Development of a new multi-wavelength confocal surface profilometer for in-situ automatic optical inspection (AOI) Liang-Chia Chen 1#, Chao-Nan Chen 1 and Yi-Wei Chang 1 1. Institute of Automation Technology, National Taipei University of Technology. No.1, Sec. 3, Chung-Hsiao East Road, Taipei, Taiwan 1061 # Corresponding Author / E-mail: lcchen@ntut.edu.tw, TEL: +886-2-2771-2171-4321, FAX: +886-2-8773-3217 KEYWORDS : Automatic optical inspection (AOI),confocal measurement, chromatic multi-wavelength, three-dimensional measurement An in-situ 3-D surface profilometer for inspecting micro surface profiles with a long vertical range and high resolution was successfully developed by using innovative slit-scan multi-wavelength confocal surface profilometry. In conventional confocal microscopy, vertical scanning of a tested surface by either stage depth movement or shifting of optical objectives is of time consuming, thus making unacceptable measurement efficiency for in-situ inspection. To overcome this, in the research, a multi-wavelength confocal system employing a broad band light source in combination with a chromatic dispersion objective was developed to generate an accurate wavelength-to-depth conversion for in-situ 3-D profile measurement. The specially designed objective is capable of modulating a broadband light to produce axial chromatic dispersion with respect to a corresponding series of axial focusing depths along the vertical-axis and then to obtain the corresponding reflected light spectrum from the object s surface. By designing the objective, the depth measurement range can be ranged between few to several hundred micrometers and the depth measurement resolution can theoretically reach to 0.03% of the overall detection range. From the experimental test, it was found that the maximum measurement error and repeatability can be controlled well within 0.099% of the overall measurement range and 52 nm in one standard deviation, respectively. By integrating a high-speed image acquisition unit, the measurement efficiency can be further enhanced for in-situ automatic optical inspection (AOI). Manuscript received: July 15, 2009 / Accepted: August 15, 2009 1. INTRODUCTION Confocal microscopy has become a powerful measurement method due to its unique optical sectioning capability. The confocal measurement utilizes the geometrical matching of two conjugate focal points corresponding to both the object surface and the point detector defined by a pinhole [1,2]. Conventional laser confocal measurement has been widely used to reconstruct 3-D surface contours in biomedical application. Laser light has its advantages in overcoming large variation of surface reflectivity while other methods may suffer from the limitation severely. However, this kind of the method has also been restricted by its inefficient scanning rate and in general has been difficult to achieve in-situ surface profilometry. In recent years, a digital micromirror device (DMD) can be deployed as an array of pinholes for confocal lateral scanning. The DMD used as a spatial light modulator has a number of independently controlled micro mirrors for generating defined spatial fringe modulations of light field as structured light in modern 3-D profilometry [3]. To increase scanning efficiency, several methods have also been developed to fasten profile measurement. Most importantly, the Nipkow disk, as a way to transmit images electrically using a pair of spinning metal disks or the use of a diffractive lens illuminated with a tunable light source, has been utilized to improve confocal scanning speed [4]. Using micro-lens arrays, it can be employed to obtain full-field 3-D information by vertical scanning [5]. A non-moving agile scanning confocal microscope system employing variable-focal-length (VFL) micro-lenses also provided a way to axially scan the foci across a sample [6]. Meanwhile, a method and apparatus using multiple imaging paths with respect to different object viewing angles was developed to measure object height and volume using a combination of optical means and image processing techniques [7]. Another potential technique is the chromatic confocal microscopy, which takes the advantage of light for avoiding vertical scanning completely. A chromatic confocal microscope generally equips with a white-light source in combination with a diffractive lens as well wavelength-to-depth coding for profile measurement of a three-dimensional surface [8]. Nevertheless, none of the existing chromatic methods has so far explored the possibility of developing a compact reflective-lens objective for achieving axial light dispersion required for depth evaluation. Therefore, the research developed a chromatic confocal profilometer by designing a reflective-lens objective and line-slit confocal optical configuration for achieving line scanning surface profilometry. The chromatic objective was designed by using optimizing light refractive property with different materials of optical lenses and various optical configurations. A system calibration was also implemented to obtain accurate mapping between light wavelength and depth. The details of the system design and measurement principle are provided in the following sessions.
