Spectroscopy of hydrothermal reactions. 10. Evidence of wall effects in decarboxylation kinetics of 1.00 m HCO2X (X = H, Na) at 280-330 °C and 275 bar

Article abstract of DOI:10.1021/jp981165l

Decomposition kinetics of 1.00 m formic acid and sodium formate to form CO₂ (and H₂) were probed under hydrothermal conditions of 280-330 °C and 275 bar. Flow reactor spectroscopy cells constructed from different metals (316 stainless steel, 90/10 Pt/Ir alloy, and grade 2 Ti) with diamond and sapphire windows were used to determine the rate of formation of CO₂ in situ by IR spectroscopy. A single fluid phase was maintained. CO₂ was produced at different rates depending on the metal and, in the case of 316 stainless steel, with a lot-dependent, concentration-time profile. The values of E_a = 87-121 kJ/mol resemble those previously reported for surface- and H₂O-catalyzed decarboxylation of formic acid. Much higher E_a values are reported for unimolecular decomposition. Sodium formate decarboxylates more slowly than formic acid. In contrast to formic acid, the rate of decarboxylation of malonic acid in all of the cells is the same within error, which is consistent with a homogeneous unimolecular reaction.

Full text of DOI:10.1021/jp981165l

5886
J. Phys. Chem. A 1998, 102, 5886-5891
Spectroscopy of Hydrothermal Reactions. 10. Evidence of Wall Effects in Decarboxylation
Kinetics of 1.00 m HCO2X (X ) H, Na) at 280-330 °C and 275 bar
P. G. Maiella and T. B. Brill*
Department of Chemistry and Biochemistry, UniVersity of Delaware, Newark, Delaware 19716
ReceiVed: February 17, 1998; In Final Form: May 11, 1998
Decomposition kinetics of 1.00 m formic acid and sodium formate to form CO2 (and H2) were probed under
hydrothermal conditions of 280-330 °C and 275 bar. Flow reactor spectroscopy cells constructed from
different metals (316 stainless steel, 90/10 Pt/Ir alloy, and grade 2 Ti) with diamond and sapphire windows
were used to determine the rate of formation of CO2 in situ by IR spectroscopy. A single fluid phase was
maintained. CO2 was produced at different rates depending on the metal and, in the case of 316 stainless
steel, with a lot-dependent, concentration-time profile. The values of Ea ) 87-121 kJ/mol resemble those
previously reported for surface- and H2O-catalyzed decarboxylation of formic acid. Much higher Ea values
are reported for unimolecular decomposition. Sodium formate decarboxylates more slowly than formic acid.
In contrast to formic acid, the rate of decarboxylation of malonic acid in all of the cells is the same within
error, which is consistent with a homogeneous unimolecular reaction.
Introduction
(Ea is in units of kJ/mol).25 Reaction 2 was reported to have
Ea ) 258-280 kJ/mol in another study.31 Shock-tube kinetics
at still higher temperatures are available.26,30 Reaction 1 was
found to have second-order rate behavior with Ea ) 272-284
kJ/mol,30 whereas reaction 2 had Ea ) 259-27230 or 169 kJ/
mol.26 These reactions have also been analyzed quantum
mechanically,8,26-29,32 and it is found that eq 1 has a higher
activation energy (range ) 272-372 kJ/mol; average ) 304
kJ/mol) than eq 2 (range ) 259-284 kJ/mol; average ) 276
kJ/mol). Consistent with early experimental data,22 the inclusion
of H2O in the transition state lowers the activation energy of
reaction 1 substantially.8,27 One H2O lowers the Ea of reaction
1 to 157,8 200 kJ/mol,32 or 204 kJ/mol,27 whereas two H2O
molecules lowers Ea to 90 kJ/mol8 or 189 kJ/mol.32 Therefore,
computational and experimental evidence reveals that the
presence of H2O and/or a metal surface substantially lowers
the Ea of reaction 1 compared to the unimolecular decomposition
pathway.
