Environmental Pollution 197 (2015) 287e294
Contents lists available at ScienceDirect
Environmental Pollution journal homepage: www.elsevier.com/locate/envpol
Adsorptive fractionation of dissolved organic matter (DOM) by carbon nanotubes Maya Engel, Benny Chefetz* Department of Soil and Water Sciences, Faculty of Agriculture, Food and Environment, The Hebrew University of Jerusalem, P.O. Box 12, Rehovot 7610001, Israel
a r t i c l e i n f o
a b s t r a c t
Article history: Received 1 September 2014 Received in revised form 5 November 2014 Accepted 14 November 2014 Available online 3 December 2014
Dissolved organic matter (DOM) and carbon nanotubes are introduced into aquatic environments. Thus, it is important to elucidate whether their interaction affects DOM amount and composition. In this study, the composition of DOM, before and after interactions with single-walled carbon nanotubes (SWCNTs), was measured and the adsorption affinity of the individual structural fractions of DOM to SWCNTs was investigated. Adsorption of DOM to SWCNTs was dominated by the hydrophobic acid fraction, resulting in relative enhancement of the hydrophilic character of non-adsorbed DOM. The preferential adsorption of the HoA fraction was concentration-dependent, increasing with increasing concentration. Adsorption affinities of bulk DOM calculated as the normalized sum of affinities of the individual structural fractions were similar to the measured affinities, suggesting that the structural fractions of DOM act as independent adsorbates. The altered DOM composition may affect the nature and reactivity of DOM in aquatic environments polluted with carbon nanotubes. © 2014 Elsevier Ltd. All rights reserved.
Keywords: Carbon nanomaterial Fraction Sorption Composition Characterization
1. Introduction Carbon nanotubes (CNTs) are produced by rolling sheets of graphene into a cylinder along a lattice vector in the graphene plane. They consist of sp2-bonded carbon atoms forming a network of aromatic rings. Two main types of CNTs are manufactured: single-walled (SWCNTs) and multi-walled (MWCNTs) (Ajayan et al., 1999). Their nano-scale size and morphology confer unique properties, such as extremely high tensile strength, exceptional electrochemical activity, metallic and semi-conducting features and high thermal conductivity (Ajayan and Zhou, 2001; Niyogi et al., 2002). Thus, a sharp increase in the use of CNTs has been reported for a wide range of applications, such as energy storage, microelectronics, composite materials and biomedicine (Liang and Chen, 2010; Mauter and Elimelech, 2008; Schnorr and Swager, 2011). CNTs are known as toxic materials and their hazardous effect is size-dependent; manufactured and combustion sources of CNTs in the environment were suggested to have adverse effects on human health (Liu et al., 2013).
* Corresponding author. E-mail address: [email protected] (B. Chefetz). http://dx.doi.org/10.1016/j.envpol.2014.11.020 0269-7491/© 2014 Elsevier Ltd. All rights reserved.
Their wide use implies their potential release into the environment (Aschberger et al., 2010; Petersen et al., 2011), where they might interact with organic pollutants (Yang and Xing, 2006). In addition, CNTs have been found to interact with dissolved organic matter (DOM), modifying behavior of both DOM and CNTs (Hyung et al., 2007; Lin and Xing, 2008a; Lin et al., 2009; Sun et al., 2012; Yang and Xing, 2010). To better understand the environmental implications of CNTs release into water bodies, one must study DOM adsorption by CNTs (Wang et al., 2011, 2008; Yang and Xing, 2009) and more importantly, evaluate DOM fractionation during adsorption. Both processes (adsorption and fractionation) affect the quantity and quality of DOM remaining in surface water. DOM is a heterogeneous mixture of soluble aromatic and aliphatic compounds, and is considered the most reactive fraction of natural organic matter (NOM). The aromatic moieties of DOM might include conjugated benzene rings with carbonyl, hydroxyl and amide substituents which impact the electron-donating and accepting properties of the benzene derivatives (Imai et al., 2001; Quails and Haines, 1991). Several studies have examined adsorption of DOM surrogates to CNTs (Li et al., 2014; Lin and Xing, 2008a; Lin et al., 2012), but only few reports have addressed the interactions between naturally occurring DOM and CNTs (Chappell et al., 2009; Hyung et al., 2007; Lu and Su, 2007). For example, tannic acid (used as a DOM surrogate) was reported to stabilize CNT
288
M. Engel, B. Chefetz / Environmental Pollution 197 (2015) 287e294
suspension while adsorption affinity was correlated with CNT surface area. A two-stage adsorption model was proposed: primary adsorption of tannic acid by the CNTs via pep interactions until monolayer formation, and further adsorption of tannic acid to the adsorbed phase by polar interactions (Lin and Xing, 2008a). The adsorption of DOM by MWCNTs was reported to increase with decreasing pH and increasing ionic strength. Moreover, higher molecular weight fractions of DOM-like materials were preferentially adsorbed to MWCNTs (Hyung and Kim, 2008; Wang et al., 2013). Another study reported favorable adsorption of aromaticrich polar fulvic acid fractions of high molecular weight by CNTs (Yang and Xing, 2009). The abovementioned studies suggest that DOM fractionation might occur upon adsorption to CNTs, consequently changing its conformation and composition. However, very little is known about configurational changes and fractionation of native DOM upon adsorption to CNTs (Pan et al., 2008; Yang and Xing, 2010). Thus, the main objective of this study was to elucidate the adsorptive fractionation of DOM with SWCNTs and to quantify the adsorption affinities of the structural fractions of DOM to SWCNTs. Adsorption of DOM by SWCNTs is expected to be governed by pep interactions, and therefore we hypothesize that highly polar fractions of the DOM will be less prone to surface interactions. The latter materials are thus expected to dominate the residual, non-adsorbed DOM, whereas the DOM coating the CNTs is likely to be enriched with higher molecular weight and aromatic-rich fractions. Consequently, the composition of the non-adsorbed DOM is likely to differ from that of the bulk DOM. To check our hypothesis we examined the interactions of two types of naturally occurring DOM at environmentally relevant concentrations (Baghoth et al., 2011; Doll and Frimmel, 2003; Imai et al., 2002) with SWCNTs which are known as synthetic, high surface area and low density adsorbents (Mauter and Elimelech, 2008; Pan and Xing, 2008).
UVeVIS absorbance was measured with an Evolution 300 spectrophotometer (Thermo Scientific, Waltham, MA, USA) in order to obtain SUVA (UV absorbance at 254 nm divided by DOC concentration), E2/E3 (UV absorbance at 250 nm/365 nm) and E4/E6 (UV absorbance at 465 nm/665 nm) parameters. The E2/E3 and E4/E6 ratios were used to evaluate aromaticity, averaged molecular weight and polarity of the DOM samples (Chen et al., 1977; Peuravuori and Pihlaja, 1997). Functional group analysis was conducted by titration procedure (Inbar et al., 1990). The ionic content of the DOM extract was measured by an Arcos-EOP ICP/AES (Spectro, Kleve, Germany) after digestion of the DOM solution with 65% HNO3 and 35% HCl at 150 C for 2 h. 2.3. DOM separation into structural fractions
Pristine SWCNTs (outer diameter 1e2 nm and length 5e30 mm) were purchased from Chengdu Organic Chemistry Co., Ltd. (Chengdu, China). The surface area of the SWCNTs (420 m2 g1) was determined through N2 adsorption by a surface area analyzer (NOVA 3200E Quantachrome, Hook, UK). Purity of the SWCNTs (94%) was determined by thermal gravimetric analysis (TGA Q500, TA instruments, New Castle, DE, USA). Supelite DAX-8, Amberlyst 15 and Amberlyst A-21 resins were purchased from SigmaeAldrich (St. Louis, MO, USA). Suwannee River natural organic matter (SRNOM) was purchased from the International Humic Substances Society (IHSS) (St. Paul, MN, USA). Deionized water and analytical grade solvents (SigmaeAldrich) were used in all experiments.
DOM separation to structural fractions was performed at the end of the adsorption trials in three equilibrium concentrations: 13, 22 and 66 mg C L1. At the end of the trial, the non-adsorbed DOM (i.e., DOM in the supernatant) and the bulk DOM (i.e., DOM at the same concentration as the non-adsorbed DOM, agitated under the same conditions without SWCNTs) were separated into five structural fractions: hydrophobic acid (HoA), hydrophobic neutral (HoN), hydrophilic acid (HiA), hydrophilic base (HiB) and hydrophilic neutral (HiN). The procedure was performed in triplicates according to a resin separation procedure developed by Leenheer (1981) with several modifications (Amery et al., 2009; Chefetz et al., 1998a). Briefly, separation between hydrophilic and hydrophobic fractions was conducted by loading the DOM onto Supelite DAX-8 resin. The DOM-to-DAX-8 ratio was set to 1 mg C to 2 mL DAX-8 resin (based on preliminary experiments), with approximately 6 mg DOC loaded in each batch. This DOM-to-DAX-8 ratio was kept constant in all experiments. Acidified DOM (pH z 3) was loaded onto DAX-8 resin; at this stage the hydrophilic fractions were eluted from the column and hydrophobic fractions were retained on the DAX-8 resin. The HoA fraction was then eluted from the DAX-8 resin using 0.1 M NaOH while the HoN fraction was retained. The pH of the hydrophilic fraction was adjusted (pH z 3) with NaOH and loaded onto cation-exchange resin (Amberlyst 15). The HiA and HiN fractions were eluted while the HiB fraction was retained on the resin. The pH of the eluted solution was readjusted to 3 with 3 M NaOH and loaded onto anion-exchange resin (Amberlyst A-21). At this stage, the HiA fraction was retained on the resin while the HiN fraction was eluted. The amounts of cation- and anion-exchange resins were set to 200 and 300% of the ionic content of the sample, respectively. DOC concentration was measured for each of the eluted fractions. SRNOM separation into fractions was performed using the same method as for DOM at 20 mg C L1. The benefit of the resin separation technique is that the distribution of DOM fractions in both the adsorbed and the non-adsorbed DOM phases may be quantified.
2.2. DOM isolation and characterization
2.4. Isolation and purification of structural fractions of DOM
DOM was extracted from composted biosolids which has been € gel-knabner used as a representative model for natural DOM (Ko et al., 2000; Niemeyer et al., 1992). The compost was sieved (2 mm) and agitated (250 rpm) overnight at 25 C with deionized water at a 1:10 solid-to-water ratio. The suspension was then centrifuged (12,000 g, 20 min) and filtered through a 0.45-mm filter using an Acrodisc Supor membrane (PALL Corp., Ann Arbor, MI, USA) (Chefetz et al., 1998b; Maoz and Chefetz, 2010; Smith et al., 2012). Fresh aqueous extract (i.e., DOM) was prepared for each experiment. Carbon concentration of DOM (i.e., dissolved organic carbon; DOC) was determined using a VCSH total organic carbon analyzer (Shimadzu, Japan; detection limit was 0.5 mg C L1).
