WO2002082140A1

WO2002082140A1 - Optical microring resonator, optical multiplexer and optical switching apparatus using deformable waveguide segments

Info

Publication number: WO2002082140A1
Application number: PCT/IL2002/000275
Authority: WO
Inventors: Dan Haronian; Rahav Cohen
Original assignee: Galayor Inc.
Priority date: 2001-04-04
Filing date: 2002-04-02
Publication date: 2002-10-17

Abstract

Micro-Electro-Mechanical System (MEMS) devices are described which perform functions in optical circuits by means of bending or stressing of waveguide sections of the circuit, in the plane of the device substrate. The devices utilize sections of suspended waveguide which are bent or stressed by means of micro-actuators. One example of such a device is a variable ring resonator (14) for use as a variable filter, in which a section of the resonator ring is stretched. Another example is a dispersion correction circuit in which one arm (254) of a split waveguide circuit is stretched to change the optical path difference between the arms. A third example is a bi-stable latching switch, operated by bending of interlocking arms (310, 318) attached to the substrate at only one end.

Description

BENDING AND STRESSING OPTOMECHANICAL DEVICES

FIELD OF THE INVENTION

The present invention relates to the field of the application of Micro-Electro-Mechanical System (MEMS) devices, to perform functions in optical circuits by means of bending or stressing of waveguide sections, and especially as applied to resonator circuits, and dispersion correction circuits and bi-stable latching switches.

BACKGROUND OF THE INVENTION

Many common MEMS applications use the motion generated by the device in order to accomplish switching functions in optical waveguides, such as by mode coupling of the light from one waveguide to the other. Such applications have been described in co-pending PCT patent application No. PCT/ILOl/00787, for "Mode Coupled Optomechanical Devices", to the inventor in the present application, published as International Publication No. WO 02/17004, and in the references quoted therein. This patent document is hereby incorporated by reference in its entirety.

Though the movement of the waveguides taking part in the switching process involves bending of the waveguides, the bending action is not an end in itself, and does not take an operative part in the device operation, but is simply an integral part of the waveguide deformation required to move the waveguides into contact. To the best of the applicant's knowledge, very few applications exist, if at all, in which the bending deformation of the waveguide itself or the strain thus induced by the deformation, are used to accomplish the desired physical operation of the device.

One type of currently available integrated optics variable resonator, whose main use is as a variable wavelength filtering element, is the electro-optically tuned ring resonator. In this device, the index of refraction of the ring resonator, and hence the resonant frequency of the resonator, is changed by application of an applied electric field through the mechanism of an electro-optical affect. However, the resonant frequency is a sensitive function of the refractive index, and the center frequency of the filter is thus likely to drift even from small environmental temperature changes. Consequently, such filters generally have to be temperature stabilized. There therefore exists a serious need for a compact, planar, variable wavelength filter device, which overcomes some of the disadvantages of such prior art filters.

In another area of optical circuit design, it is known that the performance of many optical communication systems is limited by the presence of Polarization Mode Dispersion (PMD) and/or Chromatic Dispersion (CD). PMD occurs when the fast and the slow polarization directions of a wave propagating down a fiber experience different propagation conditions, resulting in different group velocities, which causes broadening of the pulses containing the information. Methods of compensating for PMD have been proposed, such as that described in the article by John Grady in the February 2001 issue of WDM Solutions. The input state-of- polarization (SOP) is rotated to one of its principal axes on the Poincare sphere by splitting the optical signal into two orthogonal polarization components by means of a polarization beam splitter (PBS). The two orthogonal polarization directions can be the TE and TM modes of a linearly polarized beam, or two elliptically polarized beams with their major axes orthogonal. The two polarization components may have a phase difference between them ranging between zero and 2π as a result of the difference of their group velocity. By changing the relative phase and intensity, a principle polarization state results with one group velocity, and information is thus conveyed through the system with a single group velocity. This means that the PMD is reduced to zero. After inducing the required phase and/or intensity shift, the beams are then recombined. In the prior art system described in the above-mentioned publication, the phase difference between the two paths is generated by means of a MEMS micro-mirror, which physically diverts one of the beams along a longer path. A similar arrangement, using a chromatic dispersion element instead of a PBS, can be used to compensate for chromatic dispersion. This prior art method is essentially implemented in a free space configuration, and is thus bulky and not amenable to integration into planar optical circuits. There therefore exists a need for a method of performing such dispersion compensation in an integratable, on-chip, planar configuration.

In another area of optical and microelectronic circuit component design, it is known that bi-stable switches based on MEMS technology are required in many applications for providing mechanically-based switching capabilities between latched switched states, and without the need for a holding voltage. In addition to their added simplicity and flexibility of operation, in that no holding voltage is necessary for their operation, such latched mechanical switches have higher reliability than pure electronic switches in noisy electrical environments, such as in the presence of voltage spikes. Bi-Stable MEMS switching can be used for optical applications as well as RF switching, Micro-relays, cellular communication, testing and other MEMS or MOEMS microelectronic applications.

A number of methods of actuation have been commonly used in the prior art to accomplish mechanical switching, including magnetic and thermal actuation. MEMS based optical switches are known, for example, that use thermal actuators to move optical fibers. Such prior art methods have a number of disadvantages. Thermally actuated switches have a comparatively long switching time, usually require comparatively high operating powers, are generally temperature sensitive, and often require provision of a holding voltage. Magnetic actuation devices, though not generally susceptible to all of the aforementioned disadvantages, can be complicated to manufacture, and the required structures are generally larger than thermally actuated devices. Thermal and magnetic actuators both generally require a multi-step manufacturing process and sometimes, complex assembly procedures, and both require a comparatively large area on the substrate. In addition the operating mechanisms of prior art MEMS micro-switches often involve friction between moving parts, which is a potential source of unreliability, or at least, variable reliability.

A recent prior art example of a thermally actuated optical MOEMS bistable switch is described in the article entitled "Bistable 2x2 and Multistable 1x4 Micromechanical Fibre-optic Switches on Silicon", by P. Kopka et al., published in the Proceedings of the 3rd International Conference on Micro Opto Electro Mechanical Systems, MOEMS'99, Mainz, 1999, pp. 88-91. Thin film heaters, dissipating up to 1.5 Watts are needed in these devices, and the switching time achieved is reported as < 100ms. Thermal effects are used to operate both for the fiber motion, and the latching mechanism.

A recent prior art example of an electro-magnetically actuated optical MOEMS switch, with latching properties provided by means of a permanent magnet, is described in the article entitled "Modeling and Optimization of Bi-stable Optical Switch" by H. Maekoba et al., published in the Technical Proceedings of the MSM 2000 International Conference on Modeling and Simulation of Microsystems, San Diego, CA, USA, March, 2000, published by Applied Computational Research Society, Cambridge, MA, USA. A switching time of 2 ms. was achieved, with a switching power of lOmW, but a 1 A current was required to switch the device OFF from its latched ON position.

A recent prior art example of a thermally actuated MOEMS bistable microswitch is described in the US Patent Application Publication No. US 2001/ 0010488 Al, to R. Sjohn Minners, for "Bistable micro-switch and method of manufacturing the same". The described device requires currents in the range of from 40 to 160 mA to switch it, and bistable operation is achieved by the use of shape memory alloy elements attached to the actuating arm. The device dimensions of this switch in the plane of the substrate are reported as having a length of between about 500-3000 μm, and a width of between about 200-1200 μm.

In view of the comparative disadvantages of switches such as those mentioned above, there therefore exist a need for a novel, bi-stable, latching microelectronic switch mechanism, which will overcome many of the disadvantages of the prior art MEMS latching switches.

The disclosures of all of the publications mentioned in this section and in other sections of the specification, are hereby incorporated by reference, each in its entirety.

