Release Kinetics of Multiwalled Carbon Nanotubes Deposited on

Article pubs.acs.org/est

Release Kinetics of Multiwalled Carbon Nanotubes Deposited on Silica Surfaces: Quartz Crystal Microbalance with Dissipation (QCMD) Measurements and Modeling Peng Yi†,‡ and Kai Loon Chen*,† †

Department of Geography and Environmental Engineering, Johns Hopkins University, Baltimore, Maryland 21218-2686 S Supporting Information *

ABSTRACT: Understanding the kinetics of the release of carbon nanotubes (CNTs) from naturally occurring surfaces is crucial for the prediction of the environmental fate and transport of CNTs. In this study, the release kinetics of multiwalled CNTs (MWNTs) from silica surfaces was investigated using a quartz crystal microbalance with dissipation monitoring (QCM-D). MWNTs were first deposited on silica surfaces under favorable deposition conditions (1.50 mM CaCl2 and pH 7.1) and the deposited MWNTs were then rinsed at different electrolyte solutions to induce the release of MWNTs from the primary energy minimum. The kinetics of MWNT release was shown to be first order with respect to the deposited MWNTs when complete release took place. A model that accounts for the releasable and unreleasable components of MWNTs was used to fit the experimental data in order to derive the release rate coefficients. When the CaCl2 concentration in the eluent was decreased, a larger fraction of deposited MWNTs was released and the release rate coefficient of the releasable MWNTs also increased. The rise in the surface charges of both MWNTs and silica surfaces with the drop in CaCl2 concentration likely resulted in the decrease in the height of the energy barrier, thus facilitating the release of the nanotubes. Moreover, when the initial surface concentrations of deposited MWNTs were over 1000 ng/cm2, the release rate coefficient was lower than expected. The reduced release kinetics was likely due to the formation of large surface-bound MWNT clusters which had considerably lower diffusion coefficients than dispersed MWNTs or MWNT aggregates.

■

INTRODUCTION Carbon nanotubes (CNTs) are currently used in consumer and industrial products due to their superior mechanical, electronic, and thermal conductive properties. For instance, baseball bats are reinforced with multiwalled carbon nanotubes (MWNTs) to enhance their strength.1 In the electronics industry, CNTbased films have been developed for applications in flexible displays and touch screens.2,3 During the life cycle of CNTincorporated products, CNTs are likely to be released into the environment through mechanical abrasion and degradation (e.g., photodegradation, thermodegradation, hydrolytic degradation, and biodegradation) of the CNT-based products.4,5 Since recent evidence has shown that CNTs are toxic to aquatic organisms,6−9 it is crucial to understand the fate and transport of CNTs in aquatic systems in order to assess their risk in these environments. The mobility of CNTs in surface waters and subsurface environments is dependent on the deposition and release behavior of these nanomaterials.10,11 In these systems, CNTs can deposit on naturally occurring surfaces, such as sand, rocks, and sediments, and hence be removed from the aqueous phase.12−14 Subsequently, when the solution chemistry changes, the deposited CNTs may be released from the solid surfaces and re-enter the aqueous phase. © 2014 American Chemical Society

Although the deposition of CNTs has been studied extensively,12−19 only a few studies have been conducted on the release of CNTs from environmental surfaces.13,19−21 Jaisi et al.13 and Tian et al.19 showed that CNTs that were deposited in quartz sand porous media in the presence of KCl can be partially released from the column when rinsed with deionized (DI) water. In another study by Khan et al.21 on the transport of CNTs in model solid-waste porous media in the presence of humic acid, 19% of the CNTs deposited at 200 mM NaCl were released from the column when the NaCl concentration was decreased to 0 mM. In all three studies, the release of CNTs from the collector surface was attributed to the elimination of the secondary energy minimum.13,19,21 Recently, we investigated the release of MWNTs from the primary energy minimum using a quartz crystal microbalance with dissipation monitoring (QCM-D).20 The flow cell of the QCM-D had a geometry similar to that of a parallel-plate flow chamber which only allows for nanotube deposition in the primary minimum. Nanotubes that are deposited in secondary energy minimum will slide along the crystal of the QCM-D and Received: Revised: Accepted: Published: 4406

