Mass Crystallization at Multiple Levels

A. M. Kuzmin

Experimental investigations of the phenomena of mass crystallization have as their aim not only the acquisition of data for solving theoretical questions, but also the acquisition of data for solving practical problems that are posed, on the one hand, by mineralogy and petrography and, on the other, by a number of industries for which the phenomena of mass crystallization play a very important role. In the present article I should like to set forth only the factual material of one experiment, the execution of which was intended to study the phenomenon of “stratification of solutions,” first established by P. A. Zemyatchensky.

This major expert on crystallogenesis and experimenter was the first in world practice to carry out the following original experiments. In order to study the behavior of crystals in a saturated solution, he took three crystals of alum, which were suspended at different levels. The first of them was suspended in the solution in its upper part; the third was located in the deep part of the solution; the second—at the middle level of the crystallizer. Data from one of such experiments are given in Table 1.

Table 1

From the table presented it is clearly seen that after four days the upper crystal No. 1 had noticeably decreased in size, the middle one had increased somewhat relative to its initial size, and the lower one during the same time had increased by 14.7%. After six days we see that the state of the solutions continues to change. This conclusion follows from the fact that the upper crystal, dissolving, had decreased considerably in size. The middle crystal, in comparison with its size at the stage of four days of the experiment, had decreased by 4.2%, while the lower—No. 3—continued, in comparison with the other two crystals, to grow and to increase in size.

The phenomenon just described made it possible for P. A. Zemyatchensky to say that the density of the solution changes along the vertical and increases from top to bottom. The initially saturated solution after four days becomes unsaturated in the upper part, and therefore crystal No. 1 situated at this level dissolves. In the intermeж—

At the exact horizon at the level of crystal No. 2, the solution proves to be weakly supersaturated and naturally causes weak growth of the crystal. At the horizon of crystal No. 3, the solution proves to be relatively strongly supersaturated.

After six days, the unsaturated state of the solution advances downward deeper than the level of crystal No. 2; moreover, from the behavior of crystals No. 1 and No. 2 it is evident that the undersaturation of the solution increases upward. Toward the level of the third crystal the concentration increases, and the solution remains supersaturated.

Thus, observations of the behavior of crystals located in the solution at different levels allowed P. A. Zemyatchensky confidently to speak, as it were, of the stratification of saturated solutions, of the change in the density of solutions from top to bottom as an obligatory property of theirs, which is revealed with the aid of suspended crystals.

However, A. K. Boldyrev doubted the correctness of P. A. Zemyatchensky’s conclusions. With the participation of V. I. Mikheev he carried out a series of experiments in order to verify P. A. Zemyatchensky’s data. A. K. Boldyrev and V. I. Mikheev, studying the density of alum solutions without crystals suspended in them, did not obtain convincing data that would confirm the conclusion that the density of the solution under investigation increased at all appreciably from top to bottom. But if crystals are placed in such a solution at the corresponding levels, their behavior in the solution generally repeats P. A. Zemyatchensky’s observational data. In connection with his experiments A. K. Boldyrev believes that stratification of solutions is produced only under conditions of crystallization and with the indispensable participation of concentration currents arising above crystals that have precipitated on the bottom of the crystallizer and causing an unsaturated state of the upper horizons of the solutions.

We set up our experiments on mass crystallization from solutions with the aim of revealing the state of saturated solutions when they are cooled within the range from 80–100° to 15–20°. In order to obtain convincing data, 40 experiments on mass crystallization from solutions of potassium aluminum alum and potassium chromium alum, potassium bichromate, and copper sulfate were carried out. The experimental part was performed with great care, initiative, and great diligence by second-year student I. M. Teplyakov.

A solution of one or another salt in the crystallizer was heated to 80–100° and brought to complete saturation. Then the heated crystallizer was placed in a hot water bath, and the entire system was cooled over the course of 24 hours.* In order to obtain comparable figures for characterizing the state of solutions at different horizons, a rack was placed in the crystallizer, the shelves of which were made of glass, measured 9×12 cm, and were spaced 17 mm from one another. The upper glass shelf of such a rack, being at a depth of one centimeter from the surface of the solution, served as a protective screen, received upon itself all the crystals that had precipitated from the overlying layer and from the surface part of it, whose regime was regulated to a considerable degree by evaporation phenomena. For this reason, the crystals that precipitated from the solution onto the very top shelf were not taken into account. The number of counted (working) plates in the rack was determined by the dimensions of the crystallizer.

