This section is from the book "Experimental Cookery From The Chemical And Physical Standpoint", by Belle Lowe. Also available from Amazon: Experimental cookery.
In making icings, frostings, or candy like fondant and fudge, it is necessary to crystallize the sugar solution. For crystallization to occur, nuclei must form in the solution. To these nuclei the material of the solution is added to form crystals. Both the rate of formation of nuclei and the rate of crystallization are affected by the nature of the crystallizing substance, the concentration, the temperature, agitation, and the impurities present in the solution.
Nature of the crystallizing substance. Some substances like salt crystallize readily from water solution. It requires only a very slight super-saturation to start nuclear formation, and all excess salt in the solution beyond the saturation point is precipitated as crystals. Some substances do not form nuclei or crystallize so readily as salt. With sucrose it is often necessary to have a considerable degree of supersaturation before crystallization commences. Sucrose crystallizes more readily than levulose.
Formation of nuclei. Nuclei cannot form and crystallization cannot occur except from a supersaturated solution. The formation of nuclei, that is the uniting of atoms to form nuclei, is influenced by several factors. If a solution is left to stand, a few nuclei may form spontaneously in various places, and from these nuclei crystallization proceeds. When only a few nuclei develop spontaneously in the solution, the crystals grow to large size. Usually nuclei formation and crystallization do not begin immediately after supersaturation occurs. The rate of nuclear formation may be favored by specks of dust in the solution. Agitation or stirring of a solution increases the rate of nuclear formation. A drop in temperature at first favors, and then retards, the formation of nuclei. Instead of spontaneous formation of nuclei, seeding a solution may be used to start crystallization.
Seeding. When crystals of the same material are added to start crystallization the process is called seeding. These crystals serve as nuclei for crystal growth. If the quantity of crystals added is large and the size of the crystals small, it serves as many nuclei in the solution, and the resulting crystals are small. If the quantity of material added is very small, the nuclei formed are few in number and the crystals formed are large. One may think of all crystals as being large enough to be visible, whereas many of them may be very small, so small in fact that they may float in the air. Tutton tells of crystals difficult to obtain, but after being obtained several times in the laboratory they were then easily obtained. If crystals are floating in the air there is the possibility that they may serve to seed solutions, and thus start crystallization.
Rate of crystallization. To the nuclei formed in the solution new molecules from the solution are deposited, in a regular order or manner, so that each crystal has a typical shape. One side or face of a crystal may grow more rapidly than another. The rate at which the nuclei grow to larger size is called the rate of crystallization. This rate may be favored by the concentration of the solution and its temperature; it may be hindered by foreign substances.
Concentration of the solution. A more concentrated solution favors the formation of nuclei. A fondant sirup cooked to 114°C. contains less water and is more concentrated than one cooked to 111°C. Thus nuclei form more readily in the one cooked to 114°C. Large, well-shaped crystals form more readily if the degree of supersaturation is not too great. The most favorable supersaturation for crystal growth, of a sucrose solution boiled to 112°C, is that between 70° and 90°C. Although crystallization occurs in a very short time when the sirup is stirred at these temperatures, the crystals formed are larger than when the sirup is cooled to a lower temperature. See Figs. 1 and 3. Supersaturation and a low temperature are desirable for the development of small crystals. The viscosity of a very supersaturated sirup delays crystal growth.
Temperature at which crystallization occurs. It is a well-known fact that, in general, chemical precipitates come down more coarsely crystalline if crystallized at high temperatures. Barium sulfate is a good example of a substance that crystallizes in large crystals at high temperatures. The sugars follow this general rule. Other things being equal, i.e., concentration, etc., the higher the temperature at which crystal formation occurs, the coarser the crystals formed.
A drop in temperature at first favors the formation of nuclei, and then hinders it. Crystallization is favored in sugar sirups by cooling to a certain temperature, but is hindered when cooled to a lower temperature. Since the viscosity of a saturated sugar solution becomes increasingly greater as the temperature falls below 70°C, crystal formation is also slower as the temperature falls.
Agitation. Stirring a solution favors the formation of nuclei and hinders the depositing of the material of the solution on the nuclei already formed. Hence, crystals in solutions that are stirred do not develop to the size that they do in spontaneous crystallization. Bancroft states that "The mean size of crystals is determined by the total amount of the material crystallizing and number of crystals. The really important thing therefore is the number of nuclei which are formed under any given conditions." If small crystals are desired, then the conditions should be such that many nuclei are formed. Small crystals are obtained in sirups of definite concentration and temperature, if the sirup is stirred until the mass is kneadable. However, if the sirup is stirred for only a short time, some nuclei are formed, but after agitation is stopped, the formation of new nuclei is not favored and crystal growth is favored. Fig. 2 shows crystals from fondant stirred for only a few seconds. The crystals are much larger than those in Fig. 1, which shows crystals from fondant that was stirred until the mass could be kneaded. The same thing is illustrated in Fig. 10, which shows crystals from divinity that would not quite hold its shape but spread when dropped on oil paper. The crystals from the same divinity (Fig. 9) which was stirred until it would hold its shape, but was glossy and not dry in appearance when dropped on oil paper, are small. Divinity with the small crystals is very smooth and velvety on the palate. This emphasizes the importance of stirring candy and icing sirups until practically all the material is crystallized, if small crystals are desired.