Open Access
Add to BibSonomy
Abstract
When fabricating curved microstructure with DMD-based digital lithography, the buried microstructure formed in photoresist will introduce a distinct diffraction effect, which hinders the improvement in the fabrication fidelity. In this paper, a multiple exposure method is demonstrated to reduce the effect of buried microstructure diffraction. In this method, a high-space-frequency curved microstructure is decomposed into multiple low-space-frequency sub-microstructures, whose corresponding digital masks are successively exposed at the same position of substrate. Lithography experiments are implemented by introducing the multiple exposure method within the DMD-based digital lithography system. The experimental results show that the effect superimposed on the lithography pattern caused by the buried microstructure diffraction can be effectively reduced. In addition, we discuss the influence factors of buried microstructure diffraction by using FDTD method. The experimental results suggest that this fabrication method is potentially suitable for deep microstructure, particularly curved microstructure.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
A precisely controlled three-dimensional (3D) surface profile is especially attractive for the fabrication of curved microstructure [1–4]. 3D lithography has the distinct advantages of low cost and fast turn-around, so it is frequently used to realize a desired profile of curved microstructure [5–7]. However, fabrication error introduced by diffraction effect always hinders the improvement in the fabrication fidelity [8–10]. So researchers devote to improving the lithography resolution recently [11,12]. One of the important and meaningful researches has been focused in optimizing the mask to offset the image error introduced by diffraction. PSM (Phase Shift Mask) [13], OPC (Optical Proximity Correction) [14–16] and ILT (Inverse Lithography Technique) [17,18] have been developed to improve the lithography quality. In recent years, these methods have been applied in the digital lithography [19,20]. These methods, based on mask optimization, can reduce the mask diffraction, which happens when ultraviolet light passes through the mask in traditional lithography or virtual mask in digital lithography.
When fabricating a deep profile, we find that the buried microstructure which is formed by the redistribution of refractive index in the exposure region after the photoresist absorbs energy, may introduce diffraction effect as well. Twice diffractions happen during the whole lithography process, the first is the mask diffraction and the second is the buried microstructure diffraction. Although the second diffraction has no obvious effect on fabrication fidelity for fabricating a shallow microstructure, it has a significant effect on a deep one. Particularly for the fabrication of curved microstructure, the second diffraction effect will be more magnified. Although the above mentioned methods can successfully solve the question of mask diffraction, they are incapable of eliminating the buried microstructure diffraction. In addition, in the digital lithography, it is difficult to apply the above-mentioned methods to fabricate curved microstructure due to the pixel structure of DMD. In view of this, a multiple exposure method is proposed to weaken the effect of buried microstructure diffraction.
2. The buried microstructure diffraction in fabricating curved microstructure
Figure 1 provides an example of the formation of buried microstructure diffraction in fabricating planar circular grating. As shown in Fig. 1(a), the ultraviolet light is modulated by DMD before reaching the photoresist layer in the lithography process. A UV light source with fixed power is adopted in lithography, so the exposure energy toward photoresist is mainly modulated by the exposure time. The photochemical reaction occurs as the photoresist absorbs the light. Then the inconsistent distributed region of refractive index is formed within photoresist. As shown in Fig. 1(b), as the photoresist absorbs a certain exposure dose, the buried circular grating with the thickness d is formed in the surface layer of photoresist. d will increase with the continued exposure, until the desired depth is achieved. When the UV light passing through the buried circular grating included region, it will deviate from the straight-line path due to diffraction modulation of the buried circular grating, as shown in Fig. 1(c). Consequently, the fabricated circular grating shown in Fig. 1(d) is superimposed by the diffraction modulation of buried circular grating in photoresist.
Fig. 1 Illustration of buried microstructure diffraction in fabricating planar circular grating. (a) Exposure process within photoresist. (b) Buried circular grating formed in the surface layer of photoresist. (c) Diffraction modulation of the buried circular grating. (d) Fabricated circular grating superimposed by the buried microstructure diffraction.
Although the buried microstructure diffraction has no obvious effect on fabrication of a shallow microstructure, it will be more magnified in fabricating a curved microstructure. The conventional proximity lithography is not suitable for fabricating curved microstructure due to its structure particularity. Besides the point-by-point fabrication method, the DMD-based digital lithography can be used to fabricate the curved microstructure. Figure 2 illustrates the formation of buried microstructure diffraction in fabricating a curved microstructure. In Fig. 2(a), the buried microstructure within photoresist is composed of a high-space-frequency grating part and a low-space-frequency curved part where the grating part is distributed. Diffraction occurs as the UV light passes through the high-frequency grating part, and then it is obviously magnified by the low-frequency curved part which corresponds to a lens shown in Fig. 2(b). Therefore the diffraction effect resulting from the microstructure part is significantly enhanced by the curved part. Accordingly, the enhanced diffraction effect will be overlayed onto the lithographic pattern of curved microstructure, as shown in Fig. 2(c).
