- Email: [email protected]
hydrogen peroxide process by decomposition by-products
PII: S0043-1354(01)00087-2
Wat. Res. Vol. 35, No. 15, pp. 3587–3594, 2001 # 2001 Elsevier Science Ltd. All rights reserved Printed in Great Britain 0043-1354/01/$ - see front matter
EVALUATION OF THE TREATMENT PERFORMANCE OF A MULTISTAGE OZONE/HYDROGEN PEROXIDE PROCESS BY DECOMPOSITION BY-PRODUCTS KOJI KOSAKA1*, HARUMI YAMADA1, KENICHI SHISHIDA2, SHINYA ECHIGO3, ROGER A. MINEAR3, HIROSHI TSUNO1 and SABURO MATSUI1 1 Research Center for Environmental Quality Control, Kyoto University, 1-2, Yumihama, Otsu, Shiga 5200811, Japan; 2 Takuma Co. Ltd., 2-2-33, Kinrakuji-cho, Amagasaki, Hyogo 6600806, Japan and 3 Department of Civil and Environmental Engineering, University of Illinois at Urbana-Champaign, and 1101 West Peabody Drive, Urbana IL 61801, USA
(First received 3 July 2000; accepted in revised form 19 January 2001) Abstract}The performance of a multistage ozone/hydrogen peroxide (O3/H2O2) process was evaluated with respect to total organic carbon (TOC) removal of waste waters. An aqueous humic acid solution (5.2 mg C l1 as TOC) and a sand filtered secondary sewerage effluent (5.6 mg C l1 as TOC) were used as model waste waters. Appropriate range of hydrogen peroxide (H2O2) dose at each stage depended upon the components of the tested solutions that changed as the process proceeded. Higher hydrogen peroxide dose was required at later stages in which low reactivity compounds with hydroxyl radical (HOd), low molecular fatty acids, were predominant. And, oxalic acid concentration related to H2O2 demand at later stages. This was assumed that the slow decomposition of oxalic acid was rate-determining step for TOC removal after its accumulation. Also, it is important to maintain dissolved ozone at low concentration for efficient TOC removal because rapid ozone consumption is required for the rapid formation of hydroxyl radical (HOd). # 2001 Elsevier Science Ltd. All rights reserved Key words}ozone/hydrogen peroxide process, multistage injection system, TOC removal, ozone, hydroxyl radical, humic acid, sand filtered secondary sewage effluent, oxalic acid
INTRODUCTION
It is indispensable to reduce organic matter loading (e.g., chemical oxygen demand (COD) and total organic carbon (TOC)) from waste water to protect the aquatic environment. However, conventional biological treatment process is not sufficient for COD and TOC removal when a high fraction of organic compounds in waste water is recalcitrant. For such effluent, tertiary treatment by physicochemical treatment processes are required to reduce the level of organic matter. Activated carbon treatment is an effective option, but it requires regeneration of the activated carbon. Ozonation is widely applied for disinfection and color removal. However, for COD or TOC removal, ozonation is not necessarily the best option because the reactivity toward organic compounds highly dependent upon the nature of organic compounds in waste water (e.g., ozone reacts with saturated organic compounds very slowly (Neta and Huie, 1988)).
*Author to whom all correspondence should be addressed. Tel.: +81-77-527-6730; fax: +81-77-524-9869; e-mail: [email protected]
An ozone/hydrogen peroxide (O3/H2O2) process has recently shown promise for this objective due to its capability to produce high levels of hydroxyl radical (HO), a strong oxidant (Gulyas et al., 1995; Shishida et al., 1999b). Chemical oxidation process using HO like O3/H2O2 process are comprehensively referred to as advanced oxidation processes (AOPs) (Glaze et al., 1987). In AOPs, the O3/H2O2 process is considered to be one of the most practical because of its simplicity (Masten and Davies, 1994). The development of a high performance system has been investigated from the point of application. A few studies showed that a multistage injection system (reactors in series) was one high performance method in the O3/H2O2 process (Furukawa et al., 1997; Shishida et al., 1999a). Among them, Shishida et al. reported that multiple addition of H2O2 was better for TOC removal than singlet addition of H2O2 only at the inlet (Shishida et al., 1999a). When a multistage injection system is applied, it is assumed that consideration of the change of the organic solutes during the process is required for effective treatment since the difference of the organic component affect the treatment condition in AOPs. For example, it was reported that an appropriate range of H2O2 concentration existed during the
3587
3588
Koji Kosaka et al.
O3/H2O2 process (Echigo et al., 1998) because H2O2 acts not only a HO generator, but also a HO scavenger during the O3/H2O2 process (Yao et al., 1992). The scavenging reactions are as follows: H2 O2 þ HOd ! H2 O þ HOd2
ð1Þ
d d HO 2 þ HO ! OH þ HO2
ð2Þ
HO2d
where is peroxyl radical (Yao et al., 1992). It was also reported that the appropriate range of H2O2 concentration depended upon the type of solutes and their concentrations (Echigo et al., 1998). Thus, it is assumed that the appropriate range of H2O2 concentration changes at each stage of the multistage system due to the change of the components of the solutes during the process. However, there is no detailed information available on the effect of the change of the solutes during the multistage injection system on the treatment efficiency or operating factors. In the present study, the effect of the change of the organic solutes (progression of the oxidation) during a multistage O3/H2O2 process on the effective TOC removal of humic acid, one main component of natural organic matter (NOM), was investigated. In particular, the relationship between the H2O2 demand at each stage and the progression of the oxidation during the process was examined. To represent the progression of the oxidation of the organic solutes, oxalic acid, a major oxidation byproduct from AOPs, was used. This is because oxalic acid accumulates easily in the system due to its lower reactivity with HO (Buxton et al., 1988) and assumes to be one of the key reactions for effective TOC removal in AOPs. Also, the importance of monitoring of dissolved ozone concentration to determine H2O2 demand at each stage was investigated. Furthermore, a multistage O3/H2O2 process was applied to a sand filtered secondary sewage effluent as an application to real water matrices.
