Comparative Performance Evaluation of Commercial Packing Materials for Malodorants Abatement in Biofiltration

Packing materials used in biofiltration of gaseous pollutants represent a key design parameter, as a proper selection might not only determine the adequate performance of the system but also its cost-effectiveness. This study systematically assessed and compared the performance of a conventional plastic carrier with that of two novel clay-based materials from SAINT GOBAIN for the abatement of a model odorous stream composed of H2S, methylmercaptan and toluene. The packing materials were tested in biotrickling filters, biofilters and a two-phase biotrickling filter. SAINT GOBAIN materials exhibited a higher adsorption potential under abiotic conditions, a higher buffer capacity and a superior performance compared to conventional plastic rings when implemented in biotrickling filters operating at gas residence times as low as 7.5 s. Among the materials tested in biofilters, Filtralite Air AC supported almost complete H2S and toluene removals at a gas residence time of 20 s, while successfully eliminating methylmercaptan at values of ~80%. Interestingly, under most of the conditions tested, clay-based materials also showed comparable pressure drop values than those of plastic rings.


Introduction
Odour abatement from industrial activities, in particular from wastewater or waste management activities, still represents a significant concern to plant operators due to the gradual encroachment of cities on industrial park and the ever stricter environmental regulations [1]. Despite odorant prevention strategies are applied at source, the implementation of active end-of-the-pipe technologies is often needed to mitigate odour nuisance. In this context, physical-chemical technologies such as incineration, activated carbon adsorption or chemical scrubbing exhibit high operating costs and environmental impacts, which has recently paved the way to odour-treatment biotechnologies [2]. Biotechnologies are based on the biocatalytic action of fungi and bacteria to oxidize, at ambient pressure and temperature, the malodorous compounds present in industrial emissions at significantly lower operating costs and environmental impacts than their physical-chemical counterparts. Bioscrubbers, biofilters and biotrickling filters rank among the most popular biotechnologies by the absence of nutrients, whose addition is required to ensure efficient biodegradation performance, the challenging adhesion of microorganisms (increasing the start-up period) and their low water retention capacity [15,16].
Therefore, the development of new tailored-made packing materials that exhibit the merits of both organic and inorganic carriers (i.e., by providing a good specific surface area, water retention capacity and structural stability) while being competitive in the market, is of paramount importance to ensure a cost-effective operation of biofilters and biotrickling filters for odour control. In this context, the company SAINT GOBAIN has developed two novel clay-based packing materials, namely Filtralite ® AIR 10-20 mm and Filtralite ® Air AC, that meet the necessary requirements for odorants biofiltration. Both materials have been conceived at an industrial level in such a way that they can be produced in the different expanded clay plants that Saint Gobain has in Europe. This study comparatively assessed the performance of conventional plastic rings and both materials from the company SAINT GOBAIN for the abatement of an odorous emission in biofilters (BFs) and biotrickling filters (BTFs) operated at multiple gas residence times. The influence of the packing material and operation mode on the structure of the microbial community was also assessed.

Inoculum and Mineral Salt Medium
The BTFs and BFs were inoculated with 0.2 L of fresh activated sludge from Valladolid wastewater treatment plant (Spain). The mineral salt medium (MSM) used in the bioreactors was composed of (g/L): K 2 HPO 4 (0.7); KH 2

Experimental Setup and Operational Procedure
The experiment was divided into two operational phases as described below. Phase I. The experimental set-up for malodorous air treatment consisted of three identical PVC biotrickling filters (internal diameter = 0.1 m, height = 1 m, total working volume = 2.5 L) packed with three different materials: Filtralite AIR 10-20 mm (BTF-F), Filtralite AIR AC (BTF-FAC) and Kaldnes K2 plastic rings (BTF-K) (Figure 1) ( Table 1). The air stream entering the system was first pumped through an activated carbon filter, where any residual pollutant from the atmospheric air was eliminated. The air was then humidified in a water column, entering afterwards a mixing chamber. In the mixing chamber, the clean and humidified air stream was mixed with a synthetic mixture of H 2 S, toluene and methyl mercaptan in N 2 (purchased to Abelló Linde, Barcelona, Spain). The resulting polluted stream was divided into three identical streams of 1.25 L min −1 by means of rotameters. The individual streams were then fed at the bottom of each biotrickling filter. The clean air exiting the biotrickling filters was discharged outside the room using an extraction fan. Three 0.5-L holding tanks magnetically stirred were used as liquid mineral salt medium (MSM) reservoirs, containing the required nutrients for the microbial community. The liquid was continuously recycled through the packed beds at a flowrate of 0.15 L min −1 countercurrently with the polluted air stream (liquid trickling velocity of 2 m h −1 ). analysis and replaced with fresh MSM. This MSM renewal also maintained the pH above inhibitory levels. The mineral salt medium exchange rate was increased when decreasing the gas residence time, following an exchange ratio of 2.5 L MSM (g H2S fed) −1 .
