Impact of chlorine on the dissolved organic matter : SUVA, repartition of hydrophobic and hydrophilic substances, fluorescence, pyrolysis and desinfection by products (DBPs)
D. RECKHOW, B. MOUSSET, J. McCollan and Y.
Department of Civil Engineering, University of Massachusetts, at Amherst
Amherst, MA 01003
Abstract
C:\DAR\RES\Epa2\Mousset\pap3\redaction.doc
Keywords
Natural organic matter, Raw and chlorinated waters, Repartition hydrophobic and hydrophilic substances, SUVA, pyrolysis and DBPs.
Introduction
The natural organic matter (NOM) was characterized by non-specific parameters such as COT, DOC, and UV absorbance, by reactivity with coagulant and oxidant and most recently by biodegradable organic matter (BDOC), by sophistical analytical such as 13C and 15N RMN, fluorescence and pyrolysis. However, those sophistical techniques are easily applicable to pure substances. Their applicability to NOM is more uncertain, because the salts could be disturbed the results and need a high quantity of sample. So it’s necessary to extract, to concentrate and to purify the NOM. Different isolation procedures have been used during the last decades, including adsorption into XAD, anionic and cationic resins, filtration with membranes (ultrafiltration, nanofiltration and reverse osmosis), and rotary evaporation. But the basis of the most widely used procedures is proposed by Thurman and Malcolm (1981) or by Leenheer (1981). The organic natural matter separated based on charge and hydrophobic properties. The fractionation scheme divided the organic matter into two major categories : hydrophobic and hydrophilic and at least three subcategories : acids, bases and neutrals.
The ratio of UV absorbance/DOC, named SUVA, is very useful and used by many people because it represents the aromaticity of the organic matter and can correlate with other parameters, such as the chlorine demand, organohalogenated compounds (TOX) and trihalomethanes (THM). The organic matter reacts with ozone and chlorine to form disinfection by-products (DBPs) such as THM and haloacetic acids with chlorine and aldehydes and ketoacids with ozone.
The fluorescence is useful because the maximum wavelength increase with the molecular weight shown by Belin et al. (1993) and Coble (1996) used it to characterize the water of different origins.
The pyrolysis is a very promising method that can already be used to estimate the overall composition of the organic matrix of waters. This method is a thermal degradation method and fragments the organic matter in reproducible and significant products, which analyzed by gas chromatography. Those fragments could be compared with model compounds and can be related to the structure of the undegraded products such as polysaccharides, proteins, lignin, and aromatic and polyhydroxyaromatic compounds. The pyrolysis is a semi-quantitative method. However the fragments could be shown correlation between the results of the pyrolysis and the results obtained with other quantitative methods (Martin, 1995 and Labouyrie-Rouiller, 1997). Examples, the relation between the aminoacids and the structure derived by proteins, the polysaccharides and the sugars or the polyhydroxyaromatic structures and the UV absorbance characterization of the aromatic groups. And more, the fragments obtained by pyrolysis show information on the origin of the water, aquagenic or pedogenic organic mater than unlined Bruchet (1990) and Biber et al. (1996). But the pyrolysis and pyrolysis/methylation have been used together, because they present each other some limitations.
The aim of the present work was to evaluate the impact of chlorine on different parameters : DOC, UV absorbance, SUVA, fluorescence repartition hydrophobic and hydrophilic substances and characterization of the fractions with SUVA, fluorescence and reactivity with chlorine (diminution of UV absorbance at 272 nm, chlorine demand, TOX, THM and HAA).
Materials and Methods
DOC measurement
Dissolved organic carbon (DOC) was measured using a Shimadzu Model TOC-500A with ASI-5000A autosampler, calibrated with a potassium hydrogen phthalate standard (C8H5O4K) solution containing 2, 5 and 10 mg C l-1. For each sample, a minimum of triplicate measurements was made.
UV absorbance measurement
UV absorbance of all fractions was measured at 254 nm in one or five path length quartz cell using a spectrophotometer (Lambda 3A UV/Vis spectrophotometer, Perking Elmer corporation, Norwalk CT).
Extraction and fractionation of NOM
NOM fractions, used in this study, were according to the procedure developed by Leenheer and Noyes (1984), Reckhow et al. (1993) and Thurman and Malcolm (1981).
The NOM was filtered on tow filters : the first is type DH rated the retain 98% of particles 25 mm in diameter and the second filter unit contains filter tube (type AAH) rated to retain 98% of particles 0.3 mm in diameter.
The extraction system consisted of two steps. The first one, around 200 liters of water was filtered at pH neutral through the three resin columns of 2 liters (XAD-8, MSC-1 and A-7) connected in series at the rate of 6 liters per hour (figure 1).
