Evaluation and modeling of benzalkonium chloride inhibition and biodegradation in activated sludge

w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 1 2 3 8 e1 2 4 6

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/watres

Evaluation and modeling of benzalkonium chloride inhibition and biodegradation in activated sludge Chong Zhang a, Ulas Tezel b, Kexun Li a,*, Dongfang Liu a, Rong Ren a, Jingxuan Du a, Spyros G. Pavlostathis b a b

The College of Environmental Science and Engineering, Nankai University, Tianjin 300071, China School of Civil and Environmental Engineering, Georgia Institute of Technology, Atlanta, GA 30332-0512, USA

article info

abstract

Article history:

The inhibitory effect and biodegradation of benzalkonium chloride (BAC), a mixture of

Received 17 March 2010

alkyl benzyl dimethyl ammonium chlorides with different alkyl chain lengths, was

Received in revised form

investigated at a concentration range from 5 to 20 mg/L and different biomass concen-

30 August 2010

trations in an activated sludge system. A solution containing glucose and mineral salts was

Accepted 29 September 2010

used as the wastewater in all the assays performed. The inhibition of respiratory enzymes

Available online 7 October 2010

was identified as the mode of action of BAC as a result of oxygen uptake rate analysis performed at BAC concentrations ranging between 5 and 70 mg/L. The glucose degradation

Keywords:

in the activated sludge at different BAC and biomass concentrations was well-described

Benzalkonium chloride

with Monod kinetics with competitive inhibition. The half-saturation inhibition constant

Inhibition

(KI) which is equivalent to EC50 of BAC for the activated sludge tested ranged between 0.12

Biodegradation

and 3.60 mg/L. The high KI values were recorded at low BAC-to-biomass ratios, i.e. less than

Modeling

10 mg BAC/g VSS, at which BAC was almost totally adsorbed to biomass and not

Activated sludge

bioavailable. BAC degradation started as soon as glucose was totally consumed. Although BAC was almost totally adsorbed on the biomass, it was degraded completely. Therefore, BAC degradation was modeled using two-phase biodegradation kinetics developed in this study. This model involves rapid partitioning of BAC to biomass and consecutive degradation in both aqueous and solid phases. The aqueous phase BAC degradation rate was twenty times, on average, higher than the solid phase degradation rate. The specific aqueous (kI1) and solid (kI2) phase BAC utilization rate constants were 1.25 and 0.31 mg BAC/g VSS h, respectively. The findings of this study would help to understand the reason of extensive distribution of quaternary ammonium compounds in wastewater treatment plant effluents and in natural water systems although QACs are biodegradable, and develop strategies to avoid their release and accumulation in the environment. ª 2010 Elsevier Ltd. All rights reserved.

1.

Introduction

Benzalkonium chloride (BAC) is a mixture of alkyl benzyl dimethyl ammonium chlorides with C8 to C18 alkyl groups. BAC, which is a group of quaternary ammonium compounds

(QACs), is the active ingredient of many pharmaceutical formulations, cosmetics, commercial disinfectants, industrial sanitizers and food preservatives (Tezel and Pavlostathis, 2009). About 75% of the QACs consumed in domestic and industrial applications annually are released into wastewater

* Corresponding author. Tel./fax: þ86 22 23501117. E-mail address: [email protected] (K. Li). 0043-1354/$ e see front matter ª 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2010.09.037

w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 1 2 3 8 e1 2 4 6

treatment systems. BAC is the most frequently found QAC group worldwide in municipal wastewater at concentrations ranging between 20 and 300 mg/L (Martinez-Carballo et al., 2007; Clara et al., 2007). The QACs present in the wastewater upset activated sludge process (Boethling, 1984). The EC50 values for hexadecyl trimethyl ammonium bromide and dodecyl benzyl dimethyl ammonium chloride obtained from a respirometric assay conducted with activated sludge ranged between 10 and 40 mg/L (Reynolds et al., 1987). The EC50 of a mixture of alkyl trimethyl ammonium chlorides (C14e18) for unacclimated activated sludge determined based on the inhibition of [14C] glucose uptake was 28 mg/L (Larson and Schaeffer, 1982). Another study showed that didecyl dimethyl ammonium chloride inhibited the COD removal in a rotating biological contactor at concentrations above 20 mg/L and the biofilm was totally eliminated at 160 mg/L. A variety of physiologically different microorganisms participate in the wastewater treatment process, therefore the response of each species to QAC inhibition is expected to be different. For instance, QACs are particularly toxic to nitrifiers. Benzalkonium chloride was inhibitory to a mixed nitrifying culture at 10e15 mg/L with a non-competitive inhibition coefficient equal to 1.5 mg/L (Yang, 2007). Overall, these studies suggest that QACs are inhibitory to activated sludge microbial community at concentrations higher than what is found in the wastewater. However, sudden discharges of QACs resulting in temporarily high levels in treatment plants could upset plant function. BACs rapidly and strongly adsorb onto biomass or are biodegraded during the biological wastewater treatment. Therefore, adsorption and biotransformation are the main routes of BAC removal from the wastewater. Average removal up to 99% by means of adsorption and biodegradation is reported in wastewater treatment systems (Clara et al., 2007; Boethling, 1984). Microorganisms that utilize QACs as the carbon and energy source at high concentrations have been identified in the activated sludge. The majority of the QAC degraders in the activated sludge are classified in the genus Pseudomonas (Dean-Raymond and Alexander, 1977; Geftic et al., 1979; van Ginkel et al., 1992; Nishihara et al., 2000; Kaech and Egli, 2001; Nishiyama and Nishihara, 2002; Takenaka et al., 2007; Liffourrena et al., 2008). Other species that can catabolize various QACs are Xanthomonas sp. (DeanRaymond and Alexander, 1977) and Aeromonas sp. (Patrauchan and Oriel, 2003). Until recently, few studies had focused on the biotransformation/biodegradation of BAC (Patrauchan and Oriel, 2003; van Ginkel, 2004; Qin et al., 2005). According to the results of these studies, BAC biotransformation commences with the fission of the alkyl group from the quaternary nitrogen resulting in the formation of benzyl dimethyl amine as the first intermediate. Benzyl dimethyl amine is then converted to ammonia through either two demethylation followed by a debenzylation or a debenzylation followed by two demethylation processes. Although biodegradation potential and mechanism of BAC and other QACs have been elucidated, none of the studies presented above reported the biodegradation kinetics of QACs. In spite of the fact that, the information on the inhibitory effects and biodegradation of BAC as well as its adsorption to activated sludge is present in the literature, the interaction of these processes, i.e. adsorption, inhibition and biodegradation,

