Simultaneous thawing and freezing can be demonstrated by a simple experiment that can be performed as a parlor trick. Float an ice cube in a glass of water. Have a shaker of salt within arm's reach. Hand someone a paper match and challenge him to remove the ice cube from the water without lifting it with any implement other than the match. When he gives up, bend the head of the match to a right angle, place the body of the match flat on top of the ice and cover it with a thin layer of salt.

The match will promptly freeze to the cube. Lift the cube from the glass by the head of the match.

Inspection will disclose that the salt melted some ice all around the edge of the match. Ions of sodium and chlorine gained freedom of motion when they dissolved in the film of water on the surface of the cube, and motion was induced by heat drawn from the ice. Replacement energy then flowed from the region under the match that was protected from the salt. As a result the temperature of the film of fresh water in contact with the lower surface of the match dropped below its freezing point and turned into ice that cemented the cube to the match.

Common salt can similarly lower the temperature of water a degree or two, even in the absence of ice. Other salts are more effective. A good freezing solution can be made by dissolving in fresh water sodium thiosulfate (the photographer's "hypo"). The temperature of the solution can drop as much as 25 degrees Fahrenheit. To demonstrate the effect wet the bottom of a thin glass container with fresh water, stand the container on a base of wood or some other material that is a poor conductor of heat and fill the container with equal weights of water and hypo. Stir the mixture with a wooden stick for a minute or two. The wet bottom will freeze and the glass will stick to the base.

A stoppered vial of fresh water placed in the water-hypo solution will also freeze. If the vial is full, the expanding ice may exert enough pressure to break the glass, fresh water having the unusual property of being denser as a liquid than as a solid.

Recently Edward M. Little, who is a senior scientist with the Arctic Sciences Division of the Naval Undersea Warfare Center at San Diego, Calif., measured the pressure exerted by both freshwater ice and seawater ice. He found that the presence of salts in seawater reduced the pressure exerted by sea ice to a negligible amount, whereas freshwater ice developed pressures of more than 200 pounds per square inch. Little describes his work as follows:

"The structure of sea ice includes 'brine cells' that prevent high pressures from developing when the water solidifies. Sea ice is about as permeable to other substances as sandstone. (We are interested here in permeability, not porosity; a substance may be quite porous, but unless the holes in it are connected so that a liquid can flow through the material it is not permeable.) Sea ice turns out to be much more permeable than freshwater ice-probably at least 100 times more permeable.

"I made measurements of pressure by freezing water in a two-pound coffee can that was equipped with a strain gauge. A simple strain gauge consists of a wire .001 inch in diameter bent into a zigzag grid; the grid is cemented to bond paper and covered with felt for protection. The gauge is fitted with a pair of larger wires that serve as connections to the grid. Opposing forces applied to the sides of the gauge put the paper backing and the wires in tension. If the force is sufficiently large, the wire stretches and its electrical resistance increases. The increase in resistance is proportional to the strain, which is defined as a fractional increase in dimension, such as in length.

"Commercial strain gauges usually include more than one wire grid. The commercial gauge I used is of the rosette type: two grids, separated by an insulator, are bonded to the paper backing at right angles to each other, and a third grid, similarly insulated and bonded, is set at an angle of 45 degrees to the other two [see Figure 1]. I used only the two grids set at right angles to measure stresses in the can. The grids are labeled A and C in the illustration.

"The gauge was positioned so that the wires of grid A were parallel to the axis of the can. The wires of grid C extended in the direction of the can's circumference. The difference between the strains of the two grids provided a measure of the pressure in the can. The grids were connected in a circuit as opposite arms of a Wheatstone bridge. For this reason changes in resistance caused by variations of room temperature canceled each other, thus providing automatic temperature compensation. Grid C functioned as the active gauge, because it always had twice the strain of grid A.

"Strain gauges are often made of constantan wire, which consists of 60 percent copper and 40 percent nickel. The electrical resistance of constantan is relatively independent of changes in temperature. On the other hand, the resistance of constantan does change somewhat when the wire is stretched. The wire is said to have a 'gauge factor' of two. The phrase means that the percent change in the wire's electrical resistance corresponds to twice the percent change in length.

"Another popular material for strain gauges is isoelastic wire, an alloy of iron, nickel, chromium and other metals. It has a large temperature coefficient and a gauge factor of 3.5. I recommend that amateurs buy the Type AR-1 strain gauge for this experiment, because the substance to be studied must be cooled far below room temperature. The resistance of this gauge at room temperature is about 120 ohms; the gauge factor is two. The unit is manufactured by Baldwin BLH Electronics, Inc., 42 Fourth Avenue, Waltham, Mass. 02154, and is also available from distributors of scientific supplies.

"The can should be pre-bulged by freezing fresh water in it at less than 20 degrees centigrade. The circumference and thickness of the metal are then measured as accurately as possible. Bond the paper side of the strain gauge to the can with plastic cement and connect the leads into a Wheatstone bridge. The arrangement is depicted in the accompanying illustration [Figure 2 ]. I used Manganin resistors as the two fixed arms of the bridge, but if resistors of this type are not available, any precision units of the metal-glaze variety such as the IRC Type RG 20 can be substituted. The resistors need not be capable of dissipating more than 1/4 watt.

"I measured the output of the bridge by means of an amplifier and automatic pen recorder connected as shown in the illustration, but a simple indicating meter will work as well at some cost in convenience. The meter should have an internal resistance of about 1,000 ohms and a sensitivity of .125 microampere per scale division. (It should be a Leeds & Northrup Type 2310-d galvanometer or the equivalent.) In other words, the resistance of the meter should be large with respect to the resistance of the bridge circuit. Note that the bridge includes a one-ohm slide-wire potentiometer. A good wire to use is Welch No. 2809. The sliding contact of this improvised potentiometer can be an alligator clip. A protective shunt should be connected across the meter and through a switch for increasing the scale to 0-1 milliampere when the battery is switched on and the bridge is unbalanced.

"To measure the pressure exerted by ice, fill the can with water to within an inch of the top, connect the gauge to the bridge circuit, balance the bridge with the one-ohm potentiometer so that the meter reads 0, place the can in the freezer and close the battery switch. Make a table of two columns, one for time and the other for meter readings. Readings should be taken at regular intervals. At a freezing temperature of-12.5 degrees C. a complete test run will take two full days; at-27 degrees the run will last about six hours.

"All that remains is to calibrate the meter in terms of pressure, a simple calculation that takes into account the size of the can, the elasticity of the metal and the voltage of the battery that energizes the bridge. The current I_g that is indicated by the meter in microamperes is equal to the excess pressure p developed in the can, multiplied by the radius r of the can in centimeters and by the battery voltage V_b, divided by twice the elasticity E of the metal of which the can is made, multiplied by the thickness of the metal t in centimeters and by the internal electrical resistance of the meter R_g. The result is equal to the pressure (in atmospheres) per microampere of current, as indicated by the meter. Expressed symbolically, the relation is I_g = prV_b/(2EtR_g).

"The can I used was 12.4 centimeters in diameter, so that its radius was 6.2 centimeters. The thickness of the metal was .0109 inch, or .0277 centimeter. A reasonable value for the elasticity of steel of the kind used in most tin cans is about 2 x 10¹² dynes per square centimeter. I used a six-volt battery. One atmosphere of excess pressure in the can is equal to 106 dynes per square centimeter. The internal resistance of my meter is 1,000 ohms. The calibration is completed by substituting these quantities for the symbols in the formula and doing the arithmetic: I_g = 10⁶ X 6.2 x 6/(2 x 2 x 10^l2 x .0277 x 1,000) = 3.3 x 10^-7 amperes, or .33 microampere per atmosphere of pressure. Apparatus that differs in size from my setup can be similarly calibrated by substituting appropriate quantities in the formula.

"Although an indicating meter is adequate for the experiments, I used an automatic pen recorder as a convenience. The instrument was calibrated in terms of relative strain instead of pressure. Six microamperes indicated a strain of .001 inch per inch, or .1 percent. Relative strain was plotted against time for each of two experiments, one run at a freezing temperature of -12.5 degrees C. and the other at -27 degrees. Pressure corresponding to the indicated strain was then calculated.

"When making the graphs, I subtracted .0003 from the strain readings of the two sea-ice recordings so that the plots for freshwater ice and seawater ice would not overlap. The strain gauge used for the experiment at -12.5 degrees C. had a gauge factor of 3.5. The output of the gauge drove the pen of the recorder off scale. The instrument included a compensating dial that I set at 2. The setting in effect multiplied the output by 2/3.5, or a factor of .58. The relative increase in strain due to the freezing of fresh water at -12.5 degrees C. turned out to be .00593 - .0056, or .00033, which is .033 percent. In the case of this experiment p equals 1.787 times 10¹⁰ times .58, or 3.41 x 10⁶ dynes per square centimeter, which is equivalent to about three atmospheres [see Figure 4].

"I was puzzled at first by the graph of freshwater ice frozen at -27 degrees C. [see Figure 5 ]. It has many sharp peaks that indicate abrupt changes in pressure. The explanation became clear when the ice was melted from the can. The soldered joint along the seam of the can had been torn apart. Evidently when the solder was stressed by the expanding ice, it gave way in a series of breaks. It is possible to estimate the maximum pressure that would have developed in an unyielding can by adding the individual peaks of the graph. The sum indicates a maximum pressure of about 15 atmospheres, or 235 pounds per square inch.

"In spite of the experimental uncertainty, the graphs show clearly that seawater does not behave like fresh water when it freezes. The difference stems mainly from the higher permeability of seawater ice. Incidentally, chilled seawater in sealed cans makes an effective and convenient refrigerant for preserving foods in picnic ice chests. The leftover seawater surges harmlessly through the already frozen sea ice instead of distorting the containers. To provide room for the expansion the cans should not be filled to more than 80 percent of their capacity."

James Bailey of Milwaukee also experiments with saline fluids, but instead of crystallizing water by freezing he applies controlled heat to the solution for growing large single crystals of the salt. Bailey writes:

"Generations of amateurs have made a hobby of growing large single crystals of various salts such as copper sulfate, alum, hypo and similar chemicals. Traditionally two techniques have been used. One involves cooling, the other evaporation.

"Both begin with the preparation of a saturated solution. Salt is mixed with water until no more will dissolve. The solution is filtered, preferably through a sheet of permeable paper made specially for the purpose. Filtering removes minute crystals that could otherwise remain in solution and grow at the expense of the desired single crystal.

"A single crystal is always grown on a seed of some kind-a speck of foreign matter, a minute crystal of the salt or even a cluster of molecules. The seed is suspended by a thread and lowered into a bath of distilled water for a few seconds, just long enough to dissolve any microscopic crystals that may have collected on its surface. It is then transferred to the filtered solution. The concentration of the filtered solution is increased by either cooling or evaporation. When the temperature of a saturated solution is thus lowered, the water becomes supersaturated. Excess salt must leave the solution and reappear in the form of one or more crystals.

"To grow crystals by the cooling technique the saturated solution can be made at a temperature higher than the room temperature. A prepared seed is lowered into the bath and the heat is slowly reduced. Alternatively the cooling can be achieved by putting a room-temperature solution in a refrigerator.

"Supersaturation can also be induced by letting the water evaporate. The technique is slow but popular because of its simplicity. Weeks or even months may be required to grow large single crystals by evaporation, depending on the nature of the salt. During this interval the solution may be contaminated accidentally by foreign particles from the air or other sources.

"I have been experimenting with a third technique, known as the hydrothermal process. It provides positive control over the rate of crystal growth and protects the solution from dust and other airborne contamination. Manufacturers have used the method extensively in recent years to grow crystals of quartz. The equipment consists of a source of heat, an insulated vessel that holds the solution, an excess of the chemical to be crystallized and a thermostat for maintaining the solution at a constant average temperature.

"If the apparatus is operated in a room that is maintained at about 70 degrees F., heat must be provided to the solution at the rate of about 80 watts per liter. An introductory apparatus can consist of a 400-milliliter beaker heated by a 40-watt incandescent lamp [see Figure 6]. The beaker must be wrapped with a sheet of urethane foam (or an equivalent insulating material) that covers 90 percent of the solution. The top of the beaker can be closed by a sheet of transparent plastic that contains two holes-one for admitting the seed crystal to the solution and the other for supporting a glass-enclosed thermistor (Keystone RL 20 EIT, Keystone Carbon Co., Thermistor Division, St. Marys, Pa. 15857, or equivalent). Do not use a vessel of thick glass that might break when heated.

"The thermistor acts as a temperature-sensitive resistor in a circuit that automatically adjusts the power to the incandescent lamp as necessary to maintain the solution at a constant average temperature. A Zener diode applies a constant voltage to a string of resistors that includes the thermistor. The voltage that appears across the thermistor and the 2,200-ohm fixed resistor of the string varies with the temperature of the solution. This voltage is applied to the emitter of a unijunction transistor, which in turn energizes the gate electrode of a silicon-controlled rectifier. The silicon-controlled rectifier functions as a throttle for applying power to the incandescent lamp.

"When the temperature of the solution, as sensed by the thermistor, increases or decreases beyond predetermined limits, the silicon-controlled diode adjusts the energy of the lamp just enough to compensate for the change, thereby maintaining the solution at a constant temperature. Conversely, above a predetermined temperature the silicon-controlled rectifier switches the lamp off. The temperature limits within which the lamp is thus controlled can be raised or lowered by adjusting the 2,500-ohm rheostat connected in series with the thermistor.

"The C-106B1 silicon-controlled rectifier is designed for a maximum current of two amperes and will switch incandescent lamps rated up to 150 watts. When lamps larger than 40 watts are used, a two-watt resistor of 10,000 ohms should be substituted for the one-watt unit specified in the accompanying illustration [above]. All other resistors are rated at one watt.

"Silicon-controlled rectifiers of any type can be substituted for the specified unit if more or less switching capacity is desired. The silicon-controlled rectifier must be equipped with a metal heat sink. The heat sink operates at the potential of the power line and must therefore be enclosed by and insulated from a housing. The components of my unit were assembled in a small plastic box.

"The insulated beaker is heated by a hot plate. I use an inverted coffee can, with a hole in the bottom, that houses the lamp bulb and socket. When setting up the apparatus, I usually prepare the concentrated solution in a separate vessel at a temperature of about 110 degrees F., slightly above the growing temperature. A layer of salt about one centimeter thick is placed in the growing vessel. The saturated solution is filtered on top of the chemical. The growing vessel is placed on the hot plate and covered. The warmed solution rises by convection and cools as it streams upward and returns along the walls of the container.

"During the excursion the solution reaches saturation as it approaches the top of the vessel and becomes supersaturated at some point during its return along the walls. The seed crystal is suspended for growth in the region of supersaturation, which must be located by experiment. I usually apply heat to freshly prepared solutions for about five hours before introducing the seed crystal. During this interval the circulation and the region of supersaturation become stable.

"A thin layer of fresh water also forms on the surface as the result of condensation that collects on the lower surface of the cover and drips back into the container. This layer is useful because it dissolves from the seed any microscopic crystals that may have formed after the seed was cleaned. I use nylon thread for suspending seed crystals.

"When possible, I buy salts in crystalline form and select the most perfectly formed crystals for use as seeds. This is a painstaking operation that often requires the use of a magnifying glass. Some salts are sold in powdered form. These are dissolved and crystallized by evaporation. Seeds are then selected from the resulting mass of small crystals.

"A perfectly formed crystal for use as a seed can be grown by suspending any fragment of a broken crystal in a concentrated solution of the salt. Use a thread to suspend the crystal. The container must be airtight and must also be maintained at a constant temperature to prevent any change in the concentration of the solution. The conversion of the fragment into a perfect crystal may require days or weeks, depending on the nature of the salt.

"Harvesting mature crystals is easy. They are simply pulled from the growing solution and wiped dry with a soft cloth. Some crystals are sensitive to abrupt changes in temperature and may crack if they are pulled from the warm bath when the air in the room is abnormally cold. They can be harvested safely by first cooling the solution to room temperature. Incidentally, I have been trying without much success to grow large crystals of calcium carbonate and should appreciate any tips that fellow enthusiasts might pass along. Indeed, I should be happy to exchange information on any aspect of the hobby. My address is 5606 South 34th Street, Milwaukee, Wis. 53221."

Bibliography

EXPERIMENTAL INVESTIGATIONS ON THE ICE-FORMING ABILITY OF VARIOUS CHEMICAL SUBSTANCES. Norihiko Fukuta in Journal of Meteorology, Vol. 15, No. 1, pages 17-26; February, 1958.

The Society for Amateur Scientists (SAS) is a nonprofit research and educational organization dedicated to helping people enrich their lives by following their passion to take part in scientific adventures of all kinds.