Modeling Analysis of Two Common Organic Pollutants' Adsorption and Desorption from Activated Carbon,C 60,and Soil Organic Carbon

Abstract

As one of the most important elements in the environment, carbon dictates the transport and fate of hydrophobic organic pollutants. Adsorption is the most common mechanism of organic pollutant attachment to carbonaceous materials, a process that can be illustrated by an adsorption isotherm. In this study, a series of adsorption–desorption experiments have been carried out to investigate the desorption hysteresis and modeling fitting of naphthalene and 1,2-dichlorobenzene with three types of carbonaceous materials, including fullerene C₆₀, activated carbon, and soil organic carbon. Sorption models evaluated include Freundlich, Langmuir, two-compartment dual-equilibrium desorption, and Polanyi–Manes isotherms. Results of model fitting of adsorption–desorption data with each isotherm were compared and discussed. Moreover, the possible mechanism controlling the adsorption process can be deduced according to the Polanyi–Manes model. This is the first report concerning the comparison of model fitting of these four different adsorption isotherms based on experimentally obtained organic pollutant adsorption–desorption data. This reported study can expand our knowledge of sorption and especially desorption characteristics of organic pollutants with different forms of carbonaceous materials.

Introduction

Carbon is one of the most important elements in the environment and is the fourth most abundant element in the universe by mass (Stumm and Morgan, 1996; Schwarzenbach et al., 2003). Various forms of carbonaceous materials, such as black carbons, soot, coals, and activated carbon (AC) have strong affinity for environmental pollutants, particularly hydrophobic organic pollutants, making these materials strong adsorbents and hence delivering vehicles for pollutants (Allen-King et al., 2002; Chiou, 2002; Sawyer et al., 2002; Cornelissen and Gustafsson, 2004; Bello et al., 2012; Abdul et al., 2015; Grzegorczuk-Nowacka and Anielak, 2017). As a matter of fact, the fate and transport of hydrophobic organic pollutants are mainly controlled by the presence of carbonaceous materials in the environment (Stumm and Morgan, 1996; Sawyer et al., 2002; Schwarzenbach et al., 2003). The first step toward the mobilization of environmental pollutants with carbonaceous materials is the attachment of pollutants to these materials. Adsorption is the adhesion of pollutant molecules to the surface of solid and is the most common mechanism of organic pollutant attachment to carbonaceous materials (Stumm and Morgan, 1996; Adamson and Gast, 1997; Kleineidam et al., 1999; Atkins and de Paula, 2001; Accardi-Dey and Gschwend, 2002; Sawyer et al., 2002; Schwarzenbach et al., 2003; James et al., 2005). In general, carbonaceous materials have a significant impact on the fate, transport, and behavior of hydrophobic organic pollutants in the environment. As for bulk carbonaceous materials, such as black carbons, the high surface area of the material provides enormous surface sites for a given mass of carbon (Chiou and Kile, 1998; Bucheli and Gustafsson, 2000). The organic pollutants will attach to the micropores of carbonaceous materials by physiochemical interactions. In addition, colloidal carbonaceous materials are ubiquitous in the environment and they can considerably enhance the pollutant adsorption, and hence its transport and bioavailability (Hiemenz and Rajagopalan, 1997; Evans and Wennerström, 1999; Clark, 2009). Since the discovery of fullerene (C₆₀) and carbon nanotubes, numerous carbonaceous nanomaterials and their derivatives have been manufactured to explore the functionality of these materials (Vollath, 2013). Introduction of these carbon nanomaterials raises significant concerns of the environmental impact of these nanomaterials, particularly their impact on the fate, transport, and behavior of environmental pollutants (Colvin, 2003). Extensive studies have been carried out to understand the fate, transport, and toxicity of these fabricated nanocarbon materials and their interactions with the environmental pollutants (Cheng et al., 2004, 2005a, 2005b; Sahu and Casciano, 2009; Wang et al., 2015).

The process of organic adsorption to carbonaceous materials is usually described by adsorption isotherm as the amount of adsorbate onto the surface of adsorbent as a function of concentration at constant temperature (Adamson and Gast, 1997; Atkins and de Paula, 2001). The sorption of hydrophobic organic contaminants by carbonaceous materials in soils and sediments is reported to be predominated by partitioning of dissolved solute between water and naturally occurring organic matter (Schwarzenbach et al., 2003). Therefore, linear partitioning models have been used to describe the sorption processes (Stumm and Morgan, 1996; Weber and Digiano, 1996; Sawyer et al., 2002; Schwarzenbach et al., 2003). However, this simple partitioning model has been shown to be inconsistent with the observed sorption behavior. For example, nonlinear sorption isotherms and hysteretic desorption are often observed in laboratory sorption studies (Kan et al., 1994, 1998; Weber and Huang, 1996; Huang et al., 1998). In these cases, the relationship between the adsorbed concentrations and dissolved concentrations of each hydrophobic organic contaminant cannot be described by a single linear isotherm. Therefore, more complicated models of Freundlich isotherm and Langmuir isotherm have been proposed to characterize the adsorption process with additional considerations of surface distribution of adsorption energies and limited adsorption capacities, respectively (Stumm and Morgan, 1996; Adamson and Gast, 1997; Atkins and de Paula, 2001; Sawyer et al., 2002). A major drawback of all of these more complicated adsorption models is that, to date, there is no theory capable of predicting the required constants or exponents. Therefore, each chemical and adsorbent combination must be measured experimentally (Schwarzenbach et al., 2003).

Researchers are also interested in the desorption process of the adsorbed hydrophobic organic pollutants from the adsorbent materials (Fu et al., 1994; Huang et al., 1998; Weber et al., 1998; Chen et al., 1999, 2000). Such process determines the kinetics and thermodynamics of the release of pollutants back into the environment after the adsorption process. Desorption hysteresis is the phenomenon that stems from the deviation of the desorption process from the adsorption process in terms of the shape of the isotherms during these two consecutive processes (Fu et al., 1994; Kan et al., 1994, 1998; Weber and Huang, 1996; Huang et al., 1998; Weber et al., 1998; Chen et al., 1999, 2000). The existence of desorption hysteresis is attributed to the fact that a certain amount of the adsorbed pollutants resists desorption from adsorbent. Adsorption hysteresis has been observed by many researchers on studying adsorption and desorption of various compounds to and from a wide range of adsorbent materials. Particularly, carbonaceous materials, such as ACs, have been extensively investigated concerning the desorption hysteresis with hydrophobic organic pollutants (Fu et al., 1994; Weber et al., 1998; Rathousky and Zukal, 2000; Braida et al., 2003). In light of the presence of desorption hysteresis, additional isotherms have been proposed to account for the irreversibility of desorption of organics from their corresponding adsorption process. Some notable isotherms derived to characterize desorption hysteresis include dual-equilibrium desorption (DED) model (Kan et al., 1998; Chen et al., 2002) and dual reactive domain model (Xing and Pignatello, 1996, 1997; Weber et al., 1998).

