| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a State Key Laboratory of Organic Geochemistry, Guangzhou Institute of Geochemistry, Chinese Academy of Science, Wushan, Guangzhou 510640, China
b School of Environmental Science, Engineering, and Policy, Drexel University, Philadelphia, PA 19104
c School of Civil Engineering, Purdue University, West Lafayette, IN 47907-1284
* Corresponding authors (yran{at}gig.ac.cn; weilin.huang{at}drexel.edu)
Received for publication September 10, 2001.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
Abbreviations: DCB, 1,2-dichlorobenzene • FP, acid-treated peat • HOC, hydrophobic organic contaminant • PP, Pahokee peat • SOM, soil organic matter • TCB, 1,3,5-trichlorobenzene • TOC, total organic carbon • TP, base- and acid-treated peat
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
Recent investigations reveal the importance of microporosity in SOM and its role in the HOC sorption. The CO2 sorption on SOM indicates that SOM has appreciable internal microporosity not detected with N2 (De Jonge and Mittelmejer-Hazeleger, 1996). The SOM microporosity has been correlated with the degree of nonlinearity and competitive effects for chlorinated benzene sorption from solution (Xing and Pignatello, 1997). Xia and Ball (1999) applied Polanyi-based isotherm model to describe uptake of nine HOCs by natural solids. They suggested that the adsorption contribution was probably from pore filling of HOCs within microporous domains.
This study examines the effect of SOM heterogeneity in the sorption of HOCs. We hypothesize that peat and sediments may contain different types of condensed SOM, and that the pore filling in the condensed SOM domain may be an important mechanism for the sorption of HOCs. We tested this hypothesis by investigating equilibrium sorption with a peat and its two fractions, and two sediments as the sorbents and four organic solutes as the sorbates. The role of the condensed SOM in equilibrium sorption and the observed sorption–desorption hysteresis for the selected HOCs are presented in this paper.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
The contents of black carbon in the sediments and the original peat were determined following a three-step chemical procedure (Wolbach and Anders, 1989; Bird and Gröcke, 1997). In brief, about 3 g of a given sample were placed in a plastic bottle and treated with 30 mL of 10% (w/w) HCl for 12 h to remove CaCO3. The solid residue was then treated with 10 mL of 40% (w/w) HF at 80°C for 12 h to remove silicates. After digestion, the remaining solid consisting primarily of particulate organic matter was treated with 10 mL of 1 M K2Cr2O7 and 2 M H2SO4 at 60°C two times for 120 h for the sediments, and 15 times for 240 h for the original peat. By definition, the organic matter left after this oxidation process is black carbon material.
The total organic carbon (TOC) of the peats, sediments, and the isolated black carbon materials were analyzed with Heraeus (Hanau, Germany) CHN-O elemental analyzer and the results are listed in Table 1. The peat samples were first incinerated at a low temperature and then muffled at 750°C for 4 h. The ash contents are also listed in Table 1.
|
|
|
H), supercooled liquid-state solubilities (Sscl), molar volumes (Vm), densities (
), octanol–water partition coefficients (Kow), and organic carbon normalized distribution coefficient (Koc) values are listed in Table 3 (Mackay et al., 1992).
|
Sorption and Desorption Experiments
Sorption and desorption equilibria were measured with batch systems and following a withdraw–refill procedure described in our prior studies (Huang and Weber, 1997). The experiments consisted of both preliminary and final tests. The preliminary test was designed to determine both solid–solution contact times for achieving apparent sorption–desorption equilibria and appropriate sorbent-to-solution ratio for each sorbent–solution system that yields a 40 to 70% reduction in aqueous-phase solute concentration in each reactor at the completion of sorption experiment. Our preliminary experiment demonstrated that 4 and 2 wk are adequate to reach the apparent sorption and desorption equilibria for the systems tested. Desorption was performed by replacing the majority of the supernatant with the HOC-free background solution for all the sorbate–sorbent systems. The final test was conducted to collect the sorption and desorption data presented in this study. The completely mixed batch reactors (CMBRs) consisted of flame-sealed glass ampoules (10 or 20 mL; Kimble, New Brunswick, NJ). Each reactor contained a predetermined amount of sorbent, was filled with initial aqueous solution up to the shoulder, and was immediately flame-sealed. The reactors were mixed continuously in a rotary shaker set at 125 rpm on a horizontal mode at 23°C. After mixing, the ampoules were set upright for 2 d to allow suspended solids to settle; our preliminary test showed that such a period was sufficient to separate solids from solution. After settling, each reactor was opened in a flame, weighed, and approximately 3 mL of the supernatant was withdrawn and mixed with approximately 2 mL of methanol in a 5-mL vial capped with a Teflon top.
