当前位置:首页 >> 英语学习 >>

Adsorption of Organic Compounds by Carbon Nanomaterials in Aqueous Phase


Chem. Rev. 2010, 110, 5989–6008

5989

Adsorption of Organic Compounds by Carbon Nanomaterials in Aqueous Phase: Polanyi Theory and Its Application
Kun Yang?,? and Baoshan Xing*,?
Department of Environmental Science, Zhejiang University, Hangzhou 310028, China, and Department of Plant, Soil and Insect Sciences, University of Massachusetts, Amherst, Massachusetts 01003 Received February 19, 2010

Contents
1. Introduction 2. Polanyi Theory 3. Morphology and Possible Adsorption Spaces of CNPs 4. Interactions of Organic Compounds with CNPs 4.1. Hydrophobic Interaction 4.2. π-π Bonding Interaction 4.3. Hydrogen Bonding Interaction 4.4. Covalent Bonding Interaction 4.5. Electrostatic Interaction 5. Isotherm Models and Their Fitting 6. Polanyi Theory-Based Dubinin-Ashtakhov (DA) Model and Its Parameters 7. In?uences of CNP Structures and Surface Properties on Adsorption 7.1. Surface Area 7.2. Functional Groups on CNP Surface 7.3. The Pores in CNP Aggregates 7.4. Surface Curvature and Defects 8. Effects of Organic Chemical Structures and Properties on Adsorption 8.1. Molecular Structure and Functional Groups 8.2. Molecular Size 8.3. Chemical Solvatochromic Parameters 8.4. Dissociation of IOCs 9. Competitive Sorption and the Roles of DOM and Surfactants in Adsorption 9.1. Competitive Sorption of PAHs 9.2. Competitive Sorption with Metal Ions 9.3. Competitive Sorption with DOM and Surfactants 9.4. Other In?uence of DOM and Surfactants on Organic Chemical Adsorption 10. Adsorption-Desorption Hysteresis 10.1. Hysteresis Phenomenon 10.2. Desorption Hysteresis on Fullerene 10.3. Desorption Hysteresis and Immobilization on CNTs 11. Conclusions and Perspectives 12. Acknowledgments 13. References 5989 5991 5991 5992 5992 5993 5993 5993 5993 5993 5995 5996 5996 5997 5999 5999 6000 6000 6000 6001 6001 6003 6003 6003 6003 6003 6004 6004 6004 6005 6005 6006 6006

1. Introduction
Since their ?rst discovery,1-3 carbon nanomaterials, including fullerenes and carbon nanotubes (CNTs), have attracted special attention in terms of potential applications in catalyst supports, optical devices, quantum computers, and biomedical use.4,5 The unique and tunable properties of carbon nanomaterials also enable a broad range of their applications in identifying and addressing environmental challenges: sorbents, high-?ux membranes, depth ?lters, antimicrobial agents, environmental sensors, renewable energy technology, and pollution prevention strategy.5 For example, the theoretical surface area of fullerene and singlewalled CNTs is about 3000 m2/g,6,7 which makes them potential superior adsorbents. Carbon nanomaterials are now produced and used in signi?cant quantities, for example, the annual global production capacity of CNTs in the year 2007/ 2008 was estimated as exceeding 350 tons5 and increased future production and application are expected. A recent market report estimated the business opportunities of CNTs in electronics applications reaching $3.6 billion by 2009.8 Because of the current and increasing future investments and the potential widespread use, carbon nanosized particles (CNPs) will inevitably enter the environment from their production, transport, handling, use, and disposal.9-11 Although the concentrations of CNPs measured in the environment such as air and soils are still small at the present time,12 it has to be expected that their concentrations in the environment will increase considerably in future when their price falls and their applications in consumer products become more widespread.13 CNPs were observed to be taken up by different cell types and evoked diverse effects in the cells and, consequently, may be toxic to human beings, plants, animals, and other organisms.14,15 Furthermore, CNPs may accumulate along the food chain because they are totally insoluble in water and lipophilic in pristine form and possibly one of the least biodegradable man-made materials ever devised.16,17 In addition to the toxicity, therefore, the fate, transport, and bioavailability of CNPs and their effects on common environmental contaminants need to be evaluated, which will be fundamental in determining overall environmental and health impact of CNPs once escaped into the environment18-21 Adsorption of organic compounds by CNPs in aqueous phases is a critical process for environmental application of CNPs in water treatment5,22,23 and solid-phase extraction24-27 as superior sorbents and in synthesis of functionalized CNPs with desired and tunable physical and chemical properties.28 This process is also important for determining the environmental and health impacts of both CNPs and organic

* To whom correspondence should be addressed. Phone: (413) 545-5212. Fax: (413)545-3958. E-mail: bx@pssci.umass.edu. ? Department of Environmental Science, Zhejiang University. ? Department of Plant, Soil and Insect Sciences, University of Massachusetts.

10.1021/cr100059s ? 2010 American Chemical Society Published on Web 06/02/2010

5990 Chemical Reviews, 2010, Vol. 110, No. 10

Yang and Xing

Dr. Kun Yang received his B.Sc. degree in 1999 and his Ph.D. degree in 2004 in Environmental Science from Zhejiang University, China. In 2004, he joined the Zhejiang University, where he is an active faculty member in teaching and research, and was promoted to Associate Professor in 2006. Dr. Yang was a postdoctoral fellow from 2005 to 2006 and a visiting scientist from 2008 to 2010 in the group of Professor Baoshan Xing at the University of Massachusetts, Amherst. He won the second prize of the Natural Science Award in the University of China and the National Excellent Doctoral Dissertation Nominations of China in 2007. He was selected as a New Century Excellent Talent in University by the Education Ministry of China in 2008. As the author and coauthor, Dr. Yang has published more than 30 articles in peer-reviewed technical journals, and his publications are frequently cited by his peers. His current research interests focus on the environmental behavior and toxicity of engineered nanomaterials, the sorption and fate of organic chemicals in soils and sediments, the synthesis and characterization of porous adsorbents, and the surfactant-enhanced remediation of contaminated soils with organic chemicals.

chemicals.10,23 First, adsorption of natural organic matter and surfactants on CNPs would enhance their suspension and stability in water and, consequently, alter the environmental fate and transport of CNPs.29-34 Second, adsorbed organic chemicals such as hydrophobic organic compounds (HOCs) may add additional toxicity to that of CNPs when these CNPs are taken up by and accumulated in organisms.23,35,36 Third, CNPs would act similarly to other “hard” carbon materials such as charcoal and soot37,38 in the environment and alter the fate, transfer, and bioavailability of environmental organic pollutants by adsorption.23,39 For example, the bioavailability and mobility of organic pollutants into organisms could be reduced and, consequently, their persistence in the environment would be enhanced if they are adsorbed by a CNP which cannot be taken up by or accumulated in the organisms.40 Differences between CNPs and the most common carbon adsorbents (i.e., activated carbon) need to be recognized for the proper interpretation of adsorption phenomena of CNPs. First, the activated carbon consists of micropores with different sizes which provide sites for adsorption, while individual CNPs provide adsorption sites only along the external surface.6 Second, activated carbon contains carbons of varying degree of saturation and oxidation state as well as functional groups formed during activation process,41 while CNPs consist only of globally conjugated unsaturated carbons in three-dimensional arrays and contain few functional groups except for the carbons on their surface that are intentionally treated by oxidation techniques.42-45 Polanyi theory, originally used to describe gas adsorption by activated carbon46,47 and later applied to adsorption from aqueous phases,48,49 was observed to mechanistically capture the adsorption process of nonionic polycyclic aromatic hydrocarbons (PAHs) by CNPs in our previous work.23 Later, this theory was successfully used to describe the competitive

Prof. Baoshan Xing is Professor of Environmental and Soil Chemistry (since 2004) at College of Natural Sciences, University of Massachusetts, Amherst, MA, where he has been actively involved in teaching and research since 1996. Dr. Xing received his Ph.D. degree from University of Alberta, Canada, in 1994. Dr. Xing’s current research includes sorption and fate of organic chemicals in soils and sediments, environmental behavior and toxicity of engineered nanomaterials, interfacial processes in the environment, natural organic matter characterization and chemistry, and application of spectroscopic and microscopic techniques in environmental and soil sciences. Along with students and colleagues, he has published over 180 peer-reviewed articles; his research work is ranked in the top 1% of cited authors for journals in environmental sciences and ecology. Currently, Dr. Xing is teaching (1) Environmental Soil Chemistry, (2) Advanced Soil Chemistry, (3) Inorganic Contaminants in Soil, Water, and Sediment, and (4) Environmental Impact of Manufactured Nanomaterials. He has been invited to present his research results at many universities and institutions. Dr. Xing is a technical editor of Journal of Environmental Quality and an editorial board member of Environmental Pollution, Chemosphere, International Journal of Phytoremediation, Communications in Soil Science and Plant Analysis, and Pedosphere. Dr. Xing won the outstanding research award of his college (Natural Sciences) in 2003 and of Soil Science Society of America and American Society of Agronomy (Northeast Branch) in 2006. He was selected as a Cheung Kong Scholar by the Education Ministry of China in 2007. He was also early promoted to Associate Professor with tenure in 2000 and to Full Professor in 2004. His research program is recognized internationally.

sorption of PAHs on CNTs50 as well as the adsorption of ionizable organic compounds (IOCs) such as phenols and anilines.51 Several other works45,52-55 also applied this theory in describing organic chemical adsorption by CNPs. Polanyi theory-based models have been successfully used to ?t the adsorption isotherms.23,45,50-55 Adsorption mechanisms of organic compounds on CNTs from aqueous solution have been partly summarized by Pan and Xing.56 Carbon nanotubes as solid-phase extraction material and sorbents have also been partly reviewed in previous works.10,26,27 However, Polanyi theory and its application in describing organic chemical adsorption by CNPs are not discussed in these review papers. Empirical relationships, relating the adsorption capacity and af?nity of CNPs with the CNP/organic chemical structures and physicochemical properties, are also not reviewed, but these empirical relationships can be helpful to explore the adsorption mechanisms and to estimate the adsorption capacity and af?nity, and then one can evaluate the potential application of CNPs as sorbents and in?uence of adsorption on the environmental risks of both toxic organic chemicals and CNPs. With the latest publications and current understanding, a review, focused on the aqueous adsorption of organic chemicals by CNPs, therefore, is needed and will be timely provided in this paper based on the Polanyi theory and the Polanyi theory-based Dubinin-Ashtakhov (DA) model. Parameters of the DA model, identifying the adsorption capacity and af?nity of organic chemicals on CNPs, will

Polanyi Theory and Its Application

Chemical Reviews, 2010, Vol. 110, No. 10 5991

be addressed in relation with the CNP/organic chemical structures and physicochemical properties, the interaction strength between organic chemicals and CNPs, and the newly created surface characteristics as a result of surface modi?cation.

2. Polanyi Theory
Adsorption of an adsorbate by adsorbents should be described in two aspects: adsorption capacity and af?nity. Adsorption capacity is limited by the potential space of a sorbent available for adsorption of a given adsorbate, while adsorption af?nity is dependent on the strength of attractive forces between adsorbate and adsorbent. The Polanyi theory46-48,57 has been recognized as the most powerful available theory for dealing with both gas and aqueous adsorption on energetically heterogeneous surfaces such as activated carbon. This theory assumes that for a molecule located within the attractive force ?eld (i.e., adsorption space) of a solid, there exists an (attractive) adsorption potential between the molecule and the solid surface. The adsorption potential is higher if the molecule is more close to the solid surface. This attraction derives from the induced dipoleinduced dipole force of the molecule and surface atoms, which is a short-range in nature. The adsorption potential, ε ) -RT ln(Ce/Cs), at a particular location in the adsorption space is de?ned as the energy that is required to remove the molecule from that location to a point outside the attractive force ?eld of the solid surface; where R is the universal gas constant, T is the absolute temperature, Cs and Ce is the water solubility and the equilibrium solution phase concentration of an adsorbate, respectively, for aqueous adsorption. For a molecule in this space, therefore, the magnitude of adsorption potential varies within the adsorption space depending on its proximity to the solid surface atoms. A series of equipotential surfaces can be obtained once the points in adsorption space with the same ε are connected. In addition, the Polanyi theory assumes that the adsorption potential is independent of temperature and the adsorbed liquid has the similar properties as the corresponding bulk liquid. A direct consequence of these assumptions is that for a given adsorbent, a plot of adsorbed volume (qv) or adsorbed mass (qe) against equilibrium adsorption potential (ε) should yield a curve (called as “characteristic curve”) that is temperatureinvariant and determined by the structure of the adsorbent.48 This temperature-invariant characteristic curve was observed for the PAH adsorption by CNPs,23 which indicates that Polanyi theory captures this aqueous adsorption process mechanistically. Characteristic curves of phenanthrene adsorption on an multiwalled carbon nanotube sample (i.e., MWCNT15) at 25, 40, and 55 °C, as an example reported by Yang et al.,23 are shown in Figure 1. Application of Polanyi theory in describing the adsorption process of adsorbates by adsorbents needs a mathematical equation to relate the experimental qv (or qe) values with the Ce values of an adsorbate (i.e., isotherm) and extract the isotherm characteristics by employing and analyzing the parameters of the mathematical equation because the apparent relationships between qv (or qe) and Ce for different adsorbates and adsorbents are generally different. The most bene?t to develop a mathematical equation for Polanyi theory is that the parameters of the mathematical equation for certain pairs of adsorbate and adsorbent could be related with the physicochemical properties of both adsorbate and adsorbent. Then it is possible to explore the roles of the physicochemical properties of adsorbates and adsorbents in adsorption and to

Figure 1. Characteristic curves of phenanthrene adsorption on MWCNT15 at 25, 40, and 55 °C. Reprinted with permission from ref 23. Copyright 2006 American Chemical Society.

estimate the adsorption capacity and af?nity from the physicochemical properties. Several mathematical equations including the DA model23,49,57 have been developed by employing the adsorption potential (ε) in the mathematical equations since the Polanyi theory was suggested.46,47

3. Morphology and Possible Adsorption Spaces of CNPs
Monomers of fullerene and CNTs have distinct geometry, showing different spaces for adsorption. The monomer structure of fullerene is a closed graphite ball, while CNTs are rolled-up graphite sheets forming a coaxial tube (Figure 2). One single rolled-up graphite sheet forms single-walled CNTs (SWCNTs), while two or more rolled-up graphite sheets form multiwalled CNTs (MWCNTs) (Figure 2). For the spherical fullerene monomer, the available adsorption space is their external surface.6 Besides the external surface, both inner cavity (i.e., the spaces in the innermost graphite tube) and interwall spaces (i.e., the spaces between the coaxial tubes of MWCNTs) exist in the cylindrical SWCNT or MWCNT monomer (Figure 2).6 However, the interwall spacing would be inaccessible for nitrogen molecule with a size of 0.354 nm and common organic molecules with bigger sizes as compared with the narrow interwall spacing, 0.335 nm.58 The inner cavities would also be inaccessible for organic molecules because these cavities would commonly be blocked by impurities such as metal catalyst and amorphous carbons.6 Although the blocked inner pores can be opened up by acid treatment using HCl44-59 to eliminate metal catalysts located at the end of CNTs or using H2O2,43 nitric acid,60 base,61 or heat treatment44 to remove the amorphous carbon, the functional groups would be created in the opened inner pores upon these treatments and bind water molecules to increase the diffusion resistance for organic molecules, thus prohibiting the access of organic molecules into the opened inner pores.59 Therefore, available spaces of CNP monomers for adsorption are limited to their external surface. CNP monomers conglobe spontaneously to form aggregates that exist stably in the environment.29,62-64 In the aggregation process, CNP monomers form small aggregates ?rst and then big aggregates (Figure 3).6,62,63 A typical size of big aggregates of fullerene, for example, was observed to be from 20 to 50 ?m, while small aggregates obtained by

5992 Chemical Reviews, 2010, Vol. 110, No. 10

Yang and Xing

Figure 2. Schematic structures of fullerene (A), single-walled carbon nanotubes (B), and multiwalled carbon nanotubes (C), showing inner cavity, interwall spaces, and external surface. Fullerene (C60) has only external surface. Reprinted with permission from ref 6. Copyright 2007 Elsevier.

