METHANE TRI-REFORMING OVER NICKEL CATALYSTS

Transcription

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 trireforming 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
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
i
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
37
CATALIZADORES
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
REACTION AND THE TRI-REFORMING PROCESS
ii
DRY
REFORMING
67
Resumen
69
Abstract
70
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
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
2.1. INTRODUCTION
101
2.2. EXPERIMENTAL
103
103
104
105
105
105
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
3.1. INTRODUCTION
127
iii
3.2. EXPERIMENTAL
129
129
129
130
3.2.4. Experimental design
131
133
133
3.3.2. Statistical analysis
133
3.3.3. Influence of the feedstock composition on the H 2/CO molar ratio
137
3.3.4. Optimization of the reaction conditions
141
3.4. CONCLUSIONS
145
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
iv
149
Resumen
151
Abstract
152
4.1. INTRODUCTION
153
4.2. EXPERIMENTAL
155
155
155
156
156
156
164
4.3.3. Influence of the Mg/Ni molar ratio.
166
4.4. CONCLUSIONS
171
4.5. REFERENCES
172
CHAPTER 5: PREPARATION OF Ni-MG/-SIC CATALYSTS FOR THE
METHANE TRI-REFORMING: EFFECT OF THE ORDER OF METAL
175
IMPREGNATION
Resumen
177
Abstract
178
5.1. INTRODUCTION
179
5.2. EXPERIMENTAL
180
180
181
182
183
183
189
5.3.3. Characterization after reaction
194
5.4. CONCLUSIONS
197
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
6.1. INTRODUCTION
205
6.2. EXPERIMENTAL
206
206
206
207
6.2.4. Kinetic analysis
207
209
209
213
220
6.4. CONCLUSIONS
226
v
6.5 REFERENCES
CHAPTER 7: GENERAL CONCLUSIONS AND RECOMMENDATIONS
231
7. 1. GENERAL CONCLUSIONS
233
7.2. RECOMMENDATIONS
234
LIST OF PUBLICATIONS AND CONFERENCES
vi
228
237
Publications
239
Conferences
240
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
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
vii
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. CH 4 ( ) and
CO2 ( ) consumption rates vs. time on stream (left axis), and H 2/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. CH 4 ( ) 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.
85
Table 1.3. Main catalytic results.
87
Figure 1.9. Tri-reforming catalytic activity at 1073 K. Reaction conditions: CH 4
= 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/Al 2O3, 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
viii
1
106
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
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
balance, total flow rate = 100 mL min-1. CH4 (
) and CO2 (
rates vs. time on stream (left axis), and H2/CO molar ratio (
) consumption
) 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
ix
139
141
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 (
axis), and H2/CO molar ratio (
) consumption rates vs. time on stream (left
) 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
CO2 ( ) consumption rates vs. time on stream (left axis), and H 2/CO molar ratio
x
( ) vs. time on stream (right axis).
165
Figure 4.5. TPR profiles for Mg promoted and Ni/SiC catalysts.
166
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
169
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
170
catalysts.
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
185
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 H 2/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 H 2/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
xi
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
211
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) CO 2-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.
xii
226
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 PPII100045-5875).
1
Descripción del trabajo realizado
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
2
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 CO 2 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 CO 2 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.
3
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
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
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
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
Productos ligeros
Productos medios
Productos pesados
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].
4
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
5
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 FischerTropsch (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 (H 2), y
muy a menudo algo de dióxido de carbono (CO 2). 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.
6
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.
7
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.
8
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.
9
 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 CO 2 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
10
= 172,2 kJ/mol
(Ecuación A.1)
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 FischerTropsch 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 CO 2, que según la
ecuación A.9, utiliza mayor cantidad de hidrógeno [12].
11
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 (H 2CO) 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
12
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 H 2, 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 H 2 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 CH 4 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.
13
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 CO 2 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
14
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 H 2S).
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
= -282,6 kJ/mol
(Ecuación A.16)
H2 + O2 ⇄ H2O
= -241,2 kJ/mol
(Ecuación A.17)
15
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
16
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 H 2/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
17
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 CO 2. 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 trireformado 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]
18
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, (H 2O/(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 CH 4: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 H 2O/CH4 de 0 a 0,5. Song y col. [21] analizaron la
influencia de la concentración de oxígeno manteniendo constante la relación molar H 2O/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
19
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 CO 2 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
20
eficiencia del 49%, hallándose que se podría reducir hasta en un 50% las emisiones de CO 2 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
21
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).
22
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.
23
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).
24
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.
25
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
26
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.
27
Tabla A.4. Publicaciones sobre la utilización de SiC como soporte catalítico.
Reacciones
Catalizador
Autores y año
NiS2/-SiC
N. Keller y col.
(2002) [47]
Fe2O3-SiC
P. Nguyen y col.
(2011) [48]
VPO/-SiC
M.J. Ledoux y
col. (2001) [41]
BETA
zeolite/-SiC
G. Winé y col.
(2007) [49]
Ga/SBA15/-SiC
F. Z. El Berrichi y
col. (2008) [50]
TiO2-SiC
Y. Nishida y col.
(2005) [51]
Fe/-SiC
H. M.T. Galvis y
col. (2012) [52]
Co-SiC
A. R. de la Osa y
col. (2011) [53]
Ni/-SiC
S. M. Kim y col.
(2012) [54]
Ni/-SiC
J. M. GarcíaVargas y col.
(2012) [55]
Oxidación de
H2S
Oxidación de
n-butano
Friedel-Crafts
fotocatálisis
Síntesis FT
Reformado
Principales conclusiones
La alta dispersión y la falta de
microporosidad en el soporte
producen una gran actividad y
selectividad.
Un mayor tamaño de poro en el
soporte favorece el acceso de los
reactivos a los centros activos.
Interacción óptima entre VPO y SiC.
SiC mantiene la temperatura del
lecho catalítico uniforme.
Utilizar SiC como soporte permite
un control preciso de la estructura
macroscópica del catalizador, lo que
disminuye los problemas
difusionales.
-SiC mejora la transferencia de
materia y calor, aumentando la
selectividad y estabilidad.
TiO2 soportado sobre-SiC muestra
mayor actividad catalítica que
soportado sobre SiO2 o no
soportado.
Actividad catalítica estable durante
60 h, con una selectividad a C2-C4
mayor del 50%.
-SiC como soporte mejora la
conversión de CO y aumenta
selectividad a hidrocarburos pesados
de interés comparado con Al2O3.
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.
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
28
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
Deshidratación de
alcoholes
Espumas
ZSM-5/-SiC
Y. Jiao y col.
(2012) [56]
Pirólisis catalítica
Al2O3/SiC
S. G. Jeon y col.
(2012) [57]
Hidrogenación de 4carboxybenzaldehido
CDC–SiC
Y. Zhou y col.
(2012) [58]
Fotocatálisis
Grafeno
cubierto de
polvo de SiC
K. Zhu y col.
(2012) [59]
Reformado
CeO2/Pt–SiC
R. Frind y col.
(2012) [60]
Electrólisis de agua
IrO2/SiC-Si
A. V. Nikiforov y
col. (2011) [61]
Síntesis de amoniaco
Ru-C/SiC
Y. Zheng y col.
(2009) [62]
Fischer-Tropsch
Al2O3–SiC
M. Lacroix y col.
(2011) [63]
Principales conclusiones
Mayor rendimiento y
estabilidad gracias a mejor
transferencia de calor y
materia.
Mejor comportamiento
comparado con catalizador
no soportado o soportado
sobre SiC.
El composite mejora la
dispersión de Pd respecto a
soportes tradicionales.
Con el composite se dobla
la actividad fotocatalítica
comparado con SiC.
Los composites con mayor
cantidad de catalizador
mostraron la mayor
estabilidad contra oxidación
y una buena actividad
catalítica.
Mayor actividad debido a
una mejora de las
propiedades superficiales de
IrO2.
El composite preparado
mostró buenas condiciones
como soporte para la
síntesis de amoniaco.
Usando el composite como
soporte se obtuvieron
mayores selectividades a
C5+ que con Al2O3.
29
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 trireformado 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 H 2/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).
30
 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.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).
31
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
32
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.
FIC
H2
FIC
CH4
Reactor
FIC
O2
FIC
Peltier
N2
FIC
Horno
CO2
Saturador
BROOKS
INSTRUMENT
Controlador
de caudal
300
30
300
03 0 0
32
Controlador
temperatura
Campana
Caudalímetro
Microcromatógrafo de
gases
PC-1
Figura B.1. Instalación experimental para los experimentos catalíticos.
33
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: N 2, 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
34
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 trireformado, 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.
35
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.
36
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
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 N 2 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
37
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
38
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, d hkl
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.
39
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 H 2 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 H 2 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.
40
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):
ds 
n d
i
3
i
i
n i d i2
(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.
41
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)
42
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 H 2. 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 H 2 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.
43
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 (NiAl 2O4)
[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
44
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/YSZO2.
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
45
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 H 2/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 trireformado 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.
46
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 CeO 2 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 CeO 2, cuando se
utilizó cloruro o citrato de níquel como precursor, mostraron buenos valores iniciales de
47
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 H 2/CO, se pudo apreciar la influencia que el
soporte tuvo sobre este parámetro, ya que los catalizadores soportados sobre CeO 2 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
48
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 H 2/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 H 2/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 H 2/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 CH 4 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,
49
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 H 2/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 trireformado. 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 H 2/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
50
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].
51
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,
52
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 NiOMgO. 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
53
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 Ni 2Si 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
54
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 watergas 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
55
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 H 2/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.
56
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 H 2/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.
57
 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.
58
 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.
59
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) 139148.
[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).
60
[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) 133137.
[23] H.-t. Jiang, H.-q. Li, Y. Zhang, Journal of Fuel Chemistry and Technology. 35 (2007) 7278.
[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.
61
[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) 567579.
[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.
62
[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, eJournal 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) 152158.
[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.
63
[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) 157166.
[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) 283294.
64
[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) 191200.
[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.
65
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
1.4. 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
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/Al 2O3
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.
69
Chapter 1
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 trireforming process.
70
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 H 2/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
(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

