Open Access
Add to BibSonomy
Abstract
We propose and demonstrate an angularly offset multiline (AOML) dispersive silicon nitride optical phased array (OPA) that enables efficient line beam scanning with an expanded field of view (FOV) and plateau envelope. The suggested AOML OPA incorporates multiline OPA units, which were seamlessly integrated with a 45° angular offset through a thermo-optic switch based on a multimode interference coupler, resulting in a wide FOV that combines three consecutive scanning ranges. Simultaneously, a periodic diffraction envelope rendered by the multiline OPA units contributes to reduced peak intensity fluctuation of the main lobe across the large FOV. An expedient polishing enabling the angled facet was diligently accomplished through the implementation of oblique polishing techniques applied to the 90° angle of the chip. For each dispersive OPA unit, we engineered an array of delay lines with progressively adjustable delay lengths, enabling a passive wavelength-tunable beam scanning. Experimental validation of the proposed OPA revealed efficient beam scanning, achieved by wavelength tuning from 1530 to 1600 nm and seamless switching between multiline OPAs, yielding an FOV of 152° with a main lobe intensity fluctuation of 2.8 dB. The measured efficiency of dispersive scanning was estimated at 0.97°/nm, as intended.
© 2024 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Over the past decade, optical phased arrays (OPAs) have gained significant attention due to their swift and inertia-free beam scanning, versatile nonmechanical beam shaping, and compatibility with large-scale complementary metal-oxide semiconductor (CMOS) manufacturing processes. These attributes render them highly valuable in various state-of-the-art technical applications, including holographic displays, autonomous vehicles, and light detection and ranging (LiDAR) [1–6]. Silicon photonics, compatible with CMOS technology and offering a high index contrast to silica, has emerged as a leading platform for OPA design. However, silicon-based devices face challenges, such as nonlinear absorption effects, including free-carrier absorption and two-photon absorption. Silicon nitride (SiN) presents a compelling alternative, offering enhanced high-power processing capabilities, lower nonlinearity, and broader transparency. Nevertheless, the relatively weak light confinement in dielectric SiN waveguides imposes limitations on achieving excessively dense arrangements, complicating the dense layout and restricting aliasing-free beam scanning [4,7–11]. Various OPA devices have been conceived to address limitations in the beam scanning range. Aperiodic antenna arrays, which suppress side lobes during beam scanning, have become a significant alternative [12,13]. However, the power of the main lobe is dispersed throughout the field of view (FOV) during beam scanning. Alternative approaches include half-wavelength spacing, dispersion-engineered waveguides with varied widths, and photonic crystal structures for antenna segregation, all of which require high processing precision [13–18]. During beam scanning in these configurations, the main lobe intensity dramatically declines along the far-field diffraction envelope produced by a single antenna element, resulting in a significant limitation on beam detection quality and long-distance transmission for both uniform and aperiodic OPAs. Previously, a wider envelope could be achieved by reducing the dimensions of individual emitters and generating resonance through an oxide offset between the waveguide end facet and the chip edge, albeit requiring refined and complicated fabrication processes [19].
Typically, OPA structures capitalizing on an antenna array of surface gratings have been recognized as viable options for 2-dimensional (2D) emitters, enabling 2D point-beam raster scanning by adjusting the wavelength and phase gradient in both vertical and horizontal directions [7,10–15]. However, raster beam scanning encounters significant challenges, including high power loss due to gratings and extended scanning times. Therefore, OPAs supporting line beam scanning have emerged as an attractive platform for proof-of-concept, providing higher emission efficiency and simpler operation [13,20]. In general, electro-optic or thermo-optic phase modulators are deployed across the emitter channels to control the lateral movement of the beam. These OPAs necessitate the integration of active components such as phase modulators with heaters, and electronic drive circuits for precise phase control, which escalates the complexity pertaining to both manufacturing and operational processes. To mitigate this issue, dispersive OPAs were attempted to tailor the wavelength to achieve efficient horizontal beam scanning [21,22]. The passive nature of the dispersive devices could unequivocally yield advantages in terms of the reduced operational complexity and lower power consumption.
In this study, we propose and realize an integrated SiN-based angularly offset multiline (AOML) dispersive OPA, providing efficient beam scanning with a large FOV and plateau envelope. The optical switch based on a multimode interference (MMI) coupler is meticulously designed and integrated with multiline OPA units that are positioned with an angular offset, facilitating beam path selection via a phase shifter (PS). An enlarged scanning region can be secured by performing beam switching between the OPA units, each covering different scanning ranges. Furthermore, the designed AOML OPA effectively minimizes intensity fluctuations of main lobe throughout the beam-scanning process, thanks to the presence of a periodic diffraction envelope. Each separate OPA unit, consisting of a delay line (DL) array incorporating a series of waveguides of varying lengths, enables dispersive wavelength-tunable beam scanning due to phase gradient along the horizontal direction. An end-fire uniform array, devoid of surface gratings, has been designed for deployment as a line beam emitter (LBE), resulting in high emission efficiency and rapid beam scanning. The AOML OPA has been rigorously designed and fabricated based on the SiN planar lightwave circuit platform with the help of expedient oblique polishing techniques. To the best of our knowledge, the presented device constitutes the pioneering demonstration of a multiline OPA based on an angular offset, facilitating the passive line beam scanning that exhibits no significant fluctuations in main lobe intensity and a wide FOV that seamlessly combines three consecutive scanning ranges.
