Heat transfer is the trading of heat between physical frameworks. At the microscale, heat transfer measurements in both single-phase [16] and two-phase [17] flows have always been a matter of consideration for industrial manufacturers. Regular applicable areas in microfluidics related to heat transfer measurements are cooling applications [16,18,19,20,21,22,23,24,25,26,27] with the cooling systems needed for refrigerators, air conditioners, and microelectronics [28,29]. Research has been continuously carried out to provide better heat transfer performance to accommodate all these mentioned applications, to make them outstanding in their performance. The plan optimization of micro pin-fin shapes is also in the picture for cooling applications [30].

In the literature, productive work has been conducted in the field of refrigeration. Refrigerant flow is where the fluids are used as coolants, and these coolants are made to flow in micro pathways to collect the heat from around the refrigerator. Various fluids such as FC-72 dielectric fluid [17], R245fa [31], R134a [25,32,33,34,35,36], R32 [31,35], R410A [31], R600a [34], R290 [34], R1270 [34], R1234ze(E) [37] and R410a [38] are employed for the same. The local heat transfer coefficient can be estimated from the local wall temperature and liquid saturation temperature [39]. Different concentrations of graphene oxide particles-water nanofluids impact heat transfer for pin-fin microchannel and pulsating inlet velocity. For different Reynolds numbers (272, 407, and 544) and mass fractions (0.02%, 0.05%, 0.1%, 0.15%, 0.2%), the pulsating flow (pulsatile frequencies (0–5) Hz) has a greater impact on heat transfer at a lesser averaged Reynold’s value [40]. Microscale heat transfer is carried out using different newly fabricated microchannels [41], microchannels with branch spacing [42], heatsinks with multiple channels varying in microchannel dimensions [43], and an Interconnected microchannel Net (IMN) [41]. IMN yielded higher heat transfer than that of the rectangular microchannel. When it comes to mall distribution and instabilities in the parallel channel in a two-phase heat transfer, flow excursion between channels is the main reason for maldistribution, and maldistribution is the reason for a higher level of pressure down in two-phase heat sinks [23,44,45,46,47,48]. Heat transfer coefficient investigations are carried out under extreme operating conditions of heat flux, mass flux, and micro gap [49]. The heat transfer coefficient is strongly dependent on the mass flux [41], heat flux [49], and vapor flume quantity [49]. Heat transfer coefficient varies for both single-phase and two-phase. It also varies for gravitational aspects such as horizontal and vertical [39,50] orientations of microchannels and the direction of flows. Adding ethanol into the water upgraded heat transfer resulting in higher heat transfer coefficients than for either of its pure components [39]. Additionally, the heat transfer coefficient upgrades with an increase in the oblique angle of microchannel fins [17]. The heat transfer coefficient enhancement by applying magnetic quadrupole is decreased with increasing the Reynolds number [51]. Nanofluids flow along by application of magnetic field has come to the picture for enhancement of heat transfer coefficient, and enhancement depends on Reynolds number and magnetic field applied for Fe₃O₄/water magnetic nanofluid flow in the presence of the quadrupole magnets located at different axial installation positions [52]. Heat transfer coefficients of fluids such as R134a [49], R1234ze(E), R1234yf, and R600a are affected by vapor qualities [53]. The two-phase heat transfer coefficient of the pin-fin micro gap (after accounting for the additional surface area of the fins) is four times greater than the bare microcap under extremely high conditions of heat flux, mass flux, and the confined path of 10 µm [49]. Both microchannel geometry and microflow Reynold’s number will influence the heat transfer measurement [54].

Flow boiling is where the fluid flow is forced over a surface by external means such as pumps and buoyancy effects and is implemented as one of the most promising cooling methods. Flow boiling can be studied by using fluids such as R134a [49,55], and Hożejowska et al. developed a flow boiling heat transfer model based on the Trefftz method for better analyzing the flow boiling in rectangular, vertical, and asymmetrically heated mini channels with varying aspect ratio measurements [56]. Segmented finned channels overtake uniform cross-section microchannels with low response time and magnified heat transfer. The response time is less for inflow boiling (regimes) conditions compared to the single-phase flow of coolant [16]. The effects of applying a magnetic field on the pressure drop and heat transfer of Fe₃O₄/water magnetic nanofluid (ferrofluid flow) along the length of the tube resulted in maximum enhancements of 23.4%, 37.9%, and 48.9% in the local heat transfer coefficient [51]. Orientation of microchannel [57] and heatsinks [58] was also found to be influencing the flow boiling process. In different flow boiling investigations, the chaotic pressure oscillations flow pattern sequences relied upon the mean worth of the pressure drop [59]. Universal correlations are also created utilizing many data sets and heat transfer anticipating tools for the condensing and stream boiling process in a microchannel for annular stream and the other for slug and bubbly streams [60].

