Compact optical design for dual-axes confocal endoscopic microscopes Michael J. Mandella*a, Jonathan T. C. Liua, Wibool Piyawattanamethab,c, Hyejun Raa, Pei-Lin Hsiungb, Larry K. Wonga, Olav Solgaarda, Thomas D. Wangb, Christopher. H. Contagb, Gordon S. Kinoa a Ginzton Laboratory, Stanford University, Stanford, CA, 94305 *Email:
[email protected] b Clark Center for Biomedical Engineering and Sciences, Stanford University, Stanford, CA 94305 c National Electronics and Computer Technology Center (NECTEC), Pathumthani 12120, Thailand ABSTRACT Here we describe a simple optical design for a MEMS-based dual-axes fiber optic confocal scanning microscope that has been miniaturized for handheld imaging of tissues, and which is capable of being further scaled to smaller dimensions for endoscope compatibility while preserving it’s field-of-view (FOV), working distance, and resolution. Based on the principle of parallel beams that are focused by a single parabolic mirror to a common point, the design allows the use of replicated optical components mounted and aligned within a rugged cylindrical housing that is designed for use as a handheld tissue microscope. A MEMS scanner is used for high speed scanning in the X-Y plane below the tissue surface. An additional design feature is a mechanism for controlling a variable working distance, thus producing a scan in the Z direction and allowing capture of 3-D volumetric images of tissue. The design parameters that affect the resolution, FOV, and working distance are analyzed using ASAPTM optical modeling software and verified by experimental results. Other features of this design include use of a solid immersion lens (SIL), which enhances both resolution and FOV, and also provides index matching between the optics and the tissue. Keywords: confocal, endoscope, microscope, SIL, MEMS, dual-axes, fiber, scanner, tissue
1. INTRODUCTION The motivation for this new design of a dual-axes confocal microscope is to produce a microscope that is capable of being miniaturized to dimensions that are compatible for use with endoscopes and which also meets the imaging performance goals set for in vivo imaging of gastrointestinal tissues. The first important improvement of the traditional confocal scanning microscope for endoscopic applications was achieved by using optical fibers and a micromachined scanning mirror, which allowed the required miniaturization of the microscope for in vivo imaging1. Also, the MEMS mirror provided faster scanning, which is another important requirement for in vivo imaging. But in spite of these improvements, the ability to resolve cellular details in highly scattering tissue has proven to be the next big technical challenge that must be addressed before a successful in vivo microscope could be realized. The need to image through many layers of scattering tissue has historically been addressed by increasing the numerical aperture (NA) of the objective lens. But, to get a sufficient working distance from a high NA objective lens, the objective’s diameter must become larger, thus technical contradictions exist between the need to use a high NA objective, the need to have a long working distance, and the need to make the microscope smaller. Additional imaging properties that are also important for endoscopic applications include having a sufficient dynamic range and field-of-view (FOV) to capture 3-D volumetric images of tissue. Furthermore, if in vivo confocal imaging is to become a widely accepted tool for clinicians and researchers, then the instrument must be of a simple design that uses only inexpensive replicated components, and which is also easy to assemble and align. In view of all the above requirements, we have chosen to develop a new class of microscopes based on the dual-axes confocal architecture. This architecture has been shown to have the ability to remove many of the existing design contradictions of traditional single-axis confocal microscopes, thus allowing many of the above requirements to be realized while the microscope is scaled to smaller dimensions.2,3,4 For example, the dualaxes confocal microscope using a post-objective scanning mirror can provide a relatively large FOV, and long working distance, as well as high axial resolution.5,6 Also, the dual-axes confocal configuration is shown to be more sensitive to ballistic photons from deep within tissue because the separation of the illumination and collection beams provides increased rejection of multiply scattered photons, and thus provides greater dynamic range and better tissue penetration.7,8,9 These are all properties which are important for obtaining vertical cross-sectional images in tissue in endoscopic applications where the goal is to diagnose disease in tissue as far deep as 500 microns below the surface.
Three-Dimensional and Multidimensional Microscopy: Image Acquisition and Processing XIV, edited by Jose-Angel Conchello, Carol J. Cogswell, Tony Wilson, Proc. of SPIE Vol. 6443, 64430E, (2007) 1605-7422/07/$18 · doi: 10.1117/12.699484 Proc. of SPIE Vol. 6443 64430E-1
focusing surface
Collimated beam
SIL index-matching interface
2-D Scan mirror
Focused spot scanned in x-y plane parabolic mirror Fig. 1. Showing the primary design features of the MEMS-based dual-axes post-objective scanning confocal microscope, which includes two collimated beams focused by two respective surface regions of a single parabolic mirror to a common focal point, and a 2-D MEMS scanner comprising two mirror surface regions on a single gimbalmounted mirror for synchronously deflecting the focused beams back towards the focusing mirror through a centrally located hole in the parabolic mirror. The beams are coupled into the tissue by a hemispherical SIL located in the central hole of the parabolic mirror.
