Algal softening followed by ozonation: The fate of persistent micropollutants and natural organic matter in groundwater

Abstract

A proof-of-concept study evaluates the performance of a novel strategy using photosynthetic microorganisms to soften groundwater instead of using caustic chemicals. The microalga Scenedesmus quadricauda was used to increase the pH of the groundwater via natural photosynthesis. This work applied softening as a pretreatment to ozonation of hard groundwater and mainly focused on investigating the multiple effects of algal softening on the degradation of persistent micropollutants upon subsequent ozonation. The algae-induced alkaline conditions (pH >10) were favorable to catalyze the formation of OH radicals directly from O₃ molecules. Moreover, algal softening removed the strong radical-scavenging carbonate species (HCO₃⁻ and CO₃²⁻) to a much greater extent than that achieved by chemical softening, which was attributed to the combination of mineral carbonation and metabolic CO₂ reduction. The fate of the natural organic matter (NOM) was characterized with spectroscopy, chromatography, and bioassay, which indicates that algal treatment decomposed the NOM to be less susceptible to attack by OH radicals. Consequently, the ozonation of alkaline groundwater achieved a better removal of the micropollutant residues in groundwater. Carbamazepine and diclofenac were used as model chemicals of persistent groundwater contaminants and were almost completely removed with an addition of 1.25 mg O₃ L⁻¹ (0.63 mg-O₃ mg-C⁻¹).

Graphical Abstract

1. Introduction

In 2015, groundwater in the United States was withdrawn for public supply at a rate of 15,200 million gallons per day (69 × 10⁶ m³ d⁻¹), accounting for 39 % of all public supply withdrawals nationwide, which are mostly delivered to customers for domestic, commercial, and industrial needs (Dieter et al., 2018). Groundwater typically contains a wide range of micropollutants that are degradable with varying success during surface water infiltration and/or groundwater recharge (Lapworth et al., 2012). The occurrence of unregulated micropollutants (e.g., pharmaceuticals, fragrances, and flame retardants) in groundwater and in drinking water have been reported by Glassmeyer et al. (2017), and the associated concerns have also been raised over their potential adverse impacts on public health. Since long-term exposure to micropollutants may lead to cancer, hormone variation, endocrine disruption in the human body (Ngo et al., 2020), a robust removal of micropollutant residues in water sources is emphasized along with regulating chemical discharges into the water bodies.

Chemical oxidation, such as ozonation, is considered as a viable option to address the increased demand for degradation of a wide range of persistent micropollutants in potable water sources. Ozonation has been identified as a promising strategy to treat various sources of water, and it is increasingly combined with photo- and chemical-catalytic techniques for advanced oxidation processes (AOPs) that involve generating highly reactive radical intermediates (e.g., OH radicals) (Ikehata and Li, 2018). O₃-based AOPs are generally more effective than ozonation alone due to the increase in generation of OH radicals (Ikehata et al., 2006) and are thus frequently recommended as robust alternatives to remove refractory organic compounds in water and waste-water. Ozonation at a high pH can be one of AOPs since OH-induced O₃ decomposition can be significant over pH 10 where the observed half-life of O₃ would be expected to be in the range of minutes (Gerrity et al., 2017). Therefore, this would only be practical if the pH was raised for other purposes and carbonate was removed, such as for softening (Crittenden et al., 2012). Precipitative softening is the most commonly used method to reduce hardness in centralized water treatment facilities, especially when using groundwater as a source (Randtke, 2011). Although hardness removal has traditionally been considered for aesthetic issues without any regulatory limits, approximately 15 % of the water treatment plants in the United States for domestic supply employ some type of the precipitative softening process (Elder et al., 2012). Waters with combined calcium and magnesium ion concentrations above 180 mg L⁻¹ as CaCO₃ are considered very hard (USGS, 2020), and most utilities that soften produce a finished water hardness between 75 and 150 mg L⁻¹ as CaCO₃, corresponding to moderately hard water (Randtke, 2011).

