Magnetic Colloidosomes Derived from Nanoparticle Interfacial Self-Assembly

NANO LETTERS

Magnetic Colloidosomes Derived from Nanoparticle Interfacial Self-Assembly

2005 Vol. 5, No. 5 949-952

Hongwei Duan,† Dayang Wang,*,† Nelli S. Sobal,‡ Michael Giersig,‡ Dirk G. Kurth,† and Helmuth Mo1 hwald† Max Planck Institute of Colloids and Interfaces, D-14424, Potsdam, Germany, and Center of AdVanced European Studies and Research, D-53175, Bonn, Germany Received March 21, 2005; Revised Manuscript Received April 14, 2005

ABSTRACT Based on the interfacial self-assembly of magnetite nanoparticles, we demonstrate the formation of colloidosomes with shells predominantly composed of monolayers of liquid-like, close-packed nanoparticles. The gelation of aqueous phase with agarose leads to robust and waterdispersible nanoparticle colloidosomes, allowing encapsulation of various water soluble materials. The cutoff of the nanoparticle colloidosomes obtained is primarily defined by the nanoparticle size. This controllable permeability should be of great importance for the encapsulation application.

In this communication, we demonstrate the fabrication of magnetic colloidosomes by directing self-assembly of magnetite (Fe3O4) nanoparticles (NPs) at the interfaces of waterin-oil droplets. Their shell permeability can be tuned by the size of the NPs used. Microencapsulation is of considerable interest for both fundamental studies and technical applications.1 Encapsulation vehicles with defined permeability, flexibility of loading, possibility of further property-tailored modification, and in some cases facility of manipulation are in great demand. Numerous approaches, relying mainly on self-assembly and the use of templates, have been developed to create microcapsules with diverse functionalities both on the wall and in the interior.2 According to their flexibility of tuning the permeability over a broad size range, colloidosomes, microcapsules composed of shells of close-packed colloidal particles, have recently been recognized as promising potential vehicles.3 Their permeability is expected to be tunable over the nanometer to micrometer scale by varying the size of the colloidal particles used. Nevertheless, nanoscale selective permeability is far from being achieved as it remains hard to stabilize emulsion droplets with NPs, due to the fact that the spatial fluctuation of NPs trapped at interfaces suffices to remove them into bulk phases.4-6 Most recently, it has been realized to direct NP selfassembly at the water/oil interface by a proper hydrophobic surface coating.5 Lin and co-workers stabilized water-intoluene droplets with CdSe NPs and observed particle-sizedependent interfacial attachment.5a,d They further cross-linked * Corresponding author. Fax: +49 331 567 9202. Email: dayang@ mpikg-golm.mpg.de. † Max Planck Institute of Colloids and Interfaces. ‡ Center of Advanced European Studies and Research. 10.1021/nl0505391 CCC: $30.25 Published on Web 04/23/2005

© 2005 American Chemical Society

the ligands on NPs attached on the water/toluene interface, creating ultrathin membranes.6 They also demonstrated the potential of encapsulating water droplets with these crosslinked NP membranes.6 Recently we have successfully implemented interfacial self-assembly of Au and Fe3O4 NPs by capping them with ligands bearing a 2-bromo-2-methylpropionate terminus.5c Based on this, herein we demonstrate the formation of magnetic colloidosomes by assembling Fe3O4 NPs of different sizes at the interface of water-intoluene droplets. By gelating the water phase with agarose, robust magnetic NP colloidosomes are obtained and their permeability is analyzed by luminescent water-soluble CdTe NPs. For this work, 4.0, 5.0, and 8.0 nm Fe3O4 NPs capped with 2-bromo-2-methylpropionic acid were prepared based on the method reported proviously.5c To the toluene dispersions of these NPs water was added. After emulsification by vigorously shaking for a few minutes, Fe3O4 NP colloidosomes, NP-stabilized water-in-toluene droplets were obtained (Scheme 1a). Figure 1a shows a typical confocal microscopy image of colloidosomes derived from 8 nm Fe3O4 NPs, in which water droplets of around 100 µm in diameter are obviously surrounded with dark thin shells. Wrinkles are clearly visible, which are due to the collapse of the NP shells of colloidosomes upon drying. The dried colloidosomes were visualized by transmission electron microscopy (TEM). Figure 1b reveals that the colloidosome shell is composed of close-packed 8.0 nm Fe3O4 NPs. TEM images of the broken areas of the resulting NP colloidosomes identify that the colloidosomes are enclosed by shells of 8.0 nm Fe3O4 NP monolayers, where the NPs are liquid-like closely packed (Figure S1, Supporting Information), con-

