Metal-Integrated Monolithic High-Contrast Grating

• MetalMHCG is a grating platform that monolithically integrates metal to enable co-designed optical, electrical, and absorptive functionalities.
• It leverages guided-mode resonance and effective Fabry–Pérot behavior to achieve near-unity transmission, low sheet resistance, and high X-ray absorption.
• Fabrication techniques such as Au–Sn microcasting and e-beam lithography yield scalable devices for infrared electrodes, visible couplers, and non-Hermitian photonic applications.

Searching arXiv for the cited metalMHCG papers and closely related work to ground the article in published sources. I’m checking arXiv for the specific papers on metal-integrated monolithic high-contrast gratings to ensure the article is grounded in the cited literature. Metal-integrated monolithic high-contrast grating (metalMHCG) denotes a class of subwavelength grating structures in which a high-contrast grating is monolithically integrated with metal so that optical functionality and electrical or absorptive functionality are co-designed within a single patterned platform. In the cited literature, the term spans several realizations: Au–Sn-filled silicon absorption gratings for X-ray phase-contrast interferometry (Romano et al., 2017), GaAs high-contrast gratings with embedded or groove-filled gold for infrared transparent conductive electrodes (Ekielski et al., 2023, Bogdanowicz et al., 30 Jul 2025), dielectric waveguides equipped with a metal grating that support Friedrich-Wintgen bound states in the continuum and exceptional points (Kikkawa et al., 2020), and embedded-metal visible grating couplers in Si$_3$N$_4$ photonics (Smith et al., 2021). Across these realizations, the common principle is that metal is integrated into a monolithic grating geometry so that the electromagnetic field is redistributed relative to the metal, rather than simply passing through a continuous conductive or absorptive film.

1. Definition and scope

Within photonics and X-ray optics, a high-contrast grating (HCG) is a subwavelength, single-layer dielectric diffraction grating in which the refractive-index contrast between the grating material and its surroundings supports leaky-wave resonances (Ekielski et al., 2023). The metalMHCG extends this concept by integrating metal directly into the grating architecture. In the infrared transparent-electrode implementation, metal stripes are embedded into a GaAs HCG so that the resonant field is funneled through the dielectric stripes while largely avoiding the metal, thereby suppressing free-carrier absorption and Fresnel reflections (Ekielski et al., 2023). In the X-ray implementation, the metal is not a perturbative conductive inclusion but the high-density absorbing medium itself, cast directly into deep silicon templates to form monolithic metal high-contrast gratings with very high aspect ratios (Romano et al., 2017).

The term therefore encompasses more than one device family. In one family, the structure is designed to maximize optical transmission while retaining very low sheet resistance, as in infrared transparent conductive electrodes on GaAs (Ekielski et al., 2023, Bogdanowicz et al., 30 Jul 2025). In another, the integrated metal grating modifies waveguide-mode coupling and non-Hermitian spectral structure, enabling bound states in the continuum (BICs) and exceptional points (EPs) (Kikkawa et al., 2020). In a further visible-photonics variant, a buried Au grating beneath a Si$_3$N$_4$ core increases the attainable refractive-index contrast and reduces grating-coupler footprint (Smith et al., 2021). This suggests that metalMHCG is best understood as a platform concept rather than a single fixed geometry.

2. Physical principles

The underlying optical principle in infrared metalMHCGs is guided-mode resonance in the deep-subwavelength regime, where only the zeroth diffraction order propagates and the resonance condition is approximately

$m\lambda \simeq 2 n_{\rm eff}\Lambda ,$

with $m$ the mode order, $\lambda$ the free-space wavelength, $\Lambda$ the grating period, and $n_{\rm eff}$ the effective index of the resonant mode in the high-index stripes (Ekielski et al., 2023). In this regime, the grating can behave as an antireflective transmissive structure rather than as a conventional reflective metal pattern. For a flat air–GaAs interface, Fresnel reflection is

$R = \left(\frac{n-1}{n+1}\right)^2 \simeq 30\%$

for $_4$0, which sets a strong reflection penalty for unstructured GaAs (Ekielski et al., 2023). The metalMHCG is designed to suppress that penalty.

