Hydroperoxyl radical/superoxide anion radical (HO₂·/O₂ -·, pK_a = 4.8) as an oxidant is of considerable interest toward the fundamental understanding of the multiple roles in its chemical reactions [1-7]. In numerous advanced oxidation processes employing UV/H₂O₂, O₃, Fenton, radiolysis, and electron beam process, HO₂·/O₂ -· has been considered as a propagator and/or intermediate of the ensuing chain reactions of radicals during the photochemical and oxidation processes [8-11]. Furthermore, H₂·/O₂ -· has been considered to be indicative of a source of the hydroxyl radical (·OH) as a strong oxidant in biological system [12-16]. Thus, it is very important to understand physicochemical characteristics of HO₂·/O₂ -· in various oxidation processes.

In order to elucidate characteristics of HO₂·/O₂ -·, various methods have been used in the detection of HO₂·/O₂ -·. The most direct techniques of these can be electron spin resonance (ESR) and the optical absorption detection by means of spin trapping and characteristic UV spectra of HO₂ and O₂ - (240？260 nm), respectively [3,17]. However, ESR can be detected only in ice at very low temperature, whereas UV absorbance can lead to the possibility of complicated reactions arising from interference of the solute, i.e., spectrum overlapping as many chemical species, with the primary processes [3,18]. Hence, these methods are of very low sensitivity and low selectivity in practical applications [5]. The most widely used method of detecting HO₂·/O₂ -· is through the use of such tetranitromethane, nitroblue tetrazolium, and cytochrome C which form products with intense optical absorbance [3]. However, these spectrophotometric methods have also suffered from low sensitivity [5,19,20].

Recent studies have reported a chemiluminescent method using luminol and enzyme, which are disadvantageous as these reagents are unstable and expensive [21-23]. In addition, Kwon and Lee [5] have reported a kinetic method by HO₂·/O₂ -·-driven Fenton-like chemistry in UV/H₂O₂ process, which detected only o- and m-hydroxybenzoic acid (OHBA isomers) as fluorescent hydroxylated derivatives produced from benzoic acid (BA). However, the kinetic method for HO₂·/O₂ -· by using BA as a probe has produced various products including p-OHBA, decarboxylated product (i.e., phenol), and dihydroxybenzoic acid (i.e., 2,3- , 2,4-, and 2,5-diOHBA), and so on, which would interfere with the analysis of o- and m-OHBA isomers [24-26]. First above all, these products could not be detected in the fluorometer employed in the kinetic method. As a result, the sensitivity of the kinetic method by using BA probe might be decreased by various products. In continuation of our earlier works on the kinetic method by using BA [5], we studied the enhanced sensitivity for determining HO₂·/O₂ -· compared to BA. Therefore, an alternative method of detecting HO₂·/O₂ -· with high sensitivity is necessary to be developed to elucidate characteristics of HO₂·/O₂ -· having

very low concentration in oxidation processes.

In the present paper, we reported investigations on a more advanced kinetic method for HO₂·/O₂ -· determination by employing terephthalate or terephthalic acid (TA) as a probe. TA was used to trap ·OH in that TA might produce only hydroxyterephthalic acid or hydroxyterephthalate (OHTA) in the calibration procedure, which can be determined by fluorescence measurements.

BA, 3%H₂O₂, NaOH, Fe(III)-ethylenediaminetetraacetic acid (Fe³⁺ -EDTA), HCl, and so on were the American Chemistry Society (ACS) grade (Sigma-Aldrich, St. Louis, MO, USA) and used as received. Owing to its impurity, the purification of disodium terephthalate (Aldrich, 96%) was checked by UV/visible spectrophotometer (UV1601; Shimazu, Kyoto, Japan) that had been recrystallized several times with hot water and then by drying in a dry oven (60 ± 5℃), and no impurities were observed.

A commercial product of OHTA failed to find information on its properties. Hence, OHTA was synthesized from previous studies [27-29]. The OHTA synthesized could be checked by the spectra of a fluorescence detector (LS 50B; Perkin-Elmer, Buckinghamshire, UK), as shown in Fig. 1. Furthermore, as shown in Fig. 2, the OHTA synthesized showed the linearity of the response of the measurement systems with fluorescence detector with a good correlation coefficient (R² = 0.997), which was performed with working solutions and pure deionized water instead of working solutions, respectively. In this study,synthesized was used to only identify OHTA formation with a fluorescence detector (474 model; Waters, Milford, MA, USA) in the apparatus used in this study.

