Confocal principle for macro- and microscopic surface and defect analysis

Transcription

1 Confocal principle for macro- and microscopic surface and defect analysis Hans J. Tiziani, FELLOW SPIE Michael Wegner Daniela Steudle Institut für Technische Optik Pfaffenwaldring Stuttgart, Germany tiziani@ito311.ito.uni-stuttgart.de Abstract. A confocal setup based on microlenses for shape investigation, defect analysis and surface topography measurement is presented. The major advantage of this technique is its high light efficiency and the possibility to realize larger object fields without reducing the numerical aperture. Different variations of the setup for different applications are presented. An increased working distance yields a greater variety of its applications. Furthermore, the arrangement of the microlenses on a rotating disk leads to an increased spatial sampling and a high scanning rate. The axial resolution is the same as in a comparable confocal microscope based on a Nipkow disk. The microlens confocal system enables measurements on large field sizes down to microscopic ones. In addition, by using chromatic aberrations it is possible to achieve realtime images with color-coded height information. The topography of the sample can be determined from one color image, which leads to a reduction in measuring time. To reduce measuring times on curved surfaces, for instance, it is useful to adapt the focal lengths of the microlenses to the individual shape of the object. Hence, only the difference between the focal distribution and the real shape must be determined Society of Photo-Optical Instrumentation Engineers. [S (00) ] Subject terms: confocal microscopy; microlenses; surface and shape analysis. Paper SM-03 received May 20, 1999; revised manuscript received July 20, 1999; accepted for publication July 20, Introduction Different techniques for measuring surface topography are based on, e.g., triangulation, projected fringes, interferometry, and holography. In addition, the confocal principle is also suitable, as it leads to a true depth discrimination. 1 3 For the confocal scanning principle, those parts of the object that are located in the focus appear sharp and bright in the corresponding image plane, whereas parts that are outside the focus appear blurred and dark. Even for phase objects with no amplitude variations, the 3-D structure can be analyzed. High lateral and depth resolutions can be obtained with the confocal principle when high-numerical-aperture objectives are used. Furthermore, for parallel processing, diffraction-limited 2-D spot arrays can be projected onto the object, as first proposed by Petran et al. 4 and Xiao et al. 5 They used a Nipkow disk to generate the pinholes. For high spatial and depth resolutions, a high-numericalaperture lens must be used for the projection of light spots onto the object. The price paid for a highly magnifying microscope objective with a very high numerical aperture 0.8 is a low field size. For contouring large object fields with high depth discrimination, a microlens array in which each microlens has a high numerical aperture is suitable. A large number of microlenses can be considered for a large object field without a reduction in the depth discrimination. The numerical aperture and the object field are no longer directly coupled. When a microlens array is used, the numerical aperture is defined by the size of the individual lenses, and the object field is defined by the size of the array and the number of detector elements. Some options to extend the principle of confocal microscopy to the topography of optical and some technical surfaces in mechanical engineering as well as for surface characterization are presented. The use of a rotating microlens disk enables fast and parallel scanning with high light efficiency and without loss in depth discrimination in comparison to Nipkow disk based systems. Furthermore, a strong reduction of measuring time can be achieved using the chromatic length aberration of the optical system for the determination of the topography. Another way to reduce the amount of necessary height steps, especially on curved surfaces, is to adapt the focal distribution of the microlens array to the shape of the object. Depending on the individual object and the measurement task, different setups with varying field sizes from 100 m to centimeters can be applied. 2 Confocal Principle with Microlens Array To extend the object field and to improve the light efficiency a microlens array with as many as elements was used, as shown in Fig. 1. Light coming from a laser or, alternatively, from a xenon arc lamp is collimated by a lens. The parallel beam, as shown schematically in Fig. 1, is focused by the microlenses to a 2-D point array. The microlenses focus the spots nearly into a plane. A parallel beam is formed from object points in focus when passing the microlenses a second time on the return path. The light spots reflected from the object points in focus are 32 Opt. Eng. 39(1) (January 2000) /2000/$ Society of Photo-Optical Instrumentation Engineers

