Estudio de catalizadores basados en platino para la reacción de PROX: influencia de los depósitos de carbono.

Estudio De Catalizadores Basados En Platino Para La Reacción De PROX: Influencia De Los Depósitos De Carbono Estudio De Catalizadores Basados En Platino Para La Reacción De PROX: Influencia De Los Depósitos De Carbono D. Rafael Castillo Barrero Tesis Doctoral 2015-2018 Universidad de Sevilla & Normandie Université A mis dos madres. La que nunca se fue y la que me besa todas las noches Main Motivation. Estudio De Catalizadores Basados En Platino Para La Reacción De PROX: Influencia De Los Depósitos De Carbono Main motivation I Low working temperatures, high energy density, low weight, compactness or rapid start-up/ Shut-down cycles doing Polymer-electrolyte membrane fuel cell (PEMFC) development research is one of the most attractive topics in the search of clean energy for mobile devices. Hydrogen flow from the hydrocarbon reforming feed the PEMFC. However, reforming process involve significant amount of carbon monoxide formation that must be reduced below 10 ppm for avoiding the poison of the Pt catalyst on the fuel cell anode surface. Catalytic methanation process and preferential oxidation of CO in excess of H2 stream (PrOx reaction) have been considerate the most promising ways for obtaining CO-free hydrogen flow1,2. In the case of PrOx process, two competitive reactions are involved: (1) CO + O2 → CO2 H = -283 kJ/mol (2) H2 + O2 → H2O H = -242 kJ/mol Pt based catalysts present noticeable PROX activities, especially low loading Pt/Al2O3 catalysts3. Manasilp et al.4 observed CO conversions around 80% at 170ºC with selectivity values of 50%. However, they probed the dramatically CO conversion decrement with the increment of CO2 concentration. Considering this work, Avgouropoulos et al.5 reported a Pt/Al2O3 Catalyst with 100% of CO conversion for at least 110 h at 100ºC in a reaction flow composition of 1% CO, 1.25% O2, 50% H2, 15% CO2 and the balance He. On the other hand, Cu based catalyst also present excellent catalytic properties for the preferential oxidation of CO. T. Reina et al.6 carried out a comparison of CuOCeO2 based catalysts behavior in the CO clean up reactions (WGS and PrOx). They reported CuOCeO2/Al2O3 series catalyst with full conversion and selectivity at 130ºC for PrOx reaction. Main motivation II Nowadays, catalysts development for PrOx reaction applied in mobile devices is focused in Platinum catalysts modified by copper promoters which exhibit a remarkable enhanced of the PrOx activity and selectivity in the working temperature of PEMFCs1. Komatsu et al.7 showed the different catalytic behavior of Pt/Al2O3, Cu/Al2O3 and PtCu/Al2O3 catalysts (figure A). Those authors concluded that thanks to the electrophilic character of the alloy copper atoms in the presence of CO, PtCu/Al2O3 catalyst get full conversions and higher selectivity values than the monometal catalysts supported on Al2O3. Figure A. Catalytic activity evolution of Pt/Al2O3, Cu/Al2O3 and PtCu/Al2O3 catalysts. The application of PtCu/Al2O3 catalysts in mobile devices do necessary the conditioning catalyst in own reaction mixture atmosphere. For that, R. Farrauto3 patented the formulation of PtCu/Al2O3 systems with almost 1:3 stoichiometric ratio with extremely high catalytic activity and selectivity to CO2 after conditioning treatment in the reaction mixture environment. However, the wonderful performance of PtCu3/Al2O3 catalysts is not totally controlled. Previous works carried out in our group by Romero-Sarria et al.8 concluded that Main motivation III PtCu/Al2O3 catalytic behavior depends of conditioning treatments parameters and reaction mixture composition. Those authors showed the effect of conditioning treatment atmosphere, the heating rate in the conditioning treatment and the reaction mixture composition in the CO conversion (figure 1.B). Figure B. Influence of conditioning treatment parameters in the catalytic activity of PtCu3/Al2O3. A comparison of the CO conversion after conditioning treatment at 200ºC (line red) and 300ºC (blue line) is shown (a). Figure A.b. shows the CO conversion as a function of the conditioning treatment heating rate. Figure B.a. shows the evident effect of the conditioning treatment parameters in the catalytic activity of this system. Interesting, lower conditioning treatment temperatures involve higher CO conversion while higher heating rates supposed an enhanced of the catalytic behavior (figure B.b). Parallel, CO conversion of PtCu3/Al2O3 catalyst as a function of the conditioning treatment environment is shown in figure XX. Main motivation IV Figure C. CO conversion of PtCu3/Al2O3 as a function of the conditioning treatment environment. Figure C. shows the high effect in the CO conversion of the conditioning treatment atmosphere. Oxidizing environment presents lower CO conversion values with an almost increasing linear trend as a function of the temperature. When conditioning environment is mainly composed by reducer atmosphere (reaction mixture is 50% H2), catalytic activity obtains CO conversions above 75% at 120ºC. From this point, CO conversion reaches chemical equilibrium values. Moreover, second reaction cycles mean higher initial CO conversion values in all the samples pretreated under reducer atmospheres. Trying to understand the untypical catalytic behavior, S. Palma9 carried out a simulation of the conditioning treatment under reaction mixture atmosphere by operando FTIR (figure D). Main motivation V Figure D. FTIR spectra of Pt/Al2O3 catalyst surface submitted to different temperatures under a reaction mixture composition of 50% H2, 0.25% CO, 0.5% O2, 17% CO2, 15% H2O and N2 balance. Conditioning treatment at 200ºC with heating rate of 15 ºC/min. Red spectrum was acquired before the treatment under Ar atmosphere. Blue color spectra were taken during the ramping temperature and grey color spectra are the temporal surface evolution under reactive flow at 200ºC. They found initial bands at 1460, 1574, 2976 and 2931 cm-1 correspond to absorbed acetates species from the synthesis step. Catalysts were synthetized by support impregnation with acetic acid. When catalyst was exposed under reaction environment, bands characteristic to CO2 gas phase and a new band at 2032 cm-1 assigned to CO absorbed on metallic platinum become evident. During the isothermal period at 200ºC, different surface modifications were observed. First, the band at 2032 cm-1 was shifted to 2028 cm-1 while new bands at 2870, 2927 and 2967 cm-1 correspond to ν(CH) stretching vibrational modes increase with time. They related it with the presence of surface carbon deposits, from the reducer or capping agents used in synthesis step previously detected, which are hydrogenated due to the high H2 concentration in the reaction mixture. The increment of hydrocarbons species on the surface catalyst involves an electronic interaction with the metallic site showed by the shit to lower Chapter 1. Experimental Estudio De Catalizadores Basados En Platino Para La Reacción De PROX: Influencia De Los Depósitos De Carbono Experimental 7 1.1 Ambient Pressure XPS X-Ray photoelectron spectroscopy (XPS) provides essential information for the total understanding of materials. Elemental composition, oxidation state of the surface or surface dynamics could be elucidated by this technique. XPS is based on the photoelectric effect1. Figure 1.1 shows a schematic illustration of the photoemission process. Incident energy excites an atom which absorbs a photon of energy hν. Then, core electron of binding energy Eb with respect to the Fermi Level EF is ejected with kinetic energy EK by Eq (1.1): Ek = hν - Eb - (1) Where Ek is the kinetic energy of the photoelectron, h is the Planck´s constant, ν is the frequency of the exciting radiation and is the work function of the spectrometer. Figure 1.1.Schematic Photoemission process illustration. XPS has been traditionally performed under ultra-high vacuum conditions (UHV). However, M. Salmeron and R. Schlorgl2 designed a new tool for surface science and catalysis breaking the pressure gap between the working conditions and the UHV conditions in the analysis chamber of the conventional XPS system. They developed a differentially pumped electron energy analyzer which is able to acquire spectra under working conditions at pressures from UHV to almost 20 millibar at the same time that Experimental 8 the UHV conditions is keeping in the detector. This development allows the study os solid-gas interfaces under working conditions and it was called “Ambient pressure X- ray photoelectron spectroscopy”. The APXPS experiments of the present manuscript were carried out in CIRCE beamline of the ALBA synchrotron light source (Barcelona, Spain). PHOIBOS 150 NAP electron energy analyzer (SPECS) equipped with differential pumping system was employed. The working pressure range involves from UHV to 20 mbar. Synchrotron radiation provides very high photon flux than in conventional sources which leads to acquired spectra with higher resolution. CIRCE beamline is an undulator beamline with photon energy range 100-2000 eV. The beam spot size at the sample is 20 x 100 m2. Spectra were acquired with pass energy of 20 eV which leads to energy resolution lower than 300 meV full width at half-maximum (FWHM)3,4. Figure 1.2.a) CIRCE Beamline at ALBA synchrotron light source. b) PHOIBOS 150 NAP electron energy analyzer (SPECS) equipped with differential pumping system Experimental 9 1.2 CO as Probe molecule Besides of the non-reactive character and the high quality information that can be recorded by the use of this probe molecule, the vibrational spectrum of CO is simple. The (CO) stretching vibration mode and the extinction coefficient depend of the formed bond nature with the surface. Electronic structure of CO based in sp hybridization. In figure 1.3 can be observed the filling of bonding orbitals , 4 and 1 (1 x and 1 y) and the anti-bonding orbital 5 as well as the empty molecular orbitals 2 and 6 . Therefore, highest occupied molecular orbital (HOMO) is the anti-bonding orbital 5 and the low unoccupied molecular orbital (LUMO) are the degenerated 2 orbitals5. Figure 1.3. Molecular orbital diagram of CO. 1s orbitals of carbon and oxygen called 1 and 2 are not shown. After, CO interacts with metal ions by three possible pathway; simple electrostatic interaction of the CO molecule with the field of the cation, by formation of bonds Experimental 10 with CO as electron donator and finally, by formation of bonds with CO as electron acceptor6. Electrostatic interaction of the charge distribution of the CO molecule and the field of the cation effects are connected with stabilization of all orbitals of CO in the positive electrostatic potential generated by the cation. It contributes with a positive charge to the interaction energy which is also compensated by the Pauli repulsion between the electrons orbitals of the cation and CO coverage7. However, according with the Frontier orbitals approach, the most important interaction in the CO coordination to cationic centers occurs between their HOMO and LUMO when the symmetry of the corresponding orbitals fits. For that, first take place the interaction between the HOMO of CO (anti-bonding orbital 5 ) and LUMO of the metal cation with a suitable spatial orientation. As a consequence, a new bonding and anti-bonding molecular orbitals are formed8,9. Then, take place the second part of the frontier orbital interaction. It is based in the interaction between the HOMO of the metal cation (dxz and dxy orbitals) and the LUMO of the CO ligand molecule (anti-bonding 2 orbitals). In this case, a new bond is formed leading a transfer of electron density from the metal cation to the ligand which is known as -backdonation transfer (figure 1.4). The strength of the new bond depends of the site electron density, since electron density increases, the bond strength increases. Therefore, an important synergistic effect of and bonds is observed. The bonds formation involve an electron enrichment of the metal ion which means stronger bonds formation while at the same time, formation of bonds transfer electron density from the metal cation to the CO molecule meaning stronger bonds. Experimental 11 Figure 1.4. Schematic representation of the orbital interactions in the formation of and bonds between CO ligands and the metal cation. Strong shifts of the absorbed carbonyl on the metallic site are observed due to the different nature of the bonds. This make possible the study of the different properties (coordination site geometry, oxidation state, acidity,…)10. In situ FTIR system The typical in situ FTIR system (figure 1.5) is composed by almost three big compounds; a glass vaccum system, an IR cell and the spectrometer. Figure 1.5. In situ FTIR system. Experimental 12 Glass system is connecting to two vacuum pumps which keep the system under pressure from 10-3 mbar to ultra-high vacuum around 10-7 mbar. Probe molecules as well as different gases for the conditioning treatment are introduced in the system by two SVT valves. Desired gaseous amount is controlled by two vacuum gauges with different measurement ranges. Gases evacuation is carry out by the vacuum system through liquid-nitrogen traps. Also IR cell can be separated in three compounds (figure 1.6); bottom part, top part and the sample holder. KBr windows, oven and thermocouple are located in the quartz bottom part which joining with the top part, based in glass squared pipe, by viton o-ring. Here, the used small volume in the probe molecule adsorption by small doses is located. The controlled volume is split and dosed by SVT valves. Finally, spectrometer is a Nicolet 6700 Fourier Transform IR spectrometer equipped by DTG detector (Deuter Triglyceride sulphate). Spectra were collected by 128 scans with 4 cm-1 of resolution. Figure 1.6. IR cell scheme. Experimental 13 REFERENCES (1) Niemantsverdriet, J. W. Spectroscopy in Catalysis; 2007. (2) Salmeron, M.; Schlögl, R. Surf. Sci. Rep. 2008, 63, 169–199. (3) Huck-Iriart, C.; Soler, L.; Casanovas, A.; Marini, C.; Prat, J.; Llorca, J.; Escudero, C. ACS Catal. 2018, 8 (10), 9625–9636. (4) Pérez-Dieste, V.; Aballe, L.; Ferrer, S.; Nicolàs, J.; Escudero, C.; Milán, A.; Pellegrin, E. In Journal of Physics: Conference Series; 2013; Vol. 425. (5) West, A. R. Solid State Chemistry and its applications. (6) Hadjiivanov, K. I.; Vayssilov, G. N. Adv. Catal. 2002, 47, 307–511. (7) Pacchioni, G.; Cogliandro, G.; Bagus, P. S. Int. J. Quantum Chem. 1992, 42, 1115–1139. (8) Hadjiivanov, K. I.; Vayssilov, G. N. Adv. Catal. 