The extents of desorption were measured for phenanthrene in separate reactor systems by repeating the above withdraw–refill procedure for three times. The time for attainment of apparent desorption equilibrium in each cycle was 2 wk.
The aqueous solute concentration of each run was determined from the solute concentration of the supernatant–methanol mixture measured on HPLC and a dilution factor calculated based on density data for the mixtures (Weber and Huang, 1996). The solid phase solute concentrations were computed based on a mass balance of solute between the two phases.
Reactors containing no sorbent were run simultaneously to assess loss of solutes to reactor components during sorption and desorption tests. Average system losses were shown to be consistently less than 4% of initial concentrations for the four HOCs, hence, no correction was made during reduction of sorption and desorption data. To check possible biodegradation of the solutes in the reactor systems, four separate preliminary tests with a standard plating procedure (see Huang and Weber, 1998) were initiated, each test corresponding to one of the four solutes. The results showed no measurable biodegradation of the target HOC solutes in the CMBRs.
Sorption and Desorption Data Modeling
All sets of sorption and desorption data were fitted with the following two sorption isotherm models; the Freundlich model having a form of:
 | [1] |
 | [2] |
For less polar and sparely soluble compounds, Cr is related to solute activity (a) in the water phase referenced to their respective pure liquid or supercooled liquid state at a given temperature condition (Schwarzenbach et al., 1993; Carmo et al., 2000). Specifically, for HOCs that are liquid under tested temperature conditions, Cr is the ratio of Ce to Sw. In this case, K'F represents the mass of HOC sorbed per unit mass of sorbent at the solute saturation concentration (Ce = Sw) (Carmo et al., 2000). For HOCs that are solid under tested temperature conditions, Cr is the ratio of Ce to Sscl. The Sscl values for TCB, naphthalene, and phenanthrene at T = 296 K can be estimated with the following equations (Schwarzenbach et al., 1993; Chiou et al., 1982):
 | [3] |
 | [4] |
The values of K'F calculated with Cr can be further normalized to organic C 
. The values of K'F and KF are correlated as following:
 | [5] |
A linear regression procedure with SYSTAT software (Version 8.0) (SPSS, 1998) was used for fitting Eq. [1] and [2], respectively, to the transformed sorption and desorption data. Values of resulting model parameters, along with their respective 95% confidence intervals and the number of observations involved in their determination, are presented in Tables 4 and 5. Representative data and model fits to the sorption data are shown in Fig. 2 and representative data of phenanthrene sorption and desorption are shown in Fig. 3 .
|
|
|
|
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
Elemental analysis shows that the elemental carbon or black carbon in Pahokee peat accounts for 1.62% of its TOC, consistent with a recent report (Pignatello, 1999). In that report, the black carbon content of Pahokee peat quantified following a two-step combustion procedure (Gustafsson et al., 1997) was 1.5 ± 0.20 of the TOC. The black carbon contents of the MC37 and MC45 sediments are respectively 13.2 and 27.2% of their total organic carbon (Table 1).
The structural groups of the three peats measured by 13C NMR indicate that the SOM of TP and FP was altered variously due to base and/or acid treatment. Each spectrum is divided into five regions, aliphatic C (0–45 ppm), methoxyl C (45–63 ppm), oxygenated aliphatic C (63–108 ppm), aromatic C (108–160 ppm), carboxylic C (160–183 ppm), and carbonyl C (183–225 ppm) (Table 2). The three peaks at 56, 72, and 104 ppm in the oxygenated aliphatic carbon region can be assigned to methoxyl groups (50–65 ppm), O-alkyl groups (63–95 ppm) (i.e., –CH(OH)– or –CH2–O–C), and deoxygenated-alkyl groups (95–110 ppm) (i.e., acetal and ketal) (Wilson et al., 1983; Mao, 2001). Methoxyl groups are originated from lignin, and the later two groups from carbohydrate. The 13C-NMR spectra show that the O-alkyl group of TP at 63 to 95 ppm is significantly increased, whereas aliphatic and methoxyl carbons decrease, indicating that, under oxygen-containing and varied pH conditions during treatment, aliphatic carbons and lignin may be partially removed, and/or altered due to hydrolysis reactions. The carboxylic and carbonyl carbons of FP at 160 to 225 ppm are obviously elevated, whereas the peaks in the ranges of polysaccharide and lignin carbon structures are decreased.
