filtered water caused a significant decrease in dissolved organic carbon (DOC), while it ... Key words--water treatment plant, BDOC, ozonation, double filtration.
Wat. Res. Vol. 31, No. 3, pp. 639-649, 1997
Pergamon PII: S0043-1354(96)00301-6
© 1997 ElsevierScienceLtd Printed in Great Britain.All rightsreserved 0043-1354/97 $17.00+ 0.00
EFFICIENCY OF VARIOUS WATER TREATMENT PROCESSES IN THE REMOVAL OF BIODEGRADABLE AND REFRACTORY ORGANIC MATTER F. RIBAS l*, J. F R I A S 2, J. M. H U G U E T I and F. L U C E N A 2 tSocietat General d'Aigiies de Barcelona, Passeig de Sant Joan 39,08009, Barcelona, Spain and 2Department of Microbiology, University of Barcelona, Av. Diagonal 645,08028, Barcelona, Spaint (First received October 1994; accepted in revised form September 1996)
A~tract--A new drinking water treatment train has been developed in the water treatment plant at Sant Joan Despi (Barcelona). It takes the raw water from the Llobregat river, which supplies water to Barcelona city and surroundings. We studied the efficiency of the different steps in this plant for the removal of biodegradable dissolved organic carbon (BDOC). The processes, at the beginning of the monitoring, were break-point prechlorination, flocculation-sedimentation, filtration through granular active carbon (GAC) and postchlorination. A sand filtration step and an ozonation step, between sand filtration and GAC filtration, were then added. Results show that the combined effect of the processes between raw and sand filtered water caused a significant decrease in dissolved organic carbon (DOC), while it had no effect on BDOC. The ozonation process significantly increased BDOC. Finally, GAC filtration significantly decreased DOC and BDOC. © 1997 Elsevier Science Ltd. All rights reserved Key words--water treatment plant, BDOC, ozonation, double filtration
INTRODUCTION
complex compounds into simpler molecules. Sequential treatment by ozonation and G A C adsorption or sand filtration has been incorporated in water treatment plants to improve their efficiency. Over the last few years, there has been increasing interest in the measurement of biodegradable organic matter (BOM) in natural waters (Huck, 1990). Bacterial growth in the distribution network may occur even at low concentration, and drinking water may harbor a large variety of bacteria. Several species of a large variety of genera have frequently been isolated from tap water (van der Kooij et al., 1982). Hence, different methods for measuring the biodegradability of the organic matter present in
Water engineers are now required to develop advanced treatment systems to increase the quality of drinking water and to counteract increasing pollution of raw water. The conventional treatment of raw water (break-point prechlorination, flocculationsedimentation and sand filtration) is enough to remove suspended and colloidal material from raw water. However, the presence of organic compounds which could be major sources for the formation of mutagenes and trihalomethanes, and organic cornpounds which could be used by microorganisms as a source of carbon, has necessitated the development of new systems to remove compounds of this kind. Granular activated carbon (GAC) has been used for this purpose because it is the ideal substrate to remove organic compounds due to its large specific area (Servais et al., 1991). To increase the performance of G A C the use of ozonation before G A C filtration is recommended (Maloney et al., 1985; Yasui and Miyaji, 1992). However, the ozonation process has been shown to decrease the ability of G A C to remove halogenated organics (Maloney et al., 1985). Ozone has been used to increase the biodegradability of dissolved organic carbon (DOC) by breaking *Author to whom all correspondence should be addressed, [Tel.: (393)2658011]. tTel.: (393)4021489.
water have recently been developed. These techniques measure the BOM by means of the growth of an inoculum (assimilable organic carbon, AOC)(van der Kooij et al., 1982; Werner, 1985) or by means of the activity (DOC consumption)of indigenous bacteria (biodegradable dissolved organic carbon, BDOC) (Frias et al., 1992; Joret and L6vi, 1986; Ribas et aL, 1991; Servais et al., 1987, 1989). The determination of AOC and BDOC in water is of particular concern to the water industry because biodegradable organic matter is almost the only limiting factor for heterotrophic bacterial growth in drinking water distribution systems which produce changes in the water quality (van tier Kooij et al., 1982). The source of these AOC and BDOC could be the initial water or any step of the treatment processes and it is important to determine the evolution of these
