- Email: [email protected]
Maintenance of phosphorus removal in an EBPR system under permanent aerobic conditions using propionate
Biochemical Engineering Journal 43 (2009) 288–296
Contents lists available at ScienceDirect
Biochemical Engineering Journal journal homepage: www.elsevier.com/locate/bej
Maintenance of phosphorus removal in an EBPR system under permanent aerobic conditions using propionate M. Vargas, C. Casas, J.A. Baeza ∗ Departament d’Enginyeria Química, ETSE, Universitat Autònoma de Barcelona, 08193 Bellaterra, Barcelona, Spain
a r t i c l e
i n f o
Article history: Received 10 June 2008 Received in revised form 22 September 2008 Accepted 18 October 2008 Keywords: Aerobic processes Enhanced biological phosphorus removal (EBPR) Phosphorus accumulating organisms (PAO) Propionic acid Wastewater treatment
a b s t r a c t Enhanced biological phosphorus removal (EBPR) is an efficient and sustainable technology to remove phosphorus from wastewater preventing eutrophication in natural waters. It is widely accepted that EBPR requires an optimal anaerobic hydraulic retention time to obtain stable P-removal from wastewater. Thus, it is suggested that deterioration of the EBPR efficiency regularly observed in full-scale wastewater treatment plants (WWTPs) is normally caused by an excessive aeration of activated sludge that increments the amount of oxygen recycled to the anaerobic reactor and consequently, the anaerobic conditions are not totally preserved. Furthermore, it has been reported a progressive decrease in P-removal capacity in an EBPR lab-scale system enriched with acetate as the sole carbon source under permanent aerobic conditions. Hence, to evaluate the stability of P-removal with a different carbon source, an EBPR-SBR was operated with propionate under permanent aerobic conditions. As a result, net P-removal was successfully accomplished in the SBR without any anaerobic phase during 46 days of aerobic operation. Moreover, the system was shifted after this period to the standard anaerobic–aerobic conditions and reliable Premoval was maintained. FISH (fluorescence in situ hybridisation) analysis showed a significant presence of Accumulibacter (70, 50 and 72%, in different periods) and the absence of Competibacter. The results indicate that using propionate as carbon source it is possible to maintain in a long term an enriched culture of phosphorus accumulating organisms (PAO) able to remove phosphorus under permanent aerobic conditions. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Enhanced biological phosphorus removal (EBPR) is an efficient and sustainable biological technology to remove phosphorus from wastewater. It is based on the enrichment of activated sludge with phosphorus accumulating organisms (PAO) by introducing alternating anaerobic–aerobic conditions, which favour PAO growth over ordinary heterotrophic organisms (OHO). Under anaerobic conditions, PAO store volatile fatty acids (VFA) as polyhydroxyalkanoates (PHA) with concomitant P-release into solution derived from the hydrolysis of polyphosphate (poly-P) reserves. The required reducing equivalents are provided by the catabolism of stored glycogen [1]. In a subsequent aerobic (or anoxic) stage, PHA is used as carbon and energy sources for growth and replenishment of the internal glycogen and poly-P pools. An amount of phosphate higher than that released is taken up by PAO and the poly-P pools are restored. As a result, net P-removal is accomplished by periodic sludge waste after this phase.
∗ Corresponding author. Tel.: +34 93 581 1587; fax: +34 93 581 2013. E-mail address: [email protected] (J.A. Baeza). 1369-703X/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.bej.2008.10.013
On the other hand, even though the two most common VFA present in municipal wastewater are acetate and propionate, most of works on EBPR consider acetate as sole carbon source. However, the metabolism of propionate by PAO has attracted recently considerable attention [2–4] and it is suggested that it could be a more favourable substrate for EBPR [5–9]. Although many works are focused on this complex process, the complete metabolic behaviour of the EBPR sludge is still unclear and it is prone to unpredictable failures, primarily because the process design is highly empirical due to an incomplete understanding of sludge microbial ecology [10]. For example, Brdjanovic et al. [11] hypothesised that one cause of EBPR efficiency deterioration could be excessive aeration, which would result in a gradual depletion of PHB and/or saturation of the biomass by poly-P. The two most common causes of excessive aeration in continuous wastewater treatment plants (WWTPs) are heavy rainfall periods and failures in the aeration control, which increment the amount of oxygen recycled to the anaerobic reactor. In both scenarios, the anaerobic conditions are reduced and the aerobic phase is extended leading to a detriment in P-removal. To the best of our knowledge, successful EBPR operation without a physical separation between the electron donor (organic matter)
M. Vargas et al. / Biochemical Engineering Journal 43 (2009) 288–296
289
Fig. 1. Time course of TSS, VSS, VSS/TSS ratio and changes in the feed composition during both SBR periods studied. (䊉) VSS, () TSS, (♦) VSS/TSS, () phosphate, () propionic acid.
and the electron acceptor (oxygen) has not been reported yet in full scale WWTP. In conventional continuous WWTP, the inclusion of a first anaerobic reactor allows the separation of organic matter and oxygen. These conditions favour PAO against OHO because PAO are more efficient to anaerobically uptake VFA due to their polyphosphate reserves. Nevertheless, some recent studies showed the feasibility to achieve net P-removal under strictly aerobic conditions in well-controlled lab-scale EBPR reactors [12–14]. In these studies, aerobic acetate uptake was linked to phosphate release, PHA storage and glycogen degradation. After substrate depletion, phosphate was taken up linked to PHA degradation, glycogen synthesis and PAO growth. The phases with and without available external substrate are known in the literature [15–16] as feast and famine, respectively. According to the pattern observed, these stages were analogous to the anaerobic–aerobic conditions of the typical EBPR process [13]. Moreover, Pijuan et al. [17] studied the P-removal deterioration in enriched PAO sludge subjected to permanent aerobic conditions. An aerobic SBR was operated using acetate as carbon source, observing the abovementioned feast/famine phases. However, net P-removal was only maintained during the first 4 days of operation, showing that the tested conditions did not provide a stable operation for aerobic EBPR. Finally, Ahn et al. [18] proposed recently a new operational scheme for aerobic phosphate removal from wastewater consisting of an aerobic SBR where the feed stage (acetate addition) was temporally separated from phosphate addition, which begins the famine stage. The authors showed that this process was capable to remove phosphate from 10 to 12 mg P/L to less than 0.1 mg P/L over an extended period, using wastewater with low COD content. This paper aims to propose an alternative for stable aerobic P-removal. It tests if EBPR can be maintained in a SBR under permanent aerobic conditions using propionate as the sole carbon source. Additionally, after this strict aerobic operation, the stability of net P-removal is examined when the SBR is changed to conventional EBPR anaerobic–aerobic operation.
2. Materials and methods 2.1. Equipments and experimental design The 10 L SBR used for the experiments was seeded with sludge obtained from an EBPR anaerobic–aerobic SBR [19]. It was operated at four cycles per day with a hydraulic residence time (HRT) of 12 h.
Each cycle consisted of 120 min of anaerobic phase (5 initial min for feeding), 180 min of aerobic phase, 55 min of settling and 5 min of extraction of 5 L of supernatant. The sludge residence time (SRT) was kept at 10 days by periodic sludge wastage during the end of the aerobic period but before stirring was stopped. After 5 months of stable and efficient EBPR operation, the SBR was changed to permanent aerobic operation (from day 1 to 46) to investigate the short-term and long-term aerobic P-removal capability of PAO. Each cycle consisted of 300 min of aerobic phase (5 initial min for feeding), 55 min of settling and, in the last 5 min, withdrawing of 5 L of supernatant. SRT was kept also at 10 days. After the 46 days of aerobic operation, the SBR was operated with standard anaerobic–aerobic cycles for 59 days (from day 47 to 105) to evaluate the behaviour of the system regarding P-removal stability. Temperature in the SBR was kept at 25 ◦ C and the pH was maintained below 7.5 with HCl 1 M. Dissolved oxygen (DO) was controlled during the aerobic phase between 3.5 and 4.5 mg DO/L with an on/off controller to avoid oxygen limitations. Oxygen Uptake Rate (OUR) was estimated on-line by measuring the DO decrease rate between 4.5 and 3.5 mg DO/L during the off position of the aeration valve. The average biomass concentration maintained in the reactor during the whole period of study was around 4 g VSS/L, although some variations of TSS and VSS were observed (Fig. 1).
2.2. Synthetic media Two separate solutions called “concentrated feed” (constituting 0.25 L per 5 L synthetic wastewater) and “P-water” (constituting 4.75 L per 5 L synthetic wastewater) collectively formed the synthetic wastewater used in this study. The “concentrated feed” consisted of (g/L RO water): 12 (or 7.2) propionic acid (to achieve 300 mg/L or 180 mg/L in the SBR after feeding), 2.0 NH4 Cl, 0.88 MgSO4 ·7H2 O, 3.2 MgCl2 ·6H2 O, 0.84 CaCl2 ·2H2 O, 0.4 yeast extract, 0.10 allylthiourea (ATU) to inhibit nitrification and 6 mL of nutrient solution. The P-water consisted of 1:0.6 KH2 PO4 /K2 HPO4 (molar ratio). The concentration of P and propionic acid changed along the study (see Fig. 1). Prior to the permanent aerobic cycles performed in the SBR, it was operated with initial concentrations of 15 mg/L of P and 180 mg/L of propionic acid. Afterwards, in the third day of permanent aerobic operation, they were incremented up to 80 and 300 mg/L, respectively, for a better monitoring of the variables measured in the study cycles. Later on (day 19), phosphate concen-
290
M. Vargas et al. / Biochemical Engineering Journal 43 (2009) 288–296
Fig. 2. Experimental profiles during a standard anaerobic–aerobic SBR operation (left) and the first day under aerobic conditions (right): (䊉) propionic acid, () phosphate, () PHA, () glycogen.