produced from 99.99% oxygen by an ozone generator (OS1, Mitsubishi Co.), and was introduced from the bottom of the reactor through a Pyrex fritted glass dispersion tube. A magnetic stirrer was used for mixing. Concentration of ozone gas, its flow rate, and its dose were maintained at 0.56 mM (27 mg l1), 0.50 l min1, and 0.12 mM min1 (5.9 mg l1 min1), respectively, throughout this study. All experiments were performed at 20 28C. Experimental procedure The pH of the humic acid solution was adjusted at 7.0 using phosphate buffer solution (1.5 mM as final solution). Initial TOC of humic acid aqueous solution was 5.2 mg C l1. The sand filtered secondary sewerage effluent was tested without pH adjustment. To simulate a multistage injection system with a semi-batch ozone contactor, intermittent ozonation was performed. That is, ozone gas introduction (10 min) was repeated four times followed by a 5 min intermission after each ozone gas introduction. Mixing of the solution was continued during the experiment. The treatment time was defined as the total time of ozone gas introduction in this study. Hydrogen peroxide solution was added before starting the reaction and reintroducing ozone gas. Hydrogen peroxide for the first stage was mixed with the sample solution before introducing the sample solution into the reactor. For other stages, 1.0–2.0 ml of H2O2 solution was added from a top of the reactor with a glass syringe. Samples were collected from a side port outlet of the reactor. Sample analysis (except for dissolved ozone concentration) was conducted after nitrogen gas purge for removing dissolved ozone. Analytical methods Acetic, formic, and oxalic acids were analyzed using an ion chromatograph (DX-500, Dionex Co.) with a conductivity detector and an AS9-HC column (Dionex Co.) protected by AG9-HC guard column (Dionex Co.). The eluent was 9.0 mM sodium carbonate solution. Concentration of the H2O2 stock solution (30 wt%) purchased from Wako Chem. Co. was determined by its UV absorption at 240 nm (Bader and Hoigne´, 1988). Hydrogen peroxide concentration of the sample solutions was measured by
METHODOLOGY
Materials All reagents used in this experiment were of analytical grade. Phosphate buffer solution (pH 7.0) was prepared from sodium phosphate dibasic and potassium phosphate monobasic (Wako Chem. Co.). Humic acid was purchased from Aldrich Chem. Co. A sand filtered secondary sewage effluent was filtered through a glass filter (GF/C (1 1.2 mm), Whatman Co.) and the filtrate was used as a sample solution. TOC, IC, and pH of the filtrate were 5.6 mg C l1, 15 mg C l1, and 7.0, respectively. Deionized distilled water prepared by an ultrapure water system (CPW-200, Advantec Toyo Co.) was further purified by another ultrapure water system (Easypure RF, Barnstead Co.) and was used in this experiment. Apparatus Figure 1 shows the schematic of an ozone reactor used in this study. The volumes of a glass reactor and sample solution were 3.0 and 2.3 l, respectively. Ozone gas was
Fig. 1. Schematic of the ozone reactor.
TOC removal with a multistage O3/H2O2 process
3589
Figure 2 shows the profiles of the ratios of residual TOC to the initial TOC (residual TOC ratio) of humic acid aqueous solution during ozonation and the O3/H2O2 process. H2O2 dose at each injection was constant (150 mM). At 10 min, the residual TOC ratios were 0.68 for ozonation and 0.62 for the O3/H2O2 process. The difference of the residual TOC ratios between two processes was not significant, although the residual TOC ratio during the O3/H2O2 process was slightly lower than that during ozonation. On the other hand, a significant difference between two processes was observed at 40 min, i.e., the residual TOC ratio during the O3/H2O2 process was much lower than that during ozonation. Those values were 0.32 and 0.50, respectively. This result showed that the effect of H2O2 addition for TOC removal becomes more significant as oxidation proceeds. The similarity of the TOC profiles at the first stage of the two processes is attributed to the reactivity of humic acid. Humic acid contains reactive functional groups with ozone (e.g., phenolic groups), and sufficient amounts of promoters of radical chain reactions (e.g., H2O2 and superoxide anion (O2)) to yield HO are produced during the oxidation of those functional groups even without the addition of H2O2 (Staehelin and Hoigne´, 1985). The similar dissolved ozone concentration profiles of the two processes (Fig. 3) at the first stage supports that similar reactions occur regardless of the addition of H2O2.
Hence, in the first stage (e.g., during the initial 10 min), the difference of the TOC removal is not significant between the two processes. Also, to explain the better performance of the O3/ H2O2 process at later stages (i.e., at 30 and 40 min), acetic, formic, and oxalic acid concentrations were monitored (Fig. 4(a) and (b)). These compounds were assumed to consist of one major fraction of residual TOC at these stages, and the decompositions of these compounds were considered to be the rate-determining step of TOC removal. This assumption was based on the following considerations: (1) in general, the reactivity of oxidation by-products including low molecular fatty acids with ozone are relatively low except for formic acid (Buxton et al., 1988; Neta and Huie, 1988), (2) the reaction rate constants of acetic and oxalic acids with HOd are significantly low (8.5 107 and 7.7 106 M1 s1, respectively), compared with other organic compounds (Buxton et al., 1988)), (3) these compounds tend to accumulate in the O3/H2O2 process, since a greater number of HOd has to be yielded for the oxidation of a unit number of these carboxylic acids than other more reactive compounds due to their low reactivities in this competitive reaction system. The concentration profiles of the three carboxylic acids during ozonation and the O3/H2O2 process (Fig. 4(a) and (b)) verify the above assumption. The three oxidation by-products consisted of a considerable fraction of TOC after 20 min (more than 0.4 after 30 min during both two processes). Also, the behaviors of these three compounds were different after their formation. That is, the concentrations of these carboxylic acids were lower during O3/H2O2 process than during ozonation. For example, the concentration of acetic acid during ozonation and the O3/H2O2 process were around 7 and 4 mM, respectively. In case of oxalic acid, its concentration during ozonation increased with time and was more than 40 mM at 40 min. But that during the O3/H2O2 process decreased after 20 min and was less than 30 mM at 40 min. With these observations, it is reasonable to attribute greater TOC removal
Fig. 2. Profiles of the residual TOC ratios of humic acid solution during ozonation and the O3/H2O2 process.
Fig. 3. Dissolved ozone concentration profiles during ozonation and the O3/H2O2 process of humic acid solution.
the spectrophotometric method using copper(II) ion and 2,9-dimethyl-1,10-phenanthroline (Aldrich Chem. Co.) (Baga et al., 1988; Kosaka et al., 1998). Ozone concentration in the gas phase was measured by iodometry after trapping ozone gas in a potassium iodide solution. Dissolved ozone concentration was determined by the indigo method described by Bader and Hoigne´ (1981). TOC and IC were measured by a TOC analyzer (TOC-5000A, Shimadzu Co.).
RESULTS AND DISCUSSION
TOC removal of humic acid during ozonation and O3/H2O2 process
3590
Koji Kosaka et al.
efficiency of the O3/H2O2 process to its better oxidation efficiency of these carboxylic acids. Faster decompositions of oxidation by-products by the O3/H2O2 process were also in agreement with the dissolved ozone concentration profiles (Fig. 3). This suggests faster HOd production in the O3/H2O2 process. Dissolved ozone concentration was always higher during ozonation than during the O3/H2O2 process. The differences of dissolved ozone concentration between the two processes at 10 and 40 min were 25 and 80 mM, respectively. The difference of the dissolved ozone concentrations between the two processes became greater with time. This lower dissolved ozone concentration during the O3/H2O2 process indicates that ozone is converted to HOd more rapidly in the O3/H2O2 process than in ozonation, especially after the completion of the direct reactions between ozone and reactive functional groups of humic acid (note that the greater difference was observed at 40 min). In other words, the accumulation of ozone in the solution implies that the system does not produce HOd sufficiently. Also, it is to be noted that some oxidation byproducts that accumulate with time act as inhibitors of the radical chain reactions (Staehelin and Hoigne´, 1985). Thus, HOd formation by the addition of H2O2d becomes more important for TOC removal as oxidation proceeds.
Fig. 4. Profiles of the concentrations of acetic, formic, and oxalic acids during (a) ozonation and (b) the O3/H2O2 process of humic acid solution.
