Environ. Sci. Technol. 2001, 35, 690-695
Electron-Rich Phenols for Probing the Photochemical Reactivity of Freshwaters SILVIO CANONICA* AND MATTHIAS FREIBURGHAUS Swiss Federal Institute for Environmental Science and Technology (EAWAG), CH-8600 Du ¨ bendorf, Switzerland
Different Swiss freshwater samples spiked with 3,4dimethoxyphenol (DMOP) or 2,4,6-trimethylphenol (TMP) were irradiated using UV-A and visible light from a mediumpressure mercury lamp. The kinetics of depletion of both phenols at pH 8 revealed that in almost all samples the probe phenols disappeared more rapidly at 0.1 µM than 5 µM initial concentration. Pseudo-first-order rate coefficients were on average 2-3 times greater at the lower initial phenol concentration. A comparable effect was observed using buffered solutions of Suwannee River fulvic acid, which was also used as a model photosensitizer to study the influence of various parameters on such rate coefficients. Sensitizer concentration and photon fluence rate were found to be directly proportional to the rate coefficients for DMOP transformation at both initial concentrations. For both phenols, the rate coefficients increased with increasing pH in the range 4-10, but such an increase was much more pronounced at 0.1 µM than at 5 µM initial phenol concentration. The observed kinetic behavior is compatible with the assumption that electron-rich phenols are transformed by photooxidants of different lifetimes generated by photosensitization from the dissolved organic matter (DOM).
Introduction Several reactive transient species are produced upon irradiation of natural waters by sunlight (1, 2). An important class of these transients may be referred to as photooxidants because they induce oxidative transformations of natural water components and micropollutants. A major obstacle in characterizing aquatic photooxidants is the very small concentration at which they occur, which makes it difficult to detect them directly by spectroscopic methods. Consequently, they are usually detected indirectly by using selective probe molecules. In natural waters, the presence of only a few photooxidants, like the hydroxyl radical (3, 4) or singlet molecular oxygen (1∆g) (5-7), has been well-documented and their photostationary state concentration measured or calculated. On the other hand, a great variety of aquatic photooxidants that are formed upon photosensitization by dissolved organic matter (DOM) is still largely unidentified due to the complex and variable chemical composition of DOM. These photooxidants are nevertheless important as it has been shown that DOM-derived photooxidants other than the hydroxyl radical or singlet oxygen dominate the photosensitized transformation of various substituted phenols in fulvic and humic acid solutions (8, 9). Such photosensitized * Corresponding author phone: +41 1 823 5453; fax: +41 1 823 5210; e-mail:
[email protected]. 690
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 4, 2001
transformations also occurred in lake waters (8-10) and in seawater (11). Thus, substituted phenols have the potential of being used as probe molecules to characterize and quantify the reactivity of such DOM-derived photooxidants occurring in natural waters. The transformation kinetics of electron-donor-substituted phenols photosensitized by DOM presented, under stationary illumination, an increasing pseudo-first-order rate coefficient with decreasing initial concentration of the phenol in the micromolar and submicromolar range (10, 12). We showed that the extent of the concentration dependence of the rate coefficients varied with the type of DOM and with the probe phenol used (10). In particular, methoxy-substituted phenols exhibited a stronger effect in the initial concentration range 2.5-10 µM than methyl-substituted phenols. The observed concentration dependence was interpreted as being due to a variety of photooxidants, part of which had a photostationary state concentration that was depressed by reaction with the probe molecule (the substituted phenol) and possibly with reaction products (10). It was hypothesized that shortlived excited triplet states of the DOM (expected lifetime in air-saturated water e 2 µs) could contribute significantly to the transformation rate coefficient observed at the higher probe molecule concentrations, whereas other much longerlived radical species should be responsible for the increased rate coefficient at the lower concentrations (9, 10). In this study we improved the procedure for assessing DOM-photosensitized transformation rate coefficients at environmental concentrations by using a 0.1 µM initial probe phenol concentration and compared such rate coefficients with those obtained using 5 µM probe phenol. We used 3,4dimethoxyphenol (DMOP) and 2,4,6-trimethylphenol (TMP) as the probe molecules because of their rapid transformation rates and their differing kinetic behavior as explained above and extended our investigations to various Swiss freshwater samples. With the aim of further characterizing such transformations at the higher and lower probe phenol concentration, we studied the effect of photon fluence rate, sensitizer concentration, and pH on the transformation rates in Suwannee River fulvic acid solutions.
Experimental Section Materials and Samples. Samples from lakes and ponds were generally taken near the shore, 10-20 cm below the water surface, and collected in 2.5 L amber glass bottles. Samples were refrigerated and filtered under vacuum within a day by using washed cellulose nitrate membrane filters of 0.45 µM pore size. The filtrate was stored in amber glass bottles in the dark at 4 °C. Samples from Lake Lucerne and Lake Lugano were taken from a boat, and the River Rhine sample was obtained from the special 5-point sampling device which collects a mixed sample across transept of the river at Rheinu ¨ berwachungsstation Weil, Germany. Table 1 lists the water bodies together with their geographic location and selected physical chemical parameters. Humic and fulvic acid stock solutions (see Table 1) were prepared as described elsewhere (10), except that the Suwanne River fulvic acid (SRFA, type “standard”) was obtained from the International Humic Substances Society. 3,4-Dimethoxyphenol (DMOP, Aldrich, 99%), 2,4,6-trimethylphenol (TMP, EGA-Chemie, 99%), N,N-dimethylaniline (DMA, Fluka, 99.5%), and 2-propanol (Fluka, >99.5%) were used as received. For other details on chemical substances employed see ref 10. Photoirradiations. The pH of the natural waters was adjusted to 8.0 ( 0.1 by addition of hydrochloric acid or sodium hydroxide immediately before starting irradiations. 10.1021/es0011360 CCC: $20.00
2001 American Chemical Society Published on Web 01/11/2001
TABLE 1. Characteristics of the Water Samples sample no. water body or sensitizer 1 2 3 4 5 6 7 8 9 10
Lake Lucerne Lake Lugano Lake Zurich River Rhein Greifensee Lu¨ tzelsee Melin doˆ le Craˆ t Etang de la Grue` re Suwannee River fulvic acid (SRFA) Fluka humic acid (FHA)
sampling place
coordinates
elevation (m o.s.l.)
sampling date
pHa
Kastanienbaum Melide Limmatquai Weil Glatt at Fa¨ llanden Swimming baths Outflow Dam
47°01′N/8°21′E 45°57′N/8°57′E 47°22′N/8°32′E 47°35′N/7°35′E 47°23′N/8°39′E 47°15′N/8°45′E 47°17′N/7°07′E 47°14′N/7°03′E
434 271 406 225 435 500 935 998
2 Oct. 1995 3 Oct. 1995 11 Oct. 1995 4 Oct. 1995 13 Oct. 1995 10 Oct. 1995 10 Oct. 1995 10 Oct. 1995
8.55 8.87 8.62 8.28 8.60 8.75 8.27 7.81
a pH value measured in the laboratory at 22-23 °C. at the wavelength 280 nm.
b
conductivity DOC a(366 nm)b a*(280 nm)c (µS cm-1) (mg/L) (m-1) (L mg4 m-1) 173 190 220 295 370 330 410 120
0.9 1.4 1.2 2.0 2.5 3.9 9.7 16.2 2.30
0.21 0.31 0.30 0.76 0.76 2.21 8.06 21.56 2.33
1.5 1.3 2.2 1.7 2.1 3.0 3.7 4.8 3.2
0.85
2.58
7.0
Absorption coefficient at the wavelength 366 nm. c DOC-specific absorption coefficient
FIGURE 1. Depletion kinetics of the probe phenols in the water sample from Lake Zurich and corresponding linear regression fits. To avoid inner filter effects, the highly colored, high DOC samples #7 and #8 were diluted 1:4 and 1:10, respectively, using bidistilled water prior to pH adjustment. The appropriate amount of aqueous stock solution of the probe molecule was then added. Beside DMOP and TMP, DMA and 4-methylphenol (4-MP) were also used as probe molecules in irradiated fulvic acid solutions. Dilution of the natural waters by pH adjustment and probe spiking was negligible. The pH of the humic and fulvic acid solutions was fixed using phosphate buffers (total phosphorus concentration 50 mM). The merry-go-round photoreactor equipped with a 700 W medium-pressure mercury lamp and the conditions and procedures for irradiating the samples have been described in detail elsewhere (10, 13). All irradiations were performed at a temperature of 25.0 ( 0.2 °C. For all experiments but those concerning the dependence of the rate coefficients on photon fluence rate, lamp power was set to 500 W and light was filtered using a Duran glass jacket and a sodium nitrate filter solution to obtain radiation with λ > 320 nm. To investigate the dependence of the rate coefficients on photon fluence rate, the latter was varied by changing the lamp power and by applying a UV-A band-pass filter (9, 13). The photon fluence rate at the 366 nm line of the mercury lamp was determined by chemical actinometry using aqueous 4-nitroanisole and pyridine (14). Pseudo-first-order rate coefficients for the photosensitized transformation were corrected for direct photolysis and intensity variations in the photoreactor, that were mainly due to aging of the filter solution. Analyses. The residual concentration of the probe molecules after different irradiation times were determined by duplicate injection of typically 100 µL water samples in a high-performance liquid chromatography (HPLC) system.
