Wet peroxide degradation of atrazine

Article abstract of DOI:10.1016/S0045-6535(03)00701-X

The high temperature (150-200 °C), high pressure (3.0-6.0 MPa) degradation of atrazine in aqueous solution has been studied. Under these extreme conditions atrazine steadily hydrolyses in the absence of oxidising agents. Additionally, oxygen partial pressure has been shown not to affect atrazine degradation rates. In no case mineralisation of the parent compound was observed. The addition of the free radical generator hydrogen peroxide to the reaction media significantly enhanced the depletion rate of atrazine. Moreover, partial mineralisation of the organics was observed when hydrogen peroxide was used. Again, oxygen presence did not influence the efficiency of the promoted reaction. Consecutive injections of hydrogen peroxide throughout the reaction period brought the total carbon content conversion to a maximum of 65-70% after 40 min of treatment (suggesting the total conversion of atrazine to cyanuric acid). Toxicity of the effluent measured in a luminometer decreased from 93% up to 23% of inhibition percentage. The process has been simulated by means of a semi-empirical model.

Full text of DOI:10.1016/S0045-6535(03)00701-X

Chemosphere 54 (2004) 71–78
www.elsevier.com/locate/chemosphere
Wet peroxide degradation of atrazine
ꢀ
*
Eva M. Rodr ꢀı guez, Pedro M. Alvarez, F. Javier Rivas ,
Fernando J. Beltr ꢀa n
Departamento de Ingenier ꢀı a Qu ꢀı mica y Energ eꢀ tica, Facultad de Ciencias, Universidad de Extremadura,
Avenida de Elvas S/N, 06071 Badajoz, Spain
Received 27 November 2001; received in revised form 5 February 2003; accepted 3 July 2003
Abstract
The high temperature (150–200 ꢀC), high pressure (3.0–6.0 MPa) degradation of atrazine in aqueous solution has
been studied. Under these extreme conditions atrazine steadily hydrolyses in the absence of oxidising agents. Addi-
tionally, oxygen partial pressure has been shown not to affect atrazine degradation rates. In no case mineralisation of
the parent compound was observed. The addition of the free radical generator hydrogen peroxide to the reaction media
significantly enhanced the depletion rate of atrazine. Moreover, partial mineralisation of the organics was observed
when hydrogen peroxide was used. Again, oxygen presence did not influence the efficiency of the promoted reaction.
Consecutive injections of hydrogen peroxide throughout the reaction period brought the total carbon content con-
version to a maximum of 65–70% after 40 min of treatment (suggesting the total conversion of atrazine to cyanuric
acid). Toxicity of the effluent measured in a luminometer decreased from 93% up to 23% of inhibition percentage. The
process has been simulated by means of a semi-empirical model.
ꢁ
2003 Elsevier Ltd. All rights reserved.
Keywords: Wet peroxide oxidation; Hydrogen peroxide; Kinetic modelling; Water treatment; Herbicides
1
. Introduction
substances from contaminated water and tertiary treat-
ments are required.
The increasing use of herbicides in agriculture, non-
Adsorption on activated carbon is so far the pre-
ferred technology to deal with herbicides in drinking
water plants. Thus, a number of studies have been re-
ported on the use of this method both in bench and pilot
scale (Baldauf, 1993; Gicquel et al., 1997; Thacker et al.,
1997; Baup et al., 2000). Also, data can be found for
water supply plants (Griffini et al., 1999). Adsorption on
activated carbon is an efficient method showing high
levels of pollutant removal at acceptable rates. However,
it is also an expensive technology demanding a feasible
and cost-effective carbon regeneration step, able to
eliminate the adsorbates from the exhausted carbon
surface without significant loss in adsorption capacity.
Among the technologies established for activated
carbon regeneration purposes, the thermal treatment is
widely used in most common applications, nevertheless,
cropped industrial lands and on fallow lands has led to
the inevitable contamination of the natural raw waters.
Thus, post-application seasonal concentrations of dif-
ferent herbicides have been reported to exceed the
maximum contaminant level permitted for surface wa-
ters (Nelson and Jones, 1994).
Broadly speaking, herbicides show rather low water
solubility values and, in most of cases, they are toxic for
microorganisms. As a result, conventional biological
remediation processes are not suitable to remove these
*
Corresponding author. Fax: +34-924-271304.
E-mail address: fjrivas@unex.es (F.J. Rivas).
