Use of completely mixed flow-through systems: Hydrolysis of phenyl picolinate

Article abstract of DOI:10.1021/es970440z

Experimental equipment was developed to study the hydrolysis of environmentally relevant compounds in completely mixed flow-through reactors. The hydrolysis of a carboxylic acid ester, phenyl picolinate (PhP), was investigated in homogeneous solutions and in the presence of titanium dioxide particles. In the absence of solids, the hydrolysis of PhP is dominated by the acid-catalyzed reaction below pH 4 and by the base-catalyzed reaction above pH 5, while the neutral hydrolysis is of minor importance across the whole pH range investigated (22 L^-1 (≈420 m² surface area L^-1), the neutral, surface-catalyzed reaction dominated the overall hydrolysis process in the entire range from pH 2 to pH 9. Steady-state rates of the heterogeneous hydrolysis were about 1 order of magnitude slower in completely mixed flow-through systems than initial hydrolysis rates in batch systems. This was attributed to reaction products occupying reactive sites at the mineral surface, thus hindering the adsorption of the educt PhP. Experimental equipment was developed to study the hydrolysis of environmentally relevant compounds in completely mixed flow-through reactors. The hydrolysis of a carboxylic acid ester, phenyl picolinate (PhP), was investigated in homogeneous solutions and in the presence of titanium dioxide particles. In the absence of solids, the hydrolysis of PhP is dominated by the acid-catalyzed reaction below pH 4 and by the base-catalyzed reaction above pH 5, while the neutral hydrolysis is of minor importance across the whole pH range investigated (2 < pH < 8). However, in presence of 10 g of TiO₂ L^-1 (?420 m² surface area L^-1), the neutral, surface-catalyzed reaction dominated the overall hydrolysis process in the entire range from pH 2 to pH 9. Steady-state rates of the heterogeneous hydrolysis were about 1 order of magnitude slower in completely mixed flow-through systems than initial hydrolysis rates in batch systems. This was attributed to reaction products occupying reactive sites at the mineral surface, thus hindering the adsorption of the educt PhP.

Full text of DOI:10.1021/es970440z

Environ. Sci. Technol. 1997, 31, 3692-3701
influence of such chelate-induced catalysis may be most
Use of Completely Mixed
Flow-Through Systems: Hydrolysis
of Phenyl Picolinate
pronounced in the pH range where the half-life in homo-
geneous solutions is longest. In most environmental systems,
however, the concentration of free metal ions is probably too
low to play an important role in degradation of organic
compounds (1, 5). Mineral surfaces, however, are present in
such abundance that they may exert a significant catalytic
effect.
Hydrolysis reactions of organic compounds relevant in
the environment have usually been studied in batch reactors
measuring initial rates. Advantages of batch methods are
that they generally require only low-cost equipment and are
easy to use. They are of limited use for the investigation of
long-term effects that become manifest in the time span
exceeding the instant reaction phenomena observed at the
beginning of an experiment. One possible long-term effect
is the inhibition of a surface-catalytic process due to adsorp-
tion of reaction products on the reactive sites of the mineral.
This effect may be difficult to observe in batch reactors, unless
reaction products are added at the beginning of an experiment
or the reaction solution is repeatedly spiked with the reaction
educt. The determination of initial rates from batch experi-
ments is often difficult, particularly if the reaction mechanism
is complex or changes with time.
We have therefore developed an alternative method for
the investigation of hydrolytic degradation of pesticides in
model soil systems. The experimental setup is based on the
completely mixed flow-through reactors described by Furrer
et al. (6). For convenience, we subsequently replace the term
“completely mixed flow-through reactor” by “mixed flow
reactor” as it is used in chemical engineering (7). If the
material fluxes across the boundaries of this type of reactor
remain constant, steady state may eventually be reached.
Steady-state data can be analyzed using the computer
program SteadyFit, which allows the simultaneous determi-
nation of kinetic and thermodynamic constants (8). One of
the advantages of using mixed flow reactors is that reaction
rates can be determined directly from the concentrations of
the reactants in the outflow. Furthermore, inhibition by co-
adsorbed species can be more easily observed in mixed flow
reactors than in batch reactors (9).
M I C H A E L G F E L L E R , *
G E R H A R D F U R R E R , A N D
R A I N E R S C H U L I N
Institute of Terrestrial Ecology, Swiss Federal Institute
of Technology (ETHZ), Grabenstrasse 3,
CH-8952 Schlieren, Switzerland
Experimental equipment was developed to study the hy-
drolysis of environmentally relevant compounds in completely
mixed flow-through reactors. The hydrolysis of a carboxylic
acid ester, phenyl picolinate (PhP), was investigated in
homogeneous solutions and in the presence of titanium dioxide
particles. In the absence of solids, the hydrolysis of PhP
is dominated by the acid-catalyzed reaction below pH 4
and by the base-catalyzed reaction above pH 5, while the
neutral hydrolysis is of minor importance across the whole
pH range investigated (2 < pH < 8 ). However, in
presence of 10 g of TiO L-1 (=420 m2 surface area L-1 ),
2
the neutral, surface-catalyzed reaction dominated the
overall hydrolysis process in the entire range from pH 2 to
pH 9. Steady-state rates of the heterogeneous hydrolysis
were about 1 order of magnitude slower in completely mixed
flow-through systems than initial hydrolysis rates in batch
systems. This was attributed to reaction products
occupying reactive sites at the mineral surface, thus
hindering the adsorption of the educt PhP.
1. Introduction
Many widely used pesticides are carboxylic acid esters,
phosphoric acid esters, and carbamates. Under certain pH
conditions, they readily hydrolyze in water. Among other
transformation reactions, hydrolysis is important for the
mineralization of pesticide compounds in the environment.
Dissolved metal ions and mineral surfaces are known to
increase hydrolysis rates of organic esters (1, 2). Dissolved
metal ions or surface metal centers may catalyze this reaction
by direct polarization of chemical bonds in the organic
compound. In both cases, the water molecule with its dipole
character may be sufficiently nucleophilic to induce the
hydrolysis reaction. If the organic ester and the metal ion
form bidentate complexes, polarization may be enhanced as
compared to the monodentate complex. Fife and Przystas
(3) have investigated the influence of divalent metals (Ni2+
and Cu2+) on the hydrolysis rate of picolinate esters. They
have found substantial rate enhancement if the pyridine
nitrogen is in ortho position to the ester function, leading to
the formation of a chelate with the dissolved metal ion. A
surface-catalytic effect on the hydrolysis of the ester phenyl
picolinate (PhP; Figure 1) has been reported by Torrents and
Stone (4) in a heterogeneous systems containing TiO2 or
goethite (FeOOH). In other words, metal ion and surface
catalysis may lead to an increase in reaction rates in a broad
pH range, depending on the speciation of the chelate. The
The setup has been tested by re-analyzing the surface-
catalyzed hydrolysis of the carboxylic acid ester phenyl
picolinate (PhP) in the presence of TiO
2, which has been
extensively studied in batch systems by Torrents and Stone
(4, 10, 11). The hydrolysis of PhP occurs via homogeneous
as well as heterogeneous parallel reactions as shown in Figure
1. Hydrolysis products are picolinic acid and phenol. Figure
2 shows the structures of all organic compounds used in the
experiments, including protonated and deprotonated species.
