Virtual input device with diffractive optical element

Transcription

1 Virtual input device with diffractive optical element Ching Chin Wu, Chang Sheng Chu Industrial Technology Research Institute ABSTRACT As a portable device, such as PDA and cell phone, a small size build in virtual input device is more convenient for complex input demand. A few years ago, a creative idea called virtual keyboard is announced, but up to now there s still no mass production method for this idea. In this paper we ll show the whole procedure of making a virtual keyboard. First of all is the HOE (Holographic Optical Element) design of keyboard image which yields a fan angle about 30 degrees, and then use the electron forming method to copy this pattern in high precision. And finally we can product this element by inject molding. With an adaptive lens design we can get a well correct keyboard image in distortion and a wilder fan angle about 70 degrees. With a batter alignment of HOE pattern lithography, we re sure to get higher diffraction efficiency. Keywords: DOE, HOE, diffractive, virtual keyboard. INTRODUCTION There are three key components of a virtual input system, which are image projection device, image capture device and image processing device. In this paper we ll only focus on the image projection device. And will introduce about the image generator and a simple lens system for enhancing the image and so on.. SIMULATION OF A VIRTUAL KEYBOARD IMAGE DOE For a miniaturized virtual input device, there are three key components which are keyboard image generator, lens system and the mechanical parts. In this presentation we use the Adaptive Additive Fourier Transform [] to compute the phase of the DOE for the keyboard image generator, sometimes called CGH [][3][4] (Computer Generated Hologram). And also use a simple lens system to expand and correct the distortion of the projected image. With an easy used mechanical design, the device could be shortened into a smaller size for carrying about. (Fig. ) is the Adaptive Additive Fourier Transform flow chart. The known light source amplitude distribution A 0, the Holography, Diffractive Optics, and Applications II, edited by Yunlong Sheng, Dahsiung Hsu, Chongxiu Yu, Byoungho Lee, Proceedings of SPIE Vol (SPIE, Bellingham, WA, 005) X/05/$5 doi: 0.7/

2 goal image distribution A 0 and the initial phase distribution 0 should be given first. With the given data we can get the input field E which is equal to ' i A e i Ae. After the fast Fourier transform we can get the image field distribution E which is equal to, and the related image intensity distribution. By checking the difference intensity distribution between the goal distribution and the related image intensity distribution, we can make decision whether stop the iteration or not. If continue the iteration, will go into the adaptive additive part. Here, we use the field A c A ( c) A ' 0 distribution to replace the for the image field distribution for further inverse Fourier Transform, and the coefficient c is larger or equal and less or equal to. By controlling the c value, we can find a batter result for the iteration. In this case we use c value equal to, which means enhance the goal energy distribution and reduce the noise very strongly. After getting the A field and keeping the phase of the related image phase, we can get a replaced image field E where ' A A ' E A e i. By performing the inverse Fourier Transform, we can get the related input field distribution E. And replace the A by A 0, now we can go into the next iteration. This is the whole main process for the Adaptive Additive Fourier Transform we used in this paper. With a normal design of four levels phase distribution, we can get 6% efficiency, (Fig. ). And with a distortion corrected design, we can get 58% efficiency, (Fig. 3). We use 00x500 and 00x880 pixels for each keyboard image, and 0.5um in square for each pixel for fabrication [5][6] and get an image fan angle about 30 degrees. 3. VIRTUAL INPUT DEVICE FABRICATION AND EXPERIMENT RESULTS Nanolithography technology and nanoimprint [7] is able to fabricate the DOE of virtual input device. The details of the fabrication are as follows, (Fig. 4) is the flow chart explaining the processes of the fabrication: First, using e-beam lithography or photolithography such as the stepper UV exposure system, we can make the nano scale pattern of DOE in photoresist. But the traditional photolithography such as the stepper exposure system has two parameters which have to be corrected [8]. One is optical and process correction (OPC), the other is ISO-dense bias. For the defects mentioned above, we use e-beam lithography to fabricate photoresist pattern that the linewidth of DOE is set to be 500nm. The advantages of using it are that it does not change shapes and it has a high resolution. The (Fig. 5) shows the SEM photo of the photoresist of DOE with a linewidth of 500nm. Following we obtain the virtual keyboard device using two ways. One is to etch the photoresist of DOE on glass substrate to obtain the device directly; the other is using electroforming of the photoresist sample and than use the 448 Proc. of SPIE Vol. 5636