Sorption Isotherm and Desorption Hysteresis
The data presented in Tables 4 and 5 for the sorbent–sorbate systems tested show that (i) both sorption and desorption isotherms are variously nonlinear and can be quantified with the Freundlich model in the two forms (Eq. [1] and [2]); (ii) although the sorption capacity parameter KF varies dramatically, the log K'Foc values for the sorption of the four sorbates on the peats fall in a narrow range of 4.5 to 5.0, and for that of phenanthrene and naphthalene on the peats and sediment in a range of 4.5 to 5.1; (iii) hysteresis exists in the sorption–desorption of the two sorbates having higher Kow values. The measured n values range from 0.669 to 0.990 for most of the sorbate–sorbent systems. The n values of the sorption isotherms for phenanthrene and naphthalene are within the range of 0.679 to 0.852 and 0.789 to 0.903, respectively, whereas the n values for 1,3,5-TCB and 1,2-DCB are 0.813 to 1.038 and 0.904 to 1.157, respectively. It appears that, for a given sorbent and the same class of HOCs, the n value is inversely correlated with the hydrophobicity of the HOCs.
One of the characteristics of isotherm nonlinearity is that the single-point Koc value measured at low Ce for a specific sorbent–sorbate system is several times greater than that at high Ce values. For instance, the single-point Koc values of the nonlinear isotherms at Ce/Sw = 0.001 are 2.50 to 7.81, 1.83 to 3.71, 1.48 to 3.20, and 1.06 to 1.82 times those at Ce/Sw = 0.5 for phenanthrene, naphthalene, TCB, and DCB, respectively. Some early literature also reported the same trend (Huang et al., 1997; Huang and Weber, 1998; Xing, 2001).
The single-point Koc values listed in Table 4 are greater than those in Table 3 predicted for the same solutes with empirical log Koc–log Kow correlations based on a linear sorption model (Karickhoff et al., 1979; Means et al., 1980; Schwarzenbach and Westall, 1981). For example, the Koc values calculated from the sorption isotherm data on the peats and sediments at Ce/Sw = 0.001 for the four sorbates are respectively 1.0 to 28 and 0.57 to 7.37 times those estimated by Schwarzenbach and Westall (1981) and Karichoff et al. (1979). Depending upon aqueous-phase solute concentration and solute, the Koc values listed in Table 4 for the sorption isotherms of phenanthrene, naphthalene, TCB, and DCB on the peats and sediments are respectively 6.68 to 28.0, 2.94 to 5.93, 2.1 to 4.2, and 1.0 to 2.8 times the reported Koc values by Schwarzenbach and Westall (1981). The Koc value of the original peat (PP) can respectively be up to 1.8 times and up to 1.5 times greater than that of the base- (TP) or acid-treated (FP) peats at three Ce/Sw levels of a specific HOC solute tested.
One of the advantages of using Eq. [2] for fitting the isotherm is that the normalized K'Foc values change slightly for the four sorbates. The log K'Foc values of the four chemicals range from 4.8 to 5.0 and are not different at the 95% confidence interval between PP and FP, but those for TP are significantly lower than for PP, suggesting that HCl–HF treatment for FP did not change the sorption capacity, but the alkaline treatment for TP greatly did. However, the densities of phenanthrene and naphthalene are very similar and are lower than those of 1,2-DCB and 1,3,5-TCB (Table 3). If the difference in the sorbate density is considered, the below section will show that the four sorbates have different sorption volumes.
