Kinetics and mechanism study of direct ozonation organics in aqueous solution

Abstract

In this study, the kinetics and mechanism of direct ozonation organics in aqueous solution were explored. Phenoxyacetic acid was selected as the model pollutant and ozonation experiments were performed in a bubble batch reactor to determine the rate constants for the direct reaction. Two kinetic methods were used for determination of different kinetic rate constants (k_app and k_i). The first group of results showed that degradation of phenoxyacetic acid followed pseudo-first-order kinetics. A simplified model was derived related to the operational parameters on phenoxyacetic acid degradation, and the apparent rate constant k_app was obtained. The reaction was proved in the slow kinetics of the gas–liquid reaction, and the kinetic constant k_i was built. Influence of pH on k_app and k_i, the O₃ dosage, and the initial phenoxyacetic acid concentration were carefully analyzed.

---

1. Introduction

Development in industrial production has caused amounts and categories of pharmaceutical production to increase. Some pharmaceuticals, such as antibiotics and pharmaceutical intermediates, are persisting organic pollutants and cannot be treated with traditional biological technology if directly discharged into the environment.^1–4 Thus, identifying a suitable method for pharmaceutical wastewater treatment is one of the focuses of environmental workers and the government, with the aim of reducing the intolerable burden on the environment.^5–8

Some effective efforts have been made in reduction of the toxicity of pharmaceutical wastewater, such as advanced oxidation processes (AOPs).^9–11 Ozonation is a useful AOP in water treatment technology because ozone is an oxidizing agent showing high reactivity with different kinds of pharmaceuticals.^12,13 During ozonation, ozone either reacts selectively with organic pollutant or decomposes in water to form free hydroxyl radical, which is another stronger oxidizing agent with high potential (2.80 eV) leading to indirect reactions.¹⁴

The reaction of target pollutant B with molecular ozone (O₃) and hydroxyl radicals (˙OH) during ozonation is expressed in eqn (1) and (2) parallel reactions:¹⁵

B + O₃ → products (1)

B + ˙OH → products (2)

Thus, for an ozone-reacting compound B, the chemical reaction rate r_B can be expressed as follows:

r_B = −(zk_O3[B][O₃] + k_OH˙[B][˙OH]) (3)

The stoichiometric ratio z is the number of moles of compound B consumed per mole of ozone consumed. k_O3 is the second-order rate constant between ozone and compound B, and k_OH is the second-order rate constant between hydroxyl radical and compound B. [B], [O₃], and [˙OH] are the concentrations of target compound B, ozone, and hydroxyl radical, respectively.

Ozonation is a heterogeneous reaction and the kinetics of heterogeneous reactions are governed by absorption theories of gases in liquids along with chemical reactions.¹⁶ So far, the two most commonly applied absorption theories are the film and surface renewal theories. Two-film theory is relatively mature and has been widely applied. The basis of the theory is mainly diffusion law and the gas–liquid mass transfer model. Thus the main aim of ozonation kinetics is to derive the mass transfer coefficients and reaction rate constants. Eqn (4) represents the mass transfer of ozone from the gas phase to the liquid phase, where k_La is the volumetric mass transfer coefficient.

In view of the competition between ozone decomposition and the intermediates produced in the processes, absorption rate equations are unsuitable for irreversible second-order reactions.¹⁷ It is necessary to add a radical scavenger into the reactor to avoid ozone decomposition. The ozone decomposition mechanism involves an initiation reaction, a propagation reaction, and a termination reaction. The free radical species are exactly formed in the initiation or propagation reactions of ozone with agents such as hydroxyl and hydroperoxide ion. Among these, the hydroxyl radical is the main threat.¹⁸ tert-Butyl alcohol (TBA) reacts with hydroxyl radicals quickly (k = 5 × 10⁸ M⁻¹ s⁻¹) and O₃ slowly (k = 0.03 M⁻¹ s⁻¹),^15,19 therefore, TBA can be used as a hydroxyl radical scavenger.

Although there have been some studies on oxidation of relevant pharmaceuticals,¹² the developing kinetic aspects are still limited. In this study, phenoxyacetic acid (PAA) is selected as the model pollutant because it is an important pharmaceutical intermediate and widely applied in the pharmaceutical industry. It can be used as the parent molecule of herbicides, highly toxic compounds applied in the agricultural fields, and also in preparation of cefazolin, penicillin, and meclofenoxate.