2. Design of the proposed AOML OPA
Figure 1 illustrates the configuration of the proposed integrated SiN-based AOML OPA, comprising a spot size converter (SSC), optical switch, MMI power splitter, Ω-shaped DL array, and LBE. Incident light from a tunable laser is injected into the OPA chip by the SSC coupler and then directed to three OPA units (OPA 1, OPA 2, and OPA 3) via a 1 × 3 optical switch featuring the MMI coupler. By employing the PS, which is induced by a heater positioned above the waveguides, the propagation path of the light can be selected among the three OPA units, each arranged with an offset angle α from the others. Light propagating along a single dispersive OPA is evenly split into 32 channels through a five-stage MMI power splitter. The output light at the end of the power splitter is efficiently delivered to the waveguides of the DL array, where the phase of the propagating light is adjusted due to the length difference (ΔL) across the channels. Subsequently, the guided mode, exhibiting a specific phase difference across channels, is coupled to the LBE in the form of an end-fire uniformly spaced SiN-based waveguide array, which is implemented to facilitate beam scanning along the horizontal direction as the input wavelengths vary from λa to λb. Ultimately, beam scanning is achieved by switching between the three OPA units, achieving a wide FOV by combining the three consecutive scanning ranges. Here, ΔΨ1, ΔΨ2, and ΔΨ3 represent the individual scanning ranges of OPA 1, 2, and 3, respectively; ΔΨ denotes the total scanning range of the integrated AOML OPA; and Δθ indicates the vertical beamwidth.
Fig. 1. Schematic of the proposed integrated AOML OPA enabling a large FOV and plateau envelope along the horizontal direction.
The suggested AOML OPA is anticipated to integrate multiline OPA constituents through an optical switch based on MMI couplers. The configuration of a 1 × 2 optical switch, comprising a 1 × 2 MMI coupler (MMI 1), 2 × 2 MMI coupler (MMI 2), and thermo-optic PS, is depicted in Fig. 2(a). Initially, the incident light is divided into two channels via MMI 1. By manipulating the phase difference (ΔφPS) through a PS located on the upper arm, an asymmetric field distribution is generated within the MMI region of MMI 2, which allows for flexible adjustment of the power distribution at the output ports. When the ΔφPS is set to ±π/2, it is observed that beam path can be efficiently selected at the outputs [23,24]. Figure 2(b) displays the electric-field profiles of MMI 1, where light is efficiently split into two channels. The electric-field profiles of MMI 2 are shown in Figs. 2(c), (d), and (e) for ΔφPS values of 0, -π/2 (3π/2), and π/2, respectively. The beat length (Lπ) of the MMI coupler is approximately calculated by the following equation [25]:
Fig. 2. (a) Configuration of the 1 × 2 optical switch, consisting of MMI 1, MMI 2, and PS. (b) Electric-field profiles of MMI 1. Electric-field profiles of MMI 2 when ΔφPS is (c) 0, (d) -π/2, and (e) π/2, respectively.
The designed optical switch was analyzed and optimized using a simulation tool based on the beam propagation method (BeamPROP; Synopsys Inc.). The widths of the MMI regions (WMMI 1, 2) were fixed at 8 µm, and the corresponding lengths (LMMI 1 and LMMI 2) were chosen to be 107 µm and 143 µm, respectively. The taper width of the input port (Wi), the taper width of the output port (Wo), the taper length (Ltaper), and the spacing of the two output ports (Λtaper) were properly determined. The designed parameters are presented in Table 1. Figure 3(a) presents a schematic of the proposed 1 × 3 optical switch, comprising two cascaded 1 × 2 optical switches (SW1 & SW2), facilitating the path selection among the multiline angularly offset OPA units. Figures 3(b), (c), and (d) illustrate the electric-field distribution when the beam is transmitted from the three output ports (O1, O2, and O3). The phase differences between the lightwaves propagating in the upper and lower arms, ΔφPS1 and ΔφPS2, are generated by capitalizing on PS1 and PS2, respectively. For ΔφPS1 = -π/2, the incident light is emitted from O1 through SW1 while ΔφPS2 obviously has no effect on beam selection, as depicted in Fig. 3(b). For ΔφPS1 = π/2, the light wave is directed toward SW2 and then emanating from O2 and O3 when ΔφPS2 is -π/2 and π/2, as shown in Fig. 3(c) and (d), respectively. Figures 3(e), (f), and (g) plot the simulated insertion loss over the wavelength range spanning from 1530 to 1600 nm when light is emerging from O1, O2, and O3, respectively. The switching device consistently exhibited no significant optical loss within the 70-nm wavelength range, with simulated insertion losses running from 0.3 to 0.7 dB, 0.4 to 0.9 dB, and 0.8 to 1.1 dB for the three cases. It was also observed the extinction ratio (ER) was calculated to surpass 30 dB throughout the spectral band, underpinning a highly efficient broadband performance.
Fig. 3. (a) Configuration of the proposed 1 × 3 optical switch. The simulated electric-field profiles at a wavelength of 1550 nm when the beam is transmitted from (b) O1, (c) O2, and (d) O3. The corresponding simulated transmittances for the three cases are portrayed in (e), (f), and (g) when the wavelengths vary from 1530 to 1600 nm.
