Environ. Sci. Technol. 2007, 41, 2548-2553
The ROH,UV Concept to Characterize and the Model UV/H2O2 Process in Natural Waters ERIK J. ROSENFELDT AND KARL G. LINDEN* Department of Civil and Environmental Engineering, Duke University, Box 90287, Durham, North Carolina 27708-0287
A new concept is introduced to characterize and model the UV/H2O2 advanced oxidation process (AOP) in water. Similar to the RCt concept used to describe OH radical exposure per ozone dose, the ROH,UV concept is defined as the experimentally determined •OH radical exposure per UV fluence. ROH,UV was determined by examining the destruction of a probe compound, para-chlorobenzoic acid in four different waters: DI water and three natural waters. ROH,UV was found to be affected greatly by water quality, specifically background •OH radical scavenging, which competed for the formed •OH radical with the probe compound, and background UV absorbance, which screened UV irradiation from the hydrogen peroxide. The ROH,UV values determined in the experiments using low-pressure Hg lamp were greater than those for the medium-pressure Hg lamp . Finally, the ROH,UV concept was utilized to calculate an overall scavenging factor for each water matrix, and this was successfully utilized in conjunction with the steadystate •OH radical model to improve the prediction of the oxidation of endocrine-disrupting compounds 17-R-ethinyl estradiol and 17-β-estradiol in the natural waters.
Introduction Because of the effectiveness of ultraviolet (UV) radiation for disinfecting drinking water, many water utilities will be installing UV capability in the coming years. As interest in UV radiation technology for disinfection of microbial contaminants increases, water utilities that are also concerned with treating chemical contaminants of emerging concern have begun looking into using UV in combination with H2O2. Studies have shown that UV and UV/H2O2 processes are capable of oxidizing many organic contaminants, including taste and odor-causing compounds such as methylisoborneol (MIB) and geosmin (1) as well as methyl tert butyl ether (MTBE) (2) atrazine (3), and N-nitrosodimethylamine (NDMA) (4). Addition of UV energy in the presence of H2O2 is an advanced oxidation process (AOP), which means much of the micropollutant oxidation occurs via a highly reactive intermediate, the hydroxyl radical (·•OH). When UV radiation is absorbed by H2O2, the molecule splits apart into two •OH radicals. However, due to recombining effects, the quantum yield of •OH radical production in the bulk solution is unity. Once the hydroxyl radical is formed, it will rapidly undergo reactions with water matrix constituents, including carbonate species (HCO3-, CO3-2), natural organic matter (NOM), other organic compounds present, and H2O2. Given this nonselec* Corresponding Author phone: (919) 660-5196; fax: (919) 6605219; e-mail:
[email protected]. 2548
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 7, 2007
tive nature of •OH radicals, design and modeling of AOPs must account for water quality when determining the effectiveness of the process toward degrading a specific contaminant. Several approaches for modeling UV/H2O2 advanced oxidation processes (AOPs) exist, including the steady-state • OH radical method (5) as well as the commercially available AdOx method (6). While these approaches can approximate chemical destruction for the UV/H2O2 AOPs, they rely on theoretical assumptions for rate constants to determine the amount of •OH radical mediated oxidation. Problems arise because simplifications, such as using a universal rate constant for dissolved organic carbon as a surrogate for NOM, are utilized in these models. The ROH,UV concept, defined as the experimentally determined •OH radical exposure per UV fluence for a given water matrix and initial H2O2 concentration, can characterize the effectiveness of the UV/H2O2 AOP within a specific water matrix. The probe compound utilized to investigate the ROH,UV concept is para-chlorobenzoic acid (pCBA), and was chosen due to the widespread use of this compound to characterize other AOPs (7-9). In addition to providing insight about the effectiveness of advanced oxidation in a water matrix, the ROH,UV parameter can also be utilized to model the destruction of micropollutants of concern in the water. Theoretical Development of ROH,UV. ROH,UV is developed similar to the RCt concept (7) used in ozone studies. Equation 1 is the kinetic expression describing the rate of pCBA decay via the UV/H2O2 AOP, and accounts for both direct photolysis and •OH radical oxidation.