The above discussion led us to speculate that suppressing rapid ozone accumulation is one requirement for effective TOC removal by the O3/H2O2 process. However, low dissolved ozone does not always guarantee efficient TOC removal. The evaluation of the system has to incorporate H2O2 concentration because excess H2O2 acts as a HOd scavenger and consumes ozone ineffectively. Also, at early stage, molecular ozone may be more effective than HOd due to its selectivity for reaction sites of humic compounds. The effect of H2O2 concentration at each stage will be discussed more quantitatively in the next section. Effect of H2O2 addition modes on TOC removal of humic acid during the O3/H2O2 process The comparison between the O3/H2O2 process and ozonation in the previous section implies that optimal H2O2 dose increases with reaction time (number of stage) and the increase is related to (1) the change of H2O2 consumption (i.e., the difference of the H2O2 concentration at the beginning of the stage and residual H2O2 concentration after each ozone introduction) and (2) the formation and decomposition of low molecular fatty acids. To test this hypothesis, three different modes of H2O2 addition were compared: (1) H2O2 dose was decreased stepwise as 250 ! 200 ! 100 ! 50 mM (D mode); H2O2 dose was increased stepwise as 50 ! 100 ! 200 ! 250 mM (I mode); H2O2 dose was constant at 150 mM (C mode). Total H2O2 doses of three different modes were fixed at 600 mM. The order of TOC residual ratios of the three modes after 40 min treatment was I mode (TOC residual ratio was 0.26)5C mode (0.32)5D mode (0.36) (Fig. 5). This observation supports our hypothesis that optimal H2O2 dose increases with reaction time (number of stages) at a fixed total H2O2 dose. Next, the relationships between the increase of optimal H2O2 dose and its two reasons described in the earlier part of this section are discussed in detail below.
Fig. 5. Effects of H2O2 addition modes on the residual TOC ratios of humic acid solution during the O3/H2O2 process.
TOC removal with a multistage O3/H2O2 process
3591
Table 1. Profiles of the amount of H2O2 consumption of three H2O2 addition modes at each stage Number of stage D mode
I mode
[H2O2]0a
c
4
250
364
209
61 60 0.99
86 0.34
255 0.70
198 0.95
[H2O2]0a
50
118
206
255
DH2O2b
32 0.64
112 0.95
201 0.98
248 0.97
222
182
153
190 0.86
179 0.98
152 0.99
[H2O2]0a (mM) DH2O2 [H2O2]1c 0
b
3
(mM) DH2O2 [H2O2]1c 0
DH2O2b a
2
DH2O2b
(mM) DH2O2 [H2O2]1c 0
C mode
1
150 78 0.52
Initial concentration of H2O2. The amount of H2O2 consumption. Consumption ratio of H2O2.
To take a closer look of H2O2 consumption, the profiles of the initial H2O2 concentration of each stage, H2O2 consumption, and consumption ratio are summarized in Table 1. The initial H2O2 concentration is the sum of the residual H2O2 at previous stage and H2O2 dose at next stage. It is to be noted that H2O2 consumption ratio was higher at later stages regardless of the H2O2 addition modes. Also, despite the same total H2O2 dose (600 mM), the H2O2 consumptions were different among three modes from 30 to 40 min (at the last stage). H2O2 consumptions of I, C, and D modes during the fourth stage were 248, 152, and 60 mM, respectively. On the other hand, however, H2O2 consumptions of D and C modes were significantly higher than that of I mode at some early treatment stages, although the residual TOC ratios and the dissolved ozone concentrations were not quite different (Table 1, Figs 5, and 6) among the three modes. Therefore, it is assumed that H2O2 doses for D and C modes at some early treatment stages were in excess and H2O2 was ineffectively consumed while H2O2 was in short supply during the fourth stages. It is also to be noted that with an extremely high H2O2 concentration, TOC removal efficiency may be decreased since ineffective consumption of H2O2 includes ineffective consumptions of other oxidants (e.g., HOd). But, such an extreme effect by the excess H2O2 was not observed clearly in this experiment. Dissolved ozone concentrations at 40 min supported the order of TOC residual ratios of the three modes after 40 min treatment by suggesting that the HOd formation rate was the fastest in the I mode at the last stage (30–40 min). The order of dissolved ozone concentration inversely corresponded to the order of TOC removals and the order of H2O2 consumptions (Fig. 6). Since low dissolved ozone concentration indicates rapid formation of HOd, it is considered that from 30 to 40 min the amount of HOd formation of I mode was highest and hence the TOC residual ratio of I mode was lowest at 40 min. From
Fig. 6. Effects of H2O2 addition modes on dissolved ozone concentration during O3/H2O2 process of humic acid solution.
the above discussion on the profiles of the TOC, H2O2, and dissolved ozone, it was verified that under the constant ozone dose it was better to increase H2O2 dose stepwise for effective TOC removal because H2O2 demand also increased with time. Let us turn to the formation and decomposition of low molecular fatty acids. The increase of H2O2 demand with time was presumed to be related to the accumulation of oxidation by-products because they are less reactive with HOd than humic acid itself and greater amount of HOd is required for the decompositions of those less reactive compounds. Based on this presumption, the relationship between H2O2 demand and the accumulation of oxidation byproducts was investigated. Among the oxidation by-products, oxalic acid was chosen as an indicator of the extent of oxidation and the accumulation of the oxidation by-products was assumed to be expressed by the ratio of oxalic acid concentration (in mg C l1) to TOC. This is because oxalic acid accumulates as oxidation proceeds due to its significantly low reactivity with HOd among the major oxidation by-products during AOPs. This
3592
Koji Kosaka et al.
To test the practical applicability of the findings obtained from the experiments using humic acid (i.e., the better performance of the O3/H2O2 process on TOC removal over ozonation and the increase of H2O2 demand with time), a sand filtered secondary sewerage effluent was treated by ozonation and the O3/H2O2 process. The profiles of the residual TOC ratio given in Fig. 8 clearly showed the effect of the O3/H2O2 process on the treatment of the sand filtered secondary sewerage effluent compared to ozonation. In this experiment, H2O2 dose of each injection was constant (150 mM) and total H2O2 dose was 600 mM. The residual TOC ratio after the O3/H2O2 process was much lower than that after ozonation (i.e., at 40 min those values during the O3/H2O2 process and ozonation were 0.46 and 0.78, respectively). However, the effect of H2O2 addition on TOC removal of the sand filtered secondary sewerage effluent was
different from that for humic acid aqueous solution. For the humic acid solution, the effect of H2O2 addition was observed only at later stages, but for the sewerage effluent the effect of H2O2 addition was clearly observed from the first stage. Considering the similarities of pH and TOC levels of these two solutions, there are two possibilities to explain this difference of the effect of H2O2 addition on TOC removal. The first explanation is that the sand filtered secondary sewerage effluent contained bicarbonate ion (the IC was 15 mg Cl1), an inhibitor of radical chain reactions. As a result of the inhibition by bicarbonate, the effect of H2O2 addition can be observed from earlier stages because more initiator for radical chain reaction is required in this condition. The second possible reason is the difference of the characteristics of organic solutes. As seen from Figs 4 and 9, for the humic acid solution, only oxalic acid accumulated at high concentration; on the other hand, for the sand filtered secondary sewerage effluent, both acetic and oxalic acids were highly accumulated. This significant difference of the profiles of the low molecular fatty acids indicates the significant difference of their components of organic solutes. The difference of organic solutes of the two solutions is assumed to affect the difference of the effect of H2O2 addition on TOC removal. Figure 10 shows the profiles of dissolved ozone concentration during ozonation and the O3/H2O2 process. From Figs 8 and 10, the relationship between TOC removal efficiency and dissolved ozone concentration for humic acid solution was found to be applicable to the sand filtered secondary sewerage effluent. That is, dissolved ozone concentration during ozonation was always higher than during the O3/H2O2 process while its TOC removal efficiency was lower. For example, at 10 min, TOC removal during ozonation was negligible and the dissolved ozone already reached 70 mM. Therefore, it was confirmed that the profiles of dissolved ozone reflected the HO formation efficiency during these
Fig. 7. Relationship between ratio of oxalic acid to TOC at the beginning of each stage and H2O2 consumption during the stage.