Table 1. Structural parameters of the designed optical switch
The OPA, comprising an array of coherent light sources, can accomplish beam scanning by introducing variable phase gradients across the channels. Conventional phase modulators, which are instrumental in introducing phase gradients, necessitate complex processing and high-power consumption due to their electro-/thermo-optic operation. In this work, we have attempted to realize a dispersive OPA by inducing a phase gradient stemming from a passive DL array, imposing a flexible phase accumulation depending on the wavelength. The phase difference (Δφ) between adjacent waveguide channels is governed by their length difference (ΔL), expressed as $\mathrm{\Delta \varphi =\\ 2\pi {n}_{\textrm{eff}}\;\Delta L/\;\lambda}$, where neff and λ represent the effective refractive index of the guided mode of the waveguides and the wavelength, respectively. The schematic of the dispersive OPA is illustrated in Fig. 4(a). The waveguides comprising the DL efficiently accept the output light coming from the MMI power splitter, thus leading to an ΔL-dependent phase gradient across the channels. The LBE tethered to the DL array, consisting of uniformly spaced SiN waveguides, facilitates efficient beam scanning by tuning the wavelength. For the wavelength-tunable OPA unit incorporating the DL array, the beam-scanning rate ($\mathrm{d\Psi /d\lambda }$) can be approximately given by the following equation [22]:
Fig. 4. (a) Schematic of the dispersive OPA unit enabling the passive wavelength-tunable beam scanning. (b) Configuration of the DL array.
For each dispersive OPA unit, the included LBE can be embodied by incorporating the uniformly spaced end-fire SiN waveguides. The maximum scanning angle in response to phase differences of ±π can be theoretically estimated by ${\mathrm{\Psi }_{\textrm{max}}} = \textrm{si}{\textrm{n}^{\textrm{ - 1}}}\mathrm{(\lambda /2}{\mathrm{\Lambda }_{\textrm{ch}}}\textrm{)}$. The maximum offset angle (αmax) is determined by the $\mathrm{\Psi }$max of each OPA unit, expressed as αmax < 2$\mathrm{\Psi }$max, representing the maximum feasible angle which allows the scanning ranges of the three OPA units to overlap, so that the entire scanning could be fulfilled consistently with no interruption. This relationship allows for overlapping scanning ranges of the separate OPA units, as shown in Fig. 5(a). The αmax is 45.6° and 62.2° for Λch = 2 and 1.5 µm. To investigate the scanning performance, the AOML OPA was analyzed with the aid of a simulation tool based on the finite-difference time-domain (FDTD) method (Lumerical Inc.). The far-field patterns for Λch = 2 µm and 1.5 µm were mainly simulated. Figures 5(b) and (c) display the polar far-field patterns for applied phase differences of Δφ = 0 and π for Λch = 2 µm. The grating lobes were located at ±50.8° under the scanning angle of 0°. The positions of the grating lobes were determined by $\textrm{sin}{\mathrm{\Psi }_{\textrm{grating}}}\mathrm{\;\ =\ \;\ \pm n\lambda /}{\mathrm{\Lambda }_{\textrm{ch}}}$, where n signifies the order of the grating lobes. Notably, there were no grating lobes for Λch = 1.5 µm under Δφ = 0, as shown in Fig. 5(d). The corresponding maximum scanning ranges were ±22.8° and ±31.1° for the two cases, thus limiting beam scanning.
Fig. 5. (a) Calculated maximum scanning angle (Ψmax) of the OPA and corresponding maximum offset angle (αmax). Far-field patterns emerging from the OPA for Δφ of (b) 0 and (c) π when Λch is 2 µm. Far-field patterns emerging from the OPA for Δφ of (d) 0 and (e) π when Λch is 1.5 µm.
The emission response and dispersion characteristics of the OPA units with Λch = 2 µm and 1.5 µm were further explored. The waveguide dispersion for the core SiN waveguide with a thickness of 400 nm and width of 1 µm was calculated to be 2.2 × 10−4/nm, when the refractive indices of silicon oxide (SiO2) and SiN films were set to 1.45 and 1.97 in simulations. Figures 6(a) and (b) illustrate the horizontal cross sections of the emitted beam for Λch of 2 µm and 1.5 µm, respectively. In this scenario, ΔL was fixed at 20 µm, and λ varied from 1530 to 1600 nm in 10 nm steps, covering the C and L bands. It is mentioned that the light confinement in SiN waveguides is relatively weak due to their lower index contrast compared to its silicon counterpart. Therefore, it is imperative to conduct crosstalk analysis for the densely arranged waveguides. The crosstalk effect for the entire line beam emitter was rigorously investigated in terms of the far-field beam scanning characteristics depending on the phase difference for longer waveguide lengths of 50 µm and 100 µm under Λch = 2 µm and 1.5 µm. The emitted beams from the proposed OPA aligns with the anticipated scanning patterns. The simulated scanning angles exhibit a favorable concordance with the theoretical results, indicating that for the line beam emitter with Λch = 2 µm and 1.5 µm, the crosstalk exerts no detrimental effect on the beam scanning characteristics. Nevertheless, the crosstalk can be mitigated by precisely tailoring the polishing process to shorten the waveguide antenna or introducing nonuniform widths of waveguide channels [26]. The beam scans in the negative $\mathrm{\Psi\ }$ direction as the wavelength increases, and the simulated scanning rates are 0.73°/nm and 0.97°/nm, which are consistent with theoretical results. Notably, the scanning rate increases with decreasing Λch under a fixed ΔL. In addition, the scanning range (ΔΨ) and full width at half maximum (FWHM) along the horizontal direction (ΔΨFWHM) were investigated in detail, serving as crucial factors in beam-scanning accuracy. The scanning angle hinges on Λch and λ, in accordance with $\mathrm{sin\Psi \;\ =\ \;\ \lambda \Delta \varphi /\;\ (2\pi }{\mathrm{\Lambda }_{\textrm{ch}}})$. The available FOVs were 45.6° and 62.2° for Λch = 2 µm and 1.5 µm, respectively, closely matching theoretical predictions. The ΔΨFWHM slightly increased from 1.2° to 1.5° as the Λch decreased, which is attributed to the reduced aperture size of the emitters. The required wavelength range (Δλ) can achieve phase differences of ±π for both cases. And the flexible scanning ranges and resolutions can be achieved by modifying the configuration of the dispersive OPA unit.