639
640
F. Ribas et al.
parameters through the water treatment train as early as possible and to detect the steps at which B D O C appears, in order to facilitate decisions about possible treatment strategies and to give rapid responses to changes in water quality. A comparison of several methods for A O C and B D O C determination, using various types o f water showed that the biodegradable matter values given by A O C measurements are substantially lower than those obtained using B D O C methods but that the various B D O C methods, including a dynamic method (Ribas et al., 1991), do not differ significantly in their results (Frias et al., 1995). This dynamic method, especially suited to the needs of the water treatment industry, was recently proposed by the authors (Lucena et al., 1990; Ribas et al., 1991). The method measures the B D O C of
circulating water continuously pumped through a fixed biomass bioreactor system o f two glass columns arranged in series, which are filled with an inert standard support. In this case the refractory D O C (rDOC), which is the fraction o f D O C not biodegradable, is the D O C at the outflow o f the reactor and the B D O C is the difference between inflow and outflow D O C . The method has given good performance over long periods o f time and, in addition to the almost instantaneous determination of B D O C , it has the advantage of the continuous monitoring for B D O C values for different types of water over time (Lucena et al., 1990).
Therefore, this bioreactor is especially suitable to monitor the processes in the drinking water treatment train. Due to the few studies performed on this field and the interest in measuring this parameter, the purpose of the present study was to monitor the evolution o f B D O C throughout the various steps o f a water treatment plant in which the processes were changed during the monitoring period, The water treatment plant monitored is in Sant Joan Despi (Barcelona). It takes the raw water from the Llobregat river, which supplies water to Barcelona city and surroundings and is severely polluted, bearing effluents from various industries and domestic wastewater, The processes at the beginning of monitoring were prechlorination, flocculation-sedimentation, filtration through granular active carbon (GAC) and postchlorination. In that period, as the bioreactor was not yet set up, the B D O C was measured by batch methods. M o r e recently, when the bioreactor method had been developed, an additional sand filtration step and an ozonation step between sand filtration and G A C filtration were incorporated to the process. This was to bring the treatment into line with A3 river water quality as defined by EC legislation. As the ozonation equipment was constructed after the new filters, it was possible, in an intermediate treatment train, to study the effect of double filtration (sand and G A C ) without ozonation.
MATERIALS AND METHODS A O C determination
AOC was measured according to the method from van der Kooij et al. (1982) which uses the growth of Pseudomonas fluorescens P17 as a measurement of this parameter. BDOC determination
BDOC was measured according to several methods previously described. The method from Joret and Lrvi (1986)uses as an inoculum sand from a treatment plant filter and measures BDOC after 10 days of incubation. The method from Servais et al. (1987) use water sample as inoculum, measuring BDOC after 21 days of incubation. Finally, the dynamic method to determine BDOC in waters (Ribas et al., 1991) uses a bioreactor with two columns arrayed in series and filled with an inert support. Water samples were continuously pumped upwards by a positive displacement diaphragm pump, at a low flow rate. The BDOC was measured as the difference between the dissolved organic carbon (DOC) at the inlet of the reactor and at the outlet, after 2 h (time of residence). The experiment wasperformedatroomtemperature(21 + 2°C). In addition refractory DOC (rDOC) can be defined as the fraction of DOC not biodegradable by the bacterial inoculum, corresponding to the minimum value of the DOC observed. Thiosuifate is commonly used to dechlorinate samples of water before biological analysis, including AOC (APMA., 1995); that was the reason why in chlorinated samples a sterile thiosulfate solution of 1 g/L was continuously dosed to circulating water to neutralise chlorine (1-2ppm). All the determinations of the dissolved organic carbon (DOC) were performed with a Model 700 TOC Analyzer, O.I. Corporation, College Station, Texas, calibrated between 0 and 10 ppm of carbon, with a precision of at least 0.1 ppm. Before the analysis, samples were filtered through a porous membrane (0.45 #m) previously washed with 400 mL of distilled water to avoid organic carbon contamination. Processes and sampling sites
