Environ. Sci. Technol. 1999, 33, 4256-4262
Photokinetics of Azaarenes and Toxicity of Phototransformation Products to the Marine Diatom Phaeodactylum tricornutum S A S K I A W I E G M A N , * ,† PETER L. A. VAN VLAARDINGEN,‡ WILLIE J. G. M. PEIJNENBURG,‡ SEBASTIAAN A. M. VAN BEUSEKOM,† MICHIEL H. S. KRAAK,† AND WIM ADMIRAAL† Department of Aquatic Ecology and Ecotoxicology, ARISE, University of Amsterdam, Kruislaan 320, 1098 SM, Amsterdam, The Netherlands, and National Institute for Public Health and the Environment (RIVM), P.O. Box 1, 3720 BA Bilthoven, The Netherlands
UV radiation is absorbed by PAHs, structurally altering these compounds into a variety of oxygenated products. Until recently, only hazards of parental PAHs in the environment were investigated. This study aims to determine the fate and effects of azaarenes (N-heterocyclic PAHs) together with their photoproducts in marine environments. Photoreaction kinetics of eight azaarenes, ranging from tworinged to five-ringed structures, were examined using two different light sources: one with an emission peak at 300 nm (UV-B) and the other with an emission peak at 350 nm (UV-A). Azaarenes degraded rapidly in the presence of short-waved light, UV-B being more effective than UVA. Especially preexposure of azaarenes to UV-A radiation led to products toxic to the marine diatom Phaeodactylum tricornutum. Since UV-A constitutes a larger fraction of sunlight at the earth surface and in the water column, photolysis by UV-A may increase the toxic risk of aromatic compounds in the marine environment.
Introduction Several PAHs, including the azaarenes, strongly absorb radiation especially in the ultraviolet (UV) region. This sorption results in two different types of photochemical reactions: photosensitization (1-3) and photomodification (photooxidation and/or photolysis) (2, 3). In photosensitization reactions, PAHs in excited state funnel energy to molecular oxygen, primarily forming singlet oxygen or other radicals (4-6). Such radicals have a very short lifespan but are extremely reactive when formed within an organism because of their capability of oxygenating and oxidizing many different (bio)molecules (7, 4). The second mode of action, photomodification, structurally alters PAHs to a variety of products, mainly oxygenation products (mostly with increased water solubility), via unstable endoperoxide and/or peroxide intermediates (8-11). Some of these oxygenation products are similar to those produced by biodegradation * Corresponding author e-mail:
[email protected]; tel: +31 20 5257717; fax: +31 20 5257716. † University of Amsterdam. ‡ National Institute for Public Health and the Environment. 4256
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 23, 1999
(12, 13, 3). Recently it has been found that many of these photomodified products are more toxic than the parent PAHs (10, 14, 11, 3). This study aims to determine the effects of azaarenes together with their photoproducts formed by UV irradiance in marine environments. First, the kinetics of photolysis in water were determined in laboratory studies and extrapolated to field situations to assess the role of direct photochemical reactions of azaarenes in aquatic environments. To this purpose, UV absorption spectra of the compounds from which the molar absorptivity was derived were measured. Next the photoreaction rate constants of the compounds were determined. From these rate constants the quantum yields, the ratio between the number of molecules undergoing photoreaction and the number of photons absorbed, were calculated. The molar absorptivity of the compounds and their quantum yields were used for estimating the rates of direct photochemical reactions of azaarenes in surface waters. Finally, the acute toxicity of eight different azaarenes and their photochemical transformation products to the marine diatom Phaeodactylum tricornutum was determined. Because algae depend on light for photosynthesis and inhibition of photosynthesis is often a key mechanism of toxicant action in plants and algae (15, 16, 3), inhibition of 14C photosynthetic activity was chosen as an effect parameter.