Fig. 8. Profiles of the residual TOC ratios of the sand filtered secondary sewage effluent during ozonation and the O3/H2O2 process.
expression seems to be valid since fraction of oxalic acid increased with time until 40 min. Fig. 7 shows the relationship between the ratio of oxalic acid to TOC at the beginning of each stage and H2O2 consumption (i.e., H2O2 dose required) during the stage. The results of Fig. 5 and Table 1 were used after the exclusion of some results (the results of D and C modes at the fourth stage were excluded because of lower TOC removal than I mode, and that of D mode at the second stage were excluded because the significant amount of H2O2 is assumed to be consumed ineffectively). The linear relationship was observed between H2O2 demand and the ratio of oxalic acid concentration to TOC. Therefore, it was shown that the change of H2O2 demand of each stage is correlated to the progression of the oxidation and the progression of the oxidation means the accumulation of oxidation by-products with low reactivity with HOd like oxalic acid. TOC removal of a sand filtered secondary sewage effluent
TOC removal with a multistage O3/H2O2 process
3593
of oxalic acid to TOC is assumed to represent the accumulations of low reactive compounds with HOd (the progression of oxidation). Thus, it was shown that H2O2 demand of the sand filtered secondary sewage effluent increased with time and the increase of H2O2 demand was due to the progression of the oxidation like the case of humic acid solution. CONCLUSIONS
Fig. 9. Profiles of the concentrations of acetic and oxalic acids during (a) ozonation and (b) the O3/H2O2 process of the sand filtered secondary sewage effluent.
This research investigated the effect of the change of the organic solutes on the TOC removal efficiency or operating condition during a multistage injection O3/H2O2 system. During the process H2O2 demand at each stage increased under constant ozone dose. This was because the low molecular fatty acids, low reactivity compounds with HOd, became main components of organic solutes as the progression of the oxidation. In this experimental condition, the degree of the progression of the oxidation could be represented by the ratio of oxalic acid to TOC. This tendency was observed for both humic acid solution and a sand filtered secondary sewage effluent. Dissolved ozone concentration reflected the amount of ozone consumed for the reaction during the process. Therefore, an appropriate operational condition can be obtained by monitoring dissolved ozone and finding the condition that dissolved ozone was maintained at low concentration. In this experiment, the effect of the change of the organic solutes during the process was investigated from the point of H2O2 demand. As a future work, we would like to examine relationship between the change of the organic solutes during the process and other parameters (e.g., ozone demand for unit TOC reduction (DO3/DTOC)). REFERENCES
Fig. 10. Dissolved ozone concentration profiles during ozonation and the O3/H2O2 process of the sand filtered secondary sewage effluent.
processes. It is to be noted that during the O3/H2O2 process dissolved ozone was greatly increased after 20 min and became 85 mM at 40 min. Considering the accumulation of dissolved ozone and the constant H2O2 dose of each injection, it is assumed that H2O2 dose required during the O3/H2O2 process greatly increased after 20 min. The profiles of oxalic acid to TOC also greatly increased after 20 min (the ratios at 0, 10, 20, and 30 min were 0, 0.08, 0.21, and 0.24, respectively). As described in earlier sections, increase
Bader H. and Hoigne´ J. (1981) Determination of ozone in water by the indigo method. Water Res. 15, 449–456. Bader H. and Hoigne´ J. (1988) Photometric method for the determination of low concentrations of hydrogen peroxide by the peroxidase catalyzed oxidation of N,Ndimetyl-p-phenylenediamine (DPD). Water Res. 22, 1109–1115. Baga A. N., Johnson G. R. A., Nazhat N. B. and SaadallaNazhat R. A. (1988) A simple spectrophotometric determination of hydrogen peroxide at low concentrations in aqueous solutions. Anal. Chim. Acta 204, 349–353. Buxton G. V., Greenstock C. L., Hwlman W. P. and Ross A. B. (1988) Critical review of rate constants for reactions of hydrated electrons, hydrogen atoms and hydroxyl radicals (OH/O) in aqueous solution. J. Phys. Chem. Ref. Data 17(2), 513–886. Echigo S., Kosaka K., Yamada H., Matsui M. and Shishida K. (1998) The effect of water quality characteristics on the optimal hydrogen peroxide dose for the ozone/hydrogen peroxide process. J. Jpn Soc. Water Environ. 21, 444–449 (in Japanese). Furukawa S., Yasunaga N., Kawaai Y and Hirotsuji J. (1997) Highly efficient ozone-hydrogen peroxide reactor with multiple injection ports. Proceedings of the 13th Ozone World Congress, Vol. 1, pp. 499–503.
3594
Koji Kosaka et al.
Glaze W. H., Kang J. and Chapin D. H. (1987) The chemistry of water treatment process involving ozone, hydrogen peroxide and ultraviolet radiation. Ozone Sci. Technol. 9, 335–352. Gulyas H., von Bismarck R. and Hemmerling L. (1995) Treatment of industrial wastewaters with ozone/hydrogen peroxide. Water Sci. Technol. 32(7), 127–134. Kosaka K., Yamada H., Matsui S., Echigo S. and Shishida K. (1998) Comparison among the methods for hydrogen peroxide measurements to evaluate advanced oxidation processes: application of a spectrophotometric method using copper(II) ion and 2,9dimethyl-1,10-phenanthroline. Environ. Sci. Technol. 32, 3821–3824. Masten S. J. and Davies S. H. R. (1994) The use of ozonation to degrade organic contaminants in wastewaters. Environ. Sci. Technol. 28, 181A–185A. Neta P. and Huie R. E. (1988) Rate constants for reactions of inorganic radicals in aqueous solution. J. Phys. Chem. Ref. Data. 17(3), 1027–1284.
Shishida K., Echigo S., Yamada H. and Matsui S. (1999a) Study of the development of the high performance ozone/ hydrogen peroxide water treatment system for TOC removal: effect of the operational parameters. J. Jpn Soc. Water Environ. 22, 199–205 (in Japanese). Shishida K., Haruki H., Kamio K., Yamada H., Kosaka K. and Matsui S. (1999b) Comparison of the ozone/ hydrogen peroxide process, the ozone /ultraviolet process, and the hydrogen peroxide/ultraviolet process in the treatment of landfill leachate. J. Jpn Soc. Waste Manage. Experts 10(5), 247–256 (in Japanese). Staehelin J. and Hoigne´ J. (1985) Decomposition of ozone in water in the presence of organic solutes acting as promoters and inhibitors of radical chain reactions. Environ. Sci. Technol. 19, 1206–1213. Yao C. C., Haag W. R. and Mill T. (1992) Kinetic features of advanced oxidation processes for treating aqueous chemical mixtures. Chemical Oxidation: Technology for the Nineties, Proceedings of the Second International Symposium, Vol. 2, pp. 112–139.
