SCIENCE CHINA

Physics, Mechanics & Astronomy

Article October 2013 Vol.56 No.10: 17
doi: 10.1007/s11433-013-5209-z
doi: 10.1007/s11433-013-5209-z

An experimental investigation into the icing and melting process

of a water droplet impinging onto a superhydrophobic surface
JIN ZheYan1*, JIN SongYue1 & YANG ZhiGang2,3
1
School of Aerospace Engineering and Applied Mechanics, Tongji University, Shanghai 200092, China;
2
School of Automotive Studies, Tongji University, Shanghai 201804, China;
3
Shanghai Automotive Wind Tunnel Center, Shanghai 201804, China

Received June 29, 2013; accepted August 1, 2013

The freezing and melting process of a small water droplet on a superhydrophobic cold surface was investigated using the Laser
Induced Fluorescence (LIF) technique. The superhydrophobic surface was prepared using a sol-gel method on a red copper test
plate. From the obtained fluorescence images, the phase transition characteristics during the freezing and melting process of a
water droplet were clearly observed. It was found that, at the beginning of the droplet freezing process, liquid water turned into
ice at a very fast rate. Such phase transition process decreased gradually with time and the volume of frozen ice approached a
constant value at the end of the icing process. In addition, the freezing time was found to reduce with the decrease of the test
plate temperature. Besides, when the test plate temperature is relatively high, the effect of droplet volume on the freezing time
is very significant. Over all, we provide some tentative insights into the microphysical process related to the icing and melting
process of water droplets.

sol-gel, icing, melting, superhydrophobic surface, water droplet

PACS number(s): 44.10.+i, 68.08.-p, 68.18.Jk

Citation: Jin Z Y, Jin S Y, Yang Z G. An experimental investigation into the icing and melting process of a water droplet impinging onto a superhydrophobic
surface. Sci China-Phys Mech Astron, 2013, doi: 10.1007/s11433-013-5209-z

1 Introduction
When humid air encounters a cold surface whose tempera-
ture is below both the dew point of the air and the freezing
point of water, frost may accrete on the surface. The occur-
rence of frosting is undesirable for most industrial applica-
tions such as air conditioners and refrigerators. A frost layer
formed on an evaporator not only increases the air side
thermal resistance, but also reduces the airflow area, both of
which lower the evaporator performance.
Hayashi et al. [1] divided the frost formation process into
three different periods: that is, the crystal growth, the frost
layer growth, and the frost layer full growth periods. The
*Corresponding author (email: zheyanjin@tongji.edu.cn)
frost crystal growth period, i.e., the freezing process of wa-
ter droplets, is of significant importance even though the
process is very fast. This period is the starting point of the
frost deposition process and has attracted many researchers
over the years. Carte [2] experimentally investigated the
freezing process of water droplets on various surfaces by
using a microscope method. The relationship between the
lowest freezing temperatures with surfaces, the volume of
the water droplets, and the rate of cooling were obtained.
Visagie [3] employed a small quartz Bourdon tube to meas-
ure the pressures that developed inside freezing water drop-
lets and found that the cracking pressure tended to increase
with the shell thickness and was dependent on the freezing
rate. Feuillebois et al. [4] numerically studied the two-step
freezing processes of a subcooled liquid droplet, without
considering the combined effect of a variation in mass den-

Science China Press and Springer-Verlag Berlin Heidelberg 2013 phys.scichina.com www.springerlink.com
2 Jin Z Y, et al. Sci China-Phys Mech Astron
sity during freezing with the Stefan problem. Hoke [5] ob-
served the frost deposition on a variety of test plates and
found that the water distribution on the test plate at the end
of the condensation period was strongly dependent on test
plate temperature, humidity ratio, and the test plate contact
angle. Tabakova and Feuillebois [6] modeled the solidifica-
tion and subsequent cooling of a supercooled liquid droplet
that was lying on a cold solid test plate after impact, with
the assumption that solidification occurred for a given fixed
droplet shape. Hindmarsh et al. [7] used magnetic resonance
imaging (MRI) to provide spatially resolved structural and
chemical composition characterization of freezing droplets.
Recently, Wang et al. [8] experimentally investigated the
freezing processes of water and peanut oil droplets on a cold
surface and the protrusion phenomenon observed during the
tests was explained in detail. Wu et al. [9] illustrated the
condensation and freezing process of water droplets on a
cold surface. Suzuki et al. [10] studied the effects of smooth
and rough coatings of FAS-17 on contact angles (CA), slid-
ing angles, and freezing temperatures of the water droplets.
Nitsch experimentally studied water freezing and super-
cooling by methods of thermal analysis in a water droplet of
5 mm in diameter and found the degree of water supercool-
ing was influenced by previous thermal treatment but only
to a certain extent [11]. Jin and Hu [12,13] implemented a
lifetime-based Molecular Tagging Thermometry (MTT)
technique to reveal the time evolution of the unsteady heat
transfer and dynamic phase changing process within small
icing water droplets.
More recently, superhydrophobic surfaces, because of
their outstanding water repellency, have attracted the inter-
est of many researchers to investigate their icephobic be-
havior. Kulinich and Farzaneh [14] measured the adhesion
strength of artificially created glaze ice on rough fluoro-
polymer-based hydrophobic surfaces with different CA and
wetting hysteresis. Their results showed that, in general, the
ice adhesion strength on rough hydrophobic surfaces has no
correlation with the values of water CA. Cao et al. [15]
conducted experiments to study the anti-icing capabilities of
superhydrophobic surfaces prepared by using nanoparti-
cle-polymer composites. They demonstrated that certain
coatings were able to prevent ice formation upon impact of
supercooled water both in laboratory conditions and in nat-
urally occurring environments. Tourkine et al. [16] found
that when compared to normal copper test plates, the freez-
ing process of water droplets could be delayed on superhy-
drophobic copper test plates. Mishchenko et al. [17] inves-
tigated the behavior of dynamic droplets impacting super-
cooled nano- and microstructured surfaces. Their results
showed that highly ordered superhydrophobic materials
could be designed to remain entirely ice-free down to ca.
25 to 30C, due to their ability to repel impacting water
before ice nucleation occurred. He et al. [18] found that
frost formation was greatly retarded on the superhydropho-
bic i-PP surface. Kulinich et al. [19] reported that the anti-
October (2013) Vol. 56 No. 10
icing efficiency of superhydrophobic surfaces was signifi-
cantly lower in a humid atmosphere, as water condensation
both on top of and between surface asperities took place,
leading to significantly larger values of ice adhesion strength.
Jung et al. [20] observed that surfaces with nanometer-scale
roughness and higher wettability display unexpectedly long
freezing delays, at least 1 order of magnitude longer than
typical superhydrophobic surfaces with larger hierarchical
roughness and low wettability. Bahadur et al. [21] developed
a physics-based modeling frame work to predict ice for-
mation on cooled superhydrophobic surfaces resulting from
the impact of supercooled water droplets. Alizadeh et al. [22]
observed that the reduction of water-surface contact area on
superhydrophobic surfaces could reduce both heat transfer
and the probability of heterogeneous nucleation at the wa-
ter-test plate interface. Jung et al. [23] found that humidity
had a drastic effect on the roll-off/sliding of a supercooled
droplet on superhydrophobic surfaces under shear flow.
Even though extensive researches have been conducted
in previous studies [223], to the authors best knowledge,
the complete freezing and melting process of a water drop-
let on a superhydrophobic surface has not yet been experi-
mentally investigated. Considering that water droplets
eventually freeze on a superhydropobic surface when the
relative humidity of the surrounding air is high [19,23], a
further investigation on the freezing and melting process of
water droplets on superhydropobic surfaces would be de-
sirable for engineers and scientists to develop more efficient
de-icing methods. In this study, we report an experimental
visualization of the freezing and melting process of a small
water droplet on a cold superhydrophobic surface by using
the Laser Induced Fluorescence (LIF) technique. The
change of the volume of frozen ice with time during the
droplet freezing process was obtained and the effect of the
test plate temperature on the freezing time of different water
droplets was also analyzed. The present study is aimed to
elucidate the underlying fundamental physics to improve
our understanding about the important microphysical pro-
cesses pertinent to the icing phenomena.
2 Experimental
2.1 Pretreatment of red copper test plate
A red copper test plate (50.0 mm W 50.0 mm L 1.5 mm T)
was pretreated by grinding them with 1500# sand papers
and ultrasonically cleaning them in acetone and ethanol for
10 min and 30 min, respectively. After that, the red copper
test plates were washed with distilled water and then dried
in an oven.
2.2 Preparation of coatings
The superhydrophobic surface was prepared using a sol-gel
 Jin Z Y, et al. Sci China-Phys Mech Astron
method on a red copper test plate [24]. Chemicals used were
aluminum isopropoxide, toluene, ethyl acetoacetate, stearic
acid, methanol, N-hexane, and phosphoric acid.
The superhydrophobic silica coating was synthesized by
following four main steps: (1) alumina sol preparation, (2)
pretreatment of the test plate, (3) dip coating, and (4) sur-
face modification. Firstly, aluminum isopropoxide (AlP)
was dissolved in toluene under stirring at room temperature.
Ethyl acetoacetate (EAcAc) was added to the solution as a
chelating agent, and the solution was stirred for 3 h. Water
was then added to the solution for hydrolysis. This solution
was stirred vigorously for 12 h and used for coating. The
molar ratios of toluene, EAcAc, and H2O to AlP were 28, 1
and 4, respectively. The pretreated copper test plate was
dipped into the coating solution for 30 s and subsequently
withdrawn vertically at a constant speed of 2.5 mm/s. The
coatings obtained were kept at room temperature for 30 min
and then dried in an oven at 60C for 24 h. Stearic acid was
used as a water-repellent agent for surface modification [25].
The as-prepared coatings were dipped into 10 mmol/L stea-
ric acid/hexane solution for 24 h and then washed thor-
oughly with hexane. After drying in an oven at 60C, the
superhydrophobic coatings obtained were translucent.
The florescence image of a water droplet (V=4.65 L) on
the coating under room temperature (T=21.8C) was shown
in Figure 1. Water droplets were placed at seven positions
and the average contact angle value was taken as the final
value, which is 152.6.
2.3 Experimental setup
Figure 2 shows the schematic of the experimental setup
Figure 2 (Color online) Experimental setup.
October (2013) Vol. 56 No. 10 3
used in the present study. A droplet generator was used to
generate water droplets to impinge onto a red copper test
plate with superhydrophobic coating. The water droplets
contained a mixture of Rhodamine B (5106 mol/L) and DI
water. The temperature of the test plate, which was moni-
tored by using a thermocouple, was kept constant at a
pre-selected low temperature level by using a Constant
Temperature Bath Circulator (AC 150-A25, Thermo Sci-
entific). A laser sheet (~600 m in thickness) from a pulsed
Nd:Yag laser at a wavelength of 532 nm was used to illu-
minate the Rhodamine B molecules along the middle plane
of the water droplets. A filter (570FG05-50, Maxlevy) was
placed in the front of a CCD camera (Sensicam Qe, Cooke
Corp.) so that only fluorescence emissions could be cap-
tured by the camera. The camera and the pulsed Nd:Yag
laser were connected to a workstation via a digital delay
generator (BNC 575-8C), which controlled the timing of the
Figure 1 Florescence image of a water droplet on the coating under room
temperature.
4 Jin Z Y, et al. Sci China-Phys Mech Astron
laser illumination and the image acquisition. It is noted that,
due to the low concentration of the tracer molecules used
for the measurements, the effects of the molecular tracers on
the physical properties of water and icing process are be-
lieved to be very small. Since the low energy level and rep-
etition rate of the pulse laser excitations are used for the
fluorescence measurements, the effects of the energy depo-
sition due to the pulsed laser excitation on the icing process
is expected to be negligible.
3 Results and discussions
During the experiment, the relative humidity, temperature,
and pressure of the surrounding air were kept at 80.2%,
21.2C, and 101005 Pa, respectively. The upper surface of
the test plate was covered by a plexi-glass cover in order to
avoid the condensation of water vapor on the surface. Once
the test plate temperature reached a desired steady value, the
cover was removed and a water droplet was released from
the droplet generator. The arrival speed of the droplet right
before it impacted the surface was estimated to be 0.66 m/s.
In general, the droplet stayed at the supercooled state for a
Figure 3 (Color online) Fluorescence images of droplet icing process under free convection.
October (2013) Vol. 56 No. 10
period of time and then started to freeze.
Figure 3 shows the time sequence of the acquired fluo-
rescence images of a freezing water droplet (V=4.79 L)
after it impinged onto a frozen cold superhydrophobic test
plate (Tw= 8.2C). The contact angle of the droplet is
about 115.8, which is less than that of the droplet under
room temperature. Such a difference might be due to the
impingement effect and the thin water film formed on the
surface. The freezing process of the water droplet was under
free convection. Initially, the bulk of droplet was liquid wa-
ter and kept in a supercooled state for about 500 s, as shown
in Figure 3(a). Once the freezing process began, a darker
region appeared at the bottom of the droplet, as shown in
Figure 3(b). Such a darker region at the bottom indicates
solid phaseice; while the brighter region in the upper
portion of the droplet represents liquid phasewater. Inter-
estingly, air bubbles were found to be nucleated and trapped
in the solid ice, as shown in Figure 3(d). As time went by,
the interface between the liquid water and solid ice was
found to move upward gradually. In addition, the leftmost
point and rightmost point of the interface were almost on
the same height relative to the test plate during this process.
About 18 s later, the droplet was found to completely turn
 Jin Z Y, et al. Sci China-Phys Mech Astron October (2013) Vol. 56 No. 10 5

Figure 4 (Color online) Fluorescence images of the frost formation and ice crystal melting processes.
6 Jin Z Y, et al. Sci China-Phys Mech Astron October (2013) Vol. 56 No. 10

into a solid ice crystal (as seen in Figure 3(i)). It should be

noted that the droplet was initially shaped like part of a
sphere. Once it was frozen, it looked more like a cone and
seemed to be symmetrical with respect to its centerline.
Once the freezing process was completed, the constant
temperature bath circulator was kept running for another 40
s and then shut down. The temperature of the test plate in-
creased as time went by, while the ice crystal first under-
went the frost formation process and then the melting pro-
cess. The whole process, from the time when the bath cir-
culator was shut down to the time when the melting process
was finished, was recorded by fluorescence images as
shown in Figure 4. At the beginning, the vapor in the moist
air de-sublimated (vapor-to-ice transformation) onto the Figure 5 (Color online) Volume of frozen ice vs. time.
surface of the ice crystal and frost was formed. However,
apparent frost could only be found on the top of the ice
crystal, while the rest of the ice crystal surface remained
nearly unchanged (Figure 4(a)). The frosts grew with time
in the direction normal to the surface of the test plate (Fig-
ures 4(b)(d)). Then, the frosts started to melt (Figures
4(e)(g)), followed by the protrusion part of the ice crystal
(Figure 4(h)). The melted water slid down and started to
merge with the water film formed on the test plate due to
the condensation process (Figure 4(i)). Such process would
cause a bottom region of the droplet not recordable with the
camera (Figure 4(k)). It is of interest to mention that a sol-
id-liquid interface, similar to that shown in Figure 4, ap-
peared at the bottom of the ice crystal (Figure 4(l)). How-
ever, in this case, liquid water stayed below the interface
due to its higher density than that of the ice. Besides, air Figure 6 (Color online) Freezing time of water droplet vs. test plate
temperature.
bubbles which were locked inside the ice crystal gradually
appeared and then disappeared. As time advanced, the in-
terface between the liquid water and the solid ice was found cooled down further (e.g., Tw=11.2C).
to gradually move upward until the whole melting process
was finished (Figures 4(m)(o)).
Basing on the above fluorescence images (as seen in 4 Conclusions
Figure 3), the change of the volume of frozen ice with time
during the droplet freezing process was obtained and plotted In summary, we reported an icing physics study to visualize
in Figure 5. At the beginning, liquid water turned into ice at the freezing and melting process of a small water droplet on
a very fast rate. Such a rapid phase transition process de- a superhydrophobic cold surface by using the LIF technique.
creased gradually with time and the volume of frozen ice The superhydrophobic surface was prepared with a sol-gel
approached a constant value at the end of the icing process. method on a red copper test plate. It was found that, during
The freezing time, which is defined as the time interval the freezing process, the liquid water at the bottom of the
between the moment when a dark region was formed inside droplet was frozen and rapidly turned to solid ice, while the
a water droplet and the moment when the water droplet was upper portion of the droplet was still in liquid state. As time
found to turn completely to an ice crystal, can also be de- went by, the interface between the liquid water and the solid
termined based on the time sequences of the measurements ice gradually moved upward and at last an ice crystal with a
as those reported above. Figure 6 shows the variations of cusp-like top was formed. After the shutdown of the bath
the freezing time of the water droplets with the temperature circulator, frost appeared on the top of the ice crystal and
of test plate changing from 11.2C to 5.3C. As was ex- continued to grow for a short period of time. Then the frosts
pected, the freezing time was found to reduce with decreas- started to melt, followed by the protruding portion of the ice
ing test plate temperature. Besides, when the test plate tem- crystal. After that, a solid-liquid interface was formed inside
perature is relatively high (e.g., Tw=5.3C), the effect of the ice crystal and continued moving upward until the ice
droplet volume on the freezing time is very significant. crystal completely melted. In addition, at the beginning of
Such an effect becomes less apparent as the test plate is the droplet freezing process, liquid water turned into ice at a
 Jin Z Y, et al. Sci China-Phys Mech Astron
very fast rate. Such a phase transition process decreased
gradually with time and the volume of frozen ice ap-
proached a constant value at the end of the icing process. As
was expected, the freezing time was found to reduce with
the decrease of the test plate temperature. Besides, when the
test plate temperature is relatively high, the effect of droplet
volume on the freezing time is very significant. Such an
effect becomes less apparent as the test plate was cooled 10
down further.
Nowadays, superhydrophic surfaces have widely been 11
investigated as a promising method for various anti-icing
and de-icing applications. However, since water droplets 12
eventually freeze on a superhydropobic surface when the
relative humidity of the surrounding air is high, a further
13
understanding of the freezing and melting process of water
droplets on superhydropobic surfaces would be desirable for
engineers and scientists to develop more efficient anti-icing 14
and de-icing methods. The present work provides some ten-
tative insights into the microphysical process related to the
15
icing and melting process of water droplets. The effects of
air humidity on the water droplet freezing and melting pro- 16
cesses are recommended to be investigated in future studies.
Furthermore, active and passive methods to accelerate the 17
melting process of iced crystals on superhydropobic surfac-
es should also be taken into consideration for future study.
18
This work was supported by the Scientific Research Foundation for the 19
Returned Overseas Chinese Scholars, State Education Ministry.
20
1 Hayashi Y, Aoki A, Adachi S, et al. Study of frost properties corre- 21
lating with frost formation types. J Heat Trans, 1977, 99: 239245
2 Carte A E. The freezing of water droplets. Proc Phys Soc Sect B,
1956, 69: 10281037 22
3 Visagie P J. Pressures inside freezing water droplets. J Glaciol, 1969,
8(53): 301309 23
4 Feuillebois A, Lasek A, Creismeas P, et al. Freezing of a subcooled
liquid droplet. J Colloid Interf Sci, 1995, 169: 90102 24
5 Hoke J L. The Interaction Between the Test Plate and Frost Layer
Through Condensate Distribution. Dissertation for Doctoral Degree.
Urbana: University of Illinois at Urbana Champaign, 2000. 10 25
6 Tabakova S, Feuillebois F. On the solidification of a supercooled liq-
uid droplet lying on a surface. J Colloid Interf Sci, 2004, 272:
October (2013) Vol. 56 No. 10 7
225234
7 Hindmarsh J P, Buckley C, Russell A B, et al. Imaging droplet freez-
ing using MRI. Che Engi Sci, 2004, 59: 21132122
8 Wang J, Liu Z, Gou Y, et al. Deformation of freezing water droplets
on a cold copper surface. Sci China Ser E-Technol Sci, 2006, 49:
590600
9 Wu X, Dai W, Xu W, et al. Mesoscale investigation of frost for-
mation on a cold surface. Exp Therm Fluid Sci, 2007, 31: 10431048
Suzuki S, Nakajima A, Yoshida N, et al. Hydrophobicity and freezing
of a water droplet on fluoroalkylsilane coatings with different rough-
nesses. Langmuir, 2007, 23: 86748677
Nitsch K. Thermal analysis study on water freezing and supercooling.
J Therm Anal Calorim, 2009, 95: 1114
Jin Z, Hu H. Quantification of unsteady heat transfer and phase
changing process inside small icing water droplets. Rev Sci Instrum,
2009, 80(5): 054902
Hu H, Jin Z. An icing physics study by using lifetime-based molecu-
lar tagging thermometry technique. Int J Multiphas Flow, 2010, 36:
672681
Kulinich S A, Farzaneh M. How wetting hysteresis influences ice
adhesion strength on superhydrophobic surfaces. Langmuir, 2009, 25:
88548856
Cao L, Jones A K, Sikka V K, et al. Anti-icing superhydrophobic
coatings. Langmuir, 2009, 25: 1244412448
Tourkine P, Maerrer M L, Quere D. Delayed freezing on water repel-
lent materials. Langmuir, 2009, 25: 7214 7216
Mishchenko L, Hatton B, Bahadur V, et al. Design of ice-free
nanostructured surfaces based on repulsion of impacting water drop-
lets. ACS Nano, 2010, 4(12): 76997707
He M, Wang J, Li H, et al. Superhydrophobic film retards frost for-
mation. Soft Matter, 2010, 6: 23962399
Kulinich S A, Farhadi S, Nose K, et al. Superhydrophobic surfaces:
Are they really ice-repellent? Langmuir, 2011, 27: 2529
Jung S, Dorrestijn M, Raps D, et al. Are superhydrophobic surfaces
best for icephobicity? Langmuir, 2011, 27: 30593066
Bahadur V, Mishchenko L, Hatton B, et al. Predictive model for ice
formation on superhydrophobic surfaces. Langmuir, 2011, 27:
1414314150
Alizadeh A, Yamada M, Li R, et al. Dynamics of ice nucleation on
water repellent surfaces. Langmuir, 2012, 28: 31803186
Jung S, Tiwari M K, Doan N V, et al. Mechanism of supercooled
droplet freezing on surfaces. Nat Commun, 2012, 3: 615
Shi M, Xi J, Wang H, et al. A translucent and superhydrophobic sur-
face prepared with a sol-gel method based on alumina nanoparticles.
J Adhes Sci Technol, 2008, 22: 311318
Wu X, Zheng L, Wu D. Fabrication of superhydrophobic surfaces
from microstructured ZnO-based surfaces via a wet-chemical route.
Langmuir, 2005, 21(7): 26652667