nano tube vs organik nano
DESCRIPTION
nanoteknologiesTRANSCRIPT
ww.sciencedirect.com
wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 3 4e5 5
Available online at w
ScienceDirect
journal homepage: www.elsevier .com/locate /watres
Review
Adsorption of synthetic organic contaminants bycarbon nanotubes: A critical review
Onur Guven Apul, Tanju Karanfil*
Department of Environmental Engineering and Earth Sciences, Clemson University, 342 Computer Court, Anderson,
SC 29625, United States
a r t i c l e i n f o
Article history:
Received 27 May 2014
Received in revised form
23 September 2014
Accepted 24 September 2014
Available online 5 October 2014
Keywords:
Adsorption
Carbon nanotube
Nanoparticle
Predictive model
Sorption
Synthetic organic compound
* Corresponding author. Tel.: þ1 864 656 100E-mail address: [email protected] (T.
http://dx.doi.org/10.1016/j.watres.2014.09.0320043-1354/© 2014 Elsevier Ltd. All rights rese
a b s t r a c t
In last ten years, a large number (80þ) of articles regarding aqueous phase adsorption of a
variety of synthetic organic compound (SOC) by CNTs were published in peer-reviewed
journals. Adsorption depends upon the physicochemical properties of the adsorbates
and CNTs as well as the background water chemistry. Among all properties reported in the
literature, no parameter was reported as solely controlling SOC adsorption by CNTs. In this
article, these contributing parameters were reviewed and the associated explanations were
discussed. This comprehensive literature survey provides (i) a thorough CNT character-
ization summary, (ii) a discussion of adsorption mechanisms of SOCs by CNTs and (iii) a
summary of the statistical adsorption model development efforts. It also includes dis-
cussions of agreements and differences in the literature, and identifies some research
needs.
© 2014 Elsevier Ltd. All rights reserved.
Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2. General features of SOC adsorption by CNTs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3. Characterization of CNTs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4. Adsorption of SOCs by CNTs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.1. Influence of CNT properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.2. Influence of SOC properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.3. Influence of background solution properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
5. Predictive models for adsorption of SOCs by CNTs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
6. Comparison of carbon nanotubes with activated carbon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
7. Future research needs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
8. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
5; fax: þ1 864 656 0672.Karanfil).
rved.
wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 3 4e5 5 35
Supplementary data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
1. Introduction
Graphitic carbon is sp2 hybridized solid phase of pure carbon
where three of the four valance electrons are covalently
shared in a two-dimensional plane and the fourth valance
electron is delocalized among all atoms present as a weak p
bond in the third dimension (Ajayan, 1999; Terrones, 2003,
2004). Carbon nanotubes (CNTs) can be visualized as
graphitic carbon sheets rolled into hollow cylinders with
nanometer scale diameters and micrometer scale lengths
(Terrones, 2003, 2004; Iijima, 1991). There are two types of
CNTs: single-walled (SWCNT) and multi-walled (MWCNT)
(Ajayan, 1999). Owing to their unique properties, the produc-
tion and use of CNTs has been growing rapidly (Lam et al.,
2006). The CNT market estimates were approximately 90.5
million dollars in 2010, and global revenues are projected to
exceed 1 billion dollars by 2015 (The Global Market for Car,
2010). CNTs are now synthesized at larger scales (i.e. more
than 3000 metric tons/year), and used in many electronic,
medical, space and military applications (Klaine et al., 2008;
Mauter and Elimelech, 2008; Keller et al., 2013). In addition,
superior hydrophobicity, high specific surface area, and hol-
low and layered structures of CNTs make them also particu-
larly promising adsorbents (Upadhyayula et al., 2009; Zhang
et al., 2011). Adsorption of many compounds in water such
as various classes of synthetic organic contaminants (Gotovac
et al., 2006, 2007a, 2007b; Yang et al., 2006a; Yang et al., 2006b;
Cho et al., 2008; Wang et al., 2008; Chen et al., 2007; Lin and
Xing, 2008a; Oleszczuk et al., 2009; Ji et al., 2009a; Gupta
et al., 2013; Pyrzynska et al., 2007) (see Table S1 in
Supplementary Information), natural organic matter (NOM)
(Su and Lu, 2007; Yang and Xing, 2009; Ajmani et al., 2014;
Smith et al., 2012), and metallic ions (Li et al., 2005, 2003; Rao
et al., 2007) by CNTs has been widely reported.
From a natural system perspective, the remarkable in-
crease in production and use of CNTs raises health and envi-
ronmental concerns, particularly upon release into the
environment (Lam et al., 2006; Powell and Kanarek, 2006;
Johnston et al., 2010; Petosa et al., 2010). CNTs may enter the
environment through either intentional or unintentional re-
leases (e.g. atmospheric emissions and solid or liquid waste
streams) from production facilities, causing damage to plant
and animal life at the cellular level (Lam et al., 2006; Klaine
et al., 2008; Petosa et al., 2010). The toxicity of CNTs may
also be enhanced by the adsorbed organic contaminants in
the environment (Ferguson et al., 2008; Xia et al., 2010a; Yang
and Xing, 2007). Therefore, it is of critical importance to un-
derstand adsorption of SOCs by CNTs to adequately assess
their environmental impact. Froman engineering perspective,
understanding the adsorption of SOC by CNTs is also impor-
tant for evaluating the feasibility of using CNTs as alternative
adsorbents to conventional activated carbons in water and
wastewater treatment systems.
The main objective of this article was to critically review
the published peer-reviewed literature, and to articulate some
future research needs on the aqueous phase adsorption of
SOCs by CNTs with (i) a thorough CNT characterization sum-
mary, (ii) a discussion of adsorption mechanisms of SOCs by
CNTs and (iii) a summary of the statistical adsorption model
development efforts.
2. General features of SOC adsorption byCNTs
Adsorption of SOCs from water by activated carbon and other
porous carbon materials has been studied for several decades
(Dobbs and Cohen, 1980; Giusti et al., 1974; Moreno-Castilla,
2004; Kutics and Suzuki, 1993). On the other hand, adsorp-
tion of SOCs by CNTs has been rapidly growing (Iijima, 1991;
Yang et al., 2006b). More than 80 articles have been pub-
lished in peer-reviewed journals on SOC adsorption by CNTs
over the last ten years. These studies included the adsorption
of a multitude of organic contaminants by CNTs: polycyclic
aromatic hydrocarbons (PAHs) (Yang et al., 2006a, 2006b; Yang
and Xing, 2007; Zhang et al., 2010a), benzene derivatives
(Wang et al., 2008; Chen et al., 2007, 2008a; Peng et al., 2003),
phenolic compounds (Lin and Xing, 2008a; Shen et al., 2009;
Yang et al., 2008; Salam and Burk, 2008; Chen et al., 2009a),
pharmaceuticals (Chen et al., 2011; Ji et al., 2010a; Peng et al.,
2012; Yang et al., 2012; Zhang et al., 2010b; Fu et al., 2011; Ji
et al., 2010b; Cai and Larese-Casanova, 2014; Li et al., 2014;
Yu et al., 2014), polychlorinated biphenyls (Velzeboer et al.,
2014; Beless et al., 2014), dialkyl phthalate esters (Wang
et al., 2014), proteins (Kharlamova et al., 2013), insecticides
(Peng et al., 2009), herbicides (Pyrzynska et al., 2007; Chen
et al., 2009b), organic dyes (Gupta et al., 2013), aliphatics (Lu
et al., 2006; Wang et al., 2010a) and dioxin (Long and Yang,
2001).
Multiplemechanisms, of varying relative importance, have
been proposed to control the adsorption of SOCs by CNTs. The
quantification of these individual contributions is a chal-
lenging task, and it has yet to be addressed in any significant
fashion in the current CNT adsorption literature. However,
some previous studies indicated that certain parameters are
more predominant in controlling adsorption than others.
According to Yang et al. (2006b), Kow of SOCs were strongly
correlated with adsorption capacity of CNTs. Evidently, hy-
drophobic driving forces play important roles; however they
cannot completely explain adsorption. Electrostatic in-
teractions, pep interactions and hydrogen bonding also in-
fluence adsorption interactions considerably (Chen et al.,
wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 3 4e5 536
2007; Wang et al., 2014). According to the mechanism pro-
posed by Chen et al. (2007), p-electron donor and acceptor
interactions influence adsorption of aromatic SOCs by CNTs.
Similarly, among CNT parameters, specific surface area was
reported to be a controlling parameter of adsorption capacity,
though not the only controlling factor (Zhang et al., 2010a).
Pore volume, pore size distribution and functional groups of
CNTs were also influential on adsorption. As a result,
describing SOC adsorption by CNTs may require considering
multiple adsorption mechanisms and factors (Yang et al.,
2006b; Cho et al., 2008; Wang et al., 2008; Chen et al., 2007;
Lin and Xing, 2008a).
In addition to the already complicated nature of SOC
adsorption, CNTs are prone to aggregation. Aggregation
(homo- and hetero-) is a characteristic that differentiate CNTs
from other carbonaceous adsorbents, further complicating
the adsorption properties. According to Saleh et al. (2008),
CNTs exhibited DerjaguineLandaueVerweyeOverbeek
(DLVO) interactions that are predominantly similar to most
other aqueous colloidal particles. Aggregation may reduce the
surface area during the formation of interstitial channels be-
tweennanotubes and grooves on the periphery of the bundles.
The outermost surface, interstitial channels, inner cavities
and grooves are the four proposed adsorption sites for CNT
bundles (Fig. 1) (Agnihotri et al., 2006, 2008). The SOCs first
attach to high-energy adsorption sites and then spread to
lower energy sites among all available sorption sites
(Agnihotri et al., 2008).
In natural waters, the behavior of CNTs and the in-
teractions between organic compounds and CNTs are further
complicated by the presence of natural organic matter (NOM)
(Zhang et al., 2011). The suspension stability of CNTs in
aqueous solutions may be improved by the presence of NOM
due to the increased electrostatic and steric repulsion among
NOM-coated nanotubes, while NOM molecules compete with
Fig. 1 e Schematic representation of a typical CNT bundle
and its adsorption sites (1) inner cavities, (2) interstitial
channels, (3) external grooves, (4) outermost surface
(Agnihotri et al., 2006).
SOC molecules for adsorption sites (Long and Yang, 2001; Liu
et al., 2014).
The aqueous phase adsorption of SOCs by CNTs depends
on the physicochemical properties of the adsorbate and CNTs
as well as the background water chemistry (Ma et al., 2011). In
order to simplify the complicated nature of intermolecular
adsorption interactions, to gain amore comprehensive insight
into the adsorption mechanism, understand the interactions
of CNTs and SOCs and systemize the available literature, the
influence of adsorbent (CNT), adsorbate (SOC) and background
solution were investigated separately, as detailed in the
following sections. A literature review of the characteristics of
CNTs is also provided below prior to the review of adsorption
mechanisms.
3. Characterization of CNTs
Adsorption of organic contaminants is influenced by both CNT
morphology and surface chemistry. To examine the CNT
properties, a comprehensive literature survey was conducted
and tabulated in Table 1. The compiled information indicated
that the properties of two types of CNTs (i.e. SWCNT and
MWCNT) varied significantly in literature. In the 57 journal
articles reviewed, adsorption of SOCs was investigated using
18 SWCNTs and 81 MWCNTs as adsorbents. Five studies,
however, did not report the type of CNT used. The abundance
of MWCNT in these adsorption studies was likely due to the
higher market prices of SWCNTs than MWCNTs. A compila-
tion of prices in 2013 from 44 commercially available lab grade
CNT products revealed that the average market price for
MWCNTs was approximately 10 $/g, and the average price for
SWCNTswas higher than 100 $/g. However, it should be noted
that obtaining cheaper SWCNTs andMWCNTs (e.g., 45 $/g and
1.5 $/g, respectively) is possible depending on the purity and
properties sought (outer diameter, oxygen content etc.). In
2008, Cho et al. reported the average price of MWCNT as 140
$/g, which shows the remarkable decrease of its price
recently. In the articles reviewed, approximately 30 different
CNTswere synthesized in the lab, whereas remaining 75 CNTs
were obtained frommanufacturers, indicating thewidespread
use of commercially available CNTs. It should be noted that
the activated carbon is available for less than 5 $/kg (Babel and
Kurniawan, 2003), thus the price of CNTs should significantly
decrease to be competitive with activated carbons.
The most common CNT characteristics reported in the
adsorption literature were specific surface area (SSA), pore
volume (PV), pore size distribution (PSD), purity, elemental
analysis (EA) and morphological information such as length,
the inner and outer diameter and the number of walls and
types of functional groups. The mean SSA of SWCNTs and
MWCNTswere 373 ± 146m2/g and 216 ± 159m2/g, respectively
(calculated from the data tabulated in Table 1). Considering
various manufacturing, purification, and surface modification
techniques, high standard deviations of SSA are reasonable.
The theoretical SSA of SWCNTs with open ends was calcu-
lated as 2630 m2/g (Peigney et al., 2001). The theoretical SSA
calculations assumed that CNTs were composed of perfect
sheets of carbons that covalently formhexagonal arrays. If the
ends were closed, the inner cavity of the tube would be
Table 1 e Summary of CNTs employed for SOC adsorption in the literature, their characteristics and characterization methods.
Source Adsorbentname
Supplier Pretreat. Surface area Pore volume Inner diameter Outer diameter Length Purity
Value(m2/g)
Mthd Value(cm3/g)
Mthd Value(nm)
Mthd Value(nm)
Mthd Value(mm)
Mthd Value (%) Mthd
Yang et al. (2006b); Yang
and Xing (2009)
MWCNT 8 Chengdu, China HNO3, H2SO4 348 BET 0.816 BET 2e3 TEM <8 TEM 10e50 TEM 95þ TGA
Yang et al. (2006a); Yang
et al. (2006b); Yang and
Xing (2009); Shen et al.
(2009); Wang et al. (2014);
Tournus et al. (2005)
MWCNT 15 Chengdu, China HNO3, H2SO4 174 BET 0.665 BET 3e5 TEM 8e15 TEM 10e50 TEM 95þ TGA
Yang et al. (2006b); Yang
and Xing (2009); Wang
et al. (2014)
MWCNT 30 Chengdu, China HNO3, H2SO4 107 BET 0.317 BET 5e15 TEM 15e30 TEM 10e50 TEM 95þ TGA
Yang et al. (2006b); Yang
and Xing (2009); Wang
et al. (2014); Yan et al.
(2008)
MWCNT 50 Chengdu, China HNO3, H2SO4 95 BET 0.253 BET 5e15 TEM 30e50 TEM 10e50 TEM 95þ TGA
Lin and Xing (2008a);
Oleszczuk et al. (2009);
Peng et al. (2009); Yan
et al. (2008)
MWCNT 10 Shenzen, China HNO3, H2SO4 357 BET 1.093 BET NR e 9.4 (10) TEM (MC) 1e2 TEM 95þ MC
Oleszczuk et al. (2009); Peng
et al. (2009); Yan et al.
