Theoretical Estimation of the Apparent Rate Constants for Ozone

Decomposition in Gas and Aqueous Phases Using Ab Initio

Calculations

Mehrshad Parchei Esfahani ,1* Chongchong Wu1 and Alex De Visscher1,2

1. Department of Chemical and Petroleum Engineering, Schulich School of Engineering, University of Calgary, 2500 University Drive NW,
Calgary, AB, T2N 1N4, Canada
2. Department of Chemical and Materials Engineering, Gina Cody School of Engineering and Computer Science, Concordia University, 1455 de
Maisonneuve Blvd. E, Montréal, QC, H3G 2W1, Canada

An ab initio study, using the coupled cluster calculations (CCSD) method was conducted to investigate the kinetics of the ozone degradation in
gas and aqueous phases considering the reaction of ozone with the hydroperoxyl radical. Two potential transition state paths, oxygen and
hydrogen transfer, are studied and compared. It was revealed by the ab initio quantum chemical calculations that the calculated overall rate
constant in the gas phase differs by approximately an order of magnitude from measured values. However, the calculated selectivity (branching
fraction), which was measured directly with isotope studies of hydrogen atom transfer, is almost exactly equal to the experimental value at
298.15 K. The sensitivity analysis showed that adding the reaction between ozone and hydroperoxyl radical to the kinetic model accelerates the
decomposition process by more than four times in the aqueous phase (pH = 7–8.5), and for an order of magnitude change in the rate constant of
this reaction, the decomposition half‐life changes by 20–45 %. This result might affect our understanding of atmospheric ozone chemistry.

Keywords: ab initio, ozone degradation, hydroperoxyl radical, coupled cluster calculations

INTRODUCTION phase have not been able to predict the decomposition process
well in a broad range of pH, leaving a gap in our understanding of

O
zone can destroy a wide range of pollutants and is a
widely used oxidant.[1,2] A number of studies show that the process.[14,28] Furthermore, the decomposition of ozone is
the reaction between molecular ozone and species in influenced by a large number of factors, particularly in the pre-
water tends to be highly selective.[3,4] On the other hand, reac- sence of other compounds, and the occurrence of small amounts
tions of the hydroxyl radical, an ozone decomposition by‐product, of impurities.[29,30] Consequently, kinetic studies of the decay of
in water are non‐selective, resulting in faster reaction rates and ozone lead to contradictory results, deviating by up to three or-
thus, rapid degradation of compounds.[5–7] The mechanism of ders of magnitude. For instance, different reaction orders are re-
ozone decomposition in aqueous phase includes a chain of re- ported for the decomposition of ozone in water at the same
actions yielding highly reactive radicals.[8–12] The efficiency of an pH.[14,31]
ozone‐based wastewater treatment depends on ozone concentra- For example, the reported reaction order of ozone decomposi-
tion and is linked to its rate of decomposition.[13] Therefore, it is tion in water in the pH range of 8–10 varies from 1–2.[9,32,33]
important to study and clarify the kinetics of ozone decomposi- Kilpatrick et al.[33] explored the kinetics of ozone decomposition
tion in water. in pH range of 8–10, 298.15 K , and with an ozone initial con-
The mechanism of aqueous ozone decomposition has been centration of 3 × 10−4 M. They reported the reaction order to be 2.
studied for several decades.[10,12,14,15] However, most of these Staehelin and Hoigne[9] tested the ozone decomposition in water
studies are primarily based on overall kinetics found for pure at the pH range of 8–10 and initial ozone concentration of 3 ×
water and the mechanism is kinetically complex due to the pro- 10−6 M. The reaction order is measured to be 1 in this study.
duction of free radicals, which may either be scavenged by bi- Hence, the kinetics of the ozone decomposition in aqueous so-
carbonate and organic solutes or react with the remaining ozone lutions are not clearly understood yet.
molecules to yield more free radicals.[9,16] This process gradually A comparison between a chemical kinetic model based on the
accelerates the decomposition of O3.[9] Recent applications of reactions reported in the literature and the results of the con-
ozone in water include nitric acid and 2‐propanol treatment, ducted experimental studies in this area indicates a considerable
disinfection by‐products, food preservation, and groundwater re-
mediation.[17–22] Other areas of active research are process in-
* Author to whom correspondence may be addressed. E‐mail address:
tensification, and optimization of the peroxone process.[23,24] A [email protected]
recent field of interest for applications of ozone include ozone Can. J. Chem. Eng. 1–7, 2019
therapy in infantile atopic dermatitis, chronic venous in- © 2019 Canadian Society for Chemical Engineering
sufficiency, and diabetes.[25–27] DOI 10.1002/cjce.23594
Despite many studies conducted in this area, the mechanism of Additional supporting information may be found in the online version of
this article at the publisher’s web-site
the decay of ozone is not yet entirely understood. Consequently, Published online in Wiley Online Library
the existing kinetic models of ozone decomposition in aqueous (wileyonlinelibrary.com).

