High-Directionality Silicon Nitride Antenna Based on Distributed Bragg Reflector for Optical Phased Array

Abstract

Optical phased arrays (OPAs) have great potential in the fields of integrated solid-state light detection and ranging. The ranging distance of an OPA can be further enlarged by improving the directionality of the grating antenna. A high-directionality silicon nitride grating antenna with a distributed Bragg reflector (DBR) is proposed. The DBR consists of a stack of silicon nitride and silicon dioxide layers, which are utilized as the bottom reflectors to further reduce downward radiation. In a simulation, the directionality of the antenna exceeded 71.6% within the wavelength range of 1420–1740 nm. Additionally, the directionality of the antenna can achieve 97.6% at 1550 nm. Compared to a grating antenna without a DBR, the directionality is improved by 1.52 dB. Moreover, the proposed silicon nitride grating antenna has a large fabrication tolerance and is compatible with CMOS fabrication techniques, showing great potential for enhancing the performance of the integrated optical phased array.

1. Introduction

The integrated optical phased array (OPA) which uses silicon photonic technology has attracted tremendous attention as a non-mechanical beam steering device due to its high level of integration and compatibility with CMOS technology [1,2,3]. It has been widely used in various applications that require precise beam steering and high directionality, including light detection and ranging (LiDAR) [4,5], optical communication [6], imaging [7,8], and biosensors [9]. The grating antenna is one of the key components of the integrated OPA which can convert the light in the waveguide to free-propagating optical radiation and vice versa [10]. Normally, the maximum ranging distance of an OPA is directly determined by the output power of the grating antenna [11,12,13]. However, the silicon-based grating antenna cannot sustain a very high input optical power due to its nature of two-photon absorption and free carrier absorption [14], which limits its capability in handling high-power signals. In contrast, silicon nitride exhibits high-power-handling capabilities with low nonlinear absorption and low-loss-propagating properties in a broad wavelength range [15,16], making it a promising alternative material to enhance the performance of the grating antenna.

Moreover, the output power can be further enhanced by fully using the input optical power, i.e., by improving the directionality of the grating antenna. The directionality is defined as the ratio of upward light power to the total out optical power; however, there is an unavoidable downward radiation of power to the substrate, resulting in poor directionality. To solve this problem, various approaches have been proposed. The maximum upward directionality of the antenna can reach over 90% through the use of a dual-level or three-layer grating antenna to suppress the bidirectional radiation, but the fabrication complexity is significantly increased [13,17]. The other approach is to use bottom reflectors such as metal mirrors or bottom gratings to block the downward radiation [18,19]. However, these traditional reflectors are fragile at high temperatures, and the fabrication process is relatively complex and incompatible with CMOS fabrication.

In this paper, a high-directionality silicon nitride grating antenna is proposed which consists of a silicon nitride grating with a low nonlinear effect and large damage threshold and a bottom reflector. The DBR is composed of a stack of silicon nitride, and silicon dioxide is used as the bottom reflector to reduce the downward radiation. The upward directionality can be improved to 97.6% at a wavelength of 1550 nm, offering a promising means of increasing the output power of the integrated optical antenna, with great potential for the integrated OPA.

2. Structure and Principle

A three-dimensional (3D) schematic of the silicon nitride antenna is shown in Figure 1a. It consists of the total thickness of the silicon nitride H_total = 400 nm and a distributed Bragg reflector (DBR). A cross-section of the silicon nitride antenna is illustrated in Figure 1b, where a DBR consisting of a multilayer stack is deposited on the silicon substrate, and a shallow etched silicon nitride grating is placed on the buried oxide (BOX) layer (H_BOX). Additionally, a silicon dioxide cladding layer is deposited to protect the device.

Beam steering on the $\theta$-axis can be achieved by changing the wavelength of the input light, and the corresponding relationship can be expressed by the first-order Bragg condition [20]:

$$sin\theta =\frac{\mathsf{\Lambda }·n_{eff}-\lambda }{\mathsf{\Lambda }·n_c}$$

(1)

where n_c is the refractive index of the cladding layer, $\lambda$ is the wavelength of the input light, $\mathsf{\Lambda }$is the period of grating, n_eff = ff∙n_eff1 + (1−ff)∙n_eff2 is the effective index of the waveguide grating, ff is the filling factor, which is defined as the ratio of grating teeth to the period, n_eff1 is the refractive index of the grating teeth (Si₃N₄), and n_eff2 is the refractive index of the etched region.

Beam steering on the $\phi$-axis can be achieved by changing the phase difference ($∆\varphi$) of the adjacent waveguides. If the phase difference between the adjacent waveguides is equal, the beam steering at $\phi$-axis can be calculated as [21]:

$$sin\phi =\frac{\lambda ·∆\varphi }{2\pi (d+w)}$$

(2)

where $∆\varphi$ is the uniform phase difference between adjacent channels, and d is the spacing width between the waveguides.

The beam width of a 2D array can be estimated by [21,22]:

$$\mathsf{\Delta }{\theta }_{FWHM}=\frac{0.886\lambda }{Lcos\theta },\mathsf{\Delta }{\phi }_{FWHM}=\frac{0.886\lambda }{Ndcos\phi }$$

(3)

where $\Delta {\theta }_{FWHM}$ and $\Delta {\phi }_{FWHM}$ are the full widths at half-maximum (FWHM) of the main lobes along the $\theta$-axis and $\phi$-axis, respectively, N is the number of the antennas, and L is the length of the antenna. Equation (3) shows that the beam width is determined by the aperture (Nd) of the optical antenna array. A narrow beam can be realized by increasing the number of optical antennas and the spacing width; however, the beam steering range is inversely proportional with the spacing width. Therefore, a trade-off is needed.

