Article pubs.acs.org/est
Influence of Bioselector Processes on 17α-Ethinylestradiol Biodegradation in Activated Sludge Wastewater Treatment Systems Ryan M. Ziels, Mariko J. Lust, Heidi L. Gough, Stuart E. Strand, and H. David Stensel* Department of Civil and Environmental Engineering, University of Washington, Seattle, Washington 98195-2700, United States S Supporting Information *
ABSTRACT: The removal of the potent endocrine-disrupting estrogen hormone, 17α-ethinylestradiol (EE2), in municipal wastewater treatment plant (WWTP) activated sludge (AS) processes can occur through biodegradation by heterotrophic bacteria growing on other organic wastewater substrates. Different kinetic and metabolic substrate utilization conditions created with AS bioselector processes can affect the heterotrophic population composition in AS. The primary goal of this research was to determine if these changes also affect specific EE2 biodegradation kinetics. A series of experiments were conducted with parallel bench-scale AS reactors treating municipal wastewater with estrogens at 100−300 ng/L concentrations to evaluate the effect of bioselector designs on pseudo first-order EE2 biodegradation kinetics normalized to mixed liquor volatile suspended solids (VSS). Kinetic rate coefficient (kb) values for EE2 biodegradation ranged from 5.0 to 18.9 L/g VSS/d at temperatures of 18 °C to 24 °C. EE2 kb values for aerobic biomass growth at low initial food to mass ratio feeding conditions (F/Mf) were 1.4 to 2.2 times greater than that from growth at high initial F/ Mf. Anoxic/aerobic and anaerobic/aerobic metabolic bioselector reactors achieving biological nutrient removal had similar EE2 kb values, which were lower than that in aerobic AS reactors with biomass growth at low initial F/Mf. These results provide evidence that population selection with growth at low organic substrate concentrations can lead to improved EE2 biodegradation kinetics in AS treatment.
■
a broader physiological capability at slower growth rates.15 Improved micropollutant removal by biodegradation would be expected for longer SRTs due to increased contact time and biomass for treatment.16 However, an evaluation of municipal AS treatment EE2 performance data found no clear improvement in EE2 removal with increased SRT.17 EE2 removal efficiencies varied from 60% to 80% within an SRT range of 5 to 12 days, and from 25% to 70% within an SRT range of 20 to 42 days.17 Since many factors could have contributed to this variation in EE2 removal efficiency, it is important to consider how other operational parameters besides SRT could also affect the AS microbial population composition and potentially its associated EE2 biodegradation activity. The heterotrophic microbial population in AS treatment can be affected by the type of growth substrate, operating conditions (such as SRT and dissolved oxygen concentration), and reactor configuration.18 A common AS process design approach is to use bioselectors to manipulate heterotrophic populations to obtain targeted WWTP performance goals. Such bioselector applications typically utilize reactor configurations
INTRODUCTION Natural and synthetic estrogens of human origin are endocrinedisrupting compounds (EDCs) capable of harming the reproductive processes of aquatic wildlife and fish.1−3 Biodegradation is the major removal mechanism for estrogen hormones in activated sludge (AS) treatment,4 with considerably lower biodegradation rates observed for the synthetic estrogen 17α-ethinylestradiol (EE2) than for the natural estrogens estrone (E1) and 17β-estradiol (E2).5 EE2 also exerts an estrogenic potency that is more than 1.25 times that of E2 and over 15 times that of E1.6,7 Other influent wastewater substrates are needed to support the growth of EE2-degrading bacteria because the low (ng/L) influent EE2 concentrations cannot support biomass growth.8 Both heterotrophic bacteria and ammonia-oxidizing bacteria (AOB) have been associated with EE2 removal, and more recent findings suggest that heterotrophic bacteria are of major importance for EE2 removal in municipal AS processes.9−11 A number of heterotrophic bacteria have also been shown to degrade EE2 in pure-culture.12−14 In spite of this understanding, there is no clear trend established between EE2 removal efficiency and AS process parameters. Increased micropollutant removal efficiency with longer AS solids retention times (SRT) has been proposed, and was attributed to the ability to enrich for more diverse bacteria with © 2014 American Chemical Society
Received: Revised: Accepted: Published: 6160
November 30, 2013 March 19, 2014 May 8, 2014 May 8, 2014 dx.doi.org/10.1021/es405351b | Environ. Sci. Technol. 2014, 48, 6160−6167
Environmental Science & Technology
Article
Table 1. Operating Conditions for the SBRs in the Three Experimental Phasesa,b,c parameter
phase I
SRT (d) temp. (°C) metabolic type kinetic type F/Mf (g COD/g VSS/d) fill time (min)
6.0 21.3 (1.7) aer. low 0.7 300
aer. high 15.3 15
phase II
aer.d highd 15.0 15
aer. high 10.2 15
phase III
8.7 17.5 (1.7) anox./aer. high 11.4 15
anaer./aer. high 9.6 15
aer. high 84.5 5
4.0 20.0 aer. low 1.5 300
anox./aer. low 1.2 300
a Average values shown with standard deviation in parentheses. bThe effect of kinetic bioselection was evaluated in Phase I, metabolic bioselection in Phase II, and combined effects in Phase III. cKey: SBR = sequencing batch reactor; F/Mf = feeding food to mass ratio; MLVSS = mixed liquor volatile suspended solids; SRT = solids retention time; Aer. = aerobic; Anox. = anoxic; Anaer. = anaerobic; High = high-F/Mf feeding conditions; and Low = low-F/Mf feeding conditions. dReceived a lower fraction of influent readily biodegradable COD.
coefficients normalized to mixed liquor volatile suspended solids (MLVSS) were used to characterize specific EE2 biodegradation kinetics for the different bioselector operating conditions.
and/or redox conditions to provide kinetic or metabolic selective growth pressures that enrich for bacterial communities with improved sludge settling characteristics and/or accomplish biological nutrient removal (BNR).18,19 Kinetic bioselectors promote substrate uptake at high initial food to mass ratio (F/Mf) feeding conditions and elevated substrate concentrations, and have been used to enrich for floc-forming bacteria and prevent filamentous sludge bulking.19 Reactor configurations with initial anoxic (no oxygen) or anaerobic (no oxygen or nitrate) contact zones result in metabolic bioselection by favoring substrate uptake and growth of bacteria capable of nitrate/nitrite reduction or high intracellular polyphosphate storage, respectively.19 Common metabolic bioselector designs for BNR include sequential anoxic and aerobic reactors for nitrogen removal, and sequential anaerobic and aerobic reactors for enhanced biological phosphorus removal (EBPR). The effect of the aforementioned AS population selection mechanisms on the activity of EE2-degrading microorganisms is poorly understood. A comparison of estrogen removal at two full-scale WWTPs with and without an anaerobic bioselector revealed differences in EE2 biodegradation activities,20 and emphasized the need to further evaluate the influence of AS metabolic bioselector designs on EE2 biodegradation kinetics. To the best of our knowledge, no study has compared the types of metabolic bioselectors commonly used in WWTPs using the same influent municipal wastewater source and SRT to determine if distinct heterotrophic BNR populations express different EE2 biodegradation kinetics. Nor has a study compared the EE2 biodegradation kinetics of AS systems operating with identical SRTs, but with substrate uptake occurring under high or low concentrations. In this study, we hypothesized that different AS reactor configurations that use kinetic and metabolic growth pressures to select for distinct heterotrophic populations would have different specific EE2 biodegradation kinetics. The objective of this study was to investigate the effect of metabolic and kinetic biological selection processes with similar SRTs on EE2 biodegradation kinetics during AS treatment of municipal primary effluent wastewater with EE2 at concentrations representative of that in municipal WWTP influents. Metabolic bioselection was examined by operating sequencing batch reactors (SBRs) with aerobic, anoxic, or anaerobic redox conditions during the fill and react period. Kinetic bioselection conditions were produced by feeding SBRs over a short duration with a high initial contact F/Mf for substrate uptake at a high concentration, or conversely feeding over the entire SBR react period with a low F/Mf for substrate uptake at a low concentration. Pseudo first-order EE2 biodegradation rate
■
MATERIALS AND METHODS Activated Sludge Reactor Configuration and Operation. Three parallel AS SBRs were operated within three experimental phases to examine the effects of metabolic and kinetic population selection mechanisms on EE2 biodegradation kinetics. Each SBR consisted of a 4-L Pyrex glass Erlenmeyer flask. The SBRs in phases I and II were operated at the West Point WWTP in Seattle, WA, and were fed from a primary effluent feed container that was replenished daily at the WWTP. The phase III SBRs were operated in a 20 °C walk-in environmental chamber at the University of Washington, and were fed King County South WWTP primary effluent wastewater that was replenished daily and was stored at 4 °C after its delivery every 7 days. Aeration was provided with an aquarium pump and a ceramic stone sparger, and mixed liquor nitrogen sparging was provided from a gas cylinder at 300 mL/ min during the anoxic and anaerobic metabolic bioselector periods to eliminate oxygen surface transfer. Magnetic stir bars mixed the AS SBRs throughout the fill and react periods. A summary of the operating conditions for the three SBR phases is given in Table 1, and further details on the reactor operation are given in the Supporting Information (SI). Phase I examined the impact of high and low F/Mf feeding conditions under aerobic conditions. Phase II compared aerobic, anoxic, and anaerobic metabolic bioselection under high F/Mf feeding. Phase III compared the effect of high and low F/Mf feeding conditions under aerobic conditions, and also compared the effect of aerobic and anoxic metabolic conditions under low F/Mf feeding conditions. The SBRs were seeded with activated sludge from local WWTPs at the start of each phase: City of Port Orchard, WA AS (aerobic process) for phase I, City of Snoqualmie, WA oxidation ditch (anoxic and EBPR processes) for phase II, and King County South WWTP (Renton, WA) AS (aerobic process) for phase III. The parallel reactors in all operating phases were fed equal volumes of wastewater daily, while the feed duration was selected to control the initial F/Mf loading (g COD/g VSS/d), as calculated according to Tchobanoglous et al.:16 F/M f =
(Q feed)(TCODin ) (MLVSS)(V )
(1)
where Qfeed is the flow rate during the feed period (L/d), TCODin is the average influent total chemical oxygen demand 6161
dx.doi.org/10.1021/es405351b | Environ. Sci. Technol. 2014, 48, 6160−6167
Environmental Science & Technology
Article
of the SBR cycle. The in situ biodegradation tests were not spiked, and thus the removal of EE2 during normal feeding conditions was followed in phase II. The average VSS concentration of the AS in each biodegradation test was determined from measurements taken at the beginning and end of each test. A total of three batch biodegradation tests were conducted with each AS reactor. Example plots of total EE2 concentration versus time obtained during batch biodegradation tests are given in the SI (Figure S1). Estrogen concentrations were not extensively monitored throughout the metabolic bioselector periods since negligible EE2 biodegradation has been shown to occur under anaerobic and anoxic conditions,21−25 which was corroborated by our observations (data not shown). Modeling EE2 Biodegradation and Solids Partitioning. EE2 biodegradation is described by a pseudo-first order kinetic model as a function of soluble estrogen and VSS concentrations.21,26
concentration (COD) (g/L), MLVSS is the average reactor mixed liquor volatile suspended solids concentration (g/L), and V is the reactor volume (L). A high F/Mf feed condition was implemented using short feeding durations that promoted substrate uptake at high initial concentrations (Table 1). Conversely, an operating condition with substrate uptake at a low concentration was conducted by adding feed every 3 to 5 min throughout the 5 h react/fill period to simulate a low-loaded complete-mix AS (CMAS) reactor (Table 1). These two feeding methods will be hereby referred to as “high-F/M f” and “low-F/M f” operating conditions, respectively. The SBR primary effluent wastewater feed was augmented with acetate and the estrogens E1, E2, and EE2. Equal quantities of acetate were added to the parallel SBRs for all operating phases at approximately 120 mg/L for phases I and II and 150 mg/L for phase III. The exception was one high-F/Mf reactor in phase I, which did not receive influent acetate in order to examine its performance with a lower fraction of influent readily biodegradable COD (rbCOD) and high-F/Mf feeding (SI). Sodium nitrate (10 to 25 mg N/L) was added to the feed of the anoxic/aerobic SBRs only (Phases II and III) to allow an increase in the amount of organic substrate removal during the anoxic period, and thus promote a higher fraction of denitrifiers in the heterotrophic biomass of the anoxic/aerobic bioselectors (SI). Phosphate (4 mg P/L) as mono and dibasic potassium phosphate was added to the feed of all SBRs in phase II to ensure growth of polyphosphate accumulating organisms during the anaerobic period of the anaerobic/aerobic bioselector reactor (SI). All SBRs had a total cycle time of 6 h, and the feed volume for each cycle was 1/4 of the total liquid volume for phases I and II and 1/5 for phase III. A quiescent settling time of 1-h occurred before supernatant was decanted throughout the final 8 min of the cycle with peristaltic pumps. Details of the AS SBR average influent and effluent characteristics are provided in the SI (Table S1). Influent concentrations of E1 and EE2 to the phase I SBRs averaged 192 ± 35 ng/L and 118 ± 25 ng/L (n = 12), respectively, while E2 was somewhat reduced in the phase I feed at 70 ± 54 ng/L (n = 12) possibly due to transformation during storage. Influent concentrations of E1, E2, and EE2 to the phase II SBRs averaged 213 ± 41 ng/L, 176 ± 28 ng/L, and 164 ± 25 ng/L, respectively (n = 21). Influent concentrations of E1, E2, and EE2 in phase III averaged 448 ± 100 ng/L, 348 ± 75 ng/L, and 301 ± 17 ng/L, respectively (n = 23). EE2 Batch Biodegradation Tests. All reactors were operated for a minimum time of three SRTs before conducting EE2 batch biodegradation tests, in which the total estrogen (soluble plus sorbed to solids) concentration was followed with time under aerobic conditions. Because of the semicontinuous feeding strategy for the low-F/Mf operation employed in phases I and III, EE2 biodegradation tests for those reactors were conducted as ex-situ batch tests. The ex-situ tests were conducted by transferring 500 mL of mixed liquor at the end of the aerobic SBR period to aerated and mixed (magnetic stirrer) 1-L flasks that were then spiked with 300 to 400 ng/L E1, E2, and EE2 (from stock solutions in Milli-Q water). Total estrogen concentrations were monitored for 5 h during the phase I batch tests and for 12 h during the phase III batch tests. EE2 biodegradation tests for the phase II reactors were performed as in situ batch tests, in which total estrogen concentrations were monitored throughout the aeration period
