top of page
Writer's pictureLatitude Design Systems

Integrated Optical Phased Array with On-Chip Amplification for Programmable Beam Shaping

Introduction

Optical phased arrays (OPAs) are an emerging technology that enables beam forming and steering in free space without any moving parts. These solid-state devices rely on the interference between multiple optical antennas and their relative phase differences to form a beam in any desired direction. OPAs offer several advantages over mechanical beam-steering techniques, such as faster beam steering, increased stability against vibrations, and the potential for reduced size, weight, and power consumption.

OPAs have been demonstrated on various photonic integration platforms, including Silicon, Silicon Nitride, and Indium Phosphide (InP). These integrated devices have achieved impressive performance metrics, such as high image resolution, sub-0.1° angular resolution, wide field of view (over 120°), and broad wavelength operation range (up to 120 nm). However, a common challenge faced by OPAs is the optical losses that limit the output power levels, which are typically caused by propagation losses, insertion losses, and fiber-to-chip coupling losses.

In this article, we will explore an InP-based OPA circuit that embeds on-chip optical amplifiers to overcome these losses and achieve high output power levels. Additionally, we will demonstrate how the independent control of amplitude and phase in the OPA arms enables programmable beam shaping, opening up new possibilities for applications such as LiDAR and optical wireless communication (OWC).

OPA Circuit Design

The OPA circuit is based on an InP generic photonic integration platform, which allows the monolithic integration of passive building blocks (waveguides, splitters, combiners) and active components such as semiconductor optical amplifiers (SOAs) and photodetectors. The circuit features an 8-arm OPA with a booster amplifier and independently controlled SOAs and phase modulators in each arm.

Figure 1 shows the schematic of the OPA circuit design. The input consists of a waveguide optical port for fiber coupling and a 500 μm long booster SOA. The light is then split into 8 equivalent arms through a star coupler with a Gaussian power distribution over its output waveguides. Each OPA arm comprises a 500 μm long SOA followed by a 2.2 mm long electro-optic phase modulator (EOPM). The OPA emits light in free space through a uniformly distributed array of edge-emitting waveguides optimized for beam scanning within a ±20.5° field of view (FoV).

(a) OPA circuit design schematic (not to scale). (b) Photograph of the OPA assembly showing the PIC on an Aluminium mount and wire-bonded to PCBs on each side. The chip contains two OPA PICs; the one used in this work is shown in the left half of the chip.
Figure 1. (a) OPA circuit design schematic (not to scale). (b) Photograph of the OPA assembly showing the PIC on an Aluminium mount and wire-bonded to PCBs on each side. The chip contains two OPA PICs; the one used in this work is shown in the left half of the chip.

On-Chip Amplification and Beam Steering in a Wide Wavelength Range

The net on-chip gain of the OPA was measured for wavelengths between 1465 nm and 1600 nm, with input power levels ranging from -20 dBm to 8.5 dBm and various SOA driving conditions. The net on-chip gain is defined as the difference between the absolute SOA gain and the insertion losses of passive components in the circuit.

Figure 2a shows the measured net gain, reaching up to 21.5 dB at 1525 nm wavelength, which is the highest gain reported for an OPA with on-chip amplification. Figure 2b illustrates the total output power level at the output facet of the PIC, with a maximum of 35.5 mW (15.5 dBm) at 1530 nm wavelength. This is also the highest output power level recorded for an active OPA.

The OPA demonstrated positive net on-chip gain for more than 125 nm, comparable to the state-of-the-art wavelength range of operation for OPAs. This broad wavelength operation is of interest for applications such as dispersion-based 2D beam steering and spectral LiDAR.

