Modeling Combined Thermal, Electrical, Optical and Mechanical Response for MEMS Spectroscopic Gas Sensor Based On Photon

  • PDF / 230,873 Bytes
  • 6 Pages / 612 x 792 pts (letter) Page_size
  • 75 Downloads / 203 Views

DOWNLOAD

REPORT


J12.3.1

Modeling Combined Thermal, Electrical, Optical and Mechanical Response for MEMS Spectroscopic Gas Sensor Based On Photonic Crystals Anton C. Greenwald, Martin U. Pralle, Mark P. McNeal, Nicholas Moelders, Irina Puscasu, James T. Daly and Edward A. Johnson Ion Optics, Inc., Waltham, MA 02452 USA Abstract A new type of gas sensor was developed that combines the principles of bolometric infrared detectors with photonic crystals.1,2 This paper describes a quantitative model used to optimize the materials, geometry, and electrical properties of this suspended membrane MEMS device. Fundamentally the model is concerned with the thermal response of the device using temperature dependent thermal conductivity, specific heat, and electrical resistance to calculate conduction, convection, and radiation losses for a negative temperature coefficient of resistance material. Variations in the electrical drive circuit, dc and ac response, low and high frequency sinusoidal and random noise, along with an exacting calculation of expected signal were used to improve design. The model follows the time evolution of the system. We show how look-up tables with scaling (derived from exact, off-line finite element models for thermal conduction, spectral emission, etc.) provided sufficiently accurate estimates with rapid calculation to enable running the model on a standard PC type computer. The simulations matched the experimental results, accurately predicted the unstable operating regimes, and maximized the signal to noise ratio for the device. Introduction The gas sensor is a micro-hotplate filament designed to emit thermally generated infrared photons within a discrete narrow band defined by a photonic crystal.3, 4, 5 Emitted light is reflected back onto the filament that acts as it own bolometer sensor. By tuning the wavelength of emission to the absorption energy of a gas, light is attenuated, cooling the filament and a resistance change is measured as the signal. Emission of light from the surface as a function of surface structure has been modeled using the transfer matrix method in a computationally intensive approach requiring multiple, parallel processors. Using a finite element approach, ANSYS is an example, the distribution of current, resultant temperature profiles, and mechanical stress in the device can be calculated on a work station. Fluid dynamic programs can be used to compute the thermal losses due to convection. Electrical response of a circuit including such a device can be modeled through SPICE. Typically, all of these calculations are performed separately, non-interactively. Our problem was how to perform all of these calculations simultaneously with many varying parameters to optimize device design. Preferably, using a program that can run on a PC. Also, the model had to account for wide variations due to fabrication tolerances. Our approach was to use a FORTRAN like code (IDL)5 with prior calculated lookup tables for the results of the different problems described.

J12.3.2

Figure 1 Cad drawing of MEMS s