International Journal of Hydrogen Energy 51 (2024) 1130–1141

Contents lists available at ScienceDirect

International Journal of Hydrogen Energy

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

Effects of the wall temperature on the boiling process and the molecular
dynamics behavior of the liquid hydrogen on a flat aluminum wall
Xiaodan Liu a, Yusong Yu a, *, Chaoran Hou a, Jiawei Ding b
a
Hydrogen Energy and Space Propulsion Laboratory (HESPL), School of Mechanical, Electronic and Control Engineering, Beijing Jiaotong University, Beijing, 100044,
China
b
Xi’an Aerospace Propulsion Test Technique Institute, Xi’an, 710100, China

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

Handling Editor: Ibrahim Dincer Understanding the nucleation mechanism and boiling heat transfer characteristics of liquid hydrogen is a key
issue in the development of liquid hydrogen storage systems. However, the current methods of experimental and
mesoscopic or macroscopic scale continuous flow simulations suffer from the problems of large study scales and
undetermined initial boundary conditions. The purpose of this study is to reveal the bubble nucleation mecha­
nism and bubble development process of liquid hydrogen on the surface of heating plate at the molecular scale,
and to investigate the effect of different heating plate temperature on bubble growth, during thin film boiling. In
this research, molecular dynamics simulation is used to combine molecular position distribution with energy
analysis and explain the molecular behavior. The effect of heating temperature on different phase transition
stages of liquid hydrogen is also investigated by varying the heating surface temperature. In this study, the
bubble nucleation, bubble development and merging of liquid hydrogen on the heating surface were successfully
observed. It is shown that increasing the heating temperature has a small effect on the bubble growth rate and
liquid layer thickness in the nucleation boiling stage, and mainly affects the gas film growth rate in the film
boiling stage. The bubbles are extremely unstable at the initial stage of generation. Within 1 fs of the first
appearance of bubbles, which are associated with random micro-pits caused by thermal motion of solid atoms,
bubbles move on the surface of the heating plate. This study found a phenomenon that bubbles firstly generated
and then disappeared. This paper provides a new idea for the nucleation mechanism and boiling heat transfer
study of liquid hydrogen. Moreover, it provides a solution to the problems of liquid hydrogen phase change
research.

1. Introduction At room temperature and atmospheric pressure, liquid hydrogen has

a density of about 800 times that of gaseous hydrogen. It also has high
The increasing prominence of global environmental issues has driven gravimetric and volumetric efficiencies. Therefore, liquid hydrogen is
countries to promote the transformation of their energy structures to currently the main form of large-scale transport and storage of energy.
clean fuels. With the advancement of science and technology, the clean As a cryogenic fluid, liquid hydrogen boils at about 20.3 K under at­
energy industry, represented by the hydrogen energy industry, has been mospheric pressure. Since the storage and transportation processes of
booming. As a clean and efficient renewable energy source, hydrogen liquid hydrogen are particularly sensitive to heat leakage [12–15],
energy has been attracting more and more attention in the world. The bubble nucleation and boiling heat exchange phenomena of liquid
application fields of hydrogen are extremely wide. Hydrogen can be hydrogen are particularly likely to be encountered.
used as a power source, such as cathode fuel for fuel cell vehicles [1–6], Driven by the demand for high-temperature superconductivity and
and fuel for aerospace propulsion systems [7–10]. Hydrogen can also be aerospace exploration, the problems related to the boiling and heat
used as a coolant for various applications, such as cold neutron transfer of liquid hydrogen have received continuous attention. Many
moderator materials for large superconducting magnets and scattered experimental studies of liquid hydrogen boiling and heat transfer pro­
neutron sources [11]. cesses have been reported in the literatures. Class et al. [16] gave

* Corresponding author.
E-mail address: ysyu@bjtu.edu.cn (Y. Yu).

https://doi.org/10.1016/j.ijhydene.2023.11.073
Received 27 March 2023; Received in revised form 29 October 2023; Accepted 5 November 2023
Available online 15 November 2023
0360-3199/© 2023 The Authors. Published by Elsevier Ltd on behalf of Hydrogen Energy Publications LLC. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
X. Liu et al. International Journal of Hydrogen Energy 51 (2024) 1130–1141

