- Email: [email protected]
Potential contributions of clay minerals and organic matter to pentachlorophenol retention in soils
Chemosphere 65 (2006) 497–505 www.elsevier.com/locate/chemosphere
Potential contributions of clay minerals and organic matter to pentachlorophenol retention in soils Yan He, Jianming Xu *, Haizhen Wang, Qichun Zhang, Akmal Muhammad College of Environmental and Natural Resource Sciences, Zhejiang University, Hangzhou 310029, China Received 3 June 2005; received in revised form 9 January 2006; accepted 10 January 2006 Available online 14 February 2006
Abstract Sorption of pentachlorophenol (PCP) by pure minerals and humic acids were measured to obtain additional perspective on the potential contributions of both clay minerals and soil organic matter (SOM) to contaminants retention in soils. Four types of common soil minerals and two kinds of humic acids (HAs) were tested. The sorption affinity for PCP conformed to an order of HAs K-montmorillonite Ca-montmorillonite goethite kaolinite. Such a difference in sorption capacity could be attributed to the crucial control of HAs. Clay minerals also had their contribution, especially K-montmorillonite, which played an important, if not dominant, role in the controlling process of PCP sorption. By removing 80% (on average) of the organic carbon from the soils with H2O2, the sorption decreased by an average of 50%. The sorption reversibility had been greatly favored as well. Considering the uncharged mineral fractions in soil before and after H2O2-treated, the main variation in sorption behavior of the soil might thus be related to the removed organic carbon and the reduced pH. This testified rightly the interactive effect of SOM and clay minerals on PCP sorption as a function of pH. 2006 Elsevier Ltd. All rights reserved. Keywords: Pentachlorophenol (PCP); Sorption; Minerals; Humic acids (HAs)
1. Introduction Pentachlorophenol (PCP, C6Cl5OH) is a kind of ionizable hydrophobic organic contaminants (HOCs). As a pesticide, herbicide, and antiseptic, it was once used worldwide, and has been designated as a priority pollutant and a probable human carcinogen (US Environmental Protection Agency, 1980). In the 1970s PCP was popularly used in China in fighting against snail fever and as a herbicide. Its biodegradation is slow; hence it is not surprising that it still caused environmental problems at many locations (Sanjoy et al., 1996). The highest concentrations of PCP are usually found in soil and aquatic sediments (Maatela et al., 1990; Abrahamsson and Klick, 1991; Sung-Kil and Angela, 2003). PCP can remain for a very long time. Understanding the behavior of PCP requires an assessment
*
Corresponding author. Tel./fax: +86 571 8697 1955. E-mail address: [email protected] (J. Xu).
0045-6535/$ - see front matter 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2006.01.020
of the processes influencing its fate, transport, bioactivity and persistence in soils. The investigation on sorption and desorption behaviors, and the controlling key factors involved may be a possibly helpful and effective way. It has been well established that the retention of HOCs in soil–water systems is strongly correlated with soil organic matter (SOM) content and that soil mineral fractions play a comparatively minor role except in the absence of water. SOM is viewed as providing a partition phase for the uptake of HOCs (Kile et al., 1995; Weber and Huang, 1996; Xing and Pignatello, 1997; Xia and Ball, 1999). And the SOMpartition model appears valid for organic contaminants containing nonpolar or slightly polar functional groups (e.g., –Cl). It is, however, frequently extended to organic contaminants in general, including those containing polar functional groups such as many pesticides. This is illustrated by the common use of soil organic matter (or carbon)-normalized sorption coefficients (KOM, KOC) to predict pesticide mobility in soils. It is now clear that for important categories of pesticides (e.g., triazines,
498
Y. He et al. / Chemosphere 65 (2006) 497–505
carbamates, nitrophenols) and organic contaminants (e.g., nitroaromatic compounds), sorption by clays may equal to or exceed that by SOM (Haderlein et al., 1996; Sawhney and Singh, 1997; Boyd et al., 2001; Sheng et al., 2002; Li et al., 2003). For example, nitro-substituted aromatics, including explosives and some dinitrophenol pesticides, are also significantly adsorbed by clays (Haderlein and Schwarzenbach, 1993; Haderlein et al., 1996; Weissmahr et al., 1998). K+-saturated smectite (i.e., SWy-2) was a more effective sorbent for pesticides such as 4,6-dinitro-o-cresol and 2,6-dichlobenil compared to sorption by an organic soil (Sheng et al., 2001). Using sorption data from soils of variable clay and SOM contents, Karickhoff (1984) attempted to evaluate conditions under which mineral-phase sorption of organic contaminants in whole soils was important. For the N-heterocyclic simazine and biquinoline, mineral contribution to overall sorption became apparent (i.e., greater than that predicted from KOM values) at clay/organic matter ratios of >30. Sorption of pyrene, however, was not affected by clay content. It was then concluded that the contribution of mineral-phase sorption was a direct function of the polarity of the compound. The studies cited above, along with experimental evidence from earlier studies (Bailey and White, 1970), clearly indicate the ability of pure minerals to effectively bind organic molecules. However, such exact ability of clay mineralogy compared to SOM varied with contaminant types, which was less direct and has not been forthcoming for ionizable HOCs such as PCP. The objectives of this study were (1) to provide sorption data for PCP by both pure minerals and humic acids over a wide range of PCP concentrations, (2) to determine the contribution to PCP sorption of common soil minerals (kaolinite, montmorillonite, goethite) and (3) to determine the contribution of SOM to PCP sorption, and thereby improve our understanding of soil parameters that control sorption of ionizable HOCs. 2. Materials and methods 2.1. Materials Pure mineral phases. Four kinds of pure minerals were used to determine their contributions to sorption from common soil minerals. Kaolinite represented nonexpendable two-layer clay minerals. Homoionic K- and Ca-montmorillonite represented strongly expandable three-layer clay minerals with different valent exchangeable cations. Goethite represented common iron oxides. The selected minerals belonged to silicate and oxide minerals and therefore could represent different surface functional groups. The synthetic goethite was prepared according to the methods of Boily et al. (2001). It was well crystalloid and had a high purity according to the X-ray analysis (data not shown). In order to obtain homoionic montmorillonite, the <2 lm clay-sized fractions were obtained by wet sedimentation and subsequently subject to K+ and Ca2+ satu-
ration, by dispersing 10 g clay samples in 1 l of 0.5 M KCl or 0.5 M CaCl2 solution. The clay suspensions were shaken for 24 h, and then fresh chloride solutions were used to displace the original solutions after centrifugation. This process was repeated three times to ensure complete K- or Ca-saturation. The excess KCl or CaCl2 was removed by repeatedly washing with Milli-Q water until Cl was negatively determined by reacting with AgNO3 solution. The clay suspensions were then air-dried, and stored for later use. Humic acid phases. Two kinds of humic acids (HAs) were used. One was purchased from Jufeng Co. (Shanghai, China) and designed as HAs1. The other designed as HAs2 was collected from a paddy soil with the total C of 20.7 g kg1, total N of 2.7 g kg1, and pH of 4.80 (Jinhua, Zhejiang, China). The detailed extraction and purification procedure were described by Chen and Pawluk (1995). Briefly, soil was mixed with 0.1 M Na4P2O7 solution (1:10 w/v) under N2 and shaken overnight. The alkali-soluble SOM was separated from mineral and humin fractions by high speed (20,000g) centrifugation. The supernatant was acidified with 6 M HCl to pH 1.0 to precipitate the HAs. The HAs were then treated with a dilute (0.5% V/V) HF/HCl solution to remove ash, dialyzed in Milli-Q water, freeze-dried, and ground to fine powders. Selected physicochemical properties of the materials above are listed in Table 1. Organic carbon (OC) content was performed using a TOC analyzer (TOC-500, Shimadzum, Japan) with auto sampler after acidic digestion. Surface area (SA) was measured by multiple point nitrogen BET to be 4.0 m2 g1. Cation exchange capacity (CEC) was determined according to the procedure of Hendershot and Duquette (1986). And the pH was measured from the equilibrium supernatant after sorption. H2O2-treated soil phases. Six referenced soils (Xie et al., 2003), from Zhejiang Province in southeastern China, were used for further investigation of the potential contribution of SOM to sorption by comparing the different retention behaviors before and after the removal of organic carbon with H2O2 treatment. To remove organic matter from the soils, each 10 g of the air-dried soils was treated with approximately 25 ml of 35% H2O2 at room temperature for 4 days. This treatment was repeated twice, and the sediments were washed with Milli-Q water for four times and then air-dried. The organic carbon content was measured before and after removal of the organic carbon. Chemical. PCP with purity more than 98% from Aldrich Chemical Co. was used as the model compound in this study. It is a highly chlorinated ionizable HOCs and a weak acid with a molecular weight of 266.5 g mol1, pKa of 4.75, and log Kow of 5.01 (Ferro et al., 1994; Joel and Jerald, 1998). 2.2. Sorption experiments Sorption isotherms. Sorption was conducted in glass screw-PTFE-cap centrifuge vials. About 10–1000 mg (on
Y. He et al. / Chemosphere 65 (2006) 497–505
499
Table 1 Characteristics of minerals and humic acids, Freundlich isotherm parameters, and sorption distribution coefficient (Kd) for PCP sorption Sorbents
OCa (g kg1)
SA (m2 g1)
CEC (cmol (+) kg1)
pHb
Sorption isotherm parameters KF
Mineral Goethite Kaolinite Ca-Montmorillonite K-Montmorillonite Humic acids HAs1 HAs2 a b c
1.03 2.58 2.55 1.06 368.27 574.20
N
R2c
Kd
59 18 635 644
3.1 7.8 95.0 92.0
7.91 8.55 8.37 8.35
8.46 0.85 34.19 193.69
0.9391 1.3890 0.4597 0.6895
0.9792** 0.9882** 0.9848** 0.9954**
6.85 2.43 10.84 127.13
– –
– –
5.26 4.94
3412.39 5666.31
0.7695 0.7400
0.9903** 0.9862**
2842.59 4828.54
OC: organic carbon; SA: surface area; CEC: cation exchange capacity. The pH of the equilibrium solution after PCP sorption. ** Correlation is significant at 0.01 probability level.
