668

JOURNAL OF MICROELECTROMECHANICAL SYSTEMS, VOL. 16, NO. 3, JUNE 2007

Stroboscopic Imaging Interferometer for MEMS Performance Measurement

Josh A. Conway, Member, IEEE, Jon V. Osborn, Member, IEEE, and Jesse David Fowler
AbstractThe insertion of microelectromechanical systems (MEMS) components into aerospace systems requires advanced testing to characterize performance in a space environment. Here, we report a novel stroboscopic interferometer test system that measures nanometer-scale displacements of moving MEMS devices. By combining video imagery and phase-shift interferometry with an environmental chamber, rapid visualization of the dynamic device motion under the actual operational conditions can be achieved. The utility of this system is further enhanced by integrating the interferometer onto the chamber window, allowing for robust interferometric testing in a noisy environment without requiring a oating optical table. To demonstrate these unique capabilities, we present the time-resolved images of an electrostatically actuated MEMS cantilevered beam showing the rst-order to sixth-order plate modes under vacuum. [2006-0264] Index TermsInterferometry, microelectromechanical devices, microelectromechanical systems (MEMS) metrology, stroboscopic interferometer system, vacuum systems.

I. I NTRODUCTION HE HARSH environments and stringent reliability requirements of aerospace systems demand detailed knowledge of the motion of all mechanical and electromechanical devices [1]. The functional performance must be well understood, and the failure modes must be catalogued. While this may entail a straightforward analysis for macroscopic devices, microscopic electromechanical structures require new tools to study their response under various conditions. In the existing art, static, quasi-static, and vibrometer-based methods [2], [3] are employed, and the motion of microelectromechanical systems (MEMS) is then analytically reconstructed. These inferential results can be unacceptable for the requirements of modern aerospace applications. For instance, quasi-static measurements are often lacking in precision when related to dynamic operation, particularly near a mechanical resonance. Vibrometer-based measurements, on the other hand, provide tremendous precision for dynamic measurement but are not able to yield a complete picture of the device motion. While

Manuscript received November 22, 2006; revised February 21, 2007. This work was supported by The Aerospace Corporation under the Mission Oriented Investigation and Experimentation program, which was funded by the U.S. Air Force Space and Missile Systems Center under Contract FA8802-04-C-0001. Subject Editor O. Solgaard. J. A. Conway and J. V. Osborn are with The Aerospace Corporation, Los Angeles, CA 90009 USA (e-mail: Josh.A.Conway@Aero.org; Jon.V. Osborn@Aero.org). J. D. Fowler is with the Department of Mechanical and Aerospace Engineering, University of California, Los Angeles, CA 90095 USA (e-mail: jessefowler@hotmail.com). Digital Object Identier 10.1109/JMEMS.2007.896710

vibrometry is able to map amplitude and phase of vibration across a surface, it cannot yield static topographic information. Direct measurement of moving parts makes new demands on the test system. To study the motion of microscopic mechanical parts, one cannot simply attach sensors, and noncontact measurement techniques are required. The environments in which MEMS operate further complicate the measurement process. The characterization of MEMS devices for terrestrial applications typically must be performed at partial pressures of gas and over a range of temperatures to replicate packaged environments and to optimize their performance. Emerging applications of MEMS in space systems, however, require functionality under extreme environmental conditions. Such applications include spatial light modulation on the cryogenic focal plane of imaging systems [4], switching and phase modulation for massively parallel phased arrays [5], and thermal control directly on the skin of surfaces [6]. In these situations, packaging may be minimal, temperature ranges may be extreme, and pressure may be as low as the vacuum of space. The testing conditions must therefore match the operational environment for MEMS technology to establish itself in this market. In response to these challenges, various test systems have been reported in the literature. There are those specically designed for space applications, which recreate the temperatures and pressures of the space environment but employ probe or scanning techniques that do not give a complete picture of the device dynamics or topology [1], [7]. In contrast, optical stroboscopic MEMS test systems have been reported [8][12], which characterize devices in motion. These yield a complete picture of the dynamic motion and surface structure of the device down to the nanometer scale, but those reported in the literature only operate under ambient conditions. Neither of these classes of test systems can make dynamic measurements while simulating the pressures that are encountered in space and the upper atmosphere. This is a critical point because MEMS dynamic response changes markedly when the damping of atmospheric pressure is removed, particularly near mechanical resonant frequencies [13][17]. Addressing these issues, our solution is based on a fusion of these designs. By integrating an environmental test chamber with a stroboscopic imaging interferometer, we have created a system capable of generating a complete picture of the device dynamics with variable environmental control. This combination of techniques, however, goes further than simply adding new environmental controls to standard stroboscopic phaseshift interferometry. In fact, the system becomes much more than the sum of its parts. Our results show that the integration of the interferometer directly onto the environmental chamber

