Cyanobacterial biomass: a striking factor to decrease polyaluminium chloride (PACl) coagulation efficiency during a successive bloom

Lakes and reservoirs are important water resources for urban water supply systems, but global climate change and eutrophication will enhance the outbreak frequency of harmful cyanobacterial blooms in source waters (O'Neil et al. 2012). Recently, ecologists have proposed the concept of ‘a successive bloom’ according to the long-term observation of cyanobacterial blooms in several lakes (Tang et al. 2018; Wilhelm et al. 2020). Generally, cyanobacterial blooms (e.g., in Taihu lake) can be classified into three stages: development, maintenance and decay stages (Tang et al. 2018). From the development to maintenance stage, the cell-density of cyanobacteria is increasing (Tang et al. 2018). The pH values of natural waters also increased, as cyanobacteria can utilize CO₂ as carbon source to biosynthesize organic matter (Visser et al. 2016). For example, Ji et al. (2020) reported that cell-density increased from 10⁵ to 10⁷ cells mL⁻¹, and pH values also increased above 8.0 (max. 9–10) for several months during a successive Microcystis bloom. These results suggested that these changes in cell-density and pH values of cyanobacteria-laden waters may strongly challenge water treatment processes in drinking water treatment plants (DWTPs).

In DWTPs, coagulation/sedimentation is a conventional drinking water treatment process to remove colloidal particles in source waters and it is also an important barrier to prevent the breakthrough of cyanobacterial cells. PACl is the most common coagulant, and it contains the stable preformed tridecameric polymer Al₁₃O₄(OH)₂₄(H₂O)⁷⁺ referred to as Al₁₃ polymer aluminium species that is thought to be more effective at charge neutralization than alum due to a higher charge density (Lin & Ika 2020). Previous studies have demonstrated that PACl coagulation could effectively remove cyanobacterial cells and algal organic matter (AOM) in varying degrees (Tang et al. 2017). However, cyanobacterial cells have average negative charge of −20 to −40 mW and AOM could react with PACl. These characteristics can result in a decrease of PACl coagulation efficiency and an increase of PACl demands (Tang et al. 2017). Over-use of PACl may also cause aluminium contamination in drinking water, leading to chronic toxicity to the human nervous system (Wang et al. 2019). Thus, investigating the effects of pH and cell-density on cyanobacterial removal is quite important for water supplies to optimize PACl application during a successive bloom.

The pre-chlorination process has been widely employed to treat cyanobacteria-laden source waters (Lin et al. 2018). It has been proven to be effective in promoting coagulation efficiency to remove cyanobacteria, since free chlorine can inactivate cyanobacterial cells via disrupting cell structures (Fan et al. 2013; Lin et al. 2018). However, some negative effects have also been reported, that chlorine can induce membrane damage, leading to the release of intracellular organic matter (IOM) (Lin et al. 2018). This released IOM may act as precursors to produce more disinfection by-products (DBPs) after chlorination, and it may hinder cyanobacterial removal by coagulation (Zamyadi et al. 2012; Zhou et al. 2014). Hence, during a successive bloom, whether the pre-chlorination process prior to coagulation is essential to treat cyanobacteria-laden resource waters was worthy of investigation.

To our knowledge, few studies have investigated water treatment processes to remove cyanobacteria during a successive bloom. In this study, we focused on typical characteristics of elevated cell-density and pH values of a successive bloom. PACl coagulation experiments were conducted to investigate cyanobacterial removal with varied pH and cell-density. Then, pre-chlorination prior to the coagulation process was performed to evaluate the overall removal ratio of cyanobacteria via measuring turbidity, chlorophyll-a, phycocyanin and dissolved organic carbon (DOC). This study aimed to provide an important reference for water supplies to choose proper water treatment processes to treat cyanobacteria-laden source waters during a successive bloom.

Microcystis species has been one of the most common and problematic species (Harke et al. 2016), and thus, Microcystis aeruginosa FACHB-915 was employed for conducting the experiments. It was purchased from the Institute of Hydrobiology, Chinese Academy of Sciences. The strain was cultured in BG11 medium at 28 °C with a 12 h:12 h light–dark cycle under light intensity of 35 μmol m⁻¹ s⁻¹ (Li et al. 2020a, 2020b). A culture volume of 1 L was employed with an inoculation volume of 1:10, and these cultures were shaken once at 10:00am on each day. After 14 d cultivation, the Microcystis samples were collected for subsequent experiments.