Fractions were extracted from compost DOM to obtain a sufficient amount of material to be used as sorbates in batch-adsorption trials. Bulk DOM was fractionated as described above. The obtained HoA fraction was purified by loading it onto Hþ-loaded cationexchange resin (Amberlyst 15) to obtain protonated HoA. The HiA fraction was eluted from the anion-exchange resin (Amberlyst A21) with 1 M NaOH. The obtained HiA fraction was then dialyzed (Spectra/Por 100e500 Da, cellulose ester dialysis membranes) against distilled water in order to remove excess Naþ ions. The HoN fraction was extracted from the DAX-8 resin by Soxhlet extraction with methanol for 24 h. The excess methanol was evaporated at 35 C until approximately 2 mL remained. Distilled water was then
2. Materials and methods 2.1. Materials
M. Engel, B. Chefetz / Environmental Pollution 197 (2015) 287e294
were measured. The complete removal of the SWCNTs by filtration was validated via measuring the absorbance at 800 nm (Hyung et al., 2007; Smith et al., 2012). Each concentration, including blanks (i.e., bulk DOM without SWCNTs) was conducted in triplicate. The amount of DOM adsorbed to SWCNTs was calculated by mass differences; mass losses for blank samples were negligible. Adsorption trials for DOM fractions (HoA, HoN, HiA, HiB and HiN) were performed by the same procedure as for the DOMeSWCNT system but within a narrower initial concentration range (80-5 mg C L1). Ionic strength and composition were adjusted as needed to obtain conditions similar to those in the DOMeSWCNT system. Adsorption of HoA was also examined at pH 10.
added (100 mL) and the sample was freeze-dried to remove residual methanol. The HiB fraction was obtained from a previous fractionation procedure (Chefetz et al., 1998a), in which HiB was desorbed from the cation-exchange resin (Amberlyst 15) by elution with one pore volume of 0.1 M NH4OH. Excess NHþ 4 ions were removed by evaporation at 40 C and the sample was freeze-dried. 2.5. Adsorption experiments Adsorption trials were conducted using batch-equilibration technique at 25 C. DOM was diluted to a series of initial concentrations (100-3 mg C L1) which were added to Pyrex vials with Teflon screw caps containing 4 mg SWCNTs. The solid-to-water ratio was set to 1:4 (mg:mL) in order to achieve at least 20% adsorption. The background solution was prepared using chloride salts to simulate DOM extract composition (Ca2þ, 4.4 mg L1; Mg2þ, 3.8 mg L1; Naþ, 12.5 mg L1 and Kþ, 28.8 mg L1). Sodium azide (100 mg L1) was added to inhibit microbial activity. This enabled us to maintain a constant ionic strength (3.6 mM) and solution composition at all DOM initial concentrations. The pH in all solutions was 7e8. For the adsorption trials, DOM samples were agitated (120 rpm) for 5 days with SWCNTs (according to adsorption kinetic experiments). At the end of the equilibrium time the supernatants were filtered (0.45 mm filter) and DOC concentrations 60%
Bulk
289
2.6. Data analysis Langmuir (Eq. (1)) and Freundlich (Eq. (2)) models were employed to fit the adsorption data and to recalculate the adsorbed amount of each fraction in different concentrations. The distribution coefficient (Kd) was calculated using Eq. (3), and Kd of the bulk DOM was compared to Kd calculated as the sum of Kd values of each fraction adjusted by its relative percentage in the DOM using Eq. (4).
Non-adsorbed DOM conc. 13 mg C L-1
40%
Fraction percentage [%]
20%
0% HoA
HoN
HiB
HiA
HiN
60%
SRNOM conc. 20 mg C L-1
DOM conc. 22 mg C L-1 40%
20%
0% HoA
HoN
HiB
HiA
HiN
HoA
HoN
HiB
HiA
HiN
60%
DOM conc. 66 mg C L-1 40%
20%
0% HoA
HoN
HiB
HiA
HiN
DOM fractions Fig. 1. Composition of non-adsorbed and bulk dissolved organic matter (DOM) and Suwannee River natural organic matter (SRNOM) at different concentrations. Mean values and standard errors for three or nine replicates are presented. DOM fractions include: hydrophobic acid (HoA), hydrophobic neutral (HoN), hydrophilic acid (HiA), hydrophilic base (HiB) and hydrophilic neutral (HiN).
290
M. Engel, B. Chefetz / Environmental Pollution 197 (2015) 287e294
Table 1 Parameters of bulk and non-adsorbed dissolved organic matter (DOM). Mean ± standard error (for three or nine replicates) are presented. DOM type
Conc. [mg C L1]
SUVA [L mg C1 m1]
E2/E3
E4/E6
Total acidity [mmol g1]
Bulk Non-adsorbed
13
3.6 ± 1.8$102 2.6 ± 1.3$101
3.4 ± 1.1$102 3.7 ± 1.5$102
5.2 ± 3.0$102 4.2 ± 4.8$102
9.4 ± 6.5$101 7.1 ± 9.3$102
Bulk Non-adsorbed
22
3.4 ± 2.7$101 2.6 ± 2.3$101
3.2 ± 1.3$102 3.7 ± 2.3$102
4.8 ± 2.0$101 4.6 ± 2.2$102
6.5 ± 2.6$101 5.5 ± 2.4$101
Bulk Non-adsorbed
66a
3.6 ± 8.3$103 2.8 ± 1.8$104
3.2 ± 3.8$103 3.6 ± 5.6$103
5.2 ± 1.2$102 4.8 ± 4.0$102
4.3 ± 1.4$101 3.9 ± 2.3$101
Spectroscopic parameters were obtained by sample dilution to 20 mg C L1.
q¼
qmax $KL $Ce 1 þ KL $Ce
where q (mg C g1) is the adsorbed DOM, qmax is the maximum adsorption capacity, KL (L mg C1) is the adsorption affinity parameter and Ce (mg C L1) is the aqueous equilibrium concentration of DOM.
q¼
Kf $CeN
(2)
where Kf [(mg C g1)/(mg C L1)N] is the adsorption affinity coefficient and N is the Freundlich exponential coefficient.
Kd ¼
q Ce
C L-1 Ce=13 mg C/L
75%
(1)
Ce=22 mg C/L C L-1 Ce=66 mg C/L C L-1
60%
Fraction percentage [%]
a
45%
30%
15%
0%
(3) HoA
HiN
HiA
HiB
DOM fractions
KdðDOMÞ ¼ KdðHoAÞ $f þ KdðHoNÞ $f þ KdðHiNÞ $f þ KdðHiAÞ $f þ KdðHiBÞ $f
HoN
(4)
where Kd(DOM), Kd(HoA), Kd(HoN), Kd(HiN), Kd(HiA), Kd(HiB) are the distribution coefficients of the DOM, HoA, HoN, HiN, HiA and HiB fractions, respectively; and f is the relative percentage of each fraction in the bulk DOM. Data analysis was performed using Matlab software (v. 8.1.0 MathWorks Inc., Natick, MA, USA). Statistical significance (p < 0.05) was determined by Student's t-test.
3. Results and discussion 3.1. DOMeSWCNT: DOM adsorptive fractionation Adsorptive fractionation of DOM upon interaction with SWCNTs was observed and found to be concentration-dependent (Fig. 1). Interestingly, an increase in DOM concentration resulted in an increase in the relative percentage of the hydrophobic fractions (HoA and HoN) for both non-adsorbed and bulk DOM (i.e., control). At elevated concentrations, DOM molecules are forced to interact due to reduced free space for their movement, forming larger, more hydrophobic clusters (Zsolnay, 2003). In our study, these clusters are not likely to form micelle-like domains since the working concentrations were at least one order of magnitude lower than the critical micelle concentration reported for humic acid (Engebretson and von Wandruszka, 1994; Guetzloff and Rice, 1994). The altered composition of the DOM was supported by the significant decrease in DOM total acidity as the DOM concentration increased (Table 1). Based on the observed differences in DOM compositional properties with concentration increase, the effect of SWCNTs on the chemical composition of DOM was evaluated for bulk (control) and non-adsorbed DOM at equal concentrations.
Fig. 2. Mean distribution of adsorbed dissolved organic matter (DOM) fractions at different DOM equilibrium concentrations. Bars represent standard errors for three or nine replicates. DOM fractions include: hydrophobic acid (HoA), hydrophobic neutral (HoN), hydrophilic acid (HiA), hydrophilic base (HiB) and hydrophilic neutral (HiN).
Our results for the two types of DOM show distinct differences between the bulk and non-adsorbed DOM samples, indicating a significant change in composition of DOM induced by adsorption (Fig. 1). The spectroscopic parameters (SUVA, E2/E3 and E4/E6) also indicated differences in DOM properties due to interaction with the SWCNTs (Table 1). Non-adsorbed DOM exhibited significantly lower SUVA values than those of the corresponding bulk DOM, suggesting preferential adsorption of aromatic-rich DOM components by SWCNTs, as reported for the adsorption of NOM to MWCNTs (Hyung and Kim, 2008). The E2/E3 and E4/E6 values of non-adsorbed DOM were significantly higher and lower, respectively, than those of bulk DOM. Similar differences were observed between spectroscopic parameters of bulk and non-adsorbed SRNOM (SUVA value decreased from 4.0 to 3.0 L mg C1 m1). These differences imply that DOM fractions of higher molecular weight containing aromatic and polar moieties were preferentially adsorbed (Peuravuori and Pihlaja, 1997). Similarly, higher E2/E3 ratios have been reported for non-adsorbed fulvic acid after interaction with SWCNTs (Yang and Xing, 2009). Total acidity was significantly lower for non-adsorbed DOM at 13 and 22 mg C L1 (Table 1), probably due to adsorption of acidic hydrophobic compounds. At 66 mg C L1, the total acidity values did not significantly differ between the bulk and non-adsorbed DOM. The overall observations (Table 1 and Fig. 1) suggest alterations in DOM composition due to its interaction with SWCNTs. The HoA fraction exhibited significantly lower percentage in the non-adsorbed DOM than in the bulk DOM at all tested concentrations (Fig. 1). This strongly suggests preferential adsorption of this
M. Engel, B. Chefetz / Environmental Pollution 197 (2015) 287e294
fraction to the SWCNTs. Indeed, analysis of the composition of the adsorbed DOM (Fig. 2) revealed that HoA was the major fraction (>50%) of the adsorbed materials at all studied equilibrium concentrations. The preferential adsorption of the HoA fraction likely resulted from the type of HoAeSWCNTs interactions. HoA contains the highest aromatic content (SUVA value of 2.8 L mg C1 m1) of all other fractions (max. SUVA value of 1.5 mg C1 m1 was obtained for the other fractions). HoA is known to contain high amounts of humic and fulvic acids (Aiken et al., 1992; Kaiser, 2003) which are both composed mainly of aromatic moieties that are rich in polar functionalities, such as carboxyl and hydroxyl (Quails and Haines, 1991), creating both electron-deficient and electron-rich hosting benzene rings, respectively. These rings may act as efficient peelectron acceptors and donors, respectively (Wang et al., 2011), interacting with the SWCNTs which act as both peelectron donors and acceptors (Lin and Xing, 2008b). Electrostatic attraction was probably negligible in our system due to the high pH (7e8) in which the HoA carboxylic functional groups (6.2 mmol g1) were deprotonated, facilitating electrostatic repulsion with the pseudonegative charge of the SWCNTs (Niyogi et al., 2002). This electrostatic repulsion seemed to have a minor influence on the adsorption. Moreover, at elevated pH, the carboxylate anions lose their Hbond-donating abilities and may not interact with the H-bondaccepting graphene sheets of the SWCNTs. However, H-bonds may form between the phenolic groups of HoA (1.8 mmol g1) and the SWCNT sheets serving as H-bond acceptors (Yang and Xing, 2009; Yang et al., 2008). Therefore, the HoA fraction has higher capability for interaction with SWCNTs, which leads to preferential adsorption. As opposed to HoA, the HiN fraction was present at a significantly higher percentage in the non-adsorbed DOM compared to the bulk DOM at identical concentration. It is important to note that the percentage of HiN in the non-adsorbed fraction increased following adsorption, despite it being substantially adsorbed itself (Figs. 1 and 2). This suggests that this fraction has relatively low affinity to the solid phases thus it is less likely to compete over adsorption sites with other fractions. HiN consists mainly of poly and oligosaccharides (Chefetz et al., 1998a; Maoz and Chefetz, 2010; Quails and Haines, 1991). It contains few charged functional groups, and it therefore interacts with SWCNTs primarily through Van der Waals interactions. Based on the properties of HiN, one would assume low adsorption to SWCNTs. Nonetheless, its adsorbed amount was substantial because HiN is ubiquitous in the DOMeSWCNT system (>40%). At the initial DOM concentration of 22 mg C L1, HiN and HoA accounted for 46 and 28% of the DOM, respectively. Despite the initially higher amount of HiN in the DOM solution, at equilibrium, the amount of adsorbed HoA was nearly double that of HiN, emphasizing HoA's superior adsorption (Fig. 2). The adsorption of HiN and HoA fractions may also have occurred to different adsorption sites on the SWCNT aggregates. The HoA fraction contains larger, more aromatic compounds than the HiN fraction, as was implied by SUVA (2.8 versus 0.6 L mg C1 m1, respectively) and E2/E3 values (3.5 versus 4.5, respectively). Therefore, certain adsorption sites may not be accessible to HoA but mainly to smaller hydrophilic compounds such as those which compose the HiN fraction. For HoN and HiB fractions, the differences between the percentages obtained in the bulk versus non-adsorbed compost DOM samples of corresponding concentrations were not significant (Fig. 1). Low percentages of these fractions in compost DOM and the insignificant differences in their distribution due to interactions with SWCNTs produced large errors in their adsorbed amounts (Fig. 2). Therefore, no clear conclusions could be drawn regarding the adsorptive trends of HoN and HiB fractions of compost DOM.