It is evident from Table 4 that log KF and/or n values of the sorption isotherms for both phenanthrene and 1,3,5-TCB are lower than those of their respective desorption curves, indicating that apparent hysteresis exist for the two sorbates. The differences in the log KF and/or n values between the sorption and desorption curves are statistically insignificant for both naphthalene and 1,2-DCB, suggesting that both sorbates exhibit little or no sorption–desorption hysteresis. One exception for 1,2-DCB on TP may be related to the precision for measuring the sorption and desorption data. It was noted that the 95% confidence intervals of log KF and n values for 1,2-DCB on the peats and sediments and for naphthalene on the sediments are relatively large. Hence, we are not sure of the hysteresis of naphthalene and 1,2-DCB on the sediments. The below section suggests that naphthalene and 1,2-DCB rather than phenanthrene and 1,3,5-TCB on the peats reach sorption equilibrium. It was reported that the hysteresis is probably attributable to slow desorption and entrapment of sorbing molecules within SOM matrices (Weber et al., 1998; Altfelder et al., 2000; Yuan and Xing, 2001). Recent investigations indicated that the effect of SOM matrix expansion on the phenanthrene sorption rate and tricholomethane hysteresis appears to be small for Pahokee peat (Braida et al., 2001; Xia and Pignatello, 2001). The above discussions possibly suggest that nonequilibrium sorption of phenanthrene and 1,3,5-TCB is the major reason for the desorption hysteresis. Successive desorption data for three times at the highest concentrations of phenanthrene on the peats and sediments show relatively large data scatter (Fig. 3). Therefore, the accumulating errors for the successive desorption steps of phenanthrene on the peats may mask the desorption hysteresis. However, the one-step desorption procedure employed nearly constant dilution factors (98.5 ± 0.68%) for all of the sorbate–sorbent systems. The sorbed concentration errors caused by the residual solutions were deducted according to the residual solution weights and the solute concentrations in the residual solutions. Hence, the one-step desorption curve is equivalent to the reestablishment of the sorption equilibrium from the desorption direction, and may demonstrate that hysteresis appears at different sorbed concentrations.
Nonlinear Sorption and Its Mechanistic Implications
The n values for the four HOCs decline in the following order: FP > PP at 95% confident intervals except for 1,3,5-TCB (at one standard deviation level), and PP > TP at 95% confident intervals except for naphthalene (at one standard deviation level) and 1,2-DCB (Table 4). As the TP treatment partially removed humic acid, and partially oxidized and/or altered SOM, the difference in the n values for the four HOCs are not consistently significant at the 95% confidence interval. Xing and Pignatello (1997) showed that the n values for 1,2-DCB and 1,3-DCB decline in the following order: PP > humin (insoluble solid remaining after extraction of PP with Na4P5O7 solution). The higher n value of TP than of PP for 1,2-DCB in this study may relate to the precision for measuring the sorption isotherm of 1,2-DCB (see the above discussion), as illustrated in Fig. 2 for TP and FP. However, the n values for the sorption of phenanthrene and 1,3,5-TCB on the peat fractions suggest the tendency that the degree of SOM condensation appears to follow the order: FP < PP < TP.
The isotherm nonlinearity observed in this study may be attributable to the HOC dual partitioning–adsorption behavior exhibited by the condensed SOM. Black carbon is formed during incomplete combustion of fossil fuels or biomass (Goldberg, 1985), and is a form of highly condensed carbon. Several recent studies showed that the condensed SOM exhibits nonlinear sorption behavior via adsorption or a "hole filling" process. In a recent study we reported that the physically condensed and chemically reduced kerogen is responsible for isotherm nonlinearity, higher sorption capacity, and sorption–desorption hysteresis for a sandy aquifer (Borden, ON, Canada), even though the TOC content of the sand is very low (approximately 0.021% w/w) (Ran et al., unpublished data, 2002). Table 4 shows that MC-45, having a higher black carbon content, exhibits more nonlinear sorption isotherms for naphthalene and phenanthrene than does MC-37, which has a lower black carbon content.
The nonlinear sorption behavior contributed by humic substances and kerogen is evident for PP, which has a much lower content of black carbon particles than MC-45, but has approximately the same n values as does MC-45. The decrease of the n value on the order of FP > PP > TP for phenanthrene and TCB may further suggest that the peat treated differently has varied condensation and conformation and hence exhibits different isotherm nonlinearity. Yuan and Xing (2001) reported that cation (particularly multivalence) exchanged humic acids exhibit decreased isotherm linearity for both naphthalene and phenanthrene, suggesting that cations can cause more condensed configurations for humic acids. Mineral surfaces can exert similar influence on the structures and molecular configurations of bound humic substances, such as alteration of the size or accessibility of the hydrophobic domain (Murphy et al., 1990). The 13C-NMR spectra in Table 2 indicate that the functionality of the peat organic matter was certainly changed. Consequently, removal of mineral components by acid digestion may have expanded the SOM matrix of the peat and lowered the condensation of the relatively condensed organic domain, probably resulting in a more linear isotherm for FP than PP for a given sorbate. In contrast, partial removal of the humic acid by base extraction may have increased the condensation of the residual SOM since the humic acid preferentially removed is polar and less condensed (Hayes et al., 1989). The lower n value of the isotherms measured for TP than PP may suggest a net increase of SOM condensation from the combined base extraction and acid digestion.
