Tracking the behavior of different size fractions of dissolved organic matter in a full-scale advanced drinking water treatment plant

Tracking the changes of dissolved organic matter throughout the city water supply system with optical indices

Dissolved organic matter (DOM) is important in determining the drinking water treatment and the supplied water quality. However, a comprehensive DOM study for the whole water supply system is lacking and the potential effects of secondary water supply are largely unknown. This was studied using dissolved organic carbon (DOC), absorption spectroscopy, and fluorescence excitation-emission matrices-parallel factor analysis (EEM-PARAFAC). Four fluorescent components were identified, including humic-like C1-C2, tryptophan-like C3, and tyrosine-like C4. In the drinking water treatment plants, the advanced treatment using ozone and biological activated carbon (O3-BAC) was more effective in removing DOC than the conventional process, with the removals of C1 and C3 improved by 17.7%-25.1% and 19.2%-27.0%. The absorption coefficient and C1-C4 correlated significantly with DOC in water treatments, suggesting that absorption and fluorescence could effectively track the changes in bulk DOM. DOM generally remained stable in each drinking water distribution system, suggesting the importance of the treated water quality in determining that of the corresponding network. The optical indices changed notably between distribution networks of different treatment plants, which enabled the identification of changing water sources. A comparison of DOM in the direct and secondary water supplies suggested limited impacts of secondary water supply, although the changes in organic carbon and absorption indices were detected in some locations. These results have implications for better understanding the changes of DOM in the whole water supply system to help ensure the supplied water quality.

Formation and transformation of pre-chlorination-formed disinfection byproducts in drinking water treatment process

As pre-chlorination is increasingly adopted in drinking water treatment plant (DWTP), an attractive question emerged: how the disinfection by-products that formed during pre-chlorination (preformed DBPs) would be transformed in the drinking water treatment process? This study investigated the DBP formation kinetics and molecular characteristics in chlorinated source water, DBP transformation and removal in practical DWTP. It was found that the formation of trihalomethanes (THMs) followed pseudo first-order kinetic model and the intensified Br- exposure facilitated the transformation of TCM into TBM. As Br- concentration shifted from 0.5 mg L-1 to 2.0 mg L-1, the predicted maximum yield of TBM was doubled to 53.7 μg L-1 with the increase of formation rate constant (k-value) from 0.249 h-1 to 0.336 h-1. Besides known DBPs, the molecular-scale investigation unveiled that the preformed unknown Cl-DBPs were a cluster of unsaturated aromatic DBPs ((DBE-O)/Cwa = 0.16, AImod, wa = 0.36) with high H/C (H/Cwa = 1.25). Pre-ozonation exhibited a preferential removal pattern towards condensed aromatic preformed Cl-DBPs with high H/C (AImod ≥ 0.67, H/C > 1.2 and O/C < 0.3). However, the removal of Cl-DBPs in coagulation-clarification process was limited with 56 more unknown Cl-DBP formulas identified. O3-biological activated carbon process exhibited effective removal of preformed DBPs featured with low MW (carbon number ≤ 13), high unsaturation (DBE ≥ 7), condensed aromaticity (AImod ≥ 0.67), and higher H/C (H/C > 1.6). When the pre-chlorination process is adopted, the removal of preformed DBPs during the conventional treatment process is limited, while advanced treatment process can effectively remove these preformed DBPs.

Characterization of dissolved organic matter for understanding the adsorption on nanomaterials in aquatic environment: A review