1057-7157/$25.00 2007 IEEE

CONWAY et al.: STROBOSCOPIC IMAGING INTERFEROMETER FOR MEMS PERFORMANCE MEASUREMENT

669

Fig. 1.

Optical layout of the MEMS stroboscopic interferometer.

greatly reduces the measurement errors that are incurred from external vibrations, which has limited the utility of many interferometric measurement systems. By keeping all of the interferometer optics rigidly attached to the faceplate of the environmental chamber, deep-subwavelength out-of-plane resolution has been achieved in a very noisy laboratory without the use of a oating optical table. This single advance has the capability of bringing interferometric MEMS characterization out of the controlled laboratory environment and onto the factory oor. II. O PTICAL D ESIGN Our test system is based on the stroboscopic Michelson interferometer. It achieves nanometer-scale out-of-plane resolution interferometrically and diffraction-limited lateral resolution (1 m) using microscope objectives. Although it relies on sampling measurements during cyclical motion of the device, only ve intensity images are used to reconstruct each frozen surface per time step of motion. At standard video refresh rates of 15 Hz, this makes for rapid frame acquisition, which is robust in the face of environmental noise. Because a video camera is employed as a detector array, no lateral scanning is required. This popular optical design has been reported in the literature [18], and our instrument design evolved out the work of Muller [8][10]. In our attempt to recreate Mullers stroboscopic measurement system, we found that those designs proved too susceptible to vibrations to yield reproducible results in our noisy laboratory environment. Even with 1.5-in optical post assemblies, a oating optical table, and heavy laser isolation curtains, noise continued to plague this system. Resolution of this problem led to several important improvements, which are to be discussed later in the text. Although we have machined hardware and developed software to perform the data acquisition and processing operations, the design is simple enough to be replicated in other laboratories, unlike white-light and more complicated interferometers [19]. The layout for the test system is illustrated in Fig. 1. The light source is a ber-coupled diode laser operating at 635 nm (Melles Griot 57ICS062/SP/HS). This laser is directly modulated such that the beam is pulsed on and off with a programmable delay in synchrony with the MEMS driving signal. In doing so, this stroboscopic source is able to freeze the device

at any phase of its high-speed motion, even when employing slow cameras and detectors. To keep the image from blurring with the MEMS motion, the duty cycle (dened here as the optical pulse duration/period of the MEMS driving signal) was kept below 0.01. This low duty cycle, however, has the effect of greatly reducing the optical power falling on the camera (Pulnix TM-1020). Because the out-of-plane resolution depends strongly on the number of digitized bits in the detected signal [20], it is necessary to cover the 8-bit dynamic range of the camera. To correct for this without introducing expensive high-power lasers, focusing optics were introduced into the optical path to concentrate the beam onto the region of interest on the device. It is important to note that microscope objectives should not be used for this task. High-numerical-aperture optics make calibration very difcult due to their rapidly changing phase front near the focus. This is because the radius of curvature of the phase must change from a very small quantity near the focus, to innity at the actual focal spot [21]. If the reference mirror and test device are not equidistant from their objective lenses, the phase fronts will have different curvatures. After unwrapping, this curvature translates into warped surface reconstructions even if the sample and reference mirrors are both perfectly planar. Although the peak output power of our laser source is 7 mW, we have found that a low-magnication lens system can easily correct phase-curvature problems while concentrating enough light to saturate the charge-coupled device (CCD) camera. After the focusing optics, the beam encounters a nonpolarizing beam splitter. This sends half of the light into the environmental chamber and onto the MEMS device under test. The remaining half of the optical beam is sent to a reference mirror on a lead zirconate titanate transducer, forming the second arm of a Michelson interferometer. These beams then recombine and interfere, allowing one to determine out-ofplane displacement. The nature of the measurement can allow for measurement sensitivity down to the nanometer scale [19]. However, as seen from a different perspective, this sensitivity makes the measurement prone to error from noise and vibration. Nanoscale displacement or vibration in any part of the optical path in either arm is translated into measurement error. The reader is reminded that the phase-unwrapping process can greatly amplify these errors by incorrectly adding or subtracting multiples of /2. It is precisely for this reason that external noise has plagued optical interferometric measurements and why most systems only operate in highly isolated laboratory environments. To correct for this, we have integrated the interferometer onto the faceplate of the environmental chamber. The beam splitter is xed with optical epoxy to the chamber window, and the reference mirror is bolted directly to the common chamber faceplate. Because the interferometer measures relative displacement between the two arms, the measurement is exceedingly sensitive to noise in this region. To correct for this, the optical path of each arm was kept shorter than 2 cm. These simple design changes had a profound effect on the performance of the interferometer. Although the majority of the data reported in this paper was taken on a oating optical table, it was later found that the performance was not signicantly

670

JOURNAL OF MICROELECTROMECHANICAL SYSTEMS, VOL. 16, NO. 3, JUNE 2007

absolute phase, which can be calculated directly from the wrapped phase [25] using x (x, y) = sin (x, y) [cos (x + dx, y) cos (x, y)] + cos(x, y) [sin (x + dx, y) sin (x, y)] . (2)

The gradient in the y-direction is determined in a similar manner by permuting the x and y dimensions in (2). Combining the two equations yields an initial value for the absolute phase. From (1), it can be seen that the dc term of the Fourier expansion must be dropped (qx=0 = qy=0 = 0) because of the resultant null in the denominator. To resolve this ambiguity, the wrapped phase is then subtracted from this unwrapped phase, and the resultant difference is rounded to the nearest 2. The wrapped phase is added back, and the absolute phase is obtained. III. R ESULTS To demonstrate the capabilities of this MEMS stroboscopic interferometer, we have examined the motion of a MEMS cantilever beam test structure. The sample device is composed of two electrical bond pads that are connected to two-doped 150 150 m overlapping polysilicon layers, with an isolating 2-m air gap, supported by a 134 30 m polysilicon beam, as also shown in Fig. 2. The holes in the paddle, which are designed to aid in MEMS oxide release, are also evident. Under electrical bias, electrostatic deection downward of the upper 2-m-thick cantilevered polysilicon region of interest occurs. A sinusoidal electric drive signal results in a cyclical device motion proportional to drive signal amplitude and in-phase with the driving waveform. This device was chosen for analysis because exible polysilicon beam structures of this type are very common in MEMS radio frequency switches, rate sensors, accelerometers, and many other MEMS devices. For a given frequency, drive voltage, and chamber pressure, the MEMS stroboscopic interferometer can generate a surface reconstruction of the cantilever at various frozen phases of its motion. This is illustrated in Figs. 3 and 4, which show the rst and second sets of three resonant modes of the device at 10 mtorr operating pressure. These resonant modes were stimulated with sinusoidal drive signals at 74 640, 217 150, 475 450, 880 100, 1 313 430, and 1 452 230 Hz, respectively. This is the rst time that plate modes on this scale have been directly imaged [26], as known to the authors. Because the quality factor (Q-factor) varies between modes, the amplitude of the drive signal was changed to give a clear picture of the device motion. The deformation of the cantilever can be clearly seen in all sets of surface reconstructions. Also evident from the surface reconstructions is the ridge across the center of the device. This artifact of fabrication is approximately 10 nm high and is very clear from this measurement. To verify that we have correctly identied each of the rst six resonant modes of the cantilever test structure, we have performed nite-element modeling (FEM) of the complete 3-D device structure. This modeling was performed using FEMLab Multiphysics software from COMSOL, Inc. [27]. This mechanical 3-D model is constructed from the original mask les of

