Colloidal Stability of Magnetic Iron Oxide Nanoparticles: Influence of

Jun 8, 2011 - Two different molecular weight PAAs were also used for coating. ... data is made available by participants in Crossref's Cited-by Linkin...
0 downloads 0 Views 3MB Size
ARTICLE pubs.acs.org/Langmuir

Colloidal Stability of Magnetic Iron Oxide Nanoparticles: Influence of Natural Organic Matter and Synthetic Polyelectrolytes Saikat Ghosh,† Wei Jiang,† Julian D. McClements,‡ and Baoshan Xing*,† †

Department of Plant, Soil and Insect Sciences and ‡Biopolymer and Colloids Research Laboratory, Department of Food Science, University of Massachusetts, Amherst, Massachusetts 01003, United States ABSTRACT: The colloidal behavior of natural organic matter (NOM) and synthetic poly(acrylic acid) (PAA)-coated ferrimagnetic (γFe2O3) nanoparticles (NPs) was investigated. Humic acid (HA), an important component of NOM, was extracted from a peat soil. Two different molecular weight PAAs were also used for coating. The colloidal stability of the coated magnetic NPs was evaluated as a resultant of the attractive magnetic dipolar and van der Waals forces and the repulsive electrostatic and steric electrosteric interactions. The conformational alterations of the polyelectrolytes adsorbed on magnetic γFe2O3 NPs and their role in colloidal stability were determined. Pure γFe2O3 NPs were extremely unstable because of aggregation in aqueous solution, but a significant stability enhancement was observed after coating with polyelectrolytes. The steric stabilization factor induced by the polyelectrolyte coating strongly dictated the colloidal stability. The pH-induced conformational change of the adsorbed, weakly charged polyelectrolytes had a significant effect on the colloidal stability. Atomic force microscopy (AFM) revealed the stretched conformation of the HA molecular chains adsorbed on the γFe2O3 NP surface at pH 9, which enhanced the colloidal stability through long-range electrosteric stabilization. The depletion of the polyelectrolyte during the dilution of the NP suspension decreased the colloidal stability under acidic solution conditions. The conformation of the polyelectrolytes adsorbed on the NP surface was altered as a function of the substrate surface charge as viewed from AFM imaging. The polyelectrolyte coating also led to a reduction in magnetic moments and decreased the coercivity of the coated γFe2O3 NPs. Thus, the enhanced stabilization of the coated maghematite NPs may facilitate their delivery in the groundwater for the effective removal of contaminants.

’ INTRODUCTION The rapid growth of nanotechnology during the past decade drives its widespread application and the development of new materials for advanced use. Metal oxide nanoparticles (NPs) are one of the major groups of engineered NP with numerous industrial applications. Among the metal oxide NPs, magnetic iron oxide NPs have drawn considerable attention from researchers across the scientific disciplines. The progress in materials research has shown that these magnetic iron oxide NPs have possible utilization in magnetic data storage, cellular therapy, and magnetic resonance imaging (MRI).1,2 Significant research effort has been expended by the environmental scientists in using zero-valent iron and its oxide NPs in groundwater remediation.3,4 These investigations highlighted that the efficacy of these NPs relied heavily on their colloidal stability in a particular environment. The colloidal stability of maghematite (γFe2O3) NPs in an aquatic medium depends not only on the environmental parameters but also on their intrinsic magnetic properties. The longrange magnetic dipolar attraction operative in these NPs favors strong aggregation in contrast to that of their nonmagnetic counterparts. The surface modification of magnetic Fe and its oxide NPs may enhance their colloidal stability as well as bioavailability and toxicity. Maghematite and zero-valent iron NPs (nZVI) can be effectively used for the removal of As(III) r 2011 American Chemical Society

from water and in degrading the toxic non-aqueous-phase liquids.5,4 Saleh et al. reported that grafting triblock copolymers to nZVI significantly improved their delivery to oil/water interfaces.4 However, in spite of their enormous possible applications, magnetic Fe2O3 NPs also caused toxicity to nerve cells.6 Moreover, different surface modifiers on magnetic NPs interacted differently with cells.7 These findings clearly asserted the role played by the surface modifiers as well as inherent properties emerging from the magnetic core to achieve the targeted result. Aquatic and soil environments are enriched with natural organic matter (NOM) with varied structures and compositions. The magnetic NPs exposed to a natural aquatic environment may produce a coreshell structure because of NOM sorption on the NP surface. Environmental parameters such as pH, ionic strength, and NOM concentration may considerably influence the properties of the NOM shell around the magnetic core. The stabilization of these surface-modified magnetic γFe2O3 NPs would depend upon the delicate balance of magnetic dipolar, steric, and DLVO intractions.8,9 In addition to the inherent nature of the core NPs, tailoring of the polyelectrolyte shell around the core NPs can change their colloidal behavior. The Received: February 28, 2011 Revised: May 24, 2011 Published: June 08, 2011 8036

