Search

In this paper, introductory to the Symposium, an attempt is made to summarize and rationalize the available information on the reforming reactions of hydrocarbons on metals and alloys. It is shown that three important aspects of hydrocarbon adsorption, viz. chemisorption bond strength, reactivity towards C-H bonds and the ability to form multiple bonds with the group VIII metals, are not dramatically changed by alloying. Therefore, various changes due to alloying observed in the selectivity of alloys are ascribed to the availability of ensembles of active sites on the alloy surface (ensemble size effects). It is shown that also on pure metals various intermediates are identified which differ in the number of C- atoms and metal-atoms involved in the formation of activated complexes. By alloying a consistent shift is observed from mechanisms and intermediates requiring big ensembles to mechanisms which can also operate on small ensembles. As a rule mechanisms involving big ensembles are faster. The identity of alloy components in their chemisorption and catalytic behaviour is preserved to a high extent also in alloys and is a more important feature in catalysis than the effects caused by mutual influencing of alloy components.

In this review, the most relevant aspects of skeletal isomerization of n-butene to isobutene are discussed: the nature of the active sites, the prevailing mechanism of the skeletal isomerization, and the relation of this to that of n-butane. It is concluded that the prevailing mechanism of skeletal isomerization of n-butene is monomolecular (in contrast to butane isomerization) and requires Br⊘nsted acid (OH) active sites. The selectivity and catalytic stability can be influenced by the shape selectivity of zeolites and zeotypes. These effects are explained on the basis of the knowledge on the prevailing mechanism.

Adsorption of acetic acids has been studied by FT-IR on a series of adsorbents, known as good catalysts for the selective hydrogenation of carboxylic acids to corresponding aldehydes: SnO2, TiO2, Pt SnO 2 , Pt TiO 2 . Very similar adsorbed species are observed by FT-IR with all mentioned catalysts. The catalysts differ mainly by their behaviour upon increasing temperature program. The strongly bound unidentate or the bidentate, bound to the oxide, are the possible intermediates of the catalytic reaction to acetaldehyde.

The present study gives the results of our attempts to gain more knowledge on the interaction between an aromatic carboxylic acid and an oxide catalyst. There is quite a difference in the coordination of benzoic (and salicylic) acid with an acid-base catalyst (γ-Al2O3) or a redox one (α-Mn3O4). When a substituent is incorporated into the ring, the difference becomes even larger. These differences can perhaps explain the difficulties in the catalytic reduction from (substituted) benzoic acids to the corresponding aldehydes.

The structure of metallic copper oxide La2Sr6Cu8O18-δ (δ≈0.4) has been determined by electron diffraction and neutron powder diffraction. The correct space group was found to be P 4/mbm with lattice parameters a=10.8502 (3) A≈2√2ap and c=3.7658 (1) A≈ap, where ap is the lattice parameter of the cubic perovskite. The oxygen vacancies are ordered both in the basal plane, along [001], and in the z = 1 2 plane, along [110]. The idealised framework of [Cu8O18] consists of square planar CuO4 groups and CuO5 square pyramids by sharing their corner oxygen, leading to the formation of hexagonal tunnels along the c-axis. La3+ and Sr2+ ions are located in these tunnels in a disordered manner.

Reactions of n-hexane on three bimetallic (alloy) systems: Pt-Re, Pt-Ir, Pt-Co have been studied with sulphur-free and sulphided catalysts. A shift is observed in the selectivity pattern from “metallic: (on sulphur free catalysts) to “acid catalysed” reactions (on sulphur modified catalysts), which is explained by the shift in the formation of certain intermediates of the reaction.

The main topics of the presentations and the discussions in this session were synchrotron radiation techniques, in particular X-ray absorption spectroscopy, (EXAFS and NEXAFS), infra-red spectroscopy and scanning tunnelling microscopy. Transient methods are discussed in session 5 by Frennet and UV/visible spectroscopy is mentioned briefly by Rooney, (session 1). Some other in-situ techniques, like electron microscopy on catalytic reactors, positron emission tomography and the wider use of radio-active isotopes in studies on the mechanism, were therefore outside the scope of the workshop. However, the choice of the techniques to be discussed was not just incidental and the discussion confirmed the merits and a good future for the above mentioned techniques in catalysis.

