Diffusive Flux and Magnetic Manipulation of Nanoparticles through

Mar 17, 2010 - Analytical methods for separating and isolating magnetic nanoparticles. Jason R. Stephens , Jacob S. Beveridge , Mary Elizabeth William...
0 downloads 0 Views 4MB Size
Download PDF
Anal. Chem. 2010, 82, 3155–3160

Diffusive Flux and Magnetic Manipulation of Nanoparticles through Porous Membranes Jason R. Stephens, Jacob S. Beveridge, Andrew H. Latham, and Mary Elizabeth Williams* Department of Chemistry, 104 Chemistry Building, The Pennsylvania State University, University Park, Pennsylvania 16802 Measurement of transport of nanometer scale particles through porous media is important to begin to understand the potential environmental impacts of nanomaterials. Using a diffusion cell with two compartments separated by either a porous alumina or polycarbonate membrane as a model system, diffusive flux through mesoporous materials is examined. Experiments are performed as a function of particle size, pore diameter, and solvent, and the particle fluxes are monitored by the change in absorbance of the solution in the receiving cell. Using the measured extinction coefficient and change in absorbance of the solution as a function of time, the fluxes of 3, 8, and 14 nm diameter CoFe2O4 particles are determined as they are translocated across pores with diameters 30, 50, 100, and 200 nm in hexane and aqueous solutions. In general, flux decreases with increasing particle size and increases with pore diameter. We find that fluxes are faster in aqueous solutions than in hexane, which is attributed to the hydrophilic nature of the porous membranes and differences in wettability. The impact of an applied magnetic flux gradient, which induces magnetization and motion, on permeation is also examined. For larger membrane pore diameters, applied magnetic fluxes increase the rate of transport of 14 nm CoFe2O4 particles more than that of 3 or 8 nm diameter particles, reflecting their differences in susceptibility. However, larger particles are excluded from membranes with small diameter pores, consistent with magnetic interparticle attractions that reversibly induce magnetic aggregation. Burgeoning interest in the use of nanomaterials for applications ranging from drug delivery to odor control in socks has driven basic research into an increasing number of commercially available products.1 There is growing recognition that the novel electronic, catalytic, and magnetic properties that arise at the nanoscale could also potentially engender deleterious effects in the environment.2 Since greater amounts of nanomaterials are being released into waste streams, either directly or by leaching from solid waste dumps, significant concern exists regarding their * To whom correspondence should be addressed. E-mail: [email protected]. (1) (a) The Nanotech Report; http://www.luxresearchinc.com/tnr.php. (b) Ahmad, F. J.; Khar, R. K. Pharma Rev. 2005, 4, 49–56. 10.1021/ac901770k  2010 American Chemical Society Published on Web 03/17/2010

