Microbolometer Design and Simulation

From night vision cameras used by the military and law enforcement to medical imaging technologies such as digital infrared thermal imaging, thermal cameras are an integral part of the systems and technologies that surround us. With the increasing commercial applications of this technology, the thermal camera business has been growing nicely and this growth is expected to accelerate over the next decade.

Figure 1 Uncooled thermal camera market forecast, millions units. (Yole Development)

The most widely used sensing technology employed in thermal (infrared) cameras is based on the uncooled Microbolometer. The traditional Bolometer, built from thermistor across and a Wheatstone bridge^[1], was invented in 1878. Microbolometers were developed in the late 1970s and work by measuring the change in resistance in a heat sensor caused by incident infrared radiation. They are typically positioned together in an array or grid so that the temperature distribution can be represented graphically as an image or picture.

Design Challenges

The Microbolometer is primarily a thermal sensor and so understanding its electro- thermal characteristics is one of the key challenges faced by the Designer. The current-voltage (IV), voltage-temperature (VT), thermal time constant (TC) and operating response are the key performance characteristics that need to be simulated and optimized.

Coventor provides a complete design flow for exploring and simulating these characteristics. The flow is based upon MEMS specific solid modeling and finite element tools. First, material properties, process information and layout are defined using the CoventorWare’s Designer module. Designer uses this data to create a 3D model, which is then automatically meshed for simulation in CoventorWare’s Analyzer module. Results can be visualized in table format, 2D graphs or 3D via a dedicated post processor.

Figure 2 Coventor Microbolometer Design flow (design from the University of Pretoria)

The material database includes temperature dependent properties of typical bolometer materials e.g. vanadium pentoxide and titanium. The material data is coupled to a definition of the foundry process used to fabricate the Mircobolometer. The process can be user-defined or selected from one of Coventor’s Foundry Design Kits. Layout is defined in a comprehensive layout drawing tool which provides engineers with the tools to easily draw individual bolometer pixels and create parameterized arrays.

CoventorWare’s Analyzer module is employed to perform thermal and electrothermal simulations on the Microbolometer model. This versatile solver allows engineers to assign voltage and temperature profiles as inputs to simulate the steady state and transient responses. The effect of heat transfer due to conduction and convection are also easily incorporated. The results from these simulations are studied from the tables, graphs and the 3D images generated by the post processor.

1. MEMS Design ExplorationTo simulate IV and VT curves, temperature boundary conditions are assigned. The operating voltage is varied over a user-defined range and the current flow and associated Joule Heating simulated using the CoventorWare Analyzer module. Understanding the IV curve (Figure 3a), whose slope is the resistance of the Microbolometer, is important as it dictates the sensitivity of the device. The V-T curve (Figure 3b) allows Designer to understand the effect of Joule Heating on the bolometer and determine the temperature gain of the bolometer at steady state when a given potential is applied across it.

   To simulate the thermal TC (see Figure 4a), mechanical and temperature boundary conditions are assigned together with a voltage pulse. A time domain electro-thermal simulation is then performed, with the CoventorWare Analyzer module. The thermal TC is the time taken for the temperature to reach 66% of its steady state value when a voltage pulse is applied. This parameter gives designers information about the response time of the sensor and helps them design the appropriate control electronics and readout circuits. The curves below show that the predicted time constant for the example bolometer used here is few milliseconds.

   CoventorWare Analyzer is also used to simulate the operating response with time. A heat flux is first applied to model the incident infrared radiation. When the steady state temperature is reached, a voltage is input across the Microbolometer in order to simulate the resistance, which is related to the incident radiation, see Figure 4b.

   

A Complete Platform

Coventor’s solution offers a MEMS specific design flow to easily simulated, tune, and optimize MEMS Microbolometers. Pre-defined process design kits from commercial foundries are also available to accelerate the design process and reduce human error. CoventorWare Analyzer is able to predict the key device performance parameters using it’s MEMS specific multi-physics solvers and visualization tools. Critical performance parameters such as noise equivalent temperature difference and thermal capacitance can be calculated based on Analyzer’s simulation results.