0
045-6535/$ - see front matter ꢁ 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0045-6535(03)00701-X

72
E.M. Rodr ꢀı guez et al. / Chemosphere 54 (2004) 71–78
a continuous loss of 5–15% of the initial carbon amount
is experienced per regeneration cycle (Sheintuch and
Matatov-Meytal, 1999). An emerging technology which
uses milder conditions than the thermal regeneration is
the wet air regeneration. This process consists of two
steps. In the first stage (desorption) pollutants are stea-
dily transferred to the aqueous phase by means of ele-
vated temperatures and pressures typically found in this
environment (Mishra, 1995). In the second stage, or-
ganics are oxidised to non-adsorbable and harmless
polar compounds or, finally mineralised to carbon di-
oxide and water. In the later case, re-adsorption of oxi-
dised intermediates is obviously avoided and complete
regeneration of the activated carbon is expected. Addi-
tionally, regeneration of AC may be accomplished onto
the surface of the solid, so the desorption stage is not
needed. For these type of processes use of oxidation
catalysts supported onto the AC surface can also be used
spectively. The system was pressurised directly from a
gas cylinder containing either nitrogen or air. The oper-
3
ating procedure was as follows: 350 cm of an unbuffered
atrazine aqueous solution (ꢀ2–4 · 10ꢁ M) was placed in
the batch reactor and the system pressurised to 1.0 MPa.
Then, the stirrer was initiated and the reactor heated to
the required operating temperature. When the desired
temperature was achieved, typically in less than 15 min,
the pressure was adjusted at the predetermined working
conditions using the gas cylinder pressure. This time was
taken as time zero. At appropriate intervals, a sample
was withdrawn from the reactor. The sampling line was
initially flushed twice with a quantity of liquid equivalent
to the sample line volume (2 ml). The sample was cooled
in the sampling line prior to pressure let-down to prevent
flashing of the liquid across the valve. Following sam-
pling, reactor pressure was restored to the initial value.
For hydrogen peroxide promoted experiments,
4
(
Sheintuch and Matatov-Meytal, 1999).
Although the wet air regeneration has been reported
2 2
commercial 33% H O was used to prepare the appro-
priate solutions. Five millilitres of the previous solutions
were injected by means of a stainless steel cylinder
coupled to the autoclave once the working temperature
had been attained. Therefore decomposition of this re-
agent before time zero was prevented. The hydrogen
peroxide concentration at time zero oscillated in the
as a promising alternative to recover exhausted activated
carbon, information on the oxidation kinetics of de-
sorbed species is scarce. In this sense, most of wet oxi-
dation studies have been focused on the kinetics of
aromatic compounds, with special emphasis on phenol
type substances, and to a lesser extent on other chemical
structures (aliphatics, inorganics, etc.) (Mishra, 1995).
No studies are found in the past and recent literature
on the main features on pesticide wet air oxidation.
Thus, if carbon regeneration is meant to be completed
by the wet air process, the behaviour of pesticides in this
environment should be previously assessed. Therefore,
in the present work, the wet air oxidation of atrazine has
been carried out. Also, the promoted oxidation of atr-
azine by hydrogen peroxide has been considered. Hy-
drogen peroxide is known to directly generate highly
reactive species capable of oxidasing most of organics in
a non-selective way (Rivas et al., 1999).
ꢁ4
ꢁ2
range 1 · 10 to 10 M.
2.2. Analytical methods
Atrazine and first degradation intermediates (dee-
thylatrazine, deisopropylatrazine and deethyldeisopro-
pylatrazine were obtained from Dr. Ehrenstorfer
laboratory (D 86199 Augsburg, Germany) and used as
received. The concentration of atrazine was monitored
with a Hewlett-Packard HPLC system equipped with a
1100 Series pump, an injector and a UV–VIS detector set
at 220 nm. In each run, 20 ll of sample was injected via
an automatic injection system. The mobile phase (flow-
3
ꢁ1
Atrazine is a selective s-triazine herbicide used to
control broadleaf and grassy weeds in different crops.
Because it does not adsorb strongly to soil particles and
has a lengthy half-life, it has a high potential for
groundwater contamination despite its moderate solu-
bility in water (Briggs, 1992).
rate 1.0 cm min ) was a solution of 40/60 v/v aceto-
nitrile–water and the stationary phase was a Nova-Pack
C18 column (Waters, 150 · 3.9 mm). The detector was set
at 210 nm. The separation was made with the following
acetonitrile–water gradient elution program: at room
temperature, 100% of pH 7.2 buffered pure water (Milli
Q, Millipore) at time 0 for 3 min, 87% by vol. up to 11
min and 60% by vol. up to 30 min, with a flow rate of 1.0
3
ꢁ1
2
. Experimental
cm min . Total carbon (TC) was analysed by means of
a Dohrmann DC-190 carbon analyser.
