*4.1. Stability Mechanism and Evaluation*

The stability of nanofluids is of major concern for maintaining the thermophysical properties of the mixture [169]. Specifically, the stability of the suspension combines several aspects such as dispersion stability, kinetic stability, and chemical stability [120,170]. The dispersion stability deals with nanoparticles aggregation within the colloidal, while the kinetic stability describes the Brownian motion of nanoparticles hosted by the base fluid (i.e., sedimentation of randomly agglomerated particles due to gravity). As for the chemical stability, it is associated with the chemical reactions that occur between the nanoparticles themselves and between the nanoparticles and the surrounding base fluid. However, it is essential to note that chemical reactions in a nanofluid are minimized or halted at low temperature conditions (i.e., below the temperature point of a chemical reaction). Hence, agglomeration and sedimentation of nanoparticles would be the primary aspects concerned with suspension stability. When a nanofluid is physically unstable, the formed sedimentation can have one of three behaviors, namely; 1—dispersed sedimentation, 2—flocculated sedimentation, or 3—mixed sedimentation [8]. Figure 12 shows a schematic illustration of the realistic reflection for the three types of sedimentation behaviors. In addition, the speed at which the sediment forms and settles within an unstable suspension can be classified into two main regions. The first is known as the rapid settling region, which occurs at the beginning stage of the separation of the particles from the hosting base fluid; and the following stage is called the slow settling region, where the changes in sediment formation and settling becomes insignificant along the shelving lifetime [171]. Figure 13 demonstrates an example of the two sedimentation speed formation regions from Witharana et al. [171] investigation. Furthermore, there are about eight techniques that can be used to evaluate the stability of nanofluids, such as 1—sedimentation photographical capturing method, 2—dynamic light scattering (DLS) approach, 3—zeta potential analysis, 4—3-ω approach, 5—scanning electron microscopy (SEM) analysis, 6—TEM characterization, 7—spectral analysis, and 8—centrifugation method. From the previous stability evaluation methods, the sedimentation photographical capturing approach is considered as the most reliable route between them all, but at the expense of time (i.e., it takes a very long time to conduct and analyze). The DLS approach usually over-predicts the size of the particles, especially when using a non-ionized base fluid (e.g., deionized water), where the analysis can show larger values (from 2 to 10 nm more) than the actual particle size [172]. Such results are very problematic and misleading when analyzing nanofluids, especially when the dispersed particles are 10 nm or less in size, where the oversized prediction can incorrectly indicate an instability state. On the other hand, the zeta potential analysis should only be used as a supportive characterization tool. This is because if the nanoparticles and/or the base fluid are non-polar or even of low polarity, there may be other mechanisms affecting the suspension stability [172]. Thus, it is highly recommended to use multiple approaches (e.g., three methods) to determine the stability of the nanofluid. A detailed description of each of the experimental stability evaluation approaches, and their advantages and limitations can be found in the work published by Ali et al. [13]. Other than the previous stability evaluation approaches, Carrillo-Berdugo et al. [173] have proposed a novel theory-based design framework for determining the polarity between the solid and liquid interface, which can be used to adjust the interface tension by adding the required number of dispersive components to meet those of the dispersed nanomaterial.

*Nanomaterials* **2021**, *11*, x FOR PEER REVIEW 21 of 79

*Nanomaterials* **2021**, *11*, x FOR PEER REVIEW 21 of 79

**Figure 12.** The three types of sedimentation behaviours, where (**a**) shows their schematic mechanism from the starting time (to) to the finishing time (tf) [8], (**b**) demonstrates the dispersed and flocculated sedimentation behaviors from Ali et al. [8] experimental work, and (**c**) represents the mixed sedimentation behavior shown in Ma and Alain [174] investigation. **Figure 12.** The three types of sedimentation behaviours, where (**a**) shows their schematic mechanism from the starting time (to) to the finishing time (t<sup>f</sup> ) [8], (**b**) demonstrates the dispersed and flocculated sedimentation behaviors from Ali et al. [8] experimental work, and (**c**) represents the mixed sedimentation behavior shown in Ma and Alain [174] investigation. **Figure 12.** The three types of sedimentation behaviours, where (**a**) shows their schematic mechanism from the starting time (to) to the finishing time (tf) [8], (**b**) demonstrates the dispersed and flocculated sedimentation behaviors from Ali et al. [8] experimental work, and (**c**) represents the mixed sedi-

mentation behavior shown in Ma and Alain [174] investigation.

**Figure 13.** A demonstration of the two sedimentation regions in terms of settling speed, where (the left side) shows the rapid region in which the sediment height changes rapidly, and (the right side) illustrates the slow region, where the changes in the sediment height are very slow to the point where it can be negligible [13]. **Figure 13.** A demonstration of the two sedimentation regions in terms of settling speed, where (the left side) shows the rapid region in which the sediment height changes rapidly, and (the right side) illustrates the slow region, where the changes in the sediment height are very slow to the point **Figure 13.** A demonstration of the two sedimentation regions in terms of settling speed, where (the left side) shows the rapid region in which the sediment height changes rapidly, and (the right side) illustrates the slow region, where the changes in the sediment height are very slow to the point where it can be negligible [13].

