Environ. Sci. Technol. 2001, 35, 2114-2121
Solar Oxidation and Removal of Arsenic at Circumneutral pH in Iron Containing Waters STEPHAN J. HUG,* LAURA CANONICA, MARTIN WEGELIN, DANIEL GECHTER, AND URS VON GUNTEN Swiss Federal Institute for Environmental Science and Technology (EAWAG), Postfach 611, U ¨ berlandstrasse 133, CH-8600 Du ¨ bendorf, Switzerland
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 µg/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 hydrogenperoxide 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 µM citrate or 100-200 µL (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.
Introduction The current health crisis in Bangladesh caused by arsenic containing groundwater calls for the rapid development of emergency and long-term solutions. Tens of thousands of people already show symptoms of arsenic poisoning. The current limiting value for arsenic in drinking water in Bangladesh and in most countries is 50 µg/L, the new European Community and WHO recommended guide line value is 10 µg/L, and the US-EPA recently proposed a maximum contaminant level of 5 µg/L. In a recent survey in Bangladesh, 35% of 2022 samples representative of the over 3 million hand pumped tubewells were above 50 µg/L, 25% were above 100 µg/L, and 8.4% were above 300 µg/L (1). On the basis of available data and statistics, it is estimated that around 40 million people are at risk of chronic arsenic poisoning (2). * Corresponding author phone: +41-1-823-5454; fax: +41-1823-5210; E-mail:
[email protected]. 2114
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 10, 2001
The source of the arsenic in Bangladesh is geogenic (1, 3, 4). According to the now most accepted hypothesis, As is released from alluvial sediments to anoxic groundwater by reductive dissolution of As-containing Fe(III)(hydr)oxides. Reduction is driven by sedimentary organic matter in concentrations of up to 6% (3-5). Hand pumped well water can typically contain 0-20 mg Fe(II)/L and 0 to >1000 µg/L As(III), often with a correlation between As and Fe (1, 3). At the groundwater pH values of 6.5-8.0, Fe(II) is oxidized within tens of minutes to seconds upon exposure to air and precipitates in the form of Fe(III)(hydr)oxides (6). A fraction of As(III) (H3AsO3) remains unoxidized for many days (7) and adsorbs only weakly on Fe(III)(hydr)oxides (8). In contrast, As(V) (H2AsO4- and HAsO42-) adsorbs more strongly [forming bidentate and monodentate surface complexes (911)] and can be removed almost quantitatively on the precipitated Fe(III)(hydr)oxides (8, 12, 13) at circumneutral pH. Many conventional arsenic removal procedures involve an oxidation step, followed by an adsorption step. As(III) oxidation is carried out by oxidants such as chlorine, ozone, permanganate, or with MnO2 solid bed filters (14). Fe(II), Fe(III), or Al(III) are then added to adsorb and precipitate As(V) with the forming Fe(III)- or Al(III)(hydr)oxides (8, 14). In rural areas of Bangladesh, larger water treatment plants that rely on costly equipment and chemicals are not a feasible option in the short term. Small scale and simple adapted water treatment with locally available materials have the best chance to improve the situation. Instead of adding a chemical oxidant, it would be advantageous to produce oxidants in situ, for example, by exposing the water to sunlight, followed by precipitation of the As(V) with the naturally present iron. Photochemical oxidation of As(III) with UV-light from highpressure Hg-lamps was found to be possible (12). It is well-known that reactive oxidants such as superoxide (O2•-/HO2•), hydrogenperoxide (H2O2), and hydroxyl radicals (•OH) are produced by photolysis of Fe(III)-complexes and subsequent dark reactions. The role of Fe(III) photochemistry in atmospheric and surface waters (15-19) and in advanced oxidation processes for water and wastewater treatment (2022) has been studied over the last two decades. In 1997, a patent was issued for an As-removal procedure that uses addition of Fe(II,III) to contaminated water, followed by exposure to UV or solar light (23). This procedure was initially developed to remove arsenic in acidic mining effluents. A paper with more detailed information and a discussion of reaction mechanisms at pH values between 0.5 and 2.5 has just been published upon receipt of the proofs for this manuscript (24). In circumneutral and carbonate containing waters, the speciation of Fe(II) and Fe(III) differs significantly from low pH conditions, and both Fe(II) oxidation and Fe(III)-photochemistry are expected to occur over different pathways and at different rates. The quantum yield for the production of •OH and Fe(II) from photolysis of FeIII(OH)2+ is 0.017 at 360 nm (25). Similar quantum yields for other inorganic Fe(III) complexes such as Fe(III)(OH)2+, Fe(OH)3, polynuclear Fe(III) complexes, and for solid iron phases have not been quantified to our knowledge but appear to be much lower. As the fraction of FeIII(OH)2+ becomes negligible above pH 5 and iron precipitates, Fe(III)-mediated photochemical degradation of pollutants becomes less efficient. In various studies addressing photochemical degradation of organic compounds with initially dissolved iron (26) or in suspensions of hematite (22) or lepidocrocite (27), degradation rates generally decreased with increasing pH. The useful pH range can be 10.1021/es001551s CCC: $20.00