dx.doi.org/10.1021/la200772e | Langmuir 2011, 27, 8036–8043

Langmuir development of a spherical polyelectrolyte brush by grafting weakly or strongly charged polyelectrolytes around a metallic NP core has strengthened their potential application.10 The majority of previous research has focused on assessing the role of different polymeric or biopolymeric materials in the suspension stability of nZVI. However, when exposed to an aquatic environment or under high-humidity conditions, the surface layer of nZVI is oxidized to form an oxide layer. Therefore, the surface oxide layers on nZVI contribute substantially in their interactions with other environmental components (e.g., contaminants and NOM). Therefore, the objectives of our study were to assess the impact of a natural polyelectrolyte, humic acid (HA), and synthetic PAA coatings (differing in chain length) on the colloidal stability of γFe2O3 NPs. Moreover, the role of polyelectrolyte sorption in the magnetic properties of these NPs and the pH-induced conformational change of the adsorbed polyelectrolyte with respect to suspension stability have not been addressed extensively. The aggregation behavior of the PAA- and HA-coated γFe2O3 NPs was therefore determined in the presence of monovalent and divalent cations (Naþ and Ca2þ). The conformation of the polyelectrolyte layer adsorbed on the NPs surface was also examined using atomic force microscopy (AFM) as a function of pH and substrates with varying surface charge. These conformational alterations may simulate their actual behavior under natural conditions and demonstrate their potential influence on overall stabilization. Furthermore, the magnetic properties of the pure and polyelectrolyte-coated NPs were evaluated in order to understand the effect of the magnetic force on the colloidal stability of γFe2O3 NPs.

’ MATERIALS AND METHODS Materials. The humic acid used for this experiment was obtained after seven sequential extractions of Amherst peat soil (MA, USA). The detailed extraction procedure was described previously.11 Synthetic sodium salts of poly(acrylic acid) (Na-PAA) solutions were purchased from Polyscience Inc. (PA) with molecular weights of 1800 and 50 000 g/mol, and they are abbreviated as PAA2K and PAA50K henceforth. γFe2O3 NPs were purchased from Nanoamor (Houston, TX) with a diameter range of 2550 nm. Complex Preparation. HA was dissolved in a minimum volume of 0.5 mol/L NaOH to obtain a 500 mg/L solution. Na-PAA solutions were diluted with deionized water to obtain the same concentration as that of HA. NaCl was added to the solutions to keep the background electrolyte concentration at 0.1 mol/L, followed by the adjustment of pH to 4.0. γFe2O3 NPs (0.5 g) were added to 100 mL of each solution, followed by sonication in a sonicator bath for 1 h. Then the mixtures were allowed to shake for 3 days. The mixture was then separated first by a magnet and subsequently centrifuged at 12 000g. The same coating procedure was repeated to ensure a thorough coating of the γFe2O3 NP surface. The complexes were then washed repeatedly with deionized water and redispersed in 100 mL of deionized water, followed by adjustment of the pH to 5. The coating of γFe2O3 NPs with HA and PAAs led to the formation of stable suspensions.

Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFT). A Perkin-Elmer Spectrum One Fourier transform infrared spectrometer (FTIR) was used to obtain DRIFT spectra of the HA- and PAA-coated and pure γFe2O3 NPs. The sample preparation method for the DRIFT analyses was described in our previous publications.12,13 The DRIFT spectra of the coated γFe2O3 NPs were subtracted from those of the pure γFe2O3 NPs to obtain the differential spectra of coated polyelectrolytes using Spectrum software (PerkinElmer).

ARTICLE

Determination of Zeta Potential (ξ) and Hydrodynamic Diameter (DH). The zeta potential (ξ) and hydrodynamic diameter

(DH) of the coated γFe2O3 NPs were measured with a Malvern NanoZS (Malvern, Wores, U.K.) instrument. The ξ of the coated γFe2O3 NPs was determined as a function of equilibrium pH using a universal zeta dip cell. A 50 mg/L suspension concentration was used for the ξ measurements. The effect of particle concentration on the size distribution of HAand PAAs-coated γFe2O3 NPs was determined at pH 5. The early-stage aggregation of HA- and PAA50K-coated NPs was investigated in the presence of Naþ or Ca2þ at pH 5 and 9. A suspension concentration of 20 mg/L was maintained in all aggregation experiments by diluting the original concentrated suspension with deionized water. The suspension pH was then adjusted with a minimum volume of 0.1 M HCl and 0.1 M NaOH solutions to pH 5 or 9. The diameter (DH) of the aggregates formed under the influence of Ca2þ or Naþ was plotted against time (seconds). Atomic Force Microscopy. AFM imaging of the HA- and PAAcoated γFe2O3 NPs was carried out on corundum (Al2O3) (100 orientation), freshly cleaved mica, and silicon wafer substrates. The influence of the substrate surface charge on the conformational behavior of the polyelectrolytes adsorbed on the NP surface was evaluated. Corundum provides a positively charged surface, and mica surface has a net negative potential.14 Polyelectrolyte-coated γFe2O3 NPs were negatively charged at all experimental pH values. The substrates were then immersed in diluted suspensions of the coated NPs to facilitate sorption and washed subsequently by dipping in deionized water. The samples were then dried in a desiccator filled with silica gel. The AFM sample that was used to study the effect of the pH-induced conformational change of the HA coating on γFe2O3 NPs at pH 9 was prepared by placing 80 μL of the suspension on a silicon wafer substrate, and sorption was allowed for 20 min. The excess liquid was removed from the substrate under N2 flow, followed by drying.1517

DC-Superconducting Quantum Interference Device (DCSQUID) Magnetic Measurements. Saturation magnetizations of the pure and polyelectrolyte-coated NPs were measured with a DCSQUID magnetometer (Quantum Design Inc., model MPMS, USA). Weighed amounts of powdered samples were poured into gelatin capsules for the measurements.18 The measurements were carried out at 300 K under an applied magnetic field of (3 T. The saturation magnetization (Ms), hysteresis (H), and coercivity (HC) of the hysteresis loop were calculated from the magnetization data.

