Interpreting Deposition Behavior of Polydisperse Surface-Modified Nanoparticles Using QCM-D and Sand-Packed Columns

Abstract

The rising use of surface-modified engineered nanoparticles (ENPs) will result in their increased presence in aquatic environments; hence, a better understanding of their environmental fate is needed. In this study, silicon nanocrystals (Si-NCs) capped with organic acids of varying alkyl-chain length were used as model functionalized ENPs. Particle deposition kinetics were evaluated using sand-packed columns and a quartz crystal microbalance with dissipation monitoring (QCM-D). In general, an increase in solution ionic strength resulted in increased particle deposition in both columns and the QCM-D. However, the overall trends in Si-NC deposition with regard to alkyl-chain length differed in the two experimental systems, revealing how the system geometry can play a key role in defining the contribution of different particle retention mechanisms. To interpret these differences in the Si-NC deposition behavior, multiple characterization techniques were used: dynamic light scattering, nanoparticle tracking analysis, scanning ion occlusion sensing, and laser Doppler velocimetry. QCM-D also revealed insights into the influence of the particle surface coatings on particle stability. The ratio of the two QCM-D output parameters revealed that the rigidity of the particle-collector interfacial bonds varied with the alkyl-chain length, whereby particles capped with longer alkyl chains were less rigidly attached to the silica surface. Moreover, it is shown that the interpretation of ENP deposition behavior using QCM-D is limited by the presence of large-particle aggregates (≥700 nm in this study) which do not fully couple to the QCM-D sensor. Under such conditions, QCM-D measurements of ENP deposition should be interpreted with caution as the microbalance response cannot be directly considered as deposited mass. This study improves our understanding of the role that surface modifiers and ENP aggregates play in ENP deposition kinetics in efforts to predict the transport and fate of ENPs in natural and engineered aquatic environments.

Introduction

The unique physical and optical properties of engineered nanoparticles (ENPs) are being exploited for diverse applications such as biomedical imaging, solar cells, or electronic devices (Klaine et al., 2008; Algar et al., 2011). To target specific needs, ENPs are commonly coated with different organic molecules (e.g., surfactants, polymers, or polyelectrolytes) (Hotze et al., 2010; Phenrat et al., 2010). For instance, TiO₂ nanoparticles stabilized with polyacrylic acid conjugated with antibodies can be used for molecular recognition (Kanehira et al., 2008). Amphiphilic polymers render quantum dots water dispersible and ready for use in biological sensing and intracellular labeling (Algar et al., 2011). Likewise, polymeric coatings (e.g., carboxymethyl cellulose, polyvinylpyrrolidone, and guar gum) improve the stability of reactive zerovalent iron nanoparticles (nZVI) and the capability of bimetallic FePd nanoparticles to degrade chlorinated hydrocarbons (Sakulchaicharoen et al., 2010; Raychoudhury et al., 2012). However, an increasing number of products containing surface-modified ENPs (Roco et al., 2011) will likely result in their incidence in the environment (Westerhoff et al., 2013), and since different ENPs have exhibited toxic effects (Klaine et al., 2008), understanding their transport and fate is crucial for assessing their contamination potential (Petosa et al., 2010).

Silicon-based nanocrystals (Si-NCs) are one example of ENPs that exhibit luminescent properties and are promising as biomedical tracers (O'Farrell et al., 2006; Park et al., 2009). However, despite the apparent innocuous properties of Si, some studies with silicon-based nanoparticles have revealed toxic effects related to their physicochemical properties (i.e., particle size, surface charge, and surface coating) (Fujioka et al., 2008; Choi et al., 2009; Ruizendaal et al., 2009). Since the eventual manufacturing of consumer products containing surface-modified silicon nanoparticles is imminent (www.nanotechproject.org), a clear understanding of the colloidal stability and mobility of silicon-based nanomaterials (i.e., Si-NCs) is needed to predict their fate in aquatic environments.

In general, the mobility of ENPs has been studied using water-saturated columns packed with granular materials (e.g., quartz sand) representative of aquifer sediments or engineered (deep-bed) filtration systems (Petosa et al., 2010). In column transport experiments, the extent of ENP retention is evaluated by monitoring the influent and effluent particle concentrations, and the results are commonly interpreted using clean-bed colloid filtration theory (CFT) (Yao et al., 1971). Several studies of ENP transport in packed columns report an enhanced mobility of surface-modified ENPs due to the electrosteric stabilization imparted by the polymers or polyelectrolytes used to coat them (Phenrat et al., 2010; Petosa et al., 2012; Mitzel and Tufenkji, 2014). Nonetheless, generalization to the myriad of surface-modified ENPs introduced into the market every year (Roco et al., 2011) is complex, as the afforded level of electrosteric stability is associated to the properties of the molecules used (i.e., molecular weight, chemistry, adsorbed mass, density, and extended layer thickness) (Phenrat et al., 2010). Thus, a better mechanistic understanding of the role that surface modifiers play is needed to more accurately predict the transport and fate of ENPs in natural and engineered aquatic environments.

The quartz crystal microbalance with dissipation monitoring (QCM-D) is another useful technique for evaluating the deposition behavior of ENPs on model aquifer grain surfaces (e.g., SiO₂, Fe₂O₃ or Al₂O₃) (Chen and Elimelech, 2006; Quevedo and Tufenkji, 2009; Lin and Wiesner, 2012; Liu et al., 2012). Briefly, as particles deposit onto the QCM sensor, an increase in mass on the collector surface results in a measurable decrease in the crystal's resonance frequency (i.e., negative frequency shifts) (Sauerbrey, 1959). In addition, dissipative energy losses, induced by the deposited mass, are measured in terms of oscillation decay time. The applicability of QCM-D as a mass sensor to assess the deposition kinetics of ENPs is well known (Chen and Elimelech, 2006; Quevedo and Tufenkji, 2009; Fatisson et al., 2010), whereas it was only recently shown that the ratio of the two QCM-D output parameters (dissipation/frequency) can be used to better understand the conformation of the surface coatings on deposited particles (Quevedo et al., 2013). Namely, the dissipation/frequency ratio revealed whether the conformation of a polyelectrolyte (i.e., polyacrylic-acid derivative) used to coat a CdTe QD was extended or compressed upon deposition onto alumina surfaces.

Regardless of the experimental technique utilized (e.g., column studies or QCM-D), nanoparticle deposition kinetics are generally dependent on the water chemistry (e.g., electrolyte species, ionic strength [IS], and pH) as well as on collector and particle surface properties (Petosa et al., 2010). However, changes in water chemistry often result in particle aggregation, which can be challenging to interpret (Domingos et al., 2009; Petosa et al., 2010). For instance, the retention of particle aggregates in packed granular matrices may increase due to deposition in the secondary energy minimum of the particle-collector interaction energy profile (Hahn et al., 2004; Pelley and Tufenkji, 2008; Saleh et al., 2008) or physical straining in pore spaces that are too small to enable passage of the aggregates (Bradford et al., 2007; Petosa et al., 2012). Likewise, in QCM-D experiments, the deposition of aggregated ENPs might not always be straightforward to interpret using the conventional approach of mass loading (Fatisson et al., 2009; Pomorska et al., 2010; Lin and Wiesner, 2012; Liu et al., 2012). For instance, positive frequency shifts associated with large particles or ENP aggregates attached to the sensor via weak bridges have been reported (Fatisson et al., 2009; Pomorska et al., 2010), and this atypical response appears to be linked to the size of the particle aggregate (Pomorska et al., 2010). Hence, laboratory studies to assess the transport and deposition of ENPs in packed-bed columns and QCM-D require an appropriate physicochemical characterization to adequately interpret the results.

