Trihalomethane formation potential of aquatic and terrestrial fulvic and humic acids: Sorption on activated carbon.

Science of the Total Environment 521–522 (2015) 293–304

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Trihalomethane formation potential of aquatic and terrestrial fulvic and humic acids: Sorption on activated carbon Mohamed Y.Z. Abouleish a, Martha J.M. Wells b,c,⁎ a b c

Department of Biology, Chemistry, and Environmental Sciences, American University of Sharjah, Sharjah, P.O. Box 26666, United Arab Emirates Center for the Management, Utilization, and Protection of Water Resources, Department of Chemistry, Tennessee Technological University, Cookeville, TN 38505, United States EnviroChem Services, 224 Windsor Drive, Cookeville, TN 38506, United States

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• Ability of surrogate parameters DOC and UVA 254 to predict THMFP was compared. • Better removal of THMFP at pH 6 than pH 9 attributed to electrostatic interactions. • THMFP can exist after UVA254 is depleted; no THMFP existed after DOC was depleted. • THMFP remaining when UVA254 was zero ranged from 0–24%. • Not all carbon-containing THMFPforming organic matter were detectable at UV254.

a r t i c l e

i n f o

Article history: Received 30 January 2015 Received in revised form 21 March 2015 Accepted 21 March 2015 Available online xxxx Editor: D. Barcelo Keywords: Disinfection by-products Drinking water treatment Chlorination Isotherms Factorial analysis Dissolved organic carbon

a b s t r a c t Humic substances (HSs) are precursors for the formation of hazardous disinfection by-products (DBPs) during chlorination of water. Various surrogate parameters have been used to investigate the generation of DBPs by HS precursors and the removal of these precursors by activated carbon treatment. Dissolved organic carbon (DOC)- and ultraviolet absorbance (UVA254)-based isotherms are commonly reported and presumed to be good predictors of the trihalomethane formation potential (THMFP). However, THMFP-based isotherms are rarely published such that the three types of parameters have not been compared directly. Batch equilibrium experiments on activated carbon were used to generate constant-initial-concentration sorption isotherms for well-characterized samples obtained from the International Humic Substances Society (IHSS). HSs representing type (fulvic acid [FA], humic acid [HA]), origin (aquatic, terrestrial), and geographical source (Nordic, Suwannee, Peat, Soil) were examined at pH 6 and pH 9. THMFP-based isotherms were generated and compared to determine if DOC- and UVA254-based isotherms were good predictors of the THMFP. The sorption process depended on the composition of the HSs and the chemical nature of the activated carbon, both of which were influenced by pH. Activated carbon removal of THM-precursors was pH- and HS-dependent. In some instances, the THMFP existed after UVA254 was depleted. © 2015 Elsevier B.V. All rights reserved.

⁎ Corresponding author at: EnviroChem Services, 224 Windsor Drive, Cookeville, TN 38506, United States. E-mail addresses: [email protected], [email protected] (M.J.M. Wells).

http://dx.doi.org/10.1016/j.scitotenv.2015.03.090 0048-9697/© 2015 Elsevier B.V. All rights reserved.

294

M.Y.Z. Abouleish, M.J.M. Wells / Science of the Total Environment 521–522 (2015) 293–304