2001 American Chemical Society Published on Web 04/05/2001
extended by addition of suitable organic ligands such as oxalate (26, 28), and very strongly complexing ligands such as EDTA are photodegraded at neutral pH even in suspensions of Fe(III) solid phases such as lepidocrocite (27). On the basis of the available evidence, photolysis of dissolved organic Fe(III) complexes or surface Fe(III) complexes provides the most efficient pathways in Fe(III)-mediated oxidation of pollutants at pH values between 5 and 8. The goal of our studies was to understand photoinduced As(III,V) redox reactions and to develop a simple solar arsenic removal procedure that works at circumneutral pH with locally available materials, without pH-adjustment and addition of chemicals. Arsenic(III) is to be photochemically oxidized by exposing the water to sunlight, after which As(V) is adsorbed and precipitated together with the Fe(III)(hydr)oxides formed from the naturally present iron. Poly(ethylene terephthalate) (PET) bottles have been successfully used to expose drinking water to sunlight for solar disinfection (29). Closed bottles have the advantage over open containers to protect the water from airborn dust and pathogens during treatment and they are readily available. In previous studies on Cr(VI) reduction, we found that citrate, a strongly Fe(III) complexing ligand, greatly improved photoproduction of Fe(II) at circumneutral pH (30). Reported quantum yields for the Fe(II) formation from photolysis of Fe(III)citrate are 0.280.21 at pH 4-6 with 436 nm light (16) and 0.4-0.2 at pH 5-7 at 366 nm (31). Citrate is readily available in Bangladesh in the form of lemon or lime juice, which contains between 5 and 10% of citric acid. Thus, addition of a few drops of lemon juice per liter of water might be an option to improve photooxidation of As(III) at circumneutral pH.
Experimental Section Chemicals. All chemicals, As2O3, NaH2AsO4‚2H2O, FeSO4‚7H2O, Na3Citrate‚2H2O, NaOH, H2SO4, CaCO3, MgCO3, NaHCO3, (NH4)6Mo7O24‚4H2O, ascorbic acid, K(SbO)C4H4O6‚ 0.5H2O (K-antimonyl-tartrate), and NaBH4, were reagent grade from Aldrich or Fluka and were used as received. High purity 18 MΩ water (Q-H2O grade Barnstead Nanopure) was used for the preparation of synthetic groundwater. As(III) and As(tot) Analysis. As(V) was measured spectroscopically with the molybdenum-blue method (32). Care was taken to avoid contamination with phosphate. Absorbance spectra were recorded from 500 to 900 nm and fitted with the distinct spectra of molybdenum blue formed from phosphate and arsenate (33), such that unwanted contributions of phosphate could be recognized and excluded. As(III) and total As were measured on a Perkin-Elmer 5000 AAS equipped with a batch MHS-20 mercury/hydride generator. Nitrogen carried AsH3 to a 900 C quartz cell where As was quantified at 197.2 nm (more stable signal at the 193.7 nm main line). As(III) was selectively detected by hydride generation in a pH 5 citrate buffer (10 mL of 1.2 M Na2HCitrate, 50-1000 µL sample and 5 mL of 3% NaBH4, purge/ reaction/purge times 25/10/40 s). HCl was used for total As (13 mL of 2.5 M HCl, 50-1000 µL sample, 5 mL of 3% NaBH4). As the hydride generation kinetics for As(III) and As(V) differ, samples for total As were oxidized with KMnO4 and calibration curves measured with As(V) were used. Calibration curves were regularly measured during photochemical experiments. The detection limits for As(III) and As(tot) were 0.1 µg/L measured by AAS and 5 µg/L for As(V) measured by the molybdenum blue method. The maximum relative standard deviation for both methods was (10%. All laboratory experiments were repeated 2-4 times, and the variations between experiments are shown by the error bars in the figures. In field experiments, As(tot) was measured with a new portable instrument, the Arsenator 510 (http://www.arsenator.com, reported detection limit 0.5 µg/L). Briefly, AsH3 is generated by adding pressed tablets containing NH2-
TABLE 1. Composition of Bangladesh Well Water, Du1 bendorf Tap Water, and Synthetic Water Bangladesh well water initial pH Ca mg/L (mM) Mg mg/L (mM) Na (mM) total Ca and Mg (mM) HCO3- mg/L (mM) Cl- mg/L SO42- mg/L H4SiO4 mg/L PO43- mg/L TOC mg/L Fe(tot) mg/L (µM) As(III) µg/L As(tot) µg/L Mn (tot) mg/L
Du1 bendorf tap water
synthetic groundwater
pH 7.0-7.2 pH 7.0-7.3 pH 7.0-7.3 50 (1.25) 20 (0.82) nma nm (2.1) (2.5-4.6) (2-3) (2.07) 305-561 244-366 253 (5.0-9.2) (4-6) (4.15) 0.5-20 4-7 108 1.5-3.5 15-32 0.1-8.8 35-70 5.1-15 50b 0.1-1.4
