3.1. Determination of Ozone Rate Constants for the Selected Antibiotics
Kinetics of both FF and OA ozonation indicate second-order reactions of the oxidant (ozone or bromine) and target compound, as shown in Equation (2) and (3).
where the terms [Antibiotic]
t, and [Ozone]
t represent the total concentrations of the selected pharmaceuticals FF and OA, and ozone (i.e., sum of the concentrations of the different species present in the solution), respectively, and k
app is the apparent second-order rate constant. When [Ozone]
t is in significant excess with respect to [Antibiotic]t, a pseudo-first-order rate constant k
obs can be introduced, which is equal to k
app[Ozone]
t. The second part of Equation (3) can be integrated to yield Equation (4) as follows.
This rate constant typically shows a pH dependence that can be explained by the presence of several species of both the oxidant and antibiotics in the aqueous solution. To determine the species-specific rate constants of FF and OA with ozone, ozonation at various pH conditions ranging from pH 3 to 9, were carried out. The apparent rate constant at each pH value was determined for the pseudo first-order reaction, using the ozone exposure method.
The target antibiotics assume only a single pKa value each in this study. The pKa values for FF and OA are 9 and 6.8, respectively. Considering the chemical characteristics of these antibiotics, the species of FF and OA can be expected to be either neutral or anionic (PhH, Ph
−). Thus, two species can be formed as shown in Equations (5) and (6) below:
The two species of antibiotics react with ozone according to Equations (7) and (8).
Hence, the removal of antibiotics can be expressed by Equation (9) below. Considering the equilibrium of the dissociation reactions of FF and OA, the apparent second-order rate constant k
app for FF and OA are given by Equation (10).
where,
,
,
.
The kinetic model represented by Equation (10) allows the determination of the intrinsic rate constants k1 and k2 for the elementary reactions of each pharmaceutical species with each oxidant species.
Because the reaction of FF with ozone presented intermediate or slow reactions in the pre-test phase of FF removal by ozone, the measuring rate constant, kapp, for FF was determined under the pseudo first-order condition with FF in excess ([FF]t >> [O3]t). In general, rate constants for ozonation depend on the speciation of the antibiotic compounds. The ease with which ozone reacts with a molecule will be determined by the structural and chemical properties of the molecule. Being an oxidizing agent, ozone is highly electrophilic in nature. Therefore, it is attracted to antibiotic molecules with the highest electron density. With increasing pH value, the decomposition rate constant of FF with ozone increases. Because the pKa value of FF is 9, FF has a neutral form at pH values lower than 9 and has an ionized form at pH values higher than 9. Therefore, the rate constants for the reaction of FF and OA with ozone increase with the pH value.
The experimental and calculated values for the apparent rate constants for the ozonation of FF are shown in
Figure 2a. The apparent rate constant for the second-order equation, was determined by the pseudo first-order method in the pH range 3–9. The red solid line represents the speciation of FF. The experimental values and calculated values matched well. From the calculated values, the species-specific rate constants k
1 and k
2 were determined. k
1 and k
2 were determined as 0.43 M
−1s
−1 and 33.5 M
−1s
−1, respectively. k
1 represents the neutral form of the reaction of FF with ozone and k
2 represents the ionized form of the reaction of FF with ozone. The k
2 value is approximately 80 times higher than the k1 value. That is, the ionized form of FF reacts with ozone faster than the neutral form of FF, because deprotonated species react faster with the electrophilic ozone.
To determine the rate constant for the reaction of OA with ozone, the ozone exposure method was used. This method is necessary to measure the ozone decomposition, and the data were evaluated by plotting ln ([OA]/[OA]
0) versus ozone exposure.
Figure 2b shows the reaction rate of OA with ozone, where ozonation experiments at different pH values from 3 to 8 were performed using the ozone exposure method. The slope of the represents the rate constant for the reaction of OA with ozone. In the case of OA, the apparent second-order rate constant was determined by the ozone exposure method conducted at a pH range of 3–9. The determined apparent rate constants for the reaction between ozone and OA are fitted with respect to their dependence on pH, in
Figure 2b. Like FF, the species-specific rate constant of OA was also determined by the nonlinear least-squares regression of the experimental value of the apparent rate constant at each pH, for ozonation. The experimental values and calculated values (dotted line) agree well. Further, the red solid line represents the speciation of OA. The species-specific rate constants k
1 and k
2 were determined to be 120 M
−1s
−1 and 2700 M
−1s
−1, respectively. k
1 represents the neutral form of the reaction of OA with ozone, and k
2 represents the ionized form of the reaction of OA with ozone.
