Environ. Sci. Technol. 2010, 44, 6233–6241
Modeling Hydroxyl Radical Distribution and Trialkyl Phosphates Oxidation in UV-H2O2 Photoreactors Using Computational Fluid Dynamics D O M E N I C O S A N T O R O , * ,† MEHRDAD RAISEE,‡ MOSTAFA MOGHADDAMI,‡ JOEL DUCOSTE,§ MICHEAL SASGES,† LORENZO LIBERTI,| AND MICHELE NOTARNICOLA| Department of Chemical and Biochemical Engineering, University of Western Ontario, London, Ontario N6A 5B9, Canada, Department of Mechanical Engineering, University of Tehran, P.O. Box 11365/4563, Tehran, Iran, Department of Civil, Construction and Environmental Engineering, North Carolina State University, Mann Hall 319c, Box 7908, Raleigh, North Carolina 27695, and Technical University of Bari, Via A. De Gasperi, 6, 74100 Taranto, Italy
Received March 1, 2010. Revised manuscript received July 11, 2010. Accepted July 13, 2010.
Advanced Oxidation Processes (AOPs) promoted by ultraviolet light are innovative and potentially cost-effective solutions for treating persistent pollutants recalcitrant to conventional water and wastewater treatment. While several studies have beenperformedduringthepastdecadetoimprovethefundamental understanding of the UV-H2O2 AOP and its kinetic modeling, Computational Fluid Dynamics (CFD) has only recently emerged as a powerful tool that allows a deeper understanding of complex photochemical processes in environmental and reactor engineering applications. In this paper, a comprehensive kinetic model of UV-H2O2 AOP was coupled with the Reynolds averaged Navier-Stokes (RANS) equations using CFD to predict the oxidation of tributyl phosphate (TBP) and tri(2chloroethtyl) phosphate (TCEP) in two different photoreactors: a parallel- and a cross-flow UV device employing a UV lamp emitting primarily 253.7 nm radiation. CFD simulations, obtainedforbothturbulentandlaminarflowregimesandcompared with experimental data over a wide range of UV doses, enabled the spatial visualization of hydrogen peroxide and hydroxyl radical distributions in the photoreactor. The annular photoreactor displayed consistently better oxidation performance than the cross-flow system due to the absence of recirculation zones, as confirmed by the hydroxyl radical dose distributions. Notably, such discrepancy was found to be strongly dependent on and directly correlated with the hydroxyl radical rate constant becoming relevant for conditions approaching diffusion-controlled reaction regimes (kC,OH > 109 M-1 s-1).
* Corresponding author phone: (519)457-3400; fax: (519)457-3030; e-mail:
[email protected]. † University of Western Ontario. ‡ University of Tehran. § North Carolina State University. | Technical University of Bari. 10.1021/es1000962
2010 American Chemical Society
Published on Web 07/27/2010
1. Introduction While reclamation and reuse of municipal wastewater is becoming an accepted practice especially in arid and semiarid regions (1), emerging micropollutants such as endocrine disruptors and pharmaceutically active compounds are increasingly being detected in limited water sources such as surface water and groundwater. These micropollutants, which are continuously released into the environment after little or no treatment, could be oxidized by Advanced Oxidation Processes (AOPs) such as UV-H2O2. Stefan et al. (2) conducted a study on the kinetics and reaction mechanism of acetone degradation by UV-H2O2 in dilute aqueous solution. The investigators proposed a kinetic model that successfully predicted reactant and intermediate profiles. Later, Crittenden et al. (3) advanced a UV-H2O2 dynamic kinetic model for a completely mixed batch reactor to predict pollutant removal demonstrating good agreement between observed and predicted data. Watts and Linden (4) reported on the UV-H2O2 oxidation of tri-n-butyl phosphate (TBP) and tri(2-chloroethyl) phosphate (TCEP) using a monochromatic (253.7 nm) collimated beam UV reactor. They observed that direct photolysis of both contaminants was not a relevant