All experiments were conducted with proper buffer solutions, ammonium acetate buffer, acetate buffer, borate buffer and carbonate buffer, by addition of the respective 1 M H₂SO₄ or 0.5 M NaOH because experimental solution on the pH-dependent rate constant (k_obsd) [3] should be kept at constant pH during the reaction of HO₂·/O₂ -· in actual analysis. All solutions were prepared in

triple deionized water, which was further purified by an aqua- Max water system (Young Lin Co., Anyang, Korea).

The apparatus and procedures for HO₂·/O₂ -· measurement employed in the present study was almost similar to a previous study [5], except for a home-made glass debubbler. Briefly, all solutions were delivered by using a peristaltic pump (Ismatec, Glattbrugg, Switzerland) with polytetrafluoroethylene (PTFE) tubing (Cole-Parmer, Vernon Hills, IL, USA; i.d., 0.8 mm). The H₂O₂ solution (0.42 mL/min) was photolyzed by UV irradiation, which was equipped with a 4 W low pressure Hg lamp (λ_max = 254 nm; Philips, Amsterdam, Netherlands). The HO₂·/O₂ -· stream produced from the photolysis of H₂O₂ in a quartz coil-type reactor was passed through a knotted tube reactor (KTR), which was able to control the concentration of HO₂·/O₂ -· and, at the same time, to completely premix the solutions.

Then, Fe³⁺-EDTA solution (0.23 mL/min) was mixed with TA solution (0.23 mL/min) in KTR and then joined with the HO₂·/O₂ - ·-H₂O₂ stream. The reduction of Fe³⁺-EDTA by HO₂·/O₂ -· resulted in the production of Fe²⁺-EDTA, which reacted further with residual H₂O₂. A Fenton-like reaction of H₂O₂ and Fe²⁺-EDTA produced the ·OH radicals, which were then scavenged by TA to produce hydroxylated products in KTR. After 0.05 N NaOH (0.23 mL/min) was added to raise the pH level above 11, the fluorescence intensity of OHTA could be maintained at a maximum level.

The mixed solution occasionally caused the formation of air bubbles in the effluent stream. Air bubbles were then removed by a home-made glass debubbler prior to the entry of a fluorometer in order to prevent noise signal by the air bubbles. The fluorescent signal was transferred to a data acquisition system, Autochro-Win (Young Lin Co.), consisting of an analog-to-digital converter with a personal computer.

The OHTA fluorescence was monitored with a fluorometer (Waters 474 model) equipped with a 16 μL flow-through cell using 315 nm (excitation)/425 nm (emission) with slit-width of 40 nm. The fluorescent intensity of the OHTA was recorded with the fluorescence detector as a function of time.

The OHBA fluorescence was measured with a fluorometer equipped with a 16 μL flow-through cell using 320 nm (excitation)/400 nm (emission) with a slit-width of 40 nm.

The prominent feature of the kinetic method was the ability to calibrate the concentration of HO₂ ·/O₂ -· without a primary standard [5]. Briefly, to calibrate the HO₂·/O₂ -· measurement system, all working solutions were passed through the appropriate ports under UV lamp-off and the base lines were monitored. The H₂O₂ solution placed in UV photolysis under UV lamp-on was photolyzed and then passed through the gradient-length of KTR, which was controlled step-wise as 0, 1, 2, 3, and 4 m. In the absence of additives, HO₂· and O₂ -· in each KTR were disproportionated by self-reactions of R3-R5 (Table 1) according to the empirically observed pH-dependent rate constant, k_obs [3], assuming that reaction R5 is negligible due to its very small rate constant (k₅ < 0.3/M/sec):

where k_obs can be calculated using k₃ = (8.3 ± 0.7) × 10⁵/M/sec, k₄ = (9.76 ± 0.6) × 10⁷/M/ s, and K_HO2 = 1.6 × 10^-5/M as recommended values [3]. A stream containing HO₂· and O₂ -· produces ·OH by the mechanism including a Fenton-like reaction. In this manner, each fluorescence signal of OHTA produced from the ·OH radical reaction with respective TA was proportional to a produced HO₂·/O₂ -· concentration. The half-life (t_1/2) of the HO₂·/O₂ -· was obtained by plotting a simple linear relationship of the signal ratio vs. reaction time. The half-life in a second-order reaction was inversely proportional to the initial concentration. Thus, the initial concentration of HO₂·/O₂ -· could be kinetically calculated from the disproportionation reaction of HO₂·/O₂ -· based on the half-life and calculated k_obs at a given pH.