a dry-weight basis) of the sorbents was added, followed by adding 10 ml of different PCP-spiked level background electrolyte solutions containing 1 mM CaCl2, 0.1 mM MgCl2, 0.5 mM Na2B4O7 Æ 10H2O (Kan et al., 1997), with 200 mg l1 NaN3 to inhibit aerobic biodegradation. The solid to liquid ratio was adjusted to achieve 30–70% of total solute sorbed. Solutes were added via methanol carrier, keeping the methanol percentage in the range of 0.04–1%. Although methanol content varied slightly over an isotherm, there was no effect on bulk solution properties. Each series comprised 15 samples. For each batch experiment, one series of vials without sorbent as control were also prepared and monitored. The control series did not indicate any significant PCP degradation or sorption losses on the glassware during the course of the experiment. Average PCP losses in the control series reached 2.57% within the experiment period, thus the missing PCP from the solution phase can be considered safely sorbed onto the solid phase. The vials were shaken on an orbital shaker at 200 rpm at 20 ± 2 C for 24 h in the dark. Preliminary tests indicated that 24 h was sufficient to reach apparent equilibrium. After equilibration, the vials were centrifuged at 1932g for 20 min. An aliquot of 1 ml of the supernatant liquid was filtered through a syringe of 0.22 lm Millipore membrane (ANPEL, B13 mm) prior to determination. HPLC analysis was carried out on a Waters Alliance 2695-2487 HPLC system fitted with a Symmetry C18 (5 mm, 3.9 · 150 mm, Waters, USA). Chromatography was performed at 30 C. The flow rate was maintained at 1.00 ml min1 by a gradient controller and solvent delivery system. The mobile phase was methanol:1% acetic acid 9:1 (V/V), and the injection volume was 20 ll. PCP was detected at 220 nm, and readings were integrated by the Millennium32 software system (Waters, USA). The PCP concentration was quantified with an external standard method. The detection limits of the methods are 5 lg l1. Sorbed mass was calculated by the difference between total and solution-phase mass. Sorption–desorption hysteresis. Cycle of sorption and desorption consisting of sorption followed by a single-dilution desorption step was conducted to investigate the sorp-
tion–desorption hysteresis. Sorption was performed as above. Concurrently, after the sorption period, desorption was achieved by withdrawing 96% of the supernatant and replenishing the samples with fresh PCP-free background electrolyte solution (actual amounts were determined by weight), and the vials were sealed and shaken for 24 h at 200 rpm at 20 ± 2 C in the dark, then centrifuged at 1932g for 20 min, and the supernatant was analyzed for solution-phase PCP concentration by HPLC. 2.3. Data analysis All sorption data were fitted to the logarithmic form of the Freundlich equation: log qe ¼ log K F þ N log C e
ð1Þ
where qe is the solid-phase concentration (mg kg1) and Ce is the liquid-phase equilibrium concentration (mg l1). The parameters KF for sorption capacity coefficient [(mg kg1)/ (mg l1)N] and N (dimensionless) indicating isotherm nonlinearity were determined by linear regression of log-transformed data. The isotherm data of the experiments with pure minerals and HAs were also analyzed assuming a linear sorption isotherm Kd ¼
qe Ce
ð2Þ
where Kd [l kg1] is the sorption distribution coefficient, qe [mg kg1] is the equilibrium sorbed concentration, and Ce [mg l1] is the equilibrium solution phase concentration. The sorption–desorption hysteresis was quantified for each sorbent-solute-solution system using the hysteresis index (HI) defined by Huang et al. (1998) and Weber et al. (1998) qd qs HI ¼ e s e ð3Þ qe T ;Ce where qse and qde refer to the solid-phase solute concentration for the single-cycle sorption and desorption experiment, respectively, and the bracket sub-scripts T and Ce
500
Y. He et al. / Chemosphere 65 (2006) 497–505
specify conditions held constant, i.e., temperature and residual solution phase concentration.
and 0.6895), the homoionic montmorillonite with expandable 2:1 type structure may have a more heterogeneous structure or composition than the kaolinite (1.3890).
3. Results and discussion 3.1. Sorption isotherm of pure minerals and HAs
3.2. Potential contributions of pure minerals and HAs on PCP sorption respectively
The low solubility and high sorption potential of PCP caused a problem in using a broad range of different initial concentrations. A low initial PCP concentration would leave an undetectable amount of PCP in solution. The result was a very narrow working PCP concentration range, from which the isotherm might just represent a small portion of a truly isotherm. One way to circumvent this problem is to use a suitable background electrolyte solution to increase the solubility of PCP as much as possible. For this reason, we compared lots of empirical background electrolyte solution in traditional sorption studies and at last the method of Kan et al. (1997) was chosen. This would allow for equilibrium aqueous concentrations of PCP close to the solubility limit, and two orders of the concentration magnitude spanning in the sorption isotherms could then be achieved, thus confirming a reliable isotherm. Typical sorption isotherms of PCP on pure minerals and humic acids were fitted to the logarithmic form of the Freundlich equation. Linear fitting of log-transformed data was justified over direct nonlinear curve fitting in this paper for two reasons: (1) concentrations were spread evenly over the log scale, thus, nonlinear curve fitting would underestimate the importance of the low concentration data, and (2) the relative uncertainty in the measurement was not greatly dependent on concentration (Xing, 2001). As plotted in Fig. 1, each of the six isotherms contained 15 points. The Freundlich model captured the inflection well, giving a good fit to individual isotherm (Table 1). According to Table 1, N values were in general less than 1, resembling L-type isotherms (Giles et al., 1960). A L-type curve indicated a relatively high PCP-sorbent affinity at low concentration; that is, the extent of sorption decreased as the PCP concentration increased, because it became increasingly more difficult for PCP to find vacant adsorption sites. Only the kaolinite isotherms showed the value of N > 1, indicating an S-type curve. This shape indicated a low PCP-kaolinite affinity at low concentration. Sorption became easier as the PCP concentration increased. With the 1:1 type of clay minerals and equaled abundance of Si and Al sites, this behavior in most cases might be attributed to the strong competition between the water molecules and PCP for the sorption sites. In certain cases, especially in the sorption of organic contaminants, a S-type isotherm might be due to cooperative interactions among adsorbed organic species that stabilized the sorbate and enhanced its affinity for the surface (Sposito, 1984). Weber et al. (1992) suggested that N value can be taken as an index of site energy distribution (i.e., the smaller the N, the more heterogeneous the sorption site). Therefore, among the clay minerals, with smaller N values (0.4597
The sorption coefficient (KF) ranged from 0.85 for kaolinite to 5666 for HAs2. However, precise comparison could not be made between KF values because of their different units (Table 1) as a result of nonlinearity (Chen et al., 1999). Therefore, the linear sorption Kd-values were calculated from the solution equilibrium concentration range where the curvature in the Freundlich equation was insignificant with 95% confidence limits and where the residuals were normally distributed about zero (Clausen et al., 2004). The results showed that the affinity and the sorption capacity of the tested sorbents for PCP increased in the order of kaolinite < goethite < Ca-montmorillonite K-montmorillonite HAs1 < HAs2. Both kaolinite and goethite showed a lowest sorption capacity, which were much smaller than that of HAs and K-montmorillonite. This might indicate that the mineral surfaces of both kaolinite and goethite could therefore only be of limited importance for the sorption of PCP. It could also be postulated that the soil under deep weathering condition might be likely to have poor sorption capacity due to blockage of reactive sites on clay minerals by the presence of oxidized minerals (e.g., iron oxides). Table 1 also illustrated the effect of adsorbed cations on the ability of the clays to adsorb PCP. Obviously, the type of cations adsorbed to clay minerals was a crucial factor that determined their affinity for PCP. Sorption of PCP to clays was high when the exchangeable cations at the clays were K+, but was negligibly small for homoionic Ca2+. Similar effects were observed by Haderlein and Schwarzenbach (1993) and Haderlein et al. (1996) for the adsorption of nitroaromatics by clays saturated with different cations. This appeared to be a manifestation of the comparatively weak hydration of K+. The enthalpy of hydration for K+ was 314 kJ mol1, much smaller than that of Ca2+ (1580 kJ mol1). Uncharged siloxane surfaces between charged sites on montmorillonite surface were relatively hydrophobic and could interact with nonpolar or less polar moieties of organic contaminants when they were able to access these mineral surfaces (Jaynes and Boyd, 1991; Laird et al., 1992; Laird and Fleming, 1999). In contrast, when montmorillonite was saturated with strongly hydrated divalent cations (e.g., Ca2+), the hydration sphere surrounding exchangeable cations diminished the size of adsorptive domains between cations (Haderlein et al., 1996; Weissmahr et al., 1997; Weissmahr et al., 1998; Sheng et al., 2002) and reduced the strength of interactions between exchangeable ions and polar functional groups in organic contaminants (Boyd et al., 2001; Johnston et al., 2001, 2002; Li et al., 2003). Differences in swelling behavior between K- and Ca-montmorillonite might also contribute
Y. He et al. / Chemosphere 65 (2006) 497–505
300
501
200
160 Goethite
Ca-montmorillonite
Kaolinite
250 120
150
200 35 80
Equilibrium sorbed concentration (mg kg-1)
150 25
100
100
6 4
35
40 50 0
15 0
0
20
2
40
50 2
4 0
60
0
20
K-montmorillonite
2000
55
0
2.5
40
60
15
5 80
HAs1
0
0
10
20
30
4
40
50
25000
1600 15000
20000 4000
1200 200
10000
2000
15000
1000
10000
800 100
2000
5000
400 0 0
2
HAs2
30000
20000
0
0
10
20
0
1 30
0
2 40
50
0
0
5
10
5000 0
0.3 15
20
0
0.6 25
0
0
5
10
0
0.2 15
20
0.4 25
Equilibrium solution phase concentration (mg l-1) Fig. 1. Sorption isotherms of pentachlorophenol on minerals and humic acids spanning two orders of magnitude in concentration. The symbols denote experimental datum points. Low concentration range datum points are shown in the insets for clear isotherm display.