PACl used in this study was purchased from Tianjin Guangfu Fine Chemical Research Institute (Al₂O₃ content >28%, basicity 70%–75%). Sodium hypochlorite commercial solutions for the chlorination experiments were analytical grade (Sigma, German). Moreover, all solutions were prepared using ultra-pure water purified to a resistivity of 18 MΩ cm by a Milli-Q water purification system (Millipore Pty Ltd, USA).

Ohio EPA have selected chlorophyll-a thresholds of 2, 5, and 50 μg/L (or 4,000, 10,000, and 100,000 cells mL⁻¹) for minor, moderate, and severe blooms, respectively (Ohio EPA 2014). Microcystis cells were collected via centrifugation at 5,000 g for 10 min (extracellular organic matter was removed), resuspended with ddH₂O and achieved a final cell-density of 9.0 × 10⁵ cells mL⁻¹ characterized as severe blooms. During a severe Microcystis bloom, the pH value increased above 8.0 (max. 9–10), and it was about 7.0–8.0 without cyanobacterial bloom in natural freshwaters (Tang et al. 2018; Ji et al. 2020).

To investigate the effects of pH values on Microcystis removal by PACl coagulation, Microcystis samples were adjusted to pH of 7.5, 8.5, and 9.5 using either sodium hydroxide or hydrochloric acid, respectively. For coagulation experiments, 50 mL of cyanobacterial cultures were treated with various dosages of PACl (0, 2.5, 5, 10, 20, 30, 60, and 90 mg L⁻¹) in 200 mL glass conical flasks. Then, it was stirred continuously during the coagulation process (800 rpm min⁻¹, 1 min; 300 rpm min⁻¹, 4 min; 100 rpm min⁻¹, 15 min) with six magnetic agitators, and kept motionless for 40 min. Finally, Microcystis samples were taken for the measurements of turbidity, chlorophyll-a and phycocyanin. The details of the analytical methods are described below. These coagulation experiments were conducted at the same room-temperature (22 ± 2 °C).

During a successive Microcystis bloom, cell-density ranged from 10⁵ to 10⁷ cells mL⁻¹ for several months (Visser et al. 2016). To investigate the effect of cell-density on Microcystis removal by PACl coagulation, these collected Microcystis cells were resuspended with ddH₂O and achieved different cell-densities of about 9.0 × 10⁵, 3.4 × 10⁶, and 1.0 × 10⁷ cells mL⁻¹, respectively. Cell-densities of Microcystis samples were measured by a flow cytometer (FlowSight, Merck Millipore, USA) (Li et al. 2020a). For coagulation experiments, 50 mL of cyanobacterial cultures were treated with various dosages of PACl (0, 2.5, 5, 10, 20, 30, 60, and 90 mg L⁻¹) in 200 mL glass conical flasks. Then, it was stirred continuously during the coagulation process (800 rpm min⁻¹, 1 min; 300 rpm min⁻¹, 4 min; 100 rpm min⁻¹, 15 min) with six magnetic agitators, and kept motionless for 40 min. Finally, Microcystis samples of surface water were taken for the measurements of turbidity, chlorophyll-a and phycocyanin, respectively. These coagulation experiments were conducted at the same room temperature (22 ± 2 °C).

Hypochlorite solutions were prepared from sodium hypochlorite (NaClO) commercial solutions. Free chlorine concentration was measured using the N,N,diethyl-p-phenylenediamine (DPD) method.Microcystis cells were collected, as described in the section ‘Investigating the effects of pH values on cyanobacteria removal via PACl coagulation’. For pre-chlorination experiments, Microcystis samples of 50 mL were treated with various dosages of chlorine (0.5, 1, 2, 4, and 8 mg L⁻¹) in 200 mL glass conical flasks. After a contact time of 60 min, residual chlorine was quenched with sodium thiosulfate (Na₂S₂O₃) and these samples were taken for the measurement of turbidity, chlorophyll-a, phycocyanin, and dissolved organic matter (DOC), respectively. Then, various dosages of PACl (0, 2.5, 5, 10, 20, 30, 60, and 90 mg L⁻¹) were added to the Microcystis samples, as described above in the section ‘Coagulation experiments’. After PACl coagulation, Microcystis samples were taken for the measurements of DOC, turbidity, chlorophyll-a and phycocyanin, respectively.

Both Microcystis cells and algal organic matter (AOM) could contribute to the majority of turbidity in these samples, and thus, turbidity was employed to be the indicator of the removal of both Microcystis cells and AOM. Turbidity was measured with turbidometer (Orion AQ4500, USA). Chlorophyll-a and phycocyanin constituted the photosynthetic apparatus in Microcystis cells, and thus, both chlorophyll-a and phycocyanin were used as the parameters of Microcystis removal. The two parameters were measured using a two-channel fluorometer (AmiScience, USA). Microcystis samples were filtered with a 0.45 μm Millipore filter before the measurement of dissolved organic carbon (DOC) and DOC concentration was measured by the persulfate wet oxidation technique (Shimadzu TOC-V WP, Japan), as also described by Li et al. (2020a, 2020b). Cellular size of Microcystis was measured with a Bettersize2600E laser particle size distribution instrument (Jinke, China). Zeta potential of Microcystis was measured using a multi-angle particle size and high sensitivity zeta potential analyzer (NanoBrook Omni, USA).

