Crystallization from a Two-Layer Solution: Ethyl Alcohol—Aqueous Salt Solution

A. M. Kuzmin

One of the very important questions in the mineralogy and geology of ore deposits is the state of mineralized solutions and what conditions are necessary for their supersaturation in order to ensure the formation of mineral associations and the sequence of their formation. To the known factors I would like to add one more, which suggests itself as a consequence arising from experiments on crystallization from two-layer solutions. A similar process belongs to the processes of “salting out,” in which, from solutions of any degree of saturation of one salt, by adding a solution of another, the formation of nuclei and subsequent growth of crystals may occur owing to the formation of new compounds with a lower degree of solubility. Thus, from a saturated solution of barium chloride, when a solution of $\mathrm{Ca(NO_3)_2}$ is added, $\mathrm{Ba(NO_3)_2}$ precipitates (is salted out):*

\[ \begin{gathered} \mathrm{BaCl_2 + Ca(NO_3)_2 \longrightarrow Ba(NO_3)_2 + CaCl_2.}\\ 35{,}7 \qquad 129{,}0 \qquad 9{,}2 \qquad 74{,}5. \end{gathered} \]

For example, $\mathrm{KF}$ and $\mathrm{CaCl_2}$ behave similarly in solutions, promoting the formation of a $\mathrm{CaF_2}$ precipitate,

\[ \begin{gathered} \mathrm{2KF + CaCl_2 \longrightarrow CaF_2 + 2KCl.}\\ 95 \qquad 74{,}5 \qquad 0{,}0016 \qquad 34{,}0. \end{gathered} \]

In the examples given, the greater part of $\mathrm{Ba(NO_3)_2}$ and $\mathrm{CaF_2}$, forming a solid phase, will precipitate.

It must be supposed that salting-out processes of this type, with the formation of mineral associations under natural conditions, are widespread. For example, from solutions of sodium tungstate, calcium tungstate compounds

\[ \begin{gathered} \mathrm{Na_2WO_4 + Ca(HCO_3)_2 \longrightarrow CaWO_4 + Na_2CO_3}\\ 72 \qquad 0{,}11 \qquad 0{,}00064 \qquad 21{,}5 \end{gathered} \]

or iron and manganese tungstates are salted out in the form of scheelite or in the form of wolframite $(\mathrm{Fe}, \mathrm{Mn})\mathrm{WO_4}$, but in the geological literature these processes have not received due attention. In chemical practice, salting-out processes are well known. This type of reaction also includes the process of salting out salts from solutions in the presence of alcohol. This latter process, in the form of a two-layer solution—alcohol–aqueous

* The numbers under the “salts” are solubilities in grams per 100 cubic cm of water.

solution was used to obtain crystals of a number of salts soluble in water.

G. N. Khodalevich and A. M. Kuz’min developed a method for salting out salts from aqueous solutions with the formation of crystals under conditions of mutual diffusion from two-layer solutions, as stated above, and also from two-layer solutions of ethyl alcohol—aqueous saturated (and unsaturated) solution of the corresponding salts [7].

The essence of this method is that anhydrous ethyl alcohol is added to an aqueous solution, saturated with the substance \((A \cdot n\mathrm{H}_2\mathrm{O})\), of composition* \(p(A \cdot n\mathrm{H}_2\mathrm{O}) + m\mathrm{H}_2\mathrm{O} + Q\), containing \(p\) molecules. The latter, as is known, mixes well with water in any proportions and dissolves almost none of the salts with which the crystallization experiments from two-layer solutions were carried out.

The reaction of interaction between alcohol and the saturated aqueous solution may be represented in the following form:

\[ p\bigl[(A \cdot n\mathrm{H}_2\mathrm{O} + m\mathrm{H}_2\mathrm{O} + Q\bigr] + x[\mathrm{C}_2\mathrm{H}_5\mathrm{OH}] \rightarrow (p-x)[A \cdot n\mathrm{H}_2\mathrm{O}) + m\mathrm{H}_2\mathrm{O} + Q] + \]
\[ + x[\mathrm{C}_2\mathrm{H}_5\mathrm{OH} + m\mathrm{H}_2\mathrm{O} + Q] + x(A \cdot n\mathrm{H}_2\mathrm{O}). \]

It follows from the equation given that the removal by alcohol of \(x(A \cdot n\mathrm{H}_2\mathrm{O})\) of solvent leads to the result that some part of the solution becomes supersaturated, and this part promotes the appearance of nuclei and the growth of crystals at the expense of the substance being salted out, \(x(A \cdot n\mathrm{H}_2\mathrm{O})\).