The 2025 GaAs implementation formulates this behavior using Rigorous Coupled-Wave Analysis (RCWA) and an effective-medium Fabry–Pérot model (Bogdanowicz et al., 30 Jul 2025). In the deep-subwavelength limit, the grating behaves as a homogeneous slab of thickness $_4$1 with polarization-dependent effective index $_4$2, forming a Fabry–Pérot etalon between substrate and air interfaces. Fitting the height-periodicity of numerically computed maxima at $_4$3 gives $_4$4 and $_4$5, both between air and GaAs, so the MHCG acts as a low-index antireflection layer for both polarizations (Bogdanowicz et al., 30 Jul 2025). The quarter-wave intuition,

$_4$6

is not exactly met; instead, modal dispersion permits simultaneous near-zero reflection for TE and TM at a common height slightly differing from a true $_4$7 coating (Bogdanowicz et al., 30 Jul 2025).

In the waveguide-with-metal-grating configuration, the decisive physics is mode coupling rather than transparent-electrode transport. There, temporal-coupled-mode theory describes two resonators with internal coupling $_4$8, radiative loss $_4$9, and intrinsic loss $_3$0 (Kikkawa et al., 2020). A Friedrich-Wintgen BIC occurs when the two modes are degenerate and their radiation amplitudes cancel:

$_3$1

An EP occurs when the discriminant $_3$2 of the secular equation vanishes, with both $_3$3 and $_3$4 (Kikkawa et al., 2020). In that setting, varying metal-grating thickness tunes the anti-crossing gap, the sign of $_3$5, the BIC-bearing branch, and the EP location.

For X-ray gratings, the relevant physical principle is not guided-mode resonance but high-density absorption combined with capillary- and pressure-assisted metal filling of deep trenches. With $_3$6, the capillary pressure

$_3$7

assists pressure-driven flow of molten Au–Sn into $_3$8–$_3$9 deep trenches in the low-viscosity regime $_4$0 (Romano et al., 2017).

3. Materials, geometries, and fabrication

The material systems vary by application, but each implementation relies on monolithic integration of metal with a patterned high-contrast structure.

For X-ray phase-contrast interferometry, the selected alloy is the Au–Sn eutectic, $_4$1, with eutectic melting temperature $_4$2 and density $_4$3 (Romano et al., 2017). A thin metal wetting layer, $_4$4 of Ir or Au, is deposited before casting because native Si surfaces are hydrophobic to molten Au–Sn, whereas the conformal coating renders the surface hydrophilic so that capillary forces assist filling (Romano et al., 2017). Ir is deposited by atomic layer deposition on Bosch-etched or MACE templates; Au is deposited by seedless electroplating on high-resistivity $_4$5 Si templates (Romano et al., 2017). The hot-embossing tool heats at $_4$6 under vacuum $_4$7–$_4$8; a “touch” force of $_4$9 is first applied as $m\lambda \simeq 2 n_{\rm eff}\Lambda ,$0, then ramped to full embossing force in $m\lambda \simeq 2 n_{\rm eff}\Lambda ,$1 once $m\lambda \simeq 2 n_{\rm eff}\Lambda ,$2 (Romano et al., 2017).

For infrared transparent conductive electrodes, the 2023 device uses a $m\lambda \simeq 2 n_{\rm eff}\Lambda ,$3-thick, double-side-polished undoped GaAs substrate with target grating parameters $m\lambda \simeq 2 n_{\rm eff}\Lambda ,$4, duty cycle $m\lambda \simeq 2 n_{\rm eff}\Lambda ,$5, dielectric stripe width $m\lambda \simeq 2 n_{\rm eff}\Lambda ,$6, and metal groove width $m\lambda \simeq 2 n_{\rm eff}\Lambda ,$7 (Ekielski et al., 2023). The GaAs stripe height is $m\lambda \simeq 2 n_{\rm eff}\Lambda ,$8, while gold stripe thickness $m\lambda \simeq 2 n_{\rm eff}\Lambda ,$9 ranges from $m$0 to $m$1, with best performance at $m$2 (Ekielski et al., 2023). Fabrication uses a SiO$m$3 $m$4/Cr $m$5/SiO$m$6 $m$7 mask stack, e-beam lithography in AR-P 6200.9 resist, sequential plasma etching into the hard mask, ICP-RIE of GaAs in BCl$m$8/Ar to form a $m$9-deep grating with slightly concave side-walls, stripwise Au deposition by e-beam PVD, and mask lift-off in HF (Ekielski et al., 2023).

