METHANE TRI-REFORMING OVER NICKEL CATALYSTS

UNIVERSIDAD DE CASTILLA-LA MANCHA

FACULTAD DE CIENCIAS Y TECNOLOGÍAS QUÍMICAS

DEPARTAMENTO DE INGENIERÍA QUÍMICA

METHANE TRI-REFORMING OVER NICKEL

CATALYSTS

Memoria que para optar al grado de Doctor en Ingeniería Química

presenta

JESÚS MANUEL GARCÍA VARGAS

Directores: Dr. Fernando Dorado Fernández

Dra. Paula Sánchez Paredes

Composición del tribunal: Dr. José Luis Valverde Palomino

Dr. Luis José Alemany Arrebola

Dr. De Chen

Profesores que han emitido informes favorable de la tesis: Dr. George Marnellos

Dra. Sonia Gil Villarino

Ciudad Real, Octubre de 2014

D. Fernando Dorado Fernández, Profesor Titular de Ingeniería Química de la

Universidad de Castilla-La Mancha, y Dª. Paula Sánchez Paredes, Catedrática de

Ingeniería Química de la Universidad de Castilla- La Mancha,

CERTIFICAN: Que el presente trabajo de investigación titulado: “Methane tri-

reforming over nickel catalyts”, constituye la memoria que presenta D. Jesús Manuel

García Vargas para aspirar al grado de Doctor en Ingeniería Química y que ha sido

realizada en los laboratorios del Departamento de Ingeniería Química de la

Universidad de Castilla-La Mancha bajo su supervisión.

Y para que conste a efectos oportunos, firman el presente certificado

En Ciudad Real a 17 de Octubre de 2014

Fernando Dorado Fernández Paula Sánchez Paredes

i

TABLE OF CONTENTS

TABLE AND FIGURE CAPTIONS VII

DESCRIPCIÓN DEL TRABAJO REALIZADO 1

A. INTRODUCCIÓN 2

A.1. PROBLEMÁTICA AMBIENTAL Y GAS DE SÍNTESIS 2

A.1.1. Combustibles fósiles y contaminación atmosférica 2

A.1.2. Combustibles sintéticos a partir de gas de síntesis 5

A.1.3. Métodos de producción y aplicaciones industriales del gas de síntesis 6

A.1.3.1. Proceso Fischer-Tropsch 10

A.1.3.2. Producción de metanol 11

A.2. PROCESOS DE REFORMADO 12

A.2.1. Reformado con vapor 12

A.2.2. Reformado seco 14

A.2.3. Otros procesos de reformado 15

A.2.3.1. Oxidación parcial 15

A.2.3.2. Reformado combinado 16

A.2.3.3. Reformado autotérmico 16

A.2.4. Tri-reformado 17

A.2.4.1. Revisión bibliográfica del proceso de tri-reformado 18

A.2.4.2. Integración del tri-reformado con otros procesos 19

A.3. CARBURO DE SILICIO 22

A.3.1. Propiedades físico-químicas 22

A.3.1.1. Estructura 22

A.3.2. Principales métodos de producción 24

A.3.3. Aplicaciones del carburo de silicio 25

A.3.4. Carburo de silicio como soporte catalítico 26

A.4. OBJETO Y ALCANCE DEL PRESENTE TRABAJO 30

B. MATERIALES Y MÉTODOS 31

ii

B.1. REACTIVOS EMPLEADOS 31

B.1.1. Reactivos 31

B.1.2. Gases 32

B.2. INSTALACIONES EXPERIMENTALES 32

B.2.1. Preparación de catalizadores 32

B.2.2. Ensayos catalíticos 33

B.2.2.1. Sistema de alimentación 34

B.2.2.2. Sistema de reacción 34

B.2.2.3. Sistema de análisis 35

B.3. EQUIPOS DE ANÁLISIS 35

B.3.1. Microcromatografía de gases 35

B.4. TÉCNICAS DE CARACTERIZACIÓN DE SOPORTES Y

CATALIZADORES 37

B.4.1. Espectroscopía de emisión atómica de inducción de plasma acoplada

(ICP-AES) 37

B.4.2. Adsorción-desorción de nitrógeno 37

B.4.3. Difracción de rayos X 38

B.4.4. Reducción a temperatura programada 40

B.4.5. Microscopía electrónica de transmisión 41

B.4.6. Desorción de dióxido de carbono a temperatura programada 41

B.4.7. Oxidación a temperatura programada 42

B.4.8. Quimisorción estática de hidrógeno 42

B.4.9. Espectroscopía Raman 43

C. RESULTADOS Y DISCUSIÓN 44

D. CONCLUSIONES 57

E. RECOMENDACIONES 59

F. BIBLIOGRAFÍA 60

CHAPTER 1: INFLUENCE OF THE SUPPORT ON THE CATALYTIC

BEHAVIOR OF Ni CATALYSTS FOR THE DRY REFORMING

REACTION AND THE TRI-REFORMING PROCESS 67

Resumen 69

Abstract 70

iii

1.1. INTRODUCTION 71

1.2. EXPERIMENTAL 72

1.2.1. Catalyst preparation 72

1.2.2. Catalyst characterization 73

1.2.3. Catalyst activity measurements 74

1.3. RESULTS AND DISCUSSION 75


1.3.2. Dry reforming catalytic activity 82

1.3.3. Tri-reforming catalytic activity 88

1.4. CONCLUSIONS 92

1.5. REFERENCES 94

CHAPTER 2: PRECURSOR INFLUENCE AND CATALYTIC BEHAVIOR

OF Ni/CeO2 AND Ni/SiC CATALYST FOR THE TRI-REFORMING

PROCESS 97

Resumen 99

Abstract 100








2.3.2. Catalytic activity 112

2.4. CONCLUSIONS 117

2.5. REFERENCES 119

CHAPTER 3: METHANE TRI-REFORMING OVER A Ni/-SIC-BASED

CATALYST: OPTIMIZING THE FEEDSTOCK COMPOSITION 123

Resumen 125

Abstract 126


iv





3.2.4. Experimental design 131



3.3.2. Statistical analysis 133

3.3.3. Influence of the feedstock composition on the H2/CO molar ratio 137

3.3.4. Optimization of the reaction conditions 141


3.5. REFERENCES 146

CHAPTER 4: INFLUENCE OF ALKALINE AND ALKALINE-EARTH

COCATIONS ON THE PERFORMANCE OF Ni/SIC CATALYSTS IN

THE METHANE TRI-REFORMING REACTION 149

Resumen 151

Abstract 152









4.3.3. Influence of the Mg/Ni molar ratio. 166


4.5. REFERENCES 172

v

CHAPTER 5: PREPARATION OF Ni-MG/-SIC CATALYSTS FOR THE

METHANE TRI-REFORMING: EFFECT OF THE ORDER OF METAL

IMPREGNATION 175

Resumen 177

Abstract 178









5.3.3. Characterization after reaction 194


5.5. REFERENCES 198

CHAPTER 6: CATALYTIC AND KINETIC ANALYSIS OF THE

METHANE TRI-REFORMING PROCESS USING A Ni-Mg/-SiC

CATALYST 201

Resumen 203

Abstract 204






6.2.4. Kinetic analysis 207




6.3.3. Kinetic analysis 220


vi

6.5 REFERENCES 228

CHAPTER 7: GENERAL CONCLUSIONS AND RECOMMENDATIONS 231

7. 1. GENERAL CONCLUSIONS 233

7.2. RECOMMENDATIONS 234

LIST OF PUBLICATIONS AND CONFERENCES 237

Publications 239

Conferences 240

vii

TABLE AND FIGURE CAPTIONS

DESCRIPCIÓN DEL TRABAJO REALIZADO

A. INTRODUCCIÓN

Tabla A.1. Demanda global de productos del petróleo (millones barriles/dia). 4

Tabla A.2. Reservas probadas de petróleo. 5

Figura A.1. Principales aplicaciones del gas de síntesis. 8

Figura A.2. Influencia de la composición del gas de síntesis en su aplicación

final. 8

Figura A.3. Estructura 4H-SiC. 23

Figura A.4. Estructura (α) 6H-SiC. 23

Figura A.5. Estructura (β) 3C-SiC. 24

Tabla A.3. Principales propiedades del carburo de silicio. 26

Tabla A.4. Publicaciones sobre la utilización de SiC como soporte catalítico. 28

Tabla A.5. Aplicaciones catalíticas de composites basados en SiC. 29

B. MATERIALES Y MÉTODOS

Figura B.1. Instalación experimental para los experimentos catalíticos. 33

Figura B.2. Equipo de análisis de gases. 36

Tabla B.1. Características y condiciones de análisis del microcromatógrafo de

gases. 36

CHAPTER 1: INFLUENCE OF THE SUPPORT ON THE CATALYTIC

BEHAVIOR OF NI CATALYSTS FOR THE DRY REFORMING

REACTION AND THE TRI-REFORMING PROCESS

Figure 1.1. XRD profiles where (+) denotes nickel oxide diffraction peaks, (^)

denotes metallic nickel diffraction peaks, (#) denotes -Al2O3 diffraction peaks,

(*) denotes CeO2 diffraction peaks and (º) denotes -SiC diffraction peaks. a)

Ni/Al2O3, b) Ni/CeO2, c) Ni/SiC. 76

viii

Figure 1.2. XRD profiles where (+) denotes nickel oxide diffraction peaks, (^)

denotes metallic nickel diffraction peaks and (-) denotes YSZ diffraction peaks.

a) Ni/YSZ, b) Ni/YSZ-O2. 77

Table 1.1. Physical properties of the catalysts. 77

Figure 1.3. XR TEM images. a) Ni/Al2O3, b) Ni/CeO2, c) Ni/SiC, d) Ni/YSZ, e)

Ni/YSZ-O2. 78

Figure 1.4. Temperature-programmed reduction profiles. 80

Figure 1.5. Raman spectra of YSZ-supported catalysts. 81

Figure 1.6. CO2 Temperature-programmed desorption profiles. 82

Table 1.2. Basicity of the catalysts determined by CO2-TPD. 82

Figure 1.7. Dry reforming catalytic activity at 1073 K. Reaction conditions:

CH4 = 4%, CO2 = 4%, N2 balance, total flow rate = 100 Nml/min. CH4 ( ) and

CO2 ( ) consumption rates vs. time on stream (left axis), and H2/CO molar ratio

( ) vs. time on stream (right axis). a) Ni/Al2O3, b) Ni/CeO2, c) Ni/SiC. 84

Figure 1.8. Dry reforming catalytic activity at 1073 K. Reaction conditions:

CH4 = 4%, CO2 = 4%, N2 balance, total flow rate = 100 Nml/min. CH4 ( ) and


( ) vs. time on stream (right axis). a) Ni/YSZ, b) Ni/YSZ-O2. 85

Table 1.3. Main catalytic results. 87

Figure 1.9. Tri-reforming catalytic activity at 1073 K. Reaction conditions: CH4

= 6%, CO2 = 3%, H2O = 3% O2 = 0.6%, N2 balance, total flow rate = 100

Nml/min. CH4 ( ) and CO2 ( ) consumption rates vs. time on stream (left axis),

and H2/CO molar ratio ( ) vs. time on stream (right axis). a) Ni/Al2O3, b)

Ni/CeO2, c) Ni/SiC. 89

Figure 1.10. Tri-reforming catalytic activity at 1073 K. Reaction conditions:

CH4 = 6%, CO2 = 3%, H2O = 3%, O2 = 0.6%, N2 balance, total flow rate = 100

Nml/min. CH4 ( ) and CO2 ( ) consumption rates vs. time on stream (left axis),

and H2/CO molar ratio ( ) vs. time. a) Ni/YSZ, b) Ni/YSZ-O2. 90

CHAPTER 2: PRECURSOR INFLUENCE AND CATALYTIC BEHAVIOR

OF Ni/CeO2 AND NI/SIC CATALYST FOR THE TRI-REFORMING

PROCESS 1


ix

Figure 2.1. TEM images. a) Ni-NC, b) Ni-AC, c) Ni-CC, d) Ni-CiC, e) Ni-NS,

f) Ni-AS, g) Ni-CS, h) Ni-CiS. 107

Figure 2.2. XRD profiles where (*) denotes reflection of nickel oxide and (+)

denotes reflection of nickel metallic. a) Ni-NC, b) Ni-NS. 108

Figure 2.3. Temperature-programmed reduction profiles. a) ceria-based

catalysts, b) SiC-based catalysts. 109

Figure 2.4. CO2 Temperature-programmed desorption profiles. a) ceria-based

catalysts, b) SiC-based catalysts. 111

Table 2.2. Basicity of the catalysts determined by CO2-TPD. 111

Figure 2.5. Catalytic activity at 1073 K for: a) Ni-NC, b) Ni-AC, c) Ni-CC, d)

Ni-CiC. Reaction conditions: CH4 = 6%, CO2 = 3%, H2O = 3%, O2 = 0.6%, N2

balance, total flow rate = 100 mL min-1

. CH4 ( ) and CO2 ( ) consumption

rates vs. time on stream (left axis), and H2/CO molar ratio ( ) vs. time on

stream (right axis). 113

Figure 2.6. Catalytic activity at 1073 K for: a) Ni-NS, b) Ni-AS, c) Ni-CS, d)

Ni-CiS. Reaction conditions: CH4 = 6%, CO2 = 3%, H2O = 3%, O2 = 0.6%, N2


. CH4 ( ) and CO2 ( ) consumption

rates vs. time on stream (left axis), and H2/CO molar ratio ( ) vs. time on

stream (right axis). 114

Table 2.3. Catalytic performance at 20% of conversion: temperature needed to

get this level of conversion and H2/CO molar ratio obtained at this level of

conversion. 117

CHAPTER 3: METHANE TRI-REFORMING OVER A Ni/-SIC-BASED

CATALYST: OPTIMIZING THE FEEDSTOCK COMPOSITION

Table 3.1. Physical properties of the catalyst. 133

Table 3.2. Factor levels. 133

Table 3.3. Central composite design results. 134

Table 3.4. Factorial design statistical analysis. 136

Figure 3.1. Experimental vs. predicted H2/CO molar ratio. 137

Figure 3.2. Effect of the CH4, O2 and H2O volume flows on the H2/CO molar

ratio at 2.0 NmL min-1

CO2 volume flow. The shaded area indicates the region

with a value of H2/CO molar ratio ranging from 1.9 to 2.1. 138

x









Table 3.5. Factorial design for the reaction heat results and optimized variables. 142

Figure 3.5. Experimental vs. predicted overall reaction heat. 143

Figure 3.6. Catalytic activity at 1073 K. Reaction conditions: CH4 = 3.59%,

CO2 = 4.12%, H2O = 1.39%, O2 = 2.11%, N2 balance, total flow rate = 100

NmL min-1

. CH4 ( ) and CO2 ( ) consumption rates vs. time on stream (left

axis), and H2/CO molar ratio ( ) vs. time on stream (right axis). 144

CHAPTER 4: INFLUENCE OF ALKALINE AND ALKALINE-EARTH

COCATIONS ON THE PERFORMANCE OF Ni/SIC CATALYSTS IN

THE METHANE TRI-REFORMING REACTION

Table 4.1. Main physical properties of the catalysts. 158

Figure 4.1. XRD profiles of a) Catalyst support, b) Ni:Na = 2/1, c) Ni:K = 2/1,

d) Ni/SiC, e) Ni:Na = 10/1, f) Ni:K = 10/1, where (^) denotes reflection of

SiC, (+) denotes reflection of metallic nickel, (*) denotes reflection of nickel

oxide and (º) denotes reflection of -cristobalite. 159

Figure 4.2. XRD profiles a) Ni:Mg = 2/1, b) Ni:Ca = 2/1, c) Ni:Mg = 10/1, d)

Ni:Ca = 10/1, where (^) denotes reflection of SiC, (+) denotes reflection of

metallic nickel, (*) denotes reflection of nickel oxide and (-) denotes reflection

of quartz. 161

Figure 4.3. TPR profiles: a) Reference, Na and K promoted catalysts, b) Mg

and Ca promoted catalysts. 163

Figure 4.4. Catalytic activity at 1073 K for: a) Ni/SiC, b) Ni:Ca = 10/1, c)

Ni:Mg = 10/1, d) Ni:Mg = 2/1. Reaction conditions: CH4 = 6%, CO2 = 3%, H2O

= 3%, O2 = 0.6%, N2 balance, total flow rate = 100 NmL min-1

. CH4 ( ) and


( ) vs. time on stream (right axis). 165

Figure 4.5. TPR profiles for Mg promoted and Ni/SiC catalysts. 166

xi

Figure 4.6. Catalytic activity at 1073 K for: a) Ni:Mg = 4/1, b) Ni:Mg = 1/1.

Reaction conditions: CH4 = 6%, CO2 = 3%, H2O = 3%, O2 = 0.6%, N2 balance,

total flow rate = 100 mL min-1

. CH4 ( ) and CO2 ( ) consumption rates vs. time

on stream (left axis), and H2/CO molar ratio ( ) vs. time on stream (right axis). 167


Figure 4.7. CO2-TPD profiles for Mg promoted and Ni/SiC catalysts. 169

Figure 4.8. TPO profiles after reaction for Mg promoted and Ni/SiC

catalysts. 170

CHAPTER 5: PREPARATION OF Ni-MG/-SIC CATALYSTS FOR THE

METHANE TRI-REFORMING: EFFECT OF THE ORDER OF METAL

IMPREGNATION

Figure 5.1. XRD profiles, where (^) denotes reflection of SiC and (º) denotes

reflection of metallic nickel. 184


Figure 5.2. TEM pictures. a) Ni/SiC, b) Ni/Mg/SiC 1/10, c)Ni/Mg/SiC 1/1. 186

Figure 5.3. Temperature Programmed Reduction profiles. 188

Figure 5.4. CO2 Temperature Programmed Desorption profiles. 189

Figure 5.5. Catalytic activity at 1073 K for: a) Ni/SiC, b) Mg/Ni/SiC 1/10 c)

Ni/Mg/SiC 1/10 d) Ni-Mg/SiC 1/10. Reaction conditions: CH4 = 6%, CO2 =

3%, H2O = 3%, O2 = 0.6%, N2 balance, total flow rate = 100 mL min-1

. CH4 ( )

and CO2 ( ) consumption rates vs. time on stream (left axis), and H2/CO molar

ratio ( ) vs. time on stream (right axis). 190

Figure 5.6. Catalytic activity at 1073 K for: a) Mg/Ni/SiC 1/1 b) Ni/Mg/SiC 1/1

c) Ni-Mg/SiC 1/1. Reaction conditions: CH4 = 6%, CO2 = 3%, H2O = 3%, O2 =

0.6%, N2 balance, total flow rate = 100 mL min-1

. CH4 ( ) and CO2 ( )

consumption rates vs. time on stream (left axis), and H2/CO molar ratio ( ) vs.

time on stream (right axis). 191

Table 5.2. Reaction and characterization after reaction parameters. 193

Figure 5.7. Temperature Programmed Oxidation profiles after reaction. 194

Figure 5.8. XRD profiles, where (^) denotes reflection of SiC, (º) denotes

reflection of metallic nickel, (*) denotes reflection of nickel oxide and (#)

denotes reflection of Ni2Si. 196

xii

CHAPTER 6: CATALYTIC AND KINETIC ANALYSIS OF THE

METHANE TRI-REFORMING PROCESS USING A NI-MG/-SIC

CATALYST

Table 6.1. Feed composition (NmL min-1

). 208

Figure 6.1. Nitrogen adsorption–desorption isotherms of catalyst and support. 210


Figure 6.2. Characterization results a) XRD profiles, where (º) denotes

reflection of SiC, (^) denotes reflection of metallic nickel and (+) denotes

reflection of nickel oxide, b) TPR profile, c) CO2-TPD profiles. 212

Table 6.3. CH4 conversion values obtained for each experiment. 214

Figure 6.3. CH4 conversion values for each experiment and temperature. 215

Table 6.4. CO2 conversion values obtained for each experiment. 216

Figure 6.4. CO2 conversion values for each experiment and temperature 217

Figure 6.5. H2/CO molar ratio values for each experiment and temperature. 218

Table 6.5. H2/CO molar ratio values obtained for each experiment. 219

Table 6.6. Statistical significance for each kinetic constant vs temperature. 222

Table 6.7. Kinetic model results. 223

Figure 6.6. Comparison between experimental and modelled molar flows for

the 432 experiment adjustment a) CH4 molar flows b) CO2 molar flows. 224

Table 6.8. Second kinetic model. 225

Figure 6.7. Comparison between experimental and modelled molar flows for

the 356 experiment adjustment a) CH4 molar flows b) CO2 molar flows. 226

1

DESCRIPCIÓN DEL TRABAJO

REALIZADO

Este trabajo forma parte de un programa de investigación sobre la preparación,

caracterización y uso de catalizadores en reacciones de interés industrial, que se está

desarrollando en el Departamento de Ingeniería Química de la Universidad de Castilla-la

Mancha.

En particular, esta Tesis Doctoral tiene como objetivo el estudio y mejora de catalizadores

de níquel aplicados al proceso de tri-reformado, buscando obtener un soporte, precursor y

promotor óptimo a la hora de preparar dichos catalizadores. Así mismo, se ha estudiado la

influencia de la composición del alimento y la temperatura sobre el proceso, modelizando

finalmente los resultados obtenidos. Este proyecto ha sido financiado por la Consejería de

Ciencia y Tecnología de la Junta de Comunidades de Castilla-La Mancha (Proyecto PPII10-

0045-5875).

Descripción del trabajo realizado

2

A. INTRODUCCIÓN

A.1. PROBLEMÁTICA AMBIENTAL Y GAS DE SÍNTESIS

A.1.1. Combustibles fósiles y contaminación atmosférica

En la actualidad existe un deterioro importante del medio ambiente, provocado en gran

parte por la acción humana. Las emisiones de origen industrial juegan un papel trascendente,

ya que afectan tanto al aire como al suelo y al agua. Como consecuencia, se ve directamente

afectado todo el sistema natural, incluyendo la flora y fauna, las cuales sustentan la economía y

biodiversidad global.

Con la industrialización y los avances relacionados con el transporte se incrementó el uso

y demanda de los combustibles fósiles como fuente energética, siendo utilizados como

combustible para propulsión, generación de electricidad y calefacción. Además, las industrias

química y petroquímica dependen de los hidrocarburos como materia prima para obtener sus

productos, ya que la mayor parte de las sustancias químicas sintéticas proceden del petróleo.

Los principales combustibles fósiles son el carbón, el petróleo y el gas natural. Todos

ellos están compuestos por hidrocarburos con distintas proporciones de carbono e hidrógeno.

Su formación proviene de la acumulación de restos de vegetales y animales, que sufren una

serie de reacciones biológicas (actuando microorganismos aerobios y anaerobios) a unas

condiciones muy específicas durante un largo periodo de tiempo.

La principal problemática ambiental que se deriva del uso de combustibles fósiles

consiste en la contaminación atmosférica que genera la combustión de éstos, debido a la

emisión de CO2 y otros compuestos contaminantes.

Vinculado estrechamente con el problema de la contaminación, está el fenómeno del

calentamiento global o efecto invernadero, el cual consiste en un aumento de la temperatura

promedio de la Tierra. Este calentamiento sucede como consecuencia de que ciertos gases de la

atmósfera dejan pasar radiaciones de longitud de onda corta (más energéticas) que llegan a la

superficie terrestre y se absorben, pero cuando la Tierra y los océanos se enfrían e irradian

calor en forma de radiación infrarroja, estos gases la reflejan y es redirigida hacia la superficie

de la Tierra y reabsorbida, calentando la superficie y el aire. Los principales gases causantes


3

del efecto invernadero son el dióxido de carbono, el metano, el óxido nitroso, el ozono y el

vapor de agua.

A menudo se utiliza el término “cambio climático” en lugar de calentamiento global. Esto

es debido a que no solo se produce un aumento de la temperatura media de la Tierra sino que

además los vientos y las corrientes oceánicas que mueven el calor alrededor del globo sufren

también alteraciones, de modo que pueden enfriar algunas zonas, calentar otras y cambiar la

cantidad de lluvia y de nieve. Como consecuencia, el efecto invernadero afecta de manera

diferente en diferentes áreas.

La influencia del efecto invernadero sobre la temperatura global es conocida desde 1824,

cuando Joseph Fourier calculó que la Tierra sería más fría si no hubiera atmósfera. Este efecto

invernadero es lo que hace que el clima en la Tierra sea apto para la vida, ya que sin él la

temperatura media del planeta estaría en torno a unos -18 ᵒC. En 1895, el químico suizo Svante

Arrhenius descubrió que los humanos podrían aumentar el efecto invernadero produciendo

dióxido de carbono. Enunció que el equilibrio térmico de la Tierra dependía en gran medida de

la capa protectora de CO2, calculando que aumentar al doble la concentración de éste,

provocaría un aumento de entre 5-6 ᵒC en la temperatura media de la superficie terrestre [1].

Los niveles de gases de efecto invernadero (GEI) así como las temperaturas medias

globales han aumentado y descendido durante la historia de la Tierra. No obstante, se ha

observado un gran aumento en la concentración de dichos gases y en las temperaturas medias

globales desde inicios del siglo XX. El aumento observado de CO2 en la concentración

atmosférica se debe principalmente a actividades humanas. Este gas representa

aproximadamente un 60% del total de gases de efecto invernadero originados por el ser

humano, por lo que la atención se ha centrado principalmente en él [1].

Por otro lado, el CO2 también puede representar una importante fuente de carbono para la

obtención de combustibles y productos químicos, existiendo diversos estudios que apuntan que

esta fuente será de enorme importancia en el futuro [2]. Actualmente, el uso de CO2 para su

conversión y utilización comienza con su purificación. En general, el CO2 puede ser separado,

recuperado y purificado mediante procesos de absorción, adsorción o el uso de membranas. No

obstante, incluso cuando la fuente de CO2 es concentrada, su purificación requiere

considerables cantidades de energía [3], lo cual penaliza económicamente todo el proceso.


4

A estos perjuicios medioambientales ocasionados por la combustión de los hidrocarburos

hay que sumar el aumento de precio de los mismos y la problemática de abastecimiento que

ocasiona el incremento en la demanda global de energía y la escasez de reservas probadas de

estos compuestos. En la Tabla A.1 se puede observar el consumo en 2012 de diversos

productos derivados del petróleo, incluyendo la gasolina y el diésel, los dos combustibles

líquidos de mayor consumo en el transporte; y la predicción del consumo de estos compuestos

para diferentes años elaborada por la Organización de Países Exportadores de Petróleo (OPEP)

[4]. Se puede observar cómo la demanda total tanto de petróleo como de combustibles líquidos

se prevé que aumente durante los próximos 20 años.

Tabla A.1. Demanda global de productos del petróleo (millones barriles/dia).

2012 2015 2020 2025 2030 2035

Productos ligeros

Etano/GLP 9,7 10,0 10,5 10,9 11,2 11,5

Nafta 5,9 6,2 6,8 7,3 7,9 8,5

Gasolina 22,7 23,3 24,4 25,5 26,5 27,5

Productos medios

Queroseno 6,5 6,7 7,1 7,4 7,7 8,1

Diésel 25,8 27,3 30,0 32,2 34,1 36,0

Productos pesados

Fuel residual 8,2 7,8 7,1 6,6 6,3 6,0

Otros 10,0 10,2 10,5 10,7 10,8 10,9

Total 88,9 91,6 96,3 100,7 104,6 108,5

A estas previsiones hay que sumarle la escasez de las reservas probadas, que según un

informe elaborado por la compañía petrolera BP publicado a mediados del año 2013 se estima

en 1,67·1012

barriles de petróleo (Tabla A.2), lo que equivale a una duración de las reservas de

53 años al ritmo de extracción actual [5].


5

Tabla A.2. Reservas probadas de petróleo.

Reservas probadas ·10

-9

(barriles)

Porcentaje de las reservas

globales (%)

Venezuela 298 18

Arabia Saudí 266 16

Canadá 174 10

Irán 157 9

Irak 150 9

Kuwait 102 6

Emiratos Árabes 98 6

Rusia 87 5

Libia 48 3

Nigeria 37 2

Total mundial 1669 100

La conjunción de todos estos factores hace atractiva la producción de combustibles

sintéticos, ya que reduciría el consumo de petróleo, disminuyendo así la dependencia respecto

de los países productores y de la volatilidad de su precio. Además, en función de la materia

prima seleccionada para producir estos combustibles sintéticos, se evitaría la emisión a la

atmósfera de compuestos perjudiciales para el medio ambiente.

A.1.2. Combustibles sintéticos a partir de gas de síntesis

Como se ha comentado anteriormente, la obtención de combustibles líquidos con

propiedades similares a los obtenidos a partir de la destilación del petróleo ofrece interesantes

alternativas a los combustibles tradicionales. Estos combustibles líquidos sintéticos suelen ser

producidos a partir de gas de síntesis, mezclas gaseosas de hidrógeno y monóxido de carbono,

mediante las tecnologías X-To-Liquids (XTL). Según la procedencia del gas de síntesis se

distingue entre proceso Gas-To-Liquids (GTL), cuando el gas de síntesis proviene del gas

natural; proceso Coal-To-Liquids (CTL), cuando el gas de síntesis proviene del carbón; y

proceso Biomass-To-Liquids (BTL), cuando el gas de síntesis procede de la biomasa. El interés


6

en estas tecnologías está aumentando en los últimos años debido a la disponibilidad de gas

natural y carbón, las ventajas medioambientales derivadas del uso de la biomasa, y los

inconvenientes e incertidumbres que existen alrededor del petróleo.

Además, los combustibles obtenidos a partir de los procesos XTL presentan una serie de

ventajas con respecto a los obtenidos a partir de las fuentes de energía fósiles [6, 7]:

Emiten menos CO2.

No emiten partículas.

Emiten menos NOx.

Presentan un poder energético mayor (medido en mayores valores de índices de

octano para la gasolina y cetano para el diésel).

Las tecnologías XTL constan de tres grandes etapas, dos de las cuales (II y III) son

comunes para todos los procesos mientras que la primera depende de la materia prima usada en

la producción de gas de síntesis. Son las siguientes:

Etapa I: Obtención del gas de síntesis.

Etapa II: Obtención de hidrocarburos líquidos a partir del gas de síntesis vía Fischer-

Tropsch (F-T).

Etapa III: Craqueo de los hidrocarburos líquidos para obtener combustibles

comerciales.

A.1.3. Métodos de producción y aplicaciones industriales del gas de síntesis

Como se ha visto anteriormente, la producción de gas de síntesis es la primera etapa en

los procesos de producción de combustibles líquidos sintéticos vía F-T, siendo además

determinante en el coste total del proceso, ya que esta etapa puede representar más del 50% de

los costes totales de inversión y una gran parte de los costes operativos [8].

El gas de síntesis es un combustible gaseoso obtenido a partir de sustancias ricas en

carbono (hulla, carbón, coque, nafta, biomasa) sometidas a un proceso químico a alta

temperatura. Contiene cantidades variables de monóxido de carbono (CO) e hidrógeno (H2), y

muy a menudo algo de dióxido de carbono (CO2). La forma más habitual de obtenerlo es a

partir de metano, aunque también puede obtenerse con facilidad a partir de otras fuentes

carbonosas como el etano, propano o butano.


7

El nombre de “gas de síntesis” proviene de su uso para la obtención de gas natural

sintético (GNS) y para la producción de amoniaco o metanol. El hidrógeno presente en dicho

gas, una vez purificado, se puede utilizar directamente en pilas de combustible tanto para la

generación de electricidad como para vehículos eléctricos de propulsión.

Aunque puede ser utilizado como combustible, posee menos de la mitad de densidad de

energía que el gas natural. Por ello se usa principalmente en la producción de combustibles

para el transporte y como producto intermedio para la síntesis de otros compuestos químicos.

Además, el gas de síntesis producido en las grandes instalaciones para la gasificación de

residuos puede ser utilizado para generar electricidad in situ, disminuyendo los costes

operativos de estas plantas.

Los procesos de gasificación del carbón se utilizaron durante muchos años para la

fabricación de gas de alumbrado (gas de hulla) que alimentaba el alumbrado de gas de las

ciudades y en cierta medida, la calefacción, antes de que la iluminación eléctrica y la

infraestructura para el gas natural estuvieran disponibles.

Este gas también se utiliza como producto intermedio en la producción de petróleo

sintético, para su uso como combustible o lubricante a través de la síntesis de Fischer-Tropsch,

y previamente al proceso Mobil para convertir metanol en gasolina. En la Figura A.1 se

resumen las principales aplicaciones del gas de síntesis.


8

Figura A.1. Principales aplicaciones del gas de síntesis.

La relación molar H2/CO es un parámetro fundamental a la hora de clasificar el gas de

síntesis, ya que en función de esta relación se podrán sintetizar unos compuestos determinados

u otros, tal y como se muestra en la Figura A.2.

Figura A.2. Influencia de la composición del gas de síntesis en su aplicación final.


9

Históricamente, según los distintos métodos de producción del gas de síntesis, éste recibía

diferentes nombres:

Gas de alumbrado o gas de hulla: Se produce por pirólisis, destilación o pirogenación

de la hulla en ausencia de aire y a alta temperatura (1473-1573 K), o bien, por pirólisis

del lignito a baja temperatura. En estos casos se obtiene coque (hulla) o semicoque

(lignito) como residuo, que se usa como combustible. Este gas fue utilizado como

combustible para el alumbrado público (luz de gas) a finales del siglo XIX y

comienzos del siglo XX. Contiene un 45% de hidrógeno, 35% de metano, 8% de

monóxido de carbono y otros gases en menor proporción [9].

Gas de coque o gas de coquería: Se obtiene por calentamiento intenso y lento de la

hulla (hulla grasa) con una combinación de aire y vapor, a alta temperatura, en las

coquerías. A parte del coque sólido fabricado, de gran interés para la industria

siderúrgica y la síntesis de acetileno, se forma un gas que contiene hidrógeno,

monóxido de carbono, nitrógeno y dióxido de carbono.

Gas de generador de gasógeno o gas de aire: Se obtiene haciendo pasar aire a través de

una capa gruesa de gránulos de carbono de coque incandescente. A mayor

temperatura, mayor proporción de monóxido de carbono y menor proporción de

dióxido de carbono. Tiene escaso poder calorífico, debido principalmente a la dilución

con el nitrógeno atmosférico [10].

Gas de agua: Se obtiene haciendo pasar vapor de agua sobre coque a alta temperatura.

Su llama es de color azul por lo que también se llama gas azul. Este gas se puede

transformar en metanol o alcanos empleando catalizadores heterogéneos apropiados.

Esta reacción es fuertemente endotérmica por lo que requiere temperaturas muy altas.

Gas pobre: Se obtiene haciendo pasar alternativamente vapor de agua y aire sobre

carbón incandescente (alternancia de chorros de vapor y aire). Es una mezcla de los

dos métodos anteriores. Cuando el lecho de coque se ha enfriado a una temperatura a

la que la reacción endotérmica ya no puede continuar, el vapor de agua es

reemplazado por un chorro de aire. La formación inicial de dióxido de carbono

(exotérmica) aumenta la temperatura del lecho de coque y va seguida por la reacción

endotérmica en la que éste (CO2) se convierte en monóxido de carbono (CO). La

reacción global es exotérmica, originando "gas pobre". El oxígeno puro puede

sustituir al aire para evitar el efecto de dilución, y en este caso el poder calorífico es

más alto.


10

Gas de agua carburado: Se obtiene mezclando gas de agua con petróleo gasificado en

un carburador. Posee un poder calorífico más alto que los anteriores.

Gas ciudad: Se obtiene a partir de la oxidación de petróleo o algún derivado (fuel-oil,

nafta) mediante vapor de agua y aire. Se debe eliminar el azufre para evitar la

corrosión, y también el monóxido de carbono por su toxicidad. Ha sido reemplazado

por el gas natural y los gases licuados del petróleo (GLP, como butano o propano)

para todo tipo de fines, pues estos poseen un poder calorífico doble. A veces se llama

gas ciudad a cualquier gas de síntesis producido para abastecer el consumo doméstico

y distribuido mediante redes de tuberías, ya sea obtenido a partir de carbón o de

petróleo [11].

En la actualidad estos procesos clásicos de producción del gas de síntesis han sido

sustituidos por los procesos de reformado, los cuales se comentarán en profundidad en la

sección A.2.

A.1.3.1. Proceso Fischer-Tropsch

El proceso Fischer-Tropsch, conocido también como licuefacción indirecta del carbón,

fue desarrollado por los químicos alemanes Franz Fischer y Hans Tropsch entre 1920 y 1925.

En el proceso original, la producción del gas de síntesis se realizaba mediante la

gasificación de carbón por oxidación parcial utilizando oxígeno como oxidante y vapor de agua

como moderador a altas temperaturas. La relación molar del gas de síntesis está determinada

por las proporciones de oxidante y moderador utilizadas. La posibilidad de utilizar mezclas de

carbón y biomasa residual o residuos de carácter orgánico, como fuente de carbono, permite

desarrollar procesos de carácter cada vez más neutro con respecto al CO2 atmosférico.

En la síntesis de Fisher-Tropsch, se obtiene una mezcla compleja de hidrocarburos de

cadena lineal y ramificada. También aparecen productos oxigenados como alcoholes, aldehídos

y ésteres, aunque los mayoritarios son parafinas lineales y α-olefinas. Los hidrocarburos

obtenidos con punto de ebullición en el rango de gasolina y diésel son de alta calidad, debido a

que no presentan heteroátomos en su estructura. La fracción de destilado medio presenta un

índice de cetano elevado.

Las reacciones que intervienen en el proceso de gasificación son las siguientes:

C(s) + CO2 ⇄ 2CO = 172,2 kJ/mol (Ecuación A.1)


11

C(s) + H2O ⇄ CO + H2 = 131,4 kJ/mol (Ecuación A.2)

CO + 3H2 ⇄ CH4 + H2O

= -206,3 kJ/mol (Ecuación A.3)

CO + H2O ⇄ CO2 + H2 = -41,1 kJ/mol (Ecuación A.4)

En una segunda etapa el gas de síntesis se transforma mediante un proceso catalítico casi

exclusivamente en parafinas y α-olefinas lineales, mediante un mecanismo en cadena en el que

la probabilidad de crecimiento de la cadena es prácticamente independiente de su tamaño.

También es usual obtener productos oxigenados como alcoholes.

2n H2 + n CO → CnH2n + n H2O n-olefinas (α-olefinas) (Ecuación A.5)

(2n+1) H2 + n CO → CnH2n+2 + n H2O n-parafinas (Ecuación A.6)

nCO + 2n H2 → CnH2nOH + (n-1) H2O alcoholes (Ecuación A.7)

Mediante hidrocraqueo o reformado catalítico de los productos de la síntesis Fischer-

Tropsch pueden obtenerse combustibles diésel o kerosenos adecuados para el transporte

comercial terrestre, marítimo y aéreo, dada su elevada densidad energética.

A.1.3.2. Producción de metanol

Se conocen distintos procesos de producción de metanol. Por un lado está el proceso

Chem System, el cual es un proceso de alta presión (400-600 atm) y una temperatura de 673 K

usando como catalizador Cu o Zn-óxido de cromo. Por otro lado se conoce el proceso Lurgi, el

cual trabaja a baja presión (50-100 atm), con una temperatura de 523-573 K y en presencia de

un catalizador de Cu, Zn y un compuesto para aumentar la resistencia al envejecimiento.

El metanol se produce en unidades de gran capacidad de producción a partir de gas de

síntesis por reacción catalítica del hidrógeno con el CO según la ecuación A.8.

2H2 + CO ⇄ CH3OH = -92 kJ/mol (Ecuación A.8)

Si el gas de síntesis se obtiene a partir de reformado con vapor, la composición del mismo

(CO + 3 H2), no es la estequiometría adecuada. En este caso se adiciona CO2, que según la

ecuación A.9, utiliza mayor cantidad de hidrógeno [12].


12

3H2 + CO2 ⇄ CH3OH + H2O = -50 kJ/mol (Ecuación A.9)

Las reacciones se llevan a cabo a una presión del orden de 50 bares y con grados de

conversión relativamente bajos, por lo que es preciso recircular al reactor el gas no convertido,

una vez separado por condensación el producto de la reacción en forma de solución acuosa de

metanol.

La importancia de la producción de metanol reside en los productos que se pueden

obtener a partir de éste, ya que es posible sintetizar materias primas, obtener combustibles,

carburantes o componentes de mezclas para bencina, y además sirve como fuente de C para la

obtención de petroproteínas. El principal derivado del metanol es el formaldehído (H2CO) que

se comercializa en disolución acuosa, como trioxano (un trímero del mismo) o como

paraformaldehído (un oligómero) con destino a la preparación de resinas acetálicas y como

reactivo de las reacciones de alcoholización para preparar alcoholes multifuncionales. El

metanol se utiliza también como reactivo en la fabricación de éteres mejoradores del número

de octano, como el metil-terbutil-éter (MTBE) y para la síntesis del ácido acético por vía

carbonilación.

A.2. PROCESOS DE REFORMADO

A.2.1. Reformado con vapor

Hoy en día, el procedimiento más empleado a escala industrial para la producción de gas

de síntesis es el reformado con vapor (steam reforming) del gas natural y del gas de refinería.

En el reformado con vapor una fuente de metano, como el gas natural o subproductos del

refino del petróleo, reaccionan en un proceso altamente endotérmico con vapor de agua sobre

un catalizador, típicamente níquel, para producir hidrógeno, monóxido de carbono y una

cantidad relativamente pequeña de dióxido de carbono.

CH4 + H2O ⇄ CO + 3H2 = 206,3 kJ/mol (Ecuación A.10)

Esta reacción suele estar seguida en la práctica de la reducción del vapor de agua con

monóxido de carbono mediante water-gas shift (WGS, Ecuación A.4). Como último paso se

necesitaría una etapa de purificación para eliminar el dióxido de carbono y otras impurezas,

obteniéndose una corriente prácticamente pura de hidrógeno. Este proceso de reformado de


13

vapor se puede aplicar no solamente al metano sino a otros hidrocarburos como etanol,

propano o incluso gasolina.

Estas reacciones se llevan a cabo simultánea y consecutivamente en uno o varios

reactores, por lo que el gas producido consiste en una mezcla de H2, CO y CO2, además de

vapor de agua, algo de CH4 sin reaccionar y los gases inertes presentes en el gas de

alimentación. El producto de la reacción en conjunto se conoce como “gas de síntesis” y la

concentración de los distintos componentes depende de las condiciones de reacción:

temperatura, presión y relación H2O/CH4. No obstante, la cantidad de gas de síntesis generada

a partir del metano aumenta al aumentar la temperatura y disminuir la presión. Al aumentar la

temperatura, la reacción WGS se hace menos dominante y los productos principales son el H2 y

CO. Esta tecnología es la más utilizada para la producción de hidrógeno y gas de síntesis a

nivel industrial, obteniéndose una relación H2/CO de 3 o superior.

El reformado de metano puede verse afectado por la formación de carbón que se deposita,

en forma de hollín o coque, en el catalizador reduciendo su actividad o en partes del equipo,

pudiendo producir atascos. El carbón se puede formar por la descomposición del CH4 o la

desproporción del CO.

CH4 ⇄ C + 2H2 = 75,7 kJ/mol (Ecuación A.11)

2CO ⇄ C + CO2 = -172,2 kJ/mol (Ecuación A.12)

En la práctica, la formación indeseada de coque puede prevenirse utilizando vapor en

exceso y tiempos de residencia cortos en el reactor.

En resumen, los problemas más importantes a los que se enfrenta el reformado con vapor

son:

Es una reacción relativamente lenta.

Se produce un sobrecalentamiento del vapor de agua a 1073 K.

Reacción muy endotérmica. Es necesario un gran aporte de calor.

La relación teórica molar H2/CO obtenida en el proceso es igual a 3, por lo que no es la

óptima para procesos como la producción de diésel por Fischer-Tropsch o la producción de

metanol, donde es necesaria una relación molar H2/CO igual a 2.


14

A.2.2. Reformado seco

La mayoría del gas de síntesis, a nivel industrial, se produce por reformado con vapor,

pero esta tecnología presenta algunas limitaciones, ya comentadas en el apartado anterior. Para

tratar de evitar estos inconvenientes, se están estudiando en los últimos años otros procesos de

reformado, como el reformado de metano con CO2. El gas de síntesis también puede producirse

mediante la reacción del CO2 con gas natural u otros hidrocarburos alifáticos, recibiendo este

proceso el nombre de reformado con CO2 o reformado seco, ya que no se utiliza vapor.

La reacción que tiene lugar en el reformado seco es:

CO2 + CH4 ⇄ 2CO + 2H2 = 247,3 kJ/mol (Ecuación A.13)

Esta reacción produce un gas con una relación molar H2/CO de 1. El reformado seco de

metano está recibiendo una gran atención ya que hace posible la utilización de gases de efecto

invernadero (dióxido de carbono y metano) para obtener productos químicos de elevado valor

añadido.

La reacción es muy endotérmica, favorecida a bajas presiones y altas temperaturas, que se

ve influida en la mayoría de los casos por la reacción inversa a la de water-gas shift:

CO2 + H2 ⇄ CO + H2O = 41 kJ/mol (Ecuación A.14)

En relación con lo anterior, la conversión y utilización de CO2 son elementos importantes

en la investigación química sobre desarrollo sostenible ya que representa una importante fuente

de carbón para producir combustibles y otros productos en la industria química. La tendencia

general es usar CO2 puro para este fin. El CO2 puede ser separado, recuperado y purificado a

partir de fuentes concentradas del mismo en dos o más etapas basadas en absorción, adsorción

o separación con membranas. El principal inconveniente es que estos procesos requieren un

gran aporte de energía.

El reformado seco presenta dos grandes problemas como son la desactivación del

catalizador por la formación de coque y el elevado consumo de energía debido a que el proceso

es endotérmico. Por estas razones se han probado varios metales como catalizadores.

Los catalizadores más activos para el reformado con CO2 son, en general, los metales

preciosos y los metales de transición de los grupos VII y VIII. Los metales preciosos son los


15

que presentan mayor actividad, además de menor formación de coque que para el níquel. La

principal desventaja de estos catalizadores es que no son adecuados para su uso comercial,

debido a su elevado coste y baja disponibilidad [13]. Por otro lado, la desactivación que sufren

los metales no preciosos en esta reacción debido a la formación de coque hace que la

aplicación comercial de este proceso se vea muy limitada [14].

La formación de coque se ve favorecida por la ausencia de agua y la elevada relación C/H

en la alimentación. Por ello, el riesgo de formación de coque se puede minimizar mediante

alguna de las siguientes opciones:

Elevando la relación H/C u O/C de la alimentación por adición de vapor u O2.

Utilizando un catalizador que minimice la producción de coque.

Aumentando la relación estequiométrica CO2/CH4 en la alimentación.

Envenenando selectivamente la fase activa (pasivación con H2S).

A.2.3. Otros procesos de reformado

A.2.3.1. Oxidación parcial

La oxidación parcial de metano permite obtener gas de síntesis a partir de metano con un

proceso exotérmico que genera un gas de síntesis con una relación molar H2/CO de 2, útil para

el proceso de Fischer-Tropsch.

CH4 +

O2 ⇄ CO + 2H2

= -35.9 kJ/mol (Ecuación A.15)

Esta reacción, que puede llevarse a cabo con o sin catalizador, presenta el problema de

que los productos, CO y H2, pueden seguir oxidándose para formar CO2 y agua en reacciones

altamente exotérmicas, dando lugar a relaciones H2/CO menores de 2. La producción de

grandes cantidades de energía en forma de calor tampoco es deseada, pues se desaprovecha si

no se le puede dar un uso inmediato en otro proceso. La combinación de una elevada

conversión de metano y alta velocidad espacial da lugar a una fuerte liberación de calor, que

puede dañar el catalizador y convertir el proceso en peligroso y difícil de controlar.

CO +

O2 ⇄ CO2


H2 +

O2 ⇄ H2O



16

A.2.3.2. Reformado combinado

Otro de los procesos de reformado que se puede encontrar en la literatura científica es el

reformado combinado. Éste proceso consiste en la reacción simultánea del metano con CO2 y

H2O y presenta ciertas ventajas frente al reformado empleando sólo CO2 o vapor de agua. Las

reacciones principales que tienen lugar en el reformado combinado son las representadas por

las ecuaciones A.10 y A.13. Simultáneamente, también ocurre la reacción water-gas shift

(Ecuación A.4)

En general, entre las ventajas que presenta el reformado combinado se pueden destacar

las siguientes:

Se pueden producir relaciones molares H2/CO en el rango de 1-3 ajustando la relación

CO2/H2O en la alimentación.

La adición de H2O resulta beneficiosa para aumentar la resistencia a la deposición de

carbón en el catalizador, ya que inhibe la formación de carbón por tener lugar la

reacción A.2.

A.2.3.3. Reformado autotérmico

El reformado autotérmico es la combinación del reformado con vapor y la oxidación

parcial del metano. El proceso global es exotérmico y el consumo energético se reduce, ya que

la oxidación de parte del metano aporta el calor necesario para la reacción de reformado. La

temperatura a la salida del reactor se encuentra entre 1223 y 1373 K, y la presión puede llegar a

ser de 100 atm.

El proceso de reformado autotérmico se lleva a cabo en un reactor a presión con un lecho

de catalizador, por el que se hace pasar la mezcla de gas y vapor de agua junto con oxígeno de

origen criogénico o, cuando se produce gas de síntesis para la fabricación de amoníaco, aire

comprimido.

Cuando el gas de síntesis está destinado a la fabricación de metanol, deberá mantenerse

una relación de 2 a 1 entre el H2 y el CO, por lo que no procede someter al gas reformado a un

proceso de conversión del CO mediante la reacción de water-gas shift.

Los inconvenientes que presenta este método tienen que ver con el tamaño de todo el

equipo en conjunto, ya que al llevar asociado un sistema de recuperación del calor, el equipo es


17

muy grande y costoso, y con la necesidad de purificar los gases, lo que reduce la eficiencia

total y añade costes significativos al proceso [15].

Aun así, el reformado autotérmico es una de las tecnologías que más se están estudiando

en los últimos tiempos, puesto que combina las ventajas e inconvenientes del reformado con

vapor y la oxidación parcial.

A.2.4. Tri-reformado

El proceso de tri-reformado consiste en una combinación sinérgica de la reacción

endotérmica del reformado seco (Ecuación A.13), junto con el reformado con vapor (Ecuación

A.10) y la reacción exotérmica de oxidación parcial de metano (Ecuación A.15). Con este

proceso el CO2, H2O y O2 provenientes de los gases de combustión de plantas de energía

basada en combustibles fósiles o de la gasificación de residuos sólidos pueden ser utilizados

como reactivos para el tri-reformado de gas natural para la producción de gas de síntesis [16].

El proceso de tri-reformado representa una combinación sinérgica de los tres procesos

anteriormente mencionados, los cuales tienen lugar en un mismo reactor, en presencia de un

catalizador. Como resultado se produce gas de síntesis con una razón H2/CO entre 1,5 y 2, lo

que lo hace adecuado para el proceso de Fischer-Tropsch [17]. Además, existen multitud de

reacciones laterales que implican la formación (Ecuaciones A.11 y A.12) y la destrucción de

coque (Ecuaciones A.1, A.2 y A.18), lo cual favorece que la desactivación sufrida por el

catalizador debido a la presencia de coque sea muy inferior a la sufrida en el proceso de

reformado seco.

C + O2 → CO2 = -393,7 kJ/mol (Ecuación A.18)

La presencia de H2O y O2 reduce la deposición de coque respecto al reformado seco

(Ecuaciones A.2 y A.18). Además, la incorporación de O2 en la reacción genera calor in-situ,

que puede ser utilizado para incrementar la eficiencia energética del proceso, gracias a la

reacción exotérmica de oxidación parcial.

La combinación del reformado seco con el reformado con vapor cumple dos objetivos

importantes como son: producir gas de síntesis con una relación H2/CO deseada, adecuada para

la síntesis de metanol y el proceso Fischer-Tropsch, y además mitigar el problema de la

formación de coque, que es especialmente significativo en el reformado seco. A su vez, la


18

incorporación de O2 en la reacción también reduce la formación de coque, aumentando así la

vida del catalizador y mejorando la eficiencia del proceso al combinarse reacciones

exotérmicas y endotérmicas.

Por lo tanto, la propuesta del tri-reformado puede resolver algunos de los problemas más

importantes que se encuentran en los distintos procesos de reformado.

A.2.4.1. Revisión bibliográfica del proceso de tri-reformado

El proceso de tri-reformado fue propuesto en fechas relativamente recientes por el grupo

del profesor Song [18], como una forma de valorizar el CO2 producido en diversas

instalaciones industriales y poder aprovechar su potencial como fuente de carbono mediante su

transformación en gas de síntesis.

Considerando las ventajas de este proceso, Halmann y col. [19, 20] estudiaron el

tratamiento de los gases efluentes de centrales térmicas de carbón y de gas natural, así como

los de otras industrias como las cementeras y metalúrgicas, en términos de eficiencia

energética, viabilidad económica, ahorro de combustible y reducción de emisiones de CO2. En

sus cálculos, también consideraron la posterior transformación del gas de síntesis en productos

útiles como amoniaco, metanol e hidrógeno. Sus conclusiones indican que efectivamente el tri-

reformado es una alternativa perfectamente viable, con importantes reducciones en las

emisiones de CO2 y ahorro de combustible, llegando hasta una reducción del 70% en las

emisiones de CO2 en centrales térmicas de carbón y de un 66% en centrales térmicas de gas.

Dado que este proceso combina reacciones endotérmicas (Ecuaciones A.10 y A.13) y

exotérmicas (Ecuación A.15), pueden aparecer problemas de transmisión de calor y

homogeneidad de la temperatura en el lecho de catalizador. La incertidumbre en la

determinación de la temperatura del lecho hace difícil evaluar la actividad del catalizador y la

influencia de diversos parámetros, así como comprender el mecanismo de la reacción.

Para este proceso se han empezado a estudiar los catalizadores empleados en los procesos

convencionales de reformado. Debido al elevado coste y disponibilidad limitada de metales

nobles como Pt, Rh y Ru, las investigaciones se han orientado hacia el estudio de los

catalizadores basados en metales de transición. Así, durante muchos años, los catalizadores de

níquel han demostrado ser los más adecuados para el reformado de hidrocarburos [13],

fundamentalmente aquellos basados en Ni como fase activa. En este sentido, Song y Pan [21]


19

comprobaron que el catalizador Ni/MgO era el más adecuado para incrementar la conversión

de CO2 en las condiciones típicas del tri-reformado, debido a la interacción del CO2 con el

MgO. Asimismo, comprobaron que el catalizador Ni/MgO/CeZrO, si bien menos activo que el

anterior, era más estable. Recientemente, se han publicado trabajos con otros soportes para el

níquel, como Ni/Al2O3 [22, 23], NiO-YSZ-CeO2 [13] o Ni/MgxTi1-xO [24]. En todos ellos se

concluye que el tri-reformado con catalizadores de níquel es muy prometedor, pero hasta el

momento no se ha podido formular un catalizador definitivo, combinando una elevada

actividad con una suficiente resistencia a la desactivación.

Diversos autores también han analizado la influencia de la composición del alimento en el

tri-reformado de metano, aunque no de forma sistemática. Huang y col. [25] mostraron que un

aumento en la cantidad de agua que contiene el alimento, (H2O/(CH4+CO2+H2O)) de 1/9 a 4/9,

y manteniendo constante la relación molar CH4/(CO2+H2O) en 2/2,5, la conversión de metano

del reformado seco + reformado con vapor aumenta ligeramente desde 97,7% hasta 98,9%

mientras que la conversión de CO2 cayó del 94,1% al 84%. Sun y col. [26] también

investigaron cómo diferentes cantidades de agua a presión atmosférica influyen en la actividad

catalítica, usando un alimento con unas relaciones molares CH4:CO2:O2 de 2:1:0,6. Sus

resultados mostraron que la adición de vapor de agua provoca un ligero aumento en la

conversión de metano y una gran caída en la conversión de dióxido de carbono (desde un 86%

a un 57,3%) al cambiar la relación molar H2O/CH4 de 0 a 0,5. Song y col. [21] analizaron la

influencia de la concentración de oxígeno manteniendo constante la relación molar H2O/CO2

en 1 y la relación molar (H2O + CO2 + 2O2)/CH4 en 1,2. Observaron que la concentración de

oxígeno en el alimento afecta claramente a la actividad catalítica, provocando una disminución

en la conversión de dióxido de carbono desde el 78,4% para un alimento CH4:H2O:CO2:O2 =

1:0.6:0.6:0 hasta el 67,8% para un alimento CH4:H2O:CO2:O2 = 1:0.27:0.27:0.33.

A.2.4.2. Integración del tri-reformado con otros procesos

Los gases de combustión procedentes de centrales térmicas de carbón, gas o fuel, así

como de diversas industrias pesadas, como las de producción de cal, hierro y cemento, son la

principal fuente de emisión de CO2 relacionada con la actividad humana. En concreto, los

gases procedentes de centrales térmicas contribuyen en un 47% a las emisiones anuales

mundiales de dióxido de carbono mientras que los procedentes de las industrias pesadas

suponen hasta un 10% del total anual, derivadas principalmente de la combustión de


20

combustibles fósiles para proporcionar calor en procesos a elevadas temperaturas [19]. El

proceso de tri-reformado presenta una gran importancia tanto a nivel industrial como

medioambiental, en un intento de reducir estas emisiones.

De esta forma el proceso de tri-reformado puede acoplarse o combinarse con diversos

procesos industriales a gran escala como se verá a continuación.

Tratamiento de gases de combustión de centrales térmicas de carbón

Los gases emitidos por estas centrales térmicas están compuestos mayoritariamente por

CO2, H2O, O2 y N2, en una proporción media de 13:9:4:74 en volumen. Añadiendo a este

efluente CH4, H2O y aire (20, 11 y 24 partes) resulta una mezcla, que, a 1000 K y 1 atm puede

producir un gas de síntesis con una razón molar H2/CO resultante de 2,06, lo que la hace

adecuada para la síntesis de metanol y el proceso Fischer-Tropsch. Según el esquema anterior y

tomando como referencia una central térmica convencional de 500 MW con una eficiencia del

45%, Halmann y Steinfeld [20] estudiaron los beneficios que aportaría el tratamiento de los

gases de combustión de la misma mediante tri-reformado y el posterior uso del gas de síntesis

producido en la producción de metanol, hidrógeno, amoniaco o urea. Dichos beneficios fueron

cuantificados en cuanto a reducción en las emisiones de CO2 y ahorro de combustible,

estimándose una reducción de hasta el 46,7% en las emisiones de CO2 y de hasta el 75% en el

consumo de combustible. Además, se realizó una evaluación económica preliminar del coste

del proceso y el precio de venta del producto, encontrándose que el proceso sería rentable en

todos los casos.

Tratamiento de gases de combustión de centrales térmicas de gas

La concepción es similar al caso del tratamiento de los gases de combustión de centrales

de carbón. En este caso, los gases de combustión producidos están compuestos por CO2, H2O,

O2 y N2 en una proporción en volumen de 9:19:2,5:69,5. Añadiendo gas natural y aire (15 y 19

partes respectivamente), la condición de termoneutralidad se alcanza a 1100 K y una atmósfera

de presión. La razón H2/CO obtenida es de 2,03, adecuada al igual que en el caso anterior para

la síntesis de metanol o Fischer-Tropsch. Por tanto, el gas de síntesis obtenido mediante el

proceso de tri-reformado puede emplearse para producción de metanol, hidrógeno, amoniaco y

urea al igual que en el caso anterior. En este caso, para los cálculos se tomó como referencia

los gases de combustión emitidos por una central térmica de gas moderna de 400 MW, con una


21

eficiencia del 49%, hallándose que se podría reducir hasta en un 50% las emisiones de CO2 y

un 74,9% el consumo de combustible mediante el tratamiento de los gases efluentes por el

proceso de tri-reformado [19]. Asimismo se confirmó la bondad económica que tendría la

producción de los compuestos químicos anteriormente comentados mediante este proceso

Tratamiento de residuos sólidos urbanos

La cantidad de residuos sólidos urbanos ha crecido significativamente tanto en los

países industrializados como aquellos en vías de desarrollo en los últimos años, apareciendo el

consiguiente problema de su gestión y eliminación [27]. Enormes cantidades de dinero y

energía son necesarios para el tratamiento de dichos residuos, por lo que existe un creciente

interés por su reutilización. Una alternativa es recobrar su energía química, y utilizarlos como

sustitutos o complemento de los combustibles fósiles. La forma más sencilla para conseguir lo

anterior es la incineración de los residuos, con lo que además se consigue reducir su peso y

volumen. Sin embargo, este método, si bien ha demostrado ser viable en términos de

generación de calor y electricidad, tiene el serio inconveniente de la emisión de compuestos

tóxicos, como diversos gases ácidos (SOx, HCl, HF, NOx), compuestos orgánicos volátiles,

dioxinas y metales pesados [28-30]. Tomando además en cuenta las restricciones legislativas

medioambientales, cada vez más estrictas, los estudios más actuales han sido enfocados a la

búsqueda de otras alternativas para el tratamiento de los residuos urbanos que presten mayor

atención a los aspectos medioambientales, siempre que se conserve la viabilidad económica y

eficiencia energética. Entre estas alternativas, cobra cada vez mayor fuerza el proceso de

gasificación de los residuos [31, 32].

Este tratamiento de gasificación consiste en la conversión termoquímica de un

material que contiene carbono mediante la adición de calor y vapor de agua, aire o una mezcla

de los mismos. La gasificación reduce significativamente los problemas de emisión de tóxicos,

obteniéndose como productos principales, aparte de cenizas y aceites, diversos gases,

fundamentalmente CO, H2, CO2 y CH4 [33] que a su vez pueden ser utilizados en otros

procesos posteriores. Aunque la composición concreta de los gases efluentes de la gasificación

varía según factores como temperatura, presión, catalizador, etc., entre dichos gases se

encuentran los reactantes principales del proceso de tri-reformado. A partir de los gases

efluentes y mediante este proceso, podría obtenerse gas de síntesis, el cual, según se ha

justificado anteriormente, puede utilizarse para sintetizar productos de gran interés. Por lo tanto


22

el proceso de tri-reformado cobra un gran interés en la transformación de residuos en productos

de elevada utilidad.

A.3. CARBURO DE SILICIO

A.3.1. Propiedades físico-químicas

El carburo de silicio es uno de los compuestos cerámicos no óxidos más importantes, ya

que presenta diversas aplicaciones industriales gracias a la combinación de diversas

propiedades químicas y termomecánicas deseables. Es reutilizable, tiene alta estabilidad

térmica, alta resistencia mecánica con una dureza de ~9 en la escala de Mohs [34] similar al

diamante, alta conductividad térmica, es inerte químicamente y es resistente a la oxidación en

atmósfera pura de O2 hasta los 700 ºC. Todas estas cualidades hacen que el carburo de silicio

sea un candidato muy interesante para ser utilizado como abrasivo, refractario, material

cerámico y otras numerosas aplicaciones de alto rendimiento.

Principalmente, su buena conductividad térmica, su estabilidad a alta temperatura y su

resistencia mecánica han hecho que el carburo de silicio se use como soporte de catalizadores

en algunas reacciones con condiciones extremas, como procesos que operan a altas

temperaturas o en ambientes oxidantes. Sin embargo, la baja superficie específica de este

material limitó su aplicación hasta las últimas décadas donde fueron desarrollados varios

procesos para producir SiC con mayor área superficial.

A.3.1.1. Estructura

El carburo de silicio existe en un gran número de estructuras cristalinas diferentes pero

estrechamente relacionadas, normalmente conocidas como “Politipos”. Entre los más de 200

politipos de SiC que se han encontrado hasta la fecha el más común incluye el 3C, 4H, NH, o

15R, donde (C), (H), y (R) son las tres categorías cristalográficas básicas, cúbica, hexagonal y

romboédrica, respectivamente.

Los politipos 4H, 6H (representados en las Figuras A.3 y A.4 respectivamente) y el 15R

se denominan colectivamente como α-SiC [35] Es el más comúnmente encontrado y es estable

a temperatura elevada (>1700 ºC).


23

Figura A.3. Estructura 4H-SiC.

Figura A.4. Estructura (α) 6H-SiC.

El politipo 3C, representado en la Figura A.5 y también llamado β-SiC, tiene una

estructura cúbica similar a la estructura del diamante. Se forma a temperaturas inferiores a

1700 ºC [35] y se caracteriza por tener un área superficial mayor que la forma alfa (entre 25 y

30 m2/g), una menor diferencia de banda (~2,4 eV), una mayor movilidad de electrones (~800

cm2·V

-1·s

-1), y buena resistencia mecánica, resistencia a la oxidación, resistencia a ácidos

fuertes y bases, y una elevada conductividad térmica. Por esta razón la forma beta ha cobrado

mayor interés como soporte para catalizadores heterogéneos en los últimos años.


24

Figura A.5. Estructura (β) 3C-SiC.

A.3.2. Principales métodos de producción

En el siglo XIV se desarrolló el proceso Acheson, que es el predominante en la industria

para la síntesis de α-SiC. Consiste en la reacción entre la arena de sílice y coque del petróleo a

muy alta temperatura (>2500 ºC). El producto obtenido no tenía una alta pureza pero era útil

para aplicaciones abrasivas y de corte. Las desventajas más importantes de este proceso son la

alta energía requerida y la baja pureza del producto obtenido.

SiO2 (s) + 3C (s) → SiC (s) + 2CO (g) (Ecuación A.19)

El desarrollo de un proceso industrial para producir β-SiC atrajo una gran atención a

finales de los años 90, debido a su mayor área superficial y su posible uso como soporte

catalítico. Moene y col. sintetizaron por primera vez un carburo de silicio con áreas

superficiales de 30-80 m2/g [36-40]. El proceso se basa en una reacción sólido-gas de dos

pasos:

C(s) + H2 (g) → CH4 (g) (Ecuación A.20)

SiCl4 (g) + CH4 (g) → SiC (s) + 4HCl (Ecuación A.21)

Ledoux y col. [41] desarrollaron otro método de síntesis basado en la transformación de

un esqueleto preformado de carbón en SiC. Este proceso, llamado “shape memory process”,

genera carburo de silicio con grandes áreas superficiales (10-100 m2/g).

http://upload.wikimedia.org/wikipedia/commons/4/4f/SiC3Cstructure.jpg


25

Si + SiO2 → 2SiO (Ecuación A.22)

SiO + 2C → SiC + CO (Ecuación A.23)

CO + 2Si → SiO + SiC (Ecuación A.24)

Siguiendo esta última idea, SICAT construyó en el 2003 una planta industrial para la

síntesis de β-SiC. El proceso consiste en tres etapas:

Mezclado de silicio, carbón y una resina (metilcelulosa, polivinilalcohol, silicato

sódico) en forma de pasta.

Prensado de la pasta en la forma requerida.

Síntesis de β-SiC mediante un tratamiento térmico (<1673 K) en atmósfera

controlada.

Actualmente el β-SiC se usa como soporte en reacciones como, por ejemplo, la oxidación

directa de H2S a S por medio de oxígeno, la oxidación directa de butano en anhídrido maleico,

en reacciones fotocatalíticas, reacciones de Friedel-Craft o en las reacciones de reformado

comentadas anteriormente.

A.3.3. Aplicaciones del carburo de silicio

La estructura y composición del carburo de silicio hacen de éste un material cerámico

muy duro con propiedades eléctricas y térmicas muy interesantes. Su densidad es 1/5 de la del

wolframio, tiene una dureza en la escala Mohs de 9, próxima a la del diamante (10) y una

resistencia a la erosión de 9,15, siendo 10 la del diamante. Este material presenta asimismo una

gran pureza química y una alta resistencia a los ataques químicos, tanto de ácidos como de

bases o sales fundidas, a temperaturas de hasta 800 ºC. En atmósfera de aire sufre oxidación a

temperaturas superiores a 1600 ºC, formándose una lámina externa de SiO2 que funciona como

capa protectora frente a la oxidación del resto del material. Además presenta una excelente

conductividad térmica, comparable a la del cobre y una brecha energética estrecha (2,2 eV para

el -SiC y 3,3 para el -SiC) [34, 42], lo que indica que es un semiconductor eléctrico. Las

principales propiedades del carburo de silicio aparecen en la Tabla A.3.


26

Tabla A.3. Principales propiedades del carburo de silicio.

Propiedad Valor

Densidad (g·cm-3

) > 3,2 (20-27 ºC)

Resistencia a la flexión (MPa) > 400

Dureza Vickers (GPa) > 24

Porosidad (%) < 0,1

Pureza (%) > 97

Coeficiente expansión (ºC-1

) 2,8-4·10-6

(40-400 ºC)

Conductividad térmica (cal·cm-1

·s-1

·ºC-1

) 0,22

Módulo de compresibilidad (dyn·cm-2

) 2,5·1012

(27 ºC)

Modulo de elasticidad (GPa) > 400

Punto de fusión (ºC) 2800

Estas propiedades le convierten en un material interesante para aplicaciones en química,

física, ingeniería, medicina y biomedicina comparado con otros materiales convencionales.

Actualmente es ampliamente utilizado en abrasivos, materiales refractarios, cerámicas y

numerosas aplicaciones de alta tecnología. Su carácter de semiconductor eléctrico lo convierten

en un material excelente en la producción de aparatos de alta potencia y que trabajen a alta

temperatura, sistemas microelectromecánicos (MEMs), optoelectrónica y aplicaciones

biomecánicas [42, 43]. Además, el reciente desarrollo de la forma porosa del carburo de silicio

(β-SiC) lo hacen adecuado para su aplicación como soporte catalítico, gracias a su alta

conductividad térmica, resistencia mecánica y resistencia a la oxidación [39, 44-46].

A.3.4. Carburo de silicio como soporte catalítico

Un catalizador heterogéneo suele estar formado por una fase activa, un promotor y un

soporte. El papel del soporte es fundamental, afectando a la dispersión y estabilidad de la fase

activa. Los materiales más comúnmente usados como soportes catalíticos son la alúmina, el

óxido de silicio, el óxido de titanio, las zeolitas o soportes carbonosos. Teniendo en cuenta las

propiedades anteriormente comentadas (especialmente su alta resistencia mecánica, su gran

conductividad térmica, su resistencia a la oxidación y el ser químicamente inerte) han hecho


27

que en los últimos años el -SiC atraiga cierta atención como un material novedoso que pueda

sustituir a los materiales tradicionales como soporte catalítico en ciertas reacciones. Entre las

reacciones en las que se ha probado el -SiC como soporte catalítico se encuentran la

oxidación del H2S a azufre elemental, la oxidación de n-butano a anhídrido maleico, reacciones

de Friedel-Craft, fotocatálisis, síntesis Fischer-Tropsch y procesos de reformado. En la Tabla

A.4 se han recogido algunos de los principales trabajos científicos en los que se ha utilizado el

-SiC como soporte catalítico.


28

Tabla A.4. Publicaciones sobre la utilización de SiC como soporte catalítico.

Reacciones Catalizador Autores y año Principales conclusiones

Oxidación de

H2S

NiS2/-SiC N. Keller y col.

(2002) [47]

La alta dispersión y la falta de

microporosidad en el soporte

producen una gran actividad y

selectividad.

Fe2O3-SiC P. Nguyen y col.

(2011) [48]

Un mayor tamaño de poro en el

soporte favorece el acceso de los

reactivos a los centros activos.

Oxidación de

n-butano VPO/-SiC

M.J. Ledoux y

col. (2001) [41]

Interacción óptima entre VPO y SiC.

SiC mantiene la temperatura del

lecho catalítico uniforme.

Friedel-Crafts

BETA

zeolite/-SiC G. Winé y col.

(2007) [49]

Utilizar SiC como soporte permite

un control preciso de la estructura

macroscópica del catalizador, lo que

disminuye los problemas

difusionales.

Ga/SBA-

15/-SiC

F. Z. El Berrichi y

col. (2008) [50]

-SiC mejora la transferencia de

materia y calor, aumentando la

selectividad y estabilidad.

fotocatálisis TiO2-SiC Y. Nishida y col.

(2005) [51]

TiO2 soportado sobre-SiC muestra

mayor actividad catalítica que

soportado sobre SiO2 o no

soportado.

Síntesis FT

Fe/-SiC H. M.T. Galvis y

col. (2012) [52]

Actividad catalítica estable durante

60 h, con una selectividad a C2-C4

mayor del 50%.

Co-SiC A. R. de la Osa y

col. (2011) [53]

-SiC como soporte mejora la

conversión de CO y aumenta

selectividad a hidrocarburos pesados

de interés comparado con Al2O3.

Reformado

Ni/-SiC S. M. Kim y col.

(2012) [54]

El catalizador mostró una muy alta y

estable conversión de glicerol

durante 60 h, con una relación molar

H2/CO menor que Ni/Al2O3 o

Ni/CeO2.

Ni/-SiC J. M. García-

Vargas y col.

(2012) [55]

El catalizador soportado sobre SiC

mostró mejores propiedades que el

soportado sobre CeO2.

Existen también algunas aplicaciones en las que se requiere la combinación de varios

materiales para mejorar las características del soporte catalítico, formándose los denominados

composites. Existen numerosos composites obtenidos a partir de SiC que se han aplicado como


29

soporte catalítico para reacciones como la síntesis FT, reacciones Friedel-Crafts, síntesis de

amoniaco, procesos de reformado, pirolisis catalítica, etc. En la Tabla A.5 aparecen algunas de

las publicaciones en las que se han utilizado SiC composites como soporte.

Tabla A.5. Aplicaciones catalíticas de composites basados en SiC.

Reacciones Composite Autores y año Principales conclusiones

Deshidratación de

alcoholes

Espumas

ZSM-5/-SiC

Y. Jiao y col.

(2012) [56]

Mayor rendimiento y

estabilidad gracias a mejor

transferencia de calor y

materia.

Pirólisis catalítica Al2O3/SiC S. G. Jeon y col.

(2012) [57]

Mejor comportamiento

comparado con catalizador

no soportado o soportado

sobre SiC.

Hidrogenación de 4-

carboxybenzaldehido CDC–SiC

Y. Zhou y col.

(2012) [58]

El composite mejora la

dispersión de Pd respecto a

soportes tradicionales.

Fotocatálisis

Grafeno

cubierto de

polvo de SiC

K. Zhu y col.

(2012) [59]

Con el composite se dobla

la actividad fotocatalítica

comparado con SiC.

Reformado CeO2/Pt–SiC R. Frind y col.

(2012) [60]

Los composites con mayor

cantidad de catalizador

mostraron la mayor

estabilidad contra oxidación

y una buena actividad

catalítica.

Electrólisis de agua IrO2/SiC-Si A. V. Nikiforov y

col. (2011) [61]

Mayor actividad debido a

una mejora de las

propiedades superficiales de

IrO2.

Síntesis de amoniaco Ru-C/SiC Y. Zheng y col.

(2009) [62]

El composite preparado

mostró buenas condiciones

como soporte para la

síntesis de amoniaco.

Fischer-Tropsch Al2O3–SiC M. Lacroix y col.

(2011) [63]

Usando el composite como

soporte se obtuvieron

mayores selectividades a

C5+ que con Al2O3.


30

A.4. OBJETO Y ALCANCE DEL PRESENTE TRABAJO

En los apartados anteriores se ha puesto de manifiesto la importancia de los procesos de

producción de gas de síntesis a la hora de obtener combustibles líquidos sintéticos a partir de

fuentes renovables, como es la biomasa o el biogás. Dentro de estos procesos se han destacado

las ventajas del tri-reformado, que lo sitúan como una de las principales alternativas a la hora

de diseñar procesos que permitan la obtención de estos combustibles sintéticos de forma

competitiva.

Asimismo, el diseño de un catalizador que combine a la vez actividad, resistencia frente a

la desactivación y bajo coste es fundamental en los procesos de reformado. En este punto, no

solo es importante elegir una fase activa que combine estas propiedades, sino también buscar

materiales que cuenten con las características físico-químicas más adecuadas para el proceso.

Considerando todo lo expuesto anteriormente, se decidió estudiar el proceso de tri-

reformado de metano, optimizando el catalizador empleado en el proceso y la composición del

alimento al reactor. Además se realizó un estudio de la influencia de dicha composición sobre

diversos parámetros a diferentes temperaturas y se estableció un modelo cinético capaz de

reproducir el proceso.

Para tal fin, se planteó el siguiente programa de investigación:

Revisión bibliográfica y puesta a punto de las distintas instalaciones experimentales

(equipos de análisis, equipos de caracterización, calibración de gases, equipos de

reacción, etc.).

Estudio de la influencia del soporte catalítico en la reacción de reformado seco y el

proceso de tri-reformado.

Análisis de diferentes precursores de Ni en la actividad de catalizadores para el

proceso de tri-reformado.

Optimización de la composición del alimento para el proceso de tri-reformado,

analizando la influencia de este parámetro sobre la relación molar H2/CO del gas de

síntesis producido y el consumo energético del proceso, buscando obtener un gas de

síntesis con una relación molar H2/CO en el rango 1,9-2,1 (valor óptimo para su

aplicación en procesos Fischer-Tropsch y de producción de metanol).


31

Caracterización y estudio catalítico de catalizadores de Ni soportado sobre -SiC,

promocionados con diferentes metales alcalinos (Na y K) y alcalinotérreos (Ca y Mg),

analizando la influencia que tiene la adición de diferentes cantidades de promotor en

las propiedades del soporte y la actividad catalítica.

Análisis de la influencia del orden de impregnación promotor-fase activa en las

propiedades catalíticas de catalizadores aplicados al proceso de tri-reformado.

Modelización del proceso de tri-reformado.

B. MATERIALES Y MÉTODOS

B.1. REACTIVOS EMPLEADOS

B.1.1. Reactivos

A continuación se detallan los reactivos utilizados, indicando su concentración o pureza y

la empresa suministradora:

Nitrato amónico de cerio, (NH4)2Ce(NO3)6, pureza 99,99 %, SIGMA-ALDRICH.

Agua destilada y desionizada obtenida en nuestros laboratorios, conductividad < 10

μs.

Nitrato de níquel (II) hexahidratado, Ni(NO3)2·6H2O pureza 97%, (PANREAC).

Óxido de zirconio estabilizado con óxido de itrio al 8% (YSZ) (IONOTEC).

Carburo de silicio (SICAT CATALYST).

−alúmina (MERCK).

Acetato de níquel (II) tetrahidratado, Ni(C4H6O4)·4H2O pureza 99% (ALDRICH

CHEMISTRY).

Cloruro de níquel (II) hexahidratado, NiCl2·6H2O pureza 98% (SIGMA-ALDRICH).

Citrato de níquel (II) monohidratado, Ni3(C6H5O7)2·H2O pureza 98%, (ALFA

AESAR).

Hidróxido de calcio, Ca(OH)2 pureza 95%, (PANREAC).

Hidróxido de magnesio, Mg(OH)2 pureza 95%, (PANREAC).

Hidróxido de sodio, NaOH pureza 98%, (PANREAC).

Hidróxido de potasio, KOH pureza 95% (RIEDEL-DE HAËN).


32

B.1.2. Gases

Argón, envasado en botellas de acero a 200 bares con pureza superior al 99,996% y

suministrado por la empresa PRAXAIR.

Nitrógeno, envasado en botellas de acero a 200 bares con pureza superior al 99,999%

y suministrado por la empresa PRAXAIR.

Oxígeno, envasado en botellas de acero a 200 bares con pureza superior al 99,99% y


Helio, envasado en botellas de acero a 200 bares con pureza superior al 99,5 %,


Hidrógeno, envasado en botellas de acero a 200 bares con pureza superior al 99,999%,


Mezcla metano-nitrógeno (10% en volumen metano), envasado en botellas de

aluminio a 150 bares, suministrado por la empresa PRAXAIR.

Mezcla dióxido de carbono-nitrógeno (10% en volumen dióxido de carbono),

envasado en botellas de aluminio a 150 bares, suministrado por la empresa

PRAXAIR.

Metano, envasado en botellas de acero a 150 bares con pureza superior al 99,5%


Dióxido de carbono, envasado en botellas de aluminio a 150 bares con pureza superior

al 99,999%, suministrado por la empresa PRAXAIR.

B.2. INSTALACIONES EXPERIMENTALES

B.2.1. Preparación de catalizadores

Para preparar los catalizadores utilizados en esta tesis doctoral se seleccionó el método de

impregnación, utilizando para la preparación un rotavapor marca RESONA TECHNICS,

modelo LABO-ROTA S300.

Esta técnica consiste en poner en contacto el soporte catalítico con una disolución que

contenga un compuesto del metal que constituye la fase activa del catalizador. Para ello, en

primer lugar se coloca el material que va a ser utilizado como soporte catalítico en un matraz,

en el que se hace vacío durante 2 h gracias al rotavapor, a fin de eliminar las sustancias que

estén adsorbidas sobre el soporte. Transcurrido ese tiempo, se añade la disolución que contiene


33

el precursor de la fase activa. Posteriormente se elimina el disolvente por evaporación a vacío a

363 K durante 2 h. Tras este paso el catalizador se seca en un horno durante 12 h a 403 K.

En este trabajo se utilizaron diferentes precursores de níquel y también precursores de

metales alcalinos y alcalinotérreos, usados como promotores, como se detallará en los

siguientes capítulos.

B.2.2. Ensayos catalíticos

Los ensayos catalíticos de reformado seco y tri-reformado fueron llevados a cabo en la

instalación esquematizada en la Figura B.1.

Figura B.1. Instalación experimental para los experimentos catalíticos.

Campana

CH4

O2

N2

FIC

FIC

FIC

FIC

H2

PC-1

Caudalímetro

3 0 0

3 0 0

30

0

32

300

BROOKS

INSTRUMENT

Saturador

Microcromatógrafo de

gases

Controlador

temperatura

Peltier

Controlador

de caudal

Horno

Reactor

CO2

FIC


34

En esta instalación se pueden distinguir 3 zonas diferenciadas que se detallan a

continuación:

Sistema de alimentación.

Sistema de reacción.

Sistema de análisis.

B.2.2.1. Sistema de alimentación

El sistema de alimentación estaba constituido por cuatro líneas de flujo continuo,

análogas e independientes, para la alimentación de los gases de reacción: N2, CO2, CH4 y O2; y

una quinta línea que permitía la alimentación del H2 necesario en la etapa de reducción del

catalizador. El metano y el dióxido de carbono se alimentaban desde balas que contienen cada

gas a alta presión, procediendo el oxígeno, el nitrógeno y el hidrógeno de una canalización

general de gases. Cada línea de flujo estaba compuesta por una tubería de acero inoxidable de

1/8’’ con un manorreductor, un filtro de acero inoxidable, un controlador indicador de caudal

másico (FIC) y una válvula antirretorno.

Los controladores másicos, de la marca BROOKS INSTRUMENTS (modelo 5850E para el

nitrógeno y modelo 5850S para el resto de gases), estaban constituidos por un sensor de

conductividad térmica, un indicador controlador y una electroválvula y eran controlados

mediante un software informático.

Para introducir el vapor de agua necesario para el proceso de tri-reformado se dispuso de

un baño y un sistema de saturadores. Mediante el baño se controlaba la temperatura a la cual se

produce la saturación, y en consecuencia el porcentaje de vapor de agua introducido en la

corriente gaseosa.

B.2.2.2. Sistema de reacción

El sistema de reacción estaba constituido por un reactor tubular de lecho fijo y flujo

descendente construido en cuarzo, de 1 cm de diámetro interno y 45 cm de longitud. En la

parte superior se insertaba la conducción de entrada de los gases reaccionantes.

El termopar para medir la temperatura del lecho catalítico se ubicó de forma que el

extremo del mismo quedara situado al mismo nivel que el lecho de catalizador, situado sobre el

soporte. De este modo la temperatura medida por el termopar coincidía con la de reacción. Este


35

termopar estaba conectado a un controlador de temperaturas y horno eléctrico LENTON

THERMAL DESIGN, que permitía monitorizar y controlar la temperatura en el reactor, así

como realizar rampas de calentamiento y enfriamiento. El reactor se situó dentro del horno

eléctrico.

B.2.2.3. Sistema de análisis

A la salida del reactor, el efluente atravesaba un peltier Bühler PKE 511. En él se

condensaba el agua que quedaba sin reaccionar durante los experimentos catalíticos de tri-

reformado, separándose así de la corriente gaseosa para evitar su entrada en los equipos de

análisis, dado su perjudicial efecto en las columnas cromatográficas.

La corriente gaseosa resultante, libre de productos líquidos, se analizó en un micro

cromatógrafo de gases VARIAN CP-4900. El micro cromatógrafo de gases permitía el análisis

del H2, N2, O2, CH4, CO y CO2. Los cromatogramas obtenidos fueron almacenados y

cuantificados por el propio software informático suministrado con el equipo. En el siguiente

apartado se describe más detalladamente este equipo, así como las condiciones de análisis

empleadas.

B.3. EQUIPOS DE ANÁLISIS

B.3.1. Microcromatografía de gases

La microcromatografía de gases es una reciente tecnología que permite la miniaturización

del sistema cromatográfico. El principio básico es exactamente igual al de la cromatografía

gaseosa. Permite el análisis de muestras gaseosas o en forma de vapor, reduciendo

considerablemente el tiempo de análisis y dotando al sistema de una extraordinaria sensibilidad

debido a la utilización de columnas microcapilares.

En la presente investigación se ha empleado un microcromatógrafo de gases VARIAN

CP-4900 (Figura B.2) con dos columnas analíticas (tamiz molecular y columna poraplot Q)

conectadas a sendos detectores de conductividad térmica TCD.


36

Figura B.2.Equipo de análisis de gases.

Cada una de las columnas permite la separación y, por tanto, el análisis de determinados

componentes de la muestra. Así, el tamiz molecular permite la separación de los componentes

más ligeros de la muestra (H2, O2, N2, CH4, CO), mientras que la columna poraplot Q separa

los componentes más pesados (CO2, C2H6, C2H4 y C3H6).

Las características y condiciones de análisis de cada una de las columnas empleadas en la

presente investigación se muestran en la Tabla B.1.

Tabla B.1. Características y condiciones de análisis del microcromatógrafo de gases.

Tamiz molecular Columna poraplot Q

Longitud (m) 10 10

Diámetro de columna (mm) 0,32 0,15

Tiempo de análisis (s) 120 120

Temperatura columna (°C) 80 70

Temperatura inyector (°C) 110 110

Tiempo de inyección (ms) 100 15

Tiempo backflush (s) 5,5 0

Presión inicial (psi) 20 20


37

El análisis cuantitativo de los productos de reacción se lleva a cabo empleando los

factores de respuesta de cada uno de los componentes en el detector, obtenidos

experimentalmente mediante diferentes calibrados. La integración numérica de cada uno de los

picos se realizó con el software informático VARIAN STAR.

B.4. TÉCNICAS DE CARACTERIZACIÓN DE SOPORTES Y CATALIZADORES

B.4.1. Espectroscopía de emisión atómica de inducción de plasma acoplada (ICP-AES)

El contenido metálico de los catalizadores se determinó mediante espectrofotometría de

absorción atómica, usando para ello un equipo VARIAN, modelo SpectrAA 220.

La espectroscopía de emisión atómica se fundamenta en la excitación de los átomos

metálicos mediante un plasma de argón, capaz de alcanzar 10000 K, asegurando la completa

atomización de la muestra en estado líquido. Al cesar la excitación, tiene lugar la emisión de

radiación por parte del metal para volver al estado energético fundamental. La intensidad de

dicha emisión permite cuantificar la concentración del elemento ya que depende de la cantidad

de átomos del mismo. La ventaja principal del plasma es la alta temperatura alcanzable, que

asegura la completa atomización de la muestra.

Previamente al análisis, las muestras sólidas fueron sometidas a un tratamiento de

digestión ácida con ácido clorhídrico, peróxido de hidrógeno y fluorhídrico para conseguir la

disolución de los metales. Para luego determinar su concentración, era necesario obtener

previamente las curvas de calibración correspondientes a cada metal en el intervalo de

concentración adecuado. Para cada muestra se realizaron cinco puntos de calibración. Las

disoluciones se prepararon a partir de disoluciones patrón certificadas para análisis de emisión

atómica de 1000 mg/L en medio ácido nítrico.

B.4.2. Adsorción-desorción de nitrógeno

Los análisis de área superficial y volumen total de poros se realizaron mediante

adsorción-desorción de N2 a 77 K, empleando un equipo de la marca QUANTACHROME,

modelo QUADRASORB 3SI, con seis puertos de desgasificación y tres de análisis, con un

software que recogía los valores de presión relativa para cada volumen de N2 dosificado.

Esta técnica se basa en la adsorción-desorción física de gases (adsorbatos) en sólidos

(adsorbentes). Cuando una cierta cantidad de adsorbente se pone en contacto con un volumen


38

dado de una mezcla gaseosa que contiene el soluto a adsorber, se produce la retención de

soluto en la superficie del sólido acompañada de una disminución de la concentración del

mismo en la mezcla, hasta alcanzar el equilibrio de adsorción. El adsorbato retenido puede ser

posteriormente desorbido del adsorbente por una corriente de gas caliente o por reducción de la

presión. Si se mantiene constante la temperatura, la relación entre la cantidad de soluto

adsorbida y la concentración en la disolución se denomina isoterma de adsorción-desorción.

Dicha isoterma puede determinarse de manera volumétrica, calculando la cantidad adsorbida

mediante la aplicación de las leyes de los gases a la presión y volumen de adsorbato antes y

después de la adsorción-desorción.

Mediante las isotermas adsorción y desorción de un gas inerte en la superficie del

material y el tratamiento matemático con diferentes modelos de adsorción, se pueden estimar

parámetros texturales tales como la superficie específica, el volumen de poros, el diámetro de

poro medio, etc.

Previamente al análisis, las muestras se desgasificaron a 453 K aplicando vacío durante 9

horas. Una vez desgasificadas las muestras, la adsorción se realizó añadiendo cantidades

crecientes de nitrógeno, para abarcar todo el intervalo de presiones relativas hasta aproximarse

a la saturación (P/P0 = 0,995). Alcanzada la saturación, la desorción se llevó a cabo por vacío,

reduciendo la presión relativa escalonadamente. El tiempo de estabilización de cada medida se

fijó en 5 s.

La superficie específica se estimó utilizando el modelo Brunauer Emmett-Teller (BET)

aplicado a la rama de adsorción de nitrógeno en el intervalo de presiones parciales

seleccionado para cada catalizador, de forma que no se produzca en ningún momento

condensación capilar en mesoporos.

B.4.3. Difracción de rayos X

Los ensayos de difracción de rayos X se llevaron a cabo en un difractómetro PHILIPS

modelo PW-1711, con radiación CuKα (λ = 1,5404 Å).

La difracción de rayos X aporta información directa de la estructura ordenada de los

materiales. Los rayos X son una radiación de longitud de onda comprendida entre 10-3

y 100 Å

producida por el frenado de electrones de elevada energía o por transiciones electrónicas entre

niveles atómicos internos. Cuando los rayos X son dispersados por un entorno ordenado, tiene


39

lugar la difracción, debido a que las distancias entre los centros de dispersión son del mismo

orden que la longitud de onda de la radiación incidente. La difracción se produce por efecto

acumulativo de la dispersión generada cuando el haz de rayos X interacciona con las diferentes

capas ordenadas que se encuentran a la misma distancia. Para observar la difracción, se

requiere que los centros de dispersión estén distribuidos en el espacio de manera regular,

formando planos con orientaciones específicas. Además, es necesaria una distancia similar

entre los planos responsables de la difracción que constituyen una familia de planos. Así, cada

familia de planos con la misma orientación espacial da lugar a una señal de difracción si se

cumple la Ley de Bragg (Ecuación B.1):

n· = 2·dhkl·sen (Ecuación B.1)

donde n es el orden de difracción, λ es la longitud de onda de la radiación incidente, dhkl

es la distancia interplanar correspondiente a cada familia de planos denotadas por los índices de

Miller correspondientes (h, k, l) y θ es el ángulo de difracción.

La información directa que proporcionan estos experimentos es el ángulo (θ) y la

intensidad de los haces difractados a lo largo de cualquier círculo alrededor de la muestra,

teniendo un diámetro contenido en la recta definida por el haz incidente.

El método está basado en el ensanchamiento de las líneas de difracción que se produce al

disminuir el tamaño medio de partícula por debajo de 1000 Å. La relación entre el tamaño

medio de partícula y el ensanchamiento medio que produce el pico de difracción elegido a la

altura media del mismo, puede establecerse mediante la ecuación de Debye-Scherrer (Ecuación

B.2):

λ

β θ (Ecuación B.2)

siendo d, el diámetro medio de la partícula en la dirección normal a los planos que

difractan la radiación; λ, la longitud de onda de radiación; θ, el ángulo de difracción

correspondiente al pico considerado; y K, una constante de proporcionalidad cuyo valor

depende principalmente de la magnitud adoptada para la definición de la anchura total

observada en el pico de difracción. En la práctica, el ensanchamiento total observado en la

línea de difracción, B, está relacionado con β mediante la ecuación B.3, donde b es el

ensanchamiento debido a efectos instrumentales del difractómetro.


40

β (Ecuación B.3)

B.4.4. Reducción a temperatura programada

Los análisis de reducción a temperatura programada (TPR), se realizaron en un equipo

Micromeritics (Modelo TPD/TPR 2900 Analyzer), que consta de un sistema de control de

temperatura de las líneas de gases, detector de conductividad térmica (TCD), un sistema para la

adquisición y manipulación de los datos Micromeritics 2900, válvulas de gases, un sistema de

control de temperatura del horno de calefacción, medidores de flujo, panel de control de

presión y flujo de gases y un bucle calibrado para la inyección controlada de distintos gases a

la muestra.

La reducción a temperatura programada es una técnica extremadamente sensible, que

permite estudiar el proceso de reducción de un catalizador o precursor reducible al exponerlo a

un flujo de una mezcla gaseosa reductora (típicamente un pequeño %vol de H2 en un gas

inerte), mientras se aumenta la temperatura linealmente. El principio de operación de la técnica

de TPR se basa en el cambio químico que experimenta un sistema redox cuando se expone a un

ambiente reductor. Dado que los materiales a caracterizar son óxidos, en presencia de H2 se

reducen para obtener el correspondiente metal según la reacción (Ecuación B.4):

(Ecuación B.4)

El proceso de reducción se sigue continuamente por medida de la composición (contenido

en H2) de la mezcla gaseosa reductora a la salida del reactor. Para la realización de los análisis,

en primer lugar, se coloca el sólido en un reactor tubular de lecho fijo fabricado en cuarzo y,

previamente al análisis, se desgasifica para eliminar las sustancias adsorbidas físicamente.

Posteriormente, se enfría la muestra y se procede a la realización del análisis, haciendo pasar

una corriente de gas reductor (17 % vol de H2 en Ar, 100 mL min-1

) con un calentamiento hasta

1173 K a una velocidad linealmente programada de 10 K min-1

. Antes de alcanzar el detector,

para eliminar el agua formada y poder realizar el seguimiento de la cantidad de H2 consumido

en cada momento, el gas efluente se hace pasar a través de una trampa fría, consistente en una

mezcla frigorífica de isopropanol y nitrógeno líquido a una temperatura de 193 K.


41

B.4.5. Microscopía electrónica de transmisión

La microscopía electrónica de transmisión utiliza un haz de electrones que, manejado a

través de lentes electromagnéticas, se proyecta sobre una muestra muy delgada, situada en una

columna de alto vacío. El haz de electrones atraviesa la muestra, que ha sido contrastada con

átomos pesados. Se pueden dar dos situaciones básicas: que los electrones del haz atraviesen la

muestra o que choquen con un átomo de ésta. De este modo, se obtiene información estructural

específica de la muestra según las pérdidas específicas de los diferentes electrones del haz. El

conjunto de electrones que atraviesan la muestra son proyectados sobre una pantalla

fluorescente, formando una imagen visible, o sobre una placa fotográfica, registrando una

imagen latente.

Los experimentos de microscopía electrónica de transmisión (TEM) se realizaron en un

equipo JEOL JEM-4000EX con un voltaje de aceleración de 400 kV. Las muestras se

prepararon mediante dispersión ultrasónica en acetona. Una gota de la suspensión se evaporaba

sobre una malla de carbono. La distribución de tamaños de partícula metálica se evaluó

mediante el cálculo del diámetro medio basado en la superficie (Ecuación B.5):

2

ii

i

3

ii

sdn

dn

d

(Ecuación B.5)

siendo ni el número de partículas con diámetro di (Σi ni ≥ 400).

B.4.6. Desorción de dióxido de carbono a temperatura programada

El equipo utilizado para la desorción a temperatura programada o TPD, es el mismo que

el equipo de reducción a temperatura programada o TPR, descrito en el apartado B.4.4.

La desorción a temperatura programada se basa en la quimisorción de un gas o un líquido

sobre un sólido y su posterior desorción mediante aumento de la temperatura. La cantidad de

especies adsorbidas se puede evaluar mediante diferentes tipos de detectores, ya sea un TCD

(detector de conductividad térmica) o un espectrómetro de masas. En función de las

características de la superficie, el gas se puede adsorber dando lugar a distintas especies, de

manera que la desorción se producirá a diferentes temperaturas, según la fuerza de la

interacción entre el gas y el centro en cuestión.


42

Cuando las posiciones activas de la superficie del sólido adsorben gas, se forman varias

capas de adsorción, dándose la quimisorción entre la primera capa de gas y la superficie del

sólido. Por tanto, antes de realizar un experimento, se deben eliminar las multicapas formadas

por fisisorción calentando a bajas temperaturas. Posteriormente, se calienta progresivamente la

muestra de forma que se va produciendo la eliminación de las especies adsorbidas, las cuales se

conducen hasta un detector, pudiendo así obtener un registro de las especies desorbidas en

función de la temperatura. Previamente al proceso de adsorción-desorción se redujo el

catalizador bajo las mismas condiciones que antes de la reacción. Una vez adsorbido el CO2, se

calentó la muestra a una velocidad de 10 K min-1

hasta 1173 K, haciéndo pasar una corriente de

gas inerte (He).

B.4.7. Oxidación a temperatura programada

El equipo utilizado para la oxidación a temperatura programada o TPO, es el mismo que

el equipo utilizado en la reducción a temperatura programada o TPR y la desorción a

temperatura programada o TPD, descrito por tanto en el apartado B.4.4.

La técnica TPO se utiliza, entre otras aplicaciones, para determinar los depósitos

orgánicos adsorbidos en el catalizador tras ser utilizado en una prueba experimental.

Tras la colocación de la muestra dentro del reactor (aproximadamente 150 mg) se

programa el software con una rampa de calentamiento de 10 K min-1

hasta 1173 K. A

continuación se conecta un flujo continuo de gas (O2), que cumple dos funciones: proporcionar

el reactivo oxidante para transformar las especies adsorbidas y arrastrar todos los compuestos

que se desprendan de la muestra hacia la entrada del sensor de conductividad térmica.

Finalmente, en el ordenador se registra una señal invertida del oxígeno consumido. El

consumo de O2 desde los perfiles de TPO se obtuvo por integración de los picos de oxidación.

B.4.8. Quimisorción estática de hidrógeno

La quimisorción es una técnica que permite obtener el grado de dispersión de un metal en

un soporte a partir de la cantidad de gas quimisorbido en los centros metálicos. Así, la

dispersión se calcula mediante la expresión:

(Ecuación B.6)


43

donde: D es la dispersión metálica (%) y f el factor estequiométrico, cuyo valor para el H2

es de 2 [64].

El equipo utilizado para realizar las medidas de quimisorción estática con H2 fue un

MICROMERITICS, modelo ASAP 2010 Sorptometer, dotado de una unidad de quimisorción.

El procedimiento llevado a cabo consistía, en primer lugar, en la reducción del catalizador para

obtener los centros metálicos donde se va a quimisorber el H2. Para ello se calentó la muestra

en una corriente de helio hasta la temperatura de reducción (dependiente del catalizador

empleado) y posteriormente el flujo de helio se sustituía por hidrógeno para reducir la muestra

durante 2 h. Esta etapa era seguida de una etapa de desgasificación a la temperatura de

reacción, en la cual se aplicaba vacío durante 30 min. Hecho esto, la muestra se enfrió hasta

308 K y se realizó una nueva desgasificación a esta temperatura durante otros 30 min. Una vez

realizadas estas etapas previas se procedía a analizar la cantidad de H2 quimisorbido por las

muestras, mediante el análisis de la cantidad de H2 adsorbido a diferentes presiones,

manteniendo constante la temperatura. Se llevó a cabo una primera isoterma de adsorción de

H2, y tras una nueva evacuación de la muestra a 308 K, se realizó una segunda isoterma. La

primera isoterma hace referencia a la cantidad de H2 retenido por la muestra de forma

reversible (fisisorbido) e irreversible (quimisorbido), mientras que la segunda, solamente

representa la cantidad de H2 adsorbida de forma reversible (fisisorbido). Así, la cantidad de H2

quimisorbida se calculó como la diferencia entre las dos isotermas [65].

B.4.9. Espectroscopía Raman

Esta técnica de caracterización se basa en el estudio de la luz que dispersa un material

cuando se hace incidir sobre él un haz de luz monocromático (radiación láser). Una porción de

la luz es dispersada inelásticamente, sufriendo ligeros cambios de frecuencia que resultan ser

característicos del material analizado e independiente de la frecuencia de la luz incidente.

El equipo Raman utilizado (modelo BRUKER SENTERRA) utilizaba una radiación

infrarroja con una longitud de onda láser de excitación de 532 nm, analizando el espectro en el

rango 0-800 cm-1

.


44

C. RESULTADOS Y DISCUSIÓN

En este trabajo se ha estudiado el proceso catalítico del tri-reformado de metano

utilizando catalizadores de níquel, analizando: la influencia del soporte, el precursor de níquel

y la adición de diferentes promotores. Además se ha evaluado la influencia de la composición

del alimento y se ha modelizado el proceso de tri-reformado de metano.

En cuanto a los soportes, en una primera etapa se estudiaron diferentes materiales, tanto

convencionales como no convencionales. Dentro de estos últimos el carburo de silicio mostró

unas propiedades muy interesantes, lo que unido a lo novedoso que resultaba su estudio hizo

que se eligiese como soporte catalítico para gran parte de los estudios llevados a cabo.

En el Capítulo 1 se muestran los resultados relativos al estudio de la influencia de

diferentes materiales como soportes catalíticos, para la reacción de reformado seco y el proceso

de tri-reformado. Se eligieron como soportes catalíticos materiales que son habitualmente

utilizados en estudios de reformado, como la gamma-alúmina (-Al2O3); y otros que no lo son

tanto, como el óxido de cerio (CeO2), el beta carburo de silicio (-SiC) y el óxido de zirconio

estabilizado con itrio (YSZ). No sólo se estudiaron las diferencias obtenidas en el

comportamiento catalítico al utilizar varios materiales como soporte, sino que también se

analizaron las diferencias obtenidas al someter el catalizador preparado con YSZ a dos

tratamientos diferentes de calcinación. Uno de los catalizadores fue preparado siguiendo el

mismo procedimiento de calcinación que para el resto de catalizadores, mientras que el otro fue

preparado calcinándolo en una atmósfera pobre en oxígeno.

La caracterización llevada a cabo mostró sensibles diferencias entre los catalizadores

preparados. Los resultados de reducción a temperatura programada indicaron una gran

dependencia de la temperatura a la que se producía la reducción del níquel en función del

soporte utilizado. De este modo, el catalizador Ni/Al2O3 mostró el pico de reducción a mayor

temperatura, probablemente debido a la formación de compuestos tipo aluminato (NiAl2O4)

[66, 67] durante la etapa de calcinación. También se observaron diferencias entre los dos

catalizadores soportados sobre YSZ, preparados con diferente método de calcinación. En

ambos casos se observaban dos picos principales de reducción. El obtenido a temperatura más

baja se asignó a la reducción de NiO con poca interacción con el soporte, mientras que el

obtenido a mayor temperatura se atribuyó a la reducción del NiO que interaccionaba más


45

fuertemente con el soporte. Estos picos se relacionaron con los picos I y II observados por

Mori y col. [68] para un catalizador Ni/YSZ preparado por impregnación.

Comparando ambos catalizadores se pudo observar cómo el catalizador Ni/YSZ-O2,

preparado al calcinar en atmósfera pobre en O2, mostraba una mayor cantidad de especies

fácilmente reducibles, ya que la relación entre el tamaño del pico a baja temperatura con el

tamaño del pico a alta temperatura era mayor, llegando a desaparecer prácticamente este

segundo pico. Esta mayor reducibilidad del catalizador calcinado en atmósfera pobre en

oxígeno parecía estar relacionada con una mayor cantidad de vacantes de oxígeno en la

superficie de la YSZ [69]. Las vacantes de oxígeno aparecieron en la matriz de YSZ debido a

la sustitución de los cationes Zr4+

por cationes Y3+

, lo que provocaba que aparecieran estas

vacantes para mantener la neutralidad eléctrica. La presencia de un mayor número de vacantes

de oxígeno en el catalizador Ni/YSZ-O2 fue demostrada mediante experimentos de

espectrometría Raman en los que se analizó la anchura del pico obtenido a 260 cm-1

, ya que

una mayor anchura de pico está relacionada con una mayor cantidad de estas vacantes.

