The Electrokinetic Properties of Colloidal Magnetic Iron Oxides

Article pubs.acs.org/Langmuir

The Electrokinetic Properties of Colloidal Magnetic Iron Oxides I. M. Metcalfe† and T. W. Healy* Particulate Fluids Processing Centrè, Department of Chemical & Biomolecular Engineering, University of Melbourne, Vic. 3010, Australia S Supporting Information *

ABSTRACT: A novel electrokinetic streaming potential technique has been used to determine the ζ potential behavior of three magnetic iron oxides, (Fe3O4, γ-Fe2O3, and CoFe2O4) as a function of pH and salt concentration. These colloidal materials, (nanosize in one dimension), are held in the form of a plug by means of external magnets. The streaming potential (E) is measured as a function of fluid flow induced by a pressure drop (ΔP) across the plug. The magnetically held plug is found to obey the requirements of the streaming potential technique; in each case an iso-electric point, (iep) independent of salt concentration is observed. However, if one uses the appropriate quantities in the standard formula, the calculated ζ potentials are very much lower than for oxides such as silica, alumina or goethite and other colloidal oxide, latex, etc. particulates in aqueous salt solutions. Furthermore, at a given pH, the measured ζ potentials anomalously increase in magnitude rather than decrease as observed conventionally as the salt concentration is increased. This apparent anomalous behavior could not be eliminated by incorporating surface conductance effects. However by including a conductance pathway, independent of pH or salt concentration, through the magnetic particle network itself, the anomaly was removed. Confirmation of the role of a conductance pathway through the magnetic particle network was obtained by using silica coated magnetic particles which displayed normal electrokinetic behavior. Finally, we have redesigned the plug-electrode assembly to allow measurement of streaming current, a technique know to eliminate contributions from plug network conductances of any kind. The resulting ζ potentials, derived from this streaming current technique are normal.

1. INTRODUCTION Colloidal nanomagnetic iron oxide particles, once used extensively in information storage-reproduction technologies, are finding increasing use in the preparation and delivery of biomolecular effects. A biomolecule, adsorbed or chemically bound at the magnetic particle−water interface can be reacted chemically and the magnetic particles held or moved by external magnets, then washed, resuspended, and then further treated to provide novel properties for use in biosensors and other such platforms.1−5 Our goal in the present paper is to measure the fundamental, underlying electrokinetic properties of the magnetic particle− water interface itself so as to better control the uptake of biomolecular species and fragments for subsequent use, in particular, in novel biosensor platforms. To do this we have developed an electrokinetic streaming potential technique that relies on the ability of an external magnet to hold a bed or plug of particles in a fixed position. The particle plug is collected from a suspension of particles within the flowing or circulating aqueous suspension. The same chemical processing option used in biomolecule chemistry is used in our streaming potential technique. A suspension of the magnetic particles in an electrolyte solution at a given pH can be stirred and brought to equilibrium. By pumping that suspension through a glass pipe, external magnets can be used to “trap” a plug of particles across which a pressure drop, (ΔP) appears. Electrodes on either side © 2012 American Chemical Society

of the plug record a potential difference (E). From the slope of an E vs ΔP plot, the electrokinetic or ζ potential can be determined via standard equations.6,7 The ratio of the streaming potential to the pressure difference (E/ΔP) is proportional to the potential at the plane of shear or ζ potential by the Helmholtz−Smoluchowski equation

E/ΔP = (ζεε0)/ηλ

(1)

where ε is the permittivity of the plug liquid, ε0 is the permittivity of free space, η is the fluid viscosity in the plug, and λ is the conductivity of the plug, which can be equated to the conductivity of the bulk electrolyte under conditions where “surface conductance” effects are unimportant. The derivation of this equation is given elsewhere.6,7 In summary the key assumptions are that the flow of electrolyte in the plug must be laminar and the plug must be rigid. For aqueous solutions we have ζ = 84.88 × 107η(E /ΔP)(λ /ε) (mV)