Reverse-phase C18 columns (Merck LiChroCART 125-4, LiChrospher 100 RP-18 for the analysis of DMOP and TMP, and Astec Polymeric C18, Cat.-No. 56103 for the analysis of DMA) were used. For samples with 5 µM initial concentration of the probe molecule, a Hewlett-Packard HPLC 1050 system equipped with a quaternary gradient pump and a variable wavelength detector was used. For 0.1 µM initial concentration samples a Hewlett-Packard HPLC 1090 system equipped with a Hewlett-Packard 1046A fluorescence detector was used. Analyses conditions are described in detail in Table 2. UV spectra were recorded on a Kontron Instruments Uvikon 940 spectrophotometer using cuvettes with Suprasil I quartz windows and up to 10 cm optical path length. Dissolved organic carbon (DOC) analyses of the samples, filtered as previously described, were performed on a total organic carbon analyzer Shimadzu model TOC-5000 in the “nonpurgeable organic carbon (NPOC)” mode. Samples were acidified prior to analysis. For calibration, standard solutions of potassium hydrogen phthalate (Kanto Chemical Co., Tokyo, Japan) in bidistilled water were employed.
Results Photoreactivity of Various Waters. Typical first-order kinetic curves for the depletion of the probe phenols at 0.1 and 5.0 µM initial probe concentration are shown in Figure 1. The pseudo-first-order rate coefficients, keff, obtained by linear regression and normalized to a photon fluence rate of 3 mEinstein m-2 s-1 at 366 nm, are shown in Figure 2 for different water samples and DOM solutions. Relative errors in keff are estimated to be about 10%. For all but one sample (#8) the rate coefficients for 0.1 µM initial probe phenol concentration are greater than those for 5 µM. The firstorder decay at the lower initial probe phenol concentration is on average 2.6 times faster than at the higher concentration (excluding sample #8 and #10). For sample #10, a solution of Fluka humic acid, the effect is even more pronounced: rate coefficients for 0.1 µM initial probe phenol concentration are 18 times (for DMOP) and 8 times (for TMP) greater than at 5 µM. Effect of Sensitizer Concentration and Photon Fluence Rate. Figure 3a represents the variation of keff for the depletion of DMOP as a function of SRFA concentration up to a value of 4.6 mgC/L. The keff values were slightly corrected for the inner-filter effect (maximum correction factor ×1.09) to obtain values at infinitely small optical density. Direct proportionality between keff and SRFA concentration holds at both initial DMOP concentrations, as observed in previous studies using other DOM types and probe phenol concentrations of at least 0.5 µM (8, 10). These results cuncur with the usual assumption that the rate for photosensitized transformation, and thus the photostationary-state concenVOL. 35, NO. 4, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
691
TABLE 2. Parameters Used for HPLC Analysis compound initial concentration eluent composition (%) water methanol acetonitrile acetic acid, 1% vol ammonium acetate, 0.1 M, pH 9 absorption or excitation wavelength emission wavelength flow rate fluorescence signal integration time retention time internal standard retention time of internal standard a
TMPa
DMOP 5 µM
5 µM
50 40
30 60
10
10
285 nm
277 nm
DMA 5 µM
70 30 252 nm
TMPa
DMOP 0.1 µM
0.1 µM
65
40
25 10
50 10
230 nm 325 nm 1.0 mL/min 1.0 mL/min 1.0 mL/min 0.3 mL/min 1s 1.8 min 2.6 min 4.5 min 5.6 min 4-MP 9.6 min
230 nm 325 nm 0.3 mL/min 1s 12.0 min 4-MP 4.4 min
DMA 0.1 µM
70 30 252 nm 340 nm 0.3 mL/min 1s 13.6 min 1,4-dimethoxybenzene 9.2 min
The same methods with a slightly modified eluent composition were also used for analysis of 4-MP.