Fig. 3. CLSM FISH micrographs of the EBPR sludge hybridised with probes specific for “Candidatus Accumulibacter phosphatis” (in pink, Cy3-labelled PAOMIX probes) and probes for bacteria (in blue, Cy5-labelled EUBMIX probes) in different periods. (A) Standard SBR operation; (B) aerobic period, day 46; (C) anaerobic–aerobic period, day 105. Bar = 20 m
M. Vargas et al. / Biochemical Engineering Journal 43 (2009) 288–296
291
Chromatography (HPLC). A Hewlett Packard 1050 equipped with a Bio Rad Aminex HPX-87H ion exclusion column at 25 ◦ C was used. Glycogen was determined by a modification of the method of Smolders et al. [20]. A volume of 5 mL of 0.6 M HCl was added to each 20 mg of lyophilised sludge sample, and then heated at 100 ◦ C for 6 h. After cooling and filtering through 0.22 m filter (Millipore), the concentration of glucose was measured using an YSI Model 2700 Select Biochemistry Analyser (Yellow Springs Instrument, Yellow Springs, Ohio, USA). Triplicates of each sample were done. PHA was measured according to the methods of Oehmen et al. [21]. 40 mg of lyophilised sludge samples was digested and methylated with 4 mL of acidulated methanol (10% H2 SO4 ) and 4 mL of chloroform during 20 h at 100 ◦ C. Benzoic acid was used as internal standard. The calibration of the method was performed using a 3hydroxybutyric acid and 3-hydroxyvaleric acid copolymer (7:3) as standard for PHB and PHV (Fluka, Buchs SG, Switzerland) and 2hydroxycaproic acid as standard for PH2MV (Aldrich). The analyses were performed in a GC system, operated with a Hewlett Packard 5890 column (30 m length × 0.53 mm I.D. × 1.00 m film). Triplicates of each sample were done. Total suspended solids (TSS) and volatile suspended solids (VSS) were analysed according to Standard Methods [22]. 2.4. Microbial analyses
Fig. 4. Experimental profiles during permanent aerobic cycles monitored in the SBR: (䊉) propionic acid, () phosphate, () glycogen; (A) day 4; (B) day 8 and (C) day 13.
tration was decreased to 20 mg/L to avoid possible precipitation of this element in the bulk solution.
Representative samples were withdrawn from the SBR to perform fluorescence in situ hybridisation (FISH) technique. FISH was performed as in Amann [23] with Cy5-labelled EUBMIX probes for most bacteria [24], and either Cy3-labelled PAOMIX probes for “Candidatus Accumulibacter phosphatis”, comprising equal amounts of probes PAO462, PAO651 and PAO846 [25] or Cy3-labelled GAOMIX for “Candidatus Competibacter phosphatis”, comprising equal amounts of probes GAOQ431 and GAOQ989 [26]. Sludge was fixed by 4% paraformaldehyde solution in phosphate buffer 0.03 M. Hybridisation was performed with 35% formamide solution at 46 ◦ C for 2 h. FISH preparations were visualised with a Leica TCS SP2 AOBS confocal laser scanning microscope (CLSM). Accumulibacter or Competibacter were quantified as a proportion of all bacteria using image analysis with the methodology below summarised, performed with specific software developed with Matlab Image Processing Toolbox [27]. To avoid colony formers underestimation, each sample was homogenised and just one layer of sludge in each well of the slide was applied when FISH was performed. Forty randomly chosen CLSM fields from different x, y, and z coordinates were imaged (600× magnification) from each sample. Three images obtained with the same sludge and procedures, but without addition of probes, were used for evaluating the sample autofluorescence. The threshold in pixel intensity for removing autofluorescence was selected to eliminate the 99.99% of pixels with positive signal of these images. To account for the different amount of biomass in each image, the area of specific cell signal was calculated with the addition of all the detected pixels with intensity above the previously evaluated threshold for Cy3 in the 40 images divided by the total amount of pixels above its corresponding threshold detected for Cy5.
2.3. Chemical and biochemical analyses
3. Results and discussion
Analyses of phosphate in filtered samples were performed with an Electrophoresis Capillary System (Quanta 4000E CE – WATERS). The electrolyte used was the commercial solution Ionselect High Mobility Anion Electrolite. Propionic acid was measured by Liquid
3.1. Standard reactor performance The enriched PAO sludge obtained under standard EBPR operation (anaerobic–aerobic) showed a good performance of P-removal.
292
M. Vargas et al. / Biochemical Engineering Journal 43 (2009) 288–296
Fig. 5. On-line measurements during one aerobic SBR cycle on day 8.
Fig. 2 (left) depicts the experimental profiles for some significant measurements (phosphate, propionic acid, PHA and glycogen) obtained during a cycle of the SBR after 5 months of operation. The typical behaviour of EBPR sludge is clearly observed. In the anaerobic phase, propionic acid uptake was linked to phosphate release derived from the hydrolysis of poly-P stored, glycogen degradation and PHA storage. Under aerobic conditions, PHA was consumed and glycogen accumulated meanwhile phosphate was removed from the bulk solution, leading to a net P-removal in the system. FISH analysis of sludge samples withdrawn from this SBR cycle showed that Accumulibacter accounted for 70% of all bacteria (see Fig. 3A). This significant percentage of population of Accumulibacter is similar to others obtained in lab-scale SBR sludges enriched with propionate under anaerobic–aerobic conditions (e.g. 55% in Pijuan et al. [7]; 65% in Oehmen et al. [28] and 90% in Lu et al. [9]). In contrast, a few Competibacter (less than 1% of all bacteria) were observed in each of the samples analysed. 3.2. Aerobic operation After the period of standard operation with alternating anaerobic–aerobic phases, the enriched PAO sludge was shifted to permanent aeration operation. The first aerobic cycle of the SBR was monitored (Fig. 2, right) and the results obtained were similar to other previous experiments with simultaneous VFA and oxygen presence [13,14,17]. Two phases can be easily distinguished according to the propionic acid presence. Firstly, a feast phase where propionate uptake was simultaneous to phosphorus release and PHA accumulation. Then, a famine phase took place after external substrate depletion, where PHA utilization, glycogen build up and phosphorus uptake were observed concurrently. Both phases were analogous to the typical anaerobic–aerobic phases of conventional EBPR cycles. Fig. 4 shows three examples of different cycles during the first days of permanent aerobic operation, where propionic acid uptake linked to low P-release and low glycogen consumption are observed during the feast phase, while complete P-uptake is observed during the famine phase. Fig. 5 shows the online data recorded during the SBR cycle monitored on day 8. DO was always controlled between 3.5 and 4.5 mg DO/L to avoid oxygen restrictions and it never limited the aerobic operation of PAO. OUR measurements also corroborate the aerobic operation of the reactor. Fig. 6 shows phosphate profiles in several additional cycles monitored during the aerobic operation and all of them achieved net
P-removal. The profiles on days 20 and 22 were higher because some phosphate precipitation was observed due to the high phosphate concentration in the feed the first days of operation (Fig. 1). This fact was also observed in the decreased ratio VSS/TSS during this period as shown in Fig. 1. Feed phosphate concentration was decreased after day 20 to avoid this problem. The initial ratio VSS/TSS was recovered around 10 days after decreasing its concentration. The last aerobic cycle monitored on day 44 showed the maintenance of biological P-removal with the observed behaviour: P-release during the feast phase and P-uptake during the famine phase of the cycle. The time course of P-release and P-uptake is depicted on Fig. 7. Throughout this period, P-uptake was higher than P-release, indicating stable aerobic P-removal. Lower utilization of glycogen was observed in the aerobic cycles when compared to conventional anaerobic–aerobic cycles. Moreover, the amount of intracellular stored glycogen dropped significantly during the aerobic operation period (Fig. 8), suggesting a lower reducing power requirement from this reserve polymer under aerobic conditions. The tricarboxylic acid cycle (TCA) would be more active under these conditions, producing reducing equivalents and decreasing the glycogen requirements to maintain the redox balance.
Fig. 6. Phosphate concentration profiles during several cycles under the aerobic SBR operation. () Day 20, () day 22, () day 28, (䊉) day 30, () day 44.
M. Vargas et al. / Biochemical Engineering Journal 43 (2009) 288–296
293
Fig. 7. P-release () and P-uptake (䊉) under aerobic and anaerobic/aerobic conditions of operation.
Besides, at the end of the aerobic period, the FISH quantification results showed that approximately 50% of the total bacterial community was Accumulibacter (Fig. 3B). This indicated that the overall PAO population was reduced, but this decrease was not correlated with an increase in Competibacter population since they were almost undetectable (<1% of all bacteria). Microscopic observation of sludge samples also showed the presence of different bacterial morphotypes not targeted by PAOMIX probes, such as filamentous bacteria, that were not present during the anaerobic–aerobic operation. These heterotrophic bacteria probably grow using some of the propionate present at the beginning of the cycle and also using some lysis products. These experimental results suggest that P-removal ability by PAO sludge in the aerobic operation period was linked to an adaptation of their metabolism to permanent aerobic conditions, what would allow them to outcompete other heterotrophic bacteria. Furthermore, this capability was maintained under permanent aerobic conditions during a long period (46 days). 3.3. Anaerobic–aerobic operation The permanent aerobic SBR was switched to conventional alternating anaerobic–aerobic operation on day 47. Experimental
Fig. 9. Experimental profiles of the different off-line measurements during some anaerobic–aerobic cycles monitored in the SBR: (䊉) propionic acid, () phosphate, () PHA, () glycogen. (A) Day 50 (B) day 56 and (C) day 105.