Fig. 6. Simulated far-field horizontal beam patterns emerging from dispersive OPA unit for Λch of (a) 2 µm and (b) 1.5 µm.
To further inspect the emission response of the AOML OPA, the channel spacing (Λch) and the offset angle of the OPA elements were set at 1.5 µm and 45°, respectively, by scrutinizing the entire scanning region and beamwidth. The simulated emission angles of the AOML OPA are depicted in Fig. 7. The three OPA units (OPA 1, OPA 2, and OPA 3) were put together through an optical switch as a means to select the propagation path, giving rise to a total scanning region of approximately 152°. The solid line in the figure alludes to the fitted values, indicating the scanning efficiency. When the scanned beam exceeds the limitation of the negative FOV, the beam located on the positive side intensifies and becomes the primary beam, engendering a sudden shift in the emission angle. The horizontal cross sections of the calculated beam profiles are shown in Fig. 8(a). Beam scanning was fulfilled by suitably tuning the wavelength and switching between the three OPA units. A beam-scanning range of approximately 152° was ultimately attained, as intended. The envelope of the main lobe appeared periodically due to the integration of OPA units, leading to the suppressed main lobe peak variation in intensity with a large scanning angle. Figure 8(b) shows the peak intensity of the main lobe as a function of the scanning angle. The fluctuation of main lobe peak intensity (ΔI) was discovered to be about 2.5 dB when the scanning angle varied from -76° to +76°, representing a plateau envelope. The black dotted line represents the Gaussian-shaped envelope as determined by the single emitter channel, exhibiting the ΔI variation of ∼14.6 dB over the identical scanning range. It is hence demonstrated the proposed device can provide an enhanced performance in terms of scanning range and uniformity of the main lobe intensity. Due to the variations in insertion losses among the three OPA units (OPA 1, 2, and 3), the peak intensity of the emitted main lobe exhibits discrepancies. To ensure a consistent basis for comparison, the maximum peak intensity of main lobe among the three OPA units is established as a reference. Throughout the experiment, the highest emission efficiency was monitored to emanate from OPA 2. The main lobe peak intensity at a scanning angle of 0° was designated as the reference point, indicating the maximum value in the entire AOML OPA. Subsequently, the beam intensity of the other OPA units was normalized relative to the peak level observed in OPA 2, facilitating a coherent representation of the corresponding intensity across the entire structure. Our integrated AOML OPA ultimately achieved an entire scanning region of approximately 152° with ΔI of ∼2.5 dB by adjusting the wavelength and the mediation of the optical switch. As previously elucidated, our endeavor has been directed towards the realization of a dispersive OPA, which is accomplished through the induction of a phase gradient originating from a passive DL array, thereby enforcing a phase accumulation contingent upon the wavelength. The primary source of error is principally believed to arise from the fabrication tolerance of the SiN waveguides, leading to phase errors attributed to slightly varying effective refractive indices of the waveguides. To investigate the influence of phase errors resulting from fabrication errors, we explored the operational characteristics of the far-field beam under phase errors of π/3 and 2π/3. The corresponding variation in scanning angle was less than 2°, supporting that the proposed device can provide a stable scanning performance despite relatively large phase errors.
Fig. 8. (a) Simulated horizontal cross-sectional beam profiles of the integrated AOML OPA in terms of the scanning angle. (b) Corresponding main lobe peak intensity.
3. Fabrication of the integrated AOML OPA and its demonstrated performance
The AOML OPA chip was constructed using a planar lightwave circuit platform provided by Applied Nanotools Inc. (ANT, Canada), where direct-write electron-beam lithography was utilized to define the SiN waveguide patterns exhibiting width variations of ±20 nm, which may be perceived as a primary source of phase errors across the channels. The core SiN waveguide, with a thickness of 400 nm, was encapsulated by a 4.5-µm-thick SiO2 buried oxide (BOX) layer and a 3-µm-thick SiO2 upper cladding layer. A titanium-tungsten (TiW) alloy heater was then positioned on top of the waveguide, facilitating the operation of the optical switch. Deep trench processing was subsequently applied to the manufactured sample for the generation of a smooth cross-section, resulting in flawless edge coupling.
