An enhanced biological phosphorus removal (EBPR) - Water Science ...

n24

13/2/01

12:34 pm

Page 183

C.Y. Dassanayake* and R.L. Irvine* * Department of Civil Engineering and Geological Sciences, University of Notre Dame, Notre Dame, IN 46556, USA Abstract A control strategy was developed for enhanced biological phosphorus removal (EBPR) in a Sequencing Batch Reactor (SBR). Unlike past research that focused on maximizing polyhdroxyalkanoate (PHA) formation during the anaerobic period, this study investigated some of the factors that govern aerobic PHA dynamics and its efficient regulation during phosphate (P) uptake. Influent COD, influent P, and the time for aeration were critical factors that governed PHA use and P uptake during aerated react. Unnecessary PHA oxidation (i.e., in the absence of extracellular P) occurred if the time for aerated react exceeded the time required for P uptake. By adjusting the aeration time to that required for P uptake, residual PHA was sustained in the SBR and excess phosphate uptake reaction potential (PRP) was generated for use during transient influent excursions in P. Unlike space oriented systems, the time for react is simply adjusted in the SBR. Because residual PHA is easily maintained once achieved, high influent COD events can be harnessed to increase or sustain excess PRP for management of expected variations in influent P. Keywords Activated sludge; EBPR; nutrient removal; phosphorus removal; Sequencing Batch Reactor; SBR

Introduction

Research during the past three decades to advance the state of knowledge of enhanced biological phosphorus removal (EBPR) allowed development of stable EBPR processes and control strategies. Nevertheless, fundamental questions regarding the organisms needed and metabolic pathways involved remain (Wentzel et al., 1991; Jenkins and Tandoi, 1991; Smolders et al., 1994; Philipe and Daigger, 1998). This lack of fundamental information has limited the development of sophisticated start-up and general operation EBPR control strategies. Selection and enrichment strategies for stable EBPR performance in Sequencing Batch Reactors (SBRs) were investigated and developed by Manning and Irvine (1986). Research has also focused on the metabolic factors that contribute to stable and optimal EBPR performance. For example, the extent of PHA formed in the anaerobic period ultimately governs the extent of P uptake in the aerobic period (Wentzel et al., 1991; Philipe and Daigger, 1998). Accordingly, recent EBPR research was aimed at maximizing PHA formation with different organic substrates and reactor configurations (Wentzel et al., 1985; Carucci et al., 1995; Randall et al., 1997; Tasli et al., 1997). Other factors affecting PHA formation were also investigated including disruption of EBPR because of substrate competition between EBPR organisms and denitrifiers and how the nature and dynamics of intracellular polyphosphate pools can maximize substrate uptake and PHA formation (Barker and Dold, 1996; Lindrea et al., 1998, Dassanayake et al., 1998). PHA formation under anaerobic conditions has been studied more extensively than under aerobic conditions. The efficient use of PHA during P uptake in the subsequent aerobic period, however, is also critically important in the development of EBPR control strategies. Specifically, control strategies that will minimize the ratio between the PHA oxidized and P removed and/or maintain and regulate residual PHAs (if present after P uptake) will enhance performance of EBPR systems. The potential for residual P uptake will provide

Water Science and Technology Vol 43 No 3 pp 183–189 © IWA Publishing 2001

An enhanced biological phosphorus removal (EBPR) control strategy for sequencing batch reactors (SBRs)

183

n24

13/2/01

12:34 pm

Page 184

C.Y. Dassanayake and R.L. Irvine

excess P uptake reaction capacity and serve as a valuable control parameter during transient influent P excursions. The predominant EBPR metabolic pathway present in the consortium enriched will dictate the extent of P removal and the PHA dynamics under aerobic conditions. If that pathway involves glycogen, PHAs needed for P uptake are used instead to replenish glycogen reserves depleted during PHA formation that took place during the previous anaerobic period (Smolders et al., 1994; Dassanayake et al., 1998). Control strategies that curtail unnecessary oxidation of PHA in the absence of P for uptake is also a concern that must be addressed. This study investigates some of the factors governing PHA dynamics in the aerobic period and develops a control strategy for EBPR SBR similar to that described by Manning and Irvine (1986). Materials and methods