Fig. 8. Intracellular concentration of glycogen at the end of the cycles under aerobic and anaerobic/aerobic conditions of operation.
profiles of different off-line measurements for some significant cycles analysed during this period are depicted in Fig. 9. In the anaerobic phase, propionic acid uptake was linked to phosphate release, glycogen degradation and PHA storage. However, it is also observed some anaerobic P-release linked to PHA storage after VFA depletion. Under aerobic conditions, PHA was consumed, glycogen was produced and phosphate was taken up, leading to net P-removal in the system.
294
M. Vargas et al. / Biochemical Engineering Journal 43 (2009) 288–296
Table 1 Ratios calculated from the experimental results obtained in the SBR operation. Ratios (C-mol or P-mol/C-mol)
PHA/VFA GLY/VFA Carbon recovery ratioa P-release/VFA-uptake
Anaerobic–aerobic operation
Permanent aerobic operation
Anaerobic–aerobic operation
Day 0
Day 1
Day 5
Day 12
Day 50
Day 56
Day 105
0.88 0.38 0.64 0.54
1.05 0.07 0.98 0.99
n.a. 0.27 n.a. 0.22
n.a. 0.03 n.a. 0.21
2.66 0.15 2.32 0.46
1.76 0.13 1.55 0.53
1.18 0.58 0.75 0.56
n.a.: not available. a Carbon recovery ratio = (PHA-synthesised/(VFA-uptake + Glyc-degraded)).
Table 1 shows that the amount of PHA stored in the sludge was higher in the anaerobic–aerobic cycles than in the permanent aerobic cycles, with a PHA-produced/VFA-uptake ratio of 2.66 on day 50. This is probably due to the presence of organic substrates dissolved in the bulk solution, which came from the lysis of microorganisms, since a fraction of the aerobic biomass could be lysed after the shift to the anaerobic–aerobic operation [30]. Biomass decay would provide additional organic substrates available for PAO and lead to an increase in their intracellular storage of PHA. This explanation was in agreement with numerous protozoa and filamentous microorganisms detected by light microscopy examination during the aerobic period and its presence highly decreased during the anaerobic–aerobic operation. In addition, the values obtained (Table 1) for the carbon recovery ratio (calculated as PHA-synthesised/(VFA-uptake + Glyc-degraded)) increased from 0.64 C-mol/C-mol, in standard conditions, to 2.32 (day 50) at the beginning of the second anaerobic–aerobic period, with a progressive reduction along this period (1.55 and 0.75 on days 56 and 105, respectively). These values higher than 1 show that, even taking into account the increased utilization of glycogen, PAO had additional carbon sources available in the reactor during the anaerobic phase. Other explanation would be the excretion of some by-products by some microorganisms, which could be used by PAO to produce additional PHA. Glycogen stored increased progressively as observed in Fig. 8. This increase would be related to higher reducing power and energy requirements in anaerobic–aerobic cycles, when the TCA cycle is not fully effective as it was in permanent aerobic conditions. Some previous studies [2,29] reported that PAO, under anaerobic conditions, were able to degrade a small part of VFA through the TCA cycle to obtain extra reducing power. However, this partial anaerobic activation of TCA would not provide enough reducing power and energy, increasing anaerobic glycogen utilization with respect to the aerobic operation as it was detected in the present study. These results contradict the hypothesis suggested by García Martín et al. [10] that a novel cytochrome encoded in the Accumulibacter genome would allow full performance of the TCA cycle under anaerobic conditions. FISH microbiological analysis showed that more than 72% of all bacteria corresponded to Accumulibacter (Fig. 3C), whereas there was not significant presence of Competibacter cells. Therefore, the amount of Accumulibacter substantially increased (from 50 to 72%) after the change from only aerobic to anaerobic–aerobic conditions, since they were more competitive in anaerobic–aerobic operation than other microorganisms present in the system. 3.4. Comparison between aerobic and anaerobic–aerobic operation Glycogen and PHA were the most affected parameters when the operational conditions were changed in the SBR, as it was previously discussed. Table 1 summarises the experimental ratios obtained during all the SBR operation. The main PHA fractions
synthesised were PH2MV and PHV, which is in agreement with previous experimental results obtained with propionate as carbon source [31]. PHA synthesised under anaerobic operation was 0.04 PHB, 0.45 PHV and 0.54 PH2MV (C-mol synthesised per C-mol of propionic acid consumed). These values agree with the experimental results obtained by Satoh et al. [32]: <0.05, 0.60 and 0.58, and Oehmen et al. [3]: 0.04, 0.55 and 0.65. Similar values were obtained for the aerobic operation (0.02, 0.32 and 0.52). The P-release/VFA-uptake (P-mol/C-mol) ratios calculated were around 0.54 in the standard anaerobic–aerobic operation. This ratio was 0.99 the first day of aerobic operation, but it decreased and stabilised around 0.2 during the aerobic operation (Table 1). Hence, the overall amount of poly-P hydrolysed in the permanent aerobic operation was lower than under anaerobic–aerobic conditions. These results and the lower glycogen consumption suggested that the utilization of the TCA cycle and the electron transport system in the permanent aerobic conditions provided more energy than the less efficient glycogen degradation predominant under anaerobic conditions. In addition, the aerobic operation would allow some direct oxidation of VFA simultaneously to PHA storage [13,33] and hence, it would reduce the total amount of P-release required. At the beginning of the second period of anaerobic–aerobic operation (day 50), this ratio was 0.46 and then it rose progressively up to 0.56. This fact was in agreement with the enrichment in Accumulibacter detected with FISH analysis (from 50 to 72%). Finally, the value 0.56 of P-release/VFA-uptake in anaerobic–aerobic conditions was in the range to the ones found for SBRs also enriched with propionate as a sole carbon source, e.g. 0.33 in Satoh et al. [32] and 0.42 and 0.58 in Oehmen et al. [3,28], respectively. The time course of P-release and P-uptake throughout the study is depicted on Fig. 7. P-uptake was higher than P-release under aerobic and anaerobic–aerobic operation, indicating stable P-removal independently of the applied conditions. 3.5. Stability of aerobic P-removal using propionate as carbon source The results obtained in this study differ from the results shown in Pijuan et al. [17], where it was observed net-P removal deterioration under permanent aerobic conditions using acetate as carbon source. P-release linked to acetate uptake increased slightly during the 11 days of aerobic operation, but P-uptake during the famine phase decreased along the same experimental period. Thus, the observed result was the deterioration of P-removal after 4 days of aerobic operation. Nevertheless, in the present study, feast Prelease and famine P-uptake were maintained during 46 days of aerobic operation. The only difference between both studies was the utilization of propionate as carbon source and consequently, it seems to be a useful substrate to maintain P-removal activity under aerobic operation. On the other hand, Ahn et al. [18] recently proposed an alternative configuration for aerobic P-removal of wastewater with low COD content. The authors showed that when acetate addition was
M. Vargas et al. / Biochemical Engineering Journal 43 (2009) 288–296
temporally separated from phosphate addition, aerobic P-removal capacity was maintained. However, these results were obtained with a different microbial population (22.6% of Accumulibacter, using PAOMIX probes) in a SBR fed with acetate synthetic wastewater. In the present experiments, 50% of Accumulibacter (detected also with PAOMIX probes) were quantified under aerobic conditions, indicating that propionate could be added simultaneously to phosphate without losing aerobic P-removal capacity and therefore, the temporal separation was not necessary. The reason why propionate seems to be a better carbon source under aerobic operation is not clear, although this substrate was already shown to be favourable for PAO growth under conventional anaerobic–aerobic operation [3,7]. The main difference between acetate and propionate under aerobic conditions is that feast Prelease was maintained in both cases but famine P-uptake was progressively lost when operating with acetate [17]. During the famine phase, the stored PHA is used for growth, recovery of glycogen pools and phosphate storage as polyphosphate. Nevertheless, the aerobic operation with acetate or propionate demonstrated that glycogen is slightly consumed. Therefore, the only difference between both cases is the PHA composition: acetate is stored mainly as PHB while propionate produces PHV and PH2MV. This different composition may be related to the stability observed with propionate because PHA composition has been reported as potentially a key aspect of aerobic phosphorus uptake in EBPR [34]. 4. Conclusions The performance of the EBPR process under permanent aerobic conditions using propionate was investigated. The most important conclusions that can be drawn from this study are: • Propionate enriched PAO population developed under standard anaerobic–aerobic cycles was able to maintain biological phosphate removal when it was subjected to strict aerobic conditions during a long period (46 days). Famine P-uptake was achieved in all aerobic cycles resulting in a net P-removal in the system. • The drop observed in the stored glycogen under aerobic operation suggests that propionate oxidation through the TCA cycle can replace the role of glycogen in the generation of reducing equivalents. • The configuration tested shows the feasibility to treat wastewater with simultaneous presence of phosphate and propionate under permanent aerobic conditions. • PAO maintained the EBPR activity during the subsequent operation under standard anaerobic–aerobic conditions. For this purpose, they increased the utilization of glycogen, recovering the typical anaerobic PAO metabolism. • Accumulibacter cells were the dominant population in all periods studied. Its percentage started at 70%, decreased to 50% under aerobic operation and finally recovered to 72% under the anaerobic–aerobic operation. The aerobic decrease was associated with an increase of filamentous bacteria. Acknowledgements This work was supported by the Spanish Ministerio de Educación y Ciencia (CTM2004-02569/TECNO and CTQ2007-61756/PPQ). The authors are members of the GENOCOV group (Grup de Recerca Consolidat de la Generalitat de Catalunya). The Department of Chemical Engineering at UAB is a unit of Biochemical Engineering of the Xarxa de Referència en Biotecnologia de Catalunya (XRB), Generalitat de Catalunya. We also acknowledge the technical support from the “Servei de Microscopia” of the Universitat Autònoma de Barcelona.
