Biotechnology Letters, Vol 20, No 11, November 1998, pp. 1085–1090
The fate of phosphate under anoxic conditions in biological nutrient removal activated sludge systems Nazik Artana, Rüya Taslia,*, Nevin Özgürb and Derin Orhona a
Environmental Engineering Department, Istanbul Technical University, I.T.U. Insaat Fakültesi, 80626 Maslak, Istanbul, Turkey *e-mail:
[email protected] b Trakya University, Çorlu Engineering Faculty, Çorlu, Tekirdag, Turkey The nitrogen removal potential of phosphate accumulating organisms under anoxic conditions has been evaluated using a laboratory scale sequencing batch reactor fed with synthetic wastewater and operated in a sequence of anaerobic, anoxic and aerobic periods. The phosphate uptake rate under anoxic conditions was lower than that under aerobic conditions. However, in the presence of an external substrate such as glucose and acetate, the fate of phosphate was dependent on the substrate type; phosphate release occurred in the presence of nitrate as long as acetate was present and glucose did not cause any phosphate release. The nitrate uptake rate was also much lower with glucose than acetate. The results implied that poly-hydroxyalkanoates could be oxidized by nitrate and phosphate uptake during the anoxic phase should be introduced into process modeling. Keywords: Anoxic conditions, biological phosphate removal, denitrifying phosphate removing bacteria, sequencing batch reactor.
Introduction In recent activated sludge models (Gujer et al., 1995) involving enhanced biological phosphate removal (EBPR), it is assumed that the phosphate accumulating organisms, PAOs, do not possess any denitrifying capability and may release phosphate, pending upon the existence of suitable external substrate. This assumption implies that phosphate uptake can only be observed under aerobic conditions. However, experimental evidence indicates the existence of denitrifying phosphate accumulating organisms which are capable of phosphate uptake under anoxic conditions. Gerber et al. (1987) compared the phosphate uptake rates under aerobic and anoxic conditions and stated that the observed rate under anoxic conditions was significantly lower than that under aerobic conditions. Kerrn-Jespersen and Henze (1993) attributed the lower uptake rate to the fact that a fraction of the PAOs also take up phosphate under aerobic conditions. They also showed a linear relationship between the amount of acetate taken up in the anaerobic phase, the denitrification rate and the phosphate uptake rate. Kuba et al. (1993, 1994, 1996) observed a good phosphate removal in a sequencing batch reactor, SBR, operated in a sequence of anaerobic-anoxic periods in which denitrifying PAOs were enriched without oxygen and they stated that phosphate uptake activities under anoxic and aerobic con© 1998 Chapman & Hall
ditions could be identical. Sorm et al. (1996) also reported that the anoxic phosphate uptake with simultaneous denitrification after following anaerobic substrate uptake could significantly reduce the extent of competition for organic substrate between phosphate accumulating organisms and denitrifiers. In biological nutrient removal, the essential problem is a shortage of readily biodegradable organic matter in wastewaters. Thereby, the degree of denitrification capability of phosphate accumulating organisms is a significant issue that influences process design. If PAOs do not possess any denitrifying capability, polyhydroxyalkanoates (PHA) stored can only be used in aerobic conditions. Then enhanced biological phosphate removal will decrease the denitrification potential of the system. On the other hand, if phosphate accumulating organisms are capable of denitrification, the enhanced biological phosphate removal will not affect nitrogen removal efficiency provided that a sequence of anaerobic and anoxic phases are maintained in the process. The main objective of this study is to explore whether PAOs have denitrifying capability and to determine to what extent they can denitrify. For this purpose, a laboratory scale SBR was operated with pre- and post- anoxic periods using different synthetic substrates, in order to investigate the effect of nitrate on both the release and uptake mechanism of phosphate. Additional support was Biotechnology Letters ⋅ Vol 20 ⋅ No 11 ⋅ 1998
1085
N. Artan et al. obtained from batch tests, the first part of which was carried out with the mixed liquor taken from the end of the aerobic period of the SBR for assessing the effect of nitrate entering to the anaerobic zone on the extent of phosphate release. The results from the second series of batch tests conducted with the mixed liquor taken from the end of anaerobic period were evaluated in order to examine the behaviour of PAOs in anoxic periods following the anaerobic ones.