La basicidad de los catalizadores fue analizada mediante experimentos de desorción a

temperatura programada de dióxido de carbono. El catalizador Ni/Al2O3 mostró una gran

cantidad de centros básicos a baja temperatura (300-650 K) pero ningún pico de desorción en

el rango de alta temperatura (650-1100 K), lo que indicaba que sus centros de adsorción eran

de basicidad débil. El catalizador Ni/CeO2 fue el que mayor cantidad de CO2 adsorbió,

observándose picos de desorción tanto en el rango de basicidad débil como en el de basicidad

fuerte, mientras que para el catalizador Ni/SiC no se observaron picos notables de desorción.

En cuanto a los catalizadores preparados sobre YSZ, éstos mostraron un perfil de desorción

similar, aunque con una menor cantidad de CO2 desorbido en el caso del catalizador Ni/YSZ-

O2.

La actividad catalítica de los catalizadores preparados fue analizada para la reacción de

reformado seco. Se observó una baja velocidad de reacción tanto de metano como de dióxido

de carbono para el catalizador Ni/Al2O3, probablemente debido a un menor grado de reducción,

lo que implicó una menor presencia de especies Ni0, la fase activa en los procesos de

reformado [70]. Las muestras Ni/CeO2 y Ni/SiC mostraron una mayor actividad catalítica, con

velocidades de reacción para el CO2 bastante similares a pesar de la diferencia observada en la

capacidad de adsorción de CO2. Al comparar los catalizadores soportados sobre YSZ se


46

apreció cómo el catalizador Ni/YSZ mostró una menor actividad catalítica que el catalizador

Ni/YSZ-O2, lo cual parecía estar relacionado con la mayor presencia de vacantes de oxígeno en

este catalizador. Estas vacantes son capaces de promocionar la actividad en reformado seco

debido a que favorecen la reactividad de la molécula de CO2 [71].

Por último se realizaron experimentos para comprobar la actividad catalítica de los

catalizadores preparados en el proceso de tri-reformado. Los catalizadores preparados usando

CeO2 y SiC como soportes mostraron los mayores valores de velocidad de conversión de

metano, siendo estos valores bastante estables durante la duración del experimento. Estos

catalizadores mostraron grandes diferencias en cuanto a la velocidad de conversión de CO2,

siendo mayor para el catalizador Ni/CeO2. Este hecho se relacionaba con una mayor

contribución del reformado seco al proceso global del tri-reformado para este catalizador, lo

cual estaba relacionado con la mayor basicidad de este catalizador mostrada por los

experimentos TPD-CO2. Esta mayor contribución del reformado seco se vio reflejada en la

relación molar H2/CO, ya que la estequiometría de esta reacción hace que se genere menos H2

y más CO en comparación con el resto de reacciones de reformado que conforman el proceso

de tri-reformado. Así, se obtuvo una menor relación molar H2/CO para el catalizador Ni/CeO2

comparado con el catalizador Ni/SiC. El catalizador Ni/Al2O3 mostró una baja velocidad de

reacción de CO2 y una alta relación H2/CO, a pesar de la gran cantidad de centros de basicidad

débil que se observaron para este catalizador. Este comportamiento se debía a que los centros

de basicidad débil no tenían una gran influencia en el reformado seco, ya que no poseen

suficiente influencia para cambiar el carácter ácido-base [72], y a la gran cantidad de centros

ácidos que presenta la alúmina [73, 74]. Los catalizadores Ni/YSZ y Ni/YSZ-O2 volvieron a

mostrar sensibles diferencias, con una mayor actividad catalítica y una menor relación molar

H2/CO para el catalizador calcinado en atmósfera pobre en oxígeno. Estas diferencias son

atribuidas a la mayor reducibilidad y presencia de vacantes de oxígeno de este último

catalizador, lo que favorecía la presencia de la fase activa del Ni y la mayor reactividad del

CO2, lo que aumentaba la contribución del reformado seco al proceso global del tri-reformado.

En el Capítulo 2 se analizó la influencia del precursor de Ni en el proceso de tri-

reformado utilizando CeO2 y -SiC como soportes. Para ello se seleccionaron cuatro

precursores diferentes (nitrato, acetato, cloruro y citrato de níquel) y se prepararon ocho

catalizadores, uno con cada precursor sobre cada uno de los dos soportes.


47

La caracterización realizada mostró una gran dependencia del tamaño de partícula de

níquel con el precursor empleado en la preparación del catalizador. El tamaño de partícula fue

calculado a partir de los datos de difracción de rayos X y de las imágenes TEM. El orden

obtenido en los tamaños de partícula de níquel fue el mismo para los dos soportes,

obteniéndose los menores tamaños de partícula para los catalizadores preparados con nitrato y

acetato de níquel, mientras que los preparados con cloruro y citrato de níquel produjeron

catalizadores con un mayor tamaño de partícula del metal. La gran volatilidad del cloruro de

níquel en presencia de hidrógeno pudo ser la responsable del mayor tamaño obtenido al

emplear este precursor, ya que durante las primeras etapas de la reducción del catalizador se

pudo producir la vaporización y posterior depósito de las partículas de níquel [75]. En el caso

del citrato de níquel, la mayor acidez de la disolución preparada durante el proceso de

impregnación provocó que se obtuviesen partículas de níquel con un mayor tamaño, debido a

la mayor dispersión del Ni en medios básicos [76].

Los experimentos de reducción a temperatura programada no mostraron diferencias en el

perfil de reducción para los catalizadores soportados sobre CeO2 en función del precursor

seleccionado, pero sí para el caso de los catalizadores soportados sobre SiC. Estos

experimentos mostraron una mayor interacción entre el Ni y el SiC, en el caso de los

catalizadores preparados utilizando nitrato y acetato como promotor, ya que se observó un

desplazamiento de los picos de reducción a temperaturas mayores. Los perfiles de reducción

también indicaban que el Ni interaccionaba más fuertemente con el SiC que con el CeO2, ya

que el primer pico de reducción, que indica la reducción de níquel interaccionando débilmente

con el soporte, aparecía a temperaturas más bajas para este último material. Los experimentos

de desorción a temperatura programada de CO2 mostraron grandes diferencias en la capacidad

de adsorción de CO2 para los catalizadores soportados sobre CeO2 respecto a los soportados

sobre SiC, siendo mucho mayor para los primeros, como ya ha sido comentado para el anterior

capítulo.

Los experimentos de reacción mostraron que, para los catalizadores soportados sobre

CeO2, el mejor precursor era el nitrato de níquel, ya que con él se obtuvieron los mejores

resultados de actividad catalítica, que además apenas disminuyó durante el tiempo del

experimento. En cambio, el catalizador preparado utilizando acetato de níquel mostró el menor

valor de velocidad de reacción de metano. Los catalizadores soportados sobre CeO2, cuando se

utilizó cloruro o citrato de níquel como precursor, mostraron buenos valores iniciales de


48

actividad catalítica, pero sufrían una clara desactivación a lo largo del experimento. En el caso

del catalizador obtenido a partir de cloruro de níquel la desactivación parecía estar relacionada

con la presencia de iones cloruro en la superficie del catalizador y el mayor tamaño de las

partículas de metal.

En el caso de los catalizadores soportados sobre SiC, el catalizador obtenido a partir de

cloruro de níquel mostró también una considerable desactivación, apareciendo además ciclos

de actividad que parecían estar relacionados con el estado de oxidación del níquel [77, 78]. Los

catalizadores preparados utilizando nitrato y acetato como precursores mostraron los mejores

valores de actividad en este grupo, siendo además muy estables. El catalizador obtenido a partir

de citrato de níquel presentó una menor actividad inicial y además se observó una ligera

desactivación del mismo.

Al comparar los resultados obtenidos en función del soporte se observó cómo los

catalizadores soportados sobre CeO2 mostraron una mayor desactivación, lo que parecía estar

relacionado con la menor interacción metal-soporte que presentaban respecto a los

catalizadores soportados sobre SiC [76, 79], ya que éste ha sido señalado como uno de los

factores principales que determinan la desactivación que sufre un catalizador en los procesos

de reformado. En cuanto a la relación molar H2/CO, se pudo apreciar la influencia que el

soporte tuvo sobre este parámetro, ya que los catalizadores soportados sobre CeO2 mostraron

un menor valor del mismo. Como se ha comentado anteriormente, los catalizadores soportados

sobre CeO2 mostraron una mayor capacidad de adsorción de CO2, lo que estaba relacionado

con una mayor contribución del reformado seco al proceso de tri-reformado, implicando una

menor producción de H2 y, por lo tanto, una menor relación molar H2/CO. Este parámetro

también estaba vinculando con el tamaño de partícula de níquel, ya que un mayor tamaño

favorece la reacción de craqueo de metano [80], lo cual hace que aumente la relación molar

H2/CO.

En el Capítulo 3 se estudió la influencia de la composición del alimento sobre la

conversión de metano y la relación molar H2/CO del gas de síntesis producido en el proceso de

tri-reformado. Para ello se seleccionó como catalizador uno de los anteriormente estudiados, en

concreto aquel en el que se utilizó -SiC como soporte y nitrato de níquel como precursor. Para

estudiar esta influencia se eligió la metodología del diseño factorial de experimentos, ya que

permite obtener el máximo de información posible con la mínima cantidad de experimentos. Se


49

seleccionaron como variables independientes los caudales de CH4, CO2, H2O y O2; y como

variables dependientes la conversión de CH4 y la relación molar H2/CO. A cada una de las

variables independientes se le asignaron dos niveles, realizándose 4 replicaciones del punto

central, lo que originaba en una primera etapa un total de 20 experimentos. Posteriormente fue

ampliado este diseño factorial para analizar la curvatura del modelo mediante la adición de

puntos estrella, lo cual añadía otros 8 experimentos.

En primer lugar se analizó la influencia de las variables independientes sobre las

dependientes mediante un análisis estadístico con un nivel de confianza del 95%. Se observó

que, para la conversión de metano y al nivel de confianza elegido, solamente la influencia del

caudal de O2 es estadísticamente significativa entre las interacciones directas. Sin embargo, en

el caso de la relación molar H2/CO las cuatro variables independientes resultaron

estadísticamente significativas. Así, un aumento del caudal de metano o del caudal de dióxido

de carbono produjo una disminución de la relación molar H2/CO, mientras que un aumento del

caudal de agua u oxígeno provocó un aumento de dicha relación molar. Dado que la influencia

de las variables independientes sobre la conversión de metano era despreciable, se continuó el

estudio considerando como única variable dependiente la relación molar H2/CO. Para

determinar si el sistema podía ser descrito mediante un modelo lineal o era necesario un

modelo cuadrático se estudió la curvatura de los datos obtenidos, obteniéndose que este valor

era superior al del intervalo de confianza, lo que indicaba la necesidad de utilizar un modelo

cuadrático.

A partir del análisis estadístico efectuado se calculó un modelo cuadrático, capaz de

predecir la relación molar H2/CO del gas de síntesis obtenido mediante el tri-reformado de

metano, en función de los caudales molares de cada uno de los compuestos del alimento. Con

este modelo se realizaron representaciones 3D para poder determinar el rango de

composiciones del alimento que producían un gas de síntesis con una relación molar H2/CO

adecuada para la producción de diésel mediante síntesis FT (1,9-2,1). De este análisis se pudo

deducir que el aumento en el caudal de metano del alimento permitía un mayor rango de

caudales para el resto de compuestos. Esto era debido a que, cuando hay poco CH4 en el medio

de reacción, el CO2 debe competir con el H2O y el O2 por reaccionar con él, pero debido a que

termodinámicamente se encuentra favorecida la reacción de CH4 con H2O [26] el reformado

seco pierde importancia dentro del proceso global de tri-reformado, lo que hace que la relación

molar H2/CO aumente por encima del límite deseado. A pesar de lo comentado anteriormente,


50

a caudales muy altos de metano aparecía una región para valores bajos del resto de caudales

que no cumplía con el valor deseado de relación molar H2/CO, probablemente debido a que la

poca presencia de reactivos y el exceso de metano permitía el craqueo de éste, lo que

solamente produce H2 y no CO, por lo que la relación molar H2/CO obtenida es superior al

valor deseado.

En el caso del caudal de agua, al aumentar éste disminuyó el rango de caudales del resto

de compuestos que permiten obtener el gas de síntesis buscado, lo que viene determinado por

la estequiometria de la reacción de reformado con vapor, que hace que aumente sensiblemente

la relación molar H2/CO al aumentar la contribución de esta reacción al proceso global del tri-

reformado. El caudal de oxígeno tuvo una influencia similar a la del caudal de agua sobre la

relación molar H2/CO. Al aumentar este caudal también se produce una disminución del rango

de caudales válido para el resto de compuestos, lo que está relacionado con la preeminencia de

la reacción de oxidación parcial sobre el resto, y la competencia por el metano disponible entre

el agua y el dióxido de carbono, que como se ha justificado anteriormente hace disminuir la

contribución de la reacción de reformado seco al proceso global, aumentando el valor de la

relación molar H2/CO. En cuanto al caudal de CO2, un aumento del mismo aumenta el rango

admisible para el resto de caudales. A altos caudales de CO2 y baja cantidad de H2O y O2 se

observó una pequeña región para la cual el valor de la relación molar H2/CO estaba por debajo

del rango deseado, debido a la preponderancia de la reacción de reformado seco.

Por último se buscó un óptimo energético dentro de la región de caudales que permitieran

obtener un gas de síntesis con una relación H2/CO adecuada para la producción de diésel

mediante el proceso FT. Para ello se calcularon los calores de reacción para cada uno de los

experimentos mediante el simulador Aspen HYSYS. Las condiciones óptimas implicaban

valores altos para los caudales de oxígeno y dióxido de carbono, y valores bajos para los

caudales de metano y agua. Para comprobar la eficacia del modelo desarrollado, se realizó un

experimento con los caudales de alimento determinados mediante la optimización anterior. Se

comprobó que la relación molar H2/CO obtenida experimentalmente estaba muy próxima al

valor determinado por el modelo, lo que confirmó la valía del mismo.

En el Capítulo 4 se buscó aumentar la resistencia del catalizador frente a la desactivación

por coque. Para ello se añadieron distintas cantidades de promotores alcalinos (Na, K) y

alcalinotérreos (Mg, Ca) al catalizador de Ni soportado sobre -SiC, manteniendo una relación


51

molar Ni/M de 10/1 ó 2/1, siendo M el promotor seleccionado en cada caso. Una vez

preparados los catalizadores, éstos fueron probados en la reacción de tri-reformado durante 24

horas, analizando la cantidad de coque depositada en cada uno de ellos mediante análisis de

oxidación a temperatura programada. Posteriormente se analizó la influencia de la cantidad de

Mg en la actividad catalítica y la estabilidad del catalizador, preparando dos catalizadores más

con relaciones molares Ni/Mg de 4/1 y 1/1.

Los análisis de difracción de rayos X mostraron cómo los catalizadores promocionados

con Na y K sufrieron una pérdida de la estructura del -SiC y la aparición de -cristobalita,

una de las fases del óxido de silicio. La transformación del -SiC en -cristobalita se debe a un

proceso de oxidación ocurrido probablemente durante la calcinación del catalizador, proceso

que se ve acelerado debido a la presencia de estos dos metales alcalinos. Este cambio en la

estructura del soporte catalítico también fue observado al analizar los resultados de las pruebas

de adsorción de N2, ya que el valor de área superficial de los catalizadores promocionados con

Na y K fue mucho menor que el valor para el soporte, debido al colapso de la estructura porosa

del -SiC y la mayor cristalinidad de la -cristobalita. En el caso de los catalizadores

promocionados con Ca, se observó la aparición de cuarzo en la muestra preparada con una

mayor cantidad de Ca, lo que indicaba que grandes cargas de este metal alcalinotérreo también

favorecían la oxidación del -SiC. Tanto el catalizador promocionado con baja cantidad de Ca,

como los catalizadores promocionados con Mg, mostraron únicamente los picos de difracción

correspondientes al -SiC y a los diferentes estados de oxidación del Ni.

Los experimentos de reducción a temperatura programada llevados a cabo indicaron que

el colapso del -SiC y posterior aparición de -cristobalita impedían el acceso de las moléculas

de H2 a gran parte del níquel presente en el catalizador, ya que los perfiles de reducción

mostraron picos muy anchos y poco definidos, lo cual también ocurría en el caso del

catalizador promocionado con una elevada cantidad de Ca. Para el resto de los catalizadores, el

promocionado con Ca en baja cantidad y los dos promocionados con Mg, se observó un

desplazamiento de los picos de reducción a temperaturas más altas comparado con el

catalizador Ni/-SiC sin promocionar, lo que se relacionó con una mayor interacción del Ni

con el sistema soporte-promotor. El catalizador promocionado con una elevada cantidad de Mg

mostraba un pico de reducción a muy alta temperatura, probablemente debido a la formación

de una disolución sólida NiO-MgO, fase que suele aparecer en procesos de alta temperatura

donde están presentes Ni y Mg [81, 82].


52

Solamente el catalizador promocionado con baja cantidad de calcio y los dos

catalizadores promocionados con magnesio fueron probados en el tri-reformado, debido a los

cambios sufridos por el soporte en el resto de catalizadores. Estos tres catalizadores

promocionados mejoraron su estabilidad con respecto al catalizador de referencia Ni/-SiC, en

experimentos de 24 horas de duración. En cuanto a la velocidad de reacción de metano, los

catalizadores de Mg mejoraron el comportamiento del catalizador de referencia, mientras que

el catalizador promocionado con Ca mostró valores similares al mismo.

Como se ha comentado anteriormente, visto que los catalizadores con Mg mostraron

buenas propiedades, se decidió ampliar el estudio con dos nuevas muestras promocionadas con

Mg, para poder analizar la influencia de la cantidad de este metal sobre el comportamiento

catalítico. Se pudo observar que la adición de pequeñas cantidades de Mg provocaba la

aparición de dos picos de reducción a temperaturas bajas, mientras que el catalizador

promocionado con mayor cantidad de magnesio mostró un único pico de reducción a alta

temperatura. Esto es debido, como se ha comentado anteriormente, a la formación de una

disolución sólida entre el NiO y el MgO, dando lugar a una especie que es difícilmente

reducible. La formación de esta especie se vio favorecida por la adición de Mg, y ha sido

confirmada mediante difracción de rayos X, ya que provoca un desplazamiento del pico

principal característico del NiO hacia valores menores de ángulo de difracción. El tamaño de

partícula de níquel, calculado mediante los datos de difracción de rayos X, estaba influido por

la cantidad de magnesio añadido, obteniéndose valores más pequeños para el tamaño de

partícula de níquel al aumentar la proporción de magnesio.

Los resultados de reacción mostraron un aumento de la velocidad de reacción de metano

al añadir magnesio respecto al catalizador de referencia, observándose además una menor caída

de la actividad con el tiempo. Los mejores resultados se obtuvieron para los catalizadores con

relaciones molares Ni/Mg 2/1 y 1/1. Al contrario de lo que cabría esperar a priori se observó un

aumento de la relación molar H2/CO y una disminución de la velocidad de reacción de CO2 al

aumentar la cantidad de Mg en el catalizador, lo cual entraba en contradicción con el aumento

de basicidad que este tipo de promotores suele aportar [83]. Esta disminución de la velocidad

de reacción de CO parecía estar relacionada con la gran fortaleza de los centros básicos que

aparecen en los catalizadores Ni/-SiC promocionados con Mg [84]. Mediante la oxidación a

temperatura programada se evaluó la cantidad de coque generada durante los experimentos de

tri-reformado. La adición de Mg disminuye notablemente la cantidad de coque formada,


53

obteniéndose los menores valores para los catalizadores con relaciones molares Ni/Mg 2/1 y

1/1.

En el Capítulo 5 se muestran los resultados obtenidos al analizar la influencia del orden

de impregnación en la actividad catalítica de catalizadores de Ni soportados sobre -SiC y

promocionados con Mg. Se prepararon catalizadores con dos relaciones molares Ni/Mg

distintas, 10/1 y 1/1. Además, para cada una de esas razones molares se prepararon tres

catalizadores diferentes, variando el orden de impregnación de los metales. Así, en uno de ellos

se impregnó primero el Ni y luego el Mg, en el siguiente a la inversa, mientras que en el

tercero ambos metales fueron impregnados y calcinados al mismo tiempo.

Los datos de difracción de rayos X mostraron, al igual que se observó en el capítulo

anterior, un desplazamiento hacia valores menores del ángulo de difracción asociado al pico

principal del NiO, lo que está relacionado con la formación de una disolución sólida de NiO-

MgO. Los catalizadores preparados impregnando Ni en primer lugar mostraron un menor

desplazamiento de este ángulo, lo que indicaba que la formación de este compuesto no está tan

favorecida. En cuanto a los perfiles de reducción, los tres catalizadores preparados con una

relación Ni/Mg 10/1 mostraron un perfil muy similar, con dos picos principales de reducción.

Sin embargo, para los catalizadores preparados con relación Ni/Mg 1/1 se encontraron grandes

diferencias entre el catalizador preparado impregnando Ni en primer lugar con los otros dos. El

primero seguía manteniendo un perfil similar al de los catalizadores con menor cantidad de

promotor, con dos picos principales de reducción a temperaturas medias, mientras que los

preparados impregnando Mg en primer lugar o simultáneamente ambos metales, mostraron un

único pico de reducción a alta temperatura, lo que está relacionado con una mayor formación

de la fase NiO-MgO, como se ha comentado para los resultados de difracción de rayos X.

Estos catalizadores fueron probados en la reacción de tri-reformado de metano. Los

catalizadores preparados con menor carga de Mg e impregnando en primer lugar Mg o ambos

metales simultáneamente presentaron una mayor velocidad de conversión de metano que el

catalizador de referencia (Ni/-SiC), siendo además ésta mucho más estable durante la

duración del experimento. En cambio, la velocidad de reacción de dióxido de carbono fue

inferior, especialmente para el catalizador en el que se impregnó Mg en primer lugar, lo cual

estaba relacionado con la gran fortaleza de los centros básicos que se generan en estos

catalizadores [84], como se comentó en el capítulo anterior. El catalizador preparado


54

impregnando Ni en primer lugar y con la menor cantidad de Mg mostró un comportamiento

catalítico muy pobre, con valores de velocidad de reacción de CH4 y CO2 muy inferiores a los

del catalizador de referencia, lo que parecía ser debido a la mayor interacción de las partículas

de níquel con el soporte en vez de con el promotor [85], así como al posible bloqueo de los

centros activos al impregnar el magnesio sobre el níquel.

Para el caso de los catalizadores preparados con mayor carga de magnesio, aquellos

donde se impregnó en primer lugar Mg, o ambos metales simultáneamente, mostraron valores

similares de velocidad de reacción de metano que los mismos catalizadores preparados con

menor relación Ni/Mg, pero una mayor estabilidad. Esta mayor estabilidad parecía estar

relacionada con la mayor extensión de la formación de la disolución sólida NiO-MgO, que

favorecía una menor deposición de coque. En cuanto al catalizador preparado con una relación

molar Ni/Mg 1/1 y en el que fue impregnado en primer lugar el Ni, mostró los peores valores

de actividad catalítica.

La caracterización después de reacción mostró cómo la adición de Mg disminuía la

cantidad de coque generado durante la reacción de tri-reformado, siendo esta disminución

mayor en el caso de los catalizadores con mayor carga de Mg, probablemente debido al menor

tamaño de partícula y a la interacción existente entre el Ni y el Mg. Los catalizadores

preparados impregnando Mg en primer lugar fueron los que produjeron menor cantidad de

coque, seguidos de aquellos en que se impregnaron ambos metales simultáneamente. Por

último, fueron analizadas las especies presentes en los catalizadores después de reacción

mediante difracción de rayos X. Se apreció la presencia de picos correspondientes a Ni2Si en

los catalizadores en los que se impregnó Ni en primer lugar, como consecuencia de la reacción

entre el níquel y el carburo de silicio a altas temperaturas durante el proceso de calcinación.

Este compuesto parecía ser responsable de la menor actividad de estos catalizadores y su

mayor desactivación. Otra de las especies presentes en los catalizadores después de reacción es

el NiO, presente en todos menos en aquellos en los que se impregnaron simultáneamente el Ni

y el Mg. Este compuesto favorece la reacción de water-gas shift [86], lo que podía justificar la

menor velocidad de reacción de CO2 y la mayor relación molar H2/CO observada para los

catalizadores donde estaba presente el NiO.

En el Capítulo 6 se presentan los resultados relativos al estudio de la influencia de la

temperatura y la composición del alimento en el proceso de tri-reformado, utilizando un


55

catalizador Ni-Mg/-SiC en forma de pellets. Una vez obtenidos los datos experimentales de

caudal molar de cada uno de los compuestos presentes en la corriente de salida del reactor para

cada experimento, se realizó un modelado de dichos datos, considerando las ecuaciones

cinéticas correspondientes al reformado con vapor, al reformado seco y a la reacción de water-

gas shift.

Se llevaron a cabo 36 experimentos con diferente composición del alimento, obteniéndose

datos a 12 temperaturas diferentes para cada uno de ellos. Se observó un claro aumento de la

conversión de metano al aumentar la temperatura, lo que estaba relacionado con la elevada

endotermicidad de las reacciones principales del tri-reformado. Esta dependencia de la

conversión con la temperatura es aún más evidente en el caso de la conversión de dióxido de

carbono, observándose para determinados experimentos a temperaturas bajas un caudal de

salida de CO2 superior al caudal de entrada. Esto implica que no solo no se observa conversión

de CO2 en el proceso, sino que se está generando dicho compuesto debido a la concurrencia de

la reacción de water-gas shift, ya que a bajas temperaturas la reacción de reformado seco se ve

muy desfavorecida a favor de dicha reacción, produciéndose una generación neta de CO2. La

temperatura también condicionó la relación molar H2/CO del gas de síntesis producido,

disminuyendo notablemente el valor de este parámetro al aumentar la temperatura. Ello es

consecuencia de la preponderancia a bajas temperaturas de reacciones que favorecen la

producción de H2 (water-gas shift), mientras que a altas temperaturas predominan las

reacciones que favorecen la producción de CO (reformado seco).

Se observó un aumento de la conversión de metano al disminuir el caudal de este gas o

aumentar los caudales de agua u oxígeno, como era de esperar. Sin embargo, no se observó una

relación tan directa entre la conversión de CH4 y el caudal de CO2, ya que en aquellos

experimentos en los que la cantidad de agua es baja sí se observó un aumento de la conversión

de CH4 al aumentar el caudal de CO2, pero esto no ocurría cuando el caudal de agua era

elevado. Esta diferencia de comportamiento parecía estar relacionada con la competencia entre

el H2O y el CO2 por el CH4 disponible en el medio de reacción, viéndose favorecida la reacción

entre CH4 y H2O [26]. Así, un aumento en el caudal de CO2 no tenía un efecto notable sobre la

conversión de CH4, ya que éste está reaccionando predominantemente con el H2O.

En cuanto a la conversión de CO2, ésta se ve muy afectada por la contribución de la

reacción de water-gas shift al proceso global, obteniéndose los menores valores de conversión


56

al alimentar un bajo caudal de CO2 y un elevado caudal de H2O, condiciones que favorecen

esta reacción, y que por lo tanto, favorecen la generación de CO2.

El caudal de agua alimentado afectó a la relación molar H2/CO del gas de síntesis

producido, observándose un aumento de dicha relación al aumentar el caudal de agua (lo que se

debe tanto a una mayor contribución de la reacción de water-gas shift a baja temperatura),

como el hecho de que se favorezca la reacción de reformado con vapor en todo el tramo de

temperaturas. Un mayor caudal de oxígeno también favoreció el obtener relaciones molares

más altas, mientras que un mayor caudal de metano o dióxido de carbono favorecieron la

aparición de relaciones molares más bajas, debido a un aumento de la contribución de la

reacción de reformado seco al proceso global.

Como se ha comentado anteriormente, utilizando los datos experimentales obtenidos se

diseñó un modelo cinético que representase el proceso de tri-reformado. Para ello se consideró

que la reacción de oxidación parcial alcanzaba la conversión total de oxígeno, ya que no se

observó este compuesto en la corriente de salida del reactor en la práctica totalidad de los

experimentos. Para representar las reacciones de reformado con vapor y reformado seco se

utilizaron las ecuaciones de Wei e Iglesia [87], en las que se considera la activación del enlace

C-H como la etapa determinante en la cinética de estas reacciones. También se consideró

necesario tener en cuenta la contribución de la reacción de water-gas shift, para lo cual se

añadió al modelo la ecuación que en un trabajo anterior de nuestro grupo [88] se eligió como

óptima.

Los resultados del ajuste del modelo mostraron que la totalidad de los valores obtenidos

para los parámetros elegidos fueron estadísticamente significativos, así como el modelo en sí

mismo. Los valores de energías de activación obtenidos para el reformado con vapor, el

reformado seco y la reacción de water-gas shift se encontraban dentro del rango publicado en

la bibliografía, siendo los valores de las dos primeras reacciones cercanos al límite inferior. Se

obtuvo un buen ajuste de los datos experimentales con los valores calculados mediante el

modelo, especialmente para los caudales de metano y dióxido de carbono en la corriente de

salida.


57

D. CONCLUSIONES

De los resultados obtenidos en esta investigación se desprenden las siguientes

conclusiones finales:

La naturaleza del soporte tuvo una clara influencia en el comportamiento catalítico de

catalizadores aplicados al proceso de tri-reformado. La formación de aluminatos en el

catalizador Ni/Al2O3 fue perjudicial al disminuir la cantidad de especies reducidas.

Los catalizadores soportados sobre CeO2 y -SiC mostraron las mejores propiedades.

La composición de los gases utilizados durante la etapa de calcinación tuvo una gran

influencia en el comportamiento catalítico de catalizadores Ni/YSZ. Al calcinar en

atmósfera pobre en oxígeno aparecían una mayor cantidad de vacantes de oxígeno en

la superficie del soporte, lo que mejoraba la actividad y estabilidad del catalizador.

Los catalizadores soportados sobre CeO2 presentaron mayores tamaños de partícula de

Ni y una mayor capacidad de adsorción de CO2, comparados con los soportados sobre

-SiC. Una mayor interacción metal soporte en estos últimos catalizadores se

relacionó con una menor desactivación.

Al utilizar nitrato o acetato como precursores se obtuvieron menores tamaños de

partícula de Ni comparados con el cloruro o el citrato. Nitrato y acetato también

produjeron catalizadores en los que se observó una mayor interacción entre el Ni y el

soporte, especialmente en los catalizadores en los que se utilizó -SiC. Los

catalizadores preparados utilizando cloruro como precursor mostraron una

desactivación más intensa que el resto, probablemente debido a la presencia de iones

cloruro en la superficie del catalizador y al mayor tamaño de partícula de níquel

obtenido. Los mejores resultados en cuanto a actividad catalítica y estabilidad al

comparar diferentes precursores se obtuvieron para los catalizadores preparados a

partir de nitrato de níquel o acetato de níquel.

Al analizar la influencia de la composición del alimento sobre la conversión de

metano y la relación molar H2/CO del gas de síntesis obtenido mediante un diseño

factorial de experimentos, se observó cómo ninguno de los caudales estudiados tenía

un efecto estadísticamente significativo sobre la conversión de metano, mientras que

sí lo tuvieron todos ellos sobre la relación molar H2/CO. Un aumento del caudal

volumétrico de H2O u O2 produjo un aumento de esta relación molar, mientras que un

aumento del caudal volumétrico de CH4 o CO2 produjo una disminución de la misma.


58

Al adicionar Na o K como promotores a catalizadores Ni/-SiC se observó un

aumento de la velocidad de oxidación del -SiC durante la etapa de calcinación, lo

que produjo la transformación de este compuesto en -cristobalita, una de las fases

del óxido de silicio, lo que a su vez disminuía notablemente el área superficial del

soporte. También se observó un aumento de la velocidad de oxidación del -SiC para

el catalizador preparado impregnando una elevada cantidad de Ca, obteniéndose en

este caso cuarzo como resultado de la oxidación.

La adición de Mg como promotor aumentó la actividad y la estabilidad del

catalizador, disminuyendo el tamaño de partícula de Ni y aumentando la basicidad del

catalizador. Se observó que un aumento en la carga de Mg provocaba una disminución

en el tamaño de partícula y una mayor resistencia frente a la reducción, desplazando el

pico de reducción a temperaturas más altas, probablemente debido a la formación de

una disolución sólida NiO-MgO. Los catalizadores con una mayor concentración de

Mg mostraron una menor desactivación y produjeron un gas de síntesis con una mayor

relación molar H2/CO.

Al evaluar la influencia del orden de impregnación en catalizadores Ni/SiC

promocionados con Mg se observó cómo los catalizadores en los que se impregnó Ni

en primer lugar mostraron los peores resultados en cuanto a actividad catalítica,

probablemente debido a una pobre interacción entre Ni y Mg, al bloqueo parcial de las

partículas de Ni por el Mg y a la formación de Ni2Si durante la reacción, lo que

provocaba una disminución del número de centros activos. Los catalizadores

preparados impregnando Mg en primer lugar generaron menor cantidad de coque

durante la reacción, mientras que los catalizadores preparados impregnando

simultáneamente Ni y Mg mostraron los mejores resultados en cuanto a actividad

catalítica, probablemente debido a la fuerte interacción entre Ni y Mg, formándose la

disolución sólida NiO-MgO.

La temperatura tuvo una influencia notable, tanto en la conversión de metano como en

la conversión de dióxido de carbono del proceso de tri-reformado, especialmente en

éste último, ya que a bajas temperaturas se observó una producción neta de CO2,

mientras que a altas temperaturas se observaron conversiones elevadas. Estas

diferencias se debieron principalmente a la diferente contribución de la reacción de

water-gas shift sobre el proceso global, ya que al ser ésta una reacción exotérmica se

ve favorecida a bajas temperaturas.


59

El modelado de los datos catalíticos permitió la obtención de un modelo cinético para

representar el proceso de tri-reformado, considerando ecuaciones cinéticas para el

reformado con vapor, el reformado seco y la reacción de water-gas shift. Los valores

de energía de activación calculados se encontraban dentro de los intervalos habituales

indicados en la literatura, observándose una buena correlación entre los datos

modelados y los datos experimentales.

E. RECOMENDACIONES

Con objeto de ampliar y completar los resultados obtenidos en esta investigación se

recomienda:

Estudiar distintos metales como fase activa del proceso y/o diferentes métodos de

preparación del catalizador.

Analizar el efecto de los compuestos más comúnmente presentes en el gas natural que

puedan actuar como venenos del catalizador en el proceso de tri-reformado, como el

H2S.

Probar la actividad catalítica en el proceso de tri-reformado de los catalizadores

estudiados al utilizar corrientes reales, como gas natural o biogás.

Estudiar la influencia de la presión tanto en la velocidad de conversión de metano

como sobre la relación molar H2/CO del gas de síntesis producido.

Comprobar la estabilidad del catalizador en experimentos de muy larga duración,

analizando la cantidad de coque generada y otros procesos de desactivación que se

puedan dar.

Escalar el proceso a escala planta piloto para comprobar su viabilidad, utilizando una

configuración más cercana a la industrial.

Analizar el comportamiento del modelo diseñado en este trabajo mediante la

aplicación del mismo en la simulación de una planta industrial de tri-reformado de

metano.


60

F. BIBLIOGRAFÍA

[1] G.A. Olah, A. Goeppert, G.S. Prakash, Beyond oil and gas: the methanol economy, John

Wiley & Sons, 2009.

[2] C. Song, S.T. Srimat, W. Pan, L. Sun, ACS Division of Fuel Chemistry, Preprints. 46

(2001) 101-103.

[3] US-DOE, Draft Report, Feb 1999 (1999).

[4] Organization of the Petroleum Exporting Countries, 2013.

[5] British Petroleum, 2013.

[6] P.K. Swain, L.M. Das, S.N. Naik, Renewable and Sustainable Energy Reviews. 15 (2011)

4917-4933.

[7] D.A. Wood, C. Nwaoha, B.F. Towler, Journal of Natural Gas Science and Engineering. 9

(2012) 196-208.

[8] D. Wilhelm, D. Simbeck, A. Karp, R. Dickenson, Fuel Process. Technol. 71 (2001) 139-

148.

[9] L. Mayer, Métodos de la industria química en esquemas de flujo en colores: una visión

panorámica y. Orgánica, Reverté, 1987.

[10] Á.V. Ortuño, Introducción a la química industrial, Reverté, 1999.

[11] A. Parker, Contaminación del aire por la industria, Reverté, 1983.

[12] H.J.A. K. A. Weissermel, Quimica Orgánica Industrial, Editorial Reverte S.A., 1981.

[13]J.S. Kang, D.H. Kim, S.D. Lee, S.I. Hong, D.J. Moon, Applied Catalysis A: General. 332

(2007) 153-158.

[14] A. Erdohelyi, J. Cserenyi, F. Solymosi, J. Catal. 141 (1993) 287-299.

[15] T. Riis, E.F. Hagen, P.J. Vie, Ø. Ulleberg, IEA Hydrogen Implementing Agreement (IEA,

Paris, 2005) (2005).


61

[16] C. Song, W. Pan, 1 ed., Anaheim, CA., 2004.

[17] H. Jiang, H. Li, H. Xu, Y. Zhang, Fuel Process. Technol. 88 (2007) 988-995.

[18] C. Song, Chemical Innovation. 31 (2001) 21-26.

[19] M. Halmann, A. Steinfeld, Energy. 31 (2006) 3171-3185.

[20] M. Halmann, A. Steinfeld, Catal. Today. 115 (2006) 170-178.

[21] C. Song, W. Pan, Catal. Today. 98 (2004) 463-484.

[22] S.H. Lee, W. Cho, W.S. Ju, B.H. Cho, Y.C. Lee, Y.S. Baek, Catal. Today. 87 (2003) 133-

137.

[23] H.-t. Jiang, H.-q. Li, Y. Zhang, Journal of Fuel Chemistry and Technology. 35 (2007) 72-

78.

[24] H.-t. Jiang, H. Li, Y. Zhang, Chinese Journal of Catalysis. 28 (2007) 193.

[25] B. Huang, X. Li, S. Ji, B. Lang, F. Habimana, C. Li, Journal of Natural Gas Chemistry. 17

(2008) 225-231.

[26] D. Sun, X. Li, S. Ji, L. Cao, Journal of Natural Gas Chemistry. 19 (2010) 369-374.

[27] S. Sakai, S. Sawell, A. Chandler, T. Eighmy, D. Kosson, J. Vehlow, H. Van der Sloot, J.

Hartlen, O. Hjelmar, Waste Manage. (Oxford). 16 (1996) 341-350.

[28] B. Gullett, A. Sarofim, K. Smith, C. Procaccini, Process Safety and Environmental

Protection. 78 (2000) 47-52.

[29] H. Huang, A. Buekens, Chemosphere. 44 (2001) 1505-1510.

[30] S. Maken, J. Hyun, J.-W. Park, H.-C. Song, S. Lee, H. Chang, J Sci Ind Res (India). 64

(2005) 198-204.

[31] A. Björklund, M. Melaina, G. Keoleian, Int. J. Hydrogen Energy. 26 (2001) 1209-1221.

[32] T.-H. Kwak, S. Lee, S. Maken, H.-C. Shin, J.-W. Park, Y.D. Yoo, Energy & fuels. 19

(2005) 2268-2272.


62

[33] M. He, Z. Hu, B. Xiao, J. Li, X. Guo, S. Luo, F. Yang, Y. Feng, G. Yang, S. Liu, Int. J.

Hydrogen Energy. 34 (2009) 195-203.

[34] R. Gerhardt (Ed.), Properties and Applications of Silicon Carbide., InTech, Rijeka, 2011.

[35]T. Muranaka, Y. Kikuchi, T. Yoshizawa, N. Shirakawa, J. Akimitsu, Science and

Technology of Advanced Materials. 9 (2008) 044204.

[36] R. Moene, L.F. Kramer, J. Schoonman, M. Makkee, J.A. Moulijn, in: J.M.B.D.P.A.J. G.

Poncelet, P. Grange (Eds.), Stud. Surf. Sci. Catal., Elsevier, 1995, pp. 371-380.

[37] R. Moene, H.T. Boon, J. Schoonman, M. Makkee, J.A. Moulijn, Carbon. 34 (1996) 567-

579.

[38] R. Moene, L.F. Kramer, J. Schoonman, M. Makkee, J.A. Moulijn, Applied Catalysis A:

General. 162 (1997) 181-191.

[39] R. Moene, M. Makkee, J.A. Moulijn, Applied Catalysis A: General. 167 (1998) 321-330.

[40] R. Moene, E.P.A.M. Tijsen, M. Makkee, J.A. Moulijn, Applied Catalysis A: General. 184

(1999) 127-141.

[41] M.J. Ledoux, C. Crouzet, C. Pham-Huu, V. Turines, K. Kourtakis, P.L. Mills, J.J. Lerou,

J. Catal. 203 (2001) 495-508.

[42] M. Adams.

[43] L. Chair Man Hi-Tech CO.

[44] M.A. Vannice, Y.L. Chao, R.M. Friedman, Applied Catalysis. 20 (1986) 91-107.

[45] M.B. Kizling, P. Stenius, S. Andersson, A. Frestad, Applied Catalysis B, Environmental. 1

(1992) 149-168.

[46] M.J. Ledoux, C. Pham-Huu, CATTECH. 5 (2001) 226-246.

[47] N. Keller, C. Pham-Huu, C. Estournès, M.J. Ledoux, Applied Catalysis A: General. 234

(2002) 191-205.

[48] P. Nguyen, C. Pham, Applied Catalysis A: General. 391 (2011) 443-454.


63

[49] G. Winé, Z.E. Berrichi, C. Pham-Huu, J. Mol. Catal. A: Chem. 278 (2007) 64-71.

[50] F.Z.E. Berrichi, B. Louis, L. Cherif, M.J. Ledoux, C.P. Huu, 2008, pp. 1307-1310.

[51] Y. Nishida, T. Omichi, I. Katayama, H. Yamashita, M. Narisawa, Y. Matsumura, e-

Journal of Surface Science and Nanotechnology. 3 (2005).

[52] H.M. Torres Galvis, J.H. Bitter, C.B. Khare, M. Ruitenbeek, A.I. Dugulan, K.P. De Jong,

Science. 335 (2012) 835-838.

[53] A.R. De La Osa, A. De Lucas, A. Romero, J.L. Valverde, P. Sánchez, Catal. Today. 176

(2011) 298-302.

[54] S.M. Kim, S.I. Woo, ChemSusChem. 5 (2012) 1513-1522.

[55] J.M. García-Vargas, J.L. Valverde, A. De Lucas-Consuegra, B. Gómez-Monedero, P.

Sánchez, F. Dorado, Applied Catalysis A: General. 431-432 (2012) 49-56.

[56] Y. Jiao, C. Jiang, Z. Yang, J. Zhang, Microporous Mesoporous Mater. 162 (2012) 152-

158.

[57] S.G. Jeon, N.S. Kwak, N.S. Rho, C.H. Ko, J.G. Na, K.B. Yi, S.B. Park, Chem. Eng. Res.

Des. 90 (2012) 1292-1296.

[58] Y. Zhou, X. Li, X. Pan, X. Bao, J. Mater. Chem. 22 (2012) 14155-14159.

[59] K. Zhu, L. Guo, J. Lin, W. Hao, J. Shang, Y. Jia, L. Chen, S. Jin, W. Wang, X. Chen,

Appl. Phys. Lett. 100 (2012).

[60] R. Frind, L. Borchardt, E. Kockrick, L. Mammitzsch, U. Petasch, M. Herrmann, S.

Kaskel, Catalysis Science and Technology. 2 (2012) 139-146.

[61] A.V. Nikiforov, A.L. Tomás García, I.M. Petrushina, E. Christensen, N.J. Bjerrum, Int. J.

Hydrogen Energy. 36 (2011) 5797-5805.

[62] Y. Zheng, Z. Li, H. Yu, R. Wang, K. Wei, J. Mol. Catal. A: Chem. 301 (2009) 79-83.


64

[63] M. Lacroix, L. Dreibine, B. De Tymowski, F. Vigneron, D. Edouard, D. Bégin, P.

Nguyen, C. Pham, S. Savin-Poncet, F. Luck, M.J. Ledoux, C. Pham-Huu, Applied Catalysis A:

General. 397 (2011) 62-72.

[64] C. Bartholomew, R. Pannell, J. Catal. 65 (1980) 390-401.

[65] G. Poncelet, M. Centeno, R. Molina, Applied Catalysis A: General. 288 (2005) 232-242.

[66] B. Valle, B. Aramburu, A. Remiro, J. Bilbao, A.G. Gayubo, Applied Catalysis B:

Environmental. 147 (2014) 402-410.

[67] P. Kim, Y. Kim, H. Kim, I.K. Song, J. Yi, Applied Catalysis A: General. 272 (2004) 157-

166.

[68] H. Mori, C.-j. Wen, J. Otomo, K. Eguchi, H. Takahashi, Applied Catalysis A: General.

245 (2003) 79-85.

[69] J.D.A. Bellido, E.M. Assaf, Applied Catalysis A: General. 352 (2009) 179-187.

[70] H.M. Swaan, V.C.H. Kroll, G.A. Martin, C. Mirodatos, Catal. Today. 21 (1994) 571-578.

[71] J.D.A. Bellido, J.E. De Souza, J.-C. M’Peko, E.M. Assaf, Applied Catalysis A: General.

358 (2009) 215-223.

[72] O.W. Perez-Lopez, A. Senger, N.R. Marcilio, M.A. Lansarin, Applied Catalysis A:

General. 303 (2006) 234-244.

[73] M. Digne, P. Sautet, P. Raybaud, P. Euzen, H. Toulhoat, J. Catal. 226 (2004) 54-68.

[74] R. Mieville, J. Catal. 100 (1986) 482-488.

[75] C. Hoang-Van, Y. Kachaya, S. Teichner, Y. Arnaud, J. Dalmon, Applied catalysis. 46

(1989) 281-296.

[76] M. Kong, J. Fei, S. Wang, W. Lu, X. Zheng, Bioresour. Technol. 102 (2011) 2004-2008.

[77] X. Zhang, D. Hayward, D.M. Mingos, Catal. Lett. 86 (2003) 235-243.

[78] X. Zhang, C.S.-M. Lee, D.O. Hayward, D.M.P. Mingos, Catal. Today. 105 (2005) 283-

294.


65

[79] K.O. Christensen, D. Chen, R. Lødeng, A. Holmen, Applied Catalysis A: General. 314

(2006) 9-22.

[80] J.-H. Kim, D.J. Suh, T.-J. Park, K.-L. Kim, Applied Catalysis A: General. 197 (2000) 191-

200.

[81] V.R. Choudhary, B.S. Uphade, A.S. Mamman, J. Catal. 172 (1997) 281-293.

[82] A. Parmaliana, F. Arena, F. Frusteri, N. Giordano, J. Chem. Soc., Faraday Trans. 86

(1990) 2663-2669.

[83] O. Yamazaki, T. Nozaki, K. Omata, K. Fujimoto, Chem. Lett. 21 (1992) 1953-1954.

[84] L. Pino, A. Vita, F. Cipitì, M. Laganà, V. Recupero, Applied Catalysis B: Environmental.

104 (2011) 64-73.

[85] Z. Cheng, Q. Wu, J. Li, Q. Zhu, Catal. Today. 30 (1996) 147-155.

[86] S.H. Kim, S.-W. Nam, T.-H. Lim, H.-I. Lee, Applied Catalysis B: Environmental. 81

(2008) 97-104.

[87] J. Wei, E. Iglesia, J. Catal. 224 (2004) 370-383.

[88] A. De la Osa, A. De Lucas, A. Romero, J. Valverde, P. Sánchez, Int. J. Hydrogen Energy.

36 (2011) 9673-9684.

CHAPTER 1 Influence of the support on the catalytic

behavior of Ni catalysts for the dry reforming reaction and the tri-reforming process

Resumen

Abstract

1.1. INTRODUCTION

1.2. EXPERIMENTAL

1.2.1. Catalyst preparation

1.2.2. Catalyst characterization

1.2.3. Catalyst activity measurements

1.3. RESULTS AND DISCUSSION


1.3.2. Dry reforming catalytic activity

1.3.3. Tri-reforming catalytic activity

1.4. CONCLUSIONS

1.5. REFERENCES

Influence of the support on the catalytic behavior of Ni catalysts for the dry reforming reaction

and the tri-reforming process

69

Resumen

En este capítulo se analizó la influencia que diferentes materiales utilizados como

soporte (alúmina, ceria, -carburo de silicio y oxido de zirconio estabilizado con itrio)

tenían en el comportamiento catalítico de catalizadores de Ni, aplicados a la reacción

de reformado seco y al proceso de tri-reformado. Asimismo, se analizó la influencia

de la composición de la atmósfera en la que se llevó a cabo el proceso de calcinación

de catalizadores Ni/YSZ. Los experimentos de reducción a temperatura programada

mostraron notables diferencias en los perfiles de reducción y el grado de reducción

obtenido, en función tanto del soporte elegido como de las condiciones de calcinación.

El catalizador Ni/YSZ-O2, catalizador obtenido tras calcinarlo en una atmósfera pobre

en oxígeno, presentó una mayor reducibilidad debido a que presentaba un mayor

número de vacantes de oxígeno en la superficie del soporte. El catalizador Ni/Al2O3

dio los peores resultados en cuanto a velocidades de reacción de CH4 y CO2, debido a

su baja reducibilidad al formarse aluminatos de níquel. El catalizador Ni/CeO2 mostró

la relación molar H2/CO más baja en el proceso de tri-reformado. Este resultado se

pudo explicar teniendo en cuenta la mayor basicidad de este catalizador, que se

demostró mediante experimentos de desorción a temperatura programada de CO2. El

catalizador Ni/YSZ-O2 mostró la mayor velocidad de reacción de CH4 y CO2 en los

experimentos de reformado seco, obteniéndose valores ligeramente inferiores para los

catalizadores Ni/CeO2 y Ni/-SiC. Los catalizadores preparados utilizando CeO2 y -

SiC como soporte se posicionaron como los mejores para el proceso de tri-reformado.

Chapter 1

70

Abstract

The influence of different support materials (alumina, ceria, -silicon carbide and

yttria-stabilized zirconia) on the catalytic behaviour of Ni catalysts for the dry

reforming reaction and the tri-reforming process has been studied in the present

chapter. The influence on the catalytic performance of the composition of the

atmosphere surrounding the Ni/YSZ catalyst during the calcination step was also

analysed. Temperature-programmed reduction experiments showed remarkable

differences in the reduction profile and the degree of reduction of the catalysts as a

function of both the support used and the calcination conditions. Ni/YSZ-O2, the

catalyst calcined under an oxygen-poor atmosphere, presented a higher reducibility as

a consequence of the higher number of oxygen vacancies in the surface of the support.

The Ni/Al2O3catalyst gave the lowest CH4 and CO2 reaction rates as a consequence of

its low reducibility due to the formation of Ni aluminate. The Ni/CeO2 catalyst

showed the lowest H2/CO molar ratio for the tri-reforming process. This result can be

explained on considering the higher basicity of this catalyst, as shown by CO2-TPD

experiments. Ni/YSZ-O2 showed the higher reaction rate of CH4 and CO2 in the dry

reforming experiments, showing the Ni/CeO2 and Ni/-SiC slightly lower values. The

CeO2 and -SiC catalysts had the best characteristics as catalytic supports for the tri-

reforming process.



71

1.1. INTRODUCTION

The production of synthesis gas is of great interest as a first step in the manufacture of

liquid fuels through the Fischer–Tropsch synthesis as well as for other interesting chemical

compounds like methanol or dimethyl ether. One of the main characteristics that determines

the possible applications of synthesis gas is the H2/CO molar ratio. For most applications this

parameter is too high when the synthesis gas is obtained by steam reforming. Hence, it is of

interest to study alternative reforming processes that can yield the desired H2/CO molar ratio.

Dry reforming of methane (Equation 1.1) has been reported in numerous papers [1-3] as a way

to obtain synthesis gas using methane and carbon dioxide:

CO2 + CH4 → 2CO + 2H2 ΔHº = 247.3 KJ∙mol

-1 (Equation 1.1)

As a result of the stoichiometry of this reaction, the synthesis gas obtained is richer in CO

and this makes the gas more suitable for different applications. However, rapid deactivation is

commonly observed and the process has a high energy consumption.

