Alternated phenol and trichloroethylene biodegradation in an aerobic granular sludge reactor

Biochemical Engineering Journal 106 (2016) 1–10

Contents lists available at ScienceDirect

Biochemical Engineering Journal journal homepage: www.elsevier.com/locate/bej

Regular article

Alternated phenol and trichloroethylene biodegradation in an aerobic granular sludge reactor Yi Zhang a,∗ , Joo Hwa Tay b a b

Department of Environmental Science and Engineering, Fudan University, 220 Handan Road, Yangpu District, Shanghai 200433, China Department of Civil Engineering, University of Calgary, Calgary AB T2N 1N4, Canada

a r t i c l e

i n f o

Article history: Received 13 June 2015 Received in revised form 25 October 2015 Accepted 30 October 2015 Available online 5 November 2015 Keywords: Biodegradation Bioreactor Immobilized cells Wastewater treatment Aerobic granule Co-metabolism

a b s t r a c t Co-metabolic removal of a synthetic solvent, trichloroethylene (TCE) was conducted. Phenol was used as the primary substrate, and aerobic granules as the biocatalyst. An airtight reactor was designed and constructed with glass, and the configuration was verified of its applicability with volatile solvent. Phenol and TCE were fed into the reactors alternately, and both were efficiently removed during 6 weeks of operation. Compared with the control reactor receiving no TCE, biomass in the TCE-fed reactors had lower phenol-dependent specific oxygen utilization rate, but higher concentration, bigger size, and better settling ability. The morphology of the biomass in TCE and control reactors exhibited distinct characteristics. Granules in the TCE reactors slightly broke up early in the operation, but generally retained their shape and structure throughout. The biomass in the control reactor, however, lost its granular structure and totally disintegrated into flocs. Therefore, TCE co-metabolism likely improved the structural integrity of aerobic granular sludge, and co-metabolic degradation of TCE by phenol-grown aerobic granules showed long term stability. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Aerobic granule is a relatively new form of immobilized cells, usually cultivated in fully aerated column reactors operated under sequencing batch mode [1]. Aerobic granular sludge (AGS) is characterized by its near spherical shape, compact structure, high density and settling speed [2]. AGS reactors can retain abundant biomass within the reactors. In addition, granules’ size and structure provide diffusion barrier to the inner layer cells [3]. As a result, AGS reactor can withstand high concentrations and stressful loadings of toxic substances [4]. For example, various researchers have successfully used AGS reactors to treat toxic compounds like phenol [5], p-nitro phenol [6], phthalic acid [7], and halogenated phenols [8–10]. These substances were all treated as the sole or partial carbon and energy sources. However some pollutants, usually synthetic, cannot sustain microbial growth. Sometimes they can be removed via non-growth linked mechanisms, such as co-metabolism [11]. Trichloroethylene (TCE), a synthetic solvent and major groundwater pollutant, is often studied as the model compound for aerobic co-metabolism. It can hardly support aerobic microbial growth,

∗ Corresponding author. Fax: +86 21 65643597. E-mail address: [email protected] (Y. Zhang). http://dx.doi.org/10.1016/j.bej.2015.10.026 1369-703X/© 2015 Elsevier B.V. All rights reserved.

but can be effectively transformed by cultures grown on methane [12], ammonia [13], toluene [14] and phenol [15]. The non-specific enzymes synthesized in the degradation of these primary substrates, e.g. methane monooxyganse, ammonia monooxygenase, toluene mono- and di-oxygenases, and phenol hydroxylase, can also fortuitously catalyze TCE epoxidation, which would lead to its further degradation and mineralization [16]. However, several obstacles hinder the application of aerobic co-metabolism in TCE removal. One is the exhaustion of reducing power, as TCE catalysis exerts a net drain on intracellular NAD(P)H pool [17]. Another factor is the TCE product toxicity (TPT). Aerobic co-metabolism of TCE generates highly reactive intermediates or byproducts that can bind to the non-specific enzymes and other cellular materials, causing structural change and function loss [18]. The worst scenario resulted in cell death without potential of recovery. Therefore the ratio of the amount of TCE transformed to the biomass before total inactivation is defined as “transformation capacity” (TC ) [18], the determination of which could take several hours to several days. To overcome these obstacles, it is necessary that an energy generating substrate is provided to the microorganisms, to replenish the reducing power and to regenerate the damaged cellular materials. However, if the primary substrate and TCE are present together, competitive inhibition might occur, as they are catalyzed by the same enzyme. Therefore it is proposed that physical or temporal separation should be imposed on TCE

2

Y. Zhang, J.H. Tay / Biochemical Engineering Journal 106 (2016) 1–10

and the primary substrate, that they be degraded in two different reactors, or in one reactor operated in a sequential mode [19]. For example, Segar et al. used phenol as the growth substrate, and tested various combinations of rejuvenation frequency, rejuvenation duration, hydraulic retention time and phenol concentrations in a sequencing batch biofilm reactor [20]. They found that appropriate strategies were necessary to maintain enzyme activity and active biomass profile. As AGS has shown good potential in toxic substance removal, it is of practical interest to explore its application in aerobic TCE cometabolism. AGS reactor has the advantages of small footprint, easy operation and good biomass-liquid separation [21]. In addition, they have mostly been operated in SBR (sequencing batch reactor) modes. By adding a phase of non-aerated TCE co-metabolism, alternative primary substrate and TCE degradation could easily be achieved. In previous studies, AGS was formed on phenol as the growth substrate [22], and the kinetics, rate limiting factors and the toxic effect of TCE co-metabolism on the phenol-grown AGS were systematically studied [23–24]. The objective of this study is, therefore, to further explore the applicability of AGS in

TCE co-metabolism. An AGS reactor was designed and constructed, and operated in a mode where phenol and TCE are alternatively degraded. The feasibility, limitations, and performance of the reactor, and characteristics of the granules were studied to evaluate the efficiency and stability of the system.

