Content uploaded by Roberto Rosal
Author content
All content in this area was uploaded by Roberto Rosal
Content may be subject to copyright.
Hdb Env Chem Vol. 5, Part S/2 (2008): 127–175
DOI 10.1007/698_5_103
©Springer-Verlag Berlin Heidelberg
Published online: 7 February 2008
Ozone-Based Technologies in Water
and Wastewater Treatment
A. Rodríguez1·R.Rosal
1·J.A.Perdigón-Melón
1·M.Mezcua
2·A.Agüera
2·
M. D. Hernando1·P.Letón
1·A.R.Fernández-Alba
2(u)·E.García-Calvo
1
1Department of Chemical Engineering, University of Alcalá,
28871 Alcalá de Henares, Madrid, Spain
2Department of Analytical Chemistry, University of Almería, 04120 Almería, Spain
amadeo@ual.es
1 Fundamentals of Ozonation Processes .................... 129
1.1 TheMoleculeofOzone ............. ................ 129
1.2 SolubilityofOzoneinWater .......................... 130
1.3 OzoneMassTransfer .............................. 131
1.4 DecompositionofOzoneinWater....................... 134
1.5 OzoneReactionswithOrganicCompounds.................. 138
2 Ozone Uses in Water Treatment ........................ 140
2.1 PrecipitationofOxides ............................. 141
2.2 DisinfectionofDrinkingWater......................... 143
2.3 NaturalWaterandWastewaterTreatment................... 145
2.4 CatalyticOzonation............................... 153
2.4.1HomogeneousCatalyticOzonation ...................... 155
2.4.2CatalysisbyMetalsandMetalOxides ..................... 156
2.5 ApplicationsintheTreatmentofIndustrialWastewater ........... 163
2.6 Removal Efficiency of Pharmaceuticals in Wastewater: A Case Study . . . . 166
3Conclusions................................... 171
References ....................................... 172
Abstract Ozone is a strong oxidant that can be used in the potabilization of surface or
ground water as well as in wastewater treatment to remove microorganisms, inorganic
ions and organic pollutants. The oldest use of ozone is as a biocide in drinking water
potabilization. The integral ozone exposure required for a given degree of disinfection
can be calculated from the deactivation kinetic constant of the microorganism. Ozone re-
moves iron, manganese and arsenic from water by oxidation to an insoluble form that is
further separated by filtration. Both processes require ozone in molecular form, but the
removal of organic pollutants that are refractory to other treatments can be possible only
by exploiting the indirect radical reactions that take place during ozonation. Ozone de-
composes in water, especially when hydrogen peroxide is present, to yield the hydroxyl
radical, the strongest oxidizer available in water treatment. Models for the ozonation pro-
cess are required to adjust the ozone dosing to the desired degree of removal of a given
pollutant or an aggregate measure of pollution. Mineralization, defined as the removal of
organic carbon, has been accomplished in wastewaters from urban and domestic treat-
ment plants. The results show that the logarithmic decrease of TOC as a function of the
integral ozone exposure usually presents two zones with different kinetic parameters.
128 A. Rodríguez et al.
Among advanced oxidation processes, a promising alternative currently under develop-
ment is the use of ozone in combination with solid catalysts. The mechanism of catalytic
ozonation is not clear, but in the case of metal oxides, the adsorption of ozone or or-
ganic compounds on Lewis acid sites is only possible near the point of zero charge of the
surface. Activated carbon seems to behave as an initiator of ozone decomposition, a role
that may also occur with other types of catalysts. Some results on the mineralization
of water with the drugs naproxen (non-steroidal anti-inflammatory) and carbamazepine
(anticonvulsant) are presented using titanium dioxide as catalyst.
Keywords Advanced oxidation processes ·Disinfection ·Kinetic models ·Ozonation ·
Solid catalysts
Abbreviations
aSpecific gas–liquid interfacial area [m–1]
Alk Alkalinity [mg CaCO3L–1]
cAConcentration of a given compound [M]
CO3Concentration of dissolved ozone in water [M]
C∗
O3Equilibrium concentration of dissolved ozone in water [M]
csBulk concentration of catalyst [kg m–3 ]
ctConcentration of surface sites of catalyst [m ol kg–1]
ctO3,ct10 Concentration–time exposure parameter for ozone [M s]
dbBubble diameter [m]
DO2Diffusivity of oxygen [m2s–1]
DO3Diffusivity of ozone [m2s–1]
EEnhancement factor
Ha Hatta number
HeHenry’s law constant [atm mole fraction–1]
iIonic strength [M–1]
k1,k2Rate constants for the catalytic decomposition of ozone [m3kg–1 s–1]
kaKinetic constant of adsorption [L kg–1
cat s–1]
k–a Kineti c constant of desor ption [mol kg–1
cat s–1]
kcKinetic constant of catalytic ozonation [L kg–1
cat s–1]
kdKinetic constant of ozone decomposition [units depending on the order of
reaction]
kD,kDi Kinetic constants for direct reaction w ith ozone [L mol–1 s–1 ]
kHO·Kinetic constant for reactions with hydroxyl radical [L mol–1 s–1]
kHO–Kinetic constants of the hydroxide initiation of ozone decomposition [M–1 s–1 ]
kHO2–Kinetic constants of the hydroperoxide initiation of ozone decomposition
[M–1 s–1]
kLLiquid phase individual mass transfer coefficient [m s–1 ]
kLaVolumetric mass transfer coefficient [s–1]
kNKinetic constant for microorganism deactivation [M–1 s–1 ]
koKinetic constant of the surface oxidation process [Lkg–1
cat s–1]
kO3Kinetic constant for direct reaction with ozone [L mol–1 s–1]
krKinetic constant of termination reactions [Lmol–1 s–1]
KaAdsorpt ion equ ilibrium const ant [L mol–1 ]
Kox Equilibrium constant for the surface oxidation process [L mol–1]
NO3Absorption rate or flux of ozone [mol m–2 s–1]
pHPZC pH of the point of zero charge of a surface
Ozone-Based Technologies in Water and Wastewater Treatment 129
PO3Partial pressure of ozone in gas [Pa]
rdRate of ozone decomposition [mol m–3 s–1]
RKinet ic consta nt for TOC removal dur ing ozonation [L mol–1 s–1]
Rct Hydroxyl ozone ratio defined by Eq. 29
Sc Schmidt number [µLρ–1
LD–1
O3]
TOC Total organic carbon [mg L–1]
TOCcOrganic carb on refra ctory to ozonation [mg L–1]
TOC∗
cOrganiccarboninoxalate,acetateandformiate[mgL
–1]
TOCoInitial total organic carbon [mg L–1]
TOD Tot al ozone dos e t ransferred [mol L–1]
ugSuperficial gas velocity [m s–1]
XOzone dose transfer at the beginning of the ozonation [mol L–1 ]
zStoichiometric coefficient
Greek letters
εgGas holdup
µLLiquid v iscosity [kg m–1 s–1]
ρLLiquid density [kg m–3]
σLSurface tension [N m–1]
τHydraulic retention time [s]
θUnit fraction of catalyst occupied sites
1
Fundamentals of Ozonation Processes
1.1
The Molecule of Ozone
Ozone is a bluish coloured gas with a boiling point of 161.3 K(–111.9 ◦C)
and a melting point of 80.7 K(–192.5 ◦C). Mixtures of ozone and oxygen with
more than 20% ozone become explosive. In practice, the risk of explosion
does not exist because corona discharge commercial ozone generators pro-
duce much lower concentrations.
From microwave spectroscopy, it has been shown that the molecule of
ozone has an O–O distance of 127.2 pm and an O–O–O angle of 116.78◦.The
structure of the ozone molecule has been represented by resonance theory by
two main equal weighting open structures plus a cyclic form revealed by the
electron diffraction method (Fig. 1).
Fig. 1 Resonance structures for the molecule of ozone
130 A. Rodríguez et al.
Table 1 Oxidation potential for common oxidants referred to a normal hydrogen electrode
Oxidant Potential E0(V, 25 ◦C)
Fluoride 3.06
Hydroxyl radical 2.80
Atomic oxygen 2.42
Ozone 2.07
Hydrogen peroxide 1.78
Permanganate 1.68
Chlorine dioxide 1.57
Hypochlorous acid 1.49
Chlorine 1.36
Oxygen 1.23
The central atom in the open structures forms an sp2hybridization with
one lone pair and positive charge that explains the strong electrophilic be-
haviour of the molecule. Ozone has a dipole moment of 0.5337 D, a conse-
quence of the electron density of the open structures that strongly influences
the chemistry of ozone. Ozone is a very reactive molecule, with a redox oxi-
dation potential of 2.07 V. In fact it is one of the strongest oxidizers available
for water treatment (Table 1).
1.2
Solubility of Ozone in Water
The rate and extent of oxidation/mineralization of water pollutants depends
on the concentration of dissolved ozone, CO3. It is, therefore, an essential
parameter in the design of water treatment facilities. The ozone mass bal-
ance in a volume element of aqueous phase during an isothermal ozonation
process controlled by the chemical step is shown in Eq. 1. The value of CO3
is determined by the ozone solubility in water, C∗
O3, the volumetric trans-
fer coefficient, kLa, and the ozone decomposition kinetic constant, kd,dueto
reactions between ozone and water and the compound dissolved in it:
dCO3
dt=kLaC∗
O3–CO3–kdCn
O3.(1)
At low pressure, ozone is only slightly soluble in water and if ideal gas be-
haviour and a negligible ozone transfer resistance in the gas phase are as-
sumed, the relationship between the partial pressure of ozone, PO3,andits
solubility in water can be expressed by Henry’s law:
PO3=C∗
O3He.(2)
Ozone-Based Technologies in Water and Wastewater Treatment 131
Due to decomposition of ozone in water, the experimental determination of
parameters from Eq. 1 is not easy. It is usual to calculate C∗
O3by means of Eq. 2
taking Hevalues from published correlations, such as those of Andreozzi
et al. and Rischbieter et al. [1, 2]:
log He=A–B
T,(3)
where Aand Bare parameters that depend on the ionic strength of the solu-
tion; Roth and Sullivan [3], where He(atm mole fraction–1) is expressed as
afunctionoftemperatureandpHofwater:
He=3.84 ×107C0.035
OH–exp –2428
T,(4)
or Sotelo et al. [4], in which He(kPa mole fraction–1) depends on tempera-
ture, pH, ionic strength (i) and type of salt dissolved in water:
He=1.03 ×109C0.012
OH–exp –2118
Texp(0.96 i). (5)
Equation 5 corresponds to sodium phosphate solutions and 0 ≤T≤20 ◦C,
2≤pH ≤8.5, and 10–3 M≤i≤10–1 M. The C∗
O3values estimated from these
or similar equations are close to the ozone solubility values in real wastew-
ater, although in cases where an important deviation between the estimated
and real solubility values is expected, the Henry’s law constant must be exper-
imentally measured [5].
1.3
Ozone Mass Transfer
The absorption rate of ozone in water, NO3, can be expressed as:
NO3a=kLaC∗
O3–CO3,(6)
where kLis the ozone mass transfer coefficient and athe specific gas–liquid
interfacial surface inside the ozonation reactor. As indicated in the preced-
ing section, the concentration of dissolved ozone depends also on the rate of
ozone decomposition, rd:
rd=kdCn
O3.(7)
Expressions similar to Eq. 7 permit an easy estimation of the ozone consump-
tion rate in complex systems, such as those that occur in the ozonation of
wastewater (see Sect. 2.4). The parameter kdin Eq. 7 is not a real kinetic con-
stant because, besides temperature, its value depends on the properties of the
water matrix: organic and inorganic matter dissolved, pH, alkalinity and ionic
strength.
132 A. Rodríguez et al.
The ozonation kinetics may be controlled either by physical absorption or
by chemical reaction. The value of the Hatta number gives a rule to determine
the rate-controlling process for a set of given conditions [6]. The Hatta num-
ber is calculated with the values of kLand kdand assuming the double film
model of Lewis–Whitman [7]:
Ha =2
n+1DO3kdCn–1
O3
kL,(8)
where DO3is the ozone diffusivity in water and nthe ozone decomposition
kinetic order. For Ha < 0.3, the rate of ozone absorption is higher than
the ozone decomposition rate and, therefore, chemical kinetics controls the
ozonation process (Eq. 1). Operation conditions should guarantee that the
process is controlled by the chemical reaction step in order to provide a max-
imum flow of oxidant. At operational conditions with Ha > 0.3 values, the
ozonation decomposition rate is so high that the concentration of ozone in
water cannot be measured and the absorption step controls the overall ozona-
tion process. In these cases, the ozone transfer model must take into account
the contribution of the chemical reaction to the absorption expressed by the
enhancement factor (E), either calculated by the general approach [7, 8] or by
means of experiments [9]:
NO3a=kLaC∗
O3E.(9)
The diffusivity of ozone can be calculated by the Wilke–Chang type corre-
lation of Haynuk et al. or by means of ozone-specific expressions, such as
those proposed by Matrozov et al. (A=4.27 ×10–10 )orJonsonandDavis
(A=5.9 ×10–10) with the following expression [10]:
DO3=AT
µL,(10)
where DO3is in m2s–1,Tin K and µL, the solution viscosity, in poise. The
mass transfer coefficient, kL, can be estimated from equations such as that
proposed by van Dierendonck for stirred tanks in which µLand ρL,thevis-
cosity and density of the aqueous solution, are expressed in SI units:
kL=0.42 3
µLg
ρLSc–0.5 .(11)
In bubble columns and for bubble sizes db<2mm, Calderbank proposed the
same equation to estimate kLand Eq. 12 for bubbles with db>2mm:
kL=kL(db=2mm)500 db.(12)
The bubble diameter can be estimated from the operation parameters ug
(superficial gas velocity) and εg(gas holdup into the column) and the liquid-
Ozone-Based Technologies in Water and Wastewater Treatment 133
phase properties ρL(density) and σL(surface tension):
61–εg
db=2ρLg
σL0.5 ug
σLg
ρL0.25 .(13)
If all bubbles are spheres with the same size, acan be calculated by:
a=6εg
db.(14)
Although in Eqs. 10–14 the contribution of pollutants present in wastewa-
ter has not been taken into account, the values of kL,DO3and aobtained
from them will be used to characterize the transfer phenomenon taking place
in ozonation processes. According to Beltrán [11], kL,DO3and acan be ex-
perimentally determined in the wastewater where ozonation processes take
place provided the appropriate kinetic regime is chosen. The experiments
performed to determine kLaand kdconsist in bubbling a continuous gas
flow containing ozone through the wastewater in a stirred tank or bubble
column where ozonation takes place. Figure 2 shows the concentration of dis-
solved ozone during the ozonation of a wastewater (Table 5, D070208) from
the secondary clarification of a municipal wastewater treatment facility. The
experiment was carried out at 25 ◦Cina5-L stirred tank agitated at 1000 rpm
with a four-blade turbine. The gas, a mixture of ozone and oxygen with
a45.9 gNm
–3 ozone concentration, was bubbled at a rate of 0.36 Nm3h–1.
