Exceeding The Diffraction Limit With Single-photon Photopolymerization And Photo

Transcript

Exceeding the diffraction limit with single-photon photopolymerization and photo-induced termination Benjamin A. Kowalskib Timothy F. Scotta, Christopher N. Bowmana, , Amy C. Sullivanc and Robert R. McLeodb* a Department of Chemical and Biological Engineering, University of Colorado, UCB 424, Boulder, CO, 80309; b Department of Electrical and Computer Engineering, University of Colorado, UCB 425, Boulder, CO, 80309; c Department of Physics, University of Colorado, UCB 390, Boulder, CO 80309; ABSTRACT The fabrication of 3D microstructures has been realized by numerous researchers using two-photon polymerization. The premise of these studies is that the confinement provided by localized, two-photon absorption results in polymerization only near the focal point of the focused write beam and unwanted polymerization due to superposition of the out-offocus exposures is significantly reduced, enabling the fabrication of complex structures with features below the diffraction limit. However, the low cross-section of two-photon absorbers typically requires excitation by pulsed Ti:Sapphire laser at 800 nm, resulting in polymerized features that are actually larger than those created by one-photon absorption at half the wavelength. Here we describe a single photon photolithographic technique capable of producing features not limited by the physics of diffraction by utilizing a resin which is able to be simultaneously photoinitiated using one wavelength of light and photoinhibited using a second wavelength. Appropriate overlapping of these two wavelengths produces feature sizes smaller than the diffraction limit and reduces polymerization in the out-of-focus regions while avoiding the high light intensities demanded by multi-photon initiation. Additionally, because the photoinhibiting species are non-propagating radicals which recombine when the irradiation is ceased, memory effects typical of photochromic initiators are avoided, allowing rapid and arbitrary patterning. Keywords: Two photon polymerization, photochromic polymers, lithography 1. BACKGROUND The fabrication of 3D, sub-micron structures has been realized by numerous researchers using two-photon polymerization1,2,3,4. The premise of these studies is that the confinement provided by localized, two-photon absorption results in polymerization only near the focal point of the focused write beam and unwanted polymerization due to superposition of the out-of-focus exposures is significantly reduced, enabling the fabrication of complex structures with feature sizes below the one-photon diffraction limit of the wavelength employed. However, the small two-photon crosssection prevents application of these materials in many applications including large-area exposures with proportionally lower intensities and high-speed applications such as data storage. An approach to significantly increasing the two-photon cross-section is resonant enhancement in which the intermediate state is metastable. An extreme example of this approach is photochromic initiation in which the intermediate “colored” state thermally decays at a rate typically measured in minutes and the two photons are usually of different wavelengths5. The separation of colors provides greater freedom to design the 3D exposure geometry6 but is often inappropriate for rapid 3D exposure. This is because as the minimum time-scale of multiple exposures becomes less than the lifetime of the intermediate state, the confinement of the material response shifts from that of two-photon towards that of one-photon in a geometry-dependent manner. * [email protected] Organic 3D Photonics Materials and Devices II, edited by Susanna Orlic, Proc. of SPIE Vol. 7053, 70530E, (2008) · 0277-786X/08/$18 · doi: 10.1117/12.796978 Proc. of SPIE Vol. 7053 70530E-1 Downloaded from SPIE Digital Library on 07 Jun 2012 to 198.11.25.124. Terms of Use: http://spiedl.org/terms The microscopy community has developed many methods for exceeding the traditional diffraction limit, including the use of two or more photon absorption. Single-photon methods include near-field optics in which the sample to be imaged (for microscopy) or modified (for lithography) must be placed less than a wavelength from the optics. This is inconvenient for many sample geometries and impossible in three-dimensional fabrication. A two-color, one-photon absorption method that does not suffer this limit is stimulated-emission-depletion (STED) fluorescence7. In this scheme, the first color excites fluorescence in diffraction-limited spot which is then spatially refined by a second color which depletes the fluorescence in the periphery of the original spot. The resulting region of fluorescence is therefore reduced to much smaller than the diffraction limit. Here we describe a photopolymer material and lithography method inspired by the STED scheme, although using very different photochemistry. In our scheme, shown in Figure 1, a standard photoinitiator generates polymerizing radicals in response to a first color (473 nm). Simultaneously, a second color (364 nm) photo-generates terminating radicals in a Gauss-Laguerre “donut” pattern, strongly suppressing polymerization at the periphery of the diffractionlimited region. The resulting polymerized feature can be significantly smaller than the diffraction limit both transversely and in depth. The latter is important for suppression of out-of-focus exposure during dense 3D lithography or data storage. 