Formic acid, HCO2H, is the simplest carboxylic acid and, as
such, is regarded to be fundamentally important as an intermedi-
ate in the oxidation of various hydrocarbons.1-4 Recent interest
has also focused on the possible role of HCO2H as an
intermediate in the water-gas shift reaction in supercritical
water5-10 and as a component in wet-air oxidation.11 As an
example of a simple carboxylic acid, HCO2H has been widely
employed in prototypical studies of the formation and stability
of intermediates on clean metal and metal oxide surfaces and
for comparing the catalytic behavior of materials.12-18 These
latter studies may also be pertinent to the corrosion of stainless
steel by formic acid.19,20
The decomposition pathways 1-3 have been indicated for
HCO2H:12,13
HCO2H f CO2 + H2
f CO + H2O
(1)
(2)
Direct spectral observations of the species during hydrother-
molysis of HCO2H at high pressure and temperature had not
been reported when this work began.33 Such studies should
provide more insight into the homogeneous pathway or a
surface-mediated pathway when a single-phase solution is
present. The unusual CO2 formation profile observed in a 316
stainless steel (SS) cell33 suggested that the reactor surface may
influence the rate. Design features of the reactor could,
therefore, be important in the kinetics obtained for formic acid.
If competition exists between the homogeneous and wall-
mediated processes, then a large variation in the surface-to-
volume ratio (S/V) should shift the balance of these two general
processes. In addition, the role of surface catalysis should be
detectable by changing the materials used in the construction
of the cell. In the present study a large S/V reactor with a
spectroscopically accessible zone34,35 was used, and the reaction
was observed in real time by transmission IR spectroscopy
through a thin sheet of solution. It is experimentally impractical
to vary S/V within our cell, but the material of construction can
be changed while maintaining the same cell design. To this
f 1/2CO2 + 1/2H2CO + 1/2H2O
(3)
Reactions 1 and 2 are surface-catalyzed in ratios that depend
on the substrate, whereas reaction 3 does not occur on a metal
surface. The Arrhenius activation energy for first-order de-
composition of HCO2H by reactions 1 and 2 on clean Ni
surfaces is 100-109 kJ/mol,14,21 with A ) 1015 s-1 17
This range
.
of Ea values resembles Ea ) 105-113 kJ/mol22 for HCO2H
with 1% H2O and Ea ) 86 kJ/mol for 0.022 m HCO2H at 320-
420 °C in a C-276 Hastelloy tube reactor.23 Reactions 1 and 2
also are indicated to be decomposition pathways in the gas
phase.24-31 For example, the rate constant for the first-order
decarboxylation reaction 1 is 1012.47 exp(-203/(RT)) s-1, while
dehydration by reaction 2 was second order below 600 °C with
a rate constant of 1011.44 exp(-133/(RT)) cm3 mol-1 s-1 and
1.5 order above 670 °C with k ) 1015.39 exp(-253/(RT)) s-1
* Corresponding author. E-mail: brill@udel.edu.
S1089-5639(98)01165-7 CCC: $15.00 © 1998 American Chemical Society
Published on Web 06/19/1998

Decarboxylation of HCO2X (X ) H, Na)
J. Phys. Chem. A, Vol. 102, No. 29, 1998 5887
end, cells were constructed of SS, Ti, and Pt/Ir alloy containing
sapphire and diamond windows. Unlike experiences with
amines36,37 and other carboxylic acids,37,38 the formation rate
of CO2 from HCO2H depended significantly on the cell wall
composition, which suggests that heterogeneous reactions can
be important, especially at high S/V values. Such details may
be important in interpreting and understanding practical hydro-
thermal reactions, such as occur in geochemical events,39,40
biomass conversion,41,42 and waste-stream remediation.43,44
Experimental Section
The kinetic experiments were conducted with spectroscopy
cells constructed from four types of metals: 90/10 Pt/Ir with
diamond windows,35 types 304 and 316 stainless steel with
sapphire windows,34 and grade 2 titanium with sapphire
windows. The precise characteristics of the metal surface in
these cells are difficult to specify and could even be viewed as
another variable. The SS and Ti cells are expected to have
mostly a metal oxide surface. The design of the cells is the
same in terms of the flow pattern and was made by the same
machinist. S/V is 20-50 cm-1. The Ti cell was constructed
with fittings rather than Ti-Ti welded joints because it was
found that aqueous CO2 rapidly corroded the welded spots,
causing poor flushing and eventual leaking. This was true even
when care was taken to eliminate O2 during the welding process.
In each case the cell is designed for transmission IR spectros-
copy in which the path length is controlled by a gold-foil spacer
having a thickness of 30 ( 5 µm. The exact path length in
each case was determined from the interference fringe spacing
in the IR spectrum.34,35 The heat transfer and fluid mechanics
characteristics of these cells have been modeled,35 and the flow
controls are described elsewhere.34,35 Plug-flow conditions
exist at higher flow rates (lower degree of conversion); however,
laminar flow is a better description at low flow rates (higher
degree of conversion). As a result of the desire to assume plug-
flow, the degree of conversion studied was 40% or less.
The flow rate was controlled by a pulseless, dual-piston, LDC-
Analytical HPLC pump or an Isco syringe pump. The results
of experiments with two different pumps were the same,
indicating that pump characteristics are not a factor in the kinetic
data. Five electrical cartridge heaters inserted concentrically
in the cell body were used to heat the cell to a temperature that
is controlled at (1 °C and monitored by Omega PID units
attached to 0.8 mm K-type thermocouples near the flow path.
The pressure was maintained at 275 ( 1 bar by using an air-
actuated bleed-and-lock system. The pressure was actively
monitored with an Omega melt-pressure transducer. Software
written in Visual Basic controlled and recorded the temperature,
pressure, and flow rate. Spectra were taken only after the system
had stabilized at the desired conditions.
Figure 1. Mid-IR spectra of 1.00 m HCO2H in the Pt/Ir diamond flow
cell at 330 °C under 275 bar as a function of the residence time.
then ratioed against a spectrum of pure water at the same
conditions. Figure 1 shows a series of IR spectra recorded with
the Pt/Ir diamond cell. The interference fringes are apparent
in the higher energy region and provide an accurate measure-
ment of the path length of the cell (vide supra). The CO2
absorbance was followed and, by the absence of gas-phase
features, a single (liquid) phase existed at all conditions.
Postreaction examination of the cells, even after use for extended
periods, did not reveal evidence of corrosion except in the case
of 304 SS. Only a slight discoloration of the surface occurred
in the other cells.
Conversion of the ν3 (CO2) absorbance areas into concentra-
tions at hydrothermal conditions required the use of calibration
data based on the Lambert-Beer law.34,35 The band area of ν3
(CO2) was obtained by fitting each spectrum with a four-
parameter Voigt function (Peakfit, Jandel Scientific). A program
written in Visual Basic was used to remove the interference
fringe pattern. Multiple data sets were collected to determine
the error in the rate constant k. A weighted least-squares
regression was then performed where the statistical weight, ωi,
used was 1/σ2, where σ is the standard deviation of the
concentration values at each time i. Where necessary, ωi was
approximated as k2ωi′.46 In a first-order rate constant and
Arrhenius analysis, the error limits were translated into log space
as σ/xj, where xj is the average CO2 concentration.46
Results and Discussion
The decomposition rate of HCO2H by reaction 1 is apparent
in all of the vibrational modes in the IR spectrum (Figure 1).
Several modes are, however, more useful than others. The C-H
stretch of HCO2H (2940 cm-1) could be used, but its low
intensity produces considerable uncertainty in the concentration.
The decrease in the intensity of the predominantly C-O stretch/
O-H bending mode47 centered at 1175 cm-1 could be followed
in the Pt/Ir cell but was not used because it is outside the IR
band-pass of the sapphire windows used in the other cells. The
most useful mode is ν3 [CO2(aq)] at 2343 cm-1 because we
have considerable experience with the line shape and absor-
bance-concentration relation in the hydrothermal medium,34,45
and it can be seen with all of the cells. There is no experimental
evidence for reaction 2 in part because, unlike in the gas phase,
this reaction is less important than reaction 1 in H2O11 and in
part because the IR absorbance of CO is 0.047 times that of
ν3(CO2). Likewise, reaction 3 seems unlikely by the absence
A 1.00 molal (m, moles/kg) solution of formic acid (Sigma
Chemical Company, 99%) was prepared using HPLC-grade
water that had been sparged with Ar to remove atmospheric
gases. Sodium formate solutions (1.00 m) were made by adding
1 equiv of NaOH. The hydrothermolysis reaction was studied
between 280 and 330 °C using 10-40 flow rates at each
temperature depending upon the cell and pump employed.
Residence times in the range 2.5-20.6 s in the SS cells, 0.8-
7.0 s in the Pt/Ir cell, and 4.8-71.4 s in the Ti cell existed after
correcting for the change in density of the solution with
temperature (i.e., multiply by Fo/FT). Spectra were collected
using a Nicolet 60SX FTIR spectrometer equipped with a
MCT-B detector. The resolution was 4 cm-1 with 32 interfer-
rograms being summed at each condition. Each spectrum was
of the CH2 bending absorbance of CH2O at about 1500 cm-1
.

5888 J. Phys. Chem. A, Vol. 102, No. 29, 1998
Maiella and Brill
Figure 3. CO2 concentration profiles from 1.00 m HCO2H in the 316
SSa cell under 275 bar pressure showing anomalous but reproducible
behavior.
Figure 2. Concentration of CO2 from 1.00 m HCO2H at 310 °C and
275 bar as a function of residence time in four different cell materials.
On the other hand, CO2 is produced by two reactions (5 and 6)
in the hydrothermal medium because of the equilibrium reaction
4:
HCO2H h HCO2- + H+
(4)
(5)
(6)
k1
HCO2H
98 CO2 + H2
k2
HCO2- + H2O
9
8 CO2 + H2 + OH-
Consequently, as will be discussed later, decarboxylation of
HCO2H in H2O necessitates consideration of the kinetics from
-
both the HCO2H and HCO2 sources.
Figure 4. First-order rate plots for decarboxylation of 1.00 m HCO2H
A central feature of this work is the observation that the rate
of formation of CO2 depends on the material used to construct
the cell. The gold foil used as a spacer in every cell is not a
variable. The diamond and sapphire windows are neither
catalytic nor a variable. On the other hand, the cell body, which
was made from two grades of SS, 90/10 Pt/Ir alloy, and grade
2 Ti, is a major variable. Figure 2 shows the rate of formation
of CO2 from 1.00 m HCO2H at 310 °C under 275 bar in two
316 SS cells (hereafter called 316 SSa and 316 SSb), the Pt/Ir
alloy cell, and the Ti cell. Data from the 304 SS cell are not
shown because, unlike the other materials, this cell was severely
corroded by HCO2H and quickly developed a leak. The inset
in Figure 2 is an expansion of the 0-7 s region showing that
small differences exist for the 316 SS and Pt/Ir cells in the early
stage.
Two major features are apparent in Figure 2. First, the rate
of CO2 formation is markedly retarded in the Ti cell. We
considered the possibility that CO2 could migrate into the TiO2
surface layer on Ti by the formation of Ti(CO3)2. There was
no evidence that this took place in the form of surface corrosion.
Moreover, all of the CO2 was instantly flushed from the cell
by pure H2O as opposed to displaying a more gradual decrease
that would be expected from shifting of the Ti(CO3)2 h TiO2
+ 2CO2 equilibrium. It should be noted that welded Ti parts
could not be used in this work because they were found to be
rapidly corroded by CO2-H2O at 300 °C, perhaps because of
the forementioned equilibrium.
at 275 bar in the 316 SSb cell.
TABLE 1: First-Order Rate Constants for Decarboxylation
of 1.00 m HCO2H (k1) and HCO2Na (k2) under 275 barc
cell type
316 SSaa
T, °C
k1, s-1 (×103)
k2, s-1 (×103)
280
290
310
320
330
290
300
310
320
300
310
320
330
290
300
310
320
5.9 ( 0.8
12.3 ( 1.6
25.7 ( 3.6
38.5 ( 6.3
46.0 ( 7.5
7.7 ( 0.1
10.5 ( 0.1
14.7 ( 0.1
25.0 ( 0.4
11.0 ( 0.8
20.7 ( 1.3
34.6 ( 0.8
44.7 ( 1.3
1.8 ( 0.1
2.8 ( 0.3
3.5 ( 0.1
3.9 ( 0.4
3.2 ( 0.3
4.7 ( 0.5
9.9 ( 1.9
13.7 ( 1.9
19.5 ( 1.9
4.6 ( 0.3
5.7 ( 0.3
316 SSb
Pt/Ir
13.8 ( 0.7
15.0 ( 1.1
17.8 ( 2.2
Ti
b
a Initial stage before the quiescent stage (see Figure 2). b Rate was
too slow to be measured on the time scale of the experiment. c Errors
are standard deviations at 90% confidence interval.
cells. Figure 3 shows, however, that a reproducible quiescent
stage developed in the 316 SSa cell followed by an “explosive”
stage. Considerable data scatter occurred in the explosive stage,
suggesting that the reaction is not well controlled by the reactor.