291
To validate the fractionation trends obtained for the compost DOM, we evaluated the fractionation of a different type of DOM (i.e., SRNOM). SRNOM was more hydrophobic in character, as compared to the compost DOM. The HoA, HoN, HiN, HiA and HiB fractions composed 48%, 11%, 19%, 22% and 0% of SRNOM, versus 29%, 1%, 46%, 22% and 3% of compost DOM, respectively. Although the two types of DOM exhibited different composition, both showed similar adsorptive fractionation trends e significantly lower percentage of HoA and higher percentage of HiN in the nonadsorbed phase compared to the bulk materials (Fig. 1 middle). Moreover, significantly lower percentage of HoN and higher percentage of HiA was present in the non-adsorbed SRNOM compared to the bulk SRNOM, respectively. Due to the low percentage of HoN in the compost DOM, precise results for this fractionation could not be obtained. The HoN level in the SRNOM was much higher, enabling more accurate measurements. Based on the SRNOM data it is clear that HoN, similar to HoA, is likely to be adsorbed by the SWCNTs, resulting in its significant percentage reduction in the non-adsorbed SRNOM. The SRNOM data for HiA further supports the low percentage of HiA at the adsorbed phase (Fig. 2) suggesting that this fraction is either not adsorbed at a concentration higher than 13 mg C L1 or is characterized as HoA at the adsorbed phase. 3.2. DOMeSWCNT: adsorption isotherms of DOM and the isolated fractions Adsorption of DOM to SWCNTs exhibited a nonlinear isotherm (Fig. 3A), likely due to heterogeneous adsorption sites on the SWCNTs (Zhang et al., 2010) and the heterogeneous DOM mixture (Hyung and Kim, 2008; Zhang et al., 2011). In aqueous media, SWCNTs tend to bundle due to strong Van der Waals interactions, forming heterogeneous adsorption sites which include external surfaces, grooves and interstitial regions (Ajayan et al., 1999; Pan and Xing, 2008). The maximum adsorption capacity for the DOM in our system was 90 mg C g1 (Table 2); when normalized to surface area (420 m2 g1), the obtained maximum adsorption capacity was within the range of values reported for adsorption of tannic acid to SWCNTs (Lin and Xing, 2008a) and peat humic acid to MWCNTs (Wang et al., 2011). On the other hand, maximum adsorption capacity of Aldrich humic acid to MWCNTs (Wang et al., 2013) was over four times higher than the value obtained in our work. This is probably since Aldrich humic acid is rich in aromatic moieties (Chappell et al., 2009). The adsorption isotherms of fulvic acid (Yang and Xing, 2009) and SRNOM (Hyung and Kim, 2008) to MWCNTs were less curvature than the isotherm obtained for DOM used in the current study. Various interactions, such as Van der Waals forces, electrostatic interactions, H-bonding and peelectron donor/acceptor interactions have been proposed as adsorption mechanisms of DOM and DOM analogs by SWCNTs (Lin and Xing, 2008b; Wang et al., 2009; Yang and Xing, 2010; Zhang et al., 2010). The adsorptive fractionation of DOM (Figs. 1 and 2) suggests that some of the DOM structures are adsorbed while others are less likely to interact with the solid phases of the SWCNTs. This implies that the obtained adsorption parameters (capacity and affinity) for DOM cannot be used to study mechanistic interactions. Therefore, in this study we have investigated the adsorption behavior of the individual fractions. Similar to the bulk DOM, HoA and HoN exhibited nonlinear isotherms (Fig. 3A) and were all best fitted by the Langmuir model (Table 2). The nature of the isotherm was found to be dependent on the hydrophobicity of the adsorbed fraction. For the hydrophobic fractions (HoA and HoN), which are likely to form specific interactions (peelectron donor/acceptor and H-bonds) with the SWCNTs, adsorption ultimately reached maximum amount.
292
M. Engel, B. Chefetz / Environmental Pollution 197 (2015) 287e294
Fig. 3. Adsorption isotherms (A) and distribution coefficients (B) of dissolved organic matter (DOM) and DOM fractions by single-walled carbon nanotubes (SWCNTs). DOM fractions include: hydrophobic acid (HoA), hydrophobic neutral (HoN), hydrophilic acid (HiA), hydrophilic base (HiB) and hydrophilic neutral (HiN).
Table 2 Adsorption parameters for dissolved organic matter (DOM) and DOM fractions by single-walled carbon nanotubes (SWCNT). DOM fractions include: hydrophobic acid (HoA), hydrophobic neutral (HoN), hydrophilic acid (HiA), hydrophilic base (HiB) and hydrophilic neutral (HiN). Mean ± standard error presented. Langmuir model
DOM HoA HoN HiA HiN HiB
KL [L mg C1]
qmax [mg C g1]
0.095 ± 0.004 0.11 ± 0.015 0.065 ± 0.008 0.032 ± 0.006 0.017 ± 0.003 0.009 ± 0.003 Freundlich model
90 77 121 98 122 91
KF [(mg C g1)/(mg C L1)N] DOM HoA HoN HiA HiN HiB
14.42 13.84 13.41 4.20 3.18 1.37
± ± ± ± ± ±
1.32 1.60 1.40 0.62 0.32 0.25
± ± ± ± ± ±
1 4 6 12 16 20
± ± ± ± ± ±
0.03 0.04 0.03 0.04 0.02 0.05
Maximum adsorption capacity followed the order HoN > bulk DOM > HoA (Table 2), which may be explained by differences in molecular size. HoN exhibited smaller size than the DOM and HoA (as implied by E2/E3 values of 7.5, 3.4 and 3.5, respectively) and therefore had the highest qmax value (Table 2). The higher qmax value of DOM in comparison to HoA can be explained by adsorption of additional fractions such as HiN (Fig. 2), which also make up a substantial amount of the adsorbed DOM. Langmuir adsorptioneaffinity parameters were not significantly different between bulk DOM and HoA, likely owing to similar adsorption interactions of pep origin. The Langmuir adsorption affinity parameter of HoN was slightly lower than those of HoA and DOM, possibly due to its' lower aromaticity (HoN SUVA value of 1.5 mg C1 m1). Adsorption of the hydrophilic fractions (HiA, HiB and HiN) was best fitted by the Freundlich model (Table 2; Fig. 3A). Hydrophilic fractions contain less aromatic moieties (SUVA values of 1.5, 0.6 and 0.4 L mg C1 m1 for HiA, HiN and HiB, respectively) and are therefore likely to interact mostly via Van der Waals interactions. The differences between the types of isotherms of the isolated DOM fractions suggest that they behave like a heterogeneous mixture in the bulk DOM. The Freundlich adsorption coefficient of HiB was lower than those of HiN and HiA, perhaps due to the high solubility of the HiB fraction, which consists mainly of aliphatic amines and amino acids (Barber et al., 2001; Quails and Haines, 1991). No statistical differences were found between the Freundlich parameters of HiN and HiA. The distribution coefficients (Kd, Eq. (3)) were significantly higher for DOM and its hydrophobic fractions than for its hydrophilic fractions (Fig. 3B). Similarly, Chen et al. (2007) reported that the adsorption affinity was higher for nonpolar aromatic compared to nonpolar aliphatic fractions. The Kd of the hydrophilic fractions showed little dependence on adsorbed amount (Fig. 3B). Their binding, governed by nonspecific interactions, occurred mainly to the external surfaces of the SWCNTs. On the other hand, Kd for the DOM and hydrophobic fractions was dependent on adsorbed amounts, with a high affinity noted for low adsorbed amounts. Kd decreased as the adsorbed amount increased due to reduction in vacant adsorption sites. The results of this study emphasize that DOM is a heterogeneous mixture of weakly clustered constituents which interact differently with SWCNTs. Therefore, both SWCNTs
Table 3 Distribution coefficients (Kd) and weight-adjusted Kd values of hydrophobic acid (HoA), hydrophobic neutral (HoN), hydrophilic acid (HiA), hydrophilic base (HiB) and hydrophilic neutral (HiN) fractions. Mean ± standard error are presented. Ce (DOM) [mg C L1]
Fraction Kd (fraction)a [L g1]
13
HoA HoN HiN HiA HiB
6.8 7.7 1.9 3.1 2.1
± ± ± ± ±
1.2 2.8 0.3 0.5 1.1
14.1 1.5 57.3 24.8 1.0
± ± ± ± ±
0.3 1.0 0.3 0.1 0.5 1.1 0.3 0.8 0.5 0.02
± ± ± ± ±
0.2 0.05 0.2 0.1 0.01
22
HoA HoN HiN HiA HiB
5.8 7.6 1.8 2.7 1.4
± ± ± ± ±
0.9 5.7 0.3 0.5 0.7
17.7 2.3 50.3 23.9 4.1
± ± ± ± ±
0.3 1.0 1.5 0.2 1.1 0.9 1.2 0.7 1.5 0.06
± ± ± ± ±
0.2 0.1 0.2 0.1 0.04
66
HoA HoN HiN HiA HiB
2.8 5.6 1.4 2.4 1.3
± ± ± ± ±
0.4 1.2 0.3 0.7 0.4
26.9 9.4 45.8 13.8 1.8
± ± ± ± ±
0.7 0.8 1.2 0.5 1.5 0.6 2.1 0.3 0.2 0.02
± ± ± ± ±
0.1 0.1 0.1 0.1 0.01
N 0.44 0.43 0.52 0.74 0.75 0.78
a
Calculated from Eq. (3).
f (fraction) [%]
Weight-adjusted Kd [L g1]
(fraction)
M. Engel, B. Chefetz / Environmental Pollution 197 (2015) 287e294
and DOM are likely to contribute to the nonlinear adsorption isotherm of DOM. The contribution of H-bonds formed between the phenolic groups of HoA and SWCNTs was established by evaluating adsorption at pH 10 which is close to the pKa of the phenolic groups of DOM. At this pH, approximately half of the phenolic groups are deprotonated (Lin and Xing, 2008b), and therefore unable to act as H-bond donors. At pH 10 adsorption capacity and affinity were significantly lower as compared to pH 8, both parameters reduced by approximately 25%. Similar results were observed by Yang et al. (2008) who reported higher adsorption affinity for solutes with Hbond-donating abilities to CNTs. Thus H-bond interactions may indeed play an important role in the adsorption of HoA and therefore of DOM to SWCNTs. However, electrostatic repulsion also increases with elevated pH and may have contributed to the reduced adsorption. Further investigation is needed to clarify the main reason for the reduced adsorption with pH increase. The Kd value of the bulk DOM was compared to that calculated as the sum of the DOM fractions' Kd values adjusted by each fractions' relative percentage (Eqs. (3) and (4), Table 3). This calculation is based on the assumption that the bulk DOM is a physical mixture of the isolated fractions which do not interact with each other at a molecular level. Thus the DOM adsorption properties can be described by independent adsorption compartments according to the relative percentage of each fraction. The measured Kd values of the bulk DOM were 3.8, 2.8 and 1.2 L g1 at DOM equilibrium concentrations of 13, 22 and 66 mg C L1, respectively. At these equilibrium concentrations the calculated values (i.e., Kd (DOM) sum, Eq. (4)) for the bulk DOM were 3.0, 2.8 and 2.3 L g1, respectively. The observed values and the calculated ones were close, suggesting that DOM may be considered a physical mixture of independent structural fractions serving as adsorbates in our experimental system. The over estimation of Kd observed only for the high DOM concentration may suggest that the DOM fractions are clustered in by weak interactions (H-bonds, Van der Waals and pep) that suppress the ability of the fractions to act as independent adsorbates. These clusters probably break apart upon interaction with SWCNTs. The incompatible Kd values obtained for the high DOM concentration may also be explained by competition between fractions which may potentially interact with the SWCNTs via similar adsorptive interactions.
Ce=13 mg C/L C L-1 45%
L-1 Ce=22 mg C C/L
Fraction percentage [%]
Ce=66 mg C C/L L-1
30%
15%
0% HoA
HoN
HiN
HiA
HiB
DOM fractions Fig. 4. Calculated composition of adsorbed dissolved organic matter (DOM) fractions at different equilibrium concentrations. Bars represent calculated errors. DOM fractions include: hydrophobic acid (HoA), hydrophobic neutral (HoN), hydrophilic acid (HiA), hydrophilic base (HiB) and hydrophilic neutral (HiN).
