Broadband polarization- and angle-insensitive metamaterial absorber based on a hexagonal web lattice with Ni–AlN–Ni tri-layer structure for enhanced solar energy harvesting

Abstract

Broadband metamaterial absorbers (MMAs) are widely explored for photonic applications, yet many designs exhibit limited visible-range efficiency or polarization sensitivity. This work presents a nickel–aluminum nitride–nickel (Ni–AlN–Ni) MMA with a hexagonal web lattice resonator with radial connectivity in a metal–insulator–metal configuration. Finite-element electromagnetic simulations show an average absorption of 98.1% over 150–1400 nm (vacuum ultraviolet to near-infrared). In the visible range (380–780 nm), the average absorption is 99.4%, with peak absorption of 99.95% at 570 nm. Unlike conventional multiring resonators, the hexagonal web lattice with radial connectivity introduces multiple hybridized plasmonic modes and suppresses polarization conversion, which collectively enable the achievement of near-unity visible absorption. The absorber maintains absorption above 90% for incident angles up to 60° and shows polarization insensitivity. For terrestrial solar energy harvesting under the AM1.5 spectrum (280–2500 nm), the weighted absorption efficiency is 96.5%, rising to 98.656% in the visible range. These simulated performance metrics indicate potential applications in solar energy harvesting, thermal emission control, optical sensing, stealth technologies, and photothermal catalytic applications.

A Ni–AlN–Ni hexagonal web lattice MTM absorber achieves 98.1% average absorption (150–1400 nm) and 99.4% in the visible region, with 96.5% efficiency under the AM1.5 solar spectrum, ideal for solar harvesting and photonic applications.

1. Introduction

Metamaterials (MMs) are artificially engineered structures with extraordinary electromagnetic (EM) properties not found in natural materials. They enable precise control of EM wave propagation, attracting significant research attention and showing strong potential in antenna engineering,¹ stealth technologies,² biomedical devices,³ and satellite communications,⁴ demonstrating broad application prospects.

Among these developments, surface-based metamaterial absorbers (MMAs)⁵ have gained particular importance due to their versatility and efficiency. The first MMA was introduced by Landy et al. in 2008,⁶ laying the foundation for rapid advancements in MMA designs across multiple EM spectra, including narrowband⁷ and broadband⁸ configurations. The strong ability of MMAs to precisely manipulate EM waves has stimulated growing interest in extending their applications beyond the microwave–millimeter-wave–THz regime into the optical domain. In the optical regime, MMAs are widely applied in thermal imaging,⁹ filtering and detection,^10,11 sensing,¹² laser systems,¹³ and solar energy harvesting¹⁴ due to their ability to achieve strong light–matter interaction and highly selective absorption. Recent advances in this direction include a dual-stable ZnO–Ti–Si–W metamaterial absorber designed for medium-temperature solar thermal harvesting, demonstrating the growing interest in chemically robust and thermally stable absorber platforms.¹⁵

Significant advancements have been achieved in optical-regime MMA design; however, the Ni–SiO₂–Ni absorber reported in ref. 16, despite exhibiting high peak absorption, remains confined to a narrow visible wavelength range (415.2–625.1 nm), thereby excluding the UV and NIR regions. Similarly, the Ni–AlN–Ni absorbers in ref. 17 and 18 extend partially into visible, NIR, and SWIR regimes; however, both lack ultraviolet coverage and maintain moderate average absorption levels (93% and 90%, respectively), indicating incomplete broadband performance. The visible-focused designs in ref. 19 and 20 further emphasize this limitation. Although ref. 19 achieves a relatively high average absorption (97.17%), it is strictly confined to the visible range (380–780 nm), whereas ref. 20 provides only a narrow visible response (361.8–505 nm) with moderate efficiency and increased fabrication complexity due to the use of noble metals. Moving toward longer wavelengths,²¹ targets NIR–SWIR–MWIR regions but sacrifices UV-visible absorption and exhibits reduced angular stability (≤50°). Although ref. 22 reports extended spectral coverage (400–4000 nm), the absorption remains partial across bands and lacks a UV response, leading to incomplete utilization of the solar spectrum. The W–SiO₂–W absorber in ref. 23 achieves improved angular stability (≤80°) and covers visible-NIR wavelengths; however, it does not extend into the deep-UV region and introduces fabrication challenges associated with tungsten. Similarly, ref. 24 partially covers UVA-visible-NIR wavelengths (342–832 nm) but suffers from low average absorption (90%). Designs such as ref. 25 and 26 extend into the NIR–SWIR regime but exclude UV and parts of the visible spectrum, or exhibit only moderate absorption efficiency (≤93%) with uneven spectral response.

Most existing solutions either offer limited coverage in the visible spectrum or exhibit low absorbance efficiency. While some designs have successfully extended absorption into the IR region, they often compromise performance in the UV range. This trade-off is critical because a substantial portion of solar energy lies within the 280–1400 nm wavelength range, encompassing the UV, visible, and NIR regions. Therefore, there is a need for a high-efficiency broadband absorber that covers the full solar spectrum. Addressing this gap is essential to advancing solar energy-harvesting technologies.

In contrast, this study introduces a Ni–AlN–Ni MMA structure with an average absorption of 98.1% spanning the 150–1400 nm range, covering the vacuum ultraviolet (VUV) (partially), UVA, UVB, visible, and NIR (partially) spectral regions. Notably, it achieves an average absorption of 99.4% specifically in the visible range (380–780 nm), outperforming previous visible-focused absorbers in both breadth and efficiency. Additionally, it maintains polarization insensitivity and angular stability of ≥60°. This combination of a broad spectral range and strong angular and polarization robustness establishes this work as a significant advancement in optical absorber design, with high relevance for applications in solar energy harvesting.

The specific novelty of the present work lies not merely in the use of a hexagonal shape but in the radially connected web lattice topology. While previous hexagonal absorbers use isolated concentric rings, our design incorporates three hexagonal rings interconnected by radial lines. This web-like connectivity introduces additional electromagnetic coupling paths, leading to hybridized resonant modes and suppressed cross-polarization reflection. As will be demonstrated in Section 3.3, the progressive addition of rings and radial connections increases average absorption from ∼75% to >98%, proving that topological connectivity, rather than merely geometric complexity, is the key performance driver.