MONTH 2019 THE CANADIAN JOURNAL OF CHEMICAL ENGINEERING 1

deviation from experimental results. This finding suggests the
potential existence of a number of other reactions that have not O3 + HO−2 ⇒ HO5− (R3)
yet been considered in the mechanism. Oxygen‐hydrogen chem-
istry studies of aqueous phase atmospheric reactions, show that O3 + HO. ⇒ HO.2 + O2 (R4)
the reaction O3 + HO.2 → HO. + 2O2 has the potential to solve this
problem.[31,34] Especially, in combination with the reaction HO−2 + HO. ⇒ HO.2 + OH− (R5)
O3 + HO. ⇒ HO.2 + O2, this would create a chain reaction me-
chanism of ozone decomposition.
Although the chemistry of the reaction of ozone with hydro- OH− + HO. ⇒ H2 O + O. − (R6)
peroxyl radical is well studied in gas phase, there are only a limited
number of studies reporting a rate constant of the reaction in O. − + H2 O ⇒ HO. + OH− (R7)
water.[35–38] Chameides and Davis[35] estimated a rate constant of
4 × 106 m3 kmol−1 s−1 for the reaction based on a 1:40 ratio for O. − + HO−2 ⇒ O.2− + OH− (R8)
reaction rates (O3 + HO.2 / O3 + HO.), and Graedel et al.[38] used this
rate constant in their model for droplet chemistry. Sehested et al.[34]
O. − + O2 ⇒ O.3− (R9)
reported the rate constant to be as low as 104 m3 kmol−1 s−1.
However, the validity of these rate constants was questioned by
later studies. Moreover, the reported rate constants differ by several O3 + O.2− ⇒ O.3− + O2 (R10)
orders of magnitude. In gas phase, a value of 2.0 × 10−15 cm3
molecule−1 s−1, equivalent to 1.2 × 106 m3 kmol−1 s−1, has been O.3− ⇒ O. − + O2 (R11)
reported.[39,40] The reaction of hydroperoxyl radical with ozone is
known to exist in the gas phase, however, most of the conducted
HO−4 ⇒ HO.2 + O.2− (R12)
studies in water have only taken into account the reaction with the
dissociated peroxyl radicals, and this reaction has not yet been
considered in any water phase study. HO.2 + O.2− ⇒ HO−4 (R13)
Quantum chemical reaction calculations are powerful tools to
. .−
calculate the optimized geometries for reactants and products, lo- HO−
5 ⇒ HO2 + O3 (R14)
cate intermediates and transition states, and obtain reaction ki-
netics for different reaction pathways with a reasonable accuracy to O.3− + O. − ⇒ 2O.2− (R15)
examine how the path and orientation of approaching reactants
lead to a chemical reaction.[41,42] To fill the existing knowledge gap,
the purpose of this study is to calculate the rate constant of the HO.2 + O.3− ⇒ HO−
5 (R16)
O3 + HO.2 → HO. + 2O2 reaction in the aqueous medium using the
Gaussian 09W package and GaussView 5 molecular modelling HO−
5 ⇒ 2O2 + OH
−
(R17)
software.[43] A sensitivity analysis is conducted to explore the effect
of the reaction rate constant on the overall kinetics of the ozone HO. + HO.2 ⇒ H2 O + O2 (R18)
decomposition process in aqueous phase.
The reaction with the obtained rate constant was considered in
a detailed model of ozone degradation in aqueous phase in- O. − + O.2−[+ H2 O] ⇒ O2 + 2OH− (R19)
cluding a total of 22 reactions. The current study can improve our
understanding of the reactions of ozone in water and provide O.3− + HO. ⇒ O3 + OH− (R20)
solutions to minimize the deviation of rate constant between si-
mulation and experimental studies. Additionally, it is an oppor- HO. + O.2− ⇒ OH− + O2 (R21)
tunity to compare the rate constants in the gas and aqueous
phases. Understanding the mechanism of ozonation in aqueous The aqueous system is assumed without buffer solution and the
phase could also broaden our knowledge of ozone chemistry and reactions of different species with buffer ions are disregarded. The
pollutant degradation in aerosols and, hence, of photochemistry rate constant of Reactions (R2)–(R21) are listed in Table 1.
in the atmosphere.[44–46] Based on the kinetic scheme (Reactions (R1)–(R21)), a system of
differential equations, describing the time dependence of the
concentrations of different species involved in the ozone de-
CHEMICAL KINETIC MODEL composition process was composed. Solving the system of dif-
A plausible mechanism was developed based on the reactions ferential equations provides us with apparent rate constant for
involved in the ozone decomposition in the water phase and was ozone decomposition. The differential equations were solved with
reported in the available literature.[10,11,14,47] The following re- Gear’s method as implemented in MATLAB R2015b.
actions are considered in the model:
Equilibrium:

H2 O ⇔ H+ + OH− (R1) MOLECULAR STRUCTURES AND ENERGIES

Kinetics: Optimization of Reactants and Products

The reaction between ozone and hydroperoxyl radical (R22) was
O3 + OH− ⟹ HO−4 (R2) evaluated by using the Gaussian 09 molecular modelling

2 THE CANADIAN JOURNAL OF CHEMICAL ENGINEERING MONTH 2019

 Table 1. Rate (k i) and equilibrium constants (Ki) of the reactions Table 2. Calculated and experimental bond characteristics and
participating in decomposition process of ozone in water thermodynamics of optimized ozone under different levels of theory
in gas phase
Constant Value Unit Reference
Bond Bond angle
K1 10−14 −
Level of theory length (× 1010 m) (degrees)
k2 175 m3 kmol−1 s−1 Gardoni et al.[14]
k3 2. 8 × 106 m3 kmol−1 s−1 Gardoni et al.[14] CCSD(T)/6‐311 + + G 1.320 109.500
Buehler et al.[48] (3df,3pd)
k4 2. 0 × 109 m3 kmol−1 s−1
CCSD/aug‐cc‐pvdz 1.259 117.230
k5 5. 6 × 109 m3 kmol−1 s−1 Fischbacher et al.[49]
QCISD 6‐311 + + (d,p) 1.255 117.935
k6 1. 2 × 109 m3 kmol−1 s−1 Christensen et al.[50]
QCISD 6‐311 + (d) 1.255 117.935
k7 9. 6 × 107 s−1 Gardoni et al.[14] Experimental data 1.278 116.8
k8 1. 0 × 109 m kmol−1 s−1
3 Ignatiev et al.[11]
k9 3 . 5 × 109 m3 kmol−1 s−1 Gardoni et al.[14]
k10 1. 5 × 109 m3 kmol−1 s−1 Sehested et al.[10] Table 3 compares the calculated bond lengths and angles of
k11 1. 9 × 103 s−1 Gardoni et al.[14] reactant and products with the corresponding experimental va-
k12 107 s−1 Gardoni et al.[14] lues in gas phase.[54] As this table demonstrates, the calculated
k13 5. 0 × 109 m3 kmol−1 s−1 Gardoni et al.[14] values are within the range (maximum 3.2 % deviation) and are
k14 107 s−1 Gardoni et al.[14] mainly larger than the experimental values. Hence, the CCSD/
k15 7. 0 × 108 m kmol−1 s−1
3 Buxton et al.[47] aug‐cc‐pvdz level of theory is sufficiently accurate and can be
k16 5. 0 × 109 m3 kmol−1 s−1 Gardoni et al.[14] used for optimization in the aqueous phase. The stable structures
k17 107 s−1 Gardoni et al.[14] of reactants and products in the gas and aqueous phases using the
k18 6. 6 × 109 m kmol−1 s−1
3 Buxton et al.[47] chosen level of theory are in the Supplementary Information.
k19 6. 0 × 108 m3 kmol−1 s−1 Buxton et al.[47]
The Transition State (TS) and Relative Energies
k20 2. 5 × 109 m3 kmol−1 s−1 Gardoni et al.[14]
k21 5. 0 × 109 m3 kmol−1 s−1 Buxton et al.[47] After the optimization of the reactants and products, the transi-
tion state was calculated by the quasi‐Newton synchronous transit
package. The molecular structures were plotted by the Gauss- (QST3) method and confirmed by the occurrence of a single
View software: imaginary frequency through frequency analyses. Furthermore,
the intrinsic reaction coordinate (IRC) calculations were con-
O3 + HO.2 → 2O2 + HO. (R22) ducted to ensure that the transition state connects the reactants
and products. All of the calculations in the aqueous phase were
Regarding the literature, CCSD(T)/6‐311++G(3df,spd), CCSD/ carried out by using the implicit solvation model based on density
aug‐cc‐pvdz, QCISD/6‐311++(d,p), and QCISD/6‐311++(d) are the (SMD). The relative energy and rate constants were calculated at
commonly used basis sets and methods to optimize the structure 298.15 K and 1 atm (1 atm = 101 325 Pa).
of ozone.[51–53] Since no experimental data has been reported for There are two possible reaction paths for the transition state:
the ozone structures in aqueous phase, although the reaction oxygen atom transfer (HO—O…O—O2); and hydrogen atom transfer
occurs in water, in order to explore which level of theory is more (O2—O…H—O2). Both paths were considered and the results were
suitable for the optimization of the reaction, the optimization of compared with the available rate constants in the literature obtained
ozone in gas phase was performed as well. The most suitable through experimental procedures. These two paths for the transition
method and basis set were determined through a comparison of state were assumed as parallel reactions, occurring simultaneously. In
ozone bond lengths and angles of calculated stable structures in this study, oxygen atom transfer and hydrogen atom transfer path-
gas phase with the corresponding values of the experimental ways are referred to as oxygen transfer and hydrogen transfer.
data.[54] The results of the optimization are shown in Table 2.
Comparing the calculated results (117.2°; 1.259 × 10−10 m) with
the experimental findings in gas phase (116.8°; 1.278 × 10−10 m), Oxygen atom transfer
it was concluded that the bond length and bond angle when using Figure 1 shows the structure of the transition state of the oxygen
the CCSD/aug‐cc‐pvdz level of theory is the closest to the ex- transfer in gas phase. The two O–O bond lengths in the ozone
perimental data; therefore, CCSD/aug‐cc‐pvdz was chosen for molecule are 1.33 × 10−10 m and 1.21 × 10−10 m, whereas the cal-
optimizing other species of the reaction.[54] culated gas phase O–O bond length in Table 3 is ∼1.26 × 10−10 m.

