Metamaterial microbolometer

FDTD HEAT Metamaterial Plasmonics A & D

Microbolometers are used as image sensor in thermal cameras. Infrared radiation is absorbed by the sensor, inducing a change in its electrical resistance. In this article, we simulate a metamaterial based microbolometer heat sensor. We will calculate the temperature changes in the sensor due to IR absorption.

download example

Overview

Understand the simulation workflow and key results

The main elements of metamaterial microbolometer are an absorber and a thermistor. The absorber is composed of a silicon layer sandwiched between a periodic gold disks and a gold mirror. Under the absorber, there is a thin silicon nitride (Si3N4) film followed by the vanadium oxide (VOx) thermistor layer. The thin silicon nitride layer delivers heat from the absorber to the thermistor layer and serves as electrical insulation. The device is suspended by a Si3N4 bridge for thermal isolation.

Typical design goal for metamaterial microbolometers is to reduce the pixel size without sacrificing detector sensitivity and the thermal time constant. This requires both optical and thermal simulations.

Step 1: Optical simulation – Calculate Reflection/Transmission and Absorption spectra

At this step we carry out an optical simulation to calculate the absorption due to IR radiation with FDTD solver . In addition , the optical absorption profile is calculated and used in the next step as input.

Step 2: Thermal simulation – Calculate Temperature change (ΔΤ)

Next, we take the optical absorption profile and use it as a heat source in the subsequent heat transport simulation in HEAT solver to calculate the temperature change in the thermistor layer. In addition, because the HEAT solves the coupled system of equations for heat transport and conductive electrical transport, we can also look at the electrothermal effect due to an applied voltage and calculate the IV response for the microbolometer.

Run and Results

Instructions for running the model and discussion of key results

Step 1: Optical simulation

Open simulation file “meta_microbolo.fsp”
Open and run script "meta_microbolo_r.lsf"

Since the absorption is mostly confined in the absorber region, we can consider only this part of the device in optical simulation and ignore the rest of the underlying layer. Assuming that the optical behavior of each pixel is identical, we can further reduce the simulation volume by only including a single unit cell of the array and impose periodic boundary conditions as if there are infinite number of such arrays. This will significantly reduce the simulation time required. If the radiation has a non-uniform distribution and each pixel experiences different amount of absorption, then more unit cells in the device will need to be included in the optical simulation.

When applicable, the simulation time can be further reduced by taking advantage of the symmetry of the structure. As seen in the figure below, a mesh override region is used to resolve the thin layers as the thickness of the dielectric layer is critical to the absorption.

The absorption resonance of this device can be tailored by changing the radius of the metal disk as shown in the left following image (reference Fig 2. (b) and (a)), which can be created by running the “ meta_microbolo_r.lsf” [1]. A radius of 530 nm is used where an absorption resonance peak at roughly 9.3 um, that corresponds to the human body radiation, can be depicted in the right plot below.

Check the radius value in "model" and run the simulation if needed
Open and run script “refl_transm_absorp_script.lsf”

The “refl_transm_absorp_script.lsf” will calculate and plot the reflection, transmission and absorption spectra depicted below (left figure). It will also visualize the cross-sectional electric/magnetic field profiles (reference Fig 2. (a) and (d)) and absorption profile at resonance 9.3 um [1]. These results can be manually visualized from the 'xz' and 'yz' monitors. The spatial distribution of absorption can be visualized by right-clicking 'pabs' analysis group and selecting 'Pabs'. You can also set the 'Parameter' setting as below and use the slicer to get the cross-sectional view at the desired locations. The results show that the magnetic field is confined mainly to the dielectric layer, and this strong magnetic resonance is responsible for the high absorption. The 'Pabs' analysis group in the optical simulation saves the absorption profile into a .mat file. In the next step we describe how the .mat file can be imported into the heat transport simulation by using the 'Import Heat' object and selecting the specified file.

Step 2: Heat transport simulation

We are going to study both the steady-state and transient characteristics of the system in response to the heat source. For the heat transport simulation, we will be considering the following structure based on the paper by Du [1], where the metamaterial absorber is rooted on a silicon substrate by two legs. The schematic diagram of the integrated microbolometer, and its perspective view in HEAT, is depicted below. To take advantage of the symmetry for shorter simulation time, only half of the 5x5 pixel structure is simulated in HEAT, as marked by the orange simulation region.

Open model “meta_microbolo_steady.ldev”
Open and run script “meta_microbolo_steady_pixel.lsf”

For the steady-state simulation, the solver mode is set to 'steady state' in the Edit Heat Solver window. After running the simulation, right-click on the 'HEAT' object and visualize the 'thermal' result. In the visualizer, select 'T' in the attribute. Since the absorbed power is relatively small, the resulting temperature change in the microbolometer is very small and is almost uniform in the absorber region. The left image below shows the steady-state thermal distribution (reference Fig. 6 (b)) for 5 um x 5 um pixel size [1]. On the other hand, the right image depicts the 15 um x 15 um pixel size steady-state temperature distribution.

Run the script “meta_microbolo_steady_pixel.lsf ” and the change in temperature, as a function of pixel size, will be plotted (plot below). Users can modify in the script the number of disks (disk_size), or the pixel size range, at the first script lines. As expected, the dT increases quadratically as a function of pixel size [1].

The simulation number out of the total number of iterations , element array size and the pixel size (um) of each simulation will be printed out at the "Script prompt" window during the simulation sweep run, portrayed at the image below.

Open model “meta_microbolo_transient.ldev”
Open and run script “meta_microbolo_extract.lsf”.

For the transient simulation, the solver type is set to 'transient' in the Edit Heat Solver window. The on/off setting of the import source can be adjusted in the 'Transient' tab of the same window. Run the script “ meta_microbolo_extract.lsf” to obtain the transient temperature response. The script runs two simulations for two different pixel sizes, one for 5 um and another one for 10 um. The following plot was obtained by running optical and heat transfer simulations for pixel sizes of 5 um and 10 um. It is worth noting that while a larger pixel size gives a larger temperature change, it takes longer to reach a steady state level.

Important model settings

Description of important objects and settings used in this model

Users can choose the number of the metamaterial absorber unit cells and create an array that consists of a pixel. A script is included in the “Setup window” of the model that automatically creates the correct number of heat sources and imports the “heat_source.mat” in each disk. The left image below shows the setup parameters of the model, where the size of the pixel can be defined. The right image depicts the path (line 29) where the “heat_source.mat” file is imported. The size of the pixel can be roughly calculated by multiplying the number of cells with the period.

For both, steady and transient simulation cases, a fixed temperature of 293.15 K is applied to the two pillars supporting the microbolometer as shown in the following figure. The image below depicts the boundary condition settings in HEAT solver.

Taking the Model Further

Information and tips for users that want to further customize the model

Multiphysics CMOS sensor

An extension of the Microbolometer application, including CMOS image sensor and coupling to SPEOS simulation tool , can be part of taking this model further. In that way an advanced opto-thermal-electrical coupling can be covered. The micro-bolometer absorbs infrared radiation and converts it to heat where its temperature change will modulate the CMOS read-out. Then, the resulting signal can be analyzed in the SPEOS workstation. The description of how CMOS image sensors are modelled in SPEOS ( CMOS-Sensor-Camera-Image-Quality-Analysis-in-a-3D-Scene ) shows relevant workflow for coupling imaging sensors. In that way, full system-level predictions of IR sensor performance under realistic scenes and environmental conditions are enabled. A Multiphysics modelling framework for CMOS-SOI-MEMS IR sensors (absorber, thermal modelling, mechanical simulation) gives insight into advanced sensor modelling [2].

Additional Resources

Additional documentation, examples and training material

Related Publications

Kaikai Du et al., "Wavelength and thermal distribution selectable microbolometers based on metamaterial absorbers," IEEE Photonics Journal, 7, 6800908 (2015).
F. Forsberg et al., "CMOS-Integrated Si/SiGe Quantum-Well Infrared Microbolometer Focal Plane Arrays Manufactured With Very Large-Scale Heterogeneous 3-D Integration," IEEE Journal of Selected Topics in Quantum Electronics, vol. 21, no. 4, pp. 30-40, July-Aug. 2015, doi: 10.1109/JSTQE.2014.2358198.