473 nm 364 nm Gauss-Laguerre VIHY Inhibition Initiation Polymerization Gauss-Laguerre binary hologram Figure 1. . Optical layout (left) for demonstration of polymer recording at less than the transverse diffraction limit. 364 nm light from an Argon:ion laser is transformed into a Gauss-Laguerre “doughnut” mode (see inset in plot) by a binary amplitude hologram. This inhibiting halo is combined with the 473 nm writing stylus and the two are focused into the photopolymer volume by a single objective. The plot on the right shows a representative initiation profile, inhibition profile and final polymer profile (see Model section). This approach has a number of advantages for two and three-dimensional lithography. First, the response at the focus is related to the difference of the two intensities (see below), resulting in significantly smaller features than twophoton or photochromic initiation in which the response is proportional to the product of two intensities. Second, similar to photochromic initiation, the two intensities can be individually shaped, resulting in significant flexibility to shape the recorded feature. Third, the method provides independent control over feature size and the degree of polymerization which is critical for mechanical robustness in 3D nanofabrication. Finally and most critically, the individual processes are high-efficiency one-photon absorptions but without a corresponding requirement of long lifetimes. The system thus has the potential to create features significantly smaller than the diffraction limit at high speed and in thick volumes. 2. MATERIALS The key components of the material system are the photoinitiator and photoinhibitor. A primary requirement is that their absorption spectra can be excited independently, as shown in Figure 2. Application of rapid 3D microlithography also requires that the photoinhibiting species be rapidly eliminated in the absence of the corresponding exitation wavelength. Proc. of SPIE Vol. 7053 70530E-2 Downloaded from SPIE Digital Library on 07 Jun 2012 to 198.11.25.124. Terms of Use: http://spiedl.org/terms Such a photoinhibition scheme may be provided by non-initiating radicals produced from the photolysis of an otherwise inert photoinhibitor; however, these radicals are able to rapidly recombine with and terminate the growing polymer chain. Additionally, unlike the propagating radicals, the inhibiting radicals may be small and their termination kinetics can then be unconstrained by reaction diffusion. As a result, they are very short-lived, rapidly recombining with each other, thus their concentration drops precipitously in the absence of photolysing irradiation. Ideally, both the photoinhibitor and the species formed from recombination reactions are inert, not interfering with the polymerization reaction so there is no photoinhibition memory effect after the photolysing irradiation has ceased. 3.0 TED CQ × 10 2.5 Absorption 2.0 1.5 1.0 0.5 0.0 300 350 364 400 450 473 500 550 600 Wavelength [nm] Figure 2. UV-VIS absorption spectra of the photoinitiator (CQ) and the phototerminator (TED) with the two laser lines marked for convenience. Note that the CQ spectrum has been scaled by 10 for ease of comparison. Both are measured at 0.1w/v% in CHCl3. A demonstration material that matches these requirements has been designed and tested. The monomer, triethylene glycol dimethacrylate (TEGDMA), can be cured to form a crosslinked, gelled polymer via a free radical mechanism. The camphorquinone (CQ)/ethyl 4-(dimethylamino)benzoate (EDAB) photoinitiation system was chosen in combination with tetraethylthiuram disulfide (TED) as the photoinhibitor for this study as CQ has an absorbance peak around 470 nm (visible, blue) and absorbs poorly in the near UV whereas TED does not absorb at 470 nm while it does absorb strongly in the near UV (see Figure 2). Irradiating TEGDMA formulated with CQ/EDAB and TED with blue light excites the CQ and initiates the polymerization via carbon centered radicals whereas irradiation with UV photocleaves the TED, producing sulfur centered dithiocarbamyl (DTC) radicals8 which can terminate the polymerization. The feasibility of this system is demonstrated in Figure 3. TED is not an ideal photoinhibitor as chain transfer through its sulfur-sulfur bond dramatically slows the polymerization while the photoiniferter dithiocarbamate species formed are able to reinitiate polymerization under UV irradiation. Despite these imperfections, the figure shows how the presence of UV light rapidly decreases the polymerization induced by the blue light. The six-fold maximum contrast is sufficient to constrain polymerization to significantly below the one-photon diffraction limit, as we show in the following sections. Proc. of SPIE Vol. 7053 70530E-3 Downloaded from SPIE Digital Library on 07 Jun 2012 to 198.11.25.124. Terms of Use: http://spiedl.org/terms 20 Conversion rate [%/min] 18 16 14 12 ~6× reduction in conversion rate. 10 8 6 4 2 0 0 20 40 60 80 100 UV [mW/cm^2] Figure 3. Methacrylate conversion due to 473 nm initiation versus the intensity of the inhibiting UV light. This material supports a 6fold reduction of polymerization rate as the UV intensity is increased from zero to 20 mW/cm2. 