Environ. Sci. Technol. 2008, 42, 5897–5903
Compound-Specific Factors Influencing Sorption Nonlinearity in Natural Organic Matter S A T O S H I E N D O , * ,† P E T E R G R A T H W O H L , † STEFAN B. HADERLEIN,† AND T O R S T E N C . S C H M I D T †,‡ Center for Applied Geoscience (ZAG), Eberhard-Karls-University of Tu ¨ bingen, Sigwartstrasse 10, D-72076 Tu ¨ bingen, Germany, and Instrumental Analytical Chemistry, University of Duisburg-Essen, Lotharstrasse 1, D-47048 Duisburg, Germany
Received January 15, 2008. Revised manuscript received May 18, 2008. Accepted May 20, 2008.
Nonlinear sorption by natural organic matter may have a significant impact on the behavior of organic contaminants in soils and sediments. This study presents a molecular probe approach based on linear solvation energy relationships (LSERs) to identify and quantify the molecular interactions causing concentration-dependent sorption and proposes estimation methods for sorption nonlinearities. Sorption isotherms ranging over concentrations of more than 4 orders of magnitude were determined in batch systems for 23 and 16 chemically diverse probe compounds in a lignite sample and a peat soil, respectively. Each sorbent showed characteristic nonlinear sorption with Freundlich exponents (1/n) being 0.7-1. The LSERbased analysis revealed that the strength of nonspecific interactions did not vary with concentration for both sorbents. In lignite, specific interactions did not affect sorption nonlinearity either, suggesting that compound-independent factors of lignite were responsible for the nonlinear sorption. In the peat soil, by contrast, the specific interactions related to the solute polarizability/dipolarity parameter (S) decreased with increasing concentration. Consequently, compounds of higher S values were more susceptible to nonlinear sorption in the peat soil. Phenol probes have shown that hydrogen bond donating properties of sorbate compounds have a substantial impact on the overall strength of sorption with organic matter, but no significant influence on sorption nonlinearity. Heterocyclic aromatic compounds appear to undergo additional interactions that are not accounted for by the LSER. These additional interactions considerably enhance both sorption capacity and nonlinearity.
Introduction Sorption to solid matrices is one of the key processes that determine the fate of contaminants in the environment. In many cases sorption of nonionic organic compounds is dominated by sorption to organic matter (ref 1, p 108; ref 2, p 291), which has been frequently assumed to be linear. However, more and more studies have reported that sorption * Corresponding author phone: +49-7071-29-74693; fax: +49-70715059; e-mail:
[email protected]. † Eberhard-Karls-University of Tu ¨ bingen. ‡ University of Duisburg-Essen. 10.1021/es8001426 CCC: $40.75
Published on Web 07/09/2008
2008 American Chemical Society
in soils and sediments can be rather nonlinear and is stronger than expected from the linear assumption especially at low concentrations, even for nonpolar compounds (3–5). Enhanced sorption at low concentrations reduces bioavailability of contaminants in sediments and may limit bioaccumulation in biota (6). Nonlinear sorption in groundwater systems causes tailing of solutes and prolongs remediation times (7). Further, various degradation processes occurring in water are expected to be less effective at low concentrations due to decreased fractions of contaminants in the aqueous phase. Therefore, to elucidate the causes and mechanisms of nonlinear sorption and thereupon establish quantitative methods for estimating nonlinearities of various compounds is of great importance. Conceptual models proposed to explain the concentration-dependent nature of sorption can be largely sorted into two types. First, gel-like organic matter is assumed to absorb compounds and show strictly linear sorption, whereas relatively small quantities of carbonaceous materials (e.g., soot, charcoal) adsorb compounds, resulting in nonlinear sorption (8–10). Carbonaceous sorbents are generally composed of rigid, condensed aromatic structures and characterized by high surface areas and microporosities. Other models consider organic matter itself as a dual-sorbent, analogous to a glassy polymer, which contains flexible and rigid regions in its structure (11, 12). Flexible regions accommodate compounds similar to a dissolution (absorption) process, which leads to linear sorption, whereas rigid regions contain “nano-holes” having strong affinity for guest molecules but a limited capacity and are thus responsible for nonlinear sorption. Both types of models lead to similar mathematical equations (superposition of a linear and a nonlinear term), and can describe experimental nonlinear sorption data reasonably well. However, differences in nonlinearity among compounds of differing structures are not well understood and, consequently, so far no general method exists that predicts the extent of nonlinear sorption for different compounds. Note that inorganic components such as clay minerals are known to show strong nonlinear sorption with specific polar compounds (13), but this is beyond the scope of this paper as our study focuses on organic-rich sorbents. The purpose of this work is to present and evaluate a molecular probe approach based on the linear solvation energy relationships (LSERs, see the Materials and Methods section for details) to understand, quantitatively, the nonlinear sorption by organic matter in terms of sorbent-sorbate molecular interactions. To this end, sorption isotherms were experimentally determined for 23 probe compounds in a lignite sample and 16 compounds in a peat soil. Previous studies reporting sorption isotherms on soils were often limited to a relatively small number of compounds of environmental concern belonging to a single chemical class (e.g., hydrophobic compounds). This study considers a set of chemically diverse compounds as probes irrespective of their environmental significance to elucidate which compound-specific factors (e.g., polar interactions) determine sorption nonlinearities. Measured sorption coefficients for different compounds at varying concentrations are compared with LSER descriptors to assess molecular interactions between sorbate molecules and organic matter, which may change along with the increasing loading in the sorbent. The results are compared with literature data to validate our approach. Furthermore, estimation of sorption nonlinearities for other compounds and mechanistic implications for nonlinear sorption by organic matter are discussed. To our VOL. 42, NO. 16, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5897
TABLE 1. Characteristics of Studied Sorbentsa sorbent
abbreviation
Oberlausitz lignite Pahokee peat soil
BK-II PP
moisture contentb wt%
specific surface areac m2/g
carbon content wt%
nitrogen content wt%
black carbon content wt% of TOCd
11.1 ( 0.4 10.2 ( 0.2
3.1 1.3
53.5 46.1
0.6 3.3
1.9e 1.5f, 1.6g
a Values are for our pulverized materials, except for black carbon content for PP. b Weight loss of air-dry samples after 24 h at 105 °C (mean ( standard deviation, n ) 3). c N2-BET method. d TOC, total organic carbon. e Determined by a chemothermal oxidation method at 375 °C (CTO-375) according to ref 15. f From ref 16, CTO-375. g From ref 17, chemical oxidation method.
TABLE 2. Freundlich Parameters for Measured Sorption Isotherms on BK-II and PP BK-II (lignite) compound
abbreviation
n-pentane n-hexane n-heptane n-octane n-nonane cyclopentane cyclohexane cyclooctane isohexane di-n-propyl ether di-n-butyl ether benzene toluene n-propyl benzene trichloromethane tetrachloromethane trichloroethylene tetrachloroethylene 2-octanone hexanenitrile 1-nitrohexane 4-ethylphenol 2,6-dimethylphenol naphthalene a
(mg/kg)/(mg/L)1/n.
nPEN nHEX nHEP nOCT nNON cPEN cHEX cOCT iHEX DNPE DNBE BENZ TOL PrBZ TrCM TeCM TCE PCE OCTON HXNTRL NTRHX 4EtP 26DMP NAPH b
1/n
R2
Nb
log KFra
2.58 3.11 n.a. 4.04 4.50 2.22 2.53 3.29 2.92 1.80 2.62 1.97 2.42 3.14 1.77 2.29 2.35 2.76 2.63 1.99 2.77 2.53 2.46 3.32
0.923 0.929
0.9960 0.9980
44 44
0.903 0.876 0.947 0.931 0.946 0.939 0.883 0.846 0.899 0.865 0.850 0.922 0.916 0.882 0.878 0.776 0.845 0.818 0.879 0.876 0.816
0.9988 0.9985 0.9984 0.9986 0.9994 0.9992 0.9985 0.9976 0.9987 0.9975 0.9987 0.9992 0.9987 0.9980 0.9985 0.9973 0.9948 0.9969 0.9975 0.9958 0.9970
24 24 32 52 24 22 23 24 30 21 24 24 23 24 24 19 24 22 20 19 24
n.a. 2.45 2.96 3.49 4.01 n.a. 1.89 2.81 n.a. n.a. 1.72 n.a. 1.79 2.45 n.a. 1.59 n.a. 2.13 1.93 n.a. 2.02 2.02 1.83 2.72
Materials and Methods Materials. The lignite sample (BK-II) was collected from the Oberlausitz area in Saxony, Germany. Pahokee peat soil (PP) was purchased from the International Humic Substances Society. These humified materials were pulverized with a zirconium oxide planet ball mill (Laborette, Fritsch) to shorten the time necessary to attain sorption equilibrium. Particle size distribution analysis for BK-II using laser diffraction showed that 50% were
