5530
Langmuir 2000, 16, 5530-5533
Free-Standing Layer-by-Layer Assembled Films of Magnetite Nanoparticles Arif A. Mamedov and Nicholas A. Kotov* Chemistry Department, Oklahoma State University, Stillwater, Oklahoma 74078 Received April 12, 2000 A new technique for preparation of free-standing ultrathin films is presented. These films were made by layer-by-layer (LBL) deposition process, which is utilized for coatings on solid substrates and colloids. A film composed of alternating layers of magnetite nanoparticles and poly(diallyldimethylammonium bromide) was assembled on cellulose acetate, which was subsequently dissolved in acetone. From the suspended state, the LBL film can be transferred onto any solid or porous support. The strength of the film was observed to significantly increase when every other layer of magnetite was replaced with a layer of alumosilicate sheets serving as a molecular framework for the assembly.
Introduction The layer-by-layer (LBL) assembly is a method of thin film deposition, which is often used for oppositely charged polymers.1 Recently, it was successfully applied to the preparation of thin films of nanoparticles.2-10 Its simplicity and universality complemented by the high quality of the films make the LBL assembly an attractive alternative to many thin film deposition techniques such as spin coating, Langmuir-Blodgett deposition, electrophoresis, and others. The LBL assembly is especially suitable for the production of stratified thin films in which layers of nanometer thickness are organized in a specific predetermined order. The LBL assembly of nanoparticles can be described as the sequential adsorption of (mono)layers of nanoparticles on oppositely charged layers of polyelectrolytes. The deposition of the films can be performed in a cyclic manner, which is made possible by the overcompensation of surface charge when high molecular weight species are adsorbed at a solid-liquid interface.11 Typically, LBL films are assembled on glass slides or silicon wafers. The feasibility of their use in a number of advanced electronic and photonic applications has been already demonstrated.4,5,9,10,12-14 The LBL films assembled on * To whom correspondence may be addressed. E-mail:
[email protected]. (1) Decher, G. Science 1997, 277, 1232-1237. (2) Kotov, N. A.; Dekany, I.; Fendler, J. H. J. Phys. Chem. 1995, 99, 13065-13069. (3) Correa-Duarte, M. A.; Giersig, M.; Kotov, N. A.; Liz-Marzan, L. M. Langmuir 1998, 14, 6430-6435. (4) Aliev, F.; Correa-Duarte, M.; Mamedov, A.; Ostrander, J. W.; Giersig, M.; Liz-Marzan, L.; Kotov, N. Adv. Mater. 1999, 11, 10061010. (5) Liu, Y. J.; Claus, R. O. J. Am. Chem. Soc. 1997, 119, 5273-5274. (6) Rosidian, A.; Liu, Y. J.; Claus, R. O. Adv. Mater. 1998, 10, 1087. (7) Liu, Y. J.; Wang, A. B.; Claus, R. O.; Appl. Phys. Lett. 1997, 71, 2265-2267. (8) a) Caruso, F.; Lichtenfeld, H.; Giersig, M.; Mohwald, H. J. Am. Chem. Soc. 1998, 120, 8523-8524. Donath, E.; Sukhorukov, G. B.; Caruso, F.; Davis, S. A.; Mohwald, H. Angew. Chem., Int. Ed. 1998, 37 (16), 2202-2205. Caruso, F.; Caruso, R. A.; Mohwald, H. Science 1998, 282 (5391), 1111-1114. (9) Gao, M. Y.; Richter, B.; Kirstein, S.; Mo¨hwald, H. J. Phys. Chem. B 1998, 102, 4096-4103. (10) Gao, M.; Zhang, Xi.; Yang, B.; Li, F.; Shen, J. Thin Solid Films 1996, 284-285, 242-245. (11) Schlenoff, J. B.; Ly, H.; Li, M. J. Am. Chem. Soc. 1998, 120, 7626-7634. (12) Mattoussi, H.; Radzilowski, L. H.; Dabbousi, B. O.; Thomas, E. L.; Bawendi, M. G.; Rubner, M. F. J. Appl. Phys. 1998, 83, 7965-7974. (13) Cassagneau, T.; Mallouk, T. E.; Fendler, J. H. J. Am.Chem. Soc. 1998, 120, 7848-7859.
