The adsorption of gold vapor onto MgO(100) films grown on Mo(100) was studied at 300 and 100 K using single crystal adsorption calorimetry (SCAC). The Au particle morphology was investigated using He+ low-energy ion scattering spectroscopy (LEIS) and X-ray photoelectron spectroscopy (XPS). The LEIS data combined with particle shape measurements from the literature (Benedetti, S.; Myrach, P.; di Bona, A.; Valeri, S.; Nilius, N.; Freund, H.-J. Phys. Rev. B 2011, 83 (12), 125423) reveal that, at both 300 and 100 K, Au grows as 2D islands with bilayer thickness (∼0.41 nm) up to a diameter of ∼7 nm. At higher coverage, the islands thicken with little increase in diameter. The island densities are 3.0 × 1011 and 5.4 × 1011 per cm2 at 300 and 100 K, respectively. The initial sticking probability of Au is 0.90 at 300 K and 0.95 at 100 K. The surface residence time of the Au atoms that do not stick is <10 ms, implying that gold monomers bind to MgO(100) weakly (<69 kJ/mol). The adsorption energies indicate that Au particles of the same size bind more strongly to MgO(100) when grown at 300 K than at 100 K, which we attribute to Au binding to step edges or defects at 300 K, but at perfect MgO(100) terraces at 100 K (because Au diffusion is too slow to find defects). The adsorption energy of Au onto ∼30-atom Au clusters is 285 kJ/mol at 300 K, ∼68 kJ/mol higher than at 100 K, attributed to the difference between particles on defects versus terraces. Similarly, the adhesion energy of Au nanoparticles to MgO(100) extracted from the adsorption energies at 300 K is much higher (1.8 J/m2 for ∼7 nm particles at defects) than at 100 K (0.3 J/m2 for ∼7 nm particles at terraces). This 100 K adhesion energy is close to that estimated from electron-microscopy shape measurements of Au particles at terraces on MgO(100) (0.45–0.67 J/m2). The heat of Au adsorption and Au chemical potential change by >100 kJ/mol as gold’s 2D island size increases from 0.7 to 7 nm diameter, implying dramatic changes in catalytic activity and sintering rates with 2D diameter. This is the first experimental measurement of any metal adsorption energy on any oxide as a function of island diameter when making 2D islands, as well as the first direct comparison of any adhesion energy found from calorimetric adsorption energies to that from particle shape analysis.Keywords: adhesion energy; gold; heat of adsorption; metal adsorption; metal/support interaction; nanoparticle catalysis; oxide support;

The heat of adsorption and sticking probability of methyl iodide were measured on Ni(111) at 100 and 160 K using single-crystal adsorption calorimetry (SCAC). At 100 K, methyl iodide adsorbs molecularly with a heat of 102 kJ/mol on terrace sites in the low-coverage limit, giving a standard enthalpy of formation (ΔHf°) of CH3I(ad) of −87 kJ/mol. A heat of 122 kJ/mol is also measured on defect sites, probably step edges. Calorimetry of the dissociative adsorption of methyl iodide on Ni(111) at 160 K yielded an integral heat of adsorption of −270 kJ/mol at 0.04 ML, providing the energetics of adsorbed methyl, with ΔHf°[CH3(ad)] = −71 kJ/mol and a CH3–Ni(111) bond enthalpy of 218 kJ/mol. This is 22 kJ/mol stronger than the reported value for H3C–Pt(111) bonds, explaining the greater activity of Ni catalysts for hydrogenolysis in comparison to Pt. The measured energetics for methyl were compared to density functional theory (DFT) calculations from previous literature, showing that these methods systematically underestimate the bond energy of methyl to Ni(111).Keywords: adsorbed methyl; heat of adsorption; hydrogenolysis; methane activation; nickel; partial oxidation; platinum; steam re-forming;