In the previous studies, desorption hysteresis was observed for the adsorption and desorption of two common hydrophobic organic pollutants, that is, naphthalene and 1,2-dichlorobenzene (1,2-DCB), with two forms of carbons, that is, C₆₀ and AC (Cheng et al., 2004, 2005b). In one study, the adsorption experiment employed C₆₀ particles in the form of large aggregates, thin film, and small aggregates, as well as AC (Cheng et al., 2004). In another study, organic adsorption experiment was carried out using nano-C₆₀ suspension (Cheng et al., 2005b). As a result, the obtained experimental data cannot be fitted well with a single linear isotherm, as a consequence of desorption hysteresis. In this study, a series of adsorption–desorption experiments were conducted using three different types of carbonaceous materials of AC, C₆₀, and soil organic carbon (OC). The AC and C₆₀ materials were further processed to produce nanomaterial suspensions for adsorption–desorption experiments. The hydrophobic organic compounds used in this study were naphthalene and 1,2-dichlorobenzene, two common organic contaminants. The obtained experimental data were fitted with a number of adsorption isotherms, including Freundlich, Langmuir, two-compartment DED, and Polanyi–Manes isotherms. Model fitting is compared for different models. To the best of our knowledge, this is the first report concerning the comparison of model fitting of aforementioned four different adsorption isotherms based on experimentally obtained organic pollutant adsorption–desorption data. This reported study can expand our knowledge of sorption and especially desorption characteristics of organic pollutants with different forms of carbonaceous materials.

Experimental Protocols

Materials

In this study, three carbon-based adsorbents were employed: AC, fullerene (C₆₀), and Anacostia river sediment (ARS). AC particles were originally purchased from Calgon Carbon Corp. (Pittsburgh, PA). C₆₀ solids (purity >99.5%) were purchased from SES Research (Houston, TX). ARS used in this study was obtained from Anacostia River in Washington D.C. area. The sediment sample was first dried at room temperature followed by pulverization with a pestle and mortar. The soil OC content of the ARS was measured to be ca. 3.7% (Shimadazu Corp.).

¹⁴C-radiolabeled naphthalene and ¹⁴C-radiolabeled 1,2-DCB were purchased from Sigma-Aldrich (St. Louis, MO) and diluted in methanol (high performance liquid chromatography grade) to make stock solutions. The specific activities of naphthalene and 1,2-DCB are 8.1 and 8.6 μCi/μmol, respectively. Liquid scintillation cocktail for scintillation counting was supplied by Beckman Coulter, Inc. (Fullerton, CA). Biological grade sodium chloride and sodium azide (>98%) were purchased from Fisher Scientific and Eastman Kodak, respectively. Sodium azide (0.02 M) was used to inhibit bacterial growth. Anodisc membrane filters (20 nm pore size; Whatman) were used to separate nanoscaled C₆₀ and nanoscaled AC particles from aqueous solution. The 20 nm membrane filter has a precise pore structure with no lateral crossovers between individual pores, allowing a sharp molecular weight cutoff. Deionized (DI) water was prepared by reverse osmosis followed by a four-stage ion exchange water purification process, consisting of a high-capacity cation/anion column, two ultra pure ion exchange columns, and an organics removal column (Barnstead Internationals, Dubuque, IA).

Adsorption and desorption experiments using AC

As-received AC particles

To study naphthalene adsorption and desorption with as-received AC particles, five sample vials were used, each filled with 40 mg of AC and 42 mL of DI water. ¹⁴C-radiolabeled naphthalene/methanol stock solution was injected into each sample vial with a microsyringe so that the initial naphthalene concentrations in the vials were 0.97, 1.94, 2.93, 4.86, and 9.55 μg/mL, respectively. The headspace in each vial was less than 0.1 mL. The advantage of injecting a naphthalene/methanol solution into DI water is to facilitate a uniform distribution of naphthalene into aqueous solution. Solution phase methanol volume fraction in each vial was less than 0.5%, which was not expected to materially impact the adsorption experiment. Another five control vials were set up the same way as the sample vials, except that no AC particle was added. Control vials were designed to account for the loss of naphthalene from the aqueous phase due to volatilization or adsorption to the vessels. The samples were then stirred mildly on magnetic stirrers for at least 3 days for the adsorption equilibrium in the dark at room temperature (25 ± 1°C). At the end of the adsorption experiment, samples were centrifuged at 6,000 rpm on a centrifuge and the supernatants were analyzed on a liquid scintillation counter (Beckman Coulter, Inc.) to determine the aqueous naphthalene concentrations. The scintillation counter detection limit for naphthalene is ca. 1.4 × 10⁻⁴ μg/mL. Over 95% of the supernatants were then removed from each sample vial. Subsequently, naphthalene-free DI water was filled into each vial (headspace <0.1 mL) to induce the first step of desorption. Multistep desorption experiments were induced in a similar manner by successively removing supernatants and refilling with DI water. At the end of each desorption step, solution-phase naphthalene concentrations were analyzed on the liquid scintillation counter. The equilibrium time for each desorption step varied from 3 days for the first desorption step to 14 days for the last desorption step. The control experiments suggest that about 7.5% of naphthalene was adsorbed onto the 20 nm membrane filters. Thus, the experimentally obtained sorption data were corrected for adsorption of organics to the 20 nm membrane filter.

Nano-AC particles

In the laboratory, the purchased AC particles were first added into DI water to a concentration of about 0.5 mg/mL, and the suspension was then subjected to sonication for about 1 – 2 h with a high-energy sonication probe (Sonifier Cell Disruptor; Heat Systems-Ultrosonics, Inc., Farmingdale, NY). A portion of the AC particles was broken into nanometer-sized particles after a prolonged sonication process. The resulting suspension was filtered through a 0.45 μm pore size membrane filter (Millipore Corp.). The nanometer-sized AC particles in the filtrate were used as adsorbent for sorption of naphthalene and are referred to as nano-AC particles. The concentration of the obtained nano-AC particles in water was measured by total organic carbon analyzer to be 12.35 mg/L.

