Environ. Sci. Technol. 2004, 38, 148-155
Sorption of Phenanthrene to Environmental Black Carbon in Sediment with and without Organic Matter and Native Sorbates GERARD CORNELISSEN AND O ¨ RJAN GUSTAFSSON* Institute for Applied Environmental Research (ITM), Stockholm University, 10691 Stockholm, Sweden
Strong sorption to soot- and charcoal-like material (collectively termed black carbon or BC) in soils and sediments is possibly the reason for recent observations of elevated geosorbent-water distribution ratios, slow desorption, limited uptake, and restricted bioremediation. We evaluated the role of environmental BC in the sorption of phenanthrene (PHE) to a polluted lake sediment from a Rhine River sedimentation area. Sorption isotherms were determined over a wide concentration range (0.0005-6 µg/ L) for the original sediment (with organic matter or OM, native sorbates, and BC), sediment from which we had stripped >90% of the native sorbates (only OM and BC), and sediment combusted at 375 °C (only BC). The sorption isotherms of the original and stripped sediments were almost linear (Freundlich coefficient or nF > 0.9), whereas the isotherm of the BC remaining after the sediment combustion was highly nonlinear (nF ) 0.54). At low concentrations (ng/L range), PHE sorption to BC in the combusted sediment was found to exceed the total PHE sorption in the original and stripped sediments. This implies that it may not be possible to use a BC-water sorption coefficient measured in combusted sediment to estimate total sorption to the original sediment. This “intrinsic” BC-water sorption coefficient after combustion was calculated to be 9 times larger than the “environmental” one in the untreated sediment. Competition between the added PHE and the native PAHs and/or OM may explain this difference. It appears that, at low aqueous PHE concentrations (ng/L and below), BC is the most important geosorbent constituent with respect to sorption. At higher concentrations (µg/L), BC sorption sites become saturated and BC sorption is overwhelmed by sorption to the other OM constituents. Because sorption is a central process affecting contaminant behavior and ecotoxicity, understanding this process can strongly contribute to risk assessment and fate modeling.
Introduction The sorption of hydrophobic organic chemicals (HOCs) is an important process because it governs the fate, transport, and ecotoxicological risks of soil- and sediment-bound chemicals. Around 1980, it was discovered that the organic matter (OM) in soils and sediments was the principal factor controlling sorption of HOCs, and it was proposed to * Corresponding author phone: +46-8-6747317; fax: +46-86747638; e-mail:
[email protected]. 148
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 1, 2004
normalize sediment-water distribution ratios to the total organic carbon (TOC) content (1, 2). In the 1990s and 2000s, several findings such as nonlinear sorption isotherms (3-9), multiphasic desorption kinetics (10-12), and strongly elevated TOC-water distribution ratios (KTOC) in the field (1315) led to the suggestion of a dual-mode sorption concept (4, 10, 15). In this concept, the OM is regarded to be composed of two domains, one showing linear absorption and one showing nonlinear adsorption. The absorption domain has been proposed to consist of amorphous OM like humic/ fulvic substances and lignin (4), whereas harder, more condensed moieties such as coal and kerogen contribute to the adsorption domain (3, 4, 6, 7, 9). The results of previous research on sorption properties of various geosorbent constituents have been summarized in excellent reviews (16-18). One particularly strongly sorbing form of TOC is black carbon (BC) (14, 15, 19-25). BC is formed by combustion processes; the two main forms are soot (BC formed by condensation reactions) and charcoal (charred residues of the original fuel). Part of the BC in the environment stems from combustion of biomass (forest fires, residential wood burning); part of it is from fossil fuel combustion (traffic, industry, coal, oil) (26, 27). So far, sorption to BC has mainly been studied for pure soots and charcoals (19-24). This research has shown that sorption to BC can be exceptionally strong with BC-water distribution ratios (KBC) exceeding OC-water ones by a factor of 100 or more. Generally, BC contents are about 1-15% of TOC (28-30), so in several cases BC can be expected to contribute more strongly to overall sorption than all the other OM constituents. This also depends on the sorptive strength of the other OM constituents and the aqueous HOC concentration. In some studies (20, 21, 23, 24), it has been shown that planar