Characterization of particles and their relation with residual aluminum in water treated with pulsating floc blanket clarifiers and conventional clariflocculators using PACl

Conventionally, physico-chemical unit processes like coagulation, flocculation and clarification have been carried out in separate basins in series. Here, the clarification process is effected mostly by sedimentation under gravity alone. Apart from the space requirements, conventional clariflocculators (CCs) employ mechanical components at various stages, adding to complexity as well as running and maintenance costs (Dhabadgaonkar 2008). CCs have also been struggling to meet the drinking water regulatory requirements, especially in low turbidity (<10 NTU) and low alkalinity (<50 mg/L as CaCO₃) conditions as their dominant mechanism for particle removal; namely, sweep floc action suffers adversely (Cheng et al. 2008; Edzwald 2020; Srivastava et al. 2020). A zone of densely concentrated suspension of flocculated particles in fluidized state, or in other words, a floc blanket, can be advantageously used in water treatment, since the condition of low influent turbidity is compensated by it. Though floc blanket clarification is one of the principal processes, many of its aspects are less understood (Hendricks 2006). For the purpose, Srivastava et al. (2020) had described the pilot scale experimental setup for comparatively evaluating the upflow pulsating floc blanket technology and conventional upflow clariflocculation.

The destabilization power of aluminum-based coagulants (due to trivalent Al³⁺ ions) is multiple times more than that of coagulants which furnish monovalent/divalent ions and given their cost effectiveness, are widely used in water treatment facilities. Amongst these, poly aluminum chloride (PACl) has been popularly used in treatment of low turbidity water owing to speciation, stability, transformation and mechanistic advantages primarily related to relatively abundant polymeric aluminum species like ‘Keggin’ structured cation [Al₁₃O₄(OH)]_n⁷⁺ (Wang et al. 2004; Casey 2006; Lv et al. 2021). However, optimal utilization of such coagulants is crucial to address the serious issue of residual aluminum in treated water which has been implicated as a potential neuro-toxin (Edzwald 2020; Exley & Clarkson 2020). Thus according to Indian regulations (BIS 2012), the residual aluminum levels should ideally be below 30 parts per billion (ppb) in drinking water, while the maximum permissible is 200 ppb in the absence of an alternative source. Even among the family of PACls, the percentage of monomeric, polymeric and colloidal pre-formed aluminum species can vary considerably. The behavior of these forms is affected differently by factors like pH and consequently performance parameters like turbidity removal efficiency, floc properties, and particulate and dissolved fractions of residual aluminum also vary. Hence, much contemporary research is focused towards tweaking the basicity, speciation and other properties of PACls for performance improvement (Kimura et al. 2013; Shu-xuana et al. 2014; Lin & Ika 2019a, 2019b; Solanki et al. 2020). However, the physico-chemical significance of the role the reaction environment provided by a reactor may play on the efficiency of a given coagulant like PACl for the provided control set of influent conditions has not been studied in much detail, and therefore the same constitutes the core subject matter of this paper.

Srivastava et al. (2020) had shown the existence of a strong direct correlation between the turbidity of clarified water and residual aluminum. The results also showed a weak negative correlation between influent turbidity and residual aluminum, but the importance of higher influent turbidity was less for pulsating floc blanket clarifier (PFBC) compared to CC since the blanket provides that environment. Coagulation-flocculation behavior exhibited under the two fundamentally different treatment processes influences inter-related parameters like particle aggregation and destabilization; surface adsorption and coverage; floc characteristics; microfloc formation and kinetics; and finally, the performance of subsequent unit operations. The ensuing floc structure, size, fractal dimensions, density and morphology not only play an important role in determining settleability and hence removal efficiency of flocs, but also give key indications about the efficiency of water treatment often measured in terms of parameters like turbidity and concentration of residuals (Yu et al. 2011; Wang et al. 2016; Tsai et al. 2020; Lv et al. 2021). Moreover, since particulate aluminum species can play an important role in elevating total residual aluminum levels (Driscoll & Letterman 1995), it is of interest to evaluate the role a floc blanket may have on the properties of fine colloidal suspensions. The inquiry was extended to examination of the distinguishing charactersitics of flocs formed by the process of pulsating floc blanket clarification and conventional clariflocculation. The study helps understand the solid separation mechanism in a PFBC, explicates the fundamental differences between the two processes, and evaluates the effect of associated physical aspects on residual aluminum.

