T2F.004
SUPERPARAMAGNETIC PHOTOSENSITIVE POLYMER NANOCOMPOSITE FOR MICROACTUATORS 1
M. Suter1, S. Graf1, O. Ergeneman2, S. Schmid1, A. Camenzind3, B. J. Nelson2 and C. Hierold1 Micro and Nanosystems, 2Institute of Robotics and Intelligent Systems, 3Particle Technology Laboratory Department of Mechanical and Process Engineering, ETH Zurich, Switzerland
ABSTRACT We present a photosensitive polymer composite with superparamagnetic characteristics for the fabrication of microactuators. A uniform distribution of particles in the photosensitive polymer matrix SU-8 is achieved by applying superparamagnetic nanoparticles with the aid of a surfactant. The composite contains Fe3O4 nanoparticles up to 3 vol.% (12 wt.%) with diameters of around 13 nm. Superparamagnetic composite microcantilevers are successfully fabricated and actuated in resonance by the magnetic field of an external coil.
Ni F
F
ferromagnetic Ni
Mr
Fe3O4 superparamagnetic Fe3O4
KEYWORDS Superparamagnetic composite, photosensitive composite, magnetic polymer, magnetic SU-8, magnetic resonant structure, magnetic cantilever, magnetic microactuator
INTRODUCTION
Figure 1: Magnetization measurements of pure ferromagnetic Ni nanoparticles with (dNi_APS § 20 nm) and pure superparamagnetic Fe3O4 nanoparticles (dFe3O4_TEM § 13 nm) taken by a vibrating sample magnetometer at room temperature. Ni particles have a high remanent magnetization Mr and agglomerate in a low-viscosity dispersion due to strong magnetic forces between particles. In contrast, superparamagnetic Fe3O4 particles have negligible remanent magnetization, and therefore low magnetic interaction in dispersion as the direction of the magnetization of particles flips by thermal energy.
Polymers are promising materials for microactuators due to their low Young’s modulus compared to silicon or metals. An advantage of polymer actuators is that large deflection can be achieved with an applied external force. Polymers have favorable chemical properties and allow for simple fabrication processes such as direct structuring by photolithography. This makes the final devices more costeffective than conventional Si-based actuators [1]. The surface properties of polymers can also be chemically engineered if functionalization is required. Magnetic actuators can be actuated from large distances (few milimeters) which makes them attractive for wireless applications. Furthermore, they can have larger actuation forces than electrostatic microactuators [2]. Additionally, they can be operated in a variety of environments. Magnetic polymer composites are promising materials for microactuators. They combine the advantages of polymers and magnetic materials which lead to actuators with large deflections and the possibility of wireless actuation [3]. Photosensitive polymers with incorporated magnetic particles present a novel class of materials which allows for low-cost fabrication of microstructures with few process steps. Dispersing magnetic particles in a photosensitive polymer matrix has been investigated by several groups [4, 5, 6]. Most of them have used ferromagnetic particles.
978-1-4244-4193-8/09/$25.00 ©2009 IEEE
NANOPARTICLE EVALUATION For the fabrication of a composite to be used in magnetic microstructures on the order of < 3 ȝm, the particles have to be much smaller than the smallest structure dimensions. They also need to be well dispersed to obtain uniform mechanical and magnetic properties. For the fabrication of thin layers (< 3 ȝm) by spin coating, the viscosity of the composite must be low (< 50·10-6 m2/s). Ferromagnetic and superparamagnetic particles have been investigated for their suitability for the polymer composite. The superparamagnetic phenomenon for magnetic particles occurs below a critical, material-dependent particle size, when the thermal energy exceeds the magnetic anisotropy energy [7]. At room temperature the Fe3O4 particles have to be smaller than ~ 20 nm. Ferromagnetic Ni particles from NanoAmor (Houston, USA), with a diameter dNi_APS of 20 nm (specified by the supplier) and superparamagnetic Fe3O4 particles from Chemicell GmbH with a diameter dFe3O4_TEM of around 869
Transducers 2009, Denver, CO, USA, June 21-25, 2009
Authorized licensed use limited to: ETH BIBLIOTHEK ZURICH. Downloaded on June 14,2010 at 09:33:54 UTC from IEEE Xplore. Restrictions apply.
