Environ. Scl. Technol.

1982, 16, 676-681

Printing Office No. 040-000-00337-1. U.S.Environmental Protection Agency-Region 11, Draft Environmental Impact Statement for the Hudson River PCB Demonstration Reclamation Project, May 1981.
Fed. Regist. 1976, 41, 7552. Choi, P. S. K.; Nack, H.; Flinn, J. E. Bull. Environ. Contam. Toxicol. 1974, 1 1 , 12-7. Law, L. M.; Goerlitz, D. F. Pestic. Monitor. J. 1974,8,33-6. Carey, k E.; Gower,J. A.; Tai, H.; Mitchell, W. G.; Wiersma, G. B. Pestic. Monitor. J. 1979, 12, 209-29.

G. B. Pestic. Monitor. J. 1979, 13, 17-22.

Received for review November 11,1981. Accepted May 19,1982. LDGO Contribution No. 3356. Support for this research was supplied by National Oceanic and Atmospheric Administration (Contract NA79-RACOOl26) and the National Science Foundation (Contract OCE-7909249). Additional support was provided by the New York State Department of Environmental Conservation (Contract NYS-C-125638) to implement Section 3 of an agreement between the DEC and the General Electric Co. related to the discharge of PCBs into the Hudson River.

Carey, A. E.; Douglas, P.; Tai, H.; Mitchell, W. G.; Wiersma,

Decomposition of Ozone in Water: Rate of Initiation by Hydroxide Ions and Hydrogen Peroxide
Johannes Staehelln and Jurg HolgnQ
Federal Institute for Water Resources and Water Pollution Control (EAWAG), Swiss Federal Institute of Technology, CH-8600 Dubendorf, Switzerland

The initiation of ozone decomposition in pure water is first order in O3 and OH- concentration and k03,0Hbecomes 70 f 7 M-l s-l when a sequence of reactions is assumed by which three molecules of O3 are eliminated per primary event. The difference between this value and higher values reported in the earlier literature may be explained by interferences by radical chain reactions which have not been totally inhibited in those studies. H202also reacts with O3 when present as an anion, H02-. k03,HOzis (2.8 f 0.5) X lo6 M-l s-l when it is assumed that two molecules of O3 are eliminated per primary event. Therefore, whenever [H,02] > lo- M (pH <12), H02- has a greater effect than OH- on the decomposition rate of O3 in water. The high reactivity of HO, explains that H202 is observed as a significant intermediate in the ozonated water only if the pH is low, e.g., <6.
W

Introduction

Ozone is used in many drinking water plants for the oxidation of organic micropollutants and manganese and for disinfection. Since mixing processes and many of the desired direct chemical reactions of O3 with dissolved solutes M are often rather slow ( l ) ozonation , processes are typically designed for process times lasting in the order of 10 min. A significant fraction of the O3 dosed to the water can be lost during this time. Often this loss is kinetically controlled by the decomposition of O3rather than by ozonation of dissolved substances (see Figure 1). A characterization of the parameters that regulate this decomposition has been of great interest ever since O3 has been used for water treatment. Most of the studies on the decomposition of O3 in water are largely based on phenomenological descriptions of overall kinetics found for pure water. However, these systems are kinetically complex because a primary decomposition of O3 produces free radicals (2-4) which may either become scavenged by bicarbonate and organic solutes or which may react with further O3to yield more free radicals ( 3 , 5 ) ,thus accelerating the decomposition of O3 (see Figure 1). The kinetics of such chain reactions depend on many parameters that have not been separated and identified so far: a comparison of the rate data reported in the literature is therefore rather difficult, and the results from different authors seem contradictory. The proposed rate equations even differ in the reaction order to be considered for the concentrations of O3 or OH-. Discus676