1. Introduction Humic substances (HSs) are not inherently hazardous and actually have many beneficial environmental effects. HSs act as a reservoir of organic carbon in the soil as part of the global carbon cycle (Aiken et al., 1985), prevent soil erosion (Stevenson, 1982), slowly release plant nutrients (MacCarthy et al., 1990), reduce toxic levels of pollutants (Carter and Suffet, 1982; Li et al., 1997), act as oxidizing or reducing agents under specific environmental conditions and affect photochemical processes (Hu et al., 1999), and interact with metal ions, organic molecules, minerals, and microorganisms (Senesi et al., 1996). However, humic substances (HSs) are precursors for the formation of potentially hazardous disinfection by-products (DBPs) during chlorination of water (Richardson and Ternes, 2014). In 1974, Rook (1974) linked chlorination of natural organic matter (NOM) in water to production of trihalomethanes (THMs), prompting research of ways that reduce NOM prior to chlorination. One treatment method used to remove HSs from water is contact with activated carbon before disinfection with chlorine. Therefore, it is important to understand the sorption process of HSs by activated carbon to facilitate their removal and consequently reduce formation of disinfection by-products (DBPs), leading to more effective and economical treatment of water. General appreciation for the relationship between DBP formation and NOM exists, but more specific understanding of the trihalomethane formation potential (THMFP) of various components of NOM, and of the relationships with surrogate parameters used for predicting THM formation, is needed to improve water treatment. The objective of the present research is to study the influence of the chemical character of HSs during activated carbon sorption. Two hypotheses are tested in this research: (1) variation in origin (aquatic or terrestrial), type (fulvic or humic), and pH (6 or 9) affects adsorption of NOM on activated carbon, and (2) the surrogate parameters dissolved organic carbon (DOC) and ultraviolet absorbance at 254 nm (UVA254) do not always adequately predict the THMFP. To test the hypotheses of this research, a 23 factorial design matrix was developed to investigate the main effect variables—origin, type, and pH. Factorial analysis was previously applied by Platikanov et al. (2010) and Rodrigues et al. (2007) to identify factors relevant to THM formation upon chlorination. Three parameters were measured concomitantly in all samples—DOC, UVA254, and the THMFP. All parameters were expressed relative to initial state concentrations to facilitate direct comparisons among data. Well-characterized samples of organic matter from the International Humic Substances Society (IHSS)—representing differences in origin (aquatic and terrestrial), geographical source, and type (FAs and HAs) of HSs—were investigated at pH 6 and pH 9. Although the primary emphasis in this research is the prediction and interpretation of DBP formation in aquatic systems, it was considered important to investigate both aquatic and terrestrial fulvic and humic acids. Due to weathering and erosion, even surface waters are composed of mixtures of these components. Quoting Hayes and Clapp (2001): “Because HSs are found in all soils, it is inevitable that they will also be found in all waters.”

Table 1 Factorial experimental design matrix and sample concentration. Identity

Geographical sourcea

Typeb

Originc

pHd

DOC (mg C L−1)

NO-FA-6 NO-FA-9 SW-FA-6 SW-FA-9 PE-FA-6 PE-FA-9 SO-FA-6 SO-FA-9 NO-HA-6 NO-HA-9 SW-HA-6 SW-HA-9 PE-HA-6 PE-HA-9 SO-HA-6 SO-HA-9

1 1 2 2 3 3 4 4 1 1 2 2 3 3 4 4

−1 −1 −1 −1 −1 −1 −1 −1 +1 +1 +1 +1 +1 +1 +1 +1

−1 −1 −1 −1 +1 +1 +1 +1 −1 −1 −1 −1 +1 +1 +1 +1

−1 +1 −1 +1 −1 +1 −1 +1 −1 +1 −1 +1 −1 +1 −1 +1

6.0 8.3 6.3 29.5 11.2 17.5 17.4 18.5 7.5 8.1 7.6 22.5 11.7 6.6 6.2 6.6

a b c d

Nordic = 1; Suwannee = 2; Peat = 3; Soil = 4. Fulvic = −1; Humic = +1. Aquatic = −1; Terrestrial = +1. pH 6 = −1; pH 9 = +1.

