Progress in Energy and Combustion Science 103 (2024) 101160

Contents lists available at ScienceDirect

Progress in Energy and Combustion Science

journal homepage: www.elsevier.com/locate/pecs

Direct absorption solar collectors: Fundamentals, modeling approaches,

design and operating parameters, advances, knowledge gaps, and
future prospects
Alabas Hasan a, Anas Alazzam a, b, **, Eiyad Abu-Nada a, *
a
Department of Mechanical and Nuclear Engineering, Khalifa University, Abu Dhabi, United Arab Emirates
b
System on Chip Lab, Khalifa University, Abu Dhabi, United Arab Emirates

A R T I C L E I N F O A B S T R A C T

Handling editor: Christof Schultz Direct absorption solar collectors (DASCs) based on nanofluids offer a promising solution for achieving the dual
goals of solar energy utilization: maximizing solar absorption and minimizing thermal losses. In contrast to
conventional surface absorption solar collectors, which suffer from substantial heat losses, DASCs operate by
replacing elevated-temperature absorption surfaces with nanofluid bulk for volumetric absorption. To bridge the
gap between theoretical research and commercialization, a comprehensive understanding of DASCs is essential.
This includes modeling approaches, the impact of design and operational parameters, recognizing limitations,
and evaluating future prospects. This study provides a comprehensive review with a focus on resolving dis­
agreements regarding low-flux DASC responses to specific design and operational variations that have sparked
conflicting interpretations in the literature. This review, by addressing these discrepancies, serves as an
invaluable resource for researchers seeking a more nuanced understanding of this evolving field, facilitating its
advancement into practical applications.
This review comprehensively examines the field of DASCs across eight distinct sections. Section 1 provides an
overview of solar energy’s potential, the evolution of solar collectors, and the rationale for the review. Section 2
focuses on theoretical modeling approaches for simulating colloidal suspensions in solar thermal systems,
including optical properties, radiative transport, and heat transfer mechanisms. The strengths and limitations of
these models are critically evaluated to assist researchers in selecting the most suitable one for specific colloidal
systems. Additionally, a critical assessment of analytical and numerical studies in the existing literature is pre­
sented in this section. Section 3 offers a detailed view and critical assessment of experimental efforts in the field.
The stability of nanofluids is discussed in section 4, while sections 5 and 6 analyze the impact of operating
conditions, geometry, design parameters, and flow properties on DASC performance criteria. We address con­
tradictions and ambiguities in the effects of some operating variables in the DASC literature, considering state-of-
the-art simulation techniques. Section 7 focuses on economic and environmental analyses related to DASCs,
providing insights into their feasibility and sustainability. Finally, Section 8 synthesizes conclusions from the
reviewed literature, identifies research gaps, and proposes future directions based on recent advancements in
DASC technology.

* Corresponding author.
** Corresponding author. Department of Mechanical and Nuclear Engineering, Khalifa University, Abu Dhabi, United Arab Emirates.
E-mail address: [email protected] (E. Abu-Nada).

https://doi.org/10.1016/j.pecs.2024.101160
Received 2 February 2024; Received in revised form 9 April 2024; Accepted 13 May 2024
Available online 24 May 2024
0360-1285/© 2024 Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

Nomenclature
alternative energy resources and thus the focus of researchers’ attention
for the past few decades.
Symbols Greek

A Collector area (m2 ) α Particle size parameter 1.1. Why solar energy?
Am Solar-weighted absorption αt Thermal diffusivity
fraction (m2 /s)
Solar energy is renewable, free, abundant, clean with nearly zero-
Aλ Absorbance γr Radiation propagation
cosine emission conversion processes, and available for appropriate periods
AR Aspect ratio of the DASC η Efficiency in most of the world countries [1]. It offers considerable amounts of safe
cp Specific heat capacity (J/kg ​ K) ηPC Photo-thermal energy for different small and large energy applications reliably,
conversion efficiency affordably, and practically [2]. Solar incidence on the Earth’s surface in
Dp Particle diameter (m) ηІ First-law efficiency
EX Exergy rate (W) ηІІ Second-law efficiency
1 h is greater than the whole energy consumed by humans in one year,
fV Volume fraction κ Absorptive index and the incidence in one year (∼ 3, 400, 000 EJ) is more than 7500
g Gravitational acceleration, g = kB Boltzmann constant, times the whole primary energy consumed during the same interval
9.8 m/s2 1.38 × 10− 23 J/K (∼ 450 EJ) [3]. Aforementioned advantages explain why the depen­
Gs Irradiation (W/m2 ) λ Wavelength (μm) dence on solar energy increases at an accelerated pace with an annual
h Planck constant, h = 6.626 × Dynamic viscosity
μ
growth of around 8.3 % [4]. Solar energy is expected to be the
10− 34 J s (kg/m s)
fastest-growing power production resource within the period
hL /hE Combined heat loss coefficient ρ Density (kg/m3 )
(W/m2 K) 2012–2040 [4] reaching to the greatest share of universal electricity
H Collector height/nanofluid film Superscripts and subscripts production in 2050 [5]. Moreover, it is planned that it accounts for more
thickness (m) than two thirds of the universal energy consumption by 2100 [5,6,7].
I Spectral radiative heat flux a Absorption/absorbed The remarkable attention paid to solar energy is also attributed to the
(W/m2 μm)
continuous declination in solar systems’ costs that lower the prices of
k/ kc Thermal conductivity (W/m K) amb Ambient
vp Particle velocity (m/ s) bf/f Base fluid electricity production, in high solar-incidence countries like the UAE in
Ka Absorption coefficient (μm− 1 ) e Extinction 2018, to $0.024/kWh, i.e., less than the electricity produced by con­
Ke Extinction coefficient (μm− 1 ) eff Effective ventional methods [8].
Ks Scattering coefficient (μm− 1 ) in Inlet Solar energy can be converted into three different forms of useful
L Collector length (m) max Maximum energy: chemical energy through the photosynthesis process as in plants;
m Complex refraction index nf Nanofluid
ṁ Mass flow rate (kg/s) np Nanoparticle
thermal energy through a photo-thermal conversion process via solar
n Refractive index o Incident thermal systems; and electricity through a photoelectric conversion
P Pressure (kPa) out Outlet process via photovoltaic (PV) cells [9]. For electric power generation
Pe Peclet number p Particle purposes, PV cells have a privilege over other techniques as they directly
Pr Prandtl number s Scattering
convert the incident solar radiation within the ultraviolet–visible
qr Total radiative heat flux (W/ Acronyms
m2 )
(UV-VIS) spectrum into electric power [10], which significantly lowers
Qa Absorption efficiency DASC Direct absorption solar the electricity prices. On the other hand, solar thermal systems can
collector convert solar radiation in the entire spectrum into thermal energy
Qe Extinction efficiency DI water De-ionized water through a photothermal conversion process [11] and deliver it via a heat
Qs Scattering efficiency DLS Dynamic light scattering
exchanger to a working fluid like water, oil, air, ..etc. Then, thermal
Q̇u Useful power (kW) DNS Direct Numerical
Simulation energy can be employed for household use, thermal applications like
Rλ Reflectance EG Ethylene glycol HVAC, or industrial processes. Alternatively, it can be converted into
Re Reynolds number IR Infra-red electricity via a power cycle which requires a concentrated solar power
t Time (s) LSPR Localized surface (CSP) technology [12]. The hybrid technology, i.e., the
plasmon resonance
T Temperature (K) MWCNT Multi-wall carbon
photovoltaic-thermal (PV-T) technology, combines the two virtues of
nanotube photoelectric and photothermal conversions in one system and yields
Tλ Transmittance NSE Navier-Stokes equation highly efficient solar systems, as reviewed by Herrando et al. [13]. Based
u, v x and y velocity (m/ s) RSM Response Surface on the synergistic effect, such a hybrid system obtains a higher overall
Methodology
efficiency than a PV or a solar thermal system alone. However, in
V Volume (m3 ) SASC Surface absorption solar
collector comparison with the photoelectric conversion, the photothermal con­
W Collector width (m) SWCNT Single-wall carbon version is an efficient, straightforward, and less complex strategy for the
nanotube solar energy utilization [14]. Besides, solar thermal technologies pro­
x, y, z Spatial Coordinates UV Ultra-violet vide energy for several thermal applications that account for a consid­
X, Y x
Normalized position, X = , erable portion of the universal demand, even including water
L
y desalination [15]. This justifies the significant increase in the
Y =
H
solar-to-thermal applications scale and the considerable attention paid
to photothermal conversion systems to become the most common solar
energy utilization form [16].
1. Introduction
1.2. Solar thermal collectors
The continuously increasing demand on energy, due to the rapid
population and civilization growth, is accompanied by environmental Solar collectors harvest solar radiation, convert it into thermal en­
concerns because of the problematic dependence on conventional en­ ergy, and convey it to a working fluid. Depending on their operating
ergy resources such as the fossil fuel that contributes the biggest portion temperature, solar collectors are categorized into three groups: low
of the world energy. Climate change and depletion of conventional en­ temperature collectors (up to 100 ◦ C), medium (between 100 ◦ C and
ergy resources concerns have pushed the scientific community to 400 ◦ C), and high (above 400 ◦ C) [17]. Each category applies a different
transmit the world from dependence on conventional resources to technology and working fluid. Water represents a reasonable choice for
dependence on alternative green and sustainable ones such as solar operating temperatures up to 90 ◦ C, oils, glycols and glycerol for tem­
energy. Many aspects have made solar energy one of the most promising peratures between 90 and 175 ◦ C, and molten salts or gases for higher

2
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

temperatures [18]. A further classification of solar collectors can be

found in the literature [19]: non-concentrating (stationary) collectors in
which absorption of direct solar flux is performed by an absorbing sheet
or volume (such as flat plate solar collectors (FPC) [20] and evacuated
tube solar collectors (ETC) [21]), and concentrating collectors in which
reflecting mirrors are employed to collect and focus the incident sunlight
onto a smaller area where sun tracking systems are needed for efficient
collection and concentration (such as parabolic trough collector (PTC)
[22], linear Fresnel collector (LFC) [23], parabolic dish concentrator
(PDC) [24], and solar tower with a heliostat field [25]). Moreover, the
collector geometry plays a main role in the temperature that can be
reached. For instance, FPCs and ETCs attain relatively low temperatures,
PTCs attain medium temperatures, and central towers and PDCs can
attain elevated ones [26].
Academic investigations of solar collectors involve efforts to attain
maximum solar absorption and minimum heat losses, thereby maxi­
mizing the useful energy acquired by the working fluid and thus the
photothermal conversion efficiency (ηPC ). ηPC can be defined as the ratio
of useful thermal energy gained by the working fluid to solar energy
input [27]:
Fig. 1. Thermal resistance circuit schematic for (a) SASC, and (b) DASC.
ṁcp (Tout − Tin )
ηPC = (1)
Gs A throughout the working fluid, instead of being localized at the top like in
SASC. Fig. 1 shows the thermal resistance (R) network in the (a) SASC
where ṁ, cp , Tin , and Tout are the mass flow rate, specific heat, inlet
and (b) DASC, and elucidates how direct volumetric absorption in
temperature, and outlet temperature of the working fluid, respectively.
DASCs mitigates the severe convective and radiative thermal losses
Gs and A are the irradiation (solar radiative heat flux per unit area) and
area of the solar collector, respectively. These objectives represent the (qʹ and qʹ , respectively) since the DASC’s upper surface temperature
conv rad
essence of solar thermal technology enhancements either for surface or is less than that of the absorber plate placed at the top of the SASC
volumetric absorption technologies as being discussed later. system. Also, it shows that the conductive and convective thermal re­
sistances (Rcond and Rconv , respectively) between the absorber and the
1.2.1. Conventional surface absorption solar collector working fluid no longer exist in the DASC system due to the absence of
In such systems, incident radiation is absorbed by an outer surface an absorber [31]. This, in theory, results in a more homogeneous
coated with a highly absorptive material designed for the solar spec­ working fluid temperature field in the DASC [32] and thus lower losses
trum, such as a conductive metal sheet painted with TiNOx, where it is and higher ηPC [33]. However, the radiative absorption thermal resis­
then converted into thermal energy. Thereafter, it is transferred to the tance (Rabs ) influences both systems.
working fluid flowing behind the sheet by conduction and convection. Introducing “black” liquids in the photothermal conversion tech­
However, the collector absorption surface (absorber) undergoes nology as an absorbing medium opened up new research paths to
elevated temperatures creating a great temperature difference with the improve solar thermal systems. The ηPC of the volumetric absorption
surroundings leading to severe losses and thus degradation in ηPC . The collector is governed by the working fluid absorption ability which is
declination of ηPC with very high absorber temperatures is attributed to poor for most typical thermal fluids [34]. For instance, more than 50 %
the relatively-high emissive losses occurring by virtue of the thermal re- of the incident radiation was transmitted through a water-ethylene
radiation’s fourth-order dependence on the absorber temperature [28]. glycol mixture when used as a working fluid in a volumetric absorp­
As shown above, conventional surface absorption solar collectors tion solar collector with a depth of 5/16 in [30]. As a solution, it was
(SASCs) suffer from high amounts of heat losses by conduction, con­ proposed to enhance the technology by changing the absorbing medium
vection, and radiation that increase with temperature [29], and from from a conventional purely fluidic medium to a particle-laden fluid. The
thermal resistances through the conductive and convective heat transfer effect of microparticles in different heat transfer applications was first
between the absorber and working fluid. These limitations drove the discussed by Hunt [35]. Suspended microparticles were found very
research community to seek different solar energy harvesting techniques useful in photothermal conversion technologies where the suspension of
with mitigated losses. proper micron-sized particles in the working fluid enhances the working
fluid absorption ability [34]. Abdelrahman et al. [34] attributed that to
1.2.2. Volumetric absorption collectors the unique absorption properties of microparticles. However, despite the
Toward suppressing the thermal losses from which conventional enhancement witnessed in the volumetric absorption collector perfor­
SASCs suffer, Minardi and Chuang [30], in an experimental work in mance by virtue of adding micron-sized particles, some issues emerged
1974, proposed a novel solar collector called Direct Absorption Solar in the system, such as fouling of pipes and clogging of pumps and valves
Collector (DASC). The researchers utilized highly absorbent “black” due to particle clustering and agglomeration in the working fluid.
liquids (liquids blended with India ink) flowing in transparent channels Studies have indicated that such phenomena cause the absorption
to directly absorb solar radiation without the need to heat any other spectra to be governed by the bulk fluid medium rather than dispersed
parts in the system. In addition, the absence of a metallic absorber particles [30,34].
eliminates the corrosion issues that threaten SASCs. Also, authors
observed that the new collector significantly lowers thermal losses. 1.2.3. Nanofluid-based direct absorption solar collectors
Unlike SASC where heat is absorbed on the top surface, DASC offers a Nanotechnology has revealed superior physical and chemical nano­
distinct advantage. In DASC, the working fluid directly absorbs solar materials’ attributes that are distinctive from their bulk counterparts
energy throughout its volume. This leads to a more even distribution of [36]. As a result, the proposition emerged to decrease the particles
heat within the fluid, rather than concentrating it at the top surface. In initially sized at the microscale to the nanoscale range. This trans­
simpler terms, with DASC, the hottest parts are, in theory, spread formation has led to the development of novel working fluids known as

3
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

"nanofluids". Nanofluids, also referred to as nanodispersions and nano­ working fluid.

colloids, are colloidal suspensions of engineered nanoparticles, with at Early pioneer numerical [46,47] and experimental [48] efforts were
least one dimension of the nanoparticles falling within the range of reported in the literature to analyze nanofluid-based DASCs. In­
1–100 nm. The synthesis of nanofluids was initially achieved by Masuda vestigations affirmed that the presence of such ultra-small particles in
et al. [37], with the term ’nanofluid’ coined by Choi and Eastman [38]. the working fluid enhances the scattering and absorption of incident
Nanoparticles offer a significant advantage over micro-size particles radiation (i.e., enhances the light extinction) and, consequently, ηPC
in conventional piping and pumping systems due to their minimal [49]. Researchers realized that optical characteristics of nanofluids
clogging potential. Additionally, they pose fewer risks of erosion or should be sufficiently strong to be capable of capturing solar radiation,
instability. Furthermore, nanoparticles can dramatically alter the optical and their radiative behavior should be clearly understood for a proper
and thermophysical properties (OTPs) of the working fluid [34,39,40]. design of the nanofluid-based DASC systems [49]. Systematic in­
An additional advantage of nanofluids emerges when using metallic vestigations of complex physics-based models to study the radiative
nanoparticles that might be smaller than the mean free path of the behavior of nanofluids can be found in the literature [50,51], and will be
conduction electrons in their bulk metallic material, which induces the thoroughly addressed in section 2.1. Besides the radiative properties,
Localized Surface Plasmon Resonance (LSPR) phenomenon. LSPR is an many pieces of research showed that the presence of nanoparticles en­
optical phenomenon that occurs when light interacts with conductive hances the thermophysical properties of the working fluid such as
nanoparticles that are smaller than the incident wavelength. In LSPR, thermal [52–54] and mass [55] diffusivities. Phelan et al. [56] reviewed
the electric field of incident light collectively excites electrons in the the early research works and discussed nanofluids prospect in several
conduction band of the nanoparticles. This resonant interaction leads to heat transfer applications. Benefiting from attempts discussed earlier,
coherent localized plasmon oscillations, and the resonant frequency Tyagi et al. [57] laid the groundwork for numerically modeling
strongly depends on the nanoparticle composition, size, geometry, nanofluid-based DASCs by presenting and solving a 2D heat transfer
dielectric environment, and separation distance. The electrons oscilla­ model (coupled energy balance and radiative transport equations) to
tion causes the surface plasmon to be excited due to the resonant transfer predict the ηPC of a nanofluid-based DASC including the medium scat­
of energy from the incident photon to the plasmon yielding a significant tering and absorption impacts for comparison with a conventional SASC.
increase in the radiative energy absorption [41]. Thus, the spectrum of Authors found that the absorption ability of an Al/water nanofluid is
light that can be absorbed is expanded while keeping a distinguished nine times higher than that of pure water. Additionally, they stated that
absorption peak, which opens the door for further improvement in the nanofluid-based DASC showed a 10 % higher ηPC than a typical SASC
solar radiation absorption efficiency [42]. Therefore, it can be stated under the same conditions. It is evident that the 10 % increase in ηPC is
that nanofluid-based collectors outperform conventional ones under the considered slight compared to the ninefold increase in the absorption
same working conditions [43]. Employing nanofluids in DASCs repre­ ability of the working fluid. This discrepancy can be attributed to the
sented a breakthrough in the volumetric absorption technology and heightened thermal losses, which occur alongside the temperature gain
announced the launch of nanofluid-based DASC research field. There­ in the DASC. Further elaboration on this topic can be found in sections
after, huge experimental, analytical, and numerical research efforts have 5.1, 5.5, and 6.2.
been put to investigate the nanofluid-based DASC performance at After Tyagi et al. [57] had introduced a numerical model for a
different operating variables, determine their optimum values, and nanofluid-based DASC system, it was immediately followed by an
propose enhancement techniques. experimental research work conducted by the same research group,
The key differences between the conventional collector and Otanicar et al. [34]. The authors verified the benefits of using nanofluids
nanofluid-based DASC can be observed in Fig. 2. It can be seen that solar in DASCs and their potential for improvement by adjusting operational
radiation is absorbed by an absorbing surface in the conventional col­ parameters to achieve optimal volumetric light absorption. Thereafter,
lector and then energy is delivered to the working fluid running inside the field has witnessed very fast growth. A huge number of research
the channel by conduction and convection. A vacuum chamber between papers tackled the DASC system from many different viewpoints and
the absorbing surface and outer glass is placed to eliminate conductive compared nanofluid-enhanced DASCs with conventional systems using
and convective losses to surroundings. However, this does not prevent different approaches. For instances, Luo et al. [58] experimentally and
the severe re-radiative losses due to the absorbing surface’s high tem­ numerically compared a pure-fluid-based collector with a
perature. On the other hand, radiation is directly absorbed by the bulk nanofluid-based DASC and found that increments of up to 25 % and 100
volume of the nanofluid working in the DASC, which keeps the system K in DASC photothermal conversion efficiency and outlet temperature
from surface hot spots that boost severe re-radiative losses. Also, this gain, respectively, can be obtained by utilizing nanofluids instead of a
reduces the conductive and convective losses to surroundings because, pure fluid. Another comparison by Delfani et al. [59] showed 10–29 %
in theory, the temperature peak exists within the working nanofluid in enhancement in DASC performance when using nanofluids. Similar
the middle of the collector rather than on an external surface as in observations were demonstrated by Refs. [34,60]. Vakili et al. [61]
conventional solar collectors [44,45]. In addition, volumetric solar ab­ experimentally confirmed that using graphene/deionized-water
sorption eliminates the conductive and convective resistances (see (DI-water) nanofluid in DASC enhanced the collector ηPC by up to 33
Fig. 1b), that conventional SASCs experience (see Fig. 1a), due to the % due to improving the extinction coefficient (Ke ) of the working fluid.
absence of heat transfer processes from an absorbing sheet to the The extinction coefficient is an intrinsic property of the material that

Fig. 2. Schematic of radiation interactions in a conventional collector and a nanofluid-based DASC.

4
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

determines how strongly it reflects or absorbs light at a certain wave­ thoroughly examined in the current review. This review presents all
length. Also, Gupta et al. [62] could obtain an improvement of 39.6 % in possible behaviors and discusses all viewpoints with deep justifications
ηPC of a DASC under certain operating conditions of Al2O3/water in light of the state-of-the-art advances in modeling techniques that try
nanofluid mass flow rate (ṁ) and volume fraction (fV ). fV is the ratio of to capture the exact response of DASCs to these problematic parameters.
the volume of dispersed nanoparticles to the base fluid volume. More­ Detailed explanations of the dissimilarities of DASC performance criteria
over, Gorji and Ranjbar [63] experimentally and numerically investi­ behavior, which emerge when gathering certain parameters together
gated a DASC employing three nanofluids with DI-water host (base) with certain values, are also provided.
fluid and graphite, magnetite, and silver nanoparticles. They obtained Due to the vast and diverse literature on nanofluid-based DASCs, a
an enhancement of up to 57 % and 20 % in energetic and exergetic ef­ critical review is necessary to categorize, compare, and analyze existing
ficiencies, respectively. research. This work fills this gap by providing a comprehensive over­
view encompassing various perspectives. The first part delves into the
1.3. The need for a review paper in the field of nanofluid-based DASC state-of-the-art experimental, analytical, and numerical modeling ap­
systems proaches used to investigate DASC performance and nanofluid proper­
ties and stability. It critically examines these methods, highlighting
Numerous authors have either reviewed nanofluid-based DASCs limitations and inconsistencies in reported results. Additionally, an
specifically or incorporated them into their broader reviews on nano­ overview of nanofluid stability is presented. The second part explores
fluid applications in solar systems. Some of these reviews [64–66] the influence of factors like nanofluid properties (base fluid type,
focused on recent advances in nanofluid-based solar collectors, nanoparticle characteristics), collector geometry (including recent
including FPCs, ETCs, DASCs, and PV-Ts, addressing the physical design advancements), and operating conditions (flow rate, tempera­
mechanisms, forces affecting the heat transfer in nanofluid flows and the ture, dimensionless numbers) on key performance indicators (photo­
thermophysical properties, heat transfer models, and evaluation criteria thermal conversion efficiency, temperature gain, energy losses). Finally,
of nanofluid, and compared them with ones employing conventional the review addresses the economic feasibility and environmental impact
working fluids. Also, they underscored the importance of integrating the of nanofluid-based DASC. It concludes by highlighting key advance­
utilization of nanofluids and turbulence generators within solar collec­ ments, remaining challenges, knowledge gaps, and future research di­
tors to enhance the photothermal conversion efficiency and maximize rections in light of the latest developments.
the extraction of useable solar energy. However, it has been observed
that the DASC system was mentioned as an exemplar among solar col­ 2. Theoretical analysis of DASC systems
lectors employing nanofluids without delving comprehensively into
deep discussions concerning researchers seeking an exhaustive under­ The basic principle behind the nanofluid-based DASC is that the
standing of the in-depth knowledge of DASC matters specifically. The working fluid, whose OTPs are enhanced by virtue of the suspended
current review expands on the investigation of DASCs and proposes a nanoparticles, directly absorbs solar radiation, as illustrated in the DASC
roadmap for developing a reliable DASC with optimal performance schematic in Fig. 3, and acquires thermal energy throughout a single-
under realistic conditions. Such a roadmap is urgently needed when step solar-to-thermal conversion process. The presence of nano­
transmitting the technology from labs to the real world. particles considerably enhances the OTPs of typical host heat transfer
Recent developments in the field of nanofluids used in DASCs have fluids and thus the ability of solar absorption and conversion into useful
been reviewed, focusing on preparation techniques, optical behaviors, thermal energy [68,69]. Theoretical modeling approaches often used in
solar thermal energy conversion performance, and physical and thermal the literature to simulate colloidal suspensions and reveal their radiative
stability [9]. The article is structured around key aspects such as the and thermal behaviors are discussed in this section.
effects of nanoparticle material types on DASC performance, as well as
the stability and optical characterization of nanofluids. Additionally,
state-of-the-art solar heat installations for industrial processes and
conventional SASCs were reviewed, along with a comprehensive sum­
mary of studies addressing concentrating and non-concentrating DASCs
[32].
Previous reviews have only superficially addressed the development
of modeling techniques and their limitations, resulting in inconsistencies
in the behavior of DASCs under various operating parameters, as
documented in the literature. There is also a knowledge gap regarding
the detailed effects of various design and operating variables on DASC
performance, which is crucial for enriching researchers’ understanding
in this area.
This review offers a comprehensive discussion of all aspects of
DASCs, from the physics of the photothermal conversion process to the
effects of all design and operating variables affecting system perfor­
mance. The goal is to provide readers with a detailed document covering
most details of DASCs’ modeling and parametric investigations. The
review traces the evolution of the DASC system, starting from its
modification of the conventional SASC to the latest advances in
modeling and investigation attempts and state-of-the-art design
modifications.
The impacts of different design and operating parameters on the
performance of DASCs have been discussed, and optimum ranges have
been suggested [67]. However, a single point of view was adopted, and
the contradictions evident in the literature regarding the DASC response
to several parameters were not discussed. These contradictions are
attributed to the limitations of modeling approaches, which are more Fig. 3. Schematic diagram of a nanofluid-based DASC.

5
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

Fig. 4. Spectrum of Electromagnetic radiation [70]. (Reused with permission from Wiley).

2.1. Optical characterization of nanofluids scattering/absorption coefficient indicates a material with a more
frequent scattering/absorption process, leading to a more rapid decrease
Fig. 4 illustrates the complete spectrum of electromagnetic radiation. in light intensity as it travels through. The complex refractive index of a
Gamma rays, X-rays, and UV radiation with shorter wavelengths are material incorporates the material’s ability to both refract and absorb
particularly relevant to high-energy physical and engineering applica­ light. The real part (n), so-called refractive index, is a measure of how
tions, whereas microwaves and radio waves with longer wavelengths much the speed of light is altered when it enters the material (i.e. the
(exceeding 105 μm) are of interest to electrical engineers. The interme­ ratio of the speed of light in vacuum to its speed in that material). Ma­
diate segment of the spectrum, ranging from approximately 0.1 to 100 terials with a higher refractive index bend light more than materials
μm, encompasses a portion of the UV spectrum as well as the entire with a lower index. The imaginary part of the complex refractive index
visible and infrared (IR) spectrum. This intermediate range is referred to (κ), so-called absorptive index, describes how much light is absorbed by
as thermal radiation, as it both is generated by and influences the the material as it travels through. A higher κ value indicates stronger
thermal state or temperature of matter. Nevertheless, in the context of absorption. Many factors have direct effects on the out-scattering and
nanofluid-based DASC systems, the wavelength range that is used in the absorption of light passing through a nanofluid such as the base fluid
volumetric heat absorption calculations is from 200 to 2000 nm [41]. material type, and nanoparticle size, shape, loading, and material type.
The interaction between solar radiation or any light wave and a Also, the incident light intensity has a main role in determining the
medium containing fine particles, i.e. nanofluid, can be described by a nanofluid’s optical behavior, where experiments by Vakili et al. [61]
phenomenon called the extinction, i.e., attenuation of light while pass­ demonstrated that the investigated DASC areas, where incident radia­
ing through a medium by out-scattering and absorption [71]. A solution tion intensity was high, experienced less transmittance. The theoretical
to the light scattering by such small particles is the Rayleigh regime. The models proposed by researchers to optically characterize nanofluids
Rayleigh regime refers to a specific scenario in light scattering where the depend on such factors. This section is dedicated to describing the basic
particles causing the scattering are much smaller compared to the theoretical optical approaches most widely accepted in the literature to
wavelength of the light itself (i.e., the particles responsible for scattering determine the optical characteristics of nanofluids such as the Mie
have a diameter less than about one-tenth the wavelength of the light scattering, Rayleigh scattering, Discrete Dipole Approximation (DDA),
[72]). However, if the particle size is equal to or larger than the wave­ and Maxwell-Garnett approaches.
length of light, a different scattering regime, called Mie scattering, be­
comes more relevant. Useful parameters are employed in literature to 2.1.1. Rayleigh scattering approach
quantify such phenomena. Scattering efficiency (Qs ) describes how For justified adoption of the Rayleigh scattering approximation for
effective a particle is at redirecting light from its original path. In optical characterization of nanofluids, two conditions should be avail­
contrast to scattering efficiency, absorption efficiency (Qa ) describes able: dilute suspension and very fine spherical particles. When the sus­
how effective a particle is at absorbing (i.e. capturing and retaining) the pension is dilute enough, the impacts of dependent scattering can be
incident light’s energy. This absorbed energy can be converted into heat. ignored. Namely, light extinction by a particle is assumed to be unaf­
Both scattering/absorption efficiencies are dimensionless quantities, fected by the adjacent particles. This facilitates the evaluation of the
often expressed as percentages, and representing the ratio of the power radiative characteristics and light extinction within a particle cloud
of the scattered/absorbed light to the power of the incident light. The through a summation of those radiative characteristics or light extinc­
dominant mechanism governing radiation extinction by ultra-fine tion of individual particles. Independent scattering implies that the ratio
nanoparticles, particularly in the Rayleigh regime, is absorption. The between the particle diameter (Dp ) and the incident light wavelength
dominance of absorption, in Rayleigh regime, is attributed to the fact (λ), i.e., the particle size parameter (α = πDp /λ), is very small (α≪1). To
that the scattering efficiency is much lower, by several orders of explain more, the existence of neighboring particles can categorize
magnitude, than the absorption efficiency [73]. Such physical matters scattering in a participating medium into dependent and independent. If
will be detailed in this section. light extinction (i.e., absorption and scattering) of a particle is influ­
Optical characterization of a nanofluid means the determination of enced by scattering from neighboring particles, scattering is dependent.
its extinction coefficient (Ke ), represented by the sum of scattering co­ However, if particles are adequately distant from each other (i.e.,
efficient (Ks ) and absorption coefficient (Ka ), complex refractive index interparticle clearance-to-wavelength ratio ≫ 1) where scattering by a
(n + iκ), and other optical constants. The scattering/absorption coeffi­ particle is not influenced by the existence of neighboring particles, then
cient describes how quickly the intensity of a light beam weakens due to independent scattering assumption is valid. Interparticle clearance (CIP )
a scattering/absorption process within the working fluid. A higher is the space or distance between particles, defined as: CIP =

6
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

2.77Dp fV− 0.381 [74]. Thus, based on the independent scattering assump­ function of fV . Therefore, it possibly, depending on the particle size,
tion, the overall intensity of light scattered from a particle cloud can be allows for nanofluids with higher concentrations (i.e., fV > 0.006) to be
considered as the sum of the intensities of light scattered from individual considered as independent scatterers, thus violating Tien and Drolen’s
particles [75]. Furthermore, under the independent scattering assump­ [79] limit condition. Kerker et al. [82] reported that the error associated
tion, amounts of radiation scattered and absorbed by a cloud of very fine with the Rayleigh approximation (i.e., the assumption that dependent
spherical particles (α≪1) is directly proportional to fV [71]. Dependent scattering effects are negligible in the Rayleigh scattering regime) is less
scattering effects can be neglected in case of solar nanofluids as they, than 4 % when α < 0.3. Again, the assumption of independent scattering
which are the nanofluids employed in solar applications, as they can is commonly accepted when fV < 0.01 and particles are sufficiently
enhance the optical characteristics of the solar system, by orders of small [34,57]. However, in a separate investigation conducted by Kumar
magnitude, with tiny volume fractions (fV < 0.001 [76]) and ultrafine and Tien [83], a discrepancy of approximately 5 % in absorption effi­
nanoparticles (Dp < 40 nm [77]). Fig. 5 represents the ranges of inde­ ciency was identified between independent and dependent approaches
pendent and dependent scattering according to fV (on the horizontal axis when considering Rayleigh particles with α between 0.01 and 0.1 and at
as a value between 0 and 1) and α, and shows the location of solar fV of 0.06. Conversely, disparities in scattering efficiency between in­
nanofluids on the scattering regime map. dependent and dependent scattering regimes were observed to be more
According to Tien and Dorlen’s [79] condition, the independent pronounced and notably contingent on the specific dependent scattering
scattering approximation is applicable in nanofluid optical character­ model employed. Nevertheless, recognizing absorption as the predom­
ization studies if fV < 0.006. Nevertheless, Otanicar et al. [68] incor­ inant mechanism for radiation extinction by small absorbing particles in
porated dependent and multiple scattering effects into their model, the Rayleigh regime—wherein the scattering efficiency is frequently
comparing the DASC photothermal conversion efficiency in scenarios of disregarded due to its substantially lower magnitude compared to ab­
independent scattering only and dependent and multiple scattering. sorption efficiency—it can be asserted that adopting fV < 0.06 as a
Investigating 30-nm graphite nanoparticles, the study revealed that the threshold permits the neglect of dependent effects in solar nanofluids
impact on collector photothermal conversion efficiency due to surpass­ (for sufficiently small particles Dp < 40 nm [84]) with an error of no
ing the independent scattering limit proposed by Tien and Drolen [79] more than 5 % [85]. Consequently, numerous studies in the literature
(i.e., fV < 0.006) remains below 1 % at fV of up to around 0.05. Addi­ have applied Rayleigh scattering theory to assess nanofluid optical
tionally, Wei et al. [80] conducted a comparison between the indepen­ properties with fV exceeding the limit of Tien and Drolen [79] (i.e., fV =
dent scattering approximation and other approaches considering 0.006). This practice was applied in the works of Tyagi et al. [57] (fV up
multiple and/or dependent scattering effects. Their conclusion asserts to 0.05), Otanicar et al. [34] (fV up to 0.01), and Saidur et al. [44] (fV up
that, for fV > 0.05, in instances where the base fluid is a weak absorber to 0.08).
(e.g., water in the visible spectrum) and small nanoparticles, all models Such an assumption simplifies the optical characteristics’ calcula­
result in comparable outcomes. These observations by Refs. [68,80] tions. The nanoparticle extinction coefficient (Knp e ) can then be calcu­
align more closely with an alternative independent scattering limit lated, based on the Rayleigh scattering approach, as follows [73]:
condition introduced by Howell et al. [81]:
3fV Qe
( ) Knp
e = (3)
2.77 λ Dp
Dp 0.1 + ( )0.381 > (2)
100fV 2
where Qe is the nanoparticle extinction efficiency which accounts for the
This limit condition is more-relaxed than that introduced by Tien and total of absorption efficiency (Qa ) and scattering efficiency (Qs ):
Drolen [79] as it is stated based on Dp and CIP which, in turn, is a Qe = Qa + Qs (4)

where Qa and Qs , based on the Rayleigh model, can be given, respec­

tively, as follows [69]:
{ 2 [ ( ) ]}
m − 1 α2 m2 − 1 m4 + 27m2 + 38
Qa = 4αIm 1+ (5)
m +2
2 15 m + 2
2 2m + 3
2

⃒( )⃒2
8 ⃒ m2 − 1 ⃒⃒
Qs = α4 ⃒⃒ (6)
3 m +2 ⃒
2

m is the nanofluid’s relative complex refractive index given as follows

[86]:
nnp + iκnp
m= (7)
nbf

Scattering effects can be neglected for some liquids that are classified
as strong absorbers, and thus absorption is considered as the only
attenuation source in the nanofluid [69]. Consequently, the base fluid’s
extinction coefficient (Kbfe ) is assumed to be equal to the absorption
coefficient of the base fluid and calculated as:
4πκbf
Kbf
e =
(8)
λ
Finally, the effective extinction coefficient of the nanofluid (Knf
e ) is
calculated based of the first-order method in which Kbf e and K e are
np

combined together to result in Knfe in a satisfactory manner [87]:

Fig. 5. Scattering regime map. CIP is the inter-particle clearance [78]. (Reused Knf (9)
e = Ke + Ke
bf np
with permission from Elsevier).

7
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

2.1.2. Maxwell-Garnett approach

This approach is frequently used in the literature to optically char­
acterize composite materials and compute the complex dielectric func­
tions (ε) of nanofluids. The dielectric function, ε = εʹ + iεʹ , shows how
the optical characteristic of a material depends on the wavelength of the
incident radiation. εʹ and εʹ are the real and imaginary dielectric func­
tions, respectively, given as follows [71]:

εʹ = n2 − κ2 (10)

εʹ = 2nκ (11)

The effective values of n and κ for the nanofluid is determined based

on information related to the nanofluid’s complex dielectric function
(εnf ) that is given, under the independent scattering assumption [34], as
follows [73]:
⎡
εnp − εbf ⎤
3fV
⎢ εnp + 2εbf ⎥ ⎥ = εʹ + iεʹ
⎣1 +
εnf = εbf ⎢
εnp − 2εbf ⎦ nf nf (12)
1 − fV
εnp + 2εbf
The effective real (nnf ) and imaginary (κnf ) parts of the nanofluid’s
relative refractive index (effective refractive and absorptive indices,
respectively) are given, respectively, as follows [73]:
1
⎛( )1 ⎞2
εʹnf2 + εʹnf2 + εʹ
2
nf (13)
nnf = ⎝ ⎠
2

1
⎛( )1 ⎞2
εʹnf2 + εʹnf2 − εʹnf Fig. 6. Algorithm of Maxwell-Garnett approach.
2
(14)
κnf = ⎝ ⎠
2
with fine spherical particles embedded in a host fluid when the particle
Using Eqs. (10), and (11), n and κ values of the nanoparticle and base size is too large to use the Rayleigh approximation (which requires α≪1)
and too small to use geometric optics (requiring α≫1). The Mie scat­
fluid are first converted into εʹ and εʹʹ values. These values are then used
tering equations are preferred due to their ability to yield intricate
in Eq. (12) to obtain the εnf , which, in turn, is used in Eqs. (13, and 14) to
scattering parameters, including efficiencies of scattering, absorption,
find the nanofluid effective values of nnf and κnf . This algorithm is better
and backscattering (i.e., a phenomenon wherein radiation is scattered at
explained in Fig. 6.
angles exceeding 90◦ relative to its original incidence direction). Knp
s and
In DASC systems, reflection losses occur at the interface of the
Knp
e can be, respectively, obtained based on the Mie scattering model as
nanofluid with the glass plate and at the interface of the glass plate with
follows [71,73]:
air. The reflectance at an interface is determined by applying the Fresnel
[ ]
relation under normal incidence between media A and B as follows [88]: 1.5fV 2 ∑ ∞ ( )
2 2
Knp = (2n + 1) |a n | + |bn | (17)
(nA − nB )2 + (κA − κB )2
s
Dp α2 n=1
RA/B = (15)
(nA + nB )2 + (κA + κB )2 [ ]
1.5fV 2 ∑
∞

One can use nnf and κnf obtained from the Maxwell-Garnett model in Knp
e = (2n + 1)Re(a n + b n ) (18)
Dp α2 n=1
Eq. (15) to obtain the reflectance at the nanofluid/glass interface.
nf
Finally, the nanofluid’s effective extinction coefficient (Ke ) can be where an and bn are the Mie coefficients given as follows:
obtained based on the Maxwell-Garnett theory using the imaginary part
of the refractive index of the nanofluid (κnf ), effective absorptive index, mψ n (mα)ψʹn (α) − ψ n (α)ψʹn (mα)
an = (19)
as follows: mψ n (mα)ξʹn (α) − ξn (α)ψʹn (mα)
4πκnf
Knf
e =
(16) ψ n (mα)ψʹn (α) − mψ n (α)ψʹn (mα)
λ bn = (20)
ψ n (mα)ξʹn (α) − mξn (α)ψʹn (mα)
It can be observed that Knf e can be calculated, using the Maxwell-
Garnett approach, based on the effective absorptive index of the nano­ where ψ n (z) and ξn (z) are the spherical Riccati-Bessel functions [71],
fluid (κnf ). In contrast, Knf
e is obtained in the Rayleigh approach as the and m is the relative refractive index defined earlier. Bohren and Huff­
sum of Kbf e and Knp
e (see Eq. (9)) because no effective optical constants, man [73] provided computer programs for calculations of particle op­
such as κnf , can be obtained for the nanofluid using the Rayleigh tical characteristics based on the Mie approach.
approach.
2.1.4. Discrete dipole approximation (DDA)
2.1.3. Mie scattering theory Purcell and Pennypacker [89] established the DDA to calculate the
The Mie theory is dedicated to describing how light waves interact absorption and scattering characteristics of arbitrary-geometry particles

8
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

by replacing them with an arrangement of N point dipoles. The dipoles metallic nanofluids, where smaller particles can exhibit enhanced light
should be arranged such that spacing between them is short relative to absorption compared to their larger counterparts. Therefore, for
the incident light wavelength. The interaction between the incident light metallic nanofluids and situations where size plays a significant role, the
waves and the electric fields generated by the arrangement of dipoles Drude model provides a more complete picture compared to the limi­
results in oscillating polarization in each dipole. These polarizations can tations of the Rayleigh scattering approach.
be resolved by solving a group of coupled linear equations derived from Otanicar et al. [34] adapted the Drude model to account for the
the equations that govern the electric field in the dielectric scatterer. size-dependent effects on the optical properties of nanoparticles. This
Draine [90] summarized the DDA theory and introduced the DDSCAT adjustment was incorporated into the numerical methodology by
software with the help of Flatau [91]. Draine [90] compared the DDA introducing modifications to the extinction and absorption coefficients
results with N of up to 2320 with the exact Mie scattering approach within the radiative transport equation (RTE). When the incident radi­
calculations for two cases: m = 1.7 + 0.1i and m = 3 + 4i. The agree­ ation triggers surfaces of electrons of metallic nanoparticles whose di­
ment between results was excellent. ameters are less than the mean free path of the conduction electrons in
Optical characterization of non-spherical nanoparticles like gold their bulk metallic material (for instances, 53.3, 39.9, and 18.9 nm for
nanorods (AuNRs) is frequently performed using the DDA [92,93]. Ag, Cu, and Al, respectively [94]), the electrons oscillation causes the
Brioude et al. [92] employed the DDA to present a simulation of the surface plasmon to be excited due to the resonant transfer of energy from
absorption efficiency of gold nanoparticles considering their real shapes. the incident photon to the plasmon yielding a significant increase in the
Authors concluded that the DDA is a very useful prediction tool for radiative energy absorption [41]. The optical characteristics of such
nanorods’ radiative characteristics since the presented model-based re­ nanoparticles are then a function of the nanoparticle size. Also, the
sults had high agreement with experimental results and theoretical ones nanofluid optical characteristics are altered due to the oscillation
obtained using classical electrostatic predictions and ellipsoidal particle interaction at the metal-dielectric-medium interface [69]. Such plas­
assumption. Additionally, authors found that the optical characteristics monic nanofluids can be optically characterized using the Drude theory
of nanorods were highly dependent on the nanocrystal aspect ratio, and that has been numerically solved and experimentally validated [95].
that the resonance plasmon bands arising in the absorption spectra could The mathematical formulation of the Drude model is demonstrated in
be clearly explained using the DDA. Similarly, Jeon et al. [93] Otanicar et al. [34] research.
concluded, based on the DDA, that the aqueous nanofluid containing This model presents a nanofluid’s effective complex dielectric func­
AuNRs exhibits an elevated extinction efficiency with an increase in the tion depending on nanoparticle size [42,95], which, in turn, can be used
aspect ratio of the AuNRs. Authors prepared a nanofluid with AuNRs of in the Maxwell-Garnett equations to obtain size-dependent refractive
three different aspect ratios to create a nanofluid with tunable absorp­ and absorptive indices of the nanofluid.
tion ability. The predicted spectral extinction coefficient of the blended Optical characteristics of nanofluids containing nanoparticles
nanofluid was reported in the visible and near IR spectral band and composed of dielectric cores and metallic shells, such as the Si-core and
found that it was in good agreement with experimental data as shown in Cu-, Au-, Al-, or Ag-shell nanoparticles, were investigated by Lv et al.
Fig. 7. [96] based on the Rayleigh approach assuming a quasi-steady state.
Moreover, the effect of dependency on nanoparticle size in optical
2.1.5. Comments on basic optical theories properties calculations was included based on the Drude model because
While the Rayleigh scattering approach offers a simple and effective the metallic nanoparticle’s dielectric function depends on its size, and it
method for predicting the optical properties of nanofluids containing can be clearly distinguished from the dielectric function of the bulk
small particles, it falls short when dealing with size-dependent effects, fluid. Authors concluded that such core-shell nanoparticles are good
particularly for metallic nanofluids. Here is where the Drude (free choices to create appropriate nanofluids for functional solar energy
electron) model comes in. The Drude model incorporates the size- harvesting systems, especially, knowing that such nanofluids have
dependent properties of metallic nanoparticles, specifically accounting tunable capability of radiation energy absorption.
for the collective motion of free electrons within them. This allows for a Back to the Rayleigh approach once again, it is frequently used in the
more accurate prediction of how the optical response, including ab­ literature to theoretically obtain the optical characteristics of a rela­
sorption and scattering, changes as the nanoparticle size decreases. This tively wide range of nanofluids of low concentrations [44,87,97–100].
size dependence is crucial for understanding the behavior of many However, it results in more accurate predictions for the general optical
performance of nanofluids composed of carbon-based or core-shell
particles than those composed of metallic or metal oxide particles,
particularly in certain wavelength bands. More specifically, the Rayleigh
nf
approximation results in accurate Ke predictions for graphite-based
nanofluids, overestimations for metal-based nanofluids within the
relatively short wavelengths (< 700 nm), and underestimations for the
metal-based nanofluids within longer wavelengths, as discussed by
Taylor et al. [87]. Those authors experimentally and numerically char­
acterized nanofluids composed of strong carbon-based nanoparticles as
graphite, and metallic nanoparticles as Al, Cu, and Ag. Authors justified
the deviation of the Rayleigh model results, in the case of metallic
nanoparticles, by the agglomeration and the oxidation of their surfaces.
For a more thorough investigation on carbon-based nanofluids, Gan
et al. [101] optically characterized two Ethanol-based nanofluids, one is
seeded by MWCNT and the other by carbon nanoparticles (CNPs), based
on the Rayleigh approximation while satisfying the Rayleigh assumption
condition (i.e., α≪1) in the studied radiation wavelength range. Authors
found that the Rayleigh model was capable of predicting the extinction
coefficients of both nanofluids accurately in the visible light wavelength
band ( 380 − 780 nm), but it failed in the UV range.
Fig. 7. Experiment- and DDA-model-based spectral Ke of a blended nanofluid at As endorsement for the assertions provided by Taylor et al. [87],
fV of 1 ppm [93]. (Reused with permission from OPTICA Publishing group).

