To create a sealed system to measure the flow rates, the chip is confined between a circuit board and a fluidic adapter. The adapter is made from polymethyl methacrylate (PMMA) and screwed onto the circuit board, thus clamping the µValve chip in-between. The bottom part of the fluidic adapter contains holes for electrical connections to the circuit board, which is established via pogo pins. The top part has fluidic inlet and outlets. A PDMS layer seals the fluidic paths from these ports to the chip (Fig. 5). The fluidic adapter also seals the filling holes used to fill the gaps between the actuators, which prevents leakage.
The measurement setup is shown in Fig. 6. The flow rate measurements are done with pressurized air. An Elveflow OB1 pressure controller regulates the applied pressure up to 300 kPa. The flow rate is measured by a Bronkhorst Mini Cori-Flow ML120V00, which is placed between the pressure controller and the chip. The outlet of the chip is left open and is exposed to the environment. The voltage to control the NED actuators is generated by a Keithley 2470 Source-Measurement-Unit (SMU). The output of the SMU is connected to an electric adapter, which can individually disconnect specific actuator groups. For these measurements, all actuators are used. The laboratory is fitted with a climate control and environment sensors, so that we can assure a controlled environment with a pressure of 101 kPa and temperature of 21°C during the measurements.
First, leakage measurements for a passive chip, i.e. when the actuators are not supplied with voltage, were carried out. The pressure was increased from 10 kPa to 300 kPa and the flow was measured. Within these measurements no mechanical damage to the actuators was detected.
The chips were electrically connected to the SMU and the voltage was increased stepwise from 0 V to 45 V and then back down to 0V, while monitoring the current. The NED actuators are electrostatic structures. Thus, low parasitic currents are expected and currents above a certain threshold (100 nA) are an indicator for malfunction. Likely, a short circuit or even defective structures cause high electrical currents. Furthermore, if the current does not increase as the voltage increases, there is a defect in the connections or bonding of the chip. Using this approach, we verified that the enclosed actuators worked as intended by ensuring, that the current is below 100 nA for voltages of up to 45 V.
For each step of the measurement, a constant fluidic pressure and voltage is applied. Before starting the measurements, the flow is allowed to stabilize for the duration of one minute. Then, the flow rate measurement is triggered, and 500 values are recorded to calculate mean and deviation. The voltage is increased stepwise starting at 0 V, then 10 V, 20 V, 30 V, 40 V, and finally 45 V. Then, the voltage is decreased in the same steps. This is done to detect any hysteresis behavior in the response of the actuators. The process is repeated for pressures ranging from 6 kPa to 75 kPa for 3 different chip and up to 200 kPa for one of these chips. While the structures could withstand pressures of up to 300 kPa without mechanical damage in the non-actuated test, these high pressures have not been further investigated, since a significant leakage is already seen at lower pressures. Since the Bronkhorst Mini Cori-Flow ML120V00 measures a mass flow, the values are converted to SCCM according to Spitzer [9] employing the following formula:
$$SCCM=6\cdot {10}^{7}\frac{Z\cdot {R}_{u}\cdot T}{p\cdot M}\cdot \dot{m}$$
With \(Z\), the compressibility factor of air at room temperature (1), \({R}_{u}\), the universal gas constant (\(\text{8,312}\frac{J}{K\cdot mol}\)), T, the environment temperature (\(294K\)), p, the environment pressure (\(101kPa\)) plus applied pressure by the pressure controller, M, the molecular weight of air (\(28.97\frac{kg}{kmol}\)), and \(\dot{m}\), the measured mass flow in \(\frac{kg}{s}\).
An overview of the measured flow rates at different voltage-pressure combinations is shown in Fig. 7 and Fig. 8. When no voltage is applied, one can see the maximum possible airflow through the narrow channels inside the µValve chip. As the voltage increases, the NED actuators move the plunger against the flow and narrows the channel. This allows regulating the flow rate based on the applied voltage. Since the deflection scales with V² due to the electrostatic nature [1], the response it not linear.
Table 1
Maximum flow rate and leakage for a given pressure. The data are given as a mean and standard deviation for 3 chips each.
Pressure [kPa]
|
Max. Flowrate
[SCCM]
|
Leakage
|
6
|
2.5 ± 0.1
|
9.1% ± 5.3%
|
10
|
6 ± 0.2
|
7.8% ± 3.2%
|
20
|
12.5 ± 0.2
|
10.7% ± 2%
|
30
|
18.2 ± 0.3
|
10.9% ± 2%
|
40
|
23.5 ± 0.6
|
11.0% ± 2.4%
|
50
|
28.7 ± 0.8
|
12.3% ± 2.2%
|
75
|
37.3 ± 0.5
|
22.2% ± 2.6%
|
The maximum flow rate and relative leakage are given in Table 1. The maximum flow rate is the flow rate at the given pressure and no applied voltage. The leakage is calculated as ratio between the maximum flow rate and the minimum flow rate at maximum applied voltage (45 V). The leakage of the valve is as low as 9%, depending on the applied pressure. It is possible to regulate the flow rate between about 10% and 100%, depending on the applied voltage. A full blockage of the flow could not be achieved and a leakage of a least about 10% does remain. The leakage is mainly caused by the plunger, since is not able to close the channel completely in the current design.
In terms of maximum flow rate, our design allows higher flow rates than the designs presented by Bae et al. and Yoshida et al. However, the leakage of our design is with above 10% higher than the reference designs. The possible operation pressure is only 30 kPa and low compared to the design of Bae et al. (126 kPa), but our design operates on voltages below 50 V instead of 140 V. Compared to the design of Yoshida et al., our design has only half the size und half the thickness. Overall, the presented valve design sits in between the values given by other designs.