Molybdenum oxides as catalyst for biomass-derived compounds: a theoretical approach

• Autores: Marcos Rellan Piñeiro
• Directores de la Tesis: Nuria Lopez Alonso (dir. tes.)
• Lectura: En la Universitat Rovira i Virgili ( España ) en 2018
• Idioma: español
• Tribunal Calificador de la Tesis: Jan Rossmeisl (presid.), Cecilia Fondelli (secret.), José Javier Plata Ramos (voc.)
• Programa de doctorado: Programa de Doctorado en Ciencia y Tecnología Química por la Universidad Rovira i Virgili
• Materias:
  • Química
• Texto completo no disponible (Saber más ...)
• Resumen

  • Fossil resources are fundamental in our society. They provide most of the energy supply in the world. Moreover, around 90% of the organic chemicals are produce from petroleum and natural gas, which means that the most of the things that we use every day came from fossil feedstock. However, this dependence has an important cost, the global warming, whose effects are obvious nowadays. In addition to the environmental concern, the depletion of the fossil resources and the consequent rising in the crude barrel price force the chemical industry to search a new energy and chemical feedstock. In the last decades, biomass has emerged as a green and renewable alternative resource of chemical compounds. It is generated in photosynthesis from CO2 and H2O using sunlight as energy. It is estimated that around 170 billion metric tons of biomass are generated per year in nature, and only 3-4% of these compounds are used by humans. Therefore, biomass has a high potential grow as raw material for chemical industry with the added advantage that it is distributed over the entire world. The use of biomass as transportation fuel is a very attracting industrial challenge, however, its low carbon content (that means low energy density) limits its use for this purpose. Instead, biomass has a high potential as renewable feedstock for the production of valuable chemicals. It is formed by highly functionalized molecules with a wide chemical diversity, which coverts biomass in a excellent raw material for high-value compounds. A major challenge of the biobased industry is the development of catalyst active, selective, and stables for transformations of biomass-derived compounds in valuable chemicals There are many research works focused to upgrade the biomass-derived molecules, and catalysts as metals and metal oxides were reported to perform chemical transformations as, isomerization, hydrogenation, dehydrogenation, dehydroxylation/hydrogenolysis, deoxygenation, and oxidation reactions. These processes combined acid/base-catalyzed reactions such as hydrolysis and dehydration reactions to achieve a multistep conversion in one-pot process.

    The research will be performed in the framework of Density Functional Theory.The increase of computational power and the development of Density Functional Theory (DFT) allow calculating the electronic structure of the heterogeneous catalytic systems. Thus, DFT has become in one of the main tools in heterogeneous catalysis research since the study at atomic scale plays an important role in characterization, understanding and prediction of the catalytic processes. For instance, the description of the chemical bond between reactants and surface is key to understand and predict the catalyst reactivity. The catalysts often present a complex relationship between surface structure and catalytic activity and selectivity that cannot be evident by the studies with experimental techniques; but nevertheless, theoretical simulations can give insights about them. Thus nowadays, it is accepted that the combination of experiments and simulations is the most efficient way to catalysis development. Simulations on MoO3 present some challenges. It is a reducible transition metal oxide, and pure DFT cannot reproduce correctly the electron self-interaction of its electrons at 4d orbital. The DFT+U method, which add a repulsion term to DFT, has emerged as a possible solution, but its it needs to be enhanced. Despite the advantages of DFT+U, the fitting of the U parameter is arbitrary and there is no an accepted value in literature, which triggers that studies with different U, or pure DFT method have been published giving scattered results. Moreover, due its layered structure simple DFT methods cannot reproduce the interaction between layers overestimating the interlayer gap, between other problems. A more accurate description can be obtained with the addition of van der Waals interaction in DFT calculations through a semiempirical correction as the Grimme’s D2 and D3 methods.

    The general objective of this thesis is to contribute to the development of a most sustainable and environmental friendly chemical industry through the theoretical study of selective biomass conversion catalyzed by metal oxides. In particular, this thesis is focused on the reactivity of hydrocarbons on molybdenum oxides. To date, only few theoretical studies were focused to the study of these catalytic systems.

    In Chapter 3 the selective oxidation of methanol to formaldehyde is studied. Formaldehyde is an important precursor to many other materials and chemical compounds, it is predicted its production will exceed 52 million tonnes in 2017. The most efficient industrial route to obtain formaldehyde is the oxidation of methanol, CH3OH(g) + ½ O2 →CH2O + H2O, through the Formox process, where the catalyst is a mixture of MoO3 and the iron molybdate Fe2(MoO3)4. To obtain a high selectivity an excess of MoO3 should be added to the reactor. The two phases are essential; some experimental studies have determined that the Fe content promote the catalyst activity and the Mo-phase promote the selectivity, but their exact role and the synergy between them are under discussion. In addition, the nature of the active site and its oxidation state also are unknown. The objective of the present work is to elucidate these issues. There are many experimental works focused to solve these issues, since remove the drawbacks of the process, as the catalyst stability, is an important industrial challenge.

    In Chapter 4 the nature of the surface vacancy defects is studies through the DFT+U optimized method. In the last years, molybdenum oxides, mainly MoO3, have become the catalyst of choice for a wide range of applications. In addition to its catalytic properties, a wide range of chemical and physical applications have been reported for them. The most of these properties are due to their redox properties that can be fine-tuned by addition of cations, hydrogen or by oxygen removal. Specifically, in catalyst, most of the reactions are carried out in the reduced oxide, being vacancies the reactive center. Therefore, a characterization of the structural and electronic properties of the vacancies is essential to predict and to control the properties of the material. This study is focuses to the vacancy formation on the α-MoO3 (010) surface. This surface has three different types of oxygens bonded to one, two and three molybdenum centers. It is important to know which type of oxygen is removed during reduction and the electronic structure of the reduced surface, as they affect the catalytic properties of the oxide. The different Ueff results in a wide range of reaction energies and electron localization and so, the results are not comparable. Therefore, a more thorough investigation of the dependence of the chemical properties with Ueff is thus necessary.

    In Chapter 5 the selective reduction of glycerol to propylene is studied. The depletion of oil resources and the increase of the environmental concerns force the chemical industry to search new renewable resources and green processes to obtain key chemical intermediates. Thus, biomass has emerged as an alternative renewable feedstock 126 and initiatives to convert non-edible fractions and waste into chemical platforms are being explored. However, biomass is a complex material and includes a wide range oxygenated as alcohols, aldehydes, ketones, carboxylic acids and phenols. Therefore, the preferential C-O bond cleavage over the other bonds C-C and C-H is a must to the upgrading of these molecules. In this direction one of the most viable processes to reduce the oxygen content is hydrodeoxygenation, HDO.

    It achieves the reduction of the oxygen content in biomass derived molecules by its reaction with H2 at high temperatures, thus producing water. Molybdenum oxides showed good selectivity in HDO of alcohols, linear ketones and cyclic ethers to produce olefins and cyclic ketones and phenols to form aromatic hydrocarbons. In these reactions, MoO3 presents a preferential C-O bond cleavage pattern leading the formation of hydrocarbons and olefins while keeping of the carbon chain. Among the biomass derivatives glycerol plays an important role since it is a by-product in the transesterification of vegetal oils for biodiesel production. Around 10 tonnes of glycerol are produced per 100 tonnes of biodiesel. Due the increase in biodiesel production, glycerol is an abundant and cheap compound, and finding new efficient and environmental friendly routes to upgrade it is a challenging industrial target.

    Therefore, it was included as one of the top twelve building blocks by DOE. The deoxygenation of glycerol can be a source for propanediols and propanols and more interestingly propylene. Propylene is the precursor to polypropylene, acrylonitrile, propylene oxide, oxo alcohols, cumene, isopropyl alcohol, and others. The current propylene production is around 90 million tons and it is forecasted to grow up to 130 million tonnes worldwide by 2023. Nowadays, it is mainly obtained as by-product of naphtha steam cracking, but the emergence of shale-gas feedstocks will likely produce a shortage in the next years in particular geographical areas, mainly in US. Therefore, an on-purpose technology for propylene production is an attractive industrial target and its production from glycerol remains a promising route.

    In Chapter 6 the selective epimerization of glucose to mannose catalyzed by the polyoxometalate H3PMo12O40 was studied. Epimers are a steroisomers pair with several quiral center that differ in the configuration at only one. Typical molecules with epimer pair are monosaccharides, more commonly called sugars, a renewable biomass source. Some of them, as mannose, lyxose and ribose, appear in small quantities in nature and, thus, are known as rare sugars.

    Epimerization can be used to obtain these rare sugars from their more abundant epimers; glucose, xylose and arabinose. Rare sugars have a high potential in the chemical industry, as they can be employed as chemical platforms to synthesize complex compounds with the Inorganic catalyst as bases, lewis acids, and zeolites have been reported as active catalyst for epimerization reactions, but they offer a poor selectivity. The best performance for activity and selectivity is obtained with Mo-based compounds. In the 1970s, Bilik and coworkes reported molybdates as highly selective catalyst for the epimerization reaction at C2 position. This reaction was experimentally studied and a mechanism, where the epimerization is produced through a 1,2 C-shift, was proposed. The apparent activation energy was reported to 126 kJ/mol. Since then, epimerization reactions through 1,2 C-shift are known as Bilik reactions. More recently, the Mo-based polyoxometalate (POM), the Keggin cluster H 3 PMo 12 O 40 , 12 and niobium molybdates have also been reported as active catalysis, showing a better performance than molybdates.

    Conclusions: The biomass-derived compounds conversion catalyzed by Mo-oxides was studied by means of density functional theory, DFT. Three important reactions were studied: i) the oxidation of methanol to formaldehyde (Chapter 3), ii) the glycerol reduction to propylene (Chapter 5), and iii) the glucose epimerization to mannose (Chapter 6). In addition, the optimization of DFT+U method and the study of the nature of vacancy defects were performed (Chapter 4). Through these studies several unraveled questions were solved. Therefore, the studies reported in this thesis pave the way for a better understanding of Mo-oxides catalytic behavior that will allow a development of more active, selective, and stable catalysts. The detailed conclusions of each work are detailed in the following: Chapter 3. The FormoxTM process: selective methanol oxidation to formaldehyde • The reaction pathway for the methanol oxidation to formaldehyde is the following: CH3OH → CH3OH* → CH3O* → CH2O* → CH2O • The catalytic activity is determined by the vacancy formation, the highest point in reaction profiles, and catalyst recovery.

    • Both surfaces α-MoO3(010) and r-MoO2(110) are actives for the oxidation of methanol, but the only selective surface is α-MoO3(010). The presence of reduced Mo centers on this surface is detrimental to selectivity. • The presence of iron as dopant reduces the vacancy formation energy, which increases the activity. This reduction is due in vacancy formation one electron is accommodated in the iron. • The presence of iron in the surface layer forms vacancies spontaneously, reducing the selectivity. However, the presence of iron in subsurface layer does not form spontaneous vancancy defects and keeps the selectivity.

    • In the industrial catalyst MoO3 is the catalytic phase and the role of Fe is to increase the activity reducing the reaction energy of redox steps. • The best performance can be obtained with the selective iron doping in subsurface positions. Surface layer of MoO3 selectively catalyzes the reaction whereas the subsurface iron increases the activity without to affect the selectivity.

    Chapter 4. On the nature of oxygen vacancies on the α-MoO3 (010) surface through the optimized DFT+U method • The Ueff of DFT+U method was optimized to a value of 3.5 eV by fitting with the HSE06 hybrid functional and CCSD(T) reaction energies and electron localization. The previously most used value, Ueff =6.3 eV, strongly underestimates the reaction energies.

    • During α-MoO 3 (010) surface reduction vacancies will be produced only in terminal positions.

    • The reduced surface can be in two different local electronic configurations: i) two Mo (V) polarons in the undercoordinated Mo and in a neighboring Mo center along x axis; and ii) a single Mo(IV) dipolaron in the undercoordinated Mo.

    • The two electronic configurations were characterized. Bond distances, magnetization, Bader charges, Raman frequencies and XPS shifts were reported. • The two electronic configurations are close in energy; being the two Mo(V) polarons the most stable.

    • The conversion between the two electronic structures between an electron hopping has an energy barrier of 0.30 eV, easily overcomed at typical reaction temperatures. Therefore, the two oxidation states will be present during catalytic activity. • The vacancy formation energy of 2xMo(V) configuration increases with the coverage. Whereas, the vacancy formation energy of Mo(IV) is constant until a coverage of 0.5. At high coverage the inversion of the stability between the two electronic states was found. • The formation of Magnèli phases is explained by the impossibility of the vacancy accumulation along x axis.

    • The adsorption energy of molecules on vacancy positions can be influenced by the electronic state.

    Chapter 5. Propylene production from glycerol: a first principles study of hydrodeoxygenation (HDO) process on MoO3−x • Hydrogen molecules dissociatively adsorb on surface and they can diffuse over the surface with low energy barriers. The surface hydrogens can recombine with one surface oxygen to form water molecules that are released generating vacancy defects on the surface. • The high energy barriers for hydrogen adsorption explain why high H2 pressure is needed in experiments to reach high activity. • The hydrocarbons are deoxygenated on vacancy defects. They adsorb forming a Mo-O bond. A selective C-O bond cleavage produces the deoxygenated compound while the oxygen atom remains bonded to Mo center healing the vacancy and closing the catalytic cycle. • Glycerol is reduced to propylene through a reaction network with 16 intermediates interconnected by 20 reactions. All the intermediates experimentally observed are reachable in the reaction network. • In the reaction network there are three types of reaction: i) dehydrations, ii) keto-enol equilibriums, and iii) hydrogenations. The mechanism was elucidated for the three reaction types and all the elementary steps of reaction network were calculated. • The high reaction barrier and the necesity of two neigboring empty vacancy defects for the C-C bond cleavage explain the selectivity for this HDO process.

    Chapter 6. Selective C 2 epimerization of aldoses catalyzed by Mo-based compounds• The mechanism for the epimerization of aldoses on H3PMo12O40 was unraveled. The epimerization encompasses three elementary steps through a single Mo coordination: i) dissociative adsorption, ii) 1,2 C-shift, and iii) associative desorption.

    • During 1,2 C-shift step C 1 and C 2 exchange its functional groups, aldehyde and alkoxy. This large electronic rearrangement is assisted by the Mo centers. • The same mechanism can be applied to the catalyzed epimerization on larger POMs and α-MoO3(010) surface.• The presence of water as solvent reduces the energy barrier of the three elementary steps. In dissociative adsorption and associative desorption one water molecule acts as proton shuttle. During the 1,2 C-shift two water molecules assist the carbon chain rearrangement by the interaction with OH groups.• Different heteroatom and protonation grades in the Keggin cluster have an influence in the charge transfer between the heteroanion and the Mo12O36 framework. This charge transfer modifies the redox properties of the Mo atoms influencing the barrier for the 1,2 C-shift. • Regarding the POM stability the electrostatic interactions between the heteroanion and the Mo12O36 framework avoid the leaching observed in the bulk catalyst.• Reducibility, calculated as hydrogen addition energy (HAE), is an electronic descriptor for adsorption energies and 1,2 C-shift energy barrier. The representation of the rate as function of hydrogen addition energy (HAE) is a volcano plot. Its maximum indicates the HAE of the compound that would have the maximum rate.