Fig. 2 Illustration of buried microstructure diffraction in fabricating curved microstructure. (a) Exposure process within photoresist. (b) Diffraction modulation of the buried curved microstructure. (c) Fabricated curved microstructure overlayed by the enhanced diffraction.
To further confirm the diffraction effect observed in experiment, the diffraction of buried circular grating including planar and curved ones have been calculated using FDTD method. The main parameters are as follows: i-line (λ = 365 nm) dominated UV light source is set. The pitches of planar and curved circular grating with 50% duty ratio are both 10 μm. In curved circular grating, the curved part where the grating is distributed is a spherical crown having a curvature radius of 51.202 mm and a depth of 4 μm. In the grating part, the radius of the outermost ring is 384 μm, which is the same as that of the planar grating. Moreover, we define two types of materials. One type denotes the unexposed photoresist and the other denotes the exposed photoresist. The measurement method of the refractive index can be described as follows: Firstly, we measure the film thickness of unexposed photoresist and the corresponding optical path, respectively by using AFM (NTEGRA, NT-MDT Co.) and White Light Interferometer (MicroXAM-100, KLA-tencor). And then the refractive index of unexposed photoresist can be easily calculated. The measurement results by AFM and White Light Interferometer are shown in Fig. 3. From Fig. 3, we can see that the photoresist film thickness is 2.57 μm and the corresponding optical path is 4.435 μm. The refractive index of unexposed photoresist is approximately 1.73. Then a region with film thickness of 4.4 μm is chosen to be exposed. The optical path difference between the exposed region and the unexposed region is 423.2 nm, as shown in Fig. 4. The refractive index of exposed photoresist can be calculated to be 1.83. The calculated results are shown in Fig. 5. Figure 5(a) shows a planar circular grating without the buried microstructure diffraction. Figures 5(b) and 5(c) show the calculated results when i-line light passes through the planar and the curved circular grating respectively. By comparing Figs. 5(b) and 5(c), it’s obvious that for the planar circular grating, the diffraction effect certainly exists but is relatively weak. For the curved circular grating, the above-mentioned diffraction effect is significantly enhanced and will have a greater influence on the fabrication fidelity.
Fig. 3 The measurement of refractive index of unexposed photoresist. (a) Measurement of photoresist film thickness by AFM. (b) Measurement of optical length by White Light Interferometer.
Fig. 4 The measurement of optical path difference between the exposed region and the unexposed region.
Fig. 5 Calculation results for buried microstructure diffraction using FDTD method. (a) Planar circular grating without the buried microstructure diffraction. (b) Planar circular grating superimposed by the buried microstructure diffraction. (c) Curved circular grating superimposed by the buried microstructure diffraction.
3. Multiple exposure process
The diffraction effect introduced by buried microstructure is mainly related to the spatial frequency of the buried microstructure in photoresist. In order to weaken or even eliminate the diffraction effect introduced by buried microstructure, we present a multiple exposure process in this study. A curved microstructure including high-space-frequency part is first decomposed into multiple low-space-frequency sub-microstructures. The digital masks for all the curved sub-microstructures are needed to be calculated independently. Then they are successively exposed at the same position of substrate and so that the profile of the superposed UV dose can coincide with the desired profile. The spatial frequency of each sub-microstructure is decreased in comparison with the original curved microstructure, therefore the diffraction effect introduced by buried microstructure can be effectively weakened.
In order to illustrate the multiple exposure process, Fig. 6 demonstrates an example of two-step exposure for fabricating the curved circular grating. A curved circular grating G(x, y, z) in Fig. 6(a) with pitch of a is decomposed into two sub-microstructures in Fig. 6(b) with pitch of 2a, namely the spatial frequency of sub-microstructure decreases to half of that of the original curved circular grating. The two sub-microstructures are expressed by G1(x, y, z) and G2(x, y, z) respectively. Then the digital masks M1(i, j) and M2(i, j) corresponding to G1(x, y, z) and G2(x, y, z) are calculated respectively, as shown in Fig. 6(c). Finally, the two digital masks are successively exposed at the same position to realize the superimposed exposure.