9
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

Zhang et al. [99] optically characterized ionic liquid-based nanofluids Back to Otanicar et al. [97], the results showed that the temperature
seeded by copper, nickel, and carbon-coated nickel nanoparticles using increase led to changes in the host fluid’s refractive index which caused
the Rayleigh theory. Results showed almost the same assertions pro­ deterioration in the effective absorption coefficient of the nanofluid
vided by Taylor et al. [87] about the aforementioned wavelength ranges a ) and thus an increase in the light transmission. Also, authors stated
(Knf
nf
over which Rayleigh theory overestimates or underestimates Ke of that the nanoparticle size, at higher temperatures, had inconsiderable
metal-based nanofluids. Moreover, for the radiation of wavelength less impacts on the nanofluid absorption ability confirming the aforemen­
than 1200 nm, as illustrated in Fig. 8, the theory was not accurate tioned conclusions in this regard.
nf
enough in Ke predictions, especially for copper and nickel nanoparticles As for metal-oxide nanoparticle characterization, Said et al. [104]
because such metals are subject to surface oxidation. Metallic nano­ experimentally, using Beer Lambert method, and theoretically, using the
particle surface oxidation causes oxidized particles to exhibit distinct Rayleigh theory, investigated water/Al2O3 and water/TiO2 nanofluids
optical properties compared to their unoxidized counterparts. Also, within the spectral range from 200 to 1100 nm. For Al2O3, comparable
nf
nanoparticle agglomeration deforms their spherical shape and enlarges trends in Ke curves were identified as predicted through analytical and
their size over the initial (rated) one used in calculations, which leads to experimental analyses. Nevertheless, a substantial disparity between the
inconsistencies between the experimental results and Rayleigh theory experimental and analytical values of the extinction coefficient was
predictions. observed. Experimental data exceeded analytical estimations by over
Otanicar et al. [97] studied the Au-based nanofluid’s optical char­ tenfold in the UV region, whereas in the visible and IR regions, the
acteristics using the Rayleigh approach with inclusion of the tempera­ difference between the results decreased from tenfold to 1 time.
ture increase impacts on the refractive index of the host fluid and the Also, Sajid et al. [98] stated that the Rayleigh theory fails to predict
effect of nanoparticle size, utilizing an approximation proposed by the water/Al2O3 nanofluid optical characteristics in the UV radiation
Bashkatov and Genina [102] and a model proposed by Turcotte and band due to the increasing scattering effects caused by nanoparticle
Schubert [103], respectively. Bashkatov and Genina [102] presented a agglomeration, which indicates that the computed extinction was
simple approximation of the refractive index of water based on the mainly due to scattering rather than absorption. It should be noted, in
temperature and wavelength in the range (200–1000 nm). The de­ this regard, that the scattering efficiency of a nanoparticle exhibits 1/λ4
pendency of refractive index on temperature and wavelength was per­ dependence on the incident radiation wavelength, as illustrated in Eq.
formed by curve fitting of data published in the literature. The data (6). And since α≪1, the scattering efficiency is, in theory, orders of
showed that the water refractive index decreases with increasing both magnitude less than the absorption efficiency, as clarified by Eqs. (5)
the temperature and wavelength. Also, since the increase in the nano­ and (6). This justifies neglecting the scattering efficiency of nano­
fluid temperature increases the nanoparticle size that, in turn, changes α nf
particles in Ke theoretical calculations.
and fV , the optical properties obtained by the Rayleigh theory will be As for the Maxwell-Garnett approach, it was observed that it presents
influenced. Therefore, Turcotte and Schubert [103] modeled the final inaccurate predictions of nanofluids’ optical characteristics [69]. Taylor
diameter of a particle at a certain temperature, D’ (T), as: et al. [87] stated that the Maxwell-Garnett theory underestimates the
nf
( )1 absorptive indices of nanofluids leading to underestimated Ke , as shown
Dʹ(T) = αT (T − 20)D3o + D3o 3 (21)
in Fig. 9.
nf
Moreover, a comparison between the Ke of MWCNT nanofluid ob­
Where Do is the nominal particle diameter at room temperature and αT is
tained based on the Maxwell-Garnett theory and that obtained based on
the volumetric thermal expansion coefficient of the particle material in
the Rayleigh approach was made by Lee et al. [105]. Authors concluded
1/◦ C. Also, the new volume fraction (fʹ ) can be obtained using the new
V that the former method that considered the nanoparticle shape effect as
diameter and the nominal volume fraction at room temperature (fV o ) as a cylinder failed to yield an accurate Ke prediction, but the latter one
follows: could result in good predictions by adjusting the nanoparticles’ hydro­
3 dynamic diameters.
Dʹ × fV o
fVʹ = (22) The Rayleigh theory is preferable, due to its simplicity, as long as
D3o
α≪1. However, when α ≈ 1, the Mie theory is a better choice [106].
nf
Several researchers stated that Mie theory yields more accurate Ke

Fig. 8. Ke of ionic nanofluids based on Rayleigh approach for different metallic

nanoparticles of 40 nm diameter and 10 ppm nanoparticle loading [99]. Fig. 9. Extinction coefficient of different water-based nanofluids using the
(Reused with permission from Elsevier). Maxwell-Garnett theory [87]. (Reused with permission from Springer).

10
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

predictions especially when the model considers the nanoparticle size energy harvesting applications performance. In contrast, it was reported
and dynamic light scattering (DLS) [86,107], which is a non-destructive in literature that the nanofluid with a lower stability factor exhibits
material characterization technique used to determine the size and size lower solar absorption and storage abilities [114] (this will be detailed
distribution of particles in suspensions. For example, Song et al. [107] in section 4). However, agglomeration deteriorates the stability and
used both Rayleigh and Mie scattering theories to describe the optical thermal conductivity (kc ) of the nanofluid. This phenomenon can
properties of Water/Al2O3 and Ethyl-Alcohol/Al2O3 nanofluids. Authors adversely affect the efficiency of solar collectors, as the uniform
found that experimental and Mie theory-based absorption coefficients dispersion of nanoparticles is crucial for enhancing the thermal prop­
were reasonably agreed when considering nanoparticles aggregation. erties of the suspension. Agglomeration of nanoparticles deteriorates the
However, Rayleigh theory overestimated optical properties with several performance of the system by reducing the effective surface area for heat
orders of magnitude. Strong agreement between Mie theory predictions transfer and increasing the viscosity of the nanofluid, leading to higher
and experimental results is observed when the particles within the pumping power requirements. Moreover, when particle agglomeration
nanofluid are spherical and act as independent scatterers [108–110]. occurs, the effects of dependent scattering between agglomerated par­
Kameya and Hanmaura [108], assuming independent scattering, ticles cannot be overlooked. The independent scattering model is no
examined the absorption coefficient of Ni/alkyl naphthalene using the longer acceptable and new optical theories are necessary. Therefore,
Mie theory and observed high agreement between experimental and researchers stated that Rayleigh approach is no longer valid for evalu­
theoretical results in the visible and near IR spectral region (see Fig. 10). ation of nanofluids’ optical properties when agglomeration occurs due
However, considerable deviations between theory and experiment re­ to the increased particle size [107,115]. The small nanoparticle size, in
sults can occur due to nanoparticle agglomeration, where particles DASCs, translates to negligible scattering compared to the significant
clump together resulting in larger particle clusters [110–112]. In such absorption efficiency, as evidenced by a wavelength-averaged absorp­
cases, the Mie theory might overestimate the spectral transmittance of tion to scattering efficiencies ratio of around 104 in solar nanofluids
the nanofluid. Zhu et al. [111] reported that the Mie theory over­ containing spherical nanoparticles [84,116]. This absorption dominance
estimated the (AlN/water) nanofluid’s spectral transmittance due to holds true even for non-spherical particles like carbon nanohorns
agglomeration that caused the actual size of particles to become eight (CNHs), where scattering contributes a mere 5 % to the total reduction
times greater than the manufacturer-rated size. Furthermore, the accu­ in light intensity [117]. However, size plays a crucial role. As nano­
racy of the Mie theory can be impacted by the size of nanoparticles particles increase in size or cluster together, the balance between scat­
[113], where predictions become less accurate for particles exceeding a tering and absorption shifts dramatically. For instance, CNHs reaching
certain size at specific spectral ranges as detailed in Ref. [109]. around 150 nm in size can scatter a significant portion (30 %) of the light
One of the causes that explain why such theoretical nanofluid [118], and for gold spheres exceeding 88 nm, scattering can even sur­
characterization techniques cannot present accurate predictions of pass absorption, becoming the primary interaction with light [119].
nanofluids’ optical characteristics is that researchers take the nano­ Bohren and Huffman [73] stated a formula to predict the intensity of
particle size in calculations as the average size given by the manufac­ scattered light (Is ) when a nanofluid is exposed to an incident light with
turer ignoring the changes coming over it due to clustering and intensity of Io :
agglomeration. Particle agglomeration is usually caused by the two-step ⃒ ⃒2
nanofluid preparation technique, especially at fV of greater than 0.01% Is π4 Ns D6p ⃒⃒m2 − 1⃒⃒ ( )
≈ 1 + cos2 θ (23)
and when nanoparticles do not experience a proper functionalization Io 8λ r m + 2⃒
4 2 ⃒ 2

process [69]. This phenomenon presents a double-edged sword for

nanofluid-based solar systems. While it might potentially cause en­ where Ns , θ, and r are the number of scattered particles, scattering angle,
hancements in the solar extinction ability of nanofluids, it adversely and distance at which scattered light is observed, respectively. The
influences their overall OTPs. Du and Tang [112] examined the greater the nanoparticle size the more severe the agglomeration phe­
extinction coefficient of Ag/water nanofluid and observed that nomenon is, and thus the amount of scattering increases as seen in the
agglomeration increases the long-wave scattering and decreases ab­ correlation above. Consequently, the numerical models dedicated for
sorption peaks in short wavelength bands. Authors concluded that nanofluid optical characterization that assume independent scattering
agglomeration enhances the nanofluid’s extinction coefficient at cannot accurately predict the nanofluid radiative behavior anymore.
large-wavelength (> 1 μm) for fV of 0.5 − 2%, which improves the solar DLS nanoparticle size measurements were considered in radiation
scattering calculations in some investigations in the literature to yield
more accurate optical behavior predictions [105,107,109]. Further­
more, the nanomaterial impurity and oxidization are considered as main
reasons for the discrepancy between the experimentally-measured and
numerically-predicted nanofluids’ optical characteristics. These phe­
nomena are caused by the reduction of metal particle size, specifically,
to the nanoscale, which causes a sharp increase in their surface energy.
However, the dielectric constants and refractive indices of nanoparticles
are assumed to be equal to their bulk materials in optical calculations
[69].
Selecting the most appropriate optical characterization theory de­
pends on several considerations. Mie theory, while offering a compre­
hensive solution for spherical scattering that accommodates both
absorbing and non-absorbing particles without specific size constraints,
is constrained by the necessity for particles to be similar and isotropic
[120]. Despite its efficacy in elucidating various spherical scattering
particle systems, including Rayleigh scattering, Mie theory’s intricacies
often lead to a preference for Rayleigh scattering when applicable,
owing to its simpler nature. For simplicity, the Rayleigh theory is
favored, but its applicability should be ensured, namely, the suspension
should be with low volume fraction and small particle size. The DDA
Fig. 10. Experimental and Mie theory-based Ka of Ni-based nanofluids [108].
(Reused with permission from Elsevier).
stands out as a versatile method with broad applicability, limited only by

11
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

available computational power. Also, as seen earlier, Mie theory en­ the past models were simplified and the nanofluid-based DASC was
counters limitations when the nanofluid experiences agglomeration, particularly modeled by Tyagi et al. [57]. And then, Otanicar et al. [68]
necessitating alternative approaches like the Diffusion Limited Cluster built on their model to optically characterize nanofluids in DASC sys­
Aggregation (DLCA), Generalized Multi-particle Mie (GMM) and Finite tems taking into consideration the multiple radiation scattering and the
Difference Time Domain (FDTD) methods. In these methods, effects of nanoparticle size. Although the aforementioned numerical modeling
dependent scattering between particles, particle agglomeration, and attempts provided good understanding of DASCs, its optimal design
particle size and loading on the nanofluid’s extinction coefficient are features, until that moment, were unclear. For instance, the past math­
accounted for [112]. ematical modeling contributions [46,47,57,68] had just taken into
However, the sensitivity of nanoparticles’ optical characteristics to consideration specific incorporations of certain nanoparticles and host
nanoparticle agglomeration is well-established. Precise predictions of fluids in a particular geometry. On the other hand, even simple volu­
these optical properties highly depend on cluster’s size and shape. metric absorption collectors have many different parameters to vary,
Therefore, it is crucial to correlate the structure of clusters with their causing difficulties in the collector behavior prediction using the
optical characteristics using methods that are appropriate for coupled radiative and energy equations.
engineering-scale systems. This task grows increasingly intricate when
dealing with nonspherical particles, like carbon nanotubes (CNTs) or 2.2.1. Radiative transport equation (RTE)
coated nanoparticles. It is equally complex when considering the pres­ The RTE mathematically describes the interaction and propagation
ence of stabilizing agents like polymers and surfactants, as well as the of radiation within a certain medium, that could be scattering,
impact of fluid shear on deformation, fragmentation, and restructuring absorbing, and/or emitting. It is given, in the most general form, as
of clusters. In existing literature, there is a notable gap in both experi­ follows [125]:
mental and theoretical research addressing how these elements influ­ ∫
dIλ Ks 4π
ence the clustering kinetics and cluster structure, which in turn affects = − Ka Iλ (S) − Ks Iλ (S) + Ka Iλ,b (S, Tm ) + Iλ (S, Ωi )Φ(λ, Ω, Ωi ) dΩi
the optical characteristics of nanofluids [77]. However, given the dS 4π Ωi =0
infrequency of aggregation in DASC systems due to the very low volume (24)
fraction of solar nanofluids, this study was confined to the basic optical
Iλ (S), Ka , Ks , Iλ,b , Tm , Ω, and Φ are the intensity of spectral radiation in
theories.
the direction S, absorption coefficient, scattering coefficient, blackbody
radiation intensity, medium temperature, solid angle, and scattering
2.2. Governing equations of DASC systems
phase function. The terms on the right-hand side of the RTE account for
radiative energy lost from or gained by the incoming light beam: the
Developing an efficient nanofluid-based DASC requires a rigorous
radiative absorption loss, out-scattering loss, emissive gain, and in-
understanding of its radiative and thermal behaviors. To that end, many
scattering gain in the S direction, respectively.
questions about the interaction of the sunlight and nanofluid in the
The RTE can be significantly reduced if the scattering terms are
collector, attainable temperatures, and collector ηPC should be
dropped according to the negligible scattering effect approximation
answered. Therefore, a mathematical model that can thoroughly
when fV is very low and particles are ultra-small, as discussed earlier.
describe these aspects is essential. Such a model would not only enable
Moreover, the utilization of nanofluids with low fV is more useful for the
accurate predictions of DASC performance but also facilitate the opti­
collector because out-scattering, that increases with fV , does not cause
mization of various design and operating parameters.
any increase in the thermal energy of the system. The simplified y-di­
Unlike conventional SASCs in which the temperature difference be­
rection RTE according to this approximation is frequently used in the
tween the absorber and working fluid could reach 100 ◦ C or more,
literature with the form:
particles seeded in the working medium in DASCs are nearly with the
same temperature as the host fluid’s due to the huge particle surface- dIλ (y)
= − Ke Iλ (y) (25)
area-to-volume ratio [28,35,47,63]. I.e., it can be assumed that there dy
is no temperature slip between particles and the base fluid. Such an
approximation entails more efficient and instantaneous overall heat In addition to the conventional numerical solutions of differential
transfer between the nanoparticles and host fluid [63], and thus negli­ equations using discretization techniques such as the finite volume
gible temperature difference between them, which justifies the method (FVM) [63,126–129], finite element method (FEM) [41,130],
single-phase isotropic nanofluid assumption. Such heat transfer mech­ and finite difference method (FDM) [34,46,57,131], which can be used
anisms can be resolved using a coupled radiative and convective heat to solve the RTE, the RTE is frequently solved in the literature based on
transfer model to predict the collector thermal response. Researchers stochastic and deterministic solution techniques including the Monte
have initially presented numerical models to investigate volumetric Carlo (MC), PN , and differential discrete ordinate (DDO) methods.
absorption collectors of opaque bottom walls and non-scattering
semi-transparent working media [121,122]. Then, Kumar and Tien 2.2.1.1. Monte Carlo (MC) method. The RTE is converted, in the MC
[123] proposed a simple analytical framework combining the spatial method, into a sequence of probabilistic problems through several
and spectral optical characteristics of particle-laden fluids besides the numbers generated randomly. As a nanofluid is exposed to light, light
boundary conditions. This work introduced a fundamental modeling photons experience partial scattering and absorption during propaga­
contribution in the volumetric absorption investigations. The work tion within the nanofluid. The probability density function (P(S))
presented an analytical solution to the coupled radiative transport and describing the absorbed or scattered photons depending on the distance
energy equations that govern the DASC system. Chronologically, Miller they crossed is [132]:
and Koenigsdroff, in two separate pieces of research [47,124], expanded
these previous contributions to involve carbon particles oxidation in a P(S) = Ke e− Ke l
(26)
3D mathematical model. They proposed a thermal model for a volu­
where l is the path length or distance crossed by the photon before it is
metric absorption central solar collector utilizing a working fluid
scattered or absorbed, i.e. the free path length. l can be calculated by the
composed of micron carbon particles blended in a gaseous medium. The
inverse transform sampling:
RTE was solved to state the spectral and regional distribution of solar
flux in the collector including the impacts of scattering and emission ln R
l=− (27)
from the working fluid and collector walls. Besides, the energy equation Ke
was solved to obtain the temperature profile of the flow. Later, in 2009,

12
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

where R is a number between 0 and 1 uniformly and randomly gener­ predicting the photothermal conversion performance of a DASC based
ated via a quasi-random sequence. As discussed, to determine the optical on the DDO method. The proposed model presents the previous con­
characteristics of a heterogeneous nanofluid, it should be divided into siderations. The authors addressed the RTE of a fluid using the DDO
several sublayers with the possibility of neglecting the sub-layers’ method in finite-volume form and introduced a self-programming model
interfacial effects due to the low nanoparticle loading in case of DASC utilizing Python script for the photothermal conversion of DASC. Com­
parisons between the outcomes obtained from the novel and conven­
systems. The distance a photon crosses determines the ith layer in which
tional model under conditions of both low and high temperatures were
the photon is positioned. The distance the photon crosses from the layer i
conducted to ascertain the accuracy and advantages of the novel
boundary is denoted as lb , and the photon is considered that it left the
approach. Radiative heat loss was assessed across diverse collector
layer i if l > lb . The photon’s remaining motion step remains updated as
temperatures and solar concentration ratios. Additionally, the impact of
long as the ith layer boundary is not the working fluid boundary. How­
the angle of incidence on conversion efficiency was analyzed. The pro­
ever, because of the difference between the working fluid and envi­
posed model extends the applicability of the DASC numerical model
ronment media and thus the difference between their refractive indices,
from low to high temperatures. The authors presented comparisons
the photons could be refracted or reflected as they pass through the
between their proposed DDO model and the common conventional
boundary interface. The proportion of the reflected photons (Rr ) can be
model when simulated a DASC with an operating temperature of 500 ◦ C
calculated using the Fresnel equation as:
(very close to the limit condition (i.e., 750 K) below which radiative
[( )2 ]
Rs + Rp 1 m1 cos Θi − m2 cos Θt
)2 (
m1 cos Θt − m2 cos Θi losses can be neglected). They showed that the conventional model
Rr = = + overestimated the outlet temperature and ηPC by 27.6 ◦ C and 18.9 %,
2 2 m1 cos Θi + m2 cos Θt m1 cos Θt + m2 cos Θi
respectively. Results of the proposed model present a reference for ηPC
(28) enhancements in DASCs at high temperatures.
where Rs , Rp , m1 , m2 , Θi and Θt are the proportion of s-polarized pho­
tons, proportion of p-polarized photons, refractive index of the envi­ 2.2.1.4. Limitations of existing RTE solution techniques. A literature sur­
ronment, refractive index of the nanofluid, angle of incidence, and angle vey showed that solutions to the RTE can be accepted with reduced
of transmission, respectively. One more random number (R1 ) is uni­ accuracy in many cases such as the very dilute suspension case at which
formly generated to check if the photon experiences an internal reflec­ the in-scattering impacts can be ignored [34,57,130,131]. Also, the gray
tion, which is true if Rr > R1 and its position and velocity are updated. participating medium assumption can be stated to investigate nanofluids
But if Rr ≤ R1 , the photon’s motion direction is determined. [28,139]. Also, the in-scattering augmentation effects are neglected
If l < lb , the photon is still inside the working medium. An additional because the scattering efficiency is very small compared to the absorp­
random number (R2 ) is generated to determine if the photon is absorbed tion efficiency [34,57]. The reduction of the general form of RTE to the
or scattered. If Ka /Ke > R2 , the photon is absorbed and its packet weight simplified RTE used in DASC studies is conducted based on such ap­
decreases, and thus absorption counter is updated. But if Ka / Ke ≤ R2 , proximations. Given that the thickness of the suspension layer is
the photon is scattered and its new motion direction is determined as significantly smaller than the length of the collector [46] and the inci­
described in Refs. [41,132] in which the MC method is detailed. Finally, dent radiation is assumed to be normal to the collector surface, varia­
the number of reflected, transmitted, and absorbed photons are divided tions in radiative heat flux throughout the suspension are considered to
by the total number of photons to determine the nanofluid’s reflectance occur along the negative vertical axis only. Also, the in-scattering term is
(Rλ ), transmittance (Tλ ), and absorptance (Aλ ), respectively, over the dropped in case of dilute suspensions used in DASCs [34,57,130,131],
entire solar spectrum. Thus, the nanofluid’s absorption efficiency (ηA ) and the gray participating medium assumption can be applied for
can be calculated as follows: colloidal suspensions [28]. Moreover, the emissive radiation loss term is
neglected when the operating temperature is below 750 K as it is
∫ λmax
Io,λ .Aλ dλ responsible for less than 5 % of the overall radiative losses [136,137].
ηA = λ∫minλmax (29) The general RTE is simplified accordingly.
I dλ
λmin o,λ
Radiative transfer numerical methods can be categorized into two
main groups: stochastic simulation methods and deterministic methods,
2.2.1.2. PN method. The basic principle behind the PN method is the typically formulated based on the integral or differential form of the
approximation of transfer relations as a determinate group of high-order RTE. Stochastic methods, such as the MC Method, belong to the first
moment equations for the RTEs to be converted from the integral form to group, offering versatility and reliability but often demanding in terms
the differential form. It is useful when dealing with anisotropic scat­ of computational time due to the tracing of a substantial number of
tering in complex geometries [133]. photons. They also face challenges when coupled with conduction and
convection solvers, which typically employ deterministic methods like
2.2.1.3. Differential discrete ordinate (DDO) method. This method’s goal the FDM, FVM, and FEM. On the other hand, the DDO method belongs to
is to introduce the generic transfer equations by discretizing an equa­ the second group. FVM can be viewed as a specific type of DDO method
tions bundle for the mean intensity over a few angular zones and where discrete ordinate equations are derived using FVM principles.
assuming that the radiation intensity is constant throughout these zones FEM, known for its superior accuracy compared to FVM, stands out for
[46,134]. its versatility and holds promise for simulating multiphysics systems,
Considerations regarding radiation and heat loss especially at encompassing radiative energy transport [140].
elevated temperatures are overlooked when using the conventional As approximate techniques, numerical methods encounter various
model (namely, the simplified y-direction RTE explained in section types of errors. The MC method grapples with statistical errors, while the
2.2.1) commonly used in the literature to resolve the radiative behavior DDO method, FVM, and FEM contend with errors stemming from space
of DASCs [57,131,136–138]. Most pieces of research in the DASC field and angular discretization. These errors can introduce unphysical
share a common simplification of radiative heat transfer. For instance, characteristics to the solution, rendering it challenging to interpret and,
solar irradiance was assumed to vary solely along the y-axis, and the in some cases, undermining the solution entirely. Therefore, under­
emission and scattering of radiation by the working fluid were neglec­ standing the origin and nature of numerical errors is crucial. Such un­
ted. Consequently, the impact of radiation was simplified to a heat derstanding aids in the interpretation of numerical simulation results
source dependent solely on the y coordinates. Radiative heat loss is and facilitates the development of strategies to mitigate or eliminate
addressed through an effective coefficient, along with convective heat errors. The DDO method, for instance, has long been recognized to suffer
loss. Very recently, Zhu et al. [135] developed a numerical model

13
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

from false scattering and ray effects, prompting the proposal of strate­ ation, different CFD techniques can be used. Macroscale, mesoscale, and
gies to alleviate these issues. Consequently, false scattering and ray ef­ microscale techniques, respectively, may be used as the length and time
fects emerge as two general types of numerical errors that require in- scales of the system range from large to small scales, as shown in Fig. 11.
depth comprehension. An in-depth grasp of limitations of numerical Macroscale techniques include the fully-continuous and continuous-
techniques used to solve the RTE and numerical errors associated with discrete approaches. Mesoscale and microscale techniques employ ap­
its solutions is presented in Ref. [141]. proaches such as the Lattice Boltzmann method (LBM), dissipative
particle dynamics (DPD) method, smoothed particle hydrodynamics
2.2.2. Colloidal suspension governing equations: applicability and (SPH) method, and molecular dynamics (MD) simulation.
limitations The fully continuous approach is the simplest and most widely used
The energy equation form differs according to the geometry and flow in DASC investigations. It is based on handling the suspension as a
conditions of the investigated system. In case of a 2D (x and y di­ single-phase continuous medium, where nanoparticles are homoge­
rections), steady flow and y-directional radiative transport, the energy neously distributed, with averaged (effective) thermophysical proper­
equation can be given as follows [46]: ties. Or it assumes that the dispersed solid phase and fluid phase are two
( ) ( 2 ) different interpenetrating continua. Then, it is called the Eulerian-
∂T(x, y) ∂T(x, y) ∂ T(x, y) ∂2 T(x, y) ∂qr (x, y) Eulerian approach. On the other hand, the hybrid continuous-discrete
ρcp u +v = kc + −
∂x ∂y ∂x2 ∂y2 ∂y approach is based on treating the base fluid as a continuous medium
(30) and the dispersed particles as discrete points moving in the flow field.
The Lagrangian model offers detecting the discrete particle trajectories
where ρ, cp , kc , u, v, T, and qr are the density, specific heat, thermal
while moving in the flow domain. In the fully Lagrangian approach, both
conductivity, x-velocity, y-velocity, medium temperature, and total
the base fluid and nanoparticles are treated as individual discrete par­
radiative heat flux, respectively. The negative divergence of the total ticles moving through space and time.
radiative heat flux term, ∂qr (x, y)/∂y, is used to couple the energy
The continuous base fluid is always investigated based on the
equation with the RTE and thus the temperature distribution within the Eulerian approach when dealing with the macroscale level. However,
working fluid can be obtained. The total radiative heat flux, qr (x, y), is
the appropriate approach for handling the dispersed solid particle phase
defined based on the spectral solar radiation flux (Iλ ), as follows [85]: is determined based on the particle size (Dp ) or response time (τ),
∫ λmax compared to the characteristic time (τc ) and length (L) scales of the base
qr (x, y) = [Iλ (x, y, γr > 0) − Iλ (x, y, γr < 0)] dλ (31) fluid, and the particle loading. The particle response time (τ) is the time
λmin
taken for a particle to reach a new equilibrium state after a stimulus is
γ r is the directional cosine of radiation propagation with respect to the applied. It is given for spherical particles and Stokes regime as [142]:
vertical direction. γ r < 0 indicates radiation from bottom to top.
ρp D2p
Conversely, γ r > 0 indicates radiation from top to bottom (i.e., inci­ τ= (33)
18μf
dence). The reduced energy equation governing a planar DASC is
frequently given in the literature, assuming a plug flow, as follows: The characteristic time scale of the base fluid (τc ) is defined as the
2 time it takes the base liquid to pass through a channel of length L:
∂T ∂T
ρUcp = kc 2 + q̇ʹʹʹ(y) (32)
∂x ∂y L
τc = (34)
u
where U is the mean velocity, and q̇ʹʹ(y) is the volumetric heat generation
where u is the base fluid’s superficial velocity calculated from the fluid
within the working fluid given as − ∂qr (y)/∂y. Modeling the flows in
flow’s free cross-sectional area. Now, the Stokes number (St) can be
systems employing solid-liquid suspensions, including DASCs and
calculated as the ratio between τ and τc :
parallel-plate micro-channels, can be performed using the fully contin­
uous (Eulerian), continuous-discrete (Eulerian-Lagrangian), and fully St =
τ
(35)
discrete (fully-Lagrangian) approaches. Due to the high nonlinearity of τc
equations governing the flow and heat transfer of colloidal suspensions, Being St≪1 means that dispersed particles have enough time to
CFD techniques are utilized to solve them rather than analytical ones. respond to variations in the base fluid velocity. Consequently, the par­
Depending on the length and time scales of the system under consider­ ticle velocity approaches the base fluid’s one (i.e., velocity equilibrium).
The dispersed particles, whereupon, are considered as fluid tracers in the
flow field [142]. However, if St > 1, dispersed particles do not have
enough time to respond to changes in the base fluid velocity, and thus,
the particle velocity exhibits differences from the base fluid one. For
such large St’s, dispersed particles separate from the base fluid
streamlines and move in their own directions [142].
Another quantity is to evaluate how long it takes for a particle to
respond to the change in the continuous phase temperature, called the
thermal response time (τT ). Based on this quantity, the thermal Stokes
number (StT ) is calculated. When StT ≪1, dispersed particles shortly
approach the continuous phase temperature. But when StT ≫1, a
considerable relative temperature between particles and the continuous
phase will be observed. τT is given as follows [142]:

Cp ρp D2p
τT = (36)
12kf

St, Dp , and fV of the case under investigation have the main role in
Fig. 11. The position of simulation approaches with regard to the length and determining the appropriate approach by which a certain colloidal
time scales. suspension flow should be modeled. In this section, the mathematical

14
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

models offered in the literature will be discussed and then a roadmap ( ) ( )

∂ ρeff Ceff Tm ∂ ρeff Ceff um,j Tm ∂2 Tm
will be proposed to guide researchers to the most appropriate mathe­ + = keff + Sva (43)
∂t ∂xj ∂xj ∂xj
matical approach that should be used under certain conditions to result
in accurate predictions.
where Sva represents the source term in the energy equation. In case of
DASC systems, Sva represents the volumetric solar radiation absorption
2.2.2.1. Eulerian single-phase homogeneous model. Although a colloidal within the working fluid, which is the negative divergence of the total
suspension is inherently composed of two different phases, it can be radiative heat flux as explained earlier. In solar nanofluids, the particles
treated as a single-phase homogeneous mixture, i.e. nanoparticles are exceedingly small in both size and spatial occupation. Consequently,
exhibit a homogeneous and uniform distribution in the base fluid, under the velocity and temperature of the dispersed particles closely align with
certain conditions. If the suspension is dilute (i.e., fV < 6.5% for solar the corresponding fluid values. This equivalence, however, does not
nanofluids [143]) with small-size particles and low St (i.e., < 0.1), hold when dispersed particles are larger. But since solar nanofluids are
dispersed particles act as fluid tracers and will be in momentum and highly dilute and exhibit a macroscopic appearance akin to homoge­
thermal equilibria with no velocity or temperature slips. Also, particle neous fluids, and considering that their thermophysical attributes have
interactions and collisions will not have a considerable effect on the flow been calculated using this model, the homogeneous model stands as a
hydrodynamics and heat transfer of the dispersion. Then, the Eulerian preferable choice for solar nanofluid applications. The accuracy of this
single-phase homogeneous model is applicable. This model is the set of governing equations is deemed acceptable as it encompasses
simplest and most widely used in the literature in DASC investigations. It conservation laws, with the exception being non-Newtonian nanofluids.
treats the suspension as a homogeneous mixture with unique effective The boundary conditions applied to Eqs. 41–43 are same as those for the
thermophysical attributes depending on those of the suspension con­ fluid (i.e., no penetration, no slip, and equal temperature). Given the
stituents and fV . However, it was observed that several DASC studies highly dilute nature of solar nanofluids and the infrequent particle-wall
simply use the base fluid attributes rather than effective ones arguing collisions, these boundary conditions offer accurate representations of
that solar nanofluids employ negligibly small fV values [28,136, nanofluids. It is reasonable to assert that comprehensive information for
144–146]. But for higher-accuracy predictions using the fully contin­ employing computational models in nanofluids is available and highly
uous single-phase model, the effective nanofluid thermophysical prop­ accurate. Consequently, the justified use of the homogeneous model in
erties are given in the literature based on several correlations. The solar nanofluid applications is anticipated to yield sufficiently accurate
effective density of the nanofluid (ρeff ) is given based on the theory of results.
mixing as [147]: On the other hand, the single-phase homogeneous model relies on
( )
ρeff = 1 − fV ρf + fV ρp (37) the average attributes of the mixture, offering insights solely into these
averages. It cannot capture any differences between the dispersed par­
Theoretical ρeff values exhibited good agreement with experimental ticles and base fluid. Moreover, due to the uniform particle distribution
ones [148]. The effective heat capacity of the nanofluid (Ceff ) is obtained assumption, it cannot address inquiries such as the relative velocity of
based on the theory of mixing assuming thermal equilibrium between particles, particle distribution and migration impacts, or the impact of
the particles and base fluid by Xuan and Roetzel [149]: particle aggregation on the performance of the nanofluid. Bahiraei
( ) [152] inferred that contradictions, observed in the literature, in the
1 − fV ρf Cf + fV ρp Cp nanofluid performance predictions are attributed to overlooking the
Ceff = (38)
ρf particle migration impacts and ignoring the velocity and temperature
slips between dispersed particles and the host fluid. This is due to the
Also, several techniques were followed in the literature to calculate
implicit assumption in the single-phase homogeneous model that there
the nanofluid effective viscosity (μeff ) as a function of fV based on
is no momentum or energy exchange between the solid and fluid phases
empirical correlations or numerically based on Taylor series expansion
and there are momentum and thermal equilibria between them. These
[150]. The Brinkman’s formula is given as follows [150]:
limitations drove researchers to rather use different approaches to pro­
μf vide more details on the nanofluid performance and deep justifications
μeff = ( )2.5 (39) for its behavior.
1 − fV

Maxwell presented a correlation for a solid-fluid suspension and then 2.2.2.2. Eulerian-Eulerian two-fluid model. The Eulerian-Eulerian two-
it was justified by Buongiorno et al. [151] to be used for the nanofluid’s fluid model is the most common two-phase modeling approach in
effective thermal conductivity (keff ) calculations: nanofluid investigations [153]. It treats both the dispersed particles and
( / )
kp + 2kf + 2 kp kf fV base fluid as two different interpenetrating continua coexisting in the
keff = ( ) (40) same volume. Each continuum is separately governed by its mass, mo­
kp + 2kf − kp − kf fV
mentum, and energy conservation equations. Interactions between the
Correlations for μeff and keff are given in the literature with more two continua are represented by additional terms (i.e., source terms),
details accounting for the temperature and dispersed particle size and incorporated into the respective governing equations. For instance, a
shape effects. However, we selected the relations above for DASC in­ force term accounting for the drag imposed on the base fluid by particles
vestigations as they are generic (i.e., not limited to specific nanoparticle is included in the momentum equation. Conservation equations for the
types) and simple (as fV is very tiny in solar nanofluids). base fluid encompass terms representing hydrodynamic forces at the
Accordingly, the continuity, momentum, and energy equations to solid–fluid phases interface; and energy transfer from the solid to fluid
solve for the space-averaged velocity (um ) and temperature (Tm ) of the phases. Empirical closure equations are then used to individually model
homogeneous single-phase solid-liquid Newtonian mixture are given these terms. Applied to Newtonian fluids, the comprehensive equation
based on the single-phase model, respectively, as follows: set for the Eulerian-Eulerian model is outlined as follows.
( )
∂ρeff ∂ ρeff um,j - Conservation of mass for fluid and solid phases, respectively:
+ =0 (41)
∂t ∂xj [( ) ] [( ) ]
∂ 1 − fV ρf ∂ 1 − fV ρf uj
( ) ( ) + =0 (44)
∂ ρeff um,i ∂ ρeff um,j um,i ∂P ∂2 um,i ∂t ∂xj
+ = ρeff gi − + μeff (42)
∂t ∂xj ∂xi ∂xj ∂xj

15
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

( ) [ ]
∂ f V ρp ∂ fV ρp vj that this approach fails to provide details on particle dispersion, ag­
+ =0 (45) gregation, trajectories, and spatial distribution, effects of particle shape
∂t ∂xj
and size (where particles of different sizes are deemed as different
phases [142]), and particle-particle and particle–wall interactions. The
- Conservation of momentum for fluid and solid phases, respectively: Eulerian- Eulerian approach rather offers a statistical averaged
[( )
∂ 1 − fV ρf ui
] [( )
∂ 1 − fV ρf uj ui
] [ ] description for the velocity and concentration of particulate phase as a
( ) ∂P ∂2 ui
+ = 1 − fV ρf gi − + μf + Fi function of space and time. Hence, there exists a critical necessity for a
∂t ∂xj ∂xi ∂xj ∂xj
multi-scale modeling approach that can effectively capture the discrete
(46) nature of dispersed particles to shed light on the aforementioned phys­
[ ] [ ] [ ] ical phenomena while maintaining computational feasibility.
∂ fV ρp vp,i ∂ fV ρp vp,j vp,i ∂(P + Pc ) ∂2 vp,i
+ = f V ρp gi − + μp − Fi (47)
∂t ∂xj ∂xi ∂xj ∂xj
2.2.2.3. Eulerian-Lagrangian model. The continuous base fluid in this
approach is treated from the Eulerian modeling perspective. However,
- Conservation of energy for fluid and solid phases, respectively: dispersed particles are studied from the Lagrangian perspective. The
[( ) ] [( ) ] Lagrangian approach is intuitively and physically justified when the
∂ 1 − fV ρf Cf Tf ∂ 1 − fV ρf Cf uj Tf ( ) ∂2 Tf suspension is dilute (fV < 6.5% [143]) and dispersed particles are small
+ = 1 − fV kf + q + Sva
∂t ∂xj ∂xj ∂xj and discrete. In contrast to conventional fully-continuous models, fully
(48) discrete or hybrid continuous-discrete techniques offer the capability to
[ ] [ ] account for the discrete nature of nanoparticles. Such an approach en­
∂ fV ρp Cp Tp ∂ fV ρp Cp vp,j Tp ∂2 Tp ables the accurate modeling of intricate physical phenomena, encom­
+ = f V kp − q + Sva (49)
∂t ∂xj ∂xj ∂xj passing temperature and velocity slips between the continuous and
particulate phases, suspensions involving multiple particle types,
The force term (Fi ) represents hydrodynamic interactions between
non-uniformities in particle spatial distribution, interactions involving
particulate and continuous phases, including drag and lift forces. The
collisions between particles, lateral migration of particles, interactions
heat source term (q) accounts for the total enthalpy transfer per unit
with and deposition onto surfaces, individual particle position and
volume from solid to fluid phases, and Sva represents the volumetric
residence time, impacts stemming from particle size distribution, par­
solar radiation absorption within the nanofluid given as the negative
ticle aggregation, considerations of non-spherical particles, and turbu­
divergence of the total radiative heat flux. The pressure term (Pc ) reflects
lent diffusion.
the impact of particle collisions, often negligible in nanofluids as they
Due to the high computational cost associated with fully discrete
are insignificant. Transport properties for the particulate phase (kp ) and
modeling methods, such as MD and SPH, the Eulerian-Lagrangian model
(μp ) should be defined. They are either given using empirical formulas or
is preferable since it offers the ability to capture the physical phenomena
commonly assumed to be equal to those of the base fluid. mentioned above at a relatively low computational cost. Interactions
Regarding the initial and boundary conditions employed in such a between fluid and particulate phases (i.e., particle-fluid and particle-
modeling framework, uniform velocity, fV , and temperature distribu­ particle interactions) determine the forces individually applied to a
tions can serve as inlet boundary and initial conditions. At the outlet particle in the flow. These interactions are called coupling between
boundary, a zero field of relative pressure, zero temperature-gradient, phases. The degree of coupling determines to what extent the model is
and uniform fV distribution are specified [154]. Following the capable of accommodating these phenomena effectively. In one-way
approach of Otanicar et al. [34], heat loss is considered from the upper coupling, only the fluid phase can influence the particulate one. This
surface only, with a prescribed heat transfer coefficient, while the bot­ is acceptable when the suspension is ultra-dilute with (fV < 10− 3 [142]),
tom wall is insulated. Although the no-slip condition is conventionally where the space that particles occupy is sufficiently small to neglect the
applied to solid walls as a reliable condition for fluids, it is not neces­ forces that particles may exert on the fluid phase. In two-way coupling,
sarily the appropriate boundary condition for the solid phase. Indeed, both phases can mutually influence each other. This occurs when the
numerous experimental and detailed Direct Numerical Simulation
volume fraction is higher (10− 3 > fV > 10− 1 [142]), where particles
(DNS) studies have demonstrated significant slip of the particulate phase
occupy a greater volume and the forces they exert on the continuous
at solid boundaries. The slip behavior of solid-phase Partial Differential
phase can no longer be ignored. The particle force source terms are then
Equations (PDEs) is determined by closure equations derived from
involved in the continuous phase momentum equation and applied in
experimental data or more comprehensive DNS computations that
the control volume in which particles are moving through the time step.
establish the slip behavior of the particulate phase based on the funda­
In three-way coupling, disturbances caused by a particle in the contin­
mental first principles [155].
uous phase can affect another particle. In four-way coupling, the model
There are three Eulerian-Eulerian approaches offered in the litera­
takes into account the particle-particle and particle-wall interactions
ture, namely, the volume of fluid (VOF), mixture, and Eulerian ap­
[157]. This occurs when the suspension is highly dense, where collisions
proaches. These approaches are detailed in a review by Biharaei [156].
between particles and other frictional contacts become too important to
It is essential to mention that in order for the particulate phase to be
be neglected. But in highly dense suspension flows, the
deemed as a continuum and thus the Eulerian-Eulerian approach is
Eulerian-Lagrangian model becomes impractical as too many particles
applicable, it is imperative that the mean inter-particle distance (Lp− p ) is
then need to be tracked, requiring extremely high computational cost.
sufficiently smaller than the computational cell resolution (Lp− p ≪ Δx).
The Eulerian-Eulerian model would rather be preferable to treat such
This means that each computational cell should contain a considerable
dense suspensions.
number of particles. However, for solar nanofluids, characterized by
In the Eulerian-Lagrangian approach, the particles being tracked
dilution and extensive inter-particle distances, this condition imposes
could either be actual particles governed by deterministic soft-sphere
stringent constraints on the resolvable fluid characteristics. Conse­
collision models, known as the discrete element approach, or compu­
quently, the notable simplicity of the single-phase homogeneous model
tational parcels representing a group of real particles with hard-sphere
renders it a preferable approach over the Eulerian-Eulerian one for solar
collision models, referred to as the discrete parcel approach. The
nanofluid investigations.
discrete element approach finds common usage in highly dense granular
However, even if the use of the Eulerian-Eulerian approach is justi­
flows (i.e., the flow of grains or particles, such as sand, or powders, in a
fied in terms of the problematic particulate phase continuum condition,
manner distinct from conventional fluid flow) featuring relatively large
the discrete nature of the disperse phase is still overlooked. This means

16
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

particle sizes, such as fluidized beds [158], where interactions with the force from the particle [162]. Furthermore, a reduction in fluid volume
fluid are often negligible. A fluidized bed refers to a condition wherein becomes apparent in the Eulerian cell containing a particle.
solid particles exhibit fluid-like behavior. This state is attained by For moderate and dense suspensions (fV > 6.5 %), the influence of
introducing a gas or liquid upwards through the particles at a velocity particle collisions becomes significant in shaping particle trajectories.
sufficient to overcome their weight. Consequently, the force exerted by Basic deterministic collision models derived from fundamental princi­
the fluid balances the particles’ weight, allowing them to remain sus­ ples, including conservation of momentum and partial mechanical en­
pended and move freely. While the appearance may resemble that of ergy dissipation during the collision, prove inadequate in accurately
boiling liquid, it is essentially a blend of fluid and solid particles. depicting particle interactions, particularly in flows involving non-
However, for nano-sized particles, employing the discrete element spherical particles and particles of varying sizes. Consequently,
approach becomes computationally impractical, limited to very small various probabilistic collision models have been introduced and are
domains with a restricted number of particles and short simulation time presently applied to address these complexities.
intervals [159]. The Lagrangian point-source model stands out for its robustness, ease
In the Eulerian-Lagrangian approach, two description strategies are of implementation, and accuracy in scenarios where particle motion and
available: the point-source modeling and resolved particle modeling. energy exchange minimally impact the fluid’s velocity and temperature
fields. A specialized application of this model is found in Monte Carlo
- The Lagrangian Point-source model (MC) simulations, originally devised to investigate the influence of
turbulence and time-varying velocity and temperature fields on the
Dispersed particles in the Lagrangian point-source modeling are trajectories of both coarse and fine particles. In MC simulations, initial
addressed as discrete points moving in the flow domain as sources of Eulerian computations solve for the average host fluid velocity and
mass, momentum, and energy for the fluid. They are tracked individu­ temperature fields, assuming an absence of particles. Subsequently, a
ally based on Newton’s second law considering that the center of co­ probabilistic methodology is employed to represent fluid velocity and
ordinates is at the particle center of gravity. The Lagrangian point- temperature fluctuations, typically sampled from a random distribution.
source model is preferable as it offers ordinary differential equations A multitude of particles is introduced into the flow field, and their
(ODEs) representing the motion and energy exchange of particles with momentum and heat exchange with the host fluid are determined using
the continuous phase. These equations can be integrated to determine Eqs. 50 and 51. Consequently, ensemble averages for variables of in­
the trajectory of a particle and the temperature profile. The force bal­ terest (e.g., velocity, temperature, and spatial distribution) are
ance and energy equation that govern a particle are, respectively, as computed for the entire particle ensemble in the flow field. In line with
follows: the ergodic hypothesis, these ensemble averages are deemed equivalent
to the time averages of the corresponding variables for the entire particle
dvi ∑
Mp = FBi + FSi + FCi (50) set.
dt
A notable limitation of the simple MC simulations lies in their
dTp ∑ inherent neglect of the impact of particles on fluid velocity and tem­
Mp Cp = Q̇ (51) perature fields, along with the interactions among particles. Conven­
dt
tionally applying this method results in a one-way coupling model,
Mp is the particle mass. The summation of forces in the force balance restricting its accuracy to dilute suspension flows. Consequently, the
(Eq. (50)) involves the body force (FB ), surface force (FS ) which includes results derived from MC simulations are only qualitatively representa­
the lift, steady drag, history, and added mass forces, and collision or tive of the average particle behavior in moderate and dense flows or
other interaction forces with the boundaries or other particles (FC ). regions anticipating higher particle loadings, such as stagnation areas
Similarly, when accounting for the overall heat transfer (Q̇) between with particle accumulation. While MC simulations are anticipated to
phases, both the steady-state and the history terms must be incorpo­ yield accurate descriptions for very dilute nanofluids, their efficacy di­
rated. The overall heat transfer should include the source term repre­ minishes in cases involving more complex particle-fluid interactions.
sented by the volumetric heat absorption. The angular momentum of
particles becomes physically meaningless due to the assumption deem­ - The Lagrangian resolved particle model
ing dispersed particles as points moving in the flow domain.
As for the boundary conditions for the Lagrangian phase, an outflow The Lagrangian resolved-particle model is the model utilized in DNS
periodic boundary condition is applied, where a nanoparticle exiting the of particles. Within this model, the detailed resolution of governing
outlet is reintroduced at the inlet with a randomized y-location, while its equations for continuous-phase velocity and temperature fields occurs in
position, y-velocity, and temperature are reset to zero, and its x-velocity the presence of particles. This facilitates the determination of particle-
is set to the inlet velocity. In contrast, for the particle-wall boundary fluid interactions from fundamental principles, eliminating the need
condition, a particle can either adhere to or rebound from the wall upon for empirical closure equations like drag and convective heat transfer
collision, with the resulting velocity determined by the particle’s kinetic coefficients. Notably, the computational cell resolution of the contin­
energy and the interaction potential between the particle and the wall uous phase in this model, unlike the point-source model, is considerably
[85]. As for the boundary conditions of the continuous phase flow, they smaller than the particle size (Δx≪Dp ). Due to the dense computational
are same as conventional ones used in the single-phase homogeneous mesh, the resolved-particle model demands significantly greater
approach. computational resources for equivalent particle numbers and sizes than
The prerequisites for the validity of the point-particle assumption its counterpart. Consequently, the use of this model for large engineer­
encompass: characteristics of rigid particles, Dp ≪Δx, low particle Rey­ ing systems with a multitude of discrete particles is constrained.
nolds number (Rep ), and low fV [160], and particularly in turbulent Solutions of the continuous phase governing equations provide in­
flows, the particles should be notably smaller than the smallest turbu­ formation about the pressure, velocity, and temperature profiles of the
lence scale [161]. The condition (Dp ≪Δx) ensures that particles expe­ fluid surrounding each particle. By integrating the shear stress, and
rience exceedingly weak gradients in the continuous phase over the temperature gradient profiles over the particle surface (As ), the hydro­
length scale associated with their diameters, allowing the local ambient dynamic surface force, and rate of heat transfer between particulate and
flow around a particle to be reasonably deemed as uniform [162]. When continuous phases can be evaluated using Eqs. 52 and 53, respectively:
(Dp ∼ Δx), the undisturbed fluid velocity at the particle’s location sub­ ∫ [ (
∂ui ∂uj
)]
stantially deviates from the fluid velocity perturbed by the feedback FS,i = − Pδij + μf + nj dAs (52)
∂xj ∂xi

17
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

δij is the Kronecker delta. Finally, the heat transfer rate from continuous ultra-small particles, whose volume does not affect the continuous
phase, the Eulerian single-phase approach can be used to model such
to particulate phases is:
nanofluid systems. In contrast, nanofluids with fV exceeding 0.1 % are
∫ ( )
∂Tf further classified into collision-dominated, when collisions between
Q̇ = − kf n dA (53)
∂xj j s particles affect the nanofluid behavior at intermediate volume fractions,
and contact-dominated, when nanoparticles experience contact rather
The primary benefit of the Lagrangian resolved particle model lies in
than collisions in highly-dense suspension flows. Particle-particle in­
its independence from empirical closure equations governing fluid-
teractions are significant in collision-dominated nanofluid flows, when
particle interactions, as these interactions are determined from funda­
collisions are too frequent to be neglected. This necessitates the use of
mental principles. The only needed empirical input is confined to par­
the Eulerian-Lagrangian approach to capture the discrete nature of
ticle–particle interactions and particle–wall collisions, which are
nanoparticles and account for the effects of particle-particle and
infrequent in solar nanofluids. Additionally, a secondary advantage of
particle-fluid interactions on the nanofluid performance as previously
the resolved-particle model, shared with the point-source model, is its
discussed. However, when the nanofluid is highly dense with fV > 10%
flexibility to employ various well-established models for the continuous
(contact-dominated), the use of the Eulerian-Lagrangian approach be­
phase, including FDM, FEM, DNS, and LBM.
comes impractical since detecting such a huge number of collisions
On the other hand, because it is extremely computationally expen­
demand an unfeasibly high computational cost. Consequently, the
sive to resolve the surface stresses (i.e., shear stress and pressure fields)
Eulerian-Eulerian approach is recommended whereupon. Also, in case of
and flow features at the particle interface, the resolved-particle model is
a dilute suspension with large-size nanoparticles, the Eulerian-Eulerian
only applicable to simple flows containing a single particle or very dilute
approach is recommended since the effect of the nanoparticle volume on
suspensions containing notably large-size particles. This is extremely
the continuous phase cannot be neglected. Fig. 12 represents a guiding
unfeasible for nanofluid systems where a large number of nanoparticles
roadmap for researchers to select the most appropriate approach for
need to be tracked.
real-world nanofluid simulations at the macroscale level at given con­
Although the very low fV is a common condition for the use of the
ditions of fV , Dp , and St.
Eulerian-Lagrangian approach, it can be simply extended to model dense
suspensions. However, an inherent limitation of this model lies in its
demand for substantial memory and computational cost. The consider­ 2.2.2.5. Fully discrete approaches. Although matter is made up of
able multitude of nanoparticles as well as the low St introduce numerical discrete atoms and molecules, most science and engineering systems are
stiffness that requires small time steps. This need arises due to stability amenable to the continuum hypothesis under a certain size scale. The
requisites in explicit numerical techniques and in the implicit ones for justification for the continuum hypothesis validity does not originate
accurate particle drift computation. Consequently, the Eulerian- from specific physical or mathematical principles; instead, it is deduced
Lagrangian model is considered impractical for dense suspension from the absence of conflicts between the resulting continuum
flows. Moreover, the continuous phase equations generally do not description of materials and empirical observations. Indeed, the con­
incorporate considerations for the volume occupied by dispersed parti­ tinuum theory finds support in all existing empirical data for physical
cles. systems containing a sufficiently large number of molecules to be treated
as continua [143]. However, if the volume of space wherein a material is
placed is arbitrarily small, an atom or subatomic particles constituting
2.2.2.4. A guiding roadmap to the appropriate macroscale approach.
the atom might or might not be existent within this space. Then, the
Recent research stated that nanofluids with fV below 0.1 % are classified
uncertainty principle may be recalled to investigate the presence of
as very dilute mixtures, where interparticle forces can be disregarded,
matter within such a tiny volume using quantum mechanics and could
assuming no adhesive forces are present [163]. As such, and when
be further treated using statistical mechanics to derive a probability

Fig. 12. The recommended approach for certain conditions of fV , Dp and St.