*4.2. Stability Enhancements*

*4.2. Stability Enhancements*

where it can be negligible [13].

Several approaches have been shown to improve the stability of nanofluids successfully. These methods are subdivided into two main categories, which are in the form of

Several approaches have been shown to improve the stability of nanofluids successfully. These methods are subdivided into two main categories, which are in the form of

#### *4.2. Stability Enhancements Nanomaterials* **2021**, *11*, x FOR PEER REVIEW 22 of 79

Several approaches have been shown to improve the stability of nanofluids successfully. These methods are subdivided into two main categories, which are in the form of physical and chemical routes. The physical approach involves the employment of high energy forces such as ultrasonication, magnetic stirring, homogenizer (or probe sonicator), or even ball milling, which is rarely reported [117,175]. Figure 14 shows the four previous physical stability methods. Unlike the ultrasonication and homogenization methods, the magnetic stirring approach is considered as the most basic route that can be applied to break-down clusters of nanoparticles, within the suspension, with very low performance effectiveness when compared to the other two physical methods [176]. Furthermore, in the literature [177], high pressure homogenization was shown to provide better stability characteristics than ultrasonication to the as-produced nanofluids. In addition, the mixing duration and intensity used in the sonicator device were commonly seen to vary from one research work to another in an attempt to physically stabilize the nanofluid. A good explanation for the aforementioned method is that the mixing power cannot be maintained constant throughout the process due to the voltage fluctuation that the device experienced. Therefore, Yu et al. [178] suggested relying on the relation between the suspension absorption spectra against the total energy supplied to the mixture as a relative solution to the sonication time. physical and chemical routes. The physical approach involves the employment of high energy forces such as ultrasonication, magnetic stirring, homogenizer (or probe sonicator), or even ball milling, which is rarely reported [117,175]. Figure 14 shows the four previous physical stability methods. Unlike the ultrasonication and homogenization methods, the magnetic stirring approach is considered as the most basic route that can be applied to break-down clusters of nanoparticles, within the suspension, with very low performance effectiveness when compared to the other two physical methods [176]. Furthermore, in the literature [177], high pressure homogenization was shown to provide better stability characteristics than ultrasonication to the as-produced nanofluids. In addition, the mixing duration and intensity used in the sonicator device were commonly seen to vary from one research work to another in an attempt to physically stabilize the nanofluid. A good explanation for the aforementioned method is that the mixing power cannot be maintained constant throughout the process due to the voltage fluctuation that the device experienced. Therefore, Yu et al. [178] suggested relying on the relation between the suspension absorption spectra against the total energy supplied to the mixture as a relative solution to the sonication time.

**Figure 14.** Physical dispersion stability enhancement devices, where (**a**) shows the ultrasonic bath sonicator, (**b**) demonstrate the magnetic stirrer, (**c**) illustrates the homogenizer/prob sonicator, and (**d**) shows the ball milling device. Reproduced with permission from [116]. Elsevier, 2020. **Figure 14.** Physical dispersion stability enhancement devices, where (**a**) shows the ultrasonic bath sonicator, (**b**) demonstrate the magnetic stirrer, (**c**) illustrates the homogenizer/prob sonicator, and (**d**) shows the ball milling device. Reproduced with permission from [116]. Elsevier, 2020.

On the other hand, the chemical route stabilizes the suspension by declustering the agglomerated nanoparticles by alternating the pH value of the base fluid or the mixture, adding surfactant(s) to the solid–liquid matrix, or modifying the surface of the nanoparticles. Nanofluids pH alteration affects the level of free cations or anions charges in the media surrounding the dispersed particles, and hence the hydrophilicity or hydrophobicity nature of the particles changes causing the colloidal to either stabilize or destabilize [179,180]. The disadvantage of the previous method is that fabricating suspensions of high

On the other hand, the chemical route stabilizes the suspension by declustering the agglomerated nanoparticles by alternating the pH value of the base fluid or the mixture, adding surfactant(s) to the solid–liquid matrix, or modifying the surface of the nanoparticles. Nanofluids pH alteration affects the level of free cations or anions charges in the media surrounding the dispersed particles, and hence the hydrophilicity or hydrophobicity nature of the particles changes causing the colloidal to either stabilize or destabilize [179,180]. The disadvantage of the previous method is that fabricating suspensions of high or low pH values may be corrosive for high heat flux applications. In addition, surfactants are essential when dispersing nanomaterials of hydrophobic nature (e.g., CNTs and graphene) in a polar base fluid (e.g., water), and vice versa [181,182]. This is because the added surfactant would act as a bridge between the nanoparticles and the hosting fluid, and therefore would improve the dispersion stability of the particles through increasing the repulsive force between the particles themselves and reducing the interfacial tension between the base fluid and the hosted particles. Surfactants are categorized based on their head group charge as cationic, non-ionic, anionic, and amphoteric. Table 3 shows some of the surfactants used in the nanofluids preparation process according to their head group charge [118].

**Table 3.** Examples of surfactants used in nanofluids fabrication categorized by classifications based on their head group charge.