(2)

where E is in mV, λ is in Sm−1,η is in centipoises, and ΔP is in cmHg. Received: August 26, 2011 Revised: April 20, 2012 Published: May 11, 2012 7897

dx.doi.org/10.1021/la3010486 | Langmuir 2012, 28, 7897−7903

Langmuir

Article

Figure 1. Streaming potential apparatus with external magnets at M trapping a plug of magnetic particles. Two magnet configurations were employed to give plugs where the network of magnetic particles is formed at right angles to the flow (see inset a), and by using a doughnut magnet, the network forms parallel to the flow (inset b). (Further dimensions are given in text. See also Supporting Information.) measurements it was suspected that this product contained an unsaturated fatty acid analogue commonly used in manufacture as a crystal growth control additive. Treatment by using repeated Soxhlet extractions with chloroform-ethanol resulted in a reproducible isoelectric point of pH 8.2. That the material was indeed magnetite was confirmed by infrared spectroscopy on KBr-magnetite disks. The particles are ferromagnetic but most likely multidomain,8 with respect to their magnetic properties. The bulk conductivity of the magnetite particles is expected to be in the range 1 to 1.5 Sm−1.9 Toda CoFe2O4 (Sample B). Cobalt doped magnetite, also supplied by Toda Korbishi Corporation of Japan has the stoichiometric formula CoxFe(1‑x)O·Fe2O3. The precise ratio (via isomorphic substitution) of cobalt to ferrous ions and the manufacture of this material is proprietary information. This magnetic material appeared slightly brown in color but had physical properties entirely consistent with the black Toda magnetite described above. From knowledge of the recording industry, we may assume this is a cobalt impregnated magnetite and was formed by laying down a CoO oxide coating on a Fe3O4 substrate. The cobalt is then incorporated into the lattice structure by heating in a nonreactive atmosphere. The typical result is less than 1% of the ferrous ions are replaced by this method. Bayer Maghemite γ-Fe2O3 (Sample C). Maghemite was supplied by the Bayer Group, Germany. Electron micrographs indicate this light brown material to consist of 200 nm (approximately) cubes. Infrared results for this material compared favorably with the spectra for maghemite. Silica-Coated Magnetite SiO2.Fe3O4 (Sample D). In order to produce a magnetic core−shell nonconducting colloid, known amounts of sample A were treated to produce a composite particle with a coherent surface layer of polymeric so-called “dense” silica. Following a method similar to that of Furlong et al10 and that of Spuch-Calvar et al.,1 10 g of magnetite was added to 60 mL of triply distilled water in order to make an aqueous slurry which was subsequently heated to 85 °C. Over a period of 2 h, concurrent additions of 40 mL (0.25 mol·dm−3) technical grade caustic sodium silicate solution and 3 mol·dm−3 sulfuric acid were made to maintain the slurry at a pH of 10.0 ± 0.2. Throughout the procedure and for one hour after all additions, the solution was stirred vigorously. Finally the slurry was cooled to room temperature and the solid filtered, washed and vacuumdried. The filtrate was analyzed for dissolved silica. In all cases the low dissolved silica results obtained, using the reduced α-molybdosilicic acid

If one chooses to use the bulk electrolyte conductance, then ζ = 84.88 × 107η(E/ΔP)(λ 0 /ε) + 2λs /a (mV)

(3)

where λ0 is the bulk electrolyte conductivity and λs is the interfacial or surface conductance and a is the mean radius of the capillaries in the plug.