Operating Strategy IV developed by Manning and Irvine (1986) was used to enrich EBPR organisms in an SBR from a municipal seed. This strategy employed an 8 h cycle having an instantaneous fill, 2 h mixed react, 4 h aerated react, 1 h settle, and 1 h for draw plus idle. The hydraulic retention time was 0.5 d, sludge retention time, 20 d, and minimum and maximum liquid volumes, 1 L and 3 L, respectively. A synthetic waste consisting of the following constituents (ppm) was used: CH3COONa · 3H2O (230.6), C6H12O6 (100), hydrolyzed casein (100), FeCl2 · 4H2O (3.5), CaCl2 · 2H2O (3.7), MgCl2 · 6H2O (8.3), MnSO4 · H2O (3.1), K2SO4 (2.2), KH2PO4 (45.3), chemical oxygen demand (COD) (330), TKN (13), and soluble phosphate, as P (13). Nitrification and denitrification were minimized by omitting the 27 ppm NH3–N used by Manning and Irvine (1986). Soluble phosphate was determined using Standard Method 4500- P D (stannous chloride) (APHA); COD, using COD digestion vials commercially available through Hach Chemical company (Sunnyville, CA); pH, using a calibrated Accumet gel electrode and Orion pH meter (Fisher Scientific, IL); and nitrate, using the aromatic nitration procedure of Rodier (1975); and ammonia, using Nesslerization (Hach, 1992). Unless otherwise stated, 1.5–3.0 mL samples of mixed liquor were collected, placed in 1.5 mL Eppendorf centrifuge vials (Fisher Scientific, IL), and centrifuged for 1 minute at 10,000 rpm in a 5415 C Eppendorf centrifuge (Fisher Scientific, IL) prior to the analysis of soluble constituents. PHB and glycogen were measured from the resultant centrifuged pellets (Braunegg et al., 1978; Scott and Melvin, 1953). 600 mL of 4500 mg/L MLSS was taken at the end of aerobic react and used for both anaerobic and aerobic EBPR metabolic studies. The mixed liquor was settled for 1 hour, drawn down to 200 mL, dosed with 400 mL of a stock solution of acetate to 600 mg/L, and mixed under anaerobic conditions for 3.5 hours. In order to remove residual soluble phosphate and COD to below detection levels the mixed liquor was mixed for one minute with the mineral constituents of the feed minus P and settled for one hour before withdrawing the resulting supernatant. The mixed liquor volume was brought up to 600 mL with P free mineral medium and aerated for 4 hours. Samples were taken every half-hour and every hour during the anaerobic and aerobic periods, respectively, and analyzed for extracellular (COD, P) and intracellular compounds (PHB and glycogen). To determine the phosphate uptake reaction potential (PRP), 100 mL of 4500 mg/L MLSS was taken at the end of aerobic react, placed in a 150 mL Erlenmeyer flask, dosed with a P stock solution to 20 mg/L as P and then aerated. Samples were taken at half-hour intervals until residual soluble phosphate plateau was reached. Presentation of results

184

Figure 1 shows typical P, COD, ammonia, and nitrate profiles during fill and react for the EBPR consortium enriched in the SBR. During the 2 hour anaerobic mixed react period

n24

13/2/01

12:34 pm

Page 185


Figure 1 Typical cycle in the EBPR SBR

Figure 2 Polyphosphate, carbohydrate and PHB dynamics in the EBPR mixed liquor

following instantaneous fill, P levels increased from 8.6 to 58 mg/L as P (Figure 1A) and COD decreased from 220 mg/L to 40 mg/L (Figure 1B). The 8.6 mg/L as P at the end of instantaneous fill resulted from dilution of the influent 13 mg/L P. During mixed react, ammonia as N was produced to about 2 mg/L (Figure 1C). At the end of aerobic react, the effluent P was below detection (7 µg/ L as P) while effluent soluble COD was about 10 mg/L. The ammonia produced during the mixed react period was nitrified to nitrate (Figure 1D) during aerated react which was, in turn, denitrified within the first few minutes of mixed react of the next cycle. The EBPR consortium was removed from the SBR and tested under anaerobic conditions in the presence of excess acetate. As shown in Figure 2, the EBPR mixed liquor released 90 mg/L polyphosphate P, produced 400 mg/L PHB (as 3-hydroxybutyrate), and consumed 250 mg/L (as glucose) carbohydrate during the 3.5 hour anaerobic period while the extracellular COD decreased from 600 mg/L to 320 mg/L. In order to destroy the sludge’s ability for P uptake, it was rinsed free of extracellular P and COD immediately after the 3.5 h anaerobic period and aerated. As depicted in Figure 2, PHB dropped to below detection as the glycogen reserves were restored.