The downside from using surfactants is that the nanofluid becomes more viscous; starts to generate foam when being heated or cooled down; can be lost at high temperatures, and would reduce the overall thermal conductivity of the suspension. As for the nanoparticles surface modification technique, the particles are either initially functionalized (before the dispersion process), or the functionalized materials themselves are added to the colloidal (where they get grafted to the surface of the segregated particles), and therefore forming a new particle surface exposure to the hosting base fluid [183,184]. The drawback of using functionalized materials as stabilizers is that they tend to reduce the overall thermal conductivity of the produced nanofluid due to having a significantly lower thermal conductivity than the dispersed nanoparticles. Figure 15 recaps all of the nanofluid stability improvement methods that were mentioned earlier in this section.

methods.

tion (8)):

**Figure 15.** Nanofluids stability improvement methods categorized by their physical and chemical **Figure 15.** Nanofluids stability improvement methods categorized by their physical and chemical methods.

### **5. Stability Effect on Thermophysical Properties 5. Stability Effect on Thermophysical Properties**

The thermophysical properties govern the heat transfer rate that the nanofluid can provide to the system in which it is employed as a working fluid. Nanofluids thermal properties, such as the thermal conductivity, greatly depend on the type of base fluid, nanoparticles material, morphological characteristics of the particles, nanoparticles concentration, and homogeneity of nanoparticles dispersion in the hosting base fluid. The dispersion characteristics of the suspension are subjected to alteration with the change in stability of the particles in their surrounding environment (i.e., base fluid). For such reason, the stability of a nanofluid is considered as a significant factor to maintain the heat transfer rate from and to the colloidal. This section covers the influence of stability on nanofluids effective thermal conductivity and effective viscosity. It is important to highlight that the effect of suspension stability, as a parameter, on the effective density was not reported across the literature, but rather the added surfactants and particles concentration were seen responsible for the changes caused to nanofluids densities [185–187]. This is because nanofluids effective density (ρnf) is constrained by its overall volume and The thermophysical properties govern the heat transfer rate that the nanofluid can provide to the system in which it is employed as a working fluid. Nanofluids thermal properties, such as the thermal conductivity, greatly depend on the type of base fluid, nanoparticles material, morphological characteristics of the particles, nanoparticles concentration, and homogeneity of nanoparticles dispersion in the hosting base fluid. The dispersion characteristics of the suspension are subjected to alteration with the change in stability of the particles in their surrounding environment (i.e., base fluid). For such reason, the stability of a nanofluid is considered as a significant factor to maintain the heat transfer rate from and to the colloidal. This section covers the influence of stability on nanofluids effective thermal conductivity and effective viscosity. It is important to highlight that the effect of suspension stability, as a parameter, on the effective density was not reported across the literature, but rather the added surfactants and particles concentration were seen responsible for the changes caused to nanofluids densities [185–187]. This is because nanofluids effective density (*ρn f*) is constrained by its overall volume and mass, where it can be directly calculated from extending the rule of mixtures (i.e., Equation (8)):

$$
\rho\_{nf} = f\_{\rm V} \times \rho\_{np} + (1 - f\_{\rm V}) \times \rho\_{bf} \tag{8}
$$

ρnf = ƒ<sup>V</sup> × ρnp + (1 − ƒ<sup>V</sup> ) × ρbf (8) where ƒ<sup>V</sup> is the particles volumetric fraction, ρnp is the density of the nanoparticles, and ρbf is the density of the base fluid. Similarly, the effective specific heat capacity of the colloidal was not shown to be linked to the dispersion stability. The main parameter that where *f*<sup>V</sup> is the particles volumetric fraction, *ρnp* is the density of the nanoparticles, and *ρb f* is the density of the base fluid. Similarly, the effective specific heat capacity of the colloidal was not shown to be linked to the dispersion stability. The main parameter that affects nanofluids effective specific heat capacity is the particles concentration included

mass, where it can be directly calculated from extending the rule of mixtures (i.e., Equa-

in the mixture. This is because increasing the nanoparticles concentration would result in enhancing the overall thermal performance of the suspension, and hence less heat would be required to raise the temperature of the fabricated nanofluid, and vice versa [188]. In general, nanofluids effective specific heat capacity is lower than their base fluids [126,189]. According to Ali et al. [13] and other researchers [190–193], the most accurate theoretical model for calculating the effective specific heat capacity of a nanofluid (*Cpn f* ) is the following equation:

$$\mathbb{C}\_{p\_{nf}} = \frac{\rho\_{bf} \times (1 - f\_{\rm V})}{\rho\_{nf}} \times \mathbb{C}\_{p\_{bf}} + \frac{\rho\_{np} \times f\_{\rm V}}{\rho\_{nf}} \times \mathbb{C}\_{p\_{np}} \tag{9}$$

where *Cpb f* and *Cpnp* are the specific heat capacities of the base fluid and the nanoparticles, respectively. Experimentally, the *Cpn f* can be determined using the differential scanning calorimetry (DSC) technique, which basically measures the amount of heat required to be delivered to both test sample and reference source, of well-known heat capacity, so that a temperature rise can be achieved [188].
