US 20090290835 A1 Abstract A resonator structure includes an input waveguide and an output waveguide. In one embodiment, the resonator structure also includes at least one resonator that couples the input waveguide to the output waveguide and a directional coupler that optically couples the input waveguide to the output waveguide. In another embodiment, the resonator structure includes a plurality of ring resonators that couple the input waveguide to the output waveguide. The plurality of ring resonators include a sequence of ring resonators that form a coupling loop. Each ring resonator in the sequence is coupled to at least two other ring resonators in the sequence and the first ring resonator in the sequence is coupled to the last ring resonator in the sequence so as to form the coupling loop.
Claims(11) 1-18. (canceled)19. An optical resonator structure, comprising:
an input waveguide; an output waveguide; at least one resonator coupling the input waveguide to the output waveguide; and a directional coupler optically coupling the input waveguide to the output waveguide. 20. The structure of 21. The structure of 22. The structure of 23. The structure of 24. The structure of 25. The structure of 26. The structure of 27. The structure of 28. The structure of Description The invention generally relates to optical resonator structures. More particularly, the invention relates to integrated optical (microphotonic) resonant structures that comprise loop-coupled cavities having an associated loop coupling phase, and may include, for example, microring resonators, photonic crystal microcavities, and standing wave dielectric resonators. Coupled resonators have previously been used in optics, and particularly in integrated optics, for the design of various amplitude and phase filtering structures. Previous work on coupled-cavity resonant structures includes: series-coupled-cavity (SCC) structures ( Series-coupled-cavity structures (also referred to as coupled-resonator optical waveguides—CROWs—in slow-light literature), such as the structure With reference to With reference to Referring to The optical structures of the present invention solve the problem of providing transmission zeros, with positions in frequency (or, more generally, on the complex-frequency plane) controllable in design, in the drop-port response(s) of a resonant structure. The inventive optical structures do not require the use of 3 dB couplers or any strong direct coupling between the input and output waveguide, and therefore have spectral responses, including rejection ratios, that are highly insensitive to fabrication errors and design nonidealities, in contrast to AD structures. The inventive structures also allow non-minimum-phase responses, including flat-top, linear-phase filters, unlike SCC structures which are limited to all-pole responses. In one embodiment, the structures of the present invention, which include at least an input port and a drop port, enable the design of spectral responses with transmission zeros in the drop port (at real and complex frequency detunings from resonance). This enables optimally sharp filter responses (using real-frequency zeros), and dispersion-engineered spectral responses (using complex-frequency zeros). The latter responses are non-minimum-phase and therefore are not subject to the well known Kramers-Kronig (Hilbert transform) constraint between the amplitude and phase spectral responses. In particular, the structures of the present invention permit the design of spectral responses with passbands having a flat-top amplitude response at the same time as a nearly linear phase response, without the need for additional dispersion compensation following the structure (e.g., by all-pass filters). Such responses are optimal, in the sense that a minimum number of resonant cavities are used for a given amplitude response selectivity and phase linearity. Applications of the disclosed structures include: i) channel add-drop filters for high spectral efficiency photonic networks for telecommunication applications as well as for intrachip photonic networks for next-generation microprocessors; ii) dispersion-compensated filters; iii) slow-wave resonator-based structures for sensors, channelized modulators, amplifiers, wavelength converters, and coupled-cavity nonlinear optics in general. Another example application is in microwave photonics, where the flat-top, linear-phase microphotonic filters may be used in combination with an optical modulator to replace microwave satellite transponder filters, thereby reducing the size and weight of the payload. The flat-top, linear-phase microphotonic filters may also be used in terrestrial microwave filtering, such as in spectral slicing filters in cellular telephone towers. The optical structures of the present invention generally include one or more optical cavities and at least two waveguides. In one embodiment, an input port and a through (reflection) port are defined in the first waveguide, and a drop (transmission) port is defined in the second waveguide. Each of the first and the second waveguides may be coupled to at least one optical cavity, and the direct coupling between the first and second waveguides, i.e., the optical power coupled at a wavelength far from the resonance frequency of any optical cavity in the system, is, in one embodiment, less than about 50%. Preferably, the direct coupling between the first and second waveguides is less than about 10%. However, the direct coupling between the first and second waveguides may or may not be substantially negligible, as dictated by the particular design. In various embodiments, the optical structures of the present invention include one or more of the following features: -
- 1) The use of cavity modes having at least one node in the spatial electric field pattern at any one time (referred to herein as “high-order cavity modes”).
