PHYSICAL CHEMISTRY

V. V. RACHEV, L. M. KOVBA, E. A. IPPOLITOVA

A CONTRIBUTION TO THE STUDY OF THE $\mathrm{UO_2—UO_3}$ SYSTEM

(Presented by Academician V. I. Spitsyn, June 30, 1964)

The $\mathrm{UO_2—UO_3}$ system has been intensively studied in recent years ($^{1-13}$). However, only in one work ($^1$) was X-ray photography at high temperatures used (up to $960^\circ$). Other authors studied quenched samples* ($^{9-12}$), or, in constructing the condensed-state diagram, used data on the dependence of oxygen pressure on composition, etc. ($^{4-7}$). In the present work the results are presented of a study of the $\mathrm{UO_2—U_3O_8}$ system by the method of high-temperature X-ray photography, and the possibility is considered of constructing a phase diagram on the basis of the data obtained and a critical review of those published earlier.

Samples of uranium oxides of various composition were prepared by annealing mixtures of $\mathrm{UO_{2.02}}$ and $\mathrm{U_3O_8}$ for 20 h in sealed quartz ampoules at $1000^\circ$. Photography was carried out in a high-temperature chamber ($^{14}$) in the range $850—1150^\circ$ (for samples $\mathrm{UO_{2.02}—UO_{2.40}}$) and $20—1000^\circ$ (for samples $\mathrm{UO_{2.50}—UO_{2.667}}$). Temperature fluctuations did not exceed $\pm 5^\circ$C; the radiation was CuK ($\mathrm{UO_{2.02}—UO_{2.40}}$) and CuK$_\alpha$ ($\mathrm{U_{2.50}—UO_{2.667}}$). The accuracy in determining the lattice parameters of cubic oxides was $\pm 0.001$ Å; for the parameters $b$ and $c$ of the rhombic subcells $\mathrm{U_3O_8}$ and $\mathrm{U_5O_{13 \pm x = 1}}$ it was $\pm 0.002$ Å, and for the period $a$, $\pm 0.005$ Å; deviations of composition did not exceed 0.003 in the oxygen index.

For determining the phase boundaries, data from qualitative phase analysis were used; for cubic phases in the $\mathrm{UO_{2.20}—UO_{2.25}}$ region, semiquantitative phase analysis and calculation of compositions from lattice parameters were used (using the linear dependence of the parameter $a$ of the $\mathrm{UO_{2+x}}$ phase on composition). The dependence of the lattice parameter on the composition of the $\mathrm{UO_{2+x}}$ phase at high temperatures is linear (Fig. 1); the coefficient of linear expansion is practically constant for this temperature interval, although it changes somewhat with composition.

Fig. 1. Dependence of the lattice parameters of cubic phases $\mathrm{UO_{2+x}}$ and $\mathrm{U_4O_9}$ on temperature: 1 — $\mathrm{UO_{2.02}}$; 2 — $\mathrm{UO_{2.17}}$; 3 — $\mathrm{UO_{2.20}}$; 4 — $\mathrm{UO_{2.22}}$; 5 — $\mathrm{U_4O_9}$.

The upper boundary of the homogeneity region of the $\mathrm{UO_{2+x}}$ phase shifts with temperature:

\[ \begin{array}{lll} 850^\circ & x = 0.19; & 950^\circ \quad x = 0.20; \quad 1100^\circ \quad x \ge 0.22.\\ 900^\circ & x = 0.195; & 1000^\circ \quad x = 0.21; \quad 1150^\circ \quad x \ge 0.25. \end{array} \]

In the oxide $\mathrm{U_4O_9}$ the lattice parameter also depends linearly on temperature up to $1100^\circ$, while in the interval $1100—1150^\circ$ an abrupt increase of the parameter occurs, and the oxide $\mathrm{U_4O_9}$ merges with the homogeneity region of the $\mathrm{UO_{2+x}}$ phase (an order–disorder type transition); the transition temperature is thus $1125 \pm 25^\circ$. The results obtained are in general agreement with the data of Roberts and Walter ($^5$) and differ greatly from the results of Shener ($^9$). The $\mathrm{U_4O_9}$ phase coexists with the rhombic phase $\mathrm{U_5O_{13-x}}$. The phase composition of samples $\mathrm{UO_{2.55}—UO_{2.667}}$ is given in Table 1.

* In view of the high rate of phase transformations in the $\mathrm{UO_2—U_3O_8}$ system, such data cannot be used to judge the phase composition of samples at high temperatures.

The phases \( \mathrm{U_3O_{8-x}} \) and \( \mathrm{U_5O_{13+x}} \) differ in the axial ratio \(a/b\) (1.685 and 1.695), in the character of their thermal expansion, and in the type of superstructure. With increasing temperature (Fig. 2B), the parameters \(a\) and \(b\) of the \( \mathrm{U_3O_{8-x}} \) phase change differently—the parameter \(a\) increases, while \(b\) decreases (in agreement with the data of previous works \((^{1,8})\)), and a gradual transformation into the hexagonal modification takes place.

Table 1

Phase composition of \( \mathrm{UO_{2.55}} \)—\( \mathrm{UO_{2.667}} \) samples at various temperatures*

* A — \( \mathrm{U_4O_9} \); B — \( \mathrm{U_5O_{13+x}} \); C — \( \mathrm{U_3O_{8-x}} \) (orthorhombic phase); D — \( \mathrm{U_3O_{8-x}} \) (hexagonal phase).
** 950°.

The thermal expansion of the \( \mathrm{U_5O_{13\pm x}} \) phase (Fig. 2A) is extremely small, and the transition \( \mathrm{U_5O_{13+x}} \) to \( \mathrm{U_3O_{8-x}} \) (hexagonal) in the case of the \( \mathrm{UO_{2.65}} \) and \( \mathrm{UO_{2.63}} \) samples occurs discontinuously.

For the \( \mathrm{U_5O_{13\pm x}} \) phase, which has a considerable homogeneity range, the lattice parameters practically do not change with composition (\(a = 6.733\); \(b = 3.665\); \(c = 4.142\)), but they differ noticeably from the parameters of \( \mathrm{U_3O_8} \) (\(a = 6.713\); \(b = 3.990\); \(c = 4.147\)). The parameter values agree with the literature data \((^{1,2,10,12})\).

As can be seen from Fig. 2, in the case of \( \mathrm{U_3O_8} \) above 900° there is a sharp increase in the parameter \(a\); this is evidently explained by a partial loss of oxygen. If the lattice parameters of the hexagonal phases \( \mathrm{UO_{2.65}} \) (500 and 600°) and \( \mathrm{UO_{2.63}} \) (950°) are plotted on the graph, they will lie on a straight line parallel to the analogous line for \( \mathrm{U_3O_8} \), on whose continuation fall the values of the parameters of the \( \mathrm{U_3O_8} \) phase at 1000 and 1100°.

If only stable phases are considered, then the \( \mathrm{U_3O_8} \)—\(\gamma\)-\( \mathrm{UO_3} \) region is two-phase, and both oxides have narrow homogeneity ranges \((^{13})\). On the basis of the foregoing, one may construct a condensed state diagram of the \( \mathrm{UO_2} \)—\( \mathrm{UO_3} \) system that includes only stable oxides.

Fig. 2. Dependence of lattice parameters on temperature: A — for the \( \mathrm{U_5O_{13}} \) phase, B — \( \mathrm{U_3O_8} \)
(1 and 2 — data for the compositions \( \mathrm{UO_{2.65}} \) and \( \mathrm{UO_{2.63}} \), respectively)

All the other uranium oxides are metastable, and their formation in a number of processes (exclusively low-temperature ones) is connected with kinetic factors and is due to the closeness of the structures of the metastable phases to the structures of the initial oxides: \( \mathrm{UO_{2+x}} \) and \( \mathrm{U_3O_7} \) to \( \mathrm{UO_2} \), \( \alpha\)-\( \mathrm{UO_3} \), \( \varepsilon\)-\( \mathrm{UO_3} \), \( \mathrm{U_2O_5} \) to \( \mathrm{U_3O_8} \). Thus, as a result of oxidation of unstable uranium dioxide at room temperature, cubic oxides \( \mathrm{UO_{2+x}} \), with \(x < 0.40\), are formed \((^{16})\). Probably in a number of cases they were mistakenly taken for products formed upon reduction of uranium oxides: \( \mathrm{U_8O_{17}} \) oxide in work \((^{17})\). Stable dioxide, upon oxidation, gives tetragonal oxides, close in composition to \( \mathrm{U_3O_7} \) (\(\alpha\)-\( \mathrm{U_3O_7} \) and \(\beta\)-\( \mathrm{U_3O_7}\)). During thermal decomposition of amorph-

of uranium trioxide, a hexagonal\({}^{16}\) or pseudohexagonal\({}^{13}\) oxide \( \mathrm{UO}_{3-x} \) is formed. As we have shown earlier,\({}^{15}\) the formation of various modifications of uranium trioxide (\(\alpha\) and \(\varepsilon\), in particular) during the oxidation of uranium dioxide is explained by the structural similarity of these oxides to the original uranium dioxide.

Metastable \( \mathrm{U_2O_5} \) oxide is likewise formed only as a result of a low-temperature process—acid treatment of uranium dioxide.\({}^{16}\)

The oxide \( \mathrm{U_3O_{8+x}} \), which is also formed as a result of the decomposition of amorphous \( \mathrm{UO_3} \) and \( \mathrm{UO_{3-x}} \), is extremely unstable—as a result of more prolonged calcination it is converted into \( \mathrm{U_3O_8} \).\({}^{16}\)

None of these oxides is obtained as a result of interaction between stable oxides,* and they should not be placed on the phase diagram. As is known,\({}^{2,18}\) uranium dioxide above \(900^\circ\) loses oxygen, giving an oxide of composition \( \mathrm{UO_{2.64}} \), which is very close to the lower boundary of the hexagonal phase \( \mathrm{U_3O_{8-x}} \). We observed that samples \( \mathrm{UO_{2.667}} \)—\( \mathrm{UO_{2.63}} \), calcined above \(900^\circ\), upon quenching give the metastable phase \(\beta\)-\( \mathrm{U_3O_8} \) (sometimes together with the stable phase \( \mathrm{U_5O_{13+x}} \) or \( \mathrm{U_3O_{8-x}} \)). On the heating curves of such samples an exothermic effect is observed at \(130^\circ\). X-ray phase analysis showed that this effect corresponds to the transition of \(\beta\)-\( \mathrm{U_3O_8} \) into \( \mathrm{U_5O_{13+x}} \) or \( \mathrm{U_3O_{8-x}} \). The formation of the \(\beta\)-\( \mathrm{U_3O_8} \) phase upon quenching samples of \( \mathrm{U_3O_8} \) calcined above \(900^\circ\) in sealed ampoules confirms the supposition of its partial decomposition under these conditions. It seems to us that, for \( \mathrm{U_3O_8} \) itself, such a transition is unlikely because of the gradual nature of the transformation of rhombic \( \mathrm{U_3O_8} \) into hexagonal, whereas in the oxide \( \mathrm{UO_{2.65}} \) it occurs abruptly; thus, \(\beta\)-\( \mathrm{U_3O_8} \) has a composition close to \( \mathrm{UO_{2.65}} \)—\( \mathrm{UO_{2.66}} \).

Fig. 3. Phase diagram of the \( \mathrm{UO_2} \)—\( \mathrm{UO_3} \) system:
\(I\)—\( \mathrm{UO_{2+x}} \), \(II\)—\( \mathrm{U_4O_{9\pm x}} \), \(III\)—\( \mathrm{U_5O_{13\pm x}} \), \(IV\)—\( \mathrm{U_3O_{8-x}} \) (rhomb.), \(V\)—\( \mathrm{U_3O_{8-x}} \) (hexagonal), \(VI\)—\(\gamma\)-\( \mathrm{UO_3} \). \(A\)—two variants of the possible structure of point \(A\); the dashed line indicates the assumed boundaries of the two-phase region.

Moscow State University
named after M. V. Lomonosov

Received
15 V 1964

REFERENCES

1. F. Grønvold, J. Inorg. and Nucl. Chem., 1, 357 (1955).
2. H. R. Hoekstra, S. Siegel et al., J. Phys. Chem., 59, 136 (1955).
3. H. R. Hoekstra, S. Siegel, Report P/737 USA to the International Conference on the Peaceful Uses of Atomic Energy, Geneva, 1955.
4. P. E. Blackburn, J. Phys. Chem., 62, 897 (1958).
5. L. E. J. Roberts, A. J. Walter, J. Inorg. and Nucl. Chem., 22, 213 (1961).
6. S. Aronson, J. Belle, J. Chem. Phys., 29, 151 (1958).
7. K. Kiukkola, Acta chem. scand., 16, 327 (1962).
8. S. Siegel, Acta crystallogr., 8, 617 (1955).
9. B. E. Shaner, J. Nucl. Materials, 2, 110 (1960).
10. H. Hering, P. Perio, Bull. Soc. chim. France, 1952, 351.
11. K. B. Alberman, J. S. Anderson, J. Chem. Soc., 1949, Suppl. 303.
12. R. E. Rundle, N. C. Baenziger et al., J. Am. Chem. Soc., 70, 99 (1948).
13. H. R. Hoekstra, S. Siegel, J. Inorg. and Nucl. Chem., 18, 154 (1961).
14. V. G. Kuznetsov, ZhNKh, 1, 1548 (1956).
15. L. M. Vidavskii, E. G. Lavut et al., DAN, 154, 1371 (1964).
16. L. M. Kovba, E. U. Kuzmicheva, E. A. Ippolitova, Vestn. Moscow Univ., No. 3, 34 (1964).
17. P. Margottin, W. Stuckens, R. Durand, C. R., 252, 4005 (1961).
18. P. P. Budnikov, S. G. Tresvyatskii, V. I. Kushakovskii, DAN, 128, 85 (1959).

* As we have shown,\({}^{16}\) Perio’s data,\({}^{10}\) concerning the formation of tetragonal oxides as a result of the reaction of \( \mathrm{UO_2} \) with \( \mathrm{U_3O_8} \), are inaccurate—the only reaction product is \( \mathrm{U_4O_9} \).