Heterogeneous Catalysis

---

CHEMICAL RATE PROCESSES

Heterogeneous Catalysis M. BOUDART Princeton University, Princeton, N. J.

URRENT problems of interest in contact catalysis are centered around the following questions. Are all equivalent

made (11B) to correlate such data with variations in heats of adsorption. According to this analysis, the fall in adsorption heat with coverage can be predicted, in order of magnitude, if the strength of the surface double layer due to adsorbed species ie known. The effect of the latter would then seem t o be predominant, whereas heterogeneity or lateral interaction would be responsible for second-order changes in heats of adsorption at low or high coverage, respectively. The difficulty of obtaining clean iron surfaces is also emphasized by Tompkins (88B)whose criterion of purity is provided by the accommodation coefficient technique. The latter is reliable ( W B ) but not sensitive enough for the study of chemisorption processes (88B). A study of silver surfaces by means of the weighed wire method (16B)illustrates the well-known tenacity of oxygen films: measurements of surface tension of solid silver at partial pressures of oxygen between 0.0001 and 0.2 atmosphere and the Gibbs adsorption isotherm indicate that between 870" and 942' C. there are approximately 1.4 atoms of adsorbed oxygen per silver atom a t the surface. Extrapolation of the data a t 932" C. shows that the film would start evaporating a t a partial pressure of oxygen equal to atmospheres, whereas the decomposition pressure of solid silver oxide a t the same temperature is 1.7 X lo4 atmospheres. A different picture emerges from recent work on carbon surfaces. Following Duval ( 2 2 B ) , a carbon filament cleaned by flashing a t 2500' C. is not contaminated by oxygen under a pressure of about mm. between 850' and 1700' C. This result, obtained by measurements of thermionic emission, is confirmed by two groups of workers working independently on the oxidation reactions of very pure carbon. Gulbransen and *4ndrew (34B, 35B) show that "surface oxides" disappear above 700" C. in high vacuo. Their formation is not essential to the carbon dioxide reaction on carbon, between 500" and 900" C. Similar conclusions are reached by Wynne-Jones et al. (99B). The same authors also report strong catalytic effects due to impurities. For instance, pure carbon does not react a t 700" C. with carbon dioxide, but if 0.079% of iron is added a very strong reaction takes place. The iron is believed to be elemental or present as carbide since no catalytic effect is shown by iron oxide before its reduction by hydrogen (35B). In the same connection, the data of Long and Sykes (6.93) relative to the carbon-steam reaction and the carbon monoxide-water conversion on carbon, imply a L1specific interaction" between carbon and its metallic impurities. For the conversion, however, a slow decrease of reactivity at high hydrogen partial pressures suggests that the catalytic impurities are oxides, the reduction of which by hydrogen is detrimental to the process. More data on carbon oxidation can be found in papers by Wicke (97B) and Reif (68B). If there is no rapid surface migration of the adsorbate. the question of heterogeneity can be decided in principle by isotopic exchange techniques. Using radioactive carbon, Eischens ( 2 4 B ) studied exchange between adsorbed carbon monoxide and carbon monoxide in the gas phase. His kinetic data can be interpreted if his iron surface consists of a small number of patches which differ from each other but are homogeneous within themselvee with respect to the exchange process. More work along the same lines led Eischens and Webb (65B) to a striking observation:

C

sites on a given surface equally entitled to work in a catalytic reaction or is it necessary to prepare and maintain a few privileged active centers which solely contribute to the process considered? What is the nature of the slow uptake of gas observed in the overwhelming majority of studies on rates of chemisorption? When a given reaction is run over a series of catalysts, is it possible to correlate apparent activation energies and frequency factors vith structural parameters of the solid (geometrical factor) or is it more fruitful to take into account variations in bond properties with the help of available knowledge in solid state physics (electronic factor)? What can be said about the existence of a "catalytic domain"-Le., a critical particle size manifesting itself when activity is measured as a function of the degree of dispersion on a support? With polyfunctional catalysts, such as those used for Fischer-Tropsch and related syntheses or in hydroforming processes, what are the factors determining selectivity, that is, the activity of each particular function? The papers covered in this review present diversified answers to these questions. Except for a few foreign contributions of late 1951 and American papers of early 1953, they were published during 1952. Homogeneity or Heterogeneity. Conceivably, this question is best studied on clean surfaces which retain the attention of many investigators. Crowell ( I Q B )has given a review of previous work relative to adsorption of gases on metal filaments, evaporated films, and single crystals. Adsorption details which formerly could be inferred only by indirect means-e.g.. mobility of the adsorbate-may now be observed directly by means of Muller's field emission microscope ( 7 B ) . This instrument also provides the most exacting criterion of surface purity. Thus, Schleicher (76B) observes a pattern on an iron point that is probably due to a surface oxide (Fes04?)which cannot be removed by reduction in hydrogen during 30 minutes between 750" and 850' C. The resolution power of the projection microscope has been examined by Gomer ( S I B ) . This problem is important from the viewpoint of adsorption, since claims have been made ( 4 B ) that the instrument is capable of showing the dissociation of adsorbed oxygen molecules into atoms. Gomer and Speer ( S 2 B ) studied the current emerging from single molecules of Bine phthalocyanine on tungsten. They see irregularities of 5 to 10 A. on the best and cleanest of their surfaces and suggest that their findings support the essentially heterogeneous nature of surfaces. A new mass spectrographic technique that permits in principle the determination of positive ion work functions has been developed by Brady and Zemany ( I S B ) . They measure K + ions desorbing from platinum surfaces around 800" C. and find that hydrogen increases the desorption rate while oxygen, carbon tetrachloride, bromine, and helium decrease the latter. The effect of helium is hard to understand. Suhnnann (84B),on the other hand, uses the photoelectric effect as a means of studying adsorption of hydrogen and oxygen on platinum surfaces. Contact potential differences (66B)provide additional information on the polarity of chemisorption bonds and an attempt has been

898

May 1953

INDUSTRIAL AND ENGINEERING CHEMISTRY

Using again a reduced iron surface but two differently labeled carbon monoxides ( W O and C**O),they were able to show, by desorbing W1*O, that significant exchange between carbon and oxygen takes place even a t -33’ C. and becomes practically complete around 150’ C. Obviously, chemisorption of molecules susceptible to dissociation at the surface is a complex process, the details of which are still largely unknown. Activated Adsorption, Surface Migration, Bulk Diffusion. Thus, while slow sorption phenomena can be observed on

all types of surfaces, the nature of the slow step is still obscure in particular cases. A new technique adaptable to the study of very slow adsorption processes consists in measuring-e.g.,’ by means of a Geiger counter-the electrons emitted from a metal surface as a result of the formation of chemisorption bonds (33B, S8B) or other changes in surface structure (4IB). Paproth, Rathje, and Stranski (62%) measured the electron current emitted during chemisorption of iodine molecules. The exponential rise of the emission with temperature suggests a whole spectrum of activation energies for adsorption (see also S8B). When iodine is adsorbed on the silver surface, it becomes partially dissociated and a metastable Ag-I bond is formed. After some time, 1, the latter may become a normal chemisorption bond Ag+-I- by electron transfer and dissociation of the molecule. But during time t, the Ag-I bond is bombarded by metallic electrons. A collision of the second kind may then take place and the energy liberated by the transition Ag-I Ag+-I- goes to a metallic electron which now has enough energy to leave the metal. The estimated electron yield is circa 0.07y0. Another interesting approach to the problem of activated dissociative chemisorption is that of Schiifer (76B) who measurea accommodation coefficients of ethane on a series of platinum copper alloys. He finds a striking parallelism between catalytic activity of the alloys for dehydrogenation of ethane a t higher temperatures and the accommodation coefficients of the normal modes of vibration corresponding to the mutual approach of two hydrogen atoms from the two methyl groups. Moreover, a high catalytic activity is associated with an easy accommodation of these normal vibrations but a poor accommodation of the other normal modes. Otherwise, the energy required for the surface dissociation would be distributed over the entire molecule and although the activation energy might decrease, the chance of dissociation taking place during the adsoSption time would also become smaller. I n the considerations of Stranski and Schafer, the activation energy for molecular chemisorption is provided by the solid adsorbent. This picture is also favored by other workers. Taylor and Thon (86B)analyze a number of kinetic adsorption isotherms that can be represented by the relation dg. dt = ae-a4 where p is the