2
.1. Experimental apparatus and procedures
3
Experiments were completed in a 600 cm , 316
3. Results and discussion
stainless-steel autoclave operated in batch mode. Reactor
walls were lined with a PTFE layer so the only metallic
surface in contact with the reaction media was the stirrer.
Runs were carried out over a temperature and total
pressure ranges of 150–200 ꢀC and 3.0–6.0 MPa, re-
3.1. Hydrolysis of atrazine
In a first stage the wet degradation of atrazine was
conducted under an inert atmosphere of nitrogen to

E.M. Rodr ꢀı guez et al. / Chemosphere 54 (2004) 71–78
73
0
0
0
0
.9
.7
.5
.3
0.9
0.7
0.5
0.3
0
20 40 60 80 100 120 140 160 180
0
20
40
60
80
100 120
Time, min
Time, min
Fig. 2. Wet air degradation of atrazine. Influence of oxygen
partial pressure on atrazine conversion. Experimental condi-
Fig. 1. Wet degradation of atrazine under inert atmosphere.
Influence of temperature on atrazine conversion. Experimental
ꢁ4
conditions: P
T
¼ 6:0 MPa, pH ¼ 5–6, Catrazo ¼ 2 ꢂ 10ꢁ M (av-
4
tions: T ¼ 200 ꢀC, pH ¼ 5–6, C
¼ 2 ꢂ 10 M (average
atrazo
value). (r) P
T
¼ 4:0 MPa (air); (j) P
T
¼ 6:0 MPa (air); (d)
erage value). (N) 200 ꢀC; (j) 185 ꢀC; (d) 150 ꢀC (dotted
lines ¼ model calculations).
P
T
¼ 6:0 MPa (nitrogen).
3
.3. Effect of hydrogen peroxide addition
ascertain the existence of non-oxidative reactions lead-
ing to the parent compound disappearance from the
media. Thus, experiments were carried out in the ab-
sence of oxygen in the temperature range from 150 to
Based on the previous results, it was decided to
study the influence of the addition of a free radical
generator to the reaction media. Therefore, hydrogen
peroxide influence was tested by injecting this species
into the reactor at concentration values ranging from
2
00 ꢀC. Fig. 1 depicts the evolution of the normalised
remaining atrazine concentration with time for this
experimental series. As inferred from this plot, some
non-oxidative processes develop leading to atrazine
depletion. Given the conditions used in this set of
experiments, hydrolysis of atrazine (I) to form hydroxy-
atrazine (II) (6-hydroxy-N-2-ethyl-N-4-isopropyl-1,3,5-
triazine-2,4-diamine) is likely to occur (Widmer et al.,
0
to 0.01 M.
Fig. 3 shows the effect of hydrogen peroxide addition
on atrazine elimination profiles. As inferred from this
plot, a significant enhancement of the process was ex-
perienced at increasing doses of this reagent. At the sight
of the curves found for the normalised atrazine con-
centration profiles, it is observed how the decay rate of
the herbicide slows down as the reaction progresses.
Thus, from Fig. 3 it can be seen how after the initial fast
degradation stage (oxidation due to radical species),
1
993). Also, as expected, no change in the total carbon
content (TC) was experienced in the absence of oxidising
agents.
3
.2. Effect of oxygen partial pressure
The following step, considered in this study, was to
1
assess the effect of oxygen presence on both atrazine and
TC conversion rates. Consequently, a series of experi-
ments was conducted at different air total pressure and
the results compared to similar runs completed under
nitrogen atmosphere. Fig. 2 shows the experimental
atrazine conversions obtained when working at 4.0 and
0.8
0.6
0.4
0.2
0
6
.0 MPa by using either air or nitrogen. As observed
from this plot, at the conditions used in this work,
oxygen addition did not result in the enhancement of the
process in terms of atrazine removal. Moreover, TC
remained unvaried throughout the reaction regardless of
the oxygen partial pressure used. This indicates that
oxygen was unable to attack either atrazine and/or hy-
droxyatrazine and, subsequently, begin the typical radi-
cal chain mechanism developed in wet air oxidation
processes, eventually leading to the partial mineralisa-
tion of the parent compound.