These cells were constructed from different lots of 316 SS,
although both lots were found to have the same elemental
A second major difference in Figure 2 is the different behavior
of the CO2 formation rate observed with 316 SSa33 and 316
SSb. Initially, CO2 is produced at about the same rate in both

Decarboxylation of HCO2X (X ) H, Na)
TABLE 2: Arrhenius Parameters for Decarboxylation of 1.00 m HCO2H and HCO2Na, and 1.07 m Malonic Acida
HCO2H HCO2Na HO2CCH2CO2H HO2CCH2CO2Na
Ea, kJ/mol Ea, kJ/mol Ea, kJ/mol Ea, kJ/mol
ln(A, s-1 ln(A, s-1 ln(A, s-1 ln(A, s-1
117 ( 5 28.3 ( 1.5
J. Phys. Chem. A, Vol. 102, No. 29, 1998 5889
cell type
)
)
)
)
316 SSa
316 SSb
Pt/Ir
113 ( 10
107 ( 12
120 ( 23
82 ( 12
19.6 ( 2.2
17.9 ( 2.6
21.0 ( 4.6
11.2 ( 2.4
100 ( 1
16.1 ( 0.2
2.3 ( 1.5
98 ( 5
24.0 ( 1.3
21.4 ( 7.9
21.7 ( 1.2
21.4 ( 2.1
90 ( 31
91.5 ( 5
93 ( 8
31.4 ( 7.5
Ti
a Errors are at 90% confidence interval.
composition by ICP-MS. They had a somewhat different usage
history but no different from any of the other compounds we
have studied where reproducible results have been
obtained.33-39,48,49 We searched unsuccessfully for several years
to explain the unusual behavior of the 316 SSa cell but feel
obliged to report the results. It is interesting to note that the
behavior of 316 SSa is similar to the pattern of CO2 production
at 50-125 °C when formic acid is adsorbed on a clean Ni-
(110) at about 10-10 Torr.14 The explanation at these ultrahigh
vacuum (UHV) conditions is that the initial first-order stage
results from normal catalytic decomposition of HCO2H on a
metal surface.12 As effective complete coverage of the surface
-
by formate develops, the HCO2 ions impede decomposition
of one another and the CO2 production rate is slowed. This
event is responsible for the quiescent period. The “explosive”
stage results when “islands” of bare metal appear and catalyze
the reaction very rapidly. There is no evidence of a formic
anhydride intermediate.50 It would be amazing if the same
mechanism were to apply also to the surface exposed to fluid
H2O at high pressure and temperature, but the CO2 production
profile is similar. Although the surfaces of the 316 SSa and
316 SSb cells were not visually different after usage, a feature
of decarboxylation of HCO2H that differs from other acids
studied in our work is the production of H2. Possibly diffusion
of H2 through the steel surface is a factor. The 304 SS cell
was severely corroded by HCO2H but not by malonic acid,
which liberates no H2 upon decarboxylation.
Figure 5. Arrhenius plots for hydrothermal decarboxylation of HCO2H
and HCO2Na (eqs 5 and 6) in cells of the same design but different
materials of construction. The Arrhenius constants are given in Table
2.
Figure 2 does not prove that reaction 1 is solely wall-catalyzed
and does not occur homogeneously. Rather, Figure 2 indicates
that the decomposition of HCO2H as measured by the formation
of CO2 is affected by the metal surface. In fact, the global rate
of decomposition is likely to be a mixture of wall-catalyzed
and homogeneous (perhaps H2O-catalyzed) rates. In the present
study, wall catalysis is a major factor. Another indication about
the process comes from the determination of the kinetic
constants, which is discussed next.
Kinetic constants were extracted from the rate of formation
of CO2 at e40% conversion. This condition enables a plug-
flow model of the reactor to be used.35,51 Because of the use
of relatively low extents of conversion, however, the order of
the process is not firmly established. Indeed, zeroth-, first-,
and second-order rate plots for decomposition of HCO2H and
HCO2Na were all found to be relatively linear. First-order
behavior has been indicated previously for the decomposition
of HCO2H in H2O,11 as well as for acetic acid,52 and will,
therefore, be adopted.
