Available online at www.sciencedirect.

com

Particuology 7 (2009) 141–150

Review of nanofluids for heat transfer applications

Dongsheng Wen a,∗ , Guiping Lin b , Saeid Vafaei a , Kai Zhang c
a School of Engineering and Materials Science, Queen Mary University of London, E1 4NS, United Kingdom
b School of Aeronautics Engineering, Beijing University of Aeronautic and Astronautics, Beijing, China
c State Key Lab of Heavy Oil Processing, China University of Petroleum, Beijing 102249, China

Abstract
Research on nanofluids has progressed rapidly since its enhanced thermal conductivity was first reported about a decade ago, though much
controversy and inconsistency have been reported, and insufficient understanding of the formulation and mechanism of nanofluids further limits
their applications. This work presents a critical review of research on heat transfer applications of nanofluids with the aim of identifying the limiting
factors so as to push forward their further development.
© 2009 Chinese Society of Particuology and Institute of Process Engineering, Chinese Academy of Sciences. Published by Elsevier B.V. All
rights reserved.

Keywords: Nanofluids; Nanoparticles; Process intensification; Thermal conductivities; Boiling; Convective heat transfer

1. Introduction passive techniques and active techniques. External energy input

is needed for active techniques, viz. mechanical mixing, rota-
Energy has been rated as the single most important issue tion, vibration, and addition of external electrostatic or magnetic
facing humanity in the next 50 years, in particular, the serious field, which have been successfully used to intensify heat and
terawatt challenge while mitigating carbon emission and global mass transfer, leading to efficient heat pumps, separators and
warming (Smalley, 2005). Securing clean energy has become reactors and faster processing. However external energy input
the top priority of most developed countries. Fundamentally, is costly and inconvenient under compact situations. For pas-
energy conversion and transportation occur at atomic or molecu- sive techniques, heat transfer intensification can be achieved by
lar levels, nanoscience and nanotechnology are expected to play modifying fluids property, surface shape, roughness or attaching
a significant role in revitalizing the traditional energy indus- objects to increase surface area and turbulence. However their
tries and stimulating the emerging renewable energy industries. inherent limitation is the relatively low thermal conductivity of
Among all the forms of energy we are using today, over 70% heat transfer fluids. Water possesses the highest thermal conduc-
are produced in or through the form of heat. In many industrial tivity among all the fluids we use today, though it amounts to
systems, heat must be transferred either to input energy into a merely ∼0.6 W/m K under room temperatures, which is several
system or to remove the energy produced in a system. Consider- orders lower than most metals or metal oxides, as shown in Fig. 1.
ing the rapid increase in energy demand worldwide, intensifying It would perhaps be logical to add certain solid particles into a
heat transfer process and reducing energy loss due to ineffective base liquid to increase its thermal conductivity k, a concept that
use have become an increasingly important task. Heat removal has been practiced for a long time. Most of these early studies,
and control are the demanding challenges in some high heat flux however, used suspensions of millimeter- or micrometer-sized
systems such as nuclear fission, fusion, micro/nanoelectronics particles, which led to problems such as poor suspension sta-
mechanical systems (MEMS/NEMS), process intensification bility and hence channel clogging, often serious for systems
and micro chemical reactions. consisting of small channels.
Methods that have been proposed to intensify heat transfer Nanofluids are dilute suspensions of functionalized nanopar-
in different situations can be classified into two main groups: ticles smaller than 100 nm, which belong to a new type of
functional composite materials developed about a decade ago
with the specific aim of increasing the thermal conductivity of
∗ Corresponding author. Tel.: +44 20 78823232; fax: +44 20 89831007. heat transfer fluids (Choi, 1995), which have now evolved into a

1674-2001/$ – see inside back cover © 2009 Chinese Society of Particuology and Institute of Process Engineering, Chinese Academy of Sciences. Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.partic.2009.01.007
142 D. Wen et al. / Particuology 7 (2009) 141–150

nanofluids formulation, nanofluids applications and mechanistic

understanding, as will be detailed below.

2. Nanofluids formulation

Though various aspects of nanoparticles have been exten-

sively investigated over the past few decades; little attention
has been paid to the formulation of nanofluids, i.e. the mak-
ing of stable nanoparticle suspensions suitable for heat transfer
applications. The properties and behaviour of nanofluids much
depend on the properties of the base liquid and the dispersed
phases, particle concentration, size and morphology, as well as
the presence of dispersants or surfactants (called stabilizer in
Fig. 1. Comparison of the thermal conductivity of common liquids, polymers general in following sections). Formulation of stable nanoparti-
and solids. cle suspensions has been widely practised in the colloid industry,
and even in ancient times, a number of ultrafine particles such
promising nanotechnological area. Such thermal nanofluids for as carbon and gold were dispersed into liquid media for painting
heat transfer applications represent a class of its own different and glass coating. However, the requirement of nanofluids for-
from conventional colloids for other applications. Compared to mulation is more demanding than conventional colloids due to
conventional solid-liquid suspensions for heat transfer intensi- the harsh application environment, i.e. typically under high shear
fications, properly engineered thermal nanofluids possess the and high temperature conditions. Formulating stable nanofluids,
following advantages: with controlled properties such as thermal conductivity, viscos-
ity and wettability for heat transfer applications, still presents
• High specific surface area and therefore more heat transfer a challenge for the nanofluid community. Two methods can
surface between particles and fluids. be generally used for such formulation, namely, the top-down
• High dispersion stability with predominant Brownian motion method through size reduction (the two-step method), and the
of particles. bottom-up approach through simultaneous production and dis-
• Reduced pumping power as compared to pure liquid to persion of nanoparticles (the one-step method).
achieve equivalent heat transfer intensification.
• Reduced particle clogging as compared to convention slurries, 2.1. The two-step method
thus promoting system miniaturization.
• Adjustable properties, including thermal conductivity and For the two-step method, dried nanoparticles have to be
surface wettability, by varying particle concentrations to suit either synthesized or purchased in the form of dry powders.
different applications. Nanoparticles synthesized through different routes have dif-
ferent morphologies, structures and hence different physical
Fig. 2 shows the rapid growth of nanofluids research in recent and chemical properties, which will affect the quality of the
years, including US, EU, and China, and more recently, India final nanofluids. Currently a wide range of vapor, liquid and
and South Korea. The potential market for nanofluids for heat mechanical routes are available for synthesizing nanoparticles.
transfer applications is estimated to be over 2 billion dollars per The mechanical route, such as milling and grinding, is straight-
year worldwide (CEA, 2007), with prospect of further growth in forward but energy intensive. The liquid route such as the
the next 5–10 years. A critical review of state-of-the-art nanoflu- sol–gel and wet chemical methods typically involves chemi-
ids work for heat transfer intensification is both necessary and cal reactions of a few reagents with one desired product while
timely, for which this study will focus on three main aspects: normally producing other impurities that are difficult to control.
The vapor-phase route involves producing a reactive vapor in
which nucleation and growth occur through combustion, pyrol-
ysis, plasmas, laser ablation, laser pyrolysis, or chemical vapor
deposition. From both economic and quality considerations, the
gas phase route has become the most favorable method for large-
scale production of nanoparticles. Due to strong Van der Waals
attractive force, however, nearly all nanoparticles are in the form
of dried agglomerates with much larger dimensions than the
primary particles, as shown in Fig. 3 using alumina and nickel
particles as examples. The degree of agglomeration changes with
the nature of the manufacturing, handling and storage process.
As a consequence, careful selection of synthesis methods and
proper characterisation of as-received dried nanoparticles are
Fig. 2. Growth of publications by the nanofluids community. highly necessary for controlled formulation of stable nanofluids.
 D. Wen et al. / Particuology 7 (2009) 141–150 143

Fig. 5. Samples of ␣-Al2 O3 particle size profile in situ.

nizers and bead mills. The processing time and intensity could
significantly influence the dispersion effect. Weakly bonded
agglomerates could generally be broken to the primary sizes
by high shearing; though nanoparticles tend to re-agglomerate
due to the attractive London van der Waals force. So electro-
static repulsion or steric hindrance is required to overcome such
attractive force in order to form stable dispersions. Certain sta-
bilizers are generally used during the formulation of nanofluids
to establish steric barriers among nanoparticles, in order to sta-
bilize the nanofluids. Nanofluids formulated without stabilizers
could be used instantly but stability would change rapidly with
time, as shown in Fig. 4, where alumina nanoparticles become
completely separated after 5 h. In this example, the aqueous
nanofluids were made from commercial dried alumina powders
with primary particle size of 20-nm as shown in Fig. 3a, and
formulated by the two-step method at a concentration of 2.5%
by weight through a homogenizer without any stabilizers. The
duration of stability of these nanofluids depends on the parti-
Fig. 3. SEM Samples of dried nanomaterials ␣-Al2 O3 (a) and nickel (b). cle surface properties, size and morphology. It has to be realized
that in the dispersed state, the actual particle size and its distribu-
tion are generally quite different from that of the dried particles.
Compared to the bottom-up approach, dispersing dry parti- For example, Fig. 5 illustrates the particle size distribution of
cles into liquids has been more frequently used by researchers freshly made alumina nanofluids shown in Fig. 4, measured by
in the nanofluid field. It is essential to disperse large agglomer- the dynamic light scattering (DLS) method through the Malvern
ates of nanoparticles to their primary particle size to formulate a Nanosizer, to be approximately 130 nm, which is much larger
stable dispersion. The process, by which particles are dispersed than the primary particle size ∼20 nm shown in Fig. 3a. This
in the liquid, plays a critical role for adequate dispersion under might be caused by one of two reasons. One is the existence of
the thermal and hydraulic conditions imposed on it. The gen- small, difficult-to-break agglomerates which were closely coa-
eral equipment used in dispersing dried nanoparticles includes lesced or sintered together during manufacture. The other is that
ultrasonic baths, magnetic stirrers, high-shear mixers, homoge- the Malvern Nanosizer measures the hydrodynamic size accord-

Fig. 4. Samples of ␣-Al2 O3 nanofluids (without any stabilizer) stability change with time.
144 D. Wen et al. / Particuology 7 (2009) 141–150

Fig. 6. Two-step method for nanofluids formulation.

ing to the Stokes–Einstein equation, which is expected to be 2.2. The one-step method
larger than the actual size. Such a comparison calls for caution
when interpreting the particle size effect based only on the nom- The formulation of nanofluids by the bottom-up approach
inal size distribution provided by the manufacturer or observed through physical or chemical reactions has been gaining increas-
by electron microscopy. To elucidate the effect of particle size ing interests. Such a method has been practised for a long time
and agglomeration structure, it is essential to distinguish the in the colloid industry, viz. colloidal gold, which has attracted
particle size difference in dried form and in dispersed state, for considerable attention from a fundamental and practical point of
which, as mentioned above, little is provided in most published view due to its attractive electronic, optical, thermal and catalytic
studies on nanofluids. properties. The study of colloidal gold started since Michael
To overcome the stability problem, stabilizers are often Faraday’s work in the mid nineteen century. Colloidal gold is
introduced during the formulation process, as illustrated generally synthesized by the chemical reduction of gold salts
schematically in Fig. 6, that involves three steps: (a) preparing such as chloroauric acid (HAuCl4 ) in aqueous, organic phase or
pure liquid, (b) mixing suitable stabilizers to the pure liquid (base two phases, and surface functionalized with various materials.
liquid) and (c) dispersing nanoparticles into the base liquid. In The established synthesis methods include citrate reduction, the
the nanofluid community, the heat transfer capability of nanoflu- Brust-Schiffrin method (Brust, Walker, Bethell, Schriffrin, &
ids is generally compared with that of pure liquid, which is quite Whyman, 1994), and the modified Brust-Schiffrin method that
misleading. The stabilizers, especially those loose ones as shown contains different sulfur-containing ligands. The size of gold
in Fig. 6c could alone alter the values of surface tension, wet- nanoparticles can be tuned by controlling the ratio of thiol or
tability and viscosity. These factors, as will be discussed below, other ligands to Au ions used in the synthesis. Gold nanofluids
will change significantly the flow and heat transfer behaviour have only been investigated by a limited number of studies due
of nanofluids. As a consequence, the effects of nanoparticles to the high cost (Tsai et al., 2004). A number of other nanoflu-
and of stabilizers are mixed together. The common approach in ids were formulated such as copper and iron nanofluids through
nanofluid community of comparing this mixed effect with pure a modified physical vapor deposition method and hydrother-
liquid only could be one of the major reasons responsible for the mal chemical reduction of salts (Eastman, Choi, Li, Yu, &
observed large data scattering. A proper comparison base has to Thompson, 2001; Zhu, Lin, & Yin, 2004). Further formulation
be built to differentiate the effect of stabilizer and nanoparticles. of nanofluids should be referred to the experience developed in
There are some commercial nanoparticles with modified the colloid community.
surfaces, either as functional layers or stabilizers introduced The stability of nanofluids can be improved through proper
onto the particle surface during synthesis. If the bonding of sta- surface functionalization without involving mechanical facili-
bilizer and nanoparticles is strong enough, a proper dispersion ties. The bottom-up approach however suffers the disadvantage
process may eliminate the loose stabilizers in the liquid phase that residual reactants are left in the nanofluids due to incom-
of Fig. 6c, so as to establish a proper comparison base between plete reaction or stabilization. It is difficult to elucidate the
functional particles and pure liquids. However this needs to nanoparticle effect without eliminating this impurity effect. The
be carefully assessed to prove that there is no breakage of the functionality of stabilizers under high temperature is also a big
bonding. Another significant issue for these functionalized concern, especially for high temperature applications.
particles is the temperature effect. For biological or colloidal It appears that the controlled formulation of stable nanofluids
application, these nanoparticle dispersions are normally stable still presents a challenge for heat transfer intensification appli-
at room temperature, though at elevated temperatures the cations. The lack of proper control of both solid phase and liquid
stabilizers may fail, thus affecting the stability, such as occurs phase could be one of the most important reasons for the large
for some nano-dispersions from commercial suppliers, as was data scattering reported in the literature, and also presents as a
verified by Wen and Ding (2004a) for carbon nanotube nanoflu- big barrier for industrial adoption of nanofluids. Close collabo-
ids using sodium dodecylbenzene sulfonate (SDBS) as the ration between colloid and material scientists and heat transfer
stabilizer. engineers should be strengthened in the future.
 D. Wen et al. / Particuology 7 (2009) 141–150 145

3. Nanofluids for heat transfer application after a certain concentration threshold (Tseng & Lin, 2003). In
the meantime, stabilizers, especially some surfactants strongly
While the original idea of nanofluids was to enhance the ther- tend to reduce the viscosity of the base liquid, the effect of which
mal conductivities of some typical heat transfer fluids including is compounded with that of solid particles to further complicate
water, mineral oil and ethylene glycol, the influence of nanopar- the situation. The decrease of effective specific heat of nanofluids
ticles has been found even more profound than mere thermal is understandable, and was indeed observed in recent experi-
conductivity effect. While most current research is still focused ments for a sample Al2 O3 /H2 O (Zhou & Ni, 2008) nanofluid.
on thermal conductivity, neglecting the alteration of other prop- It clearly showed that the effective specific heat decreased with
erties by nanofluids especially viscosity and surface properties. increasing nanoparticle concentrations. Research into the influ-
A large number of inconsistent experiments reported for nanoflu- ence of nanoparticles on surface tension, or more generally, the
ids applications under flow conditions either with or without wettability is still at an early stage. It is generally believed that
phase change. This section will discuss the influence of nanopar- nanoparticles could increase the wettability of a base liquid and
ticles on effective properties, while using boiling heat transfer as might be responsible for the large increase in the critical heat
an example to reveal the contradictions among different research flux (CHF) as observed in experiments (Kim, Bang, Buongiorno,
groups and commenting on the effect of nanofluids on flow and & Hu, 2006, 2007a, 2007b; Vassallo, Kumar, & Damico, 2004;
heat transfer applications. You, Kim, & Kim, 2003). However further research is clearly
needed.
3.1. Influence of nanofluids on effective properties It is clear that the increase of thermal conductivity might be
h=h(cp,, , )
offset by the increase of viscosity, the decrease of effective spe-
The effectiveness of heat transfer is described by the heat cific heat or the variation of wettability. An impartial assessment
transfer coefficient (HTC), h, which is a function of a num- of the application of nanofluids should be conducted without
ber of thermo-physical properties of the heat transfer fluids, the focusing on the effective thermal conductivity only.
most significant ones being thermal conductivity, k, heat capac-
ity, cp , viscosity, μ, density, ρ, and surface tension, σ, if phase
change is involved. The general trend is that the HTC increases 3.2. Influence of nanofluids on boiling heat transfer
with effective thermal conductivity, specific heat and density, but
decreases with viscosity. For instance,
 for the
√ classical transient On the application of nanofluids, convective heat transfer
conduction heat transfer, q ∼ kρcp T/ t, the HTC can be of nanofluids with and without phase change has been exten-
expressed as h ∼ k0.5 ρ0.5 cp0.5 /t 0.5 ; and for convective heat trans- sively investigated. However similar to the investigation on the
fer of steady state homogenous flow, the HTC can be generally effective properties, these results are far from being conclusive.
expressed as h ∼ ka ρb cpc /μd σ e , where a, b, c, d, e are empirical While qualitative agreement has been reached on convective heat
or theoretical constants that depend on different boundary and transfer with nanofluids, many conflicting experimental results
geometrical conditions, and the constant e is normally zero for have been reported for pool boiling heat transfer. For instance,
convections without phase change. The addition of nanoparticles when compared with boiling of pure fluids, significant boiling
into a base liquid could affect all these properties thus affecting heat transfer enhancement has been reported for nanofluids (Liu,
the effectiveness of heat transfer. Xiong, & Bao, 2007; Park & Jung, 2007a, 2007b; Wen & Ding,
Besides the effective density that is easily measurable, the 2005; Wen, Ding, & Williams, 2006; Witharana, 2003), which
quantitative influence of nanoparticles on all other four effective however contrasts significantly with others where heat transfer
properties has not yet been established. A large number of stud- deterioration was reported (Bang & Chang, 2005; Das, Putra,
ies reported that nanofluids could possess much higher effective & Roetzel, 2003a, 2003b; Kim et al., 2007a, 2007b; Chopkar,
thermal conductivities than the base fluids and the predictions Das, Manna, and Das, 2008). Fig. 7 summarizes the trends of
from the effective media theory. Typically a ∼10% increase in these observations in terms of the boiling heat transfer coeffi-
the thermal conductivity for ∼1% particle load by volume can cient ratio, defined as the HTC of nanofluids over that of the
be achieved (Yu, France, Routbort, & Choi, 2008), although base fluids. The results are very inconsistent even for the same
quantitative agreement is still lacking and the reason for such a nanoparticles under similar experimental conditions. Most of
large data scattering as well as the underlying mechanism are the explanations on the observed experimental results have been
still unclear. Nanofluids may suffer two other inherent limita- speculative, especially on the interaction of nanoparticles, bub-
tions, the increased effective viscosity due to the addition of ble nucleate cavities and heating surfaces.
nanoparticles and the decreased effective specific heat as most Compared to the contradictory nucleate boiling results with
of the solid materials added, such as metal and metal oxides, have nanofluids, more qualitatively consistent observations have been
much lower specific heat than conventional heat transfer fluids reported on the enhancement of the Critical Heat Flux (CHF)
such as water. An increasing number of studies has shown that by nanofluids (Bang & Chang, 2005; Kim et al., 2007a, 2007b;
the increase of effective viscosity is also beyond prediction from Milanova & Kumar, 2005; Vassallo et al., 2004; You et al., 2003;
conventional theories, especially for high particle concentrations Kim, Kim, and Kim, 2006). However the quantitative results
(Chen et al., 2008). Similar to the mechanisms related to effec- still differ: the CHF enhancement has been reported to vary
tive thermal conductivity, the morphology and particle structure from 10–40% (Bang & Chang, 2005) to 200–400% (Milanova
could also affect viscosity, which could increase exponentially & Kumar, 2005; You et al., 2003) as summarized in Fig. 8.
146 D. Wen et al. / Particuology 7 (2009) 141–150

observed by Das et al. (2003a). On the content of nanofluids, the

introduction of stabilizers, especially loose stabilizers in the liq-
uid phase, significantly influence both the rheological behaviour
of the fluids and the boiling heat transfer (Cheng, Mewes, &
Luke, 2007; Hetsroni, Gurevich, Mosyak, Rozenblit, & Segal,
2004; Wasekar & Manglik, 2000; Wen & Wang, 2002). How-
ever, most published studies on pool boiling of nanofluids did
not clearly specify if surfactants or dispersants were used. For
example, the nanofluids used by Vassallo et al. (2004) and Kim et
al. (2007a) were prepared by diluting a commercial concentrated
dispersion, which very likely contained surfactants/dispersants.
The variability of the solid phase, especially the particle surface
properties, either hydrophobic or hydrophilic, and their interac-
tion on the heating surface, will significantly change the boiling
behavior.
This example of boiling heat transfer with nanofluids illus-
trates the current weak understanding of nanofluids, and calls
for more systematic and strategic work.

4. Mechanisms
Fig. 7. Summary of literature data on boiling heat transfer with nanofluids.
4.1. Conduction
It is difficult to explain such wide differences. As reviewed
in the nanofluid formulation section, the most probable rea- The nanofluid community has intensively investigated the
sons reside in the stability of nanofluids and the variability of thermal conductivity of nanofluids, proposing a number of
both the solid and the liquid phases. Most nanofluids used in mechanisms for its enhancement including the interfacial layer-
boiling experiments tend to be unstable at high temperatures ing, Brownian motion, ballistic transport of energy carriers, and
due to the failure of stabilizers. The stabilizers and particles the structure effect (Eapen, Li, & Yip, 2007; Keblinski, Phillpot,
could agglomerate and deposit on the heating surface, which cer- Choi, & Eastman, 2002; Nie, Marlow, & Hassan, 2008), though
tainly modifies the heating surface and brings about a number of there are yet debates on the effect of particle Brownian motion
unforeseen effects, including the change of active nucleate cav- (Keblinski et al., 2002; Li & Peterson, 2007), the interfacial
ities, the modification of surface wettability and the formation resistance (Tillman & Mill, 2007; Xie, Fujii, & Zhang, 2005;
of extra thermal resistance on the heating surface that prevents Xu et al., 2006), and more recently the particle aggregation and
direct contact of liquid with the boiling surface. Some of these percolation effect (Evans et al., 2008; Lee, Kim, & Kim, 2006;
factors could explain the deterioration of boiling heat transfer Prasher, Phelan, & Bhattacharya, 2006).
behavior observed by Das et al. (2003a, 2003b) and Bang and A comparison of the time constant between the particle Brow-
Chang (2005), as is indeed reflected by the images of the boiling nian motion and thermal diffusion within the base liquid shows
surface after experiments (Bang & Chang, 2005), and directly that the direct movement of nanoparticles contributes negligibly

Fig. 8. Summary of literature data on CHF enhancement ratio.

 D. Wen et al. / Particuology 7 (2009) 141–150 147

Comeau, 2006). As discussed earlier, considering the effective

thermal conductivity effect alone is not sufficient; and convective
heat transfer could only be modelled reasonably well if proper
models involving thermal conductivity, viscosity and specific
heat are selected. However no convincing models employing
these effective properties have yet been established, especially
involving thermal conductivity and viscosity as reviewed above.
Some experiments have actually suggested that the viscous
effect is not negligible, and the gain from thermal conduc-
tivity increase could even be offset by increased viscosity
(Buongiorno, 2006; Pak & Cho, 1999). There is a strong impli-
cation that particle size and shape could play an important role
in the level of heat transfer enhancement. Under similar con-
Fig. 9. Aggregation effect on the effective thermal conductivity.
ditions, the enhancement of heat transfer coefficient is much
higher for carbon nanotubes (Ding, Alias, Wen, & Williams,
to thermal conduction. The effect of micro-convection associ- 2006) than that for disc-shaped TiO2 nanofluids (Yang, Zhang,
ated with Brownian motion has recently been emphasized. These Grulke, Anderson, & Wu, 2005). Besides the effective property
models, however, generally employ the kinetic theory under a influence, a number of other mechanisms have been proposed by
number of idealized assumptions; and the prediction contradicts considering the two-phase nature of nanofluids, which includes
the direct molecular dynamic simulation (Keblinski et al., 2002). the thermal dispersion effect (Xuan & Roetzel, 2000) and par-
The enhanced mass transport visualized through the observation ticle migration effect (Wen & Ding, 2005), but none of them is
of dye diffusion in nanofluids (Krishnamurthy, Bhattacharya, conclusive (Wen, 2008a). Further study is urgently required.
Phelan, & Prasher, 2006), which is analogous to heat transfer in
nanofluids and has been taken as a strong evidence for the role of 4.3. Boiling heat transfer
micro convection by Brownian motion of particles, could be due
to the stabilizer effect as the surfactant introduced could signif- The mechanisms of boiling heat transfer of pure fluids have
icantly reduce viscosity. It is however yet unclear if the thermal been developed during the past few decades, though little mech-
conductivity enhancement is caused by Brownian motion anistic analysis of boiling heat transfer with thermal nanofluids
particles or the reduced viscosity due to the surfactant. Both the has been conducted. Nanofluids may behave more as pure flu-
Brownian motion and micro-convection effect were criticized ids than conventional suspensions of micrometer or millimeter
recently from a fundamental heat transfer point of view (Nie particles, but they are two-phase in nature and should possess
et al., 2008). Even more recently, the effect of aggregation has some features of conventional suspensions. The interaction of
been proposed by a number of researchers (Evans et al., 2008; nanoparticles with heating surface and subsequent modification
Prasher et al., 2006), who emphasized that the enhancement of of surface properties can change the boiling behavior; however
thermal conductivity is a function of nanoparticle aggregation, such an influence is difficult to predict as it involves the stabilizer
as summarized schematically in Fig. 9. According to this effect, the stability effect and the alteration of the boiling surface
mechanism, there exists an optimized aggregation structure to before and after experiments. Most of the explanation on experi-
achieve maximum thermal conductivity, which is far beyond the mental result summarized in Fig. 7 is based on speculation only.
prediction from homogeneous dispersions. Such an argument Besides these conventional factors, latest development in col-
excludes thermal conductivity as an inherent physical property. loids science suggested the importance of a structure disjoining
The loose aggregation structure may be destroyed under flow pressure.
and heating conditions. Its validation requires further careful As the boiling heat transfer is largely affected by flow and
examination. The potential influence of particle aggregation heat transfer within the microlayers of bubbles at microm-
on conduction emphasizes that colloid chemistry will play eter/nanometer thickness, the capillary force, intermolecular
a significant role in optimizing the thermal conductivity of interactions (disjoining pressure), and fluid and surface inter-
nanofluids. actions (wetting phenomena) become important. Consequently,
the understanding of thermal nanofluids on nucleate boil-
4.2. Convection without phase change ing needs to consider the influence of nanoparticles on the
flow and heat transfer within the microlayer. Besides the
Both homogeneous and heterogeneous, or combined, effects of evaporation, vapor recoil, capillary force and the
approaches have been employed in modeling convective heat conventional disjoining pressure derived from the classical
transfer using nanofluids. For most of the convective model- Derjaguin–Landau–Verwey and Overbeek (DLVO) theory, a
ing work based on the homogeneous approach, an effective long range structure disjoining pressure normal to the wall could
thermal conductivity model taking into account the effect of arise in the microlayer due to the presence of nanoparticles
nanoparticles has always been assumed, as the main mecha- (Trokhymchuk, Henderson, Nikolov & Wasan, 2001; Wasan
nism for convection (Wen & Ding, 2004b; Koo & Kleinstreuer, & Nikolov, 2003). Different from the conventional disjoining
2005; Maïga, Nguyen, Galanis, & Roy, 2004; Roy, Nguyen, & pressure, which results from London-Van der Waals force of a
148 D. Wen et al. / Particuology 7 (2009) 141–150

short-range nature, it has been demonstrated that a structural to mechanism understanding. Engineering suitable nanofluids
disjoining force is generated from the ordering of nanoparti- with controlled particle size and morphology for heat transfer
cles in a confined wedge (structuring) and the influence can applications is still a big challenge. The current uncertainties on
extend to a film depth of a few nanoparticle diameters (long the content of nanofluids, including both solid phase and liq-
range) (Wasan & Nikolov, 2003; Sefiane, 2006. The origin of uid phase, could be one of the main reasons responsible for the
the structural disjoining pressure is believed to be caused by controversy and inconsistency reported. Besides thermal con-
the confinement of the particles in the film region as opposed to ductivity effect, future research should consider other properties,
their greater freedom of location in the bulk liquid. When the film especially viscosity and wettability, and examine systematically
thickness is large enough to accommodate more than one layer their influence on flow and heat transfer. An in-depth under-
of particles, the particles tend to arrange themselves in regular standing of the interactions between particles, stabilizers, the
layers. Such layering arrangement of particles gives rise to an suspending liquid and the heating surface will be important for
excess pressure in the film, the structural disjoining pressure. A applications. Inter-disciplinary collaboration among nanomate-
result of such a structure force is that nanofluids could exhibit a rials, colloid science, physics and engineering researchers will
super spreading capability in a confined space. Such forces have be necessary to engineer suitable nanofluids and accelerate their
been observed to be capable of changing the macroscopic con- applications.
tact angle of a liquid droplet (Kim et al., 2007a, 2007b; Vafaei
et al., 2006), stabilizing liquid films (Sethumadhavan, Nikolov,
& Wasan, 2001) and lifting an oil droplet from a wall in an Acknowledgement
aqueous solution (Chengara, Nikolov, & Wasan, 2001; Wasan
and Nikolov, 2003). Further inter-disciplinary studies covering The authors would like to extend their thanks to EPSRC for
nanomaterials, colloidal science and engineering applications financial support (Grant No.: EP/E065449/1).
are required to elucidate such an influence, especially regarding
concentration effect and transient effect (Wen, 2008b).
References
5. Outlook Bang, I. C., & Chang, S. H. (2005). Boiling heat transfer performance and
phenomena of Al2 O3 –water nano-fluids from a plain surface in a pool.
A recent work reported a switchable thermal conductivity of International Journal of Heat and Mass Transfer, 48, 2407–2419.
ferrofluids through an external magnetic field (Philip, Shima, Brust, M., Walker, M., Bethell, D., Schriffrin, D. J., & Whyman, R. J. (1994).
& Raj, 2008). An extraordinary enhancement of thermal con- Synthesis of thiol-derivated gold nanoparticles in a 2-phase liquid–liquid
system. Journal of Chemical Society: Chemistry Communications,
ductivity, ∼300%, is observed when linear chain-like structures 801–802.
are generated and uniformly dispersed in the base fluid, while Buongiorno, J. (2006). Convective transport in nanofluids. Journal of Heat
negligible enhancement is obtained for well-dispersed particles. Transfer, 128, 240–250.
The effective thermal conductivity of nanofluids can be adjusted CEA. (2007). Nanofluids for heat transfer applications. France: Marketing Study
by proper control of the external magnetic field to generate dif- Unit, CEA.
Chopkar, M., Das, A. K., Manna, I., & Das, P. K. (2008). Pool boiling heat
ferent nanoparticle structures. Such a novel finding supports the transfer characteristics of ZrO2 –water nanofluids from a flat surface in a
particle aggregation and structure theory for effective thermal pool. Heat and Mass Transfer, 44, 999–1004.
conductivity and opens a new window for engineering unique Cheng, L., Mewes, D., & Luke, A. (2007). Boiling phenomena with surfactants
nanofluid properties for different applications. However caution and polymeric additives: A state-of-the-art review. International Journal of
is needed for practical heat transfer applications as the loose Heat and Mass Transfer, 50, 2744–2771.
Chen, H. S., Yang, W., He, Y. R., Ding, Y. L., Zhang, L. L., Tan, C. Q., et al.
structure could be easily broken under flow and heating con- (2008). Heat transfer and flow behaviour of aqueous suspensions of titanate
ditions. The influence of particle structure on other effective nanotubes (nanofluids). Powder Technology, 183, 63–72.
properties, especially surface tension (wettability), viscosity and Chengara, A., Nikolov, A. D., & Wasan, D. T. (2004). Spreading of nanofluids
specific heat, need to be assessed carefully. The gain from ther- driven by the structural disjoining pressure gradient. Journal of Colloid and
mal conductivity could be offset by the viscosity and specific Interface Science, 280, 192–201.
Choi, S. U. S. (1995). Enhancing thermal conductivity of fluids with nanopar-
heat effects. Other areas of future research should pay more ticles. In Proceedings of the 1995 ASME International Mechanical
attention to the linkage between rheology and thermal properties, Engineering Congress and Exposition San Francisco, CA, USA.
and nanoparticle interactions and particle-fluids-surface inter- Das, S. K., Putra, N., & Roetzel, W. (2003a). Pool boiling characteristics
actions, which call for inter-disciplinary collaboration among of nano-fluids. International Journal of Heat and Mass Transfer, 46,
nanomaterial, colloid science and engineering researchers. 851–862.
Das, S. K., Putra, N., & Roetzel, W. (2003b). Pool boiling of nano-fluids on
horizontal narrow tubes. International Journal of Multiphase Flow, 29,
6. Conclusions 1237–1247.
Ding, Y. L., Alias, H., Wen, D. S., & Williams, R. A. (2006). Heat transfer of
A critical review of the state-of-the-art nanofluids research aqueous suspensions of carbon nanotubes (CNT nanofluids). International
for heat transfer application was conducted in this work, which Journal of Heat and Mass Transfer, 49, 240–250.
Eastman, J. A., Choi, S. U. S., Li, S., Yu, W., & Thompson, L. J. (2001). Anoma-
showed that our current understanding on nanofluids is still quite lously increased effective thermal conductivities of ethylene glycol-based
limited. There are a number of challenges facing the nanoflu- nanofluids containing copper nanoparticles. Applied Physics Letters, 78,
ids community ranging from formulation, practical application 718–720.
 D. Wen et al. / Particuology 7 (2009) 141–150 149

Eapen, J., Li, J., & Yip S. (2007). Mechanism of thermal transport in dilute Smalley, R. (2005). Future global energy prosperity: The terawatt challenge.
nanocolloids. Physical Review Letters, 98, 028302. MRS Bulletin, 30, 412–417.
Evans, W., Prasher, R., Fish, J., Meakin, P., Phelan, P., & Keblinski, P. (2008). Sethumadhavan, G., Nikolov, A., & Wasan, D. T. (2001). Stability of liquid films
Effect of aggregation and interfacial thermal resistance on thermal conduc- containing monodisperse colloidal particles. Journal of Colloid Interface
tivity of nanocomposite and colloidal nanofluids. International Journal of and Science, 240, 105–112.
Heat Mass Transfer, 51, 1431–1438. Tsai, C. Y., Chien, H. T., Ding, P. P., Chan, B., Luh, Y., & Chen, P. H. (2004).
Hetsroni, G., Gurevich, M., Mosyak, A., Rozenblit, R., & Segal, Z. (2004). Boil- Effect of structural character of gold nanoparticles in nanofluid on heat pipe
ing enhancement with environmentally acceptable surfactants. International thermal performance. Materials Letters, 58, 1461–1465.
Journal of Heat and Fluid Flow, 25, 841–848. Tseng, W. J., & Lin, K. C. (2003). Rheology and colloidal structure of aqueous
Keblinski, P., Phillpot, S. R., Choi, S. U. S., & Eastman, J. A. (2002). Mechanisms TiO2 nanoparticle suspensions. Materials Science and Engineering A, 355,
of heat flow in suspensions of nano-sized particles (nanofluids). International 186–192.
Journal of Heat Mass Transfer, 45, 855–863. Tillman, P., & Mill, M. (2007). Determination of nanolayer thickness for a
Kim, H., Kim, J., & Kim, M. H. (2006). Effect of nanoparticles on CHF enhance- nanofluid. International Communications in Heat and Mass Transfer, 34,
ment in pool boiling of nano-fluids. International Journal of Heat and Mass 399–407.
Transfer, 49, 5070–5074. Trokhymchuk, A., Henderson, D., Nikolov, A., & Wasan, D. T. (2001). A simple
Kim, S. J., Bang, I. C., Buongiorno, J., & Hu, L. W. (2006). Effects of nanopar- calculation of structural and depletion forces for fluids/suspensions confined
ticle deposition on surface wettability influencing boiling heat transfer in in a film. Langmuir, 17, 4940–4947.
nanofluids. Applied Physics Letters, 89, 153107. Vafaei, S., Borca-Tascius, T., Podowski, M. Z., Purkayastha, A., Ramanath, G.,
Kim, S. J., Bang, I. C., Buongiorno, J., & Hu, L. W. (2007a). Surface wet- & Ajayan, P. M. (2006). Effect of nanoparticles on sessile droplet contact
tability change during pool boiling of nanofluids and its effect on critical angle. Nanotechnology, 17, 2523–2527.
heat flux. International Journal of Heat and Mass Transfer, 50, 4105– Vassallo, P., Kumar, R., & Damico, S. (2004). Pool boiling heat transfer exper-
4116. iments in silica–water nano-fluids. International Journal of Heat and Mass
Kim, S. J., Bang, I. C., Buongiorno, J., & Hu, L. W. (2007b). Study of pool Transfer, 47, 407–411.
boiling and critical heat flux enhancement in nanofluids. Bulletin of the Wasan, D. T., & Nikolov, A. D. (2003). Spreading of nanofluids on solids. Nature,
Polish Academy of Sciences: Technical Sciences, 55, 211–216. 423, 156–159.
Koo, J., & Kleinstreuer, C. (2005). Laminar nanofluid flow in microheat-sinks. Wasekar, V. M., & Manglik, R. M. (2000). Pool boiling heat transfer in aqueous
International Journal of Heat and Mass Transfer, 48, 2652–2661. solutions of an anionic surfactant. ASME Journal of Heat Transfer, 122,
Krishnamurthy, S., Bhattacharya, P., Phelan, P. E., & Prasher, R. S. (2006). 708–715.
Enhanced mass transport in nanofluids. Nano Letters, 6, 419–423. Wen, D. S. (2008a). Mechanisms of convective heat transfer of nanofluids. In
Lee, D., Kim, J. W., & Kim, B. G. (2006). A new parameter to control heat ASME Micro/Nanoscale International Heat Transfer Conference Taiwan,
transport in nanofluids: Surface charge state of the particle in suspension. 2008.
Journal of Physics and Chemistry B, 110, 4323–4328. Wen, D. S. (2008b). Mechanisms of thermal nanofluids on enhanced critical
Li, C. H., & Peterson, G. P. (2007). Mixing effect on the enhance- heat flux (CHF). International Journal of Heat and Mass Transfer, 51,
ment of the effective thermal conductivity of nanoparticle suspensions 4958–4965.
(nanofluids). International Journal of Heat and Mass Transfer, 50, 4668– Wen, D. S., & Ding, Y. L. (2005). Experimental investigation into the pool boiling
4677. heat transfer of aqueous based-Al2 O3 nanofluids. Journal of Nanoparticle
Liu, Z. H., Xiong, J. G., & Bao, R. (2007). Boiling heat transfer characteristics of Research, 7, 265–274.
nanofluids in a flat heat pipe evaporator with micro-grooved heating surface. Wen, D. S., Ding, Y., & Williams R, L. (2006). Pool boiling heat transfer of
International Journal of Multiphase Flow, 33, 1284–1295. aqueous based TiO2 nanofluids. Journal of Enhanced Heat Transfer, 13,
Maïga, S., Nguyen, C., Galanis, N., & Roy G. (2004). Heat transfer behaviours 231–244.
of nanofluids in a uniformly heated tube. Superlattices and Microstructures, Wen, D. S., & Ding, Y. L. (2004a). Effective thermal conductivity of aqueous
35, 543–557. suspensions of carbon nanotubes (nanofluids). Journal of Thermophysics
Milanova, D., & Kumar, R. (2005). Role of ions in pool boiling heat transfer of and Heat Transfer, 18, 481–486.
pure and silica nanofluids. Applied Physics Letters, 87, 233107. Wen, D. S., & Ding, Y. L. (2004b). Experimental investigation into convec-
Nie, C., Marlow, W. H., & Hassan, Y. (2008). Discussion of proposed mech- tive heat transfer of nanofluids at the entrance region under laminar flow
anisms of thermal conductivity enhancement in nanofluids. International conditions. International Journal of Heat and Mass Transfer, 47, 5181–
Journal of heat and Mass transfer, 51, 1342–1348. 5188.
Pak, B. C., & Cho, Y. I. (1999). Hydrodynamic and heat transfer study of dis- Wen, D. S., & Wang, B. X. (2002). Effects of surface wettability on nucleate
persed fluids with submicron metallic oxide particles. Experimental Heat pool boiling heat transfer for surfactant solutions. International Journal of
Transfer, 11, 151–170. heat and mass transfer, 45, 1739–1747.
Park, K. J., & Jung, D. (2007a). Enhancement of nucleate boiling heat transfer You, S. M., Kim, J. H., & Kim, K. H. (2003). Effect of nanoparticles on critical
using carbon nanotubes. International Journal of Heat and Mass Transfer, heat flux of water in pool boiling heat transfer. Applied Physics Letters, 83,
50, 4499–4502. 3374–3376.
Park, K. J., & Jung, D. (2007b). Boiling heat transfer enhancement with carbon Witharana, S., 2003. Boiling of refrigerants on enhanced surfaces and boil-
nanotubes for refrigerants used in building air-conditioning. Energy and ing of nanofluids. Ph.D. Thesis. Royal Institute of Technology, Stockholm,
Buildings, 39, 1061–1064. Sweden.
Philip, J., Shima, P. D., & Raj, B. (2008). Evidence for enhanced thermal con- Xie, H. Q., Fujii, M., & Zhang, X. (2005). Effect of interfacial
duction through percolating structures in nanofluids. Nanotechnology, 19, nanolayer on the effective thermal conductivity of nanoparticle-fluid
305706. mixture. International Journal of Heat and Mass Transfer, 48, 2926–
Prasher, R., Phelan, P. E., & Bhattacharya, P. (2006). Effect of aggregation kinet- 2932.
ics on the thermal conductivity of nanoscale colloidal solutions (nanofluid). Xu, J., Yu, B. M., Zou, M. Q., & Xu, P. (2006). A new model for heat conduc-
Nano Letters, 6, 1529–1534. tion of nanofluids based on fractal distribution of nanoparticles. Journal of
Roy, G. C., Nguyen, C. T., & Comeau M. (2006). Numerical investigation of Physics D: Applied Physics, 39, 4486–4490.
electronic component cooling enhancement using nanofluids in a radial flow Xuan, Y. M., & Roetzel, W. (2000). Conceptions for heat transfer correlation
cooling system. Journal of Enhanced Heat Transfer, 13, 101–115. of nanofluids. International Journal of Heat and Mass Transfer, 43, 3701–
Sefiane, K. (2006). On the role of structural disjoining pressure and contact 3707.
line pinning in critical heat flux enhancement during boiling of nanofluids. Yang, Y., Zhang, Z., Grulke, E., Anderson, W., & Wu, G. (2005). Heat
Applied Physics Letters, 89, 044106. transfer properties of nanoparticle-in-fluid dispersions (nanofluids) in lam-
150 D. Wen et al. / Particuology 7 (2009) 141–150

inar flow. International Journal of Heat and Mass Transfer, 48, 1107– Zhu, H. T., Lin, Y. S., & Yin, Y. S. (2004). A novel one-step chemical method for
1116. preparation of copper nanofluids. Journal of Colloid and Interface Science,
Yu, W. H., France, D. M., Routbort, J. L., & Choi, S. (2008). Review and com- 277, 100–103.
parison of nanofluid thermal conductivity and heat transfer enhancements. Zhou, S. Q., & Ni, R. (2008). Measurement of the specific heat capacity of
Heat Transfer Engineering, 29, 432–460. water-based Al2 O3 nanofluids. Applied Physics Letters, 92, 093123.