1239

and how it affects the overall fate of BAC in the activated sludge is not well understood. Given the toxicity of BAC to aquatic organisms and the role in the induction of antibiotic resistance in the environment (Gaze et al., 2005), BAC has to be removed completely in the wastewater treatment systems (i.e., activated sludge) before wastewater and the residual (i.e., sludge) are discharged to the environment. As the ultimate biodegradation of BAC is the main goal, BAC inhibition and biodegradation kinetics in activated sludge systems need to be well-understood. The objectives of this study were to: (a) investigate the potential inhibitory effect and biodegradation of BAC in activated sludge; and (b) develop a comprehensive dynamic model to elucidate the fate and effect of BAC in activated sludge. All the experiments were carried out with a mixed aerobic culture at a range of BAC and biomass concentration.

2.

Materials and methods

Details on the a) properties and characterization of benzalkonium chloride; b) mixed aerobic heterotrophic culture used in all assays; c) analytical methods; d) model simulations and parameter estimation; and e) adsorption kinetics and isotherm assays are given in the Supplementary Material (Text S1eS5).

2.1.

Respirometric assay

Inhibition of BAC in activated sludge was investigated based on the oxygen uptake rate. A 100 mL sample of mixed aerobic heterotrophic culture in the endogenous growth phase was transferred into a series of Erlenmeyer flasks. A glucose solution, which served as carbon/energy source, and BAC at desired concentrations were added and the total liquid volume was adjusted to 100 mL with culture media. The glucose COD in the bottles was about 300 mg/L. The culture series included six bottles that were amended with BAC resulting in total BAC concentrations of 5, 10, 20, 30, 50 and 70 mg/L. Two additional flasks were prepared: seed blank and reference which consisted of only seed and culture media and seed, culture media, and glucose (300 mg COD/L), respectively. Dissolved oxygen in each flask was measured during the time course using a DO meter while the content was continuously mixed. The oxygen uptake rate (OUR) of each culture at different BAC concentrations was determined by calculating the slope of DO versus time curve using a linear regression performed by using Sigma Plot Version 10 software (Systat Software Inc., San Jose, CA, USA). The specific oxygen uptake rate (SOUR) of each culture was determined by normalizing the OUR to the volatile suspended solids (VSS) concentration in each individual flask. At the end of the assay, BAC concentration in each flask was measured to verify that BAC was not degraded during the course of the assay.

2.2. Batch inhibition assay using the mixed aerobic culture The inhibitory effect of BAC on glucose utilization and BAC biodegradation in the mixed heterotrophic culture was tested

1240

w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 1 2 3 8 e1 2 4 6

in a batch assay. The assay was performed in 1.5-L glass reactors stirred with a Teflon-coated stirring bar and aerated with compressed air. A sample of mixed liquor from the mixed aerobic heterotrophic culture was transferred to each reactor. Glucose and NH4Cl were added as the carbon and nitrogen source. The initial total glucose concentration was about 300 mg COD/L in the reactor. The cultures were then amended with BAC resulting in a total initial BAC concentration of 5, 10, and 20 mg/L, respectively. The total liquid volume was adjusted to 1 L with the mineral media. The initial pH in the culture was 7.0 and the reactor was maintained at 25  C. During the incubation period, the DOC and BAC concentration was measured at pre-specified time intervals. pH, TSS, VSS were measured at the beginning and at the end of the incubation period. Another batch assay testing the effect of biomass concentration on the inhibition and biodegradation of BAC in the mixed heterotrophic culture was performed using the same methodology described above. The biomass concentration in each reactor was adjusted by diluting the stock mixed culture with mineral media. The biomass concentration tested in this study ranged from 180 to 1300 mg VSS/L and the BAC concentration applied was either 5 or 10 mg/L. DOC and BAC concentration, VSS and pH were measured during the test period. All the assays described above were performed in duplicate. A dynamic model delineating the effect of BAC on substrate (glucose) utilization and BAC degradation was developed using the results of the assays described above. The model simulations and parameter estimation procedures used are given in Text S4.

2.3.

A DISSOLVED OXYGEN (mg/L)

8

3.

Results and discussion

3.1.