Wat. Res. Vol. 35, No. 15, pp. 3587–3594, 2001 # 2001 Elsevier Science Ltd. All rights reserved Printed in Great Britain 0043-1354/01/$ - see front matter
EVALUATION OF THE TREATMENT PERFORMANCE OF A MULTISTAGE OZONE/HYDROGEN PEROXIDE PROCESS BY DECOMPOSITION BY-PRODUCTS KOJI KOSAKA1*, HARUMI YAMADA1, KENICHI SHISHIDA2, SHINYA ECHIGO3, ROGER A. MINEAR3, HIROSHI TSUNO1 and SABURO MATSUI1 1 Research Center for Environmental Quality Control, Kyoto University, 1-2, Yumihama, Otsu, Shiga 5200811, Japan; 2 Takuma Co. Ltd., 2-2-33, Kinrakuji-cho, Amagasaki, Hyogo 6600806, Japan and 3 Department of Civil and Environmental Engineering, University of Illinois at Urbana-Champaign, and 1101 West Peabody Drive, Urbana IL 61801, USA
(First received 3 July 2000; accepted in revised form 19 January 2001) Abstract}The performance of a multistage ozone/hydrogen peroxide (O3/H2O2) process was evaluated with respect to total organic carbon (TOC) removal of waste waters. An aqueous humic acid solution (5.2 mg C l1 as TOC) and a sand filtered secondary sewerage effluent (5.6 mg C l1 as TOC) were used as model waste waters. Appropriate range of hydrogen peroxide (H2O2) dose at each stage depended upon the components of the tested solutions that changed as the process proceeded. Higher hydrogen peroxide dose was required at later stages in which low reactivity compounds with hydroxyl radical (HOd), low molecular fatty acids, were predominant. And, oxalic acid concentration related to H2O2 demand at later stages. This was assumed that the slow decomposition of oxalic acid was rate-determining step for TOC removal after its accumulation. Also, it is important to maintain dissolved ozone at low concentration for efficient TOC removal because rapid ozone consumption is required for the rapid formation of hydroxyl radical (HOd). # 2001 Elsevier Science Ltd. All rights reserved Key words}ozone/hydrogen peroxide process, multistage injection system, TOC removal, ozone, hydroxyl radical, humic acid, sand filtered secondary sewage effluent, oxalic acid
INTRODUCTION
It is indispensable to reduce organic matter loading (e.g., chemical oxygen demand (COD) and total organic carbon (TOC)) from waste water to protect the aquatic environment. However, conventional biological treatment process is not sufficient for COD and TOC removal when a high fraction of organic compounds in waste water is recalcitrant. For such effluent, tertiary treatment by physicochemical treatment processes are required to reduce the level of organic matter. Activated carbon treatment is an effective option, but it requires regeneration of the activated carbon. Ozonation is widely applied for disinfection and color removal. However, for COD or TOC removal, ozonation is not necessarily the best option because the reactivity toward organic compounds highly dependent upon the nature of organic compounds in waste water (e.g., ozone reacts with saturated organic compounds very slowly (Neta and Huie, 1988)).
*Author to whom all correspondence should be addressed. Tel.: +81-77-527-6730; fax: +81-77-524-9869; e-mail: [email protected]
An ozone/hydrogen peroxide (O3/H2O2) process has recently shown promise for this objective due to its capability to produce high levels of hydroxyl radical (HO), a strong oxidant (Gulyas et al., 1995; Shishida et al., 1999b). Chemical oxidation process using HO like O3/H2O2 process are comprehensively referred to as advanced oxidation processes (AOPs) (Glaze et al., 1987). In AOPs, the O3/H2O2 process is considered to be one of the most practical because of its simplicity (Masten and Davies, 1994). The development of a high performance system has been investigated from the point of application. A few studies showed that a multistage injection system (reactors in series) was one high performance method in the O3/H2O2 process (Furukawa et al., 1997; Shishida et al., 1999a). Among them, Shishida et al. reported that multiple addition of H2O2 was better for TOC removal than singlet addition of H2O2 only at the inlet (Shishida et al., 1999a). When a multistage injection system is applied, it is assumed that consideration of the change of the organic solutes during the process is required for effective treatment since the difference of the organic component affect the treatment condition in AOPs. For example, it was reported that an appropriate range of H2O2 concentration existed during the
3587
3588
Koji Kosaka et al.
O3/H2O2 process (Echigo et al., 1998) because H2O2 acts not only a HO generator, but also a HO scavenger during the O3/H2O2 process (Yao et al., 1992). The scavenging reactions are as follows: H2 O2 þ HOd ! H2 O þ HOd2
ð1Þ
d d HO 2 þ HO ! OH þ HO2
ð2Þ
HO2d
where is peroxyl radical (Yao et al., 1992). It was also reported that the appropriate range of H2O2 concentration depended upon the type of solutes and their concentrations (Echigo et al., 1998). Thus, it is assumed that the appropriate range of H2O2 concentration changes at each stage of the multistage system due to the change of the components of the solutes during the process. However, there is no detailed information available on the effect of the change of the solutes during the multistage injection system on the treatment efficiency or operating factors. In the present study, the effect of the change of the organic solutes (progression of the oxidation) during a multistage O3/H2O2 process on the effective TOC removal of humic acid, one main component of natural organic matter (NOM), was investigated. In particular, the relationship between the H2O2 demand at each stage and the progression of the oxidation during the process was examined. To represent the progression of the oxidation of the organic solutes, oxalic acid, a major oxidation byproduct from AOPs, was used. This is because oxalic acid accumulates easily in the system due to its lower reactivity with HO (Buxton et al., 1988) and assumes to be one of the key reactions for effective TOC removal in AOPs. Also, the importance of monitoring of dissolved ozone concentration to determine H2O2 demand at each stage was investigated. Furthermore, a multistage O3/H2O2 process was applied to a sand filtered secondary sewage effluent as an application to real water matrices.