2. SYSTEM DESIGN The primary design features of the dual-axes confocal post-objective scanning microscope are shown in figure 1. Two collimated beams are each focused to a common focal point by a single parabolic focusing surface and then deflected towards the tissue by a 2-D MEMS scanner.10,11,12,13 The scanning mirror controls the position of the intersection of the two beams within the tissue along two transverse directions X and Y. The scanning mirror rotates about the X and Y axes at predetermined sinusoidal frequencies to produce a raster scan of the common focal point within an approximately horizontal imaging plane inside the tissue. Also in figure 1, a solid immersion lens (SIL) index-matching element is shown mounted within a central hole in the parabolic focusing mirror. The flat side of the SIL is placed against the tissue surface for coupling the beams into the tissue and minimizing aberrations of the focused beams. We have found through simulation and experiment that fused silica with an index of 1.46 is a suitable material for the SIL, which provides sufficient index matching to the tissue. This index matching is important for reducing aberrations which normally occur when focused light passes through an interface having a significant index difference. The MEMS chip with dimensions 3.2 mm × 2.9 mm is designed to be mounted inside a 5 mm diameter package but is presently being used in a larger package to facilitate alignment of the two beams and can be used for handheld imaging. The parabolic mirror used in the first prototype microscope has been fabricated using a molding process, which provides a replicated component having a surface accuracy and smoothness needed for diffraction-limited focusing of the collimated beams. For design flexibility, our first prototype uses a 10 mm diameter replicated parabolic mirror. Future devices will use a 5 mm diameter parabolic focusing mirror in a smaller package for endoscopic compatibility. Miniaturization of the dual-axes confocal architecture requires a package design that allows precise mounting of the following primary optical elements: two fiber-coupled collimator lenses, a MEMS scanner, a parabolic focusing mirror, and a hemispherical index-matching SIL. In addition to these primary optical elements, four rotating wedges (Risley prisms) are used for alignment of the collimated beams, as shown in figure 2. A cut-away view of the housing is shown in figure 3, which shows both
Proc. of SPIE Vol. 6443 64430E-2
collimators located by a pair of precision machined v-grooves. The combined precision of the v-grooves and the preassembled fiber collimators allows the collimated beams to be parallel to each other within 0.05 degree accuracy. These alignment wedges have a small (0.1 degree) wedge angle, which steers the collimated beam over a maximum range of about 0.05 degrees in any direction as each wedge is rotated. It requires two wedges in each beam to achieve complete cancellation of this beam steering effect, which occurs when two prisms are rotated 180 degrees relative to each other about their common axis. A total of four rotating wedges are thus required for maximum flexibility to bring any two collimated beams into a parallel relationship for proper alignment. The rotating wedges are then used to provide fine steering of the collimated beams to bring the system into final alignment. Once the collimated beams are aligned parallel to each other, then the parabolic mirror provides a “self-aligning” property for rest of the system, which forces the focused beams to intersect at a common focal point within the tissue. This property allows the parabolic mirror and MEMS mirror to be repeatedly removed and replaced without requiring re-alignment of the beams. The housing also provides cylindrical pockets for holding the rotating alignment wedges. The package design must also accommodate a slider mechanism, which is a mechanical feature used for Z-scanning the MEMS scanner to provide a variable working distance within the tissue during imaging. This feature allows different imaging depths to be selected during imaging for collection of 3-D volumetric images of tissue. The slider mechanism is shown in figure 4, which comprises three rods, that slide within respective holes drilled into the housing. The MEMS scanner chip is glued to a PCB as shown in figure 4, which allows the wires from the bondpads to be connected to the external control wires for energizing the scan mirrorr. The slider has a mounting surface for the PCB / MEMS assembly, and is movable in the Z direction by a piezoactuator (not shown) for varying the working distance of the focal point within the tissue with micron accuracy. The PCB is mounted to the slider mechanism using miniature screws and washers. The mounting holes in the PCB are oversize to provide 200 microns of adjustment in X and Y to insure that the MEMS chip is properly positioned so that each of the beams hit the respective center regions of the dual mirror surface of the scanner.
