Dec. 4, 2002 -- Sludge bulking and foaming caused by filamentous organisms are well known phenomena of activated sludge systems. But if this filamentous growth has poor settling properties, the effluent may deteriorate to the level where it fails to meet acceptable standards.
Waste water treatment plants for advanced nitrogen and phosphorus removal are operated under low-load conditions with sludge ages of ten days or more. The nutrient removal activated sludge process utilizes the metabolic properties of several types of microorganisms, which can be controlled by substrate leveling and provision of different electron acceptors.
The subdivision of the bioreactor into anaerobic, anoxic and aerobic zones and a proper design of flow rates and tank volumes are necessary preconditions. However, even a perfect biochemical performance results in a poor effluent quality if effective sludge separation cannot be achieved.
Sludge sedimentation has turned out to be the crucial step in the treatment process. The floc properties (size, density and shape) are the governing factors for the sedimentation process and cannot be reliably predicted.
The basic aim of dosing aluminium in the treatment process is to remove phosphorus by precipitation, but, in addition, a floc structure improvement can be expected. Two mechanisms are known for aggregations of colloids and creation of stronger flocs:
• Positively charged ions, such as Al3+, lower the colloidal surface charge, enabling aggregation. This occurs at low pH values, when the ions are soluble and do not form neutral hydroxides.
• At higher pH values hydroxides are formed and act as sweeping flocs.
These processes are instantly effective, because of their physical/chemical nature. Recent research and operational experience indicate a drawback of filamentous growth when aluminium is added to the activated sludge system. When filamentous bacteria proliferate, they grow out of the floc, slowing the settling due to increased drag forces, and may interconnect with other flocs, thus creating a network where a single floc cannot move independently. Moreover, gas bubbles entangle in the web, leading to buoyancy and foaming sludge. The filamentous decrease due to aluminium dosing is not well understood, but it may turn out to be a strong instrument for improving the floc structure.
The most common filament is Microthrix parvicella with the following selection factors known for mixed culture systems:
• low food to microorganism ratio or long sludge age
• alternating availability of oxygen and nitrate as electron acceptors
• low temperature
• availability of long chain fatty acids (sole carbon source).
Although these selection factors are well known, they cannot be effectively altered to enable filament control. The first two are preconditions for nutrient removal systems, and raising the temperature consumes too much energy. Blocking the access of the carbon source is the most promising approach but, surprisingly, no literature on the removal efficiency of grease traps is available.
The objective of this study was to investigate the performance of AlCl3 and PAX 18 regarding their instant physical/chemical and delayed physiological effects on the floc structure.
Materials and methods
(1) Origin: sludge and wastewater was taken from a municipal nutrient removing waste water treatment plant (WWTP) in Blankenloch, Germany.
(2) Waste water parameters: suspended solids were measured according to standard methods. COD, N and P were measured with test kits.
Floc and settling properties
(1) Settling velocity: this was measured in a column with 0.25 m diameter and a height of 0.8 m. AlCl3 and PAX 18 was added to 1 L of activated sludge. After 15 minutes of gently stirring, the suspension was diluted with 20 L effluent from the WWTP and allowed to settle in the column. Samples were taken for solids content analyses 40 cm below the surface at the beginning, and after two, five, ten, 20, and 40 minutes. The calculated settling velocity distributions represent single floc settling.
(2) Zeta potential: this was measured with Pem Kem 501.
(3) Filamentous levels: filamentous growth was classified into four classes (Class 0 - no filamentous growth; Class 1 - filaments grow out of the floc; Class 2 - filaments connect flocs; Class 3 - filamentous web contains flocs) using a dark field microscope at 480 x magnification.
Lab scale SBR treatment system
The sequencing batch reactor consisted of a 10 L column serving as aerobic, anoxic and phase separation reactor. A second reactor of 0.6 L volume was placed before the main reactor to produce anaerobic conditions prior to the aerobic/anoxic treatment. The raw water mixed with settled sludge was fed to the anaerobic reactor then transferred to the main reactor in the subsequent filling. The retention time within the anaerobic reactor therefore equals the time lag between two transfers to the reactor and was expected to be sufficient for biological phosphorus removal. The entire system was cooled down to 10C - 12C and the F/M ratio was kept below 0.2 g COD/gMLSS.d in order to provide good growth conditions for filamentous bacteria.
Two lab-scale treatment plants were operated in parallel, one for experiments, the other as control. Both operated for 20 days in parallel before the experiments.