experimental data on the boiling heat exchange of liquid hydrogen over wettability and superheat of the substrate. The bubble nucleation pro­
a large plane under different conditions. Furthermore, they analyzed cess of liquid argon at different temperatures and wettability surfaces
previous experimental data on hydrogen in the literatures. The results was investigated by Zhou et al. [27]. It was found that at lower wall
showed that these heat transfer data are not very consistent. This may temperatures, bubbles tended to nucleate first in the hydrophobic part of
indicate that obtaining boiling curves for liquid hydrogen is challenging, the wall. However, as the wall temperature increases, the position of
due to the lack of reliable experimental data. Seader et al. [17]con­ bubble nucleation gradually moves from the hydrophobic region to the
ducted an extensive study on the heat transfer of liquid hydrogen and hydrophilic region. Chen et al. [28]used a molecular dynamics approach
reviewed the existing theoretical methods and experimental data for to analyze the nucleation position of liquid argon. They found that
predicting boiling phenomena. Shirai et al. [18] used the experimental bubble nucleation first forms in the groove region. This conclusion is
methods to investigate the heat transfer from a flat plate facing upward consistent with the classical non-homogeneous nucleation theory. In
immersed in liquid hydrogen pool. They obtained the pressure corre­ addition to studying monatomic molecule workings, such as argon,
sponding to the maximum critical heat flux. Shiotsu et al. [19] investi­ molecular dynamics simulation (MDS) method is also a powerful tool to
gated the transient heat transfer characteristics of boiling in a liquid study the phase transition processes of polyatomic molecules such as
hydrogen pool under saturated and subcooled conditions. They analyzed water at the nanoscale [29–32].
a flat plate heater with a 5 mm in width, 60 mm in length and 0.5 mm in Some scholars have used molecular dynamics methods to study
thickness. The results show that the, transient CHF was higher for higher hydrogen. Jiang et al. [33] studied the storage of molecular hydrogen in
degree of subsoiling condition. The application of liquid hydrogen a novel 3D carbon structure – pillared graphene bubble system under
usually requires an adequate pre-cooling operation. Moreover, different various environments using molecular dynamics (MD) method。Fang
heat transfer mechanisms, including film boiling, transition boiling and et al. [34] studied hydrogen permeation behavior in liner polymer ma­
nucleate boiling, successively dominate the cooling process. Available terials of Type IV hydrogen storage vessels using the Materials Studio
studies showed that film boiling plays a dominant role in the cooling (MS) software.
process. The temperature decrease caused by film boiling is 95.5 % of Unlike the monatomic molecule of argon, the hydrogen molecule is
the total cooling [20]. diatomic molecular. It has 3 degrees of freedom at low temperature and
Some scholars have studied the liquid hydrogen boiling problem 5 at room temperature. Since hydrogen atom is the smallest atom with a
using a two-phase flow simulation method based on the continuity diameter of 0.1 nm. Furthermore, its minimum radius of electron or­
assumption. Wang et al. [21] developed a mathematical model of liquid bitals is only 0.053 nm. The van der Waals forces of hydrogen molecules
hydrogen film boiling using the volume-of-fluid method (VOF) com­ have a small range of action. The above properties of hydrogen lead to
bined with Lee’s phase transition model. It was shown that the increase the boiling characteristics of liquid hydrogen may be different from
in the size of the bubbles released from the undulating gas-liquid other liquids. At present, there is a lack of molecular dynamics simula­
interface led to a decrease in the heat transfer performance. Through tion studies for liquid hydrogen boiling. This has led to a lack of insight
experiments and RANS-VOF simulation method, Kazuma et al. [22] into the boiling mechanism of liquid hydrogen.
found that pressure recovery caused by boiling delay occurs in liquid It has been shown that the molecular dynamics approach provides
hydrogen at the early stage of tank unloading. new ideas for the study of liquid hydrogen phase nucleation and bubble
Above studies provide valuable results for the study of liquid development. Therefore, in this study, molecular dynamics method was
hydrogen boiling mechanism. However, most of the current studies on used to investigate the bubble nucleation, development and boiling
liquid hydrogen phase change and boiling problems use experimental phenomenon of liquid hydrogen under different heating temperatures.
and mesoscopic or macroscopic continuous flow simulations. The scale The boiling process of liquid hydrogen studied in this research is the
of experimental objects is usually large, and the operation is hazardous. process from the beginning of heating to the formation of film boiling.
The experimental phenomena need to be combined with nucleation Our study mainly observed the bubble nucleation positions and bubble
theory for a convincing explanation. As well as, the phase change model kinetic behaviors.
used in CFD simulations needs to provide initial boundary conditions for The structure of this paper is as follows. In Section II, the computa­
bubble nucleation, such as the initial number and diameter of bubbles. tional model used for molecular dynamics simulations and the detailed
The bubble growth rate in the phase change model also needs to be computational parameters are presented. Section III focuses on the ef­
determined. fects of different heating temperatures on the bubble nucleation,
Molecular dynamics simulations are an effective method to study the development and boiling phenomenon of liquid hydrogen, as well as the
bubble nucleation mechanism and to determine the bubble growth rate. surface morphology of the heated plate. In this section, the paper firstly
The method can simulate the microscopic motion behavior at the mo­ observes the effects of different heating temperatures from the macro­
lecular scale. The microscopic and macroscopic properties of the system scopic perspective of molecular distribution. In addition, the liquid
can be obtained by performing statistics on the molecular states. The hydrogen boundary surfaces are extracted by Matlab and Ovito to
microscopic process of boiling phase transition has been studied by observe the bubble morphology. Then, in this part, the average kinetic
many scholars using molecular dynamics methods. Hens et al. [23] energy, potential energy, and total energy in the computational domain
investigated bubble nucleation on the surface of smooth substrates at are counted. By comparing the difference in energy distribution along
inhomogeneous temperatures. The results showed that bubble nucle­ the Z-direction, the nucleation of the liquid hydrogen phase on the
ation could be observed in the region with higher temperature. heated surface and the bubble development are then investigated.
Nagayama et al. [24] investigated the bubble nucleation behavior at the Finally, this part correlates the location of bubble formation on the
nanoscale. The results showed that the nucleation behavior of bubbles heated surface with the heated surface morphology.
was significantly different on smooth substrates with different wetta­
bility. Furthermore, they found that the Yang-Laplace equation is not 2. Simulation system and method
applicable at the nanoscale. Maruyama et al. [25] successfully observed
nucleation at the wall surface by changing the interfacial wettability at 2.1. Simulation system
the nanoscale. Yamamoto et al. [26] studied the initial stages of bubble
nucleation on smooth substrates. The initial stages of bubble nucleation Fig. 1 shows the simulation system’s initial configuration and heat
on smooth substrates with inhomogeneous superheat and surface transfer schematic. As shown in Fig. 1 (a), the computation is a cubic box
wettability were studied, respectively. Under these two conditions, with a size of 9.75 nm × 9.75 nm × 45 nm in three dimensions,
bubble nucleation was successfully generated near the substrate. In the respectively. The length in the Z direction is set large enough to ensure
vicinity of the substrate, the nucleation time was related to the the accuracy and reasonableness of the simulation results. Periodic

1131
X. Liu et al. International Journal of Hydrogen Energy 51 (2024) 1130–1141

σa + σb
σ a− b = (3)
2

where ε and σ denote the energy parameter and length parameter,

respectively. The subscripts a and b represent different types of atoms.
εa− b and σa− b are calculated by Eqs. (2) and (3) based on the Lorentz-
Berthelot combining rule [41]. The values of Lennard-Jones parame­
ters for H–H, Al–H are listed in Table 1.
The contact angles of the hydrogen droplet on the Al plate is 4.91◦ ,
which is shown in Fig. 2. According to experimental test data, the con­
tact angle of liquid hydrogen on Al6061 is between 0◦ and 4◦ [42].
The simulations are carried out by a large-scale atomic/molecular
massively parallel simulator (LAMMPS) [44], and the simulation
configuration is visualized by the OVITO software [45]. Two simulation
steps are required. First, a 500 fs equilibrium simulation with the tem­
perature of 20 K is conducted in the canonical ensemble (NVT). Then,
the substrate temperature is raised to specified temperature by a Lan­
gevin thermostat, and another 26 fs nonequilibrium simulation is con­
ducted to study the bubble nucleation process of liquid hydrogen (LH2)
in the microcanonical ensemble (NVE). During both simulations, the
position and velocity of each atom are updated every 0.0001 fs by a
Velocity-Verlet algorithm [46].

