Robots and drones are presently in the industry’s focus to serve a critical role in Industry 4.0 and the Transportation Revolution. Integration of robots and drones in these areas improves efficiency and safety, adds flexibility in operation, and reduces operating costs. However, they are still far from achieving the optimal performance needed to execute autonomous tasks at high levels. As these platforms are battery operated, all sub-systems that augment their capabilities must be low-power solutions. In the case of airborne drones, it is also critical that solutions are ultra-light weight and of small form factor. Additionally, robots will be employed in the modern working environment in tandem with humans, but adequate human-robot interaction and intention communication solutions do not currently exist. Consequently, MEMS mirrors-based sensing and interaction systems designed for robots and drones are essential as they offer solutions with the lowest power consumption, weight, and cost in high volume. However, existing MEMS Mirror based solutions have not achieved the necessary compactness and efficiency for robotics. In this paper we describe and demonstrate MEMS Mirror-based 3D perception sensing (SyMPL 3D Lidar) and animated visual messaging (Vector Graphics Laser Projection with Playzer) systems optimized for robots and drones. These sub systems each consume <1W in power, at least 10x lower than other solutions in the market, weigh <50g, and have small form factors. Furthermore, we will show that combining these two systems leads to new capabilities and functionalities that meet the demands of robot vision and human-robot interaction.
LiDAR systems in applications such as autonomous mobile robots, drones, vehicles, and other commercial applications that demand compact, low-cost, and dynamic scanning will inevitably turn to MEMS mirrors as the beam-steering component. Beam scanning-based LiDAR architectures have a significant advantage as the full power and attention of the sensor is given sequentially to each point (voxel) in the scan. Competitive LiDAR designs typically utilize scanning and are differentiated by their scanning architecture and the specific hardware utilized, with the general goal of moving away from bulky mechanical and motor-based systems and toward compact silicon-based MEMS technology.
Both single-axis and dual-axis MEMS mirrors are employed to enable two-dimensional (2D LiDAR) and three dimensional (3D LiDAR) point cloud sensing, respectively. The underlying time-of-flight sensor can be generic – a laser rangefinder or single-point LiDAR, with any typical wavelength or sensing method (pulsed ToF, AMCW, FMCW, etc.). The sensor is arranged with scanning elements which brings forth challenging trade-offs, discussed here. Architectures differ in whether transmitter and receiver are arranged coaxially or biaxially, each with its advantages and disadvantages. We present a hybrid architecture, Synchronized MEMS Pair LiDAR (SyMPL), which simplifies the coaxial design significantly and increases its efficiency by removing any beam splitting components or beam dumps. Multiple prototype LiDARs are compared and evaluated on the basis of SNR, scan speed, robustness to shock and vibration, eye safety, and resilience to mutual interference and echo signals. The work discusses the varying impacts on manufacturing and cost for applications demanding large volumes of LiDAR systems.
An updated Programmable Light System (PLS) is demonstrated using a MEMS Mirror Module (MMM), allowing users to program the brightness and shape of a projected white light in a variety of dynamic solid-state lighting applications, e.g. in automotive dynamic headlights. The MMM is a new module which consists of a fast beam steering MEMS mirror with high optical laser power handling and a smart MEMS Driver with real time monitoring of the MEMS mirror for better system safety and mirror control. The PLS consists of the MMM, a multi-Watt 445-450nm laser source with beam shaping optics, a phosphor target, and projection optics to project a white light within the field of view of up to 60°. Devices such as the 1.2mm diameter A3I12.2 and 2.0mm diameter A7M20.1 aluminum coated mirror have been tested at >8W of CW power before seeing any damage to the device. The A7M20.1 MEMS mirror has been extensively tested (>100 hours) with 4W of CW power at room temperature with no physical damage. Same 4W operation, has also been successfully tested at elevated environmental temperature of 100°C during extended tests.
PLS prototypes to date utilize only ~1W-2W laser diode sources, as limited by power of available laser diodes. The extended tests and thermal studies of the MMM however show that operation at up to 100°C with e.g. 4W CW power could be safely run for at least 10000 hours, even with MEMS mirrors with a simple aluminum coating (no protection or enhancement layers).
