A Wide Field of View and Broadband Infrared Imaging System Integrating a Dispersion-Engineered Metasurface

Abstract

We present a compact hybrid imaging system operating in the 3–5 μm spectral band that combines refractive optics with a dispersion-engineered metasurface to overcome the longstanding trade-off between wide field of view (FOV), system size, and thermal stability. The system achieves an ultra-wide 178° FOV within a total track length of only 28.25 mm, employing just three refractive lenses and one metasurface. Through co-optimization of material selection and system architecture, it maintains the modulation transfer function (MTF) exceeding 0.54 at 33 lp/mm and the geometric (GEO) radius below 15 μm across an extended operational temperature range from –40 °C to 60 °C. The metasurface is designed using a propagation phase approach with cylindrical unit cells to ensure polarization-insensitive behavior, and its broadband dispersion-free phase profile is optimized via a particle swarm algorithm. The results indicate that phase-matching errors remain small at all wavelengths, with a mean value of 0.11068. This design provides an environmentally resilient solution for lightweight applications, including automotive infrared night vision and unmanned aerial vehicle remote sensing.

1. Introduction

Infrared imaging technology, as a core component of modern optoelectronic sensing systems, plays an irreplaceable role in fields such as military reconnaissance, aerospace remote sensing, and environmental monitoring. Among them, the 3–5 μm spectral band has high radiation sensitivity to high-temperature targets and is less affected by aerosol scattering during atmospheric transmission, making it suitable for long-range detection in complex environments. However, with the continuous increase in the performance requirements for optical systems in application scenarios, traditional refractive infrared optical systems face multiple contradictions in terms of field of view (FOV), volume, and aberration correction capabilities [1,2]. The demand for a wide FOV often comes with issues such as increased system structural complexity, weight gain, and deterioration of chromatic aberration, which severely restricts the deployment capabilities of infrared payloads in space-constrained scenarios such as unmanned aerial vehicles and satellite platforms.

The emergence of metasurfaces provides an innovative solution to this long-standing challenge. Metasurfaces are capable of flexibly and efficiently manipulating the amplitude, phase, and polarization states of light [3,4,5,6,7,8], and have demonstrated great potential in aspects such as abnormal reflection and refraction [9,10], imaging [11,12], and holographic display [13,14,15]. In recent years, metasurfaces have received significant attention in applications of wide FOV imaging [16,17,18]. However, the inherent material dispersion and angular sensitivity fundamentally limit the operational bandwidth of wide FOV metalenses [19,20,21]. To address the above challenges, on the one hand, deep learning-driven image reconstruction techniques have been employed to improve the imaging quality of metalenses [22,23]. This strategy can significantly boost the imaging signal-to-noise ratio and enrich image details [24]. Nevertheless, it is highly dependent on the network training condition. On the other hand, the limitations of the traditional forward design method can be overcome by reverse design, but the bandwidth and FOV of the metalens are only broadened to a limited extent [25,26].

In our previous work, we introduced a mid-wave infrared (MWIR) imaging system featuring an ultra-wide FOV based on a hybrid refractive meta-optical architecture [27]. A co-designed triplet refractive assembly and a metasurface component collectively achieve an optimal trade-off between compactness and performance. The system provides polarization-sensitive imaging over a 178° FOV within the 3–5 μm spectral range, while maintaining athermal performance across an extended temperature range from –40 °C to 60 °C. Refractive and meta-optics hybrid designs offer a promising pathway to circumvent the inherent constraints of single-element optics. Such configurations facilitate broadband aberration correction [28,29] and system miniaturization [30], as evidenced by recent advances in augmented reality displays [31] and novel optical systems [32,33,34], underscoring the viability and scalability of this integrated approach.

Building upon our previous research, this work presents a refractive and meta-optical hybrid imaging system featuring an ultra-wide FOV and polarization-insensitive operation for MWIR. Here, a dispersion-engineered metasurface was designed based on the propagation phase principle using polarization-insensitive cylindrical unit cells and optimized using the particle swarm optimization (PSO) algorithm, demonstrating excellent agreement between the optimized results and the target phase profile. Compared with the geometric metasurface adopted in earlier studies, the current system eliminates the need for the circular polarizer, resulting in a more compact structure, polarization-insensitive performance, and higher imaging efficiency. The study successfully validates the unique capability of metasurfaces technology in simplifying optical system architecture, providing a critical technical pathway for developing next-generation compact infrared imaging devices with broad spectral coverage and exceptional environmental adaptability.

2. Discussions and Results

2.1. Design and Analysis of the Optical Imaging System

This system employs only three refractive lenses and one metasurface to balance structural complexity with imaging performance. These refractive lenses serve as the primary imaging components, effectively bending light rays and enhancing imaging precision. The metasurface, positioned at the last surface, performs phase compensation. The optical performance of the imaging system was simulated using Ansys Zemax OpticStudio 2024 R1 software, with the final system layout shown in Figure 1. Key specifications include an operational waveband of 3–5 μm, F-number of 2, FOV of 178°, and effective focal length of 3.39 mm. In contrast to previous work [27], the circular polarizer is eliminated to simplify the system architecture. The configuration closely approximates an image-space telecentric design, which helps minimize measurement errors introduced during focus adjustments. Considering the rotational symmetry of the metasurface in this system, a binary 2 surface was selected to model its required phase distribution. The phase profile can be expressed as:

$$\phi =M_{order}{\displaystyle \sum _{i=1}^N{A}_i}{\rho }^{2i}$$

(1)

Here, φ represents the phase, M_order denotes the diffraction order, N is the number of terms in the polynomial series, ρ refers to the normalized radial aperture coordinate, and A_i is the coefficient of the 2i-th power of ρ. In this case, M_order is set to 1st order, N is 4, and ρ is the ratio of r to R, where r is the radius corresponding to different positions on the metasurface. R is the normalized radius, with R set to 1 mm. As a result, the values of A₁~A₄ are determined to be −23.759853, 1.831419, −0.091739, and 1.531296 × 10⁻³, respectively.

The detailed parameters of the system architecture are summarized in

Table 1. System structural parameters.

Surface Number	Surface Type	R/mm	T/mm	Material
S1	Plane	Infinity	1.944	KCl
S2	Sphere	4.348	7.778	——
S3	Stop	Infinity	0.972	——
S4	Sphere	−137.818	3.352	Si
S5	Sphere	−18.749	6.635	——
S6	Sphere	8.979	4.843	KRS5
S7	Sphere	24.985	0.972	——
S8	Plane	Infinity	1	Si
S9	Binary 2	Infinity	0.757	——
S10	Image plane	Infinity	——	——

, with lens materials sourced from the INFRARED material catalogs. All refractive lenses in the system adopt either spherical or planar surface profiles, significantly reducing structural complexity. In contrast to the geometric metasurface used in previous studies [27], the rear metasurface is designed based on the propagation phase principle, thereby eliminating the need for a circular polarizer and further reducing the total system length to 28.25 mm, a reduction of 0.75 mm. Furthermore, compared to previous research [33], we further optimized the lens materials and structural configuration to ensure stable operation within an ambient temperature range of −40 °C to 60 °C.

Owing to the rotational symmetry of the optical system, only the half-field range from 0° to 89° is illustrated herein, with the corresponding modulation transfer function (MTF) and spot diagram analyses presented to validate thermal stability and imaging performance at –40 °C, 0 °C, and 60 °C, as shown in Figure 2. The system’s ultimate cutoff frequency is determined by comparing diffraction-limited (ν₁ = 1/(λF/#) = 125 lp/mm at λ = 4 μm) and detector-limited (ν₂ = 1/(2p) = 33.3 lp/mm for 15 μm pixels) frequency thresholds. MTF curves in Figure 2a,c,e demonstrate stable performance across the operational temperature range, consistently exceeding 0.54 at ν₂. The corresponding spot analyses, as shown in Figure 2b,d,f, reveal that the root mean square (RMS) radius is smaller than the airy disk radius, and the geometric (GEO) radius is also smaller than the detector pixel size. This confirms the system’s capability of achieving high-quality imaging. Compared to our previous work [27], the proposed system exhibits a 0.01 improvement in the modulation transfer function at the cutoff frequency, along with a reduced point spread showing an average decrease of over 0.15 μm. This quantitative validation establishes the refractive and meta-optics hybrid architecture’s capacity to maintain diffraction-proximity performance under extreme environmental conditions while achieving compact system dimensions.