Fig. 2. MEMS cantilever device (a) imaged by a scanning electron microscope (SEM) and (b) imaged directly in interferometer system showing a fringe pattern. (c) Graphical rendering of the cantilever beam illustrating the prole.

degraded when the table was not oating. In addition to this, the system was able to take data with nanometer-scale out-of plane resolution in an exceptionally noisy environment and while the chamber was rigidly attached to an operating roughing pump and turbo pump. Unlike the holographic techniques or the use of specialized interferometers [11], our design achieves this high stability without the use of custom components. The power of this architecture is that it can be assembled from common laboratory equipment. After the beams recombine at the beam splitter, a long working distance objective is used to image the device onto a CCD camera. This system takes an entire image at a time; thus, no lateral scanning is required. The captured image is covered with fringes, as shown in Fig. 2(b). To obtain the actual phase data from the fringe pattern, we used phase-shift interferometry [18]. This entails translating the reference mirror longitudinally by a xed displacement between each frame. Of the various recipes for this, we employed Hariharans algorithm [22], which uses intensity data at ve known positions of the reference mirror to generate phase data. This yields the relative phase modulo 2. Converting this wrapped phase to absolute phase in two dimensions is still considered an unsolved problem in the eld [23]. To unwrap the phase over its full range, we used the method of Volkov and Zhu [24]. The absolute phase (denoted by ) is calculated using = Re 1 1 F (x )qx + F (y )qy F 2 2 2i qx + qy (1)

where F and F 1 represent the Fourier and inverse Fourier transforms, respectively, and q represents the wave vector in a given direction. Equation (1) requires the gradient of the

CONWAY et al.: STROBOSCOPIC IMAGING INTERFEROMETER FOR MEMS PERFORMANCE MEASUREMENT

671

Fig. 3.

Several phases of the MEMS cantilever for the rst three resonant modes. All units are in micrometers.

the cantilever device. We have used the COMSOL materials property database library values for polysilicon. When running the model, a Youngs modulus of 130 GPa was chosen such that the device resonance at the second eigenfrequency most closely matched that of the stroboscopic measurement. This is in agreement with empirically derived values of polysilicon, which have been measured between 132 and 174 GPa [28]. All other material properties were set to their default values. The COMSOL eigenfrequency 3-D Solid, StressStrain solver was selected, with boundary conditions set such that all stationary surfaces were xed. The rst six eigenmodes are shown in Fig. 5. These modeled results are in good agreement with both our measured frequencies and mode shapes. Table I shows the values of measured and modeled results, all with errors less than 5%. The strength of the resonant mechanical modes is also a critical parameter for these devices. To quantify this, we measured the mechanical response to a sinusoidal drive voltage around the rst resonance frequency. Cantilever devices of this type have a nonlinear quasi-static response to the driving signal and pull-in toward the lower electrode at a given threshold voltage [29]. To minimize this effect, the amplitude of the drive voltage was changed at each pressure to keep the peak displacement below 2 m at resonance. The motion of a small region near the free edge of the cantilever is illustrated in Fig. 6(a). Due to de-

vice destruction during temperature cycling, a different MEMS device was used for these tests. This device is identical to that of Fig. 2, except that the single 134 30 m polysilicon beam is replaced by a pair of 64 30 m beams, as shown in the inset of Fig. 6(b). The mechanical response was then recorded about the rst resonance peak, and the Q-factor was computed from the central frequency divided by the spectral full-width at half-maximum of the mechanical displacement amplitude. The Q-factor for the fundamental mode at several pressures is plotted in Fig. 6(b). This plot not only illustrates the transition to intrinsic damping near 10 mtorr but also underscores the critical importance of the environmental chamber. At atmospheric pressures, this mode is heavily overdamped and would be very difcult, if not impossible, to study quantitatively. Thus, we see that without the combination of vacuum pressures, nanoscale sensitivity, and high-speed measurement, the complete motion of the 3-D modes of Figs. 3 and 4 could not be studied. The authors know of no reported measurement system employing a single measurement that is capable of this metrology. There are reports of vibrometer test systems with vacuum chambers [30], but these systems cannot image surface topology directly [9] and place a great deal of strain on the scanning mechanism. It is nontrivial to translate the focusing optics or the MEMS device inside a vacuum chamber, although a confocal solution has been implemented. Digital holographic solutions

672

JOURNAL OF MICROELECTROMECHANICAL SYSTEMS, VOL. 16, NO. 3, JUNE 2007

Fig. 4. Several phases of the MEMS cantilever for the second three resonant modes. All units are in micrometers.

such as the Lynce Tec DHM R1000 can provide a tool that is able to measure mechanical transients, which is in contrast to the repeated motion that is required here. Unfortunately, this system is limited to drive frequencies below 100 kHz and, therefore, cannot reproduce the modes in Figs. 3 and 4. Whitelight interferometry systems such as the Wyko DMEMS, which can strobe up to 1 MHz [31], suffer from similar high-frequency limitations. This is not to preclude the possibility of altering the test equipment to achieve higher frequency operation [32], employing a system that uses a combination of white-light interferometry and vibrometry such as the Polytec MSA-400 or a system of white-light and phase-shift interferometries such as Micro Photonics Zoomsurf 3D. We stress that although we know of no reported system that has all of the capabilities that are required to measure these plate modes, one could be built from competing technologies. The novelty and power of this system are in its simplicity, ease of operation, accessibility to the researcher, and noise immunity implicit in the design. Unlike packaged measurement systems, changing out the parts to add new capabilities is simple. For instance, the maximum device frequency that can be measured is determined by the laser modulation bandwidth. Telecom lasers with modulated

bandwidths greater than 1 GHz are now widely available and can be added to this system. To quantify the vertical (interferometric) resolution in the presence of noise in real system operation, we undertook a statistical analysis. For this paper, we acquired 200 phaseunwrapped surfaces of the static cantilever device over the course of 10 min. This timescale is longer than the typical measurement but was needed to achieve a large statistical sample. It was seen that the device, as a whole, drifted on the order of 10 nm during the experiment. To compensate for this, the motion of the substrate was calculated and then subtracted from that of the cantilever. After removing the drift, the standard deviation of out-of-plane position was calculated at 4224 individual points on the cantilever from these 200 surface reconstructions. The median standard deviation was calculated to be 2 nm, which is a surprisingly small result in a very noisy laboratory. Most of this noise can be attributed to the camera itself, as we used an uncooled CCD that demonstrated a rootmean-square (rms) dark noise of approximately 11.2 out of the full-scale 255 (8 bits) when there was no illumination. Monte Carlo simulations using a Gaussian t to this noise showed that the camera added approximately 1.5 nm of out-of-plane

CONWAY et al.: STROBOSCOPIC IMAGING INTERFEROMETER FOR MEMS PERFORMANCE MEASUREMENT

673

IV. C ONCLUSION We have demonstrated the design and functionality of a novel stroboscopic imaging interferometer for MEMS metrology. The nanoscale precision and high-frequency capabilities make this an ideal MEMS characterization tool for devices that can be electrically actuated. By making the environmental chamber an integral part of the measurement system, we have ruggedized our optical interferometric instrument. The interferometer can function in noisy environments, greatly extending the utility of this high-precision technique. The combined attributes of our system enable the direct observation and measurement of large-area ( 500 500 m) complex 3-D motions and high-order 3-D plate modes of MEMS devices dynamically operating frequencies in excess of 1.4 MHz. Many MEMS applications such as spatial light modulators, inertial rate sensors, and accelerometers demonstrate performance that lack in respect to their macroscopic counterparts. Measurement capability such as ours can provide critical mechanical MEMS metrology well in excess of rstorder, second-order, or third-order modal analysis. These measurement data directly provide the MEMS designer observable measurements of deformations due to high-order modes and allow proper error attribution to their sources. Our novel tool was used to image many of the resonant modes of a cantilevered beam structure. The mechanical 3-D bending modes that were observed in this MEMS device, while well known in structural mechanics of large objects, are very difcult to directly measure for micrometer-scale MEMS devices. Direct observation, measurement, and surface reconstruction of MEMS devices while under full dynamic operation provides a unique metrology capability for understanding and observing the complex behavior of micromechanical structures. R EFERENCES
[1] D. J. McComas, G. P. Miller, J. N. Mitchell, S. E. Pope, and P. W. Valek, Space applications of microelectromechanical systems: Southwest Research Institute vacuum microprobe facility and initial vacuum test results, Rev. Sci. Instrum., vol. 74, no. 8, pp. 38743878, Aug. 2003. [2] J. S. Burdess, A. J. Harris, D. Wood, R. J. Pitcher, and D. Glennie, A system for the dynamic characterization of microstructures, J. Microelectromech. Syst., vol. 6, no. 4, pp. 322328, Dec. 1997. [3] J. La, J. Choi, S. Wang, K. Kim, and K. Park, Continuous scanning laser Doppler vibrometer for mode shape analysis, Opt. Eng., vol. 42, no. 3, pp. 730737, Mar. 2003. [4] H. C. Ford, M. Clampin, G. D. Illingworth, J. E. Krist, S. S. Oliver, L. Petro, and G. E. Sommagren, Requirements for an optical 8-m space telescope with a MEMs deformable mirror to detect Earth-like planets around nearby stars, Proc. SPIE, vol. 4854, pp. 554557, 2003. [5] K. J. Vinoy and V. K. Varadan, Design of recongurable fractal antennas and RF-MEMS for space-based systems, Smart Mater. Struc., vol. 10, no. 6, pp. 12111223, Dec. 2001. [6] R. Osiander, J. L. Champion, A. M. Darrin, J. J. Sniegowski, S. M. Rodgers, D. Douglas, and T. D. Swanson, Micromachined louver arrays for spacecraft thermal control radiators, presented at the 39th Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan. 811, 2001, Paper AIAA-2001-215. [7] B. Legrand, E. Quevy, B. Stefanelli, D. Collard, and L. Buchaillot, Vacuum and cryogenic station for micro-electro-mechanical systems probing and testing, Rev. Sci. Instrum., vol. 73, no. 12, pp. 43934395, Dec. 2002. [8] M. R. Hart, R. A. Conant, K. Y. Lau, and R. S. Muller, Stroboscopic interferometer system for dynamic MEMS characterization, J. Microelectromech. Syst., vol. 9, no. 4, pp. 409418, Dec. 2000.

Fig. 5. COMSOL 3-D FEM results of the MEMS cantilevered test structure. Eigenfrequencies and mode shapes of the rst six resonant modes are shown. TABLE I EIGENFREQUENCIES FOR MEASURED AND MODELED MODES USING COMSOL MULTIPHYSICS 3-D FEM, SOLID, STRESSSTRAIN SOLVER

Fig. 6. (a) Displacement of a single point on the cantilever at the fundamental resonance (70 040 Hz). (b) Q-factor of the fundamental mode over pressure. Inset shows an SEM image of the cantilever under test.

uncertainty to our measurement. This large noise source can be rectied by either upgrading to a scientic grade lownoise camera or by averaging several data sets. The reader is also reminded, however, that these pixels represent an area of approximately 1 m2 . For any real material, there will be a height variation across the pixel of much larger than 2 nm. For instance, 2 nm across the pixel represents a perfectly planar surface that is tilted by 0.1146 from the normal or a perfectly oriented surface with 2 nm of rms roughness.

