An updated review of nanofluids in various heat transfer devices

Abstract

The field of nanofluids has received interesting attention since the concept of dispersing nanoscaled particles into a fluid was first introduced in the later part of the twentieth century. This is evident from the increased number of studies related to nanofluids published annually. The increasing attention on nanofluids is primarily due to their enhanced thermophysical properties and their ability to be incorporated into a wide range of thermal applications ranging from enhancing the effectiveness of heat exchangers used in industries to solar energy harvesting for renewable energy production. Owing to the increasing number of studies relating to nanofluids, there is a need for a holistic review of the progress and steps taken in 2019 concerning their application in heat transfer devices. This review takes a retrospective look at the year 2019 by reviewing the progress made in the area of nanofluids preparation and the applications of nanofluids in various heat transfer devices such as solar collectors, heat exchangers, refrigeration systems, radiators, thermal storage systems and electronic cooling. This review aims to update readers on recent progress while also highlighting the challenges and future of nanofluids as the next-generation heat transfer fluids. Finally, a conclusion on the merits and demerits of nanofluids is presented along with recommendations for future studies that would mobilise the rapid commercialisation of nanofluids.

Introduction

Energy is a very important quantitative property that must be transferred before any system can perform work. The transfer of energy can be done by either work or heat [1]. Heat is transferred from one system to another when there exists a temperature difference between the two systems and travels from high to low temperatures [2]. The science that describes the means and rate in which thermal (heat) energy is transferred is known as heat transfer. Heat transfer applications are experienced in our daily life; the human body, for instance, is constantly emitting heat, and humans adjust their body temperature to suit environmental conditions using clothing. Heat transfer is also used in our buildings to regulate temperature [3] and is necessary for cooking, refrigeration and drying. It is also directly applied in car radiators [4] and for temperature control in electronic devices [5]. Heat transfer is used in solar thermal collectors to convert solar energy to heat and power [6, 7] and used in thermal control elements in spacecraft [8]. In many of these devices, heat needs to be dissipated at a rapid rate to ensure effective operation and maximum efficiency within the system [9]. As technology evolves, devices have become smaller and thus require better thermal management. Essentially, the more compact the size, the larger the requirement for effective cooling technology [10]. Therefore, heat transfer enhancement is a very important area in thermal engineering.

Several techniques have been considered to improve the heat transfer coefficient between the working fluids and the fluid contact surfaces [11, 12]. Conventional heat transfer fluids such as water, thermal oils and ethylene glycol/water have some limitations as their thermal properties are quite low when compared to those of solids, as shown in Fig. 1. The improvement in the thermal properties of these fluids through the addition of nanoscaled particles has led to an evolution in the study of heat transfer fluids. The suspension of these solid particles in the base fluid enhances the energy transmission in the fluid leading to improved thermal conductivity properties and better heat transfer characteristics [13]. The resultant fluids have been seen to possess higher values of thermal conductivity [14, 15]. Choi and Eastman [15] were the first to name such fluids as nanofluids. Nanofluids are the engineered colloidal suspension of nanoscaled particles (10–100 nm) in a base fluid [16]. These particles are generally metals, metallic oxides or other carbon-based elements. Over a century ago, Maxwell [17] was the first to discuss the suspension of micro-scaled particles into a fluid. However, microparticles settled rapidly in the fluid leading to abrasion and clogging in the flow channel, limiting further research into suspensions in fluids. Furthermore, these fluids did not exhibit the significant enhancement witnessed today with the use of nanofluids. The introduction of nanoparticles has allowed for further investigation into colloidal dispersion in fluids. Nanoparticles are more stable when dispersed in fluids and tend to improve on the thermal properties of the fluids. Some other properties of nanofluids which make them adequate heat transfer fluids include the Brownian motion of particles, particle/fluid nanolayers and their reduced pump power when compared to pure liquids to achieve intensified heat transfer.

Fig. 1

Bulk material thermal conductivity difference between a commonly used base fluids and b commonly used nanoparticles

Despite these benefits, nanofluids still possess some application-based limitations. Issues of sedimentation and aggregation in the fluid have been raised, although the use of ultra-sonication, pH modulation, magnetic stirring and the addition of surfactants has been recorded to improve the stability of the nanofluids [18]. Also, increasing the fluid circulation rate in the device reduces the possibilities of sedimentation, although this can lead to an erosion of heat transfer in the device or flow stream. Particles of larger sizes also tend to clog the flow channel, and there have been cases of pressure loss recorded in some devices due to the marginal increase in viscosity. Nanofluids are also expensive to prepare and toxic due to the reactive nature of the nanoparticles [19]. Over the past decade, emphasis on nanofluids research has been more apparent as shown in Fig. 2, which illustrates the number of publications involving nanofluids since 2010. These studies include those related to their preparation, characterisation, measurement of their physical properties and their utilisation in various applications. The data presented in Fig. 2 were obtained by searching the word “nanofluids” in the Scopus database against titles, abstracts and keywords over the period presented. The search illustrates that approximately 3165 papers were published in 2019 alone, and this trend is expected to increase in the coming years. Incidentally, several review papers related to the heat transfer properties and application of nanofluids were published in 2019. The reviews include solar collector applications [20, 21], a review of nanofluids in heat exchangers [22, 23], review of nanofluids in heat pipes [24], radiator cooling [25], electronic cooling [5] and also a review on various thermophysical properties of nanofluids [26, 27]. A list of some of the review papers published in 2019 is listed in Table 1.

Fig. 2

Nanofluids-related publication in the past decade

Table 1 Reviews on heat transfer-related application of nanofluids published in 2019

Owing to the increasing number of studies relating to nanofluids, there is a need for a holistic review of the progress and steps taken in 2019 concerning their application in heat transfer devices. This study adopts a retrospective look at the year 2019 by reviewing the progress made in the area of nanofluids preparation, nanofluid thermophysical properties measurements and the applications of nanofluids in various heat transfer devices including solar collectors, heat exchangers, refrigeration systems, radiators, thermal storage systems and electronic cooling. The study aims to update readers on recent progress in nanofluid synthesis and application. The study also seeks to highlight the challenges and prospects of nanofluids as the next-generation heat transfer material.

Preparation of nanofluids

The method used in the preparation of nanofluids is important in the study of the stability and thermophysical behaviour of nanofluids [57]. The preparation steps are also vital in estimating the degree to which the nanofluids are employed in heat transfer systems [58], [59]. In this section, the studies related to nanofluid stability and their synthesis techniques are discussed. Nanofluids are produced by suspending particles of nanosize dimensions in the traditional heat transfer fluids such as water, oils, acetone and glycols [60]. A wide range of nanoparticles have been utilised in the formation of nanofluids, some of these include:

• (1) Carbon nanoparticles (such as MWCNT, SWCNT, Gn, GO, graphite, diamond and fullerene).

• (2) Metal nanoparticles (such as Ag, Al, Au, Co, Cu and Fe).

• (3) Metal oxide nanoparticles (such as Al₂O₃, CeO₂, CuO, Fe₃O₄, TiO₂ and ZnO).

• (4) Others (such as Si, AlN-C, CoFe₂O₄, SiC, Field’s alloy nanoparticles, ZnBr₂ and SiO₂).

Nanofluids can be unstable due to the strong Van der Waals interactions and cohesive forces between nanoparticles. Therefore, the preparation technique used is extremely important in others to break down these forces and produce stable nanofluids. Different methods have been used to avoid nanoparticle agglomeration and improve the stability of nanofluids, such as pH control, surfactant addition, ultrasonic agitation, magnetic stirring, functionalisation and high-pressure homogenisation [61]. According to Yu and Xie [61], there are three main methods used for nanofluids preparation: one-step chemical technique, one-step physical technique and two-step technique.

The two-step technique is the most widely used for nanofluid preparation, and it is more economical for mass production. In this method, the industrial or laboratory-synthesised nanoparticles are dispersed in the base fluids by agitation, stirring or ultra-sonication [60]. A significant drawback of this technique is it often has low stability and a high tendency of agglomeration. To avoid this problem, several additional techniques, including one-step synthesis techniques and green synthesis techniques, have been used. The two-step nanofluid preparation technique is illustrated in Fig. 3.

Fig. 3

Two-step method of preparing nanofluids. (modified from [33])

Asadi et al. [62] prepared MWCNT/water nanofluid by implementing the two-step method without using surfactants, and both agitation and ultra-sonication were used in the nanofluid preparation. Their nanofluid was stable for one month. However, with the addition of surfactants, Chen et al. [63] adopted the two-step method to prepare MWCNT/DI water using polyvinylpyrrolidone (PVP) as a surfactant. They performed ultra-sonication for only one hour, and they obtained a stable solution for more than two months. Additionally, Almanassra et al. [64] used the two-step method to investigate the stability and thermophysical properties of MWCNT/water nanofluids using different surfactants PVP, SDS and AG. They prepared the nanofluids using ultra-sonication for 30 min, and the solutions were stable for more than six months. Other studies have used different surfactants such as SDBS, CTAB, oleic acid, ethyl carbamate, PEG, EBT, oleylamine, citric acid, Tween 80, Gemini’s and APTMS. The use of surfactants depends on the nature of the base fluid and type of nanoparticles. The surfactants act as a bridge between the base fluids and the nanoparticles. It increases the repulsive forces between the particles and decreases the interfacial tension between the base fluids and suspended nanoparticles. Surfactants also increase the zeta potential of nanoparticles and increase the hydrophilic properties of the suspended particles. Surfactants are classified based on the charge of their head group. Therefore, the surfactants might be cationic (positively charged) such as CTAB, distearyl dimethylammonium chloride and benzalkonium chloride; non-ionic (uncharged, neutral) such as oleic acid, PVP, AG, Tween 80 and oleylamine; anionic (negatively charged) such as SDBS and SDS; and amphoteric (negatively and positively charged) such as lecithin.