CH4  C + 2H2

H2O + C  CO + H2

ΔHº = -172.2 KJ∙mol-1
(Equation 1.4)
(Equation 1.5)
(Equation 1.6)
71
Chapter 1
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), yttriastabilized 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
72
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. Temperatureprogrammed 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 Ushaped 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 ≤
73
Chapter 1
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 ( d s ), was calculated according to:
ds 
n d
i
3
i
i
n i d i2
(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 Ktype 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 H 2 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% CH 4, 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 CP4900). The methane and carbon dioxide consumption rates were calculated as [inlet molar flow
74
– 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 NiAl 2O4 [16]. The Ni/CeO2 catalyst
showed the typical diffraction peaks related to the cubic lattice of pure CeO 2 (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.
75
76
10
20
30
#
#
40
#
50
2(º)
#
#
#
60
# #
# #
#
#
70
80
90
0
10
Support
Fresh
Reduced
b)
20
*
30
* *
**
*
40
50
2(º)
*
+ +
*
^ ^
*
*
*
*
+
60
*
*
*
70
*
*
*
*
*
*
80
*
*
*
90
*
*
*
Intensity (a.u.)
Intensity (a.u.)
Intensity (a.u.)
0
10
Support
Fresh
Reduced
c)
20
30
º
º
º
^
40
º
50
^
2(º)
+ +
º
º
60
º
º
º
+
70
º
º
º
º
º
80
+
º^
90
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.
0
Support
Fresh
Reduced
a)
Chapter 1
a)
b)
-
-
-
-
-
-
Reduced
-
^
-
- -
- -
-
10
-
-
+
-
-
-
-
-
- -
-
-
-
- -
-
-
-
+ +
- -
-
-
-
-
^ ^
Reduced
Fresh
-
Support
0
-
-
+ +
Fresh
--
-
-
Intensity (a.u.)
Intensity (a.u.)
-
-
- -
Support
20
30
40
50
60
70
80
90
0
10
20
30
2(º)
40
50
60
70
80
90
2(º)
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/YSZO2.
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
23.5
6.3
17.9
5.8
12.0
n.d.
57
52
47
49
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
Total pore volume
(cm3/g) × 102
Particle diameter from
XRD (nm)
TEM (nm)
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 NiAl 2O4 [20], a situation that
77
Chapter 1
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
Figure 1.3. XR TEM images. a) Ni/Al2O3, b) Ni/CeO2, c) Ni/SiC, d) Ni/YSZ, e) Ni/YSZ-O2.
100 nm
100 nm
a)
78
50 nm
d)
b)
50 nm
e)
c)
50 nm
consistent with the large Ni particle sizes observed by XRD and TEM.
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 NiAl 2O4. 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.
79
Chapter 1
TCD signal (a.u.)
Ni Al2O3
Ni CeO2
Ni SiC
Ni YSZ
Ni YSZO2
300
400
500
600
700
800
900
1000 1100 1200
Temperature (K)
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.
80
Intensity (a.u.)
Ni YSZO2
Ni YSZ
100
200
300
400
500
600
700
800
-1
Raman Shift (cm )
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 CO 2 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].
81
Chapter 1
TCD signal (a.u.)
Ni Al2O3
Ni CeO2
Ni SiC
Ni YSZ
Ni YSZO2
300
400
500
600
700
800
900
1000 1100 1200
Temperature (K)
Figure 1.6. CO2 Temperature-programmed desorption profiles.
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
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
82
catalytic activity, with average values for the CH 4 and CO2 reaction rates (5.60×10-4 mol s-1 gNi1
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 CO 2 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 H 2/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