December March 20, March 21, March 21,

9, 2013 2014 2014 2014

dx.doi.org/10.1021/es405471u | Environ. Sci. Technol. 2014, 48, 4406−4413

Environmental Science & Technology

Article

All experiments were performed at pH 7.1 ± 0.2 (adjusted by 0.15 mM NaHCO3). The temperature for all experiments was maintained at 25 °C. Quartz Crystal Microbalance with Dissipation Monitoring. A QCM-D system (E4, Q-Sense, Västra Frölunda, Sweden) was used to investigate the release kinetics of MWNTs that were initially deposited on silica-coated crystal surfaces. Two of the four modules in the E4 system were arranged in a parallel configuration and used to obtain duplicate data for each experimental condition. A silica-coated 5 MHz AT-cut quartz crystal sensor (QSX303, Q-Sense) was mounted in each flow module. The geometry of the flow modules was similar to that of a parallel-plate flow chamber, which allows for MWNT deposition on a silica surface to take place in the primary energy minimum and not in the secondary energy minimum.20,24,25 Before each QCM-D experiment, the crystal sensors and flow modules were cleaned with 2% Hellmanex II cleaning solution (Hellma GmbH & Co. KG, Müllheim, Germany) and then with copious amount of DI water.12 All electrolyte solutions were degassed through ultrasonication (Branson 5510R-DTH, output power 135 W, frequency 40 kHz) for 10 min and stored in a water bath at 27 °C before use. The frequency and dissipation responses of the crystal sensors at nth harmonics (n = 1, 3, 5, 7, 9, 11, 13) were monitored throughout the deposition and release experiments. Because the deposited MWNTs exhibited characteristics of a viscoelastic layer, as reflected by the high ratios of dissipation shift to normalized frequency shift (ca. 0.6 × 10−6/Hz) detected for the MWNT layers,20 the Voigt-based model, instead of the Sauerbrey relationship, was employed to quantify the concentration of MWNTs deposited on crystal surfaces.26 The Voigt model is commonly used to analyze the surface mass concentration, viscosity, and shear modulus of viscoelastic layers,26−29 while the Sauerbrey relationship is used for rigid films.30,31 Using the QTools 3 software (Q-Sense), the normalized frequency shifts (Δf) and dissipation shifts (ΔD) at the 5th, 7th, 9th, and 11th harmonics obtained from the QCM-D experiments were fitted with the Voigt-based model using the surface concentration, viscosity, and shear modulus of the MWNT layers as the fitting parameters.20 Through this approach, the surface concentrations of the MWNTs on the crystal surface during deposition and release experiments can be derived as a function of time. The fluid density and viscosity were fixed at 1.00 × 103 kg/m3 and 1.00 × 10−3 kg/(m·s), respectively, since the experiments were all conducted in aqueous solutions. The density of MWNT layers was fixed at 1.05 × 103 kg/m3, which is reasonable according to the study of Lee.32 Release of Deposited MWNTs from Silica Surfaces. Before each QCM-D experiment, the crystal surfaces were rinsed with DI water and 1.50 mM CaCl2 solutions successively until stable baselines were obtained. To investigate the release kinetics of deposited MWNTs, MWNTs were first deposited on silica-coated crystal surfaces at 1.50 mM CaCl2 for 10−80 min. The MWNTs underwent favorable deposition at 1.50 mM CaCl2 since this CaCl2 concentration is higher than the critical deposition concentration (CDC) (or the minimum electrolyte concentration that allows for favorable deposition) of the MWNTs which was determined in our previous study12 to be 1.1 mM CaCl2. Since this CaCl2 concentration is also higher than the critical coagulation concentration (CCC) (or the minimum electrolyte concentration that allows for favorable aggregation), which was determined to be 0.9 mM in our

will not attach to the crystal surface. In that study, nearly all the MWNTs deposited on silica surfaces under favorable deposition conditions (i.e., 1.50 mM CaCl2 or 600 mM NaCl at pH 7.1) were released when rinsed with DI water at pH 10, hence demonstrating that nanotube deposition in the primary minimum can be reversible.20 Although the degree of CNT release under various elution solutions has been investigated in the above studies,13,16−18 in order to quantitatively predict the release process of deposited CNTs, it is crucial to investigate the kinetics of CNT release. Specifically, the relationship between the rate of CNT release and surface concentration of deposited CNTs, as well as the key parameters that govern the release kinetics of CNTs, needs to be determined. Until now, however, no research has been conducted on the kinetics of CNT release from naturally occurring surfaces. In this study, the release kinetics of deposited MWNTs from silica surfaces was investigated using a QCM-D. MWNTs were first deposited on silica surfaces in the primary energy minimum at 1.50 mM CaCl2 and the deposited MWNTs were then rinsed with eluents of various electrolyte concentrations to induce the release of MWNTs. A model that accounts for both the releasable and unreleasable components of MWNTs was used to fit the experimental data in order to derive the release kinetics in the form of release rate coefficients. The degrees of release and release rate coefficients were obtained for eluents of different electrolyte concentrations. In addition, the effects of elevated initial surface coverages of MWNTs on the release kinetics of MWNTs were investigated.

■

MATERIALS AND METHODS Preparation and Characterization of MWNTs. The MWNTs are identical to the ones used in our previous studies.12,20,22 The untreated MWNTs were purchased from NanoLab, Inc. (PD15L5-20, Newton, MA). They were oxidized and cleaned according to the procedure described in our previous studies12,22 and also in the Supporting Information (SI). The oxidization process resulted in the formation of three oxygen-containing surface functional groups, namely, the carbonyl (CO), carboxyl (COOH), and hydroxyl (C−OH) groups, on the surface of MWNTs, as determined through Xray photoelectron spectroscopy (XPS).12,23 The predominant oxygen-containing functional groups were determined to be the carboxyl groups.12 Four batches of MWNT stock suspension were prepared by dispersing the MWNTs in DI water (Millipore), as described in the SI. By performing dynamic light scattering (DLS) measurements, the hydrodynamic diameter of the MWNTs in the stock suspensions was determined to be 105−115 nm and the stock suspensions were found to be stable to aggregation during the duration of the QCM-D experiments.12 The electrophoretic mobilities (EPMs) of the MWNTs were measured (ZetaPALS, Brookhaven) as functions of NaCl and CaCl2 concentrations at pH 7.1 in our previous study.12 An additional measurement was conducted at 0.01 mM CaCl2 ( = −1.80 × 10−8 m2/(V s)), which is one of the concentrations used in this study. The combined results showed that the EPMs of MWNTs became more negative when the CaCl2 concentration was decreased, indicating that the negative surface charge of MWNTs was enhanced when the electrolyte concentration was reduced. Solution Chemistry. Electrolyte stock solutions were prepared using ACS-grade NaCl and CaCl2 and filtered with 0.1-μm syringe filters (Anotop 25, Whatman, Middlesex, UK). 4407