Different windows—W1, W2, W3, W4, W5, W6, W7, W8, W9, and W10—are considered for various ranges of heat flux and mass flux for investigating heat transfer measurements, as highlighted in Figure 1 and Table 1. W1 to W10, shown in Figure 1, is in the increasing order of window sizes, where W1 represents the smallest window and W10 represents the largest window explored in the literature. The W1 window, being the smallest window, is explored for the dynamics and heat transfer characteristics of the flow boiling bubble train using R134a flow, where a nonequilibrium phase change model provides a way to measure the interface temperature and heat flux jump [61]. Heat transfer differs in both horizontal and vertical directions of flow. Window W2 is checked for the measurement of heat transfer in a single 5 mm inner hydraulic diameter square channel in a vertical orientation and provides the heat transfer coefficient measurement with the uncertainty of ±20 in measuring the heat transfer coefficient [39]. In window W3, half-corrugated micro-channels with bottom sinusoidal structured surfaces affected temperature and two-phase heat transfer. Channels with higher wavelength (0.5 mm) and larger wave amplitude (0.3 mm) showed better heat transfer measurement with little pressure drop [62]. Concerning the convective heat transfer coefficient, the newly fabricated interconnected microchannel net (IMN) yielded higher heat transfer than that of the rectangular microchannel for the window W4 [41].

Window W5 is used to investigate the flow boiling of R134a in a 10 µm micro gap to assess its capability to dissipate ultra-high heat fluxes [49]. Other than Window W2, Window W6 is also utilized for the vertical flow orientation with a 1 mm diameter microchannel being aligned in gravitational aspect horizontal, vertical upward, and downward orientation with R-134a. The heat transfer coefficient and pressure drop measurements increased when the refrigerant flowed vertically downward [50]. Segmented finned channels overtake uniform cross-section microchannels with less response time and enhanced heat transfer measurements in window W7 for deionized water flowing through microchannels [16]. The flow of FC-72 with absolute pressure measurement in the range (1.16–1.84) bar flowing through rectangular, vertical, and asymmetrically heated mini channels is explored in window W8 with a high aspect ratio of 40, 20, and 13.3. The heat transfer coefficient correlation using the Trefftz method helps attain a 10.1% Mean Absolute Error (MAE) for flow boiling [56]. Window W9 is explored for the heat transfer coefficient for the flow of R134a, R1234ze(E), R1234yf, and R600a through mass velocity range (200–800) kg/m² s and vapor qualities range (0.05–0.95) for the temperatures of 31 and 41 C during flow boiling inside a circular microchannel with an internal diameter of 1.1 mm [53]. W10 is the largest window explored in heat flux and mass flux ranges for the flow of FC-72 dielectric fluid through three different microchannel fins oblique angles, such as 10, 30, and 50 [17].

Microfluidics technology is used in flow manipulation, such as mixing flows [63,64] and sorting particles [65,66,67,68], based on flow parameter measurements at the micro level. Work conducted in these scopes is provided below.

Mixing time, length, throughput, and mixing index/mixing efficiency are the different parameter measurements that were found to influence microfluidic blending for varying microfluidic techniques. Table 2 provides the idea of efficiency parameters achieved for various efficiency boosting parameters, along with highest mixing efficiency achieved and time taken for the same. The pore array tube in a tube microchannel reactor allows intense micromixing performance, high throughput, and optimal micromixing time [69,70]. When studies were carried out for spiral channels, higher aspect ratio measurements showed better mixing, and higher Reynolds numbers showed the least mixing efficiency. Frequent curving of the path lines is additionally recognized to influence blending and prompts a reliance on blending on aspect ratio. The highest mixing efficiency for the spiral microchannel with 25 × 50 × 800 (1,000,000) grid cells is 90.56%. Furthermore, 99.5% is the highest mixing efficiency for the variation of Reynolds number at Re = 140 while mixing uncertainty is ±10% [63]. Almost 100% mixing efficiency is achieved at a mixing time of 167 ms [63]. In a further study, implementing a soft microchannel wall of width 0.5 mm and a height of 35 μm resulted in ultra-quick blending for Reynolds numbers as small as 226. Complete cross-stream blending is accomplished within 10 ms. Hence, the blending time becomes littler by a factor of 100,000 in contrast with that required for diffusive blending (250 s) [64]. With these parameter measurements, which are essential in serving microfluidic blending, chemical synthesis is one of the applications that employs the mixing of flows [64].

