Chemosphere 119 (2015) 1169–1175
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Influence of humic acids on sorption of alkanes by carbon nanotubes – Implications for the dominant sorption mode Thorsten Hüffer a,b,c,⇑, Sarah Schroth a, Torsten C. Schmidt a,b a
Instrumental Analytical Chemistry, University of Duisburg-Essen, Universitätsstrasse 5, 45141 Essen, Germany Centre for Water and Environmental Research (ZWU), University of Duisburg-Essen, Universitätsstrasse 2, 45141 Essen, Germany c Department of Environmental Geosciences, University of Vienna, Althanstrasse 14, 1090 Vienna, Austria b
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Sorption isotherms of alkanes by
carbon nanotubes were determined. Sorption of n-alkanes was generally
stronger than sorption of their cyclic homologs. Addition of humic substances led to decreasing sorption affinity and increasing sorption linearity. The dominant sorption mode by carbon nanotubes remained adsorption independent from the addition of humic substances.
a r t i c l e
i n f o
Article history: Received 29 April 2014 Received in revised form 24 September 2014 Accepted 30 September 2014 Available online 31 October 2014 Handling Editor: Keith Maruya Keywords: Carbon nanomaterials Adsorption vs. absorption Organic compounds Humic substances
a b s t r a c t The presence of humic substances (HS) has previously been shown to alter sorption properties of multi-walled carbon nanotubes (MWCNTs). To systematically study this process, three alkane pairs were selected as molecular probe sorbates. The influence of HS on sorption affinity, sorption linearity, and the dominant sorption mode (i.e., ad- or absorption) by MWCNTs was investigated. The addition of HS led to a continuous decrease in sorption affinity and an increase in sorption linearity with increasing HS addition. Furthermore, the comparison of distribution coefficients of n- and cycloalkanes showed that the dominant sorption mode remains to be adsorption regardless of the presence of HS on MWCNT surface. From this, it can be concluded that instead of a change in sorption mode to absorption of sorbates into HS-coated MWCNT, HS blocks high-energy sorption sites for subsequently added sorbates and that sorbates continuously sorb on the MWCNT surface. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Since the discovery of carbon-based nanomaterials (CNM) in the early 1990s (Iijima, 1991), the research and development of ⇑ Corresponding author at: Department of Environmental Geosciences, University of Vienna, Althanstrasse 14, 1090 Vienna, Austria. Tel.: +43 (1) 4277 53383; fax: +43 (1) 4277 9533. E-mail address: [email protected] (T. Hüffer). http://dx.doi.org/10.1016/j.chemosphere.2014.09.097 0045-6535/Ó 2014 Elsevier Ltd. All rights reserved.
CNM have grown steadily due to their unique physicochemical properties. Together with an expected growth in production, the question of the environmental impact of CNM has become a controversially discussed issue among researchers. To date, most data on environmental concentrations of CNM are still based on modeling (Koelmans et al., 2009; Gottschalk et al., 2009) and although engineered CNM concentrations were predicted to be lower compared to naturally occurring nanomaterials (Koelmans et al.,
1170
T. Hüffer et al. / Chemosphere 119 (2015) 1169–1175
2009), their concentrations are expected to increase in the future (Gottschalk et al., 2013). Various environmentally relevant parameters have been shown to influence the sorption behavior by CNM, such as the presence of oxygen and sunlight (Hüffer et al., 2013) or natural organic matter (NOM) (Wang et al., 2008; Zhang et al., 2011, 2012; Velzeboer et al., 2014; Kah et al., 2014). Generally, two mechanisms have been proposed on how NOM affects HOC sorption by CNM. Firstly, NOM directly competes with HOC for sorption sites on CNM surface. This seems especially relevant as commonly low sorbate and high NOM concentrations have been combined (Wang et al., 2008), which represents a realistic environmental scenario. Secondly, sorption of HOC is reduced by large NOM molecules that block the entrances of CNM pores that may otherwise be available for HOC sorption (Wang et al., 2009; Zhang et al., 2011). Recent investigations on the influence of NOM addition on HOC sorption have shown inconsistent effects on sorption strength of multi-walled carbon nanotubes (MWCNTs). A decrease in sorption strength of NOM-coated MWCNTs compared to the pristine tubes was assigned to an offset between ‘‘creating’’ new sorption sites due to increased dispersity and the reduced accessibility of polar moieties (Wang et al., 2008). In contrast, the contribution of an increase in sorption due to higher MWCNT dispersion was very recently shown to overcome the observed decreased sorption due to NOM coating (Pan et al., 2013). In addition, sorption became more linear after NOM addition. This was principally related to the presence of a co-solute, which reduces sorption of the primary solute (i.e., HOC) and thus increases sorption linearity (Zhang et al., 2011). For example, Zhang et al. studied the effect of NOM addition on sorption of phenanthrene, biphenyl, and phenylphenol by various carbon nanotubes (CNTs). From experimental data, sorption became more linear after NOM addition, while the largest change in sorption linearity was obtained for preloaded CNTs (Zhang et al., 2011). An important question in the phase transfer of HOC is to clarify the dominant sorption mode. Here, generally two sorption modes are discussed: one is the adsorption on the surface of a sorbent; the other is absorption, which is the distribution between two condensed phases. To clarify the sorption mode, Endo et al. proposed an approach to differentiate between ad- and absorption, in which the distribution coefficients of linear alkanes (Kn) and their cyclic homologs (Kc) were compared as follows (Endo et al., 2008):
K n =K c ¼ ðK d =K aw of n-alkaneÞ=ðK d =K aw of cyclo-alkaneÞ
ð1Þ
where Kd is the sorbent–water distribution coefficient and Kaw is the air–water partitioning constant. This comparison of the distribution coefficients of n- and cyclo-alkanes is based on difference in sorption energies of alkanes. First, linear alkanes preferably adsorb on a surface since all carbon atoms of planar n-hexane being able to interact with a surface while for non-planar cyclohexane only three or four carbon atoms can interact (Smiciklas et al., 2000; Onjia et al., 2001). Second, the partitioning (i.e., absorption) from air into a bulk phase of cycloalkanes is stronger than of their linear homologs due to a smaller cavity required for cycloalkanes (Goss, 2004). Note that Kn and Kc cannot be directly compared as experimentally derived Kd values additionally include molecular interactions with water, and for the characterization of the dominant sorption mode by a given sorbent, only the molecular interactions in/on the sorbent are of interest. Given Eq. (1), Kn/Kc ratios < 1 indicate absorption as the dominant sorption mode. Kn/Kc ratio 1 suggest adsorption if steric effects are negligible, while for Kn/Kc > 1 steric hindrance, e.g., from size exclusion effects, has to be considered (Endo et al., 2009). Recently, sorption of n-alkanes by MWCNTs was shown to be stronger than sorption of cycloalkanes (Hüffer et al., 2014). Thus, adsorption seems to be the dominant sorption mode for MWCNTs. Endo et al. observed that for soot, commonly regarded as an adsorber, the predominant sorption mode changed from absorption to
adsorption accompanied by an decrease in sorption linearity following the removal of extractable organic matter (Endo et al., 2009). Thus, the previously observed increase in sorption linearity by CNM following NOM addition may result from a change in sorption mode from ad- to absorption. To this end, the here presented study investigated the dominant sorption mode by MWCNTs influenced by the presence of various types and concentrations of humic substances (HS).
2. Materials and methods 2.1. Materials Multi-walled carbon nanotubes (C150HP) were purchased from Bayer Material Science (Leverkusen, Germany). Properties of MWCNTs are given in Supplemental Table S1. Four humic substances (HS) were selected (Supplemental Table S2 lists HS and selected properties). Stock solutions were prepared by dissolving 0.25–1 g HS in 100 mL 0.1 mol NaOH overnight. After filtration through a 0.45 lm cellulose membrane (Whatman), the solution was diluted to 1 L and the pH was adjusted to neutral with 0.1 mol NaCl. The concentrations of HS stock solutions were determined by total organic carbon content (TOC; Shimadzu TOC-5000, Duisburg, Germany). HS stock solutions were kept at 4 °C in the dark for no longer than six weeks and TOC concentrations were regularly monitored. Working standard HS solutions were obtained by diluting stock solutions accordingly to concentrations of 1 and 10 mg L1. Three pairs of n- and cycloalkanes (C6–C8) were selected as molecular probe compounds (all purchased from Sigma–Aldrich with >99% purity). Stock solutions of alkanes were freshly prepared in methanol (HPLC grade, Fisher Scientific, UK) every week. Selected physico-chemical properties of sorbates are displayed in Supplemental Table S3.
2.2. Sorption batch experiments Sorption isotherms were determined in a multi-phase system. 2–5 mg of MWCNTs were weighted into 20-mL amber glass vials. 10 or 15 mL of background solution were added. The background solution contained either 10 mmol CaCl2 to regulate the ionic strength or the HS solution in concentrations as given in Supplemental Table S2. Vials were then horizontally shaken for 24 h to allow the MWCNTs to be coated with HS. The samples were spiked with methanolic stock solutions resulting in 12 different sorbate concentrations ranging over 3–4 orders of magnitude. Highest equilibrium aqueous concentrations of sorbates were 20 times below their aqueous solubility. The methanol content did not exceed 0.5 vol% in water to minimize co-solvent effects. The samples were shaken for at least 24 h for equilibration at 25 °C in a temperature-controlled room. Equilibration times were pre-determined (data not shown). Subsequently, the vials were removed from the shaker and placed in the tray of the autosampler for 1 h to allow air–water equilibration at 20 °C. The analyte concentrations in the headspace of the vials were determined by direct injection of 1000 lL of the vial headspace into GC–MS. External calibration vials were prepared one day prior to analysis and contained the according background solutions. For quantification, a gas chromatography-mass spectrometry system (TraceGC Ultra/ DSQ, Thermo Finnigan) equipped with a PAL Combi-xt autosampler (CTC Analytics) was used. An HP-5 capillary column (30 m 0.32 mm 0.25 lm, J&W Scientific) and oven temperatures between 70 and 150 °C were used depending on the analyte retention times.
