Grantee Research Project Results
Final Report: Simultaneous Removal of Arsenite As (III) and Arsenate As (V) From Drinking Water Using a Novel Photoactive Adsorbent
EPA Contract Number: 68D02096Title: Simultaneous Removal of Arsenite As (III) and Arsenate As (V) From Drinking Water Using a Novel Photoactive Adsorbent
Investigators: Zeltner, Walter A.
Small Business: Microporous Oxides Science and Technology LLC
EPA Contact: Richards, April
Phase: I
Project Period: October 1, 2002 through July 31, 2003
Project Amount: $99,996
RFA: Small Business Innovation Research (SBIR) - Phase I (2002) RFA Text | Recipients Lists
Research Category: Watersheds , SBIR - Water and Wastewater , Small Business Innovation Research (SBIR)
Description:
Microporous Oxides Science and Technology, LLC (MOST), continued developing technology that was based on data obtained by Professor Marc Anderson at the University of Wisconsin-Madison (UW). Activated alumina (a commonly used adsorbent for arsenic treatment) loses its adsorption capacity for arsenate as the pH of the water being treated increases. This loss in capacity may require preacidification of the water to obtain good treatment efficiency. By adding magnesium to a suspension of -alumina nanoparticles and heating to form a magnesium aluminate spinel with a relatively high specific surface area (> 200 m2/g), these researchers demonstrated that the more alkaline spinel exhibited greater removal of arsenate (per m2) than the precursor -alumina (the primary component of many activated aluminas) at neutral pH values typical of drinking water.
Dissolved arsenic in drinking water typically is present as the charged arsenate, As(V), anion (which is easily adsorbed on numerous media) and uncharged arsenite, As(III) (which is quite difficult to remove). The general approach for removing arsenite is to first oxidize it to arsenate and then adsorb the arsenate on a suitable medium. These researchers demonstrated that photocatalytic oxidation can convert As(III) to As(V). Because of their extensive experience studying photocatalytic oxidation processes, the investigators offered several ideas for improving the efficiency of this process as part of an arsenic treatment scheme.
MOST planned to extend these studies toward eventual commercialization by pursuing several objectives, including: (1) preparing novel adsorbents as thin-film coatings on inexpensive supports, (2) evaluating the performance of these adsorbents in batch studies at selected pH values, (3) fabricating a photoreactor for performing photocatalytic oxidation studies, (4) evaluating the performance of the photoreactor for converting As(III) to As(V) in a recirculating reactor, (5) fabricating treatment systems that combine photocatalytic oxidation and advanced adsorbents, and (6) selecting a preferred treatment system.
To complete the first two objectives, an aqueous suspension of aluminum oxyhydroxide (boehmite) containing magnesium nitrate (in a 5:1 mole ratio of Al to Mg) was coated on a neutral activated alumina powder purchased from Aldrich Chemicals. Subsamples of the coated material were fired at 450°C (forming a magnesium-impregnated alumina coating) and 550°C (forming a magnesium aluminate spinel). Adsorption isotherms were measured in batch systems for the two coated materials and the uncoated substrate at two nominal pH values (7.4-7.5 and 8.1-8.2) using a short 15-minute contact time. Isotherms were obtained for different initial concentrations of arsenate and for a system that also contained two competing anions (2.5 mM bicarbonate and 2.5 mM sulfate).
Completion of the third and fourth objectives required choices to be made for two specific issues: the type of photoreactor to study, and the light source to employ.
Photoreactor. UW researchers typically employ a packed bed photoreactor for their studies and use a photocatalyst that consists of a thin coating of a titania-based catalyst on 12-mm lengths of 4-mm OD thin-wall Pyrex. For this study, however, MOST decided to utilize a waveguide photoreactor being developed at UW. In this system, both sides of the thin disks of ultraviolet (UV) transmitting acrylic are coated with the photocatalyst and stacked in a reactor with spacers between the disks. The disks are illuminated on edge by the light source so that the UV radiation is internally reflected along the disks. The evanescent wave produced at each reflection activates the photocatalyst, thus allowing light to penetrate through the reactor even if the water being treated contains UV-absorbing species.
Light Source. UW researchers generally employ commercial fluorescent UV bulbs as light sources for photocatalysis. For this project, however, MOST utilized a high-intensity, pulsed plasma UV light source provided by Novatron, Inc. Some literature studies suggest that the rates of photocatalytic processes are increased when intermittent light sources are used.
The performance of the waveguide photoreactor was evaluated by recirculating solutions of arsenite through the reactor and monitoring the concentrations of total arsenic (using inductively coupled plasma spectrometry [ICP]) and arsenite (using ICP after passing the solution through a solid phase extraction tube to remove any arsenate that formed) in subsamples taken at known times. The resulting data were used to determine rate constants for the conversion of arsenite in the reactor in the presence and absence of waveguides coated with photocatalyst. The latter test was performed to indicate the ability of Novatron, Inc.'s, pulsed plasma source to directly oxidize arsenite to arsenate.