Article

previous study,12 the MWNTs were colloidally unstable and had a tendency to undergo fast aggregation. During the deposition process, the MWNT suspension and electrolyte solution of interest were drawn separately at equal flow rates and combined in a T-junction before they were introduced into flow modules using a peristaltic pump (ISM935C, Ismatec SA, Zürich, Switzerland).12 As indicated in our previous study,12 the use of the T-junction limited the time of MWNT exposure to the electrolytes to ca. 20 s before the MWNTs entered the flow cell and thus significantly reduced the degree of concurrent aggregation during MWNT deposition. The flow rate of the combined suspension entering each flow module was 0.60 mL/ min (±0.03 mL/min) which was low enough to allow for a laminar flow in the flow module.12,20,22 The MWNT concentration in the combined suspensions entering the flow modules was ca. 0.5 mg/L. Following the deposition of MWNTs, the deposited MWNTs were first rinsed with a 1.50 mM CaCl2 solution followed by an eluent of a different solution chemistry to induce the release of deposited MWNTs. The elution process was stopped when the shifts in the normalized frequency and dissipation responses at the fifth overtone were less than 0.3 Hz and 0.15 × 10−6, respectively, over a time period of 10 min. Imaging Deposited MWNTs with Scanning Electron Microscopy. The silica-coated crystals with different surface concentrations of deposited MWNTs were imaged through scanning electron microscopy (SEM). The detailed procedures for sample preparation and SEM imaging are provided in the SI.

■

RESULTS AND DISCUSSION Release of Deposited MWNTs from Silica Surfaces. The surface concentration of deposited MWNTs on the silica surface during the deposition of MWNTs at 1.50 mM CaCl2 and the subsequent release of deposited MWNTs at 1 mM NaCl are shown in Figure 1a. SI Figure S1a and b present the experimental and fitted frequency and dissipation shifts, respectively, at the 5th, 7th, 9th, and 11th harmonics for the same deposition and release experiment. Following the stable baseline obtained during the rinsing with a 1.50 mM CaCl2 solution (0−10 min), MWNTs were deposited on the silica surface at 1.50 mM CaCl2 (10−60 min), during which the surface concentration of deposited MWNTs increased from 0 to 924 ng/cm2. When the deposited MWNTs were rinsed with a 1.50 mM CaCl2 solution (60−80 min), no change in the surface concentration of deposited MWNTs was observed, indicating that no deposited MWNTs were released. During the subsequent rinsing with a 1 mM NaCl eluent (80 min onward), however, the surface concentration of deposited MWNTs decreased from 924 to 7 ng/cm2, indicating that nearly all the deposited MWNTs (99%) were released from the silica surface. Because the geometry of E4 flow module only allows for the deposition of MWNTs in the primary minimum and because no secondary minimum should exist under favorable deposition conditions (e.g., at 1.5 mM CaCl2), the deposited MWNTs were released from the primary minimum during the elution with the 1 mM NaCl solution.20 In the theoretical work of Ruckenstein and Prieve,33 a finite primary minimum was shown to exist in the interaction energy profile between a spherical colloid and a planar collector surface when Born repulsion was included in the interaction between the colloid and planar surface. When only DLVO interactions are

Figure 1. (a) Surface concentration of deposited MWNTs during the deposition of MWNTs on a silica surface at 1.50 mM CaCl2 and their release at 1 mM NaCl. The transient spike at 10 min was due to the switch of solutions. (b) Proportionality between the initial release rate and the surface concentration of deposited MWNTs. The solution chemistries for these MWNT deposition and release experiments were the same as those used in (a).

considered (in the absence of Born repulsion), in contrast, the primary minimum has an infinite depth. With the inclusion of Born repulsion, the energy barrier for colloid release was predicted by Ruckenstein and Prieve33 to decrease when the magnitude of the surface charge of colloids and collectors increased. In our experiment presented in Figure 1a, when the 1.50 mM CaCl2 solution was replaced by the 1 mM NaCl eluent (at 80 min), both MWNTs and silica surface became more negatively charged as fewer Ca2+ cations were available to neutralize the charges on MWNTs and silica.12 The energy barrier for MWNT release became sufficiently small for all the initially deposited MWNTs to overcome and hence to be released from the silica surface. The initial rate of MWNT release (obtained through linear regression analysis on the surface concentrations within the first minute of the release process) was determined to be 193 ng/(cm2·min). Release Kinetics when Deposited MWNTs Undergo Complete Release. To examine whether the release kinetics of deposited MWNTs is first order with respect to the 4408



Article

deposited MWNTs, a series of release experiments, similar to that presented in Figure 1a, was performed by varying the surface concentration of MWNTs that were initially deposited on silica surfaces at 1.50 mM CaCl2 before the deposited MWNTs were rinsed with a 1 mM NaCl eluent. In these experiments, almost all the deposited MWNTs (95−100%) were released. The initial rate of MWNT release was plotted as a function of the surface concentration of MWNTs that were initially deposited on the silica surfaces in Figure 1b. The figure shows a linear correlation between the initial rate of release and the surface concentration of deposited MWNTs. The yintercept was determined to be 1.67 ng/cm2/min with a standard error of 10.25 ng/cm2/min and was shown through a t test not to differ from 0 at the 0.01 level of significance. Additionally, the coefficient of determination, R2, is 0.962. Thus, it can be concluded that the MWNT release rate was proportional to the surface concentration of deposited MWNTs. In other words, the release kinetics of the deposited MWNTs was first order with respect to the deposited MWNTs when the MWNTs underwent complete release. Although firstorder colloid release kinetics was predicted in the theoretical work of Dahneke34,35 and Ruckenstein and Prieve,33 the results presented in Figure 1b, to the best of our knowledge, are the first experimental evidence showing the proportionality between nanoparticle release rate and surface concentration of deposited nanoparticles. The release kinetics of deposited MWNTs from silica surfaces can therefore be described by the equations below when the MWNTs undergo complete release:34

dm = −Km dt

(1)

m = m0e−Kt

(2)