BAC inhibition assessment via oxygen consumption

The impact of BAC on the oxygen uptake rate of the mixed heterotrophic culture was assessed at different BAC concentrations ranging from 5 to 70 mg/L (Fig. 1(A)). The dissolved oxygen present in the reference culture in which there was no BAC was depleted in less than 25 min. The SOUR of the reference culture was measured as 49 mg O2/g VSS h which indicates that the culture was at the exponential growth phase. Although, nitrifiers were present in the activated sludge, their population was low enough to assume that the major fraction of the oxygen is consumed by the heterotrophic population (Fig. S1). The SOUR of the cultures decreased exponentially as the BAC concentration increased. The SOUR approached to that of seed culture, which was amended with neither glucose nor BAC, at the highest BAC concentration tested which indicates that the culture was at the endogenous respiration phase and did not utilize the added glucose (Fig. 1(B)).

6

4 BAC Conc. (mg/L) 0 30 5 50 10 70 20

2

0 6

8 10 12 14 16 18 20 22

TIME (Min) SOUR (mg O2/g VSS.hr)

B

50 40 30 20 10 0 0

20

40

60

80

BAC CONC. (mg/L)

Adsorption kinetic and isotherm assays

An adsorption kinetic assay was carried out to determine the time required for the adsorption of BAC to reach equilibrium. Subsequently, an adsorption isotherm assay was conducted to determine the BAC adsorption capacity of the activated sludge. Both assays are described in detail in Text S5.

blank

Fig. 1 e The profile of dissolved oxygen consumption (A) and specific oxygen uptake rate of activated sludge used in this study at different BAC concentrations (0e70 mg/L).

The effective BAC concentration that reduces the SOUR to half of the reference culture SOUR (EC50) is calculated as 22 mg/ L Boethling (1984) reported a range between 20 and 50 mg/L as the EC50 value of BAC for acclimated and unacclimated activated sludge in his review on cationic surfactants. Moreover, the EC50 of BAC for Pseudomonas putida was reported as 6 mg/L (Sutterlin et al., 2008). The results of the respirometric assay performed in this study revealed that BAC affects oxygen uptake rate therefore the primary mode of action of BAC in activated sludge is the inhibition of the respiratory enzymes. Inhibition of other terminal electron accepting processes (TEAPs) such as denitrification at the same concentration range was recently reported (Tezel and Pavlostathis, 2009).

3.2.

Modeling BAC inhibition and biodegradation

Glucose utilization as well as BAC biodegradation by activated sludge at 5, 10 and 20 mg/L BAC was investigated in another batch assay. The time at which half of the glucose was utilized (t1/2) in the reference culture which did not receive any BAC was calculated as 0.6 h (Fig. S2). The glucose utilization slowed

w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 1 2 3 8 e1 2 4 6

down at the higher BAC concentrations (Fig. 2). The glucose consumption t1/2 values were 1.6, 2.3 and 10.2 h at 5, 10 and 20 mg BAC/L. On the other hand, the BAC concentration was almost constant and equal to the initial concentration during the utilization of glucose (Fig. 2). Following the utilization of the major fraction of the glucose present in the cultures, BAC degradation was initiated. Given that the major fraction of the microbial community used in this study was composed of heterotrophs, and nitrifiers cannot degrade BAC (Yang, 2007), BAC was consumed by the heterotrophic microbial community in the activated sludge which also consumed glucose. The

A

400

200

15

Glucose BAC

10

Model

100

5

GLUCOSE (mg COD/L)

0

0

B

400

20

300

15

200

10

100

5

0

BAC CONC. (mg/L)

300

20

0

C

400

20

300

15

200

10

100

5

0

0 0 10 20 30 40 50 60 70 80

TIME (hr) Fig. 2 e Observed and simulated glucose and BAC utilization profiles in the cultures amended with (A) 5, (B) 10, and (C) 20 mg BAC/L at 500 mg VSS/L (Error bars represent one standard deviation of the means). In glucose utilization simulations; k [ 0.41 g COD/g VSS h, Ks [ 22 mg COD/L, Y [ 0.6 g COD/g COD and b [ 0.0025 hL1 was used and KI was estimated. In BAC utilization simulations, k, Ks, Y and b were kept constant and (A) KI [ 0.41 mg/L and kI1, kI2, KSI were estimated, (B) KI [ 0.30 mg/L, kI1 [ 0.0013 mg BAC/mg VSS-h, KSI [ 0.6 mg/L and, kI2 was estimated, (C) KI [ 0.12 mg/L, kI1 [ 0.0013 mg BAC/mg VSS-h, KSI [ 0.6 mg/L, and kI2 was estimated.