Table 3. Calculated (CCSD/aug‐cc‐pvdz) and experimental bond lengths and angles of different species in gas phase

Bond Length (× 1010 m) Angle (degrees)

Species Bond Experimental Calculated Experimental Calculated

O3 O–O 1.278 1.259 116.80 117.23

HO.2 O–O 1.333 1.357 104.29 105.57
O–H 0.971 1.002
O2 O–O 1.207 1.198 ‐ ‐
HO. O–H 0.970 0.991 ‐ ‐

MONTH 2019 THE CANADIAN JOURNAL OF CHEMICAL ENGINEERING 3

 Figure 3. Structure of transition state for the hydrogen transfer reaction
in gas phase.

Figure 1. Structure of transition state for the oxygen transfer reaction 8 %). Therefore, modelling the transition state with H transfer is
pathway in gas phase. necessary to ensure the validity of the calculated results.
Figure 3 represents the structure of the transition state (hy-
Since in the oxygen transfer pathway one of the O–O bonds in the drogen transfer) in gas phase. The length of the O–H bond in the
ozone breaks down and another becomes a double bond to create transition state is expected to be longer than the O–H bond in
an oxygen molecule, the length of the one bond is expected to be hydroperoxyl radical, as the O–H bond in the hydroperoxyl ra-
longer and another to be shorter than the O–O bond in the ozone. dical breaks down and the O–O bond becomes a double bond to
As Figure 1 represents, the O–O bond length in HO2. is larger than create an oxygen molecule. Figure 3 expresses the length of O–H
the normal O–O bond length in this radical (1.33 × 10−10 m), bond is 1.37 × 10−10 m. Comparing this value to the calculated gas
showing that the O(5)–O(6) bond is being weakened. The H–O bond phase O–H bond length in Table 3 (∼0.99 × 10−10 m) shows that
length remains almost constant because this bond will not change. the O–H bond length in the transition state is longer than the O–H
The relatively long distance between O(2) and O(5) indicates that bond length in the hydroxyl radical as these two atoms are ap-
these two oxygen atoms are approaching each other to create an proaching each other to create an oxygen molecule.
oxygen molecule. Figure 4 expresses the relative energy for the hydrogen transfer
The relative energies (i.e., energies including zero‐point energy reaction pathway in gas and aqueous phases. Based on Figures 2
correction) for the reaction O3 + HO.2 → 2O2 + HO. in the gas and 4, the difference between the energy barriers of the oxygen
and aqueous phase with reactants, transition states, and products transfer reaction pathway in the gas and water phases is larger
were calculated and presented in Figure 2. Since some original compared to the hydrogen transfer. This means that the rate
bonds are weakened and the new bonds have not yet been com- constants of the hydrogen transfer reaction pathway are antici-
pletely formed, the transition complex is not stable and the energy pated to be closer in the gas and aqueous phases compared to the
of the transition state is higher than the energy levels of reactants oxygen transfer reaction pathway.
and products. The formation of new bonds makes the transition
state more stable. The formation of bonds leads to the products
and a state with a lower energy level. Since the energy barrier is REACTION KINETICS
lower in the aqueous solution compared to the gas phase, the
reaction in the aqueous solution is faster. Reaction Rate Constant Calculations
The main reaction pathway can be expressed as follows:

Hydrogen atom transfer O3 + HO.2 ↔ transition state → 2O2 + HO. (R23)

[55]
Nelson and Zahniser, using isotopic exchange studies of the
reaction between ozone and H18O2, have shown that the reaction The reaction rate constants for the elementary reaction was
primarily proceeds by hydrogen transfer instead of oxygen calculated by using transition state theory according to Equation
transfer at room temperature (298.15 K) in gas phase (92 versus (1):[56–58]

Figure 2. Relative energy for oxygen transfer reaction path in gas and Figure 4. Relative energy for hydrogen transfer reaction path in gas and
aqueous phases. aqueous phase.

4 THE CANADIAN JOURNAL OF CHEMICAL ENGINEERING MONTH 2019

 kBT QTS ETS
k=Γ e− RT (1)
h QO3 QHO2
. Table 5. Half‐lives of ozone decomposition in aqueous phase as-
suming different rate constants for Reaction (R22) at an initial ozone
concentration of 5.0 × 10−5 M, pH = 8.5, and 298.15 K
where kB is Boltzmann’s constant; T is temperature; h is Planck’s
constant; ETS is the barrier height with respect to the separate Rate constant Half‐life (s)
reagents; Γ is the tunneling factor; and QTS , QO3 , QHO.2 are the
partition functions of the transition state, ozone, and HO.2, re- Without reaction 836
spectively.[59] The ETS was estimated by the Gaussian 09 Software Experiment upper limit 652
Experiment lower limit 63
and the k value of Reaction (R22) was calculated using Equation
4.6 × 108 67
(1). Table 5 shows the rate constant of Reaction (R22) in the gas 4.6 × 107 94
and aqueous phases considering oxygen and hydrogen transfer in 4.6 × 106 122
the transition state.[34,38–40] Since both parallel paths occur si- 4.6 × 105 183
multaneously, the overall rate constant is the sum of the constants 4.6 × 104 220
of each path. 4.6 × 103 331
As Table 4 represents, the calculated rate constants for the 4.6 × 102 428
reaction in the aqueous and gas phase change considerably by
changing the reaction mechanism. Comparing the obtained
results implies that the H transfer is faster in the gas phase, constant and equal to 8.5. Table 5 shows the half‐lives of ozone at
whereas, in the aqueous phase the O transfer is rapid in different rate constants. Since the first‐order ozone decay rate
comparison with the hydrogen transfer. Consequently, the constant and half‐life time are inversely proportional, this table
overall rate constant is influenced more by the O transfer in provides valuable information about ozone decomposition
water. Although, the calculated overall rate constant in the gas without assuming a reaction order a priori.
phase differs by almost an order of magnitude from the mea- To explore the effect of the obtained rate constant on the rate of
sured value, the calculated selectivity (branching fraction) of the ozone decomposition process and how any error in the cal-
the H transfer is 93 %, almost exactly the experimental value at culated rate constant can influence the entire process, the rate
298.15 K (92 %), measured directly with isotope studies.[39] constant has changed up to ±3 orders of magnitude. As Table 5
Since measurements of the overall rate constant are indirect indicates, including Reaction (R22) in the kinetic model sig-
and carried out in the presence of many competing reactions, nificantly accelerates the decomposition of ozone (more than four
the measured branching ratio is expected to be more accurate times faster). However, changing the rate constants of this reac-
than the measured individual rate constants. Hence, the ac- tion has a comparatively smaller effect on the rate of the entire
curate branching ratio prediction indicates that the calculated process. The rate constant of Reaction (R22) in the aqueous phase
values are accurate. Furthermore, the overall rate constant in measured by experiments is reported to be between 104–7.5 × 106,
the aqueous phase calculated here is within the range of ex- and our kinetic model predicts ∼ ±30 % deviation in the half‐lives
perimental estimations. from the calculated rate constant. Moreover, Table 5 shows that
Based on Figure 2, it was anticipated that the rate constant for for an order of magnitude change in the rate constant, the de-
hydrogen transfer in aqueous phase would be larger than in the composition half‐life changes 20–45 %. Comparing the half‐lives
gas phase as the energy of the transition states in aqueous phase obtained from experiments with the calculated half‐lives with and
were lower than in gas phase. Furthermore, the proximity of the without considering Reaction (R22) shows that adding this reac-
transition states in the water and gas phase for oxygen transfer tion to the kinetic model improves the model and the half‐lives
suggests the rate constant of the reaction in these two phases are fall within the experimental range.[14]
close. The deviation for the oxygen transfer reaction pathway rate
constant between the gas and aqueous phases is larger than that
Detailed Reaction Modelling
for the hydrogen transfer reaction pathway (Table 4), consistent
with relative energy analysis (Figures 2 and 4). To explore whether the calculated rate constant results in an
improvement of the overall kinetics of the system, the developed
kinetic model can be used, considering the new reaction. Due to
Sensitivity Analysis the impact of [OH−], and thus pH, on the overall rate of the ozone
The effect of the Reaction (R22) rate constant on the overall ki- decomposition, it is beneficial to explore the influence of pH on
netic of ozone decomposition in aqueous solutions is studied in a the kinetics of the system. Figure 5 expresses the half‐lives of
sensitivity analysis. Since pH plays a crucial role in the ozone ozone in water at different pH comparing the model predictions
decomposition process due to the impact of hydroxyl ion con- with experimental data reported in the literature.[14] At low pH
centration on the overall rate of reactions, it is assumed to be (< 6), the half‐live of ozone observed in the literature varies over a