process. They also concluded that TCEP needed an order of magnitude higher UV dose to achieve the same level of removal as TBP. Lately, the prediction of the fluence rate distribution in a UV-H2O2 reactor has also received considerable attention. Bolton (5) presented a model, which accounted for both reflection and refraction as the UV beam goes through the air/quartz/water interfaces. Liu et al. (6) investigated the performance of various refracted and nonrefracted radiation models. They concluded that the effects of refraction must be taken into consideration for accurate simulation of fluence rate distribution, as confirmed in a subsequent paper by Sasges et al. (7) who proposed and validated a Lambertian emission model. The reactor hydraulics has also been studied in depth by several investigators. Sozzi and Taghipour (8) investigated turbulent flows through complex L-shape and U-shape UVreactors and examined the influence of three different turbulence models (Standard k-ε, Realizable k-ε, and Reynolds stress model) on the predicted results. Results indicated that Realizable k-ε produced the best overall prediction of the experimental PIV measurements. Similarly, Liu et al. (9) employed six turbulence models, including variants of k-ε, k-ω, Reynolds stress transport model (RSTM), and two-fluid model (TFM), to simulate the hydraulics in a polychromatic UV reactor. Results showed that both the fluence distributions and the predicted inactivation were sensitive to the turbulence model adopted. Shao (10) and Hofman et al. (11) performed numerical computations for a 4-lamp closed vessel UV reactor using three different cross-flow configurations (staggered, squared, and trapezoidal) to investigate the role of reactor hydraulics on UV dose distribution. Among the three cases examined, the staggered lamp arrangement (e.g., UV lamps positioned with a relative vertical offset in the flow direction) predicted a UV dose distribution with the highest mean UV dose and the lowest variance. Recently, Alpert et al. (12) investigated the use of CFD to evaluate the performance of a comprehensive UV/AOP VOL. 44, NO. 16, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6233
degradation of an indicator organic contaminant (methylene blue). They found that the CFD model underpredicted the experimental percent removal of methylene blue and that its degradation was sensitive to the background dissolved organic carbon concentration. A similar study was conducted by Elyasi and Taghipour (13), who demonstrated good agreement between numerical results and experimental data for velocity vectors, fluence rate, and concentration of rhodamine WT. Although previous studies have attempted to relate AOP performance to UV dose distribution even for pollutants with low quantum yield of photolysis, the extent of degradation occurring in a photoreactor may be better described by the [OH•] dose distribution. In this paper, the steady state hydroxyl radical distribution in two hypothetical UV-H2O2 photoreactors was investigated using a CFD approach. The yield of removal and EEO efficiency (electrical energy per unit flow per log removal) were predicted for both TCEP and TBP oxidation (pollutants that are increasingly detected in municipal water and wastewater (14) and artificially recharged groundwater (15)). Given the considerable contribution of hydrogen peroxide (50 mg/L) to fluid absorption coefficient, the species transport, chemical reactions, and UV fluence rate submodels have been fully coupled in our model using a spatially dependent fluid absorption coefficient (function of the local hydroxyl peroxide concentration). The proposed CFD model attempts to advance numerical AOP models by coupling the rigorous kinetic model of UV-H2O2 oxidation (3) with transport equations to provide the detailed spatial distributions of key parameters such as hydroxyl radical, hydrogen peroxide, fluence rate, and contaminants in the continuous flow photoreactors.
