Key Role of Microbial Characteristics on the Performance of VOC

Mar 1, 2012 - When silicone oil 200 cSt (SO200) was added to the systems, the ... a factor of 8.7 and 16.3 for Consortium A (hydrophilic microorganism...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/est

Key Role of Microbial Characteristics on the Performance of VOC Biodegradation in Two-Liquid Phase Bioreactors María Hernández, Guillermo Quijano, and Raúl Muñoz* Department of Chemical Engineering and Environmental Technology, Valladolid University, Dr Mergelina, s/n, 47011, Valladolid, Spain S Supporting Information *

ABSTRACT: Despite being studied for over 20 years, little is known about the mechanisms underlying the treatment of volatile organic compounds (VOCs) from industrial off-gases in two-liquid phase bioreactors (TLPBs). Recent reports have highlighted a significant mismatch between the high abiotic mass transfer capacity of TLPBs and the low VOC biodegradation rates sometimes recorded, which suggests that a process limitation might also be found in the microbiology of the process. Therefore, this study was conducted to assess the key role of microbial characteristics on the performance of VOC biodegradation in a TLPB using three different hexane degrading consortia. When silicone oil 200 cSt (SO200) was added to the systems, the steady state hexane elimination capacities (ECs) increased by a factor of 8.7 and 16.3 for Consortium A (hydrophilic microorganisms) and B (100% hydrophobic microorganisms), respectively. In the presence of SO200, Consortium C supported a first steady state with a 2-fold increase in ECs followed by a 16-fold EC increase after a hydrophobicity shift (to 100% hydrophobic microorganisms), compared to the system deprived of SO200. This work revealed that cell hydrophobicity can play a key role in the successful performance of TLPBs, and to the best of our knowledge, this is the first report on hydrophobic VOC treatment with exclusive VOC uptake within a nonbioavailable non aqueous phase. Finally, an independent set of experiments showed that metabolite accumulation can also severely inhibit TLPB performance despite the presence of SO200.



INTRODUCTION The biological treatment of volatile organic compounds (VOCs) from off-gases is frequently challenged by the hydrophobic nature of some VOCs (e.g hexane, pentane, and terpenes), which limits the VOC transfer from the gas to the aqueous phase, where biodegradation usually takes place.1,2 These systems are also limited by toxicity problems due to the accumulation of toxic metabolites when treating high VOC loading rates.3 Two-liquid phase bioreactors (TLPBs) are based on the addition of a non-aqueous phase (NAP) with a high affinity for the target hydrophobic VOC in order to overcome mass transfer limitations and microbial inhibition.4 The presence of a NAP can provide an overall higher driving force for mass transfer and induce an increase in the gas interfacial area,5 which ultimately enhances the transfer of hydrophobic VOCs and therefore their biodegradation rates.6,7 Moreover, the NAP can maintain the concentration of toxic substrates and metabolites produced below subinhibitory levels, thus improving process robustness.8−10 Recent studies on VOC mass transfer and biodegradation in TLPBs showed that increases in the VOC mass transfer due to a NAP addition under abiotic conditions did not result in © 2012 American Chemical Society

comparable increases in VOC biodegradation. For instance, Hernandez et al.11 observed that the hexane elimination capacities achieved in a TLPB were significantly lower than the hexane transfer capacity of the system recorded under abiotic conditions. Similarly, Rocha-Rios et al.12 reported no significant enhancements in CH4 biodegradation in a TLPB implemented in an airlift bioreactor, while the same TLPB under abiotic conditions supported a 6-fold increase in the volumetric mass transfer coefficient relative to a control reactor deprived of NAP. Therefore, these recent experimental findings suggest that microbial characteristics can also play a key role in the VOC biodegradation performance of TLPBs. In this regard, some authors confirmed that microbial characteristics like cell hydrophobicity can strongly affect the biodegradation performance of hydrophobic pollutants in TLPBs. For instance, Ascon-Cabrera and Lebeault 13,14 attributed the higher biodegradation performance of 2,4,6trichlorophenol in a TLPB (compared to a control deprived of Received: Revised: Accepted: Published: 4059

November 22, 2011 February 20, 2012 March 1, 2012 March 1, 2012 dx.doi.org/10.1021/es204144c | Environ. Sci. Technol. 2012, 46, 4059−4066