1171
T. Hüffer et al. / Chemosphere 119 (2015) 1169–1175
Sorbed concentrations could be determined by mass balance approach using air–water partitioning constants (Kaw) of the analytes (see Supplemental Table S3):
C MWCNTs ¼
Vw mT CKaaw CaV a
MMWCNTs
ð2Þ
where Ca and CMWCNTs are the alkane concentrations in air (lg L1) and sorbed by MWCNTs (lg kg1), respectively; Va and Vw denote the vial volume of air (L) and background solution (L), respectively; and Mw is the mass of MWCNTs added to the vials (kg). Note that sorption of alkanes by HS was pre-determined and found to be negligible compared to sorption by MWCNTs, which is in accordance to previous reports (Chen et al., 2008; Zhang et al., 2010). Thus, ‘extrinsic’ sorption by HS and MWCNTs as a whole sorbent was considered since the HS/MWCNTs mass ratio used in the present was rather low ranging from 0.03 to 0.075 g HS/g MWCNTs (Zhang et al., 2012). Kaw values were carefully selected from literature to ensure accurate calculation of Kd values. Loss of analytes during the experiments was exemplarily determined for n-heptane, where after a six days shaking period, a mean loss of 6% was obtained (n = 5); and according corrections were made in the mass balance. 2.3. Determination of sorption mode The Freundlich model (FM) was used to fit the isotherm data:
C MWCNTs ¼ K F ðC w Þn
ð3Þ
where Cw is the aqueous concentration (lg L1), and KF and n are the Freundlich coefficient ((lg kg1)/(lg L1)n) and the Freundlich exponent [], respectively. Sigma Plot 12.0 for Windows was used for statistical analyses and model fits. Computation of model parameters and their standard errors was performed with reduced Chi-square and weighted regression with weighting to reciprocal y2. Single point distribution coefficients (K d ¼ C MWCNTs =C w ; (L kg1)) were determined by interpolation at given sorbate air concentration. All determined isotherms were non-linear; thus, Kd and Kn/Kc ratios may vary with concentration. This concentration-dependency can be accounted for by comparing the data at standardized chemical activity (e.g., by passive dosing (Mayer and Holmstrup, 2008) or reference concentration measures (Endo et al., 2009). As pointed out by Endo et al., calculating Kn/Kc ratios with Kd values occurring at the same Cair as reference concentration, may be the most appropriate (Endo et al., 2009). Kd values were hence calculated for an air-phase concentration of 10 lg L1. Propagated errors in Kn/Kc ratios were determined using Gaussian error propagation approach. 3. Results and discussion 3.1. Sorption isotherms Isotherm data were well fit with the Freundlich model (R > 0.95). It has been previously discussed that other, more complex, non-linear sorption models are more suitable for a robust interpretation of sorption isotherm data especially if large concentration ranges are considered (Kah et al., 2011). However, the aim of this study was to investigate how the presence of HS affects sorption by MWCNTs in terms of sorption affinity and sorption linearity; therefore, only the Freundlich model was taken into account for isotherm data fitting. The results of the Freundlich model fits are summarized in Table 1. Sorption isotherms of n- and cycloalkanes by MWCNTs following various kinds and concentrations of HS are shown in Fig. 1. In 2
general, sorption isotherms by all sorbent scenarios were parallel accompanied by an increase in sorption with increasing number of carbon atoms (i.e., C6 < C7 < C8). This observation is reflected by a corresponding increase of sorbate hydrophobicity parameters, such as an increase in the hexadecane–air partitioning constant (L) as displayed in Supplemental Table S3. However, it must be stressed that L refers to absorption processes, which may limit its comparability to adsorption by CNM. Sorption of n-alkanes was always up to one order of magnitude stronger compared to sorption of their cyclic homologous. The sorbate morphology (i.e., molecular size and shape) plays a major role with respect to availability of sorption sites on the CNM surface. Stronger sorption was previously explained by a better contact of linear and planar compounds with the CNM surface (Gotovac et al., 2007). Similar conclusions have been drawn for alkane sorption by flat surfaces, where linear n-hexane can interact with its six carbon atoms with the sorbent surface, while non-planar cyclohexane can only do so with three or four carbon atoms due to its non-planar shape (Smiciklas et al., 2000; Onjia et al., 2001). These differences in number of carbon atoms in contact with the CNM surface are also likely to be the case for n- and cyclooctane (Dorofeeva et al., 1985).
Table 1 Fitting parameters ± standard errors, R2, standard error of estimates (SEE) for Freundlich model fitting the isotherms of C6–C8 alkanes pairs by MWCNTs following HS addition (N: number of data points). Sorbate
KF
n
R2
SEE
N
MWCNTs nHex cHex nHep cHep nOct cOct
1.69E+05 ± 4.48E+03 1.77E+04 ± 9.14E+02 5.03E+05 ± 2.65E+04 4.57E+04 ± 2.89E+03 9.20E+05 ± 2.11E+04 1.57E+05 ± 7.21E+03
0.71 ± 0.02 0.73 ± 0.02 0.66 ± 0.03 0.68 ± 0.02 0.55 ± 0.02 0.55 ± 0.01
0.9951 0.9720 0.9828 0.9922 0.9914 0.9941
0.09 0.15 0.17 0.12 0.08 0.08
11 10 11 10 12 14
MWCNTs+1HA nHex 1.08E+05 ± 6.17E+03 cHex 1.09E+04 ± 5.53E+02 nHep 2.38E+05 ± 8.30E+03 cHep 2.07E+04 ± 1.50E+03 nOct 5.34E+05 ± 1.72E+04 cOct 6.71E+04 ± 4.86E+03
0.79 ± 0.03 0.83 ± 0.02 0.73 ± 0.02 0.78 ± 0.03 0.66 ± 0.02 0.71 ± 0.03
0.9508 0.9732 0.9793 0.9734 0.9908 0.9688
0.20 0.17 0.13 0.15 0.12 0.16
15 13 15 12 14 14
MWCNTs+10HA nHex 5.39E+04 ± 1.60E+03 cHex 5.26E+03 ± 1.97E+02 nHep 8.98E+04 ± 3.28E+03 cHep 9.78E+03 ± 7.44E+02 nOct 1.96E+05 ± 6.94E+03 cOct 2.45E+04 ± 1.69E+03
0.86 ± 0.01 0.89 ± 0.01 0.72 ± 0.02 0.79 ± 0.03 0.71 ± 0.02 0.74 ± 0.02
0.9969 0.9979 0.9794 0.9783 0.9835 0.9888
0.11 0.11 0.14 0.17 0.12 0.17
13 12 15 11 13 14
MWCNTs+LSHA nHex 5.01E+04 ± 2.70E+03 cHex 4.94E+03 ± 5.85E+02 nHep 1.69E+05 ± 8.43E+03 cHep 2.03E+04 ± 1.24E+03 nOct 4.04E+05 ± 1.61E+04 cOct 4.98E+04 ± 3.87E+03
0.83 ± 0.02 0.86 ± 0.04 0.69 ± 0.03 0.72 ± 0.01 0.71 ± 0.02 0.74 ± 0.02
0.9914 0.9513 0.9741 0.9861 0.9841 0.9813
0.17 0.23 0.15 0.11 0.10 0.13
15 14 13 14 14 14
MWCNTs+PPHAS nHex 7.78E+04 ± 7.35E+03 cHex 7.88E+03 ± 5.33E+02 nHep 1.89E+05 ± 5.31E+03 cHep 2.41E+04 ± 1.62E+03 nOct 5.37E+05 ± 2.79E+04 cOct 6.50E+04 ± 4.37E+03
0.82 ± 0.03 0.85 ± 0.02 0.67 ± 0.02 0.73 ± 0.03 0.70 ± 0.02 0.74 ± 0.03
0.9533 0.9698 0.9812 0.9728 0.9628 0.9692
0.22 0.19 0.10 0.15 0.16 0.16
14 15 15 12 12 12
MWCNTs+WPHAR nHex 5.24E+04 ± 2.88E+03 cHex 5.34E+03 ± 3.35E+02 nHep 9.67E+04 ± 2.87E+03 cHep 1.22E+04 ± 8.20E+02 nOct 2.09E+05 ± 5.51E+03 cOct 2.60E+04 ± 9.55E+02
0.84 ± 0.02 0.87 ± 0.03 0.66 ± 0.01 0.72 ± 0.02 0.68 ± 0.01 0.74 ± 0.01
0.9847 0.9716 0.9912 0.9798 0.9965 0.9968
0.13 0.17 0.10 0.16 0.08 0.07
13 13 14 11 12 13
1172
T. Hüffer et al. / Chemosphere 119 (2015) 1169–1175
Fig. 1. Sorption isotherms of (d) nHex, (s) cHex, (.) nHep, (D) cHep, (j) nOct, and (h) cOct by MWCNT (A), MWCNT + 1HA (B), MWCNT + 10HA (C), MWCNT + LSHA (D), MWCNT + PPHAS (E), and MWCNT + WPHAR (F), respectively.
3.2. HS effect on sorption affinity and linearity The sorption affinity (log KF in Table 1) was continuously decreased with increasing concentration of HA, i.e., log KF (MWCNT) > log KF (MWCNT + 1HA) > log KF (MWCNT + 10HA). This decrease in log KF was significant (p < 0.05) except for cHex. Overall, the addition of 10 mg L1 HA led to a decrease in sorption affinity ranging between 0.5 log units for nHex and 0.8 log units for cOct. Zhang et al. observed a clear linear decrease in sorption affinity of MWCNTs for pyrene following Suwannee River humic acid addition ranging from 0 to 200 mg L1; however, the obtained decrease in KF was smaller than the almost one order of magnitude effect observed here (Zhang et al., 2012).