2. Materials and methods 2.1. Reactor setup Several concerns need to be addressed when designing an AGS reactor with phenol as the growth substrate and TCE as the cosubstrate. TCE is highly volatile and dissolves plastic materials, therefore the reactor and any materials directly in contact with its content should ensure air tightness and solvent compatibility. Out of potential materials, glass was chosen for its light weight, low cost, transparency, and ease to make into different shapes. A glass column was therefore made, with an inner diameter of 40 mm, and a height of 600 mm, giving it a max. working volume of 750 mL. A lid was fitted at the top, with the contact surface grounded to

Fig. 1. Schematic of reactor setup. = Viton tubing, — silica tubing.

300

(a)

250 200 150 100 50 0

0

50

100

cycle time (min)

phenol concentration (mg L-1)

phenol concentration (m g L-1)

Y. Zhang, J.H. Tay / Biochemical Engineering Journal 106 (2016) 1–10

300

3

(b)

250 200 150 100 50 0

0

50

100

cycle time (min)

Fig. 2. Phenol degradation cycle in the reactors. (a) Day 1; (b) day 41; (䊏) TCE reactor 1; (䊉) TCE reactor 2; () control reactor. The data represent the average of duplicated samples.

ensure air tightness. A schematic drawing of the reactor set-up is shown in Fig. 1. A second concern is the feeding/withdraw of liquid and gas into/from the reactor. Phenol solution with nutrient supplements was fed with peristaltic pump from a plastic container. TCE, on the other hand, was fed with a syringe pump (New Era NE-1600) and gas-tight glass syringes (25 mL). The pump was programmed to deliver the syringes’ content in an intermittent and timed fashion, without causing leakage or creating headspace therein. In the phenol degradation phase, dissolved O2 and mixing was provided by aeration. For this purpose, a glass sinter disc was fixed at the bottom of the reactor, and its pore size was 100 ∼ 150 ␮m resulting in air bubbles of 1 ∼ 2 mm. However, in TCE degradation, aeration would cause stripping loss and must be avoided. Therefore, a liquid/gas mixture recirculation was chosen as the means to mix the content and replenish O2 . In addition, a one-way valve was installed in the air inlet line, and an electrical valve in the exit line, to prevent TCE loss due to diffusion along the tubing. The status (open/close) of the latter was controlled by its power supply (on/off). During phenol degradation the valve was switched on, and in the TCE degradation phase, it was kept off to seal the headspace of the reactor. In all the lines directly in contact with the reactor’s content, Viton tubing (Masterflex) was used as it is compatible with solvent and peristaltic pumps. For other lines silica tubing was used. Those connectors used on Viton tubing were either stainless steel or brass, and those on ordinary silica tubing were plastic. Another concern is the sampling for the sludge, and phenol and TCE concentrations. Two sets of sampling ports were installed on the reactors, one of which were Teflon valves with stoppers of 2 mm aperture size, for sludge and phenol sampling. The other was for TCE, made by melting down the bottom of a 2 mL HPLC sample bottle (Waters) and sintering it on the reactor. The bottles were fitted with 9 mm screw caps (La-Pha-Pack, with centre holes and PTFE coated septa) to seal the outlets and to ensure air-tightness. The reactors were housed in a temperature-controlled room (25 ◦ C), and the final configuration can be seen in Fig. S1. 2.2. Seed sludge Mature granules were taken from a 2 L lab scale AGS reactors that had been used to cultivate and stabilize granules for more than 4 months. The detailed granulation process was described in a previous study [22], where activated sludge was obtained from a municipal wastewater treatment plant and acclimated to phenol of 250 mg L−1 . The acclimated sludge was put in a column shape reactor of 5 cm inner diameter, and fed with phenol as the sole carbon source. It was operated in 4 h cycles and the settling time

was gradually reduced from 15 min to 3 min. At around 80 days of operation, aerobic granules started to form, and the biomass size increased from 100 ␮m to 600 ␮m. The mature granules obtained had a sludge volume index (SVI) of 20–30 mL g MLSS−1 and specific oxygen utilization rate (SOUR) 100–200 mg O2 g MLVSS−1 h−1 (with 250 mg L−1 phenol). Haldane’s equation was used to simulate phenol kinetics, with the parameters Vmax = 1.6 mg phenol g MLSS−1 min−1 , KS = 37 mg L−1 , KI = 462 mg L−1 . Michaelis-Menten kinetics described TCE transformation with the parameters of Vmax = 0.0142 mg TCE g MLSS−1 min−1 and KS = 1.02 g L−1 . In addition, the granules had a TC of approx. 11 mg g MLSS−1 , and 2.5 mol chloride ions was released per mol TCE transformed, indicating nearly total mineralization. Excessive granules were harvested from three parallel reactors operated stably, stored under 4 ◦ C for future use. In this study, they were mixed and washed with a phosphate buffer (see Section 2.5), then put into the glass reactors at a concentration of around 3 g MLSS L−1 . The first 10 days were operated without TCE, to acclimate and stabilize the granules to the smaller reactors. The operation mode was identical to that would be used in the TCE degradation operation. TCE injection was then initiated, and the TCE-phenol alternative degradation operation was conducted for 41 days. 2.3. Chemicals and media Phenol stock solution was prepared at the concentration of 50 g L−1 (Merck, Germany). TCE stock solution was made by thoroughly mixing 2 mL pure TCE (J.T. Baker) with 60 mL Milli Q water to result in the saturation concentration of 1.1 g L−1 (25 ◦ C). For phenol feeding solution, the influent contained (in g L−1 ) NH4 Cl 0.405; KH2 PO4 0.081; Na2 HPO4 0.081; MgSO4 ·7H2 O 0.048 and FeSO4 ·7H2 O 0.03, and 500 mg L−1 phenol was supplied as the sole carbon and energy source. Its pH was adjusted to over 7.5 by 1 M NaHCO3 . The phosphate buffer had the composition of (in g L−1 ): KH2 PO4 0.694, K2 HPO4 0.854, (NH4 )2 SO4 1.234, MgSO4 ·7H2 O 0.86, CaSO4 ·2H2 O 0.176 and FeSO4 ·7H2 O 0.001. 1 mL of trace elements solution as described previously [24] was supplemented to the reactor influent or the phosphate buffer, respectively. Technical grade chemicals and tap water were used to make up the reactor influent. 2.4. Operation of the phenol-TCE AGS reactor Three reactors were used, one as the control where no TCE was added, and the other two received TCE dosages (TCE 1 and 2 as duplicates). The operation volume of the reactors was 500 mL, and