The 2025 large-area GaAs device uses a semi-infinite GaAs substrate with designed period $\lambda$0, measured $\lambda$1, fill factor $\lambda$2, measured $\lambda$3, grating depth $\lambda$4, measured $\lambda$5, and Au thickness $\lambda$6, measured $\lambda$7, deposited at the bottom of each groove (Bogdanowicz et al., 30 Jul 2025). Fabrication is carried out on a $\lambda$8 GaAs wafer with nine $\lambda$9 patches covering $\Lambda$0, using a SiO$\Lambda$1/Cr/SiO$\Lambda$2 hard mask, electron-beam lithography, ICP-RIE with BCl$\Lambda$3/N$\Lambda$4 in short etch-cool cycles, e-beam evaporation of Au selectively into the groove bottoms, and buffered HF hard-mask removal (Bogdanowicz et al., 30 Jul 2025).

For visible grating couplers, the stack consists of a SiO$\Lambda$5 substrate, a buried Au grating of thickness $\Lambda$6, a Si$\Lambda$7N$\Lambda$8 core of thickness $\Lambda$9 deposited by PECVD, and air top cladding (Smith et al., 2021). The metal-integrated grating has period $n_{\rm eff}$0, duty cycle $n_{\rm eff}$1, and lateral width $n_{\rm eff}$2, followed by a $n_{\rm eff}$3 linear taper into an $n_{\rm eff}$4-wide single-mode waveguide (Smith et al., 2021). Fabrication uses PMMA-patterned e-beam lithography, thermal evaporation of $n_{\rm eff}$5 Au without adhesion layer, lift-off in acetone with low-power O$n_{\rm eff}$6 plasma de-scum, low-temperature PECVD Si$n_{\rm eff}$7N$n_{\rm eff}$8, and CSAR-62-defined Si$n_{\rm eff}$9N$R = \left(\frac{n-1}{n+1}\right)^2 \simeq 30\%$0 waveguides formed by ICP etch (Smith et al., 2021).

4. Performance regimes

The performance of metalMHCGs is application-specific, and the literature reports distinct metrics for X-ray absorption gratings, infrared transparent electrodes, and visible grating couplers.

Platform	Representative reported metrics	Primary function
Au–Sn in Si for X-ray optics	pitch $R = \left(\frac{n-1}{n+1}\right)^2 \simeq 30\%$1–$R = \left(\frac{n-1}{n+1}\right)^2 \simeq 30\%$2, depth up to $R = \left(\frac{n-1}{n+1}\right)^2 \simeq 30\%$3, aspect ratio up to $R = \left(\frac{n-1}{n+1}\right)^2 \simeq 30\%$4, $R = \left(\frac{n-1}{n+1}\right)^2 \simeq 30\%$5 area	absorption grating
GaAs/Au IR TCE	TE absolute transmittance up to $R = \left(\frac{n-1}{n+1}\right)^2 \simeq 30\%$6 at $R = \left(\frac{n-1}{n+1}\right)^2 \simeq 30\%$7, unpolarized absolute transmittance up to $R = \left(\frac{n-1}{n+1}\right)^2 \simeq 30\%$8 at $R = \left(\frac{n-1}{n+1}\right)^2 \simeq 30\%$9, $_4$00–$_4$01	transparent conductive electrode
Large-area GaAs/Au IR TCE	unpolarized transmission $_4$02 at $_4$03, $_4$04 relative to Fresnel, $_4$05, $_4$06 coverage	large-area transparent conductive electrode
Si$_4$07N$_4$08/Au visible coupler	simulated $_4$09, measured fibre-to-fibre $_4$10 at $_4$11, $_4$12 bandwidth $_4$13	compact grating coupler

In the 2023 infrared electrode, TE-polarized light reaches absolute transmittance up to $_4$14 at $_4$15, corresponding to relative transmittance $_4$16 versus bare GaAs, with a spectral bandwidth above $_4$17 of $_4$18 (Ekielski et al., 2023). Unpolarized light reaches absolute transmittance up to $_4$19 at $_4$20, $_4$21, and $_4$22 above $_4$23 (Ekielski et al., 2023). Measured sheet resistance is $_4$24 for $_4$25 and $_4$26 for $_4$27; Joule heating becomes noticeable above $_4$28 (Ekielski et al., 2023).

The 2025 large-area implementation shifts the optimization toward unpolarized mid- to far-infrared performance. It reports $_4$29 at $_4$30, relative transmission $_4$31, and a bandwidth above the Fresnel limit of $_4$32, corresponding to $_4$33 relative bandwidth (Bogdanowicz et al., 30 Jul 2025). Its measured $_4$34 is higher than the ideal bulk-Au estimate $_4$35, with the difference attributed to wire size effects and grain-boundary or impurity scattering (Bogdanowicz et al., 30 Jul 2025).

In the X-ray embodiment, the emphasis is geometrical completeness and absorptive functionality rather than spectral transmission. Pitches vary from $_4$36 to $_4$37, depths reach $_4$38, and aspect ratios reach $_4$39 (Romano et al., 2017). SEM cross-sections show bubble-free, void-free trenches over a full $_4$40 area, with excess alloy flowing sideways and leaving a uniform top surface with $_4$41 residual film (Romano et al., 2017). High-density Au–Sn gratings achieve strong absorption at $_4$42 and have been preliminarily shown to perform on par with electroplated Au gratings in phase-contrast setups (Romano et al., 2017).

For visible-wavelength Si$_4$43N$_4$44 couplers, the metalMHCG is evaluated primarily by coupling efficiency and footprint. FDTD simulation at $_4$45 gives top-plane out-coupling $_4$46, substrate leakage $_4$47, and $_4$48 for $_4$49 Au (Smith et al., 2021). Simulated Gaussian-fiber injection yields $_4$50, with metal absorption $_4$51 (Smith et al., 2021). Experimentally, fibre-to-fibre $_4$52 peaks at $_4$53 at $_4$54, the $_4$55 bandwidth is $_4$56, polarization extinction exceeds $_4$57, per-coupler insertion loss is approximately $_4$58, and back-reflection is below the measurable noise floor $_4$59 (Smith et al., 2021).

5. Resonant, non-Hermitian, and modal phenomena

A distinct line of work treats the metal-integrated grating as a platform for engineered modal singularities rather than as an electrode or absorption grating. In the dielectric waveguide equipped with a metal grating, the slit width is $_4$60, the slit is in the single-mode regime, and the metal-grating thickness is varied through $_4$61, $_4$62, $_4$63, and $_4$64 (Kikkawa et al., 2020). The dielectric slab is chosen so that only two dominant TM modes, TM$_4$65 and TM$_4$66, appear in the wavelength range $_4$67–$_4$68 (Kikkawa et al., 2020).

In the zero-slit-width or empty-lattice limit, the structure reduces to a flat metal–insulator–metal waveguide with two TM-mode dispersions $_4$69 and $_4$70; after folding into the first Brillouin zone, crossings arise near which coupling occurs (Kikkawa et al., 2020). The anti-crossing gap is approximately

$_4$71

when $_4$72, so metal-grating thickness directly controls the real-frequency splitting (Kikkawa et al., 2020). The BIC lies exactly at the empty-lattice crossing $_4$73, but its branch assignment flips according to the sign of $_4$74: if $_4$75, the BIC is on the lower-frequency branch; if $_4$76, it is on the higher-frequency branch (Kikkawa et al., 2020). In the reported calculations, this sign flips when $_4$77 crosses approximately $_4$78 (Kikkawa et al., 2020).

The same device supports EPs near the BIC only for selected grating thicknesses. Numerically observed EPs occur at $_4$79, $_4$80, and at $_4$81, $_4$82 (Kikkawa et al., 2020). The cited design guidelines emphasize that tuning $_4$83, or equivalently the slit Fabry–Pérot resonance phase, controls the internal coupling constant $_4$84, the anti-crossing gap, the sign of $_4$85, and the location of EPs in $_4$86 parameter space (Kikkawa et al., 2020). This suggests that, beyond transport electrodes, metalMHCG architectures provide a tunable non-Hermitian photonic platform.

6. Applications, limitations, and comparative context

The cited literature places metalMHCGs in several application domains. In X-ray phase-contrast interferometry, hot-embossed Au–Sn microcasting provides a rapid, low-cost, and scalable route to absorption gratings with very high aspect ratios (Romano et al., 2017). Demonstrated gratings reach $_4$87 on $_4$88 wafers, and the same approach is described as extending to full $_4$89 fields required for medical imaging (Romano et al., 2017). A full hot-embossing cycle takes $_4$90, described as orders of magnitude faster than electroplating (Romano et al., 2017).

In infrared optoelectronics, metalMHCG electrodes are proposed for electroluminescent diodes, VCSELs, quantum-cascade lasers, photodetectors, thermal imaging, automotive LiDAR, free-space IR communication, and gas sensors in the $_4$91–$_4$92 range (Ekielski et al., 2023). The 2025 large-area work further identifies high-power mid-IR LEDs and lasers, interband-cascade devices, IR photodetectors and focal-plane arrays, IR imaging and LiDAR windows, transparent IR heaters, and liquid-crystal IR modulators (Bogdanowicz et al., 30 Jul 2025).

In visible photonics, the embedded-metal grating is proposed for compact efficient visible-wavelength photonic interconnects, with a view toward cryogenic deployment for quantum photonics where space is constrained and efficiency is critical (Smith et al., 2021). The reported metalMHCG length is approximately $_4$93–$_4$94, or about $_4$95–$_4$96 teeth, and the work states that the footprint is reduced by more than $_4$97 relative to standard SiN-on-SiO$_4$98 couplers at $_4$99 (Smith et al., 2021).

Several limitations are also explicit. In the X-ray case, incomplete wetting yields partial filling, empty cavities, and bent silicon ridges, while locally imperfect wetting layers can produce isolated voids near groove bottoms (Romano et al., 2017). Gratings with $_3$00 are especially prone to silicon-ridge bending or breakage under high shear, so reducing embossing force is required to preserve the structure (Romano et al., 2017). In the 2023 infrared electrode, side-wall roughness contributes to scattering, and deeper gratings at $_3$01 are identified as promising for $_3$02 unpolarized transmission but technologically more challenging (Ekielski et al., 2023). In the visible coupler, roughness from resist lift-off causes scattering losses of about $_3$03 extra, Si$_3$04N$_3$05 thickness variation of $_3$06 induces $_3$07 efficiency variation, and duty-cycle drift from design $_3$08 to measured $_3$09 reduces $_3$10 from $_3$11 to $_3$12 (Smith et al., 2021).

Comparative claims in the cited sources are specific. The 2023 infrared study states that the sheet resistance of the metalMHCG is several times lower than any other TCE considered there, while maintaining relative transmittance well above $_3$13 (Ekielski et al., 2023). The 2025 study states that, among mid- to far-IR transparent conductive electrodes on GaAs, the metalMHCG simultaneously delivers the highest unpolarized transmission and very low sheet resistance, and that it is the first TCE to significantly exceed the Fresnel limit over a broad $_3$14 M-FIR bandwidth (Bogdanowicz et al., 30 Jul 2025). In the X-ray study, the method is characterized as low cost, fast, and easily scalable to large-area fabrication (Romano et al., 2017). These statements should be read within the comparison sets defined in the respective papers.

7. Conceptual significance

Taken together, the literature shows that metalMHCG is not a single canonical device but a recurring design strategy in which metal is integrated into a monolithic grating so that the electromagnetic function of the grating reshapes the conventional penalties of metal. In the infrared electrode case, the usual conductivity–transmittance trade-off is alleviated because guided-mode resonance concentrates optical fields in GaAs with minimal overlap in Au (Ekielski et al., 2023). In the large-area mid- to far-IR design, effective-medium and Fabry–Pérot behavior permit simultaneous near-zero reflection for TE and TM at a common geometry, yielding near-unity unpolarized transmission beyond the flat-interface Fresnel limit (Bogdanowicz et al., 30 Jul 2025). In the X-ray case, monolithic metal filling of deep silicon templates produces high-density absorption gratings with large aspect ratio and area (Romano et al., 2017). In the waveguide BIC/EP case, the metal grating becomes a control knob for radiative and internal coupling, enabling tunable singular photonic states (Kikkawa et al., 2020).

A plausible implication is that the unifying value of metalMHCG lies in modal engineering rather than in the mere addition of metal. The cited implementations repeatedly rely on the same structural logic: place metal where it enhances conductivity, absorption, or coupling control, while design the surrounding grating so that the targeted optical mode either avoids the metal, interferes through it, or uses it to tune the spectrum. That common logic connects X-ray interferometric gratings, infrared transparent electrodes, visible grating couplers, and non-Hermitian waveguide devices within a single research lineage (Romano et al., 2017, Ekielski et al., 2023, Kikkawa et al., 2020, Smith et al., 2021).