MICROMACHINED TORSIONAL MIRROR UNIT FOR OPTICAL SWITCHING AND FABRICATION METHOD THEREFOR
Field Of Invention
The invention relates generally to switching in optical communication. More particularly, the invention relates to optical switches for switching optical signals.
Background An advancing fiber optic technological field has contributed to the increasing number of applications of optical communication using optical fibers in different technologies. With the increased utilization of optical fibers, there is a need for efficient peripheral devices that assist in the transmission of data through these optical fibers, such as optical switches. An optical switch operates to selectively couple an optical fiber to one of two or more alternative optical fibers such that the two coupled optical fibers are in communication with each other.
The coupling of the optical fibers performed by an optical switch can be performed through various methods. One method of interest includes using a mirror that is placed in front of an input optical fiber to reflect optical signals from the input optical fiber to at least one of two output optical fibers. The input and output optical fibers may be either unidirectional or bi-directional fibers. In the simplest implementation of the mirror method, the input optical fiber is aligned with one of two output optical fibers, such that when the mirror is not placed in an optical path between these two aligned optical fibers, the two aligned optical fibers are in a communicating state. However, when the mirror is placed between the two aligned optical fibers, the mirror steers, i.e., reflects, optical signals from the input optical fiber to the other output optical fiber. The positioning of the mirror in and out of the optical path between the two aligned optical fibers can be accomplished by using an apparatus that mechanically moves the mirror to a desired position.
An alternative method includes using electronic cores in optical switching, which is prevalent applied in telecommunication networks using optical fiber technology in which
carriers have optical inputs and outputs but use an electronic switching matrix between the optical inputs and outputs. In such optical-to-electrical-to-optical (O-E-O) switching, optical signals are converted to electrical signals by photo-detectors. Electronic circuits in the switching matrix then direct the electrical signals to the desired outputs, and final electrical-tό-optical conversion is performed by laser diodes for onward transmission in the optical fiber network. The O-E-O switches induce substantial time-delay and equipment cost, resulting in a limitation in capacity and upgrading of the optical fiber network.
Optical switches or crosscormectors provide an all-optical link between local networks, regional networks, metro networks and Internet service providers. Using all-optical cores can help to reduce costs by eliminating a range of extra parts and interfaces that are now required to convert optical signals to electronic signals for optical switching purposes. Also, all-optical cores can allow carriers to be easily upgraded to the higher data rates needed for the growing demand for bandwidth.
To this end, numerous conventional optical switches have been proposed, and a number of patented optical switches and related technologies are described hereinafter.
U.S. Patent No. 5,208,880 to Riza et al. describes an optical switch that utilizes a piezoelectric actuator to displace a mirror to selectively couple an input optical fiber to a particular output optical signal. The piezoelectric actuator includes a number of piezoelectric bars to linearly displace the mirror. In a first embodiment, the optical switch includes N output optical fibers that are positioned perpendicularly to an input optical fiber in a side-by-side configuration. The mirror is positioned on the axis of the input optical fiber and has a reflective surface that is orientated to direct optical signals from the input optical fiber at a right angle. The mirror is coupled to the piezoelectric actuator that is able to displace the mirror along the axis of the input optical fiber to couple the input optical fiber to any one of the output optical fibers. In a second embodiment, the optical switch is configured to accommodate two input optical fibers and two output optical fibers. The optical fibers are positioned in an "X" configuration such that two output optical fibers are located in the upper portion of the configuration and the two input optical fibers are located in the lower portion of the configuration. In this embodiment, the
optical switch includes a thin mirror that has reflective surfaces on both sides. The mirror can be positioned in the optical paths between the optical fibers by the piezoelectric actuator such that when the mirror is displaced to the center of the "X" configuration, the lower left optical fiber is coupled to the upper left optical fiber and the lower right optical fiber is coupled to the upper right optical fiber (the "reflective state"). However, when the mirror is removed from the optical paths, the lower left optical fiber is coupled to the upper right optical fiber and the lower right optical fiber is coupled to the upper left optical fiber (the "passive state").