2. Design and simulation setup

2.1. Material selection

The proposed MMA operates efficiently from the VUV (partially) to the NIR (partially) region, encompassing the entire visible spectrum. It is important to note that for terrestrial solar applications, the AM1.5 spectrum contains negligible irradiance below 280 nm due to atmospheric ozone absorption. Therefore, the 150–280 nm portion of the absorber's bandwidth is primarily relevant for space-based or vacuum-UV photonic applications, while terrestrial solar harvesting utilizes the 280–2500 nm range. Given the high intensity of electromagnetic radiation in solar light, thermal stability becomes a crucial factor in absorber design. Ni has been selected for both the resonator and back metal layers due to its superior mechanical robustness, high thermal stability, and good plasmonic performance in the visible to NIR spectrum. Ni possesses a melting point of approximately 1455 °C, a thermal conductivity of 90.7 W m⁻¹ K⁻¹,²⁷ and an electrical conductivity of about 1.43 × 10⁷ S m⁻¹.²⁸ Additionally, Ni is a cost-effective and CMOS-compatible material, making it highly suitable for scalable, high-temperature photonic applications. Compared to noble metals such as gold and silver, Ni exhibits higher melting points, greater corrosion resistance, and greater chemical inertness. Aluminum nitride (AlN) is used as the dielectric spacer due to its high thermal conductivity, chemical stability, and wide bandgap. AlN maintains transparency across the 200–700 nm range and offers low optical losses in the UV-visible region, making it ideal for enhancing light confinement and facilitating strong electromagnetic coupling between the Ni layers. It features a wide bandgap of approximately 6.2 eV at room temperature, thermal conductivity of up to 285 W m⁻¹ K⁻¹, and a refractive index of around 2.12 in the optical spectrum.²⁹ Together, the Ni–AlN–Ni configuration demonstrates high absorption efficiency and environmental robustness, making it a compelling choice for broadband solar energy harvesting and optoelectronic applications.

2.2. Metamaterial absorber unit cell design

The proposed structure shown in Fig. 1 features a unit cell resembling a hexagonal web lattice that serves as a resonator. This configuration is patterned on a dielectric substrate. To ensure angular and polarization insensitivity, a symmetric design is being prioritized, leading to the development of a hexagonal web lattice structure. The design process involved placing three concentric hexagonal rings on the substrate, followed by adding three radial lines with equal angular spacing connecting the hexagons to form a web-like pattern. This design adopts a MIM-MMA architecture, with AlN as the dielectric layer and Ni as both the resonator and the back reflector. The parameter and the physical dimension of this design are depicted in Fig. 1(a) and Table 1, respectively. The square-shaped absorber, which has a length defined as (P = 200 nm). The inner hexagon diameter is defined as (P₁ = 150 nm), the middle hexagon has a diameter (P₂ = 175 nm), the outer hexagon diameter and connecting radial line length are defined as (P₃ = 200 nm), all with a consistent width of (W = 10 nm). The thicknesses of the resonator layer (t_w), dielectric substrate (t_s), and bottom metal layer (t_n) are 5, 12, and 5 nm, respectively.

Fig. 1. (a) Design composition and parameter of the proposed model. (b) Isometric view of the proposed structure.

Table 1. Parameter list of the proposed unit cell design with value in nm.

Parameters	P	P 1	P 2	P 3	W	t w	t s	t n
Value (nm)	200	150	175	200	10	5	12	5

2.3. Simulation setup

All electromagnetic simulations were performed using CST Studio Suite 2025 (Microwave and RF Studio – Frequency Domain Solver). The unit cell approach was employed with the following configuration:

2.3.1. Boundary conditions

Unit cell (master/slave) boundary conditions were applied in the x- and y-directions to model an infinite periodic array. A Floquet port with normal incident plane wave excitation (TE and TM polarizations) was placed above the structure in the z-direction. Open (absorbing) boundary conditions were used in the z-direction to eliminate reflections from the simulation domain boundaries.

2.3.2. Mesh convergence test

A tetrahedral surface-based mesh with adaptive refinement at material interfaces was used. Convergence was tested by monitoring absorbance at three representative wavelengths (310 nm, 570 nm, and 850 nm) as the total mesh cell count increased from 50 000 to 257 827. Absorption values stabilized to within 0.5% variation beyond 250 000 mesh cells. All reported simulations used a mesh with >250 000 cells, with a minimum mesh step of 0.59 nm near the Ni–AlN interfaces.

2.3.3. Optical constants

The wavelength-dependent complex refractive index ñ = n + ik (where n is the refractive index and k is the extinction coefficient) for nickel (Ni) was obtained from Palik's Handbook of Optical Constants of Solids.³⁰ For AlN, optical constants were taken from Li et al.²⁹ and cross-validated with the refractive index database.³¹ Both materials were modeled as lossy, frequency-dependent dielectric materials.

2.4. Possible fabrication method

A possible fabrication process for the proposed Ni–AlN–Ni MMA is illustrated in Fig. 2. The fabrication process of the proposed MMA follows a standard MIM architecture. Initially, a high-quality AlN substrate is prepared as the base layer due to its favorable thermal and electrical properties. A thin metallic Ni ground layer is then deposited on the substrate using DC magnetron sputtering or electron beam deposition, ensuring strong adhesion and uniform coverage. Subsequently, a second Ni layer is deposited on top, forming the continuous top metallic layer. The final absorber structure is realized by high-resolution electron-beam lithography (EBL), which patterns the desired hexagonal resonator design on the top Ni layer. This multi-step process enables the precise construction of the absorber geometry, which is critical for achieving broadband absorption across the targeted optical spectrum.

Fig. 2. Demonstration of possible fabrication process of the proposed absorber.

3. Result analysis and discussion

3.1. Absorption mechanism

The absorption performance of an MMA can be evaluated using its reflection and transmission characteristics. The general absorption formula is given by eqn (1),

A(ω) = 1 − R(ω) − T(ω) 1

where R(ω) = |S₁₁(ω)|² is the reflection coefficient, and T(ω) = |S₂₁(ω)|² is the transmission coefficient at a given angular frequency ω. For achieving perfect absorption, both reflection and transmission must approach zero. For optimal absorption, both reflection and transmission must be minimized. Transmission is effectively suppressed when the thickness of the bottom metallic layer exceeds the material's skin depth (δ), as defined by eqn (2),

2

where f is the frequency, µ is the permeability, and σ is the electrical conductivity of the metal. When the T(ω) is effectively suppressed to zero, the eqn (1) simplifies to eqn (3),

A(ω) = 1 − R(ω) 3

Hence, minimizing reflection becomes the primary design focus. Several approaches, including impedance matching and interference theory, are typically employed to reduce reflection and thus achieve near-unity absorption. By carefully engineering the geometry and material parameters of the metasurface, the impedance of the MMA can be matched to that of free space, thereby minimizing reflection and maximizing energy dissipation within the structure.