First, a single dispersive SiN OPA unit resorting to a passive DL array was characterized. The experimental setup utilized to evaluate the manufactured OPA is depicted in Fig. 9(a). The incident light was generated by a tunable laser (WSL-110, Santec), passed through a polarization controller (PC), and coupled to the OPA chip via an ultra-high numerical aperture fiber (Model UHNA4). The input light was polarized in the transverse electric (TE) mode using the PC, and a short-wave infrared (SWIR) camera (ABA-001IR-GE, AVAL GLOBAL) mounted on a three-axis motorized stage was used to entirely capture the emitted beam profiles. Figure 9(b) shows a microscopic image of the prepared OPA chip, where the SSC, five-stage power splitter, DL array, and LBE were monolithically integrated. The zoom in image of the LBE, shown in Fig. 9(b), comprises a SiN waveguide array with uniform spacing of 1.5 µm and a width of 1 µm. Figure 9(c) exhibits the captured beam profile at a scanning angle of 0° (Δφ = 0) with a corresponding wavelength of 1565 nm. The dispersive OPA engendered a total emission loss of ∼3.5 dB, attributed to losses from fiber-to-chip coupling, the MMI power splitter tree, propagation, and the LBE. The beam was scanned in the negative direction as the wavelength was gradually increased. The captured beam with a 15.6° (Δφ = π/2) angle is depicted in Fig. 9(d), with a corresponding wavelength of 1554 nm. As presented in Fig. 9(e), the maximum scanning angle reached ±31° (Δφ = π) at a wavelength of 1537 nm, aligning well with the simulation results. A single SiN OPA tapping into a passive DL was practically characterized, focusing on its beam scanning rendered by appropriately tuning the wavelengths. Figure 10(a) reveals the captured scanning beam along the Ψ direction by varying the wavelengths from 1530 to 1600 nm in increments of 10 nm. The observable maximum scanning range was roughly 62.2°, with a ΔΨFWHM of 1.5°. The scanning angle was −22.4° at the designated shortest wavelength of 1530 nm. The main lobe steered in the negative direction with increasing wavelength, and the captured beam appeared as two identical beams at a wavelength of 1537 nm, corresponding to a Δφ of ±π. Subsequently, the main lobe turned in the negative direction from +31.1° to replace the weakening main beam. As the wavelength continued to increase, the main lobe scanned continuously in the negative Ψ direction. The measured results support that the dispersive OPA can execute beam scanning across the entire available FOV within a 70-nm range. The corresponding cross sections of the beam profile are shown in Fig. 10(b). The beam profiles were normalized with reference to the output intensity at a scanning angle of 0°. The measured beam profiles imply that the ΔI decreases by ∼2.8 dB during scanning, which is determined by the aperture size of a single emitter element.
Fig. 9. (a) Test setup for measuring the emitted far-field beam profiles. (b) Microscope image of the fabricated dispersive OPA incorporating a DL array. Captured beam profiles for scanning angles of (c) 0°, (d) 15.6°, and (e) 31°.
Fig. 10. (a) Captured far-field beam profiles emitted from dispersive OPA during beam scanning. (b) Corresponding horizontal cross-sectional beam profiles.
To further scrutinize the emission characteristics of the DL array, additional OPA units with different channel spacings and length differences were constructed. The emission angles of the dispersive OPA for wavelengths ranging from 1530 to 1600 nm are shown in Fig. 11. The solid lines refer to the fitted values, indicating the scanning efficiency. The measured scanning rates (dΨ/dλ) were witnessed to be 0.73°/nm and 0.92°/nm for ΔL = 20 µm and 25 µm, respectively, under a fixed Λch of 2 µm, as shown in Fig. 11(a) and (b). All beams scan in the negative direction (-Ψ) with increasing wavelength, and a longer ΔL leads to a higher dΨ/dλ, aligning with expectations. It is noteworthy that some abrupt changes in the emission angle occurred in all four scenarios during the beam scanning. When the original main lobe scans beyond the designed FOV range in the negative direction, it is considered a grating lobe owing to its lower peak intensity. Simultaneously, a new main lobe appears at the maximum position in the positive direction of the FOV and scans continuously, causing sudden shifts in emission angles at the FOV boundary. Figures 11(c) and (d) present the measured emission angle for Λch = 3 µm and 1.5 µm, respectively, when ΔL is fixed at 20 µm. The maximum achievable scanning angles were discovered to be nearly ±15° and ±31°, with scanning rates of 0.48°/nm and 0.97°/nm. For the manufactured dispersive OPA unit, the ΔΨFWHM as well as the required wavelength region (ΔλFOV) for beam scanning to cover the entire FOV is documented in Table 2. It should be mentioned that the beam scanning performance was inspected by manually changing the incident wavelength of a tunable laser. Hence, the precise scanning speed should be ultimately estimated based on the inherent operation characteristics of the laser. This work was primarily meant to validate the scanning performance of the proposed AOML OPA. Considering a tunable laser chip can be integrated into the proposed OPA device in the future, a fast wavelength-tunable beam scanning is anticipated to be potentially accomplished [27,28].