The study further suggests that sorption within the local regions of condensed SOM may follow a pore-filling process. Table 5 listed the sorption volume 
for each sorbent calculated from the qe at Cr = 1 and the density of each solute at tested temperature conditions. The results indicate that the three peat samples, PP, FP, and TP, have sorption volumes ranging from 0.016 to 0.041 cm3/g for the four target solutes and that each of the peat samples has approximately the same sorption volume for both 1,2-DCB and naphthalene, which are smaller in molecular sizes among the four solutes. Interestingly, the sorption volume of PP at a level of 0.041 cm3/g calculated for the two smaller solutes is the same as the total volume of micropores (0.041 cm3/g) measured with CO2 gas as the adsorbate (Xing and Pignatello, 1997), suggesting that both small HOC solutes and gas adsorbate could have equal accessibility to the same microporous domain of PP. However, the calculated sorption volumes (0.026–0.029 cm3/g) of phenanthrene and 1,3,5-TCB, the larger solutes among the four target HOCs, are about 30% lower than the CO2–based microporosity for PP, and the sorption volumes of the two large HOC solutes are comparably smaller than those of the two small HOC solutes for each of the other two peat samples. As SOM has appreciable microporosity not detected with N2 and about 95 to 99% of the SOM surface area is formed by micropores (<2–0.5 nm) (De Jonge and Mittelmejer-Hazeleger, 1996), the molecular diffusion in SOM and intraparticle micropores may be very slow and sensitive to the solute size, shape, and hydrophobicity (Brusseau et al., 1991a, b; Pignatello and Xing, 1996; Xing and Pignatello, 1997). The HOC sorption into such micropores probably follows an adsorption or "hole filling" process.
The lowered pore-filling (adsorption) saturation degree, a ratio of sorption volume of HOC to that of 1,2-DCB, suggests that 1,3,5-DCB and phenanthrene has not reached the sorption equilibrium or may not enter some fraction of micropore. One possibility is that these different size molecules occupy separate holes within the condensed SOM. For instance, Xing and Pignatello (1997) found that the condensed SOM gave weak or no competition between atrazine and trichloroethene. Alternatively, the lowered saturation degree of solid adsorbates such as phenanthrene and TCB may result from varied lower packing efficiencies as they may be condensed as solids within micropores (Manes, 1998; Xia and Ball, 1999). The pore-filling saturation degrees of phenanthrene and TCB calculated for PP, TP, and FP are close to an average efficiency of 0.77 reported by Manes (1998). The saturation degree of naphthalene, an adsorbate presumably condensed as a solid, is higher and approximately the same as that of 1,2-DCB, an adsorbate presumably condensed as a liquid. Whether the adsorbed HOCs of the solid sorbate exist in the liquid state or solid state needs further experimental investigation.
|
CONCLUSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
It was found that for the small solutes 1,2-DCB and naphthalene, the sorption volumes on the peat fractions were generally very similar, and that on PP was the same as the CO2–determined micropore volume. The sorption volumes for the larger solutes phenanthrene and 1,3,5-TCB on PP were lower than the measured micropore volume. This investigation suggests that the microporosity of SOM is important for the HOC sorption capacity on the peat.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. Farenhorst Importance of Soil Organic Matter Fractions in Soil-Landscape and Regional Assessments of Pesticide Sorption and Leaching in Soil Soil Sci. Soc. Am. J., April 19, 2006; 70(3): 1005 - 1012. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. J. Smernik A New Way to Use Solid-State Carbon-13 Nuclear Magnetic Resonance Spectroscopy to Study the Sorption of Organic Compounds to Soil Organic Matter J. Environ. Qual., June 7, 2005; 34(4): 1194 - 1204. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 X. Pu, L. S. Lee, R. E. Galinsky, and G. P. Carlson Evaluation of a Rat Model versus a Physiologically Based Extraction Test for Assessing Phenanthrene Bioavailability from Soils Toxicol. Sci., May 1, 2004; 79(1): 10 - 17. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Ran, B. Xiao, W. Huang, P.'a. Peng, D. Liu, J. Fu, and G. Sheng Kerogen in Aquifer Material and Its Strong Sorption for Nonionic Organic Pollutants J. Environ. Qual., September 1, 2003; 32(5): 1701 - 1709. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Soil Science Society of America Journal | Journal of Plant Registrations | The Plant Genome | |||