Experiments have been performed in homogeneous systems
for a direct comparison of hydrolysis rates with literature
data. Heterogeneous experiments have been conducted to
study the differences between initial reaction rates obtained
from batch experiments and steady-state rates obtained using
mixed flow reactors.
2. Materials and Methods
Phenyl picolinate was synthesized according to Bald (12).
TiO2 (type P25) was provided by Degussa Corp. and used
without further treatment. It consisted of about 40% rutile
and 60% anatase (R. Giovanoli, 1992, personal communica-
tion). All solutions and suspensions were prepared with
purified water with a resistivity of 18 MΩ cm (Barnstead
NANOpure). Argon (99.998%) was obtained from PanGas.
All other chemicals were of analytical grade or better and
were obtained from either Merck or Fluka. All experiments
* Corresponding author phone: +41 1 633 60 13; fax: +41 1 633
11 23; e-mail: gfeller@ito.umnw.ethz.ch.
9
3 6 9 2 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 12, 1997
S0013-936X(97)00440-9 CCC: $14.00
1997 Am erican Chem ical Society

FIGURE 1. Reaction scheme of the hydrolysis of phenyl picolinate in heterogeneous system with TiO . No protonation equilibria are shown.
2
While in the homogeneous solution, the hydrolysis reaction occurs predominantly as H+ or OH- catalyzed (the neutral reaction is much
less important); the surface-catalyzed reaction is not pH dependent in the investigated pH range.
were performed at 25 °C and an ionic strength of 0.1 (adjusted
with NaCl). For the experiments with TiO2, reaction solutions
were protected against light by black plastic foil to prevent
photochemical reactions. The pH electrodes were calibrated
by performing acid-base titrations in the absence of particles
to obtain values for the parameters E0 and k of the Nernst
equation E ) E0 + k log [H+]. From the measured emf values
E, the pH values were calculated according to pH ) (E0 -
E)/ k.
2.1. Batch Experim ents. Batch systems were used for
acid-base titrations of TiO2 and to determine adsorption
curves for picolinic acid, phenol, and 2-benzoylpyridine on
TiO2.
Titration of TiO2. For each titration, 10 mL of a 0.2 M
NaCl solution was added to 10 mL of a ca. 25 g L-1 TiO2
suspension in a thermostated beaker. The suspension was
purged with Ar and stirred with a hanging Teflon propeller.
Acid (0.01 M HCl) or base (0.01 M NaOH) were added using
a computer-controlled dosimat (Metrohm 665). After each
addition, the concentration of the free hydrogen ions was
determined by emf measurements using a combined mi-
croglass electrode (Metrohm 6.0204.100) and a computer-
controlled pH meter (Metrohm 713). Equilibrium was
assumed when the standard deviation of 10 subsequent
measurements, taken every 60 s, was less than 0.3 mV. The
average value of these 10 measurements was used for further
evaluation. If the number of measurements exceeded 30
without the stability criteria being met, the data point was
discarded. Separate batches were used for the titration of
TiO2 with acid and base, with an initial concentration of 11.217
and 12.083 g L-1 TiO2, respectively.
Adsorption Curves. Adsorption curves for picolinic acid
(1.0 × 10-4 M), phenol (1.022 × 10-5 M), and 2-benzoylpyridine
(1.5 × 10-4 M) were determined separately in suspensions
containing 10.100, 10.583, and 9.967 g L-1 TiO2, respectively.
For each pH, separate flasks were used. The initial pH was
adjusted by adding either 0.1 M HCl or 0.1 M NaOH. After
purging with Ar for 1 h, the suspensions were equilibrated
during 24 h at 130 rpm on a shaker. The pH was measured
before analysis. Samples containing picolinic acid or phenol
were filtered through a 0.01-µm cellulose-nitrate filter
(Sartorius). Picolinic acid was analyzed at pH ) 1 (glycine-
hydrochloride buffer) at 270 nm using a spectrophotometer
(UVIKON 930, Kontron). Phenol was determined by reversed
phase HPLC (see below). Solutions containing 2-benzoylpy-
ridine were centrifuged at 20000 rpm (70000g) for 2 h
(Centrikon T-1170, rotor TST 28.28/ 17). The supernatant was
analyzed like the picolinic acid filtrate. Adsorption experi-
ments with 2-benzoylpyridine were conducted in duplicates,
phenol concentrations were measured after 1, 2, and 3 days.
The acid-base titrations and the adsorption experiments were
evaluated with the computer program GRFIT (13), applying
the constant capacitance surface complexation model (14).
2.2. Mixed Flow Experim ents. The reactors used have
been described by Furrer et al. (6). We used between four
and eight reactors concurrently. The experimental setup is
described in detail elsewhere (15). The pH was adjusted using
either HCl or NaOH. Phenyl picolinate and phenol were
analyzed by reversed phase HPLC (two pumps: JASCO PU-
980, UV-detector: JASCO UV-970, wavelength: 270 nm,
degasser: GAS TORR GT-130, column: LiChrospher 100 RP-
18 (Merck 50943.0001), eluent: 60% methanol and 40% acetic
acid-sodium acetate buffer (5 mM each), flow: 1 mL min-1).
HPLC peak areas were integrated using a chromatographic
data system (JCL6000, Jones Chromatography, U.K.). The
factors for the conversion of peak areas to concentrations
were determined by calibrations before and after an experi-
ment and used to obtain the conversion factors at time t by
linear interpolation.
Daily, the flow velocity of the reaction solution was
determined gravimetrically. The residence time (τ) was
calculated by dividing the reactor volume (ca. 60 mL), which
was determined gravimetrically after each experiment, by
the average flow rate of an experiment. The steady-state pH
for experimental series 1-5 (see Tables 1 and 2) was
determined by collecting about 10 mL of outflow solution
and subsequent measurement, usually several times during
an experiment. In series 6 and 7, the pH was determined
on-line as described by Furrer et al. (15). The electrodes
were calibrated as described above before and after the
experiments. The pH values were calculated from the emf
measurements by linear interpolation of the parameters E0
and k of the Nernst equation.