0
20
40
60
80
100
120
Time, min
Fig. 3. Wet air degradation of atrazine. Influence of hydrogen
peroxide addition on atrazine conversion. Experimental con-
ditions: P
T
¼ 6:0 MPa (air), T ¼ 200 ꢀC, pH ¼ 5–6, Catrazo
¼
¼
ꢁ4
2
ꢂ 10 M (average value). (d) CH O
o
¼ 0:0 M; (j) CH O
2
o
2
2
2
ꢁ
4
ꢁ4
ꢁ3
10 M; (N) C
H2O2 o
¼ 5 ꢂ 10 M; (.) C
H2O2 o
¼ 10 M; (r)
ꢁ2
C
H2
O
o
¼ 10 M (lines ¼ model calculations).
2

74
E.M. Rodr ꢀı guez et al. / Chemosphere 54 (2004) 71–78
atrazine depletion rate levels off to values below the rate
expected for a zero order kinetics if hydrolysis followed
this type of rate law, i.e. in case of a zero order kinetics a
faster elimination rate would develop after hydrogen
peroxide consumption has been achieved. As a conse-
quence, it might be considered that the thermal degra-
dation of atrazine does follow a first order kinetics.
(deisopropylatrazine, VI) or ethyl group removal (dee-
thylatrazine, V). Subsequent reactions of these interme-
diates may lead to complete elimination of substituents
to finally give cyanuric acid (VIII) and low molecular
weight final products. Cyanuric acid is one of the re-
fractory substances found in the oxidation processes and
it eventually accumulates in the reaction media (Beltr ꢀa n
et al., 2000). Nevertheless, according to literature data,
this compound has been demonstrated to be non-toxic
3
.4. Kinetic modelling
(
Canelli, 1974).
In the mechanism, thermal decomposition of atrazine
In the presence of added hydrogen peroxide, the high
temperature, high pressure degradation of aqueous atr-
azine was modelled by considering the mechanism pre-
sented in Scheme 1. In the adopted reaction mechanism,
atrazine (I) is considered to react with water (hydrolysis)
or with hydroxyl radicals to yield the chloride substitu-
tion (hydroxyatrazine, II), isopropyl group removal
has been assumed to proceed through simple first order
kinetics. Also, hydrogen peroxide has been considered to
undergo homolytic excision to generate two hydroxyl
radicals, capable of degrading atrazine (reaction 1). The
value assumed of kper was first taken from a previous
work (Rivas et al., 1999) although fitting of this
Scheme 1. Proposed mechanism for the wet peroxide oxidation of atrazine.

E.M. Rodr ꢀı guez et al. / Chemosphere 54 (2004) 71–78
75
parameter resulted in a slightly lower value (the fitting
session was made by just considering atrazine degrada-
tion profiles). Ineffective reaction of hydrogen peroxide
decomposition leading to non-oxidative species has also
been taken into account by means of an apparent con-
stant kineff (reaction 2). The later reaction includes ele-
mentary steps like radical–radical termination, injection
effects producing local high concentrations of hydrogen
peroxide which, at these conditions, would act as an
scavenger of hydroxyl radicals, radical deactivation
against the reactor wall, etc.
by EP. To start with the computational method initial
guess of rate constants should be accomplished. Initial
values of some rate constants of the reactions between
organics and hydroxyl radicals were taken from the lit-
erature at room temperature (Beltr ꢀa n et al., 2000). When
no available data could be found for rate constants a
9
ꢁ1 ꢁ1
generic value of 5 · 10 M
s was assumed (Buxton
et al., 1988). As observed from Scheme 1 presented, the
values of the rate constants between the hydroxyl radical
and the organic species are slightly above than those
reported at room conditions (Beltr ꢀa n et al., 2000). Thus,
typical values reported for s-triazines and hydroxyl
ꢀ
H
H
2
2
O
O
2
2
! 2OH
! 1=2O
k
Per
ð1Þ
ð2Þ
9
ꢁ1 ꢁ1
radicals are in the interval 1–5 · 10 M
s . The one
fold higher order obtained in the regression analysis may
be attributable to the higher temperature conditions
used in the WAO environment. It should be noticed that
the increase in the values of these constants is not sig-
nificant if compared (or expected) to differences in
working temperatures. This is due to the low activation
energy of the reactions of the hydroxyl radical, (i.e.
typical activation energies for hydroxyl radical reactions
are close to 5 kJ/mol, confirming the magnitude of rate
constants calculated in the fitting process).
