Le carbone organique: Son interet dans le controle de la preozonation

Abstract

The present study was performed on Seine River water and cultured water where algal growth was favoured. Multiple samples were collected during a 7-month period (November 1982–May 1983) at the Choisy-le-Roy water treatment plant. The effects of preozonation were assessed in all samples by measuring total organic carbon (TOC) and dissolved organic carbon (DOC). The latter measurements were made in ampoules on 10 ml samples after acidification and oxidation. Two different oxidation processes were tested: oxidation process No. 1 was made in the presence of 0.2 g of potassium peroxodisulphate during 5 h at 130°C; oxidation process No. 2 was made in the presence of 0.4 g of potassium peroxodisulphate during 7 h at 130°C. The CO 2 generated in the samples was measured with an infrared detector. This protocol was performed on raw water and on preozonated water. We thus had 8 variables: DOC and TOC measured in raw water and in preozonated water after two different oxidation treatments. Particular attention was paid to statistical analysis of data to improve their interpretation. Multivariate statistical analysis (principal components analysis) showed good correlations between variables ( P < 0.01) (Table 5) and the importance of tested parameters on organic carbon values 98% of the total inertia is due to sampling period (75%), filtration (12.8%), different oxidation processes used for organic carbon measurement (7.5%) and preozonation (2.6%) (Table 6, Fig. 1). Representation of the samples in the principal planes (Figs 2, 3 and 4) showed gathering of points corresponding to the same sampling period (i.e. November, January, March, April, May), the points related to November and March being opposite. It seems thus they have a different behaviour with regards to the tested parameters. This was confirmed by values in Table 7 concerning mean ± SEM values of the predefined sampling periods. The univariate statistical analysis (variance analysis and paired Student test) showed the variability of the Seine River and the presence, throughout the tested period, of a significant amount of particulate organic carbon (Table 9). The influence of oxidation performed before the organic carbon measurement is generally more important for TOC evaluation than for DOC (Table 10). Table 11 confirms this result for November. The higher preozonation effect appears in DOC measured in Seine River water after the stronger oxidation (i.e. oxidation No. 2) (Table 12). Table 13 shows that on March samples this effect appears only in DOC measured after oxidation process No. 2. Increasing the preozonation rate did not reveal this effect (Table 14). The two oxidation processes tested were not sufficient for exact determination of TOC but they allowed us to discuss preozonation effect on organic matter, this effect being different in function of the sampling period. The statistical analysis used made possible the quantitative evaluation of the preozonation effect which is generally only suggested.