Table 4. Experimental and calculated rate constants of the reaction between ozone and hydroperoxyl radical in gas and aqueous phases
(T = 298.15 K)

k (m3 kmol− 1 s−1) k (m3 kmol−1 s−1) k (m3 kmol− 1 s−1) k (m3 kmol− 1 s−1)
Phase O transfer H transfer Overall Experimental
Gaseous 1.0 × 104 1.3 × 105 1.4 × 105 1.2 × 106
Aqueous 3.7 × 105 8.8 × 104 4.6 × 105 104 to 7.55 × 106

MONTH 2019 THE CANADIAN JOURNAL OF CHEMICAL ENGINEERING 5

 other reactions affecting the behaviour of the system. Hence, to
ensure the proposed model is in line with what occurs in rea-
lity, the mechanism of the process should be investigated fur-
ther. The research indicates that ozone degradation chemistry
in aerosol droplets may be more important than is currently
understood.

ACKNOWLEDGEMENTS
The authors acknowledge Westgrid for computational support
and the NSERC Discovery Grants program for providing the fi-
nancial aid.

Figure 5. Half‐lives of ozone versus pH at 20 °C: literature experimental

REFERENCES
data (markers);[14] and modelling results without and with the modified [1] A. Irabelli, S. Jasim, N. Biswas, Ozone‐Sci. Eng. 2008, 30, 356.
rate constants (curves) at initial ozone concentration of 5 × 10−5 M. [2] S. Youmin, R. Xiaohua, C. Zhaojie, Z. Guiqin, J. Mol. Model.
2012, 18, 3821.
[3] J. Hoigné, The Chemistry of Ozone in Water, 2nd edition,
wider range compared to the higher pH values.[14] The upper limit Springer US, Boston 1988.
literature data in Figure 5 were obtained with a sulphate buffer,
[4] J. B. Duncan, Aqueous Reactions of Specific Organic Com-
whereas the lower limit data were obtained with a phosphate
pounds with Ozone, John Wiley & Sons, New York 2005.
buffer. The curves named Without R22 and With R22 refer to the
behaviour of the system without considering and considering [5] J. Prado, S. Esplugas, Ozone‐Sci. Eng. 1999, 21, 39.
Reaction (R22). [6] C. Zwiener, F. H. Frimmel, Water Res. 2000, 34, 1881.
As Figure 5 indicates, the calculated rate constant brings the [7] W. H. Glaze, F. Beltran, T. Tuhkanen, J.‐W. Kang, Water
half‐time curve into the range obtained by experiments within 6 < Pollution Research Journal Canada 1992, 27, 23.
pH < 8. Without this modification, the curve is within the range [8] C. Selçuki, V. Aviyente, P. Aplincourt, M. F. Ruiz‐López,
only in alkaline solutions (pH > 8). The ozone concentration in Molecular Modeling Annual 2000, 6, 608.
the model was assumed to be 5 × 10−5 M , at the lower end of the [9] J. Staehelin, J. Hoigne, Environ. Sci. Technol. 1982, 16, 676.
experimental concentration range. Increasing the concentration
[10] K. Sehested, H. Corfitzen, J. Holcman, C. H. Fischer, E. J.
by a factor of 10, lowers the curve by a factor of 101/2 at low pH
Hart, Environ. Sci. Technol. 1991, 25, 1589.
and does not affect the half‐life at high pH. Since the calculated