U.S. Patent No. 5,042,889 to Benzoni describes an optical switch that also uses a mirror to switch optical paths between optical fibers. In an exemplary embodiment, the optical switch is configured to accommodate four optical fibers that are positioned in an "X" configuration. In contrast to the optical switch of Riza et al., the optical switch of Benzoni utilizes an electromagnetic mechamsm, instead of a piezoelectric actuator, to move the mirror in and out of the optical paths between the optical fibers. The electromagnetic mechanism operates to create an attractive magnetic force between the mechanism and the mirror. The upper section of the mirror includes a ferromagnetic material that becomes attracted to the electromagnetic mechanism when the magnetic force is generated. The electromagnetic mechanism is located above the mirror to lift the mirror when the mechanism is activated. Initially, the mirror is positioned between the optical paths such that the four optical fibers are coupled in the reflective state. When the electromagnetic mechanism is activated, the attractive magnetic force causes the mirror to be lifted out of the optical paths to set the optical fibers in the passive state.
U.S. Patent No. 6,215,222 to Hoen describes an optical switch for steering optical beams which utilizes a surface electrostatic actuator to mechanically pivot a mirror to selectively redirect a received optical beam to a predetermined direction. In a preferred application, the optical switch can optically couple a number of first optical fibers to a number of second optical fibers in one of many configurations. The electrostatic actuator and the mirror form a switching device of the optical switch to redirect optical signals between two optical fibers such that the two optical fibers are in communication. In an exemplary embodiment, the optical switch includes sixteen switching devices in a 4X4 multiple- input-multiple-output .arrangement. The optical switch is connected to a first set of four
optical fibers that are positioned on the first side of the optical switch. The optical switch is also connected to a second set of four optical fibers that are positioned on a second side of the optical switch. The optical coupling of the optical fibers is accomplished by pivoting mirrors from non-reflective orientations to reflective orientations. A non- reflective orientation is the position of a mirror in which the reflecting surface of the mirror is generally parallel to the upper surface of the optical switch. A reflective orientation is the position of the mirror in which the reflecting surface of the mirror is perpendicular to the upper surface of the optical switch.
Although the foregoing conventional optical switches operate well for their intended purposes, these optical switches are not sufficiently compact for use in space-constraint applications, particularly in multiple-input-multiple-output switching arrangements.
To this end, micromachined mirrors or micromirrors can be used in a micro-optical switch to turn optical signals into desired directions through steering or redirecting. A micro- optical switch changes the routing of optical signals in fiber optic telecommunication systems. To meet the demand for optical communications, many telecommunication providers are looking for miniature optical switching systems, among which the micromachined mirror approach is typically adopted.
Micromirror structures are typically fabricated using microelectromechanical system (MEMS) technology and may be categorized into two classes in term of fabrication method. One class relates to surface-micromachined mirrors where in most of such cases, a micromirror is originally made from planar bulk which is parallel with a substrate and can be folded in some applications. The other class of micromirrors relates to bulk- micromachined mirrors in which the reflective surfaces are made perpendicular to the substrate.
Conventional methods of steering a mirror include the utilization of electrostatic force, magnetic force and thermal expansion. When using magnetic force it is most likely that external elements may have to be integrated and metal parts to be utilized in association with the mirror body. In the thermal expansion approach, relatively long response time is always required for temperature gain and dissipation, which may limit the switching
speed. Actuation by electrostatic force may be considered as the most appropriate way to steer a micromirror if compact design and switching time are required.
Electrostatic force is typically generated in MEMS devices by the application of direct- current (DC) voltage on opposite electrodes which cause the electrodes to repel each other. Comb-like structures are most popularly adopted owing to a large number of electrode pairs each comb-like structure can comprise and the small distance between the opposite electrodes, which may provide a relatively large electrostatic force.
Relatively modern technology enables MEMS devices to be fabricated on semiconductor substrates, typically silicon substrates. These microelectromechanical systems typically have sizes in the order of microns and may be integrated with other electrical circuits on a common substrate.
Conventional MEMS-based optical switches can operate in the plane of the substrate or normal to the substrate. An example of an in-plane optical switch using a vertical mirror is disclosed in C. Marxer et al., "Vertical Mirrors Fabricated By Reactive Ion Etching For Fiber Optical Switching Applications," Proceedings IEEE, The Tenth Annual International Workshop on MicroElectoMechanical Systems, An Investigation of Micro Structures, Sensors, Acuators, Machines and Robots (Cat. No. 97CH46021), IEEE 1997, pp.49-54. The optical switch includes a metal-coated silicon micromirror coupled to a dual comb drive actuator. The two comb actuators work in opposite directions to push and pull the micromirror into and out of, respectively, an optical path between optical fibers. The optical switch is fabricated in a single step using inductively coupled plasma-etching technology with a sidewall passivation technique.