(2008); Wu (2007a)
MWCNT 20 Shenzen, China HNO3, H2SO4 126 BET 0.415 BET NR e 21 (20) TEM (MC) 1e2 TEM 95þ MC
Wang et al. (2008);
Oleszczuk et al. (2009);
Peng et al. (2009); Yan
et al. (2008); Wu (2007a);
Apul et al. (2013b)
MWCNT 40 Shenzen, China HNO3, H2SO4 86 BET 0.319 BET NR e 28 (40) TEM (MC) 1e2 TEM 95þ MC
Oleszczuk et al. (2009); Yan
et al. (2008); Wu (2007a)
MWCNT 60 Shenzen, China HNO3, H2SO4 73 BET 0.191 BET NR e 43 (60) TEM (MC) 1e2 TEM 95þ MC
Lin and Xing (2008a);
Oleszczuk et al. (2009);
Agnihotri et al. (2008); Wu
(2007a)
MWCNT 100 Shenzen, China HNO3, H2SO4 58 BET 0.137 BET NR e 70 (100) TEM (MC) 1e2 TEM 95þ MC
Peigney et al. (2001) MWCNT-P NAM, USA e 164 BET 0.664 BET 3e5 MC 8e15 MC 10e50 MC 89.9 TGA
Kilduff et al. (1996) MWCNT-SD NAM, USA e 178 BET 0.848 BET 5e10 MC 10e20 MC 10e30 MC 99.1 TGA
Kilduff et al. (1996) MWCNT-MD NAM, USA e 127 BET 0.722 BET 5e10 MC 20e30 MC 10e30 MC 97.1 TGA
Kilduff et al. (1996) MWCNT-LD NAM, USA e 157 BET 0.702 BET 5e15 MC 30e50 MC 10e20 MC 98.6 TGA
Kilduff et al. (1996) MWCNT-SL Nanoshel, USA e 163 BET 0.728 BET NR e 4e12 MC 3e10 MC 95.0 TGA
Kilduff et al. (1996) MWCNT-ML Nanoshel, USA e 80 BET 0.367 BET NR e 4e12 MC 5e15 MC 99.2 TGA
Kilduff et al. (1996) MWCNT-LL Nanoshel, USA e 301 BET 0.978 BET NR e 4e12 MC 15e30 MC 95.3 TGA
Peigney et al. (2001) MWCNTeOH NAM, USA e 192 BET 0.765 BET 3e5 MC 8e15 MC 10e50 MC 95þ MC
Peigney et al. (2001) MWCNT
eCOOH
NAM, USA e 134 BET 0.589 BET 3e5 MC 8e15 MC 10e50 MC 95þ MC
(continued on next page)
water
research
68
(2015)34e55
37
Table 1 e (continued )
Source Adsorbentname
Supplier Pretreat. Surface area Pore volume Inner diameter Outer diameter Length Purity
Value(m2/g)
Mthd Value(cm3/g)
Mthd Value(nm)
Mthd Value(nm)
Mthd Value(mm)
Mthd Value (%) Mthd
Kilduff et al. (1996) MWCNT-S Cheap Tube, USA e 192 BET 0.925 BET 3e5 MC 8e15 MC 0.5e2 MC 97.8 TGA
Kharlamova et al. (2013) MWCNT Shenzen, China HNO3,H2SO,
Ultrasound
295 BET NR e 5e10 MC 10e30 MC 0.5e500 MC 98.9 TGA
Chen et al. (2007); Ji et al.
(2009a); Ward and Getzen
(1970)
MWCNT 1030 Shenzen, China Heat, NaClO,
Ultrasound
148 BET 0.240 BET NR e 10e30 MC 5e15 MC 95 MC
Chen et al. (2007) MWCNT 4060 Shenzen, China Heat, NaClO,
Ultrasound
74 BET 0.104 BET NR e 40e60 MC 5e15 MC 95 MC
Yu et al. (2007) r-MWCNT Shenzen, China e 300 BET 0.793 BJH NR e 20e40 MC <10 MC 97.9 MC
Wang et al. (2008) MWCNT 3.3 Nanolab, USA e 283 BET NR e 7 MC 15 MC 1e5 AFM 95þ MC
Wang et al. (2008) MWCNT Ox. Lab Modified e 287 BET NR e 7 MC 15 MC 1e5 AFM 95 MC
Pyrzynska et al. (2007) MWCNT Aldrich 40e600 MC NR e NR e 10e30 MC NR e 95þ MC
Salam and Burk (2008) MWCNT 8 Chengdu, China HNO3 559 BET 1.775 BET 2e5 TEM <8 TEM 10e30 TEM 95þ NR
Salam and Burk (2008) MWCNT 15 Chengdu, China HNO3 181 BET 0.426 BET 5e8 TEM 10e15 TEM 10e30 TEM 90þ NR
Salam and Burk (2008) MWCNT 20 Chengdu, China HNO3 167 BET 0.566 BET 5e10 TEM 10e20 TEM 10e30 TEM 95þ NR
Salam and Burk (2008) MWCNT 30 Chengdu, China HNO3 91 BET 0.300 BET 5e15 TEM 30e50 TEM 10e20 TEM 95þ NR
Salam and Burk (2008) MWCNT 50 Chengdu, China HNO3 68 BET 0.160 BET 5e10 TEM >50 TEM 10e20 TEM 95þ NR
Wang et al. (2014) MWCNT 0.85 Chengdu, China e 167 BET NR e NR e 10e20 MC NR e NR e
Wang et al. (2014) MWCNT 2.16 Chengdu, China e 185 BET NR e NR e 10e20 MC NR e NR e
Wang et al. (2014) MWCNT 7.07 Chengdu, China e 178 BET NR e NR e 10e20 MC NR e NR e
Chen et al. (2008a) MWCNT Lab Synthesized e 160 BET 0.652 BJH NR e NR e NR e 93 EA
Chen et al. (2008a) MWCNT Ox. Lab Synthesized e 274 BET 0.382 BJH NR e NR e NR e 93 EA
Yang et al. (2008) MWCNT Sun Nanotech, China e 148 BET NR e NR e 100e200 SEM NR e NR e
Yang et al. (2008) MWCNT H2O2 Lab Synthesized 144 BET NR e NR e NR e NR e NR e
Yang et al. (2008) MWCNT
HNO3
Lab Synthesized 140 BET NR e NR e NR e NR e NR e
Coughlin and Ezra (1968) MWCNT
PD15L1-5
Nanolab, USA e 200 MC NR e NR e 15 MC 1e5 MC 95 TGA
Coughlin and Ezra (1968) MWCNT
PD30L1-5
Nanolab, USA e 200 MC NR e NR e 30 MC 1e5 MC 95 TGA
Coughlin and Ezra (1968) MWCNT
PD30L5-20
Nanolab, USA e 200 MC NR e NR e 30 MC 5e20 MC 95 TGA
Chen et al. (2008b) MWCNT Chengdu, China HCl 167 BET 1.08 BET 5e10 MC 10e20 MC 10e30 MC 95þChen et al. (2008b) MWCNT
eCOOH
Chengdu, China HCl 187 BET 1.08 BET 5e10 MC 10e20 MC 10e30 MC 95þ
Apul et al. (2012) MWCNT 3.84 Chengdu, China HCl 563 BET 0.780 BET <5 MC <8 MC 10e30 TEM NR e
Apul et al. (2012) MWCNT 10.08 Lab Modified e 520 BET 0.568 BET NR e NR e <3 TEM NR e
Apul et al. (2012) MWCNT 18.01 Lab Modified e 384 BET 0.333 BET NR e NR e <0.5 TEM NR e
Apul et al. (2012) MWCNT 22.8 Lab Modified e 143 BET 0.131 BET RN e NR e <0.3 TEM NR e
Ji et al. (2009b) O-MWCNT Shenzen, China HCl 114 BET NR e NR e 10e30 MC 5e15 MC 95þ MC
Datsyuk et al. (2008) MWCNT 10 Chengdu, China e 176 BET 1.748 BJH NR e 10e20 MC NR e 95þ MC
Datsyuk et al. (2008) MWCNT 20 Chengdu, China e 130 BET 1.169 BJH NR e 20e30 MC NR e 95þ MC
Datsyuk et al. (2008) MWCNT 50 Chengdu, China e 77 BET 0.696 BJH NR e >50 MC NR e 95þ MC
water
research
68
(2015)34e55
38
Datsyuk et al. (2008) F-MWCNT 10 Chengdu, China e 174 BET 1.366 BJH NR e 10e20 MC NR e 95þ MC
Datsyuk et al. (2008) F-MWCNT 20 Chengdu, China e 123 BET 1.060 BJH NR e 20e30 MC NR e 95þ MC
Datsyuk et al. (2008) F-MWCNT 50 Chengdu, China e 77 BET 0.587 BJH NR e >50 MC NR e 95þ MC
Ji et al. (2009c) MWCNT Shenzen, China e 52 BET 0.16 BET NR e 20e40 MC NR e NR e
Wang et al. (2010b) MWCNT 2.0 Lab Synthesized Heat, Air,
HNO3
471 BET 0.64 BET NR e NR e NR e NR e
Wang et al. (2010b) MWCNT 3.2 Lab Synthesized Heat, Air,
HNO3
381 BET 0.58 BET NR e NR e NR e NR e
Wang et al. (2010b) MWCNT 4.7 Lab Synthesized Heat, Air,
HNO3
382 BET 0.58 BET NR e NR e NR e NR e
Wang et al. (2010b) MWCNT 5.9 Lab Synthesized Heat, Air,
HNO3
327 BET 0.49 BET NR e NR e NR e NR e
Ward and Getzen (1970) MWCNT in Nanothinx, Greece Ultrasound,
HNO3
525 BET NR e NR e 12e31 MC >10 MC 97 MC
Ward and Getzen (1970) MWCNT 500 Lab Modified e 320 BET NR e NR e NR e NR e NR e
Wang et al. (2007) MWCNT MH Chengdu, China Hydoxylized 228 BET NR e NR e NR e NR e 94 EA
Wang et al. (2007) MWCNT MC Chengdu, China Carboxylized 164 BET NR e NR e NR e NR e 97 EA
Wang et al. (2007) MWCNT MG Chengdu, China Graphitized 117 BET NR e NR e NR e NR e 98 EA
Zhang et al. (2012) MWCNT 800 Lab Modified e 263 BET NR e NR e NR e NR e NR e
Zhang et al. (2012) MWCNT 1200 Lab Modified e 245 BET NR e NR e NR e NR e NR e
Material Genome Initiativ,
2014)
MWCNT Shenzen, China HCl NR e NR e NR e 20e30 MC NR e NR e
Ghaedi et al. (2012a) MWCNT R.I.P.I., Iran e 280 NR NR e NR e <10 MC 5e15 MC 95þ MC
Yang et al. (2010); Chen et al.
(2008c)
MWCNT Lab Synthesized e 82 BET 1.07 BET 5e10 TEM 20e80 TEM 5e15 98þ
Rambabu et al. (2012);
Oleszczuk and Xing (2011)
MWCNT Merck, Germany HCl NR e NR e NR e NR e NR e NR e
Shi et al. (2010) MWCNT
eCOOH
Lab Modified HCl 97 NR NR e NR e NR e NR e NR e
Kotel et al. (2009); Liu et al.
(2008)
MWCNT Lab Synthesized HCl 181 BET 0.345 BET NR e 3e40 TEM NR e 95þ FTIR
Ghaedi et al. (2011) f-MWCNT Lab Synthesized Air, Acid 92 BET 0.22 BET 5e10 TEM, SEM 40e50 TEM, SEM NR e NR e
Wu (2007b) MWCNT Sun Nanotech,
China
HCl, HNO3 162 BET NR e NR e 20e50 SEM 20e50 SEM NR e
Ghaedi et al. (2012b) MWCNT Lab Synthesized e NR e NR e NR e 20 NR NR e NR e
Ghaedi et al. (2012b) O-MWCNT Lab Synthesized NR e NR e NR e 20 NR NR e NR e
Chen et al. (2011) P-SWCNT Shenzen, China e 411 BET NR e NR e <2 MC 5e15 MC 90þ MC
Chen et al. (2011) P-MWCNT Shenzen, China e 157 BET NR e NR e 10e30 MC 5e15 MC 95þ MC
Chen et al. (2011) K-SWCNT Shenzen, China e 653 BET NR e NR e NR e NR e NR e
Chen et al. (2011) K-MWCNT Shenzen, China e 423 BET NR e NR e NR e NR e NR e
Machado et al. (2012) MWCNT NR e 157 BET NR e NR e NR e NR e NR e
Machado et al. (2012) NH3-MWCNT Lab Modified e 195 BET NR e NR e NR e NR e NR e
Machado et al. (2012) HNO3-
MWCNT
Lab Modified e 152 BET NR e NR e NR e NR e NR e
Yang et al. (2006b); Wang
et al. (2014); Peng et al.
(2009)
SWCNT Chengdu, China HNO3, H2SO4 541 BET 1.021 BET 0.8e1.6 TEM 1e2 TEM 10e50 TEM 90þ TGA
Peigney et al. (2001)) SWCNT-P Cheap Tube, USA e 486 BET 0.722 BET 0.8e1.6 MC 1e2 MC 5e30 MC 88.6 TGA
(continued on next page)
water
research
68
(2015)34e55
39
Table 1 e (continued )
Source Adsorbentname
Supplier Pretreat. Surface area Pore volume Inner diameter Outer diameter Length Purity
Value(m2/g)
Mthd Value(cm3/g)
Mthd Value(nm)
Mthd Value(nm)
Mthd Value(mm)
Mthd Value (%) Mthd
Kilduff et al. (1996) SWCNT-L Cheap Tube, USA e 442 BET 0.917 BET 0.8e1.6 MC 1e2 MC 5e30 MC 96.0 TGA
Kilduff et al. (1996) SWCNT-S Chengdu, China e 413 BET 0.613 BET 0.8e1.6 MC 1e2 MC 0.5e2 MC 97.3 TGA
Peigney et al. (2001) SWCNTeOH Chengdu, China e 420 BET 0.739 BET 0.8e1.6 MC 1e2 MC 5e30 MC 91.1 TGA
Peigney et al. (2001) SWCNT
eCOOH
Chengdu, China e 386 BET 0.680 BET 0.8e1.6 MC 1e2 MC 5e30 MC 95.9 TGA
Chen et al. (2007); Ji et al.
(2009a); Peng et al. (2003)
SWCNT Shenzen, China Heat, NaClO,
Ultrasound
370 BET 0.425 BET NR e <2 MC 5e15 MC 90 MC
Yu et al. (2007) SWCNT 20 Shenzen, China e 167 BET 0.417 BJH NR e 10e20 MC 1e2 MC 95þ MC
Yan et al. (2008) SWCNT Shenzen, China HNO3, H2SO4 541 BET 1.021 BET NR e 1.4 (1e2) TEM (MC) 5e15 TEM 90 þ 50þ MC
Kato et al. (2008) SWCNT Shenzen, China e 447 BET 0.974 BET NR e NR e NR e 87 EA
Toth et al. (2012)) SWCNT Aldrich, USA e 248 BET NR e NR e 1.2e1.5 SEM 2e5 SEM 50e70 NR
Toth et al. (2012) SWCNT Ox. Lab Synthesized HNO3 284 BET NR e NR e 1.2e1.5 SEM 2e5 SEM 50e70 NR
Coughlin and Ezra (1968) SWCNT 1 Cheap Tube, USA e 407 MC NR e NR e 1e2 MC 5e30 MC 90 Raman, XRD
Coughlin and Ezra (1968) SWCNT 2 Carbon Solutions,
USA
e 355 Calculated NR e NR e 1.55 MC 1e5 MC 40e60 NIR
Coughlin and Ezra (1968) SWCNT
Carbolex AP
Aldrich, USA e 240 Calculated NR e NR e 1.4 MC 3e5 MC 66 TGA
Coughlin and Ezra (1968) SWCNT
HipCo
Carbon
Nanotechnologies,
USA
e 633 BET NR e NR e 1e2 0.4e0.7 MC NC 85 TGA
Ji et al. (2009c) SWCNT Shenzen, China e 406 BET 1.11 BET NR e <2 MC NR e NR e
Liu et al. (2008) SWCNT Lab Synthesized HCl 388 BET 0.662 BET NR e 1e2 TEM NR e 99þ FTIR
Zhang et al. (2010a) As grown
CNT
Lab Synthesized e 134 NR NR e NR e NR e NR e NR e
Zhang et al. (2010a) Graphitized
CNT
Lab Synthesized Heat, N2 126 NR NR e NR e NR e NR e NR e
Su and Lu (2007) CNT Shenzen, China e 550 BET 1.150 BET 4.5 TEM <10 MC 5e15 MC 95þ MC
Kuo et al. (2008) CNT Lab Synthesized Heat 145 BET 0.398 BET NR e 30 TEM NR e NR e
Xie et al. (1997) CNT Shenzen, China e 189 BET NR e NR e 20e40 MC 5e10 MC 97 NR
Machado et al. (2011) CNT Lab Synthesized Acid Wash,
Air
NR e NR e NR e NR e NR e NR e
Adsorbent Name: Adsorbent names were assigned by their associated authors in the original articles; Supplier: The short names for manufacturer companies and their countries (full names of
companies can be found in Table S2 in Supplementary Information); Pretreat: The pretreatment techniques that were applied to CNTs after purchase if reported by the authors.