produced from 99.99% oxygen by an ozone generator (OS1, Mitsubishi Co.), and was introduced from the bottom of the reactor through a Pyrex fritted glass dispersion tube. A magnetic stirrer was used for mixing. Concentration of ozone gas, its flow rate, and its dose were maintained at 0.56 mM (27 mg l1), 0.50 l min1, and 0.12 mM min1 (5.9 mg l1 min1), respectively, throughout this study. All experiments were performed at 20 28C. Experimental procedure The pH of the humic acid solution was adjusted at 7.0 using phosphate buffer solution (1.5 mM as final solution). Initial TOC of humic acid aqueous solution was 5.2 mg C l1. The sand filtered secondary sewerage effluent was tested without pH adjustment. To simulate a multistage injection system with a semi-batch ozone contactor, intermittent ozonation was performed. That is, ozone gas introduction (10 min) was repeated four times followed by a 5 min intermission after each ozone gas introduction. Mixing of the solution was continued during the experiment. The treatment time was defined as the total time of ozone gas introduction in this study. Hydrogen peroxide solution was added before starting the reaction and reintroducing ozone gas. Hydrogen peroxide for the first stage was mixed with the sample solution before introducing the sample solution into the reactor. For other stages, 1.0–2.0 ml of H2O2 solution was added from a top of the reactor with a glass syringe. Samples were collected from a side port outlet of the reactor. Sample analysis (except for dissolved ozone concentration) was conducted after nitrogen gas purge for removing dissolved ozone. Analytical methods Acetic, formic, and oxalic acids were analyzed using an ion chromatograph (DX-500, Dionex Co.) with a conductivity detector and an AS9-HC column (Dionex Co.) protected by AG9-HC guard column (Dionex Co.). The eluent was 9.0 mM sodium carbonate solution. Concentration of the H2O2 stock solution (30 wt%) purchased from Wako Chem. Co. was determined by its UV absorption at 240 nm (Bader and Hoigne´, 1988). Hydrogen peroxide concentration of the sample solutions was measured by
METHODOLOGY
Materials All reagents used in this experiment were of analytical grade. Phosphate buffer solution (pH 7.0) was prepared from sodium phosphate dibasic and potassium phosphate monobasic (Wako Chem. Co.). Humic acid was purchased from Aldrich Chem. Co. A sand filtered secondary sewage effluent was filtered through a glass filter (GF/C (1 1.2 mm), Whatman Co.) and the filtrate was used as a sample solution. TOC, IC, and pH of the filtrate were 5.6 mg C l1, 15 mg C l1, and 7.0, respectively. Deionized distilled water prepared by an ultrapure water system (CPW-200, Advantec Toyo Co.) was further purified by another ultrapure water system (Easypure RF, Barnstead Co.) and was used in this experiment. Apparatus Figure 1 shows the schematic of an ozone reactor used in this study. The volumes of a glass reactor and sample solution were 3.0 and 2.3 l, respectively. Ozone gas was
Fig. 1. Schematic of the ozone reactor.
TOC removal with a multistage O3/H2O2 process
3589
Figure 2 shows the profiles of the ratios of residual TOC to the initial TOC (residual TOC ratio) of humic acid aqueous solution during ozonation and the O3/H2O2 process. H2O2 dose at each injection was constant (150 mM). At 10 min, the residual TOC ratios were 0.68 for ozonation and 0.62 for the O3/H2O2 process. The difference of the residual TOC ratios between two processes was not significant, although the residual TOC ratio during the O3/H2O2 process was slightly lower than that during ozonation. On the other hand, a significant difference between two processes was observed at 40 min, i.e., the residual TOC ratio during the O3/H2O2 process was much lower than that during ozonation. Those values were 0.32 and 0.50, respectively. This result showed that the effect of H2O2 addition for TOC removal becomes more significant as oxidation proceeds. The similarity of the TOC profiles at the first stage of the two processes is attributed to the reactivity of humic acid. Humic acid contains reactive functional groups with ozone (e.g., phenolic groups), and sufficient amounts of promoters of radical chain reactions (e.g., H2O2 and superoxide anion (O2)) to yield HO are produced during the oxidation of those functional groups even without the addition of H2O2 (Staehelin and Hoigne´, 1985). The similar dissolved ozone concentration profiles of the two processes (Fig. 3) at the first stage supports that similar reactions occur regardless of the addition of H2O2.
Hence, in the first stage (e.g., during the initial 10 min), the difference of the TOC removal is not significant between the two processes. Also, to explain the better performance of the O3/ H2O2 process at later stages (i.e., at 30 and 40 min), acetic, formic, and oxalic acid concentrations were monitored (Fig. 4(a) and (b)). These compounds were assumed to consist of one major fraction of residual TOC at these stages, and the decompositions of these compounds were considered to be the rate-determining step of TOC removal. This assumption was based on the following considerations: (1) in general, the reactivity of oxidation by-products including low molecular fatty acids with ozone are relatively low except for formic acid (Buxton et al., 1988; Neta and Huie, 1988), (2) the reaction rate constants of acetic and oxalic acids with HOd are significantly low (8.5 107 and 7.7 106 M1 s1, respectively), compared with other organic compounds (Buxton et al., 1988)), (3) these compounds tend to accumulate in the O3/H2O2 process, since a greater number of HOd has to be yielded for the oxidation of a unit number of these carboxylic acids than other more reactive compounds due to their low reactivities in this competitive reaction system. The concentration profiles of the three carboxylic acids during ozonation and the O3/H2O2 process (Fig. 4(a) and (b)) verify the above assumption. The three oxidation by-products consisted of a considerable fraction of TOC after 20 min (more than 0.4 after 30 min during both two processes). Also, the behaviors of these three compounds were different after their formation. That is, the concentrations of these carboxylic acids were lower during O3/H2O2 process than during ozonation. For example, the concentration of acetic acid during ozonation and the O3/H2O2 process were around 7 and 4 mM, respectively. In case of oxalic acid, its concentration during ozonation increased with time and was more than 40 mM at 40 min. But that during the O3/H2O2 process decreased after 20 min and was less than 30 mM at 40 min. With these observations, it is reasonable to attribute greater TOC removal
Fig. 2. Profiles of the residual TOC ratios of humic acid solution during ozonation and the O3/H2O2 process.
Fig. 3. Dissolved ozone concentration profiles during ozonation and the O3/H2O2 process of humic acid solution.
the spectrophotometric method using copper(II) ion and 2,9-dimethyl-1,10-phenanthroline (Aldrich Chem. Co.) (Baga et al., 1988; Kosaka et al., 1998). Ozone concentration in the gas phase was measured by iodometry after trapping ozone gas in a potassium iodide solution. Dissolved ozone concentration was determined by the indigo method described by Bader and Hoigne´ (1981). TOC and IC were measured by a TOC analyzer (TOC-5000A, Shimadzu Co.).
RESULTS AND DISCUSSION
TOC removal of humic acid during ozonation and O3/H2O2 process
3590
Koji Kosaka et al.
efficiency of the O3/H2O2 process to its better oxidation efficiency of these carboxylic acids. Faster decompositions of oxidation by-products by the O3/H2O2 process were also in agreement with the dissolved ozone concentration profiles (Fig. 3). This suggests faster HOd production in the O3/H2O2 process. Dissolved ozone concentration was always higher during ozonation than during the O3/H2O2 process. The differences of dissolved ozone concentration between the two processes at 10 and 40 min were 25 and 80 mM, respectively. The difference of the dissolved ozone concentrations between the two processes became greater with time. This lower dissolved ozone concentration during the O3/H2O2 process indicates that ozone is converted to HOd more rapidly in the O3/H2O2 process than in ozonation, especially after the completion of the direct reactions between ozone and reactive functional groups of humic acid (note that the greater difference was observed at 40 min). In other words, the accumulation of ozone in the solution implies that the system does not produce HOd sufficiently. Also, it is to be noted that some oxidation byproducts that accumulate with time act as inhibitors of the radical chain reactions (Staehelin and Hoigne´, 1985). Thus, HOd formation by the addition of H2O2d becomes more important for TOC removal as oxidation proceeds.
Fig. 4. Profiles of the concentrations of acetic, formic, and oxalic acids during (a) ozonation and (b) the O3/H2O2 process of humic acid solution.
The above discussion led us to speculate that suppressing rapid ozone accumulation is one requirement for effective TOC removal by the O3/H2O2 process. However, low dissolved ozone does not always guarantee efficient TOC removal. The evaluation of the system has to incorporate H2O2 concentration because excess H2O2 acts as a HOd scavenger and consumes ozone ineffectively. Also, at early stage, molecular ozone may be more effective than HOd due to its selectivity for reaction sites of humic compounds. The effect of H2O2 concentration at each stage will be discussed more quantitatively in the next section. Effect of H2O2 addition modes on TOC removal of humic acid during the O3/H2O2 process The comparison between the O3/H2O2 process and ozonation in the previous section implies that optimal H2O2 dose increases with reaction time (number of stage) and the increase is related to (1) the change of H2O2 consumption (i.e., the difference of the H2O2 concentration at the beginning of the stage and residual H2O2 concentration after each ozone introduction) and (2) the formation and decomposition of low molecular fatty acids. To test this hypothesis, three different modes of H2O2 addition were compared: (1) H2O2 dose was decreased stepwise as 250 ! 200 ! 100 ! 50 mM (D mode); H2O2 dose was increased stepwise as 50 ! 100 ! 200 ! 250 mM (I mode); H2O2 dose was constant at 150 mM (C mode). Total H2O2 doses of three different modes were fixed at 600 mM. The order of TOC residual ratios of the three modes after 40 min treatment was I mode (TOC residual ratio was 0.26)5C mode (0.32)5D mode (0.36) (Fig. 5). This observation supports our hypothesis that optimal H2O2 dose increases with reaction time (number of stages) at a fixed total H2O2 dose. Next, the relationships between the increase of optimal H2O2 dose and its two reasons described in the earlier part of this section are discussed in detail below.
Fig. 5. Effects of H2O2 addition modes on the residual TOC ratios of humic acid solution during the O3/H2O2 process.
TOC removal with a multistage O3/H2O2 process
3591
Table 1. Profiles of the amount of H2O2 consumption of three H2O2 addition modes at each stage Number of stage D mode
I mode
[H2O2]0a
c
4
250
364
209
61 60 0.99
86 0.34
255 0.70
198 0.95
[H2O2]0a
50
118
206
255
DH2O2b
32 0.64
112 0.95
201 0.98
248 0.97
222
182
153
190 0.86
179 0.98
152 0.99
[H2O2]0a (mM) DH2O2 [H2O2]1c 0
b
3
(mM) DH2O2 [H2O2]1c 0
DH2O2b a
2
DH2O2b
(mM) DH2O2 [H2O2]1c 0
C mode
1
150 78 0.52
Initial concentration of H2O2. The amount of H2O2 consumption. Consumption ratio of H2O2.
To take a closer look of H2O2 consumption, the profiles of the initial H2O2 concentration of each stage, H2O2 consumption, and consumption ratio are summarized in Table 1. The initial H2O2 concentration is the sum of the residual H2O2 at previous stage and H2O2 dose at next stage. It is to be noted that H2O2 consumption ratio was higher at later stages regardless of the H2O2 addition modes. Also, despite the same total H2O2 dose (600 mM), the H2O2 consumptions were different among three modes from 30 to 40 min (at the last stage). H2O2 consumptions of I, C, and D modes during the fourth stage were 248, 152, and 60 mM, respectively. On the other hand, however, H2O2 consumptions of D and C modes were significantly higher than that of I mode at some early treatment stages, although the residual TOC ratios and the dissolved ozone concentrations were not quite different (Table 1, Figs 5, and 6) among the three modes. Therefore, it is assumed that H2O2 doses for D and C modes at some early treatment stages were in excess and H2O2 was ineffectively consumed while H2O2 was in short supply during the fourth stages. It is also to be noted that with an extremely high H2O2 concentration, TOC removal efficiency may be decreased since ineffective consumption of H2O2 includes ineffective consumptions of other oxidants (e.g., HOd). But, such an extreme effect by the excess H2O2 was not observed clearly in this experiment. Dissolved ozone concentrations at 40 min supported the order of TOC residual ratios of the three modes after 40 min treatment by suggesting that the HOd formation rate was the fastest in the I mode at the last stage (30–40 min). The order of dissolved ozone concentration inversely corresponded to the order of TOC removals and the order of H2O2 consumptions (Fig. 6). Since low dissolved ozone concentration indicates rapid formation of HOd, it is considered that from 30 to 40 min the amount of HOd formation of I mode was highest and hence the TOC residual ratio of I mode was lowest at 40 min. From
Fig. 6. Effects of H2O2 addition modes on dissolved ozone concentration during O3/H2O2 process of humic acid solution.
the above discussion on the profiles of the TOC, H2O2, and dissolved ozone, it was verified that under the constant ozone dose it was better to increase H2O2 dose stepwise for effective TOC removal because H2O2 demand also increased with time. Let us turn to the formation and decomposition of low molecular fatty acids. The increase of H2O2 demand with time was presumed to be related to the accumulation of oxidation by-products because they are less reactive with HOd than humic acid itself and greater amount of HOd is required for the decompositions of those less reactive compounds. Based on this presumption, the relationship between H2O2 demand and the accumulation of oxidation byproducts was investigated. Among the oxidation by-products, oxalic acid was chosen as an indicator of the extent of oxidation and the accumulation of the oxidation by-products was assumed to be expressed by the ratio of oxalic acid concentration (in mg C l1) to TOC. This is because oxalic acid accumulates as oxidation proceeds due to its significantly low reactivity with HOd among the major oxidation by-products during AOPs. This
3592
Koji Kosaka et al.
To test the practical applicability of the findings obtained from the experiments using humic acid (i.e., the better performance of the O3/H2O2 process on TOC removal over ozonation and the increase of H2O2 demand with time), a sand filtered secondary sewerage effluent was treated by ozonation and the O3/H2O2 process. The profiles of the residual TOC ratio given in Fig. 8 clearly showed the effect of the O3/H2O2 process on the treatment of the sand filtered secondary sewerage effluent compared to ozonation. In this experiment, H2O2 dose of each injection was constant (150 mM) and total H2O2 dose was 600 mM. The residual TOC ratio after the O3/H2O2 process was much lower than that after ozonation (i.e., at 40 min those values during the O3/H2O2 process and ozonation were 0.46 and 0.78, respectively). However, the effect of H2O2 addition on TOC removal of the sand filtered secondary sewerage effluent was
different from that for humic acid aqueous solution. For the humic acid solution, the effect of H2O2 addition was observed only at later stages, but for the sewerage effluent the effect of H2O2 addition was clearly observed from the first stage. Considering the similarities of pH and TOC levels of these two solutions, there are two possibilities to explain this difference of the effect of H2O2 addition on TOC removal. The first explanation is that the sand filtered secondary sewerage effluent contained bicarbonate ion (the IC was 15 mg Cl1), an inhibitor of radical chain reactions. As a result of the inhibition by bicarbonate, the effect of H2O2 addition can be observed from earlier stages because more initiator for radical chain reaction is required in this condition. The second possible reason is the difference of the characteristics of organic solutes. As seen from Figs 4 and 9, for the humic acid solution, only oxalic acid accumulated at high concentration; on the other hand, for the sand filtered secondary sewerage effluent, both acetic and oxalic acids were highly accumulated. This significant difference of the profiles of the low molecular fatty acids indicates the significant difference of their components of organic solutes. The difference of organic solutes of the two solutions is assumed to affect the difference of the effect of H2O2 addition on TOC removal. Figure 10 shows the profiles of dissolved ozone concentration during ozonation and the O3/H2O2 process. From Figs 8 and 10, the relationship between TOC removal efficiency and dissolved ozone concentration for humic acid solution was found to be applicable to the sand filtered secondary sewerage effluent. That is, dissolved ozone concentration during ozonation was always higher than during the O3/H2O2 process while its TOC removal efficiency was lower. For example, at 10 min, TOC removal during ozonation was negligible and the dissolved ozone already reached 70 mM. Therefore, it was confirmed that the profiles of dissolved ozone reflected the HO formation efficiency during these
Fig. 7. Relationship between ratio of oxalic acid to TOC at the beginning of each stage and H2O2 consumption during the stage.
