Experimental study and reaction kinetics of 1,3-butadiene synthesis from bioethanol over a Hf-Zn/SiO2 catalyst

i Departamento de Ingeniería Química y Ambiental Escuela Técnica Superior de Ingeniería Universidad de Sevilla 2020 Gracia María Cabello González Experimental study and reaction kinetics of 1,3-butadiene synthesis from bioethanol over an Hf-Zn/SiO2 catalyst DOCTORAL THESIS Author: Gracia María Cabello González Thesis supervisors: Dr. Ángel L. Villanueva Perales Dr. Fernando Vidal Barrero Dr. Manuel Campoy Naranjo SiO2 Zn2+ Zn2+ Zn2+ Zn2+ Zn2+ Zn2+ Hf4+ Hf4+ Hf4+ Hf4+ Hf4+ Hf4+ Zn0 Zn0 Coke Coke Coke Coke Coke OH OH i ii Doctoral Thesis Experimental study and reaction kinetics of 1,3-butadiene synthesis from bioethanol over an Hf-Zn/SiO2 catalyst Author: Mrs. Gracia María Cabello González Thesis supervisors: Dr. D. Ángel L. Villanueva Perales Dr. D. Fernando Vidal Barrero Dr. D. Manuel Campoy Naranjo Dpto. de Ingeniería Química y Ambiental Escuela Técnica Superior de Ingeniería Universidad de Sevilla Sevilla, 2020 iii iv Not all those who wander are lost v vi Acknowledgements This work would have never been possible without my thesis supervisors, Dr. Fernando Vidal, Dr. Manuel Campoy and Dr. Ángel L. Villanueva, for their guidance and dedication. They have been the helm in the storm. I wish also to acknowledge the invaluable help provided by Agustín Martinez and his group from the Instituto de Tecnología Química (ITQ-UPV-CSIC) for the collaboration in the realization of this study. I am particularly grateful to my family, for their unconditional support along the years, as well as to Luis, for his patience, affection and understanding. Also to my friends since always. Sergio, for being the hammer, the edge of my sword, the tip of my spear, the mail about my fist and the flight of my arrows. Thanks to Javi and Marichu for fighting alongside me against leaks and glass wool. To Sara and Bego, for being always there. I would like to express my very great appreciation to those people who have lately managed to sneak out in my life without me noticing. To Emilio for his intense conversations about time travel and cat behaviour; Isa and Max, experts in making me re/disconnect; Jose “the kinetic guy”, buen clima; Chris, for being an human grammar corrector and the best Penguin bars provider ever; Ivan, for letting me take him out of home from time to time; and last but not least, Pedro, Mr. Blue Caterpillar, for filling my mornings and afternoons with laughs. Thanks to my colleagues from L13 and other places of the ETSI, Fernando Castaño, Mari Ángeles, Miguel, Paco, and Camilo. Also Maite, Diego and Rafael. Days go faster with you. Finally, I would like to offer my special thanks to my sabonims Javier and Alejandro, and to my comrades in arms, especially to my nakamas Fran and David, but not forgetting Jesús, Natacha, Anika, Galván, Ricardo and everyone else. They taught me other ways to face the roads. Hai dong! vii viii Abstract In recent years, considerable interest in bio-1,3-butadiene has been aroused by the reduced 1,3-butadiene supply, due to the shift from oil-based naphtha cracking to shale gas-based ethane cracking as a consequence of the extreme volatilities in fossil oil price and low shale gas cost in the United Stated. Besides, the steam cracking process is associated to a high energy consumption and release considerable amount of greenhouse effect gases to the atmosphere. These facts have stimulated the interest in robust cost-efficient alternative systems that can be adapted to the new environmental policies in order to mitigate climate change and reduce dependency on importations. In this framework, ethanol raises as an attractive and environmentally friendly feedstock for the production of 1,3-butadiene as a chemical product at commercial scale as it can be cost-efficiently produced from a variety of renewable biomass and waste sources and its production is growing worldwide, what makes it the perfect candidate for integrated biorefining systems. The process includes two routes: the two-step pathway where, in the first step, acetaldehyde is generated from ethanol and, in the second step, the acetaldehyde/ethanol mixture is converted to 1,3-butadiene; and the one-step method where both reactions occur over the same catalyst. xv Además, se propuso y validó, a través de resultados experimentales, un modelo cinético de la síntesis de 1,3-butadieno a partir de etanol sobre el catalizador Hf-Zn/SiO2. El modelo es capaz de predecir la formación de los principales productos de reacción y ajustar el rendimiento del catalizador para el caso de que haya agua presente en la alimentación de etanol. Por lo tanto, es útil para optimizar el diseño del proceso de un solo paso, considerando las interacciones entre el reactor y el área de separación. Por último, la principal causa de desactivación del catalizador es la generación y deposición de especies de coque formadas en los centros deshidrogenantes asociados con la hemimorfita del catalizador (Zn2+), pero también por una pérdida de esos centros debido a la reducción de Zn2+ a Zn0. En este sentido, una mayor presencia de agua en el medio alarga la vida del catalizador, ya que inhibe la formación de compuestos pesados precursores del coque. La regeneración por calcinación con aire elimina el coque y reoxida una fracción de Zn0 a Zn2+, pero no restaura completamente el equilibrio original de centros. 1 Summary Acknowledgements vi Abstract viii Resumen xii Summary 1 List of figures 3 List of tables 5 Nomenclature 7 Outline 11 1. Introduction 13 2. Background and scope 19 3. Objetives 23 4. Methodology 25 4.1. Description of the experimental facility 25 4.2. Study of chemical equilibrium 28 4.3. Catalysts characterization 29 4.4. Determination of the reaction pathway 30 4.5. Effect of operating conditions and water content in ethanol on catalyst performance 31 4.6. Response surface analysis 33 4.7. Catalyst deactivation and regeneration 34 4.8. Kinetic model building and validation 35 2 5. Results and discussion 43 5.1. Study of chemical equilibrium for the ethanol conversion into 1,3-butadiene 43 5.2. Conversion of ethanol into 1,3-butadiene over an HfZn/SiO2 catalyst. 49 5.2.1. Catalyst selection and characterization 49 5.2.2. Determination of the reaction pathway 53 5.2.3. Effect of operating conditions 60 5.2.4. Effect of water content in ethanol 63 5.2.5. Response surface analysis 67 5.2.6. Catalyst deactivation and regeneration. 74 5.2.7. Kinetic model building and validation 81 5.3. Publications 94 6. Conclusions and future work 97 References 105 Appendix 113 3 LIST OF FIGURES Figure 1. Global Trends in Ethanol, Biodiesel and HVO/HEFA Production [12]. .... 14 Figure 2. Generally accepted pathway for ethanol conversion to 1,3-butadiene. ..... 16 Figure 3. Simplified diagram of the experimental setup. ............................................... 26 Figure 4. Molar fraction (y) versus W/F plot for 1,3-butadiene. .................................... 39 Figure 5. Variation of the equilibrium composition for the ethanol to 1,3-butadiene overall reaction: a) with temperature at 1 atm and b) with pressure at 300°C ......... 44 Figure 6. Variation of the equilibrium composition for the ethanol to 1,3-butadiene process with temperature at 1 atm taking into account the main secondary products. ..................................................................................................................................................... 45 Figure 7. Ellingham diagram (1 atm) for the reaction scheme presented in Figure 2 ..................................................................................................................................................... 46 Figure 8. Equilibrium composition reached for the ethanol to 1,3-butadiene reaction while changing the water content in the feedstock at 360°C. ........................................ 47 Figure 9. Ethanol to 1,3-butadiene main products yields as a function of ethanol conversion. ................................................................................................................................ 54 Figure 10. Proposed reaction pathway for the formation of 1,3-butadiene and side products from ethanol over a HfZn/SiO2 catalyst. ........................................................... 59 Figure 11. Effect of temperature and WHSV on the ethanol conversion over the mixed Hf-Zn catalyst. P= 1 bar, PEtOH= 0.21 bar, feed is anhydrous ethanol. ............. 60 Figure 12. Effect of temperature and WHSV on the yield of a) 1,3-butadiene, b) acetaldehyde, c) ethene plus diethyl ether, d) butenes, and e) heavy products (C6+), over the mixed Hf-Zn catalyst. ............................................................................................. 62 Figure 13. Effect of water content in ethanol feed on a) ethanol conversion, b) acetaldehyde yield, and c) 1,3-butadiene yield at 360 °C (left panels) and 380 °C (right panels) as a function of space velocity and water content (wt%) in ethanol. ............ 65 4 Figure 14. Effect of water content in ethanol feed on a) heavy compound yield and b) ethene + diethyl ether yield at 360 °C (left panels) and 380 °C (right panels) as a function of space velocity and water content (wt%) in ethanol. ................................... 66 Figure 15. Parametric curves of the response surface analysis model for 1,3-butadiene yield: a) variation with temperature and WHSV, b) variation with water content and WHSV. Temperatures in °C, WHSV in h-1, and water content in wt%. ...................... 70 Figure 16. Parametric curves of the response surface analysis model for ethanol conversion: a) variation with temperature and WHSV, b) variation with water content and WHSV. Temperatures in °C, WHSV in h-1, and water content in wt%.71 Figure 17. Parametric curves of the response surface analysis model for 1,3-butadiene selectivity: a) variation with temperature and WHSV, b) variation with water content and WHSV. Temperatures in °C, WHSV in h-1, and water content in wt%. ............. 72 Figure 18. Effect of deactivation on ethanol conversion, main product selectivities and yields. ................................................................................................................................. 75 Figure 19. Effect of deactivation on ethanol conversion and 1,3-butadiene and acetaldehyde yield. ................................................................................................................. 77 Figure 20. a) Solid-state 1H-to-13C CP-MAS NMR spectrum of spent Hf-Zn catalyst (HfZnused); b) main carbonaceous species detected by GC-MS in the coke extracted from HfZnused: 1) 4-tert-butylphenol, 2) diphenylketone, 3) 2-naphtaldehyde, and 4) dihexyl phthalate. .................................................................................................................... 78 Figure 21. Comparison plots between the experimental and model flow rates. ...... 88 Figure 22. Arrhenius plot for the eight reactions of the model. .................................... 92 5 LIST OF TABLES Table 1. Identified catalyst for the one-step ethanol to 1,3-butadiene process. ......... 50 Table 2. Textural properties of the individual HM (hemimorphite) and Hf/SiO2 solids and the Hf-Zn catalyst as measured by N2 physisorption [25]. .................................... 52 Table 3. Estimated parameters for the response variables. ............................................ 68 Table 4. Prediction of response variables for the validation data set........................... 69 Table 5. Operating conditions for maximum yield or selectivity to 1,3-butadiene. . 74 Table 6. Atomic surface composition and metal surface ratios determined by XPS. ..................................................................................................................................................... 80 Table 7. Calculated values of kinetic parameters. ............................................................ 