In this study, we have systematically evaluated and compared the deposition behavior of Si-NCs as a model surface-modified ENP, using (i) laboratory-scale columns packed with quartz sand, and (ii) QCM-D with SiO₂-coated crystals (as a model sand surface). Both techniques were used to determine the deposition kinetics of surface-modified Si-NCs over a broad range of environmentally relevant solution chemistries, including variations in IS and ion valence (K⁺, Ca²⁺). The Si-NCs were modified with carboxylic acids having different chain lengths, as an increase in carbon chain length has been found to improve their colloidal stability in aqueous suspensions (Clark et al., 2010). Various techniques were also used to characterize the physicochemical properties of the Si-NCs over the range of water chemistries examined, including laser Doppler velocimetry, to evaluate particle electrophoretic mobility (EPM), dynamic light scattering (DLS), nanoparticle tracking analysis (NTA), and scanning ion occlusion sensing (SIOS) to evaluate particle and aggregate sizes.

Experimental Protocols

Preparation of Si-NC suspensions

Si-NCs were prepared and functionalized with carboxylic acids of different alkyl-chain lengths: propionic acid (C₃), heptanoic acid (C₇), and undecanoic acid (C₁₁) and dispersed in deionized water (DI) as described in the Supplementary Data. A systematic characterization of the functionalized Si-NCs by infrared spectroscopy (FTIR) and X-Ray photoelectron spectroscopy has proved that all the carboxylic acids are attached to the particle surface (Clark et al., 2010). Si-NC stock suspensions were diluted to a working concentration of 20 or 200 mg/L in electrolytes of varying IS (1–100 mM KCl or 1–3 mM CaCl₂) for transport experiments in packed bed columns or QCM-D, respectively. Suspensions were adjusted to pH 5 (using HCl) and equilibrated for 2 h at 9°C and then at room temperature for 30 min before each experiment. All chemicals used to prepare solutions were of analytical grade.

Electrokinetic characterization of Si-NCs in electrolytes

Laser Doppler velocimetry (ZetaSizer Nano; Malvern) was used to evaluate the EPM of Si-NCs within the range of IS. EPM measurements were repeated using at least three different samples (prepared on different days) using disposable capillary cells at 20°C±0.2°C. The electrical field (E) was adjusted between 5 and 10±0.1 V/m. For comparative purposes, measured EPMs were converted to zeta potentials using the Henry equation and the expression for the retardation effect proposed by Ohshima (1994).

Sizing of Si-NCs in electrolytes

DLS (number mean) (ZetaSizer Nano; Malvern) and NTA (LM10-HS Nanoparticle Analysis System; Nanosight) were used to determine the hydrodynamic diameters of the Si-NCs suspended in electrolytes before each transport experiment. For NTA measurements, the mean square displacements of single particles were determined by tracking the scattered light, followed by analysis using the software (NanoSight v.2.2.), where each data point represents the average of at least 150 counted particles. At selected conditions, SIOS was used to confirm the size of the Si-NCs in electrolytes (qNano; IZON Science). In SIOS, a Coulter-type sensor enables single-particle measurements, as colloidal particles are driven through pores of a known size (Vogel et al., 2011). Membranes with different pore sizes (NP₁₀₀, NP₂₀₀, and NP₄₀₀) were mounted on the instrument and the translocation of Si-NCs through the stretched pores was detected as a result of changes in the background current (Vogel et al., 2011). Calibration with standard polystyrene latex beads of known sizes enabled the determination of the size of Si-NCs used in this study, following the procedure suggested by the vendor (IZON Science). In all cases, results are presented as the mean of size measurements conducted in triplicate.

Transport experiments in packed-bed columns

Experiments were conducted with 1 cm (internal diameter) glass chromatography columns (10/20; Amersham Biosciences) that were packed with high-purity quartz sand. Before use, the sand was sieved with Nylon sieves (U.S. standard mesh sizes 50 and 70, 300, and 212 μm openings, respectively) to obtain a sand sample having an average grain diameter (d_c) of 256 μm. The sand was cleaned according to the procedure previously described (Pelley and Tufenkji, 2008). Before each transport experiment, the required mass of sand was soaked in electrolytes for at least 16 h. To ensure uniform packing, sand was wet packed into the glass column using gentle vibration yielding a packed-bed porosity (ɛ) of 0.38. The final height of the packed bed was 4±0.2 cm. To condition the packed columns, 10 pore volumes (PVs) of background electrolyte (i.e., KCl or CaCl₂) were pumped at a constant approach velocity of 1.06×10⁻⁴ m/s before the injection of Si-NCs. In all experiments, 8 mL of each Si-NC suspension (equivalent to 6.6 PVs) were injected into the packed column. The effluent concentrations (C) of fluorescent Si-NCs (i.e., C₇ and C₁₁) were monitored in real time with a spectrofluorometer (Fluoromax-4, Jobin-Yvon Horiba) (EX/EM settings were 350/603 for the C₇, and 350/611 for C₁₁). The effluent concentration of the nonfluorescent Si-NC (C₃) was monitored with a UV-visible spectrophotometer (Hewlett Packard, 8453) (at a wavelength of 250 nm) and, under selected conditions, confirmed by ICP-OES (see Supplementary Data). The concentrations of the influent suspensions (C₀) were the same as those used for characterization of the particles. The absorbance or fluorescence of the injected suspensions was determined by bypassing the column apparatus. Si-NC transport experiments were conducted over a wide range of solution IS, using both monovalent (KCl) and divalent salts (CaCl₂), and conducted in duplicate for each condition.

Interpretation of transport experiments in packed-bed columns

Results from sand-packed column experiments were quantitatively compared by determining the particle attachment efficiency (α) using CFT (Yao et al., 1971) as follows: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}\alpha = - \frac { 2 } { 3 } \frac { d_c } { ( 1 - \varepsilon ) L \eta_0 } \ln \left( \frac { C } { C_0 } \right) \qquad \qquad \qquad \qquad (1) , \end{align*} \end{document}

where d_c is the average diameter of the sand grains, ɛ is the porosity of the porous medium, and L is the packed column length. The value of C/C₀ represents the normalized column effluent particle concentration obtained by numerical integration of each particle breakthrough curve. Values of the predicted single-collector contact efficiency (η₀) for each experimental condition were determined using the correlation developed by Tufenkji and Elimelech (2004).

Deposition of Si-NCs onto SiO₂ measured by QCM-D

Deposition rates of Si-NCs onto SiO₂-coated sensors (QSX–303; Q-Sense AB) were determined using a QCM-D (E4; Q-Sense AB). When the 5 MHz AT-cut crystals are mounted in the QCM-D flow modules, the injected flow is parallel to the flat SiO₂ surface. Si-NC deposition experiments were repeated using at least three different samples prepared on different days and conducted as follows: A peristaltic pump (Reglo-Digital IPC-N 4; Ismatec) was used to first inject DI (50 μL/min) to obtain a stable baseline. Next, the background electrolyte solution of interest was injected into the flow module, followed by a Si-NC suspension in the same electrolyte solution for at least 20 min. After each experiment, the SiO₂ crystals, QCM-D chambers, and tubing were cleaned following the protocol suggested by Q-Sense. First, the instrument was rinsed with 20 mL of 2% Hellmanex (Fisher Scientific) and 30 mL of DI. The crystals were further cleaned by soaking in 2% Hellmanex overnight, rinsed with DI, dried under N₂, and then exposed to UV/ozone treatment for 10 min before any additional measurement.

Interpretation of QCM-D deposition experiments

The QCM-D is commonly referred to as a mass sensor, because changes in the resonance frequency (−Δf) of a quartz crystal are proportional to the deposited mass (Δm) per unit area. Dissipative energy losses (dissipation shift, ΔD) of the deposited film are deduced from the oscillation decay times of each oscillation cycle. In the case of rigidly adsorbed films (ΔD ∼0), frequency shifts (−Δf_(n)) are proportional to the deposited mass (Δm), as described by the Sauerbrey relationship (1959): \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}- \Delta f_ { ( { \rm n } ) } = - \frac { \Delta f_ { { n } } } { n } = \frac { \Delta m } { C } \tag { 2 } \end{align*} \end{document}

In Equation (2), C is the mass sensitivity constant, equal to 17.7 ng/[cm²·Hz] for a 5 MHz crystal, and n is the resonance overtone number (n=3, 5, 7, 9, 11, 13). It should be noted that with standard software settings, the frequency shift obtained from QCM-D (Q-Sense) is already normalized (i.e., Δf_(n)=Δf_n/n). When the Sauerbrey relation (1959) is valid, the rate at which frequency shift changes during a given time period [−f_{(n)_slope}, Eq. (3)] describes the rate at which mass is deposited on the SiO₂ surface (r_d): \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}r_ { \rm d } = - f_ { ( { n } ) \_ { slope } } = - \frac { d \Delta f_ { ( { n } ) } } { dt } \tag { 3 } \end{align*} \end{document}

As discussed later in this article, this is a valid assumption if the normalized frequency shifts at all the overtones are negative (i.e., \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} $$\Delta f_{ ( 3 ) } , \Delta f_{ ( 5 ) } , \ldots \Delta f_{ ( 13 ) } \leq 0$$ \end{document} . Conversely, positive frequency shifts at high overtones (n) are a clear indication of a “coupled resonance” type response, which is linked to the bond strength between particles and collector, but cannot be interpreted as deposited mass (Pomorska et al., 2010; Olsson et al., 2012).

Results and Discussion

Characterization of capped Si-NCs in electrolyte

EPM and size of the studied Si-NCs were assessed before each transport and deposition experiment in the presence of monovalent (KCl) and divalent (CaCl₂) salts (Table 1). All the Si-NCs studied are negatively charged under the experimental conditions (pH 5, solution IS) due to the presence of carboxyl groups in the surface cappings (C₃, C₇, C₁₁). The experimental data (Table 1) reveals that as the IS increases, compression of the electrical double layer (EDL) results in lower EPM values (Elimelech et al., 1995). Between 1 and 100 mM KCl, more important variations in the EPM values are observed for Si-NCs capped with C₃ (i.e., EPM varies between −2.3 and −0.7 μm·cm/V·s), as compared with particles capped with the longest chain length, namely C₁₁ (where EPM only varies between −1.1 and −0.7 μm·cm/V·s). Variations in the electrokinetic properties of surface-modified ENPs as a function of IS are commonly associated with the properties of the capping ligands at the particle surface (Phenrat et al., 2010).

Table 1.

Electrokinetic Properties and Particle size (d_p) of Si-NCs as a Function of Solution IS in KCl or CaCl₂ at pH 5 Determined Using Different Techniques

						Particle size (d_p)
Electrolyte	Ionic strength (mM)	Si-NC	Electrophoretic mobility (μm·cm/V·s)	Zeta potential (mV)	PDI	DLS (nm)	NTA (nm)	SIOS (nm)
KCI	1		−2.3±0.1	−42±2	0.32	86±56	72±4
	10		−2.0±0.2	−29±2	0.40	478±151
	30	C₃	−1.2±0.1	−17±2	0.53	578±221	253±33
	50		−0.9±0.1	−13±2	0.53	979±460	245±189	766±43
	100		−0.7±0.1	−10±1	0.54	703±261	278±245	862±53
KCI	1		−1.8±0.2	−37±5	0.34	48±14	97±47
	10		−1.6±0.1	−25±1	0.48	135±36
	30	C₇	−0.6±0.1	−8±2	0.65	847±185	179±42
	50		−0.7±0.2	−9±2	0.74	949±308	163±85	140±20
	100		−0.7±0.1	−11±2	0.87	761±146	362±174	150±10
KCI	1		−1.1±0.1	−20±1	0.23	92±35	98±34
	10		−1.1±0.3	−18±4	0.26	70±40
	30	C₁₁	−1.0±0.5	−15±8	0.29	69±12	96±10
	50		−0.9±0.3	−14±5	0.16	85±25	92±68	140±16
	100		−0.7±0.2	−10±3	0.26	122±23	174±86	155±8
CaCl₂	0.3		−2.0±0.1	−37±3	0.34	117±10	140±6
	1.5	C₃	−1.2±0.1	−18±2	0.87	1125±339	278±32
	3		−0.7±0.1	−11±2	0.72	1066±125	268±70
CaCl₂	0.3		−1.3±0.4	−18±6	0.15	31±9	62±13
	1.5	C₇	−0.4±0.2	−5±3	0.97	810±273	135±5
	3		−0.6±0.1	−9±1	0.92	1246±393	265±41
CaCl₂	0.3		−0.9±0.1	−13±1	0.31	97±10	89±35
	1.5	C₁₁	−0.5±0.5	−7±7	0.79	481±224	162±45
	3		0.0±0.7	0±10	0.67	930±671	202±41

Data represent the mean diameter±standard deviation.

DLS, dynamic light scattering; IS, ionic strength; Si-NCs, silicon nanocrystals; SIOS, scanning ion occlusion sensing.

In divalent salt (CaCl₂), the EPM changes from −2.0 to −0.7 μm·cm/V·s for the C₃, from −1.3 to −0.4 μm·cm/V·s for the C₇, and from −0.9 to −0.03 μm·cm/V·s for the C₁₁ when the IS increases from 0.3 to 3 mM CaCl₂. In Table 1, EPM measurements are converted to zeta potential, using the particle sizes (d_p) obtained by DLS, and these values are used later in this article to interpret the deposition results within the context of the Derjaguin-Landau-Verwey-Overbeek (DLVO) theory of colloidal stability.

Particle size plays an important role in the transport, fate, and bioavailability of ENPs in aquatic environments (Petosa et al., 2010). Hence, to interpret the observed transport and deposition behavior of the Si-NCs, an appropriate selection of size characterization methods is required (Domingos et al., 2009). It has been shown that the length of the ligand chain significantly influences the colloidal stability (i.e., aggregation) of Si-NCs in polar solvents. Clark et al. (2010) determined that the size of the “as synthesized” capped Si-NCs used here was in the order of 3 nm, and after dispersion in DI, Si-NCs capped with C₃ and C₇ experienced aggregation (Supplementary Fig. S1). In Table 1, the mean diameters (d_p) for each Si-NC suspended in monovalent (KCl) and divalent electrolytes (CaCl₂) using three different methods (i.e., DLS, NTA, and SIOS) are shown. DLS measurements show that the d_p of Si-NCs capped with C₃ and C₇ increases by one order of magnitude within the range of IS investigated (1–100 mM KCl). In contrast, the particle capped with C₁₁ is more stable over the range of IS. NTA measurements generally confirm the trends observed by DLS by which Si-NCs capped with C₃ and C₇ exhibit larger d_p as compared with those capped with C₁₁. Nonetheless, the mean hydrodynamic diameters obtained by NTA are typically smaller than those obtained by DLS (except for C₁₁ suspended in KCl), and these discrepancies are likely due to the interpretation of particle size with each technique. DLS measurements are influenced by the presence of small amounts of aggregates or dust particles (i.e., scattered light intensity is proportional to d_p⁶), whereas NTA tracks individual particles and the presence of a few aggregates in suspension is not expected to bias the results as strongly as in DLS (Domingos et al., 2009). However, since the resolution of NTA is limited by the refractive index of the material and larger particles or aggregates (∼1 μm) are not tracked by this technique (Gallego-Urrea et al., 2011), the use of a third method is of interest to support the characterization of the Si-NCs (Domingos et al., 2009; Bell et al., 2012). SIOS is a recently introduced nonlight scattering technique that promises to improve the size assessment of polydisperse samples based on the Coulter principle (Vogel et al., 2011). In SIOS, particles suspended in electrolytes pass through membranes with size tunable pores when a fixed electrical potential is applied. Changes in the ionic current as a result of particles traversing the pores are defined as “blockades” and their amplitude can be related to the size of the particle passing through (Vogel et al., 2011). In Table 1, mean sizes determined by SIOS at 100 mM KCl reveal significant aggregation for the particle capped with C₃ (d_p ∼862 nm), and smaller sizes for C₇ (d_p ∼150 nm) and C₁₁ (d_p ∼155 nm). Figure 1 shows a comparison of Si-NC size distributions measured at 100 mM KCl using the three methods (DLS, NTA, and SIOS) that may help clarify the differences in the mean size of the suspensions (Table 1).

