Environ. Sci. Technol. 2010, 44, 7827–7832
Kinetics of UV-H2O2 Advanced Oxidation in the Presence of Alcohols: The Role of Carbon Centered Radicals EVGENY POPOV,† MUHAMED MAMETKULIYEV,† DOMENICO SANTORO,‡ LORENZO LIBERTI,¶ AND J U S S I E L O R A N T A * ,† Department of Chemistry and Biochemistry, California State University at Northridge, 18111 Nordhoff Street, Northridge, California 91330, Department of Chemical and Biochemical Engineering, University of Western Ontario, London, Ontario, Canada N6A 5B9, and Department of Environmental Engineering & Sustainable Development, Technical University of Bari, v. De Gasperi, Q.Re Paolo VI, 74100 Taranto, Italy
Received June 10, 2010. Revised manuscript received August 25, 2010. Accepted August 27, 2010.
UV photolysis of aqueous hydrogen peroxide samples was carried out in the presence of methanol, ethanol, or t-butanol. The concentrations of H2O2, dissolved O2, and the alcohols were monitored as a function of time, and a quantitative chemical kinetics model for the photolysis of the solutions is presented. The observed kinetics consisted of an initial rapid consumption of dissolved oxygen followed by a significant acceleration in the photodecomposition of hydrogen peroxide. The acceleration phase was identified to originate from the fast feedback reaction between hydrogen peroxide and the carbon centered radicals resulting from hydrogen atom abstraction from the primary alcohols. In tertiary butanol solutions the radical species formed are more stable and do not react directly with H2O2. As a consequence no significant acceleration of H2O2 photolysis was observed in the presence of t-butanol.
Introduction As widely documented in the literature, endocrine disruptors, personal care products, pharmaceutically active compounds, and toxic industrial contaminants are being increasingly detected in precious water sources such as lakes and groundwater (1). It is well established that the main point of entry into the environment for many of these micropollutants is via wastewater streams (2) which, after treatment, may be discharged into sensitive receiving bodies as well as reused in agriculture, industrial, recreational, and indirect potable reuse applications. As most municipal plants are not yet equipped with advanced tertiary treatment (3), advanced oxidation processes (AOP), when employed as a barrier against emerging micropollutant release to the environment, may be required to operate with a wastewater effluent quality where carbonaceous organic substrates may still represent a considerable fraction of the total effluent organic matter * Corresponding author e-mail:
[email protected]. † California State University at Northridge. ‡ University of Western Ontario. ¶ Technical University of Bari. 10.1021/es101959y
2010 American Chemical Society
Published on Web 09/09/2010
(4). Similar process conditions may apply during advanced oxidation of high strength organic liquid effluents ranging from landfill leachate to winery wastewaters (5) and even during the pretreatment of natural organic matter (NOM), where alcohol groups are typically present in both hydrophobic and hydrophilic neutral fractions (6, 7). More specifically, the presence of readily biodegradable chemical oxygen demand (rbCOD) may influence the AOP in several ways by, e.g., augmenting the oxidant demand and decay, increasing the hydroxyl radical scavenging capacity, augmenting byproduct (e.g., aldehydes) formation potential, and promoting the formation of carbon-centered radicals, which may scavenge the dissolved oxygen in the effluent. In this paper, the mechanism of reaction and the kinetic model for UV-H2O2 AOP in the presence of high concentrations of rbCOD such as methanol, ethanol, and t-butanol has been developed and validated based on the experimental batch data. In particular, the role of carbon-centered radicals in the process has been investigated, and their ability to scavenge dissolved oxygen and accelerate the decomposition of the oxidant is characterized. The validated kinetic model may also be helpful to assess the AOP potential for the remediation of groundwater contaminated by fuel oxygenates such as t-butanol (8). The reaction of hydroxyl radicals with primary alcohols (e.g., methanol, ethanol) in the gas phase has been extensively studied (9-14). In the gas phase hydroxyl radicals abstract a hydrogen atom from either the hydrocarbon part or the alcohol group with a slight preference given for the former channel at low temperatures (10). In aqueous solutions, both channels have been observed to lead to mostly carbon centered radicals because the latter reaction product rapidly rearranges to the corresponding carbon centered radical species (15). Furthermore a series of complex radical chain reactions takes place in the liquid phase, which have been primarily studied by generating the initial hydroxyl radicals by photolysis, pulse radiolysis, or sonolysis (16-20). In addition to the decomposition of H2O2, these processes also lead to a gradual degradation of the alcohols (i.e., mineralization). For example, for methanol, this radical chain reaction leads to the production of intermediate species such as aqueous formaldehyde (formaline), formic acid, and then finally gaseous end products such as H2, CO, and CO2 depending on factors such as the dissolved oxygen concentration and the conditions that the samples are exposed to (16, 19, 21). UV photolysis of aqueous H2O2 solutions provides a very selective and efficient way to produce hydroxyl radicals. This process has been extensively studied, and the corresponding kinetic model is well established in the literature (22) H2O2 + hν f ·OH + ·OH with k1 ) φH2O2Ia
(1)
H2O2 + ·OH f HO2· + H2O with k2 ) 2.7 × 107 M-1 s-1 ref (22) (2) HO2· + HO2· f H2O2 + O2 with k3 ) 8.3 × 105 M-1 s-1 ref (22) (3) H2O2 + HO2· f ·OH + O2 + H2O with k4 ) 3.0 M-1 s-1 ref (22) (4) ·OH + ·OH f H2O2 with k5 ) 5.5 × 109 M-1 s-1 ref (22) (5) where M denotes molarity (mol dm-3), φH2O2 denotes the quantum yield for photodissociation of H2O2 (dimensionless), VOL. 44, NO. 20, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7827
and Ia is the intensity of light absorbed (M s-1). Values for k1 in eq 1 in the following experiments varied between 9 × 10-6 to 5 × 10-5 M s-1 depending on the instrumental factors such as the lamp position and the reactor filling, which essentially influence the effective value of Ia.