FIGURE 2. Pseudo-first-order rate coefficients for the depletion of the probe phenols DMOP (A) and TMP (B) at 0.1 and 5 µM initial concentration in the various water samples. tration of photooxidants, is directly proportional to the rate of light absorbed (9, 15, 16). This assumption is also confirmed by the observed direct proportionality between keff and photon fluence rate (Figure 3b) for both initial concentrations of DMOP in solutions of SRFA (2.3 mgC/L, pH 8.1). The rate coefficients measured using the UV-A band-pass filter were corrected by a factor ×1.37 to account for the loss in light absorption rate due the changed irradiation spectrum. Effect of pH. Previous experiments using 20 µM TMP and Contech fulvic acid as the sensitizer showed only a slight increase in keff when increasing the pH from 7.1 to 8.3 (8). We investigated the pH dependence of keff for the depletion of DMOP and TMP for 5 and 0.1 µM initial concentrations in solutions of SRFA over the wider pH range of 4-10. As shown in Figure 4, the pseudo-first-order rate coefficients generally increased with increasing pH, but the increase was much more pronounced for 0.1 µM than for 5 µM initial concentration. A slight increase in keff is expected to result from the increase in absorption coefficient, and consequently in dose of light absorbed, which is observed for the SRFA solution with increasing pH. Compensating for such an optical effect does not change the qualitative picture displayed by Figure 4, i.e., a slight increase in keff at 5 µM probe phenol concentration and a 7-fold increase at 0.1 µM probe phenol concentration in the pH range 4-10. Further Results. The differential keff at 5 and 0.1 µM probe molecule concentration in a SRFA solution was also determined for 4-MP, a substituted phenol with lower propensity than DMOP and TMP to undergo oxidation, and DMA, a substituted aniline easy to oxidize. 4-MP (at pH 8.1) exhibited no initial concentration effect, whereas the ratio keff(0.1 µM)/ keff(5 µM) for DMA was the same as for TMP (at pH 8.0). Free 692
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 4, 2001
FIGURE 3. Dependence of the pseudo-first-order rate coefficients for the depletion of DMOP (A) on the concentration of SRFA at the photon fluence rate of 3 mEinstein m-2 s-1 (366 nm) and (B) on the photon fluence rate (366 nm) at a SRFA concentration of 2.3 mgC/L.
FIGURE 4. pH dependence of the pseudo-first-order rate coefficients for the depletion of DMOP and TMP in SRFA solution (DOC ) 2.3 mg/L). hydroxyl radical involvement in the photosensitized oxidation of DMOP was tested by using 5 mM 2-propanol as a scavenger. The presence of 2-propanol had no effect on the observed keff at both initial concentrations in solutions of SRFA (pH 8.1).
Discussion The use of the probe phenols DMOP and TMP at a concentration of 0.1 µM revealed that in almost all studied waters and solutions of humic substances there was an essentially faster probe depletion (expressed as the initial pseudo-first-order rate coefficient, keff) than at 5 µM. This confirms and significantly extends the results of previous studies (10, 12) that mainly focused on humic acids. Such a faster photosensitized transformation at the lower concentration was also observed for other organic compounds, viz. DMA (this study) and dimethyl sulfide (17). It demonstrates the need of using lowest concentrations in degradation tests designed to model half-lives of xenobiotics undergoing photosensitized transformations in surface waters. There is an apparent contradiction between the observed pseudo-first-order kinetics of phenol depletion (usually holding for at least one half-life) and the change in keff with phenol concentration, but this can be explained by using the kinetic model described in the Appendix. According to this model, the rate coefficients at 5 µM probe molecule concentration would express the effect of “short”-lived photooxidants, while