13 nm, measured by transmission electron microscopy (TEM), were characterized by a vibrating sample magnetometer (MicroMag, Model 3900). Ferromagnetic Ni particles (dash-doted line in Figure 1) indicate a saturation magnetization of ~ 600 kA/m and a remanent magnetization Mr of 210 kA/m. Due to the high remanent magnetization these particles tend to agglomerate by magnetic attraction in a liquid polymer matrix. A viscosity-increasing agent can be used to reduce this particle agglomeration [5]. However, for application by spin coating the polymer needs a low viscosity. In contrast to ferromagnetic particles, superparamagnetic particles do not retain any remanent magnetization. The solid line in Figure 1 shows the magnetic measurements of superparamagnetic Fe3O4 particles indicating negligible remanent magnetization. Therefore the particles exhibit no magnetic attraction during dispersion in a low-viscosity polymer. Ferromagnetic Ni particles have three times higher saturation magnetization than Fe3O4 particles but form bigger agglomerates (in the micrometer range) in lowviscosity materials due to magnetic attraction. Therefore, due to negligible magnetic particle interaction, superparamagnetic particles are more suitable to fabricate a magnetic composite which can be used for microstructures.
Figure 2: Light microscope images of superparamagnetic Fe3O4 nanoparticles dispersed in a SU-8 matrix. a) Without a surfactant, the nanoparticles form macro structures. b) With a surfactant, particles disperse more uniformly.
The composite magnetization shown in Figure 4 has been characterized with a superconducting quantum interference device (SQUID) (MPMS, Quantum Design). The composite has nearly zero remanent magnetization indicating superparamagnetic behavior and therefore negligible particle interactions. The saturation magnetization of the composite with 1 vol.% particle loading is 2.7 kA/m and in agreement with the approximately 100 times higher magnetization of 200 kA/m for 100 vol.% Fe3O4 particles (Figure 1). The slightly higher saturation magnetization in the composite could be explained by the imperfect determination of the nanoparticle concentration in the initial dispersion, or by a magnetic interface effect between nanoparticles and surfactant/polymer. In a nanoparticle, a large percentage of the atoms are surface atoms, which implies surface effects become more important [7].
FABRICATION AND CHARACTERIZATION OF THE COMPOSITE Fabrication For the fabrication of the composite, a stabilized nanoparticle dispersion of Fe3O4 nanoparticles (as described above) in Ȗ-Butyrolacton (GBL) from Chemicell GmbH is used. The dispersion is mixed in high-viscosity photosensitive SU-8 (with solvent GBL) from MicroChem Corp. using a planetary mixing (dual asymmetric) centrifuge and ultrasonic steps [8]. Even without magnetic interaction between the particles, the nanoparticles tend to form agglomerates to reduce the energy associated with the high surface area to volume ratio of the nanosized particles [7]. Therefore, a surfactant is used to obtain steric stabilization of the superparamagnetic Fe3O4 nanoparticles in the SU-8 matrix. Figure 2 a) and b) show TEM images of the composite with and without surfactant. Characterization The cross-section of the composite can be seen in the TEM image (Figure 3) showing a uniform dispersion with a mean particle agglomeration size of approximately 150 nm (determined by TEM). At the matrix-air interface (top surface) the particle density is slightly higher. This is probably caused by diffusion and dry-out effects of the nanoparticles in the spin-coated film.
Figure 3: TEM image of the cross-section of the spin coated polymer composite with 1 vol.% particle loading shows particle agglomerations with a mean size of approximately 150 nm.
870
Authorized licensed use limited to: ETH BIBLIOTHEK ZURICH. Downloaded on June 14,2010 at 09:33:54 UTC from IEEE Xplore. Restrictions apply.
Figure 6: SEM image of a microcantilever made of magnetic SU-8 polymer composite with 1 vol% Fe3O4 particle content.
