Unleashed Microactuators electrostatic wireless actuation for probe-based data storage

Summary A hierarchical overview of the currently available data storage systems for desktop computer systems can be visualised as a pyramid in which the height represents both the price per bit and the access rate. The width of the pyramid represents the capacity of the medium. At the bottom slow, cost effective mass data storage media (tape) and at the top fast, expensive media for storage of small amounts of data (FLASH, MRAM) can be found, while the middle section is the domain of the hard disk. Although hard disk development is rapidly reaching the limits of the current design, none of the current storage media can take its place. A new type of data storage technology, called probe storage or probe recording, is currently under development. This technology uses the parallelism of a large array of read/write heads in its advantage. The probe array is positioned above a data storage medium and small movements of either the array or the medium position the probes for reading and writing. This technology would ultimately bring forth a system that stores data at the densities exceeding that of hard disks, but with the form factor of FLASH. The capacity of such a system is limited by the size of the probe array. Because the realisation of a probe array of several square centimeters would be difficult and too expensive, such a system can not substitute a hard disk. In this work, the first small step is taken towards an alternative that decouples the array size from the size of the storage area, so that a large storage area (tile) of several tens of cm2 can be realised. This opens up the possibility to create probe recording systems with the capacity of a hard disk. In the approach, the probe array is divided into small sub-arrays that are distributed over the surface on which data is stored. The need for large actuators to move the probe array is eliminated and instead, each sub-array is equipped with a wirelessly powered, 2-DOF actuator, fabricated by silicon micromachining, enabling it to move freely over the tile like a miniature storage robot or StoBot. Probe based data storage without a fixed positioning system results in a more effective use of surface area and opens up new degrees of freedom in data handling. The StoBot architecture is fully scalable in size and the number of active StoBots on the tile, as well as the number of probes per StoBot. In case one StoBot 200 Summary ceases to function, others can take its place and ensure the functionality of the system. Intelligent file systems can be designed in which the number of active StoBots is adapted to the system load and power requirements. The first milestone on the way towards such a system is to develop a propulsion system for a StoBot. The propulsion system requires a microactuator that can be operated wirelessly and is able to move in arbitrary directions over the tile. In this research, the suitability of wireless and steerable Scratch Drive Actuators (SDA) and a 2-DOF version of the Image Charge Stepping Actuator (ICSA) is investigated. Both these actuators operate by electrostatic forces and although the principles of operation are very different, both actuators can be wirelessly driven and operated in 2D motion. Scratch Drive Actuators (SDA) This type of actuator is shaped like a capital L, rotated 90° clockwise. The long side of the L (plate) can be electrostatically attracted, and bent. This bending makes the end of the short side of the L (bushing) move forward. When the attraction stops the long side relaxes and the resulting movement is converted into forward motion. A qualitative model for the SDA has been developed, based on the cyclic motion of the actuator plate. If the actuation part of its cycle would be equal and opposite to the relaxation part, the SDA would be unable to output power. Therefore, it should go through a hysteresis loop. Two sources of hysteresis in the actuation cycle of the SDA have been identified. The first is based on the influence of the friction force of the bushing on the curvature of the plate. Because the this friction force always opposite to the direction of movement, the curvature of the plate during actuation therefore differs from the curvature during relaxation. The second springs from plastic deformation of the bushing under load. Due to the high pressures involved, the surface of the bushing deforms, thereby raising the coefficient of friction. The behaviour of this friction coefficient may introduce hysteresis. Although several parameters of the model can, within the scope of this thesis, not be determined, the model provides a new view on how an SDA moves. Several versions of the SDA, including wireless designs, were realised. Experiments with these devices showed that an SDA is subject to wear, which confirms the assumption that the bushing plastically deforms under load. A small investigation into SDA wear showed that the SDA, and especially the bushing, wears down rapidly, which led to the conclusion that the SDA is better suited for short-term use than for long-term use. This aspect makes it a less suitable candidate for StoBot propulsion. Summary 201 Image Charge Stepping Actuators (ICSA) The propulsion system of this type of actuator is based on the forces that electric fields exert on (semi) permanent charges in a poor conductor. The actuator consists of a slider and a stator. The slider might be just a piece of poorly conducting material or a complete StoBot. The stator is equipped with interdigital three-phase electrodes, on which the slider is placed. First, a static driving voltage charges the slider, then a dynamic driving voltage is applied to the stator to make the actuator move. Where the three-phase stator supports only 1D driving, a nine-phase stator was constructed to enable 2D driving. A model has been developed, which allows the calculation of propulsive and lift forces working on the slider. Together with this model, a macro version of the ICSA with an electrode pitch of 0.4mm, based on printed circuit board technology, has been designed and realised. The realisation included a control system and a linear high voltage amplifier. Measurements with this actuator showed that its stepping behaviour has excellent repeatability and a top speed of 1 m/s at 2.5 kHz driving frequency. The use of this scale model showed that several optimisations can be performed to improve the performance of this device. According to the theory, miniaturisation of the actuator would be beneficial, because with decreasing size, the strength of the electric field of the macro version can be realised at lower voltages. Miniaturised designs of the ICSA, based on surface micromachining and with a minimum electrode pitch of 4 μm, were designed and realised. These devices did not exhibit the expected forward movement and several explanations for this phenomenon can be given. These are an incorrect resistance of the device material, stiction effects, shielding of the electric fields by the material underneath the electrodes or an electrode to slider distance that is too small. Wafer to wafer transfer of microactuators An important aspect of the production of a future StoBot drive will be the integration of the recording medium and the StoBots. Because of cost effectiveness and because the fabrication processes may not be compatible, the StoBots and the recording medium will be fabricated separately. To meet this demand, a method has been developed to transport wireless microactuators from the wafer of origin (source) to another surface (target) without the use of sophisticated equipment. This method supports the transfer of groups of actuators as well as individual actuators. The technique requires that the actuators are fabricated up side down, fully released and connected to the source wafer by springs. The target surface should be equipped with electrodes to generate electrostatic clamping force. A die with finished actuators is picked up, turned up side down and brought into contact 202 Summary with the target wafer. At that moment, the actuators can be clamped by electrostatic force. When the source die is retracted, the clamped actuators remain on the target surface, making the suspending springs break at a predefined point. The use of electrostatic forces makes the transfer method independent of the type of actuator. Conductive actuators can be clamped by either a DC voltage between source and target wafer or by an AC voltage on the electrodes of the target surface. To clamp an non-conductive actuator, a DC voltage on the electrodes can be used. Experiments showed out that both the SDA and the ICSA can be transferred. The SDA’s could not be tested, but an optical microscopic inspection and inspection by SEM revealed no damaged. Optical inspection and experiments with transferred ICSA’s showed that these devices were undamaged and capable of movement. Outlook With the development of a suitable actuator and a method to transfer this actuator from one wafer to another, the first small steps are made on the road towards a StoBot drive. Future research on StoBots will be directed at reading and writing of data, the communication of StoBots with the rest of the drive and the integration of a small array of simple probes on a wireless actuator such as the ICSA. The flat, almost featureless shape of the miniaturised ICSA should facilitate the integration of simple electronics.