#4.1 Integration and testing of a SiC suspended wire on NACA airfoils
This section details the first tests that have been done to assess the SiC heater response to real flow.
At the time, the appropriate CTA circuit has not yet been developed, so the measurements were mostly performed in CC mode, using a Keithley Model 2450 SMU as the current source and a DEWESoft Sirius DAQ.
#4.1.1 Devices
The devices tested here are those whose fabrication was detailed in section…
As a quick recap, they are single-wire, suspended SiC heaters fabricated from the SiC/Si/glass stack produced by anodic bonding.
#4.1.2 TCR measurement and Joule-heating characterization
The work presented in the following paragraphs has lead to the production of a conference paper in IEEE Sensors 2024 (Kern et al. 2024)116.
The sensors were characterized in a temperature-controlled probing station, in order to measure the thermoresistive behavior of the SiC layer. Measurements were performed on a 24-bit NI PXIe-4142 source measure unit. The material resistivity at room temperature was measured at 2.04 × 10 -2 Ω·cm. A 100 µA measure current was supplied to the wire, and the voltage was measured for hot-plate temperatures varying from 30 °C to 500 °C. This enabled to measure the resistance change vs. temperature characteristic, which gave the TCR of the material (results shown in figure 2.21). Then the wire was Joule-heated with currents ranging to 100 µA to 1.2 mA, in order to determine the overheat (OH –wire temperature minus ambient) vs. input power characteristic.
The fabrication process detailed in sec ? enabled the suppression of current leakage through the substrate so that electric characterizations of the SiC wire itself could be conducted.
The resistance vs. temperature characteristic in figure 4.1 shows two regimes: one extrinsic domain where the resistance decreases due to dopant activation, and one metal-like domain where electron scattering in the lattice become prevalent, increasing the resistance. The measurements show TCR values up to 750 ppm/K. The SiC being n-doped, its resistivity \rho can be expressed relative to the negative carrier concentration n and mobility \mu_e, as follows: \rho = 1/(qn\mu_e) (Kittel 2018)117. The concentration evolves with T^{-\alpha}\cdot\text{exp}\left(-\frac{E_\text d}{k_\text B T}\right) where E_\text d is the dopant energy and k_\text B the Boltzmann constant, and the mobility evolves with T^{3/2} (Dinh et al. 2017)118. Thus the resistivity dependence in temperature
\frac{\rho}{\rho_0} \sim T^{\alpha - \frac{3}{2}} \cdot \text{exp}\frac{E_\text d}{k_\text B T},
which is fitted to the experimental data (figure 4.2) in extrinsic regime. This gives an estimation of the dopant energy E_\text d = 14.14 meV, which is within the correct order of magnitude for this type of semiconductor (Dinh et al. 2015)106.
Resistance and TCR evolution relative to the temperature for a 100 × 5 × 0.8 µm suspended 3C–SiC wire. An extrinsic domain (in green) appears, followed by an metal-like domain (in yellow).
Corresponding Arrhenius plot of the relative resistivity change vs. temperature.
#4.1.3 Integration on a NACA 0018 airfoil on a tabletop wind tunnel
#4.1.3.1 Integration and experimental setup
The sensor chip was attached and wire-bonded onto a custom-machined PCB, which was adjusted in a 14 cm long, 8 cm wide symmetrical NACA 0018 airfoil to achieve maximum possible flushness, as shown in figure 4.3. This setup was installed in the tabletop wind tunnel with a cylindrical chamber of diameter 10 cm. The wind speed setup could be varied from 0 to 15 m/s. The sensor was operated on CC mode by a stable 1 mA current source, supplying the wire with 16 mW power, overheating to 650 °C. The voltage, which gives an image of the flow velocity, was measured by a DAQ with a 25 kHz sampling rate. Measurements were also performed with a CTA controller, supplying an overheat power of 10 mW (370 °C). To monitor the free-stream velocity and give a reference measure, a Dantec hot-wire probe driven by a MiniCTA controller with a 250 °C overheat has also been placed in the wind tunnel.
Sensor chip integration on a NACA 0018 airfoil. Electrical contacts are taken from the sensor chip with wire bonding to a custom-machined integration PCB. Drawing not to scale.
#4.1.3.2 Results and conclusion
Sensor response to air flow in CC operating mode. Comparison with a Dantec miniCTA hot wire.
The results figure 4.4 show the SiC wires response to wind speeds up to 15 m/s in a tabletop wind tunnel. Due to the flush-mounting of the wire (figure 4.3), the measured voltage gives an image of the wall shear stress (Haritonidis 1989)119, generated by the viscous boundary layer around the airfoil, which explains the allure and magnitude differences when comparing to the signal of the Dantec probe. The latter is plunged in the free-stream, thus responding to the wind speed with King’s law (King 1914)7.
This work shows proof of concept of a fabrication process of suspended SiC wires, involving heteroepitaxy of 3C–SiC on Si, mechanical thinning and polishing the Si backside and anodic bonding onto a glass substrate. The under-etching of Si to its entire thickness ensured the lack of current losses through the substrate, and enabled proper SiC electrical characterizations. The resistivity behavior relative to temperatures up to 500 °C has been investigated, giving TCR measurements up to 750 ppm/K. The devices were wire-bonded and mounted on a NACA 0018 airfoil, and tested under flow up to 15 m/s in a tabletop wind tunnel. The preliminary results show the air flow sensitivity of a thermal anemometer made of a 3C–SiC suspended wire, flush mounted in a wall shear stress measurement configuration.
Further developments need to be done in characterizing these devices under well-known flows in proper wind tunnels, and testing their behavior under hot and aggressive atmospheres to demonstrate their usability in harsh environments.
#4.1.4 Integration on a NACA 0018 airfoil stall test on an open-jet wind tunnel
In order to further investigate the ability of the SiC hot wires to perform in real flow conditions, another experiment was set up by placing the same devices on the top side of a NACA 0015 airfoil, placed in an open-jet wind tunnel equipped with a lift-drag balance. The goal was to compare the sensor measurements to the lift and drag coefficient and check if the stall is visible to the sensor.
#4.1.4.1 Integration and experimental setup
The devices and integration are the same than these presented
above, as well as the power/measurement setups. The wind tunnel
is an open-jetOpen-jet means that the test section has no
walls. The flow is accelerated in the convergent nozzle, then is
“thrown” into the room, and catched by a downstream
nozzle.
installation in Centrale Lille’s buildings mainly
aimed for teaching purposes. The square convergent nozzle
measures 40 × 40 cm and the top speed is 50 m/s.
The NACA airfoil was made with stacked laser-cut acrylic sheets. A wool thread was attached on the top side to visually notice airfoil stall. The sensors were placed on the foil sagittal plane, at x = 0.25c, c being the foil’s chord. The foil was placed on a rotating axle, attached to a lift-drag balance enabling to measure the lift coefficient
c_\ell = \frac{2f_\ell}{\rho u^2 S},
where f_\ell is the measured lift force, \rho is the air’s density, u the flow speed and S the surface area obtained by multiplication of the foil’s span and chord.
Figure
4.5.
NACA 0018 airfoil equipped with sensors, in
attached and stalled flow state.
|
|
#4.1.4.2 Results and discussion
As expected, the signal caught by the SiC sensor is
SiC hot-wire voltage and lift coefficient vs. angle of attack.