2
þ H
2
O
k
ineff
In a first attempt, the rate constant of atrazine hydrolysis
was obtained by fitting experimental results acquired in
experiments completed in the absence of hydrogen per-
oxide. The fitting process was carried out by a computer
software using the evolutionary programming (EP)
which is a computational technique that mimics evolu-
tion and is based on reproduction and selection (B €a ck
et al., 1993; Mendes and Kell, 1998). Fig. 1 shows the
computed evolution of atrazine with time (dashed lines)
Subsequently, atrazine and intermediate theoretical
concentration profiles were determined and compared to
the experimental results obtained in the various runs
conducted in the presence of hydrogen peroxide. Fig. 3
shows the computed results (solid lines) of atrazine
concentration profiles obtained after optimisation of the
value of adjustable rate constants. As seen from this
plot, the model does a good job when simulating the
atrazine depletion with time. Also, Fig. 4 depicts the
experimental and computed intermediate concentra-
tions identified in this work for a run completed with
0.01 M of hydrogen peroxide. As observed, the reaction
mechanism proposed simulates quite well the main
1
after the adjustable parameter k was optimised. The
Arrhenius analysis of this parameter allowed for the
calculation of the pre-exponential factor and activa-
tion energy of the reaction. Thus, the following expres-
8
ꢁ1
sions was obtained: k
1
¼ 1:87 ꢂ 10 expðꢁ13237=TÞ s
.
Hydrolysis of atrazine has been studied at room tem-
perature and no available data was found at high tem-
perature. Extrapolation of data at ambient conditions
by means of the Arrhenius equation obtained is some-
how audacious. However the attempt was conducted
1
and the value of k calculated did corroborate the per-
sistence of this herbicide at circumneutral pH and room
temperature. Thus Widmer et al. (1993) gave a value of
ꢁ
13 ꢁ1
around 4.9 · 10
In this work a value of 2.8 · 10
using the Arrhenius expression for a temperature of
7 ꢀC.
Kinetic constants kPer and kineff are known to be
s
for a temperature range 4–30 ꢀC.
ꢁ
12 ꢁ1
s was obtained when
4x10
8x10-5
3
2
1
x10-4
1
6
4
2
x10-5
5
highly dependent on the nature and geometry of the
reactor (Rivas et al., 1998). Several values have been
reported in the literature (Shibaeva, 1969; Rivas et al.,
x10-4
x10-5
x10-4
x10-5
1
process) of 10
999). In this work, an average value (after the fitting
ꢁ1
ꢁ4
s
has been assumed for kPer. As stated
0x100
0
x100
previously, this value is lower than those reported for
the decomposition of hydrogen peroxide in the wet air
oxidation environment (Shibaeva, 1969; Rivas et al.,
0
5
10
15
20
25
30
Fig. 4. Wet air degradation of atrazine. Influence of hydrogen
peroxide addition on atrazine conversion and intermediates
1
999). The reason of this lower value of kPer is based on
the use of a reactor surface made of PTFE, which is
more inert than the metallic surface reported by the
aforementioned authors.
formation. Experimental conditions:
P
T
ꢁ4
¼ 6:0 MPa (air),
ꢁ2
T ¼ 200 ꢀC, pH ¼ 5–6, Catrazo ¼ 3:6 ꢂ 10 M, CH O
2
o
¼ 10
2
M. (d, ––), Atrazine; (r, - - - -), Deethylatrazine, (D, –ꢃ–ꢃ–),
Deisopropylatrazine; (j, – – –), Deethyldeisopropylatrazine
(symbols ¼ experimental results, lines ¼ model calculations).
Once k
1
and kPer were established, the rest of the
unknown rate constants were simultaneously evaluated

76
E.M. Rodr ꢀı guez et al. / Chemosphere 54 (2004) 71–78
by-product concentration profiles, i.e. deethylatrazine,
deisopropylatrazineanddeethyldeisopropylatrazine.Val-
idation of the reaction mechanism was conducted by
testing the kinetic model at different conditions of initial
atrazine concentration (Fig. 5A) and initial hydrogen
peroxide concentration (Fig. 5B).
following equation was obtained, Lnðkineff Þ ¼ ꢁ4:41 þ
0:40Ln (H ). It has to be pointed out that values ob-
2
O
2
tained for kineff are of equal magnitude than those re-
ported for a similar system like FentonÕs reagent (Rivas
et al., 2002) with the corresponding divergences attrib-
utable to differences in working temperature. A sensi-
tivity analysis of the proposed mechanism demonstrated
that the limiting reactions were the atrazine hydrolysis
and hydrogen peroxide decomposition. The values of
the rest of kinetic parameters is not of paramount im-
portance as far as they are considered in the normal
range of values for the reaction between the hydroxyl
radical and organic species (considering the high oper-
ating temperature). Nevertheless, these values of the rate
constants should be taken with caution since they have
been obtained by means of an optimisation process and
have to be considered as a mere approach to the actual
values.
