JEQ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 31 May 2006
Published in J Environ Qual 35:1049-1059 (2006)
DOI: 10.2134/jeq2005.0362
© 2006 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Pignatello, J. J.
Right arrow Articles by Xing, B.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Pignatello, J. J.
Right arrow Articles by Xing, B.
Agricola
Right arrow Articles by Pignatello, J. J.
Right arrow Articles by Xing, B.
Related Collections
Right arrow Humic Substances
Right arrow Chlorinated Hydrocarbons
Right arrow Pesticides
Right arrow Organic Compounds
Right arrow Soil Pollution

TECHNICAL REPORTS

Organic Compounds in the Environment

Nonlinear and Competitive Sorption of Apolar Compounds in Black Carbon-Free Natural Organic Materials

Joseph J. Pignatelloa,*, Yuefeng Lua, Eugene J. LeBoeufb, Weilin Huangc, Jianzhong Songd and Baoshan Xinge

a Department of Soil and Water, Connecticut Agricultural Experiment Station, 123 Huntington Street, P.O. Box 1106, New Haven, CT 06504-1106
b Department of Civil and Environmental Engineering, Vanderbilt University, 400 24th Avenue South, VU Station B 351831, Nashville, TN 37235
c Department of Environmental Sciences, Cook College, Rutgers University, New Brunswick, NJ 08901-8551
d State Key Laboratory of Organic Geochemistry, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, Guangzhou 510640, The People's Republic of China
e Department of Plant, Soil and Insect Sciences, University of Massachusetts, Amherst, MA 01003

* Corresponding author (joseph.pignatello{at}po.state.ct.us)

Received for publication September 20, 2005.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 STRUCTURAL PROPERTIES OF SOIL...
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Numerous studies have reported a spectrum of sorption phenomena in soils, sediments, and organic matter isolates of those materials that are inconsistent with a partition model proposed in the late 1970s and early 1980s, a model predicated on a hypothesis that sorption is linear and noncompetitive. To explain these nonideal phenomena, prior studies have proposed a hard–soft (glassy–rubbery) model for SOM (soil and sediment organic matter), while others have attributed them singularly to BC (black carbon: soot and charcoal) particles present in topsoils and sediments. In this study, we demonstrated nonideal sorption behavior (isotherm nonlinearity, competitive effects) for a group of apolar compounds in a large set of natural and model organic materials, including a commercial lignin and humic acids from different sources. Complete oxidation of samples by an acidic dichromate method was taken to signify the absence of BC. (However, polymethylene units are stable even if functionalized on both ends, making the technique unreliable for quantifying BC.) Other samples were inferred free of BC by their source and method of preparation. Characterization by thermalanalytical methods indicated the glassy character of the organic materials. The origin of the nonideal behaviors appears to be the glassy character of these materials. Sorption nonlinearity increased or decreased by changing temperature, cosolvent content, or degree of cross-linking by metal ions as predicted for organic solids in a glassy state. We conclude that macromolecular humic substances in the environment may exhibit nonideal sorption behavior in soils and sediments, quite apart from any such behaviors attributable to BC.

Abbreviations: BC, black carbon • DCB, dichlorobenzene • HA, humic acid • HOC, hydrophobic organic contaminant • NAPH, naphthalene • OC, organic carbon • PHEN, phenanthrene • SOM, soil and sediment organic matter • TCB, trichlorobenzene • TMDSC, temperature-modulated differential scanning calorimetry • XYL, xylene


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
STRUCTURAL PROPERTIES OF SOIL...
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
SORPTION by soils and sediments plays an essential role in the transport, bioavailability, and the associated risks of organic contaminants in the environment. Despite extensive studies during the last two decades, investigators are still searching for consensus on a comprehensive mechanistic concept for quantitative description and scientific interpretation of this important process. It has been unambiguously shown that the SOM fraction is the dominant sorbent for nonionic HOCs (hydrophobic organic contaminants), except for soils and sediments having very low OC (organic C) content, very low moisture content, or relatively high clay content with available hydrophobic surface area.

An early concept for HOC sorption to SOM that has achieved widespread acceptance attributes sorption to homogeneous dissolution (partition) into a gel-like SOM phase, similar to dissolution in an organic liquid (Chiou, 2002). In apparent agreement with data published at the time, this partition model predicted that sorption of nonpolar and weakly polar ("apolar") compounds would generate a linear isotherm and be noncompetitive when more than one sorbing chemical is present. Many studies have since demonstrated, however, that sorption of apolar HOCs to SOM in bulk soils or soil derivatives is nonideal, i.e., it displays nonlinear and competitive behavior. A rubbery–glassy polymer model was proposed to explain the dual-sorbent nature of SOM, labeled "dual-mode" or "dual reactive domain" models (e.g., Pignatello and Xing, 1996; Chefetz et al., 2000; Huang et al., 1997). In these models, it was postulated that there are two SOM domains: one made up of loosely knit (flexible) macromolecules, where sorption is akin to "solid-phase dissolution" and is linear and noncompetitive, and the other made up of tightly knit (rigid) macromolecules, where sorption occurs by both dissolution and "adsorption" or "hole-filling" mechanisms with nonlinear and competitive behavior. Dual-mode sorption behavior is characteristic of glassy organic polymers (Xing and Pignatello, 1997; LeBoeuf and Weber, 2000a, 2000b, and references therein). Humic substances from several sources exhibit glass transition temperatures (discussed below) and dual domain flexibilities (Mao et al., 2002; Gunasekara et al., 2003). In addition, hysteresis and other irreversible phenomena may result from changing flexibility of the glassy humic matrix as a function of sorbate concentration (Lu and Pignatello, 2004b; Sander and Pignatello, 2005; Sander et al., 2006).