Acronyms used in the table: MC: Manufacturer claimed, NR: Not reported, Mthd: Analytical method employed for obtaining the morphological information, BET: BrunauereEmmetteTeller, BJH:
BarretteJoynereHalenda, FTIR: Fourier transform infrared spectroscopy, TEM: Transmission electron microscopy, SEM: Scanning electron microscopy, TGA: Thermogravimetric analysis, EA:
Elemental Analysis, AFM: Atomic force microscopy, XRD: x-ray diffraction, NIR: near-infrared spectroscopy.
water
research
68
(2015)34e55
40
wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 3 4e5 5 41
unavailable and the theoretical SSA would decrease to
1315 m2/g. The reported SSAs of SWCNTs were significantly
lower than the calculated theoretical surface areas, a decrease
attributed to the tight aggregation of SWCNTs (Babel and
Kurniawan, 2003). Zhang et al. (2009) reported 19e37 individ-
ual tubes were forming bundles with a diameter 5e7 times
larger than an individual SWCNTs.MWCNTswere assumed to
be concentric tubular sheets with a distance of 0.45 nm be-
tween walls; therefore, the theoretical SSA is a function of
number of walls. The measured and theoretical SSA of
MWCNTs were reported to be in better agreement, which was
attributed to the looser aggregation state ofMWCNTs allowing
the tubes to behave like isolated ones (Zhang et al., 2009).
The PV of SWCNTs was slightly higher than the PV of
MWCNTs. The mean total PVs were 0.78 ± 0.23 cm3/g and
0.64 ± 0.39 cm3/g for SWCNT and MWCNT, respectively.
Theoretically, the pore volume occupied by the inner pores
of SWCNTs is considerably higher when compared to
MWCNT inner pore volume because the inner walls of
MWCNTs occupy space whereas the SWCNTs have open
inner channels. Zhang et al. (2009) reported that the avail-
ability of inner pores significantly affected the pore volume
especially for SWCNTs.
SSA and PV are similar parameters that are expected to
show an association with each other. However, the data
compiled for all CNTs from the literature showed no correla-
tion between SSA and the total PV (R2 ¼ 0.18). Investigating the
CNT types separately yielded higher correlations (R2 ¼ 0.50 for
SWCNT; R2 ¼ 0.27 for MWCNT) supporting that SWCNTs and
MWCNTs have different aggregation states. It should be also
noted that the SSA and PV of CNTs are determinedwith N2 gas
adsorption in their bulk phases. Given the aggregating nature
of CNTs in water, to better interpret adsorption experiments
performed in aqueous systems, it is important to develop
techniques to characterize the SSA and PV distribution of CNT
aggregates in water. The aggregation state of CNTs in aquatic
systems using time-resolved dynamic scattering and UVevis
spectrometer (Saleh et al., 2008, 2010; Khan et al., 2013; Yu
et al., 2007); however, no study has correlated the aggrega-
tion state of CNTs with their SSA and PVs in aquatic medium
to the best of our knowledge.
The inner and outer diameter ranges of MWCNTs were
2e15 nm and 4e200 nm, respectively. The wide range of the
outer diameter was attributed to variations in the number of
concentric tubular layers (walls) of CNTs. The diameters of
SWCNTswere in the range of 0.8e2 nmwith one exception, in
which the SWCNT diameter range was reported as 10e20 nm
as claimed by the manufacturer (Yan et al., 2008). The length
of MWCNT and SWCNT were in the ranges of 0.3e500 mm and
0.4e50 mm, respectively. Almost all reported morphological
properties (i.e. inner and outer diameter, length) were either
claimed by the manufacturer and/or measured by electron
microscopy (transmission, scanning electron microscopy)
(e.g., Oleszczuk et al., 2009, Lin and Xing, 2008b).
Approximately 50% of the CNTs reported were subject to
pretreatment for purification or targeted functionalization
prior to use (Table 1). The characterization of CNTs after pre-
treatment, and not relying only on the information provided
by the manufacturers for virgin CNTs, is critical to avoid
misinterpretation of the adsorption results. The most
common pretreatment techniques used have been acid wash
(HNO3, HCl, H2SO4), heat treatment and ultrasonication.
Several studies oxidized the CNTs to increase the surface
polarity (Salam and Burk, 2008; Chen et al., 2009b; Zhang et al.,
2010c). Functionalization of the CNTs has occurred at the
sidewalls, and either at ends of the tubes or the defect sites
through the covalent and non-covalent attachment of func-
tional groups. Non-covalent functionalization may be benefi-
cial since it has been shown not changing the CNT pore
texture (Ma et al., 2011; Datsyuk et al., 2008). Adding oxygen
containing functional groups on the CNT surface is the most
commonly used functionalization approach. The amount and
form of oxygen-containing functional groups (eOH, eCOOH,
>C]O) depend upon the type of oxidation technique and the
acid utilized for purification (Ma et al., 2011). Oxidation gives
CNTs hydrophilic moieties and removes impurities, amor-
phous carbon and hemispherical caps (Datsyuk et al., 2008).
The chemical modification of CNTs may also alter their
physical properties. There are different reports regarding the
impact of surface oxidation on the CNTmorphology. Cho et al.
(2008) oxidized MWCNTs by refluxing varying strengths
(10e70% w/w) of HNO3, KMnO4 and H2O2 solutions at elevated
temperatures (80e140 �C). The controlled oxidation produced
an array (1 pristine and 8 oxidized) of CNTs with surface ox-
ygen ranging from 3.3 to 14%. Authors reported no change in
SSAs after oxidation. Salam and Burk (2008) studied CNT
surface oxidation by 18% H2O2 and 8 M HNO3 at elevated
temperatures (80e140 �C). Attachment of ~1 mmol/g acidic
functional groups was detected; however, there was no
change in SSA. Wang et al. (2010b) also reported no SSA dif-
ference between the raw (0.2e1.0%) and oxidized (2.2e4.3%)
CNTs. In their studies, raw and oxidized CNTswere purchased
from a supplier and the oxidation techniques were not re-
ported. On the other hand, Ji et al. (2010a) reported a
‘remarkable’ increase in SSA from 410 to 650 m2/g and 160 to
420 m2/g for SWCNTs and MWCNTs, respectively, while using
dry KOH etching to functionalize SWCNTs and MWCNTs. The
dry KOH etching was described as contacting CNTs with KOH
powder and heating it to 800 �C under N2 stream yielding ~10%
oxygen containing functional groups. The increase in SSA
after oxidation was attributed to the removal of amorphous
carbon during functionalization (Shen et al., 2009; Chen et al.,
2009b). In contrast, Yu et al. (2012) reported a notable decrease
in SSA (from 471 m2/g to 327 m2/g) when the total surface
oxygen content was increased from 2.0% to 5.9%. The re-
searchers used NaClO solution at ambient temperature for
12 h to oxidize the pristine MWCNTs. Salam and Burk (2008)
also reported a 26% decrease (from 144 to 106 m2/g) in SSA
when 1MKMnO4 is applied to CNTs at 80 �C for 4 h followed by
a H2SO4 treatment. This decrease was attributed to the
strength of the oxidizing agent that damaged the MWCNTs.
These different findings indicate the importance of CNT type
as well as the oxidizing agent and its strength on alteration of
CNT surface chemistry. In addition to SSA, Cho et al. (2008)
reported no change in structure or length distribution of
CNTs after oxidation. However, Wu (2007a) reported a
decrease in the diameter of MWCNTs after oxidation and
attributed the modification of this diameter to the removal of
the amorphous carbon from the surface. These results clearly
demonstrate the necessity of extensive characterization of
wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 3 4e5 542
CNTs after pretreatment and/or functionalization for more
accurate interpretation of their adsorption behaviors.
4. Adsorption of SOCs by CNTs
4.1. Influence of CNT properties
The physical properties of CNTs play an important role in the
adsorption of organic contaminants. Yang et al. (2006b)
experimentally estimated that adsorption capacity of hydro-
phobic SOCs (i.e. PAHs with log Kow ranging 3.36e5.18) on
CNTs increases with SSA and PV. Similarly Oleszczuk et al.
(2009) showed a positive correlation of SSA, and micro- and
mesopore volumes with the adsorption capacity of MWCNTs
for two rather hydrophilic pharmaceuticals (i.e. carbamaze-
pine, log Kow �1.22 and oxytetracycline, log Kow 2.45). Both
studies presented strong linear relationships (R2 � 0.97) be-
tween the adsorption capacity of SOCs on MWCNTs and their
SSAs. Zhang et al. (2009) also reported that SSA and PV were
influential for the adsorption of hydrophobic SOCs, but they
were not the exclusive factors determining adsorption ca-
pacities. Li et al. (2013a) reported higher methylene blue
adsorption by activated carbon than CNT per unit mass, yet
the order was reversed after surface area normalization.
Similarly, Beless et al. (2014) presented single point sor-
bentewater partitioning coefficients of eleven PCB congeners,
and reported similar CNT adsorption affinities with activated
carbon and black carbon per unit area. On the other hand,
Wang et al. (2009) observed SSA dependency for adsorption of
macromolecular humic acids by MWCNTs only at higher
concentrations because at low concentrations the sorption
site availability was not a limiting factor (thus not controlling
adsorption).
Sorption sites such as inner pores and interstitial channels
can impede the penetration of especially large organic mac-
romolecules. Known as the size exclusion phenomenon, this
remarkably decreases the adsorption of organics by CNTs
(Wang et al., 2009), a similar observation has also been re-
ported on activated carbons (e.g., Kilduff et al., 1996). The ac-
cess of SOCs to the inner regions of CNTs or CNTs bundles
may also be hindered due to amorphous carbon or metal
catalysts that were introduced during synthesis (Gotovac
et al., 2007c), and water cluster formation around the oxygen
containing functional groups (Zhang et al., 2009).
The nanocurvature and diameter of CNTs also influence
SOC adsorption. In their study, Gotovac et al. (2007a) reported
that the alignment between PAH molecules and CNT surface
affected adsorption. They observed that tetracene molecules,
which have four benzene rings aligning with the SWCNT sur-
face, while phenanthrene had only 2.5 rings in alignment,
resulting in a six-fold greater increase of tetracene adsorption
over phenanthrene. Also the increasing strength of pep
interaction causedmore benzene rings to align on the surface.
Apul et al. (2012) also reported a positive correlation between
the outer diameter of CNTs and surface area normalized
phenanthrene adsorption, which was attributed to the better
alignment of adsorbed planar phenanthrenemolecules due to
the decreasing surface curvature with increasing CNT diam-
eter. Li et al. (2013b), investigated the diameter effect on the
adsorption of perfluorooctanesulfonate by SWCNT using mo-
lecular dynamics simulations. Authors reported diameter de-
pendency of binding energies for the attachment of molecules
to the inner sites of CNTs. A sharp increase in the binding en-
ergy fromþ60 to�60 kcal/mol was observed with the increase
of diameter from 0.8 to 1.0 nm. The sorption became less
favorable (�60 to�30 kcal/mol) as thediameter increased from
1.0 to 1.6 nm. A relatively steady increasing trend (�11 to
�15 kcal/mol) was observed for the binding energies at the
outer surface of SWCNTs with increasing diameters. Consid-
ering that SWCNTs have typically narrow range of diameters
(0.8e2.0 nm), similar simulations should also be investigated
for MWCNTs with larger diameter ranges (2e200 nm).
To the best our knowledge, the affect of CNT length, and
chirality (a parameter related to the angle between graphene
plane and the tube axis) on SOC adsorption have yet to be
reported in the literature.
In addition to physical characteristics, CNT surface
chemistry can also influence the SOC adsorption. The unin-
tended oxidation of the surface during manufacturing and/or
in the environment, and the intentional oxidation with
treatment are some possible causes of CNT surface oxidation.
Several studies have reported an overall decrease in the SOC
adsorption with an increase in surface oxygen content (Cho
et al., 2008; Shen et al., 2009; Salam and Burk, 2008; Yu et al.,
2014; Chen et al., 2009b; Zhang et al., 2009; Li et al., 2011).
Two mechanisms have been proposed to explain these ob-
servations: (i) the presence of oxygen on the CNT surface
making adsorption of water molecules energetically more
favorable relative to SOC adsorption, which results in water
clusters that deplete the available surface area for SOCs; (ii)
the presence of oxygen on CNT surface localizes the p elec-
trons, which reduces the pep interactions between the CNT
graphitic surface and benzene rings of aromatic SOCs. Similar
findings were presented and discussed for activated carbon
(Yu et al., 2014; Kato et al., 2008; Coughlin and Ezra, 1968). On
the other hand, in their investigation of the adsorption of
hydroxyl- and amino-substituted aromatics on oxidized
CNTs, Chen et al. (2008b) reported higher removal of amino
and hydroxyl containing SOCs (2-naphtol, 2,4-dichlorophenol
and 1-naphtylamine) than structurally similar 1,3-
dichlorobenzene and naphthalene by oxidized CNTs, which
was attributed to stronger nonhydrophobic attraction due to
hydrogen bonding and p-hydrogen bonding interactions be-
tween eOH containing adsorbates and oxidized CNT surface.
Yu et al. (2012) reported a remarkable increase (~100%) in
adsorption capacities of toluene, ethylbenzene and m-xylene
with increasing surface oxygen content per SSA, up to ~8%, for
MWCNTs. The increase was attributed to the increase in the
dispersion of CNTs, also increasing the available adsorption
sites. However, a further increase in the oxygen content per
SSA (up to 18%) showed a decreasing trend in the adsorption
capacity. The authors attribute this opposing behavior to
water cluster formation effect dominating over the CNT
dispersion for SOC adsorption.
4.2. Influence of SOC properties
The driving forces of SOC adsorption are attraction to the CNT
surface and repulsion from the background solution (i.e.
wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 3 4e5 5 43
water). The predominant repulsive force that repels SOCs
from water onto the CNT surface are hydrophobic forces that
are either associatedwith the size and polarity of themolecule
or with the protonation state of ionizable compounds. There
are also physical and chemical attractive forces between the
SOCs and CNT surface. Nonspecific interactions are major
contributors of overall attractive forces, which result from the
affinity of electron-deficient and electron-rich regions of un-
charged molecules. These nonspecific attractions are gener-
ally referred as van der Waals (vdW) interactions. Though
nonspecific attractive forces contribute to the overall inter-
molecular attraction regardless of the SOC properties, the in-
tensity of these forces depends upon the molecular size,
electric charge and polarizability. Other conditional attractive
forcesmay contribute to adsorption, such as pep interactions
between the resonating p electrons of the graphitic structure
of the CNT surface and the p electrons of aromatic SOCs.
Hydrogen bonding is a polarity dependent electrostatic
attraction between certain SOCs with functional groups (such
as eOH) and functionalized surfaces of CNTs. Electrostatic
interactions also occur between the charges of ionizable SOCs
and charged surfaces of CNTs.
Hydrophobicity, represented mostly by octanolewater
partitioning coefficient (Kow) or aqueous solubility (Cw), was
reported as the most predominant adsorbate property con-
trolling adsorption of several SOCs by MWCNTs (Yang et al.,
2006a, 2006b; Wang et al., 2008; Wang et al., 2009) and by
SWCNTs (Brooks et al., 2012). After hydrophobicity normali-
zation, MWCNT and SWCNT adsorption capacities were
relatively comparable for two SOCs (phenanthrene and
biphenyl) (Zhang et al., 2009). On the other hand, Wang et al.
(2009) reported hydrophobicity normalized adsorption ca-
pacities following the order of molecular sizes for tested SOCs
(phenanthrene > naphthalene > 1-naphtol) indicating that
larger molecules have more affinities provided that SOCs had
no hydrophobic differences. On the other hand, Yang et al.
(2006b) reported a negative correlation between the adsorp-
tion of naphthalene, phenanthrene and pyrene to molecular
size, which they attributed to the poor access of large mole-
cules to themicropores. Themolecular size directly influences
adsorption because the hydrophobic affinity cannot overcome
the steric hindrances, a finding that was confirmed by Chen
et al. in which the determined that the molecular sieving ef-
fect prevent bulky 1,2,4,5-tetrachlorobenzene molecules from
accessing the innermost surfaces (Chen et al., 2007). These
findings show the importance of the bundle structure of CNTs
and the availability of adsorption sites to SOCs. According to
Wang et al. (2010b), the primary adsorption sites of SOCs were
the external curved surfaces of CNTs since the probe mole-
cules (naphthalene) failed to penetrate into the hollow spaces
within the nanotubes. Toth et al. (2012) also concluded that
CNT adsorption occurred on the external surfaces exclusively.
Chen et al. (2007) supported the concept that sorption sites of
MWCNT were located on the innermost and outermost sur-
faces because the interlayer spacing between coaxial tubes are
impenetrable to SOCs. Kinetic experiments conducted by
Shen et al. (2009) suggest that the MWCNT pore-filling
mechanism of nitroaromatic organics has a relatively low
contribution to overall adsorption, indicating sorption to
external surfaces rather than interstitial channel spaces.