ΔHº = -37.09 KJ∙mol-1
(Equation 1.9)
83
Chapter 1
a)
3.0
7
2.0
5
4
1.5
3
1.0
2
H2/CO Molar ratio
2.5
6
-1
-1
Consumption rate (mol s gNi )·10
4
8
0.5
1
0
8
0.0
3.0
2.0
5
4
1.5
3
1.0
2
0.5
1
0
8
0.0
3.0
4
-1
c)
7
6
2.5
-1
H2/CO Molar ratio
2.5
6
2.0
5
4
1.5
3
1.0
2
H2/CO Molar ratio
Ni
7
-1
Consumption rate (mol s g
-1
)·10
4
b)
0.5
1
0
0
50
100
150
200
0.0
250
Time (min)
Figure 1.7. Dry reforming catalytic activity at 1073 K. Reaction conditions: CH 4 = 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,
84
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.
a)
3.0
7
2.0
5
4
1.5
3
1.0
2
H2/CO Molar ratio
2.5
6
-1
-1
4
8
0.5
1
0
8
0.0
3.0
2.5
6
2.0
5
4
1.5
3
1.0
2
H2/CO Molar ratio
Ni
7
-1
-1
)·10
4
b)
0.5
1
0
0
50
100
150
200
0.0
250
Time (min)
Figure 1.8. Dry reforming catalytic activity at 1073 K. Reaction conditions: CH 4 = 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 H 2/CO molar ratio
remained constant during the experiment, with values around 0.82 for Ni/YSZ and 0.87 for
Ni/YSZ-O2.
85
Chapter 1
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 Ni 0 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 CO 2 molecule [35].
86
Tri-reforming
experiments
Dry reforming
experiments
1.86
Average H2/CO molar ratio
1.71
63.49
-7.29
-24.09
38.19
2.57
1.61
99.22
-0.03
-0.21
99.72
11.13
7.84
0.88
-2.37
0.72
0.77
6.30
-9.48
1.44
5.17
5.60
Ni/CeO2
4.93
Average CO2 conversion (%)
Average TR CO2 reaction rate
(mol s-1 gNi-1) ×104
Change in CO2 reaction rate (%)
end/beginning
Average CH4 conversion (%)
Average TR CH4 reaction rate
(mol s-1 gNi-1) ×104
Change in CH4 reaction rate (%)
end/beginning
Average DR CH4 reaction rate
(mol s-1 gNi-1) ×104
Change in CH4 reaction rate (%)
end/beginning
Average DR CO2 reaction rate
(mol s-1 gNi-1) ×104
Change in CO2 reaction rate (%)
end/beginning
Ni/Al2O3
Table 1.3. Main catalytic results.
1.89
47.21
-34.09
2.25
95.33
-1.77
11.70
0.89
-4.42
6.43
-7.48
5.86
Ni/SiC
1.92
33.75
1.06
2.07
66.68
-8.93
6.14
0.82
-2.73
5.40
-5.89
4.37
Ni/YSZ
1.67
57.91
21.79
3.40
86.43
12.66
9.30
0.87
3.28
6.74
4.07
6.02
Ni/YSZ-O2
87
Chapter 1
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.
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.
88
a)
2.5
12
11
10
7
6
1.5
5
1.0
4
3
0.5
2
1
Ni
-1
)·10
4
0
13
-1
H2/CO Molar ratio
2.0
9
8
0.0
b)
12
11
2.5
10
2.0
9
8
7
6
1.5
5
1.0
4
3
H2/CO Molar ratio
-1
-1
4
13
0.5
2
1
0
13
0.0
c)
2.5
11
10
2.0
9
8
1.5
7
6
5
1.0
4
3
2
H2/CO Molar ratio
-1
-1
4
12
0.5
1
0
0
50
100
150
200
0.0
250
Time (min)
Figure 1.9. Tri-reforming catalytic activity at 1073 K. Reaction conditions: CH 4 = 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 H 2/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 CO 2
reaction rate underwent a slight decay with time on stream. On the other hand, the H 2/CO
molar ratio initially had a value of 1.89 and this increased slightly with time on stream. As the
89
Chapter 1
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.
12
a)
2.5
9
8
2.0
7
1.5
6
5
1.0
4
3
0.5
2
1
-1
0.0
b)
2.5
11
10
9
2.0
8
7
1.5
6
5
1.0
4
3
H2/CO Molar ratio
Ni
-1
)·10
4
0
12
H2/CO Molar ratio
-1
10
-1
4
11
0.5
2
1
0
0
50
100
150
200
0.0
250
Time (min)
Figure 1.10. Tri-reforming catalytic activity at 1073 K. Reaction conditions: CH 4 = 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 H 2/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.
90
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 CO 2 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 Ni 0 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.
91
Chapter 1
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 CH 4
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 CO 2 conversion. In this respect, Song et al. reported very
similar CO2 conversion values for both catalysts (61% for Ni/Al 2O3 and 60% for Ni/CeO2)
whereas we obtained a much lower value for Ni/Al 2O3 (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 NiAl 2O4 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 NiAl 2O4 metal phase. The catalyst
92
Ni/CeO2 presented the highest basicity value and this gave rise to the lowest H 2/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 trireforming process as they yielded a high methane reaction rate without significant deactivation.
93
Chapter 1
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) 365377.
[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.
[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) 176185.
94
[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) 7382.
[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) 15341539.
[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) 157166.
[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) 155159.
[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.
95
Chapter 1
[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.
[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.
96
CHAPTER 2
Precursor influence and catalytic behavior of
Ni/CeO2 and Ni/SiC catalyst for the trireforming process.
Resumen
Abstract
2.1. INTRODUCTION
2.2. EXPERIMENTAL
2.4. CONCLUSIONS
2.5. REFERENCES
Precursor influence and catalytic behavior of Ni/CeO2 and Ni/SiC catalyst for the trireforming process
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.
99
Chapter 2
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. CO 2-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.
100
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, nickelbased catalysts have been proven to be the most suitable ones for hydrocarbon reforming.
Following this idea, nickel supported on oxides, such as Al 2O3, 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)
◦
-1
(Equation 2.2)
CH4 + 1/2O2 → CO + 2H2
(H = −35.6 kJmol )
(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)
101
Chapter 2
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.
102
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 (NH 4)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
103
Chapter 2
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. Temperatureprogrammed 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 Ushaped 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 JEM4000EX 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 carbonsupported grid. Mean nickel particle size evaluated as the surface-area weighted diameter ( d s )
was computed according to (Equation 2.8):
ds 
104
n d
i
i
n i d i2
3
i
(Equation 2.8)
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% CH 4/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 H 2O 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
105
Chapter 2
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. ForSiC-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 NiNS and Ni-AS were 52 and 50 nm, respectively. This particle size order (Ni-NS ≈ Ni-AS < NiCS ≈ 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.
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
6.4
3.7
6.6
9.8
17.9
19.1
19.8
20.1
60
86
88
97
41
42
68
75
57
87
92
116
52
50
70
71
Total pore volume
(cm3/g) × 102
XRD (nm)
TEM (nm)
106
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
Chapter 2
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 NiNS 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.
111
a)
220
Intensity (a.u.)
311
200
331
Fresh
*
*
+
Reduced
0
10
20
30
40
422
420
222
*
+
50
60
70
80
90
222
*
400
2º)
111
b)
220
Intensity (a.u.)
311
*
200
*
*
Fresh
+
+
Reduced
0
10
20
30
40
50
60
70
80
90
2 (º)
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
108
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 NiCS 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.
TCD signal (a.u.)
a)
Ni-NS
Ni-AS
Ni-NS
Ni-CiS
CeO2
300
400
500
600
700
800
900
1000 1100 1200
Temperature (K)
TCD signal (a.u.)
b)
Ni-CS
Ni-CiS
Ni-AS
Ni-NS
SiC
300
400
500
600
700
800
900
1000 1100 1200
Temperature (K)
Figure 2.3. Temperature-programmed reduction profiles. a) ceria-based catalysts, b) SiCbased catalysts.
109
Chapter 2
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. CO 2–
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.
110
TCD signal (a.u.)
a)
Ni-CiC
Ni-CC
Ni-NC
Ni-AC
CeO2
300
400
500
600
700
800
900
1000 1100 1200
Temperature (K)
b)
TCD signal (a.u.)
Ni-CS
Ni-AS
Ni-CiS
Ni-NS
SiC
300
400
500
600
700
800
900
1000 1100 1200
Temperature (K)
Figure 2.4. CO2 Temperature-programmed desorption profiles. a) ceria-based catalysts, b)
SiC-based catalysts.
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 (5001000 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
111
Chapter 2
Figure 2.5 shows the catalytic performance of ceria-based catalysts in terms of methane
consumption and carbon dioxide consumption rates and H 2/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.
112
2.0
7.5
1.5
5.0
1.0
2.5
0.5
0.0
50
100
150
200
0.0
250
b)
2.5
10.0
2.0
7.5
1.5
5.0
1.0
2.5
0.5
0.0
0
50
Time (min)
2.5
10.0
2.0
7.5
1.5
5.0
1.0
2.5
0.5
0.0
100
150
200
0.0
250
200
0.0
250
Time (min)
3.0
12.5
H2/CO Molar ratio
-1 -1
4
c)
-1
-1
4
12.5
50
150
Time (min)
3.0
0
100
d)
2.5
10.0
2.0
7.5
1.5
5.0
1.0
2.5
H2/CO Molar ratio
0
3.0
12.5
H2/CO Molar ratio
-1 -1
4
2.5
10.0
-1
-1
4
a)
H2/CO Molar ratio
3.0
12.5
0.5
0.0
0
50
100
150
200
0.0
250
Time (min)
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 H 2/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
113
Chapter 2
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 Ni 0 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)