Around the same time that the dual models were being developed, findings began to appear of the potential importance of soot and charcoal particles (environmental BC) in sorption of contaminants to soils and sediments (Maruya et al., 1996; Gustafsson and Gschwend, 1997). It was suggested that, since BC materials have high surface areas and microporosities which—like activated carbons—provide adsorption sites, their presence could give rise to nonlinear and competitive behavior so often found in bulk soils (Chiou and Kile, 1998; Xia and Ball, 2000). Investigators have recently attempted to model sorption isotherms of polycyclic aromatic hydrocarbons (PAHs) and other large molecules in terms of two- or three-term equations. In these models, all of the observed nonlinearity is ascribed to the BC (Accardi-Dey and Gschwend, 2003; Lohmann et al., 2005), or BC and coal C (Cornelissen and Gustafsson, 2005) terms, and the term representing "ordinary" SOM is assumed linear. Ordinary SOM is usually taken to be "amorphous" humic and biological remnant substances. While these models can be made to work, the assumption on which they rest—that of attributing all nonlinearity to BC—is questionable.

The purpose of the this study was to document nonlinear and competitive sorption in macromolecular natural organic substances including humic acids and a wood lignin free of BC. We also present thermal analytical data consistent with the glassy character of these materials. Lignin is included because biopolymers derived from lignin are believed to be the main precursors of terrestrial and estuarine humic substances, as the lignin signature is abundantly clear in their structures.


   STRUCTURAL PROPERTIES OF SOIL ORGANIC MATTER
TOP
 ABSTRACT
 INTRODUCTION
 STRUCTURAL PROPERTIES OF SOIL...
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The complex nature of SOM is well known. Soil and sediment organic matterr may contain macromolecular substances (biopolymers, fulvic acids, humic acids, humin, kerogen-based coal), and elemental carbonaceous substances (i.e., BC), in proportions that may vary from site to site and from sample to sample within a site (Song et al., 2002). While SOM components may sorb HOCs differently, standard methods for their quantification and separation are currently not available.

Lignin is an aromatic, high molecular weight, macromolecular solid second to cellulose in abundance among plant-derived "polymers" (Glasser and Kelley, 1987). Although the structure depends on whether the lignin is from nonwoody plants, softwoods, or hardwoods (Brunow, 2001), all lignins derive from the phenylpropane unit, which may be substituted with different functional groups.

Humic substances, which include humin, humic acid, and fulvic acid, are usually the most abundant components of SOM. Humic acids weigh hundreds to thousands of daltons, are rich in carboxyl and phenolic groups, and have O/C ratios of 1 or greater (Hedges, 1988). Structural models vary considerably, but in general feature a C–C skeleton with COOH, ketone, and aliphatic and aromatic OH functional groups (Schulten and Schnitzer, 1997). Two-dimensional 1H nuclear Overhauser effect spectroscopy reveals pH-dependent conformational changes consistent with a chainlike primary structure for humic aliphatic structures, with humic aromatic regions exhibiting greater rigidity than aliphatic regions (Chien and Bleam, 1998). Recently, Mao et al. (2002) observed shifts in SOM domain mobility among a number of humic acid samples using solid-state NMR (nuclear magnetic resonance).

Kerogen is the dominant fraction of sedimentary organic matter that is insoluble in nonpolar or weakly polar organic solvents and is resistant to oxidizing acids. Coals, comprised primarily by Type III kerogen, are rich in condensed, crosslinked (usually by carboxyl and carbonyl groups) aromatic structures and, in more diagenetically advanced coals, show evidence of microcrystalline regions (Oberlin et al., 1980). A variable-temperature 1H NMR study of pyridine in Blind Canyon coal suggests a rigid macromolecular structure stabilized by interchain and intrachain crosslinks (Xiong and Maciel, 2002). Compared with humic acids, coal macromolecules generally have much higher molecular weights, often exceeding several hundred thousand daltons.

Glass-to-Rubber Transitions
Increasing temperature increases the motions of individual macromolecular chains, leading to expansion of free volume. If sufficient thermal energy is provided, individual chains may possess enough kinetic energy to break noncovalent interchain bonds, providing sufficient free volume to transcend from a more rigid, glass-like state to a more fluid, rubber-like state (McKenna, 1989). The transformation from glassy to rubbery states is complete at the glass transition temperature, Tg. While the Tg is often reported as a distinct temperature, in reality the heterogeneity of a macromolecule (e.g., molecular weight distribution, chemical composition) often provides a broadening of the glass transition over tens of degrees—the more heterogeneous the matrix, the larger the expected broadening (McKenna, 1989).