It has previously been suggested that the addition of NOM leads a reduction in sorption affinity due to the occupation of sorption sites by NOM (Chen et al., 2008; Wang et al., 2009; Zhang et al., 2012; Velzeboer et al., 2014). This seems to be the case for the here investigated sorption scenarios, where sorption by MWCNTs was reduced following the addition of HS. Furthermore, a decrease in sorption affinity for HOC may be explained by a decrease in hydrophobicity of CNTs due to the introduction of polar moieties to CNTs surface from NOM addition. This facilitates the formation of water clusters and consequently reduces the interaction with HOC due to a decrease in hydrophobicity of CNT surface (Sun et al., 2013). In contrast, an increase in sorption has been reported to occur due to a concomitant increase in available sorption sites following
T. Hüffer et al. / Chemosphere 119 (2015) 1169–1175
NOM addition (Wang et al., 2008; Gai et al., 2011). This was additionally assigned to the destabilization of CNM aggregates, e.g., resulting from sonication (Pan et al., 2013). Extensive characterization data by Zhang et al. suggest that there is no change in aggregation from the presence of NOM (Zhang et al., 2012). Furthermore, Kah et al. recently showed that the presence of NOM efficiently dispersed CNTs only when combined with a sonication treatment (Kah et al., 2014). In the present study, CNT dispersions were not sonicated; therefore, changes in aggregation between bulk sorbent material and HS-coated CNTs were not expected and differences in sorption by CNTs do not seem to result from a changed aggregation status of the sorbent. However, data on dispersion stability of CNM affected by NOM remain limited. Previous investigations have shown that the strength of interaction between CNTs and HOC depends also on sorbent structural properties. The effect of nanoscale curvature of the sorbent has been discussed for sorption of polycyclic aromatic hydrocarbons (PAHs) by single-walled carbon nanotubes (SWCNTs) Gotovac et al., 2007. However, this effect is expected to play a minor role in the present study since the investigated sorbates were comparably smaller. In addition, MWCNTs have a much larger diameter of >13 nm compared to 1 nm in diameter of single-walled nanotubes. Decreases in sorption by MWCNTs and graphite have been explained by structural differences in the sorbent material. The interstitial and groove areas of CNT aggregates may not be available for sorption of large humic substance molecules (Ji et al., 2009), while these sites remain available for sorption of comparably small alkane molecules after HS addition. Non-linear sorption (n = 0.55–0.79) was observed for all sorbates by MWCNTs without HS addition. This is generally consistent with recent investigations of sorption of non-ionic organic compounds (Yang et al., 2006; Kah et al., 2011). Commonly non-linear sorption by CNM is explained by a heterogeneous distribution of free-energy of sorption sites caused by e.g., surface defects or interstitial and groove areas of CNT aggregates (Pan and Xing, 2008). Moreover, aggregation of CNTs may lead to a heterogeneous distribution of sorption site energies (Zhang et al., 2009). The addition of 1 and 10 mg L1 HA led to a significant increase in sorption linearity (p < 0.05) with Dn ranging between 0.08 and 0.19 for sorption of nHep and cOct, respectively. Previous studies reported similar results regarding the impact of NOM on sorption linearity of HOC by CNTs (Zhang et al., 2011; Sun et al., 2013) and the increase in sorption linearity was recently explained by a more homogeneous energy distribution of sorption sites on NOM-coated CNM (Yang et al., 2006; Velzeboer et al., 2014). However, an increase in sorption linearity has not been consistently reported in literature as in some studies no change in sorption linearity was observed following the addition of NOM (Wang et al., 2008; Chen et al., 2008). Endo et al. investigated sorption properties by carbonaceous materials and found that for diesel soot commonly regarded as an adsorbing material sorption of n- and cyclooctane was linear (Endo et al., 2009). Furthermore, the extraction of organic material from reference diesel soot resulted in a significant decrease in sorption linearity of n = 0.75 and 0.85 for n- and cyclooctane, respectively (Endo et al., 2009). The here observed decrease in
Table 2 Kn/Kc ratios and errors of C6–C8 alkane pairs for sorbents. Sorbent
C6
C7
C8
MWCNT MWCNT + 1HA MWCNT + 10HA MWCNT + LSHA MWCNT + PPHAS MWCNT + WPHAR
2.01 ± 0.03 1.72 ± 0.03 1.55 ± 0.02 1.62 ± 0.03 1.62 ± 0.12 1.55 ± 0.07
1.73 ± 0.11 1.43 ± 0.03 1.17 ± 0.02 1.17 ± 0.07 1.16 ± 0.03 1.18 ± 0.03
0.96 ± 0.03 0.84 ± 0.03 0.72 ± 0.03 0.74 ± 0.02 0.78 ± 0.05 0.79 ± 0.03
1173
sorption affinity (KF) accompanied by the increasing sorption linearity implies that strong sorption sites on MWCNTs surface are occupied by HS, which (i) makes them no longer available for interaction and (ii) results in a more homogeneous distribution of sorption site energies and thus more linear sorption. The latter could also result from a change of sorption mode to absorption of alkanes into HS-coated on MWCNTs surface, which will be discussed in the next section. 3.3. Kn/Kc ratios for sorption by MWCNTs Kn/Kc ratios for sorption of C6–C8 alkane pairs by MWCNTs under varying HS addition conditions are given in Table 2. For sorption by MWCNTs, a Kn/Kc ratio for the C6-alkane pair of 2.01 was obtained. This is generally comparable to previous reports for similar carbonaceous sorbent materials, where Kn/Kc ranged between of 1.02 and 2.20 for carbon nanofibers and activated carbon, respectively (Endo et al., 2008). From literature data for sorption of organic vapors by single-walled carbon nanotubes, a C6-Kn/Kc ratio of 1.08 can be calculated (Agnihotri et al., 2005). Kn/Kc ratios of C6 in Table 2 consistently decreased to 1.72 and 1.55 following 1 and 10 mg L1 HA addition, respectively. Overall, a significant decrease (p < 0.05) in Kn/Kc ratios was observed for all alkane pairs following HS addition. However, Kn/Kc ratios of C6 and C7 alkane pairs for all sorbent scenarios remain larger than 1, which suggests adsorption to be the dominant sorption mode regardless of HS addition. Kn/Kc ratios of MWCNTs coated with the reference HS materials (i.e., LSHA, PPHAS, and WPHAR) were not significantly different (p > 0.05) from those obtained for non-certified HS material (i.e., HA), suggesting that the effect of HS on the dominant sorption mode by MWCNTs is independent from the properties of HS. In order to further investigate a potential change in sorption mode following HS addition, C8-Kn/Kc ratios were plotted against varying sorbate air-phase concentrations for all sorbents scenarios (see Supplemental Fig. S1). Sorption of n- and cyclooctane by MWCNTs shows Kn/Kc 1 over the whole concentration range, which has also been observed for other rigid carbonaceous materials, such as charcoal and activated carbon (Endo et al., 2009). This suggests that adsorption is the dominant sorption mode regardless of sorbate concentration. The addition of HS in general led to a decrease in Kn/Kc with increasing sorbate concentration. This was similarly observed for carbonaceous geosorbents containing native organic matter, e.g., soot (Endo et al., 2009), for which sorption was furthermore explained by a combination of ad- and absorption (Endo et al., 2009). It must be stressed that this effect was far more pronounced with decreasing Kn/Kc ratios from 1 to 0.3 or lower for soils (Endo et al., 2009). In the present study Kn/Kc ratios decreased from 1 to 0.6; thus, the contribution of absorption by MWCNTs following HS addition does not seem to be of major importance. From the results in Table 2, a significant decrease (p < 0.05) in Kn/Kc ratios with increasing number of sorbate carbon atoms was observed. These changes are the highest observed so far for the Kn/Kc-concept. Despite the fact that the lowest Kn/Kc ratios were obtained for sorption of n- and cyclooctane by MWCNT + 10HA, adsorption can be assumed to be the dominant sorption mode as Kn/Kc of 0.72 is above values obtained for typical absorbers, such as bulky liquids (e.g., C8-Kn/Kc of 0.23 for 1-octanol) or organic polymers (e.g., C8-Kn/Kc of 0.33 for polypropylene) (Endo et al., 2008). Overall, the observed effects of HS addition on Kn/Kc were not as pronounced as in previous reports, where the removal of extractable organic matter from a standard reference material of diesel soot clearly resulted in a change in sorption mode from ab- to adsorption (Endo et al., 2009). In fact, Kn/Kc ratios for sorption of
1174
T. Hüffer et al. / Chemosphere 119 (2015) 1169–1175
n- and cyclooctane by the untreated sorbent ranged between 0.46 and 0.66, while sorption by NOM-extracted sorbent material showed Kn/Kc values between 0.83 and 1.72. The here observed changes in Kn/Kc support the argument that the dominant sorption mode by HS-coated MWCNTs remains adsorption by the nanotubes rather than absorption into the attached organic matter. Aside from sorption onto the surface, pore-filling has been discussed as one of the major sorption mechanisms by CNM (Pan and Xing, 2008; Kah et al., 2011; Hüffer et al., 2013), which was especially relevant for low sorbate concentrations because of the high energetic sorption sites in the micropores of CNTs (Wang et al., 2008). Meanwhile, it was suggested by Wang et al. that fractionation of HS components by MWCNTs occurs dependent on the molecular weight and composition of HS and on the porosity of MWCNTs (Wang et al., 2011). Furthermore, steric effects may have to be considered as the sorption capacity of PCBs and PAHs by MWCNTs was negatively related to the sorbates’ molecular size (Yang et al., 2006; Velzeboer et al., 2014). Commonly, the inner pores of nanotubes are not available for sorption (Pan and Xing, 2008), so any discussion on the impact of pore size on sorption is restricted to potential pores of CNT aggregates. Microporous materials may exhibit size exclusion effects as pores may not be available for larger sorbates than the pore sizes. This led to C6-Kn/Kc ratios of up to 1500 for graphitized or activated carbons (Endo et al., 2008). The volume of micro- and mesopores of MWCNTs applied in this study is given in Supplemental Table S1. Since Kn/Kc ratios for the here investigated sorbent scenarios were smaller, the occurrence of size exclusion effects may be neglected. The addition of HS seems to have the following impact on sorption behavior by MWCNTs: adsorption remains to be the dominant sorption mode regardless of HS addition and size exclusion effects from pore blockage by HS may be neglected, which is further supported by the fact that C8-Kn/Kc ratios shown in Supplemental Fig. S1 remain constantly 1. In contrast, Kn/Kc ratios decreased from >1 to 1 with increasing sorbate concentration for lignite coke and activated carbon, where steric effects at low sorbate concentration were explained by higher availability of strong interaction sites for n- than for cyclooctane (Endo et al., 2009). Moreover, the humic substance concentrations were not sufficient to significantly sorb alkane sorbates, while they significantly affected sorption of alkanes by MWCNTs. Thus, competitive sorption of alkanes by MWCNTs rather than HS seems to dominate over an enhanced solubilization of sorbates from sorption by HS (i.e., ‘‘solubilizing effect’’). Previously, the suppression of sorption of PAHs by MWCNTs in the presence of pulmonary surfactants was shown to be a combination of both solubilization and competitive sorption (Zhao et al., 2012).
