AN EXPERIMENTAL CLASSIFICATION OF THE ENCAPSULATED WATER SUPERCOOLING
AN EXPERIMENTAL CLASSIFICATION OF THE ENCAPSULATED WATER
Juan J. Milón G.1, Sergio L. Braga2
Universidad Católica San Pablo. Arequipa, Perú.
Pontificia Universidade Católica do Rio de Janeiro, Rio de Janeiro, Brasil.
An experimental apparatus was developed to investigate the supercooling phenomenon in
water inside cylindrical capsules used for cold storage processes. The coolant, is a water-
alcohol mixture (50% vol.), controlled by a constant-temperature bath. Temperatures varying
within a one-second time are measured inside and outside the capsule. A cylinder with an
internal diameter 45 mm, thickness 1.5 mm and length 170 mm was made in aluminum
material. The water supercooling period and the nucleation temperature were investigated for
different coolant temperatures. The results indicate that different curves are a characteristic, in
cylindrical capsules, in the cooling process of water with supercooling presence and they can
represent a coherent classification of the phenomenon.
KEYWORDS: dendritic ice, nucleation, phase change, supercooled water, thermal storage.
Water is widely used in thermal storage systems as phase-change material (PCM) due to its
following advantages; high value in latent heat, stable chemical properties, low cost and easy
acquisition, no environmental pollution concerns, and good compatibility with the materials
used in air-conditioning equipment. However, there are also a few disadvantages in using
water as PCM. The most serious problem encountered is the supercooling phenomenon,
which occurs during the thermal storage cooling system.
While the water is cooled in an enclosed container, freezing does not occur at its regular
freezing point (0°C) at atmospheric pressure. Instead, it is normally cooled below 0°C before
ice nucleation happens. Supercooled water refers to a state of metastable liquid even though
the water temperature is below its freezing temperature. The metastable state will end when
ice nucleation occurs and the thin plate-like crystal of dendritic ice grows into the region of
supercooled water. During the dendritic ice growth process, latent heat is released from the
dendritic ice and absorbed by the supercooled water. At the end of this initial growth process,
normally the water temperature returns to its freezing point (0°C). If the metastable state
exists and remains during the thermal storage process, thermal energy can only be stored in
the form of sensible heat. In this case, the storage capacity is strongly reduced. There are
several studies about supercooling water process. Chen et al. , studied a numerical and an
experimental methods to analyze the influence of nucleation agents in the water solidification
process inside cylindrical copper capsules of different sizes. Yoon et al. , studied the
freezing phenomenon of saturated water within the supercooled region in a horizontal circular
cylinder using the holographic real time interferometry technique. Milón and Braga ,
studied the supercooling phenomenon in spherical capsules of different diameters and
determined the parameters that influence the phenomenon appearance. Okawa et al ,
studied a freezing of supercooled water on a metallic surface. Gilpin  studied the dendritic
ice formed in a water pipe during the freezing process. He showed that growth of dendritic ice
can cause blockage of the water pipe. The extent of dendritic growth is largely determined by
the temperature distribution existing in the pipe at the time of ice nucleation. Milón and Braga
, studied the dendritic ice formation in the supercooled water enclosed in cylindrical
capsules and mentioned the particularitities of water under these conditions.
A classification of the characteristic curves presented statistically in the different experiments
with water supercooling is not mentioned in the reference, for this reason; the present work is
a guide for future studies.
The experimental model, shown schematically in Figure 1, consists in a test section (a), an
observing system (b), a cooling system, which includes a CTB (Constant Temperature Bath)
(c), and a measurement and data acquisition system (d).
(d) CTB UPPER
DIGITAL RESERVOIR COMPRESSOR
CAMERA (c) (c)
Figure 1. Experimental Model.
This section contains the capsule to be analyzed. The box acrylic walls are 10 mm thick,
externally covered with a 25 mm thick insulation. A cross section of the test section is shown
in the Figure 2. The diffuser is intended to homogenize the coolant temperature (TC) in the
test section. The temperature control is carried out by a constant temperature bath which
receives the signals from a temperature sensor RTD type PT-100. K type thermocouples were
used for the circulating fluid temperatures registration. Two flanges are installed in the
vertical walls to facilitate the capsule installation. An overflow tank, working at atmospheric
pressure, was installed to compensate the volume variation during the PCM-phase-change
process. To define the volume of the PCM, a sliding disk (movable in the axial direction) is
used. K type thermocouples of 0,076 mm diameter, covered with Teflon, are also shown.
RTD PT-100 COOLANT SECONDARY
Figure 2. Test section.
The cooling system is schematically shown in Figure 1. It is composed by two constant
temperature baths (CTB), two reservoirs (upper and lower), and the coolant. The initial
temperature for the tests is set by the CTB-IC while the other CTB is set at the test
temperature. The temperature control system is of the PID type (Proportional Integral
Derivative) and it is able to maintain the temperature within a range of ±0.05 °C with a
refrigeration power of 800 W at 0 ºC and of 1000 W of electric power heating. The coolant
is an alcohol-water solution (50% in volume).
Measuring and Data Logging System
The Measuring and Data Logging system includes the data acquisition system and a personal
computer (PC). The acquisition equipment, which communicates with the PC by the RS232
communication port, receives, processes and transmits the temperature signals to the PC, for
storage and posterior analysis.
The initial water temperature inside the capsule (25.0 ºC) is set using the constant temperature
bath for the initial condition (CTB-IC). The other CTB, controls the coolant temperature in
the upper reservoir. The coolant of the upper reservoir passes through the test section setting
the test temperature, absorbing the initial thermal loads and immediately after getting to the
lower reservoir. The coolant of the CTB is addressed toward the test section until the
conclusion of the test. After each test, a pump impels the coolant toward the upper reservoir
for a new test. This procedure is carried out to maintain constant the test temperature through
each test. The length of each test varies from 30 to 60 minutes, depending on the test
conditions. The data is acquired every second. Uncertainties are presented in Table 1.
Table 1: Uncertainties of measurements
Temperature 0.1 ºC
Time 0.01 s
Length 0.01 mm
RESULTS AND DISCUSSION
Displayed Curves with an occurrence probability of more than 50%
For the classification shown below, it is necessary to define a thermocouple position (Fig 3).
The displayed curves represent most of the observed curves (a probability of more than 50%).
Figure 3. Temperature sensor position.
In the Fig. 4, the freezing process without supercooling is observed. This type of curve, with
stable change phase, normally happens in capsules of high thermal conductive and low
It is observed that the process begins with a PCM temperature with Ti. It immediately
descends releasing liquid sensible heat (a) down to the temperature of phase change and later
releases latent heat (b), finally, the last particle of PCM is ice (region around the
thermocouple) and releases solid sensible heat (e). In this process supercooling does not
a, b d e
a dendritic ice
Figure 4. Freezing process without supercooling, encapsulated water
In fig 5, the PCM is cooled releasing sensible heat (a), down to the density inversion
temperature (Tdi) and goes through the freezing temperature (Tf), immediately after, it keeps
lowering in metastable liquid state (b) below Tf. The metastable state will end when ice
nucleation occurs (T=Tn) and thin-plaque-like crystals of dendritic ice grows in the water
supercooled region (c). During the dendritic ice growth process, latent heat is released from
the dendritic ice and absorbed by the supercooled water. At the end of this dendritic growth
process, usually, the water temperature returns to its freezing point (0°C) and phase change
starts (d) later, solid sensible heat is released (e). This curve type appears commonly for
different capsule materials with coolant temperatures, relatively low.
a, b c d e
a dendritic ice
Figure 5. Supercooling and freezing process, encapsulated water
In Fig 6, is observed a supercooling process and instantaneous freezing, after the PCM is
supercooled, and when nucleation occurs (c), the energy level is so low in that area that the
PCM reaches Tf , and the phase change occurs quickly and immediately after it releases solid
sensible heat (e). This happens with high-thermal-conductivity capsules and coolant at low
In Fig 7, the hypercooling with an instantaneous phase change process can be observed.
Unlike the previous curve, when nucleation happens, energy is such, that the phase change
occurs at temperatures below Tf (d) and it immediately releases sensible heat (e).
a, b c e
a Liquid water
b c e
Figure 6. Supercooling and Instantaneous freezing process, encapsulated water.
a, b c d e
Tn c e
Figure 7. Hypercooling and instantaneous freezing process, encapsulated water
In fig 8, the permanent supercooling and non-freezing process is observed. This occurs when
the metastable liquid state (b) remains for an undetermined time and the PCM only releases
liquid sensible heat. This curve is common in high and low thermal conductivity capsules
with high coolant temperatures (near Tf)
a, b Thermocouple position
Figure 8. Permanent supercooling and non-freezing process, encapsulated water.
Special curves observed sporadically
The curves described in the Figs. 9 and 10 are special cases which appeared rarely in the tests,
but due to their particular characteristics, are here described.
Fig. 9, shows the curve for double supercooling, when the PCM presents an apparent
supercooling (a'), the PCM, returns to Tf, after a brief period of time. This probably occurs by
convective-cells generation that produces a thermal equilibrium in the region. Then the PCM
presents a higher degree of supercooling (b) and nucleation occurs (b'), and dendritic ice
appears (c). Finally, the phase change process continues (d) and the sensible heat is released
a, b c d e
Tdi dendritic ice
Tn a’ b c
Figure. 9. Special freezing process, double supercooling, encapsulated water.
In fig. 10 another particular case is observed. When the PCM is supercooled (b), nucleation
occurs and the PCM temperature returns above Tf. Later in a considerable amount of time it,
returns to Tf and quickly occurs the phase change (d) finally, sensible heat is released (e).
a, b c d e
a Thermocouple position
Tdi dendritic ice
Tf d e
Tn b c
0 200 300 400 500 600
Figure 10. Special freezing process, supercooling and nucleation above Tf, encapsulated water
In regard to the tests performed under controlled conditions, a classification of encapsulated
water cooling processes was made. Five characteristic curves appeared (with a probability of
more than 50% in 20 tests under the same experimental conditions).
Two curves appeared seldom in the tests. These were shown and analyzed
Ti initial temperature, [ºC]
Tdi inversion density temperature
Tf freezing temperature, [ºC]
Tn nucleation temperature
TC coolant temperature, [ºC]
PCM phase change material
CTB constant temperature bath
CTB-IC constant temperature bath to initial condition of the PCM
This paper was supported by the Conselho Nacional de Desenvolvimento Científico e
Tecnológico – CNPq of Brazil.
The authors also wish to thank the Agreement Pontificia Universidade Católica do Rio de
Janeiro and Universidad Católica San Pablo for motivating this research work.
 Chen, S. L., Lee, T. Z., 1998, A study of Supercooling Phenomenon and Freezing
Probability of Water Inside Horizontal Cylinders, International Journal of Heat and Mass
Transfer 41: (4-5, 3), 769-783.
 Yoon, J.J., Moon, C.G., E.Kim, Son, Y. S., Kim, J.D., Kato, T, 2001, Experimental Study
on Freezing of Water with Supercooled Region in a Horizontal Cylinder, Applied Thermal
Engineering 21Ç 657-668.
 Milón, J.J., Braga, S.L., 2003, Supercooling Water in Cylindrical capsules, Fifteenth
Symposium on Thermophysical Properties, National Institute of Standards and
Technology, Colorado, USA.
 Okawa, S., Saito, A., Minami, R., 2001,The Solidification Phenomenon of the
Supercooled Water Containing Solid Particles. International Journal of Refrigeration 24
 Gilpin, R.R., 1977, The Effect of Cooling Rate on the Formation of Dendritic Ice in a Pipe
whith no Main Flow. ASME Journal of Heat Transfer 99: 419-424
 Milón, J.J., Braga, S.L., 2004, Dendritic Ice Growth in Supercooled Water Inside
Cylindrical Capsule, International Refrigeration and Air Conditioning Conference at
Purdue, Colorado, USA.
The authors affirm that the paper has not been published elsewhere and, if the paper is
accepted, will not be published elsewhere prior to Clima2005.
If the paper is accepted, the authors will grant to Clima20005 a nonexclusive license to
publish the paper in any fashion.