The effect of QQ on the MBR performance at macroscopic scale is discussed here in terms of TMP, characteristics of the mixed liquor, and biodegradation efficiencies.
The TMP was chosen because it is without any doubt the most monitored property in studies of biofouling in MBRs, most likely because it is an excellent indicator of biofouling and it can be easily measured continuously during the MBR operation, providing key information on the biofouling kinetics. For these reasons, all the studies reporting the effectiveness of QQ to mitigate biofouling are at least partly based on the comparison of the TMP profiles. Before giving numerical results, we illustrate the general shape of the TMP profile in MBRs, then the possible modifications resulting from the implementation of bacterial QQ and the information that can be deduced from the comparison of the two shapes.
presents a schematic illustration of the TMP profiles during MBR operation at constant flux. The fouling phenomenon is now recognized by the MBR community as a three-stage process. Initially, a short-term increase in TMP is observed, from point A to point B, and can probably be attributed to initial pore blocking and the adsorption of solutes from the mixed liquor. Next, a slow long-term rise takes place between points B and C, and is due to the progressive deposition of biosolids (cells and microbial flocs) on the membrane surface and the progressive formation of biofilm. Finally, at point C, called the breaking point, a striking increase, called the TMP jump, indicates a severe loss of permeability [53
]. The occurrence of the TMP jump depends on the operating conditions and is believed to be caused by sudden changes in local flux or in the biocake architecture and EPS composition [55
]. Once point D is reached, the filtration process is stopped and the membrane is either cleaned with a view to reusing it, or changed. The time at which point D is reached is defined as a cycle, and several cycles usually take place during an MBR run.
With the implementation of QQ bacteria to reduce biofouling in MBRs, the second stage of the TMP rise-up is expected to be significantly slowed down from point B to point C’, and the occurrence of the TMP jump (at the breaking point) would thus be delayed (Figure 7
). Moreover, the TMP jump from point C’ to point D’ could be expected to be attenuated, indicating that the addition of QQ bacteria could lead to modified biofouling behavior with temporal and spatial variations in both the progressive biofouling stage and the TMP jump stage. Nevertheless, one particular case could be that the TMP jump in the QQ MBR (C’D’ section in Figure 7
) was parallel to that in the control MBR (CD section in Figure 7
), which could imply that QQ only slows down the progressive biofouling stage with no major changes in the TMP jump stage.
In the studies reported in Table 1
and Table 2
, the initial TMP did not exceed 10 kPa (it ranged from approximately 3 to 10 kPa). For all these studies, it was possible to identify the kind of TMP evolution presented in Figure 7
. However, the first short-term increase stage (AB section in Figure 7
), which takes place relatively quickly, was not clearly identifiable, especially for the studies carried out in the long-term operation.
Hence, the analysis of the TMP profiles with the latter approach, combined with the characterization of the mixed liquor and the biodegradation efficiencies, are expected to provide information about the direct effects of QQ on the physical phenomena in the MBR at macroscopic scale.
8.1. Effect on the Progressive Biofouling Stage
For the assessment of the bacterial QQ effect on the progressive biofouling stage, it may be interesting to take into account the vertical gap between the BC and BC’ sections at the arbitrary time of 1 day (Figure 7
). This time has been chosen because it is the time by which the steady progressive biofouling stage was set in all the studies considered in Table 1
and Table 2
. This gap would then represent the TMP reduction induced by QQ after 1 day of MBR operation.
After 1 day of operation, a significantly reduced TMP value was noted (references n° 1 to 3 and n° 7 to 12 in Table 1
) with more than 50% reduction. On the other hand, for the studies monitoring longer operation times (references n° 4 to 6 and n° 13 and 14 in Table 1
), the TMP reduction was less pronounced, with less than 25% reduction. With these results, it appears that Rhodococcus
sp. BH4-mediated QQ effectively slows down the progressive biofouling stage, which indicates that Rhodococcus
sp. BH4 probably expresses its QQ activity from the very early phase of the MBR operation (during the first day). However, this effect is less visible for longer operation times.
sp. 1A1, the TMPs measured in QQ-MBRs were between 15% and 25% lower than that in control-MBRs, after 1 day of operation (references n°1 to 4 in Table 2
). These values indicate that Pseudomonas
sp. 1A1-mediated QQ has a marked effect on the progressive biofouling stage, which means that the Pseudomonas
sp. 1A1 QQ activity starts relatively early in the MBR operation, slowing down the progressive biofouling stage. Nevertheless, it seems that this effect is less pronounced than that of Rhodococcus
8.2. Effect on the TMP Jump
For the evaluation of the Rhodococcus
sp. BH4-mediated QQ effect on the TMP jump, two criteria have been chosen. The first criterion is the comparison of the times t and t’ corresponding to the times for reaching the breaking points C and C’, respectively (Figure 7
), which provide a quantified information about the TMP jump postponement. The second criterion is the comparison of the slopes of the sections CD and C’D’ which were calculated with the points at 25 kPa and 40 kPa that were found to belong to the TMP jump stage for long-term operation.