The geometric and promoter effects on the selectivity to furhryl alcohol were studied in the hydrogenation of furfural over various series of platinum catalysts. The results obtained with the PtCu catalysts indicate that the size of the active site does not affect the selectivity. Promotion of the platinum catalyst can lead to a considerable rise in the selectivity to furfuryl alcohol. The hydrogenation over the series of PtSn catalysts showed the influence of the reaction conditions. The experiments under isothermic conditions resulted in up to 80% selectivity for hrfuryl alcohol, while the selectivity dropped to approximately 40% if non-isothermic conditions were applied. This change in selectivity is attributed to the self-poisoning of the catalyst at high temperatures.

The changes in work function φ upon adsorption of C 2 H 4 on clean film surfaces of six fcc metals (Ni, Pd, Pt, Cu, Au and Al) have been followed by means of photoelectron emission at 293 K. A marked difference was observed in the behaviour between Ni, Pd and Al on the one side and on Cu, Au and Pt on the other side: while with Ni, Pd and Al, φ as a function of coverage goes through a maximum, with Cu, Au and Pt, φ only decreases. In the discussion, the data obtained by work function measurements are related to other literature data. Several films covered with C 2 H 4 species were also submitted to a heat treatment while in other experiments H 2 was admitted to the surface covered by C 2 H 4 species. In some experiments C 2 h 4 was admitted to surfaces covered by H 2 . In all cases φ was measured. The experiments reveal that C 2 h 4 is absorbed only reversibly on Cu and Au. On Ni, Pd and Pt, C 2 H 4 is adsorbed initially with dissociation and this leads to an increase in φ on Ni and Pd and a decrease on Pt. Hydrogenated reactive species contribute to the lowering of φ observed with Ni, Pd and Pt. As with Cu and Au also on Ni, Pd and Pt a weakly bound C 2 h 4 is observed which leads to a decrease in φ as well. The behaviour of φ indicates that upon Al, C 2 h 4 adsorbs first dissociatively to a small extent, while the weakly bound C 2 H 4 species act as intermediates for strongly adsorbed species which were observed after some time.

Reactions of three /sup 13/C-labeled methylpentanes, namely, 2-(2-/sup 13/C)methylpentane, 2-(4-/sup 13/C)methylpentane, and 3-(3-/sup 13/C)methylpentane, were performed on a series of Pt-Cu alloy catalysts prepared by coimpregnation in order to determine the mechanistic change in isomerization caused by the dilution of Pt in Cu. The results of this research are explained by an ensemble size effect which causes a shift from multisite mechanisms for isomerization (on Pt-rich alloys) toward one-site mechanisms for isomerization (on Cu-rich alloys).

Some aspects of the characterization of alloy surfaces by two methods, viz. Auger spectrometry and thermal desorption, are discussed. Examples demonstrate how the results of the surface composition help to understand the catalytic data. The catalytic systems discussed are: Pt Pd, Pd Ni, Pt Au and Ni Cu. With regard to the last mentioned system, information is now available on more than twenty reactions. These data together with the fairly reliable information on the surface composition of Ni Cu alloys allow to draw quite general conclusions with regard to the classification of catalytic reactions and identification of factors responsible for the catalytic selectivity.

Isomerization, dehydrocyclization and hydrogenolysis reactions of n-pentane and n-hexane were studied on supported Pt and Pt-Au alloys. The alloys studied were all in the composition range where this system is completely miscible (18% Pt in Au). Characteristic changes in selectivity towards the different reactions were observed when Pt was compared with alloys in the composition range 1–12.5% Pt and at temperatures 250–400 °C: (a) extremely diluted Pt in Au (1–4%) mainly catalyses isomerization; (b) alloys with a composition around 10% Pt favor dehydrocyclization; (c) pure Pt shows mainly isomerization. From the comparison of the n-pentane and n-hexane reactions it is possible to suggest a self-consistent picture of the mechanisms involved. For the alloys in (a) above the reaction mainly proceeds via a one-site isomerization mechanism. Catalysts in (b) are comparable to small particle catalysts and show cyclization and isomerization mainly via a cyclic mechanism. Pure Pt catalyses isomerization mainly via the bond shift mechanism (αα-γ adsorption) while ring closure probably proceeds via an αα-Ω adsorbed intermediate.