To obtain crystals of the corresponding salts from saturated aqueous solutions in the presence of alcohol, the procedure is as follows. A saturated aqueous solution of the salt is poured into a measuring cylinder 25–35 mm in diameter to \(2/3\) of its height, and then alcohol is very carefully added from above to \(1/3\) of the height (with the cylinder in an inclined position). Between the aqueous solution and the column of alcohol an initially sharp interface is established. At the level of the boundary layer, or slightly below it, a grid is suspended for attachment of nuclei. Without the grid, the nuclei attach to the wall of the cylinder within the limits of the boundary layer, and, as they grow, owing to insufficient strength of attachment and to their own weight, the crystals break off and fall to the bottom of the cylinder. The use of a suspended grid makes it possible to obtain a large number of crystals up to 200 mm long.

The essence of the process of crystallization from a two-layer solution is that mutual diffusion of the alcohol into the aqueous solution and of the aqueous solution into the alcohol layer begins immediately, as soon as the cylinder with the solution is left at rest. The cylinder was closed, as far as possible, “tightly,” in order to reduce evaporation of the alcohol and to avoid the need to add it during the experiment.

The mutual diffusion that has begun is manifested first of all in the fact that the initially sharply expressed alcohol—aqueous-solution boundary broadens and then becomes diffuse and indistinct. On the walls of the cylinder, especially at the level of the boundary layer, crystals of the salt appear. Crystal nuclei also appear on the suspended grid. Subsequently diffusion proceeds readily along the axis of the crystal, ensuring its growth in length.

Experimental investigations of crystal growth by the author were carried out mainly with salts of the sodium and potassium halides (KCl, KBr, NaCl), epsomite \((\mathrm{MgSO}_4 \cdot 7\mathrm{H}_2\mathrm{O})\), potassium and sodium sulfates, and other salts. G. N. Khodalevich, from two-layer solutions alcohol—aqueous solution, also obtained crystals of carbonates, phosphates \((\mathrm{Pb}(\mathrm{NO}_3)_2, \mathrm{NiSO}_4 \cdot 7\mathrm{H}_2\mathrm{O})\), alums, and other compounds. Crystals ex-

* \(n = 0, 1, 2, 3,\ldots\), the amount of water of crystallization; \(m\) is the amount of solvent required to dissolve a molecule of the substance; \(Q\) is the heat of dissolution [6].

As a result of the experiment, crystals were obtained both within the column of saturated aqueous solution and within the column of alcohol. The best and most suitable crystals for investigation were obtained in the aqueous part of the solution. Within the alcohol column the crystals were usually small and very thin; in shape, among them there were cubic, prismatic, thin leafy, skeletal, and hairlike forms. The crystals of magnesium sulfate were fine-fibrous, up to 2–3 cm long and 0.01–0.005 mm thick. In their composition the water of crystallization—$7\mathrm{H_2O}$—was replaced, as G. N. Khodalevich established, by alcohol, and such crystals of magnesium sulfate became alcoholates.

Special attention was devoted to the study of halides and, chiefly, to the morphology of the crystals of KCl, KBr, and NaCl.

Cubic crystals of these salts of relatively large size, 1–7 mm, were obtained rarely; more often they were 0.5 mm and smaller. In addition, for the halides prismatic crystals were obtained, up to 50 mm in length and up to $5 \times 5$ mm in cross section; more rarely tabular, plate-like crystals were obtained, from $10 \times 10$ mm to $5 \times 5$ mm and smaller, with a thickness from 2 to 0.05 mm.

On the isometric crystals, in addition to the cubic forms $\{100\}$ typical of sodium and potassium halides, narrow faces of the pyramidal cube $\{210\}$ sometimes appeared. On prismatic crystals elongated in the direction $[100]$, faces of other indices, apart from cubic ones, were not recorded. The angles between $(100):(010):(001)$, for well-developed smooth faces of crystals 0.5–1.0 mm in size, are equal to $90^\circ$. X-ray analysis by the rotation method also showed that the crystals studied belong to the cubic system, and the value of the geometric constant for NaCl, $a_0 = 5.6284 \pm 0.0002$, and for KCl, $a_0 = 6.290 \pm 0.004$, correspond to tabulated data [4]. In polarized light they are optically isotropic.