The aim of this study was to establish a database with kinetic rate constants for the direct reaction of PAA degradation. Two kinetic methods for PAA degradation in the ozonation system were used and corresponding models were built. One relatively macroscopic kinetics model was established with the ozone dosage, the initial concentration of phenoxyacetic acid, and the pH value, which could be used easily in highly concentrated organic wastewater pollution treatment with ozonation. Another model studying the characteristics of the initial kinetic to reflect the reaction rate constant was also discussed. This paper provides basic data and kinetic reference for ozonation of pharmaceutical wastewater.

2. Experimental

2.1 Materials and reagents

Phenoxyacetic acid was purchased from Shanghai J&K Scientific Ltd. (China), (purity 99%). tert-Butyl alcohol (TBA) was chosen as the radical scavenger and purchased from Shanghai Lingfeng Chemical Reagent (China) Co., Ltd.

All other chemicals used in the experimental process were analytical reagent or guaranteed reagent. Ultra-pure water was used as experimental water. The pH value of the solutions was adjusted with sodium hydroxide.

2.2 Apparatus and methods

Ozone was produced by an ozone generator (CFY-3) with high purity oxygen (99.999%). The experiments were carried out in a cylindrical Pyrex glass reactor (inner circulation) with a capacity of 1.5 L at room temperature. Sample connection was set on top of the reactor. An aeration head was fixed at the center of the bottom of the reactor for gas distribution. At the beginning of the reaction, small bubbles of ozone were placed in sufficient contact with the solution through the aeration head, thus forming the gas–liquid two-phase. Samples were withdrawn at fixed times, filtered through 0.45 μm pore size membrane filters, and analyzed.

2.3 Analytical methods

Phenoxyacetic acid (PAA) was detected by high performance liquid chromatography (1200, Agilent Technologies, USA) with a UV-detector at 210 nm using a C18 reversed phase column. The column temperature was 30 °C. The mobile phase was a mixture of ammonium biphosphate (pH = 7, adjusted by ammonia water) and acetonitrile at 85 : 15 (v/v) at 0.8 mL min⁻¹ flow rate. The injection volume was 5 μL. Intermediates produced in the ozonation/TBA system were detected using HPLC, IC, and GC-MS technologies. The methods were performed as described previously.¹⁷

3. Results and discussion

3.1 Effect of tert-butyl alcohol on PAA removal

Fig. 1 shows the effect of the presence and absence of tert-butyl alcohol (TBA) in acid, neutral, and alkaline conditions for PAA removal. The initial pH was investigated in the range of 3.0–11.0. The effect of pH from 3.0 to 7.0 on PAA removal was non-significant, but showed a rising trend in alkaline environment without addition of TBA. The presence of TBA caused a negative acceleration of PAA removal compared with experiments carried out under the same operating conditions in the absence of TBA. It is likely that when pH is lower than 12.0, the ozone decomposition has to compete for the available ozone that reacts with pollutants.¹⁸ After 50 min at pH 3.0 without TBA, 79.22% of PAA was removed, compared with 67.14% with TBA. This is evident of a competitive effect of ozone decomposition. In this case, the ozone decomposition can be stopped by addition of a scavenger of hydroxyl radicals (TBA).

Fig. 1 Effect of presence and absence of TBA on degradation of phenoxyacetic acid. Experimental conditions: [PAA]₀ = 1000 mg L⁻¹, ozone dosage = 40 mg min⁻¹.

Intermediates formed during ozonation with TBA were identified by HPLC, IC, and GC-MS technologies. By comparison with intermediates formed during ozonation alone, as described in a previous article,¹⁷ it was found that TBA rarely affected the categories of intermediates.

3.2 Kinetic model of pseudo first-order reaction

Ozonation experiments were performed in the presence of scavengers of hydroxyl radicals to guarantee direct ozonation in the environment, and aimed at studying the kinetic models of direct ozone reactions between ozone and phenoxyacetic acid (PAA). It can be assumed that the second term on the right side in eqn (3) is negligible under experimental conditions with ozone in great excess.

r_PAA = −k_app[PAA] (5)

where k_app is the kinetic rate constant of pseudo first-order reaction.

