PPT Ice Spike Poster

1
An investigation of the effects of water purity
on ice spike formation. Jacob M. CanfieldAmes
High School, Ames, Iowa
Materials and methods Quantitative measurements
on ice spikes requires only simple equipment.
Water for the ice was prepared by mixing
distilled water, from the store, with fixed
amounts of tap water using measuring cups. An
ice cube tray, a ruler, a thermometer, and an
electric fan were also used. I cycled one or two
trays of ice cubes a day for several months,
giving me over 100 sets of data, each containing
information about 14 ice cubes. This large
amount of data allowed for relatively low bars of
error, and more reliable statistics.
Introduction Most people dont know about ice
spikes. Rarely found in nature, ice spikes are
sharp, jutting formations which grow out of ice
cubes when the cubes are made with pure water,
under the right conditions. I first learned of
them when I read a report by K. G. Libbrecht and
K. Lui, who, did experiments at Caltech
determining the optimum growing temperature for
ice spikes, discovered the importance of using a
fan in their growth, and also did preliminary
experiments on the effect of impurities on spike
formation. My hypothesis for my research was if
impurities, in the form of tap water, are added
to purified water in controlled amounts, then the
probability of spike formation will drop.
Results (continued)
Analysis (continued) The plot of ice spike
probability versus tap water concentration shows
a dramatic decrease in the likelihood of spike
formation as more and more tap water is added to
the distilled water. At 25 tap water the
probability of spike formation has dropped to
only 13, as opposed to the almost 50 seen when
no tap water is added. The average spike length
remained essentially constant at approximately
2.3 cm through 6 tap water, but started dropping
significantly for 12 and 25. These results are
consistent with never seeing ice spikes in a
lifetime of making ice cubes with tap water.
I.e. for 100 tap water, a data point of 0
probability with very small standard deviation
could be added to the plot.
Graph A graph from the paper by K. G. Libbrecht
and K. Lui on laboratory-grown ice spikes. They
determined that -8 C was the optimum growth
temperature, and found that using a fan greatly
increased the probability and length of
spikes. Inlay Picture of some of my ice spikes.
Ice spike placement in tray was noted using a
numbering system that allowed for quick and easy
assessment of placement in the ice cube tray.
Spike length was always measured as well.
Conclusions As hypothesized, the effect of tap
water, with its relatively small level of
impurities, on the formation of ice spikes is
drastic, reducing both the probability of spike
formation and the average spike length.
Something in the tap water obviously is hindering
spike growth. The effect that tap water has upon
the formation of the spikes leads to some
questions. In particular, which impurity or
impurities in the tap water are causing such a
drastic change in the probability of ice spike
formation? The water in Ames has scores of
chemicals (in low levels) in it. Its unclear
which ones are the salient ones in hindering ice
spike formation. Future studies that test
specific chemicals seperately would have to be
done to answer this question.
How does this work? The Bally-Dorsey model for
ice spike formation states that, assuming the
environmental temperature is sufficiently cold
(in this case around -10 C), ice first forms on
the edges of the container (in this case the ice
cube tray). Ice then grows in from the edges
until only a small hole on the upper surface
remains unfrozen. If the freezing rate is high
enough, the water, expanding as it freezes, will
be forced through this hole. The water pushed
through the hole will then freeze around the
edges and form a tube of ice. Water will be
forced through this tube and freeze on top,
lengthening the tube which ultimately becomes the
ice spike.
A table like this was constructed for each set
of data associated with different percentages of
tap water mixed into the distilled water. From
these data sets I was able to calculate the
probability of a spike forming in any given cube.
In this case the probability was 34. In
addition, I calculated the average length of the
spikes that did form. In this case the average
length was 2.4 cm.
The ice cube tray, ruler, fan, and thermometer
used in obtaining measurements. All of these
(except for the fan) are simple equipment that
can be found in most households. The fan was
bought at Radio Shack.
Fan placement in freezer. The fan was positioned
directly in front of the tray, to ensure maximum
air flow.
Analysis Error bars were found using the formula
for standard deviation, s
Based on the work of Libbrecht and Lui, I first
determined the temperature for best spike
formation. In agreement with their work, I found
that a temperature of about -10 C (a freezer
setting of 11/2) gave the highest probability of
spike formation. Early in the study I noted
that, although our freezer has a fan, it would
turn on and off erratically, giving rise to
poorly controlled growth conditions. To address
this problem I used an electric fan, which was
placed at the edge of the ice cube tray and left
on throughout each freezing cycle.
Literature cited K. G. Libbrecht and K. Lui.
2003. An Investigation of Laboratory-Grown Ice
Spikes.
Where xi is the number of spikes formed on a
given tray, N is the total number of trays frozen
for a given data set, and x (bar) is the average
number of spikes per tray for a given data set.
The final graph showing spike probability based
on percent tap water is shown below (error bars
of s in red).
Acknowledgments I thank my family for letting me
hog the ice cube tray with strange formations for
months on end, and my father for helping me to
collect and assemble the data. Thanks are also
due to Libbrecht and Lui for their inspirational
paper, and the fact that its freely available
over the internet.
Results After each ice tray was removed from the
freezer, the lengths and position of spikes in
the tray were recorded in my lab notebook. After
each data set was complete (between 9-14 trays
per dilution level) I would make a table like the
one seen in the next column.
The Bally-Dorsey model for ice spike formation.
The water forms a tube of ice, is pushed through
the tube, and then freezes on top, lengthening
the tube. This is much like the formation of a
volcano.
Examples of ice spikes grown during this
experiment.
One of the variables being measured was the
position of spikes in the ice cube tray. I
tallied up the data to see if any one position in
the tray had a higher probability of forming
spikes than another. The results showed that no
one position in the tray had a significantly
higher or lower likelihood of forming an ice
spike. E.g. there was no increased probability
closer to the fan than farther away from it, or
on the right or left side of the tray.
For further information http//www.its.caltech.edu
/atomic/snowcrystals/icespikes/icespikes.htm
This site has both the paper and other links for
more info.