The pH was measured by pH meter (Model PCM700; Orion Research Inc., Jacksonville, FL, USA).

Table 1 shows the reaction scheme considered from this study. As shown in Table 1, the reaction scheme was comprised of the previously reported reactions and second-order rate constants (k). In the R1 ofion scheme, H₂O₂ was photo-decomposed to produce two ·OH radicals by an UV irradiation at wavelength 254 nm in the quartz coil-type reactor. Most of the ·OH radicals formed in the quartz coil-type reactor under UV irradiation could react with residual H₂O₂, producing HO₂·/O₂ -·

(R2). Hence, the HO₂·/O₂ -· stream was produced from the photolysis of H₂O₂ only in a quartz coil-type reactor.

Hydroperoxyl radical formed from the quartz coil-type reactor is dependent upon the acid-base equilibrium with pK_a = 4.8 (R6). On the basis of the reaction scheme shown in Table 1, the k_obs and ratio of [O₂ -·]/[HO₂·] (and [HO₂·]/[O₂ -·]) was estimated as shown in Fig. 3. At pH 4.8, the relaxation time for the dissociation equilibrium of HO₂· was the order of 1 μs [30]. The k_obs of HO₂·/O₂ -· disproportionation was recalculated from the result of Bielski et al. [3]. In addition, owing to physicochemical properties (i.e., solubility and mobility) of HO₂·/O₂ -· on solution pH, the effective Henry’s law coefficient (H*) in Eq. (2) of HO₂·/O₂ -· was estimated, considering solution pH:

where H is Henry’s law coefficient (M/atm) of HO₂· at 25℃, and K is the equilibrium constant from R4 in Table 1. Since uncertainty in H (1.2 ~ 6.8 × 10³ M/atm) might be somewhat [31,32], we recalculated H* with updated datum of H (2.0 × 10³ M/atm) [33]. As shown in Fig. 3, for pH values smaller than 5, HO₂·/O₂ -· does not dissociate appreciably and its effective Henry’s law constant is

equal to its Henry’s law constant. However, for pH values larger than 5, the effective Henry’s law constant of HO₂·/O₂ -· depending on various pH values increased with pH, and thus its solubility and mobility would increase with pH [7].

Fe³⁺-EDTA could alter the reactions of R3 to R5. Fe³⁺ -EDTA was reduced by HO₂·/O₂ -· to produce Fe²⁺ -EDTA and O₂ with R7 [5]. Fe²⁺-EDTA and H₂O₂ led to the production of the ·OH radical and regeneration of Fe³⁺-EDTA in R8. Then, the ·OH radical produced OHTA in the presence of TA with a nearly diffusion-controlled rate constant of k₉ = 3.3 × 10⁹/M/sec (R9) [8]. Fe³⁺-EDTA, however, may compete with 1 mM TA for the ·OH radicals. To minimize the scavenging of ·OH radicals by Fe³⁺-EDTA, a low concentration of 20 μM Fe³⁺ -EDTA to keep the ratio of k₉[TA] to k₁₀[Fe³⁺-EDTA] larger than 150 was maintained. In other words, about 99.34% of the ·OH radicals will kinetically react with TA, and other competitive reactions for ·OH radicals would be negligible.

In the meantime, in the first paper by Kwon and Lee [5], the kinetic method adopted BA as a probe was used to demonstrate the HO₂·/O₂ -· detection system. Although the new setup as described in Kwon and Lee [5] has the advantage that the kinetic method was very sensitive, it introduced the potential loss of p- OHBA as a product on the detection of HO₂·/O₂ -· as well as the interference of by-products, which would lead to an error in the determination of HO₂·/O₂ -· by using BA in the kinetic method. As shown in Fig. 4, the HO₂·/O₂ -· loss in the detection system by using BA existed, compared to using TA. Hence, HO₂·/O₂ -· concentrations measured by using BA were lower than those using TA. In this case, the sensitivities of the kinetic method by using TA were always higher with 1.7？2.5 times at pH 8.0 than those by using BA (Fig. 4). This result suggests that ·OH trapping with TA instead of BA can be effective. Therefore, the magnitude of this loss on HO₂·/O₂ -· can be substituted with OHTA using TA as a probe of ·OH.

Fig. 5 shows the concentration of HO₂·/O₂ -· generated from H₂O₂ photolysis at a wavelength 254 nm under two experimental conditions: pH and H₂O₂ concentration. As shown in Fig. 5, the concentration of HO₂·/O₂ -· at the ranges of pH 5？10 was dependent upon both various pH values and H₂O₂ concentrations added. The concentration of HO₂·/O₂ -· increased with pH value, and its concentration measured at pH 5 was lower than that at pH 10. This result suggests that O₂ -· slowly decayed at high pH

conditions. To the best of our knowledge, one of the plausible explanations for the decay of HO₂·/O₂ -· is the k_obs on the disproportionation reactions of R3？R5 and the pH dependences of the acid-base equilibrium between HO₂· and O₂ -·. As shown in Table 1, self-disproportionation by reactions R3？R4 will be fast and dominant at an acidic pH range, whereas disproportionation by reaction R5 at high pH ranges (pH > 7) is very slow with its k_obs. Hence, there is a low loss of HO₂·/O₂ -· concentration at high pH ranges and then its concentration is relatively high. These results are consistent with the kinetic data for the disproportionation reaction of HO₂·/O₂ -· presented by Bielski et al. [3].

For a more profound account for the pH dependence of HO₂·/ O₂ -· concentrations generated in a continuous flow system, a series of experiments were carried out over ranges of pH 2？10 and a range of initial H₂O₂ concentration (4 mM). In each experiment, the life time of HO₂·/O₂ -· formed during H₂O₂ photolysis was investigated on the basis of the kinetic method over various pH conditions.

Fig. 6 shows the life-time of HO₂·/O₂ -· as a function of solution pH at 4 mM H₂O₂ concentration. As shown in Fig. 6, its life times were the lowest with approximately 51 sec at pH 4.8？5.0 (= near pK_a), with increases gradually on either side of this value. This trend was due to the k_obs on the disproportionation reactions of R3？R5. The basic principle for the life time of HO₂·/O₂ -· was based on the half-life (t_1/2) of HO₂·/O₂ -· decay. In order to investigate the t_1/2, the decay rate of HO₂·/O₂ -· by the reactions R3 and R4 becomes Eq. (3), assuming that the disproportionation of reaction R5 is negligible due to its very small rate constant (k₅ < 0.3/M/sec).

From Eq. (3), t_1/2 can readily be determined from Eq. (4) at a given pH [5]. Finally, the life time (t) of HO₂·/O₂ -· is Eq. (5) at a given pH.

As a result, the life time of HO₂·/O₂ -· as well as it concentration was highly pH dependent. In particular, at ranges of pH

2？10, its life-time was approximately from 51 to 422 sec as shown in Fig. 6. In particular, the increase in the life-time with pH was observed as expected. This result is consistent with the fact that is estimated with the observed second-order rate constants for spontaneous decay of HO₂·/O₂ -· by Bielski et al. [3]. At high pH > 8, the life-time of O₂ -· can be estimated to be larger than 100 sec, if [HO₂·/O₂ -·] > 10^-6 M. McDowell et al. [17] presented from 90 sec at pH 11.0 to 41 min at pH 12.5 with the half-time for decay of 100 μM O₂ -·. Bieski et al. [3] observed that, at pH 10, relatively slow rate processes for HO₂·/O₂ -· reaction could be determined because the first half-life for spontaneous disproportion of O₂ -· was relatively long into 70？100 sec. Although a little difference of the life-time of HO₂·/O₂ -· between measured and predicted exists, this result can be acceptable. The chemical life-time of O₂ -· at pH > 6 increases proportionally with pH, whereas its lifetime at pH < 4 increases inversely with decreasing pH. However, at about pH 4？6, the chemical life-time of HO₂·/O₂ -· is relatively short because of the spontaneous disproportionation reaction. Thus, these results suggest that the chemical life-time of HO₂·/O₂ -· is dependent upon the kinetics with the observed secondorder rate constant of HO₂·/O₂ -·.

We have newly adopted TA as a probe for the measurement of HO₂·/O₂ -· in the kinetic method presented in our previous study. Our study results showed that concentrations of HO₂·/O₂ -· formed from the photolysis technique of H₂O₂ at a range of pH 2？10 increased with pH. In addition, an increase in the life-time of HO₂·/O₂ -· with pH was observed. Since the kinetic method by using TA was highly sensitive, this study can contribute to understanding the basic function of HO₂·/O₂ -· in advanced oxidation processes.