3. MODEL A complete model of the chemical kinetics of the multiple interacting species under intense, localized two-color exposure is beyond the scope of this paper. Instead, we develop a simplified model in order to illustrate and motivate the properties of the material and the lithography. We start with a simplified set of rate equations for the radical initiator [Init*] and radical terminator [Term*], each generated independently by the two intensities I473 and I364. Ignoring nonidealities such as chain transfer, each radical is presumed to decay via unimolecular and bimolecular pathways and, of course, via the intended termination process: [ ] [ ] [ ] [ ][ ] 2 ∂ Init * = k Init [Init ]I 473 − k 2, Init Init * − k 1, Init Init * − k IT Init * Term * ∂t . 2 ∂ Term * = k Term [Term]I 364 − k 2,Term Term * − k 1,Term Term * − k IT Init * Term * ∂t [ ] [ ] [ ] [ ][ ] (1) Even in this simplified model, the coupling between the two excited species is complex and governed by a number of rate constants k that are not well known. To simplify the situation, let us focus on the most interesting spatial regions where the initiating light is intense and therefore the density of initiating radicals is high. The reaction of the terminating radicals and initiating radicals (the final term in the second equation) will then be the dominant term that removes terminating radicals, allowing us to ignore the second and third terms which describe the decay pathways of terminating radicals. Further making the typical steady-state assumption allows the two equations to be combined as: [ 0 = k Init [Init ]I 473 − k Term [Term]I 364 − k 2, Init Init * ] 2 [ ] − k 1, Init Init * . (2) This quadratic equation for the initiating radical density [Init*] is driven by the scaled difference of the two optical intensities, assuming the unexposed concentrations [Init] and [Term] are never significantly depleted. While the assumptions made are too strong to make it accurate under all conditions, this illustrates the fundamentals of the process. It is common in the photopolymer field to approximate the solution to this quadratic equation as a power law: ~ ~ α Init * ∝ I 473 − I 364 , [ ] ( ) (3) ~ ~ where I 473 and I 364 normalized intensities and α would theoretically vary from ½ for pure bimolecular termination (k2,Init >> k1,Init) to 1 for pure unimolecular termination of the initiating radicals. Note that even though simplistic, the Proc. of SPIE Vol. 7053 70530E-4 Downloaded from SPIE Digital Library on 07 Jun 2012 to 198.11.25.124. Terms of Use: http://spiedl.org/terms conversion data in Figure 3 shows a roughly linear decrease in conversion rate with UV intensity until saturation occurs. Based on this data, we add a phenomenological saturation to equation 3, limiting suppression of initiating radicals to a maximum factor of 6 (as shown in Figure 1, right), and take α ≈ 1. The latter simplification has only a mild impact on the size and shape of polymerized features in the cases illustrated below, so is a reasonable simplification for qualitative exploration. Blue Two—Photon a b c d —0.5 e f U 0.5 g h Figure 4. Polymer feature size predictions based on the model. All plots show a cross-section at the focal plane of a lens illuminated with a Gaussian beam (or Gauss-Laguerre beam in the case of b) over the 1/e2 intensity diameter of the initiating focus (a). Part (b) shows the unit topological charge Gauss-Laguerre inhibition intensity. Parts (e-h) show the predicted polymerization profile all with the same color scale as the peak normalized UV intensity relative to the peak normalized blue intensity is varied from 1 to 8 by factors of 2. For comparison, the expected polymerization profile using two-photon initiation at 400 nm is shown in (c). The size of this feature can be reduced by decreased intensity, but as shown in (d), this comes at the expense of less peak polymerization, unlike (e-h). Representative results for a Gaussian initiation profile and Gauss-Laguerre (charge = 1) inhibition profile are shown in Figure 4. Several important qualitative observations can be made. First, the polymerized regions are smaller than those expected from two photon excitation at the same wavelength and have sharper, more well-defined boundaries. Both features are due to the fact that the system presented here depends on the difference of two intensities, while twophoton absorption depends on the product. Second, the sequence (e-h) illustrates that, in this ideal model, the degree of polymerization and feature size can be controlled independently. In applications such as 3D microfabrication, this will enable mechanically-robust polymer structures simultaneous with minimum feature size. Finally and obviously, the potential of the approach to create features significantly below the diffraction limit is illustrated by comparing (a) and (h). 