Figure 2. Fitting the experimental surface concentration data (circles) during the release of deposited MWNTs at 0.01 mM CaCl2 using eq 2 (blue dashed curve) and eq 4 (red solid curve).

regression analysis to be 0.0364 min−1. However, as shown in Figure 2, eq 2 did not fit the experimental data well. This discrepancy between the experimental data and theoretical prediction demonstrates that eqs 1 and 2 cannot adequately describe the MWNT release process when partial nanotube release takes place. A plausible reason for the incomplete release of MWNTs is that the MWNTs may not all have the same surface charge densities.20 It is speculated here that, at 0.01 mM CaCl2, two fractions of MWNTs were presenthighly charged MWNTs that can be released with first-order kinetics and weakly charged MWNTs with low surface charge densities that cannot be released. Equations 1 and 2, both derived for the complete release of MWNTs, are hence modified to become eqs 3 and 4, respectively, which are appropriate to describe the release process for deposited MWNTs that are composed of releasable and unreleasable fractions:36

where m0 and m are the surface mass concentrations of deposited MWNTs at the start of the release process and at time t after the initiation of the release process, respectively, and K is the release rate coefficient when complete release takes place. Release Kinetics when Deposited MWNTs Undergo Partial Release. Our previous study20 showed that the partial (or incomplete) release of MWNTs from silica surfaces after deposition at 1.50 mM CaCl2 can occur when the deposited MWNTs were rinsed with eluents with CaCl2 concentrations lower than 1.50 mM (e.g., 0.10 mM or 0.01 mM CaCl2). Figure 2 presents the surface concentration of deposited MWNTs during the release process at 0.01 mM CaCl2. In this experiment, the amount of MWNTs initially deposited on a silica surface at 1.50 mM CaCl2, m0, was 917 ng/cm2. The deposited MWNTs were then sequentially rinsed with a 1.50 mM CaCl2 solution (0−24 min) and a 0.01 mM CaCl2 solution (24 min onward). No release was observed when the deposited MWNTs were rinsed with the 1.50 mM CaCl2 solution, while nanotube release occurred during the elution at 0.01 mM CaCl2. As the release process approached a steady state at ca. 60 min, there was still 290 ng/cm2 of MWNTs deposited on the silica surface (i.e., 32% of initial deposited mass). To test the validity of eqs 1 and 2 in cases where deposited MWNTs did not undergo complete release, an attempt was made to fit the surface concentration of deposited MWNTs during the release process at 0.01 mM CaCl2 (i.e., from 24 min onward) with eq 2. The K value that provides the best fitting curve (blue dashed curve in Figure 2) was determined by least-squares

dm = −k(m − mstable) dt

(3)

m = (m0 − mstable)e−kt + mstable

(4)

where mstable is the surface mass concentration of the deposited MWNTs that cannot be released when rinsed with the eluent of interest and k is the release rate coefficient of the releasable MWNTs. An equation similar to eq 4 was employed by Ryan and Gschwend36 for describing the release kinetics of hematite colloids from quartz-grain columns and, more recently, by Torkzaban et al.37 for modeling the release kinetics of quantum dots in porous media. Equation 4 was used to fit the experimental data in Figure 2 by least-squares regression analysis. The initial surface concentration, m0, as presented previously, was 917 ng/cm2, and the surface concentration of remaining deposited MWNTs at the end of release process, mstable, was determined to be 281 ng/cm2. The k value that provided the best fitting curve was then determined to be 0.118 min−1 (red solid curve in Figure 2). In contrast to eq 2, eq 4 fits the experimental data remarkably well (the residual sum of squares for eq 4 is 12% of that for eq 2). Thus, the good agreement between eq 4 and the experimental results confirms that the deposited MWNTs can indeed be divided into two fractions: releasable and unreleasable MWNTs. The releasable 4409



Article

MWNTs are expected to be considerably more negatively charged compared to the unreleasable MWNTs. As discussed in our previous study,20 the energy barrier for MWNT release is expected to decrease as the MWNTs become more negatively charged. Once the energy barrier becomes sufficiently small, the deposited MWNTs can be readily released from the primary energy minimum.20 When the 1.50 mM CaCl2 solution was replaced by a 0.01 mM CaCl2 solution, the surface charges of both releasable and unreleasable MWNTs were enhanced due to the smaller charge neutralization effect at the lower CaCl2 concentration.12 With the decrease in the CaCl2 concentration, it is plausible that the energy barrier of releasable MWNTs became sufficiently small for them to be remobilized from silica surfaces. In contrast, the energy barrier of unreleasable MWNTs was likely to be still too large for the nanotubes to be released.20 The good fit by eq 4 also verifies that the release kinetics of the releasable MWNTs is first order with respect to the releasable MWNTs that remain on the silica surface. Influence of Solution Chemistry on Release Kinetics of MWNTs from Silica Surfaces. Equations 3 and 4 are the more general forms of eqs 1 and 2, respectively, and therefore hold true when deposited MWNTs undergo either complete (mstable = 0) or incomplete release (mstable > 0). Because the surface concentration of the initial deposited MWNTs, m0, could vary between experiments, eq 4 is modified to become eq 5 so that comparisons can be made between experiments with different initial deposited concentrations: f = freleasable e−kt + (1 − freleasable )

