Preparation and Sedimentation Behavior in Magnetic Fields of

4410

Langmuir 2005, 21, 4410-4419

Preparation and Sedimentation Behavior in Magnetic Fields of Magnetite-Covered Clay Particles C. Galindo-Gonza´lez,† J. de Vicente,† M. M. Ramos-Tejada,‡ M. T. Lo´pez-Lo´pez,† F. Gonza´lez-Caballero,† and J. D. G. Durań*,† Department of Applied Physics, Faculty of Sciences, University of Granada, 18071 Granada, Spain, and Department of Physics. Faculty of Experimental Sciences, University of Jae´ n, 23071 Jae´ n, Spain Received October 21, 2004. In Final Form: February 3, 2005 This work is devoted to the preparation of magnetite-covered clay particles in aqueous medium. For this purpose, magnetite nanoparticles were synthesized by a coprecipitation method. These magnetic particles are adhered to sodium montmorillonite (NaMt) particles in aqueous suspensions of both materials, by appropriate control of the electrolyte concentrations. The best condition to produce such heteroaggregation corresponds to acid pH and ≈1 mol/L ionic strength, when the electrokinetic potentials (ζ-potential) of both NaMt and Fe3O4 particles have high enough and opposite sign, as demonstrated from electrophoresis measurements. When a layer of magnetite re-covers the clay particles, the application of an external magnetic field induces a magnetic moment in clay-magnetite particles parallel to the external magnetic flux density. The sedimentation behavior of such magnetic particles is studied in the absence or presence of an external magnetic field in a vertical direction. The whole sedimentation behavior is also strongly affected by the formation of big flocculi in the suspensions under the action of internal colloidal interactions. van der Waals and dipole-dipole magnetic attractions between magnetite-covered clay particles dominate the flocculation processes. The different relative orientation of the clay-magnetite particles (edge-to-edge, face-to-edge, and face-to-face) are discussed in order to predict the most favored flocculi configuration.

Introduction The study of colloidal magnetic fluids has an increasing interest in different technological areas, from mechanical antivibration systems to biomedical applications.1-3 These fluids are suspensions of ferro- or ferrimagnetic particles dispersed in aqueous or nonaqueous carrier liquids. Their main property consists of their field-responsive behavior under the action of an external magnetic field. This is because their internal structure can be modified by the formation of chainlike aggregates parallel to the field lines. Two kinds of magnetic fluids can be distinguished depending on the size of the magnetic particles: (i) ferrofluids (FRs), and (ii) magnetorheological fluids (MRFs). On one hand, FRs are composed of magnetic single domain nanoparticles (diameter ≈10 nm). Hence, they possess a magnetic dipole moment even in the absence of an applied magnetic field. From a rheological point of view, stable FRs exhibit a modest magnetoviscous effect; viscosity only changes by a factor of 2 or so under the application of the field. On the other hand, MRFs usually contain polydomain micrometer-sized particles. The application of an external magnetic field results in strong magnetic interactions leading to chaining. As a consequence, MRFs exhibit a fluid-solid transition characterized by a high yield stress (magnetorheological effect).4,5 * Corresponding author: J. D. G. Durań, Departamento de Fı´sica Aplicada, Facultad de Ciencias, Universidad de Granada. Tel: +34 958246103. Fax: +34 958243214. E-mail: [email protected]. † University of Granada. ‡ University of Jae ´ n. (1) Phule´, P. P.; Ginder, J. M. MRS Bull. 1998, 23, 19. (2) Ginder, J. M. MRS Bull. 1998, 23, 26. (3) Alexiou, Ch.; Schmid, R.; Jurgons, R.; Bergemann, Ch.; Parak, F. G.; Arnold, W. In Ferrofluids, Odenbach, S., Ed.; Springer: Berlin, 2002; p 233. (4) Charles, S. W. In Ferrofluids; Odenbach, S., Ed.; Springer: Berlin, 2002; p 3. (5) Bossis, G.; Volkova, O.; Lacis, S.; Meunier, A. In Ferrofluids; Odenbach, S., Ed.; Springer: Berlin, 2002; p 202.

Practical applications of magnetic colloids are strongly demanding. In the absence of an external field, a low viscosity as well as long term kinetic stability is commonly required for a correct operation. Many times, it is not possible to achieve these requirements and novel magnetic colloids are formulated; one possibility is to combine two different materials in the suspension. This practice has been well documented in the literature, and inverse ferrofluids are typical examples. They consist of nonmagnetic micrometer-sized particles dispersed in a ferrofluid. These nonmagnetic particles displace a magnetized ferrofluid, thereby creating magnetic holes with effective moment equal in size but opposite in direction to the total moment of the displaced ferrofluid. However, inverse ferrofluids have not found successful industrial applications because of their low magnetic strength. Actually, they are frequently used as model fluids in the laboratory. Different materials have been used as nonmagnetic particles in inverse ferrofluids: from the pioneering work with monodisperse polystyrene latex spheres by Skjeltorp6 to more recent works with hollow glass beads7 or monodisperse silica spheres and gibbsite platelets.8 In these works, the carrier ferrofluids used were organic liquids in which the magnetic nanoparticles (magnetite and cobalt ferrite) are covered with a surfactant to prevent aggregation by van der Waals forces. Colloidal clays are another interesting group of materials as a consequence of their ability to form gels. Actually, hydrophobic organoclays have been used as antisettling agents, which provides for a soft sediment once the magnetic particles settle out. This is of especial interest in order to minimize wear between moving surfaces.9 (6) Skjeltorp, A. T. Phys. Rev. Lett. 1983, 51, 2306. (7) Popplewell, J.; Rosenweig, R. E. J. Phys. D 1996, 29, 2297. (8) Rasa, M.; Philipse, A. P.; Jamon, A. Phys. Rev. E 2003, 68, 031402. (9) Munõz, B. C.; Adams, G. W.; Ngo, V. T.; Kitchin, J. R. United States Patent, 2001, 6,203,717 B1.