Figure 6. Arrhenius plots for hydrothermal decarboxylation of malonic
acid (HO2CCH2CO2H) at 275 bar in the same cells used for HCO2H
(see Figure 5). Table 2 gives the Arrhenius parameters.
residence times (t); [HCO2-] (0.19% at 290 °C, 0.09% at 330
°C) was obtained by an iso-Coulombic fit of the dissociation
constant of HCO2H,53,54 and k2 was determined from parallel
experiments on the decarboxylation of 1.00 m HCO2Na. The
same procedure was used to study HCO2Na as was used for
HCO2H except that its decarboxylation rate was too slow to
measure in the Ti cell. Figure 4 is a first-order rate plot for the
disappearance of HCO2H constructed from eq 7 at several
temperatures in the 316 SSb cell. Table 1 summarizes k1 and
k2 as a function of temperature and cell type. The statistical
analysis procedure was described in the Experimental Section.
It is noteworthy that the rate of decarboxylation of HCO2Na
(eq 6) in Table 1 is slower than that of HCO2H (eq 5). The
only exception is the rate in the Pt/Ir cell below 305 °C.
Otherwise, the general trend of k1 > k2 is in line with findings
for malonic acid38 and acetic acid52 and may be attributed to
the additional resonance stabilization energy in the anion.
A conventional first-order kinetic analysis by eq 7 affords k1
of reaction 5.
-
[CO2]t - [HCO2 ]0[1 - exp(-k2t)
-k1t ) ln 1 -
(7)
[
]
[HCO2H]0
In reality, this reaction is probably better described as pseudo-
first-order because of the probable role of H2O in the transition
state. In eq 7, [CO2] was determined experimentally at various

5890 J. Phys. Chem. A, Vol. 102, No. 29, 1998
Maiella and Brill
The rate constants in Table 1 afford the Arrhenius plots shown
in Figure 5. Table 2 contains the Arrhenius constants obtained
by weighted least-squares regressions. Ea and ln A for decar-
boxylation of HCO2H in Pt/Ir and the initial stage in the SS
cells are the same within the experimental error. On the other
hand, those for HCO2H in Ti are statistically different. Like-
wise, the Arrhenius parameters for HCO2Na are different or
close to being statistically different from those of HCO2H. The
general range of Ea values of 83-121 kJ/mol for HCO2H
resembles the surface-catalyzed values of 100-109 kJ/mol,14,21
and 86-113 kJ/mol when HCO2H is decomposed in the
presence of H2O.22,23 The range of Ea in Table 2 also includes
the calculated value of 90 kJ/mol when two units of H2O are
incorporated in the transition state to catalyze the decarboxy-
lation reaction.8 These Ea values are much lower than those
found for unimolecular decomposition, which are reportedly
greater than 200 kJ/mol25 and more typically greater than 250
kJ/mol.8,26,29 Hence, the Arrhenius parameters for HCO2H in
the hydrothermal medium are consistent with some form of
catalyzed decarboxylation.
References and Notes
(1) Klatt, M.; Ro¨hrig, M.; Wagner, H. Gg. Z. Naturforsch. 1992, 47A,
1138.
(2) Ross, D. S.; Hum, G. P.; Miin, T.; Green, T. K.; Mansami, R. In
Supercritical FluidssChemical and Engineering Principles and Applica-
tions; Squires, T. G., Paulaitis, M. F., Eds.; ACS Symposium Series 339;
American Chemical Society: Washington, DC, 1987; p 242.
(3) Penninger, J. M. L. Fuel 1989, 68, 983.
(4) Margolis, L. Ya. J. Catal. 1971, 21, 93.
(5) Elliott, D. C.; Hallen, R. T.; Sealock, L. J., Jr. Ind. Eng. Chem.
Prod. Res. DeV. 1983, 22, 431.
(6) Helling, R. K.; Tester, J. W. Energy Fuels 1987, 1, 417.
(7) Helling, R. K.; Tester, J. W. EnViron. Sci. Technol. 1988, 22, 1319.
(8) Melius, C. F.; Bergan, N. E.; Shepherd, J. E. Twenty-Third
Symposium (International) on Combustion; The Combustion Institute:
Pittsburgh, PA, 1990; p 217.
(9) Hirth, Th.; Franck, E. U. Ber. Bunsen-Ges. Phys. Chem. 1993, 97,
1091.
(10) Rice, S. F.; Steeper, R. R.; Aiken, J. D. J. Phys. Chem., in press.
(11) Bjerre, A. B.; Sorensen, E. Ind. Eng. Chem. Res. 1992, 31, 1574.