2. Materials and Methods
RH2O2 ) -2.303φH2O2I(εH2O2[H2O2] + εHO2-[HO2-]) k2[H2O2][OH•] -k3[HO2 ][OH•] - k4[H2O2][HO2•] - k5[H2O2][O2-•] -k8[H2O2][CO3-•] - k9[HO2-][CO3-•] + k10[OH•][OH•] +k12[HO2•][HO2•] + k13[HO2•][O2-•] (1)
9
k2[H2O2][OH•]
-k3[HO2 ][OH•] + k4[H2O2][HO2•] + k5[H2O2][O2 •]
-k6[OH•][CO23 ] - k7[OH•][HCO3 ] - 2k10[OH•][OH•]
-k11[OH•][HO2•] - k14[OH•][O2 •] - k15[OH•][CO3 •] -kTBP[OH•][TBP] - kTCEP[OH•][TCEP] - kTBEP[OH•][TBEP] (2)
RCO3-• ) k6[OH•][CO23 ] + k7[OH•][HCO3 ] - k8[H2O2][CO3 •] -k9[HO2 ][CO3 •] - k15[OH•][CO3 •] - k16[O2 •][CO3 •] -2k17[CO3 •][CO3 •]
(3) RHCO3- ) -k6[OH•][CO23 ] - k7[OH•][HCO3 ] +
k8[H2O2][CO3 •]
+k9[HO2 ][CO3 •]
+
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 16, 2010
(4)
k16[O2 •][CO3 •]
RO2-• ) k2[H2O2][OH•] + k3[HO2 ][OH•] - k4[H2O2][HO2•] -k5[H2O2][O2 •] + k8[H2O2][CO3 •] + k9[HO2 ][CO3 •]
-k11[OH•][HO2•] - 2k12[HO2•][HO2•] - k13[HO2•][O2 •] -k14[OH•][O2 •] - k16[O2 •][CO3 •]
(5) RTBP ) -kTBP[OH•][TBP] RTCEP ) -kTCEP[OH•][TCEP]
(6)
At near-neutral pH, the concentrations of species such as HO2•, HO2-, and CO32-are negligible if compared to those of their conjugate acid/base O2-• , H2O2, and HCO3-. For a buffered system, they can be estimated as follows (eqs 7-9) [CO23 ] )
[HO2] )
For brevity, details on the numerical methods and grid sensitivity analysis are included in the Supporting Information (SI). The Kinetic Model for UV-H2O2 Advanced Oxidation. Table 1 summarizes the photochemical and chemical reactions incorporated in the CFD model, along with their rate constants. The detailed UV-H2O2 kinetic model has been extensively validated in several previous studies and particularly in ref 3 for continuous flow stirred tank reactors and Li et al. (16), who implemented the UV-H2O2 reaction mechanism in the AdOx software. The kinetic model was incorporated into CFD using numerical subroutines to prescribe the net rate of generation, Ri, for each chemical species considered (eqs 1-6).
6234
ROH• ) 2 × 2.303φH2O2I(εH2O2[H2O2] + εHO2-[HO2 ]) -
10-10.3[HCO3] [H+] 10-11.6[H2O2]
[HO2•] )
[H+] [H+][O2 •] 10-4.8
(7)
(8)
(9)
Fluence Rate Model. The calculation of the fluence rate field was carried out by solving the radiative transport equation (RTE) using the discrete ordinate model (32). In the discrete ordinate (DO) model, the RTE is solved for the b,s b) spectral intensity Iλ(r r b), s b) s + aλIλ(b, r b) s )0 ∇.(Iλ(b,
(10)
where band r bare s position and direction vectors, respectively, λ is the wavelength, and aλ is the spectral, spatially dependent fluid absorption coefficient. In order to account for refraction, the air gap separating the lamp and the quartz sleeve regions has been included in the computational domain. Refractive indexes for air and water were assumed to be 1 and 1.37, respectively. UV-H2O2 Photoreactors. The oxidation of phosphate esters by UV-H2O2 was investigated in two different photoreactors, namely a single-lamp axi-symmetric annular and a single-lamp cross-flow UV system. Their schematic representations are shown in Figure 1, respectively, with dimensions given in centimeters. Each reactor consists of a 4-cm diameter, 100 cm long monochromatic UV lamp emitting at 253.7 nm. The treatment performances of the two photoreactors have been compared
TABLE 1. Kinetic Model of UV-H2O2 Advanced Oxidation of Phosphate Esters (16) reactions
rate constants, M-1 s-1
reference
1
H2O2 /HO2- + hυ f 2OH•
local rUV,H ) -2.303φH2O2I(εH2O2[H2O2] + εHO2-[HO2-]) 2O2
Baxendale and Wilson (17)
2
H2O2 + OH• f H2O + HO2•
k2 ) 2.7 × 107
Buxton et al. (18)
3
OH• + HO2- f HO2• + OH-
k3 ) 7.5 × 109
Christensen et al. (19)
4
H2O2 + HO2• f OH• + H2O + O2
k4 ) 3
Koppenol et al. (20)
5
H2O2 + O2-• f OH• + O2 + OH-
k5 ) 0.13
Weinstein and Bielski (21)
6