Environmental Science & Technology

Article

BM-B-450 strain, were combined and cultured at 30 °C and 200 rpm in 0.5-L glass-bottles initially filled with 100 mL of mineral salt medium and supplied with 50 μL of hexane (corresponding to 82 g m−3) every 2−3 days for 23 days . To furnish fresh inoculum, Consortium C was enriched in 0.5 L glass bottles as above-described for 5 days prior to experimentation. Chemicals. All chemicals for mineral salt medium preparation were purchased from Panreac (Barcelona, Spain), with a purity of at least 99%. Mineral salt medium was prepared according to Hernandez et al.11 n-Hexane (99.0% purity) was obtained from MERCK (Madrid, Spain). Silicone oil 200 cSt (dynamic viscosity = 0.19 kg m−1 s−1) and Antifoam 204 (compatible with biological applications) was purchased from Sigma−Aldrich (Madrid, Spain). Experimental Setup and Operation Mode. A sterile 3-L jacketed glass reactor (Afora S.A., Spain) equipped with two marine impellers was initially filled with 1900 mL of sterile mineral salt medium and 100 mL of the corresponding fresh bacterial inoculum. The system was operated at 300 rpm and 30 °C in the absence of SO200 for 10 days until a steady state was achieved. At day 10, 400 mL of cultivation medium were replaced with 400 mL of sterile SO200 (corresponding to a volume fraction of 20%) and the system was operated under similar conditions for 15 days more. The biomass from the 400 mL of cultivation medium drawn was returned (prior centrifugation under sterile conditions) to the bioreactor in order to avoid a microbial activity limiting scenario. Gaseous hexane at 2.1 ± 0.1 g m−3 was continuously supplied through the aeration (1 L min−1 of air filtered through a sterile 0.2 μm Millex1-FG membrane filter) resulting in a loading rate of 64 ± 1 g m−3 h−1. Sterile distilled water and sterile silicone oil were periodically added to minimize water losses by evaporation and silicone oil losses due to medium exchange, sampling and foam formation, respectively. The cultivation medium from the bioreactor was periodically replaced with fresh sterile mineral salt medium as described below to maintain the pH value above 5.5 as well as to provide nutrients, and to remove any potential inhibitory metabolites accumulated in the culture broth. Hence, when the bioreactor was operated in the absence of SO200, a dilution rate (D) of 0.025 day−1 was used. However, 24 h after SO200 was added to the system (day 11) the D was increased as follows: from day 11 to 15 the D was 0.35 day−1; from day 15 to 17 the D was 0.70 day−1 and from day 17 onward the D was 1.0 day−1. When the D was higher than 0.025 day−1, the cultivation broth drawn was centrifuged at 5000 rpm for 10 min and the biomass pellet was resuspended in fresh sterile mineral salt medium and returned into the bioreactor. Gas samples were taken using gastight syringes (Hamilton, U.S.) to monitor the inlet and outlet hexane and CO2 concentrations. Liquid samples of 50 mL were periodically drawn to record pH, culture absorbance of the aqueous phase, dissolved total organic carbon (TOC) and total nitrogen. While the TOC content of the aqueous phase was determined, the content of organic compounds potentially accumulating in the NAP was not measured due to the high affinity of SO200 for organic compounds, which makes metabolite extraction extremely difficult for any solvent. Performance of Hexane Biodegradation by Different Microbial Consortia. The performance of Consortium A, B, and C was evaluated independently using the experimental setup and operational protocols described above. In order to elucidate whether hexane transfer from the gas to the liquid

NAP) to the cell attachment to the NAP/aqueous interface. MacLeod and Daugulis15 reported that Mycobacterium PYR-1 exhibited unprecendently high polycyclic aromatic hydrocarbons biodegradation rates probably due to its ability to grow preferentially in the NAP, which increased polycyclic aromatic hydrocarbons bioavailability compared to the aqueous phase. Despite the large number of studies conducted to date in the field of TLPBs, most of them focused on the selection of the optimum NAP as a key process parameter in TLPB performance and ignored the biological aspect of this technology. This study was therefore conducted to assess the key role of microbial characteristics on the performance of VOC biodegradation in TLPBs. Hence, the performance of three different hexane degrading consortia was evaluated in a TLPB implemented in a stirred tank. Silicone oil 200 cSt (SO200) and hexane were used as model NAP and hydrophobic VOC, respectively. In addition, the effect of metabolite accumulation on hexane biodegradation (as a result of process operation at different mineral salt medium exchange rates) was further assessed in an independent set of experiments.



MATERIALS AND METHODS Microorganisms. Three bacterial consortia isolated under different conditions were used in this work: Consortium A was isolated from 2 mL of activated sludge (Valladolid sewage work, Spain) cultured in 1 L-glass bottles containing 100 mL of mineral salt medium and 5 mL of 2,2,4,4,6,8,8-heptamethylnonane. Heptamethylnonane was not directly in contact with the mineral salt medium but placed in a separate compartment in the bottle. This NAP was used for the isolation of Consortium A at low hexane concentrations due to the high affinity of heptamethylnonane for hexane (Gas phase/ heptamethylnonane partition coefficient = 0.0027). The bottles were incubated at 30 °C under magnetic agitation (200 rpm) and supplied with 10 μL of hexane (corresponding to an initial headspace concentration of 2 g m−3) every 3−4 days for 22 days. To furnish fresh inoculum, Consortium A was further enriched as above-described for 20 days prior to experimentation. Consortium B was obtained from a biotrickling filter treating a VOC mixture at trace level concentrations (0.28 ± 0.02 mg m−3 of hexane, 0.22 ± 0.03 mg m−3 of toluene, 0.23 ± 0.03 mg m−3 of α-pinene and 22 ± 2 mg m−3 of methyl mercaptan,) for approximately 5 months. The biotrickling filter was packed with polyurethane foam, inoculated with activated sludge (Valladolid sewage work, Spain) and the operational conditions can be found elsewhere.16 To furnish fresh inoculum, a 2-L glass-bottle containing 1 L of mineral salt medium was inoculated with 2 mL of the biofilm present in the biotrickling filter and continuously supplied with 1 L min−1 of gaseous hexane at 5− 10 mg m−3. The culture was incubated at room temperature under magnetic agitation (200 rpm) for 40 days. Consortium C was obtained by initially isolating hexane degrading bacteria in 120 mL glass serum bottles containing 18 mL of mineral salt medium and 2 mL of activated sludge (Valladolid sewage work, Spain) in the absence and in the presence of 10% of heptamethylnonane. The bottles were incubated at 30 °C under magnetic agitation (200 rpm) and supplied with increasing amounts of hexane (from 5 to 15 μL corresponding to initial hexane headspace concentrations from 34 to 101 g m−3, respectively) according to Hernandez et al.17 The isolated strains, together with a Pseudomonas aeruginosa 4060

dx.doi.org/10.1021/es204144c | Environ. Sci. Technol. 2012, 46, 4059−4066

Environmental Science & Technology

Article

present in the cultivation medium, the liquid supernatant was removed after centrifugation (at 5000 rpm for 10 min) and the biomass pellet was resuspended in fresh mineral salt medium to its original concentration in order to avoid silicone oil interferences in absorbance measurements. pH was measured using a CyberScan pH-510 (Eutech Instruments, Nijkerk, The Netherlands).

phase was limiting the process performance, the bioreactor was operated at a higher stirring rate (500 rpm) from days 22 to 25 and subjected to a 3 h step increase in hexane loading rate (by doubling the hexane inlet concentration and maintaining the air flow constant at 1 L min−1) at the end of each experiment. During the 3 h step increase, gas samples were taken every 1.5 h to monitor the inlet and outlet hexane and CO 2 concentrations. Hexane biodegradation performance was evaluated in terms of elimination capacity (EC, g m−3 h−1) and removal efficiency (RE, %). The main operational conditions of this section are summarized in Table 1.