2. EXPERIMENTAL SECTION 2.1. The Streaming Potential Technique. A schematic representation of the streaming potential apparatus is shown in Figure 1. A suspension of the magnetic particles in an electrolyte of known concentration and pH can be pumped in a loop between the two one liter Pyrex glass vessels identified as B in Figure 1. When equilibrium is attained, external magnets are inserted at position M trap a plug of particles inside a 7 mm i.d. glass tube to create a conventional streaming potential packed bed. For measurement of the mean radius of the capillaries of the plug, flow was in one direction and the volume flow recorded as a function of applied pressure. For determination of the streaming potential, flow was undertaken in both directions with the potential recorded when the liquid levels in the reservoirs (B) were equal. No asymmetry in the measured potential with left to right then right to left flow was observed indicating that polarization anisotropy was absent. The value of half the observed potential between the two flow directions was recorded as the steaming potential at each value of the applied pressure. The whole apparatus was housed in a Faraday cage within a temperature controlled room set at 22 ± 0.5 °C. For streaming potential measurements platinized platinum electrodes at location E in Figure 1 were employed. For streaming current measurements the simple u-tube (C) was redesigned to use two movable in-cell silver/silver chloride electrodes. These electrodes were adjusted by external micrometers to be located as close as possible to the surfaces of the trapped plug at M in Figure 1. 2.2. Materials. Toda Magnetite Fe3O4 (Sample A). Toda Korbishi Corporation of Japan supply reasonably monodispersed prolate ellipsoid particles approximately 300 nm in length with an axial ratio in the order of 10:1 used in the production of magnetic disks and tapes. A TEM of the magnetite is given in the Supporting Information. The surface area of this material, calculated from BET multipoint N2-adsorption analysis was 17.5 m2/g. This value was confirmed by electron-microscopy. From initially low iso-electric point (iep) 7898

dx.doi.org/10.1021/la3010486 | Langmuir 2012, 28, 7897−7903

Langmuir

Article

considerably higher flow rates at a given pressure as shown below in Table 1. However, the E/ΔP results, and hence the ζ potentials are, within experimental error, the same by both techniques as shown below in Table 2. Third, for the various magnetic particles we have used, all plugs form and exist as very low volume fraction columnar particle networks or clusters made up of a collection of particles (mostly needle shaped, of length 300−500 nm and “diameter” of 25−30 nm). This translates to an average pore radius within the plug of approximately (4.8 ± 0.2) × 10−6 m. Whether surface conductance effects are to be neglected or included depends on the magnitude of κa where κ−1 is the Debye length and a is the mean radius of the capillary pores in the plug. For the present system the κa values range from 165 for 10−4 mol·dm−3 to 1645 for 10−2 mol·dm−3 respectively and to 9000 for 0.3 mol·dm−3 electrolyte concentrations. For the 10−3 mol·dm−3 system in axial configuration, the κa value is 9360. These very large values of κa confirm that surface conductance effects will be negligible. That the values of the streaming potential are essentially the same for both normal, (4.10−6 m pore radius) and axial (90 × 10−6 m pore radius) configurations is a further justification for ignoring surface conductance effects. The κa results place the present plugs in the very top corner of the lines of Figure 2 of Delgrado et al,7 i.e., where the ratio of the streaming potentials with and without incorporation of surface conductance is unity. Furthermore, we have used the graphically representations given by Lyklema13 to generate possible surface conductance values of ca. 10−10 S. This translates to Dukhin numbers of 0.53, 0.0066, and 0.00028 for 10−4, 10−3, and 10−2 mol·dm−3 respectively. Again such low numbers confirm that our plug calculations can ignore surface conductance effects. 3.2. Electrokinetic Potential Results. The electrokinetic (ζ) potential of the Toda magnetite (sample A) at 10−3 mol·dm−3, as a function of pH is shown in Figure 3. The observed iep at pH 8.2 agrees with the value observed by Tombacz et al,8 but falls above the mid-pH range identified in the recent summaries of synthetic magnetites by falls just above the mid-pH range identified in the recent summary by Erdemoglu and Sarikaya.14 The importance of the chloroformethanol washing procedure is illustrated by Dixon,15 who established that the iep of mineral magnetite was shifted from circa pH 4 to above pH 7 following washing. The effect of variation in background electrolyte concentration on the electrokinetic (ζ) potential of Toda magnetite (sample A) as a function of pH is shown in Figure 4. The common intersection point at zero ζ potential, independent of electrolyte concentration identifies the iep at pH 8.2. The observed electrolyte effect of an increase in magnitude of the ζ potential at any given pH is opposite to that observed in all other oxide−water systems. Further the actual magnitude of the ζ potentials we have observed are significantly lower to those observed in oxide−water interfacial systems. The general electrokinetic properties of the cobalt doped magnetite (sample B) and the maghemite (sample C) materials are shown in Figures 5 and 6. Note that both materials shown significantly lower magnitudes of ζ potentials than expected for inorganic oxides. For the cobalt doped magnetite (sample B) the inverse salt concentration effect is also observed. The results of Morimoto and Kittaka16 also shown the anomalously low ζ potential our studies have identified.