oxygen demand to total kjeldahl nitrogen (COD/TKN) ratios over two years. The SBR performance was continuously monitored by measuring phosphate, chemical oxygen demand, oxidized nitrogen in the effluent daily. Mixed liquor suspended solids (MLSS) and mixed liquor volatile suspended solids (MLVSS) concentrations were also measured routinely.
Materials and methods Laboratory-scale SBR operations The experimental work was carried out in a laboratory scale sequencing batch reactor having a working volume (VT) of 8.8 l and a fill volume (VF) of 5.3 l. The sequencing batch reactor was operated with a cycle time of 6 h. Each cycle included a sequence of anoxic, anaerobic and aerobic periods and involved a 15 min static fill (F) and 90 min pre-mixing (M) within the anoxic/anaerobic period, followed by 2.75 h aeration (A), 30 min settling and 1 h draw and idle periods during the first part of the study (RUN I). As shown in Table 1, the synthetic wastewaters consisting of tryptic soy broth (TSB), acetate and glucose were used during the operations (RUN I-1,2,3,4). The total phosphorus concentration (TP) of the feed was kept constant at 20 mg/l for the whole experimental study.
Batch experiments Additional batch tests using the mixed liquor taken from different periods of SBR operation were conducted during RUN II-4. The mixed liquor was taken from the SBR, divided into 4 equal portions so as to contain the same amount of MLVSS in each batch and put into 500 ml beakers covered with stretch film and mixed by magnetic stirrer. The biomass was protected from ingress of atmospheric oxygen during the anaerobic and anoxic batch tests by a nitrogen purge above the mixed liquor. Anoxic conditions were sustained by KNO3 addition. The external substrate content of the sludge taken from the SBR was practically negligible. Acetate was added to some of the batches while glucose was added to the others as an external substrate which is thought to be unutilizable by PAOs under anaerobic conditions. COD, oxidized nitrogen (NOx-N) and phosphorus (PO4-P) concentrations were monitored in all batch tests.
The first two runs lasted one month each and the last two two months, each. After 6 months operation of the system under these pre-anoxic, anaerobic and aerobic conditions, a post-denitrification was added to the system by introducing an unaerated period (mixing) of 45 min into the aerobic period after 2 h of aeration (RUN II). RUN II was operated with only one feed composition (Feed 4). The mixed liquor volatile suspended solids concentration was maintained at about 2000 mg/l during whole experimental study by wasting a required amount of the sludge on the basis of selected sludge age. The sludge was wasted at the end of the aerobic period of the same cycle, once a day. Prior to this study the SBR had been operated under anoxic, anaerobic and aerobic conditions using combinations of same different substrates having different chemical
Analytical methods All analysis were carried out in accordance with the Standard Methods (APHA, 1989). The samples were immediately filtered through a Whatman GF/C glass fibre filter. Orthophosphate was measured according to the ascorbic acid method and persulphate digestion method was used for total phosphorus measurements. Nitrate was reduced to nitrite in a cadmium reduction column and measured as nitrite by means of a spectrophotometer as described in the colorimetric method. Ammonia nitrogen (NH3-N) concentrations were measured with phenate method. MLSS and MLVSS concentrations were determined using glass fibre filters (Whatman GF/C). Closed reflux titrimetric method was used for COD measurements.
Table 1 Average feed compositions.