Inlet and outlet gas concentrations of H2S, methylmercaptan and toluene were daily analysed. Liquid samples from the recycling medium were periodically withdrawn in order to determine pH and sulfate concentration. The pressure drop across the packed beds was also periodically measured.  Phase II. The experimental setup consisted of two identical PVC biofilters (internal diameter = 0.1 m, height = 1 m, total working volume = 2.5 L) packed with two different materials: Filtralite AIR 10-20 mm (BF-F) and Filtralite AIR AC (BF-FAC), and a two-phase biotrickling filter packed with Filtralite AIR 10-20 mm (2PBTF-F) ( Figure 1). The air stream entering the system was first pumped through an activated carbon filter, where any residual pollutant of atmospheric air was eliminated. The air was then humidified in a water column, entering afterwards a mixing chamber. In the mixing chamber, the clean and humidified air stream was mixed with a synthetic mixture of H2S, toluene and methyl  A 9-days abiotic test was performed prior inoculation of the BTFs in order to evaluate the possibility of physical-chemical removal of the target pollutants by the packing materials. After that, three aliquots of 200 mL of fresh activated sludge were centrifuged for 10 min at 10,000 rpm. The supernatant was removed, and the biomass was resuspended in 100 mL of fresh MSM. The inoculum was then added to the holding tanks and recycled through the bed. The BTFs were initially operated at a gas retention time (GRT) of 2 min, which was decreased to 1 min, 30 s, 15 s and 7.5 s in subsequent operating stages. Following the inoculation period, 150 mL of the cultivation medium were daily withdrawn for analysis and replaced with fresh MSM. This MSM renewal also maintained the pH above inhibitory levels. The mineral salt medium exchange rate was increased when decreasing the gas residence time, following an exchange ratio of 2.5 L MSM (g H 2 S fed) −1 .
Inlet and outlet gas concentrations of H 2 S, methylmercaptan and toluene were daily analysed. Liquid samples from the recycling medium were periodically withdrawn in order to determine pH and sulfate concentration. The pressure drop across the packed beds was also periodically measured.
Phase II. The experimental setup consisted of two identical PVC biofilters (internal diameter = 0.1 m, height = 1 m, total working volume = 2.5 L) packed with two different materials: Filtralite AIR 10-20 mm (BF-F) and Filtralite AIR AC (BF-FAC), and a twophase biotrickling filter packed with Filtralite AIR 10-20 mm (2PBTF-F) ( Figure 1). The air stream entering the system was first pumped through an activated carbon filter, where any residual pollutant of atmospheric air was eliminated. The air was then humidified in a water column, entering afterwards a mixing chamber. In the mixing chamber, the clean and humidified air stream was mixed with a synthetic mixture of H 2 S, toluene and methyl mercaptan in N 2 (purchased to Abelló Linde, Spain). The resulting polluted stream was initially divided into three identical streams of 1.25 L min −1 by means of rotameters. The individual streams were then fed at the bottom of each biofilter and the two-phase biotrickling filter. The clean air exiting the bioreactors was discharged outside the room using an extraction fan.
