**2. Microfluidic flow sensing technologies**

Microfluidic sensors are critical components for a complete system. Many research works on sensors have been dedicated to the biomedical and chemical sensing development based on electrochemical, optical, mass, or magnetic sensing principles. Electrochemical sensors are mostly studied and often composed of several electrodes that are easy to fabricate together with the microchannels. This limited integration with a simple configuration allows a fast response with reasonably good sensitivity and enables multiple reagents on a single microfluidic chip. The electrode embedded inside a microfluidic channel can also be used for cell counting and estimate the flow volume. With the assistance of a miniaturized LED, pH measurement could be achieved as well [28–31]. However, many of the proposed biosensors or chemical sensors are very specific, and most are research-orientated, as being determined by the catalytic or affinity properties of the biological recognition agent in a particular study and the sensor itself requires a sophisticated electronic system for readout or analysis. The standalone or large scale commercial applications are yet to emerge.

Flow meters using traditional thermal capillary and Coriolis measurement are commercially available for micro flow measurement before the microfluidics being intensively studied. Researches on microfluidic flow sensing approaches are for miniaturized, cost-effective, and integrable products. In this scope, both flow and pressure sensors have been extensively studied [28]. In some cases, the differential pressure sensor can be used for flow measurement. Flow measurement is one of the most important factors in microfluidic handling for data analysis and precise system control. Without the knowledge of the fluid quantity in the process, analytical results would not be easy to establish the needed and convincing statistics. The conventional flow sensors might be the first commercially available standalone

sensing products for microfluidics. The technologies are still limited, and their package formality is bulky and far off the cost target for the desired microfluidic system. Many studies proposed integrating flow sensors into the microfluidic system. However, there are still many factors that impact data acquisition. The existing sensor products on the market also have some unsolved reliability issues in applications. The commercialization route to a well-performed and cost-effective sensor is yet to be demonstrated.

The available flow sensors applied to microfluidics are classified as thermal and non-thermal sensors [1]. Thermal flow sensors have been applied to small flow measurement for both gas and liquid before the microfluidic concept emerged. Therefore thermal flow sensors are mostly studied and applied in microfluidic applications, and products with various thermal sensing principles are commercially available. Coriolis microfluidic sensor is a non-thermal sensor, and it has an even higher cost. Other "non-thermal" flow sensors are mostly at the research stages. Before the form factor, cost, and reliability issues can be solved, large scale applications are still not possible.

### **2.1 References and standards**

For the traditional flow sensors, the metrology characteristics will hardly enable a self-calibration. Therefore, a primary standard or a reference defined by an international norm governs the manufacture of a flow sensing product with specific sensing technology. The same should then apply to microfluidics. Demanding to establish an international standard for microfluidics has long been proposed [32, 33]. Still, only in recent years, an international microfluidic association has been established, and an international standard (ISO) working committee has been organized with a serial of workshops [34]. It has been proposed that the new ISO standard for the microfluidic shall be having four sub-standards, including *flow control* that addresses the key components of valves, pumps, and sensors for the system; *Interfacing* that is to standardize the connectors and other interfaces; *modularity* that will regulate the integration and *testing methods* that will define the methodology of the metrology and other related testing issues. Such a task is still at an earlier stage, and additional time will be need before the standards become available.

Several efforts to establish a primary standard or a traceable reference system for flow metrology in microfluidics applications have been made in the past years [35–37]. The widely adapted primary standards are the gravimetric and volumetric principle. The comparison of such standards among different European national metrology institutes indicated an uncertainty (*k* = 2) ranging from 0.05 to 6% for the flow rate ranges of 17 nl/min to 167 ml/min. Still, most of them can have uncertainties within 0.1% [35]. In the reported reference system, the flow generation is critical and requires a stable flow system in the sub microliter per minute flow. Other special effects such as evaporation must be considered, especially when approaching nanoliter per minute flowrates. Degassing through the system and preventing external vibration, and testing environmental control are also critical to ensure the measurement is repeatable and accurate. The flow generators used in these institutes include metallic bellow, precision syringe pump, and gear pump. Because the gravimetric measurement is achieved with high precision balances, the system is a uni-directional open loop. For sub microliter per minute flowrates, laser interferometry has been used as an alternative precise reference for the desired accuracy [38]. However, for high volume applications, a faster closed-loop calibration would be preferred. It could also result in good accuracy using a gear pump and high precision Coriolis meter with an accuracy of ±0.2% as the reference standard [39].

**59**

*2.2.1 Calorimetric flow sensing*

*Microfluidic Flow Sensing Approaches*

**2.2 Thermal flow sensing**

*DOI: http://dx.doi.org/10.5772/intechopen.96096*

For almost all flowrate ranges in microfluidics, the Reynold numbers are within 1000, indicating that the flow of interests is within the laminar flow regime. Therefore in a desired large dynamic range, the flow profile would not be the same at the different flowrates, which adds complexity to maintain the measurement accuracy. Meanwhile, the flow channels are small in micrometer dimensions. The interfaces between fluid and channel wall become pronounced, which differ from those described for laminar flow by Moody Diagram in the classic fluid dynamics. Besides, cavitation would play a critical role, and dissolution will also contribute to metrology. These are among the

The thermal mass flow measurement using calorimetric capillary sensors has been used to measure a very low flow to nanoliter per minute for quite a long time [40]. The sensors are composed of thin metal wires winded outside the wall of a tiny tube of a micrometer in diameter. The tube is usually made of thermally conductive materials such as stainless steel or fused silica. These sensors normally require a higher power to ensure the heat transfer resulting in a small dynamic measurement range and a low accuracy towards the low measurement end. The required manufacture process makes these sensors very costly without being able to be volume produced. Integration of such a sensor into a microfluidic system would be unlikely. In the following discussions, only micromachined sensors will be addressed. The micromachined sensors are mostly made on silicon or glass substrate. A microheater and plural numbers of sensing elements are deposited on a membrane structure, and the air or gas-filled cavity below the membrane provides the desired thermal isolation. The tiny sensing elements enable a fast response time. The membrane is frequently made with silicon nitride or silicon nitride and oxide combination. The sensing elements can be metals with a large temperature coefficient such as platinum, nickel, tungsten, or in the case for the process compatibility, doped polycrystalline silicon is used instead. The micromachined thermal flow sensors' structure has no moving parts, and the surface can be treated with various passivation and post-process coating for better reliability. The micromachining process for the flow sensors is well established today. Most MEMS foundries have the necessary equipment for manufacturing such sensors, which allows a very favorable cost and makes it possible for high volume applications. The first micromachined thermal flow sensor made for microfluid is used in micro gas chromatography [4]. It is for gaseous flow and not a standalone product and only manufactured in a minimal quantity as the OEM product. The commercially available micromachined thermal microfluidic flow sensors for liquid were incepted in the last decade. These commercial products utilize different thermal sensing principles [41–43] that cover the three major technologies with thermal calorimetry, anemometry, and thermal time-of-flight approaches. There are some research activities on other thermal flow sensing designs, such as thermal capacitive utilizing the temperature dependence of dielectric constants, [44] and temperature dependence of the PN-junction in a diode [45]. The measurement scheme of flowrate with these alternative thermal sensing designs could also be classified into the above three thermal sensing principles. **Figure 1** is the graphic illustration of these three measurement principles for

new challenges for the on-going metrology standards for microfluidics.

the typical micromachined thermal flow sensors on a silicon substrate.

The majority of the current micromachined commercial thermal flow sensors are utilizing the calorimetric principle. Most successful applications are for gaseous