Integration of the apparent rate constant leads to

where [PAA]₀ is the initial concentration of target compound PAA. The apparent rate constant k_app was used to reflect the effect of the ozone dosage, initial concentration of PAA, and the pH value on PAA degradation in 50 min by ozonation. Table 1 shows that the apparent rate constant k_app fits well the pseudo-first-order kinetics with linear correlation coefficient (R²) over 0.99 under various experimental conditions. Variation of the ozone dosage (16, 24, 32, and 40 mg min⁻¹) resulted in k_app of 0.51 × 10⁻², 0.79 × 10⁻², 1.28 × 10⁻², and 1.68 × 10⁻², respectively. Hence, the ozone dosage was proportional to the apparent rate constant k_app, while the initial concentration of PAA was inversely proportional. The pH value impeded the observation of any significant difference in the apparent rate constant k_app. Based on optimization of operational factors, the apparent rate constant k_app is hypothesized to be related to ozone dosage (Q_O3), the initial concentration of PAA ([PAA]₀), the pH value (OH⁻), and the temperature (T).²⁰

Table 1 Apparent rate constant based on experimental results

No.	T/K	pH	[PAA]0/mg L^−1	TBA/mmol L^−1	QO3/mg min^−1	R^2	kapp/min^−1
1	298	3	200	50	40	0.9915	9.68 × 10^−2
2	298	3	500	50	40	0.9939	3.64 × 10^−2
3	298	3	1000	50	40	0.9934	1.68 × 10^−2
4	298	3	2000	50	40	0.9962	0.62 × 10^−2
5	298	3	1000	50	16	0.9970	0.51 × 10^−2
6	298	3	1000	50	24	0.9971	0.79 × 10^−2
7	298	3	1000	50	32	0.9976	1.28 × 10^−2
8	298	5	1000	50	40	0.9952	1.75 × 10^−2
9	298	7	1000	50	40	0.9973	1.61 × 10^−2
10	298	9	1000	50	40	0.9967	1.70 × 10^−2
11	298	11	1000	50	40	0.9952	1.70 × 10^−2

---

If relatively macroscopic kinetics are established with optimization of operational factors, this will allow control of the highly concentrated organic wastewater pollution by ozonation. As dissolved ozone concentration is difficult to calculate, a relationship between k_app and O₃ dosage is supposed. Therefore, an empirical equation forms as suggested by Dai et al.:²¹

Eqn (7) expressed in logarithmic form becomes

Regression analysis was conducted between the first item and the O₃ dosage, the initial concentration of PAA, and the pH value, respectively. Corresponding α and β were equal to 1.3273 and −1.1816 (Fig. 2a and b) with linear correlation coefficients of 0.9918 and 0.9948, respectively. As a result of the nonlinear relationship between ln k_app and ln[OH⁻], γ was regarded as 0 (Fig. 2c).

Fig. 2 Pseudo first-order plots of PAA as a function of different parameters on the logarithmic scale. (a) ln k_app versus ln Q_O3, fitted curve; (b) ln k_app versus ln[PAA]₀, fitted curve; (c) ln k_app versus ln[OH⁻], fitted curve.

Thus eqn (8) can be converted to

As the experiments were done at room temperature, A exp(−E_a/RT) was obtained from substitution of all experimental data into eqn (9). The result was equal to 0.51. Finally, the kinetic model turns into

[PAA]_t = [PAA]₀exp(−0.51Q_O3^1.3273[PAA]^−1.1816₀t) (10)

3.3 Kinetic model in initial stage

3.3.1 Validation of the reaction in the slow kinetic regime. Ha is the dimensionless Hatta number for an irreversible first-order reaction. For such a reaction, the mass transfer rate of the gas–liquid system is defined as follows:where Ha indicates the relative importance of chemical reaction and D_O3 is the diffusion coefficient of O₃ in water. k_L is the individual mass transfer coefficient of ozone in the liquid phase. As a function of the Hatta number, the ozonation kinetic regimes can be classified as fast, moderate, and slow reactions. When Ha is more than 3, the reaction is in the fast kinetic regime; when Ha is between 0.3 and 3, it is in the moderate kinetic regime; the third regime is called the slow kinetic regime when Ha is less than 0.3.²²

Some research^23–25 has been done to study D_O3. In the ozonation process, the empirical correlation of Johnson and Davis²⁵ can be used to determine D_O3:

D_O3 = (5.9 ± 0.04) × 10⁻¹⁰T/μ_s (12)

where μ_s = the viscosity of the solvent. Thus, determination of D_O3, which was equal to 1.4 × 10⁻⁹ m² s⁻¹, was obtained from the calculation result of eqn (12).