U.S. Patent. No. 6,229,640 to Zhang describes a MEMS-based optical switch having improved characteristics. An optical switch includes a single comb drive actuator including a stationary comb mounted on a substrate, a movable comb interleaved with the stationary comb, and a beam structure connected between the substrate and the movable comb and a micromirror coupled to the actuator. The optical switch further includes a pair of first waveguide channels and a pair of second waveguide channels disposed on the substrate. The micromirror is capable of being moved between an extended position
interposed between the waveguide channels and a retracted position apart from the waveguide channels. The two combs apply a force capable of deflecting the beam structure and moving the micromirror to one of the extended positions or the retracted position and the beam structure returns the micromirror to the other of the extended position or the retracted position in the absence of the application of force between the two combs. Zhang also describes methods of forming the micromirror and combs for the comb drive actuator on substrates.
Although the foregoing conventional MEMS-based optical switches operate well for their intended purposes, the optical switches are also not designed for use in multiple-input- multiple-output switching arrangements.
The significant growth of optical fiber networks due to a need for fast broadband fiber communications has created a large demand for low-cost fiber optical switches. The market size for optical switches in 1995 was estimated to be $33 million, with an annual growth of about 40%. This estimation is consistent with a survey conducted by the Communications Industry Researchers (CIR) Inc. in 1999. According to the survey by CIR Inc., long-distance carriers are about to give their networks an optical overhaul, driven by "an Internet-soaked communications environment" in which bandwidth demands double every few months. The survey predicts that the total worldwide market for optical switches and cross-connectors are expected to increase from $234 million in 2000 to $7.4 billion by 2004.
Currently, most of the optical switches on the market are O-E-O switches. However, it is believed that optical switches with all optical-cores can gradually replace the current O-E- O switches in the near future, and rapid advances and potential benefits of MEMS-based optical switches can create a market of more than $1 billion by 2004. Hence, a large number of companies, universities and start-ups are working on MEMS-based optical switches, leading to the rapid development of the technology.
Manufacturing cost, performance in light switching, switching time and loss, and reliability of a miniature MEMS optical switch are typical concerns in telecommunication applications. It is therefore desirable to develop an approach by which MEMS optical
switches can be fabricated with high yield, good reliability, good switching performance and compatibility of post-packaging and assembly processes.
A major challenge in optical switch technology is to develop new all-optical core switching systems. MEMS technology has been applied to implement compact all-optical core switches. Such MEMS-based optical switches provide a number of advantages, namely high switch contrast, low insertion loss, small cross talk, miniaturization, and potentially low cost, resulting in wide applications in optical networks.
Accordingly, there is a need for compact MEMS-based optical switches which are configurable for multiple-input-multiple-output switching operations and having micromirrors which are actuable by electrostatic forces.
Summary The invention relates to a structure and the fabrication method of a MEMS-based torsional micromirror which can be used in optical switches and large scale crosscormectors, which are one of the essential components or subsystems in an optical fiber telecommunication system.
A micro-mirror which can be steered within an angular range is a suitable element for redirecting optical signals through different angles and therefore provides more flexibility in performing the function of optical switching, in comparison with conventional shutterlike mirrors that have reflective and non-reflective statuses. Especially when a large-scale crossconnector is required, the use of an optical switch based on steering mirrors can greatly reduce the number of mirrors needed and therefore, reduce the device dimension, power dissipation and manufacturing cost. Conventional steering mirrors are designed and fabricated through the surface-micromachining method, which needs extra process to fold up the mirror that may introduce uncertainties in terms of yield and cost.
Therefore in accordance with one aspect of the invention, an optical switch formed on a substrate is disclosed. The optical switch comprises a mirror disposed substantially upright in relation to a substrate, the mirror being rotatably displaceable about a pivot between a first position and a position angularly displaced from the first position. The
optical switch also comprises a comb drive actuator, which includes a stationary comb anchored to the substrate in a position substantially parallel to the plane of the substrate, and a displaceable comb interleaved with the stationary comb and coupled to the mirror in a position substantially coplanar with the stationary comb and substantially perpendicular to the plane of the mirror. In the optical switch, the stationary and displaceable combs are capable of applying a force for rotatably displacing the mirror to one of the first position and position angularly displaced from the first position.