Metal nanoparticles supported on oxide surfaces form the basis for many industrial catalysts and promise to play an ever-increasing role in future energy and environmental technologies. The chemical potential (μ) of the metal atoms in these particles depends strongly on particle size and support and is an important factor that determines their catalytic properties, including their binding strengths to adsorbed reaction intermediates and their long-term stability against sintering. We present here a method for estimating this chemical potential as a function of particle size for different metal/oxide combinations. We show that this chemical potential for late transition metals is well approximated for a particle of diameter D by μ(D) = [(3γm – Eadh)(1 + Do/D)](2 Vm/D), where γm is the surface energy of the bulk metal, Eadh is the adhesion energy at the bulk metal/oxide interface, and Do is ∼1.5 nm, and Vm is the molar volume of the bulk metal. We further show that Eadh increases with (1) increasing heat of formation of the most stable oxide of the metal from metal gas atoms plus O2(gas) per mole of metal atoms, (2) decreasing enthalpy of reduction of the oxide to its next lower oxidation state plus O2(gas), per mole of oxygen atoms, and (3) increasing density of surface oxygen atoms on the oxide surface. The linear scaling of Eadh with these properties allows estimations of Eadh for a variety of metal/oxide combinations. Using this Eadh estimate in the above equation with known values for γm allows estimates of metal chemical potential versus particle size for late transition metals on various oxide supports. This will improve our ability to understand structure–property relations in catalysis and design better catalysts.Keywords: adhesion energy; catalyst design; catalyst support; coarsening; nanoparticles; particle size effects; sintering; support effects;

•First measurement of adsorption energy for any metal-on-metal combination that forms a bulk alloy.•Heat of adsorption and adhesion energy for Cu on Pt(111).•Cu monolayer 12 kJ/mol more stable than Cu(solid), vs 83 in underpotential deposition.•Lattice strain causes decrease in heat with coverage.The adsorption energies of submonolayer amounts of one metal on the surface of another metal have been measured for decades by temperature programmed desorption. However, that method fails for metals that alloy. We report here the first measurement of the adsorption energy for any such metal-on-metal combination that forms a bulk alloy. The adsorption and interfacial energetics of vapor deposited Cu onto Pt(111) at 300 K has been studied using single crystal adsorption calorimetry (SCAC) and X-ray photoelectron spectroscopy (XPS). The Cu grows as 2D pseudomorphic islands in the first layer and its heat of adsorption decreased linearly from 358 to 339 kJ/mol. This is attributed to increasing lattice strain with island size, associated with the small lattice mismatch (8%). It adsorbs ~2 kJ/mol more weakly in the 2nd layer than above 3 ML, where it reaches the bulk heat of sublimation of Cu(solid), 337 kJ/mol. The adhesion energy of multilayer Cu onto Pt(111) is 3.76 J/m2. The extra stability of the first Cu monolayer compared to bulk Cu measured here is ~12 kJ/mol, compared to a difference of ~83 kJ/mol for underpotential deposition of Cu on a Pt(111) electrode, with the difference attributed to stronger bonding of Cu to the solvent and double layer compared to Pt.

Nanoparticles on surfaces are ubiquitous in nanotechnologies, especially in catalysis, where metal nanoparticles anchored to oxide supports are widely used to produce and use fuels and chemicals, and in pollution abatement. We show that for hemispherical metal particles of the same diameter, D, the chemical potentials of the metal atoms in the particles (μM) differ between two supports by approximately −2(Eadh,A – Eadh,B)Vm/D, where Ead,i is the adhesion energy between the metal and support i, and Vm is the molar volume of the bulk metal. This is consistent with calorimetric measurements of metal vapor adsorption energies onto clean oxide surfaces where the metal grows as 3D particles, which proved that μM increases with decreasing particle size below 6 nm and, for a given size, decreases with Eadh. Since catalytic activity and sintering rates correlate with metal chemical potential, it is thus crucial to understand what properties of catalyst materials control metal/oxide adhesion energies. Trends in how Eadh varies with the metal and the support oxide are presented. For a given oxide, Eadh increases linearly from metal to metal with increasing heat of formation of the most stable oxide of the metal (per mole metal), or metal oxophilicity, suggesting that metal–oxygen bonds dominate interfacial bonding. For the two different stoichiometric oxide surfaces that have been studied on multiple metals (MgO(100) and CeO2(111), the slopes of these lines are the same, but their offset is large (∼2 J/m2). Adhesion energies increase as MgO(100) ≈ TiO2(110) < α-Al2O3(0001) < CeO2(111) ≈ Fe3O4(111).Keywords: adhesion energy; catalyst support; gold on MgO(100); metal catalyst; nanoparticles; oxide surface; sintering;