- 2) A plurality of cavities coupled in a loop, thereby forming a “coupling loop” and defining an associated loop coupling coefficient (LCC).
- 3) For structures exhibiting feature 2), a further defined phase of the LCC, referred to as the loop coupling phase (LCP). The LCP of the inventive structures may be approximately 0, approximately 180 degrees, or any other value between 0 and 360 degrees. The LCP may be chosen by an appropriate arrangement of the cavity mode geometry. In the case of microring resonators, the LCP may be chosen by tilting the geometry of the coupled cavity loop by an appropriate angle in the plane of the resonators.
- 4) A non-zero (non-negligible) direct coupling between the input waveguide and the output waveguide, but still one that is less than about 50%, and is preferably less than about 10%, and even more preferably is less than about 1%.
The development of optical filter designs based on a family of resonator structures incorporating one or more of the above features, and using appropriate energy coupling coefficients between resonant cavities and between cavities and ports, permits the realization of various filter responses having transmission zeros. For example, optimally sharp filters, including elliptic and quasi-elliptic filters, and dispersion engineered filters, including nearly linear phase filters, may be realized. In general, in one aspect, the invention features a loop-coupled resonator structure that includes an input waveguide, an output waveguide, and a plurality of ring resonators that couple the input waveguide to the output waveguide. The plurality of ring resonators include a sequence of ring resonators that form a coupling loop. Each ring resonator in the sequence is coupled to at least two other ring resonators in the sequence and the first ring resonator in the sequence is coupled to the last ring resonator in the sequence so as to form the coupling loop. As used herein, the term “ring resonator” generally refers to any resonator that is formed by wrapping a waveguide into a closed loop. Accordingly, ring resonators include, for example, circular microring resonators and racetrack resonators. In various embodiments of this aspect of the invention, the coupling loop has an associated loop coupling coefficient, which itself has an associated loop coupling phase. The loop coupling phase may be approximately 0 degrees, approximately 180 degrees, or another amount. The coupling loop may include four ring resonators and each of the ring resonators of the coupling loop may include a substantially equal ring radii. In one embodiment, the four ring resonators forming the coupling loop are each centered at a different vertex of a rectangle. In another embodiment, the geometry of the coupling loop is tilted so that the four ring resonators forming the coupling loop are each centered at a different vertex of a parallelogram. For example, the geometry of the coupling loop may be tilted by an angle equal to approximately ⅛ of the guided wavelength of a ring resonator in the coupling loop. In yet another embodiment, half of the plurality of ring resonators are arranged in a first row and half of the plurality of ring resonators are arranged in a second row adjacent to the first row. In such an embodiment, each ring resonator in the first row may be coupled to at least one other ring resonator in the first row and to a ring resonator in the second row, and each ring resonator in the second row may be coupled to at least one other ring resonator in the second row and to a ring resonator in the first row. The inter-row couplings (for example, the inter-row energy coupling coefficients) of the resonators may be weaker than the intra-row couplings of the resonators. At least one of the plurality of ring resonators may be a microring resonator or, alternatively, a racetrack resonator. Moreover, at least one of the plurality of ring resonators may include magnetooptic media. In various embodiments, the coupling loop includes an even number of ring resonators. In such a case, at least one ring resonator (and, more specifically, every ring resonator) in the coupling loop is operated to only propagate light in a single direction within the resonator. In still another embodiment, the output waveguide includes a drop port and the loop-coupled resonator structure has a spectral response that includes transmission zeros in the drop port. In general, in another aspect, the invention features an optical resonator structure that includes an input waveguide, an output waveguide, at least one resonator that couples the input waveguide to the output waveguide, and a directional coupler that optically couples the input waveguide to the output waveguide. In various embodiments of this aspect of the invention there is a phase shift in the input waveguide light propagation between the directional coupler and the at least one resonator, relative to that between