review the Mpemba effect, where initially hot water freezes faster than
initially cold water.
the effect appears impossible at first sight, it has been seen in numerous
experiments, was reported on by Aristotle, Francis Bacon, and Descartes, and
has been well-known as folklore around the world. It has a rich and fascinating
history, which culminates in the dramatic story of the secondary school
student, Erasto Mpemba, who reintroduced the effect to the twentieth century
scientific community. The phenomenon, while simple to describe, is deceptively
complex, and illustrates numerous important issues about the scientific method:
the role of scepticism in scientific inquiry, the influence of theory on
experiment and observation, the need for precision in the statement of a
scientific hypothesis, and the nature of falsifiability. We survey proposed
theoretical mechanisms for the Mpemba effect, and the results of modern experiments
on the phenomenon.
of the observation that hot water pipes are more likely to burst than cold
water pipes are also described.
Firstly I will relate story about a surprising experiment. The
experiment is based on an observation made on a number of occasions that appears
to go against common sense. The observation is that if approximately equal amounts
of a hot and a cold liquid are placed together in a freezer, then the hot
liquid freezes first. This seems to me to be a prime example of what science
educators call a discrepant event (cognitive dissonance/cognitive conflict). A
discrepant event is a happening contrary to our current beliefs. Discrepant
events are said to be useful in enabling learners to reconstruct concepts that
have been imperfectly understood. The literature on discrepant events is
comparatively small with the following being the main easily accessible
references (Fensham &Kass, 1988; Hand, 1988; Thompson, 1989).
One of many unsolved mysteries of science, It will remain a mystery for
me to achieve satisfactory and logical results, the physics teacher advised me
to repeat Experiments to prove the results and to make sure it is real. So
basically what I’m trying to do is to explain Mpemba theory and find out is it
truth that the hot water freezes faster than the cold water? I suggest that the
origin of the Mpemba effect (the freezing of hot water before cold) is due to
freezing-point depression by solutes, either gaseous or solid, whose solubility
decreases with increasing temperature so that they are removed when water is
heated. The solutes are concentrated ahead of the freezing front by zone
refining in water that has not been heated, reducing the temperature of the
freezing front, and thereby reducing the temperature gradient and heat flux,
slowing the progress of the freezing front.
• Small freezer with an internal temperature:
to freeze the water
of?19.1° to ?18.8
• Cylindrical aluminium calorimetric vessels
measuring 65 mm in height by 48 mm in diameter
• Electric kettle: to boil the water
• Deionized water: water free from all
charged atoms or molecules, used mainly in manufacture of water-base cleaning
• Digital data logger: To record data over
• Temperature probes: To measure the initial
• Paper towels:
• Cling film: to cover the freeze water
• Masking tape: to cover area on which unwanted
substance not needed
• Mains power supply: to power the small
freezer and Electric kettle.
To obtain varying initial temperatures,
cold deionized water was made up to different depths in six aluminium vessels
(bare and open; bare and sealed on the top with cling film; insulated and open;
or some combination of these). Boiled water from a kettle was used to top the
water up so that the total volume of water in each vessel was 100 ml. Digital
temperature probes were secured with masking tape so that the 8 mm-long head of
each probe was fully submerged at the surface of the water and these were
connected to a data logger that sampled the temperature of each probe at 10 s
intervals. Each vessel was placed on an insulating layer of folded paper towel
to minimize conductive heat loss through the layer of frost on the shelf of the
freezer. A schematic of the experimental setup is presented in figure 1.
Schematic of the experimental setup, the vessels left
to right: open vessel with paper towel wrapping; bare vessel with cling film
covering; bare vessel with open top.
A number of
mechanisms have been hypothesized to explain the Mpemba Effect. Monwhea Jeng 2
and Marek Balážovi? and Boris Tomášik 3 have written excellent
overviews of the subject summarising these hypotheses. One is that the
initially hotter vessel melts the frost layer on which it sits more completely
than the colder vessel does; when this refreezes it creates a better thermal
contact that draws heat away more rapidly.
By placing the
vessels on insulating layers of folded paper towel the possibility of the
efficacy of this hypothesis was immediately eliminated.
Boiling the water
first also reduced the presence of dissolved gases, which had also been claimed
to contribute to the effect.
a liquid remains fluid below its freezing point before spontaneously becoming
solid, has also been proposed as an explanation. James D Brownridge states:
‘Hot water will freeze before cooler water only when the cooler water supercoils,
and then, only if the nucleation temperature of the cooler water is several
degrees lower than that of the hot water.
Heating water may
lower, raise or not change the spontaneous freezing temperature.’ 4
While this may be the case in some circumstances, it is not a satisfying
explanation for the following reasons. Firstly, supercoiling is a
Graph of temperatures for two insulated vessels; the
water in the vessel represented by the blue line starts only 4.45 °C hotter than that of the red line, but begins to
freeze in 15.5% less time.
Graph of time at commencement of freezing against
initial temperature. Red triangles: bare, cling film-covered vessels. Blue
squares: insulated, open vessels; this graph is similar to that presented in Mpemba
and Osborne’s 1969 paper.
Phenomenon was not observed to occur to temperatures below
approximately ?1 °C, if it occurred at all.
Impurities in the water, imperfections in the surface of the inside faces of the
vessel and even the very presence of the head of the temperature probe tended
consistently to cause nucleation and freezing at or very close to 0 °C.
Efforts were made to encourage supercoiling
(the use of deionized water and in some experiments only submerging the tip of
the temperature probe), but supercoiling was still difficult to achieve.
Secondly, Mpemba first observed the effect in ice cream, which is most unlikely
to supercool; ideally we would like a general explanation that also explains
Mpemba’s original observations.
Thirdly, the Mpemba Effect was observed a number of times without supercoiling,
as figure 2 shows.
Evaporation and convection have also been proposed
separately (by Mpemba in his original paper and by Jeng, Balážovi? and Tomášik). This investigation finds evidence to
suggest that the effect is caused by a combination of both of these mechanisms.
Mpemba’s original observation with ice cream
likely used ceramic (insulating) vessels and Osborne’s experiments used Pyrex
vessels, so it was thought that the Mpemba Effect might be related to the
insulation of the vessels in which the water is held. Aluminium vessels were
chosen because they allowed easy adaptation of the vessels for different
experiments. Leaving the vessels bare and adding a cling film covering to the
top to suppress evaporation allowed the cooling behaviour of the water when
radiation was the main mode of heat loss to be investigated. If a wrapping of
paper towel insulation were added around the side and base of the vessel, surface
radiation and evaporation would be the main modes of heat loss. This allowed
the effect of evaporation to be separated and analysed. As predicted, the
Mpemba Effect only occurred in insulated containers, suggesting that it had to
do with surface cooling effects. Figure 3 compares the graph of time at
commencement of freezing against initial temperature for insulated vessels
(blue squares) and cling film-covered vessels (red triangles). The cling
film-covered vessels behave in the intuitive way: the higher the initial
temperature, the longer it takes for the water to begin to freeze. With
insulated open vessels above a certain temperature, any further increases in
temperature cause a decrease in the freezing time. This suggests that
evaporation was a contributing factor to the Mpemba Effect. More data over the
full range of temperatures from freezing to boiling would be useful to further
investigate the shape of this graph. I would like to point out that time is an
essential element of this experiment and the basis for doing so, and all these
trials depend on time.
1 Tanzania, M.
and Tanzania, S. (2017). Physics Education. Tanzania: College of African
Wildlife Management Moshi Tanzania, pp.4, 172.
2 Rod, C.
(2001). HyperPhysics. online Hyperphysics.phy-astr.gsu.edu. Available at:
http://hyperphysics.phy-astr.gsu.edu/hbase/hframe.html Accessed 17 Nov. 2017.
3 Jeng, M.
(2006). Disorder and interactions in low-dimensional systems. Southern Illinois
University Edwardsville, pp.V74, 514.
4 Balážovi?, M. and Tomášik, B. (2012).
Physics Education. pp.V47,P5.
5 Brownridge J
D 2011 When does hot water freeze faster then cold water? A search for Mpemba effect
Am. J. Phys. 79 78
6 J. Walker, “Boiling and the
Leidenfrost Effect,” in Fundamentalsof Physics, by D. Halliday and R.
Resnick (Wiley,New York, 1988, 3rd ed.), p. E10-1-5.