Chemistry

O. N. BREUSOV, Corresponding Member of the Academy of Sciences of the USSR A. V. NOVOSELOVA
and Yu. P. SIMANOV

THERMAL AND X-RAY PHASE ANALYSES OF THE SYSTEM CsF—BeF₂ AND ITS RELATIONSHIP TO SYSTEMS OF THE TYPE MeIF—BeF₂

In the development of Goldschmidt’s concepts (¹) concerning structural models of complex oxides, systems Me+F—BeF₂ have been studied in our and in foreign laboratories; these systems may be regarded as weakened fluoride “models” of systems of the type Me2+O—SiO₂, where Me+: Li (²–⁴), Na (⁴–⁷); K (⁸); Rb (⁷, ⁹–¹¹).

From the standpoint of the change in the stability of various fluoroberyllates, it was of interest to investigate the system CsF—BeF₂. Differential-thermal and X-ray phase methods of investigation were applied.

The starting substances for obtaining cesium fluoroberyllates and intermediate compositions of the system were Cs₂CO₃, CsCl, and (NH₄)₂BeF₄.

Four compounds were found in the system: Cs₃BeF₅; Cs₂BeF₄; CsBeF₃; CsBe₂F₅. All of them were readily obtained by melting mixtures of the components taken in stoichiometric ratios, and the salts Cs₂BeF₄ and CsBeF₃ were also isolated from aqueous solutions. CsF was obtained by decomposition of Cs₂CO₃ with hydrofluoric acid and removal of excess HF by heating above 600°. When CsCl was used, the Cl⁻ ion was removed with freshly precipitated silver oxide.

Table 1

Results of analyses of the starting substances for preparing melts

Melts for thermal analysis were prepared from accurately weighed portions of analyzed CsF·HF; Cs₂BeF₄; CsBeF₃ and BeF₂. The composition of the prepared melts was monitored up to 50% BeF₂ (here and below, in mole percent) by analyses for cesium, and above 50% by analyses for beryllium. The deviation of the experimental data from the calculated values did not exceed 1 wt. % for cesium fluoride and 0.6 wt. % for beryllium fluoride.

Melts containing 0; 7; 13; 25; 31.5; 32.4; 33.3; 42.7; 47; 50; 54.7; 60.7; 66.66; 85.4; 90.8; and 100% BeF₂ were examined by X-ray diffraction.

Study of the system in the region 0–33% BeF₂ presented difficulty because of the exceptionally high hygroscopicity of cesium fluoride (for which reason, to avoid hydrolysis, CsF·HF rather than CsF was taken) and because of the peculiar phenomenon of melts creeping up the walls of the platinum test tubes used in thermal analysis. This could be avoided only by decreasing the heating rate; therefore part of the diagram (Fig. 1), up to 15% BeF₂, was constructed only from cooling curves. The remaining part of the diagram was constructed from heating curves, since on the less sensitive differential cooling curves the effects of polymorphic transformations were often not observed. The liquidus curve of quartz-like beryllium fluoride is less

reliable because of the strong glass formation in this region and the consequent indistinctness of the boundaries of the effects on the differential curves.

High-temperature X-ray analysis of the samples was carried out in “Unicam” cameras ({}^{(12)}) in capillaries made of a special silicate-free glass, stable until the appearance of a liquid phase, but not above (600^\circ).

For CsF, a melting point of (688^\circ) was determined, in agreement with the literature data ({}^{(13)}). At (14\%) (\mathrm{BeF_2}) and (598^\circ), CsF forms a eutectic with (\mathrm{Cs_3BeF_5}). (\mathrm{Cs_3BeF_5}) melts incongruently at (659^\circ). At (617^\circ) this compound undergoes a polymorphic transformation.

For the fluoroberyllate (\mathrm{Cs_2BeF_4}), a melting point of (793^\circ) was determined. This compound has a polymorphic transformation at (404^\circ).

The compound (\mathrm{CsBeF_3}) melts congruently at (475^\circ) and has two polymorphic transformations:

[
\text{liq.} \xleftarrow[\,]{475^\circ} \alpha \xrightleftarrows[\,]{360^\circ} \beta \xrightleftarrows[\,]{140^\circ} \gamma .
]

The eutectic (\mathrm{Cs_2BeF_4 + CsBeF_3}) lies at (449^\circ) and (48\%) (\mathrm{BeF_2}).