sions including lists for comparisons and literature referRoth ences have been presented by Stumm (6), Peleg (3, et al. (8), and Gurol (9). The purpose of our studies was therefore to separate, identify, and quantify the successive individual reactions that control the overall rate of 0, decomposition in drinking water and other aqueous solutions. The results will allow better predictions of how the rates will be influenced by changes in water composition (10). In this first paper of a series we will describe the kinetics by which the decomposition of O3 in water is initiated by water itself (H20, OH-, H30+) under conditions where secondary radical-type chain reactions are excluded by OH. radical scavengers. Such situations resemble those encountered when natural waters are ozonated, in which HCO; and organic solutes inhibit the radical-type chain reactions efficiently (5). In addition, HzOz is often an important intermediate product from the ozonolysis of aqueous solutions (see Figure 1). This H202formation results either directly from O3decomposition (see reaction 4) or from hydrolysis of organic ozonation products (3, 11-14). During this study it was observed that this intermediate can also influence the kinetics of O3 decomposition. This product (H202/H02-)behaves like water (H20/OH-) in that it only reacts with O3when ionized, and its reaction kinetics could be investigated by similar techniques. The results will also provide general information on the fate of H202produced by ozonation. In contrast, the subsequent chain reactions in which OH. and HO,. (.02-) act as chain carriers have been measured by rather different methods using electron pulse irradiation with kinetic spectroscopy. These systems will therefore be described in separate papers. Since the submission of this paper, Forni et al. (15) have published their work on the kinetics of OH--initiated ozone decomposition and the subsequent OH. radical formation. Their work, however, was not directly aimed at the study of drinking water processes and therefore somewhat different concepts and much higher pH regions were applied for the measurements.
Experimental Part

Chemicals. Water, deionized by ion exchangers, was distilled and then preozonized by adding 3 mg/L Os,which was allowed to decompose before the water was used. Aqueous ozone stock solution was prepared as described
0 1982 American Chemical Society

Environ. Sci. Technol., Vol. 16, No. 10, 1982

0013-936X/82/0916-0676$01.25/0

PH

primary radicals,

*02,OH*

secondary reactive radicals, R*

to2

0,
Mloxid

40

a &

100

Flgure 1. Scheme of reactions of aqueous ozone. M, solutes that consume ozone and become oxidized to MoxU (and H,O,); I, solutes that initiate the decomposition of ozone to primary radicals; S, free radical scavenger; M', solutes that react with OH. and form secondary radicals R e and become finally oxidized to Mroxu.

Flgure 2. Successive half-lives of ozone vs. concentration of carbonate for different pH values. [O3I0 = 50 pM: 1st fl,, A; 2nd t,,, = 3 p~ X; [o3I0= 0.3 PM [Polltot = 50 mM. o. [o3lO

+;

previously (3). Phosphate buffer solutions were prepared from phosphoric acid and sodium hydroxide (both suprapure grade). The solutions were preozonized with 3 mg/L 03.Sodium bicarbonate was analytical grade. Its solutions were preozonized. Nitrous acid and sodium hydroxide standard solutions were Merck Titrisol type reagents (0.1 N). HzOZsolutions were prepared from NazOz (16) or from commercial perhydrol solutions that were free from stabilizers. Both reagents gave the same results. The H202concentrations were determined by the method of Eisenberg (1 7). For methylmercury preparation see ref 18. Instrumentation, pH measurements were performed with a glass electrode. The calibration was based on Merck Buffer Titrisols for pH 7 and 10. The instrumental pH scale was calibrated for the determination of OH- concentrations by solutions of defined OH- concentration (standard sodium hydroxides and nitrous acid solutions) at pH 10-11 (Gran titration method) (19). The ionic strength for these calibrations was adjusted by NaN03 to correspond to that used for the experiments on the kinetics of the OH- reactions ( I = 0.15 M). UV absorption measurements were performed on a Beckman DK-2A or Uvikon 810 instrument in thermostated and closed cells. Methods. Procedure for Kinetic Measurements. All experiments were performed at 20 f 1 "C. Thermostating was required for measurements lasting more than a short time. Samples from the concentrated aqueous stock solution of O3were pipetted to the prepared mixtures of reactants. The flasks were closed and immediately stirred or shaken for 5 s. For kinetic measurements performed to test the effect of carbonate at pH 9 and for all series with H202,the depletion of O3vs. time was directly followed in a 5-cm UV cell a t 258 nm. In all other cases the ozonation was performed in a 1-L round bottom flask from which series of 10-mL aliquots were withdrawn at specified time intervals. The residual concentration of O3 in these samples was determined immediately by direct UV measurements in 1-cm cells (for series at pH 8) or by the indigo method (for all other series) (2~421). Procedure for Measurements of OH- Radical Yields. The yields of OH. radicals were determined from the stoichiometric factor by which methylmercury was mineralized (18). This mineralization is not disturbed by other ozonation reactions. For experimental details see ref 22.