Upon softening, the precipitation of divalent cations results in an equivalent removal of radical-scavenging inorganic carbon ions (HCO₃⁻ and CO₃²⁻), and the resulting alkaline conditions are favorable to the formation of OH radicals directly from O₃ molecules. The reactivity of ozone toward a wide range of trace organic compounds and second-order rate constants for the respective reactions have been extensively studied and reported in the literature (Gerrity et al., 2017). In this study, carbamazepine (CBZ) and diclofenac (DCF) were selected as model micropollutants in groundwater. Those chemicals are among the most frequently detected pharmaceutical residues in water bodies (Zhang et al., 2008) and have been classified as harmful to aquatic organisms (Petrie et al., 2015). The reactions of CBZ and DCF with O₃ are much slower than those with OH radicals, whereas the rate constants of OH radicals for the selected micropollutants are about the magnitude of the diffusion-limited acid-base reactions (i.e., ~10¹⁰ M⁻¹ s⁻¹) that are considered as the fastest aqueous-phase chemical reactions. Natural organic matter (NOM) has a much greater influence on retarding the reaction rate than the alkalinity. Several previous studies have reported a negative impact of humic-like substances on the degradation of trace organic chemicals by OH radicals produced with various oxidation processes (Doll and Frimmel, 2005; Vogna et al., 2004). Precipitative water softening tends to partially remove NOM that can be measured as a dissolved organic carbon (DOC) together with UV absorbance at 254 nm (UV₂₅₄) (Russell et al., 2009). The associated removals of disinfection byproduct formation potentials (DBPFPs) (Kalscheur et al., 2006) and assimilable organic carbon (AOC) (Hammes et al., 2010) have been reported in previous studies. Decreasing the amount of NOM can result in reduced chemical demand for multiple purposes, such as disinfection and oxidation in water treatment facilities. In particular, the NOM aromaticity, which can be approximated by the specific UV absorbance (SUVA) value, is positively correlated with chorine demand and the associated formation of harmful byproducts during chlorination of the produced water (Edzwald and Tobiason, 2011).

Overall, O₃-based oxidation could be more effective if pretreatment processes such as softening would be used not only to catalyze the formation of OH radicals directly from O₃, but also to remove the alkalinity from the groundwater. Nevertheless, the combination of precipitative softening and subsequent ozonation is very scarce and has not received much attention compared to other ways to improve the generation of OH radicals in O₃-based water treatment including the addition of H₂O₂, UV irradiation, and metal catalysts.

In this study, we employed an alternative strategy to increase the pH of water using living algal components that are capable of reproduction and renewal, unlike conventional precipitative softening that typically requires a large consumption of caustic chemicals and the associated production of non-reusable inorganic solids. Phototrophs are well known to produce OH⁻ during photosynthesis (Eq. 1), which can greatly increase the saturation of carbonate minerals and thus better allow for precipitation (Eqs. 2 and 3) (Power et al., 2011). We previously noticed that Scenedesmus quadricauda has considerably efficacy in removing both hardness and alkalinity during common autotrophic growth in light (Maeng et al., 2018a). A more important reason for choosing these specific “bio-agents” is that the photostimulation of S. quadricauda has been shown to induce degradation of humic-like substances that are typically resistant to microbial decomposition (Kim, 2018). To the best of our knowledge, no studies have dealt with algal softening of hard groundwater, primarily focusing on improving the performance of subsequent ozonation for degradation of recalcitrant micropollutants.

$${\text{HCO}}_3^{-}+{\mathrm{H}}_2\mathrm{O}+hv\to {\text{CH}}_2\mathrm{O}+{\text{OH}}^{-}+{\mathrm{O}}_{2(\mathrm{g})}\uparrow$$ (1)

$${\text{Ca}}^{2+}+{\text{CO}}_3^{2-}\to {\text{CaCO}}_{3(\mathrm{s})}$$ (2)

$${\text{Mg}}^{2+}+{\text{CO}}_3^{2-}\to {\text{MgCO}}_{3(\mathrm{s})}$$ (3)

This study presents a novel strategy to improve the performance of ozonation in terms of persistent micropollutant removal. It also advances the following goals during the treatment of groundwater: (1) effective removal of the carbonate species (as a strong radical scavenger) with an alternative method of softening groundwater using photosynthetic bio-agents; (2) enhanced oxidation of groundwater by catalyzing the production of highly reactive OH radicals under alkaline conditions, where the indirect reaction dominates; and (3) substantial reduction in aromaticity of NOM to mitigate the quenching of OH radicals with NOM during ozonation of alkaline groundwater. In this study, softening was combined with subsequent ozonation to achieve a robust treatment of hard groundwater in which a certain level of persistent micropollutants and humic-like organic compounds were given. Groundwater softening was conducted with two different methods using natural bio-agents and caustic chemicals, and each of those softening was followed by ozonation. The performance was evaluated with respect to the removal of DOC, UV₂₅₄, and micropollutants (CBZ and DCF) and to variations in well-known indicators of organic aromaticity (SUVA) and biodegradability (AOC). The fate of NOM was also characterized in this study using a suite of innovative analytical tools.