Construction of first MEMS-based dual axis microscope SIo-x S TA N FORD UNIVE RSITY
Fibers Rotatable wedges for beam alignment
3.2 mm mm MEMS MEMS Chip 3 Xx 32.9mm chip
2 mm Dia. Fused Silica SIL Collimating lenses
Index-matching tissue interface 10 mm Dia. Parabolic mirror F = 4.6 mm, 0.7 mm thickness
Mandella
10
Fig. 2. Shows the primary optical components of the system and the MEMS scanner. Four rotating wedges (Risley prisms) are used for insuring parallel alignment between the two collimated beams. Once this alignment step is accomplished, then the rest of the optical system is self-aligning (i.e., the parabolic focusing mirror and the MEMS mirror can be removed and replaced without requiring any realignment steps).
Proc. of SPIE Vol. 6443 64430E-3
Wire-EDM precision V-grooves
Collimator & wedge holder
Fig. 3. Microscope housing containing precision machined v-grooves for mounting of the collimating lenses and cylindrical pockets that hold the rotating wedges. The rotating wedges provide robust alignment to insure that the two collimated beams are parallel to within 0.005 degrees. After alignment, the rotating wedges are secured by clamps that prevent further rotation.
The present housing supports the 10 mm diameter focusing mirror and is small enough for handheld in vivo testing for checking the performance of the components before they are mounted into a 5mm diameter future package assembly. The MEMS and optical components have been designed to meet the conditions imposed by both, imaging performance, and packaging requirements for endoscopic applications. In this vein, the dimensions of the present GRIN fiber collimators and the MEMS chip have been chosen to allow the same components to fit inside a future 5 mm diameter housing for use in the instrument channel of an upper-GI endoscope. All the dimensions of the present 10 mm parabolic mirror, except for its outside diameter, would be conserved in this future 5 mm package. Fig. 5 demonstrates how the 5 mm diameter parabolic mirror is cut from the present 10 mm design, and also shows how the present optical and MEMS components are designed to fit into both, the 10 mm, and the 5 mm form factor packages. The purpose of constructing the 10 mm package first is to provide flexible packaging that is assembled using miniature screws, thus allowing different components to easily be installed and removed for testing. This is important for changing collimator lenses and fibers for operation at different wavelengths. Figures 6 and 7 show the primary geometrical parameters of the optical design that affect the resolution, FOV, and working distance. These design parameters were simulated using ASAPTM optical computer simulation software to determine the correct parameter choices to meet the design goals that were set for endoscopic imaging applications. Figure 8 shows a view of the ASAPTM model used for these simulations. The model consists of the following specific optical components: 1) two single-mode fibers, 2) two collimator lenses, 3) a parabolic mirror, 4) a scanning mirror, 5) a hemispherical SIL, and 6) the tissue being imaged. Since the primary application for this microscope is for in vivo tissue imaging of the esophagus, an index of refraction of 1.37 was used for the tissue in the ASAPTM model. The ASAPTM model also accounts for diffraction effects and the propagation of Gaussian beams from the single-mode illumination fiber, as well as coupling into the single-mode collection fiber.
Proc. of SPIE Vol. 6443 64430E-4
Slider mechanism for Z-scan
MEMS PCB
0 0
Fig. 4. The complete assembly including the slider mechanism for providing Z-scanning capability of MEMS mirror.
c 5 mm dia. circle
// //
Fig. 5. The parabolic focusing mirror can be modified for future reduction of the outside diameter of total package to 5 mm.