3.2. Impedance matching theory

As demonstrated by Pu et al.,³² the effective impedance approach is more physically meaningful than effective medium theory for describing the electromagnetic properties of metamaterial absorbers, particularly for wide-angle designs. According to this theory, perfect absorption occurs when the impedance of the metamaterial matches that of free space. Under such conditions, the structure reflects minimal energy. The reflection coefficient can be expressed as eqn (4),

4

here, Z_s(ω) denotes the surface impedance of the structure, and Z_free represents the impedance of free space (approximately 377 Ω). Reflection is minimized when Z_s(ω) = Z_free, while Z_free is a fixed quantity, Z_s(ω) can be engineered by altering the resonator design and material properties. Impedance (Z) is a function of both the material's permeability (µ) and permittivity (ε), as described by eqn (5),

5

Accordingly, the effective impedance (Z_eff) can be represented as eqn (6),

6

where µ₀ and ε₀ are the permeability and permittivity in the free space, respectively, and µ_r and ε_r represents relative permeability and relative permittivity of the material of the structure, respectively. From eqn (5) and (6), it is evident that the impedance depends on the relative permeability and permittivity of the materials comprising the structure. These parameters vary with frequency, leading to frequency-dependent changes in impedance and, consequently, in absorption.

Therefore, the absorption (A) as a function of frequency can be rewritten in terms of effective impedance as eqn (7),

7

This relationship underscores the significance of achieving impedance matching to optimize absorption, particularly in metamaterial absorber designs that target broad or specific spectral ranges. Eqn (7) implies that when Z_effective approaches Z_free, the absorption A(ω) reaches unity, leading to maximum absorptivity.

Based on these principles, the characteristics of the proposed MMA were evaluated using an EM simulation software, employing the impedance matching approach. The results are illustrated in Fig. 3 and 4. The broadband absorption exceeding 90% from 150 nm to 1400 nm is primarily attributed to effective impedance matching and strong intrinsic material losses. As shown in Fig. 3, the real part of the normalized impedance remains close (1–1.3), while the imaginary component stays near zero, minimizing reflection over the entire spectral range. Fig. 4 shows that the large imaginary components of permittivity (50–220) and permeability (80–400) indicate significant electric and magnetic losses, ensuring efficient energy dissipation within the absorber. Moreover, the transition of real components of permittivity and permeability from highly negative values to near-zero around the visible region enhances electromagnetic field confinement and facilitates multiple resonant absorption mechanisms.

Fig. 3. Normalized impedance (Z) with real part and imaginary part of the absorber.

Fig. 4. Demonstration of (a) permittivity (ε) and (b) permeability (µ) of the absorber.

Fig. 5 shows the proposed absorber exhibits broadband performance with transmission nearly zero across 100–3000 nm. Absorption remains above 90% from 150 nm through most of the NIR region (up to ∼1400 nm), with region-averaged absorptions of 98.1% (UV, 150–380 nm), 99.4% (visible, 380–780 nm), and 94.5% (NIR, 780–1400 nm). Beyond 1400 nm into the SWIR region (1400–2500 nm), absorption gradually decreases to 84–87% (average 84.7% over 1400–2500 nm). The statement “≥90%” applies strictly to the 150–1400 nm range; the extended SWIR range shows >80% absorption, consistent with the figure. Peak absorption is 99.95% at 570 nm for all polarization states. Moreover, the structure maintains an absorption rate above 99% across two bands: 168 nm to 175 nm and 304 nm to 590 nm, demonstrating broadband stability.

Fig. 5. Demonstration of absorption characteristics from UV to SWIR.

3.3. Design procedure

The absorption characteristics of the proposed MMA were analyzed through structural evolution, geometrical comparison, and layer stacking in Fig. 6.

Fig. 6. Demonstration of the absorbance characteristics with (a) structural evolution (b) various shapes. (c) Different layer.

Structural evaluation is done by considering four steps: (i) single hexagon ring, (ii) two concentric hexagon rings, (iii) three concentric hexagon rings, and (iv) radial connectivity of the rings (final web lattice). Fig. 6(a) illustrates the progressive improvement in absorbance. In step-1, a single hexagonal ring exhibits relatively weak absorption in the lower-wavelength region, with the absorbance initially remaining around 45–50% near 120 nm. The absorption gradually increases with wavelength and reaches approximately 70% near 180 nm and about 75% within the 250–300 nm range. Although resonance behavior is observed, the overall absorption bandwidth and efficiency remain limited due to insufficient resonance coupling.

In step-2, the structure is modified by introducing two concentric hexagonal rings. The additional ring within the first ring significantly enhances absorbance, reaching nearly 80% around 180 nm and exceeding 95% beyond approximately 250 nm. Compared to the single-ring structure, this configuration exhibits substantially improved broadband absorption.

But in step-3, where three concentric rings are used by adding an additional hexagonal ring inside two previous hexagonal rings (step-2), the absorbance is enhanced significantly. In this step, the absorbance remains above approximately 97–99% throughout most of the visible spectral region, extending from 300 nm to 800 nm. The spectral response also becomes smoother and more stable compared to the previous stages.

Finally, in step-4, three previous concentric rings with radial connecting lines form the final web lattice configuration. This optimized metasurface demonstrates near-perfect absorption across the visible-to-near-infrared region, with absorbance above 90% up to 1400 nm.

The critical novelty is observed when radial connections are added between hexagonal rings (final configuration). Compared to three isolated rings, the web lattice with radial interconnects increases absorption from 97.8% to 99.4% in the 380–780 nm range. This improvement originates from additional surface current paths (Fig. 19) that enable both electric and magnetic dipoles to contribute simultaneously, which is absent in non-connected ring designs. Thus, the hexagonal web lattice is not merely an incremental geometric change but a distinct topological architecture.

Fig. 19. Surface current distribution for TM mode at (a) 310 nm (b) 570 nm (c) 850 nm and TE mode at (d) 310 nm (e) 570 nm (f) 850 nm wavelength.

Fig. 6(b) compares the absorbance performance of square, triangular, octagonal, and proposed hexagonal resonators. The square geometry shows weaker short-wavelength absorption (50–60% at 120–150 nm), whereas triangular and octagonal shapes improve the broadband response, with 95–98% absorbance across the visible region. All the structures show less absorption performance than the proposed structure in the NIR region. The proposed hexagonal design achieves the most stable performance in confirming superior impedance matching and resonant coupling.

A clearer understanding of the absorption behavior is provided by separating the layer-specific contributions, as depicted in Fig. 6(c). In contrast, the simple Ni–AlN–Ni tri-layer, without designing the proposed top resonating layer, achieves a maximum absorption of about 88% near 300 nm but decreases to nearly 50% at longer wavelengths. The design without the top Ni layer, consisting only of AlN and Ni ground layers, shows a strong peak of approximately 98% around 700 nm, but fails to maintain high absorption across the full spectral range. However, after designing the top-layer hexagonal web lattice, the absorbance improves significantly across the entire operating range.