Figure 3. Schematic aggregation process of fullerene and carbon nanotubes: (A) carbon nanotubes and (B) fullerene. Reprinted with permission from ref 6. Copyright 2007 Elsevier.

extended stirring treatment were from 1 to 3 ?m62 and from 20 to 500 nm by solvent exchange.63 The particle size of CNP aggregates in water will also be affected by pH,65-68 acid-treatment of CNP surfaces,67-71 and added surfactants29,72-76 and dissolved organic matter (DOM).30,34,65,66,71 Aggregation of spherical fullerene will result in closed interstitial spaces in its aggregates, while cylindrical CNTs cannot form closed interstitial spaces in their aggregates due to their length (Figure 3).6 Aggregation of CNP monomers results in the decease of their exposed effective surface for adsorption.6 Therefore, the measured surface area of fullerene and SWCNTs are much lower than their theoretically calculated surface area (i.e., as high as 3000 m2/g).6,7 Decrease in surface area by aggregation is larger for fullerene than for CNTs because of the formation of closed interstitial spaces in fullerene aggregates that prohibit the entrance of adsorbate molecules unless these closed spaces are opened up during the adsorption process.6 Aggregation of CNP monomers can also form grooves in the aggregates.53 Interstitial spaces and grooves, surrounded by the external surface of CNP monomers, are the pores of CNPs that are possibly available for adsorption.6,53

dominant force for gas or vapor adsorption onto a hydrophobic adsorbent,77 which may also be signi?cant for adsorption from the aqueous phase.49,51 Considering only van der Waals forces, however, may not be applicable in cases when dipole-dipole, induced-dipole, and hydrogen-bonding donor-acceptor interactions exist, typically in aqueous phase, where these forces can be important especially for chemicals with certain functional groups.49,51 Five possible interactions including hydrophobic effect, π-π bonds, hydrogen bonds, and covalent and electrostatic interactions have been observed and are responsible for the adsorption of organic chemicals on CNP surface.51,56,78 These interactions, their strengths, and contribution to the overall sorption are a function of the properties of both organic chemicals and CNPs.51,56,78

4.1. Hydrophobic Interaction
The surfaces of CNPs are hydrophobic, as can be demonstrated by the preference adsorption of hydrocarbon (such as hexane, benzene, and cyclohexane) over alcohols (such as ethanol and 2-propanol).79 Therefore, hydrophobic interaction has been employed to interpret the adsorption of organic chemicals especially for HOCs by CNPs.43,44,51,80-82 However, hydrophobic interaction alone is not enough to interpret the observed adsorption by CNPs,51 as was also indicated by the failure in establishing a general relationship between the hydrophobic parameters of organic chemicals and their adsorption af?nity parameters on CNPs.53,56,83

4. Interactions of Organic Compounds with CNPs
The net attractive forces involving the solute, solvent, and the adsorbent are assumed to be responsible for the solute adsorption by sorbents such as CNPs and activated carbon.49,51,77 Among these forces, van der Waals force is normally the

Polanyi Theory and Its Application

Chemical Reviews, 2010, Vol. 110, No. 10 5993

4.2. π-π Bonding Interaction
π-π Bonding interaction has been used to interpret the adsorption of organic molecules with a CdC double bond or a benzene ring on the surface of CNPs because these organic molecules contain π electrons to interact with the π electrons of the benzene rings on CNP surface through the π-π electron coupling.83-91 The π-π bonding interaction between organic molecules and CNPs has been demonstrated by spectroscopic studies using Raman,92 nuclear magnetic resonance (NMR), and ?uorescence techniques.93 In addition, this π-π bonding interaction is affected by the relative position of the benzene ring of organic molecules to the hexagons on the CNP surface.88,89,94

4.3. Hydrogen Bonding Interaction
Hydrogen bonding interaction can form and play an important role in adsorption when the organic chemicals or CNPs have certain functional groups such as -COOH, -OH, and -NH2.51,95 The -COOH, -OH, and -NH2 groups of organic chemicals can act as hydrogen-bonding donors and form hydrogen bonds with graphite sheets of CNPs, where the benzene rings of graphite sheets of CNPs act as the hydrogen-bonding acceptors.51 Hydrogen bonds might also form between organic molecules and CNPs if CNPs have -COOH and -OH groups on their surfaces as hydrogen-bonding donors.96,97 The insigni?cant effects of hydrogen-bonding interaction on adsorption of solutes without hydrogen-bonding donor ability51,83 could be attributed to that the investigated CNPs have few functional groups on their surface and the graphite sheets of CNPs cannot be an effective hydrogen-bonding donor. Moreover, functional groups of CNPs can also form hydrogen bonds with water molecules.97-99 This hydrogen-bonding interaction would be stronger than that between functional groups of CNPs and organic solutes, which results in the competitive sorption of water with organic solutes97,98 and thereby the insigni?cant contribution of hydrogen-bonding interaction on adsorption of solutes without hydrogen-bonding donor ability.51,83

4.4. Covalent Bonding Interaction
Covalent bonds may occur between organic chemicals and CNPs if both of the chemicals and CNPs have certain functional groups such as -COOH, -OH, and -NH2.78,100,101 This covalent bond has been illustrated by spectroscopic studies with infrared spectroscopy (IR), X-ray photoelectron spectroscopy (XPS), and NMR techniques.78,101 Relative to the noncovalent bonding interactions (e.g., hydrophobic, π-π bonding, and hydrogen-bonding interactions), the attachment of organic molecules to CNPs via a covalent bond is much stronger and can resist any desorption. Therefore, covalent modi?cation of CNPs has been widely utilized to form a variety of nanostructures with excellent physical and chemical properties.28,102,103 Functionalizations of CNPs with covalent modi?cation are normally achieved via reactions such as carboxylation,104 amidation,105 diazonium,106 ?uorination,107 free radical chemistry,108 bingel reaction,109 esteri?cation,110 and composite formation.111,112

solution pH because the facts that NOM and phenolic chemicals can be dissociated to anions and CNPs can be negatively charged as pH increased.32 Electrostatic repulsion between the dissociated species of NOM (or phenolic chemicals) and negatively charged CNPs could result in the adsorption decrease. The electrostatic interaction is related to the charge nature of both organic chemicals and CNPs. Electrostatic attraction will occur if CNPs and organic chemicals have opposite charges; otherwise, electrostatic repulsion will occur if both CNPs and organic chemicals have the same sign of charges. The observed pH-dependent adsorption of a ?uoroquinolone antibacterial agent (nor?oxacin) on CNTs was also interpreted with the electrostatic interaction because nor?oxacin can be positively charged (cationic), negatively charged (anionic), and/or as zwitterionic ions, depending on the solution pH.55 However, it is dif?cult to separate this electrostatic interaction in?uence from the other interactions because dissociation of organic chemicals is always accompanied with the decrease of hydrophobic effect and hydrogen-bonding interaction of organic chemicals.32 Although ?ve interactions have been recognized to be responsible for adsorption of organic chemicals on CNPs, their relative contributions on adsorption of a given organic chemical are still dif?cult to be quanti?ed. The biggest challenge to determine their relative contributions is the estimation of the strength of these interactions, which depends on the ability of both organic chemicals and CNPs to form these interactions. The ability of an organic chemical to form hydrophobic interaction, π-π, and hydrogen bonds with CNPs can be identi?ed quantitatively by its water solubility, π-electron polarity/polarizability, and hydrogenbonding donor-acceptor parameters, respectively.51 However, the ability of an organic chemical to form covalent bonds with CNPs and that of a given CNP to form the mentioned ?ve interactions with organic chemicals still cannot be estimated. Multiple regression, which can correlate the adsorption of organic chemicals with their chemical parameters of ability to interact with CNPs if these ability parameters are available, is an useful method to evaluate the relative contributions of multiple interactions to the overall adsorption, as did in the previous studies.49,51 Using multiple regression, the ability of CNPs to interact with organic chemicals could also be estimated by correlating the adsorption af?nity parameters of reference organic chemicals with the property parameters of CNPs. For example, cyclohexane may be a good reference chemical to identify the ability of CNPs to have hydrophobic interactions.83 Moreover, PAHs could be good reference chemicals to identify the ability of CNPs to have π-π bonding interactions because the hydrophobic effects of organic chemicals on their adsorption can be largely eliminated by normalizing their adsorption af?nity with their KHW values or by conducting the adsorption experiments in cyclohexane solution.53,56,83

5. Isotherm Models and Their Fitting
Isotherm ?tting with model equations is a key issue to explore sorption mechanisms. Deviation of the ?tted isotherm from the true isotherm, as a statistic result, is inevitable because of experimental errors. However, large deviations of the ?tted isotherm from the true isotherm could occur if few experimental data in a narrow range of equilibrium concentrations are employed for ?tting (Figure 4). Large

4.5. Electrostatic Interaction
Electrostatic interaction has been also utilized to interpret the adsorption decrease of natural organic matter (NOM)31,32 and phenolic chemicals113 on CNPs with the increase of

5994 Chemical Reviews, 2010, Vol. 110, No. 10

Yang and Xing

Figure 4. Possible deviations of the ?tted isotherms from the true isotherm: the isotherms I, II, and III are ?tted by employing the experimental data in regions I, II, and III, respectively.

deviation could confuse the researchers to give misleading interpretations and conclusions of the observed data. In previous studies, ?ve models have been observed having good ?tting of the aqueous sorption isotherms of organic chemicals by CNPs (Table 1). They are linear, Freundlich, Langmuir, partition-adsorption, and Polanyi theory-based equations. The good ?tting of linear equations was only observed for the isotherms of naphthalene adsorption by large fullerene (C60) aggregates and thin C60 ?lms.62 It was also observed that isotherms of a given primary solute changed from nonlinearity toward linearity when competitors were added.50 The partition-adsorption equation was used to interpret the adsorption-desorption hysteresis of naphthalene and 1,2-dichlorobenzene on/from small C60 aggregates.114 Several studies reported the good ?tting of Langmuir equation to the isotherms of organic chemicals by CNPs.115-120 Others reported that the Freundlich equation ?tted the isotherms well.45,54,62,82,83,90,91,113,116,119-121 However, comparison of the studies regarding equation ?tting of isotherms
Table 1. Models Used for Good Fitting of Isotherms name linear model Freundlich model equation

indicates that Polanyi theory-based equations had better ?tting than Freundlich, Langmuir, and partition-adsorption equations for most obtained isotherms of various organic chemicals and CNPs.23,50,51,53 Signi?cant deviation of equation estimation from the experimental data was observed for Langmuir and partition-adsorption equations at relatively low concentrations.23 The observations of the good ?tting of Langmuir equation to the isotherms of organic chemicals by CNPs in some studies115-120 is due to a narrow range of equilibrium concentrations (<1 order of magnitude) and few experimental data points used in these studies. The inapplicability of the partition-adsorption model was also supported by the absence of a rubbery state domain (partition medium) of CNPs for HOC partition because CNPs are predominated by the crystalline and graphitic type of carbon.23,50 Moreover, it was observed that the partitionadsorption model failed to explain the sorption and competition data of PAHs.50 The good ?tting of the Freundlich equation was observed in some cases but varied with both the tested organic chemicals and CNPs in these model comparison studies.23,50,53 Although the Freundlich equation was well applied to ?t all the obtained isotherms in several other studies,45,54,62,82,83,90,91,113,116,119-121 this could be attributed to the narrow range of equilibrium concentrations (<1 order of magnitude) and few experiment data points investigated in these studies. Other models such as the Brunauer-Emmett-Teller (BET) and Dual-Langmuir models are also inadequate to describe the adsorption process because signi?cant deviation of model estimation from the experimental data was observed at relatively low concentrations.23 The model ?tting results indicate that the adsorption process of organic chemicals on CNPs (i) is neither monolayer formation on a homogeneous surface (i.e., Langmuir model) nor simple multilayer formation (i.e., BET model), (ii) is not a combination of partition and Langmuir-type adsorption domains (i.e., partition-adsorption model), and (iii) cannot be limited by two types of adsorption sites (i.e., Dual-Langmuir model). Polanyi theory-based equations seem not only applicable for pore ?lling but also applicable for ?at surfaces,122 supported by the fact that both the

capacity terma K [L/g] is the af?nity coef?cient. Kf [(mg/g)/(mg/L)1/n] is Freundlich af?nity coef?cient, and 1/n is the Freundlich exponential coef?cient. KL [mg/L] is the af?nity coef?cient. KP [L/g] is the partition coef?cient; KL [mg/L] is the af?nity coef?cient. εsw [kJ/mol], εsw ) -RT ln(Ce/Cs), is the effective adsorption potential, where R[8.314 × 10-3 kJ/(mol K)] and T[K], are universal gas constant and absolute temperature, respectively; Vs [cm3/mol] is the molar volume of solute; a [(cm3)b+1/(kg Jb)] and b are ?tting parameters of PMM model; E [kJ/mol] is the “correlating divisor”; and b is the ?tting parameter of DA model. 23, 62

refs

qe ) KCe
1/n qe ) KfCe

45, 54, 62, 82, 83, 90, 91, 113, 116, 119-121 115-120 114 23, 45, 50-55

Langmuir model partition-adsorption model Polanyi theory-based model

qe ) Q0Ce /(KL + Ce) qe ) KPCe + Q0Ce /(KL + Ce)
Polanyi-Manes model (PMM):

log qe ) log Q + a(εsw /Vs)
0

b

or Dubinin-Ashtakhov (DA) model:

log qe ) log Q0 + (εsw /E)b

a qe [mg/g] is the equilibrium adsorbed concentration of solute; Ce [mg/L] is equilibrium solution phase concentration of solute; Q0 [mg/g] is the maximum sorption capacity of solute; Cs [mg/L] is the solute solubility in water.