185

n24

13/2/01

12:34 pm

Page 186

C.Y. Dassanayake and R.L. Irvine Figure 3 Performance of the EBPR consortium during transient loading studies Table 1 Summary of parameter changes made during the EBPR loading studies

186

Phase

Influent COD (mg/L)

Influent P (mg/L as P)

Time for aerated react (h)

A

330.0

13.0

4.0

B

330.0

22.0

4.0

C

660.0

22.0

4.0

D

660.0

26.0

4.0

E

660.0

13.0

4.0

F

330.0

13.0

4.0

G

660.0

13.0

4.0

H

330.0

13.0

4.0

I

660.0

13.0

3.2

J

330.0

13.0

3.2

K

330.0

13.0

4.0

The EBPR consortium in the SBR was subjected to a series of transient loading conditions. The influent and effluent concentrations of COD and P and the PRP resulting from these transients are illustrated in Figure 3. The various react times and influent concentrations of COD and P imposed are summarized in Table 1 as phases B through K. Phase A represents the conditions employed before beginning the transient loading studies and is intended to serve as a benchmark or control for these studies. As can be seen from Figure 3 and Table 3, the phase A react time of 4 h and influent concentrations of COD and P of 330 mg/L and 13 mg/L, respectively, produced an effluent P that was consistently below detection and an available PRP of zero. The changes in COD, P, and time for react were designed to show how knowledge of PRP could be used to predict phosphorus removal efficiency. When the influent P was increased to 22 mg/L in phase B, the effluent P quality immediately deteriorated to 5 mg/L and the PRP remained zero. In phase C, the influent P was held at 22 mg/L and influent COD increased to 660 mg/L. This caused the effluent P to decrease below detection and the PRP to increase to 7.5 mg/L. In phase D, the influent COD was kept at 660 mg/L and the influent P was increased to 26 mg/L. As the PRP steadily declined to zero, the effluent P remained below detection. When the PRP reached zero, however, the

n24

13/2/01

12:34 pm

Page 187


effluent P began deteriorating and returned to 5 mg/L after about 4 days. As expected, the effluent P immediately improved to below detection in phase E when the influent P was decreased to 13 mg/L. Because of the COD remaining in excess at 660 mg/L, the PRP increased to about 12.0 mg/L. In phase F, the SBR was brought back to the initial conditions (i.e., those of phase A). Within three days the PRP decreased to zero while the effluent P remained below detection. A subsequent increase in influent COD to 660 mg/L in phase G for the same influent P increased the PRP to 12.1 mg/L with the effluent P remaining, of course, below detection levels. In order to test the repeatability of the results obtained for phase F, phase H was initiated. As can be seen from Figure 3, the results from phases F and H are nearly identical. The influent COD was increased from 330 to 660 mg/L in phase I just as it had in phase G. In this case, however, the time for aerated react was decreased from 4 h to 3.2 h, the time that had been needed for P to reach detection levels in phase G. Interestingly, the PRP increased to 12.5 mg/L while the effluent P remained below detection. In phase J, the influent COD was decreased to 330 mg/L and the influent P and aerated react were maintained fixed at 13.0 mg/L and 3.2 h, respectively. With this reduced time for aerobic react the effluent P remained below detection and the PRP leveled at 12.3 mg/L compared to zero in phase H when the time for aerated react was 4 h. When the aerated react was returned to 4 h in phase K, the PRP decreased to zero without deteriorating the effluent P just as it had done in phases A, F, and H. Discussion of results

When implemented in the studies described herein, operating strategy IV of Manning and Irvine (1986) successfully enriched for a stable mixed culture that displayed typical EBPR behavior. As can be seen from Figures 1 and 2, luxurious polyphosphate was released, COD was consumed (Comeau et al., 1986), and PHB was formed, all during anaerobic conditions, and P present at the end of anaerobic react was completely removed during the aerobic period. Figure 2 also supports the relationship between PHB and glycogen proposed by Wentzel et al., (1991) and Arun et al. (1988); glycogen supplies reducing equivalents for PHB formation during anaerobic conditions (presumably via glycolysis) and glycogen reserves are replenished during the aerobic period (Dassanayake et al., 1998; Dassanayake, 1999). The transient loading studies conducted on the SBR’s EBPR consortium clearly indicate that the dynamics involving PHB is a function of several factors including influent COD, influent P, glycogen (Dassanayake, 1999), and the time for aerated react. As can be inferred by reviewing Figures 1, 2, and 3, the influent COD sequestered as PHB is completely consumed during the aeration period if and whenever the final PRP turns out to be zero. The final PRP, then, is an obvious candidate for selection as a surrogate measure of PHB content. This notion is supported, for example, by a simple review of phases A and B. Since the PRP at the end of phase A was zero, an influent P greater than 13 mg/L would, as was found from experimentation, likely result in elevated levels of P in the discharge. This is expected because PHB is required for P consumption. Accordingly an increase in influent P without a concomitant increase in COD should and did result in a reduction in PRP (see phase D). In fact, there will not be a net increase in the mass fraction of PHB unless the feed COD is raised or the aerated react time reduced. Indeed, PHB should increase to a stoichiometric limit that is defined by the operating strategy selected, the specific raw waste constituents present, and the EBPR consortium enriched. Likewise, a reduction in the time for aerated react will decrease the net consumption of PHB. In the absence of extracellular P, PHB is used as an energy source for general cellular activities (Figure 2) and, except for the replenishment of glycogen (Dassanayake, 1999), not for the properties that provide EBPR consortia with their competitive edge, i.e., the transport of P and the synthesis of polyphosphates. The consistency of the interactions among PHB levels (as illustrated by