This work is distinctive from other investigations in the following ways: (1) we identified an alternative softening strategy using living algal components (as renewable bio-agents) that are capable of self-repair and reproduction, instead of adding caustic chemicals that result in a high level of total dissolved salt in the produced water; (2) demonstrated the robustness of persistent micropollutant removal using softening followed by ozonation, unlike most previous investigations allocating ozonation ahead of softening (Grasso et al., 1989; Hammes et al., 2010; Hsu and Yeh, 2003); (3) employed a suite of advanced characterization techniques (e.g., spectroscopy, chromatography, and bioassay) to better elucidate the changes in composition and functional properties of NOM and to provide further insight into the contribution and role of microalgae in treating NOM-laden groundwater.

2. Materials and methods

2.1. Preparation of synthetic groundwater

Unless specified otherwise, all chemicals used in this study were of analytical grade and were supplied by Sigma-Aldrich (St. Louis, MO). Synthetic groundwater was prepared by adding trace organic chemicals and organic matter in commercially available natural spring water (Evian, France). Springs are basically points where groundwater emerges onto the surface. CBZ and DCF were used as the model chemicals for the persistent groundwater contaminants. The physicochemical properties of the selected chemicals have been reported in detail elsewhere (Zhang et al., 2008). A diluted mixture of both chemicals was prepared using their stock solutions and then spiked into the spring water. Purified humic acid was utilized as a surrogate for NOM present in groundwater. The physicochemical characteristics of purified humic acid have been reported in detail elsewhere (Hur and Schlautman, 2003). Humic-like substances are usually derived from transformations of organic matter in chemical and biological processes. They are typically refractory to microbial degradation, biogenic, and yellow-colored organic acids. Aldrich humic acid sodium salts were purified through repeated pH adjustment, precipitation, and centrifugation to remove inorganic impurities (Hur and Schlautman, 2003). The organic concentration was shown as DOC and UV₂₅₄. NaNO₃ and NaCl were also used as additives to prepare synthetic groundwater. Nitrate is the most common inorganic contaminants derived from man-made sources and has been reported at concentrations greater than the National Primary drinking water standard (10 mg L⁻¹ as N) (Nolan and Hitt, 2006). The level of added NaCl corresponds to the median concentrations of Na⁺ and Cl⁻ in drinking-water supply wells (Mullaney et al., 2009). The final concentrations of organic and inorganic components in the synthetic groundwater are presented in Table 1. The concentrations of CBZ and DCF in the prepared groundwater were determined as 5.9 ± 0.2 μg L⁻¹ and 6.7 ± 0.4 μg L⁻¹, respectively. Pharmaceutically-active chemicals have been detected in the range of several micrograms per liter in actual groundwater (Capdeville and Budzinski, 2011), and therefore our initial levels of CBZ and DCF were deemed appropriate for these experiments.

Table 1.

Characteristics of synthetic groundwater used in this study. Stock solutions of some organic and inorganic chemicals (NaNO₃, NaCl, CBZ, DCF, and purified humic acid) were separately prepared and added in commercially available spring water (Evian, France) and the following characteristics were observed.

Parameter	Unit	Concentration
pH	–	7.5 ± 0.1
Conductivity	μS cm−1	818 ± 7
Alkalinity (as CaCO3)	mg L−1	284 ± 14
Total hardness (as CaCO3)	"	302 ± 11
Ca2+	"	79 ± 5
Mg2+	"	26 ± 2
Na+	"	23 ± 1
K+	"	1.0 ± 0.1
Cl−	"	19 ± 3
SO42−	"	12 ± 2
NO3−	"	26 ± 3
DOC	"	2.72 ± 0.08
UV254	cm−1	0.22 ± 0.01
Carbamazepine	μg L−1	5.9 ± 0.2
Diclofenac	"	6.7 ± 0.4

2.2. Softening

The alkalinity was measured by titrating 100 mL of the liquid sample with diluted sulfuric acid to pH 4.5, and the hardness was measured by titrating 100 mL of the liquid sample with diluted EDTA solution. Each measurement was carried out at least in triplicate, and the average values were reported. The pH and conductivity of the liquid sample were measured using a pH meter (Orion Star A215, Thermo Scientific, USA). Common anions (Cl⁻, NO₃⁻, and SO₄²⁻) and cations (Na⁺, K⁺, Ca²⁺, and Mg²⁺) were also measured using standard methods (APHA et al., 2017). Visual MINTEQ (model 3.0) was used to calculate the expected theoretical percent removal for both the hardness and bicarbonate alkalinity at a given pH of the groundwater samples.

