Micro/Nano gas sensors: a new strategy towards in-situ wafer-level fabrication of high-performance gas sensing chips

Abstract

Nano-structured gas sensing materials, in particular nanoparticles, nanotubes, and nanowires, enable high sensitivity at a ppb level for gas sensors. For practical applications, it is highly desirable to be able to manufacture such gas sensors in batch and at low cost. We present here a strategy of in-situ wafer-level fabrication of the high-performance micro/nano gas sensing chips by naturally integrating microhotplatform (MHP) with nanopore array (NPA). By introducing colloidal crystal template, a wafer-level ordered homogenous SnO2 NPA is synthesized in-situ on a 4-inch MHP wafer, able to produce thousands of gas sensing units in one batch. The integration of micromachining process and nanofabrication process endues micro/nano gas sensing chips at low cost, high throughput, and with high sensitivity (down to ~20 ppb), fast response time (down to ~1 s), and low power consumption (down to ~30 mW). The proposed strategy of integrating MHP with NPA represents a versatile approach for in-situ wafer-level fabrication of high-performance micro/nano gas sensors for real industrial applications.

Figure 1. The micro/nano gas sensors and the strategy of in-situ wafer-level fabrication process.

(a) For the in-situ wafer-level fabrication, a wafer of MHP fabricated by micromachining process, a wafer of polystyrene (PS) colloidal monolayer template, and a precursor solution which is used for synthesis of gas sensing material are respectively prepared in advanced. (b) In-situ synthesis of SnO₂ NPA on the MHP. With a precursor solution of SnCl₄, SnO₂ can be obtained via hydrothermal methods in basic aqueous solution: Sn⁴⁺+4OH⁻→SnO₂+2H₂O. First, A self-organized PS spheres monolayer template is lifted off from a glass wafer and then floating on the surface of precursor solution (SnCl₄). Then, such floated monolayer is transferred to the MHP wafer by a simple picking-up process. Due to the capillary effect, the PS monolayer on the wafer also contains the precursor solution in the interstitials among PS spheres and wafer. Along with subsequent drying and annealing, the organic PS template can be removed and ordered SnO₂ NPA is thus formed on the MHP wafer. In this case, a great many MHP-NPA integrated sensors are simultaneously fabricated and arranged on the wafer. (c) Sensor chip and the active area of the micro/nano gas sensor. After the in-situ wafer-level fabrication process, the wafer is diced into massive single micro/nano gas sensing chips. The active area of the micro/nano gas sensing chip consists of five layers from bottom to up: supporting layer, Pt microheater, isolation layer, IDEs, and NPA.

Figure 2. Micro/nano gas sensors.

(a) A 4-inch wafer of MHP, chip size: 3 mm × 3 mm. (b) PS template, diameter of the PS ball is 500 nm. (c) The whole 4-inch wafer of MHP covered with a monolayer of NPA. (d) Active area of the micro/nano gas sensing chip. (e) MHP 1, spacing of fingers is 24 μm. (f) MHP 2, spacing of fingers is 18 μm. (g) MHP 3, spacing of fingers is 10 μm. (h) monolayer SnO₂ NPA. (i) double layer SnO₂ NPA. (j) triple layer SnO₂ NPA.

Figure 3. Electrothermal characterization of micro/nano gas sensors.

(a) Temperature distribution on a typical micro/nano gas sensing chip with finite element method (FEM) simulations obtained by commercial analysis software Coventor. Sensors based on MHP 1, MHP 2, and MHP 3 have been introduced in the simulation and comparison. (b) Temperature distribution along beam-active area-beam of sensors based on MHP 1, MHP 2, and MHP 3 with a power consumption of 30 mW. The inset is zoom-in of the active area. Temperature gradient on the supporting beams is high, which means that a lot of heat flows through the beams from the heated membrane (active area) to the substrate. The active area achieves a homogenous temperature distribution with an average temperature gradient of about 0.14 ^oC/μm. The temperature distribution at the active area is generally uniform. Thus, the working temperature of the sensor can be well controlled by power consumption. (c) The average temperature of the sensor versus power consumption. It can reach up to 350 ^oC at a power of 30 mW. Temperature of MHP 3 is slightly lower than that of MHP 1 and MHP 2. But no significant difference in maximum temperature was observed among the three MHPs, which is also illustrated in the simulation results shown in b. (d) The warm-up time and cool-down time of the sensors. By applying an appropriate step voltage and measuring the resistance change of the Pt microheater, the warm-up time and cool-down time can be measured. It takes less than 10 ms to heat the sensor from room temperature up to 350 ^oC and cool it down back to room temperature.

Figure 4. Sensor response to ethanol.

(a) The 3D plot of the sensitivity (S) to 1 ppm ethanol as a function of the working temperature and types of MHP. It clearly indicates that gas sensor has the highest sensitivity at a working temperature of 350 ^oC. And the sensor based on MHP 3 has higher sensitivity than other two sensors. (b) The sensitivities of sensors based on MHP 3, with monolayer, double-layer, and triple-layer of NPA. Sensors with double-layer or triple-layer of NPA have lower sensitivity and larger error than sensors with monolayer of NPA. (c) Sensor response to ethanol with three different levels of concentrations: from 20 ppb to 100 ppb, from 100 ppb to 500 ppb, and from 1 ppm to 5 ppm. The response time and recover time to ethanol are around 2 s with a concentration less than 100 ppb, and decrease to around 1 s and less than 1 s when the concentration increase to 100 ppb – 500 ppb and 1 pm – 5 ppm respectively. Sensitivity to 20 ppb ethanol is 1.06. The sensitivity (S) has a linear relation to ethanol concentration (C) via linear fitting can be denote as S = 1 + 0.0023 × C. When the concentration increases to the level of several hundred ppb, sensitivity and concentration still have a good linear relationship as S = 1 + 0.0024 × C. When the concentration of ethanol increase to ppm level, there is a nonlinear relationship between sensitivity and concentration. Sensitivity tends to reach a constant value if the concentration keeps increasing.