colorimetric technique were indicative of polymeric-silica coating and consistent with the literature (Furlong et al.10). Electron micrographs showed similar physical characteristics as the precursor particles (sample A). Electrokinetic measurements made to determine the isoelectric point at pH 2.5−3.0, indicated the presence of a silica surface outer layer. Reagents. All electrokinetic experiments were performed at room temperature (ca. 25 °C). All reagents were of analytical grade or equivalent except where indicated. Triply distilled water prepared in an all-Pyrex, three-stage still; the second stage containing alkaline potassium permanganate to remove oxidizable organic impurities. The conductivity of this water was measured to be 1.0 × 10−6 Sm−1 (at 25 °C) and was utilized in the preparation of all dispersions and electrolyte solutions.

3. RESULTS 3.1. Properties of the Magnetic Particle Plugs. Plugs were formed from 0.22 g of 300 × 30 nm. rod-like, magnetic particles, suspended in various concentrations of electrolyte solutions. When trapped, the particles formed a plug of approximately one cm in length and approximately 10% by volume of particles. These plugs form as networks of interconnected clusters (similar to those of Figure S3 in the Supporting Information) and as previously shown schematically in the work of de Vicente et al11 and Safran.12 The mean capillary radius within the plugs was derived from standard flow analysis based on Darcy’s Law and the Kozeny−Carman equation. For completeness, the flow properties were determined for both normal and axially held plugs and for several electrolyte concentrations between 0.3 and 10−4 mol·dm−3 KCl and in distilled water. The effective capillary radii obtained by this procedure over a range of electrolyte concentrations yielded a pore radius of (4.8 ± 0.20) × 10−6 m. For an axially held plug in 10−3 mol·dm−3 as expected, a much larger capillary radius of 90 × 10−6 m was obtained. Three important aspects of these plugs need comment. First, the external magnets (3000 G or 300 mT) are powerful enough to give stable plugs over a wide range of flow rates. Extreme flow rates, particularly in the axial configuration, do cause the plug to collapse. Just prior to collapse we did observe slight deformation of the plug shape in both flow modes. By keeping below these limits, the E vs ΔP behavior, as shown in Figure 2, is remarkably linear and passes through the origin.

Figure 2. Streaming potential voltage (E) vs pressure (ΔP) behavior for a typical magnetic oxide (magnetite) in 10−3 mol·dm−3 KCl at two pH values; (“normal” configuration).

Second, the normal and axial plugs give different flow rates at any given pressure as might be expected, with axial plugs giving 7899

dx.doi.org/10.1021/la3010486 | Langmuir 2012, 28, 7897−7903

Langmuir

Article

Table 1. Streaming Potential Flow Rates (Q, cm3/s) as a Function of Magnetic Field Alignment (in “Normal” and “Axial” Configurations) for 10−3 mol·dm−3 KCl “normal” configuration

“axial” configuration

volume (cm3)

time (s)

pressure (cmHg)

Q

volume (cm3)

time (s)

pressure (cm Hg)

Q

10.85 6.0

326.7 202.2

1.68 1.45

0.033 0.029

68 34

30 60

0.55 0.22

2.23 0.57

Table 2. Streaming Potential Results: Potential/Pressure Ratios as a Function of pH, Electrolyte Concentration and Magnetic Field Alignment “normal” configuration

“axial” configuration

pH

[KCl] (M)

E (mV)

P (cmHg)

E/ΔP

E (mV)

P (cmHg)

E/ΔP

6.21 5.35 7.83 8.83

10−3 10−2 10−2 10−2

+1.691 +0.46 +0.075 −0.120

2.110 1.816 1.618 2.310

+0.806 +0.253 +0.046 −0.051

+0.350 +0.096 +0.018 −0.017

0.412 0.471 0.305 0.420

+0.850 +0.204 +0.058 −0.041

Figure 3. Variation of the ζ potential of Toda magnetite (sample A) as a function of pH at 10−3 mol·dm−3. The chloroform−ethanol wash is used to remove any residual fatty acid or other organic impurities.