Hydrophobic materials are generally chemically or physically pre-treated to improve the particle’s hydrophilicity and their stability in the base fluids. Liu et al. [65] used the two-step method to prepare GO/water nanofluids with particle concentration of (1–4.5 mg mL⁻¹). GO was initially prepared from graphite using the Hummers method; then, nanofluids were prepared using both agitation and ultra-sonication for 40 min and 1 h, respectively. The nanofluids showed good stability for more than three months. Albert et al. [66] used the sol–gel method using PVA as a surfactant to prepare CuO/water nanofluids. The nanofluids were ultra-sonicated for 1.75 h and remained stable for more than a year. Yang et al. [67] investigated TiO₂/water and Ag/water nanofluids for heat recovery using the two-step method of nanofluid preparation. The particles were prepared using the post-treated optimised preparation method, which involves separating the agglomerated particles and studying the stability and heat recovery of the remaining concentration of nanoparticles. The nanofluids prepared using this method were stable for more than six months.

In the single-step method, both steps, (i) formation of particles and (ii) dispersing them in the base fluids, happen simultaneously. In this method, all the intermediate processes such as storage, drying, dispersion of particles and transportation are subtracted or avoided, so the accumulation of nanoparticles decreased, and stability of nanofluids is maximised. Huang et al. [68] adopted the single-step method to prepare ethyl carbamate-modified Field’s alloy/polyalphaolefin (PAO) oil nanofluids using the nano-emulsification technique. In this technique, the Field’s alloy nanoparticles are heated up to 180 °C in the PAO oil under stirring for 3 h. The nanofluids were considered ready by observing the colour change from white to dark grey. According to the UV–Vis spectra results and zeta potential values, the nanofluids remain stable for more than 3 h.

Moreover, Du et al. [69] considered the one-step method to produce Gn/water nanofluids based on the shear exfoliation of graphite with the help of PVA surfactant. In this method, the graphite was converted to graphene within the DI water by six exfoliation processes. The authors observed a stable solution for more than 180 h by monitoring the UV–Vis spectra. Additionally, Li et al. [70] prepared silver and gold water-based nanofluids using the one-step method. The authors examined the effect of spacer length of different Gemini surfactants on the stability of silver and gold nanofluids. Furthermore, silver and gold solutions were added to the pre-prepared surfactant solution with the addition of glucose and sodium hydroxide under vigorous stirring till a red colour was observed. According to their stability study, the solutions remained stable for more than 60 days.

The one-step method can be used only in small-scale production. Using this method can produce well uniform dispersed nanofluids, although it is costly, and not all types of nanofluids can be produced by this method. For the readers’ convenience, the authors have summarised some of the works on nanofluids preparation in the year 2019 in Table 2. The table summarises the nanoparticles types, base fluids and the methods used for the nanoparticles and nanofluids synthesis, characterisation and stability measurement.

Table 2 Review of some 2019 studies on the synthesis, preparation, characterisation and stability methods of nanofluids

Properties of various nanofluids

The thermophysical behaviour of nanofluids directly affects the application of nanofluids, especially concerning heat transfer. Properties such as the viscosity, thermal conductivity, density and the specific heat capacity of the fluid are all vital in determining the effectiveness of nanofluids as heat transfer fluids. Experimentally, certain techniques and standards have been used in measuring the various thermophysical properties of nanofluids.

The three main techniques in measuring the thermal conductivity of nanofluids are the transient technique, steady-state technique and the thermal comparator [98]. The transient technique is more accurate and reliable than the steady-state technique, as it completely reduces the effect of natural convection and radiation [99]. Rotational viscometers are the main type of instruments used to measure the viscosity of nanofluids. These viscometers do not only measure the viscosity of the fluid, but they can also determine the rheological behaviour of the fluids. On the other hand, the specific heat capacity of nanofluids is usually measured by applying the differential scanning calorimetry (DSC) as it is easy to use, provides adequate accuracy and the measurement times are short. While there have been many studies that measured nanofluids thermophysical properties using the outlined techniques, many other researchers have focused their studies on proving an accurate theoretical model to predict the thermophysical behaviour on nanofluids.

Before the invention of nanofluids, several scientists have theorised the effect of particle dispersion on the thermophysical properties of conventional heat transfer fluids. As previously stated, J. C. Maxwell theorised the thermal conductivity of particle dispersions in a liquid in 1881 [17]. Later, Einstein theorised the dynamic viscosity of particle dispersions in a liquid in 1905. Both research endeavours represent the earliest postulations of the thermophysical behaviour of suspensions in fluids [100]. Several other researchers have proposed models to predict these thermophysical behaviours of suspensions in fluids. Theoretical models developed before the classification of nanofluids are known as classical models. Table 3 illustrates the classical models for predicting the thermophysical properties of nanofluids. While the classical models were accurate to a limited range, the models often did not predict the values of thermal conductivity, viscosity and specific heat capacity of the nanofluid with high enough levels of accuracy. This is mainly because, in the nanoscale, previously unconsidered factors appear to affect the thermophysical properties of fluid dispersion, while most classical models considered volume concentration and the base fluid property as the most important variables in determining the thermophysical properties as Einstein described it as “small rigid spheres suspended in a liquid”. However, research deduced that nanofluids thermophysical properties are affected by a broader range of variables which include particle size [101], volume concentration, base fluid property, particle agglomeration, packing fraction, fluid pH, nanolayers, particle distribution, temperature and mixture ratio (hybrid nanofluids). The inability of the classical models to factor in these variable conditions limits its application to only a narrow range of values.

Table 3 Classical formulas for the thermal conductivity and viscosity of dispersions in fluids

The physical properties of the nanofluids are an important parameter in predicting the heat transfer and friction factor behaviours of individual nanofluids. Mahian et al. [108], summarised the computational methods for solving the thermal transport model for nanofluids flow. These methods include the finite differential method, finite volume method, finite element method, lattice Boltzmann methods and other Lagrangian methods such as dissipative particle dynamics and molecular dynamics. Based on experimental results, Dadhich et al. [109] used the artificial neural network to develop correlations used to predict the heat transfer coefficient of Al₂O₃ and TiO₂ water-based nanofluids flowing in an annulus at 1 bar. The three input parameters used in their study were nanoparticle concentration, heat flux and mass flux. Their results show that both nanofluids performed better than water. At a nanoparticle concentration of 0.2%, Al₂O₃ nanofluid had enhancement of 155.24% and TiO₂ nanofluid had a heat transfer coefficient increase of 71.56% as compared to that of water. The heat transfer and friction factor behaviour hybrid nanofluids have also been investigated. Yang et al. [110] investigated the dynamic stability, sedimentation and time-dependent heat transfer characteristics of TiO₂ and CNT nanofluids. They discovered that for a volumetric concentration of 0.3% TiO₂ and 0.1% CNT nanofluids, the convective heat transfer coefficient is increased by 17.84% and 19.31%, respectively. Hameed et al. [111], from experimental data, compared the heat transfer and pressure drop behaviour of alumina–CNT/water nanofluids and alumina–Cu/water nanofluids. The convective heat transfer enhancement of alumina–CNT/water was higher compared to alumina–Cu/water hybrid fluid. The maximum enhancements of 30.65% and 20.48% in Nusselt number were obtained at 0.3% volume fraction of alumina–CNT/water and alumina–Cu/water hybrid nanofluid, respectively. Their study provided experimental correlation for both fluids.

Thermal conductivity of nanofluids

Yu et al. [112] conducted a review over a decade ago to compare the thermal conductivity of nanofluids and their convective heat transfer enhancement. They concluded that from 107 works of literature surveyed, a 15–40% enhancement was recorded with the oxides nanoparticles available back then. Today, numerous research investigating the use of different nanoparticles on the thermal conductivity has been conducted. Esfe and Afrand [113] extensively covered the thermal conductivity of nanofluids that predate 2019. In 2019, as it has been the trend, a majority of the studies focused on hybrid nanofluids. Akhgar et al. [114] experimentally measured the thermal conductivity of MWCNT–TiO₂/ethylene glycol nanofluid and obtained that an increase in volume concentration of the nanofluid tends to increase the thermal conductivity of the hybrid nanofluid. The study considered volume concentration between 0.05 and 1% and observed that at a volume concentration of 1%, maximum enhancement (40.1%) in thermal conductivity was obtained. The study also developed an artificial neural network (ANN) model to predict the thermal conductivity values obtained from the experiment. Also, Alarifi et al. [115] developed an ANN model from their experimental results to predict the thermal conductivity of MWCNT–TiO₂/thermal oil nanofluid.

As shown in Table 4, while many studies have begun to use the artificial neural network (ANN) models in thermal conductivity prediction, some others have proposed regression-based correlation equations to fit results obtained from their experiments. Moldoveanu et al. [116] conducted an experimental study on the thermal conductivity variation of Al₂O₃–TiO₂/water nanofluid at volume concentration between 0.25 and 1% and proposed a correlation model to predict the thermal conductivity.