18
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

density function. boundary, an identical molecule enters from the opposite boundary,
Consequently, when the system under investigation is sufficiently resulting in a net vanishing boundary force. The incorporation of
small to include a few molecules, the continuum theory is not applicable constraint and driving forces signifies a departure from pure Newtonian
and different modeling approaches should be used such as the MD, DPD, mechanics. The constraint force prevents abnormal energy levels in the
and LBM. Such modeling methods are fully discrete and use the molecular system, aligning with Gauss’s principle of minimum
Lagrangian equations to accommodate the discrete nature of dispersed constraint [143]. The force is mathematically expressed through a
phases, and thus capture the interactions and phenomena and answer closure equation, analogous to the friction equation as:
the questions that continuum-hypothesis-based approaches cannot.
Fci = ζF Mi vi (56)
Fig. 11 Demonstrates positions of the mathematical approaches dis­
cussed in this work with regard to the length and time scales. For large ζF is the Gauss friction coefficient. It possesses both positive and nega­
scales, macroscopic-level investigations can be conducted based on the tive values and typically exhibits time-dependent variation. In equilib­
continuum Navier-Stokes Equations (NSEs) whose solutions can be rium conditions, its time-averaged value is zero. As the system
provided utilizing computational techniques such as the FDM [34,46,57, approaches thermodynamic equilibrium, ζF becomes proportional to the
131], FVM [63,126–129], and FEM [41,130]. For ultra-small length and potential energy’s rate of change within the system.
time scales, the continuum hypothesis is broken down, and MD simu­ Describing non-equilibrium gradients at the molecular level poses
lations are necessary to investigate such micro/nanoscale systems. challenges when dealing with continua. The introduction of a driving
When the time and length scales are less than the continuum hypothesis force (Fdi ) serves as a tool to model these gradients and investigate the
threshold and larger than MD scales, mesoscale approaches such as the resulting transport processes arising from non-equilibrium velocity,
LBM, DPD, and SPH are recommended. temperature, or concentration gradients. As it is unfeasible to directly
measure these forces at the molecular scale, the selection of appropriate
- Molecular dynamics forces relies on insights from MD simulations and their validation
through experimental data. The driving force is determined, demon­
In nanofluid systems, the breakdown of continuum assumptions strating its effectiveness when employing weights (Wi ) for the compu­
becomes notable when dealing with extremely small length scale (≤ 10 tation of space- or time-averaged variables, such as the average particle
nm) [164], namely, when the length scale of the fluidic system is in the velocity (vp ):
same range as the mean free path (i.e., Kn ≥ 1). Such scales in nanofluid
flows are too small to use the continuum assumption for simulations, 1 ∑N

which necessitates transitioning to the atomic scale through MD simu­ vp = Wk vp,k (57)
N k=1
lations. MD simulations are based on computing the forces acting on
each molecule due to its interactions with neighboring ones and with By employing Lagrange multipliers, an expression for the driving
system boundaries. Each molecule is treated as a separate particle, force based on other forces influencing the particle and the gradient of
which follows Newton’s second law in response to the force system. By the variable in question can be derived as follows [143]:
numerically solving the motion equations for all molecules, MD simu­ Wi [ ( )]
lations reveal how the whole system of molecules behaves under the −
m
∑N
Wk dv
driving forces. The average behavior of the molecules provides infor­ Fdi = ( )2 Fak + Fbk + Fck − N (58)
N
∑ Wk k=1
mk dt
mation on the macroscopic process of interest [165]. The governing mk
k=1
equation of a molecule movement is given as follows:
Reviews [167,168] comprehensively covered recent MD simulation
d2 r i advances in systems employing nanofluids, addressing aspects such as
Mi = Fai + Fbi + Fci + Fdi (54)
dt2 nanofluid thermophysical properties, thermal wall modeling, etc. These
Mi and ri are the mass and position vector of the i-th molecule or particle reviews provided critical insights and identified essential areas for
at time t, respectively. The forces on the right hand side, Fai , Fbi , Fci and future research. MD has been applied in various nanofluid in­
Fdi , represent the intermolecular forces applied by other neighboring vestigations, exploring heat transfer characteristics, studying surface
molecules in the computational domain, forces applied by wall bound­ nanostructures, and investigating chaotic movements of nanoparticles.
aries, constraint forces, and driving forces, respectively. The force Fai is Furthermore, Wei et al. [169] highlighted that molecular motion,
usually given by the sum of the effects of potential functions that cause particularly without a temperature gradient, plays a key role in low
forces, like gravity and electrostatic or atomic interaction potentials. For viscous heat transfer as revealed by non-equilibrium MD simulations.
the latter, the interaction forces between two molecules or particles (i Also, using the LAMMPS code, Toghraie et al. [170] successfully utilized
and j) are often modeled by the Lennard-Jones potential function [166]: MD to simulate Poiseuille flow in a hybrid platinum/copper nanofluid,
[( ) where atomic interactions are governed by the Lennard–Jones potential.
( )6 ]
→ Γ
12
Γ While MD simulations can accurately depict the separation-
F LJi = 4ϵ − (55) dependent potential of mean force among particles, their highly
rij rij
demanding computational costs make them entirely not feasible for
where ϵ and Γ are the bonding energy and distance, respectively. The nanofluid systems characterized by a large number of dispersed particles
force Fbi signifies the impact of boundary molecules on the molecule and diverse length and time scales. Consequently, MD investigations of
movement within the computational domain. As Fbi embodies inter­ particulate flows are confined to oversimplified systems such as those
molecular forces, its typical expression can be given in the form of Eq. with Argon host fluids and very dilute suspensions. However, these
(55) with different constants, ϵ and Γ. In the examination of homoge­ systems do not accurately represent or mimic the real-world conditions
neous systems, the assumption is made that the system extends infinitely encountered in solar nanofluid applications.
in all directions (using the so-called periodic boundary condition), and
an MD study is conducted on a small part of this extended system. The - Dissipative particle dynamics (DPD)
MD study, then, investigates the interactions and behavior of a limited
number of molecules within a domain representing a fraction of the Despite the detailed insights provided by MD, its high computational
homogeneous fluid. Periodic boundary conditions are applied in this cost limits its applicability for modeling macroscopic system properties
scenario, ensuring that when a molecule exits the computational at the scale of real-world applications, given the substantial number of
particles and ultra-short time scales involved, even with the top-of-the-

19
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

line supercomputers. Nonetheless, when considering the length and time )

scales relevant to energy transfer in nanofluids, these fall within the →C ∑ (
f ij = aij ωC rij →e ij (62)
intermediate range between continuum and MD scales, commonly j∕
=i
referred to as ’mesoscales’. Effectively capturing these mesoscales can
)
be achieved through a technique known as ’coarse graining’ of particles, →D ∑ (
where each coarse-grained particle represents a sizable cluster of actual f ij = v ij →
− γ ij ωD rij .→ e ij (63)
j∕
=i
particles or fluid molecules. Therefore, analyzing heat transfer in
nanofluids stands to gain valuable insights by leveraging mesoscale )
coarse-grained models to mimic microscopic insights of energy transfer →R ∑ (
f ij = − σ ij ωR rij ζij Δt − e ij
1/2 →
(64)
between the dispersed particles and host medium particles/molecules. j∕
=i
These methods efficiently model relatively bulk physical systems while
capturing essential interaction details within fluid molecules. A notable where rij = ri − rj and vij = vi − vj . The coefficient aij governs the po­
recent coarse-grained approach of this kind is the DPD technique, which tential energy and the inter-particle interactions within the DPD
was found far more feasible than MD techniques with regard to framework, thereby influencing the system’s equation of state and its
computational time. compressibility. Furthermore, the parameters γij and σ ij are responsible
The DPD serves as a coarse-grained alternative method to MD, for modulating the intensity of the dissipative forces and the stochastic
wherein each DPD particle corresponds to a set of actual molecules. forces akin to Brownian motion in the DPD context, respectively. The
These DPD particles are distributed randomly throughout the flow stochastic variable (ζij ), which is characterized by a zero mean and a
domain, adhering to mass, momentum, and energy conservation prin­ variance of one, possesses the symmetry (ζij = ζji ), a condition that is
ciples. The DPD method was applied to various heat transfer techniques, essential for the conservation of overall momentum in the DPD system.
encompassing conduction, convection, and natural convection with ) ) )
ωC (rij , ωD (rij , and ωR (rij are the weight functions for the conservative,
fluid attributes varying with temperature. Thereafter, Abu-Nada [171, dissipative, and random forces, respectively. More details on the weight
172] further developed the DPD methodology for simulating heat functions calculations can be found in Ref. [171].
transfer in nanofluids. The approach successfully replicated energy
transfer processes by analyzing thermal fluctuations driven by Brownian - Lattice Boltzmann method (LBM)
motion. Additionally, the method explored the inter-particle energy
exchange between adjacent fluid particles [173]. Findings suggested The LBM is apt for addressing phenomena occurring in nanofluid
that the DPD method is aptly applicable to nano/microscale systems and systems wherein particle-fluid and particle-particle interactions are
holds potential for extension to diverse practical heat transfer systems. significant. This method uniquely incorporates MD techniques to bridge
Further details on the mathematical formulation of the DPD approach the gap between microscopic and macroscopic perspectives. It smoothly
are available in the literature [164,171,172]. integrates the microscopic interactions between fluid particles with the
The DPD method demonstrates advantages over conventional numerical model. In contrast to the NSEs, which capture the flow dy­
methods in terms of its applicability to multiphase flows, straightfor­ namics at the macroscale, the LBM is powerful at addressing interactions
ward extension from 2D to 3D, and capability for modeling inter-particle of fluid particles, providing insights into the mesoscale mechanisms of
interactions at the microscopic level, incorporating both attractive and the flow dynamics. In this approach, conservation equations are
repulsive forces between fluid molecules and the surrounding media. On resolved by discretizing the Boltzmann transport equation on a grid
the other hand, this method suffers from some limitations including the lattice by presuming that “artificial particles” that are mesoscopically
absence of a direct link between DPD parameters and the system’s positioned over the lattices and distributed according to the probability
physical ones, as thermophysical properties are not input parameters to distribution function f(x, ξ, t) [174,175]. This density distribution
the DPD model, but they are outputs of the simulation. This requires function characterizes the number density of dispersed nanoparticles
tuning of the microscopic DPD numerical parameters (conservative, with the microscopic speed (ξ) at the location (x) and time (t). The
dissipative, and random) to match the macroscopic thermophysical computational domain is discretized into a lattice that encompasses the
properties of the real systems. Additionally, the computational cost of entire geometry. The discrete lattice is employed to represent spatial
the method is high in comparison with the FDM and FVM. variations and discrete time, creating distinct nodes for the host fluid
For the mathematical formulation of the DPD approach, a multitude and dispersed particles. The fluid-particle interactions are stipulated at
of coarse-grained DPD particles are dispersed throughout the flow re­ boundary nodes or link joints proximate to these nodes. The LBM is used
gion to simulate the nanofluid. The motion and interactions among these as a particle probability distribution function, offering a robust alter­
nanofluid DPD particles are described, employing the Boussinesq native numerical method to conventional macroscopic approaches that
approximation to account for buoyancy forces, as follows [171]: discretize governing equations, like NSEs. This method features a par­
d→ri allel feature, allowing for fast implementation for complicated geome­
vi
=→ (59) tries. Furthermore, its computational time and cost are notably reduced
dt
compared to microscopic methods, rendering it an appealing choice
d→vi →C →D →R across various fields, particularly in addressing heat and fluid flow is­
= f ij + f ij + f ij + →
g β(T − To ) (60)
dt sues. The LBM equations for a multi-component system for nanofluids
are detailed in Xuan and Yao [174] work. The discretized form of the
dTi ( visc )
lattice Boltzmann equation (knowing that ei is the lattice velocity, fiσ is
Cv = qij + qcond + qRij (61)
dt ij
the particles distribution function of component σ with lattice velocity ei
in the i th direction at lattice x, fiσ,eq is the equilibrium distribution
ri and →
→ vi are the position and velocity vectors. Cv is the specific heat at
function, and τσ is the non-dimensional collision-relaxation time con­
constant volume. qvisc cond
ij , qij and qRij are the heat flux vectors accounting stant of the σ th fluid component):
for the conservative, viscous, and random heat fluxes, respectively. β is
→C →D →R 1[ ]
the thermal expansion coefficient, and f ij , f ij and f ij are the conser­ fiσ (x + ei Δt, t + Δt) − fiσ (x, t) = − fiσ (x, t) − fiσ,eq (x, t) ; σ = 1, …, S
τσ
vative force, dissipative force, and random force, respectively, given as (65)
follows [171]:
The representation of lattice configuration, Dn Qm , involves Dn as the

20
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

dimension of the lattice structure (n might be 1, 2, or 3), and Qm signifies ⎧

the number of speeds allocated to each particle at a lattice point, indi­ ⎪
⎪
⎪ 4
⎪ i=0
cating the permissible directions for particle movement within the lat­ ⎪
⎪
⎪
⎪ 9
⎪
tice domain [176]. D3 Q19 and D3 Q41 lattices are the most commonly ⎨
1
used models employed for 3D simulations [164]. Fig. 13 shows the ωi = i = 1, …, 4 (68)
⎪
⎪
⎪ 9
schematic of D3 Q19 lattice. Following the determination of the lattice ⎪
⎪
⎪ 1
⎩ 36 i = 5, …, 8
⎪
arrangement, the lattice Boltzmann equations governing flow dynamics ⎪
⎪
and heat transfer are then solved.
Using the Taylor expansion technique, the discrete Boltzmann
ωi is the weight coefficients and uσ,eq is the equilibrium velocity of the
equation can be derived from Eq. (65). A nanofluid refers to a mixture
σ -th component particles. The local number density of the σ-th compo­
comprising nanoparticles and a base liquid, denoted by σ = 1, 2. As
nent’s particles is [174]:
emphasized by Wolf-Gladrow [178], the lattice Boltzmann model con­
∑
sists of a lattice structure, an equilibrium distribution function, and a f σ (x, t) = fiσ (x, t) (69)
kinetic equation known as the lattice Boltzmann equation (LBE). i
Depending on the specific hydrodynamic phenomena under consider­
Relations between the macro-scale physical quantities of a nanofluid
ation, various lattice configurations can be chosen. Consequently,
flow and their counterpart in the micro-scale level such as the mass
several types of multi-speed lattice Boltzmann models have been pro­
density and the σ -th component’s velocity can be found in Ref. [174].
posed in the literature. For 2D scenarios, the widely utilized D2Q9 model
For nanofluid simulations using the LBM, it is essential to examine
is prominent. This model, employing a 2D octagonal lattice [178], in­
the dynamic factors influencing nanofluid flow. In the absence of
corporates nine discrete velocities, defined as follows:
external fields, nanoparticles experience buoyancy and gravitational
⎧ (FB ), Brownian (FBr ), and drag (FD ) forces, as well as dispersion force
⎪
⎪ (0, 0) i = 0
⎪
⎪
⎪ ( ) from repulsive potentials (FA ). Dominance of FBr may lead to nano­
⎨ c cos (i − 1)π , sin (i − 1)π
⎪
⎪
i = 1, …, 4 particle aggregation and cluster formation during random motion, while
ei = 2 2 (66) a dominant repulsive force can disrupt clusters, suppressing particle
⎪
⎪ ( )
⎪ √̅̅̅
⎪
⎪ (2i − 9)π (2i − 9)π aggregation [174]. Under the influence of gravitational force, nano­
⎩ 2c cos , sin i = 5, …, 8
⎪
⎪
4 4 particles tend to sediment, with larger clusters experiencing relatively
faster sedimentation.
c = Δx/Δt represents the reference lattice velocity, and the indices i = 1, FB can be given, knowing that the parameter a represents the
…, 4 and i = 5, …, 8 correspond to the orthogonal and diagonal di­ dispersed particle radius, while Δρʹ denotes the difference in mass
rections on the lattice, respectively. For the D2 Q9 model, three distinct density between dispersed particles and host fluid, as follows:
types of particles are considered at lattice site (x) and time (t): the 4 3
equilibrium distribution is described by the distribution functions f0 for FB = − πa gΔρ’ (70)
3
the stationary particles, fi (where i = 1, …, 4) for particles moving in
orthogonal directions, and fi (where i = 5, …, 8) for particles moving Knowing that Δu is the difference in velocity among dispersed par­
diagonally. ticle and surrounding liquid, FD is:
Selecting an appropriate equilibrium distribution function is crucial FD = − 6πμf aΔu (71)
to ensure that each component complies with the macroscale Navier-
Stokes equations. Specifically for the standard D2 Q9 model, the equi­ FBr represents the resultant impacts of interactions from the liquid
librium state’s single-particle distribution functions are outlined as fol­ molecules around a particle. It is given based on the equilibrium state
lows [179]: wherein the dispersed particle’s kinetic energy is evenly divided be­
[ ] tween its translational modes. Knowing that the direction of FBr is
(ei uσ,eq ) 9 (ei uσ,eq )2 3 (uσ,eq )2
fi = ωi ρ 1 + 3
σ,eq σ
+ − (67) random and Cr is a random number satisfying the normal distribution,
c2 2 c4 2 c2
FBr can be given statistically over a time interval (Δt), as [86]:
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
12πaμf kB Tf
FBr = Cr (72)
Δt

Tf is the base fluid temperature. kB Tf accounts for the thermal energy

generated by the irregular Brownian motion, and Δt is the time step. The
van der Waals attraction energy (GA ) between two particles at a distance
(Lp− p ) from each other is given as follows:
[ (2 )]
A11 2 2 s − 4
GA = − + 2 + ln (73)
6 s − 4 s
2 s2

( )
where s = 2a + Lp− p /a, and A11 is the Hamaker constant which rep­
resents the strength of the van der Waals interaction between particles or
between a particle and a wall. I.e., it reflects the attractiveness between
objects due to van der Waals forces, and depends on the refractive index
of the material used and dielectric function of the base fluid. The force
arising from the interaction potential among particles located in the
neighboring lattice containing ni particles within the D2 Q9 model is:
8
∑ ∂GA
Fig. 13. Schematic of D3 Q19 lattice [177]. (Reused with permission FA = ni (74)
i=1
∂ri
from Elsevier).

21
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

The van der Waals interactions between particles and the Hamaker lowed by integration across the entire computational field. The process
constant will be discussed in more detail in a coming section. Knowing is represented mathematically as [186]:
that V is volume of a certain lattice and n is the number of particles it ∫
contains, the resultant force vector (Fp ) accounting for the sum of forces fi = f(r)Wk (r − ri , hm )dr (77)
applied to a dispersed particle in a nanofluid with a given lattice volume Ω

(V) is given as follows: where fi , Ω, f(r), and hm denote the kernel estimate at particle i,
n computational domain, position vector, and smoothing length, respec­
Fp = (FA + FB + FBr + FD ) (75)
V tively. Wk (r − ri , hm ) is the smoothing function with an effect limited to a
domain characterized by hm . The corresponding particle approximation
Finally, the forces exerted on fluid molecules (FM ) within a specific of f is then given as follows [187]:
lattice can be obtained from the reaction of FBr and FD [180]:
Np
∑ Mj ( )
FM = − n(FBr + FD ) (76) fi = f j W k rj − ri , hm (78)
j=1
ρj
In the domain of solar energy, LBM has been utilized in several studies
[181,182] to solve the RTE and a conduction-radiation model, respec­ Np , Mj , and ρj are the total number of particles in the domain of the
tively. In a recent review [183], the existing literature on the utilization particle i, mass and density of the adjacent particle j, respectively. To
of LBM in the context of solar thermal collectors and heat transfer approximate a spatial derivative fi within the kernel framework, one can
modeling was surveyed. The review stated that the LBM has seen diverse insert the derivative into Eq. (79), apply the technique of integration by
applications in recent years and offers several advantages over con­ parts, and disregard the boundary conditions. The process is summa­
ventional CFD techniques, making it a suitable choice for conducting rized as follows [187]:
both small and large-scale computations for solar collectors. These ad­ Np
vantages include the elimination of the need to solve coupled
∑ Mj ∂Wk,i (r)
fi,φ = f (79)
pressure-velocity equations, a straightforward calculation procedure j=1
ρj j ∂rφ
with ease in handling complex geometries and treating boundary con­
ditions, simple numerical implementation, compatibility with supple­ ∂Wk,i (r)/∂rφ is the spatial derivative of the particle i along the φ direc­
mentary methods for developing hybrid approaches to solve the RTEs, tion, where φ serves as the direction index iterating from 1 up to the total
and notable numerical stability and efficiency. In light of these benefits, number of spatial dimensions. The SPH method’s kernel approximation
Siavashi et al. [184] introduced an in-house multi-relaxation time LBM for a field function’s derivative facilitates the computation of the spatial
model, for the first time in the literature, to simulate a DASC employing gradient using the function’s values and the smoothing function’s de­
a water-SWCNT nanofluid. The model was validated against results by rivatives, bypassing the need for the function’s own derivatives. Liu
Gorji and Ranjbar [126] and obtained very strong agreement with a et al. [188] have noted that such SPH kernel approximations of a field
maximum deviation of 1.29 %. function and its derivative maintain second-order accuracy within the
According to Mahian et al. [164], the LBM exhibits several advan­ core domain. However, at the boundaries or interfaces, these approxi­
tages, including scalability, stability, and compatibility with parallel mations fall below first-order accuracy due to the limited approximation
computing. It is particularly well-suited for handling complicated ge­ domain, a phenomenon commonly known as the boundary deficiency in
ometries, making it a preferred choice for simulations involving complex standard SPH methodologies. The readers are advised to refer to the
shapes. The method excels in calculating multiphase flows and finds review [187] to look into the mathematical formulation of the
application in mesoscale scenarios, such as colloidal systems and fluid early-improved and corrective SPH approach and its applications in
flow through small pores in porous media. Additionally, LBM allows for different science and engineering areas.
the incorporation of microscopic interactions between fluid molecules. Fig. 14 offers approximate assessments for the mathematical ap­
However, certain limitations exist: the modeling of variable proaches discussed above and determines the scope of each based on the
thermo-physical attributes, such as μ and k, presents challenges; the flow type characterized by Knudsen number (Kn). Kn, which is the ratio
method’s reliance on a uniform grid hinders the straightforward treat­ between the mean free path and the characteristic length of the flow
ment of fine gradients near boundaries, leading to computational de­ system, provides numerical values assessing the flow and determines
mands in areas like shear and boundary layers; incorporating bulk whether the continuum hypothesis is valid. Fig. 14 also presents com­
viscosity is indirect and related to the relaxation time; implementing parisons between the mathematical approaches in terms of the compu­
axisymmetric flows can be complex; and handling multiphase flows with tational cost, the depth of the physical insight they introduce, and length
significant density variations poses challenges. and time scales of the system.

- Smoothed particle hydrodynamics (SPH)

2.3. Thermal energy and exergy analyses
The SPH is one of the oldest meshless particle methods used for CFD
To assess the thermal performance of a nanofluid-based DASC sys­
simulations of Lagrangian nature. The fundamental concept behind
tem, a proper thermal analysis should be accomplished to evaluate the
meshless methods is to deliver accurate and stable numerical solutions
performance criteria or indicators chosen for the system. While thermal
for integrals or PDEs, encompassing a variety of potential boundary
energy analysis provides valuable insights, it falls short in capturing the
conditions through an arbitrary distribution of nodes. The SPH method
full picture. Exergy analysis represents a powerful tool that comple­
is based on converting the continuum conservation equations that
ments and enhances our understanding of such energy systems. Exergy is
govern the flow, specifically the NSEs, from PDEs to integral equations
the amount of work obtainable when a substance reaches a state of
through the utilization of an interpolation function that computes the
thermodynamic equilibrium with the natural surroundings through
kernel of the flow variables at any given location. The SPH demonstrates
reversible processes, interacting only with the components of nature.
favorable outcomes in simulating particulate two-phase flows [185].
This definition captures the system’s potential to perform useful work.
In the SPH technique, the primary PDEs are discretized through a
Exergy provides a deeper understanding of energy quality and the po­
two-step process involving kernel and particle approximations. For a
tential for useful work within a system, making it a valuable tool in
given function (f) at a specific location or particle (i), the kernel esti­
thermodynamics and engineering.
mation is derived by multiplying f with a smoothing kernel (Wk ), fol­
Authors, in this section, propose some performance criteria, found in

22
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

Fig. 14. The position of the discussed simulation approaches with regard to Kn, physical insight offered, length and time scales, and computational cost.

the literature, to evaluate the performance of DASC systems. The ab­ conditions, the exergy balance of a DASC system can be expressed in rate
sorption efficiency (ηa ) describes how efficient a nanofluid is at ab­ form as follows:
sorption of incident solar radiation (Qinc ) while traveling through the
EXinc = EXu + EXloss (86)
DASC:
The left-hand side parameter (EXinc ) represents the rate of exergy
Qa
ηa = (80) input to the DASC from the solar source. EXinc can be estimated by the
Qinc
product of Qinc and a well-established factor called the Petela-Landsberg-
The solar radiation absorbed by the nanofluid (Qa ) and Qinc are given, Press factor [189]:
respectively, as follows: [ ]
4Tamb T4
∫ λmax EXinc = Qinc 1 − + amb (87)
3Tsolar 3Tsolar
4
Qa = LW Ia dλ (81)
The Petela-Landsberg-Press factor, the expression between brackets
λmin

∫ λmax in Eq. (87), in the context of radiative processes, is a dimensionless

Qinc = LW Iinc dλ (82) parameter analogous to the Carnot efficiency for heat engines. This
factor quantifies the upper limit of theoretical work extractable from
λmin

L and W are the length and width of the collector, respectively. Ia and Iinc solar radiation. Furthermore, drawing upon the work of Evola and
are the spectral radiation heat flux absorbed and incident, respectively. Marletta [190], the recoverable exergy (EXu ) can be given as the dif­
Another performance indicator is the thermal efficiency (ηth ), which ference in the working fluid’s exergy content between the collector
is the ratio of the useful thermal power gain (Qu ) effectively carried by outlet and inlet. The mathematical expression for EXu is provided based
the nanofluid after flowing through the collector to the solar radiation on the Gouy-Stodola theorem [191] that gives the exergy carried by a
absorbed by the nanofluid (Qa ), expressed as: flowing fluid that experienced a temperature increase from Tin to Tout :
∫ y=H ( )
Qu ( ) Tamb
ηth = (83) EXu = ρnf cp,nf W unf (y) Tnf ,out (y) − Tnf ,in 1 − | dy (88)
Qa y=0 Tavg (y) x=L

The average temperature of the nanofluid, Tavg (y), is given as

Qu can be given as follows [138]:
follows:
∫ y=H
( )
Qu = ρnf cp,nf W unf (y) Tnf,out (y) − Tnf,in |x=L dy (84) Tnf ,out (y) − Tnf,in
y=0 Tavg (y) = ( )
Tnf,out (y) (89)
ln
Tnf ,in
ρnf , cp,nf , unf , Tnf,out (y), and Tnf ,in are the nanofluid’s density, specific heat
capacity, velocity field function, temperature field function at the col­ Finally, the overall second-law efficiency of the DASC system is given
lector outlet, and temperature at the collector inlet, respectively. One as follows:
more performance indicator is the collector efficiency (ηІ ), which rep­
EXu
resents the overall first-law DASC efficiency. It is obtained by the ηІІ = (90)
EXinc
multiplication of ηa and ηth , as follows:
It is crucial to acknowledge that the Petela-Landsberg-Press factor, as
Qu
ηІ = ηa ηth = (85) implemented in Eq. (87), represents the maximum achievable work
Qinc
under the assumption of undiluted direct solar radiation [192]. Conse­
Looking at the DASC through the lens of the first-law of thermody­ quently, its application in this context serves as an approximation for
namics, ηІ captures how well it performs overall. This considers both unconcentrated solar radiation, which typically encompasses a combi­
how well it absorbs light (optically) and how well it traps heat nation of direct and diffuse radiation components.
(thermally). The term ’direct beam radiation’ signifies solar radiation that im­
Let’s delve into the exergy analysis. When the effects of kinetic and pinges upon the Earth’s surface in a straight-line path originating from
potential energy changes are neglected, and under steady-state the sun. Diffuse radiation, in contrast, describes sunlight that has

23
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

undergone interactions with atmospheric constituents like particles. photothermal conversion potentials compared to other types of nano­
These interactions cause the radiation to be scattered in various di­ fluids [206]. Authors assumed that the flow field is parabolic along the
rections, ultimately reaching the Earth’s surface with a less defined path. entire channel (fully developed Poiseuille flow) [146], and nanofluid’s
The literature offers more refined exergy-to-energy conversion fac­ thermophysical characteristics are same as those of pure water as
tors that consider the disparity between the exergy content of direct and frequently stated in the literature for fV < 0.01% [146,207]. Results
diffuse solar radiation [193–196]. For instance, Press [193] reported illustrated that the mean radiative path depth of light photons’ pene­
that the exergy content of diffuse radiation is roughly 25 % lower tration through the collector channel increases with radiation scattering
compared to direct radiation. Similar findings were reported by Lands­ by nanoparticles and thus the solar-weighted absorption coefficient (Am )
berg and Tonge [196]. This discrepancy can be attributed to the loss of nf
is enhanced. On the other hand, the Ks increase causes the back scat­
directivity associated with diffuse radiation, which inherently repre­ nf
tering to increase as well, and thus, an optimum value for Ks , at which
sents an irreversible process.
light absorption by the nanofluid and collector photothermal conversion
In essence, the Petela-Landsberg-Press factor provides a simplified
efficiency are maxima, exists. Authors found that OpenFOAM software is
approach for unconcentrated solar radiation, and for more accurate
characterized by the ease of customization, capability of capturing fine
assessments, factors accounting for the difference between direct and
differences in the flow’s thermal and optical behavior, and that it pro­
diffuse radiation are recommended.
vides several solvers and applications such as the solar collector [197].
Modeling of nanofluid flows in DASCs is often conducted based on
2.4. Critical assessment of current theoretical DASC investigations the fully-continuous single-phase Eulerian approach benefiting from the
homogeneous-nanoparticle-distribution assumption [208]. The short­
DASC systems have undergone analyses through experimental, nu­ comings of this approach were detailed in the previous subsection. Das
merical, and analytical investigations in existing literature. A dedicated et al. [209] conducted a computational study to assess the performance
chapter in this research is allocated to exploring the experimental of a DASC system. Nanofluids containing TiO2 nanoparticles (95 nm in
approach. The current section is specifically designed to meticulously diameter) dispersed in water at varying concentrations (0.24, 0.4, 0.6,
evaluate the existing theoretical studies, encompassing both numerical 1.18, and 2.0 %) have been selected for the study. The DASC system is
and analytical approaches, comparing them, and exploring their po­ characterized by a constant heat flux applied to the top wall, a plug
tential for evolution and enhancement. profile specified at the inlet, and a pressure boundary condition imposed
at the outlet. Both single-phase and mixture-phase models were solved
2.4.1. Numerical studies using the ANSYS FLUENT 13.0 solver. It was assumed that all the
Numerical attempts to report the thermal and radiative behavior of thermophysical properties of the nanofluids are dependent on temper­
nanofluid-based DASCs are built based on introducing numerical solu­ ature. The findings manifested that the utilization of a multiphase model
tions to the RTE and energy equation. Tyagi et al. [57] employed the yielded more accurate results compared to a single-phase model when
FDM to present a 2D numerical solution to this system of governing comparing with experimental results. Also, shortcomings of the
equations and simulate the DASC. This solution was verified by a study single-phase approach drove Sharaf et al. [85] to analyze the DASC from
introduced by Otanicar et al. [34] in which several nanofluids were the perspective of the Eulerian-Lagrangian approach to reveal the effect
numerically investigated and the results were experimentally validated of the discrete nature of nanoparticles on the thermal features of the
by an experimental work included in their study. After these two system. They concluded that there is a threshold beyond which the
groundworks had been introduced, the pieces of research began to single-phase Eulerian approach becomes invalid (i.e., Re ≈ 1). As long as
emerge for DASCs’ numerical analyses utilizing numerical techniques Re ≤ 1, the assumption of a homogeneous single-phase fully-continuous
such as the FDM, FEM, and FVM. Each one of these methods has its own flow is valid and thus the single-phase approach can be used, and
advantages. The FDM, for instance, is characterized by the ease of consequently, thermal equilibrium and no-slip assumptions between the
implementation and building numerical codes to solve the discretized nanoparticles and base fluid can be stated. Beyond that value, the
governing equations of simple-geometry DASCs, such as the planar spatially non-uniform light extinction due to the uneven nanoparticle
collector, and different geometries. Although this method is rarely spatial distribution within the nanofluid and thus spatial variations in
employed for commercial CFD software packages [197], it is frequently the nanofluid optical characteristics yields inaccurate performance
used to analyze nanofluid-based solar thermal systems [59,198–203]. predictions. Temperature profiles generated based on the single-phase
The FVM is employed in commercial CFD packages such as ANSYS that and Eulerian-Lagrangian approaches start showing considerable de­
is frequently used in the literature [63,126–128] for simple and regular viations when Re exceeds unity, as the single-phase approach over­
geometries [197]. The computational domain is discretized into finite estimates the temperature profile with increasing Re. The single-phase
control volumes (cells), then, the system of equations is solved to approach’s inaccuracy is attributed to the overestimation of solar radi­
compute the properties at the center of each cell. Finally, the FEM is ation extinction within the vicinity of the collector’s upper surface,
more stable than both the FDM and FVM as it discretizes the computa­ particularly when Re ≥ 1. The lack of light extinction in the vicinity of
tional domain into finite elements and solves the system of governing the upper wall is due to particle migration from high shear regions at
equations to predict the flow properties at each node [130,204,205]. high speeds (see section 6.3).
Since this technique is stable and efficient to resolve the flow in different Also, the plug flow assumption (i.e. flow in which all particles in a
geometry types and conditions, it is employed in several commercial given cross-section have the same velocity’s magnitude and direction)
numerical software packages [197]. was frequently used in the founding literature [34,57,131,136]. This
A new leading open-source computational mechanics toolbox assumption limits the model ability to capture the actual flow and heat
employing the FVM called OpenFOAM software was utilized by Won transfer features in DASCs since the plug flow is valid for creeping
and Lee [129] to build a 2D model to simulate a DASC with a plasmonic (Re≪1) or inviscid flows, which are not the real case in DASCs wherein
nanofluid and numerically investigate its thermal and optical behavior Re is between 10 and 1000 [57,131]. Lee and Jang [210] found that the
revealing the effect of scattering on the DASC performance. The plug flow assumption in DASC modeling results in overestimations of
Monte-Carlo method was used to numerically solve the RTE for the ηPC , particularly when the heat loss coefficient (hL ) and Re are high. The
volumetric heat generation through the nanofluid to be obtained and fully-developed flow assumption is more realistic in DASC modeling
implemented in the model. A plasmonic nanofluid was selected for the [146,210] although it neglects the entrance region impacts on the col­
study since its optical characteristics can be tuned by selecting the lector performance.
appropriate shapes and sizes of metal nanoparticles employed, also Furthermore, several studies [44,59,136] have overlooked the
because of the inherent LSPR effect that offers high thermal gain and

24
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

impact of atmospheric attenuation on the extraterrestrial solar spec­ fluid accounts for the entire collector medium and its optical charac­
trum. They assumed a blackbody profile for incident radiation heat flux, teristics can significantly influence the system performance [44]. The
leading to inaccurate performance descriptions. This is because the model accounts for both the radiative and convective losses combined
out-scattering and absorption efficiencies of nanofluids are dependent together by a single heat loss coefficient to model the surface heat losses
on radiation wavelength, as argued in Refs. [100,137]. when solving the energy equation. Authors aimed to study several
The effects of the base fluid’s absorption, emission, and scattering are operating parameters, such as the particle loading (fV ), channel height
often neglected in the literature due to the difficulty of solving the RTE (H), and concentration of sunlight, and to optimize the thermal perfor­
in the medium. To show the importance of considering such effects in mance of the system. The approximation of the combined heat loss co­
the thermal conversion process in DASCs, Zhu et al. [211] developed a efficient (hE ) introduced by authors is valid for temperatures less than
comprehensive radiation model for a DASC accounting for these effects, 750 K. Most solar thermal systems fall under this temperature limit due
and provided a reliable method to evaluate the heat loss in the system. to limitations caused by material degradation. Severe radiative losses,
The model improved the estimation of heat loss and examined how that can no longer be neglected, occur beyond such an elevated tem­
different fluids affect the heat loss. Moreover, the model used RSM to fit perature. Solar radiation absorption by seeded nanoparticles and volu­
the heat loss and effective emissivity with a maximum error of 5.6 %. metric heat release within a flowing nanofluid was the base on which the
The model is robust even when 2 % measurement errors were intro­ analytical model was formulated. The analytical solution offered a
duced, and the results are still within the acceptable accuracy range. prediction tool to reveal how the operating variables affect the tem­
The reports on DASC systems with turbulent flow are very limited in perature distribution and ηPC of the solar system. The operating vari­
the literature although such a flow regime is possible in such industrial ables are given in the study in terms of dimensionless numbers, such as
applications [197]. Furthermore, optimization studies dedicated to the Peclet number (Pe = Pr Re = ρUcp H/kc ), Nusselt number (NuE =
DASC systems are also limited although some pieces of research have hE H/kc ), dimensionless irradiance (G = CGs H/kc Tin ), dimensionless
been recently presented to the literature. Baro et al. [212] performed a temperature difference (θ = kc (T − Tin )/CGs H), dimensionless heat
numerical multi-objective optimization study on an inclined planar
release (q̇ʹʹ(y) = q̇ʹʹ(y)H/CG ), and dimensionless coordinates
s
DASC (0.5 ≤ L ≤ 2 m, 0.5 ≤ W ≤ 1 m, 5 ≤ H ≤ 50 mm) with water/­
(x = x/Pe H and y = y/H). C, Gs , and Tin are the solar concentration
Al2O3 nanofluid of fV ranged from 0 to 0.006. Results demonstrated that
factor, irradiance, and inlet temperature, respectively. The flow was
ηPC,max of 98.4 % with temperature gain of 7.81 ◦ C were obtained at
assumed to be a plug flow, and since the nanofluid concentrations
optimum values (L = 0.584 m, W = 0.566 m, H = 0.05 m, ṁ = 0.01 kg/s, considered in the study were very low (fV < 0.006), their thermophys­
Tin , = 300.26 K, and fV = 0.0016). Also, the maximum temperature gain ical properties such as ρ, cp , and kc were assumed to be the same as the
of 71.93 ◦ C with ηPC,max of 59.9 % was obtained at optimum values (L =
base fluid’s. The dimensionless parameters were used to
0.5 m, W = 1 m, H = 0.01 m, ṁ = 0.001 kg/s, Tin = 300.19 K, and fV = non-dimensionlize the energy equation (Eqn 32) paving the way to solve
0.0016). Also, Hazra and Nandi [213] investigated a DASC’s parametric it analytically. The dimensionless form of the energy equation was
behavior using a transient FDM accounting for top surface radiation and presented as:
convection heat losses, as well as bottom surface mirror reflection. Au­
thors formulated transient energy equations for individual nodes, ∂θ ∂2 θ
= + q̇ʹʹʹ(y) (91)
incorporating the mean
nf
Ke . After conducting numerous parametric ∂x ∂y2
nf
simulations, they pinpointed optimal mean Ke values associated with The boundary conditions were given in both the dimensional and
maximum energy storage efficiency under one sun (≡1000 W/m2) dimensionless forms, respectively, as:
incident radiation. Authors established correlations describing the ⃒ ⃒
∂T⃒⃒ ∂T⃒⃒ ( )
nf
optimal mean Ke as functions of fluid height and radiation time via T|x=0 = Tin ; = 0; k c = hE Ty=0 − Tamb (92)
∂y ⃒y=H ∂y ⃒y=0
regression analyses based on peak points. Moreover, for optimization
purposes, Hachicha and said [214] utilized the Response Surface ⃒ ⃒
∂θ⃒⃒ ∂θ ⃒ ( )
Methodology (RSM) (i.e., a statistical approach that optimizes processes θ|x=0 = 0; = 0; ⃒⃒ = NuE θy=0 − θamb (93)
∂y⃒y=1 ∂y y=0
by analyzing how various factors impact one or more response variables
using math and stats to model these relationships) to identify the Since Tin was considered as the ambient temperature (Tamb ), the
optimal fV for maximizing the DASC’s ηPC while minimizing hybrid dimensionless temperature at the inlet (θ|x=0 ) is zero. Authors calculated
nanofluid costs. Employing ANSYS 2021 R1 for Design of Experiment the optimum collector length at which the maximum photothermal ef­
(DOE) and optimization module, they executed a three-step optimiza­ ficiency (ηPC,max ), which is limited by the nanofluid allowable tempera­
tion process. Initially, DOE generated sampling points within input data ture, can be achieved. Authors used the proposed analytical model to
ranges, utilizing the central composite method (CCD) to acquire three determine the optimum efficiency and corresponding length of DASCs of
values for each input, resulting in nine design points. Subsequently, the different geometrical configurations, nanofluid combinations, Nusselt
RSM established relationships between response and independent input numbers, and dimensionless volumetric heat generation values, which
variables. Finally, authors employed the Multi-Objective Genetic Algo­ helps select convenient designs of efficient DASCs. The parameters G
rithm (MOGA), a variation of Non-dominated Sorted Genetic and NuE were introduced to control the optimum efficiency of the system
Algorithm-II (NSGA-II), to determine optimal solutions based on where their values account for multi design configurations and many
objective functions and constraints, yielding a Pareto frontier or operating parameter combinations, such as the collector height and
sequence of optimal solutions for maximizing DASC’s ηPC and mini­ length, flow velocity, solar concentration, etc. Finally, authors applied
mizing nanofluid costs. the model to a collector of 1 cm height utilizing a graphite-therminol
VP1 nanofluid and found that its predicted optimum efficiency was
2.4.2. Analytical studies 0.35 and the corresponding dimensionless length was 0.86. The math­
One of the earliest analytical efforts that introduced a comprehensive ematical formulation and solution steps can be found in their work
analysis of a volumetric absorption receiver is an analytical attempt appendices [136].
presented by Veeraragavan et al. [136]. Authors derived a reduced Despite the authors introduced a ground-work for the analytical
model using the eigenfunction expansion technique, excluding the light investigation of DASC systems, their model suffers, mainly, from some
scattering impacts through nanoparticles and assuming a non-absorbing limitations. First, the plug flow assumption limits the ability of the
host medium and fixed nanoparticle radiative characteristics. The model to capture the actual performance of the system, as detailed in the
non-absorbing host medium assumption is not reasonable since the host

25
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

previous subsection. Second, the Rayleigh approximation was used to

nf
obtain the Ka considering the optical characteristics of the nanofluid at
a fixed single wavelength (500 nm). Third, the crucial key parameters
that have significant and direct effects on the DASC’s ηPC were not all
revealed.
A more realistic analytical attempt was presented by Lee and Jang
[210] to determine ηPC and the temperature profile of a DASC. Authors
assumed a fully-developed flow between the DASC plates and consid­
nf
ered experimentally-measured Ke of the DI-water/MWCNT nanofluid
within a wavelength ranged from 200 nm to 2000 nm. The analytical
solution revealed the key parameters that affect the thermal perfor­
mance of the DASC such as fV , Pe, Nu, and aspect ratio (AR = L/ H). The
fully developed velocity field of the nanofluid flow was resolved to be as
follows:
[ ] [ ]
H2 dP ( y )2 y y ( y )2
u(y) = − = 6U − (94)
2μ dx H H H H

where μ, P, and U are the dynamic viscosity, pressure distribution, and

mean velocity, respectively. The authors used the same dimensionless
parameters and boundary conditions that were used by Veeraragavan Fig. 15. The effect of Nu on the DASC efficiency for plug and fully-developed
et al. [136] to state and solve the non-dimensionalized energy equation: flows [210]. (Reused with permission from Elsevier). Note: The number [12]
shown in the figure refers to Veeraragavan et al. [136] in the reference list.
[ ] ∂θ ∂2 θ
6 y − y2 = + q̇ʹʹ(y) (95)
∂x ∂y2 210] to present analytical solutions of the ηPC and temperature field of
The irradiance (Gs ) shown in the dimensionless temperature (θ) both a conventional SASC and a nanofluid-based DASC for comparison
equation, above, was given as follows: purposes. Also, they revealed the key parameters by which these two
∫∞ systems are significantly affected. The MWCNT-nanofluid DASC inves­
Gs = Io,λ dλ (96) tigation results were similar to the past paper’s [210] in terms of the
0 affecting key engineering parameters whose impacts on the behavior of
( )
the DASC’s ηPC were stated as follows: ηPC ∼ 1 /Nu, fV , Pe, 1 /AR . Be­
Io,λ is the spectral solar incidence intensity at (y = 0) given, assuming
nf
blackbody radiation, as follows: sides, authors conducted experiments to measure the Ke and investigate
the DASC thermal performance. Comparisons showed good agreement
1 Satt Ωs (2hc0 2 ) between the experimental and analytical results as illustrated in Fig. 16
Io,λ = [ ( ) ]
λ5 hc0 (97) that shows the temperature profile of the DASC for different fV , Pe, and
exp − 1
λkB Tsolar ṁ. Also, experiments showed that the tested nanofluid-based DASC can
achieve a 10 % higher ηPC than the conventional SASC, which is the same
where Satt and Ωs are the factor of average sunlight attenuation through conclusion obtained by Tyagi et al. [57] and Taylor et al. [131].
the Earth’s atmosphere ( = 0.73) and sun’s solid angle as viewed from Cregan and Myers [137] proposed an approximate analytical solu­
the Earth (= 6.8 × 10− 5 steradians), respectively. tion to a 2D steady model for the performance of an inclined
Solving the RTE resulted in the spectral solar radiation intensity low-to-medium temperature DASC utilizing water/Al nanofluid whose
(Iλ (y)):

Iλ (y) = CIo,λ e− Ke,λ y

(98)

The dimensional and non-dimensionalized distributions of heat

generation within the working fluid (q̇’’’ (y)) and (q̇ʹʹ(y)) were described,
respectively, as follows:
( ) ∫ ∫
d ∞ d ∞( )
q̇’’’ y = − Iλ (y)dλ = − CIo,λ e− Ke,λ y dλ (99)
dy 0 dy 0
∫
1 ∞ ( ) ( )
q̇’’’ (y) = Io,λ HKe,λ exp − HKe,λ y dλ (100)
Gs 0

The power series expansion was used to express the exponential term
in Eq. (100) and the quantity (HKe,λ ) was grouped in one parameter
called the optical thickness. The detailed derivation can be found in Lee
and Jang [210] paper appendices in order for the explicit heat genera­
tion profile, dimensionless temperature distribution and ηPC of the
investigated system to be analytically obtained. Authors stated that Nu,
Pe, fV , and AR are the key parameters that have direct effects on ηPC .
Fig. 15 illustrates how the collector’s ηPC decreases with the Nu, and how
the use of plug flow assumption overestimates the collector thermal
performance.
Thereafter, Lee et al. [215] utilized the same analytical modeling Fig. 16. Comparison between the analytical and experimental temperature
technique followed in the past two investigations discussed above [136, profile of the surface receiver (SR) and volumetric receiver (VR) at different Pe,
φ (≡ fV volume fraction), and ṁ [215]. (Reused with permission from Elsevier).