2.3. Treatm ent of Data from Mixed Flow Experim ents.
In the following discussion, the average of variable c (e.g.,
concentration) from data point n0 to n1 is denoted by jcn and
1
n0
9
VOL. 31, NO. 12, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 6 9 3

TABLE 1. Experimental Conditions and SteadyFit Input Data
a
for the Homogeneous System
id
v
[H+]in
[PhP]in
[H CO ]
2
3 in
1.5
2.4
2.5
3.4
3.5
3.6
5.5
5.6
5.7
5.8
6.3
6.4
7.5
7.8
6.78E-06
1.32E-05
1.29E-05
1.39E-05
1.38E-05
1.39E-05
1.38E-05
1.36E-05
1.35E-05
1.36E-05
1.50E-05
1.37E-05
1.44E-05
1.38E-05
9.87E-04
9.92E-05
9.91E-03
2.98E-03
2.98E-04
2.98E-05
1.00E-07
-9.92E-07
-9.92E-06
-9.92E-05
-9.93E-06
-9.92E-07
9.92E-03
-4.96E-05
6.72E-05
6.15E-05
6.29E-05
8.26E-05
8.32E-05
8.14E-05
8.34E-05
8.36E-05
8.48E-05
8.25E-05
7.67E-05
8.32E-05
8.24E-05
8.32E-05
1.88E-08
1.90E-08
1.89E-08
1.89E-08
1.89E-08
1.94E-08
9.75E-08
4.83E-07
4.54E-06
4.60E-05
4.54E-06
4.83E-07
1.89E-08
2.28E-05
id
[PhP]sts
σPhP
[PhH]sts
σPhH
1.5
2.4
2.5
3.4
3.5
3.6
5.5
5.6
5.7
5.8
6.3
6.4
7.5
7.8
3.80E-05
5.76E-05
2.65E-05
5.38E-05
7.69E-05
7.46E-05
7.43E-05
7.76E-05
6.89E-05
2.95E-05
7.46E-05
8.01E-05
3.64E-05
6.56E-05
1.88E-06
1.50E-06
1.06E-06
2.00E-06
1.99E-06
1.92E-06
1.66E-06
1.29E-06
9.89E-07
1.18E-06
7.39E-07
1.62E-06
1.42E-06
1.63E-06
2.64E-05
4.60E-06
3.53E-05
2.66E-05
4.23E-06
4.44E-06
1.07E-05
5.58E-06
1.57E-05
5.15E-05
3.16E-06
4.83E-06
4.51E-05
1.87E-05
1.53E-06
7.71E-07
7.35E-07
1.58E-06
1.76E-06
1.57E-06
1.14E-06
5.43E-07
5.67E-07
1.60E-06
7.67E-07
1.42E-06
1.67E-06
1.19E-06
a
id, experim ent identification (series num ber before the period,
experim ent num ber within
a series after the period); v, volum e-
norm alized flow (s-1); [H+]in, concentration of H+ in the inflow; [PhP]in,
concentration of PhP in the inflow; [H2CO3]in, concentration of H2CO3
in the inflow; [PhP]sts ( σPhP and [PhH]sts ( σPhH, steady-state concentration
of PhP and PhH (with error). All concentrations are in m ol L-1
.
The absolute errors of PhP (RPhP) and PhH (RPhH) were set
equal to the standard deviations of [PhP] and [PhH] in the
steady-state range from data points n0 to n1:
FIGURE 2. Structures and abbreviations of all organic species.
Dioxane was used as cosolvent.
RPhP ) [PhP]nn
(3)
(4)
its standard deviation by cnn1. The term “data point” refers to
1
0
0
a pair of concentrations of phenyl picolinate (PhP) and phenol
(PhH). In this section, the notations [PhP] and [PhH] refer
to the total dissolved concentration of these compounds.
Rem oval of Outliers. A data point u (u > 1) was removed
before the determination of the steady state if at least one of
the conditions
RPhH ) [PhH]nn
1
0
The absolute errors were used as a first criterion to determine
steady state because the spread about the average concen-
trations is large if (1) concentrations steadily increase or
decrease (i.e., the rate of change is large) and (2) measure-
ments deviate from a horizontal course of concentrations.
Since the absolute error for PhP and PhH must be simulta-
neously small, it was required that
u+1
|[PhP]u - [PhP] | > faPhP
(1)
u-1
or
u+1
RPhP e aPhP
RPhH e aPhH
(5)
(6)
|[PhH]u - [PhH] | > faPhH
(2)
u-1
was fulfilled, where aPhP and aPhH denote the maximal absolute
error allowed for PhP and PhH, respectively. The constant
f was set to 3. This simple algorithm sometimes caused the
removal of data in the transient state. However, this was
never crucial for the steady-state determination.
Determ ination of Steady State. The optimization of the
rate constants using SteadyFit (8) requires the determination
of the steady-state concentrations of phenyl picolinate
([PhP]sts) and phenol ([PhH]sts) and of the measurement errors
were both true. The maximum absolute errors allowed for
PhP and PhH, aPhP and aPhH, were both set to 10-6 M.
Because the mass balance must be sufficiently close to
100% at steady state, a further criterion could be applied.
Assuming that phenyl picolinate and phenol are consumed
and produced solely by hydrolysis, the sum of their con-
centrations at steady state should be equal to the concentra-
tion of phenyl picolinate in the inflow, [PhP]in:
σ
PhP and σPhH. The determination of steady-state values was
performed with the reduced data set, i.e., after outliers were
removed.
[PhP]in ) [PhP]sts + [PhH]sts
(7)
9
3 6 9 4 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 12, 1997

is fulfilled if
TABLE 2. Experimental Conditions and SteadyFit Input Data
a
for the Heterogeneous System
n1
n1
FPhPn e rPhPn
(13)
(14)
0
0
id
[TiO ]
v
[H+]in
[PhP]in
[H CO ]
2 3 in
2
Likewise, for PhH, we obtain
1.1
2.1
2.2
2.3
9.31 6.69E-06
9.27 1.27E-05
9.38 1.29E-05
9.58 1.28E-05
9.87E-03 6.71E-05 2.91E-08
9.91E-05 6.21E-05 2.95E-08
9.92E-03 6.16E-05 2.93E-08
9.91E-04 6.38E-05 2.93E-08
2.98E-03 7.87E-05 2.93E-08
2.98E-04 7.51E-05 2.94E-08
2.98E-05 8.30E-05 3.00E-08
1.00E-07 8.30E-05 1.38E-07
n1
n1
FPhHn e rPhHn
0
0
3.1 12.69 1.44E-05
3.2 12.79 1.52E-05
3.3 13.12 1.41E-05
5.1
5.3
5.4
7.4
If the conditions imposed by eqs 5, 6, 13, and 14 were met
by several data sets, then the set with the biggest number of
data points N (N ) n1 - n0 + 1) was used. N was required
to be at least 10% of the total number of data points in the
reduced data set. If this was less than 3, N was set to 3.