Polanyi Theory and Its Application

Chemical Reviews, 2010, Vol. 110, No. 10 5995

Freundlich model and the Polanyi theory-based models had good ?ts for a few cases.23,50 The good ?tting of Polanyi theory-based models for organic chemical isotherms by CNPs was also mechanistically supported by the observed characteristic curve of Polanyi theory.23 Though the Polanyi theory-based PolanyiManes model (PMM) (Table 1) has good ?tting for the isotherms, the Polanyi theory-based DA model (Table 1) was later employed to ?t the isotherms because molar volume (Vs) alone as the abscissa scaling factor failed to obtain a single correlation curve.23 From the DA model (Table 1), it should be not surprising that the Freundlich equation is applicable for a few cases because the Freundlich equation is a special form of DA model at b ) 1.23,50 The linear model is also a special form of DA model, where b ) 1 and E ) RT ln10 ≈ 5.71 kJ/mol. Therefore, the DA ?tted isotherms with b > 1 has been called as a “typical Polanyi isotherm”, showing nonlinear isotherms even in a log-log scale plot.23,50 To avoid large deviations from the true isotherm in ?tting of sorption data as shown in Figure 4, a number of experimental data points in an extended range of equilibrium concentrations of organic chemicals should be employed. In our opinion, more than 20 experimental data points for an organic chemical distributed in the range of its equilibrium concentrations up to its solubility (Cs) from 10-5-10-3 Cs should be employed in sorption experiments and isotherm ?tting.23 However, it is not easy to get experimental data in such an extended range of equilibrium concentrations for some organic chemicals because of their low solubility and the detection limits of the commonly used analytical techniques. Polanyi theory-based DA and PMM models have

three ?tting parameters, more than that of the commonly used linear, Freundlich, and Langmuir models (Table 1). It has been reported that good isotherm ?tting of models may result from overparameterization of PMM and DA models,37 and thus the effect of overparameterization should be evaluated in isotherm ?tting using different models.23 Mean weighted square error (MWSE), equal to 1/V[(qmeasured - qmodel)2/ q2measured], could be calculated to evaluate the goodness of model ?tting and the effect of overparameterization,23 where V is the degree of freedom (V ) N - 2 for Freundlich, Langmuir, and BET models, V ) N - 3 for partitionadsorption, PMM, and DA models, and V ) N - 4 for DualLangmuir model), N is the number of experimental data points, qmeasured is the measured equilibrium sorbed concentration, and qmodel is the estimated equilibrium sorbed concentration by the respective models. With MWSE evaluation, the best ?tting of Polanyi theory-based PMM and DA models as compared with other models for organic chemical isotherms on CNPs was demonstrated to be general and not a result from the model overparameterization.23

6. Polanyi Theory-Based Dubinin-Ashtakhov (DA) Model and Its Parameters
Polanyi theory-based DA equation (Table 1) has three ?tting parameters, i.e., Q0, E, and b. Figure 5 illustrates how the values of Q0, E, and b in?uence the shape of the log-log scale plots of solute isotherms on a sorbent. The increase in E value is a result of the fact that the enhancement rate of the solute adsorption is higher at relatively low solute concentrations (Ce/Cs) than at relatively high solute Ce/Cs

Figure 5. In?uence of the increase in Q0, E, and b on the shape of the log-log scale plots of solute isotherms on a sorbent at various relative solute concentrations (Ce/Cs): (A) E ) 5.71, 10, 20, and 30 at Q0 ) 60 and b ) 1; (B) E ) 10, 15, 20, and 30 at Q0 ) 60 and b ) 1.5; (C) b ) 1, 1.5, 2.0, and 3.0 at Q0 ) 60 and E ) 20; (D) Q0 ) 20, 40, 60, and 100 at E ) 20 and b ) 1.5.

5996 Chemical Reviews, 2010, Vol. 110, No. 10

Yang and Xing

competitors are added or not because the DA model ?tted E and b value of isotherms in the presence of competitors50 were included in Figure 6 and eq 1.

E ) 9.373((1.152) × b - 1.309((1.812) r2 ) 0.6233

(1)

Figure 6. Linear intrinsic relationship (dotted line) between E and b values of DA model ?tted isotherms for adsorption of organic chemicals on CNPs (data from refs 23, 45, 50, 51).

(Figure 5A,B). The increase in b value is a result of the enhancement rate of the solute adsorption on a sorbent that is higher at relatively high solute Ce/Cs than that at relatively low solute Ce/Cs, showing the isotherms are more nonlinear (Figure 5C). The increase in Q0 value is a result of the fact that the enhancement rate of the solute adsorption on a sorbent is relatively equal in the entire solute concentration range (Figure 5D). Among the three ?tting parameters, the saturated adsorbed solute capacity, Q0, is limited by the available adsorption space of sorbents. The maximum potential adsorption space of sorbents is related only to the nature of adsorbent and independent of the nature of chemicals.123 However, the available adsorption space of sorbents can varied with chemicals because of the heterogeneous nature of most sorbents, which could result in the variation of solute Q0 with solute structures on a given sorbent. The “correlating divisor”, E, is the adsorption energy for a given solute, which represents all of the interaction forces responsible for adsorption49,51 and thus could be used to identify the adsorption af?nity. A signi?cant linear intrinsic relationship between the DA model ?tted E and b values has been observed for adsorption of aniline, phenol, and their substitutes on MWCNTs.51 This linear intrinsic relationship at the 95% con?dence intervals can also be observed even if the DA model ?tted E and b data of isotherms of other organic chemicals such as PAHs on various CNPs are included,23,45,50 as is shown in Figure 6 and eq 1. These CNPs include fullerene, SWCNTs, MWCNTs, and oxidized CNTs. Moreover, this linear relationship is independent of whether

This linear intrinsic relationship suggests that the b of the DA model is also a parameter to represent the adsorption energy and the interaction forces responsible for adsorption, as is similar to the parameter E of the DA model. According to the characteristic curve of Polanyi theory in Figure 7, the DA model ?tted E and b may also describe the distribution of sorption site energy. The increase in E value means the more increase in the fraction of relatively high sorption energy than the fraction in relatively low sorption energy (Figure 7A), with respect to the more increase of solute sorption at relatively low solute Ce/Cs than at relatively high solute Ce/Cs (Figure 5B). The increase in b value means the increase in the fraction of relatively low sorption energy and the decrease in the fraction of relatively high sorption energy (Figure 7B), respective to the increase of solute sorption at relatively low solute Ce/Cs and the decrease of solute sorption at relatively high solute Ce/Cs (Figure 5C).

7. Influences of CNP Structures and Surface Properties on Adsorption 7.1. Surface Area
Different synthesis, puri?cation, and postprocessing methods produce CNPs with different physical characteristics, which may control their adsorption for organic chemicals. Yang et al.23 found a high adsorption capacity of PAHs (i.e., pyrene, phenanthrene, and naphthalene) with different types of CNPs (i.e., SWCNTs, MWCNTs, and fullerene). In this study, the PMM model ?tted saturated adsorption capacity, Qv0 (cm3/g), for a given chemical (i.e., phenanthrene) had an order of SWCNTs >MWCNTs > fullerene and seems to have a linear relationship with the surface area (Asurf) of CNPs. Figure 8 shows the linear relationships between Qv0 and Asurf for investigated organic chemicals in several previous studies (data of Qv0 and Asurf as well as their sources are listed in Table 2). These linear relationships indicate that the maximum adsorption capacity of a given organic chemical on CNPs depends on the CNP surface area. Dispersion

Figure 7. In?uence of the increase in E and b on the shape of solute “characteristic curves” of Polanyi theory: (A) E ) 10, 15, 20, and 30 at Q0 ) 60 and b ) 1.5; (B) b ) 1, 1.5, 2.0, and 3.0 at Q0 ) 60 and E ) 20.

Polanyi Theory and Its Application

Chemical Reviews, 2010, Vol. 110, No. 10 5997

Figure 8. Maximum sorption capacity (Qv0) of organic chemicals on CNPs versus surface area (Asurf) values of CNPs. The solid and dotted lines represent the linear relationships between Qv0 and Asurf for a given organic chemical.

of large CNP aggregates to smaller aggregates can increase the surface area of CNPs and consequently be expected to enhance the organic chemical adsorption,6,56,124 although no direct experiments have been done to examine this enhancement of adsorption at the present because of the dif?culty of separating adsorbate supernatant from CNTs suspension. The slopes of linear relationships between Qv0 and Asurf are different signi?cantly for various organic chemicals (Figure 8), probably because part of the measured surface area of the given CNP could not be occupied by organic molecules due to the heterogeneity of the CNP surface.56 Moreover, the available surface area of a given CNP for adsorption would be different for organic chemicals because different chemical molecular structures have different interactions with the heterogeneous surface of CNPs, as discussed in detail in section 8, Effects of Organic Chemical Structure and Properties on Adsorption. Figure 9 shows the heterogeneous nature of CNP surface and the possible in?uence on solute adsorption. The ratio of the available surface area to the total measured surface area could also be different for CNPs, although most of these ratios of CNPs for adsorption of a given organic chemical should be close according to the linear relationships between Qv0 and Asurf (Figure 8). For example, the observed data of Asurf of SWCNT and the Qv0 of phenanthrene on the SWCNT could not follow the observed linear relationship between Qv0 and Asurf for phenanthrene on other CNPs.23

7.2. Functional Groups on CNP Surface
CNPs can be used as superior adsorbents for organic chemicals basically because of the hydrophobic nature of CNP surfaces.79 Functional groups such as -OH, -CdO, and -COOH on the CNP surface, derived from the processes

of synthesis and puri?cation of CNPs, therefore, could be mainly responsible for the variation of the ratio of the available surface area to the total measured surface area among different CNPs. Functional groups on CNP surface can also be intentionally added by oxidation treatment42-45 or removed by heat treatment.121 Functional groups of CNPs can form H-bonds with water molecules, which can consequently make CNPs more hydrophilic to suppress the adsorption of organic chemicals especially HOCs through the competition of water molecules.79,97-99,125 The competition of water molecules could be employed to interpret the reported phenomenon that oxidization of CNTs decreased their adsorption for naphthalene,45 chlorophenol,115 and resorcinol.116 Functional groups of CNPs could also be expected to act as hydrogen-bonding donors and form hydrogen bonds with organic molecules as hydrogen-bonding acceptors. Relative to the hydrogen-bonding interaction between functional groups of CNPs and water molecules,97-99 however, the hydrogen-bonding interaction between functional groups of CNPs and organic solutes should be too weak to occur in the aqueous phase, as indicated by the decreased adsorption of organic solutes from the competition of adsorbed water molecules97,98 and no signi?cant effects of hydrogen-bonding interaction on the adsorption of solutes without hydrogen-bonding donor ability.51,83 The functional groups such as carboxylic groups can also act as the CNTs electron acceptors, which could consequently promote phenanthrene adsorption via enhancing the π-π bond in the electron donator-acceptor (EDA) system between the CNTs and phenanthrene in organic solvents such as benzene.44 This may not occur in aqueous systems because of the competition of adsorbed water molecules, too.97,98,125 Among the interactions for adsorption of organic chemicals on CNPs in aqueous

5998 Chemical Reviews, 2010, Vol. 110, No. 10
Table 2. Reported Surface Area Values of CNPs and Their Maximum Sorption Capacities for Organic Chemicalsa chemicals phenanthrene CNPs MWNT15 MWNT15 fullerene MWNT8 MWNT15 MWNT30 MWNT50 MWNT40 HA-MWNT40 PH-MWNT40 PE-MWNT40 MWNT15 r-SWNT l-SWNT d-SWNT CNTs CNTs CNTs MWNT40 HA-MWNT40 PH-MWNT40 PE-MWNT40 MWNT15 MWNT15 MWCNT fullerene SWNT MWNT15 MWNT30 MWNT50 fullerene SWNT MWNT15 MWNT30 MWNT50 MWNT15 MWCNT10 MWCNT100 pristine MWCNTs as-grown CNTs graphitized CNTs SWNT MWNT15 r-SWNT l-SWNT d-SWNT r-SWNT l-SWNT d-SWNT MWCNT10 MWCNT100 MWNT40 HA-MWNT40 PH-MWNT40 PE-MWNT40 MWCNT10 MWCNT100 pristine MWCNTs Asurf (m2 g-1) 174 174 7.21 348 174 107 94.7 86.97 62.85 70.04 35.4 174 447 289 271 370 370 370 86.97 62.85 70.04 35.4 174 174 286 7.21 541 174 107 94.7 7.21 541 174 107 94.7 174 357 58 72 134 126 370 174 447 289 271 447 289 271 357 58 86.97 62.85 70.04 35.4 357 58 72 Q0 (mg/g) 41.7b 44.7c 0.0678 86.5 42.1 26.9 23.6 Qv0 (cm3/kg) 39.2 42.0 0.0637 81.4 39.6 25.3 22.2 20.8 21.6 17.2 14.9 74.0 163 96.4 80.0 108 102 93.0 21.7 19.5 20.8 10.5 51.4 40.9 96.0 0.202 273 118 79.2 69.4 2.01 495 101 64.4 86.2 60.3 388 70.1 14.8 137 99.4 236 99.8 282 183 156 405 295 267 593 110 193 170 143 49.2 266 53.7 10.6

Yang and Xing

refs 50 50 23 23 23 23 23 54 54 54 54 23 91 91 91 90 90 90 54 54 54 54 23 23 45 53 53 53 53 53 53 53 53 53 53 51 113 113 116 121 121 83 51 91 91 91 91 91 91 113 113 54 54 54 54 113 113 116

naphthalene

73.8 163I 96.1I 79.8I 108I,d 102I,e 92.7I,f

51.3g 40.7h 0.228 309 133 89.5 78.4 2.40 591 121 77.0 103 64.6 416I 75.1I 15.9j 178I 129I 307I 138 390I 253I 216I 494I 360I 326I 719I 134I

17 R-ethinyl estradiol (EE2)

bisphenol A (BPA)

phenol

1,2-dichlorobenzene (1,2-DCB)

2.4-dichlorophenol (2.4-DCP)

2-naphthol

1-naphthol

catechol

358I 72.2I 14.2j

a Asurf, surface area; Q0, maximum sorption capacity in gravimetric unit; Qv0, maximum sorption capacity in volumetric unit, which was calculated from Q0 and the density (F) of the chemical with the equation of Qv0 ) Q0/F if the Qv0 values are not directly available in the references; the density (F) of phenanthrene, naphthalene, EE2, BPA, phenol, 1,2-DCB, 2.4-DCP, 2-naphthol, 1-naphthol, and catechol are 1.063, 0.997, 1.13, 1.195, 1.071, 1.3, 1.383, 1.22, 1.213, and 1.344 g cm-3, respectively. b In presence of naphthalene as the competitor. c In presence of pyrene as the competitor. d In presence of Na+ as the competitor. e Iin presence of Cu2+ as the competitor. f In presence of humic acid as the competitor. g In presence of pyrene as the competitor. h In presence of phenanthrene as the competitor. I Q0 was the value of qe at Ce ) Cs, which was calculated from the Freundlich model ?tted isotherms. j Q0 was obtained directly from the Langmuir model ?tted isotherms.

