Frédérick Mailly

Anemometer with hot platinum thin film

The techniques of micromachining silicon are used for the manufacture of an anemometer with low electric consumption and great sensitivity. To reduce the energy consumption, a suspended membrane of silicon rich silicon nitride SiNx makes it possible to carry out the heat insulation between the heater and the substrate. Platinum (Pt) thin film (3000 A) with titanium (300 A) adhesion layer on SiNx/Si substrate is used for the hot resistor. Among the methods of Pt deposition tested, electron beam evaporation gives the best results for the temperature coefficient of resistance (TCR) of Pt. Its response time is about 6 ms. Sensitivity in laminar and turbulent flow range are respectively 4.80 mV/(m/s)0.45/mW and of 0.705 mV/(m/s)0.8/mW for about 20 mW power supplied. The experiments show that the temperature rise of the sensor is not sensitive to the ambient temperature. Moreover, sensor response shows no significant changes according to parallel or perpendicular orientation of the gas flow.

Design of a micromachined thermal accelerometer: thermal simulation and experimental results

This paper describes numerical simulation of a micromachined thermal accelerometer and experimental measurements. The sensor principle consists of a heating resistor, which creates a symmetrical temperature profile, and two temperature detectors symmetrically placed on both sides of the heater. When an acceleration is applied, the free convection is modified, the temperature profile becomes asymmetric and the two detectors measure the differential temperature. This temperature profile and sensor sensitivity according to the distance heater-detector have been studied using numerical resolution of fluid dynamics equations with the commercial code CFD2000/STORM: it shows that the optimum distance between the temperature detectors and the heater is about 300 μm. A thermal accelerometer with 3 pairs of detectors placed at 100, 300 and 500 μm from the heater was manufactured using the techniques of micromachining silicon and experimental measurements have shown a good agreement with the numerical simulations: the experimental optimum distance between heater and detectors seems to be close to 400 μm and the differential temperature of detectors is about 3 °C/g for an operating heater power of 54 mW and an heater temperature rise AT of 238 °C. The electrical sensitivity is then 2.5 mV/g.

A low power interface circuit for resistive sensors with digital offset compensation

This paper presents an innovative conditioning and read-out interface for resistive MEMS sensors. The proposed structure includes a digital offset compensation for robustness to process and temperature variations. Simulation results demonstrate an impressive resolution to power consumption ratio and a good immunity to environmental parameters. Experimental results finally demonstrate the efficiency of this promising read-out architecture.

A Fully Integrated Inertial Measurement Unit: Application to Attitude and Heading Determination

This paper introduces the monolithic association of five sensors in an Inertial Measurement Unit (IMU). Two thermal accelerometers, two piezoresistive magnetometers, and one seismic accelerometer have been co-integrated using two different industrial processes thus demonstrating the versatility of the proposed solution. The first prototype is a System in Package (SiP) with signal processing on a separate die, while the second prototype is a full-CMOS System on Chip (SoC). The proposed architecture is detailed including the front-end electronics which is optimized to address noise and power consumption issues. Based on characterization results, good resolutions are obtained for each sensing axis. Using these results, the IMU is evaluated in the context of an Attitude and Heading Reference System (AHRS). Results demonstrate the suitability of the proposed five axis architecture, for full 3D orientation determination, with acceptable resolutions.

A novel method for test and calibration of capacitive accelerometers with a fully electrical setup

Test and calibration cost is a bottleneck to reduce the overall production cost of MEMS sensors. One main reason is the cost of generating non electrical test stimuli. Hence, replacing the functional multi-domain test equipments with electrical ones arouses interest. The focus of this paper is on sensitivity testing and calibration through fully electrical measurements. A new method based on analytical expressions of the sensitivity with respect to both physical parameters of the structure and electrical test parameters is proposed The accuracy of the method is evaluated by mean of a high level model including global and intra-die mismatch variations. It is shown that an accurate estimation of the sensitivity can be achieved using only electrical measurements and that the dispersions on the sensitivity can be divided by about 7 after the calibration procedure. These results are promising enough for high volume production of low-cost sensors.

