ON THE POLYMORPHISM OF MAGNESIUM METASILICATE AT HIGH TEMPERATURE

V. N. Gurina, P. G. Usov

(Presented by the scientific seminar of the Department of Technology of Silicates, Inorganic Substances, and Electrochemical Production)

According to modern concepts, in the system \( \mathrm{MgO}—\mathrm{SiO}_2 \) there exist three crystalline forms of \( \mathrm{MgSiO}_3 \): enstatite, protoenstatite, and clinoenstatite. The question of the relationship between them is very complex and still has not found an unambiguous solution.

On the one hand, Tilley, Rorge [4], and others believe that in the high-temperature region clinoenstatite is stable, which is formed from low-temperature protoenstatite at \(1200—1270^\circ\mathrm{C}\).

On the other hand, Foster [5], Atlas [6], and others, on the basis of their investigations, assert that the only high-temperature crystalline form of \( \mathrm{MgSiO}_3 \) can be protoenstatite, which is enantiotropically related to low-temperature clinoenstatite. The latter, in turn, can transform into enstatite under certain conditions. The possibility of a direct transformation of protoenstatite into enstatite is also admitted. It should be noted that enstatite is known mainly as a natural mineral. It is not detected in the products of silicate technologies.

Protoenstatite and clinoenstatite are known and have been studied as products of the thermal decomposition of hydrous magnesian silicates; they can also be obtained artificially from a mixture of oxides and salts of Mg with silica. Both modifications are present in the ceramic body of steatite articles, making up its crystalline phase. Thus, the study of the polymorphism of \( \mathrm{MgSiO}_3 \) is of not only theoretical but also great practical interest, both in questions of production technology and in the question of the service behavior of finished articles—namely, in the question of their durability, since in explaining the aging of steatite ceramics investigators place the main emphasis precisely on the known phenomenon of instability of the protoenstatite lattice with time. A comparison of the X-ray diffraction patterns of these two modifications (Fig. 1) shows a great identity of their crystalline structures. According to data obtained by Eitel [1], the structural difference between protoenstatite and clinoenstatite consists only in ionic displacements of a definite kind. Similarly, Atlas [6] believes that in the protoenstatite–clinoenstatite transformation the change in the spatial lattice of the crystal occurs only along one axis, while in the other two directions the structure is preserved.

Without undertaking a special structural study of the crystal lattice, one can, on the basis of a comparison of the X-ray diffraction patterns of proto- and clinoenstatite, detect the following regularity:

1. On transition from the spatial grouping of clinoenstatite to protoenstatite, the ratio of the intensities of the diffraction maxima corresponding to the following interplanar spacings changes: \(d = 3.16\) Å (from \(J = 7\) to \(J = 10\)), \(d = 2.89\) Å (from \(J = 10\) to \(J = 7\)), and the splitting characteristic of clinoenstatite disappears (in the angular interval \(2\Theta = 26^\circ\)—\(33^\circ\) clinoenstatite has 4 diffraction maxima, whereas protoenstatite has only 2 (Table 1).

Fig. 1. X-ray diffraction patterns of the modifications: 1 — protoenstatite, 2 — clinoenstatite.

1. The diffraction maximum corresponding in the X-ray diffraction pattern to the interplanar spacing \(d = 2.11\) Å, \(J = 6\), disappears, and a new one appears, corresponding to \(d = 1.97\) Å, \(J = 6\).

2. The diffraction maximum corresponding to the interplanar spacing \(d = 1.52\) Å, \(J = 5\), disappears, and a new one appears, corresponding to \(d = 1.49\) Å, \(J = 6\).

Such a displacement of maxima in the X-ray diffraction patterns apparently reflects “ionic displacements” in the structural lattice of magnesium metasilicate.

Table 1

Values of interplanar spacings and intensities of diffraction maxima of clinoenstatite and protoenstatite in the angular interval \(2\Theta = 26\)—\(33^\circ\)

Experimental part

The aim of the work is to establish the actual picture of the thermal transformations of magnesium metasilicate.

Research methods: high-sensitivity thermography and X-ray analysis.

The thermographic studies were carried out on a derivatograph with a minimal load on the thermocouple. The rate of heating and cooling was \(5^\circ\) per minute.

X-ray investigations were carried out on a URS-50I diffractometer equipped with a special attachment. The attachment is a furnace with a platinum-rhodium heater. In the refractory lining and the furnace jacket, openings are made for the passage of X-rays; these are covered with nickel foil in order to reduce heat losses to the surrounding medium. The furnace design makes it possible to carry out X-ray recording in the angular interval \(2\Theta = 0—90^\circ\) [2].