4

Y. Zhang, J.H. Tay / Biochemical Engineering Journal 106 (2016) 1–10

TCE concentration (mg L-1)

2.50

(a)

2.00

1.50 1.00 0.50 0.00

0

5

10

15

20

25

operaon me (d)

30

35

40

TCE concentration (mg L-1)

2.5

45

(b)

2 1.5

purge on 1 0.5 0

0

20

40

60 cycle me (min)

80

100

120

Fig. 3. TCE degradation in the reactors. (a) Initial and effluent TCE concentrations during the reactor operation. (䊏) Initial concentration in the liquid phase; (䊉) initial concentration in the gas phase; (×) effluent concentration in the liquid phase; () effluent concentration in the gas phase. (b) A TCE degradation cycle. (䊏) concentration in the liquid () concentration in the headspace. The data represent the average of duplicated samples.

the volumetric exchange ratio 0.5. The aeration rate during phenol degradation phases was 2.5–3 L min−1 . The three reactors were operated in the same sequencing batch mode, and the cycle time was 4 h. One cycle was divided to several phases: aeration-phenol degradation, recirculation-TCE addition and degradation, aerationresidue TCE elimination, settling, effluent withdraw, and phenol feeding. The lengths of each phase were: 120, 105, 5, 5, 3, and 2 min, respectively. The first two hours of the cycle were the phenol degradation stage when aeration was provided. A previous study [22] showed that at the initial phenol concentration of 250 mg L−1 , the influent phenol could be degraded within 45 min. The additional aeration was provided to ensure complete phenol removal, and also to maintain granule structure, as certain level of shear force was found to be essential to granule integrity [1]. Another aim was to obtain high O2 level in the mixed liquor at the end of this phase, after which the air pump and the electrical valve were switched off. TCE feeding was immediately started by the syringe pump, timed to deliver 2 mL TCE stock solution every 4 h at the rate of 1.0 mL min−1 . For the control reactor, the content in the respective syringe was replaced by distilled water. As soon as TCE feeding began, the recirculation pump was turned on. Gas/liquid mixture was drawn from near the surface of the reactor content and pumped into the reactor’s bottom. The aims of recirculation were first to enhance TCE mass transfer in the reactor and provide more contact between biomass and TCE, and second to replenish O2 into the aqueous phase. 2 mL of TCE was fed each

cycle, which equals to 2.2 mg or 0.018 mmol. To completely mineralize this amount of TCE, 0.027 mmol O2 (0.87 mg) is needed. Assuming the reactor content had a dissolved O2 (DO) concentration of 8 mg L−1 , totally 4 mg O2 would be contained in the liquid, which was more than adequate for TCE mineralization. The excessive DO could be used for biomass respiration, and the continuous recirculation of gas/liquid mixture could also replenish O2 from the reactor headspace should DO become insufficient. At the end of the TCE degradation phase, a short aeration period of 5 min was provided to blow off any possible TCE residue to an exhaust. As TCE is potentially hazardous, this was intended to prevent its direct contact with human. 5 min of settling period followed the aeration. The seed sludge used for the TCE—phenol alternative fed AGS reactor was aerobic granules taken from stably operated 2 L SBRs whose settling time was 2–3 min. The reasons for this longer settling time were that TCE degradation benefits from higher biomass concentration, and the change of environment for the granules might require less harsh operational conditions. 2.5. Sampling and analytical methods Three kinds of samples were obtained from the reactor: liquid sample for biomass and phenol, liquid sample for dissolved TCE, and gas samples for gas-phase TCE. The first kind of samples was taken by simply turning on the Teflon valves and collecting around 25 mL liquid outflow during the phenol degradation stages. The liquid TCE samples were obtained from the septa on the screw caps by a 1 mL

Y. Zhang, J.H. Tay / Biochemical Engineering Journal 106 (2016) 1–10

5

8.0

120

(a) 100 80

5.0 4.0

60

3.0

40

2.0 20

1.0 0.0

0

7

10

14 18 22 25 operation time (d)

31

36

0

41

(b)

600 SOUR (mg O2 g MLVSS-1 h-1)

SVI (mL g-1)

6.0

800 700

500

600 400

500

300

400 300

200

size (um)

MLSS concentration (g L-1)

7.0

200 100 0

100 0

7

10

14

22

25

31

36

41

0

operation time (d) Fig. 4. Change of biomass characteristics in the reactors. (a) Change of MLSS concentration and SVI value in the control and TCE reactors; (䊉) MLSS of TCE reactor; (䊏) MLSS of control reactor; (×) SVI of TCE reactor, () SVI of control reactor. (b) Change of SOUR and granule size during the operation; (䊉) SOUR in the TCE reactor; (䊏) SOUR in the control reactor; (×) size in the TCE reactor; () size in the control reactor.

gastight syringe (SGE). 0.5 mL liquid sample was drawn out and injected to a pre-sealed 2 mL sampling vial. 50 ␮L headspace of the vial was injected into gas chromatography (GC) and the TCE concentration was calculated as the total mass of TCE in the vial divided by the liquid volume (0.5 mL). The gas phase TCE concentration was also determined by GC, but the sampling was done by directly drawing 50 ␮L gas phase sample from the screw caps installed above the liquid level in the reactor. The tests were duplicated. The configuration and operation conditions of GC were as described previously [23,24], so was the measurement of phenol by colorimetric method. The analysis of sludge characteristics was done according to APHA standard methods [26], and the parameters measured included mixed liquor suspended solids (MLSS), volatile suspended solids (MLVSS), sludge volume index (SVI), and specific oxygen utilization rate (SOUR). In addition, sludge size and morphology were monitored with a laser particle size analysis system (Malvern Mastersizer 2600) and an image analysis system (Quantimet 500 image analyzer, Leica Cambridge Instruments). MLSS and MLVSS samples were duplicated, but SVI and SOUR were not.