Gimbal-less dual axis point-to-point (quasistatic) MEMS mirrors have very wide bandwidths for laser beam steering, however users are often limited to only a third of the bandwidth due to the high Q-factor and use of low-pass filters in open-loop operation to avoid overshoot and oscillation. Closed-loop driving enables the use of the full bandwidth with additional complexity in optics and electronics, which can be undesirable in some low SWaP-C systems. But for many applications which require scanning repetitive patterns, such as LiDAR and biomedical imaging, bandwidth utilization and linearity can be greatly increased without any real-time feedback control by training the device and finding an optimal driving waveform using an iterative learning algorithm.
The algorithm drives the device with a trial waveform, measures the scan on a Position Sensing Device (PSD), calculates the error between the desired waveform and the measured position, and adjusts the drive waveform for the next iteration based on an approximate linear device model. This is repeated until the error is reduced to below an acceptable specification. The waveform is then saved in the MEMS Controller and can be reliably used for extended periods of operation. Multiple such drive signals can be trained and stored on the controller to perform different types of scans.
Several MEMS mirrors, including single- and dual-axis designs, were studied and three are reported here. Overall, in all cases a high accuracy of optical scans is achieved, typically to within ±0.025° of nominal. Repeatability after training, then running in open loop is better than ±0.01° - however, this measurement was limited by the lower resolution of the position detecting sensor. Scan rates achieved vary based on mirror design, but in each case are greatly improved from those achievable with basic driving approaches. Each mirror demonstrated higher quality vector graphics content at faster refresh rates and stable linear rasters at rates below resonance where lines are scanned with uniform velocity. Additionally, each mirror could achieve stable fast rasters with the line-scanning axis rates just below resonance, giving sinusoidal scans with line rates of ~1.6fres. Finally, each mirror was also demonstrated achieving rasters with rates above resonance, giving line rates of ~2.5fres. In all of those cases the other axis could scan linear and sharp sawtooth or triangle waveforms. Based on the symmetry of the MEMS design, we demonstrated the same performance at ddegerent angles, e.g. rastering at a 45° angle.
KEYWORDS: Mirrors, Microelectromechanical systems, LIDAR, Control systems, Micromirrors, Beam controllers, Laser range finders, Beam steering, Distance measurement, Linear filtering, Sensors, Silicon, Photodiodes, Signal to noise ratio
In 2016, we presented a low SWaP wirelessly controlled MEMS mirror-based LiDAR prototype which utilized an OEM laser rangefinder for distance measurement [1]. The MEMS mirror was run in open loop based on its exceptionally fast design and high repeatability performance. However, to further extend the bandwidth and incorporate necessary eyesafety features, we recently focused on providing mirror position feedback and running the system in closed loop control. Multiple configurations of optical position sensors, mounted on both the front- and the back-side of the MEMS mirror, have been developed and will be presented. In all cases, they include a light source (LED or laser) and a 2D photosensor. The most compact version is mounted on the backside of the MEMS mirror ceramic package and can “view” the mirror‟s backside through openings in the mirror‟s PCB and its ceramic carrier. This version increases the overall size of the MEMS mirror submodule from ~12mm x 12mm x 4mm to ~15mm x 15mm x 7mm. The sensors also include optical and electronic filtering to reduce effects of any interference from the application laser illumination. With relatively simple FPGA-based PID control running at the sample rate of 100 kHz, we could configure the overall response of the system to fully utilize the MEMS mirror‟s native bandwidth which extends well beyond its first resonance. When compared to the simple open loop method of suppressing overshoot and ringing which significantly limits bandwidth utilization, running the mirrors in closed loop control increased the bandwidth to nearly 3.7 times. A 2.0mm diameter integrated MEMS mirror with a resonant frequency of 1300 Hz was limited to 500Hz bandwidth in open loop driving but was increased to ~3kHz bandwidth with the closed loop controller. With that bandwidth it is capable of very sharply defined uniform-velocity scans (sawtooth or triangle waveforms) which are highly desired in scanned mirror LiDAR systems. A 2.4mm diameter mirror with +/-12° of scan angle achieves over 1.3kHz of flat response, allowing sharp triangle waveforms even at 300Hz (600 uniform velocity lines per second). The same methodology is demonstrated with larger, bonded mirrors. Here closed loop control is more challenging due to the additional resonance and a more complex system dynamic. Nevertheless, results are similar - a 5mm diameter mirror bandwidth was increased from ~150Hz to ~500Hz.