10.1021/la047393q CCC: $30.25 © 2005 American Chemical Society Published on Web 03/23/2005

Magnetite-Covered Clay Particles

Interestingly, Cousin and co-workers10 used binary mixtures of Laponite and citrate-coated maghemite particles to test the microrheology of the clay gel. Under wellcontrolled chemical conditions, the inclusion of maghemite particles did not modify the structure of Laponite suspensions which exhibit a fluid-solid transition at a given volume fraction. The aim of this work is to design magnetic particles composed by a clay core covered by a shell of magnetite nanoparticles starting from an aqueous suspension. If these composite particles are obtained, their sedimentation in aqueous media can be controlled by a variety of external and internal forces. Thus, the settling behavior will depend on the strength of the external magnetic field. Furthermore, the internal colloidal interactions between particles (electrostatic, van der Waals, and hydration or acid-base) will determine the different aggregation processes in the suspension (homo- and heterocoagulation between colloidal particles, and flocculation between claymagnetite particles) and, in consequence, the whole sedimentation velocity. Some recent works have been devoted to study the sedimentation behavior of monodisperse magnetic particles (cobalt ferrite,11 magnetite/ silica core-shell12,13); however, binary aggregates have been little studied in the literature. To our knowledge, only in a recent work14 has the heterocoagulation process between montmorillonite and micrometer-sized magnetite particles been extensively studied, demonstrating the “counterintuitive” phenomenon that the heterocoagulation of oppositely charged particles takes place only above a certain ionic strength (about 8 mmol/L NaCl). The action of short-range repulsive hydration forces (non-DLVO forces) is claimed there as the probable explanation for such unexpected behavior. As mentioned above, these forces will also be considered in the calculation of the interparticle energy balance for explaining the sedimentation behavior in this work. Experimental Section Materials. Clay Particles. The clay particles used in this work were obtained from a natural bentonite from Almerıá (Spain). First, the size fraction below ≈2 µm was separated by sedimentation from a water suspension of ground material. The following step was the homoionization of the particles, i.e., the substitution of all exchangeable cations by only one species, Na+ in our case, to obtain sodium montmorillonite (NaMt). The homoionization procedure, specific surface area (54.1 m2/g), surface and bulk chemical composition, and X-ray diffraction pattern were described elsewhere.15-17 Let us mention here that only Na+ was detected in a significant amount as exchangeable cation in the final product. Magnetite Particles. Magnetite nanoparticles were prepared following the coprecipitation method described by Massart.18 The synthesis consists of the following steps: (i) 40 mL of 1 mol/L FeCl3 solution was mixed with 10 mL of 2 mol/L FeCl2 and 2 mol/L HCl solution; (ii) this initial solution was slowly added, (10) Cousin, F.; Cabuil, V.; Levitz, P. Langmuir 2002, 18, 14661473. (11) de Vicente, J.; V. Delgado, A.; Plaza, R. C.; Durań, J. D. G.; Gonza´lez-Caballero, F. Langmuir 2000, 16, 7954. (12) Donselaar, L.; Philipse, A. P.; Suurmond, J. Langmuir 1997, 13, 6018. (13) Donselaar, L. N.; Philipse, A. P. J. Colloid Interface Sci. 1999, 212, 14. (14) Tombaćz, E.; Csanaky, C.; Ille´s, E. Colloid Polym. Sci. 2001, 279, 484. (15) Durań, J. D. G.; Ramos-Tejada, M. M.; Arroyo, F. J.; Gonza´lezCaballero, F. J. Colloid Interface Sci. 2000, 229, 107. (16) Ramos-Tejada, M. M.; de Vicente, J.; Ontiveros, A.; Durań, J. D. G. J. Rheol. 2001,45, 1159. (17) Lo´pez-Durań, J. D. G.; Khaldoun, A.; Kerkeb, M. L.; RamosTejada, M. M.; Gonza´lez-Caballero, F. Clays Clay Miner. 2003, 51, 65. (18) Massart, R. IEEE Trans. Magn. 1981, 17, 1247.