Figure 1b illustrates the various power distribution channels that occur when light is coupled to a silicon nitride antenna. P_up and P_down represent the upward and downward power, respectively, while P_T and P_R denote the transmitted and reflected power. In order to reduce P_down, a bottom reflector incorporating a multilayer DBR is employed.

3. Simulation and Optimizing

3.1. High Directionality Design

The three-dimensional (3D) finite-difference time-domain (FDTD) method is used to optimize antenna structures to achieve a high level of directionality. In a simulation, the fundamental transverse electric (TE) mode was used. Monitors were added in various directions to express the upward, downward, transmitted, and reflected power radiating from the grating antenna.

In order to obtain an optimal etching depth, the transmission power was calculated by varying H_t from 0 nm to 400 nm, as shown in Figure 2a, and H_t = 280 nm was chosen within the sweep range. Figure 2b illustrates the power distribution across various transmission directions as a function of filling factor variations. The filling factor affects the performance of the grating antenna by altering its effective index. The results indicate that an optimal filling factor is ff = 0.45.

A BOX layer between the grating antenna and the reflector was employed to prevent power leakage. The constructive interference between the reflected light and the upward radiated light can be achieved by changing the BOX thickness and therefore further reducing the loss. As shown in Figure 2c, the transmission power is highly dependent on the BOX thickness (H_box). According to the simulation, the most optimal BOX thickness is 1.53$\mathsf{\mu }$m. However, P_up varies slightly in the range of 1.53 $\mathsf{\mu }$m ± 60 nm. As shown in Figure 2d, although parameters such as the filling factor and etch depth were optimized, the downward radiation (P_down) still occupied quite a large portion of the total scattered power loss without a DBR. Then, a DBR consisting of a high-index material, Si₃N₄, and a low-index material, SiO₂, was added to further reduce the downward radiation. The maximum reflectivity can be achieved when the constructive interference occurs, and the thicknesses of the Si₃N₄ and SiO₂ layers are determined based on λ/4n [22], where λ represents the center wavelength of the source and n denotes the refractive index of materials. The thicknesses for Si₃N₄ and SiO₂ were calculated to be 194 nm and 268 nm at λ = 1550 nm, respectively.

As Figure 3a shows, the center wavelength was fixed at 1550 nm, the reflectivity increased proportionally with the number of layers N, and the curve tended to be smooth when N ≥ 6, which attained 98% at N = 8. In order to avoid the of vertically incident effect of light on the grating antenna, an angle = 9° was chosen, and the corresponding period was 1.2$\mathsf{\mu }\mathrm{m}$ according to Equation (1). The transmission spectrum is depicted in Figure 3b as wavelength sweeps, and the DBR structure exhibits a strong reflection effect in a quite large bandwidth.

The variations in the reflectivity and the thicknesses of the Si₃N₄ film and the SiO₂ film were calculated and are shown in Figure 4a,b. It can be seen that our proposed structure has quite a large fabrication tolerance.

Figure 5 illustrates the upward directionality of the antenna corresponding to the different wavelengths. It shows that the antenna exhibits an upward directionality exceeding 71.6% within the wavelength range of 1420–1740 nm. Moreover, the upward directionality value is 97.6% at a wavelength of 1550 nm. Compared to the grating antenna without a DBR, the directionality is improved by 1.52 dB. From the comparison,

Table 1. Comparison of the performances of the different antennas.

Ref.	Method	Highest Directionality	Wavelength Range
[23]	Dual-layer gratings antenna	96.7%	875–1100 nm
[24]	Dual-layer waveguide grating antenna	90%	1500–1600 nm
[25]	Two silicon nitride layers	93%	1550 nm
This work	Distributed Bragg reflector	97.6%	1420–1740 nm

, our proposed high-directionality silicon nitride grating antenna with a DBR demonstrates quite great performance.

3.2. Array Sweep Characteristics

The characteristics of the antenna array were simulated and demonstrated. Due to a server performance limitation, a four-channel OPA antenna with a spacing width (d) of 1 μm between the waveguides was adopted. Consequently, the FWHM of the main lobe along the $\phi$-axis ($\mathsf{\Delta }\phi$) of antenna appeared to be large. But by increasing the antenna aperture, the $\mathsf{\Delta }{\phi }_{FWHM}$ could be significantly reduced, as indicated by Equation (3).

The far field of the antenna at a wavelength of 1550 nm is simulated in Figure 6a, and a narrower beam in the = $\phi$-axis can be achieved by enlarging the spacing width. Figure 6b demonstrates that the proposed integrated Si₃N₄ grating antenna with a DBR can effectively control the emission angle along the $\theta$-axis by adjusting the wavelength from 1500 nm to 1600 nm. Figure 6c shows that beam steering on $\phi$-axis can be achieved by changing the phase difference of the adjacent waveguides.

4. Conclusions

In conclusion, a high-directivity Si₃N₄ grating antenna with a DBR is demonstrated. The directionality of the antenna is optimized by employing a DBR composed of a stack of silicon nitride and silicon dioxide as the bottom reflector. The results indicate that the directionality reaches 97.6% at a wavelength of 1550 nm and exceeds 71.6% within the wavelength range of 1420–1740 nm. The two-dimensional beam steering of our proposed antenna array is 4.09° × 6.4° (θ × φ) and can be further enlarged by utilizing a larger wavelength range and phase difference. Moreover, the optical antenna array with both a narrow beam and large steering range can be realized by using a nonuniform spacing width of the antenna array. This work provides a promising candidate for realizing high-directionality grating antennas in the integrated OPA.