Figure 4: Magnetization measurements of the composite filled with 1 vol.% Fe3O4 particles taken with a SQUID at room temperature. The absence of remanent magnetization indicates superparamagnetic behavior.
The strong UV absorption of Fe3O4 nanoparticles hinders the crosslinking reactions in lower regions of the polymer layer during photolithography. This limits the particle loading of the composite for a given composite thickness. For a thickness of 3 ȝm the limit of particle content was found to be 3 vol.%. A higher particle loading leads to adhesion problems and results in peeling of the composite structure.
FABRICATION OF CANTILEVERS The fabrication of microcantilever structures with the superparamagnetic composite is based on a sacrificial layer process using lift-off resist (LOR 30B from MicroChem Corp.) and a conventional photolithography step as illustrated in Figure 5. First, a layer of LOR with a thickness of 6 ȝm is spin coated on a glass wafer. After spin coating of a 2.5 ȝm thick photosensitive magnetic polymer layer, the film is patterned by photolithography with a dose of 800 – 1600 mJ/cm2, depending on the particle content in the composite. The illuminated composite is then developed with MR-Dev 600 (MicroChem Corp.). The sacrificial layer is etched with Microposit 351 Developer (Shipley Company) followed by a critical point dryer step that is required to release the microcantilever structure and prevent it from sticking to the substrate. Figure 6 shows a SEM image of a released composite cantilever loaded with 1 vol.% (4 wt.%) of Fe3O4 particles. The typical dimensions of the fabricated cantilevers are w = 14 ȝm, t = 2.5 ȝm and L = 100 – 200 ȝm.
ACTUATION OF CANTILEVERS Experimental To demonstrate the actuation of the magnetic composite, an actuation force acting on the cantilever was generated by an inhomogeneous magnetic field from an external magnetic coil. The force F (Equation 1) acting on a superparamagnetic composite is calculated with the volume of the nanoparticles, VNP [m3], the magnetization, M [A/m], and the magnetic field, H [A/m], where ȝ0 [N/A2] is the magnetic permeability of free space [9].
F = µ0VNP M (H ) ⋅ (∇ H )
(1)
To achieve a high magnetization, M(H), and a high magnetic field gradient, ∇ H, the design of the coil was optimized regarding the skin effect and self inductance for the frequency range of 1–100 kHz by FEM simulations and analytical calculations. The generated magnetic field, H, from the coil (Rin = 2 mm, Rout = 6 mm, h = 7 mm, diameter of the wire = 0.4 mm) is proportional to the current and can reach up to 3.7 kA/m (RMS) at the position of the cantilever. As illustrated in Figure 7, the chip, which contains the cantilever, is located above the coil and placed into a vacuum chamber (p = 7.4ǜ10-3 Pa) to minimize air damping. The measurement of the cantilever was performed with a laser-Doppler vibrometer (Polytec, MSA-400).
Figure 5: Schematic of the process-flow used to fabricate suspended magnetic cantilevers: 1) Spin coating of LOR layer and photosensitive magnetic polymer composite layer, UV Exposure, 2) Development of composite, 3) Sacrificial layer etch
871
Authorized licensed use limited to: ETH BIBLIOTHEK ZURICH. Downloaded on June 14,2010 at 09:33:54 UTC from IEEE Xplore. Restrictions apply.
ACKNOWLEDGEMENTS
laser-Doppler vibrometer
The authors would like to thank Elisabeth Müller from EMEZ for the TEM images and Prof. Bertram Batlogg’s group for the SQUID measurements. We also acknowledge Ann Marie Hirt for the support with the VSM measurements and Prof. Sotiris E. Pratsinis for using the equipment in PTL. Thanks to Michael Wendlandt for his fundamental inputs and Florian Umbrecht for helpful discussions. (All are from ETH Zurich). Financial support by the ETH Zurich (TH-28 06-3) and SNSF (Project No. 200020-113350) are gratefully acknowledged.