Fig. 6 Two-step exposure for fabricating the curved circular grating. (a) Curved circular grating with pitch of a. (b) Two sub-microstructures with pitch of 2a. (c) Two digital masks corresponding to two sub-microstructures in (b).
4. Experimental results and discussion
In order to verify the feasibility of this method, the curved circular grating has been fabricated based on a self-building digital lithography system [21,22], respectively by using one-step exposure and multiple exposure method. Figure 7 shows a schematic of DMD-based digital lithography system. A 200W mercury short arc lamp (OSRAM Co.) is used as a light source. The illumination light (365 nm) becomes a plane light beam through a beam expander/collimator and a homogenizer before striking the DMD. The Texas Instruments DMD consists of 1024 × 768 micromirrors having a pitch size of 13.68 μm. The 0.071 × reduction lens with numerical aperture of 0.3 is horizontally installed to reduce the DMD image. Through the reduction lens, the DMD can project an image covering an area of approximately 1.01 mm × 0.75 mm, including 1024 × 768 pixels, each of which is 0.977 μm in size. The motorized stage includes XY stages and Z stage. XY stages cover a total range of 3 × 3 cm2 with accuracy of ± 0.65 μm. Z stage covers a range of 2 mm with accuracy of ± 1 μm. A beam splitter (BS) is placed between the DMD and reduction lens and used to reflect the exposure image on a photoresist-coated substrate to a CCD (charge coupled device) monitor.
The fabricated curved part where the circular grating is distributed is a spherical crown having a curvature radius of 51.202 mm and a depth of 4 μm. The bottom of the spherical crown covers an area of 1024 μm × 768 μm. The fabricated circular grating distributed on the curved part has a duty ratio of 50% and a pitch of 10 μm. And the radius of its outermost ring is 384 μm. After passing through the exposure system, the light power toward the substrate is 3.45 mw/cm2 at the wavelength of 365 nm. A silicon substrate is spin coated with a positive photoresist (GP28,Chengdu spectrum optoelectronic technology LTD.).
In multiple exposure method, the original circular grating is sampled with the same sampling pitch. So the sub-microstructure acquired by sampling has similar curved structure as the original circular grating. In our experiment, the original circular grating is decomposed into two sub-microstructures with duty ration of 25% and pitch of 20 μm. Figure 8 shows the digital masks of curved circular grating, respectively used for one-step and two-step exposure. The experimental results are shown in Fig. 9. Figures 9(a) and 9(b) are the microscope images of curved circular grating respectively by one-step and two-step exposure. Figures 9(c) and 9(d) are the interferometric profile images respectively by one-step and two-step exposure, which are obtained by scanning the yellow square regions in Figs. 9(a) and 9(b). In Fig. 9(a), a star-shaped region surrounded by four red dashed lines appears in the lithography pattern, which is obviously resulted from the buried microstructure diffraction. However, we can see from Figs. 9(b) and 9(d) that this influence can be effectively weakened by the multiple exposure method.
Fig. 8 Digital masks of curved circular grating. (a) Mask used for one-step exposure. (b) and (c) Mask used for two-step exposure.
Fig. 9 Experimental results of the curved circular grating. (a) Microscope image of curved circular grating fabricated by one-step exposure. (b) Microscope image of curved circular grating fabricated by two-step exposure. (c) Interferometric profile image of curved circular grating fabricated by one-step exposure. (d) Interferometric profile image of curved circular grating fabricated by two-step exposure.