’ RESULT AND DISCUSSION DRIFT Spectra. The functional moieties of HA and PAA molecules adsorbed on the NP surface were determined after subtracting the pure mineral spectrum from the complex spectra (Figure 1A). The sorption of HA molecules on the NP surface can be identified by the presence of strong absorption bands at 2928 and 2848 cm1, reflecting the aliphatic functional moieties on the NP surface (Figure 1A).12,13 Previous studies have shown that the aliphatic nature of the sequentially extracted HAs increases with subsequent extractions.12,13,19 In addition to aliphatic CH absorption bands at 2928 and 2848 cm1 in PAA-coated NPs, a strong peak was observed at 1450 cm1 that was probably due to the scissoring vibration of the hydrocarbon chains of PAA.20,21 Pure HA molecules demonstrated a shoulder at 1704 cm1 and a peak at 1586 cm1 due to carboxylic CdO stretching vibrations as presented in our previous work.13,22 However, it was evident from the adsorbed HA spectrum that the shoulder at 1704 cm1 was shifted to 1683 cm1, reflecting the sorption of 8037

dx.doi.org/10.1021/la200772e |Langmuir 2011, 27, 8036–8043

Langmuir

ARTICLE

Figure 1. (A) DRIFT spectra of HA and PAA adsorbed on the γFe2O3 NP surface. The differential spectra were obtained from the subtraction of the pure mineral spectrum from that of the complexes. (B) The ξ values of PAA2K, PAA50K, and HA-coated γFe2O3 NPs were plotted as a function of the equilibrium pH of the suspension. ξ tends to decrease with increasing pH, reflecting the higher ionization of the adsorbed polyelectrolytes with increasing pH.

Figure 2. (A) Influence of suspension concentration on the size distribution (DH). The enhanced magnetic dipolar attraction with increased particle concentration and weaker steric stabilization in short-chain PAA2K-coated NPs may cause a relatively higher increment in DH. (B) The AFM phase image of the PAA2K-coated NPs at a 100 mg/L suspension concentration showed the formation of large clusters. (C) The AFM height image of HA-coated γFe2O3NPs on a positively charged corundum surface revealed the formation of an HA brush layer on the NP surface. (D) The AFM phase image of HA-coated γFe2O3NPs on a mica surface shows that HA molecules coated a single NP as well as nanoclusters produced by strong magnetic dipoledipole attraction.

carboxylate moieties. The sorption of carboxylates on the γFe2O3 NP surface can be further distinguished from the shifting of the 1586 cm1 band in pure HA to the 1562 cm1 band in adsorbed HA. Adsorbed HA displayed a band at 1654 cm1, revealing the sorption of CdO on the NP surface. The 1401 cm1 absorption band in adsorbed HA suggested the complexation of HA COO with Fe. Similar behavior of HA COO moieties was also previously observed on a goethite surface.22 Adsorbed PAA showed a strong band of unionized COOH at 1710 cm1 due to the partial dissociation of carboxylic moieties in PAA chains at pH 4. However, the sorption of ionized COO on the NP surface can be distinguished from the symmetric stretching vibration at 1401 cm1 whereas the asymmetric stretching vibration of COO was observed in the 1523 cm1 region. Kirwan et al. detected that the PAA absorption band intensity of unionized COOH was stronger than the asymmetric COO band on the hematite surface at pH 2.20 However, increased ionization of PAA molecules was observed with increasing pH from 3 to 9.23 The role of inner-sphere complexation between COO moieties in Na-PAA with the surface Al3þ of δ-Al2O3 particles upon sorption was addressed previously.21 Therefore, the sorption of PAA and HA molecules on the oppositely charged γFe2O3 NP surface was due to a combination of electrostatic and/or ligand-exchange reactions. Influence of Polyelectrolyte Coating on the Surface Charge. The pH-induced surface charge alteration of polyelectrolyte-coated γFe2O3 NPs is presented in Figure 1B. γFe2O3 NPs have positive surface potentials in the acidic pH range (isoelectric point ≈ pH 7.5).24 However, the polyelectrolyte

coating led to a reversal of surface charge because of the sorption of polar functional moieties on the NP surface, as shown from the DRIFT results. HA- and PAA- coated γFe2O3 NPs had similar ξ potentials at pH 5. However, the increment in pH from 5 to 6.5 showed a further lowering of ξ as a result of the enhanced ionization of the adsorbed polyelectrolytes. Numerous studies have demonstrated that the degree of ionization of the weakly charged polyelectrolytes increases with increasing pH. It is evident from Figure 1B that PAA-coated NPs had relatively lower ξ values compared to those of HA-coated NPs at pH 6.5. Thus, adsorbed PAA molecules possibly had a relatively higher polarity than the adsorbed HA fractions. Effect of Particle Concentration on Size Distribution. The effect of particle concentration on the average hydrodynamic diameter (DH) of PAA2K, PAA50K, and HA-coated γFe2O3 NPs was examined at pH 5 (Figure 2A). The alteration of DH by particle concentration is an estimation of the counteracting forces evolved because of DLVO and extended DLVO (steric and magnetic) interactions. Among the interacting forces, magnetic dipolar forces were strongly correlated to the particle concentrations in suspension. Saturation magnetization measurements of the pure and polyelectrolyte-coated NPs normalized by the amount of sample also demonstrated significant residual magnetization in spite of the polyelectrolyte coating (discussed below). The ferrofluids formed by coating γFe2O3 NPs with polyelectrolytes significantly enhanced the colloidal stability in contrast to that of pure γFe2O3 NPs. Short-chain, PAA2K-coated γFe2O3 NPs showed the lowest average DH compared to those of PAA50K- and HA-coated γFe2O3 NPs 8038