3.2. Determination of Bromine Rate Constants for the Selected Antibiotics
The second-order rate constant of FF for reaction with bromine was determined using the pseudo first-order method, in the bromine-excess condition. The concentration of bromine was 3 mM for pH 3 to 6, and 5 mM for pH 7 to 9 with the pH adjusted using a 0.2 M phosphate buffer. The rate constant of FF for reaction with bromine was similar for different concentrations of bromine at the same pH value (
Table 2).
The species-specific rate constants for reaction of FF or OA with bromine were determined at various pH ranges. The apparent rate constants were determined based on Equations (11)–(16) with K
HOBr = 10
–8.8, K
FF = 10
–9, and K
OA = 10
–6.8.
The removal of FF and OA through bromination reactions corresponding to rate constants k
1, k
2, k
3, and k
4, can be expressed by Equation (17).
From Equations (11)–(16), the apparent second-order rate constant k
app for FF and OA, considering the equilibrium of the dissociation reactions of FF and OA, is derived, as given by Equation (18):
where,
,
,
,
,
, and
.
The species-specific rate constants were determined by a nonlinear least-squares regression of the experimental value of the apparent rate constant at each pH for the bromination process. The experimental and calculated values for the apparent rate constants for different values of pH in the range 3–10, for the bromination of FF are shown in
Figure 3a. The symbols represent the experimental values while the dotted line represents the calculated values of the apparent rate constant; the solid line represents the speciation of FF and bromine. The experimental values and calculated values matched well. The species-specific rate constants k
1, k
2, and k
3, were obtained as 0.055 M
−1s
−1, 15.4 M
−1s
−1, and 24.4 M
−1s
−1, respectively, and k
4 was observed to be negligible.
From
Table 1, FF includes an aromatic compound that has a high second-order rate constant for reaction with bromine. However, FF itself has a relatively low rate constant for reaction with bromine because the aromatic ring is deactivated or only slightly activated during bromination, which is characterized by a low reactivity with HOBr. In the case of OA, the apparent second-order rate constant was determined by the pseudo first-order method, for pH values between 5 and 10. The determined apparent rate constants for the reaction between bromine and OA are fitted as a function of pH in
Figure 3b. The experimental values and calculated values matched well. The solid line depicts the speciation of bromine and OA. The value of the species-specific rate constant was determined to be 5000 M
−1s
−1 for k
1, which indicates the reaction of a neutral form of OA with bromine, and 290 M
−1s
−1 for k
2, which indicates the reaction of an ionized form of OA reaction with bromine.
3.3. Prediction of Antibiotics Removal Efficiency in Absence of Br—In Synthetic Water
Seawater contains approximately 18,000 mg/L chloride ions and 60 mg/L bromide ions. As mentioned in the introduction, the primary oxidant in seawater ozonation is bromine (HOBr/OBr−), and not ozone, even with this high concentration of chloride ion in seawater. The reaction rate of ozone with bromide ion (160 M−1s−1) is 53,000 times higher than that with chloride ion (kOzone,Cl− = 0.003 M−1s−1). Even though chlorine is generated by the reaction of chloride ion with ozone, the chlorine reacts with bromide ion to generate bromine (2950 M−1s−1). In fact, bromine is the primary oxidant along with trace ozone, for seawater ozonation.
The removal efficiency for the target antibiotics, FF and OA, by seawater ozonation was predicted using the reaction rate constants for the reaction of FF and OA with ozone and bromine determined from this study. To verify the value predicted using the calculated reaction rate, the removal efficiencies of FF and OA from synthetic seawater were evaluated; synthetic sea water was used owing to the difficulties of analysis under real seawater condition. Synthetic seawater was prepared using bromide ions of concentration 60 mg/L, at pH 8, without chloride ions. In this test, the effects of ozone and bromine were considered only with respect to their role in antibiotic removal. The ozone decomposition and bromine formation were measured and compared with the predicted values. Subsequently, the actual removal efficiencies for the target antibiotics, FF and OA, were compared with the predicted values.
3.3.1. Formation of Bromine in Seawater Ozonation
Figure 4 shows the bromine concentration after injecting ozone into synthetic sea water containing bromide ions at pH 8. Reacting ozone with bromide ions in synthetic seawater, 3.1 mg/L of bromine was generated. A previous study [
21] predicted ozone decomposition and bromine formation in synthetic seawater. The predicted ozone formation indicated that ozone decomposed rapidly within 20 s, while bromine was generated and became stable after 40 s. The predicted bromine formation of 3.1 mg/L is the same as the experimental value for a dosage of 1 mg/L ozone into water containing 60 mg/L bromide ions at pH 8.
To assess the effect of antibiotics in water on bromine formation, the bromine formation was evaluated in water containing 1 μM of FF or OA.
Table 3 shows the comparison of apparent rate constants at given condition. In the water containing 1 μM FF, 3 mg/L of bromine was generated with 1 mg/L ozone. However, in the water containing 1 μM OA, a lower bromine concentration (2.6 mg/L) was generated compared to that in the pH 8 buffer containing only bromide (3.1 mg/L). This is because the rate constant for the reaction of ozone with OA (2400 M
−1s
−1) is 15 times that with bromide ion (160 M
−1s
−1). Because ozone reacted competitively with bromide ion and OA, less bromine was formed in the water containing OA. In the case of FF, the reaction rate between ozone and FF was extremely low at 3.2 M
−1s
−1. Therefore, the effect of FF is negligible in bromine formation.
3.3.2. Removal of FF and OA by Ozonation and Bromination
For determining the removal efficiency of antibiotics by ozone only, ozonation was performed in a pH 8 buffer with the target antibiotic, without the presence of bromide ion (
Figure 5). Less than 5% FF was removed because the reaction rate of FF with ozone is low, i.e., 3.2 M
−1s
−1. Thus, the direct removal of FF by ozone is negligible. In the case of OA, however, almost all OA was removed within 90 s by ozonation, owing to the high reaction rate of OA with ozone (2400 M
−1s
−1). For the removal of OA and FF through ozonation, the experimental values matched well with the predicted values. To assess the effectiveness of antibiotic removal through bromination, 3 mg/L and 2.6 mg/L of bromine were injected into water containing FF and OA, respectively.
Figure 6 shows the removal of FF and OA by bromination at pH 8. Like the case of ozonation, FF exhibits little removal within 20 min owing to the low reaction rate of FF with bromine (3.5 M
−1s
−1). However, approximately 40% of OA was removed within 90 s of bromination at 4.0 × 10
2 M
−1s
−1. The predictions of FF and OA removal by bromination matched well with the experimental values. The effectiveness of both ozone and bromine for FF and OA removal could be predicted using the reaction rate.
3.4. Removal of FF and OA by Ozonation in Water Containing Br−
The removal efficiency of FF and OA by ozonation was assessed in water containing 60 mg/L of bromide ion at pH 8.
Figure 7a shows the removal of FF with 1 and 2 mg/L of ozone injection. Because of the lower reactivity between FF and the oxidants, ozone, and bromine (k
O3,FF = 3.2 M
−1s
−1 and k
Bromine,FF = 3.5 M
−1s
−1), the removal of FF by ozonation in water containing bromide ion was only 4% and 8% within 20 min. In fact, it is difficult to remove FF from seawater by ozonation owing to the lower reactivity of FF with ozone and bromine.
Figure 7 shows the effect of bromide ion on OA removal by ozonation, compared with OA removal in the absence of bromide ion. In the absence of bromide ion, ozone reacted rapidly with OA, and over 99% of OA was removed in 90 s, as shown in
Figure 5. However, when bromide ion was present, the removal efficiency of OA was lower than that for ozonation without bromide ion, i.e., less than 10%. OA decreased rapidly within 15 s, and decreased slowly after 15 s. The primary mechanism of OA removal in the first phase of 15 s was trace ozone; subsequently OA was removed by the generated stable bromine. The measured reaction rates of OA with ozone and bromine at pH 8 were as 2.4 × 10
3 M
−1s
−1 and 4.0 × 10
2 M
−1s
−1, respectively. These results indicate that trace ozone is more important than bromine for OA removal owing to the reaction rate of ozone being the six times that of bromine. As mentioned in the introduction, both chlorine and ozone have been used in the chemical process for seawater treatment. In seawater chlorination, chlorine reacts rapidly with bromide ions to produce bromine. Ozone and chlorine (bromine) cannot be suggested for FF removal owing to their low reaction rates, thus other chemicals or treatment methods should be considered. To remove OA from seawater, ozonation process can be applied rather than bromination, owing to the rapid reaction rate of ozone with OA. For longer reaction times, both chlorine and bromine can be recommended for the treatment method.