0.2

i
0

2,
\
A

20

40

time, s

60

80

Flgure 3. Relative concentration of ozone vs. time for different [O& [CO,*-] = 10 mM; at high scavenger concentration: [HC03-] [PO,], = 50 mM; pH 10.0.

Results and Discussion Decomposition Initiated by OH- Ions. Dependence on pH and Scavenger and Ozone Concentration. The concentration of the O3 added could easily be followed as a function of time in the batch-type reactors when the pH was in the range 8-10. At pH values below 8, measurements were still possible, but the rates of elimination of O3 became very small and could be disturbed by other reactions such as wall effects. At pH values above 10 the decomposition of O3 became so fast that it could not be followed accurately by the experimental techniques applied here. From the plots of O3 concentration vs. time, half-lives for the concentration of O3 have been determined for succeeding time intervals. As shown in Figure 2, these half-lives depended mainly on pH. At constant pH they increased when carbonate was added to scavenge OH. radicals (5). At low carbonate concentrations they also varied with initial O3 concentration and time. However, when sufficient carbonate was added, the half-lives approached a plateau value for each pH, and they became independent of the O3 concentration. Similar results were obtained when methylmercury was used as a scavenger for OH. radicals (18). Semi-log plots of O3 concentrations vs. time yield straight lines for those experiments performed a t high
Environ. Scl. Technol., Vol. 16, No. 10, 1982 677

10-2

10-4J

16-6

16-5

16-4

[OH-]

, M
1 0

9 PH

Figure 4. Measured pseudo-first-order rate constant for the decomposition of ozone vs. hydroxide lon concentration. A pH scale is given for comparison. [HCO,-1 [C0,2-] = 10 mM; [PO,], = 50 mM.

those reported in the literature for pure water. Even the orders of reaction are different. This result is not surprising since in most earlier studies the rate of the decomposition in pure water was accelerated by undefined chain reactions subsequent to OH- attack. The rate constant kOs,-,H-measd is small when compared with the rate constants quoted for other reactions of 03. This measured value is still different from the rate constant of the primary OH- attack on O3 molecules because it must be assumed that more than one O3 molecule is destroyed per primary initiation step. To get an estimation of the overall stoichiometry of the O3 OH- reaction (when all OH. radicals are scavenged before reacting with 0,) the following reaction sequences can be considered: Hypothesis I ((2)extended for dissociation of HOz-and electron-transfer mechanism)

(4)

scavenger concentration as demonstrated in Figure 3 for a series of measurements at pH 10. The slopes of these lines do not vary with O3concentration in the range 0.3-50 KM. Therefore, the results presented in Figures 2 and 3 for high concentrations of radical scavengers can be described by a pseudo-first-order rate law: -(d[O,]/dt)p~ = kbH-[O$llO
(1)

where HOz. + H+ + - 0 2 -

O3+

eoz-

ks

pK, = 4.8 so3-+ O2

0 2

(5)

(6) (7)

.03+ H20

2OH. + OH- +
___*

where kbH-= In 2 / t l l z . As shown in Figure 4 the pseukOgOKprim do-first-order rate constants It be increase by a factor of O3 OHHO2- + 0 2 10 per increased pH unit. For later convenience the pH where scale in Figure 4 has also been converted to a scale of increasing OH- concentration by the calibration procedure H202 HOz- H+ pK, = 11.6 described in the ExperimentalP a r t . The slope of log kbHkOaHOZ-@ w. log [OH-] becomes 0.99 f 0.05 (error range for p = 0.95) O3 HOzOH. -0,+ O2 when the regression is based on seven series of measurements performed at five different pH values. Thus, the followed by reactions 6 and 7. measured rate of decomposition of O3 can be described by Hypothesis I11 (23) - d[O3] / dt = ko,,o~-~~~[031~~[OH-]~~ (2)

Hypothesis I1 (this study and ref 11)

(4)
(8)

so2-

(4)

for which the second-order rate-constant becomes koa,OH-measd = koH-/[OH-]l.O = (210 f 20) M-l s-l