The difficulty in studying the transformations of \(MgSiO_3\) itself and in obtaining comparable results by different authors lies, in addition to the great structural similarity of the crystalline forms, also in the fact that starting materials of different purity are taken as the object of study. Impurities and additives, even in amounts of \(0.5—1\%\), as investigations [3] show, greatly influence the course of structural-transformation processes in magnesium metasilicate.

Talcs, enstatites, bronzites—natural magnesium silicates used to obtain crystalline forms of \(MgSiO_3\)—sometimes contain a large amount of impurities: up to \(15\%\) \(Fe_2O_3\), as well as \(CaO\), \(Al_2O_3\), oxides of alkali metals, and microimpurities.

By using solid-phase synthesis from pure oxides, foreign impurities can be avoided. But this method does not guarantee the production of a monomineral product, since the reaction forming \(MgSiO_3\) proceeds through the intermediate product \(Mg_2SiO_4\), which may partially remain until the end of firing. The duration of synthesis in the solid phase is several days.

Finally, it is known that \(MgSiO_3\) in the form of clinoenstatite is well synthesized from a melt. However, as our investigations show, a melt obtained from a mixture of pure components does not crystallize completely.

Even with additional heat treatment consisting in repeated fritting of the batch, up to \(10—15\%\) of an amorphous phase remains.

Thus, the question of the rational production of crystalline \(MgSiO_3\) is itself an object of study.

In the present work, the behavior upon heating is investigated for a modification of \(MgSiO_3\) obtained in the form of clinoenstatite.

Clinoenstatite is synthesized from a melt, with the yield of crystallized material amounting to \(98—99\%\). This is ensured by the mineralogical composition of the starting material—enriched talc. The advantage of the talc used (Alguy deposit) is its high purity (Table 2).

Table 2

Chemical composition of enriched Alguy talc

To obtain clinoenstatite, the talc is corrected with magnesium carbonate, introduced on the basis of complete binding of the silica released during the decomposition of talc into metasilicate. The decomposition of talc proceeds according to the reaction:

\[ 3MgO\,4SiO_2\,H_2O \to 3(MgO\,SiO_2) + SiO_2 + H_2O\uparrow . \]

The content of \( \mathrm{MgSiO_3} \) in the dehydrated product is 83%. \( \mathrm{MgSiO_3} \) melts at a temperature of about \(1550^\circ\). During melting, the order of aggregation of molecular groupings in the liquid phase corresponds to this compound. \( \mathrm{SiO_2} \) and \( \mathrm{MgO} \)—from \( \mathrm{MgCO_3} \)—dissolve in the melt. Their grouping also occurs in accordance with the arrangement of the structural units in the initial melt.

It is important not to overheat the melt (in practice the temperature does not rise above \(1600^\circ\)) and not to hold it for a long time in the molten state, since the structural bonds in the melt corresponding to magnesium metasilicate are weak and are rapidly broken, forming groupings corresponding to the island structure of forsterite. (The melting of \( \mathrm{MgSiO_3} \) is considered incongruent.)

Crystallization of the melt occurs at a temperature of about \(1550^\circ\). The subsequent lowering of the temperature is carried out gradually. The crystallized material is ground and additionally fired for 1 hour at \(700^\circ\).

The crystalline material obtained consists practically entirely of clinoenstatite. The optically determined content of the glass phase does not exceed 1–2%; X-ray diffraction sometimes records the presence of cristobalite (by a diffraction maximum corresponding to an interplanar spacing \(d = 4.03 \ \text{Å}\)).

The nature of the relationships among the crystalline forms of \( \mathrm{MgSiO_3} \) is strongly affected by the particle size of the metasilicate. For the present study, a finely ground powder of clinoenstatite with a maximum particle size of 5 microns was used.

As follows from the thermogram (Fig. 2), heating clinoenstatite up to \(1100^\circ\) is not accompanied by any thermal processes. At \(1100^\circ\) an endothermic effect occurs.

On the cooling curve, one temperature arrest is also detected—an exoeffect at \(1060^\circ\). Since a pure material, chemically stable up to the melting temperature, was taken for the study, the endothermic effect at \(1100^\circ\) can be regarded only as a polymorphic transformation from one crystalline modification into another.

Fig. 2. Differential heating and cooling curves of clinoenstatite: 1—heating curve, 2—cooling curve, 3—temperature curve.