2.2.1. Algal softening

The stock culture of S. quadricauda (AG 10003) was prepared as described in our previous work (Maeng et al., 2018a). The algal cells in the exponential growth phase were collected via centrifugation (1620 g, 10 min), washed with mineral bottled water, and inoculated at an optical density (OD) of 0.4 cm⁻¹ (680 nm) (corresponding to 6.9 ± 0.5 × 10⁵ cells per mL) into a flask containing 800 mL of synthetic groundwater (Supporting Information Fig. S1). The flask was incubated under continuous illumination while shaking at 130 rpm for 24 h. Fluorescent lamps with 6500 K color temperature were used as the source of light, and the incoming light intensity to the flask was 150 μmol photons m⁻² s⁻¹. The light intensity was measured using an MQ-500 quantum meter (Apogee Instruments, Inc., USA). During incubation, 50 mL of mixed suspension was collected from the flask at 6, 12, and 24 h for the pH measurement. The cell growth was monitored by determining the OD680 of the mixed suspension, and the physiological properties of S. quadricauda were characterized using fluorescence-based flow cytometry (Supporting Information Fig. S2). This batch test was performed at 25 °C in triplicate. The supernatant was collected from the flask at the end of the test, filtered using 0.45 μm microfilters, and stored in a refrigerator at 4 °C until use for further ozone treatment.

2.2.2. Chemical softening

Precipitative softening of groundwater with caustic soda was conducted to verify the effectiveness of algal softening by comparing the effects of two different softening methods on the performance of the subsequent ozonation. The groundwater was poured into a 1 L glass beaker fitted with a floating top that minimizes CO₂ exchange during chemical softening. While agitating using an overhead stirrer with a conventional blade, diluted NaOH solution was added into the 1 L glass beaker containing 800 mL of groundwater to achieve a softening pH of 10.2, which is equivalent to the eventual pH of algal softening. The addition of NaOH was completed in three minutes of rapid mixing (150 rpm), followed by an hour of slow mixing (30 rpm) to precipitate solids. The suspension was collected from the beaker at the end of the slow mixing period, filtered through a 0.45 μm microfilter, and analyzed to characterize the organic residues in groundwater. An additional filtered sample was withdrawn and stored in a refrigerator at 4 °C until use in further ozone treatment.

2.3. Ozone treatment

Ozone was generated (Lab2B, Ozonia, Korea) and injected into chilled deionized water (DIW), resulting in O₃-saturated H₂O with an initial concentration of 25 ± 2 mg O₃ L⁻¹. The UV absorbance of O₃-saturated H₂O was measured at 258 nm using a UV/Vis spectrophotometer (DR/5000, Hach), and the O₃ concentration was calculated using a molar extinction coefficient of 3000 M⁻¹ cm⁻¹ at 258 nm (Elovitz and von Gunten, 1999). Three different groundwater samples (i.e., raw, algal-softened, and chemically-softened) were prepared and ozonated by adding O₃-saturated H₂O to the samples. A 90-mL aliquot of each groundwater sample was mixed with O₃-free DIW and/or O₃-saturated H₂O to achieve 100 mL of the mixture, aiming for O₃-saturated H₂O contents of 0, 2, 5, and 10 % (v/v). O₃-saturated H₂O was used immediately after production, and the initial O₃ concentrations in the mixtures were 0, 0.50 ± 0.03, 1.25 ± 0.08, and 2.50 ± 0.16 mg O₃ L⁻¹ as a function of O₃-saturated H₂O dose (Supporting Information Fig. S3). The dilution ratio of all ozonated samples were identical (10 %, v/v) regardless of the O₃ dose. This ozonation method was used to eliminate any experimental biases resulting from different dilution ratios.

2.4. Analytical methods

2.4.1. Precipitates

During algal softening of groundwater, bio-solids with precipitates were collected as suspension and were subsequently characterized using field emission scanning electron microscopy (FE-SEM). The morphological features of the algal cells suspended in the groundwater were observed using a FEI NOVA NanoSEM 200 microscope equipped with a gaseous secondary electron detector using an accelerating voltage of 15 kV. Some bio-solids with precipitates were centrifuged, air-dried, and subjected to elemental analysis to verify the precipitation of Ca²⁺ and Mg²⁺. The characterization of the pre-treated solid samples was conducted using an identical microscope combined with an Ametek energy-dispersive X-ray (EDX) analyzer that allowed semi-quantitative analysis of the elements.