Langmuir, Vol. 21, No. 10, 2005 4411 under continuous mechanical stirring, to 500 mL of 0.7 mol/L NH3 solution; (iii) stirring was maintained during 20 min; (iv) the obtained suspension was dried in a convection oven at 35 °C in order to eliminate the remaining NH3. The resulting powder was repeatedly cleaned for undesired material (nonmagnetic particles and electrolytes) by magnetic sedimentation. The material not adhering to the flask wall in contact with the permanent magnet (B ) 0.42 T) was pipetted off, and substituted by ethanol. Water was not used in this step mainly because aging of the magnetite surface in water solutions can produce maghemite (γ-Fe2O3), modifying the isoelectric point (see below) of the particles.19 The process was finished after 15 cleaning steps with ethanol. The specific surface area of the magnetite particles was measured by N2 adsorption using the BET multipoint method in a Quantasorb Jr. apparatus (Quantachrome, USA). The X-ray diffraction pattern was obtained using the Debye-Scherrer method in a Phillips PW1710 powder diffractometer (The Netherlands). The magnetization, M, of the clay/magnetite aggregates and the magnetite powder was measured, at room temperature (293.0 ( 0.2 K), as a function of the applied magnetic field, H0, in a Manics DSM-8 magnetosusceptometer (France). The texture and shape of montmorillonite particles were investigated by a scanning electron microscopy (SEM) technique (Zeiss DSM 950, Germany). In the case of magnetite, the average diameter was estimated from high-resolution transmission electron microscopy (HREM, Philips CM20) pictures. Clay-Magnetite Suspensions. Different diluted suspensions with variable magnetite/clay mass ratio, and constant ionic strength (2 × 10-3 mol/L NaNO3), were prepared to study their sedimentation behavior. The dispersions labeled “NaMt” and “Fe3O4” only contain clay or magnetite as solid phase, and they have the same solid concentration (0.025% w/V). The montmorillonite concentration was maintained at 0025% w/V in all the mixed dispersions, which were prepared by mixing two magnetite and clay stock suspensions in order to reach a magnetite/clay mass ratio (mM/mC) ranging between 0.33 and 3. The pH of the suspensions was adjusted to the desired value immediately after preparation, using NaOH or HNO3 solutions. All chemicals used in the preparation of the solutions and suspensions were of analytical quality and were supplied by Panreac (Spain) or Sigma-Aldrich (Germany). Deionized and filtered Milli-Q water (Milli-Q Academic, Millipore, France) was always used. Methods. Electrophoresis. Electrophoretic mobility (ue) measurements were performed in a Malvern Zetasizer 2000 apparatus (Malvern Instruments, England) at 25.0 ( 0.5 °C, using suspensions with magnetite/clay mass ratio equal to that employed in sedimentation experiments, but diluted 1:10 in solids content. Electrophoretic measurements were carried out 24 h after preparation of the suspensions, and the pH was readjusted immediately before measuring the mobility. The estimation of the zeta potential, ζ, from ue measurements will be detailed below. Surface Free Energy. The surface free energy components of the clay and magnetite particles were estimated, from contact angle measurements, on the basis of van Oss et al.’s thermodynamic approach.20 The sample preparation was detailed in previous works (refs 15 and 17 for clay; and ref 21 for magnetite). The method essentially consisted in the deposition of a dry uniform layer of the powder (clay or magnetite) on a glass slide. Then, the contact angles formed by drops of water, formamide, and diiodomethane on the clay and magnetite layers were measured using a Rame´-Hart 100-07-00 (USA) telegoniometer. A brief description of the basic equations needed to calculate the surface free energy components of the materials can be seen in refs 15 and 17. Sedimentation. The sedimentation behavior of suspensions with different magnetite/clay mass ratios was inferred from optical absorbance measurements as a function of time. A Milton (19) Plaza, R. C.; Arias, J.; Espıń, M.; Jimeńez, M. L.; Delgado, A. V. J. Colloid Interface Sci. 2002, 245, 86. (20) van Oss, C. J. Interfacial Forces in Aqueous Media; Marcel Dekker: New York, 1980. (21) Viota, J. L.; de Vicente, J.; Tamos-Tejada, M. M.; Durań, J. D. G. Rheol. Acta 2004, 43, 645.

4412

Langmuir, Vol. 21, No. 10, 2005

Galindo-Gonza´ lez et al.

Figure 2. Magnetization curve of synthetic magnetite nanoparticles and magnetite-covered clay particles (magnetite/clay mass ratio 1:1).

Figure 1. (a) SEM picture of sodium montmorillonite particles, bar length 5 µm. (b) High-resolution TEM picture of magnetite particles, bar length 20 nm. Roy Spectronic 601 (USA) spectrophotometer was used. The center of the light beam (λ ) 550 nm) strikes a square cuvette containing the sample at 0.475 cm above its bottom. When the effect of the magnetic field was studied, a pair of coils (Phywe, Germany) was used. The characteristics of each coil were 600 turns, maximum current 1 A, resistance 1 Ω, and inductance 15 mH. The center of the cuvette was placed in the axial direction that passes through the center of the two coils. The magnetic flux density inside the cuvette position could be varied up to 15 mT and was measured with a Hall-effect teslameter (Phywe, Germany).

Results and Discussion Particle Characterization. As observed by scanning electron microscopy (SEM), the sodium montmorillonite consists of platelike particles with an edge length of about 1 µm and an aspect ratio ()particle equivalent diameter/ thickness) of about 14, as obtained from magnification of pictures such as in Figure 1a.

Magnetite particles were approximately spherical as observed by high-resolution transmission electron microscopy, and their average diameter, estimated from pictures such as that in Figure 1b, was 11.1 ( 2.0 nm. The specific surface area of the powder was 42.2 m2/g. The X-ray diffraction pattern (not showed here for the shake of brevity) demonstrated that the crystal structure of the powder corresponds to that of pure magnetite. Figure 2 shows the magnetization curve of the synthetic magnetite nanoparticles used in this work. As observed, a saturation magnetization MS ≈ 427 kA/m was obtained. This value is lower than that obtained in our laboratory for micrometer-sized synthetic magnetite (MS ) 570.7 kA/ m)22 and, as expected, larger than that of micrometersized cobalt ferrite particles (MS ) 270 kA/m).11 The value reported in the literature for pure magnetite is MS ) 510 kA/m.23 In Figure 2 is also represented the magnetization curve of the magnetite/clay aggregates obtained by heterocoagulation in suspensions at pH 3 (see below). A clear decrease of MS value is observed (MS,mix ≈ 286 kA/m) as compared to that of the magnetite in Figure 2. The saturation magnetization(MS,mix) of a mixture containing a volume fraction, φ, of magnetic particles with a saturation magnetization MS can be expressed as24 MS,mix ) φMS. From this expression, and taking into account that in the present case the volume fraction of magnetite nanoparticles in the clay-magnetite composite is φ ≈ 0.3, a saturation magnetization, MS,mix ≈ 127 kA/m, is expected. This value is considerably lower than that experimentally obtained, and it could be attributed to the formation of a multidomain composite, which confers to the suspensions more powerful magnetic properties than those corre(22) Viota, J. L.; de Vicente, J.; Durań, J. D. G.; Delgado, A. V. J. Colloid Interface Sci., in press. (23) Ashcroft, N. W.; Mermin, N. D. Solid State Physics; Saunders College: Philadelphia, 1976. (24) Rosensweig, R. E. Ferrohydrodinamics; Cambridge University Press: Cambridge, 1985.