Water suspensions containing ca. 0.2 mg of nano-AC particles were added into each of four 16 mL glass vials. ¹⁴C-radiolabeled naphthalene/methanol stock solutions were injected into each vial to make the initial naphthalene concentrations to be 0.45, 1.40, 1.87, and 2.36 μg/mL, respectively. Sample vials were closed tightly with a headspace of less than 0.1 mL. Methanol volume fraction in each vial was less than 0.2%, which was not expected to materially impact the adsorption experiment. Another four control vials were set up the same way as the sample vials, except that no nano-AC particle was added. Sample vials and control vials were stirred mildly on magnetic stirrers in the dark at room temperature (25 ± 1°C) for at least 3 days. At the end of the adsorption experiment, the suspension in each sample vial was filtered using the 20 nm membrane filter. One milliliter of filtrate was analyzed on the liquid scintillation counter to determine naphthalene aqueous concentration. About 1 mL of the solution in each control vial was also analyzed for aqueous-phase naphthalene concentrations by the liquid scintillation counter. The solid-phase naphthalene concentration in each sample vial was calculated from the difference between the solution-phase naphthalene concentration in the sample vial and that in the corresponding control vial. Another two sample vials, containing 0.1 mg of nano-AC particles in 8 mL water and 0.2 mg of nano-AC particles in 16 mL water, respectively, were set up the same way as that of the prior four samples. The initial naphthalene concentrations are 10.02 and 9.15 μg/mL, respectively. Upon completion of the adsorption, the suspension in each sample vial was filtered and the filtrates were analyzed for naphthalene aqueous concentrations, as discussed above. After filtration, the membrane filters containing AC solids were put into clean sample vials and naphthalene-free DI water was added into each of the clean vials to induce desorption for these two samples. The vials were then sealed tightly and put into a sonication bath (Solid State/Ultrasonic; Fisher Scientific) for 15 min to redisperse nano-AC particles into aqueous phase. Membrane filters were then removed carefully from the sample vials. Next, DI water was added to each sample vial (with headspace <0.1 mL) and samples were stirred in the dark at room temperature for 5 days for naphthalene desorption. At the end, the suspension in each sample vial was filtered with a 20 nm membrane filter and the aqueous-phase naphthalene concentration was determined by the liquid scintillation counter. Three steps of naphthalene desorption were conducted for each of the two samples. The retention of naphthalene to the 20 nm filter was accounted for by comparing the organic solution concentrations after being filtered through the filter with the stock solution concentrations. By the end of the desorption experiments, the values of the removal of naphthalene from the surface of nano-AC particles were ca. 63% and 50%, respectively, for the two desorption samples.

Adsorption and desorption experiments using C₆₀

The preparation of nanoscaled C₆₀ particles (hereafter referred to as nano-C₆₀, or nC₆₀) follows the sonication method, similar to the previous study (Cheng et al., 2005b). Essentially, a volume of C₆₀/toluene/DI water mixture was subjected to 4 h of continuous sonication by the use of a high-energy sonication probe, similar to the preparation of nano-AC particles. The resultant solution was in a yellowish color with the purple toluene layer evaporated. Subsequently, the nC₆₀ solution was produced by filtering the yellowish solution through a 0.45 μm membrane filter to remove large aggregates. The procedure of the nC₆₀ adsorption/desorption experiments with naphthalene and 1,2-DCB is detailed in Cheng et al. (2005b). Similar to naphthalene adsorption/desorption experiments with nano-AC particles, aliquots of ¹⁴C-radiolabeled naphthalene or ¹⁴C-radiolabeled 1,2-DCB were added to vials containing nC₆₀ solutions. Subsequently, adsorption and three-step desorption were conducted for each sample in the dark at room temperature (25 ± 1°C). The initial naphthalene concentrations were 1.10, 1.97, 2.20, 2.51, 3.30, 4.26, and 5.11 μg/mL, respectively. The initial 1,2-DCB concentrations were 0.89, 4.65, and 8.23 μg/mL, respectively. By the end of each adsorption experiment, nC₆₀ particles were separated from the solution using a 20 nm filter. Similar to the nano-AC experiments, control vials were employed to account for organic volatilization or adsorption to the vessels. The control experiments suggest that about 7.0% of 1,2-DCB was adsorbed onto the 20 nm membrane filters. Thus, the experimentally obtained sorption data were corrected for adsorption of organics to the 20 nm membrane filter. Moreover, the scintillation counter detection limit for 1,2-DCB is ca. 1.2 × 10⁻⁴ μg/mL.

Adsorption and desorption experiments using ARS

Similar to the aforementioned experiments using as-received AC particles, adsorption and desorption experiments were carried out using dried ARS powder. Two grams of dried ARS powder was added to a 40 mL glass vial and the vial was filled with 0.01 M NaCl and 0.01 M NaN₃ electrolyte solution (headspace <0.1 mL). ¹⁴C-radiolabeled naphthalene/methanol stock solution was injected into the vial to an initial concentration of 8.69 μg/mL. The sample was stirred mildly in the dark at room temperature (25 ± 1°C) for at least 3 days for adsorption. At the end of the adsorption experiment, the sample was centrifuged at 6,000 rpm on a centrifuge and the supernatant was analyzed on the liquid scintillation counter to determine the aqueous-phase naphthalene concentration. Over 95% of the supernatant was removed from the sample vial and naphthalene-free DI water was filled (headspace <0.1 mL) to induce multistep desorption. At the end of each desorption step, the sample was centrifuged and the aqueous-phase naphthalene concentration was analyzed on the liquid scintillation counter. The equilibrium time for each desorption step varied from 3 days for the first desorption step to 7 days for the last desorption step.

Sorption models

Freundlich sorption model

Freundlich model (isotherm) is an empirical relationship that is commonly used to fit experimental adsorption data with a minimum of adjustable parameters (Adamson and Gast, 1997; Atkins and de Paula, 2001; Schwarzenbach et al., 2003). It generally takes the following form: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} q = {K_F}C_w^n \tag{{ \rm Eq}. \ 1} \end{align*} \end{document}

where K_F [(μg/g) (μg/mL)⁻ⁿ] is the Freundlich sorption coefficient; n denotes the Freundlich exponent; and C_w (μg/mL) represents the adsorbate concentration in aqueous solution at equilibrium. The term q (μg/g) is the mass of adsorbed adsorbate per unit mass of adsorbent at equilibrium. The Freundlich model assumes a distribution of adsorption energies on the surface of adsorbent particles (Weber and Digiano, 1996; Adamson and Gast, 1997).

Langmuir sorption model

Langmuir model (isotherm) was originally applied to the adsorption of gases or vapors on a plane surface that contains a fixed number of identical active sites (Adamson and Gast, 1997; Atkins and de Paula, 2001; Schwarzenbach et al., 2003). The amount of the adsorbate increases monotonically until it reaches a limiting value that corresponds theoretically to the completion of a surface monolayer. In aqueous solutions, Langmuir model can be mathematically expressed as follows: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} q = { \frac { { q_m } b { C_w } } { 1 + b { C_w } } } \tag { { \rm Eq } . \ 2 } \end{align*} \end{document}

where b (mL/μg) is the Langmuir constant, q_m (μg/g) is the limiting (monolayer) adsorption capacity (the amount of the adsorbed adsorbate per unit mass of the adsorbent at the time when the adsorbent surface is covered with a complete monolayer of the adsorbate), and q (μg/g) and C_w (μg/mL) are defined as above.