Figure 5. ζ potentials vs pH results at two electrolyte concentrations: synthetic cobalt-doped magnetite, sample B.

Figure 4. ζ potentials vs pH results for various electrolyte concentrations: synthetic magnetite, sample A.

Figure 6. ζ potentials vs pH results for two synthetic maghemite products: sample C (present study) and that of Morimoto and Kittaka.16

(b) The general shape of the ζ potential vs pH curves of all magnetic materials studied is consistent with many other materials for which H+ and OH− are potential determining ions. (c) The magnitude of the ζ potentials observed at low electrolyte concentrations are significantly lower than expected. Again, low ζ potential values are reported, most often without comment, by other workers.21,22 (d) The salt concentration dependence is anomalous, in that the magnitude of the potential at any pH increases with

The electrokinetic potential behavior of the silica coated magnetite in both magnitude and in respect to the effect of increasing electrolyte concentration are “normal” as shown in Figure 7. By way of summary of the electrokinetic ζ potential results, we note that the present simple, innovative streaming potential electrokinetic technique utilizing the magnetic properties of the particles yields the following important results: (a) The iep values for the various materials are in general agreement with other published results. 7900

dx.doi.org/10.1021/la3010486 | Langmuir 2012, 28, 7897−7903

Langmuir

Article

Figure 8, showing restoration of the expected ζ potential magnitudes and the normal decrease in magnitude of the potential with increasing salt concentration.

Figure 7. Variation of the measured ζ potential of silica coated magnetite as a function of increasing salt concentration at 3.3 pH units from the pH of the iep, with comparisons of electrokinetic results of other pure quartz and vitreous silica−electrolyte interfaces due to Scales et al.,17 Wiese et al.,18 Lyons et al.,19 and Kulkarni and Somasundaran.20

Figure 8. ζ potential vs log electrolyte concentration for sample A magnetite at a pH 3.3 units from the isoelectric point. Shown are the calculations using eq 3 and a curve corrected by using an additional constant, estimated particle network (PN) conductivity of 4.4 × 10−3 Sm−1.

increase in electrolyte concentration. Again, we have found that other workers22 have reported this anomalous result, usually without comment, Others21,22 report ζ potentials that are essentially independent of background salt concentration. (e) Coating of the magnetite particles with silica produces conventional salt dependence and magnitude in the measured ζ potentials.

The estimated constant value, independent of ionic strength, overcomes the two anomalies of low ζ potentials and uncharacteristic variation of increased magnitude of ζ potentials with added salt. Since this second pathway is eliminated by the insulating silica coating, a conductance pathway could exist from the conducting magnetic oxide particle to surface groups on the particle and thence across the particle−particle gap in clusters in the plug. The conducting species in the particle− particle gap is most likely to be protons. The source of those protons are the ionizable group sites at the interface. Charge regulation process23 will require changes in ionization of the surface sites to accommodate the charge change from the electron transfer process. We now need to couple the internal particle conductivity to this gap pathway. Such coupling is likely to be via the Fe(III)/Fe(II) redox system. It is this redox couple that can no longer operate into the gap when the magnetic oxide is coated with the insulating silica. A schematic of our proposed additional conduction path is shown in Figure 9.