(a) Measured net gain of the Optical Parametric Amplifier (OPA). (b) Total output power level at the output facet of the Photonic Integrated Circuit (PIC) as a function of on-chip input power. The booster Semiconductor Optical Amplifier (SOA) and arrayed SOAs were driven at 10 kA/cm² and 7.5 kA/cm², respectively. The legend in (b) applies to (a) as well. Not all wavelengths in the sweep are shown for clarity. (c–f) Comparison of gain measured through the arrayed SOAs (solid lines) and consecutively driven SOAs (dashed lines). In (c–e), the SOAs are driven at 10 kA/cm². In (f), the tunable laser wavelength is set to 1530 nm, and the current density is swept to understand gain deterioration due to thermal effects.
Figure 2: (a) Measured net gain of the Optical Parametric Amplifier (OPA). (b) Total output power level at the output facet of the Photonic Integrated Circuit (PIC) as a function of on-chip input power. The booster Semiconductor Optical Amplifier (SOA) and arrayed SOAs were driven at 10 kA/cm² and 7.5 kA/cm², respectively. The legend in (b) applies to (a) as well. Not all wavelengths in the sweep are shown for clarity. (c–f) Comparison of gain measured through the arrayed SOAs (solid lines) and consecutively driven SOAs (dashed lines). In (c–e), the SOAs are driven at 10 kA/cm². In (f), the tunable laser wavelength is set to 1530 nm, and the current density is swept to understand gain deterioration due to thermal effects.

Thermal Investigation of the OPA Performance

The performance of the SOA optical gain is influenced by thermal crosstalk between neighboring SOAs in the OPA arms. The heat generated by continuously driven SOAs at high current densities and their proximity (160 μm spacing) can limit the maximum gain.

Figure 2c-f compares the gain measured when all arrayed SOAs are simultaneously driven (solid lines) with the gain measured by consecutively driving each SOA while the others are reverse-biased to absorb light (dashed lines). By reducing thermal crosstalk, up to 3.5 dB higher gain can be achieved compared to the simultaneous driving case.

Thermal simulations (Figure S4, S5 in the Supplementary Material) estimate an average temperature increase of 6°C in the SOA core due to thermal crosstalk at the maximum current driving condition of 10 kA/cm². This temperature variation corresponds to approximately 2 dB gain difference, which aligns with the measured gain drop shown in Figure 2c-f.

Beam Steering

The OPA's calibration and beam steering were demonstrated in the broad wavelength range using the modified rotating element electric field vector (mREV) method. This calibration routine returns the reverse bias voltages required for the EOPMs to form a beam at the desired calibration angle in the far-field.

Figure 3a-c shows the 1D cross-section of the far-field beam at different steering directions within the FoV for various wavelengths. The measured full-width at half-maximum (FWHM) of the beam ranges from 4.6° to 5.5° across the steering and wavelength range, matching well with simulations (Figure 3e).

The sidelobe suppression ratio (SLSR), measured as the power ratio between the main lobe and the highest sidelobe, has a median value above 13.6 dB, with up to 1.3 dB increase in the 1465-1575 nm wavelength range (Figure 3d). The variation in SLSR across different wavelengths and steering angles is attributed to the wavelength-dependent compression of the Gaussian distribution power ratio and the electro-absorption of the EOPMs driven at different voltages.

(a–c) 1D cross section of the far-field beam at different steering directions within the FoV at (a) 1525 nm, (b) 1465 nm, and (c) 1600 nm wavelengths. (d) Measured sidelobe suppression ratio (SLSR). For each wavelength, the median value across different steering angles is shown inside the box plots. (e) Measured beam width (FWHM, box plots). The dashed blue curve shows the simulated beam width for a fixed Gaussian distribution power ratio of 3.3 dB. The solid red curve represents the simulated beam width for a wavelength dependent Gaussian distribution power ratio.
Figure 3. (a–c) 1D cross section of the far-field beam at different steering directions within the FoV at (a) 1525 nm, (b) 1465 nm, and (c) 1600 nm wavelengths. (d) Measured sidelobe suppression ratio (SLSR). For each wavelength, the median value across different steering angles is shown inside the box plots. (e) Measured beam width (FWHM, box plots). The dashed blue curve shows the simulated beam width for a fixed Gaussian distribution power ratio of 3.3 dB. The solid red curve represents the simulated beam width for a wavelength dependent Gaussian distribution power ratio.

Programmable Beam Shaping through Amplitude and Phase Control

The independent control of SOAs in the OPA arms enables programmable beam shaping by varying the power distribution profile of the OPA emitter. This capability can be used to suppress diffraction sidelobes and generate complex beam shapes, such as Hermite-Gaussian (HG) beams.

Sidelobe Suppression through Emitter Power Distribution Shaping

By tuning the currents in the arrayed SOAs, the Gaussian power profile of the OPA emitter can be modified to achieve higher sidelobe suppression ratios (SLSR). Figure 4a shows the 1D near-field profile of the OPA emitter, where the power ratio between inner and outer arms is swept from 2 dB to 12 dB.