3. Results and discussion

Fig. 1. Simulation system set-up. The liquid hydrogens are colored red. The The description accuracy of the force fields of H2 is validated. In a
solid atoms are colored blue. (For interpretation of the references to color in liquid hydrogen simulation system, the density of the liquid phase re­
this figure legend, the reader is referred to the Web version of this article.) gion in the equilibrium stage in 20 K and 0.1 MPa is 70.301 kg/m3.
Compared to the saturated liquid density value of H2 at the same tem­
boundary conditions are applied in X and Y directions. In order to make perature and pressure condition from the NIST, the relative deviation is
the simulation more reasonable, the top of the Z-direction is set as a 1.37 %. The simulation result is in good agreement with the experi­
reflecting wall. The molecules are reflected from the top boundary mental data. Therefore, the force field used in this study can describe the
without any energy and momentum loss. The simulation box has three H2 system.
parts: substrate region, liquid region and vapor region. The aluminium
(Al), colored light blue in Fig. 1, is chosen as the metal substrate, because 3.1. The distribution and migration of the LH2 molecules during boiling
it is widely used in the hydrogen storage. Five layers of aluminium atoms
are placed with the arrangement of face-centered cubic structures (FCC) In this section, the boiling phenomenon of the LH2 on the smooth hot
with a lattice constant of 0.405 nm as the smooth surface bottom wall. wall at temperatures ranging from 100 K to 200 K were studied. In order
51,329 hydrogen molecules are arranged on the substrate surface to to analyze the spatial distribution of the number of hydrogen molecules,
form a liquid hydrogen film with a macroscopic density of 0.206 g/cm 3. the number of hydrogen molecules in the vertical direction (Z-direction)
The thickness of the liquid hydrogen layer is about 8.6 nm. According to in the calculation domain was counted. The specific statistical steps are
the available studies [35–37], this research belongs to thin liquid film 1) partitioning the computational domain along the z-direction with a
boiling. thickness of 0.25 Å, and 2) counting the total number of hydrogen
Due to the kinetic energy of the hydrogen molecules, some of the molecules in each zone within each 1 fs. The statistics of the number of
hydrogen molecules in the liquid film escape to the upper region to form hydrogen molecules in z-direction when the liquid hydrogen is heated at
gaseous hydrogen during the simulation of the thermodynamic equi­ a wall temperature of 100 K for 26 fs are shown in Fig. 3. In order to
librium process. It is noteworthy that the bottom five layers of better understand the correlation between the spatial distribution and
aluminium atoms are fixed to ensure that the solid wall is not deformed the number distribution of hydrogen molecules under the hot wall sur­
or shifted. The upper five layers of aluminium atoms are heated as a heat face, the spatial distribution of hydrogen molecules at 26 fs is placed on
source, i.e., a temperature control layer [38], whose temperature is the left side of the hydrogen molecule number curve. The results shown
controlled by a Langevin thermostat [39]. that a gaseous hydrogen layer was generated below the liquid hydrogen
at the current moment. Combined with the hydrogen molecule number
curve (the right half of Fig. 2), it can be seen that the number of
2.2. Potential models hydrogen molecules in z-direction can better quantify the density of
hydrogen molecules at different heights. This is important for dis­
A simple embedded-atom method (EAM) potential is conducted to tinguishing the gas-hydrogen and liquid-hydrogen distributions and
describe the potential between aluminium atoms. The interaction po­ analyzing the boiling state.
tential between hydrogen moleculers and aluminium atoms is described The wall temperature can have a significant effect on the boiling
by the Lennard-Jones (L-J) potential [40]. The potential energy is nucleation process of liquids [47]. According to the classical nucleation
expressed as follows:
[( ) ] Table 1
σ 12 (σ)6
φ(r)= 4ε − (1) Lennard-Jones parameters for H–H, Al–H.
r r
Interaction type ε/eV σ/Å
√̅̅̅̅̅
εa− b = εa εb (2) H–H 2.87 0.0025162 [43]
Al–H 2.75 0.01835

1132
X. Liu et al. International Journal of Hydrogen Energy 51 (2024) 1130–1141

Fig. 2. Equilibrium wettability contact angle at solid-liquid interface.

Fig. 4. The spatial distribution of hydrogen molecules in the computational

domain at different wall temperatures from 6 fs to 26 fs. From (a) to (e)
Fig. 3. The distribution of hydrogen molecules (the left half) and the number of correspond in turn to wall temperatures from 100 K to 200 K.
the hydrogen molecules along the z-direction (the right half) with the wall
temperature of 100 K at 26 fs. from molecular snapshots. Near the heated walls within the simulated
system, sporadic nuclei of bubbles are first generated, and as the bubbles
theory (CNT), when a liquid is heated up by high temperature wall grow and merge, they eventually form a gas film.
heating, the thermal motion of molecules intensifies. When the molec­ Fig. 5 shows that the distribution of hydrogen molecules at the wall
ular kinetic energy exceeds the critical potential energy barrier, the temperature of 200 K and the number of the hydrogen molecules along
bubble nucleation process is activated [48]. In this study, the heating the z-direction under the different wall temperatures during the boiling.
temperature of the solid wall was set to 100 K, 125 K, 150 K, 175 K and The aluminum wall thickness in z-direction is from 0 to 20 Å.
200 K, respectively. The results of the spatial distribution of hydrogen At the 6fs moment, no visible bubbles can be seen in the molecular
molecules in the computational domain at different wall temperatures snapshot. However, the molecular number curve shows that the density
and moments are shown in Fig. 4. According to the law of spatial dis­ of the liquid hydrogen layer near the heated walls decreases. The dif­
tribution of hydrogen molecules at different moments, it is known that ferences in molecular distribution in the calculated region are small for
the boiling process of liquid hydrogen is divided into three stages, i.e., different wall temperatures in the 6 fs-8 fs time range. The molecular
nuclear boiling, transition boiling and film boiling processes. In some number curves overlap well. At the same time, it can be found that the
areas near the wall, the liquid hydrogen first undergoes the phase density of the liquid hydrogen layer near the heated wall decreases
change and form tiny bubble nuclei. This phenomenon will be discussed further. This is due to the fact that at the beginning of heating, the ki­
in detail in section 3.3. During the bubble nucleation stage, the height netic energy of liquid hydrogen molecules increases, but does not
change of the upper surface of liquid hydrogen is not obvious. Subse­ completely break the potential energy limit. As a result, bubbles are not
quently, the number of hydrogen molecules with relatively sparse den­ formed stably.
sity between the liquid hydrogen and the solid wall surface gradually Starting from 10 fs, distinct isolated bubbles can be observed near the
increases, indicating that the bubbles below the liquid hydrogen grad­ heated walls in the molecular snapshot. At the same time, the number of
ually increase in size. The liquid hydrogen molecules are lifted up by the molecules near the heated walls decreases further. At 10 fs − 16 fs time,
gas hydrogen molecules that gradually phase change and vaporize due the bubbles undergo growth and merging. However, at this stage, a
to the continuous absorption of high kinetic atomic energy from the hot complete gas film is not formed on the surface of the heated wall surface.
wall. At the same time, the difference between the different heating temper­
References [36,37] very similar to the simulation system in this atures gradually becomes apparent as the bubbles have been formed
paper were chosen to compare the boiling phenomenon of hydrogen stably. The higher the wall temperature the faster the growth rate of
with that of argon. It was found that the molecular behavior of hydrogen bubbles. This is reflected in the graph of the molecular number distri­
molecules is very similar to that of argon atoms during boiling, as seen bution, where the peak in the number of hydrogen molecules is shifted