-

amount adsorbed a t time t. This relation could mean kinetically that adsorption sites, first produced by the gas itself, subsequently decay a t a bimolecular rate. Tompkins (88B)observes the same rate law for slow chemisorption of nitrogen on iron films and suggests that the measured activation energy corresponds to surface migration from site to site. The same mechanism is proposed by Trapnell (9OB) for nitrogen on tungsten. The thermal production of active centers (I%%) also accounts for the activation energy originating in the solid adsorbent. Bruns et al. (83B)show, however, that slow sorption of carbon monoxide on manganese dioxide, which can be represented by a relation of the type used by Taylor and Thon, corresponds to diffusion of lattice oxygen to the surface. This viewpoint is amplified by Rozen et al. (71B, 7dB) who analyze the cases where diffusion into a polydisperse solid leads to rate laws which are formally attributed to continuous distributions of energy sites, as can be found in a paper by Keier and Man’ko (&B). Still another position is defended by Schuit and de Boer (77B)who study hydrogen chemisorption on a reduced 3Ni-Si02 catalyst. They find a fast chemisorption process at -196O C. resembling that occurring on evaporated

899

films and a slow rate of gas uptake a t higher temperatures. The first type is favored by extensive reduction and decreased by oxygen chemisorption. The second type is described as taking place “on top” of oxygen remaining on the nickel surface after incomplete reduction, or introduced on purpose. A “two-type” surface is also invoked by Sastri and Ramanathan (74B)who extend to zinc-molybdenum oxide the Taylor-Liang effect of fast desorp tion followed by slow readsorption of hydrogen. Finally, “three-type” surface has been discovered by Kummer and Emmett (6OB): It is that of a reduced singly promoted iron catalyst on which a third kind of hydrogen chemisorption has been identified by adsorption isobars and isotherms and ita ability to perform the hydrogen-deuterium exchange reaction rapidly at -195” C. The inverse process-Le., hydrogen atom recombination on surfaces-has been analyzed theoretically by de Boer and van Steenis ( 8 B )and investigated experimentally by Voevodskii et al. (61B). A hydrogen atom hitting the surface combines with another adsorbed atom. Voevodskii finds that the recombination probability is particularly low on magnesium oxide. Geometrical and Electronic Factors. The relative importance of these factors of catalytic activity has been considered again by Dowden (BIB),whose account of the electronic structure of solids is particularly lucid and by Wmfield (98B),who uses molecular models to illustrate geometrical surface requirements. When a given reaction is studied on a series of catalysts, one of the dficulties in correlating the rate with some property of the solid appears to be the unknown state of the surface. For instance, McCabe a;d Halsey (64B) show that surface hydrogen is essen-, tial in maintaining reproducibility of the activity of a copper catalyst for ethylene hydrogenation. The role of surface hydrogen in maintaining the hydrogenation activity of Raney nickel is also emphasized by Freidlin and Rudneva (B7B),who deactivate or promote the catalyst by removing or readsorbing hydrogen. Similar conclusions are reached by Vandael (9dB) and Kefeli et al. (4SB). Rienacker et al. (69B),on the other hand, report activation energies, E, and activities per unit area for formaldehyde decomposition on a silver powder which is preheated a t increasing temperatufes. It is remarkable that the specific activity increases up to a preheating temperature of 600” C., whereas there is little change in tota.1 activity per unit mass. The activation energy increases from 16 to 23 kcal. per mole but the frequency factor also increases. A similar experiment for nitrous oxide decomposition on cupric oxide (I8B)shows that the activation energy increases linearly from 11 to 42 kcal. per mole with the temperature of pretreatment of the catalyst. The frequency factor, ko also increases by nine orders of magnitude. Vol’kenshtein (94B)has attempted to explain this simultaneous increase in E and ko, while Schwab (78B)has calculated by means of a simple electrostatic model the energy of various defects on different planes of ionic lattices of the sodium chloride type, A qualitative discussion of properties of semiconductors in chemisorption and catalysis (1OB)is illustrated by data of Parravano (&3B),who finds anomalies in the rate of carbon monoxide oxidation on several catalysts in the neighborhood of their ferroelectric transition. Garner, Stone, and Tiley (B8B)have studied further the catalytic oxidation of carbon monoxide on cuprous oxide a t room temperature. Combining kinetic, calorimetric, adsorption, and semiconductivity data, they conclude that the mechanism a t low temperature does not involve lattice oxygen and that the surface is predominantly covered with carbon monoxide, the most important steps being:

+ --

020

+

20ad8

cog toads cod, + 20ada CO~ada

COsods

COade

2CO2

The effect of a Schottky-Mott barrier layer at the free surface of a semiconductor on its adsorption properties and vice versa haye been considered by a variety of workers (IB, ldB, S7B, 96B).

3

INDUSTRIAL AND ENGINEERING CHEMISTRY

900

+

Kemball(45B, 46B) has studied the NHa Dz exchange reaction on a large number of metal films. The mechanism seems to be the same on all metals investigated, the frequency factor being about constant but the activation energy changing from 5.2 kcal, per mole on platinum to 14.1 kcal. per mole on silver. A correlation of the latter with metal work functions is put forward. Sheridan and Reid (8OB)present a new series of data on acetylene hydrogenation. Since the reaction is complicated by hydropolymerization, it is difficult to compare different catalysts but the authors conclude that all face centered cubic metals are appreciably active for the hydrogenation, whereas hexagonal closepacked structures are not. In order to elucidate the conditions favoring hydropolymerization, Bond and Sheridan ( 9 B ) have studied the hydrogenation of methylacetylene, allene, and their mixtures on nickel, palladium, and platinum. They observe a selective production of propene with small yields-due to steric hindrance-of reduced polymers. Their data seem t o support an associative adsorption mechanism of unsaturated hydrocarbons. T h a t catalytic hydrogenation of the latter is a very complex affair is shown by the work of Douglas and Rabinovitch (ZOB), whose first aim was to prepare cis-ethylene-& by catalytic hydrogenation of acetylene. The action of promoters has been thoroughly investigated by Brxechowski (61B), who worked with nickel-ceria catalysts for phenol hydrogenation. The various possibilities of catalyst modification which are well illustrated in this work are classified systematically by Roginslrii (?OB). Kiperman et al. (48B), on the other hand, on the basis of their kinetic interpretation of catalytic ammonia synthesis, tried to prepare a more active tungsten catalyst by reduction of the oxide with atomic hydrogen at low temperature. The greater yield of ammonia on this contact was attributed to a larger specific area. The kinetic scheme of Temkin was also verified by Kiperman et al. (47B), who showed that the rate of the synthesis could become first order in nitrogen as predicted by the theory on an osmium catalyst and by Livshits et al. (62B),who checked the formulas of Temkin and Pyzhev at pressures up to 500 atmospheres correcting for departure from ideal gas behavior. Polyfunctional Catalysts and Selectivity. A large number of contributions in the field of hydrocarbon catalysis are devoted to the study of the various functions of a given contact in complex reactions involving multicomponent mixtures. A detailed review of the work on syntheses from carbon monoxide-hydrogen mixtures has been given by Pichler (64B). Particular attention has been devoted to the mechanism of these processes by Emmett and his group (26B)and the workers a t the U.S. Bureau of Mines (,@E?). The Fischer-Tropsch synthesis is now held to proceed, not by carbide intermediates or polymerization of CHZ groups but by way of oxygenated intermediates resulting from carbon monoxide addition to a carbon chain followed by hydrogenation. An equation has been proposed (66B) for testing the hypothesis of constant growth probability of the carbon chain by stepwise addition of single carbon atoms. The special behavior of nitrided iron catalysts with respect to stability, activity, and alcohol production has been further investigated (@, 81B, 82B). Russian work (IOOB)shows, by means of a kinetic calorimetric technique, that during the first stages of reaction mixture of one part carbon monoxide to two parts hydrogen on cobalt catalysts, both carbon monoxide and hydrogen react and not only carbon monoxide, as was the case on a nickel contact. Rapoport and Levkovich (67B)present data t o prove that carbon dioxide originates in the HzO CO -.,COZ H* reaction on iron catalysts at atmospheric pressure. More details on the hydrocarbon synthesis from carbon monoxide-water mixtures have been given by Kolbel and Engelhardt (49B). I n this connection, i t is of interest to quote Mulford and Russell (6OB),who confirm the view that carbon monoxide is a critical intermediate in the reduction of carbon dioxide to oil and that methane can be produced by two mechanisms on cobalt catalysts, involving or not involving car-