dx.doi.org/10.1021/la200772e |Langmuir 2011, 27, 8036–8043

Langmuir at a 20 mg/L suspension concentration. However, at an increased suspension concentration to 50 mg/L, PAA2K-coated NPs demonstrated a higher relative increment in the average DH than did PAA50K- and HA-coated NPs. For PAA2K-coated NPs, the magnetic contribution to aggregation possibly increased significantly at high particle concentration because of the additive nature of the magnetic moment. Charge stabilization and weaker steric stabilization may not sufficiently counteract the strong magnetic influence that arises upon increasing the particle concentration. Moreover, the AFM phase image of the PAA2K-coated NPs at a 100 mg/L suspension concentration on a mica surface confirmed the formation of large nanoclusters, in agreement with our DLS data (Figure 2B). The weak steric stabilization and relatively high residual magnetization of PAA2K-coated NPs possibly align magnetic dipoles of neighboring NPs to form chain or ringlike structures. One study indicated that magnetic NPs orient their magnetic dipoles to facilitate overall attractions whereas electrostatic interactions are strongly repulsive.8 Furthermore, interparticle magnetic dipolar attractions are the predominant interaction energy among the thinorganic-layer-coated magnetic NPs.25 On the contrary, a relatively minute effect on average DH values was observed at increased suspension concentrations for PAA50K- and HAcoated NPs. Steric components of stabilization by adsorbed PAA50K and HA molecules were likely to be the preventing factor against magnetic destabilization. PAA50K and HA coatings produced much thicker adsorbed layers on the NP surfaces than did short-chain PAA2K molecules, resulting in a stronger steric stabilization in the former.26 Moreover, interparticle magnetic dipolar attractions are stronger as the distance between the centers of the two approaching particles decreases. Highmolecular-weight carboxy methyl cellulose molecules were found to enhance the stability of nZVI by generating stronger electrostatic and steric contributions.27 The AFM images of PAA50K-coated NPs confirmed the presence of a diffuse polyelectrolyte layer around the magnetic NP core (shown below). The coating of HA molecules on the γFe2O3 NP surface clearly revealed the formation of a hairy layer, where individual chains of the HA molecules were anchored to the NP surface (Figure 2C). Moreover, the anisotropy of the HAcoated γFe2O3 NPs can be clearly distinguished from the particle geometry. The AFM phase image further confirmed that during the coating process HA molecules coated not only a single γFe2O3 NPs but also small agglomerates (two to three NPs) produced from the strong magnetic dipolar attraction (Figure 2D). Therefore, it seems that in a few cases the magnetic-dipoleinduced particleparticle collisions were faster than the time required to envelope a single NP with HA, especially in the absence of strong sonication during complex preparation. Aggregation of Polyelectrolyte-Coated γ-Fe2O3 NPs in the Presence of Naþ and Ca2þ. The role of monovalent (Naþ) and divalent (Ca2þ) cations in the aggregation of the coated NPs was evaluated. At very low Naþ and Ca2þ concentrations, the average DH decreased for HA-coated NPs compared to those of the samples containing no salts (Figure 3A,B). The diminution in the average size of the coated NPs with a minute increase in ionic strength occurred and was possibly attributable to the shrinkage of the Debye length (k1).28 Al2O3 NPs coated with HA showed similar behavior at lower concentrations of Ca2þ in addition to their reduction in |ξ|.13 This is in accordance with the subsequent augmentation in the ζ potential of HA- and PAA50K-coated NPs suspensions in the presence of Ca2þ and

ARTICLE

Figure 3. (A)Aggregation behavior of HA-coated γFe2O3 NPs at pH 5 in the presence of Naþ. The average DH decreased at low Naþ concentration, reflecting the probable change in the conformation of the adsorbed HA layer. (B) Aggregation of HA-coated NPs at pH 5 in the presence of Ca2þ. The expected critical coagulation concentration (CCC) value is much lower for Ca2þ than for Naþ. (C) Timedependent aggregation of PAA50K-coated NPs in the presence of Ca2þ. The rate of aggregation is faster in PAA50K-coated NPs than in HA-coated NPs.

Naþ at pH 5. However, Ca2þ-induced charge screening was much stronger than Naþ-induced charge screening because of its enhanced binding efficiency to charged polyelectrolyte molecules (data not shown). The aggregation of the PAA50K-coated NPs at pH 5 was examined in the presence of Ca2þ (Figure 3C). The aggregation data indicated a larger aggregate formation of PAA50K-coated NPs in contrast to that of HA-coated NPs. The difference in aggregation between PAA50K- and HA-coated γFe2O3 NPs can be estimated by comparing the DH values of the coated NPs with 20 μM Ca2þ. Under this condition, the PAA50K-coated NPs showed a reaction-controlled aggregation where the aggregation rate increases with increasing Ca2þ concentration. However, HA-coated NPs did not show any aggregation under similar conditions. This is in line with the increment of ζ in PAA50Kcoated NPs compared to that of HA-coated NPs in the presence of 20 μM Ca2þ (data not shown). Therefore, the weaker charge neutralization of adsorbed HA compared to that of PAA50K and the relatively stronger steric stabilization in the former are consistent with their different aggregation behaviors. The instability of the coated NPs at lower pH was typical of a collapsed polymer brush adsorbed or grafted on a particle surface.29 In the case of weakly charged polyelectrolytes (e.g., PAA and HA), the 8039

dx.doi.org/10.1021/la200772e |Langmuir 2011, 27, 8036–8043

Langmuir

ARTICLE

Figure 4. (A) AFM height and (B) phase images of PAA50K-coated γFe2O3 NPs at pH 5 on a mica surface (preadsorbed in 50 mM Naþ) showing the presence of molecular chains of PAA entrapping the coated NPs. (C) AFM images further showed that PAA50K-coated γFe2O3 NPs formed open clusters at very high Naþ concentration (within the diffusion-controlled aggregation regime).

degree of pH-induced ionization determined their relative contribution to steric-elastic and electrosteric stabilization. Screening the electrostatic repulsion with the addition of Ca2þ and Naþ bolstered the importance of attractive van der Waals and magnetic dipolar attractions. This suggested that steric-elastic stabilization was inadequate to counter the forces of aggregation, leading to overall destabilization of the NPs suspension. The higher aggregation rate of the coated NPs at relatively lower concentrations of Naþ and Ca2þ may be associated with the depletion of the adsorbed polyelectrolyte when diluting the suspension with deionized water. AFM height and phase images of the PAA50K-coated NPs showed that the molecular chains of PAA desorbed from the NP surface entrapped the coated NPs, leading to aggregation (Figure 4A,B). AFM image of the PAA50K-coated NPs on the mica surface presoaked in 250 mM Naþ showed the formation of large clusters with an open structure (Figure 4C). This concentration of Naþ was well within the diffusion-controlled aggregation regime of the PAA50K-coated γFe2O3 NPs. The formation of open clusters possibly resulted from the enhanced van der Waals attraction between the coated particles because of the polyelectrolyte depletion. The large open-cluster formation in the diffusion-controlled aggregation was also observed in colloidal gold NPs in the presence of pyridine.30 The colloidal stability of 10 nm core-sized alkanoic acidcapped Fe3O4 NPs was primarily attributed to the balance of van der Waals and electrostatic forces.31 However, in our case, the average particle size (2550 nm) was much larger than the 10 nm core. Thus, a substantial magnetic moment must be associated with the NPs. Several researchers have shown that the saturation magnetization of these magnetic NPs decreased significantly with decreasing particle size.32,33 It was pointed out that below the transition size of 12.5-nm-diameter magnetic NPs, they lose a considerable magnetic moment.32 Hence, the ferromagnetic properties of the NPs have been reduced to superparamagnetic. The surface spin disorder in smaller NPs could reduce the overall magnetization because the relative contribution of surface atoms is larger than the core atoms in smaller NPs.34 Therefore, reduced electrostatic repulsion with the addition of Ca2þ or Naþ may enhance the relative importance of attractive van der Waals and magnetic dipolar forces between the particles and caused aggregation. Our SQUID results (discussed later) demonstrated a strong magnetic moment and hysteresis of the magnetic NPs even after being coating with polyelectrolytes. The aggregation of HA-coated γFe2O3 NPs in the presence of Ca2þ at pH 9 is shown in Figure 5A. It can be clearly seen that the

Figure 5. (A)Aggregation behavior of HA-coated γFe2O3 NPs at pH 9 in the presence of Ca2þ. The average aggregate size was much lower than that observed at pH 5. (B) AFM phase image of HA-coated γFe2O3 NPs on a silicon wafer substrate at pH 9. HA molecular chains adsorbed on the core NP were stretched significantly because of the strong intermolecular and intramolecular repulsions. The stretched conformation of the adsorbed chains imparts strong electrosteric stabilization against attractive van der Waals and magnetic dipolar forces.

average aggregate size at pH 9 was smaller than that at pH 5. Previously, we observed that critical coagulation concentrations (CCCs) of the weakly polar HA-coated Al2O3 NPs increased significantly with increasing pH.13 This can be further corroborated from the decrease in the ξ potential with the increase in pH. Therefore, increased ionization of the adsorbed HA segments with a consequent rise in pH facilitated long-range electrosteric stabilization. At lower pH, the PAA chains grafted onto the polystyrene (PS) core NPs were reported to be weakly charged and only minutely elongated because of the interactions between 8040

dx.doi.org/10.1021/la200772e |Langmuir 2011, 27, 8036–8043

Langmuir

ARTICLE

Figure 6. Conformational change of the PAA chains adsorbed on the NP surface can be assessed from the AFM height and phase images. (A, B) PAA50K-coated γFe2O3 NPs at pH 5 on the mica surface showed the stretching of the diffused PAA layer from the NP core. (C, D) However, the AFM height and phase images of PAA50K-coated γFe2O3 NPs at pH 5 on the corundum surface showed that the adsorbed PAA chains shrank because of the positive surface potential of the substrate under the experimental condition.