Three parallel experiments were conducted. Differences of DOC, turbidity, chlorophyll-a and phycocyanin used Student's t-test, and were considered significant at P < 0.05. All statistical analyses were performed using Origin 8.0.

Figure 1 shows that PACl coagulation with initial dosages of 2.5–30 mg L⁻¹ could effectively remove Microcystis at pH values of 7.5, 8.5 and 9.5. The removal ratio of turbidity, chlorophyll-a and phycocyanin at pH values of 8.5 and 9.5 was much higher than that at pH value of 7.5 with the equal initial dosage of PACl (P < 0.05), and there was no significant difference at pH values of between 8.5 and 9.5 (P > 0.05) (Figure 1). However, the removal ratio of Microcystis decreased with initial high dosages of 30, 60, and 90 mg L⁻¹ (Figure 1).

During a successive Microcystis bloom, pH value increased above 8.0 (max. 9–10) for several months (Visser et al. 2016; Tang et al. 2018), and Figure 1 demonstrates that the removal ratio of Microcystis had no significant difference at pH values of between 8.5 and 9.5 (P > 0.05). Hence, subsequent coagulation experiments for Microcystis samples were conducted at a pH value of 9.5 (Figure 2).

Figure 2 shows that removal ratio of turbidity, chlorophyll-a and phycocyanin was dependent on the initial dosages of PACl, among which the removal ratio was higher for 3.4 × 10⁶ and 1 × 10⁷ cells mL⁻¹ than that for 9 × 10⁵ cells mL⁻¹ with initial PACl dosages of 10–90 mg L⁻¹ (P < 0.05) (Figure 2). However, the improved PACl coagulation could not be attributed to the high cell-density since the removal ratio for 9 × 10⁵ cells L⁻¹ reached up to about 90% with initial PACl dosage of 5 mg L⁻¹ whereas the same treatment for 3.4 × 10⁶ and 1 × 10⁷ cells L⁻¹ was less than 50% (Figure 2). Besides, with the initial PACl dosages of <10 mg L⁻¹, the removal ratio was much lower for high cell-densities of 3.4 × 10⁶ and 1 × 10⁷ cells mL⁻¹ than that of 9 × 10⁵ cells mL⁻¹ (P < 0.05) (Figure 2).

The above study found that initial PACl of 5 mg L⁻¹ only removed 3%–5% of Microcystis with a cell-density of 1.0 × 10⁷ cells mL⁻¹ (Figure 2). To enhance Microcystis removal, pre-chlorination of 0.5, 1, 2, 4 and 8 mg L⁻¹ was employed to treat these Microcystis samples with high cell-density of 1.0 × 10⁷ cells mL⁻¹ at a pH value of 9.5 (Figure 3). Then, PACl coagulation of 5 mg L⁻¹ was further conducted to remove Microcystis cells (Figure 4).

After pre-chlorination, both chlorophyll-a and phycocyanin showed a significant decrease of 30%–50% with initial high dosages of chlorine (2, 4 and 8 mg L⁻¹) whereas about 5%–10% increased with initial low dosages of chlorine (0.5 and 1 mg L⁻¹) (Figure 3). Meanwhile, both turbidity and DOC were increasing by 5%–10% in all treatments, but this was not dependent on initial dosages of chlorine (Figure 3).

PACl coagulation of 5 mg L⁻¹ was further employed to treat chlorine-treated Microcystis samples (Figure 4). Figure 4 shows that the turbidity, DOC, chlorophyll-a and phycocyanin of Microcystis samples decreased after PACl coagulation in varied degrees. However, the removal ratio was chlorine-dosage-dependent, in which the highest removal ratio of 56.2% (turbidity), 41.6% (chlorophyll-a), 51.1% (phycocyanin) and 62.1% (DOC) was observed with pre-chlorination of 1 mg L⁻¹ whereas the initial highest dosage of pre-chlorination (8 mg L⁻¹) achieved the lowest removal ratio of Microcystis cells and DOC (Figure 4).

