Investigation of Biological Phosphorus Removal in a Sequencing Batch Reactor

INVESTIGATION OF BIOLOGICAL PHOSPHORUS REMOVAL IN A SEQUENCING BATCH REACTOR

RESEARCH FOR CEE 453:

LABORATORY RESEARCH IN

ENVIRONMENTAL ENGINEERING

Peter Burns

Alex Mikszewski

Brett Bovee

SPRING 2004

Abstract

Enhanced Biological Phosphorus Removal (EBPR) is a biological alternative to the more commonplace chemical phosphorus treatment methods. Research is ongoing to maximize EBPR removal efficiencies and make the treatment system a viable alternative. This experiment attempted to construct a functional EBPR wastewater plant in a laboratory setting to test phosphorus removal efficiency and compare it to large-scale systems. A small, software controlled treatment plant was constructed and operated for several weeks. Results showed a definite improvement in phosphorus removal, which was comparable to published values. The EBPR in this experiment attained a 70% removal rate. It is possible to improve removal efficiency by increasing microbial selection time and creating a favorable environment for phosphorus assimilating organisms, and such improvements are left for further study.

Introduction

The need for increased phosphorus removal from municipal and industrial wastewater is becoming more and more pertinent. Phosphorus is most often the limiting nutrient for algal growth in freshwater bodies. The over-fertilization of freshwater with phosphorus is known as eutrophication, which causes the unnatural stimulus of algae and weeds in water. The respiration and decomposition of algae in eutrophic waters depletes dissolved oxygen, causing “extinction zones” where no aquatic animals can survive (Lind, 1998). In order to protect freshwater bodies from destruction, efficient phosphorus removal techniques must be developed and implemented. In January 2001, the United States Environmental Protection Agency (EPA) published new recommendations for water quality nutrient criteria. The EPA threshold effluent phosphorus concentrations range from 0.0076 mg/L to as low as 0.010 mg/L depending on the sensitivity of the region (Lesjean 2003). As typical secondary effluent without phosphorus removal has total phosphorus concentrations ranging from 2.5 to 6 mg/L, advanced treatment is essential to comply with EPA recommendations (PA DEP, 2004).

A long used technique for phosphorus removal is chemical precipitation through addition of a precipitant like ferric chloride. Chemical removal processes typically get high removal efficiencies and are easily implemented. However, chemical additives are very expensive and can cause severe contamination of sewage sludges (WERF, 2003). As a result, more current wastewater engineering research is exploring the potential of enhanced biological phosphorus removal (EBPR) processes. EBPR treatment plants select for specific phosphorus accumulation organisms (PAOs) that efficiently uptake free phosphorus in the wastewater (WERF, 2003). An initial anaerobic stage allows the PAOs to adhere to organic matter, releasing cellular phosphorus through expenditure of energy. Upon aeration, the cells accumulate large amounts phosphorus for use as a substrate for energy production and storage (Weber-Shirk, 2004). EBPR processes are still mysterious in that the specific PAOs are as of now unidentifiable. Some publications single out Acineotobacter or Microlunatusphosphovorus as the predominant PAOs, but the subject is still up to debate (Cloete, 2003). The main problem with current EBPR use is that the required, ultra high removal efficiencies are not guaranteed. PAOs compete with gram negative G bacteria and other organisms, and thus can have population limits. Furthermore, exact conditions conducive to PAO growth are not conclusively known (WERF, 2003). The environmental and financial benefits associated with use of EBPR techniques over chemical precipitation merit their continued research and improvement.

Objectives

The primary objective of this experiment was to create a functional EBPR batch-sequencing reactor. This was accomplished in two stages. First an activated sludge batch sequencing reactor was constructed with stages for aeration, BOD removal, and denitrification. The reactor was run continuously for three weeks prior to the beginning of stage two. In stage two, the reactor’s phosphorus removal was enhanced via the addition of an hour-long anaerobic stage immediately following the input of the waste.

It was expected that the plant, before the addition of the initial anaerobic stage, would remove a small, but measurable, amount of phosphorus. Furthermore it was expected that a much more significant amount of phosphorus would be removed under the EBPR operating mode. It was difficult to foresee how scaling down both the size of the plant and the time frame of microbial activity would affect the phosphorus removal. However, based on published data, it was expected that the amount of phosphorus removed by the standard process would be around 15% and that of the EBPR would be less than 70%.