Polanyi Theory and Its Application

Chemical Reviews, 2010, Vol. 110, No. 10 5999

Figure 9. Possible adsorption of solute 1 (X) and solute 2 (x) on the CNP surface with different potential adsorption sites (black solid circle, site 1; dark gray solid circle, site 2; light gray solid circle, site 3; open circle, site 4). The heterogeneous nature of the CNP surface is presented by four different adsorption sites.

systems, therefore, hydrophobic interaction, hydrogen-bonding interaction, and π-bonding interaction can enhance the adsorption af?nity (E) but not the adsorption capacity (Q0) of organic chemicals on CNPs because all of these interactions can occur on the sites of CNP surface without functional groups but not the functional group occupied sites of CNP surface, which prefer to bond with water molecules. Therefore, isotherms of naphthalene on oxidized CNTs have almost a constant E value but decreased Q0 values with increased oxygen content of CNTs45 because only hydrophobic interaction and π-bonding interaction could occur for naphthalene adsorption on CNTs. However, functional groups of CNPs could form covalent bonds with organic chemicals in an aqueous system if these chemicals have certain functional groups such as -COOH and -OH,78,100,101 which may add additional adsorption for organic chemicals and enhance both of the adsorption af?nity (E) and the adsorption capacity (Q0) of organic chemicals on CNPs. In addition, increase in numbers of functional groups on CNP surface may result in the conformation changes of CNP aggregates and thus the changes of exposed CNP surface area,44,67,68,70,79,105,125 which should in?uence their adsorption for organic chemicals. The increase in adsorption of trihalomethanes119 and 1,2-dichlorobenzene121 after introducing functional groups to CNTs by oxidation treatment could possibly stem from the increase in exposed surface area of CNTs. The normalization of adsorbed capacity with the surface area of CNPs may be a useful way to rule out the in?uence of CNP surface area changes derived from oxidation on adsorption and thus help us to explore the interactions of organic solutes with functional groups of CNPs. In addition, surface functional groups on CNPs may possibly lead to the desorption hysteresis of organic chemicals from CNPs, which is discussed in detail in section 10, Adsorption-Desorption Hysteresis. Moreover, surface modi?cation of CNPs with other chemicals can alter the adsorption of organic chemicals on CNPs, which is discussed in section 9, Competitive Sorption and the Role of DOM and Surfactants in Adsorption.

closed or too narrow for the adsorbing organic molecules.6,53 This could be employed to explain why activated carbons (a similar material to CNPs) have commonly much higher surface area than CNPs but lower surface area-normalized adsorption capacity for organic chemicals relative to CNPs;51,56 because activated carbons are porous materials having a large quantity of narrow micropores, they are inaccessible for organic molecules. Also, the higher adsorption of DOM on MWCNTs can be attributed to the larger average pore diameter and volume of MWCNTs than activated carbon.126 Furthermore, the pores in CNP aggregates (especially fullerene) may possibly result in the desorption hysteresis of organic chemicals, which is discussed in detail in section 10, Adsorption-Desorption Hysteresis.

7.4. Surface Curvature and Defects
The surface curvature93 and defects of CNP monomers127 could affect their adsorption for organic solutes. Surface curvature of CNP monomers increases with their decreasing diameter. The ?at surfaces of CNPs can result in more surface atoms approaching to adsorbing organic molecules, which lead to higher adsorption potential of the ?at CNP surfaces and thus stronger interactions according to the Polanyi theory.46,57 Defects of CNP monomers can also result in more surface atoms of CNPs interacting with organic molecules and thus higher adsorption potential for organic molecules.48 For a given organic chemical, as mentioned above, its adsorption on CNPs is dependent on the CNP structures and surface properties such as surface area, surface functional groups, surface curvature, surface defects, and the pores in CNP aggregates. The good ?tting of Polanyi theory-based DA model for organic chemical isotherms on various CNPs indicates that the parameters of the DA model can be related with the CNP structures and surface properties to explore their roles on adsorption.23,45,53 These relationships have been observed between Q0 and CNP surface area,23,53 between Q0 and CNP surface oxygen content,45 and between E (or b) and the Rmeso/micro ratios of CNPs.23 Other relationships such as Q0 and E (or b) with the CNP structures and surface properties have not yet been obtained. Surface functionalization of CNPs with oxygen-containing groups may decrease the noncovalent adsorption of organic chemicals because of the competition of water but could increase the adsorption of other organic chemicals if they can form covalent bonds with these groups. Therefore, surface functionalization on CNPs is a primary step for the covalent modi?cation of CNPs to form a variety of nanostructures with excellent physical and chemical properties. However, the role of a speci?c functional group (such as hydroxyl, carboxyl, or ketone group) in the covalent adsorption is still unclear. Comparison of the adsorption properties between CNPs containing different functional groups will help us to understand the role of a speci?c functional group in covalent adsorption. According to the dependence of adsorption capacity (Q0) of

7.3. The Pores in CNP Aggregates
Yang et al.23 reported that the DA model ?tted E and b were linearly related to the ratios of mesopore volume to micropore volume (Rmeso/micro ) Vmeso/Vmicro) of CNPs. According to Polanyi theory and the “characteristic curve”, the measured micropore volume could describe approximately the site fraction with relatively high sorption energies on the solid surface, while the measured mesopore volume could describe approximately the site fraction with relatively low sorption energies. Therefore, it is not surprising that the DA model ?tted E and b have linear relationships with Rmeso/micro ratios because both E and b of the DA model describe the distribution of the sorption energies on the solid surface as mentioned above (Figure 7). The pores (e.g., interstitial spaces and grooves) in CNP aggregates may not be available for adsorption of organic molecules if they are

6000 Chemical Reviews, 2010, Vol. 110, No. 10

Yang and Xing

organic chemicals on the surface area of CNPs, the removal of organic pollutants from water by CNPs could be enhanced by enhancing CNP surface area when they are engineered.

8.2. Molecular Size
Molecular size of organic chemicals is another critical factor affecting their adsorption capacity on CNPs although it could not in?uence their adsorption af?nity (E).51 Therefore, PAHs (i.e., napththrene, phenanthrene, and pyrene) have the same types of interaction with CNPs (i.e., hydrophobic interaction and π-bonding interaction), while their adsorption capacities (Qv0) on a tested MWCNT sample are varied.23 First, the volume adsorption capacity of a given organic chemical on the planar surface of CNPs should be higher if the adsorption is a result of the attachment of the end of the organic molecule (Figure 10b) on the CNP surface rather than that of the attachment of the side of the organic molecule (Figure 10c). Second, larger molecules (Figure 10b) have higher volume adsorption capacity than relatively small molecules (Figure 10a) on the planar surface of CNPs if they have a same attachment orientation. For example, the adsorption capacity of big DOM and surfactants on CNTs32,129 is much higher than that of PAHs and other smaller organic chemicals.23,51 However, larger molecules should have lower adsorption capacity in the pores of CNP aggregates than relatively small molecules because of the bottleneck of pores for organic chemical diffusion, i.e., organic molecules cannot enter into the pores with smaller size than the molecules. For example, small molecules (i.e., water and ethanol) have higher adsorption capacity (10 times more) than the bigger molecules (i.e., hexane) on CNPs.95 Also, small molecules (i.e., water and ethanol) have lower diffusion rates than the bigger molecules (i.e., hexane) because the diffusion of small molecules into the inner pores of CNP aggregates is a ratelimiting process and because most of bigger molecules cannot diffuse into the inner pores of CNP aggregates.95 With the bottleneck effect of CNP pores, larger differences in chemical molecular size result in better separation in the mixed system of organic chemicals.130 Another possibility for larger molecules having lower adsorption capacity on CNPs is that the adsorbed phase may be different from the corresponding bulk phase.37,131 The change of organic chemicals from bulk solid to adsorbed phase, as an example, always results in the expansion of chemical volume and density decrease. The volume expansion and density decrease are greater for big molecules as compared to smaller molecules, thus the adsorbed volume capacity (calculated from solid phase density) of a big molecular chemical is lower than smaller molecular chemicals and its actual adsorbed volume.23 In addition, con?guration of adsorbed organic molecules on CNPs may be different relative to bulk organic molecules, especially for the large molecules containing long aliphatic chains, because the long aliphatic chains are relatively “soft” as compared with the aromatic rings and easier to deform

8. Effects of Organic Chemical Structures and Properties on Adsorption 8.1. Molecular Structure and Functional Groups
Different molecular structures of organic chemicals have different interactions with the heterogeneous surface of CNPs, which leads to the signi?cant variation in the slopes of linear relationships between Qv0 and Asurf for various organic chemicals (Figure 8). It has been observed in a previous study by Chen et al.83 that adsorption af?nity of polar and nonpolar organic compounds on CNTs increased in the order of nonpolar aliphatic < nonpolar aromatics < nitroaromatics. Cyclohexane is a typical chemical which cannot form hydrogen and π-π bonds with CNPs. Therefore, hydrophobic interaction derived from van der Waals force is responsible for adsorption of cyclohexane on CNPs,51,83 as shown by the linear isotherm.83 The strength of hydrophobic interaction for organic chemicals can be determined by their solubility in water.128 Normalization of chemical adsorption with their KOW or KHW values can largely eliminate their hydrophobic effects on adsorption; by this treatment, the in?uence of other interactions on adsorption may become apparent.83 For aromatic hydrocarbons, besides the hydrophobic effect, π-bonding interaction can play a signi?cant role in their adsorption on CNPs.22,43,51,83 Functional groups such as -COOH, -OH, and -NH2 in organic molecules can form hydrogen bonds with CNPs and promote the adsorption of these organic molecules.51,95 The in?uence of functional groups on adsorption of organic chemicals by CNPs depends on the type, number, and position of functional groups in the organic molecules. More functional groups in a molecule result in higher adsorption af?nity (E) on CNPs.51,83,116 Adsorption af?nity of aniline and phenol on CNTs is altered, with their substituted groups at a given position with the following order: nitro group > chloride group > methyl group.51 The position of substituted groups in an organic molecule could also in?uence the adsorption.51,116 However, these effects are weaker than that of the type of substituted groups and cannot be interpreted directly by the group position.51 Molecular con?gurations of organic chemicals may also play a role in their adsorption. For planar chemicals, SWCNTs have signi?cantly higher adsorption capacities and site energies than MWCNTs, whereas for nonplanar chemicals, the adsorption capacity and site energy differences between SWCNTs and MWCNTs became smaller with increasing concentration of organic chemicals.125

Figure 10. Adsorption of smaller molecule (a) and bigger molecules via the attachment of the end of the organic molecule (b) or the side of the organic molecule (c) on the planar surface of CNPs.

Polanyi Theory and Its Application

Chemical Reviews, 2010, Vol. 110, No. 10 6001

(or bend). These large molecular chemicals reported include enzymes,132,133 lipid derivatives,74 L-phenylalanine,101 natural organic matter,32,134,135 and surfactants.129 Conformational changes of organic molecules can promote their matching with the curvature surface of CNPs to enhance their stability on the CNP surface.74,132

8.3. Chemical Solvatochromic Parameters
Solvatochromic parameters, in general, including the intrinsic molar volume (VI), polarity/polarizability (π*), hydrogen-bonding accepting ability ( m), and hydrogenbonding donor parameter (Rm), are useful descriptors for predicting the properties, toxicity, mobility, and environmental behavior of various organic chemicals.49,51,136-139 These solvatochromic parameters of organic compounds can be estimated using quantitative structure-activity relationship (QSAR) modeling based on their functional and structural characteristics.136,138 The ability of organic chemicals to form π-π bonds with sorbents can be identi?ed by their π-electron polarizability parameter (π*).49,51 Chemicals with more aromatic rings in their molecules have higher π* value and lower water solubility,136-138 and thus they are more hydrophobic and can form stronger π-bonding interaction with CNPs.51 The ability of organic chemicals to have hydrogenbonding interaction with sorbents can be evaluated by their hydrogen-bonding acceptor or donor parameters ( m or Rm).49,51 Functional groups attached to organic molecules and the type, number, and position of functional groups in the organic molecules can alter the solubility of organic chemicals in water136,137 and their values of m, Rm, and π*.136-138 The linear solvation energy relationships (LSERs) model based on the solvatochromic parameters has been developed to predict adsorption af?nity (E) of aromatics, halogenated aliphatics, and halogenated aromatics onto solid sorbents from aqueous solution by correlating solvatochromic parameters with the adsorption af?nity values.49 These solid sorbents include coal-based activated carbons, coconut shell based activated carbon, unspeci?ed activated carbon, and synthetic polymeric adsorbents. This LSERs was observed for the adsorption of aniline, phenol, and their substitutes on a MWCNT sample (MWCNT15) as given in eq 2.51

CNPs. Chemical molecular size (VI) and hydrogen-bonding acceptor ability ( m) have almost no in?uence on their adsorption af?nity. The limited role of organic chemical m in their adsorption on CNPs could be attributed to the graphite sheet of carbon nanotubes that is not a hydrogenbonding donor.140 Moreover, functional groups of CNPs prefer to adsorb water molecules through hydrogen-bonding interaction, which is responsible for screening out the role of organic chemical m in their adsorption on the functional surface sites of CNPs.79,97-99

8.4. Dissociation of IOCs
IOCs including organic acids such as phenolic chemicals and organic bases such as aniline-type chemicals are different from nonionizable chemicals because they can exist either as neutral or dissociated species in the aqueous phase.51 The occurrence of their neutral or dissociated species depends on the solution pH in relation to their dissociation constants (pKa). At pH < pKa, the neutral and the dissociated species (bases form cations by protonation) are dominant for organic acids and organic bases, respectively, while at pH > pKa, the dissociated species (anion) is dominant for organic acids and the neutral species is dominant for organic bases. The fraction of neutral species for organic acids (fAN) and bases (fBN) can be respectively estimated by fAN ) (1 + 10 pH-pKa)-1 and fBN ) (1 + 10pKa-pH)-1, and the fraction of dissociated species for organic acids (fAI) and bases (fBI) can be estimated by fAI ) (1 + 10pKa-pH)-1 and fBI ) (1 + 10 pH-pKa)-1, respectively.51 Neutral species of IOCs have higher adsorption than dissociated ones although both of neutral and dissociated species can adsorb on CNPs.51 Therefore, adsorption of IOCs on CNPs is pH-dependent, showing that adsorption of organic acids is decreased with increasing pH51,82,113,115,141 while adsorption of organic bases is increased with increasing pH.51,142 For zwitterionic chemicals such as nor?oxacin, their adsorption on CNPs could be either increased or decreased with pH because they can be dissociated to be cationic or anionic species.55 Increasing pH can also decrease the adsorption of NOM on CNPs31,32,143,144 because NOM contains abundant -COOH and phenolic groups and they are ionizable similarly to organic acids. The lower adsorption of dissociated species of IOCs on CNPs as compared with their neutral species can largely be attributed to the disappearing of hydrogen-bonding donor ability and higher solubility (i.e., lower hydrophobic effect) of dissociated species.145 Furthermore, electrostatic repulsion between negatively charged NOM and CNP surface with increased pH could rise to decrease the NOM adsorption.31,32 For a given solute concentration, pH dependence of the apparent adsorption coef?cients Kd () qe/Ce) of organic acids and organic bases can be estimated from the following eqs 4 and 5, respectively.51 For organic acids,
N I Kd ) Kd × 1/(1 + 10(pH-pKa)) + Kd ×

E ) 11.58((2.03) × Rm + 8.15((3.64) × π* + 5.50((3.42) (r2 ) 0.785) (2)

The DA model ?tted E data of isotherms of other organic chemicals (i.e., PAHs) on MWCNT15 reported by Yang et al.23 also follow eq 2. In this equation, the constant intercept, 5.50 ( 3.42, is close to the DA model ?tted E value (5.71) of linear isotherm (b ) 1) and thus can be attributed to the hydrophobic effects of organic chemicals on adsorption.51 The linear isotherm, derived from the DA model (Table 1) where E and b are given to be 5.71 and 1, respectively (eq 3), shows that the solute adsorption is limited by their solubility values (Cs) only for a given sorbent having a constant Q0.