4. Conclusions The addition of humic substances affected sorption affinity and linearity of n- and cyclo-alkanes by multi-walled carbon nanotubes: sorption affinity was decreased while sorption linearity was increased. The ratio of distribution coefficients of n- and cycloalkanes suggested that adsorption is the dominant sorption mode by MWCNTs, which was not altered by the presence of humic substances. Acknowledgements The authors would like to thank Dr. Satoshi Endo (UFZ Leipzig, Germany) and Dr. Mélanie Kah (University of Vienna, Austria) for their helpful contributions. This work was financially supported by the Deutsche Forschungsgemeinschaft (DFG, SCHM 1372/10-1).
Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.chemosphere. 2014.09.097. References Agnihotri, S., Rood, M.J., Rostam-Abadi, M., 2005. Adsorption equilibrium of organic vapors on single-walled carbon nanotubes. Carbon 43 (11), 2379–2388. Chen, J.Y., Chen, W., Zhu, D., 2008. Adsorption of nonionic aromatic compounds to single-walled carbon nanotubes: effects of aqueous solution chemistry. Environ. Sci. Technol. 42 (19), 7225–7230. Dorofeeva, O.V., Mastryukov, V.S., Allinger, N.L., Almenningen, A., 1985. The molecular-structure and conformation of cyclooctane as determined by electron-diffraction and molecular mechanics calculations. J. Phys. Chem. 89 (2), 252–257. Endo, S., Grathwohl, P., Schmidt, T.C., 2008. Absorption or adsorption? Insights from molecular probes n-alkanes and cycloalkanes into modes of sorption by environmental solid matrices. Environ. Sci. Technol. 42 (11), 3989–3995. Endo, S., Grathwohl, P., Handerlein, S.B., Schmidt, T.C., 2009. Characterization of sorbent properties of soil organic matter and carbonaceous geosorbents using n-alkanes and cycloalkanes as molecular probes. Environ. Sci. Technol. 43 (2), 393–400. Endo, S., Grathwohl, P., Haderlein, S.B., Schmidt, T.C., 2009. Effects of native organic material and water on sorption reference diesel soot. Environ. Sci. Technol. 43 (9), 3187–3193. Gai, K., Shi, B., Yan, X., Wang, D., 2011. Effect of dispersion on adsorption of atrazine by aqueous suspensions of fullerenes. Environ. Sci. Technol. 45 (14), 5959– 5965. Goss, K.U., 2004. The air/surface adsorption equilibrium of organic compounds under ambient conditions. Crit. Rev. Environ. Sci. Technol. 34 (4), 339–389. Gotovac, S., Honda, H., Hattori, Y., Takahashi, K., Kanoh, H., Kaneko, K., 2007. Effect of nanoscale curvature of single-walled carbon nanotubes on adsorption of polycyclic aromatic hydrocarbons. Nano Lett. 7 (3), 583–587. Gottschalk, F., Sonderer, T., Scholz, R.W., Nowack, B., 2009. Modeled environmental concentrations of engineered nanomaterials (TiO2, ZnO, Ag, CNT, Fullerenes) for different regions. Environ. Sci. Technol. 43 (24), 9216–9222. Gottschalk, F., Sun, T., Nowack, B., 2013. Environmental concentrations of engineered nanomaterials: review of modeling and analytical studies. Environ. Pollut. (Barking, Essex: 1987) 181, 287–300. Hüffer, T., Kah, M., Hofmann, T., Schmidt, T.C., 2013. How redox conditions and irradiation affect sorption of PAHs by dispersed fullerenes (nC60). Environ. Sci. Technol. 47 (13), 6935–6942. Hüffer, T., Endo, S., Metzelder, F., Schroth, S., Schmidt, T.C., 2014. Prediction of sorption of aromatic and aliphatic organic compounds by carbon nanotubes using poly-parameter linear free-energy relationships. Water Res. 59, 295–303. Iijima, S., 1991. Helical microtubules of graphitic carbon. Nature 354 (6348), 56–58. Ji, L.L., Chen, W., Zheng, S.R., Xu, Z.Y., Zhu, D.Q., 2009. Adsorption of sulfonamide antibiotics to multiwalled carbon nanotubes. Langmuir 25 (19), 11608–11613. Kah, M., Zhang, X., Jonker, M.T.O., Hofmann, T., 2011. Measuring and modelling adsorption of PAHs to carbon nanotubes over a six order of magnitude wide concentration range. Environ. Sci. Technol. 45 (14), 6011–6017. Kah, M., Zhang, X., Hofmann, T., 2014. Sorption behavior of carbon nanotubes: changes induced by functionalization, sonication, and organic matter. Sci. Total Environ. 497–498, 133–138. Koelmans, A.A., Nowack, B., Wiesner, M.R., 2009. Comparison of manufactured and black carbon nanoparticle concentrations in aquatic sediments. Environ. Pollut. 157 (4), 1110–1116. Mayer, P., Holmstrup, M., 2008. Passive dosing of soil invertebrates with polycyclic aromatic hydrocarbons: limited chemical activity explains toxicity cutoff. Environ. Sci. Technol. 42 (19), 7516–7521. Onjia, A.E., Milonjic, S.K., Rajakovic, L.V., 2001. Inverse gas chromatography of chromia. Part I. Zero surface coverage. J. Serb. Chem. Soc. 66 (4), 259–271. Pan, B., Xing, B.S., 2008. Adsorption mechanisms of organic chemicals on carbon nanotubes. Environ. Sci. Technol. 42 (24), 9005–9013. Pan, P., Zhang, D., Li, H., Wu, M., Wnag, Z., Xing, B.S., 2013. Increased adsorption of sulfamethoxazole on suspended carbon nanotubes by dissolved humic acid. Environ. Sci. Technol. 47 (14), 7722–7728. Smiciklas, I.D., Milonjic, S.K., Zec, S., 2000. An inverse gas chromatographic study of the adsorption of alkanes on hydroxyapatite. J. Mater. Sci. 35 (11), 2825–2828. Sun, H.W., Song, Q., Luo, P., Wu, W.L., Wu, J.Z., 2013. Sorption of phenanthrene on single-walled carbon nanotubes modified by DOM: effects of DOM molecular weight and contact time. Environ. Sci.-Process. Impacts 15 (1), 307–314. Velzeboer, I., Kwadijk, C., Koelmans, A.A., 2014. Strong sorption of PCBs to nanoplastics, microplastics, carbon nanotubes, and fullerenes. Environ. Sci. Technol. 48 (9), 4869–4876. Wang, X.L., Lu, J.L., Xing, B.S., 2008. Sorption of organic contaminants by carbon nanotubes: influence of adsorbed organic matter. Environ. Sci. Technol. 42 (9), 3207–3212. Wang, X.L., Tao, S., Xing, B.S., 2009. Sorption and competition of aromatic compounds and humic acid on multiwalled carbon nanotubes. Environ. Sci. Technol. 43 (16), 6214–6219.
T. Hüffer et al. / Chemosphere 119 (2015) 1169–1175 Wang, X., Shu, L., Wang, Y., Xu, B., Bai, Y., Tao, S., Xing, B., 2011. Sorption of peat humic acids to multi-walled carbon nanotubes. Environ. Sci. Technol. 45 (21), 9276–9283. Yang, K., Wang, X.L., Zhu, L.Z., Xing, B.S., 2006. Competitive sorption of pyrene, phenanthrene, and naphthalene on multiwalled carbon nanotubes. Environ. Sci. Technol. 40 (18), 5804–5810. Yang, K., Zhu, L.Z., Xing, B.S., 2006. Adsorption of polycyclic aromatic hydrocarbons by carbon nanomaterials. Environ. Sci. Technol. 40 (6), 1855–1861. Zhang, S., Shao, T., Bekaroglu, S.S.K., Karanfil, T., 2009. The impacts of aggregation and surface chemistry of carbon nanotubes on the adsorption of synthetic organic compounds. Environ. Sci. Technol. 43 (15), 5719–5725. Zhang, S.J., Shao, T., Bekaroglu, S.S.K., Karanfil, T., 2010. Adsorption of synthetic organic chemicals by carbon nanotubes: effects of background solution chemistry. Water Res. 44 (6), 2067–2074.
1175
Zhang, S.J., Shao, T., Karanfil, T., 2011. The effects of dissolved natural organic matter on the adsorption of synthetic organic chemicals by activated carbons and carbon nanotubes. Water Res. 45 (3), 1378–1386. Zhang, X., Kah, M., Jonker, M.T.O., Hofmann, T., 2012. Dispersion state and humic acids concentration-dependent sorption of pyrene to carbon nanotubes. Environ. Sci. Technol. 46 (13), 7166–7173. Zhao, J., Wang, Z.Y., Mashayekhi, H., Mayer, P., Chefetz, B., Xing, B.S., 2012. Pulmonary surfactant suppressed phenanthrene adsorption on carbon nanotubes through solubilization and competition as examined by passive dosing technique. Environ. Sci. Technol. 46 (10), 5369–5377.