Scheme 1. Schematic Illustration of the Processes of Preparing Colloidosomes Based on Self-assembly of Fe3O4 NPs (golden dots) at Interfaces of Toluene and Water (a) and of Toluene and 1.5 wt % Agarose Aqueous Solution Containing CdTe NPs (red dots) (b)

Figure 1. (a) Confocal microscopy image of colloidosomes, waterin-toluene droplets stabilized with 8 nm Fe3O4 NPs. (b) TEM image of dried 8.0 nm Fe3O4 NP colloidosomes. (c) Confocal microscopy image of 8.0 nm Fe3O4 NP colloidosomes containing agarose hydrogel cores. (d) Photograph of accumulation of the colloidosomes in water, induced by a magnet.

sistent with the shell structures of colloidosomes formed from micron-sized spheres.3b,c Their shell structure, liquid-like close-packed NP arrays, should be of importance for encapsulation. The permeability of the colloidosomes obtained is hard to determine since they are extremely sensitive to the environment, tending immediately to collapse or even demulsify upon adding other materials into either of the two phases. To improve the stability of the colloidosome, herein we gelated the aqueous phase with agarose. Aqueous solutions of 1.5% agarose were emulsified in the toluene dispersion of Fe3O4 NPs at 70 °C. Cooling to room temperature caused the gelation of the aqueous phase. After collection by an external magnetic field, followed by three times washing with ethanol, the resulting robust colloidosomes were redispersed in water (Scheme 1b). The colloidosomes with agarose gel cores exhibit greatly improved mechanical stability, indicated by the fact that they keep spherical shape and do not collapse upon drying for a rather long time (Figure 1c), as compared to those without gelated cores. TEM images of their ultrathin sections reveal that the agarose gel cores of the colloidosomes obtained are enclosed by 8.0 nm Fe3O4 NP monolayers after the solvent exchange (Figure S2, Supporting Information). This suggests the strong attachment stability of Fe3O4 NPs on the gel cores during replacing the toluene phase with ethanol and eventually with water. The driving force of this robust attachment of the Fe3O4 NPs is not entirely clear by far. Our speculation is that since ethanol is the precipitant of the Fe3O4 NPs used, the addition of ethanol should not cause detachment but slight aggregation of the Fe3O4 NPs on colloidosomes, which might enhance the hydrophobic 950

interaction of the neighboring particles, thus locking NPs together and in turn preventing them from detachment from the surfaces of the agarose gel cores to ethanol. Due to their unique surface wettability with a water contact angle of close to 90°,5c furthermore, the Fe3O4 NPs used are only partially wetted by water, so they should not detach from the surface of the agarose gel core into water. Similar to the literature,3d in our work, the agarose gel core is essential to maintain the colloidosomes integrity against washing with ethanol. In the absence of the agarose gel in the cores, since ethanol is miscible both with toluene and with water, its addition immediately demulsified colloidosomes into a homogeneous mixture solution and led to aggregation and precipitation of the Fe3O4 NPs; large clumps of the Fe3O4 NP aggregated were obvious. Thanks to the Fe3O4 NP shells, these magnetic colloidosomes allow manipulation by applying an external magnetic field, as shown in Figure 1d, which may be important for targeted delivery. Based on the geometry of Fe3O4 NP colloidosomes and the Fe3O4 NP amount determined by absorption spectroscopy (Supporting Information), the NP number per colloidosome (N) may be estimated based on the following equation: N)

MFe3O4 FVwater

×

( )

DNPS 3 × 106 dNP

(1)

where MFe3O4 is the weight of the Fe3O4 NPs used (mg), F the Fe3O4 density (5.1 g/mL), Vwater the volume of water added (mL), DNPS the diameter of colloidosomes obtained (µm) and dNP the diameter of nanoparticles used (nm). Accordingly, the NP number per colloidosome is calculated Nano Lett., Vol. 5, No. 5, 2005

Figure 2. Confocal fluorescence microscopy images of 8.0 nm Fe3O4 NP colloidosomes containing agarose hydrogel cores loaded with 4.0 nm CdTe NPs, freshly transferred in water (a) and of those NP colloidosomes after immersion in water for 2 h, revealing release of the CdTe NPs trapped within the colloidosomes (b).

Figure 3. Fluorescence photographs of freshly prepared 8.0 nm Fe3O4 colloidosomes containing agarose hydrogel cores loaded with 4.0 nm CdTe NPs (a) and after immersion in water for 10 min (b), in which the aqueous solution turns red-luminescent, demonstrating the release of the loaded CdTe NPs out of gel cores.

about 7.8 × 108 for 8.0 nm Fe3O4 NP colloidosomes, 1.6 × 109 for 5.0 nm ones and 3.1 × 109 for 4.0 nm ones. The theoretical value of the NP number per colloidosome can be evaluated based on the following equation: N ) f3D

[( ) ( ) ] DNPS 3 DNPS dNP dNP

3

(2)

Providing the NPs in the colloidosomes are randomly closepacked, with a maximum packing density of 0.64,7 the NP number per colloidosome is calculated about 6.1 × 108 for 8.0 nm Fe3O4 NP colloidosomes, 1.5 × 109 for 5.0 nm ones, and 2.5 × 109 for 4.0 nm ones. The experimental and the theoretical values agree reasonably well. That the experimental values are higher by about 25% is likely due to the broad size distribution of the colloidosomes obtained. Another reason for this deviation may be that some areas are not covered by mono- but by multilayers of NPs. The above estimate does suggest that the NPs are close-packed in the colloidosome shell. In comparison with those without gelated cores, the colloidosomes with agarose cores remain intact against environmental variation, thus allowing incorporation of various water-soluble materials in the cores for determining the shell permeability. In this study, monodisperse watersoluble CdTe NPs were recruited as probes due to their welldefined size and geometry.8 CdTe NPs were well mixed in agarose solutions at 70 °C. After emulsification at 70 °C and cooling, CdTe NPs with different sizes were embedded within magnetic colloidosomes. Figure 2a shows a confocal fluorescence microscopy image of freshly prepared 8.0 nm Fe3O4 NP colloidosomes, loaded with 4.0 nm CdTe NPs with emission at 650 nm, in which spheres with homogeneous red emission color are observed. During immersion in water, one may observe that the trapped CdTe NPs were gradually released out of the Fe3O4 NP colloidosomes (Figure 3). As shown in Figure 2b, the red luminescence is seen from inside and outside the NP colloidosomes; the release of the loaded CdTe NPs into water makes the dark shells of Fe3O4 NPs apparent. In this work, furthermore, we embedded both 2.8 nm (emission at Nano Lett., Vol. 5, No. 5, 2005

Figure 4. Fluorescence photographs of 5.0 nm Fe3O4 NP colloidosomes containing agarose hydrogel cores loaded with both 2.8 and 4 nm CdTe NPs, formed in toluene (a), freshly transferred into water (b), immersed in water for 10 min (c), and eventually dispersed in water after removal the released CdTe NPs by washing with water (d). By the glass vial a magnet is laid, which induces the accumulation of the Fe3O4 NP colloidosomes on the vial wall.

540 nm) and 4.0 nm CdTe NPs in 5.0 nm Fe3O4 NP colloidosomes with a yellow emission color (Figure 4a). Once the resulting CdTe-loaded colloidosomes were transferred into water, as shown in Figure 4b and c, water gradually exhibit green fluorescence, suggesting that 2.8 nm CdTe NPs were released out of the colloidosomes. After 5 times washing with water, the resulting colloidsomes exhibit red fluorescence (Figure 4d), indicating that only 4.0 nm CdTe NPs are embedded in the gel cores and 2.8 ones are washed away. The release of the loaded 4.0 nm CdTe NPs is not observed at least within weeks. By using colloidosomes derived from 4.0 nm Fe3O4 NPs, such a release of 2.8 nm CdTe NPs were not observed within days. This demonstrates the selective permeability of the resulting colloidosomes, which depends on the sizes of the Fe3O4 NPs used. 951

In contrast, the gel network of the pure agarose gel may not confine CdTe NPs. Although one may load differently sized CdTe NPs in the agarose gel by dissolving agarose in the aqueous solution of CdTe NPs at 70 °C and then cooling to room temperature, the loaded CdTe NPs release out of the hydrogel upon immersion in water (Figure S3, Supporting Information). Nevertheless, it is worth noting that in the absence of agarose, we failed to stabilize the water droplets containing CdTe NPs with Fe3O4 NPs. Hence this testifies that the agarose gel core enables loading of water-soluble materials into the colloidosome but has a less contribution to their selective permeability, which is predominantly determined by the shell of close-packed Fe3O4 NPs. The permeability of colloidosomes is expected to be dependent on the sizes of interstices between NPs assembled in the shells. In liquid-like close-packed NP shells of colloidosomes, there coexist hexagonal, cubic, and pentagonal packing arrays, as suggested by Weitz and co-workers.3c Based on simple geometry analysis, one may evaluate the interstitial sizes of these three packing arrays (Scheme S1, Supporting Information). Obviously, the pentagonal arrangement gives the maximal cutoff of 0.7 of the NP size. The cutoff of 8.0 nm Fe3O4 NP colloidosomes is calculated as 5.6 nm, that of 5.0 nm ones as 3.5 nm, and that of 4.0 nm ones as 2.8 nm. Since the latter is comparable or smaller than the size of the CdTe NPs used, little amount of CdTe NPs leaks out of 4.0 nm Fe3O4 NP colloidosomes. While the cutoff of the 5.0 nm Fe3O4 NP colloidosomes is between 2.8 and 4.0 nm CdTe NPs used, one does expect the selective release of the small CdTe NPs from 5.0 nm Fe3O4 NP colloidosomes and entrapment of the large ones in accordance with the experiment (Figure 4). This also indicates that the NP shell has only few defects consistent with the above estimation of the NP number per colloidosome. Due to their unique size-tunable luminescence, CdTe NPs may be utilized as new fluorescent markers for biological detection.9 4.0 and 5.0 nm Fe3O4 NP colloidosomes loaded with CdTe NPs therefore integrate size-tunable magnetic and luminescent properties of NPs, certainly holding great promise in biotechnology. In summary, we fabricated magnetic NP colloidosomes with nanoscale selective permeability based on interfacial self-assembly of Fe3O4 NPs. By gelating the aqueous phase, robust and water-dispersible colloidosomes are constructed, allowing encapsulation of various water-soluble materials inside, which should pave a flexible and easy way for encapsulation. As the interfacial self-assembly of NPs mainly depends on their surface wettability rather than the NP chemical nature,5c various metal, metal oxides, and metal chalcogenide NPs with proper hydrophobic coating may self-

