This paper describes MEMS micromirror characterization in space environments associated with our space applications in earth observation from the International Space Station and earth’s orbit satellite. The performance of the micromirror was tested for shock and vibration, stiction, outgassing from depressurization and heating, and electrostatic charging effects. We demonstrated that there is no degradation of the micromirror performance after the space environment tests. A test bed instrument equipped with the micromirrors was delivered and tested in the ISS. The results demonstrate that the proposed micromirrors are suitable for optical space systems.
©2009 Optical Society of America
Optical MEMS (Micro-Electro-Mechanical Systems) devices are products ranging from a micron to a centimeter in size that combine mechanical, electrical, and optical components. Numerous types of optical MEMS devices have been used successfully with a wide range of applications, including optical communications, display systems, biomedical instrumentation, and adaptive optics [1,2]. Because optical MEMS devices have the advantages of small size, high robustness, and low power consumption, they have recently attracted particular attention for space optical systems, such as the James Webb Space Telescope .
The most commonly used optical MEMS device is the micromirror, which has rapid tilting speed and high reliability compared with large conventional mirrors. For space optical systems, research into optical MEMS-based micromirror arrays has been reported for multiobject spectroscopy  and adaptive optics [5,6]. The first sounding-rocket mission, Planet Imaging Concept Testbed Using a Rocket Experiment (PICTURE), was launched in 2007 by Supriya Chakrabarti of Boston University. Boston Micromachines, a provider of MEMS-based deformable mirror products for adaptive optics systems, is developing a 32 × 32 MEMS deformable mirror array with 9.6 mm aperture and 1.5 μm stroke that will aid NASA in the search for extrasolar planets [6–8].
An early flight demonstration of a MEMS device was given by the Jet Propulsion Laboratory (JPL) and Phillips Laboratory. The MEMS package demonstration was designed as an experiment in the electronics testbed on board the Space Technology Research Vehicle-2 (STRV-2, a joint UK/US mission). It was successfully launched in 2000 . The experiment of Microsystems and Packaging for Low-Power Electronics II (MAPLE-2) included the integration of Analog Devices ADXL02, ADXL05 (accelerometers) and a single tunneling microaccelerometer developed at JPL . The MEMS Technology Group at JPL pursues the development of advanced MEMS/NEMS technologies for space applications. These technologies have included a spider-web bolometer detector, miniature vibratory gyroscopes, an electron luminescence X-ray spectrometer, a thermal transpiration pump, a MEMS-based inchworm actuator, nanowire-based biochemical sensors, microinertial reference systems, and a system on a chip [11,12].
Given the rapid progress of MEMS technologies for space systems, the performance characterization of MEMS devices in space environments becomes critical. The failures associated with a variety of MEMS devices in the unique environments of space should be identified and analyzed. We found one manual about MEMS reliability assurance guidelines for space applications from the JPL . However, no performance characterization of MEMS devices in orbit space environments has been reported as of today.
In our earlier studies, we reported the performance and reliability of MEMS micromirrors [14,15] integrated with an obscura-type space telescope called the MEMS space Telescope for Extreme Lightning (MTEL) . In this paper, we report the design issues in section 2. Section 3 describes the results of space environment tests for operation in orbit and in the International Space Station (ISS). These include shock and vibration (subsection 3.1), stiction prevention (subsection 3.2) and electrostatic charging tests (subsection 3.3) for both environments. In particular, we discuss micromirror operation up to the maximum humidity level following the ISS instrumentation procedure in subsection 3.2. In subsection 3.4, the outgassing test required for the satellite environment is presented. This is followed by the micromirror performance checks discussed in subsection 3.5. The micromirrors were then transported to the ISS Russian Segment (RS), where micromirror performance tests were carried out in space for one week starting from April 11, 2008. The test results of the micromirror in the ISS are reported in subsection 3.6. MEMS micromirror characterization techniques and the results presented in this paper can contribute to improvement of MEMS device guideline for space optical systems.
2. MEMS space telescope and micromirrors
Our MEMS micromirrors were designed, fabricated, and tested for two space missions. First, a test bed instrument shown in Fig. 1 was launched in April 2008. The instrument was installed by the first Korean astronaut in the ISS and the response of the MEMS micromirror was measured for a week. The ISS is pressurized and maintains its inner temperature and humidity within a range appropriate for a human living space.