295
References [1] T. Mino, V. Arun, Y. Tsuzuki, T. Matsuo, Effect of phosphorus accumulation on acetate metabolism in the biological phosphorus removal, in: R. Ramadori (Ed.), Advanced in Water Pollution Control: Biological phosphate removal from Wastewater, Pergamon Press, Oxford, 1987. [2] P.C. Lemos, L.S. Serafim, M.M. Santos, M.A.M. Reis, H. Santos, Metabolic pathway for propionate utilization by phosphorus accumulating organisms in activated sludge: 13 C labelling and in vivo nuclear magnetic resonance, Appl. Environ. Microbiol. 69 (1) (2003) 241–251. [3] A. Oehmen, R. Zeng, Z. Yuan, J. Keller, Anaerobic metabolism of propionate by polyphosphate-accumulating organisms in enhanced biological phosphorus removal systems, Biotechnol. Bioeng. 91 (1) (2005) 43–53. [4] M. Pijuan, A. Oehmen, J.A. Baeza, C. Casas, Z. Yuan, Characterizing the biochemical activity of full-scale enhanced biological phosphorus removal systems: a comparison with metabolic models, Biotechnol. Bioeng. 99 (1) (2008) 170– 179. [5] C.R. Hood, A.A. Randall, A biochemical hypothesis explaining the response of enhanced biological phosphorus removal biomass to organic substrates, Water Res. 35 (11) (2001) 2758–2766. [6] M. Thomas, P. Wright, L. Blackall, V. Urbain, J. Keller, Optimisation of Nossa BNR plant to improve performance and reduced operating costs, Water Sci. Technol. 47 (12) (2003) 141–148. [7] M. Pijuan, A.M. Saunders, A. Guisasola, J.A. Baeza, C. Casas, L.L. Blackall, Enhanced biological phosphorous removal in a sequencing batch reactor using propionate as the sole carbon source, Biotechnol. Bioeng. 85 (1) (2004) 56–67. [8] Y. Chen, A.A. Randall, T. McCue, The efficiency of enhanced biological phosphorous removal from real wastewater affected by different ratios of acetic to propionic acid, Water Res. 38 (1) (2004) 27–36. [9] H. Lu, A. Oehmen, B. Virdis, J. Keller, Z. Yuan, Obtaining highly enriched cultures of Candidatus Accumulibacter Phosphatis through alternating carbon sources, Water Res. 40 (20) (2006) 3838–3848. [10] H. García Martín, N. Ivanova, V. Kunin, F. Warnecke, K.W. Barry, A.C. McHardy, C. Yeates, S. He, A.A. Salamov, E. Szeto, E. Dalin, N.K. Putnam, H.J. Shaphiro, J.L. Pangilinan, I. Rigoustos, N.C. Kyripides, L.L. Blackall, K.D. McMahon, P. Hugenholtz, Metagenomic analysis of two enhanced biological phosphorus removal (EBPR) communities, Nat. Biotechnol. 24 (10) (2006) 1263–1269. [11] D. Brdjanovic, A. Slamet, M.C.M. van Loosdrecht, C.M. Hooijmans, G.J. Alaerts, J.J. Heijnen, Impact of excessive aeration on biological phosphorus removal from wastewater, Water Res. 32 (1) (1998) 200–208. [12] J. Ahn, T. Daidou, S. Tsuneda, A. Hirata, Transformation of phosphorus and relevant intracellular compounds by a phosphorus-accumulating enrichment culture in the presence of both the electron acceptor and electron donor, Biotechnol. Bioeng. 79 (1) (2002) 83–93. [13] A. Guisasola, M. Pijuan, J.A. Baeza, J. Carrera, C. Casas, J. Lafuente, Aerobic phosphorus release linked to acetate uptake in bio-P sludge: process modelling using oxygen uptake rate, Biotechnol. Bioeng. 85 (7) (2004) 722–723. [14] M. Pijuan, A. Guisasola, J.A. Baeza, J. Carrera, C. Casas, J. Lafuente, Aerobic phosphorus release linked to acetate uptake: influence of PAO intracellular storage compounds, Biochemical Eng. J. 26 (2–3) (2005) 184–190. [15] J.J. Beun, E.V. Verhoef, M.C.M. Van Loosdrecht, J.J. Heijnen, Stoichiometry and kinetics of poly--hydroxybutyrate metabolism under denitrifying conditions in activated sludge cultures, Biotechnol. Bioeng. 68 (5) (2000) 497–507. [16] K. Dircks, M. Henze, M.C.M. van Loosdrecht, H. Mosbaek, H. Aspegren, Storage and degradation of poly--hydroxybutyrate in activated sludge under aerobic conditions, Water Res. 35 (9) (2001) 2277–2285. [17] M. Pijuan, A. Guisasola, J.A. Baeza, J. Carrera, C. Casas, J. Lafuente, Net P-removal deterioration in enriched PAO sludge subjected to permanent aerobic conditions, J. Biotechnol. 123 (1) (2006) 117–126. [18] J. Ahn, S. Schroeder, M. Beer, S. McIlroy, R.C. Bayly, J.W. May, G. Vasiliadis, R. Seviour, Ecology of the microbial community removing phosphate from wastewater under continously aerobic conditions in a sequencing batch reactor, Appl. Environ. Microbiol. 73 (7) (2007) 2257–2270. [19] A. Guisasola, M. Vargas, M. Marcelino, J. Lafuente, C. Casas, J.A. Baeza, Online monitoring of the enhanced biological phosphorus removal process using respirometry and titrimetry, Biochem. Eng. J. 35 (3) (2007) 371–379. [20] G.J.F. Smolders, J. van der Meij, M.C.M. van Loosdrecht, J.J. Heijnen, Model of the anaerobic metabolism of the biological phosphorus removal process: stoichiometry and pH influence, Biotechnol. Bioeng. 43 (6) (1994) 461–470. [21] A. Oehmen, B.K. Lehmann, R.J. Zeng, Z. Yuan, J. Keller, Optimisation of poly-hydroxyalkanoate analysis using gas chromatography for enhanced biological removal systems, J. Chromatogr. A 1070 (1–2) (2005) 131–136. [22] APHA, Standard Methods for the Examination of Water and Wastewater, 19th ed., American Publishers Health Association, 1995. [23] R.I. Amann, W. Ludwig, K. Schleifer, Phylogenetic identification and in situ detection of individual microbial cells without cultivation, Microbiol. Rev. 59 (1) (1995) 143–169. [24] H. Daims, A. Bruhl, R. Amann, K.H. Schleifer, M. Wagner, The domain-specific probe EUB338 is insufficient for the detection of all bacteria: development and evaluation of a more comprehensive probe set, Syst. Appl. Microbiol. 22 (3) (1999) 434–444. [25] G.R. Crocetti, P. Hugenholtz, P.L. Bond, A. Schuler, J. Keller, D. Jenkins, L.L. Blackall, Identification of polyphosphate accumulating organisms and design of 16S rRNA-directed probes for their detection and quantitation, Appl. Environ. Microbiol. 66 (3) (2000) 1175–1182.
296
M. Vargas et al. / Biochemical Engineering Journal 43 (2009) 288–296
[26] G.R. Crocetti, J.F. Banfield, J. Keller, P.L. Bond, L.L. Blackall, Glycogenaccumulating organisms in laboratory-scale and full-scale wastewater treatment processes, Microbiology 148 (11) (2002) 3353–3364. [27] MATLAB, Image Processing Toolbox for MATLAB, User’s Guide, Version 5, 2005, The MathWorks, Natick, MA, USA. [28] A. Oehmen, Z. Yuan, L.L. Blackall, J. Keller, Comparison of acetate and propionate uptake by polyphosphate accumulating organisms and glycogen accumulating organisms, Biotechnol. Bioeng. 91 (2) (2005) 162–168. [29] H. Pereira, P. Lemos, M. Reis, J. Crespo, M. Carrondo, H. Santos, Model for carbon metabolism in biological phosphorus removal processes based on in vivo CNMR labelling experiments, Water Res. 30 (9) (1996) 2128–2138. [30] P. Chudoba, A. Morel, B. Capdeville, The case of both energetic uncoupling and metabolic selection of microorganisms in the osa activated-sludge system, Environ. Technol. 13 (8) (1992) 761–770.
[31] M. Pijuan, J.A. Baeza, C. Casas, J. Lafuente, Response of an EBPR population developed in a SBR with propionate to different carbon sources, Water Sci. Technol. 50 (10) (2004) 131–138. [32] H. Satoh, T. Mino, T. Matsuo, Uptake of organic substrates and accumulation of polyhydroxialcanoates linked with glycolysis of intracellular carbohydrates under anaerobic conditions in the biological excess phosphate removal processes, Water Sci. Technol. 26 (5–6) (1992) 933–942. [33] G. Sin, A. Guisasola, D.J.W. De Pauw, J.A. Baeza, J. Carrera, P.A. Vanrolleghem, A new approach for modelling simultaneous storage and growth processes for activated sludge systems under aerobic conditions, Biotechnol. Bioeng. 92 (2005) 600–613. [34] A.A. Randall, Y.H. Liu, Polyhydroxyalkanoates form potentially a key aspect of aerobic phosphorus uptake in enhanced biological phosphorus removal, Water Res. 36 (14) (2002) 3473–3478.