Nanomaterials (NMs) have received tremendous attention as emerging adsorbents for environmental applications. The ever-increasing release into aquatic systems and the potential use in water treatment processes heighten the likelihood of the interactions of NMs with aquatic dissolved organic matter (DOM). Once DOM is adsorbed on NMs, it substantially modifies the surface properties, thus altering the fate and transport of NMs, as well as their toxic effects on (micro)organisms in natural and engineered systems. The environmental consequences of DOM-NMs interaction have been widely studied in the literature. In contrast, a comprehensive understanding of DOM-NM complexes, particularly regarding the controlling factors, is still lacking, and its significance has been largely overlooked. This gap in the knowledge mainly arises from the complex and heterogeneous structures of the DOM, which prompts the urgent need to further characterize the DOM properties to deepen the understanding associated with the adsorption processes on NMs. This review aims to provide in-depth insights into the complex DOM adsorption behavior onto NMs, whether they are metal- or carbon-based materials. First, we summarize the up-to-date analytical methods to characterize the DOM to unravel the underlying adsorption mechanisms. Second, the key DOM characteristics governing the adsorption processes are discussed. Next, the environmental factors, such as the nature of adsorbents and solution chemistry, affecting the DOM-NM interactions, are identified and discussed. Finally, future studies are recommended to fully understand the chemical traits of DOM upon its adsorption onto NMs.

Exploring the relative changes in dissolved organic matter for assessing the water quality of full-scale drinking water treatment plants using a fluorescence ratio approach

This study aims to extend and demonstrate the application of fluorescence spectroscopy for monitoring the water quality of three differently operated full-scale drinking water treatment plants located in the Shenzhen city (China). A ratio of fluorescent dissolved organic matter (FDOM), which describes relative changes in humic-like to protein-like fluorescence, was used to explain mechanisms behind the physicochemical processes. The fluorescence components obtained through individual and combined parallel factor analysis (PARAFAC) modeling revealed the presence of humic-like (C1) and protein-like (C2) structures in the DOM. The C1/C2 ratio provided a direct relationship between the seasonal variations and DOM composition. Wet season generated DOM enriched with humic-like fluorescence, while dry season caused a higher release of protein-like fluorescence. The fluorescence ratio presented unique patterns of DOM in treatment trains. The chemical pretreatment and disinfection unit processes showed a higher tendency to remove the humic-like fluorescence. However, the C1/C2 ratio increased during physical treatment processes such as coagulation-precipitation and sand filtration, indicating preferential removal of protein-like fluorescence. The DOM composition in influent directly (R² = 0.77) influenced the relative intensities of fluorescence components in the treated water. Compared to the dry season, the wet season caused significant changes in DOM composition and produced treated water enriched with humic-like fluorescence. This fluorescence ratio offers an approach to explore the role of different treatment units and determine the factors affecting the composition of DOM in the surface water and drinking water treatment plants.

Seasonal occurrence of N-nitrosamines and their association with dissolved organic matter in full-scale drinking water systems: Determination by LC-MS and EEM-PARAFAC

N-nitrosamines have been identified as emerging contaminants with tremendous carcinogenic potential for human beings. This study examined the seasonal changes in the occurrence of N-nitrosamines and N-nitrosodimethylamine formation potential (NDMA-FP) in drinking water resources and potable water from 10 drinking water treatment plants in a southern city of China. The changes in N-nitrosamines are well correlated with dissolved organic matter (DOM), particularly fluorophores, which were measured and compared between traditional fluorescence indices and excitation-emission matrix coupled with parallel factor analysis (EEM-PARAFAC). Four of N-nitrosamine species including N-nitrosodimethylamine (NDMA), N-Nitrosodibutylamine (NDBA), N-Nitrosopyrrolidine (NPYR), and N-Nitrosodiphenylamine (NDPhA) are found to be abundant compounds with an average of 29.5% (26.7%), 20.0% (25.2%), 18.9% (16.0%), and 9.0% (9.9%) in the source (and treated) water, respectively. The sum of N-nitrosamines concentration is recorded to be low in the wet season (July-September), whereas the dry season (October-December) provided opposite impacts. EEM-PARAFAC modeling indicated the predominance of humic-like component (C1) in the wet season while in the dry season the water was dominant in protein-like component (C2). All the N-nitrosamines excluding NDPhA and N-Nitrosomorpholine (NMOR) showed a strong association with protein-like component (C2). In contrast, humic-like C1, which was directly influenced by rainfall, was found to be a suitable proxy for NMOR and NDPhA. The results of this study are valuable to understand the correlation between different N-nitrosamines and DOM through adopting fluorescence signatures.