After optimizing the power profile and calibrating the phase distribution, up to 19.8 dB SLSR was demonstrated in the far-field main lobe (Figure 4b,c). This is 10 dB higher than typical values reported in the literature and the highest SLSR recorded for active OPAs.

(a) 1D near-field profile of the OPA emitter. The power ratio between inner and outer arms is swept from 2 dB to 12 dB. (b) Main far-field lobe (steering at 0 â—¦ ) and (c) beam properties for different Gaussian distribution power ratio values. The crosses and solid lines represent the measurement data and simulation results respectively. The triangles represent the measured FWHM and SLSR when the arrayed SOAs are driven with the same current density of 7.5 kA/cm2. All measurements in this figure were performed at 1530 nm wavelength.
Figure 4. (a) 1D near-field profile of the OPA emitter. The power ratio between inner and outer arms is swept from 2 dB to 12 dB. (b) Main far-field lobe (steering at 0 â—¦ ) and (c) beam properties for different Gaussian distribution power ratio values. The crosses and solid lines represent the measurement data and simulation results respectively. The triangles represent the measured FWHM and SLSR when the arrayed SOAs are driven with the same current density of 7.5 kA/cm2. All measurements in this figure were performed at 1530 nm wavelength.

Generation of 1D Hermite-Gaussian Beams

The independent control of amplitude and phase in the OPA arms allows the generation of complex beam shapes, such as Hermite-Gaussian (HG) beams. Figure 5a shows the theoretical amplitude and intensity profiles for 0th, 1st, and 2nd order HG beams.

By replicating these profiles in the near-field power distribution and phase of the OPA arms (Figure 5b), far-field HG beams were generated with more than 15 dB SLSR (Figure 5c). Beam steering within the FoV was also demonstrated for the 1st order HG mode (Figure 5d).

(a) Theoretical Hermite–Gauss (0, 1, 2 order) amplitude (E-field) and intensity (|E|²) profiles, with negative amplitude portions highlighted in red. (b) Near-field and (c) far-field images of the OPA with different power distributions in the arms, following Hermite–Gauss (HG) beams of orders 0, 1, and 2. Black plots show HG beams with uniform power distribution. Red plots show HG beams with controlled amplitude and phase in the OPA arms to achieve the desired beam. Near-field emitter waveguide modes with a π phase shift for negative amplitude are highlighted in red. (d) Beam steering of the HG1 beam at four angular positions.
Figure 5: (a) Theoretical Hermite–Gauss (0, 1, 2 order) amplitude (E-field) and intensity (|E|²) profiles, with negative amplitude portions highlighted in red. (b) Near-field and (c) far-field images of the OPA with different power distributions in the arms, following Hermite–Gauss (HG) beams of orders 0, 1, and 2. Black plots show HG beams with uniform power distribution. Red plots show HG beams with controlled amplitude and phase in the OPA arms to achieve the desired beam. Near-field emitter waveguide modes with a π phase shift for negative amplitude are highlighted in red. (d) Beam steering of the HG1 beam at four angular positions.

Conclusion

In this article, we have explored an InP-based integrated optical phased array (OPA) circuit with on-chip amplification, enabling programmable beam shaping. The key features and achievements of this work include:

  1. Up to 21.5 dB net on-chip gain and 35.5 mW optical output power, the highest recorded for an active OPA.

  2. Broad wavelength operation range of 125 nm (1465-1590 nm), suitable for applications like dispersion-based 2D beam steering and spectral LiDAR.

  3. Investigation of thermal crosstalk between SOAs in the OPA arms, showing a potential gain improvement of 3.5 dB by minimizing thermal effects.

  4. Demonstration of beam steering within a ±20.5° field of view, with median sidelobe suppression ratio above 13.6 dB across the wavelength range.

  5. Programmable beam shaping through independent control of amplitude and phase in the OPA arms,


Reference

[1] M. Gagino et al., "Integrated optical phased array with on-chip amplification enabling programmable beam shaping," Scientific Reports, vol. 14, no. 9590, May 2024, doi: 10.1038/s41598-024-60204-5.

Comments


bottom of page