Zhang et al. (2009) reported that rigid and planar phenan-
threne molecules can mainly attach to the external surface
area of interstitial channels of SWCNTs. Unlike phenan-
threne, nonplanar and flexible biphenyl and 2-phenylphenol
molecules have adjustable molecular configurations for bet-
ter packing in the MWCNT bundles. Similarly, Pan and Xing
(2008) reported that the easily rotating structure of
bisphenol-A allowed the molecule wedging into the groove
regions of CNTs unlike rigid 17a-ethinyl estradial molecules.
Oleszczuk et al. (2009) applied ultrasonication to disperse CNT
bundles and quantified the desorption of carbamazepine and
oxytetracycline from CNTs. After ultrasonic pretreatment (i.e.
disintegration of CNT bundles) oxytetracycline molecules
were released more from CNTs when compared to non-
sonicated CNTs, indicating the presence of molecules trap-
ped in the interstitial areas of CNTs. However, the dispersion
state of CNTs did not influence carbamazapine adsorption,
due to the attraction of adsorbed carbamazepine molecules to
free molecules that cause stacking of carbamazepine regard-
less of the aggregation state. Overall, these reports indicate
that the hydrophobicity itself may not explain the adsorption
affinity of SOCs on CNTs, and the accessibility of SOCs to
different regions of CNT bundles and pores can be important
for the adsorption of SOC molecules.
Another principle SOC property influencing adsorption by
CNTs is the p electron density of the compounds resonating in
aromatic rings (also possibly for some aliphatic chains with
double bonds). In that the graphitic surfaces of CNTs have
regions with rich and poor p electrons, the interaction of p
electrons influence adsorption (Gotovac et al., 2006; Oleszczuk
et al., 2009; Chen et al., 2009b; Long and Yang, 2001; Zhang
et al., 2009; Wang et al., 2010b). In their comparison of hy-
drophobicity normalized adsorption affinities of lindane (i.e. a
cyclic chlorinated aliphatic molecule) with atrazine and
phenanthrene (i.e. two aromatic SOCs), Wang et al. (2010b)
concluded that pep interactions influence adsorption posi-
tively. Chin et al. (2007) reported a 60% increase in the
adsorption of o-xylene onto SWCNT after purification by 3 M
nitric acid at 120�C for 3 h at pH 3. At pH 5 and 7, no notable
increase was observed. This increase was attributed to pep
attraction of xylene molecules and CNT surface. The methyl
groups on o-xylene were protonated and pushing electrons
towards the benzene ring increasing the electron density of
resonating p electrons enhancing the attraction between the
partially positive SWCNT surfaces, which becomes more
profound at pH 3. At pH 5 and 7, no notable increase was
observed because at pH 3, themethyl groups on o-xylene were
protonated and pushing electrons towards the benzene ring
increasing the electron density of resonating p electrons. Ac-
cording to Lin and Xing (Lin and Xing, 2008a), the sorption
affinity of cyclohexanol was lower than phenol, and this dif-
ference was attributed to the pep interaction because cyclo-
hexanol is missing p electrons and has comparable
hydrophobicity with even a lower solubility than phenol. Ac-
cording to Chen et al. (2007), the adsorption affinity of the two-
ring 2-naphtol was stronger than one ring 2,4-dichlorophenol.
The higher adsorption of 2-naphtol was attributed to stronger
conjugation potential of two rings resultingwith stronger pep
interactions. This study is supported by a comparison of two
aromatic compounds: phenanthrene is adsorbed more than
wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 3 4e5 544
naphthalene because it has onemore benzene ring that allows
more polarization and higher dispersive forces with highly
polarizable CNT surfaces (Wang et al., 2009). Additionally, the
electron induction effect of chlorine atoms reduces the elec-
tron density of 2,4-dichlorophenol in the benzene ring, which
suppresses its adsorption affinity. The pep stacking is ob-
tained when the benzene rings of SOCs align with the CNT
surface as demonstrated by Gotovac et al. (2007a); therefore,
the contribution of pep interaction between CNTs and the p
electrons in aliphatic compounds depends on the molecular
arrangement of molecules. Pan and Xing (2008) support the
importance of molecular configuration by demonstrating the
attachment of bisphenol-A with two benzene rings on CNTs
along the circumference, suggesting that the pep electron
donor acceptor complexes are stronger than pairs of donor or
pairs of acceptor. The thermodynamic calculations indicated
that adsorption of bisphenol-A along the outer circumference
of CNTs is very unlikely due to high energy requirement for
the steric conformation. Therefore, the presence ofpep bonds
may not always promote adsorption, and molecular configu-
ration or surface conformation may also play an important
role. Chen et al. (2007) propose pep electron donor acceptor
interactions between the p electron rich aromatic rings of
adsorbates and p electron depleted regions of CNT surface.
The ground-state hybrid structure of the pep electron donor
acceptor system consisted of electrostatic forces between sep
quadrupoles of opposing benzene rings. This mechanism in-
volves one-electron transfer from the highest occupied mo-
lecular orbital to the lowest unoccupiedmolecular orbital. The
resulting bond is a weak covalent bond formed by unpaired
electrons (Lin and Xing, 2008a). Chen et al. (2007) also reported
an “extremely strong adsorption of 1-naphthylamine on
CNTs”. This finding was attributed to the presence of the un-
shared pair of electrons of nitrogen on the amino group
(eNH2) making the benzene ring electron rich, which strongly
(stronger thaneOH) interacts with electron poor groups of the
CNT surface. Since electron depleting regions are likely to be
limited, strong adsorption affinity is expected to be more
notable at low concentrations. Similarly, Chin et al. (2007)
concluded that adsorption of xylenes depends upon the po-
sition of the methyl group on the benzene ring resulting from
the repulsive impact of the methyl group on the p electron
density of the xylene molecule. Wang et al. (2010a) tested the
adsorption of dialkyl phthalate esters and pep electron donor
acceptor interactions were proposed after hydrophobicity
normalization because of the p electron-accepting ester
functional group. Another study (Lin and Xing, 2008a) inves-
tigated the adsorption of chlorophenols onto pristine and
functionalized MWCNTs and reported a reduced capacity on
oxidized CNT surfaces, because the pep dispersion was
weakened by the oxygen containing functional groups on the
surface. The substituent on the benzene ring was found to
influence the resonance and time-dependent electron density
of the aromatic SOCs, thus influencing the pep interactions.
Adsorption simulations supported the coupling of p electrons
between CNT surface and aromatic molecules (such as ben-
zene, azulene, DDQ, pyrene and nitrotyrosine) using density
functional theory optimizing the binding energies (Zhao et al.,
2003; Tournus et al., 2005).
Hydrogen bonding is another principal dipoleedipole
attraction between a hydrogen atom and an electronegative
atom such as nitrogen or oxygen; however; it cannot be eval-
uated completely independent from pep interactions (Chen
et al., 2008b). Several CNT adsorption studies have empha-
sized the contribution of hydrogen bonding to overall adsorp-
tion (Lin and Xing, 2008a; Yang and Xing, 2009; Wang et al.,
2010a; Li et al., 2011). According to Wang et al. (2008) 1-
naphtol molecules may form hydrogen bonds with the oxy-
gen containing functional groups of CNT surface or benzene
rings of 1-naphtol can be aligned to the CNT surface, which
leaves theeOH functional group facing the aqueous phase that
forms new hydrogen bonding sites for free 1-naphtol mole-
cules in water. Li et al. (2013c) tested the adsorption of phen-
anthrene and phenol, in the range of ~2e1000 ppb and
100e1000 ppb, respectively, onto pristine (3.4% O) and oxidized
(11.3% O) MWCNTs. Adsorption of phenol by oxidizedMWCNT
was higher than pristine MWCNT. Opposite trend was
observed for adsorption of phenanthrene. Authors attributed
the difference to the hydrogen bonding between phenol mol-
ecules and the oxidized surface of MWCNTs. Yang et al. (2010)
also proposed hydrophobic attraction between 2,4-
dichlorophenol, 4-chloroaniline and oxygen containing
groups of MWCNTs, they emphasized the dependency of pH,
because once the molecules are dissociated, the hydrogen
bonding ability disappears. Yang et al. (2010) also found that
nitro-, chloride- and methyl-functional groups attached to
phenols or anilines enhance the affinity toMWCNTs. The order
of their influence followed the nitro group > chloride
group > methyl group. They also reported the dependency of
phenol substitution pattern by observing higher adsorption
when the hydroxyl group is attached in meta-position rather
than ortho- and para-position. However, the authors avoided
conclusive statements about the influence of group positions
because different functional groups had different influences
on adsorption affinity (Shen et al., 2009). On the other hand,
Chen et al. (2007) found that the contribution of hydrogen
bonding on overall CNT adsorption of nitroaromatics was not
significant. After testing the adsorption of 2,4-dinitrotoluene
onto oxidized SWCNTs (17% oxygen) at a pH range of 2.8e7.3,
they observed a slight decrease in adsorption capacity with
decreasing pH. On the contrary, an increase in adsorption ca-
pacity was expected as the pH decreases if the hydrogen
bonding was the controlling adsorption mechanism. In such a
case, the eCOO� functional groups on the CNT surface would
be protonated and eCOOH groups would act as H-bond donor
and form hydrogen bonds with H-bond accepting nitro groups.
Lin and Xing (2008b) and Oleszczuk et al. (2009) also reported
that hydrogen bonding might not be significant in that there
was no proportional increase with the number of eOH groups
of their adsorbates; the CNTs they used for adsorption, how-
ever, had a very low hydrogen and oxygen content. Yang and
Xing (2009) also determined that fulvic acid can act as a
hydrogen bond donor due to carboxylic and phenolic moieties,
while the CNT surface may act as hydrogen bond acceptors.
They also reported electrostatic interactions between the sur-
face charges of fulvic acid and CNTs, indicating that these in-
teractions would strongly depend on the pH. In demonstrating
the adsorption of 2,4,6-trichlorophenol at different pH values
wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 3 4e5 5 45
Chen et al. (2009a) found that increasing the pH values
increased the dissociated (i.e. negatively charged) fraction of
the compound, which decreased the adsorption from repul-
sion between the negatively charged surface of CNT and
anionic 2,4,6-trichlorophenol. In addition, the increased solu-
bility of the ionized form of the adsorbate reduced the
adsorption. Similarly, decreasing uptake of organic acids onto
activated carbon was reported with increasing pH (ranging
from 1 to 11) (Ward and Getzen, 1970). Shen et al. (2009) pre-
sented stronger adsorption affinities of aromatic compounds
with more nitro groups because these groups are strong elec-
tron acceptors interacting with highly polarizable electron-
donating graphitic surfaces of MWCNTs. At lower pH values,
adsorption was favored because of less ionization of contam-
inants such as acidic herbicides (Pyrzynska et al., 2007), direct
dyes (Kuo et al., 2008) or sulfonamide antibiotics (Ji et al.,
2009b). At pH levels around pKa, both electrostatic and hy-
drophobic sorption mechanisms are anticipated; however, it
was not possible to distinguish between those two mecha-
nisms (Pyrzynska et al., 2007).
4.3. Influence of background solution properties
NOM is ubiquitous in natural waters and adsorbs on CNT
surfaces influencing its SOC adsorption (Su and Lu, 2007; Yang
and Xing, 2009; Wang et al., 2007). The net influence of NOM
on SOC adsorption by CNTs is a tradeoff between two opposite
effects (Zhang et al., 2010c): (i) the competition by NOM
depleting the sorption sites for SOC adsorption (Hou et al.,
2013), and (ii) coating of NOM dispersing CNTs, thus
exposing more adsorption sites for SOC (Pan et al., 2013). Ac-
cording to Hou et al. (2013), though humic acid coated
MWCNTs were better dispersed in water forming a loosely
coiled network of tubes, the coverage of MWCNT adsorption
sites reduced the adsorption affinity of phenanthrene, 2-
naphtol and 1-naphtylamine noticeably. The reduction was
attributed to the decrease in surface area due to effective
humic acid coating of individual tubes. Chen et al. (2008a)
reported a moderate reduction of naphthalene, 1,3-
dinitrobenzene and 1,3,5-trinitrobenzene adsorption from
humic acid competition. In addition to the direct competition
for sorption sites, molecular sieving in micropores due to
steric hindrance was proposed because the suppression of
SOC adsorption was proportional to molecular sizes of tested
SOCs. Wang et al. (2008) explored the adsorption of phenan-
threne, naphthalene and 1-naphtol by MWCNTs coated with
humic acid, a-phenylalanine and peptone. Peptone coating
substantially reduced the surface area (from87 to 35m2/g) and
shifted the pore size distribution from micro- to meso- and
macro-ranges, indicating that the peptone coating was
depleting adsorption sites and blocking micropores. The au-
thors also reported the introduction of polar moieties to CNT
surfaces through peptone adsorption further reduced the
adsorption of hydrophobic contaminants. On the other hand,
the humic acid coating showed a much lower suppression of
adsorption, which was attributed to the negatively charged
polar functionalities of humic-acid dispersing MWCNTs that
increased the effective surface area. According to Pan et al.
(2013), humic acid suspended CNTs exhibited up to two or-
ders of magnitude greater adsorption capacity for
sulfamethoxazole than aggregated CNTs. The formation of a
stable CNT suspension in tannic acid solution was also pre-
viously demonstrated by Lin and Xing (2008b). The increasing
dispersion in water leading to higher surface area is likely to
counterbalance the depletion of surface area due to humic
acid coating. Another study by the same group (Wang et al.,
2009) entailed an investigation of the influence of humic
acid concentrations in competition with SOCs. At higher NOM
concentrations, the competitionwas less pronounced because
CNTs have limited number of high-energy adsorption sites
and once these high-energy sorption sites are depleted, the
competition for low-energy sorption sites was lower. Zhang
et al. (2010c) also reported competition for high-energy sorp-
tion sites that was indicated by increasing surface heteroge-
neity represented by Freundlich n values. The study also
showed a greater NOM-to-SOC ratio and a longer contact-time
that reduced the SOC adsorption capacity. The comparison of
simultaneous SOC andNOMadsorptionwith the preloading of
CNTs with NOM indicated that NOM preloading further
decreased the adsorption capacities, suggesting a slower
adsorption rate of NOM than SOCs. Wang et al. (2014) tested
competitive adsorption of dialkyl phthalate esters and humic
acid under various loading sequences such as simultaneous
loading of humic acid and SOCs, preloading of CNTs with
humic acid, preloading of CNTs with SOCs and pre-
equilibration of SOCs and humic acid prior to injection. Au-
thors observed better SOC adsorption under simultaneous
injection due to faster occupation of high energy sorption sites
by smaller and more hydrophobic SOCs than humic acid.
However, the loading sequences did not make an “obvious”
difference in the competitive sorption indicating that the
system containing SOC, CNT and NOM would equilibrate
under ambient conditions. However, more compounds with
different structural and chemical properties and different
NOM should be tested to validate this finding.
The pH of the background solution is another major factor
controlling adsorption. For organic acids, if pH < pKa, the non-
dissociated species for organic bases dominate the solution
and vice versa. Therefore, the influence of background solution
pH and ionic strength depends upon the ionizability and the
electron donor acceptor ability of SOCs. The pH change also
influences the protonation/deprotonation state of the func-
tional groups on CNT surfaces. Deprotonation of acidic func-
tional groups may increase the density of negatively charged
functional groups that may create repulsive forces between
negatively charged SOCs or may promote p-electron donor
ability of CNT surface and enhance pep electron donor in-
teractions between CNTs and SOCs. The formation of water
clusters decreasing hydrophobicity and reduction of hydrogen
bond formation decreasing adsorption affinity are other
possiblemechanisms for this increase in either repulsive forces
or the promotion of electron donor ability (Zhang et al., 2010c;
Pan and Xing, 2008). In their comparison of the adsorption of
nonionic phenanthrene and ionizable 2-phenylphenol byCNTs
under varying pH values, Zhang et al. (2010c) found that if the
background solution had pH values ranging from 4 to 10,
phenanthrene adsorption remained unaffected. However,
there was an observable decrease in adsorption of ionizable 2-
phenylphenol when the pH of the solution was over the pKa
of 2-phenylphenol. This decrease was attributed to the
wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 3 4e5 546
increased ionization and decreased hydrophobicity. Lin and
Xing (2008a) tested adsorption of three phenolic adsorbates
(phenol, pyrogallol and 1-naphtol) and one apolar adsorbate
(naphthalene) over a pH range of 2.2e11. The decrease of
adsorption affinities for phenolic compounds over their pKa
values were attributed to the increased electrostatic repulsion
between dissociated adsorbates and negatively charged
MWCNTs and the increase in hydrophilicities. In addition, the
authors also noted that the dissociation of eOH groups on
phenolic compounds might be inhibiting the formation of
hydrogen bonding. The adsorption affinities of phenolic com-
pounds increased up to their pKa values (from pH 2e6), indi-
cating the p-electron donating properties of these compounds
were altered. The change in adsorption affinitywith increasing
pH was not attributed to changing CNT properties because the
adsorption affinity of nonionic naphthalene to the same CNTs
remained constant in the same pH range. Similar results were
demonstrated byChenet al. (2008b) in thepH range of 3e11. For
two nonionic aromatics (1,3-dichlorobenzene and naphtha-
lene) adsorption was minimally affected. However, 2,4-
dichlorophenol showed a significant decrease above the com-
pound's pKa value. The authors also reported an increase in the
adsorption capacity of 2-naphtol above its pKa value due to
adsorption enhancing interactions counterbalancing the
decrease in hydrophobicity. It should be noted that single point
adsorption data were presented at pH 11 and the increase of
adsorption was relatively small. In another study, Chen et al.
