Environ. Sci. Technol. 2003, 37, 4717-4723
Evaluation of Fluorescence Quenching for Assessing the Importance of Interactions between Nonpolar Organic Pollutants and Dissolved Organic Matter DEBERA A. BACKHUS,* CHRISTINA GOLINI, AND EDWIN CASTELLANOS† Environmental Science Research Center, School of Public and Environmental Affairs, Indiana University, Bloomington, Indiana 47405-1701
The assumptions behind the fluorescence quenching (FQ) method were thoroughly evaluated to assess its potential for quickly and accurately assessing the importance of hydrophobic organic contaminant-macromolecular organic carbon interactions in aquatic systems. Perylene was used as the probe molecule to avoid problems encountered with other fluorescent probes. Results from a wide range of wetland samples suggest that static quenching dominates, that other quenchers do not interfere with analyses, and that full quenching on sorption does not occur for all samples. The latter result indicates that the quantum yield of the sorbed probe must be accounted for in quantifying the magnitude of Kmoc values by FQ. Observed Kmoc values compared favorably with those measured by the solubility enhancement method. Overall, our results suggest that FQ can be used as a quick and reliable screening tool as long as precautions are taken to ensure the validity of the results.
of magnitude (13, 23, 24). Although these studies have provided important insights into the factors affecting HOCDOM interactions, unfortunately, no universal predictive relationship has emerged to allow for the estimation of Kdoc values from readily obtained information. This creates a significant dilemma for practitioners and decision-makers who require good estimates of Kdoc to quantify, understand, and model fate, transport, and effects of HOCs in particular situations. Until such a predictive relationship is found, there is a need for a simple, quick, cheap, accurate, and reliable method to assess the importance of HOC-DOM interactions at sites of interest. Ideally, such a method could be used on aqueous environmental samples that require little or no processing prior to testing. In this paper, we examine whether fluorescence quenching (FQ) techniques may serve as such a method. FQ has been described as a quick, elegant, sensitive, precise, and reproducible method that does not require separation of equilibrated phases or large quantities of DOM. Furthermore, the method can be used for samples containing environmentally relevant levels of DOM and pollutants (25-28). Despite the many advantages of the FQ method, it is not problem-free (3, 26-32), as valid Kdoc values are only obtained if a number of assumptions hold true. To this end, the objective of this study is to thoroughly evaluate the assumptions behind the FQ method and in doing so to assess its utility for quickly, cheaply, and easily assessing the potential importance of HOC-DOM interactions on a site-by-site basis. Our investigation focuses on water samples collected directly from the environment and requiring little or no processing or alteration prior to analysis. This ensures the relevance of the measured Kdoc in the specific environment from which the sample was obtained and avoids the time and expense associated with processing large quantities of water to obtain DOM fractions. We examined a wide variety of water samples collected from wetlands to ensure that the FQ technique works well in general for environmental samples. The focus on wetlands is due to the expectation that HOC-DOM associations will be of greatest importance in these high [DOM] environments.
Introduction
Fluorescence Quenching Theory
The literature is replete with information indicating that dissolved organic matter (DOM) can influence the fate and effects of hydrophobic organic contaminants (HOC) in aqueous environments. Many researchers have shown that the bioavailability and consequently toxicity of HOCs to aquatic organisms are often reduced in the presence of DOM (1-4) and that the transport (5-9) and degradation (10-12) of HOCs can be affected by DOM. Understanding the fate and risks of contaminants in aquatic environments requires a thorough knowledge of and ability to predict the magnitude of HOC-DOM interactions. Many studies have investigated HOC-DOM partition coefficients (e.g., Kdoc, Koc, Koc,hs, or Kom) (13). Collectively these works show that the magnitude of Kdoc values depends on the hydrophobicity of the contaminant, properties of the DOM (14-20) (e.g., polarity, aromaticity, size, aliphatic or polymethylene content, and elemental composition), and solution chemistry (21, 22). For a given HOC, differences in DOM properties can lead to Kdoc variations over several orders