to their differential PCP adsorption. For homoionic Kmontmorillonite, the basal spacing was often observed at ˚ , which appeared optimal for adsorption of organic 12.3 A contaminants (Boyd et al., 2001; Sheng et al., 2002). This spacing might just be large enough to allow intercalation of the PCP, while minimizing its interaction with water molecules. This energetically favorable process occurred when the PCP solute directly contacted the opposing clay siloxane surfaces. Larger interlayer spacing associated with Ca2+ or multivalent exchangeable cations did not allow the partial dehydration of organic solute in the clay interlayers (Boyd et al., 2001; Johnston et al., 2002; Sheng et al., 2002; Li et al., 2003). In addition, as shown in Table 1, the sorption capacity of HAs for PCP was significantly high, with the Kd value of PCP on both humic acids being 100 times greater than that of pure minerals on average. Of the HAs and solution parameters studied in this work, the system pH might emerge as a primary factor that controlled the sorption behavior of PCP by affecting the surface characteristics of the HAs and the distribution of PCP species in solution. The equilibrium pH for both the HAs after PCP sorption was 5.26 and 4.94, approaching nearly the pKa value of
organic matter (5, Pusino et al., 2000) and PCP (4.75, Ferro et al., 1994), under which cases both the organic matter and the PCP molecule might be only imperceptibly partly dissociated and thus maximize extremely the PCP sorption level of HAs. With a functional group of OH, PCP might dissociate over the range of equilibrium pH and be present as two chemical species with very different physico-chemical properties. Results presented by Lee et al. (1990) suggested that for pH less than 7, PCP sorption increased with an increase in the fractions of the PCP neutral species. Simultaneously, the component of aromatics in HAs allowed the neutral PCP to sorb through hydrophobic bond formation. And the conjugation between neutral PCP and carbohydrates, and the formation of intermolecular hydrogen bonds between PCP and COOH groups were also likely involved (Paaso et al., 2002). Therefore, the pHdependent chemical characteristics of PCP and the surface properties of HAs co-contributed to the complexity of PCP behavior in the HAs–liquid interface, and thus could result in the high sorption capacity observed. Therefore, it can be prospected that the sorption of PCP on soils/sediments is significantly influenced by the SOM. The clay minerals also have their contribution especially
Data referred to Xie et al. (2003). CEC: cation exchange capacity; FA-C: fulvic acid carbon; HA-C: humic acid carbon; HM-C: humin carbon; TOC: total organic carbon; H2O2-BF: before H2O2-treatment; H2O2-AF: after H2O2treatment.
a
4 5 6
b
5.2 4.7 7.5 5.6 5.3 9.0 0.9 1.8 3.4 6.6 11.0 5.5 75.4 34.5 4.6 7.4 31.0 71.1 17.2 34.5 24.3 1.6 5.9 3.9 2.8 2.1 0.9 2.3 2.9 0.7 12.3 14.9 7.1 Kaolinite Kaolinite Illite Red sandstone Acid magmatite Marine deposit
6.2 6.5 5.8 3.2 2.9 1.5 18.3 24.3 17.9 3.0 9.3 11.6 57.0 46.4 48.0 40.0 44.3 40.4 7.7 15.7 9.4 5.8 5.6 6.1 4.8 2.9 2.5 28.5 27.6 22.0 Illite Montmorillonite Illite
Fluviogenic loamy fine paddy soil Paddy field on bole soil Paddy field on redeposit of purple sandstone soil Paddy field on red sandstone soil Yellow–red soil Coastal saline soil 1 2 3
Fluviatile deposit Basalt Purple sand shale
H2O2-BF H2O2-BF HA-C
HM-C
Silt Clay FA-C
Sand
H2O2-AF
pH Particle size (%) CECb (cmol (+) kg1) Chief mineralsa Parent materiala Soil name
Table 2 Selected properties of the soil samples and their detectable changes before and after H2O2-treatment
Additional experiment was carried out by the analysis of PCP sorption behaviors on soils before and after H2O2-treatment, with the intention to give a further illustration on the impact factors of PCP retention. The samples used were six referenced soils (Xie et al., 2003) with contrasting physico-chemical properties, among which the chief clay minerals for soils 1, 3 and 6 were the same illite component with K-enriched nonexpandable 2:1 type structure; for soil 2 was 2:1 type expandable montmorillonite; and that for soils 4 and 5 were 1:1 type of kaolinite (Table 2). Adding H2O2 seemed a violent treatment of the soil, but no significant difference in soil composition or specific surface area was involved (Clausen et al., 2004). The detected changes only referred to the soil total organic carbon (TOC) and pH (Table 2). This might be resulted from the oxidation of SOM and thus a loss of aromatics and an increase in oxygen containing groups (Forsberg and Bjoroy, 1981), which resulted likely in a higher O/C ratio and a lower content of aromatics in soil component, and therefore, a more hydrophilic surface with H+ gathering. The two indexes were approximately parallel before and after H2O2-treatment; with a great reduction of the Kd values. Because of the concentration-dependent effect of the sorption distribution coefficient of our early study results (He et al., 2006), the Kd values for comparison here were calculated from the representative initial equilibrium concentrations in the linear range of the dual-mode model. The H2O2 decreased 7.14–26.15% of pH and removed, on average, about 80% of TOC (Table 2), resulting in a probably alternative variation to the sorption capacity (in this case, the Kd value) for PCP. One was the reduction with the remove of SOM, and the other was the increase with the decrease of pH, as discussed in the above section. From Fig. 2 however, it could be see that the total variation in sorption capacity presented a reduction of more than 50%. This might indicate properly the neglectable effect of pH decrease if at most a 26.15% range on PCP sorption when compared with that of SOM. Results presented by Lee et al. (1990) also suggested that regardless of the pH dependent mechanisms,
Humus fractions (g kg1)
3.3. Sorption and desorption hysteresis on soils before and after H2O2-treatment
TOC (g kg1)
H2O2-AF
in the case when K+ dominates the exchangeable cation on the clay surfaces. And both the two mechanisms are pH-dependent. It had been shown in the early study (Sheng et al., 2001) that some organic contaminants, e.g., DNOC, carbaryl, and dichlobenil, were more effectively sorbed by K-clay than by SOM. Others such as diuron, parathion, and biphenyl were sorbed more effectively by SOM than by clay. Atrazine was sorbed by K-clay and SOM to a similar extent. Hence, the main factors that influenced the sorption behavior of organic contaminants varied with their types, which reflected rightly the complexity of organic contaminant-clay interactions.
4.8 4.8 4.9
Y. He et al. / Chemosphere 65 (2006) 497–505
Soil no.
502
Y. He et al. / Chemosphere 65 (2006) 497–505
methods. Also, pH dependent effect could be concluded if taking a further analysis as for soil 6. The removal extent of TOC was actually low, thus the increment of HI was possibly related to the change of pH value. It was reduced by 1.5 pH units, from 9.0 to 7.5 after the H2O2-treatment. These findings testified rightly the coincident results obtained in the above study of pure minerals and HAs. Knowledge on how SOM and mineral composition affected sorption as a function of pH was therefore a likely key parameter for the prediction of the fate of PCP in soils and sediments.
60 Before H2O2-Treatment After H2O2-Treatment
50
K d (l kg -1)
40
30
3.4. Environmental significance and applications
20
10
0
503
1
2
3 4 Soil No.
5
6
Fig. 2. Sorption distribution coefficients (Kd-values) for pentachlorophenol on the representative six soils measured at equilibrium concentration of the linear range of dual-mode model before and after removal of organic carbon by H2O2-treatment. The error bars represent the standard deviation.
sorption increased with increasing soil organic carbon. Thus, in this case, considering the limited contribution to sorption from any of the different surface functional groups of the investigated minerals except K-montmorillonite, it is more likely that the sorption capacity in the investigated soils was related to the organic carbon content. Meanwhile, for the convenience of comparing the irreversibility of sorption–desorption, the hysteresis index (HI) was applied. In the present study, the sorption– desorption isotherms were constructed under conditions that rule out artificial causes of hysteresis. The values of HI at 20 C and five representative concentration levels before and after H2O2-treatment were employed (Fig. 3). As shown in Table 2 and Fig. 3, HI generally decreased after the treatment of H2O2, especially for the soils 1, 2, 4 and 5. With the 86% removal and nearly the zero content of TOC, PCP desorption was completely reversible in soil 4 after treated with H2O2 (a zero or negative HI value indicates that sorption–desorption hysteresis is statistically insignificant). Although it had a great reduce of both TOC and pH values, soil 1 showed a steady but low decline leading to the inverse 10–15% increase of HI at the highest concentrations. This might be attributed to the K-enriched nonexpandable 2:1 type clay minerals with the effect of interacting irreversibly with the PCP and thus increasing its retention in soil. Note worthy was that the soils 3 and 6, of which the values increased unexpectedly at the majority of the concentrations tested. This might be explained by possible minor surface changes in clay and oxide minerals, which might not be detected by the used characterization
In considering the applicability of these results to real soils, simple ion exchange processes might be useful in the development of an environmental-friendly protocol to control the sorption, mobility, and bioavailability of organic contaminants in montmorillonite-containing soils or soils amended with montmorillonite clays. In this scenario, addition of simple electrolyte solutions could be used to manipulate the type and composition of exchange cations associated with clays, thereby modulating the release and immobilization of contaminants. Although K+ is not normally the dominant exchangeable cation, it is one of the most abundant exchangeable cations in soils and subsoils. And the phenomenon of cation demixing in clays (Levy and Francis, 1975) results in entire interlayer regions being occupied by a single exchangeable cation (e.g., K+) even in the presence of other cations (e.g., Ca2+) because of differences in cation hydration. The widely occurring phenomenon of K-fixation by soil clays (Eberl, 1980; Bouabid et al., 1991) proves that K-rich domains are common indeed. Thus, improved understanding of organic contaminants adsorption by clays makes it possible to enhance the retentive properties of subsoils and aquifer materials for certain organic contaminant by controlling the base saturation of the matrix via in situ electrolyte injection (e.g., KCl) as suggested by the recent findings of Weissmahr et al. (1999). Also, the application using K-montmorillonite as a carrier in pesticide formulations in order to extend the efficacious period is possible. Pesticide would be gradually released as Ca2+ and Mg2+ in soils gradually replaced K+ on the clay. An interesting feature of both H2O2-treated and untreated soils is that hysteresis coefficients decreased generally with the increasing PCP loading concentration, indicating that desorption hysteresis is favored and the binding sites with high-energy are occupied firstly by PCP at low sorption levels. Thus, it could be concluded that if the low polluted sites were involved such as aquifers, PCP could undergo natural sequestration through the favored sorption–desorption hysteresis processes and restore the polluted environment to clean naturally. While if in the case of a serious pollution, other remediation processes must be considered, e.g., the application of organic matter and K-enriched clays from extraneous sources.
504
Y. He et al. / Chemosphere 65 (2006) 497–505
4
4
Hysteresis index (HI)
Soil 1
4 Soil 2
Soil 3
3
3
3
2
2
2
1
1
1
0
0
0
3
4 Soil 4
4 Soil 5
Soil 6
2
3
3
1
2
2
0
1
1
-1
0.1
0.2 0.4
0.6
0.8
0
0.1
0.2
0.4
0.6
0.8
0
0.1
0.2
0.4
0.6
0.8
Equilibrium concentration (mg l-1) Fig. 3. Hysteresis index (HI) for pentachlorophenol sorption on the six soils at five representative concentrations before and after removal of the organic carbon by H2O2-treatment. Symbols: (j) before H2O2-treatment; (h) after H2O2-treatment.