2.6. Abiotic test Before the AGS reactors were operated for TCE removal, they were first tested without granular sludge under identical conditions. The purpose was to see if the conditions were suitable, e.g. if the reactor was gas-tight and no TCE was lost due to leakage or adsorption to reactor materials, and the recirculation was adequate for TCE mass transfer when aeration was turned off. The reactor was filled with 500 mL tap water and no biomass, and the test duration was 120 min. The recirculation pump was turned on, and the gas outflow valve was turned off. Changes of TCE concentrations in the reactor’s two phases, liquid and headspace, during a cycle are shown in Fig. S2. TCE concentration in the liquid phase fluctuated between 2 and 3 mg L−1 after the injection of 2 mL TCE feed solution. In the headspace, TCE concentration increased from 0.71 mg L−1 at the injection time to around 2 mg L−1 10 min later, and remained stable afterwards. At 120 min, aeration was turned on for 5 min to purge the reactor content. Sampling at 130 min showed no TCE residue in either liquid or headspace phase. As TCE concentrations in both the

6

Y. Zhang, J.H. Tay / Biochemical Engineering Journal 106 (2016) 1–10

liquid and headspace phases remained stable throughout the 2 h operation, it could be inferred that the reactor was gas-tight. There was no significant leakage at the tubing-reactor connections, the Teflon valves or the screw cap sampling ports. The design and setup of the reactors were proved to be valid for the application, and any possible TCE loss could be attributed to biological functions. In the subsequent study, TCE mass in the reactors was represented by its concentrations in both phases, instead of that in the liquid alone. 3. Results and discussion 3.1. Phenol degradation in the reactors Phenol degradation in the reactors was regularly monitored, both by measuring the concentration in the effluent and by studying degradation cycles. The phenol concentration profiles in the three reactors during one degradation cycle are shown in Fig. 2. Influent phenol could be degraded within 20 min, both on day 1 and day 41 of the operation. On day 1, the three curves of degradation were almost identical, while on day 41, the degradation rate in the control reactor (no TCE addition) seemed slightly lower. The initial concentration of phenol in the reactors was measured to be between 200 and 300 mg L−1 , averaged around 250 mg L−1 . Phenol concentration in the discharged liquid was universally below detection limit. TCE degradation was reported to have detrimental effect on biomass, by producing toxic intermediates or products that can inactivate cellular components [27]. One manifestation of the product toxicity was to lower the primary substrate degradation ability. For example, both ammonia and hydrazine-oxidizing activities of Nitrosomonas europaea was diminished with exposure to TCE [13]. Heterotrophic cultures grown on methane, toluene or phenol also exhibited limited TCE degradation, even with abundant supply of reducing energy [28]. However in this operation, reactors TCE 1 and 2 still showed remarkable phenol degradation activity after 40 days of TCE addition. The phenol degradation curves were comparable in three reactors. Therefore, the phenol/TCE alternated fed AGS reactors were proven to be sustainable. An important reason for this could be that, they were designed to operate far below the transformation capacity of the granules. Using an average biomass concentration of 5 g MLSS L−1 and the TC of 10 mg TCE g MLSS−1 , each reactor was calculated to be able to degrade 25 mg TCE before the TC was spent, while the daily dosage of TCE to the reactors was 12 mg TCE. Moreover, 250 mg L−1 phenol was provided each cycle to reactivate the granules from TCE product toxicity. The frequently switching between TCE degradation/inactivation and phenol degradation/reactivation, and the low TCE loading during each degradation phase very likely ensured that the biomass was kept in a healthy state. Other researchers confirmed the advantage of SBR mode in TCE removal [29], and the removal efficiency of the present system can be well expanded by exploiting the full transformation capacity of the granular sludge. 3.2. TCE degradation in the reactors Fig. 3 shows TCE degradation in the TCE reactors, including the fluctuation of the reactors’ influent and effluent TCE concentrations, and a sample of a TCE degradation cycle. The samples of initial TCE concentrations were taken 2–5 min after injection at the beginning of a TCE degradation cycle, when the reactor content was mixed to a degree. Fig. 3(a) shows that during six weeks of operation, the initial TCE concentrations, whether sampled from the liquid or gas phase, remained stable. The concentrations in the liquid fluctuated between 1.5 and 2.5 mg L−1 , and those in the headspace were slightly lower (around 1.0–1.5 mg L−1 ). The efflu-