4. LITHOGRAPHY DEMONSTRATION The experimental setup illustrated in Figure 1 is used to demonstrate this ability to write features that are not limited by diffraction. The objective lens, Geltech model 350610, is designed for no aberrations when focused 1.2 mm deep in a material with index ~1.5 and has effective focal length of 4 mm and back aperture of 4.8 mm diameter. This aperture is Proc. of SPIE Vol. 7053 70530E-5 Downloaded from SPIE Digital Library on 07 Jun 2012 to 198.11.25.124. Terms of Use: http://spiedl.org/terms illuminated by a 473 nm collimated beam with 5.7 mm 1/e2 intensity diameter provided by a diode-pumped Nd:YAG laser. This truncated Gaussian results in a focused spot diameter of 1.35 µm with 15 µW total power. The UV GaussLaguerre beam has 1.3 mm 1/e2 intensity diameter, 1.8 mW total power, and is produced by a single-line argon ion laser. This beam is slightly diverging at the lens aperture in order to compensate for the lateral chromatic aberration of the singlet objective lens, colocating the blue and UV focal planes in depth. The liquid monomer system is placed between glass slides that provide mechanical and optical surfaces. The focal plane is positioned on an interior glass surface using a 593 nm HeNe laser focused through the same objective to implement a confocal reflection microscope and Newport XYZ stages (models XMS and GTS). A shutter is used to create individual polymer “spots” like those of Figure 4 or the stages are directed to move the focus along the glass/monomer boundary and the shutter is opened to write polymer “lines”. After exposure, unused resin is washed away with isopropanol. Unused solvent is then removed by in a desiccator overnight. The resulting polymer features on the slide are then coated with 7 nm of gold and imaged with a thermal emission scanning electron microscope in high vacuum mode (JEOL model JSM-6480LV). The results are shown in Figure 5. The diameter of the polymer “spot” is ~180 nm, a factor of 7.5 smaller than the blue initiating diffraction-limited intensity profile. The dark halo around the polymer spot indicates strong suppression of polymerization at the periphery, matching the simple model above which predicts sharp-edged features. The large (b) and small (c) lines show that the terminating TED radials have a lifetime less than 200 msec, the time it takes the ~200 nm polymerizing region to move on to an adjacent area. a b c Figure 5. Polymer features. (a) shows a “spot” written in a 5 second static exposure, while (b) and (c) show “lines” written as the sample moves at one micron per second at two different exposure conditions. Note that the scales are different in each case. 5. CONCLUSIONS We have proposed and demonstrated a new material and two-color lithography system that can fabricate polymer features a factor of 7.5 smaller than the diffraction limit. Polymerization is driven by the difference of the two optical intensities, enabling small, sharp-edged features simultaneous with a high degree of conversion. The terminating species are demonstrated to have a lifetime under 200 msec, enabling continuous 3D lithography at micron per second velocity. The very low writing powers contrast strongly with two-photon initiation and should allow rapid single-point exposure or large area masks and holograms currently not possible. ACKNOWLEDGMENTS This material is based upon work supported by the National Science Foundation under Grant No. IIP-0637355 and by Intel Corporation. REFERENCES 1 S. Maruo, O. Nakamura, and S. Kawata, “Three-dimensional microfabrication with two-photon-absorbed photopolymerization,” Opt. Lett. 22, 132-134 (1997). Proc. of SPIE Vol. 7053 70530E-6 Downloaded from SPIE Digital Library on 07 Jun 2012 to 198.11.25.124. Terms of Use: http://spiedl.org/terms 2. 3 4 5 6 7 8 B. H. Cumpston, S. P. Ananthavel, S. Barlow, D. L. Dyer, J. E. Ehrlich, L. L. Erskine, A. A. Heikal, S. M. Kuebler, I.-Y. S. Lee, D. McCord-Maughon, J. Qin, H. Rockel, M. Rumi, X-L. Wu, S. R. Marder, and J. W. Perry, “Two-photon polymerization initiators for three-dimensional optical data storage and microfabrication,” Nature 398, 51–54 (1999) . K. D. Belfield, X. Ren, E. W. Van Stryland,, D. J. Hagan, V. Dubikovsky, and E. J. Miesak, “Near-IR TwoPhoton Photoinitiated Polymerization Using a Fluorone/Amine Initiating System,” J. Am. Chem. Soc. 122, 1217-1218 (2000). H-B Sun, T. Tanaka, K. Takada, and S. Kawata, “Two-photon photopolymerization and diagnosis of threedimensional microstructures containing fluorescent dyes,” Appl. Phys. Lett. 79, 1411 (2001) S-K. Lee and D. C. Neckers, “Two-photon Radical-Photoinitiator System Based on Iodinated Benxospiropyrans,” Chem. Mater. 3, 858-864 (1991). D. A. Parthenopoulos and P. M. Rentzepis, “Three-Dimensional Optical Storage Memory,” Science 245, 4920, (1989). S. W. Hell and J. Wichmann, "Breaking the diffraction resolution limit by stimulated emission: stimulatedemission-depletion fluorescence microscopy," Opt. Lett. 19, 780- (1994). L. G. Lovell, B. J. Elliott, J. R. Brown, C. N. Bowman, “The effect of wavelength on the polymerization of multi(meth)acrylates with disulfide/benzilketal combinations,” Polymer 42, 421-429 (2001). Proc. of SPIE Vol. 7053 70530E-7 Downloaded from SPIE Digital Library on 07 Jun 2012 to 198.11.25.124. Terms of Use: http://spiedl.org/terms