In accordance with another aspect of the invention, a method for fabricating an optical switch having a plurality of structures with differing heights and at least one of the plurality of structures being a floating structure from a material having an embedded sacrificial layer separating the material into upper and lower portions is disclosed. The method comprising the steps of providing a patterned resist layer for forming a structure having a first height referenced by the upper surface of the lower portion of the material, and using two-phase etching in which a first phase etching provides a height difference between the structure having the first height and another of the plurality of structures having a second height which is lower than the structure having the first height and a second phase etching exposes the embedded sacrificial layer. The method further comprises the step of releasing a portion of the sacrificial layer exposed by the second phase etching for forming the floating structure.
Brief Description Of Drawings
Embodiments of the invention are described in detail with reference to the drawings, in which:
Fig. 1 is a plan view of an optical switch having a single switching device in a 1 xN switching configuration;
Fig. 2 is a plan view of an optical switch having 2xN switching devices in an NxN switching configuration;
Fig. 3 is a plan view of the switching device of Figs. 1 and 2 according to a preferred embodiment of the invention;
Fig. 4 is a perspective view of a rotatable micromirror having a comb structure used in a switching device according to a second embodiment of the invention;
Fig. 5 is a perspective view of a rotatable micromirror having a comb structure and an anti -vibration structure used in a switching device according to a third embodiment of the invention;
Fig. 6 is a perspective view of a positioning structure having a positioning arm and multiple locking elements used in a switching device according to a fourth embodiment of the invention;
Fig. 7 is a perspective view of a locking element of Fig. 6; and
Figs. 8 A to 8C illustrate a process by which the structures of the switching device of Fig. 3 are fabricated according to a fifth embodiment of the invention.
Detailed Description
Embodiments of the invention are described hereinafter for addressing the need for compact MEMS-based optical switches which are configurable for multiple-input- multiple-output switching operations and having micromirrors which are actuable by electrostatic forces.
Accordingly, a MEMS-based optical switch, hereinafter more specifically known as a micromachined torsional mirror unit, is hereinafter described. Each optical unit in the MEMS-based optical switch is known as a switching device, which is capable of turning light path into multiple predetermined output ports by virtue of electrostatic force generated within the optical unit under a DC voltage loading. The switching device consists of different functional components, including a rotatable reflection subunit, an actuation subunit, and a self-latching subunit. All the parts are preferably unitarily fabricated from one substrate, such as a crystal silicon wafer, using the MEMS technology. A method for fabricating such a MEMS-based optical switch is also hereinafter described using bulk micromachining technology.
There are a number of advantages associated with embodiments of the invention. The switching device is specifically designed for use in all-optical switching products for the next generation of optical networks. From network protection and restoration to dynamic connection provisioning, an all-optical switching system can serve the entire spectrum of needs for the present, and can be scaled seamlessly for the future.
For example, small 1 xN and NxN optical switching arrays are typically used for fast network protection and restoration. Some examples of the network architectures that utilize optical add-drop multiplexers include linear add-drop for backbone dense wavelength division multiplexing (DWDM) networks, hub-rings in metro access networks, and a logical mesh ring that allows dynamic path reconfiguration based on the capacity demand of a network.
As a further example, large port-count optical crosscormectors are used in central offices for dynamic remote network provisioning, which allows the service providers to offer innovative high-bandwidth service level agreements.
The switching device and the subunits based on the technology described hereinafter advantageously cover a full range of optical switching requirements for network system and service providers.
One application of the switching devices is to steer or redirect light in optical switches using the rotatable reflection subunits for reflecting light by pre-defined angles. In the case of a l N fiber optical switch, there is only one switching device as shown in Fig.l . Fig. 1 is a plan view of an optical switch 102 having a single switching device 104 in a lxN switching configuration. The optical switch 102 is connected to a single input optical fiber 106 and multiple output optical fibers 108 in a one-to-many unidirectional optical transmission. Light from the input optical fiber 106 entering the first output optical fiber 110 when the switching device 104 is at rest at a first position 104' is redirected to the last output optical fiber 112 when the switching device 104 rotates through an angle to rest at another position 104".