To validate that the 5 nm bottom Ni layer sufficiently suppresses transmission, a separate simulation for transmission coefficient T(λ) with different Ni bottom layer thickness was performed with the same unit cell geometry but without the top resonator layer (only bottom Ni + AlN + air). Fig. 7 shows the transmission spectrum for bottom layer thicknesses of 3 nm, 5 nm (this design), and 10 nm. As shown in Fig. 7, the structural configurations with 3 nm and 10 nm of bottom layer thickness exhibit prominent localized transmittance peaks within the shorter wavelength region (λ < 250 nm) and in (1400–1800 nm) region. Conversely, implementing an optimal thickness of 5 nm significantly suppresses the peak transmittance magnitude to a minimal value at short wavelengths, ensuring superior transmission blocking performance across the entire 150–3000 nm range. Therefore, 5 nm is chosen as the optimal bottom layer thickness.

Fig. 7. Simulated transmission spectrum for different bottom Ni layer thicknesses.

3.4. Cross and co-polarization with PCR analysis

3.4.1. PCR analysis

Cross- and co-polarization analysis is a crucial step in the design process of an MMA, ensuring efficient absorption of incident electromagnetic waves. This analysis verifies that the proposed absorber functions solely as an energy harvester and does not behave as a polarization converter. To assess this, reflection for both TE and TM modes can be expressed using eqn (8) and (9),

|S₁₁(ω)|² = |S_TE,TE(ω)|² + |S_TE,TM(ω)|² = |R_yy|² + |R_yx|² 8

|S₁₁(ω)|² = |S_TM,TM(ω)|² + |S_TM,TE(ω)|² = |R_xx|² + |R_xy|² 9

here, the reflection coefficients for co-polarized waves are |R_yy|² = ∣S_TE,TE(ω)∣² and |R_xx|² = |S_TM,TM(ω)|², while the cross-polarized components are represented by |R_yx|² = |S_TE,TM(ω)|², |R_xy|² = |S_TM,TE(ω)|². To quantify the amount of polarization conversion, the polarization conversion ratio (PCR) is calculated using the eqn (10) and (11) for both TE and TM modes

10

11

To quantitatively interpret polarization conversion, the following thresholds are defined: PCR < 0.1 is considered low polarization conversion and PCR < 0.02 is considered negligible (indicating that the absorber does not function as a polarization converter). These thresholds are based on standard practices in metamaterial absorber literature.³³

As illustrated in Fig. 8, the PCR values for both TE and TM modes remain below 0.02 (negligible) across >90% of the operating wavelength range, confirming negligible polarization conversion. In TE mode, the PCR magnitude is close to zero throughout the spectrum, with only minor peaks observed around 700–800 nm and 1600–1700 nm, where the values reach approximately 0.03–0.05, which is still well below the 0.1 low-conversion threshold. The TM mode exhibits slightly higher, but still low, PCR levels, with a maximum peak of approximately 0.25 near 700 nm, followed by a secondary, weak peak near 1650 nm (∼0.15). Outside these narrow resonance regions, the PCR values approach nearly zero. The consistently low PCR response (maximum <0.3) for both polarizations demonstrates that the proposed MMA structure does not introduce significant polarization rotation or conversion and therefore functions purely as an efficient absorber with minimal polarization distortion.

Fig. 8. Polarization conversion ratio (PCR) for TE and TM modes.

3.4.2. Novelty in polarization suppression

A distinguishing feature of the proposed web lattice is its low polarization conversion ratio (PCR < 0.02 across 90% of the bandwidth, as shown in Fig. 8). Most complex resonators unintentionally rotate polarization, reducing effective absorption. Here, the symmetric web lattice with threefold radial symmetry cancels cross-polarized components, ensuring that nearly all incident energy is absorbed rather than converted. This unique property directly contributes to the 99.4% visible-region average absorption.

3.5. Polarization and angular sensitivity

Polarization-insensitive MMAs are highly desirable due to their consistent absorption performance across different polarization angles. Fig. 9 demonstrates the effect of polarization rotation from 0° to 180° under TE and TM modes. As shown in Fig. 9(a), the proposed absorber maintains high polarization stability in the TE mode, where the absorptance remains ≥95% across the 250–700 nm range for all polarization angles. A minor reduction is observed in the NIR region, where absorption gradually decreases from near-unity to approximately 85–90% at 1400 nm, depending on the polarization angle. Similarly, in Fig. 9(b), the TM mode exhibits consistent broadband absorption with peak values close to unity around 300–600 nm and maintains values above ∼0.90 up to 1000 nm. At longer wavelengths (1000–1400 nm), the absorption slightly declines to approximately 84–90%. The minimal variation between TE and TM responses confirms that the symmetric geometry of the resonator enables nearly identical electromagnetic coupling for different polarization orientations. Consequently, the absorber demonstrates robust polarization-independent broadband performance.

Fig. 9. Representation of the effect on change of polarization angle from 0° to 180° at (a) TE mode and (b) TM mode.

Furthermore, Fig. 10 presents the absorption response of the proposed MMA for varying incident angles from 0° to 70° under TE and TM polarizations. As shown in Fig. 10(a), the TE mode demonstrates strong angular stability, maintaining absorptance ≥95% within the 250–700 nm region for incident angles up to 60°. A gradual decrease is observed at higher wavelengths, where absorption reduces from near unity to approximately 84–88% at 1400 nm. At extreme oblique incidence of 70°, the absorptance decreases more noticeably, reaching 83–85% across the NIR region. In Fig. 10(b), sustaining absorption ≥95% up to 800 nm for angles ≤50° in TM mode. However, at 70°, a significant reduction is observed in the UV-visible region, with absorption dropping to 65–75% near 200 nm and 72–80% at longer wavelengths. Despite this degradation at extreme angles, the absorber maintains broadband high absorption ≥85% for incident angles ≥60° in TM mode. These results confirm that the proposed structure exhibits strong angular stability ≥60° in both TE and TM modes.

Fig. 10. Representation of the effect on absorption for change of incident angle from 10° to 70° at (a) TE mode and (b) TM mode.

3.6. Absorption performance under AM1.5 solar spectrum

To evaluate the terrestrial solar energy-harvesting capability, the proposed Ni–AlN–Ni MMA was analyzed under the standard Air Mass 1.5 (AM1.5) solar spectrum,³⁴ which represents realistic terrestrial sunlight. It is critical to note that the AM1.5 spectrum contains negligible irradiance below ∼280 nm due to atmospheric ozone absorption. Although the proposed absorber operates efficiently down to 150 nm (relevant for space/VUV applications), only the 280–2500 nm portion contributes to terrestrial solar harvesting.

