Design, characterization and structuring of Pt based catalysts for Water Gas Shift reaction

1 Design, characterization and structuring of Pt based catalysts for Water Gas Shift reaction Miriam González Castaño Ph.D. Thesis Directors Svetlana Ivanova Jose Antonio Odriozola Gordon 2 Contents General introduction and objectives ............................................................. 6 Chapter 1: Experimental techniques for characterization and catalytic activity ............................................................................................................ 42 1.1. Physicochemical characterization ........................................................ 43 1.1.1. X-Ray microfluorescence (XRF) ................................................ 43 1.1.2. Textural properties: specific surface area, pore volume and pore size. .......................................................................................................... 43 1.1.3. X-Ray diffraction (XRD) .............................................................. 45 1.1.4. Transmission and Scanning Electron Microscopy: TEM and SEM. ......................................................................................................... 47 1.1.5. Raman and Infrared (IR) spectroscopies ....................................... 48 1.1.6. UV-Vis spectroscopy (UV-Vis) .................................................... 50 1.1.7. Temperature programmed techniques: TPR, TPD and TPO ........ 51 1.1.8. Oxygen storage complete capacity (OSSC) and oxygen storage capacity (OSC) ................................................................................. 53 1.2. Catalytic activity .................................................................................. 54 Chapter 2: Synthesis and characterization of Pt based catalysts: catalytic screening ........................................................................................................ 57 2.1. Introduction .......................................................................................... 58 2.2. Synthesis of the catalysts ................................................................ 61 2.3. Physicochemical characterization of the samples ................................ 62 2.3.1. Textural properties and chemical composition ..................... 62 2.3.2. XRD analysis ........................................................................ 65 2.3.3. TPR studies ........................................................................... 68 2.4. Catalytic activity ............................................................................. 75 2.5. Partial conclusions .......................................................................... 77 3 Chapter 3: Electronic and structural promoters of Pt(4%)/CeO2/Al2O3 catalyst ............................................................................................................ 83 3.1. Introduction .......................................................................................... 84 3.2. Experimental ........................................................................................ 86 3.3. Support optimization for Pt based CexZr1-xO2/Al2O3 ...................... 87 3.3.1. Physicochemical characterization results. ..................................... 87 3.3.1.1. Textural properties and chemical composition of the prepared solids ................................................................................. 87 3.3.1.2. XRD analysis of the prepared solids .................................... 89 3.3.1.3. TPR characterization ............................................................ 92 3.3.2. Catalytic behavior. ........................................................................ 93 3.4. Structural and electronic effect of Zr and/or Fe addition over Pt-ceria catalysts ........................................................................................... 95 3.4.1. Physicochemical characterization of the prepared solids .............. 95 3.4.1.1. Textural properties and chemical composition ................... 95 3.4.1.2. XRD results ......................................................................... 97 3.4.1.3. CO chemisorption and HR-TEM microscopy. ................... 100 3.4.1.4. Raman and UV-Vis characterization .................................. 103 3.4.1.5. Oxygen Storage Capacity (OSC) and Oxygen Storage Complete Capacity (OSCC) .......................................................... 108 3.4.1.6. H2-TPR ............................................................................. 111 3.4.1.7. Temperature Programmed Desorption (TPD) experiment . 116 3.4.1.8. CO adsorption and CO-TPD followed by FTIR spectroscopy ....................................................................................................... 132 3.4.2. Catalytic behavior ....................................................................... 142 3.5. Partial conclusions .............................................................................. 145 Chapter 4: Structuring Pt(2%)/CeO2/Al2O3 WGS catalyst: introduction of a buffer layer ........................................................................................... 155 4.1. Introduction ........................................................................................ 156 4.2. Experimental ...................................................................................... 161 4 4.3. Physicochemical characterization results ........................................... 164 4.3.1. Textural properties and chemical composition of the prepared solids ...................................................................................................... 164 4.3.2. XRD of the prepared solids ......................................................... 165 4.3.3. SEM and adherence test. ............................................................. 166 4.4. Catalytic behavior .............................................................................. 168 4.5. Conclusions ........................................................................................ 189 Chapter 5: Structured WGS catalysts: Influence of the buffer nature on the water activation step ............................................................................. 196 5.1. Introduction ........................................................................................ 197 5.2. Experimental ...................................................................................... 199 5.2.1. Synthesis of the buffer solids and catalyst. Physical mixture preparation. ............................................................................................ 199 5.2.2. Washcoating process. .................................................................. 201 5.3. Physicochemical characterization ..................................................... 205 5.3.1. Textural properties and chemical composition of the prepared solids ...................................................................................................... 205 5.3.2. XRD analysis............................................................................... 208 5.3.3. Adherence test and SEM ............................................................ 210 5.3.4. IR spectroscopy and H2O-TPD .................................................. 211 5.4. Catalytic behavior .............................................................................. 220 5.5. Partial conclusions .............................................................................. 227 Chapter 6: Understanding the NM role in WGS reaction: Pt vs. Au ..... 232 6.1. Introduction ........................................................................................ 233 6.2-. Part I: Au and Pt powder catalysts over different supports for WGS reaction ..................................................................................................... 236 6.2.1-. Catalysts preparation .................................................................. 236 6.2.2-. Physicochemical characterization .............................................. 237 5 6.2.2.1-. Chemical composition and textural properties ................. 237 6.2.2.2-. XRD analysis .................................................................... 239 6.2.2.3-. UV-Vis and Raman spectroscopies ................................. 242 6.2.2.4. CO chemisorption and HR-TEM microscopy ................... 249 6.2.3. Catalytic activity ......................................................................... 250 6.2.4. Partial conclusions ....................................................................... 259 6.3. Part II: Au/CeFeAl and Pt/CeFeAl catalysts supported on metallic micromonolithic devices .......................................................................... 260 6.3.1. Experimental .............................................................................. 262 6.3.2. Physicochemical characterization ............................................... 264 6.3.2.1. Chemical composition and textural properties .................. 264 6.3.2.2. XRD analysis .................................................................... 265 6.3.2.3. SEM analysis and adherence test ...................................... 268 6.3.4. Catalytic behavior ...................................................................... 269 6.3.5. Partial conclusions ...................................................................... 276 6.4. Part III: Operando XANES-EXAFS study of Au and Pt ceria based catalysts .................................................................................................... 277 6.4.1-. Experimental: X-ray Absorption Spectroscopy ........................ 279 6.4.2. XANES results ............................................................................ 280 6.4.3. EXAFS results ............................................................................. 286 6.4.4. Partial conclusions ....................................................................... 304 General conclusions .................................................................................... 314 Resumen ....................................................................................................... 319 General introduction and objectives General introduction and objectives 7 Actual energetic system The growth of the environmental concern, in view of the global warming produced by the greenhouse gas emissions, and the predicted depletion of the energetic sources stimulated the scientific interest to search for new energyfriendly sources. The 20th century was characterized by a socio-economic enlargement being reflected on the increase of the total economic demand. The evolution of the energetic sources utilization through the years shows the total fossil fuels dependency (Figure 1) [1]. The fossil fuels based economic model demonstrates two main disadvantages: one is the geographical availability limitation derived from the unequal distribution of the sources, which are placed just in a few countries. The other important drawback concerns to the environmental consequences derived from their continuous use. Figure 1. Worldwide Energetic Mixture. Adapted from [1] General introduction and objectives 8 Therefore, it is mandatory to develop alternative energy models, as for example those based on H2 as a fuel. H2 technology is one of the most promising energy vectors to its sustainability and environmental responsibility [2]. On the other hand, H2 demand continuously increases in recent years mainly due to its application in petroleum refining processes, such as hydrotreating and hydrocracking [3], production of methanol [4], methanol- to-gasoline technology (MTG) [5], ammonia and hydrocarbon synthesis via Fischer-Tropsch processes [6]. However, H2 does not exist by itself in the nature and it must be produced. H2 containing streams can be achieved by using nuclear energy, water electrolysis or chemical methods. The most popular chemical H2 production routes are: - Gasification or partial oxidation (POX): It is defined as the reaction between a hydrocarbon or alcohol with sub-stoichiometric amount of oxygen that is needed for its complete oxidation (Equation 1.1). The tendency to total hydrocarbon’ oxidation must be controlled in order to obtain higher selectivity towards H2. Although the exothermicity of this reaction presents an advantage and no external energy sources are needed, the temperature reactor gradient and the catalyst hot spots formation should be controlled. In addition, the used oxygen usually is previously separated from air, which requires the use of expensive technologies. General introduction and objectives 9 𝐶𝑛𝐻𝑚𝑂𝑧 + 1 2(𝑛−𝑧)𝑂2 → 𝑛𝐶𝑂 + 𝑚 2𝐻2 (Equation 1) - Steam reforming (STR): Generally, it can be assumed as the reaction between hydrocarbon or alcohol and water to produce CO and H2 (Equation 1.2). Industrially, methane steam reforming is the most employed process to produce synthesis gas from natural gas. This reaction is highly endothermic being favored at higher temperatures and pressures. At industrial scale, it is normally carried out at 15-30 pressure bars and at 850-950 ºC employing a steam/carbon ratio around 3. 𝐶𝑛𝐻𝑚𝑂𝑧+(𝑛−𝑧)𝐻2𝑂 → 𝑛𝐶𝑂 + (𝑛 + 𝑚 2−𝑧)𝐻2 (Equation 2) - Dry reforming: In a general way, describes the reaction between a carbonaceous molecule and CO2 to obtain synthesis gas (Equation 1.3). This process has the advantage to transform CO2, a greenhouse gas, in other mixture with higher added value. Actually, methane dry reforming is considered as an option to profit the biogas. The negative aspect comes from reaction’s strong endothermic character requiring large amounts of energy, optimized installations and high operational costs. 𝑎𝐶𝑛𝐻𝑚+𝑏𝐶𝑂2 ↔ 𝑐 𝐶𝑂 +(𝑎·𝑚 2) 𝐻2 (Equation 3) General introduction and objectives 16 In general, the publications focused either on the WGS activity influence of the synthesis method, support modification, structural effects, oxidation state of the noble metal during the reaction or to the predominant mechanisms and reaction intermediates through which the reaction occurs. Several supports, classified as “reducible” and “non-reducible” oxides, have been studied for the Pt WGS catalysts. Olympiou et al. [19] studied the WGS reaction over Pt, Pd and Rh supported on irreducible alumina correlating the WGS activity exhibited to the surface concentration and binding energy of CO on the metal. The Pt > Rh > Pd order of activity was found and the WGS mechanism matches the one proposed by Grenoble et al. [20] for supports such alumina. In this mechanism, it was proposed that the water dissociation happens onto the support and the activated species migrates from the support to the noble metal to complete the reaction. However, none of these works consider the capacity of platinum itself to dissociate water [21]. Figure 4. Number of publications, in Scopus database using “WGS” and “Platinum”, as search strings for the 1995-2014 period General introduction and objectives 17 Support importance: ceria modifications Pt based catalysts and also other noble metals, exhibit an important enhancement of the catalytic activity exhibited when supported on reducible oxides [22–25]. The reducible oxides are not simply spectators in the WGS reaction but, in turn, play an important role [18,26–28]. Within the reducible oxides used as a supports the most studied are titania [8,26,29,30] and ceria [22,24,31–36]. Phatak et al. [37] reported for Pt supported on ceria WGS turnover rate 30 times higher than that of Pt alumina supported catalyst. In particular, Pt/CeO2 raised the scientific interest for WGS reaction, since in the early 1980s, during the development of the three-way catalysts (TWC), it was discovered that ceria is the best non-noble metal oxide promoter for a typical Pt–Pd–Rh/alumina TWC because it enhances the water-gas shift reaction [38]. The strong interaction between Pt and ceria, exemplified by the ceria surface oxygen reducibility upon metal addition, was recognized as indication of the electronic effect exercised by the oxidized Pt-Ox-Ce sites, the active sites in the TWC redox chemistry. Accordingly to Flytzani, this effect was even more evident after aging, leading to metal separation in two groups: large platinum particles (presumably inert) and non-metallic part, which are presumably atomic or subnanometric oxidized Pt-Ox clusters strongly bound to the ceria surface [39]. In this line, Tauster el al. [40] introduced the term strong-metal support interaction in order to describe the contribution of the reducible oxide support General introduction and objectives 18 to the catalytic site formation. It was discovered that some Pt supported on a reducible oxide catalysts the hydrogenation activity of the noble metal is decreased by structural effects: decoration or partial coverage of the active metal sites by reduced M-Ox species. Raman studies reported by Murrel et al. [41,42] showed that Pt form surface metal oxide structures (M-O) interacting strongly with ceria surface, being the latter present in bulk form or within an alumina host structure. Commonly, it is established that both, noble metal and support, play essential roles for the WGS reactions. It is also accepted that the catalytic activity of the supported metal particles is highly influenced by particle size, shape and metal oxidation state, as well as by the nature and reduction degree of the support and by the metal/oxide interaction. Initially, the main role of the support is to disperse the active phase. The support must be able to interact with the active phase in order to obtain highly dispersed and stable noble metal based catalysts. Typically, support modifications are undertaken to enhance the catalytic performance of the catalyst through changing its structural or electronic properties [18,43–46]. The synergy between the ceria and the modifier can be achieved in three different ways: solid solution formation [47,48], interaction between segregated oxides [49,50] or by metallic particles deposition on the ceria surface [51,52]. General introduction and objectives 19 Several studies have demonstrated that doping ceria with elements with different ionic radii or oxidation states induces effects on the ceria’ properties: reducibility, oxygen storage capacity and resistance to sintering [53,54]. Considering that partial reduction of ceria support is convenient for oxygen vacancies formation and activation of water during WGS reaction, the improvement of the latter should result in catalyst’ promotion and increased CO oxidation rate. The redox capacity of the samples should markedly determine the catalytic activity being the capacity to exchange oxygen species with the media, depending on the reaction atmosphere and the formation of Ce3+- Ov· sites, a feature closely related with the catalytic performance. Numerous studies regarding the doping agents’ introduction with unique purpose to increase the ceria redox properties and catalytic performance after metal addition could be found in the literature [26,45,50,55,56] with contradictory conclusions. For example, it was reported that the introduction of Zr (IV) to the ceria lattice, for gold based catalysts, worsens the WGS activity, an effect explained by the stabilization of cubic and tetragonal polymorphic ZrO2 forms in respect to the monoclinic one, the active to dissociate water [57]. On the contrary, the zirconia addition has been reported beneficial for ceria based catalysts specially when forming solid solution resulting in improved reducibilities and easier oxygen exchange with the media [56]. General introduction and objectives 20 In addition to the good redox properties, suitable ceria based support should also lower the barrier for ceria to metal oxygen transfer facilitating the oxidation of adsorbed CO. This influence of the ceria cations on the adsorption state of CO is related to its back-bonding possibility from the metal to the CO adsorbate. So, by varying the acidic properties of the catalyst and the CO-support interaction one can modulate the easiness of the system to react with the neighboring activated intermediate species. Duarte de Farias et al [58] reported that ceria-zirconia mixed oxide presents higher basic character than the pure ceria but are also more acid if Lewis sites are considered. They suggest that CO adsorbs preferably at Zrdefective center located at the metalsupport interface which interacts with the O2- and OH- to form reaction intermediates. Similar results has been published for other transition metals employed as doping agents [59,60] being the catalytic performances more frequently related to the redox properties modification. Deactivation of the catalysts The use of noble metal catalysts, more expensive than conventional LTS catalysts based on transition metal oxides, could be considered economically viable if they present a higher resistance to deactivation. Different catalytic behavior as a function of the particle size of the noble metal has been reported being more active the more dispersed noble metal because of their higher surface/volume ratio and/or metal-support interaction. General introduction and objectives 21 Indeed the noble metal sintering has been reported as one of the main causes of deactivation for platinum based catalysts [61–63]. This phenomenon could be avoided either by support modifications in order to achieve higher noble metal-support interactions and as a consequence higher platinum dispersion [31,64] or by changes in the method of preparation [65]. Moreover, in the case of ceria supported catalysts, support overreduction and carbonaceous species active sites blocking are also reported as important deactivation causes. The overreduction phenomena imply that the major proportion of the cerium oxide is completely reduced and therefore, the pair redox Ce4+/Ce3+ is not available for the catalytic process. The impossibility of the ceria to be reoxidized by the reactant water reduces drastically the catalytic activity [66]. On the other hand, the carbonaceous species can be also formed from the reaction intermediates hindering the WGS reaction active sites. From, the introduction of ceria provided catalysts with much better properties than the original Pt/Al2O3 systems. The strong interaction of Pt/Ceria, exemplified by the enhanced reducibility of the surface oxygen of ceria upon the metal addition was recognized as an indication of an electronic effect displayed by oxidized sites (Pt-Ox-Ce) which are probably responsible of the redox chemistry. Besides, further advances were achieved through the addition of promoters able to diminish the platinum sintering that provokes the catalyst deactivation. General introduction and objectives 22 The described features characteristic of platinum based systems converted the proper catalyst design in a tremendously interesting approach. Metal-support interactions: WGS mechanism and active species In principle, the structural and electronic properties of the catalysts and the WGS reaction mechanism should be connected. The structural and electronic properties of the active sites are usually affected by the chemical environment and do affect the creation and transformation of the formed species through a reaction mechanism [67–70]. Therefore, the design and optimization of metaloxide supported WGS catalyst requires further knowledge about how the reactants, intermediates and products interact with the active phase. The importance of the noble metal-support interactions and its essential role for the catalytic activity has received much attention reflected in the development of numerous mechanistic and kinetics studies. To date, some disagreement remains about the nature of the reaction intermediates and the role of the support. Indeed, the chemical nature of the intermediates and their true localization, e. g. support, metal-support interface, or metal surface still remain controversial. Concerning the mechanism of the WGS reaction, two main reaction pathways has been generally proposed in the literature. In the proposed redox mechanism [71], CO adsorbs on the metal surface and then diffuses to the metal-support interface and reacts with the lattice oxygen of the support to General introduction and objectives 23 form CO2. Water adsorbs and dissociates mainly on the reduced support sites, thus re-oxidizing them and subsequently producing H2. For the redox pathway, the successive oxidation and reduction steps occur as follows: CO + * ↔ *CO (Equation 6) H2O + * ↔ H2O* (Equation 7) H2O* + * ↔ OH* + H* (Equation 8) 2 HO* + * ↔ H2O* + O* (Equation 9) HO* + * ↔ O* + H* (Equation 10) CO* + *O ↔ CO2* + 2* (Equation 11) 2H* ↔ H2 + 2* (Equation 12) Where “*” represents a surface site, and therefore, “X*” stands for adsorbate bonded to this site. Within the associative mechanism, H2O and CO adsorbs on catalyst active sites to form surface “carbon-containing” intermediates which react or decompose to form CO2 and H2. The associative mechanism proposed by Iwasawa and Shido [72] happens as follows and it involves the reaction through adsorbed intermediate species such as formate, carbonate or carboxyl: CO + * ↔ *CO (Equation 6) H2O + 2 * ↔ OH* + H* (Equation 13) OH* + *CO ↔ HOOC** (Equation14a) OH* + *CO ↔ OCOH* + * (Equation 14b) General introduction and objectives 24 Which eventually dissociates HOOC** ↔ CO2 + H* + * (Equation 15a) OCOH* ↔ CO2 + H* (Equation 15b) Or reacts with OH* HOOC** + *OH ↔ CO2 + H2O* + 2* (Equation 16a) OCOH* + *OH ↔ CO2 + H2O* + * (Equation 16b) Mechanism and active species for Pt catalysts In literature, discussion for and against both mechanisms could be found. For example, Kalamaras et al. [73] concluded that the WGS reaction on Pt/CeO2 catalyst at 300ºC occurs mainly through redox mechanism with a minor contribution of the associative via formate mechanism. On the contrary, Jacobs et al. [33] proposed an associative formate pathway, using IR spectroscopy suggesting as catalytically active sites the Type II bridging OH groups on partially reduced ceria which reacts with CO and generates surface formate. The formates decomposition is then promoted by water molecules and platinum metal, caused by the higher oxygen vacancies concentration around platinum able to dissociate water and decompose the formates. General introduction and objectives 25 Solid arguments against the formate, as main reaction intermediates, were reported by Rodriguez et al. [74]. In this study by DRIFTS spectroscopy, it was observed that the temperature of maximum formate concentration (250ºC) do not correlate with the highest activity temperature found for Pt/CeO2. Actually, the maximum CO conversion was found at 300-350 ºC when the amount of detected formates was negligible. In concordance, Burch [75] affirmed that the formate species appear to be mere spectators not involved in the reaction on the basis of the differences observed between formate decomposition and WGS reaction rates. Gabrow et al.[76] reported after DFT calculations, as a possible reaction pathway for Pt (111) surface, the associative mechanism through carboxyl intermediate species, under the basis of more favorable energetic formation. The intermediates decomposition as described above was reported to be encouraged by OH species. In the Figure 5, all possible reaction pathways are presented. The Ea indicates the forward activation energy barrier and the ΔE, both in eV. The path involving the smallest activation energy barriers is highlighted. The difficulty to establish an unequivocal mechanism was analyzed by Burch et al. [75] in an interesting study, in which the large amount of published ambiguous results and the limitations is discussed. General introduction and objectives 32 Actually, the performance and durability of the micromonolithic catalysts strongly depend on the layer quality. Several methods through which a catalyst layer can be deposited over the monolithic walls as sol-gel, electrophoretic deposition (EPD), electrochemical deposition and electroless plating, etc [82]. Among them, the coating from a slurry (or suspension) has been the most employed and it is commonly labeled washcoating method. It is usually carried out with a slurry of particles of a comparable size to that of the macropores of the support [83]. Indeed, a proper performed washcoating method, and this implies a careful control of the suspension properties, allows achieving uniform catalyst layers. Apart of the catalysts, there are numerous other materials which can act as an additive, binder or adhesive to the washcoat slurry. Key steps in slurry preparation are particle size reduction and addition of appropriate acid/sol to adjust the pH, viscosity and homogeneity of the slurry in water. The above described skins make the metallic micromonoliths, in particular the FeCrAlloy based ones, more convenient than the ceramic ones for design of the structured WGS catalysts. For all described above, this PhD thesis will be focused on the development of a highly active catalyst with real potential applicability for a portable WGS process. For this purpose, the catalyst design should passes through attaining improved redox properties, platinum dispersions, water activation capacities, General introduction and objectives 33 all together within the general aim of reaching enhanced catalytic performances for long-term stabilities. The realization of the proposed goal has been approached from two different perspectives. By one hand, the physicochemical properties of the Pt ceria based catalysts were modified, and the resulted solids characterized and tested in WGS reactions conditions in order to discern the catalyst composition influence on its WGS activity. On the other hand, the necessity of structured WGS catalyst for portable fuel processor application is considered and several structured platinum ceria systems were investigated. This thesis is organized in 6 chapters. Along the first one, the techniques as well as the instrumentation employed to perform the experiments are presented on the Chapter 1. The Chapter 2, the physicochemical properties of platinum based catalysts supported on Al2O3 and CeO2/Al2O3 commercial supports are evaluated through general characterization and catalytic tests. The catalysts supported on “home-made” CeO2/Al2O3 support are also studied. In the Chapter 3, starting from a chosen platinum ceria based sample, the effect of support modifications is analyzed. Different doping agents on the ceria lattice are introduced with purpose to enhance catalyst’s activity and stability. In chapter 4, the suitability of structured Pt based catalysts is evaluated. A new concept is also introduced in this chapter e.g. the presence of pre-catalyst layerbuffer layer. The better catalytic performances exhibited General introduction and objectives 34 by the bi-layer micromonolith systems motivates the studies described in Chapter 5. Finally, in the chapter 6, it is presented an interesting comparison among Pt and Au based catalysts being mainly focused on the noble metal nature features depending on the conditions in which those are tested. For further understanding above the different catalytic features between Au and Pt, those were extensively studied by EXAFS and XANES techniques providing fundamental knowledge above the platinum and support structures in WGS conditions. It should be mentioned that this Thesis is developed under the International PhD normative. According to it, a brief summary containing some introduction, the objectives, the main results and the conclusions will be also presented in Spanish. Simultaneously, some parallel studies were performed under industrial contract supported by “Técnicas Reunidas” which results are not presented on this Thesis because of confidentiality. Finally, the experimental section and the already published papers and patent in ISI indexed journals are included in annexes at the end of the present PhD Thesis. General introduction and objectives 35 References [1] M.J. Economides, D.A. Wood, J. Nat. Gas Sci. Eng. 1 (2009) 1. [2] I.H. Son, A.M. Lane, Catal. Letters 76 (2001) 151. [3] D.C. Elliott, Energy and Fuels 21 (2007) 1792. [4] C.N. Hamelinck, A.P.C. Faaij, J. Power Sources 111 (2002) 1. [5] M. Stöcker, Microporous Mesoporous Mater. 