2. SLIT-SCAN CHROMATIC CONFOCAL SYSTEM SETUP AND PRINCIPLE 2.1 Traditional confocal and chromatic confocal principle Traditional confocal principle is a widely used method to reconstruct 3-D volumetric structures of a tested object. When the object locates on the focal plane of a confocal layout, the reflected light from the object s surface can pass thought a spatially filtering pinhole and reach to the detector with a maximum light intensity. However, when it is out of the focal plane, the detector will only obtain smaller amount of light and the detected light is attenuated significantly, as shown in figure 1. A vertical scanning can be used to establish a depth response curve of the light intensity, which primarily depends on the light wavelength and the numerical aperture of the objective. The light intensity detected in the above conjugative optical configuration can be shown in Eq. (1) [2]. Photo detector Illuminating aperture Pinhole respect to the light wavelength, an accurate mapping curve (shown in figure 4) between the light wavelength and the detected profile depth can be established by referencing a vertical-scanning reference mirror with a calibrated laser interferometer. With the knowledge of the detected wavelength with respect to its corresponding focal depth, the calibrated mapping function can be employed to detect height information without in-situ vertical scanning. White light source Spectrometer Conjugate spatial filter Spatial filter Object surface Beam splitter λ min λ m λmax Spectral Image λ min λ m λ max Wavelength Point light source Beam splitter Fig. 3 Schematic diagram of the developed chromatic confocal microscope. In-focus rays Out-of-focus rays Focal plane Sample Z-axis distance Peak Depth (μm) 50 40 30 20 Least-squares fit Focal length Fig. 1 Traditional confocal microscope 0 sin( u / 2) ( u) ( / 2) u 2 I (1) where 8 2 is the normalized axial coordinate; u Z sin 2 λ is the wave-length of the light source; z is the vertical distance; and α is the half angle of the field of view. The traditional confocal method described above definitely needs a vertical scanning operation for determining its focus peak. To avoid this time-consuming process, the chromatic confocal method employs chromatic light dispersion along the depth axis to obtain the mapping relationship between profile depth and light wavelength. In this research, a chromatic optical objective has been designed and fabricated to focus the incident white light onto a series of focal distances with respect to the corresponding wavelength, shown in figure 2. White light source λ1λ 2 λm λn Fig. 2 Illustration of axial chromatic light dispersion generated by the developed chromatic optical objective. The optical system and its spectrum intensity signal of the developed chromatic confocal system are illustrated in Figure 3. Using a spectrometer to analysis the received intensity response with 400 450 500 550 600 650 Wavelength (nm) Fig. 4 Calibrated mapping curve between light wavelength and profile depth. 2.2 Slit-scan Chromatic confocal microscope and measurement Principle Traditional chromatic confocal profilometers are in general a point-type profile scanner, which is not efficient for in-situ inspection. To achieve a better solution, a slit-scan chromatic confocal profilometer was developed here to perform continuous line-scan profile measurement. In general, the slit-scan has slightly less measurement resolution than the point scan; however, it has a larger field of view (FOV) and a higher measurement speed. The optical configuration of the developed slit-scan chromatic confocal system and its hardware setup are illustrated in Figure 5 and Figure 6, respectively. In the system, a white light supplied by a Xenon arc lamp is first coupled into an optical fiber and then focused by a cylinder lens to form a projected line-shape light source. It is then forwarded through a line aperture for refining its light shape and uniformity and further passing through a beam-splitter for redirecting light into the chromatic objective. Being chromatically dispersed by the objective, the incident light is then projected onto the tested object surface. Following the light being reflected back from the object s surface, it is then transmitted through the coaxial optical structure having a conjugate slit aperture and to be further received by a line spectrometer. The slit aperture is employed to filter out defocused light and differentiate focused light from other unfocused beams for generating a good depth detection resolution. The detected spectrum information includes the light wavelength, light intensity and