The weighted absorption (A_w) under AM1.5 was computed using eqn (12),

12

where A(λ) is the spectral absorptance (from simulation, sampled at 1 nm intervals), I_AM1.5(λ) is the solar irradiance (W m^–2 nm^–1) obtained from the NREL AM1.5 reference spectrum,³⁴ and the integration limits λ_min (nm) and λ_max (nm) is the lower wavelength and upper wavelength values in the absorber's bandwidth under AM1.5 considered for computation.

According to Table 2, the computed weighted absorption reaches 96.5% over the terrestrial-relevant 280–2500 nm range and exceeds 98.656% within the visible region (380–780 nm). Fig. 11 shows strong spectral alignment between the absorber's response and the AM1.5 irradiance profile. In addition to polarization insensitivity and angular stability up to 60°, the proposed absorber shows potential for high-performance terrestrial solar energy harvesting. For space-based or VUV applications where the 150–280 nm range is relevant, the absorber's full-bandwidth average absorption of 98.1% (150–1400 nm) remains advantageous.

Table 2. AM1.5 weighted absorption calculation details.

Wavelength range (nm)	∫IAM1.5(λ)dλ (W m−2)	∫A(λ)IAM1.5(λ)dλ (W m−2)	Weighted absorption (%)
280–1400	793.1	776.4863	97.908
380–780 (visible only)	325.98	330.42	98.656
280–2500 (full AM1.5)	892.74	948.6a	96.5a

^aNote: The 280–2500 nm (full AM1.5) integration includes the 1400–2500 nm SWIR region where absorption drops to ∼84–87%, reducing the overall weighted absorption.

Fig. 11. Solar irradiance spectrum (AM1.5, ASTM G173-03) and spectral energy absorbed by the proposed MMA over the terrestrial-relevant range of 280–2500 nm. The absorber's response below 280 nm (not shown here due to negligible AM1.5 irradiance) is relevant for space/VUV applications.

3.7. Geometric parameter sweep

A systematic parametric sweep was performed to investigate the influence of each geometric parameter on absorption performance. During each parameter sweep, the scanned parameter is varied while all other parameters are kept fixed at their nominal values. Table 3 summarizes all parametric sweep simulations, listing the scanned parameter, scan range, step size, and the fixed parameters for each case.

Table 3. Summary of parametric sweep simulations^a.

Sweep	Scanned parameter	Scan range (nm)	Step size (nm)	Fixed parameters (nm)
Fig. 12(a)	Unit cell period (P)	200, 210, 220, 230	10	P 1 = 150, P2 = 175, P3 = 200, W = 10, tw = 5, ts = 12, tn = 5
Fig. 12(b)	Outer hexagon diameter (P3)	125, 150, 175, 200	25	P = 200, P1 = 150, P2 = 175, W = 10, tw = 5, ts = 12, tn = 5
Fig. 13	Resonator arm width (W)	6, 8, 10, 12	2	P = 200, P1 = 150, P2 = 175, P3 = 200, tw = 5, ts = 12, tn = 5
Fig. 14(a)	Resonator thickness (tw)	3, 5, 7, 9	2	P = 200, P1 = 150, P2 = 175, P3 = 200, W = 10, ts = 12, tn = 5
Fig. 14(b)	Substrate thickness (ts)	8, 10, 12, 14	2	P = 200, P1 = 150, P2 = 175, P3 = 200, W = 10, tw = 5, tn = 5
Fig. 14(c)	Bottom layer thickness (tn)	3, 5, 7	2	P = 200, P1 = 150, P2 = 175, P3 = 200, W = 10, tw = 5, ts = 12

^aNote: For each sweep, the scanned parameter is varied while all other parameters are held at their nominal values (Table 1).

Fig. 12(a) illustrates the effect of square unit-cell length P (200–230 nm) on the absorption performance of the proposed MMA. The absorber maintains broadband high absorption >90% across most of the visible-NIR region for all tested value of P. Increasing the period to 210 and 220 nm introduces a slight degradation in the short-wavelength region, where absorption decreases to approximately 75–80% near 170–200 nm. For the largest P = 230 nm, a more pronounced reduction is observed, with absorption dropping to 60% in the UV region. Overall, the results confirm that P = 200 nm provides the most optimized broadband response, while larger unit-cell sizes reduces absorption efficiency particularly in the UV regime.

Fig. 12. Graphical representation of parametric sweep (a) changes of dimension (P) (b) changes of (P₃).

Fig. 12(b) represents a graph plotting absorbance against wavelength ranging from 100 to 1400 nm for various P₃ values (125 nm, 150 nm, 175 nm, and 200 nm). Absorbance remains high (approximately 0.95 to 1) throughout the visible and near-infrared spectrum. Overall, the design achieves a broadband absorption effect, although the curve experiences slight shifts as the P₃ value increases. This suggests that resonance coupling can be adjusted effectively without significantly impacting performance, even when P₃ varies. The absorber attains nearly complete absorption over a wide spectral range from UV to NIR. Changes in P₃ values indicate that absorption at specific wavelengths can be finely tuned through geometric modifications.

The width (W) of the hexagonal resonator arms in the designed MMA is crucial in determining the optical response over the wavelength range under investigation. The absorbance spectrum in Fig. 13 shows that there are slight but discernible changes in the absorbance profile as W increases from 6 nm to 12 nm. Narrower widths (W = 6 nm and 8 nm) exhibit a deeper modulation with steeper dips at shorter wavelengths (≤300 nm), whereas broader structures (W = 10 nm and 12 nm) smooth out these oscillations, indicating improved resonance coupling and decreased scattering losses.

Fig. 13. Graphical representation of parametric sweep change of width (W) of top layer resonator.

To investigate the structural sensitivity of the designed MMA, we systematically varied the thicknesses of three key layers: the top resonator layer (t_w), the substrate layer (t_s), and the bottom metallic layer (t_n), while monitoring the absorbance across the 150–1400 nm wavelength range. As depicted in Fig. 14(a), increasing the resonator thickness (t_w = 3–9 nm) altered both the peak position and absorption profile. A thinner resonator resulted in narrower absorption, while thicker resonators slightly reduced peak absorbance but extended broadband performance. From observation, it is evident that the proposed resonator thickness of 5 nm results in optimized absorption characteristics throughout the observed regions. Fig. 14(b) shows the substrate thickness (t_s = 8–14 nm) had a significant influence on the absorbance, particularly in the longer wavelength region. For t_s = 8 nm and 14 nm, absorption response degrades in the lower UV region, in contrast to the visible and NIR region. For a 10 nm response, the value is near unity in the lower region, while it degrades as the wavelength increases. The proposed dimension of 12 nm maintains stable absorption characteristics across a wavelength range of 150 nm to 1400 nm. As depicted in Fig. 14(c), variations in the bottom-layer thickness (t_n = 3–7 nm) have a minimal impact on the absorbance spectrum, with all curves overlapping closely and maintaining peak absorbance values near unity. This estimate suggests that a 3 nm thickness in the ground layer is sufficient to function as a reflective layer, effectively reducing transmission, but the bottom Ni layer thickness of 5 nm was selected based on a transmission analysis (Fig. 7). A 3 nm layer shows residual transmission peaks in the UV and NIR region, while 5 nm reduces transmission across the full 150–3000 nm range. Moreover, the 5 nm thickness also provides robustness against fabrication-induced non-uniformities.

