pubs.acs.org/Langmuir © 2010 American Chemical Society
Polymerization of Monomer-Based Ferrofluids Pei Bian and Thomas J. McCarthy* Polymer Science and Engineering Department, University of Massachusetts, Amherst, Massachusetts 01003 Received February 13, 2010. Revised Manuscript Received April 7, 2010 Ferrofluids based on oleate-stabilized nanoparticulate magnetite in cyclic olefin carrier solvents were prepared. Puddles of fluids containing 20-70 wt % magnetite in 1,5-cyclooctadiene or dicyclopentadiene can be moved or positioned on surfaces or in vessels with moderate magnetic fields (inexpensive NdFeB magnets). The surfaces of puddles with concentrations >40 wt % distort in response to magnetic fields applied perpendicular to the puddle surface and exhibit pointed liquid asperities. The liquid surface topography (asperity height and number per unit area) can be controlled by the strength of the applied field. The cyclic olefin carriers can be polymerized by ring-opening metathesis, solidifying the ferrofluid samples in the shape of the magnetically distorted monomeric solvent suspensions.
Magnetic liquids, generally called ferrofluids, are colloidal suspensions of ferromagnetic or ferrimagnetic nanoparticles stabilized by a surfactant that is appropriate for the particle and solvent.1-4 Water and saturated hydrocarbon liquids are most often employed as liquid carriers, magnetite (Fe3O4, which is ferrimagnetic) is the most common particle used, and oleic acid/ oleate is the most common surfactant. Figure 1 shows a schematic representation of a ferrofluid, indicating its important three components: carrier solvent, nanoparticles, and surfactant. Due to the small size of the magnetite particles (∼10 nm diameter), each has a single magnetic domain, they do not aggregate due to magnetic dipole interactions, and they do not settle or sediment in gravitational or magnetic fields. In the absence of a magnetic field, ferrofluid behavior is indistinguishable from that of normal liquids. They flow, exhibit surface tension, conform to the shapes of containers, rise or depress in capillary tubes, and form sessile drops, puddles, and capillary bridges. Their behavior differs in a magnetic field: the entire ferrofluid sample can respond as a ferromagnetic liquid and be held in a particular location or moved, the viscosity can increase orders of magnitude (magnetoviscous effect), and if the concentration of magnetic nanoparticles is sufficiently high, the liquid surface can distort to align with the magnetic field lines. That fluids exhibiting normal liquid behavior can be controlled, in terms of both viscosity and flow, with moderate magnetic fields has led to both fundamental and applied interest in ferrofluids. The first ferrofluid-based commercial product, which appeared shortly after the discovery of ferrofluids in the 1960s, was a rotary shaft seal that was capable of operating under either pressure or vacuum.5-7 The liquid seal is held in place between the shaft and housing by appropriate poling of a permanent magnet that is part of the housing. There have been *To whom correspondence should be addressed. E-mail: tmccarthy@ polysci.umass.edu.
(1) Raj, K.; Moskowitz, B.; Casciari, R. J. Magn. Magn. Mater. 1995, 149, 174. (2) Zahn, M. J. Nanopart. Res. 2001, 3, 73. (3) Odenbach, S. J. Phys.: Condens. Matter 2004, 16, R1135. (4) Laurent, S.; Forge, D.; Port, M.; Roch, A.; Robic, C.; Elst, L. V.; Muller, R. N. Chem. Rev. 2008, 108, 2064. (5) Papell, S. S. Low Viscosity Magnetic Fluid Obtained by the Colloidal Suspension of Magnetic Particles. U.S. Patent 3,215,572, 1963. (6) Papell, S. S. Liquid Storage Tank Venting Device for Zero Gravity Environment. U.S. Patent 3,295,545, 1967. (7) Kaiser, R. Ferrofluid Composition, Particles. U.S. Patent 3,700,595, 1970.
Langmuir 2010, 26(9), 6145–6148
numerous subsequent technological uses, but there are fields of research that have not yet exploited ferrofluids, and interface science is one of them. Perhaps because they are hidden from view in their common applications, they are not so well-known. It was estimated8 in 2006 that 500 million hard drives, 350 million loud speakers, 15 million DVD-ROM drives, and 200 000 rotary vacuum seals had been manufactured using ferrofluids. There are recent reports of the use of ferrofluids in magnetic ink,9 micropumps,10 medicine,11 and art.12 Reference 4 reviews many applications of superparamagnetic magnetite particles. We report the preparation of magnetite ferrofluids using cyclic olefins as carrier liquids and the polymerization of these monomeric olefins using ring-opening olefin metathesis catalysis. We demonstrate this process here using ferrofluid puddles with their surfaces distorted by magnetic fields. The resulting composite solids retain the shape of their field-distorted liquid surfaces. Magnetite nanoparticles were prepared using a modified literature procedure13 and characterized as summarized in Figures 2 and 3. Separate aqueous solutions of (a) a 1:2 molar mixture of FeCl2 and FeCl3 and (b) 1 M ammonium hydroxide were added dropwise into a magnetically stirred volume of water at rates to maintain a pH of 7. After the additions (over ∼10 min) were complete, the resulting suspension was stirred an additional 30 min, sodium oleate was added (0.2 g per gram of Fe3O4), and the mixture was heated with stirring at 80 °C for 30 min. After cooling to room temperature, 1 M HCl was added to reduce the pH to 5 and cause aggregation of the particles. The black solid was isolated by filtration, washed with copious deionized water, and then dried at reduced pressure overnight. The dry solid magnetite was easily dispersed in heptane, toluene, 1,5-cyclooctadiene (COD), or dicyclopentadiene (DCPD) with controllable concentration. Small amounts of aggregated magnetite remained and could be removed by allowing the aggregate to settle to the bottom of the vessel (either by gravity or by applying a magnetic field) and transferring the supernatant by pipet. Figure 2 shows a (8) Buschow, K. H. J. Handbook of Magnetic Materials; Elsevier: New York, 2006; Vol. 16. (9) Berger, P.; Adelman, N. B.; Beckman, K. J.; Campbell, D. J.; Ellis, A. B.; Lisensky, G. C. J. Chem. Educ. 1999, 76, 943. (10) Nguyen, N. T.; Chai, M. F. Micro Nanosyst. 2009, 1, 17. (11) Gupta, A. K.; Gupta, M. Biomaterials 2005, 26, 3995. (12) Kodama, S. Commun. ACM 2008, 51, 79. (13) Khalafalla, S. E.; Reimers, G. W. Sep. Sci. 1973, 8, 161.
Published on Web 04/13/2010
DOI: 10.1021/la1006617
6145
Letter
Bian and McCarthy
Figure 1. Schematic representation of a ferrofluid, indicating the three critical components: carrier solvent, superparamagnetic nanoparticles, and surfactant.
Figure 3. X-ray diffraction data and SQUID analysis of the oleate-stabilized magnetite particles.
Figure 4. Saturation magnetization values of the oleate-stabilized magnetite particles dispersed in COD as a function of weight percent magnetite. Figure 2. TEM micrograph of the oleate-stabilized magnetite particles and the particle size distribution.
transmission electron micrograph14 of a sample of oleate-stabilized magnetite that was dispersed in heptane, applied as a drop on a transmission electron microscopy (TEM) grid and allowed to dry. Also shown in Figure 2 is the distribution of particle size for this sample, determined by measuring 100 particles. The mean particle size is 7.4 nm, and the standard deviation is 2.9 nm. The TEM image shows that the particles, in general, do not contact one another but are separated by the oleate surfactant. Figure 3 shows X-ray diffraction15 and superconducting quantum interference device (SQUID)16 data that confirm that the particles are magnetite (Fe3O4)17 and that they are superparamagnetic. Minimal magnetic remanence is observed, and the saturation magnetization value is 50.1 ( 0.1 emu/g.18 Bulk magnetite exhibits a (14) TEM data were obtained on a JEOL JEM2010FX transmission electron microscope. (15) X-ray data were obtained using a PANalytical X’Pert PRO X-ray diffractometer. (16) Magnetic measurements were performed using a Quantum Design, MPMS XL magnetometer at 300 K. (17) The crystal was identified based on the Powder Diffraction File (PDF# 190629) from the International Centre for Diffraction Data (ICDD). (18) The SQUID data are very precise ((0.01 emu), and the source of error is in the measurement of mass. The data in Figure 3 are plotted as emu/g and Oe (1 emu/ g=1 Am2/kg, 1 Oe=79.6 A/m).