However, when comparing the X-ray diffraction patterns of the initial clinoenstatite and that fired at \(1150^\circ\), no substantial differences in the character of the diffraction maxima are found. After cooling of the fired clinoenstatite, the initial structure is retained. Consequently, the polymorphic transformation at \(1100^\circ\) is reversible. The exothermic effect on the cooling curve reflects the reversibility of the high-temperature polymorphic transformation. The equality of the areas of the thermal effects in the forward and reverse processes indicates the completeness of the transition of the high-temperature form into the initial one. Moreover, the transformation temperature is shifted by \(40^\circ\) toward the region of lower temperatures.

X-ray studies performed during heating of clinoenstatite make it possible to establish the nature of the high-temperature phase.

X-ray patterns are recorded from semicylindrical specimens pressed from powder of the material under study. The surface of the cylinder section is irradiated. The temperature is recorded by means of a thermocouple placed in front of the irradiated surface.

On the basis of the features described above, established by comparing the X-ray patterns of clinoenstatite and of the presumed high-temperature form of MgSiO$_3$—protoenstatite—in accordance with Foster’s point of view, it appears possible to carry out X-ray recording not over the entire range of angles, but to confine it to a narrow interval ($2\Theta = 42$–$48^\circ$), in which, during the transition from the clinoenstatite lattice to protoenstatite, the diffraction maximum corresponding to $d = 2.11$ Å disappears and a new maximum appears, corresponding to the interplanar spacing $d = 1.97$ Å.

During recording, registration is carried out continuously by constant reversing within the specified interval of angles $2\Theta$.

On the X-ray pattern of clinoenstatite, in the angular interval $2\Theta = 42$–$48^\circ$, two diffraction maxima are recorded, corresponding to interplanar spacings $d = 2.11$ Å and $d = 2.20$ Å. The intensity of the latter is three times smaller. When the material is heated, this position of the maxima is retained up to $1100^\circ$. Then a decrease begins in the intensity of the maximum corresponding to $d = 2.11$ Å, while the diffraction maximum corresponding to $d = 2.20$ Å disappears completely.

At the same time, a maximum at $2\Theta = 46^\circ$ ($d = 1.97$ Å), characteristic of the protoenstatite structure, appears and grows. At a heating rate of $2.5^\circ$ per minute, the diffraction maximum corresponding to the interplanar spacing $d = 2.11$ Å completely disappears by $1200^\circ$.

On the X-ray pattern, upon further heating, only one diffraction maximum remains, corresponding to $d = 1.97$ Å. This fully corresponds to the protoenstatite structure.

Further heating to $1250^\circ$ is not accompanied by any rearrangements of the crystal lattice.

When the system is cooled down to room temperature, magnesium metasilicate remains in the form of protoenstatite.

The amplitude of the maximum increases upon cooling owing to a decrease in the thermal vibration of the structural elements about the equilibrium position.

As the studies show, stabilization of protoenstatite obtained during the high-temperature polymorphic transformation from fine-grained clinoenstatite depends on a number of causes.

1. On the heating rate in the temperature region of intense structural rearrangement. To increase the completeness of the transformation, slow heating in the temperature interval $1100$–$1200^\circ$ is necessary.

2. On the degree of heating of the high-temperature phase. Studies show that, upon cooling of material fired at $1150^\circ$ with a three-hour hold, the clinoenstatite structure is retained, whereas material subjected to firing at $1250^\circ$ transforms into protoenstatite relatively irreversibly.

Conclusions

1. Pure clinoenstatite material with a minimal content of amorphous phase (1–2%) undergoes a polymorphic transformation when heated to $1100^\circ$. According to X-ray analysis data, this transformation is associated with the transition of clinoenstatite into protoenstatite.

1. The transition of clinoenstatite to protoenstatite upon heating is a reversible process. Upon cooling, the reverse transformation is recorded at a temperature of 1060°.

REFERENCES

1. W. Eitel. Physical Chemistry of Silicates. IL, 1962.
2. L. I. Dvorkin. Study of magnesian-dolomitic cement as a binder for building products with organic fillers. Dissertation, Tomsk, 1966.
3. L. G. Lavrent’eva, V. A. Presnov. Izv. vuzov “Fizika”, No. 5, 48–51, 1958.
4. Thilo E., Rogge G. Berichte der deutschen chemischen Gesellschaft, 1939, 341–362.
5. W. R. Foster, Journal of the American Ceramic Society, 1951, 255–259.
6. L. Atlas. The Journal of Geology, 1952, 125–147.