2002, 47, 307–511. (9) Ferrari, A. M.; Neyman, K. M.; Rösch, N. J. Phys. Chem. B 1997, 101 (45), 9292–9298. (10) Dominguez-Garcia, E. Effet du support sur la morphologie et l’activité des catalyseurs d’hydrodésulfuration, Univesity of Normandy, 2017. Chapter 2. Synthesis of PtCu3 Nanoparticles: A New Surface Cleaning Method Estudio De Catalizadores Basados En Platino Para La Reacción De PROX: Influencia De Los Depósitos De Carbono Synthesis of PtCu3 nanoparticles: a new surface cleaning method 15 2.1 Introduction Nanoparticle size and shape strongly influences the physical and chemical properties of catalysts. For instance, in an early study Narayanan and El-Sayed analyze the influence of Pt and Pd monometallic nanoparticles shape on the catalytic decomposition of oxalate and in Suzuki cross-coupling reactions1.These authors related the observed catalytic differences with the population of step and corner sites of the nanoparticles and, therefore, the particle size also affects the catalytic activity on modifying the relative proportion of these sites. Lang et al.2 demonstrate using DFT analysis of the CO oxidation on Pd clusters that the differences in CO binding energies in Pd(111) single crystals and small Pd clusters determine the reaction paths and in consequence the catalytic activity. These size effects have been also observed for gold nanoparticles supported on carbon. Megias-Sayago et al.3 found a strong correlation between particle size and shape and the catalytic activity of the oxidation of glucose to gluconic acid. Rearrangements of the catalyst surface during reaction have been observed for bimetallic RhPd and PtPd nanoparticles during ethanol steam reforming. The surface composition is altered as a function of the gaseous environment4. In the case of monometallic particles, the different reactivity of step and corner sites may result in a shape modification and hence on the catalytic activity1. A free energy decrease upon adsorption of molecules is responsible for surface modifications upon exposure to different gaseous environments. The energy gained by forming copper-carbonyl clusters on the surface of Cu (111) drives the reconstruction process5. Metallic nanoparticles are usually synthesized from salts solutions containing the metal of interest. Roucoux et al.6 provided a comprehensive review on synthetic methods of Synthesis of PtCu3 nanoparticles: a new surface cleaning method 22 support and the extremely low specific surface (around 8 m2/g) may facility the morphological study by HRTEM. TEM images reveal a monomodal sharp size distribution of almost spherical nanoparticles with an average diameter of 3.0 nm (Figure 2.5). Figure 2.5. (a) HRTEM image of PtCu nanoparticles alloy. (b) Pt-Cu nanoparticles alloy size distribution. (c) and (d) STEM images PtCu3/ Al2O3 Figure 2.5c and 2.5d show a clear crystallographic structure and the exhibited spots observed in the Fourier transform images (not shown) of some of the singles PtCu3 nanoparticles were studied. The interplanar spacing of the observed nanoparticles is around 2.17 Å, corresponded to (111) planes of the PtCu alloy with FCC structure while the interplanar distances of Pt FCC and Cu FCC show spots at 2.27 and 2.09 Å. Synthesis of PtCu3 nanoparticles: a new surface cleaning method 23 2.2.2 A new surface cleaning method However, the obtained nanoparticles may have capping agents/carbonaceous residues and trace elements from the reducing agents on the nanoparticles surface. Figure 2.6. HAADF-STEM image of PtCu3 nanoparticles supported on Al2O3 and elemental map of the single element. High Angle Annular Drakfield Scanning Transmission Electron Microscopy (HAADFSTEM) mode allows the study of metallic nanoparticles supported in low-Z atoms (figure 2.6). At the same time, EDS analysis was carried out. The analysis reveals the homogenous platinum and copper distribution on the support as well as the presence of carbon deposits associated to the PtCu metallic phase. Therefore, a careful study of the presence of these overlayers and the procedures to remove them avoiding any alteration of the nanoparticles size and morphology was Synthesis of PtCu3 nanoparticles: a new surface cleaning method 24 carried out. First at all, a thermal desorption experiment allows to detect the presence of capping/solvent molecules on the nanoparticle surface. Mass spectrometric analysis of the evolved gases reveals a complex decomposition pattern for the remaining molecules at the surface. For the sake of simplicity just mass-to-charge 16, 18, 28, 41 and 44 ratios are plotted (figure 2.7).These m/z ratios are assigned to the evolution of methane, water, carbon monoxide, methyl isocyanide, a product of the reduction of PVP, and carbon dioxide. The measurements of the nanoparticles TPD was carried out using an Ar 50ml/min flow and heating to 800 (10 /min). 20 mg of sample were placed in a U-shaped quartz reactor. Figure 2.7. Temperature programmed desorption – mass spectroscopy of PtCu3 from the room temperature to 800ºC. Black line at 350ºC represents the beginning of platinum sintering temperature. Synthesis of PtCu3 nanoparticles: a new surface cleaning method 25 The decomposition of PVP is catalyzed by Pt that reduces its decomposition temperature18 starting at 350ºC; moreover, in the presence of Pt PVP degradation accounts just for ca. 69% weight loss while pure PVP decomposes almost completely (>95% weight loss). Although, the main mechanism for decomposition is assumed that is the depolymerization the formation of amorphous carbon that remains on the Pt surface has been reported by Borodko et al.19. The observed profiles with significant gas evolution 300 and 450ºC of CO and methylsocyanate accounts for PVP decomposition; the low temperature evolution of H2O and CO2 may be indicative of the presence of ethylene glycol molecules on the nanoparticles surface. Gas evolution is observed at temperatures as high as 800ºC indicating that organic fragments remain on the surface and not only amorphous carbon. Thermal treatments at these temperatures should result in nanoparticle sintering and must be discarded. Low temperature air calcination of the nanoparticles was attempted to get nanoparticles free of carbonaceous residues. Figure 2.8 shows the PtCu3 nanoparticles in situ XRD data obtained during an oxidizing treatment in an O2-rich atmosphere from RT to 350ºC. At 220ºC is clearly evident from the figure a copper segregation process obtaining oxidized species of copper (correspond to monoclinic system crystalline and C2/C symmetry) and PtxCu3-xO species (tetragonal structure)20. At 350ºC (activation process temperature)21, PtCu alloy is completely segregate to metallic platinum species and CuO. The nanoparticles size increase until 20 nm. N Synthesis of PtCu3 nanoparticles: a new surface cleaning method 26 Figure 2.8. In situ-XRD data of Pt1Cu3 nanoparticles obtained under O2- rich atmosphere. It is clear evident from the figure the totally segregation of the metals to metallic platinum, oxidized copper species and a new PtxCu3-xO solid solution20 formed by several compounds member (from x=0.135 to x=0.355). The nature of the segregated phases was analyzed by peak deconvolution. The analysis reveals that the obtained metallic platinum phase is highly crystalline and also, a remarkable growth of the particle size, especially of the oxidized copper species (20 nm). To avoid these probable cases and for removing the rest of PVP from the synthesis process, an alternative cleansing step based in the PVP ligands exchange for others kind of ligands, which could be removed at low temperature, was designed. In this cleansing step consisted in wash process of the nanoparticles with a mixture of acetic acid, water and acetone. This step takes place before the conditioning process. The follow up process was measured by UV-Vis spectroscopy of the supernatant seeing how the absorption bands changed (figure 2.9). The experimental absorption band shows a maximum at 242 nm assigned to the π→ π* electronic transition of C=O and C=C groups of the PVP molecules around at the metallic site. Furthermore, it is Synthesis of PtCu3 nanoparticles: a new surface cleaning method 27 appreciable a shoulder at 300 nm assigned to charge-transfer transitions of excited electrons from carbonyls groups to vacant d orbitals of the bimetallic nanoparticles19. The main band at 242 nm will suffer changes while PVP molecules are being removed with successive centrifuge processes (cycles) coming from the surface of the nanoparticles. The more processes that take place, the band will be closer to a value of 234 nm which corresponds to the Acid/H2O mixture22 since most of the PVP that could be left on the nanoparticles would have been washed. As seen in figure 2.9, the more centrifuge processes that the nanoparticles suffer, the band from the supernatant washed liquid acquire a shape closer to that of the cleansing mixture. Figure 2.9. UV-Vis spectra of PtCu3 nanoparticles cleaning liquid after the different purification step; acetic acid/ water (black), PVP (yellow), supernatant after first purification (red), after second purification (blue) and after third purification (green). This could be due to the fact that the cleansing was successful. In order to prove this, the TPD-SM was repeated on the now clean hydrogenated nanoparticles (not shown). The m/z value of 41 (which corresponds to the product of the reduced PVP) is no longer present. As a result of this ligand exchange and considering the results of the second TPDS-SM carried out, acetates may be present in the surface of the nanoparticles. Synthesis of PtCu3 nanoparticles: a new surface cleaning method 28 The analysis of the nanoparticles surface by a specific and very sensitive technique as APXPS confirms a remarkable carbon signal. The carbonaceous layer that covers the nanoparticles is clearly evident. Therefore, for obtaining a completely clean nanoparticles surface, different in-situ treatments, after ligand-exchange step, were studied by APXPS. So that, the proposed surface cleaning method is shown in figure 2.10. Figure 2.10. Schematic nanoparticles cleansing step. The APXPS experiments were carried out at CIRCE beamline at ALBA synchrotron light source. The spectra were obtained at X-ray photon energy of 450 eV for C (1s) and the inelastic mean free paths (IMFP) account for the chosen photon energy is 0.45 nm in order to detect just the deposited carbon on the surface catalyst. Spectra were referred to the Fermi Level (see associated content 2.5). The Cu (2p) and Pt (4d) APXPS spectra were acquired at the same kinetic energy. The proposed treatments were: 750 mTorr H2 atmosphere, 750 mTorr H2 and H2O (50:50), 750 mTorrCO2, 750 mTorrH2 (after CO2 treatment), 750 mTorr CO2 and H2 (40:60). The sample holder temperature was 350ºC. Figure 2.11 shows the C (1s) signal evolution as a function of the atmosphere exposure and it is appreciable a substantial oscillation of the surface carbon amount as the main flow changes. Only under CO2 atmosphere exposure, the C (1s) signal disappear Synthesis of PtCu3 nanoparticles: a new surface cleaning method 29 meaning the absolutely remove of the surface carbon, resulting any addition to the main carbon dioxide atmosphere an increase in the amount of deposited carbon. Figure 2.11. a) Relative surface amount of carbon as a function of the activation treatment. b) Cu (2p) core-level APXPS spectra under different in situ treatments; vacuum, 750 mTorr H2, 750 mTorr H2+H2O, 750 mTorr CO2, 750 mTorr H2 (after CO2 treatment), 750 mTorr CO2 and H2 (40:60). c) Pt (4d) corelevel APXPS spectra under different in situ treatments. Incident photon energy is 1100 eV for Cu (2p) and 485 eV for Pt (4d). 2.3 Influence of the surface cleaning method in the electronic properties of the surface On the other hand, the new in-situ treatment for removing the surface carbon over layer cannot be affecting the nature of the active site. The not modification of Cu and Pt species after the CO2 treatment was confirmed by the Cu (2p), Pt (4d) and Pt (4f) APXPS spectra study. Synthesis of PtCu3 nanoparticles: a new surface cleaning method 30 Satellite peaks characteristic of oxidized species of copper are not present during any treatments (figure 2.11.b) and Cu (2p3/2) peak binding energy corresponds to metallic copper species in all the cases. We can observe that Cu (2p3/2) peak of the fresh sample at 932.3 eV and when the sample is exposed under H2 atmosphere, the peak was shifted to 931.1 eV. Later, due to the oxidizing character of water showed by Saveleva et al.23 in an electrochemical study by APXPS, Cu (2p3/2) peak is shifted to 932.3 eV although the satellite peak is still not present. Under the CO2 environment, the main peak of Cu (2p) is shifted to 931.1 eV. The oxidation state is the same but there is a clear electronic structure modification clearly influence by the change of extra-atomic contribution (change of neighbor atoms) in the final state relaxation energy24. It is reflected in a change of the metallic surface composition as a function of the exposure environment. Figure 2.12 shows the copper/platinum surface atomic ratio as a function of the Cu (2p3/2) peak binding energy and you can observe a change of the stoichiometry on the nanoparticles surface from Cu1Pt1 under the mixture of H2 and H20 atmosphere to Cu6Pt1under CO2 atmosphere. There is a clear relationship between the surface composition of the nanoparticles and the binding energy of the Cu (2p) photoelectrons from the surface (IMFP = 4.5 Å); higher Cu/Pt ratio means lower Cu (2p) binding energies. This trend is totally related with the conclusions of Komatsu et al.25. These authors carried out a comparative study between platinum nanoparticles and Pt1Cu1 alloy nanoparticles support on alumina as active site for the preferential oxidation of CO. They observed by the CO adsorption in both catalysts by FT-IR, a change in the frequency of the adsorbed CO in the metallic site. When CO adsorption was carried out in the Pt1Cu1 surface catalyst, they observed a decrement of the frequency refereed to the Pt catalyst from 2049 cm-1 to 2009 cm-1. They related this fact with an electron transfer from copper atoms to platinum sites in the Pt1Cu1 which supposes a decrement Synthesis of PtCu3 nanoparticles: a new surface cleaning method 31 of the electron density of the copper atoms. In this way, higher surface platinum concentration involves higher binding energies for Cu (2p) photoelectrons from the surface of the nanoparticles. Figure 2.12. Copper/platinum surface atomic ratio as a function of the Cu (2p3/2) peak binding energy. Photon energy of 1100 eV for Cu 2p was used to produce photoelectrons with kinetic energy around 170 eV. The IMFP account the chosen photon energy is 0.45 nm. Line green represents the linear fitting. After pumping the CO2 from the analysis chamber, nanoparticles were exposed under H2 flow and Cu (2p3/2) peak is not affected. Finally, the addition of CO2 to the H2 flow supposed a soft shift to higher energy of Cu 2p peak of 0.2 eV to 931.3 eV. Comparing the Pt (4d) peaks during the different treatments (figure 2.11.c), Pt (4d) signal is not appreciable until the surface carbon overlays removal with the CO2 treatment, improving thus the platinum availability as active site. Any addiction to the 931,0 931,5 932,0 932,5 0 1 2 3 4 5 6 7 H2+H2O Fresh H2+CO2 CO2 H2 Cu / Pt (molar ratio) Cu (2p3/2) Binding Energy / eV H2 (After CO2) 0 1 2 3 4 5 6 7 Synthesis of PtCu3 nanoparticles: a new surface cleaning method 38 same way, when CO2 is added to the main H2 flow, CO2 is also dissociated by hydrogen meaning the increment of the surface carbon amount and showing the characteristic peaks of CO2 (292.7 eV and 536.6 eV) and CO (291.0 eV and 537.7 eV) gas phases. The CO gas formed by CO2 dissociation is adsorbed in the metallic sites and the characteristics peaks of the adsorbed CO in the metallic sites at 286.1 eV and 286.9 eV as well as the peak assigned to adsorbed CO in electron deficient copper sites (533.4 eV and 287.9 eV) are shown again. Additionally, figure 2.16 shows a HRTEM image of PtCu3 nanoparticles alloy after H2- rich atmosphere exposure confirming the presence of the carbonaceous layer that cover all the catalyst (not detectable for conventional techniques) while the image of PtCu3 nanoparticles alloy treated with CO2 present a free-carbon surface. Figure 2.15. HRTEM images comparison: a) PtCu3 nanoparticles alloy after H2 treatment, b) PtCu3 nanoparticles alloy after CO2 treatment. 