To aid the reader in understanding the assumptions inherent in the FQ technique, we provide a quick review of the theory and equations used in translating FQ observations to valid Kdoc values. More detailed derivations and discussion can be found in the literature (5, 33-35). The Stern-Volmer equation is typically used to describe the relationship between the concentration of quencher present in a sample [Q] and the ratio of a probe molecules fluorescence in the presence (F) and absence (Fo) of that quencher:
* Corresponding author phone: (812)855-0563; fax: (812)855-7802; e-mail:
[email protected]. † Present address: Department of Biology and Environmental Studies, Universidad del Valle de Guatemala, Guatemala. 10.1021/es026388a CCC: $25.00 Published on Web 09/19/2003
2003 American Chemical Society
Fo/F ) 1 + Ksv[Q]
(1)
where Ksv is the Stern-Volmer constant. Interpretation of the Ksv differs depending on whether a static (probe is bound to the quencher) or dynamic (probe collision with the quencher) process dominates. Unlike static quenching, the dynamic process leads to a decrease in the fluorescent lifetime of the probe. This difference can be used to distinguish between these quenching mechanisms. For dynamic quenching, Fo/F is related to the probes’ fluorescence lifetime:
Fo/F ) τo/τ ) 1 + kqτo[Q] ) 1 + Ksv[Q]
(2)
where τo and τ are lifetimes in the absence and presence of VOL. 37, NO. 20, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4717
quencher, respectively, and kq is the bimolecular quenching constant. For dynamic quenching, Ksv is equivalent to kqτo. In this work, the quencher of greatest interest is the portion of DOM that is capable of sorbing the probe molecule. We operationally define this portion as DOM that is not removed by centrifugation (2000g for 20 min) and is retained by a 500 molecular weight cutoff ultrafilter, hereafter referred to as macromolecular organic carbon (MOC). It can be shown that Ksv is equivalent to Kdoc or hereafter Kmoc if several key assumptions hold true: (i) a static quenching process dominates, (ii) MOC is the only quencher both present in the sample and absent in the quencher-free solution, (iii) the sorbed probes fluorescence is fully quenched on binding, and (iv) no other processes affect the observed probe fluorescence (i.e., no losses by photodegradation, sorption to experimental vessels, volatilization, or other processes and the inner-filter effect (IFE) is either insignificant or properly corrected for). If these assumptions hold true, then eq 1 can be rewritten as
Fo/F ) 1 + Kmoc[MOC]
(3)
As discussed below, eq 3 can be used in several ways to quantify Kmoc for a given sample. By carefully choosing the probe molecule to minimize dynamic quenching, photodegradation, and IFE interferences; accounting for potential losses; and using ultrafiltered samples for Fo determinations, one can maximize the chances that the assumptions listed above are not violated. Equation 3 can be modified as follows to examine and as necessary account for the possibility that some portion of the sorbed probe remains fluorescent (5):
1 + φKmoc[MOC] F ) Fo 1 + Kmoc[MOC]
(4)
Using a nonlinear curve-fitting program, Kmoc and φ (the fluorescence quantum yield of the sorbed probe) can be determined by examining F/Fo versus [MOC] for concentrated and/or diluted environmental samples.
Experimental Section Probe Chemical. Perylene (Aldrich, red label, 99+% pure) was used without further purification. Samples were spiked with small aliquots of a 0.4 mg/L stock solution of perylene in methanol (EM Science Omnisolv, glass distilled) to create an initial concentration of 0.3 µg/L in all samples. The small volume fraction of methanol carrier present in these sample (7.5 × 10-4 volume fraction) is not expected to affect probe partitioning (5). The initial dissolved-phase concentration, prior to perylene partitioning to MOC and sorption to the experimental system (cuvette, stirbar, and air-water interface), was three-fourths of the aqueous solubility of perylene. At equilibrium, the dissolved-phase concentration was well below the probe solubility. Wetland Samples. Surface water samples were collected from three wetland systems: Old Woman’s Creek National Estuarine Research Reserve and State Nature Preserve in Huron, OH (OWC); the Suwannee River in Fargo, GA (SR); and the Great Dismal Swamp in Virginia (DS). At the OWC site, samples were collected from three different locations designated inlet, RR, and outlet at five time points between August 1995 and March 1997. All samples were collected from 10 to 35 cm below the surface using a submersible pump (Fultz Pump Inc., Lewistown, PA). Samples were collected and stored in 60- and 300-mL glass biological oxygen demand bottles, containing no headspace. All glassware used in collecting and investigating samples was thoroughly cleaned to remove all traces of organic matter (i.e., baking in a muffle furnace at 450 °C or soaking in a solution of Nochromix (Godax Laboratories, Inc.) or 50% hydrogen 4718
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 20, 2003
peroxide). Samples were stored in a cooler or refrigerator on site, transported on ice, and then stored in the dark at 4 °C until analysis. Ultrafiltration. Samples were ultrafiltered to meet three needs: (i) produce MOC-free samples for assessment of Fo, (ii) create a concentrated solution of MOC for determination of Kmoc and φ values from multipoint analyses, and (iii) assess the concentration of MOC in samples. A stirred cell ultrafiltration apparatus (Spectrum, Molecular/Por, 400-mL capacity) outfitted with a 76-mm 500 molecular weight cutoff ultrafiltration membrane (Amicon, Diaflo YC05) was used to accomplish these tasks. Membranes were soaked in Milli-Q water (Millipore, Milli-Q UV plus system), then seated in the stirred cell apparatus, and flushed with Milli-Q water (1-2 L) until the organic carbon content and fluorescence of the effluent reached background levels. Prior to ultrafiltration, about 360 mL of sample was centrifuged for 20 min at 2000g to remove any particulate material larger than clay flakes (e.g., diameter of 0.8 µm and density of 2.6 g/mL). The apparatus and ultrafilter were rinsed with 10 mL of centrifuged sample prior to ultrafiltration of about 300 mL of sample using 60 PSIG of zero-grade argon. When the sample concentrate remaining in the stirred cell reached a volume of 30-100 mL, the ultrafiltration process was terminated. The concentrate was then rinsed from the stirred cell and membrane using about 10 mL of ultrafiltered sample. Dilution series for the multipoint FQ experiments were created by mixing varying proportions of the sample concentrate and ultrafiltrate. In a few cases, MOC levels in the original samples were sufficiently high to allow for direct dilution of the sample without preconcentration. Organic Carbon Content. The total organic carbon (TOC) content of both whole and ultrafiltered samples was measured in triplicate using a Shimadzu 5500 TOC analyzer outfitted with a high-sensitivity catalyst. Samples were acidified with 2 N HCl and bubbled with purified air to remove inorganic carbon prior to analysis. Potassium biphthalate (Kanto Chemical Co., reagent grade) was used to calibrate the detector. The [MOC] was determined as the difference between the TOC measured in the whole sample, concentrates, or diluted concentrates and the corresponding sample ultrafiltrate. FQ Methods. Several types of tests were used to examine assumptions behind the FQ method. The potential role of oxygen as a second quencher was examined by comparing argon-purged and nonpurged samples, typically containing in situ levels of MOC (single-point analysis). For these tests, Kmoc was determined by a simple rearrangement of eq 3. Two types of multipoint analyses were accomplished using dilution series of the MOC concentrates produced by ultrafiltration: the traditional Stern-Volmer linear regression (LR) analysis and the quenching efficiency (QE) analysis described by Backhus and Gschwend (5). To account for loss of perylene to the cuvette system, all analyses made use of time-series fluorescence measurements to back-extrapolate data to a time zerosno wall loss point (5). All fluorescence measurements were obtained on a SLM Aminco series 2 luminescence spectrometer with slit widths at 4 nm. The instrument was outfitted with a 4-cuvette turret/ stirring mechanism. Fluorescence data was collected at Ex 434 nm/Em 467 nm in 10-mm path length quartz cuvettes containing 3 mm diameter × 6.25 mm Teflon micro flea stir bars. Absorbance measurements were obtained using a Hitachi U-2000 spectrophotometer at 434 and 467 nm for IFE correction of fluorescence data (36). Correction factors were generally around 1.1 and never exceeded 1.3. For each solution examined (whole, concentrated, diluted, or ultrafiltered sample), a common sequence was followed: (i) 2 g of solution was added to a cuvette, (ii) absorbance data were obtained, (iii) four cuvettes containing sample and a
TABLE 1. Sample Properties and Evidence for Static versus Dynamic Quenching
sample
in situ [MOC] (mg/L)
% MOCa
MOC range for multipoint expts (mg/L)
Fo/F range
Ksv dynm max/ Ksv obs (%)
R2
P value
interceptb
slope ) Kmocc (mL/gmoc ( %)
SR 3/96 DS 3/96 OWC inlet 8/95 OWC inlet 11/95 OWC inlet 8/96 OWC inlet 3/97 OWC RR 8/95 OWC RR 8/96 OWC outlet 8/95 OWC outlet 11/95 OWC outlet 8/96 OWC outlet 9/96 OWC outlet 3/97
30.7 40.7 7.6 4.5 7.3 2.9 10 8.6 4.9 3.4 4.4 2.3 2.2
90 94 89 96 94 81 83 80 88 68 69 89 64
0-30 0-22 0-31 0-22 0-21 0-16 0-48 0-39 0-32 0-20 0-7.2 0-13 0-24
1-6 1-2.5 1-2 1-1.9 1-1.8 1-1.6 1-2.3 1-2.6 1-3.6 1-1.3 1-1.5 1-1.5 1-1.7
0.05-0.1 0.1-0.2 0.3-0.4 0.3-0.5 0.2-0.3 0.3-0.4 0.4-0.5 0.2-0.3 0.1-0.2 1.0-1.5 0.1-0.2 0.2-0.4 0.3-0.5
0.988 0.927 0.939 0.718 0.731 0.799 0.851 0.941 0.924 0.513 0.582 0.690 0.857