Determination of the individual mass transfer coefficient k_L in liquid phase is dependent on type of reactors. For the bubble batch reactor in the experiment, k_L could be determined by detecting mass transfer of CO₂. Based on the absorption reaction of CO₂ in the Na₂CO₃–NaHCO₃ system, with hypochlorite ion and NaAsO₂ as the catalysts, k^′_L, the individual mass transfer coefficient of CO₂, could be obtained using the plot method of Roberts and Danckwerts.²⁶ The basic requirement was that the absorption reaction of CO₂ was the pseudo first-order reaction. The absorption rate is defined as follows:

where N_CO2 is the absorption rate per unit volume, [CO₂]* is the saturated solubility of CO₂ at equilibrium in the buffer solution, D_CO2 is the diffusion coefficient of CO₂, and k_CO2 is the kinetic rate constant of pseudo first-order reaction of CO₂ in the absorption system.

Then eqn (13) can be converted into

(N_CO2/[CO₂]*)² = (ak^′_L)² + a²k_CO2D_CO2 (14)

From eqn (14), it can be seen that (N_CO2/[CO₂]*)² ∼ k_CO2D_CO2 shows a linear relationship. k^′_L could be attained by variation of k_CO2D_CO2. D_CO2 was equal to 1.58 × 10⁻⁹ m² s⁻¹. Finally, k_L is determined by the calculation of eqn (15).²⁷

k_app, D_O3, and k_L measured under different conditions were taken into account in eqn (11) to yield the Ha constant shown in Table 2. The calculated results indicated the reaction was in the slow kinetic regime because all Ha constants were lower than 0.3 in PAA degradation by ozonation, and also indicated that the reaction of ozone decomposition took place in the bulk water.

Table 2 Factors of kinetic rate constants equation in the primary stage

No.	T/K	pH	[PAA]0/mg L^−1	QO3/mg min^−1	kL/m s^−1	a	Ha number
1	298	3	200	40	8.10 × 10^−5	95.3	1.86 × 10^−2
2	298	3	500	40	8.10 × 10^−5	95.3	1.14 × 10^−2
3	298	3	1000	40	8.10 × 10^−5	95.3	0.77 × 10^−2
4	298	3	2000	40	8.10 × 10^−5	95.3	0.47 × 10^−2
5	298	3	1000	16	6.90 × 10^−5	50.2	0.50 × 10^−2
6	298	3	1000	24	7.20 × 10^−5	62.8	0.60 × 10^−2
7	298	3	1000	32	7.60 × 10^−5	87.4	0.72 × 10^−2
8	298	5	1000	40	8.10 × 10^−5	95.3	0.79 × 10^−2
9	298	7	1000	40	8.10 × 10^−5	95.3	0.76 × 10^−2
10	298	9	1000	40	8.10 × 10^−5	95.3	0.78 × 10^−2
11	298	11	1000	40	8.10 × 10^−5	95.3	0.78 × 10^−2

---

3.3.2 Kinetic rate constant in initial stage. In general, the ozone process presents two steps in series, involving mass transfer of ozone through the film layer and chemical reaction with phenoxyacetic acid in the bulk liquid. Considering the influence of mass transfer, the rate constant of the direct reaction between ozone and phenoxyacetic acid can also be written:

r_PAA = −zk_i[PAA]_b[O₃]_b (16)

where the subindex b refers to bulk phase. In fact, the meaning of the kinetic constant in the initial stage (k_i) is the same as the rate constant (k_O3). Eqn (16) can be used if concentrations of phenoxyacetic acid and ozone are known with time. Ozone not only reacted with PAA, but also reacted with intermediates formed during the ozone–PAA direct reaction simultaneously.¹⁷ To determine the kinetic constant requires more parameters to be taken into account, which is rather difficult. As a consequence, a method of studying the characteristics of the initial kinetic to reflect reaction rate was chosen to determine k_i, which was also applied in a kinetic study of the ozonation of fluorine.²⁸

The presence of dissolved ozone in water is a symbol of the slow kinetic regime. For the slow kinetic regime, the reaction factor E (eqn (17)) is 1, which is perceived as the ratio between the actual chemical absorption rate and that of physical absorption in the gas–liquid reaction.¹⁸

where N_O3 is the physical absorption rate of ozone, [O₃]* represents the concentration of ozone in the gas–liquid system at equilibrium, and [O₃]_b is the concentration of O₃ in the bulk of the liquid.