The steady-state concentrations of PhP and PhH were thus
defined by
7.30 1.36E-05
7.52 1.35E-05 -9.92E-06 8.56E-05 5.59E-06
7.40 1.40E-05 -9.92E-05 8.16E-05 5.45E-05
9.90 1.35E-05 -4.96E-05 8.76E-05 2.75E-05
id
[PhP]sts
σPhP
[PhH]sts
σPhH
[PhP]sts ) [PhP]nn
(15)
(16)
1
1.1
2.1
2.2
2.3
3.1
3.2
3.3
5.1
5.3
5.4
7.4
1.76E-05
3.60E-05
2.37E-05
3.59E-05
4.36E-05
5.26E-05
5.26E-05
5.03E-05
5.93E-05
3.57E-05
4.82E-05
6.28E-07
1.29E-06
1.14E-06
1.28E-06
1.74E-06
1.27E-06
2.14E-06
8.03E-07
1.24E-06
1.17E-06
2.67E-06
4.96E-05
2.68E-05
3.67E-05
2.71E-05
3.35E-05
2.19E-05
2.75E-05
3.19E-05
2.53E-05
4.52E-05
3.58E-05
1.15E-06
1.33E-06
1.50E-06
1.31E-06
1.60E-06
1.24E-06
2.13E-06
8.05E-07
6.61E-07
1.18E-06
2.24E-06
0
[PhH]sts ) [PhH]nn
1
0
and their errors by
σPhP ) FPhP[PhP]sts + RPhP
σPhH ) FPhH[PhH]sts + RPhH
(17)
(18)
aid, experim ent identification (series num ber before the period,
experim ent num ber within a series after the period); [TiO2], concentra-
tion of TiO2 (g L-1); v, volum e-norm alized flow (s-1); [H+]in, concentration
of H+ in the inflow; [PhP]in, concentration of PhP in the inflow; [H2CO3]in,
where RPhP and RPhH are defined by eqs 3 and 4, respectively,
and FPhP and FPhH are defined by
n1
concentration of H2CO3 in the inflow; [PhP]sts, ( σPhP and [PhH]sts ( σPhH
steady-state concentration of PhP and PhH (with error). All concentra-
tions are in m ol L-1
,
FPhP ) FPhPn
(19)
(20)
0
.
n1
FPhH ) FPhHn
0
We thus defined the relative error for data point u in the mass
For series 6 and 7, steady-state pH was calculated from
the data in the steady-state range determined by the algorithm
outlined above. Since the pH was not used as a dependent
variable in SteadyFit, no condition with respect to pH was
included in the steady-state determination algorithm.
balance as
[PhP]in - ([PhP]u + [PhH]u)
µu )
(8)
[PhP]in
Instead of requiring that µ was smaller than a fixed number
(m), we imposed the condition that the relative errors of both
PhP and PhH were smaller than a value derived from m,
which was set to 0.05.
3. Results and Modeling
3.1. Batch Experim ents. Titration of TiO2. Surfaces of metal
oxides are generally covered with hydroxyl groups (16).
Adsorption of H+ and OH- on hydroxylated surfaces can be
attributed to protonation and deprotonation of these surface
hydroxyl groups (16). For TiO2, they can be represented by
tTiOH, where tTi denotes a surface metal center. The
The relative error of PhP was defined as
[PhP]u - [PhP]u*
FPhP,u
)
(9)
|
|
[PhP]*
u
titration data were described assuming the mass-action eqs
29, 36, and 37 (Table 3). The surface species tTiOH, tTiOH2
+
,
where [PhP]u* is the mean of the measured PhP concentration
([PhP]u) and the concentration calculated from the measured
phenol concentration ([PhP]in - [PhH]u):
and tTiO- and the solution species H+ and OH- were used
in all models describing systems with suspended TiO2. The
equilibrium constants for the formation oftTiOH2+ (K8, Table
3) and tTiO- (K7) were optimized by fitting calculated total
concentration of protons as a function of pH against
experimental values. The capacitance value was optimized
[PhP]u + ([PhP]in - [PhH]u)
[PhP]u* )
(10)
2
concurrently and yielded 1.719 Fm-2
surface area obtained from BET measurements (41.67 m 2 g-1
. Using the specific
By combining eqs 8-10, the relative error of PhP, FPhP,u, can
be linked to the relative mass balance error, µu:
)
and the assumption that there were 2.31 sites nm-2 (17), we
calculated a surface site density of 1.60 × 10-4 mol g-1
.
-[PhP]in µu
Experimental data and fitted curve are shown in Figure 3A.
Adsorption Curves. Adsorption of picolinic acid was
modeled using the mass-action eqs 29-31, 36, 37, 40, and 41
(Table 3). Experimental data and fitted curve are shown in
Figure 3B, adsorption constants for the surface species tTiP
(K11) and tTiOH-P- (K12) are given in Table 3. Ludwig (13)
reported a similar adsorption behaviour of picolinic acid on
TiO2 supplied by Bayer (Germany). Phenol adsorbed between
20 and 30% (Figure 3C). The adsorption constant for the
surface complex tTiOH-PhH (K9, eq 38) was optimized using
all experimental values obtained after 1-3 days. The surface
FPhP,u
)
(11)
|
|
2[PhP]*
u
Substitution of µu by m, the maximal allowed relative error
for the mass balance, in eq 11 yields the maximal allowed
relative error for PhP, rPhP,u
:
-[PhP]in
m
rPhP,u
)
(12)
|
|
2[PhP]*
u
This leads to the second steady-state criterion for PhP, which
9
VOL. 31, NO. 12, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 6 9 5

this range were used to calculate the steady-state concentra-
tions of PhP and PhH (horizontal solid lines) and their error
range, [PhP]sts ( σPhP and [PhH]sts ( σPhH (horizontal dashed
lines).
3.2.1. Hom ogeneous System . Of the 20 homogeneous
experiments, 14 were accepted by the steady-state criteria
defined in section 2.3 and were used in the optimization of
the kinetic parameters. The volume-normalized flow v (v )
1/ τ ) as well as the inflow concentration of protons and PhP
([H+]in and [PhP]in) were taken as independent variables for
SteadyFit (8). The steady-state concentrations of phenyl
picolinate and phenol, [PhP]sts and [PhH]sts, and their
measurement errors, σPhP and σPhH, were used as dependent
variables for the optimizations. Note that the steady-state
pH (pHsts) was not used as a dependent variable. Table 1 lists
the experimental conditions for the homogeneous experi-
ments, i.e., the values of the independent and dependent
parameters used as input for SteadyFit.
Assuming the hydrolysis reactions
FIGURE 3. Experimental data (symbols) and fitted curves (lines) for
PhP f PH + PhH
(21)
(22)
-1
acid-base titration of TiO (≈10 g L , [H+]tot is the analytical total
2
concentration of H+, panel A), adsorption of picolinic acid (1.0 ×
PhPH+ f PH + PhH + H+
10-4 M, 10.100 g L-1 TiO , panel B), phenol (1.022 × 10-5 M, 10.583
2
g L-1 TiO , panel C), and 2-benzoylpyridine (1.5 × 10-4 M, 9.967 g
2
where PhP and PhPH+ are the phenyl picolinate species, PH
denotes picolinic acid, and PhH denotes phenol, the rate of
the homogeneous hydrolysis of PhP can be described by (4,
18)
L
-1 TiO , panel D) on TiO . All experiments were performed at room
2
2
temperature and an ionic strength of 0.1. The adsorption curve of
phenol was fitted using the concentrations measured after 1, 2, and
3 days. The adsorption curve of 2-benzoylpyridine was fitted using
the data from the duplicates (experiments 1 and 2). The equilibrium
constants were determined using the program GRFIT (13).
d[PhP]dis
-
) (kA1[H+] + kN + kB[OH-])[PhP] +
A2[H+][PhPH+] (23)
dt
k
where kA1, kN, and kB are the rate constants for the acid, neutral,
and base-catalyzed hydrolysis of PhP and kA2 is the rate
constant for the acid-catalyzed hydrolysis of the protonated
PhP species, PhPH+.