FIG. 1.

Particle-size distributions determined by DLS, NTA, and SIOS of Si-NCs capped with (a) C₃, (b) C₇ and (c) C₁₁ in 100 mM KCl at pH 5. DLS, dynamic light scattering; NTA, nanoparticle tracking analysis; Si-NCs, silicon nanocrystals; SIOS, scanning ion occlusion sensing.

In Figure 1a and b, Si-NCs capped with C₃ and C₇ exhibit broad size distributions and consequently, the mean sizes obtained are expected to be biased by the limitations mentioned earlier for the DLS and NTA techniques (Domingos et al., 2009; Gallego-Urrea et al., 2011). The particles capped with C₁₁ (Fig. 1c) exhibit normal distributions with comparable mean sizes for the ensemble of characterization techniques corroborating their stability in aqueous suspensions. These results are in agreement with previous studies (Vogel et al., 2011; Bell et al., 2012), where sizes determined by SIOS have been compared with different characterization techniques; namely DLS, NTA, and transmission electron microscopy measurements.

DLS size measurements conducted with Si-NCs suspended in CaCl₂ reveal that considerable particle aggregation occurs, regardless of the chain length of the capping ligand (Table 1). For instance, the presence of 1.5 mM CaCl₂ seems sufficient to reach the critical coagulation concentration (CCC), which is supported by the large standard deviations observed at this IS. In CaCl₂, NTA measurements do not appear to track particle aggregates, and SIOS measurements are not feasible at the studied range of IS, due to the low conductivity of the suspensions. The results obtained here are consistent with previous studies where considerable aggregation of carboxyl-modified ENPs was observed, even in the presence of low concentrations of divalent salt (Liu et al., 2012; Petosa et al., 2012; Quevedo and Tufenkji, 2012).

Transport of capped Si-NCs in saturated sand-packed columns

Supplementary Figure S2 shows representative particle breakthrough curves for experiments conducted with Si-NCs suspended in KCl, and in Table 2 the normalized column effluent particle concentrations (C/C₀) are reported as a function of IS. For all three Si-NCs, the measurements show a decrease in C/C₀ with an increasing solution of IS. However, the capping ligand with the longest chain (C₁₁) seems to improve the mobility of Si-NCs up to a moderate salt concentration (30 mM IS). For instance, in experiments conducted at 30 mM KCl, the particles capped with C₁₁ exhibit lower retention (∼24%) as compared with those capped with C₃ (∼38%) and C₇ (∼56%). However, at 100 mM KCl, more than 50% of all the Si-NCs are retained.

Table 2.

Summary of Experiments Conducted in Saturated Quartz Sand Columns with KCl and CaCl₂ at pH 5

			Effluent concentration	Attachment efficiency	DLVO interaction parameters ^a		Particle size ^b	Physical straining ^c
Electrolyte	Ionic strength (mM)	Si-NC	C/C₀	α	Φ_max (k_BT)	Φ_2°min (k_BT)	d_p (nm)	d_p/d_c≥0.003
KCI	1		0.98±0.0	0.01±0.00	108	0	86±56	No
	10		0.79±0.0	0.15±0.01	100	2	478±151	No
	30	C₃	0.62±0.1	0.32±0.06	76	8	578±221	No
	100		0.48±0.0	0.51±0.02	NB		703±261	No
KCI	1		0.97±0.0	0.01±0.00	45	0.0	48±14	No
	10		0.72±0.1	0.09±0.03	29	0.6	135±36	No
	30	C₇	0.44±0.0	0.57±0.07	NB		847±185	Yes
	100		0.43±0.0	0.59±0.02	NB		761±146	Yes
KCI	1		0.96±0.0	0.01±0.01	33	0	92±35	No
	10		0.83±0.1	0.03±0.01	7	0	70±40	No
	30	C₁₁	0.76±0.0	0.04±0.00	NB		69±12	No
	100		0.37±0.0	0.24±0.02	NB		122±23	No
CaCl₂	0.3		0.94±0.0	0.01±0.00			117±10	No
	1.5	C₃	0.19±0.0	1.02±0.11			1125±339	Yes
	3		0.17±0.0	1.12±0.02			1066±125	Yes
CaCl₂	0.3		0.95±0.0	0.00±0			31±9	No
	1.5	C₇	0.17±0.0	1.24±0.00			810±273	Yes
	3		0.11±0.0	1.23±0.02			1246±393	Yes
CaCl₂	0.3		0.84±0.1	0.04±0.02			97±10	No
	1.5	C₁₁	0.07±0.0	1.70±0.13			481±224	No
	3		0.04±0.0	2.29±0.41			930±671	Yes

Calculated classical DLVO interaction energy parameters (Φ_max and Φ_2°min) reported in units of k_BT for the Si-NCs interacting with quartz sand. To calculate the retarded van der Waals forces, a Hamaker constant of 1.92×10⁻²⁰ J was selected for the nanoparticle–water–quartz system (Senden and Drummond, 1995); whereas for electrical double-layer interactions, constant-potential values were calculated using the zeta potential of the Si-NCs (Table 1) and the quartz sand in place of their respective surface potentials. The values of zeta potential used for the quartz sand were obtained from Redman et al. (2004).

The DLS measured particle sizes of the Si-NCs (Table 1) were used in the DLVO and straining calculations.

The transport experiments were conducted in fine silica sand (d_c=256 μm), and straining may play a significant role in deposition when d_p/d_c is as low as 0.003 (Bradford et al., 2007).

Conditions where straining is likely to play an important role are indicated in boldface type.

DLVO, DerJaguin-Landau-Verwey-Overbeek; NB, no energy barrier.

Figure 2 summarizes the calculated attachment efficiencies (α) for the surface-modified Si-NCs. For the sake of clarity, the α values were determined using only the d_p obtained by DLS.

FIG. 2.

Attachment efficiencies (α) calculated using eq 1 for Si-NCs capped with: (□) C₃, (○) C₇, and (◊) C₁₁ suspended in electrolyte over a range of IS (pH 5). Experiments were conducted using columns packed with clean quartz sand, and Si-NCs were suspended in (a) KCl (open symbols) or (b) CaCl₂ (full symbols). Dashed lines are included as eye guides, and error bars represent standard deviations. IS, ionic strength

When suspended in KCl (Fig. 2a), the α values for the Si-NCs increase significantly over the range of IS (α values varied from 0.003 to 0.51 for the C₃; from 0.002 to 0.59 for the C₇; and from 0.01 to 0.24 for C₁₁). Notably, α does not reach the theoretical mass-transfer limited value (∼1) even at the maximum IS (100 mM) for any of the Si-NCs investigated. Based on the physicochemical characterization of the Si-NCs (Table 1), the EPM values decrease with an increasing KCl concentration and, consequently, the Si-NCs experience less electrostatic repulsion as they approach the sand grain surface, thereby resulting in increased particle deposition (Elimelech et al., 1995).

To determine the influence of divalent cations on the transport and deposition behavior of the Si-NCs, column experiments were also carried out with Si-NCs suspended in CaCl₂. Unlike experiments conducted in KCl, all the particles (i.e., C₃, C₇, and C₁₁), regardless of their alkyl-chain length, exhibit limited transport at low IS of the divalent salt solution (1.5–3 mM IS CaCl₂) (Supplementary Fig. S3, Table 1). These trends are likely linked to the ability of divalent cations (such as Ca²⁺) to bridge negatively charged groups on the Si-NC surface with the sand surface (Table 1). The drastic increase in α at 1.5 mM IS CaCl₂ (Fig. 2b) and the high α values (>1) at 3 mM IS CaCl₂ suggest that the particles are retained by an additional mechanism of physical straining in the packed column (Tufenkji et al., 2004; Petosa et al., 2012). Physical straining occurs when ENPs aggregate to sizes (d_p) that are larger than the pore spaces in the granular matrix (Bradford et al., 2007). As observed in Table 2, this hypothesis is supported by the physical properties of the particles (significant aggregation occurs at this IS) and the porous medium; namely, the transport experiments were conducted in fine silica sand (d_p/d_c≥0.003), and the angularity of the sand grains is high (roundness factor ∼0.67), which increases the likelihood of physical straining (Tufenkji et al., 2004; Pelley and Tufenkji, 2008; Saleh et al., 2008; Petosa et al., 2012).

