The application of nanofluids to boiling heat transfer was first motivated by their abnormally enhanced thermal conductivity at nanoparticle concentrations on the order of a few percent by volume 44 . However, You et al., in their pioneering research 5 on CHF enhancement in nanofluids, reported that continued increases in CHF were not observed at concentrations higher than approximately 0.01 vol.%, which is significantly lower than the usual concentration of nanoparticles used for the enhancement of thermal conductivity in nanofluids. Thus, the observed CHF increases could not be explained in terms of the effect of nanoparticles on thermal conductivity enhancement. In addition to thermal conductivity, it was revealed that all other physical properties of dilute nanofluids, including surface tension, vapor and liquid density, viscosity, heat of vaporization, and boiling point temperature, are almost the same as the corresponding properties of pure water 28 45 46 . In summary, the transport and thermodynamic properties of nanofluids at low concentration (<0.01 vol.%) are very similar to those of pure water. It can be concluded that changes in the properties of nanofluids do not account for the enhancing effect of nanoparticles on liquid-to-vapor phase-change heat transfer.

To interpret the mechanism of CHF enhancement in nanofluids, two kinds of hypotheses on the roles of nanoparticle during nanofluid boiling were suggested in the early stage of research.

Vassallo et al. 6 (one of the initial studies in which significant CHF enhancement in nanofluids was observed) reported that a major deposition of nanoparticles (about 0.15-0.2 mm thick) occurs on the heater surface during nanofluid boiling, suggesting some possible interactions between the nanoparticles and the heated surface at high heat fluxes. Soon afterward, Milanova and Kumar 32 and Bang and Chang 47 confirmed that nanoparticles precipitate on the surface during nucleate boiling, forming a layer whose morphology depends on the nanoparticle material, and suggesting some surface effects on CHF phenomena such as the trapping of liquid near the heater surface due to porous structures and the breakup of voids near the surface. Figure 7 shows a SEM picture of NiCr wire after deposition of silica nanoparticles during nanofluid boiling.

Sefiane 48 suggested an alternative approach to clarify the mechanism by which the presence of nanoparticles affects heat transfer and CHF during boiling. He demonstrated experimentally that the nanoparticles in the liquid promote the pinning of the contact-angle line of the evaporating meniscus and sessile drops. He explained that the observed results were due to the structural disjoining pressure stemming from the ordered layering of nanoparticles in the confined wedge of the evaporating meniscus 49 (Figure 8) and suggested that an analysis of the boiling heat transfer of nanofluids could account for the strong effect of nanoparticles on the contact-line region via the structural disjoining pressure. Wen 50 51 subsequently carried out further investigations of the influence of nanoparticles on the structural disjoining pressure. He calculated the equilibrium meniscus shape in the presence of nanoparticles and found that the vapor-liquid-solid line could be significantly displaced toward the vapor phase by the presence of nanoparticles in the liquid. He therefore concluded that the structural disjoining pressure caused by nanoparticles can significantly increases the wettability of the fluids and inhibits the development of dry patches, triggering CHF.

The above-described two effects of nanoparticles (i.e., modification of the heater surface and structural disjoining pressure) both seem to be plausible hypotheses for CHF enhancement in nanofluids. However, to understand the principle mechanism of the phenomena, it is necessary to examine the single contribution of each factor to the enhanced CHF performance of nanofluids. Kim et al. 52 53 carried out an insightful experiment to separate the single effect of the nanoparticle deposition layer on the CHF of nanofluids. First, they conducted a pool-boiling test of a nanofluid using a fresh heater wire. A subsequent surface inspection confirmed the presence of a nanoparticle deposition layer on the heater wire. They then performed an additional CHF test on the nanoparticle-deposited wire submerged in pure water, which resulted in a CHF enhancement of the same magnitude as that of the nanofluids. The experimental results clearly demonstrated that the enhancement of CHF in nanofluids is due to the modification of surface topology associated with nanoparticle deposition on the heater surface during nanofluid boiling. Moreover, Golubovic et al. 19 and Kwark et al. 28 recently conducted the same experiments using both thin wire and flat-plate heaters and obtained experimental results consistent with those of Kim et al. 52 53 . Figure 9 shows the experimental results obtained by Kim et al. 53 and Kwark et al. 28 .

The preceding conclusion on the role of nanoparticle deposition is compatible with the recent work of Kim et al. 54 , who studied pool boiling heat transfer during the quenching of a hot sphere in a nanofluid. They reported that the CHF remained unchanged when a clean sphere was cooled in the nanofluid, and it was only enhanced during the cooling of a sphere with a nanoparticle layer. This result, therefore, confirmed that the deposition layer of the nanoparticles plays a critical role in effectively enhancing the CHF by modifying the heater surface. In conclusion, an understanding of the underlying mechanism should be sought to study the influence of the nanoparticle-deposited surface on CHF.