The data of η₁ and η₂ were gained from Figures 3 and 4, respectively. Figure 5 shows that chlorination-assisted PACl coagulation could achieve RI (about 200%, 400% and 800%) of turbidity, chlorophyll-a and phycocyanin, respectively. This suggested that the pre-chlorination process could improve the removal efficiency of high cell-density Microcystis by PACl coagulation. However, IR of DOC ranged from about 50% to −100%, suggesting that the DOC removal was dependent on initial chlorine dosage (Figure 5). Only low dosages of chlorine (0.5 and 1 mg L⁻¹) promoted DOC removal, while higher dosages resulted in elevated DOC even after PACl coagulation (Figure 5).

Previous studies have noted that pH values played an important role in PACl coagulation, since hydrolysis products of PACl would be varied in the range of 1–10 (Naceradska et al. 2019). Wu et al. (2020) found that the highest coagulation efficiency occurred at pH values of 7–8. This may be explained by the hydrolysis form of PACl being mainly comprised of cationic polymers (e.g., Al₁₃(OH)₃₄⁺⁵) at pH values of 7–8, and these cationic polymers contribute to cyanobacteria removal via electrical neutralization (Lin & Ika 2020). However, this study found that elevated pH values of 8.5–9.5 gained higher removal ratios of Microcystis than the pH value of 7.5. Actually, at pH value >8.0, the hydrolysis form of PACl mainly consists of [Al(OH)₄⁻] and Microcystis cells are also negative (Lin & Ika 2020). This result suggested that electrical neutralization might not be the key mechanism to enhance Microcystis removal by PACl coagulation at pH values of 8.5–9.5.

Tian & Zhao (2021) proposed that the polymerization degree of aluminium salt increases and that it forms a complex structure of hydroxyl polymer at pH value of >8.0. Moreover, a hydrolytic polymerization reaction tends to an increase of polymerization degree, and thus, polymerized aluminium hydroxyl complex ions easily form and eventually change into neutral gelatinous precipitate (Yang 2013). These gelatinous precipitates may act as a net to capture and sweep cells to enhance Microcystis removal at pH values of 8.5–9.5. Overall, during a successive bloom, PACl coagulation to remove cyanobacteria can benefit from the in situ elevated pH of the source waters.

To our knowledge, there is limited literature to investigate the effects of cell-density on Microcystis removal via PACl coagulation during a successive bloom. Henderson et al. (2010) demonstrated that cellular surface area and charge density held a strong positive correlation with coagulant demands. In this study, elevated cell-density of Microcystis exhibited larger total cellular surface area and total charge density, and thus, PACl coagulation efficiency decreased with elevated cyanobacterial biomass. Hence, during a successive bloom, cyanobacterial biomass can be a striking factor to decrease PACl coagulation.

After pre-chlorination, elevated DOC and turbidity indicated that chlorine had disrupted the cellular membrane, leading to the release of intracellular organic matter (IOM) (Figure 3; Supplementary Information, Figure S1). A similar result has also been documented by previous studies (Zamyadi et al. 2012; Zhou et al. 2014). Notably, elevated DOC/turbidity had no significantly positive correlation with initial dosages of chlorination, and a higher initial dosage of chlorine (2, 4, and 8 mg L⁻¹) did not result in a higher DOC concentration (Figure 3). This result suggested that the released IOM may be oxidized into CO₂ after dosing with sufficient chlorine exposure.

This study found that pre-chlorination enhanced Microcystis removal by PACl coagulation (Figures 4 and 5), in agreement with previous studies (Xie et al. 2016). However, associated mechanisms may differ with initial dosages of chlorine. Initial low dosages of chlorine (0.5 and 1 mg L⁻¹) did not inactivate Microcystis due to a slight increase of chlorophyll-a and phycocyanin (Figure 3). Ma et al. (2012) noted that a high molecular weight of AOM aided coagulation by favoring the formation of larger flocs. In these treatments with chlorine (0.5 and 1 mg L⁻¹), the released high molecular weight (MW) of IOM could be the main mechanism to enhance Microcystis removal. In contrast, higher dosages of chlorine (2, 4 and 8 mg L⁻¹) effectively inactivated Microcystis via destroying chlorophyll-a and phycocyanin (Figure 3). This result suggested that the activation of Microcystis could be the main mechanism to enhance Microcystis removal by PACl coagulation.