Table 4 Thermal conductivity studies of hybrid nanofluids in 2019

In terms of unique conventional nanofluids, Essajai et al. [117] studied the effect of particle shape on the thermal conductivity of nanofluids. The study was performed using a one-dimensional (1-D) network of interconnected gold nanoparticles (IAuNPs) and spherical Au nanoparticles. It was observed that IAuNPs in base fluids were more effective in improving the thermal conductivity of nanofluids than spherical Au particles suspended in a base fluid. Applying the one-step synthesis technique, the stability and the thermal conductivity measurement of MWCNTs/Jatropha seed oil nanofluid were investigated [118]; this environment-friendly nanofluid showed a thermal conductivity enchantment of 6.76% at a mass concentration of 0.8%.

ANN has also been used to predict the effect of particle aggregation on the thermal conductivity of nanofluids [119]. Mirsaeidi and Yousef [120] used ANN to predict the thermal conductivity, density and viscosity of carbon quantum dots nanofluids using water, ethylene glycol and EG–water (60:40) as base fluids. Motlagh et al. [121] used gene expression programming to propose a correlation that estimates the thermal conductivity of Al₂O₃ and CuO–water-based nanofluid based on experimental data from the literature. Going forward, it is expected that there will be an increase in research and development exploring the possible advantages of using ternary hybrid nanofluids. In this regard, the study conducted by Mousavi et al. [122] has already demonstrated that the thermal conductivity of CuO–MgO–TiO₂/water nanofluid is enhanced by 78.6% at a mass fraction of 0.1. For the reader’s convenience, the authors have summarised the work on thermal conductivity studies for hybrid nanofluids in the year 2019 in Table 4.

Viscosity of nanofluids

The viscosity of a fluid is important in understanding both the heat transfer and the flow behaviour of the fluid. Several experimental studies have been carried out to understand the behaviour of nanofluids. The available research on the topic is not limited to experiments alone as molecular dynamics simulations have also been used to explain the viscosity of nanofluids [136]. Dehghani et al. [137] analysed the effect of temperature and mass fraction of Al₂O₃ and WO₃ nanoparticles in water and liquid paraffin. Their findings showed that the viscosity of both nanofluids is increased only by adding a certain number of nanoparticles to both fluids. Regarding the shear rates, the viscosity of water-based nanofluids is constant, which indicates a Newtonian behaviour, while that of paraffin does not remain constant at different shear rates, and at a low amount of shear rate the viscosity achieves higher value, indicating a non-Newtonian behaviour for liquid paraffin-based nanofluids. Finally, they presented a correlation based on temperature, nanoparticle concentration and the physical properties of both the nanoparticle and base fluid for predicting the viscosity of aqueous and non-aqueous nanofluids. Ye et al. [138] extensively covered the viscosity of nanofluids that predate 2019. In 2019, as with thermal conductivity studies, there has been a trend towards hybrid nanofluids. The significance of viscosity in lubrication applications has been seen in many investigations related to oil-based nanofluids. Using the ultrasonic-assisted process, Barai et al. [85] synthesised graphene oxide–Fe₃O₄/water nanofluid at volume concentrations between 0.01 and 0.2%. The study obtained a maximum viscosity enhancement of 41%. Studying Fe–CuO/EG–water nanofluid, Bahrami et al. [139] obtained that the backward propagation methods presented the least error in predicting dynamic viscosity. Bahrami et al. [139] deduced that when the hybrid nanofluids volume concentration is below 0.1%, the Fe–CuO/EG–water nanofluid exhibited Newtonian behaviour. However, when Fe–CuO/EG–water nanofluid volume concentration is above 0.25% the behaviour of the fluid changed.

As shown in Table 5, while many studies have proposed various correlation models to predict the viscosity behaviour of nanofluids, the more accurate models proposed are the artificial neural network (ANN) models. Ruhani et al. [140] investigated the effects of volume concentration and fluid temperature on the viscosity of hybrid nanofluids. The correlation model proposed in this study demonstrated a 1.8% margin of deviation between experimental values and correlation results. Viscosity enhancement was about 80% when the volume fraction was 2%.

Table 5 Viscosity studies of hybrid nanofluids in 2019

Other types of conventional nanofluids were also studied; Mousavi et al. [141] conducted an experimental investigation into the viscosity measurements of MoS₂/diesel oil nanofluid at particle concentration between 0.1 and 0.7%. The study observed that the viscosity increased by 7.04% when volume concentration was 0.7%. Hameed et al. [142] synthesised an eco-friendly MWCNTs-Kapok seed oil nanofluid using a one-step method, at a constant nanoparticle concentration of 0.1%.

Considering all of the experimental viscosity measurements conducted, the relationship between viscosity and both temperature and particle concentration is apparent. Naturally, the viscosity of nanofluids increases with an increase in particle concentration, and this is observed in virtually all measured experiments. The viscosity of nanofluids decreases with an increase in temperature, and this is also observed in all measured experiments; this behaviour is expected as entropy is increased as particles gain thermal energy. However, the relationship between particle concentration and rheology is not as apparent. Considering the sample size alone as illustrated in Table 5, it can be observed that there exists no clear pattern between rheological behaviour and particle concentration in nanofluids. Rheological behaviour appears to vary from material to material.

Specific heat of nanofluids

The specific heat capacity of fluids is important in understanding both the heat transfer and the energy content of thermal systems. While significant research has focused on both viscosity and thermal conductivity, studies relating to the specific heat capacity of nanofluids are not as advanced. However, the specific heat capacity of fluid bears significance in thermal storage applications. Therefore, many studies regarding specific heat capacity often use molten salt as their base nanofluid. Moldoveanu and Minea [153] experimentally measured the specific heat of both Al₂O₃–TiO₂/water nanofluids and Al₂O₃–SiO₂/water nanofluids at volume concentration between 1 and 3.0%. A correlation model was determined from the measured specific heat capacity values. It is important to note that the correlation model had an average deviation of 11% when compared to experimental specific heat values. However, when the mixture theory model was used to predict the nanofluids’ specific heat capacity values, the deviation was as high as 19%.

The effect of particle size and volume fraction on the specific heat of SiO₂ molten salt nanofluid was investigated by Li et al. [154]. Using SiO₂ nanoparticle with sizes of 10, 20, 30 and 60 nm, SiO₂ molten salt nanofluid was synthesised at particle concentration between 0.5  and 2%. Addition of particles to molten salt increases the specific heat capacity for all of the volume concentrations and particle sizes considered. An important point to note is that the particle concentration and particle size with the most specific heat enhancements were 1% and 20 nm, respectively.

Using SiO₂, Al₂O₃ and TiO₂ nanoparticles, three conventional nanofluids were synthesised by Hassan and Banerjee [155]. The study aimed to predict the specific heat capacity of metal oxide molten nitrate salt nanofluids using a multilayer perceptron neural network (MLP-ANN). The ANN model proposed was more accurate when compared to classical prediction methods [155]. Alade et al. [156] also considered a machine learning approach by applying a support vector regression model optimised with a Bayesian algorithm to predict the specific heat capacity of Al₂O₃ ethylene glycol nanofluids. The proposed model also exhibited a high degree of accuracy with the root-mean-square error (RMSE) equivalent to 0.0047.

From Table 6, while the specific heat of molten salts increases with the addition nanoparticles, in experiments involving MWCNTs PEG 400 nanofluid, Al₂O₃–water nanofluid, Fe–water nanofluid and Al₂O₃–Fe nanofluid the specific heat of the base fluid exceeds that of the nanofluids.

Table 6 Some specific heat studies of nanofluids in 2019

Factors affecting nanofluids stability and thermophysical properties

The main factors affecting the thermophysical properties of nanofluids includes the morphology and concentration of nanoparticles, aggregation in the nanofluids and the sonication time used in its preparation [158]. The stability of nanoparticles suspended in a fluid is a very important parameter that affects both the rheological and thermophysical behaviours of the resultant nanofluids. Brownian motion causes the particles to collide with one another leading to cluster formation in the base fluid. These cluster formations or aggregation are controlled by a variety of internal forces between the base fluid and the nanoparticles such as the Van der Waals forces of attraction between the particles [159]. The aggregates begin to crystallise as their density exceeds that of the base fluid and affects the stability of the nanofluids over time [152]. Some of the factors that affect the stability of the nanofluids include the method of preparation of the nanofluids [66], agitation and sonication time [160,161,162], pH of the nanofluids [152], the addition of surfactants [163, 164] and surface charge density of the nanoparticles [158]. Asadi et al. [165] reviewed the effect of sonication on the stability and thermophysical properties of nanofluids. The study concluded that while there exists an optimum sonication time where thermal conductivity is maximum, and viscosity is least, more research is required to determine this optimum value, as it appears to differ for different nanofluids. Khan and Arasu [166] also reviewed the effects of nanoparticle synthesis techniques on the stability and thermophysical behaviour of nanofluids. The study importantly noted that there appears to be no standard method for stability measurements; this makes it difficult to compare stability across different papers. This is a problem because of the significant differences in reported fluid stability; this can range from days in some studies to months in others.