colloids are also being tested for drug delivery.8 Here, we report on a technology of preparation of free-standing LBL films which will open the door for exploitation of these assemblies as ultrathin membranes with a variety of possible applications ranging from gas separation, sensors, micromechanical devices, and advanced catalysis to artificial cell walls and organs. Preparation of such films from inorganic colloids affords a rich palette of mechanical, chemical, optical, electrical, and magnetic properties. Importantly, the layer-by-layer mode of their preparation makes possible the degree of structural organization of such membranes, which is hardly attainable by traditional methods of their production. Experimental Section Atomic force microscopy (AFM) images were taken by using a Nanoscope IIIa Multimode instrument operating in the tapping regime with TESP tips. Typically, the surface was scanned at 2 Hz with 256 lines per image resolution and 1.2-4.0 V set point. No filter technique was applied to the images presented. Magnetic measurements were performed by using a QUANTUM Design PPMS 6000 magnetometer. The magnetic field H was created by a superconducting solenoid in the persistent mode parallel to the film’s surface. For the magnetic hysteresis loops, the correct demagnetization values corresponding to the sample signals were obtained by subtracting the diamagnetic signal of the substrate from the total registered signal. The linear magnetic response of nonmagnetic components was intrapolated to the -10 kOe < H < 10 kOe region from the high field wings of magnetization curves. The total magnetization signal from each specimen was scaled to the mass of the sample and UV absorption intensity at 350 nm. X-ray photoelectron spectra (XPS) were recorded at a takeoff angle 0° using a system EscaLab 220-IXL. The KR X-ray line of Al at a spectrometer pass energy of 1486.92 eV and step resolution of 0.1 eV was used as source. The authors are indebted to XPS facilities of the University of Vigo and to Carmen Serra-Rodrı´guez for the help with XPS measurements. Transmission electron microscopy (TEM) images were taken on a JEOL-2000 FX instrument operating at 80kV. TEM samples were prepared so that a layer of magnetic nanoparticles was deposited only on one side of the carbon-coated TEM grid. The authors express their appreciation to Terry Colberg from the TEM facilities of OSU for the help with cross-sectioning. Absorption spectra in UV-vis were taken with a HewlettPackard 8453A spectrophotometer. (14) Liu, Y. J.; Wang, A. B.; Claus, R. O. Appl. Phys. Lett. 1997, 71, 2265-2267.
10.1021/la000560b CCC: $19.00 © 2000 American Chemical Society Published on Web 05/27/2000
Letters
Langmuir, Vol. 16, No. 13, 2000 5531
Figure 1. (A and B) Atomic force microscopy images of (M)1 and (C)1 films on silicon wafers. Fine-grain texture of the left image is characteristic for the magnetite film. Each white dot represents a 2D cluster of 8-10 nm Fe3O4 nanoparticles. Montmorillonite adsorption time in image B was shortened from 1 min, which was typically used for making the films, to 25 s to reveal better the shape and size of the platelets. (C and D) Model structures of magnetite-polyelectrolyte and magnetite-polyelectrolyte-clay multilayers. Drawn not to scale.
Results and Discussion The principle of making free-standing LBL assembled films is quite simple. Initially, the LBL deposition process is carried out on a solid substrate. When a degree of structural sophistication and/or desirable thickness is achieved, the prepared LBL assembly is lifted off by dissolving the substrate in a suitable organic solvent. To some extent, the idea of free-standing films resembles the preparation of polyelectrolyte microcapsules by burning or dissolving micrometer scale colloids.8 In fact, it is surprising that self-sustained versions of LBL films have not been made by now, because the “free” films without a substrate allow for the direct experimental determination of many physical properties of the LBL films of fundamental significance such as ion permeation and chain packing, which are being actively discussed in the literature with little or no quantitative data.8,11,15,16 The assemble-and-dissolve principle imposes two major requirements on the substrate: (1) the organic solvents (15) Mohwald, H.; Lichtenfeld, H.; Moya, S.; Voigt, A.; Baumler, H.; Sukhorukov, G.; Caruso, F.; Donath, E. Macromol. Symp. 1999, 145, 75-81. Stroeve, P.; Vasquez, V.; Coelho, M. A. N.; Rabolt, J. F. Thin Solid Films 1996, 284-285, 708-712. Leva¨salmi, J.-M.; McCarthy, T. J. Macromolecules 1997, 30, 1752-1757. Ackern, van F.; Krasemann, L.; Tieke, B. Thin Solid Films 327-329, 762-766. (16) Kotov, N. A.; Magonov, S.; Tropsha, E. Chem. Mater. 1998, 10, 886-895.