For modeling, the species considered were H+, OH-, PhP,
PhPH+, Ph-, PhH, P-, PH, PH2+, H2CO3, HCO3-, CO32-, and
the components chosen were H+, PhP, PH, PhH, and H2CO3
(see Figure 2 for structures and abbreviations of organic
species). The equilibrium constants for the formation of the
species from the components are given in Table 3. Values
for the rate constants kA1 (1.600 × 10-3 M-1 s-1) and kN (5.500
× 10-7 s-1) were taken from Torrents and Stone (4). In Figure
5, experimental and calculated log kobs values, where kobs is
the observed rate constant, are plotted as a function of pH.
The circles correspond to kobs, calculated from the measured
PhP concentration (kobs,PhP), and the squares correspond to
FIGURE 4. Total dissolved concentrations of phenyl picolinate and
phenol divided by [PhP]in vs time for the homogeneous experiment
3.4 (left) and the heterogeneous experiment 3.1 (right, 12.7 g L
-1
TiO ). Steady-state pH was 2.55 for both experiments. The concen-
2
trations of PhP in the inflow were 8.3 × 10-5 M and 7.9 × 10-5 M,
respectively. The residence time was 20 h for both experiments. The
dashed vertical lines delimit the steady-state range. The horizontal
lines are the steady-state concentrations (solid) and their error range
(dashed).
k
obs, calculated from the measured PhH concentrations
(kobs,PhH). The kobs values were calculated using
complexation of 2-benzoylpyridine was modeled analogous
to the adsorption of phenol. The equilibrium constants for
the protonation (K13, eq 42) and the adsorption (K14, eq 43)
reactions of 2-benzoylpyridine were optimized simulta-
neously. The adsorption was in the range of a few percent,
which was also observed by Torrents and Stone (4). Data
points and fitted curve are shown in Figure 3D.
[PhP]in - [PhP]dis
kobs,PhP
)
(24)
[PhP]dis
τ
and
[PhH]dis
([PhP]in - [PhH]dis)τ
kobs,PhH
)
(25)
3.2. Mixed Flow Experim ents. The experimental condi-
tions for the homogeneous and heterogeneous systems are
listed in Tables 1 and 2, respectively. In Figure 4, the total
dissolved concentrations of phenol, [PhH]dis, and phenyl
picolinate, [PhP]dis, are plotted versus time for a homogeneous
experiment (3.4, left) and a heterogeneous experiment (3.1,
right). The steady-state pH was 2.55 in both experiments.
The presence of TiO2 increased the rate of the hydrolysis
substantially. The steady-state ranges of these two experi-
ments, which were determined according to the procedure
given in section 2.3, are shown by the two vertical dashed
lines that delimit the steady-state range. Only data falling in
The first optimization was performed by assuming com-
plete absence of CO2 (i.e., without carbonic acid species). In
this case, the modeled log kobs values were substantially
smaller than experimental values above pH 6, meaning that
the value of kB was too small (Figure 5, line A).
Two modeling approaches were considered to account
for the effect of CO2 on solution pH:
In the first approach, a kinetic process was included in the
model to account for the diffusion of CO2 into and out of the
experimental system. Diffusion of CO2 was possible at the
9
3 6 9 6 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 12, 1997

FIGURE 5. Experimental (symbols) and calculated (lines) log kobs
values for the hydrolysis of phenyl picolinate in homogeneous
systems. The lines were calculated using the program SteadyFit.
The rate constants for the acid-catalyzed and neutral hydrolysis
were taken from Torrents and Stone (4). The other rate constants
were optimized using SteadyFit. The curve with pCO2 ) 10-3.5 atm
corresponds to the case where the aqueous solution in the reservoir
FIGURE 6. Calculated vs measured total dissolved concentrations
of phenyl picolinate (PhP), phenol (PhH), and pH. Calculations were
performed with SteadyFit after optimization of the homogeneous
rate constants kA2 and kB, and the heterogeneous rate constant kH.
is in equilibrium with atmospheric CO . Experiments were performed
2
at 25 °C and an ionic strength of 0.1.
The partial pressure of CO above the aqueous reservoirs was
2
aqueous solution-atmosphere boundary in the aqueous
reservoirs but also at the connecting pieces that joined
reactors, syringes, and tubing. Some tubes consisted of
polytetrafluorethylene and Tygon (Ismatec tubing), which
are not completely impermeable for CO2. The rate of the
assumed to be pCO2 ) 10-6.01 atm. Concentrations are given in mol
-1
L .
of the SteadyFit algorithm (eq 3 in ref 8) was minimal was
10-6.01 atm (Figure 5, line C). The optimized constants were
kA2 ) 1.982 × 10-3 ( 0.127 × 10-3 M-1 s-1 and kB ) 57.40 (
diffusion process, RCO , was assumed to be
2
6.24 M-1 s-1
.
in
CO2
out
CO2
RCO ) k
- k [H2CO3]
(26)
2
As Figure 5 shows, the log kobs curves calculated using the
first and the second approach are very close to each other
and describe the experimental data much better than the fit
without accounting for CO2. For comparison, the result of
a fit with pCO2 ) 10-3.5 is plotted as well (line D). All further
calculations were based on the second approach with a fixed
in
CO2
The first term, k , defines the rate of CO2 diffusion into the
system and is proportional to the partial pressure of CO2 in
in
CO2
the atmosphere:
k
) kCO pCO KH, where KH is Henry’s
2
2
constant (10-1.52 M atm-1, (19)), pCO is the partial pressure
2
of CO2 (10-3.5 atm), and kCO is the unknown diffusion
CO2 partial pressure pCO ) 10-6.01 atm.