•Heat of formation of bidentate formate adsorbed on Ni(111) is −449 kJ/mol.•First measurement of the energy of any oxygenate on Ni through metal-oxygen bonds.•Important benchmark for validating calculations of oxygenates on Ni catalysts.•The measured energies are crucial for many other important catalytic intermediates.Surface carboxylates are common intermediates in many catalytic reactions and are also used for organo-functionalization of surface. Herein, adsorbed formate, the simplest carboxylate, was produced by dissociative adsorption of formic acid on the Ni(111) surface and the heats of adsorption were measured by single-crystal adsorption calorimetry (SCAC). At 240 K, the integral heat of dissociative adsorption to make adsorbed bidentate formate and adsorbed hydrogen is 117.1 kJ/mol at 1/7 ML. The estimated accuracy of measured heats is within 3 kJ/mol. Therefore, the standard enthalpy of formation of bidentate formate on Ni(111) is derived as −449.3 kJ/mol and the bond enthalpy to the Ni(111) surface is 319.6 kJ/mol. Comparison to our earlier results for Pt(111) shows that bidentate formate is 54 kJ/mol more stable on Ni(111). This is the first experimental measurement of the energy of any molecular fragment bonded to any Ni surface through metal-oxygen bonds, so it is an important benchmark for validating the accuracy of quantum mechanical calculations of a wide range of adsorbed oxygenates on Ni catalysts. Compared to previous density functional theory (DFT) calculations, we found that DFT calculations systematically underestimate the bond energy of formate to Ni(111), but correctly predict that it is stronger to Ni(111) than to Pt(111). The integral heat of HCOOH adsorption to make molecularly adsorbed formic acid was also measured and found to be 63.2 kJ/mol at 0.45 ML and 120–155 K.Download high-res image (68KB)Download full-size image

The energetics of the reactions of water with metal oxide surfaces are of tremendous interest for catalysis, electrocatalysis, and geochemistry, yet the energy for the dissociative adsorption of water was only previously measured on one well-defined oxide surface, iron oxide. In the present paper, the enthalpy of the dissociative adsorption of water is measured on NiO(111)-2 × 2 at 300 K using single-crystal adsorption calorimetry. The differential heat of dissociative adsorption decreases with coverage from 170 to 117 kJ/mol in the first 0.25 ML of coverage. Water adsorbs molecularly on top of that, with a heat of ∼92 kJ/mol. Density functional theory (DFT) calculations reproduce the measured energies well (all within 17 kJ/mol) and provide insight into the atomic-level structure of the surfaces studied experimentally. They show that the oxygen-terminated O-octo(2 × 2) structure is the most stable NiO(111)-2 × 2 termination and gives reaction energies with water that are more consistent with the calorimetry results than the metal-terminated surface. They show that water adsorbs dissociatively on this (2 × 2)-O-octo surface to produce a hydroxyl-covered surface with a heat of adsorption of 171 ± 5 kJ/mol in the low-coverage limit (very close to 170 kJ/mol experimentally) and an integral heat that decreases by 14 kJ/mol up to saturation (compared to ∼30 kJ/mol experimentally). Sensitivity of this reaction’s energy to choice of DFT method is tested using a variety of different exchange correlation functionals, including HSE06, and found to be quite weak.Keywords: adsorption calorimetry; benchmark; density functional theory; heat of adsorption; nickel oxide; surface hydroxyl; water dissociation