(3)

where [OH-] is the analytical OH- concentration. When based on the experimental pH scale, the values become somewhat different because [OH-] is not exactly X 10PHM (19). Transformed to the pH scale, the numerical values become kog,pHmeasd= kbH-101~~ = (175 f 10)x 1014-PHM- s-1 (3) Ranges of errors are based on the precision of the measurements and on p = 0.95; the accuracy of the data is, however, limited by the accuracy of pH or [OH-] measurements, for which we assume a range of error of 0.05 units or 10%. Discussion. On the basis of these results we assume that the pH-dependent decomposition of Os is due to a reaction that is kinetically controlled by the OH- concentration. The phenomenon cannot be due to a pH-dependent change of O3 molecules (such as formation of H03+or O,OH-) because all other known O3reactions are strictly pH independent except in those cases where the degree of dissociation of the other reactant changes with pH (21). Also the fact that OH--initiated decomposition of O3is of first order, both in Osand in OH- concentration, is consistent with our experience with reactions of O3with other solutes (21). The kinetic rate laws now found for aqueous solutions that contain OH. radical scavengers are different from
678 Environ. Scl. Technol., Vol. 16, No. 10, 1982

followed by reaction 7. The occurrence of as an intermediate (hypotheses I and 11) is supported by qualitative chemical detections based on the reduction of tetranitromethane (22). The transient existence of 03has now been positively proved by kinetic spectroscopy (15) as well as by pulse radiolysis studies on ozonated water (kinetic spectroscopy) (24). Hypothesis I11 was proposed for observations made at extreme conditions of very high pH values (pH -14). For our system the stoichiometric yield of OH. radical formation was found to be 0.55 f 0.08 (3). This value is significantly less than 2.0 predicted by hypothesis 111, but it is close to 0.67 as predicted by hypotheses I and 11. We therefore can reject hypothesis I11 for our case. Because the observed overall reaction has been found to be first order in O3 concentration, the sequences of reactions formulated by hypotheses I and I1 must both be assumed to be rate limited only by the initiation step (reaction 4 or 4). This is expected the subsequent reaction of -0, with O3 is fast and dominant whenever the concentration of O3 is dominant compared with that of other solutes reactive toward so2-.It yields 03(1624) (the rate constant for the reaction of -0,with 03, k6,is about lo8 M-I s-l (15,24,25)).This is relatively high when compared with other rate constants of -0,(26)).Likewise, SO,and H202can only exist as a short-lived intermediates in the pH ranges here considered (see ref 24 and later paragraphs). All sequences of reactions proposed by hypotheses 1-111 can be followed by OH. radicals interacting with further

O3as formulated by reaction 10. However, scavengers such as carbonate, phosphate, or other solutes compete with reaction 10.

OH.

OHkI3 = 5 X

+HCO;

OH-

+ HC03.
+

(12)

+BP0,2-

OH- .HP04lo6 M-l s-l (29)

(13)
\--,

In contrast to OH., H02. (-02-) is not scavenged and leads to further decomposition of O3 (see eq 6). On the basis of previous work (5) we assume that the products formed from bicarbonate and carbonate ions do not interact significantly with further 03, and we assume that reaction 6 is not inhibited by carbonate (reactions between carbonate ions and H02-(-0L)have also not been observed in pulse radiolysis studies). Therefore, for each pH and each O3 concentration an appropriate carbonate concentration could be achieved that was sufficient to inhibit the subsequent chain reactions effectively. In both hypothesis I and I1 it is assumed that three molecules of O3 are decomposed per initiating step provided that eo2-, H02., so3-, and H202are not scavenged by other reactions. Herewith the rate constant of the primary reaction of OH- with O3 becomes kOa,OHJrim = 1/33k0a,0H-measd = 70 f 7 M-l s ' (14)

Flgure 5. Measured pseudo-flrst-order rate constants of decomposition of ozone vs. hydrogen peroxlde concentration for different pH values: [methylmercury] = 1.5 mM; [PO4ltOt = 50 mM.