OH• + CO32- f CO3-• + OH-
k6 ) 3.9 × 108
Buxton et al. (18)
7
OH• + HCO3- f CO3-• + H2O
k7 ) 8.5 × 106
Buxton et al. (18)
8
H2O2 + CO3-• f HCO3- + HO2•
k8 ) 4.3 × 105
Draganic et al. (22)
9
HO2- + CO3-• f CO32- + HO2•
k9 ) 3.0 × 107
Draganic et al. (22)
10
OH• + OH• f H2O2
k10 ) 5.5 × 109
Buxton et al. (18)
11
OH• + HO2• f H2O + O2
k11 ) 6.6 × 109
Sehested et al. (23)
12
HO2• + HO2• f H2O2 + O2
k12 ) 8.3 × 105
Bielski et al. (24)
13
HO2• + O2-• f HO2- + O2
k13 ) 9.7 × 107
Bielski et al. (24)
14
OH• + O2-• f O2 + OH-
k14 ) 7.0 × 109
Beck (25)
15
OH• + CO3-• f ?
k15 ) 3.0 × 109
Holeman et al. (26)
16
CO3-• + O2-• f CO32- + O2
k16 ) 6.0 × 108
Eriksen et al. (27)
17
CO3-• + CO3-• f ?
k17 ) 3.0 × 107
Huie and Clifton (28)
18
OH• + TCEP f TCEPproducts
k18 ) kTCEP ) 5.6 × 108
Watts and Linden (29)
19
OH• + TBP f TBPproducts
k19 ) kTBP ) 6.4 × 109
Watts and Linden (29)
20
H2CO3 S H+ + HCO3-
pKa1 ) 6.3
Stumm and Morgan (30)
21
HCO3- S H+ + CO32-
pKa2 ) 10.3
Stumm and Morgan (30)
22
H2O2 S H+ + HO2-
pKa3 ) 11.6
Perry et al. (31)
23
HO2• S H+ + O2-•
pKa4 ) 4.8
Perry et al. (31)
no.
VOL. 44, NO. 16, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6235
FIGURE 1. Schematic representation of the annular UV (left) and cross-flow (right) photoreactors (dimensions are given in centimeters).
FIGURE 2. Comparison of numerical and experimental removal yield for TCEP and TBP. utilizing a constant average dose to gain insight into hydraulic efficiency. Therefore, CFD simulations were conducted using two different values for lamp power (484W and a 163W, for the annular and the cross-flow reactor, respectively) to match the different average retention time of the two systems (e.g.,
37.6 and 339 s for the annular and the cross-flow reactor, respectively). The UV lamps are surrounded by a 5.5 cm diameter quartz sleeve and the gap space between the lamp and quartz sleeve is occupied by air. For both reactors, the overall UVC efficiency of the lamp was assumed to be 31% (7). The quartz was not specifically modeled since the net effect of the quartz is insignificant; the angle of refraction from the lamp through the air to the water is not affected by the quartz. While the modeled reactors are hypothetical, they can be considered the constitutive elements of large-scale, multilamp commercial photoreactors. In both photoreactors, water composition whose characteristics are constant (25 °C, phosphate esters 5 mg/L each, hydrogen peroxide 50 mg/L, bicarbonates 24 mg/L, buffered pH ) 7) were simulated by introducing the premixed solution at the reactor inlet at a volumetric flow rate ranging from 1.05 m3/h to 8.4 m3/h. Such conditions are comparable with experiments described in ref 4 for UV doses lower than 1200 mJ/cm2 (at higher values, the buffering capacity of the irradiated solution was exceeded). The removal performance of the two continuous flow systems were compared against removal yield observed in a collimated beam reactor assuming that the photoreactor
FIGURE 3. Velocity vectors for a) annular UV reactor and b) cross-flow UV reactor. 6236
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 16, 2010
FIGURE 4. Fluence rate contours for a) annular UV reactor and b) cross-flow UV reactor. mean dose will approach the reduction equivalent dose for contaminants that are difficult to oxidize (33). Governing equations were solved using the finite volume method (34). For the annular reactor, where flow is turbulent (Re) 4800-19200), the high-Re k-ε model with wall-functions is used to describe turbulence and the random walk employed to account for the effects of instantaneous turbulent velocity fluctuations on the particle trajectories. Laminar simulations were carried out for the cross-flow reactor due to the low Re number (Re ) 290 to 1160).