2.4.2. Natural organic matter

Extensive characterization of NOM was conducted as previously reported (Kim et al., 2014, 2017), and the procedures have been also described in detail elsewhere (Hammes and Egli, 2005; Huber et al., 2011). Briefly, the DOC (TOC-V CPN, Shimadzu, Japan) and the UV₂₅₄ (DR/5000, Hach, USA) were measured after filtration using flat-sheet 0.45 μm microfilters. The SUVA was calculated from the UV₂₅₄ divided by the DOC of the NOM sample. Each measurement was carried out at least in triplicate, and the mean and standard deviation values were reported. Fluorescence spectra were obtained using a Shimadzu RF-5301PC fluorescence spectrometer with a 150 W xenon lamp source. The samples were filtered (0.45 μm) and diluted (1 mg L⁻¹) prior to the spectroscopic analysis. The molecular size distribution of the NOM was determined via liquid chromatography with online organic carbon detection (LC-OCD) (DOC-Labor, Germany), for which real-time measurements of both DOC and UV₂₅₄ of the NOM effluent passing through a size-exclusion column were performed. AOC measurement was performed to assess the biodegradability of the organic residues in groundwater. All samples were pasteurized in a water bath at 70 °C for 30 min, filtered through a 0.1 μm membrane, and inoculated at ~1 × 10⁴ cells mL⁻¹ with a natural microbial consortium originating from mineral water (Evian, France). The samples were incubated at 30 °C for 48 h, and an enumeration of bacterial cells was carried out using a Partec Cube 6 flow cytometer (Partec GmbH, Germany). The concentration of AOC in the sample was calculated using a conversion factor of 10⁷ cells μg-AOC⁻¹ (Hammes and Egli, 2005).

2.4.3. CBZ and DCF

CBZ and DCF were analyzed using an Agilent 1200 high-performance liquid chromatograph. Mass spectrometry detection was performed on an Agilent 6460 triple-quadrupole mass spectrometer equipped with a dual jet stream electrospray ionization source (Supporting Information Fig. S4). The details have been described in our previous work (Maeng et al., 2018b).

3. Results and discussion

Fig. 1 shows spontaneous water alkalization induced by photosynthetic bio-agents in light. The growth of S. quadricauda resulted in a gradual increase in the pH of groundwater during the experiment, which in turn decreased both the hardness and alkalinity (determined by titrating from pH 8.3 to 4.5) in the aqueous phase. The SEM images of algal suspension collected after water softening show distinctive particles that were newly-produced and attached around the algal cells (Fig. 2). The OH⁻ release outside the algal cells can result in superalkalization localized on the cell surface, which can prompt mineral precipitation over the given circumstances. To support this observation, the results of the EDX measurements show an increase in the percent distribution of Ca²⁺ and Mg²⁺ in the elemental composition of the bio-solids after algal softening (Fig. 2). The hardness decreased by 52 % after the 24-h softening of groundwater, which was almost equal to the calculation obtained with Visual MINTEQ (model 3.0) under the given conditions. Based on the classification of water supplies by the degree of hardness (Randtke, 2011), the algal softening achieved a “moderately hard” range between 75 and 150 mg L⁻¹ as CaCO₃. The observations support the complexation of cations (Ca²⁺ and Mg²⁺) and carbonates under alkaline conditions (> pH 10) provided by natural photosynthesis, similar to that when using caustic chemicals for conventional water softening. This is the first time that microalgae are shown to soften hard groundwater via naturally occurring photosynthesis. This is significant since this process does not require the use of caustic chemicals and the whole drainage of the used bio-agents. The initial viable cells accounted for 89 ± 4% of the total cell count, and a negligible change in the percent viability of algal cells was found after 24-h of algal softening under the given conditions. This indicates that groundwater treatment could be continued without an undesirable loss of algal biomass as renewable bio-agents. The size and granularity of the algal cells decreased below those that were observed for the initial culture, likely due to cell division yielding new algal cells with simpler internal structure (Supporting Information Fig. S2). This observation coincided with increases in cell density (OD680) and chlorophyll content (FL3) during algal softening.

Fig. 1.

Variations in the optical density (OD680), pH, hardness (as CaCO₃), and alkalinity (as CaCO₃) during the algal treatment of groundwater under continuous illumination (150 μmol photons m⁻² s⁻¹) for 24 h.