1/(1 + 10(pKa-pH))
and for organic bases,

(4)

qe ) Q0Ce /Cs

(3)

N I Kd ) Kd × /(1 + 10(pKa-pH)) + Kd × 1/(1 + 10(pH-pKa)) (5)

According to eq 2, therefore, besides the constant hydrophobic effect of organic chemicals, chemical π-electron polarizability (π*) and hydrogen-bonding donor ability (Rm) can play an important role in their adsorption af?nity on

where KdN and KdI are the adsorption coef?cients of neutral species and dissociated species, respectively. These two models assume that the value of either KdN or KdI is constant

6002 Chemical Reviews, 2010, Vol. 110, No. 10

Yang and Xing

over the whole pH range considered for a given initial concentration. Adsorption increase with increasing pH was observed for pyrogallol and 1-naphthol,113 and 2-naphthol and 1-naphthylamine90 on CNPs, which is inconsistent with the reported results for other IOCs.51,82,90,115,141 The electron donor-acceptor (EDA) interactions were suggested by Chen et al.90 to be the primary mechanism for the pH-induced adsorption increase of naphthol and naphthylamine. One type of EDA interaction is the π-π EDA interaction because hydroxyl group of naphthol and amino group of naphthylamine as π-electron donating groups make the benzene ring(s) of naphthol and naphthylamine π-electron-rich, thus allowing these two compounds via their benzene ring(s) to interact more strongly with the electron-depleted surfaces of carbon nanotubes and graphite (π-electron acceptor). Another EDA interaction was the n-π EDA interaction because oxygen electron pairs of hydroxyl group and nitrogen electron pairs of amino group as n-electron donors might have directly interacted with the electron-depleted sites (π-electron acceptor) of carbon nanotubes and graphite. The strength of both π-π and n-π EDA interactions are enhanced when hydroxyl and amino groups are ionized to their dissociated species because of the fact that their dissociated species are even stronger electron-donating groups, and thus adsorption of naphthol and naphthylamine on the graphene surface of CNPs are increased as pH increases. In addition to the EDA interaction, Lewis acid-base interaction was also suggested by Chen et al.90 as an extra possible mechanism contributing to the pH increased adsorption of 1-naphthylamine, especially on the oxidized carbon nanotubes. However, the observed increase in adsorption of pyrogallol, naphthol, and naphthylamine with increased pH below their pKa could be due to other possible reasons. First, the pH-dependent effect of oxygen-induced oxidative coupling could lead to the adsorption of these compounds to CNTs/graphite.146 Second, the pH-dependent oxidization of chemicals such as 1-naphthol in solution was reported,147-149 and this oxidized loss could have accounted for increased adsorption with increasing pH, at least for naphthol. Third, measurement of IOCs in solution using the UV-vis spectrometer is generally dependent on the pH of solution because the neutral species and dissociated species of IOCs have different UV absorbance.51,128 The pH of supernatants has to be adjusted to the pH value at which the calibration curve was obtained. In addition, adsorption increase with increased pH was also reported for nonionic nitroaromatic compounds including 2,4-dinitrotoluene,83 1,3-dinitrobenzene, and 1,3,5trinitrobenzene.91 This increase was also interpreted by Chen et al.83,91 with the EDA interaction, for which the nitro-group (the strong electron-withdrawing group) of these aromatics make their benzene ring(s) electron depleted and act as π-electron acceptors while the oxygen containing functional groups (the electron-withdrawing groups) on surfaces of CNTs not only act as n-electron donors themselves but also make the CNT graphene surface to be π-electron-rich and thus π-electron donors. The pH-promoted electron-donor ability of oxygen containing functional groups on CNT surfaces as a result of their dissociation (e.g., -COOH is a weak electron donor and acceptor, while -COO- is a strong electron donor but not an acceptor) is thus suggested to be responsible for the increased adsorption of these nitroaromatics with increasing pH. However, this increase was not

Figure 11. Effects of pH on the adsorption of 1,3-dinitrobenzene on graphitized, puri?ed, hydroxylated, and carboxylated MWCNTs. The surface oxygen atomic percents of graphitized, puri?ed, carboxylated, and hydroxylated MWCNTs determined by X-ray photoelectron spectroscopy (XPS) method are 0, 1.9, 6.0, and 6.4%, respectively.

observed for other nitroaromatics (π-electron acceptors) on CNTs51 and activated carbons.150,151 For example, the adsorption of nitrobenzene on oxidized and unoxidized activated carbon surfaces did not change signi?cantly with the change of the solution pH from 2 to 12.150,151 The adsorption of 1,3dinitrobenzene, as another example, on graphitized, puri?ed, hydroxylated, and carboxylated MWCNTs also did not change signi?cantly with pH in the range of 2-12 (Figure 11). These observations suggest that acidic functional groups on CNP surfaces could not form EDA interactions with nitroaromatics in the aqueous phase. The much stronger interactions of water molecules to form hydrogen bonds with the acidic functional groups on CNT surfaces97-99 would suppress the EDA interaction of nitroaromatics with oxidized CNTs, thus lower surface area normalized adsorption on the CNTs with more surface oxygen atoms was observed (Figure 11). Adsorption of various organic chemicals on a given CNP, as mentioned above, is a function of the chemical structures and physicochemical properties. The good ?tting of Polanyi theory-based DA model for the isotherms of various organic chemicals indicates that the parameters of DA model can be related with the chemical structures and properties to probe their in?uence on adsorption. For example, empirical relationships between E (or b) and chemical solvatochromic parameters can be developed to explore the interactions between organic chemicals and CNPs and their strength for chemical adsorption.51 However, the ability of organic chemicals to form covalent bonds with CNPs, as a parameter, still cannot be predicted quantitatively. Without this parameter, deviation of adsorption estimation using empirical relationships between E (or b) and chemical solvatochromic parameters would occur for organic chemicals that can be adsorbed by CNPs with covalent bonds. Moreover, Q0 of DA model is not yet related with the chemical structures and physicochemical properties. Molecular size could be a potential parameter of organic chemicals to be related with their Q0 values. In addition, the in?uence of pH on adsorption of IOCs are also not related with Q0 and E (or b) of DA model although the pH-dependence of adsorption of IOCs can be estimated from the solution pH and chemical pKa values.51 These relationships should be developed to examine the in?uence of chemical structures and properties on adsorption and to estimate the adsorption in future investigations.

Polanyi Theory and Its Application

Chemical Reviews, 2010, Vol. 110, No. 10 6003

9. Competitive Sorption and the Roles of DOM and Surfactants in Adsorption 9.1. Competitive Sorption of PAHs
Competitive behavior of multiple organic chemicals on activated carbon and natural geosorbents has been widely reported.37,152-155 This behavior is a key issue because multiple pollutants exist concurrently in the environment. Signi?cant competitive sorption between PAHs was also observed on CNTs, showing the isotherm of a given primary solute changed from being signi?cantly nonlinear to nearly linear when competitors were added.50 Moreover, the observed competitive sorption depends on the relative equilibrium concentrations of both primary and cosolutes. Signi?cant competition was present at relatively low concentrations of primary solute and high concentrations of competitors, while competition was much weaker in the case of relatively high concentrations of primary solute and low competitor concentrations. When the relative concentration of primary solute (Ce/Cs) approached 1, competition by other solutes seems to disappear, i.e., the adsorption capacity of a given PAH chemical on CNTs may not be changed with the addition of other PAHs as the competitors. A Polanyi-based surface adsorption mechanism with multilayer adsorption has been proposed to interpret the observed competitive adsorption of multiple organic chemicals (i.e., pyrene, phenanthrene, and naphthalene) on the CNTs.50 The observed signi?cant competition for PAH chemicals on CNTs and the increase of isotherm linearity could be attributed to the weaker π-π electron interaction of the adsorbed PAH molecules with other PAH molecules than that with the CNT graphitic surface because the hydrophobic effect alone on solute adsorption by CNTs presents linear isotherms while additional interactions such as π-π electron interaction result in the isotherm nonlinearity.51

(Q0) and the adsorption af?nity (E) of these organic chemicals on CNTs.54,129 This is different from the observed competition phenomena for PAH chemicals on CNTs,50 in which the Q0 for a given PAH chemical remained practically constant when other PAH chemicals were added as competitors. This competition difference could be caused by the different molecular structure between DOM (or surfactants) and PAHs as well as the structure in?uence on the available surface area of CNTs for adsorption. For DOM and surfactant molecules, they have a long ?exible aliphatic chain in the molecule, while there are only aromatic rings in PAH molecules. Relative to the aliphatic chain, aromatic rings are hard and in?exible. The surface of CNTs occupied by DOM or surfactants could be never available for adsorption of other molecules.129 This hypothesis was supported by the decrease of speci?c surface area of CNTs after they are coated with DOM or surfactants.54,129 Additional evidence for this hypothesis was observed in other studies161,162 that the surface area values of coated clays with cationic surfactants were far smaller than that of the uncoated clays and the speci?c surface area of NOM measured using N2 adsorption-desorption is very small.135,163 If the surface of CNTs is occupied by hard PAH molecules, however, the adsorbed PAH molecule could provide an equivalent surface area to the occupied surface area of CNTs by the PAH molecule, which is available for adsorption of other molecules,50,129 and thus a relatively constant Q0 remains for a given PAH chemical on CNTs with or without other PAH chemicals as competitors.50

9.4. Other In?uence of DOM and Surfactants on Organic Chemical Adsorption
Adsorbed NOM and surfactants on CNPs can also act as a favorable media for organic chemical partition.129,134,135 The additional sorption of organic chemicals by partitioning into the adsorbed NOM and surfactants will counteract the competition induced adsorption decrease.129 Therefore, the apparent sorption of organic chemicals by NOM (or surfactant) coated CNPs can be attributed to: (i) adsorption of organic chemicals on unoccupied CNP surfaces and (ii) additional sorption of organic chemicals partitioning into adsorbed NOM or surfactants.129 Adsorbed surfactants, mainly the surfactant hemimicelles as a favorable partitioning media for organic chemical sorption, should impart linear isotherms of organic chemicals.129,162,164-166 Therefore, isotherm linearity of organic chemicals on surfactant coated CNPs increased with the increase of adsorbed surfactant amounts.129 Moreover, the apparent sorption of organic chemicals by surfactant coated CNPs can be described by the partition-adsorption model (Table 1) including a partition fraction of chemicals into adsorbed surfactant hemimicelles and an adsorption fraction of chemicals on unoccupied CNP surfaces.129 As compared with surfactants, adsorbed NOM can act not only as a favorable partitioning media (presenting linear isotherm) but also as an adsorbent (presenting nonlinear isotherm) for organic chemical sorption.128,134,135,154,155 Con?guration changes of NOM and surfactants might occur after they are adsorbed on CNPs, which can result in that the sorption af?nity and capacity of CNP-adsorbed NOM and surfactants for organic chemicals are different from that of bulk NOM and surfactants and vary with the adsorbed amounts of NOM and surfactants on CNPs.129,134,135 NOM fractionation might also occur upon their adsorption on CNPs, and consequently alter organic chemical sorption on adsorbed NOM as well as NOM-coated

9.2. Competitive Sorption with Metal Ions
Adsorption of organic chemicals such as 2,4,6-trichlorophenol, naphthalene, atrazine, 1,3-dinitrobenzene, and 1,3,5-trinitrobenzene on MWCNTs and SWCNTs can also be suppressed by metal ions (i.e., Cu2+, Pb2+, and Cd2+), as interpreted by the formation of surface complexes of metal ions with oxygen functional groups on CNTs.91,156,157 The suppression of the adsorption of organic chemicals by the metal ions is affected by oxygenated functional group density on CNTs.156 This suppression could also be expected to be more signi?cant for chemicals such as atrazine which could adsorb by forming covalent bonds with CNT functional groups than for other chemicals because: (i) CNT functional groups are mostly responsible for the adsorption of metal ions,158 and (ii) functional group conglutinated CNT surfaces cannot be available for the adsorption of most organic chemicals because of the suppression of water97-99 except for chemicals that can form covalent bonds with CNT functional groups.78,100,101

9.3. Competitive Sorption with DOM and Surfactants
DOM and surfactants can suppress the adsorption of organic chemicals such as PAHs and nitroaromatics on CNP surfaces.54,91,129,159,160 It should be noted that the presence of DOM or surfactants reduced both of the adsorption capacity

6004 Chemical Reviews, 2010, Vol. 110, No. 10

Yang and Xing

CNPs.134,135,167 Coadsorption of metal ions can also change the adsorption and con?gurations of NOM on CNPs because metal ions can bridge NOM with functional groups on CNPs, compress the double layer, and neutralize negative charges of NOM and thus weaken the repulsion between NOM molecules as well as between NOM and CNPs.31,143,144 This should affect the sorption of organic chemicals on NOMcoated CNPs, which needs more investigations. Dispersion of CNPs and solubility enhancement of organic chemicals by DOM and surfactants may also in?uence the organic chemical adsorption on CNPs.54,56,160 It is widely observed that DOM30,34,66,71 and surfactants72-76 can prompt dispersion of CNPs (i.e., deaggregation of big CNP aggregates to smaller CNP aggregates), which could lead to more CNP surface sites exposed for organic chemical adsorption.54,56,160 DOM168-171 and surfactants172-174 in solution could also decrease chemical adsorption on CNPs by enhancing the solubility of organic chemicals. Although adsorption of an organic chemical on CNPs can be suppressed by other organic chemicals, metal ions, DOM, and surfactants, the isotherms of organic chemicals in the presence of these competitors can be ?tted well by the Polanyi theory-based DA model,50,54,129,156 which indicates that the in?uence of competition on organic chemical adsorption can be described by the variation of the parameters of DA model. Even for the other in?uence of DOM and surfactants on organic chemical adsorption, they can be described by the variation of the parameters of DA model, too.129 Empirical relationships could be developed for Q0 and E (or b) with adsorbed amounts of the competitors on CNP surfaces.129 The pH and ionic strength could largely change the adsorption of DOM and surfactants on CNPs,31,32,143,144,175 and thus their impact on the sorption of organic chemicals in the presence of DOM and surfactants needs to be examined. In addition to cationic surfactants,129 anionic and nonionic surfactants have also been employed to synthesize and disperse CNPs.76,176,177 As compared with cationic surfactants, these surfactants could have different sorption on CNPs and thus different effects on sorption of organic contaminants by CNPs. Moreover, the adsorption of surfactants on CNPs and its effects on organic contaminant sorption may depend on the various structures and functional groups of CNPs. Therefore, research using different types of surfactants and CNPs are urgently needed to examine the adsorption effects of surfactants on sorption of organic contaminants.