ent concentrations were sampled at the end of the cycles, before aeration was turned on. The effluent concentrations in the liquid phase were mostly below detection limit, implying that almost all TCE in water was removed. However the headspace concentrations exhibited a fluctuation between 0 and 0.1 mg L−1 , which could be caused by inadequate mixing between the gaseous and aqueous phases, resulting in occasional residue TCE in the headspace. However, most time of the operation, the reactor achieved total TCE removal from both phases. A full cycle of TCE degradation can be seen in Fig. 3(b), confirming the TCE removal efficiency of the reactors. Nine samples from both the liquid and headspace were obtained during the cycle. The first one was taken 2 min after the injection, which showed TCE concentrations of 0.89 and 0.84 mg L−1 in the liquid and headspace, respectively. TCE concentrations increased to 2.15 and 1.55 mg L−1 at 10 min, indicating that a mixing period was needed for TCE to achieve equilibrium. They were followed by continuous decline in the next 90 min, and at 100 min were almost non-detectable. The fluctuation of TCE concentration in the headspace was more prominent than that in the liquid, indicating that the volatilization and solution of TCE between the liquid and headspace took place. It was likely that in this not-well mixed system, mass transfer of TCE was a rate limiting factor, as TCE in liquid was exhausted before that in gas. However under the present conditions, the reactor and process were able to completely remove TCE in the influent, as was evident by the TCE concentrations at the end of the cycle. The results achieved in this prototype glass reactor which utilized phenol as the growth substrate and aerobic granules as the biomass have shown promise for a bigger scale TCE co-metabolism process. The target TCE concentration is likely to be within the range of 0.1–20 mg L−1 in the liquid, taking into consideration the KS of the granules. Phenol could be provided at concentrations much lower that 250 mg L−1 to reduce growth substrate consumption. Materials other than glass could be adopted to construct the reactor, due to the low mechanical strength of glass. Better mixing and smaller headspace should be considered for future reactor design, as they have been found to be problematic in this study. In addition if the reactor size and material permit, mechanical mixing can be employed by placing a stirrer at the bottom to direct mix the liquid content. Gas recirculation by air pump is also an alternative or supplementary means to improve mass transfer between liquid and gas phases. 3.3. Biomass characteristics in the reactors The biomass in the control reactor and two TCE-degrading reactors was characterized during the operation duration. The results from the two TCE reactors were averaged, and the changes of MLSS concentration and SVI of the biomass are shown in Fig. 4(a), while those of the SOUR on phenol and the granular sludge size are shown in Fig. 4(b). Granular sludge was inoculated in the three reactors, and the starting MLSS concentrations were kept the same, around 3.2 g L−1 . MLVSS was 64% of MLSS, and SVI of the seed granules was 35 mL g MLSS−1 . The SOUR and endogenous respiration rate were 104 and 14.3 mg O2 g MLVSS−1 h−1 , respectively, and the size was approx. 950 ␮m. The morphologies of the biomass can be seen in Fig. 5, where one picture each of the seed granule and biomass at day 0 were given. Other photos compare the biomass from control and TCE reactors at various points in the operation. The inoculum for the TCE-phenol alternative degradation reactors were granules that had been cultivated in 2 L, 5 cm inner diameter reactors and stored under 4 ◦ C for 2 months. Therefore before the TCE degradation test began, a 10 day acclimation period was provided to reactivate the granules from the low temperature storage and to acclimate them to a different reactor configuration (500 mL working volume, 4 cm inner diameter). After 10 days

Y. Zhang, J.H. Tay / Biochemical Engineering Journal 106 (2016) 1–10

7

Fig. 5. Change of biomass morphologies in both TCE and control reactors.

of adaptive operation, MLSS concentrations nearly doubled (from 3.2 g L−1 to 5.3 and 6.7 g L−1 in the TCE and control reactors, respectively) and the MLVSS to MLSS ratio increased to 89%. Meanwhile, SOUR rose to around 200 mg O2 g MLVSS−1 h−1 , which, together with the changes in MLSS and MLVSS, strongly indicated that the reactivation of the granules from cold storage was successful. The granules fully recovered their active biomass fraction and phenol degradation ability. However during this period, SVI value increased from 35 to 100 and 70 mL g MLSS−1 (TCE and control respectively), and the mean size dropped to around 600 ␮m. It sug-

gested that the storage and change in the reactor configuration might have some negative effect on granule integrity and settling ability. When operation continued, SVI dropped to below that of the seed granules at day 7, and size stabilized, indicating the acclimation of granules to the smaller reactor was achieved. After TCE addition started, MLSS concentrations in control and TCE reactors were similar for the first 25 days, which fluctuated between 4 and 6 g L−1 and averaged around 5.3 g L−1 . However in the last 15 days of operation, MLSS in the control reactor showed a decreasing trend compared to that in the TCE reactors,

8

Y. Zhang, J.H. Tay / Biochemical Engineering Journal 106 (2016) 1–10

Table 1 Comparison of TCE degradation activities by biofilm reactors. Growth substrate

Culture

TCE concentration

Rate constants

Reference

Phenol Methane Methane Propane Methane Methanol Phenol Toluene Glucose Phenol Toluene Butane Phenol

Consortium, pulse resting Consortium, continuous growing Consortium, continuous growing Consortium, continuous or pulse feeding Consortium, growing Methylosinus trichosporium PP358, continuous growing Consortium, continuous growing Consortium, continuous growing Burkholderia cepacia PR123, continuous growing Pseudomonas putida BCRC 14,349, pulse resting Consortium, continuous growing Consortia, continuous growing Consortium, intermittent resting

0.1 mg L−1 in liquid 1 mg L−1 in liquid 0.01–5.1 mg L−1 in liquid 20 mg L−1 pulse in liquid 0.02–1.2 mg L−1 in liquid 0.5–3 mg L−1 in liquid 3–9 mg L−1 in the liquid 0.135 mg L−1 in the liquid 0.4–2.4 mg L−1 in the air 0.2–20 mg L−1 0.1–0.16 mg L−1 0.4–0.5 mg L−1 2 mg L−1 in the liquid

First order: 2–4 L g VSS−1 h−1 First order: 1.2 h−1 First order: 0.40–5.4 h−1 Max rate: 0.003 h−1 First order: 0.0004 h−1 Max rate: 0.5 h−1 NDa Max rate: 0.01 h−1 Max rate: 0.0001–0.011 h−1 c First order: 0.7 L g VSS−1 h−1 Max rate: 8.5 g m−3 h−1 Max rate: 0.002 h−1 b 0.0006 h−1 c

