W.Keith Hall

Hysteresis in the Palladium—Hydrogen System

When hydrogen is dissolved in palladium, the α phase forms first; the β phase nucleates and grows in this. Because of the large volume change which accompanies the α‐to‐β transition, plastic deformation of the α phase occurs as it is stretched beyond its elastic limit. Hence, absorption is accompanied by growth of the β phase under compressive stress of the α phase, but desorption take place from the β phase of the plastically deformed solid. Since this is not a thermodynamically reversible process, hysteresis results. It is shown that the absorption loop of the isotherm can be calculated from the desorption loop and the yield strength of the metal. Some reinterpretation of earlier theories is presented.

Studies of the hydrogen held by solids. XXI. The interaction between ethylene and hydroxyl groups of a Y-zeolite at elevated temperatures

The infrared spectra of the OH groups of a hydrogen-Y-zeolite have been recorded in the presence of C2H4 at pressures up to 1 atm and in the temperature range 25 to 150 °C. From these data, the heat of formation of hydrogen-bonded C2H4 was estimated to be about 9 kcal mole−1 for interaction with the 3648 cm−1 groups. Isotopic exchange between C2D4 and the two principal types of OH groups was studied at temperatures between 200 and 350 °; both exchanged at about the same rate. Equilibration of C2D4 and C2H4 proceeded via exchange with the surface OH groups. C2D4OH exchange took place 1.2 times faster than C2H4OD exchange over zeolite samples pretreated in equivalent ways. The kinetics deviated slightly from a first order dependence on the distance from equilibrium. The apparent first order rate constants reached an apparent maximum for pretreatment at 290 °C. The apparent activation energy for exchange was 16 kcal mole−1 for samples pretreated at 290 °C and 19 kcal mole−1 for those heated at 475 °C in O2 and rehydrated at 350 °C. In both instances, the reaction was first order in C2D4 pressure.

Studies of the hydrogen held by solids IV. Deuterium exchange and NMR investigations of silica, alumina, and silica-alumina catalysts

Abstract Rising temperature D 2 exchange experiments were used to characterize the several kinds of hydrogen held by solids. Results are presented which demonstrate that the temperature region (activation energy) for exchange increases in the order: alumina, silica-alumina, and silica. All of these substances have about the same surface density of terminal hydroxyl groups following overnight evacuation at 500 °C; these values fell between 1.5 and 4.5 × 10 14 OH/cm 2 . Within the ±20% experimental error, identical values were obtained for a particular silica or silica-alumina catalyst from integrated intensities of proton magnetic resonance absorption spectra. The spectra from silica-alumina were qualitatively indistinguishable from those of silica gel. Both sets of data led to the conclusion that silica-alumina is not an intimate mixture of microcrystals of silica and alumina grown together at contact areas, but a more nearly homogeneous substance. This information, together with infrared absorption data taken from the literature, is used to show that the largest fraction, and perhaps all, of the hydrogen of dehydrated silica-alumina closely resembles that of silica gel, i.e., it is chemically similar to alcoholic hydrogen. There is no evidence that any portion should be classified as acidic. The possibility that up to 20% of the total hydrogen exists as acidic AlOH could not be excluded, however, as this is the accuracy of the NMR data. Therefore, an upper limit for Bronsted acidity of 13 H + /cm 2 is suggested. The hydrogen studied was identified with that involved in well-known exchange reactions with hydrocarbons.

Studies of the hydrogen held by solids: VI. The hydroxyl groups of alumina and silica-alumina as catalytic sites

Abstract The microcatalytic technique was used in evaluating the ability of silica-alumina and alumina catalysts to crack 2,3-dimethylbutane and to isomerize cyclopropane; observations were also made of the coke formation during cracking and of the chemisorption of ammonia at several temperatures. Hydrogen contents (measured by exchange with D 2 ) were varied by carefully controlling the dehydration conditions and by substitution of fluorine for the hydroxyl groups of alumina. “Base exchange” of silica-alumina with alkali metal ions lowered the catalytic activity but did not materially affect the hydrogen content. With catalysts dehydrated at temperatures above the cracking range, the rate of decomposition of 2,3-dimethylbutane was independent of hydrogen content, but the activities of catalysts dehydrated at only 550 ° were somewhat higher. With alumina, the specific activities for cracking actually increased slightly as the hydrogen content was lowered. Alumina was as active as silica-alumina for cracking. Apparently the principal difference in the behavior of these two materials is that alumina is more readily poisoned by traces of H 2 O and/or reaction products than silica-alumina. Substitution of the hydroxyl groups of alumina with fluorine greatly enhanced its activity and otherwise rendered it more like silica-alumina. Samples of fairly pure eta and gamma alumina, as well as members of a series of mixtures of eta and gamma alumina, were found to vary in activity independently of hydrogen content; eta alumina was the more active. The amount of ammonia chemisorbed per unit area in the range of 175 ° to 500 ° was not a function of the catalyst hydrogen content. When the alumina or silica-alumina was extensively dehydrated, several times more ammonia was chemisorbed, even at 500 °, than the amount of residual hydrogen. It is therefore apparent that the strong acid sites presumed to chemisorb ammonia are not of the Bronsted type. The rate of isomerization of cyclopropane correlated with the hydrogen contents of alumina and of silica-alumina catalysts. In both cases, the activities could be further increased by added-back water; with silica-alumina, the rate increased continuously with H 2 O content, but a maximum activity occurred with alumina. As with cracking, base exchange of silica-alumina or substitution of F for OH in alumina decreased and enhanced the activity, respectively. Evidently, the two reactions proceed by different mechanisms although possibly involving the same aprotic sites.

The formation of cation radicals on the surface of silica-alumina catalysts

Abstract Optical and electron paramagnetic resonance (EPR) spectra have been obtained for a number of polynuclear aromatic hydrocarbons and phenylated amines adsorbed on a silica-alumina catalyst. Both compound types react to form cation radicals on the surface. The results for the amines suggest that the electrophilic centers involved are identical with those required for carbonium ion type reactions. The maximum number of paramagnetic species formed per unit area of surface was found to be in fair agreement with an earlier estimate of Leftin and Hall (1), obtained for triphenylmethane and related compounds. However, it was also found that the spin density was dependent upon the oxidation state of the catalyst surface. A possible relationship of the present results with those for phenylated olefins is pointed out.

The microcatalytic technique applied to a zero order reaction: The dehydration of 2-butanol over hydroxyapatite catalysts

Abstract A comparison has been made between microcatalytic and steady state flow reactor data for the dehydration of 2-butanol over a series of hydroxyapatite catalysts. The conventional steady state flow experiments revealed that the reaction was zero order in 2-butanol, but weakly inhibited by the product water. Unlike first order reactions, where identical results are derived from the two techniques, the microcatalytic results were not identical with those from the steady state experiments. Instead, they indicated that the surface-adsorbed species were not in equilibrium with the gas phase, and that products continued to desorb long after the gas pulse had passed. Thus, plots of conversion vs. reciprocal flow rate did not extrapolate to zero at infinite flow rate. Instead, the finite conversion intercept was a measure of the “site monolayer” from which products decsorbed very slowly. This fact was used to count the active sites for dehydration. Together, the two methods yield complementary information, not obtainable from one method alone. The design of a versatile, high-temperature, microcatalytic reactor is given.

Studies of the hydrogen held by solids: IX. The hydroxyl groups of alumina and silica-alumina as sites for the isomerization of butene

The rates of double-bond isomerization of butene-1 and cis-butene-2 have been measured over alumina and silica-alumina catalysts as a function of their hydrogen contents (comprised principally of terminal surface OH groups). Changes which occurred as the catalysts “lined-out” are described. The isomerization rate of butene-1 increased with hydroxyl concentration over silica-alumina, but decreased over alumina. Over silica-alumina, the ratio of cis- to trans-butene-2 was near unity in the initial product; over alumina, higher ratios were obtained. Both the activity and selectivity of alumina were increased by fluoriding. cis-Butene-2 and butene-1 isomerized at about the same rate over alumina while the latter was converted much faster than the former over silica-alumina. It was concluded that the mechanism of isomerization, and the intermediates involved, were not the same over alumina as over silica-alumina. Mechanisms consistent with the available literature are considered.

Studies of the hydrogen held by solids. Part 15.—Cation exchanged zeolites

Dehydration of multivalent ion-exchanged zeolites produced structural hydroxyl groups of varying thermal stability as evidenced by differences in the retention of their i.-r. bands on heating in vacuo. The retention increased with the electron affinity of the base-exchange cation. Although the i.-r. bands in the 3650 and 3545 cm–1 regions appeared to have the same chemical identity as those formed by heating the NH+4-zeolite, these hydroxyl groups were less reactive with pyridine (Py). Temperatures of 85–150° were required to react a substantial fraction of them to PyH+, whereas with the decationated zeolites, both bands were removed by Py interaction at room temperature although the 3545 cm–1 band could be restored to its original intensity by evacuation at 150°. Py also co-ordinated with the base-exchange cations (became Lewis bonded). The strength of this interaction (retention of PyL at elevated temperatures) increased with the electron affinity of the cation and the frequency of the corresponding i.-r. band increased concomitantly from 1443 cm–1(Na+) to 1455 cm–1(Zn2+). Addition of H2O, or H2O + CO2, effected conversion of PyL to PyH+, and again the extent of these reactions paralleled the electron affinity of the cation.

Studies of the hydrogen held by solids. Part 17.—The ortho-para H2 Conversion and H2–D2 exchange reactions over a transition alumina

The rates of the H2–D2 exchange, the o-p H2 and p-o D2 conversions were measured on the surfaces of a very pure alumina, with and without a portion of the sites selectively poisoned with CO2. These data were supplemented with results obtained from surface dehydroxylated to different extents. When the alumina was dehydroxylated below 525°, the rate of the ortho-para H2 conversion at 0° was only a little higher than the rate of the H2–D2 exchange. Both rates fell in a commensurate manner as the surface was poisoned with CO2(in a way which eliminated the strongest sites first) to less than 1 % of their initial values at about 16 × 1012 CO2/cm2. At –195°, the exchange reaction was immeasurably slow, while both conversion reactions (H2 and D2) had comparable and easily measurable rates, the D2 conversion being a little faster. These conversion rates fell as the surface was poisoned with CO2 to less than 1 % of their initial values at 18 × 1012 CO2/cm2. Both reactions were also poisoned by lowering the dehydroxylation temperature below 300°. After pretreatment at 800°, the site density for the conversion reactions increased to about 30 × 1012 CO2/cm2; that for the exchange reaction fell to about 9 × 1012CO2/cm2. n Arrhenius plots for the H2–D2 exchange were linear on poisoned and unpoisoned surfaces and indicated an activation energy of about 2.3 kcal/mol. The Arrhenius plots for the conversion reactions were non-linear and varied with the extent of poisoning with CO2. In all cases, however, they converged with the plots for the exchange reaction above –80°, but curved to lower slopes (lower activation energies) at lower temperatures. Evidently, the conversion and exchange reactions both proceeded by the same dissociative mechanism at high temperatures but only the paramagnetic component of the conversion reaction was significant near the temperature of liquid nitrogen. An extensive search was made, but no e.p.r. signal could be found even though the site density determined by CO2 poisoning was ∼1019/g. It is suggested, therefore, that the conversion reactions are catalyzed by interaction with exposed 27Al nuclei.

Studies of the hydrogen held by solids: X. Fluorided aluminas as acid catalysts

Abstract The effects of fluoriding alumina on hydroxyl content, cracking of 2,3-dimethylbutane, isomerization of cyclopropane, and chemisorption of ammonia were determined. As the fluorine content was increased from 0% to 6.0%, the hydroxyl concentration decreased from 3 × 10 14 /cm 2 to 0.06 × 10 14 /cm 2 . The rates of cracking and isomerization attained maximum values at 2.7% and 1.2% F, respectively; the corresponding rates increased by factors of 25 and 4000. The chemisorption of ammonia at 500 ° reached a maximum at 1.7% F. The close proximity of these maxima suggests a common cause; several such possibilities are discussed. For the reactions studied, the specific activity of crystalline aluminum fluoride was not greatly different from that for alumina containing 6.0% F. Moreover, AlF 3 was evolved when the fluorided aluminas were heated to 1040 °C. These observations are consistent with the idea that, as the fluorine concentration increased, the catalyst surface became progressively more like that of aluminum fluoride.