(12) Trillo, J. M.; Munuera, G.; Criado, J. M. Catal. ReV. 1972, 7, 51.
(13) McCarty, J.; Falconer, J.; Madix, R. J. J. Catal. 1973, 30, 235.
(14) Falconer, J.; Madix, R. J. Surf. Sci. 1974, 46, 473.
(15) Benziger, J. B.; Madix, R. J. Surf. Sci. 1979, 79, 394.
(16) Barteau, M. A.; Bowker, M.; Madix, R. J. Surf. Sci. 1980, 94, 303.
(17) Benziger, J. B.; Schoofs, G. R. J. Phys. Chem. 1984, 88, 4439.
(18) Shustorovich, E.; Bell, A. T. Surf. Sci. 1989, 222, 371.
(19) Sekine, I.; Senoo, K. Corros. Sci. 1984, 24, 439.
(20) Sekine, I.; Hatakeyama, S.; Nakazawa, Y. Corros. Sci. 1987, 27,
275.
(21) Giner, J.; Rissmann, E. J. Catal. 1967, 9, 115.
(22) Barham, H. N.; Clark, L. W. J. Am. Chem. Soc. 1951, 73, 4638.
(23) Yu, J.; Savage, P. E. Ind. Eng. Chem. Res. 1998, 37, 2.
(24) Blake, P. G.; Hinshelwood, C. Proc. R. Soc. London, Ser. A 1960,
255, 444.
A third comparison of data that support the contribution of
wall catalysis in the decarboxylation of HCO2H is with the
corresponding Arrhenius constants for decarboxylation of ma-
lonic acid by eq 8. In contrast to monocarboxylic acids,
dicarboxylic acids have been suggested to decarboxylate mainly
without catalysis.55 Hence, the rates of
HO2CCH2CO2H f CH3CO2H + CO2
(8)
(25) Blake, P. G.; Davies, H. H.; Jackson, G. E. J. Chem. Soc. B 1971,
1923.
(26) Saito, K.; Kakumoto, T.; Kuroda, H.; Torii, S.; Imamura, A. J.
Chem. Phys. 1984, 80, 4989.
(27) Ruelle, P.; Kesselring, U. W.; Nam-Tran, H. J. Am. Chem. Soc.
decarboxylation of malonic acid should be independent of the
materials used to construct the cell. Although most of the
experimental data for malonic acid in Table 2 were reported
before,38 a less sophisticated statistical analysis was used
previously, which affects the Arrhenius parameters. The values
in Table 2 were recalculated using the same procedure employed
for formic acid. Figure 6 shows the resulting Arrhenius plots
for malonic acid in the same cells used to study formic acid.
Although small differences exist in the rates, the Arrhenius
parameters lie within the uncertainty limits. The first-order rate
of decomposition is, however, much faster than that of formic
acid because of the higher preexponential factor and the slightly
lower average values of the activation energy. Unlike formic
acid, whose decarboxylation reaction appears to be catalyzed
by H2O and/or the reactor wall, malonic acid can form an
internal, six-membered, cyclic transition state and thereby
decarboxylate unimolecularly.38,55 We have also investigated
the decarboxylation of five additional monocarboxylic acids in
the SS and Ti cells and find that the rates are essentially the
same in these two materials of construction.56
1986, 108, 371.
(28) Goddard, J. D.; Yamaguchi, Y.; Schaefer, H. F., III. J. Chem. Phys.
1992, 96, 1158.
(29) Francisco, J. S. J. Chem. Phys. 1992, 96, 1167.
(30) Hsu, D. S. Y.; Shank, W. M.; Blackburn, M.; Lin, M. C. Nineteenth
Symposium (International) on Combustion; The Combustion Institute:
Pittsburgh, PA, 1983; p 89.
(31) Ramsonov, F. N.; Petrov, A. K.; Baklanov, A. V.; Vihzin, V. V.
React. Kinet. Catal. Lett. 1976, 5, 197.
(32) Akiya, N.; Savage, P. E. AIChE J. 1998, 44, 405.
(33) Brill, T. B.; Schoppelrei, J. W.; Maiella, P. G.; Kieke, M. L.; Belsky,
A. J. Proceedings of the 2nd International Conference on SolVothermal
Reactions; Committee of Solvothermal Technology Research: Takamatsu,
Japan, 1996; 5.