Second, our space telescope (MTEL) equipped with MEMS micromirrors is scheduled to be delivered into the earth’s orbit in February 2009 by a Russian microsatellite, Tatiana-II, to observe transient luminous events from space. The launch and orbital operation expose the satellite payload to vacuum and large temperature excursions that are substantially different from the conditions within the pressurized module in the ISS. The details of the operational principle and structure of the telescope are presented in . We report two types of MEMS micromirrors actuated by electrostatic force: a parallel-plate and a comb-drive micromirror and the space environment tests performed in environmental conditions commonly found in the ISS and the earth’s orbit satellite.
For both space missions, a one-axis parallel-plate micromirror consisting of a mirror plate, a pair of torsional springs, and two addressing electrodes are developed as shown in Fig. 2. The micromirror plate is suspended by the torsional springs. Single-crystal silicon was selected as an applicable material for the micromirror because of its properties such as negligible residual stress, high yield strength, high temperature resistance, and flat surface in comparison with various metals. Reflective material was deposited on top of the micromirror. The one-axis micromirror dimension is 250 × 250 × 4 μm3. A spring is designed to be 2 (W) × 40 (L) × 4 (H) μm3. Structurally, when the voltage difference between the mirror plate and the bottom electrode exceeds the so-called pull-in voltage, the mirror plate collapses abruptly onto the substrate because of the pull-in instability. Using this phenomenon, the mirror can have two discrete states: “on” and “off.” Resonance frequency and pull-in voltage were measured to be 5.4 kHz and 103.4 V, respectively. The maximum deflection angle is 7.44° at the pull-in state. As shown in Fig. 2(c), there are landing sites for micromirror tips to contact the grounded region. The micromirror tips also minimize the contact area between a mirror plate and landing sites. The micromirror array consists of 16 × 16 reflectors.
As shown in Fig. 3(a), a two-axis comb-drive micromirror consists of a number of tiny vertical combs that are electrostatic actuators. The core of the actuation mechanism is an actuator pair of top comb and bottom comb below a mirror plate, as shown in Fig. 3(b). The inner spring of the actuator connects the inner frame to the outer frame, while the outer spring connects the outer frame to the fixed supporting body, as depicted in Fig. 3. A DC bias at the comb electrodes attached to the outer frame provides electrical torque for tilting along the outer spring, while the DC bias at the comb electrodes located inside the frame generates torque to tilt the inner plate. Two orthogonal pairs of springs allow the mirror plate to be tilted independently in two orthogonal directions. The comb-drive micromirror is fabricated by bonding three wafers: a glass wafer (a substrate including addressing electrodes); a Silicon On Insulator (SOI) wafer (two vertical comb-drive layers); and a polished silicon wafer (a mirror plate). A mirror plate bonded onto an inner comb-drive actuator is 340 × 340 μm2. The fill-factor of an 8 × 8 array is 84%. The outer and inner comb-drive actuators are electrically controlled to achieve deflection angles of up to 4.42° and 2.90°, respectively. By combining rotations along the two axes, the tilting angle of the micromirror can be controlled analogously in any direction. The resonance frequency of the inner comb-drive actuator is 0.95 kHz and the resonance frequency of the outer actuator is 1.9 kHz; these determine the response time of the micromirror.
3. Space environment tests
3.1 Shock and vibration
Cunningham et al. studied the shock robustness of silicon microstructures and found that appropriate stress distribution contributed to the shock robustness . Shock and vibration tests on MEMS sensors were performed and discussed by Brown et al. . According to these studies, each of the elements in a MEMS device could be badly damaged by continuous vibration or severe shock. For space applications, as devices are subject to severe shock and vibration at the launching and orbital stages, consideration of and prerequisite tests for shock and vibration are very critical . Because the structure of a MEMS micromirror is complex with movable beams, comb fingers, mirror plates, and springs, the robustness of micromirrors against shock and vibration for space applications should be guaranteed.