Tri-reforming of methane is a synergistic combination of dry reforming (Equation 1.1),

steam reforming (Equation 1.2) and partial oxidation (Equation 1.3) of methane:

H2O + CH4 → CO + 3H2 ΔHº = 206.3 KJ∙mol

-1 (Equation 1.2)

1/2 O2 + CH4 → CO + 2H2 ΔHº = -35.6 KJ∙mol

-1 (Equation 1.3)

The simultaneous combination of these three reactions could avoid some of the problems

associated with the single reactions. For example, the quantity of coke deposited (Equations

1.4 and 1.5) is lower than that for the dry reforming process due to the reaction of coke with

water and oxygen (Equations 1.6 and 1.7). Furthermore, the process has lower energy

consumption than dry or steam reforming due to the occurrence of partial oxidation and, in

addition, it may be possible to modify the H2/CO molar ratio by changing the feed

composition:

2CO

C + CO2 ΔHº = -172.2 KJ∙mol

-1 (Equation 1.4)

CH4 C + 2H2 ΔH

º = 74.9 KJ∙mol

-1 (Equation 1.5)

H2O + C

CO + H2 ΔHº = 131.4 KJ∙mol

-1 (Equation 1.6)

Chapter 1

72

O2 + C

CO2 ΔHº = -393.7 KJ∙mol

-1 (Equation 1.7)

In the work described here, the dry reforming and tri-reforming processes were studied

using nickel as the active metal and four different supports, namely -alumina (-Al2O3), yttria-

stabilized zirconia (YSZ), silicon carbide (SiC) and ceria (CeO2). Nickel has been selected by

several authors [1, 4, 5] as the active phase for reforming reactions due to its high activity,

interesting redox properties and relatively low cost. Nickel has also been identified as the best

option for tri-reforming [6]. However, the catalytic activity of Ni-based catalysts is markedly

influenced by the nature of the support [7], which affects the reducibility and metal dispersion.

Alumina-based materials are frequently selected as supports for reforming catalysts due to their

mechanical and thermal resistance under the required reaction conditions [8, 9]. ZrO2 and

systems based on ZrO2-like YSZ have been investigated as supports for Ni catalysts in methane

reforming reactions and these materials show high thermal stability and a high ionic

conductivity due to the presence of defects in the crystal surface, where molecular oxygen can

be easily activated, thus increasing its reactivity [2]. CeO2 has been highlighted by several

authors as a promising promoter and/or support for Ni catalysts [10, 11]. One of the most

interesting properties of this material is its capacity to store and/or release reversibly high

quantities of O2 [12]. Silicon carbide has attracted interest as a result of the development of its

porous form (-SiC) [13]. SiC exhibits a high thermal conductivity and mechanical strength, a

low specific weight and chemical inertness. These properties are required for good catalyst

supports, especially for highly endothermic and/or exothermic reactions [14] where precise

control of the temperature within the catalyst bed is extremely important.

The aim of the work described here is to compare the catalytic behaviour of Ni-based

catalysts supported on different conventional and non-conventional materials for the dry

reforming and methane tri-reforming processes.

1.2. EXPERIMENTAL


Five supported nickel catalysts were prepared from four different supports. The supports

used were γ-alumina (MERCK), yttria-stabilized zirconia with a Y2O3/ZrO2 molar ratio of 0.08

(IONOTEC), -silicon carbide (SICAT CATALYST) and CeO2, which was obtained by

calcination of cerium ammonium nitrate (NH4)2Ce(NO3)6 (SIGMA ALDRICH) in air at 1173

K for 2 h. The catalysts were prepared by the wet impregnation method using nickel nitrate



73

Ni(NO3)2·6H2O (PANREAC) as the metal precursor. After impregnation, the catalysts were

dried in air overnight at 393 K and calcined in air at 1173 K for 2 h. An additional catalyst was

prepared with YSZ as the support, using different calcination conditions. In this case, the

calcination step was carried out in the reactor with a flow of 0.6% oxygen in nitrogen and a

total flow of 30 Nml min-1

. This catalyst is denoted as Ni/YSZ-O2.


Ni metal loading was determined by atomic absorption (AA) spectrophotometry using a

SPECTRA 220FS analyser. Samples (ca. 0.5 g) were treated with 2 ml of HCl, 3 ml of HF and

2 ml of H2O2 followed by microwave digestion (523 K). Surface area/porosity measurements

were conducted using a QUADRASORB 3SI sorptometer with N2 as the sorbate at 77 K. The

samples were outgassed at 523 K under vacuum (5 × 10–3

Torr) for 12 h prior to analysis.

Specific surface area was determined by the multipoint BET method. Specific total pore

volume was evaluated from N2 uptake at a relative pressure of P/Po = 0.99. Temperature-

programmed reduction (TPR) experiments were conducted in a commercial Micromeritics

AutoChem 2950 HP unit with TCD detection. Samples (ca. 0.15 g) were loaded into a U-

shaped tube and ramped from room temperature to 1173 K (10 K min−1

) with a reducing gas

mixture of 17.5% v/v H2/Ar (60 Nml min−1

). Raman spectroscopy was carried out on a Bruker

Senterra Raman Microscope at an excitation wavelength of 532 nm. Temperature-programmed

desorption (TPD) experiments were conducted in a commercial Micromeritics AutoChem 2950

HP unit with TCD detection. The sample (0.15 g) was loaded into a quartz tube, reduced and

pretreated in He. After cooling, 30 Nml min–1

of CO2 (99.99% purity, Praxair certified) was

passed through the sample for 30 min at a constant temperature of 323 K. Finally, the gaseous

and weakly adsorbed carbon dioxide was removed by a steady flow of He for a further 30 min.

The sample was then heated in 50 Nml min–1

of He at a heating rate of 10 K min–1

up to 1173

K. Static chemisorption experiments on the reduced samples were carried out at 308 K in the

pressure range of 100–450 Torr in a Micromeritics ASAP 2010 unit equipped with a

chemisorption controller using H2 as the titrant. The samples were then evacuated and cooled

down. Two parallel isotherms were obtained; the first one is a measure of both the physisorbed

and chemisorbed H2, whereas the second concerns the physisorbed H2 only. Assuming a 2:1

stoichiometry for H:Ni, the difference between the two isotherms was used to obtain the Ni

dispersion. XRD analyses were carried out on a Philips X’Pert instrument using nickel-filtered

Cu-Kα radiation. The samples were scanned at a rate of 0.02°step−1

over the range 5° ≤ 2θ ≤

Chapter 1

74

90° (scan time = 2 s step−1

). Transmission electron microscopy (TEM) analyses were carried

out on a JEOL JEM-4000EX unit with an accelerating voltage of 400 kV. Samples were

prepared by ultrasonic dispersion in acetone with a drop of the resulting suspension evaporated

onto a holey carbon-supported grid. Mean nickel particle size, evaluated as the surface-area

weighted diameter ( sd ), was calculated according to:

2

ii

i

3

ii

sdn

dn

d

(Equation 1.8)

where ni represents the number of particles with diameter di (∑ini ≥ 400).


The catalytic activity measurements were carried out in a tubular quartz reactor. The

reactor was 45 cm long and had a diameter of 1 cm. The catalyst was placed on a fritted quartz

plate located at the end of the reactor. The temperature of the catalyst was measured with a K-

type thermocouple (Thermocoax) placed inside the inner quartz tube. The entire reactor was

placed in a furnace (Lenton) equipped with a temperature control system. Reaction gases were

Praxair certified standards of CH4 (99.95% purity), CO2 (99.95% purity), O2 (99.99% purity)

and N2 (99.999% purity), with the latter used as the carrier gas. The gas flow was controlled by

a set of calibrated mass flowmeters (Brooks 5850 E and 5850 S). The water content in the

reaction mixture for the tri-reforming experiments was controlled using the vapour pressure of

H2O at the saturator temperature (297 K). All lines located downstream from the saturator were

heated to a temperature above 373 K to prevent condensation. The saturation of the feed stream

by water at the working temperature was verified by a blank experiment in which the amount

of water trapped by a condenser was measured for a certain time and then compared with the

theoretical value. Prior to the reaction, catalysts were reduced in a flow of pure H2 at a rate of

100 Nml min-1

at 673 K (Ni/CeO2, Ni/YSZ and Ni/YSZ-O2) or 973 K (Ni/Al2O3 and Ni/SiC).

The feed composition (by volume %) was as follows: 4% CH4, 4% CO2, N2 balance and the

total flow rate was of 100 Nml min-1

for the dry reforming experiments; 6% CH4, 3% CO2, 3%

H2O, 0.6% O2, N2 balance and the total flow rate was of 100 Nml min-1

for the tri-reforming

experiments. The weight hourly space velocity (WHSV) of the total gas mixture was fixed at

60000 Nml h-1

g-1

. Product gases were analysed with a micro gas chromatograph (Varian CP-

4900). The methane and carbon dioxide consumption rates were calculated as [inlet molar flow



75

– outlet molar flow]/weight of nickel. Changes in methane/carbon dioxide reaction rates were

calculated as [final reaction rate – initial reaction rate]/initial reaction rate. In a previous step, it

was verified that neither external nor internal diffusions were the controlling step under these

experimental conditions.



The XRD analyses obtained for the five catalysts tested in this work, both before and after

reduction, and those of the corresponding supports are shown in Figures 1.1 and 1.2. It can be

observed in Figure 1.1 a) that the catalyst Ni/Al2O3 had a very low crystallinity, as peaks

related to NiO or Ni0 due to the formation of NiAl2O4 [15] were not detected and there was

possible overlap of the peaks corresponding to NiO and NiAl2O4 [16]. The Ni/CeO2 catalyst

showed the typical diffraction peaks related to the cubic lattice of pure CeO2 (CaF2 structural

type) [17], while the diffraction peaks of the support in catalyst Ni/-SiC correspond

structurally to cubic -SiC (3C-type) [18]. In both samples NiO diffraction peaks for the fresh

catalysts and Ni0 peaks for the reduced ones were observed. Regarding the Ni/YSZ and

Ni/YSZ-O2 catalysts, clear differences between these two samples were not observed by XRD,

with both supports showing the tetragonal phase of YSZ [19]. The physical properties of the

catalysts are given in Table 1.1. There are clear differences in the surface characteristics of the

different supports. However, on comparing all of the catalysts only minor differences in the Ni

particle size, as determined by the Debye–Scherrer equation, were detected.

Chapter 1

76

01

02

03

04

05

06

07

08

09

0

###

c)b)

a)

^

Intensity (a.u.)

2

(º)

01

02

03

04

05

06

07

08

09

0

Red

uce

d

Fre

sh

Support

Suppo

rtS

upport

ºº

º

º

**

*

**

**

**

**

**

**

**

*

**

*

+

+

+

Intensity (a.u.)

2

(º)

Fre

sh

Red

uce

d

01

02

03

04

05

06

07

08

09

0

ºº

ºº

º

ºº

º

º

º

º*

*

*

*

*

*

++

++

^

Intensity (a.u.)

2

(º)

^^

^

Fre

sh

Red

uce

d

##

###

#

#

##

Fig

ure

1.1

. X

RD

pro

file

s w

her

e (+

) d

eno

tes

nic

kel

ox

ide

dif

frac

tio

n p

eak

s, (

^)

den

ote

s m

etal

lic

nic

kel

dif

frac

tio

n p

eak

s, (

#)

den

ote

s -

Al 2

O3 d

iffr

acti

on

pea

ks,

(*

) d

eno

tes

CeO

2 d

iffr

acti

on

pea

ks

and

(º)

den

ote

s

-SiC

dif

frac

tion

pea

ks.

a)

Ni/

Al 2

O3,

b)

Ni/

CeO

2,

c) N

i/S

iC.



77

Figure 1.2. XRD profiles where (+) denotes nickel oxide diffraction peaks, (^) denotes

metallic nickel diffraction peaks and (-) denotes YSZ diffraction peaks. a) Ni/YSZ, b) Ni/YSZ-

O2.

Table 1.1. Physical properties of the catalysts.

Ni/Al2O3 Ni/CeO2 Ni/SiC Ni/YSZ Ni/YSZ-O2

Ni loading (%) 5.2 4.1 3.9 4.3 3.7

Surface area (m2/g) 68.9 7.3 25.9 11.0 13.9

Total pore volume

(cm3/g) × 10

2

23.5

6.3 17.9 5.8 12.0

Particle diameter from

XRD (nm) n.d. 57 52 47 49


TEM (nm) n.d. 60 41 64 55

Ni dispersion (%) 1.00 2.83 1.56 1.50 1.57

Reduction degree (%) 67.0 78.3 99.8 78.3 89.8

TEM analysis (Figure 1.3.) was carried out with the aim of confirming the particle size of

the Ni metal (Table 1.1.) and to analyse the catalyst surface. Ni/Al2O3 catalysts consisted of a

layered phase that is usually assigned in this kind of catalyst to NiAl2O4 [20], a situation that

0 10 20 30 40 50 60 70 80 90

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

+++

^^

b)a)

Inte

nsi

ty (

a.u

.)

2(º)

Reduced

Support

Reduced

Support

Reduced

Fresh

Support

0 10 20 30 40 50 60 70 80 90

--

--

--

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

+ +

^

Inte

nsi

ty (

a.u

.)

2(º)

Reduced

Fresh

Support

Chapter 1

78

a)

b)

c)

d)

e)

50 n

m50 n

m50 n

m

100 n

m1

00

nm

made it difficult to discern unambiguously Ni particles. The Ni particle sizes determined by

TEM analysis for the other catalysts are very close to those calculated using the Debye–

Scherrer equation. The Ni metal dispersion was measured by H2 chemisorption (Table 1.1).

Large differences in the metal dispersion in the prepared catalysts were not observed, with

values between 1.00% (Ni/Al2O3) and 2.83% (Ni/CeO2). These values are very low and are

consistent with the large Ni particle sizes observed by XRD and TEM.

Fig

ure

1.3

. X

R T

EM

im

ages

. a)

Ni/

Al 2

O3,

b)

Ni/

CeO

2,

c) N

i/S

iC,

d)

Ni/

YS

Z,

e) N

i/Y

SZ

-O2.



79

TPR analyses were also carried out for all the catalysts (see Figure 1.4) and differences

can be seen as a function of the support used. The Ni/Al2O3 catalyst showed the reduction peak

at the highest temperature (1000 K), which could be related to the formation of NiAl2O4. This

spinel-type compound, which was formed at high calcination temperatures [9, 21], was very

difficult to reduce. The Ni/CeO2 catalyst showed two main reduction peaks. The peak at the

lowest temperature, which started at around 550 K, was assigned to the reduction of NiO to Ni0

[22]. The reduction peak that started at approximately 900 K is usually assigned to both the

removal of surface oxygen species and the reduction of bulk ceria [23]. Broader and smaller

peaks were obtained for the Ni/SiC catalyst. The reduction profile of this catalyst contained

three overlapped peaks with maxima at around 630 K, 760 and 860 K. In addition, a small peak

was observed at 1020 K. The peak at the lowest temperature is usually attributed to the

reduction of bulk NiO, while peaks at higher temperatures are attributed to a stronger

interaction between the metal and the support due to the formation of nickel silicate-like

species [24, 25]. The catalysts Ni/YSZ and Ni/YSZ-O2 gave rise to significantly different

reduction profiles. Two main reduction peaks, at around 600 K and 730 K, were observed for

the Ni/YSZ catalyst. These peaks are related to peaks I and II reported by Mori et al. [26] for a

Ni/YSZ catalyst prepared by impregnation. Peak I was attributed to the reduction of NiO

species with extremely weak interactions with YSZ, as the reduction temperature of this peak

is very close to that of pure NiO. Peak II was attributed by the authors to the reduction of NiO

with some interaction with the support. The reduction profile for the Ni/YSZ-O2 catalyst

showed one main reduction peak with a maximum at around 600 K. The presence of a large

tail in this peak is consistent with the existence of another peak that is overlapped by the major

one. On the other hand, the different calcination step applied to the Ni/YSZ-O2 catalyst led to a

system in which the NiO particles could be reduced more easily. Bellido et al. [2] studied the

influence of Y2O3 content on the reduction of Ni/YSZ catalysts. The results showed that the

shift in the reduction peaks towards lower temperatures is related to an increase in the surface

oxygen vacancies on the YSZ support, a situation that may promote the reduction of the

supported oxide at lower temperatures.

Chapter 1

80

Figure 1.4. Temperature-programmed reduction profiles.

Oxygen vacancies appear in the zirconia matrix when Z4+

cations are replaced by Y3+

in

the YSZ system in order to maintain electrical neutrality. These oxygen vacancies induce

structural disorder that can be quantified by measuring the full width at half maximum

(FWHM) of the Raman lines obtained near to 260 cm-1

[27, 28], with a larger width

corresponding to the samples with greater structural disorder and higher levels of oxygen

vacancies. The Raman spectra obtained for the Ni/YSZ and Ni/YSZ-O2 catalysts are shown in

Figure 1.5. The FWHM values for the peak observed at 260 cm-1

, which represents the

tetragonal phase of zirconia, were 45 cm-1

for Ni/YSZ and 49 cm-1

for Ni/YSZ-O2. These

values confirm the presence of higher levels of oxygen vacancies in the latter catalyst.

300 400 500 600 700 800 900 1000 1100 1200

TC

D s

ignal

(a.

u.)

Temperature (K)

Ni Al2O

3

Ni YSZ

Ni CeO2

Ni SiC

Ni YSZO2



81

Figure 1.5. Raman spectra of YSZ-supported catalysts.

The basicity of the catalysts was determined by CO2-TPD experiments (Figure 1.6) in

which the area of each desorption peak was measured (Table 1.2). The basic sites were

separated into weakly basic sites (desorption peak maximum in the range from 300 K to 650

K) and strongly basic sites (desorption peak maximum in the range from 650 K to 1100 K).

The Ni/Al2O3 catalyst contained the highest level of weakly basic sites, with two overlapped

desorption peaks with maxima at around 350 K and 530 K. Strongly basic sites were not

detected in this sample. The Ni/CeO2 catalyst contained the highest levels of strongly basic

sites and total basic sites. For this catalyst, a small peak in the range for weakly basic sites (370

K) was detected first and this was followed by two high desorption peaks at 800 K and 900 K.

The first peak was assigned to the desorption of CO2 from the support. The other two peaks

could arise due to the release of CO2 from metallic nickel [29]. The Ni/SiC catalyst only

showed a small desorption peak, indicating that this catalyst had a low basicity. The desorption

profiles for the catalysts Ni/YSZ and Ni/YSZ-O2 were very similar and contained two peaks:

one in the range for weakly basic sites and a second in the range for strongly basic sites.

However, the quantity of CO2 desorbed was higher for Ni/YSZ (see Table 1.2). This result is

consistent with the higher quantity of surface oxygen vacancies inferred for the Ni/YSZ-O2

catalyst. Surface oxygen vacancies have a positive charge [30, 31] and they can therefore act as

Lewis acid sites [32, 33].

100 200 300 400 500 600 700 800

Inte

nsi

ty (

a.u.)

Raman Shift (cm-1

)

Ni YSZ

Ni YSZO2

Chapter 1

82

Figure 1.6. CO2 Temperature-programmed desorption profiles.

Table 1.2. Basicity of the catalysts determined by CO2-TPD.

Basic sites (mol/g) Ni/Al2O3 Ni/CeO2 Ni/SiC Ni/YSZ Ni/YSZ-O2

Weak (300-650 K) 27.0 15.2 - 9.4 4.6

Strong (650-1100 K) - 66.1 3.0 8.7 6.6

Total 27.0 81.3 3.0 18.1 11.2

1.3.2. Dry reforming catalytic activity

The reaction rates for methane and carbon dioxide in dry reforming experiments are given

in Figures 1.7 and 1.8. The Ni/Al2O3 catalyst (Figure 1.7 a)) showed a quite stable and high

catalytic activity (4.93×10-4

mol s-1

gNi-1

and 5.17×10-4

mol s-1

gNi-1

, respectively), with

appreciable changes not observed in the CH4 and CO2 reaction rates. The H2/CO molar ratio

was 0.77. It is interesting to note that a value of 1 should be obtained for this ratio bearing in

mind the stoichiometry of this reaction, but the water gas shift reaction leads to an increase in

the CO content of the effluent gas. The Ni/CeO2 catalyst (Figure 1.7 b)) also showed a high

300 400 500 600 700 800 900 1000 1100 1200

TC

D s

ignal

(a.

u.)

Temperature (K)

Ni Al2O

3

Ni YSZ

Ni CeO2

Ni SiC

Ni YSZO2



83

catalytic activity, with average values for the CH4 and CO2 reaction rates (5.60×10-4

mol s-1

gNi-

1 and 6.30×10

-4 mol s

-1 gNi

-1, respectively) higher than those obtained for the alumina-based

catalyst. However, despite the fact that the CH4 reaction rate remained almost constant during

the experiment, a slight drop in the CO2 reaction rate was observed. In this case, the H2/CO

molar ratio was very stable, with values around 0.88. In the dry reforming process the H2/CO

molar ratio usually has values of less than 1 and this is due to the occurrence of the reverse

water gas shift reaction (Equation 1.9):

CO2 + H2 CO + H2O ΔH

º = -37.09 KJ∙mol

-1 (Equation 1.9)

Chapter 1

84

Figure 1.7. Dry reforming catalytic activity at 1073 K. Reaction conditions: CH4 = 4%, CO2 =

4%, N2 balance, total flow rate = 100 Nml/min. CH4 ( ) and CO2 ( ) consumption rates vs.

time on stream (left axis), and H2/CO molar ratio ( ) vs. time on stream (right axis). a)

Ni/Al2O3, b) Ni/CeO2, c) Ni/SiC.

The presence of CO2 in the feed, the generation of H2 during the reaction and the high

reaction temperatures all promote the occurrence of the aforementioned reaction, thus affecting

the final distribution of the products. The Ni/SiC catalyst was very active (Figure 1.7 c)), with

average CH4 and CO2 reaction rates of 5.86×10-4

mol s-1

gNi-1

and 6.43×10-4

mol s-1

gNi-1

,

0

1

2

3

4

5

6

7

8a)

Co

nsu

mp

tio

n r

ate

(mo

l s-1

g-1 N

i )·1

04

0

1

2

3

4

5

6

7

8

c)

b)

Co

nsu

mp

tio

n r

ate

(mo

l s-1

g-1

Ni )

·10

4

0 50 100 150 200 250

0

1

2

3

4

5

6

7

8

Co

nsu

mp

tio

n r

ate

(mo

l s-1

g-1 N

i )·1

04

Time (min)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

H2/C

O M

ola

r ra

tio

0.0

0.5

1.0

1.5

2.0

2.5

3.0

H2/C

O M

ola

r ra

tio

0.0

0.5

1.0

1.5

2.0

2.5

3.0

H2/C

O M

ola

r ra

tio



85

respectively. However, for this catalyst a slight decay in the catalytic activity with time on

stream was observed. On the other hand, the H2/CO molar ratio was very close to that observed

for the Ni/CeO2 catalyst (0.89) and was maintained during the transient experiment.

Figure 1.8. Dry reforming catalytic activity at 1073 K. Reaction conditions: CH4 = 4%, CO2 =

4%, N2 balance, total flow rate = 100 Nml/min. CH4 ( ) and CO2 ( ) consumption rates vs.

time on stream (left axis), and H2/CO molar ratio ( ) vs. time on stream (right axis). a)

Ni/YSZ, b) Ni/YSZ-O2.

The performance of the catalysts Ni/YSZ and Ni/YSZ-O2 is shown in Figure 1.8. Clear

signs of deactivation were not observed for Ni/YSZ, but the CH4 and CO2 reaction rates were

lower than those observed for the other catalysts (4.37×10-4

mol s-1

gNi-1

and 5.40×10-4

mol s-1

gNi-1

, respectively). The Ni/YSZ-O2 catalyst led to the highest CH4 and CO2 reaction rates

(6.02×10-4

mol s-1

gNi-1

and 6.74×10-4

mol s-1

gNi-1

, respectively), with a slight increase in the

catalytic activity with time on stream. Regardless of the catalyst, the H2/CO molar ratio

remained constant during the experiment, with values around 0.82 for Ni/YSZ and 0.87 for

Ni/YSZ-O2.

0

1

2

3

4

5

6

7

8

b)

a)

Consu

mpti

on r

ate

(mol

s-1 g

-1 Ni )

·10

4

0 50 100 150 200 250

0

1

2

3

4

5

6

7

8

Consu

mpti

on r

ate

(mol

s-1 g

-1

Ni )

·10

4

Time (min)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

H

2/C

O M

ola

r ra

tio

0.0

0.5

1.0

1.5

2.0

2.5

3.0

H

2/C

O M

ola

r ra

tio

Chapter 1

86

The lower reaction rates obtained with the Ni/Al2O3 catalyst could be associated with its

lower degree of reduction (Table 1.1). The degree of reduction was calculated by integrating

the reduction peaks obtained in the TPR experiments and relating this with the reduction

stoichiometry and the Ni load. This is an important parameter as the active Ni phase in the

reforming process is actually the metal and not the oxide [34]. XRD and TPR experiments

showed that the high calcination temperatures led to the formation of NiAl2O4, a compound

that is difficult to reduce. Consequently, a proportion of the supported Ni was not catalytically

active. The high activity of the catalysts Ni/CeO2 and Ni/-SiC could be related to their high

reduction degree, which would increase the availability of Ni0 species. The CO2 reaction rate

was comparable for both catalysts despite the fact that they have very different CO2 adsorption

capacities (as shown by the CO2-TPD experiments). The catalyst Ni/YSZ also showed low

catalytic activity when compared to that of Ni/YSZ-O2. The higher CH4 and CO2 reaction rates

obtained with the latter system can be attributed, according to the TPR experiments, to the

higher concentration of oxygen vacancies in the support, which would promote the catalytic

activity in the dry reforming reaction by activating the CO2 molecule [35].



87

N

i/A

l 2O

3

Ni/

CeO

2

Ni/

SiC

N

i/Y

SZ

N

i/Y

SZ

-O2

Dry

ref

orm

ing

exp

erim

ents

Aver

age

DR

CH

4 r

eact

ion r

ate

(mo

l s-1

gN

i-1)

×1

04

4.9

3

5.6

0

5.8

6

4.3

7

6.0

2

Chan

ge

in C

H4 r

eact

ion r

ate

(%)

end

/beg

innin

g

1.4

4

-9.4

8

-7.4

8

-5.8

9

4.0

7

Aver

age

DR

CO

2 r

eact

ion r

ate

(mo

l s-1

gN

i-1)

×1

04

5.1

7

6.3

0

6.4

3

5.4

0

6.7

4

Chan

ge

in C

O2 r

eact

ion r

ate

(%)

end

/beg

innin

g

0.7

2

-2.3

7

-4.4

2

-2.7

3

3.2

8

Aver

age

H2/C

O m

ola

r ra

tio

0

.77

0.8

8

0.8

9

0.8

2

0.8

7

Tri

-ref

orm

ing

exp

erim

ents

Aver

age

TR

CH

4 r

eact

ion r

ate

(mo

l s-1

gN

i-1)

×1

04

7.8

4

11

.13

11

.70

6

.14

9.3

0

Chan

ge

in C

H4 r

eact

ion r

ate

(%)

end

/beg

innin

g

-0.2

1

-0.0

3

-1.7

7

-8.9

3

12

.66

Aver

age

CH

4 c

on

ver

sio

n (

%)

99

.72

99

.22

95

.33

6

6.6

8

86

.43

Aver

age

TR

CO

2 r

eact

ion r

ate

(mo

l s-1

gN

i-1)

×1

04

1.6

1

2.5

7

2.2

5

2.0

7

3.4

0

Chan

ge

in C

O2 r

eact

ion r

ate

(%)

end

/beg

innin

g

-24

.09

-7

.29

-3

4.0

9

1.0

6

21

.79

Aver

age

CO

2 c

on

ver

sio

n (

%)

38

.19

6

3.4

9

47

.21

3

3.7

5

57

.91

Aver

age

H2/C

O m

ola

r ra

tio

1

.86

1.7

1

1.8

9

1.9

2

1.6

7

Ta

ble

1.3

. M

ain

cat

aly

tic

resu

lts.

Chapter 1

88

The principal catalytic results obtained in the dry reforming experiments are listed in

Table 1.3. A decay can be observed in the CH4 and CO2 reaction rates for the Ni/CeO2, Ni/SiC

and Ni/YSZ catalysts (change < 0), while the other two catalysts show a slight increase. This

loss of activity is usually related with coke deactivation and/or metal sintering in the dry

reforming process on using nickel catalysts [36]. The reason for the observed increase in the

catalytic activity for the other catalysts is not clear and requires further analysis in longer term

experiments.

1.3.3. Tri-reforming catalytic activity

The catalytic results corresponding to the tri-reforming experiments are depicted in

Figures 1.9 and 1.10. The Ni/Al2O3 catalyst showed a slightly higher CH4 reaction rate

(7.84×10-4

mol s-1

gNi-1

), which remained almost unchanged with time on stream, but a lower

CO2 reaction rate (1.61×10-4

mol s-1

gNi-1

) when compared to those obtained in the dry

reforming process. The H2/CO molar ratio obtained with this catalyst was around 1.86, which

is a much higher value than that obtained for the dry reforming reaction. These findings can be

explained by considering the occurrence of the steam reforming reaction and the partial

oxidation of methane to yield H2. Regarding the Ni/CeO2 catalyst, higher and constant values

with time on stream for the CH4 and CO2 reaction rates (11.13×10-4

mol s-1

gNi-1

and 2.57×10-4

mol s-1

gNi-1

, respectively) were obtained in comparison to those for the dry reforming process

and the H2/CO molar ratio was close to 1.7.



89

Figure 1.9. Tri-reforming catalytic activity at 1073 K. Reaction conditions: CH4 = 6%,

CO2 = 3%, H2O = 3% O2 = 0.6%, N2 balance, total flow rate = 100 Nml/min. CH4 ( ) and CO2

( ) consumption rates vs. time on stream (left axis), and H2/CO molar ratio ( ) vs. time on

stream (right axis). a) Ni/Al2O3, b) Ni/CeO2, c) Ni/SiC.

In the same way, the Ni/SiC catalyst proved to be more catalytically active, with average

CH4 and CO2 reaction rates of 11.70×10-4

mol s-1

gNi-1

and 2.25×10-4

mol s-1

gNi-1

, respectively.

With this catalyst, the CH4 reaction rate remained practically constant whereas the CO2

reaction rate underwent a slight decay with time on stream. On the other hand, the H2/CO

molar ratio initially had a value of 1.89 and this increased slightly with time on stream. As the

0

1

2

3

4

5

6

7

8

9

10

11

12

13 a)

Consu

mpti

on r

ate

(mol

s-1 g

-1 Ni )

·10

4

0

1

2

3

4

5

6

7

8

9

10

11

12

13

c)

b)

Consu

mpti

on r

ate

(mol

s-1 g

-1

Ni )

·10

4

0 50 100 150 200 250

0

1

2

3

4

5

6

7

8

9

10

11

12

13

Consu

mpti

on r

ate

(mol

s-1 g

-1 Ni )

·10

4

Time (min)

0.0

0.5

1.0

1.5

2.0

2.5

H2/C

O M

ola

r ra

tio

0.0

0.5

1.0

1.5

2.0

2.5

H2/C

O M

ola

r ra

tio

0.0

0.5

1.0

1.5

2.0

2.5

H2/C

O M

ola

r ra

tio

Chapter 1

90

H2/CO molar ratio increased, a decrease in the O2 reaction rate occurred and this indicates that

the dry reforming process was less important with time on stream with respect to the

contribution of the steam reforming and partial oxidation reactions, the importance of which

was increasingly significant.

Figure 1.10. Tri-reforming catalytic activity at 1073 K. Reaction conditions: CH4 = 6%,

CO2 = 3%, H2O = 3%, O2 = 0.6%, N2 balance, total flow rate = 100 Nml/min. CH4 ( ) and CO2

( ) consumption rates vs. time on stream (left axis), and H2/CO molar ratio ( ) vs. time. a)

Ni/YSZ, b) Ni/YSZ-O2.

The performance of the catalysts Ni/YSZ and Ni/YSZ-O2 is represented in Figure 1.10.

The latter catalyst showed better catalytic behaviour than Ni/YSZ. In fact, this catalyst showed

the lowest values for the CH4 and CO2 reaction rates (6.14×10-4

mol s-1

gNi-1

and 2.07×10-4

mol

s-1

gNi-1

, respectively) and the highest values for the H2/CO molar ratio (1.92). Regarding the

catalyst Ni/YSZ-O2, the values for the CH4 and CO2 reaction rates decreased slightly at the

beginning of the experiment before increasing with time on stream.

0

1

2

3

4

5

6

7

8

9

10

11

12

b)

a)

Consu

mpti

on r

ate

(mol

s-1 g

-1 Ni )

·10

4

0 50 100 150 200 250

0

1

2

3

4

5

6

7

8

9

10

11

12

Consu

mpti

on r

ate

(mol

s-1 g

-1

Ni )

·10

4

Time (min)

0.0

0.5

1.0

1.5

2.0

2.5

H

2/C

O M

ola

r ra

tio

0.0

0.5

1.0

1.5

2.0

2.5

H

2/C

O M

ola

r ra

tio



91

The catalysts prepared using CeO2 and SiC as supports gave rise to the highest CH4

reaction rates and they were also very stable. The presence of Ni species that were easier to

reduce, i.e., they did not interact strongly with the support, would explain this result. On the

other hand, the catalyst Ni/CeO2 led to a higher CO2 reaction rate and, hence, to a lower H2/CO

molar ratio. This finding indicates that the contribution of the dry reforming reaction was

higher when the Ni/CeO2 catalyst was used. This finding is associated, as verified by CO2-TPD

experiments, with the higher capacity of this catalyst to adsorb CO2 when compared to that of

Ni/SiC catalyst, thus improving the reactivity of this molecule. The slight increase in the value

of the H2/CO molar ratio with time on stream observed for the Ni/SiC catalyst could be due to

the occurrence of sintering processes. As pointed out by several authors, the presence of large

nickel particles leads to an increase in methane cracking [37], thus enhancing the H2

production and therefore the H2/CO molar ratio. If a sintering process does take place, there

would be an increase in the metal particle size and this would contribute to the methane

cracking process. Despite the high CO2 adsorption capacity of the Ni/Al2O3 catalyst, as

demonstrated by the CO2-TPD experiments, this system yielded a high H2/CO molar ratio and

a low CO2 reaction rate. In this case, the relative strengths of the basic sites should be

considered as the desorption temperature of CO2 observed for Ni/Al2O3 catalyst is much lower

than that of Ni/CeO2, which means that the basic sites in the former catalyst are weaker than

those in the latter. The weakly basic sites have a lower influence on the dry reforming reaction

because such sites are not significant in terms of changing the acid–base character [38]. In

addition, it is widely accepted that catalysts prepared using alumina as the support have a large

number of acidic sites [39-41]. In turn, supports with a highly acidic character usually promote

methane cracking (Equation 1.5) and the Boudouard reaction (Equation 1.4) [40], in which

higher quantities of H2 are generated and lower CO2 conversion is achieved.

As mentioned previously, the catalysts Ni/YSZ and Ni/YSZ-O2 showed different catalytic

performance in the tri-reforming process as far as CH4 reaction rate and H2/CO molar ratio are

concerned. These differences are probably related to the higher reducibility of the latter

catalyst, which would lead to an increase in the availability of Ni0 species and, hence, the

number of active sites (Table 1.1). On the other hand, the Ni/YSZ-O2 catalyst showed a higher

quantity of O2 vacancies when compared to that of the Ni/YSZ catalyst, which favoured the

reactivity of CO2 molecules and, therefore, the contribution of the dry reforming reaction. This

situation would explain the differences in the values of the H2/CO molar ratio observed for

these two catalysts.

Chapter 1

92

Catalytic data for the tri-reforming reaction are also shown in Table 1.3. Ni/Al2O3,

Ni/CeO2, Ni/SiC and Ni/YSZ show a decrease in the CH4 reaction rate, although this change is

very slight for the three first catalysts. As a general trend it can be observed that the

deactivation suffered by the catalysts is lower in the tri-reforming experiments compared to the

changes for the dry reforming experiments. This difference is probably related to the presence

of higher levels of oxidants in the reaction environment and, therefore, a smaller degree of

coke deposition. The results obtained for Ni/YSZ-O2 show an increase in the CH4 reaction rate

at the end of the reaction when compared to that measured at the beginning of the reaction.

Similar behaviour was also observed for the dry reforming experiment. There are very few

reports in the literature in which the influence of different supports in the tri-reforming of

methane is analysed. In one of the first papers on the tri-reforming process, Song et al. [42]

reported the behaviour of Ni/ZrO2, Ni/CeO2, Ni/MgO, Ni/MgO/CeZrO, Ni/CeZrO and

Ni/Al2O3 catalysts. They performed experiments with a CH4:CO2:H2O:O2 molar ratio of

1:0.48:0.54:0.1 in the temperature range 973-1123 K. At 1073 K a slightly lower CH4

conversion was observed for Ni/CeO2 and Ni/CeZrO when compared to the other catalysts.

The CH4 conversion values observed for Ni/Al2O3 (94%) and Ni/CeO2 (91%) were lower than

those obtained in our work (99.7% and 99.2% respectively, Table 1.3), but the main

differences were observed in the CO2 conversion. In this respect, Song et al. reported very

similar CO2 conversion values for both catalysts (61% for Ni/Al2O3 and 60% for Ni/CeO2)

whereas we obtained a much lower value for Ni/Al2O3 (38.2%) than for Ni/CeO2 (63.5%). The

differences in the behaviour could be due to the different preparation methods used for the

Ni/Al2O3 catalysts, as Song et al. used a commercial catalyst (ICI Synetix 23-4, R15513) that

showed a single reduction peak at 763 K, which indicates that NiAl2O4 was not formed during

the preparation method. In contrast, our catalyst system was obtained after a high temperature

calcinations process.

1.4. CONCLUSIONS

The results reported here show that the nature of the support clearly has an influence on

the catalytic behaviour of Ni catalysts for the tri-reforming reaction of methane. Calcination

conditions for the Ni/YSZ catalyst modified the reducibility of NiO and the number of surface

oxygen vacancies, which in turn had a marked influence on the catalytic behaviour for both the

dry reforming and the tri-reforming processes. The catalyst Ni/Al2O3 gave rise to the lowest

CH4 and CO2 reaction rates due to the formation of a NiAl2O4 metal phase. The catalyst



93

Ni/CeO2 presented the highest basicity value and this gave rise to the lowest H2/CO molar

ratio, which was associated with the occurrence of the dry reforming reaction to a greater

extent. Catalysts Ni/CeO2 and Ni/-SiC showed the best catalytic performance for the tri-

reforming process as they yielded a high methane reaction rate without significant deactivation.

Chapter 1

94

1.5 REFERENCES

[1] A. Kambolis, H. Matralis, A. Trovarelli, C. Papadopoulou, Applied Catalysis A: General.

377 (2010) 16-26.


[3] M.-S. Fan, A.Z. Abdullah, S. Bhatia, Applied Catalysis B: Environmental. 100 (2010) 365-

377.

[4] D.H. Prasad, H.I. Ji, H.R. Kim, J.W. Son, B.K. Kim, H.W. Lee, J.H. Lee, Applied Catalysis

B: Environmental. 101 (2011) 531-539.

[5] H. Özdemir, M.A. Faruk Öksüzömer, M. Ali Gürkaynak, Int. J. Hydrogen Energy. 35

(2010) 12147-12160.

[6] J.S. Kang, D.H. Kim, S.D. Lee, S.I. Hong, D.J. Moon, Applied Catalysis A: General. 332

(2007) 153-158.


78.

[8] M.C. Sánchez-Sánchez, R.M. Navarro, J.L.G. Fierro, Int. J. Hydrogen Energy. 32 (2007)

1462-1471.

[9] B. Valle, B. Aramburu, A. Remiro, J. Bilbao, A.G. Gayubo, Applied Catalysis B:

Environmental. 147 (2014) 402-410.

[10] S. Wang, G.Q. Lu, Applied Catalysis B: Environmental. 19 (1998) 267-277.

[11] L. Pino, A. Vita, M. Laganà, V. Recupero, Applied Catalysis B: Environmental. 148–149

(2014) 91-105.

[12] H.C. Yao, Y.F.Y. Yao, J. Catal. 86 (1984) 254-265.

[13] M.J. Ledoux, S. Hantzer, C.P. Huu, J. Guille, M.-P. Desaneaux, J. Catal. 114 (1988) 176-

185.



95

[14] H. Liu, S. Li, S. Zhang, J. Wang, G. Zhou, L. Chen, X. Wang, Catal. Commun. 9 (2008)

51-54.

[15] J.M. Rynkowski, T. Paryjczak, M. Lenik, Applied Catalysis A: General. 106 (1993) 73-

82.

[16] R. Buitrago-Sierra, J. Ruiz-Martínez, J.C. Serrano-Ruiz, F. Rodríguez-Reinoso, A.

Sepúlveda-Escribano, J. Colloid Interface Sci. 383 (2012) 148-154.

[17] S. Saitzek, J.F. Blach, S. Villain, J.R. Gavarri, physica status solidi (a). 205 (2008) 1534-

1539.

[18] G. Treffer, J. Neuhäuser, G. Marx, Microchimica Acta. 125 (1997) 325-330.

[19] J.C. Ray, R.K. Pati, P. Pramanik, J. Eur. Ceram. Soc. 20 (2000) 1289-1295.

[20] S. Wang, Industrial & Engineering Chemistry Research. 38 (1999) 2615-2625.

[21] P. Kim, Y. Kim, H. Kim, I.K. Song, J. Yi, Applied Catalysis A: General. 272 (2004) 157-

166.


[23] J. Xu, B. Xue, Y.-M. Liu, Y.-X. Li, Y. Cao, K.-N. Fan, Applied Catalysis A: General. 405

(2011) 142-148.

[24] Y.P. Tulenin, M.Y. Sinev, V.V. Savkin, V.N. Korchak, Catal. Today. 91–92 (2004) 155-

159.

[25] B. Mile, D. Stirling, M.A. Zammitt, A. Lovell, M. Webb, J. Catal. 114 (1988) 217-229.

[26] H. Mori, C.-j. Wen, J. Otomo, K. Eguchi, H. Takahashi, Applied Catalysis A: General.

245 (2003) 79-85.

[27] J. Moon, H. Choi, H. Kim, C. Lee, Surf. Coat. Technol. 155 (2002) 1-10.

[28] D.J. Kim, H.J. Jung, I.S. Yang, J. Am. Ceram. Soc. 76 (1993) 2106-2108.

[29] S.C. Dantas, J.C. Escritori, R.R. Soares, C.E. Hori, Chem. Eng. J. 156 (2010) 380-387.

Chapter 1

96

[30] W.-P. Dow, Y.-P. Wang, T.-J. Huang, J. Catal. 160 (1996) 155-170.

[31] E.C. Subbarao, H.S. Maiti, Solid State Ionics. 11 (1984) 317-338.

[32] N.B. Jackson, J.G. Ekerdt, J. Catal. 126 (1990) 31-45.

[33] A. Methivier, M. Pijolat, J. Catal. 139 (1993) 329-337.

[34] H.M. Swaan, V.C.H. Kroll, G.A. Martin, C. Mirodatos, Catal. Today. 21 (1994) 571-578.

[35] J.D.A. Bellido, J.E. De Souza, J.-C. M’Peko, E.M. Assaf, Applied Catalysis A: General.

358 (2009) 215-223.

[36] J. Guo, H. Lou, H. Zhao, D. Chai, X. Zheng, Applied Catalysis A: General. 273 (2004)

75-82.


200.

[38] O.W. Perez-Lopez, A. Senger, N.R. Marcilio, M.A. Lansarin, Applied Catalysis A:

General. 303 (2006) 234-244.

[39] M. Digne, P. Sautet, P. Raybaud, P. Euzen, H. Toulhoat, J. Catal. 226 (2004) 54-68.

[40] S. Freni, G. Calogero, S. Cavallaro, J. Power Sources. 87 (2000) 28-38.

[41] R. Mieville, J. Catal. 100 (1986) 482-488.


CHAPTER 2 Precursor influence and catalytic behavior of

Ni/CeO2 and Ni/SiC catalyst for the tri-reforming process.

Resumen

Abstract

2.1. INTRODUCTION

2.2. EXPERIMENTAL






2.3.2. Catalytic activity

2.4. CONCLUSIONS

2.5. REFERENCES

Precursor influence and catalytic behavior of Ni/CeO2 and Ni/SiC catalyst for the tri-

reforming process

99

Resumen

El trabajo contenido en este capítulo se centró en la evaluación del

comportamiento catalítico, para el proceso de tri-reformado, de catalizadores Ni/CeO2

y Ni/-SiC preparados con cuatro sales de níquel diferentes (nitrato, acetato, cloruro y

citrato). Las partículas de metal soportadas sobre CeO2 tuvieron un mayor tamaño (lo

que implica una menor interacción metal-soporte) que las soportadas sobre -SiC. Se

comprobó que el tipo de sal precursora elegida para la preparación del catalizador

tenía una gran influencia en el tamaño de las partículas de níquel. Al utilizar cloruros o

citratos se obtuvieron catalizadores con partículas de níquel mayores y un peor

comportamiento catalítico. El soporte y la sal precursora de níquel también tuvieron

una notable influencia sobre la velocidad de reacción de metano y la relación molar

H2/CO obtenida en los productos de reacción. Los experimentos de TPD de CO2

mostraron que los catalizadores en los que se seleccionó CeO2 como soporte exhibían

una mayor cantidad de centros básicos, lo que estaba relacionado con una menor

relación molar H2/CO de los productos de reacción. Se obtuvo una alta velocidad de

reacción de metano y una buena estabilidad catalítica en los catalizadores en los que se

utilizó nitrato de níquel o acetato de níquel como sal precursora y -SiC como soporte.

Por lo tanto, estos últimos catalizadores representan la opción más apropiada para

llevar a cabo el proceso de tri-reformado.

Chapter 2

100

Abstract

The aim of the work described here is to evaluate the catalytic performance in the

tri-reforming process of Ni/CeO2 and Ni/SiC catalysts prepared by using four

different nickel salts (nitrate, acetate, chloride and citrate).Metal particles supported

over ceria had bigger particle sizes (leading to lower metal-support interactions) than

those supported on SiC. It was also demonstrated that the metal salt used in the

preparation of Ni-based catalysts had a marked influence on the size of the nickel

particles. Larger particles with a worse catalytic behaviour were obtained when nickel

chloride and nickel citrate were used as the precursors of Ni supported species.

Methane consumption rate and H2/CO ratio in the effluents were influenced by the

type of support and salt precursor used in the preparation of the catalysts. CO2-TPD

proved that catalysts based on ceria as the support presented more basic sites, which

was related to a decrease of the H2/CO molar ratio in the effluents coming from the

reactor. High methane consumption rate and good catalytic stability were obtained

when nickel nitrate and nickel acetate were used to prepare Ni/SiC catalysts. The

results showed that these latter catalysts can be considered as promising ones for the

tri-reforming process.


reforming process

101

2.1. INTRODUCTION

Interest in the use of CO2 as an important source of carbon for the synthesis of fuels and

chemical products has significantly increased in recent years as a consequence of the public

concern about its negative effects on the atmosphere. Among the different processes proposed

that allow this specie to be transformed in valuable compounds, dry reforming of methane in

the presence of CO2 yielding synthesis gas (Equation 2.1) should be highlighted according to

the number of works reported in the literature [1-3]

CO2 + CH4 → 2CO + 2H2 (H◦ = 247.3 kJmol

-1) (Equation 2.1)

Despite of that, the high cost and limited availability of noble metal catalysts, such as Pt,

Rh, and Ru used in this process [4-6] limits its commercialization. For many years, nickel-

based catalysts have been proven to be the most suitable ones for hydrocarbon reforming.

Following this idea, nickel supported on oxides, such as Al2O3, MgO, TiO2, ZrO2, SiO2, CeO2

and La2O3 have been extensively investigated [7-10] as catalysts for the reforming of CH4 by

CO2. However, drawbacks as the deactivation of the catalyst by carbon formation and the high

energy consumption required as a result of the endothermic nature of the process should be

considered in future developments.

The tri-reforming process proposed by Song in 2001 [11] could avoid these problems.

This process consists of a synergetic combination of dry reforming (Equation 2.1), steam

reforming (Equation 2.2) and partial oxidation (Equation 2.3) of methane. Tri-reforming

process presents three major advantages: H2/CO molar ratio in the product can be controlled by

altering the relative amounts of gas reagents, the process is less endothermic due to the

occurrence of the partial oxidation reaction and, finally, coke formation can be reduced by the

presence of oxidants (H2O and O2; Equations (2.4-2.7)).

H2O + CH4 → CO + 3H2 (H◦ = 206.3 kJmol

-1) (Equation 2.2)

CH4 + 1/2O2 → CO + 2H2 (H◦ = −35.6 kJmol

-1) (Equation 2.3)

2CO

C + CO2 (H◦ = −172.2 kJmol

-1) (Equation 2.4)

CH4 C + 2H2 (H

◦ = 74.9 kJmol

-1)

(Equation 2.5)

C + H2O

CO + H2 (H◦ = 131.4 kJmol

-1) (Equation 2.6)

Chapter 2

102

C + O2 CO2 (H

◦ = −393.7 kJmol

-1) (Equation 2.7)

Synthesis gas produced by the tri-reforming of CH4 can be used for the production of

DME, Fischer–Tropsch synthesis fuels and high-valued chemicals [12, 13] as well as applied

to the fuel processor of Solid Oxide Fuel Cells (SOFC) and Molten Carbonate Fuel Cells

(MCFC) systems.

Nickel has been extensively studied in different reforming processes, including steam

reforming [14], dry reforming [1] and partial oxidation [15]. Ni compared with noble metals

[16] presents some advantages related to its availability and lower cost. Nonetheless, more

research is still required in order to improve its coke resistance ability for reforming of

available hydrocarbons. Trimm and co-workers [17] reported the effect of MgO and CaO for

the prevention of carbon deposition in the dry reforming of methane. Jiang and Badwal [18]

also reported the effective use of NiO–YSZ cermet as an anode catalyst in the internal

reforming in SOFC systems, showing that the sinterability of nickel was controlled by the

presence of YSZ particles. Nickel has also been identified as the best option for tri-reforming

[12], but the influence of nickel precursors has not been studied in depth.

Ni particle size and metallic dispersion are key factors affecting the carbon deposition

mechanism. It is well known that highly dispersed tiny particles avoid carbon deposition,

which could cause irreversible catalyst deactivation and/or the increase of catalytic bed

pressure [19]. Metallic dispersion could be changed by using different precursors. Aupretre et

al. [20] prepared Rh/spinel catalysts supported on alumina by impregnation with rhodium

nitrate, rhodium acetate or rhodium chloride. Metallic dispersion was markedly increased by

using nitrate and chloride. Moreover, these authors claimed that the impregnation with acetate

was very difficult and a significant part of the rhodium was not retained on the support. In spite

of this, they reported that nitrate precursors should be avoided since less stable materials were

produced due to its high acidity. Song et al. [21] studied the effect of cobalt precursor on the

performance of Co/CeO2 catalysts. The best results were obtained by using organometallic Co

precursors (especially cobalt acetyl acetonate), as they increased the Co dispersion and the

stability, thus leading to a high H2 yield. Depending on the cobalt-precursor, Llorca et al. [22]

found differences in the ethanol steam reforming performance of Co/ZnO catalysts. The

catalyst prepared from Co2(CO)8 showed the best performance due to a high degree of cobalt

reduction, with prevalence of small particles and presence of a CoO phase.


reforming process

103

Silicon carbide is a ceramic compound that has caught a lot of interest due to the recent

development of its porous form (-SiC) [23]. -SiC exhibits high thermal conductivity and

mechanical strength, low specific weight and chemical inertness. These properties are required

to be a good heterogeneous catalyst support, especially for high endothermic and/or exothermic

reactions [24], where precise control of the temperature inside the catalyst bed is extremely

important. In addition, due to its chemical inertness, the recovery of the active phase is

extremely easy, i.e. acidic or basic washing, which reduces the investment cost of the process

for the final spent catalyst disposal and the fully re-use of the support [25]. This material has

been proved in different reactions [25, 26] but not in the tri-reforming process.

Ceria support has been used in several formulations of Ni steam reforming catalysts

because its presence increases the catalytic stability by promoting carbon removal from the

metallic surface [27]. Ceria is known for its high oxygen storage/transport capacity (OSC), i.e.,

its ability to release oxygen under oxygen poor environment and quick re-oxidation under

oxygen rich environment. Unfortunately, OSC declines under high temperatures and reductive

conditions. Several studies have shown the effect of ceria support, since its presence promotes

oxidation reactions, such as CO oxidation, water-gas shift, and steam reforming of methane

[28-30].

In this work, the tri-reforming process using Ni-based catalysts, where the metal was

introduced from four different salts (nickel nitrate, acetate, chloride and citrate) and supported

over two support SiC and CeO2, was studied.