This value is in good agreement with the value recently [ H O I 1 ,lo-' M deduced by Forni et al. (15) for reactions at pH >lo. Figure 6. Measured pseudo-flrst-order rate constant of decomposition Other Possible Primary Reactions in Pure Water. of ozone vs. hydrogen peroxlde ion concentration: [methylmercury] (i). A priori we cannot exclude that the decomposition of = 1.5 mM; [P04],ot = 50 mM. aqueous O3could also be initiated by other species present in pure water such as by H30+. However, the decompocould again be deduced from the slope of these lines. At sition of O3 is not accelerated even at very low pH values. pH 2, H202reacts only very slowly with O3 (4). However, At pH 2 tlI2 is larger than 200000 s. Therefore k ~ ~ , H , o + at ~~ pH values above 5, a strong acceleration of the decommust be smaller than 4 X M -l s-l. position of O3by H202was observed. The effect increased (ii). Aqueous O3 could also decompose in the aqueous linearly with the concentration of H202 (see Figure 5). matrix by a spontaneous reaction that is not catalyzed by Therefore the reaction rate can be formulated by ions. The rate law of this reaction could be formulated -(d[031/dt)p1-1 = ~ 0 ~ , ~ , 0 , ~ ~ ~ (16) ~[031~[ by This rate increases by 1 order of magnitude per pH unit. -d[Od/dt = k [ o d (15) This observation suggests that H202also reacts with O3 On the basis the stability of O3 at low pH values, we can only when present in its ionized form: estimate that the first-order reaction rate constant for such V e r y Slow 0 3 + H202 see r e f 4 a reaction, tz,, is smaller than 5 X lo4 s-l. Thus, initiation of O3 decomposition reactions that are not controlled by OH-, such as proposed by Alder and Hill (30),cannot be of importance. Decomposition Initiated by H202 (H02- Ions). 03 + Hoedecomposition of O 3 Measurements on the rate of the decomposition of O3 in The rate of decomposition of O3 is plotted in Figure 6 vs. the presence of radical scavengers were performed on a the concentration of H02-. The concentration of H02-was series of samples that contained different concentrations calculated from the total concentration of H202and the of H202. For these measurements methylmercury hydegree of dissociation, a,calculated for pKH&, = 11.6 and droxide was used as a free radical scavenger instead of the pH applied, where carbonate (18) (carbonate would be inefficient in the low pH region applied, where it is protonated to form H2C0,). a = [HO,-I/([H20,1 + WO2-I) = [HO,-l/[H2021,, H202 was added in an excess when compared with that of (17) the initial concentration of 03.All plots of the log of the or for pH << pKHaOl concentration of residual O3 vs. time were linear. Thus, 1% a = PH - PKH202 (18) pseudo-first-order rate constants for the elimination of O3

Environ. Sci. Technol., Vol. 16, No. 10, 1982 679

Table I. Rate Constants for the H,O-0, and the Aqueous H,O,-0, Systems

H,O

k, = < 5 X

s-'

kO,.H,O+=

<4 x 10-4 M - 1
H,O,
a

s-l

ko , 210

, o ~= - ~ ~ ~
20 M-] s'l

k'OH- = 175 x 1OPH-14 s-1 a

k'H0,- = 5 . 5 X 106[H,0,]~ot10PH-PK s-'

~O,.H,O, =

<10-2 M-I

k03,H0,-m e a d =
s-l C

( 5 . 5 t 1.0)X 106M-I s-]

For ionic strength I = 0.15 M.

For ionic strength I = 0.05 M.

Reference 4.

Rate constants measured at different pH values in the presence of different relative amounts of nondissociated H202 fit in the same line. This means that the rate can be formulated as a reaction that is first order in the concentration of H02-, and the reaction rate constant can be defined by -d[O,]/dt = kko2-[03] = ko,,~o,-"""~[O3][H02-] (19) From the data in Figure 6 ko,,Ho,-measdbecomes (5.5 f 1.0) X lo6 M-l s-l (r ange of error based on precision of measurements of series, and p = 0.95; this range is larger than the uncertainty of pH and pKH,o values used for calculations). This rate constant is higi when compared with that of OH- or with other reactive anions (1,21). Therefore, even very small concentrations of H02- become kinetically effective for initiating the decomposition of O3 (see below). The stoichiometric yield factor for the production of OH. radicals in this decomposition was measured by the mineralization of methylmercury. The results showed that the formation of inorganic mercury was not significantly different (within 130%) when the decomposition of O3 was initiated by HOP-instead of OH- ions. This means that the stoichiometric yield of OH. radical formation does not significantly vary with the type of initiation (reaction 4, 4', or 9). We therefore can still accept the sequence of reactions proposed by Taube ( 4 ) where reaction 9 is followed by reactions 6 and 7. Also, in this system the intermediate species -02-will predominantly react with O3(vide supra), and subsequent chain reactions are again expected to be inhibited by OH. radical scavengers. Therefore, we assume that the rate of decomposition of O3is twice that of the primary initiation (reaction 9): = (2.8 f 0.5) X lo6 M-l s-l k0s,H02~rim = (20) Rate of Decomposition of Hydrogen Peroxide in Ozonation Processes. The interaction of H02- with O3 will lead to a consumption of H02-, which is rapidly supplied from the dissolved H202. On the basis of the sequence of reactions 9,6, and 7, we can therefore formulate the rate of consumption of H202by O3by