26
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

wavelength-dependent scattering and absorption were considered in the 3. Experimental approach for DASC investigations
model. The study presented a solution to the system of radiative trans­
port and energy equations stating some of the assumptions used in the This section dives into the experimental approaches researchers
study of Veeraragavan et al. [136] such as neglecting the nanofluid follow to study DASC systems. By examining these methods found in
radiative emission for temperatures less than 750 K. However, authors literature, their strengths and weaknesses can be considered, particu­
[137] determined the heat source term in the energy equation by larly when using experimental results to validate theoretical models.
numerically computing the spectral radiative heat flux integral and then The section is divided into two parts. The first explores how researchers
it was approximated as a power-law function. After the approximation use material characterization techniques to experimentally measure the
was used in the energy equation, the analytical solution became optical properties and extinction coefficients of nanofluids. The second
possible. Also, they considered wavelength-dependent nanofluid radia­ part delves into methods of DASC efficiency calculation, which typically
tive characteristics in the model making an advance progress in DASC involve using thermocouples to measure temperatures at various points
analytical modeling, where a wavelength-independent radiative char­ within the collector.
acteristics assumption was stated in past studies, or numerical methods
were followed to compute their wavelength-dependent values. The 3.1. Optical characterization of nanofluids/spectrophotometric
system of equations was non-dimensionalized using five governing
characterization
non-dimensional groups: two identifying the heat source, two identi­
fying the thermal losses to surroundings, and one identifying the relative
The nanofluid optical characteristics are a function of the nano­
contribution of advection and conduction. Results showed that the particle shape, size, and loading in addition to the optical characteristics
DASC’s ηPC depends mainly on fV despite there is a threshold beyond
of the host fluid and suspended nanoparticles. The refractive index (n) of
which there is no point to increase fV . Furthermore, the DASC perfor­ most fluids is known and tabulated for a wide range of wavelength
mance is enhanced with the collector depth, but does not experience a
including most of the solar spectrum. On the other hand, the absorptive
remarkable change with the nanoparticle size or collector inclination (also called extinction) index (κ) is not known for various fluids [216].
angle.
Since the complex refractive index (n + κi) cannot be directly measured,

Fig. 17. (a) Basic structure of a spectrophotometer, (b) Integrating sphere instrument schematics, and (c) three-slab system.

27
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

the transmittance, reflectance or angles of refraction may be needed to (λmin , λmax ), which is often considered at 1.5 air mass, and y is the
be measured first and then the experimental evaluation of such optical nanofluid film thickness.
constants becomes possible [73,216]. Certain equipment is needed to One of the earliest studies dedicated to optical properties measure­
conduct such measurements. ments is the experiment conducted by Otanicar et al. [216] to determine
the extinction indices of four fluids frequently used in solar systems
- Spectrophotometric characterization (therminol VP-1, propylene glycol, EG, and water). Experiments pre­
sented accurate results in the most spectral range and revealed that
Literature survey indicated that most experimental attempts to water is the strongest absorber among fluids studied although it is still a
optically characterize nanofluids in DASC studies were carried out based weak absorber (water absorbed only 13 % of the incident radiation).
on the nanofluid spectral transmittance via a spectrophotometer. The Authors conducted measurements utilizing a Shimadzu UV-3600
spectrophotometer is an instrument that indicates the amount of radi­ UV–Vis spectrophotometer of a working range from 170 to 3300 nm
ation absorbed within a substance by bombarding a sample by a beam of involving 98 % of the solar irradiation at air mass 1.5. The fluids were
light and measuring the intensity of the transmitted beam. Its basic placed in Spectrosil®(synthetic fused silica) cuvettes, i.e., straight sided
structure is illustrated in Fig. 17a. Modern spectrophotometers cover a optically transparent containers for holding liquid samples in a spec­
wide range of the electromagnetic spectrum (200 < λ < 2500 nm) [61, trophotometer or similar equipment, with pathlengths of 0.1, 2 and 10
217], and some types cover even wider ranges [216]. Spectrophotom­ mm. Authors decided to limit the studied wavelength range to
eters can be classified into two categories: the single-beam spectro­ 200–1500 nm because the infrared absorption bands in propylene gly­
photometer that determines a relative value of the light beam intensity col, EG, and water are very strong. Energy in this spectral range accounts
before and after passing through the sample, and the double-beam for around 85 % of the total solar energy [219]. The transmittance
spectrophotometer that emits two light beams, one to the test sample values, generated by the spectrophotometer for a certain path length
and the other to a reference sample and compare between them. cell, are a function of wavelength. An iterative technique to determine
Provided that particles seeded in a solar nanofluid are much smaller the optical constants of some organic fuels using Fourier Transfer
than the wavelength of the incident light, and the independent scat­ Infrared Spectroscopy (FTIR) was followed. A cell of a nearly-unity
nf
tering approximation is possible (α≪1), Ke can be calculated based on refractive index was used for authors to handle the cell as a single slab
the Beer-Lambert law [73]: eliminating any radiation absorbed or reflected from air to the window.
( ) On the other hand, neglecting the multiple reflections at interfaces of the
Isample ( ) − 4πκλ air-window-sample sequence used in Otanicar et al. [216] work is not
Tλ = = exp − Knf
e l = exp l (101)
Iref λ possible since they used quartz cuvettes whose real refractive index is
around 1.5 for most spectral range. Authors used a 3-slab system, as that
where Tλ , Isample , Iref , κλ , and l are spectral transmittance, transmitted shown in 7c, deepening the investigation by including multiple re­
light beam intensity through the sample nanofluid, light beam intensity flections, as discussed above.
through the reference, spectral extinction index, and path length of light Using a double-beam spectrophotometer (Perken Elmer Lambda
passage, respectively. 1050) for the entire wavelength range including solar radiation, Karami
Fluidic samples are placed in glass or fused quartz cuvettes. Typi­ et al. [220] measured the transmittance and other optical characteristics
nf
cally, Ka can be easily calculated when using a cuvette of 10 mm of a nanofluid containing CuO nanoparticles suspended in a (70:30 vol
pathlength [69]. Since out-scattering and absorption simultaneously %) water-EG mixture. Also, Lee and Jang [105] experimentally studied
nf
occur in heterogeneous systems, the Ke can be determined via a con­ nf
the Ke of a water/MWCNT nanofluid in a DASC. Initial measurements
ventional spectrophotometer. For very fine (nano-sized) distant disper­ were conducted using Lambert–Beer law at a fixed wavelength
sions of high absorptivity such as graphite and metallic nanoparticles, (632.8 nm). Three main parts were included in the measurement system;
nf
Ke can be identified with neglecting the scattering contribution since its a He–Ne laser with wavelength and power of 632.8 nm and 200 mW,
nf
effect is very low [218]. Consequently, the Ka with very low concen­ respectively; a sample fluid; and detecting units such as the photodiode
nf
tration and without aggregation can account for the extinction (i.e., and powermeter. For the sample fluid Ke measurement, the first step
nf
Ke ≈
nf
Ka ).But, if agglomeration occurs, large-size conglomerations of was to calculate the extinction index in the three-slab system as dis­
nanoparticles causes the scattering phenomenon to become too strong to cussed by Large et al. [221] that split the system into a set of discrete
be ignored any longer, and the scattering coefficient to grow with multi elements and computed the optical characteristics consecutively. A
orders of magnitude [69]. An integrating sphere is attached, whereupon, detailed formulation can be found in the literature [73,105,221].
between the sample cuvette and the detector, as shown in Fig. 17b, for Authors showed that incident radiation can be totally absorbed in 10
nf
the spectrophotometer to be capable of out-scattering effects detection. cm penetration depth using fV as low as 0.0005 vol% and that the Ke of
However, since the light beam experiences multiple reflection at in­ water/MWCNT nanofluid grows linearly with fV in the range from
terfaces between each consecutive elements of the air-window-sample 0.0005 to 0.005 vol%. Comparing the experimental results with theo­
nf
sequence, Ke determination requires additional calculations by retical models (Maxwell-Garnett for cylinder and the Rayleigh scattering
considering the filled cuvette as the three-slab system shown in Fig. 17c. approximation) showed that the conventional Maxwell-Garnett model
nf
Methods investigating the radiative transport in such a layered system that considers the shape effect is inaccurate for Ke predictions. On the
with the formulation to obtain the optical constants can be found in other hand, the Rayleigh scattering approximation that takes size impact
detail in the literature [68,87]. nf
into consideration can qualitatively anticipate the Ke of water/MWCNT
nf
After the Ke is identified, the nanofluid’s solar extinction and ab­ nanofluids. Furthermore, Lee and Jang [210] analytically studied the
sorption ability can be determined by finding the fraction of solar energy performance of a DASC utilizing DI-water/MWCNT nanofluid with
stored in the nanofluid (F(y)) after a y pathlength through the nanofluid. concentration ranged from 0.0005 to 0.005 % taking into account
F(y) is given as [216,218]: nf
experimentally-measured Ke by the UV/Vis/NIR spectrophotometer
∫ λmax ( ) (Cary 5000, Varian, Inc.) according to 200–2000 nm wavelength range.
I(λ)exp − Knf e y dλ
F(y) = 1 −
λmin
∫ λmax (102) The dual beam measurement technique, as detailed in the study [218],
λmin
I(λ)dλ nf
was followed and then the Ke was obtained using the Beer-Lambert law
where air was used as a reference [210]. Based on Eq. (101), the spectral
where I(λ) is the solar radiation intensity within the wavelength range nf
Ke was plotted, as shown in Fig. 18a, demonstrating that nanofluids can

28
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

nf
Fig. 18. The spectral (a) Ke at different volume fractions (φ ≡ fV ), and (b) solar intensity and albedo of the DI-water-MWCNT nanofluid [210]. (Reused with
permission from Elsevier).

absorb the incident radiation in the relatively short wavelength range

(visible and infrared), but pure DI-water cannot. However, in the region
where the wavelength is greater than 1400 nm, the nanofluids and pure
DI-water extinction coefficients are nearly congruent, which means that
nanofluids absorbed the radiation in the wavelength ranged from 200 to
1400 nm. Similar results were reported in Refs. [87,111,117,218].
nf
For accurate anticipation of the radiation absorbed in a DASC, the Ka
nf nf
should be measured instead of the Ke . However, the Ke can be
nf
considered as Ka when the scattering contribution to extinction is
insignificant, as stated earlier. To assess that, the albedo, which is the
fraction of light a surface reflects (the ratio of scattering to extinction)
should be obtained [117]. Despite Lee and Jang [210] found that the
albedo was of a relatively great value (~20 %) only in the short wave­
length range (200–300 nm), the incident radiation intensity in that
range was weak as well (see Fig. 18b). Also, they observed that the al­
bedo significantly decreases with the wavelength, which points out that
the scattering effect is insignificant in case of the studied Fig. 19. Experimental and numerically-predicted extinction coefficients of
DI-water/MWCNT nanofluid, and most extinct radiation was absorbed different nanofluids [63]. (Reused with permission from Elsevier).
by the nanofluid. Similar results were reported by Mercatelli et al. [117]
that experimentally proved that the albedo of a single-wall carbon Experimental Knf e values were obtained based on the Beer-Lambert
nanohorns seeded in water was 5 %. Consequently, evaluating the heat principle, and theoretical ones were obtained based on the Rayleigh
generation without considering the scattering assuming that Ke ≈ Ka
nf nf scattering model. Discrepancies between experimental and simulation
can be safely conducted. results were observed in their work (see Fig. 19). For a silver nanofluid,
Experiments enable researchers to correlate the performance in­ the model overestimates the Knfe before 500 nm and underestimates after
dicators of the investigated systems with the operating variables of in­ 500 nm. This observation supports Taylor et al. [87] assertions, dis­
terest. Lee et al. [222] experimentally introduced a correlation to cussed earlier, about the Rayleigh theory. Metallic nanoparticle surface
nf
measure the Ke of a nanofluid combining DI-water seeded by MWCNT oxidation can explain that metal nanoparticles have different optical
with a surfactant. Transmission electron microscopy (TEM) and scan­ characteristics than their bulk material. Also, nanoparticle agglomera­
ning electron microscopy (SEM) were used to characterize nano­ tion deforms their spherical shape and enlarges their size over the initial
nf (rated) one used in calculations, which leads to inconsistencies between
particles. The Ke measurements were conducted based on the
the experimental results and Rayleigh approximation predictions. Such
Beer-Lambert principle utilizing an in-house apparatus at a single
nf
experimental uncertainties should be considered in model validations.
wavelength (632.8 nm). The correlation for Ke within the fV range
(5–50 ppm) was presented as a function of fV as follows: 3.2. Thermal performance of DASCs
Knf
e = 7610.374fV + 0.01
(103)
This subsection discusses the experimental attempts that investigate
Also, experimental research is useful to validate theoretical optical the thermal performance of DASC systems. These studies often use
models and assess their assertions. Gorji and Ranjbar [63] presented an thermocouples to measure the nanofluid temperatures at different po­
experimental part in their research work to measure the optical prop­ sitions on the DASC. Such experiments are primarily conducted to
erties of three nanofluids composed of distilled water and graphite, validate theoretical models developed to investigate DASCs. Otanicar
magnetite (Fe3O4), and silver nanoparticles working in a planar DASC. et al. [34] examined the impacts of three nanoparticle material types
Experiments were conducted to validate presented model predictions. (carbon-nanotubes (CNTs), silver (Ag), and graphite) on the

29
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

performance of a micro-DASC. Results were used to validate a mathe­ the ηPC asymptotically increased with fV and ṁ, and that the maximum
matical model similar to that of Tyagi et al. [57]. Experimental data zero-loss ηPC occurred at fV and ṁ of 100 ppm and 0.025 kg/s, respec­
were obtained using three T-type thermocouples mounted on the DASC tively, and it was 17 % higher than that with the pure base fluid at the
to measure the temperatures at three different positions (inlet, midpoint same ṁ. Based on experimental observations, authors suggested a cor­
of bottom wall, and outlet). Data from both experimental and numerical relation for the studied system’s ηPC function of the fV ranged from 0 to
parts showed good agreement and similar behavior where both 100 ppm and volume flow rate (V̇) ranged from 54 to 90 lit/h, in
demonstrated that as fV increases, ηPC of the DASC rapidly climbs addition to the inlet (Tin ) and ambient (Tamb ) temperatures and solar
initially and then it levels off after fV reaches 0.5 %, as illustrated in irradiation (Gs ):
Fig. 20. ( )0.0565
Also, Lenert and Wang [28] obtained temperature profiles of ther­ ηPC = (0.281V̇ + 35.6) fV + 1
( )
minol VP-1 based carbon-coated cobalt nanofluid of 28 nm nanoparticle ( )− 0.0186 Tin − Tamb
size and different coating thicknesses to measure the optimum one. − (− 0.01V̇ + 20.66) fV + 1 (104)
Gs
Experimental and numerical results showed an agreement with a per­
centage error of 20 %. Authors justified this discrepancy by the The OTPs of the proposed nanofluid was introduced in another
remarkable temperature difference between the bottom and top walls, research work [220] that was discussed in the previous subsection.
which stimulates heat redistribution through conductive side walls but Delfani et al. [59] replicated the same procedure and investigations with
cannot be captured by the numerical model. Taylor et al. [87] experi­ a different nanofluid. They experimentally revealed the OTPs of
mentally and numerically characterized the optical properties of carboxyl functionalized MWCNT blended in a water-EG (70 %:30 %)
different nanofluids composed of water or therminol VP1 as a base fluid mixture working as a nanofluid in a domestic-applications DASC pro­
and seeded by graphite, silver, copper, gold, or aluminum nanoparticles totype and examined the DASC thermal performance based on EN
to examine their working capabilities in DASCs. They argued that 12975-2 standard at various V̇ (54–90 l/h), fV (0–100), and two different
around 95 % of incident radiation can be absorbed (in a nanofluid bottom walls (black and reflective). ηPC increased asymptotically with V̇
thickness ⩾10 cm) with fV values as small as 10 ppm, which means also and fV . Zero-loss ηPC,max at ṁ and fV of 0.025 kg/s and 100 ppm,
that the use of solar nanofluids with such trace particle loading has a respectively, was 29 % higher than that of the pure base fluid at the same
negligible effect on pumping power. Authors observed that experimental V̇, which means that MWCNT-based nanofluid outperformed the
and numerical results showed good endorsement in revealing optical CuO-based one in this regard.
characteristics of graphite and aluminum nanofluids. However, In the previously discussed study, Karami et al. [220] studied the
model-predicted results were inaccurate for the other metals investi­ thermal conductivity and viscosity of the CuO/water-EG nanofluid.
gated, particularly falling short in capturing the significant anticipated Based on the experimental data obtained, they presented correlations for
plasmon peaks (e.g., silver). The disparity between the numerical and thermal conductivity (knf ) and viscosity (μnf ) of the nanofluid function of
experimental results can primarily be ascribed to impurities in the ma­ the operating temperature (T) and fV in the same study ranges:
terials; for instance, oxidation or other contaminants on the nanoparticle (20–60 ◦ C) and (0–100 ppm), respectively:
surface may result in outcomes that weakly align with the theoretical
model. knf ( )0.05
= (0.0018T + 0.36) fV + 1 (105)
To exceed the laboratory scale experiments, Karami et al. [60] kbf
examined CuO/water-EG (70:30 vol%), for the first time in the litera­
μnf ( )
ture, in a full-scale DASC for domestic applications. EN12975-2 standard = 0.06fV + 30 T− 0.85
(106)
procedure was followed to build the setup and investigate the thermal μbf
performance of the collector with respect to fV , ṁ, and two different This full-scale experiment was followed by Gupta et al. [223] who
bottom walls (black and reflective). The use of the nanofluid raised the experimentally tested a full-scale DASC of 1.4 m2 gross area employing
DASC’s ηPC by up to 17 % compared to the pure carrier fluid. Also, tests water/Al2O3 nanofluid of (0.001–0.05 %) concentration and 20 nm
showed that the collector with a black bottom wall was 11.4 % more nanoparticle size. Experiments showed that collector’s ηPC peaked at
efficient than that with a reflective one. Also, experiments showed that some fV values. For instance, it reached 39.6 % at fV of 0.005 %, and then

Fig. 20. Efficiency of solar collectors experimentally tested vs. volume fraction for water as a host liquid and different nanoparticle material types. The conventional
SASC is named as ’water (black back)’ [34]. (Reused with permission from AIP Publishing).

30
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

lowered at less or greater fV values. [226]. Additionally, nanoparticles may even be engineered and
Several researchers stated that experimental uncertainties might be designed to possess properties that are not available in their material
responsible for disparities between experimental results and model bulk counterpart [227], which can lead to propitious and desirable OTPs
predictions in DASC studies. For example, Luo et al. [58] experimentally for different engineering systems [40,228]. The unique properties of
verified a model predictions obtained by solving radiation, conduction nanoparticles in the 1–100 nm range stem from two main factors. The
and convection heat transfer equations to report on efficiency and first is the high ratio of surface atoms to interior bulk atoms [77]. For
temperature distribution in a DASC. Experimental results fairly example, a 3.5-nm gold nanoparticle has approximately 40 % of its
endorsed the model predictions. Authors explained why predictions atoms located on the surface [229]. These surface atoms are less stable
were 3–9% lower than experimental values by the fact that the fluid at compared to their bulk counterparts due to fewer bonds. The second
the inlet was heated by radiation, but it was set 300 K in the simulation. factor influencing nanoparticle properties is the presence of quantum
Moreover, since thermocouples are under solar radiation, they can be effects in semiconductors and metals. These quantum effects can man­
hotter than the fluid there. Furthermore, Struchalin et al. [224] exper­ ifest as unique phenomena like the LSPR observed in metallic nano­
imentally and numerically examined the performance of a tubular particles [77]. LSPR signifies a special way in which light interacts with
DASC. Comparisons between the CFD model predictions and experi­ these ultrafine particles.
mental results showed discrepancies of up to 22 % in some cases as the The literature outlines two ways for nanoparticle synthesis: bottom-
model underpredicts the collector performance. Authors attributed the up and top-down. The bottom-up way is constructive and builds nano­
discrepancies to the asymmetric distribution of radiation from the solar materials from smaller parts of their counterparts such as molecules or
simulator that was not represented by the model, and to the mutual atoms [230]. The formation of the nanomaterials via this method occurs
radiation reflection from the tubes. through a set of chemical techniques and reactions between some
reacting agents and the material precursors that lead to self-assembly of
3.3. Limitations of the presently employed experimental methods atoms and molecules creating nanoparticles of regular shape and
well-defined size. The top-down way is destructive and yields a trans­
The surface oxidation of nanoparticles leads to disparities between formation of the bulk material into nano-sized particles by crushing
the OTPs of nanoparticles and those of their bulk counterparts. Addi­ material particles to drop from a larger size (top) to smaller one (down)
tionally, nanoparticle agglomeration can distort their spherical shape via several methods such as grinding, cutting, and ball milling. This
and increase their size beyond the initially rated dimensions used in method results in extremely small and irregularly-shaped nanoparticles.
theoretical calculations, resulting in inconsistencies between experi­ The suspension of nanoparticles in a base fluid creates a nanofluid/
mental observations and theoretical predictions. The discrepancy arises colloid/suspension. The selection of nanoparticles and base liquids for
from the agglomeration of nanoparticles within the nanofluid, leading to nanofluid combination should take into consideration the components’
an actual size different from the manufacturer-provided rated size. OTPs such as the thermal conductivity (kc ), heat capacity (cp ), viscosity
Direct detection methods, such as DLS, SEM and TEM, or similar tech­ (μ), freezing and boiling temperatures, thermal stability, tendency to
niques, directly determine the evolution of particle size distribution over agglomeration, and light extinction ability. Also, the base liquid should
time to assess nanoparticle agglomeration. However, the direct detec­ be selected such that it (1) has a non-reactive and non-corrosive nature
tion approach has operational challenges. It requires a significant to not cause degradation of nanoparticles, (2) has a low viscosity for low
quantity of samples, with some DLS instruments demanding large vol­ pumping costs, and (3) does not boost agglomeration [36]. The
umes, making it challenging for nanofluids prepared using one-step enhanced thermophysical characteristics of nanofluids can be attributed
methods with limited samples. Moreover, in cases of concentrated sus­ to several mechanisms including the liquid layering (mainly responsible
pensions, the rapid aggregation process may outpace the measurement for the nanofluid’s increased thermal conductivity), ballistic transport
time, making it difficult for direct detection methods to capture this (i.e., unimpeded flow of energy-carrying particles over relatively long
dynamic evolution [86]. distances within a suspension), Brownian motion (i.e. stochastic motion
Additionally, the exposure of thermocouples, used for temperature of nanoparticles due to the random motion of base fluid molecules),
measurements at different positions of the DASC system, to sunlight may clustering, and thermophoresis (i.e. temperature-difference-induced
lead to artificially higher temperature readings than the actual tem­ motion) [231]. Such mechanisms act on the molecular level of nano­
peratures of the nanofluid flow. Such experimental techniques’ limita­ fluids and are induced by interactions between the gravitational and
tions and uncertainties lead to a lot of discrepancies between the model- other forces exerting influence on nanoparticles. The Brownian motion
predicted performance of DASCs and the experimental results, including and thermophoresis have the main role in increasing the heat conduc­
the OTPs of nanofluids. tion within the nanofluid since such phenomena cause heat to dissipate
As concluded by Mahian et al. [163], in validation of computational over the nanofluid via micro convection currents [232]. Also, heat
models by comparing their results with such experimental results sur­ dissipation impacts are sharply amplified with the bulk fluid motion
veyed above, it is imperative to account for experimental uncertainties when nanoparticles are there due to their increased chaotic motion
arising from various factors including nanoparticle sedimentation and whereupon. Evidence on this fact can be inferred from the literature by
aggregation phenomena, and potential uncertainties and measurement observing that the increment in heat transfer accompanied by increasing
errors. Also, there are several inconsistencies between the experimental Re of a flow is higher in nanofluid flows than in conventional ones [233].
results reported in the literature and the predictions of mathematical Moreover, research has provided explanations for the higher values of
models. This indicates a dire need for more work on this area. Moreover, heat transfer coefficients enhanced by nanofluids in both free and forced
a number of parameters playing a pivotal role in the enhancement convection regimes compared to pure base fluids [234]. Authors
margin with nanofluids such as nanoparticle size, material type, volume attributed this behavior to the higher thermal conductivity due to the
fraction, and pH of the nanofluid, need to be experimentally examined presence of nanoparticles which, additionally, cause the thermal
and more thoroughly grasped [64]. boundary layer along the channel to severely disturb [235].
Besides thermal characteristics such as the kc and cp , the presence of
4. Overview of nanofluid stability nanoparticles, even with a very tiny quantity, causes significant modi­
fications on the suspension’s transport and, most crucially, optical
Nanofluids have been employed in several engineering applications characteristics, leading to considerable enhancements in the perfor­
since they have extraordinary characteristics due to the presence of mance of solar thermal systems employing nanofluids [65]. For
nanoparticles. The immense specific surface area of nanoparticles [225] instance, Zhang et al. [76] achieved 20 % enhancement in the ηPC of a
renders their characteristics governed by the surface instead of the bulk solar thermal system using 0.15 ppm of gold nanoparticles compared to

31
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

the pure base fluid. Also, Filho et al. [236] increased the stored thermal rounding medium is less dense. The Stokes drag equation, incorporating
energy of a DASC by up to 144 % using only 6.5 ppm of silver nano­ the gravitational force, can be employed to estimate the sedimentation
particles. Furthermore, Wen et al. [237] increased the photothermal velocity (vs ), as follows [239]:
conversion efficiency of a DASC by 67.1 % by less than 50 ppm of TiN
2 Δρga2
nanoparticles suspended in EG. Such enhancements are attributed to the vs = (107)
9 μbf
significant enhancements in optical characteristics of working fluids by
adding nanoparticles.
where Δρ is the difference between particle and base fluid densities, a is
Particles suspended in a fluidic medium suffer from the gravitational
the particle radius, g is the gravitational acceleration, and μbf is the base
effects that drive particles to cluster and settle at the bottom of the
channel. However, when particles are nano-sized, they can stay afloat fluid viscosity. It can be observed that larger particles show a greater
and resistant to the possible deposition at the bottom. Still, other forces tendency to sedimentation. It is reported that barium sulfate micropar­
that affect nanoparticles such as the van der Waals force, even at very ticles (1.3 μm) settle completely within 8 min, while sub-microparticles
low concentrations, can put them at risk of agglomeration that conse­ (420 nm) require approximately 100 min, and nanoparticles (70 nm)
quently leads to sedimentation and deteriorates the nanofluid perfor­ take more than 60 h [239]. It is worth noting that aggregates will
mance. Some kinds of suspensions are stable, but others have naturally settle at an even faster rate. Moreover, the settling velocity of
nanoparticles with tendency to agglomerate. Such nanofluids can be nanoparticles depends on their shape because shape influences the co­
treated by adding surface acting agents, called surfactants (surface- efficient of drag they experience in the fluid. Rectangular nanoparticles
active agents) or dispersants, to keep them from agglomeration. The of high aspect ratio show higher tendency to agglomeration. Kim et al.
surfactant gets attached to the nanoparticle external wall and prevents [242] examined the stability of water-based bohemite alumina nano­
any potential adhesion to other surfaces. However, the nanofluid’s fluids and stated that brick-shaped nanoparticles recorded high stability
prolonged exposure to intense solar flux or elevated temperatures leads compared to blade- and platelet-shaped nanoparticles.
to its destabilization, which weakens its ability to extinct solar radiation A description of the Brownian motion can be obtained from the
and thus diminishes the ηPC of the solar system [77]. Stability is one of diffusion coefficient of particles (D) [239] as given by the
the most important features that should characterize a solar nanofluid Stokes-Einstein equation:
[66]. Studies, therefore, suggested many dispersants/surfactants to kB Tbf
enhance stability and revealed their role in enhancing the absorption D= (108)
6πμbf a
capacity of the nanofluid [238].
Understanding the destabilization mechanisms and their impact on As the diffusion coefficient depends mainly on the base fluid vis­
solar nanofluids is imperative for devising resilient photothermal con­ cosity and particle size, motion of particles, of less than 10 nm in size, is
version systems, highlighting the pivotal role of stability in sustaining mainly governed by diffusion [243].
desired optical and thermal properties throughout the operational life­ Shear forces can lead to changes in colloid properties, although these
span. The extremely small size of nanoparticles makes their interactions deviations are often not directly observable at the particle level but
within the liquid they are surrounded by, such as the attraction between rather in the overall system’s macroscopic properties. In dilute suspen­
opposite charges, Brownian motion, and viscous resistance they face sions, the nanofluid’s shear viscosity is influenced by fV given by the
from the liquid, highly intricate. A commonly used theory to understand Brinkman equation (Eq. 39) or by the Einstein equation as follows:
the stability of these suspensions is the Deryagin-Landau-Verwey- ( )
μnf = μbf 1 + 2.5fV (109)
Overbeek (DLVO) theory [239]. This theory is based on several as­
sumptions. The suspension is dilute, and the suspended particles are Another factor affecting the nanofluid stability is the zeta potential,
assumed to be influenced solely by van der Waals attraction potential that signifies the magnitude of the electrokinetic potential at the slip­
(GA ) and electrostatic repulsion potential (GR ). The effects of gravity and ping plane within colloidal dispersions. This slipping plane represents
buoyancy are not considered, and the colloidal particles are uniformly the boundary between mobile bulk fluid and the stagnant fluid layer
dispersed. Finally, the ion distribution in the colloidal medium is adhering to the particle surface. The charged species present on the
determined by electrostatic interactions, Brownian motion, and disper­ particle’s surface are responsible for the emergence of zeta potential. It
sion caused by entropy. The theory proposes that two main forces con­ quantifies the extent of electrostatic repulsion between similarly
trol how nanoparticles behave: attraction (van der Waals forces) and charged particles in close proximity within the dispersion. Particles
repulsion (electrostatic forces). When nanoparticles approach each possessing a high absolute zeta potential (positive or negative) experi­
other, their energy increases until it reaches a peak (several times kB T), ence electrical stabilization, hindering aggregation. Conversely, a low
and if they become even closer, the energy rapidly decreases [239]. For zeta potential indicates that attractive forces can overpower repulsion,
instance, if gold particles (of 100 nm in radius and 10 nm distanced from leading to coagulation or flocculation of the colloidal dispersion. A zeta
each other) are suspended in water, they will exhibit a notable van der potential above ±30 mV signifies a stable colloidal suspension.
Waals interaction potential of approximately − 24kB T [240]. These Conversely, a value within ±15 mV indicates instability. Essentially,
changes in energy indicate a shift from repulsive to attractive forces, zeta potential is a critical indicator of colloidal system stability,
respectively. In a suspension, Brownian motion of particles can lead to reflecting the balance between electrostatic forces that keep particles
collisions. These collisions reduce the distance between particles, separated and attractive forces that lead to particle aggregation. How­
allowing the attractive van der Waals forces to take over and bind them ever, Chakraborti and Panigrahi et al. [244] reported that even with a
together. This process of clustering is termed Diffusion-Limited Cluster high zeta potential, sedimentation can occur if the particle and sur­
Aggregation (DLCA) [239,241]. Another way nanoparticles cluster rounding fluid have a substantial density difference. As for the effect of
together is known as Reaction-Limited Cluster Aggregation (RLCA) the medium temperature, researchers stated that zeta potential of
[241]. The main difference is that in DLCA, any collision results in several water-based nanofluids witnessed a dramatic decrease when
clustering, whereas in RLCA, nanoparticles need to collide and react to increasing the nanofluid temperature [245,246].
form clusters. In nanofluids, DLCA is generally considered the more If we return to the DLVO theory, we find that the stability of colloids
common process. relies on the total potential energy (GT ), formulated as:
In dilute suspensions such as solar nanofluids, where particles
GT = GA + GR (110)
behave individually as independent objects, various forces influence
their behavior. Notably, the gravitational field of the Earth exerts a The van der Waals attraction energy (GA ) between two particles at a
downward force on the particles, causing them to settle when the sur­

32
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

distance Lp− p from each other, which plays an important role in the potential (weak repulsion) and high particle loading can lead to GA
tendency of nanoparticles to agglomeration as discussed earlier, was dominance at a short Lp− p . This can create a secondary minimum in the
given earlier in Eq. (73). For engineering-level systems with a substan­ potential energy profile (Fig. 21a), leading to the formation of weakly
tial number of nanoparticles, it is critical to precisely determine the van bound clusters. These clusters can be frittered by external forces like
der Waals forces and the effective Hamaker constant in nanofluids. This stirring or ultrasonication (i.e., passing ultrasonic waves through a
requires suitable modeling and simulation methods. Given that numer­ nanofluid, causing vibration in the agglomerated nanoparticles and
ical methods are too time-intensive for our goals; they are not an option. leading to their fragmentation), restoring suspension stability.
The focus should be on identifying the most appropriate approximate A high zeta potential indicates a sufficient repulsive force, which
analytical formulas to assess van der Waals forces and the effective minimizes particle aggregation. Without this, undesirable flocculation
Hamaker constant for colloidal nanoparticles in solar nanofluids. (clustering of nanoparticles due to their small size and surface charges)
Although there is a vast selection of formulas for these calculations be­ can occur, compromising stability. Therefore, achieving and maintain­
tween spherical particles, it is imperative to choose formulas that are ing a high repulsive potential is crucial for ensuring long-term stability
accurate for colloidal nanoparticles in nanofluids. Readers are encour­ of nanofluid suspensions. Fig. 21b illustrates the positive impact of a
aged to consult [77] for an overview of various methods available in the high energy barrier on nanofluid stability.
literature for computing van der Waals interactions and the Hamaker The pH value of the suspension is closely linked to its zeta potential,
constant. This work provides a comparison and critical assessment of as detailed in Ref. [247]. The pH value affects the electrical charge
these methods as presented in the literature. density on the nanoparticle surface, thereby impacting nanofluid sta­
The electrostatic repulsion potential (GR ) between identical spherical bility. Additionally, the pH value influences the size and shape of
particles is given as follows [244]: nanoparticles during their preparation, further affecting stability. Li
et al. [248] examined how pH levels affect synthesized Ag nanoparticles’
GR = 2πεbf εv aς2 e− ΚLp− p
(111)
size and shape, of which stability are a strong function. In a basic
environment (pH = 11), Ag nanoparticles tend to be spherical with
where εbf , εv , ς, and Κ are the dielectric function of the base fluid,
uniform size distribution, whereas in a neutral environment, they vary
permittivity of vacuum, particle surface potential (electrical potential
in shape from spheres to rods with less control over size distribution. As
due to the uneven distribution of particles), and Debye constant,
GA , GR and thus GT are a function of the particle size, it is important to
respectively. Κ− 1 is known as the Debye length (measured in nm). For
highlight the influence of particle size on colloidal stability. In a study by
simple GR calculations, the surface potential is assumed to be same as
He et al. [249], they found that when dealing with α-Fe3O4 particles at
the zeta potential, which quantifies the magnitude of the electrostatic
the same pH, the aggregation tendency increases as the particle size
repulsion between a particle and the surrounding liquid. This equation,
decreases. This is attributed to the increased atom density on the surface
which is valid when Κa < 5 and ς ≤ 25 mV, is common for nano­
of smaller particles, which causes elevated active site density. Also,
particles. Readers are encouraged to look into different relations for GR ,
calculations offered in Ref. [244] proved that increasing the particle size
for different particle shapes and configurations under different bound­
(Dp = 12, 32, and 65 nm with corresponding zeta potential of 42.6, 40.2,
ary conditions, which are compiled and categorized in Ref. [244].
and 44.1 mV, respectively) of α-Fe3O3 leads to an increase in GR and GT
As GT and both GA and GR are a function of the separation distance
and a margin decrease in GA . Furthermore, although DLVO theory fails
between particles (Lp− p ), the colloid stability hinges on the interaction
to assess the stability of non-spherical (anisometric) particles, re­
between them. Brownian motion continuously alters Lp− p , leading to
searchers state that the particle shape has a significant effect on the
fluctuations in GT . Fig. 21 introduces an illustrative representation of the suspension stability. When Lp− p is less than the average diameter of
potential energy of interaction versus Lp− p with a potential barrier of two
(non-spherical) particles, their increased surface allows for more atoms
levels: (a) low, and (b) high. By examining the figure, one can formulate to interact causing attractive forces to overpower repulsive forces. As a
some observations: If GA dominates over GR at a particular distance, consequence, these anisometric nanoparticles are more prone to ag­
particles tend to aggregate, forming large clusters and destabilizing the gregation compared to their spherical counterparts [250].
suspension (Fig. 21a). This scenario can lead to faster sedimentation. In One more parameter affecting the colloid stability is the dielectric
contrast, a high total interaction potential creates a significant energy constant of the base fluid. It directly influences the repulsive potential
barrier between particles (Fig. 21b). This barrier hinders agglomeration, (GR ), implying that higher dielectric constants lead to greater stability.
promoting suspension stability. However, even with this barrier, high- Water exhibits the highest dielectric constant over other base fluids
energy collisions (e.g., at elevated temperatures) can overcome it. typically used in nanofluid-based solar thermal systems like EG, ben­
When the barrier is breached, attraction overpowers repulsion, pulling zene, ethanol, etc., indicating its strong repulsive potential. Addition­
particles together to form irreversible, strongly bound clusters. Low zeta ally, nanoparticle loading has a crucial effect on nanofluid stability as its

Fig. 21. Illustrative representation of the potential energy of interaction versus Lp− p with a potential barrier of two levels: (a) low, and (b) high [244]. (Reused with
permission from Elsevier).