The buried microstructure diffraction occurs as ultraviolet light passes through the buried microstructure formed in photoresist. For fabricating a shallow microstructure, exposure is performed on a thin photoresist layer, so the absorption of photoresist is easily going into saturation in a very short time. This implies that the exposure stops following with the formation of buried microstructure. Therefore, for a shallow microstructure, the influence from the buried microstructure diffraction is relatively small. However, for the fabrication of deep microstructure, the relatively long exposure time is required. An increase in exposure time leads to an increased depth of the buried microstructure, thus causing the gradual increase of the negative impact of diffraction. In order to verify the above statement, FDTD method is used to estimate the error distribution resulting from the buried microstructure diffraction. Figures 10(a) and 10(b), respectively, show the error variation with the increasing fabrication depth when a plane light wave at 365nm wavelength propagating through a buried planar and a curved circular grating. The main parameters used for error calculation are set as follows: the pitch of planar circular grating is 10 μm and the radius of its outermost ring is kept constant at 384 μm; for the grating part of the curved circular grating, the grating parameters are the same as those of the planar grating; the curved part where the grating is distributed is a spherical crown whose curvature can be adjusted by the crown height. As can be seen from Fig. 10(a), although the buried microstructure diffraction is slight at fabrication depth below 5 μm, it has a gradual increase with the increasing fabrication depth. Until the fabrication depth increases to 13 μm, the error caused by buried microstructure diffraction reaches the maximum value. It can be seen from Fig. 10(b) that the diffraction enhancement caused by buried curved circular grating is not obvious at a relatively large radius of curvature (namely a relatively small crown height). However, the negative effect of diffraction sharply increases when the crown height is larger than 13 μm (namely the radius of curvature is smaller than 5677 μm). Therefore, the diffraction effect introduced by buried planar microstructure will increase along with the increasing fabrication depth. The error will reach the maximum value when the fabrication depth increases to a certain value. The curved part of the buried curved microstructure will magnify the negative effect of diffraction. The error will increase with the decreasing radius of curvature when the crown height increases to a certain value.
Fig. 10 (a) RMS error with respect to the fabrication depth of planar circular grating. (b) RMS error with respect to the crown height of curved circular grating.
The experimental study of this paper is based on our self-building digital lithography system with an achievable minimal resolution of 1 μm. An approximately 8-μm-thick photoresist layer can be formed on the substrate, followed by a spin coating of GP28. By experiments we find that the buried microstructure diffraction has only a slight effect on the fabrication of a planar microstructure with spatial frequency of 2 μm−1 when the fabrication depth is no more than 5 μm. However, it seems to have an obvious negative effect in fabricating a curved microstructure with spatial frequency above 10 μm−1 when the fabrication depth is approximately 4 μm. And the negative effect becomes more serious with the increasing fabrication depth. Therefore, the buried microstructure diffraction shows a strong response to the spatial frequency of microstructure and the fabrication depth.
In this paper, double exposure method is used and the spatial frequency of microstructure in a single exposure is reduced to 20 μm−1, the negative effect of buried microstructure diffraction is effectively inhibited within 5 μm depth.
5. Conclusions
This work provides a multiple exposure method to weaken the effect of buried microstructure diffraction. In this method, a high-space-frequency curved microstructure is decomposed into multiple low-space-frequency sub-microstructures, the corresponding digital masks of which are successively exposed at the same position of substrate. Through experiments, the effectiveness of the proposed method has been demonstrated. The multiple exposure method, combined with the digital lithography, has advantages of low cost, high efficiency and great flexibility. The diffraction limit on fabricating microstructures using the multiple exposure method is limited by the imaging resolution of our self-building digital lithography system. This method is more suitable for fabricating periodic microstructures with feature size larger than the imaging resolution of digital lithography system. For the fabrication of a deeper microstructure with high frequency, the method is incapable due to the limitation of fabrication system.
Funding
National Natural Science Foundation of China (61464008, 61704070, 61665008); Natural Science Foundation of Jiangxi Province (20171BAB212020); Natural Science Foundation Jiangxi Provincial Department of Education (GJJ160722); Aeronautical Science Foundation of China (2016ZD56006); Program for Supporting Young Scientists of Jiangxi Province (20153BCB23037).
References
1. H. Bian, Y. Wei, Q. Yang, F. Chen, F. Zhang, G. Q. Du, J. L. Yong, and X. Hou, “Direct fabrication of compound-eye microlens array on curved surfaces by a facile femtosecond laser enhanced wet etching process,” Appl. Phys. Lett. 109(22), 221109 (2016). [CrossRef]
2. N. Tsutsumi, J. Hirota, K. Kinashi, and W. Sakai, “Direct laser writing for micro-optical devices using a negative photoresist,” Opt. Express 25(25), 31539–31551 (2017). [CrossRef] [PubMed]
3. T. Takai, H. Nakao, and F. Iwata, “Three-dimensional microfabrication using local electrophoresis deposition and a laser trapping technique,” Opt. Express 22(23), 28109–28117 (2014). [CrossRef] [PubMed]
4. D. Zhang, W. Yu, T. Wang, Z. Lu, and Q. Sun, “Fabrication of diffractive optical elements on 3-D curved surfaces by capillary force lithography,” Opt. Express 18(14), 15009–15016 (2010). [CrossRef] [PubMed]
5. H. Zhang, L. Li, D. L. McCray, D. G. Yao, and A. Y. Yi, “A microlens array on curved substrates by 3D micro projection and reflow process,” Sensor. Actuat. A-Phys. 179(3), 242–250 (2012).