* We chose 24 hours as the maximum for the experiment, during which time Lovitz-Wulff concentration currents do not undergo significant development. A good interval for the task we set is 12 hours, ensuring the greatest purity of the experiment, but owing to organizational difficulties it was not possible to carry them out.

Upon cooling of the crystallizer and the solution, the precipitation of crystals occurred both inside the tiered rack itself and outside it. Within the tiered rack, within the bounds of a horizon delimited by two parallel plates, the forming crystal nuclei were projected onto the surface of the lower-lying glass plates. Outside the tiered rack, the crystal nuclei fell to the bottom of the crystallizer around the rack. After 24 hours of the experiment had elapsed, and after air-drying the plates with crystals, the quantity of precipitated crystals on each horizon was determined in grams. Then the character of the distribution of crystals over the surface of the plate was studied. In addition, short-term experiments were conducted separately in order to determine the distribution of crystallization centers, the character of the precipitated crystals, and observations were made on the fate of the crystals during their further crystallization.

After brief methodological remarks concerning the experiments and their intended significance, we may proceed to consider the experimental material itself. The first group of experiments was carried out in flat crystallizers, in which there were only three working horizons in the tiered rack.

A crystallizer with a solution saturated at 80° was placed in an exsiccator with warm water, which was then closed with a ground-glass lid. As an illustration we present three crystallization experiments from potassium dichromate solutions (Table 2). From the examples given it is clearly seen that the total quantity of all precipitated crystals in grams increases from top to bottom. Other experiments belonging to this same group of experiments indicate the same thing.

Table 2

The next group of experiments on crystallization from solutions of the same potassium dichromate on 4 levels gave the following results. On the bottom of the crystallizer in experiment XIII, in comparison with the other experiments, there were abundant crystals in the sediment (Table 3).

Table 3

In the examples of crystallization just presented, we also see that the total quantity of precipitated crystals quantitatively increases, as a rule, from top to bottom. In experiment XIII, contrary to the general regularity—

apparently, the “dissolving” role of the crystals that had precipitated abundantly on the bottom of the crystallizer makes itself felt; moreover, this influence is manifested above all on the crystals of the fourth horizon and, apparently, somewhat more weakly on the crystals of the third horizon.

The third group of experiments on mass crystallization was carried out in larger crystallizers, in which we already had eight and nine working horizons. For illustration (Table 4) we shall present below the experiments with the crystallization of alum, which show that the regularity, observed in the previous experiments, of an increase in the number of precipitated crystals at many levels from top to bottom is clearly preserved in this group of experiments as well. But here, too, a number of interesting details is quite clearly indicated, complicating the overall picture of the phenomenon under consideration. Thus, the first detail consists in the fact that, during crystallization at many levels, there appear, as it were, violations of the general regularity. Such violations are evidently manifested in the form of certain jumps. For example, in experiment XXX such a jump is indicated in the transition from the second horizon to the third. This jump consists in the fact that the number of precipitated crystals increases from level 1 to 2; at level 3 a smaller number of crystals precipitates than at the level above; from level 3 the number of precipitated crystals increases regularly from top to bottom. It is curious to note that the jump from level 2 to 3 in the number of precipitated crystals is clearly expressed in experiments XXXI, XXXV, and others. Study of experiments XXXI and XXXV shows that the number of jumps in the smoothness of change in the number of precipitated crystals at many levels is equal to 4–5. However, the presence of jumps does not disturb the general tendency toward an increase in the number of crystals precipitated from the solution at many levels.

Table 4

The second detail in the group of experiments described, by its character, recalls the phenomena that occurred in experiment XIII. Thus, in experiment XXIX, which lasted, instead of the accepted 24 hours, exactly two days, beginning with

¹ The “dissolving role” of crystals should be understood as the ability of crystals, during crystallization, to disturb the equilibrium in the concentration of the solution and thereby to exert additional influence on the dissolved substance in the mass of the solution. Such circumstances lead to the fact that some of the precipitated crystals continue to grow, while others—in order to maintain the qualitative-quantitative states of the solution—dissolve, or at least their growth slows. In the case indicated above, around the rack at the bottom of the crystallizer there precipitated a relatively large number of crystals in comparison with the mass of crystals that precipitated on the lower shelves of the rack. This nonuniformity in the concentration of the mass of crystals at the lower level of the crystallizer also disturbs the general rhythmic regularity of mass crystallization noted above.