Evidence for glass transition behavior in a variety of SOMs is provided in the literature. The Tg of lignins varies from 66 to 190°C (Kadla et al., 2003), depending on their origin and isolation procedure. Glass transition behavior has been observed for a number of humic acids of terrestrial and aquatic origin, including two used in this study, AldHA (Aldrich humic acid; LeBoeuf and Weber, 1997) and PHA (peat humic acid; Young and LeBoeuf, 2000). Other humic materials exhibiting thermal transition behavior include Leonardite humic acid (LeBoeuf and Weber, 2001), Nordic aquatic humic acid (DeLapp and LeBoeuf, 2004), Georgetown aquatic humic acid, and Harpeth soil humic acid (DeLapp et al., 2005). Additional work by Schaumann and Antelmann (2000), DeLapp and LeBoeuf (2004), and Schaumann and LeBoeuf (2005) provide evidence that this behavior may also be prevalent in whole soils.

Glass transition behavior has been reported for numerous coals at temperatures ranging from 99 to 125°C (MacKinnon and Hall, 1995) for the lower temperature transitions, and to 307 to 359°C for higher temperature transitions (Lucht et al., 1987). Lucht et al. (1987) correlated the glass transition temperature with increasing degree of diagenetic alteration, or rank of the coal, observing that increased aromaticity of the coal corresponded with increased Tg values.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 STRUCTURAL PROPERTIES OF SOIL...
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Reagents and Sorbents
1,2,4-Trichlorobenzene (1,2,4-TCB), 1,3,5-TCB, 1,3-dichlorobenzene (1,3-DCB), 1,2-DCB, naphthalene (NAPH), phenanthrene (PHEN), o-xylene (o-XYL) and Organosolv lignin (OS-lig) were obtained in the highest purity available from Aldrich Chemical Co. (Milwaukee, WI). In some cases, 14C-labeled compounds, obtained from Aldrich-Sigma Chemical Co. (St. Louis, MO), were used. Other chemicals and solvents were reagent grade or better.

Properties of sorbents are given in Table 1. Organic C was determined in all cases by CO2 quantification on combustion of the sample. Humic acids were prepared following the standard procedures of the International Humic Substances Society (IHSS; Swift, 1996). Chelsea humic acid (CSHA) was extracted from a topsoil (Oxyaquic Hapludalf) collected from Chelsea, MI, following Swift (1996). Humic acids extracted from Florida peat soil (Lithic Medisaprist) (FlaHA), Amherst peat soil (Terric Haplosaprist; APHA), a mineral soil (Oxyaquic Dystrudept; MinHA), and a compost material (CompHA) were described earlier (Gunasekara et al., 2003; Wang and Xing, 2005; Yuan and Xing, 2001). A humic acid (H-HA) prepared from a different batch of Amherst peat soil is described in Lu and Pignatello (2004a); the alkaline soil extract was treated with a cation exchange resin to remove metal ions, precipitated, and de-ashed with HCl–HF solution. The final product H-HA contained a small amount of base-insoluble residue (see below).


View this table:
[in this window]
[in a new window]
  		Table 1. Sorbent properties.

 
Purified AldHA (Aldrich Chemical Co., Lot no. 01816HH) was obtained from Dr. Tanju Karanfil of Clemson University, Clemson, SC. Purification entailed 0.1 M NaOH extraction under N2 followed by repeated cycles of 0.1 M HCl precipitation, base dissolution, and centrifugation. The purified humic acid was then freeze-dried and stored under N2. Upon receipt, it was crushed and sieved to retain the 38- to 180-µm size fraction. The PHA (reference grade) was obtained from IHSS and was used as received.

Two Al3+-exchanged humic acids were prepared. The MinHA(Al) was prepared by saturating solid particles of MinHA with 0.1 M AlCl3 solution (Yuan and Xing, 2001). The Al-HA was obtained by flocculating redissolved H-HA with Al3+ at pH 6 as described previously (Lu and Pignatello, 2004a); the batch prepared for this study had a slightly different elemental composition.

Acid Dichromate Treatment
The acid dichromate wet chemical oxidation method (Song et al., 2002) was used to detect the possible presence of BC by selectively oxidizing humic substances. In preparation, the samples were ground, demineralized at 60°C with a mixture of HCl (6 M) and HF (22 M) for 20 h, and finally rinsed three times with 6 M HCl followed by Milli-Q water (Millipore Corp., Billerica, MA). The OS-lig was presumed to be mineral free and was not demineralized. Additionally, for H-HA, a subsample was dissolved in a minimal amount of NaOH and centrifuged at 8000g; acid dichromate treatment was performed on the centrifugate, which, after rinsing with water and drying, corresponded to 0.05 kg/kg of H-HA, and the results were related back to the amount of the original H-HA.

Standard procedure involved suspending a 50- to 75-mg sample in 40 mL of a solution of 0.1 M K2Cr2O7 in 2 M H2SO4 at 55 ± 1°C for 60 h. After oxidation, the residual solid, if any, was collected by centrifugation, dried, weighed, and analyzed for total OC content by CO2 evolution on combustion. The acid dichromate treatment residual C content of each sorbent was then calculated from this weight and the total OC content of the original solid.