Combined sorption model: DED model

A combination of two types of models has been proposed by a number of researchers to account for the observed sorption behavior (Kan et al., 1994, 1998; Xing and Pignatello, 1996, 1997; Chen et al., 2002). Kan et al. (1998) have proposed a “dual-equilibrium desorption” model (isotherm) to describe sorption–desorption hysteresis observed experimentally. The sorption of hydrophobic organic compounds to soils or sediments can be illustrated by two compartments with different equilibrium and kinetic characteristics. The first compartment characterizes mainly the adsorption and desorption at high initial concentrations. The second compartment accounts for the adsorption and desorption at lower concentration ranges. According to this model, adsorption/desorption of organic contaminants to/from adsorbents can be described by the following isotherm (Kan et al., 1998): \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} q = K_ { \rm oc } ^ { 1 { \rm st } } \cdot { f_ { \rm oc } } \cdot { C_w } + { \frac { K_ { \rm oc } ^ { 2 { \rm nd } } \cdot { f_ { \rm oc } } \cdot f \cdot q_ { \rm max } ^ { 2 { \rm nd } } \cdot { C_w } } { f \cdot q_ { \rm max } ^ { 2 { \rm nd } } + K_ { \rm oc } ^ { 2 { \rm nd } } \cdot { f_ { \rm oc } } \cdot { C_w } } } \tag { { \rm Eq } . \ 3 } \end{align*} \end{document}

where \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} $$K_{{ \rm{oc}}}^{{ \rm{1st}}}$$ \end{document} (mL/g) and \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} $$K_{{ \rm{oc}}}^{{ \rm{2nd}}}$$ \end{document} (mL/g) are the OC normalized sorption coefficients for the first and the second compartment, respectively; f_oc is the fraction of OC; \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} $$q_{{ \rm{max}}}^{{ \rm{2nd}}}$$ \end{document} (μg/g) is the maximum sorption capacity for the second compartment; f (0 ≤ f ≤ 1) is a factor representing the fraction of the second compartment that is saturated upon exposure; and q (μg/g) and C_w (μg/mL) are as defined above. As for adsorption and desorption of organic compounds with AC, nC₆₀, or OC in soil, f_oc = 1. Thus, Equation (3) becomes the following: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} q = K_ { \rm d } ^ { 1 { \rm st } } { C_w } + { \frac { K_ { \rm d } ^ { 2 { \rm nd } } \cdot f \cdot q_ { \rm max } ^ { 2 { \rm nd } } \cdot { C_w } } { f \cdot q_ { \rm max } ^ { 2 { \rm nd } } + K_ { \rm d } ^ { 2 { \rm nd } } \cdot { C_w } } } \tag { { \rm Eq } . \ 4 } \end{align*} \end{document}

Polanyi–Manes sorption model

The Polanyi adsorption potential theory (Polanyi, 1916; Adamson and Gast, 1997; Atkins and de Paula, 2001) has been considered the most powerful model (isotherm) in describing gas (or vapor) adsorption on energetically heterogeneous solids. Polanyi theory defines the existence of an (attractive) adsorption potential between the adsorbate molecule and the solid surface. At a particular location within the adsorption space, the adsorption potential may be viewed as the energy required to remove the molecule from that location to a point outside the attractive force field of the solid. According to the Polanyi theory, the following relationship holds for ideal gas (vapor) adsorption to porous adsorbents: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} \varepsilon = RT \ln { \frac { { P^0 } } { { P_g } } } \tag { { \rm Eq } . \ 5 } \end{align*} \end{document}

where ɛ (J/mol) represents the adsorption potential; P_g (Pa) is the vapor pressure in equilibrium with the adsorbed phase; P⁰ (Pa) is the saturation vapor pressure; R [J/(mol·K)] is the ideal gas constant; and T (K) is the absolute temperature. The original Polanyi model has been extended to a wide range of vapor-phase and liquid-phase systems by Manes (1998) and the effective adsorption potential can be defined as follows: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} \varepsilon = RT \ln \frac { S } { C } \tag { { \rm Eq } . \ 6 } \end{align*} \end{document}

where S (μg/mL) is the solubility of the adsorbate, and C (μg/mL) is the aqueous concentration of the adsorbate. The following empirically derived relationship between adsorbed volume and adsorption potential was proposed for sorption on ACs (Crittenden et al., 1987; Xia and Ball, 1999): \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} q ^\prime = { Q_0 } ^\prime \rho \times { 10^ { a ^\prime { { \left( { { \frac { { \varepsilon _ { } } } { { V_s } } } } \right) } ^ { b ^\prime } } } } \tag { { \rm Eq } . \ 7 } \end{align*} \end{document}

where ɛ (J/mol) is the net adsorption potential of the adsorbate in aqueous solution; Q₀′ (cm³/kg) is the maximum adsorption capacity; ρ (kg/cm³) is the liquid or solid density; a′ and b′ are fitting parameters; and V_s (cm³/mol) is the molar volume of the solute. Dubinin (1960) suggested that for a given adsorbent, a plot of the adsorbed volume (q′) against adsorption potential density (ɛ/V_s) would yield similar curves for a wide range of adsorbate chemicals and these curves were called “correlation curves.” The detailed descriptions of the DED and Polanyi–Manes sorption models can be found in the Supplementary Data.

Results and Discussion

Adsorption and desorption experimental results

The mean particle size of the as-received AC particles was reported to be about 1 μm by the supplier. Data of naphthalene adsorption to the AC particles are fitted with a Freundlich isotherm (Fig. 1) in the form of \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} $$q = {10^{5.12}}C_w^{0.87}$$ \end{document} , where the units of q and C_w are μg/g and μg/mL, respectively. As discussed above, the shape of the Freundlich isotherm suggests of a distribution of surface adsorption energy on the AC particles. It is speculated that this energy distribution originates from surface heterogeneity with varying energy on the surface of AC particles (Farrell and Reinhard, 1994). It has been reported that the reaction half-life of naphthalene desorption from sedimentary materials was less than 1.5 h (Karickhoff, 1980). In this study, the equilibrium time for each desorption step varied from 3 to 14 days. It is thus assumed that desorption equilibrium has been reached during each desorption step. Data of naphthalene desorption from as-received AC particles are also presented in Fig. 1. The difference of the desorption isotherms from the corresponding adsorption isotherm suggests the existence of hysteresis. As reported in a previous study investigating the desorption characteristics of naphthalene from carbon materials, a large fraction of the adsorbed naphthalene resisted desorption, leading to only a small amount of organics eventually being desorbed (Cheng et al., 2005b). A similar phenomenon was observed in this study that a limited amount of naphthalene was measured to be desorbed from AC particle surface (Fig. 1). For these samples, the solid–water distribution coefficient (K_d) values for the last desorption datum point for each sample are 10^6.67, 10^6.63, 10^6.37, 10^6.62, and 10^6.54 (mL/g), respectively. The desorption K_d value for each sample is more than one order of magnitude higher than the corresponding adsorption K_d values, which are 10^5.34, 10^5.48, 10^5.45, 10^5.28 (mL/g), and 10^5.28, respectively. Data of naphthalene adsorption (four samples) to and desorption (two samples) from nano-AC particles are plotted in Fig. 2. Data of naphthalene adsorption to nano-AC particles can be fitted with a linear isotherm in the form of q (μg/g) = 10^3.70 C_w (μg/mL). Desorption data show that desorption of naphthalene from nano-AC particles is also hysteretic, that is, for each desorption datum point, the adsorbed naphthalene concentration (q) is considerably higher than that predicted from the reversible linear sorption isotherm (the solid line on Fig. 2). Since the initial solid-phase and solution-phase organic concentrations of the two samples employed for desorption studies were similar, some of the experimentally observed desorption data points from these two samples appear to be overlapped in Fig. 2. Similar to the experimental results of naphthalene with nano-AC particles, adsorption and desorption of naphthalene and 1,2-DCB with nC₆₀ particles are highly hysteric. The experimental results of nC₆₀ experiments were published in a previous study (Cheng et al., 2005b). It should be noted that toluene was involved in nC₆₀ aqueous solution preparation in this study. The C₆₀/toluene mixture was sonicated in the presence of DI water for a period of 4 h until toluene was completely evaporated with no purple color left in the solution. The advantage of including toluene is to disperse C₆₀ more stably into aqueous solution (Andrievsky et al., 1995). Although toluene residue might be left in the nC₆₀ aqueous solution, the limited residual toluene in aqueous solution is not expected to materially impact the sorption/desorption experiments (Cheng et al., 2005a and 2005b). Figure 3 shows naphthalene adsorption to and successive desorption from ARS in water. The soil OC of ARS is the active component involved in the sorption of organic contaminants. The soil OC normalized distribution coefficient (K_oc) value for adsorption is 10^3.25 mL/g. On the desorption isotherm, the K_oc values change from 10^3.38 mL/g for the first desorption step, to 10^5.21 mL/g for the last desorption step. The deviation of the desorption isotherm from the expected linear isotherm and the gradual increase of K_d values indicate the occurrence of sorption hysteresis, as observed in numerous studies on desorption from natural soils and sediments.