dEtot = −k bXVSSEsol dt
(2)
where Etot is the total EE2 concentration (ng/L), t is the time (days), kb is the pseudo first-order EE2 biodegradation kinetic coefficient (L/g VSS/d), XVSS is the MLVSS concentration (g VSS/L), and Esol is the soluble EE2 concentration in the bulk aqueous phase (ng/L). To determine values of kb from total EE2 concentration over time in batch biodegradation tests, the above equation was modified assuming rapid solid−liquid equilibrium of estrogens27 and linearized as follows:26 ⎛ E ⎞ ⎛ ⎞ XVSS tot ⎟⎟ = −k b⎜ ln⎜⎜ ⎟t ⎝ 1 + KFXVSS ⎠ ⎝ Etot,initial ⎠
(3)
where Kp is the EE2 solid−liquid partitioning coefficient (L/g VSS). Values of kb were calculated by multiplying the slope of the linear trend-line of ln(Etot/Etot,initial) versus time by (1+KpXVSS)/XVSS (see Figure S2 in the SI). The apparent EE2 Kp values for the SBR AS were determined by simultaneously taking soluble and total EE2 measurements near the end of each batch biodegradation test. The specific sorbed concentration (Esorb, ng/g) is the solid phase EE2 concentration divided by the MLVSS concentration:
Esorb =
Etot − Esol XVSS
(4)
and the apparent EE2 solid−liquid partitioning coefficient for AS solids was then calculated as follows: Kp =
Esorb(time = t ) Esol(time = t )
(5)
One-way Analysis of Variance (ANOVA) and Tukey’s test were used to compare average EE2 kb and Kp values between reactors using the StatPlus add-on to Microsoft Excel (2011). Estrogen Measurements. Estrogen samples were prepared in accordance with the method described by Gaulke et al.,9 except that 50 pg of each internal standard (d4E1, d4E2, and d4EE2) was added to the 0.5 mL samples before liquid extraction. Samples were taken in duplicate for soluble and total estrogen measurements. Samples for total estrogen measurements consisted of 0.5 mL mixed liquor. Mixed liquor samples for soluble estrogen measurements (6 mL) were centrifuged for 10 min at 3200 rpm and 4 °C, after which 0.5 mL of the 6162
dx.doi.org/10.1021/es405351b | Environ. Sci. Technol. 2014, 48, 6160−6167
Environmental Science & Technology
Article
supernatant was decanted for analysis. Estrogens were extracted from the 0.5 mL sample aliquots by adding 3 mL of ethyl acetate and shaking vigorously for 10 min. The resulting organic fraction was transferred to a clean sample tube and evaporated to dryness at 40 °C under nitrogen gas. Evaporated samples were then reconstituted with 100 μL of NaHCO3 buffer (pH 10.5) and 100 μL of 1 mg/mL dansyl-chloride in ACN, vortexed for 1 min, and held for 30 min at 60 °C prior to immediate analysis. A description of all reagents used in estrogen analysis is provided in the SI. Estrogen quantification was performed using a Shimadzu LC20AD high performance liquid chromatograph (LC) (Kyoto, Japan) combined with an Applied Biosystems 4000 Q-Trap tandem mass spectrometer (MS/MS) (Foster City, CA). The protocol for estrogen separation and quantification with LCMS/MS is described in the SI. The method limit of detection (LOD) for E1, E2, and EE2 was 1 ng/L, and the method limit of quantification (LOQ) was 5 ng/L. The natural estrogens, E1 and E2, were included in the feed to simulate a municipal wastewater estrogen matrix, yet were present below the LOQ in all reactor effluents and were degraded too rapidly for accurate determination of biodegradation kinetics during batch tests (data not shown).
Figure 1. Average pseudo first-order EE2 biodegradation kinetic coefficients (kb) for batch biodegradation tests with activated sludge from phase I aerobic-only kinetic bioselector reactors comparing the effect of growth at high and low initial substrate concentrations. The average temperature was 24.0 °C and SRT was 6 days. Greek symbols (α and β) indicate significantly different kb values (p < 0.05 with Tukey’s test). n = 3 biodegradation tests for each SBR. Average EE2 kb values shown in center of bars. Error bars indicate one standard deviation. F/Mf = initial food to mass ratio.
outcompeted for rbCOD at the high concentrations applied in high-F/Mf operating conditions. These observations do not eliminate the possibility that EE2-degrading bacteria may have also grown on endogenous decay products, yet the relative fraction of substrate produced by endogenous decay would have been similar in the phase I reactors due to their identical HRT and SRT. The highest EE2 kb of the phase I SBRs was associated with the lowest initial growth substrate concentration in the lowF/Mf reactor (Figure 1, Table 1). It is important to note that the SBR operation allowed the F/Mf to be reduced in the lowF/Mf reactor while maintaining a relatively low SRT of 6-days (Table 1). These results suggest that kinetic population selection is an important operational factor governing the growth and activity of EE2-degrading microorganisms in AS systems. A recent study on E1 biodegradation by Tan et al.32 also found that kinetic selection affected the estrogen biodegradation rate. They observed reduced E1 biodegradation at high organic substrate concentrations, and postulated that such conditions favored growth of r-strategist populations incapable of E1 biodegradation. Koh et al.20 hypothesized that improved EE2 biodegradation at low substrate growth conditions could have been attributed to the growth of Kstrategist heterotrophic populations. However, the range of F/M conditions investigated in that study was too limited (difference of only 0.05 g BOD/g MLVSS/d)20 to elucidate the effects of kinetic population selection on EE2 biodegradation activity. Our study investigated a wide range of initial growth substrate concentrations (difference in F/Mf of 14.3 g COD/g MLVSS/d in phase I SBRs, see Table 1), which resulted in a higher EE2 kb in the low-F/Mf reactor by a factor of 1.4 to 1.7 (Figure 1). These results provide the first clear evidence that population selection with growth at low organic substrate concentration can increase EE2 biodegradation kinetics independent of SRT. Increased estrogen biodegradation observed in AS systems has sometimes been attributed to cometabolic transformation by ammonia-oxidizing bacteria (AOB).33−38 AOB biomass fractions in the MLVSS in the SBRs were estimated based on the amount of ammonia oxidized (see SI). In phase I, all of the reactors were operated at the same SRT and the highest fraction of nitrifying biomass was estimated at 3.9% for the
■
RESULTS AND DISCUSSION Effect of Kinetic Bioselection on Specific EE2 Biodegradation Kinetic Coefficients. In order to test the effects of kinetic population selection on EE2 biodegradation activity in aerobic AS systems in phase I, the feeding duration was adjusted in the SBRs to alter their feeding F/Mf conditions (Table 1) and promote substrate uptake at low concentrations in the low-F/Mf aerobic reactor and high concentrations in two other high-F/Mf aerobic reactors. One of the high-F/Mf aerobic reactors was operated with less influent rbCOD (difference of 117 mg COD/L, SI Table S1) to determine the effects of wastewater composition on the pseudo first-order EE2 biodegradation rate coefficient (kb) in the high-F/Mf kinetic bioselector reactors. The different initial contact F/M f conditions employed in the reactors altered the substrate concentration available during growth, which resulted in differing sludge morphologies in the high-F/Mf reactors and low-F/Mf reactor (SI Figure S3). The low-F/Mf reactor sludge was characterized by an abundance of filamentous organisms, while the two high-F/Mf reactors were dominated by nonfilamentous sludge flocs (SI Figure S3). This observation was consistent with the kinetic selection theory proposed by Chudoba et al.,28 in which filamentous microorganisms were enriched in CMAS reactor configurations, while nonfilamentous bacteria were dominant in batch fed kinetic bioselector SBRs.29−31 EE2 kb values within the phase I reactors ranged from 11.1 to 18.9 L/g VSS/d (Figure 1), and were significantly different between the reactor designs (p = 0.001, ANOVA). Comparisons using Tukey’s test revealed that the EE2 kb in the lowF/Mf reactor was significantly higher than in the high-F/Mf reactor treating the same influent wastewater (p = 0.001) and the high-F/Mf reactor operated with less rbCOD (p = 0.006). However, no statistically significant difference was detected between the EE2 kb values in the two high-F/Mf aerobic reactors (p = 0.27), despite their different fractions of influent rbCOD. These results suggest that EE2-degrading populations in AS may grow better on rbCOD at the low concentrations promoted with low-F/Mf feeding conditions, but they may be 6163
dx.doi.org/10.1021/es405351b | Environ. Sci. Technol. 2014, 48, 6160−6167
Environmental Science & Technology
Article
high-F/Mf reactor with the lower influent rbCOD, while the lowest fraction of 2.2% was estimated for the low-F/Mf reactor and the high-F/Mf reactor with acetate-augmented influent wastewater (SI Table S2). Thus, the highest EE2 kb occurred in the AS configuration with the lowest nitrifying biomass fraction. Moreover, no significant correlation was observed between EE2 kb values and the fraction of nitrifying biomass in the total MLVSS for all experimental reactors operated in this study (R2 = 0.21) (SI Figure S4). At similar ng/L EE2 concentrations, Bagnall et al.11 showed that EE2 biodegradation did not change during AS treatment when the nitrogen source for a synthetic wastewater feed was changed from ammonia to nitrate, resulting in the subsequent loss of nitrification and a 99% decrease in AOB concentration. In addition, Gaulke et al.9 observed no EE2 removal by pure cultures of Nitrosomonas europaea and Nitrosospira multiformis at an EE2 concentration of 500 ng/L and 10 mg/L NH3N. However, at about 1000fold higher EE2 concentrations, Khunjar et al.38 calculated an EE2 biotransformation rate by N. europaea of 13.6 L/g CODAOB/d (19.3 L/g VSSAOB/d) at 200 μg/L EE2 and 15 mg/L NH3N. Assuming that the above biotransformation rate coefficient observed at the high EE2 concentration applied to the AOB biomass fractions in our reactors, the amount of EE2 transformation by AOB would only account for 2% to 7% of the mixed liquor EE2 kb values ranging from 5.0 to 18.9 L/g VSS/d. Therefore, it is likely that heterotrophic bacteria were primarily responsible for the EE2 biodegradation observed in this study. Effect of High-F/Mf Metabolic Bioselection on Specific EE2 Biodegradation Kinetic Coefficients. The high-F/Mf metabolic bioselector reactors operated in phase II removed a substantial portion of influent soluble COD (sCOD) during the metabolic bioselector fill/react time, and also achieved EBPR and nitrogen removal typical of their respective BNR bioselector types (SI Table S1). The average removal of influent sCOD was 93% and 95% (n = 17) in the anoxic/ aerobic and anaerobic/aerobic reactors, respectively, during their 1.5-h metabolic bioselection periods. The anoxic/aerobic reactor removed an average of 11 mg N/L more than the aerobic reactor (n = 18) (SI Table S1). The anaerobic/aerobic reactor removed an average of 260% more influent phosphorus (3.2 mg P/L removed) than the aerobic reactor (n = 23) (SI Table S1). No statistically significant difference was detected among the EE2 kb values of the phase II high-F/Mf metabolic bioselector reactors (Figure 2) (p = 0.10, ANOVA), despite that most of the influent biodegradable substrate removal occurred in the aerobic and metabolic bioselector processes. On the basis of the observed substrate removal mechanisms in these three parallel reactors, different heterotrophic population compositions were expected. A possible explanation for the similar kb values among the BNR systems is that the growth of EE2-degrading bacteria may have occurred on residual slowly biodegradable substrate and endogenous decay products in the aerobic period after the bioselector step. Combined Effects of Kinetic and Metabolic Bioselection Processes on Specific EE2 Biodegradation Kinetic Coefficients. In phase III, aerobic-only and anoxic/aerobic reactors were operated under a low-F/Mf kinetic bioselection condition, and an aerobic-only reactor was operated with a high-F/Mf kinetic bioselection condition. This experimental design allowed the effects of metabolic bioselection on EE2 biodegradation kinetics to be elucidated under low-F/Mf
Figure 2. Average pseudo first-order EE2 biodegradation kinetic coefficients (kb) for in situ batch biodegradation tests with the phase II AS reactors comparing the effect of metabolic bioselection with growth at high initial food to mass ratio (F/Mf) conditions. The average temperature was 18.3 °C and SRT was 8.7 days. Average EE2 kb of the phase II SBRs were not statistically different at the p = 0.05 level (with Tukey’s test). n = 3 biodegradation tests for each SBR. Average EE2 kb values shown in center of bars. Error bars indicate one standard deviation.