used in the lift-off step must not damage the film and (2) the substrate must be hydrophilic with a positively or negatively charged surface. The latter will promote electrostatic attraction between the substrate and the first layer of polyelectrolyte, serving as a film foundation. These requirements are satisfied for cellulose acetate (CA), which is insoluble in water but dissolves readily in acetone at room temperature. Concomitantly, the surface of CA is fairly hydrophilic with contact angles of 50-55°. It also carries some negative charge from partial hydrolysis of surface ester groups. To facilitate the lift-off of the LBL assembly and realization of the dipping cycle, CA was supported by a glass slide. The glass surface was thoroughly cleaned in a hot H2O2/H2SO4 (1:3) mixture for 5 min. After thorough washing and drying, a few drops of 15% solution of CA in acetone were cast on the slide and allowed to spread, forming a thin uniform coating. Immediately after that, the slide was placed under an inert atmosphere and the solvent was allowed to slowly evaporate. When the film solidified, the traces of acetone were completely removed under vacuum. The LBL assembly was carried out by a cyclic repetition of the following operations: (1) dipping of the CA-coated slide in 1% aqueous solution of poly(dimethyldiallylammonium chloride), 400-500 kDa, P, for 1 min; (2) rinsing
5532
Langmuir, Vol. 16, No. 13, 2000
Figure 2. Magnetization curve obtained for 0.5 × 0.5 cm piece of (M)30 film taken at T ) 298 K.
with deionized water for 1 min; (3) dipping in a solution of negatively charged colloid for 1 min; and (4) rinsing with water. All films in this study were made by using a colloidal solution of negatively charged 8-10 nm magnetite nanoparticles, M, prepared as described elsewhere.3,4 The stability of the colloid originated primarily from the strong electrostatic repulsion of particles and, to a smaller extent, from the physisorption of bulky tetraalkylammonium cations preventing their physical contact. Electrostatic and van der Waals interaction with P layer caused destabilization of the colloid, which made adsorption virtually irreversible. At the same time, the negative charge acquired by the film surface limited adsorption to approximately a monolayer of nanoparticles. On the account of the recurring nature of the deposition process,
Letters
the film produced in n deposition cycles could be referred to as (M)n. One dipping sequence resulted in the addition of a polyelectrolyte-magnetite layer with an average thickness of 8 ( 0.5 nm as established by ellipsometry. This increment remained virtually constant for at least 50 deposition cycles as observed from the linearity of the integral optical density with n.3,4 AFM investigation of (M)1 revealed that such a film was made of densely packed nanoparticles (Figure 1A). The model structure of a polyelectrolyte-magnetite multiplayer can be represented by Figure 1C. It is necessary to point out that in many instances the layered nature of the films is obscured by interpenetration of the subsequent layers and high interlayer roughness. In fact, LBL films made from only two components with the same deposition condition for each layer can often be represented as a homogeneous mixture of, for instance, nanoparticles and polyelectrolyte. After depositing an appropriate number of layers and thorough drying, a thin CA coating with the LBL film was peeled off the glass support and immersed in acetone for 24 h. The CA substrate dissolved leaving a dark colored film freely suspended in the solution, which was transferred into a new portion of acetone to completely wash away remaining CA molecules. As expected, the obtained thin film retained the magnetic properties of nanoparticles: it moved through the solution toward a permanent magnet placed near the side of the beaker manner. The magnetic properties of the films can also be seen in the magnetization curve (Figure 2). It has a characteristic sigmoidal shape with very small hysteresis loop in the low field region. Such magnetization curves are typical for nanocrystalline magnetic materials that become su-
Figure 3. (A and B) XPS spectra of top (left) and bottom (right) surfaces of (C/M)30 film. In the inserts, the enlarged portions of the spectrum with 2p1 and 3p3 peaks of surface Fe atoms are given. (C and D) SEM images of top (left) and bottom (right) surfaces of (C/M)30 film.