An abiotic test was initially conducted for 15 days at 2 min of GRT. The BFs and 2PBTF-F were then inoculated as above described and operated at GRTs of 2 min, 1 min, 40 s and 20 s in subsequent operating stages. The biofilters were daily irrigated initially with 40 mL of MSM, increasing the flow rate to 80 mL d −1 , 120 mL d −1 and 192 mL d −1 when the BFs were operated at GRTs of 1 min, 40 s and 20 s, respectively. A 0.5-L holding tank magnetically stirred was used as liquid reservoir for the 2PBTF-F, consisting of a mixture of 70% v/v of MSM and 30% v/v of silicone oil 20 cts (Sigma Aldrich, Madrid, Spain). The liquid was continuously recycled through the packed bed at a flowrate of 0.15 L min −1 countercurrently with the polluted air stream (liquid trickling velocity of 2 m h −1 ). After inoculation, 150 mL of the cultivation medium (prior silicone oil separation) were daily withdrawn for analysis and replaced with fresh MSM in the 2PBTF-F at GRTs of 2 and 1 min, respectively, while 225 and 360 mL were daily replaced at 40 and 20 s, respectively.
Inlet and outlet gas concentrations of H 2 S, methylmercaptan and toluene were daily analysed. Liquid samples from the leachate of the BFs (when available) or from the recycling medium of the BTF were periodically withdrawn in order to determine pH and sulfate concentration. The pressure drop across the packed beds was also periodically measured.
Unless otherwise specified, the performance of the bioreactors is reported as the average removal efficiency under steady state for each operating condition (standard deviation <5%).

Analytical Procedure
Pollutants concentration was measured periodically both at the inlet and outlet of the bioreactors. Toluene was analyzed in a gas chromatograph equipped with a flame ionization detector and a HP-5-MS (30 m × 0.25 mm × 0.25 µm) column. Solid-phase microextraction (SPME) was used as a pre-concentration technique during toluene analysis using 250 mL glass bulbs as sampling ports. H 2 S and methylmercaptan were analyzed using portable detectors with specific sensors: Dräger X-am 5000 and Dräger X-am 7000, respectively.
Liquid samples were also withdrawn periodically from the holding tanks in order to determine pH and sulfate concentrations. The pH was daily measured using a pHmeter Eutech Cyberscan pH 510 (Eutech Instruments, Nijkerk, Netherlands). SO 4 2− concentrations were measured by HPLC-IC. Finally, the pressure drop across the columns was also periodically determined using a U-meter filled with water as the manometric fluid.

Microbial Community Analysis
Samples from the inoculum and the final microbial communities present in the BTFs and BFs packed with Filtralite materials (at the end of the experimentation) were withdrawn for microbial analysis. The biomass was centrifuged at 13,000× g for 10 min. The resulting pellet was used for DNA extraction using the MasterPure™ Complete DNA Purification Kit (Epicenter Biotechnologies, Madison, USA) according to the manufacturer's instructions. DNA was quantified with Qubit dsDNA broad-range (BR) assays in a QFX Fluorometer (DENOVIX, Wilmington, DE, USA) and subsequently purified. The extracted DNA was used for sequencing by Illumina Miseq. Sequencing was performed at the Foundation for the Promotion of Health and Biomedical Research of the Valencia Region (FISABIO, Valencia Spain). Amplicon sequencing was developed targeting the 16S ribosomal DNA V3 and V4 regions using the bacterial/archaeal primers described by Klindworth [17]. Illumina adapter overhang nucleotide sequences were added to the gene-specific sequences. After 16S rRNA gene amplification, library construction was carried out using the Nextera XT DNA Sample Preparation Kit (Illumina, San Diego, CA, USA). Libraries were then normalized and pooled prior to sequencing. Samples containing indexed amplicons were loaded onto the MiSeq reagent cartridge for automated cluster generation sequencing using a 2 × 300 pb paired-end run (MiSeq Reagent kit v3 (MS-102-3001)) according to manufacturer's instructions (Illumina). Quality assessment and sequence joining was performed using prinseq-lite program [18] and bioinformatic analysis was done using an ad-hoc pipeline in RStatistics environment, making use of open-source libraries (gdata, vegan. . . ). The sequence data was analyzed using qiime2 software tools. Taxonomic affiliations have been assigned using the Naive Bayesian classifier integrated in quiime2 plugins. The database used for this taxonomic assignation was the Silva138 [19].