2.2. Characterization of the Metasurface

A propagation metasurface was designed to realize the phase distribution of the binary 2 surface shown in Figure 3a. Silicon material exhibits excellent transmittance in the mid-wave infrared band and possesses remarkable thermal conductivity, which is beneficial for the realization of the athermal design of the system. Therefore, an all-silicon architecture was implemented for the metasurface, utilizing cylindrical unit cells to realize a polarization-insensitive response, which is a widely used approach for this purpose [35,36], as shown in Figure 3b. The refractive index of silicon is from Ref. [37]. Numerical simulations were performed using the finite difference time domain (FDTD) method, with the incident wavelength range set from 3 to 5 μm. Periodic boundary conditions were applied along the x- and y-directions, while perfectly matched layers were implemented along the z-direction. Additionally, to achieve the phase modulation ranging from 0 to 2π, a parameter sweep of the unit cells was conducted, and a corresponding phase library was established. Considering practical fabrication constraints, the silicon post radius was varied from 100 to 500 nm with a fixed height of 4.5 μm and a periodicity p of 1.2 μm. The maximum aspect ratio of the unit structures reaches 22.5:1, which can be fabricated using the Bosch process [38,39]. The simulated transmittance and simulated phase shift within the wavelength range of 3 to 5 μm are presented in Figure 3c and Figure 3d, respectively. The results show that although there are cases of relatively low transmittance locally, the transmittance is still greater than 0.3, and the overall transmittance is at a high level. Moreover, all the provided phase shifts are greater than 2π. In contrast to the geometric metasurface employed in previous studies [27,33], the propagation metasurface proposed in this work not only achieves polarization-insensitive operation but also relatively improves the imaging efficiency of the system.

Furthermore, the designed metasurface needs to accurately match the target phase profile of the binary 2 surface. However, material dispersion introduces phase deviations in the metasurface. To overcome this limitation, a dispersion-engineered metasurface was designed to achieve precise phase compensation. The physical model can be mathematically described as follows [40]:

$${\varphi }_{\mathrm{optimized}}(\lambda )={\varphi }_{\mathrm{binary}2}(\lambda )+C(\lambda )$$

(2)

where ϕ_optimized denotes the optimized phase profile and C(λ) represents the wavelength-dependent compensation term. ϕ_binary2 corresponds to the intrinsic phase distribution of the binary 2 surface. A PSO algorithm was employed to optimize C(λ), addressing phase mismatch induced by material dispersion over a broad bandwidth. As an efficient global optimization algorithm, PSO is particularly suitable for handling high-dimensional, nonlinear phase modulation problems, and is capable of rapidly converging to near optimal solutions within a large parameter space. The algorithm iteratively updates particle positions, which represent candidate solutions, by simulating collective intelligence behavior, with the objective of minimizing a predefined cost function. Here, taking five discrete wavelengths (3 μm, 3.5 μm, 4 μm, 4.5 μm, 5 μm) as examples, the optimized values of C(λ₁)–C(λ₅) are 3.02, 5.29, 4.77, 5.69, and 1.88, respectively.

Figure 4 illustrates the matching results between the phase profiles at five discrete wavelengths and the optimized phase distribution ϕ_optimized. Although relatively large deviations are observed at a few isolated points, the overall agreement is excellent across the entire phase surface.

Table 2. Phase-matching errors.

Wavelength/μm	3	3.5	4	4.5	5
RMS	0.10314	0.11448	0.075609	0.11447	0.1457

shows the RMS phase-matching errors corresponding to different wavelengths. The results demonstrate that the phase-matching errors at all wavelengths remain relatively small, with a mean value of 0.11068, well within an acceptable range. It is worth noting that the matching performance could be further enhanced by expanding the phase library and employing more efficient optimization algorithms.

3. Conclusions

We present a refractive metasurface hybrid polarization-insensitive imaging system specifically designed for wide-field applications in the 3–5 μm spectral band. The system achieves enhanced compactness while maintaining an athermalized performance across the temperature range of −40 °C to 60 °C, and simultaneously delivering high-quality imaging over the 178° FOV. A broadband and polarization-insensitive metasurface is achieved via the dispersion engineering, accurately matching the target phase profile with minimal phase error. Compared to the geometric metasurface adopted in the earlier design, this approach not only simplifies the system architecture and reduces the total track length, but also achieves polarization-insensitive operation and enhances the overall imaging efficiency. Nevertheless, further avenues for improvement remain, specifically with regard to optimizing the system structure and lens materials for greater miniaturization and extending the athermalization temperature window. Future studies will concentrate on investigating the system’s tunability and its extension to broader spectral bands. This work provides a foundational framework for developing next-generation compact infrared imaging systems with multispectral fusion capability and enhanced environmental adaptability.