Relationship between the Electronic and Catalytic Properties of Metal Oxides in the Oxidation of Ammonia under Pressure

N. P. Kurin, M. S. Zakharov

(Presented by Prof. A. G. Stromberg, Doctor of Chemical Sciences)

It has long been noted [1] that substances having unfilled inner electron shells occupy an important place in catalysis. Let us consider whether this regularity is observed in the case of the oxidation of ammonia at elevated pressure.

The highest degrees of oxidation of \(\mathrm{NH_3}\) to \(\mathrm{NO}\) were observed by us on oxides of manganese, cobalt, vanadium, iron, nickel, copper, cadmium, and chromium (Figs. 1, 2, 3). All these elements, it turns out, are transition elements, i.e., their atoms have unfilled inner shells.

The oxide \(\mathrm{MoO_3}\), whose metal is also a transition element, departs from this regularity. However, this oxide is inactive for the reaction of ammonia oxidation. As will be seen below, this deviation can apparently be explained by the presence in this oxide of a wide forbidden band [7].

Figure 3 presents the dependence of the maximum degree of oxidation of ammonia to \(\mathrm{NO}\) under optimal conditions on the atomic number of the metal of the oxide in Shchukarev’s table [2]. From this figure it is evident that the best catalysts are oxides of metals located predominantly in the fourth and fifth rows of this table \((\mathrm{V_2O_3}, \mathrm{MnO_2}, \mathrm{Fe_2O_3}, \mathrm{Co_3O_4}, \mathrm{NiO}, \mathrm{Ni_2O_3}\), etc.). The elements of these oxides are transition elements with \(4s\) electrons and with unfilled \(3d\)-shells. Apparently, in oxides of elements with such an electronic structure, intermediate compounds are more readily formed through the formation of chemical bonds between unpaired electrons in the \(3d\)-shell of the oxide element and atoms of the ammonia molecule.

It should be noted that the overwhelming majority of oxides active for ammonia oxidation \((\mathrm{V_2O_5}, \mathrm{CuO}, \mathrm{NiO}, \mathrm{Ni_2O_3}, \mathrm{MnO_2}, \mathrm{PbO}, \mathrm{Cr_2O_3})\) have hole conductivity. Oxides possessing electronic conductivity, by contrast, are either only slightly active, such as \(\mathrm{TiO_2}\), or completely inactive, such as \(\mathrm{Sb_2O_3}, \mathrm{Sb_2O_4}, \mathrm{Sb_2O_5}, \mathrm{WO_3}, \mathrm{MoO_3}\), and \(\mathrm{Al_2O_3}\). \(\mathrm{CdO}\) departs from this regularity. The latter is an electronic semiconductor; however, in its presence ammonia is oxidized to \(\mathrm{NO}\).

Jones, Huwras, and Schuwalter [3], who studied the oxidation of \(\mathrm{NH_3}\) under atmospheric pressure at low temperatures, also arrived in 1954 at the conclusion that hole semiconductors have an advantage for the reaction of \(\mathrm{NH_3}\) oxidation.

Vertical axis: Degree of oxidation of \(NH_3\), %

Fig. 1. Dependence of the degree of oxidation of \(NH_3\) to \(NO\) on space velocity at \(8\ \mathrm{kg/cm^2}\) in the presence of:

\(a\)—NiO \((1—720;\ 2—780;\ 3—800^\circ C)\);
\(b\)—NiO \((1—600,\ 2—620,\ 3—640,\ 4—660,\ 5—680,\ 6—700^\circ C)\);
\(c\)—\(Cr_2O_3\) \((1—600,\ 2—640,\ 3—700,\ 4—780^\circ C)\);
\(d\)—CuO \((1—640,\ 2—660,\ 3—700,\ 4—720^\circ C)\);
\(e\)—\(MnO_2\) \((1—700,\ 2—760,\ 3—800^\circ C)\);
\(f\)—\(Ni_2O_3\) \((1—660,\ 2—760,\ 3—780,\ 4—800^\circ C)\);
\(g\)—CdO \((1—740,\ 2—760,\ 3—780,\ 4—800^\circ C)\);
\(h\)—CuO \((1—740,\ 2—760,\ 3—780^\circ C)\);
\(i\)—\(V_2O_3\)—\(800^\circ C\).

The fact of better oxidation of ammonia to NO in the presence of semiconductors possessing hole conductivity, effected by ions with an excess of positive electricity, can be explained as follows.

It is known that for the oxidation of \(NH_3\) to NO it is necessary that the catalyst surface sorb oxygen well from the gas phase.

In general, when molecules or atoms of a gas approach the contact surface, the following may be observed. Electrons of the surface will pass to the gas atoms, or vice versa. When oxygen atoms approach, which lack only two electrons in their outer electron shell, electrons will pass from the contact surface to the oxygen atoms and charge them negatively. And since hole semiconductors carry an excess positive charge, oxygen ions will be strongly attracted (sorbed) to the contact surface, with formation of an intermediate catalyst–oxygen complex, to which the ammonia molecule is adsorbed by its hydrogen ions.

In the case of oxides with electronic conductivity, effected by ions with an excess of negative electricity, formation of the intermediate catalyst–oxygen complex is hindered by the fact that the oxygen ions and the catalyst surface carry charges of the same valence (negative).

It naturally follows from this that, in order to obtain the most active composite catalysts for ammonia oxidation, the promoters should be selected so that, upon addition of the latter, the catalyst acquires hole conductivity, for foreign impurities not only sharply change the magnitude of the electrical conductivity, but may also change the mechanism of electrical conductivity of the semiconductor catalyst.

As early as 1928, Roginskii and his co-workers, studying the process of catalytic decomposition of unstable solid oxygen compounds, found that the catalyzing substances are colored in the visible part of the spectrum. The same regularity was also observed in Roginskii’s laboratory for the deep oxidation of organic compounds [4].

Therefore the idea also arose for us of the possible existence of the same dependence in the case of ammonia oxidation at elevated pressure.

Indeed, intensely colored oxides proved to be good catalysts for this process: \(MnO_2\) (black), \(Bi_2O_5\) (brown), \(V_2O_3\) (black), \(Ni_2O_3\) (dark gray), \(NiO\) (light green), \(Cr_2O_3\) (dark green), \(CdO\) (yellow), \(CuO\) (black), \(PbO_2\) (brown), \(PbO\) (yellow). Next came white titanium dioxide and white zinc oxide; the reaction was not catalyzed at all by white aluminum oxide, tin oxide and dioxide, antimony trioxide, tetroxide, and pentoxide, molybdenum anhydride with a yellow tint, and lemon-colored tungsten anhydride.

When the color of an oxide is compared with the degree of oxidation in its presence (Table 1), it is seen that there is no direct dependence between the intensity of the color tone and the degree of oxidation. Thus, for example, black \(CuO\) or dark-brown \(PbO_2\) stand below gray \(NiO\). Only white \(TiO_2\) and \(ZnO\) stand below the colored oxides. The oxidation of \(NH_3\) to NO in the presence of these two oxides can apparently be explained by the presence in them of intense absorption bands in the near ultraviolet region.

Colored bodies, as Roginskii, Sapozhnikov, and Kucherenko [5] found in 1929, possess appreciable electrical conductivity even at room temperature. The latter, as we know, is effected by electrons or holes, or by both (mixed conductivity).

Further, almost all the colored oxides investigated by us (with the exception of CdO) are hole semiconductors, while the uncolored ones possess electronic conductivity. In the presence of the latter, oxidation of \(\mathrm{NH_3}\) to \(\mathrm{NO}\) almost does not occur at all (ZnO, \(\mathrm{TiO_2}\)). This circumstance leads to the conclusion that, for ammonia oxidation, the contact must possess holes or electrons that are readily transferred into a free state with the formation of holes. The latter is accomplished the more readily, the smaller the distance between the upper edge of the normal band and the lower edge of the conduction band. When this distance is not greater than \(2.4\,\mathrm{eV}\), the transfer of electrons into the free state may also take place at room temperature under the action of light quanta. According to the electronic theory of the structure of the solid state, the presence of electronic transitions also accounts for the color of a body in the visible part of the spectrum [6].

Fig. 2. Dependence of the degree of oxidation of \(\mathrm{NH_3}\) to \(\mathrm{NO}\) on pressure in the presence of:

1—\(\mathrm{Ni_2O_3}\) \((C_{\mathrm{NH_3}} = 6.0\text{–}6.2\%\ \text{by vol.};\ 700^\circ\mathrm{C};\ V = 8080)\).

2—\(\mathrm{NiO}\) \((C_{\mathrm{NH_3}} = 9.4\text{–}9.5\%\ \text{by vol.};\ 720^\circ\mathrm{C};\ V = 15100)\).

3—\(\mathrm{Cr_2O_3}\) \((C_{\mathrm{NH_3}} = 10.6\text{–}10.8\%\ \text{by vol.};\ 700^\circ\mathrm{C};\ V = 35300)\).

4—\(\mathrm{MnO_2}\) \((C_{\mathrm{NH_3}} = 10.2\text{–}10.5\%\ \text{by vol.};\ 760^\circ\mathrm{C};\ V = 47700)\).

In the figure: ordinate—degree of oxidation of \(\mathrm{NH_3}\) to \(\mathrm{NO}\), \(\alpha\%\); abscissa—pressure of the gas mixture, \(\mathrm{kg/cm^2}\).

If we consider the values of the forbidden band for various oxides (for CuO it is 0.64; \(\mathrm{V_2O_5}\)—0.45; CdO—0.4; NiO—0.55; \(\mathrm{MnO_2}\)—1.25; CoO—0.68; \(\mathrm{Cd_2O}\)—1.5; \(\mathrm{WO_3}\)—0.45; \(\mathrm{MoO_3}\)—0.5; ZnO—2.6; \(\mathrm{TiO_2}\)—3.05 eV), then the colored oxides are precisely those in which the latter is less than 2.4 eV. The color and shade of color of the oxides will depend on the magnitude of the energy absorbed by the electron in its transition from the normal band into the conduction band.

As was already indicated above, the catalytic activity of catalysts is connected with their electrical conductivity. It is therefore of interest to trace how this condition is fulfilled in our case. For this purpose we made use of the electrical-conductivity values for the metal oxides studied by us that are given in the literature [7]. It is true that, when examining the literature data on this question, it is evident that, for one and the same oxide, different authors obtained different values of electrical conductivity at the same temperature. This is due to differences in the methods of obtaining the oxides, which in each individual case lead to the formation of an oxide with a different amount of impurities. By an impurity it is customary to mean any deviation from the stoichiometric ratio in the substance. Therefore we shall trace only the qualitative relationship between electrical conductivity and catalytic activity.

The electrical-conductivity values for NiO, CuO, and CdO in the temperature interval \(600\text{–}800^\circ\mathrm{C}\) were obtained by us by extrapolation.

From comparison of the degree of oxidation \((a)\) and the electrical conductivity \((\sigma)\), it is seen that in the case of NiO and CuO, with an increase in temperature to \(700^\circ\mathrm{C}\)

Fig. 3. Dependence of the degree of oxidation of \(NH_3\) to \(NO\) on the atomic number and series of the oxide elements.

Vertical axis: Optimal degree of oxidation, \(X_{NH_3}\), to \(NO\).

Horizontal labels: \(Al_2O_3\) 13; \(TiO_2\) 22; \(V_2O_5\) 23; \(Cr_2O_3\) 24; \(MnO_2\) 25; \(Fe_2O_3\) 26; \(Co_2O_3\) 27; \(Ni_2O_3\) 28; \(CuO\) 29; \(ZnO\) 30; \(MoO_3\) 42; \(AgO\) 47; \(CdO\) 48; \(SnO\) 50; \(Sb_2O_3\) 51; \(WO_3\) 74; \(PbO\) 82; \(Bi_2O_3\) 83.

Right-hand horizontal-axis notation: atomic number; group.

Table 1

an increase of both $\alpha$ and $\sigma$ is observed. When the temperature is raised above 700°C, however, the curves $\lg \sigma = f(t)$ and $\alpha = f(t)$ diverge: the former continues to rise, while the latter falls. This divergence is due to the fact that at temperatures above 700°C (specific only for these catalysts), in the presence of CuO and NiO, the process of ammonia oxidation shifts toward the formation of elemental nitrogen, for reasons that will be given in the following communication.

A somewhat different picture is seen in the case of CdO. Here a descending character of the curve $\alpha = f(t)$ is observed, whereas the curve $\lg \sigma = f(t)$ rises throughout the temperature interval 720–800°C.

This is due to the fact that CdO is an electronic conductor, and the increase in electronic conductivity that occurs with increasing temperature will suppress the reaction of oxidation of NH$_3$ to NO. This circumstance, together with a number of those cited above, makes it possible to suppose that holes in the crystal lattice of the semiconductor take an active part in the catalytic process.

But the magnitude of the electrical conductivity is an altogether insufficient condition for characterizing the catalytic activity of a substance with respect to the reaction of oxidation of NH$_3$ to NO. Thus, for example, at the same temperature CuO and CdO have a much higher electrical conductivity than NiO; however, in the presence of all of them the degree of oxidation of NH$_3$ to NO under optimum conditions is almost the same and is, respectively, 41.6, 44.7, and 43.1%.

The investigation we have carried out of the catalytic properties of metal oxides for the reaction of oxidation of NH$_3$ to NO has shown that, according to their activity, the latter can be arranged in the following order: MnO$_2$,

($\alpha = 73.8\%$), Bi$_2$O$_5$ (92.5%), Cr$_2$O$_3$ (48.5%), Ni$_2$O$_3$ (46.6), PbO, PbO$_2$ (46.7%), CdO (44.7%), V$_2$O$_3$ (43.5%), NiO (43.1%), CuO (41.6%), TiO$_2$ (27.0%), ZnO (9.2%). In the presence of Al$_2$O$_3$, MoO$_3$, WO$_3$, SnO, SnO$_2$, Sb$_2$O$_3$, Sb$_2$O$_4$, and Sb$_2$O$_5$, oxidation of NH$_3$ to NO was not observed at all. Hence it is evident that the catalytic activity of solids is governed predominantly by their chemical properties, and not by the method of preparation. Thus, for example, substances of the compositions Cr$_2$O$_3$, CdO, CuO, PbO, on the one hand, and Al$_2$O$_3$, ZnO, and SnO, on the other, although prepared under identical conditions, differ sharply, as we have seen above, in their catalytic properties. Another example confirming this conclusion is provided by the catalytic properties of nickelous and nickelic oxides (NiO and Ni$_2$O$_3$). The former was calcined to a temperature of 1000°C, and the latter to 250°C. However, only at the first moment, in the presence of nickelic oxide, did the degree of oxidation of NH$_3$ to NO (49%) exceed that in the presence of nickelous oxide (43%). But after the nickelic oxide, as a result of thermal decomposition during contact in the contact apparatus, had decomposed to nickelous oxide, the degree of oxidation of NH$_3$ in the presence of both catalysts became almost the same and reached 33–35%.

Comparison of the catalytic activity of oxides with the magnitude of their surface area shows that the decisive influence on the catalytic activity of substances is exerted by their chemical composition, and not by the surface area. Thus, for example, the specific surface of substances (Table 2) of composition SnO$_2$, WO$_3$ is greater than that of catalysts of composition MnO$_2$, Cr$_2$O$_3$; however, in the presence of the former oxidation to NO did not occur at all, while in the presence of the latter it reached, respectively, 78.8 and 42.3%. Therefore, in further searches for nonplatinum catalysts, attention must first of all be paid to their chemical composition, and then to the magnitude of the surface area.

Table 2

Specific surface areas of oxides

On the basis of experimental data we found equations establishing the dependence of the degree of oxidation ($\alpha$) of NH$_3$ to NO on the temperature (T°K), the space velocity (V) of the gas mixture, the concentration of NH$_3$ ($C_{\mathrm{NH}_3}$) in the gas, and on the pressure (P).

1. The catalytic activity of oxides in the oxidation of \(NH_3\) under pressure is influenced predominantly by their chemical composition, and not by the method of preparation.

2. For the reaction of oxidation of \(NH_3\) to \(NO\) under pressure, oxides whose elements possess hole conductivity have an advantage.

3. The greatest activity with respect to the reaction of oxidation of \(NH_3\) to \(NO\) is shown by oxides that contain transition elements with \(4s\) electrons and unfilled \(3d\) shells (elements of the 4th and 5th rows of D. I. Mendeleev’s table according to Shchukarev).

4. Empirical equations are given that establish the dependence of the degree of oxidation of \(NH_3\) to \(NO\) on pressure, space velocity, temperature, and \(NH_3\) concentration.

5. The data presented provide some guidance in the development of more active mixed non-platinum catalysts.

LITERATURE

1. Pisazhevskii L. V. ZhRFKhO, 61, 9, 1929.

2. Problems of Kinetics and Catalysis, Publishing House of the Academy of Sciences of the USSR, VIII, 1955.

3. Johnston, Houvouras, Schowalter, Ind. Eng. Chem., 46, 702, 1954.

4. Elovich S. Yu., Zhabrova G. M., Margolis L. Ya. and Roginskii S. Z. DAN SSSR 52, 421, 1946.

5. Roginskii S. Z., Sapozhnikov L. M., Kucherenko N. N. Ukr. Khim. Zhurnal 4, 99, 1928.

6. Ioffe A. F. Semiconductors in Modern Physics, Publishing House of the Academy of Sciences of the USSR, 1954.

7. Vol’kenshtein F. F. Electrical Conductivity of Semiconductors, Gostekhizdat, 1947.