4. Conclusions The results of this study form a basis for understanding and quantifying the sorption behavior of PCP in soil/sediment environments. Specific sorption of PCP to clay minerals and its implications, particularly the effects of exchangeable cations at clay minerals, effects of HAs and its crucial controlling behaviors need to be considered and have been discussed in this paper. However, more work is necessary to evaluate the effect of other factors on PCP sorption in complex natural matrices such as the role of natural organic matter or other constituents that may influence the accessibility of clay mineral surfaces. In addition, sorption kinetics and competition behavior of PCP need to be re-evaluated in such a complex system, e.g., by batch and miscible displacement experiments. Acknowledgments This work was jointly supported by the National Nature Science Foundation for Distinguish Young Scholars of
China (40425007), and the Teaching and Research Award Program for Outstanding Young Teachers in Higher Education Institutions of China. References Abrahamsson, K., Klick, S., 1991. Degradation of halogenated phenols in anoxic marine sediments. Mar. Pollut. Bull. 22, 227–233. Bailey, G.W., White, J.L., 1970. Factors influencing the adsorption, desorption, and movement of pesticides in soils. In: Gunther, F.A., Gunther, J.D. (Eds.), Residue Review—Residues of Pesticides and Other Foreign Chemicals in Food and Feeds, vol. 32. Springer-Verlag, New York, pp. 29–92. Boily, J.F., Lu¨tzenkirchen, J., Balme`s, O., Beattie, J., Sjo¨berg, S., 2001. Modeling proton binding at the goethite (a-FeOOH)–water interface. Coll. Surf. 179, 11–27. Bouabid, R., Badraoui, M., Bloom, P.R., 1991. Potassium fixation and charge characteristics of soil clays. Soil Sci. Soc. Am. J. 55, 1493– 1498. Boyd, S.A., Sheng, G., Teppen, B.J., Johnston, C.T., 2001. Mechanisms for the adsorption of substituted nitrobenzenes by smectite clays. Environ. Sci. Technol. 35, 4227–4234. Chen, Z., Pawluk, S., 1995. Structural variations of humic acids in two Alberta Mollisols. Geoderma 65, 173–193.
Y. He et al. / Chemosphere 65 (2006) 497–505 Chen, Z., Xing, B., McGill, W.B., 1999. A unified sorption variable for environmental applications of the Freundlich equation. J. Environ. Qual. 28, 1422–1428. Clausen, L., Larsen, F., Albrechtsen, H.J., 2004. Sorption of the herbicide dichlobenil and the metabolite 2,6-dichlorobenzamide on soils and aquifer sediments. Environ. Sci. Technol. 38, 4510–4518. Eberl, D.D., 1980. Alkali cation selectivity and fixation by clay minerals. Clays Clay Miner. 28, 161–172. Ferro, A.M., Sims, R.C., Bugbee, B., 1994. Hycrest crested wheatgrass accelerates the degradation of pentachlorophenol in soil. J. Environ. Qual. 23, 272–279. Forsberg, A., Bjoroy, M., 1981. In: Bjoroy, A. (Ed.), Advances in Organic Geochemistry. Wiley, New York, pp. 60–68. Giles, C.H., MacEwan, T.H., Nakhwa, S.N., Smith, D., 1960. Studies in adsorption: Part XI. A system of classification of solution adsorption isotherms, and its use in diagnosis of adsorption mechanisms and in measurement of specific surface areas of solids. J. Chem. Soc., 3973– 3993. Haderlein, S.B., Schwarzenbach, R.P., 1993. Adsorption of substituted nitrobenzenes and nitrophenols to mineral surfaces. Environ. Sci. Technol. 27, 316–326. Haderlein, S.B., Weissmahr, K.W., Schwarzenbach, R.P., 1996. Specific adsorption of nitroaromatic explosives and pesticides to clay minerals. Environ. Sci. Technol. 30, 612–622. He, Y., Xu, J.M., Wang, H.Z., Ma, Z.H., 2006. Detailed sorption isotherms of pentachlorophenol on soils and its correlation with soil properties. Environ. Res., in press. Hendershot, W.H., Duquette, M., 1986. A simple barium chloride method for determining cation exchange capacity and exchangeable cations. Soil Sci. Soc. Am. J. 50, 605–608. Huang, W., Yu, H., Weber Jr, W.J., 1998. Hysteresis in the sorption and desorption of hydrophobic organic contaminants by soils and sediments. 1. A comparative analysis of experimental protocols. J. Contam. Hydrol. 31, 129–148. Jaynes, W.F., Boyd, S.A., 1991. Hydrophobicity of siloxane surface in smectites as revealed by aromatic hydrocarbon adsorption from water. Clays Clay Miner. 39, 428–436. Joel, G.B., Jerald, L.S., 1998. Predictive relationships for uptake of organic contaminants by hybrid poplar trees. Environ. Sci. Technol. 32, 3379–3385. Johnston, C.T., De Oliveira, M.F., Teppen, B.J., Sheng, G., Boyd, S.A., 2001. Spectroscopic study of nitroaromatic-smectite sorption mechanisms. Environ. Sci. Technol. 35, 4767–4772. Johnston, C.T., Sheng, G., Teppen, B.J., Boyd, S.A., De Oliveira, M.F., 2002. Spectroscopic study of dinitrophenol herbicide sorption on smectite. Environ. Sci. Technol. 36, 5067–5074. Kan, A.T., Fu, G.M., Hunter, M.A., Tomson, M.B., 1997. Irreversible adsorption of naphthalene and tetrachlorobiphenyl to lula and surrogate sediments. Environ. Sci. Technol. 31, 2176–2185. Karickhoff, S.W., 1984. Organic pollutant adsorption in aqueous systems. J. Hydraul. Eng. 110, 707–735. Kile, D.E., Chiou, C.T., Zhou, H., Li, H., Xu, O., 1995. Partition of nonpolar organic pollutants from water to soil and sediment organic matters. Environ. Sci. Technol. 29, 1401–1406. Laird, D.A., Fleming, P.D., 1999. Mechanisms for adsorption of organic bases on hydrated smectite surfaces. Environ. Toxicol. Chem. 18, 1668–1672. Laird, D.A., Barriuso, E., Dowdy, R.H., Koskinen, W.C., 1992. Adsorption of atrazine on smectites. Soil Sci. Soc. Am. J. 56, 62–67. Lee, L.S., Rao, P.S.C., Nkedi-Kizza, P., Delfino, J.J., 1990. Influence of solvent and sorbent characteristics on distribution of pentachlorophenol in octanol–water and soil–water system. Environ. Sci. Technol. 24, 654–666. Levy, R., Francis, C., 1975. Demixing of sodium and calcium ions in montmorillonite crystallites. Clays Clay Miner. 23, 475–476.
505
Li, H., Sheng, G., Teppen, B.J., Johnston, C.T., Boyd, S.A., 2003. Sorption and desorption of pesticides by clay minerals and humic acid–clay complexes. Soil Sci. Soc. Am. J. 67, 122–131. Maatela, P., Paasivitra, J., Sarkka, J., Paukku, R., 1990. Organic chlorine compounds in lake sediments: II. Organically bound chlorine. Chemosphere 21, 1343–1354. Paaso, N., Peuravuori, J., Lehtonen, T., Pihlaja, K., 2002. Sedimentdissolved organic matter equilibrium partitioning of pentachlorophenol: the role of humic matter. Environ. Int. 28, 173–183. Pusino, A., Braschi, I., Gessa, C., 2000. Adsorption and degradation of triasulfuron on homoionic montmorillonites. Clays Clay Miner. 48, 19–25. Sanjoy, K.B., Qiang, Y., Peikang, J., 1996. Removal of pentachlorophenol from wastewater by combined anaerobic–aerobic treatment. J. Hazard. Mater. 49, 143–145. Sawhney, B.L., Singh, S.S., 1997. Sorption of atrazine by Al- and Casaturated smectite. Clays Clay Miner. 45, 333–338. Sheng, G., Johnston, C.T., Teppen, B.J., Boyd, S.A., 2001. Potential contributions of smectite clays and organic matter to pesticide retention in soils. J. Agric. Food Chem. 49, 2899–2907. Sheng, G., Johnston, C.T., Teppen, B.J., Boyd, S.A., 2002. Adsorption of dinitrophenol herbicides from water by montmorillonites. Clays Clay Miner. 50, 25–34. Sposito, G., 1984. The Surface Chemistry of Soils. Oxford University Press, New York, pp. 1–46. Sung-Kil, P., Angela, R.B., 2003. Aqueous chemistry and interactive effects on non-ionic surfactant and pentachlorophenol sorption to soil. Water Res. 37, 4663–4672. US Environmental Protection Agency, 1980. Health and Environmental Effects Profile No. 135, Washington, DC. Weber, W.J., Huang, W., 1996. A distributed reactivity model for sorption by soils and sediments. 4. Intraparticle heterogeneity and phasedistribution relationships under nonequilibrium conditions. Environ. Sci. Technol. 30, 881–888. Weber, W.J., McGinley, P.M., Katz, L.E., 1992. A distributed reactivity model for sorption by soils and sediments. 1. Conceptual basis and equilibrium assessments. Environ. Sci. Technol. 26, 1955– 1962. Weber Jr, W.J., Huang, W., Yu, H., 1998. Hysteresis in the sorption and desorption of hydrophobic organic contaminants by soils and sediments. 2. Effects of soil organic matter heterogeneity. J. Contam. Hydrol. 31, 149–165. Weissmahr, K.W., Haderlein, S.B., Schwarzenbach, R.P., Hany, R., Nuesch, R., 1997. In situ spectroscopic investigations of adsorption mechanisms of nitroaromatic compounds at clay minerals. Environ. Sci. Technol. 31, 240–247. Weissmahr, K.W., Haderlein, S.B., Schwarzenbach, R.P., 1998. Complex formation of soil minerals with nitroaromatic explosives and other ðacceptors. Soil Sci. Soc. Am. J. 62, 369–378. Weissmahr, K.W., Hilderbrand, M., Schwarzenbach, R.P., Haderlein, S.B., 1999. Laboratory and field scale evaluation of geochemical controls on groundwater transport of nitroaromatic ammunition residues. Environ. Sci. Technol. 34, 2593–2600. Xie, Z.M., Wu, W.H., Xu, J.M., 2003. Study on fluoride emission from soils at high temperature related to brick-making process. Chemosphere 50, 763–769. Xia, G., Ball, W.P., 1999. Adsorption-partitioning uptake of nine low polarity organic chemicals on a natural sorbent. Environ. Sci. Technol. 33, 262–269. Xing, B., 2001. Sorption of anthropogenic organic compounds by soil organic matter: a mechanistic consideration. Can. J. Soil Sci. 81, 317– 323. Xing, B., Pignatello, J.J., 1997. Dual-mode sorption of low-polarity compounds in glassy poly (vinyl chloride) and soil organic matter. Environ. Sci. Technol. 31, 792–799.