2 Fig. 1 Arrangement for confocal 3-D analysis with a microlens array. formed at infinity and focused by lens L 1 to a point at the pinhole. After passing a spatial filter pinhole in the back focal plane, the light reaches the CCD chip of the camera. In this case, the pupils of the microlenses are imaged in the detector plane and not the light spots as usual. The image of the camera is processed by a computer with a frame grabber. This image is digitized and stored in memory. The depth discrimination arises in a manner similar to conventional confocal microscopy 1,2,6 : If the object is moved out of the focal plane, the single points of the 2-D point array are no longer in autocollimation. A small portion of the light passes the pinhole, and a reduced amount of light flux reaches the detector elements. The axial response Fig. 2 for a confocal system with microlenses can be approximated by the on-axis intensity of one microlens in reflection as 6 I u sin u/2 2 u/2 with u 2 l 2 4 f m 2 z, 1 where l is the diameter, f m the focal length of one microlens and the illuminating wavelength. Here z describes the linear movement of a perfect mirror as object. For the intensity to fall to half of its maximum value at a defocusing distance z 1/2, hence, the full width at half maximum FWHM 2 z 1/2 of the intensity response obtained by moving continuously through the focus is FWHM cos, Fig. 2 Axial response I(z). 2 where sin is the numerical aperture of the microlenses. To measure the focal length variation of the entire array, the brightest height section can be found for each microlens from the axial response. The height determination can be improved by taking the center of gravity of each intensity series. Compared with the determination of the maximum, the influence of the camera noise is reduced and a higher accuracy can be obtained. Using this approach, a sensitivity of 50 nm for the focal length determination was achieved for a numerical aperture of 0.3. Because we use the pinhole as a spatial-frequency filter, the diameter of the pinhole d pinhole should be small enough to maintain the optical depth sectioning of the confocal microscope as Wilson described in Ref. 2. On the other hand, we must conserve the images of each microlens on the CCD array. Therefore 2.44 f L /l d pinhole 10 f L /l, where f L is the focal length of the imaging lens. The pinhole represents the coherent detector for each microlens because the light detected on the CCD pixels is the sum over the amplitude field in the pinhole area, weighted with a phase factor, and squared. To obtain the optimum lateral resolution and to avoid crosstalk by stray light, the detector plane should be in the conjugated image of the microlens pupils. This is particularly important when diffractive optical elements Fresnel zones are used. Normally a diffractive element scatters more light than a refractive element due to the rougher surface. Furthermore, the image of the microlens pupil on the detector is not changed by defocusing. 3 Confocal Principle Using Diffractive Elements for Increased Working Distance By introducing diffractive or refractive microlens arrays the light efficiency can be increased by a factor of 80, for instance, as compared to the arrangement with a Nipkow disk. Many thousands of microlenses working in parallel lead to a field size increase to a few centimeters. Furthermore, the high axial resolution, depending on the high numerical aperture of the diffractive elements, can be maintained. Figure 3 shows our setup with diffractive Fresnel elements called microlenses arranged in a spiral shape on a rotating disk. This combines the fast scanning rate of a Nipkow disk with the high light efficiency of a microlens based system. No additional pinholes were used in our arrangement and the light source was an LED, at a wavelength of 780 nm and an output power of 35 mw, collimated to illuminate the microlenses on the rotating disk, as shown in Fig. 3. The design of the disk and the masks for fabrication were developed at the Institut für Technische Optik ITO, the microlenses were fabricated at the PSI Paul-Scherrer-Institut, now part of the Centre Suisse d Electronique et de Microtechnique CSEM, Zurich. They are arranged on the disk, with a nearly perfect fill factor, in many spirals so that the spots in the focal plane scan the hole area about 800 times/s like the pinholes on a Nipkow disk. The light spots are imaged into the object 3 Optical Engineering, Vol. 39 No. 1, January