the adjacent chains.35 At pH 9, a kinetically driven nanocluster formation was identified immediately after the addition of low concentrations of Ca2þ (increase in average DH) because of the partial collapse of the adsorbed HA chains.35,28 The AFM phase image of HA-coated γFe2O3 NPs at pH 9 on a silicon wafer substrate elicited the pH-induced stretched conformation of the adsorbed HA (Figure 5B). It was evident from the section analysis that the adsorbed HA molecular chains formed a stretched spherical polyelectrolyte brushlike structure with a magnetic NP core. Increased segmental repulsions among the adsorbed HA chains led to overall swelling.36 Single-molecule AFM imaging identified the expanded conformation of weakly charged polyelectrolyte at a relatively high ionic strength solution of NaCl.37 Moreover, the increased ionization of the silanol groups on the wafer substrate may favor the elongated conformation of the HA chains. The thickness of the adsorbed chains may also be affected by the surface roughness of the wafer substrate because the silicon wafer has a higher surface roughness than does the mica substrate.38 The core NP is seen as a bright center because of the minimal adhesion between the AFM tip and the hard NP core.39 However, a relatively lower phase shift was observed while the AFM tip interacted with the softer HA chains (a comparatively higher adhesive force). In our previous work, we also observed the swelling behavior of the pure HA molecules in solution with increases in pH.40 Therefore, at high pH the weakening of the magnetic and van der Waals attractions and increased electrosteric repulsions augmented the stabilization. The HA used in our study was composed of long-chain polysaccharide-like materials that are ubiquitous in natural aquatic environments, which manifested stronger electrosteric

Figure 7. (A, B) AFM height and phase images of HA-coated γFe2O3 NPs on a mica surface at pH 5. (B) The phase image in the inset showed that the core NPs were surrounded by the spongy HA molecules extended onto the mica surface. (C, D) Significant shrinkage of the HA layer can be seen when imaging is carried out on the corundum surface. (E, F) HA-coated NPs on the corundum surface further demonstrated stronger aggregation and lower surface coverage, when sorption was allowed for 90 min.

stabilization. Short-chain, highly polar fulvic acid (FA) fractions of NOM are also abundant in aquatic bodies. Therefore, the steric stability factor would be inadequate if oppositely charged γFe2O3 NPs were exposed to an aquatic environment enriched with short-chain, highly polar NOM fractions at a relatively high ionic strength. Effect of Substrates on AFM Imaging. The conformation of the polyelectrolytes adsorbed on the NP surface was investigated as a function of the substrate surface charge. The polyelectrolyte molecular chains will come into direct contact with the substrate during their sorption. AFM imaging of both PAA50K- and HAcoated γFe2O3 NPs was carried out on freshly cleaved mica and corundum (Al2O3) substrates. Corundum has a pH-dependent surface charge whereas mica has a net negative potential under the experimental condition of pH 5.14,15 A comparison between PAA50K-coated γFe2O3 NPs on mica and corundum surfaces revealed that the adsorbed PAA layer had a more stretched conformation on the mica surface (Figure 6). The weaker attraction between the adsorbed PAA layer and mica favored 8041

dx.doi.org/10.1021/la200772e |Langmuir 2011, 27, 8036–8043

Langmuir

Figure 8. (A) Saturation magnetization (Ms) of the pure and PAA50Kand HA-coated NPs. (A) The inset shows the hysteresis referring to the strong ferromagnetic nature of the materials. (B) Coercivity of the materials under a zero-field state. (C) The coercive energy of the PAA50K- and HA-coated NPs decreased compared to that of the pure NPs.

an elongated conformation. A similar conformation of a spherical polyelectrolyte brush was observed when PAA was grafted on the PS core NPs.41 Conversely, the AFM images of PAA50K-coated NPs (Figure 6C,D) on the positively charged corundum surface show a constriction of the adsorbed PAA chains. Strong electrostatic attraction between the carboxylates of PAA and the positively charged corundum substrate likely subdued the diffusivity of the adsorbed PAA. The adsorbed HA chains on γFe2O3 NPs surfaces also displayed similar behavior to adsorbed PAA chains. A comparison of the AFM images of the HA-coated NPs on mica (Figure 7A,B) and corundum (Figure 7C,D) surfaces showed that the adsorbed HA chains were more extended on the mica surface but were significantly quenched on the corundum surface. HA-coated γFe2O3 NPs on the corundum surface further exhibited strong aggregation with increasing sorption time to 90 min (Figure 7E,F). The average aggregate height had increased up to 65 nm. This is in accordance with the hypothesis that the charge inside the adsorbed HA brush layers was screened greatly by the corundum substrate, leading to aggregation.35 Magnetic Properties of Pure and HA- and PAA-Coated γFe2O3 NPs. Saturation magnetization (Ms) of the pure and coated γFe2O3 NPs was studied with a DC-SQUID magnetometer