Before proceeding to the assessment of surface effects on CHF enhancement in nanofluids, a prior question arises: why are nanoparticles deposited on the heater surface during nucleate boiling of nanofluids? Kim et al. 52 reported that the nanoparticle layer developed only during nucleate boiling in nanofluids, but was not caused by gravitational sedimentation or single-phase natural convection. Kim et al. 46 suggested the hypothesis that the evaporation of microlayers initially containing nanoparticles could be the reason for the formation of the porous layer. As vapor bubbles grow, the evaporating liquid leaves behind nanoparticles, which then concentrate at the base of the bubbles, forming the microlayer. As the microlayer evaporates, nanoparticles are again left behind, and they then bond to the hot heater surface. Kwark et al. 28 recently confirmed this theory by optically observing a single circular nanoparticle coating formed on a boiling surface, where a single-active bubble nucleation site was allowed to undergo several boiling cycles, as shown in Figure 10. Accordingly, nanofluid boiling itself, and specifically microlayer evaporation, is responsible for producing the nanoparticle layer on the surface.

Kim et al. 46 55 investigated the surface effect on CHF enhancement of water-based nanofluids containing alumina, zirconia, and silica nanoparticles. In their research, the deposition of nanoparticles on the heater surface significantly improved the wettability, as measured by the reduction of the static contact angle (see Figure 11). Note that no appreciable differences were found between pure water and nanofluids. They inferred that the buildup of a porous layer with oxide nanoparticles increases the adhesion tension γ _SV - γ _SL and the roughness factor r (the ratio of the effective contact area to the smooth contact area), and both effects lead to a pronounced reduction of the contact angle in accordance the modified Young-Laplace equation 56 ,

where θ and σ are the contact angle and the surface tension, respectively. A systematic review of the prevalent CHF theories then demonstrates that higher wettability can produce a CHF enhancement consistent in magnitude with the experimental observations of numerous researchers. Subsequently, a number of studies focusing on the role of a nanoparticle layer, including Coursey and Kim 57 , Liu and Liao 21 , Golubovic et al. 19 , and Jeong et al. 58 , produced the same conclusions as Kim et al. 46 in regard to the significant improvement of surface wettability and its role in the CHF enhancement of nanofluids.

Kim et al. 59 also found that a porous layer of nanoparticles significantly improved the surface wettability. On the other hand, they were the first to show that the effect of wettability alone cannot explain additional increases beyond the attainable CHF value when the contact angle approaches zero. By focusing on the role of capillarity in the CHF behavior of nanofluids, they showed that capillarity causes the liquid to rise on the nanoparticle-deposited surfaces in accordance with

where R _c and cosθ represent the microscopic structures and surface wettability of the nanoparticle layers, respectively. Capillary flow during boiling supplies fresh liquid to the dry region beneath the vapor bubbles, delaying the irreversible growth of hot spots and CHF. Kim and Kim 60 used capillarity to characterize a completely wetted nanoparticle-coated surface. They showed that the estimated heat-flux gain due to capillary liquid supply along the porous layer was of the same order of magnitude as that due to wettability enhancement (Figure 12). They concluded that the significant CHF enhancement of nanofluids during pool boiling is a consequence not only of increased surface wettability, but also of improved capillarity resulting from the surface deposition of nanoparticles.

A consensus explanation of the cause of CHF enhancement in nanofluids seems to be obtainable via an intense study focused on the effect of the nanoparticle layer. In other words, the CHF of a nanofluid is enhanced by its improved ability to actively wet the heater surface, thanks to the porous structure of the thin nanoparticle sorption layers.

The hydrodynamic instability model of Zuber 61 is the most widely used correlation to predict pool boiling CHF,

where h _fg is latent heat of evaporation. However, the initial research of the nanofluid CHF (for example, You et al. 5 and Kim et al. 46 ) discarded hydrodynamic instability theory as an interpretative tool for CHF in nanofluids because of its inability to account for surface effects. In fact, the CHF correlation of Eq. 4 does not depend on the fluid properties at all, whereas the primary reason for increased CHF in nanofluids is the change in surface characteristics associated with the deposition of nanoparticles during nanofluid boiling. You et al. 5 concluded using their nanofluid CHF studies that some important unknown factors, potentially missing from Zuber's theory, might be responsible for the increased CHF in nanofluids.