Fig. 2.

Energy-dispersive X-ray (EDX) spectra and field emission scanning electron microscopy (FE-SEM) images of bio-solids obtained at the beginning (a, 0 h) and end (b, 24 h) of the algal softening of groundwater using S. quadricauda.

Algal softening achieved 54 % removal of the alkalinity, which was much higher than the theoretical percent removal (28 %) calculated using Visual MINTEQ (model 3.0). A higher alkalinity removal than expected is not surprising and can be easily explained with inorganic carbon fixation in eukaryotic microalgae. It is worth noting that the ability of microalgae absorbing carbonate species as well as the elevation of the OH⁻ level in the aqueous phase is beneficial for subsequent oxidation processes that involve the generation of highly reactive radical intermediates. The OH radical as an intermediate is produced by O₃ decomposition, which can be regulated by the level of OH⁻ in water, and it is capable of rapidly reacting with a wide variety of organic and inorganic constituents in water (k = 10⁸ – 10¹⁰ M⁻¹ s⁻¹) (von Gunten, 2003). Although the radical-scavenging activity of the carbonate species increases with increasing pH (pKa (HCO₃⁻/CO₃²⁻) = 10.3) due to a stronger scavenger potential of CO₃²⁻ compared to HCO₃⁻, this partially offsets the more rapid rate of OH-induced initiation reaction at higher pH values (Singer and Reckhow, 2011). In this study, a significant removal of carbonate species was achieved via photosynthetic carbon reduction through metabolic pathways, in addition to the consumption of carbonates engaging in mineral carbonation under alkaline conditions. In contrast to algae-induced alkalization, chemical softening with NaOH removed only 25 ± 3% of the carbonate species in groundwater. Fig. 3 clearly shows the positive effect caused by an increase in the OH⁻ concentration prior to ozonation regardless of the softening methods used in this study, indicating a reduction in the required O₃ dose. All the softening was not effective in removing CBZ, but DCF decreased by more than 25 % with the algal treatment. The selected micropollutants were almost completely removed via alkaline ozonation at a dose of 1.25 ± 0.08 mg O₃ L⁻¹ or higher. The reactions of CBZ (8.8 × 10⁹ M⁻¹ s⁻¹) and DCF (7.5 × 10⁹ M⁻¹ s⁻¹) with OH radicals are over three orders of magnitude faster than those with O₃ (Huber et al., 2003). This suggests that the oxidation of these chemicals is enhanced at elevated pH values as alkaline ozonation expedites the decomposition of O₃ into OH radicals reacting at rates close to the diffusion-controlled limit with many organic contaminants. Likewise, the removal of NOM determined as DOC was also much greater when ozonation was applied to alkaline groundwater, no matter how the pH of the water increased (Fig. 4a). The observations support the fact that the pH affects the relative importance between the direct and indirect reactions. Our result was consistent with earlier reports that high degrees of mineralization can only be achieved at higher pH values, where the indirect reaction dominates (Gottschalk et al., 2010a).

Fig. 3.

Variations in the concentrations of CBZ and DCF via the ozonation of raw (GW) and groundwater softened using two different methods (n = 3). The difference in residual micropollutants between the algal-treated groundwater and the others is statistically significant (p < 0.05).

Fig. 4.

Effects of softening and subsequent ozonation on variations in the levels of DOC (a), UV₂₅₄ (b), SUVA (c), and AOC (d) (n = 6). Raw groundwater (GW) was softened with two different techniques, followed by ozonation using four different doses (0, 2, 5, and 10 %).