the directional coupler and the at least one resonator in the output waveguide. Alternatively, the relative phase shift may be introduced in the output waveguide between the directional coupler and the at least one resonator. Additionally, a length of the input waveguide between the directional coupler and a point at which the at least one resonator couples to the input waveguide may be substantially equal to a length of the output waveguide between the directional coupler and a point at which the at least one resonator couples to the output waveguide. The relative phase shift introduced may be implemented, for example, as a small waveguide length difference (e.g. half of the guided wavelength long, for a 180° relative phase shift) between the input and output waveguide, between the directional coupler and resonators, or as a thermally-actuated phase shifter. In one embodiment, a plurality of resonators couple the input waveguide to the output waveguide. The plurality of resonators may include, for example, a sequence of resonators that form a coupling loop. Each resonator in the sequence may be coupled to at least two other resonators in the sequence and the first resonator in the sequence may be coupled to the last resonator in the sequence so as to form the coupling loop. The optical coupling between the input waveguide and the output waveguide introduced by the directional coupler may be substantially broadband over several passband widths. For example, it may be broadband over at least three passband widths, or over at least ten passband widths. In another embodiment, the output waveguide includes a drop port and the optical resonator structure has a spectral response that includes transmission zeros in the drop port. These and other objects, along with advantages and features of the invention, will become more apparent through reference to the following description, the accompanying drawings, and the claims. Furthermore, it is to be understood that the features of the various embodiments described herein are not mutually exclusive and can exist in various combinations and permutations. In the drawings, like reference characters generally refer to the same parts throughout the different views. Also, the drawings are not necessarily to scale, emphasis instead generally being placed upon illustrating the principles of the invention. In the following description, various embodiments and implementations are described with reference to the following drawings, in which: In general, the present invention pertains to optical-coupled resonator structures that are based on loop-coupled cavities and loop coupling phase. Loop coupling of cavities may be understood as follows. For each coupling loop, a loop coupling coefficient (LCC) may be defined. Two cavity modes may be coupled via evanescent field coupling across a gap or via radiative field coupling via a traveling-wave coupling pathway. Evanescent coupling may be understood as follows. For each pair of mutually coupled cavities, an energy coupling coefficient (ECC), with units of rad/s, represents the rate of energy amplitude coupling in time between the two cavities, and is the standard measure of coupling between cavities under the commonly used coupled mode theory in time formalism. Then, the LCC of a coupling loop is defined as the product of ECCs around the loop. Strictly, the LCC is a vector, with a complex magnitude indicating the product of ECCs, and whose direction indicates the sense in which the loop is traversed according to a preset convention, for example according to the well-known right-hand-rule used in electrical engineering. In the following discussion, the direction of the loop is understood from the illustrations and is omitted in the LCC value, where only the magnitude is considered. For example, if the ECC from cavity i to cavity j is μ Since the LCC is a complex number, a loop-coupling phase (LCP) associated with each LCC may be defined as the phase of the LCC. As with any phase, the LCP has a value modulo 360°, i.e., modulo 2π radians. The LCP has a physical significance, which is illustrated in It should be noted that in the microwave engineering literature, reference is often made to a positive or negative coupling coefficient—this refers to the sign of each ECC, μ The Applicant has observed that it is possible to achieve negative coupling in optical resonators by using high-order resonant modes, i.e., modes with at least one null in the electric field pattern. Then, the positive and negative lobe of the same resonant mode may be coupled to various other modes, and a negative LCC may also be generated, as done in High-order (i.e., second-order) field pattern resonances are used in Electromagnetic reciprocity restricts the LCPs that may be obtained in a resonant system. Structures formed of reciprocal media, if also lossless, have a uniform phase across the cavity field pattern, which can thus be (for