Evaluating the contributions of different organic matter sources to urban river water during a storm event via optical indices and molecular composition

Dissolved organic matter (DOM) in river water dynamically changes with respect to its major sources during heavy rain events. However, there has been no established tool to estimate the relative contributions of different organic sources to river water DOM. In this study, the evolution in the contributions of ten different organic matter (OM) sources to storm water DOM was explored with a selected urban river, the Geumho River in South Korea, during storm events via an end-member mixing analysis (EMMA) based on fluorescence indices and Fourier-transform ion cyclotron resonance mass spectrometry (FT-ICR-MS). The OM source materials included treated effluent, road runoff, groundwater, topsoil, deep soil, leaves, reeds, riparian plants, attached algae, and suspended algae. The EMMA results provided quantitative estimates of the variations in the dominant OM sources with the progress of storms. Treated effluent was the prevailing source at the beginning period of the storm, while topsoil, leaves, riparian plants, and groundwater predominated during and after the peak period. The fluorescence indices-based evaluation was consistent with the statistical comparison of the molecular formulas derived from FT-ICR-MS conducted on the ten potential OM sources and the storm samples. The observed variations in the OM sources agreed with the typical characteristics of urban rivers in connection with anthropogenic inputs and the impact of surrounding impervious surfaces. This study demonstrates the application of intuitive and facile tools in estimating the relative impacts of OM sources in urban watersheds.

Wildfires Alter Forest Watersheds and Threaten Drinking Water Quality

Wildfires are a natural part of most forest ecosystems, but due to changing climatic and environmental conditions, they have become larger, more severe, and potentially more damaging. Forested watersheds vulnerable to wildfire serve as drinking water supplies for many urban and rural communities. The highly variable nature of wildfire behavior combined with spatially complex patterns in vegetation, landscape, and hydrologic factors create uncertainty surrounding the postfire effects on water supplies. Wildfires often cause dramatic changes in forest vegetation structure and soil conditions, and alter the watershed processes that control streamflow, soil erosion, nutrient export, and downstream water chemistry. The authors' work centers on field and laboratory studies to advance knowledge of postfire changes in soil and water chemical composition that influence drinking water treatment. High intensity postfire rainstorms typically increase runoff that erodes ash and soil from burned landscapes and dramatically elevates turbidity, nutrient, and dissolved organic carbon (DOC) levels in surface waters, which can cause short-term challenges for water providers. There is also growing evidence that water quality impacts can persist after high severity fires due to slow vegetative recovery, and nitrogen and DOC have remained elevated for 15 years following high severity fire. Low-moderate temperatures during wildfire may also influence water quality. Research by the authors showed that the solubility of organic matter, and C and N released from soils increased following soil heating at temperatures ≤ 350 °C. Further, the water extracted organic matter from soils heated at 225-350 °C included higher proportions of condensed aromatic structures, such as black carbon and black nitrogen. Short-term postfire water quality degradation following high intensity rainstorms can force water treatment plants to shut down or can significantly challenge treatment process performance. Extreme turbidity and high DOC in poststorm water, coupled with compositional organic matter changes, reduced the coagulation efficiency of postfire water supplies. Field and lab-based studies documented the formation of small, aromatic soluble compounds during wildfire that contribute to inefficient DOC removal from postfire stormwater. Due to increased postfire DOC concentrations, and poor treatability of poststorm runoff, toxic disinfection byproduct (DBP) formation increased during water treatment. Exceedance of drinking water standards for the carbonaceous DBPs, trihalomethanes and haloacetic acids, may present a critical management concern for water providers following wildfires. Further, postfire formation of nitrogen compounds and increased nitrogenous DBP precursors for haloacetonitriles and chloropicrin were discovered. N-DBPs pose a public health concern due to their toxicity, and water providers should be aware of potential increases in N-DBP formation following fire. Evidence from the authors' studies demonstrates that even partially burned watersheds and wildfires burning at moderate temperature can have significant, lasting effects on C and N exports, source water quality, drinking water treatability, and DBP formation. Both short- and long-term postfire water quality impacts can create challenges for drinking water providers as they confront variability in supply and treatability. Communities, forest managers, and potable water providers will need to adapt to more frequent, destructive wildfires and anticipate greater variability in water quality.