HSs—were evaluated at pH 6 and pH 9 according to the factorial experimental design matrix presented in Table 1. FAs and HAs were prepared in the appropriate phosphate buffer (pH 6 or 9) dissolved in HPLC-grade water (Fisher Scientific Company, Fair Lawn, NJ). The dissolved organic carbon (DOC) background content in HPLC-grade water was determined to be 0.108 ± 0.054 mg C ∗ L−1. Sample concentrations ranged from 6–30 mg C ∗ L−1 (Table 1). The solutions were pressure filtered under nitrogen using a 0.45 μ nylon supported plain filter (142 mm, MSI, Westboro, MA). The same HSs and factorial experimental design (Table 1) studied in this research were investigated previously for correlation between specific trihalomethane formation potential (STHMFP) and specific ultraviolet absorbance (SUVA) (Abouleish and Wells, 2012). The solutions studied in the previous manuscript are equivalent to the control samples in this study, i.e., the samples to which no activated carbon was added. Selection of the values of 6 and 9 for the pH levels of the 2k factorial design was based on setting the levels as far apart as possible to reduce variance and increase the likelihood of measuring an effect (Anderson and Whitcomb, 2007) within appropriate operating boundaries anticipated for source water pH ranges2 at which activated carbon treatment might occur. 2.2. Adsorbent The activated carbon (Filtrasorb 400, F400) used in this research (Calgon Carbon Corporation, Pittsburgh, PA) was passed through a U.S. standard number 200 sieve and retained on a U.S. standard number 400 sieve resulting in particles ranging from 38–75 μm that were washed with HPLC-grade water until the supernatant was light black in color. The pulverized, cleaned carbon was dried in an oven (105 °C) for approximately 24 h. 2.3. Activated carbon adsorption

2. Materials and methods 2.1. Adsorbate In this manuscript, a humic substance (HS) will refer to either a fulvic acid (FA) or a humic acid (HA). Well-characterized samples of organic matter from the International Humic Substances Society (IHSS1)—representing differences in origin (aquatic and terrestrial), geographical source [Nordic Aquatic (NO), Suwannee River (SW), Pahokee Peat (PE) and Elliot Soil (SO)], and type (FAs and HAs) of 1

International Humic Substances Society (IHSS). http://www.humicsubstances.org/.

Activated carbon adsorption procedures were adapted from Schmit and Wells (2002). Sorption of the HSs was studied through batch equilibrium experiments that were conducted for each individual substance in contact with activated carbon. The samples were prepared by mixing 80 mL of filtered FA or HA solutions with varying dosages of activated carbon that ranged from 0 (control) to 64 mg (0, 4, 16, 32, and 64 mg) to generate constant-initial-concentration isotherms. The samples were agitated for 24 h, then filtered using 0.22 μ nylon supported 2 U.S. Geological Survey. Water Quality Watch. http://waterwatch.usgs.gov/wqwatch/? pcode=00400.


plain filters (25 mm, Osmonics, Inc., Fisher Scientific Company, NJ). Previous research (Schmit and Wells, 2002) demonstrated equivalent sorption isotherm results by this technique at 24 and 48 h. 2.4. Analytical and statistical procedures Each sample was analyzed for DOC, UVA254, and THMFP according to detailed analytical and statistical procedures reported elsewhere (Abouleish and Wells, 2012) and described in the Supplementary material associated with this manuscript. 3. Results and discussion Since 1977, more than 100 reports of activated carbon-fulvic/humic isotherms have been published. Various surrogate parameters have been used to investigate the sorption of humic substances on activated carbon—dissolved organic carbon (DOC)- and ultraviolet absorbance (UVA254)-based isotherms are commonly reported, but trihalomethane formation potential (THMFP)-based isotherms are rarely published. THMFP-based isotherms are more experimentally demanding in time and cost to generate than DOC- or UVA254-based isotherms. One exception is that Philippe et al. (2010) published four THMFP-based adsorption isotherms of raw water compared to TiO2 dark adsorption and photocatalytic oxidation at 1 and 10 minute retention times in a study of granular activated carbon and biotreatments. The USEPA uses the DOC and UVA254 surrogates to indicate the presence of DBP precursors and derive the specific ultraviolet absorbance or SUVA: h i −1 −1 –1 UV254 cm =DOC mg L ð1Þ SUVAðL=mg mÞ ¼ 100 cm m by which water treatment plants are regulated. In order to prevent and reduce the formation of DBPs in drinking water, USEPA recommends a value not larger than 2 L mg− 1 m− 1 for SUVA at 254 nm (USEPA, 1999). Thus, in the scientific community, the presumption is made

295

that DOC, UVA254, and SUVA are good predictors of the formation potential for DBPs (USEPA, 2003, 2005, 2006, 2015). However, challenges to this perception have been raised (Fram et al., 1999; Weishaar et al., 2003; Abouleish and Wells, 2012), but require further investigation. Additionally, a conundrum exists in that sometimes DOC-based isotherms are not actually measured, but are predicted by calibrating nominal (as weighed) concentrations of DOC to UVA254. This type of calibration cannot account for preferential adsorption by the activated carbon and can lead to erroneous results. Therefore, in this research, the ability of the surrogate parameters DOC and UVA254 to predict the THMFP was compared by two complementary approaches: (1) fractional reduction in DOC, UVA254, and THMFP upon sorption to activated carbon (Section 3.1), and (2) preparation of DOC-, UVA254-, and THMFP-based sorption isotherms (Section 3.2). 3.1. Fractional reduction in DOC, UVA254, and THMFP upon sorption For each HS tested, the equilibrium value remaining in the liquid phase for DOC (Fig. 1), UVA254 (Fig. 2), and THMFP (Fig. 3) was measured as the mass of activated carbon varied. To facilitate direct comparisons among data, all parameters are presented relative to initial state conditions (Xe/Xo), where X represents DOC, UVA254, or THMFP. A value of 1 (0% removed) represents initial conditions at which there is no activated carbon added. A value of 0 represents complete (100%) removal of the parameter measured. The liquid phase values for the parameters decreased as the mass of activated carbon increased for each HS tested. An alternative presentation of the data is provided in the Supplementary Materials accompanying this manuscript that compares reduction in DOC, UVA254, or THMFP for each individual aquatic FA (Supplementary Fig. 1), terrestrial FA (Supplementary Fig. 2), aquatic HA (Supplementary Fig. 3), and terrestrial HA (Supplementary Fig. 4). The approach of using fractional reduction in surrogate parameters to represent the removal of DBP-generating precursors upon sorption by activated carbon was applied previously (see Fig. 7 in Schmit and Wells, 2002; Fig. 4 in Kilduff et al., 1996a).

Fig. 1. Fractional reduction of DOC on activated carbon: (a) FAs pH 6, (b) FAs pH 9, (c) HAs pH 6, and (d) HAs pH 9.

296


Fig. 2. Fractional reduction of UVA254 on activated carbon: (a) FAs pH 6, (b) FAs pH 9, (c) HAs pH 6, and (d) HAs pH 9.

3.1.1. pH-dependent sorption Previously the pH-dependent removal of HSs on activated carbon evidenced by UV and fluorescence monitoring in flow field-flow fractionation was demonstrated (see Fig. 6 in Schmit and Wells, 2002). Comparably, here we show that the formation of THMs upon

chlorination after activated carbon sorption is pH-dependent. Statistically, as a main effect parameter, pH was shown to be significant to the removal from solution of DOC, UVA254, and THMFP precursors (discussed in more detail in Supplementary Material Section 3). The effect of pH on sorption can be observed in each of Figs. 1–3 and

Fig. 3. Fractional reduction of THMFP on activated carbon: (a) FAs pH 6, (b) FAs pH 9, (c) HAs pH 6, and (d) HAs pH 9.

Fractional Liquid Phase THMFP at pH 9

M.Y.Z. Abouleish, M.J.M. Wells / Science of the Total Environment 521–522 (2015) 293–304 1 0.9 0.8 0.7 0.6

NO-FA & HA

0.5

SW-FA & HA

0.4

PE-FA

0.3

PE-HA

0.2

SO-FA & HA

0.1 0 0

0.2

0.4

0.6

0.8

1

Fractional Liquid Phase THMFP at pH 6

Fig. 4. Comparison of the fractional liquid phase THMFP at pH 6 and 9.

Supplementary Figs. 1–4 by comparing (horizontally) Fig. (a) to Fig. (b) or Fig. (c) to Fig. (d) in each figure. With the exception of PE-HA (compare Supplementary Fig. 4a to 4b), the fractional liquid phase DOC, UVA254, and THMFP values at pH 9 are greater than the comparable fractional liquid phase parameters for DBP-precursors at pH 6 (Fig. 4). If the sorption of HSs on activated carbon was independent of pH, the data would be equally distributed about the solid line (representing a 1:1 ratio) in Fig. 4. Therefore, at pH 6, a greater fraction of DOC, of components absorbing light at 254 nm (UVA254), and of components generating THMs (THMFP) was removed at equivalent activated carbon doses. Conversely, at pH 9, lesser fractions of DOC, UVA254, and THMFP were removed. Applying reverse psychology to interpret the data, the THMFP is greater at pH 9 (except for PE-HA) because fewer DBP-precursor chemicals are removed by activated carbon; and the THMFP is less at pH 6 (except for PE-HA) because, overall, more DBP-precursor chemicals are removed by activated carbon. Changes in pH affect the surface chemistry of activated carbon, as well as the hydrophilic and hydrophobic nature of the HS components. The F400 activated carbon used in this research is derived from bituminous coal,3 has a high capacity for humic acids compared to other types of carbon (Lee et al., 1981), and it is the carbon adsorbent that is most commonly used by water utilities (Kilduff et al., 1996a; Dastgheib et al., 2004). Chromatographically, activated carbon is a mixed-mode sorbent. The microcrystalline structure of activated carbon is composed of assemblies of defective graphene layers (Marsh and RodriguezReinoso, 2006) in which flat aromatic sheets are randomly crosslinked and contain free radicals or unpaired electrons (Bansal and Goyal, 2005). Various types of activated carbon are amphoteric having both acidic and basic surface sites (Marsh and Rodriguez-Reinoso, 2006) and the chemistry at the edges of layers may differ from the bulk material (Bansal and Goyal, 2005). Therefore, activated carbon differentiates among organic components by two modes of chromatographic separation: attraction (chemical) and size exclusion (physical) (Karanfil and Kilduff, 1999). Nonionized analytes are chemically attracted to activated carbon intermolecularly by van der Waals or dipole–dipole interactions, and ionizable analytes are attracted or repelled by electrostatic interactions. Physically, analytes can fractionate on activated carbon by size exclusion, because the activation process imparts porosity (and increased surface area) distributed among macropores (N 50 nm), mesopores (2–50 nm), and micropores (b2 nm) (Bansal and Goyal, 2005). Marsh and Rodriguez-Reinoso (2006) reported that small molecules such as phenol can access micropores in activated carbon, NOM can access mesopores, and bacteria can access macropores. For activated carbon, the point of zero charge, i.e., the pHPZC, is the pH at which the external surface charge is zero (Moreno-Castilla, 3 Calgon Carbon Corporation. http://www.calgoncarbon.com/media/images/site_ library/25_Filtrasorb_400_1019web.pdf.