As stated before, from the fitting process the value of
ineff could be obtained. Given the empirical definition of
k
this apparent rate constant, accounting for several ele-
mentary steps, its value changed as the concentration of
hydrogen peroxide was varied. Nevertheless, a simple
power expression could be applied to relate kineff and the
initial hydrogen peroxide concentration used, kineff
a[H O o] . After linearisation, by applying the natural
2 2
logarithm to both sides of the previous expression, the
¼
n
3
2
2
1
1
5
.0x10
Again the presence of oxygen was tested for experi-
ments initiated by hydrogen peroxide. A series of ex-
periments was therefore carried out at different air total
pressure (range ¼ 3.0–6.0 MPa) and under nitrogen at-
mosphere by adding the same amount of the promoter/
generator at the beginning of the reaction. Two sets of
.5x10-4
.0x10-4
.5x10-4
.0x10-4
ꢁ2
ꢁ3
2 2
runs were completed with 10 and 10 M of H O . The
.0x10-5
experimental results (not shown) established that oxygen
had not influence either on the final atrazine removal or
TC conversion, therefore discarding the development of
a chain mechanism involving organic radical species.
Some small differences found in atrazine conversion
(±10%) did not follow a clear trend and could be due to
sampling effects, even after flushing the sampling port
twice before a sample was taken for analysis. The evo-
lution of TC followed a parallel behaviour to that ex-
pected for the hydrogen peroxide concentration and
consequently for the hydroxyl radicals present in the
0
.0x100
0
1
2
3
4
5
6
7
8
9
10
(
A)
Time, min
.0x10-4
.5x10-4
.0x10-4
.5x10-4
.0x10-4
.5x10-4
.0x10-4
.0x10-5
4
3
3
2
2
1
1
5
reaction media. In Fig. 6 experimental TC conversion
ꢀ
and theoretical H
2
O
are plotted for the experiment carried out when adding
2
and HO radicals concentrations
ꢁ2
1
0
M of the radical generator. As seen from this figure,
0
.0x100
once hydrogen peroxide has disappeared, TC conversion
reaches a plateau of roughly 35% with no further eli-
mination after 90 min of treatment.
0
20
40
60
80
100
120
(
B)
To ascertain if the TC plateau could be attributed to
Fig. 5. (A) Wet air degradation of atrazine. Influence of hy-
drogen peroxide addition on atrazine conversion and interme-
2 2
the absence of H O or, contrarily, to the accumulation
ꢁ4
diates formation. Experimental conditions: Catrazo ¼ 2:7 ꢂ 10
2 o
of refractory species to the action of the oxidant, some
experiments were then carried out by adding different
ꢁ
2
M, CH2O ¼ 10 M, P ¼ 6:0 MPa (air), T ¼ 200 ꢀC, pH ¼
T
5
–6. (d, –––), Atrazine; (r, - - - -), Deethylatrazine, (D, –ꢃ–ꢃ–),
2 2
amounts of H O in successive steps throughout the
Deisopropylatrazine; (j, –––), Deethyldeisopropylatrazine
reaction period. As seen in Fig. 7, injection of additional
ꢁ2
(
symbols ¼ experimental results, lines ¼ model calculations). (B)
amounts of hydrogen peroxide (10 M) to the reaction
Wet air degradation of atrazine. Influence of hydrogen peroxide
addition on atrazine conversion and intermediates formation.
media led to a significant improvement of the process in
terms of TC final depletion. Therefore, the plateau
reached for the TC conversion in experiments carried
atrazo ¼ 3:6 ꢂ10ꢁ
4
Experimental conditions:
5
C
M, CH2O2
o
¼
ꢁ4
ꢂ 10 M, P ¼ 6:0 MPa (air), T ¼ 200 ꢀC, pH¼5–6. (d, ---),
T
2 2
out with a single H O injection was likely due to the
Atrazine; (r, ----), Deethylatrazine, (D, –ꢃ–ꢃ–), Deisopropy-
latrazine; (j, –––), Deethyldeisopropylatrazine (symbols¼
experimental results, lines¼model calculations).
disappearance of this reagent at the beginning of the
process. In any case, no important differences were ex-

E.M. Rodr ꢀı guez et al. / Chemosphere 54 (2004) 71–78
77
could be expected since reactions of hydroxyl radicals
are characterised by low activation energies, normally
the generation of these species being the controlling step.
