Pilot-Scale Comparison of Synthetic Biotrickling Filter Media for Hydrogen Sulfide Removal at the Los Coyotes Water Reclamation Plant

INTRODUCTION
The Los Angeles County Sanitation Districts (Districts) have operated full-scale odor control systems at their facilities for many years, using either stand-alone biotrickling filters (BTFs) or a combination of a BTF and carbon adsorber. In the combined system, the BTF removes most of the hydrogen sulfide (H2S) and the carbon adsorber polishes the air to remove the remaining odor and volatile organic compounds. Lava rock has been used as the media in the Districts' BTFs since 2005. At the Districts' Joint Water Pollution Control Plant (JWPCP), the H2S removal performance with lava rock media was excellent initially but declined after several years of service. This decline, along with concerns about stress on the media support structure from the weight of the lava rock, prompted consideration of synthetic media for the BTFs. Synthetic media have been installed in several full-scale BTFs at the Districts' facilities, but comparing their performance is difficult due to the differing conditions at each facility. A side-by-side comparison of synthetic media is required to evaluate the relative performance of different media under the same air quality/flow and operational conditions. BTF performance is typically evaluated based on the H2S percent removal efficiency for a specific reactor, or more generally as the H2S removal rate normalized by the media bed volume, i.e., the elimination capacity (EC, in units of g-H2S removed/m3-h). As H2S loading increases, the biofilm within the BTF is expected to grow, resulting in increased EC. Planning is currently underway for an odor control facility consisting of BTFs and carbon adsorbers at the Districts' Los Coyotes Water Reclamation Plant (LCWRP). To support the design of the proposed system, pilot-scale BTFs were constructed and operated for approximately two years to compare three synthetic media that have been considered or used recently in the Districts BTFs: (i) polyurethane foam cubes (PU foam), (ii) polyethylene mesh (PE mesh), and (iii) calcium silicate foamed glass (foamed glass). The objectives of this study were to establish the relationship between the H2S loading and removal/EC for each of the three synthetic media, and to identify any potential issues after extended operation, such as media deformation or loss.

METHODOLOGY
Figure 1 shows the schematic diagram of the pilot-scale BTF system. Each BTF was constructed from a PVC column with a nominal diameter of ~12 inches, filled to a packing height of 4 feet, and operated with an empty bed residence time (EBRT) of ~14 seconds. A recirculation system pumped water from a recirculation tank through a spray nozzle at the top of the BTF at a flow rate of 1.7 gpm, or a hydraulic loading rate of 2.4 gpm/sf of media bed. The pH of the recirculation water was maintained between 1.5 and 2.5 by adding make-up water to the recirculation tank. The make-up water was prepared by adding nutrients to plant washwater, stored in a make-up water tank (shared by the three BTFs), and pumped as needed to each BTF's recirculation tank. An OdaLog meter continuously monitored H2S concentrations and temperatures at the BTF inlet and outlet.

RESULTS
The EC increased linearly with H2S loading (Figure 2), with the PU foam showing the steepest increases, followed by PE mesh, then foamed glass. Figure 3 presents the H2S removal efficiency under varying H2S loadings; to more clearly show trends, data have been averaged in loading increments of 10 g/m3-h, up to a loading of 500g/m3-h. The H2S removal efficiency decreased as H2S loading increased for all three synthetic media. PU foam exhibited the highest H2S removal (≥93% across all tested loadings), followed by PE mesh (≥94% at loadings up to 200 g/m3-h, and 88-93% at loadings higher than 200 g/m3-h), then foamed glass (≥95% at H2S loadings up to 35 g/m3-h, and 80-90% removal at loadings higher than 100 g/m3-h). Upon completing approximately two years of operation of the BTFs, the packing material was removed from the BTF columns and visually inspected and photographed. No signs of compression, deformation, or deterioration were identified. The media were cut in half and the cross sections were visually inspected for biofilm build-up on internal surface or pores. Biofilms were observed throughout the PU foam and PE mesh media, and some of the internal fine structures were encapsulated by biofilm build-up. However, biofilm was primarily found on the external surface of the foamed glass, not within the fine pores of the media.

DISCUSSION
The absolute rate of H2S removal increased with H2S loading (Figure 2); this result suggests that higher H2S loadings and/or concentrations result in more biomass, better mass transfer of H2S into the biofilm, and/or faster removal kinetics. However, the concurrent decrease in H2S percent removal (Figure 3) indicates that any increase in biomass, mass transfer, or kinetics does not completely compensate for the increased loading, i.e., the biomass or contact time is insufficient to completely handle the higher H2S loadings. According to the manufacturer specifications, PU foam has a higher specific surface area than the PE mesh (140-160 vs 114 square feet/cubic foot), which enables more growth of sulfide-oxidizing biomass. PU foam also has a slightly higher porosity than the PE mesh (96-97% vs 92-94%) and therefore provides more void volume and a slightly longer actual contact time for the same EBRT. The better performance by the PU foam may be attributable to a higher biomass, in combination with the slightly longer contact time. The manufacturer specifications for foamed glass list a slightly lower porosity (80-90%) and a much higher specific surface area (600 square feet/cubic foot) than the other two media. However, the visual observation of biomass on only the external surface of the foamed glass (compared to growth throughout the PU foam and PE mesh media) indicates that the pore volume and surface area within the media are inaccessible to the foul air, recirculation water, and/or sulfide-oxidizing bacteria. Consequently, much of the nominal pore volume and surface area are not available for sulfide removal, likely resulting in the observed lower performance by the foamed glass. Of the three media, foamed glass has the lowest cost; PU foam costs ~2-3 times more, and PE mesh costs ~5 times more. Consequently, foamed glass may be the 'best' choice under low loadings, while PU foam or PE mesh may be more appropriate at higher loadings. The performance of the foamed glass media at low loadings has been confirmed at full-scale in one of the Districts' odor control facilities at the JWPCP, where the H2S loading is ~13 g/m3-h. Over the past two years, the foamed glass media in these BTFs has provided > 98% to > 99.7% removal, generally reducing inlet concentrations of 50-300 ppmv down to less than 1 ppmv. However, the expected H2S loading at the LCWRP is much higher, ~60 g/m3-h. Consequently, PU foam or PE mesh is recommended at LCWRP; the higher cost of these two media is expected to be offset by better H2S removal performance, reduced H2S loadings to the second stage carbon adsorber, and decreased costs for carbon regeneration/replacement.