+

+

Vol. 45, No. 5

bon monoxide as an intermediate. The Fischer-Tropsch synthesis has also been used to produce deuterated hydrocarbons from carbon monoxide-deuterium mixtures (87B). A thesis from the school of Prettre (9123) is concerned with activated adsorption of hydrogen and methane on nickel Fischer-Tropsch catalysts. Adsorption data of hydrogen and carbon monoxide on catalysts for methanol synthesis have been obtained by Ghosh et al. (29B). Dehydration or dehydrogenation of alcohols is a useful reaction for assessing catalyst selectivity. Rubinshtein et al. (75%) have shown how the activity and selectivity of chromia catalysts for dehydrogenation and/or dehydration of ethyl alcohol are modified by conditions of precipitation, temperature of pretreatment, and nature of the support. The mechanism of dehydration itself on alumina has been elucidated by Balaceanu and Jungers ( 6 B ) . These authors show that ether is an intermediate at not too high temperatures, and they propose a mechanism whereby ethylene and water are not formed directly from alcohol but in two successive reactions:

+ +

2CzH60H + O(CzHs)2 HzO CzHdOH o(CzH.5)~ CzHa --+

While the direct dehydration remains possible a t higher temperatures or for higher alcohols or with other catalysts, the two-step mechanism must be kept in mind because activation energies calculated from rate of ethylene formation might be greatly in error, owing to the intermediate production of ether. The classical Langmuir-Hinshelwood kinetic schemes are used in this work, as in other contributions of the Louvain school (6B). Antipina and Frost ( S B ) have measured adsorption isotherms of ethyl alcohol, ethyl ether, and watm between 150" and 310' C. on alumina heated at 550" C. and on a stabilized alumina catalyst. Their data are not represented by a Langmuir isotherm but by an expression that gives two straight lines in a plot of 1/x against l / p , where x and p are quantity adsorbed and pressure, respectively. Another test reaction that has been proposed for measuring the hydrogenating ability of a given catalyst is the hydrogen-deuterium exchange. A kinetic study of this reaction on a large variety of surfaces has been made by Holm and Blue (4OB), who found anomalous rates on chromia-alumina surfaces between 25 O and 300' C. in spite of a standard high temperature reduction which was quite adequate in other cases-e.g., for zinc oxide, about which anomalies of this kind have been reported before. Russian workers are still working on the mechanism of dehydrocyclization. The early ideas of Balandin have been set aside. Thus Ilagan et al. (4ZB) show that the rate of conversion to toluene at 484" C. on a Cr~Oa-KzO-Al20~ contact is faster for heptene than for heptane. On the same catalyst, deuterium exchange proceeds a t 200' C. in the caSe of the olefin but at 330' C. only for heptane. It is concluded that the first step toward aromatization is slower for paraffins than for olefins. Moreover, from studying the rate of conversion to toluene at 480" C. on vanadium oxide supported on alumina of the following mixtures: heptane-heptene, heptane-toluene, and heptene-toluene, Plate and Tarasova (66B) conclude that heptane and heptene aromatize on different centers of the same catalyst. Cracking catalysts of the silica-alumina type are still the object of many investigations. Thus new measurements of their Bronsted and Lewis acidity have been reported briefly (89B). After covering part of the surface of a cracking catalyst with heavy water, Hansford et al. (36B) study the deuteration of a variety of hydrocarbons exchanging with deuterium oxide. Since the exchange proceeds without measurable cracking, the degree of activation required for the exchange may not be sufficient for other subsequent reactions occurring in cracking. The mode of activation, however, may be the same. I n particular, the ionic intermediates produced by various catalysts from a given molecule may be qualitatively different in spite of having a common basis of initial formation. Similar conclusions are reached by