2.4 Conclusions Well-defined PtCu bimetallic alloy nanoparticles (1:3 molar ratio) were synthetized and characterized as a model catalyst for the preferential oxidation of CO. The presence of Synthesis of PtCu3 nanoparticles: a new surface cleaning method 39 surface carbon deposit from the use of organic capping agents has been evidenced. Previously, the effect of these typical carbon deposits in the catalytic activity was reported. For this reason, an alternative surface cleaning method has been developed which highlighted the soft conditions and the non-effect in the nature of the active site. Surface dynamic during the proposed pretreatment was studied by ambient pressure XPS. Changes in the surface restructuring were observed in response of the exposed environments. The surface carbon and oxygen species during the different conditioning treatment were detailed described for the understanding of the surface driven restructuring of bimetallic PtCu3 nanoparticles. 2.5 Associated content X-ray diffraction (XRD): To get information on the structure and composition of the obtained nanoparticles, X- ray diffraction (XRD) was employed. This analysis was performed on a X’Pert Pro PANalytical diffractometer. Diffraction patterns were recorded using Cu Kα radiation (40 mA, 45 kV) over 2θ range of 5 - 90 and a position-sensitive detector using a step size 0,05 and a step time of 1200s. Transmission electron microscopy (TEM): The nanoparticles morphology and size were studied by transmission electronic microscopy (TEM) on a 200kV JEOL JEM-2010F instrument with a structural resolution of 0,19 nm at Scherzer defocus conditions. To avoid the carbon contribution, samples were deposited directly onto Ni grids. Ultraviolet-visible spectroscopy (UV-VIS): Synthesis of PtCu3 nanoparticles: a new surface cleaning method 40 The analysis of the supernatant washed liquid were carried out in an Avantes AvaLight- DH-S-BAL spectrophotometer equipped with an optic fiber liquid sensor for wavelengths from 100 to 1000 nm. Ambient pressure X-ray photoelectron spectroscopy (AP-XPS) All ambient pressure X-Ray photoelectron spectroscopy (AP-XPS) experiments have been performed at the AP-XPS instrument installed at CIRCE beamline of the ALBA Synchrotron light source (Barcelona, Spain). AP-XPS measurements were carried out using a PHOIBOS 150 NAP Specs analyzer equipped with four differentially pumped stages. The beam spot size at sample is 100 x 20 μm2. AP-XPS data were acquired with 20 eV of pass energy and 0.1 eV of kinetic energy step. The work pressure used was 5.0 x 1 mbar (UHV) and 750 mTorr for the treatment atmospheres. The kinetic energy for all the measured photoelectron was around 170 eV and for taking account only the surface carbon species, the photon energy used for C (1s) was 450eV. For this energy, the C (1s) electrons IMFP (Inelastic mean free path) is 0.79 nm. In addition, the employed photon energies for the Cu (2p), O (1s), Pt (4d) and Pt (4f) were 1100 eV, 700 eV, 485 eV and 300 eV respectively. For the calculation of IMFP of this elements, we employed the TPP-2M formula proposed by Tanuma et al.35 considering a closed PtCu3 alloy (density of 12.08 g/cm3 and 43 valence electrons per molecule). Consequently, the IMFPs for these photon energies is 0.45 nm. The atomic ratio of Pt, C and O were obtained considering the calibrated Pt (4f5/2), C (1s) and O (1s) peaks areas. Previously, areas were normalized by the incident photon flux. Atomic ratio was calculated by the coefficient of normalized area of Pt (4f5/2) by the total area of Pt (4f5/2) and C (1s) for platinum/carbon ratio and the normalized area Synthesis of PtCu3 nanoparticles: a new surface cleaning method 41 of Pt (4f5/2) divided by the sum of total area of Pt (4f5/2) and O (1s) for platinum/oxygen ratio. Table S1. APXPS binding energies of the gas matter. Sample work function could vary the values. H2O (g) CO2(g) CO (g) C (1s) - 292.7 291.0 O (1s) 535.5 536.6 537.7 * Binding Energy in eV Table S2. APXPS binding energies and FWHM in O (1s) region of the different found species. Atomic Oxygen Pure OH OH H2-bonding CO/Cu O Subsurface electrophili c atomic oxygen B.E. (eV) 529.9 530.4 530.9 531.5 531.5 531.5 FWHM (eV) 1.3 1.4 1.4 1.4 1.4 1.4 H2O CO/Pt CO/ B.E. (eV) 532.4 532.5 533.4 FWHM (eV) 1.4 1.4 1.4 Table S3. APXPS binding energies and FWHM in C (1s) region of the different found species. M-C C-C CHX CO/Cu CO/Pt CO/ Satellite B.E. (eV) 283.3 284.6 285.2 286.1 286.9 287.9 290.9 FWHM (eV) 1.2 1.2 1.2 1.2 1.3 1.3 2.5 Synthesis of PtCu3 nanoparticles: a new surface cleaning method 42 REFERENCES (1) Narayanan, R.; El-Sayed, M. A. J. Phys. Chem. B 2005, 109 (26), 12663–12676. (2) Lang, S. M.; Fleischer, I.; Bernhardt, T. M.; Barnett, R. N.; Landman, U. ACS Catal. 2015, 5 (4), 2275–2289. (3) Megías-Sayago, C.; Santos, J. L.; Ammari, F.; Chenouf, M.; Ivanova, S.; Centeno, M. A.; Odriozola, J. A. Catal. Today 2017, No. 2016, 1–8. (4) Divins, N. J.; Angurell, I.; Escudero, C.; Pérez-Dieste, V.; Llorca, J. Science (80-. ). 2014, 346, 620–623. (5) Eren, B.; Zherebetskyy, D.; Patera, L. L.; Wu, C. H.; Bluhm, H.; Africh, C.; Wang, L.-W.; Somorjai, G. A.; Salmeron, M. Science (80-. ). 2016, 351 (6272), 475–478. (6) Roucoux, A.; Schulz, J.; Patin, H. Chem. Rev. 2002, 102 (10), 3757–3778. (7) Baker, L. R.; Kennedy, G.; Krier, J. M.; Van Spronsen, M.; Onorato, R. M.; Somorjai, G. A. Catal. Letters 2012, 142 (11), 1286–1294. (8) Park, J. Y.; Aliaga, C.; Renzas, J. R.; Lee, H.; Somorjai, G. A. Catal. Letters 2009, 129 (1–2), 1–6. (9) Lopez-Sanchez, J. A.; Dimitratos, N.; Hammond, C.; Brett, G. L.; Kesavan, L.; White, S.; Miedziak, P.; Tiruvalam, R.; Jenkins, R. L.; Carley, A. F.; Knight, D.; Kiely, C. J.; Hutchings, G. J. Nat. Chem. 2011, 3 (7), 551–556. (10) Luo, M.; Hong, Y.; Yao, W.; Huang, C.; Xu, Q.; Wu, Q. J. Mater. Chem. A 2015, 3 (6), 2770–2775. (11) Romero-Sarria, F.; Garcia-Dali, S.; Palma, S.; Jimenez-Barrera, E. M.; Oliviero, L.; Bazin, P.; Odriozola, J. A. Surf. Sci. 2015, 648, 84–91. (12) Alayoglu, S.; Nilekar, A. U.; Mavrikakis, M.; Eichhorn, B. Nat. Mater. 2008, 7 (4), 333– 338. (13) Zhan, W.; Wang, J.; Wang, H.; Zhang, J.; Liu, X.; Zhang, P.; Chi, M.; Guo, Y.; Guo, Y.; Lu, G.; Sun, S.; Dai, S.; Zhu, H. J. Am. Chem. Soc. 2017, 139, 8846–8854. (14) Avgouropoulos, G.; Ioannides, T.; Papadopoulou, C. Catal. Today 2002, 75, 157–167. (15) Jain, S. K.; Crabb, E. M.; Smart, L. E.; Thompsett, D.; Steele, A. M. Appl. Catal. B Environ. 2009, 89, 349–355. (16) Taylor, E.; Chen, S.; Tao, J.; Wu, L.; Zhu, Y.; Chen, J. ChemSusChem 2013, 6 (10), 1863– 1867. (17) Neergat, M.; Rahul, R. J. Electrochem. Soc. 2012, 159 (7), 234–241. (18) Du, Y. K.; Yang, P.; Mou, Z. G.; Hua, N. P.; Jiang, L. J. Appl. Polym. Sci. 2006, 99 (1), 23– Synthesis of PtCu3 nanoparticles: a new surface cleaning method 43 26. (19) Borodko, Y.; Sook Lee, H.; Hoon Joo, S.; Zhang, Y.; Somorjai, G. J. Phys. Chem. C 2010, 114, 1117–1126. (20) Muller, O.; Roy, R. J. less-common Met. 1969, 19, 209–214. (21) Shore, L.; F.Ruettinger, W.; Farrauto, R. J. Platinum group metal promoted copper oxidation catalysts and methods for carbon monoxide remediation. US 6,913,739 B2, 2005. (22) Olivo, G.; Barbieri, A.; Dantignana, V.; Sessa, F.; Migliorati, V.; Monte, M.; Pascarelli, S.; Narayanan, T.; Lanzalunga, O.; Stefano, S. Di; Angelo, P. D. J. Phys. Chem. Lett. 2017, 8, 2958–2963. (23) Saveleva, V. A.; Papaefthimiou, V.; Daletou, M. K.; Doh, W. H.; Ulhaq-Bouillet, C.; Diebold, M.; Zafeiratos, S.; Savinova, E. R. J. Phys. Chem. C 2016, 120 (29), 15930– 15940. (24) Moretti, G. J. Electron Spectros. Relat. Phenomena 1998, 95, 95–144. (25) Komatsu, T.; Takasaki, M.; Ozawa, K.; Furukawa, S.; Muramatsu, A. J. Phys. Chem. C 2013, 117 (20), 10483–10491. (26) Nierhoff, A.; Conradsen, C.; McCarthy, D.; Johansson, T. P.; Knudsen, J.; Chorkendorff, I. Catal. Today 2015, 244, 130–135. (27) Andersson, K.; Ketteler, G.; Bluhm, H.; Yamamoto, S.; Ogasawara, H.; Pettersson, L. G. M.; Salmeron, M.; Nilsson, A. J Am Chem Soc 2008, 130, 2793–2797. (28) Andersson, K.; Ketteler, G.; Bluhm, H.; Yamamoto, S.; Ogasawara, H.; Pettersson, L. G. M.; Salmeron, M.; Nilsson, A. J. Phys. Chem. C 2007, 111, 14493–14499. (29) Eren, B.; Heine, C.; Bluhm, H.; Somorjai, G. A.; Salmeron, M. J. Am. Chem. Soc. 2015, 137 (34), 11186–11190. (30) Chung, J.; Aksoy, F.; Grass, M. E.; Kondoh, H.; Ross, P.; Liu, Z.; Simon, B. Surf. Sci. 2009, 603 (5), L35–L38. (31) Salmeron, M.; Schlögl, R. Surf. Sci. Rep. 2008, 63, 169–199. (32) Eren, B.; Kersell, H.; Weatherup, R. S.; Heine, C.; Crumlin, E. J.; Friend, C. M.; Salmeron, M. B. J. Phys. Chem. B 2018, 122 (2), 548–554. (33) Heine, C.; Lechner, B. A. J.; Bluhm, H.; Salmeron, M. J. Am. Chem. Soc. 2016, 138, 13246–13252. (34) Gay, T. J.; Brand, J. A.; Fritts, M. C.; Furst, J. E.; Khakoo, M. A.; Mell, E. R.; Sieger, M. T.; Wijayaratna, W. M. K. P. J. Vac. Sci. Technol. 1994, 12 (5), 2903–2910. (35) Tanurna, S.; Powell, C. J.; Penn, D. R. Surf. Interface Anal. 1991, 17, 927–939. Chapter 3. Structural Changes During the Conditioning Treatment Estudio De Catalizadores Basados En Platino Para La Reacción De PROX: Influencia De Los Depósitos De Carbono Structural changes during the conditioning treatment 45 3.1 Introduction Bimetallic nanoparticles development becomes one of the most important topics in the field of catalysis. The general understanding of these materials goes through the control of the nanoparticles composition, structure, size or shape. The possibility of a control tunable chemical composition or nanostructure and, therefore, the possibility of the control of nanoparticles catalytic properties opens an interesting way in the bimetallic nanomaterials design. For instance, platinum nanoparticles based catalysts are employed in several reactions for obtaining CO free hydrogen flow from renewable source1–5 and Komatsu et al.6 studied how the copper addition increases the catalytic activity and selectivity toward CO2 conversion at lower temperature in the preferential oxidation of CO and Liu et al.7 showed how, due to the miscibility and surface free energy of the nanostructure compounds, the deposition sequence variation or the change of temperature and/or surrounding atmosphere in the addition of the second metal in the bimetallic structure bring on the obtaining of different bimetallic nanostructure, such as core-shell nanoparticles or nanoalloys. Different nanostructure supposes different atomic coordination of the active site. Wanjala et al.8 studied the correlations between the atomic structure and the electrocatalytic behavior (mass activity and specific activity) of PtNiFe/C catalyst for oxygen reduction reaction (ORR). These authors showed by a detailed XPS and XAFS analysis the influence of the pre-treatment temperature in the atomic structure of the synthesized catalysts. A higher mass activity for the catalyst treated at low temperature was observed while the catalyst treated at high temperature showed higher specific activity. They related the different catalytic behavior to a change in the atomic structure Structural changes during the conditioning treatment 46 from soft core-shell structure with a Pt-rich surface to a well-defined Pt46Ni22Fe32 by the XPS analysis and the increment of the Pt-Ni(Fe) coordination numbers. Figure 3.1. Influence of the structural transitions in the electrocatalytic behavior of PtNiFe/C catalyst for oxygen reduction reaction. In the same way, Zhang et al.9 describe findings for optimizing the nanoparticles based catalysts. These authors showed the enhance of the catalytic activity for ORR by the tuning nanoparticles structure of core-shell FePt/Pt catalyst via core structural transition from face centered cubic (FCC) structure to face centered tetragonal (FCT) structure. For the other hand, it is well-known that there are non-controller phenomena in nanoalloy system such the surface segregation. Wang et al.10 explained this phenomenon in PdxNi100-x alloy nanoparticles and they observed a preferentially migrations to the surface of Pd, by the strain energy effect, or Ni in synthesized alloy with low values of x due to the reduction potential effect. It supposed the loss of the Structural changes during the conditioning treatment 47 initial nanoalloy nature. However, Tao et al.11,12 opened a new way in the field of nanostructure engineering of nanoparticles with the controller “surface reaction-driven restructuring phenomenon”. They exposed under different in situ oxidizing or reducing surrounding environment to RhPd and PtPd core-shell nanoparticles and they studied by ambient pressure XPS (APXPS) the reversible changes in composition and oxidation state in response of the environment exposed. Figure 3.2. Reversible changes un composition and oxidation state in response of the environment exposure of RhPd nanoparticles. Inelastic mean free path of 16 Å. Obviously, the restructuring phenomenon supposes a change in lattice spacing of the nanoparticles. By aberration-corrected environmental TEM, Xin et al.13 studied the atomic structure of a single nanoparticle on 11 nm Pt0.5Co0.5 alloy nanoparticles under in situ O2-rich and H2-rich environment. Following the lattice spacing variation under oxidizing environment by HRTEM, these authors observed the Co segregation from the single nanoparticle forming cubic CoO islands on the surface alloy. This fact means a lattice spacing oscillation in the atomic layers near to the single nanoparticle surface and the main conclusion of the work is the segregation process reversibility since Structural changes during the conditioning treatment 54 3.2.3 Oxidation state dynamics The existence of different oxidation states and the surface dynamics during the conditioning process could be then elucidated by operando synchrotron-based X-ray Absorption spectroscopy (operando XAS) study, which would be crucial to observe if the surface oxidation states and/or the surface dynamics differ from the bulk one. For that, XAS spectra of the samples were recorded at Pt L3-edge and Cu K-edge. Considering the final state (contributions of the valence band and extra-atomic contributions)25 by multiple-scattering effects, Cu K-edge XANES spectra, based in transitions from Cu 1s core state to p symmetry bound states, were recorded showing the features for elucidate the Cu mean oxidation state and the local geometry. Figure 3.6 shows normalized Cu K-edge XANES spectra recorded upon the conditioning process and it is clear evident the energy of the Cu foil K-edge attributed to 1s to 4p transitions at 8981.3 eV (gray dash line) as well as the 1s to 4pxy transitions at 8983.2 eV attributed to Cu+1 species (red dash line) and the 1s to 4pz transitions at 8985.6 eV attributed to Cu+2 species (yellow dash line)26. The presence of Cu2+ species is not taken account because the edge for oxidized species appear at higher energies and the characteristic pre-edge peak of this kind of species at 8969 eV, attributed to 3d orbitals partially occupied (dxy) which induce 1s→3d transition , is not present27. Taking account the edge energy and the line shape, the analysis of the Cu K-edge spectra of the sample before and after conditioning process is in agreement with the presence of metallic copper species. A soft chemical shift of 0.3 eV to higher energies (refereed to the Cu foil K-edge) was found in the energy of the K-edge of the sample Structural changes during the conditioning treatment 55 before and after the treatment. Therefore, in both cases the line shape of the Cu K-edge spectra corresponds with the line shape of metallic Cu K-edge. Higher intensity of the Cu K-edge after the conditioning treatment means an increment of the density of state and therefore, an increment of the electronic density. Figure 3.6. a) Cu k-edge XANES spectra of copper standards and PtCu3 nanoparticles alloy before (blue line) and after (green line) the conditioning process at 350ºC for 2 hours. The inset shows a magnification of the 1s→3d transition region. 