Toxicity of samples after hydrogen peroxide treat-
ment was measured by using a luminometer (Tox-
1
8
6
4
2
x10
40
3
5
0
x10-3
8.0x10-14
3
.0x10-14
6
4
2
OH (mol l-1
x10-3
C
)
25
20
15
10
5
.0x10-14
.0x10-14
ꢂ
Alert ), this method based on the inhibition percentage
x10-3
of a bioluminiscent bacteria (Vibrio fischeri NRRL B-
11177). The percentage of inhibition decreased from
93% untreated atrazine solution to 23% after oxidation
which suggest the formation of the low toxic cyanuric
acid.
0
.0x100
x10-3
0
20
40
60
80
Time, min
0
x100
0
0
10 20 30 40 50 60 70 80 90
Fig. 6. Wet air degradation of atrazine in the presence of hy-
drogen peroxide. Evolution of experimental TC (solid line +
4. Conclusions
circle symbols), theoretical H
ꢀ
2
O
2
(dotted line) and theoretical
HO radicals (embedded figure) with time. Experimental con-
On the basis of the above experimental findings on
the wet air oxidation of atrazine the following conclu-
sions may be drawn:
ditions: P
T
¼ 6:0 MPa (air or nitrogen), T ¼ 200 ꢀC, pH ¼ 5–6,
atrazo ¼ 3:6 ꢂ 10ꢁ M, CH O
4
¼ 10 M, TC
ꢁ2
¼ 22 mg l
ꢁ1
.
C
2
o
o
2
ii(i) Atrazine hydrolyses at high temperatures in the ab-
sence of any oxidising agent. Hydrolysis of the her-
bicide does not contribute to the removal of the
organic load of the aqueous solution. Hydroxyatr-
azine likely accumulates in the reaction media
and, as a consequence, it does not seem an appro-
priate environment for carbon regeneration since
this metabolite probably re-adsorbs on the carbon
surface.
60
50
40
30
20
10
i(ii) The presence of oxygen does not enhance atrazine
degradation rate and accumulation of hydrolysis
products also occurs.
0
0
5
10 15 20 25 30 35 40
(
iii) Addition of hydrogen peroxide results in a signifi-
cant enhancement of the process. The use of this re-
agent also allows for the partial mineralisation of
the parent compound, though TC depletion stops
Fig. 7. Wet air degradation of atrazine in the presence of hy-
drogen peroxide. Evolution of experimental TC with time after
ꢁ
2
several injections of hydrogen peroxide (10 M). Conditions as
in Fig. 6. Arrows indicate the time at which 10 M of a fresh
solution of H
was injected. (j) 200 ꢀC single injection; (.)
ꢁ2
2
O
2
when the H
(iv) Consecutive injections of H
intervals lead to a final TC conversion of roughly
2 2
O
added completely decomposes.
2
00 ꢀC multiple injection; (d) 150 ꢀC single injection; (N)
2 2
O
at determined time
1
50 ꢀC multiple injection.
6
5–70% with no further increase of this parameter
M
ꢁ2
after having injected a total amount of 5 · 10
of H O . The high reduction of TC achieved and
2 2
perienced after the last injection of H
be postulated that after a given amount of H
2
O
2
. Hence, it can
has
2
O
2
likely the accumulation of small polar acids in the
media dictates that this technique should be suitable
for spent carbon regeneration.
been injected to the system, accumulation of refractory
substances occurs and no further conversion of TC is
achieved. The drop in pH (roughly 2 units) attained at
the end of the oxidation process suggests the formation
of HCl from the original atrazine molecule (either by
hydrolysis or direct attack of the hydroxyl radical to
the chloro position). Although speculative in nature
i(v) Two factors might contribute to the economy of the
process. On one hand, the amount of hydrogen per-
oxide used can be lowered since a negligible im-
provement of the TC conversion is obtained after
the third injection. On the other hand, the use of
spent carbon permits to achieve a concentration of
atrazine in the WAO environment higher than the
concentration reached after saturation at room tem-
perature. Therefore, it is likely that hydrogen per-
oxide can be more effectively used reducing the
fraction of this reagent leading to oxygen and water.
Additionally, in a real plant, H O should be fed
(
cyanuric acid has not been measured), it is worth to
signify that the plateau reached in TC conversion (ꢀ60–
5%) roughly coincides with the theoretical decrease in
6
TC calculated if all molecules of atrazine had trans-
formed into cyanuric acid (62.5%). Also, it is notewor-
thy to say that temperature does not significantly affect
TC final abatement. As stated previously, these results
2
2

78
E.M. Rodr ꢀı guez et al. / Chemosphere 54 (2004) 71–78
continuously, so no injection of this reagent is
needed facilitating, therefore, the process.
solved organic and mineral matter of natural waters.