During the experiment the pH was in the range 8.04–8.25. Three different
zones can be appreciated in Fig. 2. Zone I is characterized by a strong increase
in ozone dissolved concentration and is followed by zone II, where the ozone
concentration reaches a stationary value, CO3s. In zone III the concentration
of ozone decays after stopping the gas flow.
Assuming that the decomposition of ozone follows first-order kinetics,
Eq. 1 applied to zones II and III leads to the following expressions:
0=kLaC∗
O3–CO3s–kdCO3s,(15)
dCO3
dt=– kdCO3.(16)
The integration of Eq. 16 yielded kd. The solubility of ozone was calculated
as indicated in Sect. 1.2 and the value of kLawas obtained from Eq. 15.
The experimental value of kLacanbeuptotwoorfivetimeshigherthan
the corresponding estimation from literature correlations due to the specific
composition of wastewater [5].
For the operational conditions of the experiment that is represented in
Fig. 2, the calculated values of C∗
O3,kLaand kdwere 0.247 mM, 0.614 min–1
and 0.139 min–1, respectively. (The equation from Rischbieter et al. (A= 5.12,
B=1230 K–1) was used to determine He.)
134 A. Rodríguez et al.
Fig. 2 TOC and CO3values during the ozonation of D070208 wastewater (Table 5). pH:
8.04–8.25, T:25 ◦C, gas flow rate: 0.36 Nm3h–1, gas ozone concentration: 45.9 g/Nm3,
kLa=0.614 min–1
The decomposition of ozone in water makes the experimental determin-
ation of kLacomplex. To overcome this, and taking the surface renewable
theories into account, the ozone mass transfer coefficient can be based on the
corresponding value of some less reactive compound such as oxygen at the
same pressure and temperature:
kLaO3=kLaO2DO2
DO30.5
.(17)
1.4
Decomposition of Ozone in Water
Ozone in aqueous solution decomposes through a complex mechanism initi-
ated by reaction with a hydroxide ion and followed by formation of several
radical oxidizing species, such as HO, HO2and HO3. The structures of ozone
and HOxin liquid water remain uncertain. Chalmet and Ruiz-López [12]
combined quantum and classical computer simulations and showed that even
if ozone undergoes electron polarization, it does not participate in hydrogen
bonds with liquid water. In contrast, HOxform strong hydrogen bonds, being
better proton donors but weaker proton acceptors than water. Their electronic
and geometrical structures are significantly modified by the solvent, suggest-
Ozone-Based Technologies in Water and Wastewater Treatment 135
ing that water plays a crucial role in oxidation mechanisms initiated by ozone
in liquid water.
Concerning the mechanism and kinetics of ozone decomposition, the reac-
tion follows a chain process extensively studied by Buhler et al. [13], Staehelin
et al. [14], Tomiyasu et al. [15] and Hoigné [16]. In the absence of UV radia-
tion or solid catalysts, the initiation takes place through a reaction between
ozone and the hydroxide ion to yield a hydroperoxide (HO2·)andasuperox-
ide radical ion (O2·–):
O3+HO
–ki,1
–→HO2·+O
2·–kOH–=70 M–1 s–1 .
In the presence of hydrogen peroxide, initiation takes place by reaction of
ozone with the hydroperoxide ion, HO2–,theconjugatebaseofhydrogenper-
oxide:
O3+HO
2–ki,2
–→HO2·+O
3·–kHO2–=2.2 ×106M–1 s–1 .
Propagation involves the formation of ozonide radical ion O3·–,theradical
species HO3·and HO4·and several reactions of hydrogen peroxide, an inter-
mediate product of the degradation chain:
HO2·O2·–+H
+,
O3+O
2·–→O3·–+O
2,
O3·–+H
+HO3·,
HO3·→HO·+O
2,
O3+HO
·→HO4·,
HO4·→HO2·+O
2,
HO2–+H
+H2O2,
HO·+H
2O2→HO2·+H
2O,
HO·+HO
2–→HO2·+HO
–.
Homogeneous termination takes place by reactions consuming radicals:
HO·+O
3→O3+HO
–,
HO4·+HO
4·→H2O2·+2O
3(tentatively proposed) ,
HO4·+HO
3·→H2O2·+O
2+O
3(tentatively proposed) .
There are a wide variety of compounds able to promote or inhibit the chain-
reaction processes. Promoters of the free-radical reaction are substances
capable of regenerating the superoxide anion from the hydroxyl radical. Com-
mon organic promoters include formic and glyoxylic acids, primary alcohols
and humic acids. The inhibitors of the free-radical reaction are compounds
capable of consuming hydroxyl radicals without regenerating the superoxide
anion. These include bicarbonate and carbonate ions, tertiary alcohols like
tert-butanol and some humic substances [11, 17].
136 A. Rodríguez et al.
The formation of hydroxyl radicals from ozone can be enhanced by the
presence of solid catalysts. In the case of metal oxides, heterogeneous ozone
decomposition is determined by the presence of surface hydroxyl groups
acting as Brönsted acid sites. These sites also determine the charge of the sur-
face as a function of pH, and therefore the ion-exchange behaviour of the
catalyst. In addition to this, metal oxides have Lewis acid sites that, in an
aqueous solution, allow water molecules to coordinate on their surface [18].
The adsorption of ozone requires the displacement of coordinated water and
is strongly dependent on the presence of other bases. In the case that a Lewis
site is accessible to ozone, the mechanism for its adsorption/decomposition
on a catalytic surface would follow a mechanism similar to that used for ex-
plaining gas-phase decomposition [19]:
O3→(O3)ads (i)
(O3)ads →(O)ads +O
2(ii) .
The interaction of the ozone molecule with an oxidized site may yield ad-
sorbed or non-adsorbed oxygen:
O3+(O)ads →2O2,
O3+(O)ads →O2+(O2)ads →2O2(iii) .
In aqueous solution, the hydroxide ion is expected to act as a strong inhibitor
of the adsorption ability of the catalyst by blocking Lewis acid sites. There-
fore, the catalytic activity at high pH should proceed by a redox mechanism
involving surface hydroxyl groups. Ozone would react with them to yield an
ozone anion radical or another active species able to oxidize organic com-
pounds either in solution or on the surface.
Activated carbon is particularly efficient as an initiator in the decompos-
ition reaction of ozone in the liquid phase [20]. The capacity of activated
carbon to transform ozone into hydroxyl radicals depends on its surface prop-
erties. It has been demonstrated that metal centres, electrons from graphenic
layers and basic surface groups like chromene and pyrone are active sites for
ozone adsorption [21]. These basic Lewis sites are located at π-electron-rich
regions and behave as a Lewis base in aqueous solution [22]:
πC+2H
2O→πC–H3O++OH
–.
The molecule of ozone may attack the basic delocalized π-electron system or
lone pairs in pyrrolic groups with the generation of hydroxyl radicals [23]:
···NH + O3→···N+O–+HO
2·,
HO2·→H++O
2·–,
O2·–+O
3→O3·–+O
2,
O2·–+H
+→HO·+O
2.
Ozone-Based Technologies in Water and Wastewater Treatment 137
The generation of radicals from the interaction between ozone and activated
carbon has been studied by the Rct methodology using pCBA as probe com-
pound [24]. Sánchez-Polo et al. [21] showed that the interaction between
ozone and groups on the surface of activated carbon leads to an increase
of the concentration of superoxide radical ion enhancing ozone transform-
ation into hydroxyl radicals. As the activity of activated carbon decreased
with ozone exposure, it has been suggested that activated carbon does not be-
have as a true catalyst but rather as a conventional initiator or promoter for
the ozone transformation into radicals.
Figure 3 shows the transient response of dissolved ozone concentration
after charging a semicontinuous reactor with a catalyst concentration of
0.5 g/L. The mixture of ozone and oxygen was bubbled into the liquid by
means of a porous glass disc with a total gas flow of 240 NL/h. The catalysts
used were titanium dioxide Degussa P25 and activated carbon (AC). The con-
centration of ozone in the liquid was measured using a Rosemount 499A OZ
ozone amperometric sensor equipped with Pt 100 RTD temperature compen-
sation and checked against the Indigo Colorimetric Method (SM 4500-O3B).
The signal was continuously monitored by means of a Rosemount 1055 Dual
Input Analyzer connected to an Agilent 34970A data acquisition system.
The unsteady-state catalytic decomposition of ozone can be modelled as-
suming that simultaneous non-catalytic reaction follows a first-order kinetic
expression. Ozone was supposed to adsorb on the surface of titanium dioxide,
Fig. 3 Transient state decomposition of ozone at 25 ◦C after introducing 0.5 g/LofTiO
2
and AC while bubbling ozone (40–42 g/Nm3at 240 NL/handpH=5)
138 A. Rodríguez et al.
so that its decomposition takes place according to the mechanism based on
gas-phase reaction and described before. In the absence of data on adsorbed
and non-dissociated ozone intermediates, the first reaction can be described
as: O3→(O)ads +O
2. A further ozone molecule reacts with the oxidized site
to yield non-adsorbed products. The concentration of ozone can be calculated
by solving the following system of differential equations:
dCO3
dt=kLaE C∗
O3–CO3–k1csCO3(1 – θ)–k2csCO3θ–kdCO3,(18)
–ctdθ
dt=–k1CO3(1 – θ)+k2CO3θ,(19)
where θis the fraction of catalyst occupied surface sites, k1and k2are
the rate constants for the catalytic reactions (i + ii) and (iii) and csis
the bulk concentration of solids. Rosal et al. [25] reported the following
kinetic constants at pH = 5 and 20 ◦C for ozone decomposition on ti-
tanium dioxide: k1=7.21 ×10–3 ±3.1 ×10–4 m3kg–1 s–1,k2=2.73 ×10–4 ±
2.5 ×10–5 m3kg–1 s–1 and kd=8.74 ×10–3 ±1.3 ×10–4 s–1.Linetal.[19]
compared average rates of decomposition of aqueous ozone, showing that
oxides with lower ∆hfexhibit higher activities but always lower than those of
noble metals and much lower than those of activated carbon.
Another factor that has been pointed out is the fact that fine catalyst
particles may enhance the absorption of ozone by a “shuttle” mechanism in-
volving the physical adsorption of ozone on the surface of particles [26]. For
P25 titanium dioxide the maximum enhancement, denoted by Ein Eq. 18,
represented three times the mass transfer rate of ozone in a particle-free li-
quid [25].
1.5
Ozone Reactions with Organic Compounds
Ozonation may take place by the direct reaction of the ozone molecule with
the target compound or by means of hydroxyl radicals produced from the
decomposition of ozone in aqueous media. It has already been stated that
hydroxyl radicals are strong secondary oxidants produced as a consequence
of ozone self-decomposition in water. In practice, both direct and indirect
reactions take place simultaneously, but when an oxidation process is specif-
ically designed to enhance the concentration of HO·radicals in a solution,
one speaks of an advanced oxidation process (AOP). The data of Acero and
von Gunten [27] and Buffle et al. [28, 29] allow some insight into the order
of magnitude of the concentration of both oxidants in an ozonation pro-
cess. These researchers found that the ratio of the concentration of hydroxyl
radicals to dissolved ozone ranges from 10–6 to 10–8, the former value be-
ing encountered in AOP while the latter is typical from the ozonation of
drinking water. The hydroxyl concentration during the early stages of ozone
Ozone-Based Technologies in Water and Wastewater Treatment 139
decomposition in water is greatly enhanced by the presence of amines or
phenols through the formation ozonide (O3·–)orsuperoxide(O
2·–) radical
anions [23].
The direct reactions of ozone with organic compounds in aqueous so-
lutions are essentially limited to those taking place with unsaturated and
aromatic compounds and are governed by the dipolar structure of the ozone
molecule. The 1,3-cycloaddition to unsaturated compounds leads to the for-
mation of a primary ozonide:
In a protonic solution, the primary ozonide decomposes via a zwitterion
that yields a hydroperoxide. This three-step process is called the Criegge
mechanism.