(2007) tested adsorption of nonionic 1,2,4-trichlorobenzene
and ionic 2,4-dinitrotoluene by varying the pH from 2.8 to 7.4.
Nonionic adsorbate was independent of pH. Ionic 2,4-
dichlorotoluene showed a very slight increase as the pH
increased. Li et al. (2011) reported a decrease in adsorption of
perfluorooctanic acid (pKa ¼ �0.5) onto MWCNTs with
increasingpHfrom2to10.Thiswasattributed to the increaseof
electrostatic repulsion between deprotonated perfluorooctanic
acid and negatively charged functional groups of MWCNT at
elevatedpHvalues. In thesamestudy,Li et al. (2011) reportedan
increase in adsorption of 4-nitrophenol (pKa ¼ 10.7) with
increasing pH. In the pH range tested (2e10) 4-nitrophenol was
protonatedhence the increasewas attributed to the increase in
hydrogen bonding between the eOH group of 4-nitrophenol
and oxygen containing functional groups of MWCNTs
inducing strongerattraction. Similarly, Smithetal. (2012) tested
the adsorptionofNOMontopristine (2.2%oxygen) andoxidized
(8.1% oxygen)MWCNTs at pH 4,7, 6.0 and 9.0. The adsorption of
NOM by pristine MWCNT remained largely unaffected by pH,
however oxidized MWCNTs showed a remarkable decrease
with increasing pH. This decrease was also attributed to in-
crease of electrostatic repulsion. Yao and Strauss (1992)
observed an increasing attraction between cationic dyes and
negatively charged CNT surfaces as the pH of the solution was
increased. Among all tested ionizable compounds, the
adsorption of dissociated compounds was remarkably lower
than their non-dissociated forms.
The variance of the ionic strength of natural waters can be
another factor thatmay influence adsorption of SOCs. Organic
compounds are less soluble in aqueous salt solutions, which
are known as the salting-out effect. Salting-out may enhance
the hydrophobic interactions of SOCs with CNTs (Zhang et al.,
2010c; Xie et al., 1997). Only a limited number of studies have
been undertaken to determine the influence of ionic strength
on SOC adsorption by CNTs. Further research is needed to
examine the influence of ionic strength on CNT adsorption.
Kuo et al. (2008) reported the aggregation of dye molecules at
higher salt ion concentrations and the promotion of adsorp-
tion. Ions may penetrate into the diffuse double layer and
eliminate the repulsive energy between CNTs, however,
which in turn forms a more compact aggregation structure
that is unfavorable for SOC adsorption, which is known as the
squeezing-out effect (Zhang et al., 2010c). According to Zhang
et al. (2010c), the ionic strength (in the range of 0.001e0.1 M)
had negligible impact on adsorption of phenanthrene and 2-
phenylphenol on CNTs because the contribution of salting-
out effect was equivalent to that of the squeezing-out from
CNTs due to tighter aggregation or both the salting-out effect
and squeezing-out effect were too weak to exert any changes.
Chen et al. (2008a) also observed negligible influence of ionic
strength on SOC adsorption, which they attributed to a rela-
tively narrow range of ionic strength (0.02e0.1 M). Clearly,
tested ionic strengths exhibited no significant difference on
nonionic SOCs, which have low electronic coordination abili-
ties. Kuo et al. (2008) reported the aggregation of dye mole-
cules at higher salt ion concentrations that promoted
adsorption. In addition to ionic strength, the presence ofmetal
ions has also been the subject of separate investigations.
Wang et al. (2007) reported a significant increase in fulvic acid
adsorption on CNTs with increasing Ca2þ/Mg2þ ions. They
attributed this increase to (i) the compression of the diffuse
double layer and/or charge neutralization that decreases the
repulsive forces between NOM and CNTs; (ii) bridging of cat-
ions with negatively charged functional groups of NOM mol-
ecules; and (iii) bridging of cations between the NOM and
functional groups of the CNT surface. Similarly, Chen et al.
(2008a) found that the presence of 50 mg/l of Cu2þ reduced
adsorption of SOCs up to 20%. This reductionwas attributed to
the complexation of metal ions with surface oxygen func-
tionalities and formation of hydration shells of dense water
that competes with SOCs. Chen et al. (2009a) observed the
suppression of 2,4,6-trichlorophenol adsorption by MWCNTs
in the presence of 6.5e65 mg/L Cu2þ especially for oxidized
MWCNTs. In a separate study, Chen et al. (2008c) reported a
reduction in adsorption capacities in the presence of copper,
lead and cadmium, which they attributed to the large hydra-
tion spheres around the copper complexes. Additionally, the
cross-bridging effect of cations between the anionic func-
tional groups of CNT surfaces was suggested to form tighter
CNT bundles, which in turn shielded the sorption sites.
Adsorption of SOCs by CNTs is predominantly a
temperature-dependent process in which physical sorption
occursmostly as an exothermic process releasing energy. Shen
et al. (2009), reported negative DH values for adsorption of 1,3-
dinitrobenzene, m-nitrotoluene, p-nitrophenol and nitroben-
zene, suggesting that adsorption of nitroaromatics by
MWCNTs and oxidized MWCNTs is thermodynamically
spontaneous and exothermic. The reported absolute DH values
for adsorption of all were less than 40 kJ/mol implying physical
reaction (i.e. physisorption) except adsorption of p-nitrophenol
onto oxidizedMWCNTs (�42 kJ/mol). The higher enthalpy of p-
nitrophenol adsorption onto the oxidizedMWCNT surface was
attributed to chemical bonding via hydrogen bonding.
wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 3 4e5 5 47
However, it should be noted that DH > 40 kJ/mol is not a
definitive boundary between physisorption and chemisorp-
tion. Lu et al. (2006) reported negative DH values ranging from
�1.9 to �2.7 kJ/mol for adsorption of four trihalomethane
species (CHCl3, CHCl2Br, CHClBr2 and CHBr3) indicating
exothermic reaction. Authors also reported higher first order
rate constants for adsorption with increasing temperatures
and attributed the finding to the rise in the diffusion rate of
trihalomethanes at higher temperatures. Much smaller DH
values for trihalomethanes may be attributed to weaker
adsorption tendencies of trihalomethanes than hydrophobic
aromatic contaminants. Thermodynamic investigations for
the adsorption of nitroaromatics (Shen et al., 2009), trihalo-
methane (Lu et al., 2006), atrazine (Chen et al., 2009b; Yan et al.,
2008; Rambabu et al., 2012), NOM (Su and Lu, 2007; Wang et al.,
2007), and TEX (toluene, ethylbenzene andm-xylene) (Yu et al.,
2012) revealed that the CNT adsorption capacity decreases as
the temperature increases. Su and Lu (2007) observed faster
NOM adsorption kinetics when they increased the tempera-
ture, which in turn increased the diffusion rate. On the other
hand, the adsorption of 1,2-dichlorobenzene (Peng et al., 2003)
and pentachlorophenol (Salam and Burk, 2008), dyes (Kuo
et al., 2008; Wu, 2007b; Ghaedi et al., 2012a; Rodriguez et al.,
2010), methylene blue (Shahryari et al., 2010) exhibited
product-favored (endothermic) reactions, which again resulted
in an increase in adsorption capacity with an increase in
temperature. The commonality of all reported endothermic
sorption behavior was the potential electrostatic attraction
between the surface and the adsorbates, indicating that the
adsorption thermodynamics depends upon the nature of the
predominant sorption mechanism.
5. Predictive models for adsorption of SOCsby CNTs
Even many adsorption studies have been undertaken, these
cover only small portion of the approximately 40,000 anthro-
pogenic pollutants that are known to us (Nirmalakhandan and
Speece, 1990). While it is possible to expand conventional
adsorption studies on CNTs to amass data, these adsorption
isotherm experiments are time consuming, costly and labo-
rious. Therefore, predictive models are useful in rapidly
gathering adsorption data and more thoroughly elucidate the
mechanism of SOC adsorption on CNTs.
Of the many studies undertaken to predict the adsorption
of contaminants on activated carbon, some involved the use
of physical properties (e.g. molecular refraction, aromaticity,
parachor and the number of hydrophilic functional groups) to
predict the adsorption of chemicals on activated carbon (Abe
et al., 1983, 1981a, 1981b). Fundamental thermodynamic con-
cepts have also been used utilized to explain the adsorption
phenomena, the most ubiquitous of these being the Polanyi
Theory (proposed by Manes, 1978), the Net Adsorption Energy
Concept (proposed by Suffet and McGuire, 1978) and the Sol-
vophobic Approach (by Belfort, 1980).
Similar to those methods, the quantitative structur-
eeactivity relationship (QSAR) is a statistical model develop-
ment tool that has been employed to predict the adsorption of
organic chemicals by activated carbon (Blum et al., 1994). In
QSAR modeling, chemical properties are related to molecular
structures through molecular connectivity indices (MCI or c
index). MCIs are simple computational descriptors that do not
necessarily possess any chemical or physical meaning.
Another common predictive model development approach,
known as the Linear Solvation Energy Relationship (LSER),
involves the use of solvatochromic parameters predicting the
adsorption by activated carbon. In the LSER model, these pa-
rameters are used to relate chemical properties to solvation
energies such as cavity formation, dipolar interactions and
the hydrogen-bonding energies (Hickey and Passinoreader,
1991). First introduced by Kamlet et al. (1985), LSER has also
been used to predict the activated adsorption of carbon
(Luehrs et al., 1996; Shih and Gschwend, 2009; Dickenson and
Drewes, 2010). Solvatochromic descriptors contain chemical/
physical information about the organic molecules used to
explain the interactions between the adsorbate, the adsorbent
and the solvent by five independent descriptors. The LSER
model is expressed as follows:
log K ¼ aAþ bBþ vV þ pPþ rRþ c (1)
where A is the hydrogen bond donating ability, B is the
hydrogen bond accepting ability, V is the molecular volume or
McGowan's volume, P is the polarizability/dipolarity, R is the
excessmolar refraction, c is the regression constant and the a,
b, v, p and r are the regression coefficients.
There are several statistical predictive CNT models with
dependent and independent variables, all of which are listed
in Table 2. Yang et al. (2006b) presented single parameter
linear relationships between the adsorbed volume capacity
and the specific surface area, which is the monolayer N2
adsorption volume capacity, also known as the micropore
volume of CNTs. In all of their generated equations for five
CNTs, all coefficients of determination (r2) were above 0.99.
Similarly in their correlations between adsorption coefficients
with specific surface areas, mesopore volume, and micropore
volume of five CNTs, Oleszczuk et al. (2009) obtained a r2 > 0.96
coefficient. Yang et al. (2008) reported a single parameter
linear relationship between the hydrogen bond that donated
properties of anilines and phenols, with the rest of the pa-
rameters statistically insignificant at a ¼ 0.95. In addition, the
effort of a stepwise variable selection technique provided the
P term to the LSER equation, indicating that like hydrogen
bonding, p-electron polarizability is necessary for adsorption.
Yang and Xing (2010) reviewed Polanyi theory based Dubinin
Ashtakov model to estimate adsorption affinity and capacity
using empirical information, and they obtained a linear rela-
tionship between the E term also known as “the correlating
divisor”, that captures the adsorption energy of all responsible
adsorption forces and the b term, which is a fitting parameter.
Xia et al. (2010b) developed an LSER equation to depict
relative contributions of molecular interactions on CNT
adsorption, which (Eq. (2)) is expressed as follows:
log K ¼ �0:37A� 2:78Bþ 4:18V þ 1:75Pþ 0:043R� 1:33�n
¼ 28; r2 ¼ 0:93�
(2)
The authors employed a mixture of 28 aromatic SOCs
assuming that the SOC concentrations are low enough to
Table 2 e Summary of correlative equations between CNT adsorption descriptors and independent variables.
Source Dependent variable Independent variables Numberof data
points e n
Coefficient ofdetermination e r2
Notes
Yang et al. (2006b) PMM adsorption capacity descriptor-(Q0) Specific surface area (SSA) 5 0.99 Phenantherene on 5 different MWCNTs
Yang et al. (2006b) PMM adsorption capacity descriptor-(Q0) Micropore volume (PVmicro) 5 0.99 Phenantherene on 5 different MWCNTs
Oleszczuk et al. (2009) Adsorption coefficient (log K at Ce ¼ 0.01 Cs) Specific surface area (SSA) 5 0.97 Two aromatic pharmaceuticals on 5 different
MWCNTs
Oleszczuk et al. (2009) Adsorption coefficient (log K at Ce ¼ 0.01 Cs) Micropore volume (PVmicro) 5 0.97 Two aromatic pharmaceuticals on 5 different
MWCNTs
Oleszczuk et al. (2009) Adsorption coefficient (log K at Ce ¼ 0.01 Cs) Mesopore volume (PVmeso) 5 0.99 Two aromatic pharmaceuticals on 5 different
MWCNTs
Yang et al. (2008) Dubinin-Ashtakhov “correlating divisor” (E) Acidity (hydrogen bond
donating ability e a)
13 0.68 13 anilines/phenolic compounds on MWNCT; single
parameter correlation
Yang et al. (2008) Dubinin-Ashtakhov “correlating divisor” (E) Acidity and polarizability
(hydrogen bond donating
ability e a and polarizability e p)
13 0.68 13 aniline and phenolic compounds onMWNCT; poly
parameter correlation using stepwise parameter
selection indicated two significant parameters at 95%
level of significance
Yang and Xing (2010) Dubinin-Ashtakhov “correlating divisor” (E) b e Dubinin-Ashtakhov
fitting coefficient
39 0.62 Dubinin-Ashtakhov fitting coefficient e b was
speculated to contain information regarding the
adsorption affinity capturing sorption interactions
Xia et al. (2010b) Adsorption constant (k ¼ Cad/Ce) LSER solvatochromic
parameters (a, b, V and p)
28 0.93 Mixture of 28 aromatic compounds on MWCNT, poly
parameter correlation indicated V and B as the most
predominant contributors
Apul et al. (2013a) Adsorption descriptors (K∞ at Ce ¼ 0.002 Cs;
K0.1 at Ce ¼ 0.01 Cs; K0.01 at Ce ¼ 0.1 Cs)
QSAR parameters (c indices) 29 0.88 29 aromatic compounds were compiled from
different sources all adsorbed by MWCNTs; three
significant c indices were influential
Apul et al. (2013a) Adsorption descriptors (K∞ at Ce ¼ 0.002 Cs;
K0.1 at Ce ¼ 0.01 Cs; K0.01 at Ce ¼ 0.1 Cs)
LSER solvatochromic
parameters (a, b, V and p)
29 0.83 29 aromatic compounds were compiled from
different sources all adsorbed by MWCNTs; the V
term was the sole significant parameter with
increasing concentration A and B terms that were
also significant.
water
research
68
(2015)34e55
48
wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 3 4e5 5 49
prevent compoundecompound interactions. According to
their LSER derivative (Eq. (2)) the strongest contributor to
adsorption is hydrophobicity, which is represented by the V
term. The second most predominant factor is the hydrogen-
bond donating ability with a negative correlation, which in-
dicates an increase in the adsorbateewater interactions with
an increase in B values.
In an earlier study, we compiled MWCNT adsorption data
on 29 aromatic SOCs from the literature, including some
experimental results obtained in our laboratory (Apul et al.,
2013a) to develop predictive models using QSAR and LSER
techniques. The obtained QSAR and LSER equations are pre-
sented in equations (3) and (4), respectively (Apul et al., 2013a):
log K∞ ¼��2:98±0:52
�þ�0:18±0:12
�0cþ
�0:17±0:15
�0cv
þ�0:55±0:20
�3cp
�n ¼ 29; r2 ¼ 0:88
� (3)
log K∞ ¼� �4:34±0:56
�þ �0:05±0:32
�A� �
0:48±0:86�B
þ �4:55±0:56
�V þ �
0:61±0:34�P
�n ¼ 29; r2 ¼ 0:83
�
(4)
The molecular volume term (V) of the LSER model was the
most influential descriptor controlling adsorption at all con-
centrations, and the molecular size dependency was due to
the increasing hydrophobicity and increasing non-specific
attraction of larger molecules. Through this study we deter-
mined that QSAR and LSER techniques were effective in
developing effective models for predicting the adsorption of
organic compounds by CNTs. We validated our models were
externally using the adsorption data developed Xia et al.
(2010b). Experimentally measured CNT adsorption de-
scriptors were plotted against the predicted values by QSAR
and LSER modeling approaches, the results of which are pro-
vided in Fig. 2.
Our survey of predictive modeling efforts indicated that
both CNT (SSA, PV) and SOC properties (solvatochromic pa-
rameters) can be used to correlate CNT adsorption of SOCs.
Both single parameter and poly-parameter linear regression
studies were reported. Poly-parameter linear regression
studies were either conducted by employing a predefined set
of independent variables (solvatochromic parameters) or
through the use of a parameter selection technique (e.g.
stepwise parameter selection (Yang et al., 2008) or least
Fig. 2 e Experimentally measured adsorption descriptors vs. pre
based on Eq. (3)); and LSER models (to the right based on Eq. (4
2013a).
absolute shrinkage and selection operator) (Apul et al., 2013a).
The increasing number of data points decreased the linearity
of correlations (smaller r2) which reflected the ease fitting a
straight line to a fewer number of data points. As such, a
higher coefficient of determination (r2) need not necessarily
indicate a successful predictive model. An external validation
may be required to enhance the reliability and accuracy of the
model, however. In that modeling studies have only been
undertaken to elucidate the adsorption of aromatic organics
by MWCNTs, further research is required to model various
classes of organics and other CNT types, most particularly
SWCNTs.
6. Comparison of carbon nanotubes withactivated carbon
Both activated carbons and CNTs consist of graphitic carbon
backbone. They possess similar chemical properties (highly
hydrophobic, delocalized p-electron rich surfaces etc…),
while the pore structures of activated carbons differ greatly
from carbon nanotubes (Zhang et al., 2010a;Wang et al., 2014).
Activated carbons have a rigid pore structure with a network
of micro- and meso- and macropores where the majority of
adsorption surface area is located. On the other hand, indi-
vidual CNTs form random aggregates in water, driven by hy-
drophobic repulsion, non-specific attractive forces (van der
Waals) and pep stacking creating an ill-defined pore compo-
sition. The aggregation can also be influenced from the sur-
face chemistry of CNTs, pH, ionic strength and the presence of
NOM. For CNTs, four possible adsorption sites are external
surface of the outmost CNTs, peripheral grooves at the
boundary of CNT bundles, interstitial pore spaces between the
tubes, and inner cavities of open-ended CNTs (Zhang et al.,
2009).
Carbon nanotubes were reported to exhibit higher or
comparable adsorption capacity than activated carbons in
several studies (Ji et al., 2009a; Su and Lu, 2007; Lu et al., 2006;
Long and Yang, 2001; Pan et al., 2008; Apul et al., 2013b) while
opposite findings were reported in others (Lu et al., 2006; Ji
et al., 2009c). Ji et al. (2009c) reported higher adsorption ca-
pacity of a SWCNT and aMWCNT than an activated carbon for
tetracycline, whereas an opposite observation was made for
dicted adsorption descriptors from QSAR models (to the left
)) for training and external validation datasets (Apul et al.,
wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 3 4e5 550
naphthalene in the same study. This was attributed to the
bulky tetracycline molecules being subject to size exclusion
resulting in less adsorption by the microporous activated
carbon. Similarly, Lu et al. (2006) reported higher CHCl3 uptake
of a MWCNT than a powdered activated carbons, despite the
lower specific surface area of the former (295 m2/g) than the
latter (900 m2/g). The authors speculated that functional
groups on the CNT surface made it more favorable for the
adsorption of polar CHCl3 molecules. However, this could be
also attributed to better dispersion of functionalized CNTs and
exposing more sorption sites to the CHCl3 molecules. Su and
Lu (2007) reported higher sorption affinity of NOM molecules
by CNTs than a granular activated carbon. Long and Yang also
reported superior sorption of dioxin to a MWCNT than an
activated carbon. However, in these two reports, detailed
structural information about the adsorbents has not
been provided to assess the factors controlling the
differences in the adsorption results of activated carbons and
CNTs. Wang et al. (2014) tested adsorption of three aromatic
(4-chloronitrobenzene, 1,4-dichlorobenzene and 1,2-
dichlorobenzene) and one nonplanar aliphatic (trans-1,2-
dichlorocyclohexane) SOCs by a MWCNT, a SWCNT and
two activated carbons. Aromatic SOCs showed better
adsorption capacities onto CNTs, which was attributed to the
pep stacking with benzene rings of the aromatics and the
higher graphitized surfaces of CNTs than activated carbons.
On the other hand, better adsorption of trans-1,2-
dichlorocyclohexane onto activated carbons than CNTs in
the same study was attributed to better micropore filling. Ji
et al. (2009c) compared the adsorption of phenol, 1,3-
dichlorobenzene, 1,3-dinitrobenzene onto SWCNTs and acti-
vated carbons. Adsorption capacities were roughly correlated
with the surface areas of adsorbents. The better sorption ca-
pacity of activated carbons was attributed to the micropore
filling mechanism. Pore filling mechanism may enhance the
adsorption affinity of activated carbons for SOCs especially
when the pore and adsorbate molecule sizes are comparable.
Adsorption kinetics on CNTs have been reported to be faster
than on activated carbons (Zhang et al., 2012), although some
studies suggest that more studies are needed examining the
roles of adsorbate properties (molecular size, structure) on the
adsorption kinetics by CNTs and activated carbons. Further-
more, it has been shown that although in distilled and
deionized water, the adsorption capacity of a granular acti-
vated carbon was higher than a SWCNT and MWCNT, in the
presence of NOM, the SWCNT showed similar adsorption ca-
pacities to those of the granular activated carbon (Zhang et al.,
2012). It was also noted in the same study that NOM compe-
tition was more significant on the adsorption of a nonplanar
and hydrophilic SOC than a planar and hydrophobic SOC
indicating also the importance of adsorbate characteristics.
7. Future research needs
“System heterogeneity” plays an important role on the
adsorption process and it is manifested in several ways. The
organic (e.g., SOC and NOM) and inorganic (e.g., pH and ionic
strength) characteristics of each individual water may vary
substantially. Moreover, the inherent characteristics (e.g.,
type, dimensions, and functional groups) and aggregation
state of CNTs in water may vary greatly depending on their
preparation and water matrix. Each of these variables com-
plicates the development of universal adsorption models,
necessitating their individual consideration for further un-
derstanding and research. The greatest challenge remains to
understand andmodel adsorption of SOCs by CNTs accurately
in heterogeneous and complex systems.
The quantification of individual contributions to overall
adsorption still remains as a challenging task. Examining
adsorption of structurally related small molecular weight
SOCs (e.g., similar molecular structures with different mo-
lecular weights, isomers with different molecular configura-
tions, varying Lewis acidebase properties) may provide
further insights to the significance of different intermolecular
adsorption interactions. Testing SOC adsorption in nonpolar
organic solvents such as hexadecane may be useful to
examine intermolecular interactions in the absence of hy-
drophobicity effect. Understanding the influence and role of
heterogeneous CNT characteristics (type, dimensions, and
surface chemistry) in the SOC adsorption interactions needs
further systematic investigations. Another challengewill be to
use this knowledge in the generation of improved or new
(mechanistic and/or statistical) adsorption models including
the contribution of the individual interactions to overall
adsorption. Developing models that successfully predict the
adsorption of some classes of SOCs (e.g., aromatic and
aliphatic) on some types of CNTs (e.g., SWCNTs andMWCNTs)
will still present important advances. Furthermore, the ability
of such models predicting adsorption in natural waters in the
presence of NOM and complex backgroundwatermatrices is a
major challenge due to site specific, heterogeneous, water
chemistry (pH, calcium/ionic strength levels) dependent na-
ture of NOM components (e.g., hydrophobic/hydrophilic, low/
highmolecular weight, aromatic/aliphatic) thatmay influence
both CNT aggregation and SOC adsorption. A fundamental
understanding of SOC adsorption kinetics and desorption
behaviors in such complex systems is equally important, and
have been much less investigated as compared to adsorption
equilibrium in literature.
Detailed characterization of CNTs, especially after pre-
treatment and/or functionalization, should be included with
the research articles (e.g., in the supplementary information
sections of most journals). Furthermore, providing the
adsorption data is also important. In the “Big Data” age and
initiatives such “Materials Genome Initiative” (Material
Genome Initiativ, 2014), there is an abundance of data pro-
duction including adsorption of SOCs by CNTs, therefore
detailed reporting of the data in peer-reviewed articles,
especially in the electronic domains, can facilitate the data
mining, in-depth analysis and the development of advanced
adsorption models. There are more data available for
adsorption of SOCs byMWCNT than SWCNTs in the literature.
Therefore, developing adsorption data on SWCNTs can still be
useful. The SSA and PV have been influential for the adsorp-
tion of hydrophobic SOCs, but they were not the exclusive
factors determining the adsorption capacities. The SSA and PV
of CNTs are determined with N2 gas adsorption in their bulk
phases. Given the aggregating nature of CNTs in water, it is
important to develop novel techniques to characterize the
wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 3 4e5 5 51
accessible SSA and PV for adsorption on CNT aggregates in
water. The impact of CNT characteristics on its aggregation
behavior and impact on SOC adsorption need further
investigation.
8. Conclusions
Over the last ten years, more than 80 articles have been pub-
lished in peer-reviewed journals on adsorption of synthetic
organic contaminants (SOCs) (including polycyclic aromatic
hydrocarbons, benzene derivatives, phenolic compounds,
pharmaceuticals, insecticides, herbicides, organic dyes and
polychlorinated biphenyls) by CNTs. In these articles,
adsorption of SOCs was investigated using 18 single-walled
CNTs and 81 multi-walled CNTs as adsorbents. The abun-
dance of multi-walled CNT in these studies was likely due to
the higher market prices of single-walled CNTs. The average
market price formulti-walled CNTswas approximately 10 $/g,
and the average price for single-walled CNTs was higher than
100 $/g. Themean specific surface areas (±standard deviation)
of all reviewed single-walled CNTs and multi-walled CNTs
were 373 ± 146 m2/g and 216 ± 159 m2/g, respectively.
Both specific surface area and pore volume have been
found to be influential parameters for adsorption of hydro-
phobic SOCs, but they were not the exclusive factors deter-
mining adsorption capacities. Severalmechanisms, of varying
relative importance, have been proposed for controlling the
SOC adsorption by CNTs. The predominant driving force that
repels SOCs from water onto the CNT surface is hydrophobic
forces. Other principle SOC properties that positively in-
fluences adsorption is the p electron density of compounds
resulting in pep interaction with the p electron rich graphitic
surface of CNTs and polar moieties of SOCs enhancing dipo-
leedipole attraction. Size exclusionwas found to be important
especially for large organic macromolecules. In terms of CNT
surface chemistry, oxygen content was often investigated,
and two mechanisms were proposed influencing the SOC
adsorption: (i) the presence of oxygen on the CNT surface
making adsorption of water molecules energetically more
favorable relative to SOC adsorption, reducing the available
surface area for SOCs; and (ii) the presence of oxygen on CNT
surface localizes the p electrons, which reduces the pep in-
teractions between the CNT graphitic surface and benzene
rings of aromatic SOCs. On the other hand, the increase in
adsorption capacities as a result of surface oxidation was
attributed to better dispersion of CNTs, exposing more
adsorption sites, and increase in the ability to form hydrogen
bonds with some SOCs.
Natural organicmatter is ubiquitous in natural waters. The
presence of natural organic matter, in general, has been re-
ported to decrease the adsorption of SOCs by CNT. However,
natural organic matter coatings may disperse CNTs. revealing
more sorption sites for SOC adsorption.
The influences of pH, temperature, and ionic strength have
been examined less frequently. In general, pH and ionic
strengths exhibited no significant difference on nonionic
SOCs, which have low electronic coordination abilities. Some
dependence to pH and ionic strength were observed for
adsorption of ionizable SOCs (e.g. phenolic compounds,
anilines etc.) especially onto CNTs with ionizable functional
groups. Two counter-balancing forces influencing adsorption
were discussed; (i) salting-out effect: the decrease of SOC
solubility due to ionic strength that is favorable for adsorption
and (ii) squeezing-out effect: elimination of repulsive energies
between CNTs, which in turn forms a more compact aggre-
gation structure that is unfavorable for SOC adsorption. In
terms of temperature effect, adsorption of SOCs by CNTs was
predominantly reported as an exothermic process (i.e.
releasing energy); therefore, increase in temperature
decreased adsorption capacity.
Even several adsorption studies have been undertaken in
the past decade, they cover only small portion of the
approximately 40,000 anthropogenic pollutants that are
known to us. Statistical SOC adsorptionmodels have emerged
in recent years to minimize time consuming, costly and
laborious adsorption isotherm experiments. Our survey of
predictive modeling efforts indicated that both CNT and SOC
properties (solvatochromic parameters) can be used to corre-
late CNT adsorption of SOCs. Several single parameter and
poly-parameter linear regression studies were reported in the
literature. Poly-parameter linear regression studies were
either conducted by employing a predefined set of indepen-
dent variables (i.e. solvatochromic parameters) or through the
use of parameter selection techniques (e.g. stepwise param-
eter selection or least absolute shrinkage and selection
operator).
Overall, well-controlled laboratory studies have provided a
good insight to adsorption mechanisms of SOCs by CNTs, and
the greatest challenge still remains to further understand and
model adsorption of SOCs by CNTs accurately in heteroge-
neous and complex systems.
Acknowledgments
This review is a part of the work that was supported by a
research grant from the National Science Foundation (CBET
0967425). However the manuscript has not been subjected to
the peer and policy review of the agency and therefore does
not necessarily reflect its views.
Appendix A. Supplementary data
Supplementary data related to this article can be found at
http://dx.doi.org/10.1016/j.watres.2014.09.032.
r e f e r e n c e s
Abe, I., Hayashi, K., Kitagawa, M., 1981a. Prediction of adsorption-isotherms of organic-compounds from water on activatedcarbons. Bull. Chem. Soc. Jpn. 54 (9), 2819e2820.
Abe, I., Hayashi, K., Kitagawa, M., 1981b. Prediction of adsorption-isotherms of organic-compounds from water on activatedcarbons. 2. relative adsorbabilities of elements. Bull. Chem.Soc. Jpn. 54 (12), 3857e3858.
Abe, I., Hayashi, K., Kitagawa, M., Hirashima, T., 1983. Predictionof adsorbability of organic-compounds from aqueous-solution
wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 3 4e5 552
on activated carbon by means of the linear free-energyrelationship. Bull. Chem. Soc. Jpn. 56 (4), 1002e1005.
Agnihotri, S., Mota, J.P.B., Rostam-Abadi, M., Rood, M., 2006.Adsorption site analysis of impurity embedded single-walledcarbon nanotube bundles. Carbon 44 (12), 2376e2383.
Agnihotri, S., Kim, P., Zheng, Y., Mota, J.P., Yang, L., 2008.Regioselective competitive adsorption of water and organicvapor mixtures on pristine single-walled carbon nanotubebundles. Langmuir 24 (11), 5746e5754.
Ajayan, P.M., 1999. Nanotubes from carbon. Chem. Rev. 99 (7),1787e1799.
Ajmani, G.S., Cho, H.H., Chalew, T.E.A., Schwab, K.J.,Jacangelo, J.G., Huang, H., 2014. Static and dynamic removal ofaquatic natural organic matter by carbon nanotubes. WaterRes. 59, 262e270.
Apul, O.G., Shao, T., Zhang, S., Karanfil, T., 2012. Impact of carbonnanotube morphology on phenanthrene adsorption. Environ.Toxicol. Chem. 31 (1), 73e78.
