MEMS (Micro-Electro-Mechanical System) flow sensors based on a thermal principle allow detection of extremely small fluid flow rates with high accuracy and resolution [1]. Recently, considerable research effort is being spent to reduce the power consumption of these devices [2], following the requirements dictated by battery-powered platforms. In this work, we propose a thermal flow sensor, with integrated readout interface, designed to obtain a very low power consumption while maintaining a high dynamic range (DR), defined as the ratio between the maximum and minimum detectable flow. The device has been obtained by applying a post-processing micromachining procedure to chips fabricated with the STMicroelectronics BCD6s process. A photograph of the chip area is shown in Fig. 1, where the main blocks used in this work are indicated. The sensing structures are differential micro-calorimeters, consisting of two n-polysilicon/p-polysilicon thermopiles (4 mV/K sensitivity) symmetrically placed across two p-polysilicon resistive heaters (2 kW each). The sensor elements are placed on SiO2 membranes suspended over a cavity etched into the substrate in the post-processing phase. The heater driver feeds the heaters with two currents, whose differential component can be digitally tuned (10-bit resolution) to implement drift-free cancellation of the sensor offset [3]. The thermopile differential output voltage is amplified (gain=200) by the integrated low-noise, low-power chopper amplifier (In-Amp and oscillator blocks in Fig.1). A set of digital registers, which can be accessed by an embedded serial port, control the interface parameters. Selection of the sensors present on the chips occurs through an analog multiplexer. The sensing structure used in this work, indicated in Fig. 1, is optimized for low power consumption. This is obtained by reducing the thickness of the SiO2 membranes with respect to the total dielectric stack of the process, using an improved sensor design and post-processing approach with respect to the device described in Ref. [4]. In particular, the membranes have been defined with the second metal layer (Metal 2) instead of photoresist (Fig. 2a), exploiting the selectivity of the SiO2 etch in CF4 plasma towards aluminum. In this way, the metal mask is aligned during the chip design and all the dielectric layers above the Metal 2 are removed during the SiO2 etch (Fig. 2b) reducing the thickness of the suspended dielectric membranes with clear advantages in terms of thermal insulation. After the dielectric etch and the metal mask removal, the silicon substrate has been anisotropically etched in a TMAH solution (Fig. 2c). The chips are glued to ceramic DIL 28 packages and a PMMA (Poly-methyl-methacrylate) gas conveyor [4] is applied to the chip surface, obtaining the structure shown in Fig. 3. Note that the sensing structure is included into a flow channel with a 0.5 × 0.5 mm2 cross section. The response of the sensor to a nitrogen flow is shown in Fig. 4. The sensitivity is 13.2 mV/sccm, while the measured peak-to-peak output noise (over a 10 Hz bandwidth) is 0.2 mV. From these data, a resolution of 0.015 sccm (1 mm/s gas velocity) can be estimated. Considering the range of ± 100 sccm, the DR is nearly 82 dB. The heater current was set to around 0.3 mA, with a small differential component, applied to reduce the output offset to the same level as the output noise. The total current absorption of the chip, including the interface supply current, is 0.8 mA, which, at a supply voltage of 3.3 V, corresponds to a power consumption of 2.6 mW. The resolution by power-consumption product is 2.6 mW×mm/s, nearly three time higher (i.e. worse) than the sensors in [2], where, on the other hand, the amplifier supply current was not taken into account and a much smaller DR (26 dB) is reported.

Integrated smart gas flow sensor with 2.6 mW total power consumption and 80 dB dynamic range

Massimo Piotto
Primo
;
Francesco Del Cesta
Secondo
;
Paolo Bruschi
Ultimo
2015-01-01

Abstract

MEMS (Micro-Electro-Mechanical System) flow sensors based on a thermal principle allow detection of extremely small fluid flow rates with high accuracy and resolution [1]. Recently, considerable research effort is being spent to reduce the power consumption of these devices [2], following the requirements dictated by battery-powered platforms. In this work, we propose a thermal flow sensor, with integrated readout interface, designed to obtain a very low power consumption while maintaining a high dynamic range (DR), defined as the ratio between the maximum and minimum detectable flow. The device has been obtained by applying a post-processing micromachining procedure to chips fabricated with the STMicroelectronics BCD6s process. A photograph of the chip area is shown in Fig. 1, where the main blocks used in this work are indicated. The sensing structures are differential micro-calorimeters, consisting of two n-polysilicon/p-polysilicon thermopiles (4 mV/K sensitivity) symmetrically placed across two p-polysilicon resistive heaters (2 kW each). The sensor elements are placed on SiO2 membranes suspended over a cavity etched into the substrate in the post-processing phase. The heater driver feeds the heaters with two currents, whose differential component can be digitally tuned (10-bit resolution) to implement drift-free cancellation of the sensor offset [3]. The thermopile differential output voltage is amplified (gain=200) by the integrated low-noise, low-power chopper amplifier (In-Amp and oscillator blocks in Fig.1). A set of digital registers, which can be accessed by an embedded serial port, control the interface parameters. Selection of the sensors present on the chips occurs through an analog multiplexer. The sensing structure used in this work, indicated in Fig. 1, is optimized for low power consumption. This is obtained by reducing the thickness of the SiO2 membranes with respect to the total dielectric stack of the process, using an improved sensor design and post-processing approach with respect to the device described in Ref. [4]. In particular, the membranes have been defined with the second metal layer (Metal 2) instead of photoresist (Fig. 2a), exploiting the selectivity of the SiO2 etch in CF4 plasma towards aluminum. In this way, the metal mask is aligned during the chip design and all the dielectric layers above the Metal 2 are removed during the SiO2 etch (Fig. 2b) reducing the thickness of the suspended dielectric membranes with clear advantages in terms of thermal insulation. After the dielectric etch and the metal mask removal, the silicon substrate has been anisotropically etched in a TMAH solution (Fig. 2c). The chips are glued to ceramic DIL 28 packages and a PMMA (Poly-methyl-methacrylate) gas conveyor [4] is applied to the chip surface, obtaining the structure shown in Fig. 3. Note that the sensing structure is included into a flow channel with a 0.5 × 0.5 mm2 cross section. The response of the sensor to a nitrogen flow is shown in Fig. 4. The sensitivity is 13.2 mV/sccm, while the measured peak-to-peak output noise (over a 10 Hz bandwidth) is 0.2 mV. From these data, a resolution of 0.015 sccm (1 mm/s gas velocity) can be estimated. Considering the range of ± 100 sccm, the DR is nearly 82 dB. The heater current was set to around 0.3 mA, with a small differential component, applied to reduce the output offset to the same level as the output noise. The total current absorption of the chip, including the interface supply current, is 0.8 mA, which, at a supply voltage of 3.3 V, corresponds to a power consumption of 2.6 mW. The resolution by power-consumption product is 2.6 mW×mm/s, nearly three time higher (i.e. worse) than the sensors in [2], where, on the other hand, the amplifier supply current was not taken into account and a much smaller DR (26 dB) is reported.
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11568/891916
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact