C:\DAR\RES\Epa2\Mousset\pap1\measurement.doc
ISOLATION AND CHARACTERISATION OF NATURAL ORGANIC MATTER AND COMPARISON WITH THE ORIGINAL WATER, SO AS TO VALIDATE THE EXTRACTION PROCEDURE
MOUSSET B. and D. RECKHOW
Department of Civil and Environmental Engineering
University of Massachusetts
Amherst, MA 01003
Abstract
Aquatic organic matter was isolated and fractionated using a blend of the procedures of Leenheer (1981) and Malcolmet al. (1993). Extraction of the DOC and UV absorbance was nearly complete. Eight different fractions were obtained that was split between the hydrophobic fractions (fulvic, weak and humic acids, hydrophobic neutral and hydrophobic base) and the hydrophilic fractions (hydrophilic and ultra hydrophilic acids, hydrophilic neutral and hydrophilic base). The fractions were isolated from a Watchusett reservoir and on the end of Andover treatment plant. All the DOC and the UV absorbance were recovered. As the comparison between the fractions and the water, the organic matter was not change during the extraction procedure. Thus during the extraction, the organic matter was separated according to the different chemical propriety. Each fraction was different each other.
The DOC proportion of the hydrophobic fraction (60 and 50% of DOC) was around the same than the hydrophilic fraction (40 and 50% of DOC). The fulvic acid was the predominant fraction.
And the hydrophobic fractions represent the most fraction of UV absorbance (between 52 and 72% of UV absorbance).
The UV absorbance proportion of the hydrophobic substances was correlated to the SUVA.
For all fractions, the fluorescence emission maxima of the peak C was comprise between 484 and 380 nm, the relative UV absorbance between 5 and 0.52 m-1 l mg-1 C and the BDOC between 10.8 and 51.2% of the DOC. For each parameters of characterization a decreased order could be realized :
for SUVA : HAFA > HyB = HyA = HoBuHyA = HoN = WHoA > HyN,
for BDOC ;WHoA > uHyAAH > HyA > FA > HyB > HyN,
for fluorescence :AHFA > WHoA > HoB > uHyA > HyA > HyN > HoN = HyB.
A good relation was obtaining between the sum of polyhydroxyaromatic and the aromatic fragments with the relative UV absorbance. And a positive trend was found between the protein plus polysaccharide and the BDOC.
Any correlation between emission maxima fluorescence (peak C), the relative UV absorbance or the BDOC have been found. But the most biodegradable fraction, the weak hydrophobic acid, has a small SUVA and a little emission wavelength of peak C, contains a great quantity of protein plus polysaccharide and presents a great ratio B/A.
Keywords
Surface water, treated water, organic matter, fractionation, fluorescence, pyrolysis and BDOC.
Introduction
Characterization of natural organic matter (NOM) has been a major focus of many researchers in recent years. Because this NOM reacts with oxidant to form harmful compounds (Richardson, 1998), can form pesticide or metal complexes (Ziechmann, 1988) and is responsible of the reviviscence of bacteria (Huck, 1990 ; Le chevallier, 1990 and Blocket al., 1992).
The NOM was characterized by non-specific parameters such as COT, DOC, and absorbance, by reactivity with coagulant and oxidant and most recently by biodegradable organic matter (BDOC), by sophistical analytical such as 13Cand 15NRMN, 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.
The pyrolysis that is a semi-quantitative method could be shown correlation between the results of the pyrolysis and the results obtained with other quantitative methods (Martin, 1995 ; Croueet al., 1993 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 groupements. So the pyrolysis is an interesting method. And, the fragments obtained by pyrolysis show information on the origin of the water, aquagenic or pedogenic organic mater than unlined Bruchet (1990) and Biberet al. (1996).
The fluorescence is useful because the maximum wavelength increase with the molecular weight shown by Belinet al. (1993).
The biodegradable organic matter (BDOC) is an important parameter to control the quality of treated water so as to know the biological stability in the distribution system. But this parameter is difficult to measure in the distribution system because the concentration is very small and to know where it comes from. The responsible compounds of BDOC were unknown. Moreover during ozonation treatment, Huck (1990) have found that the increase of AOC is associate with the increase of formaldehyde and Martin (1995) have shown a relation between the increasing of the hydrophilic fractions (compounds non adsorbed on the XAD-8 resin) with the BDOC.
Moreover some of these techniques need a high quantity of sample. So to better know about NOM and about water treatment, this NOM was extracted and fractionated.