Sorting is isolating the same kind of particles in the fluid flows. Sorting is implemented with the flow of nanoparticles and the application of an external magnetic field, by adding a sheath or by pumping fluids with different flow rates, etc. Microchannel designs and their measurements influence particle migration and separation behavior [71]. Inertial focusing on planar layouts has also been investigated, which favors sorting [72]. Various segregation processes provide different measurements of accuracy, error, selectivity, throughput [73], separation efficiency [74], recovery [74], enrichment ratio [74], and purity [73]. Many research works are conducted to realize more functional geometries, varying in measurements to reduce the extensive laborious requirement in the traditional fabrication process, with more particles focusing on efficiency [74]. The spiral microchannel becomes one of the most explored microchannel geometries used to initialize segregation. A spiral microchannel provides a new dimension to the segregation process by increasing the segregation’s efficiency, where micro-sized particles switch to multiple locations at a flow rate of 4.2 µL/h [75] while viscoelastic fluid flows [63,70,76] allow size-dependent segregation.

The sorting efficiency achieved for 1 µm, 10 µm and 20 µm size particles is presented in Table 3 and Table 4. Segregation efficiencies achieved by different methods in the literature for different particle sizes are highlighted in Figure 2. Table 4 showcases the highest segregation efficiencies achieved for varying microparticle sizes. For 1 µm, size-dependent separation [68] and crossflow micro filter layouts [77] resulted in 90% of the segregation efficiency, while inertial focusing in spiral microchannels, followed by particle deflection in the straight channel [67], resulted in 99.7%. Polystyrene particles of sizes 2 µm and 3 µm in Newtonian and viscoelastic fluids co-flows attained 90% efficiency using size-dependent separation [68]. Inertial focusing in curved channels resulted in 94.5% separation efficiency for 5 µm particles [78], while inertial focusing in spiral microchannels led to 98.3% for 5.55 µm particles. For 10 µm particles, segregation efficiency varied between 94.5%, 100%, and 100% for inertial focusing in curved channels [78], passive inertial focusing with active magnetic deflection [79], and simultaneous separation [80] methods, respectively. Inertial focusing in curved channels resulted in 94.5% separation efficiency for 15 µm particles [78]. For 20 µm particles, segregation efficiency was 100% for both passive inertial focusing with active magnetic deflection [79] and simultaneous separation [80] methods. Considering the highest segregation efficiencies for each particle size, 10 µm and 20 µm particles reached 100% efficiency.

New microchannels were fabricated to improve the segregation process. Inertial focusing is influenced by many 2D layouts such as straight and spiral square spiral channels at the flow rates 1 mL/h, 5 mL/h, 10 mL/h,15 mL/h, and 20 mL/h. The tape’n roll method helped to configure helical and double-oriented spiral channels to check the unexplored inertial migration characteristics [72].

The contribution of microfluidics in the medical field has tremendously increased in the past two decades. Presently, the research is focusing on biosensing and microfluidic integration. Advancements are happening in label-free DNA biosensors, with a specific spotlight on the combination with microfluidic structures for point-of-site detections [9]. Dissipative particle dynamics (DPD) simulation mimics dynamic and rheological characteristics of different soft matters such as red platelets, vesicles, polymers, and bio macro atoms in microchannels [81]. It is seen that flow change and platelet adhesion significantly influence one another [82]. Currently, microfluidics applications cover DNA detection [83], cellular flow manipulation [9,84], reactants synthesis, biopharmaceutical productions such as drug synthesis [83], epithelial cell adhesion [85], in-vitro applications [69,86] and in-vivo applications [86]. Additionally, microfluidic droplet applications include DNA analysis [87], DNA arraying [88], blood analysis [87], and chemical reactions [87]. Nevertheless, as stated below, significant work has been conducted in medical applications about drug delivery and sorting/isolation applications.

The application of microfluidics also lies in drug delivery for treating cancer [69] and other diseases. Nowadays, chemotherapy has received less attention because of its low efficacy and adverse side effects in treating tumors that are close to the major blood vessels of the liver [89]. Drug delivery is creating a great buzz in treating arterial diseases such as atherosclerosis, tumors, and infections, and removing blood clots without the requirement of surgery, thereby avoiding post-surgical complications and hence decreasing treatment costs.