6146 DOI: 10.1021/la1006617
saturation magnetization of 92 emu/g;19 the lower value for nanoparticulate Fe3O4 has been attributed to defects due to the large surface area.20 The SQUID data shown are for a sample of magnetite that had not been dispersed with oleate. Scherer analysis21 of the 311 peak indicates a mean crystal size of 7.8 nm. A series of suspensions of oleate-stabilized magnetite nanoparticles was prepared in COD with oleate-stabilized magnetite varying from 20 to 80 wt % (5-38 vol %). The saturation magnetization of this series of samples is shown in Figure 4. The saturation magnetization of the solid oleate-stabilized magnetite (solid) was measured as 48.0 ( 0.1 emu/g, less than the 50.1 emu/g measured for the neat magnetite nanoparticles. We estimate that the stabilized nanoparticles are ∼4 wt % oleate by this difference; an excess of oleate is used in the preparation and is removed during the rinsing procedure. The saturation magnetization increases approximately linearly from 20 to 60 wt % and then increases more sharply with concentration. The 80 wt % sample is solid, and the 70 wt % sample, although apparently liquid, likely contains dispersed aggregated solid. Figure 5 shows frames of a video of a ∼25 wt % COD-based ferrofluid in a 20 mL vial being (19) Liu, J.; Wei, J; Li, S. Mater. Lett. 2007, 61, 1529. (20) Daou, T. J.; Greneche, J. M.; Pourroy, G.; Buathong, S.; Derory, A.; Ulhaq-Bouillet, C.; Donnio, B.; Guillon, D.; Begin-Colin, S. Chem. Mater. 2008, 20, 5869. (21) Hua, R.; Zang, C.; Shao, C.; Xie, D.; Shi, C. Nanotechnology 2003, 14, 588.
Langmuir 2010, 26(9), 6145–6148
Bian and McCarthy
Letter
Figure 7. Photographs of olefin metathesis polymerized ∼50 wt % COD-based (a) and ∼50 wt % DCPD-based (b) ferrofluid samples. Polymerizations were carried out in ∼0.03 T fields. The diameter of the watch glasses is 65 mm ,and the height of the tallest asperity in (a) is ∼9 mm.
Figure 5. Frames from a video of a ∼25 wt % COD-based ferrofluid in a 20 mL vial being manipulated with a ∼0.55 T magnetic field.
Figure 6. Puddle of ∼50 wt % COD-based ferrofluid in a Teflon watch glass being exposed to different strength magnetic fields.
manipulated with a ∼0.55 T magnetic field produced with a 25 25 25 mm3 NdFeB magnet wrapped in aluminum foil. The liquid flows up through a capillary bridge until the bridge ruptures. Many other obvious variations of liquid positioning (not shown here) were demonstrated with COD-based fluids. Their behavior is indistinguishable from those of aqueous or conventional hydrocarbon-based ferrofluids. COD- and DCPD-based ferrofluids containing greater than ∼40 wt % magnetite exhibit surface distortion in response to magnetic fields. The magnetic field force overcomes the surface Langmuir 2010, 26(9), 6145–6148
tension, and spike-shaped liquid asperities form. Figure 6 shows photographs of an (initially) 50 wt % COD-based ferrofluid puddle in a watch glass exposed to different strength magnetic fields. The field strength was adjusted for these photographs by placing different numbers of pads of paper between the magnet and watch glass. The conventional looking puddle (Figure 6a) becomes significantly distorted upon application of a weak (0.03 T, 6 pads of paper) field (Figure 6b). Increasing the field (by removing pads) to 0.18 and 0.23 T (Figure 6c,d) increases the number and decreases the size of the asperities. This trend is observed in conventional ferrofluids, and although initially counterintuitive22 it is rationalized when hydrostatic, magnetic, and surface energies are taken into account together. At a field of 0.55 T, the asperities disappear and the ferrofluid is attracted to the magnet as a contracted puddle (Figure 6e). Figure 6f shows the same ferrofluid sample exposed to a field parallel with the puddle surface. We note that these experiments and the ones described in Figure 7 are rather crude demonstrations done in open air in a rapidly flowing fume hood so that photographs