The reliability with regard to shock and vibration of MEMS micromirrors is calculated using Newtonian physics. We use the one-axis micromirror in Fig. 2 as an example. The mass is calculated as the product of area, thickness, and density of single-crystal silicon. The force from the expected operational maximum g-load is given by:
where m is 5.825 × 10-10 kg and a is 392 m/s2. For any material, fracture occurs if the stress applied exceeds the fracture strength, which is known to be 30 GPa  for single-crystal silicon. The stress on the simple cantilever is given by:
where σ is the applied stress, F is the force applied to the center of the micromirror with spring length L, t is the thickness of the spring in the direction of the force, and I is the moment of inertia. To calculate the moment of inertia, we can assume the spring is a rectangular beam and the moment equation can be described by:
where b is 3 μm, t is 3 μm, and L is 50 μm. Because the micromirror has a pair of springs, the force is distributed, implying that the springs could be fractured at twice the calculated Ffracture. From Eq. (1) and Eq. (4), 2Ffracture is 8758 times larger than Fapp. The safety factor (2Ffracture/Fapp) is indicated to exceed 8000, and the micromirror could not be fractured even by the maximum expected operational g-load.
The deformation and induced stress distribution of the micromirror were simulated for 40 g acceleration, using the ANSYS workbench. As shown in Fig. 4(a), the maximum deformation of the micromirror is expected to be 1.1 nm. The maximum Von-Mises stress is 499.7 MPa, much lower than the fracture strength of single-crystal silicon, as shown in Fig. 4(b). In the analysis shown in Fig. 4, force with 40 g acceleration is applied in the z-direction to stimulate vertical motion of the micromirror. The maximum deformation and stress are smaller for the x- and y-directional loadings than for the z-directional loading. The simulation results imply the high sustainability of the micromirror under the 40 g acceleration. Although the two-axis comb-drive micromirror is heavier than the one-axis micromirror, the safety factor is still a couple of hundred with the 40 g acceleration.
We also performed shock and vibration tests on the micromirrors in three perpendicular directions using the test procedure defined in . The spectral profile of the random vibration is shown in Fig. 5(a). The 40 g shock at the launching stage is the maximum operational g-load to qualify the test, as shown in Fig. 5(b). Figure 6 shows the measured resonance frequency of the parallel-plate micromirror and the outer actuator of the comb-drive micromirror. No structural damage and no resonance frequency changes of the micromirrors were observed after all of the shock and vibration tests. The simulation and test results demonstrated the survivability of our MEMS micromirrors under the space environmental shock and vibration.
Stiction is the phenomenon whereby microstructures tend to adhere to each other when their surfaces come into contact [22,23]. Typically, when the surfaces of MEMS devices come into contact, the large interfacial forces such as capillary, Van der Waals, and electrostatic, are strong enough to irrevocably bond the two surfaces. Usually, the structure does not recover its initial state once it contacts another surface, unless the restoring force of a spring is sufficient to overcome the stiction force. High humidity during actuating increases the possibility of stiction phenomena for MEMS devices. Even permanent adhesion can result from the large interfacial forces in high humidity .
Stiction can be prevented or decreased in several ways . To minimize the stiction problem caused by electrostatic attraction, the micromirror contact points should be conductive to allow the charge to be dissipated. For the micromirrors in Figs. 2 and 7, we designed aluminum landing sites grounded in common so that the mirror plate would not stick to the landing spot. In addition, field-limiting shields around actuation electrodes play a vital role in protecting the mirror from stiction, because the charge on the dielectric layer is dissipated through the shields. Figure 7 shows a pair of actuation electrodes and aluminum shields with the mirror plate removed. The field-limiting shields, which cover the revealed glass surface, surround the electrodes to suppress possible stiction induced by the dielectric layer. The shield electrodes employed in the proposed micromirror are also effective for mitigating other important effects of charging in the dielectric layer that cause the tilt angle to drift or cause malfunction of the mirror. The charging effects are discussed in the next section.
We conducted a humidity test on the micromirrors in a chamber following the test procedures illustrated in Fig. 8 . The space qualification test on humidity requires a micromirror to sustain its functionality for six hours at 95% humidity and 20 ± 2 °C. We activated all the micromirrors after a three-hour interval at 95% humidity. With the gradient of 1.5 °C/min, humidity steadily decreased to 80% and the micromirrors were kept at that humidity for four hours. None of the micromirrors showed evidence of stiction. Finally, the functionality of the micromirror was checked again. We observed no degradation in functionality for any mirror after the humidity test
3.3 Electrostatic charging
One of the most important problems of MEMS devices with dielectric layers is charging effects . Charge accumulation may cause drift of the micromirror even at zero volts, resulting in touchdown to the bottom substrate. Charge is trapped in the dielectric of a micromirror because of radiation in space, and also because of the strong electric field when the actuation voltage is applied to the electrodes. The reasons for charging and the mitigation techniques are the same, regardless of the source .