Experimental Section Selected Azaarenes. Eight azaarenes were tested (Figure 1): quinoline (Aldrich, 99% purity), isoquinoline (Aldrich, 97%), acridine (Aldrich, 97%), phenanthridine (Aldrich, >99%), benz[a]acridine, benz[c]acridine, dibenz[a,i]acridine, and dibenz[c,h]acridine (the last four compounds were obtained from the European Community Bureau of Reference with purities >99.5%). To determine the quantum yields of these azaarenes, a homocyclic PAH, fluoranthene (Fluka, >97%), was used as a photochemical standard. Fluoranthene was chosen because its quantum yields at wavelengths 313 and 366 nm have been reported by Zepp and Schlotzhauer (8). UV-Visible Absorption Spectra. For all compounds, an analytical stock solution was made in acetonitrile (Rathburn, HPLC grade, >99%), because acetonitrile is pristine at wavelengths higher than 290 nm (1) and its refractive index is very close to that of water (17). These stock solutions were kept in the dark at a temperature of 5 °C in crimp-cap (Tefloncoated) sealed bottles (10 or 20 mL) to minimize evaporation of compounds and/or solvent. The azaarene and fluoranthene stocks were diluted in Milli-Q water (Millipore) with a resistance of 18.2 MΩ‚cm-1. The azaarene and fluoranthene concentrations are given in Table 1 (UV spectra, C0). The UV absorption spectra were measured using a Perkin-Elmer Lambda 2 UV-visible spectrophotometer with quartz vessels having a path length of 1 cm. Because azaarenes dissolved in water may ionize or protonate, spectra for the most alkaline azaarenes were recorded with and without a 5 mM phosphate buffer (pH 7). Since no differences were observed between these spectra, no buffers were added in further experiments. From the UV absorption spectra, the molar absorptivity () was calculated for each compound with the LambertBeer law. Photolysis Experiments. For the photochemical reactions of azaarenes and fluoranthene, new solutions in Milli-Q water were prepared, in which the concentrations were maximal 10-5 M (Table 1) to avoid bimolecular reactions other than direct photoreactions. These aqueous solutions (100 mL in quartz test tubes) were irradiated in a Rayonet RPR-208 10.1021/es990228q CCC: $18.00
1999 American Chemical Society Published on Web 10/21/1999
FIGURE 1. Structures of fluoranthene and the eight azaarenes used in this study.
TABLE 1. Nominal Azaarene Concentrations (C0) and Ratios between Milli-Q Water and Carrier Solvent Acetonitrile (mQ/solvent), Used for Recording UV Spectra and for Photokinetics and Actual Initial Azaarene Concentrations (µM) for (photo)toxicity Tests (t0) and Ratios between Carrier Solvent Acetonitrile and Medium % (v/v) UV spectra
photokinetics
toxicity and photoinduced toxicity
compound
C0 (µM)
mQ/solvent
C0 (µM)
mQ/solvent
quinoline isoquinoline acridine phenanthridine benz[a]acridine benz[c]acridine dibenz[a,i]acridine dibenz[c,h]acridine fluoranthene
58 56 98 20 0.85 0.99 0.95 0.95 0.10
1000/1 1000/1 1000/1 1000/1 90/10 90/10 90/10 80/20 90/10
10 10 10 10 1 1 1 0.1 0.5
1000/1 1000/1 1000/1 1000/1 90/10 90/10 90/10 90/10 90/10
t0 (µM)
solvent (% v/v)
300 nma (% t0)
350 nma (% t0)
additionalb (µM)
5.99 5.52 4.75 5.95 0.04 0.06 0.07 0.01
0.07 0.07 0.07 0.07 0.67 0.67 0.67 0.67
25 24 69 46 36 18 29 28
11 28 20 70 43 32 16 42
2023 1863 55.8 108 4.36 2.62 0.95 0.024
a Percentages of initial azaarene concentration present after irradiation with 300- or 350-nm lamps (% t ). Irradiation times with respectively 0 300 and 350 nm are variable for the azaarenes, see preparations of azaarene solutions for toxicity tests. bHighest actual azaarene concentrations in 0.07% (v/v) acetonitrile and medium, for determining EC50 values in toxicity tests, referred to as additional solutions.
