Open Access
Add to BibSonomy
Abstract
Optical fibers are generally polarization-insensitive while photonic integrated circuits (PICs) often exhibit a large polarization dependence due to the high-aspect-ratio and high-index-contrast of integrated waveguides. As PICs become more mature there is an increasing need for tunable polarization management on-chip. Although micro-electro-mechanical systems (MEMS) are increasingly finding application in PICs for optical switching and phase shifting, they have so far not found wide application for polarization management. In this work we propose two optical MEMS architectures for polarization management enabling tunable polarization splitting and rotation – key functions so far lacking in PICs. The first structure consists of a directional coupler with a MEMS-tunable gap enabling a continuously-variable polarization splitting ratio. A second architecture consists of a symmetry-breaking MEMS perturber suspended over an air-cladded waveguide enabling tunable polarization rotation. For both architectures we simulate a polarization extinction exceeding 25 dB, and the operating bandwidth can be as large as 100 nm. We conclude with a discussion of actuation schemes and examine fabrication considerations for implementation in PIC foundries.
1. Introduction
Photonic integrated circuits (PICs) are increasingly used in data-/telecom [1], secure quantum communications [2], and chemical [3] and inertial sensing [4]. Wafer-scale fabrication of PICs is enabled by the emergence of optical foundries [5]. Many applications require that PICs are connected to optical fiber networks in which the polarization is unknown. Conversely, PICs generally include waveguides with a large index contrast (e.g., silicon / silicon dioxide) and an asymmetric cross-section that results in a large polarization dependence. Although polarization-maintaining (PM) fiber exists, most deployed optical fiber networks are not PM. Therefore, large-scale deployment of PICs is limited by the need for on-chip polarization management.
One approach to polarization management in PICs is polarization diversity in which an input signal is assumed to be a superposition of the two linear polarization states, TE and TM. An integrated polarization splitter [6–8] separates TE and TM into different waveguides so that TE and TM signals are sent to on-chip components that are optimized for that specific polarization. A drawback, however, is the increased complexity, optical bandwidth reduction, and increased loss with increasing number of components required. Consequently, PICs also require polarization rotation [9–14] to enable true polarization-independence [15]. Indeed, a combined polarization splitter-rotator [11] was demonstrated in a foundry process in Ref. [16].
While there has been substantial work on fixed polarization management devices, there have been few reports of tunable polarization devices. The ability to dynamically-tune the polarization enables greater flexibility and simplifies the development of large-scale PICs in which many components are present. Polarization can be illustrated by the commonly-used Poincaré Sphere [17] in Fig. 1(a). This visual representation describes linear and elliptically-polarized light, although integrated waveguides generally only allow for linearly-polarized light (quasi-TE and -TM, or mixtures of the linear polarization states) to propagate on chip. Assuming no ellipticity, the polarization is described by the polarization angle, ψ, lying in a plane of the Poincaré Sphere where S3 = sin(2χ) = 0 [Fig. 1(b)]. For polarization splitting light is decomposed into TE and TM with spherical coordinates (Stokes parameters) S1 = cos(2ψ) and S2 = sin(2ψ). Although waveguide modes can be hybridized TE/TM, for simplicity we assume that TE has ψ=0° and TM has ψ=90° as in Fig. 1(b). Polarization rotation requires changing the azimuthal angle ψ.
Fig. 1. (a) Poincaré Sphere illustrating polarization transformation. (b) For linearly-polarized light (i.e., TE and TM) there is no ellipticity (χ=0) and the polarization is governed only by the azimuth angle (ψ). For ψ=0° light is TE-polarized, and for ψ=90° it is TM. Tunable polarization rotation is enabled by changing ψ.
In this paper we describe two micro-electro-mechanical systems (MEMS)-based approaches for tunable polarization control. The first structure enables MEMS-tunable polarization beam splitting (PBS) while a second device enables MEMS-tunable polarization rotation (PR). We first review optical MEMS (section 2) and tunable polarization devices (section 3) to give some perspective. Next, we introduce our approach for MEMS-tunable polarization beam splitting and investigate some performance characteristics based on coupled-mode theory and beam propagation method (3D-BPM) simulations (section 4). Section 5 introduces an approach for tunable polarization rotation based on symmetry-breaking mode perturbation. Polarization angle rotation, insertion loss, and bandwidth are investigated using finite-difference-time domain (3D-FDTD) simulations. We conclude with a discussion of MEMS actuation and foundry integration (section 6).
2. Optical micro-electro-mechanical systems (MEMS)
2.1 Electrostatically-actuated optical MEMS
Micro-electro-mechanical systems (MEMS) [18] enable chip-scale mechanical structures whose displacements (typically <10 µm) are similar to the optical wavelength. MEMS are therefore especially useful for phase shifting and switching in PICs. At the same time, MEMS enable low-power operation since capacitive electrostatic actuators require little current (determined by the displacement current) and enable higher speeds (10’s-100 MHz) compared to thermo-optic devices. Early demonstrations focused on optical switching using electrostatically-actuated waveguides [19–21]. Other demonstrations include tunable surface-normal MEMS-Fabry-Perot cavities [22–24] enabling a record tuning range of 140 nm [24] and 150 nm [25]. A different approach uses MEMS for mode perturbation and phase shifting as has been demonstrated by many groups [26–30] enabling integration with PICs for active tuning and passive sensing in, e.g., accelerometers [31].
2.2 Optical force-actuated MEMS
Optical forces result from a change in photon momentum via radiation pressure [32] or gradient optical forces [33]; photothermal forces can also be exploited [34]. Although these forces have led to many interesting fundamental studies, they also have potential application in PICs. All-optical tuning of photonic crystal cavity filters has been demonstrated [35]. Similarly, tuning of a stacked microring cavity was demonstrated [36]. Optical forces can also be used for mechanical resonators [37] resulting in self-oscillation [38–40]. These optomechanical oscillators may find applications in radio-frequency (RF) and microwave photonics.