An NxN fiber optical crossconnecting switch can also be constructed using the switching device by lining up two arrays of switching devices with N switching devices in each array. In this application, two switching devices from the opposite arrays always form a pair of switching devices which has identical reflection angles for directing input and output rays in parallel as shown in Fig. 2. Fig. 2 is a plan view of an optical switch 202 having 2xN switching devices 204 to 214 in an NxN switching configuration. In operation, the optical switch 202 is connected to N input optical fibers 216, 220 and others, and N output optical fibers 218, 222 and others in a many-to-many unidirectional optical transmission. Light from the first input optical fiber 216 entering the optical switch 202 can be passed through to the first output optical fiber 218 when the switching devices 204 and 206 remain in their biased positions in which mirrors in the switching devices are not rotated to obstruct the path of light from the first input optical fiber 216. Light, however, from the last input optical fiber 220 entering the optical switch 202 can be redirected by the last switching device 212 in the input array to the second switching device 210 in the output array, which in turns redirect light into the second output optical fiber 222.
In relation to the fabrication method for each of the switching device in the optical switches using bulk micromachining technology, the advantages are also manifold. Firstly, using bulk micromachining, it is easy to obtain structures in the switching device with large thickness. With bulk micromachining, it is also possible to use single-crystal silicon as a substrate upon which the switching device is based, which possesses better mechanical characteristics than poly-crystal silicon.
More importantly, the fabrication method for the switching device produces better geometric control of floating or levitated structures in the switching device due to the application of the deep-reactive-ion-etching (DRIE) process, which in comparison with the an-isotropic wet etching process produces better results. In the fabrication method, a two- phase DRIE process is applied for generating structures with significant differences in thickness. Also, the fabrication method requires fewer lithography processes than conventional fabrication methods, which simplifies the fabrication of the switching device resulting in the reduction of cost and increase of yield. Furthermore, by using the fabrication method there is good electrical isolation among electrode structures in the
switching device due to the use of Silicon-On-Insulator (SOI) wafer for fabrication.
Additionally, the floating or levitated structures in the switching device are easily released because of the use of a buried sacrificial layer in the SOI wafer. Such a sacrificial layer also provides an auto-stop function for the DRIE process. As a consequence, the fabrication method provides self-alignment of a metallization process for forming electrode structures, and a releasing process for forming floating or levitated structures, which simplifies the fabrication method in terms of fewer lithography and patterning steps.
With reference to Fig. 3, the switching device is described in greater details. Fig. 3 is a plan view of a switching device 302 according to a preferred embodiment of the invention. The switching device 302, which is preferably made from Silicon-On-Insulator (SOI) wafer, is preferably based on a substrate 303 which is silicon-based, and consists of a rotatable reflection subunit, an actuation subunit, and a self-latching subunit. The switching device 302 has a circular structure and there are supplementary structures such as electrical connections and bonding pads (not shown) positioned around the switching device 302 for facilitating the application of the DC voltage. The reflective surfaces of the switching device are preferably made perpendicular to the substrate 303. The rotatable reflection subunit is preferably fabricated through the deep-reactive-ion-etching (DRIE) technology. It comprises two parts with different step heights formed by multiple DRIE steps. The lower part, which is typically closer to the substrate, comprises the actuation unit and self-latching unit. The upper part of the structure comprises the rotatable reflective subunit.
The rotatable reflection subunit consists of a mirror 304 which has reflective surfaces on both sides and is levitated above the substrate 303 and rotatable about a pivot which is preferably located at the center of the circular switching device 302. The mirror 304 is preferably held perpendicular to the substrate 303 and therefore rotates through an axis which is also perpendicular to the substrate 303 and collinear with the pivot. The mirror 304 is naturally positioned at a biased position as shown in Fig. 3, and can be rotated through a pre-defined angle to rest at another position through the operation of the actuation subunit.
The actuation subunit or comb drive actuator consists of a stationary comb which includes stationary comb fingers 306, 308 and others, and a displaceable comb which includes displaceable comb fingers 310, 312, and others. In the preferred embodiment, there are four stationary combs and the fingers of these stationary combs circumcircle the pivot and are anchored to the substrate 303 at comb anchors 314 to 320. Correspondingly, there are also four displaceable combs and the fingers of these displaceable combs oppose and interleave the fingers of the four stationary combs. The four displaceable combs also circumcircle the pivot and extend from the mirror 304, two of the four displaceable combs outwardly extending from opposite sides of the mirror 304 on each side of the pivot. The four stationary combs and four displaceable combs are preferably levitated above the substrate 303 on a plane preferably parallel with the plane of the substrate 303. The electrostatic interaction between the four stationary combs and four displaceable combs when a DC voltage is applied across both the comb anchors 314 and 320 and the mirror 304 or alternatively across both the comb anchors 316 and 318 and the mirror 304 provides electrostatic forces on the displaceable combs large enough to rotate the mirror 304 clockwise or anti-clockwise depending on where the DC voltage is applied.