10. Adsorption-Desorption Hysteresis 10.1. Hysteresis Phenomenon
Adsorption-desorption hysteresis has been widely observed for the sorption of organic compounds by soils, sediments, and charcoals.178-180 Desorption hysteresis shows greater apparent af?nity of adsorbate in desorption branches than adsorption branches (Figure 12). Two types of hysteresis are often observed: reversible and irreversible hysteresis.179,181 Reversible hysteresis means that complete desorption can be achieved without any intervention (e.g., vigorous solvent extraction), showing a closed hysteresis loop in desorption and adsorption branches of an isotherm (Figure 12A). Closed hysteresis loop is commonly observed for vapor sorption, which is attributed to formation of metastable states and network percolation effects, associated with capillary liquid condensation in ?xed mesopores.181 It was also observed for desorption of naphthalene from lignite in aqueous solutions.182 Irreversible hysteresis means that complete desorption cannot be achieved (Figure 12B). This type of hysteresis has been observed for vapor sorption in the nanopores of glassy polymers183 and the interlayers of dry clay184 and for aqueous sorption of organic compounds by soils and sediments178,185 and charcoal.179 Irreversible pore deformation has been proposed to interpret the irreversible hysteresis in these studies.

10.2. Desorption Hysteresis on Fullerene
Irreversible hysteresis was observed for the desorption of PAHs (i.e., naphthalene, phenanthrene and pyrene), and 1,2dichlorobenzene from fullerene into aqueous solutions,6,62,114 while no desorption hysteresis was observed for PAHs from CNTs into aqueous solutions.6 This difference was attributed to the distinct geometries of fullerene and CNTs.6 Spherical monomers of fullerene may result in its aggregates having closed interstitial spaces in small aggregates or between small aggregates. Deformation and rearrangement of small aggregates can result in the penetration of adsorbate molecule into closed interstitial spaces between small aggregates during adsorption and lead to the molecular entrapment and hysteresis. The cylindrical monomers of CNTs cannot form closed interstitial spaces in their aggregates due to their length, hence no adsorption-desorption hysteresis was observed for PAHs.

Figure 12. Adsorption-desorption isotherm shows (A) reversible hysteresis with closed hysteresis loop and (B) irreversible hysteresis with opened hysteresis loop.

Polanyi Theory and Its Application

Chemical Reviews, 2010, Vol. 110, No. 10 6005

10.3. Desorption Hysteresis and Immobilization on CNTs
Although no desorption hysteresis was observed for PAHs from CNTs,6 immobilization and irreversible hysteresis have been observed for other organic chemicals such as tetra-tertbutylphthalocyanines87 and bisphenol A (BPA) and 17Rethinyl estradiol (EE2)53 upon their desorption from CNTs into water. The strong π-π coupling of benzene-ringcontaining chemicals with the CNT surface was proposed by Wang et al.87 to interpret the immobilization. However, the absence of adsorption-desorption hysteresis for naphthalene, phenanthrene, and pyrene6 indicated that this π-π coupling could not result in the immobilization. Pan et al.53 proposed that the rearrangement of the bundles or aggregates of CNTs is responsible for the irreversible hysteresis of two estrogens (i.e., EE2 and BPA). Pan et al.53 also attributed the absence of adsorption-desorption hysteresis for PAHs reported by Yang and Xing6 possibly to the lower rearrangement of the bundles or aggregates of CNTs because of relatively less and weaker adsorption of PAHs as compared with that of two estrogens. This explanation may not be applicable for the signi?cant irreversible hysteresis of the desorption of atrazine on highly oxidized CNTs (with an oxygen content of 7.07%)156 because (i) the adsorption of atrazine on the highly oxidized CNTs is less than that on other CNTs with less oxygen content (<2.16%) and (ii) no desorption hysteresis of atrazine on CNTs with lower oxygen content was observed.52,156 The observed irreversible hysteresis of atrazine on highly oxidized CNTs rather than other lower oxygen-containing CNTs could lead us to conclude the important role of oxygen-containing functional groups on CNT surfaces in irreversible hysteresis. Moreover, tetratert-butylphthalocyanines, EE2, BPA, and atrazine have oxygen- or nitrogen-containing functional groups in their molecules as compared with the molecular structures of PAHs. These functional groups in the molecules may possibly form covalent bonds with the oxygen-containing functional groups such as -COOH and -OH of CNTs.78,100,101 Covalent modi?cation of CNTs has been widely utilized to form a variety of nanostructures.28,102,103 Therefore, we proposed here that the formation of covalent bonds between the functional groups on organic chemicals and CNTs may be partly responsible for the observed immobilization of tetratert-butylphthalocyanines87 and irreversible hysteresis of BPA, EE2, and atrazine.53,156 Consequently, the absence of desorption hysteresis for PAHs6 and atrazine52,156 from CNTs can be also attributed to (i) no covalent bonds can form between PAH molecules and CNTs because of the absence of functional groups in the PAH molecules and (ii) few covalent bonds can form between atrazine and CNTs because of these CNTs containing few functional groups. For organic chemicals such as PAHs having no desorption hysteresis on CNTs, their desorption from CNTs into water can be estimated from their adsorption isotherms using Polanyi theory-based DA model. Although Polanyi theory and DA model can be employed to describe and estimate the adsorption of organic chemicals on fullerene and CNTs, they seem to be inadequate to describe and estimate the irreversible hysteresis of desorption of organic chemicals because the irreversible hysteresis is a result of the deformation and rearrangement of CNPs aggregates (i.e., different thermodynamic pathways between adsorption and desorption). However, irreversible hysteresis of desorption of organic chemicals from CNTs, resulting from covalent bond

formation, could be described and estimated by Polanyi theory and DA model if there are parameters to describe the ability of functional groups (in organic chemicals and on CNP surfaces) to form covalent bonds.

11. Conclusions and Perspectives
The available spaces of CNPs and the strength of interactions between organic chemicals and CNPs determine the adsorption capacity and af?nity of organic chemicals on CNPs, respectively. External surface of CNP monomers and the interstitial spaces and grooves surrounded by the external surface of CNP monomers are the possible spaces of CNPs for organic chemical adsorption. The availability of these possible spaces for adsorption can be altered by the surface functional groups, surface modifying chemicals, surface curvature, surface defects, pore sizes in aggregates, and pore deformation and rearrangement of CNPs. The availability of these possible spaces is also dependent on the molecular structures, functional groups, and molecular sizes of organic chemicals. There are ?ve interactions (i.e., hydrophobic, π-π bonding, hydrogen bonding, covalent bonding, and electrostatic interactions) between organic chemicals and CNPs. These interactions and their strength are in?uenced by the structures and functional groups of chemicals and the functional groups on CNP surfaces. Hydrophobic graphitic surfaces are the main surfaces of CNPs responsible for organic chemical adsorption. Functional group (e.g., -COOH and -OH) conglutinated hydrophilic graphitic surfaces of CNPs are not available for most organic chemicals because of the suppression of water, and thus functional groups on CNP surfaces decrease the adsorption capacity of most organic chemicals except for chemical molecules having functional groups that are able to form covalent bonds with the functional groups of CNPs. Surface modi?cation of CNPs with “hard” organic chemicals such as PAHs containing only aromatic rings in their molecules may decrease signi?cantly the adsorption af?nity of other organic chemicals but not their adsorption capacity. However, surface modi?cation of CNPs with “soft” organic chemicals such as NOM and surfactants containing long aliphatic chains in their molecules could not only decrease the adsorption af?nity but also the adsorption capacity of other organic chemicals. Adsorptiondesorption hysteresis of organic chemicals on fullerene is irreversible and can be attributed to the closed interstitial spaces between small fullerene aggregates and the penetration and entrapment of chemical molecules into these closed interstitial spaces because of the deformation and rearrangement of small aggregates, while the irreversible hysteresis on CNTs can be possibly attributed to the formation of covalent bonds between the functional groups of both of organic chemicals and CNTs. For the application of CNPs as potential superior adsorbents, it is necessary to disperse the CNP aggregates to expose their surface area for more adsorption of organic chemicals because the adsorption is dependent on their surface area and the aggregation of CNPs always decreases their surface area. However, the common methods employed to promote the CNP dispersion such as surface functionalization with oxygen containing groups and surface modi?cation with DOM and surfactants could not be a viable way to enhance the CNP surface area for more adsorption of organic chemicals because these methods are accompanied by the competition of water and adsorbed DOM (or surfactants) with organic chemicals and thus could result in the decrease in

6006 Chemical Reviews, 2010, Vol. 110, No. 10

Yang and Xing

chemical adsorption.54,91,97,98,129,159,160 Therefore, new methods with selecting of dispersion reagents having no competition with organic chemicals on CNP surface need to be developed to expose CNP surface area for the application of CNPs as potential superior adsorbents. Moreover, dispersion of CNPs will result in the dif?culties in separating CNPs from aqueous phase, which should be solved before they are used as superior adsorbents. In addition, dispersed CNPs have been recognized to transport readily and farther with water in the environment than the undispersed ones,29-31,33,34,186,187 which may result in higher environmental and health risks as they released into the environment. Polanyi theory captures the aqueous adsorption process of organic chemicals by CNPs mechanistically, as evidenced by both of the observed characteristic curve of Polanyi theory and the excellent ?tting Polanyi theory-based sorption models such as DA model.23 DA model has excellent ?tting for isotherms of organic chemicals on CNPs although these chemicals and CNPs have various structures and physicochemical properties.6,23,45,51,53 The good ?tting of DA model has also been observed even if other chemicals including “hard” aromatic molecules (e.g., PAHs), “soft” aliphatic molecules (e.g., NOM and surfactants), and metal ions are added as competitors.50,54,129,156,160 Available adsorption spaces of CNPs for organic chemicals, varied with the CNP/ organic chemical structures and physicochemical properties and the CNP surface modi?cation with other chemicals, can be identi?ed by a capacity parameter (Q0) of DA model. Both E and b of the DA model are af?nity parameters to identify the interaction forces and their strength as well as the energy distribution of CNP surface for organic chemical adsorption. Empirical relationships have been developed for Q0 with CNP surface area,23,53 Q0 with CNP surface oxygen content,45 Q0 with adsorbed surfactant amounts on CNP surfaces,129 E (or b) with chemical solvatochromic parameters,51 E (or b) with the Rmeso/micro ratios of CNPs,23 E (or b) with adsorbed surfactant amounts on CNP surfaces,129 and even for the pHdependence of adsorption of IOCs, which can be estimated from the solution pH and chemical pKa values.51 These relationships may not only help us to understand the aqueous adsorption mechanisms of organic chemicals on CNPs but also represent a favorable step by estimating their adsorption capacity and af?nity to evaluate the potential application of CNPs as environmental sorbents and the environmental risks of both toxic organic chemicals and CNPs after they are released into the environment. Limited by the cost and experimental dif?culty of testing environmental behavior such as sorption for thousands of chemicals and materials, such estimation is usually essential when direct experimental data are unavailable. Therefore, efforts are still needed to combine these developed empirical relationships into a predictive model and develop more valid empirical relationships such as Q0 with chemical structures and physicochemical properties. In addition, the solubility data of some organic chemicals especially for the dissociated species of IOCs is not yet available at the present, which can be a main limitation for the application of Polanyi theory-based DA model in describing adsorption of organic chemicals relative to other models such as Freundlich model. Although there are still unanswered questions regarding the adsorption of organic chemicals on CNPs and the related environmental risks of both toxic organic chemicals and CNPs, the reported results summarized in this review have shown that Polanyi theory and the developed empirical

relationships (models) between the parameters of the Polanyi theory-based DA model and the CNP/organic chemical structures and physicochemical properties can help us to better understand the adsorption mechanisms of organic chemicals on CNPs and to estimate the adsorption capacity and af?nity of organic chemicals. In addition, appropriate Polanyi theory-based predictive models may help to design CNPs with speci?c needs, and to evaluate the potential application of CNPs as sorbents and the environmental risks of both toxic organic chemicals and CNPs. Therefore, this review may stimulate more detailed and systematic investigations for (i) effectively applying the Polanyi theory and DA model in describing the adsorption of organic chemicals on CNPs, (ii) searching for and measuring the values of useful physicochemical parameters of both organic chemicals and CNPs to develop predictive models to estimate the ?tting parameters of DA model, and (iii) using the Polanyi theory and DA model to determine the desorption, hysteresis, and bioavailability of adsorbed organic chemicals on CNPs.

12. Acknowledgments
This work was supported by the Program for New Century Excellent Talents in Universities of China (NCET-08-493), NSF of China (40973065, 20777065 and 20737002), the USDA Hatch Program (MAS00978), and NSF (CMMI0531171).