In this work we demonstrate a highly flexible laser imaging system for 3D sensing applications such as in tracking of VR/AR headsets, hands and gestures. The system uses a MEMS mirror scan module to transmit low power laser pulses over programmable areas within a field of view and uses a single photodiode to measure the reflected light. User can arbitrarily select the number of pixels to scan over an area and can thus obtain images of target objects at arbitrarily fast rates. The work builds on the previously presented “MEMSEye” laser tracking technology which uses a fast steering MEMS scan module with a modulated laser, and a tuned photosensor to acquire and track a retroreflector-marked object. To track VR/AR headsets, hands and similar objects with multiple markers or no markers at all, a single-point tracking methodology is not sufficient. Cameras could be more appropriate in such multi-point imaging cases but suffer from low frame rates, dependence on ambient lighting, and relatively low resolution when without zooming and panning capability. A hybrid method can address the problem by providing a system with its own light source (laser beam), and with full programmability of the pixel locations and scans such that frame rates of >100 Hz are possible over specific areas of interest. With a modest 1 Mpixel rate of measurement, scanning a sub-region of the field of view with 64 x 64 pixels results in ~200Hz update. Multiple such modules can be used to scan and image or track objects with multiple markers and fully obtain their position and attitude in a room with sub-5ms updates. Furthermore the room itself could be imaged and measured with wall markers or in conjunction with a camera for a total 3D scanning solution. Proof of concept demonstrator is presented here with pixel rates of only 30k-50k per second due to limitations of the present prototype electronics, resulting in refresh rates that are significantly lower than possible with the MEMS mirror scan modules.
2D quasistatic (point-to-point) gimbal-less MEMS mirrors enable programmable, arbitrary control of laser beam position and velocity - up to their maximum limits. Hence, they provide the ability to track targets, point lasercom beams, and to scan uniform velocity lines over objects in laser imaging. They are becoming increasingly established in applications including 3D scanning, laser marking and 3D printing, biomedical imaging, communications, and LiDAR. With the increased utility in applications that demand larger mirror sizes and larger overall angle*diameter (θ*D) figures of merit, the technology is continuously pushed against its limit. As a result we have implemented mirrors with larger diameters including 5.0mm, 6.4mm, and 7.5mm, and have designed actuators with larger torque and angles to match the Θ*D demand. While the results have been very positive in certain application cases, a limitation for their more wide-spread use has been the relatively high susceptibility of large- θ*D mirrors to shock and vibrations. On the other hand, one of the challenges of MEMS mirrors of small diameters is their lower optical power tolerance simply due to their smaller area and heat removal ability. Although they can be operated at up to 2-3W of CW laser power, new developments in dynamic solid state lighting in e.g. headlights demand operation at up to 10W or beyond.
In this work we study and present several package-level approaches to increase mechanical damping, shock robustness, and laser power tolerance. Specifically, we study back-filling of MEMS packages with different gases as well as with different (increased) pressures to control damping and in turn increase robustness and useable bandwidth. Additionally, we study the effects of specialized mechanical structures which were designed and fabricated to modify packages to significantly reduce volumes of space around moving structures.
In their standard form and packaging the MEMS mirrors tested in this study typically measure quality factors of 75-100. Increases of pressure up to 50psi have shown relatively modest reductions of the overall quality factor to the 40-50 range. Backfilling of packages with heavier inert gasses such as Ar and SF6 results in lowering of the quality factor down to 20-30 range. Mechanical modifications of the package with special structures and reduced air-gap to the window yielded the best results, reducing the quality factor to ~9-14. Combination of specialized packaging structures and gas backfill and pressure control could provide a very efficient heat transfer from the mirror and the desired near-critical damping, but has not been demonstrated yet. The increased performance does not change the compactness and low power consumption - the improved MEMS mirrors still consume <1mW. So far, designs with mirror sizes through 3.0mm diameter with increased damping have passed 500G shock tests.
In terms of improved heat removal we have found that the packaging improvement greatly increased optical power tolerance of MEMS mirrors from few Watts of CW laser power to <10 Watts. The exact numbers for the upper limit are not yet available - in samples where the heat removing structure was added and air was replaced with Helium, our setup with 3 combined lasers was not able to damage any samples.
