Accepted Manuscript

Layer-by-layer carbon nanotube coatings for enhanced pool boiling heat transfer on
metal surfaces

Seunghyeon Lee, Gwang Hyeok Seo, Sanghyeok Lee, Uiju Jeong, Sang Jun Lee,
Sung Joong Kim, Wonjoon Choi

PII: S0008-6223(16)30490-0
DOI: 10.1016/j.carbon.2016.06.039
Reference: CARBON 11079

To appear in: Carbon

Received Date: 30 April 2016

Accepted Date: 15 June 2016

Please cite this article as: S. Lee, G.H. Seo, S. Lee, U. Jeong, S.J. Lee, S.J. Kim, W. Choi, Layer-by-
layer carbon nanotube coatings for enhanced pool boiling heat transfer on metal surfaces, Carbon
(2016), doi: 10.1016/j.carbon.2016.06.039.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

1 Layer-by-Layer Carbon Nanotube Coatings for

2 Enhanced Pool Boiling Heat Transfer on Metal

PT
3 Surfaces

RI
4

SC
5 Seunghyeon Lee1‡, Gwang Hyeok Seo2‡, Sanghyeok Lee1, Uiju Jeong2, Sang Jun Lee1, Sung

U
6 Joong Kim2*, and Wonjoon Choi1* AN
7

1
8 School of Mechanical Engineering, Korea University, Seoul 136-701, Korea
M

2
9 Department of Nuclear Engineering, Hanyang University, 222 Wangsimni-ro, Seongdong-gu,
D

10 Seoul 133-791, Republic of Korea

TE

11
EP

12 *Authors to whom correspondence should be addressed: Wonjoon Choi and Sung Joong Kim

13 E-mail: [email protected], [email protected], Phone: +82 2 3290 5951, Fax: +82 2

C

14 926 9290.
AC

15 ‡ These authors contributed equally to this work.

16

17

1
 ACCEPTED MANUSCRIPT

1 Abstract

2 Micro/nanotextured coatings have enabled the manipulation of thermal characteristics in pool

3 boiling heat transfer such as the heat transfer coefficient (HTC) and critical heat flux (CHF)

PT
4 because of the ability to optimize bubble formations and departures. However, fabricating such

RI
5 coatings on substrates involves high cost-bulky setup, and is limited by the materials and

6 adhesion properties. Herein, we report layer-by-layer (LbL)-assembled polyethylenimine (PEI)-

SC
7 multi-walled carbon nanotube (MWCNTs) coatings on stainless steel (SS316) to enhance HTC

8 and CHF in pool boiling heat transfer. LbL-assembled PEI-MWCNTs coatings (10, 20, 40 bi-

9
U
layers) on SS316 were fabricated, whereas bare SS316 substrate and vacuum-filtered MWCNTs
AN
10 coating were prepared as controls. Because of the nano-cavities and the inner-nanoporous

11
M

structures, LbL coatings showed significant enhancement of HTCs and CHFs compared to bare

12 SS316. Furthermore, denser networks of MWCNTs due to electrostatic bonding and hydrophilic
D

13 nature of PEI in the LbL coatings could enhance the HTCs and CHFs compared to vacuum-
TE

14 filtered MWCNTs coating. LbL-assembled PEI-MWCNTs of 20 bi-layers showed the highest

15 improvement in HTC, whereas the 40 bi-layers coatings achieved the best enhancement ratio of
EP

16 CHF, (~147 %). Further development of the LbL-assembled coatings on metal surfaces would

17 enable potential applications for thermal management, from micro/nanoscale platforms to

C

18 macroscale systems.
AC

19

20

21

22

2
 ACCEPTED MANUSCRIPT

1 1. Introduction

2 The development of micro/nanoscale platforms has revolutionized both existing products and

3 new applications such as central processing units, memory devices, circuit systems, micro fuel

PT
4 cells, and batteries. Meanwhile, the rising temperatures due to restricted structures in the

5 micro/nanoscale platforms has emerged as a big obstacle for further enhancement of the overall

RI
6 performance and the operating stability. Furthermore, even in macroscale systems including

SC
7 power plants, thermal energy management is quite critical to improve the efficiency and to

8 maintain the thermal safety margin. For example, the failure of thermal energy dissipation in

U
9 nuclear power plants may lead to a damage of the fuel rod. Liquid-vapor phase change heat
AN
10 transfer is one of the most significant phenomena to be addressed for managing such thermal

11 energy in multiscale systems.[1] The heat transfer coefficient (HTC) of liquid-vapor phase
M

12 change, which is much higher than that of the single phase, allows efficient cooling performance
D

13 to dissipate thermal energy or to maintain operating temperatures in small devices and

TE

14 macroscale systems.[2] As well, achieving higher critical heat flux (CHF) extends the thermal

15 operability of the systems.

EP

16

17 The overall characteristics of liquid-vapor phase change heat transfer significantly depend on the
C

18 solid-surface conditions that are directly exposed to the working fluid. In general, porous
AC

19 structures are known to increase the site density of bubble nucleation and enhance the HTC.[3]

20 In addition, the wetting properties of surfaces are regarded as other crucial factors to manipulate

21 the HTC.[4] Furthermore, the surface conditions can affect another critical factor of phase

22 change heat transfer, viz. CHF, which determines the maximum applicable power density in the

23 thermal systems.[5, 6] The CHF is the upper limit of nucleate boiling heat transfer during liquid-

3
 ACCEPTED MANUSCRIPT

1 vapor phase heat transfer. When the system reaches the CHF, vapor blankets cover the contact

2 surfaces at the interface with working fluids. Such a circumstance results in a dramatic decrease

3 in the HTC because of the low thermal conductivity of the vapor phase on the surface and a

PT
4 decrease in bubble departures from the surfaces.[7] The sudden drop in the thermal energy

5 dissipation causes a decline in the cooling performance, collapsing the stability of the systems.

RI
6

SC
7 Coatings of micro-nanostructured materials are ideal candidates for the manipulation of contact

8 surfaces in liquid-vapor phase change heat transfer. Three desirable physical conditions, porosity,

U
9 [8] roughness, [9] and wettability [10] should be optimized at the interface with the working
AN
10 fluids. Various micro-nanostructured materials such as silicon, metal, metal oxide, organic, and

11 inorganic materials have been tested as contact interfaces between working fluids and the heat
M

12 transfer substrates.[11] For example, copper,[12] aluminum,[13] CuO,[14] ZnO,[15] and TiO2

13 [16] effectively improved HTC or CHF compared to the bare substrates. Moreover, coatings of
D

14 specific micro/nanomaterials, which were composed of flower-like structures [17] or vertically

TE

15 aligned nanowires,[18] allowed controlling the HTC and CHF in identical materials. In particular,

16
EP

carbon nanotubes (CNTs) have been regarded as potential candidate materials for contact

17 interfaces between the working fluid and the heat transfer substrate. CNTs have ideal fiber-like
C

18 structures for the formation of porous structures with increased surface roughness, while the
AC

19 wettability of CNTs can be modified from hydrophobic to hydrophilic by a facile chemical

20 method.[19] In addition, the superior thermal conductivity of CNTs can contribute to the

21 minimization of the thermal resistance at the contact interfaces between CNTs and the heat

22 transfer substrates.[20] In fact, the use of CNTs at contact interfaces was effective in changing

23 both the HTC and CHF in pool [21, 22] and flow boiling heat transfer.[23, 24]

4
 ACCEPTED MANUSCRIPT

2 However, there are fundamental limitations of contact interfaces using micro/nanostructured

3 materials for boiling heat transfer applications. Micro/nanofabrication processes are inevitable

PT
4 for directly developing micro/nanotextured surfaces on the substrates, despite being high-cost

5 and not applicable to large-scale surface areas.[25] Moreover, the fabricated structures are not

RI
6 strong enough to endure the high pressure or harsh operating conditions of liquid-vapor transfer

SC
7 and bubble formation-departure in boiling heat transfer.[26] Another approach is the indirect

8 coating of pre-synthesized micro-nanostructured materials on the bare substrates by simple

U
9 solution processing methods.[27] In this case, the adhesion with the substrate or among the
AN
10 micro/nanomaterials is not strong because it is based on only weak physical contacts due to

11 surface energy. Moreover, it is quite difficult to form uniform structures and surfaces along the
M

12 textures of the original heat transfer substrate. In other studies, the difficulty of fabrication and

13 surface transfer, the weak adhesion and non-uniform surfaces formation were major problems
D

14 causing high cost, long processing times, and a limit on the applicable sizes.[28] In this respect,
TE

15 the development of a new coating method of micro/nanomaterials on the heat transfer substrate is

16
EP

essential to facilitate further application for the enhanced pool boiling heat transfer.

17
C

18 Layer-by-layer (LbL) deposition is a fabrication method for uniform thin film layers on various
AC

19 material substrates.[29, 30] The repetitive alternate immersion of substrates into a positively

20 charged and a negatively charged solution forms stable film layers along the original surfaces of

21 the substrates.[31-34] Compared to conventional coating methods, LbL deposition can be used

22 on large areas and requires simple processing equipment because it is based on solution-

23 processing methods rather than high temperature processing.[35] The manipulation of properties

5
 ACCEPTED MANUSCRIPT

1 is implemented by adjusting the bi-layer number,[36] by controlling the pH of suspensions,[37]

2 or by adopting different techniques such as immersive,[38-40] spin,[41] spray,[42] or

3 microfluidic LbL.[43] The LbL methods can be used to fabricate diverse coatings that are highly

PT
4 porous to form large surface area films of micro/nanostructured materials.[44] In particular, the

5 electrostatic force among bi-layers allows formation of more uniform, adhesive, and substantial

RI
6 coating layers than films depending on physical contacts to surfaces.

SC
7

8 Herein, we report LbL-assembled polyethylenimine (PEI)-multi-walled carbon nanotubes

U
9 (MWCNTs) coatings on stainless steel (SS316) metal surface to significantly enhance HTC and
AN
10 CHF in pool boiling heat transfer (Figure 1). SS316 is one of the most widely used substrates in

11 various thermal systems because of its low-cost and anti-corrosion property, but it has a
M

12 relatively low HTC and poor CHF to satisfy the requirements for fast-cooling and high power

13 applications. Nowadays, MWCNTs can be mass produced and are ideal nanomaterials for using
D

14 on surfaces for controlling boiling heat transfer because they have excellent thermal
TE

15 properties,[45, 46] good mechanical-chemical stability,[47] and ideal fiber-like structures for the

16
EP

formation of porous layers.[48] The hydrophilic property of PEI can control wettability by

17 mixing with the hydrophobic MWCNTs. In this work, we fabricated three LbL-assembled PEI-
C

18 MWCNTs coatings (10, 20, 40 bi-layers) on SS316, and measured the enhancement in HTC and
AC

19 CHF in pool boiling heat transfer. They were compared with the control performances of bare

20 SS316 substrate and MWCNTs-coated SS316 fabricated by vacuum filtration. These two

21 controls provide a basis to evaluate the effectiveness of LbL-based PEI-MWCNTs coatings.

22

23

6
 ACCEPTED MANUSCRIPT

1 2. Experimental

2 2.1 Preparation of SS316 substrates and test heater assembly. The bare substrate, SS316, was

3 polished with sand paper of grade #800, for all heat transfer substrates for pool boiling

PT
4 experiments. This procedure retained the microscale wave-like structures on SS316 surfaces.

5 After the polishing step, the bare SS316 substrates were cleaned with deionized (DI) water,

RI
6 acetone, ethanol, and DI water, sequentially.[49] The dimensions of the heating element were

SC
7 35×10×2 mm3 in length, width, and thickness, respectively. Since thin films of the coating

8 materials were deposited onto the SS316 surface, the heated area (35×10 mm2) for all the test

U
9 samples was identical to the bare substrate.
AN
10

11 2.2 Fabrication of LbL-assembled PEI-MWCNTs films on SS316. The MWCNTs were

M

12 purchased from JEIO, Korea (95 wt% purity, 5–20 µm length, 20–40 nm outer diameter). In
D

13 order to prepare the negatively-charged MWCNTs solution for the LbL immersion process,
TE

14 MWCNTs were refluxed with H2SO4/HNO3 (3/1 v/v, 95–98%/70% concentration) at 70 °C for 2

15 h [44]. They were filtered from the acid solution and washed using DI water several times. The
EP

16 filtered MWCNTs were dried at room temperature for 24 h to remove the solvent. Because of the

17 previous steps, the surfaces of MWCNTs were functionalized with carboxylic group (–COOH),
C

18 which made the negatively charged surface. The carboxylic-functionalized MWCNTs (0.1
AC

19 mg/mL) were sonicated for 1 h and uniformly dispersed in DI water. In order to prepare the

20 positively charged layers for the LbL immersion process, polyethyleneimine (PEI, 50% w/v in

21 water) was purchased from Aldrich. PEI solution (1 mg/mL) was diluted in DI water. The

22 prepared SS316 was repeatedly immersed in the positively charged solution (PEI, 5 min), rinsing

23 solution (DI water, 1 min), negatively charged solution (carboxylic-functionalized MWCNTs

7
 ACCEPTED MANUSCRIPT

1 solution, 5 min), and rinsing solution (DI water, 1 min) for the deposition. This cycle deposited

2 one bi-layer of LbL PEI-MWCNTs on SS316. The immersion process was repeated at the same

3 conditions until the desired number of PEI-MWCNTs bi-layers was achieved. Because each

PT
4 layer adhered to the others by a strong electrostatic force, and the non-adhered molecules were

5 removed by the rinsing process, stable and uniform coatings of PEI-MWCNTs on SS316 were

RI
6 achieved.

SC
7

8 2.3 Fabrication of vacuum-filtered MWCNTs coating on SS316. The MWCNTs were

U
9 uniformly dispersed in DI water using 1.5 wt% sodium dodecyl sulfate (SDS) solution. The
AN
10 prepared solution was continuously sonicated for 12 hours using a tip-sonicator. The MWCNT

11 solution was centrifuged at 3500 rpm and the upper portion of the solution[50] was decanted.
M

12 The decanted MWCNT solution was filtered on a porous alumina membrane (anodisc, 100 nm

13 pore size, 47 mm in diameter) using a vacuum filtration setup.[51] The alumina membrane with
D

14 MWCNT film was washed with DI water several times to remove surfactants until no bubbles
TE

15 were observed. It was then floated on 3 M sodium hydroxide (NaOH) solution to gradually etch

16
EP

the alumina membrane.[52] After etching, the remaining MWCNT film was floated on a NaOH

17 solution. DI water was repetitively circulated to reduce the pH to 7, and it was pulled until the
C

18 MWCNT film physically contacted the SS316 substrate. The heat transfer substrate of the
AC

19 MWCNTs-SS316 was finally fabricated after the MWCNTs-coated SS316 was annealed at 80 Cº

20 for 20 min.

21

22 2.4 Characterization of PEI-MWCNT coatings on SS316. The surface morphologies of bare

23 SS316, vacuum-filtered MWCNTs and LbL assembled PEI-MWCNTs coatings were measured

8
 ACCEPTED MANUSCRIPT

1 using field emission scanning electron microscope (FE-SEM) with low vacuum, Quanta 250

2 FEG. The surface roughness was measured by atomic force microscope (AFM) and the

3 thicknesses of MWCNTs coatings on SS316 were measured by Alpha Step. The measurement of

PT
4 thicknesses was conducted by placing the MWCNT films on the smooth SS316 surface to

5 prevent errors caused by inherent SS316 roughness. Static contact angle measurements were

RI
6 carried out using Krüss EasyDrop, a contact angle measurement system.

SC
7

8 2.5 Pool boiling heat transfer experiments. The bare SS316 surface, vacuum-filtered

U
9 MWCNTs coating, and LbL-assembled PEI-MWCNTs coatings (10, 20, and 40 bi-layers) on
AN
10 SS316 substrates were tested as heat transfer substrates for measuring HTCs and CHFs in pool

11 boiling heat transfer. All experiments for HTCs and CHFs were conducted in a pool of DI water
M

12 under atmospheric pressure. The test heater assembly was mechanically fixed using the copper

13 electrodes and submerged in DI water. The DC power supply of 1200A-10V supplied electrical
D

14 power to the test heater through the copper electrodes. A K-type TC having a 1 mm sheath
TE

15 diameter was submerged in the boiling chamber to measure bulk temperature of the water pool.

16 The nominal uncertainty of the TCs was ±1.1 oC, guaranteed by OMEGA Engineering. Data
EP

17 acquisition systems, cDAQs from National Instruments were used to acquire all related data,
C

18 such as temperature, voltage, and current. Since the test heater system was in a series circuit, the
AC

19 data acquisition systems accurately determined the electrical current. For the measurement of the

20 voltage drop across the test heater, two conductive wires were inserted into both ends of the

21 heater, which were corresponding to voltage tap positions. To capture bubble behavior during a

22 test, visualization work was also carried out at a rate of 2,000 frame/sec using the high-speed

23 video system (HVS) equipped with Phantom v7.3 and light sources. All the temperature, voltage,

9
 ACCEPTED MANUSCRIPT

1 and current were recorded by the data acquisition system, a cDAQ from National Instruments.

2 The nominal uncertainty of the TCs was ±1.1 oC guaranteed by OMEGA Engineering. Prior to

3 boiling tests and measurements, the DI water in the boiling chamber was maintained at the

PT
4 saturated temperature. The immersion heaters were heated to effectively eliminate incondensable

5 gas bubbles. When the gas bubbles completely disappeared, the heating power gradually

RI
6 increased until the heater reached the CHF. During the gradual increase of the heating power, the

SC
7 HTC for each heater was measured, while all experimental procedures were terminated when

8 CHF was observed. The applied heat flux was evaluated by the following equation:

U
Power IV
9 q" = = = V 1I 1W −1 L−1 (1)
Heat transfer area WL
AN
10 Where, I, V, W, and L are electrical current, voltage, heater width, and heater length, respectively.
M

11 Using the propagation of error method, uncertainties of the heat flux were determined as follows:

2
 U q''   UV 2  U I 2  UW 2  U L 2
D

12  ''  =   + I  + W  + L  ,
 q   V       
TE

1/ 2
U q''   0.2 ×10−3   2 ×10−3  
2 2

= ( 0.005) + ( 0.005) +   + = 0.061

2 2
13   (2)
q ''   10 ×10−3   35 ×10−3  
 
EP

14 Where, the uncertainties on the voltage (Uv/V), current (UI/I), heated width (UW/W), and length
C

15 (UL/L) were less than 0.5, 0.5, 2.0, and 5.7 %, respectively. The resulting propagated uncertainty
AC

16 (Uq’’/q’’) was calculated to be 6.1 %. It should be noted that both sides of the test heaters were

17 completely insulated with epoxy, whereas the central heated region was only exposed. Thus,

18 boiling only occurred on the central heated region, and the heat losses through the supporting

19 structures were ruled out.

20

10
 ACCEPTED MANUSCRIPT

1 3. Results and Discussion

2 3.1. Physical Characterization of PEI-MWCNTs, MWCNTs, and SS316 Substrates.

3 Five different types of heat transfer substrates were prepared to confirm the changes of pool

PT
4 boiling heat transfer via LbL-assembled coatings; bare SS316, vacuum-filtered MWCNTs

5 coating on SS316, and LbL-assembled PEI-MWCNTs coatings (10, 20, and 40 bi-layers) on

RI
6 SS316. Dip-LbL coating method for large-area deposition was used to fabricate the LbL-

SC
7 assembled PEI-MWCNTs coatings that were bonded by electrostatic forces. This is the simplest

8 technique for preparing centimeter-scale target substrates without any further processing (Figure

U
9 1). The pool boiling heat transfer experiments were conducted for all types of substrates. In order
AN
10 to elucidate the effectiveness of the LbL-assembled coatings, thermal characteristics such as

11 HTC, CHF, bubble formation-departure were compared. Furthermore, the PEI-MWCNTs

M

12 coatings with different bi-layers (10, 20, and 40 bi-layers) were tested in order to understand the
D

13 detailed effects using LbL methods.

TE

14

15 The surface conditions of bare SS316, LbL-assembled PEI-MWCNTs with different bi-layers,
EP

16 and vacuum filtered MWCNTs on SS316 were investigated by scanning electron microscopy

17 (SEM), as shown in Figure 2. The bare SS316 substrates revealed the microscale wave-like
C

18 structures after they were polished by sand papers (grade #800) (Figure 2a). The vacuum-filtered
AC

19 MWCNTs, which were physically transferred to the bare SS316, formed the porous structures

20 from the randomly oriented MWCNTs networks. But they were not able to preserve the

21 microscale wave-like structures (Figure 2b). This indicates that it is difficult to maintain

22 complete contact between the MWCNTs coating and SS316 with such a physical transfer.

23 Therefore there would be empty space near the contact regimes, increasing the thermal resistance

11
 ACCEPTED MANUSCRIPT

1 near the contact interfaces of MWCNTs and SS316 substrate. On the other hand, LbL-assembled

2 PEI-MWCNTs coatings show highly nanoporous structures because of the randomly oriented

3 MWCNTs networks and preserve the wave-like structures in the original SS316 substrate (Figure

PT
4 2c, 2d, and 2e). All PEI-MWCNTs coatings show the similar nanoporous structures regardless of

5 the number of bi-layers. Moreover, the electrostatic bonding and rinsing allow maintaining a

RI
6 uniform thickness for one bi-layer of PEI-MWCNTs along the microtextures of the SS316,

SC
7 without any empty space near the contact interfaces. This implies that LbL-assembled coatings

8 can add additional porous structures without changing the original microtextures, and this

U
9 condition can restrain the increase in thermal resistance at the contact interfaces. However, there
AN
10 was a transition depending on the number of bi-layers. The PEI-MWCNTs coatings with a

11 moderate numbers of bi-layers (10 and 20 bi-layers, Figure 2c and 2d) nearly followed the
M

12 original microtextures in the SS316 substrates, whereas the increase in the number of bi-layers

13 (40 bi-layers, Figure 2e) revealed a little loss of the original microtextures. The relative density
D

14 of MWCNTs in different coatings can be estimated from the previous works.[44, 53] It was
TE

15 previously confirmed that degrees of adsorption and interpenetration of CNTs depended on

16
EP

charge densities. Thus, the electrostatic force in LbL assembly would be able to form the dense

17 networks of CNTs, in comparison with other physical methods.[44] Furthermore, it was

C

18 investigated that the density of MWCNTs in LbL assembled PEI-MWCNT films was
AC

19 significantly proportional to the number of bi-layers.[53] For those reasons, the LbL assembled

20 PEI-MWCNTs coatings might show the higher density of MWCNTs, than the vacuum filtered

21 coatings. Also, the higher bi-layer LbL assembled PEI-MWCNTs coatings might have the more

22 dense structures.

23

12
 ACCEPTED MANUSCRIPT

1 Atomic force microscope (AFM) images can provide further information about surface

2 roughness. The surface roughness of the bare SS316 after polishing was 161 nm, which is mainly

3 attributed to the wave-like microstructures (Figure 3a). However, in the case of the vacuum-

PT
4 filtered MWCNTs coating, although nanoporous textures appeared, the surface roughness

5 decreased from 161 nm to 102 nm because of the disappearance of the wave-like microtextures

RI
6 (Figure 3b). Because vacuum-filtered MWCNTs coating was physically transferred on SS316

SC
7 substrate after the fabrication, SS316 surface was completely covered with the film like the

8 blanket of MWCNTs networks. As a result of that, wave-like microtextures on SS316 mostly

U
9 disappeared and the surface roughness of vacuum filtered MWCNTs film only reflected the
AN
10 actual surface roughness of MWCNTs networks (Figure 3b). Therefore, the surface roughness of

11 vacuum-filtered MWCNTs film on SS316 decreased (~102 nm), while the surface roughness
M

12 from the wave-like structures in the bare SS316 was 161 nm. The LbL-assembled PEI-MWCNTs

13 coatings of different bi-layers clearly preserved the wave-like microtextures of SS316 as well as
D

14 the appearance of nanoporous surfaces from the random network of PEI-MWCNTs. The surface
TE

15 roughness of 10, 20, and 40 bi-layers PEI-MWCNTs coatings on SS316 substrates was 237 nm,

16
EP

307 nm, and 286 nm, respectively (Figure 3c, 3d, and 3e). They all showed higher values than

17 bare SS316 surface roughness (161nm), because the measured value was the summation of the
C

18 PEI-MWCNTs coating, itself and the wave-like microstructures on SS316. As shown in Figure
AC

19 3c, 3d, and 3e, the wave-like microstructures were preserved after the deposition of the LbL

20 assembled PEI-MWCNTs coatings, and it counted on the overall surface roughness. This

21 characteristic of the surface induced the relatively large surface roughness, in comparison with

22 its original thickness. This trend agrees with the qualitative analysis from the SEM images

23 (Figure 2). The 10 bi-layers may not be sufficient to fabricate complete nanoporous surfaces

13
 ACCEPTED MANUSCRIPT

1 because of the lack of density of MWCNTs, as shown in Figure 3c. In this case, the nanoporous

2 surfaces were present partially, while some parts of the surfaces showed relatively smooth

3 surfaces. For the 20 bi-layer PEI-MWCNTs coatings, both nanoporous surfaces and

PT
4 microtextures appeared uniformly, and it resulted in the further increase of surface roughness up

5 to 307 nm (Figure 3d). However, the 40 bi-layer PEI-MWCNTs coating gradually lost the

RI
6 microtextures while the nanoporous surfaces were preserved. This transition in physical

SC
7 dimensions might induce a decrease in the surface roughness to 286 nm (Figure 3e).

U
9 The thickness of the porous coating on SS316 can be another critical factor that affects the
AN
10 boiling heat transfer because it influences the capillary wicking. The thicknesses of a vacuum-

11 filtered MWCNTs coating and LbL-assembled PEI-MWCNTs coatings were compared (Figure
M

12 4). The thickness of the vacuum-filtered MWCNTs coating, 1043 nm, was used as a reference

13 value. The thicknesses of the LbL-assembled PEI-MWCNTs depend on the total number of bi-
D

14 layers. The average thicknesses of the PEI-MWCNTs coatings with 10, 20, and 40 bi-layers were
TE

15 275 nm, 618 nm, and 789 nm, respectively. It is known that the thickness growth of LbL films is

16
EP

linearly proportional to the number of bi-layers when the electrostatic force dominates between

17 the bi-layers.[39] However, when the LbL deposition reaches a critical point, the electrostatic
C

18 force is insufficient to attract a new layer. Therefore, the growth rate decreases, and the increase
AC

19 of coating thickness saturates at some point.[54] The thicknesses of the LbL-assembled PEI-

20 MWCNTs of 20 and 40 bi-layers were measured as 618 nm and 789 nm, respectively. This

21 relatively small differential increase between 20 and 40 bi-layers indicates that the LbL-

22 assembled PEI-MWCNTs probably reach saturation between 20 and 40 bi-layers, and there is a

23 sudden decrease in the growth rate of coating thickness. This factor might induce a decrease in

14
 ACCEPTED MANUSCRIPT

1 the surface roughness, as discussed in Figure 3e. Meanwhile, at the initial stage of LbL

2 deposition, there is a minimum coating thickness for the stabilization of the growth rate.[55] The

3 thickness of 10 and 20 bi-layers shows an almost linear increase with the number of bi-layers. If

PT
4 the initially stabilized growth rate of LbL coatings is considered, thicknesses of 275 nm and 618

5 nm for 10 and 20 bi-layers LbL assembled PEI-MWCNTs, respectively, agree with the linear

RI
6 tendency of well-constructed LbL films. It should be noted that thinner coatings benefit more

SC
7 from the perspective of material cost. In this work, LbL-assembled coatings used less MWCNTs

8 than the vacuum-filtered MWCNTs coating.

U
9
AN
10 The wettability of surfaces is an important factor in determining the liquid-vapor phase change at

11 the contact interfaces during the supply of thermal energy. Figure 5 shows the static contact
M

12 angles of the bare SS316 substrate after polishing, vacuum-filtered MWCNTs coating, and LbL

13 assembled PEI-MWCNTs coatings on SS316. The contact angles were measured using the Kruss
D

14 EasyDrop. The contact angle of the bare SS316 substrate was 82.9º, which was similar to a
TE

15 typical value for stainless steel alloy (Figure 5a).[56] Since MWCNTs have a hydrophobic

16
EP

nature,[57] the vacuum-filtered MWCNTs on SS316 show a higher contact angle of about 106.4º

17 (Figure 5b). Although the intrinsic contact angle of CNTs reaches 167 ± 3º,[58] the nanoporous
C

18 structure that is formed by the MWCNTs network might cause capillary wicking to imbibe the
AC

19 contacting water (Figure 5b).[59] On the other hand, LbL-assembled PEI-MWCNTs with 10, 20,

20 and 40 bi-layers show significantly lower contact angles, between 22º and 27º (Figure 5c, 5d, and

21 5e), than that of the bare SS316 and vacuum-filtered MWCNTs coating. In this measurement, it

22 was confirmed that the wettabilities of the LbL-assembled PEI-MWCNTs coatings were

23 dominated by the physiochemical properties of the base-materials, rather than the thickness of

15
 ACCEPTED MANUSCRIPT

1 the coatings. PEI is a hydrophilic polymer,[60] and coating layers of PEI-MWCNTs can provide

2 a hydrophilic nature, in addition to the hydrophobic property of MWCNTs. Because the

3 hydrophilic nature promotes capillary wicking during nucleate boiling heat transfer, it can help

PT
4 liquid circulation at the contact interfaces and promote the continuous formation of atomized

5 bubbles.[61] At the same time, it can prevent the transition to the film boiling heat transfer stage.

RI
6

SC
7 3.2. HTC and CHF for LbL-assembled PEI-MWCNTs Coatings

8 Pool boiling heat transfer experiments were conducted to evaluate the thermal characteristics

U
9 such as HTC, CHF, and bubble formation of the bare SS316 substrate, vacuum-filtered
AN
10 MWCNTs coating, and LbL-assembled PEI-MWCNTs coatings on SS316. As shown in Figure

11 6a, the test heater assembly consisted of the central heating element and two copper blocks to
M

12 supply electrical power. Except the central heating surface, every side of the heating element was

13 insulated with epoxy. On the backside of the heater, a K-type thermocouple (TC) was attached
D

14 with the solidified epoxy insulator to measure the heater temperature. Two conductive wires
TE

15 were inserted into both the copper blocks to measure the voltage drops across the heater. The

16
EP

current flowing across the heater was measured using a data acquisition system. A set of heater

17 assembly, power supply, high speed video system, boiling chamber, and data acquisition systems
C

18 are utilized for the HTC and CHF experiments (Figure 6b).
AC

19

20 During the experiment, a test substrate was heated with gradual increase in the heat flux rate for

21 each step. As the heat flux increased, a phase change from liquid to vapor occurred on the heater

22 surface at some point. The absorbed thermal energy from the SS316 heater converted liquid to

23 vapor on the surface, and led to the departure of the nucleate bubbles in the vapor phase from the

16
 ACCEPTED MANUSCRIPT

1 substrate. As the heat flux increased and reached the critical point, the density of vapor bubbles

2 was high, and a blanket of vapor phase covered the heat transfer surface. At this time, a sudden

3 drop in the heat flux occurred because of the low HTC of the vapor phase at the coating surfaces

PT
4 as well as a rapid decrease in the bubbles leaving from the substrate. The heat flux at this point

5 was measured as the CHF, while the change in the ratio between the heat flux and the wall

RI
6 superheat was converted to the HTC of the heat transfer substrate. It should be noted that the

SC
7 adhesion of coatings on the heat transfer substrates is quite important because the attached films

8 should remain stable during the pool boiling heat transfer. The generation and departure of vapor

U
9 bubbles from the substrates normally induces a strong pressure locally. Since LbL-assembled
AN
10 coatings on SS316 used electrostatic force rather than just physical contact depending on the

11 surface energy, they were able to form more solid and stable coatings on the substrates. In terms
M

12 of the stability of coatings, LbL-assembled PEI-MWCNTs coatings showed improved

13 performance compared to the vacuum-filtered MWCNTs coating. The quantitative analysis of

D

14 the surface stability was conducted by means of using micro scratch test. In this measurement,
TE

15 LbL assembled PEI-MWCNTs coatings showed the distinct section between cohesive failure

16
EP

(crack of film) and adhesive failure (detached from substrate), whereas, vacuum filtered

17 MWCNT coating just revealed only adhesive failure (Figure S1). This result indicates that LbL
C

18 assembled PEI-MWCNTs coatings are more stable and resistible against the applied force on
AC

19 surfaces. For the further characterization, the critical loads on onset of failure have been

20 measured, as well (Figure S2). The critical loads of LbL assembled PEI-MWCNTs coatings were

21 26, 70 and 88 mN in 10, 20 and 40 bi-layers, whereas that of a vacuum filtered MWCNTs

22 coating was 23 mN. It indicates that stronger critical loads have to be applied for damaging LbL

23 assembled PEI-MWCNTs coatings on SS316, compared with vacuum filtered MWCNTs

17
 ACCEPTED MANUSCRIPT

1 coatings. In fact, during the pool boiling experiments, the LbL-assembled PEI-MWCNTs coating

2 films were mostly stable, whereas the vacuum-filtered MWCNTs films were damaged sometimes.

3 Actually, the bubble nucleation temperature with nanocavities (~ 90 nm for LbL assembled PEI-

PT
4 MWCNTs nanoporous structures) was estimated about 200 ºC. Based on water-steam tables, the

5 vapor pressure at 200 ºC (bubble nucleation temperature) was 1.6 MPa.[62] The critical load of

RI
6 LbL assembled PEI-MWCNTs coatings can be also converted to the pressure with scratch tip

SC
7 radius, 20µm. The critical pressure at 50 mN is roughly converted to 39 MPa. It indicates that

8 this critical pressure is quite greater than 1.6 MPa, and the LbL assembled PEI-MWCNTs

U
9 coatings would be quite stable in the pool boiling heat transfer.
AN
10

11 Boiling curves and corresponding HTCs of the bare SS316 and LbL-assembled PEI-MWCNTs
M

12 coatings on SS316 are shown in Figure 7a and 7b, respectively. Each point of the boiling curves

13 indicates one measured point with the gradual increase in heat flux. Based on the boiling curves,
D

14 the heat fluxes were divided by the difference between surface temperature of heater and
TE

15 saturated temperature of adjacent working fluid (wall superheat), and the calculated results

16
EP

appeared as the HTCs, in Figure 7b. All LbL-assembled PEI-MWCNTs clearly improved the

17 heat flux and the HTCs compared to the bare SS316. Among the different bi-layers of the LbL
C

18 coatings, the 20 bi-layers PEI-MWCNTs coatings showed the best performance, whereas the 40
AC

19 bi-layers PEI-MWCNTs coatings showed the lowest HTCs. This indicates that there is an

20 optimized number of LbL assembled PEI-MWCNTs layers for further enhancement of HTCs.

21 The enhanced HTCs are mainly attributed to the increase in the nucleate site density from the

22 nanoporous surfaces as well as the improved surface roughness ranged from 47 % to 91 %. The

23 great number of nanopores formed via the randomly oriented networks of MWCNTs acted as the

18
 ACCEPTED MANUSCRIPT

1 origin of the nucleation sites for initial formation of vapor bubbles. Moreover, it is believed that

2 a higher value of surface roughness is strongly related to the presence of more micro-cavities.[63,

3 64] Jones et al. indicated that lower wall superheats and enhanced thermal dissipation can be

PT
4 achieved with larger unwetted cavities on a rougher surface.[64] Accordingly, the most improved

5 HTC was observed for the 20 bi-layers PEI-MWCNTs coating with the highest roughness value

RI
6 as discussed in previous section.

SC
7

8 However, although the MWCNTs possess high thermal conductivity, the excessively thick

U
9 MWCNTs coatings can decrease the overall HTC in nucleate boiling heat transfer because of the
AN
10 empty spaces in the porous structures.[56] A porous structure generally has a lower thermal

11 conductivity than a dense structure. Additionally, the thermal resistance at the contact interface
M

12 of SS316 and the MWCNTs coating can interrupt the heat transfer from SS316 to the coating

13 layers. Therefore, a highly enhanced HTC might be obtained in the 20 bi-layers PEI-MWCNTs,
D

14 which has the ideal nanoporous surface, a relatively solid contact at the contact interface via
TE

15 electrostatic bonding, as well as a moderate thickness of the MWCNTs coating. In the case of the

16
EP

10 bi-layers and 40 bi-layers PEI-MWCNTs, the incomplete nanoporous surfaces and the high

17 thermal resistance from the relatively thick MWCNTs coating could interrupt the thermal
C

18 transport from the substrates.

AC

19

20 The CHFs at specific wall temperatures can be also estimated from Figure 7a and 7b. The results

21 are summarized in Figure 8. In the case of CHF, the overall tendency is significantly different

22 from that of HTC. First, the enhancement in CHFs for heaters using LbL-assembled PEI-

23 MWCNTs was still observed in all specimens. However, the thicker film showed a greater

19
 ACCEPTED MANUSCRIPT

1 enhancement in CHF. The 40 bi-layers PEI-MWCNTs coatings improved the CHF of SS316

2 greatly. Moreover, the vacuum-filtered MWCNTs coating with the thickest film revealed a

3 moderate enhancement of CHF between 10 bi-layer and 20 bi-layer LbL coatings. For the

PT
4 quantitative analysis, the CHF of bare SS316, 912 kW/m2, was used as a reference value. Using

5 MWCNTs coatings, the CHF of vacuum-filtered MWCNTs-coated SS316 substrate was 1534

RI
6 kW/m2, while LbL-assembled PEI-MWNCTs-coated SS316 substrates showed significantly

SC
7 enhanced CHFs, 1480, 1629, and 2253 kW/m2 for 10, 20, and 40 bi-layers, respectively. The best

8 enhancement of about 147%, compared to a reference value obtained from bare SS316 substrates,

U
9 was achieved for the 40 bi-layers LbL coatings.
AN
10

11 In general, CHF depends on the surface roughness,[65, 66] the capillary wicking,[67, 68] and the
M

12 wettability of contact surfaces with a working fluid.[69, 70] Moreover, it is conventionally

13 believed that CHF enhancement is expected on a wettable, rougher, and/or porous surface.
D

14 Rewetting to hot dry spots on a heating surface can be facilitated on a hydrophilic surface. Also,
TE

15 Kandlikar’s study indicated that CHF increases monotonically with reducing contact angles.[69]

16
EP

Since the surface roughness affects not only HTC but also CHF, a rougher surface could attribute

17 to enhancement of CHF. In Chu et al.’s analysis, CHF gradually increased with a higher surface
C

18 roughness at the micrometer scale.[65] Furthermore, the distribution of cavities on surfaces

AC

19 affects the bubble formation characteristics such as density and size. Since the occurrence of

20 CHF is significantly relevant to the formation of a vapor cloud near the heating surface, for

21 further enhancement of CHF, the bubble should be atomized to prevent combination among the

22 bubbles on surfaces. Capillary wicking through porous structures restrains the formation of vapor

23 films. Moreover, this kind of capillary phenomena can be accelerated by the appropriate

20
 ACCEPTED MANUSCRIPT

1 hydrophilic nature of the networks in porous structures. Because these mechanisms induce the

2 suction of liquid as well as the liquid-vapor circulation through the networks, the bubbles easily

3 leave the heat transfer substrates and delay the vapor cloud formation. In this regard, the thicker

PT
4 porous structure in LbL-assembled PEI-MWCNTs coatings showed the enhanced CHF. All the

5 samples of LbL coatings on SS316 exhibited the wettable and rougher surface as compared to the

RI
6 bare sample, as shown in Figure 5.

SC
7

8 In addition, the effectiveness of electrostatic bonding and the wettability can be discussed

U
9 compared to vacuum-filtered MWCNTs. Even though the vacuum-filtered MWCNTs had the
AN
10 highest thickness among porous structures, the LbL-assembled PEI-MWCNTs coatings with 20

11 and 40 bi-layers showed higher CHFs than that of the vacuum-filtered MWCNTs coating. The
M

12 vacuum-filtered MWCNTs coating might be used to increase the nanoporous structures.

13 However, the capillary wicking might not be highly enhanced because of a relatively lower value
D

14 of the surface roughness and the hydrophobic nature of vacuum-filtered MWCNTs. As shown in
TE

15 Figure 3 and Figure 5, the LbL-assembled PEI-MWCNTs coatings on SS316 were rougher and

16
EP

hydrophilic. This characteristic could help improve rewetting to hot dry spots and capillary effect

17 through the nanoporous structures, and it might result in the further enhancement of CHFs.
C

18 Moreover, the electrostatic force could help to construct denser MWCNTs structures. It could
AC

19 prevent the formation of large sized bubbles before the departure from the substrates. If the LbL-

20 assembled coatings on the heat transfer substrates are thicker than a 40 bi-layers coating, the

21 CHF may be enhanced more, but there may be an adverse effect of decrease in HTCs because of

22 the additional thermal resistance. With the enhancement of HTC and CHF, the repeatability of

23 performances of LbL coatings is quite significant for technological applications. In order to

21
 ACCEPTED MANUSCRIPT

1 evaluate the repeatability of performances, HTCs of one LbL assembled PEI-MWCNTs coating

2 in 10 bi-layers on SS 316 were repeatedly measured, four times (Figure S3). The HTCs of all

3 tests were in the range of 20.16 and 20.57 kW/m2K, and quite consistent. This result indicates

PT
4 that LbL coatings have enough stability and repeatability in pool boiling heat transfer.

RI
6 The observation of generation and departure of vapor bubbles provides a detailed understanding

SC
7 of the physical mechanisms of how the LbL-assembled PEI-MWCNTs coatings can enhance and

8 manipulate both HTC and CHF (Figure 9). The high-speed images of the vapor bubbles

U
9 generated on the heat transfer substrates during pool boiling experiments were captured at the
AN
10 corresponding times for the bare SS316 and the LbL-assembled PEI-MWCNTs coatings on

11 SS316. Note that the high-speed images in Figure 9 are presented for two kinds of test samples:
M

12 the bare SS316 and LbL-assembled PEI-MWCNTs coatings with 40 bi-layers. As compared to

13 the case of the bare heater, most distinguished bubble behaviors were observed for the test
D

14 heaters of PEI-MWCNTs with 40 bi-layers. Moreover, the bubble behaviors among the test
TE

15 heaters of PEI-MWCNTs coatings were almost identical with the present visualization in Figure

16
EP

9.

17
C

18 In the low heat flux regime (Figure 9a), the bubble generation gradually increased on the
AC

19 substrates in SS316 substrates both with and without LbL-assembled PEI-MWCNTs coatings. In

20 terms of the size and shape, the substrate with LbL-assembled coatings created more atomized

21 bubbles than the bare SS316 because of the nanoporous surface. As the heat flux increased, the

22 bubbles were crowded on the substrates and the bubble density was amplified. Some isolated

23 bubbles began to merge with the others and, as a result, a vapor column was formed. When the

22
 ACCEPTED MANUSCRIPT

1 heat flux reached a specific value for each case, a new vapor column on the heating surface

2 combined with a previous one (Figure 9b). Moreover, some bubbles or vapor columns left the

3 surfaces, whereas some of them held together and formed the vapor clouds near the surface.

PT
4 However, at the similar heat flux, distinctive bubble behaviors were observed for two types of

5 samples. Although merged vapor columns or clouds were actively generated, the heating surface

RI
6 of LbL coatings on SS316 was exposed to the surrounding liquid more than that of the bare

SC
7 heater. In other words, improved rewetting on the heating surface substantially occurred at the

8 interface with the surrounding liquid. This observation implies that formation of hot dry spots

U
9 could be suppressed for the case of LbL coatings. Heat flux exceeding the critical point caused a
AN
10 huge vapor cloud that covered most areas of the substrates and induced a sudden excursion in the

11 wall temperature. The results confirm that LbL-assembled PEI-MWCNTs coatings can atomize
M

12 bubbles and improve capillary wicking. Improvements resulting from the hydrophilic

13 nanoporous structures lead to liquid-vapor circulation in the networks of MWCNTs. Furthermore,

D

14 as many studies on boiling heat transfer reported, the improved CHF is expected with the
TE

15 delayed formation of hot dry spots.[71-73] Consequently, the visual observations provide an

16
EP

experimental evidence that supports the CHF enhancement for the case of LbL coatings on

17 SS316.
C

18
AC

19 4. Conclusions

20 In summary, we reported the remarkable enhancement of HTC and CHF in pool boiling heat

21 transfer using LbL-assembled PEI-MWCNTs coatings on SS316 substrates. The bare SS316 and

22 vacuum-filtered MWCNTs on SS316 were prepared as the controls, and LbL-assembled PEI-

23 MWCNTs coatings with 10, 20, and 40 bi-layers were deposited on the SS316 substrates.

23
 ACCEPTED MANUSCRIPT

1 Because of the nanosize cavities on the surfaces and nanoporous structures that were fabricated

2 by the electrostatic force, all LbL coatings showed a significant enhancement in the HTCs and

3 CHFs compared to bare SS316. Furthermore, denser networks of MWCNTs due to electrostatic

PT
4 bonding and the hydrophilic nature of PEI in LbL coatings could effectively control the HTCs

5 and CHFs. LbL-assembled PEI-MWCNTs with 20 bi-layers show the highest improvement in

RI
6 HTC, whereas 40 bi-layers LbL coatings achieved the best enhancement ratio of CHF, up to

SC
7 147%, from 912 kW/m2 to 2253 kW/m2. The delay in vapor cloud formation via enhanced

8 capillary wicking and the atomization of bubbles via nanoporous surfaces in the LbL-assembled

U
9 PEI-MWCNTs could significantly contribute to the liquid-vapor circulation in both the surface
AN
10 and inner porous structures. Further development of the LbL coatings on SS316 would enable

11 potential applications for thermal management at multiple scales, from micro/nanoscale

M

12 platforms to macroscale systems. Understanding LbL coatings on metal surfaces and the

13 corresponding heat transfer characteristics would provide valuable information for further design
D

14 of textured surfaces using diverse micro/nanomaterials for pool boiling heat transfer.
TE

15
EP

16 Acknowledgement

17
C

This work was supported by the Basic Science Research Program through the National Research
AC

18 Foundation of Korea (NRF), funded by the Ministry of Education, Science and Technology

19 (NRF-2013R1A1A1010575), by the Ministry of Education (NRF-2015R1D1A1A01059274),

20 and by the National Research Foundation of Korea (NRF) grant funded by the Korean

21 government (MSIP: Ministry of Science, ICT and Future Planning) (No. NRF-

22 2014M2B2A9032081).

24
 ACCEPTED MANUSCRIPT

1 Figures and Captions

PT
RI
SC
3

U
AN
4 Figure 1. Scheme for fabrication of layer-by-layer (LbL-assembled polyethylenimine (PEI)-

5 multi-walled carbon nanotubes (MWCNTs) coatings on stainless steel (SS316) substrates and its
M

6 application for manipulation of pool boiling heat transfer.

D
TE
C EP
AC

25
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP

1
C

2 Figure 2. LbL-assembled PEI-MWCNTs coatings with different bi-layers and vacuum-filtered

AC

3 MWCNTs coating on SS316 substrates. Photographs and scanning electron microscope (SEM)

4 images on surfaces of (a) bare SS316 substrate after polishing, (b) vacuum-filtered MWCNTs

5 coating on SS316 substrate, LbL-assembled PEI-MWCNTs coatings with (c) 10 bi-layers, (d) 20

6 bi-layers, and (e) 40 bi-layers on SS316 substrate.

26
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M

1
D

2 Figure 3. Surface characteristics of heat transfer substrates. Atomic force microscope (AFM)
TE

3 images of (a) bare SS316, (b) vacuum-filtered MWCNTs coating, LbL-assembled PEI-MWCNTs

4 coatings with (c) 10 bi-layers, (d) 20 bi-layers, and (e) 40 bi-layers on SS316.
EP

5
C

6
AC

27
 ACCEPTED MANUSCRIPT

PT
4

RI
5

SC
6

U
AN
8

9 Figure 4. Coating thicknesses of heat transfer substrates, vacuum-filtered MWCNTs coating,

M

10 LbL-assembled PEI-MWCNTs coatings with 10 bi-layers, 20 bi-layers, and 40 bi-layers on

D

11 SS316.
TE

12

13
EP

14
C
AC

28
 ACCEPTED MANUSCRIPT

PT
RI
U SC
1
AN
2 Figure 5. Wettabilities of heat transfer substrates. Static contact angles of (a) bare SS316, (b)
M

3 vacuum-filtered MWCNTs coating, LbL-assembled PEI-MWCNTs coatings with (c) 10 bi-layers,

4 (d) 20 bi-layers, and (e) 40 bi-layers on SS316.

D

5
TE

6
EP

8
C
AC

29
 ACCEPTED MANUSCRIPT

PT
4

RI
6

SC
8

10
U
AN
11

12
M

13
D

14

15
TE

16

17
EP

18
C

19
AC

20

21

22 Figure 6. Experimental setup for pool boiling heat transfer characteristics of test heaters. (a)

23 Scheme of test heater assembly. The central heating part indicates the fabricated heater samples.

24 (b) Scheme of experimental setup for measurement of HTC and CHF.

30
 ACCEPTED MANUSCRIPT

PT
RI
1

SC
2 Figure 7. Heat transfer characteristics of heat transfer substrates. (a) Boiling curves and (b)

U
3 corresponding HTCs of bare SS316, vacuum-filtered MWCNTs coating, LbL-assembled PEI-
AN
4 MWCNTs coatings of 10 bi-layers, 20 bi-layers, and 40 bi-layers on SS316.

5
M

6
D

7
TE

8
EP

10
C
AC

31
 ACCEPTED MANUSCRIPT

PT
4

RI
5

SC
6

U
AN
8

9 Figure 8. CHFs of bare SS316, vacuum-filtered MWCNTs coating, LbL-assembled PEI-

M

10 MWCNTs coatings with 10 bi-layers, 20 bi-layers, and 40 bi-layers on SS316.

D

11
TE

12
EP

13
C
AC

32
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE

2 Figure 9. Representative high-speed images of real-time bubble generation and departure in the
EP

3 boiling process for bare SS316 and LbL-assembled PEI-MWCNTs coatings on SS316. (a) Low-

4 to-moderate heat flux regimes, and (b) high heat flux regimes.
C
AC

33
 ACCEPTED MANUSCRIPT

1 References

2 [1] Lin S, Broadbent J, McGlen R. Numerical study of heat pipe application in heat
3 recovery systems. Appl Therm Eng. 2005;25(1):127-33.
4 [2] Jiang LA, Wong M, Zohar Y. Phase change in microchannel heat sinks with integrated

PT
5 temperature sensors. J Microelectromech S. 1999;8(4):358-65.
6 [3] Chen JC. Correlation for Boiling Heat Transfer to Saturated Fluids in Convective Flow.
7 Ind Eng Chem Proc Dd. 1966;5(3):322-&.

RI
8 [4] Wang CC, Chang CT. Heat and mass transfer for plate fin-and-tube heat exchangers,
9 with and without hydrophilic coating. Int J Heat Mass Tran. 1998;41(20):3109-20.
10 [5] Hwang GS, Kaviany M. Critical heat flux in thin, uniform particle coatings. Int J Heat

SC
11 Mass Tran. 2006;49(5-6):844-9.
12 [6] Betz AR, Xu J, Qiu HH, Attinger D. Do surfaces with mixed hydrophilic and
13 hydrophobic areas enhance pool boiling? Appl Phys Lett. 2010;97(14).
14 [7] Haramura Y, Katto Y. A New Hydrodynamic Model of Critical Heat-Flux, Applicable

U
15 Widely to Both Pool and Forced-Convection Boiling on Submerged Bodies in Saturated Liquids.
16 Int J Heat Mass Tran. 1983;26(3):389-99.
AN
17 [8] Chang JY, You SM. Boiling heat transfer phenomena from microporous and porous
18 surfaces in saturated FC-72. Int J Heat Mass Tran. 1997;40(18):4437-47.
19 [9] Bang IC, Chang SH. Boiling heat transfer performance and phenomena of Al2O3-water
20 nano-fluids from a plain surface in a pool. Int J Heat Mass Tran. 2005;48(12):2407-19.
M

21 [10] Phan HT, Caney N, Marty P, Colasson S, Gavillet J. Surface wettability control by
22 nanocoating: The effects on pool boiling heat transfer and nucleation mechanism. Int J Heat
23 Mass Tran. 2009;52(23-24):5459-71.
D

24 [11] Ahn HS, Jo HJ, Kang SH, Kim MH. Effect of liquid spreading due to
25 nano/microstructures on the critical heat flux during pool boiling. Appl Phys Lett. 2011;98(7).
TE

26 [12] Liter SG, Kaviany M. Pool-boiling CHF enhancement by modulated porous-layer

27 coating: theory and experiment. Int J Heat Mass Tran. 2001;44(22):4287-311.
28 [13] Cieslinski JT. Nucleate pool boiling on porous metallic coatings. Experimental Thermal
29 and Fluid Science. 2002;25(7):557-64.
EP

30 [14] Coursey JS, Kim J. Nanofluid boiling: The effect of surface wettability. Int J Heat Fluid
31 Fl. 2008;29(6):1577-85.
32 [15] Truong B, Hu LW, Buongiorno J, McKrell T. Modification of sandblasted plate heaters
C

33 using nanofluids to enhance pool boiling critical heat flux. Int J Heat Mass Tran. 2010;53(1-
34 3):85-94.
AC

35 [16] Yan W, Lin-lin W, Ming-yan L. Antifouling and enhancing pool boiling by TiO2 coating
36 surface in nanometer scale thickness. Aiche J. 2007;53(12):3062-76.
37 [17] Hendricks TJ, Krishnan S, Choi CH, Chang CH, Paul B. Enhancement of pool-boiling
38 heat transfer using nanostructured surfaces on aluminum and copper. Int J Heat Mass Tran.
39 2010;53(15-16):3357-65.
40 [18] Lu MC, Chen RK, Srinivasan V, Carey V, Majumdar A. Critical heat flux of pool boiling
41 on Si nanowire array-coated surfaces. Int J Heat Mass Tran. 2011;54(25-26):5359-67.
42 [19] Geng HZ, Kim KK, So KP, Lee YS, Chang Y, Lee YH. Effect of acid treatment on
43 carbon nanotube-based flexible transparent conducting films. Journal of the American Chemical
44 Society. 2007;129(25):7758-+.

34
 ACCEPTED MANUSCRIPT

1 [20] Kwak HS, Kim K, Shon BC, Lee H, Han CS. Thermal Characteristics of a Transparent
2 Film Heater Using Single-Walled Carbon Nanotubes. J Nanosci Nanotechno. 2010;10(5):3512-5.
3 [21] Ujereh S, Fisher T, Mudawar I. Effects of carbon nanotube arrays on nucleate pool
4 boiling. Int J Heat Mass Tran. 2007;50(19-20):4023-38.
5 [22] Dai XM, Huang XY, Yang FH, Li XD, Sightler J, Yang YC, et al. Enhanced nucleate
6 boiling on horizontal hydrophobic-hydrophilic carbon nanotube coatings. Appl Phys Lett.

PT
7 2013;102(16).
8 [23] Khanikar V, Mudawar I, Fisher T. Effects of carbon nanotube coating on flow boiling in
9 a micro-channel. Int J Heat Mass Tran. 2009;52(15-16):3805-17.

RI
10 [24] Singh N, Sathyamurthy V, Peterson W, Arendt J, Banerjee D. Flow boiling enhancement
11 on a horizontal heater using carbon nanotube coatings. Int J Heat Fluid Fl. 2010;31(2):201-7.
12 [25] Jun S, Sinha-Ray S, Yarin AL. Pool boiling on nano-textured surfaces. Int J Heat Mass

SC
13 Tran. 2013;62:99-111.
14 [26] Tang ZH, Qiu CQ, McCutcheon JR, Yoon K, Ma HY, Fang DF, et al. Design and
15 Fabrication of Electrospun Polyethersulfone Nanofibrous Scaffold for High-Flux Nanofiltration
16 Membranes. J Polym Sci Pol Phys. 2009;47(22):2288-300.

U
17 [27] Kim H, Kim M. Experimental study of the characteristics and mechanism of pool
18 boiling CHF enhancement using nanofluids. Heat Mass Transfer. 2009;45(7):991-8.
AN
19 [28] Lu YW, Kandlikar SG. Nanoscale Surface Modification Techniques for Pool Boiling
20 EnhancementA Critical Review and Future Directions. Heat Transfer Eng. 2011;32(10):827-42.
21 [29] Jiang SP, Liu ZC, Tian ZQ. Layer-by-layer self-assembly of composite polyelectrolyte-
22 nafion membranes for direct methanol fuel cells. Adv Mater. 2006;18(8):1068-+.
M

23 [30] Dubas ST, Kumlangdudsana P, Potiyaraj P. Layer-by-layer deposition of antimicrobial

24 silver nanoparticles on textile fibers. Colloid Surface A. 2006;289(1-3):105-9.
25 [31] Schmitt J, Decher G, Dressick WJ, Brandow SL, Geer RE, Shashidhar R, et al. Metal
D

26 nanoparticle/polymer superlattice films: Fabrication and control of layer structure. Adv Mater.
27 1997;9(1):61-&.
TE

28 [32] Lvov Y, Decher G, Sukhorukov G. Assembly of Thin-Films by Means of Successive

29 Deposition of Alternate Layers of DNA and Poly(Allylamine). Macromolecules.
30 1993;26(20):5396-9.
31
EP

[33] Decher G, Hong JD, Schmitt J. Buildup of Ultrathin Multilayer Films by a Self-
32 Assembly Process .3. Consecutively Alternating Adsorption of Anionic and Cationic
33 Polyelectrolytes on Charged Surfaces. Thin Solid Films. 1992;210(1-2):831-5.
34 [34] Kong W, Zhang X, Gao ML, Zhou H, Li W, Shen JC. A New Kind of Immobilized
C

35 Enzyme Multilayer Based on Cationic and Anionic Interaction. Macromol Rapid Comm.
36 1994;15(5):405-9.
AC

37 [35] Keller SW, Kim HN, Mallouk TE. Layer-by-Layer Assembly of Intercalation
38 Compounds and Heterostructures on Surfaces - toward Molecular Beaker Epitaxy. Journal of the
39 American Chemical Society. 1994;116(19):8817-8.
40 [36] Kim SY, Hong J, Kavian R, Lee SW, Hyder MN, Shao-Horn Y, et al. Rapid fabrication
41 of thick spray-layer-by-layer carbon nanotube electrodes for high power and energy devices.
42 Energ Environ Sci. 2013;6(3):888-97.
43 [37] Peng CQ, Thio YS, Gerhardt RA, Ambaye H, Lauter V. pH-Promoted Exponential
44 Layer-by-Layer Assembly of Bicomponent Polyelectrolyte/Nanoparticle Multilayers. Chem
45 Mater. 2011;23(20):4548-56.
46 [38] Iler RK. Multilayers of Colloidal Particles. J Colloid Interf Sci. 1966;21(6):569-&.

35
 ACCEPTED MANUSCRIPT

1 [39] Decher G. Fuzzy nanoassemblies: Toward layered polymeric multicomposites. Science.

2 1997;277(5330):1232-7.
3 [40] Schaaf P, Voegel JC, Jierry L, Boulmedais F. Spray-Assisted Polyelectrolyte Multilayer
4 Buildup: from Step-by-Step to Single-Step Polyelectrolyte Film Constructions. Adv Mater.
5 2012;24(8):1001-16.
6 [41] Thomas IM. Single-Layer Tio2 and Multilayer Tio2-Sio2 Optical Coatings Prepared

PT
7 from Colloidal Suspensions. Appl Optics. 1987;26(21):4688-91.
8 [42] Schlenoff JB, Dubas ST, Farhat T. Sprayed Polyelectrolyte Multilayers. Langmuir.
9 2000;16(26):9968-9.

RI
10 [43] Picart C, Lavalle P, Hubert P, Cuisinier FJG, Decher G, Schaaf P, et al. Buildup
11 mechanism for poly(L-lysine)/hyaluronic acid films onto a solid surface. Langmuir.
12 2001;17(23):7414-24.

SC
13 [44] Lee SW, Kim BS, Chen S, Shao-Horn Y, Hammond PT. Layer-by-layer assembly of all
14 carbon nanotube ultrathin films for electrochemical applications. Journal of the American
15 Chemical Society. 2009;131(2):671-9.
16 [45] Che JW, Cagin T, Goddard WA. Thermal conductivity of carbon nanotubes.

U
17 Nanotechnology. 2000;11(2):65-9.
18 [46] Berber S, Kwon YK, Tomanek D. Unusually high thermal conductivity of carbon
AN
19 nanotubes. Phys Rev Lett. 2000;84(20):4613-6.
20 [47] Wong EW, Sheehan PE, Lieber CM. Nanobeam mechanics: Elasticity, strength, and
21 toughness of nanorods and nanotubes. Science. 1997;277(5334):1971-5.
22 [48] Yu HH, Cao T, Zhou LD, Gu ED, Yu DS, Jiang DS. Layer-by-layer assembly and
M

23 humidity sensitive behavior of poly(ethyleneimine)/multiwall carbon nanotube composite films.

24 Sensor Actuat B-Chem. 2006;119(2):512-5.
25 [49] Yuan SJ, Xu FJ, Pehkonen SO, Ting YP, Neoh KG, Kang ET. Grafting of antibacterial
D

26 polymers on stainless steel via surface-initiated atom transfer radical polymerization for
27 inhibiting biocorrosion by Desulfovibrio desulfuricans. Biotechnology and bioengineering.
TE

28 2009;103(2):268-81.
29 [50] Zhao P, Einarsson E, Lagoudas G, Shiomi J, Chiashi S, Maruyama S. Tunable separation
30 of single-walled carbon nanotubes by dual-surfactant density gradient ultracentrifugation. Nano
31
EP

Res. 2011;4(7):623-34.
32 [51] Choi W, Hong J. Rapid Electromechanical Transduction on a Single-Walled Carbon
33 Nanotube Film: Sensing Fast Mechanical Loading via Detection of Electrical Signal Change. Ind
34 Eng Chem Res. 2012;51(45):14714-21.
C

35 [52] Yang SB, Kong BS, Jung DH, Baek YK, Han CS, Oh SK, et al. Recent advances in
36 hybrids of carbon nanotube network films and nanomaterials for their potential applications as
AC

37 transparent conducting films. Nanoscale. 2011;3(4):1361-73.

38 [53] Sun JJ, Zhao HZ, Yang QZ, Song J, Xue A. A novel layer-by-layer self-assembled
39 carbon nanotube-based anode: Preparation, characterization, and application in microbial fuel
40 cell. Electrochim Acta. 2010;55(9):3041-7.
41 [54] Lvov Y, Ariga K, Ichinose I, Kunitake T. Assembly of Multicomponent Protein Films by
42 Means of Electrostatic Layer-by-Layer Adsorption. Journal of the American Chemical Society.
43 1995;117(22):6117-23.
44 [55] Hoogeveen NG, Stuart MAC, Fleer GJ, Bohmer MR. Formation and stability of
45 multilayers of polyelectrolytes. Langmuir. 1996;12(15):3675-81.
46 [56] Seo GH, Hwang H, Yoon J, Yeo T, Son HH, Jeong U, et al. Enhanced critical heat flux

36
 ACCEPTED MANUSCRIPT

1 with single-walled carbon nanotubes bonded on metal surfaces. Experimental Thermal and Fluid
2 Science. 2015;60:138-47.
3 [57] Harrison BS, Atala A. Carbon nanotube applications for tissue engineering. Biomaterials.
4 2007;28(2):344-53.
5 [58] Sethi S, Dhinojwala A. Superhydrophobic Conductive Carbon Nanotube Coatings for
6 Steel. Langmuir. 2009;25(8):4311-3.

PT
7 [59] Weibel JA, Kim SS, Fisher TS, Garimella SV. Carbon Nanotube Coatings for Enhanced
8 Capillary-Fed Boiling from Porous Microstructures. Nanosc Microsc Therm. 2012;16(1):1-17.
9 [60] Rao PS, Smitha B, Sridhar S, Krishnaiah A. Preparation and performance of poly(vinyl

RI
10 alcohol)/polyethyleneimine blend membranes for the dehydration of 1,4-dioxane by
11 pervaporation: Comparison with glutaraldehyde cross-linked membranes. Sep Purif Technol.
12 2006;48(3):244-54.

SC
13 [61] Gruener S, Hofmann T, Wallacher D, Kityk AV, Huber P. Capillary rise of water in
14 hydrophilic nanopores. Phys Rev E. 2009;79(6).
15 [62] Witharana S, Phillips B, Strobel S, Kim HD, McKrell T, Chang JB, et al. Bubble
16 nucleation on nano- to micro-size cavities and posts: An experimental validation of classical

U
17 theory. J Appl Phys. 2012;112(6).
18 [63] Sadasivan P, Unal C, Nelson R. Perspective - Issues in Chf Modeling - the Need for New
AN
19 Experiments. J Heat Trans-T Asme. 1995;117(3):558-67.
20 [64] Jones BJ, McHale JP, Garimella SV. The Influence of Surface Roughness on Nucleate
21 Pool Boiling Heat Transfer. J Heat Trans-T Asme. 2009;131(12).
22 [65] Biswas K, He JQ, Blum ID, Chun-Iwu, Hogan TP, Seidman DN, et al. High-
M

23 performance bulk thermoelectrics with all-scale hierarchical architectures (vol 489, pg 414,
24 2012). Nature. 2012;490(7421).
25 [66] Chu KH, Joung YS, Enright R, Buie CR, Wang EN. Hierarchically structured surfaces
D

26 for boiling critical heat flux enhancement. Appl Phys Lett. 2013;102(15).
27 [67] Rahman MM, Olceroglu E, McCarthy M. Role of Wickability on the Critical Heat Flux
TE

28 of Structured Superhydrophilic Surfaces. Langmuir. 2014;30(37):11225-34.

29 [68] Kim BS, Lee H, Shin S, Choi G, Cho HH. Interfacial wicking dynamics and its impact
30 on critical heat flux of boiling heat transfer. Appl Phys Lett. 2014;105(19).
31
EP

[69] Kandlikar SG. A theoretical model to predict pool boiling CHF incorporating effects of
32 contact angle and orientation. J Heat Trans-T Asme. 2001;123(6):1071-9.
33 [70] O'Hanley H, Coyle C, Buongiorno J, McKrell T, Hu LW, Rubner M, et al. Separate
34 effects of surface roughness, wettability, and porosity on the boiling critical heat flux. Appl Phys
C

35 Lett. 2013;103(2).
36 [71] Ahn HS, Kim JM, Kim T, Park SC, Kim JM, Park Y, et al. Enhanced heat transfer is
AC

37 dependent on thickness of graphene films: the heat dissipation during boiling. Sci Rep-Uk.
38 2014;4.
39 [72] Dhillon NS, Buongiorno J, Varanasi KK. Critical heat flux maxima during boiling crisis
40 on textured surfaces. Nat Commun. 2015;6.
41 [73] Seo GH, Jeun G, Kim SJ. Pool boiling heat transfer characteristics of zircaloy and SiC
42 claddings in deionized water at low pressure. Exp Therm Fluid Sci. 2015;64:42-53.
43

37