87 6 7 Nomenclature a Fitting parameters of the water corrective terms (bar-1) A Pre-exponential factor (mol/g h barΣni) Ac Acetaldehyde b Fitting parameters of surface response model BD 1,3-Butadiene But Butenes ButA Butanal ButOH Butanol C Total number of compounds CP-MAS NMR Cross Polarization Magic Angle Spinning Nuclear Magnetic Resonance C4 Butenes C6+ Heavy compounds d Diameter (mm) D Diffusion coefficient DEE Diethyl ether e residual Ea Activation Energy (kJ/mol) Et Ethene EtOH Ethanol F Total molar flow (mol/h) Fk Molar flow of compound k (mol/h) FID Flame Ionization Detector H Hydrogen HC Heavy Compounds ICP-OES Inductively coupled plasma – optical emission spectrometry IR Infrared radiation J Jacobian matrix of parameters k Mass transfer coefficient (m/s) L Length (mm) m Exponent of the water corrective term 8 MPVO Meerwein-Ponndorf-Verley-Oppenauer reduction n Reaction order N Total number of tests Ndf Number of degrees of freedom NR Number of reactions in the kinetic model OC Oxygenated Compounds P Total pressure (bar) Pk Partial pressure of compound k (bar) r Reaction rate (mol/h g) R Ideal gas constant (8.314 J/mol K) RSM Response surface methodology Sk Carbon selectivity to compound k SSEf Sum of the squared errors of molar flow rates SSEri Sum of the squared errors of the reaction rates for reaction i STEM-EDX Scanning transmission electron microscopy – Energy-dispersive X-ray t Student’s t-distribution T Reaction temperature (°C or K) TCD Thermal conductivity detector TOS Time-on-stream (h) Tr Trace of a matrix V Covariance matrix of parameters W Mass of catalyst (g) Wat Water Wc Water content (wt %) WHSV Weight hourly space velocity (h-1) Xk Conversion of compound k XPS X-ray photoelectron spectroscopy XRD X-ray powder diffraction yk Mole fraction of compound k Yk Carbon yield of compound k 9 Subscripts/ Superscript b c e i j bulk catalyst effective reactions experiments k compounds obs observed p pellets ref reference regen regenerated S surface of the catalyst ^ predicted by the model Greek letters  confidence level  density (kg/m3)  standard deviation  vector of kinetic model parameters  stoichiometric coefficient 1. Introduction 16 The ethanol to 1,3-butadiene pathway over heterogeneous catalysts is still a subject of debate. The most accepted route nowadays comprises ethanol dehydrogenation to acetaldehyde, aldol condensation of acetaldehyde to crotonaldehyde, selective reduction of crotonaldehyde by ethanol through Meerwein-Ponndorf-Verley-Oppenauer (MPVO) reaction, and finally, the dehydration of crotyl alcohol to 1,3-butadiene (Figure 2) [3]. That way, the maximum mass yield from ethanol (EtOH) to 1,3-butadiene (BD) is 0.587 kg BD/kg EtOH. Figure 2. Generally accepted pathway for ethanol conversion to 1,3-butadiene. CH3OH CH3O OH CH3O CH3OCH3OH CH2 CH2 Ethanol Acetaldehyde Crotyl alcohol 1,3-Butadiene Crotonaldehyde3-hydroxybutanal H2 OH2 OH2 CH3OH + OH2 H2 CH3OH + CH3O +CH3O 1. Introduction 17 Apart from 1,3-butadiene, the reaction main products are water and hydrogen, as well as acetaldehyde, product of the ethanol dehydrogenation. Other intermediate compounds, such crotonaldehyde and crotyl alcohol, are detected as traces. The main side-products are ethene and diethyl ether, direct ethanol dehydration products; ethyl acetate which is believed to be a product of the Tishchenko reaction; acetone, isopropyl alcohol, propylene, C4-compounds such as butanal, butanol and butenes; C6-products and heavier compounds, probably formed by the self and cross condensation of aldehydes and alcohols [3]. Although the technical feasibility of the ethanol to 1,3-butadiene route has been already demonstrated by its commercial experience, there is still room for improvement. First, the yield and 1,3-butadiene selectivity of the catalysts used in commercial plants was relatively low. Lebedev assured that the commercial catalyst reached a 1,3-butadiene selectivity carbon based of 70% but, unfortunately, they never revealed the elementary composition of the catalyst so this data could not be checked. In the two-stage process, the yield of the first stage was 92% over a copperbased dehydrogenative catalyst, while the subsequent conversion of ethanol and acetaldehyde to 1,3-butadiene reached a carbon-selectivity of 56% over a silica catalyst doped with Ta2O5 [3,13]. The catalyst improvement lies in finding an optimal distribution of acidic and basic centres that maximize 1,3-butadiene selectivity, minimizing the by-product formation and deactivation by coke deposition. In the last years, some catalysts exceeding the performance of old commercial ones have come to light with a 1,3-butadiene selectivity up to 90%. 1. Introduction 18 The impact of this improvement on the economics of the processes is uncertain yet since there exist few techno-economic analysis published in the literature but, from these analyses, it can be concluded that, although the environmental improvement seems significant, the economy of the process still depends on several variable that makes it risky to compete with the traditional oil derived route [6,7,12,15]. In this regard, it is necessary to know how the ethanol composition and its impurities affects the catalyst performance, as well as the side-products recycle. There is currently little information about the effect of water present in the ethanol feedstock on catalyst performance [14,16–18]. This knowledge would allow to know what degree of bioethanol refinement is necessary, since the lower the refinement, the cheaper the ethanol feedstock. Also, the effect of by-product recirculation is key to determine to what extent it is convenient to recirculate to find a trade-off between operating and investment costs and 1,3-butadiene productivity. That way, it is necessary to determine a kinetic model of the catalysts in order to be able to choose the reaction conditions that optimize the whole process. 2. Background and scope 19 2. BACKGROUND AND SCOPE Nowadays, the always increasing world energy and chemicals demand is mostly satisfied by the consumption of fossil resources. The main problems derived from its extensive use are the depletion of the oil reserves, the high dependence of countries with fossil resources and high supply risks, and the amount of greenhouse gas emissions in their combustion or transformation. An alternative to cope with this situation in the chemical sector is to develop biorefineries, integrated bioresourcesconversion processes used to generate biobased products (ethanol, butanol, 1,3- butadiene, acetaldehyde, aromatic compounds…), as an alternative to current petroleum refineries, changing the market to competitive bioresources produced from renewable primary energy. Within this context, the Bioenergy Group of the Chemical and Environmental Engineering Department (DIQA) of the University of Seville has collaborated on numerous research projects that contribute to the development of new processes for the conversion of bioresources in value-added products:  Renewable Fuels for Advanced Power Trains (RENEW) project (SES6-CT- 2003-502705. 2004-2007), a pan-European project, supported under the European Commission’s 6th Framework Programme, about the development of novel ways of producing automotive biofuels from lignocellulosic biomass. DIQA worked about the catalytic route to produce ethanol from synthesis gas. 2. Background and scope 20  Production of fuel grade bioethanol via thermochemical route (BIOCOM) project (ENE2005-08492. 2006-2008), a Plan Nacional project funded by the Ministry of Economy and Competitiveness, about the “Thermochemical Production of Bioethanol for Automotive”, studied the behaviour of direct synthesis catalysts in a slurry-type reactor and the economic and life-cycle technical analysis of new process alternatives.  Research and development of fuel grade bioethanol (I+DEA) project (PI- 0063/2007. 2007-2010), about the “Research and Development of Ethanol for Automotive”, was co-financed between the Centre for Technological and Industrial Development (CDTI) and Abengoa Bioenergy.  Sustainable Exploitation of CO2 (SOST-CO2) project (PI-0690/2009. 2008- 2011), co-funded by the Ministry of Economy and Competitiveness and the Spanish Society of Metallic Carbides in order to investigate an alternative to CO2 confinement underground, analysing from the capture of CO2 in emission sources, through its transportation, storage and recovery.  Experimental development of lignocellulosic biomass and other carbon sources transformation processes in various bioproducts in Andalusia (BIOANDALUS) project (PI-0955/2012. 2012-2014), co-financed between the Centre for Technological and Industrial Development (CDTI) and Abengoa Bioenergy, about the transformation of ethanol into butanol and bio jet-fuel was investigated. 2. Background and scope 21  Ethanol conversion to added-value products project (PRJ201101398. 2012- 2016), financed by Abengoa Bioenergy, aimed for the design of catalytic processes for the conversion of ethanol into products with high added value.  Thermochemical biorefinery based on DME project (ENE2012-31598. 2013-2015), funded by the Ministry of Economy and Competitiveness with the purpose of identifying the best reaction conditions and develop kinetic models of the catalyst deactivation. The framework of this study is the BIODIENE project (Biobutadiene production from bioethanol, CTQ2015-71427-R), funded by the Ministry of Economy and Competitiveness through the European Regional Development Fund. This research is encouraged by the remarkable development of 1,3-butadiene synthesis catalysts in the last 20 years and aims to deal with research areas like process design and reaction system in order to find optimum process configurations and reaction conditions, comparing these configurations with fossil-based 1,3-butadiene production from the point of view of economics and environmental sustainability. The project covers, on the one hand, the experimental research, with one and two-steps catalyst, focused on finding the optimum reaction conditions and developing kinetic models and, on the other hand, the conceptual design of new process configurations. 2. Background and scope 22 This thesis is dedicated to the experimental part of the project, being focused on the study of the industrial aspects of the one-step catalysts, such as the effect of the reaction conditions and the presence of impurities in the feed on their performance, and also their deactivation, in order to extract data for the industrial process design. That way, some catalysts were selected among those found in literature as the most promising ones and they were synthesized by the Instituto de Tecnología Química (ITQ-UPV-CSIC) in Valencia. A first screening was carried out to determine the best catalyst for the one-step route and, finally, the bimetallic HfZn/SiO2 catalyst presented in the work by Trees et al. [19] was selected. The scope of this work is to determine the best operating conditions, studying how the reaction conditions and side-products recirculation affects the catalyst performance and to determine the lifespan and catalyst regenerability. Besides, a kinetic model, including the formation of the main side-products, was created to support the development of the industrial process. Another thesis in the context of the BIODIENE project is being developed, whose goal is the study of the technical, economic and environmental feasibility of different process alternatives for the ethanol to 1,3-butadiene production. It takes the data collected in this document as a base for the design of the process that is simulated using Aspen Plus® software. That way, investment and operating costs are being estimated, as well as Life Cycle Assessment, with the objective of identifying the most economically and environmentally promising alternatives. 3. Objectives 23 3. OBJECTIVES The main goal of this thesis is to investigate critical issues of the production of 1,3-butadiene from ethanol that can make the technology competitive against the current petrochemical route, focusing on the key aspect for its industrial application such as the influence of the operating condition, the effect of possible impurities in the feed and the catalyst lifetime and regenerability. The study will cover the Lebedev one-step process in order to find optimum reaction conditions and develop a kinetic model that will help in the future to optimize the whole process both economically and environmentally. In order to achieve the main goal, the specific targets of this study are:  To review the literature on the conversion of ethanol to 1,3-butadiene in order to identify the most promising catalysts for the one-step route.  To investigate the reaction pathway from ethanol to 1,3-butadiene, as well as the possible routes to the main side-products.  To evaluate the effect of the operating condition on the catalyst performance.  To determine how the content of water and other reaction products in recycled ethanol affects the catalyst performance. 3. Objectives 24  To perform a study of the catalyst deactivation in order to determine the cause and the possible species involved. Also, the effect of the regeneration of the catalyst by air calcination will be assessed.  To develop a kinetic model for the catalyst in order to assist in the design of an industrial facility. 4. Methodology 25 4. METHODOLOGY This chapter is focused on describing the methods employed for obtaining the qualitative and quantitative data necessary to achieve the aforementioned objectives. More detail information about the operation of the experimental facility for each experimental test, such as start-up procedures and operating conditions, can be found in the publications attached at the end of this document. 