Contents lists available at ScienceDirect
Biochemical Engineering Journal journal homepage: www.elsevier.com/locate/bej
Maintenance of phosphorus removal in an EBPR system under permanent aerobic conditions using propionate M. Vargas, C. Casas, J.A. Baeza ∗ Departament d’Enginyeria Química, ETSE, Universitat Autònoma de Barcelona, 08193 Bellaterra, Barcelona, Spain
a r t i c l e
i n f o
Article history: Received 10 June 2008 Received in revised form 22 September 2008 Accepted 18 October 2008 Keywords: Aerobic processes Enhanced biological phosphorus removal (EBPR) Phosphorus accumulating organisms (PAO) Propionic acid Wastewater treatment
a b s t r a c t Enhanced biological phosphorus removal (EBPR) is an efficient and sustainable technology to remove phosphorus from wastewater preventing eutrophication in natural waters. It is widely accepted that EBPR requires an optimal anaerobic hydraulic retention time to obtain stable P-removal from wastewater. Thus, it is suggested that deterioration of the EBPR efficiency regularly observed in full-scale wastewater treatment plants (WWTPs) is normally caused by an excessive aeration of activated sludge that increments the amount of oxygen recycled to the anaerobic reactor and consequently, the anaerobic conditions are not totally preserved. Furthermore, it has been reported a progressive decrease in P-removal capacity in an EBPR lab-scale system enriched with acetate as the sole carbon source under permanent aerobic conditions. Hence, to evaluate the stability of P-removal with a different carbon source, an EBPR-SBR was operated with propionate under permanent aerobic conditions. As a result, net P-removal was successfully accomplished in the SBR without any anaerobic phase during 46 days of aerobic operation. Moreover, the system was shifted after this period to the standard anaerobic–aerobic conditions and reliable Premoval was maintained. FISH (fluorescence in situ hybridisation) analysis showed a significant presence of Accumulibacter (70, 50 and 72%, in different periods) and the absence of Competibacter. The results indicate that using propionate as carbon source it is possible to maintain in a long term an enriched culture of phosphorus accumulating organisms (PAO) able to remove phosphorus under permanent aerobic conditions. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Enhanced biological phosphorus removal (EBPR) is an efficient and sustainable biological technology to remove phosphorus from wastewater. It is based on the enrichment of activated sludge with phosphorus accumulating organisms (PAO) by introducing alternating anaerobic–aerobic conditions, which favour PAO growth over ordinary heterotrophic organisms (OHO). Under anaerobic conditions, PAO store volatile fatty acids (VFA) as polyhydroxyalkanoates (PHA) with concomitant P-release into solution derived from the hydrolysis of polyphosphate (poly-P) reserves. The required reducing equivalents are provided by the catabolism of stored glycogen [1]. In a subsequent aerobic (or anoxic) stage, PHA is used as carbon and energy sources for growth and replenishment of the internal glycogen and poly-P pools. An amount of phosphate higher than that released is taken up by PAO and the poly-P pools are restored. As a result, net P-removal is accomplished by periodic sludge waste after this phase.
∗ Corresponding author. Tel.: +34 93 581 1587; fax: +34 93 581 2013. E-mail address: [email protected] (J.A. Baeza). 1369-703X/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.bej.2008.10.013
On the other hand, even though the two most common VFA present in municipal wastewater are acetate and propionate, most of works on EBPR consider acetate as sole carbon source. However, the metabolism of propionate by PAO has attracted recently considerable attention [2–4] and it is suggested that it could be a more favourable substrate for EBPR [5–9]. Although many works are focused on this complex process, the complete metabolic behaviour of the EBPR sludge is still unclear and it is prone to unpredictable failures, primarily because the process design is highly empirical due to an incomplete understanding of sludge microbial ecology [10]. For example, Brdjanovic et al. [11] hypothesised that one cause of EBPR efficiency deterioration could be excessive aeration, which would result in a gradual depletion of PHB and/or saturation of the biomass by poly-P. The two most common causes of excessive aeration in continuous wastewater treatment plants (WWTPs) are heavy rainfall periods and failures in the aeration control, which increment the amount of oxygen recycled to the anaerobic reactor. In both scenarios, the anaerobic conditions are reduced and the aerobic phase is extended leading to a detriment in P-removal. To the best of our knowledge, successful EBPR operation without a physical separation between the electron donor (organic matter)
M. Vargas et al. / Biochemical Engineering Journal 43 (2009) 288–296
289
Fig. 1. Time course of TSS, VSS, VSS/TSS ratio and changes in the feed composition during both SBR periods studied. (䊉) VSS, () TSS, (♦) VSS/TSS, () phosphate, () propionic acid.
and the electron acceptor (oxygen) has not been reported yet in full scale WWTP. In conventional continuous WWTP, the inclusion of a first anaerobic reactor allows the separation of organic matter and oxygen. These conditions favour PAO against OHO because PAO are more efficient to anaerobically uptake VFA due to their polyphosphate reserves. Nevertheless, some recent studies showed the feasibility to achieve net P-removal under strictly aerobic conditions in well-controlled lab-scale EBPR reactors [12–14]. In these studies, aerobic acetate uptake was linked to phosphate release, PHA storage and glycogen degradation. After substrate depletion, phosphate was taken up linked to PHA degradation, glycogen synthesis and PAO growth. The phases with and without available external substrate are known in the literature [15–16] as feast and famine, respectively. According to the pattern observed, these stages were analogous to the anaerobic–aerobic conditions of the typical EBPR process [13]. Moreover, Pijuan et al. [17] studied the P-removal deterioration in enriched PAO sludge subjected to permanent aerobic conditions. An aerobic SBR was operated using acetate as carbon source, observing the abovementioned feast/famine phases. However, net P-removal was only maintained during the first 4 days of operation, showing that the tested conditions did not provide a stable operation for aerobic EBPR. Finally, Ahn et al. [18] proposed recently a new operational scheme for aerobic phosphate removal from wastewater consisting of an aerobic SBR where the feed stage (acetate addition) was temporally separated from phosphate addition, which begins the famine stage. The authors showed that this process was capable to remove phosphate from 10 to 12 mg P/L to less than 0.1 mg P/L over an extended period, using wastewater with low COD content. This paper aims to propose an alternative for stable aerobic P-removal. It tests if EBPR can be maintained in a SBR under permanent aerobic conditions using propionate as the sole carbon source. Additionally, after this strict aerobic operation, the stability of net P-removal is examined when the SBR is changed to conventional EBPR anaerobic–aerobic operation.
2. Materials and methods 2.1. Equipments and experimental design The 10 L SBR used for the experiments was seeded with sludge obtained from an EBPR anaerobic–aerobic SBR [19]. It was operated at four cycles per day with a hydraulic residence time (HRT) of 12 h.
Each cycle consisted of 120 min of anaerobic phase (5 initial min for feeding), 180 min of aerobic phase, 55 min of settling and 5 min of extraction of 5 L of supernatant. The sludge residence time (SRT) was kept at 10 days by periodic sludge wastage during the end of the aerobic period but before stirring was stopped. After 5 months of stable and efficient EBPR operation, the SBR was changed to permanent aerobic operation (from day 1 to 46) to investigate the short-term and long-term aerobic P-removal capability of PAO. Each cycle consisted of 300 min of aerobic phase (5 initial min for feeding), 55 min of settling and, in the last 5 min, withdrawing of 5 L of supernatant. SRT was kept also at 10 days. After the 46 days of aerobic operation, the SBR was operated with standard anaerobic–aerobic cycles for 59 days (from day 47 to 105) to evaluate the behaviour of the system regarding P-removal stability. Temperature in the SBR was kept at 25 ◦ C and the pH was maintained below 7.5 with HCl 1 M. Dissolved oxygen (DO) was controlled during the aerobic phase between 3.5 and 4.5 mg DO/L with an on/off controller to avoid oxygen limitations. Oxygen Uptake Rate (OUR) was estimated on-line by measuring the DO decrease rate between 4.5 and 3.5 mg DO/L during the off position of the aeration valve. The average biomass concentration maintained in the reactor during the whole period of study was around 4 g VSS/L, although some variations of TSS and VSS were observed (Fig. 1).
2.2. Synthetic media Two separate solutions called “concentrated feed” (constituting 0.25 L per 5 L synthetic wastewater) and “P-water” (constituting 4.75 L per 5 L synthetic wastewater) collectively formed the synthetic wastewater used in this study. The “concentrated feed” consisted of (g/L RO water): 12 (or 7.2) propionic acid (to achieve 300 mg/L or 180 mg/L in the SBR after feeding), 2.0 NH4 Cl, 0.88 MgSO4 ·7H2 O, 3.2 MgCl2 ·6H2 O, 0.84 CaCl2 ·2H2 O, 0.4 yeast extract, 0.10 allylthiourea (ATU) to inhibit nitrification and 6 mL of nutrient solution. The P-water consisted of 1:0.6 KH2 PO4 /K2 HPO4 (molar ratio). The concentration of P and propionic acid changed along the study (see Fig. 1). Prior to the permanent aerobic cycles performed in the SBR, it was operated with initial concentrations of 15 mg/L of P and 180 mg/L of propionic acid. Afterwards, in the third day of permanent aerobic operation, they were incremented up to 80 and 300 mg/L, respectively, for a better monitoring of the variables measured in the study cycles. Later on (day 19), phosphate concen-
290
M. Vargas et al. / Biochemical Engineering Journal 43 (2009) 288–296
Fig. 2. Experimental profiles during a standard anaerobic–aerobic SBR operation (left) and the first day under aerobic conditions (right): (䊉) propionic acid, () phosphate, () PHA, () glycogen.