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Influence of humic acids on sorption of alkanes by carbon nanotubes – Implications for the dominant sorption mode Thorsten Hüffer a,b,c,⇑, Sarah Schroth a, Torsten C. Schmidt a,b a
Instrumental Analytical Chemistry, University of Duisburg-Essen, Universitätsstrasse 5, 45141 Essen, Germany Centre for Water and Environmental Research (ZWU), University of Duisburg-Essen, Universitätsstrasse 2, 45141 Essen, Germany c Department of Environmental Geosciences, University of Vienna, Althanstrasse 14, 1090 Vienna, Austria b
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Sorption isotherms of alkanes by
carbon nanotubes were determined. Sorption of n-alkanes was generally
stronger than sorption of their cyclic homologs. Addition of humic substances led to decreasing sorption affinity and increasing sorption linearity. The dominant sorption mode by carbon nanotubes remained adsorption independent from the addition of humic substances.
a r t i c l e
i n f o
Article history: Received 29 April 2014 Received in revised form 24 September 2014 Accepted 30 September 2014 Available online 31 October 2014 Handling Editor: Keith Maruya Keywords: Carbon nanomaterials Adsorption vs. absorption Organic compounds Humic substances
a b s t r a c t The presence of humic substances (HS) has previously been shown to alter sorption properties of multi-walled carbon nanotubes (MWCNTs). To systematically study this process, three alkane pairs were selected as molecular probe sorbates. The influence of HS on sorption affinity, sorption linearity, and the dominant sorption mode (i.e., ad- or absorption) by MWCNTs was investigated. The addition of HS led to a continuous decrease in sorption affinity and an increase in sorption linearity with increasing HS addition. Furthermore, the comparison of distribution coefficients of n- and cycloalkanes showed that the dominant sorption mode remains to be adsorption regardless of the presence of HS on MWCNT surface. From this, it can be concluded that instead of a change in sorption mode to absorption of sorbates into HS-coated MWCNT, HS blocks high-energy sorption sites for subsequently added sorbates and that sorbates continuously sorb on the MWCNT surface. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Since the discovery of carbon-based nanomaterials (CNM) in the early 1990s (Iijima, 1991), the research and development of ⇑ Corresponding author at: Department of Environmental Geosciences, University of Vienna, Althanstrasse 14, 1090 Vienna, Austria. Tel.: +43 (1) 4277 53383; fax: +43 (1) 4277 9533. E-mail address: [email protected] (T. Hüffer). http://dx.doi.org/10.1016/j.chemosphere.2014.09.097 0045-6535/Ó 2014 Elsevier Ltd. All rights reserved.
CNM have grown steadily due to their unique physicochemical properties. Together with an expected growth in production, the question of the environmental impact of CNM has become a controversially discussed issue among researchers. To date, most data on environmental concentrations of CNM are still based on modeling (Koelmans et al., 2009; Gottschalk et al., 2009) and although engineered CNM concentrations were predicted to be lower compared to naturally occurring nanomaterials (Koelmans et al.,
1170
T. Hüffer et al. / Chemosphere 119 (2015) 1169–1175
2009), their concentrations are expected to increase in the future (Gottschalk et al., 2013). Various environmentally relevant parameters have been shown to influence the sorption behavior by CNM, such as the presence of oxygen and sunlight (Hüffer et al., 2013) or natural organic matter (NOM) (Wang et al., 2008; Zhang et al., 2011, 2012; Velzeboer et al., 2014; Kah et al., 2014). Generally, two mechanisms have been proposed on how NOM affects HOC sorption by CNM. Firstly, NOM directly competes with HOC for sorption sites on CNM surface. This seems especially relevant as commonly low sorbate and high NOM concentrations have been combined (Wang et al., 2008), which represents a realistic environmental scenario. Secondly, sorption of HOC is reduced by large NOM molecules that block the entrances of CNM pores that may otherwise be available for HOC sorption (Wang et al., 2009; Zhang et al., 2011). Recent investigations on the influence of NOM addition on HOC sorption have shown inconsistent effects on sorption strength of multi-walled carbon nanotubes (MWCNTs). A decrease in sorption strength of NOM-coated MWCNTs compared to the pristine tubes was assigned to an offset between ‘‘creating’’ new sorption sites due to increased dispersity and the reduced accessibility of polar moieties (Wang et al., 2008). In contrast, the contribution of an increase in sorption due to higher MWCNT dispersion was very recently shown to overcome the observed decreased sorption due to NOM coating (Pan et al., 2013). In addition, sorption became more linear after NOM addition. This was principally related to the presence of a co-solute, which reduces sorption of the primary solute (i.e., HOC) and thus increases sorption linearity (Zhang et al., 2011). For example, Zhang et al. studied the effect of NOM addition on sorption of phenanthrene, biphenyl, and phenylphenol by various carbon nanotubes (CNTs). From experimental data, sorption became more linear after NOM addition, while the largest change in sorption linearity was obtained for preloaded CNTs (Zhang et al., 2011). An important question in the phase transfer of HOC is to clarify the dominant sorption mode. Here, generally two sorption modes are discussed: one is the adsorption on the surface of a sorbent; the other is absorption, which is the distribution between two condensed phases. To clarify the sorption mode, Endo et al. proposed an approach to differentiate between ad- and absorption, in which the distribution coefficients of linear alkanes (Kn) and their cyclic homologs (Kc) were compared as follows (Endo et al., 2008):
K n =K c ¼ ðK d =K aw of n-alkaneÞ=ðK d =K aw of cyclo-alkaneÞ
ð1Þ
where Kd is the sorbent–water distribution coefficient and Kaw is the air–water partitioning constant. This comparison of the distribution coefficients of n- and cyclo-alkanes is based on difference in sorption energies of alkanes. First, linear alkanes preferably adsorb on a surface since all carbon atoms of planar n-hexane being able to interact with a surface while for non-planar cyclohexane only three or four carbon atoms can interact (Smiciklas et al., 2000; Onjia et al., 2001). Second, the partitioning (i.e., absorption) from air into a bulk phase of cycloalkanes is stronger than of their linear homologs due to a smaller cavity required for cycloalkanes (Goss, 2004). Note that Kn and Kc cannot be directly compared as experimentally derived Kd values additionally include molecular interactions with water, and for the characterization of the dominant sorption mode by a given sorbent, only the molecular interactions in/on the sorbent are of interest. Given Eq. (1), Kn/Kc ratios < 1 indicate absorption as the dominant sorption mode. Kn/Kc ratio 1 suggest adsorption if steric effects are negligible, while for Kn/Kc > 1 steric hindrance, e.g., from size exclusion effects, has to be considered (Endo et al., 2009). Recently, sorption of n-alkanes by MWCNTs was shown to be stronger than sorption of cycloalkanes (Hüffer et al., 2014). Thus, adsorption seems to be the dominant sorption mode for MWCNTs. Endo et al. observed that for soot, commonly regarded as an adsorber, the predominant sorption mode changed from absorption to
adsorption accompanied by an decrease in sorption linearity following the removal of extractable organic matter (Endo et al., 2009). Thus, the previously observed increase in sorption linearity by CNM following NOM addition may result from a change in sorption mode from ad- to absorption. To this end, the here presented study investigated the dominant sorption mode by MWCNTs influenced by the presence of various types and concentrations of humic substances (HS).
2. Materials and methods 2.1. Materials Multi-walled carbon nanotubes (C150HP) were purchased from Bayer Material Science (Leverkusen, Germany). Properties of MWCNTs are given in Supplemental Table S1. Four humic substances (HS) were selected (Supplemental Table S2 lists HS and selected properties). Stock solutions were prepared by dissolving 0.25–1 g HS in 100 mL 0.1 mol NaOH overnight. After filtration through a 0.45 lm cellulose membrane (Whatman), the solution was diluted to 1 L and the pH was adjusted to neutral with 0.1 mol NaCl. The concentrations of HS stock solutions were determined by total organic carbon content (TOC; Shimadzu TOC-5000, Duisburg, Germany). HS stock solutions were kept at 4 °C in the dark for no longer than six weeks and TOC concentrations were regularly monitored. Working standard HS solutions were obtained by diluting stock solutions accordingly to concentrations of 1 and 10 mg L1. Three pairs of n- and cycloalkanes (C6–C8) were selected as molecular probe compounds (all purchased from Sigma–Aldrich with >99% purity). Stock solutions of alkanes were freshly prepared in methanol (HPLC grade, Fisher Scientific, UK) every week. Selected physico-chemical properties of sorbates are displayed in Supplemental Table S3.
2.2. Sorption batch experiments Sorption isotherms were determined in a multi-phase system. 2–5 mg of MWCNTs were weighted into 20-mL amber glass vials. 10 or 15 mL of background solution were added. The background solution contained either 10 mmol CaCl2 to regulate the ionic strength or the HS solution in concentrations as given in Supplemental Table S2. Vials were then horizontally shaken for 24 h to allow the MWCNTs to be coated with HS. The samples were spiked with methanolic stock solutions resulting in 12 different sorbate concentrations ranging over 3–4 orders of magnitude. Highest equilibrium aqueous concentrations of sorbates were 20 times below their aqueous solubility. The methanol content did not exceed 0.5 vol% in water to minimize co-solvent effects. The samples were shaken for at least 24 h for equilibration at 25 °C in a temperature-controlled room. Equilibration times were pre-determined (data not shown). Subsequently, the vials were removed from the shaker and placed in the tray of the autosampler for 1 h to allow air–water equilibration at 20 °C. The analyte concentrations in the headspace of the vials were determined by direct injection of 1000 lL of the vial headspace into GC–MS. External calibration vials were prepared one day prior to analysis and contained the according background solutions. For quantification, a gas chromatography-mass spectrometry system (TraceGC Ultra/ DSQ, Thermo Finnigan) equipped with a PAL Combi-xt autosampler (CTC Analytics) was used. An HP-5 capillary column (30 m 0.32 mm 0.25 lm, J&W Scientific) and oven temperatures between 70 and 150 °C were used depending on the analyte retention times.
1171
T. Hüffer et al. / Chemosphere 119 (2015) 1169–1175
Sorbed concentrations could be determined by mass balance approach using air–water partitioning constants (Kaw) of the analytes (see Supplemental Table S3):
C MWCNTs ¼
Vw mT CKaaw CaV a
MMWCNTs
ð2Þ
where Ca and CMWCNTs are the alkane concentrations in air (lg L1) and sorbed by MWCNTs (lg kg1), respectively; Va and Vw denote the vial volume of air (L) and background solution (L), respectively; and Mw is the mass of MWCNTs added to the vials (kg). Note that sorption of alkanes by HS was pre-determined and found to be negligible compared to sorption by MWCNTs, which is in accordance to previous reports (Chen et al., 2008; Zhang et al., 2010). Thus, ‘extrinsic’ sorption by HS and MWCNTs as a whole sorbent was considered since the HS/MWCNTs mass ratio used in the present was rather low ranging from 0.03 to 0.075 g HS/g MWCNTs (Zhang et al., 2012). Kaw values were carefully selected from literature to ensure accurate calculation of Kd values. Loss of analytes during the experiments was exemplarily determined for n-heptane, where after a six days shaking period, a mean loss of 6% was obtained (n = 5); and according corrections were made in the mass balance. 2.3. Determination of sorption mode The Freundlich model (FM) was used to fit the isotherm data:
C MWCNTs ¼ K F ðC w Þn
ð3Þ
where Cw is the aqueous concentration (lg L1), and KF and n are the Freundlich coefficient ((lg kg1)/(lg L1)n) and the Freundlich exponent [], respectively. Sigma Plot 12.0 for Windows was used for statistical analyses and model fits. Computation of model parameters and their standard errors was performed with reduced Chi-square and weighted regression with weighting to reciprocal y2. Single point distribution coefficients (K d ¼ C MWCNTs =C w ; (L kg1)) were determined by interpolation at given sorbate air concentration. All determined isotherms were non-linear; thus, Kd and Kn/Kc ratios may vary with concentration. This concentration-dependency can be accounted for by comparing the data at standardized chemical activity (e.g., by passive dosing (Mayer and Holmstrup, 2008) or reference concentration measures (Endo et al., 2009). As pointed out by Endo et al., calculating Kn/Kc ratios with Kd values occurring at the same Cair as reference concentration, may be the most appropriate (Endo et al., 2009). Kd values were hence calculated for an air-phase concentration of 10 lg L1. Propagated errors in Kn/Kc ratios were determined using Gaussian error propagation approach. 3. Results and discussion 3.1. Sorption isotherms Isotherm data were well fit with the Freundlich model (R > 0.95). It has been previously discussed that other, more complex, non-linear sorption models are more suitable for a robust interpretation of sorption isotherm data especially if large concentration ranges are considered (Kah et al., 2011). However, the aim of this study was to investigate how the presence of HS affects sorption by MWCNTs in terms of sorption affinity and sorption linearity; therefore, only the Freundlich model was taken into account for isotherm data fitting. The results of the Freundlich model fits are summarized in Table 1. Sorption isotherms of n- and cycloalkanes by MWCNTs following various kinds and concentrations of HS are shown in Fig. 1. In 2
general, sorption isotherms by all sorbent scenarios were parallel accompanied by an increase in sorption with increasing number of carbon atoms (i.e., C6 < C7 < C8). This observation is reflected by a corresponding increase of sorbate hydrophobicity parameters, such as an increase in the hexadecane–air partitioning constant (L) as displayed in Supplemental Table S3. However, it must be stressed that L refers to absorption processes, which may limit its comparability to adsorption by CNM. Sorption of n-alkanes was always up to one order of magnitude stronger compared to sorption of their cyclic homologous. The sorbate morphology (i.e., molecular size and shape) plays a major role with respect to availability of sorption sites on the CNM surface. Stronger sorption was previously explained by a better contact of linear and planar compounds with the CNM surface (Gotovac et al., 2007). Similar conclusions have been drawn for alkane sorption by flat surfaces, where linear n-hexane can interact with its six carbon atoms with the sorbent surface, while non-planar cyclohexane can only do so with three or four carbon atoms due to its non-planar shape (Smiciklas et al., 2000; Onjia et al., 2001). These differences in number of carbon atoms in contact with the CNM surface are also likely to be the case for n- and cyclooctane (Dorofeeva et al., 1985).
Table 1 Fitting parameters ± standard errors, R2, standard error of estimates (SEE) for Freundlich model fitting the isotherms of C6–C8 alkanes pairs by MWCNTs following HS addition (N: number of data points). Sorbate
KF
n
R2
SEE
N
MWCNTs nHex cHex nHep cHep nOct cOct
1.69E+05 ± 4.48E+03 1.77E+04 ± 9.14E+02 5.03E+05 ± 2.65E+04 4.57E+04 ± 2.89E+03 9.20E+05 ± 2.11E+04 1.57E+05 ± 7.21E+03
0.71 ± 0.02 0.73 ± 0.02 0.66 ± 0.03 0.68 ± 0.02 0.55 ± 0.02 0.55 ± 0.01
0.9951 0.9720 0.9828 0.9922 0.9914 0.9941
0.09 0.15 0.17 0.12 0.08 0.08
11 10 11 10 12 14
MWCNTs+1HA nHex 1.08E+05 ± 6.17E+03 cHex 1.09E+04 ± 5.53E+02 nHep 2.38E+05 ± 8.30E+03 cHep 2.07E+04 ± 1.50E+03 nOct 5.34E+05 ± 1.72E+04 cOct 6.71E+04 ± 4.86E+03
0.79 ± 0.03 0.83 ± 0.02 0.73 ± 0.02 0.78 ± 0.03 0.66 ± 0.02 0.71 ± 0.03
0.9508 0.9732 0.9793 0.9734 0.9908 0.9688
0.20 0.17 0.13 0.15 0.12 0.16
15 13 15 12 14 14
MWCNTs+10HA nHex 5.39E+04 ± 1.60E+03 cHex 5.26E+03 ± 1.97E+02 nHep 8.98E+04 ± 3.28E+03 cHep 9.78E+03 ± 7.44E+02 nOct 1.96E+05 ± 6.94E+03 cOct 2.45E+04 ± 1.69E+03
0.86 ± 0.01 0.89 ± 0.01 0.72 ± 0.02 0.79 ± 0.03 0.71 ± 0.02 0.74 ± 0.02
0.9969 0.9979 0.9794 0.9783 0.9835 0.9888
0.11 0.11 0.14 0.17 0.12 0.17
13 12 15 11 13 14
MWCNTs+LSHA nHex 5.01E+04 ± 2.70E+03 cHex 4.94E+03 ± 5.85E+02 nHep 1.69E+05 ± 8.43E+03 cHep 2.03E+04 ± 1.24E+03 nOct 4.04E+05 ± 1.61E+04 cOct 4.98E+04 ± 3.87E+03
0.83 ± 0.02 0.86 ± 0.04 0.69 ± 0.03 0.72 ± 0.01 0.71 ± 0.02 0.74 ± 0.02
0.9914 0.9513 0.9741 0.9861 0.9841 0.9813
0.17 0.23 0.15 0.11 0.10 0.13
15 14 13 14 14 14
MWCNTs+PPHAS nHex 7.78E+04 ± 7.35E+03 cHex 7.88E+03 ± 5.33E+02 nHep 1.89E+05 ± 5.31E+03 cHep 2.41E+04 ± 1.62E+03 nOct 5.37E+05 ± 2.79E+04 cOct 6.50E+04 ± 4.37E+03
0.82 ± 0.03 0.85 ± 0.02 0.67 ± 0.02 0.73 ± 0.03 0.70 ± 0.02 0.74 ± 0.03
0.9533 0.9698 0.9812 0.9728 0.9628 0.9692
0.22 0.19 0.10 0.15 0.16 0.16
14 15 15 12 12 12
MWCNTs+WPHAR nHex 5.24E+04 ± 2.88E+03 cHex 5.34E+03 ± 3.35E+02 nHep 9.67E+04 ± 2.87E+03 cHep 1.22E+04 ± 8.20E+02 nOct 2.09E+05 ± 5.51E+03 cOct 2.60E+04 ± 9.55E+02
0.84 ± 0.02 0.87 ± 0.03 0.66 ± 0.01 0.72 ± 0.02 0.68 ± 0.01 0.74 ± 0.01
0.9847 0.9716 0.9912 0.9798 0.9965 0.9968
0.13 0.17 0.10 0.16 0.08 0.07
13 13 14 11 12 13
1172
T. Hüffer et al. / Chemosphere 119 (2015) 1169–1175
Fig. 1. Sorption isotherms of (d) nHex, (s) cHex, (.) nHep, (D) cHep, (j) nOct, and (h) cOct by MWCNT (A), MWCNT + 1HA (B), MWCNT + 10HA (C), MWCNT + LSHA (D), MWCNT + PPHAS (E), and MWCNT + WPHAR (F), respectively.
3.2. HS effect on sorption affinity and linearity The sorption affinity (log KF in Table 1) was continuously decreased with increasing concentration of HA, i.e., log KF (MWCNT) > log KF (MWCNT + 1HA) > log KF (MWCNT + 10HA). This decrease in log KF was significant (p < 0.05) except for cHex. Overall, the addition of 10 mg L1 HA led to a decrease in sorption affinity ranging between 0.5 log units for nHex and 0.8 log units for cOct. Zhang et al. observed a clear linear decrease in sorption affinity of MWCNTs for pyrene following Suwannee River humic acid addition ranging from 0 to 200 mg L1; however, the obtained decrease in KF was smaller than the almost one order of magnitude effect observed here (Zhang et al., 2012).