Typical radiation levels for earth orbiters in units of rads (1 rad = 6.25 × 107 MeV/g) vary with altitude and can reach up to 107 rad per year . The severe radiation environment in space may cause a variety of unwanted problems for MEMS devices. To remove charging effects caused by injected and trapped charges in a radiation environment, we employed field-limiting shields in the micromirrors, as shown in Fig. 7. They prevent not only exposure of dielectric material to radiation, but also charge accumulation. They also allow the micromirror to be immune to the electrostatic charging derived from the electric field associated with the actuation voltage. Figure 9 shows the measured tilt angle drift with and without shields when 83% of the pull-in voltage, equivalent to 108 V/m, is applied to the micromirror. Without any shield, the micromirror drifted by nearly 0.5°, as shown in Fig. 9(b). In contrast, less than 0.005° drift was observed in the micromirror with field-limiting shield, as shown in Fig. 9(a).
3.4 Outgassing (depressurization and thermal effect)
Micromirrors in space operate in an extremely high vacuum. Outgassing from any materials such as printed circuit boards, cables and adhesives will result in the loss of mirror reflectivity. In addition, the thermal requirements are quite severe, with repetitive cycles of low and high temperature excursions during a mission. Temperature excursions can accelerate the outgassing from materials that can contaminate the surface of the mirror.
An aluminum layer of 800-Å thickness is used as a reflective material on the micromirror. The mirror reflectivity was measured before and after being exposed to outgassing in a thermal vacuum chamber for 24 hours, as shown in Fig. 10. The chamber was depressurized down to 10-3 Torr and the temperature was set to 60 °C, which are the standard conditions required for outgassing tests of Russian microsatellites . After exposure to the outgassing, the mirror’s reflectivity degraded by less than 2%, which is tolerable in space applications.
3.5 Integration of the micromirrors to the telescope
We demonstrated the telescope-level functionality using the micromirrors after completion of all the mirror-level space environment tests. The telescope details are described elsewhere . The trigger micromirror positioned closest to the photodetector provides a wide field of view. It triggers events and toggles on and off to observe an event of interest. The trigger micromirror directs light to a photodetector and the light signal is quickly analyzed to determine the position of the light source. On finding the source position, the trigger micromirror is immediately turned “off” (directing the light out of the photodetector). The zoom micromirror is actuated in correspondence to the position information of the triggered event so that the image is aligned to the center of the photodetector.
In a laboratory test, the responses of 64 pixels of a Multi-Anode Photo-Multiplier Tube (MAPMT) to an incident UV LED light reflected by the trigger and zoom mirrors are shown in Figs. 11 and 12, respectively, where the z-axis denotes the light intensity in the analog-to-digital converter (ADC) unit. In Fig. 11(a), the event is detected in the right top corner of the MAPMT. Figures 11(b) to (f) show observed images, captured at 10 μs intervals, as the micromirror toggles off at a speed of 0.1 degree/μs. Pictures from Fig. 12(a) to (i) represent the close-up images formed by the zoom micromirror, moving gradually toward the center of the photo detector. These results demonstrate that the proposed micromirrors show fast detection and identification sequences for ultrafast transient phenomena.
3.6 Performance of the micromirror in the ISS
The performance of the micromirror in the ISS is compared with that of ground laboratory measurement. The micromirror was kept in nominal eight-hour operation every day in the ISS, and the measurement lasted for seven days (April 11–17, 2008), toward the tail end of the mission Expedition 16 . An MAPMT was used for the photodetector, providing high temporal resolution. The test micromirror toggled on and off with the period of 20 ms, as the MAPMT detector received reflected light originating from a UV LED source. The micromirror was actuated with the square wave voltage signal alternating between two levels. Figures 13(a) and (b) illustrate the response before the launch and in the ISS, respectively. The x-axis refers to time and the y-axis shows the ADC values. Shown are 1000 data samples, each taken every 10 μs. The ADC level of about -300 means that the micromirror directs the light onto the pixel in the toggle-on state. The ADC level of around -900 represents the toggle-off state of the micromirror. No degradation in the MAPMT response, that is, no degradation in the micromirror performance, was observed at the ISS compared with the performance on the ground before the launch.