Aromatic compounds do not undergo cycloaddition. Instead, the ozone
molecule attacks electrophilic positions in the aromatic ring. Electron-
donating groups like –OH or –NH2induce a high electronic density in the
ortho and para positions and, consequently, in these positions aromatic
compounds react actively with ozone. Electron-withdrawing groups such as
–COOH deactivate the aromatic ring for the substitution reaction. The reac-
tion is favoured by a resonance of the intermediate. For example, the attack in
the ortho position of phenol takes place by the following mechanism [30]:
Hydroxyl radicals initiate oxidative degradation by three reactions: hydro-
gen abstraction, hydroxyl addition and electron transfer. A saturated organic
compound may be attacked by a hydroxyl radical and may undergo hydro-
gen abstraction, a negligible reaction in compounds with aromatic rings and
140 A. Rodríguez et al.
double bonds [31]. It has been reported that the indirect oxidation of methyl-
tert-butyl ether starts with the abstraction of an α-hydrogen to form an or-
ganic radical, which reacts with oxygen to yield a peroxy radical with a large
(∼109M–1 s–1) second-order rate constant [32]. The peroxy radical can ab-
stract hydrogen to form α-hydroperoxy methyl-tert-butyl ether. In aqueous
solution, the reaction continues with the hydrolysis of the oxygen–oxygen
bond to produce tert-butyl formate and, subsequently, formic acid and tert-
butyl alcohol:
(CH3)3COCH3+HO
·→(CH3)3COC·H2O2
–→(CH3)3COCH2OOH ,
(CH3)3COCH2OOH –H2O
–→(CH3)3COCHO ,
(CH3)3COCHO +H2O
–→(CH3)3COH+HCOOH.
Unsaturated and aromatic compounds undergo hydroxyl addition, a reaction
with a very high rate (109–1010 M–1 s–1) and a product distribution markedly
affected by substituents. The hydroxyl radical is a strong electrophile and,
in the case of aromatic rings, preferably adds at electron-rich sites [33]. For
example, the attack of hydroxyl on aniline leads to ortho-andpara-hydroxy
compounds [34]. The stabilization of radical intermediates produced during
the addition of hydroxyl radicals may take place by hydrogen abstraction or
by electron transfer and proton elimination. Further reactions lead to ring
opening and the formation of open conjugated structures.
Electron transfer is the other mechanism of hydroxyl oxidation, commonly
encountered in oxidation of transition metal ions, which is also described in
organic compounds in which large substituents avoid addition reaction [35].
2
Ozone Uses in Water Treatment
Ozone is used as the only oxidant or in association with other oxidants or
energy (AOPs) in surface water, ground water or wastewater treatments. The
ozone-based technologies have the common objective of optimizing the use of
ozone to improve the disinfection or the removal of the pollutants present in
water.Thereasonisnotonlythefactthatitisanexpensiveoxidant,butalso
that it induces the generation of toxic oxidation intermediates. To reach this
Ozone-Based Technologies in Water and Wastewater Treatment 141
goal, it is necessary to develop models whose level of complexity depends on
the knowledge of the processes.
2.1
Precipitation of Oxides
Iron and manganese are undesirable in drinking water because of their ef-
fect on the appearance and taste of the water, their ability to cause black or
reddish staining or the formation of sediments. The rusty or brown stains on
fabrics are of concern because they are not removed by usual detergents. Iron
oxide deposits on tanks, water heaters and pipelines create problems of wa-
ter supply related to equipment maintenance. These pollutants are not health
threatening so the EPA does not set a mandatory water quality standard. The
guideline standards for both these metals have been established in their sol-
uble states, taking into account the growth of iron- and manganese-oxidizing
bacteria that strongly affects the overall water quality.
In the United States, the National Secondary Drinking Water Regulations
include secondary maximum contaminant levels (SMCLs) as a guideline to
avoid aesthetic effects related to odour, taste and colour. Current SMCLs
are 0.3 mg/L for iron and 0.05 mg/L for manganese. The Council Directive
98/83/EC on the quality of water intended for human consumption includes
iron and manganese in Annex I, Part C, with values fixed only for monitoring
purposes of 0.2 mg/L for iron and 0.05 mg/L for manganese.
The way iron and manganese should be removed depends on their oxida-
tion state and concentration. Both can be present in water in dissolved form
with oxidation states that depend on pH (Fe2+,Mn
2+,Fe
3+,Mn
4+)orincol-
loidal particle suspension. Ground waters, being anaerobic, have higher iron
and manganese contents than aerated water. In the latter case, the redox po-
tential of the water allows the oxidation of reduced ionic forms into insoluble
oxides. As concerns the oxidation mechanism, there is a certain controversy
over whether it consists of an oxygen transfer from ozone to the reduced
metal or an electron transfer from the reduced metal to ozone [36].
Iron and manganese are usually removed by oxidation of the dissolved
forms into an insoluble form by aeration or by chemical oxidization followed
by sand filtration. The success of removal by oxidation depends not only on
the oxidant used and its concentration, but also on pH and on the presence
of natural organic matter. Oxidation takes place at a faster rate at higher pH
values and the presence of organic matter makes removal more difficult. Both
iron and manganese tend to form bonds with humic acids and other natu-
ralorganicmattercompoundspresentinwater.Whenairisusedasoxidant
this causes removal difficulties and in this case the oxidation with ozone is
recommended. In general, the removal of iron is normally easier than that of
manganese, but a high content of iron requires treatment with several tanks
in series. It has been stated that the oxidation of iron with ozone is rapid,
142 A. Rodríguez et al.
but tends to form colloidal particles difficult to remove by sand or anthracite
filtration [37]. The ozone dose required for oxidation can be estimated stoi-
chiometrically as 0.43 mg/mg iron and 0.88 mg/mg manganese, the latter for
8.0 <pH <8.5, from the following reactions [38]:
2Fe2+ +O
3+H
2O→2Fe3+ +2HO
–+O
2,
2FeO + O3+3H
2O→2Fe(OH)3+O
2,
Mn2+ +O
3+H
2O→Mn4+ +2HO
–+O
2,
MnO + O3→MnO2+O
2.
Other oxidants may be used to remove iron and manganese by oxidation, but
the dose of oxidant is higher. Table 2 shows the usual values for precipitation
of iron and manganese from drinking water as a function of the oxidant [39].
The removal of oxides of iron and manganese may be carried out using dif-
ferent filtration media such as conventional beds of anthracite and sand with
chemically bonded manganese oxide. The most suitable, however, is man-
ganese greensand, a granular form of the zeolite mineral glauconite coated
with manganese oxide that bonds due to the ion-exchange properties of
glauconite [40]. This manganese-modified filtration medium also exhibits
a catalytic effect in the chemical oxidation of iron and manganese removal.
If necessary, the coating is regenerated by addition of potassium perman-
ganate to oxidize the MnO to MnO2. Backwashing of the greensand removes
the precipitated oxides from the bed. A prefilter to remove most of the pre-
cipitated iron prior to the manganese greensand also prolongs the service run
and reduces the pressure drop on the bed of greensand.
As concerns residual waters, ozone has been proposed to remove arsenic
from the wastewater of nonferrous metallurgical industries [41]. Arsenic is
a constituent of most sulphide ores and concentrates processed in nonferrous
metallurgical industries. Process wastes have to be treated in an environmen-
tally acceptable manner because of the environmental legal regulations. The
National Primary Drinking Water Regulations (EPA) limit the levels of arsenic
in drinking water to below 0.010 mg/L. Community water systems exceeding
0.005 mg/L (one half of the arsenic MCL) must notify their customers in their
Table 2 Commonly accepted dosing of oxidants required to remove iron and manganese
from drinking water
Oxidant Iron (mg/mg Fe) Manganese (mg/mg Mn)
Chlorine 0.62 1.27
Chlorine dioxide 1.21 2.45
Potassium permanganate 0.94 1.92
Ozone 0.43 0.88
Ozone-Based Technologies in Water and Wastewater Treatment 143
annual reports. The Council Directive 98/83/EC also includes a maximum
level of 0.010 mg/L(AnnexI,PartB,ChemicalParameters).
The first step in the removal of arsenic takes place by precipitation by sul-
phide formation. The solubility of arsenic sulphide is about 30 mg/L, high
enough to require a further treatment prior to discard to the environment.
Ozone can be used to oxidize As(III) to As(IV), which in the presence of
Mn(II) forms a precipitate with a Mn/As mole ratio around unity, believed to
be MnAsO4·nH2O. The residual arsenic concentration depends on the initial
manganese and iron concentrations and can be brought below the mandatory
limit of 0.010 mg/L. The precipitation of arsenic with manganese by ozona-
tion is also effective for removing arsenic in the pH range of 1–2 where ferric
arsenate and ferric hydroxide do not precipitate. It has been reported that the
conversion of As(III) to As(V) was fast with ozone with simultaneous oxida-
tion of iron and manganese. The sequestering effect of the resultant As(V)
played an important role. The sorption of freshly precipitated Fe(OH)3was
also significant and estimated to be 15.3 mg As/g Fe(OH)3[42].
2.2
Disinfection of Drinking Water
Ozone has been used since 1919 in drinking water disinfection. It is a strong
biocide which is able to deactivate resistant pathogen microorganisms re-
sistant to chlorine and chlorine dioxide, such as Cryptosporidium parvum
oocysts. The ozone-based technologies for drinking water disinfection try to
provide operation conditions which do not favour indirect ozone reaction via
hydroxyl radicals [43]. Ozone doses should eliminate and/or reduce the con-
centration of faecal microorganisms (faecal coliforms and Escherichia coli)to
values that exclude any risk to human health. In case some pathogen microor-
ganism refractory to ozone treatment exists, it is necessary to specifically
determine the required ozone doses. With this object the integral ct-exposure
parameter (ct) is defined by multiplying the disinfectant concentration in wa-
ter by the time that the microorganism is in contact with it:
ctO3=CO3dt.(20)
According to the microorganism deactivation model of Chick–Watson and
assuming first-order deactivation kinetics, ctO3determines the reduction of
viable microorganisms from an initial concentration Noto a final concentra-
tion N. For a given ozonation time in batch or plug flow:
log N
No=–kN
t
0
CO3dt.(21)
144 A. Rodríguez et al.
Equation 21 allows the calculation of the ozone necessary to obtain an ef-
ficient disinfection for a given microorganism with deactivation constant
kN[44]. The flow model permits the determination of the ozone concentra-
tion profile and therefore, the value of ctO3[45, 46]. When this information is
not accessible, the usual solution is to multiply the concentration of ozone at
theexitofthereactorbythetimethat10% of an inert tracer injected by pulse
is inside the reactor, ct10.
The relationship between the concentration of ozone dissolved, CO3,and
the total ozone dose (TOD) transferred can be obtained from Eq. 1, assuming
first-order kinetics for ozone decomposition [47]:
CO3=TOD–X–kd
t
0
CO3dt,(22)
TOD =
t
0
kLaC∗
O3–CO3dt,(23)
where Xis the TOD at the beginning of the ozonation, where it is possible
that the ozone decomposition is higher than the ozone absorption rate (mass
transfer control) and the ozone dissolved is not detected. Solving Eq. 22 the
following expression for Xis obtained:
X=TOD
i=
t
0
kLaC∗
O3.(24)
From Eq. 22 and for a process in which flow follows the continuous stirred
tank reactor (CSTR) model, the concentration of ozone inside the reactor is
given by:
CO3=TOD – X
1+kdτ,(25)
where τis the hydraulic retention time (HRT). From the Chick–Watson model
and Eq. 25, the extent of the ozone disinfection in a CSTR, expressed as the
relationship between the actual (N) and initial (No) concentration of microor-
ganisms, is:
N
No=1
1+kNTOD–X
1+kdττ
.(26)
This equation connects the operational parameters TOD, Xand τand the
kinetic parameters of the process (kN,kd,kLa) with the required extent of
disinfection. The deactivation constants for E. coli, Bacillus subtilis spores,
Rotavirus, Giardia lambia cysts, Giardia muris cysts and Cryptosporidium
Ozone-Based Technologies in Water and Wastewater Treatment 145
parvum oocysts, which are refractory to ozonation, have been reported [43]
but in general, data on microorganism deactivation kinetics are scarce.
2.3
Natural Water and Wastewater Treatment
The objectives of ozone-based treatments of surface water, ground water and
wastewater are disinfection and the elimination of dissolved organic matter.
Water dissolved organic compounds may present a huge variability in propor-
tion and nature of pollutants (persistent organic pollutants (POPs), personal
care products (PCPs), endocrine disruptors) depending on their source. The
ozone-based technologies for natural water and wastewater treatment provide
operation conditions favouring the direct or radical ozone reactions (AOPs).
Ozone can be used as the only technology or in combination with other pro-
cesses with the aim of improving coagulation–flocculation or biodegradabil-
ity, to remove pollutants in natural water treatments or as a tertiary treatment
in association with biological wastewater treatment [48].
Although the number of papers published on the efficiency of ozone to
eliminate POPs and PCPs is considerable, the use of ozone is less extended
in wastewater treatments than in disinfection or in natural water treat-
ments [49–57]. Ozone is an expensive oxidant and the necessary doses in
wastewater treatments are higher than in natural water, thus increasing op-
erational costs. However, the ability of ozone to mineralize organic matter,
alone or in association with other oxidants such as hydrogen peroxide, makes
it especially attractive for new developments, in particular those for which
the objective is the reuse of wastewater. The efficiency of the use of ozone re-
quires new ozone generators as well as models of the ozonation process to
optimize the ozone doses, thus reducing operational costs and avoiding toxic
intermediates.
