Evaluation Technique for Water Treatment Plant in terms of NOM Size, Structure, and Functionality (SSF Technique) as well as HAAFP/Reactivity
Sangyoup Lee1, Jaeweon Cho1, Seonha Chae2, In S. Kim1
1Dept. of Environmental Sci. and Eng., K-JIST, Kwangju 500-712, Korea,
2Water Supply & Sewerage Research Team, Water Resources Research Institute, Korea Water Resources Corporation, Taejon, Korea
Abstract
In this study, a tool was suggested to help water treatment plants evaluate either conventional or advanced treatment processes with respect to the removal of natural organic matter (NOM). This tool, called size/structure/functionality (SSF), is a technique that requires the characterization of NOM size distribution, structure, and charge density (functionality). Characterizing the NOM size distribution for each process demonstrated that the NOM removals by granular activated carbon (GAC) adsorption and sand filtration were mainly controlled by molecular diffusion and size exclusion, respectively. The formation potential (FP) of haloacetic acids (HAA) and its reactivity (= HAAFP/dissolved organic carbon (DOC)) were related to the NOM characteristics, as determined by the SSF for each process. Hydrophilic NOM exhibited the highest HAAFP reactivity, and it was preferentially removed by the GAC adsorption process compared to transphilic and hydrophobic NOM. However, tribromoacetic acids (TBAA) and trichloroacetic acids (TCAA) still remained at a high concentration after the GAC adsorption process. Functionality analysis with carboxylic/phenolic acidities measurements revealed that conventional (coagulation/sedimentation/filtration) and advanced (ozonation/GAC adsorption) processes did not use the charge interaction mechanism efficiently for NOM removal.
Key words- SSF evaluation, natural organic matter, haloacetic acids, NOM molecular size, NOM structure, NOM functionality
Introduction
A conventional water treatment plant in Korea was evaluated using NOM size/structure/functionality (SSF) analyses because conventionally treated water exhibited fairly high HAAFP and its reactivity. Meanwhile, the formation potential of trihalomethanes (THMs) was very low for the same source water. A pilot test with ozonation and GAC adsorption processes was performed and evaluated using the SSF technique. It is anticipated that the SSF techniques may be used for both the evaluation of conventional treatment and the determination of advanced treatment alternatives in terms of NOM removal and HAAFP minimization.
Materials and Methods
Water source used
The drinking water source from Suksung Reservoir and samples from each process of conventional actual plant and advanced pilot plant were taken to the laboratory. All samples were pre-filtered with a 0.45 m filter and stored in a refrigerator at a temperature of 5C. The NOM characteristics of the source water and treated waters are shown in Table 1.
Table 1 NOM characteristics of the source water used
Source
/ DOC(mg/L) / UVA254
(cm-1) / SUVA
(Lmg-1m-1) / pH / Br-
(g/L) / Conductivity
(S/cm) / Molecular Weight (daltons)
Number-avg. / Weight-avg.
Suksung Reservoir / 3.40 / 0.079 / 2.3 / 7.1 / 49.0 / 328 / 1,815 / 2,352
NOM size analysis
The NOM size distribution was measured using a high performance size exclusion chromatograph (HPSEC) with a proteinanous silica column (Waters) and a HPLC (Waters 510)/autosampler (Waters 717 plus). Polystyrene sulfonates with 1800, 4600, 8000, and 18000 daltons were used for a calibration standard equation between the MW size and retention time, and a 4 mM phosphate buffer solution with added NaCl (for a 0.1 M ionic strength) was used as an eluent solution (Chin et al., 1994).
NOM structure (hydrophobic, transphilic, and hydrophilic) analysis
Pre-filtered water samples were processed through XAD-8 followed by XAD-4 resins to isolate the hydrophobic (XAD-8 adsorbable), transphilic (XAD-4 adsorbable), and hydrophilic (neither XAD-8 nor XAD-4 adsorbable) NOM fractions. The adsorbed hydrophobic and hydrophilic (transphilic) acids were eluted using a 0.1 N NaOH solution.
