Physical and chemical impacts of conventional unit processes and coagulants on the dinoflagellate, Ceratium hirundinella

Dinoflagellates are algae, belonging to the phylum Pyrrophyta and the class Dinophyceae (Wetzel 2001). Dinoflagellates are considered as an intriguing planktonic algal group for numerous reasons, including their ability to luminesce and produce neuro-toxins (Harland 1988). Members of the group are often covered with armour-like cellulose plates underneath the cell membrane. The presence of these thecal plates differentiates dinoflagellates from other algal groups (Canter-Lund & Lund 1995). The movement of dinoflagellates is accomplished by two flagella implanted in grooves on the surface (Canter-Lund & Lund 1995; Janse van Vuuren et al. 2006). Dinoflagellates are common in freshwater, but they are most diverse in marine environments (Canter-Lund & Lund 1995). In the marine environment, dinoflagellates (e.g. Gymnodinium, Gonyaulax) are responsible for the so called ‘toxic red tides’ that may result in the death of marine life as well as humans that consume contaminated marine animals (Canter-Lund & Lund 1995). Though no freshwater ‘toxic red tide blooms’ have been reported when freshwater genera (e.g. Ceratium, Peridinium) proliferate their blooms produce a brown discoloration of the affected water.

Morphological variability in dinoflagellates is described by Taylor (1987) and has been correlated to buoyancy and cell resistance to sinking (Hamlaoui et al. 1998; Wetzel 2001). Members in this group (especially Ceratium species) may undergo seasonal changes known as poly- or cyclomorphism to enlarge the surface area of cells (Hamlaoui et al. 1998; Wetzel 2001; Gligora et al. 2003). Poly- or cyclomorphism can be recognized by lengthening of the horns as the temperature increases during warmer summer months and shortening during winter (Wetzel 2001). These adaptive changes are significant to C. hirundinella as a mechanism to reduce the sinking rate from the euphotic zone (Wetzel 2001).

The morphology and surface characteristics of C. hirundinella are complex and may interfere with coagulation. During the water treatment process, appendages to cells reduce contact between the cells, which is necessary for successful agglomeration during floc formation (Knappe et al. 2004). Furthermore, dinoflagellate cells can clog sand filters and produce taste and odour compounds (Knappe et al. 2004; Swanepoel et al. 2008a). In addition, the presence of advanced motile structures causes several challenges during conventional water treatment, because of its ability to disrupt coagulation and flocculation and avoid sedimentation (Swanepoel et al. 2008a). It was evident that C. hirundinella cells may penetrate through potable water treatment systems where they are responsible for high concentrations of organic compounds (Swanepoel et al. 2008a).

According to Palmer (1980) the control of such algae during water treatment requires optimized unit processes. However, well regulated coagulation with sedimentation will often remove 90 percent of algae and in some cases 95-96 percent removal has been reported (Palmer 1980). C. hirundinella can avoid coagulation, flocculation and sedimentation when coagulants dosed are not appropriate to promote the removal of these robust cells (Swanepoel et al. 2008a; Van der Walt 2012). Henderson et al. (2008) reported that the surface area of algal cells is a useful preliminary indicator of coagulant dose required for optimum (appropriate) cell removal. Furthermore, the physical and chemical aspects involved during conventional water treatment can cause major damage to cells, which can clearly be seen under the scanning electron microscope (SEM). The aim of this study was to investigate the physical and chemical impacts of coagulation, flocculation and sedimentation on the morphology of C. hirundinella cells and to describe the physical appearance of flocs containing C. hirundinella cells.

20,000 mg/l Hydrated lime (Ca(OH)₂), 2,000 mg/l activated silica (SiO₂) and (2,000 mg/l), and organic polymer stock solutions were prepared similar to what is used in a water treatment plant. The following dosage ranges were prepared for simulation:

• Ca(OH)₂-SiO₂ treatment: Ca(OH)₂ dosages ranged from 60 to 160 mg/l (with increments of 20). A SiO₂ dosage of 4 mg/l was added as a coagulant aid;

• Ca(OH)₂-organic polymer treatment: organic polymer dosages ranged from 4 to 14 mg/l (with increments of 2). A Ca(OH)₂ dosage of 10 mg/l was added as a coagulants aid; and

• Organic polymer treatment: dosages ranged from 4 to 14 mg/l (with increments of 2).

After conducting jar stirring tests using various coagulants, flocs that settled for 20 minutes were collected to perform SEM investigations. C. hirundinella cells collected from source water and floc samples formed during jar stirring tests were fixed in a formaldehyde solution (35%) by adding 200 μl of preservative to 10 ml of floc-sample. After washing (3× for 15 min each) in cacodilate buffer and distilled water, samples were dehydrated in an ethanol series of 80%, 90% and 2 × 100% for 15 minutes each. Samples were critical-point dried with CO₂, on SEM stubs with double-sided carbon tape and coated with 20 nm gold/palladium (66/34%) in a sputter coater.

A FEI QUANTA 250 FEG scanning electron microscope, operating at 7.30 kV and 10.00 kV, was used to investigate the morphology (characteristics and structure) of C. hirundinella cells in source water as well as in the flocs

Broken cells, with uncovered organic material attached to the sulcus area, became more fragile after flash mixing and subsequently fragile to chemical treatment, notably when the pH levels were increased by dosing hydrated lime as the primary coagulant. Ca(OH)₂, in combination with SiO₂, increased the pH levels in the source water from 7–8 to levels of 11-12. Visually larger flocs were formed compared to the floc sizes when other coagulants were used.

Medium sized flocs were formed during Ca(OH)₂-organic polymer treatment, when compared to Ca(OH)₂-SiO₂ treatment (when large flocs formed) and organic polymer alone treatment (when small flocs formed). The flocs formed during this treatment appeared smooth to dense and contained cells with damage that occurred only from high energy flash mixing. This observation was reached since cells were only observed to release organic material through the sulcus grooves (Figure 6(a)–6(c)). Relatively good C. hirundinella removal was confirmed by TPP measured in the supernatant after sedimentation.

Less dense flocs were formed compared to flocs formed with other coagulant options (Figures 5–7). Only a few C. hirundinella cells were observed in the flocs that appeared smooth with large open areas between individual cells in aggregation. Poor floc formation is also revealed in Figure 7(c) where C. hirundinella cells are lying apart from flocs. The poor removal of C. hirundinella cells by organic polymer is also confirmed by the highest TPP concentrations measured in the supernatant after sedimentation (Ewerts et al. 2014, 2015).

The removal of C. hirundinella cells was recorded at various Ca(OH)₂-SiO₂ dosages (as indicated by TPP removal) and can be explained by the formation of denser flocs in aggregation containing relatively large pieces of broken cells. Organic polymer achieved poor C. hirundinella removal during all dosages which is explained by fewer cells imbedded in less dense flocs in aggregation (Figure 7). Relatively good C. hirundinella removal was achieved by Ca(OH)₂-organic polymer which is confirmed by TPP removal. Flocs that formed after dosing Ca(OH)₂-organic polymer were dense with no further damage after high energy flash mixing.

Conventional water treatment includes a series of unit processes (coagulation, flocculation, sedimentation, sand filtration and disinfection) aimed at reducing the concentration of microorganisms and other suspended particles (Betancourt & Rose 2004). When source water contains C. hirundinella cells, many water treatment challenges can occur (e.g. clogging of sand filters) and these may be compounded by water quality problems (e.g. tastes and odours) (Swanepoel et al. 2008a). Ceratium cells penetrating into the treated water also contribute to large quantities of organic material due to their large cell size (Swanepoel et al. 2008a). Organic material entering water may result in the formation of harmful disinfection products, such as trihalomethanes, when chlorine is used as a disinfectant (Rositano et al. 2001; Rodríguez et al. 2007). Furthermore, organic material in a water distribution network also serves as a carbon source which promotes bacterial growth (Prévost et al. 2005). This study therefore focuses on addressing challenges experienced by conventional water treatment from a SEM perspective, since the objectives were set to investigate both the impacts on water treatment options on C. hirundinella cells and to analyse the flocs in aggregation.

These freshwater dinoflagellate cells are often described in literature to be ‘robust’ and therefore it was expected that cells would remain intact after the physical high energy flash mixing step. The physical flash mixing procedure forms an integral part of the conventional coagulation unit process to disperse coagulants quickly and uniformly throughout the source water (Chow et al. 1999). The impacts of flash mixing caused disruption of C. hirundinella cells, which force the cells to release organic material (Figure 4). Organic material released from C. hirundinella cells may create an increased coagulant demand (Sharp et al. 2006). Increasing coagulant demand (e.g. Ca(OH)₂) demand may impact negatively on the physical integrity of C. hirundinella cells and causes major disruption to the cell morphology which will have subsequent adverse impacts on coagulation conditions (Henderson et al. 2008).

The structure and strength of the flocs play important roles in water containing high phytoplankton biomass, notably when unit processes, such as coagulation and flocculation are affected by the selection of coagulant chemicals and optimum dosages thereof (Yuheng et al. 2011). The chemical impacts of Ca(OH)₂-SiO₂ treatment caused major damage to the cell morphology of C. hirundinella. It has become evident from SEM images that Ca(OH)₂-SiO₂ removed cells by causing serious cellular damage, as a result of high pH levels, instead of removing the intact cells (Figure 5). During a recent investigation where a bioflocculant was used to harvest microalgae, cell damage was also observed when flocculation was induced by decreasing pH. The influence of pH during water treatment is widely applied to achieve optimum coagulation and flocculation of suspended matter in water (Divakaran & Sivasankara Pillai 2002; Knappe et al. 2004; Degrémont 2007). The impact of pH adjustments during flocculation is described as feasible (either by increasing or decreasing pH) to improve flocculation efficiency (Nfikubwimana et al., 2014). However, major concerns have arisen by the impacts of high pH levels on C. hirundinella cells, notably with respect to aforementioned water quality problems.

Using organic polymer as a primary coagulant, in combination with Ca(OH)_2, formed medium sized flocs that appeared smooth in dense aggregation (Figure 6). C. hirundinella cells in floc aggregations were more intact than what was previously observed when dosing Ca(OH)₂-SiO₂. Less cell damage to C. hirundinella cell morphology can be ascribed to adjusted pH levels of 9–10, which enhanced the effective removal of C. hirundinella cells. These pH levels were sufficient to render cells immobile, thereby assisting in the coagulation and flocculation processes, since C. hirundinella cells could no longer swim to disrupt floc formation. This finding was also confirmed by Van der Walt (2012) who noted that immobilization of C. hirundinella cells will assist coagulation and flocculation. The inhibition of flagellar movement is often the result of a ‘pH shock’ as described by Ferreira & Du Preez (2012).

It has been reported that organic polymers resulted in stronger aggregates (flocs) than other coagulants and are beneficial for slow-settling flocs (Bolto & Gregory, 2007). However, during this investigation it was observed that flocs settled slowly, but were smaller than those formed with other coagulant options. The cells were separate from flocs in aggreation, giving an indication of poor coagulation and flocculation conditions. It has been reported that organic polymers as coagulants produce fragile loosely packed flocs (Wakeman & Tarleton, 1999) which can be observed in Figure 7. The mechanisms influencing floc formation, such as pH changes and charge neutralization, are different when using organic polymer and Ca(OH)₂. Because Ceratium cells are highly mobile (Grigorszky et al., 2003), they often disrupt flocs and are difficult to remove. Floc stability is directly related to the strength and number of bonds holding the floc together (Jarvis et al., 2005). It has become evident from this study that organic polymer alone causes no further damaging effects to the morphology of the cells. No pH adjustments therefore result in water treatment problems, since most of the cells remain motile in the supernatant and have the ability to interfere with floc formation by disrupting flocs. These observations are confirmed by SEM images of flocs, where fewer cells were embedded in flocs or appear separate from aggregates (Figure 7(a)–7(c)). Poor C. hirundinella cell removal during coagulation and flocculation puts major strain on the sand filters. Cell accumulation in the sand filters can result in taste and odour problems or penetration into the treated water.

The freshwater dinoflagellate C. hirundinella, with its complex morphology and robust thecal plate cell covering, is well-known to cause problems that can be divided into water treatment and water quality problems. Many water treatment plants experience poor C. hirundinella removal, as well as taste and odour related problems when dosing certain coagulant chemicals. SEM investigations revealed both physical and chemical impacts on C. hirundinella cells that occur as a result of high energy flash mixing and coagulants. Flash mixing resulted in disruption to cell morphology and subsequently caused cells to release organic material through the sulcus groove. Released organic material remained attached to the sulcus groove after flash mixing, but the addition of certain coagulants, notably Ca(OH)₂-SiO₂, with an associated increase in pH to levels above 10, resulted in major damage to cells. This coagulant option achieved the best C. hirundinella cell removal as a result of breaking up cells and the formation of dense flocs. However, coagulant options associated with no pH adjustment (e.g. organic polymer) caused no further damage to cells after flash mixing, but poor C. hirundinella cell removal and loosely packed flocs. Ca(OH)₂ in combination with organic polymer as a primary coagulant caused no further major damage to cells after flash mixing. Intact cells had more organic material attached to their sulcus grooves. This can be ascribed to adjusted pH levels below 10 that assist coagulation conditions as observed from smooth and less dense to denser flocs that contain mostly intact cells. Water treatment plants that experience problems with phytoplankton in source waters can use SEM investigations to assist in coagulant selection as it plays a major role in addressing water-related problems caused by phytoplankton, such as C. hirundinella.