3.0
4
5.0
1.0
2.5
0.5
0.0
0
50
100
150
200
0.0
250
2.5
2.0
-1
-1
1.5
H2/CO Molar ratio
2.0
7.5
b)
10.0
7.5
1.5
5.0
1.0
2.5
0.5
0.0
0
50
Time (min)
150
200
0.0
250
3.0
5.0
1.0
2.5
0.5
0.0
0
50
100
150
Time (min)
200
0.0
250
2.5
2.0
7.5
1.5
5.0
1.0
2.5
H2/CO Molar ratio
1.5
d)
10.0
-1
2.0
7.5
H2/CO Molar ratio
-1
4
4
12.5
2.5
c)
10.0
-1
-1
100
Time (min)
3.0
12.5
H2/CO Molar ratio
2.5
a)
10.0
-1
-1
3.0
12.5
4
12.5
(Equation 2.9)
0.5
0.0
0
50
100
150
200
0.0
250
Time (min)
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
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
114
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 ceriabased 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 CO 2 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 CO 2 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 H 2/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
115
Chapter 2
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 CO 2 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 CO 2 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 H 2/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 Ni 0 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
116
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
117
Chapter 2
Ni/SiC catalysts. The results showed that these latter catalysts can be considered as
promising ones for the tri-reforming process.
118
2.5 REFERENCES
377 (2010) 16-26.
[3] M.-S. Fan, A.Z. Abdullah, S. Bhatia, Applied Catalysis B: Environmental. 100 (2010) 365377.
[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.
[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.
119
Chapter 2
[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) 464477.
[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.
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) 2939.
[32] I. Chen, S.Y. Lin, D.W. Shiue, Industrial & Engineering Chemistry Research. 27 (1988)
926-929.
120
[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) 15341539.
[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.
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) 283294.
121
Chapter 2
(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.
122
CHAPTER 3
Methane tri-reforming over a Ni/-SiC-based
catalyst: Optimizing the feedstock
composition
Resumen
Abstract
3.1. INTRODUCTION
3.2. EXPERIMENTAL
3.3.3. Influence of the feedstock composition
on the H2/CO molar ratio
3.4. CONCLUSIONS
3.5. REFERENCES
Methane tri-reforming over a Ni/-SiC-based catalyst: Optimizing the feedstock composition
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.
125
Chapter 3
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.
126
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 CO 2 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 H 2/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)
◦
-1
(Equation 3.2)
CH4 + 1/2O2 → CO + 2H2
(H = −35.6 kJmol )
(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 trireforming 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
127
Chapter 3
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 H 2/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
obtained can have.
128
the synthesis gas
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 secondorder 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 N 2 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.
129
Chapter 3
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 ( d s ), according to Equation
3.8:
ds 
n d
i
3
i
i
n i d i2
(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
130
for each point of the factorial design and the necessary quantity of N 2 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.
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 = (
2 k 2 ·n  2 k 1 )
(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
131
Chapter 3
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):
k
k
i 1
i 1
k
Yi =    i ·xi    ii ·xi  
2
k
  ij ·xi ·x j
i 1 j 1
(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 2 k 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.
132
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.
Characterization parameter
Numerical value
Ni loading (%)
3.9
2
-1
Surface area (m g )
25.9
Total pore volume (cm3 g-1) x102
17.9
Particle diameter from TEM (nm)
41
Particle diameter from XRD (nm)
52
-1
Total basic sites (mol g )
4.6
Reduction Temperature (K)
973
Table 3.2 shows the range of values studied for each independent variable: CH 4, 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
CH4 NmL min-1
CO2 NmL min
high (+)
center (0)
axial (-)
axial (+)
4.5
7.5
6
3.59
8.41
-1
2
4
3
1.39
4.61
-1
2
4
3
1.39
4.61
0.5
2
1.25
0.04
2.46
H2O NmL min
O2 NmL min
low (-)
-1
133
Chapter 3
Table 3.3. Central composite design results.
Experiment Run CH4 CO2 H2O O2
24 factorial
design
4 central
points
8 star
points
134
CH4 conversion
H2/CO
(%)
molar ratio
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
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
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
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 H 2/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):
135
Chapter 3
Table 3.4. Factorial design statistical analysis.
CH4 conversion (%)
H2/CO
 CH -2.273
 CO = -0.918
 H O = -2.985
 O = 3.332
 CH = -0.259
 CO = -0.289
 H O = 0.307
 O = 0.288
 CH CO = 0.958
 CH  H O = 1.592
 CH O = -2.777
 CO H O = -1.438
 CO O = 2.500
 H OO = 3.270
2
 CH CO = 0.051
 CH  H O = -0.120
 CH O = -0.106
 CO H O = -0.046
 CO O = 0.006
 H OO = 0.111
91.470
t = 3.182
s = 1.896
 3.016
O2, H2O-O2
2.052
t = 3.182
s = 0.086
 0.137
CH4, CO2, H2O, O2
4
2
Main effects
2
2
4
Interactions
4
2
4
2
2
2
2
2
Significance test
(confidence level: 95%)
Mean responses
Standard deviation
Confidence interval
Significant variables
Significance of curvature
Curvature
Confidence curvature interval
Significance
2
2
4
2
2
2
4
4
2
2
4
2
2
2
2
2
2
2
0.181
 0.154
Yes
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 H 2/CO molar ratio was accurately
predicted, with an average error rate of 5.2%.
136
4.0
Experimental H2/CO molar ratio
3.5
3.0
2.5
2.0
1.5
1.0
1.0
1.5
2.0
2.5
3.0
3.5
4.0
Predicted H2/CO molar ratio (Equation 3.13)
Figure 3.1. Experimental vs. predicted H2/CO molar ratio.
3.3.3. Influence of the feedstock composition on the H 2/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 CO 2 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.
137
Chapter 3
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 H 2/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 CO 2 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 H 2/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 H 2/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 CH 4/(CO2+H2O) molar ratio
constant at 2/2.5, methane conversion of the dry + steam reforming rose slightly from 97.7%
138
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 CH 4: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.
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 trireforming, analyzing the influence of oxygen concentration when the H 2O/CO2 ratio was 1 and
the (H2O + CO2 + 2O2)/CH4 ratio was set at 1.2. The oxygen clearly affected catalytic
139
Chapter 3
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 CO 2, 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.
140
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.
141
Chapter 3
Table 3.5. Factorial design for the reaction heat results and optimized variables.
Experiment Run CH4 CO2 H2O
24 factorial design
4 central points
8 star points
Variable
Optimized value
142
O2
Reaction heat (kJ mol-1)
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
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
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
CH4 CO2 H2O
O2 H2/CO Reaction heat (kJ mol-1)
3.59 4.12 1.39 2.11
2.10
-496.4
(H◦ = −880 kJ mol-1)
CH4 + 2O2 → CO2 + 2H2O
(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.
3500
-1
Calculated reaction heat (KJ mol )
3000
2500
2000
1500
1000
500
0
0
500
1000 1500 2000 2500 3000 3500
-1
Predicted reaction heat (KJ mol ) (Equation 3.15)
Figure 3.5. Experimental vs. predicted overall reaction heat.
Taking into account equations 13 and 15, the target region of the H 2/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
143
Chapter 3
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 CO 2/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 CO 2 [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.
10
3.0
-1
7
6
2.0
5
1.5
4
3
1.0
2
H2/CO Molar ratio
2.5
8
-1
4
9
0.5
1
0
0
100
200
300
400
0.0
500
Time (min)
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 (
(right axis).
144
)
) vs. time on stream
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 H 2/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 H 2O 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.
145
Chapter 3
3.5 REFERENCES
(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.
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.
146
[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) 18571864.
[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) 650658.
[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.
147
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
4.4. CONCLUSIONS
4.5. REFERENCES
Influence of alkaline and alkaline-earth cocations on the performance of Ni/-SiC catalysts in
the methane tri-reforming reaction
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.
151
Chapter 4
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.
152
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
◦
-1
(H = −35.6 kJmol )
(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 4.44.7); 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 H 2/CO molar ratio
could be controlled by modifying the reagents ratio, yielding a synthesis gas with a H 2/CO
molar ratio around 2.
2CO  C + CO2

CH4  C + 2H2

(H◦ = −172.2 kJmol-1)
(H◦ = 74.9 kJmol-1)
(Equation 4.4)
(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
153
Chapter 4
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
154
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 H 2/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
155
Chapter 4
min. The sample was then heated in 50 mL min-1 of He with a heating rate of 10 K 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 CH 4
(99.995% purity), 10% CO 2/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 CP4900). 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,
156
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 SiO 2 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.
157
158
3.6
41
14.4
63
Total pore volume (cm3 g-1)x102
Ni particle diameter from XRD (nm)
0.896 0.542
-
5.46
43.22
H2 consumption (mmol g-1)
Total basic sites (mol g-1)
Diffraction angle of NiO (2
5.5
24.0
Surface area (m2 g-1)
4.8
4.5
-
-
0.378
85
1.1
0.9
4.9
-
0.8
1.5
4.5
56
12.9
18.5
4.6
-
-
-
-
5.63
-
2.82
0.556
53
8.3
11.8
4.6
43.14
11.05
0.873
39
13.1
22.5
5.2
42.9
11.84
0.623
38
14.0
21.8
5.1
43.02
13.54
0.615
33
11.2
18.0
4.6
42.9
14.39
0.553
32
12.6
21.2
5.5
Ni:K Ni:Ca Ni:Ca Ni:Mg Ni:Mg Ni:Mg Ni:Mg
2/1
10/1
2/1
10/1
4/1
2/1
1/1
0.284 0.198 0.761
-
2.9
4.7
4.7
Ni:Na Ni:Na Ni:K
10/1
2/1
10/1
Ni loading (%)
Ni
Table 4.1. Main physical properties of the catalysts.
Chapter 4
10
Fresh
20
20
30
30
^
40
50
50
+
2(º)
^
+
2(º)
^* *
^
^
40
^
60
^ *
^
60
^
70
80
80
*
^+
^
^
^
^
70
^
90
90
0
10
10
Fresh
Reduced
e)
0
Fresh
Reduced
b)
20
20
º
º
º
º
^
^
30
+
40
50
+
2(º)
*
+
2(º)
50
^ º
70
80
60
70
80
^
º
*
^ *^
^
^
60
º* º º º º º ^ º ^ º
+
º ºº º º ^ º ^ º
40
^ *
30
ºº
ºº
^
90
90
10
0
10
Fresh
Reduced
f)
0
Fresh
Reduced
c)
20
º
º
20
º
º
30
30
ºº
ºº
40
^
^
2(º)
50
2(º)
50
º ºº
º ºº
40
^
^
60
^
^
60
70
70
^
^
º^ º ^ º
º^ º ^ º
80
80
90
90
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.
0
10
Reduced
d)
0
^
Intensity (a.u.)
Intensity (a.u.)
Intensity (a.u.)
Intensity (a.u.)
Intensity (a.u.)
Intensity (a.u.)
a)
159
Chapter 4
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.
160
a)
b)
^
^
+
^
^
+
+
Reduced
Fresh
0
10
^
*
^ *
*
^
20
30
40
50
60
^
^
Intensity (a.u.)
Intensity (a.u.)
^
-
80
90
0
10
20
30
40
50
0
10
20
30
40
^+
+
^
*
50
2(º)
90
60
^
^
70
^
Intensity (a.u.)
Intensity (a.u.)
Fresh
80
^
^
^ *
*
^
70
^
^
+
60
2(º)
d)
^
*
Fresh
^
Reduced
- ^+
^ *- ^^
- ^ * *
^ -
2(º)
c)
+
^
^
Reduced
^
^*
70
+
+
*
80
^
Reduced
0
10
^
^
*
^ *
*
^
Fresh
90
+
20
30
40
50
60
^
^
70
*
80
90
2(º)
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
161
Chapter 4
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.
162
TCD signal (a.u.)
a)
Ni
Ni:Na 10/1
Ni:Na 2/1
Ni:K 10/1
Ni:K 2/1
300
400
500
600
700
800
900
1000 1100 1200
Temperature (K)
TCD signal (a.u.)
b)
Ni:Ca 10/1
Ni:Ca 2/1
Ni:Mg 10/1
Ni:Mg 2/1
300
400
500
600
700
800
900
1000 1100 1200
Temperature (K)
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 CO 2 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 Ni 0). 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
163
Chapter 4
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 trireforming 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 CO 2 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).
164
2.5
6
5
2.0
4
1.5
3
2
1.0
1
0
0.5
0
5
10
15
20
25
4.0
8
7
3.5
6
5
3.0
4
2.5
3
2
2.0
1
0
1.5
0
5
10
Time (h)
3.0
c)
-1
7
2.5
6
5
2.0
4
1.5
3
2
1.0
1
0
0.5
5
10
15
Time (h)
25
20
25
10
H2/CO Molar ratio
-1 -1
4
4
9
-1
3.5
0
20
Time (h)
10
8
15
3.5
9
3.0
8
d)
7
2.5
6
5
2.0
4
1.5
3
2
H2/CO Molar ratio
-1
7
4.5
b)
9
H2/CO Molar ratio
3.0
8
10
H2/CO Molar ratio
-1 -1
4
3.5
a)
9
-1
4
10
1.0
1
0
0.5
0
5
10
15
20
25
Time (h)
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 CO 2 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.
165
Chapter 4
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
TCD signal (a.u.)
attributed to the formation of a NiO-MgO solid solution [39, 40].
Ni
Ni:Mg 10/1
Ni:Mg 4/1
Ni:Mg 2/1
Ni:Mg 1/1
300
400
500
600
700
800
900
1000 1100 1200
Temperature (K)
Figure 4.5. TPR profiles for Mg promoted and Ni/SiC catalysts.
166
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 
3.5
a)
3.0
2.5
2.0
1.5
1.0
0.5
0
5
10
15
20
25
Time (h)
10
9
8
7
6
5
4
3
2
1
0
3.5
b)
3.0
2.5
2.0
1.5
1.0
H2/CO Molar ratio
-1
-1
H2/CO Molar ratio
10
9
8
7
6
5
4
3
2
1
0
4
-1
-1
Consumption rate (mol s gNi )·10 Consumption rate (mol s gNi )·10
4
Ni:Mg = 4/1 < Ni:Mg = 2/1 < Ni:Mg = 1/1.
0.5
0
5
10
15
20
25
Time (h)
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
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 CH 4
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 tri167
Chapter 4
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 H 2/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.
168
Ni
Average CH4 reaction rate
(mol s-1 gNi-1) ×104
Average CH4 conversion (%)
Ni:Mg 10/1 Ni:Mg 4/1 Ni:Mg 2/1 Ni:Mg 1/1
7.96
8.92
8.46
8.95
8.65
73.7
95.4
88.8
84.7
97.9
Drop in CH4 reaction rate (%)
Average CO2 reaction rate
(mol s-1 gNi-1) ×104
Average CO2 conversion (%)
14.8
2.6
2.3
1.1
0.9
2.72
2.72
2.38
1.83
1.84
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
2.00
Oxygen consumption in TPO
35.28
(mmol g-1)
Particle diameter from XRD after
54
reaction (nm)
2.09
2.23
2.36
2.76
10.70
8.23
5.71
6.24
42
38
39
34
Ni
TCD signal (a.u.)
Ni:Mg 10/1
Ni:Mg 4/1
Ni:Mg 2/1
Ni:Mg 1/1
400
600
800
1000
1200
Temperature (K)
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
169
Chapter 4
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 Cand
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.
TCD signal (a.u.)
Ni
Ni:Mg 10/1
Ni:Mg 4/1
Ni:Mg 2/1
Ni:Mg 1/1
300
400
500
600
700
800
900
1000 1100 1200
Temperature (K)
Figure 4.8. TPO profiles after reaction for Mg promoted and Ni/SiC catalysts.
170
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 H 2/CO
molar ratio was. Finally, the higher the interaction between Ni and Mg, the slower the catalyst
deactivation rate was observed.
171
Chapter 4
4.5 REFERENCES
[1] E. Iglesia, Applied Catalysis A: General. 161 (1997) 59-78.
[2] M.E. Dry, Catal. Today. 71 (2002) 227-241.
[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.
172
[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.
173
Chapter 4
(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.
174
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
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 Ni 2Si 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.
177
Chapter 5
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 Ni 2Si 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.
178
metal impregnation
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 CH 4 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.
179
Chapter 5
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. -SiCexhibits 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 trireforming [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
180
metal impregnation
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 coimpregnation (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 N 2 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 H 2/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
181
Chapter 5
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 ( d s ), was computed according to:
ds 
n d
i
3
i
i
n i d i2
(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
182
metal impregnation
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.
183
184
0
10
SiC
Ni/SiC
a)
20
^
º
40
2(º)
50
º
60
220
^
70
311
^
80
222
^º
90
400
0
10
20
^
^
30
Mg/Ni/SiC 1/10
Ni/Mg/SiC 1/10
Ni-Mg/SiC 1/10
^
40
^
^
^
50
º
º
º
2(º)
º
º
º
60
^
^
^
70
^
^
^
^
^
^
80
90
Intensity (a.u.)
Intensity (a.u.)
0
10
20
Mg/Ni/SiC 1/1
Ni/Mg/SiC 1/1
Ni-Mg/SiC 1/1
c)
30
^
^
^
40
^
º
50
º
º
2(º)
º
^ º
^ º
60
^
^
^
Figure 5.1. XRD profiles, where (^) denotes reflection of SiC and (º) denotes reflection of metallic nickel.
30
111 200
^
b)
70
^
^
^
^
^
^
80
90
Chapter 5
Intensity (a.u.)
36
43.22
76.9
43.22
57
99.8
Diffraction angle of NiO (2
Particle diameter from TEM (nm)
Total pore volume (cm g )x10
63
22.2
23.9
10.7
-1
14.4
3
Surface area (m g )
2
0.2
-
-1
Promoter loading (%)
2
4.6
4.5
73.3
51
43.14
41
11.2
21.8
0.2
4.8
99.4
-
43.14
39
13.1
22.5
0.2
5.2
75.3
-
43.14
33
13.9
21.1
1.8
4.5
57.8
32
42.94
33
11.0
16.9
1.7
4.2
59.0
-
42.9
32
12.6
21.2
1.8
5.5
Mg/Ni/SiC Ni/Mg/SiC Ni-Mg/SiC Mg/Ni/SiC Ni/Mg/SiC Ni-Mg/SiC
1/10
1/10
1/10
1/1
1/1
1/1
Ni loading (%)
Ni/SiC
metal impregnation
185
Chapter 5
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
a)
186
Figure 5.2. TEM pictures. a) Ni/SiC, b) Ni/Mg/SiC 1/10, c)Ni/Mg/SiC 1/1.
b)
c)
impregnation were smaller and better dispersed.
metal impregnation
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 NiOMgO 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
187
Chapter 5
lower values of reduction degree. This is in agreement with the higher extension of the NiOMgO solid solution, which makes the reduction process more difficult.
Ni/SiC
TCD signal (a.u.)
Mg/Ni/SiC 1/10
Ni/Mg/SiC 1/10
Ni-Mg/SiC 1/10
Mg/Ni/SiC 1/1
Ni/Mg/SiC 1/1
Ni-Mg/SiC 1/1
300
400
500
600
700
800
900 1000 1100 1200
Temperature (K)
Figure 5.3. Temperature Programmed Reduction profiles.
The desorption profiles of CO 2-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 CO 2
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.
188
metal impregnation
Ni/SiC
Mg/Ni/SiC 1/10
TCD signal (a.u.)
Ni/Mg/SiC 1/10
Ni-Mg/SiC 1/10
Mg/Ni/SiC 1/1
Ni/Mg/SiC 1/1
Ni-Mg/SiC 1/1
300 400 500 600 700 800 900 1000 1100 1200 1300
Temperature (K)
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 H 2/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 CO 2
consumption rate [26], what will increase in addition the H2/CO molar ratio. Catalysts
189
Chapter 5
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
2.5
6
5
2.0
4
1.5
3
2
1.0
1
0
0.5
0
5
10
15
20
5.0
b)
9
4.5
8
7
4.0
6
5
3.5
4
3.0
3
2
2.5
1
0
25
2.0
0
5
10
8
7
c)
4.5
4.0
6
5
3.5
4
3.0
3
2
2.5
1
0
2.0
0
5
10
15
Time (h)
15
20
25
Time (h)
5.0
20
25
10
H2/CO Molar ratio
-1 -1
4
Consumption rate (mol s gNi )*10
-1
9
-1
4
Time (h)
10
H2/CO Molar ratio
-1
7
3.0
10
3.5
9
3.0
8 d)
7
2.5
6
5
2.0
4
1.5
3
2
H2/CO Molar ratio
-1
8
3.5
a)
9
H2/CO Molar ratio
-1 -1
4
10
4
catalytic activity, probably due to the blockage of the Ni active sites after Mg impregnation.
1.0
1
0
0.5
0
5
10
15
20
25
Time (h)
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
-1
-1
7
4.0
3.5
5
3.0
4
3
2.5
2
1
2.0
100
3.5
6
3.0
5
2.5
4
3
2.0
2
1
100
9
H2/CO Molar ratio
b)
1.5
c)
4.0
8
3.5
-1
7
6
3.0
5
2.5
4
3
2.0
H2/CO Molar ratio
-1
7
4.0
9
8
4
4.5
6
-1
4
8
-1
a)
9
H2/CO Molar ratio
10
Consumption rate (mol s gNi )*10 Consumption rate (mol s gNi )*10
4
metal impregnation
2
1
1.5
0
0
5
10
15
20
25
Time (h)
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
191
Chapter 5
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 trireforming 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.
192
Average CH4 reaction rate
7.96
(mol s-1 gNi-1) ×104
Drop in CH4 reaction rate
after 24 h of time on stream 1.30
(mol s-1 gNi-1) ×104
Average CO2 reaction rate
2.72
(mol s-1 gNi-1) ×104
Drop in CO2 reaction rate
after 24 h of time on stream 0.76
(mol s-1 gNi-1) ×104
Average H2/CO molar ratio 2.00
Oxygen consumption in
35.28
TPO (mol g-1)
Ni/SiC
8.65
0.08
1.84
0.74
2.76
6.24
8.77
0.00
1.48
2.56
2.79
3.72
5.28
1.27
0.25
0.35
3.36
10.81
8.92
0.24
2.72
0.45
2.09
10.70
8.50
0.30
1.05
0.51
2.96
4.04
3.99
4.34
0.12
1.55
3.73
23.44
Mg/Ni/SiC Ni/Mg/SiC Ni-Mg/SiC Mg/Ni/SiC Ni/Mg/SiC Ni-Mg/SiC
1/1
1/1
1/1
1/10
1/10
1/10
Table 5.2. Reaction and characterization after reaction parameters.
metal impregnation
193
Chapter 5
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
Cand 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.
Ni/SiC
Mg/Ni/SiC 1/10
TCD signal (a.u.)
Ni/Mg/SiC 1/10
Ni-Mg/SiC 1/10
Mg/Ni/SiC 1/1
Ni/Mg/SiC 1/1
Ni-Mg/SiC 1/1
300 400 500 600 700 800 900 1000 1100 1200 1300
Temperature (K)
Figure 5.7. Temperature Programmed Oxidation profiles after reaction.
194
metal impregnation
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.
195
196
*
^
º
#
^º
Ni/SiC
º
^
*
*
*
^
^
^
º
*º
*
*
º
#
#
#
#
#
#
Mg/Ni/SiC 1/10
º
Ni/Mg/SiC 1/10
º
Ni-Mg/SiC 1/10
º
Intensity (a.u.)
^
^
^
*
*
*
^
^
^
*
*
*
º
º
º
#
#
#
#
#
#
Mg/Ni/SiC 1/1
º
Ni/Mg/SiC 1/1
º
Ni-Mg/SiC 1/1
º
2(º)
30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60
c)
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.
2(º)
^
#
^
^
30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60
*
º
b)
2(º)
^
Intensity (a.u.)
30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60
a)
Chapter 5
Intensity (a.u.)
metal impregnation
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 NiMg/SiC 1/1 was selected as the best one due to its high catalytic activity, great stability and
low coke production.
197
Chapter 5
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.
(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.
198
metal impregnation
[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.
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.
199
Chapter 5
(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.
200
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
6.4. CONCLUSIONS
Catalytic and kinetic analysis of the methane tri-reforming process using a Ni-Mg/-SiC
catalyst
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 trireformado 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.
203
Chapter 6
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.
204
catalyst
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 CH 4 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)


205
Chapter 6
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 temperatureprogrammed 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
206
catalyst
was removed by a 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 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% CO 2/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 min1
using 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 CH 4/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
207
Chapter 6
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):
208
catalyst
(Equation 6.6)
A VBA-Excel application was developed to solve this model [14-16]. The BaderDeuflhard 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
209
Chapter 6
(IUPAC classification) that is characteristic of macroporous and mesoporous materials [21],
without nitrogen adsorption in the micropore range (P/P 0 ≤ 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.
80
0,25
0,20
60
dV/d(log r)
3
-1
N2 adsorbed (cm g )
0,15
Support
0,10
40
Catalyst
0,05
0,00
1
10
100
1000
10000
Pore radius nm)
20
Support
0
0,00
Catalyst
0,25
0,50
0,75
1,00
Relative Pressure (P/P0 )
Figure 6.1. Nitrogen adsorption–desorption isotherms of catalyst and support.
210
catalyst
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) x102
9.1
37
Total basic sites (mol g-1)
10.51
Reduction Temperature (K)
973
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).
211
212
10
20
30
111
º
º
^
40
50
º
60
220
^
2(º)
200
++
º
º
º
º
70
311
+
º
80
222
º+
º^
90
400
º
º
300
b)
450
750
900
Temperature (K)
600
1050
c)
1200 300
TCD signal (a.u.)
TCD signal (a.u.)
450
750
900
Temperature (K)
600
1050
1200
denotes reflection of nickel oxide, b) TPR profile, c) CO 2-TPD profiles.
Figure 6.2. Characterization results a) XRD profiles, where (º) denotes reflection of SiC, (^) denotes reflection of metallic nickel and (+)
0
Support
Fresh
Reduced
a)
Chapter 6
Intensity (a.u.)
catalyst
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 CO 2 conversion
values are below 0, which actually means that there is a net production of CO 2 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 H 2/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 H 2/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)
213
Chapter 6
Table 6.3. CH4 conversion values obtained for each experiment.
Exp.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
214
Temperature
680
710
740
770
800
830
860
890
920
950
980
1010
8.36
3.48
0.00
0.00
9.95
2.39
0.00
8.45
0.26
3.00
6.83
8.83
0.00
1.58
4.26
7.55
0.00
7.21
2.91
4.33
4.83
4.26
0.00
2.46
1.00
7.65
6.18
0.00
0.62
0.00
0.00
2.28
2.28
0.00
5.63
4.25
0.00
4.08
0.19
0.38
5.28
3.15
0.25
7.39
1.92
2.75
9.15
7.04
3.87
2.56
8.67
7.25
0.00
8.18
5.72
4.79
10.32
8.67
5.24
3.29
10.77
7.64
10.18
0.00
10.00
0.51
10.93
3.39
10.69
0.18
8.57
5.75
11.01
4.89
5.31
1.24
6.94
4.35
0.00
9.59
3.54
4.54
10.52
8.11
6.29
4.49
10.39
8.57
2.04
10.62
7.78
6.11
12.42
10.39
7.19
4.53
13.40
10.46
11.41
0.00
13.22
3.18
12.12
5.36
13.66
2.76
11.42
7.53
8.77
3.16
6.99
2.73
9.31
6.11
2.15
11.86
6.88
6.81
12.05
9.32
9.01
7.14
12.59
10.83
6.51
13.57
10.41
7.99
15.27
12.59
9.10
6.50
18.19
14.46
14.45
1.82
18.06
6.72
15.22
8.06
17.64
6.30
15.16
9.75
13.35
1.19
12.34
4.91
12.38
8.65
5.02
13.68
11.08
9.35
13.95
11.44
12.85
10.64
15.61
14.05
10.68
15.90
13.48
10.99
18.92
15.61
11.74
8.83
23.37
19.43
18.02
5.04
23.83
11.30
18.44
12.41
22.43
11.21
19.30
12.94
19.86
2.25
11.99
7.27
16.03
11.79
8.84
16.20
12.64
12.85
16.35
13.47
17.38
15.24
19.14
18.29
15.52
19.89
17.31
14.54
22.59
19.14
15.49
11.56
29.83
25.48
23.26
9.13
30.60
16.63
23.24
18.24
28.51
16.40
23.87
17.32
24.81
8.28
15.35
9.14
22.14
16.27
13.06
19.04
17.27
16.51
18.87
15.21
22.88
20.78
23.02
22.09
21.55
24.51
21.19
17.59
26.69
23.02
19.81
14.01
36.36
32.38
28.93
13.29
38.01
24.29
29.09
24.51
35.20
22.87
29.72
22.39
26.90
8.94
17.90
10.21
27.50
21.29
17.69
21.17
19.24
21.05
20.01
14.31
27.92
27.31
28.06
24.59
28.76
29.65
24.60
19.15
30.78
28.06
23.80
14.45
43.72
41.31
37.03
17.98
46.70
32.59
35.78
30.92
43.48
31.05
36.27
28.13
34.72
10.28
20.15
11.79
34.09
26.92
20.77
22.69
23.05
26.20
18.96
13.56
37.57
34.73
34.24
25.96
38.53
35.62
26.87
20.29
37.18
34.24
30.63
14.41
52.88
51.87
46.11
23.25
56.03
41.30
43.19
37.03
53.24
40.38
43.93
32.07
40.27
13.54
21.46
14.74
39.70
32.73
20.78
23.41
28.23
32.13
18.72
14.32
48.24
42.59
40.90
28.56
49.75
42.20
27.73
21.02
44.99
40.90
36.11
15.58
62.12
62.45
56.42
30.25
66.91
51.84
51.87
42.72
62.34
51.46
49.57
32.23
46.63
13.80
25.08
19.05
46.82
37.81
22.16
24.82
34.76
38.32
20.70
16.73
57.02
50.94
47.59
33.38
59.15
49.80
30.44
24.19
53.44
47.59
36.73
18.22
71.34
72.35
65.94
39.83
76.19
61.44
61.15
47.66
71.45
61.80
54.49
33.32
53.90
28.27
30.14
24.53
55.02
42.32
26.26
28.45
53.52
43.31
25.13
19.38
66.72
59.07
55.70
38.82
66.51
56.79
35.55
28.40
61.82
55.70
38.10
21.56
80.01
81.70
75.51
50.31
85.41
71.46
70.71
54.46
80.10
70.88
59.29
36.28
Experiment number
catalyst
2
Methane conversion (%)
4
6
8
75
10
12
60
14
45
16
18
30
20
22
15
24
26
28
30
32
34
36
680 710 740 770 800 830 860 890 920 950 980 1010
Temperature (K)
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 O 2 and H2O feed flow. Comparing each
experiment with high O2 feed flow with the one with a similar feed composition but a low O 2
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 H 2O 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 CO 2 feed flow with the CH4 conversion, despite it is
another co-reactant and it can be forecasted that an increase in the CO 2 feed will favour the dry
reforming reaction and thereby will increase the global CH 4 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. 2Exp. 6-Exp. 10; etc) there is not a clear correlation between CO 2 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).
215
Chapter 6
Table 6.4. CO2 conversion values obtained for each experiment.
Exp.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
216
Temperature
680
710
740
770
800
830
860
890
920
950
980
1010
8.62
5.94
-0.28
0.85
4.44
-1.00
-0.43
12.04
1.92
-1.45
4.66
2.23
2.52
0.24
11.48
10.80
5.32
4.51
0.26
5.32
10.08
-216
4.85
16.41
8.89
11.59
6.09
6.55
6.88
2.72
6.96
6.85
8.34
0.36
7.13
-0.07
-5.06
6.81
-2.29
1.55
-7.30
-4.07
-7.51
8.70
-3.51
-11.5
-15.8
0.08
-4.64
-1.26
4.36
8.48
-7.53
3.22
-2.72
4.16
-20.2
-242
-19.9
0.54
1.24
8.52
1.24
4.69
-5.60
-0.18
0.54
3.04
-18.0
-5.03
-20.53
-2.27
4.32
7.77
-1.41
3.30
-10.4
-5.54
-8.94
7.64
-6.04
-21.9
-18.1
1.60
-4.49
-1.39
4.75
9.72
-10.9
1.53
-0.53
5.18
-28.5
-240
-28.6
0.06
-0.24
5.30
1.50
4.92
-9.30
-2.54
-1.11
3.57
-28.8
-13.88
-24.08
-2.29
-1.05
6.22
3.15
6.74
-12.8
-4.90
-7.14
8.82
-25.1
-30.8
-17.3
5.62
-3.73
-1.86
6.28
12.83
-15.3
-0.70
3.06
8.09
-36.5
-235
-26.6
3.47
-0.49
3.73
3.85
5.42
-12.0
-4.62
-0.21
6.29
-39.5
-27.81
-23.15
0.28
-1.10
3.87
9.35
10.23
-12.2
-3.88
-1.75
9.43
-36.3
-37.4
-10.8
13.00
-6.77
-0.89
9.07
17.24
-15.6
-1.50
8.54
13.09
-41.5
-225
-25.0
10.12
0.77
2.85
6.76
6.80
-13.4
-6.81
1.23
11.52
-47.7
-33.99
-17.72
5.28
6.28
4.53
12.19
14.24
-10.4
-0.15
4.70
13.89
-40.5
-39.5
-2.41
20.58
-5.94
-1.06
13.05
22.90
-14.1
0.57
14.92
19.32
-43.2
-211
-19.2
19.15
2.90
3.85
11.10
8.58
-12.4
-7.00
5.48
19.03
-53.6
-41.40
-9.13
15.50
12.49
6.63
17.08
17.82
-13.2
4.88
13.07
20.15
-22.9
-32.8
6.25
23.76
-4.82
2.71
17.86
27.92
-8.85
4.51
21.20
24.56
-40.2
-193
-8.18
25.34
5.92
6.40
16.63
12.01
-9.49
-6.14
11.31
26.76
-56.5
-42.97
2.24
26.77
16.79
7.98
22.41
21.01
-7.28
15.14
20.29
22.59
-13.0
-19.7
10.86
23.59
4.47
6.77
24.37
31.76
-3.27
10.89
26.80
28.14
-33.6
-170
4.11
27.60
10.19
10.46
24.23
15.88
-3.23
-3.73
18.64
36.09
-54.0
-43.71
15.68
37.09
25.42
9.86
27.01
24.95
2.71
26.34
26.86
26.92
0.95
-3.23
11.61
29.52
15.03
12.69
31.46
34.33
2.18
18.86
29.51
31.81
-21.1
-145
21.46
30.60
16.48
17.16
32.68
20.83
4.40
5.07
26.66
43.38
-46.3
-34.32
29.46
45.17
31.12
12.90
32.01
31.43
14.92
35.09
29.88
32.59
17.45
14.54
14.75
35.78
24.13
19.83
40.15
39.21
8.03
28.30
32.29
35.50
-1.83
-114
30.03
37.55
23.25
23.97
42.26
26.71
14.82
15.15
38.47
52.00
-32.8
-19.33
42.34
49.99
39.92
14.57
37.51
39.21
30.73
42.32
36.24
36.35
43.44
32.78
21.22
44.53
34.73
25.47
48.61
46.96
19.33
37.99
37.65
42.48
20.52
-83.5
40.26
45.40
29.68
31.43
50.68
36.26
24.93
25.04
48.89
59.58
-19.7
-2.61
53.36
54.76
48.74
31.41
43.89
43.84
38.25
47.88
43.74
43.33
68.44
50.97
25.20
52.47
46.29
29.62
58.23
53.68
39.05
45.95
43.49
50.19
44.57
-49.2
46.14
53.75
36.81
38.57
59.24
45.87
33.64
35.69
59.07
67.16
-5.08
14.95
63.86
62.62
Experiment number
catalyst
2
CO2 conversion (%)
4
6
8
50.0
10
30.0
12
14
15.0
16
0
18
20
-15.0
22
-30.0
24
26
28
30
32
34
36
680 710 740 770 800 830 860 890 920 950 980 1010
Temperature (K)
Figure 6.4. CO2 conversion values for each experiment and temperature.
This different behaviour is probably related with the methane availability. CO 2, 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 CO 2 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
217
Chapter 6
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 H 2/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
Experiment number
methane mitigates this adverse effect.
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
680 710 740 770 800 830 860 890 920 950 980 1010
H2/CO molar ratio
4.00
3.00
2.50
2.00
1.50
1.00
Temperature (K)
Figure 6.5. H2/CO molar ratio values for each experiment and temperature.
218
catalyst
Table 6.5. H2/CO molar ratio values obtained for each experiment.
Exp.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Temperature
680
710
740
10.02 7.90 5.92
2.26
5.28 2.82 2.05
37.21 13.46 9.82
21.53 12.95 8.71
9.11 4.77
- 27.57 9.20
7.23 3.13
34.76
- 35.41 10.06
- 20.50 6.85
9.03 5.58
14.45 9.36 6.38
9.53 4.75
12.48 4.08 2.55
- 30.45 10.08
- 109.80 19.67
7.13 3.82
9.25 4.14
57.38
9.53 4.75
52.56
- 59.52 8.22
- 35.83 12.12
22.98
7.68 3.59
27.35
129.85
8.85
- 23.54 5.42
18.58
32.49
770
800
830
860
890
920
950
980
1010
4.79
1.05
1.81
1.67
6.98
5.65
3.70
5.38
10.56
18.09
6.25
4.13
3.97
5.29
3.05
1.87
9.48
10.32
2.75
2.73
21.55
3.05
11.57
4.58
7.49
11.35
2.43
7.65
12.71
19.89
4.95
3.12
71.99
8.16
6.85
4.49
0.94
1.64
1.60
5.18
4.38
2.86
4.27
6.46
11.30
4.34
3.31
5.14
4.01
2.37
1.56
6.39
7.01
2.23
2.19
12.38
2.37
6.61
3.45
5.30
7.53
1.99
3.98
8.13
10.67
3.51
2.20
22.41
41.38
5.05
4.21
3.35
1.28
1.56
1.53
4.17
3.41
2.40
3.19
6.79
7.60
3.59
3.20
4.10
3.71
1.98
1.44
4.79
5.02
1.96
1.96
8.11
1.98
4.62
3.00
4.11
5.28
1.70
2.98
5.74
7.16
2.76
1.86
12.86
17.12
3.71
2.99
2.71
3.31
1.51
1.48
4.02
2.88
2.14
2.60
3.96
5.41
3.25
3.19
3.61
2.98
1.76
1.37
3.74
3.83
1.83
1.85
5.83
1.76
3.51
2.82
3.32
3.90
1.51
2.22
4.36
5.40
2.32
1.66
8.67
10.01
3.04
2.58
2.25
2.71
1.46
1.36
3.37
2.43
2.06
2.52
3.48
4.17
3.19
2.99
2.50
2.62
1.58
1.32
3.18
3.13
1.73
1.75
4.61
1.58
2.97
2.71
2.78
3.14
1.39
1.90
3.51
4.37
2.03
1.52
6.27
6.92
2.60
2.41
2.07
2.38
1.37
1.26
2.87
2.15
1.96
2.35
3.07
3.47
3.09
2.31
2.07
2.34
1.50
1.25
2.88
2.69
1.74
1.63
3.71
1.50
2.61
2.36
2.37
2.56
1.31
1.66
2.98
3.32
1.86
1.48
4.83
5.03
2.40
2.27
1.95
2.25
1.28
1.21
2.50
2.05
1.85
2.06
2.77
3.07
2.61
2.03
1.92
2.12
1.41
1.17
2.71
2.39
1.64
1.46
3.07
1.41
2.54
2.07
2.11
2.26
1.24
1.54
2.57
2.77
1.67
1.40
3.96
3.91
2.23
2.07
1.82
2.05
1.27
1.22
2.22
1.99
1.70
1.91
2.45
2.82
2.37
1.91
1.77
2.04
1.35
1.12
2.47
2.20
1.55
1.38
2.68
1.35
2.35
1.96
1.95
2.06
1.21
1.38
2.33
2.44
1.59
1.33
3.45
3.34
2.11
1.93
1.74
1.89
1.32
1.33
2.18
2.03
1.66
1.79
2.68
2.62
2.57
1.87
1.66
2.01
1.31
1.12
2.15
2.13
1.57
1.36
2.44
1.31
2.21
1.92
1.82
1.91
1.19
1.30
2.23
2.22
1.55
1.33
3.12
2.99
2.02
1.85
219
Chapter 6
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),
constants for each reaction,
and
is the partial pressure of species j (in kPa) and
are the kinetic
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 trireforming 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]:
220
catalyst
(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, CO 2 global consumption was
not observed in the low temperature range, obtaining higher CO 2 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 CO 2 concentration
as a result of it stoichiometry (Equation 6.10).
221
Chapter 6
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.
222
19.4
10.8
WGS
𝑜
3
𝑎3
= 54.26 kJ mol
-1
= 149.92 mol s-1 kPa-1
3.11/1.96
5.69·1013/1.96
7036.95/2.1 0.33
𝑎2
= 77.82 kJ mol-1
26.28/1.96
= 70.99 mol s-1 kPa-1 4.99·1013/1.96
𝑜
2
DR
SR
Kinetic equation
𝑜
1
𝑎1
= 74.72 kJ mol-1
47.43/1.96
= 85.77 mol s-1 kPa-1 3.46·1014/1.96
F/F-test
T/T-test
Parameters
Table 6.7. Kinetic model results.
 CH4 error (%) CO2 error (%)
catalyst
Figure 6 shows the fit obtained for the CH 4 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
223
Chapter 6
experimental ones, which could indicate a lower contribution of the water gas shift reaction to
the global process.
0.14
0.14
a)
b)
0.12
-1
Modelled molar flow (mol h )
-1
0.12
0.10
0.08
0.06
0.04
0.02
0.00
0.00
0.02
0.04
0.06
0.08
0.10
0.12
-1
Experimental molar flow (mol h )
0.14
0.10
0.08
0.06
0.04
0.02
0.00
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
-1
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 H 2O
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.
224
WGS
DR
SR
Kinetic equation
T/T-test
2.52·1013/1.96
6.00/1.96
= 26.71 mol s-1 kPa-1
= 68.67 kJ mol-1
1.92·1013/1.96
20.44/1.96
= 87.06 mol s-1 kPa-1
= 77.73 kJ mol-1
201.86/1.96
= 74.29 kJ mol-1
13
= 69.40 mol s-1 kPa-1 8.16·10 /1.96
Parameters
10.8
14.4
 CH4 error (%) CO2 error (%)
5639.82/2.1 0.28
F/F-test
Table 6.8. Second kinetic model
results.
catalyst
225
Chapter 6
0.14
0.14
a)
b)
0.12
-1
-1
0.12
0.10
0.08
0.06
0.04
0.02
0.00
0.00
0.02
0.04
0.06
0.08
0.10
0.12
-1
0.14
0.10
0.08
0.06
0.04
0.02
0.00
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
-1
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 CO 2 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
226
catalyst
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.
227
Chapter 6
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) 7582.
[7] N. Laosiripojana, S. Assabumrungrat, Applied Catalysis B: Environmental. 60 (2005) 107116.
[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) 3844.
[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.
228
catalyst
[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.
229
Chapter 6
[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.
230
CHAPTER 7
General conclusions and recommendations
7. 1. GENERAL CONCLUSIONS
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
233
Chapter 7
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 CO 2
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.
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.
234
 To perform tri-reforming experiments using actual natural gas or biogas as feed.
 To evaluate the pressure influence in both methane conversion and H 2/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.
235
LIST OF PUBLICATIONS AND
CONFERENCES
List of publications and conferences
Publications
 Precursor influence and catalytic behaviour of Ni/CeO 2 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. GómezMonedero, 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).
239
List of publications and conferences
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. GarciaVargas, 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.
240

METHANE TRI-REFORMING OVER NICKEL CATALYSTS

Transcription

Similar documents

CABERNET FRANC 2007

escapulario?

Nivel 2 - Aspectos Culturales

Board Members - Evolution Women

Monumentos - Hotel Lope de Vega

FALL 2006 149 II Festival Internacional de Artes

Boletín de novedades, 15 de octubre de 2014