293
To support these observations, the distribution of the adsorbed DOM was also computed according to each fractions' adsorbed amount calculated by its' fitted model parameters and relative concentration (Fig. 4). The calculated adsorbed percentages of HoA, HoN, HiN, HiA and HiB fractions were similar to the range obtained in Fig. 2 (51e72%, 0e15%, 26e34%, 0e18% and 0e9% for HoA, HoN, HiN, HiA and HiB fractions, respectively) for the tested equilibrium concentration range of DOM. Similar to the experimental results, HoA was the dominant fraction at the adsorbed phase, HiN was substantially adsorbed and the percentage of HiA decreased with increasing concentration. The calculated adsorbed distribution of DOM further supports the assumption that the structural fractions of DOM act as independent adsorbates in the solution and interact with the SWCNTs based on their specific adsorption affinities. 4. Conclusion The results of this study observed for two different types of DOM suggest that DOM adsorbs by SWCNTs and undergoes consequent fractionation. Adsorption of DOM was governed by the SWCNTeHoA interaction, proving the dominance of this fraction in the adsorbed phase while the non-adsorbed DOM becomes more hydrophilic. This, in turn, might change the reactivity of the DOM in systems exposed to CNTs in light of its' altered composition. As the hydrophilicity increases, it is less likely that the DOM will assist in adsorption to solid phases (via co and cumulative adsorption) and/ or enhance the solubility of hydrophobic organic contaminants. It is more likely that the DOM will compete with polar organic molecules for the adsorption sites. We anticipate similar adsorptive fractionation trends of DOM by other pristine carbon-based CNTs such as MWCNTs as they all exhibit similar surface properties. Acknowledgments This research was supported in part by BARD, the United StatesIsrael Binational Agricultural Research and Development Fund (US4656-13). References Aiken, G.R., Mcknight, I.D.M., Thorn, I.K.A., Thurman, E.M., 1992. Isolation of hydrophilic organic acids from water using nonionic macroporous resins. Org. Geochem. 18, 567e573. Ajayan, P.M., Charlier, J., Rinzler, A.G., 1999. Carbon nanotubes: from macromolecules to nanotechnology. Proc. Natl. Acad. Sci. U S A 96, 14199e14200. Ajayan, P.M., Zhou, O.Z., 2001. Applications of carbon nanotubes. In: Dresselhaus, M.S., Dresselhaus, G., Avouris, P. (Eds.), Topics in Applied Physics. Springer-Verlag, Berlin, pp. 391e425. Amery, F., Vanmoorleghem, C., Smolders, E., 2009. Adapted DAX-8 fractionation method for dissolved organic matter (DOM) from soils: development, calibration with test components and application to contrasting soil solutions. Eur. J. Soil Sci. 60, 956e965. Aschberger, K., Johnston, H.J., Stone, V., Aitken, R.J., Hankin, S.M., Peters, S.A.K., Tran, C.L., Christensen, F.M., 2010. Review of carbon nanotubes toxicity and exposure-appraisal of human health risk assessment based on open literature. Crit. Rev. Toxicol. 40, 759e790. Baghoth, S.A., Sharma, S.K., Amy, G.L., 2011. Tracking natural organic matter (NOM) in a drinking water treatment plant using fluorescence excitation-emission matrices and PARAFAC. Water Res. 45, 797e809. Barber, L.B., Leenheer, J.A., Noyes, T.I., Stiles, E.A., 2001. Nature and transformation of dissolved organic matter in treatment wetlands. Environ. Sci. Technol. 35, 4805e4816. Chappell, M.A., George, A.J., Dontsova, K.M., Porter, B.E., Price, C.L., Zhou, P., Morikawa, E., Kennedy, A.J., Steevens, J.A., 2009. Surfactive stabilization of multi-walled carbon nanotube dispersions with dissolved humic substances. Environ. Pollut. 157, 1081e1087. Chefetz, B., Chen, Y., Hadar, Y., Hatcher, P.G., 1998a. Characterization of dissolved organic matter extracted from composted municipal solid waste. Soil Sci. Soc. Am. J. 62, 326e332. Chefetz, B., Hadar, Y., Chen, Y., 1998b. Dissolved organic carbon fractions formed during composting of municipal solid waste: properties and significance. Acta Hydrochim. Hydrobiol. 26, 172e179.
294
M. Engel, B. Chefetz / Environmental Pollution 197 (2015) 287e294
Chen, W., Duan, L., Zhu, D., 2007. Adsorption of polar and nonpolar organic chemicals to carbon nanotubes. Environ. Sci. Technol. 41, 8295e8300. Chen, Y., Senesi, N., Schnitzer, M., 1977. Information provided on humic substances by E4/E6 ratios. Soil Sci. Soc. Am. J. 41, 352e358. Doll, T.E., Frimmel, F.H., 2003. Fate of pharmaceuticals-photodegradation by simulated solar UV-light. Chemosphere 52, 1757e1769. Engebretson, R.R., von Wandruszka, R., 1994. Micro-organization in dissolved humic acids. Environ. Sci. Technol. 28, 1934e1941. Guetzloff, T.F., Rice, J.A., 1994. Does humic acid form a micelle? Sci. Total Environ. 152, 31e35. Hyung, H., Fortner, J.D., Hughes, J.B., Kim, J.H., 2007. Natural organic matter stabilizes carbon nanotubes in the aqueous phase. Environ. Sci. Technol. 41, 179e184. Hyung, H., Kim, J.H., 2008. Natural organic matter (NOM) adsorption to multiwalled carbon nanotubes: effect of NOM characteristics and water quality parameters. Environ. Sci. Technol. 42, 4416e4421. Imai, A., Fukushima, T., Matsushige, K., Kim, Y.H., 2001. Fractionation and characterization of dissolved organic matter in a shallow eutrophic lake, its inflowing rivers, and other organic matter sources. Water Res. 35, 4019e4028. Imai, A., Fukushima, T., Matsushige, K., Kim, Y.-H., Choi, K., 2002. Characterization of dissolved organic matter in effluents from wastewater treatment plants. Water Res. 36, 859e870. Inbar, Y., Chen, Y., Hadar, Y., 1990. Humic substances formed during the composting of organic matter. Soil Sci. Soc. Am. J. 1323, 1316e1323. Kaiser, K., 2003. Sorption of natural organic matter fractions to goethite (a-FeOOH): effect of chemical composition as revealed by liquid-state 13C NMR and wetchemical analysis. Org. Geochem. 34, 1569e1579. Kogel-Knabner, I., Totsche, K.U., Raber, B., 2000. Desorption of polycyclic aromatic hydrocarbons from soil in the presence of dissolved organic matter: effect of solution composition and aging. J. Environ. Qual. 29, 906e916. Leenheer, J.A., 1981. Comprehensive approach to preparative isolation and fractionation of dissolved organic carbon from natural waters and wastewaters. Environ. Sci. Technol. 15, 578e587. Li, T., Lin, D., Li, L., Wang, Z., Wu, F., 2014. The kinetic and thermodynamic sorption and stabilization of multiwalled carbon nanotubes in natural organic matter surrogate solutions: the effect of surrogate molecular weight. Environ. Pollut. 186, 43e49. Liang, F., Chen, B., 2010. A review on biomedical applications of single-walled carbon nanotubes. Curr. Med. Chem. 17, 10e24. Lin, D., Li, T., Yang, K., Wu, F., 2012. The relationship between humic acid (HA) adsorption on and stabilizing multiwalled carbon nanotubes (MWNTs) in water: effects of HA, MWNT and solution properties. J. Hazard. Mater. 241e242, 404e410. Lin, D., Liu, N., Yang, K., Zhu, L., Xu, Y., Xing, B., 2009. The effect of ionic strength and pH on the stability of tannic acid-facilitated carbon nanotube suspensions. Carbon 47, 2875e2882. Lin, D., Xing, B., 2008a. Tannic acid adsorption and its role for stabilizing carbon nanotube suspensions. Environ. Sci. Technol. 42, 5917e5923. Lin, D., Xing, B., 2008b. Adsorption of phenolic compounds by carbon nanotubes: role of aromaticity and substitution of hydroxyl groups. Environ. Sci. Technol. 42, 7254e7259. Liu, Y., Zhao, Y., Sun, B., Chen, C., 2013. Understanding the toxicity of carbon nanotubes. Acc. Chem. Res. 46, 702e713. Lu, C., Su, F., 2007. Adsorption of natural organic matter by carbon nanotubes. Sep. Purif. Technol. 58, 113e121. Maoz, A., Chefetz, B., 2010. Sorption of the pharmaceuticals carbamazepine and naproxen to dissolved organic matter: role of structural fractions. Water Res. 44, 981e989. Mauter, M.S., Elimelech, M., 2008. Critical review environmental applications of carbon-based nanomaterials. Environ. Sci. Technol. 42, 5843e5859.
Niemeyer, J., Chen, Y., Bollag, J., 1992. Characterization of humic acids, composts, and peat by diffuse reflectance Fourier-transform infrared spectroscopy. Soil Sci. Soc. Am. J. 56, 135e140. Niyogi, S., Hamon, M.A., Hu, H., Zhao, B., Bhowmik, P., Sen, R., Itkis, M.E., Haddon, R.C., 2002. Chemistry of single-walled carbon nanotubes. Acc. Chem. Res. 35, 1105e1113. Pan, B., Ghosh, S., Xing, B., 2008. Dissolved organic matter conformation and its interaction with pyrene as affected by water chemistry and concentration. Environ. Sci. Technol. 42, 1594e1599. Pan, B., Xing, B., 2008. Critical review adsorption mechanisms of organic chemicals on carbon nanotubes. Environ. Sci. Technol. 42, 9005e9013. Petersen, E.J., Zhang, L., Mattison, N.T., O'Carroll, D.M., Whelton, A.J., Uddin, N., Nguyen, T., Huang, Q., Henry, T.B., Holbrook, R.D., Chen, K.L., 2011. Potential release pathways, environmental fate, and ecological risks of carbon nanotubes. Environ. Sci. Technol. 45, 9837e9856. Peuravuori, J., Pihlaja, K., 1997. Molecular size distribution and spectroscopic properties of aquatic humic substances. Anal. Chim. Acta 337, 133e149. Quails, R.G., Haines, B.L., 1991. Geochemistry of dissolved organic nutrients in water percolating through a forest ecosystem. Soil Sci. Soc. Am. J. 55, 1112e1123. Schnorr, J.M., Swager, T.M., 2011. Emerging applications of carbon nanotubes. Chem. Mater. 23, 646e657. Smith, B., Yang, J., Bitter, J.L., Ball, W.P., Fairbrother, D.H., 2012. Influence of surface oxygen on the interactions of carbon nanotubes with natural organic matter. Environ. Sci. Technol. 46, 12839e12847. Sun, K., Jin, J., Gao, B., Zhang, Z., Wang, Z., Pan, Z., Xu, D., Zhao, Y., 2012. Sorption of 17a-ethinyl estradiol, bisphenol A and phenanthrene to different size fractions of soil and sediment. Chemosphere 88, 577e583. Wang, F., Yao, J., Chen, H., Yi, Z., Xing, B., 2013. Sorption of humic acid to functionalized multi-walled carbon nanotubes. Environ. Pollut. 180, 1e6. Wang, X., Lu, J., Xing, B., 2008. Sorption of organic contaminants by carbon nanotubes: influence of adsorbed organic matter. Environ. Sci. Technol. 42, 3207e3212. Wang, X., Shu, L., Wang, Y., Xu, B., Bai, Y., Tao, S., Xing, B., 2011a. Sorption of peat humic acids to multi-walled carbon nanotubes. Environ. Sci. Technol. 45, 9276e9283. Wang, X., Tao, S., Xing, B., 2009. Sorption and competition of aromatic compounds and humic acid on multiwalled carbon nanotubes. Environ. Sci. Technol. 43, 6214e6219. Wang, Z., Chen, J., Sun, Q., Peijnenburg, W.J.G.M., 2011b. C60-DOM interactions and effects on C60 apparent solubility: a molecular mechanics and density functional theory study. Environ. Int. 37, 1078e1082. Yang, K., Wu, W., Jing, Q., Zhu, L., 2008. Aqueous adsorption of aniline, phenol, and their substitutes by multi-walled carbon nanotubes. Environ. Sci. Technol. 42, 7931e7936. Yang, K., Xing, B., 2006. Adsorption of polycyclic aromatic hydrocarbons by carbon nanomaterials. Environ. Sci. Technol. 40, 1855e1861. Yang, K., Xing, B., 2009. Adsorption of fulvic acid by carbon nanotubes from water. Environ. Pollut. 157, 1095e1100. Yang, K., Xing, B., 2010. Adsorption of organic compounds by carbon nanomaterials in aqueous phase: Polanyi theory and its application. Chem. Rev. 110, 5989e6008. Zhang, S., 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, 2067e2074. Zhang, S., 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, 1378e1386. 2003. Dissolved organic matter: artefacts, definitions, and functions. Zsolnay, A., Geoderma 113, 187e209.