297

2004)—the surface is positively charged at pH b pHPZC and negatively charged at pH N pHPZC (Marsh and Rodriguez-Reinoso, 2006). The pHPZC for Filtrasorb 400 has been reported (Radovic et al., 2001) to be 6.4 (Zaini et al., 2009), 6.5 (Bjelopavlic et al., 1999), 7.1 (Huang and Smith, 1981), 7.2 (Al-Degs et al., 2000), 8.82 (Valdes et al., 2003), and 10.4 (Corapcioglu and Huang, 1987). In this experiment, the activated carbon is presumed to be excessively positively charged at pH = 6 and excessively negatively charged at pH = 9, thereby affecting differentially the interaction of ionizable and polar analytes at the two pH extremes. The fulvic-like and humic-like chemical constituents studied in this research are ionized and negatively charged in the pH range of 6–9. These HSs are ionizable analytes having pKa values averaging 4.1 for protonation of carboxyl groups (Table 2a) and 9.7 for protonation of phenolate groups (Table 2b). The acid–base properties of the IHSS materials were investigated by Ritchie and Perdue (2003, 2008) indicating that the phenolic-to-carboxylic ratio is approximately 1:4 (Ritchie and Perdue, 2003) averaging one acidic hydroxyl group per two phenyl rings (Ritchie and Perdue, 2008). Combining the knowledge that (1) the surface charge of the activated carbon is excessively positive at pH 6 and excessively negative at pH 9 with (2) the electron-rich ionization status of the HSs in the pH range of 6–9 (Tables 2a and 2b), leads to the explanation of why a greater fraction of THMFP precursors are removed at pH 6 than at pH 9 (except for PE-HA). The aliphatic and aromatic regions of the HS molecules are attracted to activated carbon by van der Waals interactions at both pH 6 and pH 9. In particular, the aromatic HS regions will be attracted to the aromatic graphene-like structure of activated carbon by π–π attractions. However, the electrostatic interactions between the HS molecules and the activated carbon used in this experiment are different at pH 6 than at pH 9. At pH 9, the activated carbon is negatively charged and repulsive to the carboxylic acid groups (100% negatively charged) and the phenolic groups (11–23% negatively charged) of the HSs. At pH 6, the positively charged activated carbon will attract the near neutral electron-rich phenolic groups (non-bonding electron pairs) and the carboxylic acid groups (97-99% negatively charged) by electrostatic forces. Therefore, the pH-dependent sorption of fulvic and humic chemical constituents observed in this research resulted from attractive intermolecular interactions at pH 6 between the acidic FA/HA functional groups and the activated carbon studied in this research, and repulsive intermolecular interactions at pH 9. The reversal in behavior of PE-HA relative to the other HSs examined is attributed to an undetermined uniqueness in the chemical character of this HS (which was also observed in the sorption isotherms presented in Section 3.2). 3.1.2. Actual versus predicted comparisons Linear regression modeling, in which THMFP represents the actual value (y-axis) and DOC or UVA254 represents the surrogate values used to predict the THMFP (x-axis), is a straightforward way to evaluate whether fractional reduction in DOC or UVA254 is a good predictor of the THMFP. Data for the fractional removal of DOC relative to THMFP (Fig. 5a), of UVA254 relative to DOC (Fig. 5b), and of UVA254 relative to THMFP (Fig. 5c) were compared. Overall, UVA254 explained 95% of the variability in THMFP (Fig. 5c) but the intercept was statistically different than zero; whereas, DOC explained 82% of the variability in THMFP (Fig. 5a) and the intercept was not statistically different than zero. The data are interpreted to mean that DOC was significant to the models for predicting THMFP (Fig. 5a), and UVA254 was significant to the model for predicting DOC (Fig. 5b), but substantial amounts of variability in the data were not described. However, the overall appearance of the scatter in the data presented in Fig. 5 fails to give the true impression that the actual fit is good for each individual substance (Table 3). 3.1.2.1. THMFP was zero when DOC was extrapolated to zero. No THM formation was predicted when DOC was extrapolated to zero. The

298


Table 2a Carboxylic functional groups of selected IHSS samples. Sample

Fulvic acids Nordic Lake Suwannee River I Pahokee Peat I Elliott Soil I Humic acids Nordic Lake Suwannee River Pahokee Peat Elliott Soil a b c d

ID

Carboxylic functional groups Carboxylic (meq/g C)a

pKacarboxylica

% carboxylic ionized at pH 6b

% carboxylic ionized at pH 9b

1R105F 1S101F 1S103F 1S102F

11.16 11.44 13.34 ndc

3.79 3.80 3.99 nd

99.39 99.37 99.03 –d

100 100 100 –

1R105H 1S101H 1S103H 1S102H

9.06 9.59 9.01 8.28

4.32 4.42 4.22 4.36

97.95 97.44 98.37 97.76

100 100 100 100

Data abstracted from http://www.humicsubstances.org/acidity.html and Ritchie and Perdue (2003). Data calculated by the Henderson–Hasselbalch equation. Not determined. Not calculated.

Table 2b Phenolic functional groups of selected IHSS samples. Sample

ID

Phenolic functional groups Phenolic (meq/g C)a

pKaphenolica

% phenolic ionized at pH 6b

% phenolic ionized at pH 9b

Fulvic acids Nordic Lake Suwannee River I Pahokee Peat I Elliott Soil I

1R105F 1S101F 1S103F 1S102F

3.18 2.91 2.32 ndc

9.67 9.52 9.57 nd

0.02 0.03 0.03 –d

17.61 23.20 21.21 –

Humic acids Nordic Lake Suwannee River Pahokee Peat Elliott Soil

1R105H 1S101H 1S103H 1S102H

3.23 4.24 1.91 1.87

9.89 9.68 9.86 9.80

0.01 0.02 0.01 0.02

11.41 17.28 12.13 13.68

a b c d

Data abstracted from http://www.humicsubstances.org/acidity.html and Ritchie and Perdue (2003). Data calculated by the Henderson–Hasselbalch equation. Not determined. Not calculated.

statistical evaluation of these plots demonstrated that (overall) when the fractional liquid phase DOC is extrapolated to zero, the fractional liquid phase THMFP is predicted to be zero—intersecting at the origin (Fig. 5a). Because the HSs are carbon-based compounds, intuitively, it is expected that when the DOC is zero, the THMFP will be zero, as was demonstrated by hypothesis testing of the model: THMFP = m(DOC) + b (Table 3) and in Fig. 5a. 3.1.2.2. THMFP was not always zero when UVA254 was extrapolated to zero. Conversely, when the fractional liquid phase UVA254 is extrapolated to zero, the overall fractional liquid phase DOC is not equivalent to zero (Fig. 5b); and, when the fractional liquid phase UVA254 is extrapolated to zero, the overall fractional liquid phase THMFP is not equivalent to zero (Fig. 5c). UVA254 accurately predicts the THMFP for HSs when actual versus predicted comparisons demonstrate a slope = 1 and a yintercept equal to zero (EQ 0) (mid-section of Table 3). If UVA254 were a good representative of the THMFP for all HSs, it is expected that, when the UVA254 is zero the THMFP would be statistically not different than zero, i.e., intersect at the origin. However, this was demonstrated to be substance and/or pH specific (mid-section of Table 3). The lines for the sixteen substances modeled by THMFP = m (UVA254) + b are drawn in Fig. 5d for the data points in Fig. 5c. Statistically, eight of these lines are equivalent to a 1:1 prediction, while eight are not. The data (Fig. 5c, d, and Table 3) are interpreted to mean that, when the y-intercept of the THMFP versus UVA254 plot is not zero, a component(s) of the HS (a) is forming THMs, (b) is not being efficiently removed by activated carbon, and (c) is not strongly absorbing light at

UV254. Because the data represent fractional responses of the liquid phase, the y-intercept of a THMFP versus UVA254 plot that is NE 0 can be expressed as a percent to indicate the degree of THMFP that is unexplained by UVA254 (Table 4). The data for the percentage of the THMFP remaining when UVA254 is EQ 0 provide a means to covertly observe the signature left by compounds that cannot be directly observed by the UVA254 measurements. The data confirm previous findings in this laboratory (Aboul Eish and Wells, 2006) and in the research of others (Fram et al., 1999; Weishaar et al., 2003) that non-aromatic DOC that does not absorb strongly at 254 nm can contribute to the formation of THMs and may account for some degree of the inability of UVA254 to describe the variability in the THMFP equilibrium liquid fraction. Iriarte-Velasco et al. (2008) also reported THM formation by NOM fractions not sorbed on activated carbon and undetected by DOC. For all FAs, the THMFP was zero when the UVA254 was zero at pH 9; however, for all FAs the THMFP was not zero when the UVA254 was zero at pH 6 (Table 4). The nature of the non-aromatic DOC that does not absorb light strongly at 254 nm and does not adsorb strongly to activated carbon at pH 6, yet contributes to the formation of THMs cannot be determined directly from this data. However, the known IHSS constituents4 that fit the description of primarily non-aromatic DOC character are the protein-like components and the polysaccharide-like components composed of amino acid and carbohydrate building blocks,