2
2
coefficient of CO2 of the experimental system. The second
The differences between measured and calculated values
of [PhP]dis, [PhH]dis, and pH for the homogeneous experiments
were generally small (Figure 6, left column). The 45° line in
Figure 6 was included as eye-guide in each graph and
corresponds to equal measured and calculated values. Since
pH was not used as a dependent variable in the optimization
of the rate constants, the excellent agreement between
calculated and measured values indicates that the chosen
model gives a consistent description of the experimental
system. The deviation between calculated and measured pH
values in the pH range 5-7 is due to the experimental
uncertainties of pH determination caused by CO2 and
electrode response in this range.
out
CO2
term, k [H2CO3], describes the diffusion of CO2 out of the
out
CO2
system. Since k
is equal to kCO , eq 26 becomes
2
RCO ) kCO (pCO KH - [H2CO3])
(27)
2
2
2
The constant kCO2 was optimized by SteadyFit together with
kA2 and kB (kA2 ) 1.982 × 10-3 ( 0.127 × 10-3 M-1 s-1, kB
)
+
7.008 × 101 ( 0.690 × 101 M-1 s-1, kCO ) 1.257 × 10-4
2
0.1424 × 10-4 s-1). The corresponding log kobs curve is shown
in Figure 5 (line B).
In the second approach, it was assumed that CO2 entered
the system only through the aqueous reservoirs and that
purging them with Ar did not remove CO2 sufficiently. The
rest of the experimental system was considered to be closed
toward the atmosphere. For each experiment, the total
concentration of H2CO3, [H2CO3]tot, in the reservoir was
calculated with MICROQL (20, 21). [H2CO3]tot in the reservoir
corresponds to the concentration of H2CO3 in the inflow of
the reactor, [H2CO3]in. The logarithm of the partial pressure
3.2.2. Heterogeneous System . Of the 20 heterogeneous
experiments, 13 fulfilled the steady-state criteria, of which 11
were used for the determination of the rate constant for the
surface-catalyzed hydrolysis. Experiments 6.1 and 6.2 were
discarded because it was uncertain whether steady state was
reached. The experimental conditions for the heterogeneous
experiments, and hence the input data for SteadyFit, are listed
in Table 2.
pCO was varied in steps of 0.01, and for each pCO2 value, the
2
hydrolysis constants kA2 and kB were optimized with SteadyFit.
The partial pressure for which the least-squares function ø2
In the model for the surface-catalyzed hydrolysis, the
immobile species tTiOH, tTiOH2+, tTiO-, tTiP, tTiOH-
9
VOL. 31, NO. 12, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 6 9 7

P-, tTiOH-PhH, and tTiOH-PhP were added to the species
used in the homogeneous model, and the second approach
for dealing with CO2 was used. The additional immobile
components are tTiOH and EXP, the exponential term used
in the constant capacitance surface complexation model. The
equilibrium constants of the surface species and the mass-
action equations are listed in Table 3. The equilibrium
constant K10 for the adsorption of phenyl picolinate on the
surface of TiO2 was assumed to be equal to the adsorption
constant for 2-benzoylpyridine (K14).
The rate of surface-catalyzed hydrolysis can be described
by a term proportional to the concentration of the surface
complex tTiOH-PhP, kH[tTiOH-PhP] (22). Concentrations
of surface species are given in mol L-1
. Adding the term
kH[tTiOH-PhP] to eq 23 yields
d[PhP]dis
-
) (kA1[H+] + kN + kB[OH-])[PhP] +
A2[H+][PhPH+] + kH[tTiOH-PhP] (28)
dt
k
FIGURE 7. Experimental (symbols) and calculated (lines) log kobs
values for the hydrolysis of phenyl picolinate in heterogeneous
systems, and calculated curve for the homogeneous hydrolysis of
PhP. Curves have been calculated using SteadyFit after the
optimization of rate constant of the surface-catalyzed hydrolysis.
for the overall hydrolysis of phenyl picolinate in the hetero-
geneous system with suspended TiO2. Formulation of the
surface-catalyzed hydrolysis using the term kH[tTiOH-PhP]-
[H+] instead of kH[tTiOH-PhP] did not yield a satisfactory
description of the experimental data. In Figure 7, experi-
mental log kobs values and calculated log kobs curves are plotted
as a function of pH. Since the concentration of TiO2 was
generally different for the various experiments (between 7
and 14 g L-1, see Table 2), two log kobs curves were calculated:
one with 7 g L-1 and the other one with 14 g L-1 TiO2.
Because the concentration of suspended TiO in the experiments
2
-1
-1
was between 7 and 14 g L , one curve was calculated for 7 g L
and one was calculated for 14 g L-1 TiO . The partial pressure of
2
CO above the aqueous reservoirs was assumed to be pCO2 ) 10-6.01
2
atm. Experiments were performed at 25 °C and an ionic strength of
0.1.
Calculated versus measured values of [PhP]dis, [PhH]dis
,
zero, and the system of differential-algebraic equations
becomes a system of algebraic equations, which can be solved
for the free activity of the components [i] by applying a
modified Newton-Raphson algorithm (8, 23). For homo-
geneous systems, the system of equations is reduced by any
terms and equations relating to immobile components.
and pH are plotted in Figure 6 in the right column. As for
the homogeneous experiments, the agreement between
calculated and measured values was very good, except for
three pH values in the pH 6-9 range.
The model describing the hydrolysis of phenyl picolinate
in the presence of suspended TiO2 particles in mixed flow
reactors is defined by the differential eqs 44-48 (Table 4) for
the mobile components and by the mass-balance eqs 54 and
55 (Table 5) for the immobile components. After substitution
of the total dissolved concentrations ([i]dis) of each component
by eqs 49-53 (Table 5) and replacement of species concen-
tration by the respective mass-action equations (eqs 29-41,
Table 3), we obtain the system of differential-algebraic
equations needed to calculate the speciation as a function of
time. At steady state, the derivatives d[i]dis/ dt are equal to
In Figure 8, log kobs curves calculated with parameter values
determined in this work are compared with curves calculated
with parameters taken from the literature. The two curves
representing heterogeneous systems in batch and mixed flow
reactors were calculated for 10 g L-1 suspended TiO2.
Remarkably, the value of log kobs obtained from batch
experiments is about 1 order of magnitude higher than the
one obtained from the mixed flow experiments. This can be
interpreted as the result of a feedback effect caused by the
+
TABLE 3. Mass-Action Equations and Conditional Stability Constants for Formation of Chemical Species from Components H , PH,
a
PhH, PhP, BP, tTiOH, and EXP
log K
[OH-]
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
KW
K1
K2
K3
K4
K5
K6
K7
K8
[H+]-1
[PH]
[PH]
[PhH]
-13.78
-5.32
0.77
-9.77
1.9
-6.14
-16.07
-7.705
3.599
2.299
1.350
4.337
-1.735
1.942
1.350
(29)
(30)
(31)
(32)
(33)
(34)
(35)
(36)
(37)
(38)
(39)
(40)
(41)
(42)
(43)
[P-]
[H+]-1
[H+]
+
[PH2
]
[Ph-]
[H+]-1
[H+]
[PhPH+]
[PhP]
[H+]-1
[H+]-2
[H+]-1
[H+]
[PhH]
[PhP]
[PH]
-
[HCO3
]
[H2CO3]
[H2CO3]
[tTiOH]
[tTiOH]
[tTiOH]
[tTiOH]
[tTiOH]
[tTiOH]
[BP]
2-
[CO3
]
[tTiO-]
[tTiOH2
[EXP]-1
[EXP]
+
]
[tTiOH-PhH]
[tTiOH-PhP]
[tTiP]
K9
K10
K11
K12
K13
K14
[tTiOH-P-]
[PH]
[H+]-1
[EXP]-1
[BPH+]
[H+]
[tTiOH-BP]
[tTiOH]
[BP]
a
EXP is the exponential term used in the constant-capacitance m odel. Values for the stability constants KW, K3, K5, and K6 were taken from ref
25, for K1 and K2 from ref 26, and for K4 from ref 18. Values for K7, K8, K9, K11, K12, K13, and K14 were determ ined using the program GRFIT (13). K10
was estim ated from 2-benzoylpyridine adsorption experim ents (this work).