horizontal position. Figure 7 shows an example of the spectrum image of a step-height surface. By using the mapping function established in session 2.1, the cross-section surface profile (illustrated in Figure 7) of the measured step-height surface can be obtained. A full-field 3-D map can be further achieved by moving a translation stage along the scanning axis. Computer Line-spectrometer Cylindrical Lens White-Light Source 100W Xe Lamp Input slit aperture Translation stage Conjugate slit aperture Sample Beam splitters Wavelength Peak Fig. 5 Schematic diagram of the developed slit-scan chromatic confocal surface profilometer. Fig. 7 Measurement results of a step-height surface by using the developed slit-scan chromatic confocal system: the detected spectrum image; and the reconstructed crosssection profile. 2.3 Design of the chromatic objective To obtain a desired specification of the axial chromatic light dispersion with a specific measurement depth range and minimal image aberration, an optical lens design and optimization using commercial optics software was implemented by minimizing potential image aberrations and lens fabrication costs. The light disperse range can be accurately controlled by careful selection of various material properties of optical lenses with geometric parameters, the focal distance and the working distance. In a design example of a 20x chromatic objective by using biconvex lenses having various positive and negative meniscuses, an optical layout can be designed with its ray tracing shown in Figure 8. In this case, a light spectral bandwidth range between 400 nm and 700 nm was chosen for its optical simulation. With an optical simulation and optimization, the optical design can be optimized to have its focus spots approximately close to the light diffraction limit, illustrated in Figure 9. The circle diameter of the plot diagram represents the diameter of the Airy disc being equivalent to the diffraction limit of the light. Fig. 6 Hardware setup developed for the slit-scan chromatic confocal surface profilometry. Fig. 8 Optical simulation of a 20x chromatic objective Spectral axis 400nm 500nm 600nm 700nm Spatial axis Fig. 9 Optical simulations of the light focus spot diagrams with respect to various light wavelengths for the design of the chromatic objective. 2.4 Flow chart of the measurement method The flow chart of the developed measurement method is revealed in Figure 10. In here, a system calibration procedure including light intensity optimization, positioning verification of the translation stage and control of light exposure time for best image contrast is
performed to ensure an optimal spectrum imaging condition for chromatic scanning. Following this, implementation of a vertical scanning calibration using the measured object s surface is provided to establish the accurate mapping function between profile depth and light wavelength. With the calibrated function, the object can then be scanned to obtain its cross-section profiles by lateral scanning. 3. EXPERIMENTAL PROCESS AND RESULTS To attest the measurement accuracy of the developed measurement approach for multi-wavelength confocal surface profilometry, we conducted an experimental measurement on a calibrated step-height surface, shown in Figure 11. The overall measurable depth is 350 micrometers in the current system layout with the developed chromatic objective. By performing the measurement procedure described in the above session, Figure 11, (c) and (d) display the 3-D map, top view and cross-section profile being reconstructed from the step-height surface, respectively. From the analysis of the measured result, Table 1 illustrates the measurement accuracy and repeatability being obtained from a 30 time repeatability test. It indicates that the maximum measured error was less than 0.099 % of the overall measurement range. In the current system design, the depth