Fig. 8. Profiles of the residual TOC ratios of the sand filtered secondary sewage effluent during ozonation and the O3/H2O2 process.
expression seems to be valid since fraction of oxalic acid increased with time until 40 min. Fig. 7 shows the relationship between the ratio of oxalic acid to TOC at the beginning of each stage and H2O2 consumption (i.e., H2O2 dose required) during the stage. The results of Fig. 5 and Table 1 were used after the exclusion of some results (the results of D and C modes at the fourth stage were excluded because of lower TOC removal than I mode, and that of D mode at the second stage were excluded because the significant amount of H2O2 is assumed to be consumed ineffectively). The linear relationship was observed between H2O2 demand and the ratio of oxalic acid concentration to TOC. Therefore, it was shown that the change of H2O2 demand of each stage is correlated to the progression of the oxidation and the progression of the oxidation means the accumulation of oxidation by-products with low reactivity with HOd like oxalic acid. TOC removal of a sand filtered secondary sewage effluent
TOC removal with a multistage O3/H2O2 process
3593
of oxalic acid to TOC is assumed to represent the accumulations of low reactive compounds with HOd (the progression of oxidation). Thus, it was shown that H2O2 demand of the sand filtered secondary sewage effluent increased with time and the increase of H2O2 demand was due to the progression of the oxidation like the case of humic acid solution. CONCLUSIONS
Fig. 9. Profiles of the concentrations of acetic and oxalic acids during (a) ozonation and (b) the O3/H2O2 process of the sand filtered secondary sewage effluent.
This research investigated the effect of the change of the organic solutes on the TOC removal efficiency or operating condition during a multistage injection O3/H2O2 system. During the process H2O2 demand at each stage increased under constant ozone dose. This was because the low molecular fatty acids, low reactivity compounds with HOd, became main components of organic solutes as the progression of the oxidation. In this experimental condition, the degree of the progression of the oxidation could be represented by the ratio of oxalic acid to TOC. This tendency was observed for both humic acid solution and a sand filtered secondary sewage effluent. Dissolved ozone concentration reflected the amount of ozone consumed for the reaction during the process. Therefore, an appropriate operational condition can be obtained by monitoring dissolved ozone and finding the condition that dissolved ozone was maintained at low concentration. In this experiment, the effect of the change of the organic solutes during the process was investigated from the point of H2O2 demand. As a future work, we would like to examine relationship between the change of the organic solutes during the process and other parameters (e.g., ozone demand for unit TOC reduction (DO3/DTOC)). REFERENCES
Fig. 10. Dissolved ozone concentration profiles during ozonation and the O3/H2O2 process of the sand filtered secondary sewage effluent.
processes. It is to be noted that during the O3/H2O2 process dissolved ozone was greatly increased after 20 min and became 85 mM at 40 min. Considering the accumulation of dissolved ozone and the constant H2O2 dose of each injection, it is assumed that H2O2 dose required during the O3/H2O2 process greatly increased after 20 min. The profiles of oxalic acid to TOC also greatly increased after 20 min (the ratios at 0, 10, 20, and 30 min were 0, 0.08, 0.21, and 0.24, respectively). As described in earlier sections, increase
Bader H. and Hoigne´ J. (1981) Determination of ozone in water by the indigo method. Water Res. 15, 449–456. Bader H. and Hoigne´ J. (1988) Photometric method for the determination of low concentrations of hydrogen peroxide by the peroxidase catalyzed oxidation of N,Ndimetyl-p-phenylenediamine (DPD). Water Res. 22, 1109–1115. Baga A. N., Johnson G. R. A., Nazhat N. B. and SaadallaNazhat R. A. (1988) A simple spectrophotometric determination of hydrogen peroxide at low concentrations in aqueous solutions. Anal. Chim. Acta 204, 349–353. Buxton G. V., Greenstock C. L., Hwlman W. P. and Ross A. B. (1988) Critical review of rate constants for reactions of hydrated electrons, hydrogen atoms and hydroxyl radicals (OH/O) in aqueous solution. J. Phys. Chem. Ref. Data 17(2), 513–886. Echigo S., Kosaka K., Yamada H., Matsui M. and Shishida K. (1998) The effect of water quality characteristics on the optimal hydrogen peroxide dose for the ozone/hydrogen peroxide process. J. Jpn Soc. Water Environ. 21, 444–449 (in Japanese). Furukawa S., Yasunaga N., Kawaai Y and Hirotsuji J. (1997) Highly efficient ozone-hydrogen peroxide reactor with multiple injection ports. Proceedings of the 13th Ozone World Congress, Vol. 1, pp. 499–503.
3594
Koji Kosaka et al.
Glaze W. H., Kang J. and Chapin D. H. (1987) The chemistry of water treatment process involving ozone, hydrogen peroxide and ultraviolet radiation. Ozone Sci. Technol. 9, 335–352. Gulyas H., von Bismarck R. and Hemmerling L. (1995) Treatment of industrial wastewaters with ozone/hydrogen peroxide. Water Sci. Technol. 32(7), 127–134. Kosaka K., Yamada H., Matsui S., Echigo S. and Shishida K. (1998) Comparison among the methods for hydrogen peroxide measurements to evaluate advanced oxidation processes: application of a spectrophotometric method using copper(II) ion and 2,9dimethyl-1,10-phenanthroline. Environ. Sci. Technol. 32, 3821–3824. Masten S. J. and Davies S. H. R. (1994) The use of ozonation to degrade organic contaminants in wastewaters. Environ. Sci. Technol. 28, 181A–185A. Neta P. and Huie R. E. (1988) Rate constants for reactions of inorganic radicals in aqueous solution. J. Phys. Chem. Ref. Data. 17(3), 1027–1284.
Shishida K., Echigo S., Yamada H. and Matsui S. (1999a) Study of the development of the high performance ozone/ hydrogen peroxide water treatment system for TOC removal: effect of the operational parameters. J. Jpn Soc. Water Environ. 22, 199–205 (in Japanese). Shishida K., Haruki H., Kamio K., Yamada H., Kosaka K. and Matsui S. (1999b) Comparison of the ozone/ hydrogen peroxide process, the ozone /ultraviolet process, and the hydrogen peroxide/ultraviolet process in the treatment of landfill leachate. J. Jpn Soc. Waste Manage. Experts 10(5), 247–256 (in Japanese). Staehelin J. and Hoigne´ J. (1985) Decomposition of ozone in water in the presence of organic solutes acting as promoters and inhibitors of radical chain reactions. Environ. Sci. Technol. 19, 1206–1213. Yao C. C., Haag W. R. and Mill T. (1992) Kinetic features of advanced oxidation processes for treating aqueous chemical mixtures. Chemical Oxidation: Technology for the Nineties, Proceedings of the Second International Symposium, Vol. 2, pp. 112–139.