4. DISCUSSION The electrokinetic effect requires a conduction pathway within the streaming potential plug. This conductivity consists of conductivity through the free bulk electrolyte within the plug and any other pathways, such as those within the normal (diffuse layer) and anomalous (through the Stern layer) electrical double layer. The present experiments with the uncoated and silica-coated magnetic iron oxides suggest that an additional conduction pathway within the conducting magnetic oxide particle−particle network occurs which adds to that of the bulk electrolyte. The insulating silica coating prevents such a pathway and results for the silica coated particle are normal and show no evidence of a surface conductance correction, which again confirms our earlier assessment that our very large κa values also exclude surface conductance effects. To explore this suggestion we recalculated the ζ potentials using the bulk electrolyte conductivity plus a reasonable estimate of 4.4 × 10−3 Sm−1 for this additional particle network conductivity. (This compares with a bulk conductivity of 1.5 × 10−3 Sm−1 for 10−4 mol·dm−3 KCl and of 1.15 × 10−1 Sm−1 in 80 × 10−3 mol·dm−3 KCl.) The application of this estimated value for the particle network conductivity results in ζ potentials which are comparable in behavior and magnitude to those of other iron oxides. The estimated value of 4.4 × 10−3 Sm−1 for the additional particle network conductivity is a reasonable one given that we would expect about 1.5 Sm−1 for bulk magnetite and the “chain & cluster” particle network value would be considerably lower than this. Using capillary radii of 4 × 10−6 m. (normal alignment) and 90 × 10−6 m (axial alignment), we can convert this value to “surface conductance units” of 1.76 × 10−8 and 3.9 × 10−7 S respectively. These values are 100 to 1000 times the values listed by Lyklema13 for anticipated contributions of surface conductance. The effect of this additional pathway of constant conductivity is shown in

Figure 9. Schematic representation of the additional conduction path through the plug from crystal to interface to solution to interface to crystal.

We are not suggesting that the particles touch; desolvation energies of the hydrated interfacial groups are prohibitively large to allow such intimate ohmic contact. However our analysis suggests that conduction occurs across the gap solution. This model of particle conduction, surface redox and site dissociation 7901

dx.doi.org/10.1021/la3010486 | Langmuir 2012, 28, 7897−7903

Langmuir

Article

while speculative does have support from recent studies of reduction of compounds by magnetite suspensions.24 We note a very early suggestion25 wherein the anomalous conductivity of magnetite in streaming potential plugs was observed. By way of summary, we note the following: • The use of the “estimated” conductance value is simply to illustrate the order of magnitude of the effect we see. Any additional conduction pathway would have to be (a) relatively independent of bulk electrolyte concentration and (b) of a large magnitude (certainly much greater than could be accounted for by classical surface conductance). • If the anomalous behavior was due to purely structural causes (ie: shape and size of the particle networks and pore sizes) then we would expect the silica coated magnetite to show similar anomalous behavior with electrolyte concentration. • We are indeed suggesting that (a) adding silica blocks any redox derived conduction pathways and (b) that these magnetic iron oxides simply have anomalous behaviordue to the possibility for additional conduction pathways. For the streaming potential technique, we use a large external resistance (∼1014 ohm). Thus any surface or other conductivity will become a significant contribution which when not accounted for, results in an underestimate of the total system conductivity. This streaming potential dilemma can be resolved by switching to “streaming current” format where an external variable resistance is allowed to go, in the limit, to zero from the infinite value in the streaming potential configuration. Now any contributions to the plug resistance due to conductance within the diffuse layers are no longer significant.6,13 The comparison of the two techniques is shown in Figure 10 for the ζ potential

Figure 11. Variation of the ζ potential vs electrolyte concentration for magnetite (sample A) at a fixed pH, as determined by both streaming potential and streaming current techniques.

ensure no influence of surface conductivity. Rather they suggest that in their SAM systems the most plausible explanation, as we have argued herein, is a contribution to the bulk conductivity from the underlying gold substrate. They add further that because the ζ potential from the streaming current, measured by shorting the electrodes through a zero resistance shunt, is independent of the surface or substrate conductivity, they then used steaming current data in the further interpretation of ζ potential data. The apparent maximum shown in Figure 11 in the magnitude of the ζ potential at 10−3 mol·dm−3 from streaming current measurements is probably due to experimental error introduced by the more complicated (but necessary) adjustable in-cell electrode assembly. For both streaming potential and streaming current it was observed that there were some difficulties in keeping the pH at the same constant value, at each ionic strength. These problems produce uncertainties in the measured ζ potentials of ±2 mV for streaming potential results and ±10 mV for streaming current results.