kinetic population selection conditions (in contrast to phase II), and validated the effects of kinetic population selection found in phase I using a different AS inoculum and primary effluent source. As shown in Figure 3, the EE2 kb value in the aerobic lowF/Mf reactor was more than double that of the high-F/Mf
Figure 3. Average pseudo first-order EE2 biodegradation kinetic coefficients (kb) for batch biodegradation tests with activated sludge from the phase III reactors comparing the effect of high and low initial food to mass ratio (F/Mf) aerobic feeding and the effect of aerobic versus anoxic metabolic bioselection conditions under low F/Mf feeding. The average temperature was 20.0 °C, and the SRT was 4.0 days. Greek symbols (α and β) indicate significantly different kb values (p < 0.05 with Tukey’s test). n = 3 biodegradation tests for each SBR. Average EE2 kb values shown in center of bars. Error bars indicate one standard deviation.
aerobic reactor (p = 0.0002). This result validated the observation of higher EE2 kb values observed in the aerobic low-F/Mf reactor in phase I (Figure 1), indicating that the effects of low substrate process selection on EE2 biodegradation kinetics were repeated with a different wastewater source. However, the EE2 kb of the low-F/Mf anoxic/aerobic reactor was not significantly different from the parallel high-F/Mf aerobic reactor (p = 0.06), suggesting that the growth of EE2 degraders under low substrate feeding conditions was hindered by anoxic versus aerobic redox conditions. As such, the EE2 kb value in the aerobic low-F/Mf reactor was 1.6 times that of the aerobic EE2 kb value in the anoxic/aerobic reactor (p = 0.0006), which was operated with the same low F/Mf (Figure 3). 6164
dx.doi.org/10.1021/es405351b | Environ. Sci. Technol. 2014, 48, 6160−6167
Environmental Science & Technology
Article
Table 2. Solid-Liquid EE2 Partitioning Coefficients (Kp) for Activated Sludge from Each SBR Experimental Operating Phasea,b parameter metabolic type kinetic type EE2 (Kp) (L/kg-VSS)
phase I SBRs (n = 5) aer. high 405 (163)
aer. low 682 (203)
phase II SBRs (n = 4)
aer.c highc 440 (154)
aer. high 554 (112)
anox./aer. high 684 (62)
phase III SBRs (n = 6)
anaer./aer. high 408 (57)
aer. high 303 (138)
aer. low 867 (76)
anox./aer. low 856 (142)
a Average values shown with standard deviation in parentheses. bKey: SBR = sequencing batch reactor; Aer. = aerobic; Anox. = anoxic; Anaer. = anaerobic; High = high-F/Mf feeding conditions; and Low = low-F/Mf feeding conditions cReceived a lower fraction of influent readily biodegradable COD
Table 3. Average Total EE2 Concentrations and EE2 Biodegradation Removal Efficiencies Based on Samples Taken at the End of the Aerobic Period of the SBR Cycle for the Three Experimental Phasesa,b parameter metabolic type kinetic type end-of-cycle total EE2 concentration (ng/L)d EE2 biodegradation removal efficiency (%)d,e
phase I SBRs (n = 7)
phase II SBRs (n = 5) c
phase III SBRs (n = 9)
aer. high 16 (2.3)
aer. low 20 (1.1)
aer. highc 18 (2.1)
aer. high 29 (9.4)
anox./aer. high 24 (4.8)
anaer./aer. high 30 (5.2)
aer. high 153 (34)
aer. low 139 (20)
anox./aer. low 189 (28)
88 (1.4)
85 (1.0)
80 (3.0)
81 (6.5)
83 (3.2)
80 (3.3)
52 (4.2)
57 (3.8)
42 (6.8)
a Average values shown with standard deviation in parentheses. bKey: SBR = sequencing batch reactor; Aer. = aerobic; Anox. = anoxic; Anaer. = anaerobic; High = high-F/Mf feeding conditions; and Low = low-F/Mf feeding conditions. cReceived a lower fraction of influent readily biodegradable COD. dSamples averaged over last 1.5 SRTs of operation. eBiodegradation efficiency calculated as (1-EE2end‑of‑cycle/EE2influent).
The calculation of EE2 kb values using Eq. 3 relies on the EE2 Kp value. A sensitivity analysis of the first-order EE2 biodegradation model revealed that a 50% decrease in the EE2 Kp of the phase III aerobic low-F/Mf sludge would yield an 8% decrease in its average EE2 kb. Even at this lower EE2 kb, the observed difference between the aerobic low-F/Mf reactor and high-F/Mf reactor would still remain significant (p = 0.0002). Therefore, the higher EE2 biodegradation rate coefficients observed in the aerobic low-F/Mf AS systems were not an artifact of their elevated EE2 Kp values. Effect of AS Reactor Configuration on EE2 Biodegradation Removal Efficiency. The average EE2 biodegradation removal efficiencies for the SBRs in the three experimental phases ranged from 42% to 88% (Table 3). While operation under aerobic low substrate growth conditions resulted in higher EE2 biodegradation kinetics (Figures 1 and 3), the EE2 removal efficiencies were not increased (Table 3). Despite the 1.7-fold increase in the first-order EE2 kb of the phase I low-F/Mf reactor (Figure 1), its semicontinuous feeding resulted in an EE2 biodegradation removal efficiency that was similar to the batch fed high-F/Mf aerobic reactors operated in parallel (p > 0.10) (Table 3). Similarly, the 2.2-fold increase in EE2 kb in the phase III aerobic low-F/Mf reactor relative to the high-F/Mf reactor (Figure 3) did not result in a significantly greater EE2 biodegradation removal efficiency (p > 0.10) (Table 3). The lack of correlation between EE2 biodegradation kinetics and removal efficiency is related to different reactor hydraulic characteristics simulated by the lowF/Mf and high-F/Mf reactor operations. The high-F/Mf SBR operation simulated a plug flow reactor (PFR), in which the rapid fill-time (Table 1) was followed by batch reactor biodegradation kinetics with a high initial EE2 concentration that decreased during the remaining aeration period. Greater EE2 removal rates occur at higher EE2 concentrations due to the first-order biodegradation kinetics (Eq. 2). However, the low-F/Mf SBR operation simulated a CMAS process with a relatively constant low EE2 concentration due to the semicontinuous feeding throughout the react period. Even though the specific EE2 biodegradation kinetic coefficients were greater for the low-F/Mf aerobic systems, the amount of
However, when the anoxic/aerobic and aerobic-only reactors were operated with a high F/Mf in phase II, the aerobic EE2 kb values were similar (Figure 2). Therefore, it appears that aerobic EE2 biodegradation kinetics were negatively impacted when biomass growth occurred at high initial substrate concentrations or under anoxic conditions. To the best of our knowledge, this is the first reported difference in aerobic EE2 biodegradation rate coefficients between anoxic/aerobic and aerobic AS systems operated with identical SRT, HRT, influent wastewater, and feeding regime. Effect of AS Reactor Configuration on EE2 Partitioning to Biosolids. We hypothesized that the kinetic and metabolic selective pressures applied in the AS reactor configurations in this study would enrich for microbial communities with differing physiological properties. The apparent EE2 solid−liquid partitioning coefficient values (Kp) of the low-F/Mf reactor AS in phases I and III were greater than the Kp values of the AS in the parallel high-F/Mf reactors (Table 2). This effect was highly significant in phase III (p < 0.0001, ANOVA), as the average EE2 Kp values of the AS in the low-F/Mf aerobic and anoxic/aerobic reactors were more than double the Kp values of the AS in the high-F/Mf aerobic reactor (p < 0.0001, p < 0.0001, respectively). However, the EE2 Kp values of the two low-F/Mf reactor sludges in phase III were not significantly different (p = 0.99), suggesting similar AS sorption characteristics in the two systems operating with low substrate concentrations. The observation of variation among EE2 partitioning coefficients with changes in floc or biomass growth conditions is consistent with previously reported results. For example, higher EE2 Kp values have been associated with smaller mean floc particle sizes and higher specific surface areas of AS biomass.39 EE2 Kp values were also reported to be greater for a membrane bioreactor (MBR) AS relative to SBR AS, an effect which was attributed to greater floc surface area and cell hydrophobicity in the MBR sludge.40 In the current study, we observed elevated EE2 biodegradation kinetic coefficients for AS with greater EE2 Kp values, but the relationship of enhanced EE2 biodegradation activity to sorption remains unclear. 6165
dx.doi.org/10.1021/es405351b | Environ. Sci. Technol. 2014, 48, 6160−6167
Environmental Science & Technology
■
removal was not improved due to operation only at a lower EE2 concentration. This has been illustrated in the classic case of requiring longer reaction times in a continuously stirred tank reactor (CSTR) to achieve similar effluent concentrations as a PFR operated with the same first-order reaction rate coefficient.41 An important finding of this research is that the AS reactor configuration affects the EE2 removal efficiency in two ways: (1) the microbial population selection determines the EE2 biodegradation kinetics, and (2) the reactor EE2 concentration affects its removal rate. The EE2 first-order biodegradation kinetics also point to improved EE2 removal in reactors with increased staging or plug-flow characteristics. As priority micropollutants such as EE2 may become subject to future water quality regulations (for example, see European Union suggested Environmental Quality Standards),42 it is increasingly important to understand the relationships between AS process design and EE2 biodegradation in WWTPs. This study showed for the first time that anoxic versus anaerobic metabolic bioselection did not impact EE2 biodegradation kinetics in AS, whereas kinetic population selection had a significant effect. BNR systems and high-F/Mf aerobic bioselectors achieved similar EE2 biodegradation kinetics, which were both lower than in low-F/Mf aerobic reactors. These results indicate the potential benefit of employing MBR processes (no gravitational settling) with low-F/Mf growth conditions for improved EE2 biodegradation kinetics that may also enrich for filamentous sludge with poor settling properties. Likewise, oxidation ditch processes can provide low-F/Mf substrate removal conditions that might favor higher EE2 biodegradation kinetics, while maintaining acceptable sludge settling characteristics.