SUMMARY OF THE INVENTION

The present invention seeks to provide novel uses and applications of bending and stressing deformations of optical waveguides, to achieve novel integrated optical devices. One such device is a variable optical wavelength filter, constructed and operative according to a preferred embodiment of the present invention, comprising a waveguide ring resonator coupled to input and output waveguides. A multi-wavelength input signal is passed down the input waveguide in proximity to a waveguide resonator ring. A segment of the resonator ring is suspended to allow in-plane movement. The movement is preferably implemented by means of an in-plane microactuator which exerts tension on this segment of the resonator ring, inducing a strain therein, such that the resonator ring changes its physical length. This change in physical length results in a change in the resonant frequency of the ring. The resonant frequency can be controlled over a predetermined range by means of the force applied by the microactuator. The system thus behaves as a tunable filter, the wavelength of the signal being transferred from the input waveguide to the output waveguide being controllable by the motion of the microactuator. For rings of radius of the order of from 100 to 500 μm, a change of length of the ring of only a few tenths of a micron is sufficient to controllably tune the ring between 1.5μm and 1.55μm.

According to another preferred embodiment of the present invention, a segment of either or both of the input and output waveguides in the vicinity of the coupling gap to the resonator ring is suspended to allow in-plane movement, this movement also preferably being implemented by means of an in-plane microactuator. According to this embodiment, the change in physical proximity of the suspended segment to the resonator ring results in a change in the coupling to or from the ring. This change causes a change in the profile or Q-factor of the filtering properties of the device according to this embodiment.

There is also provided, in accordance with another preferred embodiment of the present invention, a polarization mode dispersion compensator, in which the input optical signal to be compensated is split into two orthogonal polarization components by means of a polarization beam splitter located at a junction in the input waveguide, from which two branches diverge. The two polarization components travel along these two branch waveguides and are then recombined at a second junction to the output waveguide. Such an arrangement of waveguides behaves like a Mach-Zehnder interferometer. A segment preferably of one branch, is suspended, such that it can be stretched by means of an in-plane micro-actuator waveguide. The phase change thus engendered in the light with that particular polarization component passing down that branch, when recombined with the light passing through the unchanged branch, is operative to compensate for the polarization mode dispersion. In order to induce a phase change of 2π in the light passing down a moveable branch of a silicon waveguide, for a light wave having a 1.55 μm wavelength in free space, an elongation of approximately 0.5 μm is required, which is readily attainable with in-plane MEMS actuators of the type utilized in the present invention.

Such a dispersion compensator, according to another preferred embodiment of the present invention, can be used to compensate for chromatic dispersion, if the polarization beam splitter located at the junction in the input waveguide is replaced by a wavelength splitting device.

There is also provided in accordance with another preferred embodiment of the present invention, a micromechanical switch mechanism in which the bending of cantilever arms is utilized for performing the switching operation, and in which bistable operation is achieved by mutually interaction of the arms, such that a mechanical latching action of one arm when in its bent position, is performed by appropriate positioning of the other. The timing of the mutual motion of the cantilever arms for correct operation of the switching mechanism is achieved, according to another preferred embodiment of the present invention, by means of an actuation timing control circuit.

This simple switching arrangement results in friction-free latching operation, since only bending is used to actuate the latching mechanism, and no rotating parts or joints are required. In addition, the switch mechanism operates using only in-plane motion degrees of freedom, and may be realized using surface micromachining. No assembly is required in the construction. For these reasons, the switch mechanism of the present invention is particularly suited to VLSI production methods, and is compact and cost-effective. The switching between the stable states is preferably achieved by means of electrostatic actuators, though it is understood by persons skilled in the art that other actuation methods are possible. However, to exploit the full advantages of the in-plane switch geometry of the present invention, an in plane actuator should preferably be used.

In general, any of the devices according to the preferred embodiments of the present invention may be constructed with waveguides made of silicon, silicon dioxide, lithium niobate, gallium arsenide, gallium nitride, any III-V semiconductor, or any other suitable optical material.

There is further provided in accordance with yet another preferred embodiment of the present invention an optical resonator comprising a substrate, comprising a first waveguide for inputting an optical wave having a plurality of wavelengths, a waveguide ring with a resonant frequency and having at least one segment suspended from the substrate, the ring being disposed relative to the first waveguide such that light is coupled from the first waveguide to the ring, and a second waveguide disposed relative to the ring such that light is coupled from the ring into the second waveguide, wherein the at least one segment is distensible such that the resonant frequency of the ring is adjustable.

In accordance with still another preferred embodiment of the present invention, there is provided an optical resonator comprising a substrate, comprising a first waveguide for inputting an optical wave having a plurality of wavelengths, a waveguide ring with a resonant frequency, the ring being disposed in proximity to the first waveguide such that light is coupled from the first waveguide to the ring, and a second waveguide disposed in proximity to the ring such that light is coupled from the ring into the second waveguide, wherein at least one of the first waveguide and the second waveguide have at least one segment suspended from the substrate, the at least one segment being distensible such that coupling of the light between the ring and at least one of the first and the second waveguides is adjustable.

There is further provided in accordance with still another preferred embodiment of the present invention a tunable multiplexer comprising a first waveguide for inputting an optical wave having a plurality of wavelengths, a plurality of waveguide rings each having a resonant frequency and each having at least one segment suspended from the substrate, the rings being disposed along the first waveguide and relative to the first waveguide such that light is coupled from the first waveguide to the rings, and a plurality of second waveguides, each second waveguide crossing the first waveguide in the vicinity of one of the rings such that light is coupled from at least one of the rings into the one of the plurality of second waveguides disposed in the vicinity of the at least one ring, wherein at least one of the at least one segment is distensible such that the resonant frequency of the ring having the distended segment is adjustable.

In accordance with a further preferred embodiment of the present invention, there is also provided a tunable demultiplexer comprising, a plurality of first waveguides for outputting optical waves, each waveguide having a different one of a plurality of wavelengths, a plurality of waveguide rings each having a resonant frequency and each having at least one segment suspended from the substrate, each of the rings being disposed along a different one of the first waveguides and relative to the first waveguides such that light is coupled from each of the first waveguides to the ring associated with the waveguide, and a second waveguide crossing the plurality of first waveguides in the vicinity of one of the rings such that light is coupled from at least one of the rings into the second waveguide, wherein at least one of the at least one segment is distensible such that the resonant frequency of the ring having the distended segment is adjustable.

There is provided in accordance with yet a further preferred embodiment of the present invention an optical dispersion compensator comprising a substrate comprising, an input multiplexer, an output demultiplexer, a network of switches comprising an array of input waveguides connected to outputs of the multiplexer, and an array of output waveguides connected to inputs of the demultiplexer, wherein the network of switches comprises at least one switch operated by an in-plane microactuator. In the above mentioned preferred embodiment, at least one of the multiplexer and the demultiplexer may be tunable. Furthermore, at least one of the multiplexer and the demultiplexer may comprise an array of waveguide resonator rings, at least one of the resonator rings having at least one segment suspended from the substrate and distensible by means of an in-plane microactuator

There is even further provided in accordance with a preferred embodiment of the present invention an optical dispersion compensator comprising a substrate, comprising a first waveguide for inputting an optical wave, a first junction in the first waveguide, first and second branch waveguides issuing from the junction, such that light input into the first waveguide is split into the branch waveguides, at least one of the branch waveguides having at least one segment suspended from the substrate, a second junction recombining the branch waveguidess into an output waveguide, wherein the at least one segment is distensible such that the optical path difference between the branches is adjustable.

Furthermore, in accordance with yet another preferred embodiment of the present invention, there is provided a micromechanical switching mechanism comprising a microelectronic substrate, a first and a second flexible element on the substrate, at least one of the elements having a first end fixed to the substrate, and each of the elements having a second end free to move generally parallel to the plane of the substrate, each of the flexible elements being associated with an actuator operative to impart movement to the second free end of each element in a direction generally parallel to the plane of the substrate, and a mechanical latching mechanism associated with at least one of the second free ends such that the at least one second free end may be mechanically latched, wherein the mechanically latched mechanism is activated by operation of the actuators according to a predetermined sequence.

In the above-mentioned mechanism, the first element may preferably have its first end fixed to the substrate, and the second element may preferably have its first end fixed to the second free end of the first element, and its second free end capable of being mechanically latched to a fixed object in the substrate. Alternatively and preferably, the first element and the second element may both have their first ends fixed to the substrate, and the latching mechanism may be such as to latch the second free end of the first element to the second free end of the second element.

