First-principles study of small palladium clusters on NiAl(1 1 0) alloy surface

Theoretical calculations focused on the geometry, stability, electronic and magnetic properties of small palladium clusters Pd_n (n=1–5) adsorbed on the NiAl(1 1 0) alloy surface were carried out within the framework of density functional theory (DFT). In agreement with the experimental observations, both Ni-bridge and Al-bridge sites are preferential for the adsorption of single palladium atom, with an adsorption energy difference of 0.04 eV. Among the possible structures considered for Pd_n (n=1–5) clusters adsorbed on NiAl(1 1 0) surface, Pd atoms tend to form one-dimensional (1D) chain structure at low coverage (from Pd₁ to Pd₃) and two-dimensional (2D) structures are more stable than three-dimensional (3D) structures for Pd₄ and Pd₅. Furthermore, metal-substrate bonding prevails over metal–metal bonding for Pd cluster adsorbed on NiAl(1 1 0) surface. The density of states for Pd atoms of Pd/NiAl(1 1 0) system are strongly affected by their chemical environment. The magnetic feature emerged upon the adsorption of Pd clusters on NiAl(1 1 0) surface was due to the charge transfer between Pd atoms and the substrate. These findings may shade light on the understanding of the growth of Pd metal clusters on alloy surface and the construction of nanoscale devices.     

Thermodynamics and kinetics of CO and benzene adsorption on Pt(111) studied with pulsed molecular beams and microcalorimetry

The adsorption and desorption of the system CO/Pt(111) and C₆H₆/Pt(111) at 300 K has been investigated with a pulsed molecular beam method in combination with a microcalorimeter. For benzene the sticking probability has been measured in dependence of the coverage θ. For coverages θ > 0.8 transient adsorption is observed. From an analysis of the time-dependence of the molecular beam pulses the rate constant for desorption is determined to be 5.6 s^? 1. With a precursor-mediated kinetic adsorption model this allows to obtain also the hopping rate constant of 95.5 s^? 1. The measured adsorption enthalpies could be best described by (199 ? 77θ ? 51θ²) kJ/mol, in good agreement with the literature values. For CO on Pt(111) also transient adsorption has been observed for θ > 0.95 at 300 K. The kinetic analysis yields rate constants for desorption and hopping of 20 s^?1 and 51 s^?1, respectively. The heats of adsorption show a linear dependence on coverage (131 ? 38θ) kJ/mol between 0 ≤ θ ≤ 0.3, which is consistent with the desorption data from the literature. For higher coverage (up to θ = 0.9ML) a slope of ?63 kJ/mol describes the decrease of the differential heat of adsorption best. This result is only compatible with desorption experiments, if the pre-exponential factor decreases strongly at higher coverage. We found good agreement with recent quantum chemical calculations made for (θ = 0.5ML).     

H2 splitting on Pt,Ru and Rh nanoparticles supported on sputtered HOPG

The equilibrium hydrogen exchange rate between adsorbed and gas phase hydrogen at 1 bar is measured for Pt, Ru and Rh nanoparticles supported on a sputtered HOPG substrate. The particles are prepared by Electron Beam Physical Vapor Deposition and the diameter of the particles varies between 2 and 5 nm. The rate of hydrogen exchange is measured in the temperature range 40–200 °C at 1 bar, by utilization of the H–D exchange reaction. We find that the rate of hydrogen exchange increases with the particle diameter for all the metals, and that the rate for Ru and Rh is higher than for Pt. In the case of Pt, the equilibrium dissociative sticking probability, S, is found to be nearly independent of particle diameter. For Ru and Rh, S is found to depend strongly on particle diameter, with the larger particles being more active. The apparent energy of desorption at equilibrium, E_app, shows a dramatic increase with decreasing particle diameter for diameters below 5 nm for Ru and Rh, whereas E_app is only weakly dependent on particle diameter for Pt. We suggest that the strong variation in the apparent desorption energy with particle diameter for Ru and Rh is due to the formation of compressed hydrogen adlayers on the terraces of the larger particles. Experiments are also carried out in the presence of 10 ppm CO. Pt is found to be very sensitive to CO poisoning and the H–D exchange rate drops below the detection limit when CO is added to the gas mixture. In the case of Ru and Rh nanoparticles, CO decreases the splitting rate significantly, also at 200 °C. The variation of the sensitivity to CO poisoning with particle diameter for Ru and Rh is found to be weak.     

Adsorption of thiophene on silica-supported Mo clusters

The adsorption/decomposition kinetics/dynamics of thiophene has been studied on silica-supported Mo and MoS_x clusters. Two-dimensional cluster formation at small Mo exposures and three-dimensional cluster growth at larger exposures would be consistent with the Auger electron spectroscopy (AES) data. Thermal desorption spectroscopy (TDS) indicates two reaction pathways. H₄C₄S desorbs molecularly at 190–400 K. Two TDS features were evident and could be assigned to molecularly on Mo sites, and S sites adsorbed thiophene. Assuming a standard preexponential factor (ν = 1 × 10¹³/s) for first-order kinetics, the binding energies for adsorption on Mo (sulfur) sites amount to 90 (65) kJ/mol for 0.4 ML Mo exposure and 76 (63) kJ/mol for 2 ML Mo. Thus, smaller clusters are more reactive than larger clusters for molecular adsorption of H₄C₄S. The second reaction pathway, the decomposition of thiophene, starts at 250 K. Utilizing multimass TDS, H₂, H₂S, and mostly alkynes are detected in the gas phase as decomposition products. H₄C₄S bond activation results in partially sulfided Mo clusters as well as S and C residuals on the surface. S and C poison the catalyst. As a result, with an increasing number of H₄C₄S adsorption/desorption cycles, the uptake of molecular thiophene decreases as well as the H₂ and H₂S production ceases. Thus, silica-supported sulfided Mo clusters are less reactive than metallic clusters. The poisoned catalyst can be partially reactivated by annealing in O₂. However, Mo oxides also appear to form, which passivate the catalyst further. On the other hand, while annealing a used catalyst in H/H₂, it is poisoned even more (i.e., the S AES signal increases). By means of adsorption transients, the initial adsorption probability, S₀, of C₄H₄S has been determined. At thermal impact energies (E_i = 0.04 eV), S₀ for molecular adsorption amounts to 0.43 ± 0.03 for a surface temperature of 200 K. S₀ increases with Mo cluster size, obeying the capture zone model. The temperature dependence of S₀(T_s) consists of two regions consistent with molecular adsorption of thiophene at low temperatures and its decomposition above 250 K. Fitting S₀(T_s) curves allows one to determine the bond activation energy for the first elementary decomposition step of C₄H₄S, which amounts to (79 ± 2) kJ/mol and (52 ± 4) kJ/mol for small and large Mo clusters, respectively. Thus, larger clusters are more active for decomposing C₄H₄S than are smaller clusters.     

Visible light induced photodesorption of NO from the α-Cr2O3(0001) surface

Nitric oxide chemistry and photochemistry on the Cr-terminated surface of α-Cr₂O₃(0001) were examined using temperature programmed desorption (TPD), sticking coefficient measurements and photodesorption. NO exposed to α-Cr₂O₃(0001) at 100 K binds at surface Cr cation sites forming a strongly bound surface species that thermally desorbs at 320–340 K, depending on coverage. No thermal decomposition was detected in TPD in agreement with previous results in the literature. Sticking probability measurements at 100 K indicated near unity sticking for NO up to coverages of ~ 1.3 ML, with additional adsorption with higher exposures at decreased sticking probability. These results suggest that some Cr cation sites on the α-Cr₂O₃(0001) surface were capable of binding more than one NO molecule, although it is unclear whether this was as separate NO molecules or as dimers. Photodesorption of adsorbed NO was examined for surface coverages below the 1 ML point. Both visible and UV light were shown to photodesorb NO without detectable NO photodecomposition. Visible light photodesorption of NO occurred with a greater cross section than estimated using UV light. The visible light photodesorption event was not associated with bandgap excitation in α-Cr₂O₃(0001), but instead was linked to excitation of a surface Cr^3 +–NO^? charge transfer complex. These results illustrate that localized photoabsorption events at surface sites with unique optical properties (relative to the bulk) can result in unexpected surface photochemistry.     

Magnetic characterisation and hydrogen absorption characteristics of Pr3(Fe,Ti)29Hx

An interstitial Pr₃(Fe,Ti)₂₉ hydride was synthesised by gas-phase hydrogenation on Pr₃(Fe,Ti)₂₉ powder using H₂. The reaction kinetics between Pr₃(Fe,Ti)₂₉ and H₂ gases was studied in a constant-volume reactor. The sample starts to rapidly absorb hydrogen, interstitially, at about 533 K. Absorption passes through a maximum at about 598 K (1.6 H/f.u) and then interstitial hydrogen desorption takes place up to the temperature of 673 K. By cooling to room temperature, the sample absorbs more hydrogen, interstitially, reaching the value of 3.6 H/f.u. By remaining at room temperature, the sample absorbs even more hydrogen reaching the value of 5.2 H/f.u. The lattice expansion observed is 2.1% and the Curie temperature, T_C, increased from 392 to 518 K. The hydride exhibits saturation magnetisation, M_S, of 145.4 and 157.5 Am²/kg at room temperature (RT) and at 5 K, respectively, anisotropy field, H_A, of 2.1 T (RT) and 4.5 T (5 K) and average hyperfine field, H_eff, of 23.3 T (RT). The magnetic anisotropy of Pr₃(Fe,Ti)₂₉ hydride is the same as that of the parent compounds, easy-cone-like, changing only in the cone angle (from 34° to 26°).     

Interaction of atomic H and CO with Cu(1 1 0) and bimetallic Ni/Cu(1 1 0)

Co-adsorption of hydrogen and CO on Cu(1 1 0) and on a bimetallic Ni/Cu(1 1 0) surface was studied by thermal desorption spectroscopy. Hydrogen was exposed in atomic form as generated in a hot tungsten tube. The Ni/Cu surface alloy was prepared by physical vapor deposition of nickel. It turned out that extended exposure of atomic hydrogen leads not only to adsorption at surface and sub-surface sites, but also to a roughening of the Cu(1 1 0) surface, which results in a decrease of the desorption temperature for surface hydrogen. Exposure of a CO saturated Cu(1 1 0) surface to atomic H leads to a removal of the more strongly bonded on-top CO (α₁ peak) only, whereas the more weakly adsorbed CO molecules in the pseudo threefold hollow sites (α₂ peak) are hardly influenced. No reaction between CO and H could be observed. The modification of the Cu(1 1 0) surface with Ni has a strong influence on CO adsorption, leading to three new, distinct desorption peaks, but has little influence on hydrogen desorption. Co-adsorption of H and CO on the Ni/Cu(1 1 0) bimetallic surface leads to desorption of CO and H₂ in the same temperature regime, but again no reaction between the two species is observed.     

Diffusion-promoted-desorption mechanism for D2 desorption from Si(100) surfaces

Temperature-programmed-desorption (TPD) spectra and isothermal desorption rates of D₂ molecules from a Si(100) surface have been calculated to reproduce experimental β_1, A-TPD spectra and isothermal desorption rate curves. In the diffusion-promoted-desorption (DPD) mechanism, hydrogen desorption from the Si(100) (2 × 1) surfaces takes place via D atom diffusion from doubly-occupied Si dimers (DODs) to their adjacent unoccupied Si dimers (UODs). Taking a clustering interaction among DODs into consideration, coverages θ_DU of desorption sites consisting of a pair of a DOD and UOD are evaluated by a Monte Carlo (MC) method. The TPD spectra for the β_1, A peak are obtained by numerically integrating the desorption rate equation R = ν_A exp(? E_d, A / k_BT)θ_DU, where ν_A is the pre-exponential factor and E_d, A is the desorption barrier. The TPD spectra calculated for E_d, A = 1. 6 eV and ν_A = 2.7 × 10⁹ /s are found to be in good agreement with the experimental TPD data for a wide coverage range from 0.01 to 0.74 ML. Namely, the deviation from first-order kinetics observed in the coverage dependent TPD spectra as well as in the isothermal desorption rate curves can be reproduced by the model simulations. This success in reproducing both the experimental TPD data and the very low desorption barrier validates the proposed DPD mechanism.     

Microcalorimetry of O2 and NO on flat and stepped platinum surfaces

Using the single-crystal adsorption calorimeter (SCAC), coverage-dependent heats of adsorption and sticking probabilities are reported for O₂ and NO on Pt{1 1 1}, Pt{2 1 1} and Pt{4 1 1} at 300 K. At low coverage, oxygen adsorption is dissociative for all Pt surfaces. The highest initial heat of adsorption is found on Pt{2 1 1}, with a value of 370 kJ/mol, followed by those on Pt{4 1 1} (310 kJ/mol) and Pt{1 1 1} (300 kJ/mol). We attribute this relatively large difference in the dissociative heat of adsorption at low coverage to the step character of the {2 1 1} surface. Initial sticking probabilities, s_o, are similar for the three surfaces, 0.22 on Pt{1 1 1}, 0.17 on Pt{2 1 1} and 0.18 on Pt{4 1 1}, rapidly decreasing as the oxygen coverage increases. For nitric oxide, the initial heats of adsorption are very similar and consistent with either dissociative or molecular adsorption, with values of 182 kJ/mol on Pt{1 1 1}, 192 kJ/mol on Pt{2 1 1} and 217 kJ/mol on Pt{4 1 1}. The s_o value is virtually identical for all three systems, with values ranging from 0.82 to 0.85, suggesting that the initial sticking probability is insensitive to the surface structure and adsorption is intrinsically precursor mediated. SCAC data are also used to evaluate pre-exponential factors, ν, for first-order desorption at high coverage where adsorption is non-dissociative. Values of 3 × 10¹⁸, 6 × 10¹⁸ and 2 × 10¹⁸ s^?1 for O₂, and 4 × 10¹⁹, 6 × 10¹⁷ and 2 × 10²⁰ s^?1 for NO on Pt{1 1 1}, Pt{2 1 1} and Pt{4 1 1}, respectively, are found. These unexpectedly high values are rationalised in terms of conventional transition state theory entropy changes.     

Coverage effects on the adsorption of sulfur on Co(0 0 0 1): A DFT study

The adsorption of sulfur on Co(0 0 0 1) was studied using density functional theory calculations at coverage from 0.11 ML to 1.0 ML. Calculated results indicate that atomic S favors in hollow sites with bond S–Co dominated at lower coverage and at higher coverage the strong adsorbate S–S interaction leads to the formation of S₂ species. It was shown that the adsorption energy generally increases (gets weaker) with the coverage in a near linear fashion for the most stable configurations. In addition, modification of the surface electronic properties has been discussed and some discrepancy are found between our calculations and the findings of O adsorption on Au(1 1 1) and Pt(1 1 1) surfaces.     

Surface alloy formation,short-range order,and deuterium adsorption properties of monolayer PdRu/Ru(0 0 0 1) surface alloys

The formation and structure of monolayer PdRu/Ru(0 0 0 1) surface alloys and their adsorption properties with respect to deuterium adsorption were investigated by atomic resolution scanning tunneling microscopy and by temperature programmed desorption. Surface alloys, prepared by deposition of up to one monolayer of Pd and flash annealing to 1150 K show (i) negligible loss of Pd by desorption or diffusion into the Ru bulk during this procedure and (ii) dominant phase separation into 2D Pd and Ru islands, in contrast to the random distribution in PtRu/Ru(0 0 0 1) surface alloys [H.E. Hoster, A. Bergbreiter, P.M. Erne, T. Hager, H. Rauscher, R.J. Behm, Phys. Chem. Chem. Phys. 10 (2008) 3812]. 2D short-range order parameters and the abundance of specific adsorption ensembles were evaluated for different Pd contents, effective pair interaction (EPI) energies were derived from Monte Carlo simulations. Deuterium adsorption on Pd monolayer films shows a pronounced weakening of adsorption bond, which is attributed to compressive strain and metal–metal interactions between Pd and underlying Ru atoms (‘vertical ligand effect’). Mixed adsorption ensembles containing both Pd and Ru atoms give rise to D₂ desorption in the intermediate temperature regime. The impact of the specific lateral distribution of the two metal species on the chemical surface properties is illustrated by comparison with deuterium adsorption on dispersed PtRu/Ru(0 0 0 1) surface alloys [T. Diemant, H. Rauscher, R.J. Behm, J. Phys. Chem. C 112 (2008) 8381].     

The adsorption of acetic acid on clean and oxygen-covered Au/Pd(100) alloy surfaces

The adsorption of acetic acid is studied on clean and oxygen-covered Au/Pd(100) alloys as a function of gold content by temperature-programmed desorption and reflection–absorption infrared spectroscopy. Au/Pd(100) forms ordered alloys such that, for gold coverages above ~ 0.5 monolayers, only isolated palladium atoms surrounded by gold nearest neighbors are present. Predominantly molecular acetic acid forms on Au/Pd(100) alloy surfaces for gold coverages greater than ~ 0.56 ML, and desorbs with an activation energy of ~ 59 kJ/mol. Heating this surface also forms some η¹-acetate species which decompose to form CO and hydrogen. On alloy surfaces with palladium–palladium bridge sites, η¹-acetate species initially form, but rapidly convert into η²-species. They thermally decompose to form CO and hydrogen, with a small portion rehydrogenating to form acetic acid between 280 and 321 K depending on gold coverage. The presence of oxygen on both Pd(100) and Au/Pd(100) alloys facilitates acetate dehydrogenation so that only η²-acetate species form on these surfaces. The presence of oxygen also serves to stabilize the acetate species.     

Activation of methyl acetate on Pd(111)

The adsorption and activation of methyl acetate (CH₃COOCH₃), one of the simplest carboxylic esters, on Pd(111) have been studied using self-consistent periodic density functional theory calculations. Methyl acetate adsorbs weakly through the carbonyl oxygen. Its activation occurs via dehydrogenation, instead of direct C–O bond dissociation, on clean Pd(111): It is much more difficult to dissociate the C–O bonds (E_a ≈ 2.0 eV for the carbonyl and acetate–methyl bonds; E_a = 1.0 eV for the acetyl–methoxy bond) than to dissociate the C–H bonds to produce enolate (CH₂COOCH₃; E_a = 0.74 eV) or methylene acetate (CH₃COOCH₂; E_a = 0.82 eV). The barriers for C–H and C–O bond dissociation are directly calculated for enolate and methylene acetate, and estimated for further dehydrogenated derivatives (CH₃COOCH, CH₂COOCH₂, and CHCOOCH₃) based on the Brønsted–Evans–Polanyi linear energy relations formed by the calculated steps. The enolate pathway leads to successive dehydrogenation to CCOOCH₃, whereas methylene acetate readily dissociates to yield acetyl. The selectivity for dissociating the acyl–alkoxy C–O bond, which is desired for alcohol formation, is therefore fundamentally limited by the facility of dehydrogenation under vacuum/low-pressure conditions on Pd(111).     

EXAFS study of size dependence of atomic structure in palladium nanoparticles

Dependence of atomic structure of Palladium nanoparticles on supports Al₂O₃ and SiO₂ upon their size, changed from 1.3 to 10.5 nm, was studied by Pd K-edge EXAFS. To determine the structure of the interior (core) and the near surface regions of nanoparticle, the fitting technique of the Fourier-transforms F(R) of spectra was used, which enabled to overcome instabilities of the obtained structural parameters values. The processing of experimental data was performed using results of the study of features formation in │F(R)│ of Pd K-EXAFS in Pd foil. By this approach it was revealed that the local structure of Pd atoms in the core is similar to fcc structure of bulk Pd, irrespective of size. The percentage of Pd atoms, which can be attributed to the core, upon the particles size was determined and the obtained dependence was described by the “cluster size equation”. In the near surface region of nanoparticles, nearest-neighbors Pd–Pd distances show a large Debye–Waller parameters and the mean bond length slightly contracted for nanoparticles of sizes less than ~2 nm. The effect of small structural distortions in the vicinity of absorbing Pd atom in the near surface region was studied using the cluster model of nanoparticle.     

Flame spread along a thin solid randomly distributed combustible and noncombustible areas

Flame spread route in fire strongly depends on distribution of combustible materials. Two types of scenario are considered in flame spread when combustible materials randomly distributed; one case is that flame spreads and combustible materials burn out, and the other case is that flame self-extinguishes on the way. The threshold of burning out or self-extinguishing may be determined by quantity of combustible materials and their placement in space. Our objectives are to clarify the characteristics and threshold of flame spread. In this paper, we examine non-uniform flame spread in open air along a thin combustible solid with randomly distributed pores, which are considered as noncombustible space. Experimental results show that the flame spread rate for S ≦ 1 (S ≡ d/L_h, S: scale ratio, d: pore-scale, L_h: pre-heat length ahead of flame leading edge measured by using a shadowgraph method) increases with increasing the porosity and reaches maximum value approximately at 20–30% of porosity, while the flame spread rate for S > 1 is almost constant. Over 40% of porosity, the flame spread rate for both S ≦ 1 and S > 1 decreases. The flame cannot spread and completely self-extinguish over 60% of porosity independently with pore-scale and shape. The threshold of flame spread is related with the average-number of slit, N_s, which is made by connecting each pores. The N_s as the threshold of flame spread is unity for S > 1, while the modified average-number of slit (N_s × S) becomes two for S ≦ 1.     

Desorption of O2 from a stepped platinum(113) surface during 308 nm laser irradiation

Angular and velocity distributions of desorbing O₂ during irradiation of 308 nm laser pulses were studied on a stepped Pt(1 1 3) surface. With increases in the coverage, three desorption components collimated at around 12°, 30° and 50° successively appeared when the desorption angle was changed in a plane along the step edge. The translational temperature also showed maxima at these collimation angles, and the values were slightly lower than previous results for 193 nm irradiation. Some possible desorption mechanisms are discussed.     

H adsorption at Ag/Si interfaces in epitaxially grown Ag(1 1 1) films on Si(1 1 1)7 × 7 substrates

The interaction of atomic H with Ag(1 1 1)/Si(1 1 1)7 × 7 surfaces was studied by thermal desorption (TD) spectroscopy and scanning tunneling microscopy (STM) at room temperature. TD spectroscopy revealed an intense peak from mono H–Si bonds, even though the Si surface was covered by the Ag atoms. This peak was not observed from Ag-coated SiO₂/Si substrates. STM observation showed no clear change of the Ag surface morphology resulting from H exposure. All these results indicate that the atomic H adsorbs at neither the Ag surfaces nor Ag bulk sites, but at the Ag/Si interface by diffusing through the Ag film.     

Evidence for non-thermalized vibrationally excited molecule production during azomethane pyrolysis in methane desorption from Cu(001)

A Cu(001) surface was exposed to products of an azomethane pyrolysis doser at varying temperatures. In addition to methyl radical adsorption, for certain doser conditions one or more doser emergent species can undergo an activated adsorption on the copper face. Directly after exposures, temperature programmed desorption between 170 K and 500 K was used to indicate the relative concentrations of adsorbed atomic hydrogen and methyl species, and thermally induced surface reactions. Two methane desorption features were invariably observed, indicating the presence of adsorbed methyl groups (CH₃) and transient adsorbed atomic hydrogen. The deduced relative surface concentrations levels of both H and CH₃ depend on the total exposures and the operating temperatures of the azomethane pyrolysis doser. The initial H concentrations apparent at surface temperatures between 275 K and 375 K are shown to arise from defect-related methyl decomposition and, at high operating doser temperatures, from the initial adsorption of one or more activated Cu incident species. It is proposed that the distributions of vibrational energies of emergent molecular hydrogen or methane species from higher temperature dosers are non-thermal. Hence, with doser temperatures of 800 °C or above, the effects of subsequent dissociative molecular adsorption on the copper surface can dominate over Cu defect chemistries.     

Stabilization of high spin FCC Fe in (Fe/Pd)n multilayers

Polycrystalline (Fe/Pd)_n multilayers are grown onto sapphire substrates at room temperature in a UHV system. The number of periods n=40 and the thickness of Pd layers of t_Pd=4 nm are kept constant, whereas the thickness of the Fe layers is varied from 1.5 to 5 nm. Structural properties are studied by in situ reflection high energy diffraction (RHEED), scanning tunnelling microscopy (STM) and ex situ by X-ray diffraction at small angles and large angles. Analyzing the experimental data using the program SUPREX we obtain interplanar distances of d_Fe=2.03±0.01 Å for an Fe layer thickness larger than about 2.5 nm as expected for (1 1 0) planes of BCC Fe. For Fe layers with thicknesses less than about 2.5 nm the interplanar distance is d_Fe=2.1±0.01 Å, which is close to the distance between (1 1 1) planes of FCC Fe with a lattice parameter of a=3.64 Å. Magnetic susceptibility measurements at temperatures between 1.5 and 300 K for (Fe/Pd)_n multilayers with FCC Fe yield a magnetic moment per Fe atom of μ=2.7±0.1 μ_B, which is about 20% larger compared to μ=2.2 μ_B for BCC Fe. We show that the occurrence of the large magnetic moment originates from FCC Fe being in the high spin (HS) state rather than from polarization effects of Pd at Fe/Pd interfaces.