1133
X. Liu et al. International Journal of Hydrogen Energy 51 (2024) 1130–1141

molecules in the region close to the wall are subjected to high kinetic
energy aluminum atoms and receive the largest kinetic energy incre­
ment [49], resulting in the lowest number or molecular density of
hydrogen molecules. The ratio of maximum and minimum boiling gas
hydrogen molecule number is close to 6, which value varies weakly
under different wall temperature conditions. During molecular dy­
namics simulations, it is difficult to clearly delineate the gas-liquid
interface. This is because there is an interfacial region between the
liquid and the gas, which is continuous in density and composition. This
can also be observed in Fig. 5. In this research, the region where the
number of molecules accounts for 10%–90 % of the number of molecules
in the liquid layer region is considered as the gas-liquid interface region
[50], as shown in Fig. 6.
In order to study the variation of hydrogen temperature, a statistical
region (marked in green in Fig. 7) is created. The region of 10 Å thick­
ness above the heated wall was taken as the statistical region. As can be
seen in Fig. 7, the temperature variation in the statistical region is
roughly divided into four stages starting from the initial heating stage: a
smooth and slow increase stage, a fast increase stage, a fluctuating and
slow increase stage, and a stable stage. These four stages correspond to
the liquid hydrogen heat absorption stage, the isolated bubble genera­
tion stage, the bubble merging stage, and the stable gas film formation
stage, respectively. This coincides with the results in Fig. 5. It is observed
that the temperature variation pattern is similar in the monitoring area
of the wall heating temperature 100 K–125 K. The temperature variation
pattern is similar for the statistical region with the wall heating tem­
perature range of 150 K–200 K. The following is an example of the sit­
uation at 200 K wall surface temperature for analysis. Below 125 K
temperature, the temperature in the statistical region rises slowly and
steadily. During this phase, the liquid hydrogen molecules in the sta­
tistical region continuously absorb heat, but no bubbles are produced. In
the temperature range of 125 K–250 K, the temperature in the statistical
region rises rapidly and more smoothly. This is due to the fact that, at the
beginning of bubble generation, the bubbles are all isolated, which
makes the molecular number in the statistical region change more
steadily. In the temperature range of 250 K–300 K, the temperature in
the statistical region increases slowly and fluctuates widely. This is due
to the large bubble size and the merging and movement between bub­
bles. This leads to sharp fluctuations in the number of molecules in the
statistical region, which in turn leads to large fluctuations in the tem­
perature. Finally, the temperature in the statistical region remains stable
at about 360 K. At this point, the statistical region is completely gaseous.
Fig. 5. The distribution of hydrogen molecules (the top left, placed horizon­ At this point, the hydrogen molecules in the statistical region are
tally) and the number of the hydrogen molecules along the z-direction. The
horizontal coordinate is the height in the z-direction of the computational
domain shown in Fig. 1. The vertical coordinate is the number of
hydrogen molecules.

to the right by a distance.

It can be observed at 20 fs moments that the molecular number near
the hot wall surface approaches a value of 0. This represents the for­
mation of a stable gas film. In the 20 fs-26 fs time range, it can be
observed that the heating temperature has a more pronounced effect on
the growth of the gas film. The height of the liquid hydrogen layer in the
Z-direction at 100 K heating temperature is significantly lower than that
at 200 K And there is a positive correlation between the height of the
liquid hydrogen layer in the Z-direction and the heating temperature.
All in all, some of the hydrogen molecules are attached to the wall
surface by the attractive force between the aluminum atoms and the
hydrogen molecules. As time progresses, the number of attached
hydrogen molecules gradually decreases and the difference in hydrogen
molecule number distribution in z-direction at each wall temperature
condition gradually increases. As the temperature of the wall increases,
the height of liquid hydrogen lifted by the bubbles gradually increases.
The number of gas-hydrogen molecules between the wall surface and the
liquid hydrogen increases gradually along the z-direction. The hydrogen Fig. 6. Illustration for the interface thickness, 200 K at 20fs.

1134
X. Liu et al. International Journal of Hydrogen Energy 51 (2024) 1130–1141

interface at a wall temperature of 100 K from 3 fs to 24 fs. Between 3.2 fs

and 4 fs, the gas-liquid interface appeared on the surface of the
high-temperature wall. The locations of the gas-liquid interface pro­
jections are somewhat random. The number of gas-liquid interfacial
bumps increases from 4.4 fs. Between 4.4 fs and 5.0 fs, the small
gas-liquid interfacial bumps merge into larger bumps at the heated
surface. The gas-liquid interface raised structures continue to merge
with each other as the heating time increases. Between 5.0 fs and 5.6 fs,
the gas-liquid interface increases in the z-direction in the region of
formation, but instead of entering the liquid hydrogen layer, it has been
persistently retained above the heated wall. This microscopic phenom­
enon is consistent with the results of the nucleation cell boiling exper­
iment conducted by WARRIER et al. on the ISS [52]. At 8.0 fs, the
gas-liquid interface near the heated wall surface begins to flatten out and
move in the Z direction. The results show that the gas-liquid surface at
Fig. 7. The location of the statistical area and the pattern of tempera­ the microscopic level is not planar and there is a certain concave and
ture change.
convex microstructure. The surface morphology is influenced by the
local hydrogen molecule motion state. A non-vaporization molecular
completely gaseous hydrogen. In this study, the critical boiling tem­ layer tightly attached to the wall surface is observed. The phenomenon is
perature is considered to be the temperature at which isolated bubbles also described in the literature [53].
begin to be steadily generated, which is around 100 K. In this study, the volume of the gas phase region and the area of the
When comparing the temperatures of hydrogen and argon in the gas-liquid interface were counted in the time range of 2 fs - 26 fs, as
statistical region, it was found that the pattern of temperature variation shown in Fig. 9 (a) and (b). The volume of the gas phase region rises
with heating time was essentially the same in the region close to the rapidly in the time range 2 fs - 6 fs and reaches a maximum at the 6 fs
heated wall. However, the difference lies in the temperature required moment. It can be observed that the value of the area of the gas-liquid
and the time spent to enter the different stages. According to Hu et al. interface also increases rapidly in the same time range. At the early
the start time for the rapid vaporization of argon after entering the stage of bubble formation, the bubbles mostly exist independently and in
membrane boiling is 4.5 fs and the start temperature is 134 K (case1 in large numbers, and their volume grows rapidly after heat absorption,
reference) [47]. In this study, however, the start time for the rapid leading to a rapid increase in the volume of the gas-phase region and the
vaporization of hydrogen after entering the membrane boiling is 12 fs area value of the gas-liquid interface. In the time range of 6 fs - 20 fs,
and the start temperature is 300 K. It is suggested that the difference may there is a small decrease in the volume of the gas phase region. At the
be due to the fact that hydrogen is a diatomic molecule and therefore same time, the area value of the gas-liquid interface decreases slowly.
higher potential energy limits need to be broken during the phase This is due to the merging of bubbles during the bubble growth process.
transition. In the time range of 20 fs - 26 fs, there is an increasing trend in the
volume value of the gas-phase region. At the same time, there is a ten­
3.2. Evolution of the gas-liquid interface during boiling dency for the area value of the gas-liquid interface to reach a plateau.
This is due to the fact that, at this stage, a more stable gas film is formed
The results of the molecular dynamics calculations contain only in­ by the merging between the bubbles. These changes coincide with the
formation about the individual molecules and do not include the gas- results in Fig. 4. It was found that the volume of the gas phase region and
liquid interface distribution. In this study, the statistical results of the the area of the gas-liquid interface were less sensitive to the heating
spatial molecular number density at the gas-liquid interface using temperature.
Matlab software are interpolated and displayed with the aid of ovito
software. Datta et al. investigated the nucleation of nanobubbles on 3.3. The molecular kinetic and potential energy during boiling
hydrophilic and hydrophobic surfaces using a similar approach [51].
Fig. 8 shows the evolution of liquid hydrogen boiling and the gas-liquid In order to study the heat transfer phenomena during the nucleation

Fig. 8. The evolution of liquid hydrogen boiling and the gas-liquid interface (blue surface) at a wall temperature of 100 K from 3 fs to 24 fs.

1135
X. Liu et al. International Journal of Hydrogen Energy 51 (2024) 1130–1141

total energy distribution. Since the metal potential energy and total
energy in the heated plate region are much larger than the molecular
potential energy and total energy in the gas and liquid phase regions, it
is difficult to see the difference in the overall potential energy and total
energy distribution diagram of the computational domain, so the region
of 20.25 Å to 450 Å in the z-direction is taken out separately to study the
potential energy and total energy changes in the gas and liquid phase
regions. The molecular potential energy in the 20.25 Å to 25 Å region
decreases approximately linearly from − 0.027 eV to − 0.007 eV. This is
due to the fact that there is still an ungasified liquid layer between the
heating plate in the bubble domain during the nucleation boiling stage,
and the molecules far from the surface of the heating plate are subject to
lower potential energy binding and higher gasification. However, the
potential energy at 25 Å does not drop to 0 eV because no gas film is
formed at this time and the number of molecules in the tangent region is
still high. In the region of 25 Å to 125 Å is the liquid layer region, where
25 Å to 50 Å is the gas-liquid interface from the bubble region to the pure
liquid phase, the molecular density gradually increases along the z-di­
rection, and the average potential energy gradually increases in the
negative direction to − 0.024 eV. In the region of 50 Å to110 Å is the
stable liquid layer region, the average potential energy is stable at
− 0.024 eV. 110 Å-125 Å is the gas-liquid interface between the liquid
layer and the vapor layer, the average potential energy linearly de­
creases from − 0.024 eV to 0 eV. The mean potential energy decreases
linearly from − 0.024 eV to 0 eV. 125 Å-450 Å is the vapor region, where
the mean potential energy remains stable at 0eV. The negative spikes in
the mean potential energy at 230 Å and 440 Å are due to the presence of
droplet hydrogen molecular clusters at this location.
The higher the wall temperature, the more intense the phase change
and heat transfer. As a result, the number of hydrogen molecules at
various heights within the bubble region decreases (meaning the
hydrogen gas density inside the bubble decreases). As the wall tem­
perature increases, the potential energy of hydrogen molecules within
the bubble region decreases. This is because as the temperature rises, the
kinetic energy of the molecules inside the bubble increases, causing the
average distance between the molecules to increase and weakening the
Fig. 9. The change pattern of gas phase region volume (a) and gas phase region gravitational forces. Since the molecular potential energy is related to
surface aera (b). the intermolecular distance, the potential energy decreases.
Fig. 11 shows the distribution of the number of molecules, kinetic
and development of liquid hydrogen bubbles, the molecular kinetic energy, potential energy and total energy in the calculated domain for
energy, potential energy, total energy, and molecular number distribu­ different heating temperatures at 26 fs. The kinetic energy is stable at
tions along the z-direction in the computational domain under the about 0.017 eV in the region of the heated plate at 0 to 20.25 Å. The
nucleation boiling and membrane boiling stages, as well as the tem­ reason for the sudden increase in kinetic energy at 20.25 Å to 25 Å,
perature distribution cloud of the computational domain, are calculated where the number of molecules decreases and the kinetic energy in­
in this section. Fig. 10 shows the molecular number and energy distri­ creases rapidly, is that although the gas film layer has been generated
bution along the z-direction of the computational domain for different under the inscription and is in the state of film boiling, some hydrogen
heating temperatures at heating 6 fs. The kinetic energy distribution of molecules are bound by intermolecular forces to form a gas-solid
the computational domain along the Z-direction can be roughly divided interface with a certain thickness on the surface of the heating plate
into four regions: the heated plate region, the gas layer region, the liquid due to the LJ potential energy between aluminum atoms and hydrogen
layer region, and the upper vapor region. The kinetic energy stabilizes molecules, and the hydrogen molecules in the interface away from the
around 0.017 eV in the heated plate region from 0 to 20.25 Å. There is a heating surface have a higher kinetic energy because they are weaker
sudden increase in kinetic energy at 20.25 Å-25 Å. The reason for the bound. The kinetic energy of the hydrogen molecules in the interface far
decrease in the number of molecules and the rapid increase in kinetic from the heating surface is larger because of the weaker bonding. The
energy in this region may be that the heating surface is in a nucleated 25 Å to 140 Å region is the gas film region, and the kinetic energy
boiling state at this moment, but the existence of an unevaporated liquid gradually decreases and fluctuates along the increase of z-coordinate in
layer on the heating surface is the cause. The liquid layer away from the the region, and the potential energy increases a little and changes more
heating surface is relatively less bounded by the potential energy of the gently, and the total energy shows a decreasing trend. Corresponding to
heating plate, so the molecular kinetic energy is higher. The molecular the molecular number change in (b), the gas film region at 100 K–200 K
kinetic energy in the liquid hydrogen region at the height of 25 Å to 115 heating temperature is roughly 25 Å to 100 Å, 25 Å to 110 Å, 25 Å to 115
Å decreases and finally stabilizes at about 0.005 eV. Due to the wide Å, 25 Å to 135 Å, 25 Å to 140 Å. The molecular number in the gas film
range of motion and small constraints of gaseous molecules, the mo­ region has a small increase along the z-direction, while the kinetic en­
lecular kinetic energy in the vapor region of 115 Å to 450 Å fluctuates ergy decreases, the potential energy increases, and the total energy de­
considerably. Fig. 10 (c), (d), (e) and (f) show the results of potential creases. The 25 Å to 140 Å region is the gas film region, and the kinetic
energy and total energy distribution in the calculated domain at each energy gradually decreases and fluctuates along the increase of z-coor­
heating wall temperature. The law of change of potential energy dis­ dinate in the region, and the potential energy increases a little and
tribution in the calculation domain is very similar to the law of change of changes more gently, and the total energy shows a decreasing trend.

1136
X. Liu et al. International Journal of Hydrogen Energy 51 (2024) 1130–1141

Fig. 10. The molecular number and energy distribution along the z-direction of the computational domain for different heating temperatures at heating 6 fs. (d) and
(f) are obtained by amplifying the linear parts in (c) and (e).

There is a small increase in the number of molecules along the z-direc­ show local fluctuations.
tion in the gas film region, while the kinetic energy decreases, the po­
tential energy increases, and the total energy decreases due to the fact
that the hydrogen molecules close to the liquid layer are bound by the 3.4. Wall micro-pits and initial bubble nucleation due to thermal motion
molecular potential energy within the high-density liquid layer. 140 Å to of wall atoms
240 Å is the liquid layer region, in which the molecular number and
average potential energy along the z-direction increase rapidly from 10, The wall atoms change their ideal atomic arrangement structure at
0.002 eV–240 eV, − 0.017 eV, and this part is considered as the gas- the initial moment under their own thermal motion, which can lead to a
liquid interface on the bubble side; then the molecular number and random wall microstructure. Fig. 12 shows the results of liquid hydrogen
average potential energy increase slowly to a maximum of 300, -0.024 boundary morphology, aluminum atomic arrangement on the heating
eV, and this part is considered as the liquid layer; finally the molecular plate surface, heating plate temperature and stress distribution at 4 fs
number and average potential energy decrease rapidly and Finally, the with the heating temperatures of 100 K, 125 K, 150 K, 175 K and 200 K.
number of molecules and the average potential energy drop rapidly and The results of the wall micromorphology show that the atomic
remain stable around 0 eV. This part is considered as the vapor-liquid arrangement is not the same at different heating temperatures. Micro-
interface. The kinetic energy in the liquid layer region changes along bumps and micro-pits are present in localized areas. This microscopic
the z-direction with a logarithmic curve close to the bottom less than 1 rough state due to the thermal motion of the atoms cannot be avoided.
eV, and gradually approaches 0 eV. From 240 Å to 450 Å is the upper Since there is less bubble formation at the beginning of heating and most
vapor region, where the average molecular number and average po­ of the liquid hydrogen is in direct contact with the heating surface, the
tential energy are close to 0 eV, and the kinetic energy and total energy micromorphology of the solid wall surface can be reflected by the
morphology of the contact surface between the hydrogen atoms and the

1137
X. Liu et al. International Journal of Hydrogen Energy 51 (2024) 1130–1141

Fig. 11. The molecular number and energy distribution along the z-direction of the computational domain for different heating temperatures at heating 26 fs.

solid wall surface. The present study conducted a statistical analysis of LJ potential energy binding to form small gas nuclei. However, due to
the temperature distribution on the surface of aluminum. The method the small size of the pits (on the order of Å), the vaporized hydrogen
for identifying aluminum atoms located in the surface region is illus­ molecules cannot get enough energy, so the initial bubble nucleation is
trated in Fig. 13. If a hydrogen molecule exists within a specific radius unstable, leading to the phenomenon of small gas nuclei forming and
spherical region with the aluminum atom’s center as the sphere’s center, disappearing. From a classical thermodynamic point of view is that the
it indicates that the aluminum atom is on the surface. For this study, the size of these microbubbles does not reach the critical nucleation radius.
radius of the spherical region was set to 3 Å. Subsequently, based on the Fig. 14 presents the temperature results of surface aluminum atoms
vertical coordinate of the surface aluminum atoms, they were divided at a time step of 7,220,000 steps (6 fs). The orange dots represent the
into two regions: the high surface region and the low surface region temperature values of aluminum atoms in the high surface region, while
(with the midpoint height as the boundary). This division allows for the the blue dots represent the temperature values of aluminum atoms in the
determination of the instantaneous and average temperatures of the low surface region. It can be observed that the scatter plot does not
aluminum atoms in both regions. clearly reveal the temperature characteristics of the two layers of
Multiple pits and bumps appear on the solid-liquid interface (the aluminum atoms. However, when the temperature of the high and low
second column of Fig. 12). This phenomenon also appears in some layers of aluminum atoms is statistically averaged, the temperature
literature [54,55]. Since it is a microstructure caused by the offset of patterns become evident (See Fig. 15). Fig. 15 displays the average
atomic motion, pits exist near the bumps in general. Obvious bubble temperature results of the high and low layers of aluminum atoms at
nucleation exists at the location of the lower surface of the liquid four different computational time steps. The results indicate that due to
hydrogen (blue surface in the first column of Fig. 12) corresponding to the stochastic nature of collisions and energy exchange between
the wall pits. The above simulation results can explain the random hydrogen molecules and aluminum atoms during the boiling process,
boiling mechanism of liquid at high temperature for an ideal smooth the average temperature of aluminum atoms exhibits some fluctuations.
planar solid wall surface. The micro-pits restrict the movement of However, the average temperature of the high layer aluminum atoms is
hydrogen molecules to some extent, so a small number of hydrogen slightly lower than that of the low layer (pits) aluminum atoms.
molecules continue to absorb energy inside the pits and finally break the

1138
X. Liu et al. International Journal of Hydrogen Energy 51 (2024) 1130–1141

Fig. 12. The liquid hydrogen boundary morphology, aluminum atomic arrangement on the heating plate surface, heating plate temperature and stress distribution at
4 fs with the heating temperatures of 100 K, 125 K, 150 K, 175 K and 200 K.

4. Conclusions

In this paper, a molecular dynamics model of liquid hydrogen heated

on a flat plate is developed. Under different heating temperatures, the
liquid hydrogen molecule behaviors in the different stages of phase
change were investigated by molecular dynamics simulation methods.
In addition, the correlation analysis of heating temperature, heating
plate surface morphology and bubble nucleation sites was also carried
out in this paper. The main conclusions are summarized as follows:

(1) In this study, the bubble nucleation, bubble development and

Fig. 13. The diagram of the method for identifying aluminum atoms located in merging processes of liquid hydrogen on the heated surface were
the surface region. successfully observed. The number of molecules and energy
changes along the z-direction of the computational domain were
also counted. According to the results of the molecular number
distribution, the difference between different heating

1139
X. Liu et al. International Journal of Hydrogen Energy 51 (2024) 1130–1141

temperature distribution figures and the force distribution in the

z-direction were jointly analyzed. In this study, it is concluded
that the small deformation on the surface of the heated plate was
due to the thermal motion of aluminum atoms. The part sur­
rounded by the protruding atoms formed a crater. In this paper, it
is found that on the surface of the heated plate, the aluminum
atoms that made up the craters were at a higher temperature,
which made it easier for the liquid hydrogen to phase change and
form bubbles.

In addition to this, one more point is worth discussing. The molecular

dynamics method is a study of microscopic particles, which integrates
the interactions between microscopic particles, so that the behavioral
properties of microscopic particles can be simulated, and the macro­
scopic thermophysical properties can be statistically derived from the
large amount of data derived from the simulation, therefore, the method
can be used to simulate microscopic phenomena that are difficult to be
Fig. 14. The temperature results of surface aluminum atoms at a time step of
directly observed by experiments and then derive macroscopic ther­
7,220,000 steps (6 fs).
mophysical properties. However, the potential function needs to be
selected during the molecular dynamics simulation. The value of the
potential function will affect the simulation results. Therefore, this point
needs to be considered and studied in the future research process.

Declaration of competing interest

The authors declare that they have no known competing financial

interests or personal relationships that could have appeared to influence
the work reported in this paper.

Acknowledgements

The research is funded by Beijing Engineering Research Center of

Efficient and Green Aerospace Propulsion Technology and Advanced
Space Propulsion Laboratory of BICE, No: LabASP-2023-08.

Fig. 15. The average temperature results of the high and low layers of References
aluminum atoms at four different computational time steps.
[1] Khan U, Yamamoto T, Sato H. An insight into potential early adopters of hydrogen
fuel-cell vehicles in Japan. Int J Hydrogen Energ 2021;46:10589–607.
temperatures gradually becomes apparent as the bubbles form [2] Mao X, Ying R, Yuan Y, Li F, Shen B. Simulation and analysis of hydrogen leakage
steadily. The higher the wall temperature, the faster the growth and explosion behaviors in various compartments on a hydrogen fuel cell ship. Int J
of bubbles. Hydrogen Energy 2021;46:6857–72.
[3] Halder P, Babaie M, Salek F, et al. Advancements in hydrogen production, storage,
(2) Combining Lammps software with Matlab and Ovito, the liquid distribution and refuelling for a sustainable transport sector: hydrogen fuel cell
hydrogen boundary surface was plotted. Moreover, analyzing the vehicles. Int J Hydrogen Energy 2023. In press.
morphology of the liquid hydrogen boundary surface, this paper [4] Teimouri A, Zayer Kabeh K, Changizian S, et al. Comparative lifecycle assessment
of hydrogen fuel cell, electric, CNG, and gasoline-powered vehicles under real
found that the bubbles were extremely unstable at the early stage driving conditions[J]. Int J Hydrogen Energy 2022;47(89):37990–8002.
of generation. Within 1 fs time after the bubbles first appear, [5] Ahmadi P, Khoshnevisan A. Dynamic simulation and lifecycle assessment of
there was a phenomenon that bubble moved on the heated plate hydrogen fuel cell electric vehicles considering various hydrogen production
methods[J]. Int J Hydrogen Energy 2022;47(62):26758–69.
surface. Furthermore, sometimes this bubble disappeared. With
[6] Yan J, Jing J, Li Y. Hydrogen fuel cell commercial vehicles in China: evaluation of
the increasing of heating time, the number of bubbles gradually carbon emission reduction and its economic value. Int J Hydrogen Energy 2023. In
increases, and the bubble formation process reached a stable press.
[7] Petrescu RVV, Machín A, Fontánez K, Arango JC, Márquez FM, Petrescu FIT.
state.
Hydrogen for aircraft power and propulsion. Int J Hydrogen Energy 2020;45:
(3) By analyzing the average kinetic energy, potential energy, and 20740–64.
total energy statistics along the z-direction of the computational [8] Cecere D, Giacomazzi E, Ingenito A. A review on hydrogen industrial aerospace
domain, this paper used the results of energy distribution to applications[J]. Int J Hydrogen Energy 2014;39(20):10731–47.
[9] Arat HT, Sürer MG. Experimental investigation of fuel cell usage on an air Vehicle’s
explain the molecular number distribution regularity. As the wall hybrid propulsion system[J]. Int J Hydrogen Energy 2020;45(49):26370–8.
temperature gets higher, the phase transition and heat transfer [10] Farinha J., Silva L.M., Matlock J., et al. Hydrogen fuel cell integration and testing
becomes more intense. This leads to a decrease in the potential in a hybrid-electric propulsion rig[J]. Int J Hydrogen Energy 2023. In press.
[11] Tatsumoto H., Shirai Y., Shiotsu M., Hata K., Naruo Y., Kobayasi H., et al. Forced
energy and an increase in the kinetic energy of the hydrogen convection heat transfer of saturated liquid hydrogen in vertically-mounted heated
molecules within the bubble region. This causes the average pipes. 2014; 1573 (1): 44–51.
distance between the molecules to increase and the gravitational [12] Darr S, Hu H, Glikin N, Hartwig J, Majumdar AK, Leclair A, et al. An experimental
study on terrestrial cryogenic tube chilldown II. Effect of flow direction with
force to weaken. This paper also found the gas-liquid boundary respect to gravity and new correlation set. Int J Heat Mass Tran 2016;103:
layer phenomenon and explained the phenomenon using the 1243–60.
intermolecular potential energy interaction. [13] Yang N., Deng J., Wang C., et al. High pressure hydrogen leakage diffusion:
research progress[J]. Int J Hydrogen Energy 2023. In press.
(4) The liquid hydrogen boundary surface images at the early stage of
[14] Tong S., Li X., Ding H., et al. Large-scale transient simulation for consequence
heating, the snapshots of the heating plate surface, the analysis of hydrogen-doped natural gas leakage and explosion accidents[J]. Int J
Hydrogen Energy 2023. In press.

1140
X. Liu et al. International Journal of Hydrogen Energy 51 (2024) 1130–1141

[15] Haoren W, Bo W, Ruize L, et al. Theoretical investigation on heat leakage [36] Hu C, Pei Z, Shi L, et al. Phase transition properties of thin liquid films with various
distribution between vapor and liquid in liquid hydrogen tanks[J]. Int J Hydrogen thickness on different wettability surfaces[J]. Int Commun Heat Mass Tran 2022;
Energy 2023;48(45):17187–201. 135:106125.
[16] Class C, Dehaan JR, Piccone M, Cost RB. Boiling heat transfer to liquid hydrogen [37] Wang Y, Wang S, Lu G, et al. Explosive boiling of nano-liquid argon films on high
from flat surfaces. 1960. temperature platinum walls: effects of surface wettability and film thickness[J]. Int
[17] Seader JD, Miller WS, Kalvinskas LA. Boiling heat transfer for cryogenics. 1965. J Therm Sci 2018;132:610–7.
Web. [38] Deng X, Xu X, Song X, et al. Boiling heat transfer of CO2/lubricant on structured
[18] Shirai Y, Tatsumoto H, Shiotsu M, Hata K, Kobayashi H, Naruo Y, et al. Boiling heat surfaces using molecular dynamics simulations[J]. Appl Therm Eng 2023;219:
transfer from a horizontal flat plate in a pool of liquid hydrogen. Cryogenics 2010; 119682.
50:410–6. [39] Davidchack RL, Handel R, Tretyakov MV. Langevin thermostat for rigid body
[19] Shiotsu M., Kobayashi H., Takegami T., Shirai Y., Tatsumoto H., Hata K., et al. dynamics[J]. J Chem Phys 2009;130(23):234101.
Transient heat transfer from a horizontal flat plate in a pool of liquid hydrogen. [40] Zhou W, Li Y, Li M, et al. Bubble nucleation over patterned surfaces with different
2012; 1434 (1): 1059–1066. wettabilities: molecular dynamics investigation[J]. Int J Heat Mass Tran 2019;136:
[20] Merte H. Incipient and steady boiling of liquid nitrogen and liquid hydrogen under 1–9.
reduced gravity. 1970. [41] Morrow BH, Harrison JA. Interfacial properties of linear alkane/nitrogen binary
[21] Wang J, Li Y, Wang L, Xia S, Ren J, Mao H, et al. Numerical investigation on mixtures: molecular dynamics vapor–liquid equilibrium simulations[J]. J Phys
subcooled pool film boiling of liquid hydrogen in different gravities. IInt J Chem B 2022;126(23):4379–88.
Hydrogen Energy 2021;46:2646–57. [42] Bellur K, Konduru V, Kulshrestha M, et al. Contact angle measurement of liquid
[22] Tani K, Himeno T, Sakuma Y, Watanabe T, Kobayashi H, Toge T, et al. Pressure hydrogen (LH2) in stainless steel and aluminum cells[J]. Journal of Heat Transfer-
recovery during pressure reduction experiment with large-scale liquid hydrogen transactions of The Asme 2016;138:20904.
tank. Int J Hydrogen Energy 2021;46:29583–96. [43] Wilhelm E, Battino R. Estimation of lennard-jones (6, 12) pair potential parameters
[23] Hens A, Agarwal R, Biswas G. Nanoscale study of boiling and evaporation in a from gas solubility data[J]. J Chem Phys 1971;55(8):4012–7.
liquid Ar film on a Pt heater using molecular dynamics simulation. Int J Heat Mass [44] Plimpton S. Fast parallel algorithms for short-range molecular dynamics[J].
Tran 2014;71:303–12. J Comput Phys 1995;117(1):1–19.
[24] Nagayama G, Tsuruta T, Cheng P. Molecular dynamics simulation on bubble [45] Alexander S. Visualization and analysis of atomistic simulation data with
formation in a nanochannel. Int J Heat Mass Tran 2006;49:4437–43. OVITO–the Open Visualization Tool[J]. Model Simulat Mater Sci Eng 2009;18(1):
[25] Maruyama S, Kimura T. A molecular dynamics simulation of a bubble nucleation 015012.
on solid surface. Transactions of the Japan Society of Mechanical Engineers. B. [46] Al-Chaabawi MJH, Abdollahi A, Najafi M. Pool boiling heat flux of ammonia
1999;65:3461–7. refrigerant in the presence of iron oxide nanoparticles: a molecular dynamics
[26] Yamamoto T, Matsumoto M. Initial stage of nucleate boiling: molecular dynamics approach[J]. Eng Anal Bound Elem 2023;151:387–93.
investigation. J Therm Sci Tech-JPN. 2012;7:334–49. [47] Guo W, Zeng L, Liu Z. Mechanism of surface wettability of nanostructure
[27] Zhou W, Li Y, Li M, Wei J, Tao W. Bubble nucleation over patterned surfaces with morphology enhancing boiling heat transfer: molecular dynamics simulation[J].
different wettabilities: molecular dynamics investigation. Int J Heat Mass Tran Processes 2023;11(3):857.
2019;136:1–9. [48] Dinh TN, Theofanous TG. Nucleation phenomena in boiling[J]. Multiphas Sci
[28] Chen Y, Yu B, Zou Y, Chen B, Tao W. Molecular dynamics studies of bubble Technol 2003;15(1–4).
nucleation on a grooved substrate. Int J Heat Mass Tran 2020;158:119850. [49] Yi P, Poulikakos D, Walther J, et al. Molecular dynamics simulation of vaporization
[29] Fu T, Mao Y, Tang Y, Zhang Y, Yuan W. Effect of nanostructure on rapid boiling of of an ultra-thin liquid argon layer on a surface[J]. Int J Heat Mass Tran 2002;45
water on a hot copper plate: a molecular dynamics study. Heat Mass Tran 2016;52: (10):2087–100.
1469–78. [50] Luo K, Xiao G, Ma X, et al. Liquid fuel evaporation under supercritical conditions
[30] Wu N, Zeng L, Fu T, Wang Z, Lu C. Molecular dynamics study of rapid boiling of [J]. Commun Comput Phys 2017;23(4):1241–62. 2018.
thin liquid water film on smooth copper surface under different wettability [51] Datta S, Pillai R, Borg MK, et al. Acoustothermal nucleation of surface nanobubbles
conditions. Int J Heat Mass Tran 2020;147:118905. [J]. Nano Lett 2021;21(3):1267–73.
[31] Li B, Gu Y, Chen M. Cavitation inception of water with solid nanoparticles: a [52] Warrier GR, Dhir VK, Chao DF. Nucleate pool boiling eXperiment (NPBX) in
molecular dynamics study. Ultrason Sonochem 2019;51:120–8. microgravity: international Space station[J]. Int J Heat Mass Tran 2015:83781–98.
[32] Yu H, Li Z, Tao Y, Sha J, Chen Y. Thermal bubble nucleation in graphene [53] Mao Y, Zhang Y. Molecular dynamics simulation on rapid boiling of water on a hot
nanochannels. J Phys Chem C 2019,123;6:3482–90. copper plate[J]. Appl Therm Eng 2014;62(2):607–12.
[33] Jiang H, Cheng X, Zhang H, et al. Molecular dynamic simulation of high-quality [54] Lin X, Zhu X, Zhou Z, et al. Molecular dynamics simulation on phase transition
hydrogen storage in pillared bilayer graphene bubble structure[J]. Computational behaviors of R410A liquid film over solid copper surface[J]. J Mol Liquids. 2023.
and Theoretical Chemistry 2015;1068:97–103. In press.
[34] Fang Q, Ji D. Molecular simulation of hydrogen permeation behavior in liner [55] Rony MD, Islam MA, Thakur MSH, et al. Molecular dynamics data-driven study of
polymer materials of Type IV hydrogen storage vessels[J]. Mater Today Commun leidenfrost phenomena in context to liquid thin film phase transformation[J]. Int J
2023;35:106302. Heat Mass Tran 2023;209:124107.
[35] Rony MD, Islam MA, Thakur MSH, et al. Molecular dynamics data-driven study of
leidenfrost phenomena in context to liquid thin film phase transformation[J]. Int J
Heat Mass Tran 2023;209:124107.

1141