To better understand the mechanisms governing the particle deposition behavior in monovalent salt solution, interaction energy profiles between the particles and the grain surface were calculated using classical DLVO theory. Within the context of classical DLVO theory (which does not take into account the contribution of steric forces), solution chemistry and particle (or aggregate) size have important effects on the height of the repulsive energy barrier (Φ_max), as well as on the depth of the secondary energy well (Φ_2°min) (Hahn et al., 2004). Hence, in Table 2, Φ_max and Φ_2°min are presented for each capped Si-NC as a function of KCl IS. The total interaction potentials were calculated as the sum of van der Waals forces and EDL interactions using the linear superposition approximation and assuming the system can be approximated by a sphere-plate geometry (Elimelech et al., 1995).

In Table 2, the calculated height of the interaction energy barrier (Φ_max) for the three Si-NCs generally decreases as the IS of KCl increases. Likewise, each Si-NC exhibits greater retention in the sand-packed column with increasing KCl concentration (Fig. 2a). Based on these simple, idealized classical DLVO calculations, the Si-NCs capped with the longest alkyl chains (C₁₁) are expected to deposit to a greater extent onto the sand surface than the C₇, and the C₇ are, in turn, expected to deposit more than the C₃, for which greater energy barriers are predicted at any given IS. However, this is not always the case; for example, in 10 mM KCl, where α C₃>α C₇>α C₁₁. This observation can be attributed to three main factors: (i) the particle capped with C₃ forms fairly large aggregates and may, thus, deposit in a secondary energy well (Φ_2°min) at a separation distance greater than that of the energy barrier (Hahn et al., 2004) (Table 2); (ii) the larger C₃ and C₇ aggregates may be retained by physical straining; and (iii) the deposition of the Si-NCs capped with longer alkyl chains (C₇ and C₁₁) is hindered by a repulsive steric force (which is not accounted for in the DLVO calculations). Interestingly, in 30 mM KCl, α C₇>α C₃>α C₁₁. At this IS, DLVO calculations predict no energy barrier for the C₇ particle, yet the attachment efficiency does not reach the mass transport limited rate (∼1), suggesting that the longer alkyl chain at the particle surface may be exerting some steric repulsion on the approach to the sand surface. The lower energy barriers predicted for the particle capped with C₁₁ do not correlate with the low attachment efficiencies (α) measured in the transport experiments. These results suggest that the longest alkyl chain (C₁₁) exerts significant steric repulsion, thereby hindering particle deposition on the quartz sand (Hotze et al., 2010; Phenrat et al., 2010).

Certain surface modifiers (e.g., polymers, polyelectrolytes, and surfactants) are known to inhibit the attachment of nanoparticles on sand or model silica surfaces as compared with their bare counterparts (Phenrat et al., 2010; Petosa et al., 2012; Quevedo and Tufenkji, 2012). However, some studies have demonstrated that the stabilizing effects exerted by surface coatings are not always concordant (Saleh et al., 2008; Lin et al., 2012). Saleh et al. (2008) used packed column experiments to study the transport of nZVI capped with different surface modifiers. Low-molecular-weight surfactants (sodium dodecyl benzene sulfonate) did not effectively enhance the mobility of nZVI, whereas particles modified with high-molecular-weight polymers exhibited greater mobility due to electrosteric stabilizing effects. In another study, the transport of silver nanoparticles coated with different molecules (citrate, polyvinylpyrrolidone, and gum arabic) was studied in columns packed with glass beads (Lin et al., 2012). These authors observed an enhanced attachment of silver nanoparticles to glass surfaces when the hydrophobicity of the particle coatings and surfaces increased, despite the electrosteric repulsion afforded by the surface modifier. The results of our column study with the three Si-NCs suggest that the longest alkyl chain exerts considerable electrosteric repulsion on the approach to the sand surface; this mechanism also impairs the deposition of the C₇ particle, but not to the same extent. Overall, the data show that a complex interplay of DLVO and non-DLVO forces and physicochemical processes contribute to the transport and deposition behavior of the surface-modified Si-NCs.

Deposition of Si-NCs in QCM-D

In the second part of this study, SiO₂-coated sensors mounted in a QCM-D were used as model collectors representing the surface of the sand grains used in the column experiments. Since the hydrodynamic conditions as well as the physicochemical properties of the collectors used (i.e., quartz sand vs. SiO₂-coated flat sensors) differ between the columns and QCM-D experiments, the results obtained using the two systems are only qualitatively comparable. Values of the Si-NC deposition rate (r_d) when suspended in KCl or CaCl₂ were determined from initial changes in frequency shifts (Δf_(n)) as a function of time [Eq. (3)] (Supplementary Fig. S4 and Table 3). The data are presented in terms of mass units (i.e., ng/cm²·min) using the Sauerbrey relation [Eq. (2)] and frequency responses from the third overtone (i.e., Δf₍₃₎). Moreover, the dissipative energy losses (ΔD) at each experimental condition are presented as ΔD₃/Δf₍₃₎ ratios (or, the equivalent −ΔD₃__slope to Δf₍₃₎__slope) (Fig. 3). The ΔD₃/Δf₍₃₎ ratio provides useful insights about the particle-surface bond stiffness (Johannsmann et al., 2009), particle size (Olsson et al., 2013), and conformation of the polymeric coatings used to stabilize the particles (Quevedo et al., 2013).

FIG. 3.

Ratio −ΔD_{3_slope}/Δf₍_{3)_slope} for all the Si-NCs as a function of IS in KCl. The dotted line represents the limit of applicability for the Sauerbrey model according to Q-Sense (i.e., 10⁻⁷ Hz⁻¹). Dashed lines are included as eye guides, and error bars represent standard deviations.

Table 3.

Experimental Deposition Rates (r_d) from QCM-D Measurements, and Particle Size (d_p − DLS) at the Different Experimental Conditions

			QCM-D ^a	Size	Coupled resonance?
Electrolyte	Ionic strength (mM)	Si-NC	r_d (ng/cm².min)	d_p (nm)	Δf _(n)>0^b
KCI	1		3±1	86±56	No
	10	C₃	88±2	478±151	No
	30		91±9	578±221	No
	100		79±9	703±261	Yes
KCI	1		2±3	48±14	No
	10	C₇	1±0	135±36	No
	30		5±3	847±185	No
	100		6±1	761±146	Yes
KCI	1		0±0	92±35	No
	10	C₁₁	20±27	70±40	No
	30		30±23	69±12	No
	100		10±6	112±23
CaCl₂	0.3	C₃	186±129	117±10	No
	1.5		42±28	1125±339	Yes
	3		14±3	1066±125	Yes
CaCl₂	0.3	C₇	305±15	31±9	No
	1.5		18±16	810±273	Yes
	3		2±1	1246±393	Yes
CaCl₂	0.3	C₁₁	15±16	97±10	No
	1.5		2±1	481±224	No
	3		1±2	930±671	Yes

Data represent the mean±standard deviation.

In all cases, r_d was determined from 3rd overtone measurements.

Frequency shifts (Δf_(n)) obtained at different overtones (n=3, 5, 7, 11, 13) were compared and when positive responses were observed (i.e., coupled resonance), it has been indicated using boldface type.

QCM-D, quartz crystal microbalance with dissipation monitoring.