Figure 4 shows the evolution of total organic carbon (TOC) during the
ozonation (O3/H2O2system) of aqueous solutions of a number of pollu-
tants and their mixtures classified as PCPs: analgesics (dipyrone, diclophenac,
acetyl salicylic acid), anti-inflammatories (ibuprofen), antiseptics (triclosan),
antibiotics (tetracycline), antineoplastics (cyclophosphamide), anxiolytics
(carbamazepine), hormones (oestradiol) and diagnostic compounds (ac-
etamide). The ozonation processes were carried out in a semicontinuous
mode in a 20-L bubble column reactor. Gas flowed at a rate of 0.36 Nm3h–1
(kLa=5.6 ×10–3 s–1) with an ozone concentration of 29.7–40.3 g/Nm3.The
experiments were performed at pH values in the range 7.5–8.5, temperatures
of 20–30 ◦C and a concentration of hydrogen peroxide of 1.0 ×10–4 M. In all
cases TOCowas reduced at least 50%duringthefirst30 min of ozonation.
Table 3 shows representative results and operational conditions. With the aim
of determining the nature of the final refractory organic carbon (TOCc), the
contribution of oxalate, acetate and formiate, the common ozonation end
146 A. Rodríguez et al.
Fig. 4 Ozonation of PCPs in a 20-L bubble column reactor (kLa=5.6 ×10–3 s–1). Experi-
mental values of TOC versus time tfor several compounds and mixtures. pH = 7.5–8.5,
T=20–30 ◦C, [H2O2]=1.0 ×10–4 M; gas flow rate: 0.36 Nm3h–1 , gas ozone concentra-
tion: 29.7–40.3 g/Nm3
Table 3 Experimental data for the ozonation of PCPs and mixtures
Compound TOCo×104O3g TpH (TOCc/(TOC∗
c/k
(M) (mM) (◦C) TOCo)TOC
o)(min
–1)
Acetyl salicylic acid 9.7 0.84 24 7.8 0.21 0.107
Carbamazepine 7.8 0.84 24 7.8 0.42 0.30 0.123
Diclophenac 6.7 0.84 24 7.6 0.29 0.22 0.037
Dipyrone 9.4 0.84 23 8 0.30 0.27 0.049
IBU+DCF 10% pot 14.5 0.62 24 7.8 0.35 0.17 0.035
IBU+DCF 20% pot 16.3 0.84 22 7.8 0.25 0.17 0.051
Ibuprofen 7.8 0.84 20 7.8 0.27 0.17 0.062
Tetracycline 7.7 0.84 23 7.8 0.18 0.070
Triclosan 3.0 0.84 30 7.7 0.30 0.083
ASA+DYP+CZP+ 32.4 0.84 22 7.8 0.15 0.023
IBU+DCF
Cyclophosphamide 8.8 0.84 24 7.8 0.26 0.071
β-Oestradiol 2.0 0.84 24 7.8 0.39 0.026
ASA+acetamide 17.3 0.84 24 7.8 0.53 0.40 0.071
Ozone-Based Technologies in Water and Wastewater Treatment 147
products, were measured (TOC∗
cin Table 3). The experimental data of TOC
were fitted with pseudo first-order kinetic expressions like Eq. 25 and kinetic
constantsareshowninTable3:
ln TOC
TOCo=kt .(27)
As in disinfection of drinking water, the ozonation models for wastewa-
ter must relate process conditions and kinetic parameters with the ozone
dose required to remove pollutants. The supply of ozone is determined from
a selected parameter whose value must be reduced. Depending on the in-
formation available about the wastewater, the object will be the reduction
of one of various specific pollutants or to lower global parameters such as
TOC or chemical oxygen demand (COD). The basic knowledge about the
ozonation process determines how close the model is to reality. It is pos-
sible that in wastewater from a given industrial process the ozonation kinetic
constants of the main pollutants can be available. In most cases the reaction
paths of compounds present in wastewater matrices and their elimination ki-
netic constants are not known and therefore a global approach is normally
preferred.
In the experiment of Fig. 2, where ozonation by a radical pathway is not
favoured, an elimination of TOC close to 10%wasobservedduringthefirst
4min of reaction (zone I). The ozone decomposition kinetic constant was
kd=0.139 min–1 obtained by solving Eq. 28 as in Sect. 1.3:
dCO3
dt=–kdCO3.(28)
The TOC removal kinetic constant was k=0.78 min–1,obtainedbyfittingthe
experimental TOC values of Fig. 2 with a pseudo first-order kinetic equation.
Although this approach can characterize the kinetics of the process, it does
not relate the ozone dissolved concentration with TOC in order to calculate
the ozone dosing. Taking into account that the elimination of pollutants in
water by ozone is due to direct and indirect (radical) reactions, Elovitz and
von Gunten [24] proposed a model for the removal of specific micropollutants
in which the ozonation process is characterized by a parameter Rct defined as
the relationship between the integral ct-exposure to ozone and the hydroxyl
radical, the two principal oxidants in the system:
Rct =CHO·dt
CO3dt(29)
The Rct parameter characterizes the ozonation process and allows estimation
of the concentration of the hydroxyl radical in water from the concentration
of dissolved ozone. The balance of a determined pollutant (P) with CPoini-
tial concentration in water in a volume element of the reactor either in batch
or plug flow during an ozonation process follows the expression of Eq. 30.
148 A. Rodríguez et al.
The kinetic constants kO3and kHO·are linked with direct and indirect ozone
reactions, respectively. Rct connects the extent of decontamination with the
integral ct-exposure of ozone:
ln CP
CPo
=kO3
t
0
CO3dt+kHO·
t
0
CHO·dt=kO3+RctkHO·t
0
CO3dt(30)
From Eqs. 30 and 22 the ozone requirements (concentration of ozone in wa-
ter and TOD) can be linked with the elimination level of pollutants. The
low concentration values of the hydroxyl radical in water (CHO·≤10–12 M)
make its direct measurement practically impossible. However, the integral
CHO·dtmay be determined by means of probe compounds [58, 59], such
as p-chlorobenzoic acid (pCBA), whose direct and indirect kinetic constants
are known (kO3/pCBA ≈0.15 M–1 s–1,kHO·/pCBA ≈5×109M–1 s–1). A balance
to pCBA leads to the following expression to the integral ct-exposure to HO·:
ln pCBA
pCBAo
kHO·=CHO·dt.(31)
Fig. 5 Evolution of TOC (◦) and ozone concentration during treatment of D070208
wastewater (Table 4) with O3/H2O2. pH = 8.04–8.25, T=25 ◦C; gas flow rate:
0.36 Nm3h–1, gas ozone concentration: 45.9 g/Nm3;kLa=0.614 min–1 and injection of
0.15 mL of H2O2(30%w/v)every5min
Ozone-Based Technologies in Water and Wastewater Treatment 149
Table 4 Reaction conditions and kinetic parameters of the ozonation (O3/H2O2) of domestic (D) and urban (U) wastewater at T=25 ◦C, pH =
7.6–8.2 and KLa=0.614 min–1 (f: sample filtered before ozonation)
Sample TOCoTOC CO3gas CO3∗CO3IkdI RICO3II kdII RII CO3III kdIII
(ppm) removed (mM) (mM) (mM) (min–1)(mM
–1 min–1) (mM) (min–1 )(mM
–1 min–1) (mM) (min–1 )
(%)
D070206f15.51 91.3 1.010 0.233 0.038 3.26 0.823 0.050 2.32 0.125 0.198 0.112
D070208f12.77 85.8 1.039 0.240 0.075 1.36 1.084 0.091 1.01 0.375 0.204 0.109
D070308 12.17 62.7 1.021 0.236 0.120 0.42 0.996 0.134 0.33 0.084 0.190 0.104
D070417 8.45 85.4 1.019 0.235 0.049 1.51 0.337 0.070 0.94 0.192 0.199 0.072
D070419 14.04 76.4 1.021 0.236 0.050 2.25 0.496 0.078 1.22 0.047 0.206 0.088
U070205f11.89 90.5 1.000 0.231 0.027 3.20 1.226 0.043 1.85 – 0.202 0.061
U070208f8.98 100 1.035 0.239 0.037 3.20 2.130 0.056 1.91 – 0.202 0.107
U070222 8.56 92.1 1.039 0.240 0.053 2.20 1.744 0.090 1.04 0.261 0.209 0.093
U070305 11.85 88.4 1.071 0.247 0.054 2.25 0.513 0.067 1.69 – 0.210 0.111
U070308 15.54 85.6 1.034 0.239 0.038 2.04 0.220 0.051 1.42 0.165 0.188 0.105
U070416 8.76 94.2 1.035 0.239 0.043 2.77 1.270 0.065 1.63 0.159 0.201 0.115
U070419 8.38 95.0 1.031 0.238 0.051 2.48 0.525 0.068 1.69 – 0.205 0.109
150 A. Rodríguez et al.
Table 5 Characterization of domestic (D) and urban (U) wastewaters
Sample Suspended Conductivity pH COD Alkalinity NO3–PO43– SO42– Cl–NH4+
solids (µS) (mg L–1) (ppm CaCO3) (ppm) (ppm) (ppm) (ppm) (ppm)
(mg/L)
D070206 12.8 – 7.98 44 528.14 2.56 9.64 111.46 59.80 41.27
D070208 12.8 – 8.04 44 495.52 2.12 7.97 112.42 62.24 –
D070308 10.1 368 7.76 182 419.37 2.12 5.16 84.93 51.84 36.04
D070417 0.55 631 8.12 81 579.93 3.01 12.91 86.39 54.59 49.76
D070419 7.45 579 8.38 66 521.38 5.43 11.61 80.69 54.54 42.56
U070205 3.6 – 8.01 42 539.23 14.69 8.18 126.5 77.49 –
U070208 12.4 – 8.11 35 507.67 12.45 13.03 131.59 75.78 –
U070222 – 407 7.61 62 454.57 11.57 7.09 102.83 67.84 28.32
U070305 11.6 492 7.66 71 497.44 14.46 6.39 114.26 73.64 30.25
U070308 0.65 446 7.59 54 415.29 47.82 4.87 90.60 61.36 16.05
U070416 – 601 7.65 67 475.17 95.83 2.53 116.25 78.47 1.14
U070419 – 548 7.94 52 344.64 30.52 1.34 101.46 81.35 9.39
Ozone-Based Technologies in Water and Wastewater Treatment 151
Figure 5 shows the evolution of ozone and TOC during the ozonation of the
same wastewater of Sect. 2.3 (D070208, Table 5). Reaction conditions were
the same as in Sect. 2.3 but equal volumes (0.15 mL) of hydrogen peroxide
(30%w/v)wereinjectedevery5min in order to favour the TOC elimination
by a radical pathway. Three zones that correspond with three different ozone
decomposition and TOC elimination kinetics can be observed. The TOC that
remained in zone III is refractory to ozonation. In zone III injections of
hydrogen peroxide were stopped to avoid ozone decomposition by that com-
pound, thereby allowing the concentration of ozone to reach a stationary state
(CO3=0.204 mM). In the later part of zone III, the concentration of ozone
decreased once gas flow was stopped. The conditions in zone III allow the de-
termination of kdIII =0.109 min–1 and kLa=0.619 min–1.InzonesIandIIthe
ozone fluctuations are due to the decomposition induced by hydrogen perox-
ide. In these two zones two minimums CO3I=0.075 mM and CO3II =0.091 nM
may be identified as indicated in Fig. 5. Assuming that these two values could
correspond with two stationary states of the process, the ozone decompos-
ition kinetic constants for each zone can be calculated: kdI =1.36 min–1 and
kdII =1.01 min–1.
Assuming that, due to reaction conditions, the ozonation proceeds by
a radical mechanism, Eq. 30 applied to TOC leads to the following expression:
ln TOCo
TOC =RctkHO·CO3dt=RCO3dt.(32)
Figure 6 shows the logarithmic plot of TOC removal as a function of the in-
tegral ozone exposure following Eq. 32. Two different TOC removal zones,
identified with zones I and II in Fig. 5, can be observed. The corresponding
slopes are RI=1.084 mM–1 s–1 and RII =0.375 mM–1 s–1.
The parameter Rallows a kinetic characterization of the ozonation pro-
cesses even though global parameters to measure the contamination in water
such as TOC or COD are used. As said before, from Eqs. 32 and 22 the ozone
requirements for a given degree of TOC removal can be linked. Table 4 shows
the TOC removed and the experimental values of R,kdand ozone concentra-
tion at the different steps of ozonation of two kinds of wastewaters: domestic
(D) and urban (U) from the secondary clarifier of two municipal wastewater
treatment plants. Wastewaters were treated by ozonation processes with the
O3/H2O2system as indicated before. With the aim of reflecting the seasonal
variability of wastewater, the samples were collected at different dates. The
main characteristics of the wastewater are shown in Table 5.