3 Fig. 4 Graycoded topography of a one dollar note. Fig. 3 Confocal setup with rotating microlens disk. plane by objectives L 1 and L 2. The depth discrimination is provided by a pinhole in the focal plane of the imaging lens that images the pupils of the microlenses on the camera. The camera is a CCD using pixels. A polarizing beamsplitter together with a /4-plate can be used to improve the SNR and to prevent reflected light from the lower side of the microlens disk. The size of the object field can be varied by changing the objective or by changing the imaging optics and thereby changing the number of microlenses used. Different field sizes for macro- and microscopic measurements were realized. Together with photoobjectives, about 6400 microlenses were used, for instance, leading to a field size diameter of 7 mm. In the microscopic range a field diameter down to 100 m was used. To obtain the object topography a z-scan is performed by taking one camera frame per height section. The axial position of the confocal signal in each camera pixel gives the topographic value. Because maximum detection can be very noisy, we determine the center of gravity of the signal. For very large field sizes with high lateral and depth resolution, single field measurements can be combined. Between the measurements, the sample must be laterally shifted, therefore the system must be equipped with an additional x- y-stage. A small overlap is necessary to align the measurements according to each other. Of course, this yields longer measuring times. This new setup results in a very high light efficiency, due to the use of microlenses diffractive with a nearly perfect fill factor, enabling the use of an inexpensive LED instead of a Xe arc lamp. With an Olympus objective we measured an FWHM of 0.8 m instead of the theoretical FWHM of 0.5 m. The enlargement is mainly due to residual aberrations of the optical system. Nevertheless on a /20 mirror, root mean square rms height deviations of 2 and 13 nm were measured with and objectives, respectively. High light efficiency without the requirement of additional processing or loss of optical sectioning strength was obtained. A small drawback is a slight reduction in the lateral resolution of the system. Because of the imaging of the pupils of the microlenses on the CCD chip, it is about a factor 1.2 worse than when imaging the spots using pinholes on a Nipkow disk. This may improve as is shown by Ichihara et al. 7 using microlenses to focus on the pinholes of a rotating Nipkow disk. This may lead to a slightly better lateral resolution but has stronger requirements for fabrication and alignment of the two disks. A great variety of objects with different reflectivities was measured such as glass, paper, plastics and so on. Figure 4 shows an example measured on a one dollar banknote where a field size of m 2 was used. One can clearly see the fibers of the paper. The image is graycoded, the upper fibers appear white and deep holes are black. The total height of the topography is 16 m. As an example of a larger field size such as 7 7 mm 2,a measurement on black grained leather with a total height of 171 m is shown in Fig. 5. To test the reliability of the measurements, the confocal measurements using a rotating microlens disk were compared with tactile measurements on a roughness standard Fig. 5 Grained leather surface measured with diffractive rotating disk. 34 Optical Engineering, Vol. 39 No. 1, January 2000

4 Fig. 7 Chromatic confocal arrangement with four diode laser sources. Fig. 6 Comparison of confocal and tactile measurements on a roughness standard (R a 3.40 m, R z 18.1 m, R max 18.9 m). mean roughness R a 3.40 m, R z 18.1 m, R max 18.9 m, carried out by the PTB Physikalisch Technische Bundesanstalt, Braunschweig. Figure 6 shows a comparison of the confocal and tactile measurements. The first profile is a composition of six confocal measurements with a field size of 810 m 2 each and 80 m overlap between them, while the second is the tactile profile determined with a stylus radius of 4 m. 4 Chromatic Measuring Principle For a confocal microscope illuminated with different wavelengths or white light, and with an objective lens system with chromatic length aberration f, where the spherical aberration should be corrected for each wavelength, each color has its own focal plane region. The chromatic length aberration provides us with the possibility of depth registration by color decoding. For a well defined determination of height, the system must be calibrated. Therefore a perfect mirror is shifted stepwise in the axial direction. Depending on the position of the mirror, the image changes its color. Every color indicates a certain height range. By knowing the color-related height an object s topography can be measured without shifting the object or lens system axially. The chromatic confocal principle using a Nipkow disk together with a microscope objective was first described in Ref. 8. Furthermore a chromatic principle works with diffractive elements by sequential selection of the different wavelengths of the light source Fig. 7. In our experimental setup, four stabilized semiconductor diode lasers were used instead of a white light source where color filters were used. The diode lasers were focused by graded index GRIN lenses into optical fibers single mode and were coupled together to a single point source. The light of the four lasers was coupled into a monomode fiber and collimated, as shown in Fig. 7. In the microlens arrangement, microlenses were used. The advantage of diffractive microlenses zone plates is the large chromatic aberration together with an inverse proportionality between focal length f and wavelength when operating in the first order. 