952

assemble at water/oil interfaces. The extension of our procedure into other NPs is underway, generating new NP colloidosomes probably exhibiting new electronic, optical, and magnetic cooperative properties. Acknowledgment. This work is supported by the Max Planck Society. We thank W. Tong for assistance with confocal fluorescence microscopy and J. Hartmann and R. Pitschke for help with TEM. H. Zhang is acknowledged for synthesis of water-soluble CdTe nanoparticles. Supporting Information Available: Experimental details of preparation of 2-bromo-2-methylpropionic acid-capped Fe3O4 NPs, CdTe NPs, and Fe3O4 NP colloidosomes. TEM images of the broken area of a dried 8.0 nm Fe3O4 NP colloidosome, and an ultrathin section of a dried 8.0 nm Fe3O4 NP colloidosome with an agarose core. Calculation of NP colloidosome cutoffs. Fluorescence photographs of the release of 4.0 nm CdTe NPs from agarose gel. This material is available free of charge via the Internet at http:// pubs.acs.org. References (1) Hollow and Solid Spheres and Microspheres: Science and Technology Associated with Their Fabrication and Application, in Materials Research Society Proceedings; Wilcox, D. L., Berg, M., Bernat, T., Kellerman, D., Cochran, J. K., Eds.; MRS: Pittsburgh, PA, 1995; Vol. 372, and references therein. (2) (a) Donath, E.; Sukhorukov, G. B.; Caruso, F.; Davis, S. A.; Mo¨hwald, H. Angew. Chem., Int. Ed. 1998, 37, 2202. (b) Huang, H. Y.; Remsen, E. E.; Kowalewski, T.; Wooley, K. L. J. Am. Chem. Soc. 1999, 121, 3805. (c) Marinakos, S. M.; Novak, J. P.; Brousseau Ø, L. C.; House, A. B.; Edei, E. M.; Feldhaus, J. C.; Feldheim, D. L. J. Am. Chem. Soc. 1999, 121, 8518. (3) (a) Pickering, S. J. Chem. Soc. 1907, 91, 2001. (b) Velev, O.; Furusawa, K.; Nagayama, K. Langmuir 1996, 12, 2374. (c) Dinsmore, A.; Hsu, M.; Nikolaides, M.; Marquez, M.; Bausch, A.; Weitz, D. Science 2002, 298, 1006. (d) Noble, P.; Cayre, O.; Alargova, R.; Velev, O.; Paunov, V. J. Am. Chem. Soc. 2004, 126, 8092. (4) Binks, B. P. Curr. Opin. Colloid Interface Sci. 2002, 7, 21. (5) (a) Lin, Y.; Skaff, H.; Boker, A.; Dinsmore, A. D.; Emrick, T.; Russell, T. P. Science 2003, 299, 226. (b) Reincke, F.; Hickey, S.; Kegel, W.; Vanmaekelbergh, D. Angew. Chem., Int. Ed. 2004, 43, 458. (c) Duan, H.; Wang, D.; Kurth, D. G.; Mhwald, H. Angew. Chem., Int. Ed. 2004, 43, 5639. (d) Lin, Y.; Boker, A.; Skaff, H.; Cookson, D.; Dinsmore, A.; Emrick, T.; Russell, T. P. Langmuir 2005, 21, 191. (6) Lin, Y.; Skaff, H.; Boker, A.; Dinsmore, A.; Emrick, T.; Russell, T. P. J. Am. Chem. Soc. 2003, 125, 12690. (7) Jaeger, H.; Nagel, S. Science 1992, 255, 1523. (8) (a) Rogach, A.; Katsikas, L.; Kornowski, A.; Su, D.; Eychmu¨ller, A.; Weller, H. Ber. Bunsen-Ges. Phys. Chem. 1996, 100, 1772. (b) Gaponik, N.; Talapin, D.; Rogach, A.; Hoppe, K.; Shevchenko, E.; Kornowski, A.; Eychmu¨ller, A.; Weller, H. J. Phys. Chem. B 2002, 106, 7177. (9) (a) Alivisatos, A. P. Sci. Am. 2001, 285, 59. (b) Chan, W.; Maxwell, D.; Gao, X.; Bailey, R.; Han, M.; Nie, S. Curr. Opin. Biotechnol. 2002, 12, 40.

NL0505391

Nano Lett., Vol. 5, No. 5, 2005