MY GRANDMOTHER MAKES WHAT I regard as the best ice cream in the world. She relies on an old-fashioned apparatus in which the ingredients that will turn into ice cream are churned by means of a hand crank. During and after the churning the mixture is cooled and hardened by a bath of ice, water and rock salt surrounding the container. The result is a creamy, smooth dessert that one cannot resist.

My grandmother's ice-cream maker has three major parts. A metal container holding the ingredients fits into a much larger wood bucket. A hand-cranked dasher extends into the container. Between the container and the bucket are

packed alternating layers of rock salt and crushed ice, roughly four times as much ice as salt. The entire assembly is covered to exclude room heat. My grandmother churns the mixture for about 20 minutes until it becomes quite viscous. Next she repacks the ice and salt, drapes a heavy towel over the apparatus and leaves it for a few hours to freeze hard. The ice cream is then ready for the table.

To investigate the process I have chosen one of my grandmother's recipes for a moderately rich vanilla ice cream. The ingredients for one and a half quarts are listed below. Warm the cream, stirring in the sugar. Continue to heat and stir the mixture, but do not let it boil. Once it is hot and the sugar has fully dissolved cool it to room temperature and then add the vanilla extract. Pour the mixture into the metal container and put it in the refrigerator for at least an hour. Then mount the container in the bucket and insert the dasher, which you have also cooled in the refrigerator. Pack the ice and salt around the container and begin cranking.

Several aspects of the procedure have intrigued me for many years. I realize the mixture cannot be made into ice cream without ice, but is the salt really needed? Why does my grandmother caution against adding too much salt? Why must the ice be crushed? I know that the bucket should be wood to reduce the absorption of heat from the room, but must the container be metal? Do the ingredients of ice cream serve any purpose other than to make it tasty? I have also wondered (particularly during my long turns at the crank) why the ice cream must be churned. (It is certainly well mixed before it is poured into the container.) Why should the churning rate be increased as the viscosity increases? Finally, why is ice cream ruined if you let it melt and then refreeze it?

My grandmother's explanation of the purpose of the rock salt puzzled me. She maintained the salt was intended to lower the temperature of the ice so that the ice cream would freeze faster. Perhaps she was right, but the idea seemed inconsistent with another common use of salt. When a sidewalk must be cleared of ice, a sprinkling of salt over the ice soon melts it. If salt merely lowers the temperature of ice, how can it melt ice on a sidewalk? Conversely, if the salt merely melts the ice in the ice-cream maker, how can it speed up the freezing of the mixture?

To explore the matter I shall first consider the freezing of tap water (ignoring the possibility of supercooling it, that is, making its temperature drop below the normal freezing point without any freezing). In my simple model of cooling I remove energy (heat) from the kinetic energy of the molecules of the liquid. Since the molecules slow down, the water temperature is lowered. When the freezing point (zero degrees Celsius) is reached, ice begins to form at the surface. The removal of more energy freezes more water but cannot lower the temperature until the water is completely frozen.

Picture an intermediate stage where the water is only partially frozen. If I isolate the mixture to prevent any additional loss of energy, both the temperature and the amount of ice remain constant but the molecules still move, some slowly and others quickly. Along the interface of the liquid and the ice some of the molecules are also constantly changing states. Some of the molecules in the liquid phase move slowly enough so that when they collide with the ice, they stick, adding to the ice. Simultaneously some of the molecules in the solid phase (the ice) break free to become part of the liquid. When the system is isolated and the temperature is constant, an equilibrium is reached in which the rate of freezing matches the rate of melting at the interface.

An important factor in this model is the energy exchanged in the freezing and the melting. A molecule in the liquid phase that sticks to ice loses energy for two reasons. First, it must slow down, otherwise its kinetic energy would quickly carry it away from the ice. Second, it loses additional energy when it is captured by the electric forces exerted by neighboring molecules in the ice. When a gram of water freezes, the amount of energy lost collectively by the molecules joining the ice is 80 calories.