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention will be understood and appreciated more fully from the following detailed description, taken in conjunction with the drawings in which:

Fig.l illustrates schematically an optical waveguide ring resonator, constructed and operative according to a preferred embodiment of the present invention; Fig. 2 illustrates schematically an optical waveguide ring resonator, constructed and operative according to another preferred embodiment of the present invention, in which a portion of the resonator ring is suspended to allow in plane movement;

Fig. 3 illustrates schematically an optical waveguide ring resonator, similar to that shown in Fig. 2, but wherein the resonator has a non-circular form, and the input and output waveguides do not cross;

Fig. 4 is a graph illustrating the transfer function T, for different wavelengths, as a function of the elongation in μm of a ring resonator such as that shown in Fig. 2;

Fig. 5 shows the effect of change in the attenuation constant α on the transmission of a 1500 nm filter, such as that shown in the results of Fig. 4;

Fig. 6 is a graph of the peak transmission of a filter as a function of the attenuation constant α for a fixed additional length;

Fig. 7 is a graph of the position of the filter center frequency as a function of the additional length of a resonator ring;

Fig. 8 is a transmission plot for a ring resonator wherein the index of refraction of the waveguide is n = 1.5;

Fig. 9 is a plot of the peak resonance wavelength as a function of the additional length;

Fig. 10 is a graph showing the transmission as a function of the index of refraction of prior art ring resonator based on the electro-optical affect;

Fig. 11 is a graph showing the frequency resonance peak as a function of the index of refraction for the ring resonator of Fig. 10;

Fig. 12 shows the same results as those of Fig. 11, but arranged showing the relationship in order of monotonically increasing index of refraction;

Fig. 13 shows the frequency resonance peak as a function of the index of refraction for the ring resonator of Fig. 10, but having a waveguide index of refraction of around 3.5;

Fig.14 is a graph showing the transmission of a mechanically tuned ring resonator such as that shown in Fig. 2, as a function of the additional length, for different coupling efficiencies, η; Fig. 15 is a graph similar to that of Fig. 14, but for a ring resonator constructed of waveguide with a lower attenuation constant;

Fig. 16 is a schematic drawing of a mechanically tuned ring resonator, constructed and operative according to a preferred embodiment of the present invention, using the design criteria described in connection with Fig. 15;

Fig. 17 is a schematic illustration of a 4 x 4 switching network, constructed using mechanically tunable ring resonators of the present invention;

Fig. 18 is a schematic drawing of a wavelength add and drop (WAD) circuit, constructed using mechanically tunable ring resonators of the present invention;

Fig. 19 is a schematic drawing of a different embodiment of a wavelength add and drop (WAD) circuit, constructed using mechanically tunable ring resonators of the present invention;

Figs. 20A and 20B are graphs showing the effective group delay in nanoseconds generated by passage of a wave down a length of silicon waveguide, as a function of the length of the waveguide;

Fig. 21 is a schematic drawing of a dispersion compensating switching network, constructed and operative according to a preferred embodiment of the present invention; and

Figs. 22A and 22B are schematic drawings of a dispersion compensator, according to another preferred embodiment of the present invention;

Figs. 23A to 23C are a schematic illustrations of the switch mechanism, with a latching system, according to a first preferred embodiment of the present invention; Fig. 23A shows the unlatched mechanism, Fig. 23B the mechanism during the latching process, and Fig. 23C, the mechanism latched closed;

Fig. 24 is a schematic illustration of the switch mechanism, with an alternative latching system, according to another preferred embodiment of the present invention;

Fig. 25 shows the timing sequence of the pulses required to operate the switches of the present invention, according to the embodiments shown in Figs. 23 A to 23C and Fig. 24;

Fig. 26 is a schematic circuit diagram showing how the timing of the pulses illustrated in Fig. 25 may preferably be obtained; and Fig. 27 is a simplified example of an integrated circuit, constructed and operative according to another preferred embodiment of the present invention, for operation of the switch mechanisms shown in Figs. 23 A to 23C and Fig. 24.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

Reference is now made to Fig. 1, which illustrates schematically an optical waveguide ring resonator, constructed and operative according to a preferred embodiment of the present invention. Fig. 1 is used to illustrate some of the operative constituent parts of various embodiments of the present invention. The resonator is comprised of two crossing waveguides, an input waveguide 10, and an output waveguide 12, with a ring waveguide 14 located in close proximity to the crossing point. The waveguides are preferably constructed on a silicon substrate as is well known in the art. In the embodiment of Fig. 1, a section of the input waveguide 16 close to the ring resonator 14 is constructed of suspended waveguide, such that it can move in-plane so that it can be closer to or further from the ring resonator 14. The motion of the suspended waveguide section is preferably executed by means of a microactuator 18 in which is incorporated a sensor for determining the position of the microactuator. Such an arrangement is described in US Patent No. 6,128,961 for "Micro-electro mechanics systems (MEMS)", and in PCT patent application No. PCT/TLOO/00268, published as International Publication No. WO 00/71981, for "Micromachined Displacement Sensors and Actuators", both to the inventor in the present application. Constructional methods for the suspended waveguide and its coupling to the ring resonator are described in the co-pending PCT patent application No. PCT/TL01/00787, for "Mode Coupled Optomechanical Devices", to the inventor in the present application, published as International Publication No. WO 02/17004. All of these patent documents are hereby incorporated by reference, each in its entirety.

The ring resonator acts as a fixed resonant cavity having a natural resonant frequency (wavelength) defined as λ_b which is the wavelength at which constructive interference can build up in the ring because of its physical dimensions and properties. The term ring resonator, as specified and as claimed in the present application, is not necessarily meant to be limited to a circular ring resonator, but is rather understood to include any closed loop capable of carrying a resonant current, whether circular, oval, elliptical, rectangular with rounded corners, or any other suitable shape. The resonant cavity also has an input coupling element in the form of the optical coupling across the gap between the input waveguide and the ring, and an output coupling element in the form of the optical coupling across the gap between the output waveguide and the ring. The behavior of the ring resonator can thus be compared to that of a Fabry-Perot resonator with a predefined resonant frequency (and corresponding wavelength) determined by the properties and dimensions of the ring, and with end-coupling mirrors. An optical signal of any wavelength other than the resonant wavelength will continue to propagate along the input waveguide, and will not couple into the second waveguide via the ring. Consequently, if a broadband signal, comprising wavelengths λ_h λ₂, λ₃ .... is transmitted down the input waveguide 10, only the signal of wavelength λi will be preferably induced at an appreciable level in the ring resonator 14, and from there, coupled into the output waveguide 12. The system thus acts as an optical filter whose center wavelength is λj. The coupling coefficient can be varied by varying the proximity of the suspended waveguide section 16 to the ring resonator 14.

The DI sensor incorporated into the actuator 18 enables the system to be used with a feed-back mechanism, such that the coupling level can, for instance, be kept at a constant level, or at any other desired functional level.

As an alternative to the arrangement shown in Fig. 1, where the resonator ring is fixed and the input waveguide is moved, other alternative and preferable configurations can be provided wherein it is the output waveguide 12 which is suspended and performs the variable coupling, or wherein both the input and output waveguides are suspended. Alternatively and preferably, the input and output waveguides can be fixed, and the resonator ring be constructed as a suspended element, such that movement of the complete ring is used to vary the coupling.

Change in the coupling gap in the embodiment of Fig. 1 is effective only to affect the Q-factor or profile of the filter, but does not affect the frequency of the ring resonator, which is determined by the physical dimensions and properties of the ring itself. Reference is now made to Fig. 2, which illustrates schematically an optical variable wavelength filter comprising a waveguide ring resonator, constructed and operative according to another preferred embodiment of the present invention. This embodiment differs from that shown in Fig. 1 in that a portion of the resonator ring 14 is suspended to allow in-plane movement. The microactuator 20 is operative to exert tension on this section of the resonator ring, and to induce a strain therein, such that the resonator ring changes its physical length. This change in physical length results in a change in the resonant wavelength of the ring, and the outcome is that the resonant wavelength can be controlled over a predetermined range by means of the force applied by the microactuator 20. This system thus behaves as a tunable filter, the wavelength of the signal being transferred from input waveguide 10 to output waveguide 12 being controllable by the motion of the microactuator 20. It has been calculated that for rings of radius of the order of from 100 to 500 μm, a change of length of the ring of only a few tenths of a micron is sufficient to controllably tune the ring between 1.5 μm, and 1.55 μm. Examples of the wavelength characteristics obtained from such a device are given hereinbelow in Figs. 4 to 7.