For comparison, a number of reference materials were also treated. These included a prepared wood char, paraffin flakes (Aldrich no. 411671), polyethylene microspheres (Aldrich, medium density, <149-µm diam.), n-hexadecanoic acid (palmitic acid, Sigma, 99%), n-octanoic acid (Acros, 99.5%), and decanedioc acid (sebacic acid, Aldrich, 99%). The char was prepared by muffling maple wood shavings at 400°C for 2 h in a watch-glass-covered crucible; it is characterized by Braida et al. (2003). Decomposition was monitored on replicate samples by the methods below, which were validated on controls without dichromate. Decomposition of paraffin wax and polyethylene was monitored by removing acid dichromate solution, rinsing the residual solid with water, and determining the weight of the solid after drying at 55°C. For decanedioic acid, the solution was refrigerated overnight to induce precipitation and the solid was collected and then dried in an ice bath with a stream of N2. The result was corrected for recovery (90 ± 2%) determined with controls. Decomposition of hexadecanoic and octanoic acids was monitored by extracting the solution with hexanes, passing the organic phase through a bed of Na2SO4 to remove water, evaporating the organic phase to dryness in a stream of N2, and weighing. Decomposition of char was monitored by its loss of C (Galbraith Laboratories, Knoxville, TN).

Thermal Analysis
Thermal analysis included the use of differential scanning calorimetry (DSC) and TMDSC (temperature-modulated DSC). Before analysis, each sample was heated to remove physisorbed water. Samples were subjected to multiple heating–cooling cycles to ensure repeatability of individual scans. Detailed instrumentation descriptions and experimental protocols are given in DeLapp and LeBoeuf (2004).

Sorption Methodologies
Customary shake-flask batch techniques were employed as described previously: AldHA, PHA, and OS-lig in LeBoeuf and Weber (2000b); H-HA and Al-HA in Lu and Pignatello (2004a, 2004b); MinHA, MinHA(Al), CompHA, and OS-lig in Xing and Pignatello (1997); and CSHA in Weber and Huang (1996). These techniques varied slightly among laboratories and can be summarized as follows. The solid was suspended in pH-adjusted aqueous solution containing appropriate electrolyte and bio-inhibitor (100–200 mg/L NaN3 or HgCl2). The solid/liquid ratio was selected in each case to achieve 30 to 70% of uptake of the added solute. The solute was added either dissolved in water or methanol (<0.5% of solution volume). The vessel (which was filled to capacity) was then sealed and agitated gently for a time appropriate to reach apparent equilibrium (usually 7–14 d), as determined in preliminary experiments. Slight departure from complete equilibrium would have negligible effect on the degree of nonlinearity (Weber and Huang, 1996) and, we reasonably assume, of competitiveness. The liquid and solid phases were then separated (usually by centrifugation) and an aliquot of the liquid was withdrawn for analysis by gas chromatography, liquid chromatography, or liquid scintillation counting, as appropriate.

Sorption Model
Sorption equilibrium was modeled by the Freundlich equation:

Formula 1[1]
or

Formula 2[2]
where qe (mg/kg) and Ce (mg/L) are the apparent equilibrium solid-phase and solution-phase concentrations, respectively; KF [(mg/kg)(mg/L)–n] is the Freundlich affinity parameter; and n is the isotherm linearity index. The Freundlich equation is often considered empirical, but it can be derived thermodynamically (Adamson and Gast, 1997) and can be demonstrated graphically to result from sorption to a limited number of surfaces, each having its own energy and limited adsorption capacity quantified by the Langmuir equation. In general, n < 1 suggests that, with increasing concentration, sorption sites exhibit weaker affinity for sorbing molecules. The Freundlich n may, thus, be taken as an index of the degree of linearity. When n = 1, Eq. [1] is the well-known linear partition model. From a practical perspective, any isotherm with 1.05 > n > 0.95 could be considered linear; however, n is sensitive to the experimental concentration range and so should be interpreted with caution if that range varies appreciably for a given compound compared in different sorbents or for different compounds compared in the same sorbent (Xia and Pignatello, 2001).

For data reduction, we employed the linear regression of Eq. [2], rather than the nonlinear regression of Eq. [1]. The advantages of using Eq. [2] for modeling sorption data across a wide range in concentration have been discussed elsewhere (Huang et al., 1997). The underlying assumption of the linear regression is that Ce and qe are independently measured. When qe is not measured independently, this can be adequately satisfied if the mass left in solution at equilibrium is between 30 and 70% of the mass added, a condition for which a small variation of Ce has little effect on qe.

The concentration-dependent OC-normalized distribution coefficient KOC is calculated from the fractional OC content, fOC (kg/kg), and the Freundlich parameters by:

Formula 3[3]
The KOC defined in Eq. [3] is equivalent to a single-point OC- normalized distribution coefficient determined at a particular concentration, Ce, and is not the same as the slope of the linearized isotherm, as it is commonly defined in the literature.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 STRUCTURAL PROPERTIES OF SOIL...
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Characterization of Sorbents
The sorbents used in this study had little mineral matter. In addition, they were expected to be free of, or to contain only small amounts of, BC due to their sources and methods of preparation. Since CompHA was extracted from mushroom compost, whose source material is known to be free of char (Chen et al., 2000), it is reasonable to assume that CompHA is free of BC. Likewise, due to its source, OS-lig was expected to be free of BC. Reliable analytical techniques for quantifying low concentrations of BC against a high background of humic substances do not yet exist. The widely employed thermochemical oxidation method, in which the sample is combusted at 375°C (Gustafsson et al., 1997; 2001), destroys nearly all char and much of the soot (Nguyen et al., 2004). Acid dichromate oxidation, originally developed by Wolbach and Anders (1989) for separating kerogen from elemental C in sedimentary rocks, has been used to quantify BC in soils (Song et al., 2002). This method rests on the supposed ability of acid dichromate to oxidize humic substances much more rapidly than elemental C.