Proc. of SPIE Vol. 6443 64430E-5
3. SIMULATIONS Computer modeling was done in three phases to simulate the performance in three categories: resolution, FOV, and Zscan capability which produces a variable working distance. The simulation results shown in figure 9 show the axial confocal response to a moving mirror immersed in the tissue at a wavelength of 780 nm. The simulated FWHM axial resolution is 5.96 microns, which agrees with earlier diffraction theory calculations.7 The axial resolution of the assembled system can be measured experimentally from the FWHM response to the reflection off the SIL’s flat glass-air interface as the focus is “Z-scanned”. The transverse resolution can be measured experimentally by direct imaging of group 7 of a Chromium Air Force test chart placed in contact with the SIL’s flat surface. Additional simulation results have determined the FOV of the fast axis (axis Y in figure 1) of the scanning mirror to be 748 microns for mirror tilt angles +/- 6 degrees. The FOV simulations were done by comparing the resolution at different scanning mirror tilt angles to verify that the scanning of the beams does not produce substantial aberrations causing more than a 20% decrease in resolution. The FOV results for the slow axis (axis X in figure 1) were found to be similar to the results for the fast axis. Simulations of the Z-scanning feature of the microscope were used to explore the affect of different SIL designs on the total range of the variable working distance in response to the Z-displacement of the scanning mirror. The simulations begin at the scanning mirror’s neutral position (Z = 0), which is defined at an imaging working distance of 250 microns from the surface of the tissue. At this depth, the scanning mirror’s fast and slow axes scan an approximately horizontal image plane located 250 microns below the tissue surface. When the Z-position of the scan mirror is moved upward (away from the tissue) a distance of 145 microns from the neutral position, simulations show that the focus point correspondingly moves a distance of 250 microns up to the surface of the tissue. The simulations have also shown that when the scan mirror is displaced a distance of 170 microns in the other direction (towards the tissue) from the neutral position, the focus correspondingly moves to a depth 500 microns below the tissue surface. Thus, a total depth range (variable working distance) of 500 microns is produced from a total mirror displacement in the Z-direction of 315 microns, which is a ratio given by 1.6:1. Note, that in free space (n=1), this ratio would be expected to be exactly 2:1. A specific SIL design that gives the performance described above is a hyperhemisphere truncated at -250 microns short of a being a full hemisphere with an index of n = 1.37 and it’s radius of curvature is r = 3 mm, This ratio can vary between 1:1 and 2:1 depending on the specific SIL design. For example, using a fused silica full hemisphere (n = 1.46) having a radius of curvature r = 1.5 mm, this ratio becomes 1.2:1.
'4.3 deg
5deg
Fig. 6. The cone angle α = 5o and inclination angle θ = 24.3o of the beams determine the transverse and axial resolution of the system.
Proc. of SPIE Vol. 6443 64430E-6
dφ = +/- 6ο
z-scan
mm 2.37 2.5 mm
Adjustable working distance produced by the MEMS z-scan Fig. 7. Distance from the scan mirror to the focus and the mechanical angle dφ of the scan mirror are both parameters that affect the final FOV of the system. The MEMS mirror is mounted to slider for z-scan to change the working distance, and thus allowing image depth selection (variable working distance) from 0 to 500 microns.
ASAP Model (5mm probe rev_1.6)
Fig. 8. ASAPTM computer model for simulations.
Proc. of SPIE Vol. 6443 64430E-7
Axial Response Function 1 0.9 0.8 Coupling Efficiency [%]
FWHM =5.955 um 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 -30
-20
-10
0 z [um]
10
20
30
TM
Fig. 9. Graph showing the axial resolution of the MEMS-based dual-axes confocal microscope calculated from ASAP model. The confocal response is simulated using a flat horizontal test-mirror imbedded in the tissue, which is movable within the tissue along the z direction (horizontal axis of graph) to different depths. The coupling efficiency of the single-mode collection fiber at 780 nm is shown as a function of the z displacement of the test-mirror from -30 to +30 microns about a z = 0 position defined at 250 microns deep in the tissue. The calculations are done with the MEMS mirror in a fixed position thereby keeping the focus fixed at 250 microns deep in the tissue.
4. RESULTS Three prototype systems based on the design described in this paper have been assembled and are being tested. We have installed into the systems three different sets of collimators and fibers that are designed for operation at three different wavelengths. Prototype I produces reflectance images at 1300 nm.14 Prototype II produces reflectance and fluorescence images at 780 nm.8 Prototype III produces reflectance and fluorescence images at 488 nm. The prototype systems have met the initial design goals and packaging requirements set for handheld in vivo imaging applications at these different wavelengths. The systems typically operate at 4 frames per second, with a FOV of 800 × 500 microns. The working distance is variable from 0 to 500 microns by controlling the piezo-actuator.
5. CONCLUSIONS AND FUTURE DIRECTIONS We have shown that a simple and robust design of a MEMS-based handheld instrument can meet the imaging performance requirements for an in vivo endoscopic microscope, but the larger form factor packaging of the present handheld prototypes are still too large for endoscopic compatibility. The same MEMS and optical components that are installed in these prototypes have also been designed to fit inside a future 5 mm diameter package for use in the instrument channel of a therapeutic upper-GI endoscope. The only component that needs to be modified to fit in this next smaller package is the parabolic focusing mirror, and so the images produced by the smaller package are expected to have identical resolution, FOV, and working distance as the images being produced by the present prototypes. Therefore, the future packaging of a dual-axes microscope using these same components will meet both, the performance goals, and the packaging requirements for in vivo imaging of upper-GI diseases.