The DI sensor incorporated into the actuator 20 enables the system to be used with a feed-back mechanism, such that the filter center wavelength can, for instance, be kept at a constant value. The feedback signal is preferably obtained from any wavelength sensitive element on the output side of the optical system.

Reference is now made to Fig. 3, which illustrates schematically an optical waveguide ring resonator, constructed and operative according to yet another preferred embodiment of the present invention. This embodiment differs from that shown in Fig. 2 in two ways. Firstly, the resonator has a non-circular form 22. The microactuator 24 is operative to move one leg 26 of the resonator, and thus to change the resonant wavelength of the device. The shape of the resonator ring 22 can preferably be chosen to optimize Q-factor, or to optimize space utilization as a function of filter efficiency and selectivity. Secondly, the output waveguide 28, instead of crossing the input waveguide, runs on the opposite side of the ring to the input waveguide 10, and preferably parallel to it. In order to illustrate the performance achievable by a tuned filter according to the various embodiments of the present invention, the transfer function (transmission function) T, of a ring resonator needs to be calculated. This is given by :

where:

4RA

F is the finesses of the resonator, given by F

(l -RA)²

A = e^~aL

R = l - 77 φ = βL β = kn_si

, 2π k = — λ

L is the propagation length around the ring; α is the decay constant around the ring; η is the coupling efficiency between the input/output waveguides and the ring; λ is the wavelength; and n_si is the index of refraction of the waveguide, generally silicon for the examples shown.

From the above analysis, a number of conclusions can be made about the properties of the ring resonators of the present invention. Firstly, the longer the ring the larger the losses due to the attenuation factor A. It is possible to use small radius rings, but this is not practical for low index of refraction waveguides due to the increased radiation losses from the waveguides.

The larger the index of refraction (n_Sj), the larger the Q factor. This is one of the advantages of the use of SOI waveguides.

The Q factor is controlled by the finesse of the resonator. The larger the finesse the larger is the Q factor. In order to obtain large finesse values, both η and α should be small. Of these two parameters, η has the larger affect on the Q factor. Reference is now made to Fig. 4, which shows the transfer function T, for different wavelengths, as a function of the elongation in μm of a ring resonator. The results shown in Fig. 4 were obtained with a ring of radius r = lOOμm, and with η = 0.05 and α = 3 x 10^"3 /cm ( = 3 x 10^"7 / cm, or 0.013db/cm) . The wavelengths to be plotted were selected to be at intervals of 1 nm, and as is observed from the graphs of Fig. 4, it is possible to tune the filter with a resolution of considerably better than 1 nm, by controlled movements of the micro-actuator. The range of movements required of the micro-actuator for this purpose are well within the operational capabilities of these actuators.

The results obtained using a higher attenuation constant , are similar in shape and resolution to those of Fig. 4, but with lower transmission. In Fig. 5 is shown the effect of change in the attenuation constant α on the transmission of a 1500 nm filter, such as that shown in the results of Fig. 4. As is observed, as the attenuation constant increases, the peak transmission decreases, and the Q of the filter degrades accordingly.

In Fig. 6 is shown a graph of the peak transmission as a function of the attenuation constant α for a fixed additional length. SOI waveguide typically has losses in the range of 0.1 db/cm., which results in very low insertion losses. From Fig. 6, it is apparent that even for material with a 1 db/cm attenuation constant, the losses at resonance are only of the order of 3db.

Reference is now made to Fig. 7, which is an alternative method of plotting the performance of the variable filter, according to the above-described embodiment of the present invention. In Fig. 7, the position of the filter center frequency is plotted as a function of the additional length of the ring in microns. The results are given for a resonator ring having r = lOOμm, η = 0.05 and α = 3 x 10^"1 /cm ( = 1 db/cm). As is observed, a steady monotonous change of the center wavelength can be obtained as a function of the actuator motion. There is, however, overlapping between wavelengths at the ends of their useable range. Thus for instance, in the simulation used in Fig. 7, for wavelengths above 1525nm, the passbands begin to coincide with shorter wavelengths. Reference is now made to Fig. 8, which is a transmission plot for the case when the index of refraction of the waveguide is n = 1.5. In order to get the same frequency span as for the case for the refractive index of silicon, as was shown in Fig. 4, the change in the length has to be increased to the order of 1 micrometer. In addition, as shown in Fig. 9, the resonance peak is not a monotonic function of the length change, and varies in rapid cycles between 0 and 1 micrometer. This behavior may complicate the information management procedures for the tuning of such a ring resonator. Similar behavior is also found for a resonator with n = 3.5, as observed in the results of Fig. 7, but to a far lesser extent.

A comparison is now made of the performance of the actuator-tuned ring resonator filter of the present invention with prior art electro-optically tuned ring resonators, which are ring resonators where the index of refraction of the ring, and hence the resonant frequency, is changed by an electro-optical affect.

Fig. 10 is a graph showing the transmission as a function of the index of refraction of such a prior art ring resonator for different wavelengths. The wavelength separation of the individual transmission plots is lnm. The parameters used are: r = lOOμm, η = 0.05, α = 3x10^"1 /cm (1 db/cm). As is observed from the graph, a small change in the refractive index causes the resonant wavelength to change significantly. This means that the center frequency is a sensitive function of the refractive index, and it is likely that drift in the index of refraction, even that arising from small environmental temperature changes, are large enough to mandate the use of temperature stabilization in such prior art electro-optically tuned ring resonators.

Reference is now made to Fig. 11 which is a graph showing the frequency resonance peak as a function of the index of refraction for such a ring resonator, showing that the peak wavelength changes in rapid cycles between n = 1.502 and 1.504. Fig. 12 shows the same results as those of Fig. 11, but arranged showing the relationship in order of monotonically increasing index of refraction. Fig. 13 shows the situation for a range of index of refraction around 3.5. These graphs show similar behavior to that shown in the graphs of Figs. 7 to 9 for the mechanically tuned resonators of the present invention. The main advantage of the resonators of the present invention over prior art electro-optically tuned ring resonators thus remains their comparative temperature stability.

Reference is now made to Fig. 14, which is a graph showing the transmission of a mechanically tuned ring resonator such as that shown in Fig. 2, as a function of the additional length, for different coupling efficiencies, η. For the example shown, r = lOOμm and α = 3 x 10^"Vcm (1.3db/cm). The transmission curves are plotted at intervals of 0.1 coupling efficiency. For η = 1, R = 0 and T_max = 1. The passband is therefore flat and transmits all of the wavelengths equally, as would be expected from the input and output waveguides in good optical contact with the resonator ring, thereby effectively damping any resonance in the ring. As the coupling efficiency decreases the passband becomes narrower and the peak transmission also decreases.

Reference is now made to Fig. 15, which is similar to that of Fig. 14, but for a ring resonator having a value of α = 3 x 10^"2 /cm (0.13db/cm). For this value of attenuation constant, the peak transmission of the filter is hardly reduced as the coupling constant is lowered. This graph is useful because silicon-on- insulator (SOI) waveguide can be fabricated with losses in the range of 0.1 db/cm. In order to provide a high selectivity filter, the coupling between the waveguides and the ring is arranged to be in the range of 0.1. This can be simply achieved by adjusting the gap between the waveguides and the resonator ring. In addition, since a low coupling level is required, the waveguides and the ring can be ribbed waveguides, which support only a small number of modes, or even a single mode, as is known in the art. For higher coupling levels, it would be difficult to use such ribbed waveguide.