187

n24

13/2/01

12:34 pm

Page 188


PRP), influent COD, effluent P, and time for aerated react is easily seen by comparing any two adjacent phases in Figure 3. The development of a fundamental understanding of these interactions can be obtained by reviewing the PRP data. Doing so will allow ready control of EBPR performance potential because strategies that regulate the mass of oxygen consumed and maximize P removal in SBRs are easily implemented. The net change in the mass of PHB depends upon a number of factors including the time that the biosolids are aerated in excess of that required for P uptake, intracellular reducing equivalents, and the availability of suitable electron donors. Excess aeration time will not only reduce the PRP but also decrease the relative abundance of EBPR organisms via both under-utilization of P made available for consumption during transient loads and unwanted endogenous decay. The operating flexibility of the SBR can be used to minimize unnecessary oxidation of PHB by increasing either the ratio of mixing to aeration or the time for aeration by transferring time in aerated react to idle or extending the cycle time by increasing the total volume decanted. Such adjustments to the aerobic detention time are not as easily accomplished in space-oriented processes. The availability of electron donors can be supplied through the addition of supplemental COD. While potentially costly, the use of a supplemental carbon source such as acetate can be incorporated into the EBPR control strategy to increase the PRP level to that required for meeting daily P load variations, especially if concomitant biological nitrogen removal is needed. Previous studies conducted on the EBPR mixed culture enriched in this study clearly demonstrated that glycolysis plays a crucial role in PHA formation and polyphosphate release (Dassanayake, 1999). In the absence of iodoacetate (a known glycolysis and gluconeogenesis inhibitor) a polyphosphate release of about 92 mg/ L as P occurred while simultaneously producing 390 mg/ L PHA (as 3-HB) and consuming 280 mg/ L carbohydrates as glucose (Dassanayake, 1999). However, when glycolysis was inhibited with iodoacetate, no polyphosphate release, COD utilization, carbohydrate consumption, or PHA production occurred. Hence a decrease in the amount of glycogen formed reduced the amount of PHA formed, and ultimately the PRP. Under aerobic conditions, PHA was utilized for both extracellular phosphate uptake and glycogen formation (Dassanayake, 1999). Hence, when developing a reliable EBPR control strategy, the aeration time needs to be such that adequate reserves of both glycogen and PHA are maintained. For a fixed organic load and time of aeration, a change in PRP is less predictable during an increase than a decrease in influent P. For example, the reduction in PRP resulting from the increase in influent P was greater than the change in influent P (see phase D). Based on the excess PRP consumed and the decline in quality of effluent P in phase D, either the rate of PHA oxidation or the mass of PHA oxidized per mass of phosphate consumed or both may increase when the influent P is increased. A decrease in effluent P, on the other hand, caused an increase in PRP that was comparable to the change in influent P (see phase E). Additional studies that further refine PRP monitoring techniques are clearly warranted. Conclusions

188

Several conclusions can be made from the results presented in this study. • PHA oxidation occurs in EBPR organisms in the absence of extracellular phosphate. • Influent COD, influent phosphate, and the time for aerated react affect the dynamics of PHA oxidation. • Unnecessary oxidation of residual PHA during aerated react can be prevented by aligning the time for aerated react with the time required for P uptake. • PRP (i.e., residual phosphate uptake potential) is a surrogate measure of PHB content. • The ability to predict EBPR performance potential can be obtained from the review of PRP data.