Figure 3. Fitting the mass fraction of deposited MWNTs that remained on silica surfaces in different eluents using eq 5. The data points are the mass fractions of the deposited MWNTs that remained on silica surfaces obtained through QCM-D measurements. The solid lines are the corresponding best fitting curves obtained using eq 5

shown in Figure 4a, no release of MWNTs was observed during the elution with a 1.50 mM CaCl2 solution. When the CaCl2 concentration in the eluent was decreased from 0.10 to 0.01 mM CaCl2, f releasable increased from 29% to 72%. When a 0.001 mM CaCl2 solution was used as the eluent, f releasable further increased to 80%. It is expected that, when the CaCl2 concentration was decreased, fewer carboxyl groups on MWNTs and silanol groups on silica surfaces were neutralized by the Ca2+ cations. As a result, both deposited MWNTs and silica surfaces became more negatively charged and the energy barriers for MWNT release decreased.20 Consequently, more deposited MWNTs overcame the energy barrier and were hence released from the silica surface. When a 1 mM NaCl solution or DI water was used instead of a CaCl2 solution as the eluent, the neutralization effect of Ca2+ cations was absent and the surface charges of MWNTs and silica surfaces were further enhanced such that the energy barrier was small enough for almost all the deposited MWNTs to be released (98% and 99%, respectively). Figure 4b presents the release rate coefficients, k, at different solution chemistries. When the CaCl2 concentration was decreased from 0.10 to 0.01 and 0.001 mM, k increased from 0.041 to 0.119 and 0.195 min−1, respectively. In the theoretical work of Dahneke34 and Ruckenstein and Prieve,33 the firstorder release rate coefficient of deposited spherical colloids was predicted to have an Arrhenius relationship with the energy barrier for colloid release. An analogous relationship between the release rate coefficient and energy barrier is expected for MWNTs. The energy barrier for MWNT release at 0.001 mM CaCl2 was smaller than the energy barriers at 0.01 and 0.10 mM CaCl2 because the neutralization effect of Ca2+ cations was relatively weak at 0.001 mM CaCl2 and, consequently, both the MWNTs and silica surfaces were highly charged. Thus, the kinetics of MWNT release at 0.001 mM CaCl2 was higher than at 0.01 or 0.10 mM CaCl2. Likewise, when a 1 mM NaCl solution was used in place of a CaCl2 solution, the release rate coefficient, k, was higher (= 0.303 min−1) than the k values obtained in the eluents containing Ca2+ due to the absence of

(5)

where f is the fraction of originally deposited MWNTs that remained deposited on silica surfaces at time t (= m/m0), and f releasable is the fraction of originally deposited MWNTs that can be released during elution (= (m0 − mstable)/m0). To investigate the influence of solution chemistry on the release kinetics of the deposited MWNTs, MWNTs were first deposited on silica surfaces at 1.50 mM CaCl2. Note that for this series of experiments, less than 1000 ng/cm2 of MWNTs was deposited on the silica surfaces in each experiment. In the next section, the effect of an elevated initial coverage of MWNTs on the release kinetics will be discussed. In these experiments, the deposited MWNTs were rinsed with a 1.50 mM CaCl2 solution, followed by an eluent with a lower Ca2+ ion concentration (or with no Ca2+ ions), namely, 0.10 mM CaCl2, 0.01 mM CaCl2, 0.001 mM CaCl2, 1 mM NaCl, or DI water. The eluents containing low CaCl2 concentrations were prepared through the serial dilution of a 20 mM CaCl2 stock solution. For each release experiment, f releasable was first calculated after the determination of m0 and mstable. Following that, eq 5 was used to fit the experimental data during the release of the MWNTs using k as the fitting parameter. Figure 3 presents the experimental and fitted data during the release of deposited MWNTs at 0.10 mM CaCl2, 0.01 mM CaCl2, 0.001 mM CaCl2, and 1 mM NaCl. Because the experimental data for elution with DI water overlaps with the data for elution with a 1 mM NaCl solution, the data for elution with DI water are not presented for the sake of clarity. As shown in Figure 3, eq 5 fits the experimental data very well for all eluents used in this study (including DI water). Figure 4a and b present the fractions of MWNTs that are releasable, f releasable, and the release rate coefficients of the releasable MWNTs, k, respectively, when the deposited MWNTs were released at different solution chemistries. As 4410



