Biotrickling filter modeling for styrene abatement. Part 2: Simulating a two-phase partitioning bioreactor.

A dynamic model describing styrene abatement was developed for a two-phase partitioning bioreactor operated as a biotrickling filter (TPPB-BTF). The model was built as a coupled set of two different systems of partial differential equations depending on whether an irrigation or a non-irrigation period was simulated. The maximum growth rate was previously calibrated from a conventional BTF treating styrene (Part 1). The model was extended to simulate the TPPB-BTF based on the hypothesis that the main change associated with the non-aqueous phase is the modification of the pollutant properties in the liquid phase. The three phases considered were gas, a water-silicone liquid mixture, and biofilm. The selected calibration parameters were related to the physical properties of styrene: Henry's law constant, diffusivity, and the gas-liquid mass transfer coefficient. A sensitivity analysis revealed that Henry's law constant was the most sensitive parameter. The model was successfully calibrated with a goodness of fit of 0.94. It satisfactorily simulated the performance of the TPPB-BTF at styrene loads ranging from 13 to 77 g C m-3 h-1 and empty bed residence times of 30-15 s with the mass transfer enhanced by a factor of 1.6. The model was validated with data obtained in a TPPB-BTF removing styrene continuously. The experimental outlet emissions associated to oscillating inlet concentrations were satisfactorily predicted by using the calibrated parameters. Model simulations demonstrated the potential improvement of the mass-transfer performance of a conventional BTF degrading styrene by adding silicone oil.

Dynamic mathematical modelling of the removal of hydrophilic VOCs by biotrickling filters.

A mathematical model for the simulation of the removal of hydrophilic compounds using biotrickling filtration was developed. The model takes into account that biotrickling filters operate by using an intermittent spraying pattern. During spraying periods, a mobile liquid phase was considered, while during non-spraying periods, a stagnant liquid phase was considered. The model was calibrated and validated with data from laboratory- and industrial-scale biotrickling filters. The laboratory experiments exhibited peaks of pollutants in the outlet of the biotrickling filter during spraying periods, while during non-spraying periods, near complete removal of the pollutant was achieved. The gaseous outlet emissions in the industrial biotrickling filter showed a buffered pattern; no peaks associated with spraying or with instantaneous variations of the flow rate or inlet emissions were observed. The model, which includes the prediction of the dissolved carbon in the water tank, has been proven as a very useful tool in identifying the governing processes of biotrickling filtration.

Investigating bacterial populations in styrene-degrading biofilters by 16S rDNA tag pyrosequencing.

Microbial biofilms are essential components in the elimination of pollutants within biofilters, yet still little is known regarding the complex relationships between microbial community structure and biodegradation function within these engineered ecosystems. To further explore this relationship, 16S rDNA tag pyrosequencing was applied to samples taken at four time points from a styrene-degrading biofilter undergoing variable operating conditions. Changes in microbial structure were observed between different stages of biofilter operation, and the level of styrene concentration was revealed to be a critical factor affecting these changes. Bacterial genera Azoarcus and Pseudomonas were among the dominant classified genera in the biofilter. Canonical correspondence analysis (CCA) and correlation analysis revealed that the genera Brevundimonas, Hydrogenophaga, and Achromobacter may play important roles in styrene degradation under increasing styrene concentrations. No significant correlations (P > 0.05) could be detected between biofilter operational/functional parameters and biodiversity measurements, although biological heterogeneity within biofilms and/or technical variability within pyrosequencing may have considerably affected these results. Percentages of selected bacterial taxonomic groups detected by fluorescence in situ hybridization (FISH) were compared to results from pyrosequencing in order to assess the effectiveness and limitations of each method for identifying each microbial taxon. Comparison of results revealed discrepancies between the two methods in the detected percentages of numerous taxonomic groups. Biases and technical limitations of both FISH and pyrosequencing, such as the binding of FISH probes to non-target microbial groups and lack of classification of sequences for defined taxonomic groups from pyrosequencing, may partially explain some differences between the two methods.

Microbial community analysis in biotrickling filters treating isopropanol air emissions.

