Toward a sustainable and cost-efficient biological-based platform for siloxanes removal

Abstract Volatile methyl siloxanes (VMS) are persistent contaminants extensively used in industrial applications. Their presence in biogas constitutes a major hindrance for its energetic valorization or its use as renewable natural gas. Current commercial siloxanes abatement technologies are based on physical-chemical methods, whose good performance is impaired by their high investment costs, and a hefty environmental impact. Research evidences that VMS are indeed biodegradable, which opens the possibility of implementing bio-based technologies as a cost-effective and sustainable alternative for the removal of these compounds. This review uncovers the most plausible organisms and microbial pathways involved in biological VMS degradation, a relatively unexplored area. Additionally, the most commonly applied configurations and the main operating challenges are thoroughly revised and discussed, evidencing that a feasible implementation relies on the optimization and scale-up of enhanced mass transfer bioreactors. Finally, the techno-economic and environmental analysis demonstrate that, although in a very early stage, emerging biological technologies for siloxanes removal will play an important key role on the future abatement of VMS from waste gas streams such as raw biogas. Graphical abstract


Introduction
One of the most persistent anthropogenic trace components emitted to the environment are volatile methyl siloxanes (VMS) and have been recently classified as potential trophic magnification compounds in aquatic and terrestrial food web (Cui et al., 2019;Fremlin et al., 2021). VMS consist of volatile, small and medium molecular weight compounds (<500 g mol À1 ) which have in their structure silicon, carbon and oxygen atoms (Table 1) (Ru & Ku, 2015). Siloxanes are extensively used in industrial applications (i.e. construction, electronics, automotive industry, textiles, medical equipment, food packaging, medicine and cosmetics) as well as in household products (i.e. detergents, shampoos, cosmetics, paper coatings and textiles) due to their high compressibility, low flammability, low surface tension and water repelling properties, high thermal stability and a limited effect of temperature on their properties (Dewil et al., 2006). Their current production is estimated in around 8 million tons per year (Guo et al., 2021) from which an important part ends up in waste treatment facilities (Tansel & Surita, 2017). During wastewater treatment, the majority of high-molecular weight siloxanes (polydimethylsiloxanes, PDMS) do not decompose and adsorb onto the extracellular polymeric substances of the sludge, together with the fraction of the VMS that has not volatized (Dewil et al., 2006;Soreanu et al., 2011). PDMS are subsequently hydrolyzed to VMS and volatilized during anaerobic digestion of the sludge, ending up in the produced biogas together with the adsorbed VMS, mainly D4, D5 and D6 (Table 1) (Guo et al., 2021). The concentration of VMS in gas emissions of waste treatment plants is in the range of 0.1-20 mg m À3 and might increase up to 400 mg m À3 in the case of biogas ( Angeles Torres et al., 2020;Gaj & Pakuluk, 2015;Wang et al., 2019). If the gas treatment process involves combustion (for example during the energetic valorization of the produced biogas), VMS are oxidized to CO 2 and SiO 2 , resulting in deposits of abrasive micro particulates of crystalline SiO 2 . These silica particles trigger detrimental effects on the facilities equipment through the erosion and corrosion of boilers and fuel cells, internal combustion engines and/or turbines, etc. Thus, the accumulation of silica has a strong negative economic impact in waste facilities, which devalues biogas and restrains its use in production plants (for example, it has been estimated that the removal of siloxanes in a wastewater treatment plant facility can save between 2.6 and 5.7 e/(m 3 d À1 ) treated wastewater in operating costs) (McCarrick, 2012). However, Table 1. Most common siloxanes present in raw biogas and their potential biodegradation products. due to their unique physicochemical properties, siloxanes are irreplaceable in many areas. Therefore, the abatement of VMS from raw biogas emissions in waste facilities is required to obtain cost-effective and high quality biomethane that can be used as a renewable energy source.