Letters
perparamagnetic when particle sizes reach the range of few nanometers. From the suspended state, the films could be transferred onto any solid or porous substrate. However, they were quite fragile, which was not surprising considering that the thickness of prepared films, such as (M)15 and (M)30, was in the range of a few hundred nanometers. Layer-by-layer assembly affords manipulation of the order of the deposited layers. To strengthen the film, one can replace every other layer of magnetite with a layer of exfoliated montmorillonite clay platelets, thereby producing an assembly with a (C/M)n sequence (Figure 1D). Clay platelets have a thickness of 1.0 nm, while extending 150300 nm in other dimensions. On polyelectrolytes, they formed a layer of overlapping alumosilicate sheets (Figure 1B) with an average thickness of 3.8 ( 0.3 nm.16,17 Being adsorbed virtually parallel to the surface of the substrate, their large size allowed them to cover ca. 400 nanoparticles at once, whereby cementing the assembly. (C/M)30 freestanding film prepared following the procedure outlined above could be easily picked up with tweezers, transferred, cut, moved around the solid surface, and handled in any other way. Taking advantage of this architecture, we were able to make a free-standing film with as few as eight repeating C/M units, which was impossible without an alumosilicate framework. Characterizing the free-standing films, it is important to establish the identity of both surfaces of the film to ensure the completeness of the CA removal, which might have contributed to their strength. The scanning electron microscopy (SEM) and XPS data taken on the sides that were facing the solution (top) and CA (bottom) during the deposition revealed complete identity of their surfaces with respect to both composition and relief (Figure 3). In particular, the observation of the Fe 2p1 and Fe 3p3 peaks (at 1211 and 1198 eV, respectively) would have been impossible on the CA part of the film if a CA film of even a few nanometers in thickness were there. The identical intensity of the iron peaks referenced to the intensity of carbon 1s peak clearly indicates the completeness of CA removal, which proves the self-supporting nature of the films obtained after the lift-off. (C/M)30 assembly was embedded in epoxy resin and its cross sections were investigated by optical and transmission electron microscopy. The film can be clearly seen in the optical microscope as a continuous black band of uniform thickness (Figure 4A). Because of light aberrations on objects comparable with the wavelength of light, the actual thickness was determined from electron microscopy image (Figure 4B) and was found to be 350 ( 30 nm. One can also estimate this parameter by adding up the thicknesses of M and C layers. For (C/M)30 film, this gives (3.8 + 8.0) × 30 ) 354 nm, which coincides well with the TEM measurements. It was noticed that the films showed some longitudinal cracks from cross sectioning (Figure 4B). The anisotropy of the physical damage caused by the diamond knife can be attributed to the existence of montmorillonite stacks with sheets oriented in parallel to each other andswith a lower degree of rigorousnesssto the substrate. Note that the alumosilicate platelets are flexible and conform to (17) Kotov, N. A.; Haraszti, T.; Turi, L.; Zavala, G.; Geer, R. E.; Dekany, I.; Fendler, J. H. J. Am. Chem. Soc. 1997, 119, 6821-6832.
Langmuir, Vol. 16, No. 13, 2000 5533
Figure 4. (A) Optical micrograph (3667×) of the (C/M)30 cross section. Two strands of the film can be seen; some parts of the film were folded during imbedding in epoxy resin. (B and C) Transmission electron micrographs (61600× and 81000×, respectively) of (C/M)30 cross sections.
surface topography.16 Their physical properties resemble much better a nanoscale version of fiber glass fabric rather than rigid bricks. The ability of thin inorganic sheets to wrap around the nanoparticles can also be seen in ref 18 for graphite oxide colloids, which would be impossible assuming their stiffness. The films were very homogeneous with no apparent phase boundaries between the components. Their particulate nature can be seen on a especially thin cross section in Figure 4C. In conclusion, we present here a technique for the preparation of free-standing composite films, whose architecture can be controlled in the nanometer scale. This technique can be extended to a variety of other compounds utilized in LBL research, such as polymers, proteins, dyes, metal and semiconductor nanoparticles, vesicles, viruses, DNA, and others. Different colloids will lead to a palette of thin film materials and membranes whose functional properties can be tuned by varying the layer sequence. The effect of strengthening of the films by adding a molecular framework made from overlapping alumosilicate sheets, which can be traced in the ability to lift-off films of different architecture, will be reported in a separate publication. LA000560B (18) Cassegneau, T.; Fendler, J. H. J. Phys. Chem. B 1999, 103, 17891793.