2.3 Optical MEMS at NRL
In-plane, waveguide-coupled cavities enable tunable filters to be integrated in PICs. To this end we demonstrated waveguide-integrated tunable Fabry-Perot [41] and MEMS-microring cavities [42]. We also reported tunable mode-order conversion [43] for on-chip mode-division multiplexing. Beyond electrostatic actuators, optical forces and optomechanical self-oscillation can resemble an optically-pumped laser [44] in terms of linewidth reduction and oscillation amplitude. Small amounts of absorbed light lead to measurable frequency shifts in optically-pumped oscillators enabling bolometers for, e.g., absorption spectroscopy [45] or imaging.
Clearly, optical MEMS can have many applications and the integration of MEMS in PICs will especially benefit scenarios in which size, weight, and power are key. Increasingly, MEMS efforts have also focused on optical foundry integration [29,46]. So far, however, optical MEMS have not been widely pursued for polarization control in PICs – despite MEMS being well-suited for both tunable polarization splitting and rotation, as we will show.
3. Tunable polarization management in PICs
One of the first tunable polarization rotators used lithium niobate (LiNbO3) waveguides with electro-optic phase shifters enabling tunable TE0- and TM0-mode coupling [47]. However, this approach required long device lengths (millimeters) and enabled polarization rotation for only narrow bandwidths (0.5-5 nm). A foundry-compatible tunable polarization rotator was demonstrated in [48] using alternating silicon waveguide fixed polarization rotator segments and thermo-optic (TO) phase shifters. The polarization-dependent TO phase shift enabled continuously-tunable polarization rotation. However, the approach required long phase shift lengths (1 mm) resulting in a large device footprint. A third approach utilized a structure with a waveguide segment deflecting out of plane thereby inducing a polarization rotation [49]. Although this approach is in principal broadband, the limited out-of-plane deflection required a microring cavity to enhance the rotation angle resulting in a bandwidth that is limited by the cavity resonance. The proposed polarization management devices in this paper aim to improve on performance (i.e., bandwidth up to 100 nm) and footprint (device length <100 µm) for PICs.
4. MEMS-tunable polarization beam splitter (PBS)
We consider a directional coupler consisting of two identical silicon waveguides with thickness tSi = 220 nm and width wSi = 450 nm as shown in Fig. 2(a),(b). The waveguides are suspended in air and do not have an SiO2 cladding; this suspended waveguide architecture enables a variety of photonic structures to be realized [50]. The coupled-waveguide system supports even and odd modes (simulated using Comsol [51]) having effective indices neven, TE = 2.391, nodd, TE = 2.382, neven, TM = 1.859, and nodd, TM = 1.813 at a waveguide separation or gap = 376 nm and wavelength λ=1550 nm. The coupling length for complete power transfer from the input THRU to the output CROSS waveguide is lc = λ/[2(neven – nodd)] which results in calculated coupling lengths lc≈90 µm for TE0 and lc≈16.7 µm for TM0 (gap = 376 nm). Due to the polarization-dependent coupling strength and lengths it is possible to find conditions at which TE- and TM-polarizations are split between the CROSS [Fig. 2(a)] and THRU [Fig. 2(b)] outputs. Figure 2(c) shows beam propagation method simulations (3D-BPM) for a coupler with gap = 376 nm and a coupling length lc = 100 µm. We chose to fix the coupling length at 100 µm while optimizing the initial coupling resulting in gap = 376 nm. The simulations show that the directional coupler acts as a polarization splitter with TE0 λ=1550 nm light sent to the CROSS port and TM0 remaining in the THRU port. The calculated (coupled-mode theory, or CMT) lc for TE and TM are also indicated (red, blue points) showing good agreement between calculation and 3D-BPM simulation in terms of peak THRU and CROSS transmission.
Fig. 2. Polarization splitting in a suspended waveguide silicon directional coupler: (a) beam propagation method (3D-BPM [52]) simulation, TE λ=1550 nm response, (b) TM λ=1550 nm, (c) THRU and CROSS outputs vs. propagation distance for TE- and TM-polarizations.
We investigate the bandwidth of the fixed polarization beam splitter by performing 3D-BPM simulations [52] over a wavelength range λ=1500-1600 nm [Fig. 3(a)]. Since the TE and TM polarizations have very different coupling conditions, the optimal polarization splitting is limited in bandwidth unless specialized structures using, e.g., bent waveguides are used [8]. Although the maximum polarization extinction ratio (PER) is ≈23 dB at λ=1550 nm, the -20 dB bandwidth is only BWPER = 20 dB≈3.5 nm and the -10 dB bandwidth is BWPER = 10 dB≈16 nm [Fig. 3(b)]. Clearly, the bandwidth of the fixed polarization splitter is limited and a tunable splitter would enable optimal performance over a larger operating bandwidth.