The processes used in the various periods are presented in the Fig. 1. Table 1 summarizes the characteristics of the processes at successive periods and Table 2 presents the mean of the main parameters of Llobregat river water quality at those periods. In the first treatment train monitored, the BDOC was measured at four different steps in the water treatment process: 1) raw river water, 2) prechlorinated water with flocculants added, 3) sedimented water, 4) water filtered by granulated active carbon (GAC) and 5) finished (postchlorinated) water. Between February and August 1989 the AOC and BDOC of these samples were analyzed by van der Kooij et al. (1982) and by Joret and IAvi (1986) methods, respectively. Between January and September 1990 the BDOC of these samples was measured by the method of Servais et al. (1987). In the second and third treatment trains, the BDOC method used was that based on the fixed biomass reactor (Ribas et al,, 1991). In the second treatment train studied (between January and February 1992), when a sand filtration step was added to the process, the samples monitored were: 1) raw river water, 2) prechiorinated water with flocculants added, 3) GAC filtered water and 4)finished (postchlorinated)water. Finally, in the third treatment train studied, an ozonation step was added, between sand filtration and GAC filtration. In October 1992 the samples monitored were: 1) raw river water, 2) sand filtered water, 3) ozonated water, 4) GAC filtered water and 5) finished (postcfilorinated) water. Between October 1992 and April 1993 the samples
Removal of organic matter monitored were: 1) raw river water, 2) ozonated water, 3) finished (postchlorinated) water, and between March and June 1994 samples were: 1) sand filtered water, 2) ozonated water and 3)GAC filtered water,
combined effect of sedimentation and double filtration. The results of BDOC for raw water were lower than those obtained for the first treatment train, but the evolution of BDOC through the treatment was similar. Finally, the third treatment train (prechlorination + flocculation/sedimentation + sand filtration + ozonation + GAC filtration + postchlorination) was monitored, also more inten-
Statistics
Data are expressed as the arithmetic mean and coefficient of variation of the determinations corresponding to the different periods. A Student's t-test with paired samples was performed to assess the efficiencyof the removal of different DOC fractions (total DOC and biodegradable DOC) in the treatment steps and combinations of them,
sively in October 1992 than in the later periods, from October 1992 to April 1993 and from March to June 1994 (Table 5). Thus, in the first period (October 1992), as prechlorinated and sedimented water were not sampled, the comparison between raw and sand filtered water yielded information on the combined
RESULTS Table 3 summarizes the results obtained during the first treatment train (prechlorination + flocculation/ sedimentation + GAC filtration + postchlorination), During the period February-August 1989 BDOC measurement, using the method of Joret and Lrvi
effect of prechlorination, flocculation/sedimentation and sand filtration. In the second period (October 1992-April 1993), only raw, ozonated and postchlor~nated water were sampled, allowing the observation of the combined effects of prechlorination-floccula-
(1986), was compared, in the same samples, with AOC measurement using the method of van der Kooij et al. (1982). AOC values were lower than
tion/sedimentation-sand filtration-ozonation, and GAC filtration-postchlorination. Finally, in the third
BDOC values at the various steps, The relative increase in the AOC as a consequence of GAC
period, the monitoring of sand-filtered, ozonated and GAC-filtered water allowed us to check biological filtration more intensively.
filtration was higher than the increase in BDOC. In both cases a considerable removal of DOC, BDOC and AOC through the treatment was observed.
According to all the results reported in Tables 3:5, Tables 6 and 7 summarize, respectively, the efficiency of the different individual processes and possible combinations of them in the removal of different fractions of DOC. The data correspond to different treatment trains, periods and methods for AOC and,
From January to September 1990 the evolution of BDOC through the various steps was monitored in the same treatment train by the method of Servais et al. (1987). The results show that the evolution of BDOC was similar to that obtained using the method of Joret and Lrvi (1986) in the above cited period. On the other hand, these results are also higher than the AOC results corresponding to the above cited period. Table 4 shows the results obtained for the second treatment train (prechlorination + flocculation/ sedimentation + sand filtration + GAC filtration + postchlorination) using the dynamic method from Ribas et al. (1991). As sedimented and sand filtered water were not analyzed, the comparison between prechlorinated and GAC filtered water shows the
TRAIN
641
especially, BDOC determination. DISCUSSION The prechlorination process was monitored three times, during treatment trains 1 and 2. However, it is the first step in water treatment and its results that can be extrapolated to the new situation of treatment train 3, bearing in mind the differences in raw water quality.
1
[Riverwater~. ~ ~ ~ ~ ~ = ~ ~ - - , ~ . Prechl.
Floc. Sedimen.
TRAIN 2 ^ I River water~ . ~
~
Prechl. TRAIN 3 I River water~ . ~
GAC lilt.
~ ~
~
.
^ Finishedwat.er
Floc. Sand flit, Sedimen .
Postch.
.
GAC filt.
Postch.
~
Prechl. Floc. Sandflit. Ozon. GAC lilt. Postch. Sedimen. Fig. 1.