curve reproduces the slope of the upper limit data, the change in [11] A. N. Ignatiev, A. N. Pryakhin, V. V. Lunin, Russ. Chem.
the half‐life of ozone using the calculated rate constant is justi- Bull.+ 2008, 57, 36.
fiable. It is concluded that Reaction (R22) considerably improves [12] J. Hoigné, H. Bader, W. R. Haag, J. Staehelin, Water Res.
the ozone decomposition system within 6 < pH < 8. 1985, 19, 403.
[13] B. G. Ershov, P. A. Morozov, Russ. J. Appl. Chem.+ 2008, 81, 87.
[14] D. Gardoni, A. Vailati, R. Canziani, Ozone‐Sci. Eng. 2012,
CONCLUSIONS 34, 233.
This paper presented a theoretical approach for filling the [15] J. Weiss, T. Faraday Soc. 1935, 31, 1547.
knowledge gap of the mechanism of ozone decomposition in [16] J. Hoigné, H. Bader, Water Res. 1976, 10, 377.
aqueous phase around neutrality by developing a chemical ki-
[17] Y. Liu, W. Yang, P. Zhang, J. Zhang, Appl. Surf. Sci. 2018,
netic model and adding a chemical reaction with a rate constant
442, 640.
calculated using quantum chemical reaction calculations. Values
of 1.4 × 105 and 4.6 × 105 m3 kmol−1 s−1 were found in gas [18] E. Reisz, C. von Sonntag, A. Tekle‐Röttering, S. Naumov, W.
and aqueous phases, respectively, for the rate constant of Schmidt, T. C. Schmidt, Water Res. 2018, 128, 171.
O3 + HO.2 → OH. + 2O2 . The calculated rate constant in the aqu- [19] P. Deeudomwongsa, S. Phattarapattamawong, K.‐Y. Andrew
eous phase is used in the kinetic model and in a sensitivity Lin, Chemosphere 2017, 184, 105.
analysis to explore the influence of adding this reaction and the [20] R. Pandiselvam, S. Sunoj, M. R. Manikantan, A. Kothakota,
effect of this rate constant on the rate of ozone decomposition. K. B. Hebbar, Ozone-Sci. Eng. 2017, 39, 115.
The results showed that it improves the model, and for an order of [21] A. J. Brodowska, A. Nowak, K. Śmigielski, Crit. Rev. Food Sci.
magnitude change in the rate constant, the decomposition half‐ 2018, 58, 2176.
life changes 20–45 %. [22] L. Hu, Z. Xia, J. Hazard. Mater. 2018, 342, 446.
The developed model is relatively sensitive to the rate con-
[23] W. Shen, Y. Wang, J. Zhan, B. Wang, J. Huang, S. Deng, G.
stant of the added reaction. The overall kinetics of the ozone
decomposition is complicated and highly ozone dependent at Yu, Chem. Eng. J. 2017, 310, 249.
low pH, and dependent on the acid or buffer used to control the [24] S. Phattarapattamawong, A. M. Kaiser, E. Saracevic, H. P.
pH, so that different reaction orders are reported in this range. Schaar, J. Krampe, Water Sci. Technol. 2018, 2017, 404.
Thus, although the reaction of ozone with HO.2 should be [25] G. Qin, J. Huang, Y. Pan, Y. Xiang, C. Ou, J. Huang, L. Gao, J.
considered in the mechanism of the ozone decomposition, Lu, Journal of Central South University (Medical Sciences)
which had not been taken into account before, there might be 2018, 43, 163.