The decision to employ a waveguide photoreactor also affected the ability to perform studies for the fifth objective. Although there would be significant advantages to employing a composite photocatalyst/adsorbent that would both oxidize As(III) and adsorb As(V) in a single treatment module, the use of an acrylic waveguide introduces significant material compatibility issues. In particular, it is not possible to coat the acrylic with the precursor to the spinel and then heat the coating to 550oC to form the spinel adsorbent without destroying the acrylic. In addition, there are reasons for testing a system in which photocatalytic oxidation occurs in a photoreactor and is followed by an adsorption bed. The latter approach can utilize existing adsorption technology. In addition, if one of these treatment methods provides excellent performance but the other method does not, it will be easier to incorporate the successful approach in a staged treatment system than in a single process. Therefore, the fifth objective was not pursued further in this project. Because the selection in the sixth objective was between a single treatment module and a staged treatment system, it was hoped that a two-stage system could be tested during this project, but time constraints prevented such a test.
Summary/Accomplishments (Outputs/Outcomes):
An initial experiment was performed to estimate the kinetics for the uptake of arsenate on the different adsorbents using a test solution of approximately 2.65 ppm As(V) in 0.01 M NaNO3 at pH 7. This experiment also provided an opportunity to evaluate the ICP method employed to quantitate arsenic concentrations and its associated quality assurance/quality control protocols. Results of this study indicated that 60-70 percent of the initial As(V) was adsorbed on the coated substrates in the first 15 minutes of exposure, but that the essentially complete removal of the arsenic required about 1 day of contact. Interestingly, the substrate itself removed more than 80 percent of the As(V) in the first 15 minutes of contact, with complete removal occurring in about 3 hours. A 15-minute contact time was chosen for future adsorption experiments.
Problems observed in measuring the pH of test solutions containing only arsenate and sodium nitrate led to the inclusion of 0.2 mM HEPES (an organic buffer) in future adsorbate solutions. Literature references indicate that inclusion of HEPES at this level in the test solution does not affect the adsorption characteristics of arsenate on typical adsorbents. After implementing this change, adsorption isotherms for all three adsorbents (neutral activated alumina as a substrate, Mg-impregnated alumina coated on the substrate, and spinel coated on the substrate) were measured for several adsorbates. These systems included a 25 ppm As(V) solution (in an effort to estimate the maximum adsorption density for As(V) on these materials) and an approximately 1 ppm As (V) solution that contained other anions (2.5 mM HCO3- and SO42-) that could compete with arsenate for adsorption sites. The adsorbate solutions were prepared at two different pH values (ca. 7.4 and ca. 8.1) to check for any differences in adsorption.
Under the conditions employed for these short-term tests, all three adsorbents displayed similar adsorption characteristics. There was little difference in measured adsorption densities at the two pH values employed for these tests. The presence of competing anions decreased the adsorption densities of all adsorbents roughly by a factor of two. Adsorption densities were on the order of a few µmoles As/g for initial arsenate concentrations near 1 ppm, and in the range of 25-40 µmoles As/g at an arsenate concentration of 25 ppm. Adsorption capacities for arsenic on commercial adsorbents generally are reported to be a few mM As/g. Part of this disparity may be caused by the short contact time employed for these tests. It also is possible that higher adsorption densities would be observed if even higher initial concentrations of arsenate were employed in these batch tests.
A complicating factor in interpreting these results is that separate measurements of pH in these systems during the 15-minute contact time indicate that pH changes during contact, sometimes by more than 0.5 pH units. Generally, pH increases during contact, with larger increases noted for the coated adsorbents than for the substrate. Because the pH increases were larger for the test solutions initially near pH 7.4 than for the solutions near pH 8.1, it is not surprising that the adsorption densities were similar for the adsorbates, even though their initial pH values differed by 0.7 pH units.
A further complication is that the coated adsorbents were found to release magnesium into the adsorbate solutions during these short-term tests. UW researchers have observed that adding magnesium to an adsorbate containing arsenate increased the adsorption density of the arsenate. More magnesium is released from the Mg-impregnated alumina coating than from the spinel coating, because some of the added magnesium is incorporated in the spinel structure and is not available for release back into the test solution.
Direct photooxidation of 3.5 L of a solution containing 0.35 ppm arsenite was observed with the Novatron pulsed plasma UV source. The fit of the data to a first-order rate expression (r2 = 0.9812) was slightly better than to a zero-order rate expression (r2 = 0.9748). The direct photooxidation, however, proceeded fairly slowly, as the first-order rate constant for the conversion was 2.86 x 10-3 min-1 (± 0.46 x 10-3 min-1 within a 95 percent confidence interval). This value corresponds to a half-life of about 4 hours for a reactor volume of 2.6 L and a flow rate of 154 mL/min. One reason for the slow reaction was the slow pulse rate of the source (one pulse every 10 seconds). Even at this low rate, the temperature of the test solution rose from 22 to 28°C during a 5-hour testing period.