Langmuir, Vol. 21, No. 10, 2005 4413

Figure 3. (a) Electrophoretic mobility (ue, left axis) and zeta potential (ζ, right axis) of sodium montmorillonite and magnetite particles as a function of the pH of the solution, at 2 × 10-3 mol/L NaNO3 ionic strength. (b) Electrophoretic mobility of NaMt and Fe3O4 particles as a function of NaNO3 molar concentration; pH ) 4.

sponding to a simple mixture, without aggregated particles. The formation of magnetite-covered clay particles will be checked by additional experiments (see below). For weak magnetic field strengths (H ) 0-50 kA/m) the magnetization curves (Figure 2) can be fitted to Rayleigh’s law by the expression25

M ) 9.03 + 9.93H - 0.097H2 (magnetite)

(1a)

M ) 38.80 + 5.33H - 0.048H2 (magnetite-clay) (1b) Electrophoretic Mobility. The electrophoretic mobility, ue, of diluted suspensions of clay and magnetite particles was also studied at different pH values and salt concentrations. In Figure 3a we represent ue as a function of pH at 2 × 10-3 mol/L NaNO3 ionic strength. Also, in Figure 3a the corresponding estimated values for the electrokinetic potential (ζ) are shown. The ζ-potential of the clay particles was obtained by means of the simple Smoluchowski’s formula,26 while the O’Brien and White27 method was used for magnetite because of its spherical shape. As can be seen in Figure 3a the ζ-potential of NaMt is negative in the whole pH range studied, and it is essentially pH-independent. This behavior has been repeatedly observed in bentonite clays,15,28,29 in which the electrokinetic properties are dominated by the constant negative charge of the faces (generated by cationic substitution in the crystal lattice). The electrochemical characteristics of the particle edges have a negligible effect on the ζ-potential of the particles because their surface area is very small as compared to that of the faces. Nevertheless, the influence of face-to-edge interactions is fundamental to explaining the flocculation and the rheological properties of NaMt suspensions.15,30 (25) Herpin, A. The´ ory du Magne´ tisme; Presses Universitaires de France: Paris, 1968. (26) Hunter, R. J. Foundations of Colloid Science; Oxford University Press: Oxford, 2001. (27) O’Brien, R. W.; White, L. R. J. Chem. Soc., Faraday Trans. 1978, 274, 1607. (28) Luckham, P. F.; Rossi, S. Adv. Colloid Interface Sci. 1999, 82, 43. (29) Sondi, I.; Pravdic´, V. in Interfacial Electrokinetics and Electrophoresis; Delgado, A. V., Ed.; Marcel Dekker: New York, 2002; p 773. (30) Ramos-Tejada, M. M.; Arroyo, F. J.; Perea, R.; Durań, J. D. G. J. Colloid Interface Sci. 2001, 235, 251.

On the other hand, the electrokinetic behavior of magnetite suspensions as a function of the pH of the solution (Figure 3a) shows the well-known role of H+ and OH- as potential-determining ions in metal oxide interfaces.31 The isoelectric point (iep; pH value for which ζ ) 0) is about pHiep ≈ 6.5, in good agreement with previous works dealing with micrometer-sized ferrites.11,32,33 Figure 3a suggests that if the pH of the aqueous solution is lower than the pHiep of magnetite, it would be possible to provoke a strong heteroaggregation between clay and magnetite particles. Such aggregation would be favored by the electrostatic attraction between particles with opposite zeta potential. Further evidence comes from Figure 3b, where ue is represented as a function of salt concentration (10-5, 10-1 mol/L) for a constant pH value (pH ) 4). Nevertheless, prior to any premature conclusion about the possible heteroaggregation between oppositely charged particles, we have to take into account the findings reported by Tombaćz et al.14 about the high stability observed at low ionic strength ( pHiep(magnetite) both kinds of particles are negatively charged and, as a consequence, one could expect for the electrophoretic mobility some value between clay and magnetite suspensions mobilities. However, in some (31) Matijevic´, E. In Interfacial Electrokinetics and Electrophoresis; Delgado, A. V., Ed.; Marcel Dekker: New York, 2002; p 199. (32) Regazzoni, A. E.; Matijevic´, E. Corrosion 1982, 38, 212. (33) Go´mez-Lopera, S. A.; Plaza, R. C.; Delgado, A. V. J. Colloid Interface Sci. 2001, 240, 40.

4414


Figure 4. (a) Electrophoretic mobility as a function of the pH of the solution for clay, magnetite, and clay-magnetite suspensions (obtained from suspensions with initial pH ) 3). The different magnetite/total solids mass ratio (mM/mC) is indicated on the solid lines. (b) Electrophoretic mobility vs pH for mM/mC ) 3 suspensions, obtained from stock suspensions with initial pH ) 3 or, alternatively, pH ) 10. Ionic strength: 2 × 10-3 mol/L NaNO3.