Current commercially available technologies to treat VMS from gas emissions consist of conventional physical-chemical processes (Soreanu et al., 2011), which support a good abatement performance, although at the expense of high investment costs, and a hefty environmental impact associated to a significant consumption of raw materials, organic solvents, chemicals and energy. Recent research evidences that volatile methyl siloxanes are indeed biodegradable, which opens the possibility of implementing bio-based technologies as a cost-effective and sustainable alternative for the removal of these compounds (Boada et al., 2020;Li et al., 2014). In fact, technologies based on the biological degradation of VMS through the action of microorganisms can achieve removal efficiencies (REs) up to 90% . Nevertheless, their implementation is still in development and their application is still scarce. Only lab-scale biological technologies have been configured and operated. Additionally, there is still a huge lack of knowledge about the specific organisms and the degradation processes involved in VMS biological treatment technologies.
This review aims at critically discussing the state-of-the art of current technologies for the abatement of siloxanes present in raw biogas, while analyzing existing knowledge on the fundamentals and implementation of their biological alternatives. The potential role of microorganisms undergoing biological VMS degradation and the most feasible metabolic pathways involved in the process are also presented. Finally, a comparative techno-economic and environmental feasibility analysis of most representative physical-chemical and biological technologies was performed and the results critically evaluated.

Commercial technologies for VMS abatement
The industrial solutions currently on the market for VMS abatement only involve physical-chemical processes: adsorption, absorption, catalytic processes, membrane separation, refrigeration/condensation and cryogenic methods, being the main marketable process the adsorption of VMS on fixed bed reactors (Table 2) and the main sorption materials, activated carbon and silica gel.
The use of adsorbents supports a high removal performance ensuring concentration values below 0.1 mg m À3 , together with an easy operation and maintenance and an extensive practical experience in industrial applications. However, physical-chemical technologies are characterized by high operating costs and a strong carbon footprint. In particular, the high energy consumption and the extreme operating conditions, the regeneration or replacement of the adsorbent material, the use of organic solvents and the deactivation of the catalysts are the main contributors to the operating costs and environmental impact. Similarly, the negative impact on climate change is associated to process energy requirements, production and transportation of adsorbents or packing materials, and the use of chemicals (Ajhar et al., 2010;de Arespacochaga et al., 2015;Gaj, 2017;Shen et al., 2018).
With the aim of overcoming the high operating costs of physical-chemical technologies for VMS abatement, recent advances on physical-chemical technologies have focused on the development and optimization of commercial adsorbents, as the main challenge relies on increasing the breakthrough adsorption capacity of the sorption material. For example, enhancing the activation process or the pore width of activated carbon increases the adsorption capacity, while reducing the replacement cost. In the case of silica gel it has been observed that micro/mesoporous silica gels can increase the adsorption capacity of conventional technologies (Meng et al., 2020b). Other critical parameters are the physical-chemical properties of the adsorbent and its regeneration. Textural properties such as the presence of functional groups or inorganic impurities like alkali metals (K, Na) determine the amount of hydrophilic sites, that ultimately reduce VMS adsorption capacity in the presence of humidity and the regeneration capacity distinctive of siloxanes polymerization (Tran et al., 2019b(Tran et al., , 2019a(Tran et al., , 2018. Moreover, the use of activated carbon with functionalized mixed transition metal oxides (such as the commercial SulfaTrap R8V R ) seems to have superior VMS removal capacities (Calbry-Muzyka et al., 2019). In the case of silica gels, the meso-/microporous structure and hydrophobic surface of acetylated and methyl-functionalized silica gels showed adsorption levels and breakthrough times up to 18 times higher compared to conventional silica gel Meng et al., 2020a). Overall, these tailored materials have demonstrated enhanced siloxanes' adsorption capacities as compared to conventional adsorbent materials (C. C. u. Bak et al., 2019). However, further steps on adsorption processes must target at increasing the sustainability of the process. For instance, the valorization of solid waste as raw material (wood waste, lignocellulosic waste) for sorbents production has been recently assessed. These studies suggested that the adsorbents obtained could achieve performances comparable or even higher to those of commercial activated carbons in an environmentally friendly way (Papurello et al., 2019(Papurello et al., , 2018Santos-Clotas et al., 2019).
Although these novel advances in physical-chemical technologies are targeting the economic and environmental sustainability of the VMS removal process, they still show high investment and operating costs (see Table 2 and Sec. 4), and high environmental impacts which goes against the principles of sustainability. Thus, biodegradation technologies, while in an embryonic stage when applied for siloxanes abatement, will certainly play a key role to decrease the negative impacts of physical-chemical processes.