standing-wave-cavity modes) assumed a real number distribution with a particular choice of reference phase. Hence, only positive or negative coupling coefficients, and thus only a positive or negative LCC, or 0 or 180° LCP, may be produced. An exception is degenerate resonators, which are addressed below. Non-reciprocal resonators, which include magneto-optic media, as well as reciprocal traveling-wave resonators, permit arbitrary-phase coupling and thus permit an arbitrary LCP. Non-reciprocal lossless structures do not have the restriction of uniform phase across the resonant mode field. Reciprocal traveling-wave (including microring) resonators behave the same way because they have two degenerate resonances. Either two degenerate, orthogonal standing-wave modes (excited 90° out of phase to simulate a traveling-wave resonance), or two (a clockwise and anti-clockwise) propagating traveling-wave modes may be chosen as a basis. When the latter point of view is chosen, arbitrary LCPs may be obtained. Whether LCPs can be arbitrary depends on the choice of basis because the LCC depends on the ECCs, which in turn are defined with respect to the resonant modes; so, the choice of basis of resonant modes is ultimately that with respect to which the LCP has a meaning. This is explained in greater detail below. A second way to obtain the value of the LCP is to begin at any resonator Finally, since tilting the geometry of coupled microring (or other traveling-wave) cavities may change the LCC from positive to negative, tilts other than ⅛ Now, several exemplary loop-coupled structures with various LCP configurations are shown and their properties described. In one embodiment, the structures Then, in this embodiment, the number of transmission zeros introduced into the input With regard to coupling coefficients used to describe the devices of the present invention, two descriptions are used: energy coupling coefficients (ECCs), and power coupling coefficients (PCCs). Each may be converted to the other. ECCs are representative of a general concept, whose definition is well known in the framework of coupled mode theory in time, used to describe the coupling: (a) between a first and a second resonant cavity, or (b) between a resonant cavity and a waveguide or optical access port. Coupling between a resonant cavity mode and a waveguide or port is described by an energy-amplitude decay rate, τ When using traveling-wave cavities, including microring and racetrack resonators, the cavity-cavity and cavity-waveguide coupling are typically achieved by evanescent coupling across directional coupler regions. Directional couplers are described by a PCC, which determines the fraction of power incident into the input port of one waveguide that is coupled to the cross-port, i.e., to the output port in the other waveguide. In this case, the ECCs of a resonant structure can be translated to PCCs corresponding to the directional couplers in the structure. The conversion is done in two steps: ECC to PCC conversion, and PCC correction for finite FSR. We note, as known in the literature, that for purposes of the coupling coefficient formulas the FSR is defined not as the actual frequency spacing between two adjacent resonant orders, but rather as FSR≡v where FSR where FSR Once the PCCs are obtained from the ECCs, a second mapping, also known in the art, is applied to account for the finite FSR of the cavities. Each PCC is replaced by a scaled version of itself as:
In the obtained PCC set, all numbers are unitless power coupling fractions, one per directional coupler. Referring again to
Note that equation (4) holds only for 4-cavity coupling loops, such as those shown in It is of interest to create filters with a full N poles and N zeros (per FSR) with controllable frequency positions, for an N-cavity system, because the sharpest achievable spectral response for bandpass filters is known to be the elliptic function response, which requires N poles and N zeros. Realizing this response function is optimal in the sense of achieving the sharpest filter rolloff with a given number of cavities, i.e., a given order of the resonant system. The previous discussion makes clear that a filter with an equal number of poles and zeros (e.g., four) will have no rolloff at large enough detuning, because all the poles and zeros cancel when observing the passband wavelength from the vantage point of a large wavelength detuning. This is consistent with the spectral shape of an elliptic filter response function, which reaches a constant level at large detuning. Referring now to In one embodiment, the optical coupling between the input waveguide The structures The design parameters of the filter designs shown in For the loop-coupled structure For the loop-coupled structure The described 4 The structures The structure The design of a flat-top, linear-phase, high-order (30 cavities) resonant structure A mapping between electrical resonators and couple mode parameters can be found by physical arguments. The capacitors