cases (see mM/mC < 2, Figure 4a) the resulting ue is found to be more negative than that corresponding to clay particles. This fact suggests that to some extent claymagnetite aggregates could persist when the pH of the suspensions is increased from the starting value (pH ) 3) up to basic pH. If it is actually the case and some magnetite-clay aggregates exist at neutral-basic pH, ue values should be larger than the ones corresponding to a suspension composed by individual clay and magnetite particles. This is because of the very high negative charge density that could bear the magnetite-clay complexes. In a previous work devoted to illite-iron oxide suspensions, Ohtsubo et al.34 demonstrated that the oxide particles that adhered at pH ) 3 were retained on the face surfaces of illite when the pH increased up to pH ) 9. We also observed a similar behavior with colloidal ZnS and glass surfaces.35 In this system, even under unfavorable circumstances (electrostatic repulsion), a significant (34) Ohtsubo, M.; Yoshimura, A.; Wada, S. I.; Young, R. N. Clays Clay Miner. 1991, 39, 347. (35) Durań, J. D. G.; Ontiveros, A.; Chibowski, E.; Gonza´lez-Caballero, F. J. Colloid Interface Sci. 1999, 214, 53.


amount of ZnS particles remained adhered on active patches of the glass substrate. In the classical work by O’Brien and White,27 and more recently by Ohshima,36 it was shown that the electrophoretic mobility of spherical or cylindrical particles tends to a nonzero constant value in the limit of very high zeta potential for values of κa g 10 (κ-1, thickness of the electric double layer; a particle radius). The mobility data for clay and clay-magnetite aggregates shown in Figure 4a, in the region of high pH values, could be considered as being in the vicinity of that limiting electrokinetic region and, therefore, affected by some degree of uncertainty. However, taking into account the parameters of our suspensions (particle radius a ≈ 500 nm; 2 × 10-3 mol/L NaNO3, temperature 298 K, κa ≈ 70), and that the maximum value for the measured electrophoretic mobility in Figure 4a is |ue| ≈ 4 (pH ) 10), two possible values for zeta potential are predicted:36 ζ1 ≈ 50 mV (practically the Smoluchowski’s value) and ζ2 ≈ 200 mV. Only if we would assume that the higher is the correct one, we would be into the limiting mobility region predicted for very high ζ-potential and the differences found in Figure 4a in the medium-high pH range could not be significant. This is not obviously the case because a potential as large as 200 mV is a very unrealistic situation for our magnetite-clay suspensions. Despite the previous arguments, our previous hypothesis on the persistence of magnetite-clay aggregates at neutral-basic pH can be reinforced by another electrophoresis experiment and electron microscopy observations. In Figure 4b we show ue vs pH measurements in suspensions with mass ratio mM/mC ) 3 prepared using two different procedures: (i) by mixing clay and magnetite at an initial pH ) 3; and (ii) at an initial pH ) 10. In both cases a series of samples with increasing (or decreasing) pH were prepared starting from the initial suspensions at pH 3 (or pH 10). As can be seen in Figure 4b, a clearly different behavior was observed. For pH < 7, suspensions prepared from an initial pH ) 3 gave mobility data closer to suspensions with 100% Fe3O4 than those prepared with an initial pH ) 10. This result demonstrates a more effective magnetite covering of clay particles in the first case (initial pH 3). On the other hand, for suspensions at pH > 7, |ue|(initial pH ) 3) > |ue|(initial pH ) 10). This is coherent with the persistence of magnetite-clay aggregates in the suspensions prepared from an initial pH ) 3. Additional proof of the previous arguments can be induced from electron microscopy pictures such as those shown in Figure 5. First, a suspension with mM/mC ) 3 was prepared by mixing magnetite and clay suspensions at pH ) 3. Then, the pH of the mixture was readjusted to pH ) 3 and a SEM picture of the particles was taken (Figure 5a). Finally, the pH of the suspension was further increased up to pH ) 10 and a new SEM picture was also taken (Figure 5b). At pH ) 3 clay particles appear to be completely coated by magnetite ones (Figure 5a). Even more, the magnetite particles seem to form big aggregates with a diameter much larger than the original synthesized particles (≈11 nm), confirming the existence of a magnetically multidomain coating, as we previously suggested from the saturation magnetization values shown in Figure 2. At pH ) 10 the coating is thinner and less homogeneous (Figure 5b). However, a partial retention of magnetite on clay particles is evident at pH ) 10. Surface Free Energy. Contact angles for water (12° ( 1°), formamide (10° ( 1°), and diiodomethane (14° ( 1°) on magnetite layers were used to calculate the surface (36) Ohshima, H. J. Colloid Interface Sci. 2003, 263, 337.


Langmuir, Vol. 21, No. 10, 2005 4415

suspensions of pure clay (labeled NaMt) and magnetite (labeled Fe3O4) are stable for a long time (for at least 50 min). In the case of the clay suspension, the formation of a stable gel by edge-to-face electrostatic and van der Waals attractions, which provoke a “house-of-cards” structure, justifies the absence of sedimentation in the suspension. This gel structure, in which the negative faces are adhered to the positive edges, avoids the sedimentation of the particles and confers a high yield stress to montmorillonite suspensions at acid pH, as demonstrated in previous works.15,30 On the other hand, to explain the absence of sedimentation in the magnetite suspensions, we must analyze the balance between the particle interactions by means of the extended DLVO theory of colloidal stability. This theory takes into account electrostatic (EL), Lifshitz-van der Waals (LW), and hydration or acid-base (AB) interactions.20,38 The electrostatic interaction between magnetite spheres (phase 1) dispersed in water solution (phase 3) is given by26

∆G131EL ) 2πr0aζ2 ln(1 + e-κh)

(2)

where a is the radius of the spherical particles, κ is the reciprocal Debye length, r is the dielectric constant of the medium, 0 is the permittivity of vacuum, ζ is the zeta potential, and h the distance between the particle surfaces. The LW attraction can be computed from39 Figure 5. SEM pictures of magnetite-covered clay particles (mass ratio mM/mC ) 3). (a) Suspensions at pH 3, obtained by mixing stock suspensions at an initial pH ) 3, bar length 300 nm. (b) Suspensions at pH 10, obtained from suspensions at initial pH ) 3, bar length 500 nm. Table 1. Surface Free Energy Components (in mJ/m2) of Magnetite and Sodium Montmorillonite (NaMt) Obtained from Contact Angle Measurementsa magnetite NaMt