transport, reactivity, and fate.3 To understand this complex set of problems, a diverse set of analytical tools is required to study particles under a wide range of conditions.4 Our interest in chemical modification and transport of nanoparticles5,6 has led us to begin to consider their potential environmental impacts. Our focus initially is on studies that aim to understand the transport of particles as a function of their size through porous media in aqueous and nonaqueous solutions. Typical analyses of particle transport in porous media utilize columns containing packed beds with porosities that vary from the nano- to microscale.7 Porous alumina and polycarbonate membranes are model systems for studying the transport of nanoscale particles in complex media, such as soil, which contain mesoscale pores. Because the membranes typically have welldefined porosities, pore diameters, and pore lengths, these are used to understand the role of pore diameter on transport. While porous membranes have been used8 to investigate transport of molecules by diffusion, iontophoresis, or electrophoresis, to the (2) (a) Farre, M.; Gajda-Schrantz, K.; Kantiani, L.; Barcelo, D. Anal. Bioanal. Chem. 2009, 393, 81–95. (b) Dunphy Guzman, K. A.; Taylor, M. R.; Banfield, J. F. Environ. Sci. Technol. 2006, 40, 1401–1407. (c) Godwin, H. A.; Chopra, K.; Bradley, K. A.; Cohen, Y.; Herr Harthorn, B.; Holden, P.; Keller, A. A.; Lenihan, H. S.; Nisbet, R. M.; Nel, A. E. Environ. Sci. Technol. 2009, 43, 6453–6457. (3) For example, (a) Marquis, B. J.; Love, S. A.; Braun, K. L.; Haynes, C. L. Analyst 2009, 134, 425–439. (b) Benn, T. M.; Westerhoff, P. Environ. Sci. Technol. 2008, 42, 4133–4139. (c) Impellitteri, C. A.; Tolaymat, T. M.; Scheckel, K. G. J. Environ. Qual. 2009, 38, 1528–1530. (d) BystrzejewskaPiotrowska, G.; Golimowski, J.; Urban, P. L. Waste Manage. 2009, 29, 2587– 2595. (4) (a) Liu, J.-F.; Chao, J.-B.; Liu, R.; Tan, Z.-Q.; Yin, Y.-G.; Wu, Y.; Jiang, G.-B. Anal. Chem. 2009, 81, 6496–6502. (b) H., X.; Nie, H.; Wang, K.; Tan, W.; Wu, X.; Zhang, P. Anal. Chem. 2008, 80, 9597–9603. (c) Tiede, K.; Hassellov, M.; Breitbarth, E.; Chaudry, Q.; Boxall, A. B. A. J. Chromatogr., A 2009, 1216, 503–509. (d) Kohlbusch, T. A. J.; Fissan, H.; Asbach, C. Nanotechnology 2008, 2, 229–266. (e) Simonet, B. M.; Valcarcel, M. Anal. Bioanal. Chem. 2009, 393, 17–21. (5) (a) Thode, C. J.; Williams, M. E. Langmuir 2008, 24, 5988–5990. (b) Latham, A. H.; Williams, M. E. Langmuir 2006, 22, 4319–4326. (c) Fleming, D. A.; Thode, C. J.; Williams, M. E. Chem. Mater. 2006, 18, 2327–2334. (6) (a) Latham, A. H.; Williams, M. E. Acc. Chem. Res. 2008, 41, 411–420. (b) Latham, A. H.; Tarpara, A. P.; Williams, M. E. Anal. Chem. 2007, 79, 5746– 5752. (c) Latham, A. H.; Freitas, R.; Schiffer, P.; Williams, M. E. Anal. Chem. 2005, 77, 5055–5062. (7) (a) Phenrat, T.; Kim, H.-J.; Fagerlund, F.; Illangasekare, T.; Tilton, R. D.; Lowry, G. V. Environ. Sci. Technol. 2009, 43, 5079–5085. (b) Dunphy Guzman, K. A.; Finnegan, M. P.; Banfield, J. F. Environ. Sci. Technol. 2006, 40, 7688–7693. (c) He, F.; Zhang, M.; Qian, T.; D., Z. J. Colloid Interface Sci. 2009, 334, 96–102. (8) (a) Martin, C. R.; Nishizawa, M.; Jirage, K.; Kang, M.; Lee, S. B. Adv. Mater. 2001, 13, 1351–1362. (b) Lee, S. B.; Martin, C. R. Chem. Mater. 2001, 13, 3236–3244. (c) Scott, E. R.; White, H. S.; Phipps, J. B. Anal. Chem. 1993, 65, 1537–1545. (d) Ervin, E. N.; White, H. S.; Baker, L. A. Anal. Chem. 2005, 77, 5564–5569.