[[Finis hedwater]l
F. Ribas et al.
642
Table 1. Summary of the characteristics of the processes at successive periods
Processes Prechlorination (break point raw water)* Floceulationsedimentation
Train 1 Period 1 Feb.-Aug. 1989 24
Train 2 Period 2 Jan-Sept. 1990
Jan.-Feb. 1992
22
17
Settlers: 88 tanks TVST = 100 m 2TST = 8800 m 2 VT = 485 m 3 TSV = 42,680 m 3 AL, APC AL, APC AL, APC FeCI3 (p) FeCI3 FeCh LT-24 NO-12
Sand filtration
Oct. 1992
Train 3 Period 2 Nov. 1992Apr. 1993
Period 3 Mar.-Jun. 1994
7
12
15
Period 1
LT-24 NO-12
LT-24 NO-12
No
AL
LT-24 NO-12
Double filters: 20 SSF = 2 x 50 = 100 m 3 TSSF = 2000 m2 SFV = 9.5 m/h SMML = 2.5 mca SH = 0.6 m
Ozonation
No
MO3D = 4 ppm nO3 = 3 MPOj = 32 x 3 = 96 kgO3/h nC = 4 nATU = 2 nO3/W = 4 nO3D = 4
Double filters: 20 GAC filatration
Postchlorination (residual chlorine finished water)*
TCFS = 2000 m 2 Uncovered filters (light) CFV = 9.5 m/h CH=Im CMLL = 2.5 inca CT = 6rain, 20 s 2.54
AS = 2.16 m/h RTT = 2 h, 14 rain AL, APC AL, APC WAC, FeCI3 LT-24 LT-24 NO-12 NO-12
CFS = 2 x 50 = 100 m 2
Covered filters (no light) CFV = 10.8 m/h CH = 1.5 m CMLL = 3 mca CT = 8 rain, 20 s
2.85
1.75
1.38
1.46
1.24
All quantitative data reported to the treatment of 5.3 m3/s. *mgCl2/L. AL, aluminium sulphate (alum); APC, aluminium polychloride; WAC, aluminium polychlorosulphate; FeCh (p), treatment with FeCI3 simultaneous to the use of polyelectrolyte; TVST, total view surface of a tank; TST, total surface of the tanks; AS, ascending speed of the water in the prismatic part of the tank; VT, interior volume of the tank; TSV, total sedimentation; RTT, retention time of the water in the tanks; SSF, surface of a sand filter; TSSF, total surface of sand filtration; SFV, sand filtration speed; SMLL, maximum load loss in the filtering sand bed; SH, thickness of the sand filtering bed; MO3D, maximum ozone dose; nO3, number of ozonizers; MPO~, maximum ozone production in the installation (3 ozonizers); nC, number of contact chambers; nO3D, number of ozone destroyers; CFS, surface of a carbon filter; TCFS, total carbon filtration surface, CFV, carbon filtration speed; CH, thickness of the carbon filtering bed; CMLL, maximum load loss in the filtering bed; CT, contact time water/carbon.
We observed that prechlorination slightly dec r e a s e s t o t a l D O C ( b e t w e e n 5.7 a n d 8 . 8 % ) , a s a r e s u l t o f a d e c r e a s e in B D O C ( b e t w e e n 35 a n d 4 8 . 9 % ) a n d a slight i n c r e a s e in r D O C ( 1 - 1 4 . 5 % ) . Parallely, t h e %BDOC/DOC also decreases. These increments and decrements are always significant, except the DOC d e c r e a s e in t h e first p e r i o d o f t h e first t r e a t m e n t t r a i n , a n d t h e r D O C i n c r e a s e in t h e s e c o n d t r e a t m e n t train. The data on BDOC disagree with some data from t h e l i t e r a t u r e , w h i c h r e p o r t a n i n c r e a s e in B D O C as a r e s u l t o f c h l o r i n a t i o n ( L e C h e v a l l i e r et al., 1991; M e r l e t et al., 1991). H o w e v e r , A O C d a t a in t h e first p e r i o d s h o w a n o n s i g n i f i c a n t i n c r e a s e in B O M , w h i c h s u g g e s t s t h a t p r e c h l o r i n a t e d p r o d u c t s a r e m o r e easily b i o d e g r a d e d b y P s e u d o m o n a s f l u o r e s c e n s P17 t h a n b y i n d i g e n o u s b a c t e r i a p r e s e n t in e i t h e r s a n d a n d w a t e r i n o c u l a o r in t h e c o l o n i z e d b i o r e a c t o r , Another explanation of the BDOC decrease could be t h a t s o m e p r e c h l o r i n a t i o n p r o d u c t s , a f t e r t h i o s u l phate addition, partially stress the indigenous bacteria,
A l t h o u g h t h e d i f f e r e n c e s a r e less significant, t h e results of postchlorination, which was monitored f o u r t i m e s , a r e v e r y s i m i l a r to t h o s e o f p r e c h l o r i n a t i o n , in spite o f t h e v e r y different q u a l i t i e s o f t h e w a t e r to be c h l o r i n a t e d a n d t h e different d o s e s o f chlorine used. Therefore, general conclusions about t h e c h l o r i n a t i o n o f w a t e r c o u l d be o b t a i n e d . I n t h e c a s e o f p o s t c h l o r i n a t i o n , t h e w a t e r to be c h l o r i n a t e d flows f r o m t h e o u t l e t o f G A C filters in t h e different treatment trains studied. A slight decrease w a s o b s e r v e d in t o t a l D O C (2.4-10%), only s i g n i f i c a n t in o n e o u t o f f o u r series o f o b s e r v a t i o n s . T h e B D O C d e c r e a s e d b e t w e e n 23.9 a n d 8 0 . 8 % , b u t this d e c r e a s e is o n l y s i g n i f i c a n t in t r e a t m e n t t r a i n s 2 a n d 3. T h e % B D O C / D O C a l s o d e c r e a s e s signific a n t l y in t h e s a m e t r e a t m e n t t r a i n s . T h e r D O C i n c r e a s e d slightly in t h r e e o u t o f f o u r o b s e r v a t i o n s b u t o n l y s i g n i f i c a n t l y in t r e a t m e n t t r a i n 3. A l t h o u g h in t h e t w o c h l o r i n a t i o n s t e p s were t h e quality of the chlorinated water and the chlorination d o s e different, t h e effects o n t h e different f r a c t i o n s o f
•Dissolved organic carbon. bHeterotrophic plate count.
Flow (m~/s) Temperature (°C) Turbidity (NTU) pH TAC (mg CaCO3/L) Conductivity ~ S / c m ) Dissolved 02 (mg/L) (%),,,~ Ammonium (mg NH3/L) DO(2' (mgC/L) HPC ~ 22°C (UFC/mL) Fecal coliforms (MPN/100mL)
4.9 19.9 125 7.82 214 3430 5.8 59 2.17 8.98 7.4 x 105 6.4 x l0 s
5.7 19.5 148 7.78 216 2507 6.2 63 1.86 7.32 1.2 x 106 1.1 x 106
Train 1 Period 1 Period 2 Feb.-Aug. 1989 Jan.-Scp. 1990 16.5 7.3 130 7.92 238 1704 10.0 84 1.49 6.80 1.1 x 106 3.3 x 10~
Jan.-Feb. 1992
Train 2
31.6 15.2 114 8.11 198 1095 7.7 72 0.56 4.39 2.7 x l0 s 3.9 x 10~
Period 1 Oct. 1992
18.9 10.0 118 7.97 232 1395 8.3 73 1.15 4.54 6.2 x 10~ 4.5 x 104
Train 3 Period 2 Nov. 1992-Apr. 1 9 9 3
Table 2. Mean of the main parameters of Llobregat river water quality at different periods of the study
5.6 17.2 84 7.95 206 1618 5.5 56 1.24 7.70 2.7 x l0 s 8.0 x 104
Period 3 Mar.-Jun. 1994
o~
~"
o< ~_. o o t~
644
F. Ribas et al. Table 3. Organic matter measurements in the first treatment train First period: February-August 1989 Water samples (n = 12) Raw Prechlorinated Sedimented GAC-filtered Postchlorinated
Joret and Ldvi (1986) method (10 days) (mgC/L) DOC rDOC BDOC %BDOC/DOC 8.98 (0.36) 5.35 (0.31) 3.63 (0.60) 39.3 (0.35) 8.47 (0.33) 6.11 (0.24) 2.36 (1.00) 24.7 (0.70) 7.34 (0.25) 5.98 (0.35) 1.36 (0.60) 19.9 (0.69) 6.54 (0.37) 4.78 (0.29) 1.76 (0.75) 25.6 (0.41) 6.35 (0.31) 5.01 (0.27) 1.34 (1.03) 19.6 (0.78)
van der Kooij et al. (1982) method (mgC/L) AOC 0.96 (0.69) 1.27 (1.52) 0.31 (1.74) 0.87 (1.15) 0.19 (1.74)
Second period: January-September 1990 Servais et aL (1987) method Water samples (mgC/L) (n = 12) DOC rDOC BDOC %BDOC/DOC Raw 7.32 (0.19) 4.54 (0.22) 2.78 (0.33) 37.6 (0.22) Prechlorinated 6.75 (0.20) 5.20 (0.26) 1.55 (0.45) 23.4 (0.46) Sedimented 6.67 (0.13) 5.12 (0.22) 1.55 (0.42) 23.8 (0.44) GAC-filtered 6.15 (0.17) 4.31 (0.22) 1.84 (0.47) 29.6 (0.39) Postchlorinated 5.52 (0.23) 4.17 (0.31) 1.35 (0.76) 29.4 (0.60) Data expressed as arithmetic mean (coefficientof variation). n, number of samples analyzed at each sampling point; DOC, dissolved organic carbon; rDOC, refractory dissolved organic carbon; BDOC, biodegradable dissolved organic carbon; %BDOC/DOC, percentage of BDOC to DOC; AOC, assimilable organic carbon.