the rate coefficients at 0.1 µM would comprise the effect of both “short”- and “long”-lived photooxidants. Although the kinetic model is able to describe the present results, we would like to point out that, considering the chemical complexity of DOM, mechanistic conclusions that may be drawn from the model bear a substantial degree of uncertainty. An alternative kinetic model, which assumed complex formation of the probe phenols with DOM, was discarded owing to the small affinity of such phenols for DOM, as it can be estimated from literature data using QSAR analysis (18-20). In one of our previous papers (9), we came to the conclusion that, at 5 µM initial probe phenol concentration, singlet molecular oxygen, the hydroxyl radical, and the superoxide radical anion were not major photooxidants initiating the observed photosensitized transformations. These conclusions also hold for the transformations at 0.1 µM probe phenol concentration. The result of the additional test using 2-propanol as a scavenger confirms that the
FIGURE 5. Dependence of the pseudo-first-order rate coefficients for the depletion of DMOP on the organic carbon content of the water samples. Regression lines (see also Table 3) refer to samples 1-6 (open symbols). Data points for samples 7-10 (closed symbols, not included in the regression) are labeled. hydroxyl radical should not play a significant role in these transformations, which are very likely to be initiated by other photooxidants derived from the DOM. In view of the complex and unresolved chemical structure of DOM, possible photooxidants may include excited triplet states of chromophoric DOM constituents as well as a variety of radical species, e.g. radical cations of aromatic structures, phenoxyl and peroxyl radicals, which may be embedded in DOM macromolecules or present as small molecular fragments. These species are expected to effect a one-electron oxidation of the phenols, as shown in the case of excited triplet states of some aromatic ketone photosensitizers (21), leading to the formation of phenoxyl radicals as primary reaction intermediates. Except for the excited triplet states of the DOM, whose lifetime cannot be longer than about 2 µs in aerated solution due to quenching by oxygen, the above-mentioned radical species may have lifetimes varying over several orders of magnitude and include much longer-lived species than the excited triplet states. The observed direct proportionality between keff and both DOM concentration and photon fluence rate, using the model photosensitizer SRFA, indicates that the quantum yields for the DMOP transformation at 5 and 0.1 µM remain constant over the investigated range of both parameters. According to the kinetic model described in the Appendix, this would imply that the photooxidant formation rates (Ri) are proportional to DOM concentration and photon fluence rate, and thus to the rate of light absorption, and that DOM is not a major scavenger of the photooxidants. Figure 5 shows how the proportionality principle between keff and DOM conVOL. 35, NO. 4, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
693
TABLE 3. Linear Regression Parameters for the Dependence of keff on DOC and a(366 nm) rate coeff keff (5 µM)
independent variable
probe phenol
slope ((s.d.)a
intercept ((s.d.) (min-1)
correlation coeff
DOC
DMOP TMP DMOP TMP DMOP TMP DMOP TMP DMOP TMP DMOP TMP
0.057 ( 0.006 0.041 ( 0.004 0.080 ( 0.013 0.058 ( 0.009 0.107 ( 0.035 0.056 ( 0.009 0.126 ( 0.069 0.072 ( 0.023 0.050 ( 0.033 0.014 ( 0.009 0.046 ( 0.057 0.014 ( 0.016
-0.023 ( 0.012 -0.029 ( 0.010 0.028 ( 0.013 0.008 ( 0.010 0.003 ( 0.078 -0.004 ( 0.020 0.118 ( 0.070 0.051 ( 0.023 0.026 ( 0.075 0.025 ( 0.020 0.090 ( 0.058 0.043 ( 0.016
0.98 0.98 0.95 0.95 0.83 0.95 0.68 0.84 0.59 0.62 0.38 0.40
a(366 nm) keff (0.1 µM)
DOC a(366 nm)
keff mLL
DOC a(366nm)
a
Units: (L mg-1 min-1) for DOC-dependence, (m min-1) for a(366 nm)-dependence.