Fig. 1. System (\mathrm{CsF—BeF_2}). Phase fields:
(1)—(\mathrm{CsF + liq.}), (2)—(\alpha)-(\mathrm{Cs_3BeF_5 + liq.}), (3)—(\beta)-(\mathrm{Cs_3BeF_5 + liq.}), (4)—(\mathrm{CsF + \beta\text{-}Cs_3BeF_5}), (5)—(\alpha)-(\mathrm{Cs_2BeF_4 + liq.}), (6)—(\alpha)-(\mathrm{Cs_3BeF_5 + \alpha\text{-}Cs_2BeF_4}), (7)—(\beta)-(\mathrm{Cs_3BeF_5 + \alpha\text{-}Cs_2BeF_4}), (8)—(\beta)-(\mathrm{Cs_3BeF_5 + \beta\text{-}Cs_2BeF_4}), (9)—(\alpha)-(\mathrm{CsBeF_3 + liq.}), (10)—(\alpha)-(\mathrm{Cs_2BeF_4 + \alpha\text{-}CsBeF_3}), (11)—(\beta)-(\mathrm{Cs_2BeF_4 + \alpha\text{-}CsBeF_3}), (12)—(\beta)-(\mathrm{Cs_2BeF_4 + \beta\text{-}CsBeF_3}), (13)—(\beta)-(\mathrm{Cs_2BeF_4 + \gamma\text{-}CsBeF_3}), (14)—(\alpha)-(\mathrm{CsBeF_3 + \beta\text{-}CsBe_2F_5}), (15)—(\beta)-(\mathrm{CsBe_2F_5 + liq.}), (16)—(\alpha)-(\mathrm{CsBe_2F_5 + liq.}), (17)—(\beta)-(\mathrm{CsBeF_3 + \beta\text{-}CsBe_2F_5}), (18)—(\gamma)-(\mathrm{CsBeF_3 + \beta\text{-}CsBe_2F_5}), (19)—(\beta)-(\mathrm{CsBe_2F_5 + liq.}), (20)—(Q_{\mathrm{v}})-(\mathrm{BeF_2 + liq.}), (21)—(\beta)-(\mathrm{CsBe_2F_5 + Q_{\mathrm{v}}\text{-}BeF_2}), (22)—(\beta)-(\mathrm{CsBe_2F_5 + Q_{\mathrm{n}}\text{-}BeF_2}).

The compound (\mathrm{CsBe_2F_5}) has a polymorphic transformation at (450^\circ), melts congruently at (480^\circ), and forms eutectics with (\mathrm{CsBeF_3}) ((393^\circ) and (58.4\%) (\mathrm{BeF_2})) and with (\mathrm{BeF_2}) ((367^\circ) and (77.5\%) (\mathrm{BeF_2})).

Noteworthy is the great similarity of the phase diagrams of the systems (\mathrm{CsF—BeF_2}) and (\mathrm{CsF—ZnF_2}) ({}^{(14)}). The individuality of the cesium fluoroberyllates was confirmed by X-ray phase analysis. In high-temperature-

In high-temperature X-ray photography of beryllium fluoride it was found that quartz-like beryllium fluoride exists up to 580°, after which it melts. The cristobalite-like form, however, was encountered only up to 535°, slowly transforming into the quartz-like form (noticeably from 150°). It remains unclear, however, whether the disappearance of the lines of the cristobalite-like form of $\mathrm{BeF_2}$ above this temperature is due to its melting. High-temperature X-ray analysis confirmed polymorphic transformations in the compounds $\mathrm{CsBe_2F_5}$ and $\mathrm{CsBeF_3}$ and showed the absence of polymorphism in cesium fluoride up to 400°. The $\alpha \to \beta$ transformation of $\mathrm{Cs_2BeF_4}$ was not confirmed, since satisfactory photographs of this compound in the high-temperature chamber could not be obtained. The temperature of the $\alpha \to \beta$ transformation of $\mathrm{Cs_2BeF_5}$ lies above the stability temperature of silica-free capillaries.