3. Results and Discussion CFD Model Validation. As shown in Figure 2, the measured and predicted values for the degradation of target species (TCEP and TBP) were in good agreement. Both reactor designs confirm the required order of magnitude higher dose for TCEP compared to TBP shown in ref 4. The UV fluence presented in this figure was calculated as the mean of the dose distribution using a Lagrangian approach. Flow Distribution. Figure 3(a),(b) displays the velocity vectors for annular and cross-flow reactors, respectively. As expected, the velocity profile of the annular reactor (Figure 3a) is almost uniform resembling the well-known plug-flow reactor hydrodynamics. For the cross-flow reactor, however, the flow accelerates as it approaches the quartz sleeve (Figure 3b) forming a large downstream recirculation region at around θ ) 90°. Fluence Rate Distribution. The detailed fluence rate distributions for both reactors employing a 484 W UVC lamp, a volumetric flow rate of Q ) 2.1 m3/h, and a 50 mg/L H2O2 concentration is presented in Figure 4. The fluence rate is solved using a fluid with spatially dependent water transmittance (UVT254) which is a function of the H2O2 and HO2molar concentration and their molar absorption coefficients. Although such an approach was deemed necessary in consideration of the substantial contribution given by the H2O2 concentration (50 ppm) to the overall fluid absorption coefficient, the specification of a constant absorbance
FIGURE 5. Normalized UV dose distribution for the annular and cross-flow photoreactors. coefficient yielded similar results (see the SI). Notably, the fluence rate is still dominated by the emission characteristic of the UV source considering the relatively small changes of the fluid absorption coefficient undergoing hydrogen peroxide photolysis. For both reactors, the fluence rate contours in proximity of the quartz sleeve (y ) 0.0275 m, r ) 0.0275 m) reveals a slight discontinuity at the air-water interface due to refraction. To gain further understanding in reactor performance, the UV dose distributions for both photoreactors were computed using a Lagrangian particle tracking approach for a constant average UV dose (e.g., the product of the theoretical residence time and the average fluence rate) of approximately 600 mJ/cm2 (Figure 5). VOL. 44, NO. 16, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6237
FIGURE 6. Molar concentration of H2O2 for a) annular UV reactor and b) cross-flow UV reactor.
FIGURE 7. Molar concentration of TCEP for a) annular UV reactor and b) cross-flow UV reactor. Due to a better hydraulic efficiency (Figure 3a), the parallel flow photoreactor delivered a higher dose distribution and better oxidation performance. In the cross-flow reactor configuration, the change in the UV dose distribution shape with increasing flow may be due to its operation in the laminar regime and the competing processes of eddy vorticity driven mixing versus molecular diffusion. As a matter of fact, when the dose distributions at different Re numbers are normalized 6238
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 16, 2010
with their mean values (Figure 5), a self-similar UV dose distribution emerged that would only be a function of the reactor geometry and fluence rate distribution and no longer a function of the Re number. H2O2 Distribution. The predicted contours of hydrogen peroxide for the annular reactor are presented in Figure 6(a). Only minor changes in H2O2 concentration occur in the first part of the photoreactors (x < 2 m) where the fluence rate