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. And recently, the isolation developed by Rybackiet al. (1998) using a combination of reverse osmosis and electrodialysis. The losses by this technique were lower than 5% and the desalination yield was higher than 99%. Vacuum rotary evaporation is an alternative method to concentrate few liters of water, but the salts were not eliminated and the NOM loses can be very high as showed Rybackiet al. (1998) on the seine river and reservoir of Cholet about 50% are lost for a sample of 400 l. Serkiz and Perdue (1990) proposed to use reverse osmosis as a rapid technique to concentrate the MON from large volumes of water. They obtained recoveries higher than 80%. But this technique does not separate the NOM. The procedure proposed by Thurman and Malcolm (1981) and the one developed by Leenheer (1981) are the basis of the most widely used method to isolate NOM. The combination of XAD-8 and XAD-4 resins was shown to isolate more than 74% of DOC (Martin, 1995) and the procedure of Leenheer (1981) could be recovery between 64 and 100% of the NOM. However, those techniques are consuming time and also present some risks of sample contamination and NOM alteration by using NaOH and HCl.
So no people do know if during the extraction the organic matter was changed. This question is very important and tries to be resolved in this paper to compare the water and the different fractions for some parameters of characterization. The extraction procedure using combines the methods of Leehneer (1981) and Malcolmet al. (1993) using Amberlite, anionic and cationic resins.
This organic matter change in each season and with the origin. Then NOM is allays unknown. But, it’s possible to compare each fraction and to obtain comparisons and correlation. So two samples were made in the same water at different times. And one attempt leaded on treated water.
The parameters used to characterize the organic matter and to compare the fractions with the original water were UV absorbance/DOC, biodegradable organic matter (BDOC), fluorescence and fragments obtained in pyrolysis.
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), Reckhowet 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 m in diameter and the second filter unit contains filter tube (type AAH) rated to retain 98% of particles 0.3 m 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 if the water wasn’t treated. The coagulation-flocculation eliminates 90% of humic acid DOC (Babcock and Singer, 1979 and Lefebvre and Legube, 1990). So it’s not enough humic acid to extract. The humic acids were slited 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 Sunet 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 Malcolmet 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 40C. The fulvic and hydrophilic acids could be drying.
Pyrolysis gas chromatography/mass spectrometry GC/MS
The method used was similar to the one published by Bruchetet 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 or by precipitation. Alcanizet al. (1989) were demonstrated that the usual salts, that found in the water, could be reduced the intensity of response without changed the pyrogram.
The temperature programmation of pyrolysis was 200C (1 s) to 700C (10 s) at the rate of 20C/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 220C at the rate of 3C/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 was followed as Bruchetet al. (1990) except that the sums were not multiplied by a corrective factor. An other aromatic fragment was added at this list as Labouyrie-Rouiller (1997).
Measurement of the biodegradable dissolved organic carbon (BDOC)
The BDOC analyses were inspired by the method proposed by Joret and Levi (1986) used a fixed biomass on sand. Here, the inoculum was fixed on expended glass (SIRAN) which is an inert material. In order to optimize the BDOC analyses, the siran were put directly in contact with Watchusett water. Autochthon bacteria of this water colonized the siran. But just before each experiment, colonized siran was watched. So the biomass was adapted and the time was reduced.
The BDOC correspond to a diminution of DOC at 8 days. The reactor contends 400 ml of sample and 25 ml of siran. The reactors were shacked and temperature was maintained at 20C. The pH was fixed between 6.5 and 7 as the raw water. The measurements were made in duplicate.
For the fractions, some milliliters of 3 mg/l sulfate ammonium and of 15 mg/l potassium phosphate (KH2PO4) were added to ensure an adequate nitrogen and phosphorus supply. Because during the fractionation, the salts necessaries to growth of bacteria were eliminated. One blank with sodium acetate was made to control the bacteria efficacy.
In addition, four attempts were made on the fractions (humic, fulvic, hydrophilic and ultra hydrophilic acids) to know the evolution of DOC with the time (figure 3). For the humic, fulvic, hydrophilic and ultra hydrophilic acids, the DOC decrease rapidly until 100 hours and than stabilized. The curve is the same for the water. So for the fractions, the water time of 8 days (192 hours) could be used.
Fluorescence measurement
The fluorecence 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.
Three kinds of experiments were conducted. The first one 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 Ewaldet al. (1983). The second one, the excitation wavelength was set at 275 nm to observe the peak B and T as Goble (1996). The third one, the emission wavelength was fixed at the maximum emission found in the first experiment and/or at 440 nm.