Natural organic matter in drinking water — The “NOM-typing project”, background and basic characteristics of original water samples and NOM isolates

The complex nature of natural organic matter (NOM), and the impact of this matter on drinking water quality, necessitate international co-operative efforts. After several decades of struggle, many experienced NOM scientists world wide recognize the need for international sample references, or better yet, to work together on the same set of samples. The aim of the “NOM-typing project” was a multi-method characterization of a limited number of NOM samples, isolated from different water sources. The background of the project and the nature of the sampling sites are described here, and some data comparing the composition of the samples, before and after the isolation, are presented. The techniques used for isolation were reverse osmosis (RO) and evaporation (EVA). The NOM samples were isolated from eight different locations in the southern part of Norway. The nature of the corresponding catchments differs, however, all samples were from areas with no influence of agriculture or local industry. The key issue of the project was that all samples were isolated with exactly the same methods and by the same team. In spite of the uniformity of the isolation technique used for the nine samples, the recoveries of the material differ from sample to sample. Generally, the loss of NOM is in the range of 10%, however, for one sample, the loss was as high as 35%. These discrepancies are probably partly because of differences in the nature of the NOM and partly due to differences in the general composition of the ambient water. There are a few remarkable differences between the nine samples. There was a high percentage of ash in isolates from acidified areas. These are also the samples where the loss of NOM during the isolation process was highest. The density of the RO-isolates differs 10-fold between the samples. The lipophilisity of the NOM samples, expressed as relative solubility in octanol, differs more than 30-fold.     

Isolation of natural organic matter — The influence on the assimilable organic carbon

The biofilm formation potential (BFP), determined by the assimilable organic carbon (AOC), was measured in water samples containing natural organic matter (NOM), and in “identical” samples reconstructed from NOM isolated by reverse osmosis (RO) and dissolved in distilled water. The results showed identical BFP in natural and reconstructed NOM water samples within the standard deviation of the AOC measurements. These results indicate that the composition of the NOM is the single most important parameter for the BFP of natural waters, and that the RO isolation of NOM produces a powder which, after dissolution in distilled water, is representative for the natural water containing NOM with regard to BFP determined by AOC.     

Capillary zone electrophoretic studies on Norwegian surface water natural organic matter

Capillary zone electrophoresis (CZE) is a useful tool for the analysis of the electrophoretic behavior of anionic polyelectrolytes like humic substances. CZE was used to compare natural organic matter (NOM) with high ash content obtained by reverse osmosis (RO) and low temperature, low pressure evaporation (Ev.) from different Norwegian surface water sources. The quantitative relation between the resulting electrophoretic signals (peak height and area) and the carbon amount of the injected NOM samples was found linear. The NOM gave homogeneous signals in CZE, with a distribution of the detection signals around an average electrophoretic mobility (AEM) corresponding to the charge density distribution of the NOM, governed by the distribution of their molecular sizes and acidities. Like humic substances, NOM only appears as anions in capillary electrophoresis (CE), but the high ash content of these samples and the presence of metals strongly influenced their mobilities at lower pH. CZE could be used as a tool for the rapid evaluation of the average net charge and the molecular radius of the NOM at pH 5.1.     

Characterization of natural organic matter from eight Norwegian surface waters: Proton and copper binding

As a part of the International NOM-Typing Project, proton and copper binding studies were conducted on ultrafiltered (>1000 dal.) and dialysed (>100 dal.) natural organic matter (NOM) samples from eight Norwegian Lakes. The NOM samples were similar with respect to proton and copper binding, although minor differences were observed among the samples of different origin. Slight variations occurred for seasonal changes and lime additions. The proton binding pK_a values grouped in three distinct classes: pK_a1 = 4.20–4.28; pK_a2 = 6.61–6.87; and pK_a3 = 9.27–9.8. The average total proton binding site concentration of the NOM samples was: 12.7 meq/g total organic carbon (TOC). Two binding sites were established for copper with average log K1 = 4.84 and log K2 = 7.17. The average site concentration for samples isolated by reverse osmosis (RO) was 0.56 μmol/mg C. Copper binding characteristics appear to be influenced by the method of sample isolation. For NOM isolated by low pressure evaporation (evap), binding site concentrations are smaller, conditional stability constants are higher, variability of the copper complexation capacity is smaller, and correlation with elemental composition (C, N, H) is better than in the case of NOM isolated by RO.     