Reference is now made to Fig. 16, which is a schematic drawing of a mechanically tuned ring resonator, constructed and operative according to a more preferred embodiment of the present invention, using the design criteria described in the above discussion regarding Fig. 15. The coupling efficiency can be simulated using the software by Rsoft, Ossnining NY to determine the required overlapping length. It should be noted that in this case, the overlapping area is made of one material and the coupling does not depend on the surface quality at the overlapping area. The input and the output coupling are preferably equal. According to yet another preferred embodiment, the coupling gap or gaps can be tuned, such that the passband shape of the filter, as shown in the graphs of Figs. 14 and 15, is modulated. This can preferably be accomplished by applying a micro-actuator to a section of suspended waveguide which couples to the ring, as shown in Fig. 1 hereinabove, either at one or at both coupling gaps.

The mechanical strain in the suspended waveguide can be simulated using a standard mechanical simulation software package, such as Ansys, available from ANSYS Inc. of Canonsburg, PA 15317. It is also possible to take into account the photoelastic constant, to provide more accurate simulation results.

Reference is now made to Figs. 17 to 19, which are schematic illustrations of different proposed applications of the mechanically tunable ring resonators of the present invention, constructed and operative according to more preferred embodiments of the present invention.

In Fig. 17 there is illustrated schematically a 4 x 4 switching network 100, such as that described in the above-mentioned WIPO International Publication No. WO 02/17004 for "Mode Coupled Optomechanical Devices". The function of such a network is to enable the connection of any of the four inputs 102-105 to any of the four outputs 106-109. Generally, each channel contains a signal of different wavelength. In order to obtain the four separate wavelength inputs from the single input signal fiber 110, a multiplexer must be used. The ring resonators of the present invention can be ideally used in a planar structure, compatible with that of the rest of the switching network, in order to multiplex the input signal 100. The multi-wavelength input signal if fed to a serial array of tunable ring resonators 111-114, such as those of the present invention shown in Figs. 1 - 3. Each of the ring resonators is operative to divert one wavelength λi, λ₂, λ₃ or λ from the input line to one of the four inputs of the switching network. The ring resonators act as wavelength selective switches, according to the tuned wavelength to which the microactuator of each one is adjusted. According to another preferred embodiment of the present invention, instead of using fixed frequency ring resonators, it is possible to adjust the filter frequency of each of the ring resonators dynamically, according to the needs of the system, such that savings in system size, and/or increases in system efficiency can be expected. According to a further preferred embodiment of the present invention, a serial array of tunable ring resonators according to the present invention, can also be used as a demultiplexer for combining several different wavelength signals into one multi- wavelength signal.

Reference is now made to Figs. 18 and 19, which are schematic drawings of wavelength add and drop (WAD) circuits, such as that described in the above-mentioned International Publication No. WO 02/17004 for "Mode Coupled Optomechanical Devices", constructed and operative according to more preferred embodiments of the present invention. In the preferred embodiment shown in Fig. 18, there is shown a three channel WAD, for wavelengths λ λ₂, λ₃ in which in each of the input port 120 and the ADD port 122, are located a serial array of mechanically tuned ring resonators, 124, 126 respectively, to multiplex the incoming broadband signal into its separate wavelength components λ_b λ₂, λ₃. Similarly, at each of the DROP port 130 and the PASS port 132, are located a serial array of mechanically tuned ring resonators, 134, 136 respectively, to demultiplex the outgoing separate wavelength components λ_b λ₂, λ₃ into a single broadband signal.

According to the preferred embodiment shown in Fig. 19, there is shown a three channel WAD, for wavelengths λ_l5 λ₂, λ₃, whose only operative elements are essentially two serial arrays of mechanically tuned ring resonators according to the present invention. The first array 140 is located serially with the IN port 142, and is operative to either pass the selected wavelength to the PASS port, or to switch it to the DROP port, according to the setting of the actuator in the resonator for that particular wavelength. The second array 144 is located such that the output waveguides of all of the ring resonators are connected in series such that they all output their signals to the PASS port, while each of the input waveguides of the individual ring resonators act as separate ADD ports 146, which are activated according to the setting of the corresponding ring resonator actuator. This WAD system is thus of very compact dimensions and simple to manufacture.

Turning now to the dispersion correction aspects of the present invention, reference is now made to Fig. 20A, which is a graph showing the effective group delay in nanoseconds generated by passage of a wave down a length of silicon waveguide, as a function of the length of the waveguide in micrometers. Fig. 20B shows a magnified portion around the zero's of Fig. 20 A, in the region of short delay and short waveguide length. It is found that to correct polarization mode dispersion or chromatic dispersion by the above-mentioned method of transmitting part of the wave down an additional optical path, an additional delay time ranging from the order of a picosecond to about a nanosecond is required. These are the generally found time differences which packets of light of different wavelengths gain in propagation down fiber systems, because of chromatic dispersion. Similar time delays are found for the up to 2π phase shift required for polarization dispersion correction, whereby light in different polarization states propagates at different group velocities leading to dispersion, which is removed by rotating the polarization of the light to one of the principle direction, as explained hereinabove. As is observed from the graphs, the lengths of silicon waveguides having such propagation time delays range from a few tens of microns to the order of a few millimeters. These lengths are in the correct range of lengths which can be simply incorporated on an integrated optics chip.

In a typical communication system application, the light signals may arrive at the receiver from widely different locations, and may have thus undergone different levels of dispersion, whether PMD or CD. An optimum solution for compensating for these dispersions is therefore to provide tuning ability to the additional path length introduced to compensate for the dispersion.

Reference is therefore made to Fig. 21, which is a dispersion compensating switching network 200, constructed and operative according to a preferred embodiment of the present invention. The switching network is preferably constructed using an array of opto-mechanical switching devices 202, such as those described in the above-mentioned WIPO International Publication No. WO 02/17004 for "Mode Coupled Optomechanical Devices" and as illustrated in the embodiment shown in Fig. 17 hereinabove. The input fiber 204 carries a wideband input signal to an on-chip multiplexer 206 which disperses the different component wavelengths λi, λ₂, λ₃ and λ₄, onto different input waveguides 210-216 of the switch network. After traversing the switching network, the individual component wavelengths λ_b λ₂, λ₃ and λ on the output waveguides 220-226 are recombined in a demultiplexer 228, such as a star coupler, and output to the output fiber 230. The dispersion compensation is performed by selecting a path through the switching network for each wavelength such that the path length induced phase shift for that particular wavelength exactly compensates for the dispersion induced phase shift. The phase shift induced by each path through the network is known. The operation of the dispersion compensator depends on knowledge of the dispersion which each wavelength has undergone in transmission through the system. Such information may preferably be obtained from the eye-diagram of the transmission characteristics, a good eye diagram indicating a low error-bit rate and low dispersion, as is known in the art. This information about the required paths is processed by the system controller to set the optical switches of the network to select the path through the network, such as to provide phase compensation for that particular wavelength. In the preferred embodiment shown in Fig. 21, the shortest path length, 2Lp, is generated between input waveguide 216 and output waveguide 220, while the longest path length, 8Lp, is generated between input waveguide 210 and output waveguide 226, where Lp is the switching pixel width.

A dispersion compensator using a fixed multiplexer at its input, though the simplest such embodiment to implement of the present invention, is limited in the correction that can be achieved, since each wavelength band is directed to a specific input waveguide, and the number of path options from that particular input waveguide to the range of output waveguides is limited. According to another preferred embodiment of the present invention, the multiplexer is capable of being tuned such that each wavelength can be directed to a suitable input waveguide according to the requirements of the processing that is to be performed on that wavelength band, as explained hereinabove. In this preferred embodiment, each wavelength band can be directed to any input waveguide, such that the complete range of path lengths present in the network is available to each of the input wavelength bands. For use in this embodiment, such a tuned multiplexer can preferably be of a tunable array waveguide grating type, such as that described in co-pending US Provisional Patent Application, Serial No. 60/286,448 for "Tunable Array Waveguide Grating" to the inventor of the present application. Alternatively and preferably, an array of tunable ring resonators can be used as a tunable multiplexer, each resonator acting as a wavelength selective switch according to the tuned wavelength to which the micro-actuator of each one is adjusted. Such a tuned multiplexer is described in connection with the embodiment of Fig. 17 hereinabove.