4b) k2-weighted FT of the EXAFS spectra. For the other hand, Choy et al.26 proposed that the features of the sample after the conditioning process at 8993.3 eV correspond to the transition │1s13d94 ⟩ and the above shoulder at 8998 eV is attributed to transitions│1s13d94 ⟩. These authors studied the Cu K-edge XAS spectra for well-defined copper containing oxides relating the shift in the features attribute to non-degenerated orbitals with changes in the Structural changes during the conditioning treatment 56 local symmetry, according to the crystal-field theory, which suppose an electronic density oscillation. In this way, copper atoms of the FCC structure were considerate as a centrosymmetric environment with the local symmetry of regular octahedron in the second shell coordination (dCu-Cu 2nd shell = lattice parameter). Choy et al. explained that this energy shift of 2.8 eV to lower energies of the feature at 9002.9 eV of metallic copper standard and the nanoparticles feature at 8999.0 eV is due to a change in the local symmetry from regular octahedron for metallic copper standard to tetragonally distorted nearly octahedron for the PtCu3 nanoparticles alloy. Figure 3.7 Involved transitions in Cu k-edge XANES spectra. However, it is necessary considerer that an octahedral complex used in crystal-field theory is modeled as six partial negative charge of an electric dipole that interact strongly with the d orbitals of the central metal ion and, in this cases, all the copper atoms of the regular octahedron have the same electric dipole and the charges interaction is almost zero. Thus, referred to the metallic copper standard, it is clear Structural changes during the conditioning treatment 57 evident that there is a change in the energy of the feature attributed to transitions│1s13d94 ⟩ and this shift means a change in local geometry of the copper atoms by the presence of platinum in the alloy. The Pt L3-edge XANES spectra before and after the conditioning treatment as well as the theoretical transition are shown in figure 3.8. T. Anniyev et at.28 studied the platinum partial d-density of states in PtCu nanoparticles and these authors explained that Pt L3-edge probes the dipole allowed transitions (2p3/2→5d3/2,5/2) and the most intense spectral feature, which represents the unoccupied d-density states (white line), is located at 11566 eV. Only 5d5/2 core level has empty states and therefore, complying J selection rules for the unoccupied d states ( = -1, 0, +1), a main transition 2p3/2→5d5/2 is expected. In this way, spectra before and after conditioning treatment show the same line shape and the same energy of the white-line correspond to metallic platinum species at 11566 eV in agreement with the last referred work. The integrated area under the Pt-L3 XANES spectrum is proportional to the d-states population29 and therefore, there is a decrement of d-states localized on PtCu3 nanoparticles after the conditioning treatment by the Pt d-states hybridization with the valence orbitals of copper (Cu 3d orbitals). This means a transfer of d-holes in the electronic density transmission in the alloy during the conditioning treatment. Structural changes during the conditioning treatment 58 Figure 3.8. Normalized Pt L3-edge XANES spectra of the fresh nanoparticles (blue line) and spectra of nanoparticles after the conditioning treatment (red line). Dashed line show the standard of metallic platinum (grey line) and oxidized species of platinum (green line and yellow line). Schematic transition from p core states to 5d states is represented in the right side of the figure 3.2.4 Local geometry study In order to identify with relative accuracy the structural changes and the local coordination environment of copper and platinum atoms, EXAFS functions analysis of Cu k-edge and Pt L3-edge was performed in ALBA Synchrotron Light Source during the in situ simulation of conditioning treatment. A general vision of the different atomic shells surrounding to the copper and platinum atoms is observed in the evolution of the Fourier transform of the k3 weighted Cu kedge and Pt L3-edge EXAFS data during the simulation of the conditioning treatment (figure 3.9). Structural changes during the conditioning treatment 59 Figure 3.9. Fourier-transformed of the Cu k-edge EXAFS data (k3-weighted) measured at room temperature from the PtCu3 nanoparticles before the conditioning treatment (red line). Cu k-edge EXAFS data measured under conditioning treatment environment at (a) 30ºC, (b) 150ºC, (c) 250ºC and (d) 350ºC. The fitting of the EXAFS data was carried out using the multiple-scattering FEFF6 theory30 applied to a FCC structure, with the average parameters modulation of coordination number (N), normalized distance (R) and the Debye-Waller factor ( 2, DWF) for each path31. FEFFs provided by the crystallography structure of well-defined PtCu alloy 1:3 molar ratio (JCPDS-00-035-1358). Structural changes during the conditioning treatment 60 . Figure 3.10. Cu k-edge EXAFS function measured from nanoparticles before the conditioning treatment at room temperature (a) and from nanoparticles after the conditioning treatment at 350ºC (b). Figure 3.11. Pt L3-edge EXAFS function measured from nanoparticles before the conditioning treatment at room temperature (a) and from nanoparticles after the conditioning treatment at 350ºC (b). Structural changes during the conditioning treatment 61 Considering the proposed structures models of Pt34Cu66 nanoalloy for the fresh sample and surface Cu-rich Pt27Cu73 nanoalloy with generally FCC structures for the sample after the conditioning treatment, the proposed fitting of Cu k-edge R-space data is based in the theoretical contribution of four different paths consistent with a close PtCu3 alloy. It means three different neighbor distances; first coordination shell of Cu-Cu and Cu-Pt neighbors around 2.6107 Å, second coordination shell Cu-Pt distance around 3.6921 Å, and third coordination shell Cu-Pt and Cu-Cu distance around 4.5219 Å. A value of the amplitude reduction factor was derived from EXAFS data analysis and it was fixed to 0.92 (within the theoretical physical values32). A range of E0min= 2 eV and E0max= 15 eV for the edge energy variation ( E0) was employed. Also, the corrections r for the model bond lengths was varied independently for the neighbors of copper and platinum calculations showing an increment of the values after the conditioning treatment. EXAFS fitting parameters are shown in table 1. Table 1. EXAFS fitting parameters for Pt L3-edge for the spectra collected before and after conditioning treatment. Cu-Pt 1st Shell (NCu-Pt =2.6) Cu-Cu 1st Shell (NCu-Cu =6.3) Cu-Cu 2nd Shell (NCu-Cu=) Cu-Pt 3rd Shell (NPt-Cu+ NCu-Cu =15) d Cu-Pt Cu-Pt d Cu-Cu Cu-Cu d Cu-Cu Cu-Cu d3rd Shell 3rd Shell Before 2.6215 0.0071(2) 2.6206 0.0177(8) 3.7020 0.1584(9) 4.5318 0.0995(2) After 2.6247 0.0108(2) 2.6107 0.0261(1) 3.6921 0.0331(7) 4.5219 0.0033(4) *distance units: Å Structural changes during the conditioning treatment 62 Analyzing the fitting parameters we can conclude different things. First of all, first-shell coordination number for Cu atoms suggest a nonstatistical distribution environment and the average coordination number of the absorbing Cu center is formed by 6.3 Cu neighbors atoms and 2.6 Pt neighbors atoms which agrees with the theoretical FEFF for an infinity FCC structural model with 12 atoms in the first-shell coordination30. Particularly, a change in the coordination number was not observed theoretical calculation the total coordination number. Total coordination number for first-shell was constrained following the theoretical calculation proposes by D.T. Tran et al. Who studied the total coordination number as a function of the crystal size. In addition, a change in NCu-Cu / NPt-Cu was observed. The DWFs for first-shell Cu-Pt and Cu-Cu paths reveal the similar behavior showing higher values for the conditioned sample measurement. The most significant fact is the oscillation of the distance for Pt and Cu neighbors. The fitting analysis concluded two different distances as a function of the neighbor nature. Both distances present different behavior. Initially, nanoparticles unit cell presents an elongation of the majority Cu-Cu bonds which are reduced to the theoretical values for well-defined PtCu (1:3 molar ratio) first-shell distance. The initial values of the Cu-Pt bonds were practically the same of initial Cu-Cu distance values ( < 0.001 Å) which agrees with the fundamental concept of a FCC structure; a=b=c. However, during the conditioning treatment, we observed an increment of the Cu-Pt distance. Identical results were obtained from Pt L3- edge EXAFS fuction fitting. In the employed theoretical structure for designing the fitting model, Cu atoms neighbors are located in the equatorial plane while Pt atoms neighbors are located in the axial plane. Therefore, the analysis concluded an equatorial Structural changes during the conditioning treatment 63 compression with the Cu-Cu bonds distance decrement and an axial elongation due to the Cu-Pt distance increment. Figure 3.10. Schematic illustration of distance modulation on first-shell coordination. Also, higher 2 values after the conditioning treatment (taking account the concomitant nonzero residual elastic strain energy) confirm the atoms rearrangements from close FCC structure with a Pt34Cu66 composition towards the tetragonal distortion FCC structure with a Pt27Cu73 composition and it is conclusive for understanding the atomic dynamic in the order-disorder transition33–35. The higher-shell structures do not support big information due to the signal quality. The low coordination number of the second and third shells is directly related to the finite size of the nanoalloy and the quality of the signal32,36. Chapter 4. Surface Characterization: Probe Molecule Adsorption Estudio De Catalizadores Basados En Platino Para La Reacción De PROX: Influencia De Los Depósitos De Carbono Surface Characterization: Probe Molecule Adsorption 71 4.1 Introduction The knowledge of the catalyst surface behavior is essential for the total comprehension of the materials. surface may be considerate such top-layers of atoms with a large number of crystal defects as edges, terraces, steps or corners totally controlled by the solid state properties. In this sense, Fourier transform infrared spectroscopy (FTIR) reveals fundamental aspect about the chemical nature of the surface or absorbed species as well as structural and strength bonds information1–3. Probe molecule adsorption by FTIR is one of the most developed surface characterization methods based in the study of the adsorption complexes formed by the contact between a react molecule and the surface catalyst through which it is possible to study step by step the possible pathways in the mechanism reactions as well as different properties of the catalyst such the basicity (proton-accepting) or acidity of the active site, the kind of site, the adsorption mode or the surface concentration. There are several probe molecules (carbon monoxide, carbon dioxide, nitrogen oxide, pyrrole, alcohols among many others) and the choice of one of them depends of what information we want to know as well as the site accessibility or the spectral response. Trying to understand the mechanism pathway in the methanol oxidation reaction, G. Busca et al.4 carried out the adsorption of formaldehyde such model intermediate on different oxides. They identified and characterized different adsorbed species from the initial formaldehyde as a function of temperature. Physisorbed HCHO species on surface OH groups were found at low temperature while at room temperature, dioxymethylene species were always observed. Higher temperatures involved the disproportionation of dioxymethylene into formate and methoxy groups opening the direct oxidation way on very strongly oxidizing surfaces by the Cannizzaro-type mechanism. Surface Characterization: Probe Molecule Adsorption 72 Another interesting work were carried out by E. Finocchio et al.5 who studied the mechanism of propane and propene oxidation on MgCr2O4 by the small doses-adsorption of different probe molecules based in C3, C2 and C1 oxygenates species. This particularly system is characterized because the solid is reduced by reaction giving rise oxidized organic species and CO2. These authors observed different oxidation pathways for both molecules where propene oxidation takes preferential place by previous activation at C1. Strongly adsorbed acrylates and acrolein species were found before of total combustion process by thermal effects. However, absorbed acetone species were observed as preferential oxidation product of propane. Going on the oxidation process, acetone gives formate and acetate species which burn at higher temperatures. The entire observed surface reactions, join to the published data for similar catalytic systems, allowed them to elucidate a general specific reaction pathway for alkane and alkene catalytic oxidation. For the other hand, acidity and basicity properties are totally related with the catalytic behavior of the solids. In this way, J.C. Lavalley6 reported the surface basicity in two of the most important materials in the petroleum chemistry, metal oxides and zeolites, by different probe molecules adsorption. Author concluded that the best probe molecule for the characterization of surface base centers sites would be an H-donating HX molecule elucidating the basicity character of the site by (H-X) shift. Avoiding steric limitations, A.M. Turek et al7 carried out the acidic properties study of different alumina-supported metal oxide catalysts by the CO2 adsorption as probe molecule relating the Brönsted acidity with a observed band at low frequency in the hydroxyl region (3460-3490 cm-1) which is due to water molecules H-bonded to the Brönsted sites. Also, different chemisorbed CO2 species were identified on the surface of alumina (figure 4.1). Surface Characterization: Probe Molecule Adsorption 73 Figure 4.1. Observed FTIR bands (cm-1) upon CO2 adsorption on pure metal oxides. However, one of the most used probe molecule is carbon monoxide which is the only known molecule which has preferential selectivity to the basic sites by carbonyls formation6. Due to the non-reactive character adsorption of the CO on the surface catalyst, CO probe molecule present excellent characteristic for the study of the oxidation and coordination states of the metal sites as well as of the stability and adsorption mode of the formed carbonyls complexes3. Thus, C. Binet et al8 studied the Lewis acid properties of polycrystalline ceria in oxidized and reduced states by the evaluation of the shift of gas phase (CO) value. These authors carried out CO adsorptions at low temperature for avoiding the surface carbonation and, refereed to the 2143 cm-1 value for the CO gas phase frequency, they observed higher shifts to higher wavenumber (+26 cm-1) in the CO adsorption on Ce3+ ions than for the Ce4+ Surface Characterization: Probe Molecule Adsorption 74 ion (+18 cm-1) which means that Lewis acidity of the cerium cations decreases in reducer states. Moreover, several studies have been reported about CO adsorption on PtCu based catalysts, object solid of this manuscript. However, before the study of the intermetallic compounds PtCu, it is necessary a total characterization of the monometal catalyst. In the cases of copper based catalyst, Cun+-CO surface carbonyls were located in a range from 2240 to 1916 cm-1. Carbonyls formed by Cu+ ions are the most stable species of Cun+-CO complexes9. It was