Crystals of $\mathrm{MgSO_4 \cdot 7H_2O}$, as a rule, crystallized exclusively in the form of prisms elongated along the axis $[001]$. With a thickness from 2 to 0.5 mm they reached 150–200 mm in length. On the crystals themselves, of the monoclinic system, the faces $\{110\}$ are predominantly developed, and at the ends—$\{111\}$ and, occasionally, $\{1,1,1\}$. Their alcoholate forms also belong to crystals of the orthorhombic system.

In studying the morphology of crystals of potassium and sodium halides, special attention was given to the structure of the surface of their faces.

In goniometric investigation, some halide crystals showed a large (up to two degrees) scatter of signals from the faces $\{100\}$, whose surface proved to be geometrically complex, consisting of blocks of different size. The latter have a plate-like growth character and polygonal outlines (Fig. 1). Some blocks, relative to their neighbors, are located slightly higher, others slightly lower. Moreover, the surfaces of these blocks are inclined to one another at a very small angle. A general idea of the structure of the crystal surface at magnifications up to $120$–$160\times$ is given in Fig. 1, which is a copy of a photograph. In Fig. 1a the structure of the end part of a crystal of prismatic habit $[100]$ is shown. The blocks of this face are generally square in outline. The boundaries between the blocks are raised in black ink. The width of the boundary lines corresponds to the width of the intervals between the blocks, not occupied by crystals but filled with mother liquor.

In Fig. 1b the block structure of the surface of one of the prismatic faces of the halides is presented. The blocks themselves have a tendency to elongate in the direction $[100]$. In the same direction the unoccupied empty intervals between the growth blocks, which are filled with black ink, are also elongated. But even the black lines at high magnifications look like narrow gaps (slits, channels) separating one block from another.

Fig. 1. Block structure of sodium (potassium) halide crystals: a) block structure of the end face of a crystal (100); b) block structure of one of the prismatic faces of the same crystal. The black lines are block boundaries, cracks, and voids not occupied by salt crystals.

Of the details, the following two features should be noted. The first: the surface of the blocks under the microscope, with the use of an additional beam of light glancing along the plane being examined, itself proves to be composed of several disoriented crystallites, mostly of square outline. The second detail: in Fig. 2a a portion of a prismatic face is shown with a clearly expressed zonal structure and with voids—pores filled with mother liquor—arranged parallel to the plane of the growth zone. In Fig. 2b a part of the crystal surface is shown which consists of blocks separated by gaps filled with mother liquor. Under the microscope, by lowering and raising the tube, one can see that the channels, cracks, and gaps between the blocks continue inward, branch, taper out, but at one level or another blind passages, tubes, and volumetric voids again appear in the crystal. If Fig. 2b illustrates the picture of the crystal surface, then Fig. 2c shows the relation of blocks and the gaps between them at a depth of approximately 0.03 mm.

The sizes of crystallites distinguishable under the microscope at a magnification of 300–600× vary within wide limits, namely from \(0.06 \times 0.06\) to \(0.005 \times 0.005\) mm and smaller. If such crystallites are packed closely, there would be from 275 to 50,000 units per \(1\ \mathrm{mm}^2\). Taking into account, on the one hand, the loose packing and, on the other, the blocks of crystallites that are unequal in size and separated by gaps, the actual num-

Fig. 2. Details of the structure of the (100) surface of crystals: a) zonal distribution of mother liquor (black); b) voids, swirls, cracks, and channels occupied by solution in the near-surface part of the crystals; c) the same voids, cracks, channels, and swirls at a depth of 0.03–0.08 mm from the crystal surface.

for crystallites over the same measurement area ranges from 100–150 to 12–20 thousand units.

Crystallites (submicrons), as A. V. Shubnikov [5] once assumed, are, as a rule, of different sizes; in the construction of a crystal they are arranged far from strictly spatially parallel, and therefore the surfaces of all crystallites parallel to \(\{100\}\) that form the crystal do not lie in one plane. An electron microscope with magnification up to \(6000\times\) showed that on the etched surface of rock salt there are protrusions rising to different heights, and between them depressions, slots, and channels of different depths. Hence, naturally, the next layer of deposited crystallites cannot be ideally even; the crystallites composing it cannot be arranged strictly spatially parallel relative to one another, and therefore a ray reflected from such a surface will to some extent be scattered.