RESULTS AND DISCUSSION The isolation strategy was mainly based on the exposure of the microorganisms to different gaseous hexane concentrations, which resulted in consortia with different macroscopic characteristics (e.g., different hydrophobicities). Other studies have shown that the operation mode can influence the cell hydrophobicity.14 Unfortunately, in this particular study the authors could not, a priori, establish a clear relationship between the isolation conditions and cell hydrophobicity. The hydrophobicity definition was based on the behavior of the cultures in the presence of the NAP in the bioreactor, since it was not possible to measure directly the cell hydrophobicity because the biomass was totally adhered to the NAP (even at a centrifugal force of 45 000 × g the biomass could not be separated from the silicone oil). However, culture absorbance measurements in the bioreactor follow the same methodology used in the BATH method, which is a well-known technique to quantify the cell hydrophobicity.18 In addition, despite no molecular-biology based characterization of the microbial consortia was performed, optical microscopic observations confirmed the bacterial nature of all consortia here tested. Performance of Hexane Biodegradation by Consortium A. Steady state ECs of 3.4 ± 0.5 g m −3 h −1 (corresponding to REs of 5.7 ± 0.6%), CO2 production rates of 4.3 ± 0.7 g m−3 h−1 and culture absorbance values of 0.27 ± 0.01 were recorded for Consortium A in the absence of SO200 (days 1−10) (Figure 1). During this stage, the pH values decreased from 7.0 to 6.2, total nitrogen decreased from 180 to 173 mg L−1 while TOC ranged from 12 to 78 mg L−1. The TOC values increased by days 9−10 likely due to the accumulation of hexane biodegradation metabolites as a result of the low dilution rate used in this stage (0.025 day−1). SO200 addition from day 11 gradually increased both the hexane ECs and CO2 production rates until a steady state was achieved by day 14. From day 14 to 22, the steady state ECs and CO2 production rates were 31.5 ± 4.3 g m−3 h−1 (REs of 49.9 ± 6.1%), and 58.7 ± 10.9 g m−3 h−1, respectively. Culture absorbance achieved a maximum value of 2.7 at day 18 and then decreased to ∼1.0 by day 22. The fact that the culture absorbance was higher than 0 in the aqueous phase suggested that consortium A was hydrophilic. The gradual increase in D (from 0.025 day−1 to 1.0 day−1 once SO200 was added) allowed maintaining the pH above 6.2, total nitrogen above 93 mg L−1, while TOC fluctuated from 27 to 183 mg L−1. The stirring rate was increased by day 22 from 300 to 500 rpm, which resulted in steady state ECs and CO2 production rates 42−43% higher, concomitant with an increase in culture absorbance. Likewise, the 2-fold increase in hexane loading rate brought about a 2-fold increase in the EC and a 1.5-fold increase in CO2 production. These results confirmed that the system was indeed limited by hexane mass transfer during operation at 300 rpm and 64 ± 1 g hexane m−3 h−1 despite the presence of SO200. Performance of Hexane Biodegradation by Consortium B. Hexane biodegradation by the hydrophobic Con-

Table 1. Operational Conditions for Consortia A, B, and C stage

period (days)

control 20% SO200

0−10 10−11 11−15 15−17 17−22 22−25 the last 3 h of the experiment

D (day‑1)

agitation rate (rpm)

hexane concentration (g m−3)

0.025 0.025 0.35 0.70 1.0 1.0 1.0

300 300 300 300 300 500 500

2 2 2 2 2 2 4

Assessing the Effect of Metabolite Accumulation on Hexane Biodegradation. Hexane biodegradation was carried out as described in the experimental setup and operation mode section using Consortium A as model microbial community but maintaining a lower D (from 0.025 to 0.4 day−1). In this experiment, the medium exchange protocol aimed at maintaining the pH value above 6.0 (avoiding microbial inhibition by pH) and then, assessing only the effect of metabolite accumulation on process performance. Hence, when the reactor was operated in the absence of SO200 (days 0−10), the D was 0.025 day−1. This value was maintained for 3 days after S0200 addition (days 10−13) and then increased up to 0.4 day−1 from days 13 to 16. Finally, the D was decreased to 0.2 day−1 from day 16 onward. The potential inhibitory effects of the metabolites accumulated in the cultivation broth were also assessed. For this purpose, 120 mL serum bottles supplied with 3 μL of hexane and containing 30 mL of cultivation medium consisting of different ratios of centrifuged cultivation broth from the bioreactor at day 16 and fresh mineral salt medium (0:30, 15:15, and 30:0) were cultivated with 1 mL of inoculum from the bioreactor at day 16 under orbital agitation at 300 rpm and 30 °C. Hexane headspace concentration was periodically monitored by GC-FID for 4 h. Analytical Procedure. The hexane gas concentration was determined using an Agilent 6890 gas chromatograph (Palo Alto, U.S.) equipped with a flame ionization detector according to Hernandez et al.11 CO2 concentrations were determined in a Varian CP-3800 gas chromatograph (Palo Alto, U.S.) coupled with a thermal conductivity detector according to Hernandez et al.11 TOC and total nitrogen were measured using a TOC-VCSH analyzer (Shimadzu, Tokyo, Japan) coupled with a total nitrogen module based on chemiluminesce detection (TNM1, Shimadzu, Japan). Culture absorbance measurements at 650 nm were performed using a HITACHI U200 UV/vis spectrophotometer (Hitachi,Tokyo, Japan). When SO200 was 4061

dx.doi.org/10.1021/es204144c | Environ. Sci. Technol. 2012, 46, 4059−4066

Environmental Science & Technology

Article

Figure 2. Time course of (a) the hexane loading rates (□), ECs (■), CO2 production rates (○), pH (-) and (b) culture absorbance at 650 nm (*), TOC (◊), and total nitrogen (▲) during hexane biodegradation by Consortium B without SO200 (days 0−10) and with SO200 (days 10−25). Vertical bars represent the standard deviation of duplicate measurements and horizontal arrows the medium exchange rates and the stirring rates.

Figure 1. Time course of (a) the hexane loading rates (□), ECs (■), CO2 production rates (○), pH (−) and (b) culture absorbance at 650 nm (*), TOC (◊), and total nitrogen (▲ ) during hexane biodegradation by Consortium A without SO200 (days 1−10) and with SO200 (days 10−25). Vertical bars represent the standard deviation of duplicate measurements and horizontal arrows the medium exchange rates and the stirring rates.

average values of 6.2 ± 0.1 as a result of the stabilization of process performance (steady ECs from day 15 to 25). While the addition of SO200 in the experiment conducted with Consortium A brought about a sustained biomass increase in the aqueous phase, when the process was inoculated with Consortium B the aqueous biomass concentration decreased from a maximum culture absorbance value of 0.6 to 0 by day 14 (Figure 2b). The fact that the culture absorbance decreased to 0 in the aqueous phase indicated that consortium B was 100% hydrophobic. Figure 5Sa (Supporting Information, SI) shows the high affinity of Consortium B for SO200, where no turbidity was observed in the aqueous phase from day 14 onward. This experimental finding clearly indicated that microbial characteristics can play a key role in the performance of TLPBs. Despite the high Ds applied, which maintained TOC between 10 to 38 mg L−1, the total nitrogen values sharply decreased to 12 mg L−1 (at day 16) as a result of the intense biomass growth in the organic phase. When the stirring speed was increased up to 500 rpm by day 22, no enhancement in process performance was recorded. These results preliminary ruled out a potential mass transfer limitation. However, the 2-fold increase in hexane loading rate did result in a 2-fold increase in the EC but no significant

sortium B in the absence of SO200 was characterized by steady state ECs of 3.6 ± 1.4 g m−3 h−1 (corresponding to RE of 5.5 ± 1.8%), CO2 production rates of 2.1 ± 0.9 g m−3 h−1 and culture absorbance values of 0.03 ± 0.01 (Figure 2). Hexane biodegradation brought about a slight decrease both in the pH values (from 7.0 to 6.3) and in total nitrogen (from 209 to 203 mg L−1), while TOC gradually increased from 8 to 27 mg L−1. SO200 addition prompted a rapid increase in the hexane EC and CO2 production rate up to steady state ECs of 59.4 ± 3.1 g m−3 h−1 (RE of 87.5 ± 3.0%) by day 15, which were ≈90% higher than those recorded for Consortium A after SO200 addition under comparable experimental conditions (Figure 2a). Similarly, CO2 production rates steadily rose after SO200 addition up to 152.2 ± 4.4 g m−3 h−1 by day 22 (∼2.5 times higher than in the case of Consortium A). The high hexane biodegradation performance mediated by the presence of SO200 also caused a rapid decrease in pH, which reached a minimum value of 5.5 by day 15 in spite of the increasing Ds. This strong pH drop indicates the high production rate of intermediates, which was masked when high D values were established. From day 15 onward, the pH increased up to 4062

dx.doi.org/10.1021/es204144c | Environ. Sci. Technol. 2012, 46, 4059−4066

Environmental Science & Technology

Article

days 10 to 17. In this period, culture absorbance increased up to a maximum value of 0.9 concomitantly with a tiny decrease in total nitrogen (from 206 to 178 mg L−1) and pH (from 7.1 to 6.9). Therefore, the culture absorbance higher than 0 in the aqueous phase indicated that consortium C was hydrophilic. Unexpectedly, culture absorbance rapidly decreased from day 17 to finally reach negligible values by day 19, which suggested that consortium C become 100% hydrophobic from day 19 onward. The fact that Consortium C was preferentially associated to SO200 from day 18 onward clearly indicated a shift in the hydrophobic characteristics of the consortium. Figure 5Sb,c (SI) shows this drastic change in the SO200 dispersion in the TLPB operated with Consortium C. It is worth noting that a more stable emulsion was formed before the hydrophobicity shift, while after the hydrophobicity shift a very viscous NAP-biomass aggregate was formed. This change in the cell hydrophobicity matched with a rapid improvement in the EC and CO2 production rate, reaching steady state values of 58.0 ± 3.1 g−3 h−1 (RE of 90.6 ± 1.8%) and 82.0 ± 7.1 g−3 h−1, respectively. Interestingly, the ECs achieved after the hydrophobicity shift were very similar to those recorded with Consortium B and were also correlated with sharp decreases in the pH (from 6.9 to 6.1) and total nitrogen (from 178 to 44 mg L−1) but no increase in TOC. These results confirmed the strong impact of the microbial characteristics on the TLPB performance and constitute, to the best of our knowledge, the first experimental evidence of the key role of the interactions between the microorganisms and the NAP on the success of TLPBs. Hydrophobicity shifts mediated by the presence of an NAP are well documented in the literature. For instance, several authors have reported an increase in cell hydrophobicity when strains of Pseudomonas are grown in the presence of 1-decanol or heptamethylnonane.18,20 However, such shifts seems very difficult to be a priori predicted since Consortium A did not show any significant change in its hydrophobicity throughout the experiment under the same experimental conditions tested for Consortium B and C. Similar to the results obtained with Consortium B, the increase in stirring rate did not cause any modification in the process performance. However, the 2-fold increase in the hexane loading rate did result in a 2-fold increase in the EC and the CO2 production rate (111.2 ± 5.0 g−3 h−1 and 146.9 ± 7.1 g−3 h−1, respectively), which confirmed that hexane mass transfer also limited hexane biodegradation by Consortium C. Under these scenarios of mass transfer limitation for the 3 consortia, the mass transfer mechanisms were likely determined by microbial hydrophobicity. The mass transfer limitation implies that the VOC concentration was zero in the aqueous phase for consortium A (uptake in the water), and zero in the NAP for consortia B and C (uptake in the NAP). This implies that the gas/NAP transfer pathway operated at its maximum driving force for hydrophobic consortia, which was superior to the gas/water pathway (Henry law constant for SO200 was 13 000 times lower than for water). Thus, 13 000 is the value for the SO200/water partition coefficient of hexane, while 7.7 × 10−5 is the water/SO200 one. However, when hexane uptake occurs in the aqueous phase (hydrophilic consortia), the gas/ NAP and NAP/water transfer pathways probably do not work at its maximum capacity since the VOC concentration in the NAP might not be 0 as previously observed by Quijano et al.21 Additionally, the increases in the CO2 production rates for consortia B and C corresponded with the EC enhancements

increase in CO2 production rates, which pointed out again toward mass transfer limitations during hexane biodegradation by Consortium B. This apparent mismatch might be explained by the fact that volumetric mass transfer coefficient in TLPBs is not a linear function of the stirring speed, low volumetric mass transfer coefficient enhancements being commonly observed by doubling the agitation rate in viscous liquid media.19 In this particular case, the strong association between the NAP and the biomass resulted in a very viscous single aggregate. Moreover, the increase in the stirring rate was not correlated with a better SO200-biomass dispersion (Figure 5Sa, SI). Performance of Hexane Biodegradation by Consortium C. In the absence of SO200, Consortium C supported steady state ECs of 3.9 ± 0.6 g m−3 h−1 (corresponding to REs of 5.9 ± 1.0%), CO2 production rates of 2.3 ± 1.6 g m−3 h−1and culture absorbance values of 0.2 ± 0.1 (Figure 3).

Figure 3. Time course of (a) the hexane loading rates (□), ECs (■), CO2 production rates (○), pH (−) and (b) culture absorbance at 650 nm (*), TOC ( ◊) and total nitrogen ( ▲ ) during hexane biodegradation by Consortium C without SO200 (days 0−10) and with SO200 (days 10−25). Vertical bars represent the standard deviation of duplicate measurements and horizontal arrows the medium exchange rates and the stirring rates.

In addition, slight variations in pH and total nitrogen were observed during the first 10 days of operation, while TOC ranged from 35 to 95 mg L−1. Surprisingly, the addition of SO200 slightly improved process performance, reaching steady state ECs and CO2 production rates of 7.1 ± 1.6 g m−3 h−1 (RE of 11.8 ± 2.8%) and 3.8 ± 1.9 g m−3 h−1, respectively, from 4063

dx.doi.org/10.1021/es204144c | Environ. Sci. Technol. 2012, 46, 4059−4066

Environmental Science & Technology

Article

when the hexane loading rate was doubled, suggesting that the NAP-microorganisms resistance to hexane transfer was negligible when microorganisms are immersed in the NAP. However, in the case of consortium A (immersed in the aqueous phase), the CO2 production rates did not correspond with the EC enhancements, and therefore in this case, the NAP- microorganisms resistance must be considered. This, together with the lower performance of the gas/water pathway, can explain the lower ECs recorded for consortium A. Despite a carbon balance would be helpful to explain the C fluxes, a global C balance could not be performed in our system since the hydrophobic biomass was totally adhered to the NAP (even at a centrifugal force of 45 000 ×g the biomass could not be separated from the silicone oil) and could not be quantified. Likewise, the contents of organic compounds potentially accumulated in the NAP could not be determined due to the high affinity of silicone oil for organic compounds, which makes metabolite extraction extremely difficult for any solvent (i.e., silicone oil is one of the preferred materials used for solid-phase micro extraction, SPME). However, the metabolite accumulation rate in the aqueous phase as well as the percentage of the inlet C converted to metabolites in the aqueous phase was determined (Table S2, SI). At low D, the percentage of inlet C converted to metabolites was similar for the 3 consortia and remained below 2%. However, at the highest dilution rate (1 day−1), the percentage was 3 times higher in the case of the hydrophilic consortium (consortium A) than in the case of hydrophobic consortia (B and C). The steady state ratios of CO2 produced per gram of hexane degraded in the presence of silicone oil at an agitation rate of 500 rpm, as well as the percentage of hexane mineralization relative to the stoichiometric yield (assuming that all the hexane is fully converted to CO2, i.e., 3.02 g CO2/g hexane degraded) are shown in Table S3 (SI). The hexane mineralization percentages obtained in this study were up to 5 times higher than those previously recorded by Muñoz et al.22 This indicates the better mineralization performance achieved in this study, especially in the case of consortium B. Full-scale TLPBs implemented in stirred tank bioreactors would entail high energy requirements as a result of the high pressure drop originated by the air diffusion through the liquid column and the agitation required to disperse the NAP,23,24 which would jeopardize their economic viability. In this context, consortium B (composed of high-performance hydrophobic microorganisms) is currently being evaluated in a two-phase partitioning compost biofilter to overcome the energy consumption limitations for TLPB implementation at full scale. Assessing the Effect of Metabolite Accumulation on Hexane Biodegradation. An additional hexane biodegradation experiment using Consortium A as model microbial community was conducted in order to assess the potential inhibitory effect of metabolite accumulation on hexane biodegradation. Consortium A was selected since it was exposed to a known metabolite concentration and because it grew preferentially in the aqueous phase. When the system was operated without SO200 (days 1−10), process performance remained similar to the previous experiments conducted without an NAP (steady state ECs and CO2 production rates of 3.5 ± 0.4 g−3 h−1 and 3.2.0 ± 0.4 g−3 h−1, respectively) (Figure 4). However, when SO200 was added, the EC and CO2 production rate rapidly increased up to a

Figure 4. Time course of (a) the hexane loading rates (□), ECs (■), CO2 production rates (○), pH (−) and (b) culture absorbance at 650 nm (*), TOC (◊ ) and total nitrogen ( ▲ ) during hexane biodegradation by Consortium A without SO200 (days 0−10) and with SO200 (days 10−25) in a system operated at a lower D in order to assess the effect of metabolite accumulation. Vertical bars represent the standard deviation of duplicate measurements and horizontal arrows the medium exchange rates.

maximum value of 52.1 g m−3 h−1 (RE of 76.6%) and 78.6 g m −3 h −1 , respectively (day 13). This intense hexane biodegradation (1.7% higher than the steady state ECs in the previous experiment conducted with Consortium A at higher Ds) was concomitant with a rapid increase in the TOC values, which suggested an important accumulation of metabolites (maximum value of 347 mg L−1). Thereafter, the EC decreased to values even lower than those observed in the absence of SO200 (1.6 g m−3 h−1). From day 18 onward, a gradual recovery in the EC was observed, achieving a final steady state at 22.6 ± 3.6 g−3 h−1 (RE of 34.1 ± 6.2%) by day 20. In this context, microbial inhibition due to pH or nitrogen limitation must be ruled out since pH and total nitrogen values remained above 6 and 23 mg L−1, respectively, during the whole experiment. However, the steady state ECs obtained in this experiment were 1.4 lower than those obtained in the first experiment conducted with Consortium A at higher D values (1400−2000 mL day−1) under comparable experimental conditions (Figure 1a). This suggests that hexane biodegradation could have been inhibited partially due to the presence of metabolites excreted as a result of the increased availability of hexane mediated by the addition of SO200. At this point, it is 4064

dx.doi.org/10.1021/es204144c | Environ. Sci. Technol. 2012, 46, 4059−4066

Environmental Science & Technology

Article

foaming suggests that biosurfactants/bioemulsifier production was limited when microorganisms were confined in the NAP. Therefore, this study suggested that foaming also depends on the specific interactions between the microbial community supporting VOC biodegradation and the NAP. The mechanisms controlling the diffusion of nutrients, water (and therefore pH), and metabolites between the aqueous phase and the NAP were not assessed in this work. However, these issues deserve further research. In summary, the performance of the three microbial consortia evaluated in the absence of SO200 was similar and in accordance with the low solubility of hexane in water (low hexane transfer rates). When SO200 was added to the systems, the steady state hexane ECs increased by a factor of 8.7 and 16.3 in the case of Consortia A and B, respectively. However, Consortium C supported a first steady state with a 2-fold increase in ECs followed by a 16-fold EC increase after the hydrophobicity shift. Our results suggest that the ability of Consortia B and C to grow immersed in SO200 (where the hexane solubility is ∼13 000 times higher than in the aqueous phase), as a result of their hydrophobic nature, supported an enhanced substrate bioavailability (by using the entire potential of the gas-NAP transfer pathway). This study revealed that cell hydrophobicity can play a key role in the successful performance of TLPBs, and to the best of our knowledge, this is the first report on hydrophobic VOC treatment with exclusive VOC uptake within a nonbioavailable NAP. Finally, it was shown that metabolite accumulation might inhibit TLPB performance despite the presence of SO200.

important to highlight that despite hexane biodegradation deteriorated, microbial growth at the expenses of previously excreted metabolites and also due to a significant hexane removal is still possible and might explain the final increase in turbidity (Figure 4). In addition, the final ECs obtained in this experiment (Figure 4) were 2 times lower than those obtained in the experiment conducted with Consortium B (Figure 1) with similar TOC concentrations at the end (around 100 mg L−1), which ruled out the hypothesis of inhibition in the experiment carried out with consortium B and confirmed the key role of the cell hydrophobicity in this study. As a result of the large D required in these systems to avoid metabolite accumulation and therefore a metabolite-mediated inhibition, an alternative operation mode including a settler in series with the bioreactor might be used. The settler implementation would allow for the phase separation (biomass containing NAP from the aqueous broth) and consequently a continuous exchange of mineral salt medium could be performed. This integrated process has recently been proposed by Darraq et al.25 Further experiments in gastight serum bottles also supported the hypothesis of a deteriorated hexane biodegradation performance due to metabolite accumulation. The hexane biodegradation rate at a ratio 30:0 (broth from bioreactor: fresh medium) was similar to that at ratio of 15:15, while hexane degradation in fresh mineral medium (ratio 0:30) was significantly (p > 0.05) higher than at the two former ratios (Figure S6, Supporting Information). A preliminary characterization of the aqueous phase according to the extraction protocol given by Bordel et al.,26 allowed the identification of hexanol and acetic acid as intermediates of the hexane biodegradation. Hexanol is a common metabolite of the aerobic hexane biodegradation pathway,27 whose partitioning coefficient NAP-water is ∼8, which indicates that hexanol was preferably dissolved in the NAP. However, the presence of acetic acid in the aqueous phase could be the responsible of the pH drop. It is important to stress that a potential inhibition due to hexane accumulation was ruled out for all consortia since the TOC values in the aqueous phase were much higher than the maximum aqueous hexane concentration (0.03 mg L−1) at an inlet gaseous hexane concentration of 2 g m−3. However, the increased loading experiments showed that the systems were limited by mass transfer in the presence of silicone oil, which means that hexane concentration was 0 in the aqueous phase for consortium A and 0 in the NAP for consortia B and C. Likewise, a potential inhibition by low pH values was ruled out since the initial pH of the three test media ranged from 6.5 to 7 and the experiments only lasted for 4 h. Foaming is one of the most important operational problems reported in TLPBs constructed with liquid NAPs such as 2undecanone, dodecane, hexadecane, and silicone oil.28−30 In our particular study, while hexane biodegradation by Consortium A was characterized by an intense foaming at high EC values, which was partially controlled by antifoam addition (maximum concentration of 250 μL/Lreactor). Consortium B supported a foam-free degradation with the subsequent benefit on process operation, which suggests that biosurfactants/ bioemulsifier production was not significant in this particular experiment. However, Consortium C was characterized by an intense foaming following SO200 addition, which disappeared right after the hydrophobicity shift. This rapid dissapearance of



ASSOCIATED CONTENT

S Supporting Information *

Figure 5S shows the TLPB in operation during hexane biodegradation by Consortium B at day 14 (a), by Consortium C at day 11 (b), and by Consortium C at day 17 after the hydrophobicity shift (c). Table S2 shows the metabolite accumulation rate in the aqueous phase and percentage of hexane converted to aqueous metabolites as TOC for the three consortia evaluated at the four different dilution rates. This material is available free of charge via the Internet at http:// pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This research was supported by the Spanish Ministry of Science and Innovation (RYC-2007-01667 and BES-2007-15840 contracts, CTQ2009-07601 and CONSOLIDER-CSD 200700055 projects) and the Regional Government of Castilla y Leon (ref VA004A11-2). Dr. Marcia Morales is also gratefully acknowledged for kindly supplying the Pseudomonas aeruginosa BM-B-450 strain used in this investigation.



ABBREVIATIONS Dilution rate of the aqueous medium without microorganisms EC Elimination capacity NAP Nonaqueous phase D

4065

dx.doi.org/10.1021/es204144c | Environ. Sci. Technol. 2012, 46, 4059−4066

Environmental Science & Technology RE SO200 TLPBs TOC VOCs



Article

removal. Proceedings of Water & Industry International Conference; Valladolid, Spain, May 1−4.2011 (17) Hernandez, M.; Gouveia, J.; Quijano, G.; Thalasso, F.; Villaverde, S.; Muñoz, R. A systematic study of the influence on microbial kinetics of the presence of an organic phase during microbial isolation. Chem. Eng. Trans. 2010, 23, 375−380, DOI: 10.3303/ CET1023063. (18) Hernandez, M.; Torre, R. M. Long-term influence of the presence of a non-aqueous phase on the cell surface hydrophobicity in two-phase partitioning bioreactors. Appl. Microbiol. Biotechnol. 2011, 89 (5), 1573−1581, DOI: 10.1007/s00253-010-2975-7. (19) Galaction, A. I.; Cascaval, D.; Oniscu, C.; Turnea, M. Enhancement of oxygen mass transfer in stirred bioreactors using oxygen-vectors. 1. Simulated fermentation broths. Bioprocess. Biosyst. Eng. 2004, 26 (4), 231−238, DOI: 10.1007/s00449-004-0353-5. (20) Neumann, G.; Cornelissen, S.; Breukelen, F. V.; Hunger, S.; Lippold, H.; Loffhagen, N.; Wick, L. Y.; Heipieper, H. J. Energetics and surface properties of Pseudomonas putida DOT-T1E in a twophase fermentation system with 1-decanol as second phase. Appl. Environ. Microbiol. 2006, 72 (6), 4232−4238, DOI: 10.1128/ AEM.02904-05. (21) Quijano, G.; Revah, S.; Gutierrez-Rojas, M.; Flores-Cotera, L. B.; Thalasso, F. Oxygen transfer in three-phase airlift and stirred tank reactors using silicone oil as transfer vector. Proc. Biochem. 2009, 44 (6), 619−624, DOI: 10.1016/j.procbio.2009.01.015. (22) Muñoz, R.; Arriaga, S.; Hernandez, S.; Guieysse, B.; a, Revah, S. Enhanced hexane biodegradation in a two phase partitioning bioreactor: Overcoming pollutant transport limitations. Proc. Biochem. 2006, 41 (7), 1614−1619, DOI: 10.1016/j.procbio.2006.03.007. (23) Aldrich, J. M.; Gillet, S.; Delvigne, F.; Blecker, C.; Lebeau, F.; Wathelet, J. P.; Manigat, G.; Thonart, P. Effect of surfactants and biomass on the gas/liquid mass transfer in an aqueous-silicone oil twophase partitioning bioreactor using Rhodococcus erythropolis T902.1 to remove VOCs from gaseous effluents. J. Chem. Technol. Biotechnol. 2009, 84, 1274−1283, DOI: 10.1002/jctb.2172. (24) Littlejohns, J. V.; Daugulis, A. J. Response of a solid-liquid twophase partitioning bioreactor to transient BTEX loadings. Chemosphere 2008, 73 (9), 1453−1460, DOI: 10.1016/j.chemosphere.2008.07.066. (25) Darracq, G.; Couvert, A.; Couriol, C.; Amrane, A.; Thomas, D.; Dumont, E.; Andres, Y.; Le Cloirec, P. Silicone oil: An effective absorbent for the removal of hydrophobic volatile organic compounds. J. Chem. Technol. Biotechnol. 2010, 85, 309−313, DOI: 10.1002/ jctb.2331. (26) Bordel, S.; Díaz, L. F.; Muñoz, R.; Villaverde, S. New insights on toluene biodegradation by Pseudomonas putida F1: Influence of pollutant concentration and excreted metabolites. Appl. Microbiol. Biotechnol. 2007, 74 (4), 857−866, DOI: 10.1007/s00253-006-0724-8. (27) Lee, E.-H.; Kim, J.; Cho, K.-S.; Ahn, Y. G.; Hwang, G.-S. Degradation of hexane and other recalcitrant hydrocarbons by a novel isolate, Rhodococcus sp. EH831. Environ. Sci. Pollut. Res. 2010, 17, 64− 77, DOI: 10.1007/s11356-009-0238-x. (28) Daugulis, A. J.; Janikowski, T. B. Scale-up performance of a partitioning bioreactor for the degradation of polyaromatic hydrocarbons by Sphingomonas aromaticivorans. Biotechnol. Lett. 2002, 24 (8), 591−594, DOI: 10.1023/A:1015080700477. (29) Janikowski, T. B.; Velicogna, D.; Punt, M.; Daugulis, A. J. Use of a two-phase partitioning bioreactor for degrading polycyclic aromatic hydrocarbons by a Sphingomonas sp. Appl. Microbiol. Biotechnol. 2002, 59 (2−3), 368−376, DOI: 10.1007/s00253-002-1011-y. (30) Yeom, S. H.; Daugulis, A. J. A two-phase partitioning bioreactor system for treating benzene-contaminated soil. Biotechnol. Lett. 2001, 23 (6), 467−473, DOI: 10.1023/A:1010337522094.

Removal efficiency Silicone oil 200 cSt Two-Liquid phase partitioning bioreactors Total organic carbon Volatile organic compounds

REFERENCES

(1) Devinny, J. S.; Deshusses, M. A.; Webster, T. S. Biofiltration for Air Pollution Control; Lewis Publishers, Boca Raton, 1999. (2) Zhu, X.; Suidan, M. T.; Pruden, A.; Yang, C.; Alonso, C.; Kim, B. J. Effect of substrate Henry’s constant on biofilters performance. J. Air Waste Manage. Assoc. 2004, 54 (4), 409−418. (3) Reardon, K. F.; Mosteller, D. C.; Bull Rogers, J. D. Biodegradation kinetics of benzene, toluene and phenol as single and mixed substrates for Pseudomonas putida F1. Biotechnol. Bioeng. 2000, 69 (4), 385−400, DOI: 10.1002/1097-0290(20000820) 69:43.0.CO;2-Q. (4) Daugulis, A. J. Two-phase partitioning bioreactors: A new technology platform for destroying xenobiotics. Trends Biotechnol. 2001, 19 (11), 457−462, DOI: 10.1016/S0167-7799(01)01789-9. (5) Quijano, G.; Rocha-Rios, J.; Hernandez, M.; Villaverde, S.; Revah, S.; Muñoz, R.; Thalasso, F. Determining the effect of solid and liquid vectors on the gaseous interfacial area in two-phase partitioning bioreactors. J. Hazard. Mater. 2010, 175 (1−3), 1085−1089, DOI: 10.1016/j.jhazmat.2009.10.020. (6) Arriaga, S.; Muñoz, R.; Hernandez, S.; Guieysse, B.; Revah, S. Gaseous hexane biodegradation by fusarium solani in two liquid phase packed bed and stirred-tank bioreactors. Environ. Sci. Technol. 2006, 40 (7), 2390−2395, DOI: 10.1021/es051512m. (7) Arriaga, S.; Revah, S. Improving hexane removal by enhancing fungal development in a microbial consortium biofilter. Biotechnol. Bioeng. 2005, 90 (1), 107−115, DOI: 10.1002/bit.20424. (8) Quijano, G.; Hernandez, M.; Thalasso, F.; Muñoz, R.; Villaverde, S. Two-phase partitioning bioreactors in environmental biotechnology. Appl. Microbiol. Biotechnol. 2009, 84 (5), 829−846, DOI: 10.1007/ s00253-009-2158-6. (9) Collins, L. D.; Daugulis, A. J. Biodegradation of phenol at high initial concentrations in two-phase partitioning batch and fed-batch bioreactors. Biotechnol. Bioeng. 1997, 55 (1), 155−162, DOI: 10.1002/ (SICI)1097-0290(19970705)55:13.3.CO;2-N. (10) Nielsen, D. R.; Daugulis, A. J.; Mclellan, P. J. Transient performance of a two-phase partitioning bioscrubber treating a benzene contaminated gas stream. Environ. Sci. Technol. 2005, 39 (22), 8971−8977, DOI: 10.1021/es051158s. (11) Hernández, M.; Quijano, G.; Thalasso, F.; Daugulis, A. J.; Villaverde, S.; Muñoz, R. A comparative study of solid and liquid nonaqueous phases for the biodegradation of hexane in two-phase partitioning bioreactors. Biotechnol. Bioeng. 2010, 106 (5), 731−740, DOI: 10.1002/bit.22748. (12) Rocha-Rios, J.; Quijano, G.; Thalasso, F.; Revah, S.; Muñoz, R. Methane biodegradation in a two-phase partition internal loop airlift reactor with gas recirculation. J. Chem. Technol. Biotechnol. 2011, 86 (3), 353−360, DOI: 10.1002/jctb.2523. (13) Ascon-Cabrera, M. A.; Lebeault, J. M. Selection of xenobioticdegrading microorganisms in a biphasic aqueous-organic system. Appl. Environ. Microb. 1993, 59 (6), 1717−1724. (14) Ascon-Cabrera, M. A.; Lebeault, J. M. Cell hydrophobicity influencing the activity/stability of xenobiotic-degrading microorganisms in a continuous biphasic aqueous−organic system. J. Ferment. Bioeng. 1995, 80 (3), 270−275, DOI: 10.1016/0922-338X(95)90828N. (15) MacLeod, C. T.; Daugulis, A. J. Interfacial effects in a two-phase partitioning bioreactor: degradation of polycyclic aromatic hydrocarbons (PAHs) by hydrophobic. Mycobacterium. Process. Biochem. 2005, 40 (5), 1799−1805, DOI: 10.1016/j.procbio.2004.06.042. (16) Lebrero, R.; Estrada, J. M.; Muñoz, R. Odour abatement in biotrickling filters: effect of EBRT on Methyl Mercaptan and VOCs 4066

dx.doi.org/10.1021/es204144c | Environ. Sci. Technol. 2012, 46, 4059−4066