Jadashecart, A; Elmorer, A; Stitou, M; Bouillot, P; Legube, B
The general purpose of this study is to try to identify the organic compounds that react with chlorine from the Seine water entering the final treatment stage at the Choisy-le-Roi potable waterworks, near Paris. We present here the part of the study concerning the chlorine consumption kinetics of samples of water taken after overall treatment. The influence of TOC, water, pH and temperature were initially examined.
The Choisy-le-Roi plant (Fig. 1) supplies 2 million inhabitants living in the southern suburbs of Paris. The drinking water process is based on biological treatment and chlorine is injected at the end of the run. Samples of water were taken after GAC filtration when the ammonia concentration was less than 30-mu-g l-1.
All operations were carried out in vessels carefully cleaned (sulfochromic mixture, milliQ water) and stored in chlorine solution (100 mg l-1 Cl2), then rinsed with milliQ water just before use. Seine fulvic acid was extracted following the method of Thurman and Malcolm (1981). Bovine serum albumin was supplied by Fluka. Chlorinations were carried out in a batch system at 20-degrees-C in darkness, by adding micro-volumes of chlorine stock solution. For each run, a blank was performed at the same pH as the water being chlorinated. The chlorine consumptions by the blanks were subtracted from the chlorine consumptions observed for the water samples. Total residual chlorine and free residual chlorine were analyzed by the DPD colorimetric method (Afnor, 1987) with minor modifications. Total organic carbon analysis was performed with a Dohrman DC 80 apparatus. Ultrafiltration was carried out with a 400 ml Amicon cell and YM 30, YM 10, YM 2 and YV 05 membranes, under 3 bar-pressure with nitrogen purified by passing through a silica gel cartridge, a granular activated carbon cartridge then a 750-degrees-C furnace.
The chlorine demand (total chlorine) of the GAC filtered waters is presented in Figs 2 and 3 for the 1987 period. The formation of combined chlorine built up quickly then remained constant at a value of under 0.2 mg l-1 Cl2. We decided to consider the results by dividing the chlorine demand into two phases (Fig. 4):
-an initial phase of immediate consumption during the first 4 h, called the initial chlorine demand (ID) -a second slow consumption phase after the first 4 h, called the long term chlorine demand (LTD). The LTD has been interpreted with the following kinetic law (Jadas-Hecart, 1989)
[GRAPHICS]
where b (unknown value) = the content in precursors of LTD or, again, the potential or LTD after 4 h a = the total residual chlorine at 4 h x = chlorine consumption after 4 h n = stoichiometry k = the rate constant alpha, beta = partial orders of reaction. Assuming n = 1, the kinetics were found to be consistent with an overall second-order rate law (alpha = beta = 1). The b value could be evaluated from the ultimate chlorine consumption of the experimental data (computer calculator leading to the best correlation). Table 1 gives values of k, b and ID for 1987 and 1989. Table 2 presents the mean values where b/TOC does not seem to be a function of sampling period (1987: 1.2 mg Cl2/mg TOC; 1989: 1.4 mg Cl2/mg TOC). However the seasonal evolution of b and TOC was found to be difficult to correlate in spite of some analogies (Fig. 5).
As for the effect of pH (Table 5), dissociation of hypochlorous acid is probably the main factor responsible for the evolution of the apparent rate constant in the pH field studied (k decreases as pH increases between 7.5 and 8.5). However, this effect is slight. An increasing temperature, between 5 and 30-degrees-C, leads to an increasing apparent rate constant (Table 6) as in the case of reactions with simple compounds. Note that long-term chlorine demand (b) also increases.
Several studies about the chlorine demand of the organic material of GAC filtered water, after ultrafiltration, showed that the apparent molecular weight fraction lower than 1 kDa represents an important part of the total demand (ID + b), nearly 60% of that of GAC filtered water (Fig. 6 and Table 7).
There are several applications for this kinetic model. The first comes directly from the mathematical expressions of the apparent rate constant
[GRAPHICS]
and of the long term chlorine demand
[GRAPHICS]
From these expressions, and from the knowledge of the kinetic parameters (k(i) and b(i)) giving the chlorine consumption of complex compounds, it is possible to determine for a solution of these compounds, the part played by each one in the chlorine consumption observed in the mixture. We have shown that it was possible to determine the kinetic parameters of the solution of fulvic acid (extracted from the Seine river) and bovine serum albumin, on the grounds of results obtained for each of them (Table 4). Let us note that for these two compounds the total demand of chlorine (ID + b) is very different, as it is 1.4 mg Cl2/mg TOC for the Seine fulvic acid, while it is 6 mg Cl2/mg TOC for protein.
The second application is more practical as it is possible to use the kinetic parameters of water as an analytical measure to optimize the different steps of the drinking water treatment. In this way, it has been shown that the treatment coupling ozonation and GAC filtration significantly reduces the total chlorine demand of the water (cf. Tables 1 and 8). This is particularly observed during summer, showing the importance played by biological activities in activated carbon as shown by Bablon et al. (1987). Finally, the knowledge of the kinetic data of GAC filtered water allows the part played by water in the disappearance of chlorine in networks to be calculated. Indeed, chlorine stability in distributed water (under real conditions of chlorination and dechlorination applied in the plant) can be predicted by a simplified kinetic law of pseudo-first-order (Figs 7 and 8). Let us note that this equation does not describe the real disappearance of chlorine in networks but rather the part played by organic matter in water. The part due to the network itself (biofilm, material ...) is still to be determined.