The values of the thermophysical properties of nanofluids are sensitive to the volume and size of nanoparticles used, the temperature of the mixture and the use of surfactants [167]. Yang et al. [168] investigated the thermal conductivity of graphene oxide/water nanofluids with a mass concentration range of 0–1.5%. Their result showed that as the mass fraction of nanoparticles increased, the thermal conductivity enhancement increased. Also, at a pH of 8, the nanofluids showed maximum stability with a maximum thermal conductivity enhancement of 48.1%. This indicated that the pH was a significant parameter in both its stability and thermal conductivity. The authors attributed the thermal enhancement observed to the increased Brownian motion of particles and molecules of the base fluid as temperature increased. Yang et al. [169] also studied the thermal conductivity behaviour of zinc nanopowder in SAE 50 engine oil and recorded an increase in the thermal conductivity of the nanolubricant as the volume concentration of nanoparticles was increased. They recorded a maximum thermal conductivity enhancement of 8.74% and attributed this to the effects of increased Brownian motion of particles in the lubricant as temperature raises. The thermophoresis effect was another factor they highlighted that affected the thermal conductivity enhancement.

Rostami et al. [71] examined the thermal conductivity of GO–CuO water/EG (50:50) hybrid nanofluid at a temperature of 25–50 °C and particle volume concentration of 0.1–1.6%. Their investigation observes a 46% enhancement in thermal conductivity, which is higher than the enhancement of using single nanomaterial. Mahyari et al. [73] investigated the thermal conductivity GO/SiC (50:50)/water hybrid nanofluid at volume concentrations between 0.05 and 1%. Their investigation reveals that the effect of the volume concentration of nanoparticles was more significant than the effect of increasing temperature. Importantly, the studies observed that the enhancement in thermal conductivity of their hybrid nanofluid was more than the reported thermal conductivity enhancement using GO or SiC individually. Hybrid nanofluids not only affect the thermal conductivity, but also enhance the stability of nanofluids.

Heat transfer mechanisms of nanofluids

The main benefit of using nanofluids is their enhanced thermal transport which results in improvements in the thermal conductivity of traditional heat transfer fluids. As previously outlined, several parameters influence the thermal conductivity enhancement and include nanoparticle type, nanoparticles size, nanoparticles concentration, temperature, type of base fluid and the thermophysical properties of both the base fluid and the nanoparticles. Over the last three decades, since the introduction of nanofluids in 1995, the explanations behind the enhanced heat transfer of nanofluids have been attributed to several mechanisms. The size and the large number of particles interacting with the base fluid present a challenge to properly understanding the nanoscale effects that support the improved thermal properties observed in the literature. Mahian et al. [108, 170] studied the mechanisms that would aid the simulation of nanofluids flow. They highlighted that forces such as drag, lift, Brownian motion, thermophoresis, Van der Waals and electrostatic double-layer forces had a significant effect on the thermal and rheological behaviours of nanofluids.

Brownian motion is defined as the uncontrollable random motion of particles within the fluid due to the collision between slow moving and higher velocity particles. Brownian motion occurs as a result of thermal diffusion, and this phenomenon is increased at higher temperatures, low viscosity and smaller particle size. As promoted by the scientific community, the random collision of particles within the fluid remains the primary reason for the thermal conductivity enhancement observed with nanofluids [73, 79, 92]. However, Jang and Choi [171] provided three types of collisions that occur due to the rising temperature of nanofluids: collisions between the molecules of the base fluid, collisions between base fluid molecules and the nanoparticles, and the collisions between nanoparticles due to Brownian motion. They concluded that the effect of Brownian motion on thermal conductivity enhancement had the least effect among the three types of collisions.

Keblinski et al. [172] was the first to introduce the idea of nanolayers and their effect in nanofluid thermophysical behaviour. The nanolayer is known as the solid-like structure or the interfacial layer between the solid surface and the first layer of the fluid in contact with the solid surface. A structured, layered arrangement of the fluid molecules around the surface of the nanoparticles was observed. These layers behaved like solids and act as a thermal bridge for the heat transfer process enhancing the overall thermal conductivity of the fluid. In the solid–solid interface, this layer acts as a barrier of heat transfer due to incomplete contact between solid surfaces. However, it is not the case for the solid–liquid interface as the aligned interfacial shell in the nanoparticle suspension would make heat transfer across the interface effective. Yu and Choi [173] presented a modified Maxwell model to account for the effect of nanolayers on the thermal conductivity of nanofluids. Their results proved that the thermal model is enhanced as a result of accounting for this factor. Xie et al. [174] investigated the effect of the nanolayer on the effective thermal conductivity of nanoparticle–fluid mixtures. It was observed that the effective thermal conductivity increases with a decrease in particle size and an increase in nanolayer thickness. It was concluded that manipulating the nanolayer structure might be an effective method to produce higher thermally conductive nanofluids.

Another factor responsible for nanofluid thermophysical behaviour is the “particle nanoclusters”. It should be noted that nanoparticles have strong Van der Waals interactions that force them to form nanoclusters, which lead to a rich zone of high thermal conductive nanoparticles that improve the bulk thermal conductivity of the fluid. However, increasing the size and mass of nanoclusters will result in nanoparticle sedimentation, which will eliminate its effect on thermal conductivity. Keblinski et al. [172] also suggest that at high loading of nanoparticles, the effect of nanoclusters was promoted due to an increase in Van der Waal’s force of attraction among the nanoparticles.

The effect of the diffusive/ballistic nature of heat transport and thermophoresis has been reported [172]. Thermophoresis is related to thermal diffusion due to the temperature gradient. It describes the movement of the nanoparticles due to the temperature gradient from the high-temperature zone to a lower-temperature zone which could influence the thermal conductivity. Thermophoresis is different from Brownian motion as the whole movement in thermophoresis is one-directional. The diffusive/ballistic nature of heat transport is an explanation for the heat transfer in crystalline solids. In solid media such as the nanoparticles used in nanofluids, the heat is transported by phonons. The thermal conductivity is significantly enhanced if a particle was influenced by a phonon that is created in another nearby particle that exists in the same liquid. This is because the mean free path of the phonon is shorter in the liquid than it is in the particles. However, the effect of Brownian motion, nanolayer and nanoclusters on thermal conductivity enhancement is more significant and the reasons mostly reported by the authors in 2019. These mechanisms have all been discussed in greater detail in review studies on heat transfer mechanisms in nanofluids [175,176,177,178,179].

Application of nanofluids in various thermal devices

Nanofluids in solar thermal collectors

Solar collectors are used in converting the radiant energy of the Sun to thermal or electrical energy, benefiting from radiative, convective and conductive heat transfer principles. The solar irradiance from the sun is absorbed by the collector with the aid of a working fluid flowing within its absorber. The common fluids used for thermal energy absorption are water, oils, ethylene glycol (EG) and salts. These working fluids have limitations that affect the overall efficiencies of various collectors. Their main limitation is in its low thermal conductivity. To obtain higher thermal conductivity, nanofluids have been proposed and tested for use in the various solar collectors. This section reviews the progress in the application of nanofluids in various solar collectors. Figure 4 presents a classification of different solar collectors that can use nanofluids as heat transfer fluids.

Fig. 4

Classification of solar collectors where nanofluids can be applied

Flat plate collector (FPC)

The flat plate collector is the most widely used solar collectors. It is a rectangular tray consisting of an absorber surface (plate) with copper tubes (raiser) positioned along its surface. An insulating material placed at the backside helps reduce heat loss due to conduction, and a glass or transparent glazing over the top of the collector helps minimise radiative and convective heat losses. Figure 5 presents a schematic representation of the flat plate collector. To enhance the efficiency of the collector, conventional fluids have been replaced with nanofluids. For instance, Choudhary et al. [180] investigated the stability of MgO nanofluids for use in a flat plate solar collector, considering the effect of volumetric concentrations between 0.08 and 0.4% on the stability of the nanofluid over time. The study demonstrated that the nanofluids achieved better stability at 0.04% volumetric concentration. Upon testing the nanofluids in the flat plate collector, the maximum thermal efficiency of 69.1% was achieved at a 0.2% volumetric concentration and 1.5 lit min⁻¹. This value represents a 16.36% enhancement in thermal efficiency when compared to EG/water.

Fig. 5

Schematic of flat plate solar collector

Ahmadlouydarab et al. [181] investigated the thermal absorption ability and the overall thermal efficiency of a flat plate collector using TiO₂–water nanofluids as an agent fluid in the outer part of the absorber of a flat plate collector. In this design, the nanofluids act as thermal insulation by utilising the high thermal capacity of these fluids. Furthermore, the TiO₂ nanoparticle was used on the outer part of the glass cover of the collector to enhance the self-cleaning properties of the glass surface. The study concluded that the new system design enhanced the thermal efficiency of the collector by 49% at a 5% nanoparticle volumetric concentration.

Saffarian et al. [182] using Al₂O₃ and CuO–water nanofluids investigated the effect of a change in the flow direction of the flat plate collector using modified U-shape, spiral and wavy pipes. The study demonstrates that the heat transfer coefficient increased by using nanofluids instead of water. The wavy and spiral geometries significantly improved the heat transfer; however, higher pressure losses were witnessed with the use of the wavy pipe. The study concluded that the use of the wavy pipes along with CuO nanofluids at 4% volume concentration increased the heat transfer coefficient by 78.25%.

Tong et al. [183] experimentally analysed the thermal performance of a flat plate collector using Al₂O₃ and CuO nanofluids. It was demonstrated that with a 1% volume concentration of Al₂O₃, the highest thermal efficiency enhancement of 21.9% was achieved. Furthermore, exergy efficiency enhancements of 56.9% and 49.6% were recorded when compared to water using Al₂O₃ at 1 Vol% and CuO at 0.5 Vol%, respectively.