Fig. 3. CLSM FISH micrographs of the EBPR sludge hybridised with probes specific for “Candidatus Accumulibacter phosphatis” (in pink, Cy3-labelled PAOMIX probes) and probes for bacteria (in blue, Cy5-labelled EUBMIX probes) in different periods. (A) Standard SBR operation; (B) aerobic period, day 46; (C) anaerobic–aerobic period, day 105. Bar = 20 m
M. Vargas et al. / Biochemical Engineering Journal 43 (2009) 288–296
291
Chromatography (HPLC). A Hewlett Packard 1050 equipped with a Bio Rad Aminex HPX-87H ion exclusion column at 25 ◦ C was used. Glycogen was determined by a modification of the method of Smolders et al. [20]. A volume of 5 mL of 0.6 M HCl was added to each 20 mg of lyophilised sludge sample, and then heated at 100 ◦ C for 6 h. After cooling and filtering through 0.22 m filter (Millipore), the concentration of glucose was measured using an YSI Model 2700 Select Biochemistry Analyser (Yellow Springs Instrument, Yellow Springs, Ohio, USA). Triplicates of each sample were done. PHA was measured according to the methods of Oehmen et al. [21]. 40 mg of lyophilised sludge samples was digested and methylated with 4 mL of acidulated methanol (10% H2 SO4 ) and 4 mL of chloroform during 20 h at 100 ◦ C. Benzoic acid was used as internal standard. The calibration of the method was performed using a 3hydroxybutyric acid and 3-hydroxyvaleric acid copolymer (7:3) as standard for PHB and PHV (Fluka, Buchs SG, Switzerland) and 2hydroxycaproic acid as standard for PH2MV (Aldrich). The analyses were performed in a GC system, operated with a Hewlett Packard 5890 column (30 m length × 0.53 mm I.D. × 1.00 m film). Triplicates of each sample were done. Total suspended solids (TSS) and volatile suspended solids (VSS) were analysed according to Standard Methods [22]. 2.4. Microbial analyses
Fig. 4. Experimental profiles during permanent aerobic cycles monitored in the SBR: (䊉) propionic acid, () phosphate, () glycogen; (A) day 4; (B) day 8 and (C) day 13.
tration was decreased to 20 mg/L to avoid possible precipitation of this element in the bulk solution.
Representative samples were withdrawn from the SBR to perform fluorescence in situ hybridisation (FISH) technique. FISH was performed as in Amann [23] with Cy5-labelled EUBMIX probes for most bacteria [24], and either Cy3-labelled PAOMIX probes for “Candidatus Accumulibacter phosphatis”, comprising equal amounts of probes PAO462, PAO651 and PAO846 [25] or Cy3-labelled GAOMIX for “Candidatus Competibacter phosphatis”, comprising equal amounts of probes GAOQ431 and GAOQ989 [26]. Sludge was fixed by 4% paraformaldehyde solution in phosphate buffer 0.03 M. Hybridisation was performed with 35% formamide solution at 46 ◦ C for 2 h. FISH preparations were visualised with a Leica TCS SP2 AOBS confocal laser scanning microscope (CLSM). Accumulibacter or Competibacter were quantified as a proportion of all bacteria using image analysis with the methodology below summarised, performed with specific software developed with Matlab Image Processing Toolbox [27]. To avoid colony formers underestimation, each sample was homogenised and just one layer of sludge in each well of the slide was applied when FISH was performed. Forty randomly chosen CLSM fields from different x, y, and z coordinates were imaged (600× magnification) from each sample. Three images obtained with the same sludge and procedures, but without addition of probes, were used for evaluating the sample autofluorescence. The threshold in pixel intensity for removing autofluorescence was selected to eliminate the 99.99% of pixels with positive signal of these images. To account for the different amount of biomass in each image, the area of specific cell signal was calculated with the addition of all the detected pixels with intensity above the previously evaluated threshold for Cy3 in the 40 images divided by the total amount of pixels above its corresponding threshold detected for Cy5.
2.3. Chemical and biochemical analyses
3. Results and discussion
Analyses of phosphate in filtered samples were performed with an Electrophoresis Capillary System (Quanta 4000E CE – WATERS). The electrolyte used was the commercial solution Ionselect High Mobility Anion Electrolite. Propionic acid was measured by Liquid
3.1. Standard reactor performance The enriched PAO sludge obtained under standard EBPR operation (anaerobic–aerobic) showed a good performance of P-removal.
292
M. Vargas et al. / Biochemical Engineering Journal 43 (2009) 288–296
Fig. 5. On-line measurements during one aerobic SBR cycle on day 8.
Fig. 2 (left) depicts the experimental profiles for some significant measurements (phosphate, propionic acid, PHA and glycogen) obtained during a cycle of the SBR after 5 months of operation. The typical behaviour of EBPR sludge is clearly observed. In the anaerobic phase, propionic acid uptake was linked to phosphate release derived from the hydrolysis of poly-P stored, glycogen degradation and PHA storage. Under aerobic conditions, PHA was consumed and glycogen accumulated meanwhile phosphate was removed from the bulk solution, leading to a net P-removal in the system. FISH analysis of sludge samples withdrawn from this SBR cycle showed that Accumulibacter accounted for 70% of all bacteria (see Fig. 3A). This significant percentage of population of Accumulibacter is similar to others obtained in lab-scale SBR sludges enriched with propionate under anaerobic–aerobic conditions (e.g. 55% in Pijuan et al. [7]; 65% in Oehmen et al. [28] and 90% in Lu et al. [9]). In contrast, a few Competibacter (less than 1% of all bacteria) were observed in each of the samples analysed. 3.2. Aerobic operation After the period of standard operation with alternating anaerobic–aerobic phases, the enriched PAO sludge was shifted to permanent aeration operation. The first aerobic cycle of the SBR was monitored (Fig. 2, right) and the results obtained were similar to other previous experiments with simultaneous VFA and oxygen presence [13,14,17]. Two phases can be easily distinguished according to the propionic acid presence. Firstly, a feast phase where propionate uptake was simultaneous to phosphorus release and PHA accumulation. Then, a famine phase took place after external substrate depletion, where PHA utilization, glycogen build up and phosphorus uptake were observed concurrently. Both phases were analogous to the typical anaerobic–aerobic phases of conventional EBPR cycles. Fig. 4 shows three examples of different cycles during the first days of permanent aerobic operation, where propionic acid uptake linked to low P-release and low glycogen consumption are observed during the feast phase, while complete P-uptake is observed during the famine phase. Fig. 5 shows the online data recorded during the SBR cycle monitored on day 8. DO was always controlled between 3.5 and 4.5 mg DO/L to avoid oxygen restrictions and it never limited the aerobic operation of PAO. OUR measurements also corroborate the aerobic operation of the reactor. Fig. 6 shows phosphate profiles in several additional cycles monitored during the aerobic operation and all of them achieved net
P-removal. The profiles on days 20 and 22 were higher because some phosphate precipitation was observed due to the high phosphate concentration in the feed the first days of operation (Fig. 1). This fact was also observed in the decreased ratio VSS/TSS during this period as shown in Fig. 1. Feed phosphate concentration was decreased after day 20 to avoid this problem. The initial ratio VSS/TSS was recovered around 10 days after decreasing its concentration. The last aerobic cycle monitored on day 44 showed the maintenance of biological P-removal with the observed behaviour: P-release during the feast phase and P-uptake during the famine phase of the cycle. The time course of P-release and P-uptake is depicted on Fig. 7. Throughout this period, P-uptake was higher than P-release, indicating stable aerobic P-removal. Lower utilization of glycogen was observed in the aerobic cycles when compared to conventional anaerobic–aerobic cycles. Moreover, the amount of intracellular stored glycogen dropped significantly during the aerobic operation period (Fig. 8), suggesting a lower reducing power requirement from this reserve polymer under aerobic conditions. The tricarboxylic acid cycle (TCA) would be more active under these conditions, producing reducing equivalents and decreasing the glycogen requirements to maintain the redox balance.
Fig. 6. Phosphate concentration profiles during several cycles under the aerobic SBR operation. () Day 20, () day 22, () day 28, (䊉) day 30, () day 44.
M. Vargas et al. / Biochemical Engineering Journal 43 (2009) 288–296
293
Fig. 7. P-release () and P-uptake (䊉) under aerobic and anaerobic/aerobic conditions of operation.
Besides, at the end of the aerobic period, the FISH quantification results showed that approximately 50% of the total bacterial community was Accumulibacter (Fig. 3B). This indicated that the overall PAO population was reduced, but this decrease was not correlated with an increase in Competibacter population since they were almost undetectable (<1% of all bacteria). Microscopic observation of sludge samples also showed the presence of different bacterial morphotypes not targeted by PAOMIX probes, such as filamentous bacteria, that were not present during the anaerobic–aerobic operation. These heterotrophic bacteria probably grow using some of the propionate present at the beginning of the cycle and also using some lysis products. These experimental results suggest that P-removal ability by PAO sludge in the aerobic operation period was linked to an adaptation of their metabolism to permanent aerobic conditions, what would allow them to outcompete other heterotrophic bacteria. Furthermore, this capability was maintained under permanent aerobic conditions during a long period (46 days). 3.3. Anaerobic–aerobic operation The permanent aerobic SBR was switched to conventional alternating anaerobic–aerobic operation on day 47. Experimental
Fig. 9. Experimental profiles of the different off-line measurements during some anaerobic–aerobic cycles monitored in the SBR: (䊉) propionic acid, () phosphate, () PHA, () glycogen. (A) Day 50 (B) day 56 and (C) day 105.