The comb drive actuator also consists of a biasing element or tethers 322 that extend from both sides of the mirror 304, the portions of the tethers 322 immediately extending from the mirror 304 being straight and perpendicular to the length of the mirror 304 in its initial position. After the straight and perpendicular portions, the tethers 322 convolve in a geometric manner forming folds until the extremities of the tethers 322 distal from the mirror 304 join with tether anchors 324 and 326. The tethers 322 are levitated above the substrate 303 while each of the tether anchors 324 and 326 is anchored to the substrate 303 on each side of the mirror 304. The tether anchor 324 is positioned between but spaced apart from the comb anchors 314 and 316, while the tether anchor 326 is positioned between but spaced apart from the comb anchors 318 and 320. When the DC voltage is applied to the comb drive actuator to rotate the mirror 304 clockwise, the DC voltage is preferably applied across any of the tether anchors 324 and 326 and both the comb anchors 314 and 320 to rotate the mirror 304. To rotate the mirror 304 anti-clockwise, the DC voltage is preferably applied across any of the tether anchors 324 and 326 and both the comb anchors 316 and 318 to rotate the mirror 304 anti-clockwise. The DC voltage is preferably applied to the bonding pads through the electrical connections to the tether and
comb anchors.
The tethers 322 and tether anchors 324 and 326 collectively form a biasing structure for biasing the mirror 304 to rotate about the pivot which is located at the center of the circular switching device 302. Due to the resilient construction of the tethers 322, the mirror 304 experiences a torsional force applied by the straight and perpendicular portions of the tethers 322 and the subsequent folds extending from the mirror 304 whenever the DC voltage is applied. When the electrostatic forces resulting from the application of the DC voltage is larger than these torsional forces, the mirror 304 is rotated from its initial position. To keep the mirror 304 in any position other than its initial position, the electrostatic forces have to be larger than resistance relating to the engagement of the mirror 304 into locking elements of a mirror positioning subunit. These torsional forces in the absence of the electrostatic forces when the DC voltage is removed are reciprocated by resistance provided by the locking elements thereby positioning the mirror 304 in positions corresponding to the locking elements engaged by the mirror 304. To return the mirror 304 to its initial position, the DC voltage is applied to rotate the mirror 304 in the opposite direction until the mirror 304 engages the corresponding locking elements for the initial position.
The switching device 302 also consists of the mirror positioning subunit and a stabilizing subunit. The mirror positioning subunit consists of positioning arms or triple folded cantilevered beams 328 and locking elements or bumps 330. The stabilizing subunit consists of a single stabilizing finger 332 and a pair of stabilizing fingers 334. Both the mirror positioning and stabilizing subunits are anchored to the substrate 303 through a pair of main anchors 336, each of the pair of main anchors 336 being anchored to the substrate 303 on each side of the mirror 304. Each main anchor 336 is also positioned adjacent to the respective comb anchors and tether anchor on each side of the mirror 304 but further away from the mirror 304. Due to this arrangement, the stabilizing fingers of the stabilizing subunit circumcircle the stationary and displaceable combs, while the positioning arms 328 circumcircle the stabilizing, subunit and the stationary and displaceable combs. Both the mirror positioning and stabilizing subunits are also preferably levitated above the substrate 303 on a plane preferably parallel with the plane of the substrate 303.
Operational details relating to the actuation of the mirror by the comb drive actuator are described with reference to Fig. 4, while further details of the mirror positioning subunit are described with reference to Fig. 5. Further details of the stabilizing subunit are described with reference to Figs. 6 and 7. Figs. 4 to 7 are illustrative of switching devices according to further embodiments of the invention but with simpler structures, .and therefore are more effective as illustrations for describing the respective operations and subunits.
The rotatable reflection subunit or mirror is driven by the comb drive actuator which generates repulsive electrostatic force between interleaving fingers of the stationary and displaceable combs which due to metallization form opposite electrodes when a DC voltage is applied across the respective stationary combs and the mirror, as shown in Fig. 4. Fig. 4 is a perspective view of a switching device 402 according to a second embodiment of the invention, in which the switching device 402 has a rotatable mirror 404 and comb structures 406 and 408. Fingers 416 of a displaceable comb 406 form movable electrodes which are attached to the mirror 404 so as to steer it circularly owing to the symmetrical comb design about the mirror 404.
