The influence of adsorbate interactions on elementary reaction
kinetics : CO with NO, N, O, or H on Rh(100)
Citation for published version (APA):
Jansen, M. M. M. (2010). The influence of adsorbate interactions on elementary reaction kinetics : CO with NO, N, O, or H on Rh(100). Technische Universiteit Eindhoven. https://doi.org/10.6100/IR672695
DOI:
10.6100/IR672695
Document status and date: Published: 01/01/2010
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
providing details and we will investigate your claim.
The Influence of Adsorbate Interactions
on Elementary Reaction Kinetics
CO with NO, N, O, or H on Rh(100)
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 woensdag 28 april 2010 om 16.00 uur
door
Maarten Mathijs Marinus Jansen
Dit proefschrift is goedgekeurd door de promotoren:
prof.dr. J.W. Niemantsverdriet en
prof.dr. B.E. Nieuwenhuys
Jansen, M.M.M.
The Influence of Adsorbate Interactions on Elementary Reaction Kinetics: CO with NO, N, O, or H on Rh(100)
Eindhoven University of Technology, 2010
A catalogue record is available from the Eindhoven University of Technology Library
ISBN: 978-90-386-2204-0
Copyright © 2010 by Maarten M.M. Jansen
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.
Photograph cover: Bart van Overbeeke Printed at Wöhrmann Print Service, Zutphen
1. Introduction and scope 1
2. Experimental and computational details 9
3. Adsorption/desorption studies of CO on a rhodium(100) surface under UHV conditions: a comparative study using XPS, RAIRS, and SSIMS
29
4. Kinetic parameters from temperature programmed desorption spectra combined with energy relations: top and bridge CO on Rh(100)
45
5. Interactions between co-adsorbed CO and H on a Rh(100) single crystal surface
65
6. The influence of nitrogen atoms on the adsorption of CO on a Rh(100) single crystal surface
81
7. CO oxidation on a Rh(100) single crystal surface: a correlation between structure and reactivity
101
8. The interaction and reaction of co-adsorbed NO and CO on a Rh(100) single crystal surface
127
9. Extra: chemistry of ethylene glycol on a Rh(100) single crystal surface 143 10. Summary, general conclusions, and future prospects 153
Samenvatting 159
Curriculum vitae 163
List of publications 165
Introduction and scope
1.1 Catalysis: definition and use
Catalysis dates back to the dawn of civilization in the production of wine and beer. During fermentation, a biocatalyst converts sugars into alcohol. In that time, chemistry was practised more like alchemy in which the chemist tried to convert scrap metal into expensive noble metals like gold. Catalytic reactions were poorly or not all understood. The term catalysis was first defined by the Swedish chemist J.J. Berzelius in 1835.1 He described catalysis as a chemical event that changes the composition of a reaction mixture, while the catalyst itself stays intact. The word catalysis originates from the Greek verb ‘kata-luein’ which means to dissolve. The main function of a catalyst is to ‘dissolve’ or break and subsequently make chemical bonds to facilitate a chemical reaction.
Nowadays, over 85-90 % of our chemical products are produced with the aid of one or more catalysts. The impact of catalysis in daily life is well illustrated by the ammonia process developed by Haber, Bosch, and Mittasch at the beginning of the 20th century. This process was developed to provide Europe with a fertilizer (as ammonia is the feedstock of fertilizers) to prevent famine. However, ammonia was later also used for the production of nitrogen based explosives. In this example, it becomes clear that catalysis and chemistry in general can enrich human life, but at the same time they can be used for more harmful purposes. Other major fields in catalysis are the production of transportation fuels, plastics, pharmaceuticals, and in environmental catalysis (e.g. automotive exhaust gas catalysis). It is estimated that every euro spend on catalysis yields about 300 euro’s in products and services.
1.2 The principle of catalysis
A catalyst is a substance that accelerates a chemical reaction by providing a different reaction pathway than the gas phase reaction. Figure 1.1 shows an example of a catalytic reaction on a heterogeneous catalyst: the oxidation of CO on a transition metal surface. The reaction occurs in a catalytic cycle. In the first step: CO and oxygen adsorb on an empty catalyst surface. Bonds are made between the substrate and adsorbates, while simultaneously the molecular O2-bond is weakened and broken. Hereafter, a new bond is created between one of the oxygen atoms and CO forming CO2. In the last step, CO2 desorbs into the gas phase, returning the catalyst into its initial stage. The next cycle can now proceed.
The catalyst does not affect the thermodynamics of the overall reaction: the initial and final state of the reaction remain the same. This means that, in the presence of a catalyst, the equilibrium between reactants and products does not shift.2 However, the kinetics of the reaction does change: compared to the reaction in the gas phase, the activation barrier for the catalyzed reaction is significantly lower resulting in a much higher reaction rate at equivalent reaction conditions. One of the advantages of a catalyst is that reaction can proceed at much milder reaction conditions with less side reactions making the process more economical. There is, however, not a universally applicable catalyst for different reactions. The interaction between the reactants and products is very important to make a catalyst successful. If the interaction is too weak, the substrate will fail to break molecular bonds, but if on the other hand, the interactions are to strong, the adsorbates will poison the surface. For each catalytic reaction, a tailor-made catalyst should be designed with the magnitude of the interaction being just right. This is called the Sabatier principle.
O O O C O OC CO O CO O Gas phase activation barrier Catalytic activation barrier Poten tial energy Reaction coordinate
Figure 1.1 Potential energy diagram for CO oxidation in a catalysed and uncatalysed
reaction. The catalysed reaction proceeds via the catalytic cycle. The activation barrier with and without a catalyst are indicated.
1.3 Spectroscopy and catalysis
The goal of catalysis research is being able to design the ultimate catalyst for a particular chemical reaction. This means that the catalyst has a very high activity and 100% selectivity towards the desired product at room temperature and atmospheric pressure. The catalyst should not deactivate and should be easily separable from the product stream. To come closer to this goal, we need to fully understand how the catalyst behaves on a molecular scale so we can improve preparation routes, activation procedures, and prevent deactivation of the catalyst. In heterogeneous
catalysis, spectroscopic techniques are deployed to study the catalyst itself and the chemical processes occurring on the catalyst surface. Commercial catalysts are highly complex systems: the active component is deposited on a porous support together with promoters. It can become problematic to gain a fundamental understanding of how commercial catalysts under industrially relevant conditions work.
Many spectroscopic techniques require high vacuum conditions (p < 10-6 mbar) to ensure a sufficiently long mean free path of ions or electrons used as excitation source or detected species.3 This means that industrially relevant conditions cannot be applied. This is called the pressure gap. Only a limited number of spectroscopic techniques can be deployed at moderate to high pressures such as Infra-Red Spectroscopy, Solid State Nuclear Magnetic Resonance (NMR) and X-ray Absorption Spectroscopy (EXAFS and XANES).
The active catalyst particles are embedded in the pore system of the support, making it difficult to use spectroscopic techniques which often require a high accessible surface area. Non-porous model supports do not have these problems. A model support typically consists of a conducting substrate coated with a thin layer of supporting material and the active material deposited on top.4,5
The active catalyst particle itself is a complex system. The particle is influenced by the surrounding material (e.g. particle-support interactions) and by additives. The particles are present in several shapes and sizes depending on the gas environment, temperature, and preparation route. Each particle exposes several crystal planes. Catalytic reactions are structure sensitive and thus each crystal plane has a different reactivity. All these factors make it difficult to correlate macroscopic properties of the catalyst with its microscopic structure.
In this thesis, a very clean and well defined single crystal surface is used under ultrahigh vacuum conditions, to greatly simplify the system under study. Although single crystal research is far away from commercial catalytic systems, information can be obtained on an atomic scale. This enables comparison with theoretical modelling. Reactivity studies on single crystal surfaces can provide kinetic parameters for elementary reaction steps and can contribute to unravelling complex reaction mechanisms. Under industrial conditions, gas pressures are high which in general results in high surface coverages. To mimic industrial conditions in ultrahigh vacuum, low sample temperatures (e.g. 100 K) are used enabling high surface coverages. At high coverages, lateral interactions between surface species become important by influencing reaction kinetics. In single crystal chemistry, nearly all spectroscopic techniques can be used with easy manipulation of the crystal temperature and at high coverages.
1.4 Lateral interactions
Phenomena like diffusion, desorption, and elementary surface reactions on transition metal surfaces have a rate equation with a rate constant following the Arrhenius equation6:
RT Eact
e
k / (1.1)
The magnitude of the rate constant depends on the temperature, the pre-exponential factor, and the activation energy. The pre-pre-exponential factor and the activation energy in this equation are not necessarily constant and can deviate as a result of lateral interactions between adsorbates. The binding strength of an adsorbate can be perturbed by neighbouring species which changes the activation energy of the reaction it is involved in. Interactions can change the pre-exponential factor by affecting the entropy of the system. This occurs when, for example, molecular rotations become hindered or when molecular vibrations are affected.
There are two classes of lateral interactions: direct and indirect interactions.7 The direct interactions have a gas or liquid phase equivalent while the indirect interactions are through-surface or surface mediated. Direct interactions are subdivided into electrostatic interactions, hybridization interactions, and Van-der-Waals interactions. Electrostatic interactions occur when the adsorbate metal bond is ionic or when a dipole exists along the surface-adsorbate axis. Equivalent to hydrogen bonds, hybridization interactions occur when two molecules are in close proximity and their wave functions overlap to form new hybrid orbitals. Van-der-Waals interactions are very small compared to the other contributions and only dominate for adsorbed noble gases. The indirect interactions are subdivided into electronic through-surface and elastic interactions. Electronic through-through-surface interactions arise when an adsorbate changes the electronic structure of the substrate and thereby changes the substrate properties for additional adsorbates. Elastic interactions occur when an adsorbate changes the geometry of the substrate. A surface reconstruction is an extreme example of this interaction. In practice, there is no clear-cut distinction between the different contributions in the interaction energy and pre-exponential factor and only the overall change is obtained.
In experiments, lateral interactions are manifested in two different ways. First, they change the reaction kinetics. For example, repulsive interactions decrease the stability of the ground state and thereby decrease the activation energy of a reaction as shown in Figure 1.2. A decrease in activation energy increases the reaction rate and can make this elementary reaction step more favourable than a side reaction that is less affected by lateral interactions. Because lateral interactions change the rate equation, the most direct way to observe the effect of interactions is in Temperature Programmed Desorption (TPD) and Temperature Programmed Reaction Spectrometry (TPRS). The second effect of lateral interactions is the formation of adlayer structures. If adsorbates repel each other, they tend to be as far apart as possible. As
coverages increase, there is a most optimal surface structure in which the repulsive interactions are minimized. Because the substrate with its structure and adlayer strive for a certain overlap, discrete changes in adlayer ordering are observed. The technique to measure these ordered patterns is Low Energy Electron Diffraction (LEED). The surface adlayer needs to be well ordered to obtain a clear LEED picture. This clarifies the drawback of this technique: disordered adlayers and defects in the adlayer structure with a low concentration cannot be detected.
CO(g)
CO(a)
0 actE
actE
intE
Desorption
0 actE
Reaction
actE
CO + O
intE
int)
1
(
E
CO
2#CO
2Figure 1.2 Energy profiles for desorption and reaction, indicating how the activation
energies are affected by repulsive lateral interactions. The black curve is without interactions and the gray curve is with interactions. For simplicity, the effect of lateral interactions on the reaction is only shown in the initial state and not for the final state.
Throughout this thesis, beside LEED, a more indirect way is used to unravel the surface structure. If two adsorbates are co-adsorbed on a single crystal surface, one is atomic and the other molecular and the atom has a stronger interaction with the surface than the molecule, the atom will affect bonding of the molecule to the surface. This is especially the case at medium to high atomic coverages, where compact groups of atoms are formed. When a molecule adsorbs within this group and the interaction between species is repulsive, the molecule will reduce the repulsive interactions by maximizing the atom-molecule distances. The molecule has shifted from the most stable site in the absence of interactions to a different, less stable site in the presence of the atomic species. Changes in binding geometry of the molecule can be observed with vibrational spectroscopies like Reflection Absorption Infra Red Spectroscopy (RAIRS). The molecular stretching frequency is characteristic for the binding geometry which in turn is indicative of the way the atoms surround the molecule. A necessary condition is that the dipole moment of the vibration changes perpendicularly to the surface. This is the case with both CO and NO, which are therefore excellent probe molecules for studying interaction phenomena on surfaces.
The effect of lateral interactions can be obtained from experiments both quantitatively and qualitatively by using TPD and TPRS. The general procedure is as follows. From a TPD experiment, the desorption energy is extracted from a species at close to zero coverage to exclude the effect of lateral interactions. In a second step, the desorption energy is determined again after adding a species that interacts with the desorbing species. The difference between the two desorption energies is the total interaction energy. By taking into account the adlayer structure determined in a direct way with LEED or indirectly with RAIRS, the number of interaction pairs can be counted and the pair-wise lateral interaction energy can be calculated. In this case, the approach is quantitative. When the desorbing species has reacted before desorption, which is for instance the case in hydrogen recombination or CO oxidation, only a qualitative approach can be used. For example, the oxidation of CO of Figure 1.1 is a two step process starting from adsorbed O and CO. First a surface reaction occurs followed by desorption of the product. If the reaction step has a higher activation barrier than desorption, then the surface reaction will be the rate determining step. Figure 1.2 shows the way lateral interactions influence the energy profiles of desorption and surface reaction. For desorption, the lateral interactions act only on the adsorbates and not on the gas phase species. For the reaction, the lateral interactions act on both the adsorbates and on the transition state complex. How much the activation energy of the reaction is affected by the interactions depends on α, the Brønsted-Evans-Polanyi coefficient8,9: ) ( 0 0 ads ads act act E E E E (1.2)
in which α is defined as in Figure 1.2. EadsEads
0 is the difference in stability due to
the effect of interactions.
To determine this coefficient experimentally is very difficult. Hence, for surface reactions, the magnitude of the lateral interaction energy is not calculated. Only a qualitative description of the effect of lateral interactions is given by indicating the reactive configurations and how active they are.
1.5 Scope and structure of this thesis
The aim of this thesis is to describe the role and magnitude of lateral interactions in the kinetics of catalytic surface reactions. The automotive exhaust gas catalyst10 is used as the source of inspiration for the systems studied. Besides platinum and palladium, rhodium is applied as catalytically active component for the oxidation of CO and hydrocarbons and for the selective reduction of NOx. The single crystal surface Rh(100) is used as the catalytic surface and interactions between CO with NO, N, O or H are examined. A combination of experimental techniques and computational tools is used for visualization and quantification of the lateral interactions. The following experimental techniques were employed: TPD (or TPRS),
RAIRS and LEED. On the theoretical side, Density Functional Theory (DFT) has been applied to obtain structures, frequencies, and adsorption energies and kinetic Monte Carlo (kMC) has been used to obtain structure information and the magnitude of the kinetic parameters.
Chapter 2 gives the physical background of the experimental techniques and computational tools and the kind of information these techniques provide. In addition, a description is given of the ultra-high vacuum set-up, the crystal, and cleaning procedure.
Chapter 3 is an introduction to the behaviour of CO on Rh(100). Ordering of the CO adlayer is discussed and the way this influences the CO bonding site. A comparison is made between the activation energy of CO desorption obtained from TPD experiments and from experiments under equilibrium conditions.
Chapter 4 describes how lateral interactions between CO molecules can be extracted from TPD using kMC. Experimental desorption traces from TPD are fitted with desorption traces from kMC simulations by varying the pre-exponential factor, adsorption energies, and lateral interaction energies. The main difficulty is reproducing the experimentally observed adlayer ordering. An magnitude estimate of the lateral interactions is derived from a fitting procedure.
Chapter 5 deals with a novel adlayer structure for CO co-adsorbed with H on Rh(100), that has not been reported before. The distribution of CO and H within this structure and the way CO and hydrogen affect each other is discussed.
In Chapter 6, the lateral interaction energy between CO and atomic nitrogen is determined. Although LEED gives a clear c(2×2) pattern of the nitrogen adlayer, the defects in this adlayer structure harbour free sites for CO to adsorb onto. From the difference in CO stability and counting the number of nitrogen atoms surrounding the CO molecules in the defects, the CO/N pair-wise lateral interaction energy is obtained.
Chapter 7 describes the reaction between CO and atomic oxygen on Rh(100). A model is proposed in which the reactivity for CO oxidation depends on the CO/O configuration. Four different reactivity regimes are identified each with a different oxygen environment surrounding the CO molecules.
In Chapter 8, the decomposition of NO and the reaction between NO and CO is discussed. The key step in the reaction between CO and NO is breaking of the NO bond. The influence of CO and the decomposition products N and O on NO decomposition is made visible with vibrational studies.
In the extra Chapter 9, the decomposition of ethylene glycol is described. Ethylene glycol is a simple model molecule for polyalcohols such as sugars. Understanding how ethylene glycol decomposes is important for biomass conversion into synthesis gas or chemical intermediates.
Finally, Chapter 10 summarizes the most important results and conclusions. Also some perspectives for future research are proposed.
References
1. J.J. Berzelius, Royal Swedish Academy of Sciences, 1835.
2. I. Chorkendorff and J.W. Niemantsverdriet, Concepts of Modern Catalysis and
Kinetics, Wiley-VCH, Weinheim, 2003.
3. J.W. Niemantsverdriet, Spectroscopy in Catalysis: An Introduction; Third,
Completely Revised and Enlarged Edition, Wiley-VCH, Weinheim, 2007.
4. P.C. Thüne, C.P.J. Verhagen, M.J.G. van den Boer and J.W. Niemantsverdriet, J.
Phys. Chem. B, 1997, 101, 8559-8563.
5. L. Coulier, V.H.J. de Beer, J.A.R. van Veen and J.W. Niemantsverdriet, Top. Catal., 2000, 13, 99-108.
6. D.A. King, Surf. Sci., 1975, 47, 384-402.
7. C.G.M. Hermse and A.P.J. Jansen, Catalysis, 2006, 19, 109-163. 8. J.N. Brønsted, Chem. Rev., 1928, 5, 231-338.
9. M.G. Evans and M. Polanyi, Trans. Faraday Soc., 1938, 34, 11-24. 10. B.E. Nieuwenhuys, Adv. Catal., 1999, 44, 259-328.
Experimental and computational details
This chapter presents an overview of the experimental and computational techniques employed to study the adsorption, desorption, reaction, and interaction of different molecules and atoms on a Rh(100) single crystal surface. The experimental techniques are Temperature Programmed Desorption (TPD), X-ray Photoelectron Spectroscopy (XPS), Reflection Absorption Infra Red Spectroscopy (RAIRS), Low Energy Electron Diffraction (LEED), and Secondary Ion Mass Spectroscopy (SIMS). The Ultra High Vacuum (UHV) system, the cleaning procedure for the Rh(100) crystal, and the principle and physical background of each technique will be described. The computational techniques are Density Functional Theory (DFT) and kinetic Monte Carlo (kMC).
2.1 The Ultra High Vacuum (UHV) set-up
2.1.1 The UHV chamber
The experiments were carried out in a home-built stainless steel UHV system with a base pressure of 1.5 × 10-10 mbar to ensure a low level of surface contaminations and low interference between the gas phase and the electrons or ions used by the analytical tools. The system contains two UHV chambers stacked on top of each other separated by a gate valve, shown in Figure 2.1. The single crystal sample can move between chambers by the manipulator arm.
Figure 2.1 Pictures of the experimental set-up used. Top left and right give an
overview and a schematic overview of the whole system and bottom shows the front side of stage 1.
Stage one is equipped with facilities for Temperature Programmed Desorption (TPD), Low Energy Electron Diffraction (LEED), X-ray Photoelectron Spectroscopy (XPS) and a sputter gun (Leybold IQP 10/63) for sample cleaning. Stage 2 is equipped with facilities for TPD and Reflection Absorption Infrared Spectroscopy (RAIRS). Each chamber is pumped by a turbomolecular drag pump: a 300 l/s pump (Pfeiffer TMU 521 YP) for stage 1 and a 210 l/s pump (Pfeiffer TMU 260) for stage 2. Gases can be introduced via several leak valves into each stage. Please note: Secondary Ion Mass Spectroscopy (SIMS) experiments were done in a separate set-up, see [1] for details.
UHV Chamber Hemispherical analyzer Turbo pump X-ray gun Crystal Manipulator Mass spectrometer Leak valve Stage 1 Stage 2 Mass spectrometer D RAIRS cell S P Ar+ gun LEED screen Stage 1 Stage 2 RAIRS Manipulator Ar+gun Leak valve Leak valve
Figure 2.2 Pictures of the single crystal at 1400 K inside the UHV system (top) and at
atmospheric pressure and at room temperature (bottom).
2.1.2 The instruments
TPD and residual gas analysis are performed with a quadropole mass spectrometer (Balzers, Prisma QMA200) with a mass range m/e of 0 - 200 amu. XPS spectra were obtained with a VG100AX spectrometer (VG Microtech) consisting of a twin anode (Al/Mg) X-ray source and a 100 mm hemispherical analyzer. RAIRS spectra were taken with a Fourier-transform infrared spectrometer (Galaxy 4020, Mattson) with a wire grid polarizer placed in the beam. A mercury cadmium telluride (MCT) IR-detector was used with a spectral range of 800 - 4000 cm-1. LEED experiments were done with a reverse-view, two-grid mini LEED with retractable optics (OCI, Vacuum Microengineering BDL450IR). LEED patterns are collected and digitized using a CCD camera (Cohu). SIMS measurements were performed using a Balzers QMA400 mass spectrometer with a differentially pumped ion gun (SPECS PU-IQE 12/38). Attached to the analyzer are a RF generator (Balzers QMH400-5) and an ion counter pre-amplifier (Balzers CP400).
2.1.3 The single crystal
The rhodium single crystal (from Surface Preparation Laboratory, Zaandam, the Netherlands), with a (100) crystal cut within 0.5º of 1.2 mm thickness was polished by standard procedures. The crystal was mounted on the manipulator by two tantalum wires of 0.3 mm diameter, pressed into small grooves on the side of the crystal, as shown in Figure 2.2. This construction enables resistive heating of the sample up to
1400 K. The sample can be cooled continuously with liquid nitrogen enabling temperatures as low as 88 K. Temperatures were measured using a chromel-alumel thermocouple spot-welded to the back of the crystal. The sample holder is gold coated to prevent adsorption.
2.1.4 Cleaning procedure
The crystal surface was cleaned by cycles of argon ion sputtering and annealing in an oxygen atmosphere. Argon ion sputtering (4 kV, 6 μA/cm2) at 920 K for 30 min was used to remove impurities, such as boron, sulphur, phosphorus, and chlorine. Near-surface carbon was removed in 2 x 10-8 mbar O2 for 1 hour at temperatures ranging from 900 to 1100 K. Oxygen was removed by flashing to 1400 K. After flashing, a small amount of oxygen was adsorbed and the crystal was flashed to 800 K to remove carbon diffusing to the surface at temperatures above 900 K. Next, CO was dosed at 550 K and the crystal was flashed to 800 K to remove excess oxygen. Finally, CO was adsorbed at 150 K and the crystal was flashed to 600 K to simulate the first measurement of the day. Often in TPD experiments, the first measurement of the day does not fit the trend found in consecutive experiments. Deviations are observed in the desorption temperature and/or more often in the area under the desorption trace. Carbon monoxide (Linde gas, 99.997% pure), oxygen (Linde gas, 20% O2 in argon), and argon (Linde gas, 99.990% pure) were used without further purification.
2.2 Temperature Programmed Desorption (TPD)
TPD is a widely used technique in surface science studies to determine relative coverages of adsorbates and kinetic parameters of the desorption process, such as the preexponential factor and the activation energy of desorption. The technique is relatively fast with a simple principle behind it: first a gas is adsorbed onto a surface kept at a low temperature after which the surface is heated with a linear heating rate. A thermocouple and mass spectrometer are used to measure crystal temperature and the rate at which species desorb from the surface, as shown in Figure 2.3. If the adsorbate does not undergo a surface reaction, the area under the desorption features is directly related to the surface coverage and the desorption temperature is a measure for the binding energy of the adsorbate. If surface reactions do occur, the technique is also called Temperature Programmed Reaction Spectroscopy (TPRS). TPRS experiments even allow for determining reaction kinetics, but only in favourable cases where the formed reaction products desorb instantaneously.
Figure 2.3 The experimental set-up for a TPD experiment. The single crystal is
heated by two heating wires pressed into the grooves at the side of the crystal, while the temperature is measured with a thermocouple at the back of the crystal and controlled with a PID controller. Desorption of gases from the surface are monitored by a mass spectrometer. A desorption trace is included. (Adapted from Niemantsverdriet [2])
If the pumping rate of the UHV system is sufficiently high and thus no readsorption occurs, the mass spectrometer signal which is linearly proportional to the desorption rate is given by the Arrhenius or Polanyi Wigner equation3:
RT E n des n des des k e des dt d r ()/ ) ( (2.1) with:
rdes the desorption rate in [ML/s]
θ the fractional adsorbate coverage in [ML]
t the time in [s]
kdes the reaction rate constant in [1/s] for n = 1
n the reaction order
υdes(θ) the pre-exponential factor in [1/s]
Thermocouple Temperaturecontroller Heating wires Pumps Mass spectrometer UHV Desorption rate (a.u.) Temperature (K)
Edes(θ) the activation energy in [J/mol]
R the gas constant in [J/mol K]
T the temperature in [K]
In this equation, time can easily be substituted by the temperature by using: T = T0 +
βt, with T0 as the initial temperature which is often the adsorption temperature and β
as the heating rate.
Over the years, various methods have been developed from equation 2.1 to determine the kinetic parameters3-6. These methods work well in the limit of zero coverage, where kinetic parameters are independent of coverage.7 However, as soon as adsorbate/adsorbate interactions are present, kinetic parameters become coverage dependant and interpretation of the data becomes more complex, which can even lead to erroneous results.8,9
Figure 2.3 shows a simple first order desorption trace without any adsorbate interactions. Normally, these lateral interactions cannot be excluded and they will influence the desorption energy. Assuming that lateral interactions are pair-wise additive, the activation energy becomes:
j i i i des tot lat des des E E E n E 0
int 0 ) ( (2.2) with: 0 desE the activation energy for desorption in the zero-coverage limit in
[J/mol]
tot lat
E int the total lateral interaction energy [J/mol]
ni the number of neighbouring species ‘i’
j i
the repulsive pairwise additive interaction energy between species ‘i' and ‘j’ [J/mol]
Lateral interactions are expressed in TPD experiments as shifts in the desorption temperature. It is even possible to recognize the presence of different surface structures when multiple peaks appear in the desorption traces. Changing the surface structure leads to a difference in adsorbate environment, resulting in other lateral interactions to start playing a role. We make use of this fact in our dynamic Monte Carlo study in Chapter 4 and to determine the interaction between nitrogen and CO in Chapter 6.
2.3 X-ray Photoelectron Spectroscopy (XPS)
2.3.1 Principle
XPS is a frequently applied technique in catalysis to determine the elemental composition and oxidation state of the elements present on and near the surface of a sample.10-12 XPS is based on the photoelectric effect: atoms absorbing X-ray photons emit core electrons. The emitted free core electrons are then called photoelectrons. The amount of photoelectrons with a specific kinetic energy is measured, as shown in Figure 2.4.
Figure 2.4 The principle of XPS (left). First, an incident X-ray photon is absorbed by the atom, while a photoelectron is emitted. The kinetic energy of the electron is measured. An XPS spectrum of the Rh(100) crystal is shown on the right.
Instead of the kinetic energy, the binding energy in XPS spectra is usually given: the energy between the core level and the Fermi level of the sample, which is independent of the X-ray source used. The binding energy is given by the following equation: k b h E E (2.3) with:
Eb the binding energy of the photoelectron with respect to the Fermi level
of the sample (eV)
h Planck’s constant (eV s)
ν the frequency of X-ray radiation (1/s)
Ek the kinetic energy of the emitted electron (eV)
φ the work function of the spectrometer (eV)
e -X-ray hȞ Eb ij Ek Core level Fermi level Vacuum level 0 100 200 300 400 500 600 700 Rh 4p Rh 4s Rh 3d 3/ 2 Rh 3d 5/ 2 Rh 3p 1/2 Rh 3s
Photoelectron intensity (a.u.)
Binding energy (eV) Rh(100) single crystal Rh 3p 3/ 2
+
The binding energy of the electron is characteristic for the atom from which the electron originates. Figure 2.4 shows a characteristic spectrum of pure rhodium, which can be used as a ‘fingerprint’ to verify if a sample contains rhodium. The binding energy of the electrons depends on the quantum level of the atomic orbital from which the electron originates. The electrons from the lowest quantum levels (closest to the nucleus) have the highest binding energy and the valence electrons are close to zero.
2.3.1 Quantification
Besides the ability to recognize elements, XPS is also used to determine elemental compositions. The general expression for the intensity of an XPS peak is:
dz e z n E E S F I z Ekz k k x
0 cos ) ( / , ) ( ) ( ) ( (2.4) with:I the area under the XPS peak
Fx the X-ray flux on the sample
S(Ek) the spectrometer sensitivity for detecting en electron at kinetic energy
Ek
σ(Ek) the elemental cross-section for photoemission
n(z) the concentration [mol/m3]
z the depth below the surface
λ(Ek,z) the mean free path of the photoelectron at kinetic energy Ek through the
material present at depth z
θ the angle between the surface normal and the angle at which the photoelectrons are collected
XPS is used in this study to determine adsorbate coverages. To determine relative adsorbate coverages, the intensity of an element of the adsorbate, is normalized with respect to in our case the rhodium intensity. At a constant sample position, the intensity can be simplified into approximately: I/IRh3d5/2x(Ek), with x being an apparatus specific sensitivity constant and θ now the adsorbate coverage. The elemental cross-sections for photoemission of all elements have been tabulated by for example Scofield13. Last, a reliable reference, like a saturation coverage, is needed to determine x and to obtain the coverage at other loadings. We have followed this procedure to determine CO coverages in Chapter 3 and the nitrogen coverage in Chapter 6.
2.4 Reflection Absorption Infra Red Spectroscopy (RAIRS)
2.4.1 Principle
Widely used techniques to study adsorbates on single crystal surfaces are vibrational spectroscopies like Electron Energy Loss Spectroscopy (EELS) and RAIRS. While EELS makes use of a monochromatic beam of low energy electrons, RAIRS uses an infrared beam to excite vibrations of molecules and atoms bonded to the surface during a single reflection near grazing angle with the surface, as shown in Figure 2.5.
Figure 2.5 The principle of RAIRS: IR-light is reflected in a single reflection near
grazing angle from the crystal surface. The electric field of the light is composed of a p- and s-polarized part. Due to a strict dipole selection rule, only p-polarized light can interact with an adsorbed molecule or atom, exciting vibrations with a component perpendicular to the surface.
Due to the dipole selection rule, only vibrations with a component perpendicular to the surface are excited by p-polarized light. Routinely, a spectral range of 800 - 4000 cm-1 is accessed: the range at which most vibrations of organic molecules are detected, but outside the range of metal-atom (e.g. H, N, and O atoms on a rhodium surface) vibrations lying below 800 cm-1, although special detectors are available which are able to detect vibrations as low as 50 cm-1. The main advantages of RAIRS are the high sensitivity, the high resolution at which spectra can be acquired, typically about 10 times better than with EELS, and the ability to perform measurements under atmospheric conditions.
2.4.2 Frequency, intensity, and line shape of infrared absorption bands
During a RAIRS experiment, absorption of IR light or the change in reflectivity is measured as a function of wavenumber. The absorbance at a particular wavenumber is given as follows: 0 0 p p p I I I R R (2.5) 5º Ep Es
with:
ΔR/R the absorbance in [%] 0
p
I
the intensity of p-polarized light of a stored background spectrump
I
the intensity of p-polarized light of the sample spectrumHence, two spectra are needed for a RAIRS experiment: a background spectrum of an empty surface and a sample spectrum. The intensity of an absorption band is proportional to the dynamic dipole moment of the adsorbed molecule from which the band arises and the number of molecules (coverage), although coverage and intensity are in general not linearly related.* At comparable coverages, CO and NO absorb typically a few percent of the infrared light at specific frequencies, whereas hydrocarbons absorb only a few tenth of a percent of IR light. Example RAIRS spectra of CO on Rh(100) are shown in Figure 2.6.
Figure 2.6 Typical RAIRS spectra of CO adsorbed on Rh(100). Absorption by CO on
top and bridge sites is observed around 2000 – 2100 and 1900 – 2000 cm-1,
respectively, while absorption by CO in the gas phase is observed around 2143 cm-1.
The vibrational frequency of the internal stretch mode of adsorbed CO is lower than gas phase CO (about 2143 cm-1). The difference in stretching frequency results from CO bonded to different sites: between 2000 and 2100 cm-1 for CO on top sites, between 1900 and 2000 cm-1 for CO on bridge sites and below 1900 cm-1 for three- and four-fold bonded positions.14
* Adsorbates on surfaces are in general within the low concentration Lambert Beer regime, where adsorption scales with number of molecules. Deviations from linearity are due to interactions between adsorbate dipoles and between adsorbate and surface image dipoles. This will be discussed further in chapter 3. 1800 1900 2000 2100 2200
A
bsorbance
'R/
R (%)
W avenum ber (cm
-1)
2.5%
C O on R h(100)
(2x)
Gas phase CO
The shift of the vibrational frequency from the gas phase value can be attributed to several factors.15 A mechanical shift due to adsorption of CO to a rigid rhodium surface is estimated from a simple model of masses and springs to increase the frequency with about 50 cm-1. The interaction of the adsorbed dipole and its own image dipole results in a downward shift of about 30 to 50 cm-1. An image dipole in the metal substrate is formed from screening of the dipole field by the conducting electrons. The strong frequency shift upon CO adsorption and the site dependency on the frequency cannot be explained by these two factors alone. This is largely the result of a chemical shift: as the molecule binds to the surface the electronic structure within the molecule changes.
Figure 2.7 The energy diagram for adsorption of CO on Rh(100). Upon adsorption
the HOMO and LUMO of CO interact with the d-band of rhodium forming new bonding and antibonding orbitals. Adapted from [16].
When CO adsorbs on Rh(100) or any other transition metal, a chemical bond is made by the interaction of the molecular orbitals of CO and the d-band of Rh(100).17 The CO 5σ (HOMO) interacts with the
2 z
d orbital of the metal creating
two new molecular orbital bands: a bonding band lying low in energy and an antibonding band lying high in energy, as shown in Figure 2.7. The interaction of the LUMO (CO 2π*) with the dxy and dxz orbitals creates another set of orbital bands.
Energy
Density of
states
E
vacE
Fs + p
d
2 z dxz xy d d ,Rh(100)
Rh(100)-CO
CO
2ʌ
*5ı
HOMO
LUMO
bonding bonding antibonding antibondingFilling of the bonding 2π*/d band contributes to the bonding of the CO molecule to the surface, but weakens the internal CO bond. This consequently lowers the stretch frequency of CO on a top position. Due to a better overlap of the 2π* orbital with the
d orbitals at higher bond order, this effect is stronger for bridge and hollow positions
resulting in larger frequency shifts.
Another trend is the increase of the vibrational frequency with the CO coverage, due to intermolecular interactions. Mainly two types of interactions play a role. First, chemical interactions: due to repulsive interactions between adsorbates the CO/metal-bond weakens, reducing back-donation into the 2π* CO-orbitals and increasing the frequency, as in the model just described. Second, interactions by vibrational coupling: this is a long range effect with similar dipoles coupling through space or via metal electrons (think of image dipoles). Dipole-dipole coupling results in a higher frequency, an increase in intensity, and narrower bands due to intensity stealing from closely related vibrations.
The line shape of an absorption band can also be affected by several processes. First, the line width can be homogenously broadened by damping of the vibrations through coupling with phonons or by electron hole/pair creation. This coupling is expected to be more pronounced at adsorption sites with stronger back-donation, present at lower coverages and at higher bond order. Second, inhomogeneous broadening of adsorption bands is caused by inhomogeneities of the intermolecular distances due to an irregular adsorbate distribution. Very sharp adsorption bands are expected for well ordered adlayers, while splitted broader peaks are expected for island formation. For more background information on RAIRS we refer to several reviews [15,18,19].
2.5 Low Energy Electron Diffraction (LEED)
LEED is a technique used to determine the structure of ordered surfaces and adsorbates.20,21 A monochromatic beam of low energy electrons (50 - 200 eV) is directed towards a single crystal surface. The electrons scatter back elastically and due to the ordered nature of the crystal surface, the electrons, which can be considered as waves, show interference patterns. The interference patterns are made visible with a fluorescent screen, as shown in Figure 2.8.
Figure 2.8 A schematic picture of a LEED set-up. A beam of monochromatic
electrons scatters elastically from the surface in all directions. Back scattered electrons show constructive and destructive interference due to the ordered structure of the surface and its adsorbates, leading to diffraction patterns made visible with a fluorescent screen.
The interference occurs according to Bragg’s law:
k eE m a nh a n sin 2 (2.6) with:
θ the angle between surface normal and scattered electron in degrees
n the diffraction order
λ the wavelength of the electron [m]
a the distance between scatterers, e.g. the lattice constant for a clean surface in [m]
h Planck’s constant [J s]
me the mass of a electron [kg]
Ek the kinetic energy of the electron [J]
The relation between the interatomic distances and diffraction is inverse, resulting in a diffraction pattern showing a ‘reciprocal lattice’: smaller distances
Surface structure Single crystal Primary beam LEED pattern + Diffraction beams Electron gun Fluorescent screen
between LEED spots correspond to large interatomic distances. Figure 2.8 shows a
p(1×1) surface structure with an c(2×2) adsorbate structure on top. The corresponding
LEED picture is a coincidence pattern of the two structures. The distances between the atoms in the c(2×2) structure are √2 larger than in the p(1×1) structure, while the distance between LEED spots become √2 smaller.
2.6 Secondary Ion Mass Spectroscopy (SIMS)
SIMS is able to monitor in real-time surface concentrations of elements and molecules.22,23 Conceptually, the principle of SIMS seems rather simple: the surface is exposed to a beam of high energy primary ions (often 0.5 – 5 keV Ar+ ions). The ions penetrate into the substrate creating a number of collision cascades. Part of the energy flows back to the surface and stimulates the ejection of atoms and multi-atomic clusters, with the majority being neutrals but also a small fraction as positive and negative ions, see Figure 2.9. These secondary ions can directly be detected with a mass spectrometer. In static mode (with a low primary ion flux), it is a powerful technique in single crystal studies in combination with TPRS to follow surface reactions: as a function of temperature the formation and decomposition of reaction intermediates are detected with SIMS while TPRS follows the evolution of gas phase products, making it possible to unravel the elementary reaction steps. In secondary neutral mass spectrometry, the neutrals are post-ionized before detection, however at the cost of sensitivity, which is orders of magnitude lower than that of SIMS.
Figure 2.9 The principle of SIMS: high energy argon ions penetrate into the surface
and cause a collision cascade. Part of the energy flows back to the surface and stimulates the emission of ions and neutrals. The ions are detected directly with a mass spectrometer.
Although the technique seems quite destructive, a low primary ion flux is used with an adsorbate removal rate of at least 1 ML/hour. The duration of a SIMS
Ar+beam
2+ +
- +
Secondary ions
and neutrals Mass
experiment typically lasts only several minutes. Hence, the damage is on the order of a few percent of a monolayer only.
The secondary ion yield is given by the following complex relation:
Y I
Is p (2.7)
with:
Is the measured flux of positive or negative secondary ion in [ions/s]
Ip the flux of primary ions in [ions/s]
θ the coverage of element X in the surface layer in [ML]
Y the sputter yield of element X from the matrix in [ions emitted/incident
ions]
α the effective ionization probability of atom X ejected from the matrix η the transmission of the mass spectrometer for ion X
The fact that the ionization process is very complex and that neutralization of the formed ion are not well understood, makes it difficult to determine the sputter yield and the ionization probability. For example, the ionization probability of atoms or molecular clusters depends on its surroundings. This phenomenon is called the matrix effect. To obtain semi-quantitative information from SIMS experiments, the ratio of the metal-adsorbate secondary ion intensity and the metal secondary ion intensity is taken as a measure of the coverage:
M X M X I I , (2.8)
The influence from the matrix effect acting on both groups of ions is cancelled out in this way.24-26 This approach is used to determine the CO coverage in Chapter 3.
2.7 Density Functional Theory (DFT)
Computational chemistry is an expanding area of chemistry over recent years thanks to the increase in calculation power of computers. Nowadays, ab-initio calculations are able to provide good estimates for numeral properties of chemical systems. In surface chemistry, properties such as adsorption energies, binding geometry, vibrational frequencies, and reaction pathways can be predicted by electronic structure calculations.
First principle quantum mechanical methods compute the total energy of a system by solving ab-initio the Schrödinger equation for all particles in the system. This can be done for a few very small systems, but approximations have to be made
for larger and more complex systems. Since the nuclei are much heavier than the electrons, the nuclei are taken stationary while the electrons move relative to them as a first approximation, which is called the Born-Oppenheimer approximation.27 Therefore, it is allowed to solve the Schrödinger equation for the wavefunctions of the electrons alone.
One of the most widely used ab-initio approaches is Density Functional Theory (DFT). It is based on the Hohenberg and Kohn theorem28 in which the total energy of the system’s wavefunctions in the ground state is a unique functional of the electron density. This energy functional is minimized in the Density Functional Theory (DFT) approach, by following the variational principle29. The Kohn-Sham equations30 are solved to obtain the total energy.
We have used the Vienna ab-initio simulation package (VASP)31,32, which performs an iterative solution of the Kohn-Sham equations in a plane-wave basis set. The many-electron wavefunction for the system is written as a product of one-electron wavefunctions. The total energy consists of a kinetic, an electrostatic, and an exchange-correlation term. The exchange-correlation energy has been calculated within the generalized gradient approximation (GGA) using the revised form of the Perdew, Burke, and Ernzerhof exchange-correlation functional33 proposed by Hammer et al.34 The electron-ion interactions for the elements are described by the projector augmented wave (PAW) method developed by Blöchl35. The reciprocal space has been sampled with a k-point grid automatically generated using the Monkhorst-Pack method36. Fractional occupancies were calculated using a first-order Methfessel-Paxton smearing function.37
2.8 Kinetic Monte Carlo (KMC)
Kinetic (or dynamic) Monte Carlo simulation is a powerful technique to model reaction kinetics and ordered structures on metal surfaces. In contrast to the mean-field approach (MFA), kMC uses the local surroundings of a process taking place, which makes it possible to include local surface diffusion and lateral interactions. KMC simulations were performed with the program CARLOS.38-40
The surface is modelled as a grid of active sites. Each site is labelled as vacant or with an adsorbate occupying it. A group of sites is called a configuration, which can change over time:
t1 t2
The master equation describes how the rate at which Pα, the probability of finding a
particular ensemble in configuration α at time t, changes over time:
k P k P dt dP (2.9)This equation is equivalent to a simple rate equation with k as the rate constant of changing the configuration α into β. The first term after the equal sign leads to the
formation of configuration α, while the second term describes the process that
decreases the amount of configuration α. In general, there are three types of processes
involved in catalysis: surface reaction, adsorption/desorption and diffusion. Except for adsorption, these processes have rate constants in the form of an Arrhenius equation:
T k E B e k / (2.10)
Monte Carlo simulations are based on solving the master equation (eq. 2.9), by carrying out these processes, one at a time. The rates determine the probability (how or with what route) and how fast (at what moment in time) configurations are formed or broken down.
Figure 2.10 Monte Carlo description of a model for CO diffusion from a top onto bridge site. The sites surrounding occupied top and bridge sites are excluded for additional CO molecules. The initial state of diffusion with CO on top and the indicated configuration of empty sites required for diffusion are indicated. Surrounding this configuration are close-range lateral interactions when adsorbates are present.
Figure 2.10 shows how one of the processes, in this case CO diffusion, is described by kMC on a (100) surface in the model used in Chapter 4. The initial configuration before diffusion is given in grey with CO occupying a top site. Due to the excluding model used as shown in the top part, CO is surrounded by empty bridge and hollow sites. In this example, CO can diffuse to the bridge site Br1, but only when the sites Top2 and Br2 are empty. Beside the interactions implied by the excluding model, finite lateral interactions are included. When CO is present on the bridge site above the hollow site in the top left corner, the interaction between the two CO molecules is ωtb. This is the interaction energy between top CO before diffusion and
-Ȧtb Ȧtb- Ȧbb Ȧtt- Ȧtb Ȧtb Br2 Hol Br Hol Ȧtb -Ȧbb Top2 Br1 COtop Br Ȧtt Br2 Hol Br Hol Ȧtb -Ȧtb Ȧtb- Ȧbb Ȧtt- Ȧtb Ȧtb × × × × × × × × × × × × × × × × × × × ×
the surrounding bridge CO. No interaction between bridge CO after diffusion and the surrounding CO molecule is included, due to the increased CO-CO distance. When a CO molecule is present on the top site next to this bridge site, two interactions are present ωtt- ωtb: the repulsive interaction between the two CO molecules before
diffusion which is positive and after diffusion which is negative.
Lateral interactions are taken as pair-wise additive similarly as in equation 2.2. More precise, the Brønsted-Evans-Polanyi relationship is used:
[0,1]
;
0
ij
i i act actE
n
E
(2.11)here the interactions ij are as described in Figure 2.10. In this work, the
Brønsted-Evans-Polanyi coefficient41,42 for desorption is assumed 1 (converting this equation to equation 2.2) and for diffusion ½. The effect of these coefficients becomes evident in Figure 2.11. The energy profiles of CO desorption from the most stable top site with and without lateral interactions are shown on the left. The interactions acting on the CO molecule on the surface destabilize it, whereas they do not affect gas-phase CO. The activation energy of desorption decreases with i j
i i
n
by the interactions. Diffusion from a top site onto a neighbouring site with the same stability is shown on the right. For simplicity, we have only included lateral interactions in the initial state. The lateral interactions facilitate diffusion from a destabilized site towards a site with no interactions and make the reverse diffusion process energetically less favourableby i j i i n
21 . This kMC approach is used in Chapter 4 to study the effect of lateral
interactions between CO molecules on CO desorption and the formation of ordered structures and in Chapter 6 to study ordering of the nitrogen adlayer.
Figure 2.11 The energy profiles for desorption and diffusion and how they are affected by repulsive lateral interactions. The black curves show the path without interaction and the grey curves with interactions.
COgas COads 0 act E j i i i n →
¦
ω ½ act E Desorption 0 act E j i i i n →¦
ω Diffusion:Į
= 0.5 act E j i i i n →¦
ωReferences
1. D.L.S. Nieskens, PhD thesis: The role of adsorbate interactions in catalysis by
metals, Eindhoven, 2006.
2. J.W. Niemantsverdriet, Spectroscopy in Catalysis: An Introduction; Third,
Completely Revised and Enlarged Edition, Wiley-VCH, Weinheim, 2007.
3. D.A. King, Surf. Sci., 1975, 47, 384-402.
4. C.M. Chan, R. Aris and W.H. Weinberg, Appl. Surf. Sci., 1978, 1, 360-376. 5. P.A. Redhead, Vacuum, 1962, 12, 203-211.
6. E. Habenschaden and J. Küppers, Surf. Sci., 1984, 138, L147-L150. 7. A.M. de Jong and J.W. Niemantsverdriet, Surf. Sci., 1990, 233, 355-365.
8. J.B. Miller, H.R. Siddiqui, S.M. Gates, J.N. Russell, J.T. Yates, J.C. Tully and M.J. Cardillo, J. Chem. Phys., 1987, 87, 6725-6732.
9. D.L.S. Nieskens, A.P. van Bavel and J.W. Niemantsverdriet, Surf. Sci., 2003, 546, 159-169.
10. S. Hüfner, Photoelectron Spectroscopy - Principles and Applications, Springer, Berlin, 1996.
11. G.C. Smith, Surface Analysis by Electron Spectroscopy: Measurement and
Interpretation, Plenum, New York, 1994.
12. D. Briggs and M.P. Seah, Practical Surface Analysis: Auger and X-ray Photoelectron
Spetroscopy, Wiley, New York, 1996.
13. J.H. Scofield, J. Electron Spectrosc. Relat. Phenom., 1976, 8, 129-137.
14. N. Sheppard and T.T. Nguyen, Adv. Infrared Raman Spectrosc., 1978, 5, 67-148. 15. F.M. Hoffmann, Surf. Sci. Rep., 1983, 3, 107-192.
16. I. Chorkendorff and J.W. Niemantsverdriet, Concepts of Modern Catalysis and
Kinetics, Wiley-VCH, Weinheim, 2003.
17. R.A. van Santen, Theoretical Heterogeneous Catalysis, World Scientific, Singapore, 1991.
18. Y.J. Chabal, Surf. Sci. Rep., 1988, 8, 211-357.
19. P. Hollins and J. Pritchard, Progr. Surf. Sci., 1985, 19, 275-350.
20. G. Ertl and J. Küppers, Low Energy Electrons and Surface Chemistry, 2 edn., VCH, Weinheim, 1985.
21. M.A. van Hove, W.H. Weinberg and C.M. Chan, Springer Series in Surface Sciences,
Vol. 6: Low-Energy Electron Diffraction: Experiment, Theory, and Surface Structure Determination, 1986.
22. J.C. Vickerman, A. Brown and N.M. Reed, Secondary Ion Mass Spectrometry:
Principles and Applications, Clarendon, Oxford, 1989.
23. A. Benninghoven, F.G. Rüdenauer and H.W. Werner, Secondary Ion Mass
Spectrometry, Basic Concepts, Instrumental Aspects, Applications and Trends, Wiley,
New York, 1987.
24. X.Y. Zhu and J.M. White, J. Phys. Chem. B, 1988, 92, 3970-3974.
25. A. Benninghoven, P. Beckmann, D. Greifendorf and M. Schemmer, Appl. Surf. Sci., 1980, 6, 288-296.
26. A. Benninghoven, K.H. Müller, M. Schemmer and P. Beckmann, Appl. Phys., 1978, 16, 367-373.
28. P. Hohenberg and W. Kohn, Phys. Rev., 1964, 136, B864. 29. A. Groß, Surf. Sci., 2002, 500, 347-367.
30. W. Kohn and L.J. Sham, Phys. Rev., 1965, 137, 1697-1705.
31. G. Kresse and J. Furthmüller, Phys. Rev. B: Condens. Matter, 1996, 54, 11169-11186. 32. G. Kresse and J. Hafner, Phys. Rev. B: Condens. Matter, 1993, 47, 558-561.
33. J.P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865-3868. 34. B. Hammer, L.B. Hansen and J.K. Nørskov, Phys. Rev. B: Condens. Matter Mater.
Phys., 1999, 59, 7413-7421.
35. P. E. Blöchl, Phys. Rev. B: Condens. Matter, 1994, 50, 17953-17979.
36. H.J. Monkhorst and J.D. Pack, Phys. Rev. B: Condens. Matter Mater. Phys., 1976, 13, 5188-5192.
37. M. Methfessel and A.T. Paxton, Phys. Rev. B: Condens. Matter, 1989, 40, 3616-3621.
38. A.P.J. Jansen, Comput. Phys. Commun., 1995, 86, 1-12. 39. A.P.J. Jansen, http://arXiv.org/cond-mat/, 2008, 0303028.
40. J.J. Lukkien, J.P.L. Segers, P.A.J. Hilbers, R.J. Gelten and A.P.J. Jansen, Phys. Rev.
E, 1998, 58, 2598-2610.
41. J.N. Brønsted, Chem. Rev., 1928, 5, 231-338.
Adsorption/desorption studies of CO on a
rhodium(100) surface under UHV conditions: a
comparative study using XPS, RAIRS, and SSIMS
*In order to explore to what extent equilibrium adsorption can be studied with surface science methods such as RAIRS, SSIMS and XPS, adsorption of CO on a Rh(100) single crystal surface has been studied in this chapter. A structural model of CO on Rh(100) at different coverages is presented. Under equilibrium adsorption conditions, we were able to determine the CO coverage and adsorption isotherms. From the isotherms, an isosteric heat of adsorption of about 160 kJ/mol was obtained, compared to an activation energy of desorption of 132 kJ/mol from TPD experiments.
* The contents of this chapter have been published: M.M.M. Jansen, F.J.E. Scheijen, J. Ashley, B.E. Nieuwenhuys, and J.W. Niemantsverdriet, Catal. Today, 2009, DOI: 10.1016/j.cattod.2009.11.014
3.1 Introduction
In catalysis, one of the most important parameters predicting surface reactivity is the interaction energy between a molecule and the surface.1-3 For validation of the accuracy of theoretical models, good quantitative data on adsorption energies is of significance. Several techniques are used in surface science studies to determine the adsorption energy. The most direct method is Single Crystal Adsorption Calorimetry (SCAC)4,5, where the heat released during adsorption of a gas on an ultra-thin metal single crystal foil results in an increase of the foil temperature. A serious drawback is that the foils contain a high density of surface defects, which in turn affects the heat of adsorption. More widely employed is Temperature Programmed Desorption (TPD). During TPD experiments, a metal single crystal with adsorbates is heated in a controlled fashion, often linearly with time, during which the desorption rate of the adsorbates is monitored with a mass spectrometer. Several methods have been derived from the Arrhenius equation to obtain an estimate of the kinetic parameters from the desorption traces, see [6,7] for a comparison between methods. A third technique is Clausius-Clapeyron analysis of equilibrium adsorption isosteres. In order to obtain the adsorption energy from equilibrium isosteres, the adsorbate coverage is measured in a gas atmosphere at different temperatures and pressures. In this case, it is not straightforward to determine the coverage, because the gas environment can affect the measurements.
In this chapter, equilibrium isosteres are obtained with the surface science methods X-ray Photoelectron Spectroscopy (XPS), which is quantitative, and the semi-quantitative techniques: Reflection Absorption Infra Red Spectroscopy (RAIRS) and Static Secondary Ion Mass Spectroscopy (SSIMS). A comparison is made among the different techniques to determine how accurate coverages and adsorption energies are obtained with each technique. CO is used as a probe molecule to study the adsorption and desorption behaviour on a rhodium(100) single crystal surface. Although the adsorption of CO has been studied extensively8-18, this is the first surface science study using as many as five different characterization techniques. First, CO is irreversible adsorbed at low temperatures and studied with Low Energy Electron Diffraction (LEED) and RAIRS. A structural model of CO on Rh(100) at different coverages is discussed, which is in excellent agreement with literature. The coverages belonging to the structural model are used to convert the areas under TPD traces into coverages, which in turn are correlated to the intensities observed in RAIRS, SSIMS and XPS experiments. Lastly, the adsorption/desorption equilibrium of CO is probed using XPS, RAIRS and SSIMS to determine the adsorption energy by using the Clausius-Clapeyron equation. The resulting isosteric heat of adsorption is compared to the activation energy of desorption from TPD experiments.
We will now give a short review of previous results. LEED studies have shown that on Rh(100), CO orders into three distinct ordered structures at different coverages. Several studies8-17 reported that CO orders in a c(2×2) structure at θ
