Tradeoff between micropollutant abatement and bromate formation during ozonation of concentrates from nanofiltration and reverse osmosis processes.

Water treatment with nanofiltration (NF) or reverse osmosis (RO) membranes results in a purified permeate and a retentate, where solutes are concentrated and have to be properly managed and discharged. To date, little is known on how the selection of a semi-permeable dense membrane impacts the dissolved organic matter in the concentrate and what the consequences are for micropollutant (MP) abatement and bromate formation during concentrate treatment with ozone. Laboratory ozonation experiments were performed with standardized concentrates produced by three membranes (two NFs and one low-pressure reverse osmosis (LPRO) membrane) from three water sources (two river waters and one lake water). The concentrates were standardized by adjustment of pH and concentrations of dissolved organic carbon, total inorganic carbon, selected micropollutants (MP) with a low to high ozone reactivity and bromide to exclude factors which are known to impact ozonation. NF membranes had a lower retention of bromide and MPs than the LPRO membrane, and if the permeate quality of the NF membrane meets the requirements, the selection of this membrane type is beneficial due to the lower bromate formation risks upon concentrate ozonation. The bromate formation was typically higher in standardized concentrates of LPRO than of NF membranes, but the tradeoff between MP abatement and bromate formation upon ozonation of the standardized concentrates was not affected by the membrane type. Furthermore, there was no difference for the different source waters. Overall, ozonation of concentrates is only feasible for abatement of MPs with a high to moderate ozone reactivity with limited bromate formation. Differences in the DOM composition between NF and LPRO membrane concentrates are less relevant than retention of MPs and bromide by the membrane and the required ozone dose to meet a treatment target.

Effect of cetyltrimethylammonium chloride on various Escherichia coli strains and their inactivation kinetics by ozone and monochloramine.

Cethyltrimethylammonium chloride (CTMA) is one of the most used quaternary ammonium compounds (QACs) in consumer products. CTMA and other QACs are only partially eliminated in municipal wastewater treatment and they can interact with bacteria in biological processes. Currently, there is only limited information on the antimicrobial efficiency of CTMA in matrices other than standard growth media and if and how CTMA influences conventional chemical disinfection. The results obtained in this study showed that the susceptibility of E. coli to CTMA was significantly enhanced in phosphate-buffered saline, lake water and wastewater compared to broth. In broth, a minimum inhibitory concentration (MIC) of CTMA of 20 mgL-1 was observed for E. coli, whereas a 4-log inactivation occurred for CTMA concentrations of about 4 mgL-1 in buffered ultra-purified water, a lake water and wastewater effluent. The impacts of the pre-exposure and the presence of CTMA on inactivation by ozone and monochloramine were tested with three different E. coli strains: AG100 with the efflux pump acrAB intact, AG100A with it deleted and AG100tet with it overexpressed. Pre-exposure of E. coli AG100 to CTMA led to an increased susceptibility for ozone with second-order inactivation rate constants (∼ 106 M-1s-1) increasing by a factor of about 1.5. An opposite trend was observed for monochloramine with second-order inactivation rate constants (∼ 103 M-1s-1) decreasing by a factor of about 2. For E. coli AG100tet, the second-order inactivation rate constant decreased by a factor of almost 2 and increased by a factor of about 1.5 for ozone and monochloramine, respectively, relative to the strain AG100. The simultaneous presence of CTMA and ozone enhanced the second-order inactivation rate constants for CTMA concentrations of 2.5 mgL-1 by a factor of about 3. For monochloramine also an enhancement of the inactivation was observed, which was at least additive but might also be synergistic. Enhancement by factors from about 2 to 4.5 were observed for CTMA concentrations > 2.5 mgL-1.

Halide removal from waters by silver nanoparticles and hydrogen peroxide.

The objective of this study was to remove halides from waters by silver nanoparticles (AgNPs) and hydrogen peroxide (H2O2). The experimental parameters were optimized and the mechanism involved was also determined. The AgNP/H2O2 process proved efficacious for bromide and chloride removal from water through the selective precipitation of AgCl and AgBr on the AgNP surface. The optimal AgNP and H2O2 concentrations to be added to the solution were determined for the halide concentrations under study. The removal of Cl- and Br- anions was more effective at basic pH, reaching values of up to 100% for both ions. The formation of OH, O2-, radicals was detected during the oxidation of Ag(0) into Ag(I), determining the reaction mechanism as a function of solution pH. Moreover, the results obtained show that: i) the efficacy of the oxidation of Ag(0) into Ag(I) is higher at pH11, ii) AgNPs can be generated by the O2- radical formation, and iii) the presence of NaCl and dissolved organic matter (tannic acid [TAN]) on the solution matrix reduces the efficacy of bromide removal from the medium due to: i) precipitation of AgCl on the AgNP surface, and ii) the radical scavenger capacity of TAN. AgNPs exhausted can be regenerated by using UV or solar light, and toxicity test results show that AgNPs inhibit luminescence of Vibrio fischeri bacteria.

