Destruction of humic acid in water by UV light—catalyzed oxidation with hydrogen peroxide

RESEARCH NOTE

DESTRUCTION OF HUMIC ACID IN WATER BY UV

LIGHTÐCATALYZED OXIDATION WITH HYDROGEN

PEROXIDE

GEN-SHUH WANG1_*M_{, SU-TING HSIEH}1 _{and CHIA-SWEE HONG}2 1_{Department of Public Health, National Taiwan University, Rm 1432, No. 1, Sec. 1, Jen-Ai Road,}

Taipei, Taiwan, Republic of China and2_{Wadsworth Center, New York State Department of Health,}

Albany, NY 12201-0509, USA

(First received 8 September 1999; accepted in revised form 24 December 1999)

AbstractÐA batch reactor was used to evaluate the advanced oxidation process of the UV/H2O2

system for control of natural organic matter (NOM) in drinking water. The light sources used include a 450 W high-pressure mercury vapor lamp and sunlight. Both quartz and Pyrex ®lters were used to control the wavelength and energy of UV light applied to the aqueous systems. The results showed that NOM oxidation and H2O2 decomposition followed ®rst-order and zero-order reaction kinetics,

respectively. The optimum H2O2dose was found to be 0.01% for the oxidation of humic acids in this

study. Carbonate and bicarbonate ions inhibited the degradation of humic acids. 7 2000 Elsevier Science Ltd. All rights reserved

Key wordsÐhumic acid, hydrogen peroxide (H2O2), advanced oxidation process (AOP)

INTRODUCTION

The presence of natural organic matter (NOM) in both surface and ground water supplies has received much public attention in recent years because toxic disinfection byproducts (DBPs) can result from chlorination procedures in the water treatment pro-cesses. The removal of NOM from raw water is lar-gely achieved by chemical coagulation and ¯occulation, using aluminium sulfate or ferrous sul-fate and lime in the presence of excess chlorine. However, the NOM removal in conventional drink-ing water treatment processes is quite low, between 10 and 50% (Jacangelo et al., 1995).

Due to the presence of a wide variety of NOM and the necessity of the chemical disinfection pro-cess to protect the public health, many water utili-ties have to face the problem of DBPs formation. For the control of DBPs, some treatment alterna-tives have been proposed (Owen et al., 1995). Among the proposed methods for the control of DBPs, advanced oxidation processes (AOP) deserve more study for potential applications in drinking water treatment. AOP can e€ectively mineralize

many organic contaminants and have become attractive for the control of synthetic organic com-pounds in wastewater treatments (Kang and Lee, 1997). AOP have many advantages in water treat-ment processes and have been proposed as an alternative for the control of DBP precursors (Symons and Worley, 1995; Eggins et al., 1997).

The UV/H2O2 process is an example of a geneous AOP. Generally, the e€ectiveness of homo-geneous light-driven oxidation processes is associated with very reactive species, such as hy-droxyl radicals, which are generated in the reaction mixture by the direct photolysis of H2O2under UV irradiation:

H2O2‡ hn42OH

The hydroxyl radicals attack organic compounds relatively non-selectively with rate constants ranging from 106_{to 10}10_Mÿ1_sÿ1_{(Buxton et al., 1988),} oxi-dizing them by hydrogen atom abstraction or by addition to double bonds. The UV/H2O2 process has been proven e€ective in treating waters contain-ing a number of aliphatic and aromatic compounds (BeltraÂn et al., 1993; Sundstrom et al., 1986). In UV/H2O2 process, the optimum H2O2 dosage should be obtained and applied since the excess H2O2dose can reduce the oxidation rate (Ku et al., 7 2000 Elsevier Science Ltd. All rights reserved Printed in Great Britain 0043-1354/00/$ - see front matter

3882 www.elsevier.com/locate/watres

PII: S0043-1354(00)00120-2

*Author to whom all correspondence should be addressed. Tel./fax: +886-2-23940612; e-mail: [email protected]. ntu.edu.tw

1998). The presence of Cu2+_{will enhance the H} 2O2 decomposition and increase the rates of organic oxi-dation (Baxendale and Wilson, 1957), however, the carbonate/bicarbonate ions will inhibit the oxi-dation of organics due to its scavenging e€ect (Liao and Gurol, 1995). Despite the increasing interest in this process for water and wastewater treatment, there is very little information available regarding destruction of the NOM by the UV/H2O2 process (Symons and Worley, 1995). The aim of this paper is to examine the in¯uence of light intensity, the in-itial concentration of humic acid, H2O2 dose, and the matrix of carbonate/bicarbonate ions on the photodegradation of humic acid in order to evalu-ate the applicability of UV/H2O2technology to the remediation of DBP precursors in water.