System Performance: Odorants Abatement
Under abiotic conditions, Filtralite AIR 10-20 supported a complete removal of H 2 S, while Filtralite AIR AC removed more than 90% of this contaminant. The performance of SAINT GOBAIN materials without inoculation outcompeted the removal efficiency of conventional plastic rings, which supported a 50% removal of H 2 S. Following inoculation, the BTF-FAC maintained a complete H 2 S removal, while the BTF-F and the BTF-K rapidly achieved similar H 2 S removals 2 and 4 days after inoculation, respectively. The three BTFs were capable of degrading inlet H 2 S concentrations of~25 ppm at GRTs of 2 and 1 min (Figure 2a). The GRT was reduced to 30 and 15 s by days 50 and 64, respectively, with no concomitant deterioration of the H 2 S removal performance observed in any of the BTFs. Nevertheless, a further reduction of the GRT to 7.5 s resulted in an increase in the outlet H 2 S concentration in BTF-F and BTF-K to 1.2 ppm (corresponding to a removal efficiency of 95%), while BTF-FAC was able to sustain a complete H 2 S abatement.
Methylmercaptan was only analyzed from day 17 onwards of the experiment due to analytical issues at the initial stages of the project. No methylmercaptan was observed in the outlet gas stream of any BTF at GRTs of 2 and 1 min. However, when the GRT was reduced to 30 s, the outlet concentration of methylmercaptan increased to 2.5 mg m −3 in BTF-F and 2 mg m −3 in BTF-K and BTF-FAC, resulting in removal efficiencies of 28 and 42%, respectively. In the last operating stages, at GRTs of 15 and 7.5 s, no methylmercaptan removal was recorded in any of the BTFs (Figure 2b). Appl. Sci. 2021, 11, x FOR PEER REVIEW 7 of 15 was inoculated with activated sludge by day 9 (Figure 2c). On the contrary, both BTF-FAC and BTF-F reached toluene removal efficiencies of >99% within 11 and 15 days after inoculation at a GRT of 2 min. Surprisingly, a decrease in the residence time to 1 min promoted the abatement of toluene in BTF-K, achieving removals of >99% from day 37 onwards at this GRT. The high toluene removal was also maintained in the three BTFs at a GRT of 30 s. However, a further reduction in the GRT to 15 s resulted in a decrease in the toluene removal performance of BTF-K, increasing the outlet toluene concentration up to 0.65 mg m −3 (which corresponded to a removal efficiency of 72%). On the contrary, both BTF-F and BTF-FAC were able to maintain toluene removal efficiencies >99%, which highlights the good performance of SAINT GOBAIN packing materials during biotrickling filtration of VOCs. Finally, the reduction in the GRT to 7.5 s resulted in a deterioration in the toluene removal performance regardless of the BTF, with average toluene outlet concentrations of 1.5 mg m −3 (corresponding to removal efficiencies of 50%) for BTF-K and BTF-FAC and 1 mg m −3 for BTF-F (corresponding to a removal efficiency of 67%). days of operation. Then, sulphate concentration fluctuated from a minimum value of 1186 mg SO4 −2 L −1 on day 51 to a maximum of 1811 mg SO4 −2 L −1 on day 36. Sulphate concentration increased from 796 to 1605 mg SO4 −2 L −1 and from 264 to 753 mg SO4 −2 L −1 by days 9 to 16 of operation in the BTF-FAC and the BTF-K, respectively. Then, the concentrations of SO4 −2 gradually increased reaching a maximum of 1897 mg SO4 −2 L −1 in BTF-FAC by day 71 of operation and 1672 mg SO4 −2 L −1 by day 64 in BTF-K. These concentrations were far below inhibitory concentrations of sulphate for H2S degrading bacteria.