The study of the surface structure of halide crystals makes it possible to understand another of their features, namely: porosity and the manner of its formation. Thus, if the crystal surface proves to be micro-uneven, then the deposition of a new layer of crystallites leads to the trapping of part of the mother liquor filling the spaces between the crystals, and to the subsequent formation of negative crystals and corresponding pores. Pores in the crystals studied are sometimes isolated, but the great majority of them are arranged regularly and parallel to the \(\{100\}\) faces in several rows (Fig. 2, a), which as a whole determine the appearance of zoning in the crystals. The sizes of the pores vary over wide limits: from 0.008 to 0.001 mm and less in cross section; from 0.001–0.004 to 0.008–0.02 mm and up to 0.5 mm in length. In addition, there are pores of larger dimensions, 1–3 mm or more in length. Cases are not rare in which “swirls” occur in the crystals—very flat “bubbles” that generally repeat the shape of the spaces—“voids”—between blocks not filled by crystallites (Fig. 2, b). The pores in freshly obtained crystals are filled with mother liquor, which evaporates with time. Evaporation of liquid inclusions from crystals of NaCl, KCl, KBr, and in alums proceeds at different rates. At least in the crystals of the indicated salts, after 2–3 years most pores were free of liquid inclusions [3].

In zonally built crystals the number of zones varies widely. Thus, in prismatic crystals of NaCl and KCl up to 2.5–3.5 mm thick, the maximum number of zones recorded was from 24 to 31. In other specimens the number of zones decreased to 10–5. In very rare cases zoning, emphasized by the presence of pores, was absent. The sharpness of the boundaries of the zones was determined by the character of the pores themselves: the smaller the pores, the closer they were situated to one another, the more sharply the zone boundaries themselves stood out. The larger the pores, the farther apart they are situated, the larger the intervals between the pores themselves. The zonal distribution of pores indicates intermittent growth of the crystals, and the very presence of pores along a zone indicates that in the past the zone boundaries corresponded to the surface of the growing crystal, while the depressions (spaces between blocks) were filled with mother liquor, overlain by the next layer (zone) of growth.

Porosity affects, above all, the density of the crystals. Thus, the density of sodium chloride is 2.168 g/cm (calculated 2.165); the bulk weight of NaCl crystals obtained experimentally decreases to 2.0–1.89. For potassium chloride, instead of 1.99 (for sylvite), it is equal to 1.9–1.8. The bulk weight of potassium bromide is 2.5 instead of 2.75 g/cm³. The crystals obtained are transparent and possess a glassy luster. After some time, owing to evaporation of the solvent of the mother—

the mother liquor, the transparent crystals “turn cloudy” and become opaque; with the loss of liquid inclusions, the intensity of the pearly luster of the crystals themselves increases.

Experimental studies on obtaining crystals from two-layer alcohol–aqueous solutions make it possible to draw a number of conclusions of theoretical and practical importance.

1. From observation of the growth of sodium and potassium halide crystals it follows that: a) crystals grow by the deposition on their surface of layer after layer of crystallites; b) crystallites of different size are arranged, relative to their neighbors, not strictly spatially parallel to one another, and between individual crystallites and blocks of crystallites there are pores, fissures, channels, and coils of various size and shape, formed at the growth stage by the mother liquor and covered over by the next layer of crystallites. In short, the general surface of crystals is not ideally even; c) etching of cleavage surfaces of halite, sylvite, calcite, and other crystals quite evidently allows one to say that the crystals studied are built up of the corresponding crystallites, readily distinguishable at magnifications of 300–600×.

2. Evaporation of solvent from mother liquors trapped by crystals indicates that the pores, as a rule, are not sufficiently well isolated from one another [1]. In crystals, in addition to pores containing mother liquor, there are pores filled with gaseous products that are not connected in time with the formation of the crystal itself and, with respect to it, are epigenetic formations. Thus, in cleavage pieces of rock salt one can often observe, in the plane of cracks, more often along the \(\{110\}\) cleavage, the presence of cubic and parallelepipedal pores whose edges are parallel to the edges of the cube. In some of them liquid inclusions have been preserved; there are deposits of red powder and films of hematite and iron oxide hydrate; most pores are free of any inclusions. Experimental studies show that liquid can be introduced into layers parallel to \(\{100\}\) if wet filter paper is placed on the cleavage surface of rock salt. In this case it turns out that, at a certain depth from the surface of the crystal, in a plane parallel to the face of the cube, pores filled with liquid are revealed, although before the experiment this liquid was completely absent from them. These latter data force a cautious approach to the conclusions of decrepitation, which are usually based on the assumption that pores with gas–water inclusions are “blindly” (hermetically) isolated from one another and from the surface of the crystal [1], and that gas–water inclusions have been preserved as they were at the moment when they were sealed off by the layer of crystallites of the growing crystal.