γLW

γ+

γ-

49.3 ( 0.2 44.2 ( 0.4

0.17 ( 0.01 0.0 ( 0.1

55.4 ( 0.3 60.6 ( 0.5

γLW, Lifshitz-van der Waals; γ+, electron acceptor; γ-, electron donor. a

free energy components (see Table 1). The corresponding values for sodium montmorillonite were determined previously,15 and they are also included in Table 1. These values will be latter used to estimate the van der Waals and hydration (or Lewis acid-base in the van Oss et al. terminology20) interaction energies between magnetite and clay particles. Data in Table 1 demonstrate that both materials have a monopolar and electron-donor character (i.e., electron-acceptor parameter, γ+ ≈ 0; electron-donor parameter, γ- > 0). The high γ- values, being γ+ ≈ 0, are coherent with their hydrophilic nature, as frequently reported in the literature for clays and metal oxides.16,17,37 Sedimentation. Magnetite-covered clay particles with different relative contents of magnetite were prepared by mixing suspensions of magnetite and clay at pH ) 3. Then, the sedimentation behavior was studied by means of optical absorbance measurements. In Figure 6, the relative optical absorbance, A/A0 (A0 is the initial absorbance), is represented as a function of time in absence (Figure 6a) and presence of a magnetic field (Figure 6b). In Figure 6a we observe that both (37) Giese, R. F.; Wu, W.; van Oss, C. J. J. Dispersion Sci. Technol. 1996, 17, 527.

∆G131LW ) A131 h(4a + h) 2a2 2a2 + ln (3) 6 h(4a + h) (2a + h)2 (2a + h)2

[

]

where A131 is the well-known Hamaker constant. Its value can be estimated, according to van Oss et al.’s theory,20 from

A131 ) 24πh02(xγ1LW - xγ3LW)2

(4)

where γ1LW and γ3LW are the Lifshitz-van der Waals components of the surface free energy of the magnetite particles (see Table 1) and water (see ref 20), respectively. h0 is the so-called minimum equilibrium distance between interfaces, and the most reasonable value is h0 ) 1.58 ( 0.08 Å.20 Finally, the acid-base interaction can be calculated from20

(

∆G131AB ) ∆G131AB(h0)πaλ exp

)

h0 - h λ

(5)

where λ is the so-called water correlation distance. A value of λ ) 1 nm is usually assumed for hydrophilic particles dispersed in aqueous media.20,38 ∆G131AB(h0) is the acidbase interfacial energy at the equilibrium distance h0. This quantity is given, as a function of the electron-donor (γi-) and electron-acceptor (γi+) parameters of the surface tension of both materials (Table 1 for magnetite, Ref. 20 for water), by

∆G131AB(h0) ) -4(xγ1+ γ1- + xγ3+ γ3- -

xγ1+ γ3- - xγ1- γ3+)

(6)

(38) Israelachvili, J. Intermolecular and Surface Forces; Academic Press: New York, 1992. (39) Gregory, J. J. Colloid Interface Sci. 1981, 83, 138.

4416



Figure 6. Optical absorbance (expressed as A/A0; A0 is the initial absorbance) of clay (NaMt), magnetite, and clay/magnetite suspensions as a function of time, at the indicated magnetite/clay mass ratio: (a) in absence of magnetic field; (b) applying a magnetic field opposite to gravitation. Ionic strength 2 × 10-3 M NaNO3 and pH ) 3.

simplified as a clay-magnetite heterocoagulation followed by the sedimentation of the clay-magnetite aggregates generated in the suspensions. In Figure 5a we got a first insight into the formation of such aggregates, but a better explanation is required on the basis of the colloidal interactions that can favor the adhesion of magnetite particles onto the clay ones. To calculate the colloidal interactions (electrostatic, Lifshitz-van der Waals, and acid-base) between magnetite (phase 1) and clay particles (phase 2) immersed in aqueous solution (phase 3), we can consider a sphereplane configuration. For this geometry the EL interaction can be calculated from41 Figure 7. Total interaction potential (sum of electrostatic, van der Waals, and acid-base) between magnetite particles as a function of surface-to-surface distance for different pH values of the solution. Ionic strength 2 × 10-3 mol/L NaNO3.

∆G132EL ) πr0a(ζ1 + ζ2)2 ln(1 + e-κh) +

The total energy of interaction between magnetite spheres, ∆G131TOT, is the sum of the electrostatic, van der Waals, and acid-base interactions (eqs 2-6). In Figure 7 we show ∆G131TOT as a function of the surface-to-surface distance for different pH values. At pH ) 3 the potential energy barrier is high enough to prevent particle aggregation. This is due to the fact that the long-range electrostatic and short-range acid-base interactions counterbalance the LW attraction. Therefore, it is expected that the magnetite nanoparticles remain in suspension as individual entities. Furthermore, the thermal energy prevents both the sedimentation of the particles (diameter ≈ 11 nm) and the agglomeration by magnetic dipoledipole attraction, as usually occurs in ferrofluids when the LW attraction is surmounted by another repulsive force (either steric or electrostatic).40 At this moment, we can return to Figure 6a and continue with the discussion of the settling behavior of the claymagnetite suspensions. We observed that the addition of magnetite particles to the clay suspension provokes a higher sedimentation rate as the relative amount of magnetic particles increase. In general, two regions are observed in the sedimentation curves (this is more evident for a mass ratio mM/mC ) 3): an initial increase in the A/A0 vs time curve, followed by a region with a high (negative) slope. If coagulation takes place in a diluted suspension, it is well-known that the turbidity should increase with the time. Actually, this is observed in the first region of the curves in Figure 6a (especially for mM/ mC ) 3). The whole sedimentation behavior could be

where ζ1 and ζ2 are the zeta potentials of magnetite and clay particles, respectively. The LW interaction is given by42

(40) Odenbach, S. Colloids Surf., A 2003, 217, 171.