As an illustrative example of the dispersion compensator, constructed and operative according to the present invention, required to cover a range of delays of up to 1 nanosecond, requires an overall network size of 5 cm. x 5 cm. Such a size is required to provide a diagonal path length of about 8 cm, sufficient to provide the required 1 ns. delay. Such a 64-channel compensator, using a pixel size Lp of 800 μm, is capable of covering delay times of from 16 ps. for the shortest path length, to 1 ns. for the longest path length, in discrete steps of 16 ps. If smaller compensation delays are required, then it is possible to reduce the size of the chip. Thus, for a range of delays of from 8 ps to 500 ps, a chip size of 2.5cm x 2.5 cm is required.

In the preferred embodiment of the network shown in Fig. 21, in order to illustrate the operation of the dispersion compensation system and method according to the present invention, a chromatic dispersion compensator application is shown. It is to be understood, however, that the same network can be used as a polarization mode dispersion compensator by the addition of a PBS at the input and output of the system.

Reference is now made to Fig. 22A, which is a schematic illustration of a dispersion compensator, according to another preferred embodiment of the present invention. The input optical signal to be compensated 250 is split into its two orthogonal polarization components by means of a polarization beam splitter 252 located at the junction of the two branches of the compensator. The two polarization components travel along different paths 254, 256, and are then recombined at junction 260, and are output from waveguide 262. Such an arrangement of waveguides behaves like a Mach-Zehnder interferometer, as is known in the art. The effect of the activation of the actuator 258 is illustrated in Fig. 22B, where it is seen that waveguide 254 undergoes a bending strain and elongates. A light wave having a 1.55 μm wavelength in free space has a wavelength in a silicon waveguide of about 0.5 μm. Therefore, to cover a phase change of 2π, waveguide 254 must be bent such that its length changes by this length, i.e. 0.5 μm. For a 200 μm long silicon waveguide beam of cross section 2μm x lOμm, an elongation of 0.5 μm is obtained when the beam is displaced by 7 μm at its center point. Calculations of the stress required to produce this lateral strain in silicon show that it is well within the capabilities of the MEMS actuators of the type described in the above mentioned US Patent No. 6,128,961 and in WIPO Publication No. WO 00/71981.

Though the major phase change in the light traversing the elongated waveguide 254 arises as a result of the change in physical length of the waveguide, another result of the bending stress is a change in the refractive index of the waveguide arising from the photo-elastic effect, which changes the refractive index of the waveguide medium. Such a change in refractive index may also contribute in providing the total change in optical length.

The waveguide in Figs. 22A and 22B is schematically drawn as normal rectangular waveguide. The use of such waveguide can result in a possible malfunction of the simple compensator, since such waveguide can support higher order modes. Therefore, after recombination at junction 260 of the two separated wavefronts, instead of the desired constructive or destructive interference expected from the Mach-Zehnder interferometer, according to the mutual phases of the two branches, it is possible that the wavefronts combine into a higher order mode, thereby by-passing the interferometric effect. The use of a single mode waveguide for the output waveguide 262 prevents the propagation of such higher order modes, and thus ensures the correct operation of the compensator. The single mode waveguide used can preferably be a ribbed waveguide, as described in the embodiments in the above-mentioned WIPO publication WO 02/17004.

The PBS 252 used in the described embodiments of the present invention, are preferably constructed of a stack of Brewster mirrors located at the correct angle in the junction 252, as is known in the art. Alternatively and preferably, it is possible to use the polarization splitters described in the articles "Polarization Independent Integrated Optical, Acoustically Tunable Double-Stage Wavelength Filter in LiNbO₃" by F. Tian et al, published in Journal of Lightwave Technology, Vol. 12, pp 1192 -1196, Jul 1994, or "An Integrated Optic Adiabatic Mode Splitter on Silicon" by R.M. deRitter et al, published in Journal of Lightwave Technology, Vol. I l, pp l806 -1811, Nov. 1993. In order to illustrate another application of bending in a micro-optomechanical device, according to more preferred embodiments of the present invention, reference is now made to Figs. 23A to 23C, which illustrate schematically a plan view of a switch mechanism with a latching system, constructed and operative according to a first preferred embodiment of the present invention. In Fig. 23 A, there is shown a first cantilevered beam 310, capable of in-plane bending motion relative to the substrate on which it is constructed, which is parallel to the plane of the drawing. The beam 310 is fixed to the substrate at one end 312, and is otherwise suspended over the substrate. At its other end, the beam has a latching protuberance 314. The beam is bent by means of an actuating mechanism 316, preferably an electrostatic actuator based on a parallel plate or a comb drive, as is known in the art. Alternatively and preferably, the actuator can be any other suitable sort of micro-actuator, such as an electromagnetic or a thermal actuator. A second cantilevered beam 318, also fixed to the substrate at one end 320 and otherwise suspended over the substrate, is located on the substrate in such a position that its free end 322 can interact with the protuberance 314 on the first cantilevered beam 310. The second cantilevered beam 318 has its own actuating mechanism 324. The situation shown in Fig. 23A is of the switch mechanism in its unengaged rest position.

The switching mechanism is operated by activating the first actuator 316 such that the first cantilevered beam 310 moves to its bent position. The second actuator 324 is then activated such that the second cantilevered beam 318 then moves to its bent position. This situation, with both of the beams bent, is shown in Fig. 23B. The first actuator is then deactivated, such that the first beam returns to its normal unbent position, thereby latching the second beam in its bent position by means of the latching protuberance 314. The second actuator 324 can then also be deactivated since the protuberance 314 now holds the second beam latched. This position is illustrated in Fig. 23C. The switching mechanism is thus locked in its latched condition. This condition can be regarded as the ON or the OFF position, depending on the particular switch configuration in which it is applied.

The switching mechanism is returned to its unlatched position by a procedure which is a variation of the reverse of the actuating procedure described above. First of all, the second actuator 324 is activated, to provide clearance between the latching protuberance 314 of the first beam 310 and the second beam 318. Then the first actuator 316 is activated to disengage the latching protuberance 314 of the first beam 310 from the end of the second beam 318, so that when the second actuator 324 is deactivated, the second beam 318 returns to its unlatched rest position. Finally, the first actuator 316 is deactivated such that the first beam 310 can return to its unlatched rest position.

In the embodiment shown in Figs. 23A to 23C, the end 322 of the second beam 318 is latched on the inward side of the latching protuberance 314 of the first beam 310. If the second beam 18 were located in its rest position to the left of the protuberance 314, then the latching operation could be accomplished by causing the activator 324 to bend the second beam to the right, such that on release, it latches onto the right hand side of the protuberance 314. It is to be understood that the embodiment shown in Figs. 23A to 23C is only illustrative of one embodiment of latching switch configuration according to the present invention, and that the invention is meant to cover any combination of simple interacting cantilevered beams which latch onto each other when activated in the correct temporal order.

Reference is now made to Fig. 24, which is a schematic illustration of an alternative preferred embodiment of the latched switching mechanism of the present invention. This embodiment differs from the type shown in Figs. 23 A to 23 C in that the first actuator arm 330 is connected at its non-latching end to the free end of the second actuator arm 332. The latching end of the first actuator arm 330 has a protuberance 334, which is able to latch behind a stub 336 fixed in the substrate.

Latching is accomplished in four steps. The steps required to latch the switch can be simply described utilizing the nominal "directions" on the page of Fig. 24, as follows:

Step 1 : Cantilever beam 330 is actuated upwards

Step 2: Cantilever beam 332 is actuated leftwards

Step 3: Cantilever beam 330 is released (returns back to its original position)

Step 4: Cantilever beam 332 is released (remains latched by the protuberance

334 on cantilever beam 330). To unlatch the switching mechanism, the above designated steps are applied in the following order:

Step 2: Cantilever beam 332 is actuated leftwards

Step 1: Cantilever beam 330 is actuated upwards

Step 4: Cantilever beam 332 is released

Step 3: Cantilever beam 330 is released. The nomenclature of upwards, downwards, left and right can also be used to more simply describe the operating stages of the embodiment of Figs. 23A to 23C.

As with the embodiment shown in Figs. 23A to 23C, the preferred embodiment of Fig. 24 can also be operated by having the protuberance 334 of the first cantilevered beam 330 latch on the opposite face of the stub 336. This embodiment could be implemented by having the stub 336 located to the right of the protuberance 334 in its rest position, and having the actuator 338 operate to pull the beam 332 such that the protuberance rests in tension on the right hand side of the stub 336 when latched.