Proc. of SPIE Vol. 6443 64430E-8
6. ACKNOWLEDGMENTS We would like to acknowledge funding support from the National Cancer Institute (NCI), through the U54 CA105296 Network for Translational Research in Optical Imaging (NTROI). Jonathan Liu is supported by a Canary Foundation / American Cancer Society post-doctoral fellowship for the early detection of cancer. Thomas Wang is supported by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) on grant number K08 DK067618. This work was also funded in part by grants from the National Institutes of Health R33 CA109988. This work has also been supported by funding through the Center for Biophotonics, a National Science Foundation Center managed by the University of California, Davis (PHY 0120999). We thank Ning Chan for optical simulation support. We thank Shai Friedland, Roy Soetikno, and Peyman Sahbaie for clinical support. We thank Wenchuan Liang, and Wah-Ping Luk for technical support.
REFERENCES 1. D. L. Dickensheets, G. S. Kino, “A micromachined scanning confocal microscope” Opt. Lett. 21, 764-6 (1996). 2. E. H. Stelzer, S. Lindek, S. Albrecht, R. Pick, G. Ritter, N. Salmon, and R. Stricker, “A new tool for the observation of embryos and other large specimens: confocal theta fluorescence microscopy” J. Microscopy 179, 1-10 (1995). 3. R. H. Webb and F. Rogomentich, “Confocal microscope with large field and working distance” Appl. Opt. 38, 4870-4875 (1999). 4. T. D. Wang, M. J. Mandella, C. H. Contag, and G. S. Kino, “Dual-axis confocal microscope for high-resolution in vivo imaging” Opt. Lett. 28, 414-16 (2003). 5. T. D. Wang, C. H. Contag, M. J. Mandella, N. Y. Chan, and G. S. Kino, “Dual axes confocal microscope with postobjective scanning and low coherence heterodyne detection” Opt. Lett. 28, 1915-1917 (2003). 6. T. D. Wang, C. H. Contag, M. J. Mandella, N. Y. Chan, and G. S. Kino, “Confocal fluorescence microscope with dual-axis architecture and biaxial post objective scanning” J. Biomed. Opt. 9(4), 735-742 (2004). 7. J. T. C. Liu, M. J. Mandella, S. Friedland, R. Soetikno, J. H. Crawford, C. H Contag, G. S. Kino, and T. D. Wang, “Dual-axes confocal reflectance microscope for distinguishing colonic neoplasia” J. Biomed. Opt. 11, 054019 (2006). 8. J. T. C. Liu, M. J. Mandella, H. Ra, L. K. Wong, O. Solgaard, G. S. Kino, W. Piyawattanametha, C. H. Contag, and T. D. Wang, “A miniature near-infrared dual-axes confocal microscope utilizing a two-dimensional MEMS scanner” Opt. Lett. 32 (in press). 9. L. K. Wong, M. J. Mandella, G. S. Kino, T. D. Wang, “Improved rejection of multiply scattered photons in confocal microscopy using dual-axes architecture” Opt. Lett. in preparation (2007). 10. D. Lee, U. Krishnamoorthy, K. Yu, O. Solgaard, "Single-crystalline silicon micromirrors actuated by self-aligned vertical electrostatic combdrives with piston-motion and rotation capability", Sensors and Actuators: A Physical, Vol. 114, issue 2-3, 1 September 2004, pp. 423-428. 11. D. Lee, O. Solgaard "Two-Axis Gimbaled Microscanner in Double SOI Layers Actuated by Self-Aligned Vertical Electrostatic Combdrive", Proceedings of the Solid-State Sensor and Actuator Workshop, pp. 352-355, Hilton Head, South Carolina, June 6-10, 2004. 12. H. Ra, W. Piyawattanametha, Y. Taguchi, O. Solgaard, "Dual-Axes Confocal Fluorescence Microscopy with a Two Dimensional MEMS Scanner", 2006 IEEE/LEOS International Conference on Optical MEMS and Their Applications, Big Sky, Montana, 21-24 August 2006, pp. 166-167. 13. H. Ra, Y. Taguchi, D. Lee, W. Piyawattanametha, and O. Solgaard, “Two-Dimensional MEMS scanner for DualAxes Confocal In Vivo Microscopy,” in MEMS 2006, IEEE Int. Conf. on Micro Electro Mechanical Systems, Turkey, 2006, pp. 862-865. 14. W. Piyawattanametha, J.T.C. Liu, M.J. Mandella, H. Ra, L.K. Wong, P. Hsiung, T.D. Wang, G.S. Kino, O. Solgaard, "MEMS Based Dual-axes Confocal Reflectance Handheld Microscope for in vivo Imaging", 2006 IEEE/LEOS International Conference on Optical MEMS and Their Applications, Big Sky, Montana, 21-24 August 2006, pp. 164-165.
Proc. of SPIE Vol. 6443 64430E-9