9 The working range is given by the chromatic length aberration of the Fresnel zone plates. It was found by computer simulation that the optimal distance between two wavelength focal lengths matches the FWHM. Hence, the wavelength chosen were , 780, 810 and 840 nm. When using a combination of diffractive and refractive elements, the flexibility of the wavelength selection can be extended. Furthermore a three color chip CCD detector is used to improve the speed of the measurement and analysis. Therefore the red, green and blue colors were chosen and adapted. Note that refractive systems have a nonlinear linear chromatic length aberration with respect to the wavelengths. However, the linear chromatic aberration of the diffractive elements has an opposite sign to the refractive elements. In our new experimental setup, similar to the one shown in Fig. 3, the shift of the axial focus position f with wavelengths from 400 to 700 nm of a Xe arc lamp is about 10 m. A typical field size diameter is 1 mm. Remember that by using chromatic length aberration no mechanical axial scanning is required. By looking at objects with a height variation smaller than f all parts appear bright and sharp simultaneously but with different colors. Each color corresponds to a different height of the object points. With a color CCD camera, real-time images that already contain the color-coded height information are obtained. They enable a direct analysis of the object topography from the live image. To determine the topography of the object from this color-coded image, it is necessary to calibrate the system. 9 Therefore a plane mirror is shifted axially and the portions of red, green and blue in the color channels of the CCD correspond to the axial position of the mirror Fig. 8 a. Each position can be described by the ratio of the three intensities. Figure 8 b shows a typical calibration curve. On the horizontal axis is the blue portion of the normalized Optical Engineering, Vol. 39 No. 1, January

5 Fig. 9 A 3-D plot of a laser textured metal sheet with a mean surface roughness of R a 1.43 m, R q 1.77 m, and R z 10.1 m. Fig. 8 (a) Measured intensity of different color channels depending on the axial position of the object point and (b) calculated look-up table with typical calibration curve. total intensity of the three colors, the vertical is the normalized red portion. I b z x b z I r z I g z I b z, I r z x r z I r z I g z I b z. The portion of the green intensity is implicitly contained due to the normalization. Each point on this calibration curve corresponds to one height position. Because of noise, measured intensity ratios may differ from this curve. In this case, the nearest point on the curve is taken. Therefore for each combination of x b and x r, with a distance smaller than 0.15 to the calibration curve, the corresponding height value is calculated. Measured combinations outside this range are set invalid. This calculation, together with the calibration curve, forms a 3-D look-up table. Figure 8 b shows such a look-up table with graycoded height values. It must be determined only once and can be used for all the following measurements. 4 5 However, different object materials may influence the spectral reflectivity and therefore change the ratios of the color intensities. This can be determined with the look-up table using an appropriate calibration object of the same material. With this look-up table, the transformation of a colored image on the CCD into the topography of the object can be very fast. The chromatic principle enables fast confocal measurements without mechanical scanning. To improve the SNR it is, however, recommended to take more than one image and average. This method with one color image leads to a slight reduction in axial resolution compared to the mechanical focusing with 200 to 300 height sections. In a measurement with a /20 mirror and a objective an rms deviation of 30 nm was found. Figure 9 shows a typical example where a laser textured metal sheet with special oil containing attributes was measured. The value of the mean surface roughness is R a 1.43 m, the variance R q 1.77 m and the 10 point height R z 10.1 m. The chromatic measurements were compared with tactile measurements of the PTB tip radius of the stylus: 0.1 m. Figure 10 presents a chromatic confocal measurement and the corresponding part of the tactile measurement. Analysis of some roughness parameters as well as a direct comparison of the profiles show good agreement considering that only one averaged color image was necessary for the measurements. The use of microlenses improves the light efficiency. Fast measurements with a slight reduction in resolution can be obtained with the chromatic principle when a white light source and a color CCD camera are used. No mechanical scanning is necessary because of the use of the chromatic length aberrations. 5 Microlens Arrays on Curved Surfaces The proposed confocal principle for shape comparison and defect analysis can be extended to different object shapes, such as curved surfaces, using microlens arrays with different adapted focal lengths. Lens arrays with spatially varying focal lengths can be used on nonplanar surfaces, the overall profile height of 36 Optical Engineering, Vol. 39 No. 1, January 2000