The thermodynamic state functions and partition functions for adsorbates on solid surfaces are often treated with two-dimensional (2D) ideal gas and 2D ideal lattice gas models. These are idealized limits of the real situation for adsorbates on solid surfaces, which are more accurately described as hindered translators. We describe a simple extension of the ideal 2D gas model to a more realistic ideal hindered translator model based on our recent approximation for the partition function of ideal hindered translators [Sprowl et al., J. Phys. Chem. C., 2016, DOI: 10.1021/acs.jpcc.5b11616]. Expressions for equilibrium constants and rate constants within transition-state theory (TST) are derived in a self-consistent formalism based on both partition functions and standard-state entropies. The mixing of these three adsorbate models (ideal 2D gas, ideal 2D lattice gas, and ideal hindered translator) within the same equilibrium or rate constant calculation is sometimes necessary but requires careful and consistent choices of standard-state concentrations. The formalism used here facilitates such mixing, using activities instead of concentrations to do so, and also to enable inclusion of nonidealities. We propose a standard state for 2D (and one-dimensional) ideal gases defined such that their translational entropy is 2/3 (or 1/3) that for the corresponding ideal three-dimensional (3D) gas, which offers intuitive advantages for estimating equilibrium and rate constants. This sets the standard-state concentration of the ideal 2D gas to be approximately the 2/3 power of the standard-state concentration of the corresponding 3D ideal gas (i.e., the concentration at 1 bar pressure). We show that in the derivation of the TST rate of elementary steps for ideal 2D lattice gases, the concentration of the transition state often increases as the adsorbate’s activity, θ/(1 – θ), rather than simply as θ, the fractional population of sites, and discuss the implications of this result.

The adsorption of vapor-deposited Au onto CeO2-x(111) thin films (x = 0.05 and 0.2) at 300 and 100 K was studied using single crystal adsorption calorimetry (SCAC). The morphology of Au on these films was investigated using He+ low-energy ion scattering spectroscopy (LEIS) and X-ray photoelectron spectroscopy (XPS) by monitoring the changes in substrate and adsorbate signals with Au coverage. Both techniques indicate that Au grows on CeO1.95(111) as three-dimensional particles in the approximate shape of hemispherical caps with a density of 2.8 × 1012 particles/cm2 at 300 K and 7.8 × 1012 particles/cm2 at 100 K. At 300 K, Au initially grows on CeO1.80(111) with a shape similar to hemispherical caps with a density of 5.4 × 1012 particles/cm2 until ∼1.6 ML Au coverage, above which the Au particles become thicker than hemispherical caps. At 300 K, the initial heat of adsorption of Au onto CeO1.95(111) is 259 kJ/mol, which is 37 kJ/mol lower than that on CeO1.80(111). This indicates stronger binding of Au to oxygen vacancies. On both surfaces, the Au heats of adsorption increase slowly with coverage, approaching the bulk heat of sublimation of Au(solid) (368 kJ/mol) by ∼2 ML (3.2 nm in diameter on CeO1.95(111) and 2.4 nm on CeO1.80(111)). The heat of adsorption remains higher on the reduced surface than on the oxidized surface at all particle sizes. At 100 K, the initial heat of Au adsorption onto CeO1.95(111) is 209 kJ/mol (50 kJ/mol lower than at 300 K), which is due to a higher fraction of Au atoms adsorbing to terraces rather than at step sites. The adhesion energy of Au(solid) to CeO1.95(111) at 300 K was found to be 2.53 J/m2 for 3.6 nm diameter particles and 2.83 J/m2 onto CeO1.80(111) for 2.5 nm diameter particles. This further indicates that Au particles bind more strongly to surfaces with a larger fraction of oxygen vacancies.