Space is a unique environment that may cause severe failures of MEMS devices. We have addressed the performance characterization of MEMS micromirrors in space environments for proposed space applications. The micromirrors were designed to sustain their performance in the shock and vibration environment of the ISS or free flyer orbit. Thanks to the sharp, grounded landing tips and field-limiting shields, no stiction of the micromirror was observed, even under 95% relative humidity. The field-limiting shields played a role in preventing the tilting angle drift and eventual touchdown of the micromirror by minimizing the electrostatic charging effects. After an outgassing test at 60 °C under the depressurized condition, the reflectivity of the micromirrors was hardly influenced by outgassing generated from various materials.
Having characterized our MEMS micromirrors under unique environments for operation in the ISS and the earth’s orbit, we measured for the first time the performance of the MEMS micromirror in the ISS. The successful operation of the micromirror in the ISS proved the validity and applicability of the ground environment test procedure (reported in this paper) for space qualification of a wide range of MEMS micromirrors to come.
This work was supported by Creative Research Initiatives (RCMST) of MEST/KOSEF. We would like to thank So-Yeon Yi, the first Korean astronaut, for taking invaluable measurement data in the ISS. We also would like to thank the members of the Korean astronaut program team in KARI, especially Gi-Hyuck Choi, Youn-Kyu Kim, Sang-Wook Kang and Mi-Hyeon Lee, for their efforts in delivering our instrument to the ISS.
References and links
1. P. F. V. Kessel, L. J. Hornbeck, R. E. Meier, and M. R. Douglass, “A MEMS-based projection display,” in Proceedings of IEEE Special Issue: Integrated Sensors, Microactuators, and Microsystems (MEMS)86, 1687–1704 (1998).
2. R. S. Muller and K. Y. Lau, “Surface-micromachined microoptical elements and systems,” in Proceedings of IEEE Special Issue: Integrated Sensors, Microactuators, and Microsystems (MEMS)86, 1705–1720 (1998).
3. A. S. Kutyrev, R. Arendt, S. H. Moseley, R. A. Boucarut, T. Hadjimichael, M. Jhabvala, T. King, M. J. Li, J. Loughlin, D. Rapchun, D. S. Schwinger, and R. F. Silverberg, “Programmable microshutter arrays for the JWST NIRSpec: optical performance,” IEEE J. Sel. Top. Quant. Elec. 10, 652–661 (2004). [CrossRef]
4. S. Waldis, P. Ayyalasomayajula, W. Noell, and N. F. dep Rooij, “Micromirrors for multiobject spectroscopy: Large array actuation and cryogenic compatibility,” in Proceedings of IEEE Int. Conf. Optical MEMS and Nanophotonics, (Huwalian, Taiwan, 2007), pp. 87–88. [CrossRef]
5. J. Porter, H. Queener, J. Lin, K. Thor, and A. Awwal, Adaptive Optics for Vision Science (Wiley-Interscience, 2006), Chap. 4. [CrossRef]
6. T. G. Bifano, J. Perreault, R. Krishnamoorthy-Mali, and M. N. Horenstein, “Microelectromechanical Deformable Mirrors,” IEEE J. Sel. Top. Quant. Elec. 5, 83–90 (1999). [CrossRef]
7. D. Content, S. Antonille, D. Rabin, and T. Wallace, “Sounding-rocket telescope uses new technology ultralight-weight mirrors,” (SPIE newsroom, 2006), http://spie.org/documents/Newsroom/Imported/260/2006070260.pdf.
8. J. A. Perreault, T. G. Bifano, B. M. Levine, and M. N. Horenstein, “Adaptive optics correction using microelectromechanical deformable mirrors,” Opt. Eng. 41, 561–566 (2002). [CrossRef]
9. L. Muller, M. H. Hecht, L. M. Miller, H. K. Rockstad, and J. C. Lyke, “Packaging and qualification of MEMS-based space systems,” Proc. MEMS, (Atlanta, USA, 1996), pp. 503–508.