the dissolved organic matter were similar. The lower decrease in absolute values for postchlorination was probably due to the lower D O C values in the water to be chlorinated. Looking at the percentages more than the absolute values, the similarity of chlorination processes is even higher, It is surprising that in the flocculation-sedimentation process no significant differences were observed in any fractions of organic matter. Further, differences between two periods of the same treatment train were detected. In period 1 important (but not significant) decreases in D O C (1.13 mgC/L) and in B D O C (1.0 mgC/L) were observed, while in the second period the D O C decrease was very slight (less than 0.1 mgC/L) and B D O C remained constant (which means a slight increase in % B D O C / D O C ) . r D O C decreases are very similar in both periods, However, in the first period the A O C decrease is significant with 0.1 > p > 0.05. Decreases of A O C (Adam and Kott, 1989; Huck et al., 1991) and both r D O C and B D O C (Merlet et al., 1992) have been reported, in agreement with our results, Due to the sampling strategies, the efficiency of sand filtration as an independent step was not monitored in any treatment train using this process. In the first period o f the third treatment train the combined effect of prechlorination + sedimentation
+ sand filtration can be evaluated, and the effect of the sand filtration can be calculated by subtracting the effect ofprechlorination + sedimentation. During this period, chlorine in the water after sand filtration reached 0.29 + 0.21 ppm (n = 347) of combined chlorine, reaching free chlorine values o f 0. In the combined processes the decrease is significant in D O C , r D O C and BDOC, and o f approximately the same percentage (between 18.8 and 19.5%). Considering that the prechlorination and sedimentation processes are more efficient in removing B D O C than rDOC, a higher elimination of B D O C in the sand filtration process is possible. The ozonation process was monitored in the third treatment train in two periods (1 and 3). The r D O C decreased significantly and the B D O C also increased significantly in both periods, but at higher levels of B D O C in period 3. In absolute values the decrease in r D O C is very similar to the increase in B D O C , but, as the latter is lower in sand filtered water, the relative increase in B D O C is higher (53.8-63.6%) than the decrease in r D O C (5-21.2%). The ability o f ozone to increase the biodegradability o f organic matter has long been known (Boere, 1990; Hozalsky et al., 1992; Maloney et al., 1985; Neukrug et al., 1984; Rittman and Snoeyink, 1984; Wang et al., 1986), as has its
Table 4. Organic matter measurements in the second treatment train Water Ribas et aL (1991) method samples* (mgC/L) (n = 12) DOC rDOC BDOC %BDOC/DOC Raw 6.80 (0.09) 5.45 (0.19) 1.35 (0.64) 20.0 (0.60) Prechlorinated 6.20 (0.09) 5.51 (0.15) 0.69 (0.71) 11.4 (0.77) GAC-filtered 5.90 (0.12) 4.64 (0.20) 1.26 (0.60) 21.5 (0.53) Postchlorinated 5.50 (0.10) 4.73 (0.15) 0.77 (0.40) 14.2 (0.44) 'Sedimented and sand-filtered water were not monitored. See footnote of Table 3 for abbreviations.
Removal of organic matter
645
Table 5. Organic matter measurementsin the third treatment train First period: October 1992 Water Ribas et al. (1991) method samples' (mgC/L) (n = 15) DOC rDOC BDOC %BDOC/DOC Raw 4.39 (0.14) 3.75 (0.18) 0.64 (0.48) 14.7 (0.46) Sand filtered 3.54 (0.17) 3.02 (0.12) 0.52 (0.56) 14.0 (0.41) Ozonated 3.67 (0.15) 2.87 (0.11) 0.80 (0.37) 21.3 (0.25) GAC-filtered 2.48 (0.11) 2.22 (0.11) 0.26 (0.77) 10.1 (0.72) Postchlorinated 2.42 (0.09) 2.37 (0.11) 0.05 (2.20) 2.1 (2.19) Second period: November 1992-April 1993 Water Ribas et samplesb (n = 12) DO(2 rDOC Raw 4.54 (0.15) 3.91 (0.21) Ozonated 4.29 (0.17) 3.16 (0.12) Postchlorinated 2.84 (0.17) 2.55 (0.13)
al. (1991) method
(mgC/L) BDOC 0.63 (0.48) 1.13 (0.41) 0.29 (1.17)
%BDOC/DOC 14.4 (0.53) 25.3 (0.29) 9.2 (1.13)
Third period: March-June 1994 Water Ribas et al. (1991) method samplesc (mgC/L) (n = 18) DOC rDOC BDOC %BDOC/DOC Sand filtered 6.34 (0.11) 4.94 (0.12) 1.40 (0.39) 21.9 (0.34) Ozonated 6.19 (0.09) 3.90 (0.11) 2.29 (0.25) 36.7 (0.20) GAC-filtered 2.94 (0.21) 2.27 (0.18) 0.67 (0.49) 22.0 (0.36) 'Prechlorinated and sedimentedwater were not monitored. bPrechlorinated,sedimented,sand filtered and GAC filtered water were not monitored. cRaw, precblorinated,sedimentedand postchlorinatedwater were not monitored. See footnote of Table 3 for abbreviations.