Methods

To test the efficiency of Enhanced Biological Phosphorus Removal (EBPR), a model was constructed to simulate the flows and processes of a real world wastewater treatment plant. The model consisted of a 5 L plastic tank with various inflow and outflow pipes. Inflows included concentrated waste solution and distilled water, which were pumped into the plant by a peristaltic pump at a rate of 178 mL/min. Outflow consisted of clarified effluent, which was drained out by gravity. The plant was mixed with a magnetic stirrer and aerated by pressurized airflow dispersed through an aeration stone. The airflow rate was determined by the procedures outlined in Weber-Shirk (2004). Dissolved oxygen was continuously measured with a dissolved oxygen probe suspended in the plant. The plant was operated as a batch reactor, treating and draining a single volume of wastewater at a time. Treatment consisted of several discrete operational stages, listed in Table 1. The parameter values used in defining these stages are found in Table 2. Process Controller software was used to define and manage these stages (Weber-Shirk, 2004). Once programmed and tested, the software was in complete control of the plant operation. Daily checks were done to ensure that the plant was functioning correctly.

The treatment plant was filled with approximately 4L of wastewater sludge to introduce BOD into the plant and foster further microbial growth. The sludge was obtained from the Ithaca wastewater treatment facility. Approximately every 6 hours, a new batch of concentrated waste would be pumped into the tank for treatment. The make-up of the concentrated waste is found in Table 3. Standard plant operation had a 6-hour treatment cycle, which consisted of the following approximate stage times: 1 minute of concentrated waste inflow, 10 minutes of water inflow, 3.3 hours of aeration/BOD removal, 1.5 hours of denitrification, 50 minutes of settling, and 10 minutes of effluent draining.

EBPR was incorporated by adding a stage after the tank was filled, but before aeration and BOD removal, in which POA microbes were selected. In this stage, the plant was simply mixed for one hour. Denitrification was removed from the program sequence to ensure that the microbes did not release the uptaken phosphorus back into the wastewater before effluent discharge.

Effluent samples were taken either from an effluent drain spigot or from the top of the plant during drainage. Influent samples were taken from the top of the plant directly after the plant was completely full and mixed. Effluent samples were taken during both standard and EBPR operation, while influent samples were only taken during EBPR operation. Samples were centrifuged to remove suspended solids and tested for phosphorus concentration with a wet calorimetric technique, as described in Weber-Shirk (2004). In using this method, samples were treated with a phosphorus dependent color reagent and tested with a spectrophotometer against known standards.

TABLE 1: Stages of the Standard Operation Treatment Plant

Operation Stage / Start Condition / End Condition / Active Controls / Comments

Off

/ - / - / - / -
Inflow of Concentrated Waste / Drain Stage Complete / Time in Stage > Concentrated Waste Inflow Time / Peristaltic Pump, Inflow Valve #1 / Pumped from Waste Stock
Inflow of Water / Concentrated Waste Inflow Complete / Time in Stage > Water Inflow Time / Peristaltic Pump, Inflow Valve #2 / Plant filled to 4 L
Initial Exertion / Water Inflow Complete / Dissolved Oxygen < Min. D.O. / Stirrer
Aeration / Initial Exertion Complete; BOD Exertion Complete / Dissolved Oxygen > Max D.O. / Stirrer, Airflow
BOD Exertion / Aeration Complete / Dissolved Oxygen < Min. D.O.; Time in Stage > BOD Exhaust Time / Stirrer / Plant will cycle b/w Aerate and BOD Exert.
Dentrification / BOD Exertion Complete / Time in Stage > Denitrification Time / Stirrer
Settle / Denitrification Complete / Time in Stage > Settle Time
Drain / Settle Stage Complete / Time in Stage > Drain Time / Effluent Valve / Cycle Complete