Results and discussion Performance of SBR Each operation was monitored long enough to investigate the stability of the system. The results shown in Table 2 reflect the long term operation averages of the system. During the RUN I operations (RUN-I,1,2,3,4) in which only pre-denitrification was sustained, nitrogen removal was limited by the volume exchange ratio, thus, the effluent NOx-N has to be 60% of the nitrification capacity which may be calculated on the basis of the influent TKN,
Feed No Parameters Acetate (mg COD/l) TSB (mg COD/l) Glucose (mg COD/l) TKN (mg/l) TP (mg/l) COD/TKN
1086
1
2
3
4
– 150 150 22 20 13.6
– 150 150 30 20 10
60 240 – 26 20 11.5
60 240 – 32 20 9.4
Biotechnology Letters ⋅ Vol 20 ⋅ No 11 ⋅ 1998
The fate of phosphate under anoxic conditions Table 2
Operation RUN RUN RUN RUN RUN
I-1 I-2 I-3 I-4 II-4
Average results of the long term SBR operations. Sludge age, θX (days)
Effluent NOx-N (mg/l)
Phosphorus Removal (mg/l)
% phosphorus of VSS
MLVSS/MLSS
13 19 8.8 8.8 8.8
4.8 11.3 7.1 10.3 8.0
6.9 4.1 16.1 14.4 15.3
6.4 3.4 12.0 11.5 12.0
0.80 0.85 0.68 0.70 0.69
COD and the sludge age of the system (Orhon and Artan, 1994). Measured effluent NOx-N values were always lower than those predicted on the basis of nitrogen mass balance in accordance with the conventional model of nitrification and denitrification. The difference between measured and predicted effluent NOx-N values was about 2 mg N/l and it increased as influent TKN was increased. The results showed that high TKN/COD ratios, resulting through introduction of higher nitrate into the unaerated period, caused lower EBPR performance when the same type of substrate was used. It was noted that the phosphate removal increased with an influent without glucose regardless of the fact that the NOx-N input increased. The effect of glucose-rich influent on system performance was discussed earlier (Tasli et al., 1997). During the RUN II-4 operation, effluent NOx-N was lower than that of the RUN I-4 operation with the same feed; this observation could be attributed to the additional nitrogen removal through post-denitrification, resulted in better phosphate removal performance. The results revealed that 1 mg/l reduction in NOx-N input to the unaerated period caused an average of 0.4 mg/l increase in phosphate removal. Cycle measurements Changes of significant parameters were also monitored in a cycle time to observe the metabolic behaviour of the activated sludge under different periods. Figure 1 shows
Figure 1 PO4-P, NOx-N, NH3-N and COD profiles versus time during RUN II-4. F: Fill period, M: Mixing period, A: Aeration period.
the COD, NOx-N, NH3-N, and (PO4-P) profiles throughout a complete cycle for RUN II-4. As can be seen from the figure, anaerobic conditions prevailed for nearly an hour after the pre-anoxic period. A considerable decrease in COD was observed in the first anoxic period and it was mostly depleted at the end of the anaerobic period. Nitrification was almost completed in the first aerobic period. Measured phosphate value at the end of the fill period was nearly equal to the concentration calculated from appropriate dilutions induced by the introduction of the influent waste stream into the mixed liquor. This implies that net phosphate release did not take place in the pre-anoxic period during the fill period. In the postanoxic period, however, phosphate uptake was observed with a slight decrease in nitrate concentration. No significant change in the slope of the phosphate profile versus time was recorded when aeration was restarted after the post-anoxic phase. Batch test results The results of the anaerobic and anoxic batch tests conducted with the sludge taken from the end of the second aerobic period of the system in order to observe the effect of nitrate on phosphate release under different substrate conditions are illustrated in Figure 2. It is shown that phosphate release occurred in the presence of nitrate as long as acetate was present, however, the phosphate release rate decreased from 37 mg PO4-P/ gVSS.h to 15 mg PO4-P/gVSS.h and the COD removal rate slightly increased in the presence of nitrate. Hence, in the anaerobic batch test, the released phosphate /utilized acetate ratio, YPO4, for PHA storage, was found to be 0.69 g PO4-P/g acetate COD whereas in the anoxic batch test, the released phosphate/utilized acetate ratio was 0.22 g PO4-P/g acetate COD; this is possibly due to the partial utilization of acetate for denitrification by heterotrophic non-polyP microorganisms. Another possibility is that phosphate uptake due to stored PHA oxidation occurs simultaneously with phosphate release for PHA storage when acetate is available under anoxic conditions, resulting in a lower net phosphate release. As a matter of fact, during the test without external substrate, slow phosphate Biotechnology Letters ⋅ Vol 20 ⋅ No 11 ⋅ 1998
