*Real-Time Tracking of Individual Droplets in Multiphase Microfluidics DOI: http://dx.doi.org/10.5772/intechopen.106796*

is achieved by elaborate patterning techniques such as micro-lithography by surface patterning and thin film deposition. Next to sensor dimensions, the compatibility of the sensors' substrate for microfluidic integration has to be given. Most preferably, substrates with hydroxyl groups (like silicon or glass) on their surface are preferred since channels can be easily sealed on. Next to rigid substrates, further functionalities are granted via alteration to flexible substrates. In the following paragraphs, several representative examples of MR sensor integration in droplet-based microfluidics are depicted.

The first conjunction of magnetism and droplet-based microfluidics using spin valve [82] and GMR [83] sensors in microfluidic channels were established by Lin and coworkers in 2015 allowing tracking of ferrofluid nanoliter droplets with various concentrations (5 mg/mL – 7.5 mg/mL) and lengths (150–750 μm). Here, a sustainable application depicts microfluidic coding and decoding assays due to high sensitivities of MR sensors to passing ferrofluid droplets. The μTAS setup contained an encoding area (droplet formation), encoded droplet pool as well as decoding area (GMR sensor platform) [83]. To further demonstrate the potential in the biotechnological and medical context, ferrofluid droplets of different concentrations were mixed with fluorescent dyes, specific for penicillin giving rise to a multidimensional microfluidic barcode (magnetic and optical). (**Figure 9a**). This system was subsequently improved in terms of information output by introduction of different concentrations of ferrofluids, thus generating binary coding signals with droplet chains of different concentrations of ferrofluid. The principle was firstly demonstrated in 2016 [84], a droplet-based micro-magnetofluidic μTAS system was developed facilitating the

#### **Figure 9.**

*Tracking and applications of droplets in microfluidics using MR sensors (a) droplet-based magnetofluidic platform with magnetic decoding using GMR sensors. Magnetic decoding (right side, M) was coupled with fluorescent decoding using fluorophore-coupled penicillin (right side, P) [83] (b) droplet-based micromagnetofluidic μTAS system for generation of microfluidic ferrofluid droplet codes with downstream decoding region (GMR platform). Codes consisted of four adjacent ferrofluid droplets and two ferrofluid concentrations [84]. (c) Improvement of (de-)coding platform using ferrofluid droplets and GMR sensors by determination of the code generation, code starting point and final readout. Adapted with permission from ref. [85].*

formation of binary droplet codes using four adjacent ferrofluid droplets via two different ferrofluid concentrations (**Figure 9b**). Downstream of the verification sensor, a GMR sensor, the droplets were further polymerized using a standard alginate and Ca2+ polymerization reaction to maintain their code for long time storage. Finally, to tackle the challenge of decoding, code spacing and interpretation, Wong et al. developed a complementary system allowing precise determination of the code generation, code starting point via introduction of a fluorescent marker droplet at each code beginning and a final readout (**Figure 9c**) [85]. In the spirit of ultra-sensitive highthroughput detection in micro- and millifluidics, state-of-the-art biosensors based on micro- and nanomaterials lead inevitably to big amounts of data points. Here, the trend of the analysis shifts from manual to automatic methods to meet the challenges of complex signal identification and interpretation. In this regard, machine-learning algorithms (MLAs) are starting to play a major role. To optimize the data interpretations from Lin's work of multiparametric detection and characterization of droplets [83], a MLA based on supervised discriminant analysis was developed. Another example is the work of Schütt and coworkers in 2020 analyzed by an MLA based on unsupervised k-means clustering algorithm [81].

In general, rigid substrates like silicon wafers or glass slides are preferred as substrates for microfluidic μTAS, since channel materials can be easily integrated onto these substrates. However, the transfer to flexible polymer-based substrates offers several advantages, for instance an increase of flexibility and efficiency in (bio-)detection. Furthermore, flexible substrate materials potentially lower material costs and weight compared to silicon-based substrates. Weight reduction helps to lower transportation costs, which is perfectly in line with the spirit of point-of-care systems. The utilization of polymer materials greatly increases the possibilities to fabricate biocompatible biosensors for potential applications in vivo. The fusion of magnetic detection technologies was transferred to flexible substrates by Lin and coworkers (**Figure 10**) [86]. Here, high performance GMR sensors were integrated between two flexible polymer layers patterned with a microfluidic channel system to create ferrofluid droplets, guided over the sensing structure. While the limit of detection was found at around 4 mg/mL, thereby allowing multiparametric detection of magnetic contents and droplet sizes, the whole device

#### **Figure 10.**

*Utilization of flexible substrates: Photographs for the droplet-based micro-magnetofluidic platform with integrated flexible GMR sensors. Adapted with permission from ref. [86].*

## *Real-Time Tracking of Individual Droplets in Multiphase Microfluidics DOI: http://dx.doi.org/10.5772/intechopen.106796*

could be bent to a radius of 2 mm, maintaining full functionality and performance. The utilization of inexpensive materials give rise to development of μTAS even in resourcelimited environment and increases the possibilities of applications, *e.g.* wearable health monitoring, point-of-care testing and implantable solutions by increased flexibility in material utilization with respect to biocompability. As mentioned above, the alternative route of flexible substrates was also conceptionally explored in tubing-based millifluidics [79] demonstrating advantages of isotropic detection in fluidic channels. While the integration in tubing-based systems is rather simple via wrapping the flexible substrate around the tubing, the approach in microfluidics is more complex.

### **4.3 MR sensors in droplet-based nanofluidics**

The research field of nanofluidics describes the miniaturization of microfluidics to the nanometer regime, dealing with channel sizes typically below 1 μm. In the last decade, droplet-based nanofluidics received high interest, since droplet volumes in these channels can be scaled down to even atto- to femtoliter volumes. These ultrasmall volumes allow capture, isolation and synthesis of single molecules and are of high interest to understand biological or chemical processes, *e.g*. in enzymatic and kinetic activities. Common nanofluidics channel material is, as for microfluidics, PDMS where the typical droplet length in these channels is located from 2 μm – 3 μm. Leman and coworkers [87] demonstrated the implementation of a PCR reaction in femtoliter droplets in nano-channels while the volumes can be further reduced to stable attoliter droplets in nanofluidics channels [88] (**Figure 11a**).

To our knowledge, MR sensor have not yet implemented into these ultra-small channels. However, they have potential for analysis of the attoliter-droplets due to their ultra-high sensing capabilities in the pico-Tesla (pT) regime. The smallest measured droplet size until now, depict picoliter droplets of ferrofluid in PDMS-based microfluidic channels. Firstly detected by Tondra et al. by integration of GMR sensors into microfluidic channels at small cross-sections (13 × 18 μm) [89]. The sensor platform where 4 spin-valve GMR sensors (20 × 4 μm2 ) with 2 sensing GMRs inside and 2 reference GMRs outside the channel in a Wheatstone bridge configuration (**Figure 11b**). For integration of MR sensors in nanometer-sized channels, drastic downscaling of sensors has to be achieved at or below the standard sizes of the droplets in the single μm regime.

#### **Figure 11.**

*(a) Conceptual scheme of a droplet generating T-junction in nanochannels. Adapted with permission from ref. [88]. (b) Integration of GMR sensors into microfluidic channels (13 × 18 μm). The droplet size was located in the picoliter regime. Adapted with permission from ref. [89].*

Still, since the droplets are in the micrometer regime, elaborate patterning techniques using optical lithography are suitable. In future, this conjunction can greatly affect the analysis of ultra-small volumes of liquids.