The analysis of the breaking points according to the first criterion reveals that the time at which the TMP jump occurs is successfully delayed with the implementation of Rhodococcus
sp. BH4-mediated QQ, by at least 240% (corresponding approximately to a threefold postponement) [49
] (references n° 3 to 6 and n° 13 and 14 in Table 1
For the second criterion, only the long-term MBR operations with running times ranging from 17 to 90 days were taken into account (references n° 5, 6 and 13 in Table 1
) since the filtration in shorter operations is usually stopped before (or right after) the breaking point is reached. The time delays to reach the two points of 25 kPa and 40 kPa with QQ were approximately the same for Kim et al. [25
] (references n° 5 and 13 in Table 1
), which refers to similar slopes (around 160 kPa/day for Kim et al. [25
] and 25 kPa/day for Kim et al. [50
]) for the sections CD and C’D’ (Figure 7
). This indicates that Rhodococcus
sp. BH4-mediated QQ seems to have no effect on the TMP jump stage in these two cases. In contrast, for Lee et al. [26
] and Maqbool et al. [27
] (references n° 6 and 14 in Table 1
), the point of 40 kPa was not actually reached and a less pronounced TMP jump was recorded (data not shown), which indicates a substantial effect of Rhodococcus
sp. BH4-mediated QQ on the TMP jump.
Studying the effect of Rhodococcus sp. BH4-mediated QQ on the TMP jump in addition to its effect on the progressive biofouling stage may give interesting insight into how the QQ effectiveness might evolve over time in the MBR. According to these results, it seems that Rhodococcus sp. BH4-mediated QQ tends to be globally more effective in inducing modifications of the TMP profile during the first progressive biofouling stage than during the TMP fast jump stage, which could be attributed to a potential loss of the QQ effectiveness. Another potential explanation could be that another type of QS-controlled biofouling (e.g., AI-2-controlled QS for interspecies communication), against which the Rhodococcus sp. BH4 cells have no effect, becomes predominant over the AHL-controlled one. However, further research would be needed to clarify these assumptions. All these results taken together with the number of cycles of the control MBRs compared to the QQ MBRs clearly indicate that the implementation of Rhodococcus sp. BH4-mediated QQ substantially reduces the membrane cleaning frequency.
Cheong et al. [38
] investigated the effectiveness of Pseudomonas
sp. 1A1 to mitigate biofouling in MBRs for running times ranging from 6 to 13 days. Concerning the comparison of the times t and t’ at which the breaking points C and C’ were reached for the control and the QQ MBR, respectively (Figure 7
), the time elapsing before the TMP jump was observed to increase by 180% (corresponding to an almost threefold postponement) (reference n° 1 in Table 2
). When comparing the slopes corresponding to the TMP jump sections CD and C’D’ (Figure 7
), the times necessary to attain the pressures of 25 kPa and 40 kPa were both seen to be delayed by approximately 3 days (reference n° 1 in Table 2
). Although these results reveal that Pseudomonas
sp. 1A1-mediated QQ does indeed have an effect on biofouling, more research is needed to unravel how Pseudomonas
sp. 1A1 mitigates biofouling in MBRs and how its effectiveness evolves over time.
8.3. Effect on the Mixed Liquor Characteristics
It is important to evaluate the effect of QQ on the mixed liquor characteristics, particularly in terms of EPS and AHL amounts, since these are good indicators of biofouling.
EPS are well-known to be closely related to biofouling in MBRs since they are the “glue” that holds the biofilm cell clusters attached to the membrane [2
]. In other words, a noticeable increase in the amount of EPS in the MBR is correlated with a heightened biofouling phenomenon. The total quantity of EPS in the mixed liquor can be divided into soluble microbial products (SMP) and EPS bound to the microbial flocs. In principle, these fractions need to be collected separately to be further analyzed (for more details on EPS extraction methods, see Domínguez et al. [61
]). With the implementation of Rhodococcus
sp. BH4 as a QQ bacterium, Maqbool et al. [27
] determined the amount of SMP in the mixed liquor by analyzing the supernatant from an AS sample (reference n° 6 in Table 1
). A 90% reduction in the SMP amount was recorded after 80 days of operation, indicating that Rhodococcus
sp. BH4-mediated QQ had a strong effect on the EPS production. In addition, Lee et al. [26
] recorded 52% and 85% reductions in the amounts of proteins and polysaccharides in the mixed liquor, respectively (reference n° 14 in Table 1
). It is important to quantify this effect since it has been shown that, in some cases, the solutes and colloids in the supernatant play a more important role in biofouling than the biological pellets [62
]. Nevertheless, it is still worth mentioning that slight reductions in the amounts of bound EPS were obtained with the application of QQ, with −32% and −5% in the loosely-bound EPS (LB-EPS) and the tightly-bound EPS, respectively [27
]. Lee et al. [26
] also recorded an average reduction of 17% in the amount of bound EPS in mixed liquor.