Basic characterization of Norwegian NOM samples — Similarities and differences

NOM (natural organic matter) samples, isolated by reverse osmosis (RO), and nine NOM samples, concentrated by low-pressure, low-temperature evaporation (EVA) from various origin in Norway, were characterized by different methods. The methods included elemental content of the freeze-dried samples (C, H, N, and O), elemental content of the redissolved samples (DOC, Na, Ca, Cu, Fe, and Mn), specific spectral absorbance at λ = 254 nm (A(254 nm)/DOC) and at λ = 436 nm (A(436 nm)/DOC), UV/Vis-spectra, specific fluorescence (excitation and emission spectra), proton capacity, Cu-complexation capacity, and content of hydrolysable amino acids and carbohydrates. In addition, chromatographic characterization by gel chromatography and reversed-phase chromatography was done. Studies on adsorption and flocculation abilities and the formation of trihalomethanes (THM) and organic halogens adsorbable on activated carbon (AOX) upon chlorination were also investigated. Comparison of the results showed that the samples were different according to the origin of the sample, isolation procedure, and time of sampling. The differences are not the same for all parameters used. Standardized procedures have to be applied to redissolve the organic material in water in order to compare data. For the samples investigated, it was shown that parameters like elemental analysis (H/C ratio), the specific spectral absorption in the UV range, proton capacity, complexation capacity and anionic particle charge, amount of hydrolysable amino acids and carbohydrates, adsorption behaviour on activated carbon, and the THM-FP are sensitive parameters to show differences concerning origin and isolation procedure.     

Fluorescence analysis for multi-site aluminum binding to natural organic matter

Natural organic matter (NOM) samples isolated from different water sources in Norway were compared using their fluorescence properties. Fluorescence surfaces were observed at pH 4.36 and deconvoluted using SIMPLISMA (Windig and Guilment 1991). There were a total of seven different fluorophores observed for these samples and each sampling site had between four and six of the fluorescent components. These components were observed to bind Al during titrations at the same pH. Multiresponse titration curves were fit using the method of Smith and Kramer (1998) and most of the binding strengths are similar to values for Suwannee River fulvic acid (logK' between 4.8 and 5.5), but there are strong sites (logK' = 7) and weak sites (logK' between 3 and 4) also observed. Results depended on the isolation method used; reverse osmosis and low pressure evaporation yielded different values but with no consistent trends.     

Biological vs. chemical properties of natural organic matter isolated from selected Norwegian lakes

Relationships among different spectroscopic, chemical, and biological properties of dissolved natural organic matter (NOM) were investigated. NOM samples were isolated by means of reverse osmosis from eight Norwegian lakes. Elemental composition, absorption spectra (UV, VIS, IR), the influence on algal growth, and accumulation rates of metals by mussels in the presence of NOM were measured. Linear regression and principal components analyses were utilised to assess relationships between the biological and chemical properties. It was concluded that the content of organic matter in the isolates and the aromaticity of NOM, as characterised by UV and VIS spectra, are best correlated with the biological properties.     

Characterization of natural organic matter from eight Norwegian surface waters: The effect of ash on molecular size distributions and CHN content

Natural organic matter (NOM) was isolated by reverse osmosis (RO) from eight surface waters in southern Norway. The freeze-dried samples were reconstituted in deionized water and part of the sample was dialysed against HCl. Samples of both the dialysed and non-dialysed materials were then ultrafiltered to give fractions: >50 000 dal., <50 000 but >10 000 dal. and <10 000 dal. nominal molecular weights (NMW). Dialysis against acid reduced the ash content of the samples by 68±14%. Non-dialysed NOM has a preponderance of carbon in the largest fraction (52.7±30%), while after dialysis, the smallest fraction of NOM has the greatest carbon content (52.2±9.7%). In addition to altering the overall apparent size distributions of the NOM fractions, dialysis generally reduces the amount of nitrogen relative to carbon in the NOM samples. The latter observation appears to be the result of losses of small molecular size organic matter that is enriched in nitrogen.     