The coupler mode effective indices depend on the gap, especially for TM-polarized modes since these are less tightly confined than TE. We compute the even and odd mode neff as a function of gap using a 3D-mode solver [Fig. 4(a)]. The lowest-order coupling length, lc(gap), is then computed from the neff, and the directional coupler transfer function is THRU(gap)=cos2(πldevice/2lc) and CROSS(gap)=sin2(πldevice/2lc) for a device length ldevice = 100 µm. The results indicate a strong polarization-dependent response for the MEMS-tunable coupler in which at, e.g., gap = 376 nm the TE CROSS output can exhibit a maximum while TM is minimized; similarly, at gap = 304 nm the TE CROSS output is minimized while TM is maximized [Fig. 4(b)]. The MEMS-tunable coupler response computed from coupled-mode theory is compared with beam-propagation method (3D-BPM) simulations showing good agreement [Fig. 4(b), points] in terms of MEMS gap-dependent response. We find some variation in extinction between the two approaches (theory vs. simulation) that can be explained by the finite mesh size and simulation window which may affect the computed mode overlap in the 3D-BPM simulation. From the polarization-dependent CROSS output we calculate the MEMS-tunable polarization extinction from the ratio between TM0 and TE0 output as shown in Fig. 4(c). The results show that the TM polarization extinction in the CROSS output can be tuned from PER=+30 dB at gap = 304 nm (dashed circle 1) to PER = -23 dB at gap = 376 nm (dashed circle 2). For other configurations, e.g., gap = 412 nm (dashed circle 3) the structure exhibits equal (i.e., polarization-independent) splitting. The actuation distance required can be readily achieved using MEMS electrostatic actuation (see section 6.1 MEMS Actuation below for details on the actuator design and actuation mechanism). Beyond tuning the extinction, adjusting the waveguide gap can also compensate for the limited polarization splitting bandwidth shown in Fig. 3 so that the maximum PER is maintained at an arbitrary wavelength. Although we focus here on a silicon device, similar performance can be achieved using other material systems (e.g., silicon nitride) provided that sufficient birefringence can be obtained to enable polarization-dependent coupling between the waveguides.
Fig. 4. MEMS-tunable polarization beam splitting: (a) even and odd mode effective index simulation vs. gap, (b) CROSS output CMT calculation vs. 3D-BPM simulation of an ldevice = 100 µm long directional coupler, (c) extracted MEMS-tunable polarization extinction in the CROSS output (3D-BPM). The insets in (c) show that at gap = 304 nm (circle 1) the TE0-mode remains in the input waveguide (THRU) while TM0 is coupled to the second (CROSS) waveguide; for gap = 376 nm TE is in the CROSS and TM is in the THRU output (circle 2); at gap = 412 nm (circle 3) there is equal TE0- and TM0-power in both waveguides.
5. MEMS-tunable polarization rotator (PR)
5.1 Basic design approach
Symmetry-breaking waveguides with an asymmetric cross-section can enable coupling between TE- and TM-polarized modes. We start off with a silicon waveguide with a cross-section tSi = 220 nm and wSi = 450 nm that supports TE0- and TM0-modes at λ=1550 nm [Fig. 5(a),(b)]. Although this initial design is for a fixed rotator, we consider a design with an air top cladding and an SiO2 bottom cladding to enable MEMS-tuning, as will be discussed subsequently. The simplest method to achieve a symmetry-breaking structure is to modify the basic waveguide design with a second silicon layer (perturber) with cross-section tSi = 220 nm and wSi = 225 nm covering half the width of the waveguide [Fig. 5(c),(d)]. This is equivalent to considering a waveguide with square cross-section and etching a channel into one half of the waveguide as was done by others [53]. The polarization angle, ψ, is found from [53]: tan(ψ) = ${\int\!\!\!\int {n\left( {x,y} \right)} ^2}{e_x}{\left( {x,y} \right)^2}dxdy/\int\!\!\!\int n {\left( {x,y} \right)^2}{e_y}{\left( {x,y} \right)^2}dxdy$, where n(x,y) is the index distribution of the waveguide, and ex and ey are the electric field distributions. We investigate this structure using finite-difference-time-domain (3D-FDTD) simulations [54]. The computed TE0-mode is launched into the structure and the output field is compared to the TE0- and TM0-mode profiles using a mode-overlap calculation. We repeated the simulation for perturber lengths lp = 1-7 µm and find that for lp = 4.1 µm the structure enables full TE-to-TM polarization rotation; i.e., a launched TE0-mode is rotated 90° and exits as a TM0-mode. Although the polarization extinction is >20 dB, the insertion loss is 3.5 dB for this fixed polarization rotator [Fig. 5(e)].
Fig. 5. Fixed polarization rotator: (a) simulated waveguide TE0-mode in the absence of a perturber (λ=1550 nm), (b) TM0 waveguide mode (no perturber), (c) top-view of fixed polarization rotator with symmetry-breaking overlay perturber, (d) cross-section, (e) 3D-FDTD simulation [54] of polarization rotation of an injected TE0-mode with full TE0-TM0 rotation for lc = 4.1 µm. The Ex and Ey fields in (a) and (b) refer to the dominant electric field component for TE- and TM-polarized modes, respectively.
Next, we considered the same basic structure as in Fig. 5 but now the MEMS perturber can be actuated vertically to create an adjustable air gap between the waveguide and the MEMS perturber [ Fig. 6(a)]. In this manner, the interaction strength between the propagating mode in the waveguide and the MEMS structure can be adjusted to tune the amount of polarization rotation. Figure 6(b) shows the 3D-FDTD simulation results for a basic MEMS-tunable polarization rotator with perturber lp = 4.1 µm. The results show that this structure can tune the polarization rotation so that a launched TE0-mode can be converted to TM0 or remain TE0 at the output, and the extinction can be tuned from PER=+27 dB to -23 dB. These simulations confirm the viability of this MEMS-tunable polarization rotation approach. However, as illustrated in Fig. 5(e), the high insertion loss (3.5 dB) is potentially too large for many applications. For this reason, we investigated a second design enabling a lower insertion loss.
Fig. 6. Basic MEMS-tunable polarization rotator: (a) cross-section indicating the variable gap separating the Si-MEMS perturber and the Si waveguide, (b) 3D-FDTD simulations demonstrating MEMS-tunable polarization rotation of a launched TE0-mode.