merry-go-round-reactor (MGRR) containing eight UV lamps. Two types of UV lamps were used: one in the spectral region of 250-350 nm (see ref 18) with an irradiation optimum around 300 nm (3000 Å Rayonet, 21 W) and the other lamp (3500 Å Rayonet, 24 W) emitted in the spectral region of 305-410 nm with its optimum around 350 nm (the relative energy distributions of both lamps are given by The Southern New England Ultraviolet Co., Hamden, CT). The lamps were cooled with air, which kept the working temperature at 27 °C. To measure the relative azaarene and fluoranthene concentrations, 1-mL water samples were analyzed by HPLC. For each sample the average of two analyses was used in the calculations. Time intervals for measuring these relative concentrations depended on the reactivity of the compound and, hence, differed for each compound. Calculation of Photolysis. Relative concentrations of azaarenes and fluoranthene, determined with HPLC, were plotted as ln(C0/Ct) against time. The slope of the linear regression line obtained through the time-concentration plot equals the pseudo-first-order photoreaction rate constant kexp. Photolysis obeyed to a first-order model, except for dibenz[a,i]acridine, which was better described by a zeroorder model. Correlation coefficients (r 2) of the regression lines fitted through these time-concentration plots were high, between 0.99 and 1.0, except for quinoline (0.97). For each compound, the overlap between the UV absorption spectrum of the chemicals and the irradiance emitted by the UV lamps, i.e., the sum of the products of
relative intensity of the light source and the molar absorptivity of a chemical over the relevant wavelength range (∑(I)λ) was calculated (18). With these two parameters and the reported quantum yields (φ) of fluoranthene (8), quantum yields can be calculated for the selected azaarenes. Since absolute measurements of light intensity (I) were not available, eq 1 was used:
∑ ∑
kexp,x (I)λ,ref φx ) φref kexp,ref (I)λ,x
(1)
Equation 1, in which φx is the quantum yield of compound x and φref is the quantum yield of fluoranthene, was derived from the following equation, in which j is a conversion factor of 1 einstein/mol (1):
kexp )
∑(I)
λ
× 2.303j-1φ
(2)
Because the decrease of dibenz[a,i]acridine was better described with a zero-order model, the quantum yield could not be calculated in the same manner as for the other compounds. Therefore this compound is excluded from Table 2. First-order photoreaction rate constants (kexp) of the threeringed azaarene acridine and the standard fluoranthene were also determined under natural light conditions on the roof of the Biological Centre in Amsterdam (52°21′ N, 4°57′ E), VOL. 33, NO. 23, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4257
TABLE 2. Photokinetics for the Azaarenes and Fluoranthene under 300 and 350 nm Irradiancea irradiated with 300-nm lamp
irradiated with 350-nm lamp
compound
Σ(IE)λ
kexp (min-1)
O (10-5)
N
Σ(IE)λ
kexp (min-1)
O (10-5)
N
quinoline isoquinoline acridine phenanthridine benz[a]acridine benz[c]acridine dibenz[a,i]acridine dibenz[c,h]acridine fluoranthene
11910 9381 10067 21060 76425 94670 186528 157789 28913
0.0104 ( 0.0004 0.0076 ( 0.00009 0.0058 ( 0.0003 0.0098 ( 0.0003 0.0010 ( 0.00004 0.0017 ( 0.00005 0.0901 ( 0.0036 0.0316 ( 0.0017 0.0014 ( 0.00002
211.8 196.1 139.9 112.3 3.32 4.46 117.2 48.59 12.0
12 10 6 10 15 11 10 14 15
6035 1868 139322 16219 145929 195884 86537 119981 110597
0.00014 ( 0.000004 0.00075 ( 0.000008 0.0019 ( 0.00006 0.0038 ( 0.0003 0.00069 ( 0.00002 0.0013 ( 0.00002 -24.330 ( 1.18b 0.01206 ( 0.00063 0.0015 ( 0.00004
0.33 5.95 0.20 3.46 0.07 0.10 ncc 1.51 0.20
15 15 16 14 15 16 16 16 15
irradiated with daylightd compound
Σ(IE)λ
kexp (day-1)
O (10-5)
fluoranthene acridine
15179 16772
0.20 0.66
1.33 3.93
a Given are the summed overlap spectra of the light source (I) used and molar absorptivity () of the compounds (M-1 cm-1), photoreaction rate constants (kexp), and quantum yields (φ). The values of kexp are derived from a pseudo-first-order model. N is the number of observations for determing kexp. The quantum yields of azaarenes in Milli-Q water irradiated with 300- or 350-nm lamps were calculated using the φ of fluoranthene from Zepp and Schlotzhauer (8) as calibration. b Derived from a zero-order reaction model (M min-1). c nc, not calculated. d For the compounds fluoranthene and acridine, photokinetics were also determined under natural sunlight (May 23-27, 1997). The summed overlap spectra were calculated with intensities of sunlight in the spectral region of 300-370 nm obtained from the Royal Netherlands Meteorological Institute, De Bilt, The Netherlands.