1241

calculated half-life of BAC in the cultures at 5, 10 and 20 mg/L BAC was 20.7, 21.5 and 36.9 h. The t1/2 value for BAC degradation are strongly correlated (r2 ¼ 0.999) to the t1/2 value for glucose utilization (Fig. S3). This result supports two conclusions: (1) the initial delay in the BAC degradation observed in all cultures was not related to acclimation; given that the inoculum used in this study was obtained from a wastewater treatment facility serving a very complex industrial area, the probability of the microbial community to have been exposed to QACs is high; and (2) the delay in the glucose degradation caused by BAC inhibition was the major reason for the retardation in the BAC degradation. An adsorption kinetic assay was performed in order to evaluate the dynamics of BAC partitioning in the activated sludge. The liquid phase BAC concentration reached equilibrium in half an hour which indicates that BAC sorption to the activated sludge at the concentration used was instantaneous (Fig. S4). The rapid attainment of equilibrium is consistent with previously published reports on the adsorption of quaternary ammonium compounds on a variety of municipal sludge (Ismail et al., 2010). Adsorption of BAC on activated sludge was investigated at BAC concentrations up to 70 mg/L. The Freundlich isotherm was used to model the equilibrium (Fig. S5). The estimated values for KF, capacity factor/sorption affinity and n, Freundlich exponent were 42.1  1.4 (mg/g VSS)(L/mg)n and 0.25  0.01, respectively (r2 ¼ 0.995). Similar constants for adsorption of BAC on activated sludge were reported in other studies (Ismail et al., 2010; Garcia et al., 2006). The Freundlich isotherm model represented well the BAC adsorption on activated sludge. The results of BAC adsorption kinetic and isotherm assays suggest that BAC is rapidly and extensively adsorb on activated sludge. Based on the adsorption isotherm, the calculated equilibrium liquid phase BAC concentration in the cultures at 5, 10, and 20 mg BAC/L was 0.003, 0.02 and 0.32 mg/L, respectively. Given that above 99% of BAC was adsorbed on the activated sludge and the desorption of adsorbed BAC was almost negligible under the test conditions (Ismail et al., 2010), the biodegradation of BAC proceeds both in the liquid and the solid phases. By using the facts demonstrated above which are: (1) BAC inhibits respiratory enzymes; (2) BAC degradation starts after the utilization of the major fraction of readily degradable COD; (3) BAC instantaneously and extensively partitions to the activated sludge; and (4) BAC gets degraded in both the liquid and solid phases, and assuming that all microbial cells in the activated sludge community are capable of BAC degradation, a model composed of four ordinary differential Eq. (1)e(4) and one algebraic Eq. (5) was developed. The Monod equation (Rittmann and McCarty, 2001) is the foundation of the model and used for modeling COD substrate utilization Eq. (1), and BAC utilization Eqs. (3) and (4) as well as the microbial growth Eq. (3). A novel approach was developed to model the biodegradation of BAC in the activated sludge and a schematic of the concept is given in Fig. 3. Based on this approach, BAC is instantaneously adsorbed on the biomass and gets degraded at different rates in the liquid Eq. (3) and the solid phases Eq. (4). We assumed that the same enzyme or group of enzymes

1242

w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 1 2 3 8 e1 2 4 6

BAC

product kI1

adsorption

aqueous phase

kI2

BAC

biomass

enzyme

Fig. 3 e Conceptual model of BAC degradation in activated sludge.

catalyzes the BAC degradation in the liquid and the solid phase, however other processes such as diffusion in the membrane, also may affect the degradation of biomass-sorbed BAC. Therefore, kI2 Eq. (4) is expressed as the observed biodegradation rate and may be composed of true maximum specific utilization rate plus membrane migration (diffusion within the membrane) rate. Biodegradation of dodecyl trimethyl ammonium chloride at different rates in the liquid and solid phases of a sediment slurry was previously demonstrated in another study (Shimp and Young, 1988), which also indicates that such a phenomenon occurs not only in biological reactors but also other natural environments in which QACs sorb. Therefore, combining phase distribution and biodegradation kinetics in such a way demonstrated above would contribute to a better understanding of the dynamics of pollutants in both engineered and natural systems. The Freundlich isotherm equation is incorporated in the model to calculate total BAC concentration as it was degraded in both the liquid and solid phases. An inhibition factor, (1 þ I/ KI), was included in both Eqs. (1) and (2) to reflect competitive inhibition as the inhibition mechanism for BAC which was justified by fact (1) given above. A switching factor, KS/(KSþS), was included in Eqs. (3) and (4) to reflect the substrate competition which was justified by fact (2) listed above. The biomass growth on BAC was assumed to be very small compared to the growth on glucose therefore this term (YdI/ dt) was neglected in Eq. (2). dS kSX  ¼  dt Ks 1 þ I þ S

(1)

dX kSX   bX ¼Y  dt Ks 1 þ I þ S

(2)

dCe kI1 Ce X Ks ¼ dt KsI þ Ce Ks þ S

(3)

dqe kI2 qe X Ks ¼ dt KsI þ qe Xinit Ks þ S

(4)

I ¼ qe Xinit þ Ce

(5)

KI

KI

In the above equations, S is the glucose concentration (mg COD/L), X is the active biomass concentration (mg VSS/L), Ce is the liquid phase BAC concentration (mg/L), qe is the solid phase BAC concentration (mg BAC/g VSS), I is the total BAC concentration (mg BAC/L). The parameters used in the model equations include: k, maximum specific glucose utilization

