deionized water nanofluid on a copper surface

(1)

Contents lists available atScienceDirect

International Journal of Thermal Sciences

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

E ﬀ ect of di ﬀ erent surfactants on the pool boiling heat transfer of SiO

₂

/ deionized water nano ﬂ uid on a copper surface

Sasan Etedali

^a,b

, Masoud Afrand

^a,b,*

, Ali Abdollahi

^a,b

aDepartment of Mechanical Engineering, Najafabad Branch, Islamic Azad University, Najafabad, Iran

bModern Manufacturing Technologies Research Center, Najafabad Branch, Islamic Azad University, Najafabad, Iran

A R T I C L E I N F O

Keywords:

Pool boiling

DI water/ SiO2nanoﬂuid Surfactant

Experimental Two phase

A B S T R A C T

The boiling heat transfer is one of the most widely applied heat transfer processes in today's industries. This study is an experimental investigation of the pool boiling heat transfer in a nanofluid composed of deionized (DI) water and silicon dioxide (SiO2) nanoparticles with added surfactants (Ps20, CTAB, SLS) on a copper surface at various nanofluid concentrations, under atmospheric pressure conditions. The nanofluid was synthesized using a two-step method and showed acceptable stability. The concentration of surfactant in nanofluid for all prepared nanofluids is ¹

20 of nanoparticles mass. For the sake of repeatability and accuracy of the experiments, the deionized water was measured three times and was found to match the relations proposed in the literature. The results show the boiling heat transfer to be enhanced in the DI water with added surfactants compared with the deionized water. Moreover, the DI water with added surfactants observed a better boiling heat transfer compared to the DI-water-based SiO₂nanoﬂuid with added surfactant. In addition, the boiling heat transfer was reduced as the concentration of the SiO2nanoﬂuid containing SLS and CTAB surfactants was increased.

Meanwhile, increasing the concentration of the SiO2nanoﬂuid containing Ps20 surfactants from 0.01 to 0.1 vol%

enhanced the heat transfer. However, increasing the concentration to 1 vol% reduced the heat transfer drastically. Therefore, at all heatfluxes, the nanofluid specimens with added SLS surfactants were associated with the highest boiling heat transfer, which was followed by the nanofluid with added CTAB, and the nanofluid with Ps20 (SLS > CTAB > Ps20). In conclusion, ranking them by impact, anionic, cationic, and nonionic surfactants (nonionic < cationic < anionic) resulted in the largest to smallest increase in the Boiling Heat Transfer Coefficient (BHTC) in the nanofluids.

1. Introduction

Cooling systems with many technological advances have en- countered a number of constraints, and even with the increasing use of technology, the use of liquid coolers, different geometries of equipment and cooling methods is insufficient for the high volume of heat, hence new methods of increasing heat transfer are required [1–5]. Also for study the increase in heat transfer for the cooling of equipment, the researchers first should study the single-phase and two-phase heat transfer [6–9]. The boiling heat transfer is one of the most effective and efficient methods of heat transfer in today's refrigeration and heating industry and different engineering applications. The increased heat- transfer rate in boiling has been attractive to the researchers in this field. Given the large latent heat of the boilingfluid, a large amount of energy is dissipated from the solid surface. This great feature of boiling is utilized in many industrial processes. Different methods of enhancing

heat transfer may be classified into active and passive methods. Passive methods include rough surfaces, extended surfaces, andfluid additives while using an electrostaticfield, and surface orfluid vibrations are an examples of the activate methods [10–18]. As a passive method, the nanofluids may be used to enhance the pool boiling heat transfer. Na- nofluids contain particles below 100 nm that are associated with a higher thermal conductivity compared with the basefluid. Many studies have addressed the boiling of nanofluids [19]. Some researchers have reported an enhancement in heat transfer by using nanofluids [20–22] while others have reported a decrease on heat transfer in nanofluids [23–25]. The boiling heat transfer may be controlled de- pending on the application, as in some, the goal is to enhance the heat transfer while in some other we seek to reduce it. This has also been addressed by some studies [26–33].

Abdollahi et al. [34] investigated the impact of ferroﬂuids on the pool boiling heat transfer and deposited surfaces. They used a ferroﬂuid

https://doi.org/10.1016/j.ijthermalsci.2019.105977

Received 22 February 2018; Received in revised form 20 September 2018; Accepted 7 June 2019

*Corresponding author. Department of Mechanical Engineering, Najafabad Branch, Islamic Azad University, Najafabad, Iran.

E-mail addresses:[email protected],[email protected](M. Afrand).

T

(2)

hances the heat transfer. Furthermore, they found the maximum boiling heat transfer taking place at 0.15 vol%. In a study on the boiling heat transfer in a nanoﬂuid composed of ZnO and ethylene glycol, Kole &

Dey [37] attained a maximum heat transfer of 40% at a concentration of 3.75 vol%. Moreover, the heat transfer coeﬃcient decreased as the number of ZnO nanoparticles were increased in the nanoﬂuid. This was attributed to the sedimentation of the nanoparticles and the blocking of heating surface.

Silicon dioxide (SiO2) is the most abundant oxide in the Earth's crust. Having a favorable specific area and a high BHTC which allows for investigating the heat transfer, and due to their thermal and hy- drodynamic characteristics, SiO2 nanoparticles have attracted much scientific attention. Dadjoo et al. [38] investigated the various in- clinations of a copper boiling plate in the SiO2-water nanofluid. Con- sidering the data, it was found that increasing the inclination of the boiling plate from 0 to 180° enhances the critical heatflux (CHF) but reduces the BHTC. Kulkarni et al. [39] investigated the BHTC of a SiO2- ethylene glycol (EG) nanofluid. They used SiO2nanoparticles of various sizes (20, 50, 100 nm) in their study. They reported the heat transfer coefficient to be increased. Vassallo et al. [40] experimentally investigated the pool boiling heat transfer of a SiO2-water nanofluid on a NiCr wire heater with a diameter of 0.4 mm. They studied SiO2nano- particles of two sizes (15 and 50 nm),finding the BHTC to be increased with increased nanoparticle size. The incorporation of nanofluids in pool boiling heat transfer has been addressed for a variety of nanoparticles (especially SiO2) and different basefluids [41–50].

Nonetheless, the impact of diﬀerent surfactants on the pool-boiling heat transfer has been less addressed [51–55]. Kumar and Milanova [56] compared the impact of surface tension on a CNT-water-surfactant nanoﬂuid to that on a water-surfactant. They examined the sodium dodecylbenzene sulfate (NaDBS) surfactant in various concentrations.

Moreover, they reported the surface tension to be smaller in water containing surfactants compared to that in the nanoﬂuid. Wasekar and Manglik [57] investigated pool boiling heat transfer of water with anionic surfactants. This study used sodium dodecyl sulfate (SDS) or sodium lauryl sulfate (SLS) as the anionic surfactant. The boiling heat transfer of water was enhanced considerably with added SDS, however, the enhancement in heat transfer was smaller at higher concentrations.

In addition, the enhancement of boiling heat transfer depends on parameters such as the size of the nanoparticles, the shape of bubbles on the heater surface, and the increase in the density of the nucleation sites. Furthermore, dynamic surface tension plays the main role in de- forming dynamic bubbles and in heat transfer by adsorption of surfactant molecules during boiling. Wasekar et al. [58,59] experimentally investigated the impact of the anionic and nonionic surfactants on the horizontal cylindrical heater and the pool-boiling heat transfer coeﬃ- cient. In this study, two anionic surfactants of SDS and sodium lauryl ether sulfate (SLES) and two nonionic surfactants (Triton X-100 and Triton X-305). They show on the impact of the surface tension of surfactants on bubble growth. The anionic surfactants were reported to oﬀer a better enhancement in boiling heat transfer compared to nonionic ones.

Sarafraz and Hormozi [11] investigated the pool-boiling of a CuO- water nanoﬂuid, using Triton X-100, SDBS, and SDS surfactants on a horizontal cylindrical heater. They found the BHTC to be reduced for all

transfer in multi-walled Carbon Nanotubes (CNTs) in pure water containing 9% anionic SDS surfactants. It was found that the addition of carbon nanotubes drastically enhanced the boiling heat transfer coefficient of the basefluid. Moreover, the BHTC was increased in the CNT- water nanofluid almost two times more than the CNT-water-surfactant nanofluid. They attributed the reduction in the BHTC to the increase in the boiling residues resulted from adding surfactants at high concentrations.

One method that researchers use to stabilize nanoparticles in the basefluid (to prepare a nanofluid) is adding surfactants to the nanofluid. Using surfactants results in a reduction in surface tension and homogenizes the nanofluid by scattering the nanoparticles in the base fluid, leading to an enhanced boiling heat transfer. A survey of previous studies shows that the impact of different types of surfactants (anionic, cationic, and nonionic) on the pool-boiling heat transfer of the SiO2- water nanofluid is yet to be reported. Therefore, this study is thefirst to investigate the pool-boiling heat transfer of a SiO2-DI water nanofluid with three surfactants including cetyl trimethyl ammonium bromide (CTAB), polysorbate 20 (Ps20), and SLS on aflat copper plate under atmospheric pressure conditions. In addition, most previous studies have only presented their observations, therefore, there is a lack of explanation for the heat-transfer behavior of nanofluids with added surfactants. In this regard, this study addresses the impact of surfactants on the BHTC in nanofluids in detail. The question here is whether the addition of surfactants to the nanofluids or ionic structure of them af- fects the boiling heat-transfer properties of the nanofluid. In better words, which of the nanofluid or the surfactants (anionic, cationic, or nonionic) is more effective on the boiling heat-transfer?

2. Materials and methods

2.1. Preparing the nanoﬂuid

In this study, in order to prepare the nanoﬂuid, a certain amount of SiO2nanoparticles (around 6.525 gr SiO2nanoparticles for making a 1 vol% SiO2-DI water nanoﬂuid) was poured into a volumetric balon where 0.163 gr of the surfactant (CTAB, Ps20, or SLS) was added to it.

DI water was then added to the mix, bringing the volume to 250 ml, and the solution was stirred at 500 rpm for 180 min. The solution was sonicated at 60% power with a period of 0.5 s for 20 min at the end.

Ultrasonication was carried out in three stages and every sample was sonicated for a total of 1 h. SiO2nanoparticles of the same diameter were used in various volume fractions to prepare the nanofluid with water as the base fluid. Nanofluids of various concentrations were prepared by diluting an initial nanofluid using DI water. Equation(1) was used to calculate the amount of basefluid required for making a nanofluid at a specific concentration (vol%).

=

ϕ . V₁ 1 ϕ . V2 2 (1)

Where ϕ₁ and V1 are the volume fraction of the initial (more concentrated) nanofluid and the required volume of the nanofluid to dilute, andϕ₂andV2are the volume fraction of the nanofluid and its volume, respectively. To prepare a nanofluid of a specific concentration, the amount of the required nanofluid is found using Eq.(1). The nanofluid is then diluted using the base fluid up to a certain volume. Then,

(3)

different concentrations of the nanofluid were prepared by diluting the solution obtained in this step (Table 1). To prepare less concentrated solutions, specific amounts of the initial solution were diluted using DI water according toTable 1. Small volumes picked using a micropipette and discharged them into a 100 ml volumetricflask.

For example, to prepare a 0.5 vol% nanoﬂuid, 75 ml of the 1 vol%

nanofluid was taken and brought to 150 ml by adding the basefluid. In this solution, the concentration of SiO2nanoparticles is 0.5 vol%. As mentioned earlier, water is used to dilute and prepare the nanofluids according toTable 1. SiO2-DI water nanofluids containing Ps20, CTAB, and SLS surfactants were prepared with four concentrations (1, 0.5, 0.1, and 0.01 vol%). It must be mentioned that all materials used in this study were acquired from Merck and Sigma Aldrich.Fig. 1shows the nanofluid containing SLS surfactants.

2.1.1. Investigating the shape of the SiO2nanoparticles using TEM In this study, TEM was employed to characterize the geometry and shape of the nanoparticles. Ethanol was added to the SiO2nanoparticles after being diluted down to 0.0001 vol% and 0.1 cc of the solution was placed on a smooth graphite sheet (with asperities of less than 2 nm).

After the added ethanol fully evaporated from the graphite surface, TEM was carried out and the heights of the nanoparticles were measured. The shape, geometry, and the diameter of the nanoparticles were speciﬁed using the low-height areas (graphite surface) and the asperities formed by the nanoparticles considering the TEM scale (Fig. 2).

The synthesized nanoparticles were found to be almost spherical.

Moreover, the diagrams show the size of the SiO2nanoparticles to be around 50 nm.

2.1.2. X-ray diﬀraction (XRD)

XRD is a test used to prove and characterize the type of

nanoparticles in the nanoﬂuid used in the experiment. The XRD results for the SiO2nanoparticles are presented inFig. 3. As evident from the results, a peak can be found at 25° which suggests the SiO2 nanoparticles have a non-crystalline structure. The crystal grain size in the nanoparticle samples may be estimated with the help of the related Table 1

Preparing nanoﬂuids with diﬀerent concentrations.

The concentration of nanoparticles in the baseﬂuid (volume fraction)

Volume (ml) picked from the 1 vol%

nanoﬂuid

Theﬁnal volume after dilution

%1 _ 150

%0.5 75 150

%0.1 15 150

%0.01 1.5 150

Fig. 1.SiO2-DI water nanoﬂuids containing 50 nm SLS surfactants in four concentrations (1, 0.5, 0.1, and 0.01 vol%).

Fig. 2.TEM results for SiO₂nanoparticles at a magniﬁcation of 40 nm.

Fig. 3.XRD results for SiO2nanoparticles.

(4)

relations and the XRD results. Moreover, a comparison between this diagram and the results reported in previous studies shows that SiO2

nanoparticles are very pure and can be assumed to only have an O–Si–O structure. XRD also shows that these nanoparticles are irregular (amorphous) and have no crystalline network (the middle peak).

Meanwhile, the peak is similar to the other silica nanoparticles, which shows that this is a pure silica nanoparticle.

2.1.3. Dynamic Light Scattering (DLS)

As it was mentioned in previous studies, one way for measuring the size of nanoparticles in a baseﬂuid is through DLS. In this study, DLS was employed to measure the size of the SiO2nanoparticles in the base ﬂuid (DI water) and to prepare nanoparticle size distribution diagrams.

The results for the measurement of the diameter of the nanoparticles is presented in Fig. 4 for the nanofluid. Given the fact that this test method works by the dynamic scattering of the light, a nanofluid with a concentration of 0.01 vol% (the lowest possible concentration) was used to prevent particles interfering with one another and resulting in deviation and error in the measurement. Moreover, DLS tests lasted for 10 min after the nanofluid samples were prepared by sonication. The DLS results for the water-SiO2nanofluid is suggestive the average diameter of the nanoparticles in the basefluid after ultrasonication to be 48.63 nm which is in agreement with TEM results.

2.1.4. Nanoparticle stability test results (Zeta potential)

A suitable criterion for investigating the stability of the SiO2na- noparticles in the DI water and to verify that particle agglomeration does not take place is the surface charge of the nanoparticles which results in electrostatic repulsive force between them. Surface charges prevent the agglomeration of the nanoparticles. Thus, the higher the surface charge of the nanoparticles (positive or negative) the more stable the nanofluid as nanoparticle agglomeration will be prevented more efficiently. In this study, the Zeta potential test was employed to investigate the stability of the SiO2nanoparticles in the basefluid. In this test, the maximum surface charge of the SiO2nanoparticles was found to be either below−40 mV or above +40 mV in all nanofluids (DI water-SiO2nanofluids containing SLS, CTAB, and Ps20 surfactants) which, considering the nature of the nanoparticles and the basefluid, means that the nanofluid is stable [62]. The results of the Zeta potential test are presented inFig. 5 through 7for the DI water-SiO2nanofluids containing SLS, CTAB, and Ps20 surfactants, 30 min after they were prepared and sonicated. The Zeta-potential diagrams show all surfactants to provide excellent nanofluid stability. Due to the stability of the nanoparticles in the basefluid, the agglomeration of the nanoparticles

and its effect on the physical properties such as the thermal conductivity of the nanofluid, nanofluid viscosity, and the specific heat can be ignored. The results of the Zeta-potential test inFigs. 5–7show the maximum surface charge for the nanofluids with SLS, CTAB, and Ps20 surfactants to be−87.4, 74.2, and−97.9, respectively, all of which satisfy the stability conditions. It is thus concluded that the oxide nanoparticles used in study are stable and are sufficiently charged to prevent their agglomeration (nanofluid stability means that nanoparticle agglomeration and sedimentation do not take place).

2.2. Experimental apparatus

The boiling heat transfer apparatus used in this study was composed of four main parts: the glass boiling chamber, the copper cylinder, the electronic power control unit, and the cooling system.Figs. 8 and 9 show a schematic view of the apparatus.

The main boiling chamber was a German-made quartz cylinder with high thermal resistance (up to 1600 °C). The 350 mm long glass chamber has an inner diameter of 40 mm and a thickness of 3 mm. The glass cylinder was made with an edge at the bottom that served for sealing. Using a silicone gasket, the glass tube was placed on it con- centrically with respect to the copper block. In order to create one- dimensional heat transfer conditions, the copper must be surrounded by a material with low thermal conductivity compared to copper so to concentrate the heatflux in the vertical direction to the top. In this regard, an Italian-made, high-grade, Polytetrafluoroethylene (PTFE) cylinder was used that was 150 mm wide and 90 mm tall. The heat transfer coefficient was measured using a Decagon KD2 Pro thermo- meter (with an accuracy of 0.01) and was found to be 0.29 w/mk. In addition, the PTFE was threaded to the size of the outer diameter of the copper block to be placed inside the PTFE. Two 3 mm holes were drilled into the Teflon casing to place the sensors at 90° with respect to one another (Fig. 10). The copper block was shaped like a 90 mm long and 40 mm wide cylinder and was made of 96% pure copper (Fig. 10).

Using a British Cussons P5687 apparatus, the thermal conductivity of copper was measured to be 401 w/mk. Considering the fact that the thermal conductivity of copper is many times larger than that of PTFE, heat transfer in the radial direction was successfully minimized. Four vertical, 6 mm wide and 40 mm long holes were embedded at the end of the copper cylinder to secure the heating elements (Fig. 10). Four 6 mm wide and 4 cm long, 250 W, steel heating elements were placed at the end of the copper block. It is important to note that all of the heating elements were completely located inside the copper block and had no contact with air. Moreover, two holes were drilled 20 mm into the Fig. 4.Dynamic Light Scattering (DLS) results for the water-SiO2nanoﬂuid containing SLS surfactants with a concentration of 0.01 vol%.

(5)

copper block (to the center), 15 mm and 38 mm away from the top of the cylinder at 90° with respect to one another. These holes served to place the PT-100 sensors. Note that the whole outer surface of the copper block is threaded so it can be placed inside the PTFE (Fig. 10).