Using fluorescence surrogates to track algogenic dissolved organic matter (AOM) during growth and coagulation/flocculation processes of green algae

Tracking the variation of the algogenic organic matter (AOM) released during the proliferation of green algae and subsequent treatment processes is crucial for constructing and optimizing control strategies. In this study, the potential of the spectroscopic tool was fully explored as a surrogate of AOM upon the cultivation of green algae and subsequent coagulation/flocculation (C/F) treatment processes using ZrCl₄ and Al₂(SO₄)₃. Fluorescence excitation emission matrix coupled with parallel factor analysis (EEM-PARAFAC) identified the presence of three independent fluorescent components in AOM, including protein-like (C1), fulvic-like (C2) and humic-like components (C3). Size exclusion chromatography (SEC) revealed that C1 in AOM was composed of large-sized proteins and aromatic amino acids. The individual components exhibited their unique characteristics with respect to the dynamic changes. C1 showed the highest correlation with AOM concentrations (R² = 0.843) upon the C/F processes. C1 could also be suggested as an optical predictor for the formation of trihalomethanes upon the C/F processes. This study sheds a light for the potential application of the protein-like component (C1) as a practical surrogate to track the evolution of AOM in water treatment or wastewater reclamation systems involving Chlorella vulgaris green algae.

Further insight into the roles of the chemical composition of dissolved organic matter (DOM) on ultrafiltration membranes as revealed by multiple advanced DOM characterization tools

This study assessed the relative contributions of different constitutes in dissolved organic matter (DOM) with two different sources (i.e., urban river and effluent) to membrane fouling on three types of ultrafiltration (UF) membranes via excitation emission matrix - parallel factor analysis (EEM-PARAFAC), size exclusion chromatography (SEC), and Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR-MS). Two polyethersulfone membranes with different pore sizes and one regenerated cellulose membrane were used as representative hydrophobic (HPO) and hydrophilic (HPI) UF membranes, respectively. Although size exclusion effect was found to be the most prevailing rejection mechanism, the behaviors of individual fluorescent components (one tryptophan-like, one microbial-humic-like, and terrestrial humic-like) and different size fractions upon the UF filtration revealed that chemical interactions (e.g., hydrophobic interactions and hydrogen bonding) between DOM and membrane might play important roles in UF membrane fouling, especially for small sized DOM molecules. Based on the molecular level composition determined by FT-ICR-MS, the CHOS formula group showed a greater removal tendency toward the HPO membrane, while the CHONS group was prone to be removed by the HPI membrane. The changes in the overall molecular composition of DOM upon UF filtration were highly dependent on the sources of DOM. The molecules of more acidic nature tended to remain in the permeate of effluent DOM, while the river DOM was shifted into more nitrogen-enriched composition after filtration. Regardless of the DOM sources, the HPO membrane with a smaller pore size led to the most pronounced changes in the molecular composition of DOM.

Effects of COD/N ratio on soluble microbial products in effluent from sequencing batch reactors and subsequent membrane fouling