Gimbal-less two-axis quasistatic MEMS mirrors have the ability to reflect optical beams to arbitrary positions and with arbitrary velocity. This technology has become established in many applications including laser based tracking, 3D scanning, biomedical imaging, free-space communication, and LiDAR. However, for certain defense applications, the total angle × diameter product, or the mirror’s effective achievable resolution (θ*D product), has not been large enough to address requirements for agile steering in large fields of regard and with a low diffraction-limited beam divergence. Two key limitations have been the relatively low forces available in electrostatic combdrive actuators and the susceptibility of large-diameter MEMS mirrors to shock and vibrations. In this work, we demonstrate that these same MEMS mirrors can have dramatically increased performance when fully immersed and packaged in dielectric liquids with highly favorable torque-increasing, damping-increasing, and optical gain-increasing properties. The rotating electrostatic combdrive has its torque multiplied by liquid’s relative permittivity of ~2.5. Furthermore, by selecting the appropriate fluid viscosity, quality factor of the device is reduced and structural damping is tuned to near critical damping. Finally, the increased scan angle due to the ~1.5-1.7 index of refraction of the fluid is an additional benefit. These numerous benefits of the fluidic packaging enabled us to double and in some cases triple the previously achieved θ*D product of two-axis quasistatic MEMS mirrors while still maintaining speeds applicable for above mentioned applications. One of the most exciting benefits of the packaging methodologies is that the damping dramatically increases shock and vibration tolerance, which will be tested next.
Firstly, we demonstrated a wirelessly controlled MEMS scan module with imaging and laser tracking capability which can be mounted and flown on a small UAV quadcopter. The MEMS scan module was reduced down to a small volume of <90mm x 60mm x 40mm, weighing less than 40g and consuming less than 750mW of power using a ~5mW laser. This MEMS scan module was controlled by a smartphone via Bluetooth while flying on a drone, and could project vector content, text, and perform laser based tracking. Also, a “point-and-range” LiDAR module was developed for UAV applications based on low SWaP (Size, Weight and Power) gimbal-less MEMS mirror beam-steering technology and off-the-shelf OEM LRF modules. For demonstration purposes of an integrated laser range finder module, we used a simple off-the-shelf OEM laser range finder (LRF) with a 100m range, +/-1.5mm accuracy, and 4Hz ranging capability. The LRFs receiver optics were modified to accept 20° of angle, matching the transmitter‟s FoR. A relatively large (5.0mm) diameter MEMS mirror with +/-10° optical scanning angle was utilized in the demonstration to maintain the small beam divergence of the module. The complete LiDAR prototype can fit into a small volume of <70mm x 60mm x 60mm, and weigh <50g when powered by the UAV‟s battery. The MEMS mirror based LiDAR system allows for ondemand ranging of points or areas within the FoR without altering the UAV‟s position. Increasing the LRF ranging frequency and stabilizing the pointing of the laser beam by utilizing the onboard inertial sensors and the camera are additional goals of the next design.
A high brightness Head-Up Display (HUD) module was demonstrated with a fast, dual-axis MEMS mirror that displays
vector images and text, utilizing its ~8kHz bandwidth on both axes. Two methodologies were evaluated: in one, the
mirror steers a laser at wide angles of <48° on transparent multi-color fluorescent emissive film and displays content
directly on the windshield, and in the other the mirror displays content on reflective multi-color emissive phosphor plates
reflected off the windshield to create a virtual image for the driver. The display module is compact, consisting of a
single laser diode, off-the-shelf lenses and a MEMS mirror in combination with a MEMS controller to enable precise
movement of the mirror’s X- and Y-axis. The MEMS controller offers both USB and wireless streaming capability and
we utilize a library of functions on a host computer for creating content and controlling the mirror. Integration with
smart phone applications is demonstrated, utilizing the mobile device both for content generation based on various
messages or data, and for content streaming to the MEMS controller via Bluetooth interface. The display unit is highly
resistant to vibrations and shock, and requires only ~1.5W to operate, even with content readable in sunlit outdoor
conditions. The low power requirement is in part due to a vector graphics approach, allowing the efficient use of laser
power, and also due to the use of a single, relatively high efficiency laser and simple optics.