33
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

decreases the distance between particles (Lp− p ), which has a significant independence, lower prices, limited impact on hydrodynamic particle
effect on both GA and GR , resulting in a nanofluid with poor stability size, and simplicity and controllability. A detailed comparison between
[251]. the steric stabilization and electrostatic stabilization techniques is
The literature offers a suitable method, known as Stability factor offered in literature [77]. Finally, electrosteric stabilization combines
(SF), to assess the stability of a solar nanofluid based on alterations in its electrostatic and steric repulsions to offset the attractive van der Waals
optical properties. The SF reflects the change in radiation transmission forces between particles. When two particles coated with charged spe­
through a container holding the nanofluid (such as a cuvette with a cies of high molecular weight approach each other, the overlap of their
transmissive bottom surface) over time. SF can be given as follows molecular coatings induces both steric and electrostatic repulsions. In
[114]: the former case, the polymer chains experience restricted orientational
freedom, while in the latter, similar charged species become crowded.
Iinc − Itrans (t)
SF = (112) Electrosteric stabilization can be achieved through physical and chem­
Iinc − Itrans (t = 0)
ical adsorption of charged macromolecules (such as ionic surfactants,
polyelectrolytes, and polyampholytes), non-ionic surfactants, or a
where Iinc and Itrans (t) are the incident and temporal transmitted spectral
combination of these methods [254]. Electrosteric stabilization,
radiative heat fluxes, respectively. A value of SF closer to zero indicates
although effective, has drawbacks beyond those of its individual com­
the presence of colloidal or chemical instabilities, whereas a value closer
ponents (electrostatic and steric). When particles have a weak surface
to one signifies stability. In a study by Choi et al. [114], the SF of
charge and the stabilizing molecules (polyelectrolytes) have a strong
MWCNTs dispersed in water using various surfactants (Sodium dodecyl
opposite charge, patchy interactions arise. This is because the poly­
benzene sulfonate (SDBS), Sodium dodecylsulfate (SDS), Cetyl­
electrolytes may not fully coat the particle surface. These patchy areas
trimethylammonium bromide (CTAB), and TX-100) were evaluated. The
with the original particle charge and the adsorbed polyelectrolyte
analysis considers two scenarios: long-term storage at room tempera­
patches with opposite charge can attract each other when particles are
ture, and heating to 85 ◦ C for a short duration. A slight decline in SF of
close to each other [255]. This attraction between oppositely charged
nanofluids over time during storage was observed and attributed to
patches can destabilize the entire suspension and cause the particles to
physical instability, such as sedimentation of particles. Conversely, a
aggregate.
sharp decrease in SF of the nanofluid with the TX-100 surfactant was
Exposure to solar radiation presents the most critical destabilization
observed upon heating, and explained by the chemical instability,
factor for solar nanofluids, triggering a multitude of detrimental
potentially due to surfactant degradation. Furthermore, the authors re­
chemical and colloidal instabilities [256]. A key challenge in this field
ported the solar-weighted absorption coefficients of the nanofluid with
lies in developing methods to shield these fluids from prolonged ultra­
various surfactants and stability factors. They concluded that a solar
violet radiation without compromising the advantageous optical prop­
nanofluid’s dispersion stability directly influences its optical properties
erties of the nanoparticles themselves.
and ability to absorb solar radiation.
Responsive treatment methods treat clustered particles by applying
Due to the significant interfacial area between their two phases,
energy to break up these clusters using fluid agitation and high stresses
nanofluids often face stability issues, prompting researchers to propose
[257]. However, the effectiveness depends on the cluster’s strength.
several stabilization strategies. These strategies can be preventive
Common techniques include ultrasonication, which uses sound waves
(protection before agglomeration happens) and responsive (breaking
for fragmenting soft clusters. Ball milling utilizes grinding media for
already agglomerated particles away). Preventive techniques are foun­
impact-based cluster breakdown, with a variation employing surfactants
ded on creating particles with strong repulsive potential that can sup­
to enhance the process, but this method might affect particle size and
press the van der Waals attraction and prevent particle agglomeration.
shape for particles below 500 nm [258]. High-pressure shear homoge­
The literature presents a bundle of these preventive techniques such as
nization is a powerful technique that forces the nanofluid through a
the electrostatic (charge-induced) stabilization, steric (polymer-
narrow gap at high pressure, effectively breaking down all cluster sizes,
induced) stabilization, and electrosteric stabilization techniques.
minimizing reformation, and producing pure suspensions, but the
Electrostatic stabilization involves charging particles, leading to
equipment is expensive, heavy, and susceptible to wear [259].
electrostatic repulsion when their electric double layers overlap [252].
Finally, it is important to make a number of comments on the
This method works well in base fluids with high dielectric constants.
nanofluid stability studies as shown in literature. In contrast to pre­
Changes in ionic species concentration in the dispersing medium affect
vailing assumptions within nanofluid stability literature, a high zeta
the degree of electrostatic repulsion, as the electric double layer either
potential does not invariably signify exceptional colloidal stability. This
thins or thickens. A stronger surface charge results in a higher energy
is demonstrably true for sterically stabilized dispersions, where a low
barrier, preventing particle contact and cluster formation. As for steric
zeta potential may be observed despite complete dispersion stability
stabilization, it involves using nonionic surfactants or polymers to create
[77]. Conversely, for electrostatically stabilized dispersions, alterations
a strong repulsion between particles’ cores in a dispersion [77]. When
in ionic strength due to factors like contamination or oxidation can
these surfactants or polymers adsorb onto the surface of particles or
induce variations in the electrical double layer thickness. Consequently,
droplets, they form an adsorbed layer with a thickness (denoted as δ).
the zeta potential value required for optimal colloidal stability also
This layer can be well-solvated (or hydrated) by the solvent molecules
undergoes modification. Therefore, a more precise interpretation of zeta
when the solvent is compatible with the surfactant or polymer layer. As
potential limits it to an indicator of electrostatic repulsion forces be­
particles or droplets approach each other within a separation distance
tween colloidal nanoparticles.
smaller than 2δ, the adsorbed layers can overlap or compress, resulting
Furthermore, the DLVO theory, while valuable, has limitations in
in two main effects: osmotic repulsion due to unfavorable mixing of the
capturing the complexities of nanofluid behavior. A multi-scale
adsorbed layers in good solvent conditions, and entropic or elastic
approach bridges this gap by combining the strengths of the, discussed
repulsion caused by the reduction of configurational entropy as the
earlier, atomistic simulations (MD) and system-level simulations (DPD,
layers overlap. The sum of these effects leads to steric repulsion, which
Eulerian-Lagrangian). MD simulations provide accurate interparticle
creates a barrier preventing particle contact and cluster formation.
force calculations at the nanoscale, crucial for understanding stability.
Effective steric stabilization requires complete coverage of particles by
However, MD becomes computationally expensive for many particles.
the polymer, strong anchoring of the polymer to particle surfaces, and a
System-level simulations handle larger numbers of particles and model
sufficiently large thickness to prevent weak flocculation [253]. Elec­
complex interactions like colloidal behavior and thermal effects, but
trostatic stabilization method shows some key benefits over steric sta­
struggle to capture intricate details. By leveraging MD for force calcu­
bilization as the former offers long-time effectiveness, temperature
lations and feeding them into system-level simulations, the multi-scale

34
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

approach creates a more comprehensive picture of nanofluid behavior, Investigating the effect of base fluids on DASC performance aside
paving the way for designing stable solar nanofluids and predicting their from other parameters and conditions is not possible. That is why one
performance in real-world applications. can find contradictions in the literature when trying to select the most
appropriate candidate base fluid. Although Sharaf et al. [262] confirmed
5. Effects of the working fluid properties and geometric the superiority of water-based nanofluids in DASCs over those employ­
parameters on DASC performance ing therminol-VP1-based ones, they had something different to say
about the superiority of water as the strongest absorber base fluid in
This section explores how nanofluids influence the thermal and nf
DASCs. They reported the Ke behavior of both water- and
radiative performance of DASC. It examines how the type of base fluid, therminol-VP-1-based graphite nanofluids with fV . They found that as fV
along with nanoparticle size, shape, material, and concentration, affect nf
increases, the spectral Ke of therminol-based nanofluid excels that of
the DASC performance. Additionally, research has shown that the
the water-based one in the wavelength region (0.28–1.5 μm). Conse­
DASC’s geometric parameters play a crucial role in its thermal perfor­
quently, the temperature levels up at the collector top surface and thus
mance. Collector geometry, particularly length, height, and design,
thermal losses to the environment increase when using a
significantly impacts the volumetric absorption of solar radiation. This
therminol-VP-1-based nanofluid with a high fV . Such a behavior causes
section delves into the effects of these parameters.
DASCs employing therminol-VP-1-based nanofluids to exhibit poor
performance compared to those employing water-based ones at high fV
5.1. Base fluid type even though the former nanofluid has a stronger radiation absorption
ability than the latter. Therefore, authors concluded that the DASC
The performance of a thermal system is significantly influenced by employing a nanofluid of stronger optical characteristics and a higher
the thermophysical properties of the working fluid. In solar thermal solar radiation extinction ability does not necessarily have the highest
systems, commonly used liquids, such as water, kerosene, ionic liquids, ηPC due to the increased thermal losses with increasing the temperature
engine oil, paraffin oil, therminol VP-1, EG, and propylene glycol, nf
gain. For instance, Gou et al. [263] reported that Ke of Ti3C2 nanofluid,
exhibit poor optical characteristics for efficient solar radiation absorp­
at 0.02 wt%, reached 10 cm− 1, which is 1000 times that of pure base
tion [216]. Consequently, there arises a critical need to modify the
fluid. However, ηPC of the system using this nanofluid was 2.7 times that
properties of these thermal fluids. Engineered nanofluids have emerged
of pure base fluid.
as a solution to address this challenge. However, since the base fluid
The response of DASC performance to the base fluid type changes
amount is much greater than nanoparticles, the nanofluid properties are
with factors such as the nanofluid concentration and bottom surface
dominated by the base fluid’s properties, while nanoparticles just cause
emissivity. Thus, one can conclude that a base fluid leads to higher
alterations [36]. The base fluid OTPs have significant impacts on the
performance at given values of fV and bottom emissivity, however,
characteristics of the nanofluid obtained and thus on the overall system
another base fluid can be better at different conditions. Sharaf et al.
performance [260].
[262], again, explained such a case by comparing efficiencies of DASCs
Otanicar et al. [216] experimentally tested four heat transfer fluids
utilizing water- and therminol-VP-1-based graphite nanofluids of
(water, therminol VP-1, EG, and propylene glycol) on the solar radiation
different concentrations and using bottom surfaces of different emis­
absorption capacity and reported their extinction indices and
sivities as illustrated in Fig. 22. The figure shows the optimum nanofluid
solar-weighted absorption coefficients (Am ) (i.e. the portion of solar
concentrations at which DASC’s ηPC is maximum. These results are in
radiation absorbed through a nanofluid layer of a certain thickness) over
agreement with the literature [34,100,130,264]. However, results
the wavelength ranged from 0.17 to 3.3 μm. Results demonstrated that
confirm that the optimum fV differs with the base fluid type and bottom
water was the strongest solar energy absorber among the four tested
surface emissivity. Also, one can state that the base fluid type and fV
fluids with an Am of 13.57 % against only 9.25, 9.06, and 2.098 % for
insignificantly affect the DASC’s ηPC when the bottom surface emissivity
EG, propylene glycol, and therminol VP-1, respectively. Furthermore,
is high. Moreover, DASCs employing water-based nanofluids always
Ladjevardi et al. [49] reported the intensity of light passing through pure
show superior performance when fV is high (≥ 0.05%), no matter what
water working in a DASC for different depths. Authors found that light
the bottom surface emissivity is, where the sensitivity of ηPC to the
absorption ability of water in the IR range is much greater than in other
bottom surface emissivity is negligible when fV is high. This can be
regions. In 1 m depth, all the incident radiation of greater than 0.75 μm
justified by the fact that at high fV , most light is absorbed at the nano­
wavelength is absorbed by water. However, in the visible and UV re­
fluid top layers. Thus, the role of the bottom wall is considered irrelevant
gions, irradiation keeps much of its primary energy. Sani et al. [261]
and the DASC behavior becomes similar to a surface absorber [34].
confirmed that water outperformed EG from the radiative behavior
Understanding roles of ionic liquid base fluids in improving the ab­
perspective as the water/single wall carbon nanohorn (SWCNH) nano­
fluid of 0.01 g/lit concentration showed a 5 % higher absorbed energy sorption characteristics of nanofluids paves the way for DASC perfor­
mance enhancements. Cao et al. [265] studied the impact of ionic liquids
fraction than that of EG/SWCNH nanofluid at the same concentration.
However, they concluded that EG is a more appropriate base fluid for base fluids, in which Fe3O4 nanoparticles are suspended, on the per­
formance of DASCs and investigated the effects of the cation (i.e.,
practical solar thermal systems than water as it has several more
appropriate thermophysical properties. For instance, the lower freezing positively-charged ions) anion (i.e., negatively-charged ions) radii on
the optical behavior. Authors found that increasing the cation or anion
point of EG keeps the solar system from damage when operating in very
cold regions (down to − 6◦ C, as in Struchalin et al. [224]). Also, its radius decreases the ionic nanofluid transmittance dramatically.
Consequently, the absorption ability of the nanofluid is enhanced.
higher boiling point allows for its utilization in high-temperature ther­
mal systems or solar systems of high concentration ratios such as those As for the effect of the base fluid thermal properties, such as thermal
conductivity, on the DASC thermal performance, it is important to
operating in desert regions. That is why in some thermal applications,
EG is mixed with water to keep thermal systems from freezing and/or to remember that the thermal conductivity of a nanofluid depends on
several parameters, including the nanoparticle size, shape, loading, and
allow for higher operating temperatures. Moreover, for nanofluid sys­
tems, the suspension of some types of nanoparticles such the SWCNH in material type, stability, medium temperature, presence of dispersant,
measurement technique, and mainly, base fluid’s thermal conductivity.
EG are easier than in water and with no need for surfactants. On the
other hand, thermal conductivity of EG is lower than that of water, but Sharaf et al. [138] stated that the enhancement in the nanofluid’s
thermal conductivity accompanied by the addition of nanoparticles
the suspension of carbon nanoparticles in EG enhances its thermal
conductivity and other heat transfer characteristics as confirmed in the causes a remarkable effect on ηPC of DASC if the base fluid is with
inherently low thermal conductivity such as therminol VP-1 and EG. On
literature [261].

35
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

Fig. 22. Efficiency of DASCs with graphene nanoparticles and (a) water and (b) therminol VP-1 base fluids at different fV and bottom wall emissivities [262]. (Reused
with permission from Elsevier).

the other hand, if the base fluid inherently has high thermal conductivity offered in literature. Sinha et al. [272] observed that while Cu escalated
such as water, the DASC’s ηPC sensitivity to the alteration of thermal the thermal conductivity of the suspension by 70 %, Fe did by only 33 %
conductivity is barely noticeable. Similar results were obtained by when both are dispersed in the same medium (EG) with the same fV (1
Hwang et al. [266]. However, Gao et al. [267] concluded that the %). As a result, Campos et al. [273] stated that Cu nanoparticles pro­
greatest thermal conductivity value of a graphene nanoplatelets nano­ vided the highest ηPC and temperature gain in a DASC versus Au and Ag
fluid can be obtained by using a (1:1) EG-water mixture as a base fluid. nanoparticles. Conversely, Parvin et al. [130] concluded that DASCs
Menni et al. [268] conducted a CFD investigation of three different employing Ag nanoparticles outperform those employing Cu ones in
nanofluids composed of Al2O3 nanoparticles and water, EG, and terms of ηPC at fV of 3 %. It was observed that Ag and Au, even with tiny
water-EG mixture as base fluids to compare them in terms of thermo­ concentrations, are excellent candidates for DASCs due to the LSPR ef­
physical and hydrodynamic characteristics. Authors nominated the EG fect. Chen et al. [199] found that Ag nanoparticles with fV of 0.03 % in a
as the superior base fluid over the other two ones since the Nu of DASC escalates ηPC up to 90 %. Also, Filho et al. [236] increased the
EG-based nanofluid flow was higher than the other two. However, the stored thermal energy of a DASC by up to 144 % using only 6.5 ppm of
sensitivity of the friction coefficient of the flow to the base fluid type is Ag nanoparticles. Zhang et al. [76] achieved 20 % enhancement in the
negligible as all nanofluids showed nearly identical friction coefficient ηPC of a DASC using 0.15 ppm of AuNPs compared to the pure base fluid.
values. On the other hand, Gholinia et al. [269] stated that the engine oil Furthermore, Au can be used as an excellent shell for core/shell nano­
excels water from the temperature gain and flow properties fluids with very high absorption ability due to the LSPR effect [274,
perspectives. 275]. In terms of Au nanofluids stability, Sharaf et al. [276] showed that
Although water was classified as a strong light-absorber base fluid, polyethylene glycol-coated AuNPs are potential materials of strong
EG and engine oil were found to have superior thermophysical and heat stability and thermal performance features for functional deployment in
transfer characteristics in comparison with water. However, blending DASCs running at minimum fV and maximum power gains. Also, Al is
EG with water enhances its thermal characteristics to be, besides its offered in the literature as a strong broadband sunlight absorber to a
optical ones, more appropriate [67]. certain extent in DASCs [57,131,137].

5.2.2. Metal-oxide nanoparticles

5.2. Nanoparticle materials type
On the other hand, Agarwal et al. [277] investigated the sensitivity
of thermal conductivity of water and EG base fluids to the suspension of
The extent of properties alteration caused by dispersed phases highly
metal oxide-based nanomaterials such as alumina (Al2O3) nanoparticles.
depends on the particle material type. The nanoparticle material types
Results showed that water- and EG-based nanofluids achieved 30 % and
employed in DASCs can be categorized into three groups: metallic
31 % enhancements, respectively, at 2 % concentration and 70◦ C tem­
nanoparticles such as iron (Fe), aluminum (Al), copper (Cu), silver (Ag),
perature. Also, Prado et al. [278] proposed MgO/n-C14 as a novel metal
gold (Au), etc., metal oxide nanoparticles such as alumina (Al2O3),
oxide nanofluid with 10 % mass fraction. Authors achieved an increase
copper oxide (CuO), magnesium oxide (MgO), zinc oxide (ZnO), tita­
in thermal conductivity by 17 % compared to that of the carrier fluid.
nium dioxide (TiO2), magnetite (Fe3O4), etc., and carbon-based nano­
Back to the effects on DASCs performance, Khatri et al. [279] achieved
particles such as graphene, graphite, diamond, carbon nanohorns,
13 % enhancement in a DASC’s ηPC by employing water/Al2O3 nanofluid
carbon nanotubes, etc.
of 0.01 % concentration. Thakur et al. [280] conducted a comparative
analysis of alumina and silica nanoparticles in fly ash-based nanofluids
5.2.1. Metallic nanoparticles
working in a microchannel-based DASC. Alumina/fly ash (80:20)
Several researchers concluded that the enhanced thermal conduc­
nanofluid exhibited a higher ηPC of 72.8 %, compared to 59.2 % for
tivity of nanofluids is the cornerstone in improving the ηPC of solar
silica/fly ash (80:20) nanofluid. However, the pumping power require­
systems [266]. A look at the literature reveals that metallic nano­
ment for alumina/fly ash was 3 % higher than that for silica/fly ash. The
particles lead to higher-performance solar thermal systems than metal
improved ηPC is attributed to the superior thermal conductivity and heat
oxide nanoparticles [270]. Besides the LSPR effect of interest to solar
capacity of alumina/fly ash. Furthermore, exergy efficiency was higher
thermal systems, this is justified by the high thermal conductivity of
for alumina/fly ash due to the higher cp of alumina nanoparticles,
metallic nanoparticles [271]. The comparison between metallic nano­
resulting in less exergy destruction compared to silica/fly ash-based
particles themselves is also essential in light of some inconsistencies

36
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

nanofluids. However, alumina is offered in the literature as a weak pyrolysis technique at 500 ◦ C and seeded in a solar glycol base fluid.
broadband sunlight absorber in DASCs [98,131,223,281]. Milanese Authors experimentally assessed the potentiality of the prepared nano­
et al. [282] conducted transmittance measurements on water-based fluid for working in DASCs. Results revealed that employing such a
metal oxide nanofluids including Al2O3, CuO, TiO, ZnO, CeO2, and nanofluid with fV of 0.6 % resulted in strong light absorption ability and
Fe2O3 at fV values ranging from 0.05 to 1 % over a spectral range of escalated ηPC up to 94.5 %. Furthermore, Gorji and Ranjbar [63], in the
200–1300 nm, utilizing a 3 mm path-length quartz cuvette. Fig. 23 il­ previously discussed work, found that graphite is the strongest direct
lustrates the spectral transmittance values of the water-based metal light absorber as it absorbed 100 % of the incident light faster than
oxide nanofluids investigated by authors. Within the UV band, the magnetite and silver nanofluids, as illustrated in Fig. 24. As inferred
transmittance values were around zero, except for Al2O3 nanofluid. In from the survey above, carbon-based nanofluids can be utilized as pro­
the IR band, the opposite happened, where transmittances climbed up to spective working fluids in DASCs.
90 %. The comparison shows that CuO and Fe2O3 nanofluids are aver­ One more evidence can be found in a comparison between the three
agely more absorbent than others even though CeO2 was the best nanoparticle types presented by Sharaf et al. [85]. Results manifested
absorbent within the IR range. Also, authors calculated the attenuation that the graphite nanofluid is a much better absorber than both the
distance for each material at various fV values, highlighting Al2O3 as the aluminum and alumina ones. Fig. 25 shows that the entire amount of
worst with considerably higher extinction distance than others, while incident radiation had been totally absorbed by the graphite nanofluid
the remaining five materials displayed significantly higher attenuation while light barely crossed 80 % of the collector height. However, the
rates, with distances of less than 0.02 m observed at fV as low as 0.05 %. aluminum nanofluid showed a better performance than the alumina one.
Moreover, TiO was chosen as the best nanofluid in terms of pumping and Despite the alumina nanofluid is frequently used in thermal systems, it is
maintenance costs, where a nanofluid with 0.05 % loading of TiO is classified as a weak absorber in solar systems. An evident can be
sufficient to completely absorb the sunlight within 1 cm depth. depicted from Fig. 25 that illustrated that variations in the alumina
In addition, magnetic nanofluids are good candidates as they can nanoparticle loading had no impact on the total radiative heat flux
yield photothermal convection in DASCs. Balakin et al. [283] examined within the fluid, which means that irradiation was absorbed by the base
a magnetic Fe2O3/water nanofluid and obtained velocity and ηPC,max of fluid rather than nanoparticles.
5 mm/s and 65 %, respectively, at 2 wt% concentration and 28 mT Furthermore, nanoparticles can cause erosion and corrosion.
magnetic field. Research shows that stainless steel is preferable for nanofluid-based
hydraulic circuits over Al or Cu due to their resistance to degradation
5.2.3. Carbon-based nanoparticles [291]. Carbon-based nanofluids exhibit less corrosion compared to
Carbon-based nanofluids are superior and exhibit higher perfor­ metal or metal oxide-based nanofluids, providing an additional advan­
mance than metal- and metal oxide-based nanofluids [58,68,128,284]. tage [292].
The research highlighted in Refs. [63,284,285]suggests that nanofluids
should be the preferred choice for DASC applications due to their robust 5.2.4. MXene and transition metal nitride nanoparticles
thermal and optical performance, even with very low particle loading MXene is a set of 2D early transition elements (metallic nitrides or
[286]. Ladjevardi et al. [49] asserted that 50 % of the radiation incident carbides) of an Mn+1XnTx composition, where M is a transition element
on a DASC can be absorbed using a water/graphite nanofluid with fV of (Ti, V, Nb, etc.), X is carbon (C) or nitrogen (N), and T is a surface
0.000025 vol% with a cost increment of only $0.0045 L-1, but pure water functional group (-OH, –F, -O-, etc.). MXene nanoparticles can be sus­
can absorb 27 % only. Because of the very high thermal conductivity of pended in base fluids to create nanofluids with strong optical properties
graphene nanofluids, they have recently received greater attention nf
for DASCs. Guo et al. [263] found that Ke of Ti3C2 nanofluid can reach
[287]. Ibrahim et al. [288] concluded that 54.6 % enhancement in 10 cm− 1 at 0.02 wt%, which is 1000 times that of its base fluid. Also,
thermal conductivity of the EG base fluid when suspending graphene Wang et al. [293] compared MXene-based and graphene-based nano­
nano-sheets with 0.5 wt fraction at 70◦ C. Vakili et al. [289] showed that fluids working in a DASC. Results showed that since MXene nano­
the DASC’s ηPC could level up to 93.2 % using a graphene-platelet/DI particles exhibit plasmonic behavior in the near-IR region, MXene-based
water nanofluid at 0.005 wt% against 69.9 % efficiency using pure nanofluids showed superior optical behavior over graphene-based
DI-water under similar conditions. nanofluids. The transmittance of Ti3C2Tx-based nanofluid, at 60 ppm
Benefiting from the possibility of creating porous nanoparticles concentration, was 0 against more than 10 % for the graphene-based one
whose surface area can reach 1000 m2/g [225], Kumar et al. [290] at the same conditions. Consequently, ηPC of a DASC, employing Ti3C2Tx,
proposed an inexpensive nanofluid prepared from porous-activated was higher than that employing graphene. The former reached ηPC,max
carbon nanoparticles synthesized from Kigelia africana leaves via the (∼ 63.3%) at 20 ppm concentration, whereas the latter needed 40 ppm

Fig. 23. The spectral transmittance of several aqueous metal oxide nanofluids Fig. 24. Absorbed energy fraction for different nanofluids at different fV values
[282]. (Reused with permission from Elsevier). versus the penetration distance [63]. (Reused with permission from Elsevier).

37
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

performance. Investigations showed that employing a

CuO-MWCNT/DI-water nanofluid with a CMR of 0.15 wt%/0.005 wt%
increased the DASC terminal temperature by 14.1 ◦ C over that
employing pure DI-water. Authors, thus, proposed the hybrid
CuO-MWCNT/water nanofluid as a potential working fluid for DASCs.
Recently, Generous et al. [299] investigated the heat transfer capabil­
ities of different kinds of hybrid nanofluids including Al2O3–Au,
Al2O3–CuO, Al2O3–TiO2, CuO–Au, CuO–TiO2, and TiO2–Au nano­
particles. They found that hybrid nanofluids involving Au nanoparticles
outperformed others in terms of heat transfer performance. Also, Wang
et al. [300] employed ZnO–Au/oil composite nanofluid in a DASC and
achieved 240 % enhancement in ηPC over that of base fluid. On the other
hand, Struchalin et al. [301] revealed that employing MWCNT nanofluid
with fV of 0.0091 % in a DASC, under given conditions, resulted in ηPC
higher by up to 31.5 % than that employing Fe3O4-MWCNT hybrid
nanofluid. Furthermore, Wen et al. [302] numerically examined the use
of a novel nanofluid composed of water-EG (40:60) mixture as a base
fluid with ZrC(20 nm)-TiN(50 nm) blended nanoparticles in a planar
DASC. Authors reported that this nanofluid is highly stable and recom­
mended for DASCs as it absorbed 99 % of the incident radiation within 1
cm penetration. Also, ηPC,max of 73.7 % was achieved at optimal fv of 160
ppm. However, Huminic and Huminic [303] concluded that since
hybrid nanofluids entail a minimum of two distinct nanoparticle types,
Fig. 25. Radiative heat flux (qr ) within the top-to-bottom radiation passes for ensuring their compatibility and stability remains a crucial factor.
various nanoparticle types and loadings (φP ≡ fV ). (The two curves of alumina Additional investigations are necessary to explore various combinations
are identical) [85]. (Reused with permission from Elsevier). of hybrid nanoparticles, assess suspension stability, investigate ther­
mophysical properties, and analyze heat transfer and flow characteris­
tics to be reliably employed in solar energy harvesting systems.
concentration to reach ηPC which was 60.92 % less than the former one.
Finally, optimization results revealed that whereas the Ti3C2Tx 5.2.6. Recent advances in nanoparticle materials
nanofluid-based DASC needed 3 cm fluid layer depth to reach its For energy storage besides solar-to-thermal conversion purposes,
maximum performance, the graphene-based one needed 3.5 cm. researchers proposed the phase change material (PCM) technology to be
Transition metal nitride nanoparticles work efficiently in DASCs due employed in DASCs. PCMs have high ability to store thermal energy
to their stability and superior light absorption ability. Wen et al. [237] from a low-temperature, renewable, or waste heat source as latent heat
increased ηPC of a DASC by 67.1 % by 50 ppm of TiN nanoparticles through a charging process and transfer it, thereafter, through a dis­
suspended in EG. Also, findings of a DASC investigation by Vital et al. charging process at constant temperature. Once the solar radiation ab­
[294] indicated that nanoparticles composed of transition metal nitrides sorption starts in a PCM-based DASC, the temperature of the PCM starts
exhibited small scattering ability and wide absorption peaks, effectively to increase up to the melting point and then energy absorption continues
aligning with the solar radiation spectrum. Specifically, collectors uti­ nearly at a constant temperature. Low melting-point PCMs are selected
lizing TiN, ZrN, and HfN nanofluids with fV of 0.00005 demonstrated for such solar applications, which means small temperature gradients
significantly-enhanced ηPC , surpassing that achieved with Au nanofluid, with surroundings, low losses, and thus high-efficiency systems. The
a reference plasmonic material for solar applications, by 50.4, 36.5, and phase change slurry (PCS), which is produced by seeding very fine PCM
41.8 %, respectively. Similarly, Farooq et al. [295] found that Am of Tin particles (micro- or nano-sized) in host fluids, and the micro- or nano-
nanofluid is 80 % greater than that of Au nanofluid when both work in a encapsulated PCM (ePCM) for solar-to-thermal conversion and storage
DASC under the same exact conditions. systems have been in the focus of researchers’ attention in the recent few
years [304–306].
5.2.5. Hybrid nanoparticles As miscellaneous suggestions found in the literature, coffee colloids
Hybrid nanoparticles are synthesized from a combination of at least were suggested as novel organic, nontoxic, inexpensive, stable, and eco-
two distinct nanoparticles with the aim of surpassing the constraints of friendly nanofluids for solar applications with competitive OTPs
single-component nanoparticles, enhancing properties, attaining novel [307–310]. Kosinska et al. [310] found that ηPC of coffee colloids was 12
properties unattainable by single nanoparticles, and/or achieving mul­ % higher than that of water in DASCs with a concentration range of
tiple functionalities within a single nanoparticle [296]. 0.005–0.04 g/ml. In contrast, Nguyen et al. [311] found that coffee
Hybrid nanofluids, incorporating nanomaterials of varied types, colloids, over fV from 0.005 to 0.02 wt%, barely outperformed water.
shapes and sizes dispersed in a fluid [297], offer superior absorption More recently, Zhu et al. [312] proposed a novel nanofluid composed of
ability in contrast to single-component nanofluids. Research demon­ DI-water and cuttlefish ink, which is a feasible abundant
strates that such hybrid nanofluids achieve broad-spectrum solar ab­ environmentally-friendly natural biomass substance, for DASCs. Their
sorption [298]. Two or more different kinds of nanoparticles are seeded experimental investigations showed that the aqueous cuttlefish nano­
in the same carrier fluid to obtain stronger OTPs. Hybrid nanofluids are a fluid with a mass fraction of 0.02 % achieved temperature gain and ηPC
potential replacement of conventional mono ones due to their wider of 33 ◦ C and 60.2 %, respectively, which were 17.4 ◦ C and 52 % higher
absorption range, higher thermal conductivity, and lower pressure drop, than those by the pure base fluid. Furthermore, Gupta et al. [313]
frictional losses, and thus required pumping power. Qu et al. [14] examined a novel hybrid nanofluid composed of water and
concluded that hybrid nanofluids can result in a DASC of a higher ηPC Ag-Azadirachta Indica leaves extract. The nanofluid, with fV of 0.005 %,
than that operating a mono nanofluid. However, the concentration was tested in a DASC. Key findings include a temperature gain and ηPC of
mixing ratio (CMR) needs to be accurately tuned, that wrong CMRs or 20 ◦ C and 62 % (i.e., 12 % higher than pure water), respectively, after 2
excessive amounts of individual nanoparticles can impoverish the DASC h of sun exposure. Importantly, no sedimentation or agglomeration

38
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

occurred even after prolonged sun exposure. nf

nanoparticle size has no effect on the Ke for wavelengths above 1.25
Overall, it can be seen from the previous survey that carbon-based μm. However, for shorter wavelengths, increasing the nanoparticle size
and 2D MXene composite plasmonic nanofluids are super functional nf
causes the Ke to remarkably increase, which means that the energy
for nanofluid-enhanced thermal performance of DASCs with distin­
absorption ability increases.
guished optical and thermal performance, and thus great ηPC even at
On the other hand, several studies offered in literature were found in
very low fV values.
disagreement with the previous discussion. For example, Howe et al.
[316] experimentally tested the radiative characteristics of ion­
ic-liquid/Al2O3 nanofluids with different sizes of spherical Al2O3
5.3. Nanoparticle size
nanoparticles (10, 30, 60, and 90 nm). Results showed that absorbance
values of such a nanofluid showed significant variations with the
Nanoparticle size is discussed in the literature as a very crucial
nanoparticle size only in short-wavelength regions. However, the
parameter for the evaluation of the nanofluid functionality in thermal
strongest absorption ability was shown by the nanofluid of the smallest
systems [270]. Attempts to understand the effect of nanoparticle size on
nanoparticle size (∼ 10 nm). The largest nanoparticle (∼ 90 nm) showed
the performance of DASCs are as early as the first DASC performance
the highest transmittance and thus the lowest Ke . This contradicts the
assessment studies. Back to the Rayleigh theory formulation, when
conclusion of Ladjevardi et al. [49] who found that increasing the
substituting Eqs. (5) and (6) in Eq. (4) which, in turn, is substituted in
graphite nanoparticle size increased the extinction coefficient, as shown
Eq. (3) to determine the spectral Knp e of nanoparticles, it can then be earlier. Furthermore, Otanicar et al. [34] experimentally reported an
observed that Knp np
e is combined from two terms, the scattering (Ks ) and inverse relationship between the nanoparticle size and the DASC’s ηPC ,
absorption (Knp
a ) coefficients. where an enhancement of 6 % in the collector’s ηPC is obtained when Ag
Formulation shows that Knp s changes as the third power of the nanoparticle is reduced from 40 to 20 nm. This is attributed to the
nanoparticle diameter (D3p ). However, as the value of Dp taken for the dramatic increase of the absorption area due to employing nanofluids in
particles is very small, the Knp
s term results in a very small value, and thus DASC where authors reported a 300 times greater surface area than the
it can be neglected. In contrast, the value of the Knp a term depends on a
DASC bottom when using nanoparticles of 30 nm size and 1 % volume
fixed part and on another part varying as D2p . Tyagi et al. [57] asserted fraction, as illustrated in Fig. 20. Also, Chen et al. [84] stated that
volumetric absorption desires smaller nanoparticles (specifically, 30–40
that the slight growth of the DASC’s ηPC with the nanoparticle size, as
nm) as their scattering cross-sections are negligible compared to their
their results show, is attributed to the term containing D2p . This means
absorption ones. Similar conclusion is found in Chen et al. [317].
that the DASC’s ηPC will be represented by a constant flat line when it is
Given these contradictions offered in literature, it can be concluded
plotted against the particle size if D2p is removed from Eq. (5). Basically, that the Rayleigh approach used in the previous theoretical studies is a
since Dp has a very small value, the fixed expression containing con­ weak tool to predict the radiative behavior of a nanofluid due to the gap
stants in Eq. (5) dominates the value of Knp e , and the expression con­ between the predictive model assumptions and actual reality, especially,
taining D2p has a very little effect demonstrated in the slight growth in the in the UV band where the high light extinction is due to light scattering
plot. Mathematically speaking, much larger nanoparticles let the second [98,318]. Therefore, optical behaviors depicted based on the Rayleigh
term dominate and as a result the Knp e and thus the DASC’s ηPC approach should be taken with caution.
dramatically increase. But, as discussed earlier, the Rayleigh scattering As for thermal properties of nanofluids, the nanoparticle size plays a
approach would not be valid anymore beyond a specific value of considerable role in the determination of their enhancements. Specif­
nanoparticle diameter (i.e., α = πDp /λ≪1), and consequently, such ically, its increase adversely affects the properties enhancements caused
equations can no longer be used to evaluate the Knp by nanoparticles [319]. It can be stated that smaller nanoparticles result
e in a nanofluid.
Similar results and justifications were presented by Cregan and Myers in nanofluids with higher thermal conductivity and higher heat transfer
[137]. rates [163,320,321]. This is justified by the fact that smaller particles
For more viewpoints, several studies provided more detailed result in a greater interfacial surface area between the particles and base
description for the nanofluid’s optical behavior with particle size, and fluid which boosts heat transfer rates [322]. Moreover, the Brownian
stated that the nanofluid response differs with the incident light wave­ motion has a stronger effect on smaller particles creating convection
length. For instance, Said et al. [314] used the Rayleigh approximation currents that enhance the thermal transport and, consequently, the
nf thermal conductivity [323]. Wang et al. [318] supported such justifi­
to evaluate Ke of a TiO2-based nanofluid working in a DASC. Results
cations built based on the Brownian motion since larger nanoparticles
revealed that nanoparticles below 5 nm in diameter have nearly iden­
nf nf
mean narrower specific surface areas and slower motion, which both
tical Ka values and that increasing the nanoparticle size caused Ka to negatively affect the heat transfer process in the nanofluid and lead to a
increase. However, for all investigated sizes (Dp ≤ 20 nm), high and reduction in the improvement rate of thermal conductivity. As for the
nf
non-zero Ka values appeared for light wavelengths below 400 nm where viscosity (μ), studies systematically examining the impact of nano­
transmissivity was recorded as 0 % in such a wavelength band, and then particle size on nanofluid μ have been reviewed by Koca et al. [324].
its value suddenly showed a sharp rise. Also, the nanoparticle size no Discrepancies in μ values have been noted among these studies, even
nf
longer affects Ks when wavelengths are above 550 nm, and Ks values
nf when fV values similar. The data gathered also reveals that, for identical
nf nanofluid samples, μ can exhibit variations of up to a 40 % increase
whereupon are very low (< 0.05). Authors concluded that as the Ke is
[325] or a 37 % decrease [326] solely due to changes in particle size.
not influenced by small scattering values, the nanoparticle size has a
nf nf As inferred from the previous survey, since light absorption ability
negligible effect on Ke . Similarly, Lenert [315] observed that Ke of a and enhancement rate in heat transfer can be increased by employing
nanofluid, basically, is governed by its concentration and the optical smaller-sized nanoparticles [67,163], the nanoparticle size should be
constants (n and κ) of its components rather than the nanoparticle size. reduced for the nanofluid to exhibit the optimal radiative and thermal
Other studies were found to agree with the previous discussion. For behavior. However, this has a negative impact in terms of nanoparticles’
instance, Ladjevardi et al. [49] investigated the effect of graphite toxicity danger to ecosystem and human life. The toxicity of nano­
nanoparticle size ranged from 50 to 300 nm for a nanofluid of 0.0001 % particles increases as their size decreases. Less than 10 nm-sized nano­
volume fraction employed in a DASC. Authors used the Rayleigh particles behave like gases that can more easily enter the human body
approximation for which to be valid the nanoparticle diameter should be through skin contact and inhaling. As a result, they harm cells with their
smaller than 159 nm and 318 nm for incident radiation wavelengths of poisonous instinct. This danger issues a warning that the harmful effects
0.5 and 1 μm, respectively. Results clearly manifested that the

39
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

of nanoparticles on the environment and human health should be taken reduction of about 31 %.
into account. Researchers suggest novel nanoparticle shapes to induce the LSPR
effect. Zou et al. [331] investigated a hollow spherical structure TiN/­
5.4. Nanoparticle shape water nanofluid with fV of 5 ppm and nanoparticles of inner diameter
and thickness of 10 and 5 nm, respectively. Notably, the nanofluid
The nanoparticle shape is one of the main factors that form the exhibited remarkable stability, with multiple LSPRs observed on both
interaction between nanoparticles and their carrier fluid, which has a inner and outer surfaces. Importantly, the ηPC,max achieved was 92.64 %,
main role in the determination of nanofluid characteristics. Although surpassing that of solid TiN by 7.18 %.
many studies considered it as a crucial parameter for thermal perfor­ As for heat transfer rate, Zahmatkesh et al. [327] concluded that the
mance control and heat transfer optimization, no consistent conclusion platelet-like nanoparticle obtained the highest heat transfer rate in the
could be drawn as of yet regarding the influence of nanoparticle shape majority of earlier experiments in the natural and forced convection
on the hydrothermal performance of thermal systems [327]. Starting regimes. However, the majority of research in the mixed convection
with the optical properties, this effect was early studied in the literature regime found that the blade-like nanoparticle had the superior heat
by investigating nanoparticles of various shapes including nano-spheres, transfer performance. As for the impacts on thermophysical properties,
nano-prisms, nano-stars, nano-wires, and nano-rods. Wang et al. [328] the literature revealed that the nanoparticle shape has significant im­
used the Mie theory and the finite difference time domain technique to pacts on the kc , μ, and ρ of nanofluids. Maheshwari et al. [321]
calculate the solar absorption efficiency factor of six different-shape Au concluded that 5.54 % of the enhancement in thermal conductivity is
nanoparticles (namely, ellipse, tetrahedron, triangular, 5-sided equilat­ due to the nanoparticle shape. Authors investigated three different
eral polygon, octahedron, cross, and star). Results showed that shapes (i.e., cubes, spheres, and rods) of TiO2 nanoparticles seeded in
tetrahedron-shaped nanoparticles resulted in the greatest absorption water with the same fV . Nanoparticles of different shapes have different
efficiency factor. Qin et al. [329] presented a numerical investigation of surface areas. The cubic particles have a greater surface area than the
the absorption efficiency of sharp-edged Ag nanoparticles employed in a cylindrical particle which, in turn, has a greater surface area than the
DASC. Results revealed that such nanoparticles can produce several spherical one. Cubic nanoparticles are preferable in thermal systems
absorption peaks over the solar spectrum as both LSPR and lightning rod since they are more stable and show less tendency to agglomerate. Re­
effects exist. Sharper edges, that can be created by decreasing the cur­ sults showed that, due to their larger surface area, cubic TiO2 nano­
vature radius or the edge angle, mean a higher lightning rod effect and particles provided the greatest thermal conductivity compared to rods
thus a greater number of absorption peaks, and thus higher absorption and spheres, as illustrated in Fig. 27a. Similar comparisons in terms of
ability [321,330]. However, creating more edges is not a guarantee of thermal conductivity showed the superiority of rectangular ZnO nano­
broader absorption. Authors stated that core/shell configuration (i.e., particles over spherical ones [332], and superiority of banana-shaped
four-edged SiO2-core/Ag-shell) further broadens solar absorption over SiO2 nanoparticles over spherical ones [333]. In the study outlined in
Ag four-edged nanoparticles, nano-rods, and nano-spheres, as shown in Ref. [321], it was found that the viscosity of the nanofluid is affected by
Fig. 26. Also, DASCs employing four-edge or four-edged core/shell the shape of the nanoparticles. Specifically, the use of cubic TiO2
nanoparticles result in Am values higher than those employing smooth nanoparticles resulted in a higher viscosity compared to spherical and
nanoparticles (namely, nano-spheres and nano-rods). Calculations cylindrical shapes, as depicted in Fig. 27b. Authors attributed that to the
revealed that Am of a DASC employing four-edged nanoparticles is 35 % higher difficulty of rotation suffered by cubic nanoparticles in compar­
and 20 % greater than those employing nano-spheres and nano-rods, ison with spherical and cylindrical ones, which causes a higher resis­
respectively, at fV 10− 6 . Furthermore, Sun et al. [275] showed that the tance to the nanofluid flow and thus a higher viscosity.
core-shell composite nanorods outperformed ordinary nanorods in Furthermore, Abbasi et al. [334] conducted a molecular dynamics
terms of solar energy absorption capacity, particularly in the visible light simulation to reveal the effect of the nanoparticle shape on the nanofluid
and near-IR regions. The 5-ppm composite nanorods suspension, with a density. The investigation revealed that planar nanoparticles possess the
20 mm optical thickness, achieved a solar absorption of up to 95.3 % of most concentrated nano-layer density among others studied (namely,
the incident light, marking a 9 % increase over ordinary Au nanorods. planar, spherical, and rod-shaped nanoparticles), potentially attributed
These results suggest that core-shell composite nanorods hold promise to their flat, curvature-free surfaces. Consequently, it is evident that, in
for improving DASC solar energy absorption efficiency. Additionally, addition to factors such as surface-area-to-volume ratio and aspect ratio,
the study highlighted the sensitivity of solar energy absorption capacity nanoparticle surface curvature is a significant parameter worthy of
to imperfect shells, with larger imperfections causing a maximum consideration. The findings of this study contribute to the advancement
of our understanding of nanofluids and their industrial applications.
In light of understanding the performance of a nanofluid based on
the dispersed nanoparticle shape, the effect of the nanoparticle shape on
the DASC performance can further be discussed. For example, Campos
et al. [273] investigated both spherical and cubic Ag based nanofluids
with the same concentration and found that the cubic nanoparticles
resulted in a higher-performance DASC in terms of the temperature gain
and ηPC , as shown in Fig. 28a and b, over that using spherical
nanoparticles.

5.5. Nanoparticle loading

The nanoparticle loading, i.e. nanofluid concentration or volume

fraction (fV ), is the most important parameter that determines the OTPs
of a nanofluid and thus the performance of nanofluid-based DASCs
Fig. 26. Spectral absorption coefficient (α) of Ag nano-spheres, nano-rods, [260]. The earliest research attempts in the DASC field took care of the
four-edged nanoparticles, and four-edged SiO2/Ag core/shell nanofluids with effect of fV on the DASC performance and could elaborate the nature of
an effective radius of 20.54 nm and 10− 6 concentration [329]. (Reused with the relationship between them. Tyagi et al. [57] stated that the collec­
permission from Elsevier). tor’s ηPC exhibits a rapid increase as fV escalates, ultimately converging

40
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

Fig. 27. The effect of the TiO2 nanoparticle shape on the (a) thermal conductivity and (b) viscosity of a 2.5 wt% TiO2/water nanofluid [321]. (Reused with
permission from Elsevier).

Fig. 28. Effects of cubic (C–Ag) and spherical (S–Ag) silver nanoparticles on DASC performance in terms of the (a) photothermal efficiency, and (b) temperature gain
[273]. (Reused with permission from Elsevier).

towards an approximate maximum of 80 % at the upper limit of the fV of 0.03 % suffices to attain 90 % of the maximum efficiency across
studied fv range (0.1–5 %). The augmentation of fV leads to an different time intervals. Therefore, sensible fV selection should be a key
augmented attenuation of incident sunlight within the collector, thereby consideration for practical DASCs, balancing the advantages of
enhancing collector’s ηPC . This relationship is governed by an expo­ enhanced ηPC against the associated increase in nanoparticle cost [199].
nential dependence on the extinction coefficient, which is, in turn, The fV variations cause many interrelated impacts on DASCs per­
directly proportional to fV according to Eq. (3) (authors used Rayleigh formance. Many of such impacts were discussed by Sharaf et al. [100],
approximation in their model). Consequently, the collector’s ηPC expe­ who presented a numerical analysis of a DASC with several nanofluids of
rienced an initial steep ascent at lower fV , eventually plateauing at fV values ranged from 0.0005 to 2 %. As depicted in Fig. 29b, the initial
higher fV values. This means that there is no point of adding more augmentation in fV results in a substantial ηPC boost, irrespective of the
nanoparticles to the suspension after a certain limit. For the particular choice of base fluid (indicated by the red narrow bars, which represent
nanofluid under investigation, composed of water and 5 nm Al nano­ results when therminol VP-1 is utilized as the base fluid rather than
particles, this critical limit stands at around 2 %. Similar behavior was water). Nevertheless, as fV reaches relatively higher levels, subsequent
observed by Otanicar et al. [34] who concluded that ηPC enhancement increments exert diminishing effects on performance. As fV increases,
continues only up to a certain fV threshold, beyond which an elevated the light absorption ability of the nanofluid increases and thus the ab­
portion of incident solar radiation is reflected by the nanofluid. Conse­ sorption path length decreases as most of the incident radiation is
quently, this reflection-induced phenomenon leads to a reduction in absorbed by y the uppermost layers of the nanofluid (near the collector
DASC’s ηPC (see Fig. 20). Also, Chen et al. [199] provided similar results top surface rather than within the nanofluid bulk volume). Conse­
and justifications when examined Ag nanofluids featuring 10 nm quently, this elevates the average temperature of the top surface,
nanoparticles and a collector height of 6 mm. They found that a realistic causing increased thermal losses to surroundings. Nonetheless, at a

41
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

Fig. 29. Impact of nanoparticle fV on (a) temperature distribution and (b) DASC efficiency for different nanoparticle materials [100]. (Reused with permission
from Elsevier).

specific fV , the disparity between thermal losses to the environment and found in the results reported in Refs. [59,85,154,209,284,335,336].
the useful temperature gain of the collector reaches its peak, resulting in As for additional interlinked factors and impacts, investigations by
the maximum collector’s ηPC . Beyond this optimal fV threshold (e.g., Sharaf et al. [138] revealed that elevating fV substantially enhances the
approximately 0.05 % for graphite, 0.5 % for copper and aluminum, and DASC’s solar absorption ability. Notably, at high fV , collector’s ηPC ex­
2 % for silver), the thermal losses to surroundings become dominant. As hibits an asymptotic rise owing to heightened thermal conductivity. This
fV continues to rise, the collector’s volumetric performance approaches behavior arises from the preferential solar energy absorption by the top
that of a conventional surface collector [34]. Authors [100] argued that nanofluid layers, which subsequently transfer heat to the lower layers,
the optimal fV greatly depends on Knp e . Consequently, a graphite nano­ thereby diminishing heat loss to the surroundings and elevating effi­
fluid with a low fV is anticipated to yield the same efficiency as a silver ciency. Conversely, at lower fV , the thermal conductivity has a minor
nanofluid with fV approximately 40 times greater. This expectation effect on DASC’s ηPC , whereas emissivity emerges as a crucial factor.
arises from the former’s superior broadband solar radiation absorption Increasing emissivity enhances heat transfer from the lower to upper
capabilities compared to the latter, as results showed. Consequently, layers, and consequently, heightened thermal conductivity leads to
they concluded that, at the optimum fV (for a constant nanofluid film reduced collector’s ηPC . Thus, at low fV and low emissivity levels, the
height), a DASC attains the same ηPC,max , irrespective of nanoparticle nanofluid thermal conductivity exerts minimal influence on DASC’s ηPC .
material type. Moreover, as illustrated in Fig. 29a, at lower fV values, the Nevertheless, higher emissivity introduces a more pronounced interplay
temperature distribution exhibits greater uniformity, and solar radiation with thermal conductivity, further impacting DASC’s ηPC .
penetrates deeper into the collector channel. In essence, the pathlength As for economic aspects, Ladjevardi et al. [49] observed that
traversed by solar radiation is extended. Conversely, with increasing fV , increasing fV of graphite nanofluid from 0.00025 to 0.0025 % resulted in
the temperature distribution becomes more irregular due to the dimin­ a $0.405 per liter increase in DASC cost, while concurrently raising the
ished solar radiation pathlength, with a greater portion of the solar non-dimensional collector outlet temperature (θ = ṁcp,nf (T −
energy being absorbed by the uppermost layers. Similar behaviors were Tin )/Q̇LW) from 0.915 to 0.95. As for pumping power requirements,

42
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

many authors stated that at high fV values, viscosity rises since the temperature of the upper fluid layers, leading to elevated thermal losses
nanofluid’s non-Newtonian behavior becomes more prevalent [324, to surroundings. Fig. 30b and d, illustrating power and temperature
337–339], and thus, the pressure drop of the nanofluid flow and needed gains, respectively, demonstrate an initial rise followed by a leveling off
pumping power increase [31,303,340]. However, Sharaf et al. [100] after a specific L. Fig. 30b reveals that the asymptotic power gain is
revealed that, for DASC systems, the increment in pumping power higher at greater flow velocities, with a longer L required to reach the
associated with employing a nanofluid due to the increase in viscosity is asymptotic power gain for higher flow velocities. Fig. 30d, on the other
by no means significant. Consequently, enhancement in DASC perfor­ hand, shows that the asymptotic temperature gain remains constant,
mance is much greater than the increase in pumping power. In conclu­ irrespective of flow velocity. Nevertheless, achieving this asymptotic
sion, fV is effective up to the value of 1 wt% beyond which the temperature gain demands a longer collector as flow velocity increases.
performance starts to deteriorate, as reviewed by Hussain et al. [67]. Thus, Fig. 30b and d imply that a proportional increase in flow velocity
is crucial when extending L. To harness thermal energy suitable for
5.6. Collector length practical applications, it is essential to consider power gain, ηPC , and
temperature gain collectively. Fig. 30c highlights the achievement of a
DASC investigations revealed that the collector length (L) has an balanced trade-off between these three performance parameters at L
important role in determining its ηPC . The DASC performance exhibits an corresponding to the maximum exergy efficiency. For instance, within
enhancement as L increases up to a certain limit beyond which the ef­ the explored parameter range, L values maximizing exergy efficiency
ficiency starts to deteriorate. Increasing L means increasing the surface yield energy efficiencies of 58 % and temperature gains of 28 ◦ C, irre­
area exposed to radiation and thus capturing more amounts of energy spective of flow velocity. Similarly, Bårdsgård et al. [154] revealed that
absorbed. However, at a certain L, the increment in heat losses to sur­ ηPC of the investigated DASC grew up to 87 % at maximum when the
rounding, due to the increase in surface area exposed to ambient, be­ flow velocity increased to 3 cm/s. However, any increase in the velocity
comes faster than the increment in energy absorbed, which foreshadows beyond this limit will result in a diminishing efficiency.
a decline in efficiency. Increasing L implies, for the working fluid, longer A comprehensive examination of existing literature reveals con­
periods of sunlight exposure, which results in higher temperatures at the flicting trends in the thermal performance characteristics of DASCs
DASC exit [281]. Studies stated that there are optimum L values at concerning the influence of L. Notably, the results discussed above and
which the efficiency is maximum. Tyagi et al. [57] confirmed the studies conducted by Qin et al. [146], Jeon et al. [144], and Lee and
aforementioned discussion in their pioneer study discussed earlier. Jang [210] have reported a consistent decrease in ηPC with an extended
Moreover, Bhalla et al. [284] reported the behavior of temperature gain, L. This stands in contrast to findings of Gorji and Ranjbar [126], who
heat loss, and ηPC of a DASC with L varying from 0.5 to 2.5 m. The au­ concluded that efficiency actually improves with an increase in L,
thors found that the temperature gain increases with L, however, the particularly at greater collector heights. In light of such contradictions,
efficiency of the system decreases due to the increase in thermal losses Sharaf et al. [138] conducted further calculations trying to spotlight the
with the increase in L. exact response of DASC performance to L variations taking into
Sharaf et al. [138], additionally, presented an optimization study consideration changes of other interlinked parameters such as fV and hL .
suggesting that a sensible flow velocity selection should be compatible Authors concluded that, at higher fV , increased L consistently reduces
with L for DASC to harness the greatest amount of useful power. In ηPC due to heightened temperatures in the top nanofluid layers, exposed
Fig. 30, authors elucidated the impact of L on energy and exergy effi­ to surroundings, where most light-absorption occurs (Fig. 31a).
ciencies, power and temperature gains, using a reflective bottom surface Conversely, at lower fV , greater light-absorption occurs within the
and graphite nanoparticles at various flow velocities. In Fig. 30a, an nanofluid bulk and at the collector bottom surface, promoting thermal
extension of L results in a declining energy efficiency (i.e., photothermal trapping and leading to a rise in ηPC with increasing L (Fig. 31c). For
conversion efficiency, ηPC ), with a steeper decline observed at lower flow moderate fV , the impact of L on ηPC exhibits a rise followed by a decline,
velocities. This trend stems from the increased radiation absorption and resulting from the combined effects of elevated top layer temperatures

Fig. 30. Changes of (a) energy (photothermal conversion) efficiency, (b) power gain, (c) exergy efficiency, and (d) temperature gain with L for different average flow
velocities (fV = 0:005 %, hL = 15 W/m2.K, H = 5 mm, W = 1 m, Dp = 20 nm) [138]. (Reused with permission from Elsevier).

43
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

Fig. 31. variations of energy (photothermal conversion) efficiency of a DASC with silver nanofluid at (a) fV = 0.6 %, (b) fV = 0.01 %, and (c) fV = 0.0005 %, for
different hL values versus collector length [138]. (Reused with permission from Elsevier).

and thermal trapping (Fig. 31b). The relationship between L and the
energy efficiency (i.e., photothermal conversion efficiency, ηPC ) is
influenced by fV and hL , potentially resulting in a rise, drop, or a rise
followed by a drop in efficiency.

5.7. Collector height

In their pioneer work, Tyagi et al. [57] systematically varied the

collector height, or thickness, (H) from 0.2 mm to approximately 5 mm.
The collector’s ηPC demonstrated an increasing trend with H, eventually
reaching an asymptotic value of around 80 %. This behavior can be
explained as follows: Observations from previous studies highlighted the
reduction in spectral sunlight intensity as it traverses through the
nanofluid [57]. Consequently, solar collectors with greater depths
experience heightened sunlight attenuation. This attenuation occurs due
to the direct absorption of solar energy by the nanofluid, leading to
temperature elevation. Consequently, any reduction in sunlight in­
tensity translates into elevated fluid temperatures and, consequently,
enhanced efficiencies. Additionally, the spectral intensity of sunlight, as
described by the RTE, follows an exponential then asymptotic attenua­
Fig. 32. The effect of collector height on the photothermal conversion effi­
tion pattern. Chen et al. [199] confirmed the findings and the asymptotic
trend discussed above, indicating the presence of an optimum H. Au­ ciency at different MXene nanoparticle loadings, T = 300 K, and V̇ = 36 lit/h
[343]. (Reused with permission from Elsevier).
thors investigated a DASC employing a silver nanofluid of 0.03 % con­
centration and reported an optimum H of 10 mm resulting in ηPC,max of
90 %. Additionally, Gorji and Ranjbar [126] obtained similar behavior height leads to a higher ηPC , which is corresponding to the optimal mass
when they analyzed a DASC employing a distilled-water/graphite fraction. This is mainly due to the difference in flow speeds at different
nanofluid with fV of 0.1 % reporting a ηPC,max of 78.06 % at the collector heights under the same V̇.
optimal H (∼ 2 cm). Bårdsgård et al. [154] reported an optimum DASC H To sum up, the studies above, in addition to others [28,110,137,146,
of 300 μm with ηPC,max of 68 %. Duan et al. [341] investigated a plas­ 294], concluded that the ηPC of the DASC increases with H. Sharaf et al.
[262] agreed the previous conclusion, however, they presented justifi­
monic core-shell (SiO2–Au) nanofluid of 0.01 % concentration and
cations considering other interrelated factors such as the bottom wall
stated that DASC reached its ηPC,max at H of 10 mm. Moreover, Delfani
type and fV . For a collector with a fully absorptive bottom, increased
et al. [342] built an artificial neural network to predict the thermal
thermal losses occur due to higher top-layer temperatures when the
performance variations of a DASC employing graphene oxide (GO)
nanofluid film is thinner (Fig. 33e) compared to thicker films (Fig. 33f).
nanoparticles against geometric parameters. They confirmed the
In contrast, for collectors with fully reflective or transparent bottoms
well-established observations regarding H effect and found that, at V̇ and relatively low-to-medium fV , escaped radiation losses increase as H
and fV of 45 lit/h and 500 ppm, respectively, the DASC efficiency and Nu decreases, resulting from reduced light travel distance through the
grew by 9 % and 58 %, respectively, when H increased from 5 mm to 15 nanofluid. At higher fV , thicker nanofluid films allow more efficient heat
mm. Moreover, as shown in Fig. 32, Li et al. [343] observed that as H transfer from hot top to cold bottom layers, as evident in the comparison
increases, the optimal mass fraction of the MXene-based nanofluid between thin films (Fig. 33a and c) and thick films (Fig. 33b and d).
gradually decreases. Furthermore, at the same V̇, a smaller the collector

44
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

Fig. 33. Temperature distributions within collector bulk (fV = 0.5 %) for fully (a, b) transmissive, (c, d) reflective, and (e, f) absorptive bottom walls. Panels (a, c, e)
depict thinner nanofluid films (0.5 mm), and panels (b, d, f) depict thicker films (10 mm) [262]. (Reused with permission from Elsevier).

Consequently, ηPC of a collector with a purely absorptive bottom is

constrained by thermal losses (irrespective of fV ), whereas that of a
collector with purely reflective or transmissive bottoms is constrained
by escaped radiation losses at low-to-medium fV and by thermal losses at
higher fV . As a consequence, collectors with thicker films consistently
exhibit higher ηPC values, irrespective of fV .

5.8. Collector design

This section discusses the impact of the DASC geometric design on its
performance and shows novel designs offered in the literature. Cregan
and Myers [137] analytically analyzed an inclined low-flux DASC, as
discussed earlier, and found that the DASC inclination angle insignifi­
cantly influences its performance. Struchalin et al. [224] assessed a
custom tubular DASC employing a carbon-based nanofluid for enhanced
performance. With a MWCNT particle loading of 0.01 %, the DASC
demonstrated an average photothermal conversion efficiency of 80 %.
Compared to a surface absorption collector with similar temperature
differences, the tubular DASC exhibited significantly higher efficiency
by up to 37.9 %. Notably, the optimal fV was found to be dependent on
the specific DASC geometry and required experimental or numerical
determination. Additionally, the tubular collector design promoted
efficient phase mixing, although there were slight discrepancies be­
tween predicted and experimental optimal fV values, with the experi­
mental value slightly higher at 0.01 % due to experimental
uncertainties. Kumar et al. [344] introduced three innovative water/­
Al2O3 nanofluid-based DASC designs and compared them to a conven­
tional solar collector. The circular-corrugated collector achieved the
highest peak efficiency of 85 % at a velocity of 0.08 m/s, surpassing the
triangular-corrugated, trapezoidal-corrugated, and flat-plate collectors
by approximately 5, 11, and 16 %, respectively. Also, authors observed
changes in the deposition rate with the collector configuration.
Wang et al. [345] investigated the photothermal characteristics of Fig. 34. Schematic of the RI-DASC proposed [345]. (Reused with permission
from Elsevier).
titanium nitride nanofluids using thermal transfer oil as the base fluid at
varying solar irradiation levels. Under 5 suns irradiation (i.e., 5000 W),
the surface layer reached around 160 ◦ C, creating a significant tem­ lower ones. Therefore, and since heat conduction is the primary heat
perature gradient within the fluid. This condition led to transfer mode in such a case, the temperature of upper fluid layers re­
lower-than-expected ηPC in the collector. To address these challenges, mains high, which results in higher losses and thus a lower efficiency. In
they introduced a novel approach called reverse irradiation DASC contrast, RI-DASCs direct irradiation towards the collector’s bottom, as
(RI-DASC), facilitating a transition from thermal conduction to free shown in Fig. 34, creating a hotter fluid layer at the base. This induces
convection heat transfer within the fluid. In conventional DASCs, solar buoyancy in the opposite direction to gravity, resulting in a unique form
irradiation penetrates the fluid from top to bottom, leading to energy of free convection known as Rayleigh-Bernard convection within the
absorption concentrated in the upper fluid layers rather than in the fluid. This phenomenon disrupts the fluid’s steady-state, with macro

45
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

convection taking precedence over thermal conduction as the dominant as a new way for employing magnetic nanofluids in DASCs. They
heat transfer mechanism. Consequently, a uniform temperature field is incorporated a rotating magnetic field as an external convective system
established, significantly enhancing the fluid’s internal energy. This to achieve non-uniform solar irradiation (Fig. 36). This modification led
mechanism ensures that the heated liquid consistently moves away from to a notable increase in ηPC , with gains of up to 12.8 %. The key factor
the irradiated surface layer, ultimately leading to superior absorption behind this improvement is the shift in the heat transfer mechanism of
performance compared to traditional DASCs. Authors, particularly, working fluids from conduction to thermal convection, resulting in
found that the investigated 4.0 cm deep conventional DASC experienced reduced heat loss to surroundings. Additionally, this method proved
temperature difference of approximately 50 ◦ C between the top and effective in mitigating flow losses and pipeline blockages. Furthermore,
bottom layers. In contrast, the RI-DASC achieved a maximum temper­ the application of an electric field was found to be beneficial in DASCs.
ature exceeding the medium-temperature for various working fluids, Chen et al. [350] examined a tubular thermal oil/Al2O3 nanofluid-based
with a minimal interlayer temperature difference of about 5 ◦ C. The DASC with electric field. Authors stated that the electric field
RI-DASC exhibited a significantly improved temperature distribution re-suspends deposited colloids and enhances heat transfer between
due to the presence of Rayleigh-Bernard convection, facilitating uniform upper and lower nanoparticles. They found that the temperature gain
heat distribution within the collector. and ηPC rose up to 62.1 ◦ C and 87.8 % when applying 10 kV electric field,
Additionally, Qu et al. [346] examined a tubular MXene Ti3C2/DI-­ which are 19.2 and 14.2 % greater than those with no electric field,
water nanofluid DASC with reverse irradiation and obtained ηPC,max of respectively.
83.9 % at 220 ppm loading and 60◦ reverse irradiation angle, i.e., 63.5 % Hatami and Jing [205] conducted thermal performance simulations
greater than that of water. Furthermore, Peng and Sadaghiani [347] of a wavy bottom wall DASC. In comparison with a conventional DASC,
introduced a novel DASC featuring two linear parabolic reflectors (LPR) results showed that local Nusselt numbers were higher at the peak points
designed to enhance its ηPC . This innovative collector incorporated an of the collector channel waves due to the angle at which waves influ­
insulated wall and a cavity-like receiver tube to minimize energy losses. enced heat absorption. In another study, same authors [351] optimized
Sun-ray tracing was conducted mathematically, with numerical simu­ the wavy bottom plate profile of a wavy DASC using response surface
lations executed using COMSOL software to explore various influential methodology (RSM) analysis. Employing central composite design
parameters. The DASC design comprised two parabolic mirrors with (CCD), they assessed collector performance across nine cases with
overlapping focus points, a cavity-like receiver tube composed of varying wave amplitudes and wave numbers. Their findings indicated
concentric glass tubes, and reflective metal sheets coating the tubes that the highest local and average Nu were attained with a smaller fV of
(Fig. 35). The utilization of dual concave mirrors in two stages effec­ Al2O3 nanofluid in the optimal case characterized by an average wave
tively concentrated incident sunlight into the receiver tube. The ad­ number and small amplitude. Apart from this design, Hooshmand et al.
vantages of this novel DASC are twofold. First, it eliminated radiative [352] studied a domestic SiC nanofluid DASC filled with SiC porous
and convective losses by virtue of the cavity-like receiver tube. Second, foam. They found that such a technology enhanced the thermal per­
it facilitated a more uniform distribution of volumetric heat flux, formance of the DASC and diminished losses. Experiments recorded the
resulting in higher efficiency. Additionally, the collector’s design mini­ maximum increase in the DASC photothermal conversion efficiency by
mized heat loss through the insulated wall and extended the sunray path virtue of porous foam and nanofluid to be 14 % and 37 %, respectively. It
length through sequential reflections, thereby improving solar energy can be found in the literature that heat transfer enhancements in such
absorption by the working fluid. The study demonstrated that the pro­ cavities depend highly on porosity [353,354] and less on the wave
posed DASC design could achieve higher ηPC by up to 7 % compared to number [355]. Nevertheless Bozorgi et al. [356] combined the previous
conventional counterparts, with the proposed DASC reaching ηPC of up two technologies together yielding a synergistic effect. They proposed a
to 81 %. novel DASC design featuring a wavy bottom shape filled with an
A further modification on the conventional DASC was performed by aluminum porous medium (Fig. 37). Bozorgi et al. [356] study
applying a magnetic field which further escalates its ηPC [348]. Wang employed COMSOL software and Design of Experiments (DoE) tech­
et al. [349] by introducing a separation tank into an experimental setup niques to optimize key parameters, including fV , bottom wave number,

Fig. 35. Schematic diagrams of the (a) proposed DASC, and (b) collector tube cross section [347]. (Reused with permission from Elsevier).

46
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

Fig. 36. Schematic of the modified DASC [349]. (Reused with permission from Elsevier).

as the heat-collecting core. The collector demonstrated photothermal

conversion efficiency and temperature gain of up to 56.5 % and 170 ◦ C,
respectively. Also, experimental results revealed that approximately 52
% of thermal energy was stored in the molten salt, significantly
enhancing thermal storage capabilities of DASCs and expanding their
potential for medium-temperature applications. The molten salt was an
interesting choice as Ali et al. [358] concluded, in their review, that
molten salts are the best candidates for solar thermal energy storage
systems.
Siavashi et al. [184] investigated, among others, the effect of pres­
ence of an Al absorption sheet in a DASC system. They stated that ηPC of
pure water based DASC increases when the absorber plate is in the DASC
bottom. However, the opposite is happening when using a nanofluid.
Simonetti et al. [45] proposed the adoption of an annular pipe creating a
triple tube DASC (Fig. 39) to enhance DASCs thermal performance from
the bulk temperature viewpoint. Investigations showed that the external
surface temperature at the exit section of the triple tube DASC was 1.5 K
higher than the average temperature of the same tube section. However,
it was 20 K higher in case of the conventional double tube configuration.
Qin et al. [359] introduced a novel collector by utilizing a
semi-cylindrical absorbing coating on a dual-tube direct absorption
parabolic trough solar collector (DAPTSC) to gather both volumetric and
surface solar radiation absorption (Fig. 40a). When applied in the AO
configuration (Fig. 40c), this absorbing coating notably enhanced Am ,
particularly when absorption coefficients are low. As a result, the AO
configuration achieved superior ηPC compared to the RO configuration
with a reflective coating (Fig. 40b) over a wide temperature range. This
innovative hybrid DAPTSC design offers benefits in terms of reduced
Fig. 37. Schematic of the proposed wavy-bottom DASC [356]. (Reused with nanoparticle concentration requirements and improved solar thermal
permission from Elsevier). harvesting in the medium-to high-temperature range.
It can be observed from the discussion above that the collector design
bottom wave amplitude, and porosity, using a full factorial experiment is a key parameter in determining the performance of DASC [67]. In
design approach. The findings revealed that reducing porosity and addition to the low-flux DASCs discussed above, Rasih et al. [360]
increasing fV led to improved collector’s photothermal conversion effi­ reviewed the studies in literature concerning the concentrating DASC
ciency. However, higher wave amplitude and wave number resulted in systems including parabolic dish, Fresnel collectors, parabolic-trough,
increased pressure drop. Ultimately, the integration of a porous medium etc.
and a wavy bottom with nanofluid enhanced the DASC’s ηPC from 52 %
to 93.7 %. 6. Effects of flow properties and operating parameters on the
Zhang et al. [357] introduced a novel DASC system using a DASC performance
GO-TiN/oil hybrid nanofluid with heat storage capability. The heat
storage core consisted of a ternary mixed molten salt (44 % Ca(NO3)2, This section discusses the impacts of flow properties and operating
12 % NaNO3, and 44 % KNO3). The DASC’s photothermal performance parameters on the optical and thermal performance of DASCs.
was evaluated using a simulated photothermal conversion system, as
depicted in Fig. 38. The collector structure comprises an outer quartz 6.1. Flow rate
container, an inner cuvette, and light-absorbing nanofluids in the central
section. The cuvette’s interior is filled with the ternary molten salt acting The rate of nanofluid flow in a DASC is one of the parameters that

47
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

Fig. 38. The proposed ternary molten salt storage/DASC system [357]. (Reused with permission from Elsevier).

Fig. 39. Left, conventional tubular double pipe DASC; right, triple pipe DASC [45]. (Reused with permission from Elsevier).

substantially influence the DASC thermal performance features temperature, following a fourth-power relation. Consequently, lower V̇
including ηPC , outlet temperature, and losses to surroundings. Delfani values result in elevated losses from DASC, leading to reduced collector
et al. [59] examined the impact of flow rate (V̇) on the collector outlet efficiencies. Conversely, at higher V̇, the temperature increment within
temperature and ηPC at a 10 mm collector height and a 100 ppm MWCNT the fluid is more modest. Consequently, the impact of the aforemen­
loading. The results revealed a decrease in outlet temperature with tioned heat losses diminishes progressively, resulting in higher collector
increasing V̇. This behavior can be explained by the residence time of the efficiencies. Interestingly, the analysis of zero-loss collector ηPC indi­
working fluid within the collector. At lower V̇, the fluid spends more cated that the reduction in outlet temperature due to higher V̇ values
time inside the collector, allowing for greater temperature elevation. As was offset by the improved efficiency associated with increased V̇.
the temperature increases, heat loss to surroundings, including both Specifically, it was observed that the enhancement in zero-loss ηPC when
convective and radiative losses, becomes more pronounced. Notably, V̇ changed from 54 to 72 l/h (∼ 7.7 %) was more substantial than when
radiative heat loss from the fluid increases significantly with it changed from 72 lit/h to 90 l/h (∼ 5 %). Authors stated a correlation

Fig. 40. Schematic of the receiver part of DAPTSC of two concentric glass tubes with (a) two transparent glass tubes (TR configuration); (b) a reflective coating on
the outer tube surface (RO configuration); and (c) an absorbing coating on the outer tube surface (AO configuration) [359]. (Reused with permission from Elsevier).

48
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

predicting the studied DASC’s ηPC based on the nanofluid’s V̇ (from 54 to by Maheshwary et al. [321], discussed earlier, it was evident that the
90 l/h) and volume fraction (from 0 to 100 ppm). Similar behaviors were thermal conductivity of the TiO2/water nanofluid tends to increase with
obtained by Gorji and Ranjbar [63]. Authors argued that, at the higher temperature, as Fig. 27a shows. This behavior is explained by the fact
outlet temperature the DASC achieves at a low V̇ (Fig. 41a), losses to that material molecules become more energized when temperature in­
surroundings increase, due to a higher temperature difference with the creases, and thus, heat transfer is enhanced. However, such a growth in
ambient, causing the ηPC to decrease (Fig. 41b). On the other hand, the thermal conductivity with temperature cannot be attributed to the
exergy efficiency decreases with increasing V̇ at a constant solar flux growth in thermal conductivity of the carrier fluid [362]. However, it is
(Fig. 41c). Using the response surface methodology (RSM), authors justified by the increased Brownian motion which, in turn, augments
found that the optimum flow rate (Fig. 41b) was 10 ml/min for all energy transport within the nanofluid [363]. Particularly at elevated fV ,
nanofluids investigated. Godson et al. [364] contended that the speed of particle Brownian
Bhalla et al. [284] agreed the discussion above and revealed the motion and the influence of thermophoresis on nanoparticles height­
ened. Thermophoresis represents particle motion driven by temperature
exponential decaying behavior of thermal losses from a DASC with V̇.
gradients in the fluid. In certain instances, thermophoresis may exert a
This behavior arises from the boundary condition stating that convective
more significant impact than Brownian motion, yet both factors
heat loss at the collector top wall is equal to the conductive one. As the
collectively amplify particle collisions. Predominantly, these collisions
heat loss directly changes with temperature gain, and the latter de­
among nanoparticles lead to an elevation in the thermal conductivity.
creases with V̇, then thermal loss decreases with V̇. Jyani et al. [361]
Additionally, Brownian motion induces stirring of water molecules,
presented the same argument. Also, they experimentally observed that
giving rise to microconvection within the liquid, thereby enhancing heat
the transition from laminar to turbulent flow regimes enhanced the
transfer. In Fig. 27b, Maheshwary et al. [321] indicated the reduction of
collector’s ηPC , as turbulent flow exhibited reduced solar absorption
the viscosity with temperature, which promises to decrease the pumping
compared to laminar flow. Consequently, laminar flow led to greater power. Similar results were presented in Refs. [67,365,366]. Lower
temperature increases, resulting in elevated heat losses. This theme is viscosity allows nanoparticles to move more easily through the base
accounted for in the literature by investigating the Re variation impacts fluid, enhancing their mobility and reducing resistance to heat transfer,
on DASC performance that will be a discussion in a coming subsection. which can contribute to an increase in the thermal conductivity.
Higher thermal conductivity and lower viscosity with temperature
6.2. Operating temperature are reasons for higher DASC thermal performance [280], but it is
well-known that increasing the nanofluid temperature means a greater
Apart from the fact that increasing the nanofluid temperature in­ difference with the ambient and thus greater losses. Also, Wang et al.
creases the convective and radiative heat losses to surroundings and thus [345] observed that elevated temperatures diminished the irradiation
deteriorates the DASC’s ηPC , as shown in Fig. 28a [273], this section absorption capacity of thermal oil L-QB320, which is the base fluid in
discusses the impact of temperature on the nanofluid OTPs. In the study

Fig. 41. The effect of flowrate on the (a) outlet temperature, (b) thermal (photothermal) efficiency, (c) exergy efficiency [63] of a DASC for different nanofluids and
concentrations. (Reused with permission from Elsevier).

49
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

their study, and Otanicar et al. [109], in the study discussed earlier, DASCs. The authors observed that the increased Re leads to an uneven
stated that the temperature increase causes the transmittance of nano­ nanoparticle spatial distribution within the nanofluid, which means
fluids to increase. However, the utilization of plasmonic nanoparticles non-uniform light extinction due to the spatial variations in the nano­
introduces the LSPR effect, which primarily governs the absorption fluid optical characteristics. Plus, the authors affirmed the increasing
ability of these particles. This effect is influenced by factors such as DASC ηPC behavior with Re as found by past studies, they described the
particle size, morphology, and metal composition. Temperature plays a detailed behavior of the DASC performance criteria with deep justifi­
significant role in altering the optical absorption properties of nano­ cations by virtue of capturing the discrete nature of nanoparticles. Au­
particles, impacting their dielectric function, refractive index of the thors reported that Re significantly influences the nanoparticle
surrounding medium, and particle size. Specifically, as temperature in­ distribution in nanofluid flow within DASCs, with substantial conse­
creases, the nanoparticles expand in volume while the plasma frequency quences for both optical and thermal properties, as well as collector
decreases. Conversely, lower temperatures weaken the LSPR effect, performance. Understanding the effect of spatial nanoparticle distribu­
leading to a decreased absorption rate. tion and local concentration could help present the full picture of the
ηPC of nanofluid-based DASCs and a commercial Rheem RS24-BP temperature and radiative heat flux distributions within the DASC.
collector are compared in Fig. 42 revealing two important points. Fig. 43a illustrates that, at low Re, there is a uniform nanoparticle dis­
First, most nanofluid-based DASCs have lower performance than a tribution within the collector. However, as Re increases (Fig. 43b and c),
conventional FPC, especially at higher temperatures. This indicates the there is a notable reduction in concentration near the collector walls due
need for more research on developing full-scale collectors that can to increased nanoparticle transverse migration away from the upper and
reduce the heat loss mechanisms. Second, some nanofluid-based DASCs lower surfaces. This migration phenomenon in particulate flows within
have higher performance at very low temperatures (marked in the channels arises from the tendency of suspended particles to move from
figure), which is due to the strong absorption capacity of nanoparticles regions with high shear rates to those with low shear rates, resulting in a
[18]. low-particle-concentration zone adjacent to the walls. At low Re, the
impact of shear-induced migration is counteracted by Brownian diffu­
6.3. Non-dimensional numbers sion, leading to nanoparticle homogenization within the collector.
Conversely, in high convection-dominated flows (i.e., high Re), the
This section is devoted to discussing the effects of non-dimensional weaker influence of Brownian diffusion in comparison to flow inertia
numbers, including Re, Nu, and Pe on the performance of DASC sys­ intensifies particle migration, as depicted in Fig. 43c.
tems. A literature survey revealed that Re is a major operating parameter Eulerian-Lagrangian modeling reveals that the nanofluid’s flow Re
with which the DASC performance criteria remarkably change. Parvin significantly influences its optical properties. Fig. 44 displays the spec­
nf
et al. [204] examined the performance of DASC employing a water/Cu tral Ke of graphite nanofluid near the collector’s upper surface for
nanofluid using the single phase model under a range of Re extended varying Re values. While Re itself does not have a direct relationship
from 200 to 1000. Results showed that the heat transfer rate enhanced nf
with Ke , it does impact the uniformity of nanoparticle distribution, as
by 31 % and ηPC has doubled when Re varied from 200 to 1000 at fV of 3 nf
previously established. Consequently, the local Ke , determined by local
%. As Re rises, inertial forces become predominant, leading to an
nanoparticle concentration, is affected by Re. Specifically, as depicted in
enhancement in convective heat transfer. Chronologically, Das et al.
Fig. 44, increasing Re leads to a reduction in the nanofluid’s solar ra­
[209] numerically investigated a DASC employing an aqueous TiO2
diation extinction capability near the collector’s upper surface – a
nanofluid using both a single phase model and a mixture model. They
crucial area in DASCs, as it directly affects thermal losses to surround­
reported that, as Re increases over the range (100–2000), the Nu,
ings. This decline is attributed to the decrease in nanoparticle concen­
average heat transfer coefficient, and pressure drop significantly in­
tration near the upper and lower walls with increasing Re, as illustrated
crease. Moreover, the multiphase model exhibited a higher endorsement
in Fig. 43.
with experimental results while the single-phase model remarkably
As for the effect of the heat loss Nu, Veeraragavan et al. [136]
underestimated the Nu at certain Re values. Furthermore, Jyani et al.
basically stated that the DASC with a higher heat loss Nu has greater heat
[369] concluded that at considerably high Re, the DASC performance
losses and thus lower ηPC . Furthermore, Lee and Jang [210], in their
grows when the flow regime transits from laminar to turbulent.
work discussed earlier, reported the effect of the heat loss Nu on ηPC and
In a more thorough investigation, Sharaf et al. [85] used the novel
temperature distribution within a DASC. They changed the value of the
coupled Eulerian-Lagrangian approach to reveal the effect of the
convective heat loss coefficient from 6.43 (the value used by Tyagi et al.
discrete nature of nanoparticles on the thermal and optical features of
[57]) to 250 W/m2K (the maximum forced convection heat transfer
coefficient for gases), which is corresponding to Nu range from 0.106 to
4.12. Results showed that as Nu increases, the temperature within the
DASC decreases. At the lowest Nu (=0.106), temperature rises uniformly
along the collector length with a small temperature gradient in the
y-direction, resulting in a negligible thermal trapping effect, akin to the
"greenhouse effect". Conversely, at the top surface with high heat loss
(Nu = 4.12), temperature increases slowly along the collector length due
to significant heat loss, leading to a pronounced thermal trapping phe­
nomenon where the top surface temperature is lower than the bottom
surface. Fig. 15 illustrates an inverse relationship between the DASC’s
ηPC and heat loss Nu. Also, it demonstrates that the predicted ηPC based
on the plug-flow assumption exceeds that with fully-developed flow.
This suggests that ηPC is more sensitive to heat loss magnitude with
fully-developed flow. Furthermore, the plug flow exhibits greater ther­
mal energy transport along the flow direction due to higher momentum
near the top surface compared to fully developed flow. Consequently,
the prediction model using plug flow may overestimate ηPC compared to
Fig. 42. The effect of operating temperature on ηPC for different nanofluids
the model with fully-developed flow, which aligns more closely with
[18]. (references [59–64] in the figure are references [33,59,60,62,367,368] in
the reference list, respectively). (Reused with permission from Elsevier).
actual conditions.

50
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

Fig. 43. Graphite nanoparticle distribution based on the Eulerian-Lagrangian approach at fV of 0.05 % and Re of (a) 0.1, (b) 1.0, and (c) 10 [85]. (Reused with
permission from Elsevier).

Fig. 45. Effect of Pe on DASC photothermal efficiency [210]. (Reused with

permission from Elsevier).

nf
plug-flow assumption, again, overestimates the DASC’s ηPC compared to
Fig. 44. The spectral Ke of 0.05%-graphite nanofluid near the collector top
the fully developed one since the plug flow generates higher momentum
wall at different Re [85]. (Reused with permission from Elsevier).
in the boundary layer near the wall, which leads to overestimations of
the useful energy carried by the working fluid.
As the Peclet number (Pe) puts convective and heat transport phe­
nomena in correlation (i.e., Pe = (convection/thermal diffusion) = Re 7. Economic and environmental analyses of nanofluid-based
Pr), it can be seen that it is not a material property since it involves Re DASC systems
that depends on the flow velocity and geometry characteristic length. Pe
serves, in DASC analytical studies discussed earlier, as a scaling Nanofluid-based DASCs have gained significant attention as a
parameter for the axial coordinates [136]. The increase of Pe indicates promising technology for harnessing solar energy as they offer several
that the collector length needed to achieve a certain temperature in­ advantages, including higher efficiency, reduced heat losses, and
crease. Furthermore, high Pe implies that the nanofluid flows rapidly in simplified design. However, before widespread adoption, it is crucial to
the collector and thus can deliver greater amounts of heat, but this fast evaluate their economic viability and environmental impact.
movement shortens the residence time in the collector. These two con­ The literature offers definitions to assess the economic and envi­
sequences lead to increase in the DASC’s ηPC and decrease in its tem­ ronmental impacts of products or systems such as the nanofluid-based
perature, as obtained by Lee et al. [215]. Similar conclusions were stated DASC, including the life cycle assessment (LCA) and offset damage
in a study by Lee and Jang [210]. However, these authors reported that cost. LCA is a systematic methodology for assessing the environmental
the Pe impacts on ηPC can be observed only at high heat losses, otherwise impacts associated with all stages of a commercial product. It serves as a
they are negligible. Moreover, at high heat loss, the ηPC curve exhibits holistic baseline for comparing carbon footprints and guides sustainable
modest gradients that could reach zero at high Pe, as shown in Fig. 45. As decision-making by assessing environmental impacts across a product’s
an explanation, the DASC’s ηPC is defined as the ratio of useful heat entire life cycle. The offset damage cost refers to the reduction in envi­
gained (Q˙u = ṁcp ΔT) to radiative power received. As Pe increases, ṁ ronmental or economic harm achieved by implementing a particular
increases, but the temperature gain (ΔT) decreases. Thus, variations of technology. In the context of nanofluid-based DASCs, it represents the
ηPC with Pe are barely noticed. Moreover, Fig. 45 shows that the positive impact of using these collectors compared to conventional

51
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

alternatives, such as the reduced emissions, embodied energy savings, conclusions were drawn by Otanicar et al. [34] in the context of graphite
faster payback period (i.e., how long it takes to recover the cost of the nanofluids for DASCs. Additionally, Faizal et al. [373] showed that,
system), and quantifying the offset. For instance, Faizal et al. [370] compared to conventional collectors, nanofluid-based collectors exhibit
showed that SiO2 nanofluid-based collectors can save 280 MJ more a shorter payback period primarily due to their reduced capital cost.
embodied energy, offset 170 kg less CO2 emissions, and achieve a faster Among nanofluids, SiO2 and CuO-based collectors have the shortest
payback period compared to conventional water-based solar collectors payback periods. Despite the higher cost of CuO nanoparticles, the
increased efficiency and reduced collector area (calculated based on Eq.
(1)) of CuO-based nanofluid collectors nearly equalize their overall cost
7.1. Economic assessment with SiO2 nanofluid collectors. On the other hand, Otanicar and Golden
[375] stated that the conventional collector may have a shorter payback
Nanofluids, with their superior thermal conductivity, have enabled period compared to the nanofluid-based collector, mainly because of the
the development of smaller and more compact solar collectors that still higher initial capital cost associated with certain types of nanoparticles
deliver the same desired output as that delivered by a larger conven­ required for the nanofluid-based system. Additionally, further cost
tional collector with a conventional working fluid. Faizal et al. [371] savings can be achieved if nanoparticle prices decrease as they become
stated that the use of Al2O3/water nanofluid instead of pure water in a more widely adopted and produced.
flat-plate solar collector can achieve the same output temperature but
with a collector size that is up to 24 % smaller. The reduction in the
system size means lower embodied energy, which is energy needed to 7.2. Environmental impact
produce a material, and lower overall material cost. In the existing
literature, emphasis is placed on the consideration of embodied energy One significant motivation for transitioning to solar energy har­
when employing nanofluids in solar collectors, as opposed to conven­ vesting systems, particularly for residential applications like water
tional pure fluids. The focus remains on the energy expended during the heating, is the reduction in emissions stemming from fossil fuel burning.
manufacturing phase of the solar collector itself, while other life cycle By conducting a thorough analysis of thermal performance, the emis­
stages (such as distribution, maintenance, and disposal) are intention­ sions offset facilitated by solar collectors can be qualitatively assessed.
ally neglected. Notably, Ardente et al. [372] assert that over 70 % of the Otanicar and Golden [375] presented an environmental assessment of a
total embodied energy within the system originates specifically from the nanofluid-based solar collector used to heat water. The assessment is
collector manufacturing process. To quantify this, researchers often based on the energy difference between the energy needed to meet the
adopt the reduction of collector area as the functional unit, significantly daily hot water load (Qload ) and the auxiliary energy supplied by a
influencing the overall weight and embodied energy. Faizal et al. [373] conventional electric or natural gas water heater (Qaux ), considering the
stated that the predominant materials, for solar collectors, include glass emissions profile associated with electricity generation in Phoenix,
and copper. They reported that, in a typical 40 kg collector, the weight Arizona. The authors stated a formula to determine the amount of
ratio stands at approximately 30 kg of glass and 10 kg of copper. emissions offset by the use of a solar collector (EMP,offset ), expressed as
Importantly, the embodied energy index varies significantly: 15.9 follows [375]:
MJ/kg for glass and a notably higher 70.6 MJ/kg for copper [34]. By N
∑
incorporating the results of size reduction, researchers can accurately EMP,offset = Aj,P Bj,P (Qload − Qaux ) (114)
calculate the weight and embodied energy for the solar collector. The j=1

potential cost savings resulting from the smaller collector size could
offset the increased price of the utilized nanofluid and thus balance where Aj,P is the amount of the pollutant (P) generated from the fuel (j)
overall cost of the nanofluid-based solar thermal system. These consid­ given in kg per MJ of the produced energy, and Bj,P is the percentage of
erations are pivotal for evaluating the sustainability and efficiency of energy produced by a certain fuel type. Qaux is given as follows:
nanofluid-based solar technologies. Qaux = Qload − (Qu − Qloss ) (115)
The economic evaluation of nanofluids in solar energy systems, as
reported in the literature [64,374], depends on several factors, including where Qu is the useful energy obtained from the collector, and Qloss is the
local energy costs, government subsidies, and system-specific consid­ energy lost to the surroundings. The environmental analysis demon­
erations. These considerations include the electrical usage of mechanical strated that manufacturing the nanofluid-based DASC system results in
circulation pumps, nanofluid costs, annual maintenance expenses (set at 34 kg less CO2 emissions compared to the conventional solar collector.
2 % of the initial cost), and an expected annual interest rate for cumu­ Additionally, during operational use, it saves 50 kg of CO2 per year.
lative savings (also at 2 %). Additionally, the systems are assumed to However, the differences in remaining emissions (SOx and NOx) are
generate continuous electricity, remaining connected to the grid. relatively small. Also, results showed that, over the 15-year expected
Solar collector capital costs typically consist of two components: lifetime of both solar collectors, the nanofluid-based system would offset
area-dependent costs (CA ) and area-independent costs (Cf ). The total more than 740 kg of CO2 compared to a conventional solar collector and
cost (Ct ), when considering the maintenance cost (Cm ) which is taken as 23,000 kg of CO2 when compared to traditional electric water heaters.
2 % of the capital cost with a 1 % increase yearly for the system lifetime, Furthermore, research indicated that if 50 % of the residents in Phoenix,
is expressed as [34]: Arizona, adopt nanofluid-based solar collectors, an impressive annual
reduction of 1,000,000 metric tons of CO2 could be achieved. Finally,
Cs = CA A + Cf + Cm (113)
the total damage cost avoided per year was $88 and $91 for the con­
To calculate the cost savings resulting from solar collector operation, ventional collector and nanofluid-based DASC, respectively.
the daily energy flow is considered alongside local electricity rates. According to Tiwari et al. [376], the utilization of Al2O3-based
Faizal et al. [373] performed an economic and embodied energy nanofluid in flat-plate solar heaters significantly reduces CO2 emissions
analyses for a nanofluid-based solar thermal system (solar heater in (in kg CO2 per kWh) by 32 % compared to conventional flat plate solar
Malaysia) utilizing different aqueous nanofluids and compared the re­ heaters. Additionally, a study by Khullar and Tyagi [377] revealed that
sults with a conventional system utilizing pure water as a working fluid. adopting nanofluid-based concentrated heaters could lead to substantial
Research findings indicated that reducing the copper and glass content energy savings. Specifically, these heaters would save 1716 kWh per
in the nanofluid-based solar collector led to an average reduction of household annually, which would otherwise come from fossil fuels
approximately 220 MJ in the embodied energy of the nanofluid-based (liquefied petroleum gas). Consequently, this energy-saving measure
collector compared to the conventional water-based collector. Similar results in a reduction of 2200 kg of CO2 emissions each year. Kim et al.