Mondragon et al. [184] tested the performance of a flat plate collector under laminar flow conditions using Al₂O₃–water nanofluids. The study demonstrated that at a 1% volume concentration of Al₂O₃ in the nanofluids, a 2.3% increase in heat transfer coefficient could be theoretically attained. However, when testing for the collector efficiency, the study observed a decrease in the collector’s efficiency from 47% using water to 41.5% when using the Al₂O₃–water nanofluids. The decrease was attributed to the formation of nanoparticle deposition layers on the absorber tube; these layers acted as an additional form of resistance to heat transfer. The authors attributed the formation of these layers to the low flow velocity of the nanofluids.

Evacuated tube solar collector (ETSC)

This type of collector is more efficient than the flat plate collectors as heat losses in the ETSC are reduced when compared to the FPC due to the presence of vacuum insulation. A vacuum between the glass tube and the evacuated tube heat pipe helps to reduce losses due to convection and conduction. The heat pipe within the tube contains an antifreeze liquid in a closed system. This pipe then extends into the manifold where the liquid flowing in the manifold condenses the antifreeze and is then returned to be heated by the heat pipe. A pictorial depiction of this collector is shown in Fig. 6.

Fig. 6

Schematic diagram of an evacuated tube solar collector

The use of nanofluids to enhance the performance of this collector has been investigated. For instance, Sarafraz et al. [185] evaluated the performance of an evacuated tube solar collector working with a carbon acetone mix in the heat pipe. The results demonstrate that the thermal efficiency of 91% was achieved, which is above that of the average thermal efficiency of 72.6% when using acetone alone.

Natividade et al. [186] experimentally evaluated an ETSC using multilayer graphene (MLG)-based water nanofluids. The ETSC was equipped with parabolic concentrators. The MLG at concentrations of 0.00045 Vol% and 0.00068 Vol% increased the thermal efficiency of the collector by 31% and 76%, respectively, when compared to the base fluid. Sadeghi et al. [187] also used a parabolic concentrator to enhance the performance of an ETSC operating with Cu₂O–water nanofluid. The experimental set-up was verified using an ANN multilayer perception model. The maximum thermal efficiency of 60% was attained at a flow rate of 50 litres/hour and 0.08 Vol% of Cu₂O. This value represented an 87.5% enhancement in the collector’s performance when compared to water.

Compound parabolic collectors (CPC)

CPCs are similar to flat plate collectors but have parabolic optics attached to each tube, which concentrates incident solar radiation onto the absorbers. Similar to flat plate and evacuated tube collectors, CPCs can be static while collecting diffuse solar radiation. There are four kinds of CPCs: flat one-sided absorbers, flat two-sided absorbers, wedge-like absorbers and tubular absorbers as shown in Fig. 7. A tubular absorber contains a parabolic collector surface and an absorber tube. Korres et al. [188] investigated nanofluids-based CPC under laminar flow regime. The study demonstrated a mean and maximum heat transfer coefficient enhancement of 16.16% and 17.41%, respectively. Factoring in the effect of pressure losses as a result of using nanofluids, the study concluded that the pressure drops observed were not a limitation to the use of the nanofluids and recorded a thermal efficiency enhancement of 2.76% when using CuO/Syltherm nanofluids.

Fig. 7

Cross section of CPC tubular absorber

Linear Fresnel reflectors (LFR)

A linear Fresnel reflector is a concentrating solar collector characterised by its ease of assembly; this makes it cheaper when compared to other concentrating solar collectors. As shown in Fig. 8, the LFR utilises mirrors whose orientation revolves around a pivot following the Sun in other to concentrate its rays towards the absorber tube [189]. This system can produce thermal energy for medium- to high-temperature applications. However, LFRs are not as widely installed collectors; therefore, there are not as many studies applying the collector with nanofluids. Ghodbane et al. [190] performed a study to assess the performance of MWCNT–water nanofluid in the LFR. The outcomes indicate that MWCNT at 0.3 Vol% resulted in a more favourable thermal efficiency of 33.81% when compared to other fluids tested and resulted in the highest pressure loss of 2.3–46 Pa. The use of the nanofluids also demonstrated a reduction in the rate of entropy generation within the system.

Fig. 8

Diagram of a linear Fresnel reflector

Parabolic trough solar collectors (PTSC)

Parabolic trough solar collectors are the most commercially deployed and studied concentrating solar collectors available. As depicted in Fig. 9, the collector utilises a parabola-shaped mirror to reflect the solar radiation from the Sun onto a cylindrical receiver. The receiver comprises a concentric absorber tube enveloped with a glass cover. Solar radiation absorbed by the receiver is transferred to the working fluid passing through it, and it is then transported to applications requiring medium to high temperatures (50 °C–400 °C). Okonkwo et al. [191] synthesised zero-valent iron and TiO₂ nanoparticles from olive leaf extracts for use in a solar parabolic trough collector. The nanoparticles were used to prepare nanofluids with Syltherm-800 as base fluid. The use of Syltherm-800/TiO₂ and Syltherm-800/ZVI produced a 42.9% and 51.2% enhancement in the heat transfer coefficient at a 3% nanoparticle volume concentration. Although the use of the nanofluids resulted in an 11.5% drop in pressure, the authors stated that a thermal efficiency enhancement of 0.51% and 0.48% was still achieved while using Syltherm-800/ZVI and Syltherm-800/TiO₂ nanofluids, respectively. Ehyaei et al. [192] examined the energy, exergy and economic analysis of a PTSC operating with water and Therminol VP1 as working fluids. These fluids were also used as base fluid with the addition of CuO and Al₂O₃ nanoparticles. The annual efficiency of the PTSC was taken with all four working fluids, and the results indicate that the annual energy and exergy efficiency of water was 10.64% and 9.07%, while the addition of Al₂O₃ and CuO in water at 5% volume concentration only increased the efficiency of the PTSC by 0.03% and 0.09%, respectively.

Fig. 9

Schematic of a parabolic trough solar collector

Malekan et al. [193] investigated the heat transfer in a PTSC working with Fe₃O₄ and CuO/Therminol-66 nanofluids under an external magnetic field. The results demonstrated that by increasing the nanoparticle concentration, the heat transfer in the collector was enhanced. The maximum heat transfer enhancement observed in Fe₃O₄/Therminol-66 nanofluids was at 4% volume concentration and nanoparticle size of 10 nm. The presence of a magnetic field enhanced the performance of Fe₃O₄/Therminol-66 more than that of CuO/Therminol-66 nanofluid although the CuO nanoparticles had a better thermal conductivity.

Bellos and Tzivanidis [194] evaluated the performance of an LS-2 collector with six different nanoparticles (Cu, CuO, Fe₂O₃, TiO₂, Al₂O₃ and SiO₂) using Syltherm-800 as base fluid. The nanoparticle concentration was varied up to 6%, and their result showed that SiO₂ nanoparticle provided the least enhancement in thermal efficiency with 0.19%, while Cu with 0.54% provided the highest enhancement at a 4% volume concentration as shown in Fig. 10. Other nanoparticles CuO, Fe₂O₃, TiO₂ and Al₂O₃ had an enhancement of 0.46%, 0.41%, 0.35% and 0.25%, respectively.

Fig. 10

Thermal efficiency and overall heat transfer coefficient enhancement at 4 vol% [194]

Direct absorption solar collector

The direct absorption solar collector (DASC) is a concentrating solar collector with fewer thermal resistance when compared to regular solar collectors. By removing the absorbing surface, the working fluid can absorb solar radiation directly. As illustrated in Fig. 11, the conductive and convective resistance as a result of the use of a surface absorber is eliminated, making the efficiency of the system dependent on the absorptivity and thermal properties of the working fluid. This modification reduces the thermal losses in the system. Qin et al. [195] stated that the direct absorption solar collectors are 5–10% more efficient than the regular parabolic trough collector. However, the challenge with these systems remains the low absorption properties of the working fluids. The use of nanoparticles dispersed in these working fluids can, however, improve the performance of the collector [30]. Tafarroj et al. [196] investigated the use of SiO₂/EG and MWCNT/EG nanofluids in a direct absorption solar collector. The outcomes suggest that at 0.6% volume concentration of nanoparticles, MWCNT/EG nanofluids provided the highest outlet temperature of 346.1 K. Simonetti et al. [197] performed a CFD study on direct volumetric absorption solar collector operating with SWCNT/EG nanofluids and compared its performance with a DASC integrated with a compound parabolic collector. The study concluded that the DASC performed better than the direct volumetric absorption solar collector.

Fig. 11

Thermal resistance network for a regular solar collector, b direct absorption collectors

Photovoltaic thermal collectors (PVT)

The cells of photovoltaic (PV) systems are affected negatively by high temperatures (˃25 °C), as the excess heat received from the Sun reduces the efficiency of the PV module. Technologies such as the hybrid PVT system have been developed to extract this heat for possible utilisation in other thermal applications, while also enhancing the electrical output of the PV module. Evident from Fig. 12, the excess heat absorbed by the cells is transferred to a heat transfer fluid which cools the collector and provides heat for use in other thermal applications.