The converse is true when a molecule leaves the solid phase. It must receive enough energy to escape the forces exerted by the neighboring molecules in the ice and to have a kinetic energy appropriate to being in the liquid phase. For a gram of water to be melted 80 calories must be supplied in order to free the molecules from the ice. Hence when equilibrium is reached in ice water, the amount of energy given up by the molecules joining the ice matches the energy taken up by the molecules leaving the ice. There is no net exchange of energy across the ice-liquid interface.

Suppose I remove a small amount of energy from a system that is in equilibrium with half of its water frozen. The average speed of the molecules in the liquid phase decreases and the freezing rate increases as the now slower molecules begin to stick to the ice. In this period the freezing rate is higher than the melting rate and there is a net release of energy.

The energy does not disappear. It shows up (through collisions) in the kinetic energy of the molecules that are still in the liquid phase. Before long enough energy is added to them so that fewer of them stick to the ice. Once again the freezing and melting rates match and equilibrium is restored.

Every time I remove a small amount of energy from the system I upset the equilibrium. Molecules slowed by the removal join the ice. After each removal the system reestablishes equilibrium. Eventually if enough energy is removed, all the molecules are in the solid phase. Even in the solid phase, however, the molecules are in motion: they vibrate around their positions in the crystal lattice of the ice. Now each removal of energy slows the vibration, and again the temperature begins to drop.

Reconsider the equilibrium where half of the water is frozen. If you add salt to the liquid, the equilibrium is momentarily. destroyed. Although a water molecule is electrically neutral, it has a strong electric dipole field because its atoms are arrayed in the form of a V. In a simple model of the molecule the oxygen atom, which is at the point of the V, pulls the electrons of the hydrogen atoms toward the point. This movement separates the centers of positive and negative charge in the molecule. The oxygen end can thus be regarded as more negative than the hydrogen ends.

The electric field of the water molecules breaks up the salt crystals into positive sodium ions and negative chloride ions. Each ion is surrounded by a loose cluster of water molecules. Around a hydrated sodium ion (Na+) the water molecules on the average present their oxygen ends to the ion. Around a negative chloride ion (Cl-) the water molecules on the average present their hydrogen ends. At least some of the water molecules in these clusters are held so tightly that they cannot freeze.

Now some of the collisions at the ice-liquid interface involve clusters, which cannot stick to the ice. Hence the equilibrium at the interface is destroyed by adding salt because the clusters diminish the frequency with which water molecules enter the ice phase. The frequency with which they leave the ice is unchanged, so that the ice begins to melt.

The melting accomplishes two things, both of which increase the freezing rate so that it again matches the melting rate. The water that melts dilutes the salty solution, increasing the probability that a . collision at the interface will be by a water molecule rather than by a hydrated ion. Melting also lowers the temperature of the system because each molecule freed from the ice must be supplied with energy by molecules in the liquid. As the liquid loses energy it cools.

When salt is sprinkled over ice on a sidewalk, the system of salty ice water is different because it is not isolated. Although the water might at first cool as it supplies energy for melting, the sidewalk and the surrounding air quickly make up the lost energy. The temperature stays constant while the ice melts.

Often the purpose of the salt is stated in terms of the freezing point of the ice water. When salt is added, the freezing point of water is lowered from its normal level of zero degrees C. The freezing process then extends over a broad range of temperatures. In Figure 6 the temperature of salty ice water is plotted as a function of the liquid's salinity, measured in parts per 1,000, specifically the number of grams of salt in 1,000 grams of liquid. The curve represents the equilibrium values at which the rates of freezing and melting are balanced.

To understand the freezing of salty water you might begin with the liquid at low salinity. As you remove energy from the mixture by cooling it the temperature eventually goes below zero degrees C., and yet the mixture does not freeze. The freezing point of the mixture is lower by several degrees than the freezing point of fresh water. The addition of the salt is said to have depressed the freezing point.

Every time energy is removed from the system the freezing rate becomes higher than the melting rate and part of the water freezes until the rates are rebalanced. Since the amount of liquid water decreases, the salinity increases. As you continue to cool the system and upset its equilibrium, it reestablishes the equilibrium and moves down the curve.