Contents lists available at ScienceDirect
Environmental Pollution journal homepage: www.elsevier.com/locate/envpol
Adsorptive fractionation of dissolved organic matter (DOM) by carbon nanotubes Maya Engel, Benny Chefetz* Department of Soil and Water Sciences, Faculty of Agriculture, Food and Environment, The Hebrew University of Jerusalem, P.O. Box 12, Rehovot 7610001, Israel
a r t i c l e i n f o
a b s t r a c t
Article history: Received 1 September 2014 Received in revised form 5 November 2014 Accepted 14 November 2014 Available online 3 December 2014
Dissolved organic matter (DOM) and carbon nanotubes are introduced into aquatic environments. Thus, it is important to elucidate whether their interaction affects DOM amount and composition. In this study, the composition of DOM, before and after interactions with single-walled carbon nanotubes (SWCNTs), was measured and the adsorption affinity of the individual structural fractions of DOM to SWCNTs was investigated. Adsorption of DOM to SWCNTs was dominated by the hydrophobic acid fraction, resulting in relative enhancement of the hydrophilic character of non-adsorbed DOM. The preferential adsorption of the HoA fraction was concentration-dependent, increasing with increasing concentration. Adsorption affinities of bulk DOM calculated as the normalized sum of affinities of the individual structural fractions were similar to the measured affinities, suggesting that the structural fractions of DOM act as independent adsorbates. The altered DOM composition may affect the nature and reactivity of DOM in aquatic environments polluted with carbon nanotubes. © 2014 Elsevier Ltd. All rights reserved.
Keywords: Carbon nanomaterial Fraction Sorption Composition Characterization
1. Introduction Carbon nanotubes (CNTs) are produced by rolling sheets of graphene into a cylinder along a lattice vector in the graphene plane. They consist of sp2-bonded carbon atoms forming a network of aromatic rings. Two main types of CNTs are manufactured: single-walled (SWCNTs) and multi-walled (MWCNTs) (Ajayan et al., 1999). Their nano-scale size and morphology confer unique properties, such as extremely high tensile strength, exceptional electrochemical activity, metallic and semi-conducting features and high thermal conductivity (Ajayan and Zhou, 2001; Niyogi et al., 2002). Thus, a sharp increase in the use of CNTs has been reported for a wide range of applications, such as energy storage, microelectronics, composite materials and biomedicine (Liang and Chen, 2010; Mauter and Elimelech, 2008; Schnorr and Swager, 2011). CNTs are known as toxic materials and their hazardous effect is size-dependent; manufactured and combustion sources of CNTs in the environment were suggested to have adverse effects on human health (Liu et al., 2013).
* Corresponding author. E-mail address: [email protected] (B. Chefetz). http://dx.doi.org/10.1016/j.envpol.2014.11.020 0269-7491/© 2014 Elsevier Ltd. All rights reserved.
Their wide use implies their potential release into the environment (Aschberger et al., 2010; Petersen et al., 2011), where they might interact with organic pollutants (Yang and Xing, 2006). In addition, CNTs have been found to interact with dissolved organic matter (DOM), modifying behavior of both DOM and CNTs (Hyung et al., 2007; Lin and Xing, 2008a; Lin et al., 2009; Sun et al., 2012; Yang and Xing, 2010). To better understand the environmental implications of CNTs release into water bodies, one must study DOM adsorption by CNTs (Wang et al., 2011, 2008; Yang and Xing, 2009) and more importantly, evaluate DOM fractionation during adsorption. Both processes (adsorption and fractionation) affect the quantity and quality of DOM remaining in surface water. DOM is a heterogeneous mixture of soluble aromatic and aliphatic compounds, and is considered the most reactive fraction of natural organic matter (NOM). The aromatic moieties of DOM might include conjugated benzene rings with carbonyl, hydroxyl and amide substituents which impact the electron-donating and accepting properties of the benzene derivatives (Imai et al., 2001; Quails and Haines, 1991). Several studies have examined adsorption of DOM surrogates to CNTs (Li et al., 2014; Lin and Xing, 2008a; Lin et al., 2012), but only few reports have addressed the interactions between naturally occurring DOM and CNTs (Chappell et al., 2009; Hyung et al., 2007; Lu and Su, 2007). For example, tannic acid (used as a DOM surrogate) was reported to stabilize CNT
288
M. Engel, B. Chefetz / Environmental Pollution 197 (2015) 287e294
suspension while adsorption affinity was correlated with CNT surface area. A two-stage adsorption model was proposed: primary adsorption of tannic acid by the CNTs via pep interactions until monolayer formation, and further adsorption of tannic acid to the adsorbed phase by polar interactions (Lin and Xing, 2008a). The adsorption of DOM by MWCNTs was reported to increase with decreasing pH and increasing ionic strength. Moreover, higher molecular weight fractions of DOM-like materials were preferentially adsorbed to MWCNTs (Hyung and Kim, 2008; Wang et al., 2013). Another study reported favorable adsorption of aromaticrich polar fulvic acid fractions of high molecular weight by CNTs (Yang and Xing, 2009). The abovementioned studies suggest that DOM fractionation might occur upon adsorption to CNTs, consequently changing its conformation and composition. However, very little is known about configurational changes and fractionation of native DOM upon adsorption to CNTs (Pan et al., 2008; Yang and Xing, 2010). Thus, the main objective of this study was to elucidate the adsorptive fractionation of DOM with SWCNTs and to quantify the adsorption affinities of the structural fractions of DOM to SWCNTs. Adsorption of DOM by SWCNTs is expected to be governed by pep interactions, and therefore we hypothesize that highly polar fractions of the DOM will be less prone to surface interactions. The latter materials are thus expected to dominate the residual, non-adsorbed DOM, whereas the DOM coating the CNTs is likely to be enriched with higher molecular weight and aromatic-rich fractions. Consequently, the composition of the non-adsorbed DOM is likely to differ from that of the bulk DOM. To check our hypothesis we examined the interactions of two types of naturally occurring DOM at environmentally relevant concentrations (Baghoth et al., 2011; Doll and Frimmel, 2003; Imai et al., 2002) with SWCNTs which are known as synthetic, high surface area and low density adsorbents (Mauter and Elimelech, 2008; Pan and Xing, 2008).
UVeVIS absorbance was measured with an Evolution 300 spectrophotometer (Thermo Scientific, Waltham, MA, USA) in order to obtain SUVA (UV absorbance at 254 nm divided by DOC concentration), E2/E3 (UV absorbance at 250 nm/365 nm) and E4/E6 (UV absorbance at 465 nm/665 nm) parameters. The E2/E3 and E4/E6 ratios were used to evaluate aromaticity, averaged molecular weight and polarity of the DOM samples (Chen et al., 1977; Peuravuori and Pihlaja, 1997). Functional group analysis was conducted by titration procedure (Inbar et al., 1990). The ionic content of the DOM extract was measured by an Arcos-EOP ICP/AES (Spectro, Kleve, Germany) after digestion of the DOM solution with 65% HNO3 and 35% HCl at 150 C for 2 h. 2.3. DOM separation into structural fractions
Pristine SWCNTs (outer diameter 1e2 nm and length 5e30 mm) were purchased from Chengdu Organic Chemistry Co., Ltd. (Chengdu, China). The surface area of the SWCNTs (420 m2 g1) was determined through N2 adsorption by a surface area analyzer (NOVA 3200E Quantachrome, Hook, UK). Purity of the SWCNTs (94%) was determined by thermal gravimetric analysis (TGA Q500, TA instruments, New Castle, DE, USA). Supelite DAX-8, Amberlyst 15 and Amberlyst A-21 resins were purchased from SigmaeAldrich (St. Louis, MO, USA). Suwannee River natural organic matter (SRNOM) was purchased from the International Humic Substances Society (IHSS) (St. Paul, MN, USA). Deionized water and analytical grade solvents (SigmaeAldrich) were used in all experiments.
DOM separation to structural fractions was performed at the end of the adsorption trials in three equilibrium concentrations: 13, 22 and 66 mg C L1. At the end of the trial, the non-adsorbed DOM (i.e., DOM in the supernatant) and the bulk DOM (i.e., DOM at the same concentration as the non-adsorbed DOM, agitated under the same conditions without SWCNTs) were separated into five structural fractions: hydrophobic acid (HoA), hydrophobic neutral (HoN), hydrophilic acid (HiA), hydrophilic base (HiB) and hydrophilic neutral (HiN). The procedure was performed in triplicates according to a resin separation procedure developed by Leenheer (1981) with several modifications (Amery et al., 2009; Chefetz et al., 1998a). Briefly, separation between hydrophilic and hydrophobic fractions was conducted by loading the DOM onto Supelite DAX-8 resin. The DOM-to-DAX-8 ratio was set to 1 mg C to 2 mL DAX-8 resin (based on preliminary experiments), with approximately 6 mg DOC loaded in each batch. This DOM-to-DAX-8 ratio was kept constant in all experiments. Acidified DOM (pH z 3) was loaded onto DAX-8 resin; at this stage the hydrophilic fractions were eluted from the column and hydrophobic fractions were retained on the DAX-8 resin. The HoA fraction was then eluted from the DAX-8 resin using 0.1 M NaOH while the HoN fraction was retained. The pH of the hydrophilic fraction was adjusted (pH z 3) with NaOH and loaded onto cation-exchange resin (Amberlyst 15). The HiA and HiN fractions were eluted while the HiB fraction was retained on the resin. The pH of the eluted solution was readjusted to 3 with 3 M NaOH and loaded onto anion-exchange resin (Amberlyst A-21). At this stage, the HiA fraction was retained on the resin while the HiN fraction was eluted. The amounts of cation- and anion-exchange resins were set to 200 and 300% of the ionic content of the sample, respectively. DOC concentration was measured for each of the eluted fractions. SRNOM separation into fractions was performed using the same method as for DOM at 20 mg C L1. The benefit of the resin separation technique is that the distribution of DOM fractions in both the adsorbed and the non-adsorbed DOM phases may be quantified.
2.2. DOM isolation and characterization
2.4. Isolation and purification of structural fractions of DOM
DOM was extracted from composted biosolids which has been € gel-knabner used as a representative model for natural DOM (Ko et al., 2000; Niemeyer et al., 1992). The compost was sieved (2 mm) and agitated (250 rpm) overnight at 25 C with deionized water at a 1:10 solid-to-water ratio. The suspension was then centrifuged (12,000 g, 20 min) and filtered through a 0.45-mm filter using an Acrodisc Supor membrane (PALL Corp., Ann Arbor, MI, USA) (Chefetz et al., 1998b; Maoz and Chefetz, 2010; Smith et al., 2012). Fresh aqueous extract (i.e., DOM) was prepared for each experiment. Carbon concentration of DOM (i.e., dissolved organic carbon; DOC) was determined using a VCSH total organic carbon analyzer (Shimadzu, Japan; detection limit was 0.5 mg C L1).