674

JOURNAL OF MICROELECTROMECHANICAL SYSTEMS, VOL. 16, NO. 3, JUNE 2007

[9] C. Rembe, R. Kant, and R. S. Muller, Optical measurement methods to study dynamic behavior in MEMS, Proc. SPIE, vol. 4400, pp. 127137, 2001. [10] C. Rembe and R. S. Muller, Measurement system for full threedimensional motion characterization of MEMS, J. Microelectromech. Syst., vol. 11, no. 5, pp. 479488, Oct. 2002. [11] P. Hariharan, Optical Interferometry. San Diego, CA: Academic, 2003. [12] S. Petitgrand and A. Bosseboeuf, Simultaneous mapping of out-ofplane and in-plane vibrations of MEMS with (sub)nanometer resolution, J. Micromech. Microeng., vol. 14, no. 9, pp. S97S101, Sep. 2004. [13] V. Mortet, R. Petersen, K. Haenen, and M. DOlieslaeger, Wide range pressure sensor based on a piezoelectric bimorph microcantilever, Appl. Phys. Lett., vol. 88, no. 13, p. 133 511, 2006. [14] R. A. Buser and N. F. De Rooij, Very high Q-factor resonators in monocrystalline silicon, Sens. Actuators A, Phys., vol. 21, no. 13, pp. 323327, Feb. 1990. [15] J. Mertens, E. Finot, T. Thundat, A. Fabre, M. Nadal, V. Eyraud, and E. Bourillot, Effects of temperature and pressure on microcantilever resonance response, Ultramicroscopy, vol. 97, no. 1, pp. 119126, 2003. [16] F. R. Blom, S. Bouwstra, M. Elwenspoek, and J. H. J. Fluitman, Dependence of the quality factor of micromachined silicon beam resonators on pressure and geometry, J. Vac. Sci. Technol. B, Microelectron., vol. 10, no. 1, pp. 1926, Jan. 1992. [17] K. Y. Yasumura, T. D. Stowe, E. M. Chow, T. Pfafman, T. W. Kenny, B. C. Stipe, and D. Rugar, Quality factors in micron- and submicronthick cantilevers, J. Microelectromech. Syst., vol. 9, no. 1, pp. 117125, Mar. 2000. [18] W. Hemmert, M. S. Mermelstein, and D. M. Freeman, Nanometer resolution of three-dimensional motions using video interference microscopy, in Proc. 12th IEEE Int. Conf. Micro Electro Mech. Syst., Orlando, FL, 1999, p. 302. [19] E. Novak, D. Wan, P. Unruh, and J. Schmit, Dynamic MEMS measurement using a strobed interferometric system with combined coherence sensing and phase information, in Proc. Int. Conf. MEMS, NANO and Smart Syst., Jul. 2023, 2003, pp. 285288. [20] K. Creath, Phase measurement interferometry techniques, in Progress in Optics, vol. XXVI, E. Wolf, Ed. Amsterdam, The Netherlands: North Holland, 1988, pp. 351391. [21] J. T. Verdeyen, Laser Electronics. Englewood Cliffs, NJ: Prentice-Hall, 1995, pp. 7073. [22] P. Hariharan, B. F. Oreb, and T. Eiju, Digital phase-shifting interferometry: A simple error-compensating phase calculation algorithm, Appl. Opt., vol. 26, no. 13, p. 2504, Jul. 1987. [23] D. C. Ghiglia and M. D. Pritt, Two-Dimensional Phase Unwrapping: Theory, Algorithms, and Software. New York: Wiley, 1998. [24] V. V. Volkov and Y. Zhu, Deterministic phase unwrapping in the presence of noise, Opt. Lett., vol. 28, no. 22, pp. 21562158, 2003. [25] M. A. Schoeld and Y. Zhu, Fast phase unwrapping algorithm for interferometric applications, Opt. Lett., vol. 28, no. 14, pp. 11941196, Jul. 2003. [26] A. Bosseboeuf and S. Petitgrand, Characterization of the static and dynamic behavior of M(O)EMS by optical techniques: Status and trends, J. Micromech. Microeng., vol. 13, no. 4, pp. S23S33, Jul. 2003. [27] COMSOL Multiphysics, COMSOL, Inc., Burlington, MA. [Online]. Available: http://www.comsol.com [28] W. N. Sharpe, S. B. Brown, G. C. Johnson, and W. G. Knauss, Round robin tests of modulus and strength of polysilicon, in Proc. Mater. Res. Soc., San Francisco, CA, 1998, vol. 518, pp. 5765. [29] B. D. Jensen, M. P. de Boer, N. D. Masters, F. Bitsie, and D. A. La Van, Interferometry of actuated microcantilevers to determine material properties and test structure nonidealities in MEMS, J. Microelectromech. Syst., vol. 10, no. 3, pp. 336346, Sep. 2001. [30] O. Holmgren, K. Kokkonen, V. Kaajakari, A. Oja, and J. V. Knuuttila, Direct optical measurement of the Q values of RF-MEMS resonators, in Proc. IEEE Ultrason. Symp., 2005, pp. 21122115.

[31] C. Orsulak, T. Coakley, and M. Zecchino, Characterizing AFM Cantilevers with dynamic MEMS measurement, Veeco Application Notes, 2005. [Online]. Available: www.veeco.com/pdfs.php/116 [32] L. A. J. Davis, D. R. Billson, D. A. Hutchins, and R. A. Noble, , Visualizing acoustic displacements of capacitive micromachines transducers using an interferometric microscope, Acoustic Res. Lett. Online, vol. 6, no. 2, pp. 7579, 2005.

Josh A. Conway (M04) received the B.S. degree in physics and the M.S. degree in electrical engineering from the University of Illinois at UrbanaChampaign in 1999 and 2001, respectively, and the Ph.D. degree in electrical engineering with an emphasis in photonics from the University of California, Los Angeles, in 2006. He is currently a member of Technical Staff at The Aerospace Corporation, El Segundo, CA. Before this, he designed photonic subsystems for intersatellite laser communications for the Boeing Company. This work has led to numerous patents in the eld of optical interferometry. Dr. Conway is a member of The International Society for Optical Engineers.

Jon V. Osborn (S85M87) received the B.S. and M.S. degrees in electrical engineering from the University of Southern California, Los Angeles, in 1985 and 1994, respectively. From 1985 to 1997, he was with the Space Sciences Laboratory, The Aerospace Corporation, Los Angeles, as a member of Technical Staff, developing space instrumentation for near-earth space physics research, including single-event effects testing and evaluation of spaceborne microelectronic devices. Major programs and projects included CRRES/MEA, POLAR/CEPPAD, and CLUSTER/IES. In 1997, he joined the Electronics Photonics Laboratory, The Aerospace Corporation, as a Research Scientist, actively investigating advanced microelectronic technologies for use in space systems, including radiation-tolerant deep-submicrometer CMOS, CMOS/SOS, and MEMS device technologies. Since 2004, he has been the Manager of the Microelectronics Reliability and Radiation Effects Section, Microelectronics Technology Department. Mr. Osborn was an Aerospace Corporation Fellow during 19931994. He has been a Registered Professional Engineer in the State of California since 1994.

Jesse David Fowler received the A.A. degree in auto/diesel mechanics from the Universal Technical Institute, Phoenix, AZ, in 1991, the B.S. degree in mechanical engineering from New Mexico State University, Las Cruces, in 1998, and the M.S. degree in mechanical engineering from the University of California, Los Angeles (UCLA), in 2003. He is currently working toward the Ph.D. degree in mechanical engineering at UCLA. He has worked at Los Alamos National Laboratory, Sandia National Laboratory, and The Aerospace Corporation in a research support capacity. He held the Materials Creation Training Program Fellowship from 2002 to 2003. His current research interests include discrete droplet microuidics, chemical microsensors, and MEMS metrology. Mr. Fowler is a member of Tau Beta Pi and Pi Tau Sigma. He is a certied Engineer-in-Training in the State of New Mexico.