from the seventh horizon downward, the number of crystals present on each of them progressively decreases. Here the dissolving role of abundant crystals, which have ended up on the bottom of the crystallizer and increase their size at the expense of the material of the crystals on the lower horizons, is manifested. If one compares this with other experimental data concerning the small numbers of crystals that formed on the eighth and ninth horizons of experiment XXXI, then it will be more correct to assume that on the eighth and ninth levels the influence of the crystals that had fallen to the bottom and surrounded the stack was undoubtedly felt.

I believe that the analysis of the examples given of mass crystallization under special conditions may be limited to this, and that the principal conclusions following from the experimental data may be set out below.

1. A large body of experimental material does indeed confirm P. A. Zemyatchenskii’s conclusions concerning the stratification of saturated solutions. This phenomenon, under our conditions, becomes evident if crystallization proceeds from cooling solutions saturated at 80–100°.

2. The time allowed for crystallization as a mass process was taken to be so short that during the experiment the role of the traps of Wulff concentration streams, as observations show, does not have time to develop its significance throughout the whole mass of the solution. The 24 hours of the experiment, as our observations show, are sufficient to resolve the doubts expressed by A. K. Boldyrev.

3. All the experimental material convincingly shows that a solution saturated at a temperature of 80–100°, on cooling, even before the onset of crystallization, has time, in accordance with the gypsometric law and the force of gravity, to settle, and its density rhythmically increases from top to bottom. Only such a regular change in the solution can ensure the precipitation of crystals in accordance with the data obtained.

4. The study of crystallization at the very first stages of formation of visible nuclei shows that, from the space between the glass plates, small—for example, in the case of potassium bichromate, elongated—crystallites, formed in some center of crystallization of the solution, fall, being projected, onto the plate. At the places where the crystallites fall one can observe small heaps, representing a disorderly piling-up of small crystallites one upon another. Such heaps of crystallites are distributed over the surface of the plate relatively disorderly. The piling-up of crystallites in heaps indicates that the centers of crystallization are held for some time in the mass of the solution, each of them at one point. The disorderly distribution of the heaps indicates that the centers of crystallization in the solution also arise disorderly. The experimental studies carried out on many levels of crystallization indicate that the number of crystallization centers increases numerically together with the increase in the density of the solutions in the same direction.

5. What has just been said quite evidently indicates the fluctuation properties of solutions.

6. Following the first stage of crystallization in the heaped formations of crystallite nuclei, recrystallization of the heaped accumulations occurs. This recrystallization leads to the formation, in place of a heap, of one or two larger crystals. Converged heaps favor the development of several converged crystals of various sizes.

7. From the moment of formation, as a result of recrystallization, of relatively large crystals, the latter, entering into interaction with the solution, begin to grow at the expense of the substance present in the solution, as occurs in all cases of crystal growth. Simultaneously

With the process of regeneration of cluster accumulations of crystals there continues the process of precipitation of seed crystals from the crystallization centers arising in the solution; however, the seed crystals that have precipitated in this case also begin to increase in size. These observations indicate, to a certain extent, that within the layer bounded by the plates there is a differentiation of the solution of the second order, with a somewhat noticeable increase in density from top to bottom. This latter circumstance leads to a complication of the pattern of layered crystallization, which we observed in experiments XXXI, XXXV, and others.

1. Finally, the study of the distribution of crystals on the surface of the plates leads to the following conclusion. Individual crystals, being in the most favorable growth conditions, become relatively large and are often surrounded by relatively small crystals. At the nodes of concentration of many crystals, as a rule, no unambiguous orientation of the crystals is observed.

Thus, the factual material presented above made it possible to draw a number of extremely important conclusions, which were theoretically expected on the basis of all our knowledge and were well confirmed by the experiments. Among the expected points one must above all include the possible differentiation in the density of a solution saturated at 80–100° and left to cool by itself to 20–10° room temperature. At the same time, the experiments made it possible to reveal in detail also certain aspects of the course of crystallization as mass crystallization.

As for the application of the data obtained to the solution of practical problems in mineralogy and petrography, I intend to return to this question in another article.