Chemical reactivity of cation-exchanged zeolites
Citation for published version (APA):Pidko, E. A. (2008). Chemical reactivity of cation-exchanged zeolites. Technische Universiteit Eindhoven. https://doi.org/10.6100/IR632877
DOI:
10.6100/IR632877
Document status and date: Published: 01/01/2008
Document Version:
Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)
Please check the document version of this publication:
• A submitted manuscript is the version of the article upon submission and before peer-review. There can be important differences between the submitted version and the official published version of record. People interested in the research are advised to contact the author for the final version of the publication, or visit the DOI to the publisher's website.
• The final author version and the galley proof are versions of the publication after peer review.
• The final published version features the final layout of the paper including the volume, issue and page numbers.
Link to publication
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal.
If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement:
www.tue.nl/taverne
Take down policy
If you believe that this document breaches copyright please contact us at:
openaccess@tue.nl
C
HEMICAL
R
EACTIVITY OF
C
ATION
‐E
XCHANGED
Z
EOLITES
PROEFSCHRIFT
ter verkrijging van de graad van doctor aan de
Technische
Universiteit
Eindhoven, op gezag van de
Rector Magnificus, prof.dr.ir. C.J. van Duijn, voor een
commissie
aangewezen
door het College voor
Promoties in het openbaar te verdedigen
op donderdag 13 maart 2008 om 16.00 uur
door
Evgeny Alexandrovich Pidko
prof.dr. R.A. van Santen
en
prof.dr. V.B. Kazansky
Copromotor:
dr.ir. E.J.M. Hensen
Evgeny A. Pidko
A catalogue record is available from the Eindhoven University of Technology Library ISBN: 978-90-386-1210-2
Copyright © 2008 by Evgeny A. Pidko
The work described in this thesis has been carried out at the Schuit Institute of Catalysis within the Laboratory of Inorganic Chemistry and Catalysis, Eindhoven University of Technology, The Netherlands. This research was partially supported by the National Computing Facilities Foundation (NCF), which provided computational facilities, with financial support from The Netherlands Organization for Scientific Research (NWO).
Cover design: Tom Bongers (Creanza Media Eindhoven B.V.) and Evgeny A. Pidko Printed at the Univesiteitsdrukkerij, Eindhoven University of Technology
C
ONTENTS
C
HEMICAL
R
EACTIVITY OF
C
ATION
‐E
XCHANGED
Z
EOLITES
Chapter 1 Introduction ... 1
Chapter 2 Confined space-controlled olefin–oxygen charge transfer in zeolites ... 15
Chapter 3 Molecular recognition of N2O4 on alkali-exchanged low-silica zeolites X ... 25
Chapter 4 The interplay of bonding and geometry of the adsorption complexes of light alkanes within cationic faujasites ... 43
Chapter 5 Molecular and dissociative adsorption of ethane on zinc and cadmium ions in ZSM-5 zeolite ... 61
Chapter 6 Catalytic dehydrogenation of light alkanes over zinc cations in Zn/ZSM-5 zeolite ... 79
Chapter 7 Ethane dehydrogenation over reduced extra-framework gallium cations in ZSM-5 zeolite ... 95
Chapter 8 Dehydrogenation of light alkanes over isolated gallyl ions in ZSM-5 zeolite ... 113
Chapter 9 Multinuclear gallium-oxo cations in high-silica zeolites ... 127
Summary ... 143 Samenvatting ... 147 Резюме ... 153 Acknowledgements ... 159 List of publications ... 161 Curriculum Vitae ... 163
"H-how can you look for a solution, where it d-does not exist? It's s-some sort of n-nonsense.”
"Excuse me, Feodor, but it's you who are reasoning strangely. It's nonsense to look for a solution if it already exists.”
Arkadi and Boris Strugatsky
C
HAPTER
1
I
NTRODUCTION
1.1.
Zeolites
eolites appeared on the scientific stage in the middle of 18 century, when the Swedish scientist Axel Frederik Cronstedt reported on an interesting behavior of a new mineral. Upon rapid heating the material came to life. Its particles began to “dance” and gas bubbles were released. Combining Greek words zein, "to boil", and lithos, "a stone", he called the mineral zeolite [1]. The researcher could not foresee that what he discovered and named would later be so widely applied to many important processes such as gas separation, softening of water, catalysis in petroleum and fine chemistry.
Zeolites are crystalline microporous alumino-silicate solids containing cavities and channels of a molecular size with an overall composition similar to that of quartz or sand (SiO2) but with some of the silicon atoms in the framework replaced by aluminum. The
resulting negative charge on the framework is compensated by an exchangeable cation. These positive extra-framework ions are rather loosely held and can readily be exchanged for others in a contact solution or via other chemical treatments.
The zeolite framework is built of silicon- or aluminum-occupied oxygen tetrahedra. The central atom of these tetrahedra is generally called a atom. The arrangement of the T-atoms follows the Löwenstein rule [2] that states that only Si-O-Si and Si-O-Al linkage are allowed, while Al-O-Al moieties cannot occur. It follows that the molar silicon to aluminum ratio (Si/Al) in the resulting solid must be larger than or equal to one.
Topologically, the zeolite framework can be considered as four-connected nets, where each vertex (T-site) is connected to its four closest neighbors via oxygen bridges (Figure 1.1). One can imagine an unlimited number of possible zeolitic structures. However, only a limited fraction of these are of interest, because they exhibit important properties, with even smaller portion being amenable to synthesis. Currently more than 150 zeolite types have been synthesized and 48 naturally occurring zeolites are known. The latter are rarely pure and are contaminated to varying degrees by other minerals, metals, quartz or other
zeolites. For this reason, natural zeolites are excluded from many important commercial applications, while synthetic zeolites have found large scale applications as ion-exchangers and catalysts.
Zeolites and related materials are widely applied in different technological fields. They act as efficient heterogeneous catalysts, mainly as solid acids, as adsorbents, and as molecular sieves in gas separation and purification. Such a broad spectrum of applications results from the possibility to rather easily tune the chemical and physical properties of the material by choosing the particular zeolite structure (topology) with required type and size of cavities and/or channels, and by introducing extra-framework species into the zeolite matrix. Depending on the type of species introduced, the physicochemical properties and the reactivity of the resulting material can vary drastically. Therefore, a deep understanding of their role in different chemical process is crucial for the rationalization of the behaviors of these materials, which can form the basis for the design of improved and novel systems.
1.2.
Extra-framework species in zeolites
The chemical reactivity of zeolites is usually associated with the existence of the exchangeable cations. If the compensation of the framework charge due to the isomorphous substitution of silicon with aluminum is provided by a proton, these protons sit on bridging oxygen atoms connecting framework silicons and aluminums. The resulting bridging hydroxyl groups exhibit pronounced Brønsted acidic properties. The strength of the thus formed solid acid depends on various factors such as the zeolite topology, the Si/Al ratio, the existence of additional extra-framework species, etc, and can be in some cases even as high as that of classic superacids [3-5]. The form of zeolites, in which the
negative framework charge is compensated by protons, is usually called the “hydrogen form”. The existence of the acid sites of a variable strength in the solid governs the field of applications of the hydrogen forms of zeolites that is acid catalysis [4,5].
Exchanging the Brønsted acid protons by metal cations, one can further influence chemical properties of zeolites. Introduction of metal ions such as zinc, cadmium, or gallium ions creates new Lewis acid sites within the zeolite host and opens a wide field of applications of these microporous materials in the reactions catalyzied by Lewis acids [4-7]. On the other hand the incorporation of cations of variable valence such as Cu+, Co2+, Fe2+, Mo6+, or V5+ enhances the activity of zeolites in redox reactions [4]. Stabilization of rather inert alkaline and alkali-earth cations in zeolite results in an enhancement of the basic properties of lattice oxygens [8]. In addition the charged species in the zeolite usually produce a rather strong electrostatic field that, in principle, can polarize molecules confined in the microporous matrix resulting in their activation [9,10].
The chemical properties of the extra-framework cations in zeolite depend strongly on the structure and topology of the zeolitic site accommodating the exchangeable species, as well as on the nature and structure of the cation itself. The zeolite framework can be viewed as a ligand that stabilizes the exchangeable cations, thereby influencing their chemical properties [11].
Univalent cations usually sit in zeolitic rings and coordinate to 2-3 lattice oxygen atoms
Figure 1.2. Compensation of the negative framework charge in cation sites of the zeolite with 5 or 6 T‐atoms: a bivalent M2+ cation in (a) a five‐ring and (b) a six‐ring; (c) a univalent (M2+‐OH‐)+ ion stabilized in a five‐ring; (d) a cationic (M2+‐O2‐‐M2+)2+ complex stabilized by two negative framework charges, and a less conventional alternative structure (e), where the charge in two adjacent five‐rings is compensated in a charge‐alternating manner by a bivalent M2+.
attached to [AlO2]– framework units. Such structures are easily realized within zeolites
independently of their Si/Al ratio, because one cation compensates for the charge of a single aluminum-occupied oxygen tetrahedron. In contrast, stabilization of multivalent cations in zeolites requires the existence of two or more closely located framework anionic sites. Thus, dependent on the Si/Al ratio and, hence, on the distance between framework Al ions, charge-compensation may be realized in various manners. For example, bivalent cations can be stabilized in the vicinity of two closely located [AlO2]– units (Figure 1.2 (a)
and (b)). Such sites are usually named conventional ion-exchangeable sites. This type of charge-compensation is generally accepted and is typical for low-silica zeolites, where almost all cation sites contain two or more aluminums. When direct charge-compensation is more difficult, as for instance in high-silica zeolites, alternatively multinuclear cationic complexes such as oxo- or hydroxo-substituted species (Mg-OH+ [12], Zn-O-Zn2+ [13], GaO+ [14], etc) can be formed (Figure 1.2 (c), (d)). In addition, it has been proposed recently [15] that bivalent mononuclear cations can, in principle, compensate for the charge of two distant [AlO2]– framework units (Figure 1.2 (e)). According to this model a
bivalent cation is located in the vicinity of one aluminum-occupied oxygen tetrahedron with additional coordination to neighboring basic oxygens of neutral [SiO2] units, whereas
the charge of the distantly placed anionic site is compensated indirectly. Depending on the type of charge-compensation, the chemical properties and catalytic reactivity of the extra-framework cationic species, and hence, of the cation-exchanged zeolite can vary substantially. However, due to a complex nature and composition, as well as inhomogeneity of most of the zeolite-based catalysts, the structure and properties of the exchangeable intrazeolite species are often unknown.
In summary, extra-framework cations modify the chemical and catalytic properties of microporous zeolites profoundly. Compared to our understanding of Brønsted acidity in zeolites and the resulting catalytic reactivity, our knowledge of the structure and reactivity of the modifying metal-containing cationic species is much more limited. Atomistic simulations of the stability of extra-framework cations in zeolites, their complexes, and their chemical reactivity are therefore instrumental to understand their catalytic properties and to improve the performance of the corresponding catalysts.
1.3.
Structural features of zeolites
Besides extra-framework species, the reactivity of zeolites is also dependent on the topological arrangement of the structure forming tetrahedra. It is well known that the selectivity of catalytic reactions over acidic zeolites often depends strongly on the size and shape of the zeolite channels. This is called shape selectivity [16]. Usually three types of shape selectivity are distinguished. The first two types, i.e. reactant and product shape selectivities, are based on the principle of molecular exclusion or molecular sieving. Selectivity in the former case is achieved by selecting a zeolite structure that can differentiate molecules of different size. Only those molecules, which dimensions are smaller than the pore entrances, can reach the internal active sites and be converted to the
Figure 1.4. Structure of mordenite. The highlighted region is the low‐symmetry smaller unit cell used in DFT calculations.
Figure 1.3. Structure of ZSM‐5 zeolite. The highlighted part of the zeolite shows the cluster model used in DFT calculations. Only the silicon‐ aluminum framework is shown, while oxygens are omitted for clarity.
desired products, whereas the larger molecules remain intact. Product selectivity, on the other hand, occurs when only certain product molecules formed in the micropore space can leave the zeolite crystals through the pores. The product molecules that remain in the zeolite undergo secondary reactions and subsequently desorb or slow down the reaction and in some cases even deactivate the catalyst. The third type of shape selectivity is transition-state selectivity that refers to steric constraints imposed by the zeolite structure on the transition state for the formation of certain product molecules. Although reactant and product molecules may freely diffuse into and out of the zeolite, the effective diameter of the channel system may hinder the formation of the bulky transition states or reaction intermediates, resulting in selective formation of only a fraction of possible reaction products. It should be noted that in all these cases the reaction selectivity is controlled only by the zeolite topology, while the nature of the active sites as well as their intrazeolite arrangement is assumed not to affect the selectivity of the catalytic process.
In the present work four types of zeolites of three different framework types are considered: ZSM-5 (MFI framework, Figure 1.3), mordenite (MOR framework, Figure 1.4) and zeolites X and Y (faujasite (FAU) framework, Figure 1.5). This selection is based on the fact that the above microporous materials are among the most widely used zeolites and are highly important for various chemical processes. In addition a vast majority of available experimental studies, needed to support the computational modeling, is devoted to the investigation of properties of these zeolite types.
The ZSM (Zeolite Socony Mobil) family zeolites have no natural analogues and have been synthesized for the first time by the Mobil Oil company in the 1970s [17]. Different types of ZSM zeolites (ZSM -4, -5, -8, -11, -12, -22, etc) can be prepared by varying the conditions of zeolite synthesis (temperature, ratio of components of the reaction mixture, type of template, etc). One of the most widely used zeolites is ZSM-5. Its unit cell is orthorhombic (Pnma, a = 20.07 Å, b = 19.92 Å, c = 13.42 Å) and has an ideal formula in the sodium form of NanAlnSi96-nO192·16H2O [18]. The atomic ratio of silicon to aluminum
Figure 1.5. Structure of faujasite. Left part shows the periodic structure (view along [110] plane). The light‐gray boxes represent the cubic Fd⎯3m unit cells; the darker boxes show a smaller rhombohedral cell used in periodic DFT calculations.
(Si/Al ratio) varies from 12 to ∞. Formation of the crystal lattice of ZSM-5 results in an intracrystalline system of intersecting channels, which have ten-membered ring openings. One channel system runs parallel to the a axis (view along [100], Figure 1.3) of the unit cell. It is sinusoidal and has elliptical openings of 5.1x5.5 Å. The other channels are straight, parallel to the b axis (view along [010], Figure 1.3) and have openings of 5.3x5.6 Å. The channel intersections have a critical dimension of nearly 9 Å.
Mordenite is a mineral zeolite, which has an ideal formula in the sodium form of Na8Al8Si40O96·16H2O (Si/Al=5). The crystal structure of mordenite has been reported for
the first time by Meier in 1961 [19]. The unit cell of mordenite is orthorhombic (Cmcm, a = 18.3 Å, b = 20.5 Å and c = 7.5 Å) with a unidimensional system of channels (Figure 1.4). The principal sorption channels are formed by twelve-membered rings of [TO2]
tetrahedra. The passageways are elliptical in shape and have openings of 6.5x7.0 Å.
Synthetic zeolites Y and X are the analogous of the naturally occurring mineral faujasite. The crystal structure has been reported for the first time by Bergerhoff et al in 1958 [20]. The unit cell is cubic (Fd⎯3m, a = b = c = 24.7 Å) [21]. An ideal formula of faujasite is
[Ca,Mg,Na2]29Al58Si134O384·16H2O. Zeolite Y has essentially the same framework structure
as zeolite X but differs in its Si/Al ratio, which varies from 2.2 to 3.0, as in natural faujasite, while this ratio in zeolite X is 1.1–1.5. The faujasite framework (Figure 1.5) may be described as a diamond array of cuboctahedral aluminosilicate units. This structural unit is usually referred to as the “sodalite unit” (cage), because of its occurrence in the mineral sodalite. Each sodalite unit is connected at its 6-memebred rings to four other sodalite units via distorted hexagonal prisms. This arrangement of structural units results in the formation of large absorption cages called “supercages” with diameter of about 13.7 Å. The faujasite unit cell consists of eight sodalite cages, and 16 hexagonal prisms. The supercages meet at windows of approximately 7.4 Å in diameter formed by 12-membered rings shared between them. One notes that the active sites located within the sodalite cages or hexagonal prisms are often inaccessible for the molecules adsorbed to the zeolite, because of the small size of the windows, while only those located within the channel system formed by the supercages are accessible and can directly influence the reactivity of the microporous material.
1.4.
Computational methods in zeolite science
Computational methods are now widely and extensively used in the chemical, physical, biomedical and engineering sciences in assisting the interpretation of experimental data and increasingly in predicting of the behavior of matter at the atomic level. They have a long and successful history of application in solid state and material science for modeling structural and dynamic properties of the bulk and surfaces of solids, and for understanding chemical reactivity, in which they are playing an increasingly important role. Their application to zeolite sciences developed strongly in the 1980s, with initial successes in modeling structure and sorption, and with an emerging capability in quantum mechanical methods.
Atomistic simulation methods can be divided into two very broad categories. The first is based on the use of interatomic potentials (force fields). These methods are usually empirical and no attempt to solve the Schrödinger equation is made. Instead the system of interest is described with mathematical functions, which express the energy as a function of nuclear coordinates. These may then be used to calculate structures and energies by means of minimization methods, to calculate ensemble averages using Monte Carlo simulations, or to model dynamical processes such as molecular diffusion via molecular dynamics simulations. The second category does not use empirical potentials to describe interactions between atoms but rather computes the electronic structure. On the basis of the quantum-mechanical description of the electrons of the chemical system, the solution to the multi-electron Schrödinger equation is approximated. Such methods are essential for the description of the elementary steps underlying catalysis, that is the making and breaking of chemical bonds. Hartree Fock (HF), Density Functional Theory (DFT), and post-Hartree Fock ab initio approaches have been commonly used in modeling zeolites, although DFT methods have been dominating the literature in the past decade.
This thesis is mainly aimed at the investigation of the chemical bonding and reactivity of the exchangeable cations or other cationic complexes encapsulated into the microporous zeolite matrix. These properties are directly related to the properties of the electrons of the chemical system, and moreover, can be described only when the electronic structure is taken into account. Here quantum mechanical methods play the pivotal role to calculate the electronic structure from which the most energetically stable geometry and the transition state for a particular reaction as well as their properties can be calculated. Such understanding is essential to interpret spectroscopic and catalytic data.
The goal of quantum-chemical methods is to predict the structure, energy and properties of a system containing many particles. In order to calculate the electronic states of the system, quantum chemical methods attempt to solve Schrödinger equation, Ĥ Ψ E Ψ, where Ψ is the wave function and E is the energy of the N-particle (electrons and nuclei) system. Ĥ is the Hamiltonian operator, which is comprised of the kinetic and potential energy operators acting on the overall wave function of the system. The exact solution for this equation can be found only for a very limited number of systems, and thus, a number
of approximations are required to solve it for larger systems. Below we present a simple overview of the most important approximations and of the most important limitations of the methods when applied to zeolite chemistry. More detail and in-depth discussion on the electronic structure calculations can be found in a number of very good references [22].
There are two important approximations to solve the Schrödinger equation for a multi-body system. The first is the Born-Oppenheimer approximation [23], which assumes uncoupling the electron motion from the nuclear motion since the mass of the latter species is much greater than the electron mass. The electronic wave function can then be solved separately from the nuclear one for a fixed set of nuclear positions. The second approximation is usually done in order to take into account repulsion between electrons. Since the electron-electron interaction cannot be directly solved, it has been common practice to consider each electron moving in a field of the other electrons of the system. The solution then requires convergence of the electronic structure via an iterative scheme. This is known as the self-consistent field approximation [22].
Electronic structure methods can be categorized as ab initio wave functions-based, density functional, and semiempirical methods. All of them can be applied to obtain insight into various issues concerning zeolite chemistry. Wave function methods start with the Hartree-Fock (HF) solution and use well-prescribed methods that can be used to increase its accuracy. One of the deficiencies of the HF theory is that it does not treat dynamic electron correlation, which refers to the fact that the motion of electrons is correlated so as to avoid one another. The neglect of this effect can cause very serious errors in the calculated energies, geometries, vibrational, and other properties.
There are numerous so-called post-Hartree-Fock methods for treating correlated motion between electrons. One of the most widely used approaches is based on the definition of the correlation energy as a perturbation. In other words, the configurational interactions are treated as small perturbations to the Hamiltonian. Using this expansion the HF energy is equal to the sum of the zero and first order terms, whereas the correlation energy appears only as a second order term. The second order Møller-Plesset perturbation theory (MP2) typically recovers 80-90% of the correlation energy, while MP4 provides a reliably accurate solution to most system. Due to the extremely high computational costs, the application of the post-HF methods in zeolite science is mainly limited to the MP2 method. A more attractive method is density functional theory (DFT). DFT is “ab initio” in the sense that it is derived from the first principles and does not usually require adjustable parameters. These methods formally scale with increase in the number of basis functions (electrons) as N3 and thus permit more realistic models compared to the higher-level post-HF methods, which usually scale as N5 for MP2 and up to N7 for such methods as MP4 and CCSD(T). On the other hand the theoretical accuracy of DFT is not as high as the higher level ab initio wave function methods.
The application of DFT is attributed to the work of Hohenberg and Kohn [24], who formally proved that the ground-state energy for a system is a unique functional of its
electron density. Kohn and Sham [25] extended the theory to practical applications by showing how the energy can be portioned into kinetic energy of the motion of the electrons, potential energy of the nuclear-electron attraction, electron-electron repulsion, which involves with Coulomb as well as self interactions, and exchange correlation that covers all other electron-electron interactions. The energy of an N-particle system can then be written as
(1.1)
Kohn and Sham demonstrated that the N-particle system can be written as a set of n-electron problems (similar to the molecular orbitals in wave function methods) that could be solved self-consistently in a manner that is similar to the SCF [25].
Although DFT is in principle an exact approach, a number of assumptions and approximations have to be made usually due to the fact that the exact expression for the exchange correlation energy is not known. The most basic one is the local density approximation (LDA), which assumes that the exchange-correlation per electron is equivalent to that in a homogeneous electron gas, which has the same electron density at a specific point r. The LDA is obviously an oversimplification of the actual density distribution and usually leads to overestimation of calculated bond and binding energies. The non-local gradient corrections to LDA functional improve the description of electron density. In this case the correlation and exchange energies are functionals of both the density and its gradient. The gradient corrections take on various different functionals such as P86 [26], B88 [27], PW91 [28], PBE [29], etc. However, the accuracy of these is typically less than what can be expected from high level ab initio methods.
One notes that the Hartree-Fock (HF) theory provides a more exact match of the exchange energy for single determinant systems. Thus, numerous hybrid functionals have been recently developed where the exchange functional is a linear combination of the HF exchange and the correlation (and exchange) calculated from LDA theory. The geometry and energetics calculated within this approach (B3LYP and B3PW91 [30], MPW1PW91 [31], PBE0 [32], etc) are usually in a good agreement with experimental results and those obtained by using post-HF methods. On the other hand, hybrid functionals still fail in describing of chemical effects mainly associated with the electron-electron correlation such as dispersion and other weak interactions [33,34].
As it was mentioned above, the energy in the DFT methods is formally a function of the electron density. However, in practice the density of the system ρ(r) is written as a sum of squares of the Kohn-Sham orbitals:
∑| | (1.2)
This leads to another approximation in both DFT and wave function-based methods that consists of the representation of each molecular orbital by a specific orthonormal basis set. The true electron structure of the model can in principle be mathematically represented by
an infinite number of basis functions. However, due to computational limitations, in practice these functions are truncated and described by a finite number of basis sets resulting in some loss in accuracy. A wide range of different basis sets currently exists and the choice for a certain one strongly depends on the solution method used, the type of the problem considered, and the accuracy required. These functions can take on one of several forms, including Slater-type functions, Gaussian functions, and plane waves. The basis sets build using these functions can be either full-electron basis sets, describing both core and valent electrons of the atoms, or so-called effective core potential (ECP) ones. In the latter case it is assumed that the core electrons do not influence significantly the electronic structure and the properties of atoms, and therefore, are being replaced with an approximate pseudopotential. Such a simplification is useful for the description of heavy atoms, because it decreases the number of basis functions, and correspondingly, the computational time without dramatic loss of accuracy.
The approximations done in order to solve Schrödinger equation by different quantum chemical methods as well as the use of finite basis set for the description of molecular orbitals are not the only factors leading to limited accuracy. When modeling zeolites, one can seldom take into account all of the atoms of the system. Typically, a limited subset of the atoms of the zeolite is used to construct an atomistic model. The size of the model used to describe reaction environment can be critical for obtaining of reliable results. A minimum requirement to the zeolite model is that it involves the reactive site or the adsorption site together with its environment, which gives rise to a so-called cluster approach. Here only a part of the zeolite containing finite number of atoms is considered, while the influence of the rest of the atoms of the zeolite lattice is neglected. Although this approach results in some los of “model” accuracy, it can be very useful for the analysis of different local properties of zeolites such as elementary reaction steps, adsorption, etc. The current progress in computational chemistry also made it possible to use rather efficiently periodic boundary conditions in DFT calculations of solids. This allows theoretical DFT studies of structure and properties of some zeolites with relatively small unit cells using an experimental crystal structure of the zeolite as a model.
In this thesis various physicochemical properties of very different zeolites are investigated by means of quantum-chemical modeling. Thus, the actual choice of the zeolite model and the theoretical method was mainly conditioned by a reasonable compromise between the accuracy needed and the computational time required to calculate particular molecular properties. In addition, the availability of theoretical methods (time-dependent DFT, post-Hartree-Fock methods, etc.) in a combination with the model limitations caused by the available quantum-chemical programs could influence the particular choice of the computational methodology. The reliability of the results obtained by a chosen method was usually tested by recalculation of the selected results at a higher level of theory and/or by comparison with the respective available experimental data.
mainly conditioned by the expected locality of the active site. For example, in the extreme case of the low-silica zeolite X modified with alkaline cations, most of the exchangeable cations of the faujasite supercage can be involved in the interaction with the adsorbed molecules. Therefore, in this case the cluster modeling approach would require a very large cluster model with the number of atoms even exceeding that in the rhombohedral periodic unit cell of faujasite. Thus, the periodic unit cell as a zeolite model becomes the natural choice. On the other hand, in the case of cation-exchanged high-silica zeolites the active site is usually expected to be rather local. Therefore, taking into account the small size of the reagents considered in this work, a reliable description of the active site together with the reaction environment could be done using a relatively small cluster model. Such a choice dramatically reduces the required computational time without a significant loss in the model accuracy, as compared to the case of the periodic model. This allows of a detailed and comprehensive analysis of various reaction paths possible over the active site of high-silica zeolite. At the same time, the application of a rather small cluster model makes it possible to use the higher level theoretical methods to improve the method accuracy.
1.5.
Scope of the thesis
This thesis deals with theoretical investigations of various aspects of chemical reactivity of cation-exchanged zeolites. Despite a wealth of literature in the field of zeolites modified with metal ions, there is still a lack of clear understanding of the role of the extra-framework species in catalytic reactions. The main goal of this thesis is to develop a deeper understanding of the structural and chemical properties of extra-framework cationic species in zeolites as well as of the mechanisms of chemical transformations catalyzed by such species.
This work is divided into two parts. The first part (Chapters 2 – 4) focuses on
investigations of the chemical properties of low-silica zeolites modified with typical hard Lewis acids such as alkaline and alkali-earth cations. Although these cations are rather inert, their high density in the zeolite can cause important chemical properties of the microporous matrix such as enhanced basicity of the framework and strong electrostatic field in the zeolite cages. More reactive soft Lewis acids such as Zn-, Cd- and Ga-cations stabilized in high-silica zeolites are discussed in the second part of the thesis (Chapter 5 –
9). In this case, special attention is devoted to the mechanism of C–H activation as well as to the stability and structure of the intrazeolitic cationic species.
Confinement and molecular recognition effects due to the specific arrangement and the size of the exchangeable cations in the zeolite matrix are discussed by examples of (i) photo-oxidation of alkenes with molecular O2, and of (ii) N2O4 disproportionation in
alkali-earth and alkaline-exchanged faujasites in Chapters 2 and 3, respectively. An attempt is made to separate effects of basicity of the framework, Lewis acidity of the
exchangeable cations, and the electrostatic field in the zeolite cage for the respective reactions.
Chapter 4 reports a combined FT-IR spectroscopic and quantum chemical investigation
of adsorption of light alkanes on magnesium- and calcium-exchanged Y-zeolites. In this chapter an attempt is made to understand how a light alkane molecule is adsorbed in a cation-exchanged zeolite cage and what the nature of the adsorption complex is. Based on the results of ab initio calculations (DFT and MP2) and topological analysis of the electron density distribution function in the framework of quantum theory of atoms in molecules, it is discussed how the geometry of the light alkane adsorption complexes depends on the extra-framework cation, which factors determine the particular adsorption fashion, and how this relates to the intermolecular interactions realized within the adsorption complex.
Adsorption properties of more reactive zinc and cadmium cations in high-silica ZSM-5 zeolites are studied in Chapter 5. The reactivity of different charge-compensating species toward heterolytic dissociative adsorption of light alkanes is investigated by means of DFT calculations. The results obtained are used to identify the factors, which control the activation of light alkanes over Zn- and Cd-exchanged ZSM-5. The influence of the perturbations of the adsorbed molecules due to interaction with the exchangeable cations on their subsequent chemical activation is discussed.
Chemical activation of light alkanes over soft Lewis acids in high-silica zeolites is further investigated in Chapters 6-8, which report comprehensive quantum chemical investigations of catalytic dehydrogenation of ethane over different cationic species in ZSM-5 zeolites modified with zinc and gallium. Various reaction paths for catalytic dehydrogenation of ethane, as a model reaction, are computed and analyzed. Chapter 9 examines the possibilities for formation of bi- and multinuclear cationic extra-framework species in high-silica zeolites modified with gallium. The concept of indirect charge-compensation is further developed in these chapters. It will be shown that such molecular understanding can help in the design of improved catalysts.
References 1 Cronstedt, A.F. Kungliga Svenska Vetenskapsakademiens Handlingar, Stockholm, 1756, 17, 120‐123. 2 Löwenstein, W. Am. Mineral. 1954, 39, 92. 3 Mirodatos, A.; Barthomeuf, D. J. Chem. Soc., Chem. Commun. 1981, 39. 4 Corma, A. J. Catal. 2003, 216, 298. 5 Stoeker, M. Microporous and Mesoporous Mater. 2005, 82, 257. 6 Hagen, A.; Roessner, F. Catal. Rev. 2000, 42, 403. 7 Serykh, A.I. Microporous and Mesoporous Mater. 2005, 80, 321. 8 Barthomeuf, D. Catal. Rev. 1996, 38, 521. 9 Blatter, F.; Sun, H.; Vasenkov, S.; Frei, H. Catal. Today 1998, 41, 297. 10 Frei, H. Science 2006, 313, 209. 11 Klier, K. Langmuir 1988, 4, 13. 12 Ward, J.W. J. Catal. 1968, 10, 34; Uytterhoeven, J.B.; Schoonheydt, R.; Liengme, B.V.; Keith Hall, W. J. Catal. 1969, 13, 425.
13 Penzien, J.; Abraham, A.; van Bokhoven, J.A.; Jentiys, A.; Müller, T.E.; Sievers, C.; Lercher, J.A. J. Phys.
Chem. B 2004, 108, 4116.
14 Dooley, K.M.; Chang, C.; Price, G.L. Appl. Catal. A: General 1992, 84, 17; Abdul Hamid, S.B.; Derouane, E.G.; Mériaudeau, P.; Naccache, C. Catal. Today 1996, 31, 327; Kazansky, V.B.; Subbotina, I.R.; van Santen, R.A.; Hensen, E.J.M. J. Catal. 2005, 233, 351. 15 Kazansky, V.B.; Serykh, A.I. Phys. Chem. Chem. Phys. 2004, 6, 3760; Kazansky, V.; Serykh, A. Microporous and Mesoporous Mater. 2004, 70, 151; Kazansky, V.B.; Serykh, A.I.; Pidko, E.A. J. Catal. 2004, 225, 369. 16 Chen, N.Y.; Degnan, Jr., T.F.; Morris Smith, C. Molecular Transport and Reaction in Zeolites. Design and Application of Shape Selective Catalysts, VCH Publishers, Inc., New York, 1994; Degnan, Jr., T.F. J. Catal. 2003, 216, 32.
17 Arguer, R.S.; Olson, D.H.; Landolt, G.R. G.B. Patent, 1969, 1161974; Argauer, R.S.; Landolt, G.R. U.S.
Patent, 1972, 3702886; Dwyer, F.G.; Jenkins, E.E. U.S. Patent, 1976, 3941871.
18 Olson, D.H.; Kokotailo, G.T.; Lawton, S.L.; Meier, W.M. J. Phys. Chem. 1981, 85, 2238. 19 Meier, W.M. Z. Kristallogr. 1961, 115, 439.
20 Bergerhoff, G.; Baur, W.H.; Nowacki, W. N. Jb. Miner. Mh. 1958, 193. 21 Baur, W.H. Am. Mineral. 1964, 49, 697.
22 Jensen, F. Introduction to Computational Chemistry, Wiley‐Interscience, New York, 1999; Leach, A.R.
Molecular Modeling: Principles and Applications, Pearson Education, Harlow, 1996; Foresman, J.B.;
Frish, A. Exploring Chemistry with Electronic Structure, 2nd ed., Pittsburg, PA Gaussian, 1996; Parr, R.G.; Yang, W. Density Functional Theory of Atoms in Molecules, Oxford University Press, New York, 1989; Young, D.C. Computational Chemistry: A Practical Guide for Applying Techniques to Real‐World
Problems, Wiley‐Interscience, New York, 2001. 23 Born, B.; Oppenheimer, J.R. Ann. Phys. 1927, 79, 361. 24 Hohenberg, P.; Kohn, W. Phys. Rev. 1964, 136, B864. 25 Kohn, W.; Sham, L. Phys. Rev. 1965, 140, A1133. 26 Perdew, J.P. Phys. Rev. B 1986, 33, 8822. 27 Becke, A.D. Phys. Rev. A 1988, 38, 3098 28 Perdew, J.P.; Chevary, J.A.; Vosko, S.H.; Jackson, K.A.; Pederson, M.R.; Singh, D.J.; Fiolhais, C. Phys. Rev. B 1992, 46, 6671. 29 Perdew, J.P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77, 3865. 30 Becke, A.D. J. Chem. Phys. 1993, 98, 5648. 31 Adamo, C.; Barone, V. J. Chem. Phys. 1998, 108, 664. 32 Adamo, C.; Barone, V. J. Chem. Phys. 1999, 110, 6158. 33 Johnson, E.R.; DiLabio, G.A. Chem. Phys. Lett. 2006, 419, 333. 34 Zhao, Y.; Truhlar, D.G. J. Chem. Theory Comput. 2005, 1, 415.
C
HAPTER
2
C
ONFINED SPACE
‐
CONTROLLED
OLEFIN
‐
OXYGEN CHARGE
TRANSFER IN ZEOLITES
DFT calculations on the initial charge-transfer step in the photo-oxidation of alkenes in cationic zeolites are presented. The used model system represents a part of Y-zeolite supercage containing two SII sites occupied by alkali-earth cations with 2,3-dimethyl-2-butene (DMB) and O2 adsorbed on them. It is found that the electrostatic field of the
zeolite cavity plays only a minor role for the stabilization of a [DMB+·O2–] charge-transfer
state, whereas the relative orientation and the distance between the DMB and O2 molecules
are the most important factors. On the basis of these results the photo-oxidation considered is due to a confinement effect, in which the adsorbed reagents are oriented in a suitable “pre-transition state” configuration.
2.1. Introduction
Recent photochemical studies of alkene-oxygen gas mixtures loaded in alkaline- and alkali-earth-exchanged zeolites have revealed that selective partial oxidation of unsaturated hydrocarbons can be induced by visible light [1-9]. The UV-Visible spectroscopic studies [6,10,11] have shown that a weak continuous absorption tail in the visible extending into the red spectral region appears only when both olefin and O2 are present in the zeolite. This
absorption is responsible for the light-induced oxidation of olefins.
The corresponding alkene·O2 contact charge-transfer bands in the liquid phase [12] and in
a solid oxygen matix [13,14] are well established and lie in the UV spectral range. For instance, the continuous absorption band attributed to the olefin·O2 charge transfer in the
case of 2,3-dimethyl-2-butene trapped in solid O2 has been detected at about 380 nm
[13,14]. On the other hand, the same molecules loaded in NaY zeolite exhibit a charge-transfer band at 750 nm [10]. It has been proposed that the interaction of the alkene·O2
contact pair with the strong electrostatic field of the cation-exchanged zeolite upon coadsorption results in a very strong stabilization of the charge-transfer state. Such stabilization is thought to cause the large red shift of alkene·O2 contact charge-transfer
transitions from the UV range into the visible range.
The excitation of the alkene·O2 contact pair results in the formation of an [alkene+·O2–]
charge-transfer state. When the hydrocarbon molecule is coordinated to a positively charged cation, this excited state will be strongly destabilized. On the other hand, one expects a very strong stabilization of the charge-transfer state when the O2 molecule is
adsorbed by the cation, while the hydrocarbon is located elsewhere far from the positively charged cationic adsorption sites and, at the same time, in the vicinity of the adsorbed dioxygen. This picture is not easily reconciled with the experimental fact that the adsorption of alkenes by the cation-exchanged zeolites is much stronger than O2 adsorption
and, hence, the alkene will replace the adsorbed oxygen.
To clarify this, a density functional theory (DFT) study of a model system containing 2,3-dimethyl-2-butene and O2 adsorbed in the Ca-, Mg- and Sr-exchanged supercage of
faujasite is performed. The calcium form of zeolite is chosen as a main object for the investigation, since the ionic radius of Ca2+ is similar to that of Na+, which has been used in the experimental studies. In addition, the cluster that models the FAU zeolite exchanged with bivalent cations can be chosen smaller, which reduces computational requirements.
2.2. Computational details
The quantum chemical calculations were carried out within the density functional theory (DFT) using the Gaussian 03 [15] program at the B3LYP/LanL2DZ level [16]. Earlier, the combination of the modest LanL2DZ basis set with the hybrid B3LYP functional was reported to provide reasonably accurate results in calculations of photochemical and adsorption properties [17-19]. The energy and the oscillator strength of the electron
Figure 2.1. Optimized structures and selected bond lengths (Å) of (a) MgZ, (b) CaZ and (c) SrZ cluster models. All of the interatomic distances presented are in angstroms .
excitations were estimated at the same computational level as the geometry optimization using the time-dependent DFT method that is implemented in the Gaussian 03 program package.
The M2Al4Si12O20H24 clusters, where M is Mg, Ca, and Sr, shown in parts (a)-(c) of
Figure 2.1, respectively, were chosen for the model DFT calculations. In the following these structures will be designated as MZ (M = Mg, Ca, or Sr). The cluster represents a part of the wall of the faujasite supercage, containing two 6T rings (SII sites) with a corresponding alkaline-earth cation in each 6T ring, connected via three adjacent 4T silicon rings. To stabilize exchangeable alkaline-earth cations, each 6T ring contains two aluminum atoms. The current MZ cluster was chosen because of the necessity of the existence of two adsorption sites in the zeolite model, while the use of the faujasite unit cell as a model was not possible due to the limitations of application of the time-dependent DFT method for the systems with periodic boundary conditions using the available quantum-chemical software.
The starting geometry of the clusters corresponds to the lattice of FAU zeolite according to X-ray diffraction (XRD) data [20]. Dangling bonds are terminated by H atoms located 1.4 Å from each terminal Si atom and 1.5 Å from each terminal Al atom oriented in the direction of the next T-site. Full geometry optimization was performed for the cluster models with adsorbed DMB and O2 molecules and for the clusters themselves, while the
positions of boundary H atoms were fixed according to the initial coordinates. Partial optimization of the DMB·O2 complex with a fixed distance between one of the O atoms
and a carbon from the C=C bond was performed in order to compare computational results with those obtained experimentally [14] for the complex stabilized in a solid O2 matrix.
Figure 2.2. Coadsorption of 2,3‐dimethyl‐2‐butene and O2 on (a) MgZ, (b) CaZ, and (c) SrZ. All of
the interatomic distances presented are in angstroms.
Table 2.1. Correcteda (∆EBSSE) and uncorrected (∆E) for the BSSE energies (kJ/mol) of individual
adsorptionb and coadsorptionc of O2 and DMB on MgZ, CaZ and SrZ clusters.
MgZ CaZ SrZ
∆E ∆EBSSE ∆E ∆EBSSE ∆E ∆EBSSE
DMB/MZ 63 44 81 60 86 65
O2/MZ 31 15 31 14 28 9
[DMB∙O2]/MZ 91 55 110 72 108 70
a
ΔEBSSE = ΔE – EBSSE b
ΔE = –{E([molecule]/MZ) – (E(molecule) + E(MZ))} c ΔE = –{E([DMB∙O
2]/MZ) – (E(DMB)+ E(O2) + E(MZ))}
The computed adsorption energies were corrected for basis set superposition error (BSSE) (EBSSE) using the counterpoise method [21]. The spin state of O2 was assumed to
be triplet in all of the computations presented bellow.
2.3. Results
Figure 2.2 (b) shows the optimized structure of 2,3-dimethyl-2-butene and dioxygen embedded in the CaZ cluster ([DMB·O2]/CaZ). The most important interatomic distances
are also displayed in Figure 2.2. Coadsorption of these molecules to a single exchangeable cation is energetically unfavourable due to the stronger adsorption of DMB (60 kJ/mol),
Table 2.2. The optimized C∙∙∙O distances (r) and the estimated energies (E) and oscillator strengths (f) of the charge transfer in the DMB⋅O2 complex. rC…O, Å E, eV f [DMB∙O2]/MgZ 3.957/3.807 0.80 0.007 [DMB∙O2] from MgZa 3.957/3.807 1.25 0.005 [DMB∙O2]/CaZ 2.877/2.860 1.67 0.078 [DMB∙O2] from CaZa 2.877/2.860 1.90 0.057 [DMB∙O2]/SrZ 3.782/3.754 1.21 0.008 [DMB∙O2] from SrZa 3.782/3.754 1.41 0.007 DMB∙O2 gas phase; partially optimizedb 2.150/2.655 2.69 0.099 4.22 0.070
a Data for the free DMB∙O
2 complex with the geometry obtained from the optimization of the [DMB∙O2]/MZ (M = Mg, Ca or Sr). b Partially optimized gas phase DMB∙O 2 complex with the constrained C∙∙∙O distance of 2.15 Å. Figure 2.3. Shape of the calculated orbitals involved in the DMB⋅O2 charge transfer, respectively, bonding
(a) and antibonding (b) molecular orbitals of the DMB⋅O2 contact complex
which suppresses that of O2 (14 kJ/mol). The energies of either individual adsorption or
coadsorption of DMB and O2 are listed in Table 2.1.
It is found that the DMB molecule is coordinated with the C1–C2 double bond to the Ca2+ cation (Ca2), whereas the O2 molecule is end-on adsorbed to another cation (Ca1).
The changes, which occur in the geometry of the adsorbed molecules due to interaction with the exchangeable cations, are insignificant. This is consistent with the mainly electrostatic nature of the interaction of either DMB or O2 molecules with the
alkaline-earth cations stabilized in the zeolitic cavity. For the adsorption complex, in which the DMB is coordinated to one Ca2+ cation and the O2 molecule to the other, the computed
energy of intermolecular DMB⋅O2 charge transfer is equal to 1.67 eV (745 nm). The
oscillator strength of it is f = 0.078 (Table 2.2). The shapes of molecular orbitals involved in this electron excitation process are shown in Figure 2.3. One can see the interaction of the highest occupied molecular orbital (HOMO) of the DMB with the lowest unoccupied molecular orbital (LUMO) of the dioxygen molecule. The orbitals involved in the charge-transfer process are, respectively, the
bonding and antibonding molecular orbitals of a DMB·O2 molecular
complex formed in the zeolitic cage. They represent a linear combination of the occupied πβ-orbital of the
DMB molecule and the unoccupied π*β-orbital of the O2 molecule. The
finite value of the oscillator strength is due to their small but significant overlap.
To investigate the effect of the electrostatic field in the zeolite cage on the stabilization of the charge-transfer state, the single-point time-dependent DFT calculation was performed for a free DMB·O2 complex with exactly the same geometry as in the [DMB·O2]/CaZ
system. In the absence of the zeolite cage, the computed charge-transfer energy (Table 2.2) slightly increases and becomes equal to 1.90 eV (653 nm), whereas the oscillator strength is only slightly lowered (f = 0.057).
Hence, it can be concluded that the low energy intermolecular charge transfer is due to the specific orientation between the DMB and O2 molecules in the ground state resulting
from their coadsorption to the exchangeable cations. To support this hypothesis, coadsorption by Mg2+ and Sr2+ stabilized in the same cluster model was also investigated. The optimized structures and the most important interatomic distances for the [DMB·O2]/MgZ and [DMB·O2]/SrZ are presented in parts (a) and (c) of Figure 2.2,
respectively. The energies of individual adsorption and coadsorption of DMB and O2 by
these clusters are listed in Table 2.1. Similar to the above-discussed case of CaZ model, coadsorption of 2,3-dimethyl-2-butene and dioxygen molecules by an isolated cation is not likely to occur, because of much stronger interaction of the alkene with the cation (Table 2.1) due to the higher basicity of the alkene molecule as compared to that of the O2.
Similar trends in the coordination of the DMB and O2 molecules to the cations are
observed. The alkene is adsorbed to the cation with the C=C bond, and the oxygen is adsorbed in an end-on fashion.
In spite of similarities in the adsorption fashions, in the case of the MgZ model, both the adsorbed DMB and O2 molecules are located closer to the cations in the cluster and the
distance between them is significantly larger. On the other hand, in the case of the [DMB·O2]/SrZ structure, the DMB molecule is coordinated to both strontium ions (Sr1 and
Sr2), while dioxygen is forced out from the cluster model, and again the distance between the adsorbed molecules is significantly increased as compared to that in the case of the CaZ. These effects are due to different ionic radii of the considered alkaline-earth cations (Sr > Ca > Mg) and, hence, to the different space between the adsorbed molecules. Besides this, a strongly different relative orientation between the alkene and O2 is detected in the
case of the magnesium containing cluster model. Instead of the end-on coordination of the O2 to the C=C bond of the DMB adsorbed to CaZ and SrZ, the interatomic contacts
between the O2’ atom from the dioxygen and H-atoms from the methyl groups of the DMB molecule are detected when coadsorbed to MgZ (Figure 2.2).
The computed properties of the intermolecular DMB⋅O2 charge transfer for the
[DMB·O2]/MgZ and [DMB·O2]/SrZ are listed in Table 2.2. One can see that the different
relative configuration of the O2 molecule to the alkene strongly influences both the energy
of the charge transfer and the oscillator strength. The increase of the distance between the DMB and O2 molecules results in substantial decrease of the oscillator strength and, hence,
of the probability of the intermolecular charge transfer. Surprisingly, increased separation of the adsorbed molecules from each other also results in significant decrease of the energy
Figure 2.4. Partially optimized structure of the isolated DMB⋅O2 complex with a
fixed C1‐O1 distance at 2.15 Å.
of the corresponding electron excitation. Most likely, it is connected to the fact that with a larger intermolecular separation no DMB·O2
complex is formed and, therefore, the ground state lies higher in energy. One should also note that a pure charge transfer between the DMB and O2 coadsorbed on MgZ is not detected, because
the excitation process additionally involves lone pairs of the basic oxygens of the cluster (O5 and O6). Nevertheless, in the absence of the zeolite framework, in all considered cases, the corresponding charge-transfer band is
blue-shifted and the oscillator strength slightly decreases (Table 2.2). The large difference between the charge-transfer energy in the free DMB·O2 complex and that embedded in the
MgZ cluster is most likely the result of the artifact connected with the involvement of the cluster model in the electron excitation process.
Hashimoto and Akimoto [14] estimated a distance of 4.1 Å between the positive (DMB) and negative ends (O2) of the dipole moment of the DMB·O2 charge-transfer complex in a
solid O2 matrix. One notes that this distance is realized when the shortest distance between
one of the carbons (C1) of the C=C bond of DMB and the interacting O1 atom equals 2.15 Å (Figure 2.4). In this case, two charge-transfer absorption bands with rather high values of the oscillator strength are computed: 2.69 eV (462 nm, f = 0.099) and 4.22 eV (294 nm,
f = 0.070). The latter value very well agrees with that reported in [14] (4.32 eV). On the
other hand the experimentally observed charge-transfer absorption spectrum [14] exhibits a very broad absorption band overlapping both values. It is also noticeable that the model used does not take into account additional interactions of the complex with the other surrounding dioxygen molecules of the O2 matrix, which can influence the geometrical and
electronic properties of the contact complex.
2.4. Discussion
A very strong decrease of the energy for intermolecular charge transfer between branched alkenes and oxygen embedded in alkaline-earth zeolites compared to the free state has been detected experimentally [6,10,11]. Indeed, UV-Visible spectroscopy of alkali- and alkaline-earth zeolite Y loaded with alkenes and O2 have revealed a visible absorption tail
to be attributed to the hydrocarbon·O2 contact charge-transfer transition [10]. On the other
hand, UV charge-transfer bands for similar contact complexes are well known in solid O2
matrixes [14]. It has been suggested that the interaction of the strong electrostatic field in cation-exchanged zeolites with the large dipole generated upon excitation of the hydrocarbon·O2 to the charge-transfer state leads to the stabilization of the excited state by
1.5-3 eV. This results in a strong red shift of the absorption from the UV region into the visible spectral region.
Indeed, the interaction of an O2 molecule with a zeolitic cation strongly increases its
electron affinity, but on the other hand, an adsorption of the hydrocarbon molecule to the exchangeable cations would lead to simultaneous increase of its ionization potential. Both of these effects cancel each other. It is known that alkenes are adsorbed on zeolites significantly stronger than oxygen. Thus, if one supposes that only the hydrocarbon molecules are adsorbed on the cation sites of zeolite and the O2 molecules are located
elsewhere, the electrostatic field will be directed opposite to the direction of the charge transfer. Therefore, the observed absorption band should be blue-shifted compared to that for the gas phase.
The results of DFT model calculations presented above show that the estimated energies and the probabilities of the charge transfer between 2,3-dimethyl-2-butene and dioxygen molecules with the same geometries are very close both in the presence and in the absence of the zeolite cage. It is evident that the major factor in the shift of the experimental absorption band is the relative geometry of the molecules. The specific relative orientation, which is due to adsorption to the closely located cation sites, results in the formation of a molecular complex with the excitation energy of visible light. Absorption bands in the absence of the zeolite matrix have been experimentally detected in a matrix of solid oxygen [13,14]. In this case the average distance between the hydrocarbon and oxygen molecules should be rather small and, in accordance with the presented theoretical results, this leads to a strong blue shift of the absorption band to UV range and to a significant increase of the probability of the electron transition. The model DMB·O2 complex (Figure
2.4) has a distance of about 4.1 Å between the positive and negative ends of the dipole, which are located at the opposite to O2 part of the alkene and at the non-interacting oxygen
atom (O2), respectively. Such contact complex shows a charge-transfer band of 4.22 eV. This value well agrees with that obtained experimentally (4.32 eV, Ref. [14]).
The specific orientation of the C=C double bond of the hydrocarbon to the O2 molecule
results in the formation of a molecular complex with an overlap of the πβ-orbital of the DMB molecule and the π*β -orbital of the O2 molecule. The optimum configuration is
found in the case of coadsorption of these molecules on calcium-exchanged faujasite. The adsorption of the DMB and O2 on the nearest exchangeable cations of the zeolite Y
supercage results in their confinement in specific orientation that is suitable for a rather effective overlap of corresponding HOMOs and LUMOs and, hence, for the intermolecular charge transfer. One notes that the C–O distances (2.877 and 2.860 Å) in the [DMB·O2]/CaZ are significantly lower than the sum of corresponding van der Waals radii
(3.1 Å, Ref. [22]). On the other hand, when due to steric factors such a suitable configuration between the adsorbed molecules cannot be realized (as is found for [DMB·O2]/MgZ and [DMB·O2]/SrZ), the molecular complex is not formed and the
effective charge transfer can not be observed. Thus, one can expect a much lower activity of MgY and SrY zeolites in the photo-oxidation of 2,3-dimethyl-2-butene in comparison with that of CaY.
2.5. Conclusions
The role of the zeolite in the photo-oxidation of alkenes with molecular oxygen is the complexation of the hydrocarbon and O2 to the extraframework cations, resulting in
confinement of these molecules with a specific relative orientation. This leads to the formation of a π-π intermolecular complex. The interaction between alkene and oxygen in this complex occurs with a finite overlap of the corresponding π and π* molecular orbitals. The formation of such a complex results in a significant transition moment of the intermolecular charge transfer. The zeolite matrix stabilizes the reagents in a suitable “pre-transition state” configuration. The role of the electrostatic field of the zeolite is only indirect. References 1 Frei, H. Science 2006, 313, 209. 2 Blatter, F.; Frei, H. J. Am. Chem. Soc. 1994, 116, 1812. 3 Blatter, F.; Sun, H.; Vasenkov, S.; Frei, H. Catal. Today 1998, 41, 297. 4 Sun, H.; Blatter, F.; Frei, H. J. Am. Chem. Soc. 1996, 118, 6873. 5 Sun, H.; Blatter, F.; Frei, H. Chem. Eur. J. 1996, 118, 6873. 6 Tang, S. L. Y.; McGarvey, D. J.; Zholobenko, V. L. Phys. Chem. Chem. Phys. 2003, 5, 2699. 7 Myli, K.B.; Larsen, S.C.; Grassian, V.H. Catal. Lett. 1997, 48, 199. 8 Larsen, R.G.; Saladino, A.C.; Hunt, T.A.; Mann, J.E.; Xu, M.; Grassian, V.H.; Larsen, S.C. J. Catal. 2001, 204, 440. 9 Xiang, Y.; Larsen, S.C.; Grassian, V.H. J. Am. Chem. Soc. 1999, 121, 5063. 10 Blatter, F.; Moreau, F.; Frei, H. J. Phys. Chem. 1994, 98, 13403. 11 Vasenkov, S.; Frei, H. J. Phys. Chem. B 1997, 101, 4539. 12 Coomber, J.W.; Hebert, D.M.; Kummer, W.A.; Marsh, D.G.; Pitts, Jr., J.N. Environ. Sci. Technol. 1971, 4, 1141. 13 Hashimoto, S.; Akimoto, H. J. Phys. Chem. 1986, 90, 529. 14 Hashimoto, S.; Akimoto, H. J. Phys. Chem. 1987, 91, 1347. 15 Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al‐Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, revision B.05; Gaussian, Inc.: Pittsburgh PA, 2003. 16 Becke, A.D. Phys. Rev. 1988, A38, 3098; Becke, A.D. J. Chem. Phys. 1993, 98, 1372; Becke, A.D. J. Chem. Phys. 1993, 98, 5648. 17 Ermakov, A.I.; Mashutin, V.Y.; Vishnjakov, A.V. Int. J. Quant. Chem. 2005, 104, 181. 18 Brocławik, E.; Borowski, T. Chem. Phys. Lett. 2001, 339, 433. 19 Song, X.; Liu, G.; Yu, J.; Rodrigues, A.E. J. Mol. Struct. (TEOCHEM) 2004, 684, 81. 20 Olson, D.J. J. Phys.Chem. 1970, 74, 2758. 21 Simon, S.; Duran, M.; Dannenberg, J. J. J. Chem. Phys. 1996, 105, 11024. 22 Bondi, A. J. Phys. Chem. 1964, 68, 441.
C
HAPTER
3
M
OLECULAR RECOGNITION OF
N
2
O
4
ON ALKALI
‐
EXCHANGED
LOW
‐
SILICA ZEOLITES
X
Adsorption and disproportionation of dinitrogen tetraoxide on sodium, potassium and rubidium exchanged zeolites X with Si/Al ratio of 1.18 are studied using density functional theory calculations with periodic boundary conditions. It is found that the stabilization and activation of most of the N2O4 isomers confined in the zeolitic cage does not follow the
differences in Lewis acidity of the extra-framework cations. This is also observed for the energetics of the N2O4 disproportionation reaction resulting in a spatially separated
NO+···NO3– ion pair. The reaction energy increases in the row NaX < RbX < KX. The
strength of perturbations and, therefore, the low-frequency shift of the N–O stretching frequency of the adsorbed NO+ cations correlate well with the basicity of the framework oxygens (RbX > KX > NaX). However, this factor is not the relevant reactivity parameter for the N2O4 disproportionation in the cationic zeolites. The higher activity for the
disproportionation as well as the stronger molecular adsorption of N2O4 on RbX and KX
zeolites as compared to that on NaX is ascribed to the features analogous to the molecular recognition characteristics of supramolecular systems. The steric properties of the zeolite cage and the mobility of the extra-framework cations induced by adsorption are essential to shape the optimum configuration of the active site for N2O4 disproportionation.