5.2 Improving the insertion loss
Rather than using an abrupt transition between the waveguide and the symmetry-breaking region, an adiabatic taper can be used. This approach results in a gradual mode evolution [9] and polarization rotation as a launched mode propagates down the structure. Our design is shown in Fig. 7(a) and consists of the base waveguide (tSi = 220 nm, wSi = 450 nm), a slanted perturber polarization rotator (tSi = 220 nm, wSi = 0-450 nm), and a mode width taper (tSi = 220 nm, wSi = 450 nm). The purpose of the slanted rotator perturber segment is to adiabatically rotate the mode while the width taper enables a low-loss mode transition from the polarization rotation segment to the base waveguide. Combined, these features aim to minimize the insertion loss of the polarization rotator. As with the basic rotator design in Fig. 5 we simulate the device performance using 3D-FDTD simulations [54]. The width taper length was fixed at ltaper = 20 µm; previous simulations have shown that this length is a good compromise between ensuring an adiabatic and low-loss transition to the base waveguide while also limiting the total device length and thus simulation time. We varied the slanted polarization rotation perturber length from ltaper = 5-15 µm and simulated the device to compute the TE0- and TM0-mode overlap at the output assuming a TE0 λ=1550 nm input. The results indicate a strong reduction in insertion loss from 3.5 dB in the basic design [Fig. 5(e)] to ≈0.1 dB for the low-loss design with a polarization rotation length lp = 10.5 µm [Fig. 7(b)]. We attribute the large reduction in loss to the slanted rotator segment and the resulting adiabatic mode evolution as well as the width taper that enables a gradual transition between the polarization rotator to the base waveguide. Although the simulations assume no sidewall roughness, we expect minimal additional loss in a fabricated device due to the complete selectivity in etching the SiO2 sacrificial layer vs. Si.
Fig. 7. Low-loss fixed polarization rotator: (a) top-view and cross-section view of device with polarization rotator and mode taper sections, (b) simulated polarization rotation of an input TE0-mode (λ=1550 nm).
We investigated the ability of this low-loss design to enable MEMS-tunable polarization rotation using the same approach as for the basic design in Fig. 5 using 3D-FDTD simulations. For the optimized design the rotator length was lp = 10.5 µm and the mode taper was ltaper = 20 µm. We launched an input TE0-mode (λ=1550 nm) and computed the TE0- and TM0-overlap integrals at the device output as the MEMS gap was varied [Fig. 8(a),(b)]. The simulations show that the structure enables complete TE-to-TM rotation (gap = 0 nm) with a polarization extinction PER = 15 dB. As the MEMS structure is raised (increasing gap) there is a continuous decrease in the PER indicating decreasing polarization rotation. For gap ≥ 18 nm we observe a slight decrease in the polarization extinction from -40.5 dB (gap ≥ 18 nm) to -38.1 dB (gap = 20 nm). Although further investigation is needed, simulations have shown that the polarization rotator can exhibit a periodic response with MEMS-waveguide gap. Nonetheless, the results in Fig. 8 confirm MEMS-tunable polarization rotation in a low-loss architecture for photonic integrated circuits. Even though the evanescent field of the waveguide modes shown in Fig. 5(a),(b) extend well beyond the silicon waveguide surface, we note that the operating gaps are small (0-20 nm) since the polarization rotator was designed to be as short as possible therefore requiring a very strong MEMS perturbation (lp = 10.5 µm for gap = 0 nm and complete TE-TM rotation; see Fig. 7(b)). The use of other materials and layer thicknesses may enable a relaxed constraint on the MEMS-waveguide gap, as discussed below.
Fig. 8. Low-loss MEMS-tunable polarization rotator: (a) 3D-FDTD simulation [54] of the polarization rotation TM-extinction vs. MEMS gap for a TE input, (b) complete TE-to-TM rotation at gap = 0 nm showing that the output field is Ey (all-TM), (c) no rotation case at gap = 20 nm with Ex at the output (all-TE). The circled gap = 0 nm point corresponds to the simulation in (b) while the circled gap = 20 nm corresponds to (c).
Next, we investigated the operating bandwidth of the basic and low-loss polarization rotator designs. The optimized structures feature a polarization rotator segment lp = 4.1 µm for the basic design from Fig. 5; for the low-loss design the rotator was lp = 10.5 µm and the mode taper was ltaper = 20 µm as in Fig. 7. For both designs we performed 3D-FDTD simulations on the fixed polarization rotation device (gap = 0 nm) for wavelengths λ=1440-1600 nm as shown in Fig. 9. The results show that the low-loss design has a substantially larger operating bandwidth of 150 nm for a 2 dB PER variation. In contrast, the basic design’s bandwidth is <50 nm for 2 dB PER variation. It should be noted that the basic design appears to have a higher maximum PER of 20 dB compared to the low-loss design of ≈16 dB. However, this may be due at least in part to the simulation window size in which the low-loss design’s device length (> 30 µm) compared to the basic design (<6 µm). Full 3D-FDTD simulations can lead to accumulating errors that are compounded for larger simulation windows and times. Nonetheless, the low-loss design appears to offer a larger PER bandwidth in addition to lower insertion loss.
Fig. 9. Polarization-rotator bandwidth: (a) basic design with lp = 4.1 µm, (b) los-loss design with lp = 10.5 µm and ltaper = 20 µm, (c) 3D-FDTD simulated polarization rotation bandwidth for basic vs. low-loss designs.