Fig. 11. Measured emission angles of the dispersive OPA for cases of (a) Λch = 2 µm, ΔL = 20 µm, (b) Λch = 2 µm, ΔL = 25 µm, (c) Λch = 3 µm, ΔL = 20 µm, and (d) Λch = 1.5 µm, ΔL = 20 µm.
Table 2. Measured emission response of the designed dispersive OPA unit
The suggested optical switch was first fabricated and experimentally characterized. The setup for testing the output beam profile and insertion loss of the manufactured optical switch is depicted in Fig. 12(a) and (b), respectively. As elucidated in the characterization of the emitted far-field profiles, the incident light generated by the tunable laser passed through the PC and was coupled to the fabricated chip via UHNA4 fiber. The output beam profiles of the three ports were concurrently monitored using a SWIR camera connected to a 20× objective lens (Mitutoyo Plan Apo), as illustrated in Fig. 12(a). Figure 12(b) shows the setup for checking the characteristics of optical transmittance. The light traversing the fabricated optical switch emerged from the bar and cross port and was then coupled back to the UHNA4 fiber through the SSC. The power measured at the ports was finally recorded using an optical power meter (PM100D; Thorlabs). Figure 12(c) displays a microscopic image of the switch device, comprising MMI couplers (MMI 1 and MMI 2), a TiW alloy heater, and electrode pads. Figures 12(d) (i)(ii)(iii) exhibit the beam profiles captured when light emerges from Ports 1, 2, and 3. For the optical switch, the launched light was directed toward the three ports by applying a current to induce a phase shift, facilitating efficient beam path selection. The TiW alloy heater was positioned atop the waveguide with a length of 1500 µm and a width of 5 µm. The measured power consumption for the π-shift (Pπ) was 158 mW. Figures 12(e), (f), and (g) show the measured transmissions for the bar and cross ports when the beam was routed toward O1, O2, and O3, respectively, for wavelengths running from 1530 to 1600 nm. The measured insertion losses for the bar ports were 0.9 ∼ 1.1 dB, 0.9 ∼ 1.5 dB, and 1.2 ∼ 1.6 dB for the three cases, respectively. The extinction ratio was consistently monitored to be over 20 dB throughout the spectral band.
Fig. 12. (a) Experimental setup for measurement of beam profiles and (b) transmittance of optical switch. (c) Microscope images of manufactured device. (d) Captured mode profiles from bar and cross channels. Measured transmittance of (e) O1, (f) O2, and (g) O3 with wavelength ranging from 1530 to 1600 nm.
In an effort to complete the AOML OPA, a simplified polishing scheme has been particularly attempted to achieve a 45° facet at the edge of the OPA chip. Toward that end, a specialized fixture was custom-built to adeptly secure chips at an oblique angle of 45° and facilitate subsequent polishing processes. The implementation of this bevel polishing technique is judged to be unequivocally useful for creating a uniformly flat polished facet. Additionally, the impact of uneven chip facets on the emission efficiency was observed through simulations. The emission loss was discovered to be below 0.8 dB when the unintended gap between the actual and desired edges is no more than 4 µm, underpinning a stable emission efficiency under the uneven chip facet. Figure 13 shows a microscopic image of the constructed AOML OPA, consisting of three OPA units, each offset by 45° from the others. As previously mentioned, the propagating guided beam was first routed through an optical switch and subsequently transmitted through a dispersive OPA unit, which incorporates an SSC, five-stage MMI power splitter, and DL array. The measured emission angles of the proposed device with respect to the wavelength are given in Fig. 14. The three constituting OPA elements were combined and adaptively addressed through an optical switch, thereby initiating the selection of the beam propagation path. As a result, we could attain a total scanning region of 152°. The scanning regions of the three OPA units were assessed to encompass −76° to −14°, −31° to 31°, and 14° to 76°, yielding an overlapping scanning region. The periodic diffraction envelope generated by the multiline OPA units serves to mitigate peak intensity fluctuations within the main lobe across the expansive FOV. And a uniform scanning rate of 0.97°/nm was experimentally identified. The solid line denotes the fitted values of the measurements, reflecting the scanning efficiency. As anticipated, the experimental outcomes were in close correlation with the simulation results, underscoring the accuracy and effectiveness of the AOML OPA design.