10. M. Griffin, B. Swinyard, and L. Vigroux, “SPIRE - Herschel’s submillimetre camera and spectrometer,” Proc. SPIE 4850, 686–697 (2002). [CrossRef]
11. T. George, “MEMS/NEMS development for Space Applications at NASA/JPL,” Proc. SPIE 4755, 556–567, (2002). [CrossRef]
12. L. M. Miller, “MEMS for space applications,” Symp. on design, test and microfabrication of MEMS/MOEMS, (Paris, France, 1996), pp. 1–11.
13. B. Stark, “MEMS reliability assurance guidelines for space applications,” in JPL publication 99-1 (1999).
14. Y. H. Jang, K. N. Lee, and Y. K. Kim, “Characterization of a single crystal silicon micromirror array for maskless UV lithography in biochip applications,” J. Micromech. Microeng. 16, 2360–2368 (2006). [CrossRef]
15. B. W. Yoo, J. H. Park, Y. H. Jang, and Y. K. Kim, “A low-drift, open-loop controlled, single crystalline silicon micromirror with floating field-limiting shields,” J. Micromech. Microeng. 18, 035031 (2008). [CrossRef]
16. S. Nam, S. Artikova, T. Chung, G. Garipov, J. A. Jeon, S. Jeong, J. Y. Jin, B. A. Khrenov, J. E. Kim, M. Kim, Y. K. Kim, P. Klimov, J. Lee, H. Y. Lee, G. W. Na, S. J. Oh, M. Panasyuk, I. H. Park, J. H. Park, Y.-S. Park, B. W. Yoo, and H. J. Yoo, “A telescope for observation from space of extreme lightning in the upper atmosphere,” Nucl. Instr. and Meth. A 588, 197–200 (2008). [CrossRef]
17. S. J. Cunnigham, D. G. Mclntyre, J. S. Carper, P. D. Jaramillo, and W. Tang, “Microstructures designed for shock robustness,” Proc. SPIE 2880, 99–107 (1996).
18. T. G. Brown and B. Davis, “Dynamic high-g loading of MEMS Sensors: ground and flight testing,” Proc. SPIE 3512, 228–235(1998). [CrossRef]
19. D. M. Tanner, J. A. Walraven, K. Helgesen, L. W. Irwin, F. Brown, N. F. Smith, and N. Masters, “MEMS reliability in shock environments,” in Proceedings of IRPS, (San Jose, USA, 2000), pp. 129–138.
20. C. Shepherd, M. Zalalutdinov, T. Alan, H. Craighead, A. Zehnder, and J. Parpia, “Fracture strength of single crystal silicon at micron dimensions,” (Department of Physics, University of Florida, 2003), http://www.hgc.cornell.edu/Nems%20Folder/fracture%20paper.pdf.
21. A. N. Perminov, For the space mission of a space flight participant to the Russian segment of the International Space Station (ISS) Description of ISS RS and general technical requirements (Russian Federal Space Agency, 2007), Appendix I. [PubMed]
22. J. A. Walraven, “Failure mechanism in MEMS,” in Proceedings of ITC international test conference, (Charlotte, NC, 2003), pp. 808–833.
23. W. M. van Spengen, R. Pures, and I. De Wolf, “A physical model to predict stiction in MEMS,” J. Micromech. Microeng. 12, 702–713 (2002). [CrossRef]
24. D. M. Tanner, “The effect of humidity on the reliability of a surface micromachined microengine,” in Proceedings of IRPS, (San Diego, Calif., 1999), pp. 189–197.
25. W. M. van Spengen, “MEMS reliability from a failure mechanisms perspective,” Microelectronics Reliability 43, 1049–1060 (2003). [CrossRef]
26. H. R. Shea, “Reliability of MEMS for space applications,” Proc. SPIE 6111, 61110A (2006). [CrossRef]
27. J. A. Wibbeler, G. Pfeifer, and M. Hietschold, “Parasitic charging of dielectric surfaces in microelectromechanical systems (MEMS),” Sensors and Actuators A 71, 74–80 (1998). [CrossRef]
28. J. R. Wertz, Space Mission Analysis and Design, W. J. Larson ed. (Kluwer Academic Publishers, 1991), pp. 183–193.
29. NASA, “International Space Station, expedition 16,” http://www.nasa.gov/mission pages/station/expeditions/expeditionl6/index.html.