Moreover, after PACl post-coagulation, Microcystis removal was not dependent on initial chlorine dosage, in which initial high dosage of chlorine (2, 4, and 8 mg L⁻¹) exhibited a lower removal ratio of Microcystis and DOC than initial dosages of 0.5 and 1 mg L⁻¹(P < 0.05) (Figure 4). Safarikova et al. (2013) found that low-MW proteins exhibited stronger inhibitory effects on the coagulation process than high-MW proteins. This suggested that the possible formation of low-MW organic matter after chlorination (2, 4, and 8 mg L⁻¹) strongly decreases the removal of Microcystis. Meanwhile, this low-MW organic matter was also difficult to remove by PACl coagulation, leading to an increase of DOC. In contrast, initial low dosage of chlorine (0.5 and 1 mg l⁻¹) mainly induced the release of high-MW IOM, and this IOM could not be oxidized into low-MW IOM due to insufficient chlorine exposure. This high-MW IOM aided Microcystis removal, and was easily removed by PACl coagulation, leading to an decrease of DOC. Moreover, the removal ratio of Microcystis that occurred with pre-chlorination (0.5 mg L⁻¹) was lower than that with the dosage of 1.0 mg L⁻¹, possibly attributable to its lower amounts of released high-MW IOM. Overall, moderate pre-chlorination (1 mg L⁻¹) could be a promising option to enhance the removal of high cell-density Microcystis (10⁷ cells mL⁻¹) by PACl coagulation.

The concept of ‘a successive bloom’ is essential for water supplies to determine proper water treatments to treat cyanobacteria-laden source waters, but it was always ignored by previous studies. During a successive Microcystis bloom, changes in cell-density and pH value were significant characteristics of source waters. This study suggested that Microcystis removal by PACl coagulation could benefit from in situ elevated pH (>8.5) of source waters. However, elevated cyanobacterial biomass strongly hindered PACl coagulation and the inhibitory effects on PACl coagulation could not be offset by in situ elevated pH value. These results demonstrated that cyanobacterial biomass could be a more important factor to affect PACl coagulation than elevated pH value. Thus, it is quite important for water supplies to monitor cyanobacterial biomass in real-time to determine the effective drinking water treatment process during a successive bloom.

To our knowledge, whether pre-chlorination was essential to treat cyanobacteria-laden source waters was a heated argument, attributable to its apparent advantages and drawbacks (e.g., severe membrane destruction). During a successive bloom, this study proposed that the application of pre-chlorination was mainly dependent on cyanobacterial biomass. For high cell-density of >10⁷ cells mL⁻¹, moderate pre-chlorination (1 mg L⁻¹) was essential to promote PACl coagulation efficiency to remove Microcystis. For low cell-density of 10⁵–10⁶ cells mL⁻¹, coagulation alone was sufficient to remove Microcystis, since it can benefit from in situ elevated pH value of source waters. These results provide a useful reference for water supplies to choose the proper water treatment process to treat cyanobacteria-laden source waters during a successive bloom.

Although this study demonstrated that moderate chlorination (1 mg L⁻¹) eliminated the risk of elevated DOC after PACl coagulation, other undesirable metabolites (e.g., cyanotoxins; taste/odor compounds) would be released after cell rupture (Zamyadi et al. 2012; Fan et al. 2013; Li et al. 2020a, 2020b). Notably, disinfection by-products would be another troubling problem after chlorination, since previous studies found that AOM could be important precursors to form DBPs after chlorination (Zamyadi et al. 2012; Zhou et al. 2014; Wu et al. 2019). Consequently, more studies are required to systematically assess the advantages and drawbacks of moderate chlorination-assisted PACl coagulation before it is employed to remove a high cell-density of Microcystis.

During a successive bloom, Microcystis removal via PACl coagulation can benefit from in situ elevated pH (>8.5), and thus, coagulation alone is recommended to treat cyanobacteria-laden source waters with low cell-density of 10⁵–10⁶ cells mL⁻¹. However, elevated cyanobacterial biomass was a striking factor in decreasing cyanobacterial removal by PACl coagulation, since its inhibitory effects on coagulation could not be offset by in situ elevated pH value. Moderate chlorination-assisted (1 mg L⁻¹) coagulation is recommended to treat cyanobacteria-laden source waters with high cell-density of >10⁷ cells mL⁻¹, but more studies should be done to assess the advantages and drawbacks for removing cyanobacteria with high cell-density in the future.

This work was supported by the National Training Program of Innovation and Entrepreneurship for Undergraduates (201910397006) and Fujian Natural Science Foundation Project (2020J01417). Special thanks for a scholarship provided by the University of Chinese Academy of Sciences (UCAS).

Weijun Song: Investigation, Writing/revising manuscript.

Yu Xie: Data Curation, Revising manuscript.

Xunfang Wu: Investigation, Revising manuscript.

Jie Zeng: Revising manuscript, Polishing language.

Xi Li: Experimental design, Writing/revising manuscript, Conceptualization, Project management.

Data cannot be made publicly available; readers should contact the corresponding author for details.