Prokaryotes involved in siloxanes degradation
Episodically, there have been some attempts to uncover if microorganisms are able to use siloxanes as carbon source (Table 3), however, the routes for the biological degradation of siloxanes by microorganisms still remain uncertain.
The first study that suggested the feasibility of the biological degradation of organosilicon compounds by microorganisms was conducted in 1967 by Fessenden and Fessenden (Fessenden & Fessenden, 1980). In this research, it was claimed that Pseudomonas species were likely able to degrade silane. Thereafter, different species of this genus have been reported as able to grow on complex organosiloxanes. Wasserbauer and Zadak (1990) attested degradation of high molecular polydimethylsiloxane and low molecular weight siloxanes (i.e., L 4 , L 3 , L 2 ) using two species of Pseudomonas (P. putida and P. fluorescens) in the presence of O 2 (Wasserbauer & Zadak, 1990). Similarly, P. aeruginosa grew on oligo-and poly-organosiloxanes in the presence of glucose (Ro sciszewski et al., 1998) (Table 3). In both studies cell growth was observed, however, CO 2 evolution was not recorded and siloxane breakdown products were not identified. Moreover, the high concentration of other organic compounds in the soil used as inoculum and/or the addition of glucose could have promoted bacterial growth not related to siloxanes degradation, while siloxanes metabolization might have been associated to Si À O chemical cleavage (Ro sciszewski et al., 1998).
Further on, Accettola et al. (2008) studied the degradation of D4 as the only carbon source aerobically using three types of inocula: (1) a culture of P. putida, (2) activated sludge from an urban wastewater treatment plant and (3) activated sludge from the wastewater treatment plant of a company producing silicones. The protein analysis and optical density measurements showed that P. putida didn't grow as a pure culture on D4 (Accettola et al., 2008). However, both enrichments from wastewater were able to grow using D4 as the only carbon source. In both cases, the predominant bacteria detected belonged to the genus Pseudomonas, being identified as P. aeruginosa, P. putida and P. citronellosis together with other genera with lower abundance (Rhodanobacter, Zoogloea, Mesorhizobium, Xanthomonada) (Fig. 1). Thus, the co-culture of different species seemed to be an important factor to obtain efficient D4 degradation. The main metabolite identified in this study was dimethylsilanediol (a concentration of DMSD of 14 ppm was detected in the co-cultures in comparison with 7 ppm in the negative controls), although negligible amounts of silicates were detected and CO 2 evolution was not recorded. Likewise, Li et al. (2014) observed aerobic degradation of D4 using an isolate of P. aeruginosa strain S240. L 3 (from DMSD) and trimethylmethoxysilane (from methanol) were detected, which indicated the presence of DMSD and methanol from the degradation of D4. Silicic acid and CO 2 were also recorded as degradation products. This last study proved that P. aeruginosa was able to use D4 as the only carbon source in the presence of O 2 , however, the complete biological degradation of D4 could not be demonstrated since methylsilanetriol or its derivative methyltris(trimethylsiloxy)silane was not detected. The authors suggested that these compounds were transient intermediates of DMSD degradation (Li et al., 2014). The last study that pointed Pseudomonas as a siloxanes consumer was conducted by Boada et al. (2020), who isolated bacterial species using D4 as the sole carbon source in anaerobic conditions (Table 3). Phylogenetic analysis of the 16S rRNA gene revealed that the main isolate found belonged to the genus Pseudomonas. However, CO 2 consumption or resting metabolites were not measured, which combined with the lack of negative controls, compromised the evidence of biological degradation of D4.
Pure strains from other genera different than Pseudomonas have been also identified as siloxanes degraders. Sabourin et al. (1996Sabourin et al. ( , 1999 described biodegradation of DMSD and methylsilanetriol in the presence of oxygen and a primary carbon source (dimethylsulfone) by a new species of the genus Arthrobacter. In both studies, 14 C-DMSD degradation was linked to a small production of 14 CO 2, 14 C-methylsilanetriol and inorganic silicate, while 14 C-methylsilanetriol degradation to 14 CO 2 could not be proven (Lehmann et al., 1998;Sabourin et al., 1999). Members of Phyllobacterium were also linked to siloxanes degradation in aerobic conditions. In this research the authors claimed that Phyllobacterium myrsinacearum in pure culture efficiently removed D4 during the operation of a biotrickling filter (Wang et al., 2014b). Nevertheless, the lack of quality check analysis of the isolate's purity prevents from ensuring the degradation of D4 by a single strain. Similar to Li et al. (2014), the authors identified in the liquid phase L 3 , L 4 , DMSD, and some small amounts of tetramethyldisiloxane-1,3-diol and silicic acid, potential by-products of D4 degradation (Wang et al., 2014b).