rate constant (mg COD/mg VSS h); KS, glucose half-saturation coefficient (mg COD/L); Y, true yield coefficient (g VSS/g COD); b, biomass decay rate constant (h1); KI, “observed” inhibition coefficient (mg BAC/L); kI1, maximum specific liquid phase BAC utilization rate constant (mg BAC/mg VSS h); kI2, “observed” solid phase BAC utilization rate constant (mg BAC/ mg VSS h); KSI, BAC half-saturation coefficient (mg BAC/L); and Xinit, initial biomass concentration (g VSS/L). Before simulating the effect of BAC and its biodegradation in the activated sludge, the key parameters, i.e. k, KS, Y and b, of Monod growth equations were estimated. The estimation of each parameter was done by using the glucose consumption profile in the reference culture (Fig. S2). The range for each parameter value was limited by typical parameter values reported for activated sludge (Tchobanoglous et al., 2003). The RMSD of the fit was 24.7 (9.7% of the initial conc.) and the estimated values for each parameter are given in Table 1. The estimated parameter values were kept as constants for the rest of the simulations. The identifiability of each parameter estimated was determined using local sensitivity functions obtained by Sensitivity Toolbox of Berkeley-Madonna (Gujer, 2008) (Fig. S6). According to the sensitivity analysis, sensitivity is largest for k which indicated that a minor change in the k would have the largest effect on the model output S. From the visual inspection of the sensitivity figure (Fig. S6), it was obvious that the sensitivity of Y and b had exactly the same form, while k and Ks were different. Given that Y and b were the parameters describing mainly the biomass growth, the change of one of Y or b can be compensated by an appropriate adjustment of the other. Thus, these two parameters could not be identified uniquely from the data used. On the other hand, the values estimated for k and Ks were absolute. The identifiability of Y and b was less of concern in this study because the sensitivity of S to these parameters dominated after the major fraction of the substrate was utilized. Moreover, curve fitting performed with 15 randomly selected initial estimate values for each of four parameters within the constraints specified in Table 1 resulted in the estimation of the same value for Y and b. Overall, Y and b could be identifiable within the constraints used in the curve fitting. The glucose and BAC utilization at different BAC concentrations were simulated and KI, kI1 and kI2 were estimated using the glucose and BAC utilization profiles in the cultures amended with 5, 10 and 20 mg BAC/L (Fig. 2). The estimation was performed in two steps.

Table 1 e Estimated model parameters and previously reported range of typical parameter values used for parameter estimation in this study. Parameter k, g COD/g VSS h Ks, mg COD/L Y, g VSS/g COD b, h1 c2

Estimated value

Typical value rangea

0.41  0.06 22  19.8 0.6  0.9 0.0025  0.321 3086

0.08e0.41 10e60 0.3e0.6 0.0025e0.0060

a (Tchobanoglous et al., 2003).

1243

w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 1 2 3 8 e1 2 4 6

2009; van Ginkel and Kolvenbach, 1991; Ying, 2006). On the other hand, Tezel (2009) reported more than an order of magnitude higher degradation rate for a BAC, tetradecyl benzyl dimethyl ammonium chloride (C14BDMA-Cl) in a BACenriched culture. The Monod-type specific C14BDMA-Cl utilization rate constant for the BAC enrichment culture, which utilizes BAC as the sole carbon and energy source for almost three years, was 0.03 mg C14BDMA-Cl/mg VSS h. The BACenriched culture is mainly composed of Pseudomonas spp. which is the primary species known to degrade various types QACs. Pseudomonas spp. accounts for 2e12% of the activated sludge microbial community (Dias and Bhat, 1964). The low BAC utilization rate constant estimated by assuming all of the

A

400 300 200

8 Glucose BAC

6

Model

4

100

2

0

0

B

400

In all simulations the previously estimated parameter values at each BAC concentration were kept constant. The initial estimation of the aforementioned parameters, i.e. KSI, kI1 and kI2, was done using the BAC utilization profiles of the culture amended with 5 mg BAC/L (Fig. 2(A)). The estimated KSI, kI1 and kI2 values were 0.6 mg/L, 0.0013 mg BAC/mg VSS h and 0.00028 mg BAC/mg VSS h, respectively (RMSD: 0.15 (2.9%)). The sensitivity analysis revealed that each parameter was identified uniquely from the data sets used (Fig. S7). The liquid phase BAC degradation was twenty times faster than the solid phase BAC degradation (Fig. S8). However, given the high adsorption affinity of BAC, the solid phase transformation is the limiting step in the BAC removal in activated sludge systems. The specific BAC utilization rate constants are consistent but lower than the first order liquid and solid phase dodecyl trimethyl ammonium chloride, a monoalkonium chloride (MAC), degradation rate constants obtained in experiments performed using sediments which were 0.0032  0.0008 h1 and 0.0009  0.0002 h1, respectively (Shimp and Young, 1988). The lower rate constants obtained in the present study are attributed to the type of QAC used which was confirmed by other studies indicating that BACs are less biodegradable than the MACs (Tezel and Pavlostathis,

10 8

300

6 200 4 100

2

0

3.2.2. Step 2 e estimation of BAC biodegradation constants KSI, kI1 and kI2

10

BAC CONC. (mg/L)

Step 1 e estimation of inhibition constant, KI

Since the BAC concentration was almost constant throughout the glucose utilization period, KI was estimated using only Eqs. (2) and (3) in the first step (Fig. 2). During KI estimations, only glucose consumption profiles at different BAC concentrations were used (Fig. 2). Since only KI was estimated, its value was identified uniquely from the data sets used for curve fitting. Altogether three KI values were obtained for cultures tested at the three BAC concentrations. These KI values were 0.383  0.015, 0.292  0.014 and 0.120  0.006 mg/L at 5, 10 and 20 mg BAC/L, respectively. The corresponding RMSD (the values in parentheses represent the coefficient of variation with respect to the initial concentration) of the fits was 7.9 (2.6%), 10.7 (3.4%) and 17.1 (5.2%), respectively. The mean KI was calculated as 0.28  0.15 mg/L. The mean KI value for BAC is at least two orders of magnitude lower than what is observed for conventional activated sludge which is attributed to the lower biomass concentration (450e650 mg VSS/L) used in our experiments compared to conventional activated sludge process (ca. 2000 mg VSS/L). On the other hand, inhibitory concentrations ranging from 0.2 to 6 mg QAC/L were reported for dilute activated sludge systems (Boethling, 1984). However, the biomass concentrations used in these studies were unknown. In addition, Microtox was used to determine the acute inhibitory concentration of the BAC mixture used in our experiments. The 5-min and 15-min EC50 values were 0.22 mg/ L (r2 ¼ 0.95, and 95% confidence range ¼ 0.17e0.27 mg/L) and 0.14 mg/L (r2 ¼ 0.88, and 95% confidence range ¼ 0.10e0.20 mg/ L). These results suggest that the mean KI obtained in the batch inhibition assay is the “minimum inhibitory concentration” for the activated sludge tested. KI may increase as the biomass concentration increases. This phenomenon is discussed in a subsequent section.