NOM functionality (charge density) analysis
Using a micro titrator (Metroh, CH-910), the carboxylic and phenolic acidities were measured for XAD-isolated hydrophobic and hydrophilic acids. The pH of a titration sample was lowered to 3.0 and carbonate in the sample was sparged out with nitrogen gas for at least ten minutes prior to actual titration. A NaOH solution of 0.05 N was used to increase the pH of the sample, and the amount of NaOH consumed to increase pH 3-8 and 8-12 corresponds to carboxylic and phenolic acidities, respectively.
HAA measurement
The source and treated samples were chlorinated using a concentrated HOCl solution with a chlorine concentration based on 3[DOC concentration] + 7.5[NH4+ concentration]. The chlorinated samples were incubated at a temperature of 20C for 72 hours, and the HAAFP was extracted using the Modified EPA 552 micro-extraction method with a diazo methane addition. The HAA was measured using a GC (HP 5890 Series II Plus) with an autosampler (HP 6890 Series).
Results and Discussion
NOM size
A comparison of the MW distributions of the source and treated water samples is shown in Fig. 1(a), and the NOM fractional rejections (FR) (Cho et al., 1999) calculated from the MW distribution for each process are represented in Fig. 1(b). NOM with larger sizes exhibited higher FR than NOM with smaller sizes for conventional treatment, but the larger NOM molecules (around 1500 daltons) exhibited even less removal than the smaller NOM molecules (between 1000 and 1100 daltons) for the ozonation/GAC adsorption processes. Through these results, it can be concluded that (1) size exclusion (or steric hindrance) was a major mechanism for NOM removal by conventional treatment (including alum coagulation, sedimentation, and sand filtration), and (2) the diffusion transport of small size molecules was a dominant mechanism for NOM removal by ozonation followed by GAC adsorption processes.
Fig. 1. (a) NOM MW distribution comparison of source, conventionally treated (alum coagulation, sedimentation, and sand filtration), ozonated, and GAC treated water samples; (b) Fractional NOM rejections by conventional treatment and ozonation/GAC adsorption.
The FRs of three NOM fractions provided by conventional treatment and ozonation/GAC adsorption are shown in Figs. 2 (a)-(c). Through these results, it can be determined which size ranges of NOM molecules were removed preferentially by a certain treatment alternative compared to other sizes of NOM molecules. For example, the smaller molecules of hydrophobic NOM were preferentially removed by O3/GAC adsorption, which is in agreement with the results of Kilduff et al. (1996).
Fig. 2. Fractional NOM rejections by conventional treatment and ozonation/GAC adsorption for (a) hydrophobic NOM, (b) transphilic NOM, and (c) hydrophilic NOM.
NOM structure
Figs. 3 and 4 represent the NOM fractions for the source and treated samples by each process as well as the removal of each NOM fraction by each process, respectively. It was shown that the hydrophilic NOM fraction was preferentially removed by both the conventional and advanced treatment processes.
Fig. 3. NOM structure informaion for source and treated waters.
Fig. 4. Removal percentage of each NOM fraction by each process.
NOM functionality
For the conventionally treated samples, the acidities of both hydrophobic and hydrophilic acids were higher compared to the source water, representing that conventional treatment options did not use charge interaction mechanism for NOM removal efficiently (see Table 2). However, the hydrophobic acidity was less for the GAC treated sample due to the preferential removal of charged hydrophobic NOM during the process.