Potential contributions of clay minerals and organic matter to pentachlorophenol retention in soils Yan He, Jianming Xu *, Haizhen Wang, Qichun Zhang, Akmal Muhammad College of Environmental and Natural Resource Sciences, Zhejiang University, Hangzhou 310029, China Received 3 June 2005; received in revised form 9 January 2006; accepted 10 January 2006 Available online 14 February 2006
Abstract Sorption of pentachlorophenol (PCP) by pure minerals and humic acids were measured to obtain additional perspective on the potential contributions of both clay minerals and soil organic matter (SOM) to contaminants retention in soils. Four types of common soil minerals and two kinds of humic acids (HAs) were tested. The sorption affinity for PCP conformed to an order of HAs K-montmorillonite Ca-montmorillonite goethite kaolinite. Such a difference in sorption capacity could be attributed to the crucial control of HAs. Clay minerals also had their contribution, especially K-montmorillonite, which played an important, if not dominant, role in the controlling process of PCP sorption. By removing 80% (on average) of the organic carbon from the soils with H2O2, the sorption decreased by an average of 50%. The sorption reversibility had been greatly favored as well. Considering the uncharged mineral fractions in soil before and after H2O2-treated, the main variation in sorption behavior of the soil might thus be related to the removed organic carbon and the reduced pH. This testified rightly the interactive effect of SOM and clay minerals on PCP sorption as a function of pH. 2006 Elsevier Ltd. All rights reserved. Keywords: Pentachlorophenol (PCP); Sorption; Minerals; Humic acids (HAs)
1. Introduction Pentachlorophenol (PCP, C6Cl5OH) is a kind of ionizable hydrophobic organic contaminants (HOCs). As a pesticide, herbicide, and antiseptic, it was once used worldwide, and has been designated as a priority pollutant and a probable human carcinogen (US Environmental Protection Agency, 1980). In the 1970s PCP was popularly used in China in fighting against snail fever and as a herbicide. Its biodegradation is slow; hence it is not surprising that it still caused environmental problems at many locations (Sanjoy et al., 1996). The highest concentrations of PCP are usually found in soil and aquatic sediments (Maatela et al., 1990; Abrahamsson and Klick, 1991; Sung-Kil and Angela, 2003). PCP can remain for a very long time. Understanding the behavior of PCP requires an assessment
*
Corresponding author. Tel./fax: +86 571 8697 1955. E-mail address: [email protected] (J. Xu).
0045-6535/$ - see front matter 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2006.01.020
of the processes influencing its fate, transport, bioactivity and persistence in soils. The investigation on sorption and desorption behaviors, and the controlling key factors involved may be a possibly helpful and effective way. It has been well established that the retention of HOCs in soil–water systems is strongly correlated with soil organic matter (SOM) content and that soil mineral fractions play a comparatively minor role except in the absence of water. SOM is viewed as providing a partition phase for the uptake of HOCs (Kile et al., 1995; Weber and Huang, 1996; Xing and Pignatello, 1997; Xia and Ball, 1999). And the SOMpartition model appears valid for organic contaminants containing nonpolar or slightly polar functional groups (e.g., –Cl). It is, however, frequently extended to organic contaminants in general, including those containing polar functional groups such as many pesticides. This is illustrated by the common use of soil organic matter (or carbon)-normalized sorption coefficients (KOM, KOC) to predict pesticide mobility in soils. It is now clear that for important categories of pesticides (e.g., triazines,
498
Y. He et al. / Chemosphere 65 (2006) 497–505
carbamates, nitrophenols) and organic contaminants (e.g., nitroaromatic compounds), sorption by clays may equal to or exceed that by SOM (Haderlein et al., 1996; Sawhney and Singh, 1997; Boyd et al., 2001; Sheng et al., 2002; Li et al., 2003). For example, nitro-substituted aromatics, including explosives and some dinitrophenol pesticides, are also significantly adsorbed by clays (Haderlein and Schwarzenbach, 1993; Haderlein et al., 1996; Weissmahr et al., 1998). K+-saturated smectite (i.e., SWy-2) was a more effective sorbent for pesticides such as 4,6-dinitro-o-cresol and 2,6-dichlobenil compared to sorption by an organic soil (Sheng et al., 2001). Using sorption data from soils of variable clay and SOM contents, Karickhoff (1984) attempted to evaluate conditions under which mineral-phase sorption of organic contaminants in whole soils was important. For the N-heterocyclic simazine and biquinoline, mineral contribution to overall sorption became apparent (i.e., greater than that predicted from KOM values) at clay/organic matter ratios of >30. Sorption of pyrene, however, was not affected by clay content. It was then concluded that the contribution of mineral-phase sorption was a direct function of the polarity of the compound. The studies cited above, along with experimental evidence from earlier studies (Bailey and White, 1970), clearly indicate the ability of pure minerals to effectively bind organic molecules. However, such exact ability of clay mineralogy compared to SOM varied with contaminant types, which was less direct and has not been forthcoming for ionizable HOCs such as PCP. The objectives of this study were (1) to provide sorption data for PCP by both pure minerals and humic acids over a wide range of PCP concentrations, (2) to determine the contribution to PCP sorption of common soil minerals (kaolinite, montmorillonite, goethite) and (3) to determine the contribution of SOM to PCP sorption, and thereby improve our understanding of soil parameters that control sorption of ionizable HOCs. 2. Materials and methods 2.1. Materials Pure mineral phases. Four kinds of pure minerals were used to determine their contributions to sorption from common soil minerals. Kaolinite represented nonexpendable two-layer clay minerals. Homoionic K- and Ca-montmorillonite represented strongly expandable three-layer clay minerals with different valent exchangeable cations. Goethite represented common iron oxides. The selected minerals belonged to silicate and oxide minerals and therefore could represent different surface functional groups. The synthetic goethite was prepared according to the methods of Boily et al. (2001). It was well crystalloid and had a high purity according to the X-ray analysis (data not shown). In order to obtain homoionic montmorillonite, the <2 lm clay-sized fractions were obtained by wet sedimentation and subsequently subject to K+ and Ca2+ satu-
ration, by dispersing 10 g clay samples in 1 l of 0.5 M KCl or 0.5 M CaCl2 solution. The clay suspensions were shaken for 24 h, and then fresh chloride solutions were used to displace the original solutions after centrifugation. This process was repeated three times to ensure complete K- or Ca-saturation. The excess KCl or CaCl2 was removed by repeatedly washing with Milli-Q water until Cl was negatively determined by reacting with AgNO3 solution. The clay suspensions were then air-dried, and stored for later use. Humic acid phases. Two kinds of humic acids (HAs) were used. One was purchased from Jufeng Co. (Shanghai, China) and designed as HAs1. The other designed as HAs2 was collected from a paddy soil with the total C of 20.7 g kg1, total N of 2.7 g kg1, and pH of 4.80 (Jinhua, Zhejiang, China). The detailed extraction and purification procedure were described by Chen and Pawluk (1995). Briefly, soil was mixed with 0.1 M Na4P2O7 solution (1:10 w/v) under N2 and shaken overnight. The alkali-soluble SOM was separated from mineral and humin fractions by high speed (20,000g) centrifugation. The supernatant was acidified with 6 M HCl to pH 1.0 to precipitate the HAs. The HAs were then treated with a dilute (0.5% V/V) HF/HCl solution to remove ash, dialyzed in Milli-Q water, freeze-dried, and ground to fine powders. Selected physicochemical properties of the materials above are listed in Table 1. Organic carbon (OC) content was performed using a TOC analyzer (TOC-500, Shimadzum, Japan) with auto sampler after acidic digestion. Surface area (SA) was measured by multiple point nitrogen BET to be 4.0 m2 g1. Cation exchange capacity (CEC) was determined according to the procedure of Hendershot and Duquette (1986). And the pH was measured from the equilibrium supernatant after sorption. H2O2-treated soil phases. Six referenced soils (Xie et al., 2003), from Zhejiang Province in southeastern China, were used for further investigation of the potential contribution of SOM to sorption by comparing the different retention behaviors before and after the removal of organic carbon with H2O2 treatment. To remove organic matter from the soils, each 10 g of the air-dried soils was treated with approximately 25 ml of 35% H2O2 at room temperature for 4 days. This treatment was repeated twice, and the sediments were washed with Milli-Q water for four times and then air-dried. The organic carbon content was measured before and after removal of the organic carbon. Chemical. PCP with purity more than 98% from Aldrich Chemical Co. was used as the model compound in this study. It is a highly chlorinated ionizable HOCs and a weak acid with a molecular weight of 266.5 g mol1, pKa of 4.75, and log Kow of 5.01 (Ferro et al., 1994; Joel and Jerald, 1998). 2.2. Sorption experiments Sorption isotherms. Sorption was conducted in glass screw-PTFE-cap centrifuge vials. About 10–1000 mg (on
Y. He et al. / Chemosphere 65 (2006) 497–505
499
Table 1 Characteristics of minerals and humic acids, Freundlich isotherm parameters, and sorption distribution coefficient (Kd) for PCP sorption Sorbents
OCa (g kg1)
SA (m2 g1)
CEC (cmol (+) kg1)
pHb
Sorption isotherm parameters KF
Mineral Goethite Kaolinite Ca-Montmorillonite K-Montmorillonite Humic acids HAs1 HAs2 a b c
1.03 2.58 2.55 1.06 368.27 574.20
N
R2c
Kd
59 18 635 644
3.1 7.8 95.0 92.0
7.91 8.55 8.37 8.35
8.46 0.85 34.19 193.69
0.9391 1.3890 0.4597 0.6895
0.9792** 0.9882** 0.9848** 0.9954**
6.85 2.43 10.84 127.13
– –
– –
5.26 4.94
3412.39 5666.31
0.7695 0.7400
0.9903** 0.9862**
2842.59 4828.54
OC: organic carbon; SA: surface area; CEC: cation exchange capacity. The pH of the equilibrium solution after PCP sorption. ** Correlation is significant at 0.01 probability level.