ability to promote bacterial regrowth (Buydens, 1972). More recently, the same p h e n o m e n o n has been formulated as an increase in A O C (Huck, 1990; Huck et al., 1991; Krasner et al., 1993; LeChevallier et al., 1991, 1992; Miltner et al., 1991; Reasoner and Rice, 1989; Yasui and Miyaji, 1 9 9 2 ) o r BDOC (Bonnet et al., 1992; Merlet et al., 1992; Miltner et al., 1991; Prevost et al., 1992; Servais et al., 1991a, b, 1992; Volk et al., 1993). The interpretation of these results assumes that part of the refractory carbon is transformed into biodegradables, as a consequence of the breakdown and chemical transformation of complex molecules by ozonation. Indeed, ozone reacts with natural organic matter, generating some lower molecular weight, highly oxidized compounds that are nutrients for bacteria (Miltner et al., 1992). For example, humic acids produce alkanes, aliphatic aldehydes, ketones and fatty acids as by-products (Killops 1986; Lawrence e t al., 1980), which are easily removed by biotreatment (Miltner et al., 1991). Low molecular weight compounds are more easily transported across the cell membrane and attacked by metabolic enzymes (Hozalsky et al., 1992). In the first period a slight, but significant, increase (mean 0.13 ppm) in total D O C was also observed, This observation agrees with previous reports in which no explanation of the phenomenon is offered (Bonnet et al., 1992). Ozonation was, in that period, the only step in which an increase in absolute values of D O C was observed. In this case, if the rDOC data are observed, it is very interesting to point out that this fraction substitutes the total D O C as the only
fraction that continuously decreases through all the sampling points sequence. As, apparently, sand filtered water does not contain particulate carbon, an easy interpretation of D O C increase, assuming that some particulate carbon is transformed into dissolved carbon by ozone, is not probable. A better interpretation could be that organic matter obtained by ozonation is more easily oxidized in the TOC meter used in this work, which does not use combustion but chemical oxidation. However, in the second period a slight, but significant, decrease of D O C (mean 0.15 ppm) was observed. The G A C filtration process has a different performance after ozonation than after flocculationsedimentation, as the quality of the inflow water to the G A C filters is also different: either sedimented water with some turbidity due to residual flocks or sand-filtered water, which is not turbid and has been ozonated. On the other hand, the ecological conditions inside the filter are also different, as the sedimented water does not confer aerobic conditions through the whole of the filter, while the ozonated water confers aerobic conditions to all depths of the G A C filters (biological filters). The increase in biodegradation rates for G A C filters relative to nonabsorbing media may be due to a more efficient use of sorbed substrates, higher surface area and a more favorable acclimation environment (LeChevallier et al., 1991). An increase of BDOC for trains 1 and 2 during G A C filtration was observed, probably due to the biological activity of the filters. Due to the adsorbent properties of G A C filters, in
rDOC
--0.15 -0.65 0.15
DOC
0.13 -1.19 -0.06
Period 1
--1.20 0.23
%
5.70 -6.0
BDOC BDOC/DOC NT NT NT 0.28 7.30 --0.54 -11.2 --0.21 --8.10
0.40 -0.42
The same units from Table 2 (mgC/L for AOC, DOC, rDOC and BDOC). In bold type significant increments (p < 0.05). "Significant increment 0.1 < p < 0.05. NT, not tested.