Multi-site proton interactions with natural organic matter

Water samples from various locations in Norway were used to isolate natural organic matter (NOM). The NOM was isolated using both reverse osmosis and low-pressure low-temperature evaporation for each sample site. These samples were titrated from −log[H⁺] 3 to 11 at 0.1 unit intervals. The data were analyzed using the Discrete Site Analysis (DISI) technique for pK_a intervals of 0.2. The acidity constants are grouped into four classes: strong (pK_a < 5), intermediate strong (5.1 < pK_a < 7.5), intermediate weak (7.6 < pK_a < 9.2), and weak (pK_a > 9.3). All samples, regardless of isolation method, were found to contain strong and weak ligands, along with some intermediate ligands. For the same sample site, the concentration of these ligands were found to be dependent on isolation method and titration direction (acid then base vs. base then acid). In addition, the concentration and classes of ligands present were found to vary between sample sites. Suwannee River fulvic acid was analyzed as a reference sample, and was qualitatively similar to the NOM samples but quantitatively different. Overall, the differences in pKa spectra due to isolation method and/or titration direction are almost as significant as differences between sample location, but there are no consistent trends in effects of isolation method or titration direction on characterization of NOM.     

Changes of AhR-mediated activity of humic substances after irradiation

Humic substances (HS) and natural organic matter (NOM) are natural organic compounds ubiquitous in the environment. However, some studies indicate that both HS and NOM can act as xenobiotics, e.g. induce hormone-like effects in fish, amphibians and invertebrates. Molecules of these substances contain a number of aromatic rings and conjugated double bonds--the so called chromophores. Irradiation of dissolved HS and NOM can lead to a series of photochemical reactions which can act on these substances itself, or on other substances present in aquatic environment along with HS and NOM such as e.g. xenobiotics. In our previous study, we have found significant interactions of five humic acids (HA) with cytosolic aryl hydrocarbon receptor (AhR) in an in vitro bioassay based on H4IIE-luc cells. In the present study, we have studied the changes in AhR-mediated activities both of HS and NOM after irradiation that simulated natural solar light. Nine different HS and two NOM samples were irradiated in Pyrex tubes with a medium-pressure mercury lamp for a duration of 0 to 52 h (which corresponds to 0-52 d natural solar radiation). Original concentrations of the samples were 50 mg L(-1), and the greatest concentration of HS and NOM photoproducts subsequently tested in the bioassay was 17 mg L(-1), which is an environmentally relevant concentration. After irradiation the absorbances of all the samples were less than the original materials. The AhR-mediated activity of the HA-Fluka and HA Sodium Salt were partially decreased by irradiation. The activities of other HS and NOM, that were either AhR-active or -inactive were not changed by irradiation. The results of the study demonstrate that AhR-mediated activities of two active HA is caused by both photo-stable and photo-labile AhR activators, while the other three active HA contain only photo-stable AhR activators. Potential mechanisms of the observed irradiation-induced changes in AhR-mediated activities are discussed.     

Monitoring natural organic matter in water with scanning spectrophotometer

This study uses scanning ultraviolet-visible (UV-Vis) spectrophotometer to monitor natural organic matter (NOM) in water. The results showed that the area under the UV-Vis spectra is a good surrogate to monitor the concentration of the aqueous NOM. No apparent difference was found between the spectra of the three commercially available humic acids used in this study. The use of the scanning spectra compensates the potential random error in the absorbance determined at single wavelength due to the heterogeneous NOM compositions. The major interference of the proposed method comes from the formazine turbidity and nitrate nitrogen (NO3- -N). Although filtration with a 0.45-microm filter can remove most of the interference from formazine, some interference is still present at a wavelength less than 250 nm. Also it is found that the presence of the NO3- -N greatly affect the spectra of the NOM. In order to monitor the NOM in water with minimized interference, it is recommended that the area under the spectra between 250 and 350 nm should be used as a surrogate for concentration of NOM in water.     