ARTICLE

(Figure 8A). The obtained Ms values were normalized by the weight of the samples used for the SQUID measurements. Pure γFe2O3 NPs had a saturation magnetization of 70 emu/g, reflecting strong ferrimagnetic behavior. This value was much lower than the crystalline body-centered-cubic (bcc) iron.32 The decreasing Ms value of the γFe2O3 NPs was well correlated with the smaller core size of this NPs compared to that of bulk iron. However, this Ms value is much higher than the small magnetic iron NPs, which are superparamagnetic. The comparison between the Ms values for the pure and HA-coated γFe2O3 NPs showed that the coating of HA molecules decreased the Ms. The formation of a magnetically dead layer (demagnetization) on a strongly magnetic material by oxygen and other elements is well documented.32 Several researchers observed the decrease in the saturation magnetization of polymer-coated coreshell particles. This behavior is likely due to the loss of the net magnetic moment at the interface between the oxide core and the polyelectrolyte shell. Studies have found that the magnetic properties of the NPs diminished because of their interactions with chromophores attached to the surfactant molecules.32,42 Pure and coated γFe2O3 NPs displayed strong hysteresis under an applied magnetic field of (3 T at 300 K (Figure 8A). Hysteresis of the pure and coated NPs expressed the strong ferrimagnetic nature of the materials. The magnetic coercivity (Hc) of the hysteresis loop has been defined as the intensity of the applied magnetic field needed in the reverse direction to bring the field to zero after achieving saturation magnetization.18 The coercivity (Hc) diagram of the pure and PAA50K- and HA-coated NPs manifested the relative decrease in the coercive energy of the coated NPs from the pure γFe2O3 NPs (Figure 8C). Therefore, coating the NP surface with high-molecular-weight PAA and HA fractions reduced the interparticle magnetic force and hence enhanced the stabilization of the γFe2O3 NPs.

’ CONCLUSIONS The colloidal stability of the PAA- and HA-coated γFe2O3 NPs as affected by the suspension concentration and nature of electrolytes was evaluated. The interplay of repulsive electrostatic, steric, and attractive magnetic dipolar and van der Waals forces determined the overall colloidal stability of the coated NPs. The sorption of polyelectrolytes on γFe2O3 NPs enhanced the colloidal stability manifold compared to that of the pure γFe2O3 NPs. The electrostatic and steric factors introduced by coating the polyelectrolytes on the NP surface were instrumental in stability enhancement. Long-chain, high-molecular-weight fractions of the adsorbed PAA molecules induced higher steric stabilization compared to that of their low-molecular-weight counterpart. However, relatively higher stabilization was observed for weakly polar HA-coated γFe2O3 NPs than for PAAcoated γFe2O3NPs. At low pH, strong destabilization of the coated NPs was observed at relatively lower ionic strength. The desorption of polyelectrolytes during dilution with deionized water possibly decreased the suspension stability. The colloidal stability of the coated NPs was enhanced significantly with increasing pH. The higher degree of ionization of the weakly charged polyelectrolytes adsorbed on the NP surface with increased pH enhanced the electrosteric contribution exceedingly to achieve colloidal stability. AFM images of HA-coated NPs at pH 5 and 9 manifested the conformational differences of the adsorbed HA on the NP surface. At low pH, a decrement in the electrostatic repulsions by the addition of Naþ and Ca2þ 8042

dx.doi.org/10.1021/la200772e |Langmuir 2011, 27, 8036–8043

Langmuir triggered the short-range van der Waals and magnetic dipolar attractive forces. However, at high pH the long-range electrosteric stabilization could counterbalance the long-range attractive magnetic dipolar interaction. Therefore, the pH-induced conformation change in the polyelectrolytes adsorbed on the NP surface played a pivotal role in terms of the colloidal stability of the coated γFe2O3 NPs. However, the maintenance of the steric stabilization factor is imperative to retaining the overall stability and efficacy of the coated NPs.

’ AUTHOR INFORMATION Corresponding Author

*Phone: 413-545-5212. Fax: 413-545-3958. E-mail: bx@pssci. umass.edu.

’ ACKNOWLEDGMENT This project was supported by the Massachusetts Agricultural Experimental Station (MA 00978) and the NSF (CMM10531171). We express our sincere gratitude to Prof. Mark Tuominen, Department of Physics (UMass Amherst) for his approval to use the AFM and SQUID instruments. We also sincerely acknowledge the help from Ms. Lizzy Wang and Dr. Nihar R. Pradhan, Department of Physics (UMass Amherst) for their assistance with SQUID measurements. ’ REFERENCES (1) Lam, U. T.; Mammucari, R.; Suzuki, K.; Foster, N. R. Ind. Eng. Chem. Res. 2008, 47, 599–614. (2) Gangopadhyay, P.; Gallet, S.; Franz, E.; Persoons, A.; Verbiest, T. IEEE Trans. Magn. 2005, 41, 4194–4196. (3) Liu, J.; Zhao, Z.; Jiang, G. Environ. Sci. Technol. 2008, 42, 6949–6954. (4) Saleh, N.; Phenrat, T.; Sirk, K.; Dufour, B.; Ok, J.; Sarbu, T.; Matyiaszewski, K.; Tilton, R. D.; Lowry, G. V. Nano Lett. 2005, 5, 2489–2494. (5) Auffan, M.; Rose, J.; Proux, O.; Borschneck, D.; Masion, A.; Chaurand, P.; Hazemann, J.; Chaneac, C.; Jolivet, J.; Wiesner, M. R.; Van Geen, A.; Bottero, J. Langmuir 2008, 24, 3215–3222. (6) Pisanic, T. R.; Blackwella, J. D.; Shubayevb, V. I.; Fi~nonesc, R. R.; Sungho, J. Biomaterials 2007, 28, 2572–2581. (7) Berry, C. C.; Wells, S.; Charles, S.; Curtis, A. S. G. Biomaterials 2003, 24, 4551–4557. (8) Min, Y.; Akbulut, M.; Kristiansen, K.; Golan, Y.; Israelachvili, J. Nat. Mater. 2008, 7, 527538. (9) McClements, D. J. Langmuir 2005, 21, 9777–9785. (10) Ballauff, M. Prog. Polym. Sci. 2007, 32, 1135–1151. (11) Pan, B.; Ghosh, S.; Xing, B. Environ. Sci. Technol. 2008, 42, 1594–1599. (12) Kang, S. H.; Xing, B. S. Environ. Sci. Technol. 2005, 39, 134–140. (13) Ghosh, S.; Mashayekhi, H.; Bhowmik, P.; Xing, B. Langmuir 2010, 26, 873–879. (14) Xu, J.; Stevens, M. J.; Oleson, T. A.; Last, J. A.; Sahai, N. J. Phys. Chem. C 2009, 113, 2187–2196. (15) Bergkvist, M.; Carlsson, J.; Oscarsson, S. J. Biomed. Mater. Res., Part A 2003, 64A, 349–356. (16) Bickmore, B. R.; Nagy, K. L.; Sandlin, P. E.; Crater, T. S. Am. Mineral. 2002, 87, 780–783. (17) Vinelli, A.; Primiceri, E.; Brucale, M.; Zuccheri, G.; Rinaldi, R.; Samori, B. Microsc. Res. Tech. 2008, 71, 870–879. (18) Lu, Z. H.; Prouty, M. D.; Guo, Z. H.; Golub, V. O.; Kumar, C. S. S. R.; Lvov, Y. M. Langmuir 2005, 21, 2042–2050. (19) Khalaf, M.; Kohl, S. D.; Klumpp, E.; Rice, J. A.; Tombacz, E. Environ. Sci. Technol. 2003, 37, 2855–2860.

ARTICLE

(20) Kirwan, L. J.; Fawell, P. D.; van Bronswijk, W. Langmuir 2003, 19, 5802–5807. (21) Verm€ohlen, K.; Lewandowski, H.; Narres, H., -D.; Schwuger, M. J. Colloids Surf., A 2000, 163, 45–53. (22) Kang, S.; Xing, B. Langmuir 2008, 24, 2525–2531. (23) Boisvert, J. -P.; Malgat, A.; Pochard, I.; Daneault, C. Polymer 2002, 43, 141. (24) Yu, S.; Chow, G. M. J. Mater. Chem. 2004, 14, 439–448. (25) Eberbeck, D.; Wiekhorst, F.; Steinhoff, U.; Trahms, L. J. Phys.: Condens. Matter 2006, 18, S2829–S2846. (26) Wisniewska, M. Powder Technol. 2010, 198, 258–266. (27) He, F.; Zhao, D. Environ. Sci. Technol. 2007, 41, 6216–6221. (28) Chen, K. L.; Mylon, S. E.; Elimelech, M. Environ. Sci. Technol. 2006, 40, 1516–1523. (29) Fritz, G.; Schadler, V.; Willenbacher, N.; Wagner, N. J. Langmuir 2002, 18, 6381–6390. (30) Weitz, D. A.; Huang, J. S.; Lin, M. Y.; Sung, J. Phys. Rev. Lett. 1985, 54, 1416–1419. (31) Shen, L. F.; Stachowiak, A.; Fateen, S. E. K.; Laibinis, P. E.; Hatton, T. A. Langmuir 2001, 17, 288–299. (32) Burke, N. A. D.; Stover, H. D. H.; Dawson, F. P. Chem. Mater. 2002, 14, 4752–4761. (33) Culliity, B. D. Introduction to Magnetic Materials; Addison-Wesley: Reading, MA, 1972. (34) Goya, G. F.; Berquo, T. S.; Fonseca, F. C.; Morales, M. P. J. Appl. Phys. 2003, 94, 3520–3528. (35) Guo, X.; Ballauff, M. Phys. Rev. E 2001, 64, 051406. (36) Wittemann, A.; Drechsler, M.; Talmon, Y.; Ballauff, M. J. Am. Chem. Soc. 2005, 127, 9688–9689. (37) Roiter, Y.; Trotsenko, O.; Tokarev, V.; Minko, S. J. Am. Chem. Soc. 2010, 132, 13660–13662. (38) Kiriy, A.; Gorodyska, G.; Minko, S.; Tsitsilianis, C.; Jaeger, W.; Stamm, M. J. Am. Chem. Soc. 2003, 125, 11202–11203. (39) Namjesnik-Dejanovic, K.; Maurice, P. A. Geochim. Cosmochim. Acta 2001, 65, 1047–1057. (40) Ghosh, S.; Mashayekhi, H.; Pan, B.; Bhowmik, P.; Xing, B. Langmuir 2008, 24, 12385–12391. (41) Mei, Y.; Wittemann, A.; Sharma, G.; Ballauff, M.; Koch, T.; Gliemann, H.; Horbach, J.; Schimmel, T. Macromolecules 2003, 36, 3452–3456. (42) Kataby, G.; Cojocaru, M.; Prozorov, R.; Gedanken, A Langmuir 1999, 15, 1703–1708.

8043

dx.doi.org/10.1021/la200772e |Langmuir 2011, 27, 8036–8043