29 (1999) 3. [6] M.E. Dry, Catal. Today 71 (2002) 227. [7] Y. Wang, K.S. Chen, J. Mishler, S.C. Cho, X.C. Adroher, Appl. Energy 88 (2011) 981. [8] C. Ratnasamy, J.P. Wagner, Catal. Rev. 51 (2009) 325. [9] A.R. Dubrovskiy, E. V. Rebrov, S.A. Kuznetsov, J.C. Schouten, Catal. Today 147 (2009) 198. [10] T. Giroux, S. Hwang, Y. Liu, W. Ruettinger, L. Shore, Appl. Catal. B Environ. 56 (2005) 95. [11] S.M. Opalka, T.H. Vanderspurt, R. Radhakrishnan, Y. She, R.R. Willigan, J. Phys. Condens. Matter 20 (2008) 064237. [12] D.L. Trimm, Appl. Catal. A Gen. 296 (2005) 1. [13] M.A. Edwards, D.M. Whittle, C. Rhodes, A.M. Ward, D. Rohan, M.D. Shannon, G.J. Hutchings, C.J. Kiely, Phys. Chem. Chem. Phys. 4 (2002) 3902. General introduction and objectives 36 [14] C. Rhodes, G.J. Hutchings, A.M. Ward, Catal. Today 23 (1995) 43. [15] M.S. Spencer, Top. Catal. 8 (1999) 259. [16] G. Jacobs, S. Ricote, P.M. Patterson, U.M. Graham, A. Dozier, S. Khalid, E. Rhodus, B.H. Davis, Appl. Catal. A Gen. 292 (2005) 229. [17] G. Jacobs, S. Ricote, B.H. Davis, Appl. Catal. A Gen. 302 (2006) 14. [18] M. Gonzalez Castaño, T.R. Reina, S. Ivanova, M.A. Centeno, J.A. Odriozola, J. Catal. 314 (2014) 1. [19] G.G. Olympiou, C.M. Kalamaras, C.D. Zeinalipour-Yazdi, A.M. Efstathiou, Catal. Today 127 (2007) 304. [20] D.C. Grenoble, M.M. Estadt, D.F. Ollis, J. Catal. 67 (1981) 90. [21] S. Seong, A.B. Anderson, 3654 (1996) 11744. [22] A. Martinez-Arias, J.M. Coronado, R. Cataluna, J.C. Conesa, J. Soria, J. Phys. Chem. B 102 (1998) 4357. [23] S. Damyanova, J.M.C. Bueno, Appl. Catal. A Gen. 253 (2003) 135. [24] A.P. Ferreira, J.C.S. Araújo, J.W.C. Liberatoria, S. Damyanova, D. Zanchetc, F.B. Noronha, J.M.C. Bueno, Stud. Surf. Sci. Catal. 167 (2007) 433. [25] J.Z. Shyu, K. Otto, J. Catal. 115 (1989) 16. [26] G. Dutta, U. V. Waghmare, T. Baidya, M.S. Hegde, K.R. Priolkar, P.R. Sarode, Chem. Mater. 18 (2006) 3249. General introduction and objectives 37 [27] C.I. Vignatti, M.S. Avila, C.R. Apesteguía, T.F. Garetto, Catal. Today 171 (2011) 297. [28] G. Jacobs, S. Ricote, U.M. Graham, B.H. Davis, (2014). [29] V. Idakiev, Z.-Y. Yuan, T. Tabakova, B.-L. Su, Appl. Catal. A Gen. 281 (2005) 149. [30] J.A. Rodriguez, S. Ma, P. Liu, J. Hrbek, J. Evans, M. Pérez, Science 318 (2007) 1757. [31] P. Panagiotopoulou, J. Papavasiliou, G. Avgouropoulos, T. Ioannides, D.I. Kondarides, Chem. Eng. J. 134 (2007) 16. [32] K.R. Hwang, S.K. Ihm, J.S. Park, Fuel Process. Technol. 91 (2010) 729. [33] G. Jacobs, U.M. Graham, E. Chenu, P.M. Patterson, A. Dozier, B.H. Davis, J. Catal. 229 (2005) 499. [34] G. Jacobs, L. Williams, U. Graham, D. Sparks, B.H. Davis, J. Phys. Chem. B 107 (2003) 10398. [35] G. Jacobs, E. Chenu, P.M. Patterson, L. Williams, D. Sparks, G. Thomas, B.H. Davis, Appl. Catal. A Gen. 258 (2004) 203. [36] G. Jacobs, P.M. Patterson, L. Williams, E. Chenu, D. Sparks, G. Thomas, B.H. Davis, Appl. Catal. A Gen. 262 (2004) 177. [37] A.A. Phatak, N. Koryabkina, S. Rai, J.L. Ratts, W. Ruettinger, R.J. Farrauto, G.E. Blau, W.N. Delgass, F.H. Ribeiro, Catal. Today 123 (2007) 224. [38] G. Kim, Ind. Eng. Chem. Prod. Res. Dev. 21 (1982) 267. General introduction and objectives 38 [39] M. Flytzani-Stephanopoulos, B.C. Gates, Annu. Rev. Chem. Biomol. Eng. 3 (2012) 545. [40] S.J. Tauster, S.C. Fung, R.L. Garten, J. Am. Chem. Soc. 100 (1978) 170. [41] L.L. Murrell, S.J. Tauster, D.R. Anderson, e 10th International Congress on Catalysis Budapest (1993), pp. 19–24. [42] L.L. Murrel, S. J. Tauster, D. R. Anderson Catal. Automot. Pollut. Control II (1991) 275. [43] J. Kugai, J.T. Miller, N. Guo, C. Song, J. Catal. 277 (2011) 46. [44] P.C.A. Brito, D.A.A. Santos, J.G.S. Duque, M.A. Macêdo, Phys. B Condens. Matter 405 (2010) 1821. [45] C.T. Campbell, C.H.F. Peden, Science 309 (2005) 713. [46] N. Acerbi, S. Golunski, S.C. Tsang, H. Daly, C. Hardacre, R. Smith, P. Collier, J. Phys. Chem. C 116 (2012) 13569. [47] F.J. Pérez-Alonso, M. Lopez Granados, M. Ojeda, P. Terreros, S. Rojas, T. Herranz, J.L.G. Fierro, Chem. Mater. 17 (2005) 2329. [48] G. Xiao, S. Li, H. Li, L. Chen, Microporous Mesoporous Mater. 120 (2009) 426. [49] G. Avgouropoulos, T. Ioannides, J. Mol. Catal. A Chem. 296 (2008) 47. [50] A.S. Reddy, C.Y. Chen, C.C. Chen, S.H. Chien, C.J. Lin, K.H. Lin, C.L. Chen, S.C. Chang, J. Mol. Catal. A Chem. 318 (2010) 60. General introduction and objectives 39 [51] T. Tabakova, L. Ilieva, I. Ivanov, R. Zanella, J.W. Sobczak, W. Lisowski, Z. Kaszkur, D. Andreeva, Appl. Catal. B Environ. 136-137 (2013) 70. [52] T. Tabakova, V. Idakiev, D. Andreeva, I. Mitov, Appl. Catal. A Gen. 202 (2000) 91. [53] T.R. Reina, S. Ivanova, J.J. Delgado, I. Ivanov, V. Idakiev, T. Tabakova, M.A. Centeno, J.A. Odriozola, ChemCatChem 6 (2014) 1401. [54] O.H. Laguna, M.A. Centeno, M. Boutonnet, J.A. Odriozola, Appl. Catal. B Environ. 106 (2011) 621. [55] B.M. Reddy, A. Khan, Catal. Surv. from Asia 9 (2005) 155. [56] S. Damyanova, B. Pawelec, K. Arishtirova, M.V.M. Huerta, J.L.G. Fierro, Appl. Catal. A Gen. 337 (2008) 86. [57] J. Li, J. Chen, W. Song, J. Liu, W. Shen, Appl. Catal. A Gen. 334 (2008) 321. [58] A.M. Duarte de Farias, D. Nguyen-Thanh, M. a. Fraga, Appl. Catal. B Environ. 93 (2010) 250. [59] G. Lafaye, J. Barbier, D. Duprez, Catal. Today (2015) 1. [60] Y. Peng, W. Yu, W. Su, X. Huang, J. Li, Catal. Today 242 (2015) 300. [61] S.Y. Choung, M. Ferrandon, T. Krause, 99 (2005) 257. [62] W. Ruettinger, O. Ilinich, R.J. Farrauto, J. Power Sources 118 (2003) 61. [63] X. Wang, R.J. Gorte, J.P. Wagner, J. Catal. 212 (2002) 225. General introduction and objectives 40 [64] Y. Nagai, K. Dohmae, Y. Ikeda, N. Takagi, T. Tanabe, N. Hara, G. Guilera, S. Pascarelli, M.A. Newton, O. Kuno, H. Jiang, (2008) 9303. [65] M. El Doukkali, A. Iriondo, P.L. Arias, J. Requies, I. Gandarías, L. Jalowiecki-Duhamel, F. Dumeignil, Appl. Catal. B Environ. 125 (2012) 516. [66] T.R. Reina, S. Ivanova, V. Idakiev, J.J. Delgado, I. Ivanov, T. Tabakova, M.A. Centeno, J.A. Odriozola, Catal. Sci. Technol. 3 (2013) 779. [67] F.F. Tao, M. Salmeron, Science 331 (2011) 171. [68] C.T. Campbell, Top. Catal. 1 (1994) 353. [69] G.A. Somorjai, Langmuir 7 (1991) 3176. [70] J.K. Nørskov, T. Bligaard, B. Hvolbaek, F. Abild-Pedersen, I. Chorkendorff, C.H. Christensen, Chem. Soc. Rev. 37 (2008) 2163. [71] P. Liu, J.A. Rodriguez, J. Chem. Phys. 126 (2007) 164705. [72] T. Shido, Y. Iwasawa, J. Catal. 141 (1993) 71. [73] C.M. Kalamaras, S. Americanou, A.M. Efstathiou, J. Catal. 279 (2011) 287. [74] W. Xu, R. Si, S.D. Senanayake, J. Llorca, H. Idriss, D. Stacchiola, J.C. Hanson, J.A. Rodriguez, J. Catal. 291 (2012) 117. [75] R. Burch, A. Goguet, F.C. Meunier, Appl. Catal. A Gen. 409- 410 (2011) 3. General introduction and objectives 41 [76] L.C. Grabow, A.A. Gokhale, S.T. Evans, J.A. Dumesic, M. Mavrikakis, J. Phys. Chem. C 112 (2008) 4608. [77] J.A. Rodriguez, J.C. Hanson, P.J. Chupas, In Situ Characterization of Heterogeneous Catalysts, wiley, 2013. [78] A. Bourane, O. Dulaurent, D. Bianchi, J. Catal. 196 (2000) 115. [79] D.I. Panagiotopoulou, P. Kondarides, J. Catal. 225 (2004) 327. [80] K. Shyu, J Z, Otto, Appl. Surf. Sci. 32 (1988) 246. [81] Q. Fu, H. Saltsburg, M. Flytzani-Stephanopoulos, Science 301 (2003) 935. [82] V. Meille, Appl. Catal. A Gen. 315 (2006) 1. [83] P. Avila, M. Montes, E.E. Miró, Chem. Eng. J. 109 (2005) 11. Experimental techniques for characterization and catalytic activity Chapter 1 48 Additionally, both techniques allow the study of the chemical composition through Energy dispersive X-Ray spectroscopy (EDX) from the radiation emitted by the sample when it is excited with electromagnetic radiation. The High-Resolution Transmission Electron Microscopy (HRTEM) and High Angle Annular Dark Field-Scanning Transmission Electron Microscopy (HAADF-STEM) images presented on this Thesis were recorded on a JEOL2010F instrument. The spatial resolution at Scherzer defocus conditions in HRTEM mode is 0.19 nm meanwhile the HAADF-STEM studies were performed using an electron probe of 0.5 nm of diameter and diffraction camera length of 10 cm. The employed EDX instrument was Oxford Instrument, Inca Energy-200. Loose power samples were supported on a holey carbon-coated copper grid without using any solvents. The excess of the sample was removed with N2. This approach allows obtaining electron transparent ultrathin regions. On the other hand, the SEM microscopies presented in this memory were obtained in Hitachi® S-4800 SEM-FEG provided with secondary and retrodispersed electron detectors. 1.1.5. Raman and Infrared (IR) spectroscopies The energy of the molecules can be associated mainly to three components: i) the electron movement, ii) the vibration of the constituent atoms and iii) the rotation of the molecule. This classification obeys to the different time scales Experimental techniques for characterization and catalytic activity Chapter 1 49 in which the each kind of vibration happens being the electronic vibration significantly faster than the rotational ones. The vibrational transitions observed in Raman or Infrared spectroscopy can be obtained after the excitation of a molecule with an adequate radiation source in a long wave range (λ) comprised between 10-2 and 10-4. The observed vibrational transitions proceed from the atomic cores which constitute the molecules. However, no all the transitions are allowed but for each molecule the transition number depends on its symmetry and on the selection rules. Despite of the different physical origin of Raman and IR spectroscopies, both techniques provide structural information of the studied sample because of the relation between the observed bands in the respective spectra and their molecular symmetry. The IR adsorption spectra are produced by photons in the IR region which are absorbed by two vibrational levels of the molecule in its energetic basal state. On the other hand, Raman spectra are originated because of the polarization caused after the sample irradiation with energies (ultraviolet, visible or near-infrared). So, the Raman experiment were carried out in a dispersive microscopy Horiva Jobin Yvon® (HR800) with a confocal aperture of 1000 m, using an objective 50x and a green laser (λ = 532.14 nm). The laser potency was reduced to 5 mW through an optical density filter D0.6. The IR spectroscopies were performed in transmission mode being the IR radiation passed through the sample. The samples were pilled without any Experimental techniques for characterization and catalytic activity Chapter 1 50 diluent presenting a diameter of 15 mm. FTIR spectra were recorded using THERMO NICOLET Avatar 380 FT-IR Spectrophotometer, equipped with a DTGS/ KBr detector, and accumulating 128 scans at a spectral resolution of 4 cm−1. The experiments were performed in situ using a purpose-made IR cell connected to a conventional vacuum adsorption apparatus with a residual pressure lower than 10−5 mbar. 1.1.6. UV-Vis spectroscopy (UV-Vis) Ultraviolet-visible spectroscopy (UV-Vis) refers to absorption or reflectance spectroscopy in the ultraviolet-visible spectral range corresponding to electronic transitions in the measures samples. This implies, that r uses light in the visible and adjacent regions, near-UV and near-infrared (NIR). The absorption or reflectance in the visible range directly affects the perceived color of the samples. This technique is complementary to the fluorescence spectroscopy. Meanwhile fluorescence deals with the transitions from the excited to the ground state, the absorption measurements transitions from the ground to the excited state. There three main transitions that can be considered: - d-d transitions: They are typical of transition metals ions. These electronic transitions principally give rise to adsorption in the visible range and are therefore responsible for the color of transition metal complexes. Experimental techniques for characterization and catalytic activity Chapter 1 51 - Charge-transfer transitions: They involve an electronic transfer from the occupied orbital towards empty ones. - π – n* and n – π* transitions: They normally occur on organic molecules due to the electronic transitions between n and π orbitals. The UV-Vis spectra were recorder on a Varian spectrometer model Cary 100, equipped with integrating sphere and using BaSO4 as reference. All the spectra were recorded in a diffuse reflectance mode and transformed to a magnitude proportional to the extinction coefficient through the Kubelka- Munk function F(R∞)[2]. 𝐹(𝑅∞)= (1−𝑅∞)2 2 𝑅∞∝𝜅 𝜎 Equation 1.3 In practice, R∞, is the reflectance of a layer so thick that further increase of the thickness does not change the reflectance. κ is the absorbance in cm-1 and σ is the dispersion factor that it is assumed to be independent of the wavelength for the particles with grain size higher than the light wavelength. 1.1.7. Temperature programmed techniques: TPR, TPD and TPO TPD studies are very useful in catalysts characterization providing useful information above the surface properties which will depend on the gas composition employed and the operative conditions. For the sample analysis, the sample is exposed to certain as stream and the temperature is increased following a temperature program. The TPD experiments were carried out in a PID Eng &Tech using a conventional quartz reactor. Experimental techniques for characterization and catalytic activity Chapter 1 52 The Temperature Programed Reduction (TPR) experiments provide information about the reducibility of the species [3] and they were performed using 5 % H2 in Ar and the temperature was increased at 10 ºC/min. The feed stream is bifurcated in two lines; one of them is passed through the sample contrary to the other one which is employed as reference. The difference between the H2 inlet and the outlet is the measured as a function of the temperature by using a thermal conductivity detector (TCD). A molecular sieve 13X was used to retain the H2O produced during the reduction. For the quantitative analysis the TCD signal was calibrated with a CuO pattern (Strem Chemicals 99.999%). The Temperature Programmed Desorption/Oxidation (TPD/TPO) allows obtain information related to the nature of the adsorbed species [4]. TPD /TPO measurements were carried out at atmospheric pressure in a conventional U-shape quartz reactor. The gas composition was analyzed at the exit of the reactor by a mass spectrometer PFEIFFER Vacuum PrismaPlus controlled by the program Quadera®. For the TPD experiments, the reactive gas stream employed flushed through the sample was Argon 99.999 (Air Liquid) meanwhile a 5 % O2 (Air Liquid) balanced in Ar was used for the TPO tests. Experimental techniques for characterization and catalytic activity Chapter 1 53 1.1.8. Oxygen storage complete capacity (OSSC) and oxygen storage capacity (OSC) Surface migration of oxygen can be of utmost importance in oxidation processes in which the catalysis are submitted to repeated changes in the gas composition [5]. In this sense, the Oxygen storage complete capacity (OSSC) and the Oxygen storage capacity (OSC) provide very useful information about the redox properties of the samples. For the OSCC 100 mg of catalysts was loaded and firstly oxidized into a U- shaped quartz reactor using a 5% O2/Ar flow until 350 ºC. Then, the sampled is set to the desired temperature and submitted to ten equivalent CO pulses. The OSCC is obtained from the total CO2 produced in all CO pulses. The samples are degassed in Ar flow and subjected to a new series of oxidizing (ten O2 pulses). For the OSC experiment, the sample is exposed to 5 alternating pulses (O2- CO-O2-CO-O2-CO). The OSC is determined by the average amount of CO2 per pulse formed after the first CO pulse of the alternated ones. This method is based on the one presented by Duprez et al [5,6]. The gas composition was analyzed by a mass spectrometer PFEIFFER Vacuum PrismaPlus controlled by Quadera® software. Experimental techniques for characterization and catalytic activity Chapter 1 54 1.2. Catalytic activity The Water Gas Shift reaction was performed using the home-made equipment shown in the Figure 1.2. The temperature range was raised in the 180-350 ºC range using two reaction conditions labeled model or ideal and realistic conditions. For the model WGS reaction just water and CO diluted in N2 were used while for the realistic WGS reaction a reformate gas surrogate containing H2 and CO2. The later conditions were chosen in order to simulate the outlet of a typical ethanol reformer and they are also known as CO clean-up conditions. For the entire catalytic test, the solids were sieved and the 600-800 m fraction retained. The GHSV was estimated as the ratio feed flow to catalyst bed volume and expressed in h-1. The CO and CO2 content was analyzed with Figure 1.2. Experimental set-up for the WGS reaction Experimental techniques for characterization and catalytic activity Chapter 1 55 an ABB gas analyzer and the activity expressed in terms of CO conversion. The CO conversion was estimated according to the equation 1.4. 𝐶𝑂 𝑐𝑜𝑛𝑣𝑒𝑟𝑠𝑖𝑜𝑛 (%)= 𝐶𝑂𝑖𝑛−𝐶𝑂𝑜𝑢𝑡 𝐶𝑂𝑖𝑛 𝑥 100 Equation 1. 4 Experimental techniques for characterization and catalytic activity Chapter 1 56 References [1] G. Leofanti, M. Padovan, G. Tozzola, B. Venturelli, Catal. Today 41 (1998) 207. [2] H.G. Völz, 13 (1985) 153. [3] S.D. Robertson, B.D. McNicol, J.H. De Baas, S.C. Kloet, J.W. Jenkins, J. Catal. 37 (1975) 424. [4] D.P. Woodruff, T.A. Delchar, (1986). [5] S. Kacimi, J.B. Jr, R. Taha, D. Duprez, Catal. Letters 22 (1993) 343. [6] S. Royer, D. Duprez, ChemCatChem 3 (2011) 24. Chapter 2 Synthesis and characterization of Pt based catalysts: catalytic screening Summary In this chapter, the synthesis parameters are described, e.g. platinum deposition method and loadings and support nature. The goal was to select the features for the starting catalyst of the thesis. With that purpose, platinum was deposited on two different commercial supports: a high surface -Al2O3 and on a CeO2/Al2O3 (Puralox) both supplied by Sasol. For each support, two platinum contents were chosen: 2wt. % and 4wt. %. A homemade CeO2/Al2O3 support with its corresponding catalyst (4% platinum content) was also prepared. The catalysts were tested in WGS reaction and some general characterization techniques such as XRMF, SBET, XRD and TPR were employed to comprehend the observed catalytic behavior and to relate it to the support nature and platinum content. Synthesis and characterization of Pt based catalysts: catalytic screening Chapter 2 64 comparison between the commercial supports, it can be observed narrower pore size distribution for the pure alumina containing solids. When Pt is deposited on alumina, no matter the content, the pore size distribution does not change, being the maximum placed at around 7.5 nm. Contrary to the alumina supported catalysts, the platinum loadings increase modifies the pore size distribution when Puralox is employed as a support. The pore size distribution becomes wider and the pore size maximum is shifted to higher pore diameter values, when platinum content changes from 2 wt.% to 4 wt.%. Even so, all the prepared solids maintain their mesoporous character. The chemical composition determined by -XRF is listed in Table 2.2. Although very close to the desired nominal values the chemical analysis shows some variation in the platinum contents. Figure 2.3: Pore size distribution of the prepared solids. 0 5 10 15 20 25 30 35 40 0,0 0,5 1,0 1,5 2,0 2,5 3,0 dV/dlog(D)PoreVolume (cm3/g) Pore Diameter (nm) Pt(2%)/Al2O3 Pt(4%)/Al2O3 510 15 20 25 30 35 40 0,0 0,4 0,8 1,2 1,6 dV/dlog(D) Pore Volume (cm3/gA) Pore Diameter (nm) Pt(2%)/Puralox Pt(4%)/Puralox Pt(4%)/CeAl HM Synthesis and characterization of Pt based catalysts: catalytic screening Chapter 2 65 All the prepared samples present chemical compositions very close to the targeted ones indicating the suitability of the synthesis method employed either for the commercial support or for the homemade one. 2.3.2. XRD analysis The XRD diffraction patterns for alumina and ceria based catalysts are presented in the Figure 2.4.a and 2.4.b respectively, being the corresponding primary support also included. All the characterized solids were highly amorphous as detected from the broad signals exhibited in all the cases. The employed alumina shows the characteristic diffractions of the -Al2O3 phase (JCPDS: 00-050-0741). Regarding to the alumina supports, the platinum incorporation does not lead to any important changes in the registered diffractograms. No platinum diffractions were detected which indicates that Table 2.2: Chemical composition of the prepared solids. % wt. Al2O3 % wt. CeO2 % wt. Pt Al2O3 100 - - Puralox 79.3 20.7 - CeAl HM 80.1 19.9 - Pt(2%)/Al2O 3 98.3 - 1.7 Pt(4%)/Al2O 3 96.5 - 3.5 Pt(2%)/Puralox 79.4 18.8 1.8 Pt(4%)/Puralox 77.5 18.1 4.4 Pt(4%)/CeAl HM 76.5 19.7 3.8 Synthesis and characterization of Pt based catalysts: catalytic screening Chapter 2 66 the platinum phase should be well dispersed with sizes lower than the detection limit of the technics, around 4 nm. Figure 2.4: XRD diffraction patterns: a) -Al2O3 support and alumina supported platinum catalysts; b) CeO2-Al2O3 supports and ceria supported platinum based catalysts In Figure 2.4.b, it can be observed the typical diffractions corresponding to - Al2O3 (JCPDS: 00-050-0741) and CeO2 (JCPDS: 00-004-0593). The ceria lattice parameters and crystallite size were estimated by using 𝑎 = √ℎ2+ 𝑘2+ 𝑙2· ( 𝜆 2 𝑠𝑖𝑛𝜃) and Scherrer`s equation respectively (Table 2.3). Both calculations were carried out considering the (111) ceria diffraction above 2= 28.5º (where no contribution of -Al2O3 phase is observed). Both supports, commercial or home-made, present a similar CeO2 lattice parameter, typical for those kind of solids [27] and different crystallite size, being the home-made support slightly larger in size. 10 20 30 40 50 60 70 80 90 -Al2O3 Pt(4%)/Al2O3 Pt(2%)/Al2O3 R. I. (a. u.) 2 Al2O3 10 20 30 40 50 60 70 80 90 -Al2O3 CeO2 Pt(2%)/Puralox Pt(4%)/CeAl HM SCeAl HM Puralox R. I. (a. u.) 2 Pt(4%)/Puralox Pt Synthesis and characterization of Pt based catalysts: catalytic screening Chapter 2 67 Moreover, the effect of platinum addition on the ceria phase was analyzed. When lower platinum loadings are employed, Pt(2%)/Puralox, no modifications for ceria lattice parameter are detected. On the contrary, for higher platinum contents lattice parameter diminution is observed. Regarding the ceria crystallite size, minor increment for all the systems is observed with the incorporation of platinum. No diffraction lines corresponding to PtOx species are observed for the lower platinum loading. However the increase of the platinum loading up to 4 wt. %, leads to signal corresponding to metallic platinum at around 2= 39.5º. This suggests: i) partial platinum reduction from Pt2+ to Pt0 happens during the calcination step and ii) at least, the presence of Pt0 fraction with particle size larger than 5 nm. Besides, this peak intensity appears to depend on the support, being significantly lower in the case of the CeAl HM support. The Pt particles size was estimated by Scherrer’s equation and presented in Table Table 2.3: Crystallite size for Pt and CeO2 species. Lattice parameters for CeO2. PtCrystal Size (nm) CeO2 Lattice parameter (Å) CeO2 Crystal Size (nm) Puralox - 5.41 3.9 CeAl HM - 5.41 4.7 Pt(2%)/Puralox (not observed) 5.41 4.7 Pt(4%)/Puralox 25.6 5.39 4.4 Pt(4%)/CeAl HM 13.6 5.40 5.5 Synthesis and characterization of Pt based catalysts: catalytic screening Chapter 2 68 Figure 2.5: TPR profiles: a) -Al2O3 support and platinum alumina based catalysts; b) CeO2-Al2O3 support and platinum ceria based catalysts 0100 200 300 400 500 600 700 800 900 C B Pt(4%)/Al2O3 330oC 201oC 178oC TCD signal normalized by gram (a. u.) Temperature (oC) 91oC Pt(2%)/Al2O3 339oC 196oC 156oC 96oC 0.5 a.u. A 100 200 300 400 500 600 700 800 900 BC 324oC 106oC 185oC 358oC 226oC 176oC Puralox Pt(2%)/Puralox TCD Signal normalized by gram (a.u.) Temperature (oC) Pt(4%)/Puralox 129oC 352oC 173oC 137oC 104oC CeAl HM Pt(4%)/CeAl HM A 2 a.u. 2.3. The Pt particle size for the Pt(4%)/CeAl HM sample is almost twice lower than that of the Pt(4%)/Puralox sample. The substantial platinum particle decrement could indicate stronger platinum support interaction able to stabilize low sized Pt particles. 2.3.3. TPR studies The reducibility of the samples was studied by H2-TPR experiments. H2 consumption profiles are presented in the Figure 2.5 in which a tentative peak deconvolution was made. It should be noted that this data processing does not represent an adequate peak deconvolution process. For that, the kinetics of each process must be taken into account. Therefore, it is just a visual aid to interpret the results. In general, the behavior of the platinum based catalysts during reductive (and/or oxidative) thermal treatment strongly depends on several parameter Synthesis and characterization of Pt based catalysts: catalytic screening Chapter 2 69 such as platinum loading, metal dispersion, temperature of the treatment, synthesis method and nature of support on which the platinum is deposited Considering that the Al2O3 is a non-reducible oxide, the H2 consumption observed for the platinum/alumina samples should be solely attributed to the platinum species reduction processes (Figure 2.5.a). For both metal loadings, the platinum reduction takes place between 50 and 500ºC. The reduction profiles exhibit three H2 consumption zones labeled A, B and C, in ascending order of temperature. In addition, the alumina based catalysts present two overlapped reduction processes in zone B. The discussion about the reducibility of the platinum species will be developed mainly considering platinum particle sizes and noble metal support interactions being both features closely related. At this point, it should be highlighted that no platinum oxidized species are detected on XRD data. Hence, when referring to larger platinum particle sizes implies particles lower than 4 nm. Mills et al. [28] suggested that highly dispersed Pt on Al2O3 rapidly oxidized to Pt4+ in air to form PtO2. So, the higher H2 uptake could reflect greater platinum oxidation degree after the calcination process where Pt2+ species are oxidized to form small PtO2 particles more easily reduced than isolated Pt2+ species [29]. Synthesis and characterization of Pt based catalysts: catalytic screening Chapter 2 70 Damyanova et al. [30] reported for Pt/alumina catalysts that PtOx particles were almost entirely reduced to Pt0 at 86 ºC, although, some minor fraction can be reduced to Pt2+ species. In our alumina based systems, the reduction centered at 91ºC could indicate the formation and reduction of PtO2 species. This process contribution seems to increase with Pt loading, being higher the relative H2 consumption for the Pt(4%)/Al2O3 sample indicating also higher PtOx population. In addition, Fukushima et al. [31] proposed that the degree of oxidation of Pt species increase with the diameter of Pt crystallites. In our case, it is not unexpected that the highest platinum content leads to bigger platinum particle size. Moreover, in this zone (zone A) H2 spillover phenomena should be also considered being commonly assumed that the first reduction step involves H2 dissociation. This step is facilitated when two hydrogen adsorption sites are close to each other, e.g. on higher size platinum particles. Merler et al. [32] observed two H2 consumptions centered at around 100 ºC and 300 ºC associated to PtOx of different particle sizes and noble metalsupport interaction: stronger the interaction, higher the temperature of reduction. This zone for Pt(2%)/Al2O3 is constituted by two different H2 consumptions centered at 156 and 196ºC. Both of them could be related to the reduction of different crystallite size PtOx species weakly interacting with the Al2O3 support to Pt0 [33–37]. Synthesis and characterization of Pt based catalysts: catalytic screening Chapter 2 71 Same processes are observed also for the Pt(4%)/Al2O3. So, increased Pt contents could lead to reduction processes shifted to higher temperatures possessing each one, different relative H2 consumptions. No matter the content, the intermediate reduction zone appears to be composed by two kind of platinum species being lower its reduction temperature when the interaction is weaker and/or when the size is bigger. The observed changes on the relative H2 consumptions on the intermediate reduction zone can be explained considering that higher platinum content may entail bigger platinum particles, including higher percentage of PtOx species reduced in that zone B. Highly dispersed PtOx species are more difficult to reduce due to the strongest interaction with the support. Tiernan [38] described a PtOx reduction step at ca. 420ºC due to the strong interaction between highly dispersed PtOx and Al2O3 (PtOx-Al2O3) and its high capacity to stabilize the oxidized platinum particles [39–47]. The H2-TPR profiles of the ceria based systems are presented in Figure 2.5.b. In general, cerium oxide presents two reduction processes, associated to the surface and bulk reduction process [48,49]. In our case, both ceria/alumina samples, the commercial and homemade, show one broad reduction step constituted of surface reduction process. The homemade oxide presents higher temperature of reduction. No bulk reduction process is observed for any support indicating either that the whole ceria is behaving as surface ceria or Synthesis and characterization of Pt based catalysts: catalytic screening Chapter 2 72 that the bulk reduction process overlaps with the surface one. However, the possibility that the bulk ceria is no reduced even at 900 ºC cannot be totally discarded. The platinum addition to those supports strongly enhances the reducibility, mainly by shifting the H2 take up towards lower temperatures. Platinum/Puralox based samples present two reduction zones. The reduction step situated at lower temperature (A) is composed by few contributions - reduction of PtOx species weakly interacting with the support, spillover effects and support reduction processes. The second reduction zone (B) is associated with PtOx-CeO2 interface reduction, lower the Pt particle size higher the noble metal-support interactions and higher the temperature of reduction. The main difference regarding the reducibility among the commercial and the homemade support remains on the third reduction step only observed for the Pt/homemade sample. While A and B zones should comprise similar reduction processes as described for Puralox based catalysts, the new C zone appear to be mainly due to pure support reduction process. It is also possible to calculate and compare the reducibility percentage (RP) within the samples series by using the Equation 2.1. 𝑅𝑃 =𝐸𝐻𝐶 𝑇𝐻𝐶 ×100 Equation 2.1 Synthesis and characterization of Pt based catalysts: catalytic screening Chapter 2 73 The PR relates the experimentally measured H2 consumption (EHC) to the total theoretical H2 consumption (THC). The THC will depend on the oxidation state considered for the reducible species. So, Ce4+ species were assumed as the initial oxidation state for ceria ones. As it will be shown on Chapter 6, Pt species in ceria presence were identified mainly as Pt2+. However, a possible oxidation reaction from Pt2+ to Pt4+ species during the calcination step cannot be definitely excluded in the case of the alumina supported catalysts. Hence, the reduction reactions considered for each oxide are: 2 𝐶𝑒𝑂2+ 𝐻2 → 𝐶𝑒2𝑂3+ 𝐻2𝑂 Equation 2. 2 𝑃𝑡𝑂 + 𝐻2 → 𝑃𝑡 + 𝐻2𝑂 Equation 2. 3 PtO2+ 2H2 → Pt + 2H2O Equation 2. 4 The calculated RPs of the samples are presented in Table 2.3 and the relative H2 consumption in the differently defined reduction zones in Figure 2.6. Except the alumina based catalysts when Pt4+ species are considered, all the samples present reducibility percentages close to 100%. Since metallic platinum species were detected for the calcined samples, those higher RP values mean H2 overconsumption which could be assigned to hydrogenation reactions of carbon deposits around Pt particles as well as organic rests from the synthesis method. Moreover, the detected H2 overconsumptions should be also related to spillover effects that happen on Pt particles and lead to a reversible dynamic surface equilibrium through which the H2 molecules are Synthesis and characterization of Pt based catalysts: catalytic screening Chapter 2 80 [15] M. Gonzalez Castaño, T.R. Reina, S. Ivanova, M. A. Centeno, J. A. Odriozola, J. Catal. 314 (2014) 1. [16] D. Çakír, O. Gülseren, J. Phys. Chem. C 116 (2012) 5735. [17] T.X.T. Sayle, S.C. Parker, C.R.A. Catlow, J. Phys. Chem. 98 (1994) 13625. [18] J.L. Ma, F. Ye, D.R. Ou, L.L. Li, T. Mori, J. Phys. Chem. C 116 (2012) 25777. [19] C. Zhang, A. Michaelides, D. A. King, S.J. Jenkins, J. Phys. Chem. C 113 (2009) 6411. [20] A. Sepúlveda-Escribano, F. Coloma, F. Rodrıguez-Reinoso, J. Catal. 178 (1998) 649. [21] C. Hardacre, R.M. Ormerod, R.M. Lambert, J. Phys. Chem. 98 (1994) 10901. [22] D. Andreeva, I. Ivanov, L. Ilieva, M. V. Abrashev, Appl. Catal. A Gen. 302 (2006) 127. [23] G. Germani, P. Alphonse, M. Courty, Y. Schuurman, C. Mirodatos, Catal. Today 110 (2005) 114. [24] T.R. Reina, W. Xu, S. Ivanova, M.A. Centeno, J. Hanson, J.A. Rodriguez, J.A. Odriozola, Catal. Today 205 (2013) 41. [25] M.A. Centeno, C. Portales, I. Carrizosa, J.A. Odriozola, Catal. Letters 102 (2005) 289. [26] G. Leofanti, M. Padovan, G. Tozzola, B. Venturelli, Catal. Today 41 (1998) 207. [27] O.H. Laguna, M. A. Centeno, G. Arzamendi, L.M. Gandía, F. Romero-Sarria, J. A. Odriozola, Catal. Today 157 (2010) 155. Synthesis and characterization of Pt based catalysts: catalytic screening Chapter 2 81 [28] G.A. Mill, S. Weller, E.B. Cornelius, Actes Du Zieme Congr. Intern. Catal. 2 (1962). [29] C.P. Hwang, C.T. Yeh, J. Mol. Catal. A Chem. 112 (1996) 295. [30] S. Damyanova, J.M.C. Bueno, Appl. Catal. A Gen. 253 (2003) 135. [31] T. Fukushima, J.R. Katzer, D.E. Sayer, J.Cook, 7th Intern. Congr. On Catal. Tokyo (1980). [32] E. Merlen, P. Beccat, J.C. Bertolini, P. Delich, 188 (1996) 178. [33] M. El Doukkali, A. Iriondo, P.L. Arias, J. Requies, I. Gandarías, L. Jalowiecki-Duhamel, F. Dumeignil, Appl. Catal. B Environ. 125 (2012) 516. [34] A. Iriondo, J.F. Cambra, V.L. Barrio, M.B. Guemez, P.L. Arias, M.C. Sanchez-Sanchez, R.M. Navarro, J.L.G. Fierro, Appl. Catal. B Environ. 106 (2011) 83. [35] V. A. Mazzieri, J.M. Grau, J.C. Yori, C.R. Vera, C.L. Pieck, Appl. Catal. A Gen. 354 (2009) 161. [36] P. Ciambelli, V. Palma, A. Ruggiero, Applied Catal. B Environ. 96 (2010) 18. [37] M. El Doukkali, A. Iriondo, J.F. Cambra, L. Jalowiecki-duhamel, A.S. Mamede, F. Dumeignil, P.L. Arias, Journal Mol. Catal. A, Chem. 368- 369 (2013) 125. [38] M.J. Tiernan, O.E. Finlayson, Appl. Catal. B Environ. 19 (1998) 23. [39] H.C. Yao, M. Sieg, J. Plummer H.K., J. Catal. 59 (1979) 365. [40] G. Lietz, H. Lieske, H. Spindler, W. Hanke, J. Völter J.Catal. 81 (1983) 17. Synthesis and characterization of Pt based catalysts: catalytic screening Chapter 2 82 [41] S.H. Park, M.S. Tzou, W.M.H. Sachtler App. Catalysis 24 (1986) 85. [42] J.Z. Shyu, K. Otto, J. Catal. 115 (1989) 16. [43] K. Shyu, J Z, Otto, Appl. Surf. Sci. 32 (1988) 246. [44] N.W. Hurst, S.J. Gentry, B.D. Mcnicol, Catal. Letters 24 (1982) 233. [45] C. H.C. Yao, H.S. Gandhi, M. Shelef, B. Imelik, P. Naccache, G. Coudurier, H. Praliaud, P. Meriaudeau, P. Gallezot, G.A. Martin, J.C. Vedrine, Stud. Surf. Sci. Catal. 11 (1982) 159. [46] H.C. Yao, M. Sieg, H.K. Plummer Jr. J. Catal. 59 (1979) 365. [47] A. P. Ferreira, J.C.S. Araújo, J.W.C. Liberatoria, S. Damyanova, D. Zanchetc, F.B. Noronha, J.M.C. Bueno, Stud. Surf. Sci. Catal. 167 (2007) 433. [48] W.Y. Hernández, O.H. Laguna, M. A. Centeno, J. A. Odriozola, J. Solid State Chem. 184 (2011) 3014. [49] T.R. Reina, E. Papadopoulou, S. Palma, S. Ivanova, M. A. Centeno, T. Ioannides, J. A. Odriozola, Appl. Catal. B Environ. 150-151 (2014) 554. [50] W.C. Conner, J.L. Falconer, Chem. Rev. 95 (1995) 759. [51] G. Jacobs, U.M. Graham, E. Chenu, P.M. Patterson, A. Dozier, B.H. Davis, J. Catal. 229 (2005) 499. [52] A. Badri, C. Binet, J. Lavalley, 92 (1996) 4669. Chapter 3 Electronic and structural promoters of Pt(4%)/CeO2/Al2O3 catalyst Summary An electronic and structural promotion of the Pt/CeO2/Al2O3 catalytic system was pretended and achieved by the addition of two different ceria promoters, Zr and Fe. Zr appears to be very efficient structural promoter resulting in lower platinum particle size and stronger Pt-CO interactions meanwhile Fe improves the electronic properties of the system reflecting oxygen vacancies concentration and the band gap energy alterations. The combination of both doping agents maintains the beneficial effects provided by both separately but also manifest a synergistic effect reflected on significant changes on the band gap energies, chemisorptive properties as well as catalyst’s behavior regarding deactivation phenomena. Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 84 3.1. Introduction Along the previous Chapter, the beneficial effect provided by the ceria addition to the platinum catalysts was clearly established. The enhanced activity of the ceria supported NM catalysts is frequently attributed to the ceria ability to undergo fast storage/release oxygen cycles via reversible Ce4+ to Ce3+ species conversions [1,2]. These redox cycles imply oxygen vacancies (Ov) formation mechanism which normally correlates to an enhanced catalytic activity [3,4]. As shown in the Figure 3.1, no matter the mechanism through which the reaction happens the role of the oxygen vacancies is markedly important. Simplified, the redox mechanism consists of two consecutive steps where i) the CO is oxidized to CO2 resulting on Ce3+- Ov couples on the support and ii) the water molecules are dissociated on the Ov entailing the reoxidation of Ce3+ to Ce+4 species. On the other hand, the associative mechanism proposes the reaction of activated CO species with the OH groups already created on the support, through water dissociation process, leading to intermediates carbonaceous species which subsequently decomposes to CO2 and H2. The created oxygen vacancies are then replenished by the water and the cycle repeats. Therefore, an oxygen vacancy increment on the surface should lead to enhanced catalytic performances. Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 85 In the case of CeO2, it has been widely demonstrated that the concentration of oxygen vacancies can be enhanced by the modification of its structure through doping with different cations such as Zr [5–9], Fe [10–13], Eu [12] and Zn [14,15] among others. Several studies published the use of Zr as doping agent on the basis of its good thermal stability, ability to decrease the ceria sintering [16,17] or interesting acid-base and redox properties [18]. A significant improvement of the catalytic behavior in the presence of Zr, in comparison with Pt/ CeO2 samples, was reported [19,20]. On the contrary when Fe is used as dopant, either decrease [21], increase [22] or even no effect [23] were reported for the WGS reaction. Although the both dopants could be successfully used for promoting the CeO2 properties, its properties largely depend on the preparation method, the nature of the present phases and their relative content [6,24,25]. Actually, the Figure 3.1. Possible role of ceria vacancies on WGS reaction depending on the reaction mechanism reaction Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 86 influence of these features may explain the diversity of conclusions published in several works regarding the Ce doped systems. In this context, a careful design of Pt-mixed oxides catalyst is needed for apprehension the role of the dopants in the WGS reaction. For this reason, in the present chapter, a platinum based catalyst series was prepared employing as support ceria zirconia mixed oxide with increased Zr amounts. These catalysts were characterized and tested in WGS conditions. With the optimal Ce/Zr relation established, the sample was employed as a starting point for the second part of the chapter. New Zr and Fe doped system were prepared in order to promote the Pt/CeO2 activity. Both doping agent are used separately or in a mixed manner in order to study their synergism in ceria promotion but always maintaining the Ce/doping metal proportion. In addition, for all the samples prepared in this thesis, the ceria quantity was optimized by dispersing it on Al2O3 matrix with the aim to diminish the ceria waste and to increase its surface to volume ratio. 3.2. Experimental - Catalyst synthesis Six different supports were synthesized, CeO2/Al2O3, Ce0.9Zr0.1O2/Al2O3, Ce0.75Zr0.25O2/Al2O3 Ce0.5Zr0.5O2/Al2O3, Ce0.9Fe0.1O2/Al2O3 and Ce0.9Zr0.05Fe0.05O2/Al2O3. The precursor quantities were calculated to obtain cerium oxide or cerium/promoter mixed oxide loading of 20 wt.% over Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 87 alumina in the final material. Nominally, the desired cerium - doping agent proportions were calculated in molar ratio. The supports preparation was carried out by wetness impregnation on a high surface ɣ-Al2O3 powder (Sasol). The metal oxide precursors were Ce(NO3)3·6H2O, Fe(NO3)3·9H2O and ZrO(NO3)2·xH2O (Aldrich). After drying, the solids were calcined at 450ºC for 4h with a 10ºC/min heating rate. In the adopted nomenclature, the oxygen and the subscripts were omitted. The supports CeO2/Al2O3, Ce0.9Zr0.1O2/Al2O3, Ce0.75Zr0.25O2/Al2O3, Ce0.5Zr0.5O2/Al2O3, Ce0.9Fe0.1O2/Al2O3 and Ce0.9Zr0.05Fe0.05O2/Al2O3 are then labeled CeAl, Ce0.9ZrAl, Ce0.75ZrAl, Ce0.5ZrAl, Ce0.9FeAl and Ce0.9ZrFeAl, respectively. The nominal platinum content was fixed to 4 wt. %. The platinum deposition was carried out by wetness impregnation using as a precursor tetrammonium nitrate platinate (Johnson-Matthey) slightly modified by 1M acetic acid in a 1:1.1 ratio. The solids were dried and, then, calcined at 350 ºC for 8h at 5 ºC/min. The resulting samples were named Pt/CeAl, Pt/Ce0.9ZrAl, Pt/Ce0.75ZrAl, Pt/Ce0.5ZrAl, Pt/Ce0.9FeAl, and Pt/ Ce0.9ZrAl. 3.3. Support optimization for Pt based CexZr1-xO2/Al2O3 3.3.1. Physicochemical characterization results 3.3.1.1. Textural properties and chemical composition of the prepared solids Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 88 The textural properties of the prepared solids are presented on the Table 3.1. It can be observed that the ceria incorporation to alumina leads to lower specific surface areas, average pore volume and diameter. Regarding the zirconia doped samples, higher zirconia contents result in increase of the specific surface areas and pore volumes values. On the other hand, as a general trend, the platinum incorporation to the supports leads to surface area diminutions and increase of the pore volume except for the Pt/CeAl catalyst which exhibits an important decrement of its pore volume in comparison to its parent support. Table 3.1. Textural properties and chemical composition of the prepared solids SBET (m2/g) Vpore (cm3/g) Al2O3 (wt. %) CeO2 (wt. %) ZrO2 (wt. %) Pt (wt. %) Al2O3 202 0.51 100 --- --- --- CeAl 186 0.41 74.3 24.7 --- --- Ce0.9ZrAl 160 0.36 80.1 17.7 2.2 --- Ce0.75ZrAl 167 0.38 80.0 18.0 2.0 --- Ce0.5ZrAl 173 0.39 80.8 16.7 2.5 --- Pt/CeAl 147 0.31 72.0 24.2 --- 3.8 Pt/Ce0.9ZrAl 150 0.36 74.7 18.8 1.6 4.9 Pt/Ce0.75ZrAl 156 0.38 75.9 17.4 2.0 4.7 Pt/Ce0.5ZrAl 166 0.38 76.7 16.2 2.4 4.7 Regarding the chemical compositions, the supports and their corresponding catalysts present chemical compositions closer to the desired ones. Nevertheless, all the catalysts showed slightly higher platinum contents in Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 89 24 26 28 30 32 34 36 CeAl Ce0.9ZrAl Ce0.75ZrAl Ce0.5ZrAl R. I. (a. u.) 2 20 30 40 50 60 70 80 90 Al2O3 CeO2 ZrO2 CeAl Ce0.9ZrAl Ce0.75ZrAl Ce0.5ZrAl R. I. (a. u.) 2 respect to the nominal loadings value (4 wt.%) suggesting some support dissolution during the preparation process. 3.3.1.2. XRD analysis of the prepared solids The XRD patterns of the prepared solids are presented in the Figure 3.2.a. The characteristic γ-Al2O3 (JCPDS# 00-050-0741) and CeO2 (JCPDS# 00-004- 0593) diffractions are present in all prepared solids. Moreover, zirconia typical diffractions were not observed. In Figure 3.2.b, the main ceria diffraction (111) is amplified. As it can be observed, the increased zirconia loadings lead to pronounced shift of this diffraction towards higher 2 in comparison to the bare CeAl support. The observed shifts can be explained with an substitutional solid solution formation. Indeed, ceria lattice contraction should occur when a replacement of ceria cations by lower ionic radii zirconia cations takes place (0.97 Å and Figure 3.2. XRD of the synthesized supports a)  10-95 2range; b) Main diffraction of ceria lattice (111) Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 96 Although, the changes do not seem very significant for VPore and DPore, opposite tendencies are observed. While the zirconia addition increases the VPore and DPore, the Fe decreases them. When both doping agents are present, the surface area of the sample results in an intermediate value between the monometallic substitutions. The VPore and DPore show the same values as for the Fe-doped sample. The Zr-doped sample presents very similar textural properties to the parent CeO2/Al2O3 support, which could indicate higher similarities of the samples on molecular level, i.e. solid solution formation. The decrease of the SBET and DPore for the Fe-doped sample, however, suggests the deposition of a second phase but do not exclude the solid solution formation. The chemical composition of the studied systems is presented in Table 3.3. Table 3.3. Chemical composition of the synthetized solids Al2O3 (wt.%) CeO2 (wt.%) ZrO2 (wt.%) Fe2O3 (wt.%) Pt (wt.%) CeAl 75.3 24.7 --- --- --- Ce0.9ZrAl 80.1 17.7 2.2 --- --- Ce0.9FeAl 81.4 17.4 --- 1.2 --- Ce0.9ZrFeAl 81.0 18.0 0.6 0.4 --- Pt/CeAl 72.0 24.2 --- --- 3.8 Pt/ Ce0.9ZrAl 74.7 18.8 1.6 --- 4.9 Pt/ Ce0.9FeAl 78.3 15.9 --- 1.1 4.7 Pt/ Ce0.9ZrFeAl 78.8 16.5 0.5 0.4 3.8 Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 97 10 20 30 40 50 60 70 80 90 CeO2 Ce0.9FeZrAl Ce0.9FeAl R. I. (a. u.) 2 ZrO2 Fe2O3 Al2O3 CeAl Ce0.9ZrAl 26 28 30 32 R.I. (a. u.) 2 CeAl Ce0.9ZrAl Ce0.9FeAl Ce0.9ZrFeAl Although a slight enrichment in Zr observed for the CeZr sample, probably due to the cerium content loss during the preparation; the obtained compositions are very close to the nominal values. 3.4.1.2. XRD results The XRD patterns of the employed supports are presented in Figure 3.7.a where Al2O3 is also included for clarity. All the supports show the characteristic diffractions of the cubic CeO2 fluorite structure (JCPDS# 00- 004-0593) and γ- Al2O3 phase (JCPDS# 00-050-0741). Neither Fe2O3 nor ZrO2 XRD signals are detected probably due to the small amount of doping agents or to the inclusion of these cations in the ceria lattice forming solid solution. Iron or zirconium doped cerium solid solutions systems are widely described, especially in this range of concentration [11,13,34]. In fact, the diffractions of the doped-ceria supports shift towards higher angles, as illustrated in Figure 3.7.b. The calculated parameters are Figure 3.7. a) XRD of the prepared supports; b) (111) ceria diffraction line Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 98 presented in the Table 3.4, being the lattice parameter similar for all doped systems and inferior to that of the parent CeAl. This behavior is caused by the substitution of Ce4+ with Fe3+ or Zr4+ ions. Taking into account, the ionic radii (0.97Å, 0.64Å and 0.84Å for Ce4+, Fe3+ and Zr4+ respectively) a higher contraction of the ceria lattice is expected when iron is used as doping agent. However, the same lattice contraction is observed for all doped samples no matter the doping metal. This fact could indicate a partial segregation of the iron oxide species forming interstitial solid solution, as reported by Reina et al. [14,24]. Although the segregated Fe oxide is not detected by XRD because of its low concentration, its presence cannot be discarded. Table 3.4. For supports, ceria lattice parameters and crystallite size. CeO2 Crystal Domain (Å) CeO2 Lat. parameter (Å) CeAl 36 5.41 Ce0.9ZrAl 40 5.39 Ce0.9FeAl 35 5.39 Ce0.9ZrFeAl 41 5.39 The crystallite size was also calculated by using Scherrer equation (Table 3.4). A slight increase of the ceria particle size is observed when Zr is used as dopant and, on the contrary, the Fe addition does not provoke variations on the particle size. When both doping agents are present, the observed particle size is similar to that of the Zr-doped system. Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 99 010 20 30 40 50 60 70 80 -Al2O3 CeO2 Pt/Ce0.9FeZrAl Pt/Ce09FeAl R. I. (a. u.) 2 Pt/CeAl Pt/Ce0.9ZrAl 010 20 30 40 50 60 70 80 90 Pt/CeAl Pt/Ce0.9ZrAl Pt/Ce0.9FeAl Pt/Ce0.9ZrFeAl R.I. (a. u.) 2 Al2O3 CeO2 The XRD patterns for the calcined platinum containing samples are shown in the Figure 3.8.a. For all Pt doped samples, narrowing in the main cerium oxide diffraction (111) was detected being associated to ceria particle size increment during the platinum addition. This increase might be the consequence of a partial re-dissolution of the support during the platinum deposition, resulting in reorganization of the latter with higher particle size or to the sintering during the second calcination process after Pt deposition. In the case of Pt/Ce0.9FeAl, an additional ceria lattice contraction from 5.39 to 5.38 Å was detected during the second calcination, which could indicate that some of the interstitial Ce-Fe solid solution converts into substitutional one. The XRD diffractions of the reduced catalysts, in the conditions of pretreatment prior the catalytic test, are presented in Figure 3.8.b. The main Pt0 diffraction (111) is detected at 2=39.7º for all the catalysts. Some important differences can be observed depending on the support. Minor Pt Figure 3.8. XRD of the prepared catalysts a) calcined; b) reduced at 350ºC during 2h Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 100 particles size is observed for the Pt/Ce0.9ZrAl sample suggesting a strong metal support interaction able to stabilize the Pt particles upon reduction. The platinum particle size calculated from XRD for all the samples is shown in Table 3.5. It should be mentioned however that at 39.7º some alumina diffractions are also present, which implies that the Pt particle size estimations is only approximated. Table 3.5: Ceria crystal domain and lattice parameters for the fresh catalysts and platinum crystal domain estimated for the reduced catalysts CeO2 Crystal Domain (Å) CeO2 Lat. parameter (Å) Pt Crystal Domain (nm) Pt/CeAl 36 5.395 21.1 Pt/Ce0.9ZrAl 43 5.385 7.7 Pt/Ce0.9FeAl 38 5.379 11.7 Pt/ Ce0.9ZrFeAl 45 5.381 10.8 All the doped systems present lower platinum sizes, in comparison with the unpromoted sample. Both dopants provoked significant decrement on the resulting Pt size, being Pt/Ce0.9ZrAl the sample with the lowest crystalline domain. 3.4.1.3. CO chemisorption and HR-TEM microscopy. The metal dispersion was determined by CO chemisorption at liquid N2 temperature in order to avoid the ceria contribution through CO oxidation. The metal dispersion as the number of surface exposed Pt atoms was Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 101 calculated from the total amount of adsorbed CO. More precisely, it was assumed that each surface platinum atom only chemisorbs one CO molecule and that the particles are cubic with one non-exposed face (Shape Factor = 5). From the metal dispersion, the average metallic sizes were estimated by Spinadel and Boudart method being the details of the used model reported elsewhere [28]. Both, dispersion results and estimated platinum particle size are shown in Table 3.6. The calculated particle size values are significantly lower than those obtained from XRD. Table 3.6. Estimated Pt dispersion and particle size from CO chemisorption. Dispersion (%) Pt calculated size (nm) Pt/CeAl 59 2.7 Pt/Ce0.9ZrAl 56 2.7 Pt/ Ce0.9FeAl 58 2.7 Pt/ Ce0.9ZrFeAl 53 2.7 Considering that great difference, the HR-TEM was employed in order to determine the real platinum particle size. In Fig. 3.9 are shown the HADAF images obtained for the reduced Pt/CeAl and Pt/Ce0.9ZrFeAl catalysts. A very homogeneous Pt particle size distribution was observed with an average size estimated at 1 ± 0.2 nm. The Pt size distribution appears to be even lower than the size calculated from the CO chemisorption experiments and much inferior to the crystallite size Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 102 Pt t Pt a b estimated by the XRD’s Scherrer equation. Although not observed in HRTEM, the existence of some agglomerates should not be excluded, as the Pt (111) diffraction was detected or higher Pt sizes were calculated by the chemisorption study. Therefore, all the catalyst should present two different platinum species: a minor fraction constituted by platinum agglomerates and a major one composed by highly dispersed platinum species. Within the series the doping agent introduction leads to lower fraction of agglomerated Pt, as confirmed by XRD. On the other hand no matter the agglomerates formation the results obtained from HR-TEM and CO chemisorption indicates that the obtained Pt size is similar for all the samples. From this moment one the Pt dispersion on these samples will be considered always the same. Figure 3.9. HR-TEM microphotographs of reduced samples: a) Pt/CeAl; b) Pt/CeZrFe Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 103 1 2 3 4 5 Kubelka Munk Function (a. u.) e. v. CeAl Ce0,9ZrAl Ce0,9FeAl Ce0,9ZrFeAl 3.4.1.4. Raman and UV-Vis characterization An approach to the electronic properties of the studied materials is obtained by DR UV-Vis spectroscopy. The DR UV-Vis spectra of the synthesized supports are shown in the Fig. 3.10.a. All the supports present a broad band, centered at ca. 280 nm, associated to O2p  Ce4f charge transfer transition characteristic for CeO2 [35]. The iron addition leads to the widening of this band, which is commonly attributed to the contribution of the iron oxide absorption features at around 530 nm [3]. In the case of Ce0.9ZrFeAl system, a substantial widening in the absorption band is observed. The indirect band gap energies of prepared supports estimated by extrapolating to zero the energy dependence of the [F(R)hν]1/2 are presented in Table 3.7. The corresponding band gap energies are not calculated for the catalysts because of the addition of the platinum plasmon band [36]. The latter Figure 3.10. a) UV-Vis spectra of the prepared supports; b) represented in Kubelka-Munk mode 200 300 400 500 600 700 800 900 Absorbance (a. u.) Wavelength (nm) CeAl Ce0.9FeAl Ce0.9ZrAl Ce0.9ZrFeAl Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 104 appears at around 500-600 nm, as a consequence of a collective oscillation of the conduction electrons stimulated by the incident light and does not permit the accurate calculation of the band gaps [37]. The CeAl sample presents indirect band gap energy of 2.36 eV, in good agreement with the literature [34]. The measured band gap for our ceria sample is smaller than that found for ceria monocrystal (6 eV.) and could be associated to higher Ce3+ to Ce4+ species ratio indicating partially reduced ceria on the surface [34,38]. The addition of doping metal lowers the CeAl band gap energies. The zirconium promoted sample presents a minor band gap energy decrease, in comparison to the iron promoted one. In the case of Ce0.9ZrFeAl, a significant decrease in the indirect band gap energies values is observed (1.65 eV). Several factors, as particle sizes and/or introductions of electronic levels between the valence and the conduction bands, could involve changes in the electronic properties and, hence, in the band gap energies [39,40]. Raman Table 3.7. Band gap energies calculated for the synthetized supports BG (eV) CeAl 2.36 Ce0.9ZrAl 2.33 Ce0.9FeAl 2.01 Ce0.9ZrFeAl 1.65 Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 105 550 600 650 R I. (a. u.) Wavenumber (cm-1) Ov spectroscopy was also used for electronic properties investigation. The Raman spectra of the studied supports are shown in the Figure 3.11. As it can be observed, CeAl system shows a wide band at 458 cm-1 which is assigned to F2g Raman-active mode of the CeO2 fluorite structure due to the symmetric oxygen breathing vibration around Ce4+. It should be mentioned that no bands attributed to segregated iron species were discerned indicating that the segregation although not excluded, should not be so important. Additionally, in the Figure 3.11.b, a small band at around 600 cm-1 ascribed to the presence of oxygen vacancies (Ov) is perceived [12,13,41]. It has been established in the literature that the Ov/F2g bands area ratio is a correct way to evaluate the surface oxygen vacancies concentration [13]. 