[20] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] This study

a

Not determined. Assuming that 1 g biomass contains 0.5 g protein. c Estimated with the data of the 1st hour of degradation, under initial TCE concentration of 1.5–2 mg L−1 in liquid and gas phases, using average SS concentration in the TCE reactors of 5.37 g SS L−1 . b

which remained stable. MLSS at the end of the operation were 4.5 and 5.3 g L−1 for the control and the TCE reactor, respectively. SVI in both reactors dropped sharply in the first week to around 30 mL g MLSS−1 , and were almost identical in the next week. However from day 14, a steady increase of SVI in the control reactor was observed, which reached the highest value of 97 mL g MLSS−1 at day 31 and decreased to 60 mL g MLSS−1 in the next 10 days. In the meantime, SVI in the TCE reactors was on a steady decline, and dropped to 26 mL g MLSS−1 at the end of the operation. Accompanying these phenomena, biomass morphology also exhibited changes, more drastically in the control reactor. Granules in both TCE and control reactors showed sign of disintegration, e.g. the mean size decreased during the 10 days of acclimation. However, in the TCE reactor, biomass retained the round and entire shape at the end of the operation, and smaller granules were observed, probably formed from the disintegrated biomass particles. In the control reactor, however, biomass continued to disintegrate. At day 41 the sludge therein mainly consisted of flocs, and hardly any granules could be observed. Therefore, the granules in the control reactor began to deteriorate around 2 weeks into the operation, and the decreased settling ability caused wash out of biomass, hence the drop in MLSS and rise in SVI. This observation is further supported by the change of biomass size. At day 0, the means size was 624 and 588 ␮m in the control and TCE reactors, respectively. Subsequently this parameter showed different trends in the two reactors, except in the first week when it decreased to 563 (control) and 537 ␮m (TCE). Biomass size in the control reactor was on a continuous decline, and reached 383 ␮m at the end of the operation. In the TCE reactor, the size remained stable for 25 days, and increased in the last 2 weeks. The mean size in the TCE reactor at day 41 was 522 ␮m, comparable to day 0. In terms of the biomass’ phenol degradation activity, before day 10, SOUR in both reactors increased in a similar fashion, from around 200 to 564 (control) and 475 (TCE) mg O2 g MLVSS−1 h−1 . However, from day 10 to the end of the operation, SOUR in the TCE reactors was declining. Concurrently, SOUR in the control reactor remained stable before day 22, and decreased mildly afterwards. At the end, SOUR in the TCE reactor was 93 mg O2 g MLVSS−1 h−1 , even below that of the seed granules, and that in the control reactor was 271 mg O2 g MLVSS−1 h−1 . The higher SOUR of the control reactor can probably be associated with its biomass morphology, which was almost all smaller flocs. A previous study has shown that, phenol and TCE degradation activities of the biomass were negatively correlated with biomass size, and smaller flocs and granules exhibited much higher degradation rates [23].

The biomass in the TCE reactors exhibited lower phenol degradation activity, which might not be a disadvantage. Lower primary substrate degradation rate likely resulted in slower TCE transformation, which also entails slower production of toxin(s), giving the biomass time to recover. Though repetitively affected by TCE product toxicity, granules in the TCE reactors were able to retain structural integrity, and reasonable levels of MLSS and SVI. The sustainability of the reactor seemed to have been improved by TCE addition. Other researchers had observed the morphology change of activated sludge in a TCE-phenol alternatedly fed SBR [30]. It was found that compared to reactor receiving phenol only, the phenolTCE reactor produced biomass of larger size and apparent better flocculation. The flocs in the phenol-TCE reactor showed sign of auto-aggregation, and lower but more stable phenol-hydroxylase activity. These findings coincide with the results obtained in this study, that granules in the TCE reactor were better preserved of their structural integrity, and showed lower phenol-dependent SOUR than the control reactor. 3.4. TCE degradation reactors with attached-growth cells Various bioreactors for TCE co-metabolism have been tested, encompassing various operational modes (continuous or batch), primary substrates (aliphatic or aromatic) and culture state (growing or resting, pure or mixed, dispersed or attached). As this study focuses on aerobic granules, a form of attached-growth cells, some previous experiments on TCE degradation by mixed or pure culture biofilm are summarized in Table 1 for comparison. Methane, propane, butane, phenol and toluene had been used as aliphatic and aromatic primary substrates. Methanol and glucose were provided not as the inducers of the non-specific enzymes but growthsupporting and energy generating substances, probably to avoid competitive inhibition between the primary substrate and TCE. The TCE concentration used in this study was well within the range of those generally employed. The initial rate of TCE removal was calculated based on the cycle data shown in Fig. 3, using linear regression on the decline of total TCE mass in the reactor. Though in the lower range, the rate is comparable to some of the studies (e.g. methane and butane oxidizers), indicating a moderate ability in sustainable TCE removal. However it should be noted that the capacity of the present AGS system is not sufficiently explored, and the operation was far from optimized. Higher TCE concentration and/or more frequent addition can be used to further improve the efficiency of the system, and other phenol concentrations can be tested. In addition, an SBR for TCE co-metabolism using phenol as the growth substrate, mixed culture aerobic granules as the bio-