6. A. Rammohan, P. K. Dwivedi, R. Martinez-Duarte, H. Katepalli, M. J. Madou, and A. Shama, ““One-step maskless grayscale lithography for the fabrication of 3-dimensional structures in SU-8,” Sensor. Actuat. B.-,” Chem 153(1), 125–134 (2011).
7. X. Ma, Y. Kato, F. Kempen, Y. Hirai, T. Tsuchiya, F. Keulen, and O. Tabata, “Multiple patterning with process optimization method for maskless DMD-based grayscale lithography,” Procedia Eng. 120, 1091–1094 (2015). [CrossRef]
8. H. C. Liu, Y. F. Luo, W. J. Kong, K. P. Liu, W. J. Du, C. W. Zhao, P. Gao, Z. Y. Zhao, C. T. Wang, M. B. Pu, and X. G. Luo, “Large area deep subwavelength interference lithography with a 35 nm half-period based on bulk plasmon polaritons,” Opt. Mater. Express 8(2), 199–209 (2018). [CrossRef]
9. J. Fischer and M. Wegener, “Three-dimensional optical laser lithography beyond the diffraction limit,” Laser Photonics Rev. 7(1), 22–44 (2013). [CrossRef]
10. T. Horiuchi, S. Koyama, and H. Kobayashi, “Simple maskless lithography tool with a desk-top size using a liquid-crystal-display projector,” Microelectron. Eng. 141, 37–43 (2015). [CrossRef]
11. M. Seo and H. Kim, “Influence of dynamic sub-pixelation on exposure intensity distribution under diffraction effects in spatial light modulation based lithography,” Microelectron. Eng. 98, 125–129 (2012). [CrossRef]
12. X. Y. Ding, Y. X. Ren, L. Gong, Z. X. Fang, and R. D. Lu, “Microscopic lithography with pixelate diffraction of a digital micro-mirror device for micro-lens fabrication,” Appl. Opt. 53(24), 5307–5311 (2014). [CrossRef] [PubMed]
13. N. Yao, Z. Lai, L. Fang, C. Wang, Q. Feng, Z. Zhao, and X. Luo, “Improving resolution of superlens lithography by phase-shifting mask,” Opt. Express 19(17), 15982–15989 (2011). [CrossRef] [PubMed]
14. S. Oh, D. Han, H. B. Shim, and J. W. Hahn, “Optical proximity correction (OPC) in near-field lithography with pixel-based field sectioning time modulation,” Nanotechnology 29(4), 045301 (2018). [CrossRef] [PubMed]
15. X. Ma, Z. Wang, Y. Li, G. R. Arce, L. Dong, and J. Garcia-Frias, “Fast optical proximity correction method based on nonlinear compressive sensing,” Opt. Express 26(11), 14479–14498 (2018). [CrossRef] [PubMed]
16. X. Ma and G. R. Arce, “Pixel-based OPC optimization based on conjugate gradients,” Opt. Express 19(3), 2165–2180 (2011). [CrossRef] [PubMed]
17. X. Ma and G. R. Arce, “PSM design for inverse lithography with partially coherent illumination,” Opt. Express 16(24), 20126–20141 (2008). [CrossRef] [PubMed]
18. X. Ma and G. R. Arce, “Generalized inverse lithography methods for phase-shifting mask design,” Opt. Express 15(23), 15066–15079 (2007). [CrossRef] [PubMed]
19. J. Y. Hur and M. S. Seo, “Optical proximity corrections for digital micromirror device-based maskless lithography,” J. Opt. Soc. Korea 16(3), 221–227 (2012). [CrossRef]
20. Y. Lim, J. Hahn, and B. Lee, “Phase-conjugate holographic lithography based on micromirror array recording,” Appl. Opt. 50(34), H68–H74 (2011). [CrossRef] [PubMed]
21. Z. M. Zhang, Y. Q. Gao, N. N. Luo, and K. J. Zhong, “Fast fabrication of curved microlens array using DMD-based lithography,” AIP Adv. 6(1), 015319 (2016). [CrossRef]
22. N. N. Luo and Z. M. Zhang, “Fabrication of a curved microlens array using double gray-scale digital maskless lithography,” J. Micromech. Microeng. 27(3), 035015 (2017). [CrossRef]