The diameter of the glass chamber matched that of the copper block. It must be kept in mind that the glass chamber is brittle and cannot withstand a large pressure. Therefore, to seal the device, a stainless steel holder with a diameter of 45 mm was used that had three holes 120° apart from one another to be riveted to the Teflon. The holes reach back to the end of the Teflon piece where a threaded pin is connected to them. It must be noted that before connecting the glass, a silicone gasket (with an inner diameter of 38 mm and an outer diameter of 65 mm) was placed on the copper block. Moreover, the holder was attached to the screws using a polyurethane gasket after the glass chamber was put in place. This gasket allows for applying more pressure to the glass without breaking it. The holder screws were selected from the M16 size and were 10 mm long. The space between the heater and heating block is perfectly filled with silicone adhesive (approxi- mately between 0.1 mm) with an estimate of the volume (distance) between the heater and heating block and silicone adhesive and additional adhesive out of space between them we ensure that the space between the heater and thermal block has been filled with adhesive.

After using silicon adhesive, the space between the thermal block and surrounding air is completely covered with insulating stone adhesive.

This will cause minimizing the heat loss. Similarly we used silicone adhesive for holes where there is a thermocouple as before (the space between the heater and heating block) and to ensure the layer of suitable insulating stone adhesive andﬁberglass around PTFE. The reason for using silicon adhesive is due to its high thermal conductivity coef- ﬁcient. Also the main application of this adhesive is for computing

processors such as CPUs and etc.

In this study, 30 cm long PT-100 resistive temperature detectors (RTD) with an accuracy of 0.1 K were used. The sensors were calibrated to the temperature measurement range. Two Autonics TC4Y temperature displays were also used. With an error of only 0.1, the 4-character displays were very accurate. They also allowed for adjusting the temperature so the contactor could be directly commanded to connect/

disconnect the elements for device safety. In addition, the dimensions of the electrical enclosure were 35 in 40 cm. The ammeter and voltmeter used in this study were both products of Micromax.

Given the fact that a uniform heatflux is required for accurate test conditions, an auto transformer device was required for maintaining the voltage and the current. The single-phase auto transformer attached to the device was a product of Goldstar and provided an output voltage of 0–300 V. The manually adjustable auto transformer worked with AC voltage and was directly connected to the mains. A cooling system was embedded into the boiling chamber to maintain the surface of thefluid and prevent its evaporation. The system was composed of a 200 mm long glass condenser that was connected to the glass chamber with a 40 mm polyurethane gasket. In addition, after the condenser was placed on the glass chamber using the gasket, a special glue was used to ensure the device is perfectly sealed. The cooling system is separate from the device. The pump inside the cooling vessel sends ice water into the condenser by passing through the rotameter which condenses the va- pors. The condensed vapor then returns to the boiling chamber and water circulates back to the cooling vessel again. Furthermore, the bottom side of the heating elements was perfectly insulated using glass wool to prevent energy dissipation. The condenser chamber was open to the atmosphere, therefore, the boiling chamber was at atmospheric pressure. Comparing the input power and the increase in the energy of Fig. 5.Zeta-Potential test results for SiO2nanoparticles with added SLS surfactants with a concentration of 0.01 vol% in DI water as the basefluid.

Fig. 6.Zeta-Potential test results for SiO2nanoparticles with added CTAB surfactants with a concentration of 0.01 vol% in water as the baseﬂuid.

(6)

the cooling system energy dissipation is less than 10% under the op- erational conditions. Surface roughness was measured using a Mitutoyo (U.S.A) SJ210 surface roughness tester. The standard uncertainty of surface-roughness measurement is around 10%. The roughness diagram is presented inFig. 11. It is evident that surface roughness was constant (0.319μm or 319 nm) at the start of every experiment. To ensure the surface roughness is similar at the start of every experiment, the copper was polished with a sandpaper P600.

2.3. Measurement

To obtain the boiling curves, the boiling surface temperature (Ts) and thefluid saturation temperature must be known and the heatflux (q″) applied to thefluid must be calculated. Now, reading the temperature directly from the sensor on the surface is associated with a considerable error which is due to the additional nucleation site. In this study, the boiling surface temperature was measured indirectly through extrapolation. According toFig. 12, two holes were embedded into the copper block 15 and 38 mm away from the top side of the copper cylinder, 120° apart. 150 cc of either 0.01, 0.1, 0.5 or 1 vol% nanofluid was charged into the boiling chamber for each experiment. Then, the nanofluid was heated at pre-specified voltages and currents. To maintain a constant heat flux and ensure the stability of the system, the

temperature was recorded every 10 min (^dT =0.01

dt K min)

As discussed earlier, given the fact that the thermal conductivity of copper is much higher than that of PTFE, the temperature gradient in the radial direction can be ignored and the heat transfer may be assumed one-dimensional. The heatﬂux transported to the boiling surface is calculated as follows:

′ = −

− q′ KT T

d d

2 1

2 1 (2)

The heatflux may be obtained using Eq.(2): After the heatflux was determined, the difference between the temperature of the boiling surface and the saturation temperature can be determined using Eq.(3).

⎜ ⎟

= − = ⎛

⎝

− ′ ⎞

⎠

−

′

ΔT T T T q d

K T

S S sat 1 1 sat

(3) The BHTC was then calculated from the heatﬂux and the superheat temperature of the wall according to Eq.(4):

= ′^′ h q

ΔTS (4)

Fig. 7.Zeta-Potential test results for SiO₂nanoparticles with added Ps 20 surfactants with a concentration of 0.01 vol% in water as the baseﬂuid.

Fig. 8.The actual view of the apparatus.

(7)

3. Results and discussion 3.1. Boiling curve and repeatability

To ensure the accuracy of measurement, ﬁrst, the DI water was tested three times and was compared to the Rohsenow's Curve. The experimental relation that was proposed by Rohsenow may be ex- pressed as follows [63–65]:

⎜ ⎟

= ⎡

⎣

⎢ ⎛

⎝ −

⎞

⎠

⎤

⎦ ΔT h ⎥

C C q"

μ h σ

g(ρ ρ ) Pr

s fg p,l

sf

l fg l v

0.5 n 13

(5) hfgrepresents the latent heat theﬂuid,ΔTsis the temperature diﬀerence on the surface,σis the surface tension, andρlandρvare the liquid and vapor densities,Pris the Prandtl number and Csfand n are constants that, for the copper surface, are 0.013 and 1, respectively. Some Fig. 9.The schematic view of the apparatus.

Fig. 10.View of the copper cylinder.

(8)

researchers have measured the temperature directly and there have been a large of numbers of studies where the surface temperature was measured by extrapolation [19,20,32–34,38,63].

The uncertainty analysis based on the method proposed by Moﬀat [66] was carried out according to Eqs.(6)–(8).

⎜ ⎟ ⎜ ⎟

′ = ⎡

⎣⎢⎛

⎝ −

⎞

⎠ + ⎛

⎝ −

⎞

⎠

⎤

⎦⎥

′

− −

U ′

q

U T T

U d d

q T T d d

2 1

2

2 1

2

2 1 2 1

1 2

(6)

⎜ ⎟

⎜ ⎟ ⎜ ⎟

=⎡

⎣

⎢

⎛

⎝ −

⎞

⎠ + ⎛

⎝

⎞

⎠ + ⎛

⎝ ′

⎞

⎠

⎤

⎦

⎥

′

−

′

U ′

ΔT U T T

U d

U q

ΔT s

T T sat

d q

1 2

1

2 2

s 1 sat 1

1 2

(7)

⎜ ⎟ ^⎜ ^⎟

=⎡

⎣

⎢

⎛

⎝ ′

⎞

⎠ + ⎛

⎝

⎞

⎠

⎤

⎦

⎥

′

U ′

h U

q

U

, , ΔT .

h q ΔT

s

2 2

s 1 2

(8) The uncertainty for measurement of temperature is 0.1 K and 0.1 mm for the measurement of lengths. Also standard uncertainty for measuring surface roughness is about 10%. The results are suggestive the largest uncertainty in the calculation of the heatflux and the BHTC to be 6.5 and 8.4%.The heatflux is plotted inFig. 13against the wall superheat for the DI water. The heat transfer diagrams were identical after the experiments were repeated for three times in different days.

This means that the results will remain in this range regardless of how many times the experiments are repeated.

Fig. 11.Testing the roughness with a SJ 210 surface roughness tester (R_a= 0.319μm).

Fig. 12.Arrangement of PT-100 resistive sensors.

Fig. 13.The heat ﬂux with respect to the wall superheat in the DI water compared to the Rohsenow's correlation.

(9)

3.2. The boiling curve of DI water-surfactant

The heat flux is plotted Fig. 14against the wall superheat in DI water compared to the DI water-SLS surfactants. As evident from the figure, the results show that the diagram of the heat transfer coefficient of the DI water-surfactant is shifted to the left. Therefore, the BHTC for the DI water-surfactant solution is higher compared to the DI water for all heat fluxes. Adding surfactants has reduced the surface tension which facilitates the separation of the bubbles from the surface, enhancing the heat transfer. Fig. 14shows the boiling diagram for the specimen with added SLS surfactants, the diagram corresponding to specimens with Ps20 and CTAB are also similar. This means that for all heatfluxes, the heat transfer coefficient for the specimens with DI water and a surfactant is larger than that of DI water. To avoid redundancy, only the diagram corresponding to SLS is shown here. Moreover, it is evident that as the heatflux is increased and boiling continues in the nucleate boiling region, the BHTC is increased to a larger extent which is probably due to the acceleration of the separation of bubbles from the surface and the increased number of nucleation sites in the nucleate boiling region.

3.3. The boiling curve of SiO2- DI water nanofluid with SLS surfactant Fig. 15shows the boiling curve for the SiO2-water nanofluid containing SLS surfactants at four nanoparticle volume fractions (0.01, 0.1, 0.5, and 1 vol%) compared with DI water and DI water-surfactant specimens. As mentioned earlier, the incorporation of surfactants enhances heat transfer. The enhancement in the heat transfer is due to the nature of surfactants that reduce thefluid surface tension, facilitating the separation of bubbles from the surface and enhancing the heat transfer. Therefore, according toFig. 15which shows the variations in the heatflux with respect to the wall superheat, the heat transfer of DI water-surfactant is increased compared with DI water. Moreover, it is observed that increasing the concentration of the nanofluid (especially up to 0.5 and 1 vol%) reduces the heat transfer due to increased sediment. Sediment on the boiling surface at high heatfluxes increases the thermal resistance and reduces the heat transfer. In addition, the capacity of the surfactants (for reducing surface tension) is reduced at higher heatfluxes in comparison with the increase in the concentration of the nanoparticles for changing the heat transfer. Therefore, as the concentration of the nanoparticles is increased, the density of

nucleation sites is reduced which in turn reduces the heat transfer. The concentration of surfactant in nanoﬂuid for all prepared nanoﬂuids is₂₀¹ nanoparticles. Mass used.

3.4. The boiling curve of SiO2- DI water nanofluid with CTAB surfactant Fig. 16shows the boiling curve for the SiO2-water nanofluid containing CTAB surfactants at four nanoparticle volume fractions (0.01, 0.1, 0.5, and 1 vol%) compared with DI water and DI water-surfactant specimens. Similar to what was mentioned earlier, the incorporation of CTAB also reduces the surface tension and enhances heat transfer. Re- duced surface tension means the bubbles are separated from the surface more easily, thus enhancing the heat transfer. Therefore, according to Fig. 16, the boiling heat transfer in DI water-CTAB solution is enhanced Fig. 14.The heatflux against the wall superheat in the DI water compared to

the Rohsenow's correlation and the DI water-surfactant solution.

Fig. 15.Heatflux against the wall superheat for the nanofluid containing SLS and different concentrations in comparison with DI water and DI water-surfactant specimens.

Fig. 16.Heatflux against the wall superheat for the nanofluid containing CTAB and different concentrations in comparison with DI water and DI water-surfactant specimens.

(10)

compared with DI water. It is evident from thefigure that, similar to the case with SLS, as the volume concentration of the nanofluid is increased, heat transfer is reduced as a result of the sediment, especially at 0.5 and 1 vol% concentrations. The thermal resistance of the surface is increased by sediment on the boiling surface at higher heatfluxes, reducing the heat transfer. Similar to the case with SLS, it is evident that at higher heatfluxes and concentrations, the capacity of the surfactant (in reducing the surface tension) for overcoming the effect of the increased nanoparticle concentration is reduced, which results in fewer nucleation sites, and thus, reduced heat transfer.

3.5. The boiling curve of SiO2- DI water nanofluid with Ps20 surfactant The main application of Ps20 is in the food and cosmetics industry as an additive. Also one of the advantages of using it is the non-toxicity of this substance, which can be used in various cases, especially under sensitive and hazardous conditions.Fig. 17shows the boiling curve for the SiO2-water nanofluid containing Ps20 surfactants at four nanoparticle volume fractions (0.01, 0.1, 0.5, and 1 vol%) compared with DI water and DI water-surfactant specimens. Similar to previous discus- sions, the incorporation of Ps20 also reduces the surface tension and enhances heat transfer. Reduced surface tension means the bubbles are separated from the surface more easily, thus enhancing the heat transfer. Therefore, according toFig. 17, the boiling heat transfer in DI water-Ps20 solution is enhanced compared with DI water. According to Fig. 17, it is evident that the heat transfer coefficient was increased with the concentration increasing from 0.01 up to 0.1 vol%, However increasing the concentration to 1 vol% reduced the heat transfer drastically. The heat transfer is enhanced as the concentration is increased from 0.01 to 0.1 vol% as a result of the dominance of the impact of Ps20 which reduces the surface tension of the nanofluid. However, when the concentration was increased up to 1 vol%, the reduction in the surface tension caused by the surfactant became less significant compared to the increase in the concentration of the nanoparticles, sediment was increased, the number of nucleation sites was reduced, and the surface resistance was increased, thus reducing the heat transfer. Given that Ps20 surfactant is a non-ionic surfactant, it cannot help the stability of nanofluids than other surfactants, so it can bond nanoparticles and precipitation more than other surfactants on the surface, thereby reducing BHTC. Another reason may be that Ps20 surfactant has a very

long chain that causes the long chain adheres to the surface and causes more adhesion and precipitation on the surface, thereby reducing BHTC.

3.6. Comparing the boiling curve the SiO2-Water nanoﬂuid with diﬀerent surfactants

In order to investigate the impact of different anionic, cationic, and nonionic surfactants on the BHTC of the SiO2-DI water nanofluid, three surfactants were used (SLS, CTAB, and Ps20).Fig. 18shows the boiling heat transfer curve for the DI-water-based SiO2nanofluid with Ps20, CTAB, and SLS surfactants. As evident from thefigure, at all heatfluxes, the nanofluid specimens with added SLS surfactants were associated with the highest boiling heat transfer, which was followed by the nanofluid with added CTAB, and the nanofluid with Ps20. Therefore, it is safe to say that nonionic, cationic, and anionic surfactants have the largest to smallest impact on the boiling heat transfer of the nanofluids.

This is also in agreement with theﬁndings of previous studies [67].

Ultimately, ranking them by their impact on the heat transfer of the SiO2-DI water nanoﬂuid, SLS surfactants, followed by CTAB, and then Ps20 were the most to least eﬀective surfactants to enhance the heat transfer (SLS > CTAB > Ps20).

4. Conclusion

This study is an experimental investigation of the pool-boiling heat transfer in a nanofluid composed of deionized (DI) water and silicon dioxide (SiO2) in with added anionic, cationic, and nonionic surfactants (SLS, CTAB, Ps20) at various nanofluid concentrations and under atmospheric pressure conditions. The nanofluid was synthesized using a two-step method and showed acceptable stability. The presence of the surfactants in the nanofluid reduces the surface tension, thus enhancing the Boiling Heat Transfer Coefficient (BHTC) through facilitating the separation of the bubbles from the boiling surface. The results of this study may be summarized as follows:

•

The incorporation of surfactants increases the heat transfer coeﬃ- cient. In addition, at higher heat ﬂuxes and concentrations, heat transfer is reduced.

•

The BHTCs of the specimens composed of DI water and any of the Fig. 17.Heatﬂux against the wall superheat for the nanoﬂuid containing Ps20

and diﬀerent concentrations in comparison with DI water and DI water-surfactant specimens.

Fig. 18.Heat ﬂux with respect to the wall superheat for the SiO₂-DI water nanoﬂuid with surfactants compared with DI water-surfactant at 0.5 vol%

concentration.

(11)

surfactants were higher than that of the DI water.

•

The BHTCs of the specimens composed of DI water and any of the surfactants were also higher than that of the SiO2-DI water-surfactant nanoﬂuid.

•

The BHTC is reduced by increasing the concentration of the SiO2

nanoﬂuids containing SLS and CTAB surfactants.

•

In the specimen with the Ps20 surfactant, the heat transfer coeﬃ- cient was increased with the concentration increasing from 0.01 up to 0.1 vol%, However increasing the concentration to 1 vol% reduced the heat transfer drastically.

•

This study is theﬁrst to use the nonionic Ps20 surfactant to stabilize the SiO2-DI water nanoﬂuid for measuring its BHTC.

•

At all heatfluxes, the nanofluid specimens with added SLS surfactants were associated with the highest boiling heat transfer, which was followed by the nanofluid with added CTAB, and the nanofluid with Ps20 (SLS > CTAB > PS20).

•

It can be thus concluded that anionic, cationic, and nonionic surfactants (nonionic < cationic < anionic) resulted in the largest to smallest increase in the BHTC of the nanoﬂuids.

References

[1] M.M. Sarafraz, A. Arya, F. Hormozi, V. Nikkhah, On the convective thermal performance of a CPU cooler working with liquid gallium and CuO/water nanoﬂuid: a Comparative study, Appl. Therm. Eng. 112 (2017) 1373–1381 S1359-4311 32896-4.

[2] A. Arya, M.M. Sarafraz, S. Shahmiri, S.A.H. Madani, V. Nikkhah, S.M. Nakhjavani, Thermal performance analysis of aflat heat pipe working with carbon nanotube- water nanofluid for cooling of a high heatflux heater, Heat Mass Transf. 54 (4) (2017) 985.

[3] M.M. Sarafraz, V. Nikkhah, M. Nakhjavani, A. Arya, Fouling formation and thermal performance of aqueous carbon nanotube nanoﬂuid in a heat sink with rectangular parallel microchannel, Appl. Therm. Eng. 123 (29) (2017).

[4] M. Nakhjavani, V. Nikkhah, M.M. Sarafraz, S. Shoja, M. Sarafraz, Green synthesis of silver nanoparticles using green tea leaves: experimental study on the morpholo- gical, rheological and antibacterial behavior, Heat Mass Transf. 53 (10) (2017) 3201.

[5] M.M. Sarafraz, V. Nikkhah, M. Nakhjavani, A. Arya, Thermal performance of a heat sink microchannel working with biologically produced silver-water nanoﬂuid: experimental assessment, Exp. Therm. Fluid Sci. 91 (2018) 509–519.

[6] M.M. Sarafraz, A.S. Fazel, Y. Hasanzadeh, A. Arabshamsabadi, S. Bahram, Development of a new correlation for estimating pool boiling heat transfer coeﬃ- cient of MEG/DEG/water ternary mixture, Chem. Ind. Chem. Eng. Q. CICEQ 18 (2012) 11–18.

[7] G.L. Morini, Single-phase convective heat transfer in microchannels: a review of experimental results, Int. J. Therm. Sci. 43 (2004) 631–665.

[8] M. Sarafraz, F. Hormozi, Qualitative investigation of the convective boiling heat transfer of dilute Al2O3-water/glycerol solution inside the vertical annuli, Bulg.

Chem. Commun. 46 (2014) 645–651.

[9] T. Paramatthanuwat, S. Boothaisong, S. Rittidech, K. Booddachan, Heat transfer characteristics of a two-phase closed thermosyphon using de ionized water mixed with silver nano, Heat Mass Transf. 46 (2010) 281–285.

[10] S. Aghakhani, A.H. Pordanjani, A. Karimipour, A. Abdollahi, M. Afrand, Numerical investigation of heat transfer in a power-law non-Newtonianfluid in a C-Shaped cavity with magneticfield effect usingfinite difference lattice Boltzmann method, Comput. Fluids 176 (2018) 51–67.

[11] M.M. Sarafraz, F. Hormozi, Pool boiling heat transfer to dilute copper oxide aqueous nanoﬂuids, Int. J. Therm. Sci. 90 (2015) 224–237.

[12] T. Halon, B. Zajaczkowski, S. Michaie, R. Rulliere, J. Bonjour, Experimental study of low pressure pool boiling of water from narrow tunnel surfaces, Int. J. Therm. Sci.

121 (2017) 348–357.

[13] B.A.F. Dehkordi, A. Abdollahi, Experimental investigation toward obtaining the effect of interfacial solid-liquid interaction and basefluid type on the thermal conductivity of CuO-loaded nanofluids, Int. Common. Heat. Mass. Transf. 97 (2018) 151–162.

[14] W. Ding, E. Krepper, U. Hampel, Quantitative prediction of critical heatﬂux in- itiation in pool andﬂow boiling, Int. J. Therm. Sci. 125 (2018) 121–131.

[15] M.H.K. Darvanjooghi, M. Pahlevaninezhad, A. Abdollahi, Investigation of the effect of magneticfield on mass transfer parameters of CO2 absorption using Fe3O4‐water nanofluid, Alche J. 63 (2017) 2167–2186.

[16] A. Karimi, M. Afrand, Numerical study on thermal performance of an air-cooled heat exchanger: eﬀects of hybrid nanoﬂuid, pipe arrangement and cross section, Energy Convers. Manag. 164 (2018) 615–628.

[17] M. Afrand, Experimental study on thermal conductivity of ethylene glycol containing hybrid nano-additives and development of a new correlation, Appl. Therm.

Eng. 110 (2017) 1111–1119.

[18] Y. Dehghani, A. Abdollahi, Experimental investigation toward obtaining a new correlation for viscosity of WO3 and Al2O3 nanoparticles-loaded nanoﬂuid within aqueous and non-aqueous baseﬂuids, J. Therm. Anal. Calorim. (2018) 1–16.

[19] M.R. Raveshi, A. Keshavarz, M.S. Mojarrad, S. Amiri, Experimental investigation of pool boiling heat transfer enhancement of alumina–water–ethylene glycol nano- ﬂuids, Exp. Therm. Fluid Sci. 44 (2013) 805–814.

[20] A. Abdollahi, M.R. Salimpour, N. Etesami, Experimental analysis of magneticfield effect on the pool boiling heat transfer of ferrofluid, Appl. Therm. Eng. 111 (2017) 1101–1110.

[21] S. Choi, D.A. Siginer, H.P. Wang (Ed.), Enhancing Thermal Conductivity of Fluids with Nanoparticles, Developments and Applications of Non-newtonian Flows, vol.

66, ASME, 1995, pp. 99–105 231/MD vol.

[22] C.Y. Yang, D.W. Liu, Eﬀect of nano-particles for pool boiling heat transfer of re- frigerant 141B on horizontal tubes, Int. J. Microscale Nanoscale Therm. Fluid Transp. Phenom. 3 (1) (2010) 233–243.

[23] V. Trisaksri, S. Wongwises, Nucleate pool boiling heat transfer of TiO2-R141b na- noﬂuids, Int. J. Heat Mass Transf. 52 (issues 5–6) (2009) 1582–1588.

[24] I.C. Bang, S.H. Chang, Boiling heat transfer performance and phenomena of Al2O3–

water nano-ﬂuids from a plain surface in a pool, Int. J. Heat Mass Transf. 48 (issue 12) (2005) 2420–2428.

[25] V. Sajith, R.M. Madhusoodanan, C.B. Sobhan, An experimental investigation of the boiling performance of water-based nanoﬂuids, Proceedings of the Micro/

Nanoscale Heat Transfer International Conference, Jan. 2008.

[26] S.K. Das, N. Putra, W. Roetzel, Pool boiling characterization of nano-ﬂuids, Int. J.

Heat Mass Transf. 46 (2003) 851–862.

[27] S. Witharana, Boiling of Refrigerants on Enhanced Surfaces and Boiling of Nanoﬂuids, PhD thesis Royal Institute of Technology, Stockholm, Sweden, May of 2003.

[28] H. Kim, J. Kim, M. Kim, Experimental study on the characteristics and mechanism of pool boiling CHF enhancement using nano-ﬂuids, ECI International Conference on Boiling Heat Transfer, Spoleto, 7–12 May, 2006.

[29] D. Wen, Y. Ding, Experimental investigation into the pool boiling heat transfer of aqueous based c-alumina nanoﬂuids, J. Nanoparticle Res. 7 (2005) 265–274.

[30] H. Kim, J. Kim, M.H. Kim, Eﬀect of nanoparticles on CHF enhancement in pool boiling of nano-ﬂuids, Int. J. Heat Mass Transf. 49 (2006) 5070–5074.

[31] R.A. Taylor, P.E. Phelan, Pool boiling of nanoﬂuids: comprehensive review of ex- isting data and limited new data, Int. J. Heat Mass Transf. 52 (2009) 5339–5347.

[32] A. Abdollahi, M.R. Salimpour, Experimental investigation on the boiling heat transfer of nanofluids on aflat plate in the presence of a magneticfield, Eur. Phys. J.

Plus 131 (2016) 414.

[33] M.R. Salimpour, A. Abdollahi, M. Afrand, An experimental study on deposited surfaces due to nanoﬂuid pool boiling: comparision between rough and smooth surfaces, Exp. Therm. Fluid Sci. 88 (2017) 288–300.

[34] A. Abdollahi, M.R. Salimpour, N. Etesami, Experimental analysis of pool boiling heat transfer of ferroﬂuid on surfaces deposited with nanoﬂuid, Appl. Modares Mech. Eng. 16 (No. 2) (2016) 19–30.

[35] Z. Shahmoradi, N. Etesami, M. Nasr Esfahany, Pool boiling characteristics of na- noﬂuid onﬂat plate based on heater surface analysis, Int. Commun. Heat Mass Transf. 47 (2013) 113–120.

[36] B. Raei1, F. Shahraki, M. Jamialahmadi, S.M. Peyghambarzadeh, Experimental investigation on the heat transfer performance and pressure drop characteristics of γ-Al2O3/water nanoﬂuid in a double tube counterﬂow heat exchanger, Trans.

Phenom. Nano Micro Scales 5 (1) (2017) 64–75.

[37] M. Kole, T. Dey, Investigations on the pool boiling heat transfer and critical heat ﬂux of ZnO ethylene glycol nanoﬂuids, Appl. Therm. Eng. 37 (2012) 112–119.

[38] M. Dadjoo, N. Etesami, M.N. Esfahany, Influence of Orientation and Roughness of Heater Surface on Critical Heat Flux and Pool Boiling Heat Transfer Coefficient of Nanofluid, Applied Thermal Engineering, 2017.

[39] D.P. Kulkarni, P.K. Namburu, H. Ed Bargar, D.K. Das, Convective Heat Transfer and Fluid Dynamic Characteristics of SiO2 Ethylene Glycol/Water Nanoﬂuid, (2011).

[40] P. Vassallo, R. Kumar, S. D'Amico, Pool boiling heat transfer experiments in silica- water nano-ﬂuids, Int. J. Heat Mass Transf. 47 (2004) 407–411.

[41] E. Forrest, E. Williamson, J. Buongiorno, L.W. Hu, M. Rubner, R. Cohen, Augmentation of nucleate boiling heat transfer and critical heatﬂux using nanoparticle thin-ﬁlm coatings, Int. J. Heat Mass Transf. 53 (2010) 58–67.

[42] W. Wu, H. Bostanci, L.C. Chow, Y. Hong, M. Su, J.P. Kizito, Nucleate boiling heat transfer enhancement for Water and FC-72 on titanium oxide and silicon oxide surfaces, Int. J. Heat Mass Transf. 53 (2010) 1773–1777.

[43] J. Bravo, L. Zhai, Z. Wu, R.E. Cohen, M.F. Rubner, Transparent superhydrophobic ﬁlms based on silica nanoparticles, Langmuir 23 (2007) 7293–7298.

[44] S. You, J. Kim, K. Kim, Eﬀect of nanoparticles on critical heatﬂux of water in pool boiling heat transfer, Appl. Phys. Lett. 83 (2003) 3374.

[45] S. Witharana, Boiling of Refrigerants on Enhanced Surfaces and Boiling of Nanoﬂuids, Ph.D. thesis The Royal Institute of Technology, 2003.

[46] S.J. Kim, I.C. Bang, J. Buongiorno, L.W. Hu, Surface wettability change during pool boiling of nanofluids and its effect on critical heatflux, Int. J. Heat Mass Transf. 12 (2007) 4105–4116.

[47] B.H. Truong, Determination of Pool Boiling Critical Heat Flux Enhancement in Nanoﬂuids, Undergraduate Thesis, MIT, May 2007.

[48] M.A. Ahmed, M.Z. Yusoﬀ, K.C. Ng, N.H. Shuaib, Numerical and Experimental Investigations on the Heat Transfer Enhancement in Corrugated Channels Using SiO2-Water Nanoﬂuid, Case Studies in Thermal Engineering, 2015.

[49] D. Milanova, R. Kumar, Role of ions in pool boiling heat transfer of pure and silica nanoﬂuids, Appl. Phys. Lett. 87 (23) (2005) 233107.

[50] D. Milanova, R. Kumar, Heat transfer behaviour of silica nanoparticles in pool boiling experiment, J. Heat Transf. 130 (2008) 0424011–6.

[51] B.H. Shah, R. Darby, The effect of surfactant on evaporative heat transfer in vertical filmflow, Int. J. Heat Mass Transf. 16 (1973) 1889–1903.

[52] C.S. Chang, J.R. Maa, Y.M. Yang, Dynamic surface eﬀect and nucleateﬂow boiling

(12)

Phys. Lett. 94 (2009) 073107.

[57] V.M. Wasekar, R.M. Manglik, Pool Boiling Heat Transfer in Aqueous Solutions of an Anionic Surfactant, Thermal-Fluids and Thermal Processing Laboratory, 2000.

[58] V.M. Wasekar, R.M. Manglik, The inﬂuence of additive molecular weight and ionic nature on the pool boiling performance of aqueous surfactant solutions, Int. J. Heat Transf. 45 (2002) 483–493.

[59] V.M. Wasekar, Heat transfer in nucleate pool boiling of aqueous SDS and triton X- 100 solutions, Heat Mass Transf. 45 (2009) 1409–1414.

Transf. 53 (2010) 85–94.

[65] W.M. Rohsenow, A method of correlating heat transfer data for surface boiling li- quids, Trans. ASME 74 (1951) 969–979.

[66] R.J. Moﬀat, Describing the uncertainties in experimental results, Exp. Therm. Fluid Sci. 1 (1988) 3–17.

[67] W.T. Wu, Y.M. Yang, J.R. Maa, Enhancement of nucleate boiling heat transfer and depression of surface tension by surfactant additives, J. Heat Transf. 117 (1995) 526–529.