Yurteri and Gurol [60] related the ozone decomposition kinetic constant
kdwith pH, alkalinity (Alk) and TOC. These researchers found a deviation
within ±25% between kdestimated by the empirical Eq. 33 and the experi-
mental values determined by Eq. 34 in ozonation processes of surface water
152 A. Rodríguez et al.
Fig. 6 Determination of kinetic parameters Rfor the ozonation. Treatment of D070208
wastewater (Table 4) with O3/H2O2. pH = 8.04–8.25, T=25 ◦C; gas flow rate:
0.36 Nm3h–1, gas ozone concentration: 45.9 g/Nm3;kLa=0.619 min–1 and injection of
0.15 mL of H2O2(30%w/v)every5min
and wastewater (pH = 6.8–9.0, TOC = 0.3–5.3, Alk = 10–500 mg/LCaCO
3).
log kd= – 3.98 + 0.66pH + 0.61 log TOC – 0.42 log(Alk/10) (33)
–dCO3
dt=kdCO3(34)
The removal of specific pollutants during an ozonation process can be per-
formed by considering separately the direct reaction with ozone and the
radical chain propagation with hydroxyl radical:
dCM
dt=–(zkDCO3+kHO·CHO·)CM.(35)
The corresponding balance to ozone may take into account the initiation and
termination reactions as well as reaction with the organic intermediates:
dCO3
dt=kLa(C∗
O3–CO3)–kHO–CHO–+kHO2–CHO2–CO3
–(kDCM+kDiCMi)CO3–krCrCO3.(36)
Ozone-Based Technologies in Water and Wastewater Treatment 153
The usefulness of these models to determine the requirement of ozone de-
pends on knowledge about (1) the stoichiometry of the direct ozone reaction,
(2) the direct kinetic constants of ozone with M and with reaction prod-
ucts Mi, (3) the kinetic constant of hydroxyl reaction with M and (4) the
ozone decomposition kinetic constants due to other radical species in wa-
ter, kr. Glaze and Kang [61,62] and Beltran et al. [63] solved a similar set
of equations to determine the removal of low molecular weight halogenated
compounds, polynuclear aromatics and nitroaromatic hydrocarbons. The
concentration of radical species in solution was determined in all cases by
assuming a stationary state.
2.4
Catalytic Ozonation
The homogeneous rate of production of hydroxyl radicals from ozone is
strongly dependent on pH, since the active species in the initiation of the
ozone decomposition mechanism are HO–and HO2–,theconcentrationsof
which are directly related to the concentration of hydroxide [15,16]. However,
the ozonation under alkaline conditions presents an important drawback in
the case of water with bromine levels higher than 50 µg/Lduetotheforma-
tion of bromate as oxidation by-product [43]. Excessive bromate formation is
a major concern due to its potential carcinogenicity, which imposed a limit
of 10 µg/L in drinking water standards both in the United States and Europe.
Besides limiting ozone exposure, a recommended strategy to reduce bromate
formation is to use pH <7 because bromate formation is also strongly pH
dependent [64]. On the other hand, under acidic conditions the formation
of hydroxyl radicals and the rate of mineralization are much lower than in
conventional ozonation. In this situation, a catalyst may be used to promote
ozone decomposition, ozonation reactions or both. Other advanced technolo-
gies for water and wastewater treatment currently under development may
avoid bromate formation. Sonochemical methods, photolysis or photocataly-
sis, Fenton processes or certain combinations with conventional technologies
like adsorption, wet oxidation, membrane separations or biological treatment
might compete with ozonation. To date, however, only ozone-based technolo-
gies have been widely used in water treatment plants, which justifies the effort
to develop catalytic processes able to operate in acidic media.
Another important drawback of conventional homogeneous ozonation
that might be overcome by using a catalyst is the inhibition due to the pres-
ence of carbonates, bicarbonates and other radical scavengers. The case of
carbonates and bicarbonates is especially obvious not only from their pres-
ence in natural water and wastewater, but also because they are products of
the mineralization reactions. Inhibitors reduce the efficiency of an ozonation
process and cause a poor mineralization due to the persistency of interme-
diate oxidation products. Short-chain carboxylic acids represent a class of
154 A. Rodríguez et al.
organic compounds particularly refractory to the oxidation by ozone. Acids
such as pyruvic, glyoxalic or oxalic are normally produced during the ozona-
tion of complex organic molecules, their refractory character being respon-
sible for most of the organic content of treated wastewaters. A number of
attempts have been made to remove these compounds during ozonation using
catalysts in either the homogeneous or heterogeneous phase [65–69]. An im-
portant question about the behaviour of carboxylic acids in a heterogeneous
catalytic system is the ability of the catalyst to adsorb the organic substrate.
This point will be discussed below.
It has been suggested that the combination of ozonation and adsorption
on activated carbon in a single process is an alternative to the treatment of
wastewaters containing organic contaminants [70]. As revealed before, the
ozonation efficiency for carbon removal is limited due to the formation of
refractory short-chain carboxylic acids. On the other hand, activated car-
bon becomes saturated easily when treating wastewaters with high organic
content, requiring frequent regeneration or replacement [71]. The ozonation
on activated carbon may allow these limitations to be overcome because of
a high adsorption capacity combined with high surface area and catalytic
activity due surface metals and other surface chemical properties. The cata-
lytic mechanism of ozonation on activated carbon is still unclear, but most
results suggest that the role played by carbon is essentially to promote the
decomposition of ozone with a subsequent increase in the production of rad-
icals. The hydroxyl radicals formed would not be bonded to the surface, being
free to react in the aqueous phase. Therefore, activated carbon would behave
as an initiator of the radical-type chain reaction that transforms ozone into
hydroxyl radicals.
In what follows, attention will be focused on homogeneous catalytic sys-
tems and the ozonation on metals and metal oxides. By far the most com-
monly tested catalysts for the ozonation of organic compounds are supported
and unsupported metals and metal oxides, especially titanium oxide and
manganese oxide [66, 67, 72]. There is a certain controversy on the mechan-
ism of ozonation on ionizable surfaces. Some authors assume the formation
of surface oxidation sites able to interact with organic compounds [73]. Ma
andGraham[74]suggestedamechanismbasedontheinitiationofozone
decomposition by hydroxide ions linked to the negatively charged surface of
manganese oxide. The interaction of organic solutes with charged surfaces
must be relevant and will be discussed below. Ozonation on supported metals
has been less studied. Lin et al. [19, 75] reported a considerable efficiency for
the removal of formic acid on Pd and Pt on alumina. Noble metals, especially
when supported on SiO2, also showed appreciable activity for the decompos-
ition of ozone in water and are candidates to catalyse an ozonation process.
The ozonation of carboxylic and chlorinated carboxylic acids on Ru/CeO2and
Ru/CeO2–TiO2have been reported by Karpel et al. [76] and Fu et al. [77]. The
results of ozonation on metals and metal oxides have in common a strong de-
Ozone-Based Technologies in Water and Wastewater Treatment 155
pendence of the reaction rate on the mode of preparation of the catalyst and
on support pre-treatment. These variables should affect the interaction of the
molecule with the surface and its adsorption capability for ozone or organic
molecules.
2.4.1
Homogeneous Catalytic Ozonation
Earlier works showed that certain metals in solution are able to increase
the removal of organics from aqueous solution with respect to non-catalytic
ozonation [78]. The catalytic activity of Fe(II), Fe(III), Mn(II), Ni(II), Cr(III),
Ag(I), Cu(II), Co(II), Zn(II) and Cd(II) have been reported [79].
It has been proposed that the mechanism of homogeneous ozone–metal
systems is based on the generation of hydroxyl radicals through an ozone
decomposition reaction [80]:
Fe2+ +O
3→FeO2+ +O
2,
FeO2+ +H
2O→Fe3+ +HO
·+HO
–.
On the other hand, Novell and Hoigné [81] indicated that the production of
hydroxyl radicals cannot be directly related to the interaction of ozone with
the transition metal and the latter reaction should be substituted by the fol-
lowing one:
FeO2+ +Fe
2+ +2H
+→2Fe3+ +H
2O.
In addition to this, the iron-catalysed ozonation may share reactions with the
classical Fenton homogeneous process. The interaction of ozone and water is
known to produce hydrogen peroxide, which may produce hydroxyl radicals:
Fe2+ +H
2O2→Fe3+ +HO
·+HO
–,
Fe2+ +HO
·→Fe3+ +HO
–,
RH + HO·→H2O+R
·,
R·+Fe
3+ →R++Fe
2+ .
Organic radicals should play an important role in the reduction of Fe(III), but
the regeneration of the catalyst may take place by other mechanisms with the
intervention of hydroperoxy radicals, HO2·[82].
Oxalic acid tends to form complexes with transition metals such as man-
ganese, iron and cobalt. The formation of these complexes plays an import-
ant role in the catalytic mechanism of ozonation [83]. Andreozzi et al. [66]
studied the ozonation of glyoxalic acid catalysed by manganese salts and sug-
gested a mechanism of oxidation mediated by Mn(III):
Mn(II)+O
3+2H
+→Mn(IV)+O
2+H
2O,
Mn(II)+Mn(IV)→2Mn(III).
156 A. Rodríguez et al.
Mn(III) reacts with an acid moiety originating the abstraction of CO2with
the consequent reduction of Mn(III) to Mn(II).
Mn(III)+ HOOC–R→Mn(III)– OOC–R+H
+
Mn(III)–OOC–R→Mn(II)+CO
2+R
·
The formation of complexes has also been proposed for the ozonation of ox-
alic acid using Co(II) by Pines and Reckhow [83]. In this case the Co(II)
complex is oxidized by ozone to a Co(III) form with subsequent elimination
of Co(II), resulting in an oxalate radical anion which further decomposes.
2.4.2
Catalysis by Metals and Metal Oxides
The mechanism of catalytic ozonation on oxides and metals may involve the
adsorption of ozone, but not necessarily the adsorption of organic pollu-
tants. It has been demonstrated that dissolved ozone adsorbs and decomposes
on many solid surfaces other than activated carbon, the resulting radicals
being responsible for indirect oxidation reactions [25, 75]. This mechanism
relies on the well-known result that gas-phase ozone adsorbs on solid surfaces
to yield different molecular or ionic species. Dhandapani and Oyama [84]
reported that ozone decomposition on p-type oxides is consistent with the
formation of superoxide (O2–)orperoxide(O
22–)speciesonthesurface.Bul-
lanin et al. [85] suggested that on the stronger Lewis sites ozone dissociates
after adsorption to yield a surface oxygen atom. With weaker sites, ozone
molecules form a coordination bond via one of the terminal oxygen atoms.
Another possibility is the formation of weak hydrogen bonds with surface OH
groups:
O3+SO2+O–S,
O3+SO=O–O–S,
O3+SOHO=O–O···HO–S.
In the case of metals and metal oxides, the catalytic reaction may also in-
volve the adsorption of organic molecules or ions on surface sites leading to
two additional mechanisms. The first possibility is an Eley–Rideal interaction
between an adsorbed organic molecule and ozone or radicals from the bulk
aqueous phase. On the other hand, an adsorbed organic molecule can react
with adsorbed ozone or the products of surface ozone decomposition. Legube
and Karpel [36] proposed a redox version of the latter for heterogeneous
ozonation on metals, in which the ozone oxidizes surface metal atoms with
the generation of hydroxyl radicals. Organic molecules are oxidized by elec-
tron transfer from the catalyst yielding back the reduced form of the metal
Ozone-Based Technologies in Water and Wastewater Treatment 157
and organic radical species:
Meredn++O
3→Meoxn+1 +O
3·–,
O3·–+H
+HO3·,
HO3·→HO·+O
2,
Meoxn+1 +A→Meoxn+1 –A→Meredn+A
·.
The adsorption of neutral organic compounds on Lewis acid sites is diffi-
cult due to the competitive adsorption of water molecules on the surface.
Moreover, at basic pH, the hydroxide anion should prevent any adsorption
on Lewis sites. Surface interaction is easier for ionizable organic molecules in
aqueous solution if the surface is charged and allows ion exchange. The sur-
face of metal oxides exhibits ion-exchange properties and the hydroxyl groups
formed behave as Brönsted acid sites and dissociate depending on the pH of
the solution.
M–OH + H+M–OH2+
M–OH + HO–M–O–+H
2O
If K1and K2are the ionization constants for the preceding surface equilibria,
the point of zero charge (PZC) represents the pH at which the surface is not
charged:
pHPZC =pK1+pK2
2.(37)
A neutral organic compound may adsorb on metal oxide surfaces provided
it is a strong enough Lewis base and the pH of the solution is near pHPZC
of the oxide. Otherwise, it is reasonable to assume that only ionizable sub-
stances would be capable of interacting with charged surfaces. Carboxylates,
for example, adsorb on positively charged surfaces by exchanging the corres-
ponding counteranion [18].
The kinetics of adsorption may play an important role in the ozonation
process. Figure 7 shows some results for the adsorption of naproxen (pKa=
4.60) and carbamazepine (pKa= 14.0) on TiO2Degussa P25 with pHPZC of
6.8 [86] and TiO2/Al2O3with pHPZC of 8.3, prepared by impregnation with ti-
tanium isopropoxide following the method of Zhaobin et al. [87]. As expected
from the PZC of the surface, adsorption is favoured for naproxen under
acidic conditions at which the surface behaves as an anion exchanger [88].
Carbamazepine does not dissociate in acidic solutions and consequently its
adsorption pattern is similar for pH values in the range 3–7. Similar results
have been published for other acidic solutes [67, 89]. The results indicate that
the rate of adsorption may be slow enough to control the overall kinetics. For
the drugs mentioned above, the adsorption ranged only 5–15%fromtheir
equilibrium value during the first hour. Even for naproxen at pH = 3, below its
pKathe adsorption was slow and equilibrium required a day or more in most
158 A. Rodríguez et al.
Fig. 7 Dimensionless concentration of naproxen during adsorption on TiO2/Al2O3at
pH = 5 (), on TiO2P25atpH=3(◦), on TiO2P25atpH=7() and carbamazepine on
TiO2P25 at pH = 5 (•)andpH=7(). Temperature 25 ◦C, catalyst loading 1g/L, initial
concentration 6.0–6.5 ×10–5 M
cases. This agrees with several pieces of data showing that the adsorption of
acid pollutants on metal catalysts supported on alumina is slow, taking from
hours to days to complete [68, 89]. A kinetic model of ozonation should take
into account this possibility.
Without explicit consideration of the surface charge, the ozonation mech-
anisms that involve the adsorption or ion exchange of an organic compound
start with the bonding of the adsorbate to a vacant site (S). The adsorbate (A)
would displace coordination water and bond to the surface:
A+SA–S.