characterized by A. Dandekar et al.10 in a wide range from 2160 to 2080 cm-1 during the CO adsorption on supported Cu catalyst, finding the carbonyl on supported Cu on Al2O3 around 2130 cm-1 . Conversely, Cu0-CO complexes formed by bond with the copper atoms and carbonyls of Cu+2 ions are unstable at room temperature being detectable by CO adsorption at low temperature (77 K)11. In addition, Lewis acid sites are present on alumina in tetrahedral or octahedral coordination9 showing typical bands assigned to CO on Al3+ at 2245-2215 cm-1 when samples are activated at higher temperatures(tetrahedral coordination)12, bands around 2200 cm-1 when CO adsorption is carry out under equilibrium conditions and bands around 2170 cm-1 when CO is absorbed on Al3+ cations in octahedral coordination13. CO adsoption on supported platinum catalyst has been studied by many authors due to the importance of platinum based catalyst in industry. C.W. Olsen and R.I. Masel14 carried out the CO adsorption on Pt (111) single crystal finding two different carbonyl adsorption mode on platinum sites; linear band around 2080 cm-1 and two-fold bridge bound CO at 1850 cm-1. A shift to higher wavenumber was observed by the CO coverage. For the other hand, CO adsorption may provide information about metal coordination, geometry or adsorption mode on the active site. H.J. Jänsch et al.15 studied the surface Surface Characterization: Probe Molecule Adsorption 75 migration of absorbed CO species on Pt(335) crystal by electronic excitation (electron stimulated migration). They distinguished two different adsorption modes; absorbed CO on step sites and Absorbed CO on terrace sites which their vibrational frequencies of CO stretch mode are clearly separated around 25 cm-1. Then, J. Xu and J.T. Yates Jr.16 assigned under low coverage conditions band around 2072 cm-1 to CO adsorbed on step sites. After, since coverage increased, a new band at 2100 cm-1 was observed due to the terminal-CO on the terrace sites. In the same way, P. Bazin et al.17 reported an exhaustive study about the CO adsorption on small particles of platinum on ceria support. Typical IR spectrum after the CO adsorption on supported platinum catalyst shows the linearly adsorbed CO band in a range from 2090 cm-1 to 2000 cm-1 and the band associated with CO bridge between 1860 cm-1 and 1760 cm-1 (figure 4.2a)18. However, after the CO adsorption on reducer Pt/CeO2 catalyst (figure 4.2b), they showed three different bands in the region from 2100 cm-1 to 2000 cm-1 and a new one band at 1937 cm-1, untypical region for the CO adsorption bands. They proved the nature of those bands by 12CO/13CO isotopic exchange, thermal analysis and CO adsorption at low temperature concluding; the band at highest wave number (2096 cm-1) correspond to the CO adsorption on Pt+ atoms, bands around 2054 cm-1 and 2010 cm-1 correspond with the linearly CO adsorption on platinum particles with different size while the new band at 1937 cm-1 appears for the bridged CO adsorption at the metal-support interface of Ce sites and small Pt particles (<1.5 nm). Surface Characterization: Probe Molecule Adsorption 76 Figure 4.2. FTIR spectra after CO adsorption at room temperature on non-reduced Pt/CeO2 catalyst (a) and H2- reduced Pt/CeO2 catalyst at 673K (b). Platinum atoms with different particle size also could be located in different environments. A.Y. Stakheev et al.19 studied the electronic state and localization of supported Pt atoms by CO adsorption. Those authors observed three bands at 2000, 2060 and 2075 cm-1 and they concluded that the frequency shift was due to the CO absorbed on platinum sites with different environment. They explained how CO absorption on more accessible platinum sites located near to the exposure surface involve different degree of metal-support interaction or different particles shapes and they related this kind of site with the band at 2075 cm-1. The other two bands at 2060 and 2000 cm-1 were assigned to metallic site with lower accessibility and also to CO adsorbed on high dispersed metallic platinum atoms Finally, the better catalytic behavior of intermetallic PtCu compounds than monometal catalyst has been demonstrated in the preferential oxidation of CO20. This fact and the unknown active site nature of one of the most useful commercial catalyst for mobile devices21 Surface Characterization: Probe Molecule Adsorption 77 does that several authors focus their work in the search of a suitable answer. For instance, T. Komatsu et al.22 showed the relationship of the metallic site electron density with the catalytic activity by the evaluation of the (CO) shift after the CO adsorptions on supported Pt and Cu catalyst as well as supported intermetallic compound with different atomic ratio (Pt3Cu1 and Pt1Cu3) (figure 4.3). Also, F. Romero-Sarria et al.23 showed by operando FTIR study and CO adsorption on PtCu/Al2O3 catalyst, the effect of organic deposit (from the synthesis process) in the electron density of the active site and therefore, the effect in the catalytic activity. Figure 4.3. FTIR spectra of adsorbed CO on alumina supported intermetallic compound PtCu as well as on both supported monometal catalyst. Taking account this background, in chapter 5 “Surface Characterization: Probe molecule adsorption”, an exhaustive study of the active site is carried out by the absorption of CO and NO probe molecules after the conditioning treatment on the intermetallic compound PtCu (1:3 and 3:1 molar ratio) as well as on the monometal catalysts. Oxidation state of the active site, adsorption modes, interaction of the nanoparticles with the support, surface species dynamic and other chemistry properties were described in the different sections. Surface Characterization: Probe Molecule Adsorption 78 The most important point is the not modification of the active site nature during the surface cleaning method. Also, it is necessary the complete understanding of the kind of the active site, the surface availability and the electron density site. The main goal of this chapter is the study of the active site by the probe molecule adsorption after different conditioning treatments. Specially, CO adsorptions were carried out after hydrogen treatments (environment based in the conventional conditioning treatment for PtCu catalyst applied to mobile devices) and after the alternative surface cleaning method based in CO2 environment on the PtCu intermetallic compound with different molar ratio (Pt3Cu1 and Pt1Cu3) as well as on the monometallic nanoparticles. 4.2 Surface characterization of Pt/ Al2O3 4.2.1 Pt/ Al2O3 First of all, unsupported nanoparticles are black so the transmittance is almost zero and the metal concentration is so high which supposes the saturation of the signal. For avoiding this problem, nanoparticles were supported on Al2O3 (226 m2/g) by a physic mixture minimizing the chemical interactions metal-support. Moreover, due to the observed surface composition modification (chapter 2) and the electron transfer during the conditioning treatment in the PtCu nanoparticles (chapter 3), it is necessary the study of the monometallic nanoparticles which should reveals the role of each compound. Thus, several CO adsorptions on Pt/ Al2O3 experiments were carried out after H2 treatment and CO2 treatment. Surface Characterization: Probe Molecule Adsorption 79 Figure 4.4. Background spectra (at r.t) of the fresh Pt/ Al2O3 sample (yellow color) and after CO2 treatment (red color) and H2 treatment (blue color). The insert shows a magnification of a found band around 2055cm-1 after conditioning treatments. The background spectra of the Pt/ Al2O3 are shown in Figure 4.4. The presence of residual carbonaceous species and adsorbed –OH and –CHx groups were observed on the fresh catalyst surface. Spectra after two proposed treatments showed similar surface groups. Specially, it draws attention the presence of –CHx groups in the region between 2900 cm-1 and 3000 cm-1 as well as the presence a characteristic band at 2055 cm-1 of absorbed carbonyls species on the platinum nanoparticles surface before start the CO adsorption24. It could be due to the carbons deposit decomposition during the simulated conditioning treatment in the in situ cell. This treatment is carried out in static conditions and employed environment only can be removed around 200°C for avoiding possible effects in the oxidation states. At this temperature, residual carbon species are reabsorbed on the platinum surface25. 4.2.2 CO adsorption on Pt/ Al2O3 Then, CO adsorptions after in situ treatments were carried out. Spectra acquired during the CO adsorption after H2 and CO2 treatment showed the same groups of bands. Therefore, both Surface Characterization: Probe Molecule Adsorption 86 Figure 4.9. Schematic simulation of the CO adsorption on Pt/ Al2O3 after H2 treatment. CO dissociation is avoided by carbon deposits. Black carbon atoms sham carbons deposits from the synthesis step and blue carbons atoms shams carbons deposits from the partial CO dissociation. During the CO adsorption after CO2 treatment, platinum is totally availability. CO is dissociated by Pt atoms and the increment of the CO concentration means an increment of CO2 gas phase band cause by CO oxidation. The small decrement of CO2 gas phase band intensity could be caused by the carbon deposits from the CO dissociation but due to the high surface reactivity (total availability of platinum active site) the CO2 gas phase is always present, including under CO saturation conditions. 4.2.3 Oxygen addiction on Pt/ Al2O3 For the other hand, the subsequent oxygen introduction in to the cell after the pre-treatment should reveal the possible carbon deposit presence by direct oxidation. Thus, oxygen small doses addiction were carried out on the non-conditioned Pt/ Al2O3 (fresh catalyst) and after H2 and CO2 conditioning treatments. The main goal of this study was the evaluation of the amount of surface carbon deposits in the three proposed situation and the effects of adsorbed molecular oxygen from the CO dissociation on the platinum oxidation state. First, the evolution of the CO2 bands was followed during the oxygen small doses addiction (figure 4.10). Since added oxygen amount increases, CO2 production increases by direct gasification of surface carbon deposit (Eq.(8)). (8) C* + O2 (g) → CO2 (g) + * In the case of non-conditioned catalyst and after H2 treatment, there is a remarkable surface carbon amount. Conditions during the oxygen addiction (exposure time, temperature, oxygen amount, etc.) are not enough for the total surface carbon removing by gasification process, so CO2 bands are present for all the experiment. In the addiction of oxygen small doses on Surface Characterization: Probe Molecule Adsorption 87 Pt/ Al2O3 after CO2 treatment, surface carbon amount is almost zero. Therefore, since added oxygen amount increases, the low surface carbon amount decreases by directly gasification and the CO2 bands intensities becoming zero. Relative to the metallic site, absorbed CO from the partial carbon gasification and CO2 dissociation was found during the oxygen addition after H2 and CO2 treatment. Figure 5.7.b shows bands at 2064, 2130 and 2174 cm-1 assigned to linearly absorbed CO on metallic platinum nanoparticles, CO absorbed on platinum atoms with low electron density ( atoms) and absorbed CO on (Al+3)Oct. Bands observed on the Pt/ Al2O3 catalyst after CO2 conditioning treatment present lower intensity due to the less surface carbon deposit presence. Therefore, effects on the oxidation state by the oxygen exposure at room temperature were not observed. Figure 4.10. a) Evolution of the CO2 band area during the small doses oxygen addiction on non-activated Pt/ Al2O3 (yellow color) and on Pt/ Al2O3 after H2 treatment (blue color) and CO2 treatment (red color). b) Absorbed CO after oxygen addiction on Pt/ Al2O3 pretreated under H2 atmosphere (blue color) and CO2 atmosphere (red color). Surface Characterization: Probe Molecule Adsorption 88 4.3 Surface characterization of Cu/ Al2O3 4.3.1 Cu/ Al2O3 Figure 4.11 shows the background spectra of Cu/ Al2O3 after H2 treatment and CO2 treatment. At lower wavenumber, initial bands at 1578, 1465 and 1264 cm-1 were observed. Davydov36 and B.A. Sexton37 assigned those bands by acetic adsorption on different oxides to the υas(COO), υs(COO) and υ(C-C) vibrational modes characteristic to absorbed acetates species. This species can be discriminate by bridge acetates species and bidentate acetate species. υ(COO) depends of the (OCO) angle being from 171 to 523 cm-1 for bridge acetate species and between 65 and 115 cm-1 for bidentate acetate species. υ(COO) for the found bands at 1578 cm-1 (υas(COO),) and 1465 cm-1 (υa(COO)) is 113 cm-1 so, bidentate acetate species are present on the surface of copper catalyst. Contrary to Pt/ Al2O3, CO adsorbed on the metallic sites after both different treatments was not found. Acetates species presence involves vibrational modes on the –CHx region. Different bands were observed at 2842 and 2914 cm-1 characteristic to the υs(CH3) and υas(CH3). Also, bands at 2869 and 2959 cm-1 correspond to υ(CH) vibrational modes. Surface Characterization: Probe Molecule Adsorption 89 Figure 4.11. Background spectra of Cu/ -Al2O3 after H2 conditioning treatment (blue color) and CO2 conditioning treatment (red color). Finally, vibrational modes correspond to –OH group were observed. It can be distinguished four different bands at 3580, 3678, 3730 and 3757 cm-1. Morterra et al.12,38 described the OH stretching spectra of alumina and they assigned band at 3757 cm-1 with terminal OH groups coordinated onto surface tetrahedral aluminium centers. Moreover, Dignet et al.28 carried out a simulation of surface OH stretching environment by DFT study and those authors described the coordination of above band at 3757 cm-1 such OH groups linked to single aluminium atom which is surrounded by six different oxygen atoms. Band at 3730 cm-1 was assigned by Morterra et al. 12,38 to OH groups bibridged across Al-Al pair while Dignet et al.28 calculations related this band with terminal OH groups bonded to single aluminium atom which is surrounded by five oxygen atoms. Contrary to Morterra et al.12,38, those authors and Busca et al.39 related the band at 3678 cm-1 to OH groups bridge between two aluminium atoms which are surrounded by five oxygen atoms. Last, band at 3580 cm-1 is related with OH group tribridged to aluminium atoms surrounded by six oxygen atoms. Surface Characterization: Probe Molecule Adsorption 90 4.3.2 CO adsorption on Cu/ Al2O3 Comparison of difference FTIR spectra after CO adsorption on Cu/ Al2O3 is showed in figure 4.12. Figure 4.12. Difference FTIR spectra after CO adsorption on Cu/ Al2O3 pretreated under H2 atmosphere (blue color) and CO2 atmosphere (red color). 