Table 2 confirms previous reports (Masiello et al., 2002; Wolbach and Anders, 1989) that char is fairly stable to acid dichromate; only ~17% of C of a prepared wood char (Braida et al., 2003) was oxidized during the standard 60-h treatment period, increasing to about 63% after an extended period of 23 d. Table 1 shows that OS-lig, APHA, and CSHA contained no C that survived the standard 60-h treatment, and that two additional humic acids, FlaHA and PHA, left only a trace of flocculant residual material after this time. These sorbents were, thus, essentially free of BC. Table 1 indicates that two other humic acids, MinHA and H-HA, left 4.7% and 1.0 to 3.3% unoxidized C, respectively, after standard treatment.


View this table:
[in this window]
[in a new window]
  		Table 2. Results of acid dichromate treatment of reference materials.

 
The residual C of H-HA is probably not BC, however. The base-insoluble fraction of H-HA, expected to be highly enriched in BC if any were present, contained 18.8% C. Sorption of 1,2,4-TCB to the base-insoluble fraction was only about a factor of two greater than sorption to the whole H-HA (KOC equal to 6.2 x 103 L/kg OC vs. 2.9 x 103 L/kg OC at 1 x 10–7 mol/L 1,2,4-TCB; Lu and Pignatello, 2004b), but more than two orders of magnitude lower than sorption to the wood char (KOC = ~1.3 x 106 L/kg OC) at the same concentration (Zhu and Pignatello, 2005). Acid dichromate treatment of the base-insoluble fraction of H-HA oxidized away about half the C within 60 h and the rest remained stable to the reagent during the next 100 h. The final residue had a C content of 17%, yet was pale gray in color, inconsistent with the color expected of a material having a char content of ~25%. Rather, we believe that the residual C of H-HA (and possibly MinHA, as well) is made up of aliphatic hydrocarbon material present in the original sample. This is supported by the data in Table 2, which show that oxidation of aliphatic chains—either unfunctionalized (paraffin wax, polyethylene microspheres), monofunctionalized (n-octanoic acid, n-hexadecanoic acid), or difunctionalized (decanedioic acid)—is extremely slow in acid dichromate. This is true even when the compound is melted (as was octanoic acid) or soluble (as was decanedioic acid) in the hot acid dichromate mixture. This should be taken as a cautionary note when applying the acid dichromate method to natural solids, since polymethylene chains originating from cutins, cutans, suberans, and related substances are common components of humic substances (Kang and Xing, 2005; Almendros et al., 1996). No attempt was made to remove aliphatic components from the original HA samples.

Figure 1 provides TMDSC scans for MinHA and CSHA that indicate thermal transitions. The MinHA exhibits transitions near 45 and 69°C that consolidate to a single transition near 52°C during a subsequent heating cycle. The CSHA exhibits a low-temperature transition between 3 and 6°C and a second transition between 63 and 67°C. Presence of multiple transition temperatures is not uncommon in macromolecules, and can be explained in terms of either (i) the presence of ß- and {alpha}-transitions, where the lower temperature ß-transitions are associated with side-chain mobility and the higher temperature {alpha}-transitions are attributed to mobilization of the main chain associated with the Tg; or (ii) the presence of regions having different chemical and physical properties, thus manifesting different mobilities and glass transition temperatures, as observed in inhomogeneous solids such as block copolymers. Also note that these transition behaviors apply to the dry solid. Water-wet glass transition behavior for AldHA was observed near 43°C (LeBoeuf and Weber, 1997). Transitions for PHA were observed at 62°C (Young and LeBoeuf, 2000). The TMDSC transitions were observed for H-HA at 56 or 61°C (duplicate measurements) and for Al-HA at 59 or 62°C (duplicates). The OS-lig exhibits glass transition behavior near 70°C (LeBoeuf and Weber, 2001), although a temperature near 115°C was reported by Glasser et al. (1998).


Figure 1
View larger version (25K):
[in this window]
[in a new window]
  		Fig. 1. Temperature-modulated differential scanning calorimetry traces of (a) Amherst soil humic acid and (b) Chelsea soil humic acid; T1 is thermal transition temperature 1; {Delta}Cp is the change in specific heat capacity.

 
Isotherms under "Baseline" Conditions
Baseline in this context refers to a temperature of 20 to 25°C and reactions performed in water in the absence of cosolvents, competing solutes, or cross-linking metal ions. Data collected under baseline conditions and presented in Table 3 and in subsequent figures indicate nonlinear sorption behavior for five different compounds in 10 different humic acids and a lignin, with n values ranging from 0.791 to 0.986. The precision of these values is high—the standard error is typically better than ±0.01. As will be discussed below, the degree of linearity tends to respond predictably to changes in conditions that alter the physical glassy–rubbery state of the solid.


View this table:
[in this window]
[in a new window]
  		Table 3. Isotherm parameters.