Fig. 12

Hybrid nanofluid-cooled PVT

Sangeetha et al. [198] experimented to determine the performance of a hybrid PVT system utilising different nanoparticles dispersed in water. The study evaluated the performance of MWCNT, Al₂O₃ and CuO in water and demonstrated that nanofluids improved the electrical efficiency of the PVT when compared to water. The use of MWCNT and CuO nanofluids decreased cell temperature by 19%. MWCNT, Al₂O₃ and CuO nanofluids enhanced the electrical efficiency of the PV by 60%, 55% and 52%, respectively. Similarly, Alous et al. [199] investigated the performance of MWCNT and graphene nanoplatelets (GNPs) dispersed in water as coolant in a PVT system. The study concluded that the addition of the thermal module improved the exergetic efficiency of the system by 53.4% using water, 57.2% using MWCNT–water and 63.1% using GNP–water. An 18.6% enhancement in energy efficiency was recorded with the use of GNP–water nanofluids in the PVT collector. This represented the highest observed enhancement in energy efficiency in their study. Fudholi et al. [200], on the other hand, examined the use of TiO₂ water nanofluids on a PVT. The study concluded that at a mass concentration of 1%, the TiO₂ nanofluid recorded an 85–89% performance enhancement when compared to water with 60–76% at a mass flow rate of 0.0255 kg/s. Abdelrazik et al. [201] studied the effect of optical filtration along with nano-enhanced phase change materials (PCMs) on the performance of a PVT collector, demonstrating that optical filtration, and the use of nano-PCM, increased the overall efficiency of the collector by 6–12%. A combined PVT/PCM system using nanofluids has proved to be an effective coolant in enhancing the thermal conductivity of PVT collectors [202]. Other studies related to nanofluids in solar collectors investigated the photothermal properties of various mono and hybrid nanofluids [203,204,205,206,207,208,209,210,211,212], the impact of magnetic fields on the thermal performance of nanofluids in a solar collector [213], the forced convective behaviour of nanoparticles inside a solar collector [214] and more recently the application of ANN models for the prediction of nanofluids performance in solar collectors [215,216,217]. Other studies investigating the application of nanofluids in various solar collectors are presented in Table 7.

Table 7 Studies related to the application of nanofluids in various solar collectors

Nanofluids in heat exchangers

Heat exchangers (HX) are devices used for the transfer of heat between two or more fluids. The use of nanofluids in the different kinds of heat exchangers has been investigated and discussed below.

Double-tube heat exchanger

Double-tube heat exchange is a system is widely used in industries. This type of heat exchanger consists of two concentric tubes, as illustrated in Fig. 13. Researchers have investigated various methods of improving the efficiency of these heat exchangers. Some of these include modifications in dimension, design of much larger systems and the use of a more powerful pump.

Fig. 13

Double-tube heat exchanger

A novel method that has been recently promoted is the use of nanofluids [33]. Different nanofluids have been investigated for use in improving the performance of the double-tube heat exchanger. The performance of TiO₂–water nanofluid was experimentally investigated in a double-tube HX; the study observed that the heat transfer rate was improved by 14.8%; however, the pressure drop also increases by 51.9% [261]. The heat transfer coefficient in the double-tube HX was improved by 35% using MWCNT–water nanofluids [262]. The Al₂O₃/water nanofluid was also used in a double-tube HX, and the result demonstrated a more favourable thermal efficiency of 16% compared to pure water [263]. A study that investigated the heat transfer and pressure drop in the laminar flow regime using silver-coated silica demonstrated that the heat transfer coefficient could be improved from 7% to 50% [264]. The turbulent flow was also investigated for Al₂O₃/water nanofluid, and it was observed that the Nusselt number and Reynolds number increased by 23.2% and 32.23%, respectively [265].

Plate heat exchangers

The plate heat exchanger illustrated in Fig. 14 is a type of compact heat exchanger that is widely used in industries. The application of this type of heat exchanger has been recently spread in many industries. However, there is a need to improve thermal performance and efficiency; the use of nanofluids encourages a higher heat transfer rate within the same dimensions. Multiple studies investigate the use of nanofluids in plate heat exchangers [33]. The effects of using hybrid nanofluids on plate HX performance were numerically investigated [266], where heat transfer augmentation of approximately 16–27% was apparent for Al₂O₃–CuO/water nanofluid and Al₂O₃–TiO₂/water nanofluid, respectively. Using an experiment conducted on a plate HX with Al₂O₃/water nanofluid, new correlations for Nusselt number and heat transfer enhancement rates were derived [267].

Fig. 14

Image of a plate heat exchanger

A study investigating the heat transfer enhancement when fly ash nanofluids are used as the working fluid concluded that the heat transfer rate was improved by 6–20% as the concentration increases. The maximum enhancement was achieved using nanoparticle mass concentration of 2% [268]. The effect of particle size of metal oxide nanofluids on plate HX was experimentally investigated: Al₂O₃–water with particle sizes of 20 and 40 nm, TiO₂/water with a particle size of 10–25 nm and SiO₂–water with a particle size of 20–30 nm. When SiO₂–water nanofluid at a mass concentration of 0.2% was applied, the maximum heat transfer enhancement was achieved, while Al₂O₃–water nanofluid achieved the minimum heat transfer enhancement at a mass concentration of 0.1% [269]. The use of carbon-based nanofluids on brazed plate HX and its characteristic was investigated [270]. The results demonstrate a slight decrease in the pressure, while the heat exchange capacity and system efficiency factor were increased by 9.19% and 7.28%, respectively, at a mass concentration of 0.6%.

Shell and tube heat exchangers

The shell and tube heat exchanger is a type of heat exchanger that allows for larger surface contact when compared to other types of heat exchangers. It consists of a large outer tube which is the shell and bundles of inner tubes. Figure 15 illustrates a cross-sectional view of this type of heat exchanger. The rate of heat transfer with these heat exchangers is much higher due to their large contact area, although the low thermal conductivity of many of the heat transfer fluids used allows for the use of nanofluids with higher thermal conductivities. The heat transfer performance of carbon-based nanofluids on shell and tube HX was numerically investigated [271]. The study concluded that the nanofluid used improved the thermal performance; however, the pressure drop also increases as the particle volume concentration increased. The effects of using non-Newtonian metallic oxides nanofluids in the shell and tube HX energy-saving and effectiveness were experimentally investigated. Using Fe₂O₃, Al₂O₃ and CuO nanoparticles with water as the base fluid, the highest energy saving was achieved using CuO [272].

Fig. 15

Cross section of a shell and tube heat exchanger

A study investigating the heat transfer characteristics of TiO₂–EG nanofluids in a shell and tube HX determined that the heat transfer rate increases as the flow and volume concentration increases [273]. The study obtained the best volume concentration and flow rate for optimum heat transfer, where the best heat transfer rate achieved was 0.277 J at 0.075% nanoparticle concentration and a volumetric flow rate of 0.6 l/min. Said et al. [274] conducted an experimental and numerical study on the use of CuO/water as heat transfer fluid. The outcome demonstrates an increase in the heat transfer coefficient and convective coefficient by 7% and 11.39%, respectively. Moreover, a 6.81% reduction in the area could be achieved. The heat transfer improvement on the thermal performance of the shell and tube heat exchanger by the use of Al₂O₃/water and TiO₂/water nanofluids was studied. The maximum heat transfer coefficient enhancement achieved by Al₂O₃/water was 41%, while the maximum heat transfer coefficient using TiO₂/water was 37% [275]. Further studies related to the application of nanofluids in heat exchangers are presented in Table 8.

Table 8 Studies related to nanofluids application in heat exchangers

Nanofluids in electronic cooling

The advent of the miniaturisation of electronic devices and the need for effective heat management in such devices have pioneered new innovative research areas. The heat generated per unit volume of electronic devices has continued to increase, attributed to the flow of current through a resistance resulting in heat generation. The design of proper thermal management systems in such electronic devices is essential for the efficient and reliable operation of such devices. The use of microchannel heat exchangers in cooling electronic devices is one of the best options available. The channels are small, and as such, they increase the convective heat transfer from the electronic components. These types of heat sinks are used in the thermal management of devices such as supercomputers and batteries and are also used in data centres. The use of nanoparticles to enhance these microchannel heat exchangers has received much attention. The forced and natural convective heat transfer behaviours of nanofluids in various mediums have been studied. Such studies as the heat transfer behaviour of nanofluids in cavities [309,310,311,312,313], porous materials [314,315,316] and jet impinging [317] all show an increase in the dimensionless heat transfer parameter with the addition of nanoparticles. These studies prove the tremendous potentials of nanofluids in the electronics and data storage industries. Also, the heat transfer behaviour of nanofluids in magnetic fields has shown promising potentials [318].

Vishnuprasad et al. [319] experimentally evaluated the cooling performance of microwave-assisted acid-functionalised graphene (MAAFG) in water. The characterisation of the nanofluid showed that the MAAFG nanofluid had a 55.38% enhancement in thermal conductivity. The effect of varying the flow rate and nanoparticle volume concentration on the heat transfer coefficient and processor temperature was studied, and the results show that at 0.2 Vol%, there was an increase in the convective heat transfer coefficient by 78.5%. The processor temperature was also decreased by 15%, although a 5% pressure drop was recorded at 0.2 Vol% and a flow rate of 10 mL s⁻¹. Joy et al. [320] investigated the use of Cu–water and Al–water to increase the critical heat flux (CHF) limit in a heat pipe for electronic cooling. The result of the study demonstrated that nanofluids increased the CHF by 140% at a mass concentration of 0.01%. Both nanoparticle concentrations represented the optimum value of CHF for both nanofluids without preheating. Zing and Mahjoob [321] theoretically investigated the use of single- and multijet impingements through a porous channel for electronics cooling applications. The study evaluated the effect of two different coolants in their system: water and TiO₂–water nanofluids at a volume concentration of 5%. Results demonstrate that the use of TiO₂ nanofluid decreased the base temperature of the device more effectively than using water. For enhanced heat transfer in electronic cooling, Bezaatpour and Goharkhah [322] designed a mini heat sink with porous fins operating with a magnetite nanofluid of Fe₃O₄–water at volume concentrations up to 3%. The study recorded an increase in heat transfer of 32% with the use of the ferrofluids at 3 Vol% and Re of 1040. The pressure drop also recorded a decrease of 33% with the use of the ferrofluids.