Fig. 8. Intracellular concentration of glycogen at the end of the cycles under aerobic and anaerobic/aerobic conditions of operation.
profiles of different off-line measurements for some significant cycles analysed during this period are depicted in Fig. 9. In the anaerobic phase, propionic acid uptake was linked to phosphate release, glycogen degradation and PHA storage. However, it is also observed some anaerobic P-release linked to PHA storage after VFA depletion. Under aerobic conditions, PHA was consumed, glycogen was produced and phosphate was taken up, leading to net P-removal in the system.
294
M. Vargas et al. / Biochemical Engineering Journal 43 (2009) 288–296
Table 1 Ratios calculated from the experimental results obtained in the SBR operation. Ratios (C-mol or P-mol/C-mol)
PHA/VFA GLY/VFA Carbon recovery ratioa P-release/VFA-uptake
Anaerobic–aerobic operation
Permanent aerobic operation
Anaerobic–aerobic operation
Day 0
Day 1
Day 5
Day 12
Day 50
Day 56
Day 105
0.88 0.38 0.64 0.54
1.05 0.07 0.98 0.99
n.a. 0.27 n.a. 0.22
n.a. 0.03 n.a. 0.21
2.66 0.15 2.32 0.46
1.76 0.13 1.55 0.53
1.18 0.58 0.75 0.56
n.a.: not available. a Carbon recovery ratio = (PHA-synthesised/(VFA-uptake + Glyc-degraded)).
Table 1 shows that the amount of PHA stored in the sludge was higher in the anaerobic–aerobic cycles than in the permanent aerobic cycles, with a PHA-produced/VFA-uptake ratio of 2.66 on day 50. This is probably due to the presence of organic substrates dissolved in the bulk solution, which came from the lysis of microorganisms, since a fraction of the aerobic biomass could be lysed after the shift to the anaerobic–aerobic operation [30]. Biomass decay would provide additional organic substrates available for PAO and lead to an increase in their intracellular storage of PHA. This explanation was in agreement with numerous protozoa and filamentous microorganisms detected by light microscopy examination during the aerobic period and its presence highly decreased during the anaerobic–aerobic operation. In addition, the values obtained (Table 1) for the carbon recovery ratio (calculated as PHA-synthesised/(VFA-uptake + Glyc-degraded)) increased from 0.64 C-mol/C-mol, in standard conditions, to 2.32 (day 50) at the beginning of the second anaerobic–aerobic period, with a progressive reduction along this period (1.55 and 0.75 on days 56 and 105, respectively). These values higher than 1 show that, even taking into account the increased utilization of glycogen, PAO had additional carbon sources available in the reactor during the anaerobic phase. Other explanation would be the excretion of some by-products by some microorganisms, which could be used by PAO to produce additional PHA. Glycogen stored increased progressively as observed in Fig. 8. This increase would be related to higher reducing power and energy requirements in anaerobic–aerobic cycles, when the TCA cycle is not fully effective as it was in permanent aerobic conditions. Some previous studies [2,29] reported that PAO, under anaerobic conditions, were able to degrade a small part of VFA through the TCA cycle to obtain extra reducing power. However, this partial anaerobic activation of TCA would not provide enough reducing power and energy, increasing anaerobic glycogen utilization with respect to the aerobic operation as it was detected in the present study. These results contradict the hypothesis suggested by García Martín et al. [10] that a novel cytochrome encoded in the Accumulibacter genome would allow full performance of the TCA cycle under anaerobic conditions. FISH microbiological analysis showed that more than 72% of all bacteria corresponded to Accumulibacter (Fig. 3C), whereas there was not significant presence of Competibacter cells. Therefore, the amount of Accumulibacter substantially increased (from 50 to 72%) after the change from only aerobic to anaerobic–aerobic conditions, since they were more competitive in anaerobic–aerobic operation than other microorganisms present in the system. 3.4. Comparison between aerobic and anaerobic–aerobic operation Glycogen and PHA were the most affected parameters when the operational conditions were changed in the SBR, as it was previously discussed. Table 1 summarises the experimental ratios obtained during all the SBR operation. The main PHA fractions
synthesised were PH2MV and PHV, which is in agreement with previous experimental results obtained with propionate as carbon source [31]. PHA synthesised under anaerobic operation was 0.04 PHB, 0.45 PHV and 0.54 PH2MV (C-mol synthesised per C-mol of propionic acid consumed). These values agree with the experimental results obtained by Satoh et al. [32]: <0.05, 0.60 and 0.58, and Oehmen et al. [3]: 0.04, 0.55 and 0.65. Similar values were obtained for the aerobic operation (0.02, 0.32 and 0.52). The P-release/VFA-uptake (P-mol/C-mol) ratios calculated were around 0.54 in the standard anaerobic–aerobic operation. This ratio was 0.99 the first day of aerobic operation, but it decreased and stabilised around 0.2 during the aerobic operation (Table 1). Hence, the overall amount of poly-P hydrolysed in the permanent aerobic operation was lower than under anaerobic–aerobic conditions. These results and the lower glycogen consumption suggested that the utilization of the TCA cycle and the electron transport system in the permanent aerobic conditions provided more energy than the less efficient glycogen degradation predominant under anaerobic conditions. In addition, the aerobic operation would allow some direct oxidation of VFA simultaneously to PHA storage [13,33] and hence, it would reduce the total amount of P-release required. At the beginning of the second period of anaerobic–aerobic operation (day 50), this ratio was 0.46 and then it rose progressively up to 0.56. This fact was in agreement with the enrichment in Accumulibacter detected with FISH analysis (from 50 to 72%). Finally, the value 0.56 of P-release/VFA-uptake in anaerobic–aerobic conditions was in the range to the ones found for SBRs also enriched with propionate as a sole carbon source, e.g. 0.33 in Satoh et al. [32] and 0.42 and 0.58 in Oehmen et al. [3,28], respectively. The time course of P-release and P-uptake throughout the study is depicted on Fig. 7. P-uptake was higher than P-release under aerobic and anaerobic–aerobic operation, indicating stable P-removal independently of the applied conditions. 3.5. Stability of aerobic P-removal using propionate as carbon source The results obtained in this study differ from the results shown in Pijuan et al. [17], where it was observed net-P removal deterioration under permanent aerobic conditions using acetate as carbon source. P-release linked to acetate uptake increased slightly during the 11 days of aerobic operation, but P-uptake during the famine phase decreased along the same experimental period. Thus, the observed result was the deterioration of P-removal after 4 days of aerobic operation. Nevertheless, in the present study, feast Prelease and famine P-uptake were maintained during 46 days of aerobic operation. The only difference between both studies was the utilization of propionate as carbon source and consequently, it seems to be a useful substrate to maintain P-removal activity under aerobic operation. On the other hand, Ahn et al. [18] recently proposed an alternative configuration for aerobic P-removal of wastewater with low COD content. The authors showed that when acetate addition was
M. Vargas et al. / Biochemical Engineering Journal 43 (2009) 288–296
temporally separated from phosphate addition, aerobic P-removal capacity was maintained. However, these results were obtained with a different microbial population (22.6% of Accumulibacter, using PAOMIX probes) in a SBR fed with acetate synthetic wastewater. In the present experiments, 50% of Accumulibacter (detected also with PAOMIX probes) were quantified under aerobic conditions, indicating that propionate could be added simultaneously to phosphate without losing aerobic P-removal capacity and therefore, the temporal separation was not necessary. The reason why propionate seems to be a better carbon source under aerobic operation is not clear, although this substrate was already shown to be favourable for PAO growth under conventional anaerobic–aerobic operation [3,7]. The main difference between acetate and propionate under aerobic conditions is that feast Prelease was maintained in both cases but famine P-uptake was progressively lost when operating with acetate [17]. During the famine phase, the stored PHA is used for growth, recovery of glycogen pools and phosphate storage as polyphosphate. Nevertheless, the aerobic operation with acetate or propionate demonstrated that glycogen is slightly consumed. Therefore, the only difference between both cases is the PHA composition: acetate is stored mainly as PHB while propionate produces PHV and PH2MV. This different composition may be related to the stability observed with propionate because PHA composition has been reported as potentially a key aspect of aerobic phosphorus uptake in EBPR [34]. 4. Conclusions The performance of the EBPR process under permanent aerobic conditions using propionate was investigated. The most important conclusions that can be drawn from this study are: • Propionate enriched PAO population developed under standard anaerobic–aerobic cycles was able to maintain biological phosphate removal when it was subjected to strict aerobic conditions during a long period (46 days). Famine P-uptake was achieved in all aerobic cycles resulting in a net P-removal in the system. • The drop observed in the stored glycogen under aerobic operation suggests that propionate oxidation through the TCA cycle can replace the role of glycogen in the generation of reducing equivalents. • The configuration tested shows the feasibility to treat wastewater with simultaneous presence of phosphate and propionate under permanent aerobic conditions. • PAO maintained the EBPR activity during the subsequent operation under standard anaerobic–aerobic conditions. For this purpose, they increased the utilization of glycogen, recovering the typical anaerobic PAO metabolism. • Accumulibacter cells were the dominant population in all periods studied. Its percentage started at 70%, decreased to 50% under aerobic operation and finally recovered to 72% under the anaerobic–aerobic operation. The aerobic decrease was associated with an increase of filamentous bacteria. Acknowledgements This work was supported by the Spanish Ministerio de Educación y Ciencia (CTM2004-02569/TECNO and CTQ2007-61756/PPQ). The authors are members of the GENOCOV group (Grup de Recerca Consolidat de la Generalitat de Catalunya). The Department of Chemical Engineering at UAB is a unit of Biochemical Engineering of the Xarxa de Referència en Biotecnologia de Catalunya (XRB), Generalitat de Catalunya. We also acknowledge the technical support from the “Servei de Microscopia” of the Universitat Autònoma de Barcelona.