The Netherlands, on May 23-27, 1997. Solutions of acridine and fluoranthene in 100-mL quartz test tubes with initial concentrations of 10-5 M and 1% (v/v) acetonitrile in Milli-Q water for acridine and 5 × 10-7 M and 10% (v/v) acetonitrile in Milli-Q water for fluoranthene were exposed to 50 h of sunlight. From the decrease in acridine and fluoranthene concentration, half-lives were calculated with the equation t1/2 ) ln(2)/kexp. To generate half-lives of the azaarenes in near-surface waters under clear skies, molar absorptivities () and quantum yields were incorporated in the program GCSOLAR (19). In this program, variables such as latitude, time of year, thickness of the ozone layer, and thickness of the water layer are taken into consideration for the calculations of photolysis rates under realistic environmental conditions. These estimated environmental reaction rates are based on the overlap of spectra with midseason sunlight (from 300 to 420 nm) in near-surface waters. Although it was assumed for all compounds that the quantum yield is independent of the wavelength (1, 20), some of the measured quantum yields appeared to depend strongly on the wavelength. To incorporate both quantum yields in GCSOLAR, molar absorptivities in the spectral region of 330-420 nm were corrected with the ratio of the corresponding quantum yields at 350 and 300 nm. Preparations of Azaarene Solutions for Toxicity Tests. For the toxicity tests with azaarenes, new solutions in Milli-Q water were prepared, in which the ratios of acetonitrile and Milli-Q water were lower than 1:10 (v/v) (Table 1). A pilot study had shown that 1% and lower (v/v) acetonitrile gave no additional toxicity, but 10% (v/v) was lethal to algae. The freshly prepared aqueous solutions, which were irradiated in the same way as described above, were sampled at times t ) 0 (untreated sample) and at about t ) 2t1/2 to generate samples to perform later on 14C photosynthesis tests with the marine diatom P. tricornutum. Theoretically, 75% of the parent compound would be expected to exist in the photomodified state at time t ) 2t1/2. At both sampling times, 1-mL samples from each azaarene solution were collected for HPLC analysis, and 20-mL samples were taken for 14C photosynthesis experiments. HPLC samples were stored at 5 °C; photosynthesis samples were immediately frozen in liquid nitrogen and stored at -70 °C. 4258
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 23, 1999
With these samples, inhibition of the photosynthetic activity was determined, and the toxicity of unmodified (t0) and modified azaarenes was compared. Some of the initial azaarene concentrations (t0, Table 1) were thought to be below effect concentrations, which would hamper a comparison between toxicity of unmodified azaarenes and toxicity of azaarenes irradiated with respectively the 300- and 350-nm lamp. To achieve full-scale doseresponse curves for unmodified azaarenes, higher concentration series were prepared in Milli-Q water with a constant acetonitrile/Milli-Q water ratio (0.1% v/v acetonitrile), and these are referred to as additional solutions (Table 1). Toxicity Tests. The diatom Phaeodactylum tricornutum (obtained from AquaSense BV, Amsterdam) was cultured in the laboratory in a continuous culture in artificial seawater (21) at 20 °C at a dilution rate of 24 mL/h. The culture was illuminated with circular fluorescent tubes (100 µeinstein m-2 s-1, Philips TLE 32W/33) in a light:dark regime of 16:8 h. The algal density in the continuous culture was (2 × 106 cells/mL. The unmodified (t0 and additional solution) and irradiated solutions (Table 1) were diluted with medium to a series of 66.6%, 20%, 6.6%, 2%, 0.66%, 0.2%, and 0% (control) solution in 20-mL scintillation vials in duplicate. Macro salts (NaCl, MgCl2, CaCl2, Na2SO4, K2SO4, NaHCO3) were added in order to