The reaction may then take place between adsorbed organic solute and an
oxidized site on the catalytic surface following a Langmuir–Hinshelwood
mechanism:
O–S+A–S→AOx+2S.
Alternatively, dissolved ozone or hydroxyl radicals in solution may react with
the adsorbed organic compound through an Eley–Rideal interaction:
HO·+A–S→AOx+S.
Some organic compounds react with dissolved ozone at a high rate. This is
the case for drugs such as naproxen and carbamazepine mentioned above.
Anyway, the mineralization rate is slow in non-catalytic systems in acidic
Ozone-Based Technologies in Water and Wastewater Treatment 159
conditions, so that the design of catalytic reactors is focused on refractory
compounds under conditions in which direct reactions can be ignored. The
rate of reaction of an organic compound combines the homogeneous reaction
with hydroxyl radicals and the heterogeneous reaction of hydroxyl (or ozone)
with solute with at least one reagent adsorbed.
If the reaction takes place between adsorbed species and hydroxyl radicals
fromthebulk,theEley–Ridealrateexpressionwouldbeasfollows:
–dCA
dt=kHO·CHO·CA+kcCsCHO·θ,(38)
where θis the fraction of surface sites occupied by adsorbate and Csthe bulk
concentration of solids. Accepting the Rct concept of Elovitz and von Gun-
ten [24], the ratio of hydroxyl radicals and ozone at any time is constant over
awiderangeoftheozonationprocess:
Rct =CHO·
CO3
.(39)
If the catalytic reaction with the adsorbed organic compound is the rate-
limiting process, an adsorption equilibrium exists at any time. Using kaand
kdto denote the adsorption and desorption kinetic constants, Eq. 38 can be
rewritten as follows:
–dCA
dt=kHO·RctCO3CA+kcCsRctCO3
kaCA
kaCA+k–a .(40)
If surface coverage is low, Eq. 40 can be simplified by assuming that kaCA
k–a:
–dCA
dt=kHO·Rct +kcCsRctKaCO3CA.(41)
The adsorption equilibrium constant, Ka, becomes included in a group of
constants with a linear dependence on catalyst load. Equation 41 can be in-
tegrated to obtain explicitly the time-integrated concentration of ozone:
ln CA,o
CA=kHO·Rct +kcCsRctKaCO3dt.(42)
A similar result would be obtained for a reaction between adsorbed organic
compounds and oxidized catalyst sites, provided that the elementary surface
step is rate controlling and an adsorption equilibrium exists at any time for
both reagents:
–dCA
dt=kHO·RctCO3CA+kcCsCO3CA
1+KaCs1+KoxCO3,(43)
where Kox represents the equilibrium constant of the surface oxidation step.
If the equilibrium constants are small, the surface step would be first order in
160 A. Rodríguez et al.
the oxidant and in the organic compound and the differential and integrated
rate equation would be similar to Eqs. 41 and 42, respectively.
A surface redox mechanism such as that involved in ozonation has some-
times been described by means of a Mars–van Krevelen rate expression. The
rate of catalytic reaction would depend on the rate of the oxidation process,
ko, and the rate of the organic compound with the oxidized catalyst, kc.As-
suming that the ozone is the oxidant and ignoring the surface stoichiometric
coefficient, the rate of organic depletion would be as follows:
–dCA
dt=kHO·RctCO3CA+kokcCsCO3CA
koCO3+kcCA.(44)
If the rate of catalyst oxidation is low, the catalytic reaction would be zero
order in the organic compound. Conversely, for a high rate of surface oxida-
tion, the reaction rate would be independent of the concentration of oxidant
and first order in the oxidized organic compound. The resulting equations
can be integrated and yielded equations somewhat different from Eq. 42.
Although relatively common in catalysis, the Mars–van Krevelen approach
recently received some criticism concerning its fundamental background.
Vannice [90] proved that the kinetic expression is incorrect and that the
obtained reaction orders can be derived from the Langmuir–Hinshelwood
equation under more transparent assumptions.
On the other hand, if adsorption of organics is so slow that it controls the
overall kinetics, the rate of the catalytic process would be independent of the
concentration of ozone or other oxidants:
–dCA
dt=kHO·RctCO3CA+kaCsCA.(45)
The integration of Eq. 45 leads to an expression in which the logarithmic
decrease of the organic compound is not linear in the time-integrated concen-
tration of ozone:
ln CA,o
CA=kHO·Rct CO3dt+kacst.(46)
The mechanisms described in Eqs. 42 and 46 can be discriminated by using
kinetic data. For example, Eq. 42 predicts that a change in the concentration
of ozone should have no effect in the logarithmic decrease of the concentra-
tion of a given compound, while Eq. 46 suggests that decreasing the ozone
dose would lead to a time-integrated concentration in more time and, there-
fore, to a greater conversion of the organic compound.
If an aggregate such as TOC is used instead of the concentration of a sin-
gle compound, the kinetic expressions would show the evolution of miner-
alization with the time-integrated concentration of ozone. Figure 8 shows
experimental data corresponding to the ozonation of naproxen (6.5 ×10–5 M
in pure water) using titanium dioxide Degussa P25 as catalyst and in non-
catalytic runs performed in the same conditions. Ozone was continuously
Ozone-Based Technologies in Water and Wastewater Treatment 161
Fig. 8 Mineralization during the ozonation of naproxen (6.5 ×10–5 M) on TiO2P25 at
pH=3()andpH=5(•)at25 ◦Candcatalystloadingof1g/L. Empty symbols cor-
respond to non-catalytic runs under the same conditions. The units of the integral ozone
exposure are mM s
bubbled from a corona discharge ozone generator and the steady-state con-
centration of ozone in the liquid was 0.230 mM.
Data show the existence of two different mineralization periods. During
the first period most of the TOC decay takes place, the reaction being con-
siderably accelerated by the presence of catalyst. The second period was es-
sentially independent of pH and reflects the slow mineralization of refractory
compounds. The similitude between catalytic and non-catalytic plots hides
at least one fundamental difference between both processes. Figure 9 shows
the concentration of oxalate measured by ion chromatography (DIONEX, DX-
120 ion chromatograph) and reveals that, even on a neutral surface, oxalate
is mineralized in conditions at which the rate of reaction is very slow. In
fact, the higher degree of reaction of oxalate takes place with neutral sur-
face charge and neutral pH at which the mineralization is not particularly
deep.
For the ozonation of naproxen in the runs reported in Fig. 9, the global
extent of mineralization was about 50% in non-catalytic runs at pH 5–7 and
reached over 60% using catalyst. The ozonation of carbamazepine allowed
a deeper mineralization with 73%TOCreductionafter120 min. In fact, most
of the TOC decay takes place during the first 10–20 min where the removal of
the more oxidizable compounds takes place. In non-catalytic ozonation, ox-
alate accounted for as much as 30% of the organic carbon remaining in the
162 A. Rodríguez et al.
Fig. 9 Concentration of oxalate during the non-catalytic ozonation of naproxen at pH = 5
() and the catalytic ozonation on TiO2P25ofnaproxenatpH=3(), carbamazepine at
pH = 3 (◦) and carbamazepine at pH = 5 (•). Catalyst loading 1g/L; initial concentration
of naproxen 6.5 ×10–5 M
reaction mixture, in contrast with a maximum of 12% encountered in non-
catalytic runs. This pattern reveals that the use of a catalyst favours not only
the reactions leading to oxalate but also the mineralization of oxalate itself.
In fact, oxalate was not detected in runs performed at pH = 7 with a min-
eralization degree of about 50%. At pH = 3, at which the surface of P25 is
positively charged and may interact with oxalate anion in solution, neither
the rate of mineralization nor the removal of oxalate from the solution are
particularly high Therefore, the mineralization of oxalate is not favoured by
positive surface charge, a result that seems to exclude a mechanism based on
the ion exchange of oxalate.
The results also proved that the rate of ozone decomposition is inhib-
ited by the catalyst over a neutral surface. At pH = 7, the homogeneous rate
constant for the ozone self-decomposition is 8.83 ×10–3 s–1 which lowers to
1.27 ×10–3 s–1 in the presence of 1g/LofP25TiO
2.Thecatalyticmineral-
ization rate is a maximum at pH = 5, and the reaction is also inhibited by
higher pH values. Anyway, the inhibition effect of increasing the concentra-
tion of hydroxide anion is much greater on the mineralization reaction than
on ozone decomposition. The best results for the removal of reaction interme-
diates were obtained for slightly positive surface charge, suggesting that the
adsorption of organics on Lewis sites may be the mechanism of the catalytic
ozonation of naproxen and carbamazepine. In catalytic runs, the degree of
mineralization was not directly linked to the accumulation of low molecular
weight carboxylic acids. In all runs, low levels of acetate, formiate and other
Ozone-Based Technologies in Water and Wastewater Treatment 163
low weight acids were detected in most cases, but without an accumulation
pattern linked to the evolution of TOC. These results point towards a stronger
surface interaction with the first ozonation products than with the more oxi-
dized carboxylic acids. The dependence of the degree of mineralization on the
rate of adsorption expressed by Eq. 46 was not tested and could be confirmed
by experiments performed in conditions at which the integral ozone exposure
is not linear with time.
2.5
Applications in the Treatment of Industrial Wastewater
The use of ozone-based technologies for the treatment of pollutants in water
has been the focus of attention in the literature during the last few years. Ex-
amples of their application as eco-effective alternatives have been presented
for several types of contaminants in water, in particular for pharmaceutical
residues or PCPs.
Data in scientific publications concerning the use of ozone for industrial
wastewater treatments can be found to a lesser extent. As has been men-
tioned before, ozone is an expensive oxidant and the treatment of industrial
or wastewater effluents needs doses higher than those for the treatment of
natural waters. In consequence, its use may be limited, but the ability of ozone
to mineralize organic matter, alone or in association with other oxidants such
as hydrogen peroxide, may be attractive for new developments such as in the
reuse of wastewater. Combinations with other oxidation techniques, such as
UV irradiation or ultrasonic techniques, can also be of interest due to a higher
efficiency and lower cost. The effectiveness of ozone-based technologies has
been evaluated in scientific publications for different industrial sectors, such
as textiles, petroleum refineries/phenols, pulp and paper, and electroplating
wastes.
Recent publications compile and review the studies performed for the
treatment of recalcitrant pollutants by the use of AOPs, where possible mech-
anisms responsible for synergistic action are described [91]. The advantages
of combined treatments are the reduction in the time of treatment and higher
removal efficiency. In fact, most common approaches have made use of chem-
ical (O3or H2O2) or photochemical (UV) based processes where the oxi-
dation power for the degradation of organic pollutants can be significantly
enhanced.
Table6showsasummaryofthereviewedliteratureconcerningAOPsap-
plied to the treatment of industrial wastewater effluents. The effectiveness of
combined photocatalytic and ozonation processes has been probed for tex-
tile effluents [92]. The removal efficiency of phenols from wastewater using
aUV/H
2O2/O3process (pH = 7, c(H2O2)=10 mM) is complete (100%) within
aperiodof30 min of treatment [93]. The colouring matter is almost com-
pletely eliminated, achieving a high reduction of TOC.
164 A. Rodríguez et al.
Table 6 Applications of ozone-based processes for the treatment of industrial wastewater
Industrial Pollutants/related Treatment Removal Refs.
activities pollutants efficiency
and others (%)
Petroleum Phenols UV/H2O2/O3100% Phenol [91]
refineries 58%TOC
Textiles Colouring matter TiO2/UV/O3-BAC 90% Dyes [93]
50%TOC
Colouring matter O3and electro- >80% [94]
flocculation
PAEs (DMP, DEP, TiO2/UV/O3-BAC >94.9% [95]
DBP, DEHP)a
POPs (HCB, PBB003, TiO2/UV/O3-BAC >89.3% [95]
PBB10, PBB18, PBB52,
PBB103)b
Paper HMWcO3/biological treatment 80% [96]
Lignin products (e.g. O3/UV Complete [97]
phenols)
Organic load due to Partial
starch (e.g. (no data)
saccharides, carboxylic
acids)
Domestic/ LAS O3Complete [98]
other (aqueous
industries solution)
Agriculture Pesticides O3Complete [99]
O3/UV (aqueous
O3/H2O2solution)
Pharmacy Antibiotics O3>90% [100]
Steroid hormones,
beta-blockers,
X-ray contrast
media
aDimethyl phthalate (DMP), diethyl phthalate (DEP), dibutyl phthalate (DBP) and di(2-
ethylhexyl) phthalate (DEHP)
bHexachlorobenzene (HCB), 4-bromobiphenyl (PBB003), 2,6-dibromobiphenyl (PBB10),
2,2,6-tribromobiphenyl (PBB18), 2,2,5,5-tetrabromobiphenyl (PBB52) and 2,2,4,5,6-
pentabromobiphenyl (PBB103)
cHMW: organic compounds of high molecular weight
Organic pollutants such as phthalate esters (PAEs) and other POPs in raw
water are also efficiently eliminated by various processes which combine the
use of catalyst, UV radiation and biological activated carbon, TiO2/UV/O3-
BAC [94].
The treatment of wastewater effluents generated in the industry of pulp
and paper also needs the use of advanced technologies. Pulp mill wastew-
Ozone-Based Technologies in Water and Wastewater Treatment 165
ater contains a significant amount of complex organic compounds of high
molecular weight (MW >1000 Da). It means that the treatment is not feas-
ible by biological methods. The use of ozone-based processes for this type
of effluent has demonstrated the capacity of this technology in enhancing
the biodegradability, decreasing the toxicity and increasing the removal of
organic pollutants from the effluents [95].
The advantage of the combination of a pre-treatment by ozone (dosage
of 0.7–0.8 mg O3/mL wastewater) followed by biological treatment allows the
conversion of organic compounds of high molecular weight (HMW) to low
molecular weight (LMW), increasing the biodegradability from 5 to 50% [96].
AnimportantfactorinthisprocessistheeffectofpH.Forthetreatmentofal-
kaline bleach plant effluent, a superior performance of the ozonation process
is under basic pH conditions. It is due to the reaction of organic compounds
with molecular ozone and with oxidizing radicals, including the hydroxyl rad-
ical, which are effectively formed at high pH. When the paper industry does
not employ wood to obtain the pulp and uses recycled paper, the composition
of the effluent and pH is not the same, it is less basic.
Another example of the application of advanced technologies using ozone
is the treatment of effluents which contain organic compounds arising from
the degradation of starch (e.g. saccharides, carboxylic acids), phenolic com-
pounds derived from lignin and smaller amounts of other pollutants that can
be persistent in the environment and are detected in fresh water (e.g. surfac-
tants). The studies performed with O3and/or UV have shown the utility of
this approach. By this procedure, the complete degradation of lignin prod-
ucts and their diminishment has been demonstrated, but the organic load
due to starch is not removed. The results obtained in these studies show that
toxic or inhibitory compounds (e.g. phenols) are more easily oxidized than
the highly biodegradable ones (e.g. glucose, fatty acids) by ozone-based tech-
nologies [97].
Linear alkylbenzene sulphonates (LASs) are anionic surfactants which are
discharged into wastewaters through different sources (domestic or indus-
trial), reaching aquatic compartments given their wide use. As a reference,
in 1995 the world production of LASs was ca. 2.8 ×106tonnes but now more
than 4×106tonnes are consumed globally every year. Few reports can be
found on LAS degradation. AOPs have been considered as strong oxidation
procedures for the degradation of such organic contaminants. The use of
ozonation has been proved as the most efficient approach for degrading the
typical LAS mix present in municipal and industrial wastewaters where the
typical pH values are slightly basic [98].
Hazardous organic contaminants such as pesticides when discharged into
the environment represent a risk for human health and for the ecosystem due
to their toxicity and persistence. Several publications have shown the effec-
tiveness of ozone for removing pesticides in aqueous solution [99]. Ozonation
appears to be a more efficient technique which can be easily implemented
166 A. Rodríguez et al.
with UV and/or H2O2for treating wastewater with high organic loads. The
efficiency depends, to a great extent, on the nature of the pollutant, and up
to now few experiences have been explored and documented for real con-
ditions [99]. AOPs are technologies whose position in the water treatment
processes for industry still need to demonstrate levels of reliability and full-
scale implementation. For the assessment of the removal efficiency of AOP
processes, common procedures are based on the measurement of global pa-
rameters (i.e. TOC). The use of techniques such as gas chromatography–mass
spectrometry (GC-MS) or liquid chromatography–mass spectrometry (LC-
MS) provides analytical information appropriate for further efficiency evalua-
tion and facilitates identification of by-products which can be of environmen-
tal concern. The use of toxicity assays in combination with chemical analysis
has been considered as a strategic approach for overall assessment [99].
The effluents originating from the pharmaceutical industry can show low
biodegradability since they contain active substances. In particular, certain
antibiotics, anti-tumour agents and analgesics are neither degradable nor ad-
sorbable on sewage sludge. AOPs applied to remove pharmaceuticals based
on the use ozone [100] are able to completely oxidize recalcitrant compounds,
rendering them less harmful and forming easily biodegradable components,
but also combinations of AOPs have enlarged the possibilities to treat tar-
get recalcitrant pollutants [100]. Investigations carried out with antibiotics,
steroid hormones, beta-blockers or X-ray contrast media have demonstrated
the removal capacity of ozone-based processes, achieving significant elimina-
tion of those pharmaceutics in effluents (>90%).
2.6
Removal Efficiency of Pharmaceuticals in Wastewater: A Case Study
As has been commented above, it is well established that urban wastewaters,
which include domestic and some industrial waters, among others, represent
a significant source of contamination with a strongly contaminating effect
on natural aquatic systems [101–103]. Even when they are submitted to bi-
ological treatment, it has been demonstrated by many studies that multiple
organic compounds, such as pharmaceuticals, PCPs, hormones and other dis-
rupting compounds, escape conventional wastewater treatments and some of
them are becoming ubiquitous in the environment [104]. The presence of
these contaminants in treated waters also endangers their reuse in diverse
applications, an aspect which is of special interest since the availability of wa-
ter of good quality is a critical issue and represents an essential component
for sustainable socio-economic development. Consequently, the application
of more exhaustive wastewater treatment protocols, including the use of new
andimprovedtechnologies,theapplicationofwiderandintegratedqual-
ity control strategies comprising chemical, microbiological and toxicological
analysis, and the study and development of wastewater reuse strategies are
Ozone-Based Technologies in Water and Wastewater Treatment 167
goals which it is necessary undertake [105]. With this aim, an ambitious
programme (TRAGUA) financed by the Spanish Government has been initi-
ated, which attempts to tackle the different aspects involved in the reuse of
wastewater coming from sewage treatment plants (STPs). As a part of this
programme, results regarding the efficiency of O3and O3/H2O2treatments
in the removal of organic contaminants in a municipal wastewater treatment
plant (WWTP) effluent are presented.
A suitable analytical methodology was developed in order to obtain an
adequate evaluation of the processes. Two LC-MS systems equipped with
modern and sensitive mass spectrometers, hybrid triple-quadrupole linear
ion trap (QTRAP) and time-of-flight (TOF), were used with this aim. The
joint application of both techniques provided very good results in terms of ac-
curate quantification and unequivocal confirmation of the organic pollutants
present in the samples. Quantification was performed by LC-QTRAP-MS op-
erating in the selected reaction monitoring (SRM) mode with both positive
and negative electrospray ionization, in order to cover a broad range of ana-
lytes. Limits of detection reached by the optimized method were between 0.04
and 50 ng L–1, thus guaranteeing an exhaustive evaluation of the samples.
Unequivocal analyte confirmation was provided by LC-TOF-MS analysis,
which allows accurate mass measurements of the identified compounds to
be obtained with errors lower than 2ppm.Figure10showsasanexample
the identification by TOF-MS of codeine and acetaminophen, based on the
accurate mass of their molecular ions and of their main fragments.
With the application of the developed method, up to 40 compounds were
identified in a wastewater effluent after the application of a conventional
biological treatment. They include mainly pharmaceuticals of different thera-
peutic groups, such as analgesics/anti-inflammatories, antibiotics, lipid regu-
lators, beta-blockers, antidepressants, anti-epileptics/psychiatrics, ulcer heal-
ing compounds, diuretics and bronchodilators.
The occurrence of many of these compounds has already been reported
in environmental waters [106, 107]. Also of interest was the presence of some
of their metabolites, such as 1,7-dimethylxanthine (paraxanthine) or fenofib-
ric acid, and especially the metabolites of the antipyretic drug dipyrone
and its active product 4-methylaminoantipyrine (4-MAA), such as N-acetyl-
4-aminoantipyrine (4-AAA), N-formyl-4-aminoantipyrine (4-FAA) and an-
tipyrine, which were detected at a high level of concentration. Finally, the
disinfectant chlorophene and the pesticide diuron completed the group of
detected compounds.
All of them are listed in Table 7, where the concentrations found in
an effluent sample of a municipal WWTP are also shown. Concentration
values ranged from 2 to 6590 ng L–1. The stimulant caffeine, the diuretic hy-
drochlorothiazide, the beta-blocker atenolol, the analgesic/anti-inflammatory
naproxen, the antibiotic ciprofloxacin and the metabolites of dipyrone (4-
AAA and 4-FAA) were the compounds present at the highest concentration
168 A. Rodríguez et al.
Fig. 10 Example of the identification by TOF-MS of codeine and acetaminophen in an
effluent sample based on accurate mass measurements
Ozone-Based Technologies in Water and Wastewater Treatment 169
Table 7 Compounds and concentrations present in a WWTP effluent and their removal
efficiency after O3and O3/H2O2treatment
Compound Concentration Removal Removal
in the effluent efficiency efficiency
(ng L–1)(%)O
3(%) O3/H2O2
Erythromycin 341 98 100
Ciprofloxacin 2559 95 99
Sulfamethoxazole 243 97 100
Mepivacaine 2 90 100
Caffeine 600 93 98
Omeprazole 104 98 100
Carbamazepine 140 98 100
Codeine 657 100 100
Ketorolac 465 98 100
Paraxanthine 132 87 98
Atenolol 1443 99 100
Naproxen 1990 98 100
Indomethacin 37 100 100
Propanolol 59 96 100
4-MAA 18 100 100
Diazepan 5 100 100
Metoprolol 53 100 100
Ranitidine 336 99 100
Fluoxetine 782 91 100
Trimethoprim 157 98 100
Metronidazole 185 100 100
4-FAA 3191 100 100
Antipyrine 17 100 100
4-AAA 6590 98 100
Ofloxacin 316 97 100
Salbutamol 10 100 100
Ketoprofen 590 98 99
Mefenamic acid 64 96 100
Sotalol 25 100 100
Terbutaline 11 100 100
Fenofibric acid 165 100 100
Furosemide 531 98 99
Diclofenac 33 100 100
Benzafibrate 61 90 96
Gemfibrozil 143 88 100
Hydrochlorothiazide 1310 96 100
Chlorophene 88 86 87
Diuron 9 89 95
Ibuprofen 52 84 96
170 A. Rodríguez et al.
Fig. 11 Total charge of contaminants present in a WWTP effluent before and after the
application of O3and O3/H2O2treatments
levels (>1µgL
–1). This group of compounds represents about 75%ofthe
total charge of these contaminants in the effluent, as can be observed in
Fig. 11.
The characterized effluent from the secondary clarification of a municipal
wastewater treatment facility (Table 7) was submitted to treatment with O3
and O3/H2O2.TheO
3treatments were carried out at 25 ◦Cina5-L stirred
tank agitated at 1000 rpm with a four-blade turbine. The gas, a mixture
of ozone and oxygen with a 45.9 gNm
–3 ozone concentration, was bubbled
at a rate of 0.36 Nm3h–1. During the experiment the pH was in the range
8.04–8.25. The same experimental conditions were maintained in O3/H2O2,
but now equal volumes (0.15 mL)ofhydrogenperoxide(30%w/v)werein-
jected every 5min in order to favour pollutant elimination by the radical
pathway.
The results obtained demonstrated that ozonation of wastewaters degrades
pharmaceuticals with a high efficiency. Removals higher than 90%were
reached for most of the target analytes. Only a group of five compounds, gem-
fibrozil, chlorophene, diuron, ibuprofen and mefenamic acid, yielded lower
removal efficiencies, which were higher than 84% in all cases. The com-
bination of ozone and hydrogen peroxide still enhances oxidizing ability,
providing almost total elimination of the contaminants in most cases.
Considering the total charge of compounds initially present in the waste-
water, a reduction of 97%wasobservedafterO
3treatment and their al-
most total elimination (99%) was reached by the application of O3/H2O2,
as is shown in Fig. 11. With these results, it can be concluded that both
treatments can be considered as promising alternatives for pharmaceuti-
cals and related compounds which persist through conventional biological
treatments.
Ozone-Based Technologies in Water and Wastewater Treatment 171
3
Conclusions
Ozone is an efficient oxidant of organic matter but its production is expensive.
To optimize the use of ozone it is coupled with coagulation and filtration pro-
cesses in surface or ground water treatments. It is used alone or coupled with
UV in water disinfection processes or it is coupled with other oxidants, energy
forms or catalysts in AOPs based on ozonein industrial wastewater treatment.
The process of hydroxyl radical generation from ozone/hydrogen perox-
ide was modelled in the 1980s, which made it possible to optimize the use of
ozone in the elimination of hazardous pollutants, such as pesticides, PAHs,
etc. Now the system ozone/hydrogen peroxide is a new choice for water recla-
mation and potable reuse. The use of the ozone/hydrogen peroxide system
as a tertiary treatment of domestic and urban wastewater could provide re-
claimed water to use in agriculture or industrial processes. The model of these
processes connects the grade of elimination of TOC and ozone doses with
the ct-exposure parameter, being the rate of TOC elimination described by
a first-order kinetic equation with a kinetic parameter Rwhich is obtained
by multiplying the hydroxyl-to-ozone ratio, Rct, and the elimination kinetic
constant of TOC, kHO·.
The efficiency of homogeneous catalytic ozonation has been reported for
several metals, especially iron and manganese. The reaction mechanism in-
volves the oxidation of a reduced form of the metal by ozone, hydroxyl
radicals or hydrogen peroxide followed by interaction with the organic com-
pounds. Heterogeneous catalytic ozonation is a complex process whose un-
derlying chemistry is not well known. Several mechanisms have been pro-
posed for describing it that can be classified according to the kind of surface
interaction proposed. A Langmuir–Hinshelwood rate expression may account
for a reaction between adsorbed organics and oxidized catalytic sites, while
an Eley–Rideal model can explain the direct oxidation of an adsorbed organic
compound by hydroxyl radicals from the bulk. The ozonation on activated
carbon seems to be based on the role of the surface as initiator of the radical
chain reactions that transform ozone into radicals.
