Model Systems in Heterogeneous Catalysis: Selectivity Studies at the Atomic Level
- 1.1k Downloads
Reaching 100% selectivity is the theme of the twentyfirst century in heterogeneous and heterogenized homogeneous catalysis. To study chemical reactivity at the atomic scale, model systems have been prepared and characterized. We discuss selectivity of hydrogenation and dehydrogenation reactions on supported Pd particles, methanol oxidation at vanadium oxide model catalysts and, at last, the design of model catalysts with a well defined charge state of the metal, i.e., Au catalyst model systems.
KeywordsCatalysis Selectivity Nanoparticles Adsorption Reaction
Activity has always been an important characteristic for a heterogeneous or homogeneous catalyst . Modifying a catalyst in such a way that it becomes 100% selective for a given chemical reaction is a task for the twentyfirst century. Understanding selectivity at the atomic scale is therefore important. In general, heterogeneous catalysts are complex materials and it is difficult to achieve this goal due to the presence of many different, possibly active sites. In this respect, a homogeneous catalyst represents a rather well-defined material exposing single sites, thus favoring high selectivity. There have been many attempts to heterogenize homogeneous catalysts by binding a homogeneous active complex to a solid surface in order to combine the advantages of homogeneous and heterogeneous catalysts . In order to get insight into the atomic detail of such classes of catalysts strategies to circumvent the complexity of the “real” catalyst and to design models with a sufficient degree of complexity are a necessity.
Here, we will discuss a few recent examples from the author’s laboratory, i.e., methanol dehydrogenation on Pd nanoparticles [8, 9], methanol oxidation on vanadia surfaces , ethene hydrogenation on Pd nanoparticles [11, 12, 13, 14], and CO adsorption on Au atoms and clusters [15, 16, 17, 18].
The experiments were performed in a variety of ultrahigh vacuum systems built and situated at the Fritz Haber Institute of the Max Planck Society in Berlin. For details please consult the literature [5, 19].
The first example discussed here involves the oxidative dehydrogenation of methanol to CO/CO2 and H2O. A detailed kinetic study using molecular beams  has demonstrated that the selectivity of the reaction
So far we have discussed particle size, the presence of particle-specific defects and surface modifiers as principles to control the reactivity and selectivity of a catalyzed reaction. There may be other parameters as well that would be important, to control. One such parameter is the charge on the supported metal. It has recently been proposed that the charge may be controlled by depositing metals on oxide thin films of well-defined thickness which are themselves deposited on a metallic support . The concepts are similar to those used in semiconductor industry. Learning to modify band offsets at underlying interfaces by controlling interfacial dipoles may hold great promise in catalysis as well.
In the following I will briefly discuss the concepts and the presently available knowledge.
It has been recognized that the thickness of oxide films, as they are grown on metal substrates, may be used as a design parameter to create materials of potential in catalysis. A concept to control the catalytic activity of a dispersed metal by the thickness of an insulating oxide layer was introduced some 20 years ago by Maier and co-workers using silica layers covering Pt [56, 57, 58, 59]. In this case the catalytic performance for dehydrogenation of cycloalkenes depends significantly on the thickness of the silica films, which was explained by the decreasing transport of hydrogen atoms produced at the Pt surface through the silica film. The main experimental challenge, however, is to provide undisputable proof for the proposed mechanisms which is usually hampered either by the complexity of the samples investigated and/or the lack of appropriate methodology to exclude interference with alternative mechanisms. Developing experimental control together with realistic theoretical modeling, with respect to the thickness and structure of the oxide films , allows these structural properties to be used to control their functional characteristics and, thus, the catalytic properties of a metal deposited onto them. Here, control may concern transport of species, e.g., hydrogen or oxygen, through the film or it may concern electronic interaction of the interior metal–oxide interface with adsorbed species on top.
On the basis of density functional theory (DFT) calculations it was recently proposed by Pacchioni and his co-workers that charge transfer may also occur for metal atoms (as opposed to oxygen in the case of the Cabrera-Mott mechanism) and metal clusters adsorbed on a supported thin oxide film, provided that the adsorbed metal exhibits a high electronegativity and the oxide film does not exceed a few mono-layers [15, 18]. In particular, Au atoms adsorbed on thin MgO(001) films grown on Mo(001) and Ag(001) are expected to be negatively charged in contrast to their counterparts on the corresponding bulk MgO (or thick MgO films) [15, 18]. In the latter case it was proven that Au atoms are essentially neutral . Thus, the thickness of the oxide film may serve as a parameter to tune the electronic properties of supported metals. Pacchioni and co-workers discussed the phenomenon in terms of the modification of the work function in the thin film system and the electronegativity (electron affinity) of the adsorbed species . It is also connected to the ability of thin films to structurally relax upon charge transfer, i.e., the phenomenon contains a polaronic component stabilizing the charge transfer .
Recently, it was demonstrated through modeling how different oxides may lead to drastic changes of the potential step . The reasons for the predicted changes are very different, depending on the type of oxide.