Fractions were extracted from compost DOM to obtain a sufficient amount of material to be used as sorbates in batch-adsorption trials. Bulk DOM was fractionated as described above. The obtained HoA fraction was purified by loading it onto Hþ-loaded cationexchange resin (Amberlyst 15) to obtain protonated HoA. The HiA fraction was eluted from the anion-exchange resin (Amberlyst A21) with 1 M NaOH. The obtained HiA fraction was then dialyzed (Spectra/Por 100e500 Da, cellulose ester dialysis membranes) against distilled water in order to remove excess Naþ ions. The HoN fraction was extracted from the DAX-8 resin by Soxhlet extraction with methanol for 24 h. The excess methanol was evaporated at 35 C until approximately 2 mL remained. Distilled water was then
2. Materials and methods 2.1. Materials
M. Engel, B. Chefetz / Environmental Pollution 197 (2015) 287e294
were measured. The complete removal of the SWCNTs by filtration was validated via measuring the absorbance at 800 nm (Hyung et al., 2007; Smith et al., 2012). Each concentration, including blanks (i.e., bulk DOM without SWCNTs) was conducted in triplicate. The amount of DOM adsorbed to SWCNTs was calculated by mass differences; mass losses for blank samples were negligible. Adsorption trials for DOM fractions (HoA, HoN, HiA, HiB and HiN) were performed by the same procedure as for the DOMeSWCNT system but within a narrower initial concentration range (80-5 mg C L1). Ionic strength and composition were adjusted as needed to obtain conditions similar to those in the DOMeSWCNT system. Adsorption of HoA was also examined at pH 10.
added (100 mL) and the sample was freeze-dried to remove residual methanol. The HiB fraction was obtained from a previous fractionation procedure (Chefetz et al., 1998a), in which HiB was desorbed from the cation-exchange resin (Amberlyst 15) by elution with one pore volume of 0.1 M NH4OH. Excess NHþ 4 ions were removed by evaporation at 40 C and the sample was freeze-dried. 2.5. Adsorption experiments Adsorption trials were conducted using batch-equilibration technique at 25 C. DOM was diluted to a series of initial concentrations (100-3 mg C L1) which were added to Pyrex vials with Teflon screw caps containing 4 mg SWCNTs. The solid-to-water ratio was set to 1:4 (mg:mL) in order to achieve at least 20% adsorption. The background solution was prepared using chloride salts to simulate DOM extract composition (Ca2þ, 4.4 mg L1; Mg2þ, 3.8 mg L1; Naþ, 12.5 mg L1 and Kþ, 28.8 mg L1). Sodium azide (100 mg L1) was added to inhibit microbial activity. This enabled us to maintain a constant ionic strength (3.6 mM) and solution composition at all DOM initial concentrations. The pH in all solutions was 7e8. For the adsorption trials, DOM samples were agitated (120 rpm) for 5 days with SWCNTs (according to adsorption kinetic experiments). At the end of the equilibrium time the supernatants were filtered (0.45 mm filter) and DOC concentrations 60%
Bulk
289
2.6. Data analysis Langmuir (Eq. (1)) and Freundlich (Eq. (2)) models were employed to fit the adsorption data and to recalculate the adsorbed amount of each fraction in different concentrations. The distribution coefficient (Kd) was calculated using Eq. (3), and Kd of the bulk DOM was compared to Kd calculated as the sum of Kd values of each fraction adjusted by its relative percentage in the DOM using Eq. (4).
Non-adsorbed DOM conc. 13 mg C L-1
40%
Fraction percentage [%]
20%
0% HoA
HoN
HiB
HiA
HiN
60%
SRNOM conc. 20 mg C L-1
DOM conc. 22 mg C L-1 40%
20%
0% HoA
HoN
HiB
HiA
HiN
HoA
HoN
HiB
HiA
HiN
60%
DOM conc. 66 mg C L-1 40%
20%
0% HoA
HoN
HiB
HiA
HiN
DOM fractions Fig. 1. Composition of non-adsorbed and bulk dissolved organic matter (DOM) and Suwannee River natural organic matter (SRNOM) at different concentrations. Mean values and standard errors for three or nine replicates are presented. DOM fractions include: hydrophobic acid (HoA), hydrophobic neutral (HoN), hydrophilic acid (HiA), hydrophilic base (HiB) and hydrophilic neutral (HiN).
290
M. Engel, B. Chefetz / Environmental Pollution 197 (2015) 287e294
Table 1 Parameters of bulk and non-adsorbed dissolved organic matter (DOM). Mean ± standard error (for three or nine replicates) are presented. DOM type
Conc. [mg C L1]
SUVA [L mg C1 m1]
E2/E3
E4/E6
Total acidity [mmol g1]
Bulk Non-adsorbed
13
3.6 ± 1.8$102 2.6 ± 1.3$101
3.4 ± 1.1$102 3.7 ± 1.5$102
5.2 ± 3.0$102 4.2 ± 4.8$102
9.4 ± 6.5$101 7.1 ± 9.3$102
Bulk Non-adsorbed
22
3.4 ± 2.7$101 2.6 ± 2.3$101
3.2 ± 1.3$102 3.7 ± 2.3$102
4.8 ± 2.0$101 4.6 ± 2.2$102
6.5 ± 2.6$101 5.5 ± 2.4$101
Bulk Non-adsorbed
66a
3.6 ± 8.3$103 2.8 ± 1.8$104
3.2 ± 3.8$103 3.6 ± 5.6$103
5.2 ± 1.2$102 4.8 ± 4.0$102
4.3 ± 1.4$101 3.9 ± 2.3$101
Spectroscopic parameters were obtained by sample dilution to 20 mg C L1.
q¼
qmax $KL $Ce 1 þ KL $Ce
where q (mg C g1) is the adsorbed DOM, qmax is the maximum adsorption capacity, KL (L mg C1) is the adsorption affinity parameter and Ce (mg C L1) is the aqueous equilibrium concentration of DOM.
q¼
Kf $CeN
(2)
where Kf [(mg C g1)/(mg C L1)N] is the adsorption affinity coefficient and N is the Freundlich exponential coefficient.
Kd ¼
q Ce
C L-1 Ce=13 mg C/L
75%
(1)
Ce=22 mg C/L C L-1 Ce=66 mg C/L C L-1
60%
Fraction percentage [%]
a
45%
30%
15%
0%
(3) HoA
HiN
HiA
HiB
DOM fractions
KdðDOMÞ ¼ KdðHoAÞ $f þ KdðHoNÞ $f þ KdðHiNÞ $f þ KdðHiAÞ $f þ KdðHiBÞ $f
HoN
(4)
where Kd(DOM), Kd(HoA), Kd(HoN), Kd(HiN), Kd(HiA), Kd(HiB) are the distribution coefficients of the DOM, HoA, HoN, HiN, HiA and HiB fractions, respectively; and f is the relative percentage of each fraction in the bulk DOM. Data analysis was performed using Matlab software (v. 8.1.0 MathWorks Inc., Natick, MA, USA). Statistical significance (p < 0.05) was determined by Student's t-test.
3. Results and discussion 3.1. DOMeSWCNT: DOM adsorptive fractionation Adsorptive fractionation of DOM upon interaction with SWCNTs was observed and found to be concentration-dependent (Fig. 1). Interestingly, an increase in DOM concentration resulted in an increase in the relative percentage of the hydrophobic fractions (HoA and HoN) for both non-adsorbed and bulk DOM (i.e., control). At elevated concentrations, DOM molecules are forced to interact due to reduced free space for their movement, forming larger, more hydrophobic clusters (Zsolnay, 2003). In our study, these clusters are not likely to form micelle-like domains since the working concentrations were at least one order of magnitude lower than the critical micelle concentration reported for humic acid (Engebretson and von Wandruszka, 1994; Guetzloff and Rice, 1994). The altered composition of the DOM was supported by the significant decrease in DOM total acidity as the DOM concentration increased (Table 1). Based on the observed differences in DOM compositional properties with concentration increase, the effect of SWCNTs on the chemical composition of DOM was evaluated for bulk (control) and non-adsorbed DOM at equal concentrations.
Fig. 2. Mean distribution of adsorbed dissolved organic matter (DOM) fractions at different DOM equilibrium concentrations. Bars represent standard errors for three or nine replicates. DOM fractions include: hydrophobic acid (HoA), hydrophobic neutral (HoN), hydrophilic acid (HiA), hydrophilic base (HiB) and hydrophilic neutral (HiN).
Our results for the two types of DOM show distinct differences between the bulk and non-adsorbed DOM samples, indicating a significant change in composition of DOM induced by adsorption (Fig. 1). The spectroscopic parameters (SUVA, E2/E3 and E4/E6) also indicated differences in DOM properties due to interaction with the SWCNTs (Table 1). Non-adsorbed DOM exhibited significantly lower SUVA values than those of the corresponding bulk DOM, suggesting preferential adsorption of aromatic-rich DOM components by SWCNTs, as reported for the adsorption of NOM to MWCNTs (Hyung and Kim, 2008). The E2/E3 and E4/E6 values of non-adsorbed DOM were significantly higher and lower, respectively, than those of bulk DOM. Similar differences were observed between spectroscopic parameters of bulk and non-adsorbed SRNOM (SUVA value decreased from 4.0 to 3.0 L mg C1 m1). These differences imply that DOM fractions of higher molecular weight containing aromatic and polar moieties were preferentially adsorbed (Peuravuori and Pihlaja, 1997). Similarly, higher E2/E3 ratios have been reported for non-adsorbed fulvic acid after interaction with SWCNTs (Yang and Xing, 2009). Total acidity was significantly lower for non-adsorbed DOM at 13 and 22 mg C L1 (Table 1), probably due to adsorption of acidic hydrophobic compounds. At 66 mg C L1, the total acidity values did not significantly differ between the bulk and non-adsorbed DOM. The overall observations (Table 1 and Fig. 1) suggest alterations in DOM composition due to its interaction with SWCNTs. The HoA fraction exhibited significantly lower percentage in the non-adsorbed DOM than in the bulk DOM at all tested concentrations (Fig. 1). This strongly suggests preferential adsorption of this
M. Engel, B. Chefetz / Environmental Pollution 197 (2015) 287e294
fraction to the SWCNTs. Indeed, analysis of the composition of the adsorbed DOM (Fig. 2) revealed that HoA was the major fraction (>50%) of the adsorbed materials at all studied equilibrium concentrations. The preferential adsorption of the HoA fraction likely resulted from the type of HoAeSWCNTs interactions. HoA contains the highest aromatic content (SUVA value of 2.8 L mg C1 m1) of all other fractions (max. SUVA value of 1.5 mg C1 m1 was obtained for the other fractions). HoA is known to contain high amounts of humic and fulvic acids (Aiken et al., 1992; Kaiser, 2003) which are both composed mainly of aromatic moieties that are rich in polar functionalities, such as carboxyl and hydroxyl (Quails and Haines, 1991), creating both electron-deficient and electron-rich hosting benzene rings, respectively. These rings may act as efficient peelectron acceptors and donors, respectively (Wang et al., 2011), interacting with the SWCNTs which act as both peelectron donors and acceptors (Lin and Xing, 2008b). Electrostatic attraction was probably negligible in our system due to the high pH (7e8) in which the HoA carboxylic functional groups (6.2 mmol g1) were deprotonated, facilitating electrostatic repulsion with the pseudonegative charge of the SWCNTs (Niyogi et al., 2002). This electrostatic repulsion seemed to have a minor influence on the adsorption. Moreover, at elevated pH, the carboxylate anions lose their Hbond-donating abilities and may not interact with the H-bondaccepting graphene sheets of the SWCNTs. However, H-bonds may form between the phenolic groups of HoA (1.8 mmol g1) and the SWCNT sheets serving as H-bond acceptors (Yang and Xing, 2009; Yang et al., 2008). Therefore, the HoA fraction has higher capability for interaction with SWCNTs, which leads to preferential adsorption. As opposed to HoA, the HiN fraction was present at a significantly higher percentage in the non-adsorbed DOM compared to the bulk DOM at identical concentration. It is important to note that the percentage of HiN in the non-adsorbed fraction increased following adsorption, despite it being substantially adsorbed itself (Figs. 1 and 2). This suggests that this fraction has relatively low affinity to the solid phases thus it is less likely to compete over adsorption sites with other fractions. HiN consists mainly of poly and oligosaccharides (Chefetz et al., 1998a; Maoz and Chefetz, 2010; Quails and Haines, 1991). It contains few charged functional groups, and it therefore interacts with SWCNTs primarily through Van der Waals interactions. Based on the properties of HiN, one would assume low adsorption to SWCNTs. Nonetheless, its adsorbed amount was substantial because HiN is ubiquitous in the DOMeSWCNT system (>40%). At the initial DOM concentration of 22 mg C L1, HiN and HoA accounted for 46 and 28% of the DOM, respectively. Despite the initially higher amount of HiN in the DOM solution, at equilibrium, the amount of adsorbed HoA was nearly double that of HiN, emphasizing HoA's superior adsorption (Fig. 2). The adsorption of HiN and HoA fractions may also have occurred to different adsorption sites on the SWCNT aggregates. The HoA fraction contains larger, more aromatic compounds than the HiN fraction, as was implied by SUVA (2.8 versus 0.6 L mg C1 m1, respectively) and E2/E3 values (3.5 versus 4.5, respectively). Therefore, certain adsorption sites may not be accessible to HoA but mainly to smaller hydrophilic compounds such as those which compose the HiN fraction. For HoN and HiB fractions, the differences between the percentages obtained in the bulk versus non-adsorbed compost DOM samples of corresponding concentrations were not significant (Fig. 1). Low percentages of these fractions in compost DOM and the insignificant differences in their distribution due to interactions with SWCNTs produced large errors in their adsorbed amounts (Fig. 2). Therefore, no clear conclusions could be drawn regarding the adsorptive trends of HoN and HiB fractions of compost DOM.