Eventually the system reaches a state called the eutectic point. Further cooling then solidifies the entire system into ice and hydrated sodium chloride. The temperature of the eutectic point depends on the type of salt. For sodium chloride it is about -21.1 degrees C., with a salinity of 233 parts per 1,000. A solution of calcium chloride has a eutectic point with a temperature of-55 degrees, which is why calcium chloride is much better than sodium chloride for melting ice on a sidewalk in cold weather. The calcium chloride depresses the freezing point of water more than sodium chloride does.

The same type of graph serves as a guide for adding sodium chloride to the ice in an ice-cream maker. Suppose you have added a small amount of salt to the ice so that the salty ice water is approximately in equilibrium at a temperature not far below zero degrees C. This point is labeled A in Figure 7. Since the temperature is not very low, the ice-cream mixture freezes slowly. Tired of cranking, you decide to put more salt in the ice.

The additional salt momentarily destroys the equilibrium in the system because the new supply of sodium and chloride ions binds more of the water molecules into clusters. Since the freezing rate decreases, part of the ice melts as the system heads back toward equilibrium. And since the system is almost isolated, the energy required for melting must come from the kinetic energy of the water molecules. They slow down, and so the temperature decreases.

Eventually enough of the ice melts so that the freezing rate again matches the melting rate. Equilibrium is reestablished, but at a lower point on the curve. In spite of the melting of some of the ice the salinity is now higher than it was at the outset because of the additional salt. More important, the temperature is now lower because energy was removed in the melting process.

In an ice-cream maker the salty ice water is not an isolated system. Energy leaks at a low rate through the wood walls. Even more energy comes from the mixture that is to be frozen. Still, the point is that the addition of salt to the ice water drives the temperature down by melting some of the ice. The cold liquid. bath can then remove energy from the container of ingredients. The salty ice water remains cold even though the ice-cream mixture continues to supply energy. The container should be metal so that energy can be conducted into the bath to freeze the mixture in a reasonable amount of time.

I checked the ability of salt to decrease the temperature of ice water by salting ice water in a Thermos bottle. I monitored the temperature with an unmounted type K thermocouple connected to a thermocouple thermometer bought from the Cole-Parmer Instrument Company (7425 North Oak Park Avenue, Chicago, Ill. 60648). As I added more salt, ice melted and the temperature dropped until it was-15 degrees C. There it stopped, presumably because heat was leaking into the bottle.

Is ordinary ice suitable for freezing an ice-cream mixture? It is if enough of it makes good contact with the container. Ice cubes do not pack well. Chopped ice makes more contact but still not enough to keep the churning time reasonably short. Indeed, the churning time might be so long that butter separates from the ice-cream mixture and no ice cream results. Far better contact is made when the container is surrounded by a bath of ice water. Without salt in the bath, however, the ice water can be no colder than zero degrees C.

With salt in the bath the container will be surrounded by a liquid that is colder than zero degrees C. The more salt, the lower the temperature. My grandmother cautions, however, that one should not add too much salt or the ice cream will be ruined. Salt and ice should be in a ratio of about 1:4. More salt creates too low a temperature in the ice-water bath. Since the difference in temperature between the ice water and the ice-cream mixture is then greater, heat is conducted through the wall of the container at a higher rate. Heat is not conducted through the ice-cream mixture as quickly. Hence the layer of mixture adjacent to the wall cools and freezes rapidly whereas mixture farther from the wall does not.

Churning the mixture helps to prevent the early freezing of the outside layer. Still, if the ice-water bath is cold enough, early freezing may set in. The dasher is then difficult to turn even though the center of the ice cream is not yet viscous. The person cranking, misled by the resistance of the dasher to turning, concludes that the mixture needs no more churning and allows it to sit for a while. As a result ice crystals grow, giving the product a hard, grainy texture.

The initial mixture of ingredients is enough to fill about 70 percent of the container. As you churn the mixture it expands enough to fill the container, partly because water expands when it freezes. Much more of the expansion, however, is due to the air beaten into the mixture by the churning.

When I first begin to churn, I turn the handle slowly because the viscosity of the mixture is too low to trap air bubbles. Vigorous turning would be wasted effort. Besides, I do not want to churn butter out of the cream. When the solution has cooled and become thicker, I must turn the handle faster to work air bubbles into the mixture. They are trapped in the fluid by the viscosity and the freezing. Their effect-is to lighten the ice cream, which otherwise would be as heavy as ice.