FIGURE 8. Molar concentration of [OH•] through the (a) annular and (b) cross-flow reactors. is very low (Figure 4, a) while a considerable spatial variation in H2O2 photolysis takes place in the central portion of the system (2 m < x < 3 m). Cross-flow photoreactor results, illustrated in Figure 6(b), showed that hydrogen peroxide concentration starts to decrease as fluid approaches the irradiated zone around the cylindrical quartz sleeve. In this reactor configuration, significant gradients of peroxide concentration are observed in the radial direction. TCEP and TBP Distribution. Contaminant contours are presented in Figures 7 (TCEP) and S5 (TBP). In both cases, the steady state contaminant and hydrogen peroxide distributions showed considerable similarity (Figures 6, 7, and S5). Moreover, the yield of TBP removal was found to be considerably higher than that of TCEP due to its higher hydroxyl radical rate constants (kTBP ) 6.4 × 109 M-1 s-1, kTCEP ) 5.6 × 108 M-1 s-1). OH• Distribution. Contours of hydroxyl radical concentrations are presented in Figure 8. In both cases, the regions of high hydroxyl radical concentration resulted from both the complex generation and termination reactions and from the transport phenomena occurring in the UV-H2O2 photoreactors with significant concentration gradients displayed in the radial direction. The corresponding results for the cross-flow reactor, shown in Figure 8(b), confirmed higher hydroxyl radical levels in proximity of the sleeve (see Figure 4b). Contours of hydroxyl radicals are elongated in the direction of the flow due to convective transport occurring in the system. The [OH•] dose distribution was computed using a particle tracking approach to assess the contaminant exposure to hydroxyl radicals (Figure 9). The [OH•] dose distributions suggest that an enhancement in process performance will most likely result from an improved reactor hydraulics. For the reactor configurations examined, the log removal and electrical energy per order (EEO) are shown in Figure 10. By recalling that the two reactors, despite having the same average dose, displayed a considerably different mean dose (600 and 470 mJ/cm2), it can be concluded that the higher log removal occurring in the parallel flow photoreactor is the
FIGURE 9. [OH•] dose distribution for annular and cross-flow photoreactors. result of the absence of recirculation zones where the oxidation kinetics become diffusion-controlled. Such a behavior is clearly emphasized for contaminants with a hydroxyl radical contaminants kC,OH > 109 M-1 s-1. On the contrary, at the investigated conditions (e.g., high UV transmittance), the cross-flow configuration showed higher photon utilization efficiency (e.g., lower EEO) as a result of a longer mean UV path length. Unlike TCEP which at 50 mg/L initial hydrogen peroxide concentration requires an EEO ranging from 0.511 (crossflow reactor) kWh/m3 to 1.407 (parallel-flow reactor) log TCEP-1, the oxidation of TBP was achieved in a cost-effective manner using an order of magnitude lower energy consumption (EEO ) 0.113 and 0.063 kWh/m3 log TBP-1). Such opposite trends in efficiencies displayed by the two photoreactors suggest that a hybrid configuration with VOL. 44, NO. 16, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6239
Literature Cited
FIGURE 10. Impact of hydroxyl radical rate constant on EEO and reactor log removal. angled-to-flow UV lamps may indeed be optimal for AOP applications.
Acknowledgments This work has been supported (in part) by the Grant # 091 “Integrated Strategies for Productive Reuse of Municipal Wastewater”, funded by the Puglia Region (Italy). Special thanks are due Dr. Jussi Eloranta and Dr. Mihaela Stefan for the technical support offered throughout the investigation.