291
To validate the fractionation trends obtained for the compost DOM, we evaluated the fractionation of a different type of DOM (i.e., SRNOM). SRNOM was more hydrophobic in character, as compared to the compost DOM. The HoA, HoN, HiN, HiA and HiB fractions composed 48%, 11%, 19%, 22% and 0% of SRNOM, versus 29%, 1%, 46%, 22% and 3% of compost DOM, respectively. Although the two types of DOM exhibited different composition, both showed similar adsorptive fractionation trends e significantly lower percentage of HoA and higher percentage of HiN in the nonadsorbed phase compared to the bulk materials (Fig. 1 middle). Moreover, significantly lower percentage of HoN and higher percentage of HiA was present in the non-adsorbed SRNOM compared to the bulk SRNOM, respectively. Due to the low percentage of HoN in the compost DOM, precise results for this fractionation could not be obtained. The HoN level in the SRNOM was much higher, enabling more accurate measurements. Based on the SRNOM data it is clear that HoN, similar to HoA, is likely to be adsorbed by the SWCNTs, resulting in its significant percentage reduction in the non-adsorbed SRNOM. The SRNOM data for HiA further supports the low percentage of HiA at the adsorbed phase (Fig. 2) suggesting that this fraction is either not adsorbed at a concentration higher than 13 mg C L1 or is characterized as HoA at the adsorbed phase. 3.2. DOMeSWCNT: adsorption isotherms of DOM and the isolated fractions Adsorption of DOM to SWCNTs exhibited a nonlinear isotherm (Fig. 3A), likely due to heterogeneous adsorption sites on the SWCNTs (Zhang et al., 2010) and the heterogeneous DOM mixture (Hyung and Kim, 2008; Zhang et al., 2011). In aqueous media, SWCNTs tend to bundle due to strong Van der Waals interactions, forming heterogeneous adsorption sites which include external surfaces, grooves and interstitial regions (Ajayan et al., 1999; Pan and Xing, 2008). The maximum adsorption capacity for the DOM in our system was 90 mg C g1 (Table 2); when normalized to surface area (420 m2 g1), the obtained maximum adsorption capacity was within the range of values reported for adsorption of tannic acid to SWCNTs (Lin and Xing, 2008a) and peat humic acid to MWCNTs (Wang et al., 2011). On the other hand, maximum adsorption capacity of Aldrich humic acid to MWCNTs (Wang et al., 2013) was over four times higher than the value obtained in our work. This is probably since Aldrich humic acid is rich in aromatic moieties (Chappell et al., 2009). The adsorption isotherms of fulvic acid (Yang and Xing, 2009) and SRNOM (Hyung and Kim, 2008) to MWCNTs were less curvature than the isotherm obtained for DOM used in the current study. Various interactions, such as Van der Waals forces, electrostatic interactions, H-bonding and peelectron donor/acceptor interactions have been proposed as adsorption mechanisms of DOM and DOM analogs by SWCNTs (Lin and Xing, 2008b; Wang et al., 2009; Yang and Xing, 2010; Zhang et al., 2010). The adsorptive fractionation of DOM (Figs. 1 and 2) suggests that some of the DOM structures are adsorbed while others are less likely to interact with the solid phases of the SWCNTs. This implies that the obtained adsorption parameters (capacity and affinity) for DOM cannot be used to study mechanistic interactions. Therefore, in this study we have investigated the adsorption behavior of the individual fractions. Similar to the bulk DOM, HoA and HoN exhibited nonlinear isotherms (Fig. 3A) and were all best fitted by the Langmuir model (Table 2). The nature of the isotherm was found to be dependent on the hydrophobicity of the adsorbed fraction. For the hydrophobic fractions (HoA and HoN), which are likely to form specific interactions (peelectron donor/acceptor and H-bonds) with the SWCNTs, adsorption ultimately reached maximum amount.
292
M. Engel, B. Chefetz / Environmental Pollution 197 (2015) 287e294
Fig. 3. Adsorption isotherms (A) and distribution coefficients (B) of dissolved organic matter (DOM) and DOM fractions by single-walled carbon nanotubes (SWCNTs). DOM fractions include: hydrophobic acid (HoA), hydrophobic neutral (HoN), hydrophilic acid (HiA), hydrophilic base (HiB) and hydrophilic neutral (HiN).
Table 2 Adsorption parameters for dissolved organic matter (DOM) and DOM fractions by single-walled carbon nanotubes (SWCNT). DOM fractions include: hydrophobic acid (HoA), hydrophobic neutral (HoN), hydrophilic acid (HiA), hydrophilic base (HiB) and hydrophilic neutral (HiN). Mean ± standard error presented. Langmuir model
DOM HoA HoN HiA HiN HiB
KL [L mg C1]
qmax [mg C g1]
0.095 ± 0.004 0.11 ± 0.015 0.065 ± 0.008 0.032 ± 0.006 0.017 ± 0.003 0.009 ± 0.003 Freundlich model
90 77 121 98 122 91
KF [(mg C g1)/(mg C L1)N] DOM HoA HoN HiA HiN HiB
14.42 13.84 13.41 4.20 3.18 1.37
± ± ± ± ± ±
1.32 1.60 1.40 0.62 0.32 0.25
± ± ± ± ± ±
1 4 6 12 16 20
± ± ± ± ± ±
0.03 0.04 0.03 0.04 0.02 0.05
Maximum adsorption capacity followed the order HoN > bulk DOM > HoA (Table 2), which may be explained by differences in molecular size. HoN exhibited smaller size than the DOM and HoA (as implied by E2/E3 values of 7.5, 3.4 and 3.5, respectively) and therefore had the highest qmax value (Table 2). The higher qmax value of DOM in comparison to HoA can be explained by adsorption of additional fractions such as HiN (Fig. 2), which also make up a substantial amount of the adsorbed DOM. Langmuir adsorptioneaffinity parameters were not significantly different between bulk DOM and HoA, likely owing to similar adsorption interactions of pep origin. The Langmuir adsorption affinity parameter of HoN was slightly lower than those of HoA and DOM, possibly due to its' lower aromaticity (HoN SUVA value of 1.5 mg C1 m1). Adsorption of the hydrophilic fractions (HiA, HiB and HiN) was best fitted by the Freundlich model (Table 2; Fig. 3A). Hydrophilic fractions contain less aromatic moieties (SUVA values of 1.5, 0.6 and 0.4 L mg C1 m1 for HiA, HiN and HiB, respectively) and are therefore likely to interact mostly via Van der Waals interactions. The differences between the types of isotherms of the isolated DOM fractions suggest that they behave like a heterogeneous mixture in the bulk DOM. The Freundlich adsorption coefficient of HiB was lower than those of HiN and HiA, perhaps due to the high solubility of the HiB fraction, which consists mainly of aliphatic amines and amino acids (Barber et al., 2001; Quails and Haines, 1991). No statistical differences were found between the Freundlich parameters of HiN and HiA. The distribution coefficients (Kd, Eq. (3)) were significantly higher for DOM and its hydrophobic fractions than for its hydrophilic fractions (Fig. 3B). Similarly, Chen et al. (2007) reported that the adsorption affinity was higher for nonpolar aromatic compared to nonpolar aliphatic fractions. The Kd of the hydrophilic fractions showed little dependence on adsorbed amount (Fig. 3B). Their binding, governed by nonspecific interactions, occurred mainly to the external surfaces of the SWCNTs. On the other hand, Kd for the DOM and hydrophobic fractions was dependent on adsorbed amounts, with a high affinity noted for low adsorbed amounts. Kd decreased as the adsorbed amount increased due to reduction in vacant adsorption sites. The results of this study emphasize that DOM is a heterogeneous mixture of weakly clustered constituents which interact differently with SWCNTs. Therefore, both SWCNTs
Table 3 Distribution coefficients (Kd) and weight-adjusted Kd values of hydrophobic acid (HoA), hydrophobic neutral (HoN), hydrophilic acid (HiA), hydrophilic base (HiB) and hydrophilic neutral (HiN) fractions. Mean ± standard error are presented. Ce (DOM) [mg C L1]
Fraction Kd (fraction)a [L g1]
13
HoA HoN HiN HiA HiB
6.8 7.7 1.9 3.1 2.1
± ± ± ± ±
1.2 2.8 0.3 0.5 1.1
14.1 1.5 57.3 24.8 1.0
± ± ± ± ±
0.3 1.0 0.3 0.1 0.5 1.1 0.3 0.8 0.5 0.02
± ± ± ± ±
0.2 0.05 0.2 0.1 0.01
22
HoA HoN HiN HiA HiB
5.8 7.6 1.8 2.7 1.4
± ± ± ± ±
0.9 5.7 0.3 0.5 0.7
17.7 2.3 50.3 23.9 4.1
± ± ± ± ±
0.3 1.0 1.5 0.2 1.1 0.9 1.2 0.7 1.5 0.06
± ± ± ± ±
0.2 0.1 0.2 0.1 0.04
66
HoA HoN HiN HiA HiB
2.8 5.6 1.4 2.4 1.3
± ± ± ± ±
0.4 1.2 0.3 0.7 0.4
26.9 9.4 45.8 13.8 1.8
± ± ± ± ±
0.7 0.8 1.2 0.5 1.5 0.6 2.1 0.3 0.2 0.02
± ± ± ± ±
0.1 0.1 0.1 0.1 0.01
N 0.44 0.43 0.52 0.74 0.75 0.78
a
Calculated from Eq. (3).
f (fraction) [%]
Weight-adjusted Kd [L g1]
(fraction)
M. Engel, B. Chefetz / Environmental Pollution 197 (2015) 287e294
and DOM are likely to contribute to the nonlinear adsorption isotherm of DOM. The contribution of H-bonds formed between the phenolic groups of HoA and SWCNTs was established by evaluating adsorption at pH 10 which is close to the pKa of the phenolic groups of DOM. At this pH, approximately half of the phenolic groups are deprotonated (Lin and Xing, 2008b), and therefore unable to act as H-bond donors. At pH 10 adsorption capacity and affinity were significantly lower as compared to pH 8, both parameters reduced by approximately 25%. Similar results were observed by Yang et al. (2008) who reported higher adsorption affinity for solutes with Hbond-donating abilities to CNTs. Thus H-bond interactions may indeed play an important role in the adsorption of HoA and therefore of DOM to SWCNTs. However, electrostatic repulsion also increases with elevated pH and may have contributed to the reduced adsorption. Further investigation is needed to clarify the main reason for the reduced adsorption with pH increase. The Kd value of the bulk DOM was compared to that calculated as the sum of the DOM fractions' Kd values adjusted by each fractions' relative percentage (Eqs. (3) and (4), Table 3). This calculation is based on the assumption that the bulk DOM is a physical mixture of the isolated fractions which do not interact with each other at a molecular level. Thus the DOM adsorption properties can be described by independent adsorption compartments according to the relative percentage of each fraction. The measured Kd values of the bulk DOM were 3.8, 2.8 and 1.2 L g1 at DOM equilibrium concentrations of 13, 22 and 66 mg C L1, respectively. At these equilibrium concentrations the calculated values (i.e., Kd (DOM) sum, Eq. (4)) for the bulk DOM were 3.0, 2.8 and 2.3 L g1, respectively. The observed values and the calculated ones were close, suggesting that DOM may be considered a physical mixture of independent structural fractions serving as adsorbates in our experimental system. The over estimation of Kd observed only for the high DOM concentration may suggest that the DOM fractions are clustered in by weak interactions (H-bonds, Van der Waals and pep) that suppress the ability of the fractions to act as independent adsorbates. These clusters probably break apart upon interaction with SWCNTs. The incompatible Kd values obtained for the high DOM concentration may also be explained by competition between fractions which may potentially interact with the SWCNTs via similar adsorptive interactions.
Ce=13 mg C/L C L-1 45%
L-1 Ce=22 mg C C/L
Fraction percentage [%]
Ce=66 mg C C/L L-1
30%
15%
0% HoA
HoN
HiN
HiA
HiB
DOM fractions Fig. 4. Calculated composition of adsorbed dissolved organic matter (DOM) fractions at different equilibrium concentrations. Bars represent calculated errors. DOM fractions include: hydrophobic acid (HoA), hydrophobic neutral (HoN), hydrophilic acid (HiA), hydrophilic base (HiB) and hydrophilic neutral (HiN).