13. References
(1) Kroto, H. W.; Heath, J. R.; O’Brien, S. C.; Curl, R. F.; Smalley, R. E. Nature 1985, 318, 162. (2) Iijima, S. Nature 1991, 354, 56–58. (3) Iijima, S.; Ichihashi, T. Nature 1993, 363, 603. (4) Colvin, V. L. Nat. Biotechnol. 2003, 21, 1160. (5) Mauter, M. S.; Elimelech, M. EnViron. Sci. Technol. 2008, 42, 5843. (6) Yang, K.; Xing, B. S. EnViron. Pollut. 2007, 145, 529. (7) Yin, Y. F.; Mays, T.; McEnaney, B. Langmuir 1999, 15, 8714. (8) NanoMarkets Carbon nanotube electronics: a technology analysis and market forecast; NanoMarkets, 2005. (9) Kohler, A. R.; Som, C.; Helland, A.; Gottschalk, F. J. Clean. Prod. 2008, 16, 927. (10) Nowack, B.; Bucheli, T. D. EnViron. Pollut. 2007, 150, 5. (11) Schmid, K.; Riediker, M. EnViron. Sci. Technol. 2008, 42, 2253. (12) Maynard, A. D.; Baron, P. A.; Foley, M.; Shvedova, A. A.; Kisin, E. R.; Castranova, V. J. Toxicol. EnViron. Health, Part A 2004, 67, 87. (13) Mueller, N. C.; Nowack, B. EnViron. Sci. Technol. 2008, 42, 4447. (14) Helland, A.; Wick, P.; Koehler, A.; Schmid, K.; Som, C. EnViron. Health Perspect. 2007, 15, 1125. (15) Lam, C.-W.; James, J. T.; McCluskey, R.; Arepalli, S.; Hunter, R. L. Crit. ReV. Toxicol. 2006, 36, 189. (16) Lam, C.-W.; James, J. T.; McCluskey, R.; Hunter, R. L. Toxicol. Sci. 2004, 77, 126. (17) Wu, Y.; Hudson, J.; Lu, Q.; Moore, J.; Mount, A.; Rao, A. M.; Alexov, E.; Ke, P. C. J. Phys. Chem. B 2006, 110, 2475. (18) Dunphy Guzman, K. A.; Taylor, M. R.; Ban?eld, J. F. EnViron. Sci. Technol. 2006, 40, 1401. (19) Oberdorster, E.; McClellan-Green, P.; Haasch, M. L. In NanomaterialssToxicity, Health and EnVironmental Issues, Nanotechnologies for the Life Sciences, Vol. 5; Kumar, C., Ed.;WileyVCH: New York, 2006. (20) Petersen, E. J., Jr.; Huang, Q. G.; Weber, W. J. EnViron. Health Perspect. 2008, 116, 496. (21) Petersen, E. J., Jr.; Huang, Q. G.; Weber, W. J. EnViron. Sci. Technol. 2008, 42, 3090. (22) Long, R. Q.; Yang, R. T. J. Am. Chem. Soc. 2001, 123, 2058. (23) Yang, K.; Zhu, L. Z.; Xing, B. S. EnViron. Sci. Technol. 2006, 40, 1855. (24) Liu, X. Y.; Ji, Y. S.; Zhang, Y. H.; Zhang, H. X.; Liu, M. J. Chromatogr., A 2007, 1165, 10. (25) Niu, H. Y.; Cai, Y. Q.; Shi, Y. L.; Wei, F. S.; Liu, J. M.; Mou, S. F.; Jiang, G. B. Anal. Chim. Acta 2007, 594, 81. (26) Pyrzynska, K. Sep. Purif. ReV. 2008, 37, 374. (27) Kaur, A.; Gupta, U. J. Mater. Chem. 2009, 19, 8279.

Polanyi Theory and Its Application
(28) Riggs, J. E.; Guo, Z.; Carroll, D. L.; Sun, Y. P. J. Am. Chem. Soc. 2000, 122, 5879. (29) Lecoanet, H. F.; Botterro, J.-Y.; Wiesner, M. R. EnViron. Sci. Technol. 2004, 38, 5164. (30) Hyung, H.; Fortner, J. D.; Hughes, J. B.; Kim, J. H. EnViron. Sci. Technol. 2007, 41, 179. (31) Hyung, H.; Kim, J. H. EnViron. Sci. Technol. 2008, 42, 4416. (32) Yang, K.; Xing, B. S. EnViron. Pollut. 2009, 157, 1095. (33) Chappell, M. A.; George, A. J.; Dontsova, K. M.; Porter, B. E.; Price, C. L.; Zhou, P. H.; Morikawa, E.; Kennedy, A. J.; Steevens, J. A. EnViron. Pollut. 2009, 157, 1081. (34) Lin, D. H.; Xing, B. S. EnViron. Sci. Technol. 2008, 42, 5917. (35) Oberdorster, G.; Oberdorster, E.; Oberdorster, J. EnViron. Health Perspect. 2005, 113, 823. (36) Li, N.; Sioutas, C.; Cho, A.; Schmitz, D.; Misra, C.; Sempf, J.; Wang, M. Y.; Oberley, T.; Froines, J.; Nel, A. EnViron. Health Perspect. 2003, 111, 455. (37) Sander, M.; Pignatello, J. J. EnViron. Sci. Technol. 2005, 39, 1606. (38) Bucheli, T. D.; Gustafsson, O. EnViron. Sci. Technol. 2000, 34, 5144. (39) Cornelissen, G.; Gustafsson, O.; Bucheli, T. D.; Jonker, M. T. O.; Koelmans, A. A.; Van Noort, P. C. M. EnViron. Sci. Technol. 2005, 39, 6881. (40) Ferguson, P. L.; Chandler, G. T.; Templeton, R. C.; Demarco, A.; Scrivens, W. A.; Englehart, B. A. EnViron. Sci. Technol. 2008, 42, 3879. (41) Sontheimer, H.; Crittenden, J. C.; Summers, R. S. ActiVated Carbon for Water Treatment; DVGW-Forschungsstelle: Karlsruhe, Germany, 1988. (42) Ago, H.; Kugler, T.; Cacialli, F.; Salaneck, W. R.; Shaffer, M. S. P.; Windle, A. H.; Friend, R. H. J. Phys. Chem. B 1999, 103, 8116. (43) Gotovac, S.; Hattori, Y.; Noguchi, D.; Miyamoto, J.; Kanamaru, M.; Utsumi, S.; Kanoh, H.; Kaneko, K. J. Phys. Chem. B 2006, 110, 16219. (44) Gotovac, S.; Yang, C. M.; Hattori, Y.; Takahashi, K.; Kanoh, H.; Kaneko, K. J. Colloid Interface Sci. 2007, 314, 18. (45) Cho, H.-H.; Smith, B. A.; Wnuk, J. D.; Fairbrother, H.; Ball, W. P. EnViron. Sci. Technol. 2008, 42, 2899. (46) Polanyi, M. Verh. Dtsch. Phys. Ges. 1916, 18, 55. (47) Brunauer, S. Adsorption of Gases and Vapors; Princeton University Press: Princeton, NJ, 1945. (48) Manes, M. In: Encyclopedia of EnVironmental Analysis and Remediation; Meyers, R. A., Eds.; John Wiley & Sons: New York, 1998; p 26. (49) Crittenden, J. C.; Sanongraj, S.; Bulloch, J. L.; Hand, D. W.; Rogers, T. N.; Speth, T. F.; Ulmer, M. EnViron. Sci. Technol. 1999, 33, 2926. (50) Yang, K.; Wang, X. L.; Zhu, L. Z.; Xing, B. S. EnViron. Sci. Technol. 2006, 40, 5804. (51) Yang, K.; Wu, W. H.; Jing, Q. F.; Zhu, L. Z. EnViron. Sci. Technol. 2008, 42, 7931. (52) Yan, X. M.; Shi, B. Y.; Lu, J. J.; Feng, C. H.; Wang, D. S.; Tang, H. X. J. Colloid Interface Sci. 2008, 321, 30. (53) Pan, B.; Lin, D. H.; Mashayekhi, H.; Xing, B. S. EnViron. Sci. Technol. 2008, 42, 5480. (54) Wang, X. L.; Lu, J. L.; Xing, B. S. EnViron. Sci. Technol. 2008, 42, 3207. (55) Wang, Z. Y.; Yu, X. D.; Pan, B.; Xing, B. S. EnViron. Sci. Technol. 2010, 44, 978. (56) Pan, B.; Xing, B. S. EnViron. Sci. Technol. 2008, 42, 9005. (57) Polanyi, M. Verh. Dtsch. Phys. Ges. 1914, 16, 1012. (58) Maddox, M. W.; Gubbins, K. E. Langmuir 1995, 11, 3988. (59) Onyestyak, G.; Otvos, Z.; Valyon, J.; Kiricsi, I.; Rees, L. V. C. HelV. Chim. Acta 2004, 87, 1508. (60) Gotovac, S.; Song, L.; Kanoh, H.; Kaneko, K. Colloid Surf., A 2007, 300, 117. (61) Datsyuka, V.; Kalyva, M.; Papagelis, K.; Parthenios, J.; Tasisb, D.; Siokoua, A.; Kallitsisa, I.; Galiotisa, C. Carbon 2008, 46, 833. (62) Cheng, X. K.; Kan, A. T.; Tomson, M. B. J. Chem. Eng. Data 2004, 49, 675. (63) Fortner, J. D.; Lyon, D. Y.; Sayes, C. M.; Boyd, A. M.; Falkner, J. C.; Hotze, E. M.; Alemany, L. B.; Tao, Y. J.; Guo, W.; Ausman, K. D.; Colvin, V. L.; Hughes, J. B. EnViron. Sci. Technol. 2005, 39, 4307. (64) Murr, L. E.; Soto, K. F. Mater. Charact. 2005, 55, 50. (65) Saleh, N. B.; Pfefferle, L. D.; Elimelech, M. EnViron. Sci. Technol. 2008, 42, 7963. (66) Lin, D. H.; Liu, N.; Yang, K.; Zhu, L. Z.; Xu, Y.; Xing, B. S. Carbon 2009, 47, 2875. (67) Smith, B.; Wepasnick, K.; Schrote, K. E.; Bertele, A. H.; Ball, W. P.; O’Melia, C.; Fairbrother, D. H. EnViron. Sci. Technol. 2009, 43, 813. (68) Smith, B.; Wepasnick, K.; Schrote, K. E.; Cho, H. H.; Ball, W. P.; Fairbrother, D. H. Langmuir 2009, 25, 9767. (69) Chen, J.; Hamon, M. A.; Hu, H.; Chen, Y. S.; Rao, A. M.; Eklund, P. C.; Haddon, R. C. Science 1998, 282, 95.

Chemical Reviews, 2010, Vol. 110, No. 10 6007
(70) Georgakilas, V.; Bourlinos, A.; Gournis, D.; Tsou?s, T.; Trapalis, C.; Mateo-Alonso, A.; Prato, M. J. Am. Chem. Soc. 2008, 130, 8733. (71) Schierz, A.; Zanker, H. EnViron. Pollut. 2009, 157, 1088. (72) Islam, M. F.; Rojas, E.; Bergey, D. M.; Johnson, A. T.; Yodh, A. G. Nano Lett. 2003, 3, 269. (73) Matarredona, O.; Rhoads, H.; Li, Z. R.; Harwell, J. H.; Balzano, L.; Daniel, E.; Resasco, D. E. J. Phys. Chem. B 2003, 107, 13357. (74) Richard, C.; Balavoine, F.; Schultz, P.; Ebbesen, T. W.; Mioskowski, C. Science 2003, 300, 775. (75) Ham, H. T.; Choi, Y. S.; Chung, I. J. J. Colloid Interface Sci. 2005, 286, 216. (76) Vaisman, L.; Wagner, H. D.; Marom, G. AdV. Colloid Interf. Sci. 2006, 128, 37. (77) Adamson, A. W. Physical Chemistry of Surface; 5th ed.; John Wiley & Sons: New York, 1990; Chapter XV. (78) Ke, G.; Guan, W. C.; Tang, C. Y.; Hu, Z.; Guan, W. J.; Zeng, D. L.; Deng, F. Chin. Chem. Lett. 2007, 18, 361. (79) Bittner, E. W.; Smith, M. R.; Bockrath, B. C. Carbon 2003, 41, 1231. (80) Balavoine, F.; Schultz, P.; Richard, C.; Mallouh, V.; Ebbesen, T. W.; Mioskowski, C. Angew. Chem., Int. Ed. 1999, 38, 1912. (81) Chen, R. J.; Bangsaruntip, S.; Drouvalakis, K. A.; Kam, N. W. S.; Shim, M.; Li, Y. M.; Kim, W.; Utz, P. J.; Dai, H. J. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 4984. (82) Pyrzynska, K.; Sta?ej, A.; Biesaga, M. Microchim. Acta 2007, 159, 293. (83) Chen, W.; Duan, L.; Zhu, D. Q. EnViron. Sci. Technol. 2007, 41, 8295. (84) Star, A.; Stoddart, J. F.; Steuerman, D.; Diehl, M.; Boukai, A.; Wong, E. W.; Yang, X.; Chung, S. W.; Choi, H.; Heath, J. R. Angew. Chem., Int. Ed. 2001, 40, 1721. (85) Sumanasekera, G. U.; Pradhan, B. K.; Adu, C. K. W.; Romero, H. E.; Foley, H. C.; Eklund, P. C. Mol. Cryst. Liq. Cryst. 2002, 387, 255. (86) Sumanasekera, G. U.; Pradhan, B. K.; Romero, H. E.; Adu, K. W.; Eklund, P. C. Phys. ReV. Lett. 2002, 89, 166801. (87) Wang, X. B.; Liu, Y. Q.; Qiu, W. F.; Zhu, D. B. J. Mater. Chem. 2002, 12, 1636. (88) Zhao, J. J.; Lu, J. P.; Han, J.; Yang, C. K. Appl. Phys. Lett. 2003, 82, 3746. (89) Woods, L. M.; Badescu, S. C.; Reinecke, T. L. Phys. ReV. B 2007, 75, 155415. (90) Chen, W.; Duan, L.; Wang, L. L.; Zhu, D. Q. EnViron. Sci. Technol. 2008, 42, 6862. (91) Chen, J. Y.; Chen, W.; Zhu, D. Q. EnViron. Sci. Technol. 2008, 42, 7225. (92) Chen, J.; Liu, H. Y.; Weimer, W. A.; Halls, M. D.; Waldeck, D. H.; Walker, G. C. J. Am. Chem. Soc. 2002, 124, 9034. (93) Gotovac, S.; Honda, H.; Hattori, Y.; Takahashi, K.; Kanoh, H.; Kaneko, K. Nano Lett. 2007, 7, 583. (94) Tournus, F.; Charlier, J. C. Phys. ReV. B 2005, 71, 165421. (95) Vermisoglou, E. C.; Georgakilas, V.; Kouvelos, E.; Pilatos, G.; Viras, K.; Romanos, G.; Kanellopoulos, N. K. Microporous Mesoporous Mater. 2007, 99, 98. (96) Franz, M.; Arafat, H. A.; Pinto, N. G. Carbon 2000, 38, 1807. (97) Ahnert, F.; Arafat, H. A.; Pinto, N. G. Adsorption 2003, 9, 311. (98) Ania, C. O.; Parra, J. B.; Pis, J. J. Fuel Sci. Technol. 2002, 77, 337. (99) Muller, E. A.; Rull, L. F.; Vega, L. F.; Gubbins, K. E. J. Phys. Chem. 1996, 100, 1189. (100) Huang, W. J.; Taylor, S.; Fu, K. F.; Lin, Y.; Zhang, D. H.; Hanks, T. W.; Rao, A. M.; Sun, Y. P. Nano Lett. 2002, 2, 311. (101) Piao, L. Y.; Liu, Q. R.; Li, Y. D.; Wang, C. J. Phys. Chem. C 2008, 112, 2857. (102) Banerjee, S.; Wong, S. S. Nano Lett. 2002, 2, 195. (103) Hussain, C. M.; Saridara, C.; Mitra, S. Analyst 2009, 134, 1928. (104) Wang, Y. B.; Iqbal, Z.; Malhotra, S. V. Chem. Phys. Lett. 2005, 402, 96. (105) Chen, Y. H.; Mitra, S. J. Nanosci. Nanotechnol. 2008, 8, 5770. (106) Bahr, J. L.; Tour, J. M. Chem. Mater. 2001, 13, 3823. (107) Mickelson, E. T.; Huffman, C. B.; Rinzler, A. G.; Smalley, R. E.; Hauge, R. H.; Margrave, J. L. Chem. Phys. Lett. 1998, 296, 188. (108) Ying, Y. M.; Saini, R. K.; Liang, F.; Sadana, K.; Billups, W. E. Org. Lett. 2003, 5, 1471. (109) Coleman, K. S.; Bailey, S. R.; Fogden, S.; Green, M. L. H. J. Am. Chem. Soc. 2003, 125, 8722. (110) Kim, W. J.; Kang, S. O.; Ah, C. S.; Lee, Y. W.; Ha, D. H.; Choi, I. S.; Yun, W. S. Bull. Korean Chem. Soc. 2004, 25, 1301. (111) Shen, J. F.; Huang, W. S.; Wu, L. P.; Hu, Y. Z.; Ye, M. X. Compos., Part A: Appl. S. 2007, 38, 1331. (112) Daniel, S.; Rao, T. P.; Rao, K. S.; Rani, S. U.; Naidu, G. R. K.; Lee, H. Y.; Kawai, T. Sens. Actuators B 2007, 122, 672. (113) Lin, D. H.; Xing, B. S. EnViron. Sci. Technol. 2008, 42, 7254. (114) Cheng, X. K.; Kan, A. T.; Tomson, M. B. J. Nanopart. Res. 2005, 7, 555. (115) Liao, Q.; Sun, J.; Gao, L. Carbon 2008, 46, 553.