processes Prechlorination Floccul./Sedim. Sand filtration Ozonation G A C filtration Postehlorinafion
Individual
-0.80 -0.19
--14.7 -4.80
G A C filtration Postchlorination
-1.27" - 1.00
-0.51 - 1.13
Preehlorination Floceul./Sedim. Sand filtration Ozonation
0.76 -0.13
BDOC/DOC
DOC
processes
BDOC
% rDOC
Treatment train 1
Individual
Period 1
DOC
Period 2
0.56 ~ -0.68
0.31 -0.96 ~
PI 7
AOC
rDOC
--0.52 --0.63
--0.57 -0.08
DOC
Period 2
0.29 -0.49
--1.23 0.00
BDOC
%
DOC NT NT NT --0.15 --3.25 NT
Period 3
5.80~ -4.40
--14.2 0.40
BDOC/DOC
BDOC BDOC/DOC NT NT NT NT NT NT
%
Treatment train 3
--0.81 -0.14
0.66 -0.08
rDOC
DOC
-- 1.05 --1.63
rDOC
-0.40
--0.60
Treatment train 2
0.89 --1.62
BDOC
0.09
0.06
rDOC
14.8 --14.8
%
--7.20
-8.60"
BDOC/DOC
BDOC/DOC
%
NT --0.49
NT NT
--0.66
BDOC
Table 6. Summary of the efficiency of different processes in the removal of different fractions of D O C (different treatments, periods and methods for BDOC determinations)
.~
o~
DOC
BDOC --0.75 --0.26
-0,12 -0.47 0.16 --1.53 --0.59
--0.5 -0.8
-0.73 -0.65 --0.88 -1.91 --1.38
--2.29
-0.34
rDOC
-2.27 -0.60 -0.02 -1.87 - 1.02
BDOC
0.63 --1.33 -0.97" -0.57 --1.10
rDOC
Treatment train I
The same units from Table 2 (mgC/L for AOC, DOC, r D O C and BDOC). In bold type significant increments (p < 0.05). •Significant increments 0.1 < p < 0.05. NT, not tested.
Period I Combined processes D(X~ Prechl. + Floc./sed. Floc./sed. + G A C tilt. -1.25 G A C tilt. + postchl. + ozon. + G A C tilt -1.06 Prechl, + floc./sed. + G A C flit. Floc./sed. + GACfilt. + postchl. Floc,/sed. + sand tilt. + G A C flit. Prechl. + floc./sed. + sand tilt. --0.85 Ozon. + G A C flit. + postchl. --1.12 Prechl. + floc./sed, + sand tilt, + G A C tilt. Prechl. + floc./sed. + sand flit. + ozon. --0.72 Prechl. + floc./sed. + sand tilt. + ozon. + G A C flit. All treatment --1.97
Prechl. + Floc./sed. --1.64 Floc./sed. + G A C flit. --1.93 G A C flit. + postchl. + ozon. + G A C tilt -0.99" Prechl. + floc./sed. + G A C tilt. --2.44 Floc./sed. + GACfilt. + postchl. --2.12 Floc./sed. + sand flit. + G A C tilt. Prechl. + floc./sed. + sand tilt, Ozon. + G A C tilt. + postehl. Prechl. + floc./sed. + sand flit. + G A C tilt. Prechl. + floc./sed. + sand tilt. + ozon. Prechl. + floc./sed. + sand flit, + ozon. + G A C flit. All treatment --2.63
Combined processes
Period 1
6.5 --0.38 --12.7
-0.7 --12.0
--19.2 -3.9
% BDOC/DOC
--19.7
-19.4 0.70 -0.30 -13.7 -2.9
% BDOC/DOC
-0.25 -4.6 --1.7
-1.45
DOC
Period 2
--0.77
-0.65 -0.40 -0.12 -0.09 - 1.08"
AOC P17
--1.36
--0.75
-0.61
rDOC
--1.8
--0.65 -0.60 --1.15 --1.17 --1.23
DOC
Period 2
--1.43
--1.23 0.29 -0,20 -0.94 -0.20
BDOC
0.50 NT --0.34
NT NT
-0.84 NT
BDOC
-5.2
10.9
-16.1
NT NT NT
NT NT
NT --3.40
I:K)C
Period 3
--13.3
-13.8 6.20" -0.30 --8.4 - 1.0
% BDOC/DOC
% BDOC/DOC
Treatment train 3
-0.37
0.6 --0.89 --0.95 -0.23 --1.03
rDOC
-2.67
rDOC
--0,73
BDOC
-0.72"
-0.81
-0.9
--1.3
-0.86
rDOC
-0.29"
DOC
Treatment train 2
0.57'
0.1
-5.8
1.5
10.1
% BDOC/DOC
% BDOC/DOC
-0.58'
-0.09
NT
NT
NT
BDOC
Table 7. Summary of the efficiency of differeent combinations of processes in the removal of different fractions of D O C (different treatments, periods and methods for B D O C determination)
~.
o
o[~ ~