2.2. EXPERIMENTAL


Eight nickel-supported catalysts were prepared from different nickel salts and supports.

The supports used were SiC, provided by SICAT CATALYST, and CeO2, obtained by

calcination in air of ammonium cerium nitrate (NH4)2Ce(NO3)6 (SIGMA ALDRICH) at 1173

K for 2 h. The catalysts were prepared by the impregnation method using nickel nitrate

Ni(NO3)2·6H2O (PANREAC), nickel acetate C4H6NiO4·4H2O (Sigma Aldrich), nickel chloride

NiCl2·6H2O (Sigma Aldrich) and nickel citrate Ni3(C6H5O7)2·H2O (Alfa Aesar). The catalysts

were named to as Ni-XY, where X indicates the precursor nitrate (N), acetate (A), chloride (C)

or citrate (Ci) and Y the support ceria (C) or silicon carbide (S) employed in each case. For

Chapter 2

104

instance, Ni-NC was a nickel catalyst obtained from nitrate as the precursor and ceria as the

support. After impregnation, the catalysts were dried in air overnight at 393 K and calcined in

air at 1173 K for 2 h. The Ni loading was fixed at 5 wt%.


Ni metal loading was determined by atomic absorption (AA) spectrophotometry, using a

SPECTRA 220FS analyzer. Samples (ca. 0.5 g) were treated in 2 mL HCl, 3 mL HF and 2 mL

H2O2 followed by microwave digestion (523 K). Surface area/porosity measurements were

conducted using a QUADRASORB 3SI sorptometer apparatus with N2 as the sorbate at 77 K.

The samples were outgased at 523 K under vacuum (5×10–3

torr) for 12 h prior to the analysis.

Specific surface areas were determined by the multi point BET method. Specific total pore

volume was evaluated from N2 uptake at a relative pressure of P/Po = 0.99. Temperature-

programmed reduction (TPR) experiments were conducted in a commercial Micromeritics

AutoChem 2950 HP unit with TCD detection. Samples (ca. 0.15 g) were loaded into a U-

shaped tube and ramped from room temperature to 1173 K (10 K min−1

), using a reducing gas

mixture of 17.5% v/v H2/Ar (60 cm3 min

−1). Temperature-programmed desorption (TPD)

experiments were conducted in a commercial Micromeritics AutoChem 2950 HP unit with

TCD detection. 0.15 g of sample were loaded in a quartz tube, reduced and pretreated in He.

After cooling, 30 mL min-1

of CO2 (99.99% purity, Praxair certified) was passed through the

sample for 30 min at a constant temperature of 323 K. Finally, the gaseous and weakly

adsorbed carbon dioxide was removed by a steady flow of He for another 30 min. The sample

was then heated in 50 mL min-1

of He with a heating rate of 10 K min-1

up to 1173 K. XRD

analyses were conducted with a Philips X’Pert instrument using nickel-filtered Cu Kα

radiation; the samples were scanned at a rate of 0.02°step−1

over the range 5° ≤ 2θ ≤ 90° (scan

time = 2 s step−1

). Transmission electron microscopy (TEM) analyses employed a JEOL JEM-

4000EX unit with an accelerating voltage of 400 kV. Samples were prepared by ultrasonic

dispersion in acetone with a drop of the resultant suspension evaporated onto a holey carbon-

supported grid. Mean nickel particle size evaluated as the surface-area weighted diameter ( sd )

was computed according to (Equation 2.8):

2

ii

i

3

ii

sdn

dn

d

(Equation 2.8)


reforming process

105



Measurement of the catalytic activity was carried out within a tubular quartz reactor. The

dimensions of the reactor were 45 cm length and 1 cm diameter, with the catalyst placed on a

fritted quartz plate located at the end of the reactor. The temperature of the catalyst was

measured with a K-type thermocouple (Thermocoax) placed inside the inner quartz tube. The

entire reactor was placed in a furnace (Lenton) equipped with a temperature-programmed

system. Reaction gases were Praxair certified standards of 10% CH4/N2, 10% CO2/N2, O2

(99.99% purity), and N2 (99.999% purity), the latter used as the carrier gas. The gas flow was

controlled by a set of calibrated mass flowmeters (Brooks 5850 E and 5850 S). The water

content in the reaction mixture was controlled using the vapour pressure of H2O at the

temperature of the saturator (297 K). All lines placed downstream from the saturator were

heated above 373 K to prevent condensation. The saturation of the feed stream by water at the

working temperature was verified by a blank experiment in which the amount of water trapped

by a condenser was measured for a specific time and compared with the theoretical value. The

feed composition (by volume %) was as follows: 6% CH4, 3% CO2, 3% H2O, 0.6% O2, N2

balance, and the total flow was of 100 mL min-1

. This composition was determined based on

previous studies [11, 31] and the conditions imposed by our setup, to get a molar ratio in the

feed of CH4/CO2/H2O/O2 = 1/0.5/0.5/0.1. The weight hourly space velocity (WHSV) of the

total gas mixture was fixed at 60000 mL h-1

g-1

. The catalytic activity was evaluated at 1073 K

and atmospheric pressure for 4 h. Gas effluents were analyzed with a micro gas chromatograph

(Varian CP-4900). Methane and carbon dioxide consumption rate were calculated as [inlet

molar flow – outlet molar flow]/nickel weight.



The most important characterization results are given in Table 2.1. The textural properties

of the catalysts strongly depended on the surface area and pore volume of the support, and are

independent of the precursor used in each catalyst. Catalysts based on SiC presented higher

values of both surface area and total pore volume, which could be related to the lower size of

Ni particle, if compared to that of ceria-based catalysts. The metal particle size (obtained with

Chapter 2

106

the Debye–Scherrer equation using the data obtained from the XRD patterns and confirmed

with TEM pictures, as commented below) strongly depended on the nickel precursor. Thus,

very different particle sizes were obtained for ceria-supported catalysts: from 57 nm for Ni-NC

to 116 nm for Ni-CiC. ForSiC-supported catalysts, larger Ni particles were obtained for the

samples Ni-CS (70 nm) and Ni-CiS (71 nm), whereas the metal particle sizes for catalysts Ni-

NS and Ni-AS were 52 and 50 nm, respectively. This particle size order (Ni-NS ≈ Ni-AS < Ni-

CS ≈ Ni-CiS) is similar to that obtained when CeO2 is used as support and is in agreement with

other results reported in literature [32-34]. The higher size of Ni particles obtained when nickel

chloride was used as precursor could be due to the high volatility of nickel chloride in the

presence of hydrogen and hydrogen chloride, which in the initial stages of the reduction

produce the vaporization of microscopic nickel particles and then the deposit of this nickel on

the nearest neighbours. Therefore, large crystals tend to grow whereas the small ones tend to

vaporize and disappear completely [35]. When nickel citrate was used as precursor, bigger

metal particles are obtained due to the more acid character of the solution formed during the

impregnation process, as Ni dispersion is higher in basic environments [36]. Particle sizes

obtained from TEM images (Figure 2.1) were very close to that obtained from XRD.

Table 2.1. Physical properties of the catalysts.

Ni-NC Ni-AC Ni-CC Ni-CiC Ni-NS Ni-AS Ni-CS Ni-CiS

Ni loading (%) 4.1 4.0 4.7 4.5 3.9 3.7 3.7 4.8

Surface area (m2/g) 7.3 8.3 8.8 9.8 25.9 23.2 26.7 23.7

Total pore volume

(cm3/g) × 10

2

6.4

3.7 6.6 9.8 17.9 19.1 19.8 20.1


XRD (nm) 60 86 88 97 41 42 68 75


TEM (nm) 57 87 92 116 52 50 70 71


reforming process

107

Fig

ure

2.1

. T

EM

im

ages

. a)

Ni-

NC

, b

) N

i-A

C,

c) N

i-C

C,

d)

Ni-

CiC

, e)

Ni-

NS

, f)

Ni-

AS

, g

) N

i-C

S,

h)

Ni-

CiS

.

Chapter 2

108

Figures 2.2 a) and 2.2 b) show the XRD patterns for samples Ni-NC and Ni-NS before

and after reduction. All peaks associated with the support in catalyst Ni-NC presented

diffraction peaks with the following Miller indices related to the cubic lattice of pure CeO2

(CaF2 structural type): (111), (200), (220), (311), (222), (400), (331) and (420), which is also in

good agreement with data reported in literature [37]. Peaks related to the support in catalyst Ni-

NS are also indicated in Figure 2.2 b) and structurally correspond to cubic SiC (3C-type)

[38]. In the same Figures peaks associated with NiO and Ni0 are also represented.

Figure 2.2. XRD profiles where (*) denotes reflection of nickel oxide and (+) denotes

reflection of nickel metallic. a) Ni-NC, b) Ni-NS.

TPR profiles for all the catalysts considered in this study are plotted in Figures 2.3 a) and

2.3 b). Ceria-based catalysts exhibited similar reduction behaviour. The main peak around 600

K was assigned to the reduction of NiO to Ni0 [11]. Peak at approximately 1050 K was

assigned to the removal of surface oxygen species and the reduction of bulk ceria [39]. TPR

0 10 20 30 40 50 60 70 80 90

+

+

****

400222

311

220

200

200

111

Fresh

Reduced

Inte

nsi

ty (

a.u

.)

2 (º)

b)

0 10 20 30 40 50 60 70 80 90

++

**

422420

331

222

311220

Inte

nsi

ty (

a.u

.)

2º)

a)

*

Fresh

Reduced

111


reforming process

109

profiles for SiC-based catalysts were different among them depending on the metal

precursor. A peak was obtained at 633 K for samples Ni-CS and Ni-CiS, corresponding to the

reduction of bulk NiO without interaction with the support [40]. Three minor reduction peaks

were also obtained at 723, 873 and 1143 K. The former was attributed to the reduction of metal

particles with stronger interaction with the silicon carbide. The peaks at higher temperatures

(873 and 1143 K) could be attributed to the reduction of nickel silicate [41]. Reduction peaks at

similar temperatures were observed for the two other catalysts (Ni-AS and Ni-NS). Thus, a

higher one was observed at 750 K, which would be a consequence of a stronger metal-support

interaction as a consequence of the lower metal particle size, if compared to that of samples Ni-

CS and Ni-CiS. On the other hand, it seems clear, attending to the fact that the reduction peak

of bulk NiO in SiC-based catalysts appeared at higher temperatures (633 K) than that of

ceria-based ones (600 K), that the metal-support interactions in the former were stronger that

those in ceria-based ones.

Figure 2.3. Temperature-programmed reduction profiles. a) ceria-based catalysts, b) SiC-

based catalysts.

300 400 500 600 700 800 900 1000 1100 1200

Ni-CS

Ni-CiS

Ni-AS

Ni-NS

TC

D s

ign

al (

a.u

.)

Temperature (K)

SiC

b)

300 400 500 600 700 800 900 1000 1100 1200

Ni-NS

Ni-AS

Ni-NS

Ni-CiS

CeO2

TC

D s

ign

al (

a.u

.)

Temperature (K)

a)

Chapter 2

110

CO2–TPD curves, characterizing the surface basicity of the catalysts, are collected in

Figures 2.4 a) (CeO2-based catalysts) and 2.4 b) (SiC-based catalysts). The weak,

intermediate and strong basic sites present in the samples were estimated from the area under

the TPD curves for the temperature ranges of 325-500, 500–1000, and >1000 K, respectively.

Both the CO2 adsorbed on weaker sites was desorbed at low temperature whereas that adsorbed

on strong sites was desorbed at high temperature [42]. A quantitative distribution of basic sites

according to their strength is included in Table 2.2 for the different catalysts. The total amount

of desorbed CO2 was noticeably influenced by the support and the precursor utilized in each

catalyst, this being higher for ceria-based catalysts. Regarding the parent supports, it is shown

that ceria desorbed CO2 at low temperature whereas SiC did not desorb any amount of CO2

in the range of temperature analyzed, showing the basic character of the former support. CO2–

TPD profile in ceria-based catalysts showed a first desorption peak at 350 K and another one at

around 900 K. The former was assigned to a CO2 release, which was adsorbed on the support.

The latter could be attributed to the CO2 release, which was adsorbed on metallic nickel [43].

The incorporation of nickel seemed to increase the formation of intermediate basic sites and to

slightly decrease the quantity of weak basic ones. Regarding the SiC-based catalysts, the

incorporation of nickel favoured the presence of strong basic sites, but the total amount of

basic sites was lower than that in ceria-based ones.


reforming process

111

Figure 2.4. CO2 Temperature-programmed desorption profiles. a) ceria-based catalysts, b)

SiC-based catalysts.

Table 2.2. Basicity of the catalysts determined by CO2-TPD.

Basic sites (mol/g) CeO2 Ni-NC Ni-AC Ni-CC Ni-CiC SiC Ni-NS Ni-AS Ni-CS Ni-CiS

Weak (325-500 K) 5.6 3.4 4.7 3.6 2.9 - - - - -

Intermediate (500-

1000 K) 2.0 14.8 3.4 9.5 14.6 - 0.7 - - -

Strong >1000 K - - 1.4 - - - 4.0 6.4 6.4 4.9

Total 7.6 18.3 9.4 13.1 17.5 - 4.6 6.4 6.4 4.9

300 400 500 600 700 800 900 1000 1100 1200

Ni-CiC

Ni-CC

Ni-NC

Ni-AC

CeO2

b)

Ni-CS

Ni-AS

Ni-CiS

Ni-NS

TC

D s

ign

al (

a.u

.)

Temperature (K)

SiC

300 400 500 600 700 800 900 1000 1100 1200

a)

TC

D s

ign

al (

a.u

.)

Temperature (K)

Chapter 2

112


Figure 2.5 shows the catalytic performance of ceria-based catalysts in terms of methane

consumption and carbon dioxide consumption rates and H2/CO molar ratio. It can be observed

that in all cases both the initial methane consumption rate and the catalyst deactivation were

dependent on the nickel precursor used. Catalyst Ni-NC had the best catalytic performance

with an initial value of methane consumption rate close to 11.5×10-4

mol s-1

gNi-1

, which did not

practically change with the time on stream. Catalyst Ni-AC showed the lowest methane

consumption rate (around 4.5×10-4

mol s-1

gNi-1

). Sample Ni-CC showed a good initial methane

consumption rate, 8.4×10-4

mol s-1

gNi-1

, but suffered a deep deactivation, decreasing both the

methane and carbon dioxide consumption rates with the time on stream. This deactivation is

linked to the methane cracking reaction (Equation 2.5), which is favoured at high temperature

respect the Boudouard reaction, as the methane cracking is an endothermic reaction. It could be

related to the occurrence of chloride ions on the catalyst surface coming from the metal

precursor used (NiCl2·6H2O) [44] and the higher size of the metal particles in this catalyst. It

has been reported that the metal particle size is an important factor that leads to an increase in

coke formation [36, 45]. The larger the size of the particles, the higher the level of deactivation

observed. The presence of ion chloride after the calcination and reduction of the parent

catalysts was probed by ICP analysis, yielding a concentration of 6.6 wt.%. Sample Ni-CiC

showed an initial methane consumption rate close to 9.9×10-4

mol s-1

gNi-1

, but its activity

decreased clearly during the time on stream.


reforming process

113

Figure 2.5. Catalytic activity at 1073 K for: a) Ni-NC, b) Ni-AC, c) Ni-CC, d) Ni-CiC.

Reaction conditions: CH4 = 6%, CO2 = 3%, H2O = 3%, O2 = 0.6%, N2 balance, total flow rate

= 100 mL min-1

. CH4 ( ) and CO2 ( ) consumption rates vs. time on stream (left axis), and

H2/CO molar ratio ( ) vs. time on stream (right axis).

Figure 2.6 shows the catalytic performance of SiC -based catalysts in terms of methane

consumption and carbon dioxide consumption rates and H2/CO molar ratio. It can be observed

that for sample Ni-CS the initial methane consumption rate was close to 12×10-4

mol s-1

gNi-1

.

This catalyst suffered a clear deactivation. This way, at the end of the experiment the methane

consumption rate was lowered to 10.8×10-4

mol s-1

gNi-1

. This deactivation could be related, as

previously explained, to both the metal precursor (the presence of ion chloride after the

calcination and reduction steps was also probed by ICP analysis, yielding a concentration of

6.2 wt.%) and again the higher size of the metal particles. In this case, it is remarkable the

0 50 100 150 200 2500.0

2.5

5.0

7.5

10.0

12.5

d)

Co

nsu

mp

tio

n r

ate

(m

ol

s-1 g

-1 Ni )

·10

4

Time (min)

a)

0 50 100 150 200 2500.0

2.5

5.0

7.5

10.0

12.5b)

Co

nsu

mp

tio

n r

ate

(m

ol

s-1 g

-1 Ni )

·10

4

Time (min)

0 50 100 150 200 2500.0

2.5

5.0

7.5

10.0

12.5

c)

Co

nsu

mp

tio

n r

ate

(m

ol

s-1 g

-1 Ni )

·10

4

Time (min)

0 50 100 150 200 2500.0

2.5

5.0

7.5

10.0

12.5

Co

nsu

mp

tio

n r

ate

(m

ol

s-1 g

-1 Ni )

·10

4

Time (min)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

H

2/C

O M

ola

r ra

tio

0.0

0.5

1.0

1.5

2.0

2.5

3.0

H2/C

O M

ola

r ra

tio

0.0

0.5

1.0

1.5

2.0

2.5

3.0

H

2/C

O M

ola

r ra

tio

0.0

0.5

1.0

1.5

2.0

2.5

3.0

H

2/C

O M

ola

r ra

tio

Chapter 2

114

higher carbon dioxide conversion observed and the oscillatory behaviour, which is related to

changes in the oxidation state of nickel, which can easily shift from Ni0 to NiO and vice versa

[40, 41], causing the occurrence of the water gas-shift reaction (Equation 2.9):

CO + H2O

CO2 + H2 (H◦ = –37.09 kJ/mol) (Equation 2.9)

Figure 2.6. Catalytic activity at 1073 K for: a) Ni-NS, b) Ni-AS, c) Ni-CS, d) Ni-CiS. Reaction

conditions: CH4 = 6%, CO2 = 3%, H2O = 3%, O2 = 0.6%, N2 balance, total flow rate = 100 mL

min-1

. CH4 ( ) and CO2 ( ) consumption rates vs. time on stream (left axis), and H2/CO molar

ratio ( ) vs. time on stream (right axis).

This reaction is concomitant to the reforming process when Ni-based catalysts are

used [31]. Oscillations in the rate of partial oxidation of methane, one of the three reactions

involved in the tri-reforming process, have been observed for a variety of catalysts including

0 50 100 150 200 2500.0

2.5

5.0

7.5

10.0

12.5

d)

a) b)

Co

nsu

mp

tio

n r

ate

(m

ol

s-1 g

-1 Ni )

·10

4

Time (min)

0 50 100 150 200 2500.0

2.5

5.0

7.5

10.0

12.5

Co

nsu

mp

tio

n r

ate

(m

ol

s-1 g

-1 Ni )

·10

4

Time (min)

0 50 100 150 200 2500.0

2.5

5.0

7.5

10.0

12.5

c)

Co

nsu

mp

tio

n r

ate

(m

ol

s-1 g

-1 Ni )

·10

4

Time (min)

0 50 100 150 200 2500.0

2.5

5.0

7.5

10.0

12.5

Co

nsu

mp

tio

n r

ate

(m

ol

s-1 g

-1 Ni )

·10

4

Time (min)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

H

2/C

O M

ola

r ra

tio

0.0

0.5

1.0

1.5

2.0

2.5

3.0

H2/C

O M

ola

r ra

tio

0.0

0.5

1.0

1.5

2.0

2.5

3.0

H

2/C

O M

ola

r ra

tio

0.0

0.5

1.0

1.5

2.0

2.5

3.0

H

2/C

O M

ola

r ra

tio


reforming process

115

supported palladium, palladium wire and foil, supported nickel, nickel wire and foil,

nickel/chromium alloy and cobalt wire and foil [46]. It is generally thought that these

oscillations arise from repetitive cycles of oxidation and reduction of the metal surface, the

detailed nature of each oscillation depending upon the relative activities of the oxidised and

reduced surfaces.

Samples Ni-AS and Ni-NS showed a similar catalytic performance, with a higher

methane consumption rate and very low deactivation. Catalyst Ni-CiS presented a lower initial

methane consumption rate (9.8×10-4

mol s-1

gNi-1

) and a light deactivation. Anyway, the

deactivation observed in SiC-based catalysts was less pronounced than that seen in ceria-

based ones. This difference could be attributed to the higher metal-support interaction observed

in the formers [36, 45]. This fact has been reported to be one of the principal factors to limit the

deactivation in reforming processes.

The H2/CO molar ratio in the effluent depends mainly on the CO2 and H2O conversion in

the tri-reforming process. The more the H2O and the lesser CO2 are converted, the higher the

H2/CO molar ratio is attained. Therefore, the H2/CO molar ratio is a good indicator for

comparing the ability to convert CO2 in the presence of H2O over different catalysts [31]. The

kind of support seemed to affect this ratio. Thus, the higher H2/CO molar ratio were obtained

when SiC-based catalysts were used since ceria-based ones promoted the CO2 conversion,

which is related to their superior capacity for CO2 adsorption as evidenced by the CO2-TPD

results. The metal particle size also affected the H2/CO molar ratio obtained. Several studies

have pointed that larger nickel particles favoured the methane cracking [47], leading to higher

H2/CO molar ratios. This fact is in good agreement with the resulted here reported: catalysts

with higher nickel particle sizes produced a synthesis gas with higher H2/CO molar ratio. The

methane cracking would also explain the light increase in the H2/CO molar ratio with the time

on stream. If a sintering process occurred, an increase of the metal particle sizes also did,

contributing to the methane cracking process. It is also interesting to note that the water gas

shift equilibrium (Equation 2.9) also played an important role in the H2/CO molar ratio since

nickel is an active metal in this reaction.

However, there are two catalysts that did not follow this general trend: Ni-AC and Ni-CS.

The H2/CO molar ratio observed for sample Ni-AC was lower than the expected value, which

would imply that this catalyst promoted the dry reforming reaction. This fact could explain

Chapter 2

116

attending to the basicity of this catalyst. Ceria is more basic than SiC, as was probed by

CO2-TPD, but its basicity is mitigated by the acidity of the dilutions used during the

impregnation process (pH of nickel nitrate, chloride and citrate solution equal to 4.8, 4.8 and

3.0, respectively; pH of the nickel acetate solution equal to 6.2). Since CO2 has a more acid

character than water, the adsorption and reaction of the former molecule would be favoured,

being lower the H2/CO molar ratio in the effluents. This behaviour is also in agreement with

CO2-TPD results, as the peak corresponding to the desorption of CO2 obtained at the lower

temperature in ceria as the support is slightly higher than that observed in catalyst Ni-AC,

which indicates that the weak basic character of this support was not altered by the presence of

the metal. In addition, a desorption peak around 1150 K would in turn demonstrate the

presence of strong basic sites in this catalyst. Despite of that, the peak corresponding to the

desorption of CO2 obtained at intermediate temperatures, assigned to the basicity of nickel

metal particles, is lower than that desorbed in the rest of ceria-based catalysts, which is related

to its low activity in the tri-reforming process and higher selectivity towards the dry reforming

reaction.

Catalyst Ni-CS is also out of this general trend, as it led to lower molar H2/CO ratios and

an oscillatory behaviour. In this case, the oscillations are again related to the water gas shift

equilibrium, as an increase in the hydrogen production (not shown in Figures) and a decrease

in the carbon monoxide production and the carbon dioxide consumption were observed, while

the methane consumption remained almost the same. As above mentioned, this behaviour

could be related to the oxidation of Ni0, as NiO plays the role of a promoter in the water gas

shift reaction [48]. In these catalysts, the formation of NiO would be favoured due to the

presence of both chloride in the catalyst (demonstrated via ICP) and O2 in the feed gas.

However, the high quantity of hydrogen present in the reaction environment would cause its

reduction, leading to an oscillation process from Ni0 to NiO and vice versa. These oscillations

were not observed in sample Ni-CC due to the high oxygen storage and transport capacity of

ceria, which favoured the stability of the reduced state of nickel.

The reaction temperature which is necessary to obtain a conversion level of 20% as well

as the H2/CO molar ratio obtained at this level of conversion are summarized in Table 2.3 for

all the catalytic systems used in this study. It can be seen that the temperature required to reach

that value of conversion was higher for those samples with lower activity (Ni-AC, Ni-CC and

Ni-CS). Sample Ni-NC was the catalyst with the lowest temperature needed to reach this value


reforming process

117

of conversion, despite of not having the best catalytic behaviour at 1073 K. This could be

attributed to the higher contain of nickel in this sample, if compared to that of catalysts with

the best performance: Ni-NS and Ni-AS. Ceria-based catalysts led to the higher H2/CO molar

ratios, which can be attributed to their superior activity for the water-gas shift reaction

(Equation 2.9), one of the main functions of ceria in automobile three-way catalysts [49, 50]. In

all cases this molar ratio presented higher values than those obtained at 1073K, which could be

related to the exothermicity of the water-gas shift reaction.

Table 2.3. Catalytic performance at 20% of conversion: temperature needed to get this level of

conversion and H2/CO molar ratio obtained at this level of conversion.

Ni-NC Ni-AC Ni-CC Ni-CiC Ni-NS Ni-AS Ni-CS Ni-CiS

Temperature (K) 732 1035 1007 805 813 813 843 812

H2/CO molar ratio 10.0 10.8 12.9 12.1 8.1 8.9 4.2 4.0

Reaction conditions: CH4 = 6%, CO2 = 3%, H2O = 3%, O2=0.6%, N2 balance, total flow rate =

100 mL min-1

.

2.4. CONCLUSIONS

XRD and TEM analysis shown that both precursor and support had a great influence in

the metal particle size. Larger particles were obtained when ceria was used as the support, and

smaller particles when nitrate and acetate were used as precursor. Thereby, Ni-NS and Ni-AS

catalysts showed the smaller particle size. The TPR experiment for samples Ni-NS and Ni-AS,

showed a profile in which the major peak was obtained at high temperatures. This finding

means that, for these samples, there is a strong interaction between nickel and the support. A

strong metal support interaction was obtained when using SiC as support, which inhibited the

deactivation process due to coke formation. Thus, Ni-NS and Ni-AS are the catalysts with the

bigger metal support interaction, yielding a low deactivation rate. Catalysts supported over

ceria enhanced the adsorption of CO2, as proved by CO2-TPD, which yielded to a synthesis gas

with lower H2/CO molar ratio. Catalysts prepared using chloride as the precursor showed a

more intense deactivation than the other catalysts. It could be related to the occurrence of

chloride ions on the catalyst surface coming from the metal precursor used and the higher size

of the metal particles obtained when using this salt. High methane consumption rate and good

catalytic stability were obtained when nickel nitrate and nickel acetate were used to prepare

Chapter 2

118

Ni/SiC catalysts. The results showed that these latter catalysts can be considered as

promising ones for the tri-reforming process.


reforming process

119

2.5 REFERENCES


377 (2010) 16-26.


[3] M.-S. Fan, A.Z. Abdullah, S. Bhatia, Applied Catalysis B: Environmental. 100 (2010) 365-

377.

[4] A. Erdohelyi, J. Cserenyi, F. Solymosi, J. Catal. 141 (1993) 287-299.

[5] D. Qin, J. Lapszewicz, Catal. Today. 21 (1994) 551-560.

[6] M.F. Mark, W.F. Maier, J. Catal. 164 (1996) 122-130.

[7] S. Wang, G. Lu, Energy & fuels. 12 (1998) 248-256.

[8] Z. Xu, Y. Li, J. Zhang, L. Chang, R. Zhou, Z. Duan, Applied Catalysis A: General. 210

(2001) 45-53.

[9] H.J. Seo, E.Y. Yu, J. Ind. Eng. Chem. 3 (1997) 85-92.

[10] H.S. Roh, K.W. Jun, S.E. Park, J. Ind. Eng. Chem. 9 (2003) 261-266.



(2007) 153-158.


137.

[14] D.H. Prasad, H.I. Ji, H.R. Kim, J.W. Son, B.K. Kim, H.W. Lee, J.H. Lee, Applied

Catalysis B: Environmental. 101 (2011) 531-539.


(2010) 12147-12160.

[16] J. Rostrupnielsen, J.B. Hansen, J. Catal. 144 (1993) 38-49.

Chapter 2

120

[17] O. Sidjabat, D. Trimm, Top. Catal. 11 (2000) 279-282.

[18] S. Jiang, S. Badwal, Solid State Ionics. 123 (1999) 209-224.

[19] A. Effendi, Z.-G. Zhang, K. Hellgardt, K. Honda, T. Yoshida, Catal. Today. 77 (2002)

181-189.

[20] F. Aupretre, C. Descorme, D. Duprez, D. Casanave, D. Uzio, J. Catal. 233 (2005) 464-

477.

[21] H. Song, B. Mirkelamoglu, U.S. Ozkan, Applied Catalysis A: General. 382 (2010) 58-64.

[22] J. Llorca, P.R.r. de la Piscina, J.-A. Dalmon, J. Sales, N.s. Homs, Applied Catalysis B:

Environmental. 43 (2003) 355-369.


185.


51-54.

[25] D.L. Nguyen, P. Leroi, M.J. Ledoux, C. Pham-Huu, Catal. Today. 141 (2009) 393-396.


(2011) 298-302.

[27] D. Trimm, Catal. Today. 37 (1997) 233-238.

[28] A. Yee, S. Morrison, H. Idriss, J. Catal. 186 (1999) 279-295.

[29] L. Mattos, F. Noronha, J. Catal. 233 (2005) 453-463.

[30] N. Laosiripojana, S. Assabumrungrat, Applied Catalysis B: Environmental. 66 (2006) 29-

39.


[32] I. Chen, S.Y. Lin, D.W. Shiue, Industrial & Engineering Chemistry Research. 27 (1988)

926-929.


reforming process

121

[33] P. Kim, H. Kim, J.B. Joo, W. Kim, I.K. Song, J. Yi, J. Mol. Catal. A: Chem. 256 (2006)

178-183.

[34] F. Li, X. Yi, W. Fang, Catal. Lett. 130 (2009) 335-340.

[35] C. Hoang-Van, Y. Kachaya, S. Teichner, Y. Arnaud, J. Dalmon, Applied catalysis. 46

(1989) 281-296.


[37] S. Saitzek, J.F. Blach, S. Villain, J.R. Gavarri, physica status solidi (a). 205 (2008) 1534-

1539.


[39] J. Xu, B. Xue, Y.-M. Liu, Y.-X. Li, Y. Cao, K.-N. Fan, Applied Catalysis A: General. 405

(2011) 142-148.

[40] X. Zhang, D. Hayward, D.M. Mingos, Catal. Lett. 86 (2003) 235-243.


159.


104 (2011) 64-73.

[43] S.C. Dantas, J.C. Escritori, R.R. Soares, C.E. Hori, Chem. Eng. J. 156 (2010) 380-387.

[44] H.J. Park, S.H. Park, J.M. Sohn, J. Park, J.-K. Jeon, S.-S. Kim, Y.-K. Park, Bioresour.

Technol. 101 (2010) S101-S103.


(2006) 9-22.

[46] X. Zhang, C.S.-M. Lee, D.O. Hayward, D.M.P. Mingos, Catal. Today. 105 (2005) 283-

294.


200.

Chapter 2

122


(2008) 97-104.

[49] Y. Li, Q. Fu, M. Flytzani-Stephanopoulos, Applied Catalysis B: Environmental. 27 (2000)

179-191.

[50] A. Haryanto, S.D. Fernando, S.F. To, P.H. Steele, L. Pordesimo, S. Adhikari, Energy &

fuels. 23 (2009) 3097-3102.

CHAPTER 3 Methane tri-reforming over a Ni/-SiC-based

catalyst: Optimizing the feedstock composition

Resumen

Abstract

3.1. INTRODUCTION

3.2. EXPERIMENTAL




3.2.4. Experimental design



3.3.2. Statistical analysis

3.3.3. Influence of the feedstock composition

on the H2/CO molar ratio

3.3.4. Optimization of the reaction conditions

3.4. CONCLUSIONS

3.5. REFERENCES

Methane tri-reforming over a Ni/-SiC-based catalyst: Optimizing the feedstock composition

125

Resumen

En este capítulo se ha probado cómo el tri-reformado de metano es un proceso

muy eficiente a la hora de obtener un gas de síntesis que pueda ser utilizado en el

proceso Fischer-Tropsch y la síntesis de metanol. Se ha estudiado la influencia de la

composición de la corriente alimento en la conversión de metano, la relación molar

H2/CO del gas de síntesis obtenido y el calor de reacción del proceso de tri-reformado

con un catalizador Ni/-SiC. En primer lugar, se eligió la metodología del diseño

factorial para analizar la influencia de la composición del alimento sobre las variables

dependientes elegidas. A continuación, con los datos experimentales obtenidos, se

construyó un modelo cuadrático que relacionaba variables dependientes e

independientes. Se observó que tanto el caudal volumétrico de agua como el de

oxígeno tenían un efecto positivo sobre la relación molar H2/CO, mientras que el

caudal volumétrico de metano y dióxido de carbono tenían un efecto negativo. Por

último, se calculó la influencia de las variables independientes estudiadas previamente

sobre el calor de reacción del proceso, obteniéndose un óptimo energético.

Chapter 3

126

Abstract

Tri-reforming of methane has proved to be a highly efficient process for

obtaining synthesis gas suitable for use in the Fischer Tropsch process and methanol

synthesis. In this chapter the influence of the feedstock composition on methane

conversion, the H2/CO molar ratio of the synthesis gas obtained by tri-reforming of

methane and the heat released or supplied to the system with a Ni/SiC catalyst are

all described. Firstly, a factorial plus central composite design of experiments was

chosen in order to optimize the independent variables selected. Then, using the

experimental data obtained, a quadratic model was built. It was observed that the

effect of both water and oxygen volume flow on the H2/CO molar ratio was positive

while that of methane and carbon dioxide volume flow was negative. Finally, in order

to obtain an energetic optimum inside the target region, the influence of the

independent variables studied previously on the overall reaction heat was calculated.


127

3.1. INTRODUCTION

The process of activating carbon dioxide and transforming it into useful products has

gained more and more importance in recent years due to the environmental problems related to

its release into the atmosphere and the subsequent well-known global warming effects. One of

the processes currently being studied as a way of obtaining useful products from CO2 is dry

reforming of methane (Equation 3.1), which allows synthesis gas to be obtained as the main

derivative. However, this process has two main drawbacks that are hindering its use in

industry: firstly, coke formation causes the catalyst to rapidly deactivate and, secondly, a great

deal of energy is consumed as a consequence of the endothermic nature of this process.

CO2 + CH4 → 2CO + 2H2 (H◦ = 247.3 kJmol

-1) (Equation 3.1)

In order to address these problems, tri-reforming was put forward by Song in 2001 [1].

Tri-reforming of methane consists of a synergetic combination of dry reforming (Equation

3.1), steam reforming (Equation 3.2) and partial methane oxidation (Equation 3.3). Compared

with dry reforming, catalyst deactivation by carbon formation is lower in this process due to

the presence of water and oxygen (Equations (3.4-3.7)). Moreover, energy consumption is

much lower as a result of exothermic partial methane oxidation. In addition to this, said process

has a further advantage over dry reforming as the H2/CO molar ratio depends on the relative

amounts of each feedstock component, so it can be controlled by varying these gases:

H2O + CH4 → CO + 3H2 (H◦ = 206.3 kJmol

-1) (Equation 3.2)

CH4 + 1/2O2 → CO + 2H2 (H◦ = −35.6 kJmol

-1) (Equation 3.3)

2CO

C + CO2 (H◦ = −172.2 kJmol

-1) (Equation 3.4)

CH4 C + 2H2 (H

◦ = 74.9 kJmol

-1) (Equation 3.5)

C + H2O

CO + H2 (H◦ = 131.4 kJmol

-1) (Equation 3.6)

C + O2 CO2 (H

◦ = −393.7 kJmol

-1) (Equation 3.7)

As previously stated, the H2/CO molar ratio of the synthesis gas obtained from tri-

reforming depends on the feedstock composition, which makes it possible to obtain a wide

range of H2/CO molar ratios and therefore a high number of applications for this synthesis gas,

such as: production of dimethyl ether, production of fuels and high-value chemicals by means

of the Fischer Tropsch synthesis [2, 3] and fuel processing in solid oxide fuel cell and molten

Chapter 3

128

carbonate fuel cell systems. It has already been mentioned in previous papers [4, 5] that the

optimum H2/CO molar ratio for diesel production by means of the Fischer-Tropsch synthesis is

approximately 2, which is also the most suitable feedstock for methanol synthesis [5, 6].

In this study, we looked at tri-reforming using nickel as the active metal and silicon

carbide as the support. Nickel is one of the most studied metals as active phase in catalysts for

reforming processes, including steam reforming [7], dry reforming [8] and partial oxidation

[9]. It is widely accepted now that [10] despite the sound properties noble metals contain,

especially their high activity, Ni catalysts are more favourable due to the high costs and

limited availability of the former. Furthermore, Ni has appealing redox properties and, as it is

relatively inexpensive, it is easily accessible.

The development of the porous form of silicon carbide (-SiC) and its use as a catalyst

support [11] has attracted a great deal of attention within the scientific community. The

properties this material contains that make it more appealing are low specific weight, high

thermal conductivity, high mechanical strength and chemical inertness. Together, these boost

the effectiveness of the catalyst when used as a support in a heterogeneous catalyst in highly

endothermic or exothermic reactions [12], especially where it is crucial to control the

temperature inside bed catalysts. The chemical inertia of this material facilitates easy recovery

of the active phase by acidic or basic washing, thereby reducing the cost of the process [13].

Moreover, the -SiC could be re-used without any impediment as a support after recovery is

complete. This material has been selected as a support because of its aforementioned

characteristics and the satisfactory results obtained in different reactions [14, 15]. Additionally,

in a previous study our group undertook [16], reported in addition in the previous chapter;

when -SiC was tested for the first time as a catalyst support in tri-reforming and was

compared to ceria, there was stronger metal support interaction, which inhibited the

deactivation process which coke formation causes.

As previously mentioned, one of the main advantages of tri-reforming with regards to

other reforming processes is the ability of the former to control the H2/CO molar ratio of

products by adjusting said ratios of the different feedstock components, which is a highly

appealing characteristic as this property determines the potential uses the synthesis gas

obtained can have.


129

Consequently, the aim of this research was to evaluate how the molar ratios of different

reagents influenced methane conversion and the H2/CO molar ratio in the effluent gas given off

by the reactor by using the factorial design of experiments to obtain a synthesis gas suitable for

use in the Fisher Tropsch process and methanol synthesis. We selected this methodology due to

the advantages it offers when compared to classical approaches as the quantity of experiments

needed to determine the influence of the same number of factors with the same precision is

lower and, in addition to this, the interaction between factors can be determined. This approach

has been employed in several chemical systems [17-19] and in this study it was used for

determining the best feedstock composition in tri-reforming for the first time. The values of the

independent variables studied in this research were selected according to both the best

operating conditions reported elsewhere [20] and the inherent limitations of the experimental

set up here used for catalysts testing. The experimental results obtained were fitted to a second-

order polynomial equation by using the experimental design methodology.

3.2. EXPERIMENTAL


The catalyst used in this study was prepared using the impregnation method with nickel

nitrate Ni(NO3)2·6H2O (Panreac) as the precursor and -silicon carbide (SICAT) as the

support. After impregnation, the catalyst was dried in air overnight at 393 K after which

calcination in air took place at 1173 K for 2 h. The Ni loading was set at approximately 5 wt%

and Ni metal loading was determined by atomic absorption (AA) spectrophotometry, using a

Spectra 220FS analyzer. For this reason, the catalyst (ca. 0.5 g) was previously treated in 2 mL

HCl, 3 mL HF and 2 mL H2O2 and then by microwave digestion (523 K).


The fresh catalyst surface areas and pore volumes were analyzed using the N2 adsorption–

desorption isotherm and the BET surface area and total pore volume of the catalyst were

determined by nitrogen adsorption/desorption isotherms measured at 77 K using a Quadrasorb

3SI sorptometer apparatus. Prior to gas adsorption measurements, the catalyst was degassed at

523 K under vacuum for 12 h. Finally, the total pore volume was calculated at a relative

pressure of P/Po = 0.99.

Chapter 3

130

A Temperature programmed reduction (TPR) analysis in a 17.5%H2/Ar (60 cm3 min

−1)

gas mixture was carried out in a Micromeritics AutoChem 2950 HP unit with TCD detection.

The samples (ca. 0.15 g) were then analyzed in a U-shaped reactor and heated in the gas

mixture from room temperature to 1173 K with a heating rate of 10 K min−1

.

CO2 temperature-programmed desorption (CO2-TPD) was also carried out in the

Micromeritics AutoChem 2950 HP unit, equipped with a TCD detector. 0.15 g were first

reduced and then pre-treated in He. After cooling to 323 K, CO2 adsorption was carried out

with a 30 mL min-1

flow of CO2 (99.99% purity, Praxair certified) for 30 min. Finally, the

weakly adsorbed carbon dioxide was removed by a steady flow of He for another 30 min. The

sample was then heated in 50 mL min-1

of He at a rate of 10 K/min up to 1173 K.

A Philips X’Pert instrument was used to carry out the XRD analyses, using nickel-filtered

Cu Kα radiation. Afterwards, samples were scanned at a rate of 0.02°step−1

over the 5° ≤ 2θ ≤

90° range (scan time = 2 s step−1

).

Conventional TEM analysis was carried out with a JEOL JEM-4000EX unit operating at

400 kV. Next, samples for this analysis were suspended in acetone, diluted by ultrasonic

dispersion and placed on copper grids with a holey-carbon film support. Mean nickel particle

size was then calculated as the surface-area weighted diameter ( sd ), according to Equation

3.8:

2

ii

i

3

ii

sdn

dn

d

(Equation 3.8)



The catalytic behaviour was tested in a tubular quartz reactor (45 cm long and 1 cm

across), with the catalyst placed on a fritted quartz plate located at the end of the reactor. The

reaction temperature was then measured with a K-type thermocouple (Thermocoax) placed

inside the inner quartz tube. The flow of the reaction gases was controlled by mass flowmeters

(Brooks 5850 E and 5850 S) and the water content was regulated by a saturation system. The

reaction feedstock was formed with the necessary quantity of CH4, CO2, H2O and O2 required


131

for each point of the factorial design and the necessary quantity of N2 to keep the total

feedstock flow constant at 100 NmL min-1

. Next, the reactor was loaded with 0.1 g of catalyst,

which yielded a GHSV of 60000 h-1

. The experiments were carried out at 1073 K and at

atmospheric pressure, over 4 hours, which was deemed sufficient time to reach a steady state;

to obtain the catalytic results used in the factorial design. No significant deactivation was

observed in any of the experiments. The -SiC (used as the catalyst support) had already

proved to be inert in a previous test, and, hence, no catalytic activity was observed. Finally, a

micro gas chromatograph (Varian CP-4900) with two analytical columns (with each one

having its own TCD analyzer) was used as the analysis system.

3.2.4. Experimental design

The main advantage factor design offers with this kind of research is that a lower quantity

of experiments are needed than with the classical one-at-a-time experiments to obtain the same

precision. In addition with the former it is possible to estimate any interaction between factors,

which cannot be determined with the latter. According to Deming and Palasota [21], for

chemical systems in which there is more than one factor, the application of properly designed

experiments and adequate multifactor models, such as a central composite design, is essential.

This consists of a two-level full factorial design extended with a star design which allows the

intercept, slope, curvature, and interaction terms to be estimated and can be used to describe

the system with the significance of each term being characterized by the statistical beta

coefficients. In a coded factor space, the star points, which determine the effect of the quadratic

terms in the mathematical model that link independent and dependents variables; are usually

located at a distance (Equation 3.9)

2 = (

12 2·2 kk n ) (Equation 3.9)

from the centre, where k is the number of factors (4) and n is the total number of

experiments which is determined using the expression (Equation 3.10):

n = 2k + 2·k + C (Equation 3.10)

where C is the repeated number of experiments at the centre point. Three different groups

of points represent this design: two-level factorial (coded 1), star (coded ), and centre

points (coded 0). The star points allow the curvature in the model to be estimated and the

Chapter 3

132

centre point is repeated C times to estimate the experimental reproducibility. The results from a

central composite design can be transferred to a linear model or response surface, which is

adequate for describing a wide variety of multifactor chemical systems which take the

following form (Equation 3.11):

Yi = ii

k

ix·

1

2

1· iii

k

ix

jiij

k

j

k

ixx ··

11

(Equation 3.11)

In this case, Yi are the responses; 0 is the intercept (which is the fitted response value at

the design centre); i are the slopes with respect to each of the four factors; ii are curvature

terms; ij are the interaction terms; and xi are the factors being studied. The response function

coefficients were determined by regression using the experimental data and Statgraphics

Centurion XVI.I. software.

Firstly, it is important to determine whether the independent variables had a significant

influence over the dependent ones or not. For this purpose, a 2k factorial design was carried out

(which included 4 centre points in order to estimate the extent of experimental error and also to

determine whether the linear model was adequate or not) by calculating the curvature effect

[22]. This parameter could be calculated as the average result given by the centre points minus

the average result for each one of the experiments that the factorial design was made up of

[23]. This value must be compared with the confidence curvature interval and if the curvature

has a higher value (in absolute value terms) then a linear model does not adequately describe

the experimental data and it would be more appropriate to use a quadratic term model. In this

study we compiled 28 experiments, including the 24 experiments that corresponded to the

factorial design, 4 central points and 8 star points. The independent variables studied were the

CH4, CO2, H2O and O2 volume flows in the feedstock, whilst the total flow was kept constant.

The order in which the experiments were carried out was determined randomly, thus avoiding

possible systematic experimental errors in the results obtained. The dependent variables

studied were methane conversion and the H2/CO molar ratio in the reactor exit stream. As

mentioned in the introduction, the H2/CO molar ratio is a key factor in determining the

potential uses of the synthesis gas obtained.


133



The main characterization results can be seen in Table 3.1 and further details about the

characterization of this catalyst can be found in the previous chapter.

Table 3.1. Physical properties of the catalyst.

Characterization parameter Numerical value

Ni loading (%) 3.9

Surface area (m2 g

-1) 25.9

Total pore volume (cm3 g

-1) x10

2 17.9

Particle diameter from TEM (nm) 41

Particle diameter from XRD (nm) 52

Total basic sites (mol g-1

) 4.6

Reduction Temperature (K) 973

3.3.2. Statistical analysis

Table 3.2 shows the range of values studied for each independent variable: CH4, CO2,

H2O and O2 volume flow. Tri-reforming catalytic results from the 2k full factorial experiments

and the central points appear in Table 3.3 (Experiments 1-20).

Table 3.2. Factor levels.

Variable low (-) high (+) center (0) axial (-) axial (+)

CH4 NmL min-1

4.5 7.5 6 3.59 8.41

CO2 NmL min-1

2 4 3 1.39 4.61

H2O NmL min-1

2 4 3 1.39 4.61

O2 NmL min-1

0.5 2 1.25 0.04 2.46

Chapter 3

134

Table 3.3. Central composite design results.

Experiment Run CH4 CO2 H2O O2 CH4 conversion

(%)

H2/CO

molar ratio

24 factorial

design

1 6 7.50 4.00 4.00 2.00 88.29 2.23

2 12 7.50 4.00 4.00 0.50 83.54 2.07

3 2 7.50 4.00 2.00 2.00 89.01 1.97

4 5 7.50 2.00 4.00 2.00 89.26 2.27

5 10 4.50 4.00 4.00 2.00 95.92 2.58

6 9 7.50 4.00 2.00 0.50 91.79 1.67

7 16 7.50 2.00 4.00 0.50 88.68 2.22

8 14 7.50 2.00 2.00 2.00 87.11 2.32

9 11 4.50 4.00 4.00 0.50 78.67 2.00

10 8 4.50 4.00 2.00 2.00 93.71 2.17

11 20 4.50 2.00 4.00 2.00 90.87 3.27

12 17 7.50 2.00 2.00 0.50 87.43 2.09

13 19 4.50 4.00 2.00 0.50 89.60 2.03

14 18 4.50 2.00 4.00 0.50 87.03 2.46

15 3 4.50 2.00 2.00 2.00 93.36 2.22

16 1 4.50 2.00 2.00 0.50 94.13 2.18

4 central

points

17 4 6.00 3.00 3.00 1.25 91.29 2.14

18 7 6.00 3.00 3.00 1.25 91.81 1.97

19 13 6.00 3.00 3.00 1.25 91.77 1.99

20 15 6.00 3.00 3.00 1.25 91.00 2.11

8 star

points

21 27 8.41 3.00 3.00 1.25 75.46 2.13

22 23 3.59 3.00 3.00 1.25 64.55 2.38

23 25 6.00 4.61 3.00 1.25 53.74 1.98

24 22 6.00 1.39 3.00 1.25 49.43 2.81

25 28 6.00 3.00 4.61 1.25 82.02 2.53

26 24 6.00 3.00 1.39 1.25 79.18 2.04

27 21 6.00 3.00 3.00 2.46 91.66 2.26

28 26 6.00 3.00 3.00 0.04 77.67 2.17


135

The results obtained were processed using statistical analysis with a 95% confidence

level. In this way, we have calculated the effect of each one of the independent variables on the

dependent variables studied and their interactions (Table 3.4). These values must be higher

than the confidence interval ( 3.016 for methane conversion and 0.137 for the H2/CO

molar ratio) in order to be statistically significant, but, as we can see in the tables, as regards

methane conversion, only the O2 volume flow and the H2O-O2 interaction were statistically

significant, because their effect was higher than that of the confidence interval. However, all

the variables were significant in the case of the H2/CO molar ratio, although their interactions

were not. As for the methane and carbon dioxide volume flow effects, these were negative,

which brought about a lower H2/CO molar ratio when these variables increased in value, which

can probably be attributed to the predominance of dry reforming (Equation 3.1) under these

conditions. However, the effects of the water and oxygen volume flows were positive, which

brought about a higher H2/CO molar ratio, as a consequence of the predominance of steam

reforming (Equation 3.2) and oxidation of CH4 to CO2 and H2 (Equation 3.12).

CH4 + 3/2O2 → CO2 + 2H2 (H◦ = −518.74 kJmol

-1) (Equation 3.12)

Therefore, the influence of the different independent variables on methane conversion

was negligible. Subsequently, the H2/CO molar ratio was the only dependent variable

considered for the remainder of this study.

It must be pointed out that curvature and its confidence interval are two highly important

factorial design parameters as they determine whether a linear model is sufficient to describe

the system or whether a quadratic model is needed. A curvature of 0.181 and a confidence

curvature interval of 0.153 were obtained in the H2/CO molar ratio analysis. Thus, the

curvature turned out to be significant and a quadratic model was required. The H2/CO molar

ratio and methane conversion obtained for the experiments corresponding to the 8 star points of

the composite design are shown in Table 3.3 (experiments from 21 to 28). Then, by including

the star experiments, the quadratic model representing the H2/CO molar ratio appeared as

follows (Equation 3.13):

Chapter 3

136

Table 3.4. Factorial design statistical analysis.

H2/CO = 3.220 – 0.080·CH4 – 0.686·CO2 + 0.152·H2O + 0.157·O2 + 0.011·(CH4)2 +

0.017·CH4·CO2 – 0.040·CH4·H2O – 0.047·CH4·O2 + 0.079·(CO2)2 – 0.023·CO2·H2O +

0.004·CO2·O2 + 0.036·(H2O)2 + 0.074·H2O·O2 + 0.018·(O2)

2 (Equation 3.13)

where CH4, CO2, H2O and O2 denote the volume flow (NmL min-1

) of each feedstock

component and H2/CO denotes the molar ratio of these in the synthesis gas (exhaust stream

from the reactor). Figure 1 shows the experimental H2/CO molar ratio was accurately

predicted, with an average error rate of 5.2%.

CH4 conversion (%) H2/CO

Main effects

4CH -2.273 4CH = -0.259

2CO = -0.918 2CO = -0.289

OH2 = -2.985

OH2 = 0.307

2O = 3.332 2O = 0.288

Interactions

24 COCH = 0.958 24 COCH = 0.051

OHCH 24 = 1.592

OHCH 24 = -0.120

24 OCH = -2.777 24 OCH = -0.106

OHCO 22 = -1.438

OHCO 22 = -0.046

22 OCO = 2.500 22 OCO = 0.006

22 OOH = 3.270 22 OOH = 0.111

Significance test

(confidence level: 95%)

Mean responses 91.470 2.052

Standard deviation t = 3.182

s = 1.896

t = 3.182

s = 0.086

Confidence interval 3.016 0.137

Significant variables O2, H2O-O2 CH4, CO2, H2O, O2

Significance of curvature

Curvature 0.181

Confidence curvature interval 0.154

Significance Yes


137

Figure 3.1. Experimental vs. predicted H2/CO molar ratio.