295
References [1] T. Mino, V. Arun, Y. Tsuzuki, T. Matsuo, Effect of phosphorus accumulation on acetate metabolism in the biological phosphorus removal, in: R. Ramadori (Ed.), Advanced in Water Pollution Control: Biological phosphate removal from Wastewater, Pergamon Press, Oxford, 1987. [2] P.C. Lemos, L.S. Serafim, M.M. Santos, M.A.M. Reis, H. Santos, Metabolic pathway for propionate utilization by phosphorus accumulating organisms in activated sludge: 13 C labelling and in vivo nuclear magnetic resonance, Appl. Environ. Microbiol. 69 (1) (2003) 241–251. [3] A. Oehmen, R. Zeng, Z. Yuan, J. Keller, Anaerobic metabolism of propionate by polyphosphate-accumulating organisms in enhanced biological phosphorus removal systems, Biotechnol. Bioeng. 91 (1) (2005) 43–53. [4] M. Pijuan, A. Oehmen, J.A. Baeza, C. Casas, Z. Yuan, Characterizing the biochemical activity of full-scale enhanced biological phosphorus removal systems: a comparison with metabolic models, Biotechnol. Bioeng. 99 (1) (2008) 170– 179. [5] C.R. Hood, A.A. Randall, A biochemical hypothesis explaining the response of enhanced biological phosphorus removal biomass to organic substrates, Water Res. 35 (11) (2001) 2758–2766. [6] M. Thomas, P. Wright, L. Blackall, V. Urbain, J. Keller, Optimisation of Nossa BNR plant to improve performance and reduced operating costs, Water Sci. Technol. 47 (12) (2003) 141–148. [7] M. Pijuan, A.M. Saunders, A. Guisasola, J.A. Baeza, C. Casas, L.L. Blackall, Enhanced biological phosphorous removal in a sequencing batch reactor using propionate as the sole carbon source, Biotechnol. Bioeng. 85 (1) (2004) 56–67. [8] Y. Chen, A.A. Randall, T. McCue, The efficiency of enhanced biological phosphorous removal from real wastewater affected by different ratios of acetic to propionic acid, Water Res. 38 (1) (2004) 27–36. [9] H. Lu, A. Oehmen, B. Virdis, J. Keller, Z. Yuan, Obtaining highly enriched cultures of Candidatus Accumulibacter Phosphatis through alternating carbon sources, Water Res. 40 (20) (2006) 3838–3848. [10] H. García Martín, N. Ivanova, V. Kunin, F. Warnecke, K.W. Barry, A.C. McHardy, C. Yeates, S. He, A.A. Salamov, E. Szeto, E. Dalin, N.K. Putnam, H.J. Shaphiro, J.L. Pangilinan, I. Rigoustos, N.C. Kyripides, L.L. Blackall, K.D. McMahon, P. Hugenholtz, Metagenomic analysis of two enhanced biological phosphorus removal (EBPR) communities, Nat. Biotechnol. 24 (10) (2006) 1263–1269. [11] D. Brdjanovic, A. Slamet, M.C.M. van Loosdrecht, C.M. Hooijmans, G.J. Alaerts, J.J. Heijnen, Impact of excessive aeration on biological phosphorus removal from wastewater, Water Res. 32 (1) (1998) 200–208. [12] J. Ahn, T. Daidou, S. Tsuneda, A. Hirata, Transformation of phosphorus and relevant intracellular compounds by a phosphorus-accumulating enrichment culture in the presence of both the electron acceptor and electron donor, Biotechnol. Bioeng. 79 (1) (2002) 83–93. [13] A. Guisasola, M. Pijuan, J.A. Baeza, J. Carrera, C. Casas, J. Lafuente, Aerobic phosphorus release linked to acetate uptake in bio-P sludge: process modelling using oxygen uptake rate, Biotechnol. Bioeng. 85 (7) (2004) 722–723. [14] M. Pijuan, A. Guisasola, J.A. Baeza, J. Carrera, C. Casas, J. Lafuente, Aerobic phosphorus release linked to acetate uptake: influence of PAO intracellular storage compounds, Biochemical Eng. J. 26 (2–3) (2005) 184–190. [15] J.J. Beun, E.V. Verhoef, M.C.M. Van Loosdrecht, J.J. Heijnen, Stoichiometry and kinetics of poly--hydroxybutyrate metabolism under denitrifying conditions in activated sludge cultures, Biotechnol. Bioeng. 68 (5) (2000) 497–507. [16] K. Dircks, M. Henze, M.C.M. van Loosdrecht, H. Mosbaek, H. Aspegren, Storage and degradation of poly--hydroxybutyrate in activated sludge under aerobic conditions, Water Res. 35 (9) (2001) 2277–2285. [17] M. Pijuan, A. Guisasola, J.A. Baeza, J. Carrera, C. Casas, J. Lafuente, Net P-removal deterioration in enriched PAO sludge subjected to permanent aerobic conditions, J. Biotechnol. 123 (1) (2006) 117–126. [18] J. Ahn, S. Schroeder, M. Beer, S. McIlroy, R.C. Bayly, J.W. May, G. Vasiliadis, R. Seviour, Ecology of the microbial community removing phosphate from wastewater under continously aerobic conditions in a sequencing batch reactor, Appl. Environ. Microbiol. 73 (7) (2007) 2257–2270. [19] A. Guisasola, M. Vargas, M. Marcelino, J. Lafuente, C. Casas, J.A. Baeza, Online monitoring of the enhanced biological phosphorus removal process using respirometry and titrimetry, Biochem. Eng. J. 35 (3) (2007) 371–379. [20] G.J.F. Smolders, J. van der Meij, M.C.M. van Loosdrecht, J.J. Heijnen, Model of the anaerobic metabolism of the biological phosphorus removal process: stoichiometry and pH influence, Biotechnol. Bioeng. 43 (6) (1994) 461–470. [21] A. Oehmen, B.K. Lehmann, R.J. Zeng, Z. Yuan, J. Keller, Optimisation of poly-hydroxyalkanoate analysis using gas chromatography for enhanced biological removal systems, J. Chromatogr. A 1070 (1–2) (2005) 131–136. [22] APHA, Standard Methods for the Examination of Water and Wastewater, 19th ed., American Publishers Health Association, 1995. [23] R.I. Amann, W. Ludwig, K. Schleifer, Phylogenetic identification and in situ detection of individual microbial cells without cultivation, Microbiol. Rev. 59 (1) (1995) 143–169. [24] H. Daims, A. Bruhl, R. Amann, K.H. Schleifer, M. Wagner, The domain-specific probe EUB338 is insufficient for the detection of all bacteria: development and evaluation of a more comprehensive probe set, Syst. Appl. Microbiol. 22 (3) (1999) 434–444. [25] G.R. Crocetti, P. Hugenholtz, P.L. Bond, A. Schuler, J. Keller, D. Jenkins, L.L. Blackall, Identification of polyphosphate accumulating organisms and design of 16S rRNA-directed probes for their detection and quantitation, Appl. Environ. Microbiol. 66 (3) (2000) 1175–1182.
296
M. Vargas et al. / Biochemical Engineering Journal 43 (2009) 288–296
[26] G.R. Crocetti, J.F. Banfield, J. Keller, P.L. Bond, L.L. Blackall, Glycogenaccumulating organisms in laboratory-scale and full-scale wastewater treatment processes, Microbiology 148 (11) (2002) 3353–3364. [27] MATLAB, Image Processing Toolbox for MATLAB, User’s Guide, Version 5, 2005, The MathWorks, Natick, MA, USA. [28] A. Oehmen, Z. Yuan, L.L. Blackall, J. Keller, Comparison of acetate and propionate uptake by polyphosphate accumulating organisms and glycogen accumulating organisms, Biotechnol. Bioeng. 91 (2) (2005) 162–168. [29] H. Pereira, P. Lemos, M. Reis, J. Crespo, M. Carrondo, H. Santos, Model for carbon metabolism in biological phosphorus removal processes based on in vivo CNMR labelling experiments, Water Res. 30 (9) (1996) 2128–2138. [30] P. Chudoba, A. Morel, B. Capdeville, The case of both energetic uncoupling and metabolic selection of microorganisms in the osa activated-sludge system, Environ. Technol. 13 (8) (1992) 761–770.
[31] M. Pijuan, J.A. Baeza, C. Casas, J. Lafuente, Response of an EBPR population developed in a SBR with propionate to different carbon sources, Water Sci. Technol. 50 (10) (2004) 131–138. [32] H. Satoh, T. Mino, T. Matsuo, Uptake of organic substrates and accumulation of polyhydroxialcanoates linked with glycolysis of intracellular carbohydrates under anaerobic conditions in the biological excess phosphate removal processes, Water Sci. Technol. 26 (5–6) (1992) 933–942. [33] G. Sin, A. Guisasola, D.J.W. De Pauw, J.A. Baeza, J. Carrera, P.A. Vanrolleghem, A new approach for modelling simultaneous storage and growth processes for activated sludge systems under aerobic conditions, Biotechnol. Bioeng. 92 (2005) 600–613. [34] A.A. Randall, Y.H. Liu, Polyhydroxyalkanoates form potentially a key aspect of aerobic phosphorus uptake in enhanced biological phosphorus removal, Water Res. 36 (14) (2002) 3473–3478.