A stationary comb 408 consists of fingers 410 and these are coupled to a comb anchor 412 which is anchored to a substrate 414 on which the switching device 402 is based. Fingers 416 of the displaceable comb 406 extend away from the mirror 404 while fingers of the stationary comb 408 extend from the comb anchor 412 toward the mirror 404. The fingers 410 of the stationary comb 408 are interleaved with the fingers 416 of the displaceable comb 406.
In the switching device 402, adjacent interleaved fingers of the respective stationary 408 and displaceable 406 combs are spaced apart. Gaps or spaces 418 provided between adjacent fingers 410 of the stationary comb 408 ensure that the fingers 416 of the displaceable comb 406 interleaving with the fingers 410 of the stationary comb 408 are not in contact thereby avoiding short-circuiting the DC voltage applied.
Each finger 410 or 416 is an electrode having electrical conductivity due to metallization, and each finger 410 of the stationary comb 408 is electrically connected to the comb anchor 412, which is also electrically conductive due to metallization. Each finger 416 of the displaceable comb 406 is electrically connected to the mirror 404, which is also electrically conductive due to metallization. To actuate the mirror 404, a DC voltage is applied across the stationary and displaceable combs to apply an electrostatic force between the stationary and displaceable combs.
To provide rotational displacement of the mirror 404 when actuated by the application of the DC voltage, each finger 410 or 416 is circularly curved and concentric with respect to each other in relation to a pivot about which the mirror 404 is rotated.
As shown in Fig. 5, gaps provided between guide elements or stabilizing fingers in the stabilizing subunit are narrower than those between fingers in the actuation subunit or comb drive actuator. This is to constrain the displacement of the movable electrodes or displaceable fingers within a safe range and therefore avoid crashing of the stationary and displaceable combs which can lead to short-circuit and structural failure. Fig. 5 is a perspective view of a switching device 502 according to a third embodiment of the invention having a rotatable mirror 504, comb structures 506 and 508, and an anti- vibration structure 510.
The anti- vibration structure or stabilizing subunit 510 includes a pair of stabilizing fingers 512 which extends from a stabilizing anchor 514 which is anchored to a substrate 516 on which the switching device 502 is based. The pair of stabilizing fingers 512 is coupled to the stabilizing anchor 514 and therefore anchored to the substrate 516. The stabilizing subunit 510 also includes a guide element or single stabilizing finger 518 which is coupled to and extends away from the mirror 504. The single stabilizing finger 518 is sandwiched between the pair of stabilizing fingers 512 with little space in between to enable the single stabilizing finger 518 to slide into and out of the pair of stabilizing fingers 512 along the length of the pair of stabilizing fingers 512. This arrangement of the stabilizing unit 510 is for limiting lateral displacement of the mirror 504 to only angular displacement.
To this end, the pair of stabilizing fingers 512 and single stabilizing finger 518 are circularly curved and concentric with respect to each other in relation to a pivot about which the mirror 504 is rotated. Also, the pair of stabilizing fingers 512 and single stabilizing finger 518 are positioned substantially parallel with the plane of the substrate 516. To prevent short-circuit from occurring, the stabilizing subunit 510 is isolated from the stationary comb 508 by a gap 520 so that when the pair of stabilizing fingers and 512 and single stabilizing finger 518 make contact during operation, the comb structures 506 and 508 are not short-circuited via the stabilizing subunit 510.
Circular cantilevered beams are also allocated symmetrically around both sides of the reflective subunit or mirror as shown in Fig.6. The mirror is positioned during steering of the mirror through snap-on latching using indentations provided between pairs of locking elements as shown in Fig.7. The displacement angles of the mirror are pre-determined by the positions of the latching elements which are placed symmetrically about the mirror.
Fig. 6 is a perspective view of a switching device according to a fourth embodiment of the invention having a mirror positioning subumt 602 which consists of a positioning arm 604 and multiple locking elements 606 for positioning a mirror 608, and Fig. 7 is a perspective view of a pair of locking elements 702 and 704.