(HyN : hydrophilic neutral, HyB : hydrophilic base, HyA : hydrophilic acid, HoN : hydrophobic neutral, HoB : hydrophobic base, WHoA : weak hydrophobic acid, HS : humic substances, FA : fulvic acid, HA : humic acid, uHyA : ultra hydrophilic acid)
Figure 1 : Fractionation procedure
The organic natural matter separated based on charge and hydrophobic properties.
The fractionation scheme divided the organic matter into two major categories : hydrophobic and hydrophilic and at least three subcategories : acids, bases and neutrals.
The second step corresponds of an elution at 0.85 l/h of fractions : the hydrophilic base (HyB), the hydrophobic base (HoB) and the weak hydrophobic acid (WHoA) were eluted by NaOH or HCl in XAD-8 and MSC-1 resins. The hydrophobic neutral (HoN) were extracted in XAD-8 by methanol in soxhlet. On resin A-7 were eluted together, fulvic acid (FA) and humic acid (HA) corresponded of humic substances (SH), hydrophilic acid (HyA) and ultra hydrophilic acid (uHyA). The fractions, that have a DOC more than 200 mg C/l, were acidified at pH 1 for the separation of humic acids that precipitated. The humic acids were separated by centrifugation (30 min, 500 rpm). The salts were staying in the suspended fraction. So all of the humic acids were recovered and the procedure avoids the precipitation of this fraction on the H+ exchange resin. This problem of the precipitation was unlighted by Sun et al. (1995).
The procedure which was represented in figure 2, consisted by the separation of fulvic, hydrophilic and ultra hydrophilic acids at pH 2 on 0.5 liter of each resins XAD-8 and XAD-4 in series ; than by the elution and finely by the purification. The rates of filtration are 0.8 l/h in XAD-8 and 0.4 l/h in XAD-4 and the volume is 12 liters with a DOC < 10 mg C/l in accordance with the procedure of Malcolm et al. (1993).
(FA : fulvic acid, HyA : hydrophilic acid, uHyA : ultra hydrophilic acid)
Figure 2 : Concentration and purification of fulvic and hydrophilic acids
The ultra hydrophilic acid and the hydrophilic neutral fractions were concentrated by rotary evaporation at 40°C. The fulvic and hydrophilic acids could be drying.
Fluorescence measurement
The fluorescence spectra of all extracts were recorded with a spectrofluorophotometer model RF-540 (Shimadzu) equipped with a monochromator both on excitation and emission (Off-plane concave diffraction, aperture : f/2.6). The signal of photons was detected by a photometric photomultipliplier R452-01.
The first sample was introduced in a 1 cm square cell at the temperature of the room (20 degree C) and the wavelength of excitation was set at 313 nm as mentioned by Ewald et al. (1983).
Pyrolysis gas chromatography/mass spectrometry GC/MS
The method of pyrolysis alone used was similar to the one published by Bruchet et al. (1990). The fractions were concentrated by different methods reported upstairs. The sample must be have got a DOC more than 100 mg C/l. A few milliliters were transformed, under a nitrogen stream, to a solid fraction. Then the pyrolysis GC/MS experiments were run on around 50 mg of sample deposited into quartz tube, which was inserted into a filament pyrolyzer. The salts were lost during the evaporation on the wall of the tube. Alcaniz et al. (1989) were demonstrated that the usual salts, that found in the water, could be reduced the intensity of response without changed the pyrogram.
For the pyrolysis/methylation, one step of methylation was added before pyrolysis. A derivatizing reagent used was the tetramethylammonium hydroxide (25 wt % solution in water) noted TMAH. This method is described in detail elsewhere (Challinor, 1989, Saiz-Jimmenez et al., 1993 ; De Leeuw and Baas, 1993 ; Del Rio and Hatcher, 1996) and is briefly explained here. Approximately 5 ml of the TMAH was placed with the sample in the quartz tube. The all was inserted into a filament pyrolyzer.
The temperature programmation of pyrolysis was 200°C (1 s) to 700°C (10 s) at the rate of 20°C/ms with a CDS 1500 pyroprobe 2000. At this temperature, the organic matter produced different fragments separated on a 30 cm length DB WAX capillary column by Hewlett Packard 5890A gas chromatography flushed with helium gas. The oven programmation was from 25 to 220°C at the rate of 3°C/min and the identification was made by a Hewlett Packard 5988A mass spectrometer operated at 70 ev and scanned from 20 to 400 amu at 1 scan/s.