Finally, beam scanning was successfully fulfilled by finely tuning the wavelength and inducing a phase shift via the thermo-optic effect of the optical switch. The captured horizontal cross sections of the emitted beam profiles are shown in Fig. 15(a). A scanning range of about 152° was ultimately realized, validating the combined scanning region of the three OPA units as intended. Figure 15(b) presents the measured peak intensity of the main lobe in terms of the scanning angle, which implies that the fluctuation of ΔI varies within about 2.8 dB over the scanning range from −76° to 76°, indicating an enlarged FOV with plateau envelope. Consequently, it was confirmed the AOML OPA could achieve efficient beam scanning by adjusting the wavelength while adaptively selecting the beam path via the optical switch, rendering a FOV of 152° with ΔI of approximately 2.8 dB. The corresponding maximum scanning range was ±31.1° under a channel spacing of 1.5 µm. Regarding the proposed AOML OPA, the offset angle (α) of the OPA units (OPA 1, 2, and 3) were set at 45° to comprehensively examine the entire emission response. Consequently, a total scanning region of 152° was attainable through the application of the optical switch. The scanning regions of the three OPA units were measured to span from −76° to −14° (OPA 1), −31° to 31° (OPA 2), and 14° to 76° (OPA 3), thus providing overlapping scanning ranges. It should be mentioned that the scanned beam is sequentially emitted from OPA 1, 2, and 3 by applying the optical switch and exhibits no mutual influence. It is known that the peak intensities of the grating lobes vary according to a Gaussian-shaped envelope as determined by a single emitter channel belonging to an OPA with uniform channel spacing. The maximum scanning angle in the negative direction appears for Δφ = -π. As Δφ changes from -π to 0, the intensity of the main lobe gradually increases along the envelope whereas the intensity of the grating lobes decreases. It is observed at Δφ = 0 that the grating lobes are extremely tenuous relative to the main lobe. As Δφ changes from 0 to π, the intensity of the main lobe gradually diminishes whereas the intensity of the grating lobes increases. The intensity levels of the main beam and grating lobe become equivalent for Δφ = ±π, beyond which the main lobe cannot be practically discerned as against the grating lobe from the perspective of the detector. In this context, the FOV is effectively determined to be the angular region between the two identical beams corresponding to Δφ = ±π. Here, the side mode suppression ratio (SMSR) is defined as log10(Imain_max/Igrating_max), where Imain_max and Igrating_max show the maximum intensity of the main beam and grating lobe, respectively. The SMSR decreases from 33 to 0 dB when Δφ increases from 0 to π. For the proposed AOML OPA, the smallest SMSR is 0 dB and 6 dB leading to FOVs of ±76° and ±67.5°, respectively. An alias-free beam scanning, immune to detrimental grating lobes, can be achieved by capitalizing on an aperiodic antenna array or incorporating more OPA units with smaller offset angles [12,13].
Fig. 15. (a) Captured horizontal cross-sectional beam profiles of integrated AOML OPA. (b) Corresponding main lobe peak intensities.
4. Conclusion
The proposed AOML OPA has been successfully demonstrated to provide efficient dispersive beam scanning with a substantial FOV and plateau envelope by integrating the multiline OPA units, which were arranged at an offset angle of 45° through an optical switch. The constituting OPA units with an angular offset yielded a broad horizontal FOV by consolidating the capabilities of three consecutive scanning ranges through the path selection. And the presence of periodically appearing diffraction envelopes significantly contributed to reducing the intensity fluctuation of the main lobe throughout the beam scanning process. Each OPA unit was outfitted with a dispersive DL array featuring incrementally delay lengths, facilitating wavelength-tunable beam scanning with an efficiency of 0.97°/nm. Finally, the suggested AOML OPA achieved an FOV of 152° with ΔI of less than 2.8 dB experimentally by switching the propagation path and using wavelengths ranging from 1530 to 1600 nm. The optical switch, featuring MMI couplers, was experimentally characterized, demonstrating exceptional broadband performance with insertion loss of less than 1.5 dB within a spectral bandwidth of 70 nm. In prospective developments, the augmentation of the offset angle and diminution of Λch within LBE can be employed to systematically enhance the FOV. Additionally, the reduction in both thickness and width of the SiN core holds the potential for generating a more expansive diffraction envelope, thereby leading to a consequential decrease in the main lobe intensity. The suggested AOML OPA, capable of efficient dispersive beam scanning with a wide FOV and plateau envelope, is anticipated to be a promising strategy for advanced LiDAR system applications and rapid line beam scanning.
Funding
National Research Foundation of Korea (2020R1A2C3007007); Agency for Defense Development (915027201).
Acknowledgment
This work was supported by the Challengeable Future Defense Technology Research and Development Program through the Agency for Defense Development (ADD) funded by the Defense Acquisition Program Administration (DAPA) in 2023 (No. 915027201) and the National Research Foundation of Korea (NRF) funded by the Ministry of Science and ICT (No. 2020R1A2C3007007).
Disclosures
The authors declare no conflicts of interest.
Data availability
The datasets generated and/or analyzed during the current study are available from the corresponding author upon reasonable request.