Removal of NOM in the different stages of the water treatment process

Natural organic matter (NOM) is abundant in natural waters in Finland and in many ways affects the unit operations in water purification. In this study, the organic matter content in water in different stages of a full-scale treatment process over 1 year was measured. The full-scale treatment sequence, studied at the Rusko water treatment plant in Tampere, Finland, consisted of coagulation, flocculation, clarification by sedimentation or flotation, activated carbon (AC) filtration, and disinfection. High-performance size exclusion chromatography (HPSEC) was used for separation to determine changes in the humic substances content during the purification process. In addition, total organic carbon (TOC), KMnO4-number, and UV-absorbance at wavelength 254 nm (UV254) were measured. High molecular weight (HMW) matter was clearly easier to remove in coagulation and clarification than low molecular weight (LMW) matter. Furthermore, depending on the regeneration of the activated carbon filters, activated carbon filtration was effective to a degree but did not remove most of the lowest molecular weight compounds. Significant correlation was established among HPSEC, KMnO4, UV254 absorbance, and TOC. HPSEC proved to be a fast and relatively easy method to estimate NOM content in water and, in fact, gave more information than traditional methods on the type of NOM in a water sample. It also helped the process performance follow-up.     

Pre-assessment of the speciation of 60Co, 125Sb, 137Cs and 241Am in a contaminated aquifer

A sample of contaminated groundwater was analyzed using a combination of wet techniques to obtain geochemical information on the mobile species of 60Co, 125Sb, 137Cs and 241Am. The techniques were combined in a scheme to determine the predominant character of the radionuclides in negative or positive fractions, size separation by ultrafiltration, and their association with natural organic matter (NOM). The analyses indicated that the radionuclides of interest were predominantly in the negative fraction. Most of the 60Co and 125Sb were in the small size fraction (<5000 Da), and 137Cs and 241Am were found with the larger, colloidal-sized material. Antimony-125 and 60Co were predominantly in the hydrophilic fraction, while 137Cs and 241Am were found in hydrophobic fractions. Our analysis indicated that 137Cs is found in the same fraction as the large-sized colloidal (hydrophobic) material, suggesting an association with NOM. The results suggested that 60Co and 241Am were associated with NOM, in different size fractions, suggesting that these two nuclides are bound to different sites. Finally, the 125Sb results were inconclusive, whether this nuclide is associated with NOM, or it is inorganic.     

Basic characterization of reference NOM from Central Europe — Similarities and differences

The FA and HA fractions represent 20 to 50% of the DOC of the original samples, depending on the origin, and serve as reference materials. The original aqueous samples and the aqueous solutions of the reference materials were characterized with methods suited for measuring samples at natural concentrations. The DOC normalized values of the spectral absorbances (λ = 254 nm; λ = 436 nm) in the original water samples were highest for the brown water and brown coal water and lowest for the secondary effluents. The FA fractions were similar in the UV-absorbance with the exception of the low-absorbing groundwater FA, reflecting a low density of double bonds. The DOC-normalized proton capacity of the FA fractions (10 to 20 μmol/mg) was highest for the brown coal, brown water, and soil-derived samples and again lowest for the groundwater FA. The high values are due to a higher contribution of the weak acids of phenolic type. Specific data for copper complexation capacity (0.5 to 2.5 μmol/mg) went parallel to the proton capacity. The characterization of the FAs by gel chromatography with organic carbon detection revealed three major fractions. The refractory part at short retention time dominated in the brown water and soil-derived samples, whereas the fractions for the lower molecular mass acids at long retention times were dominant in the secondary effluent and the brown coal isolate. Comparison with UV-detected chromatograms revealed a strong decrease in the specific absorbance of the secondary effluent. Up to 3% of the total mass of the FAs could be identified as hydrolyzable amino acids with the highest yield in the sample from secondary effluent and the lowest one in the groundwater sample. The nitrogen content was always higher in the FAs compared to the HAs of the same origin. The application of gel chromatography fluorescence and fluorescence decay time distribution analysis turned out to be a useful tool to differentiate between the different fractions of NOM, the isolates obtained by different methods (e.g., XAD-8, membrane separation) and the fractions of the samples from different origin and age. This is of special value, for the information can also be obtained for samples with natural NOM concentrations, even though further investigations required for full understanding of the data. This work is part of a multi-disciplinary research program (ROSIG: Refraktäre Organische Säuren in Gewässern), funded by the German Research Association (DFG).     

