No.4091 March 27, 1948 NATURE 461

ADSORPTION AND HETEROGENEOUS CATALYSIS* By PRoF. ERIC K. RIDEAL, O.B.E., F.R.S.

E now recognize two forms of adsorption: that Such experiments reveal the fact that many cases W in which the dispersive forces alone are operative of slow uptake of gas, which had been attributed to -the so-called van der Waals or physical adsorption the chemi-sorptive process, possessing a definite and -and adsorption in which an electron switch has large energy of activation, were in reality due to taken place between adsorbent and adsorbate- other causes ; for example, solution or displacement chemi-sorption. It is important to obtain criteria for of one surface layer by another, since the chemi- distinguishing between these two forms so that we sorptive process proceeded on the clean metal can identify the type or types involved in any surface without any appreciable energy of activation. particular case of heterogeneous catalysis. While there may be some doubt in particular cases The literature concerning adsorption is very as to the exact magnitude of the energy of activation voluminous, and we may refer to Brunauer's recent involved, it is relatively small compared to the volume on physical adsorption to get an idea of its bond-str-engths of the ordinary diatomic gases which magnitude and extent. Here we merely observe that undergo dissociation on chemi-sorption. The metal- if we are to obtain a complete picture of such adsorp- electron substrate must act in a manner analogous to tion, more attention will have to be paid to the a free radical. That an electron switch has taken forces operative with different and definite sub- place during chemi-sorption is revealed by the strong strates. A commencement has already been made, dipole character of the chemi-sorbed complex; for but we require more systematic investigation of the example, adsorption of polarizable and non-polarizable adsorb- + +- + - ants on substrates which are respectively ionic, W CS or W 0, W H. covalent and metallic in lattice structure. One may In several cases, for example, oxygen and ethylene, cite as an example of the former the adsorption of the reacting electron is probably the 1t-electron of the rare gases on one hand, and of iodine on the the adsorbate ; whereas many catalytic metals, for other, on crystalline potassium chloride, whereby the example, those in the transition group, have a narrow theoretical treatment can be examined in the light d-band overlapped by a broad s-band of electron- of experimental data. As an example of the latter, levels, and it is possibly an electron-level in this band we may mention the experimental work on the which is involved. While a number of experiments adsorption of vapours at the surface of liquid mer- have been carried out to determine the thickness of cury, which reveals the shortcomings of the various metal required for the emergence of 'metallic' theoretical expressions which have been derived for characteristics, the great experimental difficulties such interactions with metallic surfaces. involved in maintaining the substrate, in the effects Apart from the complications introduced by the of surface tension, in formation and preservation of fact that the adsorbing surface is scarcely ever a thin continuum in lattice form, have so far hindered nniform, it is by no means improbable that many progress in this direction. It is possible that further cases of physical adsorption on what is believed to advances in our knowledge of the nature of the be a metallic substrate may well be taking place on chemi-sorbed complex may come through physical the top of a chemi-sorbed monolayer such, for examination of snch systems, for example, the example, as hydrogen on tungsten, on which there is determination of specific heats of surface films, of already a chemi-sorbed monolayer of gas such as the dielectric constant and para- or dia-magnetic tungsten oxide or tungsten hydride. The slow uptake susceptibilities. Experimentally, such investigations of gas which results on the admission of a gas to a are proving to be within the compass of modern bulb containing finely divided reduced metal, or technique. other solid, may be due to the formation of a chemi- One of the most important problems connected sorbed layer on its surface, or to a process of solution with adsorptive phenomena is that concerning surface and diffusion into the interior of the solid. The data mobility. Under what conditions are adsorbed derived from a study of the kinetics of such processes layers, held by dispersive forces and chemi-adsorbed have been interpreted on a basis both of 'activated' layers respectively, mobile over the surf.we? The adsorption and of 'activated' diffusion. It is clearly surface field is an nndulating periodic one, and important to be able to distinguish between these movement from one site to another may involve two different processes, since either of them might evaporation and recondensation, or moV ment prove to be the controlling reaction in a case of over a potential hill. The adatom or adion heterogeneous catalysis. Valuable information may be raised into a migratory level. Here it may be obtained by studying the processes on surfaces be free to move all over the surface, or may be freshly prepared in vacuo. So far, tungsten wire and limited to flight in one line only. It may move only mirrors of other metals deposited by thermal evapora- into the next trough, or have a relatively long range tion have been employed. of flight. We can develop equations to satisfy each Different physical methods are available for of these conditions, but the experimental evidence examining the surface properties of a wire or a mirror. so far is too inexact and too scanty to make any Thus the accommodation coefficient for a gas, the decision. heat change, the thermionic and photo-electric proper- In the case of the alkali metals on tnngsten, the ties all provide information on the presence or data favour the view that the ion in the migratory absence and, indeed, on the structure also of chemi- level is restricted to a linear path and can traverse sorbed monolayers. several tnngsten atoms before terminating its flight. The growth of crystals from the vapour phase and • Substance of a course of three lectures delivered at the Royal Institution on February 12, 19 and 26. the migration of organic molecules and metallic atoms © 1948 Nature Publishing Group 462 NATURE March 27, 1948 Vol. 161 over glass surfaces provide further evidence for In view of the repulsive forces which we have shown lateral mobility. Such evidence as is available at to be operative and the great heat of adsorption even present suggests that lateral mobility of physically in a closely packed film, it is clear that many diffi- adsorbed molecules, but not of the chemi-sorbed culties have still to be overcome before this hypothesis species, can occur within the range of temperature as to the mode of action can be freely accepted. at which the ca1oalytic reactions occur. We cannot, An alternative proposal involving an interaction however, eliminate the possibility of some mobility between a physically adsorbed molecule situated over even in the chemi-sorbed species. one of the vacant sites, and a neighbouring chemi- Several hypotheses concerning surface catalysis sorbed adatom to form a transition complex, presents involve the concept of two-dimensional gas reactions. us with a mechanism involving exchange of valence The experimental evidence available suggests that in forces rather than rupture of any primary bonds, surface reactions chemi-sorbed reactants, and not and goes far to overcome the difficulties associated physically adsorbed reactants, are involved ; thus with the other two proposed mechanisms. Many of the question of mobility becomes one of great the heterogeneous catalytic reactions can be readily importance. Apart from the fact that this mobility interpreted by means of such a mode of action, but for the chemi-sorbed species appears to be non- experimental re-examination of some of the catalytic existent or extremely limited, we must note that the exchange reactions is necessary before an unequivoc- adatoms have dipole moments, and as a result strong able decision is possible. repulsive fields exist between them. It is interesting The mechanism proposed, it will be seen, involves to compare this repulsive or spreading pressure of a what is, in fact, the half-hydrogenated state which film of adatoms where the pressure is due to a free in the catalytic chemistry of the hydrocarbons has two-dimensional gas, that is, one which obeys the proved such a useful concept. The hypothesis of formal relationship vacant sites inaccessible to particular reactants F = crKT, suggests their accessibility to other species, and provides us with an explanation of the broken linear where cr is the surface density of adatoms, with the character of the curves connecting poisoning with pressure due to a free two-dimensional array of catalytic activity, an explanation alternative to a dipoles, that is, series of patches of different activity. Dissociation F 1 = 4·51 cr51•, in chemi-sorption involves not only an electron where !L is the dipole moment. At values of cr = switch, but also the problem of interatomic spacing; 0·8 x l014/cm. 2 and 4·6 x l014fcm. 2, the values ofF the relation of both these factors to the energetics of and F 1 are for formic acid on mercury 14 and surface action provides yet another field of inquiry 52 dynes/em. for F, and for potassium on tungsten as yet almost unexplored. Provisional theoretical 26·0 and 307 dynesfcm. for F 1 respectively. The examination of the effects of surface spacing on the value of the dipole moment !L likewise changes as we activation energy have not taken the electron dis- alter the surface density of the film, due to mutual tribution in the substrate into consideration ; while interaction. attempts to evaluate energies of activation by Examination of the heats of adsorption givEs us examination of metals which form a continuous series some indication as to the extent of mutual inter- of solid solutions and thus provide a gradual change action which is repulsive in the chemi-sorbed layer, in the mean lattice spacing, require a more detailed as well as on the mobility of the film. It is interesting examination of the composition of the Gibbs surface to note that, in the physically adsorbed layer, below layer and its dimensions before we can draw justifiable the critical temperatures attractive forces appear to conclusions. predominate ; for example, in the condensation of metallic and other vapours on glass. In the chemi-sorption of a diatomic gas such as hydrogen on the surface of a metallic lattice, the THE MAIN GEOMAGNETIC FIELD molecule undergoes dissociation to form a pair of metallic hydride molecules. If the film ·be immobile HE main geomagnetic field and its secular during the process of filling up the lattice, some T variation were the subject of a Geophysical eight per cent of the lattice points will be left isolated Discussion held at the Royal Astronomical Society as singlets. These singlets may play an important on February 27. The largely attended meeting, at part in heterogeneous catalysis. Two of the simplest which the Astronomer Royal was chairman, was cases of heterogeneous catalysis are the ortho-para addressed by Dr. E. C. Bullard, of the Department conversion ofhydrogen or dauterium and the exchange of Geodesy and Geophysics, Cambridge (who is reaction between hydrogen and deuterium at the shortly to take up the post of head of the Department surface of non-paramagnetic metals. The reactions of Physics in the University of Toronto), Dr. J. MeG. proceed smoothly at quite low temperatures with Bruckshaw, in charge of the sub-department of small but definite energies of activation. In the cases applied geophysics at the Imperial College of Science which have been examined, the metal is covered and Technology, London, Mr. D. W. Bishopp, with a chemi-sorbed layer of hydride broken only by director of the Geological Survey of Eire, and Mr. the gaps already referred to. The formal kinetics of S. K. Runcorn, of the Physics Department in the the reaction and reference to the corresponding University of Manchester. homogeneous reactions preclude the possibility of Dr. Bullard outlined a new theory of the secular the reaction proceeding by a two-dimensional magnetic variation. He first recalled the main char- bimolecular gas reaction in a mobile physically acter of this variation, as exemplified by the longest adsorbed phase on the top of the chemi-sorbed series of observations, for London, which suggests a atomic or hydride layer. One of the current views is cycle of change of magnetic direction, of which about that the reaction pr0ceeds by interaction of two three-quarters has been completed in four centuries. neighbouring adsorbed hydrogen adatoms with sub- Other places, however, show variations that suggest sequent evaporation of the newly formed molecule. shorter periods. The accumulated data discredit the

© 1948 Nature Publishing Group