9
3 6 9 8 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 12, 1997

TABLE 4. Rate of Change of Total Dissolved Concentration (dis) of Each Mobile Componenta
d[H+]dis
)
)
)
)
)
v[H+]in - v[H+]dis
(44)
(45)
(46)
(47)
(48)
dt
-v[PhH]dis + kA1[H+][PhP] + kA2[H+][PhPH+] + kN[PhP] + kB[OH-][PhP] + kH[tTiOH-PhP]
v[PhP]in - v[PhP]dis - kA1[H+][PhP] - kA2[H+][PhPH+] - kN[PhP] - kB[OH-][PhP] - kH[tTiOH-PhP]
-v[PH]dis + kA1[H+][PhP] + kA2[H+][PhPH+] + kN[PhP] + kB[OH-][PhP] + kH[tTiOH-PhP]
v[H2CO3]in - v[H2CO3]dis
d[PhH]dis
dt
d[PhP]dis
dt
d[PH]dis
dt
d[H2CO3]dis
dt
a
At steady state, the derivatives are zero, yielding the flux-balance equations for each com ponent.
TABLE 5. Total Dissolved Concentrations of Mobile Components (dis) and Total Concentrations of Immobile Components (tot)a
[H+]dis
[PH]dis
[PhP]dis
[PhH]dis
[H2CO3]dis
[tTiOH]tot
[EXP]tot
)
)
)
)
)
)
)
[H+] - [OH-] - [P-] + [PH2+] - [Ph-] + [PhPH+] - [HCO3-] - 2[CO3
]
(49)
(50)
(51)
(52)
(53)
(54)
(55)
2-
[P-] + [PH] + [PH2
[PhP] + [PhPH+]
[Ph-] + [PhH]
]
+
[H2CO3] + [HCO3-] + [CO3
]
-2
[tTiOH] + [tTiOH2+] + [tTiO-] + [tTiP] + [tTiOH-P-] + [tTiOH-PhH] + [tTiOH-PhP]
-[tTiO-] + [tTiOH2+] - [tTiOH-P-]
a
[tTiOH]tot was a known experim ental quantity, and [EXP]tot was calculated assum ing the constant-capacitance m odel (14).
FIGURE 9. Contribution of the individual rates to the overall rate for
the hydrolytic degradation of PhP for the homogeneous system (left)
and the heterogeneous system (right, 10 g L-1 TiO ) in mixed flow
2
reactors. In the heterogeneous system, the surface-catalyzed
hydrolysis is dominant across the whole pH range. The partial
pressure of CO is fixed at 10-6.01 atm. R ) (kA1[H+] + kN
+
2
all
FIGURE 8. Calculated log kobs curves for the hydrolysis of phenyl
picolinate in homogeneous and heterogeneous systems. The
temperature was 25 °C except where otherwise noted. The curve
for homogeneous hydrolysis at 25 °C was calculated with SteadyFit
using the rate constants for acid-catalyzed and neutral reaction
published by Torrents and Stone (4) and the rate constants kA2 and
kB determined in this work. Homogeneous hydrolysis is assumed to
be identical in batch and mixed flow reactors. The partial pressure
kB[OH-])[PhP] + kA2[H+][PhPH+] + kH[tTiOH-PhP], used without
the term kH[tTiOH-PhP] in the calculation for the homogeneous
system. RA1 ) kA1[H+][PhP], RA2 ) kA2[H+][PhPH+], RN ) kN[PhP], RB
) kB[OH-][PhP], RH ) kH [tTiOH-PhP].
catalytic process (Figure 9). The other two curves have been
calculated based on results published by Felton and Bruice
(19). They measured the kinetics of the hydrolytic degradation
of PhP at an ionic strength of I ) 1.0 at T ) 30 °C and at T
)55 °C. Even though the results cannot be compared directly,
they do not appear to be contradictory.
of CO above the aqueous reservoir solutions was assumed to be
2
pCO ) 10-6.01 atm for the homogeneous as well as the heterogeneous
2
mixed flow system. The curves for the heterogeneous hydrolysis
were calculated for 10 g L-1 TiO . The rate constant for the surface-
2
catalyzed hydrolysis in the batch system was taken from Torrents
and Stone (4), and for the mixed flow system, the constant determined
in this work was used. Rate constants for the hydrolysis of PhP in
homogeneous solutions at 30 °C (base-catalyzed hydrolysis only)
and 55 °C were published by Felton and Bruice (18) for an ionic
strength of 1.0.
3.2.3. Validation: Calculation of Transient State. The
model describing the hydrolytic degradation of PhP in
homogeneous and heterogeneous systems was tested by
comparing calculated concentration-time series leading to
steady state (the transient state) with experimental data. For
this purpose, the system of differential-algebraic equations
(Tables 4 and 5) was solved for each experiment using the
double precision differential-algebraic sensitivity analysis
code DDASAC (24). Figure 10 shows the DDASAC results
together with measured data for experiments 3.1 and 3.4.
The calculated curve of the phenol (PhH) concentration for
adsorption of products on the surface of TiO2, which
consequently blocks the reactive sites for the surface-catalyzed
hydrolysis of the ester. Obviously, in the presence of TiO2,
the hydrolytic degradation of PhP is dominated by the surface-
9
VOL. 31, NO. 12, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 6 9 9

reaction products are involved in reactions influencing the
transformation rate of the educt, the steady-state approach
presented in this work appears to be the appropriate method
for the investigation of chemical kinetics. This method may
thus help to understand and predict the long-term fate of
pollutants in the environment.
Acknowledgments
Some of the results reported here were obtained by use of
GREGPAK, developed at the University of Wisconsin under
the supervision of Warren E. Stewart. Helen Scha¨rli performed
all batch experiments. We thank Martin Keller, Hanspeter
La¨ser, and Philippe Pe´risset for their contribution to the
experimental setup and Rene´ Schwarzenbach and Jo¨ rg
Klausen for valuable comments. This work was supported by
the ETH Research Grant 41-2713.5.
FIGURE 10. Experimental data (symbols) and calculated curves (lines)
of the normalized concentrations of phenyl picolinate (PhP) and
phenol (PhH) and their sum for experiments 3.4 (left) and 3.1 (right).