The relative ratios of chemical oxygen demand (COD) to nitrogen (N) in wastewater are known to have profound effects on the characteristics of soluble microbial products (SMP) from activated sludge. In this study, the changes in the SMP characteristics upon different COD/N ratios and the subsequent effects on ultrafiltration (UF) membrane fouling potentials were examined in sequencing batch reactors (SBR) using excitation emission matrix-parallel factor analysis (EEM-PARAFAC) and size exclusion chromatography (SEC). Three unique fluorescent components were identified from the SMP samples in the bioreactors operated at the COD/N ratios of 100/10 (N rich), 100/5 (N medium), and 100/2 (N deficient). The tryptophan-like component (C1) was the most depleted at the N medium condition. Fulvic-like (C2) and humic-like (C3) components were more abundant with N rich wastewater. Greater abundances of large size biopolymer (BP) and low molecular weight neutrals (LMWN) were found under the N deficient and N rich conditions, respectively. SMPs from various COD/N exhibited a greater degree on membrane fouling following the order of 100/2 > 100/10 > 100/5. C1 and C2 had close associations with reversible and irreversible fouling, respectively, while the reversible fouling potential of C3 depended on the COD/N ratios. No significant impact of COD/N ratio was observed on the relative contributions of SMP size fractions to either reversible or irreversible fouling potential. However, the COD/N ratios likely altered the BP foulants' composition with greater contribution of proteinaceous substances to reversible fouling under the N deficient condition than at other N richer conditions. The opposite trend was observed for irreversible fouling. Our results provided further insight into changes in different SMP constitutes and their membrane fouling in response to microbial activities under different COD/N ratios.

Fast tracking the molecular weight changes of humic substances in coagulation/flocculation processes via fluorescence EEM-PARAFAC

The removal of a commercial humic acid (HA) and changes in its chemical composition were examined for coagulation/flocculation (C/F) processes based on jar tests using two different coagulants at a wide range of pH. ZrCl₄ showed a better performance in eliminating HA than Al₂SO₄ with the same removal rates at lower dosages. The highest removal rates were found at a neutral pH range (5.0-6.5). The HA was further decomposed into three different humic-like components (C1, C2, and C3) by excitation emission matrix coupled with parallel factor analysis (EEM-PARAFAC). Although the removal rates of all three components generally followed those of dissolved organic carbon, the relative removals of the individual components depended on the coagulant's doses and the solution pH. The fluorescent components of five ultrafiltered size fractions of the HA revealed that the peak with a longer emission wavelength could be associated with larger sized molecules. The C1/C3 ratios of the size fractions exhibited a significant linear relationship with the logarithmic values of the average molecular weight (MW) measured by size exclusion chromatography, which made it possible to predict the HA MW value changes upon the C/F using EEM-PARAFAC alone. Irrespective of the coagulant types and the pH, larger sized HA molecules were removed to a greater extent than smaller sized fractions. The preferential removal was more pronounced for ZrCl₄ versus Al₂SO₄ and at a neutral pH range. Our study suggests a great potential of EEM-PARAFAC in fast tracking the MW of humic substances in conventional C/F processes.

Fate and fouling characteristics of fluorescent dissolved organic matter in ultrafiltration of terrestrial humic substances

Ultrafiltration (UF) membrane fouling caused by terrestrial input of dissolved organic matter (DOM), especially during high flood periods, is poorly understood. In this study, we examined the fouling characteristics of three different terrestrial humic substances (HS) on regenerated cellulose (RC) UF membranes with the pore sizes of 30 k-3 kDa via conventional bulk HS measurements as well as an advanced fluorescence spectroscopy. The fluorescence excitation-emission matrix coupled with parallel factor analysis (EEM-PARAFAC) identified one protein-like (C1) and three humic-like fluorescent components (C2-C4) from soil and leaf-derived HS. The fate of the different fluorescent components was individually tracked for the UF processes. The higher removal rates were found generally on the order of high molecular weight (HMW) C1 to smaller sized humic-like components (C4 > C3 > C2) regardless of the HS sources, implying the importance of HS molecular sizes on the UF operation. Among the humic-like components, C2 was the most associated with irreversible fouling, while other two humic-like components contributed more to reversible fouling. For soil-derived HS, C4 can be suggested as a good surrogate for membrane fouling, as evidenced by the highest correlation between the removal rates and the total fouling indices among the tested HS variables including conventional bulk parameters. Our study demonstrated a promising application of EEM-PARAFAC for probing membrane fouling of terrestrial DOM, which provided additional insight into the fate of different fluorescent components on the UF processes.