In experiments conducted in KCl, the Si-NCs capped with the longer alkyl chain lengths (C₇, C₁₁) exhibit low deposition onto the SiO₂ surface in the QCM-D; whereas the C₃ particle deposition rate (r_d) increases with an increasing salt concentration up to 30 mM IS (Table 3). Between 1 and 30 mM KCl, r_d values for Si-NCs capped with C₇ and C₁₁ increase from 2 to 5 and 0 to 30 ng/cm²·s, as compared with C₃ (from 3 to 91 ng/cm²·s). The ratio −D_{3_slope}/f_{(3)_slope} tends to decrease, which suggests that the particles are more rigidly deposited onto the SiO₂ surface at the highest salt concentration. Namely, the calculated ratio −D_{3_slope}/f_{(3)_slope} is closer to the limit of applicability proposed for the Sauerbrey model (Q-Sense) at the highest IS. The ratio −D_{3_slope}/f_{(3)_slope} also reveals that the particles capped with longer alkyl-chain cappings (C₇ and C₁₁) dissipate more energy (higher −D_{3_slope}/f_{(3)_slope}) and are, thus, less rigidly attached to the quartz sensor, as compared with the particles capped with the shortest alkyl chain capping (C₃). Notably, at the highest IS examined, larger error bars in −D_{(3)_slope}/f_{(3)_slope} are observed for C₃ as compared with the C₇ and C₁₁, and this behavior can be linked to the aggregated state of C₃ at this condition (Table 3) (Quevedo et al., 2013).

Interestingly, when comparing the deposition behavior of the Si-NCs using the two different experimental systems (QCM-D and packed columns), different trends in deposition behavior are observed (Fig. 2 and Table 3, respectively). In the packed column (Fig. 2a), the particle capped with C₇ experienced greater retention than the C₃ and C₁₁ (at 30 and 100 mM KCl). This observation can be attributed to the important contribution of physical straining in the retention of the aggregated C₇ Si-NC in the packed granular matrix; this mechanism is absent in the QCM-D, where particles are deposited onto a flat surface (Table 3). Differences in deposition behavior can also be observed in CaCl₂ experiments (Fig. 2b and Table 3). All the Si-NCs reach the theoretical maximum deposition rate at 1.5 mM CaCl₂ in the packed columns (α=1); however, they exhibit much lower deposition rates in the QCM-D when suspended in the same electrolytes. Here again, the observed differences can be attributed to the different controlling mechanisms in each experimental system (i.e., physical straining of larger aggregates in the column vs. reduced convective-diffusive transport and stronger bonding of those same aggregates to the collector surface in QCM-D). This latter concept will be discussed in greater detail further in this article.

QCM-D experiments conducted with particles suspended in CaCl₂ (Table 3) show that the r_d values are generally higher in the presence of the divalent cation, likely due to Ca²⁺ bridging between negatively charged groups on the surface of the Si-NCs and the QCM crystal surface. However, in experiments conducted near the CCC (∼1.5 mM IS CaCl₂), all the Si-NCs exhibited a drastic increase in their size (d_p), and a drastic decrease in deposition rate is observed as compared with experiments conducted at 0.3 mM IS CaCl₂ (Table 3). In experiments conducted in KCl at the maximum IS (100 mM KCl), we also observe a decrease in r_d for particles capped with C₃ and C₇, as compared with experiments conducted at 30 mM KCl (Table 3). In agreement with these results, other QCM-D studies have reported that when particle aggregation is highly favorable, a lessened convective-diffusive transport of aggregated particles to the surface can result in lower measured r_d values (Chen and Elimelech, 2006; Quevedo and Tufenkji, 2009; Quevedo et al., 2013). To better understand this phenomenon, experimental deposition rates (r_d) were normalized by the theoretical particle deposition rate (r_d^SL) based on purely convective-diffusive transport (Adamczyk and van de Ven, 1981) (Fig. 4). At low solution IS (where repulsive electrostatic interactions predominate), r_d is expected to be significantly lower than the theoretical mass transport limited deposition rate (r_d^SL), but as the IS increases, the energy barriers between the Si-NCs and the SiO₂ surface are expected to decline and r_d/r_d^SL should approach unity. However, in experiments conducted with highly aggregated Si-NCs; namely, at 100 mM KCl and at 1.5 mM CaCl₂ (Fig. 4a, b, respectively), we observe that r_d in the absence of energy barriers is considerably lower than the theoretical deposition rates (r_d/r_d^SL<<1).

FIG. 4.

Normalized retention presented as a ratio of the experimental deposition rate (r_d) to the theoretical particle deposition rate (r_d^SL) for QCM-D experiments conducted with Si-NCs suspended in (a) KCl (open symbols) and (b) CaCl₂ (full symbols). r_d was determined from 3rd overtone measurements (Table 3), and r_d^SL was estimated using the Smoluchowski–Levich approximation (Adamczyk and van de Ven, 1981; Quevedo and Tufenkji, 2009) using DLS-measured particle sizes (d_p) (Table 1). Data represent the mean±standard deviation. Dashed lines are included as eye guides. QCM-D, quartz crystal microbalance with dissipation monitoring.

The interpretation of aggregated nanoparticle systems using QCM-D has proved to be challenging (Pomorska et al., 2010; Lin and Wiesner, 2012; Liu et al., 2012), and in some cases, positive frequency shifts have been reported (Fatisson et al., 2009; Pomorska et al., 2010), which are counterintuitive to the principle of QCM-D as a sensitive mass sensor. Positive frequency shifts can be explained in the context of “coupled resonance theory” (Dybwad, 1985). In brief, Δf_(n) can be either negative or positive depending on how well the particle couples to the oscillation of the sensor during the measurement. Firmly coupled particles decrease the crystal resonance frequency as per conventional mass-loading theory (i.e., inertial loading); whereas weakly coupled particles possess a resonance of their own that, depending on the bond stiffness, might increase the crystal resonance frequency (i.e., elastic loading) (Pomorska et al., 2010). The latter case has been shown to occur when micron-sized particles are deposited on a QCM sensor (Berglin et al., 2008; Fatisson et al., 2009; Olsson et al., 2012), and in some instances, a transition from inertial loading (negative frequency shifts) at low overtones to elastic loading (positive frequency shifts) at high overtones can be observed along with a maximum in the dissipation factor (ΔD_n) (Pomorska et al., 2010; Olsson et al., 2012). In Figure 5a and b, we have plotted the frequency and dissipation response, at different overtones (n) for Si-NCs capped with C₁₁ in CaCl₂. The results can be compared with the particle-size distributions and d_p values obtained using DLS (Fig. 5c).

FIG. 5.

(a) Normalized frequency shifts (Δf_(n)) and (b) dissipation shifts (ΔD_n) from experiments conducted with Si-NCs capped with C₁₁ in CaCl₂ at different overtones (n=3, 5, 7, 9, 11, 13). The dotted line in (a) indicates the zero frequency shift. (c) Particle-size distributions for suspended Si-NCs and mean hydrodynamic diameters (d_p) from DLS measurements.