System Performance: Odorants Abatement
Under abiotic conditions, 2PBTF-F and BF-FAC supported a complete removal of H 2 S by day 0, while the H 2 S removal efficiency recorded in the BF-F gradually dropped from 90% by day 1 to~55% by day 15 of abiotic operation (Figure 4a). The three bioreactors were capable of biologically degrading inlet H 2 S concentrations of 19.9 ± 2.2 ppm, completely removing the H 2 S immediately after inoculation of the systems. The decrease in GRT from 2 min to 1 min and 40 s did not affect the H 2 S removal performance, no H 2 S being detected in the treated air exiting the bioreactors throughout these operational stages (regardless of the packing material or presence of trickling aqueous-silicone oil solution). Similarly, BF-FAC and 2PBTF-F supported a complete H 2 S removal at 20 s of GRT, whereas the steady H 2 S removal efficiency of BF-F slightly decreased during this last experimental period to 97.5 ± 0.3%. Appl. Sci. 2021, 11, x FOR PEER REVIEW 10 of 15  Methylmercaptan concentration was reduced from 3 to 1.5 ppmv regardless of the packing material and bioreactor configuration from day 8 of the abiotic period (Figure 4b). After inoculation, MM removal remained similar in both biofilters, while the two-phase BTF packed with Filtralite AIR 10-20 mm achieved a complete removal of MM at a GRT of 2 min. Surprisingly, a decrease in the GRT from 2 to 1 min resulted in a higher MM abatement in BF-FAC (up to~80%), which supported a removal efficiency >78% at 40 and 20 s of GRT. On the contrary, when the GRT was reduced to 1 min, a concomitant decrease in the MM abatement performance to~50% was observed in the BF-F and the 2PBTF-F. A further decrease in the GRT to 40 s resulted in average removal efficiencies of 63 ± 7% and 42 ± 9% in BF-F and 2PBTF-F, respectively. Finally, both bioreactors supported average MM removals of~45% at 20 s of GRT.
In the particular case of toluene, a negligible removal was observed in the biofilters packed with Filtralite AIR 10-20 mm or Filtralite AIR AC during the 15 days of abiotic operation (Figure 4c). On the contrary, the 2PBTF-F reached toluene removal efficiencies >99% by day 7 under abiotic conditions. After inoculation, toluene was completely removed regardless of the bioreactor, although 6 days were necessary for BF-F to achieve a complete removal. A decrease in the GRT from 2 min to 1 min and 40 s slightly reduced toluene removal in the 2PBTF-F to~98 and 96%, respectively, while no effect was observed in the biofilters performance. Finally, the biofilter packed with Filtralite AIR AC (BF-FAC) was able to completely remove the toluene from the polluted air stream, while the toluene removal efficiency of BF-F and 2PBTF-F dropped to 96 ± 3% and 89 ± 4%, respectively, at the lowest GRT of 20 s.

Analysis of the Liquid Phase and Pressure Drop across the Bed
The pH measurements started during the abiotic phase of the process in the recycling liquid of the 2PBTF, gradually decreasing from 7.36 to 4.73 by the end of the operation at a GRT of 1 min and stabilizing at~3.0 during the last two operational stages (Figure 5a). This acidification of the 2PBTF-F recycling liquid was attributed to the oxidation of H 2 S to sulphate. In the particular case of biofilters, the pH of the leachate was monitored from day 31 onwards (when leachate appeared as a result of packing media saturation with water), remaining stable in BF-F at a value of 2.2 ± 0.5 and slightly dropping from 6.6 liquid of the 2PBTF, gradually decreasing from 7.36 to 4.73 by the end of the operation at a GRT of 1 min and stabilizing at ~3.0 during the last two operational stages (Figure 5a). This acidification of the 2PBTF-F recycling liquid was attributed to the oxidation of H2S to sulphate. In the particular case of biofilters, the pH of the leachate was monitored from day 31 onwards (when leachate appeared as a result of packing media saturation with water), remaining stable in BF-F at a value of 2.2 ± 0.5 and slightly dropping from 6.6 to 4.2 by day 55 in BF-FAC, stabilizing afterwards at 2.5 ± 0.3.