d[pCBA] ) -(k′d + k′i)[pCBA] dt
(1)
where k′d is the pseudo-first-order rate constant of pCBA destruction by direct UV photolysis (s-1) (valid for low micropollutant concentrations), k′i is the pseudo first-order rate constant of pCBA destruction via the •OH radical mechanism (s-1), described by eq 2.
k′i ) kOH,pCBA[•OH]
(2)
Equation 3 is the result of the integration of eq 1 incorporating eq 2, and if both sides are divided by time, eq 4 results, incorporating the total observed, direct, and •OH radical oxidation rate constants into one equation.
{
ln
}
[pCBA]o [pCBA]t
) k′dt + kOH,pCBA
∫ [ OH]dt t •
0
(3)
and,
∫ [ OH]dt t •
k′T ) k′d + kOH,pCBA
0
t
(4)
In evaluating UV systems, it is important to convert timebased rate constants into fluence-based rate constants (10). In this case, dividing both sides of eq 4 by the average UV fluence rate Eo (mW cm-2), converts the rate constants, and results in eq 5.
k′ D T -
k′ D d
)
kOH,pCBA
∫ [ OH]dt t •
0
Eo‚t
(5)
The superscript “D” denotes the rate constant is a “fluencebased” rate constant, with units of cm2 mJ-1. UV fluence H 10.1021/es062353p CCC: $37.00
2007 American Chemical Society Published on Web 03/01/2007
(mJ cm-2) is simply the product of average fluence rate (E0) and time, so upon rearrangement of eq 5, ROH,UV (M s cm2 mJ-1), can be determined (eq 6).
∫ [ OH]dt t •
ROH,UV )
0
H
)
D k′ D T - k′ d kOH,pCBA
(6)
As a note, the fluence rate is assumed constant as the actual absorbance of the water is not significantly changed over the course of the UV exposure. Also, these equations were developed for monochromatic UV emission sources, and several of the measured parameters are wavelength dependent. In expanding these equations to a polychromatic lamp application (e.g., medium pressure UV sources), E0, k′dD, the fraction of light absorbed by H2O2 (impacting ∫0t [•OH]dt), and the effect of variations in wavelength emission spectra between lamps, must be considered. This work presently focuses on the merits of ROH,UV as a tool for measuring •OH radical production by a well characterized MP lamp. Future work will pursue the question of how to expand the concept to include fluctuations between lamp emission spectra.
Experimental Section Chemicals and Waters. Para-chlorobenzoic acid (pCBA, Aldrich Chemical Co., Milwaukee, WI) was spiked into a water matrix at 5 µM as the probe compound. Model validation was performed using 17-R-ethinyl estradiol (EE2, Steraloids, Inc., Newport, RI) and 17-β-estradiol (E2, Steraloids, Inc., Newport, RI), which display second-order rate constants with OH radical of kE2,OH ) 1.08 × 1010 M-1 s-1 and kEE2,OH ) 1.41 × 1010 M-1 s-1 (11). Several waters were utilized for this experiment. Deionized (DI), purified water was taken from a laboratory water purifying system (Hydro Services and Supplies, Inc., Research Triangle Park, NC). Three natural water samples were taken from water treatment plant inlets, before any treatment. Samples were then filtered through a 0.45 µm filter to remove particles, and water quality parameters including DOC, alkalinity, pH, UV absorbance, and turbidity were measured to characterize the waters. Analytical Equipment and Methods. A Varian Pro Star high performance liquid chromatography (HPLC) workstation (Walnut Creek, CA), with a model 330 polychromatic diode array detector was used to detect pCBA and