FIG. 1.

Adsorption and desorption of naphthalene to and from “as-received activated carbon” particles (five samples). ♦, adsorption data for five samples; □, Δ, × , ⋄, and o represent four steps of desorption data for each of the five samples.

FIG. 2.

Adsorption and desorption of naphthalene to and from nano-AC particles. ♦, adsorption data; +, desorption data for the first desorption sample; ⋄, desorption data for the second desorption sample. AC, activated carbon.

FIG. 3.

Adsorption and desorption of naphthalene to and from Anacostia sediment. ▴, adsorption data; Δ, successive desorption data.

Model fitting for experimental data

Experimental data of sorption and desorption of naphthalene and 1,2-DCB with nC₆₀, naphthalene with AC, and naphthalene with soil OC in ARS are fitted with the above described four sorption models and are shown in Figs. 4 –7: Freundlich model (Fig. 4), Langmuir model (Fig. 5), DED model (Fig. 6), and Polanyi–Manes model (Fig. 7). Values of model predicted parameters were determined using the software of SigmaPlot and are listed in Table 1 with the R² values for each model fit. The linear isotherms (dashed lines in Figs. 4 –7) for the adsorption of naphthalene to nC₆₀, 1,2-DCB to nC₆₀, naphthalene to nano-AC particles, and naphthalene to soil OC are q (μg/g) = 10^3.74C_w (μg/mL), q (μg/g) = 10^3.48C_w (μg/mL), q (μg/g) = 10^3.70C_w (μg/mL), and q (μg/g) = 10^3.25C_w (μg/mL), respectively. As can be seen from Figs. 4 to 7 that, for the three forms of carbon tested in this study (nC₆₀, AC, and soil OC), the linear isotherms fail to predict the desorption data. This is because the adsorbed adsorbate concentration for each desorption datum point is one to two orders of magnitude higher than that predicted from the reversible linear isotherm, indicating the occurrence of sorption hysteresis. This means that the simple linear isotherm may not be appropriate to predict the observed sorption/desorption behavior. By comparing the R² values obtained from different model fits and the model fitting curves in Figs. 4 –7, it can be argued that the Polanyi–Manes model fits most desirably with the experimental data (Fig. 7). On the other hand, the overall data are not well described by the Langmuir model, indicated by the low R² values and the deviation of experimental data from model predicted values illustrated in Fig. 5. The poor performance of Langmuir model is probably due to the failure of this model to predict the high sorption affinity of all three forms of carbon at low aqueous adsorbate concentration ranges.

FIG. 4.

Freundlich model data fitting. Adsorption and desorption of (a) naphthalene with nC₆₀; (b) 1,2-dichlorobenzene with nC₆₀; (c) naphthalene with nano-AC particles; and (d) naphthalene with soil OC in Anacostia sediment. Filled symbols: adsorption data; open symbols: desorption data; dashed lines: linear isotherm assuming reversible desorption without hysteresis; solid lines: isotherms fitted by Freundlich model. □, Δ, × , ⋄, and + represent desorption data for each sample. OC, organic carbon.

FIG. 5.

Langmuir model data fitting. Adsorption and desorption of (a) naphthalene with nC₆₀; (b) 1,2-dichlorobenzene with nC₆₀; (c) naphthalene with nano-AC particles; and (d) naphthalene with soil OC in Anacostia sediment. Filled symbols: adsorption data; open symbols: desorption data; dashed lines: linear isotherm assuming reversible desorption without hysteresis; solid lines: isotherms fitted by Langmuir model. □, Δ, × , ⋄, and + represent desorption data for each sample.

FIG. 6.

DED model data fitting. Adsorption and desorption of (a) naphthalene with nC₆₀; (b) 1,2-dichlorobenzene with nC₆₀; (c) naphthalene with nano-AC particles; and (d) naphthalene with soil OC in Anacostia sediment. Filled symbols: adsorption data; open symbols: desorption data; dashed lines: linear isotherm assuming reversible desorption without hysteresis; solid lines: isotherms fitted by DED model. □, Δ, × , ⋄, and + represent desorption data for each sample. DED, dual-equilibrium desorption.

FIG. 7.

Polanyi–Manes model data fitting. Adsorption and desorption of (a) naphthalene with nC₆₀; (b) 1,2-dichlorobenzene with nC₆₀; (c) naphthalene with nano-AC particles; and (d) naphthalene with soil OC in Anacostia sediment. Filled symbols: adsorption data; open symbols: desorption data; dashed lines: linear isotherm assuming reversible desorption without hysteresis; solid lines: isotherms fitted by Polanyi–Manes model. □, Δ, × , ⋄, and + represent desorption data for each sample.

Table 1.

Parameters of Model Fitting to Experimental Data of Sorption–Desorption of Naphthalene and 1,2-Dichlorobenzene with nC₆₀, Naphthalene with Activated Carbon, and Naphthalene with Soil Organic Carbon in Anacostia River Sediment

Model fitting No.	Isotherm	Sorbent	Adsorbate	K_F [(μg/g)·(μg/mL)⁻ⁿ]	n	R²
1.1	Freundlich	nC₆₀	Naphthalene	10^4.26	0.44	0.9897
1.2		nC₆₀	1,2-DCB	10^3.33	0.09	0.9977
1.3		Nano-AC	Naphthalene	10^4.54	0.13	0.9583
1.4		ARS	Naphthalene	10^3.43	0.20	0.8675

Model fitting No.	Isotherm	Sorbent	Adsorbate	q_m (μg/g)	b (mL/μg)	R²
2.1	Langmuir	nC₆₀	Naphthalene	10^4.71	0.90	0.9669
2.2		nC₆₀	1,2-DCB	10^3.28	412.81	0.7757
2.3		Nano-AC	Naphthalene	10^4.63	35.00	0.8041
2.4		ARS	Naphthalene	10^3.37	74.74	0.4237