COD, ammonia, nitrate and phosphorus were measured of both influent and effluent. Sludge samples were taken from the reactors for determination of MLSS. Sludge volume was read directly at the reactor columns after the 45 minute settling phase. Thus the calculated sludge volume index does not match with the determination according to standard methods and must be considered as a plant specific relative value.
PAX-18 long term performance
PAX-18 was dosed at day 0 only. The initial Al/MLSS ratio was set to 30 mg PAX-Al per gMLSS. Due to the dosage the pH increased to 5.5 but recovered with the next filling. Initially the effluent quality deteriorated and the COD reached 160 mg/L and was accompanied by a high turbidity of 200 NTU. The system recovered within three days and no more significant variations of the effluent quality were recorded compared with the control reactor.
Filamentous levels were considered by microscopic observation. No filaments were detected within the experimental reactor ten days after dosage, whereas filamentous growth proceeded in the control reactor. These levels were maintained for a further 20 days without additional dosage. The depression of filamentous growth was reflected by the SVI measurements to some extent, although the decreasing tendency in the control reactor did not reflect the increase of filament growth. When the experiment was terminated, the descent of the sludge blanket was recorded in a 100 ml cylinder. The results indicate a faster and more intense thickening even though the Al application dated back 30 days.
AlCl3 long term performance
In the preliminary stages of the AlCl3 experiment the filamentous growth was very poor. Only a few short filaments grew out of the flocs, representing a filament level somewhere below One, in the classes described above. The Al dosage was therefore set to 6 mgAl/gMLSS, which led to a pH drop to 6. In contrast to PAX-18, AlCl3 had no significant effect on effluent quality. In fact, there was no significant variation in the filamentous levels in either the control or the experimental reactor during the 15 day experiment. Although the filament levels within the experimental reactor were found to be slightly below the reference reactor, the difference was marginal and was not reflected by the sludge volume index measurements.
Zeta potential and settling velocity
Zeta potential measurements were carried out with an activated sludge suspension at pH 4.5, pH 5.5 and pH 6.5. PH values were adjusted by adding HCl prior to the addition of AlCl3 and PAX-18. AlCl3 was least effective for colloid destabilization at pH 6.5. At lower pH values the destabilization with AlCl3 was faster and 2 mgAl/L were sufficient to reach zeta 0. PAX-18 turned out to be effective even at pH 6.5.
In review of the settling velocity distributions of activated sludge flocs treated with AlCl3 and PAX-18, generally both precipitants effected a settling velocity increase. This increase can be attributed to the floc destabilization and agglomeration. The application of PAX led to a sharp rise of the zeta potential, as compared to the addition of AlCl3 at pH 6.5, which was the level during settling. The lower PAX dosage (6.4 gAl/gMLSS) led to a complete destabilization and maximum agglomeration, therefore no further increase was achieved by adding a higher dose. The effects of AlCl3 on zeta potential and settling velocity correspond in almost the same manner.
The zeta potential measurements demonstrate that the destabilization of biocolloids by PAX-18 occurred in an almost neutral pH range, whereas a destabilization with AlCl3 requires either a higher dosage or an additional pH adjustment. The settling velocity increase corresponded with the degree of floc destabilization and agglomeration, although sweeping floc effects must have contributed. The presence of soluble charged Al species - the proportion is dependent on pH values - is probably the governing factor for biological inhibition within the activated sludge process.
If this turns out to be the case, the dosing strategy of aluminium for depression of filament growth should target on the concentration of free aluminium rather than the total dose.
The results of the long-term experiments show that a single dosage of PAX-18 led to long lasting filament depression. As for the AlCl3 application, the result is less clear. However, a dosing strategy, which brings together the presented results and assumptions could be realized with a low-pH-bypass-dosing. A small portion of the return sludge flow could be diverted to a dosing reactor, where the pH is reduced and either PAX-18 or AlCl3 is added.
The treated bypass sludge is released back to elsewhere in the aerobic/anoxic system. It is assumed that the treated portion would not suffer from filamentous growth subsequently as long as one sludge age. Furthermore, the proposed dosing strategy may reduce the chemicals demand as compared to the treatment of the entire return sludge, since the percentage of active (free) aluminium would be greater due to the lower pH value.
The proposed dosing strategy will be the subject of further research. The optimum flow rate, retention time and pH/Al-ratio shall be investigated in order to achieve a sustainable filament depression and to minimize the chemicals demand.
The authors of the original paper are by J. Kegebein, E. Hoffman, and H. H. Hahn. J. Kegebein is from the Institute for Aquatic Environmental Engineering at the University of Karlsruhe, Germany, and can be contacted at email@example.com.