\
4
6 PH

Flgure 7. Calculated half-llves of hydrogen peroxide in the presence of Constant ozone Concentrations at different pH values (all OH- radicals scavenged, see eq 22).

OH- increase proportionally with their concentrations. The effect of H02- on the rate of decomposition of O3 becomes larger than that of OH- if
ko,,o~-~~~~< ~ [k0a,~0~-meaadLy[H20~]tot OH-] (23) Below pH 11, [H02-]and [OH-] vary the same way with pH (see eq 18). We can therefore compute a critical H202 concentration, for which the kinetic effect of H02- (or of [H202],) at all pH values below about 11becomes larger than that of OH-: [H202lc1it =

/(

~ o ~ , H o > ~ lo-' ~ ~ M ~ ~ L ~ )

(24)
This critical H202concentration is low and independent of the pH at pH values significantly below pKHzoz = 11.6.

Conclusion In pure water, the decomposition of O3is initiated only by its reaction with OH- (see Table I). The apparent rate = 210 f 20 M-l s-l constant for this reaction is kos,OH'measd and the apparent pseudo-first-order rate constant for the decomposition of O3by OH- becomes kbH- = 175 X 10PH-14 s-l. This value is low when compared with the rate of decomposition of O3 reported for "pure water". The discrepancy is due to secondary chain reactions of nondefined chain lengths which probably controlled the kinetics re-d[HzOzItot/dt = ~ 0 , , ~ 0 , ~ " ~ ~ [ H z 0 2 1 t (21) ~ t [ 0 ~ 1 ported in earlier studies for solutions containing insufficient concentrations of OH. radical scavengers. and if the concentration of O3is kept constant, the lifetime H202reacts with O3 only when present as the anion, of H202in a semi-batch-type reactor becomes H02- (see Table I). The rate constant k03,H0zmeasd = (5.5 f 1.0) X lo6 M-l s-l is high when compared to that for OHt1/2 = (In ~ ~ / ~ ~ 0 3 , H O z ~ P r i " (22) ~ ~ ~ ~ ~ ~ (see Table I). In aqueous solutions where the concentration of H202becomes larger than lo-' M, the decompoFigure 7 shows that the half-life of H202becomes short sition of O3is initiated by H02- faster than by OH- (this for high pH values even in presence of radical scavengers. relation holds up to pH <11.6). This is the reason why H202,even when formed as a kiAt high pH values H202will not accumulate even when netically important intermediate, will only accumulate it is produced as an intermediate ozonolytic product beduring the ozonation process if low pH values (pH <6) are cause already in the neutral pH region, where a small maintained. This is in good agreement with reported obdegree of dissociation of H202becomes apparent, the deservations on H202formation in ozonized solutions (3,11, composition of H202by reaction of O3with HOC proceeds 12). fast (see Figure 7). The decomposition of H202increases Efficiency of HO, Relative to That of OH- in Iniby a factor of 10 per pH unit increase. tiating Decomposition of Ozone. The roles of HOT and
680

Environ. Scl. Technol., Vol. 18, No. 10, 1982

105

\
\

6
PH

IO

1 1 OH. Figure 8. Half-ilves (f,,*) of ozone in different waters: I, a radicals scavenged (this work); 11, distilled water, [PO,], = 50 mM; 111, water of the lake of ZiSrich (30); IV, groundwater, infiltrated from Rhlne (Basle, Waterwork Hard) (37). Reactions of OH-and H02- with O3can initiate radical chain reactions. The kinetic chain length of such reactions depends on the relative rate by which the radicals formed react with O3compared with with other solutes present in the solution. Therefore such chain reactions can be more important in pure water than in real drinking waters or in some types of waste waters where bicarbonate and organic impurities may significantly scavenge OH. radicals. In Figure 8, the half-lives of O3 measured in different types of natural waters are compared with those measured in our model solutions. This comparison also demonstrates that our conclusions are applicable to measurements on real waters.