Appendix A absorption coefficient [m-1] fluence rate [W/m2] rate constant [M-1 s-1] or turbulent kinetic energy [m2 s-2] pressure [pa] equilibrium constant [-] turbulent production [m2 s-3] position vector [m] species source term [s-1] Reynolds number [-] direction vector [m] molecular Schmidt number [-] turbulent Schmidt number [-] Reynolds stress tensor components [m2 s-2] turbulent mass flux [m s-1] stream wise coordinate [m] coordinate system [m] cross-stream wise coordinate [m] or fluctuating mass fraction [-] mean mass fraction [-] mean velocity [m s-1]
a I k P pK Pk b r Ri Re b s Sc ScT ujiuj ujiyj x xi y Yi Uj
GREEK SYMBOLS ε λ
ν νt σk, σε F φ
molar absorption coefficient [M-1 cm-1] or dissipation rate of k [m2 s-3] wavelength [nm] kinematic viscosity [m2 s-1] turbulent kinematic viscosity [m2 s-1] turbulent Prandtl number for k and ε [-] density [kg m-3] quantum yield [mol ein-1]
Supporting Information Available Formulation of governing equations and details on the numerical methods. This material is available free of charge via the Internet at http://pubs.acs.org. 6240
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 16, 2010
(1) Asano, T.; Burton, F.; Leverenz, H.; Tsuchihashi, R.; Tchobanoglous, G. Water Reuse: Issues, Technologies, and Applications; Metcalf & Eddy/AECOM Press & McGraw Hill Professional: 2007. (2) Stefan, M. I.; Hoy, A. R.; Bolton, J. R. Kinetics and Mechanism of the Degradation and Mineralization of Acetone in Dilute Aqueous Solution Sensitized by the UV Photolysis of Hydrogen Peroxide. Environ. Sci. Technol. 1996, 30 (7), 2382–2390. (3) Crittenden, J. C.; Hu, S.; Hand, D. W.; Green, S. A. A kinetic model for H2O2/UV process in a completely mixed batch reactor. Water Res. 1999, 33, 2315–2328. (4) Watts, M. J.; Linden, K. G. Photooxidation and subsequent biodegradability of recalcitrant tri-alkyl phosphates TCEP and TBP in water. Water Res. 2008, 42 (20), 4949–4954. (5) Bolton, J. R. Calculation of ultraviolet fluence rate distributions in an annular reactor: significance of refraction and reflection. Water Res. 2000, 34 (13), 3315–3324. (6) Liu, D.; Ducoste, J.; Jin, S.; Linden, K. Evaluation of Alternative Fluence Rate Distribution Models. J. Water Supply: Res. Technol.--AQUA 2004, 53 (6), 391–408. (7) Sasges, M. R.; van der Pol, A.; Voronov, A.; Robinson, J., A Standard Method for Quantifying the Output of UV Lamps. In International Ozone Association Joint Congress; International UV Association: 2007. (8) Sozzi, D. A.; Taghipour, F. Computational and experimental study of annular photo-reactor hydrodynamics. Int. J. Heat Fluid Flow 2006, 27 (6), 1043–1053. (9) Liu, D.; Wu, C.; Linden, K.; Ducoste, J. Numerical simulation of UV disinfection reactors: Evaluation of alternative turbulence models. Appl. Math. Modell. 2007, 31 (9), 1753–1769. (10) Shao, L. Degradation of 4TBP by AOP, UV Reactor Modeling and Validation, Master Thesis, Delft University of Technology, 2007. (11) Hofman, J.; Shao, L.; Wols, B.; Uijttewaal, W.; Pelaar, G. I. J.; Beerendonk, E.; van Dijk, H., Design of UV reactors by, C. F. D.: Model Development and Experimental Validation. In Proceedings of the COMSOL Users Conference, Grenoble, 2007. (12) Alpert, S.; Knappe, D.; Ducoste, J. J. Modeling of UV/Hydrogen Peroxide Advanced Oxidation Processes using Computational Fluid Dynamics. Water Res. 2010, 44 (6), 1797–1808. (13) Elyasi, S.; Taghipour, F. Simulation of UV Photoreactor for Degradation of Chemical Contaminants: Model Development and Evaluation. Environ. Sci. Technol. 2010, 44 (6), 2056–2063. (14) Reemtsma, T.; Weiss, S.; Mueller, J.; Petrovic, M.; Gonzalez, S.; Barcelo, D.; Ventura, F.; Knepper, T. P. Polar Pollutants Entry into the Water Cycle by Municipal Wastewater: A European Perspective. Environ. Sci. Technol. 2006, 40 (17), 5451–5458. (15) Kolpin, D. W.; Furlong, E. T.; Meyer, M. T.; Thurman, E. M.; Zaugg, S. D.; Barber, L. B.; Buxton, H. T. Pharmaceuticals, Hormones, and Other Organic Wastewater Contaminants in U.S. Streams, 1999-2000: A National Reconnaissance. Environ. Sci. Technol. 2002, 36 (6), 1202–1211. (16) Li, K.; Crittenden, J. C.; Hand, D. W.; Hokanson, D. R. Advanced Oxidation Process Simulation Software (AdOx) Manual, Version 1.0; Michigan Technological University: 2002. (17) Baxendale, J. H.; Wilson, J. A. The photolysis of hydrogen peroxide at high light intensities. Trans. Faraday Soc. 1957, 53, 344–356. (18) Buxton, G. V.; Greenstock, C. L.; Helman, W. P.; Ross, A. B. Critical review of rate constants for reactions of hydrated electrons, hydrogen atoms and hydroxyl radicals in aqueous solution. J. Phys. Chem. Ref. Data 1988, 17 (2), 513–886. (19) Christensen, H. S.; Sehensted, H.; Corfitzan, H. Reactions of hydroxyl radicals with hydrogen peroxide at ambient and elevated temperatures. J. Phys. Chem. 1982, 86, 55–68. (20) Koppenol, W. H.; Butler, J.; Van Leeuwen, J. W. The HaberWeiss cycle. Photochem. Photobiol. 1978, 28 (4-5), 655–660. (21) Weinstein, J.; Bielski, B. H. J. Kinetics of the interaction of HO2 and O2- radicals with hydrogen peroxide. The Haber-Weiss reaction. J. Am. Chem. Soc. 1979, 101 (1), 58–62. (22) Draganic, Z. D.; Negron-Mendoza, A.; Sehested, K.; Vujosevic, S. I.; Navarro-Gonzales, R.; Albarran-Sanchez, M. G.; Draganic, I. G. Radiolysis of aqueous solutions of ammonium bicarbonate over a large dose range. Radiat. Phys. Chem. 1991, 383, 21–317. (23) Sehested, K.; Rasmussen, O. L.; Fricke, H. Rate constants of OH with HO2, O2 - and H2O2+ from hydrogen peroxide formation in pulse-irradiated oxygenated water. J. Phys. Chem. 1968, 72 (2), 626–631. (24) Bielski, B. H. J.; Cabelli, D. E.; Arudi, R. L.; Ross, A. B. Reactivity of HO2/O2 - radicals in aqueous solution. J. Phys. Chem. Ref. Data 1985, 14 (4), 1041–1100.
(25) Beck, G. Elektrishe leitfaehigkeits messungen zum nachweis geladener zwischenprodukte der pulsaradiolyse. Int. J. Radiat. Phys. Chem. 1969, 1 (3), 361–371. (26) Holeman, J.; Bjergbakke, E.; Sehested, K. The importance of radical-radical reactions in pulse radiolysis of aqueous carbonate/bicarbonate. Proc. Tihany Symp. Radiat. Chem. 1987, 61, 53–149. (27) Eriksen, T. E.; Lind, J.; Merenyi, G. On the acid-base equilibrium of the carbonate radical. Radiat. Phys. Chem. 1985, 26 (2), 197–199. (28) Huie, R. E.; Clifton, C. L. Temperature dependence of the rate constants for reactions of the sulfate radical, SO4-, with anions. J. Phys. Chem. 1990, 94 (23), 8561–8567. (29) Watts, M. J.; Linden, K. G. Advanced Oxidation Kinetics of Aqueous Trialkyl Phosphate Flame Retardants and Plasticizers. Environ. Sci. Technol. 2009, 43 (8), 2937–2942.
(30) Stumm, W.; Morgan, J. J. Aquatic Chemistry; Prentice-Hall: Engelwood, NJ, 1996. (31) Perry, R. H.; Green, D. W.; Maloney, J. D. Perry’s Chemical Engineers’ Handbook, 5th ed.; McGraw-Hill: New York, 1981. (32) Modest, M. F. Radiative Heat Transfer; Series in Mechanical Engineering, McGraw-Hill: 1993. (33) Wright, H. B.; Lawryshyn, Y. A. An Assessment of the Bioassay Concept for UV Reactor Validation. In Proceedings of the Water Environment Federation; 2000; pp 378-400. (34) Patankar, S. V. Numerical Heat Transfer and Fluid Flow; Hemisphere Publishing: McGraw Hill: 1980.
ES1000962
VOL. 44, NO. 16, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6241
