PHYSICAL CHEMISTRY

A. M. RUBINSHTEIN and A. A. SLINKIN

MAGNETIC PROPERTIES OF Cr₂O₃—Al₂O₃ CATALYSTS

(Presented by Academician B. A. Kazanskii, January 5, 1959)

The magnetic properties of Cr₂O₃—Al₂O₃ catalysts prepared by impregnating γ-Al₂O₃ with chromic acid followed by reduction with H₂ at 360° and containing 0.1–40 wt. % Cr₂O₃ were studied by Selwood and co-workers (¹). They found that the magnetic moment at low Cr₂O₃ concentrations is ~3.7 μB, while in the interval from ~5 wt. % to 40% Cr₂O₃ it is equal to 3.2 μB; the authors associate this with the presence of Cr⁶⁺ ions. At ~8% Cr₂O₃ a sharp break was observed in the course of the susceptibility curve (χ), and χ increased rapidly as the Cr₂O₃ concentration decreased. On the curve of the dependence of the Weiss constant (Δ) on composition, a break occurred at the same point, and Δ fell to zero as the Cr₂O₃ concentration decreased. The authors related these changes in the concentration dependence of χ and Δ to the character of the distribution of Cr₂O₃ over the surface of γ-Al₂O₃.

A study of the process of oxidation by oxygen of deposited Cr₂O₃—Al₂O₃ and Cr₂O₃—ZnO catalysts, by constructing curves of the dependence of magnetic susceptibility on the degree of oxidation, was carried out by Matsunaga (²,³). He was able to show that the oxidation processes of Cr³⁺ ions to Cr⁶⁺ ions proceed most intensively in the initial region of Cr₂O₃ concentrations (approximately up to 10 wt. %), and that oxidation becomes more difficult as the surface area of the support decreases. Similar results were obtained by him in a magnetochemical study of the process of thermal decomposition of CrO₃ deposited on γ-Al₂O₃ (⁴), and in work (⁵) on the study of CrO₃—SiO₂ catalysts.

It seemed of interest to us to investigate the processes of oxidation of Cr³⁺ ions to Cr⁶⁺ ions for coprecipitated catalysts over a wide range of Cr₂O₃ concentrations, and to determine how the content of Cr⁶⁺ ions in the catalysts changes under the conditions of the catalytic reaction carried out on them, as a function of the Cr₂O₃ concentration. In addition, on the basis of a detailed magnetic analysis of the system, we intended to give an approximate picture of the distribution of Cr₂O₃ within the bulk of Al₂O₃.

Cr₂O₃—Al₂O₃ catalysts with contents of 0; 1.48; 4.4; 7.28; 14.2; 20.8; 33.2; 60.0; 80.0; 93.2 and 100 wt. % Cr₂O₃ were prepared by coprecipitation of Al and Cr hydroxides from mixtures of 10% solutions of the nitrates of these metals with a 10% NH₄OH solution. By heat treatment in air of the washed precipitates, dried at 110°, for 6 hours at 450° and 600°, two series of catalyst samples were obtained. The catalytic activity of the resulting samples was investigated in the decomposition reaction of iso-C₃H₇OH in the temperature range 275–320°. We carried out a magnetochemical study of the catalysts of both series before and after the catalytic experiments, as well as additional measurements of χ for fresh catalysts subjected to sintering by calcination in air at 1000° for 6 hours.

Figure 1 presents the change in magnetic susceptibility, calculated per 1 g Cr ($\chi_{\mathrm{Cr}}\cdot 10^6$), as a function of the concentration of Cr₂O₃ in the cata-

catalysts calcined at up to 450°, before and after operation. A sharp difference is evident in the values of \(\chi_{\mathrm{Cr}}\) for used and unused catalysts at \(\mathrm{Cr_2O_3}\) concentrations up to \(\sim 40\) wt. %. It is characteristic that the concentration range of \(\mathrm{Cr_2O_3}\) in which a sharp decrease in \(\chi_{\mathrm{Cr}}\) is observed for unused catalysts, as compared with catalysts reduced in the course of the reaction—in our case jointly precipitated oxides—is much broader than in the case of impregnated catalysts. This is apparently connected with the greater dispersity of \(\mathrm{Cr_2O_3}\) in jointly precipitated catalysts than in impregnated catalysts. Further, as is seen from Fig. 1, \(\chi_{\mathrm{Cr}}\) of the catalysts before and after carrying out the alcohol-decomposition reaction, beginning at \(\sim 60\) wt. % \(\mathrm{Cr_2O_3}\),

Fig. 1. Dependence of \(\chi_{\mathrm{Cr}}\) on composition for catalysts calcined at 450°.
\(a\)—before operation; \(b\)—after operation

Fig. 2. Dependence of \(\mu\) \((1, 2)\) and \(\Delta\) \((3, 4)\) on composition for catalysts calcined at 450°.
\(1, 3\)—before operation; \(2, 4\)—after operation

are practically identical. This, however, does not mean that in this range of \(\mathrm{Cr_2O_3}\) concentrations there are no \(\mathrm{Cr^{6+}}\) ions. On the basis of the temperature dependence of \(\chi_{\mathrm{Cr}}\), we calculated the magnetic moments \((\mu)\) of the catalysts before and after operation. The corresponding data are given in Fig. 2. Unused catalysts containing up to 40 wt. % \(\mathrm{Cr_2O_3}\) have a lower \(\mu\) than after operation, which is due to the presence in the former of a considerable quantity of \(\mathrm{Cr^{6+}}\) ions. The magnetic moments of catalysts with 60, 80, and 93.2 wt. % \(\mathrm{Cr_2O_3}\), both before and after operation, are the same (\(3.3\,\mu_{\mathrm{B}}\)) and lower than the theoretical value of \(\mu\) for the \(\mathrm{Cr^{3+}}\) ion. Consequently, these catalysts also contain \(\mathrm{Cr^{6+}}\) ions (which it was possible to confirm by qualitative chemical analysis), which are not reduced in the course of the reaction. The amount of \(\mathrm{Cr^{6+}}\) ions in the series of catalysts calcined at 450°, calculated from the values of \(\mu\), is presented in Fig. 3.

Analogous data were obtained for catalysts calcined at 600°, but in this case the limiting \(\mathrm{Cr_2O_3}\) content up to which a significant concentration of \(\mathrm{Cr^{6+}}\) ions is present in unused catalysts is reduced to 20 wt. %, and the concentration of \(\mathrm{Cr^{6+}}\) ions in catalysts calcined at 600° is approximately 1.5 times smaller than in the corresponding preparations calcined at 450°. In this series as well, the value of \(\chi_{\mathrm{Cr}}\) at \(\mathrm{Cr_2O_3}\) contents of 60 wt. % and higher is the same for used and unused catalysts; i.e., \(\mathrm{Cr^{6+}}\) ions are not reduced during catalyst operation under conditions of alcohol dehydrogenation–dehydration.

To elucidate the reasons that may cause such behavior of preparations with a \(\mathrm{Cr_2O_3}\) content \(\geq 60\) wt. %, we considered the possible pha-

phase composition of the $\mathrm{Cr_2O_3}$—$\mathrm{Al_2O_3}$ catalysts, based on magnetic data. Let us turn again to Fig. 1. We see that $\chi_{\mathrm{Cr}}$ decreases with increasing concentration of $\mathrm{Cr_2O_3}$ not smoothly, and at 60 wt. % $\mathrm{Cr_2O_3}$, $\chi_{\mathrm{Cr}}$ is equal to $\chi_{\mathrm{Cr}}$ at 33 wt. % $\mathrm{Cr_2O_3}$ (this corresponds to the maximum susceptibility calculated per 1 g of catalyst). The same is observed for catalysts calcined at 600°. On going from 33 wt. % $\mathrm{Cr_2O_3}$ to 60 wt. %, a sharp increase in $\Delta$ from 125 to 250° is also observed (Fig. 2). An increase in $\Delta$ on going from 33 to 60% $\mathrm{Cr_2O_3}$ also occurs in the case of catalysts calcined at 600 and 1000°: $\Delta$ increases, respectively, from 125 to 300° and from 130 to 460°; however, in catalysts calcined at 1000°, no increase in $\chi_{\mathrm{Cr}}$ is observed on going from 33 to 60 wt. % $\mathrm{Cr_2O_3}$.

Fig. 3. Dependence of the ion content on composition for catalysts calcined at 450°.
1 — before operation; 2 — after operation

A characteristic change in $\Delta$ is observed for all three series of catalysts in the interval 1.5–33 wt. % $\mathrm{Cr_2O_3}$. For catalysts calcined at 450°, $\Delta$ in the interval 1.5–14% $\mathrm{Cr_2O_3}$ changes from 75 to 100°, and in the interval 14–33% $\mathrm{Cr_2O_3}$ from 100 to 130° (Fig. 2). Catalysts calcined at 600° in the interval 1.5–14 wt. % $\mathrm{Cr_2O_3}$ are characterized by $\Delta$ varying from 30 to 70°, and in the interval 14–33% $\mathrm{Cr_2O_3}$ from 70 to 130°. In catalysts calcined at 1000°, $\Delta$ in the interval 1.5–14% $\mathrm{Cr_2O_3}$ is close to zero, and in the interval 14–33% $\mathrm{Cr_2O_3}$ its value changes from 100 to 130°.

Thus, we have three regions of $\mathrm{Cr_2O_3}$ concentrations in which $\Delta$, depending on the calcination temperature, changes in completely different ways: 1) the region 1.5–14% $\mathrm{Cr_2O_3}$, where $\Delta$ decreases approximately to 0 with increasing catalyst calcination temperature; 2) the region 14–33% $\mathrm{Cr_2O_3}$, where $\Delta$ practically does not change with increasing calcination temperature, remaining at the level of 100–130°; 3) the concentration region $\geq 60\%$ $\mathrm{Cr_2O_3}$, where $\Delta$ increases sharply with increasing calcination temperature. This last region of $\mathrm{Cr_2O_3}$ concentrations is also of interest because the catalyst with 93% $\mathrm{Cr_2O_3}$ calcined at 600° is antiferromagnetic, whereas pure $\mathrm{Cr_2O_3}$ under these preparation conditions is paramagnetic. When the calcination temperature is increased to 1000°, catalysts with 93 and 100% $\mathrm{Cr_2O_3}$ become antiferromagnetic. On the basis of the foregoing, the following approximate picture can be given of the distribution of $\mathrm{Cr_2O_3}$ in $\mathrm{Al_2O_3}$ for the catalysts studied.

In the interval 0–14 wt. % $\mathrm{Cr_2O_3}$ there is a solid solution of $\mathrm{Cr_2O_3}$ in $\mathrm{Al_2O_3}$, which becomes ordered as the calcination temperature is increased, leading to a decrease in $\Delta$. In the interval 14–33% $\mathrm{Cr_2O_3}$ there is a mixture of different phases. Along with the solid solution of $\mathrm{Cr_2O_3}$ in $\mathrm{Al_2O_3}$, substitutional solid solutions of $\mathrm{Al_2O_3}$ in $\mathrm{Cr_2O_3}$ apparently also form here, i.e., there are phases rich and poor in $\mathrm{Cr_2O_3}$. In some of them $\Delta$ may decrease with increasing calcination temperature, and in others it may increase. On average, these two processes may lead to an unchanged value of $\Delta$. In the concentration interval 60–93% $\mathrm{Cr_2O_3}$, a phase of free $\mathrm{Cr_2O_3}$ appears for the first time; moreover, in catalysts calcined at 450 and 600°, this phase is in a sufficiently dispersed state, and its appearance (at 60% $\mathrm{Cr_2O_3}$) is accompanied by an increase in $\chi$. In addition, in this region of $\mathrm{Cr_2O_3}$ concentrations, an interstitial solution of $\mathrm{Al_2O_3}$ in $\mathrm{Cr_2O_3}$ apparently appears, with which the appearance of antiferromagnetism in catalysts with a high $\mathrm{Cr_2O_3}$ content, calcined at 600 and 1000°, is connected.

It is probable that, in forming an interstitial solid solution, the $\mathrm{Al}^{3+}$ ions produce local displacements of $\mathrm{O}^{2-}$ ions in the lattice and thereby create favorable conditions for the occurrence of indirect exchange between $\mathrm{Cr}^{3+}$ ions as a result of the valence angle $\mathrm{Cr}^{3+}—\mathrm{O}^{2-}—\mathrm{Cr}^{3+}$ approaching $180^\circ$.

In light of the above, the absence of noticeable differences in the magnetic properties of fresh and used catalysts with $\mathrm{Cr}_2\mathrm{O}_3 \geq 60$ wt. % can be explained by the fact that the number of surface $\mathrm{Cr}^{6+}$ ions is very small in comparison with their number contained in the bulk of the catalyst crystals. Surface $\mathrm{Cr}^{6+}$ ions are readily reduced. The ions located in the bulk of the crystal, however, are not reduced under the given conditions ($T$, time) because of the absence of direct contact with the gas phase and the slow diffusion of vapors and gases into the crystals.

In addition to the $\mathrm{Cr}_2\mathrm{O}_3—\mathrm{Al}_2\mathrm{O}_3$ catalysts described above, we carried out a series of measurements for $\mathrm{Cr}_2\mathrm{O}_3—\mathrm{SiO}_2$ catalysts obtained* by depositing $\mathrm{CrO}_3$ on $\mathrm{SiO}_2$ followed by thermal treatment. Treatment of $\mathrm{CrO}_3—\mathrm{SiO}_2$ was carried out in two ways: 1) with ethyl alcohol at room temperature, and then with $\mathrm{H}_2$ at $515^\circ$ for 1 hour; 2) treatment with $\mathrm{H}_2$ or air, also at $515^\circ$ for 1 hour. These two groups of catalysts differed sharply in their magnetic and crystallographic properties. The former proved to be paramagnetic with $\mu = 3.2\mu_{\mathrm{B}}$ and X-ray amorphous. The latter were antiferromagnetic, and their X-ray patterns contained distinct lines of crystalline $\mathrm{Cr}_2\mathrm{O}_3$. We consider it possible that this difference in the magnetic behavior of the two groups of catalysts is explained by the formation of chromium silicates in the case of treatment of $\mathrm{CrO}_3—\mathrm{SiO}_2$ with alcohol, owing to the large heat evolution in the oxidation–reduction process. More detailed data on these catalysts will be published, together with a characterization of their catalytic activity, in one of the forthcoming communications.

Institute of Organic Chemistry
named after N. D. Zelinsky
Academy of Sciences of the USSR

Received
2 I 1959

REFERENCES CITED

1. R. P. Eischens, P. W. Selwood, J. Am. Chem. Soc., 69, 2698 (1947).
2. Y. Matsunaga, Bull. Chem. Soc. Japan, 30, 868, 984 (1957).
3. Y. Matsunaga, Bull. Chem. Soc. Japan, 31, 58 (1958).
4. Y. Matsunaga, Bull. Chem. Soc. Japan, 31, 745 (1958).
5. R. Mihail, P. Corlateanu, A. G. Ionesco, J. Chim. Phys., 56, 568 (1959).

* Together with M. I. Rozengart.