*2.4.4 Optical microfluidic flow sensing*

*Advances in Microfluidics and Nanofluids*

microfluid subject to measurement.

*2.4.3 Microwave microfluidic flow sensing*

sensor is to utilize a cantilever or a diaphragm [66]. The mechanical deflection can be read out with an optical microscope or photodiode. Another approach to measuring the deflection is to utilize the piezoresistive or piezoelectric elements embedded at the positions where maximal deformation could occur at the designed cantilever or diaphragm. To increase the measurement sensitivity, the Fabry Perot spectrum's fringe shift was used to measure the cantilever movement correlated flowrate, which, however, complicated the data acquisition and limited the package options [67]. The materials used to make the micro-cantilevers are silicon nitride, SU8, and polydimethylsiloxane (PDMS). An integrated micro-cantilever inside the microfluidic channel via the microfluidic favorable PDMS process achieved a capability of detecting 200 μL/min flowrate but only have a small 5:1 dynamic range [68]. Most of the micro-cantilevers measure microliter per minute flowrate, even though nanoliter per minutes sensitivity was reported, but the required optical readout often makes the fine readings and subsequent digitization a challenge [69]. While piezoresistive or piezoelectric configuration is more preferred as no optical assistance in readout will be needed. On the other hand, as the piezoelectric cannot detect a static flow, piezoresistive is considered a better choice. The cantilever sensors are more sensitive than diaphragm sensors, but there are still concerns for their reliability and repeatability per the moving cantilever. The sensitivity of these sensors also requires meaningful pressure or critical mass to activate the deformation of the cantilever or diaphragm. Such pressure is not necessarily existing in the

The non-invasive approach is always preferred in microfluidic applications, for which life science is the major focus. A microwave microfluidic flow sensor is reported [70] to achieve a large dynamic range of 1-300 μL/min with a high resolution of 1 μL/min. The detection of the flowrate with the microwave is via the measurement of a membrane that was a part of a microfluidic channel and on which the fluid is flowing over, causing the deflection of the membrane. Therefore, it could also be a type of differential pressure sensing. The measurement element is the microwave resonator that detects the effective capacitance because of the changes in the deflected thin membrane's effective permittivity due to the channel pressure changes by the flowing fluid at different flowrates. This configuration is much easier to be packaged with the microfluidic channel, and it is a true noncontact detection that can be miniaturized compared to the optical assisted readout. The microwave flow sensor is consisting of two critical components. One is the microfluidic channel with the membrane that was micromachined with PDMS soft lithography and replica molding. PDMS is a preferred material for microfluidics for its compatibility, and more importantly, it is transparent to microwave with a low loss. The membrane is about 1.5 to 3 mm in diameter and 100 μm in thickness, strong enough to hold the fluidic pressure inside the microfluidic channel. Simultaneously, it is thin enough for the sensitivity of the resonator function needed for the measurements. The second component of the flow sensor is the microwave resonator, designed into an open-ended half-wavelength ring resonator with a microstrip structure on a high-performance microwave substrate made of a 35 μm copper layer on top and bottom surfaces. The resonator operated at a 4 GHz resonant frequency. The fabrication is via the cost-effective conventional printed circuitry board processing. However, the integration with the microchannel made strong application dependence and difficulties in controlling the cost. Also, the metrological performance of this sensor

**66**

was not well documented.

Optical flow sensing is attractive to the microfluidic application for its noninvasive and high accuracy features. Laboratory flow measurements such as particle image velocimetry, infrared thermal velocimetry, and laser interferometry are reported for microfluidic metrology studies [71–74]. These optical technologies are all having complicated bulk settings and require the microfluidic channels to be optically transparent. While the miniaturization efforts continue to focus on microfluidics, optofluidics is now a dedicated field for the studies of the combined optics and microfluidics with targeted miniaturized optical integration sensing functions into a single microfluidic chip. In a microscale optical flow sensor report, [75] an optical fiber structure was fabricated in the form of a drag force cantilever to measure the microfluidic flow. A stripped single-mode optical fiber was positioned across a microfluidic channel and aligned with a multi-mode fiber receiver. The microfluidic flow in the perpendicular directions will displace the fiber cantilever tip, causing the light intensity change at the aligned receiver. The reported sensor had achieved a measurement dynamic range over 60:1 and a minimal detection of 7 μL/min. However, the making of the sensor would be quite complicated with the fiber alignment, and direct contact of the flowing fluid with the fiber cantilever is also required. In another report using the optical approach for flow sensing, miniaturized fluorescence sensing is attempted for micro molecular tagging velocimetry in microfluidics [76], but these methods are not cost-effective and yet to reach the small footprint.

In an alternative optical sensing approach, [77] a collimated light beam was employed to excite the surface plasmon resonance at a gold film on top of a polymethyl methacrylate (PMMA) microfluidic channel. The fluidic flow will cause the temperature redistribution inside the microfluidic channel, which alters the refractive index above the metal film. The refractive index is inversely proportional to the temperature. By acquiring and analyzing the image of the excited surface plasmon, the flowrate could be measured. However, since surface plasmon resonance is very sensitive to temperature, and the response is nonlinear, a full functional measurement scheme and affirmation of metrology parameters will need additional efforts.
