3180
OPTICS LETTERS / Vol. 30, No. 23 / December 1, 2005
Fabrication of three-dimensional structures by three-photon polymerization M. Farsari, G. Filippidis, and C. Fotakis Institute of Electronic Structure and Laser, Foundation for Research and Technology—Hellas, P.O. Box 1527, 71110 Heraklion, Crete, Greece Received June 7, 2005; revised manuscript received July 11, 2005; accepted July 30, 2005 We report the fabrication of three-dimensional structures of submicrometer resolution by three-photon polymerization. This resolution has been achieved by polymerizing ORMOCER, a UV photocurable organic– inorganic hybrid material, with an ultrafast laser irradiation at 1028 nm. To our knowledge, this is the first demonstration of three-photon polymerization, a process that may allow the fabrication of components of very high resolution. © 2005 Optical Society of America OCIS codes: 140.7090, 160.5470, 220.4000, 270.4180.
Nonlinear optical microstereolithography based on multiphoton polymerization of polymeric mixtures allows the fabrication of three-dimensional (3D) structures with a submicrometer resolution. When the beam of an ultrafast infrared laser is tightly focused into the volume of a photosensitive material, the polymerization process can be initiated by nonlinear absorption within the focal volume. By moving the laser focus three dimensionally through the resin, 3D structures can be fabricated. The technique has been used with a variety of acrylate and epoxy materials,1–5 and several components and devices have been fabricated, such as photonic crystals,5,6 microrotors driven by laser tweezers,7 mechanical devices,8,9 and microscopic models.10,11 The highest resolution of 120 nm was reported by Tanaka et al.9 Until now the phenomenon that has been employed is two-photon polymerization by use of Ti:sapphire femtosecond lasers operating at 750– 800 nm. The only exception is Straub and co-workers,4,10 who use a Ti:sapphire laser combined with an optical parametric oscillator to achieve a pump wavelength of 600 nm. In this Letter we show that 3D devices can be built with three-photon polymerization. For this purpose, we employ a compact femtosecond oscillator operating at 1028 nm and a UV photocurable organic– inorganic hybrid known as ORMOCER. To our knowledge, this is the first time three-photon polymerization has been demonstrated, a process that may lead to the construction of components of even higher resolution than those reported to date. The material used for the fabrication of the 3D nanostructures is the organic–inorganic hybrid ORMOCER (Micro Resist Technology), a material developed for optical applications that shows high transparency in the visible and near-infrared ranges. ORMOCER contains a highly crosslinkable organic network and inorganic components leading to high optical quality and mechanical and thermal stability. The polymerization process is initiated by the reaction of the radical photoinitiator Irgacure 369 (Ciba), which is the same as the one used by Straub in his two-photon polymerization experiments at 0146-9592/05/233180-3/$15.00
600 nm.4,10 ORMOCER has been extensively studied as a material for two-photon polymerization, and several devices have been made with them, the most notable being a photonic crystal built up from individual rods with a diameter of 200 nm and a spacing between the rods of 250 nm.11–13 It is a material interesting not only because of its properties as an optical material but also because it is biocompatible.12 The setup for the fabrication of 3D microstructures by three-photon microstereolithography is shown in Fig. 1. The laser used was an Amplitude Systems t-pulse laser femtosecond oscillator operating at 1028 nm. This source is a compact diode-pumped femtosecond laser oscillator delivering a train of high-energy, short-duration pulses. The average power of the laser is 1 W, a pulse duration of less than 200 fs, and a repetition rate of 50 MHz. The photopolymerized structure was generated by using an x – y-galvanometric mirror digital scanner (Scanlabs Hurryscan II), controlled by SAMLight (SCAPS) software. The scanner has been adapted to accommodate a high-numerical-aperture focusing microscope objective (Edmund Scientific 100⫻, N.A. 1.25) with immersion oil being used for index matching 共noil = 1.515兲. The beam waist of the focused laser
Fig. 1. Optical system for three-photon microfabrication. © 2005 Optical Society of America
December 1, 2005 / Vol. 30, No. 23 / OPTICS LETTERS
Fig. 2. Computer design of the component built.