The quenching of OH radicals with NOM is usually more important than that with carbonate species (Crittenden et al., 2012), which drastically affects the required O₃ dose. The degradation of the selected micropollutants could be retarded by the presence of NOM competing for OH radicals. The rate constants for the reaction of OH radicals with NOM obtained from a wide variety of water sources ranged between 1.4 and 4.5 × 10⁸ M-C⁻¹ s⁻¹ (Westerhoff et al., 2007), which is relatively lower than that with the selected micropollutants close to the diffusion-controlled limit (i.e., ~10¹⁰ M⁻¹ s⁻¹). Unfortunately, the concentration of NOM (as DOC) in this study was over several orders of magnitude higher than that of the micropollutants targeted for destruction. Nevertheless, softening as a pretreatment procedure enhanced the degradation of the micropollutants upon ozonation in groundwater. The removal kinetics of both CBZ and DCF as a function of O₃ dose were the fastest in algal-softened groundwater, followed by chemically-softened groundwater (Fig. 3). In the context of the scavenger potential of NOM, it is worth noting that algal softening of groundwater achieved a dramatic decrease in the electron-rich moieties (estimated as SUVA) in the organic molecule prior to ozonation. The OH radical is a strong electrophile and preferably enters electron-rich sites (Albarran and Schuler, 2005; Latch, 2017) whereas the reaction rate of OH radicals tends to be low for aliphatic constituents without nucleophilic sites (Crittenden et al., 2012). Fig. 4 shows that the NOM decreased more when measured as UV₂₅₄ than as DOC. UV₂₅₄ is basically linked to the presence of aromatic groups with varying degrees of activation (Martin-Neto et al., 2009). SUVA has thus been adopted to estimate the organic aromaticity, which is generally proportional to the degree of condensation and is highly correlated with the extent of electron-rich sites such as aromatic functional groups and double-bonded carbon groups in an organic molecule (Kim et al., 2015b; Sillanpää et al., 2015). Therefore, our result indicates that the aromatic chromophores in the NOM molecules were significantly degraded but not completely decomposed by algal treatment, leading to the formation of “bleached” byproducts with no UV absorbance. The result of the SUVA measurements was consistent with concurrent observations using fluorescence spectroscopy (Fig. 5) and size-exclusion chromatography (Fig. 6) in characterizing NOM, as discussed later in this section. Consequently, these byproducts with lower SUVA values are less susceptible to attacks by OH radicals produced in alkaline ozonation, which could mitigate the inhibitory effect of NOM on the degradation of persistent micropollutant residues in groundwater. Similarly, the greater drop in UV₂₅₄ compared to DOC was also observed in the chemical softening of groundwater (Fig. 4). NOM is readily adsorbed onto mineral solids, such as CaCO₃ and MgCO₃ formed in water softening with typical caustic chemicals (Russell et al., 2009), and the removal of NOM tends to increase with an increase in the pH of water (Randtke, 2011). Some studies suggest that the aromatic fraction of organic matter is preferentially removed during softening (Kalscheur et al., 2006; Roalson et al., 2003).

Fig. 5.

Changes in the fluorescent characteristics of NOM by ozonation of raw (A0~10), chemically-softened (B0~10), and algal-softened (C0~10) groundwater. O₃-saturated H₂O dose (%, v/v) applied in the ozonation is denoted with the Arabic numerals on each label of the three-dimensional excitation-emission-matrix (EEM) fluorescence spectra.

Fig. 6.

LC-OCD (a) and -UVD (b) chromatograms of hydrophilic (HPI) NOM passing through a size exclusion chromatographic column. Organic compounds that do not elute during the given period of time for the analysis are not shown in the chromatograms, designated as hydrophobic (HPO) fraction. The DOC (c) and UV₂₅₄ (d) distributions of the NOM fractions are determined based on their hydrophobicity.