Environ. Technol. 18 (5), 467–478.
Griffini, O., Bao, M.L., Burrini, D., Santianni, D., Barbieri,
C., Pantani, F., 1999. Removal of pesticides during the
drinking water treatment process at Florence water supply,
Italy. J. Water Serv. Res. Technol.-Aqua. 48 (5), 177–
Acknowledgements
1
85.
Authors wish to thank the CYCIT of Spain and the
FEDER funds of European Union for the financial
support under project IFD97/2222/C02/03.
Mendes, P., Kell, D.B., 1998. Non-linear optimization of
biochemical pathways: applications to metabolic engineer-
ing and parameter estimation. Bioinformatics 14, 869–
8
83.
Mishra, V.S., 1995. Wet air oxidation. Ind. Eng. Chem. Res. 34,
–48.
2
References
Nelson, H., Jones, R., 1994. Potential regulatory problems
associated with atrazine, cyanazine, and alachlor in surface-
water source drinking-water. Weed Technol. 8 (4), 852–861.
Rivas, F.J., Kolaczkowski, S.T., Beltr ꢀa n, F.J., McLurgh, D.B.,
1998. Development of a model for the wet air oxidation of
phenol based on a free radical mechanism. Chem. Eng. Sci.
53, 2575–2586.
B €a ck, T., Fogel, D.B., Michalewicz, Z., 1993. Handbook of
Evolutionary Computation. IOP Publishing/Oxford Uni-
versity Press, Oxford.
Baldauf, G., 1993. Removal of pesticides in drinking-water
treatment. Acta Hydroch. Hydrob. 21 (4), 203–208.
Baup, S., Jaffre, C., Wolbert, D., Laplanche, A., 2000.
Adsorption of pesticides onto granular activated carbon:
determination of surface diffusivities using simple batch
experiments. Adsorp. J. Int. Adsorp. Soc. 6 (3), 219–228.
Beltr ꢀa n, F.J., Gonz ꢀa lez, M., Acedo, B., Rivas, F.J., 2000.
Kinetic modelling of aqueous atrazine ozonation processes
in a continuous flow bubble contact. J. Hazard. Mater. 80
Rivas, F.J., Kolaczkowski, S.T., Beltran, F.J., McLurgh, D.B.,
1999. Hydrogen peroxide promoted wet air oxidation of
phenol: influence of operating conditions and homogeneous
metal catalysts. J. Chem. Technol. Biotechnol. 74 (5), 390–
398.
ꢁ
Rivas, F.J., Beltr ꢀa n, F.J., Gimeno, O., Alvarez, P., 2002.
Treatment of brines by combined Fenton reagent aerobic
biodegradation. II. Process modelling. J. Hazard. Mater. B
96, 259–276.
(1–3), 189–206.
Briggs, S., 1992. A Basic Guide to Pesticides: Their Character-
istics and Hazards. Hemisphere Publishing Corporation,
Washington. pp. 8–9.
Sheintuch, M., Matatov-Meytal, Y.I., 1999. Comparison of
catalytic processes with other regeneration methods of
activated carbon. Catal. Today 53 (1), 73–80.
Shibaeva, L.V., 1969. The oxidation of phenol with molecular
oxygen in aqueous solutions. II. The role of hydrogen
peroxide in the oxidation of phenol. Kinet. Catal. 10, 1022–
1026.
Buxton, G.V., Greenstock, C.L., Helman, W.P., Ross, A.B.,
1988. Critical review of rate constants for reactions of
hydrated electrons, hydrogen-atoms and hydroxyl radicals
ꢀ
ꢀ
ꢁ
(
(
OH/ O ) in aqueous-solution. J. Phys. Chem. Ref. Dat. 17
2), 513–886.
Canelli, E., 1974. Chemical, bacteriological and toxicological
properties of cyanuric acid and chlorinated isocyanurates as
applied to swimming pool disinfection. Am. J. Public Health
Thacker, N.P., Vaidya, M.V., Sipani, M., Kalra, A., 1997.
Removal technology for pesticide contaminants in potable
water. J. Environ. Sci. Health B 32 (4), 483–496.
Widmer, S., Olson, J., Koskinen, W., 1993. Kinetics of atrazine
hydrolysis in water. J. Environ. Sci. Health B 28 (1), 19–
28.
62, 155–162.
Gicquel, L., Wolbert, D., Laplanche, A., 1997. Adsorption of
atrazine by powdered activated carbon: influence of dis-