4 International Humic Substances Society (IHSS). http://www.humicsubstances.org/ sugar.html and http://www.humicsubstances.org/aminoacid.html.


1

y = 0.9522x - 0.013 R² = 0.8219 p-intercept=0: 0.7033 p-slope=1: 0.8000

0.8

Fractional liquid phase DOC

Fractional liquid phase THMFP

1

0.6

0.4

0.2

0.8

0.6

0.4

0 0

a

0.2

0.4

0.6

0.8

1

0.2

0.4

0.6

0.8

1

Fractional liquid phase UVA 1


y = 0.9120x + 0.0758 R² = 0.9497 p-intercept=0: 0.0001 p-slope=1: 0.9891

0.8

0

b

Fractional liquid phase DOC 1


y = 0.8000x + 0.1771 R² = 0.8059 p-intercept=0:

Effects of Humic and Fulvic Acids on Silver Nanoparticle Stability, Dissolution, and Toxicity.

Influence of humic acids on sorption of alkanes by carbon nanotubes--implications for the dominant sorption mode.

The role of humic and fulvic acids in the phototransformation of phenolic compounds in seawater.

Pb(II), Cr(VI) and atrazine sorption behavior on sludge-derived biochar: role of humic acids.

The impact of humic and fulvic acids on the dynamic properties of liposome membranes: the ESR method.

Impacts of powdered activated carbon addition on trihalomethane formation reactivity of dissolved organic matter in membrane bioreactor effluent.

Contribution of coated humic acids calculated through their surface coverage on nano iron oxides for ofloxacin and norfloxacin sorption.

Effect of humic acid (HA) on sulfonamide sorption by biochars.

Sorption of perfluorooctanoic acid, perfluorooctane sulfonate and perfluoroheptanoic acid on granular activated carbon.

Cascading effects of induced terrestrial plant defences on aquatic and terrestrial ecosystem function.

Cd(II) Sorption on Montmorillonite-Humic acid-Bacteria Composites.

Effects of chlorination operating conditions on trihalomethane formation potential in polyaluminum chloride-polymer coagulated effluent.

fulvic acids as related to their chemical structures in greenhouse vegetable fields with cultivation chronosequence.

Comparative sorption and desorption behaviors of PFHxS and PFOS on sequentially extracted humic substances.

Effects of sludge retention times on reactivity of effluent dissolved organic matter for trihalomethane formation in hybrid powdered activated carbon membrane bioreactors.

Effect of humic acids with different characteristics on fermentative short-chain fatty acids production from waste activated sludge.

Activated carbon from biochar: influence of its physicochemical properties on the sorption characteristics of phenanthrene.

Effects of climate extremes on the terrestrial carbon cycle: concepts, processes and potential future impacts.

Tracking the composition and transformation of humic and fulvic acids during vermicomposting of sewage sludge by elemental analysis and fluorescence excitation-emission matrix.

Potential regulatory role of gibberellic and humic acids in sprouting of Chlorophytum borivilianum tubers.

Myoglobin oxygen affinity in aquatic and terrestrial birds and mammals.

Positional strategy of trunk muscles among aquatic, semi-aquatic and terrestrial species in Urodela.

Pulmonary ventilation: perfusion relationships in terrestrial and aquatic chelonian reptiles.

Partitioning of humic acids between aqueous solution and hydrogel: concentration profiling of humic acids in hydrogel phases.