52
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

[378] investigated various nanofluids (including Al2O3, MWCNTs, SiO2, field. By employing colloidal suspensions, DASCs are characterized by
and CuO) in a U-tube solar collector. Compared to conventional solar enhanced solar absorption ability and improved photothermal conver­
heaters, the use of MWCNT nanofluid, that escalated the photothermal sion efficiency. However, despite significant progress, there are several
efficiency of the collector up to 63 % with fV of 0.2 %, in the system challenges, unexplored areas, and research gaps that need to be
would lead to annual reductions of 6.900 kg of CO2 and 0.022 kg of SO2 addressed to harness the full potential of this technology. This section
emissions. Considering the usage of 50 solar heaters in Gwangju, South discusses the conclusions withdrawn from the literature reviewed
Korea, the adoption of nanofluid-based solar heaters would also result in above, key challenges, and future directions in the field of nanofluid-
a substantial reduction in annual coal usage by 792 kg, equivalent to a based DASCs.
decrease of 2080 kg in CO2 production and a reduction of 6.7 kg in SO2 Concluding remarks and challenges:
emissions.
On the other hand, in spite of the positive side of nanofluidic systems - Solar nanofluids are used in DASC systems as they can enhance their
in terms of the harmful emissions reduction, critical environmental optical performance by orders of magnitude with very tiny volume
concerns could be associated with nanofluids, necessitating a thorough fractions (fV < 0.001) and ultrafine nanoparticles (Dp < 40 nm).
discussion. These concerns arise from the use of nanoparticles in nano­ Dependent scattering effects in calculations of nanofluid’s optical
fluid synthesis, stemming from both the synthetic procedures involved properties are usually negligible in the literature for (fV < 0.006).
and the physicochemical properties of nanomaterials [379]. While the Nevertheless, other studies proved that adopting fV < 0.06 as a
composition of nanofluids is relevant, it is the unique features of threshold permits the neglect of dependent scattering effects in solar
nanomaterials, such as their surface area and size, that raise primary nanofluids (for sufficiently small particles Dp < 40 nm) with an error
environmental alarms. Due to their substantial surface energy, nano­ of no more than 5 %.
materials exhibit high reactivity and can readily penetrate human - The Rayleigh approach is a weak tool to predict the radiative
bodies, potentially causing harm to vital organs [380]. The size of these behavior of a nanofluid, especially, in the UV, where the strong light
nanoparticles is a critical factor; particles typically smaller than 100 nm extinction is due to light scattering, which is neglected in this
can be inhaled, posing a serious risk to lung health. Furthermore, the approach. In a bit more detail, it results in accurate extinction co­
release of nanoparticles into the atmosphere during nanofluid prepara­ efficient predictions for carbon-based nanofluids, overestimations
tion or usage can lead to severe health issues upon inhalation by humans for metal-based nanofluids in the relatively short-wavelength bands
or other mammals. Once inhaled, these nanoparticles can enter the (< 700 nm), and underestimations for metal-based nanofluids in
bloodstream, resulting in toxicity to various organs, including the lungs, longer wavelength regions. Researchers attributed the failure of the
skin, liver, and brain [381,382]. Their impact extends to cellular Rayleigh approach with metal-based nanoparticles to agglomeration
structures, affecting DNA, mitochondria, and cell membranes, ulti­ (which enlarges the particle size over the manufacturer-rated size)
mately inducing oxidative stress [383,384]. Therefore, implementing and surface oxidation (which changes the particle optical constants)
proper safety protocols is essential [379]. phenomena. Moreover, the independent scattering assumption,
More concretely, all carbon nanoparticles are toxic, but their toxicity stated for Rayleigh approximation to be valid, is justified within the
level depends on their dimensions and surface chemistry [164]. Also, Au limit conditions mentioned in the previous point, however, this was
nanoparticle size and even shape are good indicators of its level of not the case in all the studies that used the Rayleigh approach. This
toxicity. When Au nanoparticles are 15 nm or larger, they behave justifies the contradictions found in the literature regarding nano­
essentially non-toxically, but when they are 1–2 nm smaller, they fluids’ radiative behaviors with nanoparticle size variations.
become highly toxic [385]. Furthermore, spherical Au nanoparticles are - Many studies in the field of DASCs simplify radiative heat transfer.
less hazardous than rod-shaped ones [386]. Depending on the toxicity of For example, they assume that solar irradiance varies only along the
nanoparticles, they can cause minor to severe harmful implications on y-axis, neglecting radiation scattering and emission by the working
the ecosystem, as discussed earlier. As a result, one of the biggest fluid. Consequently, the impact of radiation is simplified to a heat
challenges may be the increased use of nanoparticles in various appli­ source dependent solely on y-coordinates. Radiative heat loss is
cations and potential human interaction with them. However, no addressed using an effective coefficient, alongside convective heat
benchmarking on the toxicity of the various nanoparticles utilized for loss. This simplification is based on the approximation that radiative
various applications, the procedures needed, or requirements for their losses in low-flux DASCs can be neglected as severe radiative heat
disposal has been documented as of yet [64]. losses, that cannot be neglected, occur when DASC temperatures
When nanofluids are released into water streams like rivers and seas, exceed 750 K. However, solving the full radiative transport equation
their impact on aquatic life becomes a concern. These nanoparticles can and accounting for radiative losses in high-temperature DASCs may
accumulate in organisms such as fish and shrimp, potentially affecting yield significantly different results compared to the simplified RTE.
both aquatic ecosystems and human consumers [387,388]. Researchers, For instance, considering radiative losses from a DASC operating at
instead, have proposed innovative methods to mitigate the potential 500 ◦ C (close to the literature-stated limit of 750 K) and solving the
harmful effects of nanofluids. For instance, Akhavan-Behabadi et al. full RTE revealed that the conventional model overestimated outlet
[389] advocate for recycling carbon nanotubes (CNTs) used in nano­ temperature by 27.6 ◦ C and ηPC by 18.9 % [135], which are signifi­
fluids to prevent inhalation-related health risks. Similarly, Ranjbarzadeh cant disparities. Therefore, the literature-stated limit of 750 K should
et al. [390] highlight the use of eco-friendly nano-silica derived from be taken with caution when calculating for reliable solar systems.
rice husks—a renewable agricultural waste. This sustainable nano­ - Several assumptions, commonly stated by researchers in DASC
material enhances heat transfer by up to 33 % (at 55 ◦ C with a 3 % modeling, result in inaccurate performance predictions. For instance,
volume fraction). Careful nanoparticle selection, such as opting for the plug flow assumption leads to overestimations of DASC’s pho­
CuO-based nanofluids over pure Cu that persists in the environment and tothermal conversion efficiency, particularly in case of high heat loss
makes it challenging to eliminate [391], contributes to effective recy­ coefficient and Reynolds number. Such an assumption is not valid for
cling and minimizes environmental impact. DASCs because it specifies a creeping flow which is not the real case
in DASCs. Also, this assumption overestimates the flow momentum
8. Concluding remarks, challenges, research gaps, and future near the collector walls, and thus leads to overprediction of the
directions collector performance. The fully-developed flow assumption is more
realistic in DASCs modeling although it neglects the entrance region
Nanofluid-based direct absorption solar collectors (DASC) represent impacts on the collector performance. Moreover, several pieces of
an innovative and promising technology in the solar energy harvesting research ignore the atmospheric attenuation impacts on the

53
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

extraterrestrial solar spectrum by the assumption of blackbody pro­ - The response of DASC performance to the base fluid type changes
file for incident radiation heat flux, which results in inaccurate with other factors such as the fV and bottom surface emissivity. While
performance description since the nanofluids’ scattering and ab­ a particular base fluid might excel at specific fV and emissivity
sorption efficiencies are a function of solar radiation wavelength. combinations, others may perform better under different conditions.
- The fully-continuous single-phase modeling approach based on the Interestingly, it was observed that the base fluid type and fV insig­
homogeneous nanoparticle distribution assumption, prevalent in the nificantly affect the DASC photothermal conversion efficiency when
literature, is no longer valid beyond a critical Reynolds number the bottom surface emissivity is high. Moreover, DASCs employing
(Re ≈ 1) [85]. Thus, the need for capturing the discrete nature of water-based nanofluids always show superior performance when fV
nanoparticles and considering the spatially uneven attenuation of is high (≥ 0.0005), no matter what the bottom surface emissivity is
solar radiation due to the spatially non-uniform nanoparticle distri­ (i.e., the sensitivity of the DASC performance to the bottom surface
bution within the DASC becomes necessary for accurate performance emissivity is negligible when fV is high). Such contradictions can be
criteria predictions. Such an assumption was found to overestimate resolved based on the mentioned-earlier fact that, at high fV , most
the DASC’s photothermal conversion efficiency, which affects the light is absorbed at the nanofluid top layers. Thus, the role of the
reliability of the system if it is manufactured for commercial pur­ bottom wall is considered irrelevant and the DASC behavior becomes
poses. Instead, many studies stressed that DASCs performance pre­ similar to a surface absorber.
dictions obtained based on a multi-phase model are far more - Although water was classified as a strong light-absorber base fluid
accurate and significantly closer to experimental results compared to among other heat transfer fluids [298], EG and engine oil were found
those obtained based on a single-phase model. to have superior thermophysical and heat transfer characteristics
- When comparing theoretical model predictions with experimental compared to water. However, blending EG with water enhances the
data for validation purposes, it is imperative to account for experi­ thermal characteristics of water to become as appropriate as its
mental uncertainties arising from various factors including nano­ radiative properties.
particle aggregation and sedimentation phenomena, and potential - Carbon-based, transition metal nitride, and 2D MXene composite
measurement errors. plasmonic nanofluids are super functional for nanofluid-enhanced
- The solar nanofluid’s prolonged exposure to intense solar flux or thermal performance of DASCs with distinguished optical and ther­
elevated temperatures leads to its destabilization, which weakens its mal performance, and thus great ηPC even at very low volume
ability to extinct solar radiation and thus diminishes the photo­ fractions.
thermal efficiency of the solar system. Stability is one of the most - Nanofluids composed of angular (sharp-edged) or cubic nano­
important features that should characterize a nanofluid. Research, particles manifest enhanced properties pertaining to light absorption
therefore, suggested many protective and responsive methods to and thermal conductivity when compared with spherical or rod-like
enhance the nanofluid stability. However, as exposure to solar ra­ nanoparticles. Platelet-shaped or cylindrical particles, in particular,
diation is a genuine reason for solar nanofluid instability, it is yield superior performance in terms of entropy, viscosity generation,
important to use a stability technique that does not mask nano­ and thermal conductivity. Commonly, nanoparticles characterized
particles and deteriorate their absorption ability. by anisotropic characteristics and possessing multiple surfaces
- The notion of a high zeta potential guaranteeing exceptional exhibit heightened capabilities in terms of radiation absorption and
colloidal stability is challenged by several studies. Sterically stabi­ thermal properties. However, no consistent conclusion could be
lized dispersions, for instance, can exhibit low zeta potential yet drawn as of yet regarding the influence of nanoparticle shape on the
remain entirely stable [77]. Conversely, electrostatically stabilized hydrothermal performance of thermal systems. Also, the influence of
dispersions experience fluctuations in electrical double layer thick­ nanoparticle shape on the nanofluid performance in the majority of
ness due to changes in ionic strength (e.g., contamination, oxida­ research is theoretically implemented through thermophysical
tion). This, in turn, alters the zeta potential needed for optimal properties, with relatively little experimental evidence to support it.
stability. Consequently, zeta potential should be interpreted solely as Future experimental efforts for diverse nanoparticle shapes in
an indicator of electrostatic repulsion forces between colloidal various nanofluids with single and hybrid nanoparticles are therefore
nanoparticles. necessary.
- In contrast to nanofluid thermal properties, which typically neces­ - Since light scattering effects are less and thermal properties are
sitate relatively high nanoparticle loadings for heat transfer prop­ better when using smaller-sized nanoparticles, nanoparticle size
erties to be notably enhanced, optical properties relevant to should be reduced for a higher-optical performance solar nanofluid
photothermal conversion systems, like DASCs, can exhibit significant and higher enhancement rate in heat transfer.
improvement with nanoparticle dispersion at orders of magnitude - Increasing the collector length results in adverse impacts on the
lower volume fractions (< 0.001). Research has shown that higher DASC’s photothermal conversion efficiency, as found by many
nanoparticle loadings yield near-surface absorption dominating over studies, due to increasing the exposure time to radiation which yields
volumetric absorption since a greater portion of radiation extinction elevated temperatures and greater losses. Optimization studies sug­
occurs within the uppermost nanofluid layers rather than the bulk gested that a sensible flow velocity selection should be compatible
nanofluid volume, which can deteriorate the performance of with the collector length for the DASC to harness the greatest amount
nanofluid-based DASCs. Furthermore, studies on DASC proved that of useful energy.
there exists an optimum volume fraction beyond which the solar - Contradictions, including a rise, drop, or rise-followed-by-drop, are
radiation absorption ability of the nanofluid no longer increases. This observed in the literature regarding the behavior of DASC’s photo­
value changes with the nanofluid type (e.g., optimum fV of 0.0001 thermal conversion efficiency with variations of collector length. By
for MWCNT/water nanoparticle [14]). In general, different studies taking into consideration other interlinked parameters, such as the
stated that the performance of a nanofluid starts to deteriorate volume fraction, heat loss coefficient, and type of bottom surface,
beyond fV of 1 wt% [76]. such contradictions can be resolved as follows: At high fV , the
- Due to the trace nanoparticle loading required for high-performance increased collector length consistently reduces photothermal con­
in nanofluid-based DASC systems (fV < 0.001), the pumping power version efficiency due to heightened temperatures in the top nano­
needed to circulate the nanofluid does significantly increase. This is a fluid layers, exposed to surroundings, where most light-absorption
consequence of the negligible increase in viscosity and pressure drop occurs. Conversely, at low fV , greater radiation absorption occurs
associated with such dilute nanofluids. within the nanofluid bulk and at the collector bottom surface (in case
of absorptive bottom), promoting thermal trapping and leading to a

54
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

rise in photothermal conversion efficiency with increasing collector - Nanofluid-based DASC systems present a cost-effective solution over
length. For moderate fV , the impact of collector length on efficiency the long term, despite the high initial costs of nanofluids. The use of
exhibits a rise followed by a drop, resulting from the combined ef­ nanofluids leads to more efficient and compact solar collectors,
fects of elevated top layer temperatures and thermal trapping. which lowers both the embodied energy and land requirements. This
- DASCs with thicker nanofluid films consistently exhibit higher efficiency gain not only compensates for the initial expense but also
thermal efficiencies, regardless of the fV value, because when DASC plays a crucial role in reducing CO2 emissions, aligning with the
height increases, greater sunlight attenuation occurs as light crosses goals of sustainable and eco-friendly energy. Additionally, the
the nanofluid film and translates into absorption, temperature, and enhanced efficiency reduces the need for manufacturing materials,
photothermal conversion efficiency increases. However, the surface thereby diminishing environmental pollution. Research supports
type of the bottom wall plays a main role in revealing the exact that the broader adoption of solar energy, facilitated by nanofluids,
behavior. Thinner DASCs with a fully absorptive bottom suffer from could significantly magnify these ecological advantages, high­
higher thermal losses, compared to thicker DASCs, due to higher top- lighting their value in both the economic and environmental aspects
layer temperatures. Thinner DASCs with fully reflective or trans­ of solar technology. On the other hand, nanoparticles raise envi­
parent bottoms and a low-to-moderate fV suffer from higher radiative ronmental and health concerns due to increased toxicity. These sub-
losses, compared to thicker DASCs, due to the short light travel dis­ 10nm particles act like gases, entering the body and potentially
tance through the nanofluid and thus weak absorption. However, at harming cells. Nanoparticle shape also influences toxicity, impacting
higher fV , the nanofluid allows more efficient heat transfer from hot ecosystems and ultimately human health. A major challenge lies in
top to cold bottom layers. the lack of established protocols for safe handling and disposal of
- Different novel innovative DASC designs offered in the literature these increasingly prevalent nanoparticles, necessitating robust
confirm that there is room for dramatic development in DASC toxicity investigations for safe manufacturing practices.
technology by proposing new designs with heightened performance
and mitigated losses and pipeline blockages, and that can be valid for Research gaps and future directions:
different engineering systems. Modifications, including the position
of incident radiation, shape and surface type of the bottom wall, - The Eulerian-Lagrangian approach can be expanded to involve the
employing auxiliaries such as porous media, or magnetic fields, kinetics of particle clustering in both static and flowing nanofluids.
implementing an annular pipe, as explained earlier, using thermal Accordingly, the approach will be capable of forecasting the collision
insulation layers [145], etc., can lead to remarkably higher frequency, attachment efficiency (probability of two particles stick­
performance. ing together upon collision), and colloidal stability in nanofluid-
- Low nanofluid flow rates result in a low DASC’s photothermal con­ based systems. The implementation of such an extension can be
version efficiency due to the longer residence (i.e., exposure to ra­ performed through a set of key steps. Firstly, a hard sphere collision
diation) time inside DASCs which yields high temperatures and thus model, identifying a critical particle velocity for clustering based on
high radiative and convective losses. In contrast, high flow rates momentum/energy balance pre- and post-impact, should be
result in a higher efficiency. Understanding these effects is crucial for employed. Thus, the model can determine if the collision will yield
optimizing DASC design making a trade-off between efficiency and rebound or clustering, relying on the relative kinetic and attractive/
temperature. repulsive potential energies between particles. Secondly, the char­
- Investigations into the effect of nanofluid thermal conductivity on acteristics and dynamics of clusters as well as the particle-cluster and
DASC performance yielded several key findings. First, at low nano­ cluster-cluster interactions should be identified. Lastly, modified
particle volume fractions, enhancing nanofluid thermal conductivity particle momentum and energy source terms will be implemented to
had a small negative effect due to the dominance of volumetric ab­ account for the particle size distribution impacts on the continuous
sorption of solar radiation. Conversely, at high nanoparticle volume phase.
fractions, improved thermal conductivity had a positive impact, - The mathematical model of DASC systems can be extended to
playing a crucial role in propagating absorbed energy within the accommodate high-flux scenarios through including an additional
nanofluid volume, where surface absorption dominated. Notably, term to the RTE to account for emission augmentation, taking into
changes in nanofluid thermal conductivity had minimal impact on account the blackbody radiation from the collector cover plate to the
energy efficiency when using base fluids with inherently high ther­ working fluid, and considering both the radiative and convective
mal conductivities (e.g., water). However, DASCs employing nano­ thermal losses from the collector upper plate.
fluids composed of base fluids with relatively low thermal - The development of more accurate mathematical models to predict
conductivities (e.g., therminol VP-1) were more sensitive to such the optical properties of nanofluids or extending the existing ones to
variations. accommodate the impacts of the multiple scattering, agglomeration,
- Although Re does not have a direct relation with nanofluid extinction clustering of particles, medium temperature, and dynamic features in
coefficient, increasing Re reduces the nanofluid’s radiation extinc­ nanofluids are essential.
tion ability at the vicinity of the DASC upper wall as high Re affects - Despite the usefulness of DLVO theory, it falls short when it comes to
the uniformity of nanoparticle distribution. At high Re, a traverse fully describing the intricate interactions within nanofluids. A multi-
nanoparticle migration from upper and lower walls, where shear scale approach tackles this challenge by merging the strengths of two
rates are high, to low shear rate regions is observed, which dilutes the simulation techniques: Atomistic simulations (MD) and system-level
nanoparticle concentration near DASC top and bottom walls. simulations (DPD, Eulerian-Lagrangian). MD simulations excel at
Consequently, the major light absorption and thus temperature calculating forces between individual nanoparticles at the nanoscale,
elevation occur in the nanofluid middle bulk rather than top layers, which is vital for understanding stability. However, MD becomes
which mitigates losses to surroundings and enhances the photo­ computationally demanding as the number of particles increases.
thermal conversion efficiency. At low Re, the impact of shear- System-level simulations, on the other hand, can handle larger par­
induced migration is counteracted by Brownian diffusion, leading ticle numbers and model complex interactions like colloidal behavior
to nanoparticle homogenization within the collector. Conversely, in and thermal effects, but struggle with capturing the finer details. The
high Re (i.e., convection-dominated) flows, the weaker influence of multi-scale approach bridges this gap by utilizing MD for accurate
Brownian diffusion in comparison to flow inertia intensifies nano­ force calculations, which are then incorporated into system-level
particle migration. simulations. This combined approach provides a more complete
picture of nanofluid behavior, ultimately aiding in the design of

55
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

stable solar nanofluids and predicting their performance in practical CRediT authorship contribution statement
applications.
- Hybrid plasmonic nanofluids combine nanoparticles of varying ma­ Alabas Hasan: Writing – original draft, Visualization, Software,
terials, shapes, and/or sizes to enhance the broad-spectrum photo­ Methodology, Investigation, Data curation. Anas Alazzam: Writing –
thermal conversion of solar radiation. Blending different types of review & editing, Writing – original draft, Visualization, Supervision,
nanoparticles in specific proportions generate multiple LSPR peaks Resources, Project administration, Funding acquisition, Conceptualiza­
aligning with the solar spectrum. For effective design of hybrid tion. Eiyad Abu-Nada: Writing – review & editing, Writing – original
nanofluids, it is necessary to extend the mathematical model by draft, Visualization, Supervision, Resources, Project administration,
incorporating interparticle interactions among diverse particle types. Funding acquisition, Conceptualization.
This necessitates introducing multiple equations for particle mo­
mentum and energy, along with integrating multiple particle source Declaration of competing interest
terms that are interconnected with the fluid phase. Additionally, an
expansion of the classical Mie theory becomes essential to assess the The authors declare that they have no known competing financial
optical properties of non-spherical nanoparticles. interests or personal relationships that could have appeared to influence
- An in-depth economic assessment of the nanofluid-based DASCs the work reported in this paper.
seems as a critical knowledge gap. Cost-benefit and life-cycle ana­
lyses of DASCs should be necessarily undertaken to evaluate the Data availability
feasibility of the DASC technology transition from the lab to the
commercial large-scale production. Additionally, the environmental Data will be made available on request.
implications of DASCs in terms of carbon mitigation potential and
harmful-emission reduction should be thoroughly studied. Acknowledgment
- Although nanofluid-based DASCs must be extensively tested under
real-world conditions to validate their reliability alleged in theory, a Our sincere gratitude extends to ASPIRE for their generous funding
very few studies based on field testing exist in the literature. (AARE20-358) to support this work.
- Standardization of the nanofluid-based DASC technology by devel­
oping industry protocols and standards including roadmaps for References
successful selection of the nanoparticle material type, size, shape,
and loading, the compatible base fluid, the DASC geometry di­ [1] Qiu G, Yu S, Cai W. A novel heating strategy and its optimization of a solar
mensions, the flow properties, etc., can guarantee high and consis­ heating system for a commercial building in term of economy. Energy 2021;221:
119773.
tent performance and encourage and facilitate commercialization [2] Shahsavari A, Akbari M. Potential of solar energy in developing countries for
and widespread adoption. reducing energy-related emissions. Renew Sustain Energy Rev 2018;90:275–91.
- Although recent studies suggested magnetic and electric fields to [3] Council WE. World energy resources. London World Energy Counc; 2013.
[4] Conti J, Holtberg P, Diefenderfer J, LaRose A, Turnure JT, Westfall L.
support the nanoparticle dispersion and stability in DASCs, offering International energy outlook 2016 with projections to 2040. Washington, DC
practical techniques, that support the stability and uniform distri­ (United States: USDOE Energy Information Administration (EIA); 2016.
bution of nanoparticles and keep them from agglomeration, is still a [5] Cozzi L, et al. World energy outlook 2020, vol. 2050. Fr: IEA Paris; 2020.
p. 1–461.
crucial challenge. Such techniques should guarantee efficient solar [6] Breyer C, et al. On the role of solar photovoltaics in global energy transition
absorption and volumetric heat generation in DASCs. scenarios. Prog Photovoltaics Res Appl 2017;25(8):727–45.
- The base fluid type, nanoparticle material type, size, shape, and [7] Clerici A. World energy council. World energy resources 2013 survey. 2019.
[8] Emiliano B. Dubai: tariff for large-scale PV hits new low at $0.024/kWh. PV Mag.;
loading, flow velocity, collector length, collector height, collector
2018.
bottom surface type, etc. are interrelated parameters, and their col­ [9] Moghaieb HS, Amendola V, Khalil S, Chakrabarti S, Maguire P, Mariotti D.
lective impacts on the optical and thermal performance of DASCs Nanofluids for direct-absorption solar collectors—DASCs: a review on recent
progress and future perspectives. Nanomaterials 2023;13(7):1232.
require further examination. Also, the impact of a specific parameter
[10] Green MA, Emery K, Hishikawa Y, Warta W, Dunlop ED. Solar cell efficiency
on the performance of a DASC may vary from one study to another tables (Version 45). Prog Photovoltaics Res Appl 2015;23(1):1–9. https://doi.
due to differences in other parameters. org/10.1002/pip.2573.
- For example, collector length will not influence two DASCs with two [11] Han D, Meng Z, Wu D, Zhang C, Zhu H. Thermal properties of carbon black
aqueous nanofluids for solar absorption. Nanoscale Res Lett 2011;6:1–7.
different fV values the same way, as detailed earlier. Such in­ [12] Abutayeh M, Goswami YD, Stefanakos EK. Solar thermal power plant simulation.
consistencies may create a perception that results of the different Environ Prog Sustain Energy 2013;32(2):417–24.
studies are not in agreement with each other, which means that a [13] Herrando M, et al. A review of solar hybrid photovoltaic-thermal (PV-T)
collectors and systems. Prog Energy Combust Sci 2023;97:101072.
collective impact should be considered when comparing results of [14] Qu J, Zhang R, Wang Z, Wang Q. Photo-thermal conversion properties of hybrid
different investigations. CuO-MWCNT/H2O nanofluids for direct solar thermal energy harvest. Appl
- A viable future path is integrating DASCs with current grid infra­ Therm Eng 2019;147:390–8.
[15] Abutayeh M, Li C, Goswami DY, Stefanakos EK. Solar desalination. Desalin. Water
structure and solar energy systems. Combining DASCs with other from water 2014:549–81.
renewable energy sources to create a hybrid system has the potential [16] Ge TS, et al. Solar heating and cooling: present and future development. Renew
to increase system reliability and total energy production. Energy 2018;126:1126–40.
[17] Sukhatme SP, Sukhatme SP. Solar energy: principles of thermal collection and
- As for the erosion and corrosion threats, employing nanoparticles
storage. 1996.
introduces a significant problem. Researchers, as detailed earlier, [18] Goel N, Taylor RA, Otanicar T. A review of nanofluid-based direct absorption
investigated the corrosion of metal and metal-oxide nanoparticles solar collectors: design considerations and experiments with hybrid PV/Thermal
and direct steam generation collectors. Renew Energy 2020;145:903–13.
and found that Cu and Al pipes significantly degraded whereas
[19] Kalogirou SA. Solar thermal collectors and applications. Prog Energy Combust Sci
stainless steel pipes were unaffected. As a result, stainless steel 2004;30(3):231–95.
components are advised for the hydraulic circuits of nanofluid-based [20] Farjana SH, Huda N, Mahmud MAP, Saidur R. Solar process heat in industrial
DASCs. According to various researchers, corrosion is influenced by systems–A global review. Renew Sustain Energy Rev 2018;82:2270–86.
[21] Sabiha MA, Saidur R, Mekhilef S, Mahian O. Progress and latest developments of
the type of nanoparticles, with carbon-based nanofluids showing less evacuated tube solar collectors. Renew Sustain Energy Rev 2015;51:1038–54.
corrosion potential than metal- or metal oxide-based nanofluids [22] Abutayeh M, et al. Effect of short cloud shading on the performance of parabolic
[292,392,393]. This adds one more advantage to the carbon-based trough solar power plants: motorized vs manual valves. Renew Energy 2019;142:
330–44.
nanoparticles besides their superior OTPs. [23] Rungasamy AE, Craig KJ, Meyer JP. A review of linear Fresnel primary optical
design methodologies. Sol Energy 2021;224:833–54.

56
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

[24] Kumar KH, Daabo AM, Karmakar MK, Hirani H. Solar parabolic dish collector for [60] Karami M, Akhavan-Bahabadi MA, Delfani S, Raisee M. Experimental
concentrated solar thermal systems: a review and recommendations. Environ Sci investigation of CuO nanofluid-based direct absorption solar collector for
Pollut Res 2022;29(22):32335–67. residential applications. Renew Sustain Energy Rev 2015;52:793–801.
[25] Chen J, et al. Experiment and dynamic simulation of a solar tower collector [61] Vakili M, Hosseinalipour SM, Delfani S, Khosrojerdi S, Karami M. Experimental
system for power generation. Renew Energy 2022;196:946–58. investigation of graphene nanoplatelets nanofluid-based volumetric solar
[26] Reddy VS, Kaushik SC, Ranjan KR, Tyagi SK. State-of-the-art of solar thermal collector for domestic hot water systems. Sol Energy 2016;131:119–30.
power plants—a review. Renew Sustain Energy Rev 2013;27:258–73. [62] Gupta HK, Das Agrawal G, Mathur J. Investigations for effect of Al2O3–H2O
[27] Bhowmik H, Amin R. Efficiency improvement of flat plate solar collector using nanofluid flow rate on the efficiency of direct absorption solar collector. Case
reflector. Energy Rep 2017;3:119–23. Stud Therm Eng 2015;5:70–8.
[28] Lenert A, Wang EN. Optimization of nanofluid volumetric receivers for solar [63] Gorji TB, Ranjbar AA. A numerical and experimental investigation on the
thermal energy conversion. Sol Energy 2012;86(1):253–65. performance of a low-flux direct absorption solar collector (DASC) using graphite,
[29] Vishal G, Chinmay V, Kishor R. Solar water heating systems: a review. Int J Sci magnetite and silver nanofluids. Sol Energy 2016;135:493–505.
Eng Res 2015;3(4):13–7. [64] Said Z, Hachicha AA, Aberoumand S, Yousef BAA, Sayed ET, Bellos E. Recent
[30] Minardi JE, Chuang HN. Performance of a ‘black’ liquid flat-plate solar collector. advances on nanofluids for low to medium temperature solar collectors: energy,
Sol Energy 1975;17(3):179–83. exergy, economic analysis and environmental impact. Prog Energy Combust Sci
[31] Raj P, Subudhi S. A review of studies using nanofluids in flat-plate and direct 2021;84:100898.
absorption solar collectors. Renew Sustain Energy Rev 2018;84:54–74. [65] Hamzat AK, Omisanya MI, Sahin AZ, Oyetunji OR, Olaitan NA. Application of
[32] Sainz-Mañas M, Bataille F, Caliot C, Vossier A, Flamant G. Direct absorption nanofluid in solar energy harvesting devices: a comprehensive review. Energy
nanofluid-based solar collectors for low and medium temperatures. A review. Convers Manag 2022;266:115790.
Energy 2022;260:124916. [66] Aissa A, et al. A review of the enhancement of solar thermal collectors using
[33] Karami M, Bozorgi M, Delfani S, Akhavan-Behabadi MA. Empirical correlations nanofluids and turbulators. Appl Therm Eng 2023;220:119663.
for heat transfer in a silver nanofluid-based direct absorption solar collector. [67] Hussain M, Shah SKH, Sajjad U, Abbas N, Ali A. Recent developments in optical
Sustain Energy Technol Assessments 2018;28:14–21. and thermal performance of direct absorption solar collectors. Energies 2022;15
[34] Otanicar TP, Phelan PE, Prasher RS, Rosengarten G, Taylor RA. Nanofluid-based (19):7101.
direct absorption solar collector. J Renew Sustain Energy 2010;2(3):33102. [68] Otanicar T, Taylor RA, Phelan PE, Prasher R. Impact of size and scattering mode
[35] Hunt AJ. Small particle heat exchangers. Berkeley (USA): California Univ.; 1978. on the optimal solar absorbing nanofluid. Energy Sustainability 2009:791–6.
Lawrence Berkeley Lab. [69] Gorji TB, Ranjbar AA. A review on optical properties and application of
[36] Tembhare SP, Barai DP, Bhanvase BA. Performance evaluation of nanofluids in nanofluids in direct absorption solar collectors (DASCs). Renew Sustain Energy
solar thermal and solar photovoltaic systems: a comprehensive review. Renew Rev 2017;72:10–32.
Sustain Energy Rev 2022;153:111738. [70] Incropera FP, Dewitt DP, Bergman TL, Lavine AS. Principles of heat and mass
[37] Masuda H, Ebata A, Teramae K. Alteration of thermal conductivity and viscosity transfer. 2013 (No Title).
of liquid by dispersing ultra-fine particles. Dispersion of Al2O3, SiO2 and TiO2 [71] Modest MF, Mazumder S. Radiative heat transfer. Academic press; 2021.
ultra-fine particles. 1993. [72] Lockwood DJ. Rayleigh and Mie scattering. Encycl. Color Sci. Technol. 2016:
[38] Choi SUS, Eastman JA. Enhancing thermal conductivity of fluids with 1097–107.
nanoparticles. Argonne, IL (United States): Argonne National Lab.(ANL); 1995. [73] Bohren CF, Huffman DR. Absorption and scattering of light by small particles.
[39] Wang X-Q, Mujumdar AS. Heat transfer characteristics of nanofluids: a review. Int John Wiley & Sons; 2008.
J Therm Sci 2007;46(1):1–19. [74] Taylor R. Thermal energy conversion in nanofluids. Arizona State University;
[40] Qiu L, et al. A review of recent advances in thermophysical properties at the 2011.
nanoscale: from solid state to colloids. Phys Rep 2020;843:1–81. [75] Ivezić Z, Mengüç MP. An investigation of dependent/independent scattering
[41] Lee BJ, Park K, Walsh T, Xu L. Radiative heat transfer analysis in plasmonic regimes using a discrete dipole approximation. Int J Heat Mass Tran 1996;39(4):
nanofluids for direct solar thermal absorption. 2012. 811–22.
[42] Kreibig U, Vollmer M. Optical properties of metal clusters, vol. 25. Springer [76] Zhang H, Chen H-J, Du X, Wen D. Photothermal conversion characteristics of gold
Science & Business Media; 2013. nanoparticle dispersions. Sol Energy 2014;100:141–7.
[43] Abdin Z, et al. Solar energy harvesting with the application of nanotechnology. [77] Sharaf OZ, Taylor RA, Abu-Nada E. On the colloidal and chemical stability of
Renew Sustain Energy Rev 2013;26:837–52. solar nanofluids: from nanoscale interactions to recent advances. Phys Rep 2020;
[44] Saidur R, Meng TC, Said Z, Hasanuzzaman M, Kamyar A. Evaluation of the effect 867:1–84.
of nanofluid-based absorbers on direct solar collector. Int J Heat Mass Tran 2012; [78] Mojumder JC, Aminossadati SM, Leonardi CR. Performance analysis of a
55(21–22):5899–907. concentrated direct absorption solar collector (DASC) with nanofluids using
[45] Simonetti M, Restagno F, Sani E, Noussan M. Numerical investigation of direct computational fluid dynamics and discrete ordinates radiation modelling (CFD-
absorption solar collectors (DASC), based on carbon-nanohorn nanofluids, for low DORM). Renew Energy 2023;205:30–52.
temperature applications. Sol Energy 2020;195:166–75. [79] Tien C-L, Drolen BL. Thermal radiation in particulate media with dependent and
[46] Kumar S, Tien CL. Analysis of combined radiation and convection in a particulate independent scattering. Annu. Rev. Heat Transf. 1987;1.
laden falling film. In: Heat transfer phenomena in radiation, combustion, and [80] Wei W, Fedorov AG, Luo Z, Ni M. Radiative properties of dense nanofluids. Appl
fires. Publ by ASME; 1989. p. 231–8. Opt 2012;51(25):6159–71.
[47] Miller FJ, Koenigsdorff RW. Thermal modeling of a small-particle solar central [81] Siegel R. Thermal radiation heat transfer. CRC press; 2001.
receiver. J Sol Energy Eng 2000;122(1):23–9. [82] Kerker M, Scheiner P, Cooke DD. The range of validity of the Rayleigh and
[48] Bertocchi R, Karni J, Kribus A. Experimental evaluation of a non-isothermal high Thomson limits for Lorenz-Mie scattering. JOSA 1978;68(1):135–7.
temperature solar particle receiver. Energy 2004;29(5–6):687–700. [83] Kumar S, Tien CL. Dependent absorption and extinction of radiation by small
[49] Ladjevardi SM, Asnaghi A, Izadkhast PS, Kashani AH. Applicability of graphite particles. 1990.
nanofluids in direct solar energy absorption. Sol Energy 2013;94:327–34. [84] Chen M, He Y, Ye Q, Wang X, Hu Y. Shape-dependent solar thermal conversion
[50] Prasher R. Modification of Planck black body emissive power and intensity in properties of plasmonic Au nanoparticles under different light filter conditions.
particulate media due to multiple and dependent scattering. Heat Transfer Sol Energy 2019;182:340–7.
Summer Conference 2005:163–73. [85] Sharaf OZ, Al-Khateeb AN, Kyritsis DC, Abu-Nada E. Direct absorption solar
[51] Prasher RS, Phelan PE. Modeling of radiative and optical behavior of nanofluids collector (DASC) modeling and simulation using a novel Eulerian-Lagrangian
based on multiple and dependent scattering theories. In: ASME international hybrid approach: optical, thermal, and hydrodynamic interactions. Appl Energy
mechanical engineering congress and exposition; 2005. p. 739–43. 2018;231:1132–45.
[52] Prasher R, Bhattacharya P, Phelan PE. Thermal conductivity of nanoscale [86] Jing D, Song D. Optical properties of nanofluids considering particle size
colloidal solutions (nanofluids). Phys Rev Lett 2005;94(2):25901. distribution: experimental and theoretical investigations. Renew Sustain Energy
[53] Bhattacharya P, Saha SK, Yadav A, Phelan PE, Prasher RS. Brownian dynamics Rev 2017;78:452–65.
simulation to determine the effective thermal conductivity of nanofluids. J Appl [87] Taylor RA, Phelan PE, Otanicar TP, Adrian R, Prasher R. Nanofluid optical
Phys 2004;95(11):6492–4. property characterization: towards efficient direct absorption solar collectors.
[54] Prasher R, Bhattacharya P, Phelan PE. Brownian-motion-based convective- Nanoscale Res Lett 2011;6:1–11.
conductive model for the effective thermal conductivity of nanofluids. 2006. [88] Stenzel O. The physics of thin film optical spectra. Springer; 2015.
[55] Krishnamurthy S, Bhattacharya P, Phelan PE, Prasher RS. Enhanced mass [89] Purcell EM, Pennypacker CR. Scattering and absorption of light by nonspherical
transport in nanofluids. Nano Lett 2006;6(3):419–23. dielectric grains. Astrophys J 1973;186:705–14. Vol. 186, pp. 705–714.
[56] Phelan PE, Bhattacharya P, Prasher R. Nanofluids for heat transfer applications. [90] Draine BT. The discrete-dipole approximation and its application to interstellar
Annu. Rev. Heat Transf. 2005;14. graphite grains. Astrophys J 1988;333:848–72. Part 1 (ISSN 0004-637X), vol.
[57] Tyagi H, Phelan P, Prasher R. Predicted efficiency of a low-temperature 333, Oct. 15, 1988, pp. 848–872.
nanofluid-based direct absorption solar collector. J Sol Energy Eng 2009;131(4). [91] Draine BT, Flatau PJ. Discrete-dipole approximation for scattering calculations.
[58] Luo Z, Wang C, Wei W, Xiao G, Ni M. Performance improvement of a nanofluid Josa a 1994;11(4):1491–9.
solar collector based on direct absorption collection (DAC) concepts. Int J Heat [92] Brioude A, Jiang XC, Pileni MP. Optical properties of gold nanorods: DDA
Mass Tran 2014;75:262–71. simulations supported by experiments. J Phys Chem B 2005;109(27):13138–42.
[59] Delfani S, Karami M, Akhavan-Behabadi MA. Performance characteristics of a [93] Jeon J, Park S, Lee BJ. Optical property of blended plasmonic nanofluid based on
residential-type direct absorption solar collector using MWCNT nanofluid. Renew gold nanorods. Opt Express 2014;22(104):A1101–11.
Energy 2016;87:754–64. [94] Gall D. Electron mean free path in elemental metals. J Appl Phys 2016;119(8).

57
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

[95] Westcott SL, Jackson JB, Radloff C, Halas NJ. Relative contributions to the [128] Kasaeian A, Daneshazarian R, Pourfayaz F. Comparative study of different
plasmon line shape of metal nanoshells. Phys Rev B 2002;66(15):155431. nanofluids applied in a trough collector with glass-glass absorber tube. J Mol Liq
[96] Lv W, Phelan PE, Swaminathan R, Otanicar TP, Taylor RA. Multifunctional core- 2017;234:315–23.
shell nanoparticle suspensions for efficient absorption. J Sol Energy Eng 2013;135 [129] Won KH, Lee BJ. Effect of light scattering on the performance of a direct
(2):21004. absorption solar collector. Front Energy 2018;12:169–77.
[97] Otanicar TP, Higgins BT, Brunter S, Phelan PE, Dai L, Swaminathan R. [130] Parvin S, Nasrin R, Alim MA. Heat transfer and entropy generation through
Temperature dependent optical properties of nanoparticle suspensions. In: Heat nanofluid filled direct absorption solar collector. Int J Heat Mass Tran 2014;71:
transfer summer conference. American Society of Mechanical Engineers; 2012. 386–95.
p. 145–50. [131] Taylor RA, et al. Applicability of nanofluids in high flux solar collectors. J Renew
[98] Sajid MH, Said Z, Saidur R, Adikan FRM, Sabri MFM, Rahim NA. A time variant Sustain Energy 2011;3(2):23104.
investigation on optical properties of water based Al2O3 nanofluid. Int Commun [132] Chen X, Chen M, Zhou P. Solar-thermal conversion performance of heterogeneous
Heat Mass Tran 2014;50:108–16. nanofluids. Renew Energy 2022;198:1307–17.
[99] Zhang L, Liu J, He G, Ye Z, Fang X, Zhang Z. Radiative properties of ionic liquid- [133] Bayazitoglu Y, Higenyi J. Higher-order differential equations of radiative transfer:
based nanofluids for medium-to-high-temperature direct absorption solar P3 approximation. AIAA J 1979;17(4):424–31.
collectors. Sol Energy Mater Sol Cells 2014;130:521–8. [134] Kumar S, Majumdar A, Tien CL. The differential-discrete-ordinate method for
[100] Sharaf OZ, Kyritsis DC, Abu-Nada E. Impact of nanofluids, radiation spectrum, solutions of the equation of radiative transfer. 1990.
and hydrodynamics on the performance of direct absorption solar collectors. [135] Zhu Y, Kong W, Fan J, Englmair G, Yuan Y. Numerical investigation on radiative
Energy Convers Manag 2018;156:706–22. heat loss of a direct absorption solar collector using the discrete ordinates
[101] Gan Y, Qiao L. Optical properties and radiation-enhanced evaporation of method. Appl Therm Eng 2023:121560.
nanofluid fuels containing carbon-based nanostructures. Energy Fuels 2012;26 [136] Veeraragavan A, Lenert A, Yilbas B, Al-Dini S, Wang EN. Analytical model for the
(7):4224–30. design of volumetric solar flow receivers. Int J Heat Mass Tran 2012;55(4):
[102] Bashkatov AN, Genina EA. Water refractive index in dependence on temperature 556–64.
and wavelength: a simple approximation. In: Saratov fall meeting 2002: optical [137] Cregan V, Myers TG. Modelling the efficiency of a nanofluid direct absorption
technologies in biophysics and medicine IV; 2003. p. 393–5. SPIE. solar collector. Int J Heat Mass Tran 2015;90:505–14.
[103] Turcotte DL, Schubert G. Geodynamics. Cambridge university press; 2002. [138] Sharaf OZ, Al-Khateeb AN, Kyritsis DC, Abu-Nada E. Energy and exergy analysis
[104] Said Z, Saidur R, Rahim NA. Optical properties of metal oxides based nanofluids. and optimization of low-flux direct absorption solar collectors (DASCs): balancing
Int Commun Heat Mass Tran 2014;59:46–54. power-and temperature-gain. Renew Energy 2019;133:861–72.
[105] Lee S-H, Jang SP. Extinction coefficient of aqueous nanofluids containing multi- [139] Lenert A, Perez Zuniga YS, Wang EN. Nanofluid-based absorbers for high
walled carbon nanotubes. Int J Heat Mass Tran 2013;67:930–5. temperature direct solar collectors. In: International heat transfer conference;
[106] Wu Y, Zhou L, Du X, Yang Y. Optical and thermal radiative properties of 2010. p. 499–508.
plasmonic nanofluids containing core–shell composite nanoparticles for efficient [140] Zhao JM, Liu LH. Radiative transfer equation and solutions. Handb. Therm. Sci.
photothermal conversion. Int J Heat Mass Tran 2015;82:545–54. Eng. Springer Int. Publ. Cham, Switz. 2018:933–78.
[107] Song D, Wang Y, Jing D, Geng J. Investigation and prediction of optical properties [141] Hunter B, Guo Z. Numerical smearing, ray effect, and angular false scattering in
of alumina nanofluids with different aggregation properties. Int J Heat Mass Tran radiation transfer computation. Int J Heat Mass Tran 2015;81:63–74.
2016;96:430–7. [142] Tu J, Inthavong K, Ahmadi G. Computational fluid and particle dynamics in the
[108] Kameya Y, Hanamura K. Enhancement of solar radiation absorption using human respiratory system. Springer Science & Business Media; 2012.
nanoparticle suspension. Sol Energy 2011;85(2):299–307. [143] Michaelides E. Nanofluidics. Springer; 2016.
[109] Otanicar T, Hoyt J, Fahar M, Jiang X, Taylor RA. Experimental and numerical [144] Jeon J, Park S, Lee BJ. Analysis on the performance of a flat-plate volumetric solar
study on the optical properties and agglomeration of nanoparticle suspensions. collector using blended plasmonic nanofluid. Sol Energy 2016;132:247–56.
J Nanoparticle Res 2013;15:1–11. [145] Kaluri R, Vijayaraghavan S, Ganapathisubbu S. Model development and
[110] Jin H, Lin G, Bai L, Amjad M, Bandarra Filho EP, Wen D. Photothermal conversion performance studies of a concentrating direct absorption solar collector. J Sol
efficiency of nanofluids: an experimental and numerical study. Sol Energy 2016; Energy Eng 2015;137(2):21005.
139:278–89. [146] Qin C, Kang K, Lee I, Lee BJ. Optimization of a direct absorption solar collector
[111] Zhu Q, Cui Y, Mu L, Tang L. Characterization of thermal radiative properties of with blended plasmonic nanofluids. Sol Energy 2017;150:512–20.
nanofluids for selective absorption of solar radiation. Int J Thermophys 2013;34: [147] Pak BC, Cho YI. Hydrodynamic and heat transfer study of dispersed fluids with
2307–21. submicron metallic oxide particles. Exp. Heat Transf. an Int. J. 1998;11(2):
[112] Du M, Tang GH. Optical property of nanofluids with particle agglomeration. Sol 151–70.
Energy 2015;122:864–72. [148] Delavari V, Hashemabadi SH. CFD simulation of heat transfer enhancement of
[113] Moreira LM, et al. Thermo-optical properties of silver and gold nanofluids. Al2O3/water and Al2O3/ethylene glycol nanofluids in a car radiator. Appl Therm
J Therm Anal Calorim 2013;114:557–64. Eng 2014;73(1):380–90.
[114] Choi TJ, Jang SP, Kedzierski MA. Effect of surfactants on the stability and solar [149] Xuan Y, Roetzel W. Conceptions for heat transfer correlation of nanofluids. Int J
thermal absorption characteristics of water-based nanofluids with multi-walled Heat Mass Tran 2000;43(19):3701–7.
carbon nanotubes. Int J Heat Mass Tran 2018;122:483–90. [150] Brinkman HC. The viscosity of concentrated suspensions and solutions. J Chem
[115] Song D, Hatami M, Wang Y, Jing D, Yang Y. Prediction of hydrodynamic and Phys 1952;20(4):571.
optical properties of TiO2/water suspension considering particle size distribution. [151] Buongiorno J, et al. A benchmark study on the thermal conductivity of
Int J Heat Mass Tran 2016;92:864–76. nanofluids. J Appl Phys 2009;106(9):94312.
[116] Duan H, Zheng Y, Xu C, Shang Y, Ding F. Experimental investigation on the [152] Bahiraei M. Particle migration in nanofluids: a critical review. Int J Therm Sci
plasmonic blended nanofluid for efficient solar absorption. Appl Therm Eng 2019; 2016;109:90–113.
161:114192. [153] Darzi AAR, Farhadi M, Lavasani AM. Two phase mixture model of nano-enhanced
[117] Mercatelli L, et al. Absorption and scattering properties of carbon nanohorn-based mixed convection heat transfer in finned enclosure. Chem Eng Res Des 2016;111:
nanofluids for direct sunlight absorbers. Nanoscale Res Lett 2011;6(1):1–9. 294–304.
[118] Gimeno-Furio A, Hernandez L, Barison S, Agresti F, Cabaleiro D, Mancin S. [154] Bårdsgård R, Kuzmenkov DM, Kosinski P, V Balakin B. Eulerian CFD model of
Optical characterisation of oxidised carbon nanohorn nanofluids for direct solar direct absorption solar collector with nanofluid. J Renew Sustain Energy 2020;12
energy absorption applications. Sol Energy 2019;191:323–31. (3).
[119] Baffou G, Quidant R. Thermo-plasmonics: using metallic nanostructures as nano- [155] Feng Z-G, Ponton MEC, Michaelides EE, Mao S. Using the direct numerical
sources of heat. Laser Photon Rev 2013;7(2):171–87. simulation to compute the slip boundary condition of the solid phase in two-fluid
[120] Menbari A, Alemrajabi AA. Analytical modeling and experimental investigation model simulations. Powder Technol 2014;265:88–97.
on optical properties of new class of nanofluids (Al2O3–CuO binary nanofluids) [156] Bahiraei M. A comprehensive review on different numerical approaches for
for direct absorption solar thermal energy. Opt Mater 2016;52:116–25. simulation in nanofluids: traditional and novel techniques. J Dispersion Sci
[121] Webb BW, Viskanta R. Analysis of heat transfer and solar radiation absorption in Technol 2014;35(7):984–96.
an irradiated thin, falling molten salt film. 1985. [157] Sharaf OZ, Al-Khateeb AN, Kyritsis DC, Abu-Nada E. Four-way coupling of
[122] Bohn MS, Wang KY. Experiments and analysis on the molten Salt direct particle-wall and colloidal particle-particle interactions in direct absorption solar
absorption receiver concept. 1988. collectors. Energy Convers Manag 2019;195:7–20.
[123] Kumar S, Tien CL. Analysis of combined radiation and convection in a particulate- [158] Liu D, Bu C, Chen X. Development and test of CFD–DEM model for complex
laden liquid film. 1990. geometry: a coupling algorithm for Fluent and DEM. Comput Chem Eng 2013;58:
[124] Miller F, Koenigsdorff R. Theoretical analysis of a high-temperature small-particle 260–8.
solar receiver. Sol Energy Mater 1991;24(1–4):210–21. [159] Kondaraju S, Jin EK, Lee JS. Direct numerical simulation of thermal conductivity
[125] Howell JR, Mengüç MP, Daun K, Siegel R. Thermal radiation heat transfer. CRC of nanofluids: the effect of temperature two-way coupling and coagulation of
press; 2020. particles. Int J Heat Mass Tran 2010;53(5–6):862–9.
[126] Gorji TB, Ranjbar AA. Geometry optimization of a nanofluid-based direct [160] Balachandar S, Liu K, Lakhote M. Self-induced velocity correction for improved
absorption solar collector using response surface methodology. Sol Energy 2015; drag estimation in Euler–Lagrange point-particle simulations. J Comput Phys
122:314–25. 2019;376:160–85.
[127] Moradi A, Sani E, Simonetti M, Francini F, Chiavazzo E, Asinari P. Carbon- [161] Eaton JK. Two-way coupled turbulence simulations of gas-particle flows using
nanohorn based nanofluids for a direct absorption solar collector for civil point-particle tracking. Int J Multiphas Flow 2009;35(9):792–800.
application. J Nanosci Nanotechnol 2015;15(5):3488–95.

58
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

[162] Liu K, Lakhote M, Balachandar S. Self-induced temperature correction for inter- [199] Chen M, He Y, Zhu J, Wen D. Investigating the collector efficiency of silver
phase heat transfer in Euler-Lagrange point-particle simulation. J Comput Phys nanofluids based direct absorption solar collectors. Appl Energy 2016;181:65–74.
2019;396:596–615. [200] Menbari A, Alemrajabi AA, Rezaei A. Heat transfer analysis and the effect of CuO/
[163] Mahian O, et al. Recent advances in modeling and simulation of nanofluid flows- Water nanofluid on direct absorption concentrating solar collector. Appl Therm
Part I: fundamentals and theory. Phys Rep 2019;790:1–48. Eng 2016;104:176–83.
[164] Mahian O, et al. Recent advances in modeling and simulation of nanofluid [201] Toppin-Hector A, Singh H. Development of a nano-heat transfer fluid carrying
flows—Part II: applications. Phys Rep 2019;791:1–59. https://doi.org/10.1016/j. direct absorbing receiver for concentrating solar collectors. Int J Low Carbon
physrep.2018.11.003. Technol 2016;11(2):199–204.
[165] Haile JM. Molecular dynamics simulation: elementary methods. John Wiley & [202] Freedman JP, Wang H, Prasher RS. Analysis of nanofluid-based parabolic trough
Sons, Inc.; 1992. collectors for solar thermal applications. J Sol Energy Eng 2018;140(5):51008.
[166] Frenkel D, Smit B. Understanding molecular simulation: from algorithms to [203] Garg K, Khullar V, Das SK, Tyagi H. Parametric study of the energy efficiency of
applications. Elsevier; 2023. the HDH desalination unit integrated with nanofluid-based solar collector.
[167] Jabbari F, Rajabpour A, Saedodin S. Thermal conductivity and viscosity of J Therm Anal Calorim 2019;135:1465–78.
nanofluids: a review of recent molecular dynamics studies. Chem Eng Sci 2017; [204] Nasrin R, Parvin S, Alim MA. Heat transfer and collector efficiency through a
174:67–81. direct absorption solar collector with radiative heat flux effect. Numer Heat Tran
[168] Lee SL, Saidur R, Sabri MFM, Min TK. Effects of the particle size and temperature 2015;68(8):887–907.
on the efficiency of nanofluids using molecular dynamic simulation. Numer Heat [205] Hatami M, Jing D. Evaluation of wavy direct absorption solar collector (DASC)
Tran 2016;69(9):996–1013. performance using different nanofluids. J Mol Liq 2017;229:203–11.
[169] Wei M, Song Y, Wang Y. Heat transfer of nanofluidics in hydrophilic pores: [206] Kumar S, Chander N, Gupta VK, Kukreja R. Progress, challenges and future
insights from molecular dynamics simulations. Chin J Chem Eng 2016;24(9): prospects of plasmonic nanofluid based direct absorption solar collectors–a state-
1117–21. of-the-art review. Sol Energy 2021;227:365–425.
[170] Toghraie D, Mokhtari M, Afrand M. Molecular dynamic simulation of copper and [207] Das SK, Choi SU, Yu W, Pradeep T. Nanofluids: science and technology. John
platinum nanoparticles Poiseuille flow in a nanochannels. Phys. E Low- Wiley & Sons; 2007.
dimensional Syst. Nanostructures 2016;84:152–61. [208] Dugaria S, Bortolato M, Del Col D. Modelling of a direct absorption solar receiver
[171] Abu-Nada E. Simulation of heat transfer enhancement in nanofluids using using carbon based nanofluids under concentrated solar radiation. Renew Energy
dissipative particle dynamics. Int Commun Heat Mass Tran 2017;85:1–11. 2018;128:495–508.
[172] Abu-Nada E. Dissipative particle dynamics investigation of heat transfer [209] Das PK, Santra AK, Ganguly R. Performance analysis of direct absorption solar
mechanisms in Al2O3-water nanofluid. Int J Therm Sci 2018;123:58–72. collector using multiphase model. In: Proceedings of the 23rd national heat and
[173] Abu-Nada E. Dissipative particle dynamics simulation of natural convection using mass transfer conference and 1st international ISHMT-ASTFE heat and mass
variable thermal properties. Int Commun Heat Mass Tran 2015;69:84–93. transfer conference; 2015. p. 17–20. IHMTC2015.
[174] Xuan Y, Yao Z. Lattice Boltzmann model for nanofluids. Heat Mass Tran 2005;41: [210] Lee S-H, Jang SP. Efficiency of a volumetric receiver using aqueous suspensions of
199–205. multi-walled carbon nanotubes for absorbing solar thermal energy. Int J Heat
[175] Timm K, Kusumaatmaja H, Kuzmin A, Shardt O, Silva G, Viggen E. The lattice Mass Tran 2015;80:58–71.
Boltzmann method: principles and practice. Cham, Switz: Springer Int. Publ. AG; [211] Zhu Y, Li P, Ruan Z, Yuan Y. A model and thermal loss evaluation of a direct-
2016. absorption solar collector under the influence of radiation. Energy Convers Manag
[176] Mohamad AA. Lattice Boltzmann method, vol. 70. Springer; 2011. 2022;251:114933.
[177] Aliu O, Sakidin H, Foroozesh J, Yahya N. Lattice Boltzmann application to [212] Baro RK, Kotecha P, Anandalakshmi R. Multi-objective optimization of nanofluid-
nanofluids dynamics-A review. J Mol Liq 2020;300:112284. based direct absorption solar collector for low-temperature applications. J Build
[178] Wolf-Gladrow DA. Lattice-gas cellular automata and lattice Boltzmann models: an Eng 2023;72:106258.
introduction. Springer; 2004. [213] Hazra SK, Nandi TK. Performance prediction and determination of optimum
[179] He X, Luo L-S. Theory of the lattice Boltzmann method: from the Boltzmann optical property of the nanofluid for energy storage in direct absorption solar
equation to the lattice Boltzmann equation. Phys Rev E 1997;56(6):6811. collectors. Appl Therm Eng 2023;233:121189.
[180] O. Aliu, H. Sakidin, J. Foroozesh, and N. Yahya, Lattice Boltzmann application to [214] Hachicha AA, Said Z. Numerical modeling and multi-objective optimization of
nanofluids dynamics− A review, vol. vol. 300. Elsevier, p. 112284. direct absorption solar collectors using mono and hybrid nanofluids. J Clean Prod
[181] Mishra SC, Lankadasu A, Beronov KN. Application of the lattice Boltzmann 2023:137740.
method for solving the energy equation of a 2-D transient conduction–radiation [215] Lee S-H, Choi TJ, Jang SP. Thermal efficiency comparison: surface-based solar
problem. Int J Heat Mass Tran 2005;48(17):3648–59. receivers with conventional fluids and volumetric solar receivers with nanofluids.
[182] Mondal B, Mishra SC. Simulation of natural convection in the presence of Energy 2016;115:404–17.
volumetric radiation using the lattice Boltzmann method. Numer Heat Tran 2008; [216] Otanicar TP, Phelan PE, Golden JS. Optical properties of liquids for direct
55(1):18–41. absorption solar thermal energy systems. Sol Energy 2009;83(7):969–77.
[183] Nokhosteen A, Sobhansarbandi S. Utilizing lattice Boltzmann method for heat [217] Mehrali M, Ghatkesar MK, Pecnik R. Full-spectrum volumetric solar thermal
transfer analysis in solar thermal systems: a review. Sustain Energy Technol conversion via graphene/silver hybrid plasmonic nanofluids. Appl Energy 2018;
Assessments 2021;46:101264. 224:103–15.
[184] Siavashi M, Ghasemi K, Yousofvand R, Derakhshan S. Computational analysis of [218] Sani E, et al. Carbon nanohorns-based nanofluids as direct sunlight absorbers. Opt
SWCNH nanofluid-based direct absorption solar collector with a metal sheet. Sol Express 2010;18(5):5179–87.
Energy 2018;170:252–62. [219] G. Astm. 173–03: standard tables for reference solar spectral irradiances: direct
[185] Swan JW, Wang G. Rapid calculation of hydrodynamic and transport properties in normal and hemispherical on 37 tilted surface. ASTM Int.; 2003.
concentrated solutions of colloidal particles and macromolecules. Phys Fluids [220] Karami M, Akhavan-Behabadi MA, Dehkordi MR, Delfani S. Thermo-optical
2016;28(1). properties of copper oxide nanofluids for direct absorption of solar radiation. Sol
[186] Liu G-R, Liu MB. Smoothed particle hydrodynamics: a meshfree particle method. Energy Mater Sol Cells 2016;144:136–42.
World scientific; 2003. [221] Large MCJ, McKenzie DR, Large MI. Incoherent reflection processes: a discrete
[187] Ye T, Pan D, Huang C, Liu M. Smoothed particle hydrodynamics (SPH) for approach. Opt Commun 1996;128(4–6):307–14.
complex fluid flows: recent developments in methodology and applications. Phys [222] Lee S, Kim HJ, Kim KH, Jang SP. Extinction coefficient of water-based multi-
Fluids 2019;31(1). walled carbon nanotube nanofluids for application in direct-absorption solar
[188] Liu MB, Liu G. Smoothed particle hydrodynamics (SPH): an overview and recent collectors. Micro & Nano Lett 2014;9(10):635–8.
developments. Arch Comput Methods Eng 2010;17:25–76. [223] Gupta HK, Das Agrawal G, Mathur J. An experimental investigation of a low
[189] Petela R. Exergy of heat radiation. 1964. temperature Al2O3-H2O nanofluid based direct absorption solar collector. Sol
[190] Evola G, Marletta L. Exergy and thermoeconomic optimization of a water-cooled Energy 2015;118:390–6.
glazed hybrid photovoltaic/thermal (PVT) collector. Sol Energy 2014;107:12–25. [224] Struchalin PG, et al. Performance of a tubular direct absorption solar collector
[191] Bejan A, Tsatsaronis G, Moran MJ. Thermal design and optimization. John Wiley with a carbon-based nanofluid. Int J Heat Mass Tran 2021;179:121717.
& Sons; 1995. [225] Tiwari SK, Bystrzejewski M, De Adhikari A, Huczko A, Wang N. Methods for the
[192] Petela R. Exergy of undiluted thermal radiation. Sol Energy 2003;74(6):469–88. conversion of biomass waste into value-added carbon nanomaterials: recent
[193] Press WH. Theoretical maximum for energy from direct and diffuse sunlight. progress and applications. Prog Energy Combust Sci 2022;92:101023.
Nature 1976;264(5588):734–5. [226] Naito M, Yokoyama T, Hosokawa K, Nogi K. Nanoparticle technology handbook.
[194] Chu SX, Liu LH. Analysis of terrestrial solar radiation exergy. Sol Energy 2009;83 Elsevier; 2018.
(8):1390–404. [227] Auffan M, Rose J, Bottero J-Y, V Lowry G, Jolivet J-P, Wiesner MR. Towards a
[195] Kalogirou SA, Karellas S, Braimakis K, Stanciu C, Badescu V. Exergy analysis of definition of inorganic nanoparticles from an environmental, health and safety
solar thermal collectors and processes. Prog Energy Combust Sci 2016;56:106–37. perspective. Nat Nanotechnol 2009;4(10):634–41.
[196] Landsberg PT, Tonge G. Thermodynamics of the conversion of diluted radiation. [228] Babar H, Ali HM. Towards hybrid nanofluids: preparation, thermophysical
J Phys Math Gen 1979;12(4):551. properties, applications, and challenges. J Mol Liq 2019;281:598–633.
[197] Rasih RA, Sidik NAC, Samion S. Numerical investigation of direct absorption solar [229] Shibata T, et al. XAFS studies of gold and silver–gold nanoparticles in aqueous
collector using nanofluids: a review. In: IOP conference series: materials science solutions. J Synchrotron Radiat 2001;8(2):545–7.
and engineering. IOP Publishing; 2019, 12059. [230] Abid N, et al. Synthesis of nanomaterials using various top-down and bottom-up
[198] Karami M, Raisee M, Delfani S. Numerical investigation of nanofluid-based solar approaches, influencing factors, advantages, and disadvantages: a review. Adv
collectors. In: IOP conference series: materials science and engineering. IOP Colloid Interface Sci 2022;300:102597.
Publishing; 2014, 12044.

59
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

[231] Wang JJ, Zheng RT, Gao JW, Chen G. Heat conduction mechanisms in nanofluids [263] Guo J, et al. Enhanced optical properties and light-to-heat conversion
and suspensions. Nano Today 2012;7(2):124–36. performance of Ti3C2/[BMIM] BF4 nanofluids based direct absorption solar
[232] Chiam HW, Azmi WH, Usri NA, Mamat R, Adam NM. Thermal conductivity and collector. Sol Energy Mater Sol Cells 2022;237:111558.
viscosity of Al2O3 nanofluids for different based ratio of water and ethylene [264] Gorji TB, Ranjbar AA. Thermal and exergy optimization of a nanofluid-based
glycol mixture. Exp Therm Fluid Sci 2017;81:420–9. direct absorption solar collector. Renew Energy 2017;106:274–87.
[233] Radkar RN, Bhanvase BA, Barai DP, Sonawane SH. Intensified convective heat [265] Cao P, et al. Role of base fluid on enhancement absorption properties of Fe3O4/
transfer using ZnO nanofluids in heat exchanger with helical coiled geometry at ionic liquid nanofluids for direct absorption solar collector. Sol Energy 2019;194:
constant wall temperature. Mater. Sci. Energy Technol. 2019;2(2):161–70. 923–31.
[234] Hekmatipour F, Akhavan-Behabadi MA, Sajadi B. Combined free and forced [266] Hwang YJ, et al. Investigation on characteristics of thermal conductivity
convection heat transfer of the copper oxide-heat transfer oil (CuO-HTO) enhancement of nanofluids. Curr Appl Phys 2006;6(6):1068–71.
nanofluid inside horizontal tubes under constant wall temperature. Appl Therm [267] Gao Y, Wang H, Sasmito AP, Mujumdar AS. Measurement and modeling of
Eng 2016;100:621–7. thermal conductivity of graphene nanoplatelet water and ethylene glycol base
[235] Sarode HA, Barai DP, Bhanvase BA, Ugwekar RP, Saharan V. Investigation on nanofluids. Int J Heat Mass Tran 2018;123:97–109.
preparation of graphene oxide-CuO nanocomposite based nanofluids with the aid [268] Menni Y, Chamkha AJ, Massarotti N, Ameur H, Kaid N, Bensafi M. Hydrodynamic
of ultrasound assisted method for intensified heat transfer properties. Mater Chem and thermal analysis of water, ethylene glycol and water-ethylene glycol as base
Phys 2020;251:123102. fluids dispersed by aluminum oxide nano-sized solid particles. Int J Numer
[236] Bandarra Filho EP, Mendoza OSH, Beicker CLL, Menezes A, Wen D. Experimental Methods Heat Fluid Flow 2020;30(9):4349–86.
investigation of a silver nanoparticle-based direct absorption solar thermal [269] Gholinia M, Hosseinzadeh K, Ganji DD. Investigation of different base fluids
system. Energy Convers Manag 2014;84:261–7. suspend by CNTs hybrid nanoparticle over a vertical circular cylinder with
[237] Wen J, Li X, Chen W, Liu J. Systematical investigation on the solar-thermal sinusoidal radius. Case Stud Therm Eng 2020;21:100666.
conversion performance of TiN plasmonic nanofluids for the direct absorption [270] Abbas N, et al. Applications of nanofluids in photovoltaic thermal systems: a
solar collectors. Colloids Surfaces A Physicochem. Eng. Asp. 2021;624:126837. review of recent advances. Phys. A Stat. Mech. its Appl. 2019;536:122513.
[238] Chaturvedi KR, Narukulla R, Amani M, Sharma T. Experimental investigations to [271] Younes H, Mao M, Murshed SMS, Lou D, Hong H, Peterson GP. Nanofluids: key
evaluate surfactant role on absorption capacity of nanofluid for CO2 utilization in parameters to enhance thermal conductivity and its applications. Appl Therm Eng
sustainable crude mobilization. Energy 2021;225:120321. 2022;207:118202.
[239] Cosgrove T. Colloid science: principles, methods and applications. John Wiley & [272] Sinha K, Kavlicoglu B, Liu Y, Gordaninejad F, Graeve OA. A comparative study of
Sons; 2010. thermal behavior of iron and copper nanofluids. J Appl Phys 2009;106(6):64307.
[240] Shvets AA. Theory of colloidal stabilization by unattached polymers. 2020. arXiv [273] Campos C, Vasco D, Angulo C, Burdiles PA, Cardemil J, Palza H. About the
Prepr. arXiv2010.08110. relevance of particle shape and graphene oxide on the behavior of direct
[241] Lattuada M, Sandkühler P, Wu H, Sefcik J, Morbidelli M. Aggregation kinetics of absorption solar collectors using metal based nanofluids under different radiation
polymer colloids in reaction limited regime: experiments and simulations. Adv intensities. Energy Convers Manag 2019;181:247–57.
Colloid Interface Sci 2003;103(1):33–56. [274] Xing L, Wang R, Ha Y, Li Z. Absorption characteristics and solar thermal
[242] Kim HJ, Lee S-H, Lee J-H, Jang SP. Effect of particle shape on suspension stability conversion of Fe3O4@ Au core/shell nanoparticles for a direct-absorption solar
and thermal conductivities of water-based bohemite alumina nanofluids. Energy collector. Renew Energy 2023;216:119120.
2015;90:1290–7. [275] Sun C, et al. Solar absorption characteristics of SiO2@ Au core-shell composite
[243] Loza K, Epple M, Maskos M. Stability of nanoparticle dispersions and particle nanorods for the direct absorption solar collector. Renew Energy 2022;189:
agglomeration. Biol. Responses to Nanoscale Part. Mol. Cell. Asp. Methodol. 402–11.
Approaches 2019:85–100. [276] Sharaf OZ, et al. Ultrastable plasmonic nanofluids in optimized direct absorption
[244] Chakraborty S, Panigrahi PK. Stability of nanofluid: a review. Appl Therm Eng solar collectors. Energy Convers Manag 2019;199:112010.
2020;174:115259. [277] Agarwal R, Verma K, Agrawal NK, Singh R. Sensitivity of thermal conductivity for
[245] Chen L, Liu J, Fang X, Zhang Z. Reduced graphene oxide dispersed nanofluids Al2O3 nanofluids. Exp Therm Fluid Sci 2017;80:19–26.
with improved photo-thermal conversion performance for direct absorption solar [278] Prado JI, Vallejo JP, Lugo L. A new relationship on transport properties of
collectors. Sol Energy Mater Sol Cells 2017;163:125–33. nanofluids. Evidence with novel magnesium oxide based n-tetradecane
[246] Chen L, Xu C, Liu J, Fang X, Zhang Z. Optical absorption property and photo- nanodispersions. Powder Technol 2022;397:117082.
thermal conversion performance of graphene oxide/water nanofluids with [279] Khatri R, Jangid R, Tomar PS, Sharma SS. Experimental investigations on direct
excellent dispersion stability. Sol Energy 2017;148:17–24. absorption solar flat plate collector using Al2O3 nanofluid. Int J Nanotechnol
[247] Ali N, Teixeira JA, Addali A. A review on nanofluids: fabrication, stability, and 2021;18(11–12):968–79.
thermophysical properties. J Nanomater 2018;2018. [280] Thakur PP, Khapane TS, Sonawane SS. Comparative performance evaluation of
[248] Li D, Hong B, Fang W, Guo Y, Lin R. Preparation of well-dispersed silver fly ash-based hybrid nanofluids in microchannel-based direct absorption solar
nanoparticles for oil-based nanofluids. Ind Eng Chem Res 2010;49(4):1697–702. collector. J Therm Anal Calorim 2021;143:1713–26.
[249] He YT, Wan J, Tokunaga T. Kinetic stability of hematite nanoparticles: the effect [281] Bhalla V, Khullar V, Singh H, Tyagi H. Solar thermal energy: use of volumetric
of particle sizes. J Nanoparticle Res 2008;10:321–32. absorption in domestic applications. Appl Sol Energy 2018:99–112.
[250] Vold MJ. Van der Waals’ attraction between anisometric particles. J Colloid Sci [282] Milanese M, Colangelo G, Creti A, Lomascolo M, Iacobazzi F, De Risi A. Optical
1954;9(5):451–9. absorption measurements of oxide nanoparticles for application as nanofluid in
[251] Chakraborty S, Sarkar I, Ashok A, Sengupta I, Pal SK, Chakraborty S. Thermo- direct absorption solar power systems–Part I: water-based nanofluids behavior.
physical properties of Cu-Zn-Al LDH nanofluid and its application in spray Sol Energy Mater Sol Cells 2016;147:315–20.
cooling. Appl Therm Eng 2018;141:339–51. [283] V Balakin B, Stava M, Kosinska A. Photothermal convection of a magnetic
[252] Al-Gebory L, Mengüç MP, Koşar A, Şendur K. Effect of electrostatic stabilization nanofluid in a direct absorption solar collector. Sol Energy 2022;239:33–9.
on thermal radiation transfer in nanosuspensions: photo-thermal energy [284] Bhalla V, Khullar V, Tyagi H. Investigation of factors influencing the performance
conversion applications. Renew Energy 2018;119:625–40. of nanofluid-based direct absorption solar collector using Taguchi method.
[253] Kumar DD, Arasu AV. A comprehensive review of preparation, characterization, J Therm Anal Calorim 2019;135:1493–505.
properties and stability of hybrid nanofluids. Renew Sustain Energy Rev 2018;81: [285] Yu S-P, Teng T-P, Huang C-C, Hsieh H-K, Wei Y-J. Performance evaluation of
1669–89. carbon-based nanofluids for direct absorption solar collector. Energies 2023;16
[254] Fritz G, Schädler V, Willenbacher N, Wagner NJ. Electrosteric stabilization of (3):1157.
colloidal dispersions. Langmuir 2002;18(16):6381–90. [286] Hazra SK, Ghosh S, Nandi TK. Photo-thermal conversion characteristics of carbon
[255] Tadros TF. Interfacial phenomena and colloid stability: basic principles, vol. 1. black-ethylene glycol nanofluids for applications in direct absorption solar
Walter de Gruyter GmbH & Co KG; 2015. collectors. Appl Therm Eng 2019;163:114402.
[256] Hjerrild NE, Scott JA, Amal R, Taylor RA. Exploring the effects of heat and UV [287] Barai DP, Bhanvase BA, Sonawane SH. A review on graphene derivatives-based
exposure on glycerol-based Ag-SiO2 nanofluids for PV/T applications. Renew nanofluids: investigation on properties and heat transfer characteristics. Ind Eng
Energy 2018;120:266–74. Chem Res 2020;59(22):10231–77.
[257] Taylor R, et al. Small particles, big impacts: a review of the diverse applications of [288] Ibrahim M, Saeed T, Chu Y-M, Ali HM, Cheraghian G, Kalbasi R. Comprehensive
nanofluids. J Appl Phys 2013;113(1). study concerned graphene nano-sheets dispersed in ethylene glycol: experimental
[258] Faure B, et al. Dispersion and surface functionalization of oxide nanoparticles for study and theoretical prediction of thermal conductivity. Powder Technol 2021;
transparent photocatalytic and UV-protecting coatings and sunscreens. Sci 386:51–9.
Technol Adv Mater 2013;14(2):23001. [289] Vakili M, Yahyaei M, Ramsay J, Aghajannezhad P, Paknezhad B. Adaptive neuro-
[259] Sauter C, Schuchmann HP. High pressure for dispersing and deagglomerating fuzzy inference system modeling to predict the performance of graphene
nanoparticles in aqueous solutions. Chem. Eng. Technol. Ind. Chem. Equipment- nanoplatelets nanofluid-based direct absorption solar collector based on
Process Eng. 2007;30(10):1401–5. experimental study. Renew Energy 2021;163:807–24.
[260] Karami M, Bozorgi M, Delfani S. Effect of design and operating parameters on [290] Kumar PG, et al. Exploring the photo-thermal conversion behavior and extinction
thermal performance of low-temperature direct absorption solar collectors: a coefficient of activated carbon nanofluids for direct absorption solar collector
review. J Therm Anal Calorim 2021;146:993–1013. applications. Environ Sci Pollut Res 2022:1–13.
[261] Sani E, et al. Potential of carbon nanohorn-based suspensions for solar thermal [291] Fotowat S, Askar S, Ismail M, Fartaj A. A study on corrosion effects of a water
collectors. Sol Energy Mater Sol Cells 2011;95(11):2994–3000. based nanofluid for enhanced thermal energy applications. Sustain Energy
[262] Sharaf OZ, Kyritsis DC, Al-Khateeb AN, Abu-Nada E. Effect of bottom surface Technol Assessments 2017;24:39–44.
optical boundary conditions on nanofluid-based DASC: parametric study and [292] Arshad A, Jabbal M, Yan Y, Reay D. A review on graphene based nanofluids:
optimization. Sol Energy 2018;164:210–23. preparation, characterization and applications. J Mol Liq 2019;279:444–84.

60
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

[293] Wang H, Li X, Luo B, Wei K, Zeng G. The MXene/water nanofluids with high conductivity, rheology, and convective heat transfer. Ind Eng Chem Res 2019;58
stability and photo-thermal conversion for direct absorption solar collectors: a (19):8349–69.
comparative study. Energy 2021;227:120483. [324] Koca HD, Doganay S, Turgut A, Tavman IH, Saidur R, Mahbubul IM. Effect of
[294] Vital CVP, Farooq S, de Araujo RE, Rativa D, Gómez-Malagón LA. Numerical particle size on the viscosity of nanofluids: a review. Renew Sustain Energy Rev
assessment of transition metal nitrides nanofluids for improved performance of 2018;82:1664–74.
direct absorption solar collectors. Appl Therm Eng 2021;190:116799. [325] V Timofeeva E, Smith DS, Yu W, France DM, Singh D, Routbort JL. Particle size
[295] Farooq S, et al. Thermo-optical performance of bare laser-synthesized TiN and interfacial effects on thermo-physical and heat transfer characteristics of
nanofluids for direct absorption solar collector applications. Sol Energy Mater Sol water-based α-SiC nanofluids. Nanotechnology 2010;21(21):215703.
Cells 2023;252:112203. [326] Nguyen CT, et al. Temperature and particle-size dependent viscosity data for
[296] Ma D. Chapter 1 - hybrid nanoparticles: an introduction. In: Mohapatra S, water-based nanofluids–hysteresis phenomenon. Int J Heat Fluid Flow 2007;28
Nguyen TA, Nguyen-Tri PBT, editors. Micro and nano technologies. Woodhead (6):1492–506.
Publishing; 2019. p. 3–6. https://doi.org/10.1016/B978-0-12-814134-2.00001-2. [327] Zahmatkesh I, et al. Effect of nanoparticle shape on the performance of thermal
[297] Tiwari AK, Kumar V, Said Z, Paliwal HK. A review on the application of hybrid systems utilizing nanofluids: a critical review. J Mol Liq 2021;321:114430.
nanofluids for parabolic trough collector: recent progress and outlook. J Clean [328] Wang Z, Zhang ZM, Quan X, Cheng P. A numerical study on effects of surrounding
Prod 2021;292:126031. medium, material, and geometry of nanoparticles on solar absorption efficiencies.
[298] Sreekumar S, Shah N, Mondol JD, Hewitt N, Chakrabarti S. Broadband absorbing Int J Heat Mass Tran 2018;116:825–32.
mono, blended and hybrid nanofluids for direct absorption solar collector: a [329] Qin C, Kim JB, Gonome H, Lee BJ. Absorption characteristics of nanoparticles
comprehensive review. Nano Futur 2022;6(2):22002. with sharp edges for a direct-absorption solar collector. Renew Energy 2020;145:
[299] Generous MM, Abu-Nada E, Alazzam A. Exploring the potential of hybrid 21–8.
nanofluids for enhanced heat transfer in a duct: a comprehensive study utilizing a [330] Song D, Yang Y, Jing D. Insight into the contribution of rotating Brownian motion
phases-interaction driven multiphase mixture model. Int. J. Thermofluids 2023: of nonspherical particle to the thermal conductivity enhancement of nanofluid.
100453. Int J Heat Mass Tran 2017;112:61–71.
[300] Wang X, He Y, Chen M, Hu Y. ZnO-Au composite hierarchical particles dispersed [331] Zou Y, Li X, Yang L, Zhang B, Wu X. Efficient direct absorption solar collector
oil-based nanofluids for direct absorption solar collectors. Sol Energy Mater Sol based on hollow TiN nanoparticles. Int J Therm Sci 2023;185:108099.
Cells 2018;179:185–93. [332] Jeong J, Li C, Kwon Y, Lee J, Kim SH, Yun R. Particle shape effect on the viscosity
[301] Struchalin PG, Kuzmenkov DM, Yunin VS, Wang X, He Y, V Balakin B. Hybrid and thermal conductivity of ZnO nanofluids. Int J Refrig 2013;36(8):2233–41.
nanofluid in a direct absorption solar collector: magnetite vs. Carbon nanotubes [333] Ferrouillat S, Bontemps A, Poncelet O, Soriano O, Gruss J-A. Influence of
compete for thermal performance. Energies 2022;15(5):1604. nanoparticle shape factor on convective heat transfer and energetic performance
[302] Wen J, et al. The enhanced photothermal characteristics of plasmonic ZrC/TiN of water-based SiO2 and ZnO nanofluids. Appl Therm Eng 2013;51(1–2):839–51.
composite nanofluids for direct absorption solar collectors. Renew Energy 2023; [334] Abbasi M, Heyhat MM, Rajabpour A. Study of the effects of particle shape and
206:676–85. base fluid type on density of nanofluids using ternary mixture formula: a
[303] Huminic G, Huminic A. Hybrid nanofluids for heat transfer applications–a state- molecular dynamics simulation. J Mol Liq 2020;305:112831.
of-the-art review. Int J Heat Mass Tran 2018;125:82–103. [335] Hatami M, Mosayebidorcheh S, Jing D. Thermal performance evaluation of
[304] Zhang Y, Li X, Li J, Ma C, Guo L, Meng X. Solar-driven phase change alumina-water nanofluid in an inclined direct absorption solar collector (IDASC)
microencapsulation with efficient Ti4O7 nanoconverter for latent heat storage. using numerical method. J Mol Liq 2017;231:632–9.
Nano Energy 2018;53:579–86. [336] Khatri R, Kumar M, Jiyani R. An experimental investigation on direct absorption
[305] Albdour SA, Haddad Z, Sharaf OZ, Alazzam A, Abu-Nada E. Micro/nano- solar collector using TiO 2-water nanofluid. In: 2018 2nd international conference
encapsulated phase-change materials (ePCMs) for solar photothermal absorption on green energy and applications (ICGEA). IEEE; 2018. p. 59–62.
and storage: fundamentals, recent advances, and future directions. Prog Energy [337] Alirezaie A, Saedodin S, Esfe MH, Rostamian SH. Investigation of rheological
Combust Sci 2022;93:101037. behavior of MWCNT (COOH-functionalized)/MgO-engine oil hybrid nanofluids
[306] Ma F, Zhang P. Performance investigation of the direct absorption solar collector and modelling the results with artificial neural networks. J Mol Liq 2017;241:
based on phase change slurry. Appl Therm Eng 2019;162:114244. 173–81.
[307] Essa FA, et al. Eco-friendly coffee-based colloid for performance augmentation of [338] Murshed SMS, Estellé P. A state of the art review on viscosity of nanofluids.
solar stills. Process Saf Environ Protect 2020;136:259–67. Renew Sustain Energy Rev 2017;76:1134–52.
[308] V Balakin B, Struchalin PG. Eco-friendly and low-cost nanofluid for direct [339] Motahari K, Moghaddam MA, Moradian M. Experimental investigation and
absorption solar collectors. Mater Lett 2023;330:133323. development of new correlation for influences of temperature and concentration
[309] Alberghini M, et al. Coffee-based colloids for direct solar absorption. Sci Rep on dynamic viscosity of MWCNT-SiO2 (20-80)/20W50 hybrid nano-lubricant.
2019;9(1):4701. Chin J Chem Eng 2018;26(1):152–8.
[310] Kosinska A, V Balakin B, Kosinski P. Use of biodegradable colloids and carbon [340] Asadi M, Asadi A, Aberoumand S. An experimental and theoretical investigation
black nanofluids for solar energy applications. AIP Adv 2021;11(5). on the effects of adding hybrid nanoparticles on heat transfer efficiency and
[311] V Nguyen L, Kosinski P, V Balakin B, Kosinska A. Direct absorption solar collector: pumping power of an oil-based nanofluid as a coolant fluid. Int J Refrig 2018;89:
use of nanofluids and biodegradable colloids. Int J Therm Sci 2023;190:108292. 83–92.
[312] Zhu W, Zuo X, Ding Y, Yan H, An Y, Yang W. Experimental investigation on the [341] Duan H, Tang L, Zheng Y, Xu C. Effect of plasmonic nanoshell-based nanofluid on
photothermal conversion performance of cuttlefish ink nanofluids for direct efficiency of direct solar thermal collector. Appl Therm Eng 2018;133:188–93.
absorption solar collectors. Appl Therm Eng 2023;221:119835. [342] Delfani S, Esmaeili M, Karami M. Application of artificial neural network for
[313] Gupta VK, Kumar S, Kukreja R, Chander N. Experimental thermal performance performance prediction of a nanofluid-based direct absorption solar collector.
investigation of a direct absorption solar collector using hybrid nanofluid of gold Sustain Energy Technol Assessments 2019;36:100559.
nanoparticles with natural extract of Azadirachta Indica leaves. Renew Energy [343] Li X, et al. Numerical analysis of photothermal conversion performance of MXene
2023;202:1021–31. nanofluid in direct absorption solar collectors. Energy Convers Manag 2020;226:
[314] Said Z, Sajid MH, Saidur R, Mahdiraji GA, Rahim NA. Evaluating the optical 113515.
properties of TiO2 nanofluid for a direct absorption solar collector. Numer Heat [344] Kumar R, Bharadwaj G, Kharub M, Goel V, Kumar A. Performance analysis of
Tran 2015;67(9):1010–27. nanofluid based direct absorption solar collector of different configurations: a
[315] Lenert A. Nanofluid-based receivers for high-temperature, high-flux direct solar two-phase CFD modeling. Energy Sources, Part A Recover Util Environ Eff 2022:
collectors. Massachusetts Institute of Technology; 2010. 1–16.
[316] Howe ML, Paul TC, Khan JA. Radiative properties of Al2O3 nanoparticles [345] Wang K, et al. Highly-efficient nanofluid-based direct absorption solar collector
enhanced ionic liquids (NEILs) for direct absorption solar collectors. Sol Energy enhanced by reverse-irradiation for medium temperature applications. Renew
Mater Sol Cells 2021;232:111327. Energy 2020;159:652–62.
[317] Chen M, He Y, Zhu J, Kim DR. Enhancement of photo-thermal conversion using [346] Qu J, Zhou G, Zhang M, Shang L, Yu W. Effect of reverse irradiation angle on the
gold nanofluids with different particle sizes. Energy Convers Manag 2016;112: photo-thermal conversion performance of MXene nanofluid-based direct
21–30. absorption solar collector. Sol Energy Mater Sol Cells 2023;251:112164.
[318] Wang B, Wang X, Lou W, Hao J. Ionic liquid-based stable nanofluids containing [347] Peng W, Sadaghiani OK. Thermal performance enhancement of direct absorption
gold nanoparticles. J Colloid Interface Sci 2011;362(1):5–14. solar collector using nanoparticle in sunray trap. Appl Therm Eng 2021;187:
[319] Ambreen T, Kim M-H. Influence of particle size on the effective thermal 116578.
conductivity of nanofluids: a critical review. Appl Energy 2020;264:114684. [348] Zheng D, Yao J, Zhu H, Wang J, Yin C. Optimizing photothermal conversion
[320] Nikkam N, Ghanbarpour M, Khodabandeh R, Toprak MS. The effect of particle efficiency in a parabolic trough direct absorption solar collector through
size and base liquid on thermo-physical properties of ethylene and diethylene ferrofluid and magnetic field synergy. Energy Convers Manag 2023;285:117020.
glycol based copper micro-and nanofluids. Int Commun Heat Mass Tran 2017;86: [349] Wang D, et al. Highly efficient energy harvest via external rotating magnetic field
143–9. for oil based nanofluid direct absorption solar collector. Green Energy Environ
[321] Maheshwary PB, Handa CC, Nemade KR. A comprehensive study of effect of 2021;6(2):298–307.
concentration, particle size and particle shape on thermal conductivity of titania/ [350] Chen Y, Zhang Y, Lan H, Li C, Liu X, He D. Electric field combined nanofluid to
water based nanofluid. Appl Therm Eng 2017;119:79–88. enhance photothermal efficiency of the direct absorption solar collector. Renew
[322] He Q, Zeng S, Wang S. Experimental investigation on the efficiency of flat-plate Energy 2023;215:118988.
solar collectors with nanofluids. Appl Therm Eng 2015;88:165–71. [351] Hatami M, Jing D. Optimization of wavy direct absorber solar collector (WDASC)
[323] Barai DP, Bhanvase BA, Saharan VK. Reduced graphene oxide-Fe3O4 using Al2O3-water nanofluid and RSM analysis. Appl Therm Eng 2017;121:
nanocomposite based nanofluids: study on ultrasonic assisted synthesis, thermal 1040–50.

61
A. Hasan et al. Progress in Energy and Combustion Science 103 (2024) 101160

[352] Hooshmand A, Zahmatkesh I, Karami M, Delfani S. Porous foams and nanofluids [378] Kim H, Ham J, Park C, Cho H. Theoretical investigation of the efficiency of a U-
for thermal performance improvement of a direct absorption solar collector: an tube solar collector using various nanofluids. Energy 2016;94:497–507.
experimental study. Environ Prog Sustain Energy 2021;40(6):e13684. [379] Said Z, Arora S, Bellos E. A review on performance and environmental effects of
[353] Baïri A, Martín-Garín A, Millán-García JA. Porous media to intensify natural conventional and nanofluid-based thermal photovoltaics. Renew Sustain Energy
convection in rectangular enclosures used in building techniques. Int J Mod Phys Rev 2018;94:302–16.
C 2020;31(4):2050061. [380] Dobrovolskaia MA, Shurin MR, Kagan VE, Shvedova AA. Ins and outs in
[354] Bairi A. Porous materials saturated with water-copper nanofluid for heat transfer environmental and occupational safety studies of asthma and engineered
improvement between vertical concentric cones. Int Commun Heat Mass Tran nanomaterials. ACS Nano 2017;11(8):7565–71.
2021;126:105439. [381] Ahamed M, AlSalhi MS, Siddiqui MKJ. Silver nanoparticle applications and
[355] Mourad A, et al. Galerkin finite element analysis of thermal aspects of Fe3O4- human health. Clin Chim Acta 2010;411(23–24):1841–8.
MWCNT/water hybrid nanofluid filled in wavy enclosure with uniform magnetic [382] Elder A, et al. Translocation of inhaled ultrafine manganese oxide particles to the
field effect. Int Commun Heat Mass Tran 2021;126:105461. central nervous system. Environ Health Perspect 2006;114(8):1172–8.
[356] Bozorgi M, Ghasemi K, Mohaghegh MR, Tasnim SH, Mahmud S. Optimization of [383] Pernodet N, et al. Adverse effects of citrate/gold nanoparticles on human dermal
silver/water-based porous wavy direct absorption solar collector. Renew Energy fibroblasts. Small 2006;2(6):766–73.
2023;202:1387–401. [384] Handy RD, Shaw BJ. Toxic effects of nanoparticles and nanomaterials:
[357] Zhang H, Wang K, Yu W, Wang L, Xie H. Ternary molten salt energy storage implications for public health, risk assessment and the public perception of
coupled with graphene oxide-TiN nanofluids for direct absorption solar collector. nanotechnology. Health Risk Soc 2007;9(2):125–44.
Energy Build 2021;253:111481. [385] Pan Y, et al. Size-dependent cytotoxicity of gold nanoparticles. Small 2007;3(11):
[358] Ali HM, et al. Advances in thermal energy storage: fundamentals and applications. 1941–9.
Prog Energy Combust Sci 2024;100:101109. [386] Cancino-Bernardi J, Paino I, Souza J, Marangoni V, Nogueira P, Zucolotto V.
[359] Qin C, Lee J, Lee BJ. A hybrid direct-absorption parabolic-trough solar collector Current challenges in the commercialization of nanocolloids: toxicology and
combining both volumetric and surface absorption. Appl Therm Eng 2021;185: environmental issues. In: Nanocolloids. Amsterdam, The Netherlands: Elsevier;
116333. 2016. p. 427–63.
[360] Rasih RA, Sidik NAC, Samion S. Recent progress on concentrating direct [387] Rocha TL, Gomes T, Sousa VS, Mestre NC, Bebianno MJ. Ecotoxicological impact
absorption solar collector using nanofluids: a review. J Therm Anal Calorim 2019; of engineered nanomaterials in bivalve molluscs: an overview. Mar Environ Res
137:903–22. 2015;111:74–88.
[361] Jyani L, Singh M, Bohra S. Effect of volume concentration on direct absorption [388] Brittle SW, et al. Freshwater crayfish: a potential benthic-zone indicator of
solar collector in laminar and turbulent flow using CuO–H 2 O nanofluids. Int J nanosilver and ionic silver pollution. Environ Sci Technol 2016;50(13):7056–65.
Eng Dev Res 2017;5(4):268–72. [389] Akhavan-Behabadi MA, Shahidi M, Aligoodarz MR, Ghazvini M. Experimental
[362] Arulprakasajothi M, Dineshbabu M, Jothishanmugam C, Saikrishnan V. investigation on thermo-physical properties and overall performance of
Convective heat transfer characteristics of nanofluids. In: Frontiers in automobile MWCNT–water nanofluid flow inside horizontal coiled wire inserted tubes. Heat
and mechanical engineering-2010. IEEE; 2010. p. 201–4. Mass Tran 2017;53:291–304.
[363] Apmann K, Fulmer R, Soto A, Vafaei S. Thermal conductivity and viscosity: [390] Ranjbarzadeh R, Moradikazerouni A, Bakhtiari R, Asadi A, Afrand M. An
review and optimization of effects of nanoparticles. Materials 2021;14(5):1291. experimental study on stability and thermal conductivity of water/silica
[364] Godson L, Raja B, Lal DM, Wongwises S. Experimental investigation on the nanofluid: eco-friendly production of nanoparticles. J Clean Prod 2019;206:
thermal conductivity and viscosity of silver-deionized water nanofluid. Exp Heat 1089–100.
Tran 2010;23(4):317–32. [391] Said Z, Rahman SMA, Assad MEH, Alami AH. Heat transfer enhancement and life
[365] Gao D, Bai M, Hu C, Lv J, Wang C, Zhang X. Investigating control of convective cycle analysis of a Shell-and-Tube Heat Exchanger using stable CuO/water
heat transfer and flow resistance of Fe3O4/deionized water nanofluid in magnetic nanofluid. Sustain Energy Technol Assessments 2019;31:306–17.
field in laminar flow. Nanotechnology 2020;31(49):495402. [392] Rashmi W, Ismail AF, Khalid M, Anuar A, Yusaf T. Investigating corrosion effects
[366] Malekzadeh A, Pouranfard AR, Hatami N, Kazemnejad Banari A, Rahimi MR. and heat transfer enhancement in smaller size radiators using CNT-nanofluids.
Experimental investigations on the viscosity of magnetic nanofluids under the J Mater Sci 2014;49:4544–51.
influence of temperature, volume fractions of nanoparticles and external [393] Srinivas V, Moorthy CV, Dedeepya V, V Manikanta P, Satish V. Nanofluids with
magnetic field. J Appl Fluid Mech 2016;9(2):693–7. CNTs for automotive applications. Heat Mass Tran 2016;52:701–12.
[367] Karami M. Experimental investigation of first and second laws in a direct
absorption solar collector using hybrid Fe 3 O 4/SiO 2 nanofluid. J Therm Anal
Mr. Alabas Hasan, a Mechanical Engineering Ph.D. candidate at Khalifa University under
Calorim 2019;136:661–71.
the guidance of Dr. Abu-Nada and Dr. Alazzam, earned his MSc from Jordan University of
[368] Liu J, Ye Z, Zhang L, Fang X, Zhang Z. A combined numerical and experimental
Science and Technology. His current research focuses on numerically simulating and
study on graphene/ionic liquid nanofluid based direct absorption solar collector.
modeling the enhancement of solar energy performance through the use of nano-
Sol Energy Mater Sol Cells 2015;136:177–86.
encapsulated phase change materials.
[369] Jyani L, Singh M, Bohra S, Rajpurohit L. Experimental investigation of direct
absorption solar collector UsingCuO-H2O nanofluid for laminarand turbulent
flow. Int J Res Appl Sci Eng Technol 2017;(X):1718–23. Dr. Anas Alazzam is a graduate of Concordia University – Montreal. After a brief Post­
[370] Faizal M, Saidur R, Mekhilef S, Hepbasli A, Mahbubul IM. Energy, economic, and Doctoral Fellow appointment with the Canadian Space Agency, he joined Khalifa Uni­
environmental analysis of a flat-plate solar collector operated with SiO 2 versity in 2012. He is currently an associate professor of Mechanical Engineering and the
nanofluid. Clean Technol Environ Policy 2015;17:1457–73. head of the microfluidics lab at Khalifa University. In addition, he is a visiting associate
[371] Faizal M, Saidur R, Mekhilef S, Faizal M. Potential of size reduction of flat-plate professor in the department of electrical engineering at Ecole de technologie superieure
solar collectors when applying Al2O3 nanofluid. Adv Mater Res 2014;832: ETS. Dr. Alazzam has more than one hundred peer-reviewed publications and a number of
149–53. international patents. His research expertise in microfluidics, nanofluids, Dielectropho­
[372] Ardente F, Beccali G, Cellura M, Lo Brano V. Life cycle assessment of a solar resis, graphene, and healthcare applications of microdevices. His research in the BioMEMS
thermal collector. Renew Energy 2005;30(7):1031–54. area has received substantial recognition in the form of awards and media appearance. He
[373] Faizal M, Saidur R, Mekhilef S, Alim MA. Energy, economic and environmental has been the Principal Investigator of numerous research projects with budgets exceeding
analysis of metal oxides nanofluid for flat-plate solar collector. Energy Convers $2 million. He has served as the advisor to numerous PhD and MSc students and post­
Manag 2013;76:162–8. doctoral fellows. Dr. Alazzam is a member of the editorial board of Micromachines and
[374] Maadi SR, Kolahan A, Passandideh-Fard M, Sardarabadi M, Moloudi R. PLOS ONE Journals.
Characterization of PVT systems equipped with nanofluids-based collector from
entropy generation. Energy Convers Manag 2017;150:515–31.
Dr. Eiyad Abu-Nada is a Full Professor of Mechanical Engineering in the Department of
[375] Otanicar TP, Golden JS. Comparative environmental and economic analysis of
Mechanical Engineering at Khalifa University in Abu Dhabi, UAE. He joined Khalifa
conventional and nanofluid solar hot water technologies. Environ Sci Technol
University in 2013. Prior to that, he worked as a Full Professor at King Faisal University in
2009;43(15):6082–7.
Saudi Arabia and Associate Professor at Hashemite University in Jordan. Dr. Abu-Nada is
[376] Tiwari AK, Ghosh P, Sarkar J. Solar water heating using nanofluids–a
an Alexander von Humboldt Fellow and a recipient of Abdul-Hameed Shoman Award in
comprehensive overview and environmental impact analysis. Int. J. Emerg.
Jordan. Also, he received the Distinguished Scholar Award from the Arab Fund in Kuwait.
Technol. Adv. Eng. 2013;3(3):221–4.
His recent research interests are in nanofluid heat transport and microfluidics simulations.
[377] Khullar V, Tyagi H. A study on environmental impact of nanofluid-based
concentrating solar water heating system. Int J Environ Stud 2012;69(2):220–32.

62