keep the salinity of all solutions equal to the salinity of the artificial seawater with algae. Algae were added to the scintillation vials such that the initial concentration of algae used in 14C photosynthesis experiments was 2 × 105 cells/ mL. Scintillation vials with algae and azaarene solution were kept at 20 °C, under cool-white fluorescent tubes (Osram, L13W/20; 100 µeinstein m-2 s-1) on a rotary shaker. After 1 h, 100 µL of NaH14CO3 (Amersham; 50-60 mCi mmol-1) with an activity of 0.5 µCi was added to each vial. The biological activity in the scintillation vials, incubated with 14C, was stopped after another 1-h period by adding 0.6 mL of formaldehyde (Merck, 37%). In each test series, controls with algae and 1.67% (v/v) acetonitrile were incorporated. Besides controls (algae without toxicant), two additional vials were used for determination of the incorporation of abiotic 14C (controls without algae and toxicant). Nonincorporated 14C was removed overnight, by adding 100 µL of 6 N HCl (Merck,
37%) to the vials, leading to evaporation of the carbon dioxide in a fume hood. Activity was measured after addition of 7 mL of scintillation liquid (Instagel; Hewlett-Packard), using a liquid scintillation analyzer (Packard Tricarb, model 1600 TR). Photosynthetic activity was calculated and expressed as disintegrations per minute (dpm). Photosynthetic activity in treatments was expressed as percentage of the mean of the two corresponding controls. The photosynthetic activity values were plotted against the corresponding actual azaarene concentrations in the water. EC50 values (including 95% confidence limits) were obtained by fitting the following equation (22) through dose-response plots using Kaleidagraph:
Y)
c 1 + eb(X - a)
in which Y ) inhibition (%), X ) 10log concentration (µM), a ) 10log EC50, b ) slope of the logistic curve, and c ) photosynthetic activity of controls (100%). HPLC Analysis. Samples from the highest azaarene concentration of t0 solutions, of solutions irradiated with 300- and 350-nm lamps and additional solutions (Table 1), were analyzed by HPLC using fluorescence detection (Kratos Spectroflow 980) for the (di)benzacridines and UV detection (Applied Biosystems model 785A) for the other compounds. A 150 × 4.6 mm Lichrosorb 5 µm RP-18 analytical column was used with a 4 × 4 mm Lichrosorb 5 µm RP18 guard column. The column temperature was kept at room temperature (20 °C). The flow of the mobile phase, an isocratic mixture of acetonitrile (J. T. Baker Analyzed HPLC Reagent, minimum 99.9%) and water (J. T. Baker Analyzed HPLC Reagent) with varying composition according to the compound to be analyzed was 1 mL/min. A total of 20 µL of each sample was injected automatically. Azaarene concentrations were calibrated with corresponding standards in methanol (J. T. Baker Analyzed HPLC Reagent, minimum 99.8%). From the irradiated solutions, only the concentrations of parent compound were quantified. The photoreaction products were, although observed in several HPLC chromatograms, not identified. In the case of acridine, 9(10H)acridone was identified as main photoproduct (23), but other photoproducts with almost similar retention times as acridine were present. Since a whole range of unknown products is formed, it is not possible to identify and quantify all these products. To compare the toxicity of the irradiated solutions with the toxicity of unmodified azaarenes, the toxicity of the irradiated solutions was expressed in the concentration of the parent azaarenes present after photolysis. Although azaarene concentrations of irradiated solutions diminished, EC50 values expressed as parent azaarene concentrations would not alter if the photoproducts were not toxic. All deviations from the EC50 values of the azaarenes can thus be explained by toxicity of phototransformation products.
Results Photolysis of Azaarenes. The overlap between the light absorption of azaarenes and both the spectrum of the 300and 350-nm lamp (∑(I)λ) was highest for the four- and fiveringed structures and lowest for the two-ringed compounds (Table 2). The experimentally determined photoreaction rate constants (kexp) for all compounds were higher at 300 nm than 350 nm (Table 2). The difference between the lamps was most pronounced for quinoline, which degraded slowly under the 350-nm lamp because quinoline absorbs hardly any irradiance of this lamp. In the experiments with the 300-nm lamp, the rate constant of quinoline was comparable to those of isoquinoline, phenanthridine, and acridine (Table 2). In
general, the five-ringed structures dibenz[a,i]acridine and dibenz[c,h]acridine were the most photochemical unstable structures, which was in agreement with the high overlap between the light absorption of these compounds and the spectral distribution of the lamps. Benz[a]acridine and benz[c]acridine on the contrary were the most stable compounds under both light regimes (Table 2), despite their high overlaps with the lamps used. From the spectral overlap (∑(I)λ) and photoreaction rate constants (kexp), the quantum yields (φ)sor efficiency of photochemical reactionssof the azaarenes irradiated with 300 and 350 nm were derived, using the photochemical standard fluoranthene. The quantum yield at 300 nm was highest for quinoline and isoquinoline and lowest for benz[a]acridine and benz[c]acridine. At 350 nm, the efficiency of photochemical reactions was highest for isoquinoline and phenanthridine and lowest for both benzacridines. Dibenz[a,i]acridine, irradiated with 350 nm, adhered to zero-order kinetics, and the quantum yield (φ) was not calculated (Table 2). In contrast to the assumption that φ does not change over short intervals of wavelengths, the φ of azaarenes at 300 nm differed from the ones at 350 nm. For acridine and fluoranthene, photoreaction kinetics were also determined under sunlight conditions in an outdoor experiment in early summer. The photoreaction rate constants of acridine and fluoranthene in pure water were respectively 0.66 and 0.20 day-1 (the average day length was 14.3 ( 0.1 h) in quartz vessels. In this outdoor experiment, quantum yields were calculated from the available amount of energy in the wavelength region of 286-360 nm (data obtained from Royal Netherlands Meteorological Institute, De Bilt; 52°10′ N, 5°18′ E) (Table 2). However, the calculated quantum yields obtained with 300 and 350 nm irradiance in the laboratory did not correspond with the environmental quantum yields obtained at wavelengths that varied between 286 and 360 nm (Table 2). Toxicity of Azaarenes and Phototransformation Products. In the t0 treatments, the maximal azaarene concentrations used were e10-5 M (Table 1). Only the tested concentrations of acridine, phenanthridine, and benz[a]acridine had a slight inhibitory effect on the photosynthetic activity. The tested concentrations of all other compounds had no effect. To obtain effect concentrations for all unmodified compounds, higher azaarenes concentrations were tested (additional solutions, Table 1). Clear dose-response relationships were observed for the additional solutions of quinoline, isoquinoline, acridine, phenanthridine, and benz[a]acridine (Figure 2 and Table 3). The tested concentrations of benz[c]acridine, dibenz[a,i]acridine, and dibenz[c,h]acridine had no effect. Therefore, it was concluded that the EC50 for these compounds exceeds the highest concentration tested (Table 3). Since for the EC50 values of acridine, phenanthridine, and benz[a]acridine between the t0 and additional solutions an overlap in 95% confidence limits were observed, only EC50 values based on additional solutions are presented in Table 3. In general, the toxicity of the compounds coincided with an increase in aromatic rings or lipophilicity, a pattern previously observed for azaarenes (24). After irradiation with the 300-nm lamp (UV-B), the concentrations of most azaarene solutions had no effect, only those of irradiated acridine and phenanthridine solutions (Table 3) had any effect. Since the EC50 of the acridine and phenanthridine solutions equaled the EC50 of the irradiated solutions, only unmodified acridine and phenanthridine caused the toxicity of these mixtures. After irradiation with the 350-nm lamp (UV-A), toxicity of the quinoline and isoquinoline solutions increased 2 orders of magnitude as compared to that of the unmodified compounds. Toxicity of the irradiated acridine and phenanthridine solutions also increased, although this difference VOL. 33, NO. 23, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4259
FIGURE 2. Photosynthetic activity (%) of the marine diatom Phaeodactylum tricornutum at different concentrations of the compounds quinoline, isoquinoline, acridine, phenanthridine, and benz[a]acridine before and after irradiation with the 300- or 350-nm lamps, plotted as percentage of the corresponding controls. Data of the additional solutions and t0 series (see Table 1) are plotted together as unmodified azaarenes.
TABLE 3. Calculated EC50 Values and the Corresponding 95% Confidence Limits for the Effects of Unmodified Azaarenes (Based on Additional Solutions) and Modified Azaarenes after Irradiation with 300- and 350-nm Lamps on the Photosynthetic Activity of Phaeodactylum tricornutum EC50 values (95% confidence limits) in µM compound
unmodified
300 nma
350 nma
quinoline isoquinoline acridine phenanthridine benz[a]acridine benz[c]acridine dibenz[a,i]acridine dibenz[c,h]acridine
554 (479-641) 385 (308-481) 12.6 (10.1-15.8) 14.8 (13.1-16.7) 0.29 (0.23-0.36) >0.38c >0.95d >0.02d
neb ne 10.8 (6.9-16.8) 8.13 (4.54-14.6) ne ne ne ne
1.8 (1.24-2.6) 4.3 (0.88-20.8) 2.23 (1.36-3.65) 5.85 (2.24-15.2) ne ne ne ne
a Effect concentrations are based on actual concentrations of parent azaarene. b ne, no effect at highest concentration of modified azaarenes tested. c Actual concentration. d Nominal concentration.
was only significant for acridine. Since the toxicity of irradiated (350 nm) quinoline and isoquinoline solutions increased orders of magnitudes, toxicity was mostly caused 4260
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 23, 1999
by new products. Toxicity of acridine and phenanthridine solutions irradiated with the 350-nm lamp, however, was caused by the unmodified compounds and the photoprod-
ucts jointly (Figure 2). The tested concentrations of irradiated benzacridines and dibenzacridines solutions had no measurable effect (Table 3).
Discussion Photolysis of Azaarenes. Leifer (1) assumed that quantum yields are independent of wavelength. Also Zepp (17) assumed an independency of wavelength in the region of sunlight absorption and argued that φ cannot change drastically over such a short wavelength range (8). Indeed quantum yields of the PAHs benz[a]anthracene and benzo[a]pyrene in water did not change with wavelength (25). In our study, however, a large discrepancy was observed between quantum yields (φ) of fluoranthene and azaarenes at 300 and 350 nm, demonstrating that different excitation processes played a role resulting in different reactions, each with their specific quantum yields. Only for quinoline, photoreaction parameters in outdoor experiments are reported. Mill et al. (25) measured half-lives of 17.3 days in pure water and of 4.5 days in water with humic acid in June at 40° N. They also computed half-lives of 22 days in midsummer and 160 days in midwinter in nearsurface waters. Kochany and Maquire (26) estimated halflives of 15 and 133 days, while we estimated half-lives for quinoline of 2.5 and 41 days respectively for midsummer and midwinter (with GCSOLAR). The difference between our estimated half-lives and that of the others was mainly caused by differences in determined quantum yields; these were respectively 212 × 10-5 at 300 nm (Table 2), 25 × 10-5 at 313 nm (26), 32 × 10-5 at 313 nm (25), and 0.33 × 10-5 at 350 nm (Table 2). Kochany and Maquire (26) computed a quantum yield of 91 × 10-5 for quinoline in sunlight. As for acridine and fluoranthene, this environmental quantum yield fell between the φ of 300 and 350 nm (Table 2), indicating that the characteristics of the lamps used in experiments greatly affect the determination of quantum yields. In our setup, the UV-B lamp emitted in the region 250-350 nm and the UV-A lamp emitted in that of 305-410 nm, resulting in quantum yields integrated over this wavelength region. The cited studies used a much smaller region of UV. Therefore in photoreactions experiments, outdoor tests need to be incorporated, facilitating a reliable estimation of photolysis and environmental half-lives. Toxicity of Azaarenes and Phototransformation Products. Especially UV-A radiation increased the hazard of azaarenes due to formed phototransformation products in combination with or without parent azaarenes. Although plants and algae are still assumed to be more tolerant to PAHs than evertebrates (27), recently Huang et al. (2) observed that photomodified PAHs have an unique mechanism of toxicity to photosynthesis of plants and consequently of algae, which makes them very sensitive to modified PAHs and NPAHs, as shown in this study. In our setup photosensitization, a process in which radicals oxygenate and oxidize various molecules (7, 4), did not contribute to the increase in toxicity, because azaarenes were modified prior to toxicity tests and radicals have an extremely short lifespans. Furthermore, during the toxicity tests photosensitization did not occur, since fluorescence light, which contains no radiation below 400 nm (28), was used. Thus, the observed increase in toxicity by UV-A radiation was due to phototransformation products. In contrast to toxicity of UV-A-modified azaarenes, the toxicity of UV-B-modified azaarenes is primarily caused by the parent azaarenes. There are two reasons for these particular differences between toxicity of UV-A- and UV-B-modified azaarenes. First, UV-B is a more reactive radiation source than UV-A, which is stressed by the short half-lives of azaarenes under UV-B radiation in comparison to UV-A (