Though the present invention has been described in terms of the various preferred embodiments shown in Figs. 23A to 23C and 24, it is to be understood that these embodiments are only two examples of configurations of micro- mechanically latching arm mechanisms, and that there exist other combinations of arm geometries which can be made to move interactively such that a latching mechanism is provided. It is to be understood that the present invention is not meant to be limited to the embodiments shown, but to any similar combination of mechanically latching arms, actuated by motion parallel to the plane of the substrate on which the arms are mounted.

The correct latching operation of the switching mechanism requires that the timing of the actuation of the two arms is performed correctly. Reference is now made to Fig. 25, which is a time-graph showing a preferable method of sequencing the operation of the actuators for latching the switching mechanism, by applying separate and temporally overlapping square voltage waveforms 340, 342, one voltage waveform for each actuator. The leading edge of the first waveform at time t_! actuates cantilever beam 310 of Fig. 23 A, the leading edge of the second waveform at time t₂ actuates cantilever beam 318, the falling edge of the first waveform at time t₃ releases cantilever beam 310, and the falling edge of the second waveform at time t₄ releases cantilever beam 318. The switching mechanism operates correctly so long as t₂ > t_t and t₄ > t₃. Though square pulses are shown in the embodiment represented by Fig. 25, it is to be understood that any other form of pulse shape may be used, provided that the applied voltages are sufficient to ensure that the actuators are activated, and that the voltages are above the activation threshold for the duration of the correct timing sequences as described above.

The operating pulses can either be applied by means of two external pulse sources, or they can preferably be supplied from one voltage source by means of an electrical circuit that generates a sequence of pulses with a time delay between them. Reference is now made to Fig. 26 which is a schematic diagram of a circuit which can generate such a sequence of two pulses by means of a pair of RC circuits. The actuators are represented by Al and A2. When the switch SI is closed the voltage V is applied to the actuator Al after a time delay characteristic of the time constant RCi, and to the actuator A2 after a time delay characteristic of the time constant RC₂. In order for the sequencing to operate correctly, i.e. for Al to be activated before A2, it is necessary to ensure that RC_! < RC₂ (i.e. Cι<C₂). In order to complete the latching process, the switch SI is opened and Al deactivates before A2 as required, since RC] < RC₂. In order to release the latched switching mechanism, it is necessary to first activate actuator 2, and then actuator 1, i.e. the timing sequence is reversed in comparison with the latching process. Therefore, it is necessary to ensure that RCj > RC₂ (i.e., > C₂).

Reference is now made to Fig. 27, which is a schematic diagram of a preferred circuit which enables the reversal of the characteristic time constants of the two RC circuits shown in the simple circuit of Fig. 26, to allow for latching and unlatching with a single circuit. In this circuit, Mi and M₂ represent the two actuators. Four MOS switches with gates Gj_j are used to determine which capacitors of CI and C2 are connected in circuit with the resistors R in determining the delay in the supply of activating voltage to each of the actuators Ml, M2. The holding voltages are applied to their respective gates according to a predetermined order defined by the state of the switching mechanism operating switch, marked SW1 in Fig. 27. Thus, when SW1 is closed in order to latch the switching mechanism close, Ml has to activate before M2, as previously explained. It is thus necessary to arrange that the RC circuit associated with Ml is less than that associated with M2. Thus, for example, assuming in Fig. 27 that CI < C2 the control processor is arranged such that when SW1 is closed, gates G12 and G22 are powered, such that their respective MOS switches short out the capacitors across which they are connected, such that Ml has a characteristic time delay determined by RC1, and M2 has a characteristic time delay determined by RC2. Since Cl < C2, it follows that RCK RC2. Therefore the cantilever beam connected to Ml moves before that connected to M2. In the embodiment shown in Figs. 23 A to 23C, this means that cantilever beam 310 moves upwards before cantilever beam 318 moves leftwards. This completes the first half of the switching step by sequentially activating the cantilever beams with time delay Δt = t₂ - ti as defined in Fig. 25. When SW1 is opened the voltage across the two actuators falls to zero according to the RC constants associated with each actuator. Since RCK RC2, actuator Ml responds before actuator M2 such that the time delay between the actuators for the voltage to decay to zero is now given by Δt = t₄ - 1₃. This time delay results first in the downwards movement of the cantilever beam connected to Ml, followed by the rightwards movement of the cantilever beam connected to M2.

In the unlatching step, when SW1 is closed to activate the unlatching process, the order of the movement of the cantilever beams must be reversed. This is achieved by arranging that the control processor ensures that the MOS switches associated with gates Gl 1 and G21 are closed, such that Ml has a characteristic time delay RC2 and M2 has a characteristic time delay RC1. Since RCK RC2, the significance of this is that the time-delay between the two actuators is reversed, as required for the unlatching procedure.

The switching mechanism described in the above-mentioned preferred embodiments of the present invention have a number of advantages over prior art switches. These advantages include:

1. Very small size, as small as 200μm by 200μm using currently available technology

2. Simple integration with current surface MEMS technologies

3. Simple manufacture, without the need for assembly operations 4. Fast MEMS switching times, as low as a few microseconds

5. Low power consumption, especially if use is made of electrostatic actuators, requiring no current for operation

6. Ease of integration in large matrices.

The bistable switching mechanisms described in the above-mentioned preferred embodiments of the present invention have been described in general terms without adaptation to any specific application. It is clear to anyone skilled in the art that such mechanisms can be advantageously used in many MEMS technology applications, especially those constructed using the techniques of on-surface micromachining, with an in-plane degree of freedom.

It is appreciated by persons skilled in the art that the present invention is not limited by what has been particularly shown and described hereinabove. Rather the scope of the present invention includes both combinations and subcombinations of various features described hereinabove as well as variations and modifications thereto which would occur to a person of skill in the art upon reading the above description and which are not in the prior art.

Claims

CLAIMSWe claim:

1. An optical resonator comprising a substrate, comprising: a first waveguide for inputting an optical wave having a plurality of wavelengths; a waveguide ring with a resonant frequency and having at least one segment suspended from said substrate, said ring being disposed relative to said first waveguide such that light is coupled from said first waveguide to said ring; and a second waveguide disposed relative to said ring such that light is coupled from said ring into said second waveguide; wherein said at least one segment is distensible such that said resonant frequency of said ring is adjustable.

2. An optical resonator according to claim 1 and wherein at least one of said first waveguide and said second waveguide also has at least one segment suspended from said substrate, said at least one segment being distensible such that coupling of said light between said ring and at least one of said first and said second waveguides is adjustable.

3. An optical resonator according to claim 2 and wherein said adjustable coupling is operative to change the shape of the wavelength profile of at least one of said light coupled from said first waveguide to said ring and from said ring to said second waveguide.

4. An optical resonator according to any of the previous claims and wherein said ring disposed relative to at least one of said first and said second waveguides is in proximity to at least one of said first and said second waveguides.

5. An optical resonator according to any of the previous claims and wherein said ring disposed relative to at least one of said first and said second waveguides touches at least one of said first and said second waveguides.

6. An optical resonator according to claim 1 and also comprising an in-plane microactuator attached to said at least one segment for distending said segment.

7. An optical resonator according to claim 2 and also comprising an in-plane microactuator attached to said at least one segment for distending said segment.

8. An optical resonator according to claim 6 and wherein said at least one segment is distended by tension such that its length changes.

9. An optical resonator according to any of the previous claims and wherein said first and said second waveguides cross each other.

10. An optical resonator according to any of the previous claims and wherein said first and said second waveguides are essentially parallel.

11. An optical resonator according to any of the previous claims and wherein at least one of said first and said second waveguides is a single mode waveguide.

12. An optical resonator according to any of the previous claims and wherein at least one of said first and said second waveguides is a ribbed waveguide.

13. An optical resonator according to any of the previous claims and wherein said ring is circular.

14. An optical resonator according to any of the previous claims and wherein said ring is non-circular.

15. An optical resonator according to any of the previous claims and wherein at least one of said waveguides and said ring comprises a material selected from the group consisting of silicon, silicon dioxide, lithium niobate, gallium arsenide, gallium nitride and a III-V semiconductor.