GLUCOSE (mg COD/L)

3.2.1.

0

C

400

10 8

300

6 200 4 100 0

2 0 0 10 20 30 40 50 60 70 80

TIME (hr) Fig. 4 e Observed and simulated glucose and BAC utilization profiles in the cultures amended with 5 mg BAC/L at (A) 615, (B) 394 and (C) 179 mg VSS/L (Error bars represent one standard deviation of the means). In glucose utilization simulations; k [ 0.41 g COD/g VSS h, Ks [ 22 mg COD/L, Y [ 0.6 g COD/g COD and b [ 0.0025 hL1 was used and KI was estimated. In BAC utilization simulations, k, Ks, Y and b were kept constant and (A) KI [ 2.04 mg/L, kI1 [ 0.0013 mg BAC/mg VSS-h, KSI [ 0.6 mg/L and kI2 was estimated, (B) KI [ 0.38 mg/L, kI1 [ 0.0013 mg BAC/mg VSS-h, KSI [ 0.6 mg/ L, and kI2 was estimated, (C) KI [ 0.14 mg/L, kI1 [ 0.0013 mg BAC/mg VSS-h, KSI [ 0.6 mg/L, and kI2 was estimated.

w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 1 2 3 8 e1 2 4 6

3.3. Effect of BAC-to-biomass ratio on inhibition and biodegradation The toxicity of BAC varies depending on the biomass concentration in the activated sludge most probably due to the extent of adsorption. Therefore, BAC-to-biomass ratio plays a crucial role in identifying inhibition and assessing BAC degradation since inhibition directly affects the BAC degradation by prolonging the half-life of the substrate COD utilization. The glucose and BAC utilization was tested at various biomass concentrations ranging from 179 to 1280 mg VSS/L and at 5, 10 and 20 mg BAC/L resulting in a BAC:VSS ratio ranging between 8 and 38 mg/g in a series of batch assays (Figs. 4 and 5, Fig. S9). The KI and kI2 values were estimated for each set of data following the two steps described above. The glucose and BAC utilization profiles presented in Figs. 4 and 5 were used for the estimation of these parameters. The estimated KI (mg/L) and kI2 (mg BAC/mg VSS h) values in the cultures having 615, 394 and 179 mg VSS/L and amended with 5 mg BAC/L were 2.04 and 0.0002, 0.38 and 0.0003, and 0.14 and 0.0003, respectively (Fig. 4). These parameter values in the cultures having 1280, 701 and 437 mg VSS/L and amended with 10 mg BAC/L were 3.61 and 0.0002, 0.56 and 0.0003, and 0.23 and 0.0004, respectively (Fig. 5). At both BAC concentrations, kI2 was constant having a mean value of 2.89  0.73  104 mg BAC/mg VSS-h. On the contrary, KI was high at the highest VSS concentration tested at both BAC concentrations and in a range of 0.14e0.56 mg/L at the rest of VSS concentrations tested. A comprehensive plot was created by using the estimated KI and kI2 as well as the BAC adsorption isotherm at different BAC-to-biomass (BAC:VSS) ratios tested in this study (Fig. 6). Multi-dimensional analysis of the data obtained in this study

A

400 300 200

20 15

Glucose BAC

10

Model

100

5

0

0

B

400

20

300

15

200

10

100

5

0

BAC CONC. (mg/L)

species present in the activated sludge are capable of BAC degradation may suggest that only about 4% (the ratio of BAC utilization rates of activated sludge and BAC enrichment community) can degrade BAC. The profiles of glucose and BAC utilization in the cultures at 10 and 20 mg BAC/L were simulated and only kI2 was estimated at each BAC concentration using the previously estimated parameter values as constants. The estimated kI2 values were 0.0004 (RMSD ¼ 0.24 (2.4%)) and 0.0013 (RMSD ¼ 0.63 (3.2%)) mg BAC/mg VSS h, respectively. Thus, the estimated kI2 values increased and approached the kI1 value as the BAC concentration increased from 5 to 20 mg/L. As it was discussed above, kI2 was a lumped parameter which may represent both biodegradation and membrane migration. The BAC concentration may affect the dominance of one or the other. For instance, the membrane migration process is the rate limiting step at low BAC concentrations at which BAC sorption is heterogeneous through the biomass surface, thus there is a concentration gradient on the biomass. On the contrary, the membrane migration process diminishes at high BAC concentrations at which the biomass surface is saturated, and the biodegradation rate in the solid phase, therefore, becomes equal to that in the liquid phase. The adsorption isotherm presented in this study supports that saturation of biomass was reached at 20 mg BAC/L (Fig. S5). The phenomenon presented here is discussed in detail below.

GLUCOSE (mg COD/L)

1244

0

C

400

20

300

15

200

10

100

5

0

0 0 10 20 30 40 50 60 70 80

TIME (hr) Fig. 5 e Observed and simulated glucose and BAC utilization profiles in the cultures amended with 10 mg BAC/L at (A) 1280, (B) 701 and (C) 437 mg VSS/L (Error bars represent one standard deviation of the means). In glucose utilization simulations; k [ 0.41 g COD/g VSS h, Ks [ 22 mg COD/L, Y [ 0.6 g COD/g COD and b [ 0.0025 hL1 was used and KI was estimated. In BAC utilization simulations, k, Ks, Y and b were kept constant and (A) KI [ 3.61 mg/L and kI1, kI2, KSI were estimated, (B) KI [ 0.56 mg/L, kI1 [ 0.0013 mg BAC/mg VSS-h, KSI [ 0.6 mg/L and, kI2 was estimated, (C) KI [ 0.23 mg/L, kI1 [ 0.0013 mg BAC/mg VSS-h, KSI [ 0.6 mg/L, and kI2 was estimated.

suggests that BAC does not exert an inhibitory effect to activated sludge at and below 10 mg BAC/g VSS at which almost all of the BAC is adsorbed on the biomass. The inhibition increases as BAC:VSS increases, and KI approaches to a constant value which is defined as “minimum inhibitory concentration (MIC)”. The MIC was reached at the BAC:VSS around 28 mg BAC/g VSS at which the liquid phase BAC concentration (Ce) is equal to the MIC. This implies that BAC is inhibitory only if it is in the liquid phase. The kI2 was constant around 3.08  0.83  104 mg BAC/mg VSS-h at between 7 and 28 mg BAC/g VSS. A sudden increase in the kI2 was obtained at

w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 1 2 3 8 e1 2 4 6

0.35

4

Ceq

0.0012 3

0.0006 0.0004

KI (mg BAC/L)

-1

kI2 (hr )

0.0010 0.0008

KI

0.30 0.25 0.20

2

kI2

0.15

0.05 0 5

10

15

20

25

30

35

0.00 40

BAC:VSS (mg BAC/g VSS)

Fig. 6 e Profile of liquid phase BAC concentration (Ceq: solid line e), estimated inhibition coefficient (KI: hollow circle B) and solid phase BAC utilization rate constant (kI2: upward triangle 6) at different BAC-to-biomass ratios (BAC:VSS).

32 mg BAC/g VSS at which the available biomass surface reached saturation by BAC according to the adsorption isotherm obtained in this study. The kI2 approached the kI1 value at that particular BAC:VSS at which the rate limiting step, i.e. membrane migration, diminished and biodegradation dominated. The kI2 stayed constant and equal to kI1 above 32 mg BAC/g VSS (Fig. 6).

4.

in higher than typical BAC removal efficiencies. The model, which integrates inhibition, adsorption and biodegradation processes to simulate the dynamics of BAC in the activated sludge, developed in this study may effectively be used to simulate the dynamics of other compounds (e.g., triclosan, triclocarban, linear alkyl and alkyl benzene sulfonates, perfluoroalkyl carboxylates and sulfonates etc) with properties and behavior similar to QACs.

0.10

1

0.0002 0.0000

Ceq (mg/L)

0.0014

1245

Conclusions

In this study, the dynamics of BAC in an activated sludge system were investigated. Respiratory inhibition, adsorption and biodegradation were identified as the three major processes which affect the fate of BAC in activated sludge. A comprehensive model was developed by integrating these processes into the Monod equation. The model agreed well with the data obtained from a series of batch assays performed. In conclusion, BAC inhibits oxygen uptake and use, thereby causing prolonged COD substrate utilization. BAC degradation initiates after the major portion of the readily degradable COD is utilized. Therefore, a delay in the readily degradable COD utilization causes retardation in the BAC degradation, as well. A major fraction of BAC instantly adsorbs on the biomass. Biodegradation of BAC proceeds both in the liquid and solid phases, however the solid phase BAC degradation is about twenty times slower than in the liquid phase. Given the low BAC concentrations found in municipal wastewaters, BAC is unlikely to be toxic in wastewater treatment. However, since the biodegradation rate of BAC is very slow, a major fraction of BAC is likely to be transferred and accumulated in the environment, especially in anaerobic compartments. In order to mitigate this problem, we suggest the implementation of activated sludge systems with long solid retention times such as extended aeration or employment of attached growth systems. These systems would favor the prolonged retardation of BAC by facilitating the adsorption on the biomass and increase in the degradation rate, resulting

Acknowledgements This work was financially supported by the National Water Pollution Control and Treatment Technological project (No. 2008ZX07314-002) and the National Natural Science Foundation of China (No. 50908117).

Appendix. Supplementary data Supplementary data associated with this article can be found in the on-line version, at doi:10.1016/j.watres.2010.09.037.