6 THE CANADIAN JOURNAL OF CHEMICAL ENGINEERING MONTH 2019

[26] S. Menendez‐Cepero, “Ozone Therapy in Leg Ulcers of Pa- [52] L. Vereecken, A. R. Rickard, M. J. Newland, W. J. Bloss, Phys.
tients with Chronic Venous Insufficiency,” The 5th WFOT Chem. Chem. Phys. 2015, 17, 23847.
Meeting, World Federation of Ozone Therapy, India, 18–20 [53] H. Tachikawa, S. Abe, Inorg. Chem. 2003, 42, 2188.
November 2018. [54] “NIST Chemistry WebBook,” National Institute of Standards
[27] S. Menendez‐Cepero, O. S. Leon, G. Martinez‐Sanchez, H. and Technology (NIST), 2018, accessed on 16 June 2018,
Alvarez, “Ozone Therapy in Diabetes and its Complications,” https://webbook.nist.gov/chemistry/.
The 5th WFOT Meeting, World Federation of Ozone Therapy, [55] D. D. Nelson, M. S. Zahniser, J. Phys. Chem. 1994, 98, 2101.
India, 18–20 November 2018.
[56] L. Bonnet, J. C. Rayez, Int. J. Quantum Chem. 2010,
[28] C.‐H. Kuo, L. Zhong, M. E. Zappi, A. P. Hong, Can. J. Chem. 110, 2355.
Eng. 1999, 77, 473.
[57] C. Wu, A. De Visscher, I. D. Gates, RSC Adv. 2017, 7, 35776.
[29] M. E. Lovato, C. A. Martín, A. E. Cassano, Chem. Eng. J. 2009,
[58] H. Eyring, J. Chem. Phys. 1935, 3, 107.
146, 486.
[59] Y. Zhang, J. B. Rommel, M. T. Cvitas, S. C. Althorpe, Phys.
[30] A. A. Peyghan, M. Moradi, J. Mol. Model. 2014, 20, 2071.
Chem. Chem. Phys. 2014, 16, 33021.
[31] B. G. Ershov, P. A. Morozov, Russ. J. Phys. Chem. A 2009,
83, 1795.
[32] Y. Ku, W.‐J. Su, Y.‐S. Shen, Ind. Eng. Chem. Res. 1996,
Manuscript received March 11, 2019; revised manuscript received
35, 2110.
April 24, 2019; accepted for publication April 29, 2019.
[33] M. L. Kilpatrick, C. C. Herrick, M. Kilpatrick, J. Am. Chem.
Soc. 1956, 78, 2483.
[34] K. Sehested, J. Holcman, E. Bjergbakke, E. J. Hart, J. Phys.
Chem. 1984, 88, 20.
[35] W. L. Chameides, D. D. Davis, J. Geophys. Res.‐Oceans 1982,
87, 4863.
[36] R. C. Whitten, W. J. Borucki, L. A. Capone, R. P. Turco,
Nature 1978, 275, 523.
[37] J. Holcman, K. Sehested, Radiat. Phys. Chem. 1996, 47, 357.
[38] T. E. Graedel, M. L. Mandich, C. J. Weschler, J. Geophys. Res.‐
Atmos. 1986, 91, 5205.
[39] R. Atkinson, D. L. Baulch, R. A. Cox, J. N. Crowley, R. F.
Hampson, R. G. Hynes, M. E. Jenkin, M. J. Rossi, J. Troe, J.
Troe, Atmos. Chem. Phys. 2004, 4, 1463.
[40] J. B. Burkholder, S. P. Sander, J. P. D. Abbatt, J. R. Barker, R.
E. Huie, C. E. Kolb, M. J. Kurylo, V. L. Orkin, D. M.
Wilmouth, P. H. Wine, Jet Propulsion Laboratory 2015,
18, 254.
[41] J. Foresman, E. Frisch, Exploring Chemistry with Electronic
Structure Methods, 2nd edition, Guassian Inc, Bitola 1996.
[42] D. C. Young, Computational Chemistry, John Wiley & Sons,
Inc., San Francisco 2001.
[43] Gaussian 09, Revision A.02, Gaussian Inc., Wallingford
2016, https://www.westgrid.ca/support/software/gaussian.
[44] H. Sievering, E. Gorman, T. Ley, A. Pszenny, M. Springer‐
Young, J. Boatman, Y. Kim, C. Nagamoto, D. Wellman, J.
Geophys. Res.‐Atmos. 1995, 100, 366.
[45] J. Yu, D. R. Cocker, R. J. Griffin, R. C. Flagan, J. H. Seinfeld, J.
Atmos. Chem. 1999, 34, 207.
[46] P. J. Ziemann, Faraday Discuss. 2005, 130, 469.
[47] G. V. Buxton, C. L. Greenstock, W. P. Helman, A. B. Ross, J.
Phys. Chem. Ref. Data 1988, 17, 513.
[48] R. E. Buehler, J. Staehelin, J. Hoigne, J. Phys. Chem. 1984,
88, 2560.
[49] A. Fischbacher, J. von Sonntag, C. von Sonntag, T. C.
Schmidt, Environ. Sci. Technol. 2013, 47, 9959.
[50] H. Christensen, K. Sehested, H. Corfitzen, J. Phys. Chem.
1982, 86, 1588.
[51] L. P. Viegas, A. J. C. Varandas, J. Phys. Chem. B 2016,
120, 2376.

MONTH 2019 THE CANADIAN JOURNAL OF CHEMICAL ENGINEERING 7