Al-Rashed et al. [323] evaluated the first and second law performance of a non-Newtonian nanofluid of CuO and 0.5% carboxymethyl cellulose (CMC) in water for use in a microchannel heat sink (MCHS). Figure 16 illustrates an offset strip-fin MCHS with a description of its geometric parameters and imposed boundary conditions. By varying the nanoparticle concentration and Reynolds number, the effect of the nanofluids on the surface temperature of the CPU was observed. The results demonstrate that increasing Reynolds number adversely affected the frictional entropy generation and pressure drop. The nanofluid also reduced the surface temperature of the CPU and entropy generation rate in the system. A 2.7% decrease in the entropy generation rate of the CPU was attained at 1 Vol% and Re of 300. At 1 Vol% and Re of 700, the CMC/CuO water had an optimal ratio of heat transfer to the pressure drop of 2.29. Qui et al. [324] investigated the interfacial transport between vertically aligned carbon nanotube and electronic heat sinks. Their results show that CNT reduced the thermal contact resistances from 10 mm²K/W to 0.3 mm²K/W. Other studies related to the use of nanofluid in improving heat transfer in electronic devices are detailed in Table 9.

Fig. 16

The schematic of (a) offset strip-fin microchannels, and b the computational domain of the one unit of microchannels [323], [325]

Table 9 Studies related to the application of nanofluids in electronic cooling devices

Nanofluids in automobile radiators

The thermal management of automobile engines is necessary for the effective and efficient operation of the automobile. Figure 17 illustrates a schematic diagram of a car radiator which functions as a heat exchanger that disperses the heat generated from the operation of the engines. Recently, the use of nanofluids as alternative coolants in radiators have been investigated. Elsaid [341] experimentally investigated the performance of an engine radiator using nanofluids in the hot arid climate of Cairo, Egypt. Two nanoparticles Al₂O₃ and Co₃O₄ are used in varying concentrations in a base fluid of EG/water at 0:100%, 10:90% and 20:80%. A schematic of his experimental set-up for investigating nanofluids effectiveness in radiators is illustrated in Fig. 18. The study confirms that the use of Co₃O₄/EG–water results in a more favourable thermal performance than that of Al₂O₃/EG–water. The cobalt oxide also contributed to larger energy savings when compared to alumina. The nanoparticles enhanced the Nusselt number by 31.8%; however, this was at the expense of an increase of 16% in friction factor. This reduction in friction factor resulted in the need for additional pump power for the nanofluids. It is essential to note that pump power was also intensified with the use of EG–water as the base fluid. The performance of a hybrid of Al₂O₃ nanocellulose dispersed in EG/water in a radiator was investigated by Naiman et al. [342], who recorded a maximum thermal conductivity at 0.9 Vol% and concluded that the nanofluids were more efficient than the use of EG–water. Al Rafi et al. [343] studied the heat transfer potential of Al₂O₃/EG–water and CuO/EG–water in a car radiator, revealing that the addition of EG into the water decreased the overall heat conductance by 20–25%. Moreover, experimental results demonstrate that Al₂O₃/EG–water at 0.1 Vol% and CuO/EG–water at 0.2 Vol% improved the heat transfer potential of the radiator by 30–35% and 40–45%, respectively.

Fig. 17

Schematic diagram of a car radiator

Fig. 18

Schematic diagram of the experimentation system used by Elsaid [341]

Kumar and Sahoo [344] analysed the energy and exergy performance of a wavy fin radiator using Al₂O₃–water nanofluid as a coolant. The effect of various nanoparticle shapes (spherical, brick and platelet) on the radiator’s effectiveness, pump power and heat transfer was also investigated; results show that the shape of the nanoparticles affected their performance in the radiator. Furthermore, it was observed that the spherical nanofluids had a 21.98% enhancement in heat transfer when compared to the platelet nanofluid. A 13% enhancement in the exergy efficiency of the spherical nanofluids determined that the use of spherical nanofluids performed better in comparison with nanofluids of other shapes. Contreras et al. [345] experimentally investigated the thermo-hydraulic performance of silver/EG–water and graphene/EG–water for use in a radiator. The study showed that silver/EG–water had an improved heat transfer rate of 4.7% when compared to EG–water, while the heat transfer using graphene nanofluid decreased by 11% and 3% at concentrations for 0.01 Vol% and 0.05 Vol%, respectively, when compared to water. The thermo-hydraulic performance coefficient of all nanofluids showed that nanographene at 0.1 Vol% and silver nanofluids at 0.05 Vol% had values of 1.5% and 2.5%, respectively, while graphene nanofluids at concentrations of 0.01 Vol% and 0.05 Vol% were not suitable for use in the radiator as they performed below EG–water. Other studies on the use of nanofluid in improving the performance of automobile radiators are detailed in Table 10.

Table 10 Studies related to the application of nanofluids in automobile radiators

Nanofluids in thermal storage

Thermal energy storage (TES) is a very important part of the utilisation, conservation and development of new and existing energy sources. The three forms of TES are chemical energy storage, sensible heat storage and latent heat storage. The difference between sensible and latent heat storage types is related to the phase transition of the thermal material used for storage. There is a phase transition before energy is released or stored in the Latent TES, while sensible TES does not require a phase change and operates mainly with the changing temperature of the material. Phase change materials (PCMs) can be used in both cases and is essential to the operation of the latent TES unit. The drawbacks of PCMs are their low thermal properties.

A classification of the various materials used in thermal energy storage is presented in Fig. 19. Highlighting the studies that investigate the effects of nanoparticles on the thermal performance of PCM, Bondareva et al. [357] investigated the heat transfer performance of the nano-enhanced phase change material system under the inclination influence. Studying the performance of paraffin enhanced with Al₂O₃ nanoparticles, they discovered that; for small inclinations of the cavity, when convective heat transfer dominates, an increase in the nanoparticles volume fraction leads to an increase in the melting time. Navarrete et al. [358] proposed the use of molten salt-based nanofluid for both sensible and latent energy storage. The molten salt nitrate would serve as the base fluid for the nano-encapsulated phase change materials (nePCM) consisting of Al-Cu alloy nuclei. Oxidation that occurs as a result of the metals been exposed to air would serve as an encapsulation over the nanoparticles. The study tested the resistance of the oxide shell to temperatures up to 570 °C, demonstrating that although the specific heat and by extension the sensible heat storage decreased with the presence of solid content, the phase change enthalpy and latent storage capacity increased by 17.8% at constant volume bases. Furthermore, the thermal conductivity of the salt nitrates increased with the addition of nanoparticles enhancing the heat transfer performance of the PCM nanofluid. Martin et al. [359] developed a novel nePCM from two fatty acids of capric acid (CA) and capric–myristic (CA-MA) using nSiO₂ for thermal energy management in a building. The addition of the 1.5% nSiO₂ significantly improved both the thermal conductivity and specific heat of nePCM. The thermal stability test after 2000 thermal cycles indicated that the addition of nanoparticles did not affect the thermal stability of CA, but slightly improved that of CA-MA. The sensible heat storage capacity of both fatty acids improved due to a 20% improvement in specific heat capacity at a volume concentration of 1%; however, the latent energy storage capacity of both fatty acids was lowered. The use of the nSiO₂ nanoparticles strengthens on the initial weaknesses of the fatty acids as heat storage fluids as Fig. 20 illustrates.

Fig. 19

Classification of the various thermal energy storage materials (modified from [362])

Fig. 20

Organic PCMs that plot latent heat of fusion vs thermal conductivity [359]

Ding et al. [360] studied the use of two crystal forms of TiO₂ nanoparticles (anatase referred to as A and rutile referred to as R) dispersed in water operating in a microchannel inside a PCM used to enhance the thermal storage in miniatured devices. The two nanofluids R-TiO₂–water and A-TiO₂–water were thermally tested, and both nanofluids confirmed to be stable. R-TiO₂–water was more stable than A-TiO₂, and the thermal conductivity of R-TiO₂ was found to be higher than that of A-TiO₂. The addition of TiO₂–water in the microchannel at a volume concentration of 0.5%, 0.7% and 1.0% decreased the complete melting time of paraffin by 7.78%, 16.51% and 32.90% while increasing the complete solidification time by 7.42%, 15.65% and 22.57% in the solidification process. The use of nanofluids increased the melting and solidification pressure by less than 8% in both cases. Harikrishnan et al. [361] investigated the effect of Ni–ZnO nanocomposite dispersed in oleic acid on the thermal conductivity and phase change properties of the resulting nePCM. The thermal reliability along with the freezing and melting characteristics of the nePCM was studied, and the thermal conductivity of the nanofluids was confirmed to be higher than that of oleic acid. For the mass fraction considered, 0.3, 0.6, 0.9 and 1.2% of Ni–ZnO, the complete melting and solidification processes were enhanced by 7.03%, 14.06% 24.21%, 29.69% and 7.58%, 13%, 19.13%, 28.52%, respectively. The trend confirms that the time required in melting and freezing was lowered with the use of the nano-PCMs. Other studies related to the use of nanoparticles in thermal storage units are detailed in Table 11.

Table 11 Studies related to the application of nanofluids in thermal energy storage

Nanofluids in refrigeration

Nanofluids can also be used in air conditioning and refrigeration systems. The negative environmental effect of using chlorofluorocarbons along with hydrofluorocarbons has propelled research into alternative refrigerants. Traditionally, vapour compression refrigeration systems (VCRSs) are used in the cooling industry; however, the major drawback to this system is the large compressor power requirement. An alternative heat-powered absorption refrigeration system (VARS) has been developed, although the coefficient of performance (COP) of these systems is still below those of the VCRS. Nanoparticles have been used to create new refrigerants known as nanorefrigerants which can improve the COP of both the VARS and VCRS and decrease the compression work of the VCRS.

Rahman et al. [376] analysed the effect of using nanoparticles in a refrigerant. The effect of the nanorefrigerant on the compression work and COP of the air conditioning system is observed. They observed that the addition of 5% SWCNT to R407c refrigerant at temperatures between 283 K and 308 K resulted in a reduction in the energy consumption of the compressor by 4%. Moreover, the nanorefrigerant had improved the thermal conductivity and specific heat values by 17.02% and 10.06%, respectively. The nanorefrigerant also enhanced the COP by 4.59% and reduced the compressor work by 34% when compared to conventional vapour compression refrigeration systems.

Jiang et al. [377] investigated the effect of 0.5% TiO₂ and 0.02% SDBS on the COP of ammonia absorption refrigeration system (AARS). The experimental set-up of the test rig used in their investigations is illustrated in Fig. 21. Outcomes of the experiment were compared to that of 0.1%, 0.3% and 0.5% of TiO₂ dispersed in ammonia water as a refrigerant. The results demonstrate that the addition of TiO₂ to any of the concentrations studied significantly improved the COP of the AARS. It was observed that the further addition of 0.02% of SDBS improved the stability of the mixture and enhanced the COP by 27% as shown in Fig. 22. In conclusion, the improvement in COP of the AARS was strongly dependent not only on nanoparticle concentration but also on the number of nanoparticles stably dispersed in the base fluid. Jeyakumar et al. [378] investigated the use of three nanoparticles CuO, ZnO and Al₂O₃ in the refrigerant of a vapour compression system. The nanoparticles were added to refrigerant R134 at concentrations of 0.06%, 0.08% and 0.1% with 0.1% polyester oil as a lubricant. The results demonstrate an improvement in COP of 12.2% and 3.42% using the nanorefrigerant of CuO and Al₂O₃, respectively. Also, a reduction in the power consumption of 1.39% and 0.6% with CuO and Al₂O₃, respectively, was observed. Other studies related to the use of nanoparticles in compression and absorption refrigeration systems are given in Table 12.

Fig. 21

Test rig for investigating the influence of TiO₂ nanoparticles on AARS [377]

Fig. 22

The COP of AARS with different mass fractions of TiO₂ [377]

Table 12 Studies related to the application of nanofluids refrigeration systems

The use of nanofluids in many other devices has also been studied, and some of these include the application of nanofluids in solar still [389, 390] and also in mineral oil to enhance the insulating properties of high-voltage AC and DC transformers as proposed by Rafiq et al. [391].

Challenges and future prospects

Due to stability concerns with nanofluids, exponential improvements are required for nanofluids to reach their full potential as heat transfer fluids. The problems with stability are more obvious in liquids with low viscosity than liquids with high viscosity. Most of the current methods used to increase fluid stability appear to fall short in certain regards. pH modulation has demonstrated promising signs of improving the stability of nanofluids; however, acidic and basic solutions exponentially increase corrosion in metals and would thus render heat transfer system untenable. The addition of surfactants has the potential to improve nanofluids stability, however, at high-temperature surfactants tend to foam and decrease the overall efficiency of the system. The most promising technique for increasing fluid stability is by improving the synthesis techniques used. Incidentally, the most common method for synthesising nanofluid is the worst performing method for ensuring fluid stability. Green synthesis techniques demonstrate sufficient promise in improving stability; however, the thermal performance of the green-synthesised nanofluids is not normally as high as nanofluids synthesised by the two-step technique. Furthermore, there appears no standard for reporting the stability of nanofluids. Therefore, a generic standard for measuring nanofluid stability must be developed so that easy comparisons can be made across nanofluid types.

Another significant challenge is the theoretical unpredictability of the thermophysical behaviour of nanofluids. While many studies settle for regression-based correlation models to predict thermophysical properties, intelligent computing has also been widely used in the predictions. It is the opinion of the authors that because of the almost infinite variables that affect the thermophysical behaviour of nanofluids, intelligent computing would be the most accurate predicting the thermophysical behaviour of nanofluids. Therefore, a generic standard must be developed for labelling data obtained from the experiments measuring thermophysical properties of nanofluids. Developing a global data bank will drastically improve the prediction accuracy of artificial neural network and machine learning models, saving unlimited research costs in conducting thermophysical behaviour measurements.

To improve numerical analysis models, further nanofluid heat transfer correlation studies are required for determining the Nusselt number correction equation. Many studies adopt the Nusselt number correlation equation proposed by Pak and Cho [392]; however, this model was developed for water, Al₂O₃–water and TiO₂–water nanofluids and may not be particularly accurate for other nanofluids. More experiments using other nanofluids, especially for hybrid nanofluids, will further enlighten the field and improve the accuracy of numerical studies.

Finally, the classification of nanofluids must be improved. As nanofluid research increases, several unique types of fluids are synthesised. Previously, conventional fluid often implies fluids with a single-particle material, while hybrid nanofluid refers to a fluid with more than one nanoparticle material. However, it appears that further classifications are required as nanofluid have the potential to have an nth number of significant nanomaterials types present in the fluid. Some authors have sought to classify nanofluids with two significant nanomaterials type as “binary hybrid nanofluid” and nanofluids with three significant nanomaterials type as “ternary hybrid nanofluid”. It may be beneficial if classifications are conducted along these lines.

Conclusions and recommendations

The use of nanofluids as coolants in heat transfer devices has gained attention over the years. This study presents a detailed review of studies relating to the preparation, thermophysical property measurements and application of nanofluids in a range of thermal devices requiring efficient heat transfer published in 2019. Some of the areas reviewed include thermophysical models used in determining the properties of the nanofluids, mechanisms that support the enhanced thermal behaviours of nanofluids, and the application of different nanofluids in devices such as solar collectors, heat exchangers, electronics cooling and thermal storage. Based on the articles reviewed in this study, the following recommendations are made:

On the preparation of nanofluids;

• Few studies on the preparation of nanofluids based on the one-step method are available, and this method has been proven to have better stability than the two-step method. More studies on the production of nanofluids using the one-step method are needed, as this could help in the development of more cost-effective means for the large-scale production of nanofluids.

Regarding the thermophysical properties of nanofluids:

• An increase in the nanoparticle volume concentration leads to a decrease in the specific heat capacity of nanofluids in cases where the heat capacity of base fluids is higher than those of nanoparticles. Since a higher heat capacity is needed in coolants, further studies are required to assess how this phenomenon can be improved.

• Many studies on the thermal behaviour of nanofluids were conducted for temperatures between 10 and 100 °C. The interaction mechanism of nanoparticles in base fluids for heat transfer at higher temperatures (greater than 100 °C) and cryogenic conditions requires further investigation.

• There exist huge differences between the heat transfer predicted by the single-phase homogenous model and those obtained from experiments. More studies related to the development of other models (two-phase models) are required which allude to defining the conditions where the single-phase models can be applied to provide more accurate results.

• There has been an increase in both the number and methods for developing correlation models that predict the thermophysical properties of nanofluids. However, more correlation equations that predict the heat transfer (Nusselt number) and friction factor behaviours of many nanofluids are needed.

On studying the mechanisms that influenced the properties of nanofluids:

• Knowledge of the dominant forces responsible for the behaviour of nanorefrigerants in various flow configurations requires further development.

• An understanding of the impact of nanoparticle morphology (size and shape), nanoparticle mixture ratio (for hybrid nanofluids) on heat transfer augmentation is limited. More studies are needed to understand the impact of these on the performance of nanofluids in heat transfer devices.

Investigation on the various heat transfer devices:

• Further studies are required, as there are contrasting reports on the effect of nanoparticle loading on the pressure drop and additional pump power requirement. While some studies claim that the effect of particle loading increases the pressure drop and consequently the pump power requirement of the system, others argue that when the heat transfer rate obtained using nanofluids is compared with that of conventional fluids, the nanofluids lowers the pump power requirements.

• In heat exchangers and car radiators, the constant rate of heat transfer from the use of nanofluids leads to a reduction in the heat transfer surface. This can result in an improvement in the size and volume of these devices. Such improvements would lead to a reduction in the drag forces witnessed in vehicles and increase the performance of the engine.

• The most common model used in the literature for the simulation of nanofluids remains the finite volume method. Further studies using other methods are needed for the comparison of the different numerical approaches.

• Further studies on the effects of erosion of heat transfer and corrosion of flow channels resulting from the use of nanofluids, especially in high temperatures, are required. Both the short- and long-term impacts of sedimentation and nanoparticle deposition on the efficiency of heat transfer devices require investigation.

• Few studies are available on the production cost and environmental impact of nanofluids. Such factors present huge hurdles to the commercialisation of nanofluids.

• Further information on the effect of oxidisation of metallic nanoparticles used with phase change materials on the thermal performance of the thermal storage unit is required, especially during the melting phase.