In recent studies of siloxanes biodegradation, advanced molecular tools such as next generation sequencing, have been implemented to study the enriched population of siloxanes removal biotechnologies (Boada et al., 2020(Boada et al., , 2021Pascual et al., 2020;Zhang et al., 2020). In these studies, sewage sludge was used as the inoculum for enrichment of organosiloxanes degraders. Interestingly, only Boada et al. (2020) identified Pseudomonas as one of the enriched genus after biodegradation of D4. However, the authors reported a new species of the genus Methylibium to be the most efficient D4 degrader (removal efficiencies up to 53% in anaerobic vials containing D4 as the sole carbon source). In a follow-up study, Boada et al. (2021) also identified the genus Methylibium as one of the most abundant genus in an anaerobic lab-scale biotrickling filter. Although in this case the most abundant bacteria belonged to the genera Castellaniella and Pseudomonas was not detected (Boada et al., 2021). Using D4 and D5 as carbon sources, Pascual et al. (2020) enriched an aerobic consortium dominated by members of the genera Pseudoxanthomonas (3 to 21% relative abundance), Reyranella (8 to 10% relative abundance), Chitinophaga (5 to 7% relative abundance), Flavobacterium (2 to 11.2% relative abundance) and an uncultured genus from the family Acidithiobacillaceae, KCMB-112 (7 to 20% relative abundance). However, the only product from degradation measured during operation was CO 2 . Alternatively, Zhang et al. (2020) enriched acidophilic microorganisms from anaerobic sludge during siloxanes treatment. In this study the archaeal community was analyzed for the first time, exhibiting that the enriched population was highly dominated by the archaeal genus Ferroplasma (85.5% relative abundance). The other main organisms found belonged to the genus Acidithiobacillus (9.5%) and Thermogymnomonas (4.6%) (Florentino et al., 2015). Degradation metabolites similar to those found by Li et al. (2014) and Wang et al. (2014b) were detected, although CO 2 production or consumption was not reported in this study (Table 3). However, Zhang et al. (2020) studied the degradation of siloxanes mixed with biogas, which contains CO 2 , CH 4 and H 2 S. In this sense, the authors related the growth of acidophilic prokaryotes with the use of carbon and energy sources different from VMS, and the removal of siloxanes with abiotic absorption by liquid recircularization in the biofilter, as well as adsorption onto the biomass.

Metabolic pathway for siloxanes biodegradation
The catabolic pathway for siloxanes biodegradation is still unknown: neither the genes expressed, nor the proteins involved in the degradation of siloxanes have been researched to date. It has been extensively reported that D4 and D5 chemically hydrolyzed to linear oligomer silanediols, which can be further chemically hydrolyzed to DMSD (Ru & Ku, 2015). Moreover, the reverse condensation is also observed, from the oligomer diols mixture obtained from D4, D3 and some small amounts of D5 and D6 are produced (Soreanu et al., 2011). Thus, some reports claimed that the hydrolysis of organosiloxanes during bioprocess performance is carried out chemically (throughout reactions at the organic groups and Si À C cleavage) and not biologically (Ru & Ku, 2015). However, several studies have demonstrated a biocatalyzed hydrolysis of Si À O without any Si À C cleavage in comparison with negative controls (Gr€ uMping et al., 1999;Li et al., 2014;Ohannessian et al., 2008;Ro sciszewski et al., 1998;Wang et al., 2014b). This fact makes plausible the biocatalyzation of the hydrolysis of complex organosiloxanes by microbial enzymes. According to the products obtained during the degradation of D4 and D5, two main degradation pathways have been hypothesized (Fig. 2). Gr€ uMping et al. (1999) hypothesized that D4 is enzymatically hydroxylated to DMSD by bacterial