(34) Kieke, M. L.; Schoppelrei, J. W.; Brill, T. B. J. Phys. Chem. 1996,
100, 7455.
(35) Schoppelrei, J. W.; Kieke, M. L.; Wang, X.; Klein, M. T.; Brill, T.
B. J. Phys. Chem. 1996, 100, 14343.
(36) Schoppelrei, J. W.; Kieke, M. L.; Brill, T. B. J. Phys. Chem. 1996,
100, 7463.
(37) Belsky, A. J.; Brill, T. B. J. Phys. Chem. A 1998, 102, 4509.
(38) Maiella, P. G.; Brill, T. B. J. Phys. Chem. 1996, 100, 14352.
(39) Lundegard, P. D.; Kharaka, Y. K. ACS Symposium Series; American
Chemical Society: Washington, DC.; 1990; Vol. 416, p 169.
(40) Shock, E. L. Geochim. Cosmochim. Acta 1993, 97, 3341.
(41) Katritzky, A. R.; Allin, S. M.; Siskin, M. Acc. Chem. Res. 1996,
29, 399.
(42) Savage, P. E.; Gopalan, S.; Mizan, T. I.; Martino, C. J.; Brock, E.
E. AIChE J. 1995, 41, 1723.
(43) Tester, J. W.; Holgate, H. R.; Armellini, F. J.; Webley, P. A.;
Killilea, W. R.; Hong, G. T.; Barner, H. E. ACS Symposium Series; American
Chemical Society: Washington, DC, 1993; Vol. 518, p 35.
(44) Ding, Z. Y.; Frisch, M. A.; Li, L.; Gloyna, E. F. Ind. Eng. Chem.
Res. 1996, 35, 3257.
In summary, the dependence of reaction 1 on the metal used
in cell construction, the magnitude of its activation energy, and
the comparison with noncatalyzed decarboxylation of malonic
acid all suggest that decarboxylation of HCO2H is catalyzed in
these experiments. It is logical that catalysis in the hydrothermal
medium involves both H2O and the reactor wall in a ratio that
varies with the material of construction, S/V, and perhaps the
pressure and temperature. Unfortunately, experimental limita-
tions at hydrothermal conditions make it difficult to probe the
mechanism of surface catalysis further, but plans are underway
in our laboratory to gain more fundamental insight into catalyzed
reactions in the hydrothermal medium.
(45) Maiella, P. G.; Brill, T. B. To be published.
(46) Cvetanoric, R. J.; Singleton, D. L. Int. J. Chem. Kinet. 1977, 9,
481.
Acknowledgment. We are grateful to the U.S. Army
Research Office for support of this work on DAAL03-92-G-
0174 (R. W. Shaw, program manager).
(47) Reva, I. D.; Plokhotnichenko, A. M.; Radchenko, E. D.; Sheina,
G. G.; Blagoi, Yu. P. Spectrochim. Acta 1994, 50A, 1107.
(48) Schoppelrei, J. W.; Brill, T. B. J. Phys. Chem. A. 1997, 101, 2298.

Decarboxylation of HCO2X (X ) H, Na)
J. Phys. Chem. A, Vol. 102, No. 29, 1998 5891
(49) Schoppelrei, J. W.; Brill, T. B. J. Phys. Chem. A. 1997, 101,
8593.
(50) Avery, N. R.; Toby, B. H.; Anton, A. B.; Weinberg, W. H. Surf.
Sci. 1982, 122, L574.
(53) Kortum, G.; Vogel, W.; Andrussow, K. Dissociation Constants of
Organic Acids in Aqueous Solutions; Butterworths: London, 1961.
(54) Bell, J. L. S.; Wesolowski, D. J.; Palmer, D. A. J. Solution Chem.
1993, 22, 125.
(51) Cutler, A. H.; Antal, M. J., Jr.; Jones, M., Jr. Ind. Eng. Chem. Res.
1988, 27, 691.
(52) Bell, J. L. S.; Palmer, D. A.; Barnes, H. L.; Drummond, S. E.
Geochim. Cosmochim. Acta 1994, 58, 4155.
(55) Bell, J. L. S.; Palmer, D. A. In Organic Acids in Geological
Processes; Pittman, E. D., Lewan, M. D., Eds.; Springer-Verlag: Heidelberg,
1994; p 226.
(56) Belsky, A. J.; Brill, T. B. To be published.