Apul, O.G., Wang, Q., Shao, T., Rieck, J.R., Karanfil, T., 2013a.Predictive model development for adsorption of aromaticcontaminants by multi-walled carbon nanotubes. Environ. Sci.Technol. 47 (5), 2295e2303.
Apul, O.G., Wang, Q., Zhou, Y., Karanfil, T., 2013b. Adsorption ofaromatic contaminants by graphene nanosheets: comparisonwith carbon nanotubes and activated carbon. Water Res. 47(4), 1648e1654.
Babel, S., Kurniawan, T.A., 2003. Low-cost adsorbents for heavymetal uptake from contaminated water: a review. J. Hazard.Mater. B97, 219e243.
Beless, B., Rifai, H.S., Rodrigues, D.F., 2014. Efficacy ofcarbonaceous materials for sorbing polychlorinated biphenylsfrom aqueous solution. Environ. Sci. Technol. 48 (17),10372e10379.
Belfort, G., 1980. Selective adsorption of organic homologs ontoactivated carbon from dilute aqueous-solutions - solvophobicinteraction approach-II. Abstr. Pap. Am. Chem. Soc. 179 (MAR),54eENVR.
Blum, D.J.W., Suffet, I.H., Duguet, J.P., 1994. Quantitativestructure-activity relationship using molecular connectivityfor the activated carbon adsorption of organic-chemicals inwater. Water Res. 28 (3), 687e699.
Brooks, A.J., Lim, H.N., Kilduff, J.E., 2012. Adsorption uptake ofsynthetic organic chemicals by carbon nanotubes andactivated carbons. Nanotechnology 23 (29).
Cai, N., Larese-Casanova, P., 2014. Sorption of carbamazepine bycommercial graphene oxides: a comparative study withgranular activated carbon and multiwalled carbon nanotubes.J. Colloid Interface Sci. 42, 152e161.
Chen, W., Duan, L., Zhu, D., 2007. Adsorption of polar andnonpolar organic chemicals to carbon nanotubes. Environ. Sci.Technol. 41 (24), 8295e8300.
Chen, J., Chen, W., Zhu, D., 2008a. Adsorption of nonionicaromatic compounds to single-walled carbon nanotubes:effects of aqueous solution chemistry. Environ. Sci. Technol.42 (19), 7225e7230.
Chen, W., Duan, L., Wang, L., Zhu, D., 2008b. Adsorption ofhydroxyl- and amino-substituted aromatics to carbonmanotubes. Environ. Sci. Technol. 42 (18), 6862e6868.
Chen, G.C., Shan, X.Q., Wang, Y.S., Pei, Z.G., Shen, X.E., Wen, B.,Owens, G., 2008c. Effects of copper, lead, and cadmium on thesorption and desorption of atrazine onto and from carbonnanotubes. Environ. Sci. Technol. 42 (22), 8297e8302.
Chen, G.C., Shan, X.Q., Wang, Y.S., Wen, B., Pei, Z.G., Xie, Y.N.,Liu, T., Pignatello, J.J., 2009a. Adsorption of 2,4,6-trichlorophenol by multi-walled carbon nanotubes as affectedby Cu(II). Water Res. 43 (9), 2409e2418.
Chen, G.C., Shan, X.Q., Zhou, Y.Q., Shen, X., Huang, H.L.,Khan, S.U., 2009b. Adsorption kinetics, isotherms and
thermodynamics of atrazine on surface oxidized multiwalledcarbon nanotubes. J. Hazard. Mater. 169 (1e3), 912e918.
Chen, Z., Pierre, D., He, H., Tan, S., Pham-Huy, C., Honh, H.,Huang, J., 2011. Adsorption behavior of epirubicinhydrochloride on carboxylated carbon nanotubes. Int. J.Pharm. 405 (1e2), 153e161.
Chin, C.J.M., Shih, L.C., Tsai, H.J., Liu, T.K., 2007. Adsorption of o-xylene and p-xylene from water by SWCNTs. Carbon 45 (6),1254e1260.
Cho, H.-H., Smith, B.A., Wnuk, J.D., Fairbrother, D.H., Ball, W.P.,2008. Influence of surface oxides on the adsorption ofnaphthalene onto multiwalled carbon nanotubes. Environ.Sci. Technol. 42 (8), 2899e2905.
Coughlin, R.W., Ezra, F.S., 1968. Role of surface acidity in theadsorption of organic pollutants on the surface of carbon.Environ. Sci. Technol. 2 (4), 291e297.
Datsyuk, V., Kalyva, M., Papagelis, K., Parthenios, J., Tasis, D.,Siokou, A., Kallitsis, I., Galiotus, C., 2008. Chemical oxidationof multiwalled carbon nanotubes. Carbon 46 (6), 833e840.
Dickenson, E.R.V., Drewes, J.E., 2010. Quantitative structureproperty relationships for the adsorption of pharmaceuticalsonto activated carbon. Water Sci. Technol. 62 (10), 2270e2276.
Dobbs, R.A., Cohen, J.M., 1980. Carbon Adsorption Isotherms forToxic Organics. Municipal Environmental ResearchLaboratory, U.S. Environmental Protection Agency.
Ferguson, P.L., Chandler, G.T., Templeton, R.C., DeMarco, A.,Scrivens, W.A., Englehart, B.A., 2008. Influence of sediment-amendment with single-walled carbon nanotubes and dieselsoot on bioaccumulation of hydrophobic organiccontaminants by benthic invertebrates. Environ. Sci. Technol.42 (10), 3879e3885.
Fu, H., Yang, L., Wan, Y., Xu, Z., Zhu, D., 2011. Adsorption ofpharmaceuticals to microporous activated carbon treatedwith potassium hydroxide, carbon dioxide, and steam. J.Environ. Qual. 40 (6), 1886e1894.
Ghaedi, M., Hassanzadeh, A., Kokhdan, S.N., 2011. Multiwalledcarbon nanotubes as adsorbents for the kinetic andequilibrium study of the removal of alizarin red S and morin. J.Chem. Eng. Data 56 (5), 2511e2520.
Ghaedi, M., Haghdoust, S., Nasiri-Korkhdan, S., Mihandoost, A.,Sahraie, R., Daneshfar, A., 2012a. Comparison of activatedcarbon, multiwalled carbon nanotubes, and cadmiumhydroxide nanowire loaded on activated carbon as adsorbentsfor kinetic and equilibrium study of removal of safranine o.Spectrosc. Lett. 45 (7), 500e510.
Ghaedi, M., Khajehsharifi, H., Yadkuri, A.H., Roosta, M.,Asghari, A., 2012b. Oxidized multiwalled carbon nanotubes asefficient adsorbent for bromothymol blue. Toxicol. Environ.Chem. 94 (5), 873e883.
Giusti, D.M., Conway, R.A., Lawson, C.T., 1974. Activated carbonadsorption of petrochemicals. J. Water Pollut. Control Fed. 46(5), 947e965.
Gotovac, S., Hattori, Y., Noguchi, D., Miyamoto, J., Kanamaru, M.,Utsumi, S., Kanoh, H., Kaneko, K., 2006. Phenanthreneadsorption from solution on single wall carbon nanotubes. J.Phys. Chem. B 110 (33), 16219e16224.
Gotovac, S., Honda, H., Hattori, Y., Takahashi, K., Kanoh, H.,Kaneko, K., 2007a. Effect of nanoscale curvature of single-walled carbon nanotubes on adsorption of polycyclic aromatichydrocarbons. Nano Lett. 7 (3), 583e587.
Gotovac, S., Yang, C.M., Hattori, Y., Takahashi, K., Kanoh, H.,Kaneko, K., 2007b. Adsorption of polyaromatic hydrocarbonson single wall carbon nanotubes of different functionalitiesand diameters. J. Colloid Interface Sci. 314 (1), 18e24.
Gotovac, S., Song, L., Kanoh, H., Kaneko, K., 2007c. Assemblystructure control of single wall carbon nanotubes with liquidphase naphthalene adsorption. Colloids Surf. A Physicochem.Eng. Asp. 300 (1e2), 117e121.
wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 3 4e5 5 53
Gupta, V.K., Kumar, R., Nayak, A., Salah, T.A., Barakat, M.A., 2013.Adsorptive removal of dyes from aqueous solution ontocarbon nanotubes: a review. Adv. Colloid Interface Sci.193e194, 24e34.
Hickey, J.P., Passinoreader, D.R., 1991. Linear solvation energyrelationships - rules of thumb for estimation of variablevalues. Environ. Sci. Technol. 25 (10), 1753e1760.
Hou, L., Zhu, D., Wang, X., Wang, L., Zhang, C., Chen, W., 2013.Adsorption of phenanthrene, 2-naphthol, and 1-naphthylamine to colloidal oxidized multiwalled carbonnanotubes: effects of humic acid and surfactant modification.Environ. Toxicol. Chem. 32 (3), 493e500.
Iijima, S., 1991. Helical microtubules of graphitic carbon. Nature354 (6348), 56e58.
Ji, L., Chen, W., Duan, L., Zhu, D., 2009a. Mechanisms for strongadsorption of tetracycline to carbon nanotubes: a comparativestudy using activated carbon and graphite as adsorbents.Environ. Sci. Technol. 43 (7), 2322e2327.
Ji, L., Chen, W., Zheng, S., Xu, Z., Zhu, D., 2009b. Adsorption ofsulfonamide antibiotics to multiwalled carbon nanotubes.Langmuir 25 (19), 11608e11613.
Ji, L., Liu, F., Xu, Z., Zheng, S., Zhu, D., 2009c. Zeolite-templatedmicroporous carbon as a superior adsorbent for removal ofmonoaromatic compounds from aqueous solution. Environ.Sci. Technol. 43, 7870e7876.
Ji, L., Shao, Y., Xu, Z., Zheng, S., Zhu, D., 2010a. Adsorption ofmonoaromatic compounds and pharmaceutical antibiotics oncarbon nanotubes activated by KOH etching. Environ. Sci.Technol. 44 (16), 6429e6436.
Ji, L., Chen, W., Bi, J., Zheng, S., Xu, Z., Zhu, D., Alvarez, P.J., 2010b.Adsorption of tetracycline on single-walled and multi-walledcarbon nanotubes as affected by aqueous solution chemistry.Environ. Toxicol. Chem. 29 (12), 2713e2719.
Johnston, H.J., Hutchison, G.R., Christensen, F.M., Peters, S.,Hankin, S., Aschberger, A., Stone, V., 2010. Critical review ofthe biological mechanisms underlying the in vivo andin vitro toxicity of carbon nanotubes: the contribution ofphysico-chemical characteristics. Nanotoxicology 4 (2),207e246.
Kamlet, M.J., Doherty, R.M., Abraham, M.H., Taft, R.W., 1985.Linear solvation energy relationships .33. An analysis of thefactors that influence adsorption of organic-compounds onactivated carbon. Carbon 23 (5), 549e554.
Kato, Y., Machida, M., Tatsumoto, H., 2008. Inhibition ofnitrobenzene adsorption by water cluster formation at acidicoxygen functional groups on activated carbon. J. ColloidInterface Sci. 322 (2), 394e398.
Keller, A., McFerran, S., Lazareva, A., Suh, S., 2013. Global lifecycle releases of engineered nanomaterials. J. NanoparticlesRes. 15 (6), 1e17.
Khan, I.A., Afrooz, A.R.M.N., Flora, J.R.V., Schierz, P.A.,Ferguson, P.L., Sabo-Attwood, T., Saleh, N., 2013.Chirality affects aggregation kinetics on kinetics of single-walled carbon nanotubes. Environ. Sci. Technol. 47,1844e1852.
Kharlamova, M.V., Mochalin, V.N., Lukatskaya, M.R., Niu, J.,Presser, V., Mikhalovsky, S., Gogotsi, Y., 2013. Adsorption ofproteins in channels of carbon nanotubes: effect of surfacechemistry. Mater. Express 3 (1), 1e10.
Kilduff, J.E., Karanfil, T., Chin, Y.P., Weber, W.J., 1996. Adsorptionof natural organic polyelectrolytes by activated carbon: a size-exclusion chromatography study. Environ. Sci. Technol. 30 (4),1336e1343.
Klaine, S.J., Alvarez, P.J.J., Batley, G.E., Fernandes, T.F.,Handy, R.D., Lyon, D.Y., Mahendra, S., McLaughlin, M.J.,Lead, J.R., 2008. Nanomaterials in the environment: behavior,fate, bioavailability, and effects. Environ. Toxicol. Chem. 27 (9),1825e1851.
Kotel, L.Y., Brichka, A.V., Brichka, S.Y., 2009. Adsorptionproperties of modified multilayer carbon nanotubes withrespect to benzoic acid. Russ. J. Appl. Chem. 82 (4), 569e573.
Kuo, C.Y., Wu, C.H., Wu, J.Y., 2008. Adsorption of direct dyes fromaqueous solutions by carbon nanotubes: determination ofequilibrium, kinetics and thermodynamics parameters. J.Colloid Interface Sci. 327 (2), 308e315.
Kutics, K., Suzuki, M., 1993. Adsorption of organics on surface-treated activated carbon-fibers. Hung. J. Ind. Chem. 21 (2),93e99.
Lam, C.W., James, J.T., McCluskey, R., Arepalli, S., Hunter, R.L.,2006. A review of carbon nanotube toxicity and assessment ofpotential occupational and environmental health risks. Crit.Rev. Toxicol. 36 (3), 189e217.
Li, Y.H., Wang, S.G., Luan, Z.K., Ding, J., Xu, C.L., Wu, D.H., 2003.Adsorption of cadmium(II) from aqueous solution by surfaceoxidized carbon nanotubes. Carbon 41 (5), 1057e1062.
Li, Y.H., Di, Z., Ding, J., Wu, D., Luan, Z., Zhu, Y., 2005. Adsorptionthermodynamic, kinetic and desorption studies of Pb2þ oncarbon nanotubes. Water Res. 39 (4), 605e609.
Li, X., Zhao, H., Quan, X., Chen, S., Zhang, Y., Yu, H., 2011.Adsorption of ionizable organic contaminants on multi-walledcarbon nanotubes with different oxygen contents. J. Hazard.Mater. 186 (1), 407e415.
Li, Y., Du, Q., Liu, T., Peng, X., Wang, J., Sun, J., Wang, Y., Wu, S.,Wang, Z., Xia, Y., Xia, L., 2013a. Comparative study ofmethylene blue dye adsorption onto activated carbon,graphene oxide, and carbon nanotubes. Chem. Eng. Res. Des.91 (2), 361e368.
Li, Y., Niu, J., Shen, Z., Feng, C., 2013b. Size effect of single-walledcarbon nanotube on adsorption of perfluorooctanesulfonate.Chemosphere 784e790.
Li, M., Hsieh, T.C., Doong, R.A., Huang, C.P., 2013c. Tuning theadsorption capability of multi-walled carbon nanotubes topolar and non-polar organic compounds by surface oxidation.Sep. Purif. Technol. 117, 98e103.
Li, H., Zhang, D., Han, X., Xing, B., 2014. Adsorption of antibioticciprofloxacin on carbon nanotubes: pH dependence andthermodynamics. Chemosphere 95, 150e155.
Lin, D., Xing, B.S., 2008a. Adsorption of phenolic compounds bycarbon nanotubes: role of aromaticity and substitution ofhydroxyl groups. Environ. Sci. Technol. 42 (19), 7254e7259.
Lin, D., Xing, B.S., 2008b. Tannic acid adsorption and its role forstabilizing carbon nanotube suspensions. Environ. Sci.Technol. 42 (16), 5917e5923.
Liu, C.H., Li, J.J., Zhang, H.L., Li, B.R., Guo, Y., 2008. Structuredependent interaction between organic dyes and carbonnanotubes. Colloids Surf. A Physicochem. Eng. Asp. 313,9e12.
Liu, S., Wang, C., Su, W., Wangyan, L.V., Zhu, S., Wang, F., Fu, Q.,2014. Water dispersed multi-walled carbon nanotubesmodified by tannin acid. Mater. Lett. 123, 44e47.
Long, R.Q., Yang, R.T., 2001. Carbon nanotubes as superiorsorbent for dioxin removal. J. Am. Chem. Soc. 123 (9),2058e2059.
Lu, C., Chung, Y.L., Chang, K.F., 2006. Adsorption thermodynamicand kinetic studies of trihalomethanes on multiwalled carbonnanotubes. J. Hazard. Mater. 138 (2), 304e310.