3.3.3. Influence of the feedstock composition on the H2/CO molar ratio

The influence of the feedstock composition (CH4, CO2, H2O and O2) on the H2/CO molar

ratio obtained in tri-reforming was established in accordance with the statistical analysis

carried out in the previous section.

The 3D responses of the H2/CO molar ratio as a function of the independent variables

considered here are shown in Figures 2-4. They were drawn by keeping the CO2 volume flows

constant at 2 NmL min-1

(Figure 3.2), 3 NmL min-1

(Figure 3.3) and 4 NmL min-1

(Figure 3.4).

These values correspond with the coded values for the CO2 volume flows of 1, 0 and +1,

respectively. The shaded region corresponds to the CH4, O2 and H2O volume flows in the

feedstock that yielded a synthesis gas with H2/CO molar ratio values ranging from 1.9 to 2.1.

As mentioned above, these values were in keeping with the requirements laid down for the

production of both C5+, by means of the Fischer-Tropsch process, and methanol synthesis.

1.0 1.5 2.0 2.5 3.0 3.5 4.0

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Ex

peri

men

tal

H2/C

O m

ola

r ra

tio

Predicted H2/CO molar ratio (Equation 3.13)

Chapter 3

138

Figure 3.2. Effect of the CH4, O2 and H2O volume flows on the H2/CO molar ratio at 2.0 NmL

min-1

CO2 volume flow. The shaded area indicates the region with a value of H2/CO molar

ratio ranging from 1.9 to 2.1.

The influence of methane concentration in the feedstock can be clearly observed in Figure

3.2. As one can see, when the methane volume flow shows low values, the region

corresponding to the desired H2/CO molar ratio (about 2) is very small and low values of water

and oxygen volume flows were obtained. This may be because, for the low CO2 volume flow

(2NmL min-1

) considered, the high O2 and H2O volume flows would hinder the dry reforming

reaction as they would compete with steam reforming and the partial oxidation of methane. If

the methane available reacts predominantly with water and oxygen, then the H2/CO molar ratio

obtained will be higher than the value desired if one looks at the stoichiometry of these

reactions. When the CH4 volume flow was increased, more methane would be available and

more CO2 could react, yielding, (once steam reforming and the partial oxidation have reached

their equilibrium) the desired H2/CO molar ratio. Regarding the water volume flow, its

contribution to the overall H2/CO molar ratio is determined by the steam reforming reaction, so

the more water in the feed, the higher the H2/CO molar ratio in the exhaust stream from the

reactor is detected. On the other hand, as can be seen in the central area of Figure 3.2, the

higher the water volume flow, the smaller the region that met the required H2/CO molar ratio.

Huang et al. [24] found that by increasing the quantity of water in the feedstock

(H2O/(CH4+CO2+H2O)) from 1/9 to 4/9 and by keeping the CH4/(CO2+H2O) molar ratio

constant at 2/2.5, methane conversion of the dry + steam reforming rose slightly from 97.7%


139

to 98.9% and CO2 conversion fell from 94.1% to 84%. Sun et al. [25] also researched how

different quantities of steam at atmospheric pressure influenced catalytic performance, using a

gas flow for the feedstock in which the molar ratio of CH4:CO2:O2 was 2:1:0.6. Their results

showed that the addition of steam led to a rise in CH4 conversion and a fall in CO2 conversion

(from 86% to 57.3%) while the H2O/CH4 ratio rose from 0 to 0.5. This is logical since a higher

quantity of water in the catalyst environment would favour steam reforming of methane

thereby hindering the CO2 conversion reaction. From a thermodynamics viewpoint, it is

favoured the reaction between H2O and CH4 instead of CO2 and CH4 [25] which would

indicate that dry reforming would be depressed by increasing steam volume flow values.


min-1



Likewise, the oxygen volume flow had a similar influence to that of water. An increase in

this variable yielded a higher H2/CO molar ratio, which decreased the size of the target region.

This finding can be corroborated if we observe that the partial oxidation of methane

predominated over the other reactions (dry and steam reforming). In this way, the higher the

quantity of oxygen there was, the less methane available in the reaction environment there was.

As a consequence, dry reforming would be displaced by steam reforming, yielding a higher

H2/CO molar ratio in the exhaust stream. Song et al. [20] reported a similar trend in tri-

reforming, analyzing the influence of oxygen concentration when the H2O/CO2 ratio was 1 and

the (H2O + CO2 + 2O2)/CH4 ratio was set at 1.2. The oxygen clearly affected catalytic

Chapter 3

140

behaviour, as a higher quantity of it in the feedstock meant lower carbon dioxide conversion,

which dropped from 78.4% at CH4:H2O:CO2:O2 = 1:0.6:0.6:0 to 67.8% at CH4:H2O:CO2:O2 =

1:0.27:0.27:0.33.

Despite these findings, one can observe in Figure 3.2 a region where the H2/CO molar

ratio was above 2.1 for high values of methane volume flow and low water and oxygen values.

Under these conditions, there would be an excess of methane, which would favour the methane

cracking reaction (Equation 3.5). It is well known that nickel is an active metal for the latter

reaction [26, 27]. As a consequence, the H2/CO molar ratio was higher than the desired value,

because no CO was produced in this reaction.

The target region obtained when the CO2 volume flow was 3 NmL min-1

is plotted in

Figure 3.3, and turned out to be greater than that in Figure 3.2 (where there is a lower value of

CO2 volume flow). Under these conditions, with a higher content of CO2, the dry reforming

would be promoted, allowing higher values of water and oxygen volume flow to produce the

required H2/CO molar ratio, as the major extension of steam reforming and partial oxidation of

methane is mitigated in the H2/CO molar ratio with a major extension of dry reforming.

Likewise, the influence of methane, water and oxygen volume flows on the H2/CO molar ratio

can be discussed in a similar way.

As expected, the higher the CO2 volume flow, the greater was the target area represented

(Figure 3.4). Here, lower values of the methane, oxygen and water volume flows caused the

values of H2/CO molar ratio in the exhaust stream to be lower than desired. Under these

conditions, dry reforming would predominate.


141


min-1



3.3.4. Optimization of the reaction conditions

Finally, this study was completed by calculating the reaction heat required or realized in

all the experiments carried out, as a way of perfecting the reaction conditions. Taking as a

reference the target region of the H2/CO molar ratio values (ranging from 1.9 and 2.1), the aim

of this part of the study was to select the optimum conditions (in terms of volume flows of the

different components entering the reactor) which would make more exothermic/less

endothermic the tri-reforming process. As mentioned above, tri-reforming involves both

exothermic and endothermic reactions (equations 3.1 to 3.7). As expected, any modification of

the feedstock composition could alter the energy balance of the system as listed in Table 3.5

where the reaction heat corresponding to each experiment was calculated by using the process

simulation tool Aspen HYSYS. For this purpose, the main tri-reforming reactions of the

process (Equations 3.1-3.3) and the total methane combustion (Equation 3.14) were

considered.

Chapter 3

142

Table 3.5. Factorial design for the reaction heat results and optimized variables.

Experiment Run CH4 CO2 H2O O2 Reaction heat (kJ mol-1

)

24 factorial design

1 6 7.50 4.00 4.00 2.00 1140.0

2 12 7.50 4.00 4.00 0.50 2872.0

3 2 7.50 4.00 2.00 2.00 1281.0

4 5 7.50 2.00 4.00 2.00 1021.0

5 10 4.50 4.00 4.00 2.00 -356.6

6 9 7.50 4.00 2.00 0.50 3221.0

7 16 7.50 2.00 4.00 0.50 2854.0

8 14 7.50 2.00 2.00 2.00 1316.0

9 11 4.50 4.00 4.00 0.50 1269.0

10 8 4.50 4.00 2.00 2.00 -104.5

11 20 4.50 2.00 4.00 2.00 -267.3

12 17 7.50 2.00 2.00 0.50 2174.0

13 19 4.50 4.00 2.00 0.50 1725.0

14 18 4.50 2.00 4.00 0.50 1415.0

15 3 4.50 2.00 2.00 2.00 -158.8

16 1 4.50 2.00 2.00 0.50 1679.0

4 central points

17 4 6.00 3.00 3.00 1.25 1388.0

18 7 6.00 3.00 3.00 1.25 1329.0

19 13 6.00 3.00 3.00 1.25 1397.0

20 15 6.00 3.00 3.00 1.25 1299.0

8 star points

21 27 8.41 3.00 3.00 1.25 1858.0

22 23 3.59 3.00 3.00 1.25 -296.4

23 25 6.00 4.61 3.00 1.25 1663.0

24 22 6.00 1.39 3.00 1.25 83.6

25 28 6.00 3.00 4.61 1.25 1787.0

26 24 6.00 3.00 1.39 1.25 781.1

27 21 6.00 3.00 3.00 2.46 67.9

28 26 6.00 3.00 3.00 0.04 2352.0

Variable CH4 CO2 H2O O2 H2/CO Reaction heat (kJ mol-1

)

Optimized value 3.59 4.12 1.39 2.11 2.10 -496.4


143

CH4 + 2O2 → CO2 + 2H2O (H◦ = −880 kJ mol

-1) (Equation 3.14)

With the reaction heat values calculated for all the experiments carried out previously, the

following second order polynomial was obtained:

Reaction heat (kJ mol-1

) = -559.124 + 593.887·CH4 + 487.064·CO2 - 590.437·H2O -

1288.53·O2 - 39.4551·(CH4)2 + 53.5·CH4·CO2 + 40.65·CH4·H2O + 34.0111·CH4·O2 -

2.9706·(CO2)2 - 75.6625·CO2·H2O - 76.3333·CO2·O2 + 106.058·(H2O)

2 - 33.9667·H2O·O2 +

137.557·(O2)2 (Equation 3.15)

where CH4, CO2, H2O and O2 denote the volume flow (NmL min-1

) of each component in

the feedstock. Figure 3.5 shows the degree of accuracy of the reaction heat calculated by Aspen

HYSYS data and those predicted by this equation.

Figure 3.5. Experimental vs. predicted overall reaction heat.

Taking into account equations 13 and 15, the target region of the H2/CO molar ratio

values and the limits of the independent variables (+ it is possible to establish the

optimum conditions to strive for in terms of energy consumption. For this purpose, Microsoft

Excel Solver was used, with the optimal set of values of CH4, CO2, H2O and O2 being the

volume flows collected in Table 3.5. As could be expected, the optimum condition, which gave

0 500 1000 1500 2000 2500 3000 3500

0

500

1000

1500

2000

2500

3000

3500

Calc

ula

ted

reacti

on

heat

(KJ

mo

l-1)

Predicted reaction heat (KJ mol-1) (Equation 3.15)

Chapter 3

144

rise to a H2/CO value equal to 2.1, required a high O2 concentration which was involved in the

main exothermic reactions (partial oxidation and methane combustion).

The conditions selected in this section were experimentally tested, at 1073 K and 1 atm,

for 8 h. Figure 3.6 plots the methane and carbon dioxide consumption rates and the H2/CO

molar ratio in the exhaust given out from the reactor. It can be seen that the experimental value

of the H2/CO molar ratio was close to the predicted one. In addition, catalyst deactivation can

be observed with the time on stream, with the methane consumption rate varying from 4.5·10-4

mol s-1

gNi-1

to 3.7·10-4

mol s-1

gNi-1

. It is also noteworthy that the CO2 consumption rate value

was low when compared to that of methane so a synthesis gas with a CO2/CO molar ratio close

to 1.5 was yielded, which is a parameter of great importance depending on the destination of

the synthesis gas obtained, as CO2 has no influence in Fischer-Tropsch when Fe was used as

catalyst [28] but has a negative influence in the Fischer-Tropsch process when using Co

catalysts [29]. Anyway CO2 could be hydrogenated with CO in a Fischer-Tropsch reactor over

cobalt catalyst, especially in the case of high content of CO2 [30]. As expected, the high O2

concentration in the feedstock meant less methane was available for dry reforming and

favoured both partial oxidation and total combustion of methane.

Figure 3.6. Catalytic activity at 1073 K. Reaction conditions: CH4 = 3.59%, CO2 = 4.12%,

H2O = 1.39%, O2 = 2.11%, N2 balance, total flow rate = 100 NmL min-1

. CH4 ( ) and CO2 ( )

consumption rates vs. time on stream (left axis), and H2/CO molar ratio ( ) vs. time on stream

(right axis).

0 100 200 300 400 500

0

1

2

3

4

5

6

7

8

9

10

Time (min)

Co

nsu

mp

tio

n r

ate

(m

ol

s-1 g

-1 Ni )

·10

4

0.0

0.5

1.0

1.5

2.0

2.5

3.0

H2/C

O M

ola

r ra

tio


145

3.4. CONCLUSIONS

Tri-reforming of methane with a Ni/-SiC catalyst has proven to be a sound way of

obtaining a synthesis gas suitable for the Fischer Tropsch process and methanol synthesis. In

this chapter the influence of the feedstock composition on methane conversion, the H2/CO

molar ratio of the synthesis gas obtained by tri-reforming of methane and the heat released or

supplied to the system have all been described. For this purpose, a factorial plus central

composite design of experiments was set up to optimize the variables.

Methane conversion was not influenced by the CH4, CO2, H2O and O2 volume flows in

the range of values and reactions conditions considered. However, the H2O and O2 volume

flows did have a positive influence on the H2/CO molar ratio, whereas the CH4 and CO2

volume flows had a negative influence on it.

On a final note, this study was completed by calculating the reaction heat required or

realized in all the experiments carried out, as a way of perfecting the reaction conditions. As

expected, the optimum condition which caused the H2/CO value to be equal to 2.1 required a

high O2 concentration which was involved in the main exothermic reactions (partial oxidation

and methane combustion). The conditions selected were experimentally tested with catalyst

deactivation being observed with the time on stream.

Chapter 3

146

3.5 REFERENCES



137.


(2007) 153-158.

[4] S. Tang, J. Lin, K. Tan, Catal. Lett. 51 (1998) 169-175.

[5] M.A. Peña, J.P. Gómez, J.L.G. Fierro, Applied Catalysis A: General. 144 (1996) 7-57.

[6] A.P.E. York, T. Xiao, M.L.H. Green, Top. Catal. 22 (2003) 345-358.


B: Environmental. 101 (2011) 531-539.


377 (2010) 16-26.


(2010) 12147-12160.

[10] J. Rostrupnielsen, J.B. Hansen, J. Catal. 144 (1993) 38-49.


185.


51-54.



(2011) 298-302.

[15] P. Nguyen, J.M. Nhut, D. Edouard, C. Pham, M.J. Ledoux, C. Pham-Huu, Catal. Today.

141 (2009) 397-402.


147



[17] L. Sánchez, E. Lacasa, M. Carmona, J.F. Rodríguez, P. Sánchez, Industrial & Engineering

Chemistry Research. 47 (2008) 9783-9790.

[18] H. Sharifi Pajaie, M. Taghizadeh, Chemical Engineering & Technology. 35 (2012) 1857-

1864.

[19] S. Endoo, K. Pruksathorn, P. Piumsomboon, Renewable Energy. 35 (2010) 807-813.


[21] J.A. Palasota, S.N. Deming, J. Chem. Educ. 69 (1992) 560.

[22] R.O. Kuehl, Diseño de experimentos, Principios estadísticos de diseño y análisis de

investigación, second ed., Thomson Learning, Mexico, DF, 2001.

[23] A. Casas, C.M. Fernández, M.J. Ramos, Á. Pérez, J.F. Rodríguez, Fuel. 89 (2010) 650-

658.

[24] B. Huang, X. Li, S. Ji, B. Lang, F. Habimana, C. Li, Journal of Natural Gas Chemistry. 17

(2008) 225-231.


[26] A. Amin, W. Epling, E. Croiset, Industrial & Engineering Chemistry Research. 50 (2011)

12460-12470.

[27] M. Ermakova, D.Y. Ermakov, G. Kuvshinov, Applied Catalysis A: General. 201 (2000)

61-70.

[28] T. Riedel, M. Claeys, H. Schulz, G. Schaub, S.-S. Nam, K.-W. Jun, M.-J. Choi, G. Kishan,

K.-W. Lee, Applied Catalysis A: General. 186 (1999) 201-213.

[29] S.-M. Kim, J.W. Bae, Y.-J. Lee, K.-W. Jun, Catal. Commun. 9 (2008) 2269-2273.

[30] Y. Yao, D. Hildebrandt, D. Glasser, X. Liu, Ind. Eng. Chem. Res. 49 (2010) 11061-11066.

CHAPTER 4 Influence of alkaline and alkaline-earth

cocations on the performance of Ni/SiC catalysts in the methane tri-reforming reaction

Resumen

Abstract

4.1. INTRODUCTION

4.2. EXPERIMENTAL







4.3.3. Influence of the Mg/Ni molar ratio.

4.4. CONCLUSIONS

4.5. REFERENCES

Influence of alkaline and alkaline-earth cocations on the performance of Ni/-SiC catalysts in

the methane tri-reforming reaction

151

Resumen

En este capítulo se ha analizado la influencia de promotores alcalinos (Na, K) y

alcalinotérreos (Mg, Ca) sobre el comportamiento de catalizadores Ni/SiC

utilizados en el proceso de tri-reformado de metano. Los promotores fueron añadidos

mediante el método de co-impregnación con Ni, preparando catalizadores con

diferente relación promotor/Ni. Los catalizadores fueron caracterizados mediante

AAS, TPR, adsorción de N2, TPD de CO2 y DRX después de haber sido calcinados,

así como mediante DRX y TPO después de reacción. Se analizó el efecto de los

promotores en la velocidad de oxidación del SiC, siendo ésta mayor al añadir Na o

K. La presencia de Mg permitía una mayor actividad catalítica y estabilidad (con una

menor velocidad de formación de coque) debido a una disminución del tamaño de

partícula de Ni, una fuerte interacción entre el níquel y el promotor, y un aumento de

la basicidad del catalizador. Los catalizadores con relación molar Ni:Mg 2/1 y 1/1

mostraron el mejor comportamiento en cuanto a actividad, estabilidad y formación de

coque. Estos catalizadores fueron seleccionados como buenos candidatos para llevar a

cabo el proceso de tri-reformado de metano.

Chapter 4

152

Abstract

The influence of alkaline (Na, K) and alkaline earth (Mg, Ca) cocations on the

behaviour of Ni/SiC catalyst for the tri-reforming of methane has been evaluated in

the present chapter. The cocations were loaded by co-impregnation with Ni, using

different cocation/Ni ratios. Catalysts were characterized by AAS, TPR, N2

adsorption, CO2-TPD and XRD after calcination, as well as by XRD and TPO after

reaction. It was analyzed the effect of the cocations on the SiC oxidation rate,

which was increased when Na or K were loaded. The presence of Mg led to a high

catalytic performance and stability (with a lower coke formation) since it provoked a

decrease of Ni particle size and an increase of both the interaction between nickel and

promoter and the catalyst basicity. Catalysts with Ni:Mg molar ratios of 2/1 and 1/1

showed the best performance in terms of activity and stability and formation of coke.

These catalysts were considered good candidates for the tri-reforming of methane.



153

4.1. INTRODUCTION

Tri-reforming of methane is an interesting process, as production of synthesis gas from

carbon dioxide and methane helps to resolve two main problems: first, these two gases have a

well known green house effect, so that this reaction decreases their emission to the atmosphere;

second, synthesis gas is the raw material for many chemical process, e.g. production of

valuable chemical compounds through Fischer-Tropsch synthesis [1, 2], dimethyl ether

synthesis [3], or methanol production [4, 5].

The goal of this process is to avoid the limitations related to each one of the three

reforming reactions involved in this process: steam reforming (Equation 4.1), dry reforming

(Equation 4.2), and partial oxidation of methane (Equation 4.3).

H2O + CH4 → CO + 3H2 (H◦ = 206.3 kJmol

-1) (Equation 4.1)

CO2 + CH4 → 2CO + 2H2 (H◦ = 247.3 kJmol

-1) (Equation 4.2)

CH4 + 1/2O2 → CO + 2H2 (H◦ = −35.6 kJmol

-1) (Equation 4.3)

The properties that make interesting methane tri-reforming are: a higher resistance against

coke deactivation compared to dry reforming, due to the presence in the reaction environment

of oxidants (H2O and O2) that could eliminate the coke previously generated (Equations 4.4-

4.7); it is less endothermic than dry or steam reforming as partial oxidation is very exothermic

and mitigates the endothermicity of the other two processes; and finally the H2/CO molar ratio

could be controlled by modifying the reagents ratio, yielding a synthesis gas with a H2/CO

molar ratio around 2.

2CO

C + CO2 (H◦ = −172.2 kJmol

-1) (Equation 4.4)

CH4 C + 2H2 (H

◦ = 74.9 kJmol

-1) (Equation 4.5)

C + H2O

CO + H2 (H◦ = 131.4 kJmol

-1) (Equation 4.6)

C + O2 CO2 (H

◦ = −393.7 kJmol

-1) (Equation 4.7)

Nickel has been selected as the active phase for different reforming reactions by many

authors [6-9] due to its low cost compared to other metals and its high activity. Silicon carbide

exhibits a high thermal conductivity, a high resistance towards oxidation, a high mechanical

strength, chemical inertness and average surface area (around 25 m2/g). Therefore, it is a good

candidate as a catalyst support [9]. Silicon carbide has been chosen as a support for steam

Chapter 4

154

reforming by some authors. The known high thermal conductivity of this material may be

interesting in order to improve the temperature profile of the catalyst bed and decrease the

temperature differences that the high endothermicity of this reaction can originate in the

catalyst bed [10]. In addition, many other works have shown that this material has interesting

properties as catalytic support for different reforming reactions [11, 12]. In this work, nickel

catalysts supported over -SiC, have been prepared. This kind of catalysts has shown

acceptable performance for the tri-reforming process in previous chapters. Nevertheless, it is

necessary to improve the catalytic stability, specially its resistance against coke deactivation.

Alkaline and alkaline earth oxides have been extensively studied as traditional promoters

for heterogeneous catalysts, as they are easily accessible and have a low cost. J. Juan-Juan et

al. [13] studied the influence of K load in Ni/Al2O3 catalyst for dry reforming of methane,

reporting that the presence of potassium in Ni/Al2O3 catalysts hinders the accumulation of coke

on the catalyst surface during the dry reforming of methane, but produces a decrease in the

catalytic activity. The addition of Na as promoter in Co/ZnO catalysts for steam reforming was

analyzed by A. Casanovas et al. [14] and was compared with the effect of other metals, seeing

that Na promoted catalysts have a higher activity and selectivity towards reforming products

than the original Co/ZnO catalyst. Alkaline earth metals have also been extensively studied as

catalyst promoters for different reforming reactions. It was reported that MgO-CaO mixed

oxide was an excellent support: carbon deposition was effectively prevented during the

reaction of CO2 with CH4 by supporting Ni on it [15]. The suppression of carbon deposition

was attributed to the basicity of the MgO-CaO mixed oxide. Other authors have suggested that

carbon deposition is suppressed when the metal is supported on a metal oxide with strong

Lewis basicity [16, 17]. Then, the higher the support Lewis basicity, the higher the ability of

the catalyst to chemisorb CO2 [18]. A higher concentration of adsorbed CO2 is suggested to

reduce carbon formation via CO disproportionation (Equation 4.4) by shifting the equilibrium

concentrations. However, Zhang and Verykios reported that the addition of a basic CaO

promoter to Ni/-Al2O3 increased both catalyst stability and carbon deposition in the form of

Ni carbide and/or graphitic carbon, enhancing the reactivity of these species [16]. In addition,

X-ray photoelectron spectroscopy (XPS) results by Tang et al. [19] also illustrated that the

addition of either MgO or CaO to Ni/-Al2O3 greatly increased both catalyst basicity and

carbon deposition during CO2 reforming of CH4. In the present chapter, we report the influence



155

of these traditional promoters over Ni/-SiC catalysts and how these promoters modified its

catalytic performance for the tri-reforming process.

4.2. EXPERIMENTAL


Nickel-supported catalysts were prepared by the impregnation method using nickel nitrate

Ni(NO3)2·6H2O (PANREAC). The support used was SiC, provided by SICAT CATALYST.

Na, K, Mg and Ca, used as promoters, were also added to the catalyst by the impregnation

method. This way a solution of nickel nitrate and the corresponding metal hydroxide, with the

required quantity to obtain a 5 wt% Ni catalyst and the Ni:M (M = Na, K, Mg or Ca) molar

ratio desired, was prepared. In the first part of the study, eight catalysts were prepared with a

Ni:M molar ratio fixed at 10/1 and 2/1 for each metal. For a comparison purpose, a Ni/SiC

catalyst without any promoter was also prepared. In the last part of the study, Mg was selected

as promoter. This way, two catalysts were prepared in order to complete the present study with

a Ni:Mg molar ratio fixed at 4/1 and 1/1. All the catalysts prepared in this work were

dehydrated at 393 K for 12 h and subsequently calcined in air at 1173 K for 2 h.


Ni and Na, K, Mg or Ca metal loading were determined by atomic absorption (AA)

spectrophotometry, using a SPECTRA 220FS analyzer. Samples (ca. 0.5 g) were treated in 2

mL HCl, 3 mL HF and 2 mL H2O2 followed by microwave digestion (523 K). In order to

calculate textural properties (surface area and total pore volume) samples were outgased at 453

K under vacuum for 12 h and analyzed afterwards in a QUADRASORB 3SI sorptometer

apparatus with N2 as the sorbate at 77 K. Temperature-programmed reduction (TPR)

experiments were conducted in a commercial Micromeritics AutoChem 2950 HP unit with

TCD detection. Samples (ca. 0.15 g) were loaded into a U-shaped tube and ramped from room

temperature to 1173 K (10 K min−1

), using a reducing gas mixture of 17.5% v/v H2/Ar

(60 cm3 min

−1). CO2 temperature-programmed desorption (TPD) experiments were also

conducted in the Micromeritics AutoChem 2950 HP unit. The sample (0.15 g) was loaded in a

quartz tube, reduced and pretreated. Then, a flow of 30 mL min-1

of CO2 (99.99% purity,

Praxair certified) was passed through the sample for 30 min at a constant temperature of 313 K.

Finally, the physically adsorbed carbon dioxide was removed by a flow of He for another 30

Chapter 4

156

min. The sample was then heated in 50 mL min-1


up to

1273 K. XRD analyses were conducted with a Philips X’Pert instrument using nickel-filtered

Cu Kα radiation. The samples were scanned at a rate of 0.02° step−1

over the range 5° ≤ 2θ ≤

90° (scan time = 2 s step−1

). Temperature-programmed oxidation (TPO) analyses were

performed in the Micromeritics AutoChem 2950 HP unit, flowing 50 cm3 min

-1 of pure oxygen

from room temperature to 1173 K (10 K min-1

).



internal diameter). The catalyst was placed on a fritted quartz plate located at the end of the

reactor. The reactor was heated with a furnace (Lenton) and the temperature measured with a

K-type thermocouple (Thermocoax). Reaction gases were Praxair certified standards of CH4

(99.995% purity), 10% CO2/N2, O2 (99.99% purity), and N2 (99.999% purity). The water


temperature of the saturator (297 K). The temperature of the saturator was controlled by a

heating bath. All lines placed downstream from the saturator were heated above 373 K to

prevent condensation. The saturation of the feed stream by water at the working temperature

was verified by a blank experiment in which the amount of water trapped by a condenser was

measured for a specific time and compared with the theoretical value. The feed composition

(by volume %) was as follows: 6% CH4, 3% CO2, 3% H2O, 0.6% O2, N2 balance, with a total

flow of 100 NmL min-1

. The catalytic activity was evaluated at 1073 K and atmospheric

pressure for 24 h. Gas effluents were analyzed with a micro gas chromatograph (Varian CP-

4900). Methane and carbon dioxide consumption rates were calculated as follows: [inlet molar

flow of CH4/CO2 – outlet molar flow of CH4/CO2]/nickel weight. A blank experiment carried

out with pure SiC showed no appreciable conversion in the considered conditions.



As commented above, the first part of the study corresponds to an evaluation of four

different cocations (Na, K, Mg and Ca) loaded as promoters in Ni/SiC catalysts for methane

tri-reforming. They were prepared with a M/Ni molar ratio of 1/10 and 1/2, where M represents

the cocation. These catalysts were characterized by atomic absortion spectrophotometry, XRD,



157

TPR, N2 adsorption and CO2-TPD analysis. The main results are listed in Table 4.1. Figure 4.1

shows the diffraction spectrum for the parent SiC used as support (Figure 4.1 a)), and those

obtained for catalysts Ni:Na = 2/1, Ni:K = 2/1, Ni, Ni:Na = 10/1 and Ni:K = 10/1, before and

after reduction (Figures 4.1 b)-f)). The addition of Ni did not significantly alter the SiC

structure. Only minor changes due to the presence of NiO or Ni peaks were observed. Figures

4.1 b), 4.1 c), 4.1 e) and 4.1 f), where the diagrams for the catalysts containing Na and K are

represented, show that these two alkaline metals clearly attack the structure of SiC. For both

promoters, a clear transition from SiC to cristobalite, a high-temperature polymorph of

SiO2, is observed when a high amount of promoter is introduced. Thus, the principal diffraction

peak, and many others, corresponds to one of the -cristobalite reflections (Figures 4.1 b) and

c)). The transient from SiC to -cristobalite is an oxidation process that probably occurs

during the calcination of the catalysts. Na and K have been reported to be poisons for SiC that

make it more easily oxidable. Pareek and co-workers [24] attributed this SiC oxidation rate

enhancement to the capacity of the alkali compounds to get dissolved in the SiO2 and enhance

the O2 transport to the bulk SiC, leading to a drastically increased oxidation. Riley [25] also

reported the effect of alkali environments on SiC, indicating that its oxidation is dramatically

accelerated at high temperatures. However, he observed that SiC is not very affected by alkalis

at temperatures under 1200 K. It is interesting to note that for the catalysts prepared with K as

promoter, the peaks usually assigned to NiO or Ni could not be observed, even though the

presence of this metal was confirmed by atomic absorption (Table 4.1). XRD diagrams showed

that for the Na and K promoted catalysts there are some differences depending on the promoter

load, as the relative intensity of the main -cristobalite peak compared with the intensity of the

principal -SiC peak is clearly higher for high promoter loads. This means that the quantity of

-cristobalite species, and therefore the extension of the -SiC oxidation, is higher for high

promoter loads.

Chapter 4

158

Ta

ble

4.1

. M

ain

phy

sica

l p

roper

ties

of

the

cata

lyst

s.

N

i N

i:N

a

10

/1

Ni:

Na

2/1

Ni:

K

10

/1

Ni:

K

2/1

Ni:

Ca

10

/1

Ni:

Ca

2/1

Ni:

Mg

10

/1

Ni:

Mg

4/1

Ni:

Mg

2/1

Ni:

Mg

1/1

Ni

loa

din

g (

%)

4.5

4

.8

4.9

4

.7

4.5

4

.6

4.6

5

.2

5.1

4

.6

5.5

Su

rfa

ce a

rea

(m

2 g

-1)

24

.0

5.5

0

.9

4.7

1

.5

18

.5

11

.8

22

.5

21

.8

18

.0

21

.2

To

tal

po

re v

olu

me

(cm

3 g

-1)x

10

2

14

.4

3.6

1

.1

2.9

0

.8

12

.9

8.3

1

3.1

1

4.0

1

1.2

1

2.6

Ni

pa

rtic

le d

iam

eter

fro

m X

RD

(n

m)

63

41

85

- -

56

53

39

38

33

32

H2 c

on

sum

pti

on

(m

mo

l g

-1)

0.8

96

0.5

42

0.3

78

0.2

84

0.1

98

0.7

61

0.5

56

0.8

73

0

.62

3

0.6

15

0.5

53

To

tal

ba

sic

site

s (

mo

l g

-1)

5.4

6

- -

- -

5.6

3

2.8

2

11

.05

1

1.8

4

13

.54

14

.39

Dif

fra

ctio

n a

ng

le o

f N

iO (

2

43

.22

- -

- -

- -

43

.14

4

2.9

4

3.0

2

42

.9



159

01

02

03

04

05

06

07

08

09

0

f)e)

d)

c)b

)a)

^

^^

^

^

Intensity (a.u.)

2

(º)

01

02

03

04

05

06

07

08

09

0

º*

º

ºº+

+

º

Intensity (a.u.)

2

(º)

^^

^

ºº

ºº º

ºº

º

^^

^

ºº

ºº º

ºº

ºF

resh

Red

uce

d

01

02

03

04

05

06

07

08

09

0

º

^^

^

Intensity (a.u.)

2

(º)

ºº

ººº

ºº

º

º

^^

^

ºº

ººº

ºº

ºF

resh

Red

uce

d

01

02

03

04

05

06

07

08

09

0

Fre

sh

**

*

*++

+

^

^

Intensity (a.u.)

2

(º)

^^

^

^

^^

^

^

^

Red

uce

d

01

02

03

04

05

06

07

08

09

0

ºº

**

º+

+

º^

^

^

^

^

^^

Intensity (a.u.)

2

(º)

^

Fre

sh

Red

uce

d

01

02

03

04

05

06

07

08

09

0

ºº

^^

^

^

^ ^

Intensity (a.u.)

2

(º)

Fre

sh

Red

uce

d

**

Fig

ure

4.1

. X

RD

pro

file

s o

f a)

Cat

aly

st s

up

po

rt,

b)

Ni:

Na

= 2

/1,

c) N

i:K

= 2

/1,

d)

Ni/

SiC

, e)

Ni:

Na

= 1

0/1

, f)

Ni:

K =

10

/1,

wh

ere

(^)

den

ote

s re

flec

tio

n o

f

SiC

, (+

) d

eno

tes

refl

ecti

on

of

met

alli

c n

ick

el,

(*)

den

ote

s re

flec

tion

of

nic

kel

ox

ide

and (

º) d

eno

tes

refl

ecti

on

of

-

cris

tob

alit

e.

Chapter 4

160

Figure 4.2 shows the XRD results for the catalysts prepared using Mg or Ca as promoters.

For both catalysts, Ni:Mg = 10/1 (Figure 4.2 c)) and Ni:Mg = 2/1 (Figure 4.2 a)), only the

peaks attributed to SiC, NiO or Ni could be clearly assigned. For the catalyst Ni:Ca = 2/1,

there are three peaks that cannot be assigned to the previously commented species. This fact

could indicate the most probable presence of quartz in the sample, caused by an increase in the

oxidation rate of SiC during the calcination step. This increase could be due to the presence

of a relatively high content of Ca in the catalyst. This element has been pointed to decrease the

stability against oxidation of SiC, despite it is less deleterious than alkali salts [26]. Taking into

account the calcination temperature (1173 K), SiC should be oxidized into amorphous SiO2 or

quartz [27]. However, this was not observed for Na and K promoted catalysts. They showed -

cristobalite as the main SiO2 phase, which should be obtained at higher calcination

temperatures [27]. However, Na [28] and K [24] induce crystallization of the amorphous silica

into -cristobalite at temperatures far below the normal transition temperature. In addition, the

acid SiO2, which forms the passive film that protects SiC from oxidation, will react in a higher

extension with the more basic oxides [29], leading to a higher extension of the SiC oxidation.



161

Figure 4.2. XRD profiles a) Ni:Mg = 2/1, b) Ni:Ca = 2/1, c) Ni:Mg = 10/1, d) Ni:Ca = 10/1,

where (^) denotes reflection of SiC, (+) denotes reflection of metallic nickel, (*) denotes

reflection of nickel oxide and (-) denotes reflection of quartz.

In order to analyze the influence of the promoters on the metal support interactions and

reducibility of the catalysts, TPR experiments were carried out from room temperature to 1173

K. The corresponding data, represented in Figure 4.3, showed several differences between the

reference catalyst (Ni/SiC) and those ones that incorporated any promoter. Metal support

interaction always increased after adding the promoter, shifting the reduction peaks toward

higher temperatures. Among the samples prepared with Na or K as cocations (Figure 4.3 a)),

only catalyst Ni:Na = 10/1 showed a clear reduction peak. The profiles for the other samples

were broad and small. The H2 consumption during these TPR experiments (Table 4.1) also

showed that Ni was hardly reduced. Regarding Ca and Mg promoted catalysts (Figure 4.3 b)),

sample Ni:Ca = 2/1 gave a similar response as those catalysts loaded with Na or K, being the

H2 consumption also too low. It can be noted a relation between the SiC oxidation reported

from the XRD experiments and the low extent of Ni reduction. Thus, the oxidation process

undergone by SiC and the consequent crystallization of SiO2 made a big part of the Ni to be

inaccessible by H2. The TPR profile for sample Ni:Ca = 10/1 showed just one relatively sharp

0 10 20 30 40 50 60 70 80 90

Fresh*

***

+

Inte

nsi

ty (

a.u

.)

2(º)

^

^^

^

^

^^

^

^

^

++

a)

Reduced

0 10 20 30 40 50 60 70 80 90

b)

-

--

-

-

***

+

Inte

nsi

ty (

a.u

.)

2(º)

^

^^

^

^

^^^^

^

+

+

Fresh

Reduced

0 10 20 30 40 50 60 70 80 90

c)

**

**

+++

Inte

nsi

ty (

a.u

.)

2(º)

^

^^

^

^

^^

^

^

^

Fresh

Reduced

0 10 20 30 40 50 60 70 80 90

d)

**

**

+

+

Inte

nsi

ty (

a.u

.)

2(º)

^

^

^

^

^

^^^

^

^

Fresh

Reduced

*

Chapter 4

162

reduction peak around 950 K. Comparing this profile with that of Ni/SiC, it seems that the

presence of a low amount of Ca increased the metal support interaction. For the reference

catalyst, it was obtained a profile with two overlapped peaks with maxima around 720 and 870

K, followed by a small peak obtained at 1140 K. The peak at low temperature is usually

attributed to the reduction of bulk NiO, while peaks at higher temperatures are attributed to a

higher interaction between nickel and support, originating nickel silicate [30, 31]. It can be

observed that the addition of low quantities of Mg increased the Ni dispersion and decreases

the H2 consumption during the TPR experiments (Table 4.1). The presence of Mg in catalysts

Ni:Mg = 10/1 and Ni:Mg = 2/1 led to a shift of the reduction peaks toward higher

temperatures, which was influenced by the Mg loading. The TPR corresponding to the catalyst

with the lowest quantity of Mg, Ni:Mg = 10/1, showed two main reduction peaks at 750 and

980 K. The first peak could be related to the reduction of bulk NiO, as discussed for catalyst

Ni/SiC, while the second one seems to correspond to the reduction of a NiO-MgO solid

solution, as high calcination temperature usually leads towards the formation of this phase in

catalysts where Ni and Mg are present, requiring higher reduction temperatures due to the

strong interaction between NiO and MgO [32, 33]. Comparing this catalyst with the reference,

it is clear that the addition of this low amount of Mg increased the quantity of species hardly

reducible and decreased the total amount of bulk NiO. Similarly, TPR profile for catalyst

Ni:Mg = 2/1 showed that a higher quantity of Mg almost caused the peak assigned to the

reduction of bulk NiO to disappear and led to an increase of the size of the peak assigned to the

reduction of the NiO-MgO solid solution, which implied a decrease in the reducibility of NiO.



163

Figure 4.3. TPR profiles: a) Reference, Na and K promoted catalysts, b) Mg and Ca promoted

catalysts.

Nickel particle size, surface area and basicity, in terms of CO2 desorbed in a CO2-TPD

experiment, are also listed in Table 4.1. Ni particle size was obtained with the Debye–Scherrer

equation using the data from the XRD patterns (111 reflection of Ni0). This is not a very

accurate method in order to obtain the actual metal particle size, but it could be applied in order

to analyze different catalysts and get a relative comparison of the metal particle size. As a

general trend, it could be observed a decrease in the Ni metal particle size after the promoter

addition. However, it can be seen that the catalyst prepared with the highest quantity of Na

shows the bigger particle size. It is likely related to the previously commented change in the

support structure due to the enhance oxidation of SiC during the calcination. It leads to the

formation of -cristobalite, which is a compound with a high crystallinity and a very low

surface area, facilitating the sintering of Ni particles during the calcination. It can be noted that

presence of Ca did not greatly affect to Ni particle size. On the contrary, Mg promoted

300 400 500 600 700 800 900 1000 1100 1200

a)

Ni

Ni:Mg 2/1

Ni:Mg 10/1

Ni:Ca 2/1

Ni:K 2/1

Ni:K 10/1

Ni:Na 2/1

Ni:Na 10/1

TC

D s

ign

al (

a.u

.)

Temperature (K)

Ni:Ca 10/1

300 400 500 600 700 800 900 1000 1100 1200

b)T

CD

sig

nal

(a.

u.)

Temperature (K)

Chapter 4

164

catalysts presented a lower Ni particle size than that of catalyst Ni/SiC. As it can be

observed in Table 4.1, surface area and total pore volume values were in agreement with the

XRD diagrams, being the surface area of the Na, K and Ca promoted catalysts much lower

than that of the Mg promoted and reference catalysts, as a consequence of the increased

oxidation rate of the former catalysts during the calcination and the changes in the support

structure. The presence of Mg as promoter has a clear effect on the catalyst basicity, even for

the catalyst with the lower Mg loading, increasing the basicity of the catalyst with the Mg

content. However, Ca promoted catalyst did not show the same trend, having the catalyst with

the highest quantity of Ca the lowest basicity, which seems to be related to the formation of

quartz observed for this catalyst.


Taking into account the catalyst characterization results above discussed, only catalysts

Ni:Ca = 10/1, Ni:Mg = 10/1 and Ni:Mg = 2/1 were considered to be tested for the methane tri-

reforming process, being their catalytic performance compared to that of sample Ni/SiC.

Catalytic results are plotted in Figure 4.4. It could be seen that during the 24 h experiment the

methane reaction rate of catalyst Ni/SiC (Figure 4.4 a)) drops from 8.8x10-4

to 7.5x10-4

mol

s-1

gNi-1

(14.8% less after 24 h). Activity loss for sample Ni:Ca = 10/1 (Figure 4.4 b)) was lower

(from 7.9x10-4

to 7.3x10-4

mol s-1

gNi-1

, that is 7.6% less), being its methane reaction rate quite

close to that of the reference catalyst. However, its carbon dioxide reaction rate was clearly

lower, reaching values near to 0, thus yielding a H2/CO molar ratio ranging from 2.5 to 2.9

(close to the stoichiometric value of the synthesis gas produced by the steam reforming

reaction). Regarding catalysts Ni:Mg = 10/1 (Figure 4.4 c)) and Ni:Mg = 2/1 (Figure 4.4 d)),

methane reaction rate was higher than that of the reference catalyst. The activity drop after 24 h

was also clearly lower (2.6% and 1.1%, respectively). Hence, it could be drawn that the

addition of Ca or Mg decreased, or at least did not increase, the carbon dioxide reaction rate.

This is something unexpected as both promoters increased catalyst basicity, which has been

often reported as a positive factor in dry reforming due to the higher CO2 adsorption capacity

of the catalyst [18]. Nevertheless, despite this general rule, it can be seen in Table 4.1 that the

low quantity of Ca in catalyst Ni:Ca = 10/1 did not increase its basicity (compared to

Ni/SiC).



165

Figure 4.4. Catalytic activity at 1073 K for: a) Ni/SiC, b) Ni:Ca = 10/1, c) Ni:Mg = 10/1, d)

Ni:Mg = 2/1. Reaction conditions: CH4 = 6%, CO2 = 3%, H2O = 3%, O2 = 0.6%, N2 balance,

total flow rate = 100 NmL min-1

. CH4 ( ) and CO2 ( ) consumption rates vs. time on stream

(left axis), and H2/CO molar ratio ( ) vs. time on stream (right axis).

On the other hand, catalysts containing Ca have been reported to shift the water gas shift

equilibrium towards the formation of CO2 at a temperature close to that selected in this work

[34, 35]. This process could be responsible of the carbon dioxide conversion decrease.

Catalysts prepared with Mg as the promoter showed a higher CO2 adsorption capacity than that

of the reference catalyst (Table 4.1). Thus, it can be concluded that the presence of Mg

increased the basicity of the catalyst. Anyway, as previously commented, catalysts Ni:Mg =

10/1 and Ni:Mg = 2/1 showed a similar CO2 reaction rate and a lower CO2 reaction rate,

respectively, if compared as that of the reference sample, despite having a higher basicity. A

higher basicity implies a higher CO2 adsorption capacity of the catalyst, which usually is

correlated to a higher reactivity of this molecule. However, these catalysts did not follow this

trend. This behaviour will be deeply explained in the next section.

0 5 10 15 20 25

0

1

2

3

4

5

6

7

8

9

10

a)

Consu

mpti

on r

ate

(mol

s-1 g

-1 Ni )

·10

4

Time (h)

0 5 10 15 20 25

0

1

2

3

4

5

6

7

8

9

10b)

Consu

mpti

on r

ate

(mol

s-1 g

-1 Ni )

·10

4

Time (h)

0 5 10 15 20 25

0

1

2

3

4

5

6

7

8

9

10

c)

Consu

mpti

on r

ate

(mol

s-1 g

-1 Ni )

·10

4

Time (h)

0 5 10 15 20 25

0

1

2

3

4

5

6

7

8

9

10

d)

Consu

mpti

on r

ate

(mol

s-1 g

-1 Ni )

·10

4

Time (h)

0.5

1.0

1.5

2.0

2.5

3.0

3.5

H2/C

O M

ola

r ra

tio

1.5

2.0

2.5

3.0

3.5

4.0

4.5

H2/C

O M

ola

r ra

tio

0.5

1.0

1.5

2.0

2.5

3.0

3.5

H2/C

O M

ola

r ra

tio

0.5

1.0

1.5

2.0

2.5

3.0

3.5

H2/C

O M

ola

r ra

tio

Chapter 4

166

4.3.3. Influence of the Mg/Ni molar ratio.

In order to better understand the influence of the Ni:Mg molar ratio on the catalytic

performance in the tri-reforming process, other two catalysts were prepared with Ni:Mg molar

ratios of 4/1 and 1/1. The same characterization techniques, previously commented, were

applied to these catalysts. Figure 4.5 shows the reduction profile for all the samples with Mg as

promoter. The addition of Mg increased the temperature of the reduction peaks, which could be

related to the occurrence of a higher interaction between NiO and MgO. A reduction profile

evolution was observed as a function of Mg content. Thus, catalyst Ni/SiC showed at least

two peaks overlapped with maxima around 720 and 870 K, followed by a small peak obtained

at 1140 K, whereas catalyst Ni:Mg = 1/1, with the highest Mg content, showed a reduction

peak with maximum at 1020 K. The higher the Mg content in the catalyst, the lower the

catalyst reducibility was obtained. The presence of the peak at a high temperature would

indicate that Ni and Mg were in the form of a NiO-MgO solid solution [32, 33]. Bradord et al.

[36] attributed the formation of this solid solution to the fact that Ni and Mg almost perfectly

fit the Hume-Rothery criteria for the formation of an extensive solid solution, i.e., both cations

have similar ionic radii, ca. 0.78 Å [37], the same common oxidation state (2+), and the same

bulk oxide structure, NaCl-type [38]. The formation of this NiO-MgO solid solution was also

confirmed with the XRD experiments. The diffraction angle of NiO for the different Mg

promoted catalysts is given in Table 4.1, showing a light decrease in this value with the Mg

loading. This decrease in the diffraction angle of NiO in catalysts where there is Mg is usually

attributed to the formation of a NiO-MgO solid solution [39, 40].

Figure 4.5. TPR profiles for Mg promoted and Ni/SiC catalysts.

300 400 500 600 700 800 900 1000 1100 1200

Ni

Ni:Mg 10/1

Ni:Mg 4/1

Ni:Mg 2/1

Ni:Mg 1/1

TC

D s

ignal

(a.

u.)

Temperature (K)



167

Particle size, surface area, total pore volume and catalysts basicity were also measured for

these samples (Table 4.1). As previously commented, Ni:Mg catalysts showed a lower Ni

particle size (compared with the reference sample). The results indicate that the higher the Mg

loading, the lower the Ni particle size was obtained. Basicity of the catalysts also increased

after the addition of Mg. The following order was established: Ni/SiC << Ni:Mg = 10/1

Ni:Mg = 4/1 < Ni:Mg = 2/1 < Ni:Mg = 1/1.

Figure 4.6. Catalytic activity at 1073 K for: a) Ni:Mg = 4/1, b) Ni:Mg = 1/1. Reaction

conditions: CH4 = 6%, CO2 = 3%, H2O = 3%, O2 = 0.6%, N2 balance, total flow rate = 100 mL

min-1

. CH4 ( ) and CO2 ( ) consumption rates vs. time on stream (left axis), and H2/CO molar

ratio ( ) vs. time on stream (right axis).

The influence of the Mg loading on the catalytic performance of these catalysts was also

studied. Figure 4.6 shows for catalysts Ni:Mg = 4/1 and Ni:Mg = 1/1 the evolution of CH4

reaction rate, CO2 reaction rate and H2/CO molar ratio during the 24 h tri-reforming tests.

H2/CO molar ratio is a key parameter in order to evaluate the catalytic performance in the tri-

0 5 10 15 20 25

0

1

2

3

4

5

6

7

8

9

10b)

Co

nsu

mp

tio

n r

ate

(mo

l s-1

g-1 N

i )·1

04

Time (h)

0 5 10 15 20 25

0

1

2

3

4

5

6

7

8

9

10a)

Co

nsu

mp

tio

n r

ate

(mo

l s-1

g-1 N

i )·1

04

Time (h)

0.5

1.0

1.5

2.0

2.5

3.0

3.5

H

2/C

O M

ola

r ra

tio

0.5

1.0

1.5

2.0

2.5

3.0

3.5

H

2/C

O M

ola

r ra

tio

Chapter 4

168

reforming process, as not only will determine the possible applications of the synthesis gas

obtained but also indicates the relative importance of each reaction in the tri-reforming.

Thereby, when the dry reforming activity is higher, the H2/CO molar ratio obtained from global

tri-reforming will be lower due to the stoichiometry of the dry reforming reaction (Equation

4.2). Catalytic deactivation in these catalysts was also lower than observed for the reference

one, with a drop in the CH4 reaction rate after 24 h of 2.3% for Ni:Mg = 4/1 and 0.9% for

Ni:Mg = 1/1. Similarly, the CO2 reaction rate was lower and the H2/CO molar ratio higher in

the Mg loaded catalysts than those corresponding to catalyst Ni/SiC. A comparison between

the most representative reaction parameters for the different Mg promoted catalysts and the

reference one is seen in Table 4.2. Mg promoted catalysts led to values of methane reaction

rate and stability higher than those corresponding to catalyst Ni/SiC, indicating that this

cocation had a beneficial effect over the methane tri-reforming. The best performance was

found for samples Ni:Mg = 2/1 and Ni:Mg = 1/1, which showed very close catalytic results,

with the highest CH4 rate and a very high stability. It is also remarkable to note that the

addition of Mg increased the H2/CO molar ratio, which could be related to the high strength of

the catalyst basic sites. Thus, L. Pino et al. [41] reported for Ni–La–CeO2 catalysts an increase

in the H2/CO molar ratio, which was related to the higher concentration of strong basic sites

with increasing La loads. The basic sites strengthening in our catalysts can be clearly seen in

Figure 4.7. The Mg addition led to an increase of the CO2 desorption peak at 1075 K, which

corresponded to the presence of very strong basic sites. Moreover, as above mentioned, an

increase in the Mg loading made catalyst reduction to be more difficult. Hence, the higher the

Mg loading, the higher the amount of NiO species present on the catalytic surface. In addition,

NiO could act as a promoter of the water gas shift reaction. It has been reported that this oxide

enhances the formation of surface oxygen intermediates such as Ni(OH)2 and NiOOH [42],

which would in turn lead to a lower CO2 reaction rate.



169


Ni Ni:Mg 10/1 Ni:Mg 4/1 Ni:Mg 2/1 Ni:Mg 1/1

Average CH4 reaction rate

(mol s-1

gNi-1

) ×104

7.96 8.92 8.46 8.95 8.65

Average CH4 conversion (%) 73.7 95.4 88.8 84.7 97.9

Drop in CH4 reaction rate (%) 14.8 2.6 2.3 1.1 0.9

Average CO2 reaction rate

(mol s-1

gNi-1

) ×104

2.72

2.72 2.38 1.83 1.84

Average CO2 conversion (%) 52.3 60.4 51.9 36.0 43.2

Drop in CO2 reaction rate (%) 23.75 14.6 15.1 24.2 24.4

Average H2/CO molar ratio 2.00 2.09 2.23 2.36 2.76

Oxygen consumption in TPO

(mmol g-1

) 35.28 10.70 8.23 5.71 6.24

Particle diameter from XRD after

reaction (nm) 54 42 38 39 34

Figure 4.7. CO2-TPD profiles for Mg promoted and Ni/SiC catalysts.

Catalysts deactivation was evaluated in terms of the oxygen consumption of the coke

generated after tri-reforming reaction tests in a TPO experiment. Figure 4.8 shows the TPO

400 600 800 1000 1200

Ni

Ni:Mg 10/1

Ni:Mg 4/1

Ni:Mg 2/1

Ni:Mg 1/1

TC

D s

ign

al (

a.u

.)

Temperature (K)

Chapter 4

170

profiles obtained. Table 4.2 lists the values of oxygen consumption in the TPO runs. It can be

observed that the total amount of coke deposited onto the catalysts decreased with increasing

values of the Mg loading, probably due to the decrease in metal particle size and the increase in

the interaction between Ni and Mg in the resulting catalysts. The lower coke formation was

found for catalysts Ni:Mg = 2/1 and Ni:Mg = 1/1. These two catalysts contained a similar

quantity of coke, which would indicate that the addition of Mg decreased the coke generation

rate until a given value of Ni:Mg molar ratio. Two peaks with maxima around 900 and 1000 K,

which corresponds to the occurrence of two coke species, are observed in Figure 4.8. The

oxidation temperature of these peaks matches with that reported by Zhang et al. [43] for Cand

C coke species. The former would be related with the generation of CO at high reaction

temperatures and the later would be responsible of the catalyst deactivation. Our results show

that the addition of Mg decreased the quantity of both coke species, leading to an increase of

the catalyst stability. In addition, the Ni particle size in the catalysts used, measured by XRD

analysis (Table 4.2), did not change in a meaningful way, concluding that no Ni sintering

occurred during the reaction.

Figure 4.8. TPO profiles after reaction for Mg promoted and Ni/SiC catalysts.

300 400 500 600 700 800 900 1000 1100 1200

Ni

Ni:Mg 10/1

Ni:Mg 4/1

Ni:Mg 2/1

Ni:Mg 1/1

TC

D s

ign

al (

a.u

.)

Temperature (K)



171

4.4. CONCLUSIONS

Alkaline and alkaline earth metals have been studied as cocations in Ni/SiC catalysts

for methane tri-reforming. It was found that Na and K, despite their good properties as catalytic

promoters in other reforming processes, were not useful for this process. It was due to their

ability to increase the SiC oxidation rate during the calcination step, generating -

cristobalite and decreasing the surface area of the support. High loads of Ca also affected the

final oxidation rate of the SiC catalyst.

Catalysts modified with Mg or with low load of Ca were tested in tri-reforming

experiments. The presence of Mg enhanced both activity and stability of the catalyst,

decreasing Ni metal particle size and increasing its basicity. It was observed that an increase in

Mg loading favoured the formation of Ni metal particles smaller in diameter and decreased the

reducibility of Ni, shifting the reduction peaks towards higher temperatures, likely due to the

formation of a NiO-MgO solid solution. Catalytic activity was also analyzed in terms of Mg

loading. A higher H2/CO molar ratio and a better stability were observed for those catalysts

with a higher amount in Mg. The stronger the basic sites in the catalyst, the higher the H2/CO

molar ratio was. Finally, the higher the interaction between Ni and Mg, the slower the catalyst

deactivation rate was observed.

Chapter 4

172

4.5 REFERENCES

[1] E. Iglesia, Applied Catalysis A: General. 161 (1997) 59-78.

[2] M.E. Dry, Catal. Today. 71 (2002) 227-241.


137.


[5] A.P.E. York, T. Xiao, M.L.H. Green, Top. Catal. 22 (2003) 345-358.

[6] M. Akiyama, Y. Oki, M. Nagai, Catal. Today. 181 (2012) 4-13.

[7] Y.J.O. Asencios, C.B. Rodella, E.M. Assaf, Applied Catalysis B: Environmental. 132-133

(2013) 1-12.

[8] H.t. Jiang, H.q. Li, Y. Zhang, Ranliao Huaxue Xuebao/Journal of Fuel Chemistry and

Technology. 35 (2007) 72-78.


[10] F. Basile, P.D. Gallo, G. Fornasaria, D. Gary, V. Rosetti, A. Vaccari, in: F.B. Noronha, M.

Schmal, E.F. Sousa-Aguiar (Eds.), 2007, pp. 313-318.


51-54.

[12] E. Kockrick, L. Borchardt, C. Schrage, C. Gaudillere, C. Ziegler, T. Freudenberg, D.

Farrusseng, A. Eychmüller, S. Kaskel, Chem. Mater. 23 (2011) 57-66.

[13] J. Juan-Juan, M.C. Román-Martínez, M.J. Illán-Gómez, Applied Catalysis A: General.

301 (2006) 9-15.

[14] A. Casanovas, M. Roig, C. de Leitenburg, A. Trovarelli, J. Llorca, Int. J. Hydrogen

Energy. 35 (2010) 7690-7698.

[15] O. Yamazaki, T. Nozaki, K. Omata, K. Fujimoto, Chem. Lett. 21 (1992) 1953-1954.



173

[16] Z.L. Zhang, X.E. Verykios, Catal. Today. 21 (1994) 589-595.

[17] T. Horiuchi, K. Sakuma, T. Fukui, Y. Kubo, T. Osaki, T. Mori, Applied Catalysis A:

General. 144 (1996) 111-120.

[18] G.J. Kim, D.S. Cho, K.H. Kim, J.H. Kim, Catal. Lett. 28 (1994) 41-52.

[19] S.B. Tang, F.L. Qiu, S.J. Lu, Catal. Today. 24 (1995) 253-255.





[23] J.M. García-Vargas, J.L. Valverde, A. de Lucas-Consuegra, B. Gómez-Monedero, F.

Dorado, P. Sánchez, Int. J. Hydrogen Energy. 38 (2013) 4524-4532.

[24] V. Pareek, D.A. Shores, J. Am. Ceram. Soc. 74 (1991) 556-563.

[25] F.L. Riley, 1996, pp. 1-14.

[26] K.N. Lee, R.A. Miller, Surf. Coat. Technol. 86–87, Part 1 (1996) 142-148.

[27] M.J.-F. Guinel, M.G. Norton, J. Mater. Res. 21 (2006) 2550-2563.

[28] A. Palermo, J.P.H. Vazquez, A.F. Lee, M.S. Tikhov, R.M. Lambert, J. Catal. 177 (1998)

259-266.

[29] N.S. Jacobson, J. Am. Ceram. Soc. 76 (1993) 3-28.


159.



Chapter 4

174


(1990) 2663-2669.

[34] A.H. Clemens, L.F. Damiano, T.W. Matheson, Fuel. 77 (1998) 1017-1020.

[35] Z. Zhang, M. Baerns, Applied Catalysis. 75 (1991) 299-310.

[36] M.C.J. Bradford, M.A. Vannice, Applied Catalysis A: General. 142 (1996) 73-96.

[37] R.A.T.P.K. Flinn, Engineering materials and their applications, Wiley, New York;

Chichester, 1995.

[38] V.E.C.P.A. Henrich, The surface science of metal oxides, Cambridge University Press,

Cambridge; New York, 1994.

[39] Y.-H. Wang, H.-M. Liu, B.-Q. Xu, J. Mol. Catal. A: Chem. 299 (2009) 44-52.

[40] F. Arena, F. Frusteri, A. Parmaliana, L. Plyasova, A.N. Shmakov, J. Chem. Soc., Faraday

Trans. 92 (1996) 469-471.


104 (2011) 64-73.


(2008) 97-104.


CHAPTER 5 Preparation of Ni-Mg/-SiC catalysts for the methane tri-reforming: effect of the order of

metal impregnation

Resumen

Abstract

5.1. INTRODUCTION

5.2. EXPERIMENTAL







5.3.3. Characterization after reaction

5.4. CONCLUSIONS

5.5. REFERENCES

Preparation of Ni-Mg/-SiC catalysts for the methane tri-reforming: effect of the order of

metal impregnation

177

Resumen

En el presente capítulo se analizó la influencia que tiene el orden en el que se

impregnan el Ni y el Mg sobre la actividad catalítica y la estabilidad de catalizadores

soportados sobre -SiC, probados en el proceso de tri-reformado. Los catalizadores

fueron caracterizados usando técnicas como la reducción a temperatura programada,

difracción de rayos X, microscopía de transmisión electrónica y oxidación a

temperatura programada. La adición de Mg produjo un cambio notable en los perfiles

de reducción, aumentando la temperatura necesaria para obtener Ni0. La reducción

ocurrió a temperaturas mayores cuando se impregnó en primer lugar Mg o ambos

metales fueron impregnados simultáneamente, lo que se atribuyó a una fuerte

interacción entre el Ni y el Mg en estos catalizadores. Los catalizadores preparados

impregnando Ni en primer lugar mostraron los peores resultados catalíticos,

probablemente debido a una escasa interacción entre Ni y Mg, el posible bloqueo de

las partículas de Ni por el Mg y la aparición de Ni2Si después de reacción. Los

catalizadores preparados con la mayor razón molar Mg/Ni (1/1) presentaron menor

tamaño de partícula de Ni, menor velocidad de formación de coque, mayor basicidad y

mayor interacción Ni-Mg. El catalizador Ni-Mg/SiC 1/1 fue seleccionado como el

mejor debido a su gran actividad catalítica, buena estabilidad y baja generación de

coque.

Chapter 5

178

Abstract

The influence of the order of Ni and Mg impregnation has been analyzed in terms

of catalytic activity and stability of -SiC supported catalysts for the tri-reforming of

methane. Catalysts were characterized using different techniques such as Temperature

Programmed Reduction, X-Ray Diffraction, Transmission Electron Microscopy and

Temperature Programmed Oxidation. The addition of Mg clearly changed the

reduction profile, increasing the temperature required to obtain Ni0. Higher reduction

temperatures were needed when Mg was firstly loaded or when both metals, Ni and

Mg, were simultaneously loaded, which was attributed to the occurrence of

interactions between Ni and Mg. Catalyst prepared by first Ni impregnation showed

the worst catalytic behaviours, probably due to a poor interaction between Ni and Mg,

a possible blockage of Ni particles by Mg ones and the occurrence of Ni2Si after

reaction. Catalysts prepared with the highest Mg/Ni molar ratio (1/1) showed smaller

Ni particle sizes, lower coke rate formation and higher basicity and Ni-Mg interaction.

Ni-Mg/SiC 1/1 was selected as the best catalyst due to its high catalytic activity and

stability and low coke generation.


metal impregnation

179

5.1. INTRODUCTION

Nowadays, global warming is one of the most important environmental problems, usually

related to the emission of green house gases like CH4 and CO2, and the raise observed in the

atmospheric CO2 concentration during the last century, which is assigned to human activities

and the burning of fossil fuels. Therefore, CO2 conversion and utilization are important

elements in chemical research on sustainable development, but its recovery from concentrated

sources requires substantial energy input [1], which makes interesting its conversion without

previous separation.

Tri-reforming is an interesting process from an environmental point of view as it can help

to reduce emissions of two green house effect gases like CH4 and CO2. In addition, this process

enables the production of synthesis gas from a renewable source like biogas. Biogas is a clean

and environmentally friendly fuel that is typically generated from anaerobic degradation of

biomass. Biogas, consisting mainly of CO2 and CH4, is an attractive renewable carbon source

and its exploitation would be advantageous from both financial and environmental points of

view [2]. Synthesis gas is a fundamental feedstock for refining processes, such as hydrotreating

and hydrocracking, for petrochemical processes, such as the synthesis of methanol, methanol to

gasoline, and the synthesis of ammonia [3] and for the hydrocarbon synthesis, via Fischer–

Tropsch processes [4].

Tri-reforming consists of a synergetic combination of steam reforming (Equation 5.1), dry

reforming (Equation 5.2) and partial oxidation (Equation 5.3) of methane.

H2O + CH4 → CO + 3H2 (H◦ = 206.3 kJmol

-1) (Equation 5.1)

CO2 + CH4 → 2CO + 2H2 (H◦ = 247.3 kJmol

-1) (Equation 5.2)

CH4 + 1/2O2 → CO + 2H2 (H◦ = −35.6 kJmol

-1) (Equation 5.3)

The main advantages of this process, compared to dry and steam reforming, are the less

endothermic nature of the global process (due to the presence of the exothermic methane

partial oxidation reaction), the low quantity of coke generated (Equations 5.4 and 5.5) due to

the presence of oxidants like H2O and O2 (Equations 5.6 and 5.7), and the possibility to modify

the H2/CO molar ratio by shifting the reactants ratio.

Chapter 5

180

2CO C + O2 (H◦ = −172.2 kJmol

-1) (Equation 5.4)

CH4 C + 2H2 (H◦ = 74.9 kJmol

-1) (Equation 5.5)

C + H2O CO + H2 (H◦ = 131.4 kJmol

-1) (Equation 5.6)

C + O2 CO2 (H◦ = −393.7 kJmol

-1) (Equation 5.7)

Due to the high price and non-availability of noble metals like Pt, Rh, and Ru, transition

metals have been selected as the active phase for several catalytic processes. Nickel has proven

to be the most appropriate transition metal for reforming processes [5]. Moreover, nickel has

been extensively studied in different reforming processes, including steam reforming [6], dry

reforming [7] and partial oxidation [8].

In this work a novel support, Silicon carbide (-SiC), with many interesting

characteristics, was used. -SiCexhibits a high thermal conductivity, a high resistance towards

oxidation, a high mechanical strength, chemical inertness and average surface area (around 25

m2/g) [9]. Although -SiC-based catalysts has shown acceptable performance for methane tri-

reforming [10, 11], an improvement of their catalytic stability and specially their resistance

against coke deactivation should be considered before considering them as potential catalysts

for this process. In the previous chapter, Mg was chosen as the best promoter for Ni/SiC

catalysts used in the tri-reforming process of methane. The presence of Mg enhanced both the

activity and stability of the catalyst, leading to a decrease of the Ni metal particle size and an

increase of its basicity.

In the present chapter, the influence of the order of Ni and Mg impregnation on the

catalytic performance in the methane tri-reforming process of Ni-Mg/-SiC catalysts was

studied.

5.2. EXPERIMENTAL


Catalysts were prepared by the wet impregnation method, using β-SiC (SICAT

CATALYST) as support and nickel nitrate Ni(NO3)2·6H2O (PANREAC) and magnesium

hydroxide Mg(OH)2 as precursors, adding the required quantity to an aqueous solution in order

to obtain catalysts with a Ni load of 5 wt%. Catalysts were prepared with two different Mg/Ni


metal impregnation

181

molar ratios, 1/10 and 1/1. For each molar ratio, three different sets of catalysts were prepared.

The first one was prepared by first Ni impregnation, followed by a calcination step (2 h at 1173

K) and subsequent Mg impregnation (samples Mg/Ni/SiC). The second one was prepared by

first Mg impregnation, followed by a calcination step (2 h at 1173 K) and subsequent Ni

impregnation (samples Ni/Mg/SiC). The third one was prepared by Ni and Mg co-

impregnation (samples Ni-Mg/SiC). In addition, a reference catalyst prepared by just Ni

impregnation was used in order to check the influence of the promoter on the performance of

the methane tri-reforming process. All the catalysts were dried at 393 K overnight and calcined

in air at 1173 K for 2 h.


Ni and Mg metal loadings were determined by atomic absorption (AA)

spectrophotometry, using a SPECTRA 220FS analyzer. Samples (ca. 0.5 g) were treated in 2

mL HCl, 3 mL HF and 2 mL H2O2 followed by microwave digestion (523 K). Surface

area/porosity measurements were conducted using a QUADRASORB 3SI sorptometer

apparatus with N2 as the sorbate at 77 K. The samples were outgased at 453 K under vacuum

(5×10-3

torr) for 12 h prior to the analysis. Specific surface areas were determined by the multi

point BET method. Specific total pore volume was evaluated from N2 uptake at a relative

pressure of P/Po = 0.99. Temperature-programmed reduction (TPR) experiments were

conducted in a commercial Micromeritics AutoChem 2950 HP unit with TCD detection.

Samples (ca. 0.15 g) were loaded into a U-shaped tube and ramped from room temperature to

1173 K (10 K min−1

), using a reducing gas mixture of 17.5% v/v H2/Ar (60 cm3 min

−1). CO2

temperature-programmed desorption (TPD) experiments were conducted in a commercial

Micromeritics AutoChem 2950 HP unit with TCD detection. 0.15 g of sample were loaded in a

quartz tube, reduced and pretreated in He. After cooling, 30 mL min-1

of CO2 (99.99% purity,

Praxair certified) was passed through the sample for 30 min at a constant temperature of 323 K.

Finally, the gaseous and weakly adsorbed carbon dioxide was removed by a steady flow of He

for another 30 min. The sample was then heated in 50 mL min-1

of He with a heating rate of 10

K min-1

up to 1173 K. XRD analyses were conducted with a Philips X’Pert instrument using

nickel-filtered Cu Kα radiation. The samples were scanned at a rate of 0.02° step−1

over the

range 5° ≤ 2θ ≤ 90° (scan time = 2 s step−1

). Temperature-programmed oxidation (TPO)

analyses were performed in a Micromeritics AutoChem 2950 HP unit, flowing 50 cm3 min

-1 of

pure oxygen from room temperature to 1173 K (10 K min-1

). Transmission electron

Chapter 5

182

microscopy (TEM) analyses employed a JEOL JEM-4000EX unit with an accelerating voltage

of 400 kV. Samples were prepared by ultrasonic dispersion in acetone with a drop of the

resultant suspension evaporated onto a holey carbon-supported grid. Mean nickel particle size

evaluated as the surface-area weighted diameter ( sd ), was computed according to:

2

ii

i

3

ii

sdn

dn

d

(Equation 5.8)



Experiments were carried out in a tubular quartz reactor (45 cm long and 1 cm internal

diameter). The catalyst, with particle size in the range 250-500 m and no dilution, was placed

on a fritted quartz plate located at the end of the reactor. The temperature of the catalyst was

measured with a K-type thermocouple (Thermocoax) placed inside the inner quartz tube. The

entire reactor was placed in a furnace (Lenton) equipped with a temperature-programmed

system. Reaction gases were Praxair certified standards of CH4 (99.995% purity), CO2

(99.999% purity), O2 (99.99% purity), and N2 (99.999% purity). The gas flow was controlled

by a set of calibrated mass flowmeters (Brooks 5850 E and 5850 S). The water content in the

reaction mixture was controlled using the vapour pressure of H2O at the temperature of the

saturator (297 K). All lines placed downstream from the saturator were heated above 373 K to

prevent condensation. The saturation of the feed stream by water at the working temperature

was verified by a blank experiment in which the amount of water trapped by a condenser was

measured for a specific time and compared to the theoretical value. The feed composition (by

volume %) was as follows: 6% CH4, 3% CO2, 3% H2O, 0.6% O2, N2 balance, with a total flow

of 100 mL min−1

. This composition was used in previous studies [11-14] to get a molar ratio in

the feed of CH4/CO2/H2O/O2 = 1/0.5/0.5/0.1. The weight hourly space velocity (WHSV) of the

total gas mixture was fixed at 60,000 mL h−1

g−1

.

Prior to the reaction, the catalysts were reduced in a hydrogen pure stream at 973 K. The

catalytic activity was evaluated at 1073 K and atmospheric pressures for 24 h. Gas effluents

were analyzed with a micro gas chromatograph (Varian CP-4900). Methane and carbon


metal impregnation

183

dioxide consumption rates were calculated as follows: [inlet molar flow − outlet molar

flow]/nickel weight.



XRD results for the reduced catalysts are given in Figure 5.1. In all cases the main

diffraction peaks corresponding to -SiC and Ni0 were observed. Peaks related to the support

structurally corresponded to cubic SiC (3C-type), their corresponding Miller indexes being

indicated in Figure 5.1 a). Ni particle sizes (Table 5.1) were obtained with the Debye–Scherrer

equation using data from the XRD patterns (111 reflection of Ni0). This method has some

limitations when it is used for quantitative purposes but it is useful for comparative ones. The

diffraction angle of NiO for the different Mg promoted catalysts before reduction is given in

Table 5.1. The higher the Mg content, the lower the value of this diffraction angles was. This

decrease is usually attributed to the formation of a NiO-MgO solid solution [15, 16], which is

related to the occurrence of a great interaction between Ni and Mg. In addition, for the catalysts

prepared by first Mg impregnation or by simultaneous Ni and Mg impregnation, an even lower

diffraction angle was measured if compared to that of the catalysts prepared by first Ni

impregnation. The formation of a solid solution between two different metals should meet the

criteria determined by Hume-Rothery, which happen for Ni and Mg, as both cations have

similar ionic radii, ca. 0.78 Å [17], the same common oxidation state (+2), and the same bulk

oxide structure, NaCl-type [18]. This solid solution has been observed by several authors when

preparing catalysts where Ni and Mg are present [19-21], and has shown high selectivity and

stability in different reforming processes.

Chapter 5

184

01

02

03

04

05

06

07

08

09

0

400

222

311

220

200

111

Mg/N

i/S

iC 1

/10

Ni/

SiC

SiC

c)b)

ºº

º

^

^

^

^

^ Intensity (a.u.)

2

(º)

a)

01

02

03

04

05

06

07

08

09

0

Mg/N

i/S

iC 1

/1

Ni-

Mg/S

iC 1

/10

Ni/

Mg/S

iC 1

/10

º

º

º

º

º

º

Intensity (a.u.)

2

(º)

^

^^

^

^

^

^^

^

^

^

^^

^^

^

^^

^

^

01

02

03

04

05

06

07

08

09

0

Ni-

Mg/S

iC 1

/1

Ni/

Mg/S

iC 1

/1

º

º

ºº

ºº

Intensity (a.u.)

2

(º)

^

^^

^

^

^

^^

^

^

Fig

ure

5.1

. X

RD

pro

file

s, w

her

e (^

) d

eno

tes

refl

ecti

on

of

SiC

an

d (

º) d

eno

tes

refl

ecti

on

of

met

alli

c n

ick

el.


metal impregnation

185

N

i/S

iC

Mg

/Ni/

SiC

1/1

0

Ni/

Mg

/SiC

1/1

0

Ni-

Mg

/SiC

1/1

0

Mg

/Ni/

SiC

1/1

Ni/

Mg

/SiC

1/1

Ni-

Mg

/SiC

1/1

Ni

loa

din

g (

%)

4.5

4

.6

4.8

5

.2

4.5

4

.2

5.5

Pro

mo

ter

loa

din

g (

%)

- 0

.2

0.2

0

.2

1.8

1

.7

1.8

Su

rfa

ce a

rea

(m

2 g

-1)

23

.9

22

.2

21

.8

22

.5

21

.1

16

.9

21

.2

To

tal

po

re v

olu

me

(cm

3 g

-1)x

10

2

14

.4

10

.7

11

.2

13

.1

13

.9

11

.0

12

.6

Pa

rtic

le d

iam

eter

fro

m X

RD

(n

m)

63

36

41

39

33

33

32

Dif

fra

ctio

n a

ng

le o

f N

iO (

2

43

.22

43

.22

43

.14

43

.14

43

.14

42

.94

42

.9

Pa

rtic

le d

iam

eter

fro

m T

EM

(n

m)

57

- 5

1

- -

32

-

Red

uct

ion

deg

ree

(%)

99

.8

76

.9

73

.3

99

.4

75

.3

57

.8

59

.0

Ta

ble

5.1

. P

hy

sica

l p

rop

erti

es o

f th

e ca

taly

sts.

Chapter 5

186

In order to check the accuracy of the XRD technique to measure the Ni metal particle

size, TEM analyses were also carried out (Table 5.1). Some differences between the values

reported by both techniques are noted although the trend is the same. The higher the load of

Mg in the bimetallic catalysts, the lower the Ni particle size was. This influence of the Mg load

on the Ni particle size is probably related to the formation of NiO-MgO solid solution particles,

as the aggregation of Ni metal particles is depressed during the reduction process due to the

presence of highly dispersed MgO [22]. Figure 5.2 shows TEM images obtained for samples

Ni/SiC, Ni/Mg/SiC 1/10 and Ni/Mg/SiC 1/1. Ni pa rticles in catalysts prepared by first Mg

impregnation were smaller and better dispersed.

b)c)

a)

Fig

ure

5.2

. T

EM

pic

ture

s. a

) N

i/S

iC,

b)

Ni/

Mg

/SiC

1/1

0,

c)N

i/M

g/S

iC 1

/1.


metal impregnation

187

Surface area and pore volume data (Table 5.1) also gave information about the

catalysts. The presence of Mg led to a slight decrease of the surface area and the pore volume

probably due to a partial blockage of the -SiC pores.

Figure 5.3 shows the reduction profiles obtained by the TPR technique. That of the

reference catalyst, Ni/-SiC, placed at the top of the figure, was very similar as reported

elsewhere [11], showing broad peaks and two overlapped ones around 720 and 870 K, which

are usually assigned to the reduction of bulk NiO. The small peak occurred at 1140 K was

related to the reduction of nickel species with a higher interaction with the support, which was

attributed to the formation of nickel silicate like species [23, 24]. There are no clear differences

in the reduction profiles of catalysts with a Mg/Ni molar ratio of 1/10. They showed two main

reduction peaks (at 700-750 K and 980 K). The former was assigned to the reduction of bulk

NiO whereas the later was attributed to the reduction of a NiO-MgO solid solution, phase

formed during the high temperature calcinations process and also observed in the XRD results.

This phase needs higher temperatures in order to be reduced due to the strong interaction

between NiO and MgO. It is clearly noted that the addition of Mg, regardless impregnation

order, decreased the reducibility of Ni.

On the other hand, some differences in the reduction profiles of catalysts prepared with a

Mg/Ni molar ratio of 1/1 were observed. The reduction profile in catalysts prepared by first Ni

impregnation kept closer to that of the previous catalysts, presenting two overlapped peaks:

one with a maximum at about 720 K and another one at about 900 K. Just a reduction peak at

high temperature was observed when Mg was firstly impregnated or when the resulting catalyst

was simultaneously impregnated by Ni and Mg, which was associated to the reduction of NiO-

MgO solid solution, requiring higher reduction temperatures due to the strong interaction

between NiO and MgO.

It can be observed in Table 5.1 the reduction degree obtained from the H2 consumption in

these TPR experiments, taking into account the stoichiometry of the reduction process.

Catalysts with a low Mg load showed a lower reduction degree compared to that of the

reference catalysts, except for the Ni-Mg/SiC sample. This latter catalyst showed a reduction

degree very close to that of the reference catalyst, despite having a profile where reduction

peaks are shifted towards higher temperatures. Catalysts with a high Mg load showed even

Chapter 5

188

lower values of reduction degree. This is in agreement with the higher extension of the NiO-

MgO solid solution, which makes the reduction process more difficult.

Figure 5.3. Temperature Programmed Reduction profiles.

The desorption profiles of CO2-TPD experiments showed a similar trend with a small

desorption peak around 1100 K (Figure 5.4). This desorption peak is related to adsorption

points of strong basicity, as it is required a high temperature in order to desorb the CO2

molecule. There are not significant differences in the quantity of CO2 adsorbed for the different

catalysts, as it keeps relatively low for all of them. However, as a general trend, it can be

observed an increase in the quantity of CO2 adsorbed (and therefore in the catalyst basicity)

with the Mg load increase. This effect is clearer in those catalysts prepared by simultaneous

impregnation of Ni and Mg.

300 400 500 600 700 800 900 1000 1100 1200

Ni-Mg/SiC 1/1

Ni/Mg/SiC 1/1

Mg/Ni/SiC 1/1

Ni-Mg/SiC 1/10

Ni/Mg/SiC 1/10

Mg/Ni/SiC 1/10

Ni/SiC

TC

D s

ign

al

(a.u

.)

Temperature (K)


metal impregnation

189

Figure 5.4. CO2 Temperature Programmed Desorption profiles.


Figures 5.5 and 5.6 show the performance of all catalysts in the methane tri-reforming

process. With the exception of catalysts prepared by first Ni impregnation, the addition of Mg

improved the performance of catalysts Ni/-SiC. Comparing the performance of catalysts with

a Mg/Ni molar ratio of 1/10, it can be seen that catalysts Ni/Mg/SiC 1/10 and Ni-Mg/SiC 1/10

led to the higher methane consumption rate and presented better stability. The addition of Mg

decreased the carbon dioxide consumption rate and increased the H2/CO molar ratio, especially

when Mg was firstly impregnated. This behaviour seems not to be linked with the increase in

the catalyst basicity that the presence of Mg usually induces. However, as reported in the

previous chapter, it could be related to both the strong basicity induced in the SiC support [25]

and the higher presence of NiO species in the catalysts (as the Mg loaded catalysts are more

difficult to reduce). NiO promotes the water gas shift reaction, resulting in a lower CO2

consumption rate [26], what will increase in addition the H2/CO molar ratio. Catalysts

300 400 500 600 700 800 900 1000 1100 1200 1300

Ni-Mg/SiC 1/1

Ni/Mg/SiC 1/1

Mg/Ni/SiC 1/1

Ni-Mg/SiC 1/10

Ni/Mg/SiC 1/10

Mg/Ni/SiC 1/10

Ni/SiC

TC

D s

ign

al

(a.u

.)

Temperature (K)

Chapter 5

190

Ni/Mg/SiC 1/10 and Ni-Mg/SiC 1/10 presented a remarkable stability after 24 h on stream if

compared to that of reference sample, demonstrating the positive influence of these preparation

methods for the tri-reforming process. Catalyst Mg/Ni/SiC 1/10 showed in turn a very low

catalytic activity, probably due to the blockage of the Ni active sites after Mg impregnation.

Figure 5.5. Catalytic activity at 1073 K for: a) Ni/SiC, b) Mg/Ni/SiC 1/10 c) Ni/Mg/SiC 1/10

d) Ni-Mg/SiC 1/10. Reaction conditions: CH4 = 6%, CO2 = 3%, H2O = 3%, O2 = 0.6%, N2


. CH4 ( ) and CO2 ( ) consumption rates vs. time on

stream (left axis), and H2/CO molar ratio ( ) vs. time on stream (right axis).

0 5 10 15 20 25

0

1

2

3

4

5

6

7

8

9

10a)

Consu

mpti

on r

ate

(mol

s-1 g

-1 Ni )

*10

4

Time (h)

0 5 10 15 20 25

0

1

2

3

4

5

6

7

8

9

10b)

Consu

mpti

on r

ate

(mol

s-1 g

-1 Ni )

*10

4Time (h)

0 5 10 15 20 25

0

1

2

3

4

5

6

7

8

9

10c)

Consu

mpti

on r

ate

(mol

s-1 g

-1 Ni )

*10

4

Time (h)

0 5 10 15 20 25

0

1

2

3

4

5

6

7

8

9

10

d)

Consu

mpti

on r

ate

(mol

s-1 g

-1 Ni )

*10

4

Time (h)

0.5

1.0

1.5

2.0

2.5

3.0

3.5

H2/C

O M

ola

r ra

tio

2.0

2.5

3.0

3.5

4.0

4.5

5.0

H2/C

O M

ola

r ra

tio

2.0

2.5

3.0

3.5

4.0

4.5

5.0

H2/C

O M

ola

r ra

tio

0.5

1.0

1.5

2.0

2.5

3.0

3.5

H2/C

O M

ola

r ra

tio


metal impregnation

191

Figure 5.6. Catalytic activity at 1073 K for: a) Mg/Ni/SiC 1/1 b) Ni/Mg/SiC 1/1 c) Ni-Mg/SiC

1/1. Reaction conditions: CH4 = 6%, CO2 = 3%, H2O = 3%, O2 = 0.6%, N2 balance, total flow

rate = 100 mL min-1

. CH4 ( ) and CO2 ( ) consumption rates vs. time on stream (left axis),

and H2/CO molar ratio ( ) vs. time on stream (right axis).

Table 5.2 lists the main reaction parameters obtained with all the catalysts. The reference

catalyst showed a good methane consumption rate but a remarkable deactivation after 24 h of

time on stream. The addition of Mg as promoter after Ni impregnation led to a decrease of both

the methane consumption rate and the catalyst stability. This behaviour seems to be related to

the predominant interaction of the Ni particles with the support, rather than with the promoter

[27], as well as to the more difficult access of the reaction gases to the Ni metal particles. In

0

1

2

3

4

5

6

7

8

9

10

Co

nsu

mp

tio

n r

ate

(mo

l s-1

g-1 N

i )*

10

4

a)

Co

nsu

mp

tio

n r

ate

(mo

l s-1

g-1 N

i )*

10

4

0

1

2

3

4

5

6

7

8

9

10

b)

0 5 10 15 20 25

0

1

2

3

4

5

6

7

8

9

10c)

Co

nsu

mp

tio

n r

ate

(mo

l s-1

g-1 N

i )*

10

4

Time (h)

2.0

2.5

3.0

3.5

4.0

4.5

H2/C

O M

ola

r ra

tio

1.5

2.0

2.5

3.0

3.5

4.0

H2/C

O M

ola

r ra

tio

1.5

2.0

2.5

3.0

3.5

4.0

H2/C

O M

ola

r ra

tio

Chapter 5

192

addition, the H2/CO molar ratio obtained was very high while the CO2 consumption rate was

kept very low, demonstrating that dry reforming did not greatly contribute to the global tri-

reforming process. Methane consumption rate values and catalytic stability for catalysts

Ni/Mg/SiC and Ni-Mg/SiC with a Mg/Ni molar ratio of 1/10 resembled very similar, showing

a catalytic activity slightly higher than that obtained for the reference catalyst. The presence of

Mg in these catalysts had a positive effect, which was related to the above mentioned

interaction between Ni and Mg. In addition, this interaction allowed both the presence of

smaller Ni particles, favouring a higher reforming activity and a smaller coke rate deposition

[28, 29], and the simultaneous formation of a Ni-Mg solid solution, confirmed by the XRD

results and the TPR experiments, leading to Ni to strongly interact with Mg, which in turn also

hindered coke formation [30, 31].

Figure 5.6 shows the performance of catalysts with a Mg/Ni molar ratio of 1/1. Very

high and stable catalytic activity was obtained for samples Ni/Mg/SiC and Ni-Mg/SiC, being

the values of methane consumption rate around 9x10-4

mol s-1

gNi-1

. Similar values were

obtained for catalysts with a Mg/Ni molar ratio of 1/10 but, H2/CO molar ratio in the

downstream was slightly higher. Again, the catalyst prepared by first Ni impregnation showed

the poorest catalytic behaviour, with very low methane and carbon dioxide consumption rates

and a worse stability with the time on stream.

Table 5.2 also lists the main reaction parameters obtained with all the catalysts. Catalyst

Mg/Ni/SiC 1/1 showed the worst catalytic results at all, including those of the reference

catalyst. As previously commented, the impregnation of Mg after Ni made the latter to weakly

interact with the former and hinder methane access to the active sites. As was confirmed by

XRD, the higher the amount of Mg, the stronger the interaction of both metals was. This higher

interaction seemed to improve the catalytic activity but did not balance the negative effect of

Ni particles blockage. Methane consumption rates for catalysts Ni/Mg/SiC and Ni-Mg/SiC

with a Mg/Ni molar ratio of 1/1 resembled those obtained for catalysts prepared with a Mg/Ni

molar ratio of 1/10. However, the formers presented a better catalytic performance due to the

beneficial effect of Mg on the catalyst stability, due to a lower coke formation rate, as it will be

discussed in the next section.


metal impregnation

193

N

i/S

iC

Mg

/Ni/

SiC

1/1

0

Ni/

Mg

/SiC

1/1

0

Ni-

Mg

/SiC

1/1

0

Mg

/Ni/

SiC

1/1

Ni/

Mg

/SiC

1/1

Ni-

Mg

/SiC

1/1

Av

era

ge

CH

4 r

eact

ion

ra

te

(mo

l s-1

gN

i-1)

×1

04

7.9

6

3.9

9

8.5

0

8.9

2

5.2

8

8.7

7

8.6

5

Dro

p i

n C

H4 r

eact

ion

ra

te

aft

er 2

4 h

of

tim

e o

n s

trea

m

(mo

l s-1

gN

i-1)

×1

04

1.3

0

4.3

4

0.3

0

0.2

4

1.2

7

0.0

0

0.0

8

Av

era

ge

CO

2 r

eact

ion

ra

te

(mo

l s-1

gN

i-1)

×1

04

2.7

2

0.1

2

1.0

5

2.7

2

0.2

5

1.4

8

1.8

4

Dro

p i

n C

O2 r

eact

ion

ra

te

aft

er 2

4 h

of

tim

e o

n s

trea

m

(mo

l s-1

gN

i-1)

×1

04

0.7

6

1.5

5

0.5

1

0.4

5

0.3

5

2.5

6

0.7

4

Av

era

ge

H2/C

O m

ola

r ra

tio

2

.00

3.7

3

2.9

6

2.0

9

3.3

6

2.7

9

2.7

6

Oxy

gen

co

nsu

mp

tio

n i

n

TP

O (

mo

l g

-1)

35

.28

23

.44

4.0

4

10

.70

10

.81

3.7

2

6.2

4

Ta

ble

5.2

. R

eact

ion

an

d c

har

acte

riza

tio

n a

fter

rea

ctio

n p

aram

eter

s.

Chapter 5

194

5.3.3. Characterization after reaction.

Temperature programmed oxidation was performed on the aged catalysts in order to

quantify the coke generated during the tri-reforming process (Figure 5.7). As observed, the

addition of Mg avoided the formation of coke, especially for the catalysts prepared with a

Ni/Mg molar ratio of 1/1 as a probable consequence of their lower Ni particle size and higher

interaction between Ni and Mg. The total amount of oxygen consumed in the TPO experiments

is shown in Table 5.2. Aged catalysts with the lower coke content were those prepared by first

Mg impregnation. They were followed by those prepared by simultaneous impregnation. Two

peaks with maxima around 900 and 1000 K, which would correspond to the occurrence of the

Cand C coke species reported by Zhang et al. [32], are observed in Figure 5.7. The former

would be related to the generation of CO at high reaction temperatures whereas the latter

would be responsible of the catalyst deactivation. Our results show that the addition of Mg

decreases the formation of both coke species, leading to an increase of the catalyst stability.

Figure 5.7. Temperature Programmed Oxidation profiles after reaction.

300 400 500 600 700 800 900 1000 1100 1200 1300

Ni-Mg/SiC 1/1

Ni/Mg/SiC 1/1

Mg/Ni/SiC 1/1

Ni-Mg/SiC 1/10

Ni/Mg/SiC 1/10

Mg/Ni/SiC 1/10

Ni/SiC

TC

D s

ign

al

(a.u

.)

Temperature (K)


metal impregnation

195

Figure 5.8 reports the X-ray diffraction patterns obtained for the aged catalysts. The

reference sample (Figure 5.8 a)) showed two peaks at 45.50 and 48.76 º, which can be ascribed

to the orthorhombic phase of Ni2Si [33]. This compound was also present on the surface of

aged catalysts Mg/Ni/SiC 1/10 and Mg/Ni/SiC 1/1, as observed in Figures 5.8 b) and 5.8 c),

respectively, where the corresponding peaks appeared sharp and well defined at the same 2

values. This phase, which is stable up to 1223 K, is a consequence of the direct reaction

between metallic Ni and SiC that is thermally activated over 873 K [34]. In other catalysts,

these peaks are smaller or do not exist, indicating that the simultaneous or previous Mg

impregnation decreases the interaction between Ni and Si. According to the results listed in

Table 5.2, the formation of Ni2Si (leading to a lesser availability of Ni active sites) seems to be

related to lower methane reaction rates and a higher extension of the deactivation processes.

Diffraction peaks assigned to NiO also appeared in aged catalysts, although they are less

evident in catalysts Ni-Mg/SiC 1/10 and Ni-Mg/SiC 1/1. These catalysts led to a higher

average CO2 reaction rate if compared to that of catalysts Ni/Mg/SiC 1/10 and Ni/Mg/SiC 1/1.

Consequently, it could be concluded that the presence of NiO species in the latter catalysts

should promote the water gas-shift reaction.

Chapter 5

196

30

32

34

36

38

40

42

44

46

48

50

52

54

56

58

60

**

º#

#

Mg/N

i/S

iC 1

/10

Ni/

SiC

c)b

)

ºº

º

^

^

^

^ Intensity (a.u.)

2

(º)

a)

30

32

34

36

38

40

42

44

46

48

50

52

54

56

58

60

**

***

*

# # ####

ººº

ººº

^^^

^^^

Mg/N

i/S

iC 1

/1

Ni-

Mg/S

iC 1

/10

Ni/

Mg/S

iC 1

/10

Intensity (a.u.)

2

(º)

30

32

34

36

38

40

42

44

46

48

50

52

54

56

58

60

**

***

*# # #

###º º º

ººº^ ^ ^

^^^

Ni-

Mg/S

iC 1

/1

Ni/

Mg/S

iC 1

/1

Intensity (a.u.)

2

(º)

Fig

ure

5.8

. X

RD

pro

file

s, w

her

e (^

) d

eno

tes

refl

ecti

on

of

SiC

, (º

) d

eno

tes

refl

ecti

on

of

met

alli

c n

ick

el,

(*)

den

ote

s re

flec

tio

n o

f n

ick

el

ox

ide

and

(#)

den

ote

s re

flec

tion

of

Ni 2

Si.


metal impregnation

197

5.4. CONCLUSIONS

The addition of Mg to a Ni/-SiC catalyst, whenever it is not loaded after Ni

impregnation, promoted the catalytic behaviour of the methane tri-reforming process. Catalysts

where Ni was firstly impregnated showed the worst catalytic behaviours, probably due to a

poor interaction between Ni and Mg, a possible blockage of the Ni particles by Mg and the

formation of Ni2Si, which decreased the number of Ni active sites. Catalysts prepared with a

higher Mg/Ni molar ratio (1/1) showed smaller Ni particle sizes, a lower coke rate formation, a

higher basicity and a higher Ni-Mg interaction. Catalysts where Mg was firstly impregnated

were less deactivated keeping a good catalytic behaviour. Simultaneous impregnation of Ni

and Mg yielded catalysts with the best catalytic performances, which was related to the high

interaction between Ni and Mg due to the formation of a Ni-Mg solid solution. Catalyst Ni-

Mg/SiC 1/1 was selected as the best one due to its high catalytic activity, great stability and

low coke production.

Chapter 5

198

5.5 REFERENCES

[1] U.S.D.o.E.O.o. Science, U.S.O.o.F. Energy, Carbon Sequestration : State of the Science: A

Working Paper for Roadmapping Future Carbon Sequestration R&D, 1999.

[2] J. Xu, W. Zhou, Z. Li, J. Wang, J. Ma, International Journal of Hydrogen Energy. 34

(2009) 6646-6654.



78.


(2007) 153-158.


B: Environmental. 101 (2011) 531-539.


377 (2010) 16-26.

[8] H. Özdemir, M.A. Faruk Öksüzömer, M. Ali Gürkaynak, International Journal of Hydrogen

Energy. 35 (2010) 12147-12160.


[10] J.M. García-Vargas, J.L. Valverde, A. de Lucas-Consuegra, B. Gómez-Monedero, F.

Dorado, P. Sánchez, Int. J. Hydrogen Energy. 38 (2013) 4524-4532.



[12] C. Song, W. Pan, Catalysis Today. 98 (2004) 463-484.


[14] J.M. García-Vargas, J.L. Valverde, J. Díez, P. Sánchez, F. Dorado, Applied Catalysis B:

Environmental. 148-149 (2014) 322-329.


metal impregnation

199

[15] Y.-H. Wang, H.-M. Liu, B.-Q. Xu, J. Mol. Catal. A: Chem. 299 (2009) 44-52.

[16] F. Arena, F. Frusteri, A. Parmaliana, L. Plyasova, A.N. Shmakov, J. Chem. Soc., Faraday

Trans. 92 (1996) 469-471.

[17] R.A.T.P.K. Flinn, Engineering materials and their applications, Wiley, New York;

Chichester, 1995.

[18] V.E.C.P.A. Henrich, The surface science of metal oxides, Cambridge University Press,

Cambridge; New York, 1994.

[19] Y. Hu, E. Ruckenstein, Catal. Lett. 36 (1996) 145-149.

[20] K. Tomishige, Y. Himeno, Y. Matsuo, Y. Yoshinaga, K. Fujimoto, Industrial &

Engineering Chemistry Research. 39 (2000) 1891-1897.

[21] Y.H. Hu, Catal. Today. 148 (2009) 206-211.

[22] T. Nakayama, N. Ichikuni, S. Sato, F. Nozaki, Applied Catalysis A: General. 158 (1997)

185-199.


159.



104 (2011) 64-73.


(2008) 97-104.

[27] Z. Cheng, Q. Wu, J. Li, Q. Zhu, Catal. Today. 30 (1996) 147-155.

[28] D.L. Trimm, Catal. Today. 49 (1999) 3-10.

[29] V.C.H. Kroll, H.M. Swaan, C. Mirodatos, J. Catal. 161 (1996) 409-422.


Chapter 5

200


(2006) 9-22.


[33] X. Chen, A. Zhao, Z. Shao, C. Li, C.T. Williams, C. Liang, The Journal of Physical

Chemistry C. 114 (2010) 16525-16533.

[34] F. Basile, P.D. Gallo, G. Fornasaria, D. Gary, V. Rosetti, A. Vaccari, in: F.B. Noronha, M.

Schmal, E.F. Sousa-Aguiar (Eds.), 2007, pp. 313-318.

CHAPTER 6 Catalytic and kinetic analysis of the methane tri-reforming process using a Ni-Mg/-SiC

catalyst

Resumen

Abstract

6.1. INTRODUCTION

6.2. EXPERIMENTAL




6.2.4. Kinetic analysis





6.4. CONCLUSIONS

Catalytic and kinetic analysis of the methane tri-reforming process using a Ni-Mg/-SiC

catalyst

203

Resumen

En este capítulo se ha analizado la influencia de la temperatura y la composición

del alimento en el comportamiento catalítico de un catalizador Ni-Mg/-SiC aplicado

al proceso de tri-reformado. La caracterización del catalizador incluyó las técnicas de

absorción atómica, reducción a temperatura programada, adsorción de N2, desorción

de CO2 a temperatura programada y DRX. Se realizaron 36 experimentos con

diferente composición del alimento, obteniéndose datos a 12 temperaturas diferentes.

La influencia de cada una de las reacciones que ocurren durante el proceso de tri-

reformado ha sido evaluada en función de la temperatura, observándose una mayor

contribución del reformado con vapor y la reacción de water gas shift a baja

temperatura y una mayor contribución de la reacción de reformado seco a alta

temperatura. Por último, se ha desarrollado un modelo cinético para representar los

resultados experimentales obtenidos. Para ello se han tenido en cuenta como

reacciones relevantes a nivel cinético el reformado con vapor, el reformado seco y la

reacción de water gas shift, obteniéndose un buen ajuste de los datos obtenidos del

modelo a los experimentales.

Chapter 6

204

Abstract

In the present work we have analyzed the influence of the temperature and feed

composition in the catalytic behaviour of a Ni-Mg/-SiC catalyst. The catalyst was

characterized by AAS, TPR, N2 adsorption, CO2-TPD and XRD. 36 experiments with

different feed composition were performed, obtaining catalytic data at 12 different

temperatures. It was evaluated the predominance of each one of the different reactions

that take place in the tri-reforming process depending on the temperature, with a

higher contribution of the steam reforming and water gas shift at low temperatures and

a higher contribution of the dry reforming at high temperatures. Finally, a kinetic

model was developed in order to fit the experimental data. We consider the steam

reforming, the dry reforming and the water gas shift as the kinetically relevant

equations, obtaining a good fit of the experimental data to the modelled one.


catalyst

205

6.1. INTRODUCTION

Interest in the conversion of CO2 into valuable chemical compounds has been growing in

the last years both due to the harmful effect that the emissions of this gas have in the

environment, affecting specially to the climate, being the increase in the carbon dioxide

atmospheric concentration generally accepted as the most important cause of the global

warming effect [1-4]; and the decrease in the petroleum reserves and the consequent increase in

the price of oil, what makes interesting the possibility of obtaining carbon derived compounds

from a widely abundant an relatively cheap source.

In this way, dry reforming (Equation 6.1) could be a feasible route in order to convert

CO2 into valuable chemical compounds via synthesis gas. This reaction has attracted some

interest in the last years and several groups have analyzed its characteristics [5-10]

CO2 + CH4 → 2CO + 2H2 (H◦ = 247.3 kJmol

-1) (Equation 6.1)

However, this process has two main drawbacks that are hindering its use in industry:

firstly, coke formation causes the catalyst to rapidly deactivate and, secondly, a great deal of

energy is consumed as a consequence of the endothermic nature of this process.

Tri-reforming of methane offers an alternative to the dry reforming and the other

reforming process, in order to obtain synthesis gas from CH4 and CO2. This process consists in

a synergetic combination of dry reforming (Equation 6.1), steam reforming (Equation 6.2) and

partial oxidation (Equation 6.3). It shows some advantages over the dry reforming process, as a

higher resistance against coke deactivation due to the presence of oxidants (Equations 6.4 and

6.5), the lower energy consume due to the presence of the exothermic partial oxidation and the

possibility of shifting the H2/CO molar ratio to a desired value modifying the feed composition.

H2O + CH4 → CO + 3H2 (H◦ = 206.3 kJmol

-1) (Equation 6.2)

CH4 + 1/2O2 → CO + 2H2 (H◦ = −35.6 kJmol

-1) (Equation 6.3)

C + H2O CO + H2 (H◦ = 131.4 kJmol

-1) (Equation 6.4)

C + O2 CO2 (H◦ = −393.7 kJmol

-1) (Equation 6.5)

Chapter 6

206

Chapters 4 and 5 deal with the development of a very active and stable Mg promoted

Ni/-SiC catalyst. In this stage we try to develop a predictive model for the tri-reforming of

methane which fits with the experimental results obtained with the catalyst previously referred

in the tri-reforming of methane.

6.2. EXPERIMENTAL


The selection of the catalyst was based in a previous study [11], reported also in chapter

4. A Mg promoted, Nickel supported catalyst was prepared, with a Ni/Mg molar ratio of 2/1

and using -SiC pellets (1 mm diameter) as support. The catalyst was prepared by the

impregnation method using nickel nitrate Ni(NO3)2·6H2O (PANREAC) and magnesium

hydroxide Mg(OH)2 (PANREAC). The support used was provided by SICAT CATALYST. A

solution containing both nickel nitrate and magnesium hydroxide was prepared with the

corresponding amount to yield a 5wt% Ni catalyst and a Ni/Mg molar ratio of 2. After the

impregnation procedure the catalyst was dehydrated at 393 K for 12 h and subsequently

calcined in air at 1173 K for 2 h.


Ni and Mg metal loading were determined by atomic absorption (AA) spectrophotometry,

using a SPECTRA 220FS analyzer. Samples (ca. 0.5 g) were treated in 2 mL HCl, 3 mL HF

and 2 mL H2O2 followed by microwave digestion (523 K). In order to calculate textural

properties (surface area and total pore volume) samples were outgased at 453 K under vacuum

for 12 h and analyzed afterwards in a QUADRASORB 3SI sorptometer apparatus with N2 as

the sorbate at 77 K. Temperature-programmed reduction (TPR) experiment was conducted in a

commercial Micromeritics AutoChem 2950 HP unit with TCD detection. Sample (ca. 0.15 g)

was loaded into a U-shaped tube and ramped from room temperature to 1173 K (10 K min−1

),

using a reducing gas mixture of 17.5% v/v H2/Ar (60 cm3 min

−1). CO2 temperature-

programmed desorption (TPD) experiment was also conducted in the Micromeritics AutoChem

2950 HP unit. The sample (0.15 g) was loaded in a quartz tube, reduced and pre-treated. Then,

a flow of 30 mL min-1

of CO2 (99.99% purity, Praxair certified) was passed through the sample

for 30 min at a constant temperature of 313 K. Finally, the physically adsorbed carbon dioxide


catalyst

207

was removed by a flow of He for another 30 min. The sample was then heated in 50 mL min-1


up to 1273 K. XRD analyses were conducted with a

Philips X’Pert instrument using nickel-filtered Cu Kα radiation. The samples were scanned at a

rate of 0.02° step−1

over the range 5° ≤ 2θ ≤ 90° (scan time = 2 s step−1

.



internal diameter). The catalyst was placed on a fritted quartz plate located at the end of the

reactor. The reactor was heated with a furnace (Lenton) and the temperature measured with a

K-type thermocouple (Thermocoax). Reaction gases were Praxair certified standards of CH4

(99.995% purity), 10% CO2/N2, O2 (99.99% purity), and N2 (99.999% purity). The water


temperature of the saturator required for each condition. The temperature of the saturator was

controlled by a heating bath. All lines placed downstream from the saturator were heated above

373 K to prevent condensation. The saturation of the feed stream by water at the working

temperature was verified by a blank experiment in which the amount of water trapped by a

condenser was measured for a specific time and compared with the theoretical value. The feed

composition differs for each experiment but the total flow was kept always at 100 NmL min-

1using N2 as balance. Specific composition for each experiment can be seen in Table 6.1. The

weight hourly space velocity (WHSV) of the total gas mixture was fixed at 60000 NmL h-1

g-1

.