Analytical Chemistry, Vol. 82, No. 8, April 15, 2010

3155

best of our knowledge, these have not been employed as a straightforward means to study nanoparticle transport. In this paper, we describe the translocation of 3, 8, and 14 nm diameter CoFe2O4 nanoparticles across porous membranes with pore diameters ranging from 30 to 200 nm. Using a u-tube geometry diffusion cell, permeation in hexane and aqueous solutions is measured and the data used to obtain rates of flux across the membranes as a function of pore and particle dimensions. Since the CoFe2O4 nanoparticles are superparamagnetic,6a the impact of an applied magnetic flux gradient on the rate of translocation is also examined and compared to Au nanoparticles of similar size. These initial studies provide a straightforward means with which to understand particle permeation in confined geometries or complex media such as the environment. EXPERIMENTAL SECTION Materials. All chemicals were acquired from commercial sources and used as received without further purification. Au nanoparticles were prepared according to literature methods.9 Uncoated microporous alumina membranes (Whatman Anodisc, 60 µm thick) and track-etched polycarbonate membranes (SPIPore, 6 µm thick) of varying pore diameters and thicknesses are soaked for 15 min in solvent (hexane or water) prior to insertion in the experimental apparatus. Synthesis of Water-Soluble CoFe2O4 Nanoparticles. Synthesis of CoFe2O4 nanoparticles was accomplished by the thermal reduction of metal acetylacetonate (acac) complexes.10 Briefly, 0.353 g of Fe(acac)3 (1.0 mmol), 0.129 g of Co(acac)2 (0.5 mmol), 1.29 g of 1,2-hexadecanediol (5.0 mmol), 0.848 g of oleic acid (3.0 mmol), 0.803 g of oleylamine (3.0 mmol), and benzyl ether (10 mL) were combined under N2 and heated to 200 °C for 60-120 min, followed by refluxing for 60 min. Nanoparticles were purified by precipitation with ethanol, centrifuged, and redispersed in 2.0 mL of hexane with 15 µL of both oleic acid and oleylamine. This solution was centrifuged to remove any insoluble particulates and decanted. The soluble particles were then precipitated with ethanol, isolated by centrifugation, and finally redissolved in a minimal amount of hexane. Larger CoFe2O4 nanoparticles were synthesized by a seed-mediated route using 8 nm diameter CoFe2O4 particles as “seeds” for the growth of larger particles.10 Oleic acid- and oleyl amine-coated CoFe2O4 nanoparticles (120 mg) were dispersed in a 50/50 mixture of toluene and N,Ndimethylformamide (15 mL of total volume), to which 0.1 g of citric acid was added.11 The mixture was then stirred at 100 °C for 24 h. The particles were subsequently precipitated by the addition of ethyl ether (40 mL) and recovered with an electromagnet. The particles were redispersed in acetone and reprecipitated by centrifugation to remove all traces of free citric acid. The particles were then dried and redispersed in H2O. Instrumentation. Transmission electron microscopy (TEM) samples were prepared by slow evaporation of hexane or H2O solutions of nanoparticles directly onto a grid (300 mesh, (9) Keating, C. D.; Musick, M. D.; Keefe, M. H.; Natan, M. J. J. Chem. Educ. 1999, 76, 949–955. (10) Sun, S.; Zeng, H.; Robinson, D. B.; Raoux, S.; Rice, P. M.; Wang, S. X.; Lie, G. J. Am. Chem. Soc. 2004, 126, 273–279. (11) Lattuada, M.; Hatton, T. A. Langmuir 2007, 23, 2158–2168.

3156

Analytical Chemistry, Vol. 82, No. 8, April 15, 2010

carbon-coated Cu, EM Science). Images were obtained using a JEOL-1200EXII microscope operating at 80 keV equipped with a high resolution Tietz F224 digital camera. Particle sizes are reported as the mean ± the standard deviation, based on statistical analysis of at least 300 particles taken from different areas on the TEM grid. A Varian Cary 50 scan UV-visible absorption spectrometer was used to collect absorbance spectra. Magnetic fields were applied using 2 NdFeB permanent magnets (circumference, 2.8 cm; height, 2.5 cm; field strength; 0.122 T, Engineered Concepts, Birmingham, AL). Magnetic field measurements were obtained with a hand-held gauss meter (Magnetic Instrumentation, Inc.; model 907). Experimental Design. Our studies focus on quantifying magnetic field induced motion through alumina and polycarbonate mesoporous membranes that are mounted between two halves of a u-tube shaped diffusion cell (Figure S-1 in the Supporting Information). A 1.5 cm diameter O-ring held the membrane in place and defined the area (1.77 cm2) of the membrane exposed to the feed and permeant half-cell solutions. The effective surface area of the membrane is the product of the exposed area and the membrane porosity: the polycarbonate membrane porosities are 25, 40, and 40% for the 30, 50, and 100 nm diameter pores, and the alumina membrane porosities are 35 and 45% for the 100 and 200 nm diameter pores, respectively. The feed half cell was filled with 5 mL of a solution of 5 mg/ mL CoFe2O4 nanoparticles dissolved in either hexane or water; the permeant half was filled with 5 mL of pure hexane or water. Two types of permeation experiments, in the presence and absence of a magnetic (B) field, were done using the following experimental protocol: a 5 mL amount of pure solvent (hexane or water) was added to the receiving half cell and the permeant solution (5 mL of 5 mg/mL CoFe2O4 in hexane or water) was then added to the feed half cell. Prior to absorbance measurement, the feed and receiving reservoirs were mixed. Permeation data were collected by monitoring the absorbance of CoFe2O4 at 310 nm of an undiluted 0.75 mL aliquot pipetted from the permeant half -cell every 15 min for 3 h. The aliquot was returned to the half cell following the absorbance measurement. For measurements of magnet-induced transport, two NdFeB permanent magnets are placed 4 mm away from the membrane on the permeant half-cell side (see Figure S1 in the Supporting Information) on the cell exterior. Each experiment was repeated 3 times, and statistical significance was tested with a student t test. RESULTS AND DISCUSSION Magnetic Particle Samples. The magnetic forces acting on nanoparticle mixtures of varying size relies on differences in their magnetic moments and, to a lesser extent, their diameters.6,12 To test the ability to affect the permeation of nanometer-scale particles of varying sizes, we began by preparing monodisperse samples with known magnetic properties. High-temperature synthesis was used to generate a series of CoFe2O4 nanoparticles of varying diameter; the average sizes of the as-prepared particles were by analysis of transmission electron microscopy (TEM) images. Representative images of the samples used in this study are (12) Giddings, J. C.; Yang, F. J. F.; Myers, M. N. Science 1976, 193, 1244– 1245.

Figure 1. Transmission electron micrographs of (A) 3, (B) 8, and (C) 14 nm diameter CoFe2O4 nanoparticles. Scale bars are 50 nm. Insets: plots of size distributions determined by analysis of at least 300 particles per samples. (D) UV-visible absorbance spectra of solutions of (black) 3 nm diameter, (green) 8 nm diameter, and (red) 14 nm diameter CoFe2O4 nanoparticles at a 3 mg/mL concentration in hexane. Dashed line indicates detection wavelength in permeation experiments.

shown in Figure 1A-C and reveal that particles are spherical and highly uniform in size. Statistical analysis of the nanoparticles (histograms in the insets) shows that CoFe2O4 samples in Figure 1A-C have average diameters of 3.3 ± 0.4, 8.1 ± 0.8, and 13.9 ± 1.2, respectively. Because transport across the porous membrane is monitored using changes in solution absorbance, UV-visible absorbance spectra were acquired for each of the nanoparticle samples in hexane solutions. In Figure 1D, the spectra show that the particles are broad-wavelength absorbers and have high extinction in the UV region. To monitor changes in solution concentration of particles, a wavelength of 310 nm was selected (dashed line in Figure 1D) because all three samples absorb strongly at this wavelength. Extinction coefficients at 310 nm, determined from Beer’s law plots of these, are given in Table S1 in the Supporting Information. Diffusive Flux of Nanoparticles through Porous Membranes. A set of experiments were conducted using the u-tube geometry (Figure S1 in the Supporting Information) so that the experimentally controllable variables are the size of the pore and the nanoparticle sample. A microporous membrane of fixed diameter was first soaked in solvent to wet the interior walls of the pore for 15 min prior to being sandwiched between the two half cells. A solution of magnetic nanoparticles of fixed size was

placed in the feed half cell and pure solvent in the receiving half cell. The concentration gradient across the membrane induces mass transport by diffusion that is expected to depend on a number of factors including particle size, attached surface monolayer, density gradient, solution viscosity, etc. Although the compartments are unstirred, a small amount of convection may occur on the time scale of the experiment; however, it is most likely that diffusive permeation across the membrane is the dominant means of transport. To monitor mass transport, absorbance measurements were taken of aliquots of the receiving solution every 15 min; the aliquot was returned to the solution, and this process was repeated for 3 h. Nanoparticle flux across the porous membrane (6 and 60 µm thickness for polycarbonate and alumina, respectively) causes an increase in the absorbance of the receiving solution. For example, Figure 2 compares diffusive transport of 3 nm diameter CoFe2O4 particles in water and 8 nm CoFe2O4 diameter particles in hexane. Since absorbance is proportional to concentration, these data are indicative of increasing particle concentration in the receiving solution. In all cases, the rate of change in particle concentration in the receiving compartment depends on the size of the membrane pores. For example, in Figure 2A, the absorbance of 3 nm diameter CoFe2O4 nanoparticles in hexane increases as the polycarbonate pore diameter increases Analytical Chemistry, Vol. 82, No. 8, April 15, 2010

3157

Figure 2. Plot of absorbance at 310 nm vs time in the receiving cell using feed solutions containing 5 mg/mL of (A) 3 nm diameter CoFe2O4 nanoparticles in hexane and (B) 14 nm diameter CoFe2O4 in water with (black b) 200 nm and (red b) 100 nm diameter porous alumina membranes or (blue 9) 100 nm, (green 1) 50 nm, and (yellow 1) 30 nm diameter porous polycarbonate membranes. No magnetic field is applied.

from 30 to 100 nm; transport through alumina membranes with 200 nm diameter pores is more rapid than in alumina with 100 nm diameter pores. In Figure 2B, the analogous effect is observed for larger 14 nm diameter CoFe2O4 particles in water; identical trends are observed in both hexane and aqueous solutions. These observations are, therefore, not a result of solvent viscosity or differences in hydrophilicity. We note that the 10-fold difference in the membrane thickness for the alumina vs polycarbonate membranes precludes comparison of the slopes in Figure 2. Using measured extinction coefficients for each of the particles in hexane or aqueous solutions (Table S1 in the Supporting Information), the absorbance values were converted to particle mass. The small dispersities of the nanoparticle sizes give rise to an uncertainty in the molar mass of a particle sample. However, using the average particle diameter (determined by TEM) and unit cell, the molar masses can be approximated (Table S1 in the Supporting Information). Absorbance values were, therefore, converted to molar quantities of particles that are transported in these experiments; Figure 3 compares representative results of experiments using 30 nm diameter polycarbonate pores in water and 100 nm diameter alumina pores in hexane for CoFe2O4 nanoparticles of varying size (i.e., 3, 8, 14 nm diameter). Data for hexane-soluble 2.5 nm diameter Au clusters and 13 nm diameter citrate stabilized Au particles are also shown for 3158

Analytical Chemistry, Vol. 82, No. 8, April 15, 2010

Figure 3. Plots of the amount of particles diffused across the membrane versus time (A) in hexane solutions across an alumina membrane with 100 µm diameter pores and (B) in water across a polycarbonate membrane with 30 µm diameter pores, with linear regressions. Symbols represent (black b) 3 nm diameter, (red b) 8 nm diameter, and (green 1) 14 nm diameter CoFe2O4 particles and Au nanoparticles that are (yellow 1) 2.5 nm diameter hexadecanethiol stabilized (in hexane) or 13 nm diameter citrate stabilized (in water).

comparison. In all cases (in both water and hexane and using all pore sizes), we observe an inverse dependence of the rate of permeation of particles on the particle’s size. The slopes of the plots in Figure 3 are used together with the known membrane thicknesses to calculate the flux of the particles (J, in mol cm2 hr-1) across the porous membrane (i.e., in the absence of an applied magnetic field):13 J)

dC V dt MwA

(1)

where the change in concentration (C, in mg/mL) as a function of time (t, in s) is determined using the extinction coefficient (Table S1 in the Supporting Information); V is the volume of the receiving cell (in mL); MW is the estimated molar mass of the particle (mg/mol, Table S1 in the Supporting Information); and A is the effective surface area of the membrane (in cm2) and accounts for differences in membrane porosity. The plot in (13) (a) Yang, T.; Hong, L.; Vaidyanathan, N.; Weber, S. G. J. Membr. Sci. 2009, 349, 170–176. (b) Odom, D. J.; Baker, L. A.; Martin, C. R. J. Phys. Chem. B 2005, 109, 20887–20894. (c) Yu, S.; Lee, S. B.; Martin, C. R. Anal. Chem. 2003, 75, 1239–1244. (d) Tsuru, T.; Izumi, S.; Yoshioka, T.; Asaeda, M. AIChE J. 2000, 46, 565–575. (e) Jirage, K. B.; Hulteen, J. C.; Martin, C. R. Anal. Chem. 1999, 71, 4913–4918. (f) Vroon, Z. A. E. P.; Keizer, K.; Gilde, M. J.; Ver weij, H.; Burggraaf, A. J. J. Membr. Sci. 1996, 113, 293–300.

min prior to each experiment, the results in Figure 4 further suggest that some of the observed differences in membrane partitioning may be at least in part responsible for this trend. Ongoing experiments that modify the surface chemistry of both the pores and particles, and as a function of concentration, are further examining the details of these effects. Au and CoFe2O4 particles of approximately the same size permeate the membrane at roughly the same fluxes: 13 nm diameter citrate stabilized Au and 14 nm diameter CoFe2O4 particles have comparable fluxes in water, as do 2.5 nm diameter Au clusters and 3 nm diameter CoFe2O4 particles in hexane. The Au particles can, therefore, serve as adequate measures of the CoFe2O4 diffusive flux for examining the impact of magnetic fields on the latter particles’ permeation across the membrane. Effects of Magnetic Flux Gradients. The above studies focused on quantifying motion through porous membranes in the absence of a magnetic field. When a magnetic flux is applied to superparamagnetic particles, a magnetic dipole is induced in each particle; inhomogeneous fields cause magnetized particles to move toward higher flux densities. The force exerted on the particle, FM is related to the volume magnetic moment of the particle (m) by14 FM ) (m·∇)B

Figure 4. Plots of particle flux (J) vs membrane pore diameter for CoFe2O4 nanoparticles with (b) 3 nm, (1) 8 nm, and (9) 14 nm diameter in (A) hexane and compared to (() 2.5 nm diameter Au particles and (B) aqueous solutions with (() 13 nm diameter Au particles (no magnetic field is applied). Solid symbols: polycarbonate membranes. Unfilled symbols: alumina membranes.

Figure 4 summarizes the diffusive flux for each particle and membrane pore size in both hexane and water and reveals several trends in transport in these systems. For example, 3 nm diameter CoFe2O4 nanoparticles in water permeated across a membrane with 30 nm diameter pores had J of 7.9 × 10-3 µmol/cm2 hr. As the diameter of the pore increases from 30 to 100 nm, J of the 3 nm diameter CoFe2O4 nanoparticles increased by a factor of nearly five. Similar trends are observed for each of the nanoparticle samples. In comparison with small molecules permeating across a membrane with similar pore sizes, the flux of the particles is 1 to 2 orders of magnitude slower, consistent with the much larger sizes of the particless.13b In both hexane and aqueous solutions, flux apparently decreases with increasing particle size and decreasing pore diameter. We note when comparing particles of differing size that, although the solutions were equimass, molar concentrations do vary and so also impact flux. In water, permeation across the thicker alumina membranes is slower than the polycarbonate membranes although the opposite is true in hexane solutions. We observe slightly faster transport in aqueous solutions than in hexane, which is unexpected since the viscosity of hexane (2.9 x10-3 P) is smaller than the viscosity of water (8.9 × 10-3 P). These differences could be due to varying degrees of wettability of the alumina versus polycarbonate surfaces, both of which are hydrophilic in nature. Since the membranes were each soaked in solvent for a minimum of 15

(2)

where m is the magnetic moment (A · m2) and 3B is the magnetic flux gradient (T/m). In a size monodisperse nanoparticle sample, FM is approximately the same for all particles in the population. A magnetic flux gradient applied from the side of the receiving compartment gives rise to FM that induces motion toward and across the porous membrane. According to eq 2, measurements of magnetic field-enhanced solution transport could provide a measure of particle magnetic susceptibility. However, this apparatus applies inhomogeneous magnetic fields across the membrane so that there is lateral variation in FM. Nonetheless, qualitative analysis of the impact of magnetic fields on diffusive flux is possible because, for a given magnetic nanoparticle size and experimental apparatus, the magnetic moment and, therefore, FM are constant. We, therefore, turned to measurements of particle transport across the membrane in the presence of a magnetic flux gradient. In these experiments, both a concentration gradient and a magnetic flux gradient across the membrane exist so that the total flux is expected to contain both diffusive and magnetic induced velocity components. To qualitatively assess the impact of the applied magnetic field, we estimate the increase in absorbance due to magnetic force induced motion across the porous membrane as AM ) A - AD

(3)

where A is the raw absorbance data and AD is obtained from the linear regression of the transport plots as in Figure 2. This simple model does not take into account changes in local concentration as a result of applied magnetic fields, which would in turn impact diffusive flux. Instead, it serves to compare transport across the membrane in the presence and absence of a magnetic (14) Wilson, R. J.; Hu, W.; Fu, C. W. P.; Koh, A. L.; Gaster, R. S.; Earhart, C. M.; Fu, A.; Heilshorn, S. C.; Sinclair, R.; Wang, S. X. J. Magn. Magn. Mater. 2009, 321, 1452–1458.

Analytical Chemistry, Vol. 82, No. 8, April 15, 2010

3159

Figure 5. Plots of the rate of change in absorbance resulting from an applied magnetic flux gradient to (A) (black b) 3 nm; (red b) 8 nm; and (green 1) 14 nm diameter CoFe2O4 nanoparticles and (yellow 1) 13 nm Au nanoparticles in water, translocated across a 200 nm diameter pore alumina membrane and (B) 14 nm diameter CoFe2O4 nanoparticles in water across (black b) 30 nm and (red b) 50 nm diameter porous polycarbonate or (yellow 1)100 nm and (green 1) 200 nm diameter porous alumina membranes.

field. Figure 5A is a representative comparison of the AM as a function of time for 3, 8, and 14 nm diameter CoFe2O4 particles and 13 nm diameter Au particles in water. The changes in AM with time as these are translocated across a 200 nm diameter porous membrane are, thus, a measure of the degree of enhanced transport as a result of the magnetic field. We observe no magnetic-induced flux of Au nanoparticles across the membrane when the magnetic field is applied. The data in Figure 5A further show that, as the magnetic nanoparticle size increases, the rate of magnetic translocation across the pore also increases, consistent with the known increase in magnetic moments and higher FM (eq 2) of particles as their size increases. Differences in magnetic-induced transport rate across the membrane as a function of pore diameter are also observed: as the size of the pore decreases, both the AM and AD are attenuated. Like diffusive transport, this is most pronounced for the 14 nm diameter particles, as shown in Figure 5B for aqueous solutions. Since slowed transport within the pores is already accounted for in the AD term, this effect is attributed at least in part to attractive interparticle interactions of magnetized nanoparticles when a magnetic flux is applied. There have been several reports of observations of chains,14,15 rings,16 and aggregates17 of magnetic nanoparticles while in a magnetic field. Since these would have a larger effective diameter, transport through the porous membrane would be expected to be slower: magnetically attracted aggregates approaching the dimensions of the pore can be expected to be excluded, move slowly, or block the pores. Any of these scenarios would result in an inverse relationship between magnetic transport rate and pore size. The observation that the larger 14 nm CoFe2O4 particles are more affected than smaller particles is consistent with this scenario, since their magnetic (15) Wang, H.; Chen, Q.-W.; Sun, L.-S.; Qi, H.-P.; Yang, X.; Zhou, S.; Xiong, J. Langmuir 2009, 25, 7135–7139. (16) Tripp, S. L.; Pusztay, S. V.; Ribbe, A. E.; Wei, A. J. Am. Chem. Soc. 2002, 124, 7914–7915. (17) Aqil, A.; Duguet, V. E.; Passirani, C.; Benoit, J. P.; Roch, A.; Muller, R.; Jerome, R.; Jerome, C. Eur. Polym. J. 2008, 44, 3191–3199.

3160

Analytical Chemistry, Vol. 82, No. 8, April 15, 2010

moment and interparticle interactions would be strongest. As a result, no net change in absorbance, and therefore flux across the membrane, while under a magnetic field, is observed for 14 nm particles through 30 and 50 nm diameter pores. Given these dimensions, this effect could arise from as few as 2-4 particles magnetically aggregating. In all cases, removal of the field and turbulent mixing of the solution breaks any magnetic particle interactions, and these are never observed in TEM images. The use of nanoparticle transport across porous membranes is a tool for understanding their motion and separation, and application of magnetic flux gradients provides a handle with which to probe magnetic transport in solution and in confined geometries. Our results point to magnetic-induced aggregation of larger CoFe2O4 particles while in the magnetic field, which although impede transport through porous membranes, suggests opportunities in using this to obtain size-selective separations Because these data have further implications for the transport of nanoparticles in porous media or in constrained geometries, our ongoing investigations aim to further understand the roles of surface chemistries of the pore and particles and the affects of attractive magnetic aggregation. ACKNOWLEDGMENT This work was generously supported by the National Science Foundation (CHE- 0848701). We gratefully acknowledge assistance from R. Haldeman and M. Hazen from the Penn State Huck Institute of the Life Sciences Electron Microscopy Facility. SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

Received for review August 5, 2009. Accepted March 8, 2010. AC901770K