enzymes. In their experiment negative controls without biomass and with the addition of D4 didn't produce DMSD, however, when biomass was added D4 decreased and a DMSD increase was observed. A subsequent decrease of DMSD after a long incubation period (2 months) was related to biodegradation processes. According to Gr€ uMping et al. (1999) a microbial DMSD hydroxylase transforms DMSD into hydroxy(dimethoxy)silane which is further converted into formylmethylsilanediol through the expression of a silanediol oxidase. Formylmethylsilanediol is then hydrolyzed to methylsilanetriol and formaldehyde. Likewise, Sabourin et al. (1999) proposed biocatalyzed hydroxylation followed by abiotic oxidation to the corresponding DMSD that would be finally cleaved to methylsilanetriol and formaldehyde. Methylsilanetriol is supposed to be hydroxylated to hydroxymethyl-silanetriol through the hypothetical enzyme methylsilanetriol-hydroxylase which is then oxidized to formylsilanetriol with a hypothetical hydroxymethyl-silanetriol oxidase. Finally, the hydrolysis of formylsilanetriol would completely processed DMSD to silicic acid and a molecule of formaldehyde. The two molecules of formaldehyde produced during DMSD degradation are catabolized in the C 1 metabolic cyclic characteristic of bacteria. Silicic acid is excreted to the extra-cellular environment (supplementary materials, Fig. S1).
However, Li et al. (2014), Wang et al. (2014b), and Zhang et al. (2020) reported that there are probably several chemical and biological intermediates during the transformation of D4 and D5 into DMSD. Li et al. (2014) and Wang et al. (2014b) proposed that the D4 ring is opened by hydrolysis to produce a linear intermediate, octamethyltetrasiloxane 1,7 diol. After hydrolysis and polymerization of this compound, either by chemical or biological action, hexamethyltrisiloxane 1,5 diol and tetramethydisiloxane 1,3 diol are formed and hydrolyzed again to DMSD. Similarly, Zhang et al. (2020) suggested that under acidic conditions D5 was hydrolyzed to decamethylpentasiloxane 1,9 diol, which was further hydrolyzed to DMSD through the same linear organsiloxanes intermediates proposed by Li et al. (2014). However, in this hypothetical pathway it was claimed that a precursor would be necessary to transform DMSD since there is no evidence of a direct oxidation of a Si-CH 3 to a Si-OH group. Once DMSD is formed, the different degradation pathways postulated agreed that DMSD is then converted into hydroxy(dimethoxy)silane. Nevertheless, the potential enzymes undergoing the process were not described.
Based on this assumption, DMSD is biologically oxidized into hydroxy(dimethoxy)silane which is biologically hydrolyzed or rearranged into methoxymethylsilanediol, which is then hydrolyzed to methylsilanetriol and methanol. A following oxidation and rearrangement of a second intermediate end up producing silicic acid and methanol through hydrolysis. The two molecules of methanol produced can be further oxidized to formaldehyde, and to CO 2 via the C1 metabolic cycle pathway (Fig. 2). These pathways could be carried out in aerobic and anaerobic conditions, being the main difference the C1 metabolic cycle for carbon assimilation in each case.
Although there are several studies that seem to prove the biological degradation of VMS, some research stands that the process is mainly abiotic and the degradation observed is due to biomass adsorption and chemical hydrolyzation (Ru & Ku, 2015;Zhang et al., 2020). In these sense, further studies targeting carbon assimilation of VMS in the biomass (using labeled VMS), the isolation and physiological assimilation of VMS degrading strains, and the development of multi-omic studies (transcriptomics, proteomics and metabolomics) that help to disclose the metabolic pathways involved in the process are necessary to definitively proof the biological degradation of these compounds.