When waveguides coated with photocatalyst were present in the operating photoreactor, both adsorption and photocatalytic oxidation were observed. Most of the arsenic in 2.5 L of a solution containing 0.35 ppm arsenite adsorbed on the photocatalyst before the light source was activated. Raising the concentration of arsenite in the solution to 2.5 ppm resulted in 20 percent of the arsenite being adsorbed initially. Activation of the light source initiated photocatalytic oxidation and further removal of arsenic from the test solution. The first-order rate constant for the oxidation of As(III) was 1.05 x 10-2 min-1 (± 0.06 x 10-2 min-1 within a 95 percent confidence interval), which is a half-life of 66 minutes for a reactor volume of 1.7 L and a flow rate of 154 mL/min. This process was accompanied by the removal of total arsenic from the test solution, for which the first-order rate constant was 5.28 x 10-3 min-1 (± 0.11 x 10-3 min-1 within a 95 percent confidence interval). The corresponding half-life is 131 minutes. These removal rates would be expected to increase if the pulse rate of the light source was increased. The photocatalytic oxidation/adsorption process was accompanied by a decrease in the pH of the test solution from near neutral to approximately 4.5.
Conclusions:
Under the experimental conditions employed for these tests, the three adsorbents tested for this project—a spinel-coated substrate, a magnesium-impregnated -alumina coated substrate, and the substrate itself (a neutral activated alumina consisting primarily of -alumina)—all provided similar performance for removing arsenate. No benefit was observed with the spinel. There was little difference in performance at the two different pH values employed for these tests (pH 7.4-7.5 and pH 8.1-8.2). The addition of 2.5 mM bicarbonate and sulfate to the adsorbate solution decreased the adsorption capacities of all the adsorbents by about a factor of two.
Adsorption capacities for these adsorbents at these pH values were on the order of 25-40 µmoles arsenate per gram adsorbent for a test solution containing 25 ppm As(V). Because these capacities were measured in batch tests with a 15-minute contact time, it is believed that the maximum adsorption capacities for these materials would be significantly higher if both the concentration of arsenate and the contact time were increased.
In general, the pH of the solution increased during the 15-minute contact with the adsorbents, in some cases by more than 0.5 pH units. The smallest changes were observed for the substrate, and similar changes were observed for the two coated substrates. These changes complicate the interpretation of the adsorption data.
Both coated adsorbents released magnesium into the adsorbate solution during the tests, with the spinel coating releasing less magnesium. This release of magnesium into the adsorbate solution increases the amount of As(V) that is adsorbed under a given set of experimental conditions and may be responsible for some of the pH increase noted during these tests. If these coated adsorbents were used in a packed bed treatment system, it is likely that their adsorption capacities would decrease once all the free magnesium had leached from the coatings.
The Novatron pulsed plasma UV light source can directly photooxidize arsenite to arsenate. The rate of reaction, however, is slow (a half-life of 4 hours) for a test solution containing 0.35 ppm As (III) at a pulse rate of one pulse every 10 seconds. Arsenite adsorbs on the titania-based photocatalyst employed for this study. The adsorption properties of this photocatalyst, however, were not studied in any detail.
Once a quasi-steady state was achieved after initial adsorption of arsenite, photocatalytic oxidation of a 2.5 ppm solution of arsenite in a waveguide photoreactor occurred at a faster rate (a half-life of 1 hour) than for direct oxidation. Photocatalytic oxidation of As(III) was accompanied by further removal of arsenic from the test solution. The half-life for removal of total arsenic was slightly more than 2 hours.
Supplemental Keywords:
arsenite, As(III), arsenate, As(V), arsenic removal, drinking water, photoactive adsorbent, activated alumina, magnesium, arsenic, nanoparticles, photoreactor, aluminum oxyhydroxide, ultraviolet radiation, UV, inductively coupled plasma spectrometry, ICP, spinel-based adsorbents, small business, SBIR., RFA, Scientific Discipline, Water, Environmental Chemistry, Arsenic, Analytical Chemistry, Environmental Monitoring, Drinking Water, Environmental Engineering, monitoring, public water systems, Safe Drinking Water, pathogens, risk management, photooxidation, arsenic removal, chemical contaminants, community water system, treatment, drinking water distribution system, cryptosporidium , photoactive adsorbent, arsenic exposure, contaminant removal, drinking water contaminants, drinking water treatment, water treatment, drinking water system, best available technologyThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.