A highly compact and low power consuming Q-switch module was developed based on a fast single-axis MEMS mirror, for use in eye-safe battery-powered laser range finders The module’s 1.6mm x 1.6mm mirror has <99% reflectance at 1535nm wavelength and can achieve mechanical angle slew rates of over 500 rad/sec when switching the Er/Yb:Glass lasing cavity from pumping to lasing state. The design targeted higher efficiency, smaller size, and lower cost than the traditional Electro-Optical Q-Switch. Because pulse-on-demand capability is required, resonant mirrors cannot be used to achieve the needed performance. Instead, a fast point-to-point analog single-axis tilt actuator was designed with a custom-coated high reflectance (HR) mirror to withstand the high intra-cavity laser fluence levels. The mirror is bonded on top of the MEMS actuator in final assembly. A compact MEMS controller was further implemented with the capability of autonomous on-demand operation based on user-provided digital trigger. The controller is designed to receive an external 3V power supply and a digital trigger and it consumes ~90mW during the short switching cycle and ~10mW in standby mode. Module prototypes were tested in a laser cavity and demonstrated high quality laser pulses with duration of ~20ns and energy of over 3mJ.
The SEMATECH High-NA Actinic Reticle review Project (SHARP) is a synchrotron-based, EUV-wavelength microscope, dedicated to photomask imaging, now being commissioned at Lawrence Berkeley National Laboratory. In terms of throughput, resolution, coherence control, stability and ease of use, SHARP represents a significant advance over its predecessor, the SEMATECH Berkeley Actinic Inspection Tool (AIT), which was decommissioned in September 2012. SHARP utilizes several advanced technologies to achieve its design goals: including the first Fouriersynthesis illuminator on a zoneplate microscope, EUV MEMS mirrors, and high-efficiency freestanding zoneplate lenses
with numerical aperture values up to 0.625 (4×). In its first week of operation, SHARP demonstrated approximately 150 times higher light throughput than AIT and a spatial resolution down to 55-nm half-pitch with 0.42 4×NA (i.e. the smallest feature size on our test mask.) This paper describes the current status of the tool commissioning and the performance metrics available at this early stage.
This work aims to advance 3D position input and motion sensing in a variety of human-machine interface (HMI) and
industrial robotics systems with a MEMS-mirror based optical 3D tracking approach which we termed "MEMSEye."
The goal is to enable real time interaction with computers and robotics in ways that are more intuitive, precise and
natural. Objects can be tracked which are marked either by light sources (e.g. a near-IR LED,) corner-cube retroreflectors
(CCRs,) or with retro-reflective tape. Each "MEMSEye" unit can track the object with high speed and
determine with high precision the azimuth and elevation (θX and θY) angles of the line between the unit and the object.
When two or more such units are utilized to triangulate the object, relative position can be fully determined since
distance information can also be obtained. This final XYZ position information down to sub-millimeter precision can be
obtained in relatively large volumes at update rates of >20 kHz. A demonstration system capable of tracking full-speed
human hand motion provides position information at up to 4m distance with 13-bit precision and repeatability. In another
demonstration, a vector in free space is marked by two target CCRs and the MEMSEye system measures its orientation
in space with ~0.1° precision by locating both CCRs in a time-multiplexed manner.
An important problem in silicon micromachining involves fabrication of suspended structures on a chip for both thermal as well as electromagnetic isolations. The problem becomes more pronounced as the size of a suspended structure increase. A new technique to remove the silicon from beneath a large structure, by micromachining for making a suspended microstructure both for thermal, as well as electromagnetic isolations on a CMOS chip is reported. Conventional methods require two steps; front-side etching, isotropic step, followed by an anisotropic etching step. An alternative technique is based on the backside etching process which requires extra masks and processing steps. In order to keep the post-processing steps to a minimum, a simple technique has been developed that exploits the front-side anisotropic etching to a create both under-cuttings as well as deep etching on one single step. This method uses the gate oxide and polysilicon layer in CMOS technology as the sacrificial layer for initiating the under-cutting needed to make a free standing microstructure. The micro-suspension thickness, width and length of 2 micrometers , 150 micrometers X micrometers , respectively, are made out of LPCVD oxide and have been fabricated.
SC236: Polysilicon Surface Micromachine Technology and Devices
This course is designed to introduce newcomers to micromachining technology and concepts as well as those with a basic familiarity with integrated circuit manufacturing technology about the emerging field of Micro Electro Mechanical Systems (MEMS). Both manufacturing technologies for these devices and examples of sensor and actuator devices will be presented. The course focuses on polysilicon surface micromachining, but will also include a brief overview of other MEMS manufacturing technologies and devices.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.