Implementation of biotechnologies for siloxanes abatement: first steps and future prospects
As a growing number of studies try to disentangle the microbiology, the degradation products and the degradation pathways involved in biological siloxanes abatement, there has also been an increasing interest on the implementation of biotechnologies for the removal of these compounds. The most investigated and best performing biotechnology has been biotrickling filtration (Fig. 3A). Early studies focused on the abatement of D3 and D4 in both aerobic and anoxic biotrickling filters (BTF). The maximum removals obtained were low, around 20% for D3 and 43% for D4, respectively (Accettola et al., 2008;Popat & Deshusses, 2008). Higher D4 removals ($55%) were achieved in an aerobic BTF inoculated with activated sludge obtained from the effluent of an organic silicon manufacturer operated at gas residence times (GRT) between 12 and 30 min (Wang et al., 2014a). The reinoculation of the same BTF with a previously isolated microorganism (Phyllobacterium myrsinacearum) resulted in a slight increase in the removal performance up to 60.2% at a GRT of 24 min (Wang et al., 2014b). Both studies demonstrated a detrimental effect on the system performance when decreasing the GRT, with optimum values of 15 to 24 min and a pH of the recycling liquid between 4-7. Few researchers have recently investigated the simultaneous removal of VMS and trace compounds such as H 2 S, hexane, toluene and/ or limonene (Boada et al., 2021;Zhang et al., 2020), reaching maximum removal efficiencies of D4 and D5 between 40 and 50%.
The limited performance reported for siloxanes biodegradation in BTFs has been consistently associated to the low solubility of siloxanes in the recycling mineral medium (high Henry's law coefficient; L2: 404, L3: 1440, D3: 72, D4: 252; D5: 183 (www.henrys-law.org)), which hinders the mass transfer of these hydrophobic compounds from the gas phase to the liquid phase and consequently reduces their availability to the microbial community. This entails operation at high gas residence times in order to promote mass transfer, resulting in increased reactor volumes and, therefore, higher investment costs (see Sec. 4). In this sense, the enhancement of the D4 removal up to 74% (at a GRT of 13.2 min) in a biotrickling filter inoculated with Pseudomonas aeruginosa S240 was associated to the presence of Rhamnolipids, organic biosurfactants produced by the genus Pseudomonas (Li et al., 2014).
Therefore, the feasible implementation of siloxanes biodegradation technologies relies on the optimization and scale up of enhanced mass transfer bioreactors. For instance, two-phase partitioning biotrickling filters (TP-BTF) (Fig. 3B), based on the addition to conventional biotrickling filters of a non-aqueous phase with high affinity for the target pollutants, have been recently studied as a potential alternative for VMS abatement. The superior performance of this configuration on the removal of L2, L3, D4 and D5, compared to a conventional BTF has been demonstrated under aerobic conditions at a gas residence time of 60 min. A total VMS removal <30% was recorded in the conventional BTF, increasing up to 70% in the two-phase partitioning BTF. D4 and D5 were the compounds with the highest removals in the TP-BTF, reaching values of 80 and 90%, respectively  at optimal operating conditions of 60 min (GRT) and 45% of silicone oil in the cultivation broth (Pascual et al., 2021).
Hollow fiber membrane bioreactors (MBR) (Fig. 3C) constitute another novel technology lately implemented for the removal of VMS along with other trace compounds. The highest removals obtained for D4 and D5 were $17 and $21%, corresponding to elimination capacities of $1 and $2.5 g m À3 h À1 , respectively. Despite the low performance of the MBR in terms of VMS abatement, the reduced gas retention times (18-60 s) allowed for significant lower reactor sizes compared to those of BTFs . Further research on membrane materials that support an increased mass transfer of the gaseous compounds to the microbial biofilm is necessary in order to ensure the cost-efficiency of this novel biotechnology.
As mentioned in a previous section (see Secs. 3.1 and 3.2), the bases of siloxanes removal of these processes are uncertain, however the presence of secondary metabolites in the cultivation broth, the CO 2 production observed, and the specialized bacterial community enriched during the processes suggests biological degradation of VMS.
Overall, biotechnologies are characterized by a lower performance compared to several physical-chemical processes. However, their low investment and operating cost (Table 2, Sec. 4) and the promising results recently obtained encourage further research on the optimization and scaling of these processes.