Luehrs, D.C., Hickey, J.P., Nilsen, P.E., Godbole, K.A., Rogers, T.N.,1996. Linear solvation energy relationship of the limitingpartition coefficient of organic solutes between water andactivated carbon. Environ. Sci. Technol. 30 (1), 143e152.
Ma, X., Tsige, M., Uddin, S., Talapatra, S., 2011. Application ofcarbon nanotubes for removing organic contaminants fromwater. Mater. Express 1 (3), 183e200.
Machado, F.M., Bergmann, C.P., Fernandes, T.H., Lima, E.C.,Royer, B., Calvete, T., Fagan, S.B., 2011. Adsorption of ReactiveRed M-2BE dye from water solutions by multi-walled carbon
wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 3 4e5 554
nanotubes and activated carbon. J. Hazard. Mater. 192 (3),1122e1131.
Machado, F.M., Bergmann, C.P., Lima, E.C., Royer, B., deSouza, F.E., Jauris, I.M., Calvete, T., Fagan, S.B., 2012.Adsorption of reactive blue 4 dye from water solutions bycarbon nanotubes: experiment and theory. Phys. Chem.Chem. Phys. 14 (31), 11139e11153.
Manes, M., 1978. Polanyi adsorption potential theory and itsapplication to liquid-phase adsorption. Abstr. Pap. Am. Chem.Soc. 176, 4.
Material Genome Initiative. White House, 2014. www.whitehouse.gov/mgi
Mauter, M.S., Elimelech, M., 2008. Environmental applications ofcarbon-based nanomaterials. Environ. Sci. Technol. 42 (16),5843e5859.
Moreno-Castilla, C., 2004. Adsorption of organic molecules fromaqueous solutions on carbon materials. Carbon 42 (1), 83e94.
Nirmalakhandan, N., Speece, R., 1990. Adsorption fromaqueous phase by activated carbon - a simplified applicationof the solvophobic theory. Environ. Sci. Technol. 24 (4),575e580.
Oleszczuk, P., Xing, B., 2011. Influence of anionic, cationic andnonionic surfactants on adsorption and desorption ofoxytetracycline by ultrasonically treated and non-treatedmultiwalled carbon nanotubes. Chemosphere 85 (8),1312e1317.
Oleszczuk, P., Pan, B., Xing, B.S., 2009. Adsorption and desorptionof oxytetracycline and carbamazepine by multiwalled carbonnanotubes. Environ. Sci. Technol. 43 (24), 9167e9173.
Pan, B., Xing, B.S., 2008. Adsorption mechanisms of organicchemicals on carbon nanotubes. Environ. Sci. Technol. 42 (24),9005e9013.
Pan, B., Lin, D., Mashayekhi, H., Xing, B., 2008. Adsorption andhysteresis of bisphenol A and 17a-ethinyl estradiol on carbonnanomaterials. Environ. Sci. Technol. 42, 5480e5485.
Pan, B., Zhang, D., Li, H., Wu, M., Wang, Z., Xing, B.S., 2013.Increased adsorption of sulfamethoxazole on suspendedcarbon nanotubes by dissolved humic acid. Environ. Sci.Technol. 47 (14), 7722e7728.
Peigney, A., Laurent, C.H., Flahaut, E., Bacsa, R.R., Rousset, A.,2001. Specific surface area of carbon nanotubes and bundles ofcarbon nanotubes. Carbon 39 (4), 507e514.
Peng, X.J., Li, Y., Luan, Z., Di, Z., Wang, H., Tian, B., Jia, Z., 2003.Adsorption of 1,2-dichlorobenzene from water to carbonnanotubes. Chem. Phys. Lett. 376 (1e2), 154e158.
Peng, X., Jia, J., Luan, Z., 2009. Oxidized carbon nanotubes forsimultaneous removal of endrin and Cd(II) from water andtheir separation from water. J. Chem. Technol. Biotechnol. 84(2), 275e278.
Peng, H., Pan, B., Wu, M., Liu, Y., Zhang, D., Xing, B.S., 2012.Adsorption of ofloxacin and norfloxacin on carbon nanotubes:hydrophobicity- and structure-controlled process. J. Hazard.Mater. 233, 89e96.
Petosa, A.R., Jaisi, D.P., Quevedo, I.R., Elimelech, M., Tujenkji, N.,2010. Aggregation and deposition of engineerednanomaterials in aquatic environments: role ofphysicochemical interactions. Environ. Sci. Technol. 44 (17),6532e6549.
Powell, M.C., Kanarek, M.S., 2006. Nanomaterial healtheffectsepart 1: background and current knowledge. WMJ Off.Publ. State Med. Soc. Wis. 105 (2), 16e20.
Pyrzynska, K., Stafiej, A., Biesaga, M., 2007. Sorption behavior ofacidic herbicides on carbon nanotubes. Microchim. Acta 159(3e4), 293e298.
Rambabu, N., Guzman, C.A., Soltan, J., Himabindu, V., 2012.Adsorption characteristics of atrazine on granulated activatedcarbon and carbon nanotubes. Chem. Eng. Technol. 35 (2),272e280.
Rao, G.P., Lu, C., Su, F., 2007. Sorption of divalent metal ions fromaqueous solution by carbon nanotubes: a review. Sep. Purif.Technol. 58 (1), 224e231.
Rodriguez, A., Ovejero, G., Sotelo, J.L., Mestanza, M., Garcia, J.,2010. Adsorption of dyes on carbon nanomaterials fromaqueous solutions. J. Environ. Sci. Health A Toxic/Hazard.Subst. Environ. Eng. 45 (12), 1642e1653.
Salam, M.A., Burk, R.C., 2008. Thermodynamics ofpentachlorophenol adsorption from aqueous solutions byoxidized multi-walled carbon nanotubes. Appl. Surf. Sci. 255(5), 1975e1981.
Saleh, N.B., Pfefferle, L.D., Elimelech, M., 2008. Aggregationkinetics of multiwalled carbon nanotubes in aquatic systems:measurements and environmental implications. Environ. Sci.Technol. 42 (21), 7963e7969.
Saleh, N., Pfefferle, L.D., Elimelech, M., 2010. Influence ofbiomacromolecules and humic acid on the aggregationkinetics of single-walled carbon nanotubes. Environ. Sci.Technol. 44 (7), 2412e2418.
Shahryari, Z.G., Goharrizi, A.S., Azadi, M., 2010. Experimentalstudy of methylene blue adsorption from aqueous solutionsonto carbon nano tubes. Int. J. Water Resour. Environ. Eng. 2(2), 16e28.
Shen, X.E., Shan, X.Q., Dong, D.M., Hua, X.Y., Owens, G., 2009.Kinetics and thermodynamics of sorption of nitroaromaticcompounds to as-grown and oxidized multiwalled carbonnanotubes. J. Colloid Interface Sci. 330 (1), 1e8.
Shi, B.Y., Zhuang, X., Yan, X., Lu, J., Tang, H., 2010. Adsorptionof atrazine by natural organic matter and surfactantdispersed carbon nanotubes. J. Environ. Sci. China 22 (8),1195e1202.
Shih, Y.H., Gschwend, P.M., 2009. Evaluating activated carbon-water sorption coefficients of organic compounds using alinear solvation energy relationship approach and sorbatechemical activities. Environ. Sci. Technol. 43 (3), 851e857.
Smith, B., Yang, J., Bitter, J.L., Ball, W.B., Fairbrother, D.H., 2012.Influence of surface oxygen on the interactions of carbonnanotubes with natural organic matter. Environ. Sci. Technol.46 (23), 12839e12847.
Su, F., Lu, C., 2007. Adsorption kinetics, thermodynamics anddesorption of natural dissolved organic matter by multiwalledcarbon nanotubes. J. Environ. Sci. Health A Toxic/Hazard.Subst. Environ. Eng. 42 (11), 1543e1552.
Suffet, I.H., McGuire, M.J., 1978. Activated carbon adsorptionof organics from aqueous phase. Abstr. Pap. Am. Chem. Soc.176, 1.
Terrones, M., 2003. Science and technology of the twenty-firstcentury: synthesis, properties and applications of carbonnanotubes. Annu. Rev. Mater. Res. 33, 419e501.
Terrones, M., 2004. Carbon nanotubes: synthesis and properties,electronic devices and other emerging applications. Int. Mater.Rev. 49 (6), 325e377.
The Global Market for Carbon Nanotubes to 2015: a RealisticAssessment, 2010.
Toth, A., Torocsik, A., Tombacz, E., Laszio, K., 2012. Competitiveadsorption of phenol and 3-chlorophenol on purifiedMWCNTs. J. Colloid Interface Sci. 387, 244e249.
Tournus, F., Latil, S., Heggie, M.I., Charlier, J.-C., 2005. p-stackinginteractions between carbon nanotubes and organicmolecules. Phys. Rev. B 82 (21), 3746e3748.
Upadhyayula, V.K.K., Deng, S., Mitchell, M.C., Smith, G.B., 2009.Application of carbon nanotube technology for removal ofcontaminants in drinking water: a review. Sci. Total Environ.408 (1), 1e13.
Velzeboer, I., Kwadijk, C.J.A.F., Koelmans, A.A., 2014. Strongsorption of PCBs to nanoplastics, microplastics, carbonnanotubes and fullerenes. Environ. Sci. Technol. 48 (9),4869e4876.
wat e r r e s e a r c h 6 8 ( 2 0 1 5 ) 3 4e5 5 55
Wang, S.G., Liu, X.W., Gong,W.X., Nie, W., Gao, B.Y., Yue, Q.Y., 2007.Adsorption of fulvic acids from aqueous solutions by carbonnanotubes. J. Chem. Technol. Biotechnol. 82 (8), 698e704.
Wang, X., Lu, J., Xing, B.S., 2008. Sorption of organic contaminantsby carbon nanotubes: influence of adsorbed organic matter.Environ. Sci. Technol. 42 (9), 3207e3212.
Wang, X., Tao, S., Xing, B.S., 2009. Sorption and competition ofaromatic compounds and humic acid on multiwalled carbonnanotubes. Environ. Sci. Technol. 43 (16), 6214e6219.
Wang, F., Yao, J., Sun, K., Xing, B.S., 2010a. Adsorption of dialkylphthalate esters on carbon nanotubes. Environ. Sci. Technol.44 (18), 6985e6991.
Wang, X., Liu, Y., Tao, S., Xing, B.S., 2010b. Relative importanceof multiple mechanisms in sorption of organic compoundsby multiwalled carbon nanotubes. Carbon 48 (13),3721e3728.
Wang, F., Yao, J., Yu, C., Chen, H., Yi, Z., Choi, M.M.F., 2014.Mutual effects of dialkyl phthalate esters and humic acidsorption on carbon nanotubes in aqueous environments. ACSSustain. Chem. Eng. 2, 1219e1227.
Ward, T.M., Getzen, F.W., 1970. Influence of pH on the adsorptionof aromatic acids on activated carbon. Environ. Sci. Technol. 4(1), 64e67.
Wu, C.H., 2007a. Studies of the equilibrium and thermodynamicsof the adsorption of Cu2þ onto as-produced and modifiedcarbon nanotubes. J. Colloid Interface Sci. 311 (2), 338e346.
Wu, C.H., 2007b. Adsorption of reactive dye onto carbonnanotubes: equilibrium, kinetics and thermodynamics. J.Hazard. Mater. 144 (1e2), 93e100.
Xia, X., Li, Y., Zhou, Z., Feng, C., 2010a. Bioavailability of adsorbedphenanthrene by black carbon andmulti-walled carbonnanotubes to Agrobacterium. Chemosphere 78 (11), 1329e1336.
Xia, X.R., Monteiro-Riviere, N.A., Riviere, J.E., 2010b. An index forcharacterization of nanomaterials in biological systems. Nat.Nanotechnol. 5 (9), 671e675.
Xie, W.H., Shiu, W.Y., Mackay, D., 1997. A review of the effect ofsalts on the solubility of organic compounds in seawater. Mar.Environ. Res. 44 (4), 429e444.
Yan, X.M., Shi, B.Y., Lu, J.J., Feng, C.H., Wang, D.S., Tang, H.X.,2008. Adsorption and desorption of atrazine on carbonnanotubes. J. Colloid Interface Sci. 321 (1), 30e38.
Yang, K., Xing, B.S., 2007. Desorption of polycyclic aromatichydrocarbons from carbon nanomaterials in water. Environ.Pollut. 145 (2), 529e537.
Yang, K., Xing, B.S., 2009. Adsorption of fulvic acid by carbonnanotubes from water. Environ. Pollut. 157 (4), 1095e1100.
Yang, K., Xing, B., 2010. Adsorption of organic compounds bycarbon nanomaterials in aqueous phase: Polanyi Theory andits application. Chem. Rev. 110 (10), 5989e6008.
Yang, K., Wang, X., Zhu, L., Xing, B.S., 2006a. Competitivesorption of pyrene, phenanthrene, and naphthalene onmultiwalled carbon nanotubes. Environ. Sci. Technol. 40 (18),5804e5810.
Yang, K., Zhu, L.Z., Xing, B.S., 2006b. Adsorption of polycyclicaromatic hydrocarbons by carbon nanomaterials. Environ. Sci.Technol. 40 (6), 1855e1861.
Yang, K., Wu, W., Jing, Q., Zhu, L., 2008. Aqueous adsorption ofaniline, phenol, and their substitutes by multi-walled carbonnanotubes. Environ. Sci. Technol. 42 (21), 7931e7936.
Yang, K., Wu, W., Jing, Q., Jiang, Q., Xing, B.S., 2010. Competitiveadsorption of naphthalene with 2,4-dichlorophenol and 4-chloroaniline on multiwalled carbon nanotubes. Environ. Sci.Technol. 44 (8), 3021e3027.
Yang, W., Lu, Y., Zheng, F., Xue, X., Li, N., Liu, D., 2012. Adsorptionbehavior and mechanisms of norfloxacin onto porous resinsand carbon nanotube. Chem. Eng. J. 179, 112e118.
Yao, J.H., Strauss, G., 1992. Adsorption of cationic surfactants onmedical polymers - effects of surfactant and substratestructures. Langmuir 8 (9), 2274e2278.
Yu, J., Grossiord, N., Koning, C.E., Loos, J., 2007. Controlling thedispersion of multi-walled carbon nanotubes in aqueoussurfactant solution. Carbon 618e623.
Yu, F., Ma, J., Wu, Y., 2012. Adsorption of toluene, ethylbenzeneand xylene isomers on multi-walled carbon nanotubesoxidized by different concentration of NaOCl. Front. Environ.Sci. Eng. 6 (3), 320e329.
Yu, F., Ma, J., Han, S., 2014. Adsorption of tetracycline fromaqueous solutions onto multi-walled carbon nanotubes withdifferent oxygen contents. Sci. Rep. 4, 1e8.
Zhang, S., Shao, T., Bekaroglu, S.S.K., Karanfil, T., 2009. Theimpacts of aggregation and surface chemistry of carbonnanotubes on the adsorption of synthetic organic compounds.Environ. Sci. Technol. 43 (15), 5719e5725.
Zhang, S., Shao, T., Kose, H.S., Karanfil, T., 2010a. Adsorption ofaromatic compounds by carbonaceous adsorbents: acomparative study on granular activated carbon, activatedcarbon fiber, and carbon nanotubes. Environ. Sci. Technol. 44(16), 6377e6383.
Zhang, D., Pan, B., Zhang, H., Ning, P., Xing, B.S., 2010b.Contribution of different sulfamethoxazole species to theiroverall adsorption on functionalized carbon nanotubes.Environ. Sci. Technol. 44 (10), 3806e3811.
Zhang, S., Shao, T., Bekaroglu, S.S.K., Karanfil, T., 2010c.Adsorption of synthetic organic chemicals by carbonnanotubes: effects of background solution chemistry. WaterRes. 44 (6), 2067e2074.
Zhang, S., Shao, T., Karanfil, T., 2011. The effects of dissolvednatural organic matter on the adsorption of synthetic organicchemicals by activated carbons and carbon nanotubes. WaterRes. 45 (3), 1378e1386.
Zhang, S., Shao, T., Kose, H.S., Karanfil, T., 2012. Adsorptionkinetics of aromatic compounds on carbon nanotubes andactivated carbons. Environ. Toxicol. Chem. 31 (1), 79e85.
Zhao, J., Lu, J.P., Han, J., Yang, C.-K., 2003. Noncovalentfunctionalization of carbon nanotubes by aromatic organicmolecules. Appl. Phys. Lett. 82 (21), 3746e3748.