■
REFERENCES
(1) Jobling, S.; Nolan, M.; Tyler, C. R.; Brighty, G.; Sumpter, J. P. Widespread sexual disruption in wild fish. Environ. Sci. Technol. 1998, 32, 2498−2506. (2) Purdom, C. E.; Hardiman, P. A.; Bye, V. V. J.; Eno, N. C.; Tyler, C. R.; Sumpter, J. P. Estrogenic effects of effluents from sewage treatment works. Chem. Ecol. 1994, 8, 275−285. (3) Routledge, E. J.; Sheahan, D.; Desbrow, C.; Brighty, G. C.; Waldock, M.; Sumpter, J. P. Identification of estrogenic chemicals in STW effluent. 2. In vivo responses in trout and roach. Environ. Sci. Technol. 1998, 32, 1559−1565. (4) Andersen, H.; Siegrist, H.; Halling-Sørensen, B.; Ternes, T. A. Fate of estrogens in a municipal sewage treatment plant. Environ. Sci. Technol. 2003, 37, 4021−4026. (5) Ternes, T. A.; Kreckel, P.; Mueller, J. Behaviour and occurrence of estrogens in municipal sewage treatment plantsII. Aerobic batch experiments with activated sludge. Sci. Total Environ. 1999, 225, 91− 99. (6) Gutendorf, B.; Westendorf, J. Comparison of an array of in vitro assays for the assessment of the estrogenic potential of natural and synthetic estrogens, phytoestrogens and xenoestrogens. Toxicology 2001, 166, 79−89. (7) Van den Belt, K.; Berckmans, P.; Vangenechten, C.; Verheyen, R.; Witters, H. Comparative study on the in vitro/in vivo estrogenic potencies of 17β-estradiol, estrone, 17α-ethynylestradiol and nonylphenol. Aquat. Toxicol. 2004, 66, 183−195. (8) Tran, N. H.; Urase, T.; Ngo, H. H.; Hu, J.; Ong, S. L. Insight into metabolic and cometabolic activities of autotrophic and heterotrophic microorganisms in the biodegradation of emerging trace organic contaminants. Bioresour. Technol. 2013, 146, 721−731. (9) Gaulke, L. S.; Strand, S. E.; Kalhorn, T. F.; Stensel, H. D. 17αEthinylestradiol transformation via abiotic nitration in the presence of ammonia oxidizing bacteria. Environ. Sci. Technol. 2008, 42, 7622− 7627. (10) McAdam, E. J.; Bagnall, J. P.; Koh, Y. K. K.; Chiu, T. Y.; Pollard, S.; Scrimshaw, M. D.; Lester, J. N.; Cartmell, E. Removal of steroid estrogens in carbonaceous and nitrifying activated sludge processes. Chemosphere 2010, 81, 1−6. (11) Bagnall, J. P.; Ito, A.; McAdam, E. J.; Soares, A.; Lester, J. N.; Cartmell, E. Resource dependent biodegradation of estrogens and the role of ammonia oxidising and heterotrophic bacteria. J. Hazard. Mater. 2012, 239−240, 56−63. (12) Yoshimoto, T.; Nagai, F.; Fujimoto, J.; Watanabe, K.; Mizukoshi, H.; Makino, T.; Kimura, K.; Saino, H.; Sawada, H.; Omura, H. Degradation of estrogens by Rhodococcus zopf ii and Rhodococcus equi isolates from activated sludge in wastewater treatment plants. Appl. Environ. Microbiol. 2004, 70, 5283−5289. (13) O’Grady, D.; Evangelista, S.; Yargeau, V. Removal of aqueous 17α-ethinylestradiol by Rhodococcus species. Environ. Eng. Sci. 2009, 26, 1393−1400. (14) Shi, J. H.; Suzuki, Y.; Lee, B. D.; Nakai, S.; Hosomi, M. Isolation and characterization of the ethynylestradiol-biodegrading microorganism Fusarium proliferatum strain HNS-1. Water Sci. Technol. 2002, 45, 175−179. (15) Clara, M.; Kreuzinger, N.; Strenn, B.; Gans, O.; Kroiss, H. The solids retention timeA suitable design parameter to evaluate the capacity of wastewater treatment plants to remove micropollutants. Water Res. 2005, 39, 97−106. (16) Tchobanoglous, G.; Burton, F.; Stensel, H. D. Wastewater Engineering: Treatment and Reuse; McGraw-Hill Education: New York, 2003. (17) Kreuzinger, N.; Clara, M.; Strenn, B.; Kroiss, H. Relevance of the sludge retention time (SRT) as design criteria for wastewater treatment plants for the removal of endocrine disruptors and pharmaceuticals from wastewater. Water Sci. Technol. 2004, 50, 149−156. (18) Yuan, Z.; Blackall, L. L. Sludge population optimization: A new dimension for the control of biological wastewater treatment systems. Water Res. 2002, 36, 482−490.
ASSOCIATED CONTENT
* Supporting Information S
Description of analytical methods; details of the AS SBR operation for the three experimental phases; photomicrographs of AS flocs from the Phase I SBRs; supporting calculations for the fraction of active nitrifying biomass; example plots of total EE2 concentrations throughout batch degradation tests; and example linear regression showing how EE2 biodegradation rate coefficients were derived from batch degradation test data. This material is available free of charge via the Internet at http://pubs.acs.org.
■
Article
AUTHOR INFORMATION
Corresponding Author
*Phone: (206) 852-2190; e-mail: [email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This research was supported by the National Science Foundation projects CBET 1067744 and DGE-0718124, and the King County Wastewater Treatment Division Graduate Student Research Fellowship program. We sincerely acknowledge the assistance from Songlin Wang (University of Washington) with laboratory analyses, and Pardi Sukapanpotharam, John Smyth, and Bob Bucher (King County Metro, Wastewater Treatment Division) with WWTP reactors. 6166
dx.doi.org/10.1021/es405351b | Environ. Sci. Technol. 2014, 48, 6160−6167
Environmental Science & Technology
Article
(19) Jenkins, D.; Richard, M. G.; Daigger, G. T. Manual on the Causes and Control of Activated Sludge Bulking, Foaming, and Other Solids Separation Problems; IWA Publishing: London, UK, 2004. (20) Koh, Y. K. K.; Chiu, T. Y.; Boobis, A. R.; Scrimshaw, M. D.; Bagnall, J. P.; Soares, A.; Pollard, S.; Cartmell, E.; Lester, J. N. Influence of operating parameters on the biodegradation of steroid estrogens and nonylphenolic compounds during biological wastewater treatment processes. Environ. Sci. Technol. 2009, 43, 6646−6654. (21) Joss, A.; Andersen, H.; Ternes, T.; Richle, P. R.; Siegrist, H. Removal of estrogens in municipal wastewater treatment under aerobic and anaerobic conditions: Consequences for plant optimization. Environ. Sci. Technol. 2004, 38, 3047−3055. (22) Dytczak, M. A.; Londry, K. L.; Oleszkiewicz, J. A. Biotransformation of estrogens in nitrifying activated sludge under aerobic and alternating anoxic/aerobic conditions. Water Environ. Res. 2008, 80, 47−52. (23) Li, Y. M.; Zeng, Q. L.; Yang, S. J. Removal and fate of estrogens in an anaerobic-anoxic-oxic activated sludge system. Water Sci. Technol. 2011, 63, 51. (24) Suarez, S.; Lema, J. M.; Omil, F. Removal of pharmaceutical and personal care products (PPCPs) under nitrifying and denitrifying conditions. Water Res. 2010, 44, 3214−3224. (25) Zhang, Z.; Feng, Y.; Gao, P.; Wang, C.; Ren, N. Occurrence and removal efficiencies of eight EDCs and estrogenicity in a STP. J. Environ. Monit. 2011, 13, 1366−1373. (26) Gaulke, L. S.; Strand, S. E.; Kalhorn, T. F.; Stensel, H. D. Estrogen biodegradation kinetics and estrogenic activity reduction for two biological wastewater treatment methods. Environ. Sci. Technol. 2009, 43, 7111−7116. (27) Andersen, H. R.; Hansen, M.; Kjølholt, J.; Stuer-Lauridsen, F.; Ternes, T.; Halling-Sørensen, B. Assessment of the importance of sorption for steroid estrogens removal during activated sludge treatment. Chemosphere 2005, 61, 139−146. (28) Chudoba, J.; Grau, P.; Ottová, V. Control of activated-sludge filamentous bulkingII. Selection of microorganisms by means of a selector. Water Res. 1973, 7, 1389−1406. (29) Chudoba, J.; Č ech, J. S.; Farkač, J.; Grau, P. Control of activated sludge filamentous bulking: Experimental verification of a kinetic selection theory. Water Res. 1985, 19, 191−196. (30) van Niekerk, A. M.; Jenkins, D.; Richard, M. G. The competitive growth of Zoogloea ramigera and Type 021N in activated sludge and pure culture: A model for low F:M bulking. J. Water Pollut. Control Fed. 1987, 59, 262−273. (31) Cao, Y. S.; Alaerts, G. J. Influence of reactor type and shear stress on aerobic biofilm morphology, population and kinetics. Water Res. 1995, 29, 107−118. (32) Tan, D. T.; Arnold, W. A.; Novak, P. J. Impact of organic carbon on the biodegradation of estrone in mixed culture systems. Environ. Sci. Technol. 2013, 47, 12359−12365. (33) Ren, Y.-X.; Nakano, K.; Nomura, M.; Chiba, N.; Nishimura, O. Effects of bacterial activity on estrogen removal in nitrifying activated sludge. Water Res. 2007, 41, 3089−3096. (34) Vader, J. S.; van Ginkel, C. G.; Sperling, F. M. G. M.; de Jong, J.; de Boer, W.; de Graaf, J. S.; van der Most, M.; Stokman, P. G. W. Degradation of ethinyl estradiol by nitrifying activated sludge. Chemosphere 2000, 41, 1239−1243. (35) Shi, J.; Fujisawa, S.; Nakai, S.; Hosomi, M. Biodegradation of natural and synthetic estrogens by nitrifying activated sludge and ammonia-oxidizing bacterium Nitrosomonas europaea. Water Res. 2004, 38, 2323−2330. (36) Yi, T.; Harper, W. F. The Link between Nitrification and Biotransformation of 17α-ethinylestradiol. Environ. Sci. Technol. 2007, 41, 4311−4316. (37) De Gusseme, B.; Pycke, B.; Hennebel, T.; Marcoen, A.; Vlaeminck, S. E.; Noppe, H.; Boon, N.; Verstraete, W. Biological removal of 17α-ethinylestradiol by a nitrifier enrichment culture in a membrane bioreactor. Water Res. 2009, 43, 2493−2503. (38) Khunjar, W. O.; Mackintosh, S. A.; Skotnicka-Pitak, J.; Baik, S.; Aga, D. S.; Love, N. G. Elucidating the relative roles of ammonia
oxidizing and heterotrophic bacteria during the biotransformation of 17α-ethinylestradiol and trimethoprim. Environ. Sci. Technol. 2011, 45, 3605−3612. (39) Yi, T.; Harper, W. F., Jr. The effect of biomass characteristics on the partitioning and sorption hysteresis of 17α-ethinylestradiol. Water Res. 2007, 41, 1543−1553. (40) Yi, T.; Harper, W.; Holbrook, R.; Love, N. Role of particle size and ammonium oxidation in removal of 17α-ethinyl estradiol in bioreactors. J. Environ. Eng. 2006, 132, 1527−1529. (41) Benjamin, M. M.; Lawler, D. F. Water Quality Engineering: Physical/Chemical Treatment Processes; Wiley: New York, 2013. (42) Proposal for a directive of the European parliament and of the council amending Directives 2000/60/EC and 2008/105/EC as regards priority substances in the field of water policy, COM(2011)876; European Commission: Brussels, 2011.
6167
dx.doi.org/10.1021/es405351b | Environ. Sci. Technol. 2014, 48, 6160−6167
Influence of Bioselector Processes on 17α-Ethinylestradiol Biodegradation in Activated Sludge Wastewater Treatment Systems Ryan M. Ziels, Mariko J. Lust, Heidi L. Gough, Stuart E. Strand, and H. David Stensel* Department of Civil and Environmental Engineering, University of Washington, Seattle, Washington 98195-2700, United States S Supporting Information *
ABSTRACT: The removal of the potent endocrine-disrupting estrogen hormone, 17α-ethinylestradiol (EE2), in municipal wastewater treatment plant (WWTP) activated sludge (AS) processes can occur through biodegradation by heterotrophic bacteria growing on other organic wastewater substrates. Different kinetic and metabolic substrate utilization conditions created with AS bioselector processes can affect the heterotrophic population composition in AS. The primary goal of this research was to determine if these changes also affect specific EE2 biodegradation kinetics. A series of experiments were conducted with parallel bench-scale AS reactors treating municipal wastewater with estrogens at 100−300 ng/L concentrations to evaluate the effect of bioselector designs on pseudo first-order EE2 biodegradation kinetics normalized to mixed liquor volatile suspended solids (VSS). Kinetic rate coefficient (kb) values for EE2 biodegradation ranged from 5.0 to 18.9 L/g VSS/d at temperatures of 18 °C to 24 °C. EE2 kb values for aerobic biomass growth at low initial food to mass ratio feeding conditions (F/Mf) were 1.4 to 2.2 times greater than that from growth at high initial F/ Mf. Anoxic/aerobic and anaerobic/aerobic metabolic bioselector reactors achieving biological nutrient removal had similar EE2 kb values, which were lower than that in aerobic AS reactors with biomass growth at low initial F/Mf. These results provide evidence that population selection with growth at low organic substrate concentrations can lead to improved EE2 biodegradation kinetics in AS treatment.
■
a broader physiological capability at slower growth rates.15 Improved micropollutant removal by biodegradation would be expected for longer SRTs due to increased contact time and biomass for treatment.16 However, an evaluation of municipal AS treatment EE2 performance data found no clear improvement in EE2 removal with increased SRT.17 EE2 removal efficiencies varied from 60% to 80% within an SRT range of 5 to 12 days, and from 25% to 70% within an SRT range of 20 to 42 days.17 Since many factors could have contributed to this variation in EE2 removal efficiency, it is important to consider how other operational parameters besides SRT could also affect the AS microbial population composition and potentially its associated EE2 biodegradation activity. The heterotrophic microbial population in AS treatment can be affected by the type of growth substrate, operating conditions (such as SRT and dissolved oxygen concentration), and reactor configuration.18 A common AS process design approach is to use bioselectors to manipulate heterotrophic populations to obtain targeted WWTP performance goals. Such bioselector applications typically utilize reactor configurations
INTRODUCTION Natural and synthetic estrogens of human origin are endocrinedisrupting compounds (EDCs) capable of harming the reproductive processes of aquatic wildlife and fish.1−3 Biodegradation is the major removal mechanism for estrogen hormones in activated sludge (AS) treatment,4 with considerably lower biodegradation rates observed for the synthetic estrogen 17α-ethinylestradiol (EE2) than for the natural estrogens estrone (E1) and 17β-estradiol (E2).5 EE2 also exerts an estrogenic potency that is more than 1.25 times that of E2 and over 15 times that of E1.6,7 Other influent wastewater substrates are needed to support the growth of EE2-degrading bacteria because the low (ng/L) influent EE2 concentrations cannot support biomass growth.8 Both heterotrophic bacteria and ammonia-oxidizing bacteria (AOB) have been associated with EE2 removal, and more recent findings suggest that heterotrophic bacteria are of major importance for EE2 removal in municipal AS processes.9−11 A number of heterotrophic bacteria have also been shown to degrade EE2 in pure-culture.12−14 In spite of this understanding, there is no clear trend established between EE2 removal efficiency and AS process parameters. Increased micropollutant removal efficiency with longer AS solids retention times (SRT) has been proposed, and was attributed to the ability to enrich for more diverse bacteria with © 2014 American Chemical Society
Received: Revised: Accepted: Published: 6160
November 30, 2013 March 19, 2014 May 8, 2014 May 8, 2014 dx.doi.org/10.1021/es405351b | Environ. Sci. Technol. 2014, 48, 6160−6167
Environmental Science & Technology
Article
Table 1. Operating Conditions for the SBRs in the Three Experimental Phasesa,b,c parameter
phase I
SRT (d) temp. (°C) metabolic type kinetic type F/Mf (g COD/g VSS/d) fill time (min)
6.0 21.3 (1.7) aer. low 0.7 300
aer. high 15.3 15
phase II
aer.d highd 15.0 15
aer. high 10.2 15
phase III
8.7 17.5 (1.7) anox./aer. high 11.4 15
anaer./aer. high 9.6 15
aer. high 84.5 5
4.0 20.0 aer. low 1.5 300
anox./aer. low 1.2 300
a Average values shown with standard deviation in parentheses. bThe effect of kinetic bioselection was evaluated in Phase I, metabolic bioselection in Phase II, and combined effects in Phase III. cKey: SBR = sequencing batch reactor; F/Mf = feeding food to mass ratio; MLVSS = mixed liquor volatile suspended solids; SRT = solids retention time; Aer. = aerobic; Anox. = anoxic; Anaer. = anaerobic; High = high-F/Mf feeding conditions; and Low = low-F/Mf feeding conditions. dReceived a lower fraction of influent readily biodegradable COD.
coefficients normalized to mixed liquor volatile suspended solids (MLVSS) were used to characterize specific EE2 biodegradation kinetics for the different bioselector operating conditions.
and/or redox conditions to provide kinetic or metabolic selective growth pressures that enrich for bacterial communities with improved sludge settling characteristics and/or accomplish biological nutrient removal (BNR).18,19 Kinetic bioselectors promote substrate uptake at high initial food to mass ratio (F/Mf) feeding conditions and elevated substrate concentrations, and have been used to enrich for floc-forming bacteria and prevent filamentous sludge bulking.19 Reactor configurations with initial anoxic (no oxygen) or anaerobic (no oxygen or nitrate) contact zones result in metabolic bioselection by favoring substrate uptake and growth of bacteria capable of nitrate/nitrite reduction or high intracellular polyphosphate storage, respectively.19 Common metabolic bioselector designs for BNR include sequential anoxic and aerobic reactors for nitrogen removal, and sequential anaerobic and aerobic reactors for enhanced biological phosphorus removal (EBPR). The effect of the aforementioned AS population selection mechanisms on the activity of EE2-degrading microorganisms is poorly understood. A comparison of estrogen removal at two full-scale WWTPs with and without an anaerobic bioselector revealed differences in EE2 biodegradation activities,20 and emphasized the need to further evaluate the influence of AS metabolic bioselector designs on EE2 biodegradation kinetics. To the best of our knowledge, no study has compared the types of metabolic bioselectors commonly used in WWTPs using the same influent municipal wastewater source and SRT to determine if distinct heterotrophic BNR populations express different EE2 biodegradation kinetics. Nor has a study compared the EE2 biodegradation kinetics of AS systems operating with identical SRTs, but with substrate uptake occurring under high or low concentrations. In this study, we hypothesized that different AS reactor configurations that use kinetic and metabolic growth pressures to select for distinct heterotrophic populations would have different specific EE2 biodegradation kinetics. The objective of this study was to investigate the effect of metabolic and kinetic biological selection processes with similar SRTs on EE2 biodegradation kinetics during AS treatment of municipal primary effluent wastewater with EE2 at concentrations representative of that in municipal WWTP influents. Metabolic bioselection was examined by operating sequencing batch reactors (SBRs) with aerobic, anoxic, or anaerobic redox conditions during the fill and react period. Kinetic bioselection conditions were produced by feeding SBRs over a short duration with a high initial contact F/Mf for substrate uptake at a high concentration, or conversely feeding over the entire SBR react period with a low F/Mf for substrate uptake at a low concentration. Pseudo first-order EE2 biodegradation rate
■
MATERIALS AND METHODS Activated Sludge Reactor Configuration and Operation. Three parallel AS SBRs were operated within three experimental phases to examine the effects of metabolic and kinetic population selection mechanisms on EE2 biodegradation kinetics. Each SBR consisted of a 4-L Pyrex glass Erlenmeyer flask. The SBRs in phases I and II were operated at the West Point WWTP in Seattle, WA, and were fed from a primary effluent feed container that was replenished daily at the WWTP. The phase III SBRs were operated in a 20 °C walk-in environmental chamber at the University of Washington, and were fed King County South WWTP primary effluent wastewater that was replenished daily and was stored at 4 °C after its delivery every 7 days. Aeration was provided with an aquarium pump and a ceramic stone sparger, and mixed liquor nitrogen sparging was provided from a gas cylinder at 300 mL/ min during the anoxic and anaerobic metabolic bioselector periods to eliminate oxygen surface transfer. Magnetic stir bars mixed the AS SBRs throughout the fill and react periods. A summary of the operating conditions for the three SBR phases is given in Table 1, and further details on the reactor operation are given in the Supporting Information (SI). Phase I examined the impact of high and low F/Mf feeding conditions under aerobic conditions. Phase II compared aerobic, anoxic, and anaerobic metabolic bioselection under high F/Mf feeding. Phase III compared the effect of high and low F/Mf feeding conditions under aerobic conditions, and also compared the effect of aerobic and anoxic metabolic conditions under low F/Mf feeding conditions. The SBRs were seeded with activated sludge from local WWTPs at the start of each phase: City of Port Orchard, WA AS (aerobic process) for phase I, City of Snoqualmie, WA oxidation ditch (anoxic and EBPR processes) for phase II, and King County South WWTP (Renton, WA) AS (aerobic process) for phase III. The parallel reactors in all operating phases were fed equal volumes of wastewater daily, while the feed duration was selected to control the initial F/Mf loading (g COD/g VSS/d), as calculated according to Tchobanoglous et al.:16 F/M f =
(Q feed)(TCODin ) (MLVSS)(V )
(1)
where Qfeed is the flow rate during the feed period (L/d), TCODin is the average influent total chemical oxygen demand 6161
dx.doi.org/10.1021/es405351b | Environ. Sci. Technol. 2014, 48, 6160−6167
Environmental Science & Technology
Article
of the SBR cycle. The in situ biodegradation tests were not spiked, and thus the removal of EE2 during normal feeding conditions was followed in phase II. The average VSS concentration of the AS in each biodegradation test was determined from measurements taken at the beginning and end of each test. A total of three batch biodegradation tests were conducted with each AS reactor. Example plots of total EE2 concentration versus time obtained during batch biodegradation tests are given in the SI (Figure S1). Estrogen concentrations were not extensively monitored throughout the metabolic bioselector periods since negligible EE2 biodegradation has been shown to occur under anaerobic and anoxic conditions,21−25 which was corroborated by our observations (data not shown). Modeling EE2 Biodegradation and Solids Partitioning. EE2 biodegradation is described by a pseudo-first order kinetic model as a function of soluble estrogen and VSS concentrations.21,26
concentration (COD) (g/L), MLVSS is the average reactor mixed liquor volatile suspended solids concentration (g/L), and V is the reactor volume (L). A high F/Mf feed condition was implemented using short feeding durations that promoted substrate uptake at high initial concentrations (Table 1). Conversely, an operating condition with substrate uptake at a low concentration was conducted by adding feed every 3 to 5 min throughout the 5 h react/fill period to simulate a low-loaded complete-mix AS (CMAS) reactor (Table 1). These two feeding methods will be hereby referred to as “high-F/M f” and “low-F/M f” operating conditions, respectively. The SBR primary effluent wastewater feed was augmented with acetate and the estrogens E1, E2, and EE2. Equal quantities of acetate were added to the parallel SBRs for all operating phases at approximately 120 mg/L for phases I and II and 150 mg/L for phase III. The exception was one high-F/Mf reactor in phase I, which did not receive influent acetate in order to examine its performance with a lower fraction of influent readily biodegradable COD (rbCOD) and high-F/Mf feeding (SI). Sodium nitrate (10 to 25 mg N/L) was added to the feed of the anoxic/aerobic SBRs only (Phases II and III) to allow an increase in the amount of organic substrate removal during the anoxic period, and thus promote a higher fraction of denitrifiers in the heterotrophic biomass of the anoxic/aerobic bioselectors (SI). Phosphate (4 mg P/L) as mono and dibasic potassium phosphate was added to the feed of all SBRs in phase II to ensure growth of polyphosphate accumulating organisms during the anaerobic period of the anaerobic/aerobic bioselector reactor (SI). All SBRs had a total cycle time of 6 h, and the feed volume for each cycle was 1/4 of the total liquid volume for phases I and II and 1/5 for phase III. A quiescent settling time of 1-h occurred before supernatant was decanted throughout the final 8 min of the cycle with peristaltic pumps. Details of the AS SBR average influent and effluent characteristics are provided in the SI (Table S1). Influent concentrations of E1 and EE2 to the phase I SBRs averaged 192 ± 35 ng/L and 118 ± 25 ng/L (n = 12), respectively, while E2 was somewhat reduced in the phase I feed at 70 ± 54 ng/L (n = 12) possibly due to transformation during storage. Influent concentrations of E1, E2, and EE2 to the phase II SBRs averaged 213 ± 41 ng/L, 176 ± 28 ng/L, and 164 ± 25 ng/L, respectively (n = 21). Influent concentrations of E1, E2, and EE2 in phase III averaged 448 ± 100 ng/L, 348 ± 75 ng/L, and 301 ± 17 ng/L, respectively (n = 23). EE2 Batch Biodegradation Tests. All reactors were operated for a minimum time of three SRTs before conducting EE2 batch biodegradation tests, in which the total estrogen (soluble plus sorbed to solids) concentration was followed with time under aerobic conditions. Because of the semicontinuous feeding strategy for the low-F/Mf operation employed in phases I and III, EE2 biodegradation tests for those reactors were conducted as ex-situ batch tests. The ex-situ tests were conducted by transferring 500 mL of mixed liquor at the end of the aerobic SBR period to aerated and mixed (magnetic stirrer) 1-L flasks that were then spiked with 300 to 400 ng/L E1, E2, and EE2 (from stock solutions in Milli-Q water). Total estrogen concentrations were monitored for 5 h during the phase I batch tests and for 12 h during the phase III batch tests. EE2 biodegradation tests for the phase II reactors were performed as in situ batch tests, in which total estrogen concentrations were monitored throughout the aeration period