Article

heterogeneous with respect to their surface charge density compared to the colloids used in previous studies on colloid release38−43 such that a single exponential decay function in eq 5 is sufficient to fit the MWNT release process well. Effects of Elevated MWNT Surface Coverage on Release Rate Coefficient. As presented in Figure 1b, the release rate was proportional to the surface mass concentration of MWNTs when the deposited MWNT underwent complete release at 1 mM NaCl and when the surface mass concentration of initially deposited MWNTs was less than 1000 ng/cm2. In this section, the relationship between release kinetics and surface concentration of MWNTs was studied over a broader range of surface mass concentrations to include elevated surface coverages (>1000 ng/cm2). The release rate coefficients were derived during the release processes as a function of surface concentration of MWNTs that were initially deposited at 1.50 mM CaCl2 (from 200 to 1600 ng/cm2). Two eluents, namely, 1 mM NaCl and 0.01 mM CaCl2, were used in order to investigate the effects of elevated MWNT coverage when complete (at 1 mM NaCl) and partial (at 0.01 mM CaCl2) release took place. Figure 5a presents the release rate coefficient, k, during the release process at 1 mM NaCl (squares) and at 0.01 mM CaCl2 (circles) as a function of initial surface concentration. The release rate coefficients at 1 mM NaCl were ca. 0.30 min−1 when the surface concentration of initially deposited MWNTs were less than 1000 ng/cm2 (i.e., from 201 to 924 ng/cm2). However, when the surface concentration was increased to the range of 1240−1600 ng/ cm2, the release rate coefficient decreased considerably. A similar trend was observed when 0.01 mM CaCl2 was used as the eluent. The release rate coefficient, k, was ca. 0.12 min−1 when the surface concentration of deposited MWNTs ranged from 525 to 917 ng/cm2. However, k decreased when the surface concentration of deposited MWNTs was increased to the range of 1140−1380 ng/cm2. Figure 5b presents the SEM images of MWNTs on silica surfaces at concentrations of 612, 970, and 1665 ng/cm2 following nanotube deposition at 1.50 mM CaCl2. In all three images, both single strands and aggregates of MWNTs are present on the silica surfaces. Because MWNTs underwent concurrent aggregation during the deposition processes at 1.50 mM CaCl2, the deposition of the nanotube aggregates resulted in some small surface-bound aggregates. At a low surface coverage of 612 ng/cm2, the MWNT strands and aggregates were evenly dispersed on the silica surface. Even as the surface concentration of MWNTs was increased from 612 to 970 ng/ cm2, the MWNT strands and aggregates remained dispersed on the silica surface and were not observed to come into contact with one another. However, when surface concentration was further increased to 1665 ng/cm2, the size of the surface-bound aggregates was observed to be considerably larger than those at surface concentrations lower than 1000 ng/cm2. At such high surface coverages, the deposited MWNT strands and aggregates may begin to merge on the silica surface to form large clusters. Because the diffusion coefficients of the large clusters are lower than those of single MWNT strands or small aggregates, the clusters of MWNTs underwent considerably slower transport from the silica surface compared to the dispersed nanotubes or small aggregates. As there is a noticeable decline in both the release rate coefficients at 1 mM NaCl and 0.01 mM CaCl2 when the surface concentration is higher than ca. 1000 ng/cm2, as shown in Figure 5a, a coverage of ca. 1000 ng/cm2 is expected to be high enough for the merging of the deposited

Figure 4. (a) Fraction of deposited MWNTs released from silica surfaces, f releasable, when rinsed with different eluents after deposition at 1.50 mM CaCl2. (b) Release rate coefficients, k, of released MWNTs in different eluents. *Indicates release of MWNTs was not detectable at 1.50 mM CaCl2. The error bars represent standard deviations.

charge neutralization by Ca2+ cations which, in turn, led to a lower energy barrier. However, when DI water was used in place of the 1 mM NaCl solution as the eluent, the release rate coefficient remained about the same (= 0.252 min−1), despite the decrease in ionic strength which may further decrease the energy barrier,20,33 suggesting that the release kinetics was approaching the maximum at these solution chemistries. It is conceivable that, when the MWNTs were eluded with either a 1 mM NaCl solution or DI water, the release kinetics were already approaching mass transport-limited kinetics. In addition, it is interesting to note that, although the MWNTs are expected to have a distribution of surface charge densities,20 the release process of MWNTs at all elution conditions can be fitted well using eq 5 which assumes only two populations of nanotubes (releasable and unreleasable). In several other studies on the release kinetics of colloids, in contrast, a sum of exponential decay functions with a distribution of release rate coefficients was required to obtain a satisfactory fit of the experimental data under the framework of first-order release kinetics.38−43 The approach by these studies was based on the assumption that multiple populations of colloids were released with different release rate coefficients. It is possible that the MWNTs used in this study are not as 4411



Article

be released and the release rate coefficient of releasable MWNTs can be obtained through QCM-D experiments and are important parameters for the modeling and prediction of the transport and mobility of CNTs in subsurface environments. The increase in both the kinetics and degree of MWNT release with decreasing CaCl2 concentrations observed in this study indicates that water hardness will play a critical role in controlling the reversibility of nanotube deposition in natural and engineered aquatic systems. The decline in the release rate coefficients observed at MWNT surface concentrations higher than 1000 ng/cm2 suggests that the release process of deposited CNTs can be retarded if the deposited CNT strands are interconnected to form a network of CNTs on mineral surfaces in aquatic systems. More research is needed to understand the influence of the physicochemical properties of collector surfaces (e.g., roughness and chemical heterogeneity), solution pH, natural organic matter, and hydrodynamics in the system on the release kinetics of CNTs.

■

ASSOCIATED CONTENT

S Supporting Information *

Additional figures and details for Materials and Methods. This material is available free of charge via the Internet at http:// pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]; phone: 410 516 7095. Present Address ‡

Department of Environmental Sciences, Connecticut Agricultural Experiment Station, New Haven, Connecticut 065041106. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This research was funded by the Semiconductor Research Corporation (award 425-MC-2001, project 425.041).

■

Figure 5. (a) Release rate coefficients of released MWNTs obtained at 1 mM NaCl (squares) and 0.01 mM CaCl2 (circles) were lower at elevated initial MWNT coverages (above ca. 1000 ng/cm2). The blue and red dashed lines represent the average release rate coefficients at 1 mM NaCl and 0.01 mM CaCl2, respectively, when initial surface concentrations of MWNTs were lower than ca. 1000 ng/cm2. (b) SEM images of MWNTs that were deposited on silica surfaces at 1.50 mM CaCl2. The surface concentrations of deposited MWNTs in the three images are 612, 970, and 1665 ng/cm2.

REFERENCES

(1) Miller, J. C.; Harris, D. L. The carbon nanotube patent landscape. Nanotechnol. Law Bus. 2006, 3 (4), 427−454. (2) Cao, Q.; Kim, H. S.; Pimparkar, N.; Kulkarni, J. P.; Wang, C. J.; Shim, M.; Roy, K.; Alam, M. A.; Rogers, J. A. Medium-scale carbon nanotube thin-film integrated circuits on flexible plastic substrates. Nature 2008, 454 (7203), 495−500. (3) Park, S.; Vosguerichian, M.; Bao, Z. A. A review of fabrication and applications of carbon nanotube film-based flexible electronics. Nanoscale 2013, 5 (5), 1727−1752. (4) Bello, D.; Wardle, B. L.; Yamamoto, N.; deVilloria, R. G.; Garcia, E. J.; Hart, A. J.; Ahn, K.; Ellenbecker, M. J.; Hallock, M. Exposure to nanoscale particles and fibers during machining of hybrid advanced composites containing carbon nanotubes. J. Nanopart. Res. 2009, 11 (1), 231−249. (5) 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. Potential release pathways, environmental fate, and ecological risks of carbon nanotubes. Environ. Sci. Technol. 2011, 45 (23), 9837−9856. (6) Kang, S.; Mauter, M. S.; Elimelech, M. Microbial cytotoxicity of carbon-based nanomaterials: Implications for river water and wastewater effluent. Environ. Sci. Technol. 2009, 43 (7), 2648−2653.

MWNT strands and aggregates to be significant. This observation is consistent with the theoretical predictions of Dahneke34 and Ruckenstein and Prieve33 that the release rate coefficient is proportional to the diffusion coefficient of deposited colloids. Environmental Implications. The QCM-D is used to measure the release kinetics of deposited CNTs and the same approach can potentially be used to measure the release kinetics of other classes of nanoparticles. It is experimentally verified in this study that the release of MWNTs which were deposited on silica surfaces in the primary energy minimum follows firstorder kinetics. Both the fraction of deposited MWNTs that can 4412



Article

(7) Kang, S.; Pinault, M.; Pfefferle, L. D.; Elimelech, M. Single-walled carbon nanotubes exhibit strong antimicrobial activity. Langmuir 2007, 23 (17), 8670−8673. (8) Smith, C. J.; Shaw, B. J.; Handy, R. D. Toxicity of single walled carbon nanotubes to rainbow trout, (Oncorhynchus mykiss): Respiratory toxicity, organ pathologies, and other physiological effects. Aquat. Toxicol. 2007, 82 (2), 94−109. (9) Zhao, X. C.; Liu, R. T. Recent progress and perspectives on the toxicity of carbon nanotubes at organism, organ, cell, and biomacromolecule levels. Environ. Int. 2012, 40, 244−255. (10) Chen, K. L.; Smith, B. A.; Ball, W. P.; Fairbrother, D. H. Assessing the colloidal properties of engineered nanoparticles in water: Case studies from fullerene C60 nanoparticles and carbon nanotubes. Environ. Chem. 2009, 7 (1), 10−27. (11) Petosa, A. R.; Jaisi, D. P.; Quevedo, I. R.; Elimelech, M.; Tufenkji, N. Aggregation and deposition of engineered nanomaterials in aquatic environments: Role of physicochemical interactions. Environ. Sci. Technol. 2010, 44 (17), 6532−6549. (12) Yi, P.; Chen, K. L. Influence of surface oxidation on the aggregation and deposition kinetics of multiwalled carbon nanotubes in monovalent and divalent electrolytes. Langmuir 2011, 27 (7), 3588−3599. (13) Jaisi, D. P.; Saleh, N. B.; Blake, R. E.; Elimelech, M. Transport of single-walled carbon nanotubes in porous media: Filtration mechanisms and reversibility. Environ. Sci. Technol. 2008, 42 (22), 8317− 8323. (14) Jaisi, D. P.; Elimelech, M. Single-walled carbon nanotubes exhibit limited transport in soil columns. Environ. Sci. Technol. 2009, 43 (24), 9161−9166. (15) Chang, X. J.; Bouchard, D. C. Multiwalled carbon nanotube deposition on model environmental surfaces. Environ. Sci. Technol. 2013, 47 (18), 10372−10380. (16) Hwang, Y. S.; Qu, X. L.; Li, Q. L. The role of photochemical transformations in the aggregation and deposition of carboxylated multiwall carbon nanotubes suspended in water. Carbon 2013, 55, 81− 89. (17) Yang, J.; Bitter, J. L.; Smith, B. A.; Fairbrother, D. H.; Ball, W. P. Transport of oxidized multi-walled carbon nanotubes through silica based porous media: Influences of aquatic chemistry, surface chemistry, and natural organic matter. Environ. Sci. Technol. 2013, 47 (24), 14034−14043. (18) Tian, Y.; Gao, B.; Wang, Y.; Morales, V. L.; Carpena, R. M.; Huang, Q. G.; Yang, L. Y. Deposition and transport of functionalized carbon nanotubes in water-saturated sand columns. J. Hazard. Mater. 2012, 213, 265−272. (19) Tian, Y.; Gao, B.; Wu, L.; Munoz-Carpena, R.; Huang, Q. G. Effect of solution chemistry on multi-walled carbon nanotube deposition and mobilization in clean porous media. J. Hazard. Mater. 2012, 231, 79−87. (20) Yi, P.; Chen, K. L. Influence of solution chemistry on the release of multiwalled carbon nanotubes from silica surfaces. Environ. Sci. Technol. 2013, 47 (21), 12211−12218. (21) Khan, I. A.; Berge, N. D.; Sabo-Attwood, T.; Ferguson, P. L.; Saleh, N. B. Single-walled carbon nanotube transport in representative municipal solid waste landfill conditions. Environ. Sci. Technol. 2013, 47 (15), 8425−8433. (22) Yi, P.; Chen, K. L. Interaction of multiwalled carbon nanotubes with supported lipid bilayers and vesicles as model biological membranes. Environ. Sci. Technol. 2013, 47 (11), 5711−5719. (23) Langley, L. A.; Villanueva, D. E.; Fairbrother, D. H. Quantification of surface oxides on carbonaceous materials. Chem. Mater. 2006, 18 (1), 169−178. (24) Redman, J. A.; Walker, S. L.; Elimelech, M. Bacterial adhesion and transport in porous media: Role of the secondary energy minimum. Environ. Sci. Technol. 2004, 38 (6), 1777−1785. (25) Elimelech, M.; Gregory, J.; Jia, X.; Williams, R. A. Particle Deposition and Aggregation - Measurement, Modelling and Simulation; Butterworth-Heinemann: Oxford, England, 1995.

(26) Voinova, M. V.; Rodahl, M.; Jonson, M.; Kasemo, B. Viscoelastic acoustic response of layered polymer films at fluid-solid interfaces: Continuum mechanics approach. Phys. Scr. 1999, 59 (5), 391−396. (27) Hook, F.; Kasemo, B.; Nylander, T.; Fant, C.; Sott, K.; Elwing, H. Variations in coupled water, viscoelastic properties, and film thickness of a Mefp-1 protein film during adsorption and cross-linking: A quartz crystal microbalance with dissipation monitoring, ellipsometry, and surface plasmon resonance study. Anal. Chem. 2001, 73 (24), 5796−5804. (28) Tammelin, T.; Merta, J.; Johansson, L. S.; Stenius, P. Viscoelastic properties of cationic starch adsorbed on quartz studied by QCM-D. Langmuir 2004, 20 (25), 10900−10909. (29) Yan, M.; Liu, C.; Wang, D.; Ni, J.; Cheng, J. Characterization of adsorption of humic acid onto alumina using quartz crystal microbalance with dissipation. Langmuir 2011, 27 (16), 9860−9865. (30) Sauerbrey, G. Verwendung von schwingquarzen zur wagung dunner schichten und zur mikrowagung. Z. Phys. 1959, 155 (2), 206− 222. (31) Rodahl, M.; Höök, F.; Fredriksson, C.; Keller, C. A.; Krozer, A.; Brzezinski, P.; Voinova, M.; Kasemo, B. Simultaneous frequency and dissipation factor QCM measurements of biomolecular adsorption and cell adhesion. Faraday Discuss. 1997, 107, 229−246. (32) Lee, D.; Cui, T. J. Layer-by-layer self-assembly of single-walled carbon nanotubes with amine-functionalized weak polyelectrolytes for electrochemically tunable pH sensitivity. Langmuir 2011, 27 (7), 3348−3354. (33) Ruckenstein, E.; Prieve, D. C. Adsorption and desorption of particles and their chromatographic-separation. AICHE J. 1976, 22 (2), 276−283. (34) Dahneke, B. Resuspension of particles. J. Colloid Interface Sci. 1975, 50 (1), 194−196. (35) Dahneke, B. Kinetic-theory of escape of particles from surfaces. J. Colloid Interface Sci. 1975, 50 (1), 89−107. (36) Ryan, J. N.; Gschwend, P. M. Effects of ionic-strength and flowrate on colloid release - Relating kinetics to intersurface potentialenergy. J. Colloid Interface Sci. 1994, 164 (1), 21−34. (37) Torkzaban, S.; Bradford, S. A.; Wan, J.; Tokunaga, T.; Masoudih, A. Release of quantum dot nanoparticles in porous media: Role of cation exchange and aging time. Environ. Sci. Technol. 2013, 47 (20), 11528−11536. (38) Hubbe, M. A. Detachment of colloidal hydrous oxide spheres from flat solids exposed to flow. 1. Experimental system. Colloids Surf. 1985, 16 (3−4), 227−248. (39) Hubbe, M. A. Detachment of colloidal hydrous oxide spheres from flat solids exposed to flow. 2. Mechanism of release. Colloids Surf. 1985, 16 (3−4), 249−270. (40) Grolimund, D.; Barmettler, K.; Borkovec, M. Release and transport of colloidal particles in natural porous media. 2. Experimental results and effects of ligands. Water Resour. Res. 2001, 37 (3), 571−582. (41) Grolimund, D.; Borkovec, M. Release of colloidal particles in natural porous media by monovalent and divalent cations. J. Contam. Hydrol. 2006, 87 (3−4), 155−175. (42) Grolimund, D.; Borkovec, M. Long term release kinetics of colloidal particles from natural porous media. Environ. Sci. Technol. 1999, 33 (22), 4054−4060. (43) Kallay, N.; Matijević, E. Particle adhesion and removal in model systems. 4. Kinetics of detachment of hematite particles from steel. J. Colloid Interface Sci. 1981, 83 (1), 289−300.

4413


Release Kinetics of Multiwalled Carbon Nanotubes Deposited on

Recommend Documents