Microbial Communities
The Read numbers after quality filtering and chimera filtering were 120,146 (inoculum), 134,416 (BTF-FAC), 124,396 (BTF-F), 132,622 (BF-FAC) and 30,030 (BF-F). The rarefaction curves of operational taxonomic units (OTUs) from the samples analyzed reached the plateau, demonstrating that the sequencing depth was sufficient to cover the diversity of the microbial communities in the bioreactors (Supplementary Figure S1).

Discussion
The results here obtained confirmed clay-based packing materials from SAINT GOB-AIN company as a promising platform for the optimization of the performance of biofilters and biotrickling filters devoted to the removal of odorants from different characteristics (VOC and VIC).
In this context, Filtralite materials showed better adsorption capacities when the bioreactors were operated under abiotic conditions, also supporting a more rapid startup after inoculation. The presence of water in the materials under abiotic conditions enhanced their ability to remove H 2 S, as confirmed by the superior performance of the BTFs compared to BFs. A similar H 2 S and MM abatement was supported by the three BTFs, achieving almost complete removals at 7.5 s and 1 min, respectively. On the contrary, Filtralite materials showed a superior toluene removal efficiency, exhibiting values >99% at 15 s of GRT, while toluene removal in the BTF packed Kaldnes plastic rings remained at 70%. At this point it should be highlighted that the Henry's law constants of H 2 S, toluene and methyl mercaptan are 1×10 −3 , 1.5 ×10 −3 and 3.5 ×10 −3 mol (m 3 Pa) −1 . This rules out a potential mass transfer limitation causing the deterioration of methylmercaptan removal based only on the mass transfer gradient. Gas-liquid mass transfer in biotrickling filters depends, not only on the gas-liquid concentration gradient, but also on the mass transfer coefficient (K L a), which itself is a function of the pollutant. Indeed, K L a is proportional to the (molar volume of the pollutant)ˆ0.4, where the molar volume for H 2 S, toluene and methylmercaptan are 33, 106 and 55 mL mol −1 , respectively. The results suggest that Filtralite Air AC was able to support better methylmercaptan removal efficiencies due to the higher metabolic activity of the microbial community as a result of the higher pH values provided by the packing material compared to plastic rings. In this context, the sharp initial decrease in the pH of the trickling liquid recorded in the BTF-K was likely detrimental for the microbial activity and might explain the lower MM and toluene removal performance observed in the biotrickling filter packed with plastic rings. In addition to the superior abatement performances obtained at 7-15 s (typical gas residence times for the operation of BTFs at a commercial scale), SAINT GOBAIN materials provided pressure drops as low as those of conventional plastic rings, which foresees a good performance of these materials in terms of energy demand during odour abatement. In this context, 10 cm of water column is often considered as the maximum tolerable pressure drop in biofiltration to avoid severe operating costs [20,21]. In our particular study, this value was not exceeded at any of the gas flowrates tested, even at 7.5 s of gas residence time (<3.5 cmH 2 O), which demonstrates the technical viability of the tested materials.
In biofilters, lower air velocities, and thus higher residence times are typically provided to accomplish an effective air pollutant treatment. For instance, for H 2 S and VOCs removal, typical values ranging from 40 to 60 s are usually suggested [22]. However, negligible H 2 S concentrations were recorded at the outlet of the biofilters and the two-phase biotrickling filter packed with SAINT GOBAIN materials even at the lowest GRT of 20 s. On the contrary, the two-phase BTF did not show any improvement regarding methylmercaptan and toluene removal. Previous studies have demonstrated a superior removal performance of hydrophobic VOCs such as hexane in two-phase partitioning bioreactors, due to an enhanced mass transfer of the pollutant from the gas phase to the microorganisms [4]. Similarly, Darraq et al. (2010) reported gas-silicone oil partition coefficients for sulfur compounds such as DMS and DMDS up to 30 times lower than those values for gaswater systems [23]. Therefore, it could be hypothesized that the limited methylmercaptan abatement here obtained was not associated to mass transfer limitations but rather to a low microbial activity. Overall, Filtralite AIR AC outcompeted Filtralite Air 10-20 mm for the abatement of the three model odorants, while supporting the lowest pressure drop (despite its lower particle size) and the highest buffer capacity at 40 and 20 s. Previous studies on odour treatment using compost-based biofilters have achieved an efficient abatement of odorants (>95% for H 2 S, butanone and toluene) at GRTs ranging from 30 up to 94 s, although at the expense of an important pressure drop across the filter bed (from 6 up to 33 cm of water column) due to the compaction of the organic packing material [20].
