Infrared thermochromic antenna composite for self-adaptive thermoregulation

Self-adaptive thermoregulation enables systems to maintain stable temperatures by balancing radiative heat loss and gain. Passive approaches that consume no operational energy are especially attractive and often rely on thermochromic materials that undergo temperature-driven phase changes near a critical temperature Tc (e.g., VO2, chalcogenides like GST and IST, and correlated perovskites). Through Kirchhoff’s law, temperature-dependent absorption translates into temperature-dependent emissivity, characterized by the hot-to-cold emissivity contrast Δε = εh − εc, which must be positive for passive thermoregulation. When the medium deviates from Tc, a radiative feedback pushes it back towards equilibrium. Application-relevant spectral bands include the MWIR (3–8 µm) for thermal imaging and the LWIR (8–15 µm) for radiative cooling, aligned with the atmospheric window and blackbody peaks at typical temperatures. Existing solutions often use resonant structures (Fabry–Perot cavities, metasurfaces, thin films), which can achieve adequate switching but face scalability, cost, substrate constraints, and planarity limitations. Recent spherical core–shell particle approaches improved manufacturability but yielded modest Δε (~0.26 in the atmospheric window) and are not optimal in material usage or performance because spherical geometries inherently limit absorption cross-section contrast.

Prior work on adaptive thermal emission predominantly employs engineered resonant structures: multilayer Fabry–Perot cavities, metasurfaces, and thin films leveraging thermochromic materials (e.g., VO2, GST/IST, perovskites). While these can deliver emissivity modulation, they typically require specialized deposition, high-temperature processing, and rigid substrates (silicon, quartz, sapphire), raising cost and limiting form factor to planar surfaces. Spherical or core–shell VO2 particles have been explored to improve scalability, but theoretical and experimental analyses indicate limited absorption cross-section ratios in LWIR (≈4 for optimized 2 µm spheres), bounding achievable Δε to <0.35 in composites. The literature also highlights the importance of positive differential emission (Δε > 0) for passive regulation, and identifies application windows: MWIR for imaging and LWIR/atmospheric window for radiative cooling. The need remains for scalable, low-cost, free-form solutions offering large Δε and tunable spectral response.

Materials and synthesis: VO2 antennas were synthesized hydrothermally. For rods: V2O5 (0.48 g) was dispersed in water (10 mL), H2SO4 (0.75 mL) added at 80 °C; hydrazine hydrate (270 µL) introduced to reduce V, followed by CTAB (1.8 mL, 0.1 M). Solution pH adjusted to 4.0–4.2 with NaOH (1 M), then hydrothermally treated at 220 °C for 63 h (45 mL Teflon-lined autoclave). Precipitates were washed, dried (80 °C, 1 h), and transformed from VO2(A) to VO2(M) by vacuum anneal at 550 °C, 0.2 mbar, 1 h. For stars: similar procedure with higher hydrazine (330 µL), lower CTAB (300 µL), and higher temperature (235 °C, 63 h) to obtain VO2(M) stars in one step. Antenna morphology was tuned by adjusting reductant/surfactant concentration, pH, and temperature. Antenna dimensions: rods exhibited W × L ≈ (0.47 ± 0.01 µm) × (14.6 ± 9.2 µm). PXRD/XPS confirmed VO2(M) phase; DSC measured Tc = 70.2 °C. Composite fabrication: Flexible films were prepared by mixing VO2 antennas with polyethylene and hot-pressing against aluminum foil acting as a back reflector. Experimental films used LDPE (0.8 g) with antioxidant (IRGANOX 0.01 g) and VO2 (0.002–0.02 g), compression molded at 180 °C/350 bar for 3 min to produce ~80–85 µm films; the aluminum backing prevented radiative exchange with the substrate. Volume fractions were estimated from mass ratios and densities (VO2(M) 4.23 g cm−3; LDPE 0.925 g cm−3). A spray-coating route dispersed VO2 powder in acetone (2 wt%), ultrasonicated 4 h, airbrushed onto heated aluminum (50 °C), then overcoated by dip-coating in 2 wt% PE in toluene to form free-form composites. Characterization: SEM (JEOL JSM-6701F) assessed morphology and size distributions. PXRD (STOE SEIFERT, Mo Kα) and XPS (Thermo K-alpha) verified phase and composition. DSC (Mettler Toledo) characterized IMT (heating 5 °C min−1; cooling 0.5 °C min−1). LWIR thermal imaging used a FLIR A655c camera. Spectral measurements employed FTIR (Shimadzu IRTracer-100) with a gold mid-IR integrating sphere and a thermoelectric temperature controller to capture hot/cold emissivity and hysteresis. Modeling and simulation: A Beer–Lambert composite model linked individual antenna absorption cross-section ratio (Cabs,h/Cabs,c) to composite emissivity contrast Δε, highlighting roles of particle volume fraction f, film thickness δfilm, host absorption α, and absorption mean free path Aabs = Vp/(f Cabs,c). Orientation- and polarization-averaged scattering/absorption cross sections for arbitrarily shaped antennas (rods, stars, flakes, spheres, core–shells) were computed with SCUFF-EM (AVESCATTER), using Gmsh meshes with convergence checks (absolute tolerance ≤0.05). First-principles radiative transfer was performed via mc-photon Monte Carlo simulations (1,000,000 photons per wavelength) including absorption, scattering, Fresnel reflections/refractions, and host absorption, using ensemble-averaged cross sections and asymmetry parameter g for polydisperse distributions. Host refractive index was taken as ~1.5 for transparent polymers; real PE absorption (~1.3 mm−1 in the atmospheric window) was included to assess practical performance. Antenna geometry parametrics (length L, width W, SA:V) were swept to identify resonant tuning and maximize Cabs,h/Cabs,c at target wavelengths (e.g., 10 µm in LWIR, 5–6 µm in MWIR).

• Non-spherical VO2 thermochromic antennas (rods, stars, flakes) achieve very large hot-to-cold absorption cross-section ratios, enabling strong emissivity switching with minimal loading and thin, flexible films.
• Experimental LWIR composite (VO2 rods + PE on Al) exhibited spectral emissivity change at 10 µm from εc ≈ 0.33 to εh ≈ 0.88 (Δε ≈ 0.56); average Δε over the atmospheric window was 0.44 with narrow hysteresis width ΔThyst ≈ 7.7 °C; film thickness <100 µm, fully flexible.
• Thermal imaging confirmed a clear emissivity switch near T ≈ 70 °C (Tc ≈ 70.2 °C) compared to ε ≈ 0 (aluminum) and ε ≈ 1 (carbon black) references; simulations agreed well with measurements.
• Optimization via SA:V: Spherical particles are intrinsically limited (max Cabs,h/Cabs,c ≈ 4 in LWIR for ~2 µm spheres, bounding Δε < 0.35). Increasing SA:V (disks, rods) dramatically boosts Cabs,h and reduces Cabs,c, leading to cross-section ratios exceeding 200 for rods—over 50× higher than spheres.
• Geometry tuning: For rods, reducing W (200→50 nm) increases ratio by ~4; decreasing L (2→1 µm) blue-shifts the peak from 10→5 µm. Optimal LWIR targeting at 10 µm occurs around L ≈ 2 µm, W ≈ 50 nm.
• Predicted composite performance (transparent host, n ≈ 1.5): Δε up to 0.91 with f = 0.11% v/v for optimal rods; with experimentally realized rod dimensions (conservative), Δε ≈ 0.76 at f ≈ 0.51% v/v. PE host absorption reduces Δε but the design remains effective.
• Polydispersity tolerance: With coefficient of variation (CV) = 2, Δε reductions of ~5% (optimal) and ~11% (conservative); experiments had CV < 0.63, implying ~3.5% Δε reduction.
• MWIR-targeted VO2 nanostars in PE showed Δε = 0.52 at λ = 5.4 µm with ΔThyst ≈ 7.2 °C, aligning with the geometry-induced resonance shift.
• Morphology comparison (optimized at 10 µm): absorption cross-section ratio—rods 132.4; stars 56.7; flakes 52.1; core–shell spheres 9.7; solid spheres 3.9. Scattering can enhance or diminish emissivity depending on relative normalized Csca/Vp vs Cabs/Vp; stars/flakes benefited from scattering-driven path-length enhancement, while core–shell/spheres suffered reflectance increase.
• Material efficiency and scalability: Lab-scale hydrothermal batches produced >200 mg VO2 antennas, whereas <2.15 mg of rods suffice per m² of composite; antennas can be incorporated via scalable processes (film extrusion, spray/dip coating, compression molding, electrospinning, etc.).
• Generality: Design rules extend to other IMT materials; predicted maximum Δε for rod composites: GST ≈ 0.84, IST ≈ 0.92.

The study addresses the core challenge of achieving large, tunable emissivity switching in scalable, low-cost, non-planar formats. By framing emissivity modulation through the absorption cross-section ratio (Cabs,h/Cabs,c) and linking it to composite Δε via Beer–Lambert and radiative transfer modeling, the work identifies high SA:V dipole antennas as the critical lever. Rods, stars, and flakes suppress non-radiative channels in the hot (metallic) phase and minimize polarization-current-driven absorption in the cold (insulating) phase, maximizing the differential response. Experimental composites validate this approach with substantial Δε and narrow hysteresis at thin film thicknesses and very low filler loadings, while simulations corroborate performance and provide robust design maps across geometries, sizes, concentrations, and hosts. Sensitivity analysis shows limited impact from polydispersity and clarifies the nuanced role of scattering: it can either enhance path length (benefiting stars/flakes) or increase reflectance (hurting spheres/core–shells). The findings provide a versatile, free-form pathway to self-adaptive radiators tailored to MWIR and/or LWIR windows, directly relevant to radiative cooling, thermal sensing, imaging, and camouflage. Furthermore, the framework outlines conditions for negative differential emission (Δε < 0) upon percolation, enabling thermal camouflage modalities where apparent temperature is decoupled from actual temperature.

This work introduces infrared thermochromic dipole antennas as a universal, scalable platform for self-adaptive thermoregulation. By maximizing absorption cross-section contrast via high SA:V geometries (especially rods), composites achieve large emissivity switching (measured Δε ≈ 0.56 at 10 µm; average Δε ≈ 0.44 in LWIR with narrow hysteresis) at ultralow filler loadings in thin, flexible films. Modeling predicts Δε approaching 0.9 for optimized rods in low-loss hosts, and demonstrates spectral tunability by tailoring antenna dimensions and morphology (rods for LWIR, stars for MWIR, combinations for broadband). The approach is compatible with diverse industrial processes and surfaces, enabling coatings, films, fibers, and membranes for energy-efficient thermal management, sensing, and camouflage. Future work can tailor transition temperatures (e.g., W-doping, strain, oxygen vacancies) toward near-room-temperature operation, extend antenna morphologies/materials (GST, IST, perovskites), and explore regimes beyond percolation for negative differential emissivity to realize advanced thermal camouflage and dynamic radiators.

• Host material absorption reduces achievable Δε versus ideal transparent hosts; practical polymers like PE (≈1.3 mm−1 in the atmospheric window) damp performance relative to predictions for lossless media.
• Performance relies on precise antenna geometry and dispersion; although tolerant to polydispersity, deviations in length/width and orientation could shift resonance and reduce contrast.
• Current VO2(M) transition temperature (Tc ≈ 70.2 °C) may be above desired setpoints for some applications; while tunable via doping/strain/defects, such adjustments were not implemented here.
• Scattering can either assist or hinder emissivity depending on morphology and normalized cross-sections; careful control is needed to avoid reflectance increase (notably in spherical/core–shell designs).
• Results emphasize sub-percolation regimes (Δε > 0); behavior near and beyond percolation (Δε < 0) is discussed conceptually but not experimentally validated in this work.