Oxidative bleaching, as discussed earlier, decreases the light absorption due to a split in the aromatic rings, which is consistent with the result shown for the fluorescence measurements of this study (Fig. 5). The algal softening of groundwater resulted in a substantial removal of humic-like fluorophores, indicating a decrease in the number of aromatic rings or conjugated bonds in a chain structure, or the conversion of a linear to a non-linear ring system (Kim et al., 2006). Similarly, ozonation was effective at extinguishing the humic-like fluorophores of NOM in raw and chemically-softened groundwater, but it did not completely remove them even with the highest O₃ dose examined in this study. A similar conclusion using surface waters was also drawn by Stylianou et al. (2018). The decrease in fluorescence intensity of NOM reflects either the degradation or sometimes mineralization of the corresponding fluorophores or functional groups that are present in the chemical structure of the fluorescent organic matter (Martin-Neto et al., 2009; Mostofa et al., 2013). Fig. 5 also shows that algal treatment newly formed tryptophan-like components typically originating with microbial activities. The result of preliminary tests for cultivating S. quadricauda in groundwater without adding purified humic acid revealed that any noticeable organic fluorophores did not appear, and a negligible change was also found in the DOC for the same period of cultivation (data not shown). The observations coincided with the result of the chromatographic analysis in this study (Fig. 6), showing the removal of UV₂₅₄-causing smaller organic molecules and the production of non-chromophoric biopolymers (e.g., polysaccharide-like substances) during algal softening of groundwater. Biopolymers are well known to be readily assimilated by microorganisms, unlike the highly UV-absorbing organic components that are designated for humic-like substances and/or building blocks (Kim et al., 2015b). Similarly, we have previously reported that biological treatment processes can remove tryptophan-like components to a much greater extent than humic-like components (Kim et al., 2016, 2015a), implying that the microalga under the given conditions is able to transform humic-like substances into more biodegradable forms. This result was consistent with the AOC measurements of this study (Fig. 4d). The concentration of the AOC increased tremendously with the algal treatment from 72 to 188 μg L⁻¹, which increased up to 220 μg L⁻¹ by subsequent ozonation. The oxidative reaction of O₃ and/or OH radicals with NOM induces molecular changes, but not mineralization (Supporting Information Fig. S5). These changes form the basis for the production of biodegradable metabolites and the formation of more hydrophilic organic components that tend to form fewer DBPs with chlorine disinfectant (Gottschalk et al., 2010b; Kim and Yu, 2005, 2007; Reckhow and Singer, 2011). In general, water supplies with high SUVA have greater chlorine demand, resulting in greater DBP formation (Edzwald and Tobiason, 2011). For the ozonation of raw and chemically-softened groundwater, a gradual but noticeable increase in the AOC was found as a function of the O₃ dose. The chemical softening of groundwater also resulted in an increase in the AOC to some extent, and an associated change was found in the LC-OCD chromatogram. The reason for this is unknown, but could be due to partial denaturation and/or decomposition of NOM in the high-pH environment associated with precipitative softening (Randtke, 2011). In addition to the recarbonation process following alkaline ozonation, a biologically-active process should be used as a complementary post-treatment (e.g., slow sand filtration) to remove AOC prior to supplying the produced water to the distribution system.

This study analyzed the fate of NOM using spectroscopy, chromatography, and bioassay, indicating that algae-induced alkalization of groundwater resulted in a significant increase in biodegradability of carbonaceous organic residues in groundwater. The chemical softening decreased the UV₂₅₄ of NOM to some degree, but to a far lesser extent than algal softening. This result suggests that a high pH was not a direct concern. We have reported other mechanisms with respect to algae-mediated organic deformation in our prior work (Kim, 2018). Briefly, the photostimulation of S. quadricauda has been shown to induce the degradation of humic-like substances that are typically resistant to microbial decomposition. This is not surprising at all when considering the exoelectrogenic activity of eukaryotic phototrophs (McCormick et al., 2011). Many redox enzymes are located at their surfaces. They could produce a variety of free radicals and other reactive species that can transform or degrade organic chemicals, although an understanding of the overall mechanisms related to electron movement from the cell surfaces to specific sites on NOM molecules is still lacking. Moreover, humic-like substances contain a variety of redox-active functional groups that are capable of accepting and donating electrons, which can result in an increase in organic radical contents (Aeschbacher et al., 2010). The growth of algae in light significantly changes both the water chemistry and algal physiological properties, readily affecting reactions with specific organic compounds. Further research is needed to explore the ways in which process sustainability could be enhanced through the design of an appropriate system for algal softening.

4. Conclusions

This study used a microalga S. quadricauda in the softening of groundwater via natural photosynthesis in light. This alternative strategy of increasing the pH of water was combined with subsequent ozonation to enhance the degradability of persistent micropollutants in groundwater. The algae-induced alkalization of groundwater accelerated the removal of micropollutant residues (CBZ and DCF) upon ozonation, probably due to (1) the formation of OH radicals directly from O₃ molecules under alkaline conditions (pH > 10); (2) > 50 % removal of strong radical-scavenging carbonate species (HCO₃⁻ and CO₃²⁻) through the pathways of abiotic precipitation and metabolic CO₂ reduction; (3) a significant decomposition of NOM competing for OH radicals. Algal softening followed by ozonation achieved a better removal of NOM and dramatically decreased the electron-rich moieties in the organic molecule, which can consequently guarantee a reduction in the formation of harmful byproducts during disinfection of the produced water with chlorine. Interestingly, the algal softening of groundwater resulted in a significant increase in the biodegradability of carbonaceous organic residues in groundwater. This is crucial for many cases to treat groundwater in order to increase the reliability of water treatment systems by combining it with complementary post-treatment technologies (e.g., biological filtration). The softening process characterized in this study uses living algal components that are capable of self-repair and reproduction. Additional research is necessary to fully exploit the mechanisms involved in the algae-induced organic decomposition observed in this study. More work should also be performed using a wide range of micropollutants to generalize the findings of such a study, along with economic assessments.