dEtot = −k bXVSSEsol dt
(2)
where Etot is the total EE2 concentration (ng/L), t is the time (days), kb is the pseudo first-order EE2 biodegradation kinetic coefficient (L/g VSS/d), XVSS is the MLVSS concentration (g VSS/L), and Esol is the soluble EE2 concentration in the bulk aqueous phase (ng/L). To determine values of kb from total EE2 concentration over time in batch biodegradation tests, the above equation was modified assuming rapid solid−liquid equilibrium of estrogens27 and linearized as follows:26 ⎛ E ⎞ ⎛ ⎞ XVSS tot ⎟⎟ = −k b⎜ ln⎜⎜ ⎟t ⎝ 1 + KFXVSS ⎠ ⎝ Etot,initial ⎠
(3)
where Kp is the EE2 solid−liquid partitioning coefficient (L/g VSS). Values of kb were calculated by multiplying the slope of the linear trend-line of ln(Etot/Etot,initial) versus time by (1+KpXVSS)/XVSS (see Figure S2 in the SI). The apparent EE2 Kp values for the SBR AS were determined by simultaneously taking soluble and total EE2 measurements near the end of each batch biodegradation test. The specific sorbed concentration (Esorb, ng/g) is the solid phase EE2 concentration divided by the MLVSS concentration:
Esorb =
Etot − Esol XVSS
(4)
and the apparent EE2 solid−liquid partitioning coefficient for AS solids was then calculated as follows: Kp =
Esorb(time = t ) Esol(time = t )
(5)
One-way Analysis of Variance (ANOVA) and Tukey’s test were used to compare average EE2 kb and Kp values between reactors using the StatPlus add-on to Microsoft Excel (2011). Estrogen Measurements. Estrogen samples were prepared in accordance with the method described by Gaulke et al.,9 except that 50 pg of each internal standard (d4E1, d4E2, and d4EE2) was added to the 0.5 mL samples before liquid extraction. Samples were taken in duplicate for soluble and total estrogen measurements. Samples for total estrogen measurements consisted of 0.5 mL mixed liquor. Mixed liquor samples for soluble estrogen measurements (6 mL) were centrifuged for 10 min at 3200 rpm and 4 °C, after which 0.5 mL of the 6162
dx.doi.org/10.1021/es405351b | Environ. Sci. Technol. 2014, 48, 6160−6167
Environmental Science & Technology
Article
supernatant was decanted for analysis. Estrogens were extracted from the 0.5 mL sample aliquots by adding 3 mL of ethyl acetate and shaking vigorously for 10 min. The resulting organic fraction was transferred to a clean sample tube and evaporated to dryness at 40 °C under nitrogen gas. Evaporated samples were then reconstituted with 100 μL of NaHCO3 buffer (pH 10.5) and 100 μL of 1 mg/mL dansyl-chloride in ACN, vortexed for 1 min, and held for 30 min at 60 °C prior to immediate analysis. A description of all reagents used in estrogen analysis is provided in the SI. Estrogen quantification was performed using a Shimadzu LC20AD high performance liquid chromatograph (LC) (Kyoto, Japan) combined with an Applied Biosystems 4000 Q-Trap tandem mass spectrometer (MS/MS) (Foster City, CA). The protocol for estrogen separation and quantification with LCMS/MS is described in the SI. The method limit of detection (LOD) for E1, E2, and EE2 was 1 ng/L, and the method limit of quantification (LOQ) was 5 ng/L. The natural estrogens, E1 and E2, were included in the feed to simulate a municipal wastewater estrogen matrix, yet were present below the LOQ in all reactor effluents and were degraded too rapidly for accurate determination of biodegradation kinetics during batch tests (data not shown).
Figure 1. Average pseudo first-order EE2 biodegradation kinetic coefficients (kb) for batch biodegradation tests with activated sludge from phase I aerobic-only kinetic bioselector reactors comparing the effect of growth at high and low initial substrate concentrations. The average temperature was 24.0 °C and SRT was 6 days. Greek symbols (α and β) indicate significantly different kb values (p < 0.05 with Tukey’s test). n = 3 biodegradation tests for each SBR. Average EE2 kb values shown in center of bars. Error bars indicate one standard deviation. F/Mf = initial food to mass ratio.
outcompeted for rbCOD at the high concentrations applied in high-F/Mf operating conditions. These observations do not eliminate the possibility that EE2-degrading bacteria may have also grown on endogenous decay products, yet the relative fraction of substrate produced by endogenous decay would have been similar in the phase I reactors due to their identical HRT and SRT. The highest EE2 kb of the phase I SBRs was associated with the lowest initial growth substrate concentration in the lowF/Mf reactor (Figure 1, Table 1). It is important to note that the SBR operation allowed the F/Mf to be reduced in the lowF/Mf reactor while maintaining a relatively low SRT of 6-days (Table 1). These results suggest that kinetic population selection is an important operational factor governing the growth and activity of EE2-degrading microorganisms in AS systems. A recent study on E1 biodegradation by Tan et al.32 also found that kinetic selection affected the estrogen biodegradation rate. They observed reduced E1 biodegradation at high organic substrate concentrations, and postulated that such conditions favored growth of r-strategist populations incapable of E1 biodegradation. Koh et al.20 hypothesized that improved EE2 biodegradation at low substrate growth conditions could have been attributed to the growth of Kstrategist heterotrophic populations. However, the range of F/M conditions investigated in that study was too limited (difference of only 0.05 g BOD/g MLVSS/d)20 to elucidate the effects of kinetic population selection on EE2 biodegradation activity. Our study investigated a wide range of initial growth substrate concentrations (difference in F/Mf of 14.3 g COD/g MLVSS/d in phase I SBRs, see Table 1), which resulted in a higher EE2 kb in the low-F/Mf reactor by a factor of 1.4 to 1.7 (Figure 1). These results provide the first clear evidence that population selection with growth at low organic substrate concentration can increase EE2 biodegradation kinetics independent of SRT. Increased estrogen biodegradation observed in AS systems has sometimes been attributed to cometabolic transformation by ammonia-oxidizing bacteria (AOB).33−38 AOB biomass fractions in the MLVSS in the SBRs were estimated based on the amount of ammonia oxidized (see SI). In phase I, all of the reactors were operated at the same SRT and the highest fraction of nitrifying biomass was estimated at 3.9% for the
■
RESULTS AND DISCUSSION Effect of Kinetic Bioselection on Specific EE2 Biodegradation Kinetic Coefficients. In order to test the effects of kinetic population selection on EE2 biodegradation activity in aerobic AS systems in phase I, the feeding duration was adjusted in the SBRs to alter their feeding F/Mf conditions (Table 1) and promote substrate uptake at low concentrations in the low-F/Mf aerobic reactor and high concentrations in two other high-F/Mf aerobic reactors. One of the high-F/Mf aerobic reactors was operated with less influent rbCOD (difference of 117 mg COD/L, SI Table S1) to determine the effects of wastewater composition on the pseudo first-order EE2 biodegradation rate coefficient (kb) in the high-F/Mf kinetic bioselector reactors. The different initial contact F/M f conditions employed in the reactors altered the substrate concentration available during growth, which resulted in differing sludge morphologies in the high-F/Mf reactors and low-F/Mf reactor (SI Figure S3). The low-F/Mf reactor sludge was characterized by an abundance of filamentous organisms, while the two high-F/Mf reactors were dominated by nonfilamentous sludge flocs (SI Figure S3). This observation was consistent with the kinetic selection theory proposed by Chudoba et al.,28 in which filamentous microorganisms were enriched in CMAS reactor configurations, while nonfilamentous bacteria were dominant in batch fed kinetic bioselector SBRs.29−31 EE2 kb values within the phase I reactors ranged from 11.1 to 18.9 L/g VSS/d (Figure 1), and were significantly different between the reactor designs (p = 0.001, ANOVA). Comparisons using Tukey’s test revealed that the EE2 kb in the lowF/Mf reactor was significantly higher than in the high-F/Mf reactor treating the same influent wastewater (p = 0.001) and the high-F/Mf reactor operated with less rbCOD (p = 0.006). However, no statistically significant difference was detected between the EE2 kb values in the two high-F/Mf aerobic reactors (p = 0.27), despite their different fractions of influent rbCOD. These results suggest that EE2-degrading populations in AS may grow better on rbCOD at the low concentrations promoted with low-F/Mf feeding conditions, but they may be 6163
dx.doi.org/10.1021/es405351b | Environ. Sci. Technol. 2014, 48, 6160−6167
Environmental Science & Technology
Article
high-F/Mf reactor with the lower influent rbCOD, while the lowest fraction of 2.2% was estimated for the low-F/Mf reactor and the high-F/Mf reactor with acetate-augmented influent wastewater (SI Table S2). Thus, the highest EE2 kb occurred in the AS configuration with the lowest nitrifying biomass fraction. Moreover, no significant correlation was observed between EE2 kb values and the fraction of nitrifying biomass in the total MLVSS for all experimental reactors operated in this study (R2 = 0.21) (SI Figure S4). At similar ng/L EE2 concentrations, Bagnall et al.11 showed that EE2 biodegradation did not change during AS treatment when the nitrogen source for a synthetic wastewater feed was changed from ammonia to nitrate, resulting in the subsequent loss of nitrification and a 99% decrease in AOB concentration. In addition, Gaulke et al.9 observed no EE2 removal by pure cultures of Nitrosomonas europaea and Nitrosospira multiformis at an EE2 concentration of 500 ng/L and 10 mg/L NH3N. However, at about 1000fold higher EE2 concentrations, Khunjar et al.38 calculated an EE2 biotransformation rate by N. europaea of 13.6 L/g CODAOB/d (19.3 L/g VSSAOB/d) at 200 μg/L EE2 and 15 mg/L NH3N. Assuming that the above biotransformation rate coefficient observed at the high EE2 concentration applied to the AOB biomass fractions in our reactors, the amount of EE2 transformation by AOB would only account for 2% to 7% of the mixed liquor EE2 kb values ranging from 5.0 to 18.9 L/g VSS/d. Therefore, it is likely that heterotrophic bacteria were primarily responsible for the EE2 biodegradation observed in this study. Effect of High-F/Mf Metabolic Bioselection on Specific EE2 Biodegradation Kinetic Coefficients. The high-F/Mf metabolic bioselector reactors operated in phase II removed a substantial portion of influent soluble COD (sCOD) during the metabolic bioselector fill/react time, and also achieved EBPR and nitrogen removal typical of their respective BNR bioselector types (SI Table S1). The average removal of influent sCOD was 93% and 95% (n = 17) in the anoxic/ aerobic and anaerobic/aerobic reactors, respectively, during their 1.5-h metabolic bioselection periods. The anoxic/aerobic reactor removed an average of 11 mg N/L more than the aerobic reactor (n = 18) (SI Table S1). The anaerobic/aerobic reactor removed an average of 260% more influent phosphorus (3.2 mg P/L removed) than the aerobic reactor (n = 23) (SI Table S1). No statistically significant difference was detected among the EE2 kb values of the phase II high-F/Mf metabolic bioselector reactors (Figure 2) (p = 0.10, ANOVA), despite that most of the influent biodegradable substrate removal occurred in the aerobic and metabolic bioselector processes. On the basis of the observed substrate removal mechanisms in these three parallel reactors, different heterotrophic population compositions were expected. A possible explanation for the similar kb values among the BNR systems is that the growth of EE2-degrading bacteria may have occurred on residual slowly biodegradable substrate and endogenous decay products in the aerobic period after the bioselector step. Combined Effects of Kinetic and Metabolic Bioselection Processes on Specific EE2 Biodegradation Kinetic Coefficients. In phase III, aerobic-only and anoxic/aerobic reactors were operated under a low-F/Mf kinetic bioselection condition, and an aerobic-only reactor was operated with a high-F/Mf kinetic bioselection condition. This experimental design allowed the effects of metabolic bioselection on EE2 biodegradation kinetics to be elucidated under low-F/Mf
Figure 2. Average pseudo first-order EE2 biodegradation kinetic coefficients (kb) for in situ batch biodegradation tests with the phase II AS reactors comparing the effect of metabolic bioselection with growth at high initial food to mass ratio (F/Mf) conditions. The average temperature was 18.3 °C and SRT was 8.7 days. Average EE2 kb of the phase II SBRs were not statistically different at the p = 0.05 level (with Tukey’s test). n = 3 biodegradation tests for each SBR. Average EE2 kb values shown in center of bars. Error bars indicate one standard deviation.