6 Fig. 12 Focal length distribution for a stairlike object. Fig. 10 Comparison of chromatic confocal and tactile measurements. which exceeds the vertical measurement range of a microlens array based setup but is known or calibrated. The lens array focal surface is shaped to fit the expected object shape. In the measurement, only the deviation of the object surface from the focal surface must be detected, the maximum accessible profile depth is therefore considerably increased. Furthermore, it enables even faster measurements since the total vertical scan movement is reduced. There are two principles for adapting the focal length. One method is the flexure of a flexible microlens array, where all microlenses MLs are of the same focal length. The left side of Fig. 11 shows this principle. In this case, the number of possible adaptable shapes is limited by the natural stiffness of the carrier material, in our case a polyurethane sheet. This ML array was used for a cylindrical or aspherical shape and bent in one direction only. More extended possibilities are offered if a nonflexible ML array containing a varying focal length distribution is produced. As shown in Fig. 12, a focal length distribution for a stairlike object, measured on a plane mirror to explain the principle, was designed on a field of nearly mm 2. The step between the single stairs is 50 m, the maximum height range of the ML focal length varies in a range of 300 m. The application on a suitable stairlike sample will show only defects. Figure 13 shows the measured focal length distribution of a lens array fitted on a cylinder. It consists of a height of 201 m on the measured cylinder with a radius 30 mm. For lens arrays with spatially varying optical functions, the optical design of the individual lenslets and the preparation of the laser writing exposure data requires special consideration. Having 8-bit resolution and submicrometers pixel sizes, the amount of exposure data increases to the order of several gigabytes for the lens arrays discussed here. Real-time computation of the exposure data is therefore favorable to avoid such large data files. 10,11 In the confocal condition, the light reflected from the object to be measured must pass through the same ML aperture on its way back to fulfill the measurement principle. This can be ensured by normal incidence of the light onto a curved polished surface and is slightly relaxed by a technical surface. Arrays fulfilling this condition have been designed and fabricated for focusing onto cylindrical and spherical surfaces. The normal incidence condition is fulfilled by a simultaneous variation of the focal length and the beam deflection angle. Fig. 11 Different methods for adapting the focal lengths on a curved surface. Fig. 13 Focal length distribution of an ML array adapted on a cylinder surface; field size is 7 7 mm 2. Optical Engineering, Vol. 39 No. 1, January

7 Fig. 14 Normal incidence by off-axis MLs. Fig. 16 Difference topography obtained from a measurement on a calibration sphere with a radius of mm. For off-axis MLs, the chief ray is deflected by a certain angle. With variable off-axis parameters the angle of incidence onto the surface under test can be normal over all of the measurement area, as shown in Fig. 14. Here an example for the adaptation onto a spherical or cylindrical shape is demonstrated. Due to the off-axis arrangement, the amount of reflected and detected light is independent of the slope of the test surface. The pupils of the MLs are equally illuminated in the focused state. The focal length distribution for such a spherical object with radius R is described in dependence of the radial coordinate r by f r f min R R 2 r 2 1/2, where the apex focal length is f min. To test this method an ML array was evaluated and produced in cooperation with the PSI now CSEM in Zurich. The main application will not be the testing of polished spherical surfaces, e.g., 6 lenses, which can be analyzed with classical methods such as interferometric methods. But the application to complex surface shapes such as aspherical surfaces seems to be very promising for instance, different aspheric surfaces on a substrate that must be measured together. The pseudo 3-D representation of the four aspheric surfaces is shown in Fig. 15. For comparison, a design was adapted on a planoconvex lens with a nominal focal length of 100 mm. The radius of curvature was supposed to be mm. A calibration was required to determine whether a deviation results from the object or the focal length distribution. A calibrated sphere from a Twyman-Green interferometer with a nominal radius of curvature of mm was available and served as an exactly shaped reference. The measurement is shown in Fig. 16. Due to a difference in the radius of curvature, a topography with an apex height of 48 m was obtained. Although the surface slope did not fit the focal length distribution, the measurement due to the off-axis arrangement was possible. A defect in the form of a scratch was measured on a shape adapted cylinder Fig. 17. The cylindrical shape of the object agreed, therefore the defect occurred essentially in a plane topography. The axial resolution of the confocal measurements depends on the width of the axial response. This width does increase with decreasing numerical aperture, therefore, the vertical resolution varies with the lenslets numerical aperture over the array. If a constant vertical resolution is required, constant numerical aperture lenslets must be used, leading to larger lenses at the edges and therefore less dense packing. The choice to have a regular array of sampling points was made for convenience of data generation. Fig. 15 Lens with aspherical surfaces. The variations of the surface slopes and the shouldered facets favor the method with adapted focal length distribution; field size is 7 7 mm 2. 