centration can be extended to different DOMs, using DOC as an indicator. Linear regressions, whose results are displayed in Table 3, were restricted to the less aromatic DOMs to reduce scattering in the data. We arbitrarily set the upper limit of the specific absorption coefficient, a*(280 nm) [used as an index of aromaticity (22)], to 3.0 L mg-1 m-1. This subset of data includes the most common and abundant type of Swiss surface waters. For 5 µM initial DMOP concentration a good correlation between keff and DOC is found for the data subset. For 0.1 µM concentration, data are more scattered even within the subset and the correlation is weaker than for 5 µM. The minimum long-lived component of keff (Figure 5c) is even less correlated to DOC. The same trend was observed using TMP as a probe molecule instead of DMOP as well as employing the absorption coefficient of the water at 366 nm instead of DOC as an indicator of DOM concentration (see Table 3). It can be concluded that DOC (or the absorption coefficient) can be effectively used to predict keff only within a group of waters having lowaromaticity DOM and only at the higher probe phenol concentration. DOC seems not to be the only important factor determining keff, especially at the lower concentration of 0.1 µM and for the long-lived photooxidant component. At the lower probe phenol concentration, rate coefficients in SRFA solutions strongly increased with pH in a range of values common to many freshwaters. As changes in pH are observed in many lakes, mostly due to changes in photosynthetic activity, our results predict an important seasonal variation of reactivity at 0.1 µM probe phenol (corresponding to a variation in concentration of long-lived photooxidants, according to our kinetic interpretation) in these lakes. Such a prediction is currently subject to experimental verification (23). We can only speculate on the origin of such a pH dependence, since variation in the pH could affect various parameters used in the kinetic model. Partial deprotonation of DMOP and TMP cannot explain the pH dependence, because both probes exhibit almost the same dependence but have pKa values that differ as much as 0.9. Possibly, with increasing pH the formation rate of long-lived photooxidants increases as a result of a decrease in DOM oxidation potential following deprotonation of some functional groups.
Acknowledgments We thank Michael Elovitz, Barbara Sulzberger, Yu-Ping Chin, and Penney Miller for helpful discussions and suggestions. We also thank Alberto Barbieri and Daniel Steiner for collecting some of the water samples.
the transformation is brought about by a number of different photooxidants, OXi (i ) 1, ..., N), whose photostationarystate concentration [OXi]ps is controlled by their formation rate Ri, their pseudo-first-order decay processes (rate constant ki), and their reaction with the probe molecule P (secondorder rate constant ki,P), according to eq 1.
[OXi]ps )
694
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 4, 2001
(1)
This equation neglects higher-order terms in the denominator, in particular those derived from reactions of the photooxidant with itself or other transient species, but this approximation is justified in view of the observed direct proportionality of keff to photon fluence rate and sensitizer concentration. Considering that the transformation of P after reaction with OXi occurs with some efficiency ηi,P, the rate of transformation of the probe molecule may be expressed by eq 2.
d[P]
-
dt
N
)
∑k
Riki,Pηi,P
N
i,Pηi,P[OXi]ps[P]
)
i)1
∑k + k i)1 i
i,P[P]
[P]
(2)
By transforming eq 2, initial pseudo-first-order rate coefficients at a given probe molecule concentration, [P]0, are obtained (eq 3).
|
d ln[P]
-
dt
[P])[P]0
N
)
∑k
i,Pηi,P[OXi]ps
i)1
|
) [P])[P]0
Riki,Pηi,P
N
∑k + k i)1 i
i,P[P]0
(3)
At the concentration of [P]0 ) 5 µM we assume that for part of the photooxidants (i ) 1, ..., N1; N1 < N) ki must be at least an order of magnitude greater than ki,P[P]. Using highly reactive probe molecules such as the phenols employed here, this condition is met by all photooxidants having a lifetime of less than about 2 µs (assuming a maximum ki,P of 1010 M-1 s-1) but may also be met by some longer-lived photooxidants. Neglecting the term ki,P[P]0 in eq 3 for the photooxidants satisfying this condition yields eq 4