Fig. 2. Dependence of the destruction temperatures of the crystal lattices of fluoroberyllates on cation radii. 1 — congruent melting, 2 — incongruent melting, 3 — decomposition in the solid state.

For the compounds $\mathrm{Cs_2BeF_4}$ and $\mathrm{CsBeF_3}$, the values of the lattice axial parameters, which proved to be orthorhombic, were determined from powder diagrams. For $\mathrm{Cs_2BeF_4}$ the axial parameters of the cell are: $a = 10.79\ \text{\AA}$; $b = 6.21\ \text{\AA}$; $c = 7.99\ \text{\AA}$. The pycnometric density is $d_{20} = 4.23$, whence $z = 3.91 \simeq 4$ and $d_{\text{X-ray}} = 4.35$. Oscillation X-ray photographs of a single crystal showed the cell to be primitive. It is of interest that, in the ratio of axes, $\mathrm{Cs_2BeF_4}$ is a closer crystal-chemical analogue of $\mathrm{Ba_2SiO_4}$ than is $\mathrm{Rb_2BeF_4}$ (11), despite the difference in radii.

For $\mathrm{CsBeF_3}$ the following cell parameters were found: $a = 7.18\ \text{\AA}$; $b = 4.44\ \text{\AA}$; $c = 11.96\ \text{\AA}$. The pycnometric density is $d_{20} = 3.43$; $d_{\text{X-ray}} = 3.46$.

Figure 2 shows the dependence of the destruction temperatures of the crystal lattices of fluoroberyllates (congruent melting, incongruent melting, decomposition in the solid state) on the radii of the cations. The diagram shows that compounds of the type $\mathrm{Me_2^IBeF_4}$ are the most stable. The stability of the compound $\mathrm{Me_2^IBeF_4}$ increases from Li to Rb and decreases slightly from rubidium to cesium. For compounds $\mathrm{MeBe_2F_5}$, the decomposition temperature from Li to K changes hardly at all. Compounds of this composition for rubidium and cesium are considerably stronger and melt congruently.

Moscow State University
named after M. V. Lomonosov

Received
8 X 1957

CITED LITERATURE

1. V. M. Goldschmidt, T. Barth, G. Lunde, W. Sachariasen, Skr. Akad. Oslo, No. 2, 117 (1926); No. 8, 127 (1927).
2. E. Thilo, H. Lehmann, Zs. anorg. Chem., 258, 332 (1949).
3. A. V. Novoselova, Yu. P. Simanov, E. I. Yarembash, ZhFKh, 26, 1244 (1952).
4. D. M. Roy, R. Roy, E. F. Osborn, J. Am. Ceram. Soc., 33, 85 (1950).
5. A. V. Novoselova, M. E. Levina, Yu. P. Simanov, A. G. Zhasmin, ZhOKh, 14, 385 (1944).
6. E. Thilo, H. Schröder, Zs. phys. Chem., 197, 9 (1951); 199, 125 (1952).
7. D. M. Roy, R. Roy, E. F. Osborn, J. Am. Ceram. Soc., 36, 185 (1953).
8. M. P. Borzenkova, A. V. Novoselova et al., Zhurn. neorg. khim., 1, issue 9, 2071 (1956).
9. N. A. Toropov, R. G. Grebenshchikov, Zhurn. neorg. khim., 1, issue 7, 1619 (1956).
10. N. A. Toropov, R. G. Grebenshchikov, Zhurn. neorg. khim., 1, issue 12, 2686 (1956).
11. R. G. Grebenshchikov, DAN, 114, No. 2, 316 (1957).
12. H. Lipson, A. Wilson, J. Sci. Instr., 18, 144 (1941).
13. Fluorine and Its Compounds, vol. 1, IL, 1953.
14. O. Schmidz-Du-Mont, H. Bornfeld, Zs. anorg. Chem., 287 (3), 120 (1956).
15. D. M. Roy, R. Roy, E. F. Osborn, J. Am. Ceram. Soc., 37, 7, 300 (1954).