16. An optical resonator according to any of the previous claims and wherein said light coupled to said second waveguide has a wavelength corresponding to said resonant frequency.

17. An optical resonator according to any of the previous claims and wherein said resonant frequency of said ring is adjusted to filter light of a preselected wavelength from said optical wave having a plurality of wavelengths.

18. An optical resonator according to claim 6 and wherein said microactuator is controlled by means of a signal corresponding to the wavelength of light coupled into said second waveguide, such as to stabilize the resonant frequency of said ring.

19. An optical resonator according to claim 7 and wherein said microactuator is controlled such as to control coupling of said light between said ring and at least one of said first and second waveguides.

20. An optical resonator comprising a substrate, comprising: a first waveguide for inputting an optical wave having a plurality of wavelengths; a waveguide ring with a resonant frequency, said ring being disposed in proximity to said first waveguide such that light is coupled from said first waveguide to said ring; and a second waveguide disposed in proximity to said ring such that light is coupled from said ring into said second waveguide; wherein at least one of said first waveguide and said second waveguide have at least one segment suspended from said substrate, said at least one segment being distensible such that coupling of said light between said ring and at least one of said first and said second waveguides is adjustable.

21. An optical resonator according to claim 20 and wherein said waveguide ring also has at least one segment suspended from said substrate, said at least one segment being distensible such that said resonant frequency of said ring is adjustable.

22. An optical resonator according to claim 20 and also comprising an in-plane microactuator attached to said at least one segment for distending said segment.

23. An optical resonator according to claim 21 and also comprising an in-plane microactuator attached to said at least one segment for distending said segment.

24. An optical resonator according to any of claims 20 to 23 and wherein at least one of said waveguides and said ring comprises a material selected from the group consisting of silicon, silicon dioxide, lithium niobate, gallium arsenide, gallium nitride and a III-V semiconductor.

25. A tunable multiplexer comprising; a first waveguide for inputting an optical wave having a plurality of wavelengths; a plurality of waveguide rings each having a resonant frequency and each having at least one segment suspended from said substrate, said rings being disposed along said first waveguide and relative to said first waveguide such that light is coupled from said first waveguide to said rings; and a plurality of second waveguides, each second waveguide crossing said first waveguide in the vicinity of one of said rings such that light is coupled from at least one of said rings into said one of said plurality of second waveguides disposed in the vicinity of said at least one ring; wherein at least one of said at least one segment is distensible such that said resonant frequency of said ring having said distended segment is adjustable.

26. A tunable multiplexer according to claim 25 and wherein said resonant frequency of at least one of said rings is adjusted to filter a predetermined wavelength from said first waveguide to said second waveguide disposed in the vicinity of said at least one ring.

27. A tunable multiplexer according to claim 25 and wherein said multiplexer is adapted as the input to a wavelength selective switching network.

/

28. A tunable multiplexer according to claim 25 and wherein said multiplexer is adapted as the input to a drop and add switching network.

29. A tunable multiplexer according to any of claims 25 to 28 and wherein at least one of said waveguides and said ring comprises a material selected from the group consisting of silicon, silicon dioxide, lithium niobate, gallium arsenide, gallium nitride and a III-V semiconductor.

30. A tunable demultiplexer comprising; a plurality of first waveguides for outputting optical waves, each waveguide having a different one of a plurality of wavelengths; a plurality of waveguide rings each having a resonant frequency and each having at least one segment suspended from said substrate, each of said rings being disposed along a different one of said first waveguides and relative to said first waveguides such that light is coupled from each of said first waveguides to said ring associated with said waveguide; and a second waveguide crossing said plurality of first waveguides in the vicinity of one of said rings such that light is coupled from at least one of said rings into said second waveguide; wherein at least one of said at least one segment is distensible such that said resonant frequency of said ring having said distended segment is adjustable.

31. A tunable demultiplexer according to claim 30 and wherein said demultiplexer is adapted as the output to a wavelength selective switching network.

32. A tunable demultiplexer according to claim 30 and wherein said demultiplexer is adapted as the output to a drop and add switching network.

33. A tunable demultiplexer according to any of claims 30 to 32 and wherein at least one of said waveguides and said ring comprises a material selected from the group consisting of silicon, silicon dioxide, lithium niobate, gallium arsenide, gallium nitride and a III-V semiconductor.

34. An optical dispersion compensator comprising a substrate comprising; an input multiplexer; an output demltiplexer; a network of switches comprising an array of input waveguides connected to outputs of said multiplexer, and an array of output waveguides connected to intputs of said demultiplexer; wherein said network of switches comprises at least one switch operated by an in-plane microactuator.

35. An optical dispersion compensator according to claim 34 and wherein at least one of said multiplexer and said demultiplexer is tunable.

36. An optical dispersion compensator according to claim 34 and wherein at least one of said multiplexer and said demultiplexer comprises an array of waveguide resonator rings, at least one of said resonator rings having at least one segment suspended from said substrate and distensible by means of an in-plane microactuator

37. An optical dispersion compensator comprising a substrate, comprising: a first waveguide for inputting an optical wave; a first junction in said first waveguide, first and second branch waveguides issuing from said junction, such that light input into said first waveguide is split into said branch waveguides, at least one of said branch waveguides having at least one segment suspended from said substrate; a second junction recombining said branch waveguidess into an output waveguide; wherein said at least one segment is distensible such that the optical path difference between said branches is adjustable.

38. An optical dispersion compensator according to claim 37 and also comprising a polarized beam splitter disposed at said first junction, adapted such that light passing down said first branch waveguide and said second branch waveguide have generally orthogonal polarizations, so that polarization mode dispersion is reduced.

39. An optical dispersion compensator according to claim 37 and also comprising a wavelength splitter disposed at said first junction, adapted such that light passing down said first branch waveguide and said second branch waveguide have different wavelengths, so that chromatic dispersion is reduced.

40. An optical dispersion compensator according to any of claims 37 to 39, and also comprising an in-plane microactuator attached to said at least one segment for distending said segment.

41. An optical dispersion compensator according to claim 40 and wherein said at least one segment is distended by tension such that its length changes.

42. An optical dispersion compensator according to claim 40 and wherein said branch waveguide comprises a material displaying a photo-elastic effect, and wherein said microactuator is operative to stress said branch waveguide such that said photoelastic effect causes the refractive index in said branch waveguide to change.

43. A micromechanical switching mechanism comprising: a microelectronic substrate; a first and a second flexible element on said substrate, at least one of said elements having a first end fixed to said substrate, and each of said elements having a second end free to move generally parallel to the plane of said substrate; each of said flexible elements being associated with an actuator operative to impart movement to said second free end of each element in a direction generally parallel to the plane of said substrate; and a mechanical latching mechanism associated with at least one of said second free ends such that said at least one second free end may be mechanically latched; wherein said mechanically latched mechanism is activated by operation of said actuators according to a predetermined sequence.

44. A micromechanical switching mechanism according to claim 43, wherein: said first element has its first end fixed to said substrate, and said second element has its first end fixed to said second free end of said first element, and its second free end capable of being mechanically latched to a fixed object in said substrate.

45. A micromechanical switching mechanism according to claim 43, wherein: said first element and said second element both have their first ends fixed to said substrate, and said latching mechanism is such as to latch said second free end of said first element to said second free end of said second element.

46. A micromechanical switching mechanism according to any of claims 43 to 45, wherein said predetermined sequence is such that said actuator associated with said first element imparts movement to said free end of said first element before said actuator associated with said second element imparts movement to said free end of said second element, such that said latched mechanism operates correctly.

47. A micromechanical switching mechanism according to claim 46, wherein said predetermined sequence is generated by means of electronic circuits wherein said actuators are activated with differing time delays from the application of an activating signal to said circuits.

48. A micromechanical switching mechanism according to claim 47, wherein said time delays are generated by means of resistor-capacitor networks with different time constants.

49. A micromechanical switching mechanism according to claim 48, wherein said electronic circuits are operative to reverse said resistor-capacitor networks in order to provide reversed time delays to said actuators for unlatching said micromechanical switching mechanism.