measurement range can be ranged between few to several hundred micrometers and the depth measurement resolution can reach down to 0.03% of the overall detection range. The measurement speed is determined by the light exposure time required for adequate spectrometry imaging. It is generally determined by the imaging CCD frame rate and the surface reflectivity condition of the tested object. Meanwhile, an industrial sample of a micro brightness enhance film (BEF) with a top height of 99.2 μm was also measured for verification of its feasibility on an industrial micro structure fabricated by Roll-to-Roll nanoimprinting processes. The measurement FOV was within a rectangular FOV area of 0.5*4.0 mm 2 for demonstration. The physical sample, 3-D shape, top view map and cross-section profile of the measured result are shown in Figure 12 -(d), respectively. From the measured results, it indicates that the developed method and system are capable of measuring insitu micro structures accurately. The measuring speed can reach up to 60 surface cross sections per second. System calibration Calibration curve Measuring sample Cross-section profile Lateral scanning 3D reconstruction Fig. 10 Flow chart of the developed measurement method. 4. CONCLUSIONS A slit-scan chromatic confocal surface profilometer was developed to achieve in-situ line-scan surface profilometry of micro structures without needing time-consuming vertical scanning operation. A chromatic optical objective can be designed and optimized by modulating various refractivity of multiple-wavelength light through optical lenses. A desirable measurable depth range can be obtained by designing the optical layout of the chromatic objective and it can be ranged from a few to several hundred micrometers. The detection resolution can be theoretically reaching down to 0.03% of the overall detection range. From the experimental tests, it was verified that the maximum measurement error can be controlled less than 0.1% of the overall measurement range with a repeatability of 52 nm within one standard deviation. The developed method and measurement system can be widely employed to in-situ microstructure profile measurement. The measurement speed of the system is currently limited by the frame rate of the imaging unit, which requires further development to achieve its best performance. Standard step-height surface Reconstructed 3-D map
(c) Top view of the 3-D map Reconstructed 3-D map (d) Cross-section profile Fig. 11 Measurement results of a standard step-height surface. (c) Top view of the 3-D map Table 1 Measurement accuracy evaluation from a 30-time test sample Standard height(μm) Measured height (μm) 1 2 3 4 100.9 50.6 25.3 10.10 100.96 50.64 25.32 10.11 Error (%) 0.059 0.079 0.079 0.099 Standard deviation (μm) 0.042 0.052 0.048 0.046 Sample of a micro brightness enhancement film (BEF) (d) Cross-section profile of the 3-D map Fig. 12 Measurement results of of a micro BEF sample REFERENCES 1. Minsky, M., "Microscopy Apparatus," U.S. Patent, No.:3,013,467,1961. 2. Wilson, T., "Confocal Microscopy, " Academic, London, 1990. 3. Bitte, F., Dussler, G., and Pfeifer, T., "3D micro-inspection goes DMD," Optics and Lasers in Engineering, Vol. 36, pp. 155 167, 2001. 4. Tanaami, T., Otsuki, S., Tomosada, N., Kosugi, Y., Shimizu, M., and Ishida, H., "High-speed 1-frame/ms scanning confocal microscope with a microlens and Nipkow disks," Applied Optics, Vol. 41, No. 22, pp. 4704-4708, 2002. 5. Ishihara, M., and Sasaki, H., "High speed surface measurement using a nonscanning multiple-beam confocal microscope," Optical Engineering, Vol. 38, Issue 6, pp. 1035-1040, 1991. 6. Raighne, A. M., Wang, J., Cabe, E. M., and Scharf, T., "Variable focus microlenses: Issues for confocal imaging," Proceedings of SPIE, Vol. 5827, pp. 12-22, 2005. 7. Seng, T. P., "Hybrid Confocal Microscopy," U.S. Patent, No.:5,880,844, 1999. 8. Paul, L. C., Chen, S. P., Lijun, Z., and Yeshaiahu, F., "Singleshot depth-section imaging through chromatic slit-scan confocal microscopy," Applied Optics, Vol. 37, No. 28, pp. 6764-6770, 1998. 9. Wilson, T., Hewlett, S. J., and Sheppard, C. J. R., "Use of objective lenses with slit pupil functions in the imaging of line structures," Applied Optics, Vol. 29, No. 31, pp. 4705-4714, 1990.