EDCs in water. A C-18 reverse phase column served as the stationary phase, with 1 mL min-1 of eluent consisting of an acetonitrile (ACN) to water ratio of 60:40% for EE2, 50:50% for E2, and 50% ACN, 50% H2O adjusted to pH 2 with phosphoric acid for pCBA. These HPLC methods were adapted from the literature for EE2, E2, and pCBA (7, 11, 12). Water quality parameters were measured using Standard Methods. Dissolved organic carbon (DOC) was measured in accordance with Standard Method 5310 A (combustion and detection method) (Tekmar Dohrmann Apollo 9000 total carbon analyzer). Alkalinity was measured according to Standard Method 2320 B. Residual H2O2 was measured using the I3method (13). pH was measured using an electronic pH meter (Cole Parmer pH 100 series), calibrated daily using pH 4, 7, and 10 buffers. UV absorbance (200-300 nm) of the spiked test water was measured in a spectrophotometer (Cary 100 bio-spectrophotometer, Varian, Houston, TX). UV Apparatus. Irradiation by low pressure (LP) UV was performed using a bench-scale collimated beam UV apparatus consisting of four 15 W germicidal lamps (ozonefree, General Electric no. G15T8). Irradiation by a 1 kW medium pressure (MP) UV lamp (Hanovia Co., Union NJ) was performed using a bench-scale collimated beam UV reactor provided by Calgon Carbon Corporation (CCC, Pittsburgh, PA) modified by Duke University. A general
schematic of a collimated beam UV reactor can be found in ref 11. A UV radiometer and detector (International Light Inc., model 1700/SED 240/W) calibrated at 2 nm intervals in the range of 200-400 nm was used to measure UV irradiance at the surface of the test water. Fluence Calculation. UV fluence (mJ cm-2) was calculated as the average irradiance multiplied by the exposure time. The average UV fluence rate (E0) in the completely mixed sample was determined from the incident irradiance, UV absorbance, and sample depth using an integrated form of the Beer-Lambert law (14). Incident UV irradiance (mW cm-2) was measured with a radiometer and detector at the water sample surface. Samples were exposed to UV light for specific time intervals using a shutter system above the reaction vessel. The exposure time (seconds) necessary was determined by dividing the desired UV fluence by the average UV irradiance. For the LP UV source, the fluence was calculated as the radiation emitted at 253.7 nm. All UV fluence values reported for the MP UV source were calculated as a function of the unweighted lamp output between 200 and 300 nm. Irradiation and Sampling Procedure. Water samples were spiked with the desired initial compound concentration, and a sample was taken and analyzed by HPLC to determine the initial concentration. 50 or 100 mL samples were measured into 70 × 50 mm flat bottom Pyrex petri dishes, and then spiked with H2O2 as needed. A small stir bar was added and mixing began at a rate that allowed for good mixing, but disturbed the surface of the water minimally. Initial samples were taken for absorption measurement (0.7 mL), pCBA or EDC analysis (0.3 mL) and H2O2 analysis (0.1-0.2 mL), where applicable. The samples were then irradiated in the LP or MP collimated beam UV radiation setup for the proper times corresponding to the desired delivered UV fluence (from 0 to 1000 mJ cm-2), and sub-sampled (0.3 mL) for analysis at various fluence values.
Results and Discussion Quantum Yield of pCBA. To validate the assumption that the pCBA photolysis can be neglected in calculation of ROH,UV, the quantum yield of pCBA for LP and MP UV sources was determined. Average quantum yield values of 0.013 ( 0.002 mol Es-1 and 0.018 ( 0.004 mol Es-1 were determined for direct pCBA photolysis by LP and MP UV respectively. These are relatively low quantum yields when compared to other species, including atrazine, which displays a quantum yield of 0.05 mol Es-1 for LP UV (3) and 0.046 mol Es-1 for MP UV (15), 17-R-ethinyl estradiol (0.026 mol Es-1 and 0.061 mol Es-1 for LP and MP), and 17-β-estradiol (0.043 mol Es-1 and 0.1 mol Es-1 for LP and MP) (11). These quantum yield values and the concentrations utilized in these experiments (5 µM) imply that phototransformation rate constants did not exceed 1 × 10-4 cm2 mJ-1 for MP UV photolysis in DI water. In natural waters, and with the addition of hydrogen peroxide, this rate would slow as the fraction of light absorbed by pCBA is lower. These direct UV photolysis rate constants are known to be significantly slower than UV/H2O2 AOP rate constants for the conditions tested in this study, indicating that direct photolysis should not significantly account for observed pCBA degradation. Calculating ROH,UV in DI and Natural Waters. Figure 1 displays the natural log transformed relative pCBA concentration as a function of applied LP UV fluence for a range of added hydrogen peroxide concentrations. The observed decay of pCBA using LP UV is representative of the kinetics for all conditions (waters, lamp type). The decay is first order with UV fluence, with only one kinetic regime throughout. This is a significant difference between UV/H2O2 AOP and ozone based AOPs, which display distinct, two-phase kinetics with a very rapid initial degradation, VOL. 41, NO. 7, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2549
FIGURE 1. Oxidation of pCBA as a function of LP UV fluence with variable initial hydrogen peroxide concentration in natural water. followed by a slower, steady-state first-order degradation (7). In ozone AOPs, only the kinetics of the second regime are used to describe steady-state RCt, but this limitation is avoided in the UV processes. The ROH,UV values calculated using this kinetic data describe •OH radical exposure throughout the entire UV AOP. The best-fit slope of each data set yields k′D d in the case of 0 mM H2O2, or k′D in the cases where H O was added. For 2 2 T D each initial peroxide, k′D d was subtracted from k′T and then this was divided by kOH,pCBA)5 × 109 M-1 s-1, to calculate ROH,UV (eq 6), displayed in Figure 2. ROH,UV in DI and Natural Waters: Water Comparison. Figure 2 displays the ROH,UV values in DI water, and the three natural waters, water 1, water 2, and water 3, utilizing LP UV (a) and MP UV (b) and 0.15, 0.29, 0.74, 1.47, and 2.94 mM H2O2, corresponding to doses of 5, 10, 25, 50, and 100 mg L-1 H2O2. As the initial peroxide concentration increases, the ROH,UV value increases, indicating that more •OH radical is available per UV dose when the initial peroxide concentration increases. This result seems intuitive, considering that increased peroxide levels will absorb more of the available UV energy, resulting in more •OH radical produced. However, the increase in peroxide concentration does not correspond with a linear increase in ROH,UV because of the •OH radical scavenging ability of H2O2. In other words, as peroxide concentration is increased to a point, the peroxide becomes a significant scavenging species contributing to the overall scavenging of the system, and therefore reduces the efficiency of •OH radical reacting with the target pollutant once it is formed. Comparison between the waters yields some interesting results. Table 1 displays the water quality parameters of each water. Utilizing these parameters, some of the trends observed in Figure 2 can be explained. The water quality parameters indicate that water 1 is a high DOC, low alkalinity water matrix, with the lowest nitrate level, while water 2 is a low DOC, high alkalinity water matrix, and has the highest nitrate level. Water 3 has the lowest DOC and alkalinity levels, with a nitrate level slightly higher than water 1. Assuming the following rate constants for the reaction between •OH radical and the various water quality parameters: DOC (2.5 × 104 L mg-1 s-1) (16), HCO3- (8.5 × 106 M-1 s-1), CO3-2 (3.9 × 108 M-1 s-1), and H2O2 (2.7 × 107 M-1 s-1) (17), the overall scavenging factors (Table 2) in each water matrix at each peroxide concentration can be calculated. These calculated values are utilized for qualitative comparison purposes only, and are presented acknowledging that discrepancies within the reported data exists regarding reaction rate constants, that errors in measurement of water quality parameters can significantly alter the parameter, and that this discussion is only valid for the initial conditions. Even with these qualifications, the •OH radical scavenging factors will 2550
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 7, 2007
be a useful qualitative tool to illustrate differences between waters and discuss underlying reasons for these differences. Assuming scavenging-dominated kinetics, we would expect water 3 to display the greatest ROH,UV at each peroxide level, followed by water 2, with water 1 slowest. Generally, this prediction agrees with the observed results shown in Figure 2, indicating that scavenging can predict the qualitative comparative performance of the UV/H2O2 process in natural waters. However, due to the acknowledged uncertainties involved in the theoretical calculation of the scavenging factors, utilizing this tool as a quantitative prediction parameter is not convincing. ROH,UV in DI and Natural Waters: Lamp Comparison. Another important parameter affecting ROH,UV is the ability of hydrogen peroxide to absorb light. Figure 3 displays the background absorption spectra for the three natural waters, as well as the absorption of 0.29 mM H2O2 dissolved in DI water. Hydrogen peroxide absorbs weakly throughout the UV range, especially near the main output regions for each lamp. Because of this, the background water absorbance is very important in determining the amount of UV energy absorbed by peroxide, and eventually, the amount of •OH radical formed by the UV/H2O2 AOP. Comparing the data between Figure 2a and b for each water indicates that LP UV has a greater ROH,UV in each natural water, but in DI water, the ROH,UV values for MP UV and LP UV are not statistically different at any peroxide concentration. The observed differences can be attributed to the amount of light absorbed by the hydrogen peroxide as compared to that absorbed by the entire water sample. Equation 7 is the equation describing the production of •OH radical by UV photolysis of hydrogen peroxide from ref 6.
rUV/H2O2(λ) ) -ΦOH(λ)Ep(λ)fH2O2(λ)(1 - e-A(λ))
(7)
Where ΦOH(λ) is the quantum yield describing the production of •OH by H2O2 photolysis (mol Es-1) (assumed to be 1.0 throughout the UV range), Ep(λ) is the incident photon irradiance (Es cm-2 s-1), A ) 2.303‚a(λ)‚b where a(λ) is the wavelength specific absorption coefficient and b is the solution path length, and fH2O2 (λ) is the fraction of UV energy absorbed by hydrogen peroxide at each wavelength. The rate of •OH radical formation is directly proportional to the fraction of UV energy absorbed by the hydrogen peroxide. It is important to note that the parameters describing •OH radical formation are all wavelength dependent. To account for the wavelength dependence of these properties in the polychromatic MP UV system, fH2O2 was calculated using eq 8.
fH2O2 )
∑
200-300
([
])
{Ep(λ)(1 - e-A(λ)) fH2O2(λ)}
∑
200- 300
Ep(λ)(1 - e
-A(λ)
)
(8)
Table 2 displays fH2O2 in all natural waters for the LP and MP UV systems. In every natural water at all peroxide concentrations, fH2O2 is greater for the LP UV system compared to the MP UV system, which makes sense considering the MP wavelengths include those >250 nm not well absorbed by the peroxide. In water 1, this difference is 15% at low initial peroxide concentrations, up to 23% at 2.9 mM H2O2. In waters 2 and 3, the differences in fH2O2 are approximately 25% for all peroxide levels. For comparison, in DI water, there was less than 5% difference in fH2O2 for either lamp setup. These differences correspond to the observed variation in ROH,UV shown in Figure 2. Modeling with ROH,UV. Since ROH,UV is calculated for a water matrix impacted with the probe compound pCBA, this
FIGURE 2. ROH,UV for DI water and three natural waters utilizing LP (a) and MP (b) UV sources. Values were determined for 0.15, 0.29, 0.74, 1.47, and 2.91 mM initial H2O2. The steady-state •OH radical model assumes a constant concentration of •OH radicals, and that this concentration can be estimated with eq 10.
TABLE 1. Water Quality Parameters of Each Natural Water Matrix L-1)
TOC (mg alkalinity (mM) nitrate (mg L-1) pH a(254 nm) (cm-1)
water 1
water 2
water 3
16.74 0.36 1.13 7.11 0.61
3.06 2.63 2.97 8.00 0.04
1.39 0.67 1.30 7.98 0.11
U254(
TABLE 2. Scavenging Factors (∑kOH,S[S]) and Fraction of UV Energy Absorbed by H2O2 in Each Water Matrix as a Function of Initial Peroxide Concentration fH2O2 × 100 ΣkOH,S[S] × 10-5 (s-1) [H2O2]i (mM) water 1 water 2 water 3 0.15 0.29 0.74 1.5 2.9
4.3 4.3 4.4 4.6 5
1.1 1.1 1.2 1.4 1.8
0.45 0.49 0.61 0.81 1.2
water 1
water 2
water 3
LP
MP
LP
LP
0.42 0.83 2 4 7.7
0.36 0.71 1.7 3.3 6.1
MP
MP
6.1 4.8 2.2 1.7 12 9 4.3 3.3 25 19 10 7.7 40 30 18 14 57 44 31 23
parameter alone is not enough to accurately model the UV/ H2O2 process, such as it has been utilized for an ozone system (7). Rather, our approach was to utilize ROH,UV to calculate the overall background scavenging in the water (subtracting the contribution of pCBA and H2O2), and then couple this scavenging contribution as an input into the steady-state •OH radical model. Since the background •OH radical scavenging is dependent solely on the water quality, data from the LP UV/H2O2 systems were utilized to calculate background scavenging. Although not explored in this article, the method utilized below should be useful for measuring background hydroxyl radical scavenging in water subject to advanced oxidation processes. The assumption utilized in our approach is similar to that of the steady-state •OH radical model outlined in ref 5, specifically that •OH radical is a very transient species and its concentration can be assumed to be at steady state. Utilizing this assumption, and revisiting the units of ROH,UV, it can be shown that the parameter describes the steadystate •OH radical concentration for a specific average fluence rate (E0) (eq 9).
ROH,UV
( )( )
( )
[OH]ss M M‚s M‚s ) ) mJ mW‚s Eo mW cm2 cm2 cm2
[•OH]ss )
(9)
Eo‚H2O2[H2O2]ΦOH
∑k
S,OH[S]i +
kpCBA,OH[pCBA] + kH2O2,OH[H2O2]) (10)
where E0 is in units of (mW cm-2), U254 is the energy per 1 mol of photons (Einstein) at 254 nm (J Es-1), H2O2 is the molar absorption coefficient for H2O2, ΦOH has been previously defined, and ∑ kS,OH[S]i is the background scavenging factor. Rearranging eq 10 to estimate [•OH]ss (11) to solve for [•OH]ss/EO results in eq 11.
[•OH]ss
)
Eo H2O2[H2O2]ΦOH U254(
∑k
S,OH[S]i
+ kpCBA,OH [pCBA] + kH2O2,OH[H2O2]) (11)
Inverting both sides of eq 10 and simplifying the terms results in eq 12.
∑k
U254(
Eo
)
[•OH]ss
S,OH[S]i
+ kpCBA,OH[pCBA]) ‚
H2O2ΦOH
( ) 1
[H2O2]
+
U254‚kH2O,OH Η2Ο2ΦOH
(12)
Equation 12 lends itself to a plot of EO/[•OH]ss (also 1/ROH,UV) as a function of 1/[H2O2]. When the inverse of the ROH,UV data from waters 1, 2, and 3 in Figure 2a is plotted as a function of 1/[H2O2], a linear relationship with slope (m) and yintercept (b) is observed. Further analysis of eq 12 indicates that U254/H2O2ΦOH is a common term to both slope and y-intercept, so with combination and rearrangement of the terms, one can solve for the water matrix scavenging with eq 13.
∑k
S,OH[S]i
) kH2O2,OH ‚
m b
- kpCBA,OH[pCBA]
VOL. 41, NO. 7, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
(13) 9
2551
FIGURE 3. Background absorbance of three natural waters and 0.29 mM H2O2 (in DI water) between 200 and 300 nm.
FIGURE 4. Comparison of observed (gray bars with 95% confidence intervals) and steady-state OH radical model predicted oxidation rate constants utilizing scavenging factors calculated using measured water quality parameters (black bars) or as measured by the pCBA probe method (white bars).
TABLE 3. Comparison between Scavenging Factors (∑kOH,S[S]) Measured with the pCBA Probe Method and Calculated from Water Quality Parameters ∑kS,OH[S] × 10-4 (s-1) water no.
measured with pCBA probe
calculated using water quality parameters
1 2 3
3.59 4.83 0.32
42.4 12.1 3.48
Once this matrix scavenging term was determined for each natural water, it was input into the steady-state OH radical scavenging model to predict degradation of other contaminants at any initial peroxide concentration. Table 3 compares the background scavenging determined with this method to that calculated using measured water quality parameters. The traditional method of calculating scavenging utilizes TOC and alkalinity as surrogate parameters for NOM and carbonate/bicarbonate species, respectively. Rate constants for these species and the composition of the waters can be found in Table 1 and its accompanying text. The traditional method consistently predicts greater scavenging factors than does the new ROH,UV probe compound method. To determine if the ROH,UV method accurately predicts chemical degradation, Figure 4 is presented to compare modeled predictions for LP UV/H2O2 AOP with initial EE2 and E2 concentrations of approximately 5 µM, utilizing the two methods for calculating scavenging factors. In all cases, the traditional steady-state OH radical model under-predicted the oxidation rate constants, while utiliza2552
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 7, 2007
tion of the new ROH,UV method to calculate scavenging factors predicts rate constants within the 95% confidence intervals of the observed rate constants. The rate of •OH radical formation was always the same between the models for a specific run, so the discrepancy must lie within the calculation of the scavenging factors. Because the probe compound method directly measures scavenging in each water, it does not rely on generic rate constants and surrogate parameters (such as utilizing TOC as a NOM surrogate). Additionally, the contributions of species contributing to the formation of •OH radicals, such as nitrate and other species, which may be missed by utilizing the traditional model, are included with the probe compound method. Similar to the useful ozone based RCt concept, the ROH,UV provides a method to experimentally determine •OH radical exposure per UV fluence for UV-based advanced oxidation processes. However, there are several drawbacks to the tool. First, the tool is specific for a set of water conditions. As water quality changes, scavenging and absorption characteristics are affected, and the ROH,UV will also vary. Proponents of the RCt concept have recognized this and pursued studies characterizing water sources over time. Expanding the ROH,UV concept in this way would add to its usefulness as a water characterization tool. Future work must also examine the effect of wavelength specificity and path length effects on ROH,UV. Different polychromatic lamp sources display different characteristic emission spectra, and the emission of a single lamp may change over time. Additionally, in order to apply ROH,UV in bench or full-scale reactors, the effect of changing path length on ROH,UV must be examined. The most useful application of the ROH,UV parameter may be in calculating the scavenging properties of a water matrix.
By utilizing this method, one can directly calculate the overall scavenging factor for a water matrix, and from there, notice scavenging abnormalities or examine specific water quality effects. In general, by introducing the ROH,UV parameter, we have developed a way to experimentally assess the •OH radical exposure in UV/H2O2 AOPs. This newly developed parameter can be used to compare water-matrix effects on the UV/ H2O2 AOP, compare between lamp types, and model oxidation of environmental pollutants of concern during AOP treatment in different waters. This parameter can also be useful in comparing efficiency between several AOPs, including ozone and UV based AOPs.
(6) (7) (8) (9) (10)
Acknowledgments We acknowledge Dr. Urs von Gunten, and the W&T group at EAWAG, whose help led to work inspiring the idea for ROH,UV. We would also like to thank Dr. Charles M. Sharpless for helpful discussions and the ES&T reviewers for excellent suggestions. Funding and natural waters for this work were provided in part from AWWARF project no. 2897 “Impact of UV and UV Advanced Oxidation Processes on Toxicity of Endocrine Disrupting Compounds in Water”. Additional funding for the research was provided by the American Water Works Association Abel Wolman and National Water Research Institute doctoral fellowships awarded to Erik Rosenfeldt.
(11)
(12)
(13) (14)
Literature Cited (1) Rosenfeldt, E. J.; Melcher, B.; Linden, K. G. UV and UV/H2O2 treatment of methylisoborneol (MIB) and geosmin in water. AQUA. J. Water Supply Res. Technol. 2005, 54 (7), 423-434. (2) Cater, S. R.; Stefan, M. I.; Bolton, J. R.; Safarzadeh-Amiri, A. UV/H2O2 treatment of methyl tert-butyl ether in contaminated waters. Environ. Sci. Technol. 2000, 34 (4), 659-662. (3) Beltran, F. J.; Ovejero, G.; Acedo, B. Oxidation of atrazine in water by ultraviolet-radiation combined with hydrogenperoxide. Water Res. 1993, 27 (6), 1013-1021. (4) Sharpless, C. M.; Linden, K. G. Experimental and model comparisons of low- and medium-pressure Hg lamps for the direct and H2O2 assisted UV photodegradation of N-nitrosodimethylamine in simulated drinking water. Environ. Sci. Technol. 2003, 37 (9), 1933-1940. (5) Glaze, W. H.; Lay, Y.; Kang, J. W. Advanced Oxidation Processes - A kinetic model for the oxidation of 1,2-Dibromo-3-chloro-
(15) (16) (17)
(18)
propane in water by the combination of hydrogen peroxide and UV radiation. Ind. Eng. Chem. Res. 1995, 34 (7) 23142323. Crittenden, J. C.; Hu, S. M.; Hand, D. W.; Green, S. A. A kinetic model for H2O2/UV process in a completely mixed batch reactor. Water Res. 1999, 33 (10) 2315-2328. Elovitz, M.; von Gunten, U. Hydroxyl radical ozone ratios during ozonation processes. I-The R-ct concept. Ozone: Sci. Eng. 1999, 21, 239-260. Pines, D. S.; Reckhow, D. A. Solid phase catalytic ozonation process for the destruction of a model pollutant. Ozone: Sci. Eng. 2003, 25 (1), 25-39. Han, S. K.; Nam, S. N.; Kang, J. W. OH radical monitoring technologies for AOP advanced oxidation process. Water Sci. Technol. 2002, 46 (11-12), 7-12. Bolton, J. R.; Stefan, M. Fundamental photochemical approach to the concepts of fluence (UV dose) and electrical energy efficiency in photochemical degradation reactions. Res. Chem. Intermed. 2002, 28 (7-9) 857-870. Rosenfeldt, E. J.; Linden, K. G. Degradation of endocrine disrupting chemicals bisphenol A, ethinyl estradiol, and estradiol during UV photolysis and advanced oxidation processes. Environ. Sci. Technol. 2004, 38 (20) 5476-5483. Ohko, Y.; Iuchi, K. I.; Niwa, C.; Tatsuma, T.; Nakashima, T.; Iguchi, T.; Kubota, Y.; Fujishima, A. 17 beta-estradiol degradation by TiO2 photocatalysis as means of reducing estrogenic activity. Environ. Sci. Technol.2002, 369 (19), 4175-4181. Klassen, N. V.; Marchington, D.; McGowan, H. C. E. H2O2 determination by the I3 method and by KMnO4 titration. Anal. Chem. 1994, 66 (18), 2921-2925. Bolton, J. R.; Linden, K. G. Standardization of methods for fluence (UV dose) determination in bench-scale UV experiments. J. Environ. Eng. 2003, 129 (3), 209-215. Meunier L. Swiss Federal Institute of Environmental Science and Technology (EAWAG). 2004 Personal Communication. Larson, R.; Zepp, R. Reactivity of the Carbonate Radical with Aniline Derivatives. Environ. Toxicol. Chem. 1988, 7 (4), 265274. Buxton, G.; Greenstock, C.; Helman, W.; Ross, A. Critical Review of Rate Constants for Reactions of Hydrated Electrons, Hydrogen Atoms and Hydroxyl Radicals (OH/O-) in Aqueous Solution. J. Phys. Chem. Ref. Data 1988, 17 (2), 513-886. NDRL, Notre Dame Radiation Laboratory Chemistry Data Center, www.rcdc.nd.edu/Solnkin2/ (July 31, 2005).
Received for review October 2, 2006. Revised manuscript received January 10, 2007. Accepted January 18, 2007. ES062353P
VOL. 41, NO. 7, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2553