The switching device (not shown) includes the mirror positioning subunit 602 for locking the mirror 608 in the biased position or other positions which are angularly displaced from the biased position. The mirror positioning subunit 602 includes a cantilevered beam or positioning arm 604 anchored to a substrate (not shown) on which the switching device is based through a positioning anchor 610. The mirror positioning subunit 602 also includes multiple locking elements 606 formed along the positioning arm 604. In the switching device, the positioning arm 604 is positioned for facilitating the engagement of the end of the mirror 608 furthest away from a pivot about which the mirror 608 rotates with a pair of the multiple locking elements 606.
In a preferred implementation, the positioning arm 602 is a triple folded cantilever beam which anchored to the positioning anchor 610. The positioning arm 602 preferably also forms a plane substantially parallel with the plane of the substrate. To facilitate the
engagement of the mirror 608 in the multiple locking elements 606, the positioning arm is circularly curved and the triple folds of the positioning arm are concentric with respect to each fold in relation to a pivot about which the mirror 608 is rotated.
With reference to Figs. 8A to 8C, a process for fabricating a MEMS-based microstructure such as the optical switch described hereinbefore, is described according to a fifth embodiment of the invention. For ease of illustration, the cross sections depicted in Figs. 8 A to 8C correspond to a cross section of an SOI wafer 802 from which a switching device based on a substrate is formed with a vertical floating structure 804 such as the mirror, a lower-height fixed structure 806 such as the comb or any of the main anchors, and higher-height fixed structures 808 such as the tether anchors from which lead-outs extend for providing electrical connectivity.
In this example process, a masking layer 810 which is a resist layer is formed over the wafer 802. The wafer 802 is typically formed from a semiconductor material, such as silicon, and includes a buried insulating layer 812 separating the wafer 802 into an upper portion 814 and a lower portion 816. The buried insulating layer 812 may, for example, be an oxide layer, such as silicon dioxide. The optical device structures are formed in the upper portion 814 of the wafer 802 above the insulating layer 812.
The masking layer 810 is provided to protect portions of the wafer 802 which are of higher height, such as the vertical floating structure 804 and the higher-height fixed structures 808, during subsequent etching and typically has a thickness sufficient to do so. In the illustrated process, the masking layer 810 is formed from, for example, oxide such as silicon dioxide. To obtain the patterns 818 free of the masking layer 810, a first photoresist deposition followed by lithography are applied.
To isolate the structures, a second photoresist deposition is applied after which lithography is performed for providing a pattern for a two-phase DRIE process which forms in the first phase of the two-phase DRIE process a level lower than the upper surface of the upper portion 814 known as a first level 820. The first level 820 subsequently forms gaps, separations, or spaces 824 between structures. Portions of the upper portion 814 exposed
or unprotected by the masking layer 810 but protected by the second photoresist layer form a second level 822 which is flushed with the upper surface of the upper portion 814. The second level 822 in turn forms a resist layer for forming the gaps, separations, or spaces 824, and subsequently forms the lower-height fixed structure 806 or lower-height floating structures such as comb fingers (not shown). The first phase of the two-phase DRIE process is time-controlled for achieving the desired difference between the first .and second levels.
After the first phase of the two-phase DRIE process is performed, the second photoresist layer is removed followed by the start of the second phase of the two-phase DRIE process, which forms the gaps, separations, or spaces 824 and the lower-height structures such as the lower-height fixed structure 806. The second phase of the two-phase DRIE process is stopped when the buried insulating layer 812 is reached or exposed, and thereafter the masking layer 810 is stripped.
To form levitating spaces 825, which is space freed by removing portions of the buried insulating layer 812, so as to float or release the vertical floating structure 804, the buried insulating layer 812, which forms a sacrificial layer, is etched, for example using HF buffer solution. The etching of the buried insulating layer 812 is also time-controlled so that all floating structures are properly released. To ensure that the fixed structures are anchored to the lower portion 816, the widths of areas of the buried insulating layer 812 lying under these fixed structures have to be substantially larger than the widths of areas of the floating structures which are to be released. By doing so, these fixed structures form a resist layer for forming the levitating spaces 825 while keeping the fixed structures anchored to the lower portion 816. Metallization is then performed, for example using aluminum evaporation, to form a metal layer 826 over all the structures.
In the foregoing manner, compact MEMS-based optical switches which are configurable for multiple-input-multiple-output switching operations and having micromirrors which are actuable by electrostatic forces, and fabrication method therefor, are described.
Although only a number of embodiments of the invention are disclosed, it will be apparent
to one skilled in the art in view of this disclosure that numerous changes and/or modification can be made without departing from the scope and spirit of the invention.