beam is given by r = 0.61 / N.A.= 500 nm, and its movement on the z axis is achieved with a single-axis piezoelectric stage (PI). A mechanical shutter (Uniblitz) was used for beam control. The piezoelectric stage and the shutter were computer controlled by a National Instruments LabVIEW program. Beam intensity control was achieved by using neutral-density filters. A CCD camera is mounted behind a dichroic mirror for online monitoring of the three-photon polymerization process. This is possible as the refractive index of the originally liquid ORMOCER changes during polymerization, so that the illuminated structures become visible during the building process. The structures were fabricated layer by layer from the bottom up, with the last layer attached to the coverslip. As the material is very viscous and the building process lasts less than 1 min, there are no issues with material drifting. Figure 2 shows the design of a component built by three-photon polymerization. It consists of five step-in squares, which serve to build a robust support structure, and four vertical and horizontal lines, which serve as cell dividers. The galvo scanning speed was 750 m / s. The same layer was repeated 15 times, with the piezoelectric stage moving the sample lower by 1 m after each layer. For the same galvo speed, several laser powers were investigated. The lowest laser energy per pulse that appeared to polymerize the material was found to be 0.8 nJ, which corresponds to a laser fluence of 0.1 J / cm2. When the laser energy per pulse was increased to more than 1 nJ, the material showed burn marks and was irreversibly damaged. The components presented in this Letter were made by using a laser pulse energy of 0.8 nJ. After the completion of the component build process, the sample was developed for 3 min in a 50:50 solution of isopropanol:4-methyl-2-pentanone and rinsed in isopropanol. The samples were not UV postcured. Following the washout the samples were coated with a few-nanometer-thick palladium layer, and their structural properties were investigated in a scanning electron microscope (SEM). The results are depicted in Figs. 3–5. Figure 3 shows a 1500⫻ magnification of the component where the effect of distortion due to polymer shrinkage can be clearly seen. Figure 4 shows a
3181
5000⫻ magnification of the central area of the component. The image is tilted so that the third dimension of the component is visible. Figure 5 shows a 20,000⫻ magnification of the same component. It can be seen that the resolution is at least 500 nm, which is the same as the focused beam waist. This is a clear indication that the effect is nonlinear and not thermal. The lack of exact laser pulse control combined with the small range of fluences available did not allow extensive studies of the component resolution versus fluence, which would be required in order to conclusively prove that the components were built by threephoton polymerization. However, the other two possibilities, two-photon polymerization or thermal curing, can be disregarded. Two-photon polymerization can be excluded because of the absorption properties of the material. ORMOCER containing the Irgacure 369 photoinitiator is a material designed to be highly transparent in the visible range and particularly at 514 nm, which
Fig. 3. SEM image of the component built by three-photon microstereolithography. The component was scanned using 15 layers at 750 m / s and layer spacing of 1 m. The laser pulse energy used was 0.8 nJ, which corresponds to fluence of 0.1 J / cm2.
Fig. 4. Magnification of the central area of the component of Fig. 3.
3182
OPTICS LETTERS / Vol. 30, No. 23 / December 1, 2005
Fig. 5. Magnification of a feature of the component of Fig. 3, where the 500 nm resolution can be seen.
In conclusion, we have built components employing three-photon polymerization by using a compact femtosecond laser source operating at 1028 nm. The material we used is ORMOCER, a commercially available UV photosensitive material that has been extensively studied for its two-photon photopolymerizable properties and has been used in the construction of photonic crystals and miniature models of very high resolution. The demonstration for the first time, to our knowledge, of three-photon polymerization opens the road to achieve even higher resolution with this and other similar materials, and, with our current experimental setup, 500 nm resolution was achieved. M. Farsari thanks B. N. Chichov and A. Ovsianikov of the Laser Zentrum Hannover e.V. for many helpful discussions. M. Farsari’s e-mail address is
[email protected]. References
is the wavelength for two-photon polymerization with our laser.13 In contrast, the absorption is approximately 300 dB/ cm at 343 nm, which is the threephoton polymerization wavelength with our laser. The possibility of thermal curing can never be totally excluded; however, the components constructed have a resolution of 500 nm. Considering that the waist of the focused beam is 500 nm, it is very unlikely that structures of such high resolution can be made as a result of a thermal process. While one would expect the resolution reached by three-photon polymerization to overtake that currently reported to have been achieved by two-photon polymerization, in fact our results lag behind. The two main contributing factors to this reduction in resolution is the comparatively longer wavelength of the laser and that the microscope objective used had a numerical aperture N.A. 1.25, and not N.A. 1.4 as, for example, in Refs. 9–12. The laser–objective combination gives us a lateral resolution of r = 0.61 / N.A.= 500 nm. In comparison, Serbin et al.,12 who use the same material as in this Letter—a Ti:sapphire laser and an objective with N.A.= 1.4—have a lateral resolution of r = 0.61 / N.A.= 340 nm.
1. S. Maruo, O. Nakamura, and S. Kawata, Opt. Lett. 22, 132 (1997). 2. S. Kawata, H. B. Sun, T. Tanaka, and K. Takada, Nature 412, 697 (2001). 3. X. M. Duan, H. B. Sun, K. Kaneko, and S. Kawata, Thin Solid Films 453, 518 (2004). 4. L. H. Nguyen, M. Straub, and M. Gu, Adv. Funct. Mater. 15, 209 (2005). 5. M. Deubel, G. v. Freymann, M. Wegener, S. Pereira, K. Busch, and C. M. Soukoulis, Nat. Mater. 3, 444 (2004). 6. H.-B. Sun, S. Matsuo, and H. Misawaa, Appl. Phys. Lett. 74, 786 (1999). 7. P. Galajda and P. Ormosa, Appl. Phys. Lett. 78, 249 (2001). 8. H.-B. Sun, K. Takada, and S. Kawata, Appl. Phys. Lett. 79, 3173 (2001). 9. T. Tanaka, H.-B. Sun, and S. Kawata, Appl. Phys. Lett. 80, 312 (2002). 10. M. Straub, L. H. Nguyen, A. Fazlic, and M. Gu, Opt. Mater. 27, 359 (2004). 11. J. Serbin, A. Ovsianikov, and B. Chichkov, Opt. Express 12, 5221 (2004). 12. J. Serbin, A. Egbert, A. Ostendorf, B. N. Chichkov, R. Houbertz, G. Domann, J. Schulz, C. Cronauer, L. Fröhlich, and M. Popall, Opt. Lett. 28, 301 (2003). 13. R. Houbertz, L. Frohlich, M. Popall, U. Streppel, P. Dannberg, A. Brauer, J. Serbin, and B. N. Chichkov, Adv. Eng. Mater. 5, 551 (2003).