The lines have been calculated using the double precision
differential-algebraic sensitivity analysis code DDASAC (24).
the homogeneous system (experiment 3.4) is in excellent
agreement with the measurements, while the calculated
concentrations of phenyl picolinate (PhP) at the beginning
of the experiment are higher than the measured values. This
could be explained by sorption of PhP on the hydrophobic
tubing and reactor walls, which may gradually become
saturated with the organic compound.
Literature Cited
For the initial period of the heterogeneous experiment
(experiment 3.1), the agreement between the calculated and
observed concentrations of PhP and PhH is not as good as
for later times, but the sum of the concentrations is predicted
well. It should be noted that the initial concentration of
H2CO3, and thus of H+, in the reactor as well as the kinetics
of the CO2 outgassing from the reactor are unknown. In
general, the agreement between calculated and measured
concentrations as well as the sum of the concentrations was
satisfactory.
(1) Mabey, W.; Mill, T. J. Phys. Chem. Ref. Data 1978, 7, 383-414.
(2) Stone, A. T.; Torrents, A. In Environmental Impact of Soil
Component Interactions; Huang, P., Ed.; Lewis: Chelsea, MI,
1992; pp 275-298.
(3) Fife, T.; Przystas, T. J. Am. Chem. Soc. 1985, 107, 1041-1047.
(4) Torrents, A.; Stone, A. T. Environ. Sci. Technol. 1991, 25, 143-
149.
(5) Mill, T.; Mabey, W. In Reactions and Processes; Hutzinger, O.,
Ed.; Springer: Berlin, 1988; pp 71-111.
(6) Furrer, G.; Zysset, M.; Schindler, P. In Geochemistry of Clay-pore
Fluid Interactions; Manning, D., Hall, P., Hughes, C., Eds.;
Chapman & Hall: London, 1993; pp 243-262.
In summary, we can draw the following conclusions. In
order to study surface-catalyzed hydrolysis of pesticides in
open systems at steady state, an experimental setup using
completely mixed flow-through reactors was developed.
Steady-state data were analyzed using the computer program
SteadyFit (8). The experimental setup and the application of
SteadyFit to real experimental data were tested by investigat-
ing the hydrolytic degradation of phenyl picolinate and
comparing the results with data available in the literature.
(7) Levenspiel, O. Chemical Reaction Engeneering; Wiley: New York,
1972.
(8) Gfeller, M.; Furrer, G.; Schindler, P. Geochim. Cosmochim. Acta
1993, 57, 685-695.
(9) Gfeller, M.; Furrer, G. In Transport and Reactive Processes in
Aquifers; Proceedings of the IAHR/ AIRH Symposium on Trans-
port and reactive processes in aquifers, Zu¨ rich, Switzerland, April
11-15, 1994; Dracos, T., Stauffer, F., Eds.; Balkema: Rotterdam,
1994; pp 59-65.
(10) Torrents, A.; Stone, A. T. Environ. Sci. Technol. 1993, 27, 1060-
The selection of species and components as well as the
assumed equilibrium reactions and kinetic reactions and rate
laws provided a consistent model for the description of the
hydrolytic degradation of phenyl picolinate at steady state.
The agreement between predicted and measured concentra-
tions of PhP and PhH in the transient state further indicated
that the assumptions were appropriate.
1067.
(11) Torrents, A.; Stone, A. T. Soil Sci. Soc. Am. J. 1994, 58, 738-745.
(12) Bald, E. Chem. Scripta 1978, 13, 47.
(13) Ludwig, C. Ph.D. Dissertation; University of Berne, Switzerland,
1993.
(14) Schindler, P.; Stumm, W. In Aquatic Surface Chemistry; Stumm,
W., Ed.; Wiley: New York, 1987; pp 83-110.
For the homogeneous system, observed rate constants
below pH 5 agreed with rate constants calculated using data
published by Torrents and Stone (4). For a description of the
hydrolysis above pH 5, carbonic acid species had to be
included in the model to account for the presence of
atmospheric CO2 in the reaction solutions and its effect on
solution pH. Thus, while it is difficult to completely exclude
CO2 from the experimental system, it can be dealt with
successfully in chemical models. In the presence of TiO2, the
rate of hydrolysis was substantially accelerated. In contrast
to initial rates obtained from batch experiments, steady-state
rates reflected the inhibition of the surface-catalyzed reaction
due to co-adsorbed species. This negative feedback effect
was also observed in batch reactors by initially adding one
of the reaction products, picolinic acid (10).
(15) Furrer, G.; Klausen, J.; Gfeller, M. Environ. Sci. Technol. Submitted
for publication.
(16) Schindler, P. In Adsorption of Inorganics at Solid-Liquid Interfaces;
Anderson, M. A., Rubin, A. J., Eds.; Ann Arbor Science: Ann
Arbor, MI, 1981; pp 1-50.
(17) Davis, J.; Kent, D. In Mineral-water interface geochemistry;
Hochella, M. F., White, A. F., Eds.; Reviews in Mineralogy Vol.
23; Mineralogical Society of America: Washington, DC, 1990;
pp 177-260.
(18) Felton, S.; Bruice, T. J. Am. Chem. Soc. 1969, 91, 6721-6732.
(19) Smith, R.; Martell, A. Critical Stability Constants. Volume 6:
Second Supplement; Plenum: New York, 1989.
(20) Westall, J. C. MICROQLsII. Computation of Adsorption Equilibria
in BASIC; Technical Report; Swiss Federal Institute for Envi-
ronmental Science and Technology: Du¨ bendorf, Switzerland,
1979.
(21) Westall, J. C. MICROQLsI. A Chemical Equilibrium Program in
BASIC, Version 2 for PC’s; Technical Report 86-02; Oregon State
University: Corvallis, OR, 1986.
The fate of organic pollutants in natural systems may be
determined by steady-state rates rather than initial rates,
depending on the reaction time scale and the residence time
of the compounds in the environmental compartment. If
(22) Furrer, G.; Stumm, W. Geochim. Cosmochim. Acta 1986, 50, 1847-
1860.
9
3 7 0 0 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 12, 1997

(23) Furrer, G.; Westall, J.; Sollins, P. Geochim. Cosmochim. Acta 1989,
53, 595-601.
Received for review May 21, 1997. Revised manuscript re-
ceived August 29, 1997. Accepted August 30, 1997.X
(24) Caracotsios, M.; Stewart, W. Comput. Chem. Eng. 1985, 9, 359-
365.
ES970440Z
(25) Smith, R.; Martell, A. Critical Stability Constants. Volume 4:
Inorganic Complexes; Plenum: New York, 1976.
(26) Green, R.; Tong, H. J. Am. Chem. Soc. 1956, 78, 4896-4900.
X Abstract published in Advance ACS Abstracts, October 15, 1997.
9
VOL. 31, NO. 12, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 7 0 1