This method has a disadvantage in that the value of the ozone concentration is vague. Actually, the concentration of dissolved ozone is theoretically zero at the beginning of reaction. Nonetheless, application of the mass balance of ozone in water is handled well with this problem. Eqn (18), in the case of a semi-batch reactor, is as follows:¹⁸

where the subindex j at the right of the equation represents any compound j (PAA in the experiment) present in water that reacts with ozone directly, and k_j, k_d, and k_T are the rate constants of the ozone direct, hydroxyl-ion-initiated decomposition, and other initiation reactions, respectively. The left of this equation represents the transport rate of transference from the gas phase to the liquid phase. At the beginning of ozonation, the ozone accumulation rate term d[O₃]_b/dt and intermediates reactions are regarded as negligible, so eqn (18) reduces to

From the change of eqn (19), the concentration of ozone in the initial reaction is

Combining eqn (17) and (20) allows us to determine kinetic constant k_i in the initial stage:

Table 2 shows that some parameters, such as individual mass transfer coefficients and the Hatta number, are useful to determine kinetic constant k_i in the initial stage. Comparison of k_i with different parameters, such as pH value, the initial concentration of PAA, and the ozone dosage is shown in Fig. 3. Fig. 3a shows the effect of different initial concentrations of PAA ([PAA]₀) on k_i. The greater the initial concentration of PAA, the less was k_i, which was the same as the variation of k_app with increasing initial concentration of PAA. Fig. 3b shows the effect of the ozone dosage (Q_O3) on k_i. When the ozone dosage increased from 16 to 40 mg L⁻¹, k_i could be regarded as independent of the value of the stable ozone concentration in agreement with Leitner et al.²⁹ It is clear that the effect of ozone dosage on k_app is larger than on k_i. Fig. 3c shows the effect of pH on k_i. Increasing the pH value from 3.0 to 11.0, variation tendency of k_i was not evident which is almost identical to that of k_app. This explained directly that the main sites of reaction result from molecular ozone only, and that ozone decomposition was inhibited well by TBA. Results shown in Tables 1 and 2 confirmed that the relationship between k_app and k_i. k_app changed with variation of k_i and the concentration of ozone.

Fig. 3 Initial kinetic constant k_i as a function of different parameters. (a) Tendency chart of k_i with initial PAA concentration; (b) tendency chart of k_i with different O₃ dosages; (c) tendency chart of k_i at different pH.

4. Conclusions

Kinetic models of PAA degradation by ozonation were discussed and two rate constants k_app and k_i were carefully explored. The apparent rate constant k_app was supposed to be related to the pH value, the initial concentration of PAA, the O₃ dosage, and the temperature. A kinetics model was proposed as an exponential function of [PAA]_t = [PAA]₀exp(−0.51Q_O3^1.3273[PAA]^−1.1816₀t).

The calculated results showed that all Ha constants are lower than 0.3 in PAA degradation by ozonation, so the reaction was proved to be in the slow kinetic regime of gas–liquid reaction. In kinetic study of the initial stage, establishment of the mass balance equation of ozone and PAA in water determined the kinetic constant k_i.

Comparing k_i and k_app under different conditions, the results implied that there were effects of PAA initial concentration on k_app and k_i, and that these followed the same trend. The ozone dosage did not make any difference to k_i, but did to k_app. Increasing the pH value from 3.0 to 11.0, variation tendency of k_i was not evident which is almost identical to that of k_app and proved ozone decomposition was inhibited well by TBA. The mathematical models are able to reproduce the experimental observations concerning the reaction rate of direct ozonation. This paper provides basic data and kinetic reference for ozonation of pharmaceutical wastewater.