TABLE 2: Standard values used in stage decisions and overall plant operation

Parameter /

Value

Water Inflow Rate / 178 mL/min
Airflow Rate / 6 x 10-4M/s
Concentrated Waste Inflow Time / 47 s
Water Inflow Time / 620 s
Maximum Dissolved Oxygen / 5 mg/L
Minimum Dissolved Oxygen / 2 mg/L
Denitrification Time / 5400 s
BOD Exhaust Time / 650 s
Settle Time / 3000 s
Drain Time / 600 s

TABLE 3: Make-up of concentrated waste solution

Compound / Stock Concentration (g/L)
Starch / 8.440
Casein / 12.500
Sodium Acetate / 3.190
Capric Acid / 1.160
Ammonium Chloride / 7.533
Potassium Phosphate / 0.690
Sodium Hydroxide / 17.500
Glycerol / 1.200

Results and Discussion

As anticipated, Enhanced Biological Phosphorus Removal (EBPR) achieved significant phosphorus reduction. On average, the EBPR system removed 70% of the free phosphorus (PO4-3) from the wastewater influent, as seen in Figure 1. Current EBPR plants in the United States have published phosphorus removal efficiencies between 70 and 80 %, or more, depending microbial makeup of the plant (Lind, 1998). Therefore, it is surprising that the system in this experiment achieved the phosphorus removal of real world treatment systems after only one week of microbe selection and growth. It is likely that such rapid microbe selection is due to the small scale of the experimental treatment plant. It would not take a significant amount of time for phosphate accumulating organisms (PAOs) to grow and take over.

Figure 1: Influent and effluent samples for the EBPR reactor.

It is uncertain whether microbial selection would have continued with extended operation of the plant. The estimated residence time of microbes in the treatment plant is 10 days, and thus it seems likely that microbial selection was not exhausted at the termination of the experiment. Furthermore, phosphorus removals as high as 95% can be achieved using EBPR (WERF, 2003), supporting the assumption that the greatest efficiency of the plant had not been reached. It is left for a future study, but operating a plant such as the one in this experiment for several weeks would likely give conclusive evidence on the rate and threshold of microbial selection and thus plant efficiency.

To ensure that our EBPR plant design was the major reason for phosphorus reduction, EBPR effluent phosphorus concentrations were compared to those of a standard activated sludge sequencing batch reactor. As expected, the standard activated sludge system yielded effluents much higher in phosphorus, as there was no selection phase for the PAOs. The average PO4-3 effluent concentration from the standard plant was 934 g/L as compared to an average value of 498 g/L from the EBPR plant. The EBPR plant drastically decreases effluent phosphorus concentrations, proving our design effectively selects for the PAOs. See Table 4 and Figure 2 for exact effluent phosphorus concentrations for the EBPR plant and the standard activated sludge plant. The removal efficiency of the standard plant is unknown as no influent concentrations were measured. Active wastewater treatment plants without EBPR have PO4-3 removal efficiencies of 10 to 20 %, and thus the efficiency of our standard plant should not rival our 70 % ENPR removal efficiency (EPA, 2004). The addition of the hour long stirred anaerobic stage to our batch sequence succeeds at selecting the PAOs that uptake massive amounts of phosphorus in the presence of oxygen.

TABLE 4 – Influent and Effluent Phosphorus Concentrations for Standard and EBPR Systems

Sample / Influent Conc.
(g/L) / Effluent Conc.
(g/L) / Percent Removal
Standard #1 / - / 970 / -
Standard #2 / - / 726 / -
Standard #3 / - / 810 / -
Standard #4 / - / 1,230 / -
Standard Avg. / - / 934 / -
Standard St. Dev. / - / 222 / -
EBPR #1 / 1,440 / 427 / 70.3
EBPR #2 / 1,540 / 432 / 71.9
EBPR #3 / 1,940 / 634 / 67.3
EBPR Avg. / 1,640 / 498 / 69.9
EBPR St. Dev. / 264 / 118 / 2.35

Figure 2: Effluent phosphorus concentrations in Standard and EBPR reactors.

The main limiting factor in biological phosphorus removal is competition of PAOs with Gram-negative cocci known as “G” bacteria. The G bacteria are present in the MLVSS of wastewater treatment plants, and compete with PAOs for organic matter and volatile fatty acids in the anaerobic stage (Cloete & Theron, 2003). Therefore the presence of G bacteria severely inhibits phosphorus removal. There are several methods of promoting selection of PAO’s over G bacteria in the treatment plant. Establishing a solid retention time of 3 to 5 days, lowering the feed temperature below 10C, operating at a pH greater than 7.2, and having a high ratio of chemical oxygen demand to phosphorus in the influent have all serve to enhance growth of the phosphophilic bacteria (WERF, 2003).