1087
N. Artan et al. nitrifying fraction of the heterotrophs should be considered as the most important mechanism for the deterioration of phosphate removal, possibly due to the input of nitrate to the “anaerobic” zone. Addition of glucose did not cause any phosphate release; on the contrary, phosphate uptake took place when both glucose and nitrate were present. The presence of glucose did not have any significant effect on phosphate uptake rate (PUR) but nitrate utilization rate (NUR) increased from 1.5 mg NOx-N/g VSS.h to 3.7 mg NOx-N/g VSS.h, a value which is still much lower than the level observed with acetate. The utilized NOx-N/ removed COD ratio for the anoxic growth on glucose was also lower than that on acetate, due to the fact that the yield coefficient of glucose is higher than that of acetate. Figure 3 illustrates the results of the aerobic and anoxic batch tests conducted with the sludge taken from the end of anaerobic period. The main purpose of these tests was to observe the fate of phosphorus under anoxic conditions following anaerobic conditions.
Figure 2 The results of the anaerobic and anoxic (with or without substrate) batch tests conducted with the sludge taken from the end of aerobic period. Symbols: s, anaerobic-acetate ; d, anoxic-acetate ; m, anoxicglucose ; ✶, anoxic-no external substrate.
uptake was measured together with a slight decrease in nitrate concentration since the PHA content of the activated sludge was very low. However, the decrease in phosphate release under anoxic conditions was much higher than the difference between phosphate release and uptake. This result can be evaluated to indicate that a simple competition for acetate between the PAOs and the de-
1088
Biotechnology Letters ⋅ Vol 20 ⋅ No 11 ⋅ 1998
The results showed that anoxic phosphate uptake took place together with nitrate reduction when there was no external substrate. In the presence of external substrate, the fate of phosphate was dependent on the substrate type. With acetate as substrate, phosphate release occurred simultaneously with a high nitrate utilization rate (8 mg NOx-N/gVSS.h), whereas phosphate uptake with glucose as external substrate, was observed at relatively low nitrate utilization rates (2 mg NOx-N/gVSS.h). The phosphate uptake rate under anoxic conditions was lower than that under aerobic conditions. The ratio of PUR under anoxic conditions to that under aerobic conditions was observed to be 0.46 in this experiment, whereas in other experiments within the same study, values of up to 0.82 were measured. On the basis of these observations, a correction factor, hNO3, of around 0.5 can be suggested. This ratio can be implemented as a correction factor for anoxic growth of PAOs, adjusting for either the change in maximum specific growth rate associated with anoxic conditions or for the fact that only a portion of the PAOs can denitrify like the “other” heterotrophs. The measured hNO3 values for PAOs show that these microorganisms can adapt themselves to anoxic conditions comparatively well as other heterotrophs. Anoxic batch tests, using mixed liquor both from the end of the anaerobic and aerobic periods of the same cycle of the SBR, were performed as additional support to previous observations to visualize the effect of sludge on the fate of phosphate and oxidized nitrogen. As almost all of the soluble COD was taken up after the anaerobic period, each batch had practically no external substrate and residual COD was probably inert metabolic
The fate of phosphate under anoxic conditions
Figure 4 PO4-P and NOx-N profiles versus time during the batch test conducted using mixed liquor both from the end of anaerobic and aerobic periods of the same SBR cycle. Symbols: d, sludge from aerobic period; m,sludge from anaerobic period.
serve as an electron acceptor for the oxidation of stored PHA by polyP microorganisms and concomitant phosphorus uptake takes place with denitrification.