On the other hand, Golubovic et al. 19 suggested a possible approach to interpreting the CHF mechanism in nanofluids by modifying the hydrodynamic instability models of Lienhard and Dhir 25 and Ramilison et al. 62 . They hypothesized that a change in the surface contact angle alters the size and spacing of the vapor jets above the heater surface, so that the surface effect on the nanoparticle deposition can be incorporated into the hydrodynamic CHF model proposed by Lienhard and Dhir 25 .

Recently, Park et al. 63 reported that the CHF of water-based nanofluids containing graphene/graphene-oxide nanoparticles was as high as that of alumina nanofluid, even though neither the wettability nor the capillarity of the surface was improved on the nanoparticle layers. Alternatively, they measured the dewetting wavelengths of water on heater wires and reported that the wavelength change corresponds to the CHF enhancement tendency for all tested nanofluids. Although a direct correlation between the critical instability wavelength obtained from Zuber's theory and the dewetting wavelength of the liquid is questionable, they concluded that the wavelength modulation most adequately supports the CHF enhancement of nanofluids.

Haramura and Katto 26 proposed the macrolayer dryout model. In this model, CHF occurs due to macrolayer dryout if the heat flux is sufficient to evaporate the macrolayer before the departure of the mushroom bubble. Kim et al. 46 examined the impact of the improved wettability of a nanoparticle layer (or reduction of the contact angle) on the equivalent thickness of the macrolayer. When the macrolayer thickness was calculated using the model of Sadasivan et al. 64 , they found that a contact angle reduction due to nanoparticle deposition could produce an increase in the thickness of the liquid layer enough to result in a significant increase in the CHF.

Bubble crowding at a heated surface was proposed by Rosenhow and Griffith 65 . In this model, close packing of bubbles on the heater surface is responsible for the cessation of the liquid flow toward the surface, leading to CHF. More sophisticated theories include the effect of the shear force generated by the mutual interaction of growing and departing bubbles, for example, Kolev 66 ,

where Δτ_W and Δτ_d are the bubble wait time and departure time at the heated surface, respectively. In Eq. 5, the bubble departure time is a strong function of the nucleation site density (n"). According to the Wang and Dhir 67 correlation, the site density decreases with the contact angle. Thus, the intensity of the shear stress generated by the mutual interactions of bubbles grows slowly on a heated surface with a low contact angle compared with a surface with a large contact angle. Kim et al. 46 found that according to the Kolev 66 model, a change in the contact angle θ can have a major impact on the CHF. Therefore, it could be concluded that the bubble-interaction theory supports the notion of surface wettability improvement as a plausible cause of CHF enhancement in nanofluids.

Hot spot model was first proposed by Unal et al. 68 . This model suggests that the temperature at the center of a dry patch on the heater surface is an important parameter that can trigger CHF. Ability of cooling liquid to rewet the heated dry area should make the strong impact on CHF. In accordance with this idea, Theofanous and Dinh 30 proposed the modified hot spot model with focus on the micro-hydrodynamics of the solid-liquid-vapor line at the boundary of a hot/dry spot. In their model, CHF occurs when the evaporation recoil force (which drives the liquid meniscus to recede) becomes larger than the surface tension force (which drives the meniscus to advance and rewet the hot/dry spot).

Kim et al. 46 semi-quantitatively showed that the improved wettability on the nanoparticle-fouled surface significantly increases the surface tension force to rewet the hot/dry spot, suggesting higher CHF. In this regard, they concluded that the hot/dry spot model incorporating the micro-hydrodynamics of a liquid meniscus corroborates the link between increased wettability and CHF enhancement in nanofluids. In addition, Kim et al. 69 conducted sessile-drop wetting experiments focused on the effect of a nanoparticle layer on the stability of an evaporating meniscus. They found that an individual liquid meniscus is more stable on an alumina nanoparticle layer and hence can sustain the evaporation recoil force at a higher heat flux. The evaporative heat-flux gain attainable on the nanoparticle layer was of the same order of magnitude as the CHF increases in nanofluids. Thus, these experimental results also supported the hot/dry spot theory based on the micro-hydrodynamics of a liquid meniscus.

Several recent studies have demonstrated that the CHF model proposed by Kandlikar and Steinke 70 is reasonably well correlated with measured CHF data in nanofluids as a function of contact angle (see, for example, 40 71 ). This model utilizes the evaporation momentum force and receding contact angle β as parameters. Fundamentally, this model also focuses on the micro-hydrodynamics of the vapor-liquid interface of a single bubble at the heater surface based on the force balance at the solid-vapor-liquid triple-contact line. The accuracy of this model in predicting the CHF of nanofluids supports the argument that the hot/dry spot model incorporating the micro-hydrodynamics of an evaporating meniscus is a plausible mechanism.