cantilever-chip
A HCoil ~ ICoil
REFERENCES [1]
A. Johansson, G. Blagoi, and A. Boisen, "Polymeric cantilever-based biosensors with integrated readout," Applied Physics Letters, vol. 89, p. 173505, 2006 [2] I. J. Busch-Vishniac, "The case for magnetically driven microactuators," Sensors and Actuators A: Physical, vol. 33, pp. 207-220, 1992 [3] Y. Yamanishi, Y. C. Lin, and F. Arai, "Magnetically modified PDMS microtools for micro particle manipulation," IEEE/RSJ International Conference on Intelligent Robots and Systems, Vols 1-9 New York: IEEE, pp. 759-764, 2007 [4] N. Damean, B. A. Parviz, J. N. Lee, T. Odom, and G. M. Whitesides, "Composite ferromagnetic photoresist for the fabrication of microelectromechanical systems," Journal of Micromechanics and Microengineering, vol. 15, pp. 29-34, Jan 2005 [5] K. Kobayashi and K. Ikuta, "Three-dimensional magnetic microstructures fabricated by microstereolithography," Applied Physics Letters, vol. 92, p. 3, Jun 2008 [6] M. Feldmann and S. Buttgenbach, "Novel microrobots and micromotors using Lorentz force driven linear microactuators based on polymer magnets," IEEE Transactions on Magnetics, vol. 43, pp. 3891-3895, Oct 2007 [7] A. H. Lu et al., "Magnetic nanoparticles: Synthesis, protection, functionalization, and application," Angewandte Chemie-International Edition, vol. 46, pp. 1222-1244, 2007 [8] H. Schulz, B. Schimmoeller, S. E. Pratsinis, U. Salz, and T. Bock, "Radiopaque dental adhesives: Dispersion of flame-made Ta2O5/SiO2 nanoparticles in methacrylic matrices," Journal of Dentistry, vol. 36, pp. 579-587, Aug 2008 [9] J. J. Abbott, O. Ergeneman, M. Kummer, A. M. Hirt, B. J. Nelson, "Modeling Magnetic Torque and Force for Controlled Manipulation of Soft-Magnetic Bodies", IEEE Transactions on Robotics, vol. 23, No. 6, pp. 1247-1252, December 2007
Figure 7: Schematic illustration of the measurement setup. The magnetic composite cantilever is actuated by the alternating magnetic field HCoil of an external coil. The resonant frequency of the cantilever is measured by a laser-Doppler vibrometer. The photo in the right upper corner shows the fabricated coil with the cantilever chip on top.
Results A fabricated composite microcantilever was successfully actuated with the external coil. Table 1 shows the measurement data of the magnetic actuated cantilever with a particle loading of 3 vol.% (12 wt.%). The cantilever is excited directly at a single frequency (in resonance, 16.8 kHz) by a sinusoidal current in the coil. A frequency sweep and Q factor determination has not yet been performed because of the frequency dependent induction of the coil. Table 1: Measurement data obtained from a laser-Doppler vibrometer of the superparamagnetic composite cantilever with 3 vol.% of Fe3O4 nanoparticles actuated at resonance (f = 16.8 kHz, HCoil § 3.7 kA/m) with the setup described in Figure 7 in vacuum (p = 7.4ǜ10-3 Pa). Amplitude fR measured Cantilever dimensions IRMS coil L = 189ȝm 500 mA 20 nm 16.8 kHz w = 17 ȝm t = 2.5 ȝm
CONCLUSION A photosensitive polymer composite with superparamagnetic characteristics was developed for the fabrication of magnetic microactuators. Superparamagnetic Fe3O4 nanoparticles are used to obtain a uniform dispersion of the particles in the composite. The limiting factor for the filler loading and layer thickness of the composite are given by the strong UV absorption of the Fe3O4 nanoparticles. This limits the usage of the composite layers to the range of few micrometers and a loading up to 3 vol.%, which are suitable for small microstructures like cantilevers. Deflection measurement of the fabricated microcantilever approve that such composite microstructures can be actuated off-chip by the magnetic field of an external coil.
CONTACT *M. Suter, tel: +41-44-632-55-32
[email protected] 872
Authorized licensed use limited to: ETH BIBLIOTHEK ZURICH. Downloaded on June 14,2010 at 09:33:54 UTC from IEEE Xplore. Restrictions apply.