During the course of the experiment, many difficulties were encountered. The dissolved oxygen probe performed very erratically and was a recurring problem during the operation of the plant. The problem was likely due to the growth and attachment of microbes onto the probe. To alleviate the problem, the membrane and o-ring were replaced on a weekly basis. The transition of operational states was dependent on dissolved oxygen readings, and thus a malfunctioning probe proved a significant problem that needed to be addressed.

Substantial cell growth on the walls of the plant also presented problems to be addressed. Significant attachment interferes with proper cell residence time, because cells would remain attached to the walls instead of being flushed out of the system. The plant was operated at high aeration rates, which also presented difficulties. When aerating, substantial foaming occurred at the surface of the plant. To prevent wastewater from overtopping the tank, a plastic cover was placed over the plant, which effectively eliminated the effervescence problem.

Another difficulty was obtaining proper readings from the spectrophotometer. Initial readings were highly erroneous. The root of the error was not known, but several different solutions were undertaken to address it. Recalibrating the standards and re-testing the samples produced reasonable results. Thus, there is an air of mystery about the spectrophotometer and further erratic results would not be unexpected. Taking great care in mixing standards and reagents, and having experience in using the spectrophotometer, might make consistent readings more likely.

There are also several weaknesses and errors that may be associated with this experiment. The most glaring of these is that the collection of standard influent samples was neglected. At the outset of this research it had been assumed that the influent phosphorus concentrations would be constant. It was later discovered that this is not entirely the case and that there is a significant amount of variance in influent concentrations (see Table 4). Another weakness of this research is that the number of samples that were taken was too small. However this was practically unavoidable given the short time frame and the difficulties that were encountered during the course of the experiment.

The creation and measurement of the phosphorus standards and samples were the most probable sources of error for this experiment. Based on the ingredient information provided with the waste and assuming casein has a 0.8% phosphorus concentration (Berggren 1931), the phosphorus concentration in the influent should have always been roughly 3.33 mg/L. However, our experimental data only shows it ranging up to at most 1.94mg/L, which is significantly lower than what the influent phosphorus concentration was supposed to be. There are several points during the process at which errors of this sort may have been introduced. The first of these is in the creation of the waste source. It is entirely possible that the care taken in preparing the waste was lax enough to result in unpredictable results. There is also no guarantee that the 20x concentrated waste coming out of the refrigerator was uniformly mixed, which could also make results erratic. Human error could have played a significant role in the mixing of the color reagent and the phosphorus standards. Unsatisfactory measuring or stirring during creation of the solutions could present magnified problems in absorbance measurements. In the end, the source of error causing measured influent concentrations to be substantially less than the expected calculated value remains unknown.

Conclusion

The primary focus of this experiment was to evaluate whether biological phosphorus removal was a viable treatment option in the provided laboratory environment. Based on this experiment, it is fair to say that biological removal can be achieved. As a result, in the future it would be reasonable to extend this experiment to include adjusting environmental variables in an effort to optimize the efficiency of the phosphorus removal. The efficiency of the sequencing batch reactor with EBPR in this experiment was 70 %. Preliminary research into this area indicates that increasing the plant’s pH, decreasing its temperature, or decreasing its solids retention time may improve phosphorus removal. Considering the small scale of the plant, and the limited time for EBPR microbial selection, it is surprising such a high removal rate was achieved. This efficiency should increase with more time, yet chemical precipitation might have to supplement the process to comply with EPA water quality guidelines. Current research is exploring ways to maximize biological phosphorus removal abilities, which will undoubtedly make EBPR systems more common in the future. EBPR wastewater treatment plants would remove phosphorus in an environmentally conscious and cost effective manner.

References

Berggren, Ruth. “The Phosphorus Content of Casein, Preliminary Paper.” Laboratory of Physiological Chemistry, Yale University. Yale University; New Haven, 1931.