At the lowest IS (0.3 mM CaCl₂), Si-NCs capped with C₁₁ are monodispersed (d_p ∼97 nm), and in terms of frequency shifts, all the overtones (n=3, 5, 7, 9, 11, 13) exhibit “negative” responses, in agreement with inertial (i.e., Sauerbrey) loading (Olsson et al., 2012). As the IS increases to 1.5 mM CaCl₂, the Si-NCs experience aggregation and the suspensions become more polydisperse (d_p=481±224 nm). At this condition, values of both Δf_(n) and ΔD_n decrease in comparison to those measured at lower IS. This can be explained by the fact that for particles approaching sizes resulting in a “coupled resonance” response, the bond strength (i.e., stiffness) influences the magnitude of the negative frequency shift. A stronger bond (e.g., by reduced electrostatic repulsion or Ca²⁺ bridging) reduces the frequency shifts per attached particle (Pomorska et al., 2010; Olsson et al., 2012). At 3 mM CaCl₂, the particle size becomes large enough (d_p=930±671 nm) to result in the typical “coupled resonance” response. At this condition, the deposition of the polydisperse Si-NCs results in mostly positive frequency shifts (Fig. 5a). The higher overtones (n=7, 9, 11, 13) become positive as per “elastic loading,” whereas lower overtones (n=3 and 5) remain negative or close to zero as per “inertial loading.” In addition, we observe an increase in ΔD_n when compared with the lower IS of 1.5 mM (Fig. 5b). As mentioned earlier, this is a clear indication that the traditional “Sauerbrey-type” response can gradually transform into a “coupled resonance”-type response as particle aggregation occurs. Applying this same analysis to all the experiments conducted in our study (Table 3), we note that this transition from inertial to elastic loading is likely linked to the mean size of the particle aggregates, where a critical average size (d_p ∼700 nm) appears to be the limit for this specific system.

Conclusions

Two complementary experimental techniques were used to investigate the deposition behavior of Si-NCs onto silica surfaces and provided useful insights into the role of particle surface cappings (carboxylic acids of different alkyl-chain lengths) and aggregate size on ENP deposition kinetics. Experiments conducted with packed columns show that the mobility of Si-NCs varies as the alkyl-chain length increases. The deposition behavior of particles coated with shorter chain lengths (C₃, C₇) appears to be controlled by physicochemical filtration; namely, retention in the primary energy minimum, and upon aggregation, in a secondary energy well. Physical straining also contributes to the extent of retention of the aggregated C₃ and C₇ Si-NC suspensions in higher IS KCl and in CaCl₂. However, the deposition behavior of Si-NCs capped with the longest alkyl chain (C₁₁) is controlled by electrosteric interactions. QCM-D provides further insights on nanoparticle deposition behavior, by which the ratio of the two output parameters (frequency and dissipation) indicates how rigidly the nanoparticles are bound to the SiO₂ surface. However, based on the obtained results, interpretation of ENP deposition behavior by QCM-D is limited by the presence of large aggregates (d_p ∼700 nm), which give rise to non-Sauerbrey behavior. Future studies of nanoparticle deposition using QCM-D should consider the response of all the overtones to adequately determine whether the acquired frequency shifts are proportional to the deposited mass. The combined approach used here can be considered a useful method for investigating ENP-surface interaction phenomena in an effort to increase our knowledge of the roles that particle and collector surface coatings and water chemistry play in the transport and fate of ENPs in natural and engineered aquatic environments.

Supporting information available

Representative Si-NC breakthrough curves in KCl and CaCl₂, comparative table of the deposition kinetics determined in sand columns and QCM-D, and representative ICP-OES measurements for Si-NCs capped with C₃.

Footnotes

Acknowledgments

The authors acknowledge the financial support of the Natural Sciences and Engineering Research Council of Canada (NSERC), the Canada Research Chairs program, les Fonds de recherche du Québec—Nature et Technologies, Environment Canada, and le Ministère du Développement économique, Innovation et Exportation (MDEIE). The authors also thank Ranjan Roy (McGill) for assistance with ICP-OES measurements.

Author Disclosure Statement

No competing financial interests exist.

References

Adamczyk

, and van de Ven

T.G.M.

(1981). Deposition of particles under external forces in laminar flow through parallel-plate and cylindrical channels. J. Colloid Interface Sci., 80, 340.

Algar

W.R.

, Susumu

, Delehanty

J.B.

, and Medintz

I.L.

(2011). Semiconductor quantum dots in bioanalysis: crossing the valley of death. Anal. Chem., 83, 8826.

Bell

N.C.

, Minelli

, Tompkins

, Stevens

M.M.

, and Shard

A.G.

(2012). Emerging techniques for submicrometer particle sizing applied to stöber silica. Langmuir, 28, 10860.

Berglin

, Olsson

A.L.J.

, and Elwing

(2008). The interaction between model biomaterial coatings and nylon microparticles as measured with a quartz crystal microbalance with dissipation monitoring. Macromol. Biosci., 8, 410.

Bradford

S.A.

, Torkzaban

, and Walker

S.L.

(2007). Coupling of physical and chemical mechanisms of colloid straining in saturated porous media. Water Res., 41, 3012.

Chen

K.L.

, and Elimelech

(2006). Aggregation and deposition kinetics of fullerene (C₆₀) nanoparticles. Langmuir, 22, 10994.

Choi

, Zhang

, Reipa

, Wang

N.S.

, Stratmeyer

M.E.

, Hitchins

V.M.

, and Goering

P.L.

(2009). Comparison of cytotoxic and inflammatory responses of photoluminescent silicon nanoparticles with silicon micron-sized particles in RAW 264.7 macrophages. J. Appl. Toxicol., 29, 52.

Clark

R.J.

, Dang

M.K.M.

, and Veinot

J.G.C.

(2010). Exploration of organic acid chain length on water-soluble silicon quantum dot surfaces. Langmuir, 26, 15657.

Domingos

R.F.

, Baalousha

M.A.

, Ju-Nam

, Reid

M.M.

, Tufenkji

, Lead

J.R.

, Leppard

G.G.

, and Wilkinson

K.J.

(2009). Characterizing manufactured nanoparticles in the environment: multimethod determination of particle sizes. Environ. Sci. Technol., 43, 7277.

10.

Dybwad

G.L.

(1985). A sensitive new method for the determination of adhesive bonding between a particle and a substrate. J. Appl. Phys., 58, 2789.

11.

Fatisson

, Ghoshal

, and Tufenkji

(2010). Deposition of carboxymethylcellulose-coated zero-valent iron nanoparticles onto silica: roles of solution chemistry and organic molecules. Langmuir, 26, 12832.

12.

Elimelech

, Gregory

, Jia

, and Williams

R.A.

(1995). Particle Deposition and Aggregation: Measurement, Modelling and Simulation. Oxford: Butterworth-Heinemann, p. 448.

13.

Fatisson

, Domingos

R.F.

, Wilkinson

K.J.

, and Tufenkji

(2009). Deposition of TiO₂ nanoparticles onto silica measured using a quartz crystal microbalance with dissipation monitoring. Langmuir, 25, 6062.

14.

Fujioka

, Hiruoka

, Sato

, Manabe

, Miyasaka

, Hanada

, Hoshino

, Tilley

R.D.

, Manome

, Hirakuri

, et al. (2008). Luminescent passive-oxidized silicon quantum dots as biological staining labels and their cytotoxicity effects at high concentration. Nanotechnology, 19, 415102.

15.

Gallego-Urrea

J.A.

, Tuoriniemi

, and Hassellöv

(2011). Applications of particle-tracking analysis to the determination of size distributions and concentrations of nanoparticles in environmental, biological and food samples. Trends Anal. Chem., 30, 473.

16.

Hahn

M.W.

, Abadzic

, and O'Melia

C.R.

(2004). Aquasols: on the role of secondary minima. Environ. Sci. Technol., 38, 5915.

17.

Hotze

E.M.

, Phenrat

, and Lowry

G.V.

(2010). Nanoparticle aggregation: challenges to understanding transport and reactivity in the environment. J. Environ. Qual., 39, 1909.

18.

Johannsmann

, Reviakine

, and Richter

R.P.

(2009). Dissipation in films of adsorbed nanospheres studied by quartz crystal microbalance (QCM). Anal. Chem., 81, 8167.

19.

Kanehira

, Banzai

, Ogino

, Shimizu

, Kubota

, and Sonezaki

(2008). Properties of TiO₂–polyacrylic acid dispersions with potential for molecular recognition. Colloids Surf. B, 64, 10.

20.

Klaine

S.J.

, Alvarez

P.J.J.

, Batley

G.E.

, Fernandes

T.F.

, Handy

R.D.

, Lyon

D.Y.

, Mahendra

, McLaughlin

M.J.

, and Lead

J.R.