compounds show a stronger affinity for sorption to BC than nonplanar ones with comparable hydrophobicity. There are only two studies that have so far performed sorption experiments to environmental BC in soils/sediments. In a study by Xiao et al. (31) it was shown that sorption to environmental BC/kerogen could be up to 10 times stronger than sorption to OC and that it was highly nonlinear: sorption coefficients decreased strongly with increasing concentration. However, only a combined BC/ kerogen isolate was studied. Accardi-Dey and Gschwend (25, 28) successfully applied a BC-inclusive Freundlich sorption model to express the sediment concentration (CS) of pyrene in a Boston Harbor sediment (µg/kg dry weight, dw) in an OC and a BC sorption term:
CS ) fOCKOCCW + fBCKF,BCC nWF
(1)
where fOC and fBC are the sediment mass fractions of OC and BC, respectively; KOC is the OC-water distribution coefficient (L/kg); CW is the aqueous concentration (µg/L); KF,BC is the Freundlich BC-water distribution ratio [(µg/kgBC)/(µg/L)n]; and nF is the Freundlich exponent of BC sorption. A concentration dependence appears in the BC term because of the nonlinearity of BC sorption. Accardi-Dey and Gschwend (25, 28) determined KF,BC in the following way. They removed the OM and native sorbates by combusting the sediment at 375 °C under air (15, 32) and determined KF,BC by measuring sorption isotherms in the combusted sediment that contained only minerals and BC. In the present study, the abovementioned model (eq 1) is further evaluated for environmental BC. It is investigated 10.1021/es034776m CCC: $27.50
2004 American Chemical Society Published on Web 11/18/2003
int whether the “intrinsic” KF,BC after combustion, KF,BC (for BC without OM and native sorbates), is representative of the “environmental” KF,BC in the original contaminated sediment, env KF,BC (for BC with OM and native sorbates). This was implicitly assumed in the Accardi-Dey and Gschwend studies (25, 28). In the original sediment, OM and native sorbates can potentially influence the sorption to the BC. In the present study, sorption isotherms were measured for added deuterated phenanthrene (PHE-d10) in a sediment from a PAHpolluted sedimentation area from the Rhine River. The influence of native sorbates on PHE-d10 sorption was studied by measuring PHE-d10 isotherms in “stripped” sediment where >90% of the native sorbates was removed by desorption under an air flow at 110 °C for 42 d. The sorption characteristics of the BC after combustion were studied by determining sorption isotherms of sediment that was combusted at 375 °C. In order not to overlook processes occurring at low but environmentally relevant concentrations (in the ng/L range), we determined isotherms over a wider concentration range (CW ) 0.0005-6 µg/L) than in previous BC sorption studies. The applied sorption method was the polyoxymethylene-solid-phase extraction method (POMSPE) recently developed by Jonker and Koelmans (21).
Methods Sediment: Location and Sampling. Sediment was sampled from Ketelmeer (KET), a lake in The Netherlands (52°36′ N, 5°45′ E). This freshwater lake is a sedimentation area of the Rhine River. The shores of this river are densely populated and heavily industrialized; in addition, large-scale coal mining activities used to take place. Therefore the sampling location can be considered representative of sedimentation areas of polluted rivers in many places of the world. The 10-50 cm layer was sampled in 2000 using a box corer. Sampling depth was 2.5 m. Dry weight (dw) of the homogenized sediment was determined after drying overnight (60 °C) to 47.2 ( 0.2% (n ) 6). Materials. Phenanthene-d10 (PHE-d10), 13C2-labeled PHE (both 99%), and solvents (Burdick and Jackson glass-distilled quality) were obtained from various commercial sources. Polyoxymethylene (POM) was obtained in 0.5 mm thick sheets from Vink Kunststoffen BV, The Netherlands. TOC and TON. TOC and total organic nitrogen (TON) contents were determined using a previously published procedure (15) in which the dried (60 °C) and ground sediment is first subjected to in situ microacidification (1 M HCl) and then analyzed with catalytic combustion elemental analysis at 1030 °C (Carlo Erba model 1106). SSA. Specific Surface Areas (SSA) of the original and combusted sediments were determined by N2 adsorption at -196 °C using an ASAP 2010 system (Micromeritics, Norcross, GA). Before SSA analysis, the original sediments were dried (60 °C), very carefully ground, and degassed (100 °C). With regular BET measurements, it is not possible to analyze nanoporosity (33, 34) in the range of 4-20 Å. Performance of nitrogen adsorption at low pressures (