 
Sorption of NAPH and PHEN in OS-lig under baseline conditions is nonlinear (Fig. 2, Table 3). The value of n for NAPH is 0.931 ± 0.006 and the value of n for PHEN is 0.80 ± 0.05. The lignin literature is mixed on the linearity question. Severtson and Banerjee (1996) observed linear isotherms for several chlorinated phenols in softwood pulp, although Xing et al. (1994) observed nonlinear sorption for phenol in two hardwood lignins. Linear isotherms were reported for sorption of trichloroethene and toluene to a pine lignin (Garbarini and Lion, 1986). A degraded Douglas fir lignin gave a highly nonlinear (n = 0.63) isotherm for pyrene (Greenwood et al., 2000; Chefetz et al., 2000). MacKay and Gschwend (2000) studied sorption of various monoaromatic hydrocarbons in water-saturated samples of Ponderosa pine and Douglas fir. They suggested that the roughly linear sorption (average and standard deviation of n for six cases = 0.96 ± 0.12) resulted mainly from solute partitioning in lignin components of the wood; however, their isotherms contained few data (five) and covered only one order of magnitude range in concentration.


Figure 2
View larger version (14K):
[in this window]
[in a new window]
  		Fig. 2. Isotherms of phenanthrene and naphthalene on Organosolve lignin (OS-lig).

 
Humic acids also typically exhibit nonlinear isotherms under baseline conditions (Table 2; examples are shown in Fig. 3–4). The value of n ranges from 0.791 (PHEN in PHA) to 0.986 (1,3,5-TCB in CSHA). Of 14 solute–sorbent combinations, eight give n < 0.9. Of particular significance is the finding that n is <1 for HOCs in HAs known or, inferred by their source, to be free of BC: 0.883 for PHEN in APHA; 0.845 for PHEN in CSHA (pH 2); 0.916 for PHEN in CSHA (pH 7); 0.930 for NAPH in CSHA; 0.867 for o-XYL in CSHA; 0.791 for PHEN in PHA; 0.939 for 1,2,4-TCB in H-HA; 0.877 for NAPH in H-HA; and 0.91 for PHEN in CompHA. The nonlinearity of PHEN in CompHA (shown in Fig. 4) is especially important because this material was created without soil in the absence of combustion activities. Figure 3 presents isotherms for PHEN in CSHA at pH 2 and 7. It shows that both isotherms are nonlinear, but the one measured at pH 2 is more nonlinear and reflects greater affinity of the solute for the solid. The difference may result from the so-called "third-phase effect": partitioning of PHEN in "dissolved" HA at pH 7 (as indicated by noticeably yellowish supernatant), which would result in higher liquid-phase concentrations.


Figure 3
View larger version (21K):
[in this window]
[in a new window]
  		Fig. 3. Sorption of five different compounds by Chelsea humic acid (CSHA); DCB is dicholorobenzene and TCB is trichlorobenzene.

 

Figure 4
View larger version (14K):
[in this window]
[in a new window]
  		Fig. 4. Phenanthrene sorption in compost humic acid (compost HA) and ultrafiltered Aldrich humic acid (AldHA).

 
The value of KOC (Table 3) decreases with increasing concentration at a rate that depends on n (see Eq. [3]) and that varies with the properties of both sorbent and solute. For example, in the slightly nonlinear case of NAPH sorption to CSHA (n = 0.930), KOC increases 38% overall, from 556 to 653 and to 767 as Ce decreases from 10 to 1 and to 0.1 mg/L, respectively; whereas, in the more nonlinear case of PHEN sorption to PHA, KOC increases 262% across the same concentration range. The variation in KOC is greater among different HA solids. Among the five HA solids, KOC for NAPH ranges up to a factor of almost 3, from 562 to 1582 at Ce = 0.1 mg/L and from 397 to 902 at Ce = 10 mg/L. The H-HA has a KOC for NAPH about two times greater than that for CSHA and MinHA evaluated at the same Ce. The Al form of HA has a relatively greater concentration-dependent KOC than the H form of the same sorbate; this is due to increased rigidity of the organic matter caused by Al ion cross-linking (see below).

Sorption affinity as log KOC calculated at Ce = 1 mg/L (Table 3) correlates with the hydrophobic character of the solute, expressed either as octanol–water partition coefficient, log KOW, (log KOC = 1.09 log KOW – 0.89 [r2 = 0.904]) or inverse water solubility, log Sw (log KOC = – 0.872 log SW – 0.89 [r2 = 0.929]). This is generally consistent with many prior studies of HOC sorption to natural solids (e.g., Chiou, 2002); however, this correlation by itself says little about the mechanism, since KOC–KOW or KOC–SW linear free energy relationships can be expected for both absorption and adsorption processes that are driven predominantly by weak nonspecific interactions and water-exclusion effects. The effect of solute structure on the degree of linearity is unclear from the data presented in Table 3. While the isotherms of PHEN—the largest and most hydrophobic compound among the HOCs tested—appear to be the most nonlinear, the trend of increasing nonlinearity with molecular size (or KOW) is not evident when sorption of a series of HOCs to the same solid is compared: on CSHA, n = 0.845 for PHEN, 0.930 for NAPH, 0.986 for 1,3,5-TCB, and 0.867 for o-XYL. Evidently, the relationship between molecular size and n is complex.