(2008). Nanomaterials in the environment: behavior, fate, bioavailability, and effects. Environ. Toxicol. Chem., 27, 1825.

21.

Lin

, Cheng

, Liu

, and Wiesner

M.R.

(2012). Polymeric coatings on silver nanoparticles hinder autoaggregation but enhance attachment to uncoated surfaces. Langmuir, 28, 4178.

22.

Lin

, and Wiesner

M.R.

(2012). Deposition of aggregated nanoparticles—a theoretical and experimental study on the effect of aggregation state on the affinity between nanoparticles and a collector surface. Environ. Sci. Technol., 46, 13270.

23.

Liu

, Chen

, and Su

(2012). Influence of collector surface composition and water chemistry on the deposition of cerium dioxide nanoparticles: QCM-D and column experiment approaches. Environ. Sci. Technol., 46, 6681.

24.

Mitzel

M.R.

, and Tufenkji

(2014). Transport of industrial PVP-stabilized silver nanoparticles in saturated quartz sand coated with Pseudomonas aeruginosa PAO1 biofilm of variable age. Environ. Sci. Technol., 48, 2715.

25.

O'Farrell

, Houlton

, and Horrocks

B.R.

(2006). Silicon nanoparticles: applications in cell biology and medicine. Int. J. Nanomed., 1, 451.

26.

Ohshima

(1994). A simple expression for Henry's function for the retardation effect in electrophoresis of spherical colloidal particles. J. Colloid Interface Sci., 168, 269.

27.

Olsson

A.L.J.

, Quevedo

I.R.

, He

, Basnet

, and Tufenkji

(2013). Using the quartz crystal microbalance with dissipation monitoring to evaluate the size of nanoparticles deposited on surfaces. ACS Nano, 7, 7833.

28.

Olsson

A.L.J.

, Van Der Mei

H.C.

, Johannsmann

, Busscher

H.J.

, and Sharma

P.K.

(2012). Probing colloid-substratum contact stiffness by acoustic sensing in a liquid phase. Anal. Chem., 84, 4504.

29.

Park

J.-H.

, Gu

, von Maltzahn

, Ruoslahti

, Bhatia

S.N.

, and Sailor

M.J.

(2009). Biodegradable luminescent porous silicon nanoparticles for in vivo applications. Nat. Mater., 8, 331.

30.

Pelley

A.J.

, and Tufenkji

(2008). Effect of particle size and natural organic matter on the migration of nano- and microscale latex particles in saturated porous media. J. Colloid Interface Sci., 321, 74.

31.

Petosa

A.R.

, Brennan

S.J.

, Rajput

, and Tufenkji

(2012). Transport of two metal oxide nanoparticles in saturated granular porous media: role of water chemistry and particle coating. Water Res., 46, 1273.

32.

Petosa

A.R.

, Jaisi

D.P.

, Quevedo

I.R.

, Elimelech

, and Tufenkji

(2010). Aggregation and deposition of engineered nanomaterials in aquatic environments: role of physicochemical interactions. Environ. Sci. Technol., 44, 6532.

33.

Phenrat

, Song

J.E.

, Cisneros

C.M.

, Schoenfelder

D.P.

, Tilton

R.D.

, and Lowry

G.V.

(2010). Estimating attachment of nano- and submicrometer-particles coated with organic macromolecules in porous media: development of an empirical model. Environ. Sci. Technol., 44, 4531.

34.

Pomorska

, Shchukin

, Hammond

, Cooper

M.A.

, Grundmeier

, and Johannsmann

(2010). Positive frequency shifts observed upon adsorbing micron-sized solid objects to a quartz crystal microbalance from the liquid phase. Anal. Chem., 82, 2237.

35.

Quevedo

I.R.

, Olsson

A.L.J.

, and Tufenkji

(2013). Deposition kinetics of quantum dots and polystyrene latex nanoparticles onto alumina: role of water chemistry and particle coating. Environ. Sci. Technol., 47, 2212.

36.

Quevedo

I.R.

, and Tufenkji

(2009). Influence of solution chemistry on the deposition and detachment kinetics of a CdTe quantum dot examined using a quartz crystal microbalance. Environ. Sci. Technol., 43, 3176.

37.

Quevedo

I.R.

, and Tufenkji

(2012). Mobility of functionalized quantum dots and a model polystyrene nanoparticle in saturated quartz sand and loamy sand. Environ. Sci. Technol., 46, 4449.

38.

Raychoudhury

, Tufenkji

, and Ghoshal

(2012). Aggregation and deposition kinetics of carboxymethyl cellulose-modified zero-valent iron nanoparticles in porous media. Water Res., 46, 1735.

39.

Roco

, Mirkin

, and Hersam

(2011). Nanotechnology research directions for societal needs in 2020: summary of international study. J. Nanopart. Res., 13, 897.

40.

Ruizendaal

, Bhattacharjee

, Pournazari

, Rosso-Vasic

, De Haan

L.H.J.

, Alink

G.M.

, Marcelis

A.T.M.

, and Zuilhof

(2009). Synthesis and cytotoxicity of silicon nanoparticles with covalently attached organic monolayers. Nanotoxicology., 3, 339.

41.

Sakulchaicharoen

, O'Carroll

D.M.

, and Herrera

J.E.

(2010). Enhanced stability and dechlorination activity of pre-synthesis stabilized nanoscale FePd particles. J. Contam. Hydrol., 118, 117.

42.

Saleh

, Kim

H.-J.

, Phenrat

, Matyjaszewski

, Tilton

R.D.

, and Lowry

G.V.

(2008). Ionic strength and composition affect the mobility of surface-modified Fe⁰ nanoparticles in water-saturated sand columns. Environ. Sci. Technol., 42, 3349.

43.

Sauerbrey

G.Z.

(1959). Verwendung von. Schwingquarzen zur wiigung dunner. Schichten und zur mikrowägung. Z. Phys., 155, 206.

44.

Tufenkji

, and Elimelech

(2004). Correlation equation for predicting single-collector efficiency in physicochemical filtration in saturated porous media. Environ. Sci. Technol., 38, 529.

45.

Tufenkji

, Miller

G.F.

, Ryan

J.N.

, Harvey

R.W.

, and Elimelech

(2004). Transport of Cryptosporidium oocysts in porous media: role of straining and physicochemical filtration. Environ. Sci. Technol., 38, 5932.

46.

Vogel

, Willmott

, Kozak

, Roberts

G.S.

, Anderson

, Groenewegen

, Glossop

, Barnett

, Turner

, and Trau

(2011). Quantitative sizing of nano/microparticles with a tunable elastomeric pore sensor. Anal. Chem., 83, 3499.

47.

Westerhoff

P.K.

, Kiser

, and Hristovski

(2013). Nanomaterial removal and transformation during biological wastewater treatment. Environ. Eng. Sci., 30, 109.

48.

Yao

K.M.

, Habibian

M.T.

, and O'Melia

C.R.

(1971). Water and waste water filtration: concepts and applications. Environ. Sci. Technol., 5, 1105.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.37 MB

0.10 MB

0.05 MB

0.06 MB

Interpreting Deposition Behavior of Polydisperse Surface-Modified Nanoparticles Using QCM-D and Sand-Packed Columns

Abstract

Abstract

Introduction

Experimental Protocols

Preparation of Si-NC suspensions

Electrokinetic characterization of Si-NCs in electrolytes

Sizing of Si-NCs in electrolytes

Transport experiments in packed-bed columns

Interpretation of transport experiments in packed-bed columns

Deposition of Si-NCs onto SiO2 measured by QCM-D

Interpretation of QCM-D deposition experiments

Results and Discussion

Characterization of capped Si-NCs in electrolyte

Transport of capped Si-NCs in saturated sand-packed columns

Deposition of Si-NCs in QCM-D

Conclusions

Supporting information available

Footnotes

Acknowledgments

Author Disclosure Statement

References

Supplementary Material

Deposition of Si-NCs onto SiO₂ measured by QCM-D