3. The process of salting out one substance or another from solution was not generally denied, but in practice it was not used to explain the formation of mineral associations or in geochemistry. It follows from the experiments that, in order to obtain crystals from two-layer solutions, it is not necessary for the aqueous solution to be saturated. For precipitation of a sediment from an unsaturated solution in the presence of alcohol, more of the latter is required than in the case of a saturated solution. In this case salting out will develop in two stages. First, during diffusion, the free water must be bound, and then the water-solvent \(m\mathrm{H_2O}\) must be removed, thereby promoting precipitation of the dissolved substance \((A \cdot n\mathrm{H_2O})\).

Under natural conditions there is no “alcohol,” but there are media capable of “removing” the solvent and converting an undersaturated solution into a supersaturated state and thereby promoting precipitation of the dissolved substance.

For example, such a medium may consist of clayey shales, clays, porous sandstones, and finely crushed rocks (crushing zones), which to a certain extent may be more permeable to the solvent and impermeable or only slightly permeable to the solute. That is, in this case an osmotic process takes place, in which, instead of a membrane, there are media with extremely fine porosity that favors diffusion of the solvent and is difficult for the molecule \([(A \cdot n\mathrm{H}_2\mathrm{O}) + m\mathrm{H}_2\mathrm{O} + Q]\) to penetrate. With diffusion of the solvent (\(\mathrm{H}_2\mathrm{O}\)) into the remaining solution, the concentration of the solute increases and favorable conditions are created for the precipitation and formation of vein, ore, and other minerals.

1. The walls of cracks bounding a cavity within which hydrothermal solutions move may, to a certain extent, prove accessible to diffusion of the solvent. The part of the solution remaining in the cavity becomes more concentrated and supersaturated with one or another component or components capable of being deposited in the cavity of the cracks.

It is often assumed that hydrothermal solutions moving in the cavity of cracks leach part of the substance from the enclosing rocks. Such an assumption is hardly consistent with reality, since moving solutions are under the pressure of forces that cause them to move toward the earth’s surface, of the gravity of the solutions themselves, and of forces that cause the solutions partly to diffuse into the enclosing rocks. That is, what will take place is not the leaching of particular components from the enclosing rocks, but the washing of the substance of the solutions into the mass of such rocks.

Thus, experimental studies of the crystallization of salts from two-layer solutions make it possible to understand crystal growth and a number of its features as a consequence of the formation of the crystals themselves by deposition and successive layering of crystallites (submicrons, A. V. Shubnikov), while the study of salt crystallization from aqueous solutions makes it possible to explain the mechanism of formation of mineral salts from saturated and unsaturated solutions.

REFERENCES

1. E. E. Kostyleva. Some Methods for Studying the Ore-Bearing Capacity of Quartz. “Nauka,” 1964.

2. A. M. Kuz’min. The nature of imperfections in crystals. 1. Syngenetic imperfections of crystals. Izv. TPI, vol. 165, 1969.

3. A. M. Kuz’min. The nature of imperfections in crystals. Jubilee collection dedicated to the 50th anniversary of Soviet power. Geology and Geography. Tomsk State University Press, 1968.

4. V. I. Mikheeva. X-ray Determinative Guide to Minerals. G. N.-T. Publishing House, Moscow, 1957.

5. A. V. Shubnikov. How Crystals Grow. Moscow, Publishing House of the Academy of Sciences of the USSR, 1935.

6. A. M. Kuz’min. The phenomenon of concentration currents during crystallization. Izv. Siberian Branch, Academy of Sciences of the USSR, 1958, No. 6.

7. A. M. Kuz’min. Formation of mosaic crystals during salting out at the boundary of two-layer solutions. Physicochemical Dynamics of the Processes of Magmatism and Ore Formation. “Nauka” Publishing House, Novosibirsk, 1971.