A dependable microelectronic peptide synthesizer using electrode data

The research in the area of microelectronic fluidic devices for biomedical applications is rapidly growing. As faults in these devices can have serious personal implications, a system is presented which includes fault tolerance with respect to the synthesized biomaterials (peptides). It can employ presence and purity detection of peptide droplets via current (charge) tests of control electrodes or impedance (phase) measurements using direct sensing electrodes near the peptide collector area. The commercial multielectrode array performs better in pure and impure detection of peptides in impedance and phase. Our two-electrode X-MEF case shows slightly poorer results. In both cases the phase is the best choice for contents detection. If there are presence or purity problems, the location is marked, and repeated peptide synthesis at another collector site is initiated.

Investigation of Pt/Ti bilayer on SiNx/Si substrates for thermal sensor applications

Pt/Ti thin films on SiNx/Si substrates have been investigated for thermal sensor applications on SiNx membrane. Therefore, Pt/Ti adhesion during KOH etching of silicon and high temperature coefficient of resistance (TCR) are the principal aims of this study. ac sputtering and electron beam evaporation have been investigated for metal deposition. Vacuum annealing is used to improve the Pt/Ti characteristics. Stress characterizations and adhesion strength are evaluated by an x-ray diffraction pattern and adhesive tape test, respectively. TCR and resistivity were finally measured to confirm the compatibility with good thermal sensor sensitivity. Pt/Ti films elaborated by electron beam evaporation and vacuum annealed present the best characteristics for thermal sensor applications: good adhesion is obtained even after 5 h in KOH etching solution, electrical resistivity is about 15 μΩ cm and TCR is 3.3×10−3/°C. Finally, by using these platinum thin films, a thermal accelerometer has been manufactured and tilt ...

A Behavioral Model of MEMS Convective Accelerometers for the Evaluation of Design and Calibration Strategies at System Level

This paper presents a behavioral model that can be used to improve the manufacturability of systems based on MEMS convective sensors. This model permits to handle faults related to process scattering, taking into account not only the electrical and lateral geometrical parameters but also the influence of the cavity depth. Moreover correlations between conductive and convective phenomena are included. The model is validated with respect to FEM simulations and a very good agreement is obtained between the behavioral model and FEM results. The proposed model can then be used in system-level simulations, for instance to evaluate the impact of process scattering on the performances of the sensing part and/or to investigate different design and calibration strategies with respect to the system robustness.

Sensitivity and power modeling of CMOS mems single axis convective accelerometers

In this paper, we present 3D finite element modeling and simulation of a CMOS MEMS single axis convective accelerometer. We describe the sensor architecture and present a sensor geometry model to be used in 3D FEM simulations. Differences between 3D and previously published 2D simulations are discussed. This work investigates 3D effects which give the opportunity to better predict not only sensor sensitivity but also power dissipation. Experimental sensitivity values and 3D FEM ones are compared for two different sensor geometries and two different heater temperatures. For a prototype having a heater-cavity border distance of 340?m and a heater length of 230?m, maximum sensitivity point is obtained for detectors localized at a distance of 125?m from heater center. This distance should be moved to 90?m if a 50?m heater length is used. So, detectors should be placed closer to the heater than the usually used mid distance. Moreover, optimal detectors location shifts closer to the heater as heater length shrinks. We also show that if heater length is reduced by 80% (from 230 to 50?m), then both electrical power and sensitivity decrease by 63% and 55%, respectively. So, best efficiency is obtained for shorter heaters. In addition, detectors length decrease is found to have a significant effect on sensitivity, with an increase of 58% and 87% using heater lengths of 230?m and 50?m, respectively. Here, detectors length decreased from the total side bridge length to a fraction of this length equals to 2.5%. Optimal length is obtained when detectors are implemented on the same side bridge fraction as that used to implement the heater on the central bridge.

A Comparative Study of Conditioning Architectures for Convective Accelerometers

In this paper, we compare three different conditioning and readout electronics for CMOS convective accelerometers. The work is based on both characterization results and high-level simulations. The three following architectures are evaluated: (i) a simple amplifier, (ii) a chopper stabilized amplifier, and (iii) an innovative 1st-order thermal sigma-delta modulator. Experimental data are obtained using the hybrid combination of a 0.8 mum CMOS integrated sensor and discrete electronics. Behavioral simulations were carried-out under Matlabreg/Simulinkreg using small-signal models. The sensor model includes performance-limiting thermal phenomena and 1/f noise contributions. Previous studies describing the modeling of thermal sigma-delta modulators do not address device noise modeling, since most of the time, the performance of those modulators are limited by quantization noise. In our case, we show that the performance is limited by both the noise in the devices and the quantization noise.