kinetic population selection conditions (in contrast to phase II), and validated the effects of kinetic population selection found in phase I using a different AS inoculum and primary effluent source. As shown in Figure 3, the EE2 kb value in the aerobic lowF/Mf reactor was more than double that of the high-F/Mf
Figure 3. Average pseudo first-order EE2 biodegradation kinetic coefficients (kb) for batch biodegradation tests with activated sludge from the phase III reactors comparing the effect of high and low initial food to mass ratio (F/Mf) aerobic feeding and the effect of aerobic versus anoxic metabolic bioselection conditions under low F/Mf feeding. The average temperature was 20.0 °C, and the SRT was 4.0 days. Greek symbols (α and β) indicate significantly different kb values (p < 0.05 with Tukey’s test). n = 3 biodegradation tests for each SBR. Average EE2 kb values shown in center of bars. Error bars indicate one standard deviation.
aerobic reactor (p = 0.0002). This result validated the observation of higher EE2 kb values observed in the aerobic low-F/Mf reactor in phase I (Figure 1), indicating that the effects of low substrate process selection on EE2 biodegradation kinetics were repeated with a different wastewater source. However, the EE2 kb of the low-F/Mf anoxic/aerobic reactor was not significantly different from the parallel high-F/Mf aerobic reactor (p = 0.06), suggesting that the growth of EE2 degraders under low substrate feeding conditions was hindered by anoxic versus aerobic redox conditions. As such, the EE2 kb value in the aerobic low-F/Mf reactor was 1.6 times that of the aerobic EE2 kb value in the anoxic/aerobic reactor (p = 0.0006), which was operated with the same low F/Mf (Figure 3). 6164
dx.doi.org/10.1021/es405351b | Environ. Sci. Technol. 2014, 48, 6160−6167
Environmental Science & Technology
Article
Table 2. Solid-Liquid EE2 Partitioning Coefficients (Kp) for Activated Sludge from Each SBR Experimental Operating Phasea,b parameter metabolic type kinetic type EE2 (Kp) (L/kg-VSS)
phase I SBRs (n = 5) aer. high 405 (163)
aer. low 682 (203)
phase II SBRs (n = 4)
aer.c highc 440 (154)
aer. high 554 (112)
anox./aer. high 684 (62)
phase III SBRs (n = 6)
anaer./aer. high 408 (57)
aer. high 303 (138)
aer. low 867 (76)
anox./aer. low 856 (142)
a Average values shown with standard deviation in parentheses. bKey: SBR = sequencing batch reactor; Aer. = aerobic; Anox. = anoxic; Anaer. = anaerobic; High = high-F/Mf feeding conditions; and Low = low-F/Mf feeding conditions cReceived a lower fraction of influent readily biodegradable COD
Table 3. Average Total EE2 Concentrations and EE2 Biodegradation Removal Efficiencies Based on Samples Taken at the End of the Aerobic Period of the SBR Cycle for the Three Experimental Phasesa,b parameter metabolic type kinetic type end-of-cycle total EE2 concentration (ng/L)d EE2 biodegradation removal efficiency (%)d,e
phase I SBRs (n = 7)
phase II SBRs (n = 5) c
phase III SBRs (n = 9)
aer. high 16 (2.3)
aer. low 20 (1.1)
aer. highc 18 (2.1)
aer. high 29 (9.4)
anox./aer. high 24 (4.8)
anaer./aer. high 30 (5.2)
aer. high 153 (34)
aer. low 139 (20)
anox./aer. low 189 (28)
88 (1.4)
85 (1.0)
80 (3.0)
81 (6.5)
83 (3.2)
80 (3.3)
52 (4.2)
57 (3.8)
42 (6.8)
a Average values shown with standard deviation in parentheses. bKey: SBR = sequencing batch reactor; Aer. = aerobic; Anox. = anoxic; Anaer. = anaerobic; High = high-F/Mf feeding conditions; and Low = low-F/Mf feeding conditions. cReceived a lower fraction of influent readily biodegradable COD. dSamples averaged over last 1.5 SRTs of operation. eBiodegradation efficiency calculated as (1-EE2end‑of‑cycle/EE2influent).
The calculation of EE2 kb values using Eq. 3 relies on the EE2 Kp value. A sensitivity analysis of the first-order EE2 biodegradation model revealed that a 50% decrease in the EE2 Kp of the phase III aerobic low-F/Mf sludge would yield an 8% decrease in its average EE2 kb. Even at this lower EE2 kb, the observed difference between the aerobic low-F/Mf reactor and high-F/Mf reactor would still remain significant (p = 0.0002). Therefore, the higher EE2 biodegradation rate coefficients observed in the aerobic low-F/Mf AS systems were not an artifact of their elevated EE2 Kp values. Effect of AS Reactor Configuration on EE2 Biodegradation Removal Efficiency. The average EE2 biodegradation removal efficiencies for the SBRs in the three experimental phases ranged from 42% to 88% (Table 3). While operation under aerobic low substrate growth conditions resulted in higher EE2 biodegradation kinetics (Figures 1 and 3), the EE2 removal efficiencies were not increased (Table 3). Despite the 1.7-fold increase in the first-order EE2 kb of the phase I low-F/Mf reactor (Figure 1), its semicontinuous feeding resulted in an EE2 biodegradation removal efficiency that was similar to the batch fed high-F/Mf aerobic reactors operated in parallel (p > 0.10) (Table 3). Similarly, the 2.2-fold increase in EE2 kb in the phase III aerobic low-F/Mf reactor relative to the high-F/Mf reactor (Figure 3) did not result in a significantly greater EE2 biodegradation removal efficiency (p > 0.10) (Table 3). The lack of correlation between EE2 biodegradation kinetics and removal efficiency is related to different reactor hydraulic characteristics simulated by the lowF/Mf and high-F/Mf reactor operations. The high-F/Mf SBR operation simulated a plug flow reactor (PFR), in which the rapid fill-time (Table 1) was followed by batch reactor biodegradation kinetics with a high initial EE2 concentration that decreased during the remaining aeration period. Greater EE2 removal rates occur at higher EE2 concentrations due to the first-order biodegradation kinetics (Eq. 2). However, the low-F/Mf SBR operation simulated a CMAS process with a relatively constant low EE2 concentration due to the semicontinuous feeding throughout the react period. Even though the specific EE2 biodegradation kinetic coefficients were greater for the low-F/Mf aerobic systems, the amount of
However, when the anoxic/aerobic and aerobic-only reactors were operated with a high F/Mf in phase II, the aerobic EE2 kb values were similar (Figure 2). Therefore, it appears that aerobic EE2 biodegradation kinetics were negatively impacted when biomass growth occurred at high initial substrate concentrations or under anoxic conditions. To the best of our knowledge, this is the first reported difference in aerobic EE2 biodegradation rate coefficients between anoxic/aerobic and aerobic AS systems operated with identical SRT, HRT, influent wastewater, and feeding regime. Effect of AS Reactor Configuration on EE2 Partitioning to Biosolids. We hypothesized that the kinetic and metabolic selective pressures applied in the AS reactor configurations in this study would enrich for microbial communities with differing physiological properties. The apparent EE2 solid−liquid partitioning coefficient values (Kp) of the low-F/Mf reactor AS in phases I and III were greater than the Kp values of the AS in the parallel high-F/Mf reactors (Table 2). This effect was highly significant in phase III (p < 0.0001, ANOVA), as the average EE2 Kp values of the AS in the low-F/Mf aerobic and anoxic/aerobic reactors were more than double the Kp values of the AS in the high-F/Mf aerobic reactor (p < 0.0001, p < 0.0001, respectively). However, the EE2 Kp values of the two low-F/Mf reactor sludges in phase III were not significantly different (p = 0.99), suggesting similar AS sorption characteristics in the two systems operating with low substrate concentrations. The observation of variation among EE2 partitioning coefficients with changes in floc or biomass growth conditions is consistent with previously reported results. For example, higher EE2 Kp values have been associated with smaller mean floc particle sizes and higher specific surface areas of AS biomass.39 EE2 Kp values were also reported to be greater for a membrane bioreactor (MBR) AS relative to SBR AS, an effect which was attributed to greater floc surface area and cell hydrophobicity in the MBR sludge.40 In the current study, we observed elevated EE2 biodegradation kinetic coefficients for AS with greater EE2 Kp values, but the relationship of enhanced EE2 biodegradation activity to sorption remains unclear. 6165
dx.doi.org/10.1021/es405351b | Environ. Sci. Technol. 2014, 48, 6160−6167
Environmental Science & Technology
■
removal was not improved due to operation only at a lower EE2 concentration. This has been illustrated in the classic case of requiring longer reaction times in a continuously stirred tank reactor (CSTR) to achieve similar effluent concentrations as a PFR operated with the same first-order reaction rate coefficient.41 An important finding of this research is that the AS reactor configuration affects the EE2 removal efficiency in two ways: (1) the microbial population selection determines the EE2 biodegradation kinetics, and (2) the reactor EE2 concentration affects its removal rate. The EE2 first-order biodegradation kinetics also point to improved EE2 removal in reactors with increased staging or plug-flow characteristics. As priority micropollutants such as EE2 may become subject to future water quality regulations (for example, see European Union suggested Environmental Quality Standards),42 it is increasingly important to understand the relationships between AS process design and EE2 biodegradation in WWTPs. This study showed for the first time that anoxic versus anaerobic metabolic bioselection did not impact EE2 biodegradation kinetics in AS, whereas kinetic population selection had a significant effect. BNR systems and high-F/Mf aerobic bioselectors achieved similar EE2 biodegradation kinetics, which were both lower than in low-F/Mf aerobic reactors. These results indicate the potential benefit of employing MBR processes (no gravitational settling) with low-F/Mf growth conditions for improved EE2 biodegradation kinetics that may also enrich for filamentous sludge with poor settling properties. Likewise, oxidation ditch processes can provide low-F/Mf substrate removal conditions that might favor higher EE2 biodegradation kinetics, while maintaining acceptable sludge settling characteristics.