n24

13/2/01

12:34 pm

Page 189

References Arun, V., Mino, T. and Matsuo, T. (1988). Biological Mechanism of Acetate Uptake Mediated by Carbohydrate Consumption in Excess Phosphorus Removal Systems. Wat. Res., 22(5), 565–572. Barker, P.S. and Dold, P.L. (1996). Denitrification Behaviour in Biological Excess Phosphorus Removal Activated Sludge Systems. Wat. Res., 30(4), 769–780. Braunegg, G.B., Sonnleitner, J. and Lafferty, R. M. (1978). A Rapid Gas Chromatographic Method for the Determination of Poly-β-hydroxybutyric Acid in Microbial Biomass. Bio. Technol., 6, 29–37. Carucci, A., Lindrea, K., Majone, M. and Ramadori, R. (1995). Dynamics of the Anaerobic Utilization of Organic Substrates in an Anaerobic/ Aerobic Sequencing Batch Reactor. Wat. Sci. Tech., 31(2), 35–43. Dassanayake, C.Y. (1999). Monitoring and Control of Physiological State in Mixed Cultures, Ph.D. Thesis, University of Notre Dame, Notre Dame, Indiana, USA. Dassanayake, C.Y., Irvine, R.L. and Chiavola, A. (1998). Intracellular Dynamics of Polyphosphates (PolyP) and Other Energy Reserves in Biological Phosphorus Removing Microorganisms (BioP organisms). Proceedings, European Conference on New Advances in Biological Nitrogen and Phosphorus Removal Plants for Municipal or (Agro)Industrial Wastewaters, Narbonne, France. Filipe, C.D.M. and Daigger, G.T. (1998). Development of a Revised Metabolic Model for the Growth of Phosphorus Accumulating Organisms. Wat. Env. Res., 80, 67–79. Jenkins, D. and Tandoi, V. (1991). The applied microbiology of enhanced biological phosphate removal – the accomplishments and needs. Wat. Res., 25(12), 1471–1478. Lindrea, K.C., Lockwood, G.L. and Majone, M. (1997). The Distribution and Movement of Polyphosphate and Associated Cations in Sludges from NDEBPR Plants in Different Configurations at the Pilot Scale. Wat. Sci. Tech., 37(4–5), 555–562. Manning, J.F. and Irvine, R.L. (1985). The biological removal of phosphorus in a sequencing batch reactor. J. Wat. Poll. Cont. Fed., 57(1), 87–94. Randall, A.A., Benefield, L.D., Hill, W.E., Nicol, J., Boman, G.K. and Jing, S. (1997). The effect of VFAs on enhanced biological phosphorus removal and population structure in anaerobic/aerobic sequencing batch reactors. Wat. Sci. Tech., 35(1), 153–160. Rodier, J. (1975). L’analyze de l’eaux: euax naturelles, eaux residuaires, eaux de mer, 5 edn. Scott, T.A. and Melvin, E.H. (1953). Determination of Dextran with Anthrone. Anal. Chem., 11, 1656–1661. Smolders, G.J.F., van der Meij, J., van Loosdrecht, M.C.M. and Heijnen, J.J. (1994). Stoichiometric model of the aerobic metabolism of the biological phosphate removal process. Biotechnol. Bioeng., 44, 837–842. Standard Methods for the Examination of Water and Wastewater (1995). 19th edn, American Public Health Association/American Water Works Association/Water Environment Federation, Washington DC, USA. Tasli, R., Artan, N. and Orhon, D. (1997). The influence of different substrates on enhanced biological phosphorus removal in a sequencing batch reactor. Wat. Sci. Tech., 35(1), 75–80. Wentzel, M.C., Dold, P.L., Ekama, G.A. and Marais, G.V.R. (1985). Kinetics of biological phosphorus release. Wat. Sci. Tech., 17(11–12) 57–71. Wentzel, M. C., Lötter, L.H., Loewenthal, R.E. and Marais, G.v.R. (1991). Evaluation of biochemical models for biological excess phosphorus rRemoval. Wat. Sci. Tech., 23(4–5), 567–572.


• Residual PHA can be utilized to remove excess P by increasing the time for aerated react. Without residual PHA, the PRP is zero and an increase in the time for aerated react will have no positive impact on P removal. • The operating flexibility of the SBR allows for easy adjustment of the time for aeration and hence efficient regulation of PHA during aerated react. Idle time or increased cycle time of can be used to provide additional aeration, if required. • Glycogen is needed for PHA production. When glycolysis was inhibited with iodoacetate, no polyphosphate release, COD utilization, carbohydrate consumption, or PHA production occurred. • Because residual PHA is easily maintained once achieved, high influent COD events can be harnessed to increase or sustain excess PRP for management of expected variations in influent P.

189

n24

13/2/01

12:34 pm

Page 190