Model fitting No.	Isotherm	Sorbent	Adsorbate	\documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} $${ \rm{q}}_{max}^{2nd}$$ \end{document} (μg/g)	\documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} $${ \rm{K}}_d^{1st}$$ \end{document} (mL/g)	\documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} $${ \rm{K}}_d^{2nd}$$ \end{document} (mL/g)	R²
3.1	DED	nC₆₀	Naphthalene	10^3.85	10^3.74	10^5.73	0.9896
3.2		nC₆₀	1,2-DCB	10^3.21	10^3.44	10^6.23	0.9941
3.3		Nano-AC	Naphthalene	10^4.45	10^3.70	10^6.69	0.9443
3.4		ARS	Naphthalene	10^3.18	10^3.25	10^5.99	0.9871

Model fitting No.	Isotherm	Sorbent	Adsorbate	a′	b′	Q_0′ (cm³ 10³/kg)	R²
4.1	Polanyi-Manes	nC₆₀	Naphthalene	−0.0765	0.61	10^5.21	0.9917
4.2		nC₆₀	1,2-DCB	−9.21 × 10⁻⁵	1.50	10^3.32	0.9998
4.3		Nano-AC	Naphthalene	−2.09 × 10⁻⁵	1.91	10^4.60	0.9828
4.4		ARS	Naphthalene	−1.53	0.14	10^6.22	0.9464

Model fitting #	Isotherm	Sorbent	Adsorbate	a′	b′	Q₀′ (cm³ 10³/kg)	R²
5.1	Polanyi (q′ − ɛ/V_s)	nC₆₀	Naphthalene	−0.099	0.56	10^5.27	0.9869

1,2-DCB, 1,2-dichlorobenzene; AC, activated carbon; ARS, Anacostia river sediment; DED, dual-equilibrium desorption.

In Fig. 6, adsorption and desorption data of naphthalene with nC₆₀, nano-AC particles, and soil OC, and 1,2-DCB with nC₆₀ were fitted with the DED model. This model describes the sorption of organic compounds as the combination of two different processes. Sorption of the compound to the labile compartment obeys linear isotherm and desorbs reversibly, while sorption to the resistant compartment tends to be stronger and desorption hysteresis was normally observed due to the irreversible change of the pore structure of the adsorbent and/or the entrapment of adsorbate molecules in the adsorbent (Kan et al., 1998; Lu and Pignatello, 2002; Braida et al., 2003). As is shown in Fig. 6, this model can predict the overall sorption data satisfactorily with reasonably high R² values. As is listed in Table 1, the range of the \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} $$K_{ \rm{d}}^{{ \rm{1st}}}$$ \end{document} (mL/g) is from 10^3.25 to 10^3.74; \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} $$K_{ \rm{d}}^{{ \rm{2nd}}}$$ \end{document} (mL/g) is from 10^5.73 to 10^6.69, and \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} $$q_{{ \rm{max}}}^{{ \rm{2nd}}}$$ \end{document} (μg/g) is from 10^3.18 to 10^4.45. Relatively speaking, the values of each parameter among the three types of carbons varied insignificantly for about one order of magnitude. It was reported that (Kan et al., 1998; Chen et al., 2002) for many sediment/adsorbate combinations, \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} $$K_{\rm oc}^{2 {\rm nd}}$$ \end{document} (mL/g) was measured to be close to a single constant of 10^5.92. The \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} $$K_{\rm d}^{2 {\rm nd}}$$ \end{document} values obtained in this study for model fitting of all experimental data were close to this reported \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} $$K_{\rm oc}^{2 {\rm nd}}$$ \end{document} value obtained from soils/sediments. This similarity could possibly imply similar mechanisms. The DED model may possibly be used to predict the sorption behavior of new carbonaceous materials from that of other known carbon forms, such as AC.

Model fits of experimental sorption/desorption data with the Polanyi–Manes sorption model [Eq. (7)] are illustrated in Fig. 7. In the model fitting calculation, the molar volume (V_s) values for naphthalene and 1,2-DCB were estimated as the ratio of the molecular weight and the density of the chemical in its pure form (Farrell and Reinhard, 1994). Parameters of a′, b′, and Q₀′ are obtained from model fitting and listed in Table 1. By comparing the R² values obtained from this model to those of other models, it shows that the Polanyi–Manes model is the best to describe experimental nC₆₀ and AC sorption data. As discussed above, the sorption isotherm of adsorbed volume per unit mass of adsorbent versus the adsorption potential density (“correlation curve”) is expected to fall on a single line for all liquid adsorbates studied, if there are no specific interactions between adsorbate and the adsorbent surface (Manes, 1998; Xia and Ball, 1999). This is mostly the case for sorption of nonpolar organic adsorbates in aqueous solutions (Manes and Wohleber, 1971). Therefore, the maximum volume of nonpolar hydrocarbons on the surfaces of carbon is commonly calculated to be the same for all liquids investigated (Xia and Ball, 1999). The adsorbed volume per unit mass of nC₆₀ (q′, cm³ 10⁻³ kg) versus the adsorption potential density (ɛ/V_s, J/cm) is plotted for the adsorption and desorption of naphthalene and 1,2-DCB with nC₆₀ in Fig. 8. The obtained parameters of a′, b′, and Q₀′ are also listed in Table 1. Data can be fitted with one single characteristic curve reasonably well with an R² value of 0.987. The desirable fitting of experimental data with Polanyi–Manes sorption model, as illustrated in Figs. 7a–d and 8, probably provides some evidence for the mechanism underlying the sorption of hydrophobic organic compounds to nC₆₀ particles. It can be proposed that the sorption uptake of naphthalene and 1,2-DCB by nC₆₀ may be the result of adsorbate condensation into intraparticle micropores in nC₆₀, as depicted in Fig. 9. This proposed mechanism may also explain the observed sorption hysteresis since “capillary condensation” has long been believed to be one important cause for hysteretic sorption (Adamson and Gast, 1997). Note that, according to classical aquatic chemistry definitions, adsorption refers to a surface sorption phenomenon, while absorption corresponds to a bulk (volume) sorption phenomenon (Stumm and Morgan, 1996). In light of the fact that the entrapped organic molecules in the micropores were retained by the exterior surfaces of the micropores, it is thus more appropriate to designate the observed sorption phenomenon in this study as “adsorption” to emphasize the surface sorption characteristics. In addition, considering the sorption/desorption with soil OC data, the fit of Polanyi–Manes model underperforms the DED model, judged by the R² values. This is probably because soil OC is composed of both normal OC and high sorption affinity carbon, such as black carbons, leading to a more desirable data fitting with the two-compartment DED model. Moreover, it is noted that the Freundlich model fits experimental data of naphthalene and 1,2-DCB adsorption to nC₆₀ pretty well. This is not surprising since sorption sites on nC₆₀ particles are not expected to be homogeneous and the Freundlich model is most suitable for the scenario with multiple types of sites on the heterogeneous adsorbent surface (Weber and Digiano, 1996; Chiou, 2002; Schwarzenbach et al., 2003).

FIG. 8.