INDUSTRIAL AND ENGINEERING CHEMISTRY

May 1953

I”

Gladrow et al. (SOB). The mechanism of carbon deposition on various surfaces as well as the cracking activity of soot particles have been studied by Tesner et al. (66B,86B). When a metal is dispersed on a cracking support, the resulting polyfunctional catalyst possesses new selectivity for a variety of hydrocarbon reactions. I t s isomerizing capabilities are strikingly demonstrated in a series of papers by Ciapetta et al. (16B, I7B). The nonadditivity of the separate functions of such catalysts is also shown by the work of Mills et al. and Heinemann et al. (39B,67B). The fact that a nickel-alumina-silica catalyst, for instance, does not behave (16B)either as a nickel catalyst or as a silica-alumina catalyst for the isomerization of alkanes and cycloalkanes in the presence of hydrogen, raises the question as t o the nature of the dispersion of the metal on the oxide support. This problem has been considered by van Eijk van Voorthuisen and Frazen (96B)by means of x-ray diffraction and differential thermal analysis studies of nickel-silica catalysts. It is shown that the reduction in hydrogen leads to the formation of cubic metallic nickel but that, in certain cases, part of the nickel atoms remain a t the places they occupied in the nickel hydrosilicate prior to reduction. Further work on the problem of dispersion of oxides of transition metals on alumina has been pursued in the laboratory of Selwood (68B,69B,79B). Magnetic susceptibility data and measurements of catalytic activity for the oxidation of carbon monoxide and the decomposition of hydrogen peroxide are compared. The possibility of a minimum catalytic particle size or domain is again suggested. On the basis of earlier work by Selwood, Hungarian workers ( M B )conclude from BET surface measurements that the dispersion of molybdenum oxide on alumina is similar to that of chromia if the catalyst is prepared from paramolybdate, but that the kind of dispersion is different if molybdate is used. The art of catalyst preparation is well illustrated by this example.

901

Eischens, R.jP., and Webb, A. N., J . Chem. Phys., 20, 1048 (1 952).

Emmett, P. H., and Kummer, J. T., Petroleum Congr., Proc. 3rd. Congr., Hague, 1951, Sect. IV, p. 25. Freidlin, L. Kh., and Rudneva, K. G., Doklady Akad. N a u k . S.S.S.R., 83,105 (1952). Garner, W. E., Stone, F. S., and Tiley, P. F., Proc. Roy. SOC. (London), A211,472 (1952).

Ghosh, J. C., Sastri, M. W. C., and Kanath, G. S., J . chim. phys., 49, 500 (1952).

Gladrow, E. M., Krebs, R. W., and Kimberlin, C. N . , Jr., IND. ENG.CHEM.,45, 142 (1953). Gomer, R., J. Chem. Phys., 20, 1772 (1952). Gomer, R., and Speer, D. A., Ibid., 21, 73 (1953). Grunberg, L., and Wright, X. H. R., Nature, 170, 456 (1 952).

Gulbransen, E. A., IND. ENQ.CHEM.,44,1045 (1952). Gulbransen, E. A., and Andrew, K. F., Ibid., 44, 1034, 1039, 1048 (1952).

Hansford, R. C., Waldb, P. G., Drake, L. C., and Honig, R. E., Ibid., 44, 1108 (1952). Hauffe. K.. and Enaell. H. J.. 2. Elektrochem. 56. 366 (19521. Haxel, O., Houterm-ana, F. G., and Seeger, K., 2.Physik, 136, 109 (1951).

Heinemann, H., Mills, G. A., Hattman, J. B . , and Kirsch, F. W., IND. ENQ.CHEM.,45, 130 (1953). Holm, V. C. F., and Blue, R. W., Ibid., 44,107 (f952). Houdremont, E., and Rildinger, O., Naturwisaensohaften, 39, 399 (1952).

Kagan, M. Ya., Erivanskaya, L. A,, and Trofinova, I. V., Doklady Alcad. N a u k . S.S.S.R.,82, 913 (1952).

I

Kefeli, L. M., and Lel’ohuk, 5.L., Ibid., 83, 697 (1952). Keier, N. P., and Man’ko, N. M., Ibid., 83, 713 (1952). Kemball, C . , Proc. Roy.SOC.(London),A214,413 (1952). Kemball, C., Trans. Faraday SOC.,48,254 (1952). Kiperman, S. L., and Granovskaya, V. Sh., Zhur. Fiz. Khim., 25, 557 (1951).

Kiperman,‘S. L.’, Rybakova, N. A., and Temkin, M. I.,Ibid., 26, 621 (1952).

Kalbel, H., and Engelhardt, F., Brennstoff-Chem., 33, 13 (1 952).

Kummer, J, T. and Emmett, P. H., J. Phys. Chem., 56. 258 (1952).

Literature Cited Heterogeneous Catalysis

(1B) Aigrain, P., and Dugas, C., 2.Elektrochem., 56,363 (1952). (2B) Anderson, R. B., Feldman, J., and Storch, H. H., IND. ENQ. CHEM.,44, 2418 (1952). (3B) Antipina, T. V., and Frost, A. V., Doklady Akad. N a u k S.S. S.R., 84, 985 (1952). (4B) Ashworth, F., “Advances in Electronics,’’ Vol. 111,p. 1, New York, Academic Press, 1951. (5B) Balaceanu, J. C., and Jungers, J. C., Bull. SOC. chim. Belg., 60, 476 (1951). (6B) Balaceanu, J. C., and Jungers, J. C., World Petroleum Congr.. Proc. 3rd Congr., Hague, 1951, Sect. V, p. 67 (7B) Becker, J., BeZZSystem Tech. J., 30,907 (1951). (8B) Boer, J. H. de, and van Steenis, J., Koninkl. Ned. Akad. Wetenschap., Proc., B55,572,578,587 (1952). (9B) Bond, G. C., and Sheridan, J., Trans. Faraday Soc., 48, 651. 658, 664 (1952). (10B) Boudart, M., J . Am. Chem. SOC.,74,1531 (1952). (11B) Ibid., 74, 3556 (1952). (12B) Boudart, M., and Taylor, H. S., “L. Farkas Memorial Volume,’’ p. 223, Research Council of Israel, Jerusalem, 1952. (13B) Brady, E. L., and Zemany, P. D., J . Chem. Phys., 20, 294 (1952). (14B) Brattain, W. H., and Bardeen, J., Bell System. Tech. J.,32, 1 (1953). (15B) Buttner, F. H., Funk, E. R., and Udin, H., J. Phys, Chem., 56, 657 (1952). ENQ.CHEM.,45,159, 162 (1953). (16B) Ciapetta, F. G., IND. (17B) Ciapetta, F. G., and Hunter, J. B., Ibid., 45, 147, 155 (1953). (18B) Cremer, E., and Marschall, E., Monatsh. Chem., 82, 840 (1951). b (19B) Crowell, A. D., Am. J . Phys., 20,89 (1952). (20B) Douglas, J. E., and Rabinovitch, B. S., J. Am. Chem. Soc., 74,2486 (1952). (21B) bowden, D. A., IND. ENQ.CBEM.,44, 977 (1952). (22B) Duval, X., Compt. rend., 284, 208 (1952). (23B) Epgleton, A. E. J., Tompkins, F. C., and WanfoGd, D. W. B., Proc. R o y . SOC.(London), A213,266 (1952). (24B) Eischens, R. P., J . Am. Chem. ~ o c . 74,6167 , (1952).

Lavrovskaya, G. K., and Voevodskii, V. V., Zhur. Fiz. Khim., 25, 1050 (1951).

Livshits, V. D., and Sidorov, I. P., Ibid., 26,538 (1952). Long, F. J., and Sykes, K. W., Proc. Roy. SOC.(London), A215, 100, 111 (1952).

MoCabe, C. L., and Halsey, G. D., Jr., J. Am. Chem. Soc., 74, 2732 (1952). k Manes, Milton, Ibid., 74,3148 (1952). Mignolet, J. C. P., J . Chem. Phys., 20, 341 (1952). Mills, G. A., Heinemann, H., Milliken, T. H., and Oblad, A. G., IND. ENG.CHEM.,45, 142 (1953). Mooi, J., and Selwood, P. W., J . Am. Chem. Soc., 74, 1760

Ibid., 74, 2461 (1952).

Mulford, R. N. R., and Russell, W. W., Ibid., 74, 1969 (1 ~9.521 _ _. _ _ ,

Orzechowski, A., M h . SOC. TOY. sei. Lidge, 1 2 , l (1952). Paproth, H., Rathje, W., and Stranski, I. N., 2. Elektrochem., 56, 409 (1952).

Parravano, G., J . Chem. Phys., 20, 342 (1952). Pichler, H.. Brennstoff-Chem.. 33. 289 (1952). Plate, A. F., and Tarrtsova, G. A:, Zhu;. ObshcheI Khim., 22, 765 (1952).

Rabinovich, E. Ya., Snegireva, T. D., and Tesner, P. A., Doklady Akad. Nauk. S.S.S.R., 88, 95 (1952).

Rapoport, I. B., and Levkovich, M. M., Ibid., 84, 725 (1952). Reif, A. E., J,Phys. Chem., 56,785 (1952). Rienilcker, G., Bremer, H., and Unger, S., Naturwiss. 39 (1952).

Roginskii, S. Z., Doklady Akad. Naulc. S.S.S.R.,87, 1012 (1952).

Rozen, A., and Shevelev. Ya., Ibid.. 87.817 (1952). (72B) Ibid., 87, 1017 (1952). (73B) Rubinshtein, A. M., Kulikov, S. G., and Pribytkova, N.A., Ibid., 85, 126 (1952). (74B) Sastri, M. V. C., and Ramanathan, K. V., J . Phys. Chem., 56, 220 (1952). (75B)Schtifer, K., Z . Elektrochem., 56, 398 (1952). (76B) Schleicher, H. W., 2.Naturforsch., 7a, 471 (1952). (77B) Schuit, G. C. A,, and de Boer, N. H., Rec. trav. chim., 70, 1067 (1951). (78B) Schwab, G. M., 2.Elektrochem., 56,297 (1952).

INDUSTRIAL AND ENGINEERING CHEMISTRY Selwood, P. W., and Lyon, L., J . Am. C h m . Boc., 74, 1051 (1952). Sheridan, J., and Reid, W. D., J. Chem. Soc., 1952, p. 2962. Shultz, J. F., Seligman, E., Lecky, J., and Anderson, R. B., J . Am. Chem. SOC.,74, 637 (1952). Shultz, J. F., Seligman, B., Shaw, L., and Anderson, R. B., IXD.ENQ.CHEM.,44,397 (1952). Shurmovskaya, N. A., Bruns, B. P., and Mel'nikova, Z. Ya., Zhur. Fiz. Khim., 25,1306 (1951). Suhrmann, R., 2. Elektrochem., 56, 351 (1952). Taylor, H. il., and Thon, iY,, J. Am. Chem. Soc., 74, 4169 (1952). Tesner, P. A , , and Rafal'kes, I. S., Doklady Akad. Xauk. S.S.S.R., 87, 821 (1552). Thompson, S . O., Turkevich, J., and Irsa, A. P., J . Phys. Chem., 56, 243 (1952). Tompkins, F . C., 2. Electrochem., 56,360 (1952).

Vol. 45, No. 5

(89B) Tramboure, Y., de Mourgues, L., and Perrin, M., Compt. Tend., 284, 1770 (1952). (90B) Trapnell, B. M. W., Trans. Faraday Soc., 48,160 (1952). (91B) Troesch, A., thesis, Lyon University, France, 1952. (92B) Vandael, C., I n d . chim. belge, 17, 581 (1952). (93B) Varga, J., Rabo, G y . , and Steingazner, P., Acta Chim. Acad. Sci. Hung., 1, 146 11952). (94B) Vol'kenshtein, F. F., Zhur. Pia. Kh