6008 Chemical Reviews, 2010, Vol. 110, No. 10
(116) Liao, Q.; Sun, J.; Gao, L. Colloid Surf. A 2008, 312, 160. (117) Chin, C-J. M.; Shih, L.-C.; Tsai, H.-J.; Liu, T.-K. Carbon 2007, 45, 1254. (118) Wu, C. H. J. Hazard. Mater. 2007, 144, 93. (119) Lu, C. S.; Chung, Y.-L.; Chang, K.-F. Water Res. 2005, 39, 1183. (120) Lu, C. S.; Chung, Y. L.; Chang, K. F. J. Hazard. Mater. 2006, 138, 304. (121) Peng, X. J.; Li, Y. H.; Luan, Z. K.; Di, Z. C.; Wang, H. Y.; Tian, B. H.; Jia, Z. P. Chem. Phys. Lett. 2003, 376, 154. (122) Condon, J. B. Microporous Mesoporous Mater. 2000, 38, 377. (123) Dubinin, M. M. Chem. ReV. 1960, 60, 235. (124) Liu, C. H.; Li, J. J.; Zhang, H. L.; Li, B. R.; Guo, Y. Colloid Surf. A 2008, 313, 9. (125) Zhang, S. J.; Shao, T.; Bekaroglu, S. S. K.; Karan?l, T. EnViron. Sci. Technol. 2009, 43, 5719. (126) Su, F. S.; Lu, C. S. J. EnViron. Sci. Health, Part A: Toxic/Hazard. Subst. EnViron. Eng. 2007, 42, 1543. (127) Fagan, S. B.; Souza, A. G.; Lima, J. O. G.; Mendes, J.; Ferreira, O. P.; Mazali, I. O.; Alves, O. L.; Dresselhaus, M. S. Nano Lett. 2004, 4, 1285. (128) Yang, K.; Zhu, L. Z.; Lou, B. F.; Chen, B. L. Chemosphere 2005, 61, 116. (129) Yang, K.; Jing, Q. F.; Wu, W. H.; Zhu, L. Z.; Xing, B. S. EnViron. Sci. Technol. 2010, 44, 681. (130) Mao, Z. G.; Sinnott, S. B. J. Phys. Chem. B 2001, 105, 6916. (131) Allen-King, R. M.; Grathwohl, P.; Ball, W. P. AdV. Water Resour. 2002, 25, 985. (132) Karajanagi, S. S.; Vertegel, A. A.; Kane, R. S.; Dordick, J. S. Langmuir 2004, 20, 11594. (133) Wang, Z. Y.; Zhang, K.; Zhao, J.; Liu, X. Y.; Xing, B. S. Chemosphere 2010, 79, 86. (134) Yang, K.; Xing, B. S. EnViron. Sci. Technol. 2009, 43, 1845. (135) Yang, K.; Zhu, L. Z.; Xing, B. S. EnViron. Sci. Pollut. Res. 2010, 17, 410. (136) Kamlet, M. J.; Doherty, R. M.; Abraham, M. H.; Marcus, Y.; Taft, R. W. J. Phys. Chem. 1988, 92, 5244. (137) Marcus, Y. J. Phys. Chem. 1991, 95, 8886. (138) Hickey, J. P.; Passino-Reader, D. R. EnViron. Sci. Technol. 1991, 25, 1753. (139) Nirmalakhandan, N.; Egemen, E.; Trevizo, C.; Xu, S. Ecotoxicol. EnViron. Saf. 1998, 39, 112. (140) Marcus, Y. J. Phys. Chem. B 1997, 101, 8617. (141) Biesaga, M.; Pyrzynska, K. J. Sep. Sci. 2006, 29, 2241. (142) Xie, X.; Gao, L.; Sun, J. Colloids Surf. A 2007, 308, 54. (143) Wang, S. G.; Liu, X. W.; Gong, W. X.; Nie, W.; Gao, B. Y.; Yue, Q. Y. J. Chem. Technol. Biotechnol. 2007, 82, 698. (144) Lu, C. Y.; Su, F. S. Sep. Purif. Technol. 2007, 58, 113. (145) Huang, G.-L.; Xiao, H.; Chi, J.; Shiu, W.-Y.; Mackay, D. J. Chem. Eng. Data 2000, 45, 411. (146) Jiang, J.; Pang, S. Y.; Ma, J. EnViron. Sci. Technol. 2009, 43, 3398. (147) Brackman, W.; Havinga, E. Recl. TraV. Chim. Pays-Bas 1955, 74, 1021. (148) Denisov, E. T. In ComprehensiVe Chemical Kinetics, Vol. 16: LiquidPhase Oxidation; Bamford, C. H., Tipper, C. F. H., Eds.; Elsevier Scienti?c Publishing Co.: Amsterdam, 1980; p181. (149) Karthikeyan, K. G.; Chorover, J.; Bortiatynski, J. M.; Hatcher, P. G. EnViron. Sci. Technol. 1999, 33, 4009. (150) Haghseresht, F.; Nouri, S.; Lu, G. Q. In Sustainable Energy and EnVironmental Technologies; Hu, X. J., Yue, P. L., Eds.; World Scienti?c Publishing Co.: Singapore, 2000; p 362. (151) Radovic, L. R.; Silva, I. F.; Ume, J. I.; Menendez, J. A.; Leon, C. A. L. Y.; Scaroni, A. W. Carbon 1997, 35, 1339.

Yang and Xing
(152) Crittenden, J.; Luft, P.; Hand, D.; Oravitz, J.; Loper, S.; Ari, M. EnViron. Sci. Technol. 1985, 19, 1037. (153) McGinley, P.; Katz, L.; Weber, W. EnViron. Sci. Technol. 1993, 27, 1524. (154) Xing, B. S.; Pignatello, J. J.; Gigliotti, B. EnViron. Sci. Technol. 1996, 30, 2432. (155) Xia, G.; Ball, W. P. EnViron. Sci. Technol. 2000, 34, 1246. (156) Chen, G. C.; Shan, X. Q.; Wang, Y. S.; Pei, Z. G.; Shen, X. E.; Wen, B.; Owens, G. EnViron. Sci. Technol. 2008, 42, 8297. (157) Chen, G. C.; Shan, X. Q.; Wang, Y. S.; Wen, B.; Pei, Z. G.; Xie, Y. N.; Liu, T.; Pignatello, J. J. Water Res. 2009, 43, 2409. (158) Rao, G. P.; Lu, C.; Su, F. Sep. Purif. Technol. 2007, 58, 224. (159) Hu, X. L.; Liu, J. F.; Mayer, P.; Jiang, G. B. EnViron. Toxicol. Chem. 2008, 27, 1868. (160) Wang, X. L.; Tao, S.; Xing, B. S. EnViron. Sci. Technol. 2009, 43, 6214. (161) Zhu, L. Z.; Tian, S. L.; Shi, Y. Clays Clay Miner. 2005, 53, 123. (162) Zhu, R. L.; Zhu, L. Z. J. Colloid Interface Sci. 2008, 322, 27. (163) Chiou, C. T.; Lee, J. F.; Boyd, S. A. EnViron. Sci. Technol. 1990, 24, 1164. (164) Boyd, S. A.; Lee, J.-F.; Mortland, M. M. Nature 1988, 333, 345. (165) Boyd, S. A.; Mortland, M. M.; Chiou, C. T. Soil Sci. Soc. Am. J. 1988, 52, 652. (166) Backhaus, W. K.; Kunmpp, E.; Narres, H.-D.; Schwuger, M. J. J. Colloid Interface Sci. 2001, 242, 6. (167) Kang, S.; Xing, B. S. Langmuir 2008, 24, 2525. (168) Chiou, C. T.; Malcolm, R. L.; Brinton, T. I.; Kile, D. E. EnViron. Sci. Technol. 1986, 20, 502. (169) Chiou, C. T.; Kile, D. E.; Brinton, T. I.; Malcolm, R. L.; Leenheer, J. A.; Maccarthy, P. EnViron. Sci. Technol. 1987, 21, 1231. (170) Maxin, C. R.; KogelKnabner, I. Eur. J. Soil Sci. 1995, 46, 193. (171) Burkhard, L. P. EnViron. Sci. Technol. 2000, 34, 4663. (172) Kile, D. E.; Chiou, C. T. EnViron. Sci. Technol. 1989, 23, 832. (173) Edwards, D. A.; Luthy, R. G.; Liu, Z. EnViron. Sci. Technol. 1991, 25, 127. (174) Jafvert, C. T.; Van Hoof, P. L.; Heath, J. K. Water Res. 1994, 28, 1009. (175) Monticone, V.; Treiner, C. Colloid. Surf. A 1995, 104, 285. (176) Gong, X. Y.; Liu, J.; Baskaran, S.; Voise, R. D.; Young, J. S. Chem. Mater. 2000, 12, 1049. (177) Jeevananda, T.; Siddaramaiah, K.; Kim, N. H.; Heo, S. B.; Lee, J.H. Polym. AdV. Technol. 2008, 19, 1754. (178) Kan, A. T.; Fu, G.; Hunter, M.; Chen, W.; Ward, C. H.; Tomson, M. B. EnViron. Sci. Technol. 1998, 32, 892. (179) Braida, W. J.; Pignatello, J. J.; Lu, Y. F.; Ravikovitch, P. I.; Neimark, A. V.; Xing, B. S. EnViron. Sci. Technol. 2003, 37, 409. (180) Gunasekara, A. S.; Xing, B. S. J. EnViron. Qual. 2003, 32, 240. (181) Rouquerol, F.; Rouquerol, J.; Sing, K. Adsorption by Powders and Porous Material; Academic Press: San Diego, CA, 1999. (182) Sander, M.; Pignatello, J. J. EnViron. Sci. Technol. 2005, 39, 7476. (183) Wang, J. S.; Kamiya, Y.; Naito, Y. J. Polym. Sci., Part B: Polym. Phys. 1998, 36, 1695. (184) Tvardovski, A. V.; Fomkin, A. A.; Tarasevich, Y. I.; Zhukova, A. I. J. Colloid Interface Sci. 2001, 241, 297. (185) Lu, Y. F.; Pignatello, J. J. EnViron. Sci. Technol. 2002, 36, 4553. (186) Fang, J.; Shan, X. Q.; Wen, B.; Lin, J. M.; Owens, G. EnViron. Pollut. 2009, 157, 1101. (187) Li, Y. S.; Wang, Y. G.; Pennell, K. D.; Abriola, L. M. EnViron. Sci. Technol. 2008, 42, 7174.

CR100059S


赞助商链接
相关文章:
jacs-NiMOF催化.pdf_图文
as the latter compounds are used widely in industry...of carbon dioxide by the incorporation of the ...TCPE1′; (c) CO2 adsorption isotherms of Ni?...
Adsorption of heavy metal onto activated carbon mod...
Adsorption of heavy metal onto activated carbon ...Adsorption experiments from aqueous solution ...of metals (Pb and Cu) were explored in a ...
金属纳米颗粒论文:金属纳米颗粒的性质研究及其应用
In particular, metal nanomaterials have excellent ...which is influenced significantly by size, shape ...aqueous phase for synthesis of gold nanoparticles....
职位描述
Organic and carbon nano material based electronic ...Students with research background in nanomaterials ...phase of a custom-built swept source OCT system...
Development of high sensitivity ELISA strategy and ...
1-4 nanomaterials, GOs own the advantages of ...carbon material, have drawn intense attention in ...in PBS to minimize nonspecific adsorption and ...
good纳米CaCO_3复合微粒的制备以及在PVC塑料中的应用
Key Lab for Nanomaterials, M inistry of Education...of Nano - CaCO3 in aqueous phase was researched...in the composites matrix was observed by TEM. Th...
张帼杰毕业论文终本
cyanometalate such as K4Fe(CN)6 in aqueous ...The palladium nanomaterials with three-dimensional (...by spontaneous deposition of platinum on carbon-...
英语论文标题
nitrate-contaminated groundwater using biodegradable snack ware as carbon sourc ...in aqueous solutions using boron doped diamond anodes 3、Adsorption of lead ...
更多相关标签: