Magnetophoresis of microspheres covered by magnetic nanoparticlesX. Zhao and L. E. HelsethaSchool of Physical and Mathematical Sciences, Division of Physics and Applied Physics,Nanyang Technological University, Singapore

Received 8 June 2007; accepted 15 July 2007; published online 13 September 2007

We fabricated magnetic microspheres by letting magnetic nanoparticles self-assemble on the surfaceof surfactant-covered oil droplets. The magnetophoretic velocities of individual magneticmicrospheres were measured in a defined magnetic field gradient, and their susceptibilities werefound to scale with the surface area. 2007 American Institute of Physics.DOI: 10.1063/1.2776008

I. INTRODUCTION

It has been well established that fine solid particles canspontaneously adsorb to the oil-water interface when the in-terfacial tensions of the solid/oil SO, solid/water SW, andoil/water OW satisfy SOSWOW.1 Energy is requiredto remove the particle to the oil phase or to the water phasesince adsorption is energetically favored.2 This mechanismhas been used to create Pickering emulsions,3 which are sta-bilized by fine particles at the oil-water interface,4,5 and tofabricate colloidosomes by the self-assembly of fine particlesonto the interface of emulsion droplets.1,6 Recently, colloido-somes have been recognized as promising potential vehiclesdue to precise control of size, permeability, mechanicalstrength, and compatibility. To this end, magnetic colloido-somes have attracted interest since they can be manipulatedby an external magnetic field, and may therefore be useful intargeted delivery.7,8

The magnetic properties of magnetic particles rely onmany factors, such as the shape of the particles, the micro-structure, and type, amount and distribution of magnetic ma-terial within particles, thus there are strong differences be-tween individual magnetic particles.9 Detailed knowledge ofthe properties of individual magnetic particles is required byan increasing number of biomedical applications from micro-fluidic applications in lab-on-a-chip system to drugdelivery.10 The magnetic characterization of individual mi-crometer sized solid magnetic particles has been done bymagnetophoretic analysis,9 while submicron particles havebeen studied using particle tracking of thermally fluctuatingparticles placed in a potential well.10 Here we present amethod for fabricating such colloids by self-assembly ofmagnetic nanoparticles onto the interface of emulsion drop-lets. Such micrometer-sized oil droplets encapsulated by ashell composed of magnetic nanoparticles have so far notbeen characterized using magnetophoresis. We investigatethe motion of individual magnetic microspheres in a definedmagnetic field gradient, estimate the susceptibilities of indi-vidual magnetic microspheres, and show that the magneticmoment of the oil droplets scale with their diameters.

II. EXPERIMENTAL METHODSA. Materials

All chemicals were used as received without further pu-rification. Iron II chloride tetrahydrate FeCl24H2O 99%, potassium hydroxide KOH, volumetricstandard, 1.0 M solution in water, and hydrogen peroxideH2O2, 3 wt. % solution in water were purchased from Al-drich, whereas n-hexadecane C16H34, 99% was purchasedfrom Acros and bovine serum albumin BSA was purchasedfrom Sigma. Sodium dodecyl sulfate SDS, electrophoresisgrade was purchased from Fisher. Ultrapure water was usedto prepare aqueous solutions.

B. Preparation of magnetic nanoparticlesMagnetite nanoparticles were prepared in open air at

room temperature.11 We added 4.2 mL of 1 M KOH aqueoussolution to 25 mL of 0.1 M FeCl2 aqueous solution. Then250 L of H2O2 3 wt. % aqueous solution was added intothe solution to yield a black precipitate which was stronglyattracted to a permanent magnet. The particles were sepa-rated by magnetic decantation, washed with 320 mL dis-tilled water and 220 mL acetone, and finally dried in air atroom temperature. As reported in Ref. 11, the structure of thenanoparticles is an intermediate between Fe3O4 and -Fe2O3.