EXPERIMENTAL

Humic acids extract, H2O2 (30% solution), sodium

car-bonate, and sodium bicarbonate used in this study were purchased from Nacalai Tesque (Kyoto, Japan). Milli-Q water (Millipore) was used to prepare the humic acid stock solution (which contained 1000 mg/l of non-purge-able organic carbon, NPOC). Before each experiment, aqueous solutions were prepared by diluting the humic acid stock solution and adding with known amounts of H2O2and carbonate/bicarbonate.

A 10-l stainless-steel batch reactor (20 cm diameter  30 cm depth) with a quartz window was used in this study for the photooxidation of humic acid in aqueous solutions (Hu and Yu, 1994). A 450 W high-pressure mercury-vapor lamp (Hanovia, Ace Glass Co., Vineland, NJ) was used as the light source. A UV lamp was inserted into the hollow quartz or Pyrex tube (55 mm od, 45 mm id) located at the center of the reactor. A water-cooling loop was used to prevent the lamp from overheating and to maintain the water at room temperature (258C). The UV light was turned on ten minutes before performing the experiment. Eight liters of aqueous solution was then added into the reactor. The solution's mixing was achieved with a mag-netic stirrer. Two aliquots of 30 ml were used to measure initial humic acid concentration; duplicate samples of 30 ml were withdrawn from the reactor at various time intervals for analysis. All of the experiments were con-ducted at pH=7 except the ones to study the e€ects from carbonate ions, which were conducted at pH=10. Control experiments were carried out under the same conditions in the dark.

The experiments were also carried out with direct sun-light. Four 2-l glass beakers, each ®lled with 1.5 l of water and then covered with a quartz plate, were put on the roof of a 15-story building (at 70 m elevation) in down-town Taipei between 11 am and 2 pm in July 1998. Unex-posed controls were run with each batch of test samples to establish starting humic acid concentrations and also to monitor the consistency of the results. The control reac-tors were wrapped with aluminum foil to prevent UV ex-posure; they were set outside with the test samples.

Samples were ®ltered through a 0.45 mm syringe ®lter to remove particles and then acidi®ed with two drops of 2 N HCl solution prior to TOC analysis. The NPOC concen-tration was measured by an organic carbon analyzer (Shi-madzu TOC 5000A), and the UV absorbance was measured by a high-precision, double-beam spectropho-tometer (Shimadzu UV 160A). The concentration of the hydrogen peroxide was determined by UV absorbance spectrometry at a wavelength of 260 nm (UV260). For

cor-rection of the absorbance contributed from humic acid in

the water, the UV260from humic acids was deducted from

the total absorbance.

A UV radiation spectroradiometer (MSS 2040, Elekro-nik Gmbh, Bielefeld, Germany) was used to measure the UV spectrum and energy distribution of the light source. The spectroradiometer was placed outside the quartz or Pyrex window of the reactor to record the wavelength and energy distribution of the UV light. Figure 1 shows the UV spectra of the UV/quartz, UV/Pyrex, and sunlight that were used for the oxidation of humic acids in this study. The major spectral distribution of the UV lamp was at wavelengths of 254, 265, 297, 302, 312, 365, 405, and 435 nm, if the quartz tube was used as the cooling device. However, the UV light at 254, 265, 297, and 302 nm was ®ltered out when the Pyrex cooling tube was used. When sunlight was used as the light source, a continuous UV spectrum was observed at wavelengths longer than 300 nm. The light intensities for UV/quartz, UV/Pyrex, and sun-light were 275.8, 20.8, and 23.2 W/m2_{, respectively.}

RESULTS AND DISCUSSION Direct photolysis of hydrogen peroxide

The photolysis of H2O2in pure water was studied extensively by various groups of researchers in the 1950s (Baxendale and Wilson, 1957; Weeks and Matheson, 1956; Hunt and Taube, 1952) in order to elucidate the reaction mechanism and to deter-mine the primary and overall quantum yields. The direct photolysis of H2O2generates the very highly oxidizing and reactive
OH radicals, which then participate in the steps leading to the degradation

Fig. 1. Wavelength and energy distribution of the UV sources used in UV/H2O2 oxidation of humic acid. Ð:

UV/quartz; ± ± ±: UV/Pyrex; - - -: sunlight.