3 nano-imprinting like molding. We use the nano-imprinting like fabrication and the plastic injection molding to implement the DOE, because it can improve the e-beam lithography to make it to have a repeating function and mass production ability. So the precisely electroforming is very important for the further process. The mould plays the same role as photomask does in photolithography. Then we use the Ni mould to imprint the polymer. First, the UV-curable polymer smears on the Ni mould, and then with glass or plastic material to be as substrate to cover on UV-curable polymer. After exposing by the UV light, the UV-curable polymer will be cured completely and transfered the DOE pattern, and the depth of imprinting is about 650nm. The depth corresponds to the thickness of DOE nanostructure in Ni mould. When the DOE nanostructure is imprinted on polymer, it can be measured for the performance of the diffraction. (Fig. 6) Shows the diffraction image of the virtual keyboard on the board is clearly. (Fig. 7) displays the device on the left hand side of the white board is the projected keyboard image. Proc. of SPIE Vol

4 Initial Phase 0 Initial field distribution Amplitude A 0 0 i A E Ae Image amplitude =A 0 A0 Coefficient c ' A A FFT Image field distribution E ' i A e Image field distribution ' i E A e FFT - Replaced Image field SSE ' A A 0 E A e i SSE OK NO A AA algorithm ' c A0 ( c) A, c Figure: Shows the flow chart of the Adaptive Additive Fourier Transform. 450 Proc. of SPIE Vol. 5636

5 Figure: Shows the simulation of keyboard image with 4 levels. Figure 3: Shows the simulation of distortion corrected keyboard image with 4 levels. Figure 4: The flow chart explaining the processes of the fabrication. Proc. of SPIE Vol

6 Figure 5: Photoresist linewidth of DOE pattern is 500nm using e-beam lithography. 45 Proc. of SPIE Vol. 5636

7 Figure 6: The diffraction image of the virtual keyboard on the board clearly. Figure 7: The device on the right hand side of the white board is the project of virtual keyboard. Proc. of SPIE Vol

8 4. CONCLUSCION In this paper, we present our study on the keyboard image generating DOE s fabrication process. We simulate the DOE phase for generating the keyboard image by Adaptive Additive Fourier transform method, and get a fan angle about 30 degrees. When projecting this image on a plane will get more distortion because of the HOE propriety and image plane tilt. However, we can use a plastic concave aspherical lens for correcting the distortion and to enlarge the fan angle, so we can get a batter projecting image and a shorter projecting distance. Also we find that when the pixel width of the DOE is smaller than 0.3 um it could be difficult to produced by injection molding. According to the above results, we make the DOE lithography process, plastic molding of the DOE and assemble the system for keyboard image projection. REFERENCES. Jeng-Feng Lin and Alexander A. Sawchuk, Design of diffractive optical elements with optimization of the singal-to-noise ratio and without a dummy area, Applied Optics, vol.36, no.4, 0 May 997, pp Publisher: Opt. Soc. America, USA. Michael T. Gale, Markus Rossi, Continuous-relief diffractive optical elements for two-dimensional array generation, Applied Optics, vol.3, no.4, 0 May 993, pp Publisher: Opt. Soc. America, USA 3. Wai-Hon Lee, Binary computer-generated holograms, Applied Optics, vol.8, no., November 979, pp Publisher: Opt. Soc. America, USA 4. Brown DM, Brown DR, Brown JD, High performance analog profile diffractive elements, SPIE-Int. Soc. Opt. Eng. Proceedings of Spie - the International Society for Optical Engineering, vol.3633, 999, pp USA 5. Suleski TJ, Bagget B, Delaney WF, Kathman AD, Emerging fabrication methods for diffractive optical elements, SPIE-Int. Soc. Opt. Eng. Proceedings of Spie - the International Society for Optical Engineering, vol.3633, 999, pp USA 6. Andersson H, Ekberg M, Hard S, Jacobsson S, Larsson M, Nilsson T, Single photomask, multilevel kinoforms in quartz and photoresist: Manufacture and evaluation, Applied Optics, vol.9, no.8, Oct. 990, pp USA 7. L Jay Guo, Recent progress in nanoimprint technology and its applications, J. Phys. D: Appl. Phys. 37, R3 (004). 8. Shamma et al, A method for the correction of proximity effects in optical projection lithography, in Proc. KTI Microelectronics seminar (99), Proc. of SPIE Vol. 5636