Continuous flow pilot plants based on conventional upflow clariflocculation and upflow pulsating floc blanket clarification technologies were designed and fabricated for a capacity to treat about 8,000 litres per day (LPD) and are presently installed in the Hydraulics lab of Malaviya National Institute of Technology Jaipur, India. In order to understand the mechanisms of particle removal in the two processes and comparatively assess their performance, the two pilot plants were run in parallel with identical operational conditions as depicted schematically in the combined setup (Figure 1).

Technical aspects of both pilot plants and description of functions served by different design elements have been elaborated upon elsewhere (Srivastava et al. 2020). An inline helical mixer was incorporated before admittance into the PFBC for better rapid mixing within a closed configuration. The inline helical mixer's function for PFBC (Figure 1(b)) was akin to that of the rapid mixer unit with motorized stirrer employed for CC (Figure 1(a)) and for maintaining similar control conditions, the Gt value attained with both for the given flow rate was kept equivalent at ≈ 18,000. The received actuator-solenoid valve system cycle was pre-calibrated according to the optimized values employed at full-scale PFBC facilities. A cycle time of 55 seconds (s) with ‘on’ time of 10 s and ‘off’ time of 45 s was used in the experiments without alteration for ready extrapolation of the results of the study to its field performance. The pulsed flow and continuous flow together helped maintain a homogeneously suspended blanket for a continuous supply of treated water. The designed average flow rate of 3.745 LPM was maintained throughout the experimental phase in both the pilot plants. Steady state is generally reached after about 15 times the residence time has elapsed. Thus, samples of clarified water were collected after the setup was run continuously for 6 hours to represent steady state average values.

Unlike CCs, the foremost requirement for operating most floc blanket clarifiers (FBCs), and especially PFBCs, is the formation of a stable floc blanket. After establishment of the blanket to the desired height, routine water treatment operational runs may be carried out efficiently. The initial floc blanket has been reported to form optimally in the PFBC with continuous dosing of 6.296 mg L⁻¹ PACl for an influent kaolin turbidity of 500 NTU (Hurst et al. 2017), which was followed in the present study until the blanket reached the height of 1.2 m. Assessment of performance under various influent turbidity conditions was then carried out. A series of jar tests was performed to get the respective dose of coagulant for different influent turbidity values, for which the minimum elemental Al influx was able to bring residual turbidity below 0.2 NTU. PACl doses of 0.671, 0.923, 0.671, 0.588 and 1.007 mg L⁻¹ elemental Al for influent turbidity of 2, 5, 8, 10 and 20 NTU, respectively, were used for treating SRW with identical properties on both PFBC and CC.

For imparting turbidity to the influent stream of raw water (for preparation of SRW), a stock kaolinite solution of known strength was mixed thoroughly at a set rate with a given flow rate of raw water being admitted to the pilot plants. Kaolin (Loba Chemie Pvt. Ltd, Mumbai, India) was used to effectively simulate natural turbid surface waters (Hurst et al. 2017). Polyaluminum chloride (PACl) (Aditya Birla Chemicals, Grasim Industries Ltd, Nagda, Madhya Pradesh, India) was used as coagulant for treatment of different SRW in both pilot plants. Specifications of the coagulant used in the study and the characteristics of raw tap water supplied at Malaviya National Institute of Technology Jaipur, India, have been given in Table 1. Stock aluminum solution for calibration in the testing of residual aluminum was prepared by dissolving 8.791 g aluminum potassium sulphate, AlK(SO₄)₂.12H₂O, (Merck KGaA, Darmstadt, Germany) in water and diluting it to 1 L.

Measurement of turbidity was done in accordance with the 2,130 B Nephelometric Method (APHA 2012) on a portable digital nephelometer (TU-2016, Lutron Electronic Enterprise Co. Ltd, Taipei, Taiwan) having a resolution of 0.01 NTU and 1 NTU for different ranges. All samples were tested for turbidity soon after collection to prevent changes in temperature and physical characteristics. Electrical conductivity, Total Dissolved Solids (TDS) and pH were examined using portable instrument HI-98130™ (Hanna Instruments, Rhode Island, USA).

The analysis for residual aluminum was carried out in accordance with the 3500-Al B Eriochrome Cyanine R Method (APHA 2012) on the spectrophotometer (UV 1800, Shimadzu Corp., Kyoto, Japan) at a wavelength of 535 nm using a 1 cm path length glass cuvette. Total and dissolved residual aluminum concentrations were measured in accordance with the methodology described by Van Benschoten & Edzwald (1990). To measure total residual aluminum; solubilization was done with the help of trace metal grade concentrated HNO₃ (Fisher Scientific, Massachusetts, USA). Undigested clarified water samples were first passed through 0.22 μm pore size polymeric membrane filter with the help of a vacuum pump. The filtrate so obtained was then tested for Al after acidification according to the procedure described above to get the dissolved residual aluminum. The difference between total and dissolved residual aluminum gives the particulate fraction. All reagents used in the process were prepared or stored in polypropylene ware instead of glassware to avoid any error in trace aluminium determination. Glass cuvettes and wares used in the determination step were rinsed with 1 + 1 reagent grade HNO₃ (Fisher Scientific, Massachusetts, USA) and deionized water.

Microscopic analysis was used for determining the size of flocs, which are larger and amenable to settle over time. A Zetasizer^® Nano ZS instrument (Malvern Instruments, Malvern, UK) utilizing the dynamic light scattering (DLS) technique was employed to measure zeta potential and size of fine suspended colloidal particles in the clarified water. It may be noted that the particle sizes, zeta potential and other properties were measured for the clarified water directly emanating from the stand-alone pilot-scale PFBC and CC units, which implies that the size of particles should be larger inside the respective reactors at different depths/zones and also that the results are expected to indicate the effect on the performance of subsequent unit operations like filtration.

A modified approach was followed, since sample collection and analysis of flocs amenable to breakage/aggregation is crucial. Therefore, samples were carefully collected and transferred to a slide, with the help of a large opening diameter (6 mm) micropipette to limit floc breakage, and analysed immediately under a microscope to limit aggregation (Lapointe & Barbeau 2016). Brilliant green for microscopy (Merck Millipore, Merck KGaA, Darmstadt, Germany) was used for staining flocs prior to analysis to enable better visualization of flocs and floc boundaries, and distinguish the included matrix (water) volume from the floc. CX21i Microscope (Olympus Corp., Tokyo, Japan) and digital camera assembly, Magcam HD L (Magnus Opto Systems India Pvt. Ltd, New Delhi, India) using 1/2.5’ CMOS image sensor was used to capture high resolution images (2,464 × 1,632 pixels) at ten times optical magnification (10× optical). Magvision^® software (Magnus Opto Systems India Pvt. Ltd, New Delhi, India) was used for layering, scaling and primary image analysis. Further image analysis, including particle size analysis (PSA) was done with the help of open source software ImageJ distributed by Fiji (Schindelin et al. 2015). This method allowed capture of stabilized high-resolution images since the flocs were not in motion and the problem of masked/out-of-focus flocs was minimal.

The more densely packed a structure, the higher the magnitude of its fractal dimension and it approaches the Euclidean geometrical value (for a coalesced sphere) of 2 in two dimensions and 3 in three dimensions. Different coagulation mechanisms ensuing from either variation in type of coagulant or conditions of water chemistry should leave a characteristic impact on the formed floc, which may then be elucidated by its fractal dimension.

Based on the second order moments of the two-dimensional floc image, an ellipse was fitted for each floc. The computed major axis of the fitted ellipse was used to represent the characteristic length (l) of the flocs. Since flocs were found to be branched and elongated, the three-dimensional structure of the floc was assumed to be an ellipsoid obtained by rotating the fitted ellipse about its major axis (Chakraborti et al. 2007).

On similar lines, after taking log on both sides of the above relation (Equation (4)), the ratio of settling velocity for flocs from PFBC and CC was subsequently computed. The advantage of this method was that the necessity for measurement of several physical parameters could be bypassed.

All statistical analyses reported in this study were done for a confidence interval of 95%.

For different influent conditions, respective optimised equivalent coagulant doses were administered to both pilot plants and the variation in zeta potential (ξ) and residual turbidity has been depicted in Figure 2(a). Average value of ξ was lower in magnitude for PFBC (| − 5.1|) than CC (| − 9.6|), indicating comparatively better destabilization for identical control conditions. The turbidity remaining in the clarified water by PFBC was 0.42 NTU on an average with standard deviation (σ) of ±0.4, whereas for CC, an average residual turbidity of 1.61 NTU (σ = ±0.3) was observed. The results indicate that the fluidized zone of high solids concentration provided by the contracting and expanding blanket of flocs in a PFBC improved destabilization of colloidal particles, perhaps by contact flocculation (Hendricks 2006) and charge-neutralization (Chen et al. 2016; Lin & Ika 2019b). As the freshly formed microflocs sift upwards through the floc blanket, constituted primarily of well-grown large flocs, collisions between these disparately sized particles occur, resulting in adsorption like flocculation caused by entrapment due to bridging and charge-neutralization mechanisms.

Multiple linear regression (MLR) analysis for total residual aluminum against ξ and residual turbidity yielded a non-significant correlation with ξ (p-value of 0.26 > α, where α = 0.05). However, a significant (p-value = 0.038) direct (indicated by positive b-value of 0.125) correlation between total residual aluminum and clarified water turbidity was obtained, which corroborates the existence of a strong direct link between the two parameters as reported in previous studies (Srivastava et al. 2020).

The results from PSA by DLS (Figure 2(b)) show that the average diameter of fine colloidal particles suspended in the clarified water from PFBC for different influent conditions was larger, with an average of 1,648.2 nm. Average size of colloidal particles in suspension for CC was found to be 941.9 nm, which is below 1,000 nm mark. Average polydispersity index (PdI) value of about 0.7 was recorded for samples from PFBC and CC, indicating a condition of hetero-dispersity. These results support that the blanket also abets nucleation by providing surface sites, which in turn affects floc growth/re-growth at nanoscale (Yu et al. 2011). Also, this accords well with the comparative results on destabilization by the two technologies and indicates that the pulsating action was aiding flocculation, starting at the micro levels.

This is of significance because it is expected that particle aggregation during rapid mixing and flocculation should provide the major reduction in submicron (<1 μm or 1,000 nm) colloid numbers, since the subsequent filtration unit process is less efficient for them while significant removal (>90%) has been reported only for supramicron (>1 μm) particles (Chowdhury et al. 1993). Particulate colloidal suspensions are responsible for the major fraction (>80%) of residual aluminum in finished water (Schintu et al. 2000).

Though traditionally a pore size of 0.45 μm has been used to define soluble materials, it was noted by Van Benschoten & Edzwald (1990) that even 0.22 μm may not be a cut-off size to distinguish particulate and dissolved forms of Al. The filtrate from 0.22 μm nylon membrane filtration contains an important class of aluminum hydroxide species which are 1–2-nm-sized (Casey 2006) and very active in forming bonds (Licskó 1997) with silica particles. Various low particle size alumino-hydroxy complexes may get attached to colloidal particles to form a range of aggregate sizes. Therefore, the dissolved and particulate residual aluminum fractions reported here are subject to the constraints of the operational nature of the definition.

Total residual aluminum levels in water clarified from CC ranged from 172 to 450 ppb with the dissolved fraction of residual aluminum varying from 44 to 118 ppb. Lower residual aluminum levels were found in water clarified from PFBC with values of total residual aluminum ranging from 84 to 251 ppb and the dissolved fraction of residual aluminum varying from values of 12 ppb to 86 ppb. It can be observed from Figure 3 that the average total residual aluminum level for PFBC (σ = 142 ppb) was about half that of CC (σ = 285 ppb) and below the stipulated 200 ppb regulatory requirement for the same operational conditions. Similar to the observations of previous studies (Schintu et al. 2000), particulate residual aluminum was found to comprise the major fraction (≈72%) of total residual aluminum in clarified water. Therefore, it is important to control the particulate fraction since the average dissolved residual aluminum in clarified water was well below the 100 ppb mark, and recorded at 79 ppb and 40 ppb for CC and PFBC respectively.

It was previously shown (Srivastava et al. 2020) by means of MLR modelling, the existence of a negative correlation between influent turbidity and residual aluminum levels, while Schintu et al. (2000) had reported that more than 80% of the aluminum exists in particulate form. Considering the given range of influent conditions (0–20 NTU) at very low influent turbidity, suspended particulates are also less and result in lower residual aluminum, while it is relatively easier to treat higher influent turbidities since the ready availability of particles aids the bridging mechanism and lowers the probability of charge reversal. Thus, the two observations may together help explain the reason for slightly higher residual aluminum for relatively difficult-to-treat intermediate influent turbidity levels.

The values reported in this paper are of residuals in the water clarified from standalone PFBC/CC and residual aluminum levels are expected to decline after filtration. Moreover, further retention of particulate aluminum in the clearwell could result in additional reduction of total residual aluminum by about 40% in finished water compared to filtered water (Driscoll & Letterman 1995).

After treatment, the overall variation of pH (Figure 4) was within a narrow band of 7.6–7.9 and the average difference in pH between water clarified by CC and PFBC was marginal (0.05). A low drop in pH occurred since the supplied water had high alkalinity (Table 1), a high basicity variant of PACl was employed and coagulant doses for turbidity removal with minimal aluminum influx were administered. Lin & Ika (2019a) had discussed the role of pH in dictating the precipitation of various aluminum species and hence the concentration of residual aluminum, especially the dissolved fraction with PACls. Therefore, if the PACl is considered to be constituted of pre-hydrolyzed forms of monomeric aluminum (Al_a), polymeric aluminum (Al_b) and colloidal aluminum (Al_c), then the Al_b form may be expected to remain relatively stable with varying pH while precipitation of Al_c will be facilitated at near-neutral pH. The higher the basicity of PACl, the lower the content of Al_a, which would have resulted in lower dissolved residual aluminum (Kimura et al. 2013). Variation in clarified water temperature has also been depicted in Figure 4 and the mean temperature in this study was 18.4 °C (σ = ±0.2). Furthermore, a drop in water temperature from the standard 25 °C for pH values higher than the minimum aluminum solubility range of 6.5–7 would have tended to decrease the soluble aluminum concentration.

It can be inferred prima-facie from the juxtaposed microscopic images (Figure 5) that the flocs formed in PFBC are larger and those formed in CC are mostly smaller and more dispersed. This is also evinced quantitatively from Figure 6(a), where the floc diameters in clarified water from PFBC are larger compared to those in clarified water from CC. On average, the size of PFBC flocs was found to be 73% higher than CC flocs.

If a large sphere is split into several smaller spheres, then the net combined volume of the smaller spheres will remain the same as the original sphere but their total surface area would be much higher than the consolidated sphere. On similar lines, Figure 6(b) provides evidence with regards to dispersity, wherein the average particle count of 481 and the area coverage by flocs of 5.13% is more for CC compared to PFBC's respective figures of 232 and 4.52%. Thus, it may be inferred that PFBC flocs were larger and more consolidated while CC flocs were smaller and dispersed. At higher influent turbidity and for lower magnitude of ξ, visibly more flocs were detected.

It was earlier inferred from the destabilization results that enhanced collision of microflocs with well-grown large flocs available in abundance in the blanket resulted in contact flocculation (Hendricks 2006), which promulgated agglomeration. Accordant results on fine particles in suspension indicated that the larger number of surface sites provided by the blanket and the pulsating action abetted nucleation and floc growth/re-growth (Yu et al. 2011) at the micro levels. The above phenomena may together explain the larger floc formation and improved consolidation in the case of PFBC compared to CC.

Further investigation was carried out on the shape characteristics of the flocs. Irregular morphological structure of flocs is indicated by deviation from the spherical shape measured in terms of circularity (C) of the floc. It may be observed from Figure 7 that the C values for the two technologies were comparable; however, the PFBC flocs had slightly less smooth boundaries (C_avg = 0.528) than CC flocs (C_avg = 0.575). Furthermore, these moderate C values are in line with the observations (Jarvis et al. 2005) that when coagulation is carried out with PACl in slightly alkaline conditions, then unlike the ‘fluffy’ flocs formed by alum, the flocs are comparatively smaller with smoother boundaries.

For samples from the two different water treatment technologies, viz. CC (C_1–5) and PFBC (P_1–5), for different influent conditions, the resulting fractal dimensions with their standard error (SE), coefficient of determination (R²), mean characteristic length (l), population (N) and settling velocity (W_s) ratio are mentioned in Table 2.

The range of values of D₂ and D₃ for PACl flocs is in good agreement with the respective values reported for flocs formed after unit processes of coagulation-flocculation (Chakraborti et al. 2007; Wang et al. 2016; Lv et al. 2021). Also, for all influent conditions, the fractal dimension data was highly correlated, as indicated by R² values and with the reasonably large analysed population (N > 70). Marked difference in the average D₂ value may be observed for PFBC [(D₂)_avg = 1.89] and CC [(D₂)_avg = 1.74]. D₃ values for PFBC were also higher, averaging at 2.78 compared to an average of 2.48 for CC. Fractal dimensions D₂ and D₃, representing the geometrical structure of flocs in two- and three-dimensional planes respectively, are generally lower for more branched aggregates and higher values represent a more consolidated shape. Lower values of D₃ for CC indicate that a faster rate of coagulation can take place but flocs are less amenable to settling under gravity.

Furthermore, for the sweep flocculation mechanism, the values of D₃ have been reported to be below 2. Overall, values for D₃ for both technologies together ranged from 2.25 to 2.84, indicating inter-particle bridging to be the dominant mechanism (Wang et al. 2004). This is logically backed by the fact that PACl is composed primarily of polymeric aluminum species, especially ‘Keggin’ Al₁₃⁷⁺ cation [Al₁₃O₄(OH)]_n⁷⁺, which aids in inter-linking colloidal particles. Also, the range of recorded D₃ values was greater than 2, which suggests that a reaction-limited aggregation (RLA) process was followed rather than the traditional diffusion-limited aggregation (DLA) process (Chakraborti et al. 2007). In other words, aggregation was limited by electrostatic forces on the interacting particle/cluster surfaces. This entailed that, unlike in DLA, only a small fraction of collisions could successfully result in attachment and thus the particles had to penetrate deeper into the aggregates, resulting in a more consolidated floc structure.

