Extending the validity of squeezed-film damper models with elongations of surface dimensions

Abstract

Compact models for micromechanical squeezed-film dampers with gap sizes comparable to the surface dimensions are presented. Two different models considering both the border flow and non-uniform pressure distribution effects are first derived for small squeeze numbers. In the first 'surface extension' model the border effects are considered simply by calculating the damping with extended surface dimensions, and in the second 'border flow channel' model an additional short fictitious flow channel is placed at the damper borders. Utilizing a large amount of two-dimensional (2D) FEM simulation results by varying the damper dimensions, mainly the ratio a/ h between the surface length and the air gap height, surface elongations are extracted using both elongation models. Both linear and torsional modes of motion are considered at the continuum flow regime. These results show that the 'surface extension' model is superior, since the extracted elongation a is almost constant ( a = 1.3h), leading to a very simple model. Next, the rare gas effects are included in the 'surface extension' model in the slip flow regime (Knudsen number 0 < Kn < 0.13). Considering slip boundary conditions, 2D FEM simulations are repeated and the elongations, now functions of the Knudsen number, are extracted from the results. A simple fitted equation for calculating the surface extension a finally results. The maximum relative errors of the models are smaller than 10% for a/ h > 4 in the linear motion and for a/ h > 10 in the torsional motion. The model assumes incompressible flow and thus the maximum frequency where the models are valid is limited. In typical MEMS topologies where the elongations must be considered, this means that the models are valid below frequencies of 500 kHz. To also model rectangular 2D squeezed-film dampers, these elongations are applied directly in the surface length and width used in the compact models. Comparison with three-dimensional (3D) FEM simulations shows that the new model gives excellent results, and it extends the validity range of existing compact models. The maximum relative error of the models is smaller than 10% for a/ h > 16 in the linear motion and for a/ h > 16 in the torsional motion. The new surface extension model is useful in simulating both the circuit level and the system level behavior of gas-damped microelectromechanical devices with aspect ratios greater than 2 in the time and frequency domains.

A micro-accelerometer MDO benchmark problem

“…The analysis models are obtained from Mukherjee et al (1999) and Zhou (1998), except for the mechanical spring stiffness analysis and some extensions to the damping coefficient model. The spring model is developed here from linear beam theory, and the damping model is extended to also account for the small aspect ratios in squeezed film damping using the models of Veijola et al (2005).…”

“…Squeezed film damping between the rotor and stator comb fingers (2 times N f + N s gaps) can be approximated by a damping coefficient b s (Veijola et al 2005)…”

Numerical and compact modelling of squeeze-film damping in RF MEMS resonators

“…These equations are identical with the "narrow gap" equations specified in [6], but they include an additional equation (12). The strategy to have an approximate solution for the equations above is the following.…”

“…Also, the pressure is no more constant across the gap even at small frequencies. These phenomena have been studied in [12] regime is between the squeeze-film and the gap resonance regions. Table II summarizes the resonant frequencies detected from the amplitude/phase responses of all simulated topologies.…”

A macromodel for squeeze-film air damping in the free-molecule regime

“…One is based on the modification of continuum theories. Examples in this category include, but are not limited to, the modification of Reynolds' equation by employing the effective viscosity obtained either from a linearized Bhatnagar-Gross-Krook solution of the Poiseuille flow 12,14,23 or from experimental data 24,25 via curve fitting, by using a pressure dependent effective flow rate coefficient 13 and a modified pressure boundary condition with coefficients extracted from the NavierStokes slip-jump simulations and the direct simulation Monte Carlo ͑MC͒ simulations. 15 Most of these methods have demonstrated their accuracy in the slip regime ͑Kn Յ 0.1͒, i.e., the near continuum regime.…”