293
To support these observations, the distribution of the adsorbed DOM was also computed according to each fractions' adsorbed amount calculated by its' fitted model parameters and relative concentration (Fig. 4). The calculated adsorbed percentages of HoA, HoN, HiN, HiA and HiB fractions were similar to the range obtained in Fig. 2 (51e72%, 0e15%, 26e34%, 0e18% and 0e9% for HoA, HoN, HiN, HiA and HiB fractions, respectively) for the tested equilibrium concentration range of DOM. Similar to the experimental results, HoA was the dominant fraction at the adsorbed phase, HiN was substantially adsorbed and the percentage of HiA decreased with increasing concentration. The calculated adsorbed distribution of DOM further supports the assumption that the structural fractions of DOM act as independent adsorbates in the solution and interact with the SWCNTs based on their specific adsorption affinities. 4. Conclusion The results of this study observed for two different types of DOM suggest that DOM adsorbs by SWCNTs and undergoes consequent fractionation. Adsorption of DOM was governed by the SWCNTeHoA interaction, proving the dominance of this fraction in the adsorbed phase while the non-adsorbed DOM becomes more hydrophilic. This, in turn, might change the reactivity of the DOM in systems exposed to CNTs in light of its' altered composition. As the hydrophilicity increases, it is less likely that the DOM will assist in adsorption to solid phases (via co and cumulative adsorption) and/ or enhance the solubility of hydrophobic organic contaminants. It is more likely that the DOM will compete with polar organic molecules for the adsorption sites. We anticipate similar adsorptive fractionation trends of DOM by other pristine carbon-based CNTs such as MWCNTs as they all exhibit similar surface properties. Acknowledgments This research was supported in part by BARD, the United StatesIsrael Binational Agricultural Research and Development Fund (US4656-13). References Aiken, G.R., Mcknight, I.D.M., Thorn, I.K.A., Thurman, E.M., 1992. Isolation of hydrophilic organic acids from water using nonionic macroporous resins. Org. Geochem. 18, 567e573. Ajayan, P.M., Charlier, J., Rinzler, A.G., 1999. Carbon nanotubes: from macromolecules to nanotechnology. Proc. Natl. Acad. Sci. U S A 96, 14199e14200. Ajayan, P.M., Zhou, O.Z., 2001. Applications of carbon nanotubes. In: Dresselhaus, M.S., Dresselhaus, G., Avouris, P. (Eds.), Topics in Applied Physics. Springer-Verlag, Berlin, pp. 391e425. Amery, F., Vanmoorleghem, C., Smolders, E., 2009. Adapted DAX-8 fractionation method for dissolved organic matter (DOM) from soils: development, calibration with test components and application to contrasting soil solutions. Eur. J. Soil Sci. 60, 956e965. Aschberger, K., Johnston, H.J., Stone, V., Aitken, R.J., Hankin, S.M., Peters, S.A.K., Tran, C.L., Christensen, F.M., 2010. Review of carbon nanotubes toxicity and exposure-appraisal of human health risk assessment based on open literature. Crit. Rev. Toxicol. 40, 759e790. Baghoth, S.A., Sharma, S.K., Amy, G.L., 2011. Tracking natural organic matter (NOM) in a drinking water treatment plant using fluorescence excitation-emission matrices and PARAFAC. Water Res. 45, 797e809. Barber, L.B., Leenheer, J.A., Noyes, T.I., Stiles, E.A., 2001. Nature and transformation of dissolved organic matter in treatment wetlands. Environ. Sci. Technol. 35, 4805e4816. Chappell, M.A., George, A.J., Dontsova, K.M., Porter, B.E., Price, C.L., Zhou, P., Morikawa, E., Kennedy, A.J., Steevens, J.A., 2009. Surfactive stabilization of multi-walled carbon nanotube dispersions with dissolved humic substances. Environ. Pollut. 157, 1081e1087. Chefetz, B., Chen, Y., Hadar, Y., Hatcher, P.G., 1998a. Characterization of dissolved organic matter extracted from composted municipal solid waste. Soil Sci. Soc. Am. J. 62, 326e332. Chefetz, B., Hadar, Y., Chen, Y., 1998b. Dissolved organic carbon fractions formed during composting of municipal solid waste: properties and significance. Acta Hydrochim. Hydrobiol. 26, 172e179.
294
M. Engel, B. Chefetz / Environmental Pollution 197 (2015) 287e294
Chen, W., Duan, L., Zhu, D., 2007. Adsorption of polar and nonpolar organic chemicals to carbon nanotubes. Environ. Sci. Technol. 41, 8295e8300. Chen, Y., Senesi, N., Schnitzer, M., 1977. Information provided on humic substances by E4/E6 ratios. Soil Sci. Soc. Am. J. 41, 352e358. Doll, T.E., Frimmel, F.H., 2003. Fate of pharmaceuticals-photodegradation by simulated solar UV-light. Chemosphere 52, 1757e1769. Engebretson, R.R., von Wandruszka, R., 1994. Micro-organization in dissolved humic acids. Environ. Sci. Technol. 28, 1934e1941. Guetzloff, T.F., Rice, J.A., 1994. Does humic acid form a micelle? Sci. Total Environ. 152, 31e35. Hyung, H., Fortner, J.D., Hughes, J.B., Kim, J.H., 2007. Natural organic matter stabilizes carbon nanotubes in the aqueous phase. Environ. Sci. Technol. 41, 179e184. Hyung, H., Kim, J.H., 2008. Natural organic matter (NOM) adsorption to multiwalled carbon nanotubes: effect of NOM characteristics and water quality parameters. Environ. Sci. Technol. 42, 4416e4421. Imai, A., Fukushima, T., Matsushige, K., Kim, Y.H., 2001. Fractionation and characterization of dissolved organic matter in a shallow eutrophic lake, its inflowing rivers, and other organic matter sources. Water Res. 35, 4019e4028. Imai, A., Fukushima, T., Matsushige, K., Kim, Y.-H., Choi, K., 2002. Characterization of dissolved organic matter in effluents from wastewater treatment plants. Water Res. 36, 859e870. Inbar, Y., Chen, Y., Hadar, Y., 1990. Humic substances formed during the composting of organic matter. Soil Sci. Soc. Am. J. 1323, 1316e1323. Kaiser, K., 2003. Sorption of natural organic matter fractions to goethite (a-FeOOH): effect of chemical composition as revealed by liquid-state 13C NMR and wetchemical analysis. Org. Geochem. 34, 1569e1579. Kogel-Knabner, I., Totsche, K.U., Raber, B., 2000. Desorption of polycyclic aromatic hydrocarbons from soil in the presence of dissolved organic matter: effect of solution composition and aging. J. Environ. Qual. 29, 906e916. Leenheer, J.A., 1981. Comprehensive approach to preparative isolation and fractionation of dissolved organic carbon from natural waters and wastewaters. Environ. Sci. Technol. 15, 578e587. Li, T., Lin, D., Li, L., Wang, Z., Wu, F., 2014. The kinetic and thermodynamic sorption and stabilization of multiwalled carbon nanotubes in natural organic matter surrogate solutions: the effect of surrogate molecular weight. Environ. Pollut. 186, 43e49. Liang, F., Chen, B., 2010. A review on biomedical applications of single-walled carbon nanotubes. Curr. Med. Chem. 17, 10e24. Lin, D., Li, T., Yang, K., Wu, F., 2012. The relationship between humic acid (HA) adsorption on and stabilizing multiwalled carbon nanotubes (MWNTs) in water: effects of HA, MWNT and solution properties. J. Hazard. Mater. 241e242, 404e410. Lin, D., Liu, N., Yang, K., Zhu, L., Xu, Y., Xing, B., 2009. The effect of ionic strength and pH on the stability of tannic acid-facilitated carbon nanotube suspensions. Carbon 47, 2875e2882. Lin, D., Xing, B., 2008a. Tannic acid adsorption and its role for stabilizing carbon nanotube suspensions. Environ. Sci. Technol. 42, 5917e5923. Lin, D., Xing, B., 2008b. Adsorption of phenolic compounds by carbon nanotubes: role of aromaticity and substitution of hydroxyl groups. Environ. Sci. Technol. 42, 7254e7259. Liu, Y., Zhao, Y., Sun, B., Chen, C., 2013. Understanding the toxicity of carbon nanotubes. Acc. Chem. Res. 46, 702e713. Lu, C., Su, F., 2007. Adsorption of natural organic matter by carbon nanotubes. Sep. Purif. Technol. 58, 113e121. Maoz, A., Chefetz, B., 2010. Sorption of the pharmaceuticals carbamazepine and naproxen to dissolved organic matter: role of structural fractions. Water Res. 44, 981e989. Mauter, M.S., Elimelech, M., 2008. Critical review environmental applications of carbon-based nanomaterials. Environ. Sci. Technol. 42, 5843e5859.
Niemeyer, J., Chen, Y., Bollag, J., 1992. Characterization of humic acids, composts, and peat by diffuse reflectance Fourier-transform infrared spectroscopy. Soil Sci. Soc. Am. J. 56, 135e140. Niyogi, S., Hamon, M.A., Hu, H., Zhao, B., Bhowmik, P., Sen, R., Itkis, M.E., Haddon, R.C., 2002. Chemistry of single-walled carbon nanotubes. Acc. Chem. Res. 35, 1105e1113. Pan, B., Ghosh, S., Xing, B., 2008. Dissolved organic matter conformation and its interaction with pyrene as affected by water chemistry and concentration. Environ. Sci. Technol. 42, 1594e1599. Pan, B., Xing, B., 2008. Critical review adsorption mechanisms of organic chemicals on carbon nanotubes. Environ. Sci. Technol. 42, 9005e9013. Petersen, E.J., Zhang, L., Mattison, N.T., O'Carroll, D.M., Whelton, A.J., Uddin, N., Nguyen, T., Huang, Q., Henry, T.B., Holbrook, R.D., Chen, K.L., 2011. Potential release pathways, environmental fate, and ecological risks of carbon nanotubes. Environ. Sci. Technol. 45, 9837e9856. Peuravuori, J., Pihlaja, K., 1997. Molecular size distribution and spectroscopic properties of aquatic humic substances. Anal. Chim. Acta 337, 133e149. Quails, R.G., Haines, B.L., 1991. Geochemistry of dissolved organic nutrients in water percolating through a forest ecosystem. Soil Sci. Soc. Am. J. 55, 1112e1123. Schnorr, J.M., Swager, T.M., 2011. Emerging applications of carbon nanotubes. Chem. Mater. 23, 646e657. Smith, B., Yang, J., Bitter, J.L., Ball, W.P., Fairbrother, D.H., 2012. Influence of surface oxygen on the interactions of carbon nanotubes with natural organic matter. Environ. Sci. Technol. 46, 12839e12847. Sun, K., Jin, J., Gao, B., Zhang, Z., Wang, Z., Pan, Z., Xu, D., Zhao, Y., 2012. Sorption of 17a-ethinyl estradiol, bisphenol A and phenanthrene to different size fractions of soil and sediment. Chemosphere 88, 577e583. Wang, F., Yao, J., Chen, H., Yi, Z., Xing, B., 2013. Sorption of humic acid to functionalized multi-walled carbon nanotubes. Environ. Pollut. 180, 1e6. Wang, X., Lu, J., Xing, B., 2008. Sorption of organic contaminants by carbon nanotubes: influence of adsorbed organic matter. Environ. Sci. Technol. 42, 3207e3212. Wang, X., Shu, L., Wang, Y., Xu, B., Bai, Y., Tao, S., Xing, B., 2011a. Sorption of peat humic acids to multi-walled carbon nanotubes. Environ. Sci. Technol. 45, 9276e9283. Wang, X., Tao, S., Xing, B., 2009. Sorption and competition of aromatic compounds and humic acid on multiwalled carbon nanotubes. Environ. Sci. Technol. 43, 6214e6219. Wang, Z., Chen, J., Sun, Q., Peijnenburg, W.J.G.M., 2011b. C60-DOM interactions and effects on C60 apparent solubility: a molecular mechanics and density functional theory study. Environ. Int. 37, 1078e1082. Yang, K., Wu, W., Jing, Q., Zhu, L., 2008. Aqueous adsorption of aniline, phenol, and their substitutes by multi-walled carbon nanotubes. Environ. Sci. Technol. 42, 7931e7936. Yang, K., Xing, B., 2006. Adsorption of polycyclic aromatic hydrocarbons by carbon nanomaterials. Environ. Sci. Technol. 40, 1855e1861. Yang, K., Xing, B., 2009. Adsorption of fulvic acid by carbon nanotubes from water. Environ. Pollut. 157, 1095e1100. Yang, K., Xing, B., 2010. Adsorption of organic compounds by carbon nanomaterials in aqueous phase: Polanyi theory and its application. Chem. Rev. 110, 5989e6008. Zhang, S., 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, 2067e2074. Zhang, S., 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, 1378e1386. 2003. Dissolved organic matter: artefacts, definitions, and functions. Zsolnay, A., Geoderma 113, 187e209.