The evolution of the microbial community was analysed over one year in two biotrickling filters operating under intermittent feeding conditions and treating isopropanol emissions, a pollutant typically found in the flexography sector. Each reactor was packed with one media: plastic cross-flow-structured material or polypropylene rings. The communities were monitored by fluorescence in situ hybridization (FISH) and denaturing gradient gel electrophoresis (DGGE) analysis of the 16S rRNA region. After inoculation with activated sludge, the biotrickling filters were operated using inlet loads (ILs) from 20 to 65 g C m(-3) h(-1) and empty-bed residence times (EBRTs) from 14 to 160 s. Removal efficiencies higher than 80% were obtained with ILs up to 35 g C m(-3) h(-1) working at EBRTs as low as 24 s. There was an increase in the total percentage of the target domains of up to around 80% at the end of the experiment. Specifically, the Gammaproteobacteria domain group, which includes the well-known volatile organic compound (VOC)-degrading species such as Pseudomonas putida, showed a noticeable rise in the two biotrickling filters of 26% and 27%, respectively. DGGE pattern band analysis revealed a stable band of Pseudomonas putida in all the samples monitored, even in the lower diversity communities. In addition, at similar operational conditions, the biotrickling filter with a greater relative abundance of Pseudomonas sp. (19.2% vs. 8%) showed higher removal efficiency (90% vs. 79%). Results indicate the importance of undertaking a further in-depth study of the involved species in the biofiltration process and their specific function.

Performance of a pilot-scale biotrickling filter in controlling the volatile organic compound emissions in a furniture manufacturing facility.

A 0.75-m3 pilot-scale biotrickling filter was run for over 1 yr in a Spanish furniture company to evaluate its performance in the removal of volatile organic compounds (VOCs) contained in the emission of two different paint spray booths. The first one was an open front booth used to manually paint furniture, and the second focus was an automatically operated closed booth operated to paint pieces of furniture. In both cases, the VOC emissions were very irregular, with rapid and extreme fluctuations. The pilot plant was operated at an empty bed residence time (EBRT) ranging from 10 to 40 sec, and good removal efficiencies of VOCs were usually obtained. When a buffering activated carbon prefilter was installed, the system performance was improved considerably, so a much better compliance with legal constraints was reached. After different shutdowns in the factory, the period to recover the previous performance of the biotrickling reactor was minimal. A weekend dehydration strategy was developed and implemented to control the pressure drop associated with excessive biomass accumulation.

Long-term performance of peat biofilters treating ethyl acetate, toluene, and its mixture in air.

Three laboratory-scale peat biofilters were operated at 90 s empty bed residence time (EBRT) for over a year. Biodegradation of ethyl acetate, toluene, or a 1:1 mixture were investigated. In first stage, inlet concentration was progressively increased from 0.4 to 4.5 g/m(3). The maximum elimination capacity (EC) found for ethyl acetate was 190 gC/m(3).h, and it was not affected by toluene. The maximum EC found for toluene as a sole contaminant was 150 gC/m(3).h, but the presence of ethyl acetate decreased the toluene maximum EC to 80 gC/m(3).h. From respirometry monitoring, values of 3.19 g CO(2)/gC and 3.06 g CO(2)/gC for pure ethyl acetate and pure toluene, respectively, were found, with overall yield coefficients of 0.13 g dry biomass produced per gram ethyl acetate consumed and 0.28 g dry biomass produced per gram toluene consumed. CO(2) production in the 1:1 mixture was successfully simulated. Dynamics of living and dead cells were monitored in four sections of the biofilters. Concentrations ranged between 2.6 x 10(9) and 3.0 x 10(10) cells per gram-dry peat for total bacteria, and 2.4 x 10(9)-1.9 x 10(10) cells per gram-dry peat for living bacteria. At high loads loss of bacterial density in the inlet zones, and increase in the dead cells percentages up to 60% was observed. In second stage, long-term performance at an inlet concentration of 1.5 g/m(3) was evaluated to show the process feasibility. Good agreement with previous data was obtained in terms of EC and CO(2) production. Restoration of living cells proportion was also observed.