|
d ln[P] keff([P]0 ) 5µM) ) dt
Appendix We develop here a kinetic model based on the one described in ref 10. The underlying assumption of such a model is that
Ri ki + ki,P[P]
where
N1
=
[P])5µM
∑k′ + i
i)1
N
Riki,Pηi,P
i)N1+1
ki + ki,P‚5µM
∑
(4)
k′i )
Riki,Pηi,P ki
(5)
is a pseudo-first-order rate coefficient describing the transformation of P induced by OXi. In our previous study (10), keff was found to vary only very slightly with [P]0 near [P]0 ) 5 µM (with the only exception of FHA solutions) and the depletion of P followed pseudo-first-order kinetics, which means that the second summation in the right-hand term of eq 4 should be of minor importance. The rate coefficient at 5 µM my be viewed as an upper limit for the component of keff due to short-lived photooxidants (lifetime < 2 µs). At the lower probe molecule concentration, [P]0 ) 0.1 µM (which was chosen based on analytical detection limitations), the number of photooxidants satisfying the condition ki . ki,P[P] will be larger than at 5 µM, and keff will also be larger. Ideally, at infinitely low concentration of the probe molecule one obtains the highest possible rate coefficient, eq 6. N1
eff
k ([P]0 f 0) )
N
∑k′ + ∑ i
i)1
i)N1+1
N
k′i )
∑k′
(6)
i
i)1
Because keff([P]0 f 0) cannot be measured (due to analytical constraints), we may use keff([P]0 ) 0.1 µM) as a lower limit for it. From eqs 4 and 6 we may derive the inequality 7. N
keff([P]0 ) 0.1µM) - keff([P]0 ) 5µM) e
∑
k′i
(7)
i)N1+1
The summation term at the right-hand side of this inequality is a pseudo-first-order rate coefficient related to those photooxidants OXi not satisfying the condition ki . ki,P[P] at [P] ) 5 µM. All of these OXi will have a lifetime of at least 2 µs. We may interpret this term as a lower limit for the longlived component of keff. Operationally, we define the lefthand side of the inequality 7 as the minimum long-lived component (mLL) of keff, eq 8. eff eff keff mLL ) k ([P]0 ) 0.1 µM) - k ([P]0 ) 5 µM)
(8)
Literature Cited (1) Zafiriou, O. C.; Joussot-Dubien, J.; Zepp, R. G.; Zika, R. G. Environ. Sci. Technol. 1984, 18, 358A-371A. (2) Hoigne´, J. In Aquatic Chemical Kinetics: Reaction Rates of Processes in Natural Waters; Stumm, W., Ed.; Wiley: New York, 1990; pp 43-70. (3) Zepp, R. G.; Hoigne´, J.; Bader, H. Environ. Sci. Technol. 1987, 21, 1, 443-450. (4) Zhou, X.; Mopper, K. Mar. Chem. 1990, 30, 71-88. (5) Haag, W. R.; Hoigne´, J.; Gassmann, E.; Braun, A. M. Chemosphere 1984, 5/6, 631-640. (6) Haag, W. R.; Hoigne´, J.; Gassmann, E.; Braun, A. M. Chemosphere 1984, 5/6, 641-650. (7) Haag, W. R.; Hoigne´, J. Environ. Sci. Technol. 1986, 20, 341348. (8) Faust, B. C.; Hoigne´, J. Environ. Sci. Technol. 1987, 21, 957-964. (9) Canonica, S.; Jans, U.; Stemmler, K.; Hoigne´, J. Environ. Sci. Technol. 1995, 29, 1822-1831. (10) Canonica, S.; Hoigne´, J. Chemosphere 1995, 30, 2365-2374. (11) Lin, K.; Carlson, D. J. Mar. Chem. 1991, 33, 9-22. (12) Kawaguchi, H. Chemosphere 1993, 11, 2177-2182. (13) Wegelin, M.; Canonica, S.; Mechsner, K.; Fleischmann, T.; Pesaro, F.; Metzler, A. J. Water Supply Res. Technol. - Aqua 1994, 43, 154-169. (14) Dulin, D.; Mill, T. Environ. Sci. Technol. 1982, 16, 815-820. (15) Zepp, R. G. In Dynamics, Exposure and Hazard Assessment of Toxic Chemicals; Haque, R., Ed.; Ann Arbor Science Publisher Inc.: Ann Arbor, MI, 1980; pp 69-110. (16) Zepp, R. G.; Baughman, G. L.; Schlotzhauer, P. F. Chemosphere 1980, 10, 119-126. (17) Kieber, D. J.; Jiao, J.; Kiene, R. P.; Bates, T. S. J. Geophys. Res. 1996, 101(C2), 3715-3722. (18) Ahel, M.; Giger, W. Chemosphere 1993, 26, 1471-1478. (19) Chin, Y.-P.; Weber, W. J., Jr. Environ. Sci. Technol. 1989, 23, 978-984. (20) Karickhoff, S. W.; Brown, D. S.; Scott, T. A. Water Res. 1979, 13, 241-248. (21) Canonica, S.; Hellrung, B.; Wirz, J. J. Phys. Chem. A 2000, 104, 1226-1232. (22) Chin, Y.-P.; Aiken, G.; O’Loughlin, E. Environ. Sci. Technol. 1994, 28, 1853-1858. (23) Canonica, S.; Barbieri, A., manuscript in preparation.
Received for review March 27, 2000. Revised manuscript received October 23, 2000. Accepted November 2, 2000. ES0011360
VOL. 35, NO. 4, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
695