Concerning the amount of AHL in the mixed liquor, given that these signal molecules are usually produced at very low concentrations (in the range of picograms to nanograms per liter) and that they are present as a complex mixture with different compounds, an extraction procedure is necessary before the quantification [64
]. Several quantification methods to measure the AHL concentration after their extraction have been reported to date and are summarized by Siddiqui et al. [5
]. Maqbool et al. [27
] extracted the AHLs from the supernatant of a broth sample then analyzed them using HPLC. A much smaller AHL concentration was observed in the QQ-MBR than in the control MBR (qualitative results). Hence, Maqbool et al. [27
] came to the conclusion that the implementation of Rhodococcus
sp. BH4 leads to a biofouling reduction via the destruction of AHLs in the mixed liquor, which is consistent with previous studies [24
]. This result shows that monitoring the AHL concentration in the mixed liquor could help evaluate the progress of the QQ activity in MBRs. However, no information about the evolution of this concentration during the MBR operation is provided in the studies noted in Table 1
, probably because of the very low amounts, which would make the quantification laborious.
Recently, Lee et al. [26
] showed that QQ could have an effect on the floc size of the AS. In a pilot-scale MBR of 80 L, they recorded a 17% reduction in the average floc size. However, in a three-stage MBR of a total working volume of 450 L, composed of three tanks of 150 L (anoxic, aerobic and membrane tank), QQ did not lead to major differences in floc size in each tank. Thus, there is no clear trend yet about the influence of Rhodococcus
sp. BH4-mediated QQ on the floc size of the mixed liquor.
Finally, additional research into the effect of QQ on other properties such as zeta potential, viscosity, or Sludge Volume Index (SVI) is required to provide complementary data to understand exactly how Rhodococcus sp. BH4 affects the mixed liquor characteristics.
The effect of Pseudomonas
sp. 1A1 on the sludge characteristics in an MBR has only been assessed in terms of SMP. Cheong et al. [52
] have reported a 60% reduction in the amount of polysaccharides, whereas the reduction in proteins was merely 6% (references n° 4 and 5 in Table 2
). These results suggest that Pseudomonas
sp. 1A1-mediated QQ targets the QS-controlled genes in charge of the production of polysaccharides in a more pronounced way. However, at the current stage and with the few elements known so far, it is still hard to unravel the effect of Pseudomonas
sp. 1A1-mediated QQ on the mixed liquor characteristics. Nevertheless, it is worth mentioning that some studies have investigated the QQ potential of a commercial purified acylase (porcine kidney I) that is believed to have the same mode of action as the acylase from Pseudomonas
sp. 1A1. Yeon et al. [37
] have reported reductions in the protein and the polysaccharide concentrations in the mixed liquor of approximately 60% and 20%, respectively, and Jiang et al. [34
] found a 20% reduction in both the protein and the polysaccharide concentrations. Thus, there is no clear trend revealing the kind of biofilm-related genes that are affected when acylase-mediated QQ is used.
Concerning the physical characteristics of the mixed liquor, the application of the purified acylase mentioned above resulted in a lower SVI, apparent viscosity and mean particle size, and a higher zeta potential, suggesting a better filterability [34
]. More investigations should be carried out though, to confirm these observations and to investigate exactly how Pseudomonas
sp. 1A1-mediated QQ affects the mixed liquor characteristics.
8.4. Effect on the Biodegradation Efficiencies in the MBR
The effect on the MBR performance should obviously be evaluated to make sure the implementation of Rhodococcus
sp. BH4-mediated QQ does not impair the MBR treatment capacities. For nine of the studies considered in Table 1
(references n° 5 to 13), this was assessed in terms of Chemical Oxygen Demand (COD) removal efficiencies and resulted in very negligible variations (less than a 3% modification compared to the control MBR) [25
]. The fact that the COD removal efficiencies remained practically unchanged indicates that the use of Rhodococcus
sp. BH4 as QQ bacteria induces no adverse effect on the ability of the biomass to metabolize the organic matter in the MBR. The same approach was used to assess Total Kjeldahl Nitrogen (TKN) removal and resulted in less than a 5% variation compared to the control MBR, when the Rhodococcus
sp. BH4 cells were entrapped in CEBs, microbial vessels or RMCF (data not shown) [51
]. For Lee et al. [26
], neither the total nitrogen removal efficiency nor the ammonia-nitrogen (NH4-N) removal efficiency was significantly affected when Rhodococcus
sp. BH4-meadiated QQ was used in a pilot-scale MBR. Therefore, these findings confirm that Rhodococcus
sp. BH4-mediated QQ effectively mitigates biofouling without affecting the MBR treatment performance.
The implementation of Pseudomonas
sp. 1A1 to mitigate biofouling in an MBR resulted in COD removal efficiencies exceeding 95%, which is not significantly different compared to that of the control MBR [38
]. Thus, Pseudomonas
sp. 1A1 is believed to induce no effect on the degradation of organics in MBRs. However, no information is available yet in the literature concerning the nitrogen removal efficiency.