The exchange reaction of cyclopentane and deuterium on Pt and Pt--Au alloys was studied, and information on reactions of C--H bonds and reaction mechanisms was provided. Product distribution data for the Pt and Pt--Au catalysts at various temperatures was tabulated. It was concluded by comparing the alloy and Pt films at the same conditions that the activity decreases on alloys by up to one tenth of that of pure Pt films. (DDA)

Palladium-silver alloys have been prepared by chemical reduction at 353 K and by equilibration in hydrogen at 673 K. They have then been investigated by Auger spectroscopy. The equation derived by Gallon has been used for the determination of the “monolayer intensity” and for the evaluation of data on the surface composition of alloys. The surface composition measured experimentally has been compared with the theoretical predictions. It appears that in compliance with the latter, the surface of equilibrated alloys is strongly enriched in silver. This silver enrichment is of the same order of magnitude as indicated earlier by other methods. The implications of this surface enrichment for chemisorption and catalysis are discussed.

A series of Pt-Cu alloys on a silica carrier has been prepared and investigated by infrared spectroscopy. It is observed that the position of the infrared bands is a function of the alloy composition. Comparison of the data obtained with other systems reported in the literature excludes the explanation of the shift in terms of electron transfer between alloy components (changes in n d , the number of d electrons) or in terms of bond strength to the nearest neighbors of the metal atoms in alloys. A tentative explanation is suggested according to which changes in the collective properties of metals (local density of states) are responsible for the shift of the infrared band maxima.

The IR Spectra of carbon monoxide adsorbed on different forms of platinum, iridium and copper (flat and rough surfaces) have been investigated. The results of IR transmission experiments on the silica supported metals are presented and compared with the data from IRAS (IR reflection adsorption spectroscopy) experiments on the unsupported metals. It is shown that the coverage dependency of the IR band frequency ΔνCO (θ) is independent of the state of the metal if the formation of islands of CO is taken into consideration. This was proven by isotopic dilution experiments with12 CO/13 CO mixtures. It is suggested that many apparent discrepancies and irreproducibilities in the literature can be explained by an uncontrolled formation of islands. At a given surface coverage. Group VIII metals reveal a lower νco value when CO is adsorbed on rough surfaces (or small particles) than on smooth surfaces (or large particles). The reverse is true for the lb metal Cu. The differences between supported and unsupported metals can be ascribed to differences in the electronic and geometric structure of the surface. The differences between adsorption of CO on Group VIII and lb metals are discussed in terms of a difference in bonding of CO on these two groups of metals.

Methanation of CO and CO 2 , disproportionation of CO, and hydrogenation of carbon deposited on the surface of catalysts have been studied on Ni and Ni-Cu alloy films at low pressures, at temperatures of 250–350 °C. Under these reaction conditions, two steps of the overall mechanism have been identified; namely, dissociation of CO and hydrogenation of deposited carbon by adsorbed hydrogen. CO can be dissociated only on places where carbon atoms can be bound to several Ni atoms simultaneously; therefore addition of Cu to Ni strongly reduces the rate of methanation.

The extent of adsorption at 295 and 78 K and the thermal desorption of CO from Pd, Ag, and PdAg alloy films have been measured in an ultrahigh vacuum system. The adsorption data indicated that the surface of the alloys studied in this paper was enriched by Ag. Desorption spectra revealed several adsorption/desorption states, designed as γ 1 , γ 2 , α 1 , α 2 , and β . The nonspecific, most weakly bound γ 1 -state was observed with all adsorbents. The γ 2 -state, also weakly bound, was observed only on Pd. The α 1 - and α 2 -states were observed on Pd and PdAg alloys; they were completely missing on Ag. The population of α-states decreased with increasing Ag content. This is explained by assuming that upon alloying of Pd with Ag the number of large ensembles necessary for strong adsorption decreases.