0.26µmol of CO added in both cases. It is clear evidence a higher absorbed CO band intensity on the copper site as well as in the – OH region when catalyst is treated under CO2 environment. Figure 4.13 shows the difference FTIR spectra recorded during the CO adsorption on Cu/ Al2O3 after H2 and CO2 treatment. Interestingly in both cases, a main band can be observed at 2123 cm-1 with a shoulder to lower wavenumber at 2098 cm-1. In addition, a weak band to 2165 cm-1 can be observed. It means that the nature of pre-treatment does not affect the oxidation state of the copper particles. However, this band is shaper when sample is treated under H2 atmosphere. Different possibilities could be proposed for the origin of such of Cu sites. K.I. Hadjiivanov and G.N. Vayssilvov9 described frequencies of the surface carbonyls absorbed on the copper atoms as a Surface Characterization: Probe Molecule Adsorption 91 function of the copper oxidation state concluding that the characteristic band of linearly carbonyls absorbed on Cu2+ (Cu2+-CO) is located between 2220 and 2150 cm-1, band of carbonyls absorbed on Cu+ (Cu+-CO) is located between 2160 and 2080 cm-1 while carbonyls absorbed on metallic copper (Cu0-CO) show a weak band lower than 2130 cm-1. Figure 4.13. Difference FTIR spectra during the CO adsorption on Cu/Al2O3 after CO2 treatment (a) and H2 treatment (b). For the other hand, Zecchina et al.40 carried out a CO adsorption on silica supported copper catalyst after progressive reduction step under H2 environment at different temperatures. When Cu/SiO2 was reduced at 400ºC, they observed two bands; a weak band at 2163 cm-1 and a sharper band at 2136 cm-1. They concluded that both bands are due to CO absorbed on small Cu+ atoms on extended faces and edges and steps respectively. Also, Hadjiivanov et al.9 proposed the band at 2165 cm-1 due to carbonyls absorbed on Cu2+ species. Moreover, when reduction temperature was increased during the CO adsorption on Cu/SiO2, a new shoulder around 2100 cm-1 was observed and they assigned this band to CO absorbed on metallic copper or CO absorbed on a Cu-rich non stoichiometric Cu2O phase. Thus, the CO absorption on Cu2+ species, carbonyls absorbed on small particles of Cu+ (probably located in the metalsupport interface) and CO absorption on metallic copper could be assigned to the found band Surface Characterization: Probe Molecule Adsorption 92 at 2165, 2122 and 2098 cm-1 in agreement with Zecchina et al.40 and Hadjiivanov9 conclusions. However, the presence of Cu2+ species cannot be possible after the reducer H2 treatment. It is means that band at 2165 cm-1 could be due to carbonyls absorbed on Cu+1 species or CO absorbed on the Al3+ cations from the support in octahedral coordination ((Al+3-CO)oct). Note that tetrahedral site on alumina are blocked by copper particles avoiding the CO adsorption. Thus, the attribution of bands at 2122 cm-1 to Cu+1 species after a reducer treatment could be explained with support-metal interactions by the –OH groups evaluation. Figure 4.14 ν(OH) vibrational modes of the OH groups present on Cu/Al2O3 during the CO adsorption after CO2 treatment (a) and H2 treatment (b). Figure 4.14 represents the OH vibration region during the CO adsorption after CO2 treatment (a) and H2 treatment. Bands around 3622, 3650, 3684 and 3755 cm-1 were observed in both cases. Those bands were assigned to terminal OH groups coordinated onto surface tetrahedral aluminium centers (3755 cm-1)38, to OH groups linked to tetrahedral and octahedral Al+3 cations (3684 cm-1)41 and to the ν(OH) modes of the OH groups present in the newly formed mono (3650 cm-1) and bidentate hydrogen carbonates (3620 cm-1)42. Quantitatively, lower intensity of the bands assigned to the OH vibrational modes was observed after H2 treatment, being the intensity of the metallic sites (from 1950 to 2200 cm-1) affected in the same way. Surface Characterization: Probe Molecule Adsorption 93 The similar behavior of both groups of band is related with the strength of the interaction support-metal, i.e. copper-alumina interaction is stronger after H2 treatment. According with the conclusions of K. Bechoux et al.41 who studied CO adsorptions on Ag/Al2O3 after reducing and oxidizing pretreatments, alumina hydroxyls groups present on the surface of Cu/Al2O3 after the CO2 treatment remain stable in the interface of copper nanoparticles – alumina support, while after the H2 treatment, copper atoms interact with the closer OH groups (neighboring basic oxygen) decreasing the available copper sites for the CO adsorption. Also, neighbors basic oxygen transfer electron density to the copper atoms allowing the Cu+-CO complexes and its characteristic band at 2122 cm-1. Despite the expected metallic site of copper nanoparticles, the presence of Cu+1 species could be related to the electrophilic behavior of copper atoms in the presence of CO molecules proposed by Komatsu et al.22 who concluded that Cu atoms absorb possible oxygen atoms from CO dissociation originating species. In summary, observed bands at 2098, 2122 and 2163 cm-1 after both treatments are related with the CO adsorption on metallic copper species, CO absorbed on Cu+1 species and absorbed CO on Al3+ cations with octahedral coordination ((Al+3-CO)oct). Moreover, an important detail is the not formation of CO2 bands during the CO adsorption on Cu/ Al2O3. As expected, the CO adsorption at room temperature after H2 and CO2 treatments on Cu/ Al2O3 exhibit different bands in the region between 1100 and 1800 cm-1. First, during the CO adsorption after H2 treatment, bands at 1229, 1264, 1374, 1396, 1432, 1468, 1579, 1619 and 1657 cm-1 were observed (figure 4.15). Surface Characterization: Probe Molecule Adsorption 94 Figure 4.15. FTIR spectra during CO adsorption on Cu/Al2O3 after H2 treatment. Strong carbonates bands were observed during the CO adsorption on Cu/Al2O3 after H2 treatment at 1657, 1432 and 1229 cm-1. The assignment of the bands is not simple due to the different structures showed by carbonates. The observed bands correspond to hydrogencarbonates species; νas(COO) at 1657 cm-1, νs(COO) at 1432 cm-1 and (OH) at 1229 cm-1. However, different carbonates species on the surface alumina should be expected. Weak bands at 1579, 1468 and 1264 cm-1 correspond to υas(COO), υs(COO) and υ(C-C) vibrational modes of absorbed acetates species and bands at 1374, 1396 and 1579 cm-1 characteristic of formates species were found36. However, the intensity of the secondary formed carbonates species bands are extremely weak even not showing the CH stretching vibrational mode. Thus, only hydrogen-carbonates species were considerate for the surface analysis. K. Föttinger et al.43 studied the possible pathways mechanism of carbonate formations on Pd/Al2O3 and they concluded that the formation of hydrogen-carbonates takes place via “water gas shit” reaction between CO and alumina hydroxyl groups followed by the reaction Surface Characterization: Probe Molecule Adsorption 95 of the formed CO2 with other OH groups. This fact and the strong interaction between copper atoms and hydroxyl groups could be explains the absence of water (figure 4.14b). For the other hand, figure 4.16 shows the carbonates region of Cu/Al2O3 during the CO adsorption after CO2 treatment. It clear evident the increment of the hydrogen-carbonates bands at 1657 cm-1 (νas(COO)), 1430 cm-1 (νs(COO)) and 1228 cm-1 ( (OH)) with increasing CO amounts. Figure 4.16. FTIR spectra during the CO adsorption on Cu/Al2O3 after CO2 treatment. In this case there is a remarkable species transformation. First, an isosbestic point at 1465 cm- 1 was observed. This frequency correspond to the of absorbed acetates species which are decreasing with the CO adsorption. Band correspond to υas(COO) of acetate species is not shown due to a change of slope with the formation of other species. Moreover, according to Morterra et al.44, the presence of the isosbestic point at 1465 cm-1 evidence the presence of different hydrogen-carbonates during the CO adsorption and the transformation of the surface species from hydrogen-carbonates B2, which are mainly formed by the interaction of CO2 Surface Characterization: Probe Molecule Adsorption 102 Figure 4.22. Difference FTIR spectra of CO adsorption at room temperature on Pt/ Al2O3 (orange color) and CO adsorption on Pt3Cu1/ Al2O3 after H2 treatment (blue color) and CO2 treatment (red color). Bands with solid lines correspond to species on Pt3Cu1/ Al2O3 and bands with dash lines correspond to species on Pt3Cu1/ Al2O3. After the CO adsorption on Pt/ Al2O3 pretreated under both treatments, bands at 1684, 1578, 1378 and 1290 cm-1 assigned to acrylates species were observed (see figure 4.6). However, the addiction of small amount of copper supposes different surfaces species. After the CO adsorption on Pt3Cu1/ Al2O3 pretreated under H2 and CO2 atmosphere, bands at 1656, 1432 and 1229 cm-1 assigned to hydrogen-carbonates were observed. However, when sample is pretreated under H2 atmosphere, the decrement of bands at 1578 and 1460 cm-1 characteristic of absorbed acetates species is showed. Thus, copper species in PtCu alloy (3:1 molar ratio) induce the surface hydrocarbons oxidation. Initial acrylates species formed by platinum atoms are oxidized to acetates species50. With the increment of oxidation degree, acetates species are oxidized to CO2 and H2O which involves an enhanced of the sites availability. Finally, CO2 and H2O form hydrogen-carbonates (figure 4.23)50. Surface Characterization: Probe Molecule Adsorption 103 Figure 4.23. Mechanism of hydrocarbon oxidation on the Pt3Cu1/ Al2O3 surface. The appearance of the decreasing bands of acetates species in samples pretreated under hydrogen environment means a lower reaction rates for the hydrocarbons oxidation. Small CO2 production was observed during the CO adsorption on Pt3Cu1/Al2O3 after H2 treatment. For the other hand, when sample is pretreated under CO2 environment, only the final product is appreciable. 4.4.2 CO adsorption on Pt1Cu3/ Al2O3: The role of Platinum. To elucidate the role of platinum in the intermetallic compound PtCu, CO adsorption on Pt1Cu3/Al2O3 after CO2 and H2 pretreatment were carried out (Figure 4.24). Surface Characterization: Probe Molecule Adsorption 104 Figure 4.24. Difference spectra of CO adsorption on Pt1Cu3/Al2O3 after CO2 treatment (red color) and H2 treatment (blue color). During the CO adsorption after both pretreatments, bands characteristic of absorbed carbonyls on copper species appear at the same frequency, 2098 and 2122 cm-1. Considering the Curich surface composition between 63 and 83% of copper species calculated by APXPS of the unsupported Pt1Cu3 nanoparticles (Figure 3.5, chapter 3), the surface of Pt1Cu3/Al2O3 catalyst is extremtly similar to the Cu/Al2O3 surface and the observed bands at 2098 and 2122 cm-1 could be related to carbonyls absorbed on metallic copper species and Cu+ species respectivily. The different intensties is due to the surface claneaning method effectivity. Pt1Cu3/Al2O3 surface is more availability for the carbonyls abdsortion after CO2 treatments. Moreover, as in the case of CO adsorption on Cu/Al2O3 (section 4.3.2), copper atoms present higher interaction with the alumina OH groups which supposes a less availabilty sites for the CO adsorption. Finally, band at 2172 cm-1 was assigned to carbonyls absorbed to Al+3 cations in octahedral coordination ((Al+3-CO)oct). A comparison of the FTIR spectra recorded in the 1000-1800 spectral region of Cu/Al2O3 and Pt1Cu3/Al2O3 catalyst is shown in figure 4.25. Intermetallic compound supported on alumina present similar surface species after the pretreatment in CO2 and H2 environment. There is a clear hydrocarbon oxidation process during the CO adsorption in both catalyst. Bands at 1229, 1432 and 1657 cm-1 characteristic of hydrogen carbonates species are clearly incresing while bands at 1470 and 1580 cm-1 characteristic of absorbed acetate species are decreasing. In the case of CO2 pretreatment, bindentate carbonates were not observed on the Pt1Cu3/Al2O3 surface and, contrary to the CO adsorption on Cu/Al2O3, the decrement of the band correspond to the υas(COO) vibrational mode at 1580 cm-1 is evident. Surface Characterization: Probe Molecule Adsorption 105 Finally, while CO2 production was observed during the CO adsorption on Pt3Cu1/Al2O3 after H2 treatment, in this case, weak bands characteristic to CO2 gas phase was observed during the CO adsorption on the Pt1Cu3/Al2O3 after CO2 treatment. Therefore, according with the avobe comparion, Pt1Cu3/Al2O3 after both treatment shows a low reaction rate for the surface hydrocarbons oxydation respect to the PtCu alloy 3:1 molar ratio. The acetates species oxidation to CO2, clearly evidenced with the decrement of bands at 1470 and 1580 cm-1, and the not CO2 oxidation to hydrogen-carbonates species could be explain the higher CO conversions and selectivity to CO2 for the intermetallic PtCu 1:3 molar ratio based catalyst during the preferential oxidation of CO23,51. Figure 4.25. Difference FTIR spectra comparison between the CO adsorption on Cu/Al2O3 (orange color) and Pt1Cu3/Al2O3 after CO2 treatment (red color) and H2 treatment (blue color). 4.5. Conclusions Throughout this chapter, a complete surface characterization by probe molecule absorption of nanoparticles supported on alumina has been shown. The initial monomental catalysts characterization revealed the metallic character site of the monometal platinum catalyst as well as the simultaneity presence of Cu+1 and metallic copper species on the surface of copper Surface Characterization: Probe Molecule Adsorption 106 supported on alumina. Different surface species were observed as a function of the active site nature without influence of the conditioning treatment. The efficiency of the conditioning treatments based in CO2 and H2 atmospheres was probed by the CO and O2 adsorption showing an enhance of the availability surface after CO2 treatments. Moreover, the influence of metal-support interactions on the active site was evaluated by the correlations of the alumina OH groups and carbonyls absorbed on the active site. Also, oxidation state of monometal copper catalyst was confirmed by NO probe molecule adsorption. The role of both metals in the intermetallic compound PtCu was studied by CO adsorption on supported PtCu nanoparticles alloy with different molar ratios on alumina. The surface hydrocarbons oxidation by copper effects has been concluded showing different reaction rate as a function of the conditioning treatment. For the other hand, the main role of platinum is the CO dissociation which allows the CO2 formation by different pathway mechanism. 4.6 Associated content Procedure carried out for the probe molecule adsorption was always the same. First, sample is pressed in pellet form of 15-20 mg and 2 cm2 of area (spectra are normalized by the area and weight). After, 5 Torr conditioning environment is introduced in the vacuum system at room temperature. In this point, conditioning treatment simulation stars. Sample is heating at 350ºC for 2 hours (heating rate of 10ºC/min). Then, sample is cooling down without cooling rate follow by gas evacuation at 200ºC temperature trying to enhance the evacuation efficiency. Finally, probe molecule adsorption stars at room temperature when the total vacuum in the cell is almost 10-6 mbar. Probe molecule doses of 0.1 Torr are added until surface saturation. Surface Characterization: Probe Molecule Adsorption 107 Figure 4.26. Schematic probe molecule adsorption protocol. REFERENCES (1) Gilmore, I. S.; Vickerman, J. C. Surface Analysis – The Principal Techniques; 2009. (2) Niemantsverdriet, J. W. Spectroscopy in Catalysis; 2007. (3) Hadjiivanov, K. I.; Vayssilov, G. N. Adv. Catal. 2002, 47, 307–511. (4) Busca, G.; Lamotte, J.; Lavalley, J.; Lorenzelli, V. J. Am. Chem. Soc. 1987, 106 (16), 5197–5202. (5) Finocchio, E.; Busca, G.; Lorenzelli, V.; Ingegneria, F.; Kennedy, P.; Willey, R. J. J. Chem. Soc. Faraday Trans. 1994, 90 (21), 3347–3356. (6) Lavalley, J. C. Catal. Today 1996, 27, 377–401. (7) Turek, A. M.; Wachs, E. J. Phys. Chem. C 1992, 96, 5000–5007. (8) Binet, C.; Daturi, M.; Lavalley, J. Catal. Today 1999, 50, 207–225. (9) Hadjiivanov, K. I.; Vayssilov, G. N. Adv. Catal. 2002, 47, 307–511. (10) Dandekar, A.; Vannice, M. A. J. Catal. 1998, 178 (2), 621–639. (11) Amores, J. M. G.; Escribano, V. S.; Busca, G.; Lorenzelli, V. J. Mater. Chem. 1994, 4 (6), 965–971. (12) Morterra, C.; Bolis, V.; Magnacca, G. Faraday Discuss. 1996, 92 (11), 1991–1999. (13) Zecchina, A.; Platero, E. E.; Areán, C. O. J. Catal. 1987, 247, 244–247. Surface Characterization: Probe Molecule Adsorption 108 (14) Olsen, C. W.; Masel, R. I. Surf. Sci. 1988, 201, 444–460. (15) J., J. H.; Xu, J.; Yates, J. T. J. Chem. Phys. 1993, 99, 721–724. (16) Hu, J.; Xu, J.; Zhao, Y.; Shi, L.; Li, Q.; Liu, F.; Ullah, Z.; Li, W.; Guo, Y.; Liu, L. Nature 2017, 7, 45358. (17) Bazin, P.; Saur, O.; Lavalley, J. C.; Daturi, M.; Blanchard, G. Phys. Chem. Chem. Phys. 2005, 7, 187–194. (18) Hoffmann, F. M. Surf. Sci. Rep. 1983, 3, 107–192. (19) Stakheev, A. Y.; Shpiro, E. S.; Jaeger, N. I.; Schulz-EKloff, G. Catal. Letters 1995, 32, 147–158. (20) Liu, K.; Wang, A.; Zhang, T. ACS Catal. 2012, 2, 1165–1178. (21) Shore, L.; F.Ruettinger, W.; Farrauto, R. J. Platinum group metal promoted copper oxidation catalysts and methods for carbon monoxide remediation. US 6,913,739 B2, 2005. (22) Komatsu, T.; Takasaki, M.; Ozawa, K.; Furukawa, S.; Muramatsu, A. J. Phys. Chem. C 2013, 117 (20), 10483–10491. (23) Romero-Sarria, F.; Garcia-Dali, S.; Palma, S.; Jimenez-Barrera, E. M.; Oliviero, L.; Bazin, P.; Odriozola, J. A. Surf. Sci. 2015, 648, 84–91. (24) Ding, K.; Gulec, A.; Johnson, A. M.; Schweitzer, N. M.; Stucky, G. D.; Marks, L. D.; Stair, P. C. Science (80-. ). 2015, 350 (6257), 189–192. (25) Borodko, Y.; Sook Lee, H.; Hoon Joo, S.; Zhang, Y.; Somorjai, G. J. Phys. Chem. C 2010, 114, 1117–1126. (26) Hollins, P. Surf. Sci. Rep. 1992, 16, 51–94. (27) Maugé, F.; Lavalley, J. J. Catal. 1992, 137, 69–76. (28) Digne, M.; Sautet, P.; Raybaud, P.; Euzen, P.; Toulhoat, H. J. Catal. 2002, 211, 1–5. (29) Digne, M.; Sautet, P.; Raybaud, P.; Euzen, P.; Toulhoat, H. J. Catal. 2004, 226, 54–68. (30) Finocchio, E.; Busca, G.; Lorenzelli, V.; Escribanob, V. S. J. Chem. Soc. Faraday Surface Characterization: Probe Molecule Adsorption 109 Trans. 1 1996, 92 (9), 1587–1593. (31) Busca, G.; Daturi, M.; Finocchio, E.; Lorenzelli, V.; Ramis, G.; Willey, R. J. Catal. Today 1997, 33, 239–249. (32) Finocchio, E.; Daturi, M.; Binet, C.; Lavalley, J. C.; Blanchard, G. Catal. Today 1999, 52, 53–63. (33) Rinnemo, M.; Kulginov, D.; Johansson, S.; Wong, K. L.; Zhdanov, V. P.; Kasemo, B. Surf. Sci. 1997, 376, 297–309. (34) Mccrea, K. R.; Parker, J. S.; Somorjai, G. A. J. Phys. Chem. B 2002, 106, 10854– 10863. (35) Bär, M.; Falcke, M.; Zijlicke, C.; Engel, H.; Eiswirth, M.; Ertl, G. Surf. Sci. 1992, 269– 270, 471–475. (36) Davydov, A. Molecular Spectroscopy of Oxide Catalyst Surfaces; 2003. (37) Sexton, B. A. Chem. Phys. Lett. 1979, 65 (3), 469–471. (38) Morterra, C.; Ghiotti, G.; Boccuzzi, F.; Coluccia, S. J. Catal. 1978, 51, 299–313. (39) Busca, G.; Lorenzelli, V.; Escribanob, V. S.; Guidetti, R. J. Catal. 1991, 131, 167–177. (40) Zecchina, A.; Spoto, G.; Scarano, D. J. Electron Spectros. Relat. Phenomena 1987, 45, 269–279. (41) Bechoux, K.; Marie, O.; Daturi, M.; Delahay, G.; Petitto, C.; Rousseau, S.; Blanchard, G. Catal. Today 2012, 197 (1), 155–161. (42) Köck, E.; Kogler, M.; Bielz, T.; Klötzer, B.; Penner, S. J. Phys. Chem. C 2013, 117, 17666–17673. (43) Föttinger, K.; Schlögl, R.; Rupprechter, G. Chem. Commun. 2008, 3, 320–322. (44) Morterra, C.; Zecchina, A.; Coluccia, S.; Chiorino, A. J. faraday Transit. 1977, 1544– 1560. (45) Hadjiivanov, K.; Knözinger, H. Phys. Chem. Chem. Phys. 2001, 3, 1132–1137. (46) Centi, G.; Perethaner, S.; Biblino, D.; Giamello, E. J. Catal. 1995, 151, 72–92. Surface Characterization: Probe Molecule Adsorption 110 (47) Hadjiivanov, K. I. Catal. Rev. 2000, 42 (1–2), 71–144. (48) Tanaka, T.; Okuhara, T.; Misono, M. Appl. Catal. B Environ. 1994, 4, L1–L9. (49) Venkov, T.; Hadjiivanov, K.; Klissurski, D. 2002, 2443–2448. (50) Davydov, A. A.; Mikhaltchenko, V. G.; Sokolovskii, V. D.; Boreskov, G. K. J. Catal. 1978, 55, 299–313. (51) Liu, K.; Wang, A.; Zhang, T. ACS Catal. 2012, 2 (6), 1165–1178. Chapter 5. Model Catalyst Characterization Under Reaction Conditions Estudio De Catalizadores Basados En Platino Para La Reacción De PROX: Influencia De Los Depósitos De Carbono Model Catalyst characterization under reaction conditions 118 Second catalytic cycle presents higher CO conversion values from all the reaction temperature range. Reproducibility study confirms a change in the catalytic behavior around 200ºC showed by a modification of the slope in the CO conversion line evolution. This change is also observed in the second cycle. Selectivity toward CO2 formation shows a decrement with the temperature increment. It must be considered that after H2 pretreatment, carbon deposits are present on the nanoparticles surface (see section 2.2.2, chapter 2) while after the surface cleaning method, nanoparticles surface are completely free of carbonaceous rests. Considering that after the surface cleaning method, changes in the electronic surface and the geometric structure were not observed refereed to the unsupported nanoparticles pretreated under H2-atmosphere (chapter 2) , the different CO conversion could be related with the presence or absence of carbon deposits on the nanoparticles surfaces. Therefore, free-carbon deposits surfaces involve lower CO conversion while somehow, surface carbon deposits presence are related with higher CO conversion values. However, the application of this system in mobile devices supposes the activation of the catalysts under the reaction atmosphere. For this reason, catalytic activity for unsupported PtCu3 nanoparticles pretreated under reaction mixture environment after the surface cleaning method application was evaluated (figure 5.4). Model Catalyst characterization under reaction conditions 119 Figure 5.4. Catalytic activity of the PtCu3 nanoparticles as a function of reaction temperature. Nanoparticles were conditioned under reaction mixture environment. (a) Reaction flow composition: 50% H2, 10% CO2, 1% O2, 1% CO and N2 balance. (b) Reaction flow composition: 50% H2, 10% CO2, 10% H2O, 1% O2, 1% CO and N2 balance. Figure 5.4a shows the catalytic activity of unsupported nanoparticles pretreated under reaction mixture environment at 350ºC for 2 hours in dry conditions. It is appreciable a similar catalytic behavior that nanoparticles pretreated under H2-flow (figure 5.3a). In this case, a maximum of 56% of CO conversion was observed in the first cycle while the second cycle supposed an increment of the CO conversion at lower temperature with a maximum CO conversion of 61% at 210ºC. In both cycles, CO conversion is controlled by thermodynamic equilibrium after the maximum value. In all the cases, selectivity values decrease with the increment of reaction temperature. However, when 10% H2O was added to the main flow in the conditioning treatment, unsupported PtCu3 nanoparticles showed a totally different catalytic behavior (figure 5.4b). The presence of water in the reaction mixture involves lower CO conversion with linearly positive trend with the increment of reaction temperature. Similar to the samples conditioned under CO2 atmosphere, there is a change of slope at 200ºC in the CO conversion and CO conversion after the second cycle presents higher values. Model Catalyst characterization under reaction conditions 120 Figure 5.5. APXPS C1s spectra from unsupported PtCu3 nanoparticles acquired during the reaction step at 200ºC in different reaction mixture composition; 50% H2, 10% CO2, 1% O2, 1% CO, 10% H2O and N2 balance (blue line) and same composition without water (green line). The IMFP of C1s photoelectron was 5 Å. A simulation of the conditioning treatment and reaction step under both different feed streams (dry and wet) of the unsupported PtCu3 nanoparticles was carried out by APXPS experiments. Figure 5.5 shows the C1s spectra acquired during the reaction step at 200ºC. Green line shows the evolution of C1s signal from nanoparticles surface during the reaction step in absence of water. It is appreciable a remarkable signal with a main compound at 284.6 eV correspond to C-C species from the synthesis step. However, the addition of water supposed a dramatically decrement of the C1s signal. This fact could explain the observed different catalytic behavior which showed higher CO conversion for carbon-polluted surfaces and it could accept the hypothesis proposed by Romero-Sarria et al.13 based in the positive role of carbon in the preferential oxidation of CO. Model Catalyst characterization under reaction conditions 121 In order to confirm the positive role of carbon in the conventional PtxCuy catalyst for PrOx reaction, catalytic activity for two different feed streams of nanoparticles supported on Al2O3 was evaluated. As it was concluded before, dry reaction environment involves the presence of carbon deposits on the nanoparticles surface while the presence of water in the real reaction mixture flow means almost the totally removing of surface carbon overlayers. Figure 5.6. PrOx reaction performance comparison of PtCu3/ Al2O3 under dry feed stream (a) and wet feed stream (10% H2O) (b). Figure 5.6 shows the PtCu3/ Al2O3 catalytic activity for dry and wet feed streams. As observed in figure 5.6a for dry reaction conditions, CO conversions in both tasted cycles present maximum CO conversion values around 80-85% at 150ºC being limited from this point by the thermodynamic equilibrium. In the case of wet feed streams, PtCu3/ Al2O3 presents similar global catalytic behavior showing lower CO conversion at low temperatures in the first cycle. Probably, the low CO conversion values at low temperature for wet reaction conditions correspond with free-surface of carbon deposits on the nanoparticles and support. However, the presence of support could be favoring the deposit of carbon rests on the alumina surface during the reaction Model Catalyst characterization under reaction conditions 122 which involves a change in the catalytic properties toward the observed catalytic behavior under dry conditions (see section 4.4.2, chapter 4). Doing a comparison of the catalytic activities of PtCu3/ Al2O3 and unsupported PtCu3 nanoparticles can be observed higher CO conversion values for the supported nanoparticles. The strong influence of the support on surface rearrangements of bimetallic nanoparticles was reported by Divins et al.14. Alumina support can provide new active sites and it can be affecting to the physical and chemical properties of nanoparticles. In this way, thanks to the present of Al2O3 support, CO conversion reaches the maximum value and the equilibrium values at lower temperature than unsupported nanoparticles. In all the cases, selectivity toward CO2 formation shows the same linearly trend to lower values with the increment of reaction temperature. However, the use of Al2O3 allows side reactions which induce to loss of 25% of selectivity. 5.3 Surface dynamic during PrOx reaction Come back to the main goal of this manuscript, after the study of the role of surface carbon deposits, another possible answer to the untypical catalytic behavior of PtxCuy nanoparticles based catalysts for PrOx reaction could be the surface dynamic or rearrangements of PtCu nanoparticles under working conditions. F. Tao et al.5 showed how some structures and active phases only exist under reaction conditions. In this way, surface in situ characterization is essential to identify the active sites at work as well as the restructuring phenomena as a function of the surrounding environment and temperature. Therefore, after the surface cleaning method (exchange capping ligand method and CO2 treatment) and the conditioning treatment under reaction feed stream, Model Catalyst characterization under reaction conditions 123 the electronic surface properties of the synthetized PtCu3 nanoparticles were studied by in situ APXPS under working conditions at different temperatures. Fresh unsupported PtCu3 nanoparticles were pressed to form a pellet about 13 mm in diameter. An initial characterization of the fresh sample was carried out by APXPS in ultra-high vacuum conditions in order to record the state of the synthetized nanoparticles during the surface cleaning method and before the conditioning environment exposure (see section 3.2.2, chapter 3). Next step was the introduction of the reaction environment in the analysis chamber. It must be highlighted the use of mass flow controller for the partial pressure control of the reacts. We were the first team users in the startup of the mass flow controller at the CIRCE beamtime of ALBA synchrotron light source in February 2018. Mass flows system connected to the APXPS endstation is equipped only by four gas lines. It supposed the use of mixture gas cylinder of O2 and CO2 [CO2/O2=4]. Therefore, due to the system limitations, the gaseous flux introduced by mass flow controller was 30mL/min H2 (gas line 1), 15 mL/min N2 (gas line 2), 3 mL/min CO (gas line 3) and 15 mL/min O2+CO2 (gas line 4). 10% H2O was introduced to the analysis chamber using a bubbler connected by UHV leak valve. The pressure work in the analysis chamber was 1.5 Torr. Feed stream was introduced at room temperature and sample was heating at 300ºC (heating rate 10ºC/min). As it was showed in chapter 3, after the surface cleaning method, nanoparticles present a Cu-rich surface soft core-shell structure. However, nanoparticles composition changes as a function of the surrounding environment. Thus, first of all, depth profile analysis was carried out for elucidate the nanoparticle composition during the conditioning treatment and reaction step under the same feed stream. Model Catalyst characterization under reaction conditions 124 The chosen incident photon energies for carrying out the depth profiling were 300, 548 and 1026 eV. Obviously, acquired APXPS spectra were normalized by the photon flux and the cross section for the analysis. For the inelastic mean free path (IMFPs) calculation in this APXPS experiment, we employed the TPP-2M formula proposed by Tanuma et al. considering the nanoparticles as well-defined PtCu3 alloy with an only electronic distribution. Consequently, the Pt 4f and Cu 3p IMFP for the selected photon energies were 5, 8 and 13 Å respectively. Figure 5.7. Dependence of Pt and Cu composition for the PtCu3 nanoparticles during the conditioning treatment under the reaction mixture at 300ºC. Figure 5.7 shows the depth profiling carried out under the reaction environment. Surface composition in the outer shell (excited at hν =300 eV) was 74% Cu and 26% Pt, whereas for the inner region (excited at hν =1026 eV) was 60% Cu and 40% Pt. In comparison with the Pt15Cu85 surface composition obtained under CO2 environment of the surface cleaning method (section 3.2.2, chapter 3), it is observed from figure 5.7 a migration of Pt atoms towards the surface upon reaction conditions, reaching a close Pt26Cu74 composition on the surface nanoparticles. Model Catalyst characterization under reaction conditions 125 After the measurements at 300ºC during the conditioning treatment in the reaction atmosphere, sample was cooling down to room temperature in the same atmosphere. For the preferential oxidation of CO test, Pt 4f - Cu 3p, C1s and O1s APXPS spectra were acquired at four different temperatures; room temperature, 150ºC, 200ºC y 300ºC. Photon energies of 300 eV for Pt 4f - Cu 3p, 450 eV for C1s and 700 eV for O1s were used to produce photoelectrons with kinetic energies around 170 eV for APXPS core-level spectra, corresponding to IMFPs in a range of 5 and 8 Å. Figure 5.8 shows the Pt 4f - Cu 3p spectra correspond to the surface state at increasing temperature. Figure 5.8 Ambient pressure X-ray photoelectron spectra of Pt 4f y Cu 3p recorded for unsupported PtCu3 upon reaction conditions at different temperatures. Model Catalyst characterization under reaction conditions 126 Before elucidating the nature of the Pt and Cu species, it is necessary detailing of the observed changes in the surface atomic distribution as a function of the reaction temperature (figure 5.9). As mentioned before, shell of synthetized PtCu3 nanoparticles during the conditioning treatment under reaction condition was composed by 74% of copper and 26% of platinum. However, the decrement of the temperature for the beginning of the catalytic reaction leads to higher surface Cu/Pt atomic ratio. At room temperature, restructuring phenomena took place and surface Cu/Pt atomic ratio increased from 2.9 to 4.07 reaching a surface Pt19Cu81 composition. It means a migration of the copper atoms to the surface. Then, with the increment of temperature to 150ºC, surface Cu/Pt atomic fraction remained approximately constant (around 4.12). For the other hand, a further increment of temperature involved a dramatic decline of the atomic fraction reaching values of 2.8 at 200ºC and 2.4 at 300ºC. Interesting, the observed specie at 76.9 eV only appear at the same temperatures when Cu/Pt atomic ratio is maximum. Therefore, for elucidate the nature of the observed species, C 1s and O 1s during the catalytic reaction were recorded. Figure 5.9. Surface Cu/Pt atomic ratio of the synthetized PtCu3 nanoparticles under reaction conditions after the proposed surface cleaning method. Model Catalyst characterization under reaction conditions 127 Photoemission lines of the Pt 4f and Cu 3p core-level contain in all the temperature range different components. First, single platinum specie was observed in all the range of temperature with a remarkable Pt 4f7/2 peak around 71.0 eV and the corresponding Pt 4f5/2 peak shifted 3.3 eV to higher binding energies (L-S coupling). At room temperature, the peak is shifted to 71.1 eV while the increment of temperature supposes a shift to lower binding energies (70.9 eV). For the other hand, Cu 3p spectra can be deconvoluted up to three spectral features. First found peak at 74.3 eV was assigned to Cu 3p3/2 component of metallic copper species. The Cu 3p1/2 contribution was fitted at 2.2 eV (L-S coupling). Second Cu 3p3/2 peak was found at 75.3 eV. Both species were observed in all the temperatures. However, during the reaction at room temperature and reaction at 150ºC, a new Cu 3p3/2 component at 76.9 eV was observed. Figure 5.11 shows the APXPS spectra of Pt 4f, Cu 3p, O 1s and C 1s during the catalytic reaction. In the conditioning treatment step at 300ºC (Cu/Pt atomic ratio around 3), Pt 4f and Cu 3p spectrum show contribution of Pt 4f7/2 at 71.1 eV and three contribution of copper species at 74.3, 75.3 and 76.9 eV. Due to the high hydrogen content in the environment (50%) and the temperature, conditioning treatment under reaction conditions could be considerate as reducer treatment. In this way, Cu 3p3/2 component at 74.3 was assigned to metallic copper species. The shift of 1 eV to 75.3 eV could be explained by two options: i) oxidized copper species (Cu+) or, ii) metallic copper species with deficient electron density caused by the commented electron donation in the alloy from copper to platinum atoms (see section 4.4.1). Model Catalyst characterization under reaction conditions 134 on the nanoparticles surface. The loss of carbon species with sp3 hybridization means lower reactivity of the monolayer showing a change in the positive trend in the CO conversion (figure 5.4b). 5.4 Structural transitions during PrOx reaction First of all, X-ray absorption near-edge structure spectroscopy (XANES) has been extensively applied to the study of the local electronic and geometrical environments of copper and platinum sites21–26. This technique detects spectral changes in the near edge feature which appear as resonances characterized by the energy position and the intensity of the absorption edge27. Analysis of spectra makes possible a definitive interpretation of the synthetized PtCu3 catalytic behavior. 5.4.1 Electron density transfer during PrOx reaction Figure 5.12 shows the in situ XANES spectra of synthetized unsupported PtCu3 nanoparticles after the surface cleaning method and during the preferential oxidation of CO at room temperature and 250ºC. Initial spectra after the surface cleaning method were taken under He environment. Sample was activated at 300ºC for 2 hours (heating rate 10ºC/min) by flowing 100 mL/min stream feed based in the reaction atmosphere. It is appreciable the same line shape of spectra and same white-line energy for the Pt L3- edge (a) and same edge energy for the Cu K-edge (b) meaning the identical oxidation states in all the cases. However, changes in the intensity of the Cu K-edge and Pt L3 white-line were observed. Model Catalyst characterization under reaction conditions 135 Figure 5.12. Comparison between XANES spectra for the synthetized PtCu3 nanoparticles after the surface cleaning method and during the preferential oxidation of CO at room temperature and 250ºC. a) Pt L3-edge XANES spectra. The insert of part (a) shows the magnified Pt L3 white-line. b) Cu K-edge XANES spectra. The insert shows the magnification of Cu K-edge. In particular, at the Pt L3-edge, the white-line (2p3/2 to 5d orbitals transitions) intensity increases under working conditions referred to the conditioning treatment spectrum. According to Mansour et al.28, the integrated Pt L3 white-line area is proportional to the number of unoccupied 5d states. A change in the occupied d-electron states could be expected by charge transfer with the reaction environment or with copper atoms of the alloy (see section 3.2.3, chapter 3). In this way, the nanoparticles exposure to the reaction atmosphere leads to increment of the Pt L3 white-line and spectra broaden about 8 eV above the absorption edge (figure 5.12a). This broad structure is related to hydrogen absorption and Ankudinov et al.29 explained that it is due to the change in the Pt potential by the Pt-H bonding. Model Catalyst characterization under reaction conditions 136 Reaction atmosphere is mainly composed by hydrogen (around 50%), since the electronegativity of platinum (2.28) and hydrogen (2.2) are similar in the Pauling scale30transfer process with the environment are expected. Watari and Ohnishi31 studied the changes in the electronic structure of Pt cluster during the hydrogen adsorption. They concluded that hydrogen adsorbed on platinum surface means the reduction of the white-line and the growth of the adjacent shoulder. However, an increment of the Pt L3 white-line of the PtCu3 nanoparticles during the reaction atmosphere exposure is observed. Despite XANES analysis is based in the excitation of atoms from the entire nanoparticles, total electron density of atoms is strong influenced by the surface species and the reaction-driven restructuring. Moreover, Shin et al.32 studied by DFT calculations the influence in the site electron density of CO and oxygen species adsorbed on the surface PtCu core-shell nanoparticles. They concluded the stronger CO adsorption in pure metal than in PtCu systems showing an increment of C-O distance with the copper amount. Copper atoms transfer electrons to platinum atoms increasing the backbonding between platinum and CO molecule which lead to C-O bond elongation and electron rich platinum atoms. Therefore, rearrangements of synthetized PtCu3 nanoparticles under working conditions involve changes in the electronic density of platinum and copper atoms. The increment of the Pt L3 white-line means higher number of unoccupied dstates and, therefore, lowers electron density. Of course, a modification of the platinum electron density by copper atoms is reflected in Cu K-edge spectra. As commented in section 3.2.3 (chapter 3), Cu K-edge intensity (1s to 4p transition) is proportional to the density of states so, increment of Cu K-edge means an increment of copper electron Model Catalyst characterization under reaction conditions 137 density. Thus, during the preferential oxidation of CO, an increment of reaction temperature involves higher electron density transfer from copper to platinum atoms. It is showed by the decrement of the Pt L3 white-line and the Cu K-edge intensity (filled Pt 5d orbitals and decrement of density of state of Cu). This electronic transfer elucidates the nature of the Cu 3p3/2 APXPS peak at 75.3 eV. As the electron density transfer increases with the temperature, the peak at 75.3 eV assigned to electron deficient metallic Cu species increases. For the other hand, theoretical hypothesis proposed by Shin et al.32 and Komatsu et al.33 based in the higher electron density transfer from copper to platinum atoms with the copper amount increment was verified. The optimized synthesis method described in section 2.2.1 (chapter 2) allows the synthesis of well-defined PtxCuy with different molar ratio. In this case, intermetallic Pt1Cu1 nanoparticles were synthetized and characterized by X-ray absorption spectroscopy (figure 5.13). Model Catalyst characterization under reaction conditions 138 Figure 5.13. Comparison between Pt L3-edge XANES spectra of Pt standard collected under He environment and Pt1Cu1 and PtCu3 nanoparticles acquired under working conditions at room temperature. The insert shows a magnification of the Pt L3 white-line Analysis of Pt L3-edge XANES spectra showed lower Pt L3 white-line as copper amount increased confirming the higher electron density transfer. It concludes that the difference rearrangements of PtCu3 nanoparticles observed by APXPS as a function of the reaction temperature. 5.5 Conclusions In this chapter, catalytic activity for the preferential oxidation of CO (PrOx reaction) of the synthetized PtCu3 nanoparticles after the surface cleaning method has been evaluated. Nanoparticles showed different catalytic behavior as a function of the surface carbon deposits presence or absence. In this sense, Nanoparticles conditioned in hydrogen environment or under dry reaction mixture environment reached higher values of CO conversion that nanoparticles conditioned under CO2 environment or under wet reaction mixture environment. The simulation of the catalytic reaction by in situ APXPS related the higher catalytic activity with the presence of surface carbon deposits. Surface-driven restructuring phenomenon under working conditions was studied. The exposure to the reaction environment at room temperature and 150ºC leads to surface copper enrichment. The nature of the surface carbon and oxygen species had been described by APXPS analysis. Results concluded that the interaction between nanoparticles surface and the oxidized carbon species depend on the stacking degree of the carbon multi or single layers. Maximum CO conversion values were related with the Model Catalyst characterization under reaction conditions 139 presence of surface carbon monolayers composed by carbon atoms with sp2 and sp3 hybridization. Moreover, the exposure to the reaction mixture environment also leaded to an electron density transfer from copper to platinum atoms. This electronic transfer increases with the temperature. REFERENCES (1) Knop-Gericke, A.; Hävecker, M.; Schedel-Niedrig, T.; Schlögl, R. Top. Catal. 2001, 15 (1), 27–34. (2) Haghofer, A.; Föttinger, K.; Nachtegaal, M.; Armbrüster, M.; Rupprechter, G. J. Phys. Chem. C 2012, 116, 21816–21827. (3) Föttinger, K.; Bokhoven, J. A. Van; Nachtegaal, M.; Rupprechter, G. J. Phys. Chem. Lett. 2011, 2, 428–433. (4) Müller, O.; Nachtegaal, M.; Just, J.; Lützenkirchen-Hecht, D.; Frahm, R. J. Synchrotron Radiat. 2016, 23, 260–266. (5) Tao, F.; Grass, M. E.; Zhang, Y.; Butcher, D. R.; Renzas, J. R.; Liu, Z.; Chung, J. Y.; Mun, B. S.; Salmeron, M.; Somorjai, G. A. Science (80-. ). 2008, 322, 932– 934. (6) Zhang, F.; Liu, Z.; Zhang, S.; Akter, N.; Palomino, R. M.; Vovchok, D.; Orozco, I.; Salazar, D.; Rodriguez, J. A.; Llorca, J.; Lee, J.; Kim, D.; Xu, W.; Frenkel, A. I.; Li, Y.; Kim, T.; Senanayake, S. D. ACS Catal. 2018, 8 (4), 3550–3560. Model Catalyst characterization under reaction conditions 140 (7) Vovchok, D.; Guild, C.; Dissanayake, S.; Llorca, J.; Stavitski, E.; Liu, Z.; Palomino, R. M.; Waluyo, I.; Li, Y.; Frenkel, A. I.; Rodriguez, J. A.; Suib, S. L.; Senanayake, S. D. J. Phys. Chem. C 2018, 122 (16), 8998–9008. (8) Eren, B.; Heine, C.; Bluhm, H.; Somorjai, G. A.; Salmeron, M. J. Am. Chem. Soc. 2015, 137 (34), 11186–11190. (9) Huck-Iriart, C.; Soler, L.; Casanovas, A.; Marini, C.; Prat, J.; Llorca, J.; Escudero, C. ACS Catal. 2018, 8 (10), 9625–9636. (10) Hernández, W. Y.; Centeno, M. A.; Romero-sarria, F.; Ivanova, S.; Montes, M.; Odriozola, J. A. . (11) Pozdnyakova, O.; Teschner, D.; Wootsch, A.; Kröhnert, J.; Steinhauer, B.; Sauer, H.; Toth, L.; Jentoft, F. C.; Knop-Gericke, A.; Paál, Z.; R. Schlögl. J. Catal. 2006, 237, 1–16. (12) Divins, N. J.; López, E.; Roig, M.; Trifonov, T.; Rodríguez, A.; Rivera, F. G. De; Rodríguez, L. I.; Seco, M.; Rossell, O.; Llorca, J. Chem. Eng. J. 2011, 167 (2–3), 597–602. (13) Romero-Sarria, F.; Garcia-Dali, S.; Palma, S.; Jimenez-Barrera, E. M.; Oliviero, L.; Bazin, P.; Odriozola, J. A. Surf. Sci. 2015, 648, 84–91. (14) Divins, N. J.; Angurell, I.; Escudero, C.; Pérez-Dieste, V.; Llorca, J. Science (80-. ). 2014, 346 (October), 620–623. (15) Thair, D.; Tougaard, S. J. Phys. Condens. Matter 2012, 24, 175002. (16) Rodriguez, N. M.; Anderson, P. E.; Wootsch, A.; Paál, Z.; Wild, U.; Schlögl, R.; Paál, Z. J. Catal. 2001, 197 (2), 365–377. Model Catalyst characterization under reaction conditions 141 (17) Drewniak, S.; Muzyka, R.; Stolarczyk, A.; Pustelny, T.; Kotyczka-Morańska, M.; Setkiewicz, M. Sensors 2016, 16 (106), 1–16. (18) Darmstadt, H.; Roy, C.; Kaliaguine, S. Carbon N. Y. 1994, 32 (8), 1399–1406. (19) Lamoen, D.; Persson, B. N. J. J. Chem. Phys. 1998, 108 (8), 3332–3341. (20) Liu, L.; Ryu, S.; Tomasik, M. R.; Stolyarova, E.; Jung, N.; Hybertsen, M. S.; Steigerwald, M. L.; Brus, L. E.; Flynn, G. W. Nano Lett. 2008, 8 (7), 1965–1970. (21) Borfecchia, E.; Lomachenko, K. A.; Giordanino, F.; Falsig, H.; Beato, P.; Soldatov, A. V; Bordiga, S.; Lamberti, C. Chem. Sci. 2015, 6, 548–563. (22) Martini, A.; Borfecchia, E.; Lomachenko, K. A.; Pankin, I. A.; Negri, C.; Berlier, G.; Beato, P.; Falsing, H.; Bordiga, S.; Lamberti, C. Chem. Sci. 2017, 8, 6836– 6851. (23) Lamberti, C.; Bordiga, S.; Zecchina, A.; Salvalaggio, M.; Geobaldo, F.; Areán, C. O. J. Chem. Soc, Faraday Trans. 1998, 94 (10), 1519–1525. (24) Braglia, L.; Borfecchia, E.; Martini, A.; Bugaev, A. L.; Soldatov, A. V.; Øien- Ødegaard, S.; Lønstad-Bleken, B. T.; Olsbye, U.; Lillerud, K. P.; Lomachenko, K. A.; Agostini, G.; Manzoli, M.; Lamberti, C. Phys. Chem. Chem. Phys. 2017, 19 (40), 27489–27507. (25) Behafarid, F.; Ono, L. K.; Mostafa, S.; Croy, J. R.; Shafai, G.; Hong, S.; Rahman, T. S.; Bare, S. R.; Roldan Cuenya, B. Phys. Chem. Chem. Phys. 2012, 14 (33), 11766–11779. (26) Ankudinov, A. L.; Rehr, J. J.; Low, J. J.; Bare, S. R. J. Chem. Phys. 2002, 116, 1911–1919. Model Catalyst characterization under reaction conditions 142 (27) Jiang, T.; Ellis, D. E. J. Mater. Res. 1996, 11 (9), 2242–2256. (28) Mansour, A. N.; Cook, J. W.; Sayers, D. E. J. Phys. Chem. 1984, 88, 2330–2334. (29) Ankudinov, A. L.; Rehr, J. J.; Low, J.; Bare, S. R. Phys. Rev. Lett. 2001, 86 (8), 1642–1645. (30) Murphy, L. R.; Meek, T. L.; Allred, A. L.; Allen, L. C. J. Phys. Chem. A 2000, 104, 5867–5871. (31) Watari, N.; Ohnishi, S. J. Phys. Chem. 1997, 106 (18), 7531–7540. (32) Shin, K.; Zhang, L.; An, H.; Ha, H.; Yoo, M.; Lee, H. M.; Henkelman, G.; Kim, H. Y. Nanoscale 2017, 9, 5244–5253. (33) Komatsu, T.; Takasaki, M.; Ozawa, K.; Furukawa, S.; Muramatsu, A. J. Phys. Chem. C 2013, 117 (20), 10483–10491. Model Catalyst characterization under reaction conditions 143