a dry-weight basis) of the sorbents was added, followed by adding 10 ml of different PCP-spiked level background electrolyte solutions containing 1 mM CaCl2, 0.1 mM MgCl2, 0.5 mM Na2B4O7 Æ 10H2O (Kan et al., 1997), with 200 mg l1 NaN3 to inhibit aerobic biodegradation. The solid to liquid ratio was adjusted to achieve 30–70% of total solute sorbed. Solutes were added via methanol carrier, keeping the methanol percentage in the range of 0.04–1%. Although methanol content varied slightly over an isotherm, there was no effect on bulk solution properties. Each series comprised 15 samples. For each batch experiment, one series of vials without sorbent as control were also prepared and monitored. The control series did not indicate any significant PCP degradation or sorption losses on the glassware during the course of the experiment. Average PCP losses in the control series reached 2.57% within the experiment period, thus the missing PCP from the solution phase can be considered safely sorbed onto the solid phase. The vials were shaken on an orbital shaker at 200 rpm at 20 ± 2 C for 24 h in the dark. Preliminary tests indicated that 24 h was sufficient to reach apparent equilibrium. After equilibration, the vials were centrifuged at 1932g for 20 min. An aliquot of 1 ml of the supernatant liquid was filtered through a syringe of 0.22 lm Millipore membrane (ANPEL, B13 mm) prior to determination. HPLC analysis was carried out on a Waters Alliance 2695-2487 HPLC system fitted with a Symmetry C18 (5 mm, 3.9 · 150 mm, Waters, USA). Chromatography was performed at 30 C. The flow rate was maintained at 1.00 ml min1 by a gradient controller and solvent delivery system. The mobile phase was methanol:1% acetic acid 9:1 (V/V), and the injection volume was 20 ll. PCP was detected at 220 nm, and readings were integrated by the Millennium32 software system (Waters, USA). The PCP concentration was quantified with an external standard method. The detection limits of the methods are 5 lg l1. Sorbed mass was calculated by the difference between total and solution-phase mass. Sorption–desorption hysteresis. Cycle of sorption and desorption consisting of sorption followed by a single-dilution desorption step was conducted to investigate the sorp-
tion–desorption hysteresis. Sorption was performed as above. Concurrently, after the sorption period, desorption was achieved by withdrawing 96% of the supernatant and replenishing the samples with fresh PCP-free background electrolyte solution (actual amounts were determined by weight), and the vials were sealed and shaken for 24 h at 200 rpm at 20 ± 2 C in the dark, then centrifuged at 1932g for 20 min, and the supernatant was analyzed for solution-phase PCP concentration by HPLC. 2.3. Data analysis All sorption data were fitted to the logarithmic form of the Freundlich equation: log qe ¼ log K F þ N log C e
ð1Þ
where qe is the solid-phase concentration (mg kg1) and Ce is the liquid-phase equilibrium concentration (mg l1). The parameters KF for sorption capacity coefficient [(mg kg1)/ (mg l1)N] and N (dimensionless) indicating isotherm nonlinearity were determined by linear regression of log-transformed data. The isotherm data of the experiments with pure minerals and HAs were also analyzed assuming a linear sorption isotherm Kd ¼
qe Ce
ð2Þ
where Kd [l kg1] is the sorption distribution coefficient, qe [mg kg1] is the equilibrium sorbed concentration, and Ce [mg l1] is the equilibrium solution phase concentration. The sorption–desorption hysteresis was quantified for each sorbent-solute-solution system using the hysteresis index (HI) defined by Huang et al. (1998) and Weber et al. (1998) qd qs HI ¼ e s e ð3Þ qe T ;Ce where qse and qde refer to the solid-phase solute concentration for the single-cycle sorption and desorption experiment, respectively, and the bracket sub-scripts T and Ce
500
Y. He et al. / Chemosphere 65 (2006) 497–505
specify conditions held constant, i.e., temperature and residual solution phase concentration.
and 0.6895), the homoionic montmorillonite with expandable 2:1 type structure may have a more heterogeneous structure or composition than the kaolinite (1.3890).
3. Results and discussion 3.1. Sorption isotherm of pure minerals and HAs
3.2. Potential contributions of pure minerals and HAs on PCP sorption respectively
The low solubility and high sorption potential of PCP caused a problem in using a broad range of different initial concentrations. A low initial PCP concentration would leave an undetectable amount of PCP in solution. The result was a very narrow working PCP concentration range, from which the isotherm might just represent a small portion of a truly isotherm. One way to circumvent this problem is to use a suitable background electrolyte solution to increase the solubility of PCP as much as possible. For this reason, we compared lots of empirical background electrolyte solution in traditional sorption studies and at last the method of Kan et al. (1997) was chosen. This would allow for equilibrium aqueous concentrations of PCP close to the solubility limit, and two orders of the concentration magnitude spanning in the sorption isotherms could then be achieved, thus confirming a reliable isotherm. Typical sorption isotherms of PCP on pure minerals and humic acids were fitted to the logarithmic form of the Freundlich equation. Linear fitting of log-transformed data was justified over direct nonlinear curve fitting in this paper for two reasons: (1) concentrations were spread evenly over the log scale, thus, nonlinear curve fitting would underestimate the importance of the low concentration data, and (2) the relative uncertainty in the measurement was not greatly dependent on concentration (Xing, 2001). As plotted in Fig. 1, each of the six isotherms contained 15 points. The Freundlich model captured the inflection well, giving a good fit to individual isotherm (Table 1). According to Table 1, N values were in general less than 1, resembling L-type isotherms (Giles et al., 1960). A L-type curve indicated a relatively high PCP-sorbent affinity at low concentration; that is, the extent of sorption decreased as the PCP concentration increased, because it became increasingly more difficult for PCP to find vacant adsorption sites. Only the kaolinite isotherms showed the value of N > 1, indicating an S-type curve. This shape indicated a low PCP-kaolinite affinity at low concentration. Sorption became easier as the PCP concentration increased. With the 1:1 type of clay minerals and equaled abundance of Si and Al sites, this behavior in most cases might be attributed to the strong competition between the water molecules and PCP for the sorption sites. In certain cases, especially in the sorption of organic contaminants, a S-type isotherm might be due to cooperative interactions among adsorbed organic species that stabilized the sorbate and enhanced its affinity for the surface (Sposito, 1984). Weber et al. (1992) suggested that N value can be taken as an index of site energy distribution (i.e., the smaller the N, the more heterogeneous the sorption site). Therefore, among the clay minerals, with smaller N values (0.4597
The sorption coefficient (KF) ranged from 0.85 for kaolinite to 5666 for HAs2. However, precise comparison could not be made between KF values because of their different units (Table 1) as a result of nonlinearity (Chen et al., 1999). Therefore, the linear sorption Kd-values were calculated from the solution equilibrium concentration range where the curvature in the Freundlich equation was insignificant with 95% confidence limits and where the residuals were normally distributed about zero (Clausen et al., 2004). The results showed that the affinity and the sorption capacity of the tested sorbents for PCP increased in the order of kaolinite < goethite < Ca-montmorillonite K-montmorillonite HAs1 < HAs2. Both kaolinite and goethite showed a lowest sorption capacity, which were much smaller than that of HAs and K-montmorillonite. This might indicate that the mineral surfaces of both kaolinite and goethite could therefore only be of limited importance for the sorption of PCP. It could also be postulated that the soil under deep weathering condition might be likely to have poor sorption capacity due to blockage of reactive sites on clay minerals by the presence of oxidized minerals (e.g., iron oxides). Table 1 also illustrated the effect of adsorbed cations on the ability of the clays to adsorb PCP. Obviously, the type of cations adsorbed to clay minerals was a crucial factor that determined their affinity for PCP. Sorption of PCP to clays was high when the exchangeable cations at the clays were K+, but was negligibly small for homoionic Ca2+. Similar effects were observed by Haderlein and Schwarzenbach (1993) and Haderlein et al. (1996) for the adsorption of nitroaromatics by clays saturated with different cations. This appeared to be a manifestation of the comparatively weak hydration of K+. The enthalpy of hydration for K+ was 314 kJ mol1, much smaller than that of Ca2+ (1580 kJ mol1). Uncharged siloxane surfaces between charged sites on montmorillonite surface were relatively hydrophobic and could interact with nonpolar or less polar moieties of organic contaminants when they were able to access these mineral surfaces (Jaynes and Boyd, 1991; Laird et al., 1992; Laird and Fleming, 1999). In contrast, when montmorillonite was saturated with strongly hydrated divalent cations (e.g., Ca2+), the hydration sphere surrounding exchangeable cations diminished the size of adsorptive domains between cations (Haderlein et al., 1996; Weissmahr et al., 1997; Weissmahr et al., 1998; Sheng et al., 2002) and reduced the strength of interactions between exchangeable ions and polar functional groups in organic contaminants (Boyd et al., 2001; Johnston et al., 2001, 2002; Li et al., 2003). Differences in swelling behavior between K- and Ca-montmorillonite might also contribute
Y. He et al. / Chemosphere 65 (2006) 497–505
300
501
200
160 Goethite
Ca-montmorillonite
Kaolinite
250 120
150
200 35 80
Equilibrium sorbed concentration (mg kg-1)
150 25
100
100
6 4
35
40 50 0
15 0
0
20
2
40
50 2
4 0
60
0
20
K-montmorillonite
2000
55
0
2.5
40
60
15
5 80
HAs1
0
0
10
20
30
4
40
50
25000
1600 15000
20000 4000
1200 200
10000
2000
15000
1000
10000
800 100
2000
5000
400 0 0
2
HAs2
30000
20000
0
0
10
20
0
1 30
0
2 40
50
0
0
5
10
5000 0
0.3 15
20
0
0.6 25
0
0
5
10
0
0.2 15
20
0.4 25
Equilibrium solution phase concentration (mg l-1) Fig. 1. Sorption isotherms of pentachlorophenol on minerals and humic acids spanning two orders of magnitude in concentration. The symbols denote experimental datum points. Low concentration range datum points are shown in the insets for clear isotherm display.