The energies of adsorbates containing H, N, C, O, and halogens that are of interest as intermediates, poisons, and promoters in catalytic reactions have been measured on well-defined single-crystal surfaces by equilibrium adsorption isotherms, temperature-programmed desorption (TPD), and single-crystal adsorption calorimetry (SCAC). Here we tabulate a large collection of those experimental adsorption energies which we consider to be particularly reliable based on reproducibility by other groups, comparisons to results for closely related systems, and/or reliability of other results reported by the same group. Specifically, we list the enthalpies and energies of 81 molecular and dissociative adsorption reactions that were measured on 26 different metal single-crystal faces of 12 different late transition metals, and we extract from these the standard enthalpies of formation of the adsorbates thus produced. These can serve as benchmarks for validating computational methods for estimating surface reaction energies.

•Quantitative analysis of low-energy ion scattering spectroscopy (LEIS) signals with He+ and Ne+ ions•Hemispherical cap model of 3D particle growth analyzed for any measurement geometry•Provides average particle thickness and diameter, and number density of particles•Quantitative analysis of the fraction of the surface covered by nanoparticlesNanoparticles of one element or compound dispersed across the surface of another substance form the basis for many materials of great technological importance, like catalysts, fuel cells, sensors and biomaterials. Nanoparticles also often grow during thin film deposition. The size and number density of such nanoparticles are important, often estimated with electron or scanning tunneling microscopies. However, these are slow and often unavailable with sufficient resolution for particles near 1 nm. Because the probe depth of low-energy ion scattering spectroscopy (LEIS) with He+ and Ne+ is so shallow (less than one atom), it provides quantitative information on the fraction of the surface that is covered by such nanoparticles. Combined with the total amount per unit area, this fraction provides the average particle thickness. When the ions are incident or detected at some angle away from the surface normal, macroscopic screening effects cause interpretation of LEIS signals in terms of area fraction covered to be complicated. In this paper, we report a geometric analysis of particles with the shape of hemispherical caps so that LEIS signals obtained in any measurement geometry can also be used to quantitatively determine the area fraction, average particle thickness and diameter, or number density of particles.

Nanoparticles of one element or compound dispersed across the surface of another substrate element or compound form the basis for many materials of great technological importance, such as heterogeneous catalysts, fuel cells and other electrocatalysts, photocatalysts, chemical sensors and biomaterials. They also form during film growth by deposition in many fabrication processes. The average size and number density of such nanoparticles are often very important, and these can be estimated with electron microscopy or scanning tunneling microscopy. However, this is very time consuming and often unavailable with sufficient resolution when the particle size is ~ 1 nm. Because the probe depth of electron spectroscopies like X-Ray Photoelectron Spectroscopy (XPS) or Auger Electron Spectroscopy (AES) is ~ 1 nm, these provide quantitative information on both the total amount of adsorbed material when it is in the form of such small nanoparticles, and the particle thickness. For electron spectroscopy conducted with electron detection normal to the surface, Diebold et al. (1993) derived analytical relationships between the signal intensities for the adsorbate and substrate and the particles' average size and number density, under the assumption that all the particles have hemispherical shape and the same radius. In this paper, we report a simple angle- and particle-size-dependent correction factor that can be applied to these analytical expressions so that they can also be extended to measurements made at other detection angles away from the surface normal. This correction factor is computed using numerical integration and presented for use in future modeling. This correction factor is large (> 2) for angles beyond 60°, so comparing model predictions to measurements at both 0° and ≥ 60° will also provide a new means for testing the model's assumptions (hemispherical shape and fixed size particles). The ability to compare the hemispherical cap model at several angles simultaneously also should enable more accurate estimates of surface structural parameters when elastic diffraction effects cause strong peaks in the angular distributions of emitted electrons.