400 450 500 550 Normalized intensity (a. u.) Raman shift (cm-1) CeAl Ce0.9ZrAl Ce0.9FeAl Ce0.9ZrFeAl F2g Figure 3.11. Raman spectra obtained for the synthesized supports: a) F2g b) Ov Raman band Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 112 100 200 300 400 500 600 700 800 Pt/Ce0.9FeAl Pt/Ce0.9ZrFeAl 5250C 5080C 4420C 4460C 1090C 1420C 1270C Pt/Al Pt/Ce0.9ZrAl TCD Signal Temperature (oC) Pt/CeAl 990C more significant. These shifts reveal an improvement on the reducibility, probably due to synergic effect obtained through the solid solution formation, in particular, when iron is present. The H2-TPR of the platinum based catalysts is shown in the Fig. 3.13. All platinum based catalysts exhibit three clearly differentiated reduction zones. Various factors have to be taken into consideration when comparing the relative contribution of the reduction processes; the particle size of all reducible species and their interaction within the sample. The first H2 consumption zone takes place up to 250oC, the second in the 250- 425oC range and the third above 425ºC. The first zone is attributed to the Figure 3.13. H2-TPR profiles of the supports (dot lines) compared with their respective catalysts (solid lines) Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 113 reduction of weakly interacting PtOx species. It could be explained considering that the first reduction step is H2 dissociation. The H2 dissociation is easier when two H adsorption sites are adjacent; a case more probable on higher Pt particle sizes [30,46,47]. So, the higher the particle size, the lower the temperature of reduction. Nevertheless, the absence of XRD diffraction for PtOx indicates that the highest particle size should be typically lower than 5 nm, which is in concordance with HR-TEM results. The second reduction zone could be associated to highly dispersed PtOx species strongly interacting with the support and to partial support reduction. In general, the temperature of reduction depends on particle size and the NM- support interaction [31]. Strong NM-support interaction increases the reduction temperatures, because of the NM-support synergism. The third zone could be only associated to the support reduction, which occurs at temperatures superiors to the final WGS temperature (350 ºC), indicating that the support will not be completely reduced during the WGS reaction [48,49]. The reducibility percentages (RP) calculated considering Pt2+, Fe3+ and Ce4+ species are presented on the Table 3.10. None of the support, achieves RP of 100% probably related with the bulk ceria fraction not reduced during the TPR experiments. On the other hand, the platinum incorporation leads to superior RP values for all the samples. Actually, the platinum samples exhibit RPs higher than 100% Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 114 being the H2 overconsumption especially noticeable for Pt/Ce0.9ZrFeAl sample. Indeed, the platinum metallic diffraction clearly detected for Pt/CeAl and Pt/Ce0.9FeAl and then, RPs above 100% also will indicate H2 overconsumptions. These overconsumptions could be explained by hydrogen species incorporation into the ceria matrix forming HyCeO2-x bronze like species [27,50]. Fiero et al. [51] published similar results and concluded that an important amount of H2 was incorporated in bronze like species into the ceria lattice by heating under H2 at 300 ºC. The relative H2 consumption for all the three zones is estimated and presented on the Figure 3.14. The first, second and third zone corresponds to the temperature ranges between 25-250 ºC, 250-425 ºC and 425-900 ºC, respectively. Comparing the three relative H2 consumption processes for each system, an improvement of the redox behavior is detected for the Pt/Ce0.9ZrFeAl system. For this system, the area of the first H2 consumption step is qualitatively higher than the second one, contrary to all the other Pt based catalysts. This improvement appears to be the consequence of several Table 3.10. Reducibility percentages (RP) obtained for supports and catalysts RP (%) RP (%) CeAl 76 Pt/CeAl 91 Ce0.9ZrAl 86 Pt/ Ce0.9ZrAl 105 Ce0.9FeAl 71 Pt/ Ce0.9FeAl 100 Ce0.9ZrFeAl 79 Pt/ Ce0.9ZrFeAl 118 Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 115 factors induced from both promoters. The zirconia lowers the temperature of reduction and the iron enhances the reducibility, by making easier H2 dissociations and diffusions. As a result the relative reduction degree in the lower temperature zone is enhanced and an excellent redox behavior for this sample is achieved. Considering that all the catalysts are activated in H2 at 350 ºC during 2h prior the catalytic tests, the reduction processes observed in ZI and ZII should be performed. From the TPR data, the ceria reduction degree presented for the catalysts after the activation step can be also estimated. For that different assumptions are made i) all the platinum species are reduced along the ZI and ZII processes; ii) ceria reduction happens along the three reduction zones iii) the third and highest temperature reduction zone is uniquely due to ceria support. The reduction degree (Ce4+/Ce3+) estimations obtained are presented in Table 3.11. 0 20 40 60 80 ZI ZII ZIII Pt/CeAl Pt/Ce0.9ZrAl Pt/Ce0.9FeAl Pt/Ce0.9ZrFeAl Figure 3.14. Relatives H2 consumptions as a function of the temperature zone calculated form H2-TPR Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 116 Table 3.11. % Reducibility degree of supports exhibited for the catalysts after the H2 pretreatment performed on the catalytic test Pt/CeAl Pt/ Ce0.9ZrAl Pt/ Ce0.9FeAl Pt/ Ce0.9ZrFeAl 48.6 56.1 76.0 73.6 As it can be observed, the ceria lattice modification, no matter the doping metal, leads to significant differences in the redox behavior of the samples. Meanwhile the zirconia addition leads to slightly increased Ce3+/Ce4+ ratio in comparison to the unpromoted sample, the Fe containing samples exhibits much higher reduction capacities. The iron presence induces similar reduction degree no matter the presence or not of a second metal. At this point, the modifications performed by each doping agents can be appreciated. The redox behavior is prominently enhanced through the iron incorporation to the ceria lattice while, the most remarkable effect of zirconia is attributed to the stronger noble metal support interaction achieved and confirmed by the Pt size reduction. 3.4.1.7. Temperature Programmed Desorption (TPD) experiments TPD characterization techniques can be used as a simple tool to obtain more detailed information about the gas-solid interactions during the WGS reaction. Indeed, in heterogeneous catalysts the reactive species should arrive to the catalytic surfaces, should adsorb, reacts and leave the surface liberating the active sites. Thus, knowledge about interactions of H2, CO and CO2 species Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 117 with the surface should provide useful understanding and correlation to the catalyst’ reactivity. The TPD experiments were carefully designed in order to obtain clean and similar catalytic surfaces before any gas adsorption or reaction. On first place the catalysts are preheated in inert gas (Ar) then reduced under H2 and finally, the surface species generated by the reduction were removed by secondary heating in inert atmosphere. The adsorption desorption experiment includes the absorption of the reactive species till surface saturation, removal of the weakly physisorbed species under inert flow and then desorption as a function of the temperature. The H2-TPD profiles, presented on the Figure 3.15, were obtained after H2 adsorption at room temperature. All the catalysts exhibited two similar and well discerned desorption zones with the increasing of the temperature. The low temperature H2 desorption for all the samples suggests physically adsorbed H2 on the catalysts surface meanwhile the majority of H2 desorbed at higher temperature proposes strong chemisorption [52]. Moreover, the higher temperature desorption zone appears to be composed by multiple desorption processes accounting for different states of the adsorbed species on the surface. Although, the evolved desorption processes cannot be properly separated, several desorption processes could be appreciated especially for the Pt/CeAl and Pt/Ce0.9FeAl samples. The observed wide Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 118 desorption process also indicates a strong dependence of the chemisorption properties on the surface coverage degree. Figure 3.15. H2-TPD profiles of the catalysts after H2 adsorption at 25 ºC Normally, H2 species desorbed at temperatures around 115 ºC, are ascribed to chemisorbed hydrogen on dispersed platinum particles [53]. The H2 desorption initiates at lower temperatures for Zr or Fe doped samples compared to the unpromoted sample, indicating that the presence of any of both metals can modify the platinum adsorption properties, either by modifying Pt particle size or by affecting its electronic properties. In fact, the Pt/Ce0.9ZrFeAl sample exhibits the above mentioned desorption immediately after evacuating the physisorbed H2 being the later process even overlapped. That easiness of H species desorption, probably via reverse spillover process, suggests a substantial increment of the H species transfer ability, in good agreement with the high H2 consumption observed at lower temperatures during the TPR test for this sample. 100 200 300 400 500 600 Pt/Ce0.9ZrFeAl Pt/Ce0.9FeAl Pt/Ce0.9ZrAl 432oC 365oC 66oC292oC 298oC 43oC 414oC 48oC 244oC331oC 327oC ppm H2/g Temperature (oC) 500ppm 66oC Pt/CeAl Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 119 On the other hand, the introduction of both metals into the ceria lattice may increase the electron density on the surface. Considering the electron donor character of the H2 species, the Pt-H2 bond strength should decrease and thus, the chemisorbed H2 species should be more easily removed from the surface. The evolved H2 species at higher temperatures, around 285 ºC, were related in the literature to hydrogen species adsorbed at the perimeter sites located at the metal support interface [53]. Any temperature shift of these desorption accounts for stronger electronic interactions between those sites and H2. The higher temperature desorption should be also related to reverse spillover of the H2 species from the support to the platinum [52]. Multiple processes can be observed in the 250-350 ºC temperature range, usually assigned either to desorption of strongly chemisorbed spillover hydrogen or to that of adsorbed hydrogen in the subsurface layers. The H2-TPD profiles when H2 is absorbed at 300 ºC are presented on the Figure 3.16. It should be underlined the difference in the desorbed H2 amounts which is a function of the adsorption conditions. The hydrogen amount desorbed after H2 adsorption at 300 ºC, is noticeably higher in respect to that observed at 25 ºC. This is in good agreement with Menon and Froment [54] work where the effect of the H2 treatment temperature on the platinum samples adsorption was studied. They reported that higher the treatment temperature higher the spillover capacity of the samples, suggesting different interactions at every H2 chemisorption temperatures. Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 120 Figure 3.16. H2-TPD profiles of the catalysts after H2 absorption at 300ºC Regarding the shape of the H2-TPD profiles, the main difference relays on the H2 desorbed at around 100 ºC. This desorption process is present when the H2 chemisorption is carried out at 300 ºC, but absent at room temperature adsorption. This absence at low temperature could be attributed to the dissociative H2 chemisorption through which the hydrogen species will be consumed and spillover to the support. Meanwhile, when H2 is adsorbed at high temperature, the spillover processes should be already performed and chemisorbed hydrogen species readily desorbed. Therefore, the higher H2 amounts at 300ºC could be attributed to reverse spillover effects during the whole desorption process. The observed H2-TPD results could be also correlated to the bronze like species formation. The presence of stabilized hydrogen species, as bronze like entities, must affect the adsorption and therefore the catalytic properties of the 100 200 300 400 500 600 Pt/CeAl Pt/Ce0.9ZrAl Pt/Ce0.9FeAl ppm H2/gcat Temperature (oC) 1000ppm Pt/Ce0.9ZrFeAl Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 121 samples [51]. The significantly higher H2 amounts desorbed after adsorption at 300 ºC suggests that the incorporation of hydrogen species into the support is favored by the temperature and that those hydrogen species are later desorbed via reverse spillover through the platinum metal as H2 molecules. The CO/catalyst interactions are also analyzed by CO-TPD experiments. In this case during the desorption process the CO2, H2 and CO profiles are evaluated and presented in the Figure 3.17. The m/z=28 signal, is attributed only to the CO species and the relative contribution arising from the CO2 species is properly subtracted. For all the catalysts, two CO desorption zones (signal m/z=28) can be separated as a function of temperature: one below 250 ºC and another above that temperature. On first place a difference in the relative desorbed CO amounts is detected, being the Pt/Ce0.9ZrAl and Pt/Ce0.9FeAl the samples with the highest desorbed quantities. For similar systems, comparable profiles has been described in the literature [20,55,56]. For example, Du et al. [20] report weakly interacting CO species desorbing at temperatures below 100 ºC and hardly interacting ones evolving at around 320 ºC. The CO species desorbed between 170-260 ºC are generally ascribed to moderately adsorbed CO molecules, and can be discerned for all the catalysts, except for the Pt/Ce0.9ZrFeAl system. Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 128 Apparently, all the doping agents promote the CO to CO2 oxidation in comparison to the unpromoted sample. The promotion origin depends on the doping metal, as for example, the iron presence boosts the CO reactivity via OH. In the same time, an enhance of the O species mobility occurs (reaction through support`s O species) no matter the doping agent, which is in a good agreement with the OSCC and OSC measurements for those samples. Also, important improvements concerning the deactivation species formation are achieved when both doping metals are added to the ceria matrix. While zirconia addition helps the rapid evacuation of the species susceptible to deactivate the catalysts from the surface, the iron incorporation significantly diminish their concentration. And when both are combined, the stability of the deactivation species decreases considerably, being apparently ruled by the zirconia presence. As a final TPD experiment WGS mixture containing CO and H2O in 1:1 volumetric ratio was passed through the samples at room temperature. After Table 3.12. Relative proportion of products after TPD-TPO % COOH % CO[o]sup % COtotal % COdeact Pt/CeAl 17.31 45.08 62.39 37.61 Pt/Ce0.9ZrAl 18.45 53.36 71.81 28.18 Pt/Ce0.9FeAl 26.56 55.10 81.66 18.34 Pt/Ce0.9ZrFeAl 25.00 57.00 82.00 18.00 Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 129 the adsorptions and once the signal stabilized the temperature was increased up to 600 ºC and the results resumed in Figure 3.19. Figure 3.19. CO and H2O adsorption at R. T, followed by WGS reaction During the adsorption step at room temperature, H2 formation is observed in absence of water signal suggesting that those H2 species should proceed from the re-oxidation processes performed within the oxygen vacancies proceeding through water dissociation. Once the surface saturated, the water signal starts to appear meanwhile CO and H2 signals decreased without formation of CO2. The reaching of surface reduction-oxidation equilibrium could be imagined. 0 3000 6000 9000 12000 15000 H2O adsorption WGS Reactor at R. T. Bypass ppm Time (s) Pt/Ce0.9FeAl 100 200 300 400 500 600 Temperature (oC) 0 3000 6000 9000 12000 15000 Pt/Ce0.9ZrFeAl H2O adsorption WGS Reactor at R. T. Bypass ppm time (s) 100 200 300 400 500 600 Temperature (oC) 0 3000 6000 9000 12000 15000 H2O adsorption WGS Reactor at R. T. Bypass ppm time (s) Pt/Ce0.9ZrAl 100 200 300 400 500 600 Temperature (oC) Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 130 With the start of the heating ramp, prominent water signal is observed accounting for the physically absorbed water on the surface. However, a second contribution also arises. At around 150 ºC the WGS reaction products CO2 and H2 start to appear. At the same time, the CO signal decays accompanied with a change in the slope of water desorption. Thus, the important changes of the water signal are due not only to physisorbed water but also to water consumptions via WSG reaction. At higher temperatures, a change in the behavior tendency of the H2O and CO species occurs. The water signal intensity increases, which could indicate WGS reaction limitation at higher temperatures. Actually, the H2 and CO2 signals exhibit a parallel diminution when the temperature is increased indicating the reverse WGS reaction. Besides, the H2 signal shows a diminution at around 450 ºC for all the samples which could be attributed to the support reduction in H2 presence as observed by the TPR experiments. Within the series of samples besides, the water desorption behavior change, some other differences could be also detailed in Figure 3.20. Although the water desorption process appears to be the main contribution on the water signal at lower temperature, some water species reacts with CO via WGS reaction. The maximum of activity is placed at around 350 ºC, temperature at which the water concentration reaches its minimum. Above this temperature the water concentration steadily increases following the WGS CO conversion equilibrium limitation with the temperature. Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 131 Figure 3.20. Water desorption profile under WGS conditions in the 180-600 oC temperature range The Pt/Ce0.9ZrFeAl sample exhibits the lowest water concentration between 300 ºC and 450 ºC which could be related to a possible higher activity in comparison to the other samples. Moreover, the water signal observed after the optimal WGS temperature range present different profiles. Apparently, the most active sample desorbs higher amounts of water with the increase of temperature. Considering the WGS equilibrium, higher partial pressure of water on the catalyst surface should lead to higher CO conversion. Thus, the better catalytic performance achieved for this sample could be associated to its apparently enhanced capacity to maintain more absorbed water on the surface. The water dissociation, the limiting step of the WGS reaction, should be easier when the adsorbed water concentration is maintained high, a case suggested for the Pt/Ce0.9ZrFeAl sample. 200 300 400 500 600 2000 3000 4000 5000 ppm H2O/gcat Temperature (oC) Pt/Ce0.9ZrAl Pt/Ce0.9FeAl Pt/Ce0.9ZrFeAl Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 132 3.4.1.8. CO adsorption and CO-TPD followed by FTIR spectroscopy The adsorption of CO over a catalyst’s surface is often followed by IR spectroscopy which constitutes a very powerful tool for supported transition and noble metals characterization. In the case of Pt, CO adsorption leads to IR spectra mainly composed by two (CO) adsorption regions: 2090-2000 cm-1 and 1860-1780 cm-1 characteristic respectively, to a linear CO adsorption on a single Pt atoms and to a a bridged CO adsorption on two Pt atoms. The chemical interaction CO-Pt was established by Blyholder [61,62] and it is considered that the bond proceeds through the 5σ bonding and 2π* antibonding orbitals of CO and Pt interaction. The 5σ–Pt interaction is electron donating towards Pt and strengths the Pt-C bond. The antibonding 2π*-Pt interaction corresponds to an electron transfers from Pt to the antibonding CO orbital, decreasing the C-O bond strength through increasing of the electron density on the 2π* orbital. Thus, higher the back donation from Pt to CO 2π*, lower the (CO) vibration. The (CO) band position depend on several factors such as: the number of Pt bonded to the CO molecules, the coordination of the Pt atoms in which the CO is adsorbed, the oxidation state of Pt, the CO coverage and the presence of electron donating or accepting centers in contact with the Pt particles. The combination of CO used as probe molecule, and IR spectroscopy are widely used to characterize the metal oxides such as cerium simple or mixed oxides [63,64]. Indeed, many studies have been carried out addressing the CO Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 133 ceria interaction and the Pt promotor effect. Contrasting to the assignments of Pt-CO interaction, the interactions are less clear for the ceria surfaces. For example, on highly ordered CeO2 (111) surfaces only weakly adsorbed CO was found suggesting some oxygen vacancies role on the CO-ceria interactions [65,66]. Similar conclusions on the importance of the oxygen defects were reported by Berner and Schierbaum [66]. They found that the CO adsorption was not possible on stoichiometric CeO2 supported on Pt(111) but, on a partially reduced ceria surfaces where several interactions were registered. Even though for Pt/ceria samples strong metal support interaction (SMSI) is commonly accepted there is not a scientific agreement concerning the molecular mechanism. The explanations are normally related to Pt decoration by a reduced ceria and/or to purely electronic effect arising from an intimate Pt ceria surface contact. O adsorptions till saturation followed by FTIR techniques are performed. For that, the samples were pretreated in H2 atmosphere at 350 ºC then, degassed at 500 ºC in secondary vacuum in order to clean the surface before the adsorption and then the CO was adsorbed in portions till saturation. The CO adsorption spectra are presented in the Figure 3.21 for all the samples. Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 134 Figure 3.21. Spectra obtained during the CO adsorption on the catalysts In all the samples, the CO adsorption leads to increased intensities of the Pt- CO bands at 2100 cm-1 indicating that all the platinum species were successfully reduced to Pt0 during the H2 pretreatment. Furthermore, several other species are observed in the 1800-1100 cm-1 region and in the hydroxyl region (3500-3200 cm-1). It should be noted that the bands in the former region, ascribed to carbonaceous species, increased more rapidly than the Pt- CO stretching band saturation. The bands placed at 1650, 1611, 1434, 1397, 1229 and 1219 cm-1 could be ascribed to different kinds of hydrogencarbonates, being also discerned its corresponding υ(OH) stretching vibrations Pt/Ce0.9ZrAl Absorbance (a.u.) 1200 1400 1600 1800 2000 Wavenumber (cm-1) 2 a. u. Pt/CeAl Absorbance (a.u.) 1200 1400 1600 1800 2000 Wavenumber (cm-1) 2 a.u. Pt/Ce0.9FeAl Absorbance (a.u.) 1200 1400 1600 1800 2000 Wavenumbers (cm-1) 0.05 a.u. Absorbance (a.u.) 1200 1400 1600 1800 2000 2200 Wavenumbers (cm-1) 2 a.u. Pt/Ce0.9ZrFeAl Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 135 at 3617 cm-1 [63,67,68]. In addition the bands situated at 1630 and 1550 cm-1 are attributed to carbonates species. The CO adsorption yields to bands at 1840 and 1790 cm-1 for all the catalysts, being their intensity significantly lower for the Pt/CeAl sample. The band placed at 1840 cm-1 is usually related to bridged carbonyls [69] but could also be attributed to bridged carbonates [2]. On the other hand the 1790 cm-1 band, could be related to tilt CO adsorbed on very dispersed Pt complexes [70]. Yee et al. [52] related those CO species to altered support properties and indicated that the ceria cations present in the supports influence the adsorption state of CO by enhancing the backbonding through the metal. Moreover, the main Pt-CO band exhibits significant changes on its position as a function of the introduced CO amount. As shown on the Figure 3.22.a, the adsorption of 1.4 mol of CO within the series of samples leads to different relative intensities and provokes a shift of the band position to higher wavenumbers as a function of the CO coverage. The coverage effect can be explained considering both, the static or chemical and the dynamic effect. The chemical effect considers the platinum species as electron reservoir. Thus, an increment of the adsorbed CO molecules should result in decrease of the back-donation which involves decrease of the Pt-CO strength appearing at higher (CO). The dynamic approach accounts for the coupling effect between the CO molecule dipoles which results in a shift to a Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 136 higher wavenumbers. Once the surface saturated all the samples present the same (CO) maximum at 2086 cm-1 (Figure 3.22.b) which indicates that the platinum particles distribution is for all the catalysts. The TPO experiment pointed out the formation of carbon deposits on the catalytic surfaces during the CO adsorption via Boudouard`s reaction. Although the C deposits could not be observed by IR directly an account for them could be intuited through the alteration of the CO adsorption capacity for the equally dispersed samples, i.e. samples with the same Pt particles size. If the Boudouard’s reaction occurs only on the Pt particles an indirect indication of the C formation could be obtained by the estimation of the CO adsorption capacity, in a way that higher the C deposit lower the surface of the site for the CO adsorption and thus lower the CO quantity for monolayer Figure 3.22. Pt-CO stretching at: a) 1.4 mol CO and b) saturation Absorbance (a.u.) 1960 2000 2040 2080 2120 Wavenumber (cm-1) 0.2 a.u. Pt/CeAl Pt/Ce0.9ZrAl Pt/Ce0.9FeAl Pt/Ce0.9ZrFeAl Normalized Absorbance (a.u.) 1950 2000 2050 2100 2150 Wavenumber (cm-1) Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 137 formation. Figure 3.23, shows the integrated Pt-CO stretching band area normalized by the Pt content as a function of the introduced CO. All the samples exhibit two different trends being the intersection between them commonly related to the monolayer formation. Figure 3.23. Pt-CO band area normalized by the Pt content vs. mol CO introduced From this point, it is possible to establish a relationship between the monolayer area and the Pt particles available for the adsorption of CO molecules. The following order of monolayer formation is observed: Pt/CeAl< PtCe0.9ZrAl< PtCe0.9ZrFeAl≤PtCe0.9FeAl This trend shows on first place that the introduction of any doping agents leads to improvement of the sample`s resistance to carbon deposits formation, being the iron containing solids and Pt/Ce0.9FeAl in particular especially resistant. The results obtained for the Pt/Ce0.9ZrFeAl sample are rather unexpected regarding the lowest CO2 amounts desorbed during the TPO 0123456 0 10000 20000 30000 40000 50000 CO-Pt band area normalized by gPt 0123456 0 5000 10000 15000 20000 25000 30000 Pt/Ce0.9FeAl mol CO in Pt/Ce0.9ZrFeAl 012345678 0 5000 10000 15000 20000 0 1 2 3 4 5 6 7 8 0 5000 10000 15000 20000 25000 30000 Pt/Ce0.9ZrAl CO-Pt band area normalized by gPt mol CO in Pt/CeAl Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 144 In order to elucidate the support effect, the catalytic activity was normalized by the mole of exposed platinum atoms by using the dispersion values obtained in CO chemisorption measurements. The turnover frequencies (TOF, s-1) are estimated for different GHSV and presented in Figure 3.28.b. The Pt/Ce0.9ZrFeAl sample appears now as the most active along all the tested GHSV range. However with the purpose to ensure the beneficial effect provided by the support and to discard the platinum content role, 2 wt.% of Pt was deposited on the same supports. For all the catalysts similar platinum contents were achieved. The catalytic behavior of the samples tested in real conditions at 280ºC and as a function of the GHSV is presented in Figure 3.29. Now the tendency in CO conversion terms confirms the beneficial effect of the support on the catalytic activity. However, when comparing the 2 wt.% series with the 4wt.% one similar conversions are observed suggesting that at the higher Pt loading and lower space velocity, the Pt atoms are not entirely involved in the reaction. Thus in agreement to all the characterization results the sample presenting lower cerium oxide particle size, the highest reducibility, the lowest BG energy and relatively high F2g/Ov ratio, pointing to the most important electronic and structural promotion is the most active sample. In addition lower degree of deactivation is expected for this sample on the basis of the TPD and the IR studies. Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 145 3.5. Partial conclusions Catalytic screening of a series of different ratio Ce/Zr based supports was successfully achieved. Within the series no matter the doping metal proportion the solid solution formation was detected. Although minor, the differences obtained in the catalytic performance of the samples allowed the choice of an optimal Ce/Zr molar ratio of 9. The improvement on the catalytic behavior on this catalyst was associated to the absence of platinum metallic diffraction which implies higher platinum dispersions and, as an outcome, higher catalytic activity. From that point the Ce/doping metal ratio was maintained and the zirconia and iron were incorporated as doping metals separately or in combined manner. For these samples and concerning the structural changes, the zirconium presence provokes larger ceria particle sizes meanwhile the iron causes the opposite effect. Figure 3.29. GSHV effect on the 2 wt. % Pt catalysts in % CO conversions terms 6000 12000 18000 24000 30000 36000 0 10 20 30 40 50 60 70 80 90 100 CO Conversion (%) GHSV (h-1) Pt(2%)/CeAl Pt(2%)/Ce0.9ZrAl Pt(2%)/Ce0.9FeAl Pt(2%)/Ce0.9ZrFeAl Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 146 Concerning to the Pt metal, all the samples are constituted by two types of Pt particles: one fraction highly dispersed and the other one agglomerates being the later detected on the XRD patterns (Figure 3.8). However, the HR-TEM and the CO chemisorption results suggest that the agglomerated fraction was significantly minor than the dispersed ones, above 1 nm. In good agreement to the XRD results, all the doped samples showed, on the CO-TPD/FTIR experiments, features related to highly active Pt species, strongly interacting with the support. The iron as promoter leads to especially enhanced OSC capacities, higher band gap contraction and higher oxygen vacancies population. Its promotion could be then mainly related to electronic and redox features also in concordance with the results observed by Raman spectroscopy. On the other hand, the zirconia promotion should be principally related to the Pt metal and it provokes the most important diminution of the size of the Pt agglomerates accounting for a strong metal support interaction confirmed also by the FTIR experiments. The combination of both Fe and Zr results in finely balanced structurally and electronically promoted system. The Pt/Ce0.9ZrFeAl sample shows beneficial promotions from both doping agents resulting in lower Pt sizes and the enhanced redox properties provided by Zr and Fe, respectively. Besides, the prominent band gap diminution detected from UV-Vis characterization suggesting also an important synergic effect achieved in the case of the Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 147 Pt/Ce0.9ZrFeAl sample. Taking into account the greater H2 consumption at lower temperatures and the significant H2 overconsumption showed by this sample during the TPR experiment, a superior mobility of the oxygen species was found (OH and O species). However, the oxygen vacancies population and the oxygen mobility ability for the Pt/Ce0.9FeAl sample were superior to that observed for the Pt/Ce0.9ZrFeAl ones. In addition a superior mobility of the H species was also found, resulting from a reverse spillover processes confirmed by the lower temperatures of H2 species desorption on the Pt/Ce0.9ZrFeAl sample during the H2-TPD experiment (Figure 3.15). Similar conclusions can be extracted from the FTIR experiments where the Pt/Ce0.9ZrFeAl sample exhibited higher capacities to activate and to transfer the CO species (Figure 3.22). In addition, the fact that the Pt/Ce0.9ZrFeAl sample shows higher water absorption capacity at higher temperatures should be also considered (Figure 3.20), and related to the improved H/OH species transfer. The introduction of doping agents modifies the stability of the carbonaceous species formed on the support (Figure 3.18). Furthermore, different features were obtained for the Pt/Ce0.9FeAl and Pt/Ce0.9ZrAl catalysts. While the former strongly decreases the desorbed amounts, the later leads to easier oxidations reactions at lower temperatures. Again, Pt/Ce0.9ZrFeAl combined both beneficial effects of the doping metals, exhibiting the lowest amount of desorbed CO2, issued from the carbonaceous species at similar temperatures Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 148 observed for the Pt/Ce0.9ZrAl. In good concordance, with the FTIR results where desorption occurs at secondary vacuum (Figure 3.26.b). Therefore, the combination of Zr and Fe on the Pt/Ce0.9ZrFeAl provides enough electron density to the Pt to activate the CO molecules but also a sufficient electron mobility to transfer the activated species to the support and finally, to desorb CO2 and H2. In fact, the enhanced catalytic performance of the Pt/Ce0.9ZrFeAl appears to be a conjunction of strong interaction and lower Pt particle size provided by Zr and to the prominent improvement of the redox properties offered by Fe. It should be also highlighted the strong diminution on the band gap and the easier H transferences found for this sample, which clearly indicates the achieved synergism by the combination of the promoters. As a consequence, the combination of both promoters allows not just better catalytic activity but also higher resistance towards deactivation effects. Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 149 References [1] T.R. Reina, S. Ivanova, V. Idakiev, J.J. Delgado, I. Ivanov, T. Tabakova, M.A. Centeno, J.A. Odriozola, Catal. Sci. Technol. 3 (2013) 779. [2] A. Holmgren, B. Andersson, D. Duprez, Appl. Catal. B Environ. 22 (1999) 215. [3] M.A. Centeno, M. Paulis, M. Montes, J.A. Odriozola, Appl. Catal. A Gen. 234 (2002) 65. [4] A. Trovarelli, Catal. Rev. 38 (2006) 439. [5] A. Martinez-Arias, J.M. Coronado, R. Cataluna, J.C. Conesa, J. Soria, J. Phys. Chem. B 102 (1998) 4357. [6] S. Damyanova, B. Pawelec, K. Arishtirova, M.V.M. Huerta, J.L.G. Fierro, Appl. Catal. A Gen. 337 (2008) 86. [7] M. Ozawa, H. Yuzuriha, M. Haneda, Catal. Commun. 30 (2013) 32. [8] J. Li, J. Chen, W. Song, J. Liu, W. Shen, Appl. Catal. A Gen. 334 (2008) 321. [9] Y.T. Kim, S.J. You, E.D. Park, Int. J. Hydrogen Energy 37 (2012) 1465. [10] F.J. Pérez-Alonso, M. Lopez Granados, M. Ojeda, P. Terreros, S. Rojas, T. Herranz, J.L.G. Fierro, Chem. Mater. 17 (2005) 2329. [11] T.R. Reina, W. Xu, S. Ivanova, M.Á. Centeno, J. Hanson, J.A. Rodriguez, J.A. Odriozola, Catal. Today 205 (2013) 41. [12] W.Y. Hernández, O.H. Laguna, M.A. Centeno, J.A. Odriozola, J. Solid State Chem. 184 (2011) 3014. Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 150 [13] O.H. Laguna, F. Romero Sarria, M.A. Centeno, J.A. Odriozola, J. Catal. 276 (2010) 360. [14] T.R. Reina, S. Ivanova, J.J. Delgado, I. Ivanov, V. Idakiev, T. Tabakova, M.A. Centeno, J.A. Odriozola, ChemCatChem 6 (2014) 1401. [15] J.A. Rodriguez, P. Liu, J. Hrbek, J. Evans, M. Pérez, Angew. Chemie - Int. Ed. 46 (2007) 1329. [16] J.C. Duchet, M.J. Tilliette, D. Cornet, Catal. Today 10 (1991) 507. [17] S. Damyanova, B. Pawelec, K. Arishtirova, M.V.M. Huerta, J.L.G. Fierro, Appl. Catal. B Environ. 89 (2009) 149. [18] S. Ricote, G. Jacobs, M. Milling, Y. Ji, P.M. Patterson, B.H. Davis, Appl. Catal. A Gen. 303 (2006) 35. [19] A.M. Duarte de Farias, D. Nguyen-Thanh, M.A. Fraga, Appl. Catal. B Environ. 93 (2010) 250. [20] H. Liang, H. Yuan, F. Wei, X. Zhang, Y. Liu, J. Rare Earths 29 (2011) 753. [21] T. Dhannia, S. Jayalekshmi, M.C. Santhosh Kumar, T. Prasada Rao, A. Chandra Bose, J. Phys. Chem. Solids 71 (2010) 1020. [22] P.C.A. Brito, D.A.A. Santos, J.G.S. Duque, M.A. Macêdo, Phys. B Condens. Matter 405 (2010) 1821. [23] L. Yue, X.M. Zhang, J. Alloys Compd. 475 (2009) 702. [24] O.H. Laguna, M.A. Centeno, M. Boutonnet, J.A. Odriozola, Appl. Catal. B Environ. 106 (2011) 621. [25] T.R. Reina, E. Papadopoulou, S. Palma, S. Ivanova, M.A. Centeno, T. Ioannides, J.A. Odriozola, Appl. Catal. B Environ. 150-151 (2014) Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 151 554. [26] G. Jacobs, U.M. Graham, E. Chenu, P.M. Patterson, A. Dozier, B.H. Davis, J. Catal. 229 (2005) 499. [27] D. Andreeva, I. Ivanov, L. Ilieva, M. V. Abrashev, Appl. Catal. A Gen. 302 (2006) 127. [28] M. Fadoni, L. Lucarelli, Stud. Surf. Sci. Catal. 120 (1999) 177. [29] J.Z. Shyu, K. Otto, J. Catal. 115 (1989) 16. [30] S. Damyanova, J.M.C. Bueno, Appl. Catal. A Gen. 253 (2003) 135. [31] M.J. Tiernan, O.E. Finlayson, Appl. Catal. B Environ. 19 (1998) 23. [32] O.H. Laguna, M.A. Centeno, G. Arzamendi, L.M. Gandía, F. Romero-Sarria, J.A. Odriozola, Catal. Today 157 (2010) 155. [33] H. Bao, X. Chen, J. Fang, Z. Jiang, W. Huang, Catal. Letters 125 (2008) 160. [34] M. Gonzalez Castaño, T.R. Reina, S. Ivanova, M.A. Centeno, J.A. Odriozola, J. Catal. 314 (2014) 1. [35] B.M. Reddy, P. Bharali, P. Saikia, S.-E. Park, M.W.E. Van Den Berg, M. Muhler, W. Grunert, J. Phys. Chem. C 112 (2008) 11729. [36] I. Zorić, M. Zäch, B. Kasemo, C. Langhammer, ACS Nano 5 (2011) 2535. [37] C.W.M. Castleton, J. Kullgren, K. Hermansson, J. Chem. Phys. 127 (2007) 1. [38] V. Petrovsky, B.P. Gorman, H.U. Anderson, T. Petrovsky, J. Appl. Phys. 90 (2001) 2517. Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 152 [39] N. Acerbi, S. Golunski, S.C. Tsang, H. Daly, C. Hardacre, R. Smith, P. Collier, J. Phys. Chem. C 116 (2012) 13569. [40] P. Patsalas, S. Logothetidis, L. Sygellou, S. Kennou, Phys. Rev. B 68 (2003) 1. [41] J. Spanier, R. Robinson, F. Zhang, S.-W. Chan, I. Herman, Phys. Rev. B 64 (2001) 1. [42] Y. Madier, C. Descorme, A.M. Le Govic, D. Duprez, J. Phys. Chem. B 103 (1999) 10999. [43] G. Balducci, P. Fornasiero, R. Di Monte, J. Kaspar, S. Meriani, M. Graziani, Catal. Letters 33 (1995) 193. [44] H. Vidal, J. Kašpar, M. Pijolat, G. Colon, S. Bernal, A. Cordón, V. Perrichon, F. Fally, Appl. Catal. B Environ. 30 (2001) 75. [45] P. Arnoldy, A. Moulijn, O.J. Wimmers, J. Phys. Chem 90 (1986) 1331. [46] K. Shyu, J Z, Otto, Appl. Surf. Sci. 32 (1988) 246. [47] W.C. Conner, J.L. Falconer, Chem. Rev. 95 (1995) 759. [48] S. Damyanova, C.A. Perez, M. Schmal, J.M.C. Bueno, Appl. Catal. A Gen. 234 (2002) 271. [49] T. Yamaguchi, J. Catal. 67 (1981) 324. [50] C.B.L. I, V. Cedex, A. Badri, J. Lamotte, E. Catherine, J.C. Lavalley, J. El Fallah, L. Hilaire, F. Normand, B. Pascal, S. Cedex, E. Quemere, G.N. Sauvion, 87 (1991) 1601. [51] J.L.G. Fierro, J. Soria, J. Sanz, J.M. Rojo, J. Solid State Chem. 66 Electronic and structural promoters of Pt(4%)/CeO2 /Al2O3 catalyst Chapter 3 153 (1987) 154. [52] A. Yee, S.J. Morrison, H. Idriss, 295 (1999) 279. [53] P. Panagiotopoulou, D.I. Kondarides, J. Catal. 260 (2008) 141. [54] P.G. Menon, G.F. Froment, J. Catal. 59 (1979) 138. [55] D.R. Mullins, K.Z. Zhang, Surf. Sci. 513 (2002) 163. [56] A. Holmgren, B. Andersson, J. Catal. 25 (1998) 14. [57] A. Goguet, F. Meunier, J.P. Breen, R. Burch, M.I. Petch, A. Faur Ghenciu, J. Catal. 226 (2004) 382. [58] E. Rogemond, R. Frety, V. Perrichon, M. Primet, S. Salasc, M. Chevrier, C. Gauthier, F. Mathis, J. Catal. 169 (1997) 120. [59] G. Pitchon, V. Zins, J. F., Hilaire, L., Maire, 59 (1996) 203. [60] C. Li, Y. Sakata, T. Arai, K. Domen, K. Maruya, T. Onishi, J. Chem. Soc. Faraday Trans. 1 85 (1989) 929. [61] G. Blyholder, J. Phys. Chem. 79 (1975) 756. [62] G. Blyholder, J. Phys. Chem. A 68 (1964) 2772. [63] C. Binet, M. Daturi, J.-C. Lavalley, Catal. Today 50 (1999) 207. [64] V. Sánchez Escribano, E. Fernández López, M. Panizza, C. Resini, J.M. Gallardo Amores, G. Busca, Solid State Sci. 5 (2003) 1369. [65] D. Mullins, J. Catal. 188 (1999) 340. [66] U. Berner, K. Schierbaum, G. Jones, P. Wincott, S. Haq, G. Thornton, Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 256 Confirming the specific rate trend, the TOF values for model mixture for the catalysts at 180 ºC show the superiority of gold sample. However, at 250 oC the platinum based catalysts doubled the TOF obtained for gold system with a similar dispersion, thus converting platinum in better choice in this temperature range. When the mixture is changed to realistic the difference between the gold and platinum at low temperature significantly decrease converting the platinum catalyst in the logical choice in the whole temperature range for these conditions. The evaluated gold-based catalysts presented higher CO rates and turnover frequencies at 180ºC than some well referenced Au/CeO2 and Au/CeO2-MOx systems tested in the same conditions [61,62]. Regarding the Pt samples, the specific reaction rates of our Pt based materials are superior compared to the results obtained by the group of Efstathiou using a similar WGS inlet stream [63]. Stability test in which the stream composition was varied from model to real stepwise was carried out over the Pt/CeFeAl and Au/CeFeAl samples (Figure 6.7). For both catalysts, two different temperatures were chosen in order to keep the same initials non equilibrium CO conversion with the purpose to discern the causes of catalyst’s behavior changes. In the case of Au/CeFeAl catalyst, the initial temperature (280 ºC) was maintained constant during all the experiment and, in the last step the GHSV was reduced by half. For Pt/CeFeAl sample, the initial temperature of 220 ºC was kept during the experiment, excepting the last step, where the temperature was increased to 280 ºC. Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 257 025 50 75 100 125 150 175 200 0 20 40 60 80 100 50% H2 30% H2 % CO Conversion Time (h) Model +9%CO +11%CO2 +15% H2 Real T= 280oC Figure 6.7. Stability test step by step from model to real at 4000h-1 for: a) Pt/CeFeAl and b) Adapted from Au/CeFeAl Regarding to the catalysts stability, the Au/CeFeAl presents slight and continuous decline of the catalytic activity with time on stream being especially accentuated under complete real conditions and 2000 h-1. On the other hand, the Pt/CeFeAl presents better stability in the final stage with no indications of catalyst deactivation in time on stream. Once the real mixture fed and temperature increased to 280 ºC, the catalyst exhibits CO conversion values around 80% corresponding of approximately 10% of CO conversion lost in respect to the conversion observed in the primary light off test (Figure 6.6.B). The influence of the gas stream composition, on the other hand, provides significant dissimilarities depending on the noble metal nature. The Au/CeFeAl exhibit a progressive diminution of the catalytic activity when the feed stream is changed stepwise from model to real mixture being the 050 100 150 200 250 300 350 0 20 40 60 80 100 +50% H2 +30% H2 +15% H2 +11%CO2 +9%CO Model Real CO conversion (%) Time (h) 2000 h-1 Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 258 decrement more emphasized when H2 is incorporated. The latter in concordance, with the Le Châtelier principle and with the above mentioned inability of the support to activate water molecules in H2 presence. Contrary to the Au based sample, its Pt homologue shows a significant decrease of the catalytic activity when the inlet CO concentration increases from 4.5 % to 9 %. This different behavior could be directly related to the different CO adsorption capacities of both metals. The CO adsorption strength can be associated to the sample’s band gap model (Table 6.3). Whereas the support band gap correspond to the energy difference between the O(2p) and the localized Ce(4f) states, the catalyst band gap should account for the difference in energy between the NM 5d states and the 4f levels of cerium. The catalytic activity of the samples is a function of this band gap, in such a way that the higher the band gap the higher the activity. The higher band gaps energies exhibited by the Pt/CeFeAl vs. the Au/CeFeAl suggests that the platinum 5d states fall closer to the O(2p) band which implies an easier electron transfer between oxygen atoms of ceria and the NM and therefore, this excess electron density weakens the CO bond strength through backbonding and facilitates the formation of the reaction intermediates. However, the stronger Pt-CO interaction appears to hinder excessively the CO2 desorption processes and leads to lower catalytic activities when the CO concentration is increased. Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 259 6.2.4. Partial conclusions Platinum and gold metals were supported on different oxide materials and the resulted catalysts characterized and tested in WGS conditions. From the characterization results, high surface areas were obtained in all the solids. Besides, the desired solid solution formation for CeFeAl was achieved and observed by XRD, UV-Vis and Raman spectroscopies. The catalytic test reveals enhanced performances, for both noble metals, when ceria was incorporated to the support being those improvements more emphasized when iron was used as doping agent of ceria. The advantages obtained from the iron incorporation were reflected on an enhanced catalytic activity. The improvement of the catalytic behavior of the catalysts was directly related to the electronic properties modifications achieved through solid solution formation and reflected on the band gap energies changes in a way that: higher the band gap higher the activity. Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 260 Figure 6.8. Catalytic behavior and electronic properties relationship. Adapted from [12] No matter the stream composition, the gold based catalyst present superior activity at 180ºC, shifting its operation window to the lowest temperature range. Nevertheless, the platinum based systems are much more versatile since they are able to withstand changes in the stream composition with higher TOF than its gold homologues and attained the equilibrium conversion at temperatures close to 250ºC. 6.3. Part II: Au/CeFeAl and Pt/CeFeAl catalysts supported on metallic micromonolithic devices Along the second part, the catalytic activity of selected Au and Pt ceria based catalysts was studied in structured systems. The deactivation effects are also Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 261 considered. Several deactivation processes were reported for ceria based catalysts, more precisely, the over-reduction of ceria and the formation of carbonaceous species blocking the catalysts’ active sites are the most frequents phenomena leading to activity depletion [4,26,64–67]. Although the catalytic performance can be almost fully recovered by O2/air treatment, such a thermal procedure is not suitable during continuous operational conditions [4,68,69]. On the other hand the addition of small quantities of oxygen within the WGS stream could be beneficial to mitigate the deactivation effects: the as called O2-assisted WGS. This oxygen introduction should extend the catalytic life-time leading to highly active and stable WGS catalysts. In the literature, most studies related to O2-assisted WGS concerns the bimetallic Cu-based catalysts. Some promising results were published by Kughai et al. [69] for bimetallic Pt-Cu/Ceria catalysts. This system remarkably enhances the H2 yields with the addition of oxygen, due to the fast CO oxidation to CO2 and liberation of the exposed metallic active sites, resulting in higher catalytic activity per site. The O2-assisted WGS reaction over Au/ceria and AuCu/ceria based catalysts was also studied by Gamboa-Rosales et al. [68] affirming that the water dissociation is favored by the O2 presence and highlighting the influence of the synthesis method on the gold catalysts performance. The group of Flytzani-Stephanopoulos also studied the Au and Pt powder catalysts in the O2 assisted WGS suggesting that for both metals, the active sites were nonmetallic particles but oxidized metal species [4]. However in the scientific Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 262 literature the effect of O2 on the WGS performance of structured catalysts, to the best of our knowledge, were not published. Then, this study is focused on the application of platinum and gold based structured catalysts in the WGS reaction aiming to overcome the reactor volume restrictions associated to the shift process in conventional packed bed reactors and taking advantage of the excellent mass and heat transfer properties provided by the structured systems. The effect of the structuring process on the powder catalysts is also a subject of this study. Moreover, the influence of the oxygen on the catalytic performance of both systems is studied and correlation of this behavior to the metal nature, catalyst state and/or reaction conditions are proposed. For that purpose, gold and platinum were supported on iron-doped ceria mixed oxide. The obtained catalysts were standardly characterized and supported on metallic micromonolithic devices through the washcoating method. The employed slurries were dried, calcined in the same conditions as the parent catalysts and then, characterized and tested in WGS conditions in order to detect any possible modifications induced on the initial catalysts during the washcoating method. 6.3.1. Experimental The same ceria – iron mixed oxide, Ce0.8Fe0.2O2, was supported on -Al2O3 dispersant matrix being the nominal chemical composition and the preparation method also maintained as described in section 6.2.1. 2 wt.% of platinum and gold were deposited on this support. The Pt catalyst was prepared keeping the Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 263 same synthesis method detailed on the section 2.xx but in this case, the gold catalyst was synthetized by precipitation with K2CO3. Similarly, the obtained support was labeled CeFeAl and the powder catalysts Pt/CeFeAl and Au/CeFeAl, respectively. The metallic micromonolithic devices and the slurry preparations were also prepared as explained in the Chapter 4, section 4.2. Once dried and calcined, the suspensions were called S_Pt/CeFeAl and S_Au/CeFeAl. The washcoating process was repeated till 1 g of each catalyst was deposited on the metallic substrate. The structured catalysts were labelled M_Pt/CeFeAl and M_Au/CeFeAl (where the “M” stays for micromonolith). Concerning the catalytic tests, the same GHSV=4000 h-1 was kept for all performed experiments. The powder samples were tested by employing catalyst volume of 1.5 cm3 and 600<<800 m. The WGS tests for the structured catalysts were performed in different stream composition, e.g. model mixture, O2-assisted model mixture, real mixture, O2-assisted real mixture and total CO oxidation mixture (TOX) reaction. The inlet compositions used for each experiment are summarized on Table 6.6. The platinum based catalysts were pre-activated during 2h at 350ºC in 10 vol.% H2 meanwhile the gold catalysts were not treated prior the reaction. All the catalytic data are presented in CO conversions terms. Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 264 6.3.2. Physicochemical characterization 6.3.2.1. Chemical composition and textural properties The chemical composition of the prepared solids is summarized in Table 6.7. All the samples presents very close composition to the targeted one. Table 6.7. Chemical composition and textural properties of the synthetized solids. Al2O3 (wt. %) CeO2 (wt. %) Pt o Au (wt. %) Fe2O3 (wt. %) SBet (m2/g) VPore (cm3/g) Al2O3 100 - - - 202 0.36 CeFeAl 78.9 18.9 - 2.2 172 0.38 Au/CeFeAl 77.9 17.8 2.6 1.7 182 0.40 S_Au/CeFeAl 80.2 16.2 2.3 1.3 185 0.41 Pt/CeFeAl 76.3 19.3 2.3 2.1 175 0.38 S_Pt/CeFeAl 78.1 17.7 2.2 2.0 181 0.40 Table 6.6. Experimental conditions of the catalytic tests Model Real O2-Model O2-Real TOX CO (vol%) 4.5 9 4.5 9 4.5 H2O (vol%) 30 30 30 30 - CO2 (vol%) - 6.3 - 6.3 - H2 (vol%) - 50 - 50 - N2 (vol%) Balance Balance Balance Balance Balance O2 (vol%) - - - 0.7 0.7 Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 265 The observed variations from the catalysts to slurries composition were expected because of the addition of colloidal alumina during the washcoating process, which increases the total alumina content and as a consequence decreases the contribution of the rest of the components. Textural properties of the synthetized samples are also presented in Table 6.7. Compared to bare -Al2O3, the ceria iron mixed oxide deposition decreases the specific surface area and increases the pore volume of the samples. Pt/CeFeAl and Au/CeFeAl showed intermediate values of surface area between alumina and CeFe/Al solids. However, opposite trends are observed for pore volume being increased those of gold catalyst and decreased that of platinum. The increased values for both area and pore volume for the slurries, in comparison to their corresponding support, are directly related to the use of Nyacol colloidal alumina during the slurries preparation. Despite of the detected textural variations, in general terms the textural properties are ruled by the presence of -Al2O3. Actually, the obtained values are in a good agreement with the theoretical ones when calculated supposing a physical mixture. 6.3.2.2. XRD analysis XRD patterns obtained for the obtained solids are presented in the Figure 6.9. All the samples show typical diffraction lines pertaining to -Al2O3 (JCDPS# 00-048-0267) and CeO2 (JCPDS#00-004-0593) oxides. For the CeFeAl system, no diffraction peaks are detected for iron species. The absence of the Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 272 WGS and CO oxidation reactions acting together to abate CO. In this context, the CO oxidation reaction was carried out for both noble metals employing the same CO/O2 ratio used in the O2 assisted WGS reaction. Both catalysts show the maximum possible CO conversion (above 30%) in the whole temperature range, stoichiometrically limited by the added O2 amount (0.7%). It should be noted that the conversion obtained in CO oxidation is exactly the enhanced effect obtained for the M_Pt/CeFe when O2 is added to the WGS feed stream at 180ºC with respect to direct WGS. Figure 6.13. CO conversions in obtained for the micromonolithic structures in: a) model and O2-model; b) TOX In agreement with Duprez et al. [72], the observed catalytic activity improvement appears to be uniquely due to an additive effect of both, WGS and total CO oxidation reactions. Therefore, in the case of Pt the beneficial effect of oxygen is not exactly an O2-assisted WGS but rather the result of two parallel reactions consuming CO. Moreover, a less pronounced difference 180 200 220 240 260 280 300 320 340 0 10 20 30 40 50 60 70 80 90 100 CO Conversion (%) Temperature (oC) M_Au/CeFeAl M_Pt/CeFeAl 180 200 220 240 260 280 300 320 340 0 10 20 30 40 50 60 70 80 90 100 Temperature (oC) M_Pt/CeFeAl O2- WGS M_Pt/CeFeAl WGS M_Au/CeFeAl O2-WGS CO Conversion (%) M_Au/CeFeAl WGS Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 273 regarding the catalytic behavior of the Pt catalyst depending on the conditions was observed when the temperature increases which could be explained by the preference of the M_Pt/CeFeAl catalyst to react via WGS instead through CO oxidation route. In turn, once the CO species adsorbed on the noble metal and/or the metal/support interface quickly removed via CO oxidation, the concentration of the available active sites to WGS reaction (slower reaction) should increase. On the contrary, the noticeable catalytic enhancement observed for the gold based monolith cannot be exclusively attributed to an additive effect from CO oxidation and WGS reactions. In fact, the CO conversion obtained in the O2-assisted WGS test exceeds by far the contribution of the direct CO oxidation (30%) in the whole temperature range. Consequently, this noticeable catalytic improvement can be clearly ascribed to an O2-assisted WGS reaction. The capacity exhibited by the gold based catalyst (and not for the platinum based catalysts) to be assisted by oxygen species during the WGS reaction could be partially ascribed to the redox reaction pathway described for gold ceria catalysts. Although the role of ceria support is clearly important no matter the NM (Pt or Au), it becomes absolutely necessary when gold is used as active phase due to the metal inability to dissociate water. The water dissociation is a slow step on the WGS reaction and, for gold based systems it must be carried out at the metal/support interface. The oxygen addition has been reported to favor the generation of surface OH groups leading to promoted WGS reaction [73,74]. Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 274 200 220 240 260 280 300 320 340 360 0 10 20 30 40 50 60 70 80 90 100 CO Conversion (%) Temperature (oC) M_Au/CeFeAl O2 - Real M_Pt/CeFeAl Real M_Pt/CeFeAl O2- Real The beneficial effect of the oxygen addition to the stream could be also related to the inhibition of the deactivation effects such as overreduction of the support and/or formation of carbonaceous species blocking catalytically active sites. Actually, it has been also reported that oxygen species facilitates the carbonaceous intermediate species decomposition avoiding the active sites blocking and pushing forward the WGS [7]. On the other hand, recent operando XAS study of an Au/CeO2 catalyst demonstrates that the optimum Ce oxidation state to carry out the WGS is 3.3 [55]. Given that under the WGS reaction atmosphere the ceria partial reduction is likely to happen; the presence of oxygen traces could help to maintain cerium partially oxidized and close to the optimum formal valence active in WGS. Finally, the catalytic behavior of both noble metals was tested on O2-assisted real conditions and the results are presented on Figure 6.14. Figure 6.14. CO conversions of M_Pt/CeFe and M_Au_CeFe catalysts: a) CO conversions at GHSV = 4000h-1 in WGS reaction in “real” and O2-assisted “real” conditions. Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 275 The M_Au/CeFeAl does not show any activity in the studied temperature range under real WGS mixture. The M_Pt/CeFeAl reaches the equilibrium conversions at the highest temperature. The oxygen incorporation to the real mixture does not present any advantage for the M_Pt/CeFeAl system, demonstrating the same CO conversions like in oxygen absence in the whole temperature range. It could be then suggested that the oxygen incorporation appears to affect mainly the processes occurring over the support, which is not of primordial importance when platinum is acting as active metal. Although the support capacities should be affected by the H2 presence the water dissociation ability of platinum in this case do not seems to be altered. On the other hand, the structured gold based catalyst presents even a negative tendency with the increase of the temperature, the CO conversions decrease with the temperature. Surprisingly, the highest catalytic conversion in real conditions presented by M_Au/CeFeAl is achieved at 180 ºC showing yields around 20%. This could be due to a favored H2 oxidation by gold in these conditions and, as stated above, to the impossibility of ceria reoxidation in real WGS conditions and its loss of capacity for water dissociation under these conditions. Comparing both noble metals in H2-rich streams, no matter the type of the catalytic bed, powder fixed bed or structured micromonolithic devices, the platinum based systems present clear superiority. Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 276 6.3.5. Partial conclusions An interesting comparison between structured gold and platinum noble metals supported on iron-doped ceria mixed oxide for WGS was carried out. The original powder catalysts were successfully deposited onto metallic micromonolithic structures achieving well fixed and homogeneous catalytic layer. Both metal based powders perform satisfactorily in the classical WGS, being Pt superior to Au in all cases. The scaling up from the powder to the structured catalysts seems to affect mainly to the gold based catalyst while the Pt monolith retains its excellent behavior, a fact not observed for the powder samples and reported here for the first time. O2 addition to the WGS feed stream changes the situation by boosting the WGS activity, greatly for Au and merely for Pt based monolith. Our data reveal that only the gold based monolith benefits for an actual “O2 assisted WGS” reaction meanwhile the CO conversion increment observed for the Pt monolith results from the parallel CO oxidation reaction. In any case, oxygen addition is a strategy to be considered to potentiate the WGS reaction over noble metals based structured catalysts especially for gold/ceria systems. Such an activity promotion together with the inherent features of the structured catalysts allowing smaller reactor volumes coverts our monoliths into an interesting alternative to be implemented in hydrogen fuel processors. Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 277 6.4. Part III: Operando XANES-EXAFS study of Au and Pt ceria based catalysts Under reaction conditions, the catalysts can undergo chemical transformations that severely transform their structure with respect to the initially obtained [75]. In situ/operando characterizations appear as interesting techniques allowing detailed analysis of the structural, electronic and chemical properties of the catalyst under reactions and/or operation conditions. The main difference between in-situ and operando relies on the catalytic activity measurement, which is only recorded in the operando approach. So, operando method must involve in-situ spectroscopy measurements under true catalytic reaction conditions. For that purpose, the WGS reactor cell should be designed considering the atmospheric pressures and the temperatures in which the WGS reaction is carried out. The main goal of the operando studies is to provide a complete picture of the catalytic process whereby an accurate structure-activity relationship of the substrate-catalyst species can be determined. Therefore, the purpose of operando spectroscopy is to measure the catalytic changes that happen within the reactor using time resolved and spatially resolved spectra [76]. Although time resolved spectra theoretically record the formation and disappearance of intermediate species at the active sites of the catalysts in real time, current operando instrumentation only works in second or sub-seconds scale and therefore, only relative concentrations of intermediates can be assessed. Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 278 Among the existing spectroscopic techniques, XAS (X-Ray Absorption Spectroscopy) is one of the most powerful, which provides direct information of the structural and electronic properties changes under the reaction conditions in the presence of reactants and products [77]. Concerning the WGS in-situ/operando studies, the lack of knowledge concerning basic questions such as nature of active species and reaction pathway hinders the proper design and optimization of the catalysts [59]. As mentioned in the General Introduction, an intense debate above the nature of the active species of metal based catalysts is still alive. Controversial results on the oxidation state of active gold and platinum catalysts supported on ceria can be found in literature. For example, Kim et al [78] proposed for gold based catalysts, that the active sites associated to nanocrystalline gold domains are much more active than those associated to atomically dispersed metal. These results challenge the observations obtained by the group of Flytziani-Stephanopoulous [14,21] that associates the WGS activity with atomically dispersed cationic gold species. In this regard, Rodriguez et al. [20] demonstrates that the WGS activity mainly relies on the pure gold nanoparticles oxide interface. Cationic clusters were also proposed as major active species in WGS for similar gold and palladium supported on ceria. In a similar way, the agreement on the oxidation state of the platinum species during the WGS reaction is not achieved. For low temperature WGS, Fu el al. Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 279 [14] claimed that the mechanism involves mainly cationic platinum species in the Pt-Ox-Ce sites. By XPS spectroscopy, the identified oxidation state was essentially PtII with some PtIV fraction, but not Pt0. From operando studies, opposite conclusions were stablished by Xu et al. [79] affirming that at temperatures at which the catalyst exhibits maximum conversion (300ºC), most of the platinum oxide is transformed into metallic platinum. The operando XAS study of the gold and platinum based catalysts were carried out in this work. Within the series of prepared catalysts in this thesis Au/CeAl(Puralox) and Pt/CeAl(Puralox) were the chosen ones for the XAS study. 6.4.1. Experimental: X-ray Absorption Spectroscopy Time-resolved Au, Pt and Ce L3-edges were collected in situ during the WGS reaction. Beamlines X18B of the NSLS at Brookhaven National Laboratory (BNL), BL22-CLAESS line of the CELLS-ALBA facility and SAMBA line at SOLEIL were used for the EXAFS/XANES experiments. The spectra were taken either in the transmission-yield or fluorescence mode for platinum. The samples were mounted without diluting in capillary glass reactor. A standard mass flow controllers and gas mixer managed the gas flow. H2O was introduced with a bubbler located at the exit of the gas mixer. The temperature of the inlet gas, controlled by heated gas pipes, determined H2O concentration. The reaction temperature was ramped at 10 ºC/min, from RT upwards. Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 280 6.4.2. XANES results First, the importance of the required pretreatment for the platinum based catalyst was evaluated. For that purpose, the activation conditions were reproduced being the catalyst exposed to H2 stream and heated at 10ºC/min until 350 ºC. The evolution of the Pt-L3 and Ce-L3 XANES spectra with the temperature is shown in the Figure 6.15. The characteristic PtO sharply rising absorption peak is observed at 11564 eV, (Figure 6.15.a), being its intensity diminished as the temperature is increased. This absorption peak, denominated the “white line”, directly accounts for the vacancies on 5d orbital when platinum species are oxidized. Higher the white line intensity, higher the platinum population in an oxidized state. 11560 11580 11600 11620 0,0 0,7 1,4 x Energy (eV) 25oC 100oC 200oC 300oC 50oC 150oC 250oC 350oC Pt-L3 5700 5720 5740 5760 5780 5800 0,0 0,7 1,4 Ce-L3 25oC 100oC 200oC 300oC 50oC 150oC 250oC 350oC x Energy (eV) Figure 6.15. XANES spectra obtained during the H2 preactivation treatment for: a) Pt-L3 absorption edge; b) Ce-L3 absorption edge Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 281 Therefore, for platinum metallic species the intensity of the absorption edge should decrease because of the complete occupation of the 5d orbitals and consequent minor probability of electron transitions. In fact, for platinum 5d orbitals nominally occupied, this transition should not be allowed because of the absence of holes. However, an s-p-d orbital hybridization allows the presence of low intensity white line in the XANES spectra for Pt bulk [80]. For the Pt/CeAl sample the Pt-L3 XANES results show clearly an oxidized platinum population before the introduction of H2. Moreover, the platinum species are reduced by H2 starting from room temperatures as indicated by the decrease of the intensity of the white line. At the final temperature of the treatment, which is also the temperature of pretreatment before the WGS reaction, completely reduced platinum state is reached. On the other hand, the modification of the Ce-L3 XANES absorption with the temperature when exposed to H2 rich streams is presented in the Figure 6.15.b. The XANES line shape of Ce3+ and Ce4+ are quite different [16]. At 25 ºC, Ce-L3 XANES spectrum show two absorption peaks. The lower energy absorption peak is commonly attributed to the absorption from 2p level to the 5d with the 4f orbital remaining unoccupied, resulting in Ce [2p5 4f05d1]O[2p6] state configuration. The second absorption characteristic of Ce4+ appears at higher energies and it is assigned to the transit from the oxygen valence band to the ceria 4f shell leaving a hole in the oxygen valence band, with final Ce[2p54f15d1]O[2p5] state configuration. Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 288  Ce-OH coordination number is almost free (tight bond between 3 and 5)  Coordination distances are variable, but linked to a single parameter that accounts for the variation of the lattice parameter (Δa), which is the same as the above mentioned bulk one.  Ce-OH coordination distance is free.  Debye-Waller factors are the same for each coordination shell as the bulk ones.  Ce-OH Debye-Waller factor is free. Therefore, the ceria model can be simplified as shown in the Figure 6.20 like a linear combination of bulk ceria and surface ceria. The main conceptual difference between them is the Ce-OH coordination shell accounted for the surface ceria species. Pt-L3 models The Pt model was constructed considering the XANES results where the fingerprint of the sample compared to that from metallic Pt at all temperatures Figure 6.20. Schematized model of ceria in which the ceria surface with the hydroxyl coordination shell and ceria bulk is combined Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 289 seem to be quite similar. Furthermore, the Fourier transform magnitudes of the spectra (Figure 6.24) revealed a first zone composed of what it looks like an interference from different peaks, that seem to reduce their sharpness once the temperature rises. However, as we can see in the Pt metal reference, the first Pt-Pt coordination shell gives indeed a weird arrange of peaks and shoulders by itself instead of a simple peak, so the choice of a simple Pt metal model seemed wise. The further shells are quite similar between the different spectra, so relevant changes seem to only happen at the first shell. However, the first test fits revealed a strong reduction of the coordination numbers at all Pt coordination shells, being more severe at longer distances, ranging from 30 to 70% approximately. Even with this, the fits were not as good as required. Thus, as one would expect an intimate contact between the CeO2 support and the Pt-metal active phase, an additional Pt-Ce shell was included in the fit, giving better results; with a resulting Ce mean coordination number around 1 or bigger. This means that the Pt atoms are most probably in very small nanoparticle form, attached firmly (maybe growing epitaxially) to the CeO2 crystal. There is an interesting match between the Pt-metal crystal structure and the CeO2 (100) surface. On the (100) surface of CeO2, the tetrahedral sites on the surface (where the O atoms would stand if we add a new layer on the surface) are disposed in a squared lattice, with a lattice parameter of 5.41/2 = 2.705 Å. On the other hand, the pure Pt metallic crystal has an fcc structure with a lattice parameter of 3.92 Å, which means a first coordination shell of 2.77 Å. Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 290 Taking into account that, in general, metallic nanoparticles tend to reduce their coordination distances as they reduce size, it seems that an epitaxial growth of small clusters of Pt on the (100) surface of the CeO2 support might be quite likely. In this case, the first Pt layer would have a Pt-Ce coordination number of 2. As the Ce mean coordination number calculated with the first model was around 1, we concluded that the Pt atoms will most likely grow epitaxially as bi or tri-layers on top of the (100) CeO2 surface, following the Pt-fcc structure through the [100] direction. This is, building the squared lattice at the first layer by filling the tetrahedral holes of the CeO2, and then filling the cubic holes in the second and third layers. For more flexibility, a parameter that accounts for the filling of the second and third layer respect to the first is introduced, thus allowing partially reconstructed layers, or even only monolayers. Besides, this parameter would take into account layers built on other CeO2 surfaces, like (110) and (111), which would give scattering signals similar to the (100) monolayer, so the model is not as restrictive as it seems. In order to take into account the borders of the layers due to the limited lateral size of the nanoparticles, a model of squared layers was built for simplicity. The variable taken to account for this was the total number of atoms in the nanoparticle, and then the size of each layer was calculated from the filling proportion, and supposing them squared in shape. Then, the coordination numbers are calculated for each atom, taking into account: 1. The layer at which they belong (bottom, intermediate or top) Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 291 2. Whether they are at the interior of the shell, at the border, or at the corner This results in a complex coordination number formula for each coordination shell, which, however, only depends on 3 variables, that will finally give the mean shape and size of the nanoparticle, in the squared-trilayer model. A scheme of the model can be seen in Figure 6.21. More precisely, the model was constructed like this: Pt bottom layer parameters:  The first 4 Pt-Pt coordination shells were taken both the single scattering paths and the multiple-scattering paths of the 4th shell. The Figure 6.21. Model for the Pt-L3 EXAFS analysis. Pt atoms form epitaxial squared layers growing on top of a Ce (100) surface, with different covering proportions. The atoms color codes are: white for Ce, red for O, grey for Pt Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 292 Pt atoms at the intermediate and top layers (described below) were obviously taken into account.  Coordination numbers are fixed to the ones calculated for an infinite bilayer: 8, 4, 8 and 4, respectively (crystalline 3D Pt is 12, 6, 12, 24)  Coordination distances are variable, but linked to a single parameter that accounts for the variation of the first coordination distance, and then coherent with the structure.  The Debye-Waller factors are independent for each shell.  The first two Pt-Ce coordination shells were taken, only single scattering paths. These are in positions (1/2, 1/2, -1/2 - Δh) and (1/2, 3/2, -1/2 - Δh), being Δh the increment of height of the Pt layer from the CeO2 (100) plane, starting from the O position (see below).  Coordination numbers are fixed to the ones of the (100) surface: 2 and 4  Coordination distances depend firstly the coordination distance of the Pt-Pt lattice, and secondly, on a new parameter that accounts for the height of the Pt layer hovering over the CeO2 layer. The starting positions are the O positions, thus the increment of height should be always positive, as the Pt atoms are bigger.  The Debye-Waller factors are independent for each shell. The multiple-scattering path has the same Debye-Waller factor as its corresponding single-scattering path. Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 293 Pt intermediate layer parameters: At the second layer there are 2 different Pt according to their positions respect to the Ce atoms. One of them is directly on top of a Ce, and the other in the rest of the positions. Most of the parameters are the same of the bottom layer, only changing the parameters corresponding to the Pt-Ce shells.  The first 4 Pt-Pt coordination shells were taken both the single scattering paths and the multiple-scattering paths of the 4th shell. The atoms at the first layer were obviously taken into account. Coordination numbers, distances and Debye-Waller factors are the same as the ones for the bottom layer, above described.  For the Pt atoms on positions on top of the Ce positions (half of the top layer), the two first Pt-Ce shells were taken. The first has only a single-scattering path, while the second has single and multiplescattering paths through some of the bottom Pt atoms.  Coordination numbers are fixed to the ones of the (100) surface: 1 and 4  Coordination distances depend on the coordination distance of the Pt-Pt lattice, and on a the height of the Pt layer hovering over the CeO2 layer (see above).  The Debye-Waller factors are: for the first shell, the same as the first Pt-Ce of the bottom layer. This is justified as the distances are similar. The second shell, independent, but the same for the single and multiple-scattering paths. Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 294  For the Pt atoms on rest of the positions, only the first Pt-Ce coordination shell was taken, being the second too far.  Coordination number is fixed to the one of the (100) surface: 4  Coordination distance depend on the coordination distance of the Pt-Pt lattice, and on the height of the Pt layer hovering over the CeO2 layer (see above).  The Debye-Waller factor is the same as the second Pt-Ce shell of the bottom layer, due to the similar distance. Pt top layer parameters: The third layer stands too far from the Ce atoms, and thus it is modelled as a simple Pt layer, taking into account their neighbour Pt atoms in the layers below.  The first 4 Pt-Pt coordination shells were taken both the single scattering paths and the multiple-scattering paths of the 4th shell. The atoms at the layers below are taken into account. Coordination numbers, distances and Debye-Waller factors are the same as the ones for the bottom layer, above described. Although it seems a quite sophisticated model, the number of variables is kept quite reduced due to the restrictions imposed, and the long k and R range used in the fit. There are a total of 12 variables (Number of atoms, 2 layers ratio, 2 distances, 7 Debye-Waller factors) of a possible maximum of 23 given by the Nyquist theorem, which seems quite good. Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 295 Au-L3 model A model similar to the ceria ones was developed for fitting the EXAFS function derived from the gold spectra. Distinction between the internal and the external part of the gold nanoparticles is imposed for the analysis as represented in Figure 6.22. For gold particles, the key parameter was the surface/bulk fraction which provides information about gold dispersion and dynamics during the WGS. Moreover, the intimate gold-ceria contact was also considered. So, the Au bulk was fitted using the fcc structure of metallic gold meanwhile, for the Au surface, gold atoms were placed over the oxygen atoms in the (110) plane of CeO2. More precisely, the model proposed for data analysis of the gold particles can be summarized:  Au bulk: regular Au metal fcc structure including the first five coordination spheres.  Fixed coordination numbers Figure 6.22. Model for the Au-L3 EXAFS analysis in which is represented, from left to right, a total gold dispersion, an intermediate situation and a agglomerated gold situation Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 296  Variable coordination distances, linked to a single variable (lattice parameter increment)  Variable Debye-Waller factors, one for each coordination sphere.  Au surface: dispersed Au atoms at the center of the 4 Oxygen atoms at the (110) surface plane of the CeO2 considering the first two coordination shells (the first one correspond to Au-O and the second one to Au-Ce)  Fixed Oxygen coordination number reduced by mean number of vacancies and fixed cerium coordination number  Variable coordination distances  Variable Debye Waller factors. The EXAFS oscillations for Ce-L3 for both noble metals based catalysts during the WGS reaction are presented on Figure 6.23. Both samples showed two distinct peaks depending on the R value. One at low-R, which is in fact a double peak or a peak with a shoulder that will probably be the result of the interference between the Ce-O(1) shell of the bulk CeO2 structure with the Ce- OH shell of the surface Ce, which should have different coordination distances. The second peak should be the combination of the Ce-Ce shell with the Ce-O(2) shell. The signal in between is small enough to be disregarded, and is probably due to truncated Fourier transform tails from the main peaks and to multiple-scattering signals, so it will not be fit. Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 297 The Fourier transform magnitude of the spectra (Figure 6.23.a) confirms that the ceria does not exhibit any significant change during the WGS reaction in the case of the Pt sample. So, once the sample is exposed to H2 pretreatment, the ceria structure remains similar during all the WGS reaction. This fact is opposite to that observed for the gold based catalyst in which the ceria shows prominent changes. The characteristic PtO sharply rising absorption peak is observed at 11564 eV, (Figure 6.15.a), being its intensity diminished as the temperature is increased. This absorption peak, denominated the “white line”, directly accounts for the Figure 6.23. Ce-L3 EXAFS R space spectra obtained when heated in a WGS reaction: a) Pt/CeAl; b) Au/CeAl Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 304 catalyst atains the equilibrium conversions and hence, the observed changes on the Pt-Ce shells and the increment on the catalytic performances could be related to the increment of the catalytic activity. Once more, the Debye-Waller factors for Pt-Ce and Pt-Pt shells follow quite a regular increase as temperature rises, in accordance with the higher dynamical disorder at higher temperatures. The EXAFS spectra of gold may be fitted just assuming Au-Au neighbors. The average coordination number of gold continuously decreases till 300ºC the number of gold atoms surrounding the Au scattered approaches zero, indicating that Au atoms redispersed and, even, stabilized probably by the ceria oxygen vacancies. However, this interaction must be considered electrostatic since no evidence of white line is observed in the Au XANES spectra. 6.4.3. Partial conclusions The oxygen vacancies of ceria play a multirole in the reaction. On one hand, they are active sites for water splitting and, on the other hand, they act as preferential nucleation points for noble metal deposition allowing high metallic dispersion and improving the catalytic performance. The strong metal support interaction involves a charge transfer from the structural defects towards NM atoms, modulating the activity of the catalysts. In view of the later, the creation of oxygen vacancies on ceria is desirable to achieve high Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 305 catalytic performances. It is evident, from the Figure 6.6, that the support affects more importantly the Au catalysts than the Pt ones. The average oxidation state of cerium atoms in Pt catalysts remain almost unchanged during WGS whatever the reaction temperature, while in the case of gold catalysts higher oxidation states are observed for cerium at low temperatures. This different behavior may be understood considering that the platinum catalysts are H2-reduced prior to the WGS and that the average oxidation states of cerium in gold catalysts follow the behavior of cerium atoms in the platinum catalysts upon reduction. For the reduced platinum catalysts the EXAFS oscillations account just for Pt- Pt interactions; the average coordination number of platinum remaining almost unaltered on increasing the reaction temperature (the Debye Waller factor may account for the lower intensities on increasing temperature). However, at low R values ~1.5 Å oscillations that decrease on increasing temperature are observed. These oscillations may be fitted to Pt-OH bonds accounting for water splitting on the metal surface even at the higher temperatures. Besides, the Pt particles could growth epitaxially being constituted by three layers in which the Pt atoms placed at the bottom ones occupied positions which initially should be occupied by Oxygen atoms of the ceria lattice. Variations regarding the Pt-Ce coordination were also detected being related to the different relative sizes of the intermediate and the bottom Pt atoms Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 306 layer. At lower temperatures, the intermediate layer appears to be composed by fewer atoms than the bottom ones. At 200-250 ºC and coinciding to important increment on the achieved CO conversions, both layers seem to be constituted by the same Pt atoms. The later variation on the relative sizes of the layers were accompanied by Pt-Ce distances diminutions. Concerning the gold dynamics, the proportion of Au surface atom respect to that of Au bulk increases significantly when the temperature rise up from 25 ºC to 100 ºC suggesting a complete gold redispersion. That total gold redispersion was accompanied to the increment of the Au-Ce and Au-O signals which could implies than gold atoms are located on the oxygen vacancies. The result is that metallic gold atoms may be placed at the oxygen vacancies being surrounded of oxygen atoms. Considering that the water activation steps should procced on the oxygen vacancies, gold particles at intermediate temperatures could avoid the water dissociations step. Both platinum and gold catalysts evidence the existence of a coordination sphere at ~2 Å that corresponds to the interaction of hydroxil groups with the ceria surface. In the case of gold this adsorption starts to be evident from 150 ºC upwards and decrease again once 300 ºC is reached. On the contrary for the platinum catalyst, the ceria/OH interaction remains within the fitting in the whole temperature range, but slightly reduced at higher temperatures. This behavior is consistent with the average oxidation state of cerium estimated Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 307 from the XANES spectra indicating that water adsorption/splitting is favored over ceria oxygen vacancies. Operando X-ray absorption data of NM/CeO2/Al2O3 catalysts allow establishing that the activity increase observed upon supporting noble metals on ceria modified supports must be ascribed to the ability of ceria for adsorbing/splitting water molecules. This may explain why in the case of Au, which cannot dissociate water, the presence of oxygen vacancies probably at the metal/ceria interface, promotes the process to the active site and increases the reaction rate. On the other hand, platinum may activate water by itself but the required reduction pretreatment increases the number of oxygen vacancies in ceria increasing the number of water molecules at the metal nanoparticles periphery. Since the WGS reaction is controlled by water diffusion, the increase in the number of support surface sites that may hold water molecules increase the reaction rate at lower temperatures accounting for the observed increase in the catalytic activity. Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 308 References [1] M. Haruta, Catal. Today 36 (1997) 153. [2] M.A. Edwards, D.M. Whittle, C. Rhodes, A.M. Ward, D. Rohan, M.D. Shannon, G.J. Hutchings, C.J. Kiely, Phys. Chem. Chem. Phys. 4 (2002) 3902. [3] D. Andreeva, V. Idakiev, T. Tabakova, A. Andreev, R. Giovanoli, Appl. Catal. A Gen. 134 (1996) 275. [4] W. Deng, M. Flytzani-Stephanopoulos, Angew. Chemie - Int. Ed. 45 (2006) 2285. [5] D.L. Trimm, Appl. Catal. A Gen. 296 (2005) 1. [6] C. Zhang, A. Michaelides, D.A. King, S.J. Jenkins, J. Phys. Chem. C 113 (2009) 6411. [7] T.R. Reina, S. Ivanova, V. Idakiev, J.J. Delgado, I. Ivanov, T. Tabakova, M.A. Centeno, J.A. Odriozola, Catal. Sci. Technol. 3 (2013) 779. [8] K.R. Hwang, S.K. Ihm, J.S. Park, Fuel Process. Technol. 91 (2010) 729. [9] G. Dutta, U. V. Waghmare, T. Baidya, M.S. Hegde, K.R. Priolkar, P.R. Sarode, Chem. Mater. 18 (2006) 3249. [10] J.A. Rodriguez, P. Liu, J. Hrbek, J. Evans, M. Pérez, Angew. Chemie - Int. Ed. 46 (2007) 1329. [11] W.Y. Hernández, M.A. Centeno, F. Romero-sarria, J.A. Odriozola, J. Phys. Chem. C 113 (2009) 5629. [12] M. Gonzalez Castaño, T.R. Reina, S. Ivanova, M.A. Centeno, J.A. Odriozola, J. Catal. 314 (2014) 1. Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 309 [13] A. Luengnaruemitchai, S. Osuwan, E. Gulari, Catal. Commun. 4 (2003) 215. [14] Q. Fu, H. Saltsburg, M. Flytzani-Stephanopoulos, Science 301 (2003) 935. [15] D. Tibiletti, A. Amieiro-Fonseca, R. Burch, Y. Chen, J.M. Fisher, A. Goguet, C. Hardacre, P. Hu, D. Thompsett, J. Phys. Chem. B 109 (2005) 22553. [16] G. Jacobs, E. Chenu, P.M. Patterson, L. Williams, D. Sparks, G. Thomas, B.H. Davis, Appl. Catal. A Gen. 258 (2004) 203. [17] R. Burch, Phys. Chem. Chem. Phys. 8 (2006) 5483. [18] A.A. Phatak, N. Koryabkina, S. Rai, J.L. Ratts, W. Ruettinger, R.J. Farrauto, G.E. Blau, W.N. Delgass, F.H. Ribeiro, Catal. Today 123 (2007) 224. [19] N. Dupont, G. Germani, A.C. van Veen, Y. Schuurman, G. Schäfer, C. Mirodatos, Int. J. Hydrogen Energy 32 (2007) 1443. [20] J.A. Rodriguez, S. Ma, P. Liu, J. Hrbek, J. Evans, M. Pérez, Science 318 (2007) 1757. [21] M. Flytzani-Stephanopoulos, B.C. Gates, Annu. Rev. Chem. Biomol. Eng. 3 (2012) 545. [22] P. Liu, J.A. Rodriguez, J. Chem. Phys. 126 (2007) 164705. [23] I.D. González, R.M. Navarro, W. Wen, N. Marinkovic, J.A. Rodriguéz, F. Rosa, J.L.G. Fierro, Catal. Today 149 (2010) 372. [24] R.J. Farrauto, Y. Liu, W. Ruettinger, O. Ilinich, L. Shore, T. Giroux, Catal. Rev. 49 (2007) 141. [25] W. Ruettinger, O. Ilinich, R.J. Farrauto, J. Power Sources 118 (2003) 61. Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 310 [26] X. Liu, W. Ruettinger, X. Xu, R. Farrauto, Appl. Catal. B Environ. 56 (2005) 69. [27] S. Fiorot, C. Galletti, S. Specchia, G. Saracco, V. Specchia, Int. J. Chem. React. Eng. 5 (2007). [28] S. Ivanova, C. Petit, V. Pitchon, Appl. Catal. A Gen. 267 (2004) 191. [29] F. Moreau, G.C. Bond, Appl. Catal. A Gen. 302 (2006) 110. [30] Z. Kónya, V.F. Puntes, I. Kiricsi, J. Zhu, J.W. Ager, M.K. Ko, H. Frei, P. Alivisatos, G.A. Somorjai, Chem. Mater. 15 (2003) 1242. [31] O.H. Laguna, F. Romero Sarria, M.A. Centeno, J.A. Odriozola, J. Catal. 276 (2010) 360. [32] S. Damyanova, B. Pawelec, K. Arishtirova, M.V.M. Huerta, J.L.G. Fierro, Appl. Catal. A Gen. 337 (2008) 86. [33] F.J. Pérez-Alonso, M. Lopez Granados, M. Ojeda, P. Terreros, S. Rojas, T. Herranz, J.L.G. Fierro, Chem. Mater. 17 (2005) 2329. [34] A.S. Reddy, C.Y. Chen, C.C. Chen, S.H. Chien, C.J. Lin, K.H. Lin, C.L. Chen, S.C. Chang, J. Mol. Catal. A Chem. 318 (2010) 60. [35] M.Á. Centeno, C. Portales, I. Carrizosa, J.A. Odriozola, Catal. Letters 102 (2005) 289. [36] C.W.M. Castleton, J. Kullgren, K. Hermansson, J. Chem. Phys. 127 (2007) 1. [37] V. Petrovsky, B.P. Gorman, H.U. Anderson, T. Petrovsky, J. Appl. Phys. 90 (2001) 2517. [38] P. Patsalas, S. Logothetidis, L. Sygellou, S. Kennou, Phys. Rev. B 68 (2003) 1. [39] S. Tsunekawa, T. Fukuda, A. Kasuya, Surf. Sci. 457 (2000) L437. Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 311 [40] S. Deshpande, S. Patil, S.V. Kuchibhatla, S. Seal, Appl. Phys. Lett. 87 (2005) 1. [41] L. Wu, H. Wiesmann, A. Moodenbaugh, R. Klie, Y. Zhu, D. Welch, M. Suenaga, Phys. Rev. B 69 (2004) 1. [42] O.H. Laguna, M.A. Centeno, M. Boutonnet, J.A. Odriozola, Appl. Catal. B Environ. 106 (2011) 621. [43] O.H. Laguna, M.A. Centeno, G. Arzamendi, L.M. Gandía, F. Romero-Sarria, J.A. Odriozola, Catal. Today 157 (2010) 155. [44] J. Spanier, R. Robinson, F. Zhang, S.-W. Chan, I. Herman, Phys. Rev. B 64 (2001) 1. [45] W.Y. Hernández, O.H. Laguna, M.A. Centeno, J.A. Odriozola, J. Solid State Chem. 184 (2011) 3014. [46] T. Tabakova, L. Ilieva, I. Ivanov, R. Zanella, J.W. Sobczak, W. Lisowski, Z. Kaszkur, D. Andreeva, Appl. Catal. B Environ. 136-137 (2013) 70. [47] G. Graham, J. Catal. 130 (1991) 310. [48] C.M. Kalamaras, I.D. Gonzalez, M. Navarro, L.G. Fierro, A.M. Efstathiou, J. Phys. Chem. C 115 (2011) 11595. [49] Z. V. Popović, Z.D. Dohčević-Mitrović, N. Paunović, M. Radović, Phys. Rev. B - Condens. Matter Mater. Phys. 85 (2012) 1. [50] M.A. Centeno, M. Paulis, M. Montes, J.A. Odriozola, Appl. Catal. A Gen. 234 (2002) 65. [51] C.T. Campbell, C.H.F. Peden, Science 309 (2005) 713. [52] C. Zhang, A. Michaelides, D.A. King, S.J. Jenkins, Phys. Rev. B - Condens. Matter Mater. Phys. 79 (2009) 1. [53] C. Zhang, A. Michaelides, D.A. King, S.J. Jenkins, J. Chem. Phys. Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 312 129 (2008) 1. [54] Z. Yang, Z. Lu, G. Luo, Phys. Rev. B - Condens. Matter Mater. Phys. 76 (2007) 1. [55] T.R. Reina, W. Xu, S. Ivanova, M.Á. Centeno, J. Hanson, J.A. Rodriguez, J.A. Odriozola, Catal. Today 205 (2013) 41. [56] W. Ruettinger, X. Liu, R.J. Farrauto, Appl. Catal. B Environ. 65 (2006) 135. [57] A. Martinez-Arias, J.M. Coronado, R. Cataluna, J.C. Conesa, J. Soria, J. Phys. Chem. B 102 (1998) 4357. [58] G. Bond, Gold Bull. 42 (2009) 337. [59] J.A. Rodriguez, J.C. Hanson, D. Stacchiola, S.D. Senanayake, Phys. Chem. Chem. Phys. 15 (2013) 12004. [60] A. Phatak, W. Delgass, J. Phys. Chem. C 113 (2009) 7269. [61] Z.Y. Yuan, V. Idakiev, A. Vantomme, T. Tabakova, T.Z. Ren, B.L. Su, Catal. Today 131 (2008) 203. [62] D. Andreeva, I. Ivanov, L. Ilieva, M. V. Abrashev, Appl. Catal. A Gen. 302 (2006) 127. [63] K.C. Petallidou, A.M. Efstathiou, Appl. Catal. B Environ. 140-141 (2013) 333. [64] X. Wang, R.J. Gorte, J.P. Wagner, J. Catal. 212 (2002) 225. [65] C. Li, Y. Sakata, T. Arai, K. Domen, K. Maruya, T. Onishi, J. Chem. Soc. Faraday Trans. 1 85 (1989) 1451. [66] J.M. Zalc, V. Sokolovskii, D.G. Löffler, J. Catal. 206 (2002) 169. [67] A. Gayen, M. Boaro, C. De Leitenburg, J. Llorca, A. Trovarelli, J. Understanding the NM role in WGS reaction: Pt vs. Au Chapter 6 313 Catal. 270 (2010) 285. [68] N.K. Gamboa-Rosales, J.L. Ayastuy, M.P. González-Marcos, M. a. Gutiérrez-Ortiz, Catal. Today 176 (2011) 63. [69] J. Kugai, J.T. Miller, N. Guo, C. Song, J. Catal. 277 (2011) 46. [70] V.G. Milt, S. Ivanova, O. Sanz, M.I. Domínguez, A. Corrales, J.A. Odriozola, M.A. Centeno, Appl. Surf. Sci. 270 (2013) 169. [71] O.H. Laguna, M. González Castaño, M.A. Centeno, J.A. Odriozola, Chem. Eng. J. 275 (2015) 45. [72] J. Barbier, D. Duprez, Appl. Catal. B Environ. 3 (1993) 61. [73] T. Utaka, K. Sekizawa, K. Eguchi, Appl. Catal. A Gen. 194 (2000) 21. [74] G.C. Chinchen, M.S. Spencer, K.C. Waugh, D.A. Whan, J. Chem. Soc. Faraday Trans. 1 83 (1987) 2193. [75] T. Bunluesin, R.J. Gorte, G.W. Graham, Appl. Catal. B Environ. 15 (1998) 107. [76] S.J. Tinnemans, J.G. Mesu, K. Kervinen, T. Visser, T.A. Nijhuis, A.M. Beale, D.E. Keller, A.M.J. Van Der Eerden, B.M. Weckhuysen, Catal. Today 113 (2006) 3. [77] D.C. Koningsberger, B.L. Mojet, G.E. Van Dorssen, D.E. Ramaker, Top. Catal. 10 (2000) 143. [78] C.H. Kim, L.T. Thompson, J. Catal. 230 (2005) 66. [79] W. Xu, R. Si, S.D. Senanayake, J. Llorca, H. Idriss, D. Stacchiola, J.C. Hanson, J.A. Rodriguez, J. Catal. 291 (2012) 117. [80] G. Jacobs, U.M. Graham, E. Chenu, P.M. Patterson, A. Dozier, B.H. Davis, J. Catal. 229 (2005) 499. Resumen 320 1. Introducción Ante el problema medioambiental derivado del continuo uso de los combustibles fósiles así como de su previsible agotamiento, se establece la necesidad de encontrar nuevas fuentes energéticas más respetuosas con el medio ambiente. Entre las posibles alternativas, el H2 se presenta como un vector energético prometedor debido a su alta eficiencia energética, limpieza y versatilidad ante los combustibles que los alimente. El creciente interés científico ha conllevado el desarrollo de celdas de combustible capaces de convertir la energía química de la molécula de H2 en energía eléctrica destacando las celdas de combustible tipo PEM (Polymeric Exchange Membranes) ya que han sido descritas como las más eficaces siendo capaces de trabajar a temperaturas relativamente bajas. Sin embargo, el correcto funcionamiento de las pilas de combustible tipo PEM pasa por la casi completa eliminación del CO ya que éste se depositaría en el ánodo de las celdas PEM conllevando la desactivación de la misma. Considerando que en la mayoría de los métodos de producción de H2, este viene acompañado de CO, surge la necesidad de desarrollar de procesos de purificación de corrientes de H2. La reacción de Desplazamiento del Vapor de Agua o WGS (𝐶𝑂 + 𝐻2𝑂 ↔ 𝐻2+ 𝐶𝑂2) consiste en la reacción del CO con el H2O para dar CO2 e H2. De esta manera, la reacción de WGS permite disminuir la concentración de CO a la par que incrementa el rendimiento total hacia H2. Además de ser un Resumen 321 equilibrio moderadamente exotérmico, la reacción de WGS se caracteriza por ser cinéticamente favorecida pero termodinámicamente limitada [1]. Así, el rango óptimo de temperaturas de operación deberá permitir unas altas conversiones de CO no limitadas por el equilibrio. Además, dicha temperatura ha de favorecer la cinética de la reacción que vendrá determinada por el catalizador empleado. Industrialmente, la reacción de WGS se lleva a cabo en dos reactores en serie denominados HTS y LTS (del inglés, High y Low Temperature Shift, respectivamente). Sin embargo, estos catalizadores presentan serias desventajas en cuanto a la implementación de la tecnología del H2 a dispositivos móviles [2]. Por ejemplo, los catalizadores clásicos de la unidad de HTS están basados en Cr6+ que combinando una alta toxicidad y solubilidad en agua, resulta en un metal poco apropiado para una aplicación móvil. Por otra parte, los catalizadores basados en Cu son típicos de las unidades de LTS pero son pirofóricos en contacto con el aire lo que implicaría un cuidadoso control de tales dispositivos. Además, la sustitución de dos unidades en serie HTS y LTS por una única etapa a una temperatura intermedia, MTS (Medium Temperature Shift) se presenta como una interesante propuesta que permitiría intensificar el proceso a través de una reducción importante del volumen necesario para llevarlo a cabo. Debido a los inconvenientes que presentan los catalizadores clásicos, nuevas formulaciones basadas en metales nobles se han desarrollado en las últimas Resumen 322 décadas. En este contexto, los catalizadores que usan el Pt como metal noble han sido descritos como prometedores especialmente cuando son combinados con óxidos reducibles como el CeO2. En general, las publicaciones se focalizan en la influencia del método de síntesis, las modificaciones del soporte, la importancia de los cambios estructurales, el estado de oxidación del metal noble durante la reacción o el mecanismo de reacción a través del cual se da principalmente la reacción de WGS. Olympiou et al. [3] correlacionó la actividad de catalizadores basados en Pt, Pd y Rh soportados sobre alúmina para la reacción de WGS con la concentración de metal superficial y la fuerza de enlace del CO sobre el metal. El orden de actividad propuesto fue Pt > Rh > Pd, el cual coincide con el establecido por Grenoble et al. [4]. Sin embargo, ninguno de estos estudios consideraron la capacidad intrínseca del Pt para disociar agua [5]. El comportamiento catalítico de los catalizadores basados en Pt, y también el de otros metales nobles, experimenta notables mejoras cuando son soportados sobre óxidos reducibles [6–9]. De hecho, los óxidos reducibles no son simples expectadores en la reacción de WGS sino que juegan un importante papel [10–13]. Para catalizadores basados en Pt/Cerio, Phatak et al. [14] reportó TOFs 30 veces superiores que los obtenidos para catalizadores basados en Pt/Alúmina atribuyéndose dicha mejora al par redox Ce3+/Ce4+. La fuerte interacción Pt-Cerio indicada reducibilidad del cerio superficial tras la adición del metal noble, fue reconocida como una indicación del efecto Resumen 323 electrónico ejercido por los sitios Pt-Ox-Ce, presumiblemente responsables de su elevada actividad catalítica. En la misma línea, Tauster et al. [15] introdujo el término SMSI (del inglés, strong metal support interaction) para describir la contribución del óxido reducible a la formación del sitio activo. De manera general, ha sido establecido que tanto el metal como el soporte juegan un papel esencial en la reacción de WGS. Por otra parte, también es generalmente aceptado que la actividad catalítica de las partículas metálicas soportadas viene fuertemente influenciada por el tamaño de partícula, la forma de las mismas y el estado de oxidación de que presenten. El comportamiento catalítico viene además, fuertemente modulado por la naturaleza y el grado de reducción del soporte y/o de la interfaz metal-soporte. Típicamente, las modificaciones del soporte se llevan a cabo con el objetivo de modificar sus propiedades estructurales o electrónicas de manera que se consigan mejoras en las actividades catalíticas de los catalizadores[16–20]. La sinergia entre la el óxido de cerio y el agente dopante puede con seguirse a través de la formación de solución sólida, de la interacción de óxidos segregados [21,22] o de la deposición de partículas metálicas en la superficie del cerio [23,24]. Numerosos estudios han demostrado que dopando el cerio con elementos con diferentes radios iónicos (Zr4+) o estados de oxidación (Fe3+) se afectan las propiedades electrónicas del cerio como la reducibilidad, la capacidad de Resumen 324 almacenamiento de oxígeno o la resistencia frente a fenómenos de sinterización [25,26]. La capacidad redox de las muestras determinara su actividad catalítica. De este modo, la habilidad para la formación de sitios tipo Ce3+- Ov· se ha relacionado íntimamente con la reducibilidad así como con la velocidad de oxidación de CO. Sin embargo, el uso de catalizadores basados en metales nobles, en principio más caros que los catalizadores convencionales de las unidades de LTS normalmente basados en óxidos de metales de transición, sólo será económicamente viable si presentasen mayores resistencias frente a los fenómenos de desactivación. Para catalizadores basados en Pt, la sinterización del metal noble ha sido descrita como una importante causa de desactivación [27–29]. Este fenómeno puede ser evitado modificando el soporte de manera que se logren interacciones más fuertes entre el metal y el soporte y, en consecuencia, mayores dispersiones de Pt [30,31]. Comparado con los catalizadores tipo Pt/Al2O3, la incorporación del cerio conlleva significantes mejoras catalíticas. La fuerte interacción Pt/Ceria ejemplificada por el incremento de la reducibilidad de las especies de cerio superficiales se reconoce como indicativo del efecto electrónico demostrado por sitios del tipo Pt-Ox-Ce que probablemente sean responsables de la química redox de estos sistemas [15,32]. Resumen 325 Además de la sinterización del metal noble, existen otras causas de desactivación para los catalizadores basados en cerio como la sobrereducción del soporte que eliminaría el juego redox del par Ce3+/Ce4+ [33]. Además, el bloqueo de los sitios activos del soporte por deposición de especies carbonáceas (tipo carbonatos, bicarbonatos, etc) o de carbón formado mediante la reacción de Boudouard suponen un serio inconveniente. La disminución de dichos efectos puede lograrse mediante el dopaje de la red del cerio consiguiendo una mayor resistencia frente a los fenómenos de desactivación y por lo tanto, una prolongación del tiempo de vida del catalizador. Por otra parte, junto con la actividad y estabilidad catalítica de los catalizadores, existen otras dificultades intrínsecas de la reacción de WGS que deben ser tenidas en cuenta especialmente si se desean implementar la tecnología del H2 a dispositivos portátiles. Normalmente, un requisito principal de los catalizadores de WGS para alcanzar conversiones aceptables de CO es trabajar a bajos tiempos de residencia lo cual supone un importante inconveniente ya que conlleva elevados volúmenes de reactor [34]. En comparación con las aplicaciones estáticas convencionales, una implementación tecno-económica exitosa pasa por una reducción del volumen del reactor de al menos dos órdenes de magnitud. En este aspecto, el estructurado de los catalizadores en micromonolitos aparece como un planteamiento interesante [35,36]. Los beneficios de los Resumen 326 catalizadores estructurados basados en metales nobles fueron establecidos por Wagner et al. [2] donde publicaban catalizadores estructurados capaces de alcanzar conversiones de equilibrio a 20000 h-1. En estas condiciones, los catalizadores convencionales de LTS basados en Cu-ZnO no lograrían conversiones de equilibrio. Además, el lecho de partículas de los catalizadores en polvo conllevaría unas enormes pérdidas de carga si trabajasen a tales velocidades espaciales. Los micromonolitos pueden ser cerámicos o metálicos (Figura R.1). Los micromonolitos metálicos presentan sustanciales ventajas con respecto a los cerámicos como por ejemplo: una mejor distribución del flujo, una mayor velocidad de difusión de los reactivos hasta el catalizador así como mayores resistencias mecánicas. Además, los monolitos metálicos presentan mayores conductividades térmicas lo cual implica un mayor control de la temperatura de reacción ya que disipan mejor el calor producido durante la exotérmica reacción de WGS [37]. Figura R.1. Micromonolitos metálicos comparados con cerámicos Resumen 327 Teniendo en cuenta lo descrito, los micromonolitos metálicos son excelentes candidatos para estructurar catalizadores para la reacción de WGS debido a su elevada conductividad térmica, elevada área superficial por unidad de volumen y fácil fabricación. Normalmente, los micromonolitos metálicos se preparan co-enrollando una lámina lisa y una rizada en torno a un eje de manera que se consiguen estructuras cilíndricas compuestas por canales longitudinales perpendiculares al eje de dimensiones menores de 1 mm de diámetro. Con respecto a los micromonolitos cerámicos, los metálicos presentan valores de cpsi superiores lo cual consiente depositar mayores cantidades de catalizador a la par que se mantienen un espesor de capa catalítica tal que conserve las capacidades de transferencias de materia y de calor proporcionadas por los micromonolitos. Sin embargo, los micromonolitos presentan como principal desventaja la elevada dificultad del anclaje del catalizador en las paredes del monolito debido a la baja rugosidad de las mismas. Entre los sustrato metálicos posibles, los de FeCrAlloy ofrecen la posibilidad de incrementar la rugosidad de la misma mediante el crecimiento de una capa de -Al2O3 de manera que se favorece una óptima deposición y anclaje del catalizador en polvo al substrato metálico [38]. Además, las aleaciones de FeCrAlloy muestran buenas estabilidades en cuanto a los fenómenos corrosivos de carburación que se pueden dar en atmósferas tan reductoras como las de la reacción de WGS. Resumen 328 Tras el pretratamiento inicial de las estructuras metálicas, la deposición del catalizador en las mismas se llevó a cabo empleando el método de washcoating. La óptima ejecución del método de washcoating implica un cuidadoso control de las propiedades reológicas de la suspensión o slurry [39]. Por ello, se ha de tener en cuenta la cantidad de sólido que se añade así como su diámetro partícula de manera que sea lo suficientemente pequeño como para que se mantenga en suspensión. Además, se pueden añadir aditivos como alúmina coloidal o PVA que favorezcan la estabilidad de la misma de forma que se logren capas catalíticas más homogéneas. En este contexto, la presente tesis doctoral pretende desarrollar catalizadores basados en Pt y óxido de Ce más activos y estables. Además, se presta especial atención a los procesos de estructurado de este tipo de catalizadores sobre micromonolitos metálicos. Por último, se comparan Au y Pt ya que ambos metales nobles han sido descritos como prometedores para la reacción de WGS. 2. Síntesis y caracterización de catalizadores de Pt: screening catalítico El objetivo del primer capítulo era seleccionar el catalizador de partida para el desarrollo de la tesis doctoral. Para ello, se depositaron dos cantidades distintas de Pt (2 y 4 p/p.%) sobre dos soportes comerciales proporcionados por Sasol: una -Al2O3 de alta superficie y un soporte CeO2/Al2O3 (Puralox). Además, se preparó un soporte basado en CeO2/Al2O3. Los catalizadores se Resumen 329 caracterizaron y se ensayaron en la reacción de WGS. Así, se intentó correlacionar la actividad catalítica de los sistemas preparados con la carga de platino y la naturaleza del soporte. 2.1. Síntesis de los materiales El soporte CeO2/Al2O3 “home-made” fue preparado por impregnación húmeda. Para ello, se partió del Ce(NO3)3·6H2O como precursor del óxido de cerio; éste fue disuelto en etanol y añadido a una -Al2O3 de alta superficie. A continuación, se eliminó el exceso de disolvente en el rotavapor, se seco en estufa a 120 ºC y se calcinó durante 4h a 450 ºC con una rampa de calentamiento de 10 ºC/min. Las cantidades obtenidas fueron muy próximas a las nominales siendo éstas un 20 p/p.% de óxido de cerio sobre alúmina. Este soporte se denominó CeAl HM. El Pt se depositó sobre los soportes por impregnación húmeda partiendo del precursor ([Pt(NH3)4(NO3)2]. Dicho precursor fue mezclado en proporción molar 1:1.1 con un ácido acético 1M. Tras la eliminación del exceso de disolvente en el rotavapor, los catalizadores se secaron en estufa a 120 ºC y por último, se calcinaron a 350 ºC durante 8h empleando una rampa de 5 ºC/min. La nomenclatura empleada para los catalizadores basados en alúmina fue Pt(2%)/Al2O3 y Pt(4%)/Al2O3. Para los catalizadores basados en CeO2/Al2O3 Resumen 336 010 20 30 40 50 60 70 80 90 Pt/CeAl Pt/Ce0.9ZrAl Pt/Ce0.9FeAl Pt/Ce0.9ZrFeAl R.I. (a. u.) 2 Al2O3 CeO2 5000 10000 15000 20000 25000 30000 35000 40000 0,02 0,04 0,06 0,08 0,10 0,12 0,14 0,16 0,18 0,20 0,22 0,24 0,26 0,28 TOFs (s-1) GHSV (h-1) Pt/CeAl Pt/Ce0.9FeAl Pt/Ce0.9ZrAl Pt/Ce0.9ZrFeAl La actividad catalítica de los catalizadores, en presencia de H2 y CO2, se evaluó a distintas velocidades espaciales (Figura R5). Mientras que el Fe como promotor parecía no conllevar ninguna mejora del comportamiento catalítico del sistema, el Zr sí que conseguía incrementar las conversiones de CO. Sin embargo, la combinación de ambos metales dopantes en el catalizador Pt/Ce0.9ZrFeAl lograba superiores rendimientos catalíticos. Figura R5. a) TOFs exhibidos por los catalizadores en condiciones “reales” en función de la velocidad espacial Figura R4: a) Micrografía HR-TEM del Pt/CeAl reducido; b) DRX de los catalizadores reducidos Pt a b Resumen 337 Con el objetivo de dilucidar los cambios observados en función de la naturaleza del soporte en la actividad catalítica tanto los soportes como los catalizadores fueron examinados. Así, las propiedades electrónicas de los soportes fueron analizadas por espectroscopía UV-VIS y Raman. En presencia de Fe, ambas técnicas indicaban modificaciones electrónicas importantes, por ejemplo, los cambios del espectro Raman exhibidos por el soporte Ce0.9FeAl con respecto al CeAl parecía responder a efectos electrónicos y no estructurales presentando la concentración más alta de vacantes de oxígeno en la superficie. Además, este soporte exhibía una disminución importante del valor de band gap del CeO2 (calculado empleando la función de Kubelka-Munk) con respecto al Ce0.9ZrAl. Sin embargo, la combinación de ambos agentes dopantes provocaba una brusca contracción del mismo indicando un efecto sinérgico (pues era superior a la suma de las disminuciones provocadas por los metales dopantes por separado) probablemente provocado por la incorporación de los niveles de ambos dopantes entre los del Ce y el O lo cual implica unas transferencias de carga especialmente facilitadas entre el cerio y el oxígeno de la red del CeO2. Resumen 338 Además, las reducibilidades de los catalizadores fueron estudiadas mediante TPR-H2. En todos los casos, se observó una reducción favorecida del soporte en presencia de Pt probablemente debida a la capacidad de spillover del Pt. Cuando se calcularon los porcentajes de reducibilidad, asumiendo que el Pt2+, Ce4+ y Fe3+ se reducían respectivamente a Pt0, Ce3+ y Fe0, se obtuvieron valores en torno al 100% indicando un sobreconsumo en todos los casos si tenemos en cuenta los picos de Pt metálico observados en DRX (Figura R.6.a). Estos sobre consumos se atribuyen a la incorporación de especies de H en la red del cerio formando especies tipo bronce de la forma ref. Con respecto a la reducibilidad, la combinación de ambos agentes dopantes conllevaba a sobreconsumos especialmente elevados tanto teniendo en cuenta el porcentaje de reducibilidad total (118 %) como la mayor reducibilidad relativa observada durante la primera zona de reducción que aparece con la subida de temperaturas. Tabla R.1. Energías de band gap y relaciónes Ov/F2g calculadas para los soportes preparados BG (eV) Ov/F2g CeAl 2.36 0.034 Ce0.9ZrAl 2.33 0.036 Ce0.9FeAl 2.01 0.065 Ce0.9ZrFeAl 1.65 0.056 Resumen 339 0 20 40 60 80 ZI ZII ZIII Pt/CeAl Pt/Ce0.9ZrAl Pt/Ce0.9FeAl Pt/Ce0.9ZrFeAl Asimismo, esta mayor facilidad para reducirse en presencia de H2 y/o incorporar especies de H a su red, parecía ser reversible. En el TPD- H2, el segundo y más importante proceso de desorción de H2 se solapaba con el primero sólo en el caso del Pt/Ce0.9ZrFeAl. Por otra parte, se estudió por FT-IR la interacción del CO con los catalizadores de Pt. Para ello, los catalizadores se redujeron y desgasificaron previamente a la adsorción de CO. Tras las sucesivas dosis de CO, se saturaron y se desorbieron. Cuando comparamos los espectros a bajas absorciones de CO (aproximadamente, 1.4 mol) se observó para todas las muestras el crecimiento de bandas que se pueden clasificar como: i) adsorción lineal de moléculas de CO sobre especies de Pt reducidas y ii) formación de especies carbonáceas, tipo hidrogenocarbonatos o carbonatos, en el soporte (Figura R.7) Además del desplazamiento observado en la banda de Pt-CO Table R2. Reducibility percentages (RP) obtained for supports and catalysts RP (%) Pt/CeAl 91 Pt/ Ce0.9ZrAl 105 Pt/ Ce0.9FeAl 100 Pt/ Ce0.9ZrFeAl 118 Figura R.6. a) % Reducibilidad total de los catalizadores; b) %Reducibilidad relativa Resumen 340 debido a los diferentes cubrimientos, si comparamos el desarrollo relativo de ambas regiones (Pt-CO versus especies carbonáceas en el suporte) el catalizador Pt/Ce0.9ZrFeAl parecía constituir una mayor cantidad de especies en el soporte con respecto a las que dejaba adsorbidas sobre el Pt. Esto podría indicar que el Pt/Ce0.9ZrFeAl presenta una mayor tendencia a transferir las especies de CO, una vez activadas, al soporte mediante la reacción bien con especies OH o O presentes en el soporte. Además, en la Figura R.8 se muestra la comparación de los espectros obtenidos para la los catalizadores tras la saturación (R.8.a) y cuando son sometidos a vacío secundario (R.8.b). Como puede observarse, el Pt/Ce0.9ZrFeAl parece tener una mayor capacidad para desorber las especies formadas en el soporte con respecto a las demás. Figura R.7. Comparación de los espectros obtenidos para los catalizadores cuando se adsorben 1.4 mol de CO Absorbance (a.u.) 1200 1400 1600 1800 2000 2200 Wavenumber (cm-1) 2 a.u. Pt/CeAl Pt/Ce0,9ZrAl Pt/Ce0,9FeAl Pt/Ce0,9ZrFeAl Resumen 341 De acuerdo a la observada capacidad de desorción, el catalizador Pt/Ce0.9ZrFeAl mostró, durante un TPD-CO, desorciones de CO2 e H2 a más baja temperaturas que el resto de los catalizadores indicando una mayor capacidad para descomponer especies carbonáceas hacia los productos de reacción de la WGS. De hecho, el anclaje de dichas especies en el soporte se reconoce como una causa importante de desactivación para catalizadores basados en cerio ya que bloquean los sitios activos del soporte conllevando así, a una pérdida de actividad y finalmente, a la desactivación del mismo. Algunas consideraciones más con respecto a los fenómenos de desactivación pudieron obtenerse a partir del TPO que se llevó a cabo después del TPD-CO para todas las muestras (Figura R.9). Durante el TPO, el sistema Pt/Ce0.9ZrFeAl exhibió una menor cantidad de CO2 que el resto de los Figura R.8. Comparación de los espectros obtenidos para los catalizadores: a) Saturados; b) Vacío secundario Absorbance (a.u.) 1200 1400 1600 1800 2000 2200 Wavenumber (cm-1) 5 a.u. Absorbance (a.u.) 1400 1600 1800 2000 2200 Wavenumber (cm-1) 1200 2 a.u. Pt/CeAl Pt/Ce0.9ZrAl Pt/Ce0.9FeAl Pt/Ce0.9ZrFeAl Resumen 342 100 200 300 400 500 600 Pt/Ce0.9ZrFeAl Pt/Ce0.9FeAl Pt/Ce0.9ZrAl 259oC 369oC 251oC 68oC 47oC ppm CO2/gcat Temperature (oC) 500 ppm 124oC Pt/CeAl /2 catalizadores de la serie. En este caso, el CO2 puede proceder de la oxidación de C formado mediante la reacción de Boudouard o bien de especies como bicarbonatos o carbonatos que se forman en el soporte. Por lo tanto, el catalizador Pt/Ce0.9ZrFeAl presentaba una menor tendencia total al bloqueo de los sitios activos del soporte. En conclusión, la combinación de ambos promotores conseguía un catalizador más activo y con una mayor resistencia frente a dichos fenómenos de desactivación estando ambos fenómenos relacionados con la mayor movilidad de especies de H bien fueran OH o H. Figura R.9. Evolución de la m/z=44 exhibida por los catalizadores durante el TPO Resumen 343 4. Estructurado de catalizadores Pt(2%)/CeO2/Al2O3: Introducción de una capa buffer Este capítulo de la tesis doctoral se centra en intentar mejorar la disociación del agua ya que se considera la etapa limitante de la reacción de WGS. Además, este planteamiento se combina con el proceso de estructurado de catalizadores en micromonolitos metálicos atendiendo a los beneficios que ofrecen para reacciones lentas y exotérmicas como la WGS así como el interés que presenta como etapa intermedia ante un posible escalado. De esta forma, se estructuraron cantidades crecientes de un catalizador basado en Pt/Cerio en una serie de micromonolitos metálicos y analizó si se conseguía mejorar la actividad catalítica mediante el estructurado así como la importancia de los fenómenos de transporte en los catalizadores estructurados. Tras esto, se comparó uno de los catalizadores estructurados constituidos por una sola capa de catalizador (Pt(2%)/CeO2/Al2O3) con uno constituido por dos capas diferentes: una capa de CeO2/Al2O3 depositada encima del sustrato metálico seguido de una capa de Pt(2%)/CeO2/Al2O3. Los diferentes comportamientos catalíticos obtenidos en función del catalizador estructurado y las condiciones de reacción fueron estudiados. 4.1. Síntesis de los catalizadores estructurados En primer lugar se prepararon las estructuras metálicas enrollando dos láminas de acero ferrítico, FeCrAlloy, una lisa y otra rizada en torno a un eje. Resumen 344 Tras esto, los micromonolitos se calcinaron durante 22 h a 900 ºC con el objetivo de generar una capa de -Al2O3 que incremente la rugosidad del substrato metálico y favorezca la adhesión del catalizador. Por su parte, como catalizador se usó Pt(2 p/p.%)/CeO2/Al2O3. Tanto para el soporte del catalizador como el buffer, se empleó un CeO2/Al2O3 comercial (Puralox) que contenía un 20 p/p.% de CeO2. El Pt se depositó por impregnación húmeda siguiendo el procedimiento descrito anteriormente. La nomenclatura empleada para el catalizador y el buffer fue Pt/CeAl y CeAl. Los polvos, catalizador y buffer, se impregnaron en los micromonolitos metálicos mediante el método de washcoating. Las suspensión estaban compuestas por 1.96 p/p.% de una solución de PVA (1.5 p/p.% PVA en agua), 17.65 p/p.% de Al2O3 coloidal (Nyacol, 20 p/p.% en agua), 18.14 p/p.% de catalizador y 62.25 wt.% de agua. Los sólidos añadidos a la suspensión fueron previamente molidos de manera que presentasen tamaños de en torno a 10 m. Las propiedades reologicas fueron modificadas añadiendo aditivos, PVA y alúmina coloidal, para conseguir unas deposiciones más homogéneas y suspensiones más estables ya que estos actúan como “binders” y surfactantes. Una vez depositados, los monolitos fueron calcinados en las mismas condiciones que sus polvos iniciales (350 ºC – 8h en el caso del Pt/CeAl y 450 ºC – 4h en el caso del bufer CeAl) pero empleando una rampa de calentamiento de 2 ºC/min de manera que se evitase el resquebrajamiento de Resumen 345 la capa de catalizador debido a los diferentes coeficientes de expansión que presentan estos frente al substrato metálico. Las suspensiones fueron secadas y calcinadas en las mismas condiciones. Tras esto, las suspensiones del catalizador Pt/CeAl y el buffer CeAl se denominaron S_Pt/CeAl y S_CeAl, respectivamente. Por una parte, se sintetizaron 3 micromonolitos en los que se depositaron 0.3, 0.6 y 1g de catalizador. La nomenclatura usada para estos monolitos fue M0.3, M0.6 y M1, respectivamente. Por otra, se sintetizó otro catalizador estructurado que contenía una capa de 0.3g de buffer CeAl seguida de una capa que contenía 0.3 g de catalizador Pt/CeAl. A este micromonolito se le denominó M0.3-B0.3. 4.2. Resultados y discusión En primer lugar y con el objetivo de determinar si se producen cambios importantes durante el método de washcoating, se testaron los catalizadores Pt/CeAl y S_Pt/CeAl en model mixture y a GHSV = 4000 h-1 (Figura R.10.a). De acuerdo con los resultados obtenidos en DRX, ambos sistemas presentaban actividades catalíticas similares indicando la ausencia de cambios importantes durante la preparación de la suspensión. Las mínimas diferencias podrían ser atribuidas a la menor carga de Pt presente en la suspensión debido a la adición de alúmina coloidal durante la síntesis de la misma. Resumen 352 Cuando se observan los órdenes de reacción propuestos en literatura para catalizadores de Pt/Al2O3 y/o Pt/CeO2, las mayores diferencias se detectan cuando la corriente de alimentación no incluye CO2 y H2 [4,14]. En general, cuando se incluyen dichas especies los órdenes aparentes y las energías de activación son bastante similares a pesar de las diferencias en la naturaleza del soporte, el contenido de Pt y la composición de la mezcla reactiva. La principal diferencia entre los estudios se da en el orden de reacción aparente para el agua que varía entre 0.44-1.10. Obviamente, dichas diferencias deben ser atribuidas a la naturaleza del soporte y/o a la presencia de Re. Con respecto al CO2, el orden de reacción aparente cercano a cero se explica por la débil interacción entre el CO2 con la superficie de Pt [30]. Por su parte, las diferencias observadas para el orden de reacción aparente para el CO se deben a la disminución de la energía de enlace del CO absorbido en Pt conforme aumenta el grado de cubrimiento [46–49], lo cual también explica el orden negativo para el H2: conforme aumenta el cubrimiento del CO, su energía de enlace disminuye favoreciendo la adsorción e inhibición de la reacción de WGS causada por las especies de H2 [50]. Phatak et al. [14] propuso un orden de reacción para el H2, (- 0.5 aprox.), afirmando que, una vez que la superficie está saturada, el H atómico dominará la superficie [14]. El incremento del grado de cubrimiento de la superficie por parte de las especies de H atómico, conlleva la inhibición de la reacción de WGS. Además, los diferentes órdenes aparentes de reacción calculados para el agua fueron asociados a diferentes mecanismos de reacción en función del soporte. Resumen 353 En el presente trabajo, no se pueden proponer distintos mecanismos de reacción ni distintas energías de activación puesto que el catalizador empleado es siempre el mismo. De acuerdo con los resultados observados, la velocidad de reacción del catalizador M0.3-B0.3 puede ajustarse de acuerdo a la ecuación R.1 cuando se ensaya en mezcla modelo. Sin embargo, en presencia de una corriente de reacción que simula la de salida de un reformador, la capa buffer altera el comportamiento catalítico del catalizador estructurado resultando en comportamientos catalíticos superiores. En este caso, los datos experimentales pueden ser ajustados con una ecuación cinética similar cuya única diferencia con la anterior es un orden de reacción aparente para el agua superior: −𝑟𝐶𝑂 = (1.6·108)×(𝑒−86000 𝑅𝑇 )×𝐶𝑂0.13 ×𝐻2𝑂0.70 ×𝐻2 −0.45 ×𝐶𝑂2 −0.12 × ( 1 − 𝐻2×𝐶𝑂2 𝐾𝑒𝑞×𝐶𝑂×𝐻2𝑂) (Ecuación R.2) El incremento del orden de reacción para el agua debe atribuirse a una mayor disponibilidad del soporte para activar el agua. Recientemente, Clay et al. [51] modelizó mediante DFT la cinética de un catalizador de Pt para WGS proponiendo un mecanismo con carboxilos como intermedios de reacción y afirmando que el paso limitante de velocidad era la etapa de disociación del agua. De acuerdo con su modelo, el paso limitante de velocidad es dependiente del grado de cubrimiento de la superficie por CO de tal forma que, para mayores presiones parciales de CO, el paso limitante de velocidad pasa a ser la formación del intermedio carboxilo. El cambio del paso limitante Resumen 354 de la velocidad de reacción hace referencia a la dependencia de los órdenes de reacción aparentes con respecto a las condiciones de reacción ya que estas determinarán el cubrimiento de la superficie en función de las presiones parciales superficiales. En consecuencia, un mayor cubrimiento de la superficie por especies de H2 resultará en una inhibición de la formación de los productos de reacción formados por reacción de CO + OH. Olimpiou et al. [3], que estudió catalizadores de Pt y Rh soportados sobre γ- Al2O3 empleando el método SSITKA, concluyó que los grupos hidroxilos lábiles así como las especies de H procedían de la molécula de agua y se encontraban principalmente sobre el soporte. Sin embargo, sólo una pequeña fracción de dichas especies son energéticamente capaces de difundir hasta el metal para dar H2 [52,53]. Además, otros estudios DFT sostienen que la disociación de agua esta favorecida en óxidos de cerio parcialmente reducidos, con o sin Pt [54]. Más recientemente, Anarifard et al. llevó a cabo un estudio DFT sobre Pt/Cerio postulando que la disociación del agua se da principalmente en las vacantes de oxígeno a través de la transferencia de átomos de H a átomos de oxígeno circundantes del soporte [55]. Además, el papel de las vacantes de oxígeno se relacionó con una mayor capacidad para estabilizar fragmentos originados en la disociación del agua. Así, la introducción de la capa buffer CeAl debe proporcionar un mayor número de sitios para la absorción de moléculas de agua. De hecho, la elevada Resumen 355 presión parcial del H2 deberá resultar en una superficie de cerio parcialmente reducida favoreciéndose así, la disociación de las moléculas de agua y la estabilización de las especies disociadas en la capa buffer [52–56]. Además, una mayor concentración de vacantes de oxígeno conlleva densidades electrónicas superiores en torno a las partículas metálicas [13,19,57–59]. Dicha densidad electrónica debería inhibir la adsorción del H2 debido a su carácter donador lo que conllevaría una mayor disponibilidad de sitios activos para la adsorción del CO a la vez que se disminuye su energía de adsorción en Pt. En la misma línea, Farrauto et al. [48] logró unas selectividades superiores para la reacción de PrOx a partir del a disminución de la fuerza de interacción H2-metal. Sin embargo, este efecto es independiente de la presencia de la capa buffer de manera que no puede explicar por sí mismo la mejora catalítica observada. Volviendo a la idea inicial, resulta evidente que la hipótesis basada en incrementar el número de moléculas de agua absorbidas de manera que puedan ser activadas conllevando mejores comportamientos catalíticos, ha sido alcanzada. En tales condiciones reductoras, incrementando la concentración de óxido de cerio mediante la incorporación de la capa buffer, se incrementa el número de vacantes de oxígeno de manera que se logran catalizadores estructurados con superiores comportamientos catalíticos. Sin embargo, el balance entre el carácter reductor y oxidante de la corriente de entrada controla la capacidad de mejora proporcionada por la capa buffer. Así, Resumen 356 si el carácter de la mezcla de alimentación es más oxidante, como en el caso de la mezcla modelo, las moléculas de agua pueden oxidar la capa buffer inhibiendo los beneficios proporcionados por la misma. No obstante, si el carácter reductor de la mezcla de reacción supera al oxidante, la capa buffer actúa como una reserva de grupos hidroxilos capaz de difundiros hasta los sitios activos metálicos a velocidades superiores que las de la reacción. 5. Rol de la capa buffer: importancia de la disociación de H2O en WGS Distintos sistemas estructurados fueron preparados estando todos ellos compuestos por dos capas diferenciadas: la capa buffer, justo encima del substrato metálico, y sobre esta, la capa de catalizador. En este estudio, la composición de la capa buffer fue modificada mientras que siempre se mantuvo la capa de catalizador constante. El papel de la capa buffer en la reacción de WGS fue directamente correlacionado con el agua. La composición de las capas buffers fue seleccionada en función de la capacidad para activar agua de los distintos óxidos. Así, la alúmina no es capaz de disociar agua mientras que un CeO2/Al2O3 sí que presenta dicha capacidad debido a las vacantes de O presentes en su red. Por último, un óxido de cerio dopado con Eu3+ debería ser presentar capacidades superiores para la activación de las moléculas de agua en base a una mayor concentración de vacantes provicadas tanto por la descompensación de carga como por la diferencia entre los tamaños de ambas especies. Resumen 357 Por último, se preparó un micromonolito constituido por una sola capa a partir de una mezcla física compuesta a su vez por el catalizador y por el buffer de cerio dopado con europio. La idea fue que si conseguíamos mejorar el contacto entre el buffer que activaba el agua y el sitio activo de Pt, se conseguirían mejores rendimientos catalíticos. 5.1. Preparación de los sistemas estructurados La preparación y adecuación de los micromonolitos se realizó de la misma forma que en el apartado 4.1. Por otra parte, se necesitaron 4 sólidos iniciales Al2O3, CeO2/Al2O3 y Ce0.9Eu0.1O2/Al2O3. Ambos soportes basados en cerio estaban constituidos por un 20 p/p% de óxido de cerio (dopado o sin dopar) soportado sobre una -Al2O3 de alta superficie. Estos se sintetizaron por impregnación húmeda y de forma análoga a la descrita en el apartado 3.1. En el caso del Eu, el precursor usado fue Eu(NO3)3·5H2O (Aldrich). Como catalizador se usó un Pt (2 p/p%)/CeO2/Al2O3 siendo este soporte comercial (Puralox) y también se sintetizó como se describe en el apartado 2.1. La nomenclatura empleada para los Al2O3, CeO2/Al2O3 y Ce0.9Eu0.1O2/Al2O3 fue Al, CeAl y CeEuAl mientras que el catalizador se denominó Pt/CeAl. La mezcla física se llamó PMix y fue preparada mezclando CeEuAl y Pt/CeAl en proporción 1:1. Todos los catalizadores se depositaron en los monolitos mediante el método washcoating y para la preparación de la suspensión se empleó la misma receta detallada en el apartado 4.1. En este estudio, se mantuvo constante la cantidad Resumen 358 de sólido depositado en todos los casos. Así, todos los catalizadores estructurados compuestos por dos capas contenían 0.3 g de buffer y sobre estos, 0.3 g de catalizador. Por su parte, el sistema estructurado sintentizado a partir de la mezcla física contenía 0.6g de PMix. Para denominar a las respectivas suspensiones, se empleó el prefijo “S_” seguido del nombre del correspondiente polvo inicial. Por ejemplo, a la suspensión formulada a partir del CeAl se la llamó S_CeAl. La nomenclatura empleada para los catalizadores estructurados hacía referencia al buffer incorporado; de esta forma al monolito que contenía una capa de buffer basada en CeAl se le denominó M_CeAl mientras que al que contenía la PMix se le denominó P_Mix. Con el objetivo de esclarecer, en la Figura R.15 se presenta el conjunto de catalizadores estructurados de una forma esquemática. Figura R.15. Esquemáticamente, conjunto de catalizadores estructurados preparados en este apartado Resumen 359 180 200 220 240 260 280 300 320 340 0 10 20 30 40 50 60 70 80 90 100 CO conversion (%) Temperature (oC) Pt/CeAl S_Pt/CeAl 160 180 200 220 240 260 280 300 320 340 360 0 10 20 30 40 50 60 70 CO conversion (%) Temperature (oC) M_Al M_CeAl M_CeEuAl 5.2. Resultados y discusión En la Figura R.16.a, se muestra la actividad catalítica del catalizador de Pt inicial comparado con su respectiva suspensión. Ambas muestras exhibieron similares comportamientos catalíticos indicando que no se habían producido cambios importantes en el catalizador durante el proceso de estructurado. Figura R.16. Actividad catalítica obtenida para a) Pt/CeAl y S_Pt/CeAl a 6 L/gh; b) M_Al, M_CeAl and M_CeEuAl a 80 L/gh empleando para ambos una alimentación compuesta por 7% CO + 9% CO2 + 50% H2 + 30% H2O Por otra parte, en la Figura R.16.b se observa la actividad catalítica de los micromonolitos M_Al, M_CeAl and M_CeEuAl. Las diferencias en actividad catalíticas entre las muestras en polvo y los catalizadores estructurados deben atribuirse a las diferentes condiciones catalíticas en las que fueron testados, 6 L/gh y 80 L/gh, respectivamente. Con el objetivo de determinar la contribución de la naturaleza de la capa buffer en función de la atmósfera de reacción, se analizó la variación de las Resumen 360 concentraciones relativas de cada componente en la corriente de alimentación (Figura R.17). En este caso, la concentración del CO2 fue disminuida del 9% al 3% con el objetivo de ampliar el rango de variación de los componentes. A partir de la composición inicial resultante, 7% CO + 3% CO2 + 50% H2 + 30% H2O, se variaron las presiones parciales de cada componente. La Figura R.17 muestra los resultados obtenidos para esta serie de experimentos. Teniendo en cuenta que la reacción de WGS es un equilibrio, el incremento de la concentración de los productos de la reacción a la entrada del reactor deberá conllevar una disminución de la conversión de CO. Especialmente para los buffer compuestos por óxido de cerio, la variación del % CO2 no conllevó cambios importantes en las conversiones de CO observadas. Contrariamente, la incorporación de H2 a la mezcla reactiva provocaba importantes disminuciones en las actividades catalíticas. Esto debe atribuirse tanto al desplazamiento del equilibrio termodinámico como a la competición entre las moléculas de CO e H2 para ser adsorbidos en los sitios activos metálicos. Por su parte, el aumento del % CO también disminuía las conversiones de CO en todos los casos probablemente debido al mayor número de moléculas de CO que deben ser convertidas por el mismo número de centros activos de Pt. En general, se puede afirmar que todos los cambios observados en las variaciones de % CO, % CO2 e % H2 fueron similares en todos los sistemas estructurados analizados. Resumen 361 45678910 11 12 10 20 30 40 50 CO conversion (%) CO (%) M_Al M_CeAl M_CeEuAl 0 2 4 6 8 10 12 10 20 30 40 50 CO conversion (% ) CO2 (%) M_Al M_CeAl M_CeEuAl 010 20 30 40 50 60 0 10 20 30 40 50 60 70 80 90 100 CO conversion (%) H2 (%) M_Al M_CeAl M_CeEuAl 15 20 25 30 35 40 15 20 25 30 35 40 45 CO conversion (%) H2O (%) M_Al M_CeAl M_CeEuAl Figura R.17. Efecto sobre la actividad catalítica de los sistemas M_Al, M_CeAl and M_CeEuAl cuando se variaban las presiones parciales de: a) % CO; b) % CO2; c) % H2 y d) % H2O Sin embargo, la variación de la concentración de H2O sí que conllevaba diferentes cambios en el comportamiento de los estructurados en función de la naturaleza de la capa buffer (Figura R.17.d). A bajas concentraciones de H2O todos los micromonolitos alcanzaban las mismas conversiones de CO. El incremento de la presión parcial conllevaba mejores conversiones de CO para todos los sistemas lo cual, teniendo en cuenta que la reacción de WGS es un equilibrio, parece lógico. A pesar de esto, el grado de mejora observado en Resumen 368 Gráficamente, puede discernirse como la muestra S_Al perdía el agua a menores temperaturas si la comparamos con las muestras compuestas por cerio. Así, las diferencias observadas en este rango de temperatura dben ser exclusivamente al óxido de cerio. Además, el S_CeEuAl exhibe una mayor señal de agua a lo largo de todo el rango de desorción incluyendo a las temperaturas máximas en las cuales la reacción de WGS ha sido llevada a cabo. Esto implica que esta muestra parece ser capaz de mantener absorbidas mayores cantidades de agua. Peden et al. [65] estudió las interacciones entre el agua y superficie oxidadas y reducidas de óxido de cerio concluyendo que las moléculas de agua se asociaban a sitios reducidos Ce3+. Efectivamente, la muestra dopada con Eu presentaba mayores capacidades para disociar agua y para retenerla molecularmente absorbida lo cual se debe relacionar directamente con una mayor concentración de vacantes de oxígeno. Dicha conclusión está de acuerdo con los resultados observados en los experimentos de FT-IR y debe ser relacionada directamente con la importancia del agua en la reacción de WGS así como con la hipótesis inicial. Por lo tanto, se ha demostrado que la incorporación una capa buffer conlleva mejores comportamientos catalíticos debido a una modulación de la capacidad para disociar agua en función de la concentración de vacantes de oxígeno. Así mismo, dicha disociación parece estar relacionada con una capacidad superior de mantener absorbidas las especies de agua probablemente debido a unas Resumen 369 mayores densidades electrónicas generadas por mayores concentraciones de vacantes de oxígeno. En conclusión, el estructurado de catalizadores compuestos por dos capas diferenciadas aparece como una propuesta interesante ya que consigue mejores comportamientos catalíticos, debido a procesos de disociación de agua favorecidos pero además, inhibe los efectos de desactivación provocados especies carbonáceas en este tipo de catalizadores. 6. Rol del metal noble en la reacción de WGS: Pt vs Au El último capítulo pretende arrojar luz en cuanto al papel que juegan los metales nobles en la reacción de WGS a través de la comparación de mismos. Para ello este capítulo se subdivide en tres partes, en la primera una serie catalizadores fueron preparados, caracterizados y ensayados catalíticamente. En la segunda, se estructuraron catalizadores equivalentes de Au y Pt evaluándose el efecto del estructurado para cada uno de ellos así como la capacidad de realizar la reacción de WGS asistida por O2. Por último, se llevó un estudio EXAFS-XANES que provee resultados tanto atendiendo a su estructura como a su dinámica. Parte I: Catalizadores de Au y Pt en polvo sobre diferentes soportes para la reacción de WGS Para este estudio, se soportó un 2 p/p% de Au y Pt sobre soportes basados en distintos óxidos siendo estos -Al2O3, CeO2/Al2O3 and Ce0.8Fe0.2O2/Al2O3. La Resumen 370 nomenclatura seleccionada para los soportes fue Al2O3, CeAl and CeFeAl. Para los catalizadores, se empleó una terminología análoga pero añadiéndose un prefijo que indicaba el metal. Por ejemplo, los catalizadores de Au y Pt preparados a partir del soporte CeAl, fueron denominados Au/CeAl y Pt/CeAl. Todos los catalizadores preparados presentaban propiedades texturales y composiciones químicas cercanas a las nominales. Por su parte, la dispersión de los catalizadores de determino mediante HRTEM. A pesar de que inicialmente ambos metales presentaban tamaños de partícula distintos, 4 nm y 2 nm para el Au y el Pt respectivamente, durante la reacción el Pt parecía sinterizar alcanzando tamaños de en torno a 4 nm según los resultados obtenidos por quimisorción de CO. Esto permitía analizar las atribuir las diferencias exhibidas entre los catalizadores, a la naturaleza del metal y/o su interacción con el soporte pero no a efectos estructurales relacionados con el tamaño de partícula. En la Figura R.23, se muestra la actividad catalítica obtenida por estos catalizadores cuando fueron testados en condiciones modelo así como en condiciones reales. Los test de actividad catalítica muestran notables diferencias en función del metal noble y de la naturaleza del soporte. Así, la importancia del soporte parecía ser mucho mayor para el Au que para el Pt. De hecho, mientras que el Au/Al era prácticamente inactivo, el Pt/Al conseguía incluso alcanzar las conversiones de equilibrio lo cual se debe relacionar con su capacidad intrínseca para disociar el agua. A pesar de esto, Resumen 371 la presencia de un óxido reducible, como el cerio, que lleve a cabo dicha disociación resulta en mejoras significativas de su comportamiento catalítico. Figura R.23. Actividad catalítica exhibida por los catalizadores cuando se testan en a) Condiciones modelo y b) Condiciones reales Por otra parte, el uso del CeFeAl conlleva significativas mejoras catalíticas en el caso del Au doblando casi las conversiones de CO en todo el rango de temperaturas. A pesar de que las diferencias entre el Pt/CeAl y el Pt/CeFeAl fueron bastante menos acusadas (observándose sólo leventemente en las temperaturas más bajas), los catalizadores Pt/CeFeAl y Au/CeFeAl fueron seleccionados y testados en condiciones reales (Figura R.23.b). En tales condiciones quedó la superioridad del Pt para trabajar en presencia de H2 y CO2 claramente demostrada. Con el objetivo de dilucidar a qué se debían tales diferencias se caracterizaron las propiedades electrónicas de los catalizadores a partir de espectroscopía UV-Vis. La Tabla R.1 muestra lo valores obtenidos para la energía de band del óxido de cerio para todas las muestras. Con respecto a los soportes, la 180 200 220 240 260 280 300 320 340 0 10 20 30 40 50 60 70 80 90 100 Pt/Al Au/Al Au/CeAl Pt/CeAl Au/CeFeAl Pt/CeFeAl % CO Conversion Temperature (oC) A 180 200 220 240 260 280 300 320 340 0 10 20 30 40 50 60 70 80 90 100 % CO Conversion Temperature (0C) Au/CeFeAl Pt/CeFeAl Equilibrium B Resumen 372 disminución de la energía de band gap presentada por el soporte CeFeAl con respecto al CeAl se explica por la incorporación de niveles electrónicos aportados por el Fe entre la banda de conducción y la banda de valencia del óxido de cerio. Por su parte, los catalizadores de Au presentan en todos los casos valores de band gap menores que los exhibidos por los catalizadores de Pt. Esto puede explicarse en base a las estructura electrónica de ambos sistemas. La presencia de metal induced gap states (MIGS) es el resultado de la interacción del metal noble con la superficie del óxido de cerio. La interacción iónica entre los metales nobles cargados negativamente y los clusters de vacantes superficiales resulta en la sobreposición de los niveles de 5d del Pt con los 2p del O, mientras que los niveles 5d de Au caen entre los niveles del 2p del O y los 4f del Ce. Independientemente de la configuración de los estados energéticos de ambos sistemas, los electro extra procedentes de los metales nobles resultan en un transferencia electrónica de los conjuntos Ce(Fe)- Tabla R.1. Energías de band gap calculadas Sample Direct band gap (eV) Indirect band gap (eV) CeAl 3.05 2.80 CeFeAl 2.93 2.45 Pt/CeAl 2.94 2.12 Au/CeAl 2.97 2.90 Pt/CeFeAl 2.79 2.26 Au/CeFeAl 2.11 1.50 Resumen 373 O(VO)-Ce(Fe) hacia el metal noble siendo mayor la carga negativa en el Pt que en el Au. Por lo tanto, el superior comportamiento catalítico observado en el caso del Pt/CeFeAl se relaciona directamente con las diferencias que ambos catalizadores presentaban en sus estructuras electrónicas viniendo esto relacionado en las distintas energías de band gap calculadas. Parte II: Au/CeFeAl y Pt/CeFeAl estructurados en micromonolitos metálicos Esta parte se focaliza en la aplicación de catalizadores de catalizadores estructurados de Au y Pt para la reacción de WGS con el objetivo superar las restricciones de volumen asociadas a los reactores convencionales de lecho catalítico para la reacción de WGS aprovechando las excelentes transferencias de calor y de masa porporcionadas por los micromonolitos metálicos. El efecto del proceso de estructurado se evalúa también a lo largo de esta parte. Por otra parte, se analizó el efecto de la adición de pequeñas cantidades de oxígeno que permitan paliar los efectos de desactivación por especies carbonáceas. Los resultados catalíticos se relacionaron directamente con la naturaleza del metal noble y con las condiciones de reacción. Con tal objetivo, los catalizadores, Pt/CeFeAl y Au/CeFeAl se prepararon y estructuraron mediante el método washcoating. Las suspensiones empleadas, secadas y calcinadas, fueron caracterizadas y testadas en WGS. La nomenclatura empleada para las suspensiones fue S_Pt/CeFeAl y Resumen 374 180 200 220 240 260 280 300 320 340 0 10 20 30 40 50 60 70 80 90 100 S_Au/CeFeAl Au/CeFeAl S_Pt/CeFeAl CO Conversion (%) Temperature (oC) Pt/CeFeAl S_Au/CeFeAl. Por su parte, los catalizadores estructurados se denominaron M_Pt/CeFeAl y M_Au/CeFeAl. En la Figura R.24, se presenta la actividad catalítica de los catalizadores iniciales y de las suspensiones correspondientes en mezcla modelo. Como puede observarse, el proceso de estructurado conlleva serias disminuciones en cuanto a la actividad catalítica en el caso del catalizador basado en Au. Esto puede atribuirse a cambios en el tamaño de las partículas de Au. Por su parte, el Pt parecía soportar bastante bien el método washcoating y prácticamente no se aprecian cambios significativos entre la actividad catalítica del catalizador inicial y su respectiva suspensión. Figura R.24. Conversiones de CO obtenidas para las muestras en polvo en mezcla modelo y a 4000 h-1 Por otra parte, también se ensayaron los catalizadores estructurados en condiciones modelo (Figura R.25). De acuerdo a lo esperado, el M_Pt/CeFeAl mostraba superiores comportamientos catalíticos en comparación con el Resumen 375 M_Au/CeFeAl. De hecho, el M_Au/CeFeAl presentaba pobres conversiones a lo largo de todo el rango de reacción. Figura R.25. CO conversions in obtained for the micromonolithic structures in: a) model and O2-model; b) TOX La incorporación a la corriente de reacción de pequeñas cantidades de oxígeno (O2-WGS) suponía un importante aumento en cuanto a las conversiones obtenidas por ambos catalizadores. No obstante, el incremento de actividad catalítica mostrado por el M_Au/CeFeAl era, además de sorprendente, muy superior al exhibido por el M_Pt/CeFeAl para el cual la mejora disminuye conforme aumenta la temperatura de reacción. Teniendo en cuenta que la reacción de oxidación de CO se puede dar de forma paralela y que, de hecho, los catalizadores de metales nobles son muy activo para este tipo de reacciones [66], se realizó un ensayo análogo pero en ausencia de agua sobre ambos catalizadores estructurados (Figura R.25.b). Ambos catalizadores presentaban conversiones análogas para la reacción de 180 200 220 240 260 280 300 320 340 0 10 20 30 40 50 60 70 80 90 100 CO Conversion (%) Temperature (oC) M_Au/CeFeAl M_Pt/CeFeAl 180 200 220 240 260 280 300 320 340 0 10 20 30 40 50 60 70 80 90 100 Temperature (oC) M_Pt/CeFeAl O2- WGS M_Pt/CeFeAl WGS M_Au/CeFeAl O2-WGS CO Conversion (%) M_Au/CeFeAl WGS Resumen 376 oxidación de CO alcanzando el máximo de conversión permitido por la estequiometría de la alimentación durante todo el rango de temperaturas siendo este un 30 %. El M_Pt/CeFeAl presenta un incremento máximo de conversión a 180 ºC que se corresponde exactamente con el posible incremento que se obtendría en el caso de que sí se dieran ambas reacciones de forma paralela. Similares resultados fueron obtenidos por Duprez et al. [67] quien afirmaba que el incremento de la actividad catalítica observado para catalizadores de Pt en reacciones de WGS asistidas por oxígeno se podía atribuir únicamente a dichas reacciones en paralelo. Contrariamente, en el caso del catalizador M_Au/CeFeAl, el incremento de conversión observado para la reacción WGS asistida por oxígeno no se puede atribuir al hecho de que se den, de forma paralela, la reacción de WGS y reacción de oxidación de CO. Por lo tanto, el Au como metal noble sí que parece ser capaz de dar la reacción de WGS asistida por oxígeno. En literatura ha sido descrito que el efecto beneficioso de la incorporación de O2 favorece la generación de grupos OH lo que favorecería la reacción de WGS [68,69]. Por otra parte, en el caso de los catalizadores de Au, la importancia vital del soporte ya ha sido demostrada. Así pues, el efecto beneficioso de la incorporación de oxígeno pudiera estar también relacionada con la inhibición de los efectos de desactivación del soporte tales como la sobrerreducción del soporte y/o el bloqueo de los centros activos para la disociación del agua del Resumen 377 soporte. De hecho, también ha sido reportado que la incorporación de oxígeno facilita la descomposición de las especies carbonáceas liberando los sitios activos que volverían a estar disponibles para la disociación de agua lo cual explicaría el incremento de conversión observado en la reacción de WGS [70]. En conclusión, la adición de O2 a la reacción de WGS invierte la balanza en favor de los catalizadores de Au siendo este el único de los dos metales capaz de dar una reacción de WGS asistida por oxígeno. A pesar de que la dilucidación del efecto de la adición de pequeñas cantidades de oxígeno a la reacción de WGS, esto se presenta como una interesante estrategia que potencia la reacción de WGS siendo especialmente atractiva cuando se combina con los micromonolitos metálicos debido a las importantes reducciones de volumen que permiten de manera que tales dispositivos puedan ser implementados en procesos de combustible basados en hidrógeno. Parte III: Estudios Operando XANES-EXAFS sobre catalizadores basados en Au y Pt En condiciones de reacción los catalizadores pueden llevar cabo cambios químicos que transformen, de una forma significativa su estructura con respecto a la que inicialmente presentaban [71]. Las técnicas de caracterización in In situ/operando aparecen como interesantes ya que Resumen 384 [14] A.A. Phatak, N. Koryabkina, S. Rai, J.L. Ratts, W. Ruettinger, R.J. Farrauto, G.E. Blau, W.N. Delgass, F.H. Ribeiro, Catal. Today 123 (2007) 224. [15] S.J. Tauster, S.C. Fung, R.L. Garten, J. Am. Chem. Soc. 100 (1978) 170. [16] M. Gonzalez Castaño, T.R. Reina, S. Ivanova, M.A. Centeno, J.A. Odriozola, J. Catal. 314 (2014) 1. [17] J. Kugai, J.T. Miller, N. Guo, C. Song, J. Catal. 277 (2011) 46. [18] P.C.A. Brito, D.A.A. Santos, J.G.S. Duque, M.A. Macêdo, Phys. B Condens. Matter 405 (2010) 1821. [19] C.T. Campbell, C.H.F. Peden, Science 309 (2005) 713. [20] N. Acerbi, S. Golunski, S.C. Tsang, H. Daly, C. Hardacre, R. Smith, P. Collier, J. Phys. Chem. C 116 (2012) 13569. [21] G. Avgouropoulos, T. Ioannides, J. Mol. Catal. A Chem. 296 (2008) 47. [22] A.S. Reddy, C.Y. Chen, C.C. Chen, S.H. Chien, C.J. Lin, K.H. Lin, C.L. Chen, S.C. Chang, J. Mol. Catal. A Chem. 318 (2010) 60. [23] T. Tabakova, L. Ilieva, I. Ivanov, R. Zanella, J.W. Sobczak, W. Lisowski, Z. Kaszkur, D. Andreeva, Appl. Catal. B Environ. 136-137 (2013) 70. [24] T. Tabakova, V. Idakiev, D. Andreeva, I. Mitov, Appl. Catal. A Gen. 202 (2000) 91. [25] T.R. Reina, S. Ivanova, J.J. Delgado, I. Ivanov, V. Idakiev, T. Tabakova, M.A. Centeno, J.A. Odriozola, ChemCatChem 6 (2014) 1401. [26] O.H. Laguna, M.A. Centeno, M. Boutonnet, J.A. Odriozola, Appl. Catal. B Environ. 106 (2011) 621. Resumen 385 [27] S.Y. Choung, M. Ferrandon, T. Krause, 99 (2005) 257. [28] W. Ruettinger, O. Ilinich, R.J. Farrauto, J. Power Sources 118 (2003) 61. [29] X. Wang, R.J. Gorte, J.P. Wagner, J. Catal. 212 (2002) 225. [30] Y. Nagai, K. Dohmae, Y. Ikeda, N. Takagi, T. Tanabe, N. Hara, G. Guilera, S. Pascarelli, M.A. Newton, O. Kuno, H. Jiang, (2008) 9303. [31] P. Panagiotopoulou, J. Papavasiliou, G. Avgouropoulos, T. Ioannides, D.I. Kondarides, Chem. Eng. J. 134 (2007) 16. [32] D. Pierre, W. Deng, M. Flytzani-Stephanopoulos, Top. Catal. 46 (2007) 363. [33] J.M. Zalc, V. Sokolovskii, D.G. Löffler, J. Catal. 206 (2002) 169. [34] R.J. Farrauto, Y. Liu, W. Ruettinger, O. Ilinich, L. Shore, T. Giroux, Catal. Rev. 49 (2007) 141. [35] P. Avila, M. Montes, E.E. Miró, Chem. Eng. J. 109 (2005) 11. [36] S. Ivanova, O.H. Laguna, M.Á. Centeno, A. Eleta, M. Montes, J.A. Odriozola, Renew. Hydrog. Technol. (2013) 225. [37] V.G. Milt, S. Ivanova, O. Sanz, M.I. Domínguez, A. Corrales, J.A. Odriozola, M.A. Centeno, Appl. Surf. Sci. 270 (2013) 169. [38] O. Sanz, L.M. Martínez T, F.J. Echave, M.I. Domínguez, M.A. Centeno, J.A. Odriozola, M. Montes, Chem. Eng. J. 151 (2009) 324. [39] V. Meille, Appl. Catal. A Gen. 315 (2006) 1. [40] Q. Fu, H. Saltsburg, M. Flytzani-Stephanopoulos, Science 301 (2003) 935. Resumen 386 [41] P.B. Weisz, J.S. Hicks, (1961). [42] O.H. Laguna, M. González Castaño, M.A. Centeno, J.A. Odriozola, Chem. Eng. J. 275 (2015) 45. [43] D.I. Potemkin, P.V. Snytnikov, V.D. Belyaev, V. a. Sobyanin, Chem. Eng. J. 176-177 (2011) 165. [44] R.J. Farrauto, Y. Liu, W. Ruettinger, O. Ilinich, L. Shore, T. Giroux, Catal. Rev. 49 (2007) 141. [45] G. Germani, P. Alphonse, M. Courty, Y. Schuurman, C. Mirodatos, Catal. Today 110 (2005) 114. [46] X. Zheng, J. Mantzaras, R. Bombach, Combust. Flame 161 (2014) 332. [47] S. Aranifard, S.C. Ammal, A. Heyden, J. Phys. Chem. C 116 (2012) 9029. [48] X. Liu, O. Korotkikh, R. Farrauto, Appl. Catal. A Gen. 226 (2002) 293. [49] L. Grabow, Y. Xu, M. Mavrikakis, Phys. Chem. Chem. Phys. 8 (2006) 3369. [50] P. Thormählen, M. Skoglundh, E. Fridell, B. Andersson, J. Catal. 188 (1999) 300. [51] J.P. Clay, J.P. Greeley, F.H. Ribeiro, W.N. Delgass, W.F. Schneider, J. Catal. 320 (2014) 106. [52] D. Martin, D. Duprez, J. Phys. Chem. B 112 (2006) 4428. [53] D. Duprez, Catal. Today 112 (2006) 17. [54] Y. Lykhach, V. Johánek, H. a. Aleksandrov, S.M. Kozlov, M. Happel, T. Skála, P. St. Petkov, N. Tsud, G.N. Vayssilov, K.C. Prince, K.M. Resumen 387 Neyman, V. Matolín, J. Libuda, J. Phys. Chem. C 116 (2012) 12103. [55] S. Aranifard, S.C. Ammal, A. Heyden, J. Phys. Chem. C 118 (2014) 6314. [56] G. Jacobs, P.M. Patterson, L. Williams, E. Chenu, D. Sparks, G. Thomas, B.H. Davis, Appl. Catal. A Gen. 262 (2004) 177. [57] S.M. Opalka, T.H. Vanderspurt, R. Radhakrishnan, Y. She, R.R. Willigan, J. Phys. Condens. Matter 20 (2008) 064237. [58] F.W. Poulsen, M. Glerup, P. Holtappels, Solid State Ionics 135 (2000) 595. [59] Z. V. Popović, Z.D. Dohčević-Mitrović, N. Paunović, M. Radović, Phys. Rev. B - Condens. Matter Mater. Phys. 85 (2012) 1. [60] W. Ruettinger, X. Liu, R.J. Farrauto, Appl. Catal. B Environ. 65 (2006) 135. [61] C. Li, Y. Sakata, T. Arai, K. Domen, K. Maruya, T. Onishi, J. Chem. Soc. Faraday Trans. 1 85 (1989) 1451. [62] M.I. Zaki, M.A. Hasan, F.A. Al-Sagheer, L. Pasupulety, Colloids Surfaces A Physicochem. Eng. Asp. 190 (2001) 261. [63] J.H. Kwak, R.J. Chimentao, C.H.F. Peden, J. Phys. Chem. C 111 (2007) 2661. [64] O. Thinon, K. Rachedi, F. Diehl, P. Avenier, Y. Schuurman, Top. Catal. 52 (2009) 1940. [65] M.A. Henderson, C.L. Perkins, M.H. Engelhard, S. Thevuthasan, C.H.F. Peden, Surf. Sci. 526 (2003) 1. [66] O.H. Laguna, M.A. Centeno, G. Arzamendi, L.M. Gandía, F. Romero-Sarria, J.A. Odriozola, Catal. Today 157 (2010) 155. Resumen 388 [67] J. Barbier, D. Duprez, Appl. Catal. B Environ. 3 (1993) 61. [68] T. Utaka, K. Sekizawa, K. Eguchi, Appl. Catal. A Gen. 194 (2000) 21. [69] G.C. Chinchen, M.S. Spencer, K.C. Waugh, D.A. Whan, J. Chem. Soc. Faraday Trans. 1 83 (1987) 2193. [70] T.R. Reina, S. Ivanova, V. Idakiev, J.J. Delgado, I. Ivanov, T. Tabakova, M.A. Centeno, J.A. Odriozola, Catal. Sci. Technol. 3 (2013) 779. [71] T. Bunluesin, R.J. Gorte, G.W. Graham, Appl. Catal. B Environ. 15 (1998) 107.