When a mixture of carbon monoxide and hydrogen is contacted with some metals (Co, Ru, Ni, Fe), a great number of reactions occur. Some of them are schematically described by the following stiochio-metric equations Reaction (1) with m=1 was discovered by Sabatier [1] and is usually called methanation. Reactions (1) to (3) with m < 1 are the Fischer-Tropsch [2] synthesis of hydrocarbons. There are several very good review papers [3–8] on this subject, and it is not intended to add another paper to this collection. Rather, results obtained during the last few years at Leiden will be reviewed. The mechanism suggested is still somewhat speculative; however, some experimental evidence exist for all aspects of the mechanism.

Reactions of butanal and butanol and hydrogenation of cyclopentanone have been studied in a flow apparatus in mixtures with hydrogen and at a total pressure of 1 atm. The catalysts used were Ni and NiCu powders without carrier and Pt and PtAu on silica. Hydrogenation/ dehydrogenation of the compounds is only marginally influenced by alloying, while the inter- and intramolecular dehydration (formation of olefins and ethers from alcohols) is suppressed strongly. The selectivity can be controlled by the availability of ensembles of active sites. A mechanism is suggested which can rationalize the influence of alloying, including some data from the literature.

Results are reported on the reactions of n-butane, n-pentane and n-hexane on Pt and Pt/Au alloy films. The results are compared with those obtained for Pt−Au alloys on a SiO2 carrier. It is concluded that alloying can change the selectivity of the catalyst. No indication has been found for the possible role of oxygen in the high selectivity of Pt in non-destructive reactions.

The chemical characteristics and the reactivity of carbon on polycrystalline Rh and Pt surfaces towards hydrogen have been studied by Auger Electron Spectroscopy. Auger spectra of carbon on both metals have been subjected to spectrum subtraction techniques in order to reveal the chemical characteristics of it, as reflected by the fine structure of the carbon KVV Auger spectra. Two classes of carbon can be identified by AES: (A) amorphous and graphitic carbon; (B) molecular (or molecular fragments) and carbidic carbon, the first being almost unreactive in H 2 . Carbon B transforms into A and this transformation is faster on Pt than on Rh and is also faster in the absence than in the presence of H 2 .

The Ni-Cu alloys were prepared by evaporation of the specpure metals in UHV onto a quartz substrate. Spectra were obtained from clean as well as from gas covered surfaces. The Auger signal intensity of a monolayer of both metals was determined for the low energy electrons (102–105 eV) and for the high energy electrons (716–920 eV). The overlapping peaks of Cu and Ni in the low energy region (102–105 eV) were evaluated by comparing them with computer simulated alloy spectra. The results of the sintered alloys are interpreted by means of a model by Gallon and Jackson, using the experimentally determined signal intensity of a monolayer. Several surface enrichment data were used to predict the experimentally observed Auger signal intensities. A clear indication of surface enrichment of Cu was obtained; this is in good agreement with previous conclusions based upon hydrogen adsorption and work function measurements. An explanation is suggested why previous work with AES and CO chemisorption did not reveal any surface enrichment.

Infrared reflection absorption spectroscopy (IRAS) is applied to follow the interaction of CO with O 2 on evaporated well sintered Ir films. It is confirmed that the CO reaction with oxygen in the adsorbed state is possible, and that at low temperatures domains are formed of adsorbed CO and adsorbed oxygen.

The effect of alloying on the absorption band frequency of CO adsorbed on Pt , v ( CO Pt ) , has been investigated with the alloys Pt-Pb, Pt-Sn, and Pt-Re. By use of the isotopic dilution method, it has been established that the downward shift of v ( CO Pt ) observed upon alloying Pt with Pb is caused partly by a dilution effect and partly by another, possibly electronic effect. Alloying Pt with Sn or Re, on the other hand, caused only a small shift of v ( CO Pt ) , which, moreover, could also be explained at least partly by a dilution effect. Oxidation of the alloy samples resulted in a phase separation of the components; the alloy situation could be restored again by a reduction procedure.

The ability to form multiple bonds has been studied for transition metals alloyed with aluminium. Results show that their behavior does not differ significantly from that of alloys with endothermic and weakly exothermic heats of alloy formation, i.e. aluminium does not suppress the multiple bond forming property of Ni and Pd.

Cyclopentane-deuterium exchange has been followed on pure Pd and PdNi alloy films in the temperature range 200–430 K and at hydrogen/cyclopentane ratios of ~50. Conclusions on activity changes are limited because of the self-poisoning of alloys and Ni catalysts. However, the selectivity of the catalysts is much less influenced by self-poisoning; the initial product distributions are well reproducible and independent of these side processes. The main result of alloying of Pd with Ni is the gradual disappearance of one low temperature mechanism which leads on Pd to D10 products. This fact can be easily understood if we assume that it is the “roll-over” mechanism which is responsible for these products.

The reactions of neohexane were studied on a series of Pt-Cu alloy catalysts. It appears that from the three possible adsorption modes (the αγ, αγ′, and αβ complexes) which neohexane offers when two carbon atoms are bound to the surface, the αγ′ mode is preferred to the others and this tendency is independent of alloy composition. In this way, the already known preference of Pt to form αγ and αγ′ complexes is fully confirmed. The activity and selectivity pattern found with other hydrocarbons studied earlier is also confirmed: with increasing Cu content, the selectivity for hydrogenolysis increases, whilst that for isomerisation decreases. An explanation is suggested according to which the chemisorption complex yields isomers upon further reaction on a pure Pt ensemble, whereas on a mixed ensemble (Pt and Cu) hydrogenolytic products result.

The temperature programmed desorption of hydrogen from platinum films has been measured in a U.H.V. system by means of a mass spectrometer. The parameters varied were the extent of adsorption, the duration of the adsorption process and the heating rate during desorption. For maximum coverage three desorption peaks have been observed in temperature-programmed desorption: γ (120 K), β1 (200 K) and β2 (330 K). The population of the γ state varies with the duration of pumping at 78 K before adsorption is started, and the population of the β2 state is dependent on the duration of adsorption at room temperature. The isosteric heat of adsorption for high coverages has been measured and varies from 8 to < 2 kcal/Mol for coverages from 7 to 12.1014 mol/cm2 geometric surface.

The practical importance of alloy surfaces in catalysis, corrosion andother aspects of materials performance is widely recognized. What is needed now is sufficient knowledge of the relationship between externally controllable factors — alloy composition, temperature, environment — and surface properties — composition, structure, chemical activity — to control materials performance in these applications. Our purpose here is to review progress in determining and predicting the relationship between one surface property, composition, and certain externally controllable variables: overall composition, temperature, environment and physical form. We find that theoretical treatments of metal alloy surface composition now include essentially all significant physical effects and can predict values for most parameters of interest. Though improvements are still possible, the accuracy of predictions is more often limited by uncertainties or absence of the basic data for the calculations (e.g., thermochemical values) than by the models themselves. Alloy surface composition can now be measured well. The first monolayercomposition of large alloy slabs can be determined quantitatively over a wide temperature range in ultra-high vacuum. Difficulties with specimens of practical interest still challenge experimentalists. Among these are supported catalysts, surfaces under chemisorbed layers and composition of layers below the first. Significant progress is being made and we expect the next few years will see success.

The thermal desorption of hydrogen from Pt and PtAu films has been measured in an ultra-high vacuum system by means of a mass spectrometer. On the average, hydrogen is more loosely bound on the alloys than on pure Pt. About 50% of the adsorbate is desorbed by pumping at 78 K from the alloys while only a very small percentage is desorbed from Pt at this temperature. After maximum coverage of Pt films by hydrogen adsorption three desorption peaks have been observed: γ (120 K), β 1 (200 K) and β 2 (330 K). The same peaks have been found for the alloys as well but the relative population of the various adsorption types was different. The relative peak heights vary with the alloy composition.

It is to be regretted that only few catalysts have been studied from the point of view of “materials science”. Nevertheless, important information can be obtained on the function of catalysts when, for example, the surface composition of multicomponent catalysts is determined using surface science tools. Two examples of such investigations are discussed: PdAg alloys and NiCu alloys. The data on PdAg alloys show that the vibrational spectra of CO and possibly also the catalytic behaviour of the alloys can be explained by the dilution of palladium in silver, i.e. by essentially “geometrical” effects. The data on NiCu alloys show that the Fischer-Tropsch synthesis of hydrocarbons is a reaction that requires ensembles of several active atoms to occur.

A short study on the title reaction confirms that at low hydrogen pressures no ether formation occurs, while when a higher hydrogen pressure suppresses the fast dehydrogenation, ethers are formed. A comparison of Pt and Pt alloys shows that ether formation is faster on large ensembles of Pt atoms, requirements of decarbonylation are less in this respect. Results of earlier studies on Ni/Cu alloys have thus been confirmed.

Coadsorption of ethylene or oxygen with carbon monoxide has been investigated using ir transmission spectroscopy. Information on the extent of “electronic” and “dgeometric” (dilution) effects on the ir stretching frequency of CO has been obtained by the isotopic dilution method. The increase of this frequency (blue shift) caused by oxygen can be explained mainly by a compression of the CO layer. The decrease of the frequency (red shift) caused by ethylene can be explained for more than 60% by a dilution of the CO layer.

The work function Φ, the emission constant M , and the increase in work function after CO adsorption ΔΦ have been measured as a function of the overall composition of well annealed Ni-Al alloy films. The work function of the Ni-Al alloys is not constant within the several miscibility gaps of the Ni-Al phase diagram, and is not lower than the work function of the respective pure metals, as was found with Ni-Cu and Pt-Au alloy films within the miscibility gap. Of all compositions studied only the 25% Ni alloy (NiAl 3 ) distinguishes itself by an increased Φ-value. The higher Φ is explained by the higher atomic density of the NiAl 3 compound compared with pure Al and Ni 2 Al 3 . As with other metal alloy systems, an enrichment of the surface by the component with the lower surface energy (Al) is also found with the ordered Ni-Al alloy system. From ΔΦ measurements it appears that up to about 40% Ni the surface of the Ni-Al alloy is composed entirely of Al. The behavior of Φ and ΔΦ in the Ni-rich part is explained by the enrichment of the surface by Al. The tendency to form the surface from Al atoms is strengthened by the existence of suitable crystallographic structures. The emission constant does not distinguish any of the intermetallic compounds and resembles that of the Pt-Au and Pd-Ag alloy films.

Pt-Pd alloys were prepared as powders by a chemical reduction method and as films by evaporating the specpure metals onto a quartz substrate. Results were evaluated by means of the Gallon equation; the parameters of this equation — the signal intensity of a monolayer and of the bulk material — were determined experimentally for all relevant Auger peaks. The backscattering factor for the alloy system was evaluated from experiments in which a layer of one metal was evaporated on top of the other. The value determined in this way was compared with results obtained in the literature on similar systems. The theoretical surface compositions are calculated for several models and compared with the experimental results. Experimental results reveal surface enrichment in Pd, in good agreement with the theoretical calculations. Interactions of alloys with carbon monoxide and propane are discussed and a possible explanation of the effects observed is suggested.

Catalysts containing 15% loading of 4% Pt-Au alloy on silica or ..gamma..-alumina were prepared by co-reduction with hydrazine, reduced in hydrogen, and tested for activity for n-hexane isomerization and cyclization at 350/sup 0/C. The catalysts were then subjected to two oxidation-reduction cycles, and tested with the same reaction. Over alumina-supported alloy, the initial conversion was 2.9%/g catalyst and the selectivities for isomerization and cyclization were approx. 90 and approx. 5%, respectively. After the oxidation-reduction treatment, conversion was 6.0%/g catalyst, and isomerization and cyclization selectivities were 30 and 56%, respectively. The silica-supported alloy had similar initial activity and selectivity as the alumina-supported catalyst, but did not change significantly on oxidation-reduction. A comparison with pure platinum on the same carriers suggested that the platinum segregated from the alloy during oxygen treatment and formed platinum oxide, that the platinum oxide migrated to the alumina surface where it formed separate particles on reduction, but it was unable to ''wet'' the silica surface and migrated to the alloy particle surface in the silica-supported catalyst, where it forms bulk alloy again, on reduction.

The work function Φ and the change in work function ΔΦ caused by the adsorption of CO or C2H4 have been measured as a function of the overall composition of equilibrated Ni-Cu and Ni-Au alloy films. The measurements reveal that Φ as a function of bulk (average) composition of the alloy films has the same character for alloy films annealed below the critical temperature Tc, where two phases coexist (a Ni-rich phase and a Cu- or Au-rich phase), as well as for films annealed above Tc where only one phase is found. There always appears to be a broad range of concentrations where Φ is independent of the composition. Both the Φ and ΔΦ measurements indicate the surface in this concentration range to be Cu rich (about 20 atom% Ni) or Au rich (about 10 atom% Ni). The critical temperature Tc of Ni-Cu alloys was checked. The results agree with the value predicted by Meijering. The C2H4 adsorption experiments reveal that on Ni-Au surfaces less deep dissociation occurs and more partially hydrogenated C2H4 species are present than in the case of Ni-Cu alloys. The results indicate that the surface of Ni-Cu alloys within the concentration range of constant surface composition contains larger Ni clusters than in the case of Ni-Au alloys.

Pt-Ir alloys were prepared by evaporating the specpure metals onto a quartz substrate and equilibration at 700 K. The Auger signal intensity of a monolayer of the elements was determined for the relevant peaks. The results are evaluated by means of the Gallon equation. The theoretical surface composition is calculated for the ideal solution model and compared with the experimental resultd. A pronounced surface enrichment in Pt is observed, in agreement with the theoretical predictions. Experiments with propane adsorption have been performed and a possible explanation of the results is offered. Results obtained are found to correspond well with literature data on catalytic reactions.

The results relevant for the theory of catalysis and of interest for this Conference, obtained with model alloys (active metal with an inactive one) are reviewed first. It is argued that of the two effects of alloying—ensemble size effect and the electronic (ligand) effect— the first one is more important. Recent results obtained with Pt/Re alloys (two active metals) are briefly summarized and a conclusion is made that the results can be rationalized by ideas reviewed in the first part of this paper.

Selective chemisorption of ethene and the chemisorption induced surface segregation of Pt upon the chemisorption of ethene (and its subsequent decomposition) on Pt-Cu alloys have been studied by Auger electron spectroscopy. The experimental results have also been evaluated by computer simulations of the various possible processes upon the chemisorption. In these computer simulations a modified Gallon model has been used in order to describe the variations in the intensities of the Pt and the Cu surface sensitive Auger transitions. Our present results indicate a preferential chemisorption of the carbonaceous species onto Pt sites, and a chemisorption induced surface segregation of Pt. Alloying of Pt with Cu results in a decreased coking of the catalyst.

The adsorption of CO on alumina-supported Pt-Cu alloys was studied. The results were compared with those obtained on the unalloyed Al 2 O 3 -supported metals as well as with the results on alloys supported on silica. By use of the isotopic dilution method it has been established that the downward shift of the frequency of the CO Pt band observed upon alloying Pt with Cu is “geometrical,” caused by dilution of Pt by Cu; an electronic or ligand effect was not observed (within the limits of error). The upward shift of the frequency of the CO/Cu band upon alloying Cu by Pt is most probably caused by “individualization” of the Cu atoms in the Pt layer.