The semi-quantitative interpretation of pyrograms of the fractions was followed as Bruchet et al. (1990) except that the sums were not multiplied by a corrective factor. An other fragment : the aromatics were added at this list as Labouyrie-Rouiller (1997). And the carboxylic acid methyl ester comes from fatty acids what represents an other group.
Some standards (benzene, acetone, 2,3 dimethyl naphthalene, 2 methyl naphthalene, 1 methyl naphthalene, naphthalene, phenol, pyridine, toluene, 2,3,5 trimethyl naphthalene, o xylene, m xylene and p xylene) were injected directly in the GC/MS to confirm the retention time.
Results and discussions
Presentation of the sampling waters
This study was realized on two different waters sampling on May 26 1998. The first water was sampling in the Wachusett reservoir and the second one was collected in the same place after chlorination.
The two waters were filtered on 25 and 0.3 mm. The DOC, UV absorbance, SUVA, fluorescence and the elimination of DOC and UV absorbance were presented in the table .
Table : DOC, UV absorbance (UV abs), SUVA and fluorescence (maximum emission wavelength of peak C, gmax) of the chlorinated and raw water and the elimination of DOC and UV absorbance by the chlorine
Water / DOC / UV Abs / SUVA / gmax / Elimination (%)(mg C/l) / (cm-1) / (m-1 l mg-1 C) / (nm) / DOC / UV Abs
Raw / 2.85 / 0.07 / 2.56 / 427 / - / -
Chlorinated / 2.59 / 0.0496 / 1.91 / 412 / 9 / 29
The better elimination of the UV absorbance is in accordance with the oxidation by chlorine as the ozone and with the lowest SUVA of chlorinated water.
The maximum emission wavelength of peak C was greater for the raw water than the chlorinated water. So the chlorine moved the maximum emission wavelength of peak C toward the smaller maximum emission wavelength. That means under the results of Belin et al. (1993) that the chlorine broke some links and reduced the apparent molecular weigh.
NOM recovery measurements
Two filtration of around 200 l were made follow the protocol describe just more height. those volumes correspond in volume where an augmentation of the UV absorbance was seen.
The NOM recoveries shown on the table 3, can be estimated, a various stages of a procedure, by measurements of the DOC and of UV absorbance change in the top and in the end of each column. The determination of the percent recoveries is comparing the quantities of carbon or UV absorbance of NOM filtered with the quantities of carbon or UV absorbance of the NOM fractions that adsorbed on the different columns. DOC or UV absorbance and the volume measurements determined the quantities.
Table : Organic carbon and UV absorbance recoveries of Wachusett water using filtration and column adsorption system (LEENHEER and NOYES, 1984)
Fractions adsorbed / Quantity of organic matter (mg C)* / Organic carbon recovery (%) / Quantity of UV Abs(cm-1)** / UV Abs recovery (%)
filtered / adsorbed / filtered / adsorbed
Raw
Water / XAD-8
MSC1
A7 / 55
13
362 / 131
29
396 / 239
223
102 / 1.5
0.6
12.1 / 1.4
0.6
10.9 / 94.5
100
90
Bilan*** / 591 / 690 / 115 / 15.2 / 13.9 / 91.5
Chlorinated
Water / XAD-8
MSC1
A7 / 30
23
375 / 89
29
357 / 291
125
95 / 0.7
0.6
8.2 / 0.75
0.7
8 / 108
122
97
Bilan*** / 545 / 591 / 109 / 10.4 / 10.3 / 99
(* +\- 10 mg C/l, ** +\- 0.08 cm-1, *** all fractions with hydrophilic neutrals)
This extraction procedure of the organic matter could be to extract the totality of the DOC and of the UV absorbance of each water. The values were 115% and 109% for DOC and of 91.5 and 99 for UV absorbance. Those results were acceptable in view the error made.
Repartition of hydrophobic and hydrophilic fractions
The protocol could be split the water into neigh fractions : fulvic acids (FA), weak hydrophobic acids (WhoA), humic acids (HA), hydrophobic neutral (HoN), hydrophobic
Raw water
Chlorinated water
Figure 4 : DOC and UV absorbance repartition of hydrophobic and hydrophilic substances
(FA : fulvic acid, WHoA : weak hydrophobic acid, HA : humic acid, HoN : hydrophobic neutral, HoB : hydrophobic base, HyA : hydrophilic acid, UHyA : ultra hydrophilic acid, HyN : hydrophilic neutral, HyB : hydrophilic base)
bases (HoB), hydrophilic acids (HyA), ultra hydrophilic acids (UHyA), hydrophilic neutrals (HyN), hydrophilic bases (HyB).