(ζ1 - ζ2)2 ln(1 - e-κh) (7)

∆G132LW )

[

]

A132 h + 2a 2a(a + h) ln 6 h h(h + 2a)

(8)

where the Hamaker constant A132 can be calculated from20

A132 ) -24πh02(xγ1LW γ3LW + xγ3LW γ2LW -

xγ1LW γ2LW - γ3LW)

(9)

Finally, the acid-base interaction can be calculated by means of20

(

∆G132AB ) ∆G132AB(h0)2πaλ exp

)

h0 - h λ

(10)

where

∆G132AB(h0) ) 2xγ3+(xγ1- + xγ2- - xγ3-) + xγ3+(xγ1+ + xγ2+ -

xγ3+) - xγ1+ γ2- - xγ1- γ2+

(11)

(41) Hogg, R.; Healy, T. W.; Fuerstenau, D. W. J Chem. Soc., Faraday Trans. 1966, 62, 1638. (42) Visser, J. In Surface and Colloid Science; Matijevic´, E., Ed.; Wiley: New York, 1976; Vol. 8, p 21.


Langmuir, Vol. 21, No. 10, 2005 4417

Figure 8. (a) Same as Figure 7, but now between magnetite and clay particles. (b) Total energy of interaction between magnetite and clay particles as a function of the pH at a surface-to-surface distance h ) 7 nm.

The total potential energy of interaction between magnetite and clay particles is represented in Figure 8a as a function of surface-to-surface distance for different pH values. From this plot, the total energy of interaction in the potential well at pH ) 5 (h ) 7 nm) is calculated and represented as a function of the pH of the suspension in Figure 8b. We can conclude that the heterocoagulation between magnetite and clay particles is favorable for pH < pHiep(magnetite), thus explaining the covering of clay particles by magnetite at pH ) 3, which was previously observed in SEM pictures (Figure 5a). The heavy weight of the generated magnetite-covered clay particles, and the rupture of the internal structure of the preexisting clay gel, explains the higher sedimentation rates observed in Figure 6a as the magnetite content is increased. The sedimentation behavior of the clay or magnetite particles and their mixtures in the presence of a vertical magnetic field (B ) 15 mT) is represented in Figure 6b. As expected, when a magnetic field is applied, clay suspensions remain stable (they are nonmagnetic). However, the absorbance for nanometric magnetite suspensions decreases with time. First, the magnetic moment of the particles is oriented in the magnetic field direction. Then they aggregate and, finally, migrate under the action of the gravitational force. On the other hand, binary mixtures sediment faster than individual magnetite particles (cf. Figure 6b). At short times, some sharp absorbance peaks are observed, which can be ascribed to fast heterocoagulation between clay and magnetite particles. Sharp peaks occurring at large time scales are ascribed to visible big flocculi passing through the light beam. A better understanding of the sedimentation behavior of magnetite-clay mixtures at pH ) 3 could be obtained from potential energy calculations in the presence of the magnetic field. Let us now extend the DLVO theory to include magnetic interactions between aggregates. If an isolated particle of relative magnetic permeability µp surrounded by a medium of relative permeability µm is placed in an external magnetic field, H B 0, the particle will acquire a magnetic moment43

µp - µm m b ) µ0Vp H B µm + N(µp - µm) 0

(12)

where µ0 is the permeability of vacuum and Vp is the particle volume. N is the demagnetizing factor, which takes into account that every magnet exists in a self-generated (43) Anderson, R. A.; Martıń, J. E. Am. J. Phys. 2002, 70, 11941204.

field that has a direction such as it tends to demagnetize the specimen. The value of N depends on both the particle shape and the relative orientation of the particle in the external field H B 0. Magnetite-covered clay particles will have an average relative magnetic permeability, µp, which depends on the shape of the particles. Considering the laminar shape of the montmorillonite particles, a prismatic geometry is expected for the clay-magnetite particles, which could be approximated to oblate spheroids in order to estimate µp by means of the Maxwell-Garnett theory for spheroids44

µp ) 1 + β)

φβ 1 - φNβ

(13)

µpm - 1

(14)

1 + N(µpm - 1)

where µpm is the magnetic permeability of the magnetite particles and φ the magnetite volume fraction. The value of µpm can obtained from the initial slope of the magnetization curve shown in Figure 2, giving µpm ) 10.93. This value is well into the range given in the literature for magnetite solid particles (5 e µpm e 35).45 The demagnetizing factor for an oblate spheroid, having two long axes k times the length of the axis of symmetry, and magnetized parallel to a long axis is given by46,47

N x ) Ny )

{

(

)

(k2 - 1)1/2 1 1 k2 arcsin - 2 2 [k2 - 1]3/2 k k -1

}

(15)

and for an oblate spheroid magnetized along its short axis

Nz ) 1 - 2Nx )

{

(

)}

(k2 - 1)1/2 k2 1 1 arcsin k k2 - 1 (k2 - 1)1/2

(16)

Considering the aspect ratio of the montmorillonite particles (see Experimental Section) and the density of both magnetite and clay (5.2 and 2.4 g/cm3, respectively) the volume of a clay particle is approximately 1000 nm × 1000 nm × 71.4 nm. However, a shell of magnetite nanoparticles covers each clay particle with a thickness (44) Garnett, J. C. M. Philos. Trans. R. Soc. London 1904, 203, 385. (45) Bothorz, R. M. Ferromagnetism; van Nostrand: Princeton, NJ, 1968. (46) O’Handley, R. O. Modern Magnetic Materials. Principles and Application; Wiley & Sons: New York, 2000. (47) Osborn, J. A. Phys. Rev. E 1945, 67, 351.