Techno-economic and environmental aspects of siloxanes removal: toward a more sustainable process
A simplified comparative techno-economic and environmental analysis was performed including the most commonly employed physical-chemical technologies for siloxanes removal (activated carbon adsorption and chemical absorption), the best performing biotechnology up-to-date (BTF) and a combined configuration consisting of a BTF followed by a polishing step based on AC-adsorption. The highest operating costs were calculated for AC adsorption technology (7.2 e/(m 3 /h) À1 ), in contrast to the BTF that showed the lowest operating costs (1.2 e/(m 3 /h) À1 ) (Fig. 4a). This expenditure results from the continuous replacement of the activated carbon after saturation of the adsorbent, which contrasts with the long lifespan and low pressure drop of the packing material in the BTF (10 years and 500 Pa, respectively). In this sense, media replacement constitutes the main contributor to the operating costs of these technologies (66% and 44% for the AC adsorption and BTF, respectively). Chemical absorption also presented higher operating costs (3.6 e/(m 3 /h) À1 ) compared to the BTF associated to chemicals purchase (organic solvents and concentrated alkaline or acid solutions), with a share of 69% to the total costs. Finally, the combined BTF-AC exhibited operating costs slightly higher (2.7 e/(m 3 /h) À1 ) than those of the BTF due to the replacement of the activated carbon (41% of the total operating costs). However, these values were lower than those calculated for physical-chemical technologies thanks to the preliminary reduction in VMS concentration in the BTF, which increases the lifespan of the AC on the adsorption unit. Similar results have been obtained in previous evaluation of the economics of odor removal technologies (Estrada et al., 2012).
Conversely, the highest capital investment costs were obtained for the BTF and the combined BTF-AC technology (711,698 and 898,389 e, respectively), being significantly lower for the physical-chemical processes (191,314 and 25,220 e for Chemical absorption and AC adsorption, respectively) (Fig. 4b). Nevertheless, the lower operating costs of the BTF and the combined BTF-AC alternatives might offset the initial investment expenses, thus resulting in a favorable economic balance in the long-term.
As expected, the biological technology exhibited the lowest environmental impact according to the climate change indicator (2.0 Â 10 À4 kg CO 2 eq. m À3 ), followed by the combined BTF-AC (8.0 Â 10 À4 kg CO 2 eq. m À3 ) (Fig. 4c). The energy requirement was the main contributor to the climate change indicator for the BTF, while AC production and transportation increased the impact of the combined BTF-AC. On the contrary, AC adsorption (1.9 Â 10 À3 kg CO 2 eq. m À3 ) and chemical absorption (1.4 Â 10 À3 kg CO 2 eq. m À3 ) presented the highest CO 2 footprints due to the packing material, chemical use and waste disposal in landfills (Alfons ın et al., 2015).
It should be noted that this analysis focuses on the exclusive removal of siloxanes. The simultaneous removal of other pollutants present in biogas, such as H 2 S or volatile organic compounds, would increase the operating costs and the environmental impact of the studied technologies. For instance, competitive adsorption of different compounds will reduce the AC lifespan, thus increasing its replacement frequency. Similarly, the formation of carbonates in alkaline absorption processes (derived from the reaction of CO 2 and NaOH) increases chemicals consumption and therefore, the operating cost (Shen et al., 2018). BTF will be likely the least affected technology by the simultaneous removal of other substances, since siloxanes demand higher GRTs due to their low solubility, which is in turn a determining parameter for the investment cost.
In summary, the analysis demonstrated that biotechnologies outperform physical-chemical siloxanes abatement technologies from both an economic and an environmental point of view. However, the low VMS removal efficiencies achieved to date in BTFs (70%) limit the stand-alone implementation of this technology. In this context, the combined BTF-AC configuration could be an adequate solution for VMS removal, since the BTF unit decreases the overall environmental impact and the operating costs while de AC unit increases the removal efficiency of the global process (Fig. 4d).
It is also worth to mention that recent advances in biotechnologies could change this current scenario. For instance, as previously mentioned, TP-BTFs exceed the siloxanes abatement efficiency of conventional BTFs, reaching values up to 90%. The industrial silicone oil employed in this configuration to boost the mass transfer of VMS has a current market price of 11 eL À1 , which would not significantly affect the total costs. The performance of TP-BTFs for VMS removal from raw biogas is currently being studied within the URBIOFIN project in a demo plant, an innovation project funded by the Bio Based Industries Joint Undertaking (BBI JU) under the EU Horizon 2020 programme. This technology will be validated in a 120 L unit fed with 7.2 m 3 d À1 of VMS loaded biomethane from preliminary biogas upgrading unit (P erez et al., 2020).