5. SUMMARY We present a novel electrokinetic technique utilizing the magnetic properties of the colloidal nanoparticles of magnetic iron oxides. Despite excellent determinations of the pH of the iep of the materials tested, we, and earlier workers observe that the magnitude of the ζ potentials are low compared to other oxide nonmagnetic, nonconducting materials, and the magnetic materials show an anomalous dependence of increasing magnitude in ζ potential with increasing salt concentration. Our proposal is that we can eliminate these anomalous effects by coating the particles with a layer of insulating silica, or by inserting an estimation of the conductivity due to the particle network into the streaming potential calculations or by use of the streaming current technique. We suggest that the additional conduction path involves Fe(II)/Fe(III) electron exchange from the conducting particle to surface sites, followed by charge regulation of the surface site equilibrium at the interface with proton exchange across the particle−particle cluster network gaps. We further suggest that researchers using magnetic particles for bioapplications should consider the possibility that uncoated (or even partially coated) magnetite particles might offer novel conduction pathways in network structures.

Figure 10. ζ potential−pH results for magnetite (sample A) as determined by streaming potential and streaming current techniques at a fixed electrolyte concentration of 10−3 mol·dm−3.

as a function of pH at fixed electrolyte and in Figure 11 for the ζ potential as a function of salt concentration at a fixed pH of 3.3 pH units below the pH of the iep. The anomalous behavior illustrated in the results of Figure 4 and in other systems is no longer displayed. The salt dependence and the magnitude of the potentials are both normal. The occurrence of anomalous conductivity in streaming potential measurements has been reported recently by Schweiss et al.26 in their work on the pKa values of SAM monolayers on gold substrates. They used a slit channel set up with 50 μm height to 7902

dx.doi.org/10.1021/la3010486 | Langmuir 2012, 28, 7897−7903

Langmuir

■

Article

(14) Erdemoglu, M.; Sarikaya, M. Effects of Heavy Metals and Oxalate on the Zeta Potential of Magnetite. J. Colloid Interface Sci. 2006, 300, 795−804. (15) Dixon, D. R. Magnetic adsorbents: Properties and applications. J. Chem. Techol. Biotechnol. 1980, 30, 572−578. (16) Morimoto, T.; Kittaka, S. The Electrification of Iron Oxide in Water. Bull. Chem. Soc. Jpn. 1973, 46 (10), 3040. (17) Scales, P. J.; Grieser, F.; Healy, T. W.; White, L. R.; Chan, D. Y. C. Electrokinetics of the Silica-Solution Interface: A Flat Plate Streaming Potential Study. Langmuir 1992, 8, 965−974. (18) Wiese, G. R.; James, R. O.; Healy, T. W. Discreteness of Charge and Solvation Effects in Cation Adsorption at the Oxide-Water Interface. Discuss. Faraday Soc. 1971, 52, 302−311. (19) Lyons, J. S.; Furlong, D. N.; Homola, A.; Healy, T. W. A Radial Flow Streaming Potential Apparatus for Electrokinetic Measurements of Sheet or Plate Materials. Aust. J. Chem. 1981, 34 (6), 1167. (20) Kulkarni, R. D.; Somasundaran, P. Proceedings of the Symposium on Oxide-Electrolyte Interfaces, Miami, USA, 1972; Alwitt,R. S., Ed.; The Electrochem. Soc.: New York, 1973; p 31) (21) Arias, J. L.; Harivardhan Reddy, L; Couvreur, P. Magnetoresponsive Squalenoyl Gemcitabine Composite Nanoparticlesor Cancer Active Targeting. Langmuir 2008, 24, 7512−7519. (22) Jiang, C; Wang, R.; Ma, W. The effect of magnetic nanoparticles on Microcystis aeruginosa removal by a composite coagulant. Colloids Surf. A, Physicochem. Eng. Aspects 2010, 369, 260. (23) Chan, D; Healy, T. W.; White, L. R. Electrical Double Layer Interaction Under Regulation by Surface Ionization Equilibria Dissimilar Amphoteric Surfaces. J. Chem. Soc. Faraday Trans. I 1976, 72, 2844. (24) Gorski, C; Nurmi, J; Tratnkek, P; Hofstette, T; Schere, M. Redox Behavior of Magnetite: Implications for Contaminant Reduction. Environ. Sci. Technol. 2010, 44, 55−60. (25) O’Connor, D. J.; Street, N.; Buchanan, A. S. Charge Densities and Surface Conductances at Solid-Liquid Interfaces. Aust. J. Chem. 1954, 7, 245. (26) Schweiss, R; Welzel, P. B.; Werner, C; Knoll, W. Dissociation of Surface Functional Groups and Preferential Adsorption of Ions on Self-Assembled Monolayers Assessed by Streaming Potential and Streaming Current Measurements. Langmuir 2001, 17, 4304−4311.

ASSOCIATED CONTENT

S Supporting Information *

Additional details of the streaming potential technique and TEM images of the magnetite particles. This material is available free of charge via the Internet at http://pubs.acs.org/.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Present Address †

Now at Adaptive Learning, Vic., Australia. E-mail: ian@ adaptivelearning.com.au. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS

■

REFERENCES

The authors acknowledge funding from the Australian Research Council under the Special Research Centres program in support of this work. We dedicate this paper to the memory of Bev Healy, who died on February 19, 2009, and to Trevor Metcalfe who died on May 31, 2009.

(1) Spuch-Calvar, M.; Rodríguez-Lorenzo, L.; Puerto Morales, M. P.; Alvarez-Puebla, R. A.; Liz-Marzán, L. M. Bifunctional Nanocomposites with Long-Term Stability as SERS Optical Accumulators for Ultrasensitive Analysis. J. Phys. Chem. C 2009, 113, 3373−3377. (2) Qu, H.; Caruntu, D.; Liu, H.; O’Connor, C. J. Water-Dispersible Iron Oxide Magnetic Nanoparticles with Versatile Surface Functionalities. Langmuir 2011, 27, 2271−2278. (3) Teo, B.; Chen, F.; Hatton, T. A.; Grieser, F.; Ashokkumar, M. Novel One-Pot Synthesis of Magnetite Latex Nanoparticles by Ultrasound Irradiation. Langmuir 2009, 25, 2593−2595. (4) del Campo, A.; Sen, T.; Lellouche, J.-P.; Bruce, I. J. Multifunctional magnetite and silica−magnetite nanoparticles: Synthesis, surface activation and applications in life sciences. J. Magn. Magn. Mater. 2005, 293, 33−40. (5) Nel, A.; Madler, L.; Velegol, D.; Xia, T.; Hoek, E.; Somasundaran, P.; Klaessig, F.; Castranova, V.; Thompson, M. Understanding biophysicochemical interactions at the nano−bio interface. Nat. Mater. 2009, 8, 543−557. (6) Hunter, R. J. Zeta Potential in Colloid Science-Principles and Applications. Academic Press: London, 1981. (7) Delgado, A. V.; González-Cabellero, F; Hunter, R. J.; Koopal, L. K.; Lyklema, J. Measurement & Interpretation of Electrokinetic Phenomena. Pure Appl. Chem. 2005, 77, 1753−1805. (8) Tombacz, E.; Majzik, A.; Horvat, E. Magnetite in Aqueous Medium: Coating its Surface and Surface Coated with it. Rom. Rep. Phys. 2006, 58, 281−286. (9) Fannin, P. C.; Marin, C. N.; Malaescu, I.; Stefu, N. Microwave dielectric properties of magnetite colloidal particles in magnetic fluids. J. Phys.: Condens. Matter 2007, 19, 036104. (10) Furlong, D. N.; Freeman, P. A; Lau, A. C. M. The Adsorption of Soluble Silica at the Solid/Aqueous Solution Interfaces- I Leaching from Glass - An Electrokinetic Study. J. Colloid Interface Sci. 1981, 80, 20−31. (11) de Vicente, J.; Delgado, A. V.; Plaza, R. C.; Durán, J. D.; González-Caballero, F. Stability of Cobalt Ferrite Colloidal Particles. Effect of pH and Applied Magnetic Fields. Langmuir 2000, 16, 7954− 7961. (12) Safran, S. A. Ferrofluids: Magnetic strings and networks. Nat. Mater. 2003, 2, 71−72. (13) Lyklema, J. Fundamentals of Interface and Colloid Science, Vol II, Solid-Liquid Interfaces; Academic Press: London, 1995. 7903

dx.doi.org/10.1021/la3010486 | Langmuir 2012, 28, 7897−7903