4418



that depends on the amount of magnetite added to the clay suspension. Assuming that all the magnetite particles are equally spread on the clay particle faces in a cubic close packing, we have estimated, for particles containing a magnetite/clay ratio mM/mC ) 3, a volume Vp ≈ 1000 nm × 1000 nm × 225 nm. Then, we can estimate the demagnetizing factors for clay-magnetite particles with an axes ratio k ) 1000/225, obtaining (eqs 15 and 16): Nx ) Ny ) 0.137, and Nz ) 0.725. From these values of the demagnetizing factors, the relative permeability of the clay-magnetite particles can be calculated (eqs 13 and 14), being µp,x ) µp,y ) 5.27, and µp,z ) 2.77. The corresponding magnetic moments are (eq 12): mx ) my ) 6.03 × 10-22 A‚m2 (particles magnetized parallel to long axes), and mz ) 1.73 × 10-22 A‚m2 (particles magnetized along the short axis). The magnetic interaction between magnetite-clay aggregates is strongly anisotropic. To a first approximation it can be estimated as the potential energy of interaction between two magnetic dipoles24

Edd )

[

b 1‚m b2 3 1 m - 5(m b 1‚r b)(m b 2‚r b) 3 4πµ0 r r

]

(17)

where m b 1 and m b 2 are the magnetic moments of two different magnetite-covered clay particles and r is the distance between the center of the particles. If we consider that both magnetic moments are parallel and oriented head-to-tail by the magnetic field, we obtain

Edd ) -

m1m2 2πµ0r3

(18)

where m1 and m2 have been calculated by means of eqs 12-16, r can be written as r ) r1 + r2 + h, being r1 and r2 the semiaxes of the particles and h is the surface-tosurface separation. The van der Waals, electrostatic and acid-base interactions can be estimated, for a plate-plate configuration, by means of a set of equations analogous to eqs 7-10 (see refs 20 and 41). Three possible orientations exist between the claymagnetite particles: edge-to-edge, face-to-edge, and faceto-face. In Figure 9 we represent the total energy of interaction between the clay-magnetite particles (mM/ mC ) 3) as a function of the surface-surface separation. Here, at short distances between the particles (∼10 nm), both the electrostatic and acid-base interactions are repulsive, while the van der Waals and magnetic ones are attractive. As observed in Figure 9, whatever the relative orientation between the particles, a potential well exists for h ≈ 10 nm that favors the aggregation in the suspension. Due to the larger surface area and the lower axis length, the most favorable aggregation process corresponds to the face-to-face orientation. Only in this case (Figure 9c), the magnetic interaction represents a significant contribution over the van der Waals attraction, which is the dominant interaction (for h ≈ 10 nm) in the other cases (edge-edge and face-edge, parts a and b of Figure 9, respectively). These aggregation processes explain the fast sedimentation of the magnetite-covered clay particles, deduced from absorbance measurements, in the presence of a magnetic force parallel to gravitation. Finally, let us to compare the sedimentation behavior of the clay-magnetite core-shell particles obtained in this work with a similar study carried out in our laboratory with ferrite aqueous suspensions.11 In that work, in the absence of external magnetic field, the stationary state in

Figure 9. Total interaction energy (sum of electrostatic, van der Waals, acid-base, and magnetic) between magnetitecovered clay particles, as a function of surface-to-surface distance, for the following relative orientations: (a) edge-toedge; (b) face-to-edge; and (c) face-to-face.

optical absorbance vs time curves for cobalt ferrite (diameter 850 nm) suspensions was reached for a time around 200 s, while for the magnetite-clay aggregates obtained in the present work the time needed (see Figure 6a) is larger than 1000 s. On the other hand, in the micrometer-sized cobalt ferrite suspension an external magnetic flux density B0 ≈ 2 mT was enough to observe a strong effect on the sedimentation rate. However, in the clay-magnetite aggregates a field of B0 ) 2 mT did not produce any significant effect, and only for B0 ≈ 10-15 mT perceptible changes in the sedimentation velocity was detected. These differences can be explained by the small density and the low magnetic response of the magnetiteclay suspensions. Therefore, we have a more stable (from the sedimentation point of view) MR-like fluid, although a less intense magnetorheological effect can be expected. The potential use of these composed particles to formulate new magnetorheological fluids will have to be checked in the future by measuring their magnetorheological behavior, particularly the increase in the yield stress and rigidity modulus under the action of strong magnetic fields.


Conclusions We have shown that it is possible to prepare aqueous suspensions of magnetic aggregates, composed by a core of clay and a shell of magnetite nanoparticles, by the adequate control of the electrical properties of the solid/ liquid interface. The composed particles show a magnetic saturation larger than that corresponding to a single mixture of clay and magnetite nanoparticles. This behavior allows considering the clay/magnetite particles as magnetic multidomain entities, in a similar way to that corresponding to micrometer-sized particles usually employed in magnetorheological fluids. The sedimentation rates of the heteroaggregates obtained in this work, under the action of an external

Langmuir, Vol. 21, No. 10, 2005 4419

magnetic field, are slower than those obtained in micrometer-sized ferrite suspensions. Therefore, we can consider these composed particles as a possible alternative for formulating magnetorheological fluids in which the settling and redispersion difficulties can be reduced, although it is expected that the magnetorheological effect will be less strong than in classical MR fluids. Acknowledgment. Financial support by Spanish Ministerio de Educacioń y Ciencia and European FEDER funds, under Project MAT-2004-00866, is gratefully acknowledged. LA047393Q

Preparation and Sedimentation Behavior in Magnetic Fields of

Recommend Documents