Fig. 2. Decomposition of hydrogen peroxide in the UV/ H2O2processes. [H2O2]=0.1%; UV source: UV/quartz.

of organic pollutants (Volman, 1949). Figure 2 shows the experimental data obtained during the ir-radiation of a 0.1% hydrogen peroxide aqueous sol-ution, with a rate of 0.30 mM minÿ1_{. Although} ®rst-order decomposition was observed in literature (Liao and Gurol, 1995), the initial photolysis of hydrogen peroxide follows zero-order kinetics within the H2O2addition tested in this study (0.01± 0.5%).

Degradation of humic acid sensitized by hydrogen peroxide under UV light

The production of hydroxyl radicals from hydro-gen peroxide requires a large dissociation energy (213 kJ moleÿ1_{) in order to cleave the O±O bond,} which means that short wave UV-C energy (wave-lengths of 200±280 nm) is necessary to lead to use-ful radical yields (Atkins, 1990).

Both hydrogen peroxide and humic acids are weakly absorbing compounds in the UV range (Fig. 3). In the case of H2O2, the absorption increases as the wavelength decreases. When the sol-utions was irradiated by UV light, H2O2 absorb much more light than humic acids at wavelength less than 300 nm, as shown in Fig. 3. Therefore, in dilute aqueous solutions of humic acid (010 mg/l) in the presence of various concentrations of hydro-gen peroxide (0.001±0.5%), H2O2 is the principal absorber of UV light (Beltran et al., 1997).

Figure 4 compares the destruction rate of humic acid irradiated by three UV light sources with an initial peroxide concentration of 0.1%. Approxi-mately 90% of the humic acid was oxidized within 1 h of irradiation when UV/quartz was used. How-ever, oxidation of only 20% of humic acid was observed after 2 h irradiation when UV/Pyrex or sunlight was used as the light source. The lamp with the quartz ®lter has the highest intensity in the UVC region, which favors its use for the activation of hydrogen peroxide. The employment of such a UV source will result in higher OH radical yields from the increased emission of UVC compared with the other sources, UV/Pyrex and sunlight. Of the

three di€erent light sources, the rate of humic acid decay with UV/quartz was the greatest. With UV/ quartz, the humic-acid decay follows ®rst-order kin-etics with an observed rate constant of k ˆ 0:037 minÿ1_{, while that of hydrogen peroxide}

obeys zero-order kinetics with a rate of 0.30 mM minÿ1_{. Humic acid decomposed at a slower rate} with UV/Pyrex …k ˆ 0:0025 minÿ1_{† and sunlight}

…k ˆ 0:0007 minÿ1_{), which have minor UVC}

emis-sion and comparable total UV light intensity com-pared to UV/quartz. Because the primary quantum yield of hydrogen peroxide at 254 nm is very high, f_H2O2 ˆ 0:5 molecule/photon (Baxendale and Wilson, 1957), with UV/quartz, H2O2 could undergo photolysis and oxidation of humic acid would be due mainly to the UV/H2O2process. The degradation of humic acid with UV/Pyrex or sun-light could result from both direct photolysis of humic acid and the UV/ H2O2process (Wang et al., 1997).

E€ect of H2O2concentration

Although H2O2 does not oxidize humic acid at all, as observed in this work, when combined with UV radiation the rate of humic acid oxidation increases extraordinarily compared to that of direct photolysis. Data showing the e€ect of H2O2 concen-tration on the pseudo-®rst-order rate constants for humic acid degradation are plotted in Fig. 5. The initial H2O2 concentration varied from zero to 0.5% (147 mM), the UV dose rate was 275.8 W/m2 for each run, and the initial concentration of humic acid was set at 5 mg/l. The destruction rate of humic acid increased with the increase of hydrogen peroxide concentration up to 0.01% and then decreased with further increases of H2O2 concen-tration. Similar experimental results have been reported by Ku et al. (1998). In this process, hy-droxyl radicals generated from the direct photolysis of H2O2 were the main responsible species for humic acid elimination. However, H2O2 also reacted with these radicals and hence acted as an

Fig. 4. UV/ H2O2 oxidation of humic acid with various

UV sources. [humic acid]0=5 mg/l; [H2O2]0=0.1%; (*)

UV/quartz; (Q) UV/Pyrex; (R) sunlight; solid symbol for humic acid and hollow symbol for H2O2.

Fig. 3. Absorption spectra of H2O2 and humic acid

(Nacalai Tesque) in water. (Ð) H2O2; (. . .) Humic acid;

inhibiting agent of humic acid oxidation. In ad-dition, H2O2 itself absorb lights in the system and hence the light intensity available for humic acids is reduced at higher H2O2 concentrations. As can be deduced from Fig. 5, when the concentration of H2O2 is higher than 0.01% its OH radical scaven-ging e€ect becomes of great importance and the humic acid oxidation rate decreases. It has to be noted, however, that elimination of humic acid due to hydroxyl radicals continues to be the main path-way even at the high H2O2 concentration of 0.5% for which the oxidation rate is still higher than that obtained from direct photolysis (Fig. 5, [H2O2Š0ˆ 0).

E€ect of initial humic acid concentration

At an H2O2 concentration of 0.1% (29.4 mM), the initial rate constants for humic acid degradation were similar with the initial humic acid concen-tration of 3 and 5 mg/l but slightly decreased with the initial humic-acid concentration of 8 mg/l (Fig. 6). At H2O2 concentrations of 0.3 and 0.5%, the initial rate constants of humic acids showed no signi®cant di€erence among 3, 5, and 8 mg/l of in-itial humic acid concentration. At short irradiation times, where humic acids and hydrogen peroxide can be considered the only e€ective OH radical sca-vengers, a steady-state kinetic analysis can be con-sidered on the basis of the following simple reaction scheme (Brezonik and Fulkerson-Brekken, 1998):

H2O2‡ hn42
OH fOHFG0=V …1†

OH ‡ humic acid4humic acid radical ‡ H2O

kHAˆ 2:3  104…mg of C=l†ÿ1 sÿ1

…2†
OH ‡ H2O24HO2
‡ H2O

kH2O2ˆ 2:7  107 Mÿ1 sÿ1

…3† where fOH is the primary quantum yield if
OH

radical generation, F is the fraction of UV light absorbed by H2O2, G0 is the total incident UV is the photon ¯ux and V is the total irradiated volume. In the steady-state approximation, d‰
OH Š=dt ˆ 0, and the ®nal kinetic expression ex-trapolated to time t = 0 can be written as:

ÿd‰HAŠ_dt 

tˆ0ˆ

kHA‰HAŠ0fOHFG0=V

kHA‰HAŠ0‡ kH2O2‰H2O2Š0

…4† where fÿd‰HAŠ=dtg jtˆ0 denotes the initial rate of

humic acid degradation. The concentration of H2O2 is much larger than that of humic acid in this study; therefore equation (4) can be simpli®ed to fÿd‰HAŠ=dtg jtˆ0ˆ K‰HAŠ0 when the H2O2 concen-tration and the light intensity are constant.

From Fig. 6 the initial humic acid concentration did not a€ect the apparent ®rst order rate constant, therefore the rate of humic acid destruction appears to be truly ®rst order in the range of 3±8 mg/l NPOC. However, at higher humic acids concen-tration, it should be noted that the scavenging e€ect of humic acids may in¯uence the initial rate con-stant of itself when the concentration of H2O2 is low (Liao and Gurol, 1995), as shown in Fig. 6 (top).

E€ect of bicarbonate/carbonate concentration The e€ect of HCOÿ

3/COÿ23 on the degradation of humic acid in the UV/ H2O2process was examined with the initial carbonate concentration of 100± 400 mg/l as CaCO3 (Fig. 7). A 22 and 70% re-duction of the initial rate constants for humic acid removal (0.029 and 0.011 minÿ1_{, respectively) were}

Fig. 6. In¯uence of initial humic acid concentration in UV/H2O2systems. (*) [humic acid]0=3 mg/l; (Q) [humic

acid]0=5 mg/l; (R) [humic acid]0=8 mg/l; UV source:

UV/quartz. Fig. 5. Rate constant for humic acid degradation with

var-ious H2O2 concentrations. [Humic acid]0=5 mg/l; UV

observed with initial bicarbonate and carbonate concentrations of 96 and 124 mg/l.

In reality, there may be a signi®cant amount of HCOÿ

3=CO3ÿ2 in natural water and it may compete

with organic matter and peroxide for reaction with OH radicals. Equations (5) and (6) indicate that an increase of bicarbonate/carbonate concentration will lower the OH radical concentration, so the destruction rate of target organics will decrease with an increase in bicarbonate/carbonate concen-tration. The e€ect of bicarbonate/carbonate on the oxidation rate will be signi®cant, particularly when the concentration of the target organic is low (Ligrini et al., 1993; Buxton et al., 1988; Staehelin and Hoigne, 1985): HCOÿ 3 ‡
OH4CO3ÿ
‡ H2O k ˆ 1:5  107 _Mÿ1 _sÿ1_† …5† CO2ÿ 3 ‡
OH4CO3ÿ
‡ OHÿ k ˆ 4:2  108 _Mÿ1 _sÿ1_† …6†

BeltraÂn et al. (1996) reported that the presence of bicarbonate had no e€ect on the direct photolysis of deethylatrazine and deisopropylatrazine, but it caused a signi®cant inhibition of oxidation in the UV/H2O2 system. The hydroxyl radical scavenging character of bicarbonate and carbonate (equations (5) and (6)) can explain these inhibition e€ects. Although bicarbonate and carbonate do not adsorb UV light, they react readily with hydroxyl radicals (Buxton et al., 1988; Staehelin and Hoigne, 1985) which are the primary oxidizing species in the UV/ H2O2 process. Thus OH radical scavenging is suggested to account for the observed inhibition e€ect of bicarbonate and carbonate. Although the generated carbonate radical anion has been shown

to be an oxidant itself, its oxidation potential is less positive than that of the OH radical (Ligrini et al., 1993).

Glaze et al (1995) and Liao and Gurol (1995) stu-died the e€ect of bicarbonate (0.1±2 mM) on ben-zene (initial concentration 0.64 mM) decomposition at a peroxide concentration of 3 mM: It was found that bicarbonate a€ects benzene decomposition insigni®cantly in the range of 0.5±2 mM. The lack of an e€ect of bicarbonate is due to the fact that the concentration of benzene is relatively high and the reaction rate constant of benzene with an OH radical is several hundred times greater than that of bicarbonate. The e€ect of bicarbonate/carbonate can be calculated approximately using the rate con-stant values of OH radicals with humic acid, per-oxide, and bicarbonate/carbonate ion.

CONCLUSIONS

The rate of humic acid oxidation is greatly increased in the combined UV=H2O2system but the

presence of bicarbonate/carbonate species has a negative e€ect due to the scavenging of hydroxyl radicals, especially when its concentration is high. At the experimental conditions evaluated in this study, the humic-acid decay follows ®rst-order kin-etics with a rate constant of 0.037 minÿ1_{when the} UV/quartz was used as the light source, while that of hydrogen peroxide obeys zero-order kinetics at 0.30 mM minÿ1_{. Humic acid decomposed at a} much slower rate with UV/Pyrex (0.0025 minÿ1₎ and sunlight (0.0007 minÿ1_{), which have minor} UVC emission and comparable total UV light intensity compared to UV/quartz. Hydrogen per-oxide acts as both an initiating and scavenging agent of hydroxyl radicals. The former e€ect predo-minates when the initial hydrogen peroxide concen-tration is lower than 0.01% (2.94 mM), the latter at higher concentrations. However, even at high H2O2 concentrations, oxidation of humic acids is due to hydroxyl radical attack because H2O2absorbs most of the light.

AcknowledgementsÐThis research was supported by National Science Council, Taiwan, Republic of China, under Grant No. NSC 86-2211-E002-015. The authors thanks the reviewers for valuable comments on the prep-aration of this paper.

REFERENCES

Atkins P. W. (1990) Physical Chemistry, 4th ed. Oxford University Press, p. 938.

Baxendale J. H. and Wilson J. A. (1957) The photolysis of hydrogen peroxide at high light intensities. Trans. Fara-day Soc. 53, 344.

BeltraÂn F. J., Ovejero G. and Acedo B. (1993) Oxidation of atrazine in water by UV radiation combined with H2O2. Wat. Res. 27(6), 1013±1021.

BeltraÂn F. J., GonzaÂlez M., Rivas F. J. and Alvarez P. (1996) Aqueous UV radiation and UV/TiO2 oxidation

Fig. 7. Degradation of humic acid in UV/H2O2 systems

with various HCOÿ

3 and COÿ23 concentrations. [humic

acid]0=5 mg/l; [H2O2]0=0.005%; UV source: UV/quartz;

[HCOÿ

3/COÿ23 ]: (*)=0, 0 mg/l; (T) 96, 124 mg/l; (Q) 190,

of atrazine ®rst degradation products: Deethylatrazine and deisopropylatrazine. Environ. Toxicol. Chem. 15, 868±872.

BeltraÂn F. J., GonzaÂlez M. and GonzaÂlez J. F. (1997) Industrial wastewater advanced oxidation. Part I. UV radiation in the presence and absence of hydrogen per-oxide. Wat. Res. 31(10), 2405±2414.

Brezonik P. L. and Fulkerson-Brekken J. (1998) Nitrate-induced photolysis in natural waters: controls on con-centration of hydroxyl radical photo-intermediates by natural scavenging agents. Envir. Sci. and Tech. 32, 3004±3010.

Buxton G. V., Greenstock C. L., Helman W. P. and Ross A. B. (1988) Critical review of rate constants for reac-tions of hydrated electrons, hydrogen atoms and hy-droxyl radicals in aqueous solution. J. Phys. Chem. Ref. Data 17(2), 513±886.

Eggins B. R., Palmer F. L. and Byrne J. A. (1997) Photo-catalytic treatment of humic substances in drinking water. Water Res. 31(5), 1223±1226.

Glaze W. H., Lay Y. and Kang J. W. (1995) Advanced oxidation processes. A kinetic model for the oxidation of 1,2-dibromo-3-chloropropane in water by the combi-nation of hydrogen peroxide and UV radiation. Ind. Eng. Chem. Res. 34(7), 2314±2323.

Hu S. T. and Yu Y. H. (1994) Preozonation of chlorophe-nolic wastewater for subsequent biological treatment. Ozone Sci. Eng. 16(1), 13±28.

Hunt J. P. and Taube H. (1952) The photochemical de-composition of hydrogen peroxideÐquantum yields, tracer and fractionation e€ects. J. Amer. Chem. Soc. 74(5), 5999±6002.

Jacangelo J. G., DeMarco J., Owen D. M. and Randtke S. J. (1995) Selected processes for removing NOM. J. AWWA 87(1), 64±77.

Kang J. W. and Lee K. H. (1997) A kinetic model of the H2O2/UV process for the treatment of hazardous waste

chemicals. Environ. Engi. Sci. 14(3), 183±192.

Ku Y., Wang L. S. and Shen Y. S. (1998) Decomposition of EDTA in aqueous solution by UV/H2O2 process. J.

Hazard. Mat. 60, 41±45.

Liao C. H. and Gurol M. D. (1995) Chemical oxidation by photolytic decomposition of hydrogen peroxide. Env. Sci. Tech. 29(12), 3007±3014.

Ligrini O., Oliveros E. and Bruan A. M. (1993) Photoche-mical processes for water treatment. Chem. Rev. 93, 671±698.

Owen D. M., Amy G. L., Chowdhury Z. K., Paode R., McCoy G. and Viscosi K. (1995) NOM characterization and treatability. J. AWWA 87(1), 46±63.

Staehelin J. and Hoigne J. (1985) Decomposition of ozone in water in the presence of organic solutes acting as pro-motors and inhibitors of radical chain reactions. Environ. Sci. Technol. 19(12), 1206±1213.

Sundstrom D. W., Klei H. E., Nalette T. A., Reidy D. J. and Weir B. A. (1986) Destruction of halogenated ali-phatics by UV catalyzed oxidation with H2O2. Haz.

Waste and Haz. Mater. 3(1), 101±110.

Symons J. M. and Worley K. L. (1995) An advanced oxidation process for DBP control. J. AWWA 87(11), 66.

Volman D. H. (1949) The photochemical decomposition of hydrogen peroxide. J. Chem. Phys. 17, 949.

Wang G. S., Hsieh S. T. and Chiou K. Y. (1997) Catalytic photochemical oxidation of NOM in drinking water. In Proceedings 1997 AWWA Water Quality Technology Conference. AWWA, Denver, CO.

Weeks J. L. and Matheson M. S. (1956) The primary quantum yield of hydrogen peroxide decomposition. J. Amer. Chem. Soc. 78(1), 1273±1279.