The fabrication of nanocarriers for oral drug delivery uses different approaches involving drugs/plasmids and nanocarriers flowing from different inlets into a single microchannel, mixing both using a staggered herringbone mixer in the microchannel, microfluidic droplet generator, and microfluidic processor [86]. The flow of biocompatible nanoparticles in blood flow is controlled by a magnetic field so that they can carry the drugs to the region of the disease, such as the stenosed artery, and release the drugs at that point for treatment [90,91]. Shear stress decides the residing duration of nanoparticles at the targeted points while creating the way for these medical applications [90]. Nanoparticle concentration is a function of axial velocity and temperature. Furthermore, the concentration and size of nanoparticles increase wall heat transfer and favor drug delivery [90]. Electrokinetic force and pulsatile pressure gradient are the other factors that influence nanoparticle flow [91]. Microfluidics work with the effective and minimal expense creation of different miniature and nanoparticles is on focus. This is carried from various materials and therapeutic agents, with high stacking limit and controlled delivery at a limited scale, which limits the measure of required reagents.

The nature of synthesized nanocarriers, such as morphology, drug stacking limit, and kinetic delivery boundaries, can be effectively and successfully altered and advanced by changing the microchannel geometry and stream rate [86]. Nanotube/pH-responsive polymer composite safeguards the stacked meds from unfavorable release until their pH comes to 7.4 [92]. Another drug delivery system that falls in the category is when high-intensity focused ultrasound-induced mild hyperthermia allows drug release and acoustic streaming pushes temperature-sensitive liposome particles from the vessel wall to the target area; this results in effective drug penetration into the tissue increasing by 56% compared to conventional drug delivery approaches [89]. A carrier can also utilize its properties to enter the desired intracellular compartment and release the payload at appropriate times and conditions. Two types of intracellular delivery are carrier-based methods and membrane disruption techniques [93]. Features of an ideal intracellular drug delivery system are scalability, minimal cell perturbation, suitability to cell types, safety biocompatibility, control mechanism, and cost [93].

In the drug delivery process, microfluidic technology plays different roles at different stages, from the fabrication of nanocarriers and drug encapsulation to the drug intercellular entry.

The control of miniature and nanoparticles in complex biofluids is exceptionally requested in many biomedical applications and micro-level biological environments. Detection is carried out using biomarkers [74], while the isolation of live cells [94] and cancer cells (ex: colon cancer cells) traveling in the bloodstream in various parts of the body is handled with inertial microfluidics and the design of different microchannels [95]. For instance, microchannel geometry with a fishbone shape increased the performance of separation of three different sized (2 µm, 3 µm and 13 µm) microparticles and segregation of MCF-7 from human erythrocyte mixture, with recovery efficiency greater than 98% [96]. Compared to circular, rectangular, and trapezoidal channels, a 3D-printed river meander-such as cross-section facilitated inertial particle focusing and sorting of MDA-MB 231 cells (26 µm) in whole blood, achieving 85.4% recovery [74]. The use of curved channel walls decreases cell damage, and the introduction of fabric filters resulted in ≈86% capture efficiency and ≈92% retrieval efficiency for epithelial and mesenchymal cancer cells [97]. Size-dependent separation of staphylococcus bacteria (1 µm) and platelets ((2–3) µm) with polystyrene in water and polyethylene oxide co-flow achieved >90% of separation efficiency and purity [68]. The needle-like Cytosensor is newly designed to detect and capture colon cancer cells with a detection span of (102–106)/mL for a flow speed of 10 mL/min [95], as shown in Figure 2.

In the literature, the highest segregation efficiencies attained for MDA-MB 231 [98], staphylococcus bacteria [68], platelets [68], epithelial [97,99], mesenchymal [97,99], and MCF-7 [98] circulating tumor cells are 81.2%, 90%, 90%, 93.81%, 95.13%, and 96.3%, respectively. The same is shown in Figure 2. MDA-MB 231 has attained the least segregation efficiency, and MCF-7 circulating tumor cells have attained the highest.

At the same time, the percentage of recovery achieved for exomes is at least 80% [100], and for RBC/WBC, it is the highest, with 99.5% recovery [73]. The same is plotted in Figure 3 and Figure 4, which show the purity attained by different methods in the literature for biological cells/bacteria, where EpCAM -ve circulating tumor cells have attained the lowest purity of 81.2% [98,101] and WBC/RBC cells have attained the highest purity percentage of 99.8% [73]. Besides these live-cell separations, sorting also applies to blood plasma [73,102,103] and enzymes [71].