References
1. J. C. Hulme, J. K. Doylend, M. J. R. Heck, et al., “Fully integrated hybrid silicon two-dimensional beam scanner,” Opt. Express 23(5), 5861–5874 (2015). [CrossRef]
2. S. Chung, H. Abediasl, and H. Hashemi, “A monolithically integrated large-scale optical phased array in silicon-on-insulator CMOS,” IEEE J. Solid-State Circuits 53(1), 275–296 (2018). [CrossRef]
3. C. T. Phare, M. C. Shin, S. A. Miller, et al., “Silicon optical phased array with high-efficiency beam formation over 180 degree field of view,” arXiv, arXiv:1802.04624 (2018). [CrossRef]
4. C.-S. Im, S.-M. Kim, K.-P. Lee, et al., “Hybrid integrated silicon nitride–polymer optical phased array for efficient light detection and ranging,” J. Lightwave Technol. 39(13), 4402–4409 (2021). [CrossRef]
5. J. K. Doylend, M. J. R. Heck, J. T. Bovington, et al., “Two-dimensional free-space beam steering with an optical phased array on silicon-on-insulator,” Opt. Express 19(22), 21595–21604 (2011). [CrossRef]
6. M. Raval, A. Yaacobi, and M. R. Watts, “Integrated visible light phased array system for autostereoscopic image projection,” Opt. Lett. 43(15), 3678–3681 (2018). [CrossRef]
7. P. Wang, G. Luo, Y. Xu, et al., “Design and fabrication of a SiN-Si dual-layer optical phased array chip,” Photonics Res. 8(6), 912–919 (2020). [CrossRef]
8. C.-S. Im, B. Bhandari, K.-P. Lee, et al., “Silicon nitride optical phased array based on a grating antenna enabling wavelength-tuned beam steering,” Opt. Express 28(3), 3270–3279 (2020). [CrossRef]
9. Q. Wang, S. Wang, L. Jia, et al., “Silicon nitride assisted 1×64 optical phased array based on SOI platform,” Opt. Express 29(7), 10509–10517 (2021). [CrossRef]
10. N. A. Tyler, D. Fowler, S. Malhouitre, et al., “SiN integrated optical phased arrays for two-dimensional beam steering at a single near-infrared wavelength,” Opt. Express 27(4), 5851–5858 (2019). [CrossRef]
11. W. D. Sacher, Y. Huang, G.-Q. Lo, et al., “Multilayer silicon nitride-on-silicon integrated photonic platforms and devices,” J. Lightwave Technol. 33(4), 901–910 (2015). [CrossRef]
12. D. N. Hutchison, J. Sun, J. K. Doylend, et al., “High-resolution aliasing-free optical beam steering,” Optica 3(8), 887–890 (2016). [CrossRef]
13. M. C. Shin, A. Mohanty, K. Watson, et al., “Chip-scale blue light phased array,” Opt. Lett. 45(7), 1934–1937 (2020). [CrossRef]
14. N. Dostart, B. H. Zhang, A. Khilo, et al., “Serpentine optical phased arrays for scalable integrated photonic lidar beam steering,” Optica 7(6), 726–733 (2020). [CrossRef]
15. C. V. Poulton, M. J. Byrd, P. Russo, et al., “Long-range LiDAR and free-space data communication with high-performance optical phased arrays,” IEEE J. Sel. Top. Quantum Electron. 25(5), 1–8 (2019). [CrossRef]
16. M. J. R. Heck, “Highly integrated optical phased arrays: Photonic integrated circuits for optical beam shaping and beam steering,” Nanophotonics 6(1), 93–107 (2017). [CrossRef]
17. Y. Zhang, Y.-C. Ling, K. Zhang, et al., “Sub-wavelength-pitch silicon-photonic optical phased array for large field-of-regard coherent optical beam steering,” Opt. Express 27(3), 1929–1940 (2019). [CrossRef]
18. S. A. Miller, Y.-C. Chang, C. T. Phare, et al., “Large-scale optical phased array using a low-power multi-pass silicon photonic platform,” Optica 7(1), 3–6 (2020). [CrossRef]
19. W. Xu, L. Zhou, L. Lu, et al., “Aliasing-free optical phased array beam-steering with a plateau envelope,” Opt. Express 27(3), 3354–3368 (2019). [CrossRef]
20. M. R. Kossey, C. Rizk, and A. C. Foster, “End-fire silicon optical phased array with half-wavelength spacing,” APL Photonics 3(1), 011301 (2018). [CrossRef]
21. D. Wu, Y. Yi, and Y. Zhang, “High-efficiency end-fire 3D optical phased array based on a multi-layer Si3N4/SiO2 platform,” Appl. Opt. 59(8), 2489–2497 (2020). [CrossRef]
22. B. Bhandari, C. Wang, J.-M Heo, et al., “Dispersive silicon–nitride optical phased array incorporating arrayed waveguide delay lines for passive line beam scanning,” Sci. Rep. 12(1), 18759 (2022). [CrossRef]
23. R. Cheng, D. Zhang, J. Wang, et al., “Fluorinated photopolymer cascaded MMI-based integrated optical waveguide switching matrix with encoding functions,” Opt. Express 27(9), 12883–12898 (2019). [CrossRef]
24. F. Duan, K. Chen, D. Chen, et al., “Low-power and high-speed 2× 2 thermo-optic MMI-MZI switch with suspended phase arms and heater-on-slab structure,” Opt. Lett. 46(2), 234–237 (2021). [CrossRef]
25. Y. Gao, X. Sun, P. Li, et al., “Polymer mode selecting switch based on cascaded MMI couplers,” IEEE Photonics Technol. Lett. 33(3), 147–150 (2021). [CrossRef]
26. W. Song, R. Gatdula, S. Abbaslou, et al., “High-density waveguide superlattices with low crosstalk,” Nat. Commun. 6(1), 7027 (2015). [CrossRef]
27. Y. Gao, J.-C Lo, S. Lee, et al., “High-power, narrow-linewidth, miniaturized silicon photonic tunable laser with accurate frequency control,” J. Lightwave Technol. 38(2), 265–271 (2020). [CrossRef]
28. C. Xiang, J. Guo, W. Jin, et al., “High-performance lasers for fully integrated silicon nitride photonics,” Nat. Commun. 12(1), 6650 (2021). [CrossRef]