references

Boethling, R.S., 1984. Environmental fate and toxicity in wastewater treatment of quaternary ammonium surfactants. Water Res. 18 (3), 1061e1076. Clara, M., Scharf, S., Scheffknecht, C., Gans, O., 2007. Occurrence of selected surfactants in untreated and treated sewage. Water Res. 41 (19), 4339e4348. Dean-Raymond, D., Alexander, M., 1977. Bacterial metabolism of quaternary ammonium compounds. Appl. Environ. Microbiol. 33, 1037e1041. Dias, F.F., Bhat, J.V., 1964. Microbial ecology of activated sludge: I. Dominant bacteria. Appl. Microbiol. 12 (5), 412e417. Garcia, M.T., Campos, E., Sanchez-Leal, J., Comelles, F., 2006. Sorption of alkyl benzyl dimethyl ammonium compounds by activated sludge. J. Dispers. Sci. Technol. 27 (5), 739e744. Gaze, W.H., Abdouslam, N., Hawkey, P.M., Wellington, E.M.H., 2005. Incidence of class 1 integrons in a quaternary ammonium compound polluted environment. Antimicrob. Agents Chemother. 49 (5), 1802e1807. Geftic, S.G., Heymann, H., Adair, F.W., 1979. 14 year survival of Pseudomonas cepacia in a salts solution preserved with benzalkonium chloride. Appl. Environ. Microbiol. 37, 505e510. Gujer, W., 2008. Systems Analysis for Water Technology. Springer, Berlin, Germany. Ismail, Z.Z., Tezel, U., Pavlostathis, S.G., 2010. Sorption of quaternary ammonium compounds to municipal sludge. Water Res. 44, 2303e2313. Kaech, A., Egli, T., 2001. Isolation and characterization of a Pseudomonas putida strain able to grow with trimethyl-1,2dihydroxy-propyl-ammonium as sole source of carbon, energy and nitrogen. Syst. Appl. Microbiol. 24, 252e261. Liffourrena, A.S., Lopez, F.G., Salvano, M.A., Domenech, C.E., Lucchesi, G.I., 2008. Degradation of tetradecyl trimethyl ammonium by Pseudomonas putida A ATCC 12633 restricted by accumulation of trimethylamine is alleviated by addition of Al3þ ions. J. Appl. Microbiol. 104, 396e402. Martinez-Carballo, E., Sitka, A., Gonzalez-Barreiro, C., Kreuzinger, N., Furhacker, M., Scharf, S., Gans, O., 2007.

1246

w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 1 2 3 8 e1 2 4 6

Determination of selected quaternary ammonium compounds by liquid chromatography with mass spectrometry. Part I. Application to surface, waste and indirect discharge water samples in Austria. Environ. Pollut. 145 (2), 489e496. Nishihara, T., Okamoto, T., Nishiyama, N., 2000. Biodegradation of didecyl dimethyl ammonium chloride by Pseudomonas fluorescens TN4 isolated from activated sludge. J. Appl. Microbiol. 88, 641e647. Nishiyama, N., Nishihara, T., 2002. Biodegradation of dodecyl trimethyl ammonium bromide by Pseudomonas fluorescens F7 and F2 isolated from activated sludge. Microbes Environ. 17, 164e169. Patrauchan, M.A., Oriel, P.J., 2003. Degradation of benzyl dimethyl alkyl ammonium chloride by Aeromonas hydrophila sp K. J. Appl. Microbiol. 94, 266e272. Qin, Y., Zhang, G.Y., Kang, B.A., Zhao, Y.M., 2005. Primary aerobic biodegradation of cationic and amphoteric surfactants. J. Surfactants Detergents 8, 55e58. Reynolds, L., Blok, J., Demorsier, A., Gerike, P., Wellens, H., Bontinck, W.J., 1987. Evaluation of the toxicity of substances to be assessed for biodegradability. Chemosphere 16, 2259e2277. Rittmann, B.E., McCarty, P.L., 2001. Environmental Biotechnology: Principles and Applications. McGraw-Hill, Boston. Shimp, R.J., Young, R.L., 1988. Availability of organic-chemicals for biodegradation in settled bottom sediments. Ecotoxicol. Environ. Saf. 15 (1), 31e45. Sutterlin, H., Alexy, R., Kummerer, K., 2008. The toxicity of the quaternary ammonium compound benzalkonium chloride alone and in mixtures with other anionic compounds to bacteria in test systems with Vibrio fischeri

and Pseudomonas putida. Ecotoxicol. Environ. Saf. 71 (2), 498e505. Takenaka, S., Tonoki, T., Taira, K., Murakami, S., Aoki, K., 2007. Adaptation of Pseudomonas sp strain 7-6 to quaternary ammonium compounds and their degradation via dual pathways. Appl. Environ. Microbiol. 73, 1797e1802. Tchobanoglous, G., Burton, F.L., Stensel, H.D., 2003. Wastewater Engineering: Treatment, Disposal and Reuse, Fourth ed. McGraw-Hill Professional, New York, NY. Tezel, U., 2009. Fate and effect of quaternary ammonium compounds in biological systems. Ph. D. Thesis, Georgia Institute of Technology, Atlanta, GA. Tezel, U., Pavlostathis, S.G., 2009. Transformation of benzalkonium chloride under nitrate reducing conditions. Environ. Sci. Technol. 43 (5), 1342e1348. van Ginkel, C.G., Kolvenbach, M., 1991. Relations between the structure of quaternary alkyl ammonium salts and their biodegradability. Chemosphere 23 (3), 281e289. van Ginkel, C.G., Vandijk, J.B., Kroon, A.G.M., 1992. Metabolism of hexadecyl trimethyl ammonium chloride in Pseudomonas strain-B1. Appl. Environ. Microbiol. 58, 3083e3087. van Ginkel, C.G., 2004. Biodegradation of cationic surfactants. In: Zoller, U. (Ed.), Handbook of Detergents Part B: Environmental Impact, vol. 121. Marcel Dekker, Inc., New York, USA. Yang, J., 2007. Fate and effect of alkyl benzyl dimethyl ammonium chloride in mixed aerobic and nitrifying cultures. MS Thesis. Georgia Institute of Technology, Atlanta, GA. Ying, G.G., 2006. Fate, behavior and effects of surfactants and their degradation products in the environment. Environ. Int. 32 (3), 417e431.