Transport of NOM and trace metals through macropores in the Lake Skjervatjern catchment

In the dystrophic Lake Skjervatjern, located at the west coast of Norway, periods with high precipitation coincide with an increase in the concentrations of natural organic matter (NOM) and several trace elements. The lake has no visual inlets, and, during high flow periods, a major part of the drainage follows macropores, entering the lake 10–40 cm below the surface. Water from the lake and the macropores were fractionated according to size by means of a hollow fiber ultra-filtration technique, during two periods with high precipitation. In the macropores, the high molecular weight form of NOM (nominal molecular cut off > 10 kDa) appears to be an important transporting agent for several trace elements that strongly interact with NOM (e.g., Fe, Al, Pb, Sm, La, Ce, and Eu). In contrast, anionic (V and As) and acid sensitive (Mn and Cs) elements were, generally, at a low molecular weight form (nominal molecular cut off < 3 kDa). The latter showed similar or lower concentrations in the macropores compared to the lake water, and NOM were apparently of small importance as a carrier for V, As, Mn, and Cs. Differences in concentration and size distribution of trace elements between the macropores and the lake basins indicate that parts of NOM and associated trace elements that enter the lake in a high molecular weight form, undergo aggregation and gravitational settling and thereby become a part of the lake sediments.     

Characterization of Norwegian natural organic matter: Size,diffusion coefficients,and electrophoretic mobilities

Eighteen Norwegian NOM samples were analyzed by AFM, TEM, FCS, and CE. The TEM and AFM gave complementary, but not identical, information regarding the conformation of the NOM. A large majority of the material in all samples appeared as points with a size of approximately 2 nm or less. FCS gave values for diffusion coefficients which were in the range 2.1–3.0 × 10⁻¹⁰ m² s⁻¹. This corresponds to molecular diameters of between 1.6 and 2.0 nm, in reasonable agreement with both TEM and AFM. Electropherograms, using absorbance at 200 nm, were all dissimilar except for the Gjerstad and Hellerudmyra sample pairs, for which no major differences were observed. CE was also carried out using both fluorescence (excitation at 325, 457, and 488 nm) and UV-absorbance (200, 210, 254, and 288 nm) detection on three of the NOM samples. The different modes of detection and wavelengths gave qualitatively similar electropherograms. Calculated EPMs of the major sample components were in the range -2.0 to -5.0 10⁻⁸ m² s⁻¹ V⁻¹.     

Removal of trihalomethane precursors by reverse osmosis

The effect of pre- and postchlorination on reducing trihalomethane (THM) precursors by reverse osmosis (RO) was investigated. Prechlorination of the feedwater, in contrast to postchlorination of the permeate, resulted in improved reduction of maximum total THM potential (MTP) due to adsorption. No significant difference in permeate MTP concentration occurred between the prechlorinated feed and the postchlorinated permeate. Recirculating permeate water through the RO unit was successful in restoringv flux rates and resuspending the precipitate that accumulated on the RO membranes. The small loss in free chlorine residual indicated prechlorination may be economically attractive in controlling biological fouling of cellulose acetate membranes. Reverse osmosis was not reliable in removing THM precursors and failed to produce a permeate with MTP concentration less than the maximum contaminant level of 0.10 mg/L.     

Arsenic and fluoride removal from groundwater by reverse osmosis

The reverse osmosis process was evaluated for removal of naturally occurring arsenic and fluoride from groundwater. Arsenic removal was affected by the prevalent arsenic species present in the water. Arsenic concentrations were reduced by approximately 60%–90% from nearly 80 μg/L. Fluoride concentrations were reduced by approximately 60% from nearly 1.7 mg/L.