Chemistry

A. M. Rubinshtein, Yu. A. Eltekov, and K. I. Slovetskaya

Chemisorption of Isopropyl Alcohol on Catalysts—Ferroalumogels

(Presented by Academician A. A. Balandin, 19 IV 1958)

To characterize the activity and selectivity of oxide catalysts, the decomposition reaction of isopropyl alcohol is often used as a standard. This reaction can proceed in two directions: dehydration (for example, on Al₂O₃) and dehydrogenation (for example, on metals or oxides, including Fe₂O₃). A similar study, which is the subject of a separate communication, was carried out in our laboratory with ferroalumogels. It was therefore of interest to study the adsorption of isopropyl alcohol on Fe₂O₃·Al₂O₃ catalysts, on which both of these reactions take place; this was done in the present work. The study was carried out by the vacuum method on a McBain balance at 30°. The initial samples were prepared by coprecipitation of Al(OH)₃ and Fe(OH)₃ with a 10% aqueous solution of NH₄OH. The precipitates obtained were washed, dried, and molded into tablets (diameter 1.2–1.3 mm). Separate portions of these catalysts were heated for 7 hours at 400 and 600°. Before the experiments, the samples were evacuated to a pressure of 10⁻⁵ mm Hg and a temperature of 300°. Under analogous conditions, weighed portions of the same samples were evacuated and then calcined in air at 1250° in a crucible furnace in order to determine the loss in weight. This loss in weight was attributed to removal of structural water, the content of which in the experimental samples is given in Table 1.

Table 1

Since the molecular areas and the behavior of i-C₃H₇OH on the catalysts studied were not known, the values of the specific surface area

Table 2

The samples were calculated by the BET method from adsorption isotherms of benzene vapor at 20°, taking the area per molecule in a close-packed monolayer to be 41 Ų. From these isotherms the pore diameters and the total pore volume were calculated for each sample. The measurements showed that all the samples studied have comparatively wide pores, on the order of 60–100 Å; the exceptions were samples calcined at 400° and containing 8 and 46 mol.% Fe₂O₃, for which the pore diameter was somewhat smaller—about 50 Å. The values of the specific surface area are given in Table 2.

In experiments with absolute \(u\)-C₃H₇OH, the influence of the composition and the conditions of thermal treatment of the catalysts on its adsorption was studied. Figure 1 presents experimental isotherms of adsorption and desorption of \(u\)-C₃H₇OH on catalysts with different ratios of Fe₂O₃ and Al₂O₃, calcined at 400 and 600°. Desorption is indicated by black points. It is evident from Fig. 1 that isopropyl alcohol at 30° is chemisorbed on the surface of all the samples studied. The composition of the catalysts has little effect on the magnitude of chemisorption, which depends much more strongly on the magnitude of their specific surface area. This indicates that in the surface layer of the catalyst there is either only one component, or both components, chemisorbing isopropyl alcohol to the same extent. However, the first assumption proves untenable, since the catalysts were prepared by coprecipitation. The second assumption is confirmed by phase-analysis data, which showed that the catalyst components are dissolved in one another, forming two solid-solution phases: 1) up to 8 mol.% Fe₂O₃ in Al₂O₃ and 2) up to 25 mol.% Al₂O₃ in Fe₂O₃. Figure 1 also shows that increasing the calcination temperature of each catalyst leads to a decrease both in the total adsorption of \(u\)-C₃H₇OH and in its chemisorbed amount. This latter circumstance should be taken into account in elucidating the nature of chemisorption, since with more severe heat treatment: a) the water content decreases, and b) the specific surface area of each catalyst decreases. It is necessary to discuss the question of which surface groups undergo chemisorption.

Fig. 1. Sorption isotherms of isopropyl alcohol on Fe₂O₃·Al₂O₃ catalysts. Fe₂O₃ content in the catalyst (mol.%): 1—8, 2—18, 3—32, 4—40, 5—46, 6—57. Calcination temperature: \(a\)—400°, \(b\)—600°.

In 1950, Kiselev and co-workers \(^{(1)}\) discovered irreversible sorption of CH₃OH by silica gel as a result of the formation of a surface ether. In the work of Topchieva and Ballyod \(^{(2)}\), it was established that, on interaction with aluminum oxide and an aluminosilicate catalyst, methyl alcohol forms analogous surface compounds through a surface reaction with hydroxyl groups according to the scheme:

\[ \mathrm{CH_3OH + H{-}O{-}Al <\!\!-\!\!> H_2O + CH_3{-}O{-}Al <} \]

Terenin and co-workers \(^{(3,4)}\), and later Babushkin and co-workers \(^{(5,6)}\), studying the infrared spectra of substances adsorbed on glasses, silica gel, and aluminum oxide, showed that, during adsorption of alcohols, there is a disappearance of the frequency characterizing unperturbed hydroxyl groups and the appearance of a frequency characterizing the C—O—M bond in bulk alco-

alcoholates. By optical methods, the formation of a surface complex due to the hydrogen bond

\[ \begin{array}{c} \mathrm{M{-}O{-}H}\\ \ \ \vdots\ \ \vdots\\ \mathrm{H{-}O{-}R} \end{array} \]

was also detected. In the case investigated by us, the strong adsorption of \(i\)-\(\mathrm{C_3H_7OH}\) indicates the formation of a surface alcoholate of the type: \(\mathrm{R{-}O{-}Me}\), which under catalysis conditions (with increasing temperature) can decompose, giving propylene.

Thus, it may be concluded that at room temperature chemical adsorption of isopropyl alcohol occurs on the surface of the catalysts, with the formation of surface alcoholates.

In considering the mechanism of chemisorption of isopropyl alcohol we proceeded from the assumption that hydroxyl groups are present on the catalyst surface. It is obvious that the concentration of OH groups on the surface depends on the composition of the bulk phase and on thermal treatment. From Table 1 it is seen that the water content in the catalyst decreases with increasing treatment temperature and with increasing \(\mathrm{Fe_2O_3}\) content; moreover, the amount of water decreases faster than the value of the specific surface area (cf. Tables 1 and 2). This indicates that the concentration of OH groups per unit surface area of the catalyst decreases under these changes. In this case, it would seem, a decrease in the fraction of chemical adsorption should also occur. From the desorption kinetics, we estimated the fraction of chemisorption from that part of the sorbed amount of alcohol which is not removed by evacuating the samples at room temperature and a pressure of \(5 \cdot 10^{-5}\) mm Hg for 2 h. This amount was calculated per unit surface area in order to exclude the factor of the magnitude of the surface itself, which changes both upon thermal treatment and as a function of the \(\mathrm{Fe_2O_3}\) content in the catalyst.

Table 2 gives the values of the fraction of chemisorption (\(a_x\)) and the concentration of OH groups, calculated on the assumption that all water is present on the surface in the form of OH groups. This table also includes the values \(\omega_0\) of the areas occupied by an isopropanol molecule in a monolayer on the surface of the catalyst. These values were calculated from the formula \(\omega_0 = \dfrac{s}{a_m N}\), where \(s\) is the value of the specific surface area of the catalyst, calculated by the BET equation from the adsorption isotherms of benzene vapor; \(N\) is Avogadro’s number; and \(a_m\) is the capacity of the monolayer of chemisorbed isopropanol, calculated by the same equation for the desorption branch of the sorption isotherm of \(i\)-\(\mathrm{C_3H_7OH}\).

From Table 2 it is seen that the amount of chemisorbed isopropanol remains practically unchanged at about \(4\ \mu\mathrm{mol}/\mathrm{m}^2\) and, to a first approximation, does not depend on the composition or the treatment temperature. Similarly, the concentration of OH groups on samples calcined at \(600^\circ\) also does not depend on the composition and is \(6.7\ \mu\mathrm{eq}/\mathrm{m}^2\). At the same time, this concentration in samples heated at \(400^\circ\) decreases with increasing \(\mathrm{Fe_2O_3}\) content and proves to be greater than in analogous samples calcined at \(600^\circ\). This probably indicates that there is a sufficient amount of hydroxyl groups on the catalyst surface for chemisorption to proceed. If one compares the effective cross-sectional areas of the isopropanol molecule (\(40\ \text{\AA}^2\)) and of the OH group (\(12\ \text{\AA}^2\)), it may be concluded that even upon removal of 60% of the surface OH groups covering the catalyst particles with a dense monolayer, the chemisorbed isopropyl alcohol can cover the entire surface, since the cross section of the alcohol molecule is three times greater than the cross section of the OH group. The literature also gives a value for the capacity of \(1\ \mathrm{m}^2\) for saturation with OH groups of \(\alpha = 13\ \mu\mathrm{eq}/\mathrm{m}^2\) for the surface of crystalline quartz and silica gel \((^7)\). Only for one sample is this value exceeded. In the remaining cases the calculated capacity values are less than \(13\ \mu\mathrm{eq}/\mathrm{m}^2\), and therefore for the majority

it is characteristic of the samples that there are present on the surface not only hydroxyl groups, but also oxygen atoms attached to the metal according to the scheme:

$$ \begin{array}{cccccccc} & \mathrm{HO} & & \mathrm{O} & \mathrm{H} & & \mathrm{O} \\ & | & & / \ \backslash & | & & \| \\ & \mathrm{M} & \mathrm{M} & & \mathrm{M} & & \mathrm{M} \\ / & & \backslash & / & \backslash & & / \\ \mathrm{O} & & \mathrm{O} & & \mathrm{O} & & \mathrm{O} \end{array} $$

N. D. Zelinskii Institute of Organic Chemistry
Academy of Sciences of the USSR

Received
17 IV 1958

REFERENCES CITED

1. O. M. Dzhigit, N. N. Mikos-Avgul, A. V. Kiselev, K. D. Shcherbakova, DAN, 70, 441 (1950).
2. K. V. Topchieva, A. P. Ballod, DAN, 75, 247 (1950).
3. A. N. Terenin, N. G. Yaroslavskii, DAN, 66, 885 (1949).
4. A. N. Terenin, in: Surface Chemical Compounds and Their Role in Adsorption Phenomena, Moscow, 1957, p. 206.
5. A. A. Babushkin, L. A. Ignat’eva, N. N. Grigor’ev, ibid., p. 246.
6. A. A. Babushkin, A. V. Uvarov, DAN, 110, 581 (1956).
7. A. V. Kiselev, G. G. Muttik, Colloid Journal, 19, 562 (1957).