(5) Hoigng, J.; Bader, H. Vom Wasser 1977,48, 283-304. (6) Stumm, W. Helv. Chim. Acta 1954,37, 773-8. (7) Peleg, M. Water Res. 1976,10,361-5. (8) Roth, J. S.; Sullivan, D. E.; DAmbrosi, M. Ozonews Technical Paper Section 1979,7. (99 Gurol, M. D. Dissertation University of North Carolina, Chapel Hill, NC, 1980. (10) Staehelin,J.; Hoigng, J. 5th Ozon-Weltkongress,April 1981, Wasser, Berlin, Colloquium Verlag: West Berlin, 1981;pp 623-636. (11) Joy, P.;Gilbert, E.; Eberle, S. H. Water Res. 1980,14, 1509-16. (12) Gilbert, E. Water Res. 1980,14,1637-43. (13) Kruithof, J. C. In Oxidationsverfahren in der Trinkwasseraufbereitung;Kiihn, W., Sontheimer, H., Eds.; DVGW Karlsruhe, 1978;388-416. (14) Yamamoto, Y.; Niki, E.; Shiokawa,H.; Kamiya, Y. J . Org. Chem. 1979,44, 2137-42. (15) Forni, L.;Bahnemann, D.; Hart, E. J . Phys. Chem. 1982, 86,255-259. (16) Riesenfeld, E. H. Anorg. Chem. Praktikum, Rascher-Verlag, Zurich, 15,Auflage 1940,356-7. (17) Eisenberg, G. M.Ind. Eng. Chem. 1943,15, 327-8. (18) Hoigng, J.; Bader, H. In Organometals and Organo-

metalloids, Occurrence and Fate in the Environment; Brinckmm, F. E., Bellama, J. M., Eds., American Chemical Society: Washington, D.C., 1978;ACS Symp. Ser. 82,

Acknowledgments
We thank Werner Stumm for his continuous support and interest in this study and Werner Haag, Ren6 Schwarzenbach,and Herbert Hohl for valuable discussions and reviewing the manuscript. Heinz Bader was of great help in providing many of the experimental techniques applied for these studies.

292-313. (19) Stumm, W.; Morgan, J. J. Aquatic Chemistry, 2nd ed.; Wiley: New York, 1981;pp 226-9. (20) Bader, H.; Hoigng, J. Water Res. 1981,15,449-56. (21) Hoign6, J.; Bader, H. Water Res., in press. (22) Staehelin, J. Thesis, ETH Ztirich, in preparation. (23) Gorbenko-Germanow;Koslova Dokl. Akad. Nauk SSSR 1973,210, 851-4. (24) Btihler, R.; Staehelin, J.; Hoigng, J., unpublished data, ETH-Z and EAWAG, 1982. (25) Sehested, K.; Holeman, J.; Hart, E. J., to be submitted for (26) Farhataziz; Ross, A. B. Selected Specific Rates of Reactions

publication.

of Transients from Water in Aqueous Solution, National Bureau of Standards, Washington D.C., NSRDS-NBS 59,

Literature Cited (1) Hoigng, J. In Handbook of Ozone Technology and its Application; Rice, R. G., Netzer, A., Eds.; Ann Arbor Science: Ann Arbor, MI, 1982. (2) Weiss, J. Trans. Farad. SOC. 1936,31,668-81. (3) Hoign6, J.; Bader, H. Water Res. 1976,10,377-86. 1940,62,3357-73. (4) Taube, H.; Bray, W. C. J. Am. Chem. SOC.

1977. (27) Bahnemann,D.; Hart, E. J. Phys. Chem. 1982,86,252-255. (28) Weeks, J. L.;Rabani, J. J . Phys. Chem. 1966,70,2100-6. (29) Black, E. D.; Hayon, E. J. Phys. Chem. 1970,74,3199-203. 1950,72,1884-6. (30) Alder, M. G.; Hill, G. R. J . Am. Chem. SOC. (31) Hoign6, J.; Bader, H. Ozone Sci. Eng. 1979,1, 357-72.
Received for review December 23,1981.Accepted June 9,1982.

Environ. Sci. Technol., Vol. 16, No. 10, 1982 881