Figure 1a shows a scanning electron microscopeSEM image of the magnetic nanoparticles. Image analysisrevealed that the particles have irregular shapes and are poly-

aElectronic mail: lars@ntu.ed.sg

FIG. 1. Color online a The SEM image of magnetic nanoparticles in sizeranging from 16 to 44 nm with an average diameter of d=296 nm bMagnetic moment versus magnetic field measured for powder of the mag-netic nanoparticles.

JOURNAL OF APPLIED PHYSICS 102, 054905 2007

0021-8979/2007/1025/054905/4/$23.00 2007 American Institute of Physics102, 054905-1

disperse in size ranging from 16 to 44 nm with an averagediameter of d=296 nm. The magnetic properties of thenanoparticles were measured using a vibrating sample mag-netometer Lakeshore 7404. The anhysteretic curve of mag-netic moment versus magnetic field shown in Fig. 1b sug-gested that the particles are superparamagnetic.

C. Preparation of magnetic microspheresAn amount of 3 mg magnetic nanoparticles was washed

with 1 mL ultrapure water once, dispersed in 1 mL ultrapurewater, and sonicated for 1 min to get a uniform black solu-tion by the sonicator DH.D250H. Next 15 L of 5 mM ofthe surfactant SDS was added to the solution and shaken byhand for 1 min, followed by sonication for 1 min. We thenadded 400 L of the oil hexadecane and shook the containerby hand for 3 min, followed by sonication for 1 min to formmagnetic microspheres. The surfactant SDS will cover theoil droplets, and also assist in the attachment of magneticnanoparticles to the oil droplets. Three phases were distrib-uted in the container due to their different densities. Thelower phase was water; oil droplets coated with magneticnanoparticles were in the middle layer, while the upper phasewas hexadecane. The microspheres are very stable over timeand maintain their appearance for over 7 months.

The experimental setup used for magnetophoretic mea-surements is shown in Fig. 2. The magnetic microspheresmoved toward the permanent magnet, while the motion wasobserved with a Leica polarization microscope 50microscope objective used in transmission mode and re-corded with a CCD camera 25 fps. The pixel size in amovie is 0.3 m. The recorded movies were analyzedmanually. The magnetic field was measured using a gauss-meter with a Hall effect probe Hirst GM04. An amount of2 L of the magnetic microspheres in the middle layer wasdissolved in 1 mL of 1 mg/L BSA solution as the samplewhich was used for magnetophoretic measurements. By us-ing BSA we could prevent magnetic microspheres from be-ing adsorbed to the glass slide. It is known that BSA has aneffective negative charge at pH=7,12 thus adsorbing to thepositively charged nanoparticles and preventing them fromadsorbing to the negatively charged glass slide.

III. RESULTS

In order to understand how the magnetic microspherescan be transported, we did magnetophoretic measurements ofthe individual magnetic microspheres. A typical magneto-

phoretic behavior of a magnetic microsphere is shown in Fig.3, which was made by superimposing captured images at 5 sintervals. The x component of the velocity of the micro-sphere was measured from the superimposed images.

The x components of the velocities in magnetic field B=24.7 mT, dB /dx=2.9 mT/mm for 30 microspheres whichappeared to be uniformly coated with magnetic nanoparticlesare plotted against their radius in Fig. 4. This figure showsthat larger microspheres usually have larger velocities, typi-cally up to 10 m/s. However, it is also clear that micro-spheres of the same size can have velocities differing by afactor of 4 or larger. We also measured x component of thevelocities of microspheres which had clusters attached totheir surfaces, and found that the velocities had a biggervalue, with a maximum of approximately 70 m/s.

IV. DISCUSSION

The x component of the magnetic force Fmx acting on themagnetic particle in a magnetic field can be expressed as1315

Fmx =sphere0

BBx

, 1

where sphere is the susceptibility of the particle, B is themagnetic field in the absence of the particle, and 0 is thevacuum magnetic permeability. Here we assume that the me-

FIG. 2. Color online Experimental devices for magnetophoreticmeasurements.

FIG. 3. The magnetophoretic behavior of a single magnetic microsphere inmagnetic field B=24.7 mT, B /x=2.9 mT/mm. This microphotographwas made by superimposing the images captured at 5 s intervals.

FIG. 4. Diagram of the x component of the velocities under magnetic fieldB=24.7 mT, B /x=2.9 mT/mm for 30 microspheres which appeared tobe coated with magnetic nanoparticles uniformly in the images versus theirradius. r=0.3 m and v=0.3 m/s.

054905-2 X. Zhao and L. E. Helseth J. Appl. Phys. 102, 054905 2007

dium is nonmagnetic and the magnetic particle is paramag-netic.

In the experiments, a nearly linear relation exists be-tween the magnetic moment of magnetic nanoparticles andthe magnetic field when absolute value of the magnetic fieldis less than 40.9 mT shown in Fig. 1, while the magneto-phoretic measurements were conducted in this region B=24.7 mT. The susceptibility of the medium 1 mg/L BSAsolution is negligible 7107 compared with that ofmagnetic nanoparticles 8103. Therefore Eq. 1 is agood approximation for the magnetic microspheres.

In our experiment we observed that the diameters L andvelocities v of the spherical oil droplets ranged between2 mL36 m, 0.06 m/sv10 m/s, respec-tively. From these data we calculated Reynolds number to beRe41041. For this reason the viscous drag force actson the microsphere FDx can be represented by Stokeslaw16,17

FDx = 6rvx, 2

where is the viscosity of the medium and vx is the magne-tophoretic velocity of the microsphere in the x direction andr is the radius of the microsphere. Equations 1 and 2 canbe combined in Newtons equation of motion

Fmx + FDx = max, 3

where ax and m are the acceleration and the mass of themicrosphere. Since the mass and Reynolds numbers aresmall, the inertia max is also negligibly small, thus allowingus to represent sphere as

sphere =60rvx

BBx

. 4

The calculated value of sphere of 30 microspheres accordingto Eq. 4 is shown in Fig. 5 as a function of r. This figureshows that larger microspheres usually have larger suscepti-bilities. Considering the surface area of a microsphere 4r2,the curve of sphere=4r2A A=1.13107 /4m displayedin Fig. 5 suggests that sphere is roughly proportional to thesurface area of the microsphere. This is reasonable since themagnetic nanoparticles were attached to the surface of the oildroplet, which means the larger the surface area, the moreopportunities for the magnetic nanoparticles to attach to thesurface. In optical micrographs these microspheres appear to

be relatively uniformly coated, therefore an explanation forvariation of susceptibilities can be distribution of magneticnanoparticles on the microspheres, the number of magneticnanoparticles attached to the surface of oil droplets and thestrong differences between individual magnetic nanoparticlesdue to the shape of the nanoparticles and theirnanostructure.10

For a 24.7 mT magnetic field, we calculated the averagemagnetic moment of one nanoparticle to be m1=1.961018 A m2 according to Fig. 1b. So the number n ofnanoparticles attached to the surface of the oil droplets canbe estimated by

n =sphereB0m1

. 5

Figure 6 shows the number of the magnetic nanoparticles peroil droplet versus the radius of the oil droplets for 30 micro-spheres. When the nanoparticles are close-packed on the sur-face of the oil droplets, the number of the particles per oildroplet is given by

n0 =4r2

d2. 6

We define

R =n

n0. 7

Here R is a measure of how large a fraction of the oil drop-lets is coated with magnetic nanoparticles. Diagram of ratioof number R of the oil droplets versus the radius for 30 oildroplets is shown in Fig. 7. R is in the range of 1%19%.

V. CONCLUSION

We fabricated magnetic microspheres by self-assemblyof magnetic nanoparticles onto the interface of emulsiondroplets and measured the magnetophoretic velocities of thiskind of magnetic microspheres in a defined magnetic field.We found that larger microspheres usually have larger ve-locities. However, it is also clear that microspheres of thesame size can have velocities that differ by more than afactor of 4. Susceptibilities of 235 m magnetic micro-spheres have been calculated to characterize their magneticproperties. We found that larger microspheres usually have alarger susceptibility and sphere is proportional to the surfacearea of the microsphere roughly. However, it is also clearthat microspheres of the same size can have susceptibilities

FIG. 5. Color online The susceptibility is calculated according to Eq. 6using the data in Fig. 4. b is magnified one of the circled parts of a.sphere=21018 m3. The line was obtained by a least-square method.sphere=1.131019r2, R2=0.71421.

FIG. 6. The number of the magnetic nanoparticles per oil droplet versus theradius of the oil droplets for 30 microspheres. b is magnified one of thecircled parts of a. We estimate the upper bound of error n=2104.

054905-3 X. Zhao and L. E. Helseth J. Appl. Phys. 102, 054905 2007

which differ by more than a factor of 4. Since many drugscan be dissolved in the oil phase and the magnetic micro-spheres are stable and controlled efficiently by an externalmagnetic field, we hope the current study can assist effortstoward obtaining a tool for controlled drug delivery.

ACKNOWLEDGMENTSThe authors thank Dr. Zhu Yong for help with the SEM

and Patrick Lai at the School of Materials Sciences for helpwith the VSM.

1A. D. Dinsmore, M. F. Hsu, M. G. Nikolaides, M. Marquez, A. R. Bausch,and D. A. Weitz, Science 298, 1006 2002.

2M. J. Hey and J. G. Kingston, J. Colloid Interface Sci. 298, 497 2006.3S. U. Pickering, J. Chem. Soc. 91, 2001 1907.4L. E. Helseth, R. M. Muruganathan, Y. Zhang, and T. M. Fischer, Lang-muir 21, 7271 2005.

5C. Zeng, H. Bissig, and A. D. Dinsmore, Solid State Commun. 139, 5472006.

6X. D. He, X. W. Ge, M. Z. Wang, and Z. C. Zhang, Polymer 46, 75982005.

7H. Duan, D. Wang, N. S. Sobal, M. Giersig, D. G. Kurth, and H. Mo-hwald, Nano Lett. 5, 949 2005.

8S. Melle, M. Lask, and G. G. Fuller, Langmuir 21, 2158 2005.9U. O. Hafeli, M. A. Lobedann, J. Steingroewer; L. R. Moore, and J. Riffle,J. Magn. Magn. Mater. 293, 224 2005.

10K. van Ommering, J. H. Nieuwenhuis, L. J. V. Ijzendoorn, B. Koopmans,and M. W. J. Prins, Appl. Phys. Lett. 89, 142511 2006.

11M. Tada, S. Hatanaka, H. Sanbonsugi, N. Matsushita, and M. Abe, J. Appl.Phys. 93, 7566 2003.

12J. A. Reynolds and C. Tanford, Proc. Natl. Acad. Sci. U.S.A. 66, 10021970.

13R. Gerber and R. R. Birss, High Gradient Magnetic Separation Wiley,New York, 1983, p. 69.

14G. P. Hatch and R. E. Stelter, J. Magn. Magn. Mater. 225, 262 2001.15C. X. Liu, L. Lagae, R. Wirix-Speetjens, and G. Borghs, J. Appl. Phys.

101, 024913 2007.16M. Suwa and H. Watarai, Anal. Chem. 73, 5214 2001.17S. W. Yuan, Foundations of Fluid Mechanics Prentice-Hall, Inc., Engle-

wood Cliffs, NJ, 1967, p. 286.

FIG. 7. Diagram of the surface coverage of oil droplets by magnetic nano-particles R versus the radius for 30 oil droplets. We estimate the upperbound of error R=9%.

054905-4 X. Zhao and L. E. Helseth J. Appl. Phys. 102, 054905 2007