Gold-Black Coating Predicted to Boost Microbolometer Sensitivity

Researchers have developed a coupled optical-electrothermal model to compare the performance of different infrared absorber coatings on hydrogenated amorphous-silicon microbolometers. The study, published in Discover Condensed Matter, focuses on the potential of gold-black, a highly porous nanostructured coating, to enhance the sensitivity of uncooled thermal imaging sensors. By simulating the interaction between light absorption, electrical self-heating, and thermal noise, the team identified specific geometric configurations where gold-black coatings outperform traditional dielectric alternatives. The work provides a numerical framework for preliminary design optimization, highlighting the trade-offs between absorption efficiency, thermal mass, and electrical bias constraints in next-generation thermal cameras.

The central finding of the numerical analysis indicates that gold-black-coated devices can achieve significantly higher specific detectivity than dielectric-coated counterparts under specific operating conditions. In the long-wave infrared band, spanning 8 to 15 micrometers, the selected gold-black configuration yielded a predicted detectivity of 2.635 × 10⁹ cm√Hz W⁻¹ at 30 Hz. This value is 2.97 times higher than the best-performing dielectric candidate in the same band. In the mid-wave infrared band, ranging from 3 to 5 micrometers, the gold-black device achieved a detectivity of 3.147 × 10⁹ cm√Hz W⁻¹, which is 5.17 times greater than the optimal dielectric alternative. These ratios apply to complete, independently varied device configurations rather than a simple material substitution at identical geometry.

Microbolometers are widely used in thermal imaging because they operate near ambient temperature and can be integrated with micro-electro-mechanical systems and silicon readout electronics. Hydrogenated amorphous silicon remains an attractive thermistor material because its resistance and temperature coefficient of resistance can be tailored through deposition and post-processing while maintaining compatibility with standard silicon fabrication. However, the sensing film, absorbing layers, thermal supports, bias condition, and readout form a coupled design problem. The study addresses this complexity by developing a model that explicitly includes electrical self-heating and separates absorption in the suspended membrane from absorption in the anchored reflector. This approach allows for a more accurate assessment of how different optical stacks influence the overall thermal and electrical response of the sensor.

The optical model utilizes a homogenized effective-medium description for the porous gold-black layer, based on the Landau-Lifshitz-Looyenga mixing rule. This approximation treats the complex dendritic and fractal microstructure of the deposited coating as a uniform medium with an effective dielectric function. While this method provides a compact modelling route, the authors note that it does not resolve the detailed morphology-dependent optical behavior of actual fabricated coatings. The dielectric films, consisting of silicon dioxide and silicon nitride, are represented by single Lorentz oscillators with inherited model inputs. The multilayer propagation is calculated using recursive interface amplitude coefficients, retaining both real-index phase and extinction to account for coherent interference effects within the film and cavity stack.

The thermal network represents each suspended film as a thermal node with specific heat capacity and conductance to adjacent layers. The support is modeled as a lumped equivalent nickel-chromium thermal path, rather than a resolved pair of MEMS legs, to maintain numerical reproducibility. The model assumes an ideal vacuum environment, neglecting residual-gas heat transfer. Electrical self-heating is retained in the calculations, with an ideal continuous current selected to meet a sensor temperature rise cap of 5 Kelvin. This cap is an imposed operating-window choice, not a fabrication limit or measured safety threshold. The nominal design constraint includes a time constant limit of 10 milliseconds, which is evaluated separately from the 30 Hz characterization frequency.

Parameter search and numerical verification were conducted to identify promising parameter combinations. The search bounds followed the authors’ original numerical study, with positive lower thicknesses to preserve layer topology. An isothermal screening approximation first rescored nearly 1.6 million inherited local-search vectors, followed by local refinement of over 25,000 vectors. A broader differential-evolution exploration was also performed to ensure a wide search space. The numerical implementation was checked using passive limiting cases, per-layer energy conservation, and independent nonlinear time-domain sinusoidal tests. For selected configurations, the maximum optical energy residual was below 2 × 10⁻¹⁵, and the independent reflectance discrepancy was below 1.3 × 10⁻¹⁵, confirming the implementation’s consistency.

The results show that the gold-black configurations exhibit a broader response than the selected dielectric configurations in the inherited optical model. Band-weighted membrane absorptances for the selected gold-black devices were 0.873 in the long-wave infrared band and 0.916 in the mid-wave infrared band. In contrast, the best dielectric candidates had absorptances of 0.773 and 0.737, respectively. The noise budget analysis reveals that both Johnson noise and support thermal-fluctuation contributions are relevant. Electrical self-heating is also non-negligible, with DC Joule powers ranging from approximately 0.87 to 2.08 microwatts under the selected bias policy. Removing the Joule term would change the operating point and would no longer represent the calculated conditions.

To isolate the effect of the absorber material, the researchers performed an additional calculation fixing the pixel dimensions to 30 by 30 micrometers and the sensing-film thickness to 1 micrometer. Under these common geometry constraints, the gold-black Type A configuration exceeded the best dielectric package by factors of 1.44 in the long-wave infrared band and 2.88 in the mid-wave infrared band. The reduction from the independent-device ratios of 2.97 and 5.17 demonstrates that the performance gain cannot be attributed wholly to the absorber material. Even the common-geometry test is a comparison of absorber packages, not an intrinsic bulk-material measurement, highlighting the importance of system-level optimization.

Sensitivity analysis indicates that the predicted detectivity is weakly sensitive to the assumed gold-black conductivity over the tested range. Reducing the conductivity from 200 to 0.01 W m⁻¹ K⁻¹ changes the detectivity by about -1.30% in the long-wave infrared band and -0.17% in the mid-wave infrared band. However, the 5 Kelvin temperature rise cap is consequential; reducing it to 1 Kelvin lowers the predicted detectivity by roughly 48 to 49%. The study also notes that the selected configurations touch search bounds and are not proven global optima. The thin sensing films and porous coatings require mechanical and process assessment, as the numerical search does not establish manufacturability.

The model has several limitations that must be considered when interpreting the results. The optical dispersion coefficients and porous effective-medium approximation require independent experimental assessment. Measured studies show that gold-black absorption and stability depend on morphology, porosity, uniformity, patterning, and ageing, which are not fully resolved in the homogenized approximation. The equivalent support path makes the calculation reproducible but does not verify a particular pair of physical support legs. Interface resistance, nonuniform in-plane temperature, and residual-gas conduction are unresolved. Furthermore, the continuous ideal-current operating condition and 5 Kelvin cap are deliberate numerical choices. Actual pulsed bias, electrical readout dynamics, source noise, photon noise, and calibrated 1/f noise can alter both the operating point and detectivity.

The contribution of this study is a coupled optical-electrothermal numerical design framework and a conditional comparison of absorber architectures within stated search ranges and model assumptions. It does not claim a universal fivefold material gain or a fabricated-device performance claim. The lack of experimental validation, calibrated low-frequency readout noise, and established fabrication feasibility means that these predictions remain conditional.

By combining optical, thermal, and electrical simulations, the team was able to identify specific design parameters that maximize performance. The study highlights the potential of gold-black coatings to enhance sensor sensitivity, but also emphasizes the need for experimental validation and process development.

Subject of Research: Optical-electrothermal modelling of gold-black coated a-Si:H microbolometers

Article Title: A study on gold-black coated a-Si:H microbolometer modelling

Article References: Toğrul, Z. Ç., & Gürel, Ç. S. (2026). A study on gold-black coated a-Si:H microbolometer modelling. Discover Condensed Matter, 1(1), Article 6. https://doi.org/10.1007/s44570-026-00007-y

Image Credits: AI Generated

DOI: 10.1007/s44570-026-00007-y

Keywords: microbolometer, gold-black, amorphous silicon, infrared imaging, thermal sensing, optical modelling, electrothermal network, detectivity, MEMS, absorber coating, coated, a-Si