Acknowledgements

The authors are grateful for the financial support provided by the National Natural Science Foundation of China (21306175) and the Project of Science and Technology Office of Zhejiang Province (2008C13014-6).

Notes and references

1. B. I. Escher, R. Baumgartner, M. Koller, K. Treyer, J. Lienert and C. S. McArdell, Water Res., 2011, 45, 75–92.
2. A. Joss, H. Siegrist and T. A. Ternes, Water Sci. Technol., 2008, 57, 251–255.
3. I. M. Sebastine and R. J. Wakeman, Process Saf. Environ. Prot., 2003, 81, 229–235.
4. K. Kümmerer, Chemosphere, 2001, 45, 957–969.
5. J. L. Zhou, Z. L. Zhang, E. Banks, D. Grover and J. Q. Jiang, J. Hazard. Mater., 2009, 166, 655–661.
6. Q. Z. Dai, J. Y. Wang, J. Chen and J. M. Chen, Sep. Purif. Technol., 2014, 127, 112–120.
7. H. K. Singh, M. Saquib, M. M. Haque, M. Muneer and D. W. Bahnemann, J. Mol. Catal. A: Chem., 2007, 264, 66–72.
8. S. P. Kamble, S. B. Sawant and V. G. Pangarkar, Chem. Eng. Res. Des., 2006, 84, 355–362.
9. F. J. Beltrán, A. Aguinaco, J. F. García-Araya and A. Oropesa, Water Res., 2008, 42, 3799–3808.
10. X. Liu, T. Garoma, Z. Chen, L. Wang and Y. Wu, Chemosphere, 2012, 87, 1134–1140.
11. F. J. Beltrán, A. Aguinaco and J. F. García-Araya, Water Res., 2009, 43, 1359–1369.
12. J. Rivas, O. Gimeno, A. Encinas and F. Beltrán, Chemosphere, 2009, 76, 651–656.
13. V. Farines, S. Baig, J. Albet, J. Molinier and C. Legay, Chem. Eng. J., 2003, 91, 67–73.
14. L. Zhong and C. H. Kuo, Chin. J. Chem. Eng., 2000, 8, 272–275.
15. Z. Q. Zeng, H. K. Zou, X. Li, M. Arowo, B. C. Sun, J. F. Chen, G. W. Chu and L. Shao, Chem. Eng. J., 2013, 229, 404–411.
16. C. N. Satterfield, Mass transfer in heterogeneous catalysis, Massachusetts Institute of Technology Press, Cambridge, 1970.
17. Q. Z. Dai, L. L. Chen, W. Chen and J. M. Chen, Sep. Purif. Technol., 2015, 142, 287–292.
18. F. J. Beltrán, Ozone reaction kinetics for water and wastewater systems, Taylor & Francis e-library, 2005.
19. J. Staehelin and J. Hoigné, Environ. Sci. Technol., 1985, 19, 1206–1213.
20. Z. Q. He, S. Song, H. M. Zhou, H. P. Ying and J. M. Chen, J. Chem. Ind. Eng., 2006, 57, 2965–2969.
21. Q. Z. Dai, J. Y. Wang, L. L. Chen and J. M. Chen, Ind. Eng. Chem. Res., 2014, 53, 11593–11600.
22. J. C. Charpentier, Adv. Chem. Eng., 1981, 11, 1–133.
23. C. R. Wilke and P. Chang, AIChE J., 1955, 1, 264–270.
24. R. G. Utter, J. B. Burkholder, C. J. Howard and A. R. Ravishankara, J. Phys. Chem., 1992, 96, 4973–4979.
25. P. N. Johnson and R. A. Davis, J. Chem. Eng. Data, 1996, 41, 1485–1487.
26. D. Roberts and P. V. Danckwerts, Chem. Eng. Sci., 1962, 17, 961–969.
27. W. R. Zhao, Z. B. Wu and D. H. Wang, J. Hazard. Mater., 2006, 137, 1859–1865.
28. F. J. Beltrán, G. Ovejero, J. M. Encinar and J. Rivas, Ind. Eng. Chem. Res., 1995, 34, 1596–1606.
29. N. K. V. Leitner and B. Roshani, Water Res., 2010, 44, 2058–2066.

---