Figure 3 The results of the anoxic (with or without substrate) and aerobic batch experiments conducted with the sludge taken from the end of anaerobic period. Symbols: s, aerobic-no external substrate ; d, anoxicacetate ; m, anoxic-glucose; ✶, anoxic-no external substrate.
products (Orhon et al., 1989). The intracellular PHA concentration of the sludge from the anaerobic period must be higher than that of the sludge from the aerobic period. Higher PHA concentration resulted in higher phosphate uptake rate as well as higher nitrate utilization rate, as can be seen in Figure 4. This result also shows that nitrate can
Conclusions The experimental results on a continuously operated SBR system and batch tests show that phosphate uptake by PAOs also take places in the anoxic period. On the other hand, phosphate release can be observed in the presence of nitrate as long as acetate is available. The observed release rate, YPO4, however, decreases both due to the utilization of acetate for denitrification by the non-poly P heterotrophs and the simultaneous phosphate uptake by denitrifying PAOs. From the batch test results, YPO4, was determined as 0.69 g P/g acetate COD, a value higher than the default level of activated sludge model no:2 (Gujer et al., 1995) where it is assumed that PAOs can not denitrify. Comparing aerobic and anoxic phosphorus uptake rates, it is evident that a portion of the PAOs can not denitrify. The observed values of correction factor for denitrification by PAOs, hNO3, obtained from the batch tests ranged from Biotechnology Letters ⋅ Vol 20 ⋅ No 11 ⋅ 1998
1089
N. Artan et al. 0.46 to 0.82. Although the scientific basis of defining such a factor reflecting the fraction of PAOs with a denitrification capability is to be further evaluated, the experimental results of this study provide a clear indication that PAOs can adapt themselves to anoxic conditions comparatively well as “other” heterotrophs. Therefore, anoxic growth of PAOs and the anoxic phosphate uptake should be introduced into activated sludge modeling with a correction factor, hNO3 5 0.5, for denitrification. References APHA (1989). Standard Methods for the Examination of Water and Wastewater. 17th edition, Washington,D.C. Gerber, A., Villiers De, R.H., Mostert, E.S., Van Riet, C.J.J., (1987). Biological phosphate removal wastewaters. In : Advances in Water Pollution Control 4, R. Ramadori, ed, pp 123–134, Oxford, England: Pergamon Press.
Gujer, W., Henze, M., Mino, T., Matsuo, T., Wentzel, M.C., Marais, G.v.R. (1995). Wat Sci Tech, 31(2):1–11. Kerrn - Jespersen, J.P., Henze, M. (1993). Wat Res, 27(4): 617–624. Kuba, T., Smolders, G., Van Loosdrecht, M.C.M., Heijen, J.J. (1993). Wat Sci Tech, 27(5/6): 241–252. Kuba, T., Wachtmeister, A., Van Loosdrecht, M.C.M., Heijen, J.J. (1994). Wat Sci Tech, 30(6): 213–219. Kuba, T., Van Loosdrecht, M.C.M., Heijen, J.J.(1996). Wat Sci Tech, 34(1/2): 33–40. Orhon D., Artan, N., (1994). Modelling of Activated Sludge Systems, Lancaster, U.S.A. : Technomic Publishing. Orhon, D., Artan, N., Cim¸sit, Y. (1989). Wat Sci Tech, 26(5/6): 933–942. Sorm, R., Bortone, G., Saltarelli, R., Jenicek, P., Wanner, J., Tilche, A. (1996) Wat Res, 30(7): 1573–1584. Tasli, R., Artan, N., Orhon, D . (1997). Wat Sci Tech, 35(1): 75–80.
Received: 29 July Revisions requested: 19 August/18 September Revisions received: 16 September/13 October Accepted: 13 October
1090
Biotechnology Letters ⋅ Vol 20 ⋅ No 11 ⋅ 1998
1998 1998 1998 1998