to their differential PCP adsorption. For homoionic Kmontmorillonite, the basal spacing was often observed at ˚ , which appeared optimal for adsorption of organic 12.3 A contaminants (Boyd et al., 2001; Sheng et al., 2002). This spacing might just be large enough to allow intercalation of the PCP, while minimizing its interaction with water molecules. This energetically favorable process occurred when the PCP solute directly contacted the opposing clay siloxane surfaces. Larger interlayer spacing associated with Ca2+ or multivalent exchangeable cations did not allow the partial dehydration of organic solute in the clay interlayers (Boyd et al., 2001; Johnston et al., 2002; Sheng et al., 2002; Li et al., 2003). In addition, as shown in Table 1, the sorption capacity of HAs for PCP was significantly high, with the Kd value of PCP on both humic acids being 100 times greater than that of pure minerals on average. Of the HAs and solution parameters studied in this work, the system pH might emerge as a primary factor that controlled the sorption behavior of PCP by affecting the surface characteristics of the HAs and the distribution of PCP species in solution. The equilibrium pH for both the HAs after PCP sorption was 5.26 and 4.94, approaching nearly the pKa value of
organic matter (5, Pusino et al., 2000) and PCP (4.75, Ferro et al., 1994), under which cases both the organic matter and the PCP molecule might be only imperceptibly partly dissociated and thus maximize extremely the PCP sorption level of HAs. With a functional group of OH, PCP might dissociate over the range of equilibrium pH and be present as two chemical species with very different physico-chemical properties. Results presented by Lee et al. (1990) suggested that for pH less than 7, PCP sorption increased with an increase in the fractions of the PCP neutral species. Simultaneously, the component of aromatics in HAs allowed the neutral PCP to sorb through hydrophobic bond formation. And the conjugation between neutral PCP and carbohydrates, and the formation of intermolecular hydrogen bonds between PCP and COOH groups were also likely involved (Paaso et al., 2002). Therefore, the pHdependent chemical characteristics of PCP and the surface properties of HAs co-contributed to the complexity of PCP behavior in the HAs–liquid interface, and thus could result in the high sorption capacity observed. Therefore, it can be prospected that the sorption of PCP on soils/sediments is significantly influenced by the SOM. The clay minerals also have their contribution especially
Data referred to Xie et al. (2003). CEC: cation exchange capacity; FA-C: fulvic acid carbon; HA-C: humic acid carbon; HM-C: humin carbon; TOC: total organic carbon; H2O2-BF: before H2O2-treatment; H2O2-AF: after H2O2treatment.
a
4 5 6
b
5.2 4.7 7.5 5.6 5.3 9.0 0.9 1.8 3.4 6.6 11.0 5.5 75.4 34.5 4.6 7.4 31.0 71.1 17.2 34.5 24.3 1.6 5.9 3.9 2.8 2.1 0.9 2.3 2.9 0.7 12.3 14.9 7.1 Kaolinite Kaolinite Illite Red sandstone Acid magmatite Marine deposit
6.2 6.5 5.8 3.2 2.9 1.5 18.3 24.3 17.9 3.0 9.3 11.6 57.0 46.4 48.0 40.0 44.3 40.4 7.7 15.7 9.4 5.8 5.6 6.1 4.8 2.9 2.5 28.5 27.6 22.0 Illite Montmorillonite Illite
Fluviogenic loamy fine paddy soil Paddy field on bole soil Paddy field on redeposit of purple sandstone soil Paddy field on red sandstone soil Yellow–red soil Coastal saline soil 1 2 3
Fluviatile deposit Basalt Purple sand shale
H2O2-BF H2O2-BF HA-C
HM-C
Silt Clay FA-C
Sand
H2O2-AF
pH Particle size (%) CECb (cmol (+) kg1) Chief mineralsa Parent materiala Soil name
Table 2 Selected properties of the soil samples and their detectable changes before and after H2O2-treatment
Additional experiment was carried out by the analysis of PCP sorption behaviors on soils before and after H2O2-treatment, with the intention to give a further illustration on the impact factors of PCP retention. The samples used were six referenced soils (Xie et al., 2003) with contrasting physico-chemical properties, among which the chief clay minerals for soils 1, 3 and 6 were the same illite component with K-enriched nonexpandable 2:1 type structure; for soil 2 was 2:1 type expandable montmorillonite; and that for soils 4 and 5 were 1:1 type of kaolinite (Table 2). Adding H2O2 seemed a violent treatment of the soil, but no significant difference in soil composition or specific surface area was involved (Clausen et al., 2004). The detected changes only referred to the soil total organic carbon (TOC) and pH (Table 2). This might be resulted from the oxidation of SOM and thus a loss of aromatics and an increase in oxygen containing groups (Forsberg and Bjoroy, 1981), which resulted likely in a higher O/C ratio and a lower content of aromatics in soil component, and therefore, a more hydrophilic surface with H+ gathering. The two indexes were approximately parallel before and after H2O2-treatment; with a great reduction of the Kd values. Because of the concentration-dependent effect of the sorption distribution coefficient of our early study results (He et al., 2006), the Kd values for comparison here were calculated from the representative initial equilibrium concentrations in the linear range of the dual-mode model. The H2O2 decreased 7.14–26.15% of pH and removed, on average, about 80% of TOC (Table 2), resulting in a probably alternative variation to the sorption capacity (in this case, the Kd value) for PCP. One was the reduction with the remove of SOM, and the other was the increase with the decrease of pH, as discussed in the above section. From Fig. 2 however, it could be see that the total variation in sorption capacity presented a reduction of more than 50%. This might indicate properly the neglectable effect of pH decrease if at most a 26.15% range on PCP sorption when compared with that of SOM. Results presented by Lee et al. (1990) also suggested that regardless of the pH dependent mechanisms,
Humus fractions (g kg1)
3.3. Sorption and desorption hysteresis on soils before and after H2O2-treatment
TOC (g kg1)
H2O2-AF
in the case when K+ dominates the exchangeable cation on the clay surfaces. And both the two mechanisms are pH-dependent. It had been shown in the early study (Sheng et al., 2001) that some organic contaminants, e.g., DNOC, carbaryl, and dichlobenil, were more effectively sorbed by K-clay than by SOM. Others such as diuron, parathion, and biphenyl were sorbed more effectively by SOM than by clay. Atrazine was sorbed by K-clay and SOM to a similar extent. Hence, the main factors that influenced the sorption behavior of organic contaminants varied with their types, which reflected rightly the complexity of organic contaminant-clay interactions.
4.8 4.8 4.9
Y. He et al. / Chemosphere 65 (2006) 497–505
Soil no.
502
Y. He et al. / Chemosphere 65 (2006) 497–505
methods. Also, pH dependent effect could be concluded if taking a further analysis as for soil 6. The removal extent of TOC was actually low, thus the increment of HI was possibly related to the change of pH value. It was reduced by 1.5 pH units, from 9.0 to 7.5 after the H2O2-treatment. These findings testified rightly the coincident results obtained in the above study of pure minerals and HAs. Knowledge on how SOM and mineral composition affected sorption as a function of pH was therefore a likely key parameter for the prediction of the fate of PCP in soils and sediments.
60 Before H2O2-Treatment After H2O2-Treatment
50
K d (l kg -1)
40
30
3.4. Environmental significance and applications
20
10
0
503
1
2
3 4 Soil No.
5
6
Fig. 2. Sorption distribution coefficients (Kd-values) for pentachlorophenol on the representative six soils measured at equilibrium concentration of the linear range of dual-mode model before and after removal of organic carbon by H2O2-treatment. The error bars represent the standard deviation.
sorption increased with increasing soil organic carbon. Thus, in this case, considering the limited contribution to sorption from any of the different surface functional groups of the investigated minerals except K-montmorillonite, it is more likely that the sorption capacity in the investigated soils was related to the organic carbon content. Meanwhile, for the convenience of comparing the irreversibility of sorption–desorption, the hysteresis index (HI) was applied. In the present study, the sorption– desorption isotherms were constructed under conditions that rule out artificial causes of hysteresis. The values of HI at 20 C and five representative concentration levels before and after H2O2-treatment were employed (Fig. 3). As shown in Table 2 and Fig. 3, HI generally decreased after the treatment of H2O2, especially for the soils 1, 2, 4 and 5. With the 86% removal and nearly the zero content of TOC, PCP desorption was completely reversible in soil 4 after treated with H2O2 (a zero or negative HI value indicates that sorption–desorption hysteresis is statistically insignificant). Although it had a great reduce of both TOC and pH values, soil 1 showed a steady but low decline leading to the inverse 10–15% increase of HI at the highest concentrations. This might be attributed to the K-enriched nonexpandable 2:1 type clay minerals with the effect of interacting irreversibly with the PCP and thus increasing its retention in soil. Note worthy was that the soils 3 and 6, of which the values increased unexpectedly at the majority of the concentrations tested. This might be explained by possible minor surface changes in clay and oxide minerals, which might not be detected by the used characterization
In considering the applicability of these results to real soils, simple ion exchange processes might be useful in the development of an environmental-friendly protocol to control the sorption, mobility, and bioavailability of organic contaminants in montmorillonite-containing soils or soils amended with montmorillonite clays. In this scenario, addition of simple electrolyte solutions could be used to manipulate the type and composition of exchange cations associated with clays, thereby modulating the release and immobilization of contaminants. Although K+ is not normally the dominant exchangeable cation, it is one of the most abundant exchangeable cations in soils and subsoils. And the phenomenon of cation demixing in clays (Levy and Francis, 1975) results in entire interlayer regions being occupied by a single exchangeable cation (e.g., K+) even in the presence of other cations (e.g., Ca2+) because of differences in cation hydration. The widely occurring phenomenon of K-fixation by soil clays (Eberl, 1980; Bouabid et al., 1991) proves that K-rich domains are common indeed. Thus, improved understanding of organic contaminants adsorption by clays makes it possible to enhance the retentive properties of subsoils and aquifer materials for certain organic contaminant by controlling the base saturation of the matrix via in situ electrolyte injection (e.g., KCl) as suggested by the recent findings of Weissmahr et al. (1999). Also, the application using K-montmorillonite as a carrier in pesticide formulations in order to extend the efficacious period is possible. Pesticide would be gradually released as Ca2+ and Mg2+ in soils gradually replaced K+ on the clay. An interesting feature of both H2O2-treated and untreated soils is that hysteresis coefficients decreased generally with the increasing PCP loading concentration, indicating that desorption hysteresis is favored and the binding sites with high-energy are occupied firstly by PCP at low sorption levels. Thus, it could be concluded that if the low polluted sites were involved such as aquifers, PCP could undergo natural sequestration through the favored sorption–desorption hysteresis processes and restore the polluted environment to clean naturally. While if in the case of a serious pollution, other remediation processes must be considered, e.g., the application of organic matter and K-enriched clays from extraneous sources.
504
Y. He et al. / Chemosphere 65 (2006) 497–505
4
4
Hysteresis index (HI)
Soil 1
4 Soil 2
Soil 3
3
3
3
2
2
2
1
1
1
0
0
0
3
4 Soil 4
4 Soil 5
Soil 6
2
3
3
1
2
2
0
1
1
-1
0.1
0.2 0.4
0.6
0.8
0
0.1
0.2
0.4
0.6
0.8
0
0.1
0.2
0.4
0.6
0.8
Equilibrium concentration (mg l-1) Fig. 3. Hysteresis index (HI) for pentachlorophenol sorption on the six soils at five representative concentrations before and after removal of the organic carbon by H2O2-treatment. Symbols: (j) before H2O2-treatment; (h) after H2O2-treatment.
4. Conclusions The results of this study form a basis for understanding and quantifying the sorption behavior of PCP in soil/sediment environments. Specific sorption of PCP to clay minerals and its implications, particularly the effects of exchangeable cations at clay minerals, effects of HAs and its crucial controlling behaviors need to be considered and have been discussed in this paper. However, more work is necessary to evaluate the effect of other factors on PCP sorption in complex natural matrices such as the role of natural organic matter or other constituents that may influence the accessibility of clay mineral surfaces. In addition, sorption kinetics and competition behavior of PCP need to be re-evaluated in such a complex system, e.g., by batch and miscible displacement experiments. Acknowledgments This work was jointly supported by the National Nature Science Foundation for Distinguish Young Scholars of
China (40425007), and the Teaching and Research Award Program for Outstanding Young Teachers in Higher Education Institutions of China. References Abrahamsson, K., Klick, S., 1991. Degradation of halogenated phenols in anoxic marine sediments. Mar. Pollut. Bull. 22, 227–233. Bailey, G.W., White, J.L., 1970. Factors influencing the adsorption, desorption, and movement of pesticides in soils. In: Gunther, F.A., Gunther, J.D. (Eds.), Residue Review—Residues of Pesticides and Other Foreign Chemicals in Food and Feeds, vol. 32. Springer-Verlag, New York, pp. 29–92. Boily, J.F., Lu¨tzenkirchen, J., Balme`s, O., Beattie, J., Sjo¨berg, S., 2001. Modeling proton binding at the goethite (a-FeOOH)–water interface. Coll. Surf. 179, 11–27. Bouabid, R., Badraoui, M., Bloom, P.R., 1991. Potassium fixation and charge characteristics of soil clays. Soil Sci. Soc. Am. J. 55, 1493– 1498. Boyd, S.A., Sheng, G., Teppen, B.J., Johnston, C.T., 2001. Mechanisms for the adsorption of substituted nitrobenzenes by smectite clays. Environ. Sci. Technol. 35, 4227–4234. Chen, Z., Pawluk, S., 1995. Structural variations of humic acids in two Alberta Mollisols. Geoderma 65, 173–193.
Y. He et al. / Chemosphere 65 (2006) 497–505 Chen, Z., Xing, B., McGill, W.B., 1999. A unified sorption variable for environmental applications of the Freundlich equation. J. Environ. Qual. 28, 1422–1428. Clausen, L., Larsen, F., Albrechtsen, H.J., 2004. Sorption of the herbicide dichlobenil and the metabolite 2,6-dichlorobenzamide on soils and aquifer sediments. Environ. Sci. Technol. 38, 4510–4518. Eberl, D.D., 1980. Alkali cation selectivity and fixation by clay minerals. Clays Clay Miner. 28, 161–172. Ferro, A.M., Sims, R.C., Bugbee, B., 1994. Hycrest crested wheatgrass accelerates the degradation of pentachlorophenol in soil. J. Environ. Qual. 23, 272–279. Forsberg, A., Bjoroy, M., 1981. In: Bjoroy, A. (Ed.), Advances in Organic Geochemistry. Wiley, New York, pp. 60–68. Giles, C.H., MacEwan, T.H., Nakhwa, S.N., Smith, D., 1960. Studies in adsorption: Part XI. A system of classification of solution adsorption isotherms, and its use in diagnosis of adsorption mechanisms and in measurement of specific surface areas of solids. J. Chem. Soc., 3973– 3993. Haderlein, S.B., Schwarzenbach, R.P., 1993. Adsorption of substituted nitrobenzenes and nitrophenols to mineral surfaces. Environ. Sci. Technol. 27, 316–326. Haderlein, S.B., Weissmahr, K.W., Schwarzenbach, R.P., 1996. Specific adsorption of nitroaromatic explosives and pesticides to clay minerals. Environ. Sci. Technol. 30, 612–622. He, Y., Xu, J.M., Wang, H.Z., Ma, Z.H., 2006. Detailed sorption isotherms of pentachlorophenol on soils and its correlation with soil properties. Environ. Res., in press. Hendershot, W.H., Duquette, M., 1986. A simple barium chloride method for determining cation exchange capacity and exchangeable cations. Soil Sci. Soc. Am. J. 50, 605–608. Huang, W., Yu, H., Weber Jr, W.J., 1998. Hysteresis in the sorption and desorption of hydrophobic organic contaminants by soils and sediments. 1. A comparative analysis of experimental protocols. J. Contam. Hydrol. 31, 129–148. Jaynes, W.F., Boyd, S.A., 1991. Hydrophobicity of siloxane surface in smectites as revealed by aromatic hydrocarbon adsorption from water. Clays Clay Miner. 39, 428–436. Joel, G.B., Jerald, L.S., 1998. Predictive relationships for uptake of organic contaminants by hybrid poplar trees. Environ. Sci. Technol. 32, 3379–3385. Johnston, C.T., De Oliveira, M.F., Teppen, B.J., Sheng, G., Boyd, S.A., 2001. Spectroscopic study of nitroaromatic-smectite sorption mechanisms. Environ. Sci. Technol. 35, 4767–4772. Johnston, C.T., Sheng, G., Teppen, B.J., Boyd, S.A., De Oliveira, M.F., 2002. Spectroscopic study of dinitrophenol herbicide sorption on smectite. Environ. Sci. Technol. 36, 5067–5074. Kan, A.T., Fu, G.M., Hunter, M.A., Tomson, M.B., 1997. Irreversible adsorption of naphthalene and tetrachlorobiphenyl to lula and surrogate sediments. Environ. Sci. Technol. 31, 2176–2185. Karickhoff, S.W., 1984. Organic pollutant adsorption in aqueous systems. J. Hydraul. Eng. 110, 707–735. Kile, D.E., Chiou, C.T., Zhou, H., Li, H., Xu, O., 1995. Partition of nonpolar organic pollutants from water to soil and sediment organic matters. Environ. Sci. Technol. 29, 1401–1406. Laird, D.A., Fleming, P.D., 1999. Mechanisms for adsorption of organic bases on hydrated smectite surfaces. Environ. Toxicol. Chem. 18, 1668–1672. Laird, D.A., Barriuso, E., Dowdy, R.H., Koskinen, W.C., 1992. Adsorption of atrazine on smectites. Soil Sci. Soc. Am. J. 56, 62–67. Lee, L.S., Rao, P.S.C., Nkedi-Kizza, P., Delfino, J.J., 1990. Influence of solvent and sorbent characteristics on distribution of pentachlorophenol in octanol–water and soil–water system. Environ. Sci. Technol. 24, 654–666. Levy, R., Francis, C., 1975. Demixing of sodium and calcium ions in montmorillonite crystallites. Clays Clay Miner. 23, 475–476.
505
Li, H., Sheng, G., Teppen, B.J., Johnston, C.T., Boyd, S.A., 2003. Sorption and desorption of pesticides by clay minerals and humic acid–clay complexes. Soil Sci. Soc. Am. J. 67, 122–131. Maatela, P., Paasivitra, J., Sarkka, J., Paukku, R., 1990. Organic chlorine compounds in lake sediments: II. Organically bound chlorine. Chemosphere 21, 1343–1354. Paaso, N., Peuravuori, J., Lehtonen, T., Pihlaja, K., 2002. Sedimentdissolved organic matter equilibrium partitioning of pentachlorophenol: the role of humic matter. Environ. Int. 28, 173–183. Pusino, A., Braschi, I., Gessa, C., 2000. Adsorption and degradation of triasulfuron on homoionic montmorillonites. Clays Clay Miner. 48, 19–25. Sanjoy, K.B., Qiang, Y., Peikang, J., 1996. Removal of pentachlorophenol from wastewater by combined anaerobic–aerobic treatment. J. Hazard. Mater. 49, 143–145. Sawhney, B.L., Singh, S.S., 1997. Sorption of atrazine by Al- and Casaturated smectite. Clays Clay Miner. 45, 333–338. Sheng, G., Johnston, C.T., Teppen, B.J., Boyd, S.A., 2001. Potential contributions of smectite clays and organic matter to pesticide retention in soils. J. Agric. Food Chem. 49, 2899–2907. Sheng, G., Johnston, C.T., Teppen, B.J., Boyd, S.A., 2002. Adsorption of dinitrophenol herbicides from water by montmorillonites. Clays Clay Miner. 50, 25–34. Sposito, G., 1984. The Surface Chemistry of Soils. Oxford University Press, New York, pp. 1–46. Sung-Kil, P., Angela, R.B., 2003. Aqueous chemistry and interactive effects on non-ionic surfactant and pentachlorophenol sorption to soil. Water Res. 37, 4663–4672. US Environmental Protection Agency, 1980. Health and Environmental Effects Profile No. 135, Washington, DC. Weber, W.J., Huang, W., 1996. A distributed reactivity model for sorption by soils and sediments. 4. Intraparticle heterogeneity and phasedistribution relationships under nonequilibrium conditions. Environ. Sci. Technol. 30, 881–888. Weber, W.J., McGinley, P.M., Katz, L.E., 1992. A distributed reactivity model for sorption by soils and sediments. 1. Conceptual basis and equilibrium assessments. Environ. Sci. Technol. 26, 1955– 1962. Weber Jr, W.J., Huang, W., Yu, H., 1998. Hysteresis in the sorption and desorption of hydrophobic organic contaminants by soils and sediments. 2. Effects of soil organic matter heterogeneity. J. Contam. Hydrol. 31, 149–165. Weissmahr, K.W., Haderlein, S.B., Schwarzenbach, R.P., Hany, R., Nuesch, R., 1997. In situ spectroscopic investigations of adsorption mechanisms of nitroaromatic compounds at clay minerals. Environ. Sci. Technol. 31, 240–247. Weissmahr, K.W., Haderlein, S.B., Schwarzenbach, R.P., 1998. Complex formation of soil minerals with nitroaromatic explosives and other ðacceptors. Soil Sci. Soc. Am. J. 62, 369–378. Weissmahr, K.W., Hilderbrand, M., Schwarzenbach, R.P., Haderlein, S.B., 1999. Laboratory and field scale evaluation of geochemical controls on groundwater transport of nitroaromatic ammunition residues. Environ. Sci. Technol. 34, 2593–2600. Xie, Z.M., Wu, W.H., Xu, J.M., 2003. Study on fluoride emission from soils at high temperature related to brick-making process. Chemosphere 50, 763–769. Xia, G., Ball, W.P., 1999. Adsorption-partitioning uptake of nine low polarity organic chemicals on a natural sorbent. Environ. Sci. Technol. 33, 262–269. Xing, B., 2001. Sorption of anthropogenic organic compounds by soil organic matter: a mechanistic consideration. Can. J. Soil Sci. 81, 317– 323. Xing, B., Pignatello, J.J., 1997. Dual-mode sorption of low-polarity compounds in glassy poly (vinyl chloride) and soil organic matter. Environ. Sci. Technol. 31, 792–799.