4.1. Description of the experimental facility All catalytic tests were carried out in a laboratory scale facility (Figure 3) remotely controlled through a SCADA system and divided into three main zones: reactant feeding, reaction and analysis area. The reactant feeding system allows to feed gas and liquid streams into the reaction area. The gas stream, composed by nitrogen as carrier gas, is fed from a pressurized cylinder into the reactor through a mass flow controller. On the other hand, the liquid feedstock, ethanol or and ethanolwater mixture, was charged into a tank, pressurized with an inert gas (nitrogen), and fed into the reactor using an HPLC pump or a Bronkhorst® mini cori-flow. A check valve was placed at the reactor inlet to avoid reverse flow. 4. Methodology 32 To check for catalyst deactivation, the first experiment at the highest space velocity (WHSV=9.8 or 8 h-1) was repeated at the end of the run of each temperaturewater content combination. Three extra tests were performed at 360 °C with a water content of 3.75 wt% in order to obtain more information about the water effect at the central temperature. The water content in ethanol feed was obtained by adding the required amount of deionized Milli-Q water to anhydrous ethanol (Panreac  99.8% v/v). All tests were carried out with 0.5 g of catalyst at 1 bar of total pressure with ethanol partial pressure in the feed of 0.21 bar. The nitrogen flow was adjusted to keep the ethanol partial pressure constant. When water was added to the feed, the necessary amount of nitrogen was replaced by water, therefore keeping constant the ethanol partial pressure and total volumetric flow for a given WHSV. 4. Methodology 33 4.6. Response surface analysis Response surface methodology (RSM) was used to analyse the effect of the operating condition on the catalyst performance, as well as to explore the relationship between the different variables. RSM is a statistical technique used to model and analyze systems in which one variable of interest (response variable) is influenced by others. The purpose is to design an experiment that provides values of the response variable and then determine the mathematical model that best fits the data obtained. The final objective is to establish the values of the factors that optimize the value of the response variable. Ethanol conversion, main product selectivity and 1,3-butadiene yield were set as response variables and the manipulated variables were temperature, weight hourly space velocity and the water content in the feedstock. That way, the experimental tests proposed were carried out in the experimental facility and then, the response surface analysis was performed with the software StatGraphics®. Some experiments were set apart randomly to validate the proposed model. The response variables were fitted with a second-order equation by using linear least squares regression (Equation 2), where T is the reaction temperature in 0C, Wc the water content in wt%, WHSV the weight hourly space velocity in h-1 and bi the regression coefficients. The most influential factors over the response variables were determined making use of the analysis of variance (ANOVA) and the Pareto chart. 4. Methodology 34 2 0 1 2 3 4 5 22 6 7 8 9 ˆ y b b T b Wc b WHSV b T b T Wc b T WHSV b Wc b Wc WHSV b WHSV                       Equation 2 The response surface analysis was finally used to find the operating conditions that maximize the ethanol conversion and 1,3-butadiene selectivity and yield. 4.7. Catalyst deactivation and regeneration The long-term performance of the catalyst and the cause of its deactivation was determined by carrying out long-lasting tests. The evolution of ethanol conversion and main product selectivities was studied and the experiment was halted to regenerate the catalyst when ethanol conversion decreased 20 percentage points from its initial value. After regeneration the operation was resumed. Two in-situ regeneration by air calcination were performed. That way, two experiments were carried out: the first one feeding anhydrous ethanol and the second one feeding ethanol-water azeotropic mixture (7.5% w/w of water in the feed). 4. Methodology 35 4.8. Kinetic model building and validation A kinetic model was established using the data obtained in the tests described in section 4.5, where temperature, space velocity and water content in ethanol were varied. The proposed kinetic model does not contemplate the deactivation of the catalyst by coke deposition. Some of the experiments were randomly set apart to validate the model. A reaction network (with NR reactions) was proposed taking into account the scheme inferred from the kinetic curves, grouping some of the compounds as a simplification when necessary (lumping). A power-law kinetics was assumed for each reaction (Equation 3) and corrective terms were introduced to contemplate the inhibition and promotion of certain reactions by water.                 2 11 11.... 1 ki i ref i Cn Ea k R T T k ii m i H O P r A e i NR aP Equation 3 4. Methodology 36 In Equation 3, i A is the kinetic constant at the central temperature for each reaction i, i Ea the activation energy, ki n the reaction orders and k P the reactant k partial pressure. The term (1 1 ) ref TT is derived from expressing the kinetic constant at temperature T relative to the kinetic constant at the reference temperature ( ref T = central temperature, i.e., 360 ºC). In section 5.2.4, it can be experimentally observed how water affects the different reactions promoting or inhibiting the generation of some compounds. To contemplate this effect, a term has been included into the kinetic equation. The, either inhibition or enhancement term   2 1i m i H O aP , includes i a and i m as fitting parameters, being 0 i m for inhibition terms and 0 i m for enhancement ones. Besides, 2 HO P is the water partial pressure along the reactor. The set of laboratory-scale experiments presented in Cabello et al. [20] and summarized in table A1 and A2 (Appendix) was used to regress the 39 parameters of the model, and 10 out of the 58 catalytic tests (test number 7, 19, 22, 24, 29, 33, 39, 50, 51 and 56) were randomly set apart to validate the model. Ideal plug flow reactor was assumed for the experimental laboratory scale facility since the ratio between the length (L) and the diameter (D) of the reactor tube is greater than 20 (L/D=31.1 > 20). That simplification means that the fluid is perfectly mixed in the radial direction. 4. Methodology 37 That way, the molar balance for each compound expressed as a function of the reaction rate is shown in Equation 4, being C the number of compounds, W the mass of catalyst, k F the molar flow rate of each compound k,  ki the stoichiometric coefficient of each compound k in the chemical reaction i, and i r the reaction rate.     1 1... NR k ki i i dF r k C dW Equation 4 Regression of the kinetic parameters was carried out by applying the principle of maximum likelihood, which under the following assumptions, is equivalent to minimize an objective function [21,22]. Thus, by assuming that errors in the observations are independent, normally distributed, with constant variance for each dependent variable, and that the covariance between dependent variables is negligible, the resulting objective function is the sum of the squared errors (or residuals, e) of molar flow rates (SSEf) calculated as the difference between the experimental and estimated molar flow rate at the reactor outlet for each compound divided by the variance of the experimental error. In the objective function (Equation 5), kj F is the molar flow rate of compound k at the outlet of the reactor for the experiment number j; 2 k  is the variance of the experimental error for each compound flow rate independently determined from the result of, at least, three measurements taken for each condition; N is the number of experiments and C the number of compounds. 4. Methodology 38 The molar flow rate of each compound at the reactor outlet was calculated by integration of the differential equations from the molar balance in the reactor (Equation 4).                 22 22 1 1 1 1 11 ˆ C N C N kj kj kj k j k j kk SSEf F F e Equation 5 A good initial guess of the kinetic parameters is necessary in order to help the optimization algorithm to converge to a set of the regression parameters. For that purpose, a simplified method was carried out to estimate the parameters. Constant total molar flow rate (F) along the reactor can be assumed as the flow rate of N2 is significantly higher than that of the reagents (Table A4). This assumption allows to avoid the integration of the differential equation system because the net reaction rate of compound k can be expressed as a function of the variation of the molar composition ( k y ) (Equation 6) and, therefore, be calculated as the slope of the curve k y W F (such as Figure 4). Many WF values can be chosen, and the corresponding net reaction rates be calculated from the kinetic curves of each compound. kk dF dy dW W dF    Equation 6 4. Methodology 39 That way, the net reaction rate for each compound and every experiment was determined obtaining a large set of net reaction rates for many operating conditions and, then, the gross reaction rates ( i r ) were calculated through the molar balance (Equation 4). Lastly, an initial guess of the kinetic parameters were determined separately for each reaction i, making use of non-linear regression by minimizing the sum of the squared errors of the reaction rates (SSEri) (Equation 7). Figure 4. Molar fraction (y) versus W/F plot for 1,3-butadiene. Tests 1-30 Table A4.       2 1 ˆ1.... N i ij ij j SSEr r r i NR Equation 7 y = -0.0002x2+ 0.0009x + 0.0271 y = -0.0008x2+ 0.0104x + 0.0032 y = -0.0006x2+ 0.0083x + 0.0034 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 0.0 2.0 4.0 6.0 8.0 10.0 y W/F 1,3-Butadiene 380 360 340 4. Methodology 40 Once the kinetic parameters were obtained with the rigorous method, a statistical analysis based on confidence intervals was carried out to study its uncertainty. The analysis provides a range of the values in which the value of the parameter is true with a certain probability. The determination of the confidence intervals for the parameters was calculated with (Equation 8), where  ˆ is the estimated parameter vector, /2,Ndf t  the t-Student value for an α confidence level and Ndf degrees of freedom, and   Tr V  the variance-covariance matrix transpose.   /2, ˆNdf t Tr V      Equation 8 The variance-covariance matrix was estimated through the approximation to the hessian known as the Gauss-Newton approximation. That way, the variancecovariance matrix can be approached as the inverse of the product of the transpose Jacobian matrix and the Jacobian matrix itself (Equation 9).   1 T V J J    Equation 9 Finally, a significance test was carried out and the p-value was calculated for each parameter. Those parameters whose p-value was over the chosen significance level, 0.05, were considered not significant and therefore, rejected, simplifying the model. The parameters were discarded one by one and the model optimization and statistical test was repeated after each elimination until getting a model where all the parameters were found significant. 4. Methodology 41 All the calculations were carried out using Matlab® software. The integration of the differential equations along the reactor was carried out with the ode45 routine. The function fminsearch, which implements the Nelder-Mead directsearch algorithm, was used to minimize the objective function (Equation 5) until a value close to the minimum. Then, the function fmincon, which make use of the interior-point algorithm and allows setting bounds for the parameters, was employed for a final parameter estimation. The lsqnonlin function was used to obtain the Jacobian matrix at the optimum point. For the simplified method, the minimization of the sum of the squared errors of the reaction rates (Equation 7) was carried out making use of the Solver Tool in MS Excel®. Besides, an evaluation of the possible external and internal diffusion limitation was assessed to assure that the chemical reaction was the controlling resistance in the catalytic bed. Mears criterion was applied to evaluate the influence of external mass transfer effects. This criterion assures that external mass transport limitations can be neglected in the case of a Mears number (CMears) under 0.15. CMears number is defined in Equation 10, where -robs is the observed reaction rate (kmol/kg·s), b  is the bulk density of catalyst bed (kg/m3), dp is the diameter of the pellets (m), nk is the reaction order of reactant k, k is the mass transfer coefficient (m/s), and CkB the concentration of the reactant k in the bulk gas phase (kmol/m3). It was calculated, using the reactor output values for the most unfavourable operating conditions at the three different temperatures and for each water content in the feed. 5. Results and discussion 48 To sum up, since the ethanol to 1,3-butadiene reaction is not thermodynamically favoured with pressure, operation at atmospheric pressure looks like the first choice. For the selection of the operating temperature, one should bear in mind that the most thermodynamically unfavourable reactions are the dehydrogenation to acetaldehyde and the aldol reaction of acetaldehyde to 3- hydroxy-butanal. Therefore, it seems that the operating temperature should be chosen so these reactions are favoured as much as possible but limiting possible side reactions. According to the literature, the typical operating temperature sets between 350 - 430°C [3]. Finally, the presence of water in the feedstock inhibits the 1,3- butadiene production shifting the equilibrium of the dehydration reactions so ethanol feedstock with low water content may be preferred. The acetaldehyde condensation route is the most accepted pathway. In 2013, Angelici et al. [23] compared Prins condensation, where 1,3-butadiene is formed from ethene, and aldol condensation route, concluding that the latter was thermodynamically more favourable if crotonaldehyde reduction with ethanol was considered. Later, in 2014, Makshina et al. [3] performed a thermodynamic analysis of the aldehyde condensation route with similar results to those presented here. Consequently, the aldol condensation pathway seems to be the most feasible route. However, Ochoa et al. [24] proposed an alternative route over MgO-SiO2 catalysts, where acetaldehyde reacts directly to produce crotyl alcohol that further dehydrates into 1,3-butadiene. Since these route yields similar results when compared with the aldol condensation route, it would be necessary further studies to determine the most likely path towards 1,3-butadiene. 5. Results and discussion 49 5.2. Conversion of ethanol into 1,3-butadiene over an HfZn/SiO2 catalyst. 5.2.1. Catalyst selection and characterization A bibliographic analysis was carried out with the objective of identifying the most promising catalyst for the ethanol to 1,3-butadiene one-step process. The criteria followed was to choose a cheap catalyst (absence of precious metals) with a good balance between the 1,3-butadiene selectivity and yield. That way, De Baerdemaeker et al. [19] HfZn/SiO2 catalyst was selected as it does not contain precious metals and presents one of the highest 1,3-butadiene selectivity and yield in the literature (Table 1). The HfZn/SiO2 catalyst was synthesised by the ITQ (Instituto de Tecnología Química, UPV-CSIC) according to the methodology reported in Baerdemaeker et al.[19] work. The complete characterization of the synthesized catalyst, along with the images generated in the different analysis commented in this section, can be found in the work of Cabello et al. [25]. 5. Results and discussion 50 Table 1. Identified catalyst for the one-step ethanol to 1,3-butadiene process. Doped alumina catalyst Catalyst T (ºC) X (%) SBD (%) YBD (%) References Al2O3/ZnO 400 58 31 18 Lebedev et al [13,26] 20Al2O3/80MgO 430 91 11 10 Natta et al [27] 2PbO/98Al2O3 425 - - 18 Corson et al [28] 40ZnO/60Al2O3 425 94.4 59 55.8 Bhattacharyya et al [29] 30ZnO/4SiO2/4MgO/1K2O/61γ-Al2O3 420 42.6 80 34.8 Tret’jakov et al [30] 40ZnO/60Al2O3 400 41.4 53.1 22 Ezinkwo et al [31] Doped magnesia-silica catalyst Catalyst T (ºC) X (%) SBD (%) YBD (%) References 60MgO/40SiO2 430 85 41 35 Natta et al [27] 2Cr2O5/59MgO/39SiO2 415 82 52 43 Natta et al [27] 2Cr2O5/59MgO/39SiO2 425 80 49 39 Corson et al [28] Na2O/MgOSiO2 350 100 87 87 Ohnishi et al Na2O/MgOSiO2 350 53 30 16 Kvisle et al [32] NiO/MgOSiO2 282 59 90 53 Kitayama et al [33] Ag/MgSiO 480 85 50 42 Janssens et al [34] CuO/MgSiO 425 80 49 39 Angelici et al [35] MgO/SiO2 450 95 77 73 Huang et al [36] Other catalyst Catalyst T (ºC) X (%) SBD (%) YBD (%) References CuZrZn/SiO2 375 45 65 30 Jones et al. [37] CuHfZn/SiO2 360 98.8 70 69 De Baerdemaeker et al [19] 2Ag/4ZrO2/SiO2 320 55.2 71.3 39 Sushkevich et al [38] Ag/Zr/BEA 320 30 66 21 Sushkevich et al [39] Na/ZnZrO 350 97 47 45.6 R.Baylon et al [40] Cr-Ba/MCM 41 450 100 30 30 La Salvia et al [41] Zn/Talc 400 43 49 21.2 Sekiguchi et al [42] CuTaSiBEA 325 88 73 64 Kyriienko et al [43] 1ZnO-5ZrO2/SiO2 400 95 41 39 Patil et al [44] Zn0.5–ZrO2–SiO2 310 36 87 31 Zhang et al [45] Ge-Talc 400 44 71 31 Akiyama et al [46] Zn−Y/Beta 400 95 71 67.5 Yan et al [47] 1Ag/4ZrO2/SiO2-SBA-16 325 99 71 70 Dagle et al [48] ZnTa-TUD-1 400 94 73 68.6 Pomalaza et al [49] Being T temperature, X ethanol conversion, SBD 1,3-butadiene selectivity and YBD 1,3-butadiene yield 5. Results and discussion 51 The textural parameters achieved for the individual HM and Hf/SiO2 solids and the Hf-Zn catalyst prepared by the ITQ are gathered in Table 2. As seen there, the textural properties of the Hf/SiO2 solid showed minor deviations with respect to those of the bare SiO2 support. In comparison, the hemimorphite (HM) exhibited much lower BET surface area and total pore volume, although higher average pore size. The Hf-Zn catalyst, prepared by physically mixing 85 wt% Hf/SiO2 + 15 wt% HM, displayed values that are intermediate to those of the individual components, although much closer to the Hf/SiO2 solid. According to repeated ICP-OES analyses, the Hf and Zn contents in the catalyst are, respectively, 2.5 (±0.1) wt% and 10.0 (±1.0) wt%, what means a lower Hf loading compared to the reference catalyst (2.5 < 3.5 wt% [19]). The synthesized HM sample exhibits only the characteristic diffractions for hemimorphite XRD patterns confirming its successful preparation and high purity. On the other hand, the absence of reflections related to Hf compounds in the Hf/SiO2 component evidences its high dispersion on the amorphous SiO2 carrier. The Hf-Zn catalyst shows only the X-ray diffractions associated to the hemimorphite phase, albeit with a much attenuated intensity with respect to pure hemimorphite, as expected by its high dilution with Hf/SiO2 in the physical mixture [25]. 5. Results and discussion 52 Table 2. Textural properties of the individual HM (hemimorphite) and Hf/SiO2 solids and the Hf-Zn catalyst as measured by N2 physisorption [25]. BET area Total pore volume Mean pore diameter Sample (m2/g) (cm3/g) (nm) Hf/SiO2 206 1.26 26 HM 51 0.46 42 Hf-Zn 179 1.10 28 The acid properties of hemimorphite (HM), Hf/SiO2, and Hf-Zn catalyst assessed by FTIR spectroscopy of adsorbed pyridine show that the Hf-Zn catalyst, as well as its components, exhibit essentially Lewis-type acidity, being the Lewis acid sites associated with Zn2+ species in hemimorphite of a slightly higher strength than those related to Hf4+ sites. Nevertheless, both contribute to the catalyst total Lewis acidity. The concentration of Lewis acid sites amounts to 65, 39, and 41 mol/g for HM, Hf/SiO2, and Hf-Zn catalyst, respectively. These results indicate that the Lewis acidity of the catalyst is mostly determined by that of its main component Hf/SiO2. Moreover, the amount of Brønsted acid sites should be very low and/or that their strength is too low to be detected by this method. Low temperature (-170 °C) FTIR spectroscopy of CO adsorption tests confirm the presence of OH groups (i.e. Brønsted acid sites) of different acid strength [25]. Overall, the characterization results of the synthesised catalyst are in good agreement with those reported for the equivalent Hf-Zn catalyst in De Baerdemaeker et al. [19]. 5. Results and discussion 53 5.2.2. Determination of the reaction pathway The kinetic curves (product yield vs reagent conversion) allow to distinguish stable/unstable and primary/secondary products. In order to study the reaction scheme of 1,3-butadiene formation from ethanol, kinetic curves were obtained for the main products performing several tests with different load of catalyst at several WHSV. The results of the experimental test carried out for this study are gathered in Table A1 and Table A2 of the Appendix. Acetaldehyde is the main intermediate product of the ethanol-to-1,3- butadiene reaction. The acetaldehyde curve resembles that of a primary unstable product (Figure 9a), indicating that it is directly generated by the dehydrogenation of ethanol. This reaction takes place on the Lewis acid sites associated to Zn2+ species in the hemimorphite [19]. In order to study in which reactions acetaldehyde is directly involved, a set of tests was carried out feeding pure acetaldehyde. The main product obtained was crotonaldehyde, plausibly formed by the aldol condensation of acetaldehyde on Hf sites in the Hf/SiO2 component [19]. However, almost no traces of 3-hydroxybutanal could be found among the products, probably as a consequence of its rapid dehydration to crotonaldehyde [25]. Carbon monoxide and methane were also observed in a 1:1 mole ratio, suggesting that they are formed by the decomposition of acetaldehyde. Heavy compounds (C6+) were also produced in significant amount, probably through the self- and cross-condensation of acetaldehyde and crotonaldehyde [25]. 5. Results and discussion 54 a) b) c) d) e) Figure 9. Ethanol to 1,3-butadiene main products yields as a function of ethanol conversion. a) 1,3-butadiene, acetaldehyde and crotonaldehyde; b) ethene and diethyl ether; c) butanol, butenes and butanal; d) propene, ethyl acetate, CO2 and acetone; e) heavy compounds. All tests were carried out at 360 °C and atmospheric pressure. 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 0 5 10 15 20 25 30 35 40 020 40 60 80 100 Crotonaldehyde yield (%) Yield (%) Ethanol conversion (%) Butadiene Acetaldehyde Crotonaldehyde 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 010 20 30 40 50 60 70 80 90 100 Yield (%) Ethanol conversion (%) Ethene Diethyl Ether 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 010 20 30 40 50 60 70 80 90 100 Yield (%) Ethanol conversion (%) Butanol Butenes Butanal 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 010 20 30 40 50 60 70 80 90 100 Yield (%) Ethanol conversion (%) Propene Ethyl acetate CO2 Acetone 0 5 10 15 20 25 010 20 30 40 50 60 70 80 90 100 Yield (%) Ethanol conversion (%) C6+ 5. Results and discussion 55 Also, an experiment feeding pure crotonaldehyde was performed. The results of this experiment revealed the predominant formation of heavy compounds, supporting their likely formation from crotonaldehyde via self-condensation reactions, however, no traces of 1,3-butadiene were found among the products. Conversely, an increase in 1,3-butadiene selectivity was observed in experiments cofeeding crotonaldehyde with ethanol pointing to a reaction between ethanol and crotonaldehyde to produce 1,3-butadiene [25]. This observation supports the pathway where crotyl alcohol was formed through the MPVO reduction of crotonaldehyde with ethanol and subsequently dehydrated to 1,3-butadiene [27,34,50–53]. Almost no crotyl alcohol was detected during the tests feeding only ethanol, probably due to its high reactivity. The formation of 1,3-butadiene by the dehydration of crotyl alcohol and the high rate of this reaction were confirmed by individually feeding crotyl alcohol. The conversion of crotyl alcohol (360°C and WHSV of 1.12 h-1) was full, with a selectivity of 60% to 1,3-butadiene [25]. Therefore, it can be concluded that, over the HfZn/SiO2 catalyst, both crotonaldehyde and crotyl alcohol are directly involved in the formation of 1,3-butadiene from ethanol. Indeed, the kinetic curve of 1,3-butadiene (Figure 9a) resembles that of a secondary stable product. Its stable character was confirmed by the absence of reaction in an additional experiment performed feeding only 1,3-butadiene. Similarly, no signs for the occurrence of reactions were evidenced when 1,3-butadiene was co-fed with ethanol [25]. 5. Results and discussion 56 From literature, ethene and diethyl ether are both direct products of ethanol dehydration but ethene can also be indirectly produced from diethyl ether or, maybe, both pathways can coexist [3]. According to the obtained kinetic curves of diethyl ether and ethene (Figure 9b), they both seem to be primary stable products probably coming from ethanol dehydration. However, as the kinetic curve of ethene rises exponentially, there seems to be another reaction, besides ethanol dehydration, which leads to ethene. Prins condensation route, where ethanol dehydrogenates to acetaldehyde and dehydrates to ethene and both products react to produce 1,3- butadiene, is ruled out over this catalyst as ethene is not an intermediate product. Several C4 compounds were also found amid the products such as butanol, butanal and butenes. The kinetic curve of 1-butanol (Figure 9c) clearly exhibits the characteristic shape of a secondary unstable product. Although the curve of butanal also resembles that of a secondary unstable product, this seems less clear than for 1- butanol, as a net decrease in butanal yield above a certain conversion cannot be unambiguously concluded. The fact that at low ethanol conversions (< 30%) 1- butanol is detected in the products, but not butanal, may be an indication of the dehydrogenation of 1-butanol as a plausible route leading to butanal. This assumption is compatible with the results obtained in an experiment feeding only 1- butanol, where butanal was the main product formed, especially at low 1-butanol conversions, followed by butenes. 5. Results and discussion 57 Since 1-butanal is formed from 1-butanol, crotonaldehyde hydrogenation should play a minor role as a source of butanal. Then, it seems that crotonaldehyde is mainly hydrogenated to crotyl alcohol by the MPVO reaction, which then can be further hydrogenated to 1-butanol. In fact, when only crotyl alcohol was fed, the lumped selectivity to 1-butanol and butenes was significant. This indicates that over silica doped metal catalysts, and in the absence of an external H2 source, 1-butanol likely comes from crotyl alcohol hydrogenation by H fragments present on the catalyst surface originating from previous alcohol dehydrogenations [34,38,50]. Finally, the butenes fraction comprising 1-butene, isobutene, and cis- and trans-2-butene shows the characteristic kinetic curve of a secondary stable product (Figure 9c) which, is in line with a route involving the dehydration of 1-butanol to 1- butene and its successive isomerization to cis- and trans-2-butene and isobutene [3]. The kinetic curve of ethyl acetate (Figure 9d) indicates that it is a secondary unstable product, probably formed from acetaldehyde by the Tishchenko reaction. Indeed, ethyl acetate was observed in the products of the experiment performed feeding only acetaldehyde. Sushkevich et al. [38] proposed that further transformation of ethyl acetate leads to acetic acid, which produces acetone releasing CO2 by decarboxylation. Finally, acetone can be reduced and dehydrated to propene. This reaction sequence is consistent with the shape and position of the kinetic curves of ethyl acetate, acetone, propene and CO2. Acetic acid was observed in trace amounts as it is rapidly transformed into acetone and CO2 is released. 5. Results and discussion 64 As acetaldehyde is formed by ethanol dehydrogenation and consumed in aldol condensation reactions, the increase in acetaldehyde yield points to a larger inhibition by water of aldol condensation reactions as compared to ethanol dehydrogenation. This is supported by the decrease in the 1,3-butadiene and heavy products yield, as observed in Figure 13c (left) and Figure 14a, respectively. Inhibition of aldol condensations by water is expected to occur by blocking of Lewis acid sites active for these reactions [57] which have been associated to Hf4+ species [19]. Differently, at high temperature (380 °C) and low space velocity, where the formation of heavy compounds (C6+) is significant for an anhydrous ethanol feedstock (Figure 14a, right), the presence of water results in an increase of the yield to 1,3-butadiene (Figure 13c, right). This fact indicates that, at 380 °C, formation of heavy products is inhibited to a larger degree than that of 1,3-butadiene. This could be explained considering that a reduction in the concentration of available active Hf4+ sites due to water adsorption will decrease the probability of occurrence of consecutive aldol condensation reactions through which heavy products are formed [58]. This effect can be beneficial at a high operating temperature, where acetaldehyde is so reactive that it is rapidly converted into heavy compounds unless water is present, allowing higher 1,3-butadiene yield with lower heavy compounds formation. 5. Results and discussion 65 a) 360 °C 380 °C b) 360 °C 380 °C c) 360 °C 380 °C Figure 13. Effect of water content in ethanol feed on a) ethanol conversion, b) acetaldehyde yield, and c) 1,3-butadiene yield at 360 °C (left panels) and 380 °C (right panels) as a function of space velocity and water content (wt%) in ethanol. 30 40 50 60 70 80 90 100 0 1 2 3 4 5 6 7 8 9 10 11 Ethanol conversion WHSV (h-1) 0 H2O 7.5 H2O 15 H2O 30 40 50 60 70 80 90 100 0 1 2 3 4 5 6 7 8 9 10 11 Ethanol convesrion (%) WHSV (h-1) 0 H2O 7.5 H2O 15 H2O 0 5 10 15 20 25 30 0 1 2 3 4 5 6 7 8 9 10 11 Acetaldehyde Yield (%) WHSV (h-1) 0 H2O 7.5 H2O 15 H2O 0 5 10 15 20 25 30 0 1 2 3 4 5 6 7 8 9 10 11 Acetaldehyde Yield (%) WHSV (h-1) 0 H2O 7.5 H2O 15 H2O 0 10 20 30 40 50 60 0 1 2 3 4 5 6 7 8 9 10 11 Butadiene Yield (%) WHSV (h-1) 0 H2O 7.5 H2O 15 H2O 0 10 20 30 40 50 60 0 1 2 3 4 5 6 7 8 9 10 11 Butadiene Yield (%) WHSV (h-1) 0 H2O 7.5 H2O 15 H2O 5. Results and discussion 66 a) 360 °C 380 °C b) 360 °C 380 °C Figure 14. Effect of water content in ethanol feed on a) heavy compound yield and b) ethene + diethyl ether yield at 360 °C (left panels) and 380 °C (right panels) as a function of space velocity and water content (wt%) in ethanol. On the other hand, the effect of water content in ethanol on the formation of butenes is, at any temperature, similar to that of 1,3-butadiene (not shown), an expectable result considering that self-condensation of acetaldehyde is also an intermediate reaction step in the formation of butenes [25]. 0 5 10 15 20 25 30 35 0 1 2 3 4 5 6 7 8 9 10 11 C6+ Yield (%) WHSV (h-1) 0 H2O 7.5 H2O 15 H2O 0 5 10 15 20 25 30 35 0 1 2 3 4 5 6 7 8 9 10 11 C6+ Yield (%) WHSV (h-1) 0 H2O 7.5 H2O 15 H2O 0 2 4 6 8 10 12 14 0 1 2 3 4 5 6 7 8 9 10 11 Ethene + DEE Yield (%) WHSV (h-1) 0 H2O 7.5 H2O 15 H2O 0 2 4 6 8 10 12 14 0 1 2 3 4 5 6 7 8 9 10 11 Ethene + DEE Yield (%) WHSV (h-1) 0 H2O 7.5 H2O 15 H2O 5. Results and discussion 67 Another observation is that, at any temperature, the presence of water increases the formation of direct ethanol dehydration products, ethene and diethyl ether (Figure 14b). This behaviour is due to an increase in the Brønsted acidity of the Hf-Zn catalyst by transformation of some of the Lewis acid sites into Brønsted acid sites by water chemisorption as demonstrated based on in situ IR spectroscopic experiments [20]. This alteration has been also spotted by Larina et al. [59] over Zn– La–Zr–Si oxide catalyst by situ FTIR studies and could contribute to the lower activity of the catalyst in the presence of water. Also, the inhibition of the ethanol dehydrogenation reaction results in a higher amount of ethanol available that reacts to produce ethene and diethyl ether. This last effect seems to be more noticeable as it can explain for itself the increased presence of this compounds when hydrous ethanol is fed. 5.2.5. Response surface analysis The effect of reaction conditions (temperature, content of water in the ethanol and WHSV) on ethanol conversion (X) and product selectivity (Si) and yield (Yi) over the mixed Hf-Zn catalyst was studied using a response surface analysis with the statistical software StatGraphics®. 5. Results and discussion 68 Table 3 shows the fitted parameters and the coefficients of determination (R2) of the regressions, which are relatively high (90-98%). The selectivity to 1,3- butadiene, acetaldehyde, butenes, and heavy products as well as the yield of 1,3- butadiene were better fitted when their natural logarithm was chosen as the response variable. Table 3. Estimated parameters for the response variables. Parameters X ln(YBD) ln(SBD) ln(SAC) SET SDEE ln(SC4) ln(SC6+) ao -207.108 -30.456 -18.715 6.036 - 190.691 16.224 -23.786 104.250 a1 0.838 0.194 0.136 -0.012 1.105 -0.042 0.151 -0.580 a2 -7.188 -0.475 -0.268 0.084 0.674 0.644 -0.326 -0.226 a3 -8.619 -1.300 -0.793 0.412 -2.751 -1.286 -1.030 -0.315 a4 - -0.0003 -0.0002 - -0.002 - -0.0002 0.0008 a5 0.016 0.0014 0.0008 - - -0.001 0.0009 - a6 - 0.0034 0.0022 - 0.005 0.004 0.0024 - a7 0.051 - -0.0010 -0.0024 -0.019 -0.007 0.009 a8 - -0.007 -0.0053 -0.0032 -0.025 -0.005 -0.0061 - a9 0.421 - - -0.0254 0.068 - 0.0093 0.019 R2 98.5 96.8 89.6 97.8 92.4 95.2 94.6 90.7 2 2 2 0 1 2 3 4 5 6 7 8 9 ˆ             y b bT bWc bWHSV b T b T Wc b T WHSV bWc bWc WHSV bWHSV Note: BD= butadiene; AC=Acetaldehyde; ET=ethene; DEE=diethyl ether; C4= butenes; C6+= heavy compounds 5. Results and discussion 69 Table 4. Prediction of response variables for the validation data set. Op. Cond. Value X (%) YBD (%) SBD (%) SAC (%) SET (%) SDEE (%) SC4 (%) SC6+ (%) T (°C) 340 Measured 41.56 15.76 37.93 36.25 3.33 1.64 2.81 3.85 W (%) 0 Estimated 40.76 15.82 36.36 33.98 2.53 1.68 2.62 4.81 WHSV (h-1) 6.1 Abs. Error 0.80 0.06 1.57 2.27 0.80 0.04 0.19 0.96 T (°C) 360 Measured 43.40 12.50 28.80 53.10 6.40 2.40 1.60 1.70 W (%) 15 Estimated 43.18 10.70 27.81 49.83 6.58 2.22 1.66 1.03 WHSV (h-1) 8 Abs. Error 0.22 1.80 0.99 3.27 0.18 0.18 0.06 0.67 BD= 1,3-butadiene; AC=Acetaldehyde; ET=ethene; DEE=diethyl ether; C4= butenes; C6+= heavy compounds Two tests at different temperature, water content, and space velocity were set aside to validate the statistic model. As shown in Table 4, the validation data set is reasonably well predicted, being the absolute errors small. Therefore, a good generalization capability of the model is expected within the employed range of reaction conditions. The response surface analysis allows distinguishing how reaction conditions impact variables of interest such as 1,3-butadiene yield and selectivity. For instance, when the water content in ethanol is 7.5%, the 1,3-butadiene yield is maximized at low space velocity and high operating temperature (Figure 15), which involves high ethanol conversion (Figure 16) and also high 1,3-butadiene selectivity (Figure 17). 5. Results and discussion 70 a) b) Figure 15. Parametric curves of the response surface analysis model for 1,3-butadiene yield: a) variation with temperature and WHSV, b) variation with water content and WHSV. Temperatures in °C, WHSV in h-1, and water content in wt%. 5. Results and discussion 71 a) b) Figure 16. Parametric curves of the response surface analysis model for ethanol conversion: a) variation with temperature and WHSV, b) variation with water content and WHSV. Temperatures in °C, WHSV in h-1, and water content in wt%. 5. Results and discussion 72 a) b) Figure 17. Parametric curves of the response surface analysis model for 1,3-butadiene selectivity: a) variation with temperature and WHSV, b) variation with water content and WHSV. Temperatures in °C, WHSV in h-1, and water content in wt%. 5. Results and discussion 73 Also, at 360 °C and relatively low space velocities (< 5 h-1), the 1,3-butadiene yield is almost insensitive to the water content in ethanol (Figure 15) since contour plots are flat. This is a consequence of the fact that, at 360 °C and fixed space velocity, as water content increases ethanol conversion decreases, but 1,3-butadiene selectivity increases, and both effects counterbalance. The statistical model was used to study which combination of operating variables maximized 1,3-butadiene selectivity or yield. As can be seen in Table 5, the highest 1,3-butadiene yield achievable within the bounds of the process variables studied is 50.62%, which involves operating at the highest temperature and lowest spatial velocity with the highest concentration of water in the feedstock (T= 380 °C, WHSV= 1.12 h-1 and 15 wt% water). This operating point seems to be the finest since the presence of water causes a drastic decrease in the formation of heavy compounds. The operating conditions for maximum 1,3-butadiene selectivity (T= 368 °C, WHSV= 1.12 h-1 and 15 wt% water) are close to that of the maximum 1,3-butadiene yield. In fact, similar 1,3-butadiene selectivity is achieved in both scenarios but, in this case, the 1,3-butadiene yield is lower. It should be noted that the maximum yield operating point might not be the best operating condition in economic terms, as many trade-offs in the design of the process are involved such as feedstock, reactor and separation costs. Thus, selection of the best operating point of the reactor should be done within the design of the whole process with the aid of the statistical model. 5. Results and discussion 80 After regeneration of the catalyst, the surface contents of Zn, Si, and Hf increased, approaching the values of the fresh catalyst. Alongside, the amount of carbon decreased to levels close to those in the fresh sample, supporting the removal of most of the coke from the catalyst after regeneration by calcination in air at 500 °C. Table 6. Atomic surface composition and metal surface ratios determined by XPS. Atomic surface concentration (%) Surface atomic ratios Catalyst Zn Hf Si C Zn/Hf Zn/Si Hf/Si HfZnfresh 0.50 0.20 32.0 24.3 2.6 0.015 0.006 HfZnused 0.25 0.17 19.3 40.0 1.4 0.013 0.009 HfZnregen 0.40 0.19 23.5 29.4 2.1 0.017 0.008 5. Results and discussion 81 5.2.7. Kinetic model building and validation The main reaction products observed in the catalytic tests were considered in the model: 1,3-butadiene, acetaldehyde, ethene, diethyl ether, and n-butanol. Besides, numerous minor compounds were grouped into three lumps, butenes, heavy compounds and oxygenated compounds, to simplify the model. 1-butene, cis- 2-butene, trans-2-butene and isobutene were lumped as butenes and 1-butene was selected as characteristic compound. Heavy compounds, with more than six carbons and mostly aromatic, were grouped into one lump. Diphenyl ketone was chosen as representative compound since it was identified between those molecules absorbed over the catalyst surface after reaction [25]. It also presents a molecular formula alike to the average one of all the products considered heavy compounds. Finally, the oxygenated compounds lump comprises other compounds with a molecular formula similar to butanal. It includes a lot of minority compounds of which the main ones are: acetone, ethyl acetate, butanal and 2-ethyl-butanol. The reaction network obtained in section 5.2.2 was used as a starting point to propose a set of reactions which could explain the formation of the selected reaction products. As crotonaldehyde and crotyl alcohol were presented in a really low concentration in the reactor outlet due to its quick transformation, acetaldehyde aldol-condensation to crotonaldehyde, crotonaldehyde reduction with ethanol, and crotyl alcohol dehydration have been unified in a global stage from ethanol to 1,3- butadiene. 5. Results and discussion 82 These simplifications resulted in the following set of reactions: ethanol dehydrogenation to acetaldehyde (Model reaction 1), ethanol and acetaldehyde reaction to 1,3-butadiene (Model reaction 2), ethanol dehydration to diethyl ether (Model reaction 3) and ethene (Model reaction 4), butanol production from ethanol condensation (Model reaction 5), butanol dehydration into butenes (Model reaction 6), generation of heavy compounds (Model reaction 7), which are believed to be formed by acetaldehyde aldol-condensation and further dehydrogenation and cyclization reactions [58], and oxygenated compounds production (Model reaction 8). According to the experimental results shown in 5.2.2, all reactions of the kinetic model were considered irreversible in the range of operating conditions studied. 2 5 2 4 2 C H OH C H O H Model reaction 1 2 5 2 4 4 6 2 2C H OH C H O C H H O   Model reaction 2   2 5 2 5 2 2 2 C H OH C H O H O Model reaction 3 2 5 2 4 2 C H OH C H H O Model reaction 4 2 5 4 9 2 2C H OH C H OH H O Model reaction 5 4 9 4 8 2 C H OH C H H O Model reaction 6 2 4 13 10 2 2 13 5 2 2 2 1 1 C H O C H O H O H   Model reaction 7 2 5 4 8 2 2 2 C H OH C H O H O H   Model reaction 8 5. Results and discussion 83 The effect of water over the catalyst performance has to be taken into account since, as commented on section 5.2.4, water inhibits aldol condensation by blocking of Lewis acid sites active for these reactions [57] which have been associated with Hf4+ species. Also, water blocks and/or transforms Zn2+ sites, where ethanol dehydrogenation undergoes. Consequently, the formation of 1,3-butadiene and most of the side-products will be hindered as the presence of water in the feed increases. In addition, water boosts the formation of ethene and diethyl ether due to a combined effect of (i) the increase in the Brønsted acidity of the catalyst by transformation of some of the Lewis acid sites into Brønsted acid sites by water chemisorption and (ii) the major presence of ethanol in the medium as a consequence of the inhibition of the ethanol dehydrogenation (section 5.2.4). That way, as commented in section 4.8, a power-law kinetics was assumed for each reaction and a correction term was added to consider the effect of the presence of water. Model equation 6 does not include the water corrective term as it is considered that the inhibition of butenes by water is a consequence of the inhibition of butanol (Model reaction 5), since butenes are formed from butanol dehydration (Figure 10). From Model reaction 1 to 8, i r is the reaction rate in mol/g h, i A is the kinetic constant at the central temperature in mol/g h barΣnki, R is the ideal gas constant in kJ/mol K, i Ea the activation energy in kJ/mol, T the temperature in Kelvin, k P the reactant partial pressure in bars, ki n the reaction orders, and i a (bar-1) and i m are fitting parameters. 5. Results and discussion 84   11 1 2 11 11 1 1 ref Ea n R T T EtOH m HO P r A e aP             Model equation 1   23 2 2 2 11 22 2 1 ref Ea n n R T T EtOH Ac m HO PP r A e aP              Model equation 2   34 3 2 11 33 3 1 ref Ea n R T T EtOH m HO P r A e aP             Model equation 3   45 4 2 11 44 4 1 ref Ea n R T T EtOH m HO P r A e aP             Model equation 4   56 5 2 11 55 5 1 ref Ea n R T T EtOH m HO P r A e aP             Model equation 5 6 7 11 66 ref Ea R T T n ButOH r A e P            Model equation 6   78 6 2 11 77 6 1 ref Ea n R T T Ac m HO P r A e aP             Model equation 7   89 7 2 11 88 7 1 ref Ea n R T T EtOH m HO P r A e aP             Model equation 8 5. Results and discussion 85 The parameters of the kinetic model were obtained by fitting data of catalyst performance from the design of experiments presented in section 5.2.3 and 5.2.4 (Table A4 and A5). Since the experiments were designed so the deactivation of the catalyst was as small as possible, the kinetic model is only valid for a fresh catalyst. The mole balance for each compound expressed as a function of the reaction rate is presented from Equation 12 to 22, being W the catalyst mass and Fk the mole flow of each compound k, where internal and external resistance to mass transfer is neglected. The different products considered were ethanol (EtOH), acetaldehyde (Ac), water (Wat), hydrogen (H), 1,3-butadiene (BD), diethyl ether (DEE), ethene (Et), butanol (ButOH), butenes (But), heavy compounds (HC) and oxygenated compounds (OC). Certainly, the external and internal diffusion effects were evaluated for the catalytic tests as described in section 4.8. The CMears obtained values were around 2·10−3, which is far from 0.15 and confirms that, for all the tests, no concentration gradient exists between the bulk gas and external surface of the catalyst particle (external diffusion negligible). Also, the calculated CWP values were around 1·10−5, which is far from 0.15, confirming that internal diffusion was negligible too. 5. Results and discussion 86   1 2 3 4 5 8 2 2 2 EtOH dF r r r r r r dW        Equation 12 1 2 7 13 2 Ac dF r r r dW         Equation 13        2 3 4 5 6 7 8 11 22 Wat dF r r r r r r r dW Equation 14    1 7 8 5 2 H dF r r r dW Equation 15 2 BD dF r dW  Equation 16 3 DEE dF r dW  Equation 17 4 Et dF r dW Equation 18 56 ButOH dF rr dW  Equation 19 6 But dF r dW Equation 20 7 HC dF r dW Equation 21 8 OC dF r dW Equation 22 5. Results and discussion 87 The kinetic parameters estimated with the rigorous method are shown in Table 7, along with their 95% confidence intervals. Some of the obtained reaction orders are unusually high, which might be a consequence of the simplifications made in the model building: (i) lumping of reactions and compounds; and (ii) the use of kinetic equations not derived from any reaction mechanism. Table 7. Calculated values of kinetic parameters. A (mol/g h barΣni) Ea (kJ/mol) n a (bar-1) m EtOH Ac ButOH r1 49.21±0.59 162.67±0.27 2.98±7.90·10-3 - - 511.06±15.65 0.32±1.60·10-3 r2 7.93±0.53 170.99±0.81 1.43±0.01 0.87±0.02 - 24.46±3.51 0.57±0.04 r3 0.68±0.23 144.91±5.82 3.51±0.18 - - - - r4 0.19±5.90·10-3 193.29±0.75 1.99±0.01 - - - - r5 4.82·10-3±9.44·10-5 197.91±0.94 0.20±0.01 - - 20.69±0.95 1.58±0.05 r6 1.83±0.23 230.78±1.30 - - 1.12±0.02 - - r7 1.24·10-2±1.91·10-4 69.23±4.95 - 0.63±0.01 - 3.73±1.38 3.23±1.00 r8 3.25·10-2±3.75·10-4 205.44±0.88 0.13±2.10·10-3 - - 124.12±1.36 1.14±4.70·10-3 The comparison between the prediction of the model and the experimental results can be observed in Figure 21, which represent the mole flows of each compound calculated by the model versus the experimental ones obtained in the catalytic tests. The model fitting is good, not only for the major compounds at the reactor outlet (ethanol, acetaldehyde, water, hydrogen and 1,3-butadiene), but also for most of the minor ones (ethene, diethyl ether, butanol and oxygenated compounds lump). Approximately 90% of the points lie within the ±20% error bands. 5. Results and discussion 88 Figure 21. Comparison plots between the experimental and model flow rates. (● 340°C, ▲360°C and ■ 380°C, empty symbols 0%w/w water, green 3.75%w/w water, magenta 7.5%w/w water and black 15%w/w water) 10% error band (black dash-dot line), 20% error band (red dotted line). 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0 0.02 0.04 0.06 0.08 Experimental (mol/h) Model (mol/h) Ethanol 0 0.005 0.01 0.015 0.02 0.025 0 0.005 0.01 0.015 0.02 0.025 Experimental (mol/h) Model (mol/h) Acetaldehyde 0 0.01 0.02 0.03 0.04 0 0.01 0.02 0.03 0.04 Experimental (mol/h) Model (mol/h) Hydrogen 0 0.005 0.01 0.015 0 0.005 0.01 0.015 Experimental (mol/h) Model (mol/h) 1,3-Butadiene 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 Experimental (mol/h) Model (mol/h) Water 0 1 2 3 4 5 x 10-4 0 1 2 3 4 5 x 10-4 Experimental (mol/h) Model (mol/h) Diethyl ether 0 0.5 1 1.5 2 2.5 3 3.5 4 x 10-3 0 1 2 3 4x 10-3 Experimental (mol/h) Model (mol/h) Ethene 0 1 2 3 4 5 6 7 x 10-4 0 1 2 3 4 5 6 7x 10-4 Experimental (mol/h) Model (mol/h) Butanol 5. Results and discussion 89 Figure 21 cont. Comparison plots between the experimental and model flow rates. (● 340°C, ▲360°C and ■ 380°C, empty symbols 0%w/w water, green 3.75%w/w water, magenta 7.5%w/w water and black 15%w/w water) 10% error band (black dash-dot line), 20% error band (red dotted line). In order to validate the assumptions of the regression model (that the errors are independent, follow a normal distribution and have a constant variance [64]), the residuals were analyzed. As the experiments were carried out independently, the independence of the errors is satisfied. Besides, the standardized residuals are normally distributed (Figure A1) as they fairly follow the normal-distribution line, so it can be considered that the normality hypothesis is also fulfilled. 0 0.2 0.4 0.6 0.8 1 1.2 x 10-3 0 0.2 0.4 0.6 0.8 1 1.2 x 10-3 Experimental (mol/h) Model (mol/h) Butenes 0 0.2 0.4 0.6 0.8 1 1.2 1.4 x 10-3 0 0.2 0.4 0.6 0.8 1 1.2 1.4x 10-3 Experimental (mol/h) Model (mol/h) Heavy compounds 0 1 2 3 4 5 x 10-3 0 1 2 3 4 5 x 10-3 Experimental (mol/h) Model (mol/h) Oxygenated compounds 5. Results and discussion 96 6. Conclusion and future work 97 6. CONCLUSIONS AND FUTURE WORK 1,3-butadiene is a key platform compound used as a chemical intermediate and as a monomer in the manufacture of polymers to produce synthetic rubber, resins, and elastomers. Nowadays, 1,3-butadiene is mostly produced by the petrochemical route as a by-product of the naphtha or gas oil steam cracking process. The production of 1,3-butadiene from bioethanol, which was abandoned in the 1960’s, rises as a sustainable and promising alternative in the framework of the new low-carbon policies due to the development of new catalysts whose performance exceed that of commercial catalysts. This thesis is dedicated to the study of the industrial aspects of those catalysts, such as the effect of the reaction conditions and the presence of impurities in the feed on their performance, and also their deactivation, in order to extract data for the industrial process design to support the evaluation of the sustainability and profitability of this technology. In particular, a highly selective Hf-Zn mixed catalyst was selected from the literature and its performance assessed with numerous catalytic tests to, finally, develop a kinetic model which may be used in the future to design and optimize an industrial process. 6. Conclusion and future work 98 The most likely ethanol to 1,3-butadiene pathway over the one-step catalyst includes: ethanol dehydrogenation to acetaldehyde, acetaldehyde aldolcondensation to crotonaldehyde, crotonaldehyde reduction with ethanol, and crotyl alcohol dehydration, which concurs with the most accepted route in literature. Regarding side products, ethene and diethyl ether seem to be primary stable products formed from ethanol dehydration. The decarbonylation of aldehydes has been observed over the catalyst: acetaldehyde to methane and crotonaldehyde to propene, with the associated release of carbon monoxide. Ethyl acetate is likely produced via the Tishchenko reaction and then further transformed into acetone and propene, which releases carbon dioxide. 1-Butanol seems to be formed from crotyl alcohol hydrogenation with surface H fragments originating from previous dehydrogenation reactions. 1- butanol dehydrogenation and dehydration are the main sources of butanal and 1- butene, respectively. 1-Butene can be isomerized into cis- and trans-2-butenes. The source of heavy compounds appears to be the self- and cross-condensation of C4+ aldehydes, like butanal and crotonaldehyde; and ketones, like acetone. 6. Conclusion and future work 99 Focusing on the impurities that may be recycled along with the nonconverted ethanol, the recycle of butanal and acetone should be avoided, as they are finally converted to heavy compounds by aldol condensation reactions. Heavy product formation leads to the deactivation of the catalyst, the fouling of downstream equipment and also makes it difficult to recover and recycle ethanol. Also, recycling 1-butanol should be averted, as its dehydrogenation leads to butanal and then to heavy products while 1-butanol dehydration leads to butenes, which are difficult to separate from 1,3-butadiene due to their similar boiling points, shape and polarity. Finally, the recycling of a considerable amount of diethyl ether could suppress unwanted diethyl ether formation from ethanol. The economic viability of this measure will depend on the cost of separating and recycling diethyl ether and the increase in 1,3-butadiene selectivity due to the suppression of diethyl ether formation. According to our results, this will not be practical since the needed amount of recycled diethyl ether would be quite high. A power-law kinetic model was developed with a corrective term for water effect in order to predict the formation of the major products: acetaldehyde, 1,3- butadiene, hydrogen, water, ethene, diethyl ether, butanol, and the lumps butenes, diphenyl ketone and butanal. The predictions of the model agree with the experimental validation data, so the model can be used to predict the rates of formation of the different products, when water is present in the ethanol feedstock. 6. Conclusion and future work 100 The presence of water in the ethanol feed enhances the selectivity to dehydration products, i.e., ethene and diethyl ether from ethanol and decreases ethanol conversion. The increased dehydration activity can be accounted for by the generation of new Brønsted acid sites of medium strength in the mixed Hf-Zn catalyst, as assessed by in situ IR spectroscopy, probably by reaction of Zn2+-related Lewis acid sites (active for ethanol dehydrogenation) with water at reaction conditions. Besides, blocking of Zn2+ sites by water represents a loss of dehydrogenation sites resulting in a decrease in ethanol conversion and letting more ethanol available for its dehydration to ethene and diethyl ether. Indeed, the developed kinetic model explain the rise in the ethene and diethyl ether formation in the presence of water as a result of the inhibition of ethanol dehydrogenation so, although there is a transformation of the actives sites, its influence appears to be low. Moreover, water seems to hinder aldol condensation reactions to a greater degree than dehydrogenation due to blocking of Hf4+-related Lewis acid sites. This effect can be beneficial at a high operating temperature, where acetaldehyde is so reactive that it is rapidly converted into heavy compounds unless water is present, allowing higher 1,3-butadiene yield with lower heavy compounds formation. 6. Conclusion and future work 101 In addition, the presence of water significantly decreases the rate of catalyst deactivation, which can be ascribed to the inhibition of successive condensation and dehydrogenation reactions involved in coke formation. Catalyst deactivation is quite noticeable and mostly caused by the formation and retention of aromatic-type species preferentially on the dehydrogenating Zn2+ sites of the hemimorphite phase of the catalyst and by the reduction of some of these Zn2+ sites to Zn0. The loss in Zn2+ sites induces an imbalance between Hf4+ and Zn2+ sites which results in a change in catalyst selectivity. Regeneration by calcination with air successfully removes coke and re-oxidizes a fraction of Zn0 back to Zn2+, but it does not completely re-establish the Zn2+/Hf4+ site balance as the re-oxidized Zn species are segregated from the original hemimorphite phase and provide a less favourable interaction with Hf sites. As a general conclusion, the results presented in this work may be extrapolated for most of the ethanol to 1,3-butadiene one-step catalysts. In the first place, the elucidated pathway concurs with most of the one-step catalyst presented in the literature (i.e. MgO-SiO2, ZnO-Al2O3 and Ag-ZrO2-SiO2). Besides, in the development of one-step processes the effect of co-feeding water should be taken into account since it assess the option of using a cheaper ethanol feedstock instead of anhydrous grade, and to decide to what degree water should be removed from unconverted ethanol in the conceptual design of the process in order to find a tradeoff between reactor performance and separation costs. Finally, the methodology used to assess the kinetic model can be applied to any ethanol to 1,3-butadiene catalyst. 6. Conclusion and future work 102 The results of this thesis suggest the following future work:  The study of the reaction pathway, although quite complete, could be improved by conducting more feeding and co-feeding tests at different spatial velocities with other intermediate compounds whose origin is not clear.  Since water inhibit large-chain compounds formation and deposition over the catalyst surface, limiting the catalyst deactivation, the design of experiments could be extended at higher temperatures to acquire a wider view of the effect of water over the catalyst performance.  As the catalyst loses active centres in each regeneration, a deeper study including more regeneration cycles will be necessary to draw stronger conclusions for its industrial application suitability.  There is also room for improvement in the kinetic model development. In the first place, to include new reaction products, especially those which can affect the separation process downstream the reactor or affect the catalyst performance when recirculated with the ethanol. This will improve the robustness of the model. For example, it is known that acetone and butanal affect negatively the catalyst performance, boosting the heavy compounds generation. Besides, it was found that when co-feeding diethyl ether, its generation was impaired so ethanol dehydration to diethyl ether may be taken as reversible. 6. Conclusion and future work 103  Also, butanal, butenes and diphenyl ketone lumps fitting may be improved by separating the main compounds they contain, adding new generation reactions for them and finding their kinetic parameters.  Besides, the accuracy of the kinetic model should be tested on a higher scale in order to check if it is really useful for the design of an industrial process.  The proposed kinetic model does not cover the catalyst deactivation, which is quite relevant as commented on section 5.2.6. Including the deactivation effect in the model would lead to a completer and more useful model when designing an industrial process.  It will be also interesting to study the effect of reactor configurations and locations and distribution of ethanol feed along the reactor using the developed kinetic model.  Finally, different scenarios of the industrial process should be proposed and analysed from an economic and environmental point of view in order to globally optimize the 1,3-butadiene production. 6. Conclusion and future work 104 References 105 REFERENCES [1] IPTS, Best Available Techniques (BAT) Reference Document for the Production of Wood-based, 2016. doi:10.2791/21807. [2] A.J. Nizamoff, On-Purpose Butadiene On-Purpose Butadiene, (2013). [3] E. V. Makshina, M. Dusselier, W. Janssens, J. Degrève, P.A. Jacobs, B.F. Sels, Review of old chemistry and new catalytic advances in the on-purpose synthesis of butadiene, Chem. Soc. Rev. 43 (2014) 7917–7953. doi:10.1039/C4CS00105B. [4] Mordor Intelligence, Butadiene Market. Growth, Trends, and Forecast., (2019). [5] M. Reporting, Argus Ethylene Annual 2018 Table of Contents Events, 2018 (2018). [6] D. Cespi, F. Passarini, I. Vassura, F. Cavani, Butadiene from biomass, a life cycle perspective to address sustainability in the chemical industry, Green Chem. 18 (2016) 1625–1638. doi:10.1039/C5GC02148K. [7] S. Farzad, M.A. Mandegari, J.F. Görgens, Integrated techno-economic and environmental analysis of butadiene production from biomass, Bioresour. Technol. 239 (2017) 37–48. doi:10.1016/j.biortech.2017.04.130. [8] European Commission, Going climate-neutral by 2050, Facilities. (2018). [9] European Commission, A clean planet for all. A European strategic longterm vision for a prosperous, modern, competitive and climate neutral economy., (2018). [10] R.E.N. Members, Renewables 2019 Global Status Report, 2019. https://wedocs.unep.org/bitstream/handle/20.500.11822/28496/REN2019.pdf ?sequence=1&isAllowed=y%0Ahttp://www.ren21.net/cities/wpcontent/uploads/2019/05/REC-GSR-Low-Res.pdf. References 112 [63] K.D. Jung, O.S. Joo, S.H. Han, S.J. Uhm, I.J. Chung, Deactivation of Cu/ZnO catalyst during dehydrogenation of methanol, Catal. Letters. 35 (1995) 303– 311. doi:10.1007/BF00807187. [64] N.R. Draper, H. Smith, Applied regression analysis, John Wiley & Sons, 1998. [65] H. Niiyama, S. Morii, E. Echigoya, Butadiene Formation from Ethanol over Silica-Magnesia Catalysts, Bull. Chem. Soc. Jpn. 45 (1972) 655–659. doi:10.1246/bcsj.45.655. [66] O. V. Larina, P.I. Kyriienko, S.O. Soloviev, Ethanol Conversion to 1,3- Butadiene on ZnO/MgO-SiO2 Catalysts: Effect of ZnO Content and MgO:SiO2 Ratio, Catal. Letters. 145 (2015) 1162–1168. doi:10.1007/s10562- 015-1509-4. [67] D. Dussol, N. Cadran, N. Laloue, L. Renaudot, J.M. Schweitzer, New insights of butadiene production from ethanol: Elucidation of concurrent reaction pathways and kinetic study, Chem. Eng. J. (2019). doi:10.1016/j.cej.2019.123586. [68] M. Legendre, D. Cornet, Catalytic oxidation of ethanol over tantalum oxide, J. Catal. (1972). doi:10.1016/0021-9517(72)90218-7. Appendix 113 Appendix Table A1. Catalytic tests (from 1 to 9) used to draw the kinetic curves. Ethanol conversion and carbon-based yield to products (%). All tests were performed at 360°C and an ethanol partial pressure of 0.21 bar. Only ethanol is fed. Test 1 2 3 4 5 6 7 8 9 SDmax Catalyst (g) 1 1 1 1 1 1 0.5 0.5 0.5 WHSV (h-1) 1.12 1.87 3.73 1.87 1.12 1.87 5.6 11.2 5.6 Ethanol conversion 87.0 75.7 67.3 82.2 87.2 74.9 58.8 49.6 64.8 1.6 Yield 1,3-Butadiene 36.9 33.1 28.0 36.21 37.45 29.86 25.23 14.16 20.92 1.60 Acetaldehyde 6.41 11.34 16.01 12.40 5.86 9.30 17.90 21.77 17.99 1.70 Crotonaldehyde (cis+trans) 0.13 0.12 0.20 0.16 0.04 0.05 0.32 0.50 0.40 0.10 3-Hidroxy Butanal 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Crotyl alcohol 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Ethene 3.91 2.90 2.10 3.04 3.58 2.33 1.64 1.12 1.51 0.20 Diethyl ether 1.11 1.14 1.01 1.01 0.86 0.85 0.59 0.53 0.62 0.10 Ethyl acetate 0.50 0.33 0.42 0.31 0.38 0.41 0.34 0.20 0.30 0.08 Propene 1.84 1.32 0.94 1.46 1.72 1.11 0.79 0.35 0.69 0.10 Propanal 2.20 2.10 1.92 2.36 2.14 2.03 1.69 0.92 1.49 0.10 Butenes 3.81 3.01 2.31 3.44 3.56 2.51 1.83 0.98 1.46 0.16 Butanol 0.43 0.73 1.04 0.67 0.44 0.61 1.28 1.01 1.06 0.10 Butanal 0.60 1.02 0.81 0.78 0.50 0.82 0.73 0.60 0.79 0.15 C6+ 18.60 12.31 8.13 14.33 21.01 15.84 6.09 3.45 5.51 2.20 Hexene 0.33 0.13 0.23 0.26 0.25 0.21 0.16 0.07 0.14 0.02 Hexane 0.44 0.30 0.23 0.33 0.37 0.30 0.27 0.15 0.23 0.02 Hexanol 0.28 0.22 0.20 0.25 0.29 0.23 0.17 0.09 0.14 0.04 Methanol 0.62 0.61 0.43 0.55 0.65 0.53 0.46 0.25 0.37 0.08 Isopropanol 0.01 0.11 0.02 0.05 0.15 0.13 0.07 0.00 0.00 0.02 Acetone 0.41 0.53 0.70 0.41 0.36 0.67 0.50 0.31 0.44 0.15 CO2 0.20 0.10 0.04 0.12 0.19 0.10 0.17 0.03 0.07 0.05 CO 0.03 0.02 0.01 0.02 0.02 0.01 0.01 0.01 0.01 0.00 CH4 0.09 0.06 0.05 0.06 0.06 0.04 0.05 0.04 0.05 0.01 Carbon balance error < 10% SDmax = maximum standard deviation among all catalytic tests Appendix 114 Table A2. Catalytic tests (from 10 to 18) used to draw the kinetic curves. Ethanol conversion and carbon-based yield to products (%). All tests were performed at 360°C and an ethanol partial pressure of 0.21 bar. Only ethanol is fed. Test 10 11 12 13 14 15 16 17 18 SDmax Catalyst (g) 0.5 0.5 0.5 0.5 0.1 0.1 0.1 2 2 WHSV (h-1) 11.2 5.6 3.73 7.0 11.2 30 50 1.12 1.12 Ethanol conversion 49.9 60.6 62.5 56.1 43.7 30.6 22.6 84.4 80.5 1.6 Yield 1,3-Butadiene 14.57 22.29 22.14 16.92 11.41 5.45 2.09 37.17 34.94 1.60 Acetaldehyde 21.54 18.68 16.32 19.40 17.97 16.91 16.69 6.75 6.80 1.70 Crotonaldehyde (cis+trans) 0.40 0.28 0.29 0.39 0.46 0.47 0.36 0.03 0.09 0.10 3-Hidroxy Butanal 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Crotyl alcohol 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Ethene 1.11 1.60 1.56 1.21 1.00 0.86 0.46 3.55 2.81 0.20 Diethyl ether 0.57 0.64 0.70 0.55 0.33 0.23 0.11 1.13 1.09 0.10 Ethyl acetate 0.22 0.29 0.38 0.24 0.10 0.55 0.00 0.43 0.44 0.08 Propene 0.22 0.64 0.74 0.54 0.43 0.26 0.10 1.58 1.33 0.10 Propanal 0.95 1.53 1.72 1.18 0.92 0.34 0.11 2.11 2.10 0.10 Butenes 1.21 1.63 1.62 1.41 1.00 0.87 0.49 3.46 3.01 0.16 Butanol 1.17 1.21 1.12 1.26 0.89 0.63 0.29 0.41 0.55 0.10 Butanal 0.68 0.76 0.89 0.72 0.41 0.31 0.15 0.54 0.83 0.15 C6+ 3.59 5.14 8.18 7.69 4.92 1.57 0.54 18.28 19.57 2.20 Hexene 0.07 0.15 0.16 0.10 0.08 0.01 0.00 0.26 0.24 0.02 Hexane 0.16 0.24 0.24 0.19 0.15 0.06 0.02 0.33 0.29 0.02 Hexanol 0.07 0.12 0.21 0.15 0.12 0.04 0.04 0.16 0.21 0.04 Methanol 0.26 0.42 0.42 0.29 0.23 0.20 0.27 0.43 0.64 0.08 Isopropanol 0.06 0.07 0.10 0.05 0.04 0.04 0.01 0.12 0.12 0.02 Acetone 0.27 0.41 0.44 0.48 0.22 0.09 0.05 0.35 0.32 0.15 CO2 0.08 0.03 0.12 0.01 0.01 0.02 0.01 0.15 0.13 0.05 CO 0.00 0.01 0.02 0.01 0.00 0.03 0.02 0.00 0.00 0.00 CH4 0.04 0.04 0.06 0.05 0.04 0.11 0.06 0.03 0.04 0.01 Carbon balance error < 10% SDmax = maximum standard deviation among all catalytic tests Appendix 115 Table A3. Experimental results of the conversion of ethanol to 1,3-butadiene over the one-step mixed Hf-Zn catalyst at different reaction temperatures (T), space velocities (WHSV), and water contents in the ethanol feed. Note: TOS= time on stream, X= ethanol conversion, Y= yield, BD= butadiene; ET=ethene; DEE=diethyl ether; C4= butenes; AC=Acetaldehyde; C6+= heavy compounds. “Others” comprises a mixture of minor subproducts such as acetone, ethyl acetate, butanal, butanol, 2-ethyl-hexenal, CO, CO2 and CH4. T Water WHSV TOS X YBD Selectivity (%) (°C) (wt%) (h-1) (h) (%) (%) BD ET DEE C4 AC C6+ Others 340 0 1.12 19 69.2 28.5 41.2 4.6 1.6 4.6 10.8 13.6 23.5 0 3.2 15 52.9 23.3 44.1 3.8 1.6 3.8 23.2 6.9 16.6 0 6.1 10 41.6 15.7 37.9 3.3 1.6 2.8 36.2 3.9 14.2 0 9.8 5 32.9 10.7 32.8 3.3 1.7 2.3 43.6 4.1 12.1 7.5 1.12 19 59.9 29.8 49.8 9.0 3.5 4.9 19.9 2.9 9.9 7.5 3.2 15 45.8 17.7 38.7 6.3 2.8 2.8 40.0 2.6 6.8 7.5 6.1 10 33.4 9.1 27.3 5.3 2.6 1.7 55.7 1.8 5.5 7.5 8.0 4.5 25.4 5.8 23.1 5.1 2.5 1.4 61.5 1.6 4.7 15 1.12 16 54.9 25.5 46.6 10.5 4.0 4.3 24.2 2.3 8.1 15 3.2 11 35.2 11.7 33.5 7.4 3.3 2.2 47.2 1.2 5.2 15 6.1 7 29.2 6.5 22.3 6.2 2.9 1.2 62.5 0.9 3.8 15 8.0 4 24.6 4.9 20.0 6.0 2.8 1.2 65.3 1.1 3.6 360 0 1.12 22 87.1 39.3 45.1 5.6 1.1 5.6 9.6 13.3 19.6 0 3.2 16 70.9 33.1 46.7 4.2 1.2 4.2 22.9 5.7 14.8 0 6.1 11 60.4 25.5 42.3 3.7 1.2 3.5 28.1 5.6 15.5 0 9.8 7 51.4 20.0 38.9 3.5 1.2 2.9 33.8 4.9 14.7 3.75 1.12 23 82.2 41.2 50.2 9.2 2.2 5.3 15.5 3.3 14.4 3.75 3.2 19 68.5 32.1 46.9 7.6 2.2 4.1 24.7 2.8 11.6 3.75 6.1 12 48.4 18.9 39.3 5.8 2.1 2.6 3.0 2.3 45.0 3.75 8.0 7 44.1 16.1 36.4 5.5 2.1 2.3 44.2 1.8 7.7 7.5 1.12 26 78.0 38.3 49.1 9.5 2.3 5.3 16.6 3.9 13.2 7.5 3.2 18 59.9 26.0 43.5 6.8 2.1 3.5 32.4 2.3 9.4 7.5 6.1 13 48.6 17.6 36.3 5.6 2.0 2.5 43.5 1.8 8.3 7.5 8.0 8 43.7 15.4 35.4 5.3 1.9 2.3 44.9 2.1 8.1 15 1.12 27 79.1 37.4 47.3 11.4 2.6 5.1 18.4 3.3 11.9 15 3.2 21 59.4 24.0 40.5 8.4 2.5 3.1 35.8 1.9 7.8 15 6.1 14 45.8 14.4 31.5 6.9 2.5 2.0 49.8 1.2 6.1 15 8.0 7 43.4 12.5 28.8 6.4 2.4 1.6 53.1 1.7 5.9 380 0 1.12 23 95.2 27.4 28.7 3.4 0.5 2.9 9.2 34.2 23.0 0 3.2 18 90.0 30.8 34.2 3.5 0.5 3.4 15.3 21.9 21.1 0 6.1 12 77.4 32.3 41.8 4.3 0.9 3.8 24.6 7.9 16.6 0 9.8 7 66.7 28.6 42.9 4.5 1.1 3.6 28.3 5.7 13.9 7.5 1.12 22 96.5 48.0 49.7 8.7 1.3 5.4 11.6 8.4 14.8 7.5 3.2 18 85.0 39.3 46.2 7.3 1.3 4.1 26.2 2.7 12.2 7.5 6.1 12 72.3 30.7 42.4 6.4 1.6 3.2 34.2 2.2 10.0 7.5 8.0 6 65.0 26.6 40.9 6.1 1.7 2.8 36.5 2.3 9.7 15 1.12 18 98.1 45.4 46.3 10.3 1.4 5.1 15.4 6.9 14.6 15 3.2 14 84.4 38.8 45.9 8.6 1.5 3.8 27.1 2.5 10.5 15 6.1 8 69.0 27.4 39.7 7.1 1.7 2.8 37.9 2.5 8.3 15 8.0 5 64.7 25.5 39.4 6.9 1.8 2.5 38.7 2.5 8.2 Appendix 116 Table A4. Operating condition and inlet flows of the tests gathered in Table A3. Test Catalyst load (g) WHSV (h-1) T (ºC) Mole flow (mol/h) Ethanol Nitrogen Water 1 1.0 1.12 360 0.0243 0.0911 - 2 1.0 1.87 360 0.0407 0.1527 - 3 1.0 3.73 360 0.0811 0.3027 - 4 1.0 1.87 360 0.0407 0.1527 - 5 1.0 1.12 360 0.0243 0.0911 - 6 1.0 1.87 360 0.0407 0.1527 - 7 0.5 5.60 360 0.0609 0.2277 - 8 0.5 11.2 360 0.1217 0.4554 - 9 0.5 5.60 360 0.0609 0.2277 - 10 0.5 11.2 360 0.1217 0.4554 - 11 0.5 5.60 360 0.0609 0.2277 - 12 0.5 3.73 360 0.0405 0.1527 - 13 0.5 7.00 360 0.0761 0.2839 - 14 0.1 11.2 360 0.0243 0.0911 - 15 0.1 30.0 360 0.0652 0.2438 - 16 0.1 50.0 360 0.1087 0.4071 - 17 2.0 1.12 360 0.0487 0.1821 - 18 2.0 1.12 360 0.0487 0.1821 - 19 0.5 9.80 340 0.1065 0.3964 - 20 0.5 6.10 340 0.0663 0.2464 - 21 0.5 3.20 340 0.0348 0.1286 - 22 0.5 1.12 340 0.0122 0.0455 - 23 0.5 9.80 360 0.1065 0.3964 - 24 0.5 6.10 360 0.0663 0.2464 - 25 0.5 3.20 360 0.0348 0.1286 - 26 0.5 1.12 360 0.0122 0.0455 - 27 0.5 9.80 380 0.1065 0.3964 - 28 0.5 6.10 380 0.0663 0.2464 - 29 0.5 3.20 380 0.0348 0.1286 - 30 0.5 1.12 380 0.0122 0.0455 - 31 1.0 8.00 340 0.0870 0.3054 0.0183 32 1.0 6.10 340 0.0663 0.2330 0.0139 33 1.0 3.20 340 0.0348 0.1232 0.0072 34 1.0 1.12 340 0.0122 0.0429 0.0028 35 1.0 8.00 340 0.0870 0.2866 0.0394 36 1.0 6.10 340 0.0663 0.2196 0.0300 37 0.5 3.20 340 0.0348 0.1152 0.0156 38 0.5 1.12 340 0.0122 0.0402 0.0056 39 0.5 8.00 360 0.0870 0.3080 0.0020 40 0.5 6.10 360 0.0663 0.2357 0.0016 41 0.5 3.20 360 0.0348 0.1232 0.0008 42 0.5 1.12 360 0.0122 0.0429 0.0003 43 0.5 8.00 360 0.0870 0.3054 0.0183 44 0.1 6.10 360 0.0663 0.2330 0.0139 45 0.1 3.20 360 0.0348 0.1232 0.0072 46 0.1 1.12 360 0.0122 0.0429 0.0028 47 2.0 8.00 360 0.0870 0.2866 0.0394 48 2.0 6.10 360 0.0663 0.2196 0.0300 49 0.5 3.20 360 0.0348 0.1152 0.0156 50 0.5 1.12 360 0.0122 0.0402 0.0056 51 0.5 8.00 380 0.0870 0.3054 0.0183 52 0.5 6.10 380 0.0663 0.2330 0.0139 53 0.5 3.20 380 0.0348 0.1232 0.0072 54 0.5 1.12 380 0.0122 0.0429 0.0028 55 0.5 8.00 380 0.0870 0.2866 0.0394 56 0.5 6.10 380 0.0663 0.2196 0.0300 57 0.5 3.20 380 0.0348 0.1152 0.0156 58 0.5 1.12 380 0.0122 0.0402 0.0056 Appendix 117 Table A5. Mole flows at the reactor outlet of the tests gathered in Table A3. BD=1,3-butadiene, Ac=acetaldehyde, C4=butenes, ButOH=butanol, Et=ethene, C6+=heavy compounds, DEE=diethyl ether, ButA=butanal, EtOH=ethanol. Test Mole flow (mol/h) BD Ac C4 ButOH Et C6+ DEE ButA EtOH H2O H2 1 0.0045 0.0016 0.0005 0.0000 0.0009 0.0007 0.0001 0.0019 0.0032 0.0168 0.0143 2 0.0067 0.0046 0.0006 0.0001 0.0012 0.0008 0.0002 0.0023 0.0099 0.0228 0.0206 3 0.0114 0.0130 0.0009 0.0004 0.0017 0.0010 0.0004 0.0035 0.0265 0.0362 0.0370 4 0.0074 0.0051 0.0007 0.0001 0.0012 0.0009 0.0002 0.0023 0.0073 0.0248 0.0226 5 0.0046 0.0014 0.0004 0.0001 0.0009 0.0008 0.0001 0.0018 0.0031 0.0171 0.0148 6 0.0061 0.0038 0.0005 0.0001 0.0009 0.0010 0.0002 0.0028 0.0102 0.0226 0.0215 7 0.0064 0.0109 0.0004 0.0003 0.0009 0.0005 0.0002 0.0030 0.0250 0.0209 0.0249 8 0.0086 0.0265 0.0006 0.0006 0.0014 0.0006 0.0003 0.0040 0.0614 0.0283 0.0449 9 0.0077 0.0109 0.0006 0.0004 0.0010 0.0006 0.0002 0.0031 0.0214 0.0243 0.0268 10 0.0089 0.0262 0.0007 0.0007 0.0014 0.0007 0.0003 0.0038 0.0609 0.0291 0.0449 11 0.0068 0.0114 0.0005 0.0004 0.0010 0.0005 0.0002 0.0029 0.0240 0.0216 0.0254 12 0.0045 0.0066 0.0003 0.0002 0.0006 0.0005 0.0001 0.0022 0.0152 0.0156 0.0179 13 0.0064 0.0148 0.0005 0.0005 0.0009 0.0009 0.0002 0.0029 0.0334 0.0234 0.0322 14 0.0014 0.0044 0.0001 0.0001 0.0002 0.0002 0.0000 0.0008 0.0137 0.0052 0.0082 15 0.0018 0.0110 0.0003 0.0002 0.0006 0.0002 0.0001 0.0013 0.0453 0.0072 0.0156 16 0.0011 0.0181 0.0003 0.0002 0.0005 0.0001 0.0001 0.0011 0.0841 0.0051 0.0212 17 0.0090 0.0033 0.0008 0.0001 0.0017 0.0014 0.0003 0.0033 0.0076 0.0327 0.0280 18 0.0085 0.0033 0.0007 0.0001 0.0014 0.0015 0.0003 0.0029 0.0095 0.0312 0.0279 19 0.0057 0.0153 0.0004 0.0004 0.0012 0.0002 0.0003 0.0017 0.0715 0.0171 0.0248 20 0.0052 0.0100 0.0004 0.0003 0.0009 0.0002 0.0002 0.0016 0.0387 0.0152 0.0183 21 0.0041 0.0043 0.0003 0.0002 0.0007 0.0002 0.0002 0.0014 0.0164 0.0123 0.0114 22 0.0017 0.0009 0.0002 0.0000 0.0004 0.0002 0.0001 0.0010 0.0037 0.0063 0.0052 23 0.0107 0.0185 0.0008 0.0005 0.0019 0.0004 0.0003 0.0035 0.0518 0.0315 0.0364 24 0.0085 0.0112 0.0007 0.0003 0.0015 0.0003 0.0002 0.0028 0.0263 0.0251 0.0256 25 0.0058 0.0057 0.0005 0.0001 0.0011 0.0002 0.0002 0.0017 0.0101 0.0168 0.0151 26 0.0024 0.0010 0.0003 0.0000 0.0006 0.0002 0.0001 0.0010 0.0016 0.0083 0.0064 27 0.0152 0.0201 0.0013 0.0004 0.0032 0.0006 0.0004 0.0046 0.0355 0.0450 0.0455 28 0.0107 0.0126 0.0010 0.0002 0.0022 0.0006 0.0002 0.0041 0.0150 0.0336 0.0330 29 0.0054 0.0048 0.0005 0.0001 0.0011 0.0011 0.0001 0.0032 0.0035 0.0221 0.0229 30 0.0015 0.0011 0.0002 0.0000 0.0004 0.0006 0.0000 0.0013 0.0006 0.0086 0.0094 31 0.0026 0.0136 0.0002 0.0001 0.0011 0.0001 0.0003 0.0004 0.0649 0.0260 0.0170 32 0.0030 0.0123 0.0002 0.0001 0.0012 0.0001 0.0003 0.0005 0.0442 0.0227 0.0164 33 0.0031 0.0064 0.0002 0.0001 0.0010 0.0001 0.0002 0.0004 0.0189 0.0159 0.0105 34 0.0018 0.0014 0.0002 0.0000 0.0007 0.0000 0.0001 0.0003 0.0049 0.0081 0.0039 35 0.0021 0.0139 0.0001 0.0001 0.0013 0.0000 0.0003 0.0003 0.0656 0.0461 0.0167 36 0.0022 0.0121 0.0001 0.0001 0.0012 0.0000 0.0003 0.0003 0.0469 0.0366 0.0148 37 0.0021 0.0058 0.0001 0.0001 0.0009 0.0000 0.0002 0.0002 0.0226 0.0215 0.0083 38 0.0016 0.0016 0.0001 0.0000 0.0007 0.0000 0.0001 0.0002 0.0055 0.0102 0.0036 39 0.0070 0.0169 0.0004 0.0002 0.0021 0.0001 0.0004 0.0012 0.0486 0.0214 0.0261 40 0.0063 0.0128 0.0004 0.0002 0.0019 0.0001 0.0003 0.0011 0.0342 0.0191 0.0212 41 0.0056 0.0059 0.0005 0.0001 0.0018 0.0001 0.0003 0.0013 0.0109 0.0170 0.0137 42 0.0025 0.0015 0.0003 0.0000 0.0009 0.0001 0.0001 0.0007 0.0022 0.0079 0.0052 43 0.0067 0.0171 0.0004 0.0002 0.0020 0.0001 0.0004 0.0014 0.0489 0.0372 0.0263 44 0.0058 0.0140 0.0004 0.0002 0.0018 0.0001 0.0003 0.0012 0.0341 0.0303 0.0218 45 0.0045 0.0067 0.0004 0.0001 0.0014 0.0001 0.0002 0.0009 0.0139 0.0200 0.0128 46 0.0023 0.0016 0.0003 0.0000 0.0009 0.0001 0.0001 0.0006 0.0027 0.0099 0.0050 47 0.0054 0.0200 0.0003 0.0002 0.0024 0.0001 0.0005 0.0009 0.0493 0.0554 0.0273 48 0.0048 0.0151 0.0003 0.0002 0.0021 0.0001 0.0004 0.0008 0.0359 0.0439 0.0212 49 0.0042 0.0074 0.0003 0.0001 0.0017 0.0001 0.0003 0.0007 0.0141 0.0277 0.0128 50 0.0023 0.0018 0.0002 0.0000 0.0011 0.0000 0.0001 0.0006 0.0025 0.0127 0.0050 51 0.0116 0.0207 0.0008 0.0003 0.0035 0.0002 0.0005 0.0025 0.0304 0.0508 0.0365 52 0.0102 0.0164 0.0008 0.0002 0.0031 0.0002 0.0004 0.0022 0.0183 0.0425 0.0303 53 0.0068 0.0078 0.0006 0.0001 0.0022 0.0001 0.0002 0.0017 0.0052 0.0269 0.0174 54 0.0029 0.0014 0.0003 0.0000 0.0010 0.0002 0.0001 0.0009 0.0004 0.0121 0.0065 55 0.0111 0.0218 0.0007 0.0003 0.0039 0.0002 0.0005 0.0021 0.0307 0.0709 0.0369 56 0.0091 0.0173 0.0006 0.0002 0.0032 0.0002 0.0004 0.0017 0.0205 0.0560 0.0297 57 0.0067 0.0080 0.0006 0.0001 0.0025 0.0001 0.0002 0.0015 0.0054 0.0351 0.0172 58 0.0028 0.0018 0.0003 0.0000 0.0012 0.0001 0.0001 0.0009 0.0002 0.0146 0.0066 Appendix 118 Figure A1. Normal distribution of standardized residuals. -2 -1 01234 0.01 0.02 0.05 0.10 0.25 0.50 0.75 0.90 0.95 0.98 0.99 Standardized Residuals Normal probability Ethanol -2 -1 0 1 2 0.01 0.02 0.05 0.10 0.25 0.50 0.75 0.90 0.95 0.98 0.99 Standardized Residuals Normal probability Acetaldehyde -3 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 0.01 0.02 0.05 0.10 0.25 0.50 0.75 0.90 0.95 0.98 0.99 Standardized Residuals Normal probability Hydrogen -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3 0.01 0.02 0.05 0.10 0.25 0.50 0.75 0.90 0.95 0.98 0.99 Standardized Residuals Normal probability 1,3-Butadiene -3 -2 -1 0 1 2 3 0.01 0.02 0.05 0.10 0.25 0.50 0.75 0.90 0.95 0.98 0.99 Standardized Residuals Normal probability Water -2 -1 0 1 2 0.01 0.02 0.05 0.10 0.25 0.50 0.75 0.90 0.95 0.98 0.99 Standardized Residuals Normal probability Diethyl ether -2 -1 0 1 2 0.01 0.02 0.05 0.10 0.25 0.50 0.75 0.90 0.95 0.98 0.99 Standardized Residuals Normal probability Ethene -3 -2 -1 012 0.01 0.02 0.05 0.10 0.25 0.50 0.75 0.90 0.95 0.98 0.99 Standardized Residuals Normal probability Butanol Appendix 119 Figure A1 cont. Normal distribution of standardized residuals. -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 0.01 0.02 0.05 0.10 0.25 0.50 0.75 0.90 0.95 0.98 0.99 Standardized Residuals Normal probability Butenes -1.5 -1 -0.5 0 0.5 1 1.5 2 0.01 0.02 0.05 0.10 0.25 0.50 0.75 0.90 0.95 0.98 0.99 Standardized Residuals Normal probability Heavy compounds -3 -2 -1 012 0.01 0.02 0.05 0.10 0.25 0.50 0.75 0.90 0.95 0.98 0.99 Standardized Residuals Normal probability Oxygenated compounds Appendix 120 Figure A2. Standardized residuals versus predicted values. Temperature of experiments 340°C (circle), 360°C (triangle) and 380°C (square). Water content in ethanol feed: blank: 0 wt%, cyan: 3.75 wt%, magenta: 7.5 wt%, black: 15 wt%) 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 -5 -4 -3 -2 -1 0 1 2 3 4 5 Predicted by the model (mol/h) Standardized Residuals Ethanol 0 0.005 0.01 0.015 0.02 0.025 -5 -4 -3 -2 -1 0 1 2 3 4 5 Predicted by the model (mol/h) Standardized Residuals Acetaldehyde 0 0.01 0.02 0.03 0.04 0.05 -5 -4 -3 -2 -1 0 1 2 3 4 5 Predicted by the model (mol/h) Standardized Residuals Hydrogen 0 0.005 0.01 0.015 -5 -4 -3 -2 -1 0 1 2 3 4 5 Predicted by the model (mol/h) Standardized Residuals 1,3-Butadiene 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 -5 -4 -3 -2 -1 0 1 2 3 4 5 Predicted by the model (mol/h) Standardized Residuals Water 0 1 2 3 4 5 6 x 10-4 -5 -4 -3 -2 -1 0 1 2 3 4 5 Predicted by the model (mol/h) Standardized Residuals Diethyl ether 0 0.5 1 1.5 2 2.5 3 3.5 4 x 10-3 -5 -4 -3 -2 -1 0 1 2 3 4 5 Predicted by the model (mol/h) Standardized Residuals Ethene 01234567 x 10-4 -5 -4 -3 -2 -1 0 1 2 3 4 5 Predicted by the model (mol/h) Standardized Residuals Butanol Appendix 121 Figure A2 cont. Standardized residuals versus predicted values. Temperature of experiments 340°C (circle), 360°C (triangle) and 380°C (square). Water content in ethanol feed: blank: 0 wt%, cyan: 3.75 wt%, magenta: 7.5 wt%, black: 15 wt%) 0 0.5 1 1.5 2 x 10-3 -5 -4 -3 -2 -1 0 1 2 3 4 5 Predicted by the model (mol/h) Standardized Residuals Butenes 0 0.2 0.4 0.6 0.8 1 1.2 x 10-3 -5 -4 -3 -2 -1 0 1 2 3 4 5 Predicted by the model (mol/h) Standardized Residuals Heavy compounds 0 1 2 3 4 5 6 x 10-3 -5 -4 -3 -2 -1 0 1 2 3 4 5 Predicted by the model (mol/h) Standardized Residuals Oxygenated compounds