Polanyi–Manes sorption isotherm for naphthalene and 1,2-dichlorobenzene desorption from nC₆₀. Data are plotted as adsorbed volume (q′) versus adsorption potential density (ɛ/V_s).

FIG. 9.

Pictorial illustration of adsorption and desorption processes based on mechanism suggested by Polanyi–Manes model. The adsorption process can be viewed as condensation of organic pollutants into the micropores of carbonaceous material surface.

In summary, the DED model and Polanyi–Manes model both simulate experimental data presented herein reasonably well. Although three apparent parameters need to be optimized in the DED model, model fitting results showed that the values of each parameter, \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} $$K_{ \rm{d}}^{{ \rm{1st}}}$$ \end{document} , \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} $$K_{ \rm{d}}^{{ \rm{2nd}}}$$ \end{document} , and \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} $$q_{{ \rm{max}}}^{{ \rm{2nd}}}$$ \end{document} , are similar for the three forms of carbon tested in this work. Therefore, this model may be very useful in the prediction of sorptive behavior of novel carbonaceous nanomaterials from other known carbon forms. On the other hand, all three parameters in the Polanyi–Manes model need to be obtained by model fitting, one main limitation of applying such model. However, the satisfactory fitting of experimental data with the Polanyi–Manes model provides the most convincing explanation of the sorption mechanism as adsorbate condensation into intraparticle micropores of adsorbent. The single correlation curve as shown in Fig. 8 further supports the proposed “pore-filling” mechanism. In our future studies, the fitness of the Polanyi–Manes model with experimentally obtained sorption data will be systematically investigated and multiple duplicate samples will be employed during the sorption experimental studies to understand the deviation of the sorption data.

Summaries

In this study, sorption hysteresis is observed for the adsorption and desorption of naphthalene and 1,2-dichlorobenzene with C₆₀, AC, and soil OC, that is, desorption data deviate substantially from forward-constructed adsorption isotherm. As a consequence, experimental adsorption–desorption data cannot be fitted with a single linear isotherm. Instead, they were fitted with four different sorption models, including Freundlich model, Langmuir model, DED model, and Polanyi–Manes sorption model. Results show that DED model and Polanyi–Manes model fit data reasonably well, while the overall data are not well described by the Langmuir model. Each DED model obtained parameter has similar values for different carbon forms, indicating that this model can possibly be used to predict the sorption and desorption behavior of new carbonaceous materials from that of other known carbon forms. The desirable fitting of data with the Polanyi–Manes model implied that the intraparticle pore-filling might be the mechanism controlling the adsorption of organic compounds to nC₆₀ and AC particles.

Footnotes

Acknowledgments

This work was financially supported by Brine Chemistry Consortium companies of Rice University, including Aegis, Apache, BHGE, BWA, Chevron, ConocoPhillips, Coastal Chemical, EOG Resources, ExxonMobil, Flotek Industries, Halliburton, Hess, Italmatch, JACAM, Kemira, Kinder Morgan, Nalco, Oasis, Occidental Oil and Gas, Range Resources, RSI, Saudi Aramco, Schlumberger, Shell, SNF, Statoil, Suez, Total, and the NSF Nanosystems Engineering Research Center for Nanotechnology-Enabled Water Treatment (ERC-1449500). The authors also appreciate the financial support of Start-Up Research Grant provided by University of Macau (SRG2018-00112-FST) and the sponsorship of Science and Technology Development Fund, Macao S.A.R (FDCT) (0063/2018/A2).

Author Disclosure Statement

No competing financial interests exist.

References

Abdul

, Wang

, Zhang

, and Li

(2015). Adsorption of diethyl phthalate on carbon nanotubes: pH dependence and thermodynamics. Environ. Eng. Sci. 32, 103.

Accardi-Dey

, and Gschwend

P.M.

(2002). Assessing the combined roles of natural organic matter and black carbon as sorbents in sediments. Environ. Sci. Technol. 36, 21.

Adamson

A.W.

, and Gast

A.P.

(1997). Physical Chemistry of Surfaces, 6th ed., New York, NY: Wiley-Interscience.

Allen-King

R.M.

, Grathwohl

, and Ball

W.P.

(2002). New modeling paradigms for the sorption of hydrophobic organic chemicals to heterogeneous carbonaceous matter in soils, sediments, and rocks. Adv. Water Resour. 25, 985.

Andrievsky

G.V.

, Kosevich

M.V.

, Vovk

O.M.

, Shelkovsky

V.S.

, and Vashchenko

L.A.

(1995). On the production of an aqueous colloidal solution of fullerenes. J. Chem. Soc. Chem. Commun. 1281. https://pubs.rsc.org/en/content/articlelanding/1995/c3/c39950001281#!divAbstract

Atkins

, and de Paula

(2001). Physical Chemistry, 7th ed., Oxford, UK: W. H. Freeman.

Bello

O.S.

, Fatona

T.A.

, Falaye

F.S.

, Osuolale

O.M.

, and Njoku

V.B.

(2012). Adsorption of eosin dye from aqueous solution using groundnut hull–based activated carbon: Kinetic, equilibrium, and thermodynamic studies. Environ. Eng. Sci. 29, 186.

Braida

W.J.

, Pignatello

J.J.

, Lu

, Ravikovitch

P.I.

, Neimark

A.V.

, and Xing

(2003). Sorption hysteresis of benzene in charcoal particles. Environ. Sci. Technol. 37, 409.

Bucheli

T.D.

, and Gustafsson

(2000). Quantification of the soot-water distribution coefficient of PAHs provides mechanistic basis for enhanced sorption observations. Environ. Sci. Technol. 34, 5144.

10.

Chen

, Kan

A.T.

, Fu

, and Tomson

M.B.

(2000). Factors affecting the release of hydrophobic organic contaminants from natural sediments. Environ. Toxicol. Chem. 19, 2401.

11.

Chen

, Kan

A.T.

, Fu

, Vignona

L.C.

, and Tomson

M.B.

(1999). Adsorption desorption behaviors of hydrophobic organic compounds in sediments of Lake Charles, Louisiana, USA. Environ. Toxicol. Chem. 18, 1610.

12.

Chen

, Kan

A.T.

, Newell

C.J.

, Moore

, and Tomson

M.B.

(2002). More realistic soil cleanup standards with dual-equilibrium desorption. Ground Water, 40, 153.

13.

Cheng

, Kan

A.T.

, and Tomson

M.B.

(2004). Naphthalene adsorption and desorption from aqueous C₆₀ fullerene. J. Chem. Eng. Data, 49, 675.

14.

Cheng

, Kan

A.T.

, and Tomson

M.B.

(2005a). Study of C₆₀ transport in porous media and the effect of sorbed C₆₀ on naphthalene transport. J. Mater. Res. 20, 3244.

15.

Cheng

, Kan

A.T.

, and Tomson

M.B.

(2005b). Uptake and sequestration of naphthalene and 1,2-dichlorobenzene by C₆₀. J. Nanopart. Res. 7, 555.

16.

Chiou

C.T.

(2002). Partition and Adsorption of Organic Contaminants in Environmental Systems. Hoboken, NJ: John Wiley & Sons, Inc.