It has previously been suggested that the addition of NOM leads a reduction in sorption affinity due to the occupation of sorption sites by NOM (Chen et al., 2008; Wang et al., 2009; Zhang et al., 2012; Velzeboer et al., 2014). This seems to be the case for the here investigated sorption scenarios, where sorption by MWCNTs was reduced following the addition of HS. Furthermore, a decrease in sorption affinity for HOC may be explained by a decrease in hydrophobicity of CNTs due to the introduction of polar moieties to CNTs surface from NOM addition. This facilitates the formation of water clusters and consequently reduces the interaction with HOC due to a decrease in hydrophobicity of CNT surface (Sun et al., 2013). In contrast, an increase in sorption has been reported to occur due to a concomitant increase in available sorption sites following
T. Hüffer et al. / Chemosphere 119 (2015) 1169–1175
NOM addition (Wang et al., 2008; Gai et al., 2011). This was additionally assigned to the destabilization of CNM aggregates, e.g., resulting from sonication (Pan et al., 2013). Extensive characterization data by Zhang et al. suggest that there is no change in aggregation from the presence of NOM (Zhang et al., 2012). Furthermore, Kah et al. recently showed that the presence of NOM efficiently dispersed CNTs only when combined with a sonication treatment (Kah et al., 2014). In the present study, CNT dispersions were not sonicated; therefore, changes in aggregation between bulk sorbent material and HS-coated CNTs were not expected and differences in sorption by CNTs do not seem to result from a changed aggregation status of the sorbent. However, data on dispersion stability of CNM affected by NOM remain limited. Previous investigations have shown that the strength of interaction between CNTs and HOC depends also on sorbent structural properties. The effect of nanoscale curvature of the sorbent has been discussed for sorption of polycyclic aromatic hydrocarbons (PAHs) by single-walled carbon nanotubes (SWCNTs) Gotovac et al., 2007. However, this effect is expected to play a minor role in the present study since the investigated sorbates were comparably smaller. In addition, MWCNTs have a much larger diameter of >13 nm compared to 1 nm in diameter of single-walled nanotubes. Decreases in sorption by MWCNTs and graphite have been explained by structural differences in the sorbent material. The interstitial and groove areas of CNT aggregates may not be available for sorption of large humic substance molecules (Ji et al., 2009), while these sites remain available for sorption of comparably small alkane molecules after HS addition. Non-linear sorption (n = 0.55–0.79) was observed for all sorbates by MWCNTs without HS addition. This is generally consistent with recent investigations of sorption of non-ionic organic compounds (Yang et al., 2006; Kah et al., 2011). Commonly non-linear sorption by CNM is explained by a heterogeneous distribution of free-energy of sorption sites caused by e.g., surface defects or interstitial and groove areas of CNT aggregates (Pan and Xing, 2008). Moreover, aggregation of CNTs may lead to a heterogeneous distribution of sorption site energies (Zhang et al., 2009). The addition of 1 and 10 mg L1 HA led to a significant increase in sorption linearity (p < 0.05) with Dn ranging between 0.08 and 0.19 for sorption of nHep and cOct, respectively. Previous studies reported similar results regarding the impact of NOM on sorption linearity of HOC by CNTs (Zhang et al., 2011; Sun et al., 2013) and the increase in sorption linearity was recently explained by a more homogeneous energy distribution of sorption sites on NOM-coated CNM (Yang et al., 2006; Velzeboer et al., 2014). However, an increase in sorption linearity has not been consistently reported in literature as in some studies no change in sorption linearity was observed following the addition of NOM (Wang et al., 2008; Chen et al., 2008). Endo et al. investigated sorption properties by carbonaceous materials and found that for diesel soot commonly regarded as an adsorbing material sorption of n- and cyclooctane was linear (Endo et al., 2009). Furthermore, the extraction of organic material from reference diesel soot resulted in a significant decrease in sorption linearity of n = 0.75 and 0.85 for n- and cyclooctane, respectively (Endo et al., 2009). The here observed decrease in
Table 2 Kn/Kc ratios and errors of C6–C8 alkane pairs for sorbents. Sorbent
C6
C7
C8
MWCNT MWCNT + 1HA MWCNT + 10HA MWCNT + LSHA MWCNT + PPHAS MWCNT + WPHAR
2.01 ± 0.03 1.72 ± 0.03 1.55 ± 0.02 1.62 ± 0.03 1.62 ± 0.12 1.55 ± 0.07
1.73 ± 0.11 1.43 ± 0.03 1.17 ± 0.02 1.17 ± 0.07 1.16 ± 0.03 1.18 ± 0.03
0.96 ± 0.03 0.84 ± 0.03 0.72 ± 0.03 0.74 ± 0.02 0.78 ± 0.05 0.79 ± 0.03
1173
sorption affinity (KF) accompanied by the increasing sorption linearity implies that strong sorption sites on MWCNTs surface are occupied by HS, which (i) makes them no longer available for interaction and (ii) results in a more homogeneous distribution of sorption site energies and thus more linear sorption. The latter could also result from a change of sorption mode to absorption of alkanes into HS-coated on MWCNTs surface, which will be discussed in the next section. 3.3. Kn/Kc ratios for sorption by MWCNTs Kn/Kc ratios for sorption of C6–C8 alkane pairs by MWCNTs under varying HS addition conditions are given in Table 2. For sorption by MWCNTs, a Kn/Kc ratio for the C6-alkane pair of 2.01 was obtained. This is generally comparable to previous reports for similar carbonaceous sorbent materials, where Kn/Kc ranged between of 1.02 and 2.20 for carbon nanofibers and activated carbon, respectively (Endo et al., 2008). From literature data for sorption of organic vapors by single-walled carbon nanotubes, a C6-Kn/Kc ratio of 1.08 can be calculated (Agnihotri et al., 2005). Kn/Kc ratios of C6 in Table 2 consistently decreased to 1.72 and 1.55 following 1 and 10 mg L1 HA addition, respectively. Overall, a significant decrease (p < 0.05) in Kn/Kc ratios was observed for all alkane pairs following HS addition. However, Kn/Kc ratios of C6 and C7 alkane pairs for all sorbent scenarios remain larger than 1, which suggests adsorption to be the dominant sorption mode regardless of HS addition. Kn/Kc ratios of MWCNTs coated with the reference HS materials (i.e., LSHA, PPHAS, and WPHAR) were not significantly different (p > 0.05) from those obtained for non-certified HS material (i.e., HA), suggesting that the effect of HS on the dominant sorption mode by MWCNTs is independent from the properties of HS. In order to further investigate a potential change in sorption mode following HS addition, C8-Kn/Kc ratios were plotted against varying sorbate air-phase concentrations for all sorbents scenarios (see Supplemental Fig. S1). Sorption of n- and cyclooctane by MWCNTs shows Kn/Kc 1 over the whole concentration range, which has also been observed for other rigid carbonaceous materials, such as charcoal and activated carbon (Endo et al., 2009). This suggests that adsorption is the dominant sorption mode regardless of sorbate concentration. The addition of HS in general led to a decrease in Kn/Kc with increasing sorbate concentration. This was similarly observed for carbonaceous geosorbents containing native organic matter, e.g., soot (Endo et al., 2009), for which sorption was furthermore explained by a combination of ad- and absorption (Endo et al., 2009). It must be stressed that this effect was far more pronounced with decreasing Kn/Kc ratios from 1 to 0.3 or lower for soils (Endo et al., 2009). In the present study Kn/Kc ratios decreased from 1 to 0.6; thus, the contribution of absorption by MWCNTs following HS addition does not seem to be of major importance. From the results in Table 2, a significant decrease (p < 0.05) in Kn/Kc ratios with increasing number of sorbate carbon atoms was observed. These changes are the highest observed so far for the Kn/Kc-concept. Despite the fact that the lowest Kn/Kc ratios were obtained for sorption of n- and cyclooctane by MWCNT + 10HA, adsorption can be assumed to be the dominant sorption mode as Kn/Kc of 0.72 is above values obtained for typical absorbers, such as bulky liquids (e.g., C8-Kn/Kc of 0.23 for 1-octanol) or organic polymers (e.g., C8-Kn/Kc of 0.33 for polypropylene) (Endo et al., 2008). Overall, the observed effects of HS addition on Kn/Kc were not as pronounced as in previous reports, where the removal of extractable organic matter from a standard reference material of diesel soot clearly resulted in a change in sorption mode from ab- to adsorption (Endo et al., 2009). In fact, Kn/Kc ratios for sorption of
1174
T. Hüffer et al. / Chemosphere 119 (2015) 1169–1175
n- and cyclooctane by the untreated sorbent ranged between 0.46 and 0.66, while sorption by NOM-extracted sorbent material showed Kn/Kc values between 0.83 and 1.72. The here observed changes in Kn/Kc support the argument that the dominant sorption mode by HS-coated MWCNTs remains adsorption by the nanotubes rather than absorption into the attached organic matter. Aside from sorption onto the surface, pore-filling has been discussed as one of the major sorption mechanisms by CNM (Pan and Xing, 2008; Kah et al., 2011; Hüffer et al., 2013), which was especially relevant for low sorbate concentrations because of the high energetic sorption sites in the micropores of CNTs (Wang et al., 2008). Meanwhile, it was suggested by Wang et al. that fractionation of HS components by MWCNTs occurs dependent on the molecular weight and composition of HS and on the porosity of MWCNTs (Wang et al., 2011). Furthermore, steric effects may have to be considered as the sorption capacity of PCBs and PAHs by MWCNTs was negatively related to the sorbates’ molecular size (Yang et al., 2006; Velzeboer et al., 2014). Commonly, the inner pores of nanotubes are not available for sorption (Pan and Xing, 2008), so any discussion on the impact of pore size on sorption is restricted to potential pores of CNT aggregates. Microporous materials may exhibit size exclusion effects as pores may not be available for larger sorbates than the pore sizes. This led to C6-Kn/Kc ratios of up to 1500 for graphitized or activated carbons (Endo et al., 2008). The volume of micro- and mesopores of MWCNTs applied in this study is given in Supplemental Table S1. Since Kn/Kc ratios for the here investigated sorbent scenarios were smaller, the occurrence of size exclusion effects may be neglected. The addition of HS seems to have the following impact on sorption behavior by MWCNTs: adsorption remains to be the dominant sorption mode regardless of HS addition and size exclusion effects from pore blockage by HS may be neglected, which is further supported by the fact that C8-Kn/Kc ratios shown in Supplemental Fig. S1 remain constantly 1. In contrast, Kn/Kc ratios decreased from >1 to 1 with increasing sorbate concentration for lignite coke and activated carbon, where steric effects at low sorbate concentration were explained by higher availability of strong interaction sites for n- than for cyclooctane (Endo et al., 2009). Moreover, the humic substance concentrations were not sufficient to significantly sorb alkane sorbates, while they significantly affected sorption of alkanes by MWCNTs. Thus, competitive sorption of alkanes by MWCNTs rather than HS seems to dominate over an enhanced solubilization of sorbates from sorption by HS (i.e., ‘‘solubilizing effect’’). Previously, the suppression of sorption of PAHs by MWCNTs in the presence of pulmonary surfactants was shown to be a combination of both solubilization and competitive sorption (Zhao et al., 2012).