Table 2 NOM acidities for source and treated samples by each process
Sample / Hydrophobic acid (meq/gC) / Hydrophilic acid (meq/gC)-COOH / -OH / -COOH / -OH
Source / 5.9 / 15.9 / 4.5 / 12.1
Conventionally treated / 11.6 / 14.9 / 10.3 / 18.0
Ozonated / 9.3 / 13.4 / 13.9 / 15.6
GAC treated / 7.2 / 5.9 / 13.6 / 29.5
HAA formation potential and reactivity
Conventional treatments (alum coagulation/sedimentation/sand filtration) exhibited virtually no HAAFP removal, while the ozonation/GAC process exhibited a fairly high HAAFP removal (see Fig. 5). However, the TCAA and TBAA still remained at a high concentration, even after O3/GAC treatment (see Table 3) (recall that the final regulation for HAA is 30g/L). Hydrophilic NOM exhibited the highest HAA reactivity compared to the hydrophobic and transphilic NOM fractions (see Table 4).
Fig. 5. HAAFP for source and treated samples by each process.
Table 3 HAAFP9 (TCAA, TBAA, and other 6 HAAs) for source and treated samples
HAAFP / Source (g/L) / Conventionally treated (g/L) / GAC treated (g/L)TCAA / 73.6 / 73.6 / 25.5
TBAA / 53.5 / 44.0 / 29.3
Total HAA-TCAA-TBAA / 82.5 / 77.8 / 16.7
Total HAA / 209.6 / 195.4 / 71.5
Table 4 HAA reactivity of each NOM fraction for source and treated waters
Water / Raw (g/mg) / Hydrophobic (g/mg) / Transphilic (g/mg) / Hydrophilic (g/mg)HAA6 / HAA9 / HAA6 / HAA9 / HAA6 / HAA9 / HAA6 / HAA9
Source / 15.2 / 22.1 / 28.3 / 32.3 / 14.5 / 24.3 / 61.5 / 89.3
Filtered / 19.8 / 27.5 / 16.2 / 25.7 / 14.9 / 26.9 / 94.8 / 131.8
Ozonated / 28.1 / 38.8 / 13.9 / 29.1 / 17.4 / 30.7 / 72.3 / 99.6
GAC / 8.7 / 17.6 / 12.8 / 23.4 / 8.0 / 30.9 / 38.8 / 78.0
Conclusions
-The SSF technique was successfully introduced as a tool to evaluate water treatment processes along with NOM rejection and HAAFP.
-Conventional treatment alternatives appeared to be controlled mainly by the size exclusion mechanism (no charge interactions) for NOM removal, and exhibited almost no HAAFP removal.
-Ozonation followed by the GAC adsorption process was controlled by a diffusion transport mechanism for NOM removal (based on NOM size analysis), and exhibited fairly high HAAFP removal. However, TCAA and TBAA remained at high concentrations compared to the other six HAAs, even after the O3/GAC process.
-Hydrophilic NOM exhibited the highest HAA reactivity (HAAFP/DOC; g/mg C) compared to the hydrophobic and transphilic NOM fractions.
-Hydrophilic NOM can be removed preferentially by both the conventional and advanced treatment (O3/GAC) options.
-The charge interaction for NOM removal was not used effectively by conventional treatment processes based on NOM functionality analysis, and it was used for the removal of hydrophobic acids by GAC adsorption.
Acknowledgments
This work was supported in part by the Korea Science and Engineering Foundation (KOSEF) through the Advanced Environmental Monitoring Research Center (ADEMRC) at Kwangju Institute of Science and Technology (K-JIST).
References
Chin, Y., Aiken, G., and O'Loughlin, E. (1994) Molecular weight, polydiversivity, and spectroscopic properties of aquatic humic substances. Environ. Sci. Technol. 28, 1853 –1858.
Cho, J., Amy, G. and Pellegrino, J. (1999) Membrane filtration of natural organic matter: Initial comparison of rejection and flux decline characteristics with ultrafiltration and nanofiltration membranes. Wat. Res. 33, 2517-2526.
Kilduff, J.E., Karanfil T., and Weber W.J. (1996) Competitive interactions among components of humic acids in granular activated carbon adsorption systems: effects of solution chemistry. Environ. Sci. Technol.30, 1344-1351.