Fig. 14. Demonstration of varying the thickness of resonator, substrate and bottom layer respectively in (a)–(c).

The results of the top resonator layer (t_w), the substrate layer (t_s), and the bottom metallic layer (t_n) highlight that while the bottom layer thickness has a negligible influence, both the substrate and resonator thickness are critical tuning parameters for optimizing the absorption characteristics of the proposed MMA.

3.8. Tolerance analysis: robustness against fabrication errors

Nanofabrication processes (electron-beam lithography (EBL), sputtering, and etching) introduce dimensional variations. Though the acceptable fabrication error depends entirely on the application's requirements, in high-precision nanofabrication, critical dimension (CD) errors should generally be kept within ±2. To assess the robustness of the proposed design, a tolerance analysis was performed by perturbing key geometric parameters by ±2 nm relative to their nominal values (Table 4). These tolerances are typical for EBL-defined metal–insulator–metal structures.

Table 4. Summary of tolerance analysis. Absorption values are reported as averages over the visible range (380–780 nm) and full bandwidth (150–1400 nm).

Parameter	Nominal (nm)	Variation	Visible avg. abs. (%)	Full-band avg. abs. (%)	Δ from nominal (full-band)
None (nominal)	—	—	99.4	98.1	—
Resonator thickness (tw)	5	+2 nm (7 nm)	98.73	97.13	−0.97%
		−2 nm (3 nm)	99.2	97.3	−0.8%
Substrate thickness (ts)	12	+2 nm (14 nm)	99.3	96.5	−1.6%
		−2 nm (10 nm)	98.3	97.76	−0.34%
Resonator width (W)	10	+2 nm (12 nm)	98.7	97.2	−0.9%
		−2 nm (8 nm)	99.3	97.28	−0.82%
Unit cell period (P)	200	+2 nm (202 nm)	99	96.7	−1.4%
		−2 nm (198 nm)	98.7	97.4	−0.7%
Bottom layer thickness (tn)	5	+2 nm (7 nm)	99.2	97.35	−0.75%
		−2 nm (3 nm)	99.2	97.35	−0.75%

Considering Fig. 13–15, the proposed absorber maintains visible-range average absorption above 98% and full-band average absorption above 96% for all ±2 nm variations. The most sensitive parameter is the substrate thickness (t_s), where a +2 nm (14 nm) deviation reduces full-band absorption to 96.5% (reduced by 1.6% but still >95%). Overall, the design demonstrates acceptable tolerance for EBL-based fabrication. Therefore, the detailed tolerance analysis demonstrates that the design is robust against inevitable nanofabrication errors.

Fig. 15. Absorbance due to fabrication error of unit cell period (P) at critical dimension errors of ±2 nm when the optimum dimension is 200 nm.

3.9. Absorption with different dielectric and resonator materials

Fig. 16 illustrates the influence of dielectric substrates and resonator metals on the absorption performance of the proposed MMA while maintaining identical structural geometry. As shown in Fig. 16(a), substrates such as alumina and gallium arsenide achieve superior broadband absorption, maintaining absorptance ≥95% throughout most of the visible-NIR spectrum (350–1200 nm), with a moderate reduction to 75–80% in the UV region due to impedance mismatch at shorter wavelengths. AlN demonstrates the most balanced broadband response, with near-unity absorption (99–100%) in the visible range and a gradual decrease to 90% at 1400 nm. In contrast, polyimide and polycarbonate exhibit reduced NIR performance, with absorptance decreasing to 86–88% beyond 1000 nm owing to weaker electromagnetic confinement.

Fig. 16. Characteristic of absorption for the change of (a) dielectric and (b) resonator material.

Fig. 16(b) compares different resonator metals, where nickel achieves the highest efficiency, maintaining absorptance close to unity across the visible spectrum and remaining ≥90% up to 1400 nm due to its higher intrinsic loss and enhanced energy dissipation. Conversely, metals such as Au, Cu, Al, and W show limited absorption with peak values ≤65% and significant reduction in UV response (22–30% at 150 nm), attributed to reduced ohmic loss and weaker field localization.

3.10. Electric field, magnetic field, and surface current distribution analysis

To clarify the absorption mechanism of the proposed MMA, the electric and magnetic field distributions have been analyzed at key resonant wavelengths under both TE and TM polarizations. According to Maxwell's equations, the electric and magnetic field intensities in high frequency regions are governed by time-varying electric and magnetic flux densities, as expressed in eqn (13) and (14).

13

14

where, E, H, B_m, and D_e represent electric field density, magnetic field density, magnetic flux density, and electric flux density, respectively. These are related to effective material properties through D_e = ε_effectiveε₀E and B_m = µ_effectiveµ₀H.

To analyze the absorption mechanism of the proposed absorber, the electric field, magnetic field, and surface current distributions have been analyzed at 310 nm, 570 nm, and 850 nm under TM and TE modes, as shown in Fig. 17–19.

Fig. 17. Electric field distribution for TM mode at (a) 310 nm (b) 570 nm (c) 850 nm and TE mode at (d) 310 nm (e) 570 nm (f) 850 nm wavelength.

Fig. 18. Magnetic field distribution for TM mode at (a) 310 nm (b) 570 nm (c) 850 nm and TE mode at (d) 310 nm (e) 570 nm (f) 850 nm wavelength.

As shown in Fig. 17(a), at 310 nm, the electric field is strongly concentrated at the resonator edges, indicating intense plasmonic resonances that enhance light confinement in the UV region. As depicted in Fig. 18(a), the magnetic field distribution reveals strong coupling between the metallic layers, confirming the excitation of a magnetic dipole resonance. Correspondingly, Fig. 19(a) shows that the surface current distribution exhibits anti-parallel currents in the top and bottom metallic layers.

As shown in Fig. 17(b), at 570 nm, the electric field is strongly localized inside the dielectric spacer and around the resonator boundaries. This indicates the presence of hybridized resonance modes that facilitate efficient absorption in the visible region. As depicted in Fig. 18(b), the magnetic field is symmetrically confined, while the surface current distribution that is shown in Fig. 19(b) exhibits circulating currents along the edges of the resonator.

As shown in Fig. 17(c) and 18(c), at 850 nm, the electric and magnetic fields become more uniformly spread across the structure respectively, with reduced localization compared to the UV and visible regions. This shift indicates that propagating surface plasmon resonances and cavity-like modes dominate the absorption mechanism at longer wavelengths. Fig. 19(c) shows, the surface current at this wavelength forms extended loops along the resonator, enabling sustained absorption even in the NIR region.

Overall, the combined field and current analysis demonstrates that the absorber utilizes localized plasmonic resonances at shorter wavelengths and transitions to propagating and cavity-enhanced modes at longer wavelengths. This hybrid mechanism ensures broadband absorption across the UV-NIR spectrum for both TM and TE polarizations.

4. Comparative analysis with prior work

The rigorous comparison in Table 5, based on uniformly defined metrics, clearly demonstrates the superior and genuinely novel performance of the proposed Ni–AlN–Ni hexagonal web lattice MMA. Unlike prior works that either operate within narrow visible bands (e.g., ref. 16, 19 and 20), lack UV coverage (ref. 17, 18, 22 and 26), or sacrifice visible performance for infrared extension (ref. 21 and 25), the proposed design achieves a wide operational bandwidth of 150–1400 nm (VUV partially to NIR partially) with an overall average absorption of 98.1%. Most notably, in the technologically critical visible range (380–780 nm), it attains an average absorption of 99.4% that is significantly higher than the next-best values (∼94–95% in ref. 19 and 23). Furthermore, under the strict angular stability criterion (≥80% absorption retention), the proposed MMA maintains stable performance up to 60°, comparable to or exceeding most cited works (only ref. 23 reports 80° but with a narrower bandwidth and tungsten-based fabrication challenges). Under fair comparison (same visible range definition and overall averages), the proposed design shows the highest visible-region absorption (99.4%) and the widest continuous 90% bandwidth (150–1400 nm) among all Ni-based absorbers. These combined advantages of broadband coverage, near-unity visible absorption, and robust angular stability establish the clear advancement and practical significance of the proposed absorber over existing designs, particularly for solar energy harvesting applications.

Table 5. Comparative analysis of MMA performances^a.

Ref.	Year	Material	Type of work (simulation/fabrication)	Reported spectral coverage (nm) (A ≥ 90%)†	Range covered	Visible-range avg. absorption (380–780 nm or reported range)!	Overall avg. absorption (reported range)!	Angular stability (θ)	Distinctive geometric feature
16	2024	Ni–SiO2–Ni	Sim.§	415–625	Visible (partially)	92.1%*	95.8%	60°	Simple single-ring
17	2024	Ni–AlN–Ni	Sim.§	400–4000	Visible (partially), NIR, SWIR, MWIR (partially)	89.7%*	93.0%	60°	Cross-shaped resonator
18	2024	Ni–AlN–Ni	Sim.§	400–3000	Visible (partially), NIR, SWIR	88.4%*	90.0%	60°	Cross-shaped resonator
19	2024	Ni–SiO2–Ni	Sim.§	390–780	Visible (partially)	94.2%*	97.2%	60°	Plus-shaped resonator
20	2024	Au–Ni–quartz–Au	Sim.§	361.8–505	Visible (partially)	91.3%*	93.9%	60°	Concentric octagonal elliptical ring resonator
21	2025	Ni–Al2O3–Pt	Sim.§	800–3000	NIR, SWIR, MWIR	No visible coverage	91.6%	50°	Square and circular rings resonant
22	2025	Ni–SiO2–Ni	Sim.§	400–4000	Visible (partially), NIR, SWIR, MWIR (partially)	90.8%*	95.3%	60°	Circular ring resonator
23	2025	W–SiO2–W	Sim.§	360–1200	Visible, NIR (partially)	94.1%*	96.3%	80°	Square ring
24	2025	W–quartz–W	Sim.§	342.00–832	UVA (partially), visible, NIR (partially)	87.3%*	90.1%	60°	Modified rectangular ring with stubs
25	2021	Ni–SiO2–Ni	Sim.§	820–2520	NIR, SWIR	No visible coverage	97.0%	60°	Hexagonal ring shaped fractal
26	2022	Ni–SiO2–Ni	Sim.§	400–2800	Visible (partially), NIR, SWIR	89.6%*	93.0%	60°	Circular ring
This work	—	Ni–AlN–Ni	Sim.§	150–1400	VUV (partially), UVA, UVB, visible, NIR (partially)	99.4%	98.1%	≥60°	Hexagonal web lattice with radial connectivity

^aNote: ! averages are computed over the reported operational range at 1 nm intervals. * indicates values recomputed from published spectra. Angular stability = max θ where absorption ≥80% of normal-incidence value. † For terrestrial solar harvesting, only the 280–1400 nm range is relevant due to atmospheric absorption; the full 150–1400 nm range applies to space/VUV applications. Sim.§ = simulation-only data available from the original reference.

To contextualize the simulated performance of the proposed design against experimentally demonstrated devices, the results are compared with two experimentally validated Ni-based metamaterial absorbers reported in the literature in Table 6. To our knowledge, no experimentally validated Ni–AlN structure operating across the full 150–1400 nm range is currently available. However, the following experimental studies serve as valuable benchmarks for Ni-based absorbers in overlapping wavelength ranges. Luo et al.³⁵ demonstrated a Ni-based absorber with average absorption exceeding 90% across the visible range (400–700 nm) using soft nano-imprinting lithography, with near-unity absorption (>99%) over 500–560 nm. Asad et al.³⁶ reported a low-cost Ni/SiO₂/Ni transparent solar absorber with stable performance under solar illumination, achieving approximately 90% absorption in the visible-NIR region. Both experimental studies confirm that Ni-based MIM configurations can achieve high absorption in the visible spectrum, consistent with the simulation results (99.4% visible-range average) of the proposed work.

Table 6. Comparison of simulated (this work) vs. experimentally measured performance of Ni-based metamaterial absorbers.

Ref.	Material system	Type	Wavelength range	Average absorption (reported range)
35	Ni–dielectric cylinder–Ni	Experimental	400–700 nm	>90%
36	Ni/SiO2/Ni multilayer	Experimental	400–1100 nm	∼90%
This work	Ni–AlN–Ni	Simulation	150–1400 nm	98.1%

On the other hand, the proposed design extends the operational bandwidth significantly further into the UV (down to 150 nm) and NIR (up to 1400 nm) regions, achieving an overall average absorption of 98.1% (considering absorption >90%) that is higher than the experimentally demonstrated values in the visible range alone. This extended bandwidth and enhanced performance are attributed to the radially connected hexagonal web lattice resonator, which enables hybridized plasmonic modes and suppresses polarization conversion. Though the proposed absorber has not been fabricated, the tolerance analysis (Section 3.8 Tolerance analysis: robustness against fabrication errors) confirms that the structure is robust against predictable nanofabrication errors.

5. Relevance to chemical applications

While the preceding sections have focused on the electromagnetic design and absorption performance of the proposed Ni–AlN–Ni metamaterial absorber, this section discusses its relevance to chemistry-oriented domains. The material choices, nickel (Ni) and aluminum nitride (AlN), possess intrinsic chemical and physical properties that extend the utility of this absorber beyond traditional photonic applications into photothermal catalysis, photocatalytic processes, and chemically robust energy harvesting.

5.1. Chemical properties of nickel and aluminum nitride

Nickel exhibits high corrosion resistance against oxidation and chemical attack in alkaline and neutral environments, along with a high melting point of approximately 1455 °C and thermal conductivity of 90.7 W m⁻¹ K⁻¹.²⁷ These properties ensure structural and chemical stability under intense solar illumination or high-temperature reaction conditions. Aluminum nitride, as a wide-bandgap semiconductor (E_g ≈ 6.2 eV), offers high chemical inertness, high thermal conductivity (up to 285 W m⁻¹ K⁻¹), and transparency in the UV-visible range.²⁹ Unlike conventional dielectric materials such as SiO₂, AlN does not undergo significant chemical degradation in humid or oxidative environments, making it suitable for long-term outdoor solar applications.

5.2. Photothermal catalysis

The near-unity broadband absorption (99.4% in the visible range, 98.1% overall from 150–1400 nm) enables efficient photothermal conversion, where incident solar energy is converted into localized heat within the Ni resonator layers. This photothermal effect can drive endothermic catalytic reactions without external resistive heating. For instance, the absorber could serve as a photothermal substrate for CO₂ hydrogenation (reverse water–gas shift or methanation) or ammonia synthesis, where localized hot spots at the Ni–AlN interface enhance reaction rates. The high thermal conductivity of both Ni and AlN ensures rapid heat distribution to surface-adsorbed catalytic species or co-deposited catalyst nanoparticles (e.g., Pt, Ru, or Ni nanoparticles themselves).

5.3. Photocatalytic water splitting and UV activation

The AlN dielectric spacer, with its wide bandgap (6.2 eV), absorbs deep-UV photons (<200 nm) while remaining transparent in the visible and near-infrared ranges. Under UV illumination, AlN can generate electron–hole pairs, potentially initiating photocatalytic redox reactions. Meanwhile, the Ni top layer supports surface plasmon resonances that produce hot electrons, which can be injected into the AlN or directly into adsorbed water molecules. This dual mechanism, direct photocatalysis from AlN and plasmon-driven photocatalysis from Ni, could enhance the efficiency of the hydrogen evolution reaction (HER) or the oxygen evolution reaction (OER) in water-splitting configurations. Although the present work focuses on absorption characterization, the material platform is inherently suited for integration with photoelectrochemical cells.

5.4. Chemically robust solar absorbers for harsh environments

Many solar-harvesting applications involve chemically aggressive conditions, including seawater desalination (high salinity, chlorine exposure), industrial waste-heat recovery (acidic or basic exhaust gases), and outdoor concentrated solar power (CSP) plants (oxidation at elevated temperatures). The Ni–AlN–Ni architecture offers superior resistance to oxidation and corrosion compared to noble-metal-based absorbers (Au, Ag) that suffer from dewetting or sulfide formation, and compared to oxide-based absorbers (e.g., ITO) that degrade under prolonged UV exposure. The absence of easily oxidized materials ensures long-term chemical stability, a critical factor for commercial deployment.

5.5. Fabrication and integration with chemical systems

The proposed absorber can be fabricated using DC magnetron sputtering or electron beam evaporation for Ni layers, and reactive sputtering or atomic layer deposition (ALD) for AlN. These techniques are compatible with the deposition of additional catalyst layers (e.g., Pt, Co, or Ni nanoparticles) or protective coatings. The hexagonal web lattice pattern is defined via electron beam lithography, which allows precise control over surface morphology, an important parameter for catalytic activity (e.g., surface roughness, edge site density). Future work will explore the integration of this absorber with microfluidic reactors or photoelectrochemical cells for solar-driven chemical synthesis.

6. Conclusions

In this study, a Ni–AlN–Ni metamaterial absorber with a radially connected hexagonal web lattice resonator was designed and characterized using finite-element electromagnetic simulations. The simulated absorption exceeds 95% across 150–1400 nm, with a full-bandwidth average of 98.1% and a visible-range average of 99.4% (380–780 nm). The absorber exhibits polarization insensitivity and maintains absorption above 90% for incident angles up to 60° under both TE and TM polarizations. The weighted absorption efficiency under the terrestrial AM1.5 solar spectrum (280–2500 nm) is 96.5%. The findings underscore the absorber's significant promise for energy harvesting, thermal imaging, photodetection, and optoelectronic applications, especially in contexts requiring broadband, high-efficiency absorption. Beyond its optical performance, the Ni–AlN–Ni material platform offers significant chemical advantages, including corrosion resistance, thermal stability, and a wide-bandgap semiconductor interface. These properties make the proposed absorber relevant for photothermal catalysis, photocatalytic water splitting, and chemically robust solar thermal harvesting.

On the other hand, the main limitation of the present study is its purely computational nature. Experimental validation through fabrication and optical characterization (UV-vis-NIR spectrophotometry) has not yet been performed. Also, thermal and chemical stability under operating conditions was not assessed. Therefore, these limitations can be overcome by a future directions that will focus on: (i) experimental fabrication of the proposed hexagonal web lattice using electron beam lithography and sputter deposition, (ii) optical characterization to validate the simulated absorption spectra, (iii) integration with catalytic nanoparticles (e.g., Pt, Ru) to evaluate photothermal CO₂ hydrogenation, (iv) testing under concentrated solar illumination for solar thermal applications, and (v) investigation of long-term chemical stability (oxidation, corrosion) under relevant operating conditions. The design is also planned to be extended to flexible substrates for wearable photonic devices.

Author contributions

Rahul Chowdhury: conceptualization, data curation, formal analysis, investigation, methodology, software, validation, visualization, writing – original draft, writing – review & editing; Muhammad Asad Rahman: formal analysis, methodology, project administration, resources, supervision, writing – original draft, writing – review & editing; Abu Hanif: formal analysis, resources, software, writing – review & editing; Idban Alamzadeh: supervision, software, writing – review & editing; Mohammad Adham Inan: visualization, writing – review & editing; Asif Mohammed Siddiqee: formal analysis, software, writing – original draft, writing – review & editing.

Conflicts of interest

No conflict of interest.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.