4. Conclusions The addition of humic substances affected sorption affinity and linearity of n- and cyclo-alkanes by multi-walled carbon nanotubes: sorption affinity was decreased while sorption linearity was increased. The ratio of distribution coefficients of n- and cycloalkanes suggested that adsorption is the dominant sorption mode by MWCNTs, which was not altered by the presence of humic substances. Acknowledgements The authors would like to thank Dr. Satoshi Endo (UFZ Leipzig, Germany) and Dr. Mélanie Kah (University of Vienna, Austria) for their helpful contributions. This work was financially supported by the Deutsche Forschungsgemeinschaft (DFG, SCHM 1372/10-1).
Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.chemosphere. 2014.09.097. References Agnihotri, S., Rood, M.J., Rostam-Abadi, M., 2005. Adsorption equilibrium of organic vapors on single-walled carbon nanotubes. Carbon 43 (11), 2379–2388. Chen, J.Y., Chen, W., Zhu, D., 2008. Adsorption of nonionic aromatic compounds to single-walled carbon nanotubes: effects of aqueous solution chemistry. Environ. Sci. Technol. 42 (19), 7225–7230. Dorofeeva, O.V., Mastryukov, V.S., Allinger, N.L., Almenningen, A., 1985. The molecular-structure and conformation of cyclooctane as determined by electron-diffraction and molecular mechanics calculations. J. Phys. Chem. 89 (2), 252–257. Endo, S., Grathwohl, P., Schmidt, T.C., 2008. Absorption or adsorption? Insights from molecular probes n-alkanes and cycloalkanes into modes of sorption by environmental solid matrices. Environ. Sci. Technol. 42 (11), 3989–3995. Endo, S., Grathwohl, P., Handerlein, S.B., Schmidt, T.C., 2009. Characterization of sorbent properties of soil organic matter and carbonaceous geosorbents using n-alkanes and cycloalkanes as molecular probes. Environ. Sci. Technol. 43 (2), 393–400. Endo, S., Grathwohl, P., Haderlein, S.B., Schmidt, T.C., 2009. Effects of native organic material and water on sorption reference diesel soot. Environ. Sci. Technol. 43 (9), 3187–3193. Gai, K., Shi, B., Yan, X., Wang, D., 2011. Effect of dispersion on adsorption of atrazine by aqueous suspensions of fullerenes. Environ. Sci. Technol. 45 (14), 5959– 5965. Goss, K.U., 2004. The air/surface adsorption equilibrium of organic compounds under ambient conditions. Crit. Rev. Environ. Sci. Technol. 34 (4), 339–389. Gotovac, S., Honda, H., Hattori, Y., Takahashi, K., Kanoh, H., Kaneko, K., 2007. Effect of nanoscale curvature of single-walled carbon nanotubes on adsorption of polycyclic aromatic hydrocarbons. Nano Lett. 7 (3), 583–587. Gottschalk, F., Sonderer, T., Scholz, R.W., Nowack, B., 2009. Modeled environmental concentrations of engineered nanomaterials (TiO2, ZnO, Ag, CNT, Fullerenes) for different regions. Environ. Sci. Technol. 43 (24), 9216–9222. Gottschalk, F., Sun, T., Nowack, B., 2013. Environmental concentrations of engineered nanomaterials: review of modeling and analytical studies. Environ. Pollut. (Barking, Essex: 1987) 181, 287–300. Hüffer, T., Kah, M., Hofmann, T., Schmidt, T.C., 2013. How redox conditions and irradiation affect sorption of PAHs by dispersed fullerenes (nC60). Environ. Sci. Technol. 47 (13), 6935–6942. Hüffer, T., Endo, S., Metzelder, F., Schroth, S., Schmidt, T.C., 2014. Prediction of sorption of aromatic and aliphatic organic compounds by carbon nanotubes using poly-parameter linear free-energy relationships. Water Res. 59, 295–303. Iijima, S., 1991. Helical microtubules of graphitic carbon. Nature 354 (6348), 56–58. Ji, L.L., Chen, W., Zheng, S.R., Xu, Z.Y., Zhu, D.Q., 2009. Adsorption of sulfonamide antibiotics to multiwalled carbon nanotubes. Langmuir 25 (19), 11608–11613. Kah, M., Zhang, X., Jonker, M.T.O., Hofmann, T., 2011. Measuring and modelling adsorption of PAHs to carbon nanotubes over a six order of magnitude wide concentration range. Environ. Sci. Technol. 45 (14), 6011–6017. Kah, M., Zhang, X., Hofmann, T., 2014. Sorption behavior of carbon nanotubes: changes induced by functionalization, sonication, and organic matter. Sci. Total Environ. 497–498, 133–138. Koelmans, A.A., Nowack, B., Wiesner, M.R., 2009. Comparison of manufactured and black carbon nanoparticle concentrations in aquatic sediments. Environ. Pollut. 157 (4), 1110–1116. Mayer, P., Holmstrup, M., 2008. Passive dosing of soil invertebrates with polycyclic aromatic hydrocarbons: limited chemical activity explains toxicity cutoff. Environ. Sci. Technol. 42 (19), 7516–7521. Onjia, A.E., Milonjic, S.K., Rajakovic, L.V., 2001. Inverse gas chromatography of chromia. Part I. Zero surface coverage. J. Serb. Chem. Soc. 66 (4), 259–271. Pan, B., Xing, B.S., 2008. Adsorption mechanisms of organic chemicals on carbon nanotubes. Environ. Sci. Technol. 42 (24), 9005–9013. Pan, P., Zhang, D., Li, H., Wu, M., Wnag, Z., Xing, B.S., 2013. Increased adsorption of sulfamethoxazole on suspended carbon nanotubes by dissolved humic acid. Environ. Sci. Technol. 47 (14), 7722–7728. Smiciklas, I.D., Milonjic, S.K., Zec, S., 2000. An inverse gas chromatographic study of the adsorption of alkanes on hydroxyapatite. J. Mater. Sci. 35 (11), 2825–2828. Sun, H.W., Song, Q., Luo, P., Wu, W.L., Wu, J.Z., 2013. Sorption of phenanthrene on single-walled carbon nanotubes modified by DOM: effects of DOM molecular weight and contact time. Environ. Sci.-Process. Impacts 15 (1), 307–314. Velzeboer, I., Kwadijk, C., Koelmans, A.A., 2014. Strong sorption of PCBs to nanoplastics, microplastics, carbon nanotubes, and fullerenes. Environ. Sci. Technol. 48 (9), 4869–4876. Wang, X.L., Lu, J.L., Xing, B.S., 2008. Sorption of organic contaminants by carbon nanotubes: influence of adsorbed organic matter. Environ. Sci. Technol. 42 (9), 3207–3212. Wang, X.L., Tao, S., Xing, B.S., 2009. Sorption and competition of aromatic compounds and humic acid on multiwalled carbon nanotubes. Environ. Sci. Technol. 43 (16), 6214–6219.
T. Hüffer et al. / Chemosphere 119 (2015) 1169–1175 Wang, X., Shu, L., Wang, Y., Xu, B., Bai, Y., Tao, S., Xing, B., 2011. Sorption of peat humic acids to multi-walled carbon nanotubes. Environ. Sci. Technol. 45 (21), 9276–9283. Yang, K., Wang, X.L., Zhu, L.Z., Xing, B.S., 2006. Competitive sorption of pyrene, phenanthrene, and naphthalene on multiwalled carbon nanotubes. Environ. Sci. Technol. 40 (18), 5804–5810. Yang, K., Zhu, L.Z., Xing, B.S., 2006. Adsorption of polycyclic aromatic hydrocarbons by carbon nanomaterials. Environ. Sci. Technol. 40 (6), 1855–1861. Zhang, S., Shao, T., Bekaroglu, S.S.K., Karanfil, T., 2009. The impacts of aggregation and surface chemistry of carbon nanotubes on the adsorption of synthetic organic compounds. Environ. Sci. Technol. 43 (15), 5719–5725. Zhang, S.J., Shao, T., Bekaroglu, S.S.K., Karanfil, T., 2010. Adsorption of synthetic organic chemicals by carbon nanotubes: effects of background solution chemistry. Water Res. 44 (6), 2067–2074.
1175
Zhang, S.J., Shao, T., Karanfil, T., 2011. The effects of dissolved natural organic matter on the adsorption of synthetic organic chemicals by activated carbons and carbon nanotubes. Water Res. 45 (3), 1378–1386. Zhang, X., Kah, M., Jonker, M.T.O., Hofmann, T., 2012. Dispersion state and humic acids concentration-dependent sorption of pyrene to carbon nanotubes. Environ. Sci. Technol. 46 (13), 7166–7173. Zhao, J., Wang, Z.Y., Mashayekhi, H., Mayer, P., Chefetz, B., Xing, B.S., 2012. Pulmonary surfactant suppressed phenanthrene adsorption on carbon nanotubes through solubilization and competition as examined by passive dosing technique. Environ. Sci. Technol. 46 (10), 5369–5377.