Oxidation of manganese(II) during chlorination: role of bromide.

The oxidation of dissolved manganese(II) (Mn(II)) during chlorination is a relatively slow process which may lead to residual Mn(II) in treated drinking waters. Chemical Mn(II) oxidation is autocatalytic and consists of a homogeneous and a heterogeneous process; the oxidation of Mn(II) is mainly driven by the latter process. This study demonstrates that Mn(II) oxidation during chlorination is enhanced in bromide-containing waters by the formation of reactive bromine species (e.g., HOBr, BrCl, Br2O) from the oxidation of bromide by chlorine. During oxidation of Mn(II) by chlorine in bromide-containing waters, bromide is recycled and acts as a catalyst. For a chlorine dose of 1 mg/L and a bromide level as low as 10 µg/L, the oxidation of Mn(II) by reactive bromine species becomes the main pathway. It was demonstrated that the kinetics of the reaction are dominated by the adsorbed Mn(OH)2 species for both chlorine and bromine at circumneutral pH. Reactive bromine species such as Br2O and BrCl significantly influence the rate of manganese oxidation and may even outweigh the reactivity of HOBr. Reaction orders in [HOBr]tot were found to be 1.33 (±0.15) at pH 7.8 and increased to 1.97 (±0.17) at pH 8.2 consistent with an important contribution of Br2O which is second order in [HOBr]tot. These findings highlight the need to take bromide, and the subsequent reactive bromine species formed upon chlorination, into account to assess Mn(II) removal during water treatment with chlorine.

Ozonation of iodide-containing waters: selective oxidation of iodide to iodate with simultaneous minimization of bromate and I-THMs.

The presence of iodinated disinfection by-products (I-DBPs) in drinking water poses a potential health concern since it has been shown that I-DBPs are generally more genotoxic and cytotoxic than their chlorinated and brominated analogs. I-DBPs are formed during oxidation/disinfection of iodide-containing waters by reaction of the transient hypoiodous acid (HOI) with natural organic matter (NOM). In this study, we demonstrate that ozone pre-treatment selectively oxidizes iodide to iodate and avoids the formation of I-DBPs. Iodate is non-toxic and is therefore a desired sink of iodine in drinking water. Complete conversion of iodide to iodate while minimizing the bromate formation to below the guideline value of 10 µg L⁻¹ was achieved for a wide range of ozone doses in five raw waters with DOC and bromide concentrations of 1.1-20 mg L⁻¹ and 170-940 µg L⁻¹, respectively. Lowering the pH effectively further reduced bromate formation but had no impact on the extent of iodate and bromoform formation (the main trihalomethane (THM) formed during ozonation). Experiments carried out with pre-chlorinated/post-clarified samples already containing I-DBPs, showed that ozonation effectively oxidized I-THMs. Therefore, in iodide-containing waters, in which I-DBPs can be produced upon chlorination or especially chloramination, a pre-ozonation step to oxidize iodide to iodate is an efficient process to mitigate I-DBP formation.

The basics of oxidants in water treatment. Part B: ozone reactions.

The oxidation of organic and inorganic compounds during ozonation can occur via ozone or OH radicals or a combination thereof. Ozone is an electrophile with a high selectivity. The reactions of ozone with inorganic compounds are typically fast and occur by an oxygen atom transfer reaction. Organic micropollutants are oxidised with ozone selectively. Ozone reacts mainly with double bonds, activated aromatic systems and non-protonated amines. The kinetics of direct ozone reactions depend strongly on the speciation (acid-base, metal complexation). The reaction of OH radicals with the majority of inorganic and organic compounds is nearly diffusion-controlled. The degree of oxidation by ozone and OH radicals is given by the corresponding kinetics and the ratio of the concentration of the two oxidants. Product formation from the ozonation of organic micropollutants has only been established for a few compounds. Numerous organic and inorganic ozonation disinfection/oxidation byproducts have been identified. The byproduct of main concern is bromate, which is formed in bromide-containing waters. A low drinking water standard of 10 microgL(-1) has been set for bromate. In certain cases (bromide > approximately 50 microgL(-1)), it may be necessary to use control measures to lower bromate formation (lowering of pH, ammonia addition, chlorination-ammonia process).

Photooxidation of naphthalenesulfonic acids: comparison between processes based on O(3), O(3)/activated carbon and UV/H(2)O(2).

The aim of the present study was to analyze and compare the efficacy of UV photodegradation with that of different advanced oxidation processes (O(3), UV/H(2)O(2), O(3)/activated carbon) in the degradation of naphthalenesulfonic acids from aqueous solution and to investigate the kinetics and the mechanism involved in these processes. Results obtained showed that photodegradation with UV radiation (254 nm) of 1-naphthalenesulfonic, 1,5-naphthalendisulfonic and 1,3,6-naphthalentrisulfonic acids is not effective. Presence of duroquinone and 4-carboxybenzophenone during UV irradiation (308-410 nm) of the naphthalenesulfonic acids increased the photodegradation rate. Addition of H(2)O(2) during irradiation of naphthalenesulfonic acids accelerated their elimination, due to the generation of ()OH radicals in the medium. Comparison between UV photodegradation 254 m and the advanced oxidation processes (O(3), O(3)/activated carbon and UV/H(2)O(2)) showed the low-efficacy of the former in the degradation of these compounds from aqueous medium. Thus, among the systems studied, those based on the use of UV/H(2)O(2) and O(3)/activated carbon were the most effective in the oxidation of these contaminants from the medium. This is because of the high-reactivity of naphthalenesulfonic acids with the *OH radicals generated by these two systems. This was confirmed by the values of the reaction rate constant of *OH radicals with these compounds k(OH), obtained by competitive kinetics (5.7 x 10(9) M(-1) s(-1), 5.2 x 10(9) M(-1) s(-1) and 3.7 x 10(9) M(-1) s(-1) for NS, NDS and NTS, respectively).

Understanding medicinal taste and odour formation in drinking waters.

The formation of bromophenols during chlorination of phenol- and bromide-containing waters can be critical for taste and odour problems in drinking waters. The work performed has confirmed that flavour threshold concentrations of some bromophenols are in the ng/L range. In addition, under typical drinking water conditions, kinetic experiments and model simulations performed have shown that (1) bromination is the predominant reaction pathway, (2) bromophenol reaction kinetics are rapid leading to taste-and-odour episodes that last for short periods of 10-20 min, (3) increasing phenol concentration and pH tends to increase taste and odour intensity, (4) increasing chlorine or bromide concentrations tends to shorten the duration of the taste-and-odour episode.

Ag-doped carbon aerogels for removing halide ions in water treatment.

The objective of this study was to analyze the efficiency of silver(Ag)-doped carbon aerogels for the removal of bromide (Br(-)) and iodide (I(-)) from drinking waters. Textural characterization of Ag-doped aerogels showed that an increase in the Ag dose added during the preparation process produced: (i) a reduction in the surface area (S(BET)) and (ii) an increase in mesopore (V(2)) and macropore (V(3)) volumes. Chemical characterization of the materials revealed an acidic surface (pH of point of zero charge, pH(PZC)=4.5, O(surface)=20%). The oxidation state of Ag was +1 and the surface concentration of this element ranged from 4% to 10%. The adsorption capacity (X(m)) and affinity of adsorbent (BX(m)) increased with a reduction in the radius of the halogenide. Furthermore, an increase in the adsorption capacity was observed with higher Ag concentrations on the aerogel surface. The high adsorption capacity of the aerogel may be due to the presence of Ag(I) on its surface, with the formation of the corresponding Ag halides. Our observations indicate that the halogenides adsorption on commercial activated carbon (Sorbo-Norit) is much lower than that of the Ag-doped carbon aerogels. The presence of chloride and natural organic matter (NOM) in the medium reduced the adsorption capacity of Br(-) and I(-) on Ag carbon aerogels.

Bromide and iodide removal from waters under dynamic conditions by Ag-doped aerogels.

The objective of this study was to analyze the efficiency of Ag-doped aerogels in the removal of bromide and iodide from water. To test the applicability of these aerogels in water treatment, adsorption of bromide and iodide was studied under dynamic conditions using waters from Lake Zurich and a mineral water. The results obtained by using these waters showed a high breakthrough volume (V(0.02)=0.4 L) of the columns, while the height of the mass transfer zone (H(MTZ)=6.8 cm) was low, regardless of the anion under study. Bromide- and iodide-saturated columns were regenerated with NH4OH. No change in the column characteristics was observed after two regeneration treatments, regardless of the type of water considered.

Metal-doped carbon aerogels as catalysts during ozonation processes in aqueous solutions.

The efficiency of Co(II)-, Mn(II)-, and Ti(IV)-doped carbon aerogels for the transformation of ozone into (*)OH radicals was investigated. The carbon aerogels had a markedly acid surface character (pH(PZC approximately equal) congruent with 3-4) with very high surface oxygen concentrations (O approximately equal with 20%). X-ray photoelectron spectroscopy (XPS) analyses of the samples showed the oxidation state of the metals was +2 for Co and Mn and +4 for Ti. The presence of Mn(II)-doped carbon aerogel enhanced ozone transformation into (*)OH radicals, whereas the presence of Co(II) and Ti(IV) carbon aerogels presented no activity in this process. Moreover, it was observed that an increase in the concentration of Mn in the surface of the aerogel increases its efficiency to transform ozone into (*)OH radicals, with an Rct value ([OH]/[O(3)]) of 5.36 x 10(-8) for the aerogel doped with 16% of surface Mn(II) compared to an R(ct) of 2.68 x 10(-9) for conventional ozonation. Regardless of the aerogel used, XPS analysis of the ozonated aerogel samples showed an increase in the concentration of surface oxygen when the exposure to ozone was longer. However, presence of oxidized metal species after ozone treatment was only detected in the case of the Mn-doped aerogel, (Mn(III) and Mn(IV)). CO(2) activation of carbon aerogel produced a marked increase in its efficiency to transform ozone into (*)OH radicals compared with non-activated sample. The efficiency of Mn activated carbon aerogel to transform ozone into (*)OH radicals was greater than that of Witco commercial activated carbon or H(2)O(2) in the ozonation of water from Lake Zurich (Zurich, Switzerland).

Removal of bromide and iodide anions from drinking water by silver-activated carbon aerogels.

The aim of this study is to analyze the use of Ag-doped activated carbon aerogels for bromide and iodide removal from drinking water and to study how the activation of Ag-doped aerogels affects their behavior. It has been observed that the carbonization treatment and activation process of Ag-doped aerogels increased the surface area value ( [Formula: see text] ), whereas the volume of meso-(V(2)) and macropores (V(3)) decreased slightly. Chemical characterization of the materials revealed that carbonization and especially activation process considerably increased the surface basicity of the sample. Original sample (A) presented acidic surface properties (pH(PZC)=4.5) with 21% surface oxygen, whereas the sample that underwent activation showed mainly basic surface chemical properties (pH(PZC)=9.5) with only 6% of surface oxygen. Carbonization and especially, activation process considerable increased the adsorption capacity of bromide and iodide ions. This would mainly be produced by (i) an increase in the microporosity of the sample, which increases Ag-adsorption sites available to halide anions, and (ii) a rise of the basicity of the sample, which produces an increase in attractive electrostatic interactions between the aerogel surface, positively charged at the working pH (pH(solution)

The impact of selected water quality parameters on the inactivation of Bacillus subtilis spores by monochloramine and ozone.

Selected water quality parameters-pH, dissolved organic carbon, turbidity (NTU), and temperature-were tested for their potential effects on ozone and monochloramine inactivation of Bacillus subtilis spores. In oxidant demand-free phosphate-buffer, temperature had the strongest influence on inactivation kinetics when using ozone, pH had a smaller but significant impact on B. subtilis spore inactivation with both monochloramine and ozone. Where monochloramine was applied, modeling and experimental measurements confirmed that dichloramine levels were too low to produce significant inactivation effects under these experimental conditions. It was demonstrated that oxidant demand-free phosphate buffer may not be an adequate environmental analogue for inactivation responses in natural waters.

Efficiency of activated carbon to transform ozone into *OH radicals: influence of operational parameters.

Based on previous findings (Jans, U., Hoigné, J., 1998. Ozone Sci. Eng. 20, 67-87), the activity of activated carbon for the transformation of ozone into *OH radicals including the influence of operational parameters (carbon dose, ozone dose, carbon-type and carbon treatment time) was quantified. The ozone decomposition constant (k(D)) was increased by the presence of activated carbon in the system and depending on the type of activated carbon added, the ratio of the concentrations of *OH radicals and ozone, the R(ct) value ([*OH]/[O3]), was increased by a factor 3-5. The results obtained show that the surface chemical and textural characteristics of the activated carbon determines its activity for the transformation of ozone into *OH radicals. The most efficient carbons in this process are those with high basicity and large surface area. The obtained results show that the interaction between ozone and pyrrol groups present on the surface of activated carbon increase the concentration of O2*- radicals in the system, enhancing ozone transformation into *OH radicals. The activity of activated carbon decreases for extended ozone exposures. This may indicate that activated carbon does not really act as a catalyst but rather as a conventional initiator or promoter for the ozone transformation into *OH radicals. Ozonation of Lake Zurich water ([O3] = 1 mg/L) in presence of activated carbon (0.5 g/L) lead to an increase in the k(D) and R(ct) value by a factor of 10 and 39, respectively, thereby favouring the removal of ozone-resistant contaminants. Moreover, the presence of activated carbon during ozonation of Lake Zurich water led to a 40% reduction in the content of dissolved organic carbon during the first 60 min of treatment. The adsorption of low concentrations of dissolved organic matter (DOM) on activated carbon surfaces did not modify its capacity to initiate/promote ozone transformation into *OH radicals.

Characterization of iron and manganese precipitates from an in situ ground water treatment plant.

Aquifer samples from the precipitation zone of an in situ iron and manganese removal plant that was operated for 10 years were analyzed for iron and manganese minerals. Measurements were performed by various chemical extraction techniques (5 M HCI, 0.008 M Ti(III)-EDTA, 0.114 M ascorbic acid), X-ray diffraction and Mössbauer spectroscopy. Chemical extractions showed that iron was precipitated as ferric oxides, whereas manganese was not oxidized but deposited as Mn(II) probably within carbonates. The ferric oxides in particular accumulate preferentially in the smaller grain- size fractions. This tendency was observed to a lesser extent for manganese. X-ray diffraction and Mössbauer spectroscopy showed that the ferric oxides were mainly crystalline (goethite, 50% to 100% of the iron). Ferrihydrite was found as well, but only as a minor fraction (< or = 12%). Pure manganese minerals were not found by X-ray diffraction. The precipitated amounts of iron (5 to 27 micromol/g Fe as ferric oxide) and manganese (1 to 4 micromol/g Mn) during 10 years operation of the treatment plant agree with values that were estimated from operational parameters (9 to 31 micromol/g Fe and 3 to 6 micromol/g Mn). Considering the small amounts of precipitated iron and manganese, no long-term risks of clogging of the aquifer are expected.

MTBE oxidation by conventional ozonation and the combination ozone/hydrogen peroxide: efficiency of the processes and bromate formation.

The present study investigates the oxidation of methyl tert-butyl ether (MTBE) by conventional ozonation and the advanced oxidation process (AOP) ozone/hydrogen peroxide under drinking water treatment conditions. The major degradation products identified were tert-butyl formate (TBF), tert-butyl alcohol (TBA), 2-methoxy-2-methyl propionaldehyde (MMP), acetone (AC), methyl acetate (MA), hydroxyisobutyraldehyde (HiBA), and formaldehyde (FA). The rate constants of the reaction of ozone and OH radicals with MTBE were found to be 0.14 and 1.9 x 10(9) M(-1) s(-1), respectively. The rate constants for the same oxidation processes were also measured for the degradation products TBF, MMP, MA, and HiBA (k(O3-TBF) = 0.78 M(-1) s(-1); k(OH-TBF) = 7.0 x 10(8) M(-1) s(-1); k(O3-MMP) = 5 M(-1) s(-1); k(OH-MMP) = 3 x 10(9) M(-1) s(-1), k(O3-MA) = 0.09 M(-1) s(-1), k(O3-HiBA) = 5 M(-1) s(-1); k(OH-HiBA) = 3 x 10(9) M(-1) s(-1)). Since all compounds reacted slowly with molecular ozone, only the degradation pathway of MTBE with OH radicals has been determined, including the formation of primary degradation products. In experiments performed with several natural waters, the efficiency of MTBE elimination and the formation of bromate as disinfection byproduct have been measured. With a bromide level of 50 microg/L, only 35-50% of MTBE could be eliminated by the AOP O3/H2O2 without exceeding the current drinking water standard of bromate (10 microg/L). The transient concentrations of MTBE and its primary degradation products were modeled using a combination of kinetic parameters (degradation product distribution and rate constants) together with the ozone and OH radical concentration and were in good agreement with the experimental results.

Inactivation of Bacillus subtilis spores and formation of bromate during ozonation.

Inactivation of B. subtilis spores with ozone was investigated to assess the effect of pH and temperature, to compare the kinetics to those for the inactivation of C. parvum oocysts, to investigate bromate formation under 2-log inactivation conditions, and to assess the need for bromate control strategies. The rate of B. subtilis inactivation with ozone was independent of pH, decreased with temperature (activation energy of 42,100 Jmol(-1)), and was consistent with the CT concept. B. subtilis was found to be a good indicator for C. parvum at 20-30 degrees C, but at lower temperatures B. subtilis was inactivated more readily than C. parvum. Bromate formation increased as both pH and temperature increased. For water with an initial bromide concentration of 33 microgl(-1), achieving 2-logs of inactivation, without exceeding the 100 microg l(-1) bromate standard, was most difficult at 30 degrees C for B. subtilis and at midrange temperatures (10-20 degrees C) for C. partum. pH depression and ammonia addition were found to reduce bromate formation without affecting B. subtilis inactivation, and may be necessary for waters containing more than 50 microgl(-1) bromide.

Bromate minimization during ozonation: mechanistic considerations.

Bromate formation during ozonation of bromide-containing natural waters is somewhat inversely connected to the ozone characteristics: an initial fast increase followed by a slower formation rate. During the initial phase mostly OH radical reactions contribute to bromate formation,whereas in the secondary phase both ozone and OH radicals are important. To minimize bromate formation several control options are presented: ammonia addition, pH depression, OH radical scavenging, and scavenging or reduction of hypobromous acid (HOBr) by organic compounds. Only the two first options are applicable in drinking watertreatment. By both methods a similar effect of a bromate reduction of approximately 50% can be achieved. However, bromate formation during the initial phase of the ozonation cannot be influenced by either method. Ammonia (NH3) efficiently scavenges HOBrto NH2Br. However, this reaction is reversible which leads to higher required NH3 concentrations than expected. The rate constant kNH2Br for the hydrolysis of NH2Br by OH- to NH3 and OBr- was found to be 7.5-10(6) M(-1) s(-1). pH depression shifts the HOBr/ OBr- equilibrium to HOBr and also affects the ozone chemistry. The effect on ozone chemistry was found to be more importantfor bromate formation. For a given ozone exposure, the OH radical exposure decreases with decreasing pH. Therefore, for pH depression the overall oxidation capacity for a certain ozone exposure decreases which in turn leads to a smaller bromate formation.

Solar oxidation and removal of arsenic at circumneutral pH in iron containing waters.

An estimated 30-50 million people in Bangladesh consume groundwater with arsenic contents far above accepted limits. A better understanding of arsenic redox kinetics and simple water treatment procedures are urgently needed. We have studied thermal and photochemical As(III) oxidation in the laboratory, on a time scale of hours, in water containing 500 micrograms/L As(III), 0.06-5 mg/L Fe(II,III), and 4-6 mM bicarbonate at pH 6.5-8.0. As(V) was measured colorimetrically, and As(III) and As(tot) were measured by As(III)/As(tot)-specific hydride-generation AAS. Dissolved oxygen and micromolar hydrogen peroxide did not oxidize As(III) on a time scale of hours. As(III) was partly oxidized in the dark by addition of Fe(II) to aerated water, presumably by reactive intermediates formed in the reduction of oxygen by Fe(II). In solutions containing 0.06-5 mg/L Fe(II,III), over 90% of As(III) could be oxidized photochemically within 2-3 h by illumination with 90 W/m2 UV-A light. Citrate, by forming Fe(III) citrate complexes that are photolyzed with high quantum yields, strongly accelerated As(III) oxidation. The photoproduct of citrate (3-oxoglutaric acid) induced rapid flocculation and precipitation of Fe(III). In laboratory tests, 80-90% of total arsenic was removed after addition of 50 microM citrate or 100-200 microL (4-8 drops) of lemon juice/L, illumination for 2-3 h, and precipitation. The same procedure was able to remove 45-78% of total arsenic in first field trials in Bangladesh.