Toluene degraders were mainly distributed among the Proteobacteria and Actinobacteria phyla in the BFs and BTFs, although Actinobacteria were less represented in the reactors packed with Filtrate Air 10-20 mm. Members of Pseudomonadales, Burkholderiales, Xanthomonadales (within Betaproteobacteria) and Rhizobiales (within Alfaproteobacteria) are known to biodegrade toluene [24], and these microorganisms were present in all bioreactors. Among Actinobacteria, Mycobacterium, a well-known toluene degrader [25], was a key player in toluene degradation in BFs and BTFs.
It is important to note that Filtrate Air AC packing material favored the development of Proteobacteria as the clear dominant phyla in BFs and BTFs, while the dominance of this group in BFs and BTFs packed with Filtrate Air 10-20 was less pronounced. In fact, aerobic oxidation of methyl mercaptan is performed by members of Proteobacteria such as Thiobacillus, Hyphomicrobium, Pseudomonas, Methylophaga and Klebsiella [26]. Microorganisms of the genus Thiobacillus, Halothiobacillus and Acidithiobacillus within the Proteobacteria phylum were detected and were likely involved in methyl mercaptan degradation in the BF-FAC, BTF-F and BTF-FAC. Thiobacillus thioparus is able to oxidize methyl mercaptan producing elemental sulfur and sulfate [27] and different Thiobacillus species have been previously found in BFs and BTFs treating methyl mercaptan and other sulfur odorants [15,28,29]. Some species of Thiobacillus have been reclassified in the genera Halothiobacillus and Acidithiobacillus [30] and their ability to degrade methyl mercaptan cannot be ruled out. On the contrary, BF-F did not show aerobic methyl mercaptan oxidizers. Instead, Archaea, that were exclusively present in this bioreactor, likely degraded this compound, since the ability of Archaea such as Methanomethylovorans hollandica, Methanolobus and Methanomethylovorans to anaerobically degrade methyl mercaptan has been demonstrated [31][32][33]. These results suggest that the higher performance in terms of methyl mercaptan removal in BF-FAC as compared to BF-F, could be mediated by the development of aerobic methyl mercaptan oxidizers, that were not found in the biofilter packed with Filtrate Air 10-20. Similarly, BF-FAC showed a higher population of Thiobacillus, Acidithiobacillus and Halothiobacillus (56%) than BF-F (25%), which also could have influenced the better performance of BTF-FAC at GRT of 30 s; although none of the microbial communities were able to support methyl mercaptan removal at lower GRTs. Overall, these results demonstrate the relevance of the packing material in shaping the desired microbial communities.

Conclusions
SAINT GOBAIN clay-based packing materials exhibited a superior performance compared to conventional plastic-based carriers for the abatement of model odorous compounds (H 2 S, toluene and methylmercaptan) when implemented both in biofilter and biotrickling filter configurations. Their superior performance at gas residence times as low as 7.5 s, together with their comparable pressure drops to those of plastic rings, open up new opportunities for boosting the cost-effectiveness of biotechnologies for odour abatement. Among the materials tested in biofilters, Filtralite Air AC supported an almost complete H 2 S and toluene abatement and ∼80% methylmercaptan removals at a gas residence time of 20 s. Hence, SAINT GOBAIN materials have the potential to become the core of a new generation of advanced biofilters based on the combination of high buffer capacity, high water retention, good gas-liquid mass transfer, low pressure drops and high structural resistance.