The ice-cream mixture is prepared hot so that the sugar and other ingredients dissolve and mix thoroughly. Obviously if it is still hot when you put it in the ice-cream maker, the bath of salty ice water must remove a great deal more heat. That is why you should cool the mixture, the dasher and the container in the refrigerator for an hour or even longer.

When an ice-cream mixture is frozen without churning, it is grainy. The difference in the final texture comes from the growth of ice crystals in the mixture. Suppose the mixture is cooled and frozen slowly in a freezer. A few tiny ice crystals appear. They grow as more water separates from the mixture to crystallize at those sites. When the ice cream is completely frozen, it is full of large ice crystals.

When an ice-cream mixture is churned as it freezes, the crystals are much smaller. The motion of the churning disrupts the process of crystallization, and the initial nucleating sites no longer dominate it. Instead many more sites have a chance to initiate crystals. The mixture freezes into a great many small crystals rather than a few large s ones. The result is smooth ice cream.

Churning also prevents the formation of large ice crystals in another way. It coats the newly formed crystals with some of the cream in the solution. It is then difficult for more water to reach the crystal surfaces. Crystal growth is retarded and new crystal sites can appear. Milk, eggs, honey and gelatin also serve to retard crystal growth.

If you must make ice cream in your freezer compartment, occasionally beat it with an electric mixer or blender. This procedure breaks up some of the ice crystals and coats them with the cream. In addition it beats air bubbles lil into the mixture.

I did several simple experiments with ice cream. I prepared a batch according to my grandmother's recipe and divided it between two bowls, which I covered with aluminum foil and put in my freezer. The freezer stays at approximately -12 degrees C., which is cold enough to freeze ice cream. As ice began to form along the edge of the mixtures, I re moved one batch and beat it for several minutes in a mixer at low speed. Then I put it back in the freezer.

Several hours later both mixtures were well frozen. The one that had been beaten had many small air bubbles frozen in place. Removing small samples with a spoon, I could see ice crystals a few millimeters long, large enough for me to feel them when I put the ice cream in my mouth. I could both feel and hear the crystals crunch between my teeth.

The mixture that had not been beaten was hard with ice. It was also dense because there were no air bubbles in it. The edge of the spoon made noise as it broke through the ice. When I put a sample in my mouth, some of the pieces of ice were too big to crunch readily. Eating the stuff was like chewing small ice s cubes. The same kind of failed ice cream results when I melt a good sample and then refreeze it.

I compared both mixtures with some I had made from the same recipe and churned. One sample was churned in an apparatus identical with my grandmother's except that it was driven electrically. The other sample was prepared in an electric ice-cream maker lent to me by Peter Renz, an editor at W. H. Freeman and Company. His machine fits into a freezer compartment with its electric 16 cord extending out to a wall plug. (The rubber gasket on the freezer door fits snugly around the cord.) It has a small fan that blows air past the ice cream container to speed the freezing.

Whereas my machine with salted ice takes about 20 minutes of churning to make ice cream, the machine with a blower takes more than an hour. Both machines make smooth, light ice cream. I can see ice crystals and air bubbles only when I examine samples under a magnifying lens.

Several desserts similar to ice cream could be investigated. Sherbet, which usually includes a citrus juice and tiny bits of rind, lacks the cream of ice cream. It requires churning or beating because the growth of ice crystals can make it grainy and hard. The problem can be worse because many sherbets lack an ingredient that would coat the ice crystals and retard their growth.

A sorbet is prepared in a similar way except that a fruit is pureed and then blended into a sugar syrup. A granita is almost the same except that it is not churned or beaten, resulting in deliberately large ice crystals. While the mixture is freezing you should stir it periodically with a fork to leave it sharp with crystals but not lumpy with little ice cubes.

Bibliography

CHEMICAL THERMODYNAMICS. Frederick D. Rossini. John Wiley & Sons, Inc., 1950.

THE COMPLETE DAIRY FOODS COOK BOOK. E. Annie Proulx and Lew Nichols. Rodale Press, 1982.

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.