Materials and Methods The following chemicals were used: 50% H2O2 (Fischer Scientific), Milli-Q grade water (unbuffered), and N2 gas (Airgas, Inc.) for purging the samples. The continuously stirred semibatch reactor consisted of a cylindrical glass vessel (V ≈ 80 cm3; 298 ( 2 K and 1 atm) that was exposed to UV irradiation from an unfiltered low pressure UV lamp with a quartz housing (model 11SC-1, UVP, LLC, Upland, CA). The primary actinometric standard was based on the wellestablished iodide-iodate method (23), which was used to calibrate the incident radiation to provide Ia. In the current H2O2 concentration regime (6-45 mM), which was deliberately selected to be sufficiently high to isolate the kinetics, the quantum yield of aqueous H2O2 photolysis is approximately one at 254 nm. The secondary standard was based on the kinetic model of eqs 1 to 5, which was used to obtain the effective k1 in eq 1 prior to the actual measurements of the alcohol containing solutions (i.e., photolysis of pure aqueous H2O2 solution). All the measurements were carried out in the full absorption regime where all UV photons were absorbed by the sample (i.e., optically thick samples). This condition was accomplished by choosing a relatively long optical path length in photolysis (ca. 5 cm), sufficiently high concentration for H2O2 (6-45 mM), and limiting the photolysis time accordingly. The working solution, with approximately a constant pH value of 5, was placed in the reactor and stirred continuously during photolysis. The system was exposed to the atmosphere, which means that there was a steady but slow diffusive exchange of O2 through the liquid surface. This was incorporated into the kinetic model by the following equation d[O2] ) -kr([O2] - [O2]eq) dt
(6)
where the O2 exchange rate constant kr ) 6 × 10-4 s-1 was obtained by fitting the model to the experimentally observed O2 concentration in the reactor as a function of time by observing the O2 recovery of oxygen free water solution. The equilibrium dissolved O2 concentration was determined as [O2]eq ) 276 µM. The oxygen concentration was monitored by using a dissolved O2 electrode (Van London pHoenix, Co. model 027NG15), which was immersed in the solution (24). The current output from the electrode was amplified by a current amplifier (105 gain coefficient; Keithley model 428) and digitized by a digital voltmeter (Agilent model 34401A). The initial dissolved oxygen concentration was varied by running photolysis of aqueous H2O2, which produces O2, prior to injecting the alcohol. Samples from the reactor were taken continuously by using a peristaltic pump (MasterFlex model LS 7518-10) for analysis by UV/vis spectroscopy (D2 lamp model Beckman 96280 and Newport UV CCD spectrometer model 78125) at 254 nm (ε254nm ) 20 M-1 cm-1 for H2O2) or by electron spin resonance (ESR; flat fused silica cuvette for working with aqueous samples) and returned to the reactor after analysis. The ESR experiments were carried out by using a Bruker ESP-380 X-band spectrometer with 1 G modulation amplitude, 100 kHz modulation frequency, and 4 mW microwave power. To trap the short-lived radicals, an aqueous solution containing the spin trap compound (DMPO; 5,5-dimethylpyrroline-N-oxide) was injected into the reactor such that the overall spin trap concentration was approximately 25 mM. Time vs alcohol concentration profiles were obtained by the combined gas chromatography - mass 7828
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 20, 2010
spectroscopic (GC-MS) method using the Hewlett-Packard G1800B GCD system equipped with a 15 m wax column (0.32 mm inner diameter; column temperature 40 °C; gas flow 1 mL/min; sample volume 1 µL; column model BP21, SGE, Inc.). To model the kinetics of UV irradiated aqueous H2O2alcohol solutions, we have numerically integrated the presented kinetic models by using the LSODA routine (25, 26). The simplex algorithm was used to fit simultaneously both H2O2 and O2 concentrations against the experimentally observed data to yield estimates for the rate constants (27). The time vs alcohol concentration profiles were included in the fitting process when these data were available. The ESR spectra were analyzed by using the XEMR package (28).
Results and Discussion Photolysis of Aqueous H2O2-CH3OH Solutions. When methanol is introduced into aqueous H2O2 solution, the kinetic model must be modified to include the radical reactions involving methanol CH3OH + ·OH f ·CH2OH + H2O with k7 ) 9.7 × 108 M-1 s-1 ref (17)
(7)
H2O2 + ·CH2OH h CH3OH + HO2· with k8+ ) 5.0 × 103 M-1 s-1, k8- ) 0.6 M-1 s-1 (8) H2O2 + ·CH2OH f CH2O + ·OH + H2O with k9 ) 6 × 104 M-1 s-1 ref (17)
(9)
where eq 7 is in competition with eq 2 and eqs 8 and 9 compete for producing HO2 · (pKa ) 4.8; both HO2 · and O2- · forms denoted collectively by HO2 · ) and · OH, respectively. Note that methanol and the other alcohols used in this study are not photolyzed directly by 254 nm light. Since k2