Apparently, ionic oxides such as MgO cause a decrease of the surface potential step of metals such as Ag or Mo because the O2− ions polarize the metal electrons away from the interface, effectively modifying the electron density and thus depleting the metal surface of electrons . This effect strongly depends on the distance between the ionic layer and the top metal layer. In the case of more covalent oxides this effect is less strong. Here, the oxygen ions still have some electron accepting ability leading to an accumulation of charge in the local metal substrate-oxygen bonds. This, in turn leads to an increase in the surface potential step. SiO2 appears to be a good example for the latter case. In addition, there is an interesting trend observed in the theoretical predictions on the change of level alignment as the number of oxide layers increases: except for the case of a single oxide layer, which establishes a special situation, there is little change in the level alignment as the film gets thicker than two layers . However, as the film thickens it loses its ability to structurally relax.
With all this information in mind, it is foreseen, that by controlling the thickness of an oxide layer one has a unique opportunity to control charge transfer from the oxide-modified metal support to an adsorbate without using electrical control but by pure chemical means. If one could ensure stability under reaction conditions, thin oxide films on metallic supports would present a possibility to design model catalyst supports which control the charge state of an adsorbed species depending on their electronegativity. Of course, based on what has been discussed above, the oxide film cannot be thicker than the tunneling length. When growing oxide layers with thicknesses well exceeding this length, the relation of work function and electronegativity would still be favoring charge transfer to or from the adsorbed species but the electronic interaction would be shut off.
This exemplifies another important control parameter: the buried metal/oxide interface may be used to control the shape of the adsorbed species! Recently, Ricci et al.  predicted that Au nanoparticles change shape when they are deposited on a thin MgO film supported on a metal instead of a bulk MgO surface. Apparently, this change is induced by electrostatic interaction between the underlying metal and the metal-induced excess charge at the cluster oxide interface. This is accompanied with a crossover from 3-D geometries to 2-D structures. These 2-D Au cluster structures have also been experimentally observed in the gas phase . There is now experimental evidence that corroborates this prediction for MgO thin films .
In order to open up opportunities in applications, ultrathin oxide film systems need to be prepared on granular (powder) substrates. There are thin film oxide systems exhibiting self limiting growth. For example, alumina films grown on a NiAl alloy surface show such behavior . In this respect, metal alloy systems, in general, could offer possibilities.
The examples discussed indicate that model studied may provide useful information on catalytic reactions, even under realistic, i.e., technically relevant conditions, provided the systems studied contain the appropriate degree of complexity. In particular structure/morphology—spectroscopy and—reactivity relations, which are difficult to reach at the atomic level with real powder samples, allow us to achieve a new level of insight into heterogeneous reactions. A number of parameters influencing selectivity may be varied independently for model systems allowing clarification of heuristic concepts in heterogeneous catalysis.
The author is grateful to his collaborators whose names appear in the references as well as to a number of agencies including the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie. Particular thanks go to Gianfranco Pacchioni and Joachim Sauer for stimulating discussions on theory.
This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.
- 1.Ertl G, Knözinger H, Weitkamp J (eds) (1997) Handbook of heterogeneous catalysis, chapter 1. Wiley-VCH Verlagsgesellschaft mbH, WeinheimGoogle Scholar
- 7.Freund H, Goodman DW (2007) In: Ertl G, Knözinger H, Schüth F, Weitkamp J (eds) Handbook of heterogeneous catalysis, 2nd edn. Wiley-VCH Verlagsgesellschaft mbH, WeinheimGoogle Scholar
- 31.Kaya S (2007) Thesis, Humboldt -Universität zu Berlin, BerlinGoogle Scholar
- 32.Romanyshyn Y, Guimond S, Göbke D, Sturm M, Kuhlenbeck H, Freund H-J, Ganduglia-Pirovano MV, Döbler J, Sauer J, Top Catal (submitted)Google Scholar
- 35.Risse T, Freund H-J (2005) In: Copéret C, Chaudret B (eds) Topics organometallic chemistry, vol 16. Springer Verlag, p 117Google Scholar
- 41.Fukutani K (2001) Solid State Phys 36:353Google Scholar
- 43.Wilde M, Naschitzki M, Fukutani K, Freund H-J (in press) Phys Rev BGoogle Scholar
- 55.Brandt B, Ludwig W, Fischer K, Schauermann S, Zaera F, Freund H-J J Phys Chem C (submitted)Google Scholar
- 57.McEwen AB, Maier WF, Smith DA (1987) In: Ward JW (ed) Catalysis. Elsevier Science Publishers B.V., Amsterdam, p 823Google Scholar
- 58.Fleming RH, Baumann SM, Meeker GP, Maier WF (1987) In: Benninghoven A (ed) Secondary ion mass spectrometry: SIMS VI. John Wiley and Son Ltd, ParisGoogle Scholar
- 65.Nilius N, Ganduglia-Pirovano MV, Brazdova V, Kulawik M, Sauer J, Freund H-J Phys Rev Lett (submitted)Google Scholar