6 Conclusion Different applications of the confocal principle for the analysis of shape and topography of technical surfaces were presented. The use of MLs provides a very high light efficiency. A new confocal setup with increased working distance is based on a rotating microlens disk. It was applied for different field sizes from 100 m to 7 mm. The arrangement 38 Optical Engineering, Vol. 39 No. 1, January 2000

8 6. H. J. Tiziani, R. Achi, R. N. Krämer, and L. Wiegers, Theoretical analysis of confocal microscopy with microlenses, Appl. Opt. 35 1, A. Ichihara, T. Tanaami, K. Isozaki, Y. Sugiyama, Y. Kosugi, K. Mikuriya, M. Abe, and I. Uemura, High-speed confocal fluorescence microscopy using a nipkow scanner with microlenses for 3-D imaging of single fluorescent molecule in real time, Bioimages 4 2, H. J. Tiziani and H.-M. Uhde, Three dimensional image sensing with chromatic confocal microscopy, Appl. Opt , H. J. Tiziani, R. Achi, and R. N. Krämer, Chromatic confocal microscopy with microlenses, J. Mod. Opt. 43 1, H. J. Tiziani, R. Achi, R. N. Krämer, T. Hessler, M. T. Gale, M. Rossi, and R. E. Kunz, Microlens arrays for confocal microscopy, Opt. Laser Technol. 29 2, T. Hessler, M. Rossi, J. Pedersen, M. T. Gale, M. Wegner, D. Steudle, and H. J. Tiziani, Microlens arrays with spatial variation of the optical functions, Pure Appl. Opt. 6, Fig. 17 Defect analysis result obtained with an adapted focal length distribution on a cylinder; field size is 7 7 mm 2. of the lenses in spirals on a rotating disk enables rapid and parallel measurement, and the same axial resolution as in a Nipkow disk based system was obtained. With a slight reduction in axial resolution, very rapid measurements were realized using a white light source and a color CCD camera. No mechanical scanning is necessary because of the use of the chromatic length aberrations. Finally, the adaptation of the focal distribution of the MLs on the basic shape of the object was demonstrated, which enables an extension of the range of applications and shorter measuring times. Furthermore, off-axis MLs were designed to lead to normal incidence on curved surfaces. Experimental results support the potential of the methods. Acknowledgments We acknowledge the financial support by the Deutsche Forschungsgemeinschaft DFG. Furthermore we would like to thank PSI now CSEM for the fabrication of the microlenses and the PTB for providing the tactile measurements. References 1. C. J. R. Sheppard and D. M. Shotton, Confocal Laser Scanning Microscopy, BIOS Scientific Publishers, Oxford T. Wilson, Ed., Confocal Microscopy, Academic, London H. J. Tiziani and H.-M. Uhde, Three-dimensional analysis by a microlens-array confocal arrangement, Appl. Opt. 33 4, M. Petran, M. Hadravsky, M. D. Egger, and R. Galambos, Tandemscanning reflected-light microscope, J. Opt. Soc. Am. 58, G. Q. Xiao, T. R. Corle, and G. S. Kino, Real time confocal scanning optical microscope, Appl. Phys. Lett. 53, Hans J. Tiziani received his DiplIng degrees in mechanical engineering from St. Gallen and in optics from the Institut d Optique of Paris, his PhD from the Imperial College London, and his Habilitation from the ETH Zürich. He directs the Institute of Applied Optics and is a professor with the University of Stuttgart. He was a consultant at IBM, San Jose, California, and headed the optics group at the Institute of Applied Physics, ETH Zürich. He directed the central laboratory at Wild, Heerbrugg. He is past president of European Electrooptics, past president of the optics division of the European Physical Society, and past governor and fellow of SPIE. His main experience is in image formation, optical system and transfer function analysis and measurement, interferometry, holography, metrology and inspection, and 3-D and surface profile measurements. He has written more than 200 publications on applied optics, image formation, optical metrology, and optical surface measurement techniques. Michael Wegner studied mechanical engineering at the University of Stuttgart, Germany, where he received his diploma degree in He has since been engaged in research at the Institut für Technische Optik in Stuttgart in the field of scanning confocal microscopy. His interests are methods for 3-D surface topometry and microlenses. Daniela Steudle graduated in physics in 1996 from the University of Stuttgart. Currently she is working for her PhD at the Institut für Technische Optik, Stuttgart, with Prof. Tiziani. Her main research interests are confocal microscopy and optical 3-D surface topometry. Optical Engineering, Vol. 39 No. 1, January