Y. Zhang, J.H. Tay / Biochemical Engineering Journal 106 (2016) 1–10

catalyst has several potential advantages. First, phenol itself is an organic pollutant, and using it as the primary substrate could achieve the removal of two pollutants in a single reactor. Second, phenol degraders generally exhibit moderate TCE degradation rate, but high growth rates and high tolerance to its toxicity [25]. Mixed culture, though usually not as efficient as pure cultures, is much easier and less costly to maintain. Aerobic granular sludge as a form of attached-growth cells can harbor a great diversity of microbes with different physiological traits, kinetic properties and responses to TCE and phenol toxicity [42]. This is a huge advantage, as it enables the system to be versatile and robust under different growth conditions and substrates loadings, where various microbial populations would shift and dominate. This advantage is further enhanced by the SBR mode, which is highly flexible. Cycle lengths, hydraulic retention time, as well as sludge retention time can be independently adjusted. The frequency, strength and duration of phenol/TCE feeding can be individually controlled to meet varying needs of substrate degradation. Last but not least, SBR mode also generates dynamic concentration profiles in the reactor, which seemed to be beneficial to TCE removal. Shih et al. [29] compared the effects of four phenol feeding patterns on TCE removal, and found reactors operated in continuous or semi-continuous pattern produced cells much less capable of TCE degradation, and biomass taken from the pulse-fed reactor gave the best performance. Similar phenomena were observed by Futamata et al. [43], that periodical feeding of phenol resulted in complete degradation of TCE. 4. Conclusions A glass SBR reactor was designed to facilitate alternated phenol and TCE degradation by phenol-grown aerobic granules. The reactors were tested for possible leakage and adsorption, and then inoculated with phenol-grown aerobic granules cultured in larger 2 L reactors. 10 days of acclimation period was provided, followed by 6 weeks of TCE-phenol alternately fed test. Phenol degradation in all three reactors (two TCE, one control) was found to be complete, and the phenol in the influent was degraded within 30 min. The two TCE-fed reactors also showed good TCE degradation activity, and were able to achieve 90–95% TCE removal efficiency during the entire test duration, including TCE present in the headspace. The granules in the TCE reactors exhibited lower SOUR and SVI value, and higher MLSS and size than that in the control reactor. These differences were probably caused by TCE addition, as granules in the TCE reactors were able to fully retain their structural integrity, round shape, and good settling ability after 6 weeks of operation. However, granules in the control reactor totally disintegrated and almost no granular structure was observed at the end of test. Therefore the reactor and process design were proven to be capable of alternated phenol and TCE degradation, and the process was sustainable and efficient. Acknowledgements The authors would like to express their sincere appreciation for Dr. Stephen Tay Tiong-Lee (School of Civil and Environmental Engineering, Nanyang Technological University, deceased) for his kind help and encouragement. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.bej.2015.10.026. References

9

[1] J.H. Tay, Q.S. Liu, Y. Liu, The effect of shear force on the formation, structure and metabolism of aerobic granules, Appl. Microbiol. Biotechnol. 57 (2001) 227–233. [2] S.S. Adav, D.J. Lee, K.Y. Show, J.H. Tay, Aerobic granular sludge: recent advances, Biotechnol. Adv. 26 (2008) 411–423. [3] Z.C. Chiu, M.Y. Chen, D.J. Lee, C.H. Wang, L.Y. Lai, Oxygen diffusion in active layer of aerobic granule with step change in surrounding oxygen levels, Water Res. 41 (2007) 884–892. [4] A.M. Maszenan, Y. Liu, W.J. Ng, Bioremediation of wastewaters with recalcitrant organic compounds and metals by aerobic granules, Biotech. Adv. 29 (2011) 111–123. [5] H.L. Jiang, J.H. Tay, S.T.-L. Tay, Aggregation of immobilized activated sludge cells into aerobically grown microbial granules for the aerobic biodegradation of phenol, Lett. Appl. Microbiol. 35 (2002) 1–7. [6] S. Yi, W.Q. Zhuang, B. Wu, S.T.-L. Tay, J.H. Tay, Biodegradation of p-nitrophenol by aerobic granules in a sequencing batch reactor, Environ. Sci. Technol. 40 (2006) 2396–2401. [7] P. Zeng, W.Q. Zhuang, S.T.-L. Tay, J.H. Tay, The influence of storage on the morphology and physiology of phthalic acid-degrading aerobic granules, Chemosphere 69 (2007) 1751–1757. [8] A. Carruci, S. Milia, G. de Gioannis, M. Piredda, Acetate-fed aerobic granular sludge for the degradation of 4-chlorophenol, J. Hazard. Mater. 166 (2009) 482–490. [9] A.F. Duque, V.S. Bessa, M.F. Carvalho, M.K. de Kreuk, M.C.M. van Loosdrecht, P.M.L. Castro, 2-fluorophenol degradation by aerobic granular sludge in a sequencing batch reactor, Water Res. 42 (2011) 6745–6752. [10] S.G. Wang, X.W. Liu, H.Y. Zhang, W.X. Gong, X.F. Sun, B.Y. Gao, Aerobic granulation for 2,4-dichlorophenol biodegradation in a sequencing batch reactor, Chemosphere 69 (2007) 769–775. [11] M. Alexander, Cometabolism, in: Biodegradation and Bioremediation, second ed., Academic Press, San Diego, 1999, pp. 249–268. [12] L. Alvarez-Cohen, P.L. McCarty, Effect of toxicity: aeration and reductant supply on trichloroethylene transformation by a mixed methanotrophic culture, Appl. Environ. Microbiol. 57 (1991) 228–235. [13] M.R. Hyman, S.A. Russell, R.L. Ely, K.J. Williamson, D.J. Arp, Inhibition, inactivation, and recovery of ammonia-oxidizing activity in cometabolism of trichloroethylene by Nitrosomonas europaea, Appl. Environ. Microbiol. 61 (1995) 1480–1487. [14] S. Heald, R.O. Jenkins, Trichloroethylene removal and oxidation toxicity mediated by toluene dioxygenase of Pseudomonas putida, Appl. Environ. Microbiol. 60 (1994) 4634–4637. [15] B.R. Folsom, P.J. Chapman, P.H. Pritchard, Phenol and trichloroethylene degradation by Pseudomonas cepacia G4: kinetics and interactions between substrates, Appl. Environ. Microbiol. 56 (1990) 1279–1285. [16] A.R. Harker, Y. Kim, Trichloroethylene degradation by two independent aromatic-degrading pathways in Alcaligenes eutrophus JMP134, Appl. Environ. Microbiol. 56 (1990) 1179–1181. [17] J. Dolfing, A.J. van den Wijingaard, D.B. Janssen, Microbiological aspects of the removal of chlorinated hydrocarbons from air, Biodegradation 4 (1993) 261–282. [18] L. Alvarez-Cohen, P.L. McCarty, A cometabolic biotransformation model for halogenated aliphatic compounds exhibiting product toxicity, Environ. Sci. Technol. 25 (1991) 1381–1387. [19] L. Alvarez-Cohen, P.L. McCarty, Two-stage dispersed growth treatment of halogenated aliphatic compounds by cometabolism, Environ. Sci. Technol. 25 (1991) 1387–1392. [20] R.L. Segar Jr., S.L. De Wys, G.E. Speitel Jr., Sustained trichloroethylene cometabolism by phenol-degrading bacteria in sequencing biofilm reactors, Water Environ. Res. 67 (1995) 764–774. [21] J.H. Tay, S.T.-L. Tay, Y. Liu, K.Y. Show, Biogranulation Technologies for Wastewater Treatment, first ed., Pergamon Press, Oxford, 2006. [22] Y. Zhang, J.H. Tay, Co-metabolic degradation activities of trichloroethylene by phenol-grown aerobic granules, J. Biotechnol. 162 (2012) 274–282. [23] Y. Zhang, J.H. Tay, Rate limiting factors in trichloroethylene co-metabolic degradation by phenol-grown aerobic granules, Biodegradation 25 (2014) 227–237. [24] Y. Zhang, J.H. Tay, Toxic and inhibitory effects of trichloroethylene aerobic co-metabolism on phenol-grown aerobic granules, J. Hazard. Mater. 286C (2015) 204–210. [25] H. Futamata, S. Harayama, K. Watanabe, Diversity in kinetics of trichloroethylene-degrading activities exhibited by phenol-degrading bacteria, Appl. Microbiol. Biotechnol 55 (2001) 248–253. [26] APHA, Standard Methods for the Examination of Water and Waste Water, 21st ed., American Public Health Association, Washington, D.C, 2005. [27] M.E. Dolan, M.L. McCarty, Methanotrophic chloroethene transformation capacities and 1,1-dichloroethene transformation product toxicity, Environ. Sci. Technol. 29 (1995) 2741–2747. [28] H.L. Chang, L. Alvarez-Cohen, Transformation capacities of chlorinated organics by mixed cultures enriched on methane, propane, toluene, or phenol, Biotechnol. Bioeng. 45 (1995) 440–449. [29] C.C. Shih, M.E. Davey, J.Z. Zhou, J.M. Tiedje, C.S. Criddle, Effects of phenol feeding pattern on microbial community structure and cometabolism of trichloroethylene, Appl. Environ. Microbiol. 62 (1996) 2953–2960. [30] H.L. Ayala-del-Río, S.J. Callister, C.S. Criddle, J.M. Tiedje, Correspondence between community structure and function during succession in phenol- and