would be lucid. COD and DCPD evaporate (and smell horrible), and the magnetite aggregates at the liquid-vapor interface and the puddle contact lines. The fluids exhibit these behaviors in closed vessels, but they are stable (as shown in Figure 5). COD and DCPD are monomers that very rapidly undergo ring-opening polymerization with numerous olefin metathesis catalysts.23 This monomer pair was chosen because COD forms a linear polymer, DCPD forms a cross-linked polymer, and mixtures of the two can be used to control the cross-link density of solids. Because of its relative air stability, we used the ruthenium-based initiator24 commonly called “Grubbs II” to polymerize COD and DCPD while puddles of ferrofluids were distorted in magnetic fields. Placing small drops of a dichloromethane solution of Grubbs II at four points of the puddle contact line initiated polymerization. The magnetic fields were removed after several hours. No attempt was made to promote a rapid reaction in these experiments, and a minute amount of initiator was added. One of these materials (DCPD) can be polymerized essentially instantly and is the basis of a commercial reaction injection molding (RIM) process.25 The samples shown in Figure 7 are solid objects and are in the absence of magnetic fields. Studies directed at analysis of the composition and orientation of (22) (23) (24) (25)
Lange, A.; Richter, R.; Tobiska, L. Ges. Angew. Math. Mech. 2007, 30, 171. Grubbs, R. H. Tetrahedron 2004, 60, 7117. Scholl, M.; Ding, S.; Lee, C. W.; Grubbs, R. H. Org. Lett. 1999, 1, 953. Chowdhury, B. J. Therm. Anal. 1997, 49, 325.
DOI: 10.1021/la1006617
6147
Letter
particles in the asperities to assess gradient/anisotropic composition and structure are in progress. We note the report26 of the curing of a silicone elastomer containing 3-5 μm Fe or 4-7 μm Ni particles in a magnetic field to prepare composites that exhibit magnetorestricition. We also note that others have incorporated superparamagnetic particles into materials including liposomes for MRI contrast,27 magneticfield-sensitive polymer gels,28,29 and silica.30,31 We point out that, for the systems reported here, the surfactant (oleate) contains a cis (26) Martin, J. E.; Anderson, R. A.; Read, D.; Gulley, G. Phys. Rev. E 2006, 74, 051507. (27) Martina, M. S.; Fortin, J. P.; Menager, C.; Clement, O.; Barratt, G.; Grabielle-Madelmont, C.; Gazeau, F.; Cabuil, V.; Lesieur, S. J. Am. Chem. Soc. 2005, 127, 10676. (28) Szabo, D.; Szeghy, G.; Zrinyi, M. Macromolecules 1998, 31, 6541. (29) Mayer, C. R.; Cabuil, V.; Lalot, T.; Thouvenot, R. Angew. Chem. 1999, 38, 3672. (30) Estournes, C.; Lutz, T.; Happich, J.; Quaranta, T.; Wissler, P.; Guille, J. L. J. Magn. Magn. Mater. 1997, 173, 83. (31) Bentivegna, F.; Ferre, J.; Nyvlt, M.; Jamet, J. P.; Imhoff, D.; Canva, M.; Brun, A.; Veillet, P.; Visnovsky, S.; Chaput, F.; Boilo, J. P. J. Appl. Phys. 1998, 83, 7776.
6148 DOI: 10.1021/la1006617
Bian and McCarthy
double bond and could function as a chain transfer agent in the polymerization. We did not investigate this. To summarize, we emphasize four points: (1) Ferrofluids can be prepared using polymerizable molecules (monomers). (2) The fluids can be easily manipulated (moved, positioned, injected into complex shapes) in the same way that conventional ferrofluids can. (3) These ferrofluid surfaces (liquid-air interfaces) can be distorted using perpendicular magnetic fields, and the surface area and surface structure can be controlled by applied field strength. (4) Two examples of conventional monomers, COD and DCPD, can be polymerized while functioning as a ferrofluid carrier liquid to form solid superparamagnetic magnetite/polymer composites in the shape or position determined by an external magnetic field. Acknowledgment. We thank the Centers for Materials Research Science and Engineering (DMR-0213695) and Hierarchical Manufacturing (CMMI-0531171) at the University of Massachusetts for support.
Langmuir 2010, 26(9), 6145–6148