The catalytic activity was evaluated from 680 K to 1073 K and atmospheric pressure. Gas

effluents were analyzed with a micro gas chromatograph (Varian CP-4900). Methane and

carbon dioxide consumption rates were calculated as follows: [inlet molar flow of CH4/CO2 –

outlet molar flow of CH4/CO2]/nickel weight. A blank experiment carried out with pure SiC

showed no appreciable conversion in the considered conditions.


In this work, a kinetic model for the tri-reforming process was developed in order to

represent the results obtained in the catalytic experiments reported. We consider single reaction

equations for the steam reforming, dry reforming and water gas shift, assuming that the partial

oxidation is close to the equilibrium as the experimental results showed almost no oxygen in

the effluent gas flow for all the experiments carried out. The set of equations was based in the

model proposed by J. Wei and E. Iglesia [12] for the steam and dry reforming reactions. The

Chapter 6

208

kinetic equation for the water gas shift reaction was selected based in a previous work of our

group [13].

Table 6.1. Feed composition (NmL min-1

).

Experiment Run CH4 CO2 H2O O2 Experiment Run CH4 CO2 H2O O2

1 11 50 25 20 3 19 26 37.5 16 3 3

2 5 50 25 20 0.6 20 33 37.5 16 3 0.6

3 4 50 25 3 3 21 18 37.5 7 20 3

4 3 50 25 3 0.6 22 22 37.5 7 20 0.6

5 9 50 16 20 3 23 15 37.5 7 3 3

6 2 50 16 20 0.6 24 23 37.5 7 3 0.6

7 12 50 16 3 3 25 24 25 25 20 3

8 16 50 16 3 0.6 26 30 25 25 20 0.6

9 6 50 7 20 3 27 34 25 25 3 3

10 7 50 7 20 0.6 28 13 25 25 3 0.6

11 29 50 7 3 3 29 28 25 16 20 3

12 36 50 7 3 0.6 30 14 25 16 20 0.6

13 8 37.5 25 20 3 31 19 25 16 3 3

14 1 37.5 25 20 0.6 32 32 25 16 3 0.6

15 21 37.5 25 3 3 33 20 25 7 20 3

16 35 37.5 25 3 0.6 34 17 25 7 20 0.6

17 10 37.5 16 20 3 35 27 25 7 3 3

18 25 37.5 16 20 0.6 36 31 25 7 3 0.6

The reactants flow through the reactor and catalyst bed was modelled with a

pseudohomogeneous, one-dimensional model. Isothermal conditions and no pressure drops

were assumed. Therefore, the following expression for the axial flow profiles through the

reactor for the fed gases Pi can be used (Equation 6.6):


catalyst

209

(Equation 6.6)

A VBA-Excel application was developed to solve this model [14-16]. The Bader-

Deuflhard method was used in the evaluation of the set of ordinary differential equations [17],

whereas the Marquardt-Levenberg algorithm was used in the nonlinear regression procedure

[18, 19]. The weighted sum of the squared differences between the observed and the calculated

outlet flow rates was minimized [14]:

(Equation 6.7)

The index i indicates the species considered, and the terms Fexp and Fth respectively

denote the experimental and theoretical molar flow rates for each component. An F test was

carried out in order to compare the fit of the different models to the data. The procedure was

based on the comparison between the tabulated F value (F test) and Fc, which is defined by the

following equation [18, 20]:

(Equation 6.8)

If Fc is larger than F (p, N-p, 1-a) (assuming a value of = 0.05, 95% confidence

level), the regression is considered to be meaningful, although there is no guarantee that the

model is statistically suitable since the meaningfulness of each parameter in the model must be

also evaluated. Hence, a complementary test, named t-test, was used. The t-test is a statistical

hypothesis test in which the test statistic follows a Student’s t distribution and allows verifying

if the estimate of bi (bfi) differs from a reference value (generally zero). Thus, a parameter is

meaningful each time that the following inequality occurs:

(Equation 6.9)

where [V(bf )]ii represents the diagonal jth term of the covariance matrix.



Catalyst characterization includes BET analysis. Figure 6.1 shows the N2 adsorption

isotherm of both support and catalyst. The two samples presented a type II–IV isotherm

Chapter 6

210

(IUPAC classification) that is characteristic of macroporous and mesoporous materials [21],

without nitrogen adsorption in the micropore range (P/P0 ≤ 0.03) and a small increase in the

mesopore range. It can be seen that the addition of Ni and Mg decreases the surface area, as the

volume of N2 adsorbed at a relative pressure close to 1 is lower for the catalyst. In addition, the

mesopore size distribution shows that in both cases the pores with a radius around 10 nm are

predominant, but the catalysts showed a decrease in the quantity of this kind of pores compared

with the support, which seems to indicate a partial blockage of these pores due to the

impregnation process. The main textural properties measured for the catalyst are summarized

in Table 6.2.

Figure 6.1. Nitrogen adsorption–desorption isotherms of catalyst and support.

0,00 0,25 0,50 0,75 1,00

0

20

40

60

80

N2

adso

rbed

(cm

3 g

-1 )

Relative Pressure (P/P0 )

1 10 100 1000 10000

0,00

0,05

0,10

0,15

0,20

0,25

Catalyst

Catalyst

Support

Support

dV

/d(l

og

r)

Pore radius nm)


catalyst

211


Characterization parameter Numerical value

Ni loading (%) 5.0

Mg loading (%) 0.77

Surface area (m2 g

-1) 20.5

Total pore volume (cm3 g

-1) x10

2 9.1

Particle diameter from XRD (nm) 37

Total basic sites (mol g-1

) 10.51

Reduction Temperature (K) 973

Reduction degree (%) 63.7

X-ray diffraction experiments were carried out for support and catalyst before and

after reduction. Figure 6.2 a) displays the X-ray diffractogram obtained for each sample.

Diffraction peaks of the support structurally correspond to cubic SiC (3C-type) [22]. These

peaks could be clearly identified in the other two samples. In addition, peaks corresponding to

NiO and Ni0 could be identified in the diffractogram of catalyst before reduction and after

reduction respectively. Ni particle size was obtained with the Debye–Scherrer equation using

the data from the XRD pattern (111 reflection of Ni0), yielding a value of 37 nm (Table 6.2).

Reduction behaviour was analyzed by a Temperature Programmed Reduction

experiment. The reduction profile can be observed in Figure 6.2 b), showing a principal

reduction peak obtained around 1010 K and two smaller peaks, one with maximum at 660 K

and other overlapped with the principal one and maximum at 755 K. This reduction profile is

similar to that reported by our group previously for a similar catalyst [11] in form of powder

instead of pellets. The addition of Mg as promoter to a Ni/-SiC catalyst could yield a great

interaction between Ni and Mg, with the formation of a Ni-Mg-O solid solution [23, 24] that

needs very high temperatures in order to be reduced. Reduction degree was calculated taking

into account the quantity of H2 consumed during the TPR experiments, obtaining a reduction

degree of 63.7% (Table 6.2).

Chapter 6

212

01

02

03

04

05

06

07

08

09

0º º

c)b)

a)

Intensity (a.u.)

2

(º)

400

222

311

220

200

111

ºº

º

º

+

º

ºº

º

º

º

+

++

^^

^

30

04

50

60

07

50

90

01

05

01

20

0

Red

uce

d

Fre

sh

Support

TCD signal (a.u.)

Tem

per

ature

(K

)

30

04

50

60

07

50

90

01

05

01

20

0

TCD signal (a.u.)

Tem

per

ature

(K

)

Fig

ure

6.2

. C

har

acte

riza

tion

res

ult

s a)

XR

D p

rofi

les,

wh

ere

(º)

den

ote

s re

flec

tio

n o

f

SiC

, (^

) d

eno

tes

refl

ecti

on

of

met

alli

c n

ick

el a

nd

(+

)

den

ote

s re

flec

tio

n o

f n

ick

el o

xid

e,

b)

TP

R p

rofi

le,

c) C

O2-T

PD

pro

file

s.


catalyst

213

Basicity of the catalyst was evaluated in terms of CO2 Temperature Programmed

Desorption. This analysis showed a principal desorption peak at 1115 K (Figure 6.2 c)),

yielding a quantity of basic sites of 10.51 mol g-1

(Table 6.2).


Tri-reforming experiments were performed at different temperature and feed composition.

Table 6.1 summarizes the different feed composition selected for the 36 experiments, which

were carried out randomly. Methane conversion, carbon dioxide conversion and the H2/CO

molar ratio of the synthesis gas obtained for each feed conditions and temperature can be

observed in Table 6.3, Table 6.4 and Table 6.5 respectively. In order to have a clear sight of

these results, Figures 6.3-6.5 also summarize them. In the first place, methane conversion in

the tri-reforming process has a clear dependence on the temperature, showing progressively

higher values while increasing the temperature. This effect is in concordance with the high

endotermicity of steam and dry reforming (Equations 6.1 and 6.2). Regarding CO2 conversion,

it could be seen how at low temperatures and for certain feed composition CO2 conversion

values are below 0, which actually means that there is a net production of CO2 in the process,

probably due to the concurrence of the water gas-shift equilibrium (Equation 6.10), which is an

exothermic reaction, so it is favoured at low temperatures. The influence of the water gas-shift

equilibrium and its relation with the temperature can also be observed in the H2/CO molar ratio

of the synthesis gas obtained from the tri-reforming experiments. In the low temperature range

H2/CO molar ratios above 3 were obtained, which indicates the presence of the water gas shift

reaction increasing the production of H2 and CO2 by the reaction between H2O and the CO

obtained from the reforming reactions, so the H2/CO molar ratio could surpass the maximum

theoretical value considering the stoichometry of the steam reforming reaction. As we increase

the reaction temperature, water gas shift reaction loses some importance while the endothermic

dry reforming increases its contribution to the tri-reforming process, what yields lower values

of H2/CO molar ratio and higher values of CO2 conversion at high temperatures.

CO + H2O

CO2 + H2 (H◦ = –37.09 kJ/mol) (Equation 6.10)

Chapter 6

214

Table 6.3. CH4 conversion values obtained for each experiment.

Exp. Temperature

680 710 740 770 800 830 860 890 920 950 980 1010

1 8.36 0.00 11.01 8.77 13.35 19.86 24.81 26.90 34.72 40.27 46.63 53.90

2 3.48 4.08 4.89 3.16 1.19 2.25 8.28 8.94 10.28 13.54 13.80 28.27

3 0.00 0.19 5.31 6.99 12.34 11.99 15.35 17.90 20.15 21.46 25.08 30.14

4 0.00 0.38 1.24 2.73 4.91 7.27 9.14 10.21 11.79 14.74 19.05 24.53

5 9.95 5.28 6.94 9.31 12.38 16.03 22.14 27.50 34.09 39.70 46.82 55.02

6 2.39 3.15 4.35 6.11 8.65 11.79 16.27 21.29 26.92 32.73 37.81 42.32

7 0.00 0.25 0.00 2.15 5.02 8.84 13.06 17.69 20.77 20.78 22.16 26.26

8 8.45 7.39 9.59 11.86 13.68 16.20 19.04 21.17 22.69 23.41 24.82 28.45

9 0.26 1.92 3.54 6.88 11.08 12.64 17.27 19.24 23.05 28.23 34.76 53.52

10 3.00 2.75 4.54 6.81 9.35 12.85 16.51 21.05 26.20 32.13 38.32 43.31

11 6.83 9.15 10.52 12.05 13.95 16.35 18.87 20.01 18.96 18.72 20.70 25.13

12 8.83 7.04 8.11 9.32 11.44 13.47 15.21 14.31 13.56 14.32 16.73 19.38

13 0.00 3.87 6.29 9.01 12.85 17.38 22.88 27.92 37.57 48.24 57.02 66.72

14 1.58 2.56 4.49 7.14 10.64 15.24 20.78 27.31 34.73 42.59 50.94 59.07

15 4.26 8.67 10.39 12.59 15.61 19.14 23.02 28.06 34.24 40.90 47.59 55.70

16 7.55 7.25 8.57 10.83 14.05 18.29 22.09 24.59 25.96 28.56 33.38 38.82

17 0.00 0.00 2.04 6.51 10.68 15.52 21.55 28.76 38.53 49.75 59.15 66.51

18 7.21 8.18 10.62 13.57 15.90 19.89 24.51 29.65 35.62 42.20 49.80 56.79

19 2.91 5.72 7.78 10.41 13.48 17.31 21.19 24.60 26.87 27.73 30.44 35.55

20 4.33 4.79 6.11 7.99 10.99 14.54 17.59 19.15 20.29 21.02 24.19 28.40

21 4.83 10.32 12.42 15.27 18.92 22.59 26.69 30.78 37.18 44.99 53.44 61.82

22 4.26 8.67 10.39 12.59 15.61 19.14 23.02 28.06 34.24 40.90 47.59 55.70

23 0.00 5.24 7.19 9.10 11.74 15.49 19.81 23.80 30.63 36.11 36.73 38.10

24 2.46 3.29 4.53 6.50 8.83 11.56 14.01 14.45 14.41 15.58 18.22 21.56

25 1.00 10.77 13.40 18.19 23.37 29.83 36.36 43.72 52.88 62.12 71.34 80.01

26 7.65 7.64 10.46 14.46 19.43 25.48 32.38 41.31 51.87 62.45 72.35 81.70

27 6.18 10.18 11.41 14.45 18.02 23.26 28.93 37.03 46.11 56.42 65.94 75.51

28 0.00 0.00 0.00 1.82 5.04 9.13 13.29 17.98 23.25 30.25 39.83 50.31

29 0.62 10.00 13.22 18.06 23.83 30.60 38.01 46.70 56.03 66.91 76.19 85.41

30 0.00 0.51 3.18 6.72 11.30 16.63 24.29 32.59 41.30 51.84 61.44 71.46

31 0.00 10.93 12.12 15.22 18.44 23.24 29.09 35.78 43.19 51.87 61.15 70.71

32 2.28 3.39 5.36 8.06 12.41 18.24 24.51 30.92 37.03 42.72 47.66 54.46

33 2.28 10.69 13.66 17.64 22.43 28.51 35.20 43.48 53.24 62.34 71.45 80.10

34 0.00 0.18 2.76 6.30 11.21 16.40 22.87 31.05 40.38 51.46 61.80 70.88

35 5.63 8.57 11.42 15.16 19.30 23.87 29.72 36.27 43.93 49.57 54.49 59.29

36 4.25 5.75 7.53 9.75 12.94 17.32 22.39 28.13 32.07 32.23 33.32 36.28


catalyst

215

Figure 6.3. CH4 conversion values for each experiment and temperature.

Feed composition has also an important influence over the parameters analyzed. Methane

conversion has a great dependence on the inlet methane flow as could be seen in Figure 6.3,

where the higher values of methane conversion (the orange and red region) appear from the

experiment 24 to the 36, which are those with a lower quantity of methane in the feed (Table

6.1). As can be expected, an increase in CO2, H2O and O2 feed content yields a higher CH4

conversion. This effect is more noticeable for the O2 and H2O feed flow. Comparing each

experiment with high O2 feed flow with the one with a similar feed composition but a low O2

feed flow (Exp. 1-Exp. 2; Exp.3-Exp. 4; etc) there is a clearly higher value of methane

conversion for those experiments where a higher quantity of O2 was added. The same trend can

be observed comparing the experiments performed with a higher H2O quantity in the feed with

those with a lower quantity (Exp. 1-Exp 3; Exp. 2-Exp. 4; etc). In the case of CO2, it is not so

easy to correlate the increase in the CO2 feed flow with the CH4 conversion, despite it is

another co-reactant and it can be forecasted that an increase in the CO2 feed will favour the dry

reforming reaction and thereby will increase the global CH4 conversion of the tri-reforming

process. However, when comparing results obtained for experiments in which we change the

CO2 feed flow and there is a high volume of water in the feed (Exp. 1-Exp. 5-Exp. 9; Exp. 2-

Exp. 6-Exp. 10; etc) there is not a clear correlation between CO2 feed flow and methane

conversion, which can be seen when comparing experiments with different CO2 feed flow and

low volume of water (Exp. 3-Exp. 7-Exp. 11; Exp. 4-Exp. 8-Exp 12; etc).

680 710 740 770 800 830 860 890 920 950 980 1010

36

34

32

30

28

26

24

22

20

18

16

14

12

10

8

6

4

2

Temperature (K)

Ex

per

imen

t n

um

ber

15

30

45

60

75

Methane conversion (%)

Chapter 6

216

Table 6.4. CO2 conversion values obtained for each experiment.

Exp. Temperature

680 710 740 770 800 830 860 890 920 950 980 1010

1 8.62 -5.06 4.32 -1.05 -1.10 6.28 12.49 16.79 25.42 31.12 39.92 48.74

2 5.94 6.81 7.77 6.22 3.87 4.53 6.63 7.98 9.86 12.90 14.57 31.41

3 -0.28 -2.29 -1.41 3.15 9.35 12.19 17.08 22.41 27.01 32.01 37.51 43.89

4 0.85 1.55 3.30 6.74 10.23 14.24 17.82 21.01 24.95 31.43 39.21 43.84

5 4.44 -7.30 -10.4 -12.8 -12.2 -10.4 -13.2 -7.28 2.71 14.92 30.73 38.25

6 -1.00 -4.07 -5.54 -4.90 -3.88 -0.15 4.88 15.14 26.34 35.09 42.32 47.88

7 -0.43 -7.51 -8.94 -7.14 -1.75 4.70 13.07 20.29 26.86 29.88 36.24 43.74

8 12.04 8.70 7.64 8.82 9.43 13.89 20.15 22.59 26.92 32.59 36.35 43.33

9 1.92 -3.51 -6.04 -25.1 -36.3 -40.5 -22.9 -13.0 0.95 17.45 43.44 68.44

10 -1.45 -11.5 -21.9 -30.8 -37.4 -39.5 -32.8 -19.7 -3.23 14.54 32.78 50.97

11 4.66 -15.8 -18.1 -17.3 -10.8 -2.41 6.25 10.86 11.61 14.75 21.22 25.20

12 2.23 0.08 1.60 5.62 13.00 20.58 23.76 23.59 29.52 35.78 44.53 52.47

13 2.52 -4.64 -4.49 -3.73 -6.77 -5.94 -4.82 4.47 15.03 24.13 34.73 46.29

14 0.24 -1.26 -1.39 -1.86 -0.89 -1.06 2.71 6.77 12.69 19.83 25.47 29.62

15 11.48 4.36 4.75 6.28 9.07 13.05 17.86 24.37 31.46 40.15 48.61 58.23

16 10.80 8.48 9.72 12.83 17.24 22.90 27.92 31.76 34.33 39.21 46.96 53.68

17 5.32 -7.53 -10.9 -15.3 -15.6 -14.1 -8.85 -3.27 2.18 8.03 19.33 39.05

18 4.51 3.22 1.53 -0.70 -1.50 0.57 4.51 10.89 18.86 28.30 37.99 45.95

19 0.26 -2.72 -0.53 3.06 8.54 14.92 21.20 26.80 29.51 32.29 37.65 43.49

20 5.32 4.16 5.18 8.09 13.09 19.32 24.56 28.14 31.81 35.50 42.48 50.19

21 10.08 -20.2 -28.5 -36.5 -41.5 -43.2 -40.2 -33.6 -21.1 -1.83 20.52 44.57

22 -216 -242 -240 -235 -225 -211 -193 -170 -145 -114 -83.5 -49.2

23 4.85 -19.9 -28.6 -26.6 -25.0 -19.2 -8.18 4.11 21.46 30.03 40.26 46.14

24 16.41 0.54 0.06 3.47 10.12 19.15 25.34 27.60 30.60 37.55 45.40 53.75

25 8.89 1.24 -0.24 -0.49 0.77 2.90 5.92 10.19 16.48 23.25 29.68 36.81

26 11.59 8.52 5.30 3.73 2.85 3.85 6.40 10.46 17.16 23.97 31.43 38.57

27 6.09 1.24 1.50 3.85 6.76 11.10 16.63 24.23 32.68 42.26 50.68 59.24

28 6.55 4.69 4.92 5.42 6.80 8.58 12.01 15.88 20.83 26.71 36.26 45.87

29 6.88 -5.60 -9.30 -12.0 -13.4 -12.4 -9.49 -3.23 4.40 14.82 24.93 33.64

30 2.72 -0.18 -2.54 -4.62 -6.81 -7.00 -6.14 -3.73 5.07 15.15 25.04 35.69

31 6.96 0.54 -1.11 -0.21 1.23 5.48 11.31 18.64 26.66 38.47 48.89 59.07

32 6.85 3.04 3.57 6.29 11.52 19.03 26.76 36.09 43.38 52.00 59.58 67.16

33 8.34 -18.0 -28.8 -39.5 -47.7 -53.6 -56.5 -54.0 -46.3 -32.8 -19.7 -5.08

34 0.36 -5.03 -13.88 -27.81 -33.99 -41.40 -42.97 -43.71 -34.32 -19.33 -2.61 14.95

35 7.13 -20.53 -24.08 -23.15 -17.72 -9.13 2.24 15.68 29.46 42.34 53.36 63.86

36 -0.07 -2.27 -2.29 0.28 5.28 15.50 26.77 37.09 45.17 49.99 54.76 62.62


catalyst

217

Figure 6.4. CO2 conversion values for each experiment and temperature.

This different behaviour is probably related with the methane availability. CO2, H2O and

O2 compete for the methane present in the reaction environment, following the three different

reactions that form the tri-reforming process. It has been reported [25] that water and oxygen

react with methane preferentially over CO2, which implies that when CO2 competes for the

methane with a high feed flow of H2O, an increase in the carbon dioxide amount in the feed

does not affect to the methane conversion. Feed composition has also a remarkable effect over

the CO2 conversion. As has been commented previously, in the considered conditions the water

gas shift equilibrium plays an important role in the reforming process, and this role depends

not only on the temperature but also on the reactants presence. Those experiments performed

with a low feed concentration of CO2 yielded the lowest values of CO2 conversion (green and

blue regions in Figure 6.4), effect remarked when a high volume of water is fed (Exp. 9; Exp.

10; Exp. 21; etc). A high water concentration and a low carbon dioxide concentration coupled

with low temperature values are the most favourable conditions for the water gas shift reaction,

what causes a very low CO2 conversion or even a net CO2 production. However, at high

temperature there is not a clear correlation between the CO2 conversion and the feed

composition, probably due to the concurrence of two simultaneous reactions, the dry reforming

reaction and the reverse water gas shift, both increasing the CO2 conversion at high

temperature but with different co-reactants. As a general trend, it could be observed in Figure

6.4 how the regions with a higher CO2 conversion (the red and orange coloured ones) are

680 710 740 770 800 830 860 890 920 950 980 1010

36

34

32

30

28

26

24

22

20

18

16

14

12

10

8

6

4

2

Temperature (K)

Ex

per

imen

t n

um

ber

-30.0

-15.0

0

15.0

30.0

50.0

CO2 conversion (%)

Chapter 6

218

slightly bigger for the experiments with a lower CH4 feed flow and for those where water was

added in a small quantity. H2/CO molar ratio of the synthesis gas obtained for each experiment

is also influenced for the feed gas composition. At low temperatures very high H2/CO molar

ratio values were obtained for all the experiments except in the case of those were the lowest

quantity of water was added, behaviour related with the concurrence of the water gas shift

reaction as has been commented previously. At high temperatures it can be observed the

H2/CO molar ratio is higher in those experiments were more water and oxygen is added and

lower in those where more methane and carbon dioxide were added. Taking into account the

stoichiometry of the three different reactions that form the tri-reforming process, it is

reasonable to expect higher values of H2/CO molar ratio in those experiments were the steam

reforming and the partial oxidation are predominant in the tri-reforming compared to those

were dry reforming is predominant. The influence of methane in the H2/CO molar ratio is

probably related with the already mentioned competition between water, carbon dioxide and

oxygen for the available methane in the reaction environment, where the reaction of methane

with CO2 is thermodynamically less favoured and therefore an increase in the quantity of

methane mitigates this adverse effect.

Figure 6.5. H2/CO molar ratio values for each experiment and temperature.

680 710 740 770 800 830 860 890 920 950 980 1010

36

34

32

30

28

26

24

22

20

18

16

14

12

10

8

6

4

2

Temperature (K)

Ex

per

imen

t n

um

ber

1.00

1.50

2.00

2.50

3.00

4.00

H2/CO molar ratio


catalyst

219

Table 6.5. H2/CO molar ratio values obtained for each experiment.

Exp. Temperature

680 710 740 770 800 830 860 890 920 950 980 1010

1 10.02 7.90 5.92 4.79 4.49 3.35 2.71 2.25 2.07 1.95 1.82 1.74

2 - - - 1.05 0.94 1.28 3.31 2.71 2.38 2.25 2.05 1.89

3 - - 2.26 1.81 1.64 1.56 1.51 1.46 1.37 1.28 1.27 1.32

4 5.28 2.82 2.05 1.67 1.60 1.53 1.48 1.36 1.26 1.21 1.22 1.33

5 37.21 13.46 9.82 6.98 5.18 4.17 4.02 3.37 2.87 2.50 2.22 2.18

6 21.53 12.95 8.71 5.65 4.38 3.41 2.88 2.43 2.15 2.05 1.99 2.03

7 - 9.11 4.77 3.70 2.86 2.40 2.14 2.06 1.96 1.85 1.70 1.66

8 - 27.57 9.20 5.38 4.27 3.19 2.60 2.52 2.35 2.06 1.91 1.79

9 - 7.23 3.13 10.56 6.46 6.79 3.96 3.48 3.07 2.77 2.45 2.68

10 - - 34.76 18.09 11.30 7.60 5.41 4.17 3.47 3.07 2.82 2.62

11 - 35.41 10.06 6.25 4.34 3.59 3.25 3.19 3.09 2.61 2.37 2.57

12 - 20.50 6.85 4.13 3.31 3.20 3.19 2.99 2.31 2.03 1.91 1.87

13 - 9.03 5.58 3.97 5.14 4.10 3.61 2.50 2.07 1.92 1.77 1.66

14 14.45 9.36 6.38 5.29 4.01 3.71 2.98 2.62 2.34 2.12 2.04 2.01

15 - 9.53 4.75 3.05 2.37 1.98 1.76 1.58 1.50 1.41 1.35 1.31

16 12.48 4.08 2.55 1.87 1.56 1.44 1.37 1.32 1.25 1.17 1.12 1.12

17 - 30.45 10.08 9.48 6.39 4.79 3.74 3.18 2.88 2.71 2.47 2.15

18 - 109.80 19.67 10.32 7.01 5.02 3.83 3.13 2.69 2.39 2.20 2.13

19 - 7.13 3.82 2.75 2.23 1.96 1.83 1.73 1.74 1.64 1.55 1.57

20 - 9.25 4.14 2.73 2.19 1.96 1.85 1.75 1.63 1.46 1.38 1.36

21 - - 57.38 21.55 12.38 8.11 5.83 4.61 3.71 3.07 2.68 2.44

22 - 9.53 4.75 3.05 2.37 1.98 1.76 1.58 1.50 1.41 1.35 1.31

23 - - 52.56 11.57 6.61 4.62 3.51 2.97 2.61 2.54 2.35 2.21

24 - 59.52 8.22 4.58 3.45 3.00 2.82 2.71 2.36 2.07 1.96 1.92

25 - 35.83 12.12 7.49 5.30 4.11 3.32 2.78 2.37 2.11 1.95 1.82

26 - - 22.98 11.35 7.53 5.28 3.90 3.14 2.56 2.26 2.06 1.91

27 - 7.68 3.59 2.43 1.99 1.70 1.51 1.39 1.31 1.24 1.21 1.19

28 - - - 7.65 3.98 2.98 2.22 1.90 1.66 1.54 1.38 1.30

29 - - 27.35 12.71 8.13 5.74 4.36 3.51 2.98 2.57 2.33 2.23

30 - - 129.85 19.89 10.67 7.16 5.40 4.37 3.32 2.77 2.44 2.22

31 - - 8.85 4.95 3.51 2.76 2.32 2.03 1.86 1.67 1.59 1.55

32 - 23.54 5.42 3.12 2.20 1.86 1.66 1.52 1.48 1.40 1.33 1.33

33 - - - 71.99 22.41 12.86 8.67 6.27 4.83 3.96 3.45 3.12

34 - - - - 41.38 17.12 10.01 6.92 5.03 3.91 3.34 2.99

35 - - 18.58 8.16 5.05 3.71 3.04 2.60 2.40 2.23 2.11 2.02

36 - - 32.49 6.85 4.21 2.99 2.58 2.41 2.27 2.07 1.93 1.85

Chapter 6

220

6.3.3. Kinetic analysis.

In order to obtain a kinetic expression for the tri-reforming process that fit the

experimental data obtained, we choose one kinetic expression for each reaction present in the

process. For steam reforming and dry reforming we consider the expressions pointed by J. Wei

and E. Iglesia [12] for this reactions. They found that the only kinetically relevant step in these

reactions is the activation of the C-H bond considering the forward reaction rate, with no effect

of identity or concentration of the coreactants. So, the expressions considered for these

reactions were:

(Equation 6.11)

(Equation 6.12)

where and are the CH4 net reaction rate (mol s-1

), and are the kinetic

constants for each reaction, is the partial pressure of species j (in kPa) and and are

the respective equilibrium constants. The value of these equilibrium constants depend on the

temperature, so in order to be determined we use the following equations [26]:

(Equation 6.13)

(Equation 6.14)

As previously commented, partial oxidation of methane, the other main reaction of the tri-

reforming process, was considered at equilibrium in the selected conditions, due to the fact that

almost no O2 was detected in the effluent gas. So, the expression considered for this reaction

was:

(Equation 6.15)

However, the water gas shift reaction was considered to have a key role in the catalytic

data observed, so a kinetic expression for this reaction was added, based on a previous work of

our group where different kinetic expressions for the water gas shift reaction were evaluated

[13]:


catalyst

221

(Equation 6.16)

In all the cases the kinetic constants were expressed using the Arrhenius equation, so a

pre-exponential factor and an activation energy were calculated for each expression.

(Equation 6.17)

Where is the pre-exponential factor for reaction i, is the activation energy for

reaction i, R is the universal gas constant and T is the temperature.

In order to estimate the parameter values, the Marquardt-Levenberg algorithm was

applied, using the T-test for the determination of the statistic significance of each parameter

and the F-test for the determination of the statistic significance of the global model. In the first

place, we evaluate for each temperature the statistical significance of the kinetic constant,

calculated using the previously commented algorithm. Table 6.6 shows the T values obtained

for each parameter divided by the T-test value. When this ratio is higher than 1 the value for

the parameter considered is statistically significant. It can be seen that k1 and k3 are statistically

significant in almost all the temperature range studied except for some of the lower

temperatures, where specially k3 shows some data which are not. However, k2 have a different

behaviour, with a wide range in the low and medium temperature zone where it is not

statistically significant. The experimental data obtained for the lowest temperatures (680-710

K) showed very low conversions, so the experimental and analytical error could have a great

influence in the results, what could be the reason of the not statistical significance of the kinetic

constants. However, for the k2 results we should also consider the already commented

temperature dependence for the catalytic results, which is in agreement with the required high

temperature in order to have statistical significance. In this way, CO2 global consumption was

not observed in the low temperature range, obtaining higher CO2 molar flow in the effluent

compared to that fed to the system, probably due to the high contribution of the water gas shift

reaction to the global catalytic process at low temperature, what increase the CO2 concentration

as a result of it stoichiometry (Equation 6.10).

Chapter 6

222

Table 6.6. Statistical significance for each kinetic constant vs temperature.

Temperature k1 k2 k3

680 2.31/1.96 1.48/1.96 0.00/1.96

710 2.13/1.96 2.68/1.96 0.00/1.96

740 1.32/1.96 3.18/1.96 1.44/1.96

770 2.92/1.96 0.86/1.96 2.90/1.96

800 7.66/1.96 0.01/1.96 3.91/1.96

830 11.05/1.96 0.00/1.96 4.22/1.96

860 18.63/1.96 0.01/1.96 4.81/1.96

890 25.41/1.96 0.03/1.96 6.72/1.96

920 2.25·1017

/1.96 8.92·1013

/1.96 8.91·1013

/1.96

950 2.41·1021

/1.96 7.53/1.96 7.98/1.96

980 7.17/1.96 3.68/1.96 4.73·1021

/1.96

1010 1.83·1018

/1.96 1.67·1018

/1.96 6.50/1.96

Table 6.7 shows the kinetic data for the adjustment considering all the temperatures

simultaneously. It can be seen how all the kinetic parameters have statistical significance, as

their T value are higher than the T-test. Regarding the activation energy calculated, we

obtained very close values for the dry reforming and steam reforming reactions, 74.72 and

77.82 kJ mol-1

respectively, values lower than that calculated by density-functional theory for

the C-H bond activation (85-100 kJ mol-1

[27, 28]) but close to that calculated by embedding

methods (72 kJ mol-1

[29]). These values are also in accordance with the wide range of values

reported in the literature for the steam reforming and dry reforming reactions (74-118 kJ mol-1

[30-33]. Regarding the activation energy for the water gas shift reaction, a value of 54.26 kJ

mol-1

was obtained, value inside the range reported by Newsome (48.9-89.2 kJ mol-1

) for the

water gas shift reaction [34]. The F-test also indicates that the model adjusts the experimental

data with statistical significance. The average error comparing the modelled and experimental

data for the CH4 and CO2 molar flow in the effluent gas are also depicted in Table 6.7. An

average error of 10.8% for the CH4 molar flow and 19.4% for the CO2 was obtained.


catalyst

223

Figure 6 shows the fit obtained for the CH4 and CO2 molar flow, with the values obtained

experimentally and those obtained by the model. It can be seen a reasonable accuracy between

both values, especially in the case of the methane molar flow. For the carbon dioxide it can be

observed that, especially for the lowest values of CO2 molar flow there is a great difference

between the experimental and modelled molar flows, being the modelled clearly lower than the

Kin

etic

eq

ua

tio

n

Pa

ram

eter

s T

/T-t

est

F/F

-tes

t

CH

4 e

rro

r (%

) C

O2 e

rro

r (%

)

SR

1𝑜

= 8

5.7

7 m

ol

s-1

kP

a-1

3.4

6·1

01

4/1

.96

70

36

.95

/2.1

0

.33

10

.8

19

.4

𝑎1 =

74

.72

kJ

mo

l-1

47

.43

/1.9

6

DR

2𝑜

= 7

0.9

9 m

ol

s-1

kP

a-1

4.9

9·1

01

3/1

.96

𝑎2 =

77

.82

kJ

mo

l-1

26

.28

/1.9

6

WG

S

3𝑜

= 1

49

.92

mo

l s-

1 k

Pa-1

3

.11

/1.9

6

𝑎3 =

54

.26

kJ

mo

l-1

5.6

9·1

01

3/1

.96

Ta

ble

6.7

. K

inet

ic m

od

el r

esu

lts.

Chapter 6

224

experimental ones, which could indicate a lower contribution of the water gas shift reaction to

the global process.

Figure 6.6. Comparison between experimental and modelled molar flows for the 432

experiment adjustment a) CH4 molar flows b) CO2 molar flows.

The parameters commented were obtained considering the 432 experiments carried out.

In order to better fit the experimental and modelled data, a second adjust was performed with

356 experiments, excluding those where the CO and H2O molar flow in the effluent gas were 0,

as they induced a large error in the system. The new kinetic parameters can be observed in

Table 6.8. In this case, the pre-exponential factor and the activation energy values for the steam

reforming and dry reforming equation are very close to those observed in the previous adjust.

However, a significantly lower pre-exponential factor and slightly higher activation energy

(68.67 kJ mol-1

) were obtained for the water gas shift equation, what indicates a lower reaction

rate of the water gas shift, what is in agreement with the exclusion of the data where the H2O

and CO molar flow in the effluent were 0. The T-test shows that all the kinetic parameters are

statistically significant, which was also confirmed for the global model by the F-test. The CH4

error obtained was slightly higher than that of the previous model, while the , CO2 error and

the average error were slightly lower. The correlation between the calculated and experimental

data for the methane and carbon dioxide molar flow for this adjustment can be seen in Figure

6.7. The plots obtained are very close to that of Figure 6.6, but a slightly better adjustment

could be seen for Figure 6.7, primarily in the region of low molar flow.

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

b)a)

Mo

del

led

mo

lar

flo

w (

mo

l h

-1)

Experimental molar flow (mol h-1

)

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

Mo

del

led

mo

lar

flo

w (

mo

l h

-1)


)


catalyst

225

Kin

etic

eq

ua

tio

n

Pa

ram

eter

s T

/T-t

est

F/F

-tes

t

CH

4 e

rro

r (%

) C

O2 e

rro

r (%

)

SR

=

69

.40

mo

l s-1

kP

a-1

8.1

6·1

013

/1.9

6

563

9.8

2/2

.1

0.2

8

10

.8

14

.4

= 7

4.2

9 k

J m

ol-1

2

01.8

6/1

.96

DR

=

87

.06

mo

l s-1

kP

a-1

20

.44/

1.9

6

= 7

7.7

3 k

J m

ol-1

1

.92

·10

13/1

.96

WG

S

= 2

6.7

1 m

ol s

-1 k

Pa-1

6

.00

/1.9

6

= 6

8.6

7 k

J m

ol-1

2

.52

·10

13/1

.96

Ta

ble

6.8

. S

eco

nd

kin

etic

mod

el

resu

lts.

Chapter 6

226

Figure 6.7. Comparison between experimental and modelled molar flows for the 356

experiment adjustment a) CH4 molar flows b) CO2 molar flows.

6.4. CONCLUSIONS

The catalytic behaviour of a Ni- -SiC catalyst has been evaluated at different

temperatures and feed composition. The catalytic results showed a clear dependence of the

methane conversion on the temperature, with higher values of conversion at higher

temperature, probably due to the high endothermicity of the steam reforming and dry

reforming. CO2 conversion also showed a dependence on the temperature, as the CO2 flow in

the effluent gas is higher than the CO2 flow in the feed at low temperatures, due to the

concurrence of the water gas shift reaction, which is responsible of the formation of CO2.

Temperature also affects the H2/CO molar ratio, with lower values for this last parameter at

higher temperatures.

The feed composition has a great importance in the CO2 conversion and H2/CO molar

ratio, as CO2 conversion is decreased when more H2O and O2 is present in the reaction media,

due to the competition for the methane available and the decrease in the dry reforming

contribution to the global tri-reforming process, as CO2 reaction with methane is

thermodynamically less favoured, which also increases the H2/CO molar ratio.

In the last part of this work we modelled the catalytic data obtained considering kinetic

equations for the steam reforming, dry reforming and water gas shift. The activation energies

obtained for each reaction were consistent with the values reported in the literature. It was

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

b)a)M

od

elle

d m

ola

r fl

ow

(m

ol

h-1

)


)

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

Mo

del

led

mo

lar

flo

w (

mo

l h

-1)


)


catalyst

227

observed a good accuracy between the modelled and the experimental data, with average errors

about 10 % for the methane molar flow and 14% for the carbon dioxide molar flow.

Chapter 6

228

6.5 REFERENCES

[1] P.M. Fearnside, Climatic change. 46 (2000) 115-158.

[2] L. Bernstein, P. Bosch, O. Canziani, Z. Chen, R. Christ, O. Davidson, W. Hare, S. Huq, D.

Karoly, V. Kattsov, Retrieved March. 20 (2007) 2011.

[3] M. Meinshausen, N. Meinshausen, W. Hare, S.C. Raper, K. Frieler, R. Knutti, D.J. Frame,

M.R. Allen, Nature. 458 (2009) 1158-1162.

[4] T.L. Frölicher, M. Winton, J.L. Sarmiento, Nature Climate Change (2013).

[5] X.E. Verykios, Int. J. Hydrogen Energy. 28 (2003) 1045-1063.

[6] J. Guo, H. Lou, H. Zhao, D. Chai, X. Zheng, Applied Catalysis A: General. 273 (2004) 75-

82.

[7] N. Laosiripojana, S. Assabumrungrat, Applied Catalysis B: Environmental. 60 (2005) 107-

116.

[8] A.I. Tsyganok, T. Tsunoda, S. Hamakawa, K. Suzuki, K. Takehira, T. Hayakawa, J. Catal.

213 (2003) 191-203.

[9] Y.-A. Zhu, D. Chen, X.-G. Zhou, W.-K. Yuan, Catal. Today. 148 (2009) 260-267.

[10] I.O. Costilla, M.D. Sánchez, C.E. Gigola, Applied Catalysis A: General. 478 (2014) 38-

44.

[11] J.M. García-Vargas, J.L. Valverde, J. Díez, P. Sánchez, F. Dorado, Applied Catalysis B:

Environmental. 148-149 (2014) 322-329.

[12] J. Wei, E. Iglesia, J. Catal. 224 (2004) 370-383.

[13] A. De la Osa, A. De Lucas, A. Romero, J. Valverde, P. Sánchez, Int. J. Hydrogen Energy.

36 (2011) 9673-9684.

[14] J.L. Sotelo, M.A. Uguina, J.L. Valverde, D.P. Serrano, Industrial & Engineering

Chemistry Research. 32 (1993) 2548-2554.


catalyst

229

[15] J.F. Rodríguez, A. de Lucas, J.R. Leal, J.L. Valverde, Industrial & Engineering Chemistry

Research. 41 (2002) 3019-3027.

[16] J. . alverde . . ucas . Carmona . on le J. . odr gue , Chem. Eng. Sci. 59

(2004) 71-79.

[17] B.P.F. William H. Press, Saul A. Teukolsky, William T. Vetterling, Numerical Recipes in

Fortran 77: The Art of Scientific Computing, Cambridge University Press; 2 edition

(September 25, 1992), 1992.

[18] G.F. Froment, K.B. Bischoff, J. De Wilde, Chemical reactor analysis and design, Wiley

New York, 1990.

[19] D.W. Marquardt, Journal of the Society for Industrial & Applied Mathematics. 11 (1963)

431-441.

[20] J. Valverde, L. Zumalacárregui, G. Suris, Chem. Eng. Res. Des. 75 (1997) 784-791.

[21] J.A. Díaz, M. Calvo-Serrano, A.R. de la Osa, A.M. García-Minguillán, A. Romero, A.

Giroir-Fendler, J.L. Valverde, Applied Catalysis A: General. 475 (2014) 82-89.




(1990) 2663-2669.


[26] K. Hou, R. Hughes, Chem. Eng. J. 82 (2001) 311-328.

[27] I. Ciobica, F. Frechard, R. Van Santen, A. Kleyn, J. Hafner, The Journal of Physical

Chemistry B. 104 (2000) 3364-3369.

[28] P. Kratzer, B. Hammer, J.K. No/rskov, The Journal of Chemical Physics. 105 (1996)

5595-5604.

[29] H. Yang, J.L. Whitten, The Journal of Chemical Physics. 96 (1992) 5529-5537.

Chapter 6

230

[30] S. Wang, G.Q. Lu, Energy & fuels. 12 (1998) 248-256.

[31] O. Tokunaga, S. Ogasawara, React. Kinet. Catal. Lett. 39 (1989) 69-74.

[32] T. Horiuchi, K. Sakuma, T. Fukui, Y. Kubo, T. Osaki, T. Mori, Applied Catalysis A:

General. 144 (1996) 111-120.

33 P. erreira- paricio . uerrero- ui . odr gue -Ramos, Applied Catalysis A:

General. 170 (1998) 177-187.

[34] D.S. Newsome, Catalysis Reviews Science and Engineering. 21 (1980) 275-318.

CHAPTER 7 General conclusions and recommendations

7. 1. GENERAL CONCLUSIONS

7.2. RECOMMENDATIONS

General conclusions and recommendations

233

This chapter lists the main conclusions derived from the research performed in this PhD

work. In addition, some recommendations are suggested for further studies.

7.1. GENERAL CONCLUSIONS

The obtained results from the present research work support the following main

conclusions:

The material selected as support has a great influence on the catalytic behaviour of

catalysts applied for the tri-reforming of methane. Nickel aluminates were obtained

for the Ni/Al2O3 catalyst, what decreased the quantity of Ni0 species, yielding a lower

catalytic activity. The best catalytic behaviour was observed for Ni/CeO2 and Ni/-

SiC catalysts.

The catalytic behaviour of Ni/YSZ catalysts were affected by the atmosphere

surrounding the catalyst during the calcination step. When this process was carried out

under a lean oxygen atmosphere, Ni/YSZ showed a higher quantity of oxygen

vacancies in the support surface, what increased the catalytic activity and stability.

CeO2 supported catalysts showed larger Ni particles and a higher basicity, measured

by CO2-TPD, compared to -SiC supported catalysts. A higher metal support

interaction was observed for these latter catalysts, what improved the catalyst stability.

Nickel nitrate and nickel acetate yielded the smaller Ni particles and a higher metal

support interaction when used as nickel precursors, compared to nickel chloride and

nickel citrate. Catalysts prepared using nickel chloride showed the highest

deactivation rate, probably due to the presence of chloride ions on the catalyst surface

and the higher nickel particle size. Nickel nitrate and nickel acetate yielded the

catalysts with a higher catalytic activity and stability.

The study performed using a factorial design of experiments showed that, under the

considered conditions, the effect of feed composition on the methane conversion was

not statistically significant. However, the effect of the feed composition on the H2/CO

molar ratio was statistically significant. A higher water or oxygen volume flow

increased the H2/CO molar ratio of the synthesis gas obtained, while a higher methane

or carbon dioxide volume flow decreased the H2/CO molar ratio.

When Na or K were added as promoters to Ni/-SiC catalysts, -SiC suffered a great

oxidation during the calcination process, what yields -cristobalite, one of the phases

Chapter 7

234

of SiO2, as product of the oxidation. When a high Ca load was added to a Ni/-SiC

catalyst, it was observed a great oxidation of the support, yielding quartz.

The influence of Mg as promoter was also evaluated. It was observed an increase in

the activity and stability of the catalysts, decreasing the Ni particle size and increasing

its basicity. It was observed that the higher the Mg load, the smaller the nickel

particles and the less reducible the catalyst is, shifting the reduction peak observed in

TPR experiments towards higher temperatures. This behaviour is probably related to

the formation of a NiO-MgO solid solution. Catalysts with higher Mg load were more

stable and yielded a synthesis gas with a higher H2/CO molar ratio.

Impregnation order between Ni and Mg has a key role on the catalytic behaviour.

Those catalysts where Ni was impregnated in first place showed the worst catalytic

results, probably due to a poor interaction between Ni and Mg, the partial blockage of

Ni particles by Mg ones and the formation of Ni2Si during the reaction. When Mg was

impregnated in first place, the amount of coke observed after reaction was the lowest.

Catalyst prepared by simultaneous impregnation of Ni and Mg showed the best

catalytic activity, probably due to the high interaction between Ni and Mg.

Temperature showed a great influence both in methane and carbon dioxide conversion

on the tri-reforming process. At low temperatures, it was obtained a net production of

CO2, due to the high extend of the water gas shift reaction, while high CO2

conversions were observed at high temperature, where the endothermic dry reforming

gains importance and the exothermic water gas shift reaction is less favoured.

It was developed a kinetic model in order to fit the experimental data obtained,

considering the steam reforming, dry reforming and water gas shift as the kinetically

relevant reactions. Activation energies values obtained by the model where in

agreement with those reported in the literature, showing the modelled data a good

correlation with the experimental ones.

7.2. RECOMMENDATIONS

The following proposals can be stated in order to complete and extent this research work:

To evaluate new metals and different catalyst preparation methods.

To analyze the influence of different poisons, especially those usually present in

natural gas, like H2S.

General conclusions and recommendations

235

To perform tri-reforming experiments using actual natural gas or biogas as feed.

To evaluate the pressure influence in both methane conversion and H2/CO molar ratio

of the synthesis gas obtained.

To study the catalyst stability in long term experiments, measuring the coke quantity

obtained and analyzing the deactivation process.

To develop a pilot plant in order to perform experiments in a configuration closer to

the industrial one.

To perform a simulation of a tri-reforming industrial plant using the model calculated

in the present work in order to analyze its behaviour under different conditions.

LIST OF PUBLICATIONS AND CONFERENCES

List of publications and conferences

239

Publications

Precursor influence and catalytic behaviour of Ni/CeO2 and Ni/SiC catalysts for the

tri-reforming process. J. M. García-Vargas, J. L. Valverde, A. de Lucas-Consuegra, B.

Gómez-Monedero, P. Sánchez, F. Dorado. Applied Catalysis A: General 431-432

(2012) 49-56.

Methane tri-reforming over a Ni/-SiC-based catalyst: Optimizing the feedstock

composition. J. M. García-Vargas, J. L. Valverde, A. de Lucas-Consuegra, B. Gómez-

Monedero, F. Dorado, P. Sánchez. International Journal of Hydrogen Energy 38

(2013) 4524-4532.

Influence of alkaline and alkaline-earth cocations on the performance of Ni/-SiC

catalysts in the methane tri-reforming reaction. J. M. García-Vargas, J. L. Valverde, J.

Díez, P. Sánchez, F. Dorado. Applied Catalysis B: Environmental 148–149 (2014)

322–329.

Influence of the support on the catalytic behaviour of Ni catalysts for the dry

reforming reaction and the tri-reforming process. J. M. García-Vargas, J. L. Valverde,

F. Dorado, P. Sánchez. Journal of Molecular Catalysis A: Chemical 395 (2014)

108–116.

Preparation of Ni-Mg/-SiC catalysts for the methane tri-reforming: effect of the

order of metal impregnation. J. M. García-Vargas, J. L. Valverde, J. Díez, P. Sánchez,

F. Dorado. Applied Catalysis B: Environmental 164 (2015) 316-323.

Catalytic and kinetic analysis of methane tri-reforming over a Ni-Mg/-SiC catalyst.

J. M. García-Vargas, J. L. Valverde, J. Díez, F. Dorado, P. Sánchez (Submitted to

Chemical Engineering Journal).

List of publications and conferences

240

Conferences

SECAT 11. Zaragoza (Spain), June 2011. Estudio de la influencia del precursor en

la preparación de catalizadores Ni/SiC aplicados al proceso de tri-reformado de

metano. J. M. Garcia-Vargas, F. Dorado, B. Gomez-Monedero, P. Sánchez, J. L.

Valverde. Poster. National.

Europacat X. Glasgow (United Kingdom), August 2011. Precursor influence and

catalytic behaviour of Ni/SiC catalysts for the tri-reforming process. J. M. Garcia-

Vargas, F. Dorado, B. Gomez-Monedero, A. R. de la Osa, P. Sánchez, J. L. Valverde.

Poster. International.

SynFuel2012 Symposium, Munich (Germany), June 2012. Support influence and

catalytic behaviour of Nickel catalysts for the dry reforming and tri-reforming process.

J. M. García-Vargas, J. L. Valverde, A. de Lucas-Consuegra, F. Dorado, P. Sánchez.

Poster. International.

7th International Conference on Environmental Catalysis. Lyon (France),

September 2012. Optimization of the reagents flow in a tri-reforming process over

Ni--SiC catalyst. J. M. García-Vargas, J. L. Valverde, A. de Lucas-Consuegra, B.

Gómez-Monedero, F. Dorado, P. Sánchez. Poster. International.

SECAT 13. Sevilla (Spain), June 2013. Influencia de promotores alcalinos y

alcalinotérreos en catalizadores de Ni/-SiC. J. M. Garcia-Vargas, J. L. Valverde, A.

de Lucas-Consuegra, F. Dorado, B. Gomez-Monedero, P. Sánchez. Poster. National.

E2KW 13. Energy and Environment Knowledge Week. Toledo (Spain),

November 2013. Tri-reforming of methane: Converting CO2 and CH4 into valuable

chemical compounds. J. M. Garcia-Vargas, J. L. Valverde, J. Díez, F. Dorado, P.

Sánchez. Oral presentation. International.

JJ. II. SECAT 14. Malaga (Spain), June 2014. Estudio del método de impregnación

en la preparación de Ni/Mg/SiC para el tri-reformado de metano. J. Díez, J. M.

Garcia-Vargas, J. L. Valverde, P. Sánchez, F. Dorado. Oral presentation. National.

METHANE TRI-REFORMING OVER NICKEL CATALYSTS

Documents

Transcript of METHANE TRI-REFORMING OVER NICKEL CATALYSTS