6. Discussion
6.1 MEMS actuation of polarization beam splitter (PBS) and polarization rotator (PR)
Actuation of MEMS structures can be achieved using capacitive electrostatic actuation [55], gradient electric fields [46], optical forces [45], and electro-thermal forces [43]. For the present discussion optical forces will not be considered as these forces generally do not result in large enough displacements for polarization splitting and/or rotation. Electrostatic actuation, in particular, lends itself well to the MEMS-tunable directional coupler in Fig. 4 and can be simulated by taking advantage of symmetry where each waveguide is biased at V + (V-) resulting in a virtual ground plane between the two waveguides [dashed line in Fig. 10(a)]. Alternatively, electrodes can be placed on the opposite side of the waveguide enabling actuation towards increasing gaps [Fig. 10(b)]. Note that the waveguides only need to be lightly-doped for actuation resulting in negligible optical loss, and no metal on the waveguide itself is required. The electrostatic force and resulting displacement can then be calculated [56] or simulated. We simulated the MEMS-tunable polarization splitter from Fig. 4 using Comsol [51]. The results in Fig. 10(c) show that the device can be actuated over gap = 376-304 nm (device length, lc = 100 µm) resulting in complete tunability of the polarization splitting for a bias = 0-1 V. Although the actuator is near the pull-in condition at gap = 304 nm [Fig. 10(c)], the “push-pull” actuation enabled by the configurations in Fig. 10(a),(b) can enable full tunablity of the MEMS-PBS since the gap would need to be varied only by +/-36 nm (pre-pull-in) as opposed to 72 nm (near pull-in).
Fig. 10. (a) Electrostatic actuation of MEMS polarization splitter. (b) Alternative electrostatic actuation scheme. (c) Simulated waveguide displacement vs. bias for configuration in (a). (d) Gradient electric force actuation (adapted from Ref. [55]). (e) Electro-thermal actuation (Ref. [43]). While the actuator in (a) and (b) is more suited to the MEMS-PBS, the actuators in (d) and (e) are more applicable to a MEMS-PR.
As mentioned, electrostatic actuators have a limited travel range (pull-in [56]) that limits their use in structures requiring continuous travel as in the MEMS-tunable polarization rotator from Fig. 7. Other electric actuators that can be applied to the MEMS-tunable polarization structure include gradient electric force actuators [Fig. 10(d)] in which a gradient electric field polarizes the MEMS perturber resulting in a force that actuates the MEMS downwards (as demonstrated in [55]). An alternative approach is to take advantage of thermal expansion in a metal-dielectric/semiconductor bi-morph resulting in an electro-thermal actuator Fig. 10(e)] (as in [43]). While electro-thermal actuators require more electrical power than electrostatic and gradient electric force actuators, they enable continuous actuation over large distances.
6.2 Processing and foundry compatibility
The MEMS-tunable directional coupler polarization beam splitter (Fig. 4) was designed to be foundry-compatible [5]. The silicon thickness and width (tSi = 220 nm and wSi = 450 nm) are commonly-used waveguide dimensions, although the suspended design is not. However, process customization including opening a trench to the silicon waveguide layer is offered as an add on in some foundries [57]. This leaves a release etch of the SiO2 surrounding the silicon to be completed in order to suspend the directional coupler waveguides. This is achieved using a BOE etch as has been performed for previous foundry-processed optical MEMS PICs [46]. Ongoing work may enable this release etch to be integrated in a foundry process in the future.
The MEMS-tunable polarization rotator as currently designed [Figs. 7,8] is more challenging to implement in a foundry process. While the base waveguide is standard (tSi = 220 nm and wSi = 450 nm), the MEMS perturber suspended above the waveguide is not. A double silicon-on-insulator wafer (i.e., one with two silicon device layers of tSi = 220 nm separated by an SiO2 spacer layer) is technically feasible, although these wafers are not currently standard. An alternate approach would be to fabricate the waveguide in a foundry and to then post-process the MEMS perturber and mode taper structures by depositing, e.g., a polysilicon MEMS layer and SiO2 sacrificial layer. The poly-Si MEMS would then be patterned and etched, and the MEMS structure would be released using BOE. Due to the need for a second Si layer the MEMS-tunable polarization rotator would likely still require a custom fabrication sequence that complicates foundry integration. As an alternative, stacked silicon nitride (SiN) waveguides are already available in some photonic foundries [5] and would enable one SiN layer to serve as the waveguide and the second SiN layer as the polarization rotation structure. All-dielectric structures (e.g., SiN) can be actuated using gradient electric forces [42,46,55].
Given the small feature sizes of the waveguide and MEMS structures (widths w < 500 nm), alignment becomes important. The MEMS-tunable polarization beam splitter (PBS) consists of suspended waveguides that are fabricated in the same device layer. However, the MEMS-tunable polarization rotator (PR) consists of separate waveguide and MEMS perturber layers that need to be precisely aligned. We performed simulations of the fixed polarization rotator in Fig. 7 but with a lateral offset of the MEMS perturber with respect to the waveguide and found that offets of +/-10 nm (a reasonable alignment tolerance between layers) result in changes in the polarization rotation angle of <5%. Although additional investigation is needed, lower index materials (e.g., silicon nitride) will presumably result in larger feature sizes and a relaxed alignment tolerance.
6.3 Actuation distance
The MEMS-tunable polarization beam splitter operates in a regime that is well-suited to optical MEMS (gap = 300-400 nm during typical operation). Such actuation distances (Δgap≈100 nm) can be readily achieved using, e.g., electrostatic actuators. By using lightly-doped silicon the suspended waveguides can be biased to enable capacitive electrostatic actuation.
Actuation of the MEMS-tunable polarization rotator is more challenging due to the small air gap (0-20 nm) separating the waveguide and the symmetry-breaking MEMS perturber. Two approaches can be taken here: 1) utilize a 20 nm SiO2 spacer that is etched during the MEMS release process, 2) utilize a thicker sacrificial layer/spacer (e.g., 100 nm SiO2) to ease fabrication requirements at the expense of requiring a larger actuation distance. Electrostatic actuation exhibits a nonlinear displacement with bias and has a well-known instability (“pull-in”) [56] that limits the actuation distance to ≈1/3 the initial gap. In contrast, electro-thermal actuators [43] enable a continuous and linear displacement with electrical power. For this reason, linear actuators including electro-thermal may enable a simpler realization of the MEMS-tunable polarization rotator proposed in Fig. 7. Alternatively, the silicon waveguide design results in a tightly-confined mode that limits the evanescent field interaction with the slanted MEMS perturber. Although not investigated here, the use of thinner silicon or lower index waveguide core materials (e.g., silicon nitride) may enable a lower mode confinement resulting in a larger evanescent field and – presumably – a larger required air gap separating the waveguide and MEMS perturber. Narrower waveguides can enable larger evanescent fields and may also increase the required gap that is needed for polarization rotation. Future designs can benefit from optimization of the materials and geometries used in order to optimize the required MEMS displacement to ease fabrication requirements.
7. Conclusions
Polarization management is key for integrating PICs, which are generally highly-polarization-dependent, with standard optical fiber networks. This work has proposed two approaches for MEMS-tunable polarization management in PICs that can enable an increased bandwidth of operation and optimal performance for both TE- and TM-polarizations thereby enabling a wider range of devices to be included in a foundry process development kit (PDK). The first structure enables MEMS-tunable polarization beam splitting while a second device enables MEMS-tunable polarization rotation. Both polarization management structures have been investigated via calculation and simulation (3D-BPM and 3D-FDTD) in terms of reconfigurability, polarization extinction ratio, and bandwidth. The required MEMS displacements (<100 nm) are well-suited to electrostatic, gradient electric force, and electro-thermal actuation, and some actuation schemes have been proposed. Finally, we have also discussed some fabrication considerations for potential implementation of these polarization splitters and rotators in a standard photonic foundry. Although MEMS are still an emerging photonic technology, their integration in optical foundries will enable a much-needed capability for polarization management in PICs. We expect that the MEMS-tunable polarization management components in this work will lead to an increased deployment of PICs in standard fiber optic networks.
Funding
U.S. Naval Research Laboratory (WU# 1L52).
Disclosures
The authors declare that there are no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request. Data release is subject to approval by the U.S. Naval Research Laboratory.
References
1. Y. C. Shi, Y. Zhang, Y. T. Wan, Y. Yu, Y. G. Zhang, X. Hu, X. Xiao, H. N. Xu, L. Zhang, and B. C. Pan, “Silicon photonics for high-capacity data communications,” Photonics Res. 10(9), A106–A134 (2022). [CrossRef]
2. T. K. Paraiso, T. Roger, D. G. Marangon, I. De Marco, M. Sanzaro, R. I. Woodward, J. F. Dynes, Z. L. Yuan, and A. J. Shields, “A photonic integrated quantum secure communication system,” Nat. Photonics 15(11), 850–856 (2021). [CrossRef]
3. S. A. Holmstrom, T. H. Stievater, D. A. Kozak, M. W. Pruessner, N. Tyndall, W. S. Rabinovich, R. A. McGill, and J. B. Khurgin, “Trace gas Raman spectroscopy using functionalized waveguides,” Optica 3(8), 891–896 (2016). [CrossRef]
4. Y. J. Huang, J. G. F. Flores, Y. Li, W. T. Wang, D. Wang, N. Goldberg, J. J. Zheng, M. B. Yu, M. Lu, M. Kutzer, D. Rogers, D. L. Kwong, L. Churchill, and C. E. W. Wong, “A Chip-Scale Oscillation-Mode Optomechanical Inertial Sensor Near the Thermodynamical Limits,” Laser Photonics Rev. 14, 1800329 (2020). [CrossRef]
5. N. M. Fahrenkopf, C. McDonough, G. L. Leake, Z. Su, E. Timurdogan, and D. D. Coolbaugh, “The AIM Photonics MPW: A Highly Accessible Cutting Edge Technology for Rapid Prototyping of Photonic Integrated Circuits,” IEEE J. Select. Topics Quantum Electron. 25(5), 1–3 (2019). [CrossRef]
6. I. Kiyat, A. Aydinli, and N. Dagli, “A compact silicon-on-insulator polarization splitter,” IEEE Photonics Technol. Lett. 17(1), 100–102 (2005). [CrossRef]
7. A. N. Miliou, R. Srivastava, and R. V. Ramaswamy, “A 1.3-Mu-M Directional Coupler Polarization Splitter by Ion-Exchange,” J. Lightwave Technol. 11(2), 220–225 (1993). [CrossRef]
8. D. X. Dai and J. E. Bowers, “Novel ultra-short and ultra-broadband polarization beam splitter based on a bent directional coupler,” Opt. Express 19(19), 18614–18620 (2011). [CrossRef]
9. M. R. Watts and H. A. Haus, “Integrated mode-evolution-based polarization rotators,” Opt. Lett. 30(2), 138–140 (2005). [CrossRef]
10. H. El-Refaei, D. Yevick, and T. Jones, “Slanted-rib waveguide InGaAsP-InP polarization converters,” J. Lightwave Technol. 22(5), 1352–1357 (2004). [CrossRef]
11. D. X. Dai and J. E. Bowers, “Novel concept for ultracompact polarization splitter-rotator based on silicon nanowires,” Opt. Express 19(11), 10940–10949 (2011). [CrossRef]
12. L. Chen, C. R. Doerr, and Y. K. Chen, “Compact polarization rotator on silicon for polarization-diversified circuits,” Opt. Lett. 36(4), 469–471 (2011). [CrossRef]
13. M. Aamer, A. M. Gutierrez, A. Brimont, D. Vermeulen, G. Roelkens, J. M. Fedeli, A. Hakansson, and P. Sanchis, “CMOS Compatible Silicon-on-Insulator Polarization Rotator Based on Symmetry Breaking of the Waveguide Cross Section,” IEEE Photonics Technol. Lett. 24(22), 2031–2034 (2012). [CrossRef]
14. D. Vermeulen, S. Selvaraja, P. Verheyen, P. Absil, W. Bogaerts, D. Van Thourhout, and G. Roelkens, “Silicon-on-Insulator Polarization Rotator Based on a Symmetry Breaking Silicon Overlay,” IEEE Photonics Technol. Lett. 24(6), 482–484 (2012). [CrossRef]
15. T. Barwicz, M. R. Watts, M. A. Popovic, P. T. Rakich, L. Socci, F. X. Kartner, E. P. Ippen, and H. I. Smith, “Polarization-transparent microphotonic devices in the strong confinement limit,” Nat. Photonics 1(1), 57–60 (2007). [CrossRef]
16. W. D. Sacher, T. Barwicz, B. J. F. Taylor, and J. K. S. Poon, “Polarization rotator-splitters in standard active silicon photonics platforms,” Opt. Express 22(4), 3777–3786 (2014). [CrossRef]
17. J. E. Bigelow and R. A. Kashnow, “Poincare Sphere Analysis of Liquid-Crystal Optics,” Appl. Opt. 16(8), 2090–2096 (1977). [CrossRef]
18. J. W. Judy, “Microelectromechanical systems (MEMS): fabrication, design and applications,” Smart Mater. Struct. 10(6), 1115–1134 (2001). [CrossRef]
19. E. Ollier, “Optical MEMS devices based on moving waveguides,” IEEE J. Select. Topics Quantum Electron. 8(1), 155–162 (2002). [CrossRef]
20. T. Bakke, C. P. Tigges, J. J. Lean, C. T. Sullivan, and O. B. Spahn, “Planar microoptomechanical waveguide switches,” IEEE J. Select. Topics Quantum Electron. 8(1), 64–72 (2002). [CrossRef]
21. M. W. Pruessner, N. Siwak, K. Amarnath, S. Kanakaraju, W. H. Chuang, and R. Ghodssi, “End-coupled optical waveguide MEMS devices in the indium phosphide material system,” J. Micromech. Microeng. 16(4), 832–842 (2006). [CrossRef]
22. E. C. Vail, M. S. Wu, G. S. Li, L. Eng, and C. J. Changhasnain, “Gaas Micromachined Widely Tunable Fabry-Perot Filters,” Electron. Lett. 31(3), 228–229 (1995). [CrossRef]
23. C. K. Madsen, J. A. Walker, J. E. Ford, K. W. Goossen, T. N. Nielsen, and G. Lenz, “A tunable dispersion compensating MEMS all-pass filter,” IEEE Photonics Technol. Lett. 12(6), 651–653 (2000). [CrossRef]
24. S. Irmer, J. Daleiden, V. Rangelov, C. Prott, F. Romer, M. Strassner, A. Tarraf, and H. Hillmer, “Ultralow biased widely continuously tunable Fabry-Perot filter,” IEEE Photonics Technol. Lett. 15(3), 434–436 (2003). [CrossRef]
25. V. Jayaraman, G. D. Cole, M. Robertson, A. Uddin, and A. Cable, “High-sweep-rate 1310 nm MEMS-VCSEL with 150 nm continuous tuning range,” Electron. Lett. 48(14), 867–868 (2012). [CrossRef]
26. M. Poot and H. X. Tang, “Broadband nanoelectromechanical phase shifting of light on a chip,” Appl. Phys. Lett. 104(6), 061101 (2014). [CrossRef]
27. W. Jin, R. G. Polcawich, P. A. Morton, and J. E. Bowers, “Piezoelectrically tuned silicon nitride ring resonator,” Opt. Express 26(3), 3174–3187 (2018). [CrossRef]
28. P. R. Stanfield, A. J. Leenheer, C. P. Michael, R. Sims, and M. Eichenfield, “CMOS-compatible, piezo-optomechanically tunable photonics for visible wavelengths and cryogenic temperatures,” Opt. Express 27(20), 28588–28605 (2019). [CrossRef]
29. N. Quack, H. Sattari, A. Y. Takabayashi, Y. Zhang, P. Verheyen, W. Bogaerts, P. Edinger, C. Errando-Herranz, and K. B. Gylfason, “MEMS-Enabled Silicon Photonic Integrated Devices and Circuits,” IEEE J. Quantum Electron. 56(1), 1–10 (2020). [CrossRef]
30. P. Edinger, A. Y. Takabayashi, C. Errando-Herranz, U. Khan, H. Sattari, P. Verheyen, W. Bogaerts, N. Quack, and K. B. Gylfason, “Silicon photonic microelectromechanical phase shifters for scalable programmable photonics,” Opt. Lett. 46(22), 5671–5674 (2021). [CrossRef]
31. A. G. Krause, M. Winger, T. D. Blasius, Q. Lin, and O. Painter, “A high-resolution microchip optomechanical accelerometer,” Nat. Photonics 6(11), 768–772 (2012). [CrossRef]
32. P. F. Cohadon, A. Heidmann, and M. Pinard, “Cooling of a mirror by radiation pressure,” Phys. Rev. Lett. 83(16), 3174–3177 (1999). [CrossRef]
33. M. L. Povinelli, M. Loncar, M. Ibanescu, E. J. Smythe, S. G. Johnson, F. Capasso, and J. D. Joannopoulos, “Evanescent-wave bonding between optical waveguides,” Opt. Lett. 30(22), 3042–3044 (2005). [CrossRef]
34. M. Zalalutdinov, A. Zehnder, A. Olkhovets, S. Turner, L. Sekaric, B. Ilic, D. Czaplewski, J. M. Parpia, and H. G. Craighead, “Autoparametric optical drive for micromechanical oscillators,” Appl. Phys. Lett. 79(5), 695–697 (2001). [CrossRef]
35. P. B. Deotare, I. Bulu, I. W. Frank, Q. M. Quan, Y. N. Zhang, R. Ilic, and M. Loncar, “All optical reconfiguration of optomechanical filters,” Nat. Commun. 3(1), 846 (2012). [CrossRef]
36. G. S. Wiederhecker, L. Chen, A. Gondarenko, and M. Lipson, “Controlling photonic structures using optical forces,” Nature 462(7273), 633–636 (2009). [CrossRef]
37. M. Hossein-Zadeh, H. Rokhsari, A. Hajimiri, and K. J. Vahala, “Characterization of a radiation-pressure-driven micromechanical oscillator,” Phys. Rev. A: At., Mol., Opt. Phys. 74(2), 023813 (2006). [CrossRef]
38. O. Arcizet, P. F. Cohadon, T. Briant, M. Pinard, and A. Heidmann, “Radiation-pressure cooling and optomechanical instability of a micromirror,” Nature 444(7115), 71–74 (2006). [CrossRef]
39. C. H. Metzger and K. Karrai, “Cavity cooling of a microlever,” Nature 432(7020), 1002–1005 (2004). [CrossRef]
40. T. J. Kippenberg, H. Rokhsari, T. Carmon, A. Scherer, and K. J. Vahala, “Analysis of radiation-pressure induced mechanical oscillation of an optical microcavity,” Phys. Rev. Lett. 95(3), 033901 (2005). [CrossRef]
41. M. W. Pruessner, T. H. Stievater, and W. S. Rabinovich, “In-plane microelectromechanical resonator with integrated Fabry-Perot cavity,” Appl. Phys. Lett. 92(8), 081101 (2008). [CrossRef]
42. M. W. Pruessner, D. Park, B. J. Roxworthy, D. A. Kozak, T. H. Stievater, N. F. Tyndall, and W. S. Rabinovich, “Loss reduction in electromechanically tunable microring cavities,” Opt. Lett. 44(13), 3346–3349 (2019). [CrossRef]
43. M. W. Pruessner, B. J. Roxworthy, D. A. Kozak, N. F. Tyndall, W. S. Rabinovich, and T. H. Stievater, “Micro-electro-mechanically tunable optical phase matching and mode conversion,” Opt. Lett. 48(2), 427–430 (2023). [CrossRef]
44. J. B. Khurgin, M. W. Pruessner, T. H. Stievater, and W. S. Rabinovich, “Laser-Rate-Equation Description of Optomechanical Oscillators,” Phys. Rev. Lett. 108(22), 223904 (2012). [CrossRef]
45. M. W. Pruessner, D. Park, T. H. Stievater, D. A. Kozak, and W. S. Rabinovich, “Optomechanical Cavities for All-Optical Photothermal Sensing,” ACS Photonics 5(8), 3214–3221 (2018). [CrossRef]
46. M. W. Pruessner, D. A. Kozak, N. A. Tyndall, W. S. Rabinovich, V. Deenadayalan, M. Fanto, S. Preble, and T. H. Stievater, “Foundry-processed optomechanical photonic integrated circuits,” OSA Continuum 4(4), 1215–1222 (2021). [CrossRef]
47. R. C. Alferness and L. L. Buhl, “Electro-Optic Wave-Guide Te]-]Tm Mode Converter with Low Drive Voltage,” Opt. Lett. 5(11), 473–475 (1980). [CrossRef]
48. J. D. Sarmiento-Merenguel, R. Halir, X. Le Roux, C. Alonso-Ramos, L. Vivien, P. Cheben, E. Duran-Valdeiglesias, I. Molina-Fernandez, D. Marris-Morini, D. X. Xu, J. H. Schmid, S. Janz, and A. Ortega-Monux, “Demonstration of integrated polarization control with a 40 dB range in extinction ratio,” Optica 2(12), 1019–1023 (2015). [CrossRef]
49. Q. Xu, L. Chen, M. G. Wood, P. Sun, and R. M. Reano, “Electrically tunable optical polarization rotation on a silicon chip using Berry's phase,” Nat. Commun. 5(1), 5337 (2014). [CrossRef]
50. T. H. Stievater, M. W. Pruessner, W. S. Rabinovich, D. Park, R. Mahon, D. A. Kozak, J. B. Boos, S. A. Holmstrom, and J. B. Khurgin, “Suspended photonic waveguide devices,” Appl. Opt. 54(31), F164–F173 (2015). [CrossRef]
51. Comsol, V6.0, COMSOL, Inc., Burlington, MA, 2021.
52. BeamPROP, V2021.09-1, Synopsis, Inc., Sunnyvale, CA, 2021.
53. L. X. Jia, T. Y. Liow, J. F. Song, X. S. Luo, N. Duan, S. C. Koh, Q. Fang, M. B. Yu, and G. Q. Lo, “Compact Optical Polarization Rotators Based on an Asymmetric Silicon Waveguide,” IEEE Photonics Technol. Lett. 25(22), 2229–2232 (2013). [CrossRef]
54. FullWave, V2021.09-1, Synopsis, Inc., Sunnyvale, CA, 2021.
55. M. W. Pruessner, D. Park, T. H. Stievater, D. A. Kozak, and W. S. Rabinovich, “Broadband opto-electro-mechanical effective refractive index tuning on a chip,” Opt. Express 24(13), 13917–13930 (2016). [CrossRef]
56. P. M. Osterberg and S. D. Senturia, “M-TEST: A test chip for MEMS material property measurement using electrostatically actuated test structures,” J. Microelectromech. Syst. 6(2), 107–118 (1997). [CrossRef]
57. “AIM Photonics” (2023), retrieved 29 June 2023, https://www.aimphotonics.com/.