Other mechanisms exclude adsorption equilibrium and lead to models in
which the rate of the catalytic process is not dependent on the concentra-
tion of oxidant. An adsorption-limited kinetics seems to be more realistic
considering the difficulty of adsorption encountered by organics in aqueous
solutions, especially on the surface of oxides. Depending on the pH of the so-
lution, the surface of an oxide may be charged or not. On a neutral surface,
the adsorbate must displace water coordination molecules and at basic con-
ditions, Lewis sites would be inhibited by the hydroxide anion. On oxides,
such as titanium dioxide, the reaction is probably better described by an in-
teraction between Lewis acid sites and organic molecules, with an optimum
mineralization rate obtained in slightly acidic conditions.
172 A. Rodríguez et al.
The ozonation reaction of individual compounds showed that the ozona-
tion starts with a rapid mineralization period followed by a slow decay
of the organic carbon associated with the accumulation of refractory com-
pounds. Some other circumstances complicate the modelling and description
of a catalytic ozonation process. For example, the distribution of reaction
products markedly differs from that encountered in non-catalytic reactions.
Understanding the role of catalyst in the inhibition of the ozone decompos-
ition reaction and the determination of values of the hydroxyl-to-ozone ratio,
Rct, a parameter that may change during the reactionand that the catalyst can
modify, are additional difficulties.
Acknowledgements The authors wish to express their gratitude to the Ministry of Educa-
tion of Spain (Contracts CTM2005-03080/TECNO, CTM2004-06265-C03-03 (EVITA) and
CONSOLIDER-INGENIO 2010 CSD2006-00044), the Dirección General de Universidades
e Investigación de la Comunidad de Madrid under Contract No. PAMB-000395-0505 and
the research network from Comunidad de Madrid REMTAVARES Ref. 0505/AMB-0395.
References
1. Andreozzi R, Caprio V, Ermellino I, Insola A, Tufano V (1996) Ind Eng Chem Res
35:1467–1471
2. Rischbieter E, Stein H, Shumpe A (2000) J Chem Eng Data 45:338–340
3. Roth JA, Sullivan DE (1981) Ind Eng Chem Fundam 20:137–140
4. Sotelo JL, Beltran FJ, Benitez FJ, Beltran-Heredia J (1989) Water Res 23:1239–1246
5. Beltran FJ, García-Araya JF, Encinar JM (1997) Ozone Sci Eng 19:281–296
6. Charpentier JC (1981) Mass-transfer rates in gas–liquid absorbers and reactors. In:
Drew TB, Cokelet GR, Hoopes HW Jr, Vermeulen T (eds) Advances in chemical en-
gineering, vol 11. Academic, New York, pp 3–133
7. Danckwerts PV (1970) Gas–liquid reactions. McGraw-Hill, New York, p 113
8. Onda K, Sada E, Kobayashi T, Fujine M (1979) Chem Eng Sci 25:761–768
9. Beltran FJ, Encinar JM, García-Araya JF, Alonso MA (1992) Ozone Sci Eng 14:303–
327
10. Poling BE, Prausnitz JM, O’Connell JP (2001) The properties of gases and liquids.
McGraw-Hill, New York, p 25
11. Beltrán FJ (2004) Ozone reaction kinetics for water and wastewater systems. CRC,
Boca Raton, pp 70–87
12. Chalmet S, Ruiz-López M (2006) J Chem Phys 124:1945
13. Buhler RE, Staehelin S, Hoigné J (1984) J Phys Chem 88:2560–2564
14. Staehelin S, Buhler RE, Hoigné J (1984) J Phys Chem 88:5999–6004
15. Tomiyasu H, Fukutomi H, Gordon G (1985) Inorg Chem 24:2962–2966
16. Hoigné J (1998) Chemistry of aqueous ozone and transformation of pollutants by
ozone and advanced oxidation processes. In: Handbook of environmental chemistry,
vol5,partC.QualityandtreatmentofdrinkingwaterII.Springer,Berlin
17. Staehelin S, Hoigné J (1983) Vom Wasser 61:337–348
18. Kasprzyk-Hordern B, Ziolek M, Nawrocki J (2003) Appl Catal B Environ 46:639–669
19. Lin J, Kawai A, Nakajima T (2002) Appl Catal B Environ 39:157–165
20. Logemann FP, Annee JHJ (1997) Water Sci Technol 35:353–360
Ozone-Based Technologies in Water and Wastewater Treatment 173
21. Sánchez-Polo M, von Gunten U, Rivera-Utrilla J (2005) Water Res 39:3189–3198
22. Rivera-Utrilla J, Sánchez-Polo M, Mondaca MA, Zaror CA (2002) J Chem Technol
Biotechnol 77:883–890
23. von Gunten U (2003) Water Res 37:1443–1467
24. Elovitz MS, von Gunten U (1999) Ozone Sci Eng 21:239–260
25. Rosal R, Rodríguez A, Zerhouni M (2006) Appl Catal A Gen 305:169–175
26. Beenackers AACM, van Swaaij WPM (1993) Chem Eng Sci 48:3109–3139
27. Acero JL, von Gunten U (2001) J Am Water Works Assoc 93:99–100
28. Buffle MO, Schumacher J, Meylan S, Jekel M, Gunten U (2006) Ozone Sci Eng 28:245–
249
29. Buffle MO, Schumacher J, Salhi E, Jekel M, Gunten U (2006) Water Res 40:1884–1894
30. Langlais B, Reckhow DA, Brink DR (1991) Ozone in water treatment: application and
engineering. AWWA Research Foundation and Lewis, Denver
31. Sonntag C (2007) Water Sci Technol 55:19–23
32. Barreto R, Gray KA, Andres K (1995) Water Res 29:1243–1248
33. Albarran G, Schuler RH (2005) J Phys Chem A 109:9363–9370
34. Palmisano G, Addamo M, Augugliaro V, Caronna T, Paola A, García E, Loddo V,
Marci G, Palmisano L, Schiavello M (2007) Catal Today 122:118–127
35. Fang X, Schuchmann HP, Sonntag J (2000) J Chem Soc Perkin Trans 2 1391–1398
36. Legube B, Karpel N (1999) Catal Today 53:61–72
37. Sallanko J, Lakso E, Röpelinen J (2006) Ozone Sci Eng 28:269–273
38. Langlais B, Reckhow DA, Brink DR (eds) (1991) Ozone in drinking water treatment:
application and engineering. AWWARF and Lewis, Boca Raton
39. Ramírez F (2005) Tratamiento de desinfección de agua potable. Canal de Isabel II,
Madrid, p 9
40. Viraraghavan T, Subramanian KS, Aruldoss JA (1999) Water Sci Technol 40:69–76
41. Nishimura T, Umetsu Y (2001) Hydrometallurgy 62:83–92
42. Kim MJ, Nriagu J (2000) Sci Total Environ 247:71–79
43. von Gunten U (2003) Water Res 37:1469–1487
44. USEPA (1989) National primary drinking water regulations: final rules 54:27485–
27541
45. von Gunten U, Elovitz MS, Kaiser HP (1999) J Water Supply Aqua 48:250
46. Do-Quang Z, Roustan M, Duguet JP (2000) Ozone Sci Eng 22:99–111
47. Xu P, Janex ML, Savoye P, Cockx A, Lazarova V (2002) Water Res 36:1043–1055
48. Camel V, Bermond A (1998) Water Res 32:3280–3222
49. Rice RP (1999) Ozone Sci Eng 21:99–118
50. Larocque RL (1999) Ozone Sci Eng 21:119–125
51. Matsumoto N, Watanable K (1999) Ozone Sci Eng 21:127–138
52. Kruithof CJ, Masschelin W (1999) Ozone Sci Eng 21:139–152
53. Paulose J, Langlais B (1999) Ozone Sci Eng 21:153–162
54. Böhme A (1999) Ozone Sci Eng 21:163–176
55. Pollo I (1999) Ozone Sci Eng 21:177–186
56. Geering F (1999) Ozone Sci Eng 21:1187–1200
57. Lowndes R (1999) Ozone Sci Eng 21:201–205
58. Haag WR, Hoigné J (1985) Chemosphere 14:1569–1671
59. Hoigné J (1997) Water Sci Technol 35:1–8
60. Yurteri C, Gurol DM (1988) Ozone Sci Eng 10:277–290
61. Glaze WH, Kang J-W (1989) Ind Eng Chem Res 28:1573–1580
62. Glaze WH, Kang J-W (1989) Ind Eng Chem Res 28:1580–1587
63. Beltran FJ, Rivas J, Alvarez PM, Alonso MA (1999) Ind Eng Chem Res 38:4189–4199
174 A. Rodríguez et al.
64. Legube B (2003) Ozonation by-products. In: Handbook of environmental chemistry,
vol5,partG.Springer,Berlin,pp95–116
65. Andreozzi R, Insola A, Caprio V, D’Amore MG (1992) Water Res 26:917–925
66. Andreozzi R, Marotta R, Sanchirico R (2000) J Chem Technol Biotechnol 75:59–65
67. Beltrán FJ, Rivas FJ, Montero R (2002) Appl Catal B Environ 39:221–231
68. Beltrán FJ, Rivas FJ, Montero R (2003) J Chem Technol Biotechnol 78:1225–1233
69. Beltrán FJ, Rivas FJ, Montero R (2005) Water Res 39:3553–3564
70. Rivera-Utrilla J, Sánchez-Polo M (2002) Appl Catal B Environ 39:319–329
71. Faria PCC, Órfao JJM, Pereira MFR (2005) Water Res 39:1461–1470
72. Andreozzi R, Casale MS, Marotta R, Pinto G, Pollio A (2000) Water Res 34:4419–4429
73. Gracia R, Cortes S, Sarasa J, Ormad P, Ovelleiro JL (2000) Water Res 34:1525–1532
74. Ma J, Graham NJD (1999) Water Res 33:785–793
75. Lin J, Nakajima T, Jomoto T, Hiraiwa K (1999) Ozone Sci Eng 21:241–247
76. Karpel N, Delouane B, Legube B, Luck F (1999) Ozone Sci Eng 21:261–276
77. Fu H, Karpel N, Legube B (2002) New J Chem 26:1662–1666
78. Hewes CG, Davinson RR (1972) Water AIChE Symp Ser 70:69–71
79. Gracia R, Aragües JL, Cortés S, Ovelleiro JL (1996) Ozone Sci Eng 18:195–208
80. Piera E, Calpe JC, Brillas E, Domenech X, Peral J (2000) Appl Catal B Environ 27:169–
177
81. Nowell LH, Hoigné J (1987) Interaction of iron(II) and other transition metals with
aqueous ozone, 8th Ozone World Congress, Zurich, September 1987, p E80
82. Neyens E, Baeyens JH (2003) J Hazard Mater B29:33–50
83. Pines DS, Reckhow DA (2002) Environ Sci Technol 36:4046–4051
84. Dhandapani B, Oyama ST (1997) Appl Catal B Environ 11:129–166
85. Bulanin KM, Lavalley JC, Tsyganenko AA (1995) Colloids Surf A 101:153–158
86. Fernandez P, Nieves FJDL, Malato S (2000) J Colloid Interface Sci 227:510–516
87. Zhaobin W, Xienian G, Sham EL, Grange P, Delmon B (1990) Appl Catal 63:305–317
88. Kasprzyk-Hordern B (2004) Adv Colloid Interface Sci 110:19–48
89. Beltrán FJ, Rivas FJ, Monterio R (2004) Appl Catal B Environ 47:101–109
90. Vannice MA (2007) Catal Today 123:18–22
91. Agustina TE, Ang HM, Vareek VK (2005) J Photochem Photobiol C Photochem Rev
6:264–273
92. Moraes SG, Freire RS, Duran N (2000) Chemosphere 40:369–373
93. Kusic H, Koprivanac N, Bozic AL (2006) Chem Eng J 123:127–137
94. Ciardelli G, Ranieri N (2001) Water Res 35:567–572
95. Li L, Zhu W, Zhang P, Zhang Q, Zhang Z (2006) J Hazard Mater 128:145–149
96. Bijan L, Mohsein M (2005) Water Res 39:3763–3772
97. Amat AM, Arques A, Miranda MA, López F (2005) Chemosphere 60:1111–1117
98. Fernández J, Riu J, García-Calvo E, Rodríguez A, Fernández-Alba AR, Barceló D
(2004) Talanta 64:69–79
99. Chiron S, Fernández-Alba AR, Rodríguez A, García-Calvo E (2000) Water Res
34:366–377
100. Hernando MD, Petrovic M, Radjenovic J, Fernández-Alba AR, Barceló D (2007) Re-
moval of pharmaceuticals by advanced treatment technologies. In: Petrovic M,
Barceló D (eds) Analysis, fate and removal of pharmaceuticals in the water cycle.
Elsevier Science Ltd., Oxford, UK
101. Gómez MJ, Martínez Bueno MJ, Lacorte S, Fernández-Alba AR, Agüera A (2007)
Chemosphere 66:993–1002
102. Hernando MD, Mezcua M, Fernández-Alba AR, Barceló D (2006) Talanta 69:334–342
103. Zuccato E, Castiglioni S, Fanelli RE (2005) J Hazard Mater 122:205–209
Ozone-Based Technologies in Water and Wastewater Treatment 175
104. Carballa M, Omil F, Lema JM, Llompart M, García-Jares C, Rodríguez I, Gómez M,
Ternes TA (2004) Water Res 38:2918–2926
105. Hernando MD, Ferrer I, Agüera A, Fernández-Alba AR (2004) In: Barcelo D (ed)
Emerging organic pollutants in waste waters and sludge. Springer, Berlin, pp 53–77
106. Bendz D, Paxeus NA, Ginn TR, Loge FJ (2005) J Hazard Mater 122:195–204
107. Bound JP, Voulvoulis N (2004) Chemosphere 56:1143–1155