for n=1 . . . N/2, where N is the filter order (number of cavities), with coupling coefficient indices in equation (5) referring to the numbering of the cavities The graph of The normalized ECCs (for a 2 rad/s full passband width) for the filter prototype used for the loop-coupled structure With reference again to For a chosen exemplary cavity FSR of 2 THz, typical for a microring resonator, the ECCs given above are translated to power coupling coefficients for the directional couplers. In such an embodiment, the final power coupling coefficients for the loop-coupled structure Referring to In one embodiment, the final power coupling coefficients for the typical maximally-flat SCC filter with a 40 GHz bandwidth and 2 THz FSR are, from input to output: {κ In general, the ECCs, as listed in the previous examples, scale linearly with passband width, so they are scaled to obtain different bandwidths. Equations (1) and (2), set forth above, may be applied to resonators having nonidentical FSRs. In this way, the same design may be scaled to various physical structures. The specific examples given so far have been used to illustrate the utility of loop coupling, and of the engineering of the LCP(s). More generally, coupled-cavity structures that include an input waveguide and at least one output waveguide, and that are connected by a coupled-resonator system with non-trivial coupling loops, will introduce finite, complex-frequency transmission zeros into the input-to-drop response spectrum. Exemplary embodiments of such generic structures are illustrated in In the structure An exemplary physical realization of the 4 A physical realization of a 4 The devices The approach for obtaining N transmission zeros (per FSR) in an N-cavity system by coupling the input and output waveguides directly at a power fraction equal to the desired drop-port transmission level at large detunings is valid generally. More generally, the number of transmission zeros in the drop port is equal to N−M, if N is the number of cavities, and M is the smallest number of coupled cavities that must be traversed in going from the input port to the drop port. In one embodiment, the optical coupling between the input waveguide Referring to In Optical hybrid structures may similarly be created from other loop-coupled standing-wave cavity structures by converting each of the input and output waveguides from a waveguide that ends at the resonator, having substantially high reflection (ideally approaching 100%) when none of the resonators are excited (and thus having the through port as the reflected wave relative to the input port), to a waveguide with separated ports having no substantial reflection, and using two copies of the standing-wave loop coupled structure, the two copies being coupled to each other at those cavities that are coupled to access waveguides, and optionally having other cross-coupling. The optical hybrid, besides having a non-reflecting waveguide, in general has twice as many coupling loops as the non-hybrid standing-wave structure. Thus, the embodiments provided in the invention describe standing-wave structures, such as the structure In summary, the structures presented herein enable filter responses that have, in certain embodiments, transmission-response (drop-port) zeros at complex-frequency detunings. The structures may include an input waveguide, at least one output waveguide, and a coupled-cavity system coupling the input and output waveguides. The coupled-cavity system may include cavity coupling loops, each with a loop-coupling phase determined by the geometry of the coupling loop. The loop coupling phase may be approximately 0 or 180° for standing wave resonators using high-order spatial modes, and may take on arbitrary values for reciprocal microring resonator structures operated with unidirectional excitation, or when standing- or traveling-wave resonators made of non-reciprocal (magnetooptic) media are used. By designing coupling loops with LCPs, and with the optional addition of a direct, phase-aligned coupling between the input and output port, the control of N poles and N zeros (per FSR) may be obtained in an N cavity filter. This allows for the design of optimum filters—in the sense of optimally sharp amplitude responses and non-minimum-phase designs permitting linear-phase or phase-engineered passbands—and for optical signal processing structures in a compact and robust implementation. Having described certain embodiments of the invention, it will be apparent to those of ordinary skill in the art that other embodiments incorporating the concepts disclosed herein may be used without departing from the spirit and scope of the invention. Accordingly, the described embodiments are to be considered in all respects as only illustrative and not restrictive. Referenced by
Classifications
Legal Events
Rotate |