In this study, the actual settling velocities of individual aggregates could not be reported because it is difficult to accurately measure parameters like density of flocs in dynamic systems. However, the ratio of settling velocities can be assessed indirectly with the help of proportionality equation (Equation (4)) relating W_s with fractal dimensions and characteristic length. Values of D₃ and D₂ play an interesting role in affecting the settling velocity of the flocs and it may be inferred from Equation (4) that D₃ has a positive while D₂ has a negative exponential correlation with settling velocity. The color-coded scales depicted in Table 2 allow ready comparison of the relative magnitudes of characteristic length, D₂ and D₃ of CC and PFBC flocs for mapping the combined effect on the settling velocity. Settling velocity of flocs for PFBC was found to be at least 50% higher than CC flocs. The results signify that conditions of a pulsating flow regime in the presence of fluidized flocculated particles in high concentration abet contact flocculation, which renders the overall size, structure and morphology of the formed flocs such that they were more amenable to settling.

An explanation may be given on the basis of the nucleation sites and reaction time, which play a crucial role in the relative formation of aluminum hydroxide and hydrolyzed aluminum species that are active in establishing bonds with the colloids as well as previously formed inert flocs (Licskó 1997; Wang et al. 2004). Effectively increased residence time and a high solids contact zone is offered in the lower columns of the PFBC, which affects the speciation and may have resulted in a better utilization of coagulant and more settleable flocs.

Scanning electron microscopy (SEM) showed that the sludge of PFBC (Figure 8(b)) was more compact than CC sludge (Figure 8(a)). Coagulant affects the morphology of flocs/sludge, and it was observed at high magnification that unlike the more fibrous/jagged microstructure obtained with alum, smoother boundaries were obtained with PACl. Main voluminous surface coverage is by kaolin, and Al₁₃ aggregates adsorbed on the surface of the kaolin particles extend to attach other particles in the interior of the PACl-Al₁₃ agglomerate.

It has been generally accepted that Al₁₃ species with the ε-keggin structure, which has four planar trimeric Al₃(OH)₆ groups linked to the central Al(O)₄ site via four μ₄-O (at the edges), comprise the major fraction of PACl and are chiefly responsible for particle aggregation by strong initial charge neutralization (Casey 2006; Lin et al. 2008).

It is proposed that the structure of aggregates formed was akin to the chain-like clusters formed by small Al₁₃ units bridged via anions (Casey 2006) with colloidal adsorption destabilization, inter-particle bridging, as well as by colloidal particles adhering by charge neutralization at cationic sites (Hendricks 2006) as shown in Figure 9. Given the RLA mode of aggregation and low |ξ| values (Figure 2(a)), interparticle bridging is purportedly dominant (Lin et al. 2008). It may also be noted that the microstructure shown would be mostly the same for CC and PFBC flocs with PACl, though the macro parameters like size and consolidation may vary.

Comparative performance assessment of pulsating floc blanket clarifier (PFBC) and conventional clariflocculator (CC) along with examination of flocs formed in the two processes with PACl was carried out. The residual turbidity was found to be lower by 74%, flocs and fine colloids in suspension were larger by 73 and 75% respectively, and the total and dissolved residual aluminum lower by 50 and 49% respectively on average for PFBC compared to CC. The particulate form comprised the major fraction (≈72%) of the total residual aluminum for both. Unlike for CC, fine colloidal particles suspended in the clarified water from PFBC were on average supramicron, which can be effectively removed in the subsequent filtration unit process. Destabilization of colloids improved by 47% in PFBC compared to CC under identical conditions, as characterized by their settling velocity and floc geometry and morphology, which helped improve its performance substantially both in terms of residual turbidity and residual aluminum.

We acknowledge the support of Materials Research Centre (MRC) at Malaviya National Institute of Technology Jaipur, India, in carrying out nano-particle size analysis (nano-PSA), zeta potential analysis and scanning electron microscopy (SEM) at their facilities.

The authors have no conflicts of interest to disclose.

All relevant data are included in the paper or its Supplementary Information.