
(14) Lindsay, W. T. Proc. Int. Water Conf. Eng. Soc. W. Pa. 1980, 41,
284.
(15) Marshall, W. L.; Franck, E. U. J. Phys. Chem. Ref. Data 1981, 10,
295.
(16) Uematsu, M.; Franck, E. U. J. Phys. Chem. Ref. Data 1980, 9,
1291.
(17) Houser, T. J.; Liu, X. J. Supercrit. Fluids 1996, 9, 167.
(18) Park, S. W.; Yoon, J. H.; Lee, H. Korean J. Chem. Eng. 1996, 13,
640.
(19) Marrone, P. A.; Arias, T. A.; Peters, W. A.; Testor, J. W. J. Phys.
Chem. A 1998, 102, 7013.
(20) Maiella, P. G.; Brill, T. B. J. Phys. Chem. A 1998, 102, 5886.
(21) Li, J.; Wang, X. G.; Klein, M. T.; Brill, T. B. Int. J. Chem. Kinet.
2002, 34, 271.
(22) Newman, K. E. Chem. Soc. ReV. 1994, 23, 31.
(23) Klein, M. T.; Torry, L. A.; Wu, B. C.; Townsend, S. H. J. Supercrit.
Fluids 1990, 3, 222.
(24) Li, J.; Brill, T. B. J. Phys. Chem. A 2001, 105, 6171.
(25) Marcus, Y. Chem. ReV. 1988, 88, 1475.
(26) Izzo, B. H.; Klein M. T. AIChE J. 1997, 43, 2048.
(27) Brill, T. B.; Belsky, A. J.; Gunawardena, N.; Miksa, D. High Press.
Res. 2001, 20, 429.
-
In the case of the NCCH2CO2 ion the resonance form II
does not exist and hence the cation has little effect on the
decarboxylation rate. In the case of malonate, where there exists
a systematic difference in the decarboxylation rate depending
on the cation, the probable reason is that the transition state
may be different from the other anions studied. For the malonate
ion the transition state is likely to be a six-member ring structure
which facilitates the proton-transfer step essential to decarboxyl-
ation. Thus, when M ) H+, the rate of decarboxylation is the
(28) Clark, L. W. In The Chemistry of Carboxylic Acids and Esters.
The Chemistry of Functional Groups Series; Patai, S., Ed.; John Wiley:
New York, 1969; p 589.
(29) Solomons, T. W. G. Organic Chemistry, 4th ed.; John Wiley &
Sons: New York, 1988.