17.

Chiou

C.T.

, and Kile

D.E.

(1998). Deviation from sorption linearity on soils of polar and nonpolar organic compounds at low relative concentrations. Environ. Sci. Technol. 32, 338.

18.

Clark

M.M.

(2009). Transport Modeling for Environmental Engineers and Scientists, 2nd ed., Hoboken, NJ: John Wiley & Sons.

19.

Colvin

(2003). The potential environmental impact of engineered nanomaterials. Nat. Biotechnol. 21, 1166.

20.

Cornelissen

, and Gustafsson

(2004). Sorption of phenanthrene to environmental black carbon in sediment with and without organic matter and native sorbates. Environ. Sci. Technol. 38, 148.

21.

Crittenden

J.C.

, Hand

D.W.

, Arora

, and Lykins

B.W.J.

(1987). Design considerations for GAC treatment of organic chemicals. J. Am. Water Works Assoc. 79, 74.

22.

Dubinin

M.M.

(1960). The potential theory of adsorption of gases and vapors for adsorbents with energetically nonuniform surfaces. Chem. Rev. 60, 235.

23.

Evans

D.F.

, and Wennerström

(1999). The Colloidal Domain: Where Physics, Chemistry, Biology, and Technology Meet, 2nd ed. New York, NY: Wiley-VCH.

24.

Farrell

, and Reinhard

(1994). Desorption of halogenated organics from model solids, sediments, and soil under unsaturated conditions. 1. Isotherms. Environ. Sci. Technol. 28, 53.

25.

, Kan

A.T.

, and Tomson

M.B.

(1994). Adsorption and desorption hysteresis of PAHs in surface sediment. Environ. Toxicol. Chem. 13, 1559.

26.

Grzegorczuk-Nowacka

, and Anielak

A.M.

(2017). Effect of iron and aluminum on adsorption of fulvic acids on Norit ROW 0.8 Supra carbon. Environ. Eng. Sci. 34, 659.

27.

Hiemenz

P.C.

, and Rajagopalan

(1997). Principles of Colloid and Surface Chemistry, 3rd ed. Boca Raton, FL: CRC Press.

28.

Huang

, Yu

, and Weber

W.J.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.

29.

James

, Sabatini

D.A.

, Chiou

C.T.

, Rutherford

, Scott

A.C.

, and Karapanagioti

H.K.

(2005). Evaluating phenanthrene sorption on various wood chars. Water Res. 39, 549.

30.

Kan

A.T.

, Fu

, Hunter

, Chen

, Ward

C.H.

, and Tomson

M.B.

(1998). Irreversible adsorption of neutral organic hydrocarbons-experimental observations and model predictions. Environ. Sci. Technol. 32, 892.

31.

Kan

A.T.

, Fu

, and Tomson

M.B.

(1994). Adsorption/desorption hysteresis in organic pollutant and soil/sediment interaction. Environ. Sci. Technol. 28, 859.

32.

Karickhoff

S.W.

(1980). Sorption kinetics of hydrophobic pollutants in natural sediment. In Baker

R.A.

, Ed., Contaminants and Sediments: Analysis, Chemistry, Biology. Ann Arbor, MI: Ann Arbor Science Publishers.

33.

Kleineidam

, Rugner

, Ligouis

, and Grathwohl

(1999). Organic matter facies and equilibrium sorption of phenanthrene. Environ. Sci. Technol. 33, 1637.

34.

, and Pignatello

J.J.

(2002). Demonstration of the “conditioning effect” in soil organic matter in support of a pore deformation mechanism for sorption hysteresis. Environ. Sci. Technol. 36, 4553.

35.

Manes

(1998). Activated carbon adsorption fundamentals. In Myers

R.A.

, Ed., Encyclopedia of Environmental Analysis and Remediation. New York: Wiley, p. 26.

36.

Manes

, and Wohleber

D.A.

(1971). Application of the Polanyi adsorption potential theory to adsorption from solution on activated carbon. II. Adsorption of partially miscible organic liquids from water solution. J. Phys. Chem., 75, 61.

37.

Polanyi

(1916). Adsorption of gases (vapors) by a solid nonvolatile adsorbent. Verh. Dtsch. Phys. Ges. 18, 55.

38.

Rathousky

, and Zukal

(2000). Adsorption of krypton and cyclopentane on C₆₀: An experimental study. Fullerene Sci. Technol. 8, 337.

39.

Sahu

, and Casciano

D.A.

(2009). Nanotoxicity: From In Vivo and In Vitro Models to Health Risks. Chichester, West Sussex, UK: Wiley.

40.

Sawyer

, McCarty

, and Parkin

(2002). Chemistry for Environmental Engineering and Science, 5th ed. New York: McGraw-Hill.

41.

Schwarzenbach

R.P.

, Gschwend

P.M.

, and Imboden

D.M.

(2003). Environmental Organic Chemistry. Hoboken, NJ: John Wiley & Sons.

42.

Stumm

, and Morgan

J.J.

(1996). Aquatic Chemistry, 3rd ed. Hoboken, NJ: Wiley-Interscience.

43.

Vollath

(2013). Nanomaterials: An Introduction to Synthesis, Properties and Applications, 2nd ed. Weinheim, Germany: Wiley-VCH.

44.

Wang

, Santos

, Evdokiou

, and Losic

(2015). An overview of nanotoxicity and nanomedicine research: principles, progress and implications for cancer therapy. J. Mater. Chem. B, 3, 7153.

45.

Weber

W.J.J.

, and Digiano

F.A.

(1996). Process Dynamics in Environmental Systems. New York: John Wiley & Sons.

46.

Weber

W.J.J.

, and Huang

(1996). A distributed reactivity model for sorption by soils and sediments. 4. Intraparticle heterogeneity and phase-distribution relationships under nonequilibrium conditions. Environ. Sci. Technol., 30, 881.

47.

Weber

W.J.J.

, Huang

, and Yu

(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.

48.

Xia

, and Ball

W.P.

(1999). Adsorption-partitioning uptake of nine low-polarity organic chemicals on a natural sorbent. Environ. Sci. Technol. 33, 262.

49.

Xing

, and Pignatello

J.J.

(1996). Time-dependent isotherm shape of organic compounds in soil organic matter: Implications for sorption mechanism. Environ. Toxicol. Chem. 15, 1282.

50.

Xing

, and 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.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.05 MB

Modeling Analysis of Two Common Organic Pollutants' Adsorption and Desorption from Activated Carbon,C 60,and Soil Organic Carbon

Abstract

Abstract

Introduction

Experimental Protocols

Materials

Adsorption and desorption experiments using AC

As-received AC particles

Nano-AC particles

Adsorption and desorption experiments using C60

Adsorption and desorption experiments using ARS

Sorption models

Freundlich sorption model

Langmuir sorption model

Combined sorption model: DED model

Polanyi–Manes sorption model

Results and Discussion

Adsorption and desorption experimental results

Model fitting for experimental data

Summaries

Footnotes

Acknowledgments

Author Disclosure Statement

References

Supplementary Material

Adsorption and desorption experiments using C₆₀