10

[31]

[32] [33]

[34]

[35]

[36]

[37]

Y. Zhang, J.H. Tay / Biochemical Engineering Journal 106 (2016) 1–10 phenol-plus-trichloroethene-fed sequencing batch reactors, Appl. Environ. Microbiol. 70 (2004) 4950–4960. G.W. Strandberg, T.L. Donaldson, L.L. Farr, Degradation of trichloroethylene and trans-1,2-dichloroethylene by a methanotrophic consortium in a fixed-film, packed-bed bioreactor, Environ. Sci. Technol. 23 (1989) 1422–1425. S.E. Strand, J.V. Wodrich, H.D. Stensel, Biodegradation of chlorinated solvents in a sparged, methanotrophic biofilm reactor, Res. J. WPCF 63 (1991) 859–867. L.L. Phelps, J.J. Niedzielski, R.M. Schram, S.E. Herbes, D.C. White, Biodegradation of trichloroethylene in continuou-recycle expanded-bed bioreactors, Appl. Environ. Microbiol. 56 (1990) 1702–1709. E. Arvin, Biodegradation kinetics of chlorinated aliphatic hydrocarbons with methane oxidizing bacteria in an aerobic fixed biofilm reactor, Water Res. 25 (1991) 873–881. M.W. Fitch, G.E. Speitel, G. Georgiou, Degradation of trichloroethylene by methanal-grown cultures of Methylosinis trichosporium OB3b PP358, Appl. Environ. Microbiol. 62 (1996) 1124–1128. H.S. Shin, J.L. Lim, Performance of packed-bed bioreactors for the cometabolic degradation of trichloroethylene by phenol-oxidizing microorganisms, Environ. Technol. 17 (1996) 1351–1359. J.P. Arcangeli, E. Arvin, Modeling of the cometabolic biodegradation of trichloroethylene by toluene-oxidizing bacteria in a biofilm system, Environ. Sci. Technol. 31 (1997) 3044–3052.

[38] A.K. Sun, T.K. Wood, Trichloroethylene mineralization in a fixed-film bioreactor using a pure culture expressing constitutively toluene ortho-monooxygenase, Biotechnol. Bioeng. 55l (1997) 674–685. [39] Y.M. Chen, T.F. Lin, C. Huang, H.C. Lin, F.M. Hsieh, Degradation of phenol and TCE using suspended and chitosan-bead immobilized Pseudomonas putida, J. Hazard. Mater. 148 (2007) 660–670. [40] Y. Zhao, Z.J. Liu, F.X. Liu, Z.Y. Li, Cometabolic degradation of trichloroethylene in a hollow fiber membrane reactor with toluene as a substrate, J. Membr. Sci. 372 (2011) 322–330. [41] D. Frascari, G. Zanaroli, G. Bucchi, A. Rosato, N. Tavanaie, S. Fraraccio, D. Pinelli, F. Fava, Trichloroethylene aerobic cometabolism by suspended and immobilized butane-growing microbial consortia: a kinetic study, Bioresour. Technol. 144 (2013) 529–538. [42] Y. Zhang, J.H. Tay, Physiological and functional diversity of phenol degraders isolated from phenol-grown aerobic granules: phenol degradation kinetics and trichloroethylene co-metabolic activities. 2015, submitted. [43] H. Futamata, S. Harayama, A. Hiraishi, K. Watanabe, Functional and structural analyses of trichloroethylene-degrading bacterial communities under different phenol-feeding conditions: laboratory experiments, Appl. Microbiol. Biotechnol. 60 (2003) 594–600.