■
REFERENCES
(1) Jobling, S.; Nolan, M.; Tyler, C. R.; Brighty, G.; Sumpter, J. P. Widespread sexual disruption in wild fish. Environ. Sci. Technol. 1998, 32, 2498−2506. (2) Purdom, C. E.; Hardiman, P. A.; Bye, V. V. J.; Eno, N. C.; Tyler, C. R.; Sumpter, J. P. Estrogenic effects of effluents from sewage treatment works. Chem. Ecol. 1994, 8, 275−285. (3) Routledge, E. J.; Sheahan, D.; Desbrow, C.; Brighty, G. C.; Waldock, M.; Sumpter, J. P. Identification of estrogenic chemicals in STW effluent. 2. In vivo responses in trout and roach. Environ. Sci. Technol. 1998, 32, 1559−1565. (4) Andersen, H.; Siegrist, H.; Halling-Sørensen, B.; Ternes, T. A. Fate of estrogens in a municipal sewage treatment plant. Environ. Sci. Technol. 2003, 37, 4021−4026. (5) Ternes, T. A.; Kreckel, P.; Mueller, J. Behaviour and occurrence of estrogens in municipal sewage treatment plantsII. Aerobic batch experiments with activated sludge. Sci. Total Environ. 1999, 225, 91− 99. (6) Gutendorf, B.; Westendorf, J. Comparison of an array of in vitro assays for the assessment of the estrogenic potential of natural and synthetic estrogens, phytoestrogens and xenoestrogens. Toxicology 2001, 166, 79−89. (7) Van den Belt, K.; Berckmans, P.; Vangenechten, C.; Verheyen, R.; Witters, H. Comparative study on the in vitro/in vivo estrogenic potencies of 17β-estradiol, estrone, 17α-ethynylestradiol and nonylphenol. Aquat. Toxicol. 2004, 66, 183−195. (8) Tran, N. H.; Urase, T.; Ngo, H. H.; Hu, J.; Ong, S. L. Insight into metabolic and cometabolic activities of autotrophic and heterotrophic microorganisms in the biodegradation of emerging trace organic contaminants. Bioresour. Technol. 2013, 146, 721−731. (9) Gaulke, L. S.; Strand, S. E.; Kalhorn, T. F.; Stensel, H. D. 17αEthinylestradiol transformation via abiotic nitration in the presence of ammonia oxidizing bacteria. Environ. Sci. Technol. 2008, 42, 7622− 7627. (10) McAdam, E. J.; Bagnall, J. P.; Koh, Y. K. K.; Chiu, T. Y.; Pollard, S.; Scrimshaw, M. D.; Lester, J. N.; Cartmell, E. Removal of steroid estrogens in carbonaceous and nitrifying activated sludge processes. Chemosphere 2010, 81, 1−6. (11) Bagnall, J. P.; Ito, A.; McAdam, E. J.; Soares, A.; Lester, J. N.; Cartmell, E. Resource dependent biodegradation of estrogens and the role of ammonia oxidising and heterotrophic bacteria. J. Hazard. Mater. 2012, 239−240, 56−63. (12) Yoshimoto, T.; Nagai, F.; Fujimoto, J.; Watanabe, K.; Mizukoshi, H.; Makino, T.; Kimura, K.; Saino, H.; Sawada, H.; Omura, H. Degradation of estrogens by Rhodococcus zopf ii and Rhodococcus equi isolates from activated sludge in wastewater treatment plants. Appl. Environ. Microbiol. 2004, 70, 5283−5289. (13) O’Grady, D.; Evangelista, S.; Yargeau, V. Removal of aqueous 17α-ethinylestradiol by Rhodococcus species. Environ. Eng. Sci. 2009, 26, 1393−1400. (14) Shi, J. H.; Suzuki, Y.; Lee, B. D.; Nakai, S.; Hosomi, M. Isolation and characterization of the ethynylestradiol-biodegrading microorganism Fusarium proliferatum strain HNS-1. Water Sci. Technol. 2002, 45, 175−179. (15) Clara, M.; Kreuzinger, N.; Strenn, B.; Gans, O.; Kroiss, H. The solids retention timeA suitable design parameter to evaluate the capacity of wastewater treatment plants to remove micropollutants. Water Res. 2005, 39, 97−106. (16) Tchobanoglous, G.; Burton, F.; Stensel, H. D. Wastewater Engineering: Treatment and Reuse; McGraw-Hill Education: New York, 2003. (17) Kreuzinger, N.; Clara, M.; Strenn, B.; Kroiss, H. Relevance of the sludge retention time (SRT) as design criteria for wastewater treatment plants for the removal of endocrine disruptors and pharmaceuticals from wastewater. Water Sci. Technol. 2004, 50, 149−156. (18) Yuan, Z.; Blackall, L. L. Sludge population optimization: A new dimension for the control of biological wastewater treatment systems. Water Res. 2002, 36, 482−490.
ASSOCIATED CONTENT
* Supporting Information S
Description of analytical methods; details of the AS SBR operation for the three experimental phases; photomicrographs of AS flocs from the Phase I SBRs; supporting calculations for the fraction of active nitrifying biomass; example plots of total EE2 concentrations throughout batch degradation tests; and example linear regression showing how EE2 biodegradation rate coefficients were derived from batch degradation test data. This material is available free of charge via the Internet at http://pubs.acs.org.
■
Article
AUTHOR INFORMATION
Corresponding Author
*Phone: (206) 852-2190; e-mail: [email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This research was supported by the National Science Foundation projects CBET 1067744 and DGE-0718124, and the King County Wastewater Treatment Division Graduate Student Research Fellowship program. We sincerely acknowledge the assistance from Songlin Wang (University of Washington) with laboratory analyses, and Pardi Sukapanpotharam, John Smyth, and Bob Bucher (King County Metro, Wastewater Treatment Division) with WWTP reactors. 6166
dx.doi.org/10.1021/es405351b | Environ. Sci. Technol. 2014, 48, 6160−6167
Environmental Science & Technology
Article
(19) Jenkins, D.; Richard, M. G.; Daigger, G. T. Manual on the Causes and Control of Activated Sludge Bulking, Foaming, and Other Solids Separation Problems; IWA Publishing: London, UK, 2004. (20) Koh, Y. K. K.; Chiu, T. Y.; Boobis, A. R.; Scrimshaw, M. D.; Bagnall, J. P.; Soares, A.; Pollard, S.; Cartmell, E.; Lester, J. N. Influence of operating parameters on the biodegradation of steroid estrogens and nonylphenolic compounds during biological wastewater treatment processes. Environ. Sci. Technol. 2009, 43, 6646−6654. (21) Joss, A.; Andersen, H.; Ternes, T.; Richle, P. R.; Siegrist, H. Removal of estrogens in municipal wastewater treatment under aerobic and anaerobic conditions: Consequences for plant optimization. Environ. Sci. Technol. 2004, 38, 3047−3055. (22) Dytczak, M. A.; Londry, K. L.; Oleszkiewicz, J. A. Biotransformation of estrogens in nitrifying activated sludge under aerobic and alternating anoxic/aerobic conditions. Water Environ. Res. 2008, 80, 47−52. (23) Li, Y. M.; Zeng, Q. L.; Yang, S. J. Removal and fate of estrogens in an anaerobic-anoxic-oxic activated sludge system. Water Sci. Technol. 2011, 63, 51. (24) Suarez, S.; Lema, J. M.; Omil, F. Removal of pharmaceutical and personal care products (PPCPs) under nitrifying and denitrifying conditions. Water Res. 2010, 44, 3214−3224. (25) Zhang, Z.; Feng, Y.; Gao, P.; Wang, C.; Ren, N. Occurrence and removal efficiencies of eight EDCs and estrogenicity in a STP. J. Environ. Monit. 2011, 13, 1366−1373. (26) Gaulke, L. S.; Strand, S. E.; Kalhorn, T. F.; Stensel, H. D. Estrogen biodegradation kinetics and estrogenic activity reduction for two biological wastewater treatment methods. Environ. Sci. Technol. 2009, 43, 7111−7116. (27) Andersen, H. R.; Hansen, M.; Kjølholt, J.; Stuer-Lauridsen, F.; Ternes, T.; Halling-Sørensen, B. Assessment of the importance of sorption for steroid estrogens removal during activated sludge treatment. Chemosphere 2005, 61, 139−146. (28) Chudoba, J.; Grau, P.; Ottová, V. Control of activated-sludge filamentous bulkingII. Selection of microorganisms by means of a selector. Water Res. 1973, 7, 1389−1406. (29) Chudoba, J.; Č ech, J. S.; Farkač, J.; Grau, P. Control of activated sludge filamentous bulking: Experimental verification of a kinetic selection theory. Water Res. 1985, 19, 191−196. (30) van Niekerk, A. M.; Jenkins, D.; Richard, M. G. The competitive growth of Zoogloea ramigera and Type 021N in activated sludge and pure culture: A model for low F:M bulking. J. Water Pollut. Control Fed. 1987, 59, 262−273. (31) Cao, Y. S.; Alaerts, G. J. Influence of reactor type and shear stress on aerobic biofilm morphology, population and kinetics. Water Res. 1995, 29, 107−118. (32) Tan, D. T.; Arnold, W. A.; Novak, P. J. Impact of organic carbon on the biodegradation of estrone in mixed culture systems. Environ. Sci. Technol. 2013, 47, 12359−12365. (33) Ren, Y.-X.; Nakano, K.; Nomura, M.; Chiba, N.; Nishimura, O. Effects of bacterial activity on estrogen removal in nitrifying activated sludge. Water Res. 2007, 41, 3089−3096. (34) Vader, J. S.; van Ginkel, C. G.; Sperling, F. M. G. M.; de Jong, J.; de Boer, W.; de Graaf, J. S.; van der Most, M.; Stokman, P. G. W. Degradation of ethinyl estradiol by nitrifying activated sludge. Chemosphere 2000, 41, 1239−1243. (35) Shi, J.; Fujisawa, S.; Nakai, S.; Hosomi, M. Biodegradation of natural and synthetic estrogens by nitrifying activated sludge and ammonia-oxidizing bacterium Nitrosomonas europaea. Water Res. 2004, 38, 2323−2330. (36) Yi, T.; Harper, W. F. The Link between Nitrification and Biotransformation of 17α-ethinylestradiol. Environ. Sci. Technol. 2007, 41, 4311−4316. (37) De Gusseme, B.; Pycke, B.; Hennebel, T.; Marcoen, A.; Vlaeminck, S. E.; Noppe, H.; Boon, N.; Verstraete, W. Biological removal of 17α-ethinylestradiol by a nitrifier enrichment culture in a membrane bioreactor. Water Res. 2009, 43, 2493−2503. (38) Khunjar, W. O.; Mackintosh, S. A.; Skotnicka-Pitak, J.; Baik, S.; Aga, D. S.; Love, N. G. Elucidating the relative roles of ammonia
oxidizing and heterotrophic bacteria during the biotransformation of 17α-ethinylestradiol and trimethoprim. Environ. Sci. Technol. 2011, 45, 3605−3612. (39) Yi, T.; Harper, W. F., Jr. The effect of biomass characteristics on the partitioning and sorption hysteresis of 17α-ethinylestradiol. Water Res. 2007, 41, 1543−1553. (40) Yi, T.; Harper, W.; Holbrook, R.; Love, N. Role of particle size and ammonium oxidation in removal of 17α-ethinyl estradiol in bioreactors. J. Environ. Eng. 2006, 132, 1527−1529. (41) Benjamin, M. M.; Lawler, D. F. Water Quality Engineering: Physical/Chemical Treatment Processes; Wiley: New York, 2013. (42) Proposal for a directive of the European parliament and of the council amending Directives 2000/60/EC and 2008/105/EC as regards priority substances in the field of water policy, COM(2011)876; European Commission: Brussels, 2011.
6167
dx.doi.org/10.1021/es405351b | Environ. Sci. Technol. 2014, 48, 6160−6167
