978-3-9810801-7-9/DATE11/©2011 EDAA An Electrical Tes Accelerometers A.A. Rekik 1 LIRMM - CNRS/Un 2 ENIS - University of Abstract — In this paper, an alternative test m convective accelerometers is presented. It is fi that device sensitivity can be determined wi physical test stimuli by simple electrical measu previously developed behavioral model that Monte-Carlo simulations, we have established a between electrical test parameters and d Proposed test method is finally evaluated for d that privilege yield, fault coverage or test efficien Keywords: MEMS testing, convective accelerom electrical test I. INTRODUCTION MEMS testing is a challenging issue du domain nature of MEMS devices. They ther application of physical test stimuli to verify the As a result, MEMS testing requires specific a test equipment that is more expensive than st interesting approach is to develop alternativ test procedures, and numerous solutions have the last decade for various types of M accelerometers [1-5], magnetic field senso sensors [7]… In this paper, we focus on ME accelerometers. On the one hand, accele requires expensive test equipments with mo and long test sequences due to their a accelerations only in low frequency ranges. O literature reports only methods applicable to c [1-5] where electrostatic actuation can be us alternate electrical tests. It is therefore our obj an alternative electrical test method for ME accelerometers. More specifically, our goal motionless test method that can be applied usin test stimuli and that permits to verify device s the most challenging specification to measure a calibrated acceleration. The paper is organized as follows. In sectio the convective accelerometer together with model and we introduce a list of parametric fa to process scattering) that can affect the devic alternative electrical test method, its implem evaluation are presented in section III and Finally, evaluation results are discussed in sect st Method for MEMS Co s: Development and Eva 1,2 , F. Azaïs 1 , N. Dumas 1 , F. Mailly 1 , P. Nouet 1 niv. Montpellier 2 - 161 rue Ada, 34392 Montpellier, Fran Sfax - Route Soukra, Cité Elhabib BP W 3052 Sfax, Tun method for MEMS irst demonstrated ithout the use of urements. Using a t allows efficient a good correlation device sensitivity. different strategies ncy. meter, alternative ue to the multi- refore require the eir specifications. and sophisticated tandard ATE. An ve electrical-only been proposed in MEMS such as ors [6], pressure EMS convective erometer testing ovable test heads ability to detect On the other hand, capacitive sensors sed to implement ective to propose EMS convective is to develop a ng only electrical sensitivity, as it is without applying on II, we describe h its behavioral aults (mainly due e sensitivity. The mentation and its IV, respectively. tion V. II. DEVICE UN A. Device Overview The device under test is a obtained by Front-Side Bulk Micr CMOS die fabricated in a 0.8 µm Microsystems® (Fig.1). Three thin CMOS process back-end layers (ox and nitride), are suspended over a polysilicon is used to embed resistor temperature sensing (R D1 , R D2 ). T bridge) is biased with an electrical v bubble confined in the bottom (i.e. package) cavities: the temperature heater location and minimum at the lateral dimensions are the half-wid (r 1 ) and the cavity (r 2 ), and the dista one detector (d). Figure 1. SEM picture of the prototype parameters: r1=20µm, r2=350µm, d=17 In absence of acceleration along the temperature of detectors (i.e. la identical for symmetry reasons. Un sensitive axis (AA’), the hot bub convection and a differential temper detectors. Thanks to the Temperatur (TCR) of polysilicon, this differenti Heater Amplifier A1 Heater Amplifier A1 Heater Amplifier A1 onvective aluation nce nisia NDER TEST convective accelerometer romachining (FSBM) of a m technology from Austria n bridges, composed of the xide, polysilicon, aluminum, a silicon etched cavity and rs, for both heating (R H ) and The heater R H (i.e. central voltage (U H ) to create a hot etched silicon) and top (i.e. e is then maximum at the e cavities boundaries. Main th of both the heater beam ance between the heater and and corresponding geometrical 75µm, h1=270µm, e=5.2µm g the sensor sensitive axis, ateral bridges: R D1 , R D2 ) are nder acceleration along the bble deforms due to free rature appears between both re Coefficient of Resistance ial thermal signal implies a Detectors Sensing direction (AA) Detectors Sensing direction (AA) Detectors Sensing direction (AA)