The large variations in n and KOC for the same HOC indicate that HAs in different soils have different sorption properties, and thus they should not be regarded as a simple gel-like phase. It is known that the heterogeneous properties of HA are functions of source material, biogeochemical alteration history, and pedologic and geographic conditions. Furthermore, sorption capacities and nonlinearity are affected by the contents of aliphatic-C, aromatic-C, and O functional groups, the O/C atomic ratio, the size distribution of the macromolecules, glass transition temperature, and adsorbed minerals. For example, isotherm nonlinearity was reported to increase with aromaticity of several HAs extracted from different depths of a soil (Xing, 2001). Also, adsorption of HA by minerals can cause nonlinear and competitive sorption, particularly at low HA loadings (Murphy et al., 1994; Wang and Xing, 2005).

Influence on Sorption of Experimental Conditions that Alter Sorbent Properties
Temperature
Temperature affects sorption behavior as a result of increasing macromolecular mobility, even below the Tg. Thus, sorption typically becomes more linear with temperature approaching the Tg as the solid becomes more rubbery, and as solid-phase dissolution increases in importance compared with hole filling. Evidence for this trend has been reported for synthetic organic polymers (Vrentas and Vrentas, 1996; LeBoeuf and Weber, 2000a, and unpublished data, 2006), as well as SOM samples with known Tgs (LeBoeuf and Weber, 1997, 2000a). A trend of increasing linearity with temperature for 1,3-DCB and metolachlor sorption in the high-organic-matter (~93%) Pahokee soil was reported previously (Xing and Pignatello, 1997). On the other hand, sorption nonlinearity with respect to Illinois no. 6 and Wyoming coals was little influenced by temperature between 5 and 45°C due to the high Tg of the coals (LeBoeuf and Weber, unpublished data, 2006).

The thermal analysis summarized above indicate that, at and below room temperature, the HA and OS-lig sorbents probably exist in a glassy state, but they progress toward more a gel-like state with increasing temperature. Consistent with this, Fig. 5 shows a trend of increasing n with temperature for PHEN in CSHA, AldHA, PHA, and OS-lig and for NAPH in OS-lig. The results for the Pahokee soil (Xing and Pignatello, 1997) are also plotted for comparison. At temperatures well below the Tg, OS-lig gives clearly nonlinear behavior (n = 0.678 at 5°C for PHEN and 0.931 at 23°C for NAPH). For PHEN in OS-lig, the data support no significant change in sorption capacity and nonlinearity of the isotherm between 5 and 45°C; however, for NAPH in OS-lig, at temperatures closer to the Tg of 70°C, there is a trend toward increased linearity (n = 0.987 at 60°C and 0.991 at 85°C), suggesting increasing mobility of the lignin matrix near the Tg. Sorption of PHEN by CSHA and PHEN by PHA clearly show increasing linearity with temperature. For PHEN in AldHA, there is an increase in sorption linearity between 5 and 25, and 5 and 45°C, but a decrease between 25 and 45°C. Since the water-wet glass transition of AldHA was observed near 43°C, one would expect sorption of PHEN to be linear at 45°C, yet the value of n is 0.918 at that temperature. We suggest that additional, localized regions within the heterogeneous AldHA matrix that have yet to be fully mobilized may continue to influence sorption behavior (i.e., the glass transition may not be fully complete at 43°C).


Figure 5
View larger version (28K):
[in this window]
[in a new window]
  		Fig. 5. Trend of increasing isotherm linearity (Freundlich exponent) with temperature for lignin, three humic acids, and a high-organic-content soil. The soil data are taken from Xing and Pignatello (1997).

 
In most cases, OS-lig and the HA sorbents show decreasing KOC with increasing temperature, especially at the lowest Ce where KOC was calculated (Table 3). This is consistent with exothermic sorption, and possibly indicates a reduction in the number of higher energy sites as the matrix progresses toward the Tg, increasing solubility of the HOC in water with temperature, or both.

Organic Cosolvent Addition
Similar to energy input in the form of heat to a macromolecular solid, energy input in the form of sorbate chemical potential may induce matrix expansion, promote macromolecular flexibility, and reduce the energy required for glass-to-rubber conversion (McKenna, 1989). Consistent with this is the reduction in the Tg of AldHA from 70 to 43°C due to the swelling of the matrix by water (LeBoeuf and Weber, 1997). Plasticization was also observed for a peat soil at high water content (Schaumann and LeBoeuf, 2005). (Below 12% water, however, antiplasticization is observed due to H-bond cross-linking between water molecules and soil that is disrupted at higher water content.)

Sorption of NAPH to OS-lig from solutions containing methanol cosolvent represents another instance of the effects of solvent permeation (Table 2). The linearity of the NAPH isotherm at 23°C is clearly influenced by the volumetric fraction of cosolvent, changing from n = 0.931, to 0.959, to 1.000 at methanol contents of 0, 5, and 20% , respectively. The KOC of NAPH is simultaneously reduced. This suggests that methanol progressively lowers the Tg to a temperature near or below the temperature of the experiment. Similar results were reported in earlier work for sorption of HOCs by soils (Xing and Pignatello, 1997; Bouchard, 2002) and by a soil humin (Gunasekara and Xing, 2003).

Polyvalent Metal Ion Cross Linking
Isotherms were constructed in humic acids exchanged with Al3+ ion. Aluminum ions are known to be effective for coagulating humic substances. Coagulation occurs by bridging functional groups (particularly, carboxyl) located on the same or different macromolecules. The resulting cross-linked structure is expected to be less flexible at the strand scale, and therefore to generate isotherms with lower n values. The MinHA(Al) was prepared by saturating MinHA particles with AlCl3 solution (pH ~ 3.5). The Al-HA was prepared by flocculating H-HA with Al3+ ion (Lu and Pignatello, 2004a), yielding a material in which Al ions preferentially coordinate to organic functional groups rather than incorporate in Al oxide polymers (Masion et al., 2000). Isotherms in both MinHA(Al) and Al-HA are more nonlinear than in the corresponding Ca2+- or H+-saturated form, respectively (Table 3). The link between glassiness of HA promoted by metal-ion cross-linking and enhanced nonlinearity is important because it supports the polymeric nature of HA and points away from BC as the source of nonlinearity in the original HA. By saturating HA with Al3+ ions, isotherm hysteresis is also increased (Yuan and Xing, 2001; Lu and Pignatello, 2004b), consistent with the more condensed form of Al-HA. True hysteresis is associated with matrix expansion in glassy, but not rubbery, solids (Sander and Pignatello, 2005; Sander et al., 2006). From the preceding discussion, it is evident that SOM properties (glassy character) can be altered and isotherm nonlinearity changes accordingly.

Cosolute Addition
If the solid contains a limited number of discreet sorption sites, then the addition of a cosolute can lead to a competitive effect in which there is mutual suppression of sorption. In fact, concave-down nonlinearity (n < 1) in the isotherm of a single solute can be regarded as a manifestation of self competition for sorption sites that increases in intensity with loading. Competitive sorption is expected, and observed, for sorption by glassy organic solids due to the presence of microporous holes in the solid phase (Xing and Pignatello, 1997). Competitive sorption has been observed in many soils (e.g., Lu and Pignatello, 2004a; Xing and Pignatello, 1997). Competition is not observed in partitioning behavior involving dissolution, such as partitioning between two liquids or between a fluid and a rubbery polymer (Xing et al., 1996; Sander et al., 2006).

Figure 6 shows the single-solute isotherms of TCB in H-HA and in Al-HA at pH 3 and pH 6, respectively. At specific points along the isotherm, we performed competitive sorption experiments by adding increasing amounts of the competing cosolute, 1,3-DCB, while keeping the mass of TCB constant. Increasing the concentration of 1,3-DCB resulted in decreasing sorption of TCB in both solids. Concurrently, the solution-phase concentration of TCB increased progressively with 1,3-DCB concentration, leading to the diagonal shift of the isotherm point illustrated in Fig. 6. Competitive sorption in H-HA and Al-HA by these nonspecifically interacting compounds is evidence that these two HA materials are not completely in a rubbery state.


Figure 6
View larger version (17K):
[in this window]
[in a new window]
  		Fig. 6. Sorption of 1,2,4- trichlorobenzene (TCB) by Amherst humic acid (H-HA) and Amherst Al-flocculated humic acid (Al-HA) in the absence and presence of 1,3- dichlorobenzene (DCB) as the competing solute. Dashed lines represent Freundlich fits for the entire single-solute trichlorobenzene (TCB) isotherms, only part of which is shown in each case. Arrows indicate progressively increasing initial concentration of DCB (CDCB) from 0 to 25.8 mg/L in each series at constant total TCB mass present.

 

   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 STRUCTURAL PROPERTIES OF SOIL...
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
We have shown that HAs and humic precursors free of BC sorb apolar compounds nonlinearly and that sorption is competitive when two solutes are present together. These phenomena are typical of glassy solids. Several of the HAs displayed glass transition behavior above room temperature. In selected cases, it was shown that sorption nonlinearity increases or decreases as one would predict for a solid in its glassy state: nonlinearity decreases with increasing sorption temperature, increasing cosolvent content, and reduction of metal-ion cross-linking of the organic matter. It is reasonable to conclude that macromolecular humic substances in the environment behave analogously and will exhibit nonideal sorption behavior in soils and sediments, quite apart from any such behaviors attributable to BC. Humic acid is lower in molecular weight and more water soluble than macromolecular SOM as a whole. Given that the Tg increases with molecular weight for a polymer of given subunit structure (Toi and Paul, 1982), it may be expected that SOM would be, overall, a harder material and capable of exhibiting even more pronounced nonideal behavior than HA. This study, thus, serves to caution the scientific community about generalizing the impact of BC on sorption of contaminants. Lastly, we note that polymethylene units are stable to the acid dichromate reagent used for BC determination.


   ACKNOWLEDGMENTS
 
This project was funded in part by Cooperative States Research Education and Extension Service of the USDA (NRICGP 2002-35107-12544